linux-stable/kernel/trace/trace_events_hist.c

6860 lines
168 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
/*
* trace_events_hist - trace event hist triggers
*
* Copyright (C) 2015 Tom Zanussi <tom.zanussi@linux.intel.com>
*/
#include <linux/module.h>
#include <linux/kallsyms.h>
#include <linux/security.h>
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
#include <linux/mutex.h>
#include <linux/slab.h>
#include <linux/stacktrace.h>
#include <linux/rculist.h>
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
#include <linux/tracefs.h>
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
/* for gfp flag names */
#include <linux/trace_events.h>
#include <trace/events/mmflags.h>
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
#include "tracing_map.h"
#include "trace_synth.h"
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
#define ERRORS \
C(NONE, "No error"), \
C(DUPLICATE_VAR, "Variable already defined"), \
C(VAR_NOT_UNIQUE, "Variable name not unique, need to use fully qualified name (subsys.event.var) for variable"), \
C(TOO_MANY_VARS, "Too many variables defined"), \
C(MALFORMED_ASSIGNMENT, "Malformed assignment"), \
C(NAMED_MISMATCH, "Named hist trigger doesn't match existing named trigger (includes variables)"), \
C(TRIGGER_EEXIST, "Hist trigger already exists"), \
C(TRIGGER_ENOENT_CLEAR, "Can't clear or continue a nonexistent hist trigger"), \
C(SET_CLOCK_FAIL, "Couldn't set trace_clock"), \
C(BAD_FIELD_MODIFIER, "Invalid field modifier"), \
C(TOO_MANY_SUBEXPR, "Too many subexpressions (3 max)"), \
C(TIMESTAMP_MISMATCH, "Timestamp units in expression don't match"), \
C(TOO_MANY_FIELD_VARS, "Too many field variables defined"), \
C(EVENT_FILE_NOT_FOUND, "Event file not found"), \
C(HIST_NOT_FOUND, "Matching event histogram not found"), \
C(HIST_CREATE_FAIL, "Couldn't create histogram for field"), \
C(SYNTH_VAR_NOT_FOUND, "Couldn't find synthetic variable"), \
C(SYNTH_EVENT_NOT_FOUND,"Couldn't find synthetic event"), \
C(SYNTH_TYPE_MISMATCH, "Param type doesn't match synthetic event field type"), \
C(SYNTH_COUNT_MISMATCH, "Param count doesn't match synthetic event field count"), \
C(FIELD_VAR_PARSE_FAIL, "Couldn't parse field variable"), \
C(VAR_CREATE_FIND_FAIL, "Couldn't create or find variable"), \
C(ONX_NOT_VAR, "For onmax(x) or onchange(x), x must be a variable"), \
C(ONX_VAR_NOT_FOUND, "Couldn't find onmax or onchange variable"), \
C(ONX_VAR_CREATE_FAIL, "Couldn't create onmax or onchange variable"), \
C(FIELD_VAR_CREATE_FAIL,"Couldn't create field variable"), \
C(TOO_MANY_PARAMS, "Too many action params"), \
C(PARAM_NOT_FOUND, "Couldn't find param"), \
C(INVALID_PARAM, "Invalid action param"), \
C(ACTION_NOT_FOUND, "No action found"), \
C(NO_SAVE_PARAMS, "No params found for save()"), \
C(TOO_MANY_SAVE_ACTIONS,"Can't have more than one save() action per hist"), \
C(ACTION_MISMATCH, "Handler doesn't support action"), \
C(NO_CLOSING_PAREN, "No closing paren found"), \
C(SUBSYS_NOT_FOUND, "Missing subsystem"), \
C(INVALID_SUBSYS_EVENT, "Invalid subsystem or event name"), \
C(INVALID_REF_KEY, "Using variable references in keys not supported"), \
C(VAR_NOT_FOUND, "Couldn't find variable"), \
C(FIELD_NOT_FOUND, "Couldn't find field"), \
C(EMPTY_ASSIGNMENT, "Empty assignment"), \
C(INVALID_SORT_MODIFIER,"Invalid sort modifier"), \
C(EMPTY_SORT_FIELD, "Empty sort field"), \
C(TOO_MANY_SORT_FIELDS, "Too many sort fields (Max = 2)"), \
C(INVALID_SORT_FIELD, "Sort field must be a key or a val"), \
tracing: Add support for creating hist trigger variables from literal Currently hist trigger expressions don't support the use of numeric literals: e.g. echo 'hist:keys=common_pid:x=$y-1234' --> is not valid expression syntax Having the ability to use numeric constants in hist triggers supports a wider range of expressions for creating variables. Add support for creating trace event histogram variables from numeric literals. e.g. echo 'hist:keys=common_pid:x=1234,y=size-1024' >> event/trigger A negative numeric constant is created, using unary minus operator (parentheses are required). e.g. echo 'hist:keys=common_pid:z=-(2)' >> event/trigger Constants can be used with division/multiplication (added in the next patch in this series) to implement granularity filters for frequent trace events. For instance we can limit emitting the rss_stat trace event to when there is a 512KB cross over in the rss size: # Create a synthetic event to monitor instead of the high frequency # rss_stat event echo 'rss_stat_throttled unsigned int mm_id; unsigned int curr; int member; long size' >> tracing/synthetic_events # Create a hist trigger that emits the synthetic rss_stat_throttled # event only when the rss size crosses a 512KB boundary. echo 'hist:keys=keys=mm_id,member:bucket=size/0x80000:onchange($bucket) .rss_stat_throttled(mm_id,curr,member,size)' >> events/kmem/rss_stat/trigger A use case for using constants with addition/subtraction is not yet known, but for completeness the use of constants are supported for all operators. Link: https://lkml.kernel.org/r/20211025200852.3002369-2-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-10-25 20:08:33 +00:00
C(INVALID_STR_OPERAND, "String type can not be an operand in expression"), \
C(EXPECT_NUMBER, "Expecting numeric literal"), \
C(UNARY_MINUS_SUBEXPR, "Unary minus not supported in sub-expressions"), \
C(DIVISION_BY_ZERO, "Division by zero"), \
C(NEED_NOHC_VAL, "Non-hitcount value is required for 'nohitcount'"),
#undef C
#define C(a, b) HIST_ERR_##a
enum { ERRORS };
#undef C
#define C(a, b) b
static const char *err_text[] = { ERRORS };
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
struct hist_field;
typedef u64 (*hist_field_fn_t) (struct hist_field *field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
#define HIST_FIELD_OPERANDS_MAX 2
#define HIST_FIELDS_MAX (TRACING_MAP_FIELDS_MAX + TRACING_MAP_VARS_MAX)
#define HIST_ACTIONS_MAX 8
tracing: Add support for creating hist trigger variables from literal Currently hist trigger expressions don't support the use of numeric literals: e.g. echo 'hist:keys=common_pid:x=$y-1234' --> is not valid expression syntax Having the ability to use numeric constants in hist triggers supports a wider range of expressions for creating variables. Add support for creating trace event histogram variables from numeric literals. e.g. echo 'hist:keys=common_pid:x=1234,y=size-1024' >> event/trigger A negative numeric constant is created, using unary minus operator (parentheses are required). e.g. echo 'hist:keys=common_pid:z=-(2)' >> event/trigger Constants can be used with division/multiplication (added in the next patch in this series) to implement granularity filters for frequent trace events. For instance we can limit emitting the rss_stat trace event to when there is a 512KB cross over in the rss size: # Create a synthetic event to monitor instead of the high frequency # rss_stat event echo 'rss_stat_throttled unsigned int mm_id; unsigned int curr; int member; long size' >> tracing/synthetic_events # Create a hist trigger that emits the synthetic rss_stat_throttled # event only when the rss size crosses a 512KB boundary. echo 'hist:keys=keys=mm_id,member:bucket=size/0x80000:onchange($bucket) .rss_stat_throttled(mm_id,curr,member,size)' >> events/kmem/rss_stat/trigger A use case for using constants with addition/subtraction is not yet known, but for completeness the use of constants are supported for all operators. Link: https://lkml.kernel.org/r/20211025200852.3002369-2-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-10-25 20:08:33 +00:00
#define HIST_CONST_DIGITS_MAX 21
#define HIST_DIV_SHIFT 20 /* For optimizing division by constants */
enum field_op_id {
FIELD_OP_NONE,
FIELD_OP_PLUS,
FIELD_OP_MINUS,
FIELD_OP_UNARY_MINUS,
FIELD_OP_DIV,
FIELD_OP_MULT,
};
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
enum hist_field_fn {
HIST_FIELD_FN_NOP,
HIST_FIELD_FN_VAR_REF,
HIST_FIELD_FN_COUNTER,
HIST_FIELD_FN_CONST,
HIST_FIELD_FN_LOG2,
HIST_FIELD_FN_BUCKET,
HIST_FIELD_FN_TIMESTAMP,
HIST_FIELD_FN_CPU,
HIST_FIELD_FN_STRING,
HIST_FIELD_FN_DYNSTRING,
HIST_FIELD_FN_RELDYNSTRING,
HIST_FIELD_FN_PSTRING,
HIST_FIELD_FN_S64,
HIST_FIELD_FN_U64,
HIST_FIELD_FN_S32,
HIST_FIELD_FN_U32,
HIST_FIELD_FN_S16,
HIST_FIELD_FN_U16,
HIST_FIELD_FN_S8,
HIST_FIELD_FN_U8,
HIST_FIELD_FN_UMINUS,
HIST_FIELD_FN_MINUS,
HIST_FIELD_FN_PLUS,
HIST_FIELD_FN_DIV,
HIST_FIELD_FN_MULT,
HIST_FIELD_FN_DIV_POWER2,
HIST_FIELD_FN_DIV_NOT_POWER2,
HIST_FIELD_FN_DIV_MULT_SHIFT,
HIST_FIELD_FN_EXECNAME,
tracing/histogram: Fix stacktrace key The current code will always use the current stacktrace as a key even if a stacktrace contained in a specific event field was specified. For example, we expect to use the 'unsigned long[] stack' field in the below event in the histogram: # echo 's:block_lat pid_t pid; u64 delta; unsigned long[] stack;' > /sys/kernel/debug/tracing/dynamic_events # echo 'hist:keys=delta.buckets=100,stack.stacktrace:sort=delta' > /sys/kernel/debug/tracing/events/synthetic/block_lat/trigger But in fact, when we type out the trigger, we see that it's using the plain old global 'stacktrace' as the key, which is just the stacktrace when the event was hit and not the stacktrace contained in the event, which is what we want: # cat /sys/kernel/debug/tracing/events/synthetic/block_lat/trigger hist:keys=delta.buckets=100,stacktrace:vals=hitcount:sort=delta.buckets=100:size=2048 [active] And in fact, there's no code to actually retrieve it from the event, so we need to add HIST_FIELD_FN_STACK and hist_field_stack() to get it and hook it into the trigger code. For now, since the stack is just using dynamic strings, this could just use the dynamic string function, but it seems cleaner to have a dedicated function an be able to tweak independently as necessary. Link: https://lkml.kernel.org/r/11aa614c82976adbfa4ea763dbe885b5fb01d59c.1676063532.git.zanussi@kernel.org Signed-off-by: Tom Zanussi <zanussi@kernel.org> [ Fixed 32bit build warning reported by kernel test robot <lkp@intel.com> ] Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2023-02-10 21:33:05 +00:00
HIST_FIELD_FN_STACK,
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
};
/*
* A hist_var (histogram variable) contains variable information for
* hist_fields having the HIST_FIELD_FL_VAR or HIST_FIELD_FL_VAR_REF
* flag set. A hist_var has a variable name e.g. ts0, and is
* associated with a given histogram trigger, as specified by
* hist_data. The hist_var idx is the unique index assigned to the
* variable by the hist trigger's tracing_map. The idx is what is
* used to set a variable's value and, by a variable reference, to
* retrieve it.
*/
struct hist_var {
char *name;
struct hist_trigger_data *hist_data;
unsigned int idx;
};
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
struct hist_field {
struct ftrace_event_field *field;
unsigned long flags;
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
unsigned long buckets;
const char *type;
struct hist_field *operands[HIST_FIELD_OPERANDS_MAX];
struct hist_trigger_data *hist_data;
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
enum hist_field_fn fn_num;
unsigned int ref;
unsigned int size;
unsigned int offset;
unsigned int is_signed;
/*
* Variable fields contain variable-specific info in var.
*/
struct hist_var var;
enum field_op_id operator;
char *system;
char *event_name;
/*
* The name field is used for EXPR and VAR_REF fields. VAR
* fields contain the variable name in var.name.
*/
char *name;
/*
* When a histogram trigger is hit, if it has any references
* to variables, the values of those variables are collected
* into a var_ref_vals array by resolve_var_refs(). The
* current value of each variable is read from the tracing_map
* using the hist field's hist_var.idx and entered into the
* var_ref_idx entry i.e. var_ref_vals[var_ref_idx].
*/
unsigned int var_ref_idx;
bool read_once;
unsigned int var_str_idx;
tracing: Add support for creating hist trigger variables from literal Currently hist trigger expressions don't support the use of numeric literals: e.g. echo 'hist:keys=common_pid:x=$y-1234' --> is not valid expression syntax Having the ability to use numeric constants in hist triggers supports a wider range of expressions for creating variables. Add support for creating trace event histogram variables from numeric literals. e.g. echo 'hist:keys=common_pid:x=1234,y=size-1024' >> event/trigger A negative numeric constant is created, using unary minus operator (parentheses are required). e.g. echo 'hist:keys=common_pid:z=-(2)' >> event/trigger Constants can be used with division/multiplication (added in the next patch in this series) to implement granularity filters for frequent trace events. For instance we can limit emitting the rss_stat trace event to when there is a 512KB cross over in the rss size: # Create a synthetic event to monitor instead of the high frequency # rss_stat event echo 'rss_stat_throttled unsigned int mm_id; unsigned int curr; int member; long size' >> tracing/synthetic_events # Create a hist trigger that emits the synthetic rss_stat_throttled # event only when the rss size crosses a 512KB boundary. echo 'hist:keys=keys=mm_id,member:bucket=size/0x80000:onchange($bucket) .rss_stat_throttled(mm_id,curr,member,size)' >> events/kmem/rss_stat/trigger A use case for using constants with addition/subtraction is not yet known, but for completeness the use of constants are supported for all operators. Link: https://lkml.kernel.org/r/20211025200852.3002369-2-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-10-25 20:08:33 +00:00
/* Numeric literals are represented as u64 */
u64 constant;
/* Used to optimize division by constants */
u64 div_multiplier;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
};
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
static u64 hist_fn_call(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event);
tracing: Add support for creating hist trigger variables from literal Currently hist trigger expressions don't support the use of numeric literals: e.g. echo 'hist:keys=common_pid:x=$y-1234' --> is not valid expression syntax Having the ability to use numeric constants in hist triggers supports a wider range of expressions for creating variables. Add support for creating trace event histogram variables from numeric literals. e.g. echo 'hist:keys=common_pid:x=1234,y=size-1024' >> event/trigger A negative numeric constant is created, using unary minus operator (parentheses are required). e.g. echo 'hist:keys=common_pid:z=-(2)' >> event/trigger Constants can be used with division/multiplication (added in the next patch in this series) to implement granularity filters for frequent trace events. For instance we can limit emitting the rss_stat trace event to when there is a 512KB cross over in the rss size: # Create a synthetic event to monitor instead of the high frequency # rss_stat event echo 'rss_stat_throttled unsigned int mm_id; unsigned int curr; int member; long size' >> tracing/synthetic_events # Create a hist trigger that emits the synthetic rss_stat_throttled # event only when the rss size crosses a 512KB boundary. echo 'hist:keys=keys=mm_id,member:bucket=size/0x80000:onchange($bucket) .rss_stat_throttled(mm_id,curr,member,size)' >> events/kmem/rss_stat/trigger A use case for using constants with addition/subtraction is not yet known, but for completeness the use of constants are supported for all operators. Link: https://lkml.kernel.org/r/20211025200852.3002369-2-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-10-25 20:08:33 +00:00
static u64 hist_field_const(struct hist_field *field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
return field->constant;
}
static u64 hist_field_counter(struct hist_field *field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
{
return 1;
}
static u64 hist_field_string(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
{
char *addr = (char *)(event + hist_field->field->offset);
return (u64)(unsigned long)addr;
}
static u64 hist_field_dynstring(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
u32 str_item = *(u32 *)(event + hist_field->field->offset);
int str_loc = str_item & 0xffff;
char *addr = (char *)(event + str_loc);
return (u64)(unsigned long)addr;
}
tracing: Support __rel_loc relative dynamic data location attribute Add '__rel_loc' new dynamic data location attribute which encodes the data location from the next to the field itself. The '__data_loc' is used for encoding the dynamic data location on the trace event record. But '__data_loc' is not useful if the writer doesn't know the event header (e.g. user event), because it records the dynamic data offset from the entry of the record, not the field itself. This new '__rel_loc' attribute encodes the data location relatively from the next of the field. For example, when there is a record like below (the number in the parentheses is the size of fields) |header(N)|common(M)|fields(K)|__data_loc(4)|fields(L)|data(G)| In this case, '__data_loc' field will be __data_loc = (G << 16) | (N+M+K+4+L) If '__rel_loc' is used, this will be |header(N)|common(M)|fields(K)|__rel_loc(4)|fields(L)|data(G)| where __rel_loc = (G << 16) | (L) This case shows L bytes after the '__rel_loc' attribute field, if there is no fields after the __rel_loc field, L must be 0. This is relatively easy (and no need to consider the kernel header change) when the event data fields are composed by user who doesn't know header and common fields. Link: https://lkml.kernel.org/r/163757341258.510314.4214431827833229956.stgit@devnote2 Cc: Beau Belgrave <beaub@linux.microsoft.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-11-22 09:30:12 +00:00
static u64 hist_field_reldynstring(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
u32 *item = event + hist_field->field->offset;
u32 str_item = *item;
int str_loc = str_item & 0xffff;
char *addr = (char *)&item[1] + str_loc;
return (u64)(unsigned long)addr;
}
static u64 hist_field_pstring(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
char **addr = (char **)(event + hist_field->field->offset);
return (u64)(unsigned long)*addr;
}
static u64 hist_field_log2(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
tracing: Add hist trigger 'log2' modifier Allow users to have numeric fields displayed as log2 values in case value range is very wide by appending '.log2' to field names. For example, # echo 'hist:key=bytes_req' > kmalloc/trigger # cat kmalloc/hist { bytes_req: 504 } hitcount: 1 { bytes_req: 11 } hitcount: 1 { bytes_req: 104 } hitcount: 1 { bytes_req: 48 } hitcount: 1 { bytes_req: 2048 } hitcount: 1 { bytes_req: 4096 } hitcount: 1 { bytes_req: 240 } hitcount: 1 { bytes_req: 392 } hitcount: 1 { bytes_req: 13 } hitcount: 1 { bytes_req: 28 } hitcount: 1 { bytes_req: 12 } hitcount: 1 { bytes_req: 64 } hitcount: 2 { bytes_req: 128 } hitcount: 2 { bytes_req: 32 } hitcount: 2 { bytes_req: 8 } hitcount: 11 { bytes_req: 10 } hitcount: 13 { bytes_req: 24 } hitcount: 25 { bytes_req: 160 } hitcount: 29 { bytes_req: 16 } hitcount: 33 { bytes_req: 80 } hitcount: 36 When using '.log2' modifier, the output looks like: # echo 'hist:key=bytes_req.log2' > kmalloc/trigger # cat kmalloc/hist { bytes_req: ~ 2^12 } hitcount: 1 { bytes_req: ~ 2^11 } hitcount: 1 { bytes_req: ~ 2^9 } hitcount: 2 { bytes_req: ~ 2^6 } hitcount: 3 { bytes_req: ~ 2^3 } hitcount: 13 { bytes_req: ~ 2^5 } hitcount: 19 { bytes_req: ~ 2^8 } hitcount: 49 { bytes_req: ~ 2^7 } hitcount: 57 { bytes_req: ~ 2^4 } hitcount: 74 Link: http://lkml.kernel.org/r/7ff396b246c6a881f46b979735fddf05a0d6c71a.1457029949.git.tom.zanussi@linux.intel.com Cc: Tom Zanussi <tom.zanussi@linux.intel.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Signed-off-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:55:02 +00:00
{
struct hist_field *operand = hist_field->operands[0];
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
u64 val = hist_fn_call(operand, elt, buffer, rbe, event);
tracing: Add hist trigger 'log2' modifier Allow users to have numeric fields displayed as log2 values in case value range is very wide by appending '.log2' to field names. For example, # echo 'hist:key=bytes_req' > kmalloc/trigger # cat kmalloc/hist { bytes_req: 504 } hitcount: 1 { bytes_req: 11 } hitcount: 1 { bytes_req: 104 } hitcount: 1 { bytes_req: 48 } hitcount: 1 { bytes_req: 2048 } hitcount: 1 { bytes_req: 4096 } hitcount: 1 { bytes_req: 240 } hitcount: 1 { bytes_req: 392 } hitcount: 1 { bytes_req: 13 } hitcount: 1 { bytes_req: 28 } hitcount: 1 { bytes_req: 12 } hitcount: 1 { bytes_req: 64 } hitcount: 2 { bytes_req: 128 } hitcount: 2 { bytes_req: 32 } hitcount: 2 { bytes_req: 8 } hitcount: 11 { bytes_req: 10 } hitcount: 13 { bytes_req: 24 } hitcount: 25 { bytes_req: 160 } hitcount: 29 { bytes_req: 16 } hitcount: 33 { bytes_req: 80 } hitcount: 36 When using '.log2' modifier, the output looks like: # echo 'hist:key=bytes_req.log2' > kmalloc/trigger # cat kmalloc/hist { bytes_req: ~ 2^12 } hitcount: 1 { bytes_req: ~ 2^11 } hitcount: 1 { bytes_req: ~ 2^9 } hitcount: 2 { bytes_req: ~ 2^6 } hitcount: 3 { bytes_req: ~ 2^3 } hitcount: 13 { bytes_req: ~ 2^5 } hitcount: 19 { bytes_req: ~ 2^8 } hitcount: 49 { bytes_req: ~ 2^7 } hitcount: 57 { bytes_req: ~ 2^4 } hitcount: 74 Link: http://lkml.kernel.org/r/7ff396b246c6a881f46b979735fddf05a0d6c71a.1457029949.git.tom.zanussi@linux.intel.com Cc: Tom Zanussi <tom.zanussi@linux.intel.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Signed-off-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:55:02 +00:00
return (u64) ilog2(roundup_pow_of_two(val));
}
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
static u64 hist_field_bucket(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
struct hist_field *operand = hist_field->operands[0];
unsigned long buckets = hist_field->buckets;
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
u64 val = hist_fn_call(operand, elt, buffer, rbe, event);
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
if (WARN_ON_ONCE(!buckets))
return val;
if (val >= LONG_MAX)
val = div64_ul(val, buckets);
else
val = (u64)((unsigned long)val / buckets);
return val * buckets;
}
static u64 hist_field_plus(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
struct hist_field *operand1 = hist_field->operands[0];
struct hist_field *operand2 = hist_field->operands[1];
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
u64 val1 = hist_fn_call(operand1, elt, buffer, rbe, event);
u64 val2 = hist_fn_call(operand2, elt, buffer, rbe, event);
return val1 + val2;
}
static u64 hist_field_minus(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
struct hist_field *operand1 = hist_field->operands[0];
struct hist_field *operand2 = hist_field->operands[1];
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
u64 val1 = hist_fn_call(operand1, elt, buffer, rbe, event);
u64 val2 = hist_fn_call(operand2, elt, buffer, rbe, event);
return val1 - val2;
}
static u64 hist_field_div(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
struct hist_field *operand1 = hist_field->operands[0];
struct hist_field *operand2 = hist_field->operands[1];
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
u64 val1 = hist_fn_call(operand1, elt, buffer, rbe, event);
u64 val2 = hist_fn_call(operand2, elt, buffer, rbe, event);
/* Return -1 for the undefined case */
if (!val2)
return -1;
tracing/histogram: Optimize division by a power of 2 The division is a slow operation. If the divisor is a power of 2, use a shift instead. Results were obtained using Android's version of perf (simpleperf[1]) as described below: 1. hist_field_div() is modified to call 2 test functions: test_hist_field_div_[not]_optimized(); passing them the same args. Use noinline and volatile to ensure these are not optimized out by the compiler. 2. Create a hist event trigger that uses division: events/kmem/rss_stat$ echo 'hist:keys=common_pid:x=size/<divisor>' >> trigger events/kmem/rss_stat$ echo 'hist:keys=common_pid:vals=$x' >> trigger 3. Run Android's lmkd_test[2] to generate rss_stat events, and record CPU samples with Android's simpleperf: simpleperf record -a --exclude-perf --post-unwind=yes -m 16384 -g -f 2000 -o perf.data == Results == Divisor is a power of 2 (divisor == 32): test_hist_field_div_not_optimized | 8,717,091 cpu-cycles test_hist_field_div_optimized | 1,643,137 cpu-cycles If the divisor is a power of 2, the optimized version is ~5.3x faster. Divisor is not a power of 2 (divisor == 33): test_hist_field_div_not_optimized | 4,444,324 cpu-cycles test_hist_field_div_optimized | 5,497,958 cpu-cycles If the divisor is not a power of 2, as expected, the optimized version is slightly slower (~24% slower). [1] https://android.googlesource.com/platform/system/extras/+/master/simpleperf/doc/README.md [2] https://cs.android.com/android/platform/superproject/+/master:system/memory/lmkd/tests/lmkd_test.cpp Link: https://lkml.kernel.org/r/20211025200852.3002369-7-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Suggested-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-10-25 20:08:38 +00:00
/* Use shift if the divisor is a power of 2 */
if (!(val2 & (val2 - 1)))
return val1 >> __ffs64(val2);
return div64_u64(val1, val2);
}
static u64 div_by_power_of_two(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
struct hist_field *operand1 = hist_field->operands[0];
struct hist_field *operand2 = hist_field->operands[1];
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
u64 val1 = hist_fn_call(operand1, elt, buffer, rbe, event);
return val1 >> __ffs64(operand2->constant);
}
static u64 div_by_not_power_of_two(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
struct hist_field *operand1 = hist_field->operands[0];
struct hist_field *operand2 = hist_field->operands[1];
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
u64 val1 = hist_fn_call(operand1, elt, buffer, rbe, event);
return div64_u64(val1, operand2->constant);
}
static u64 div_by_mult_and_shift(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
struct hist_field *operand1 = hist_field->operands[0];
struct hist_field *operand2 = hist_field->operands[1];
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
u64 val1 = hist_fn_call(operand1, elt, buffer, rbe, event);
/*
* If the divisor is a constant, do a multiplication and shift instead.
*
* Choose Z = some power of 2. If Y <= Z, then:
* X / Y = (X * (Z / Y)) / Z
*
* (Z / Y) is a constant (mult) which is calculated at parse time, so:
* X / Y = (X * mult) / Z
*
* The division by Z can be replaced by a shift since Z is a power of 2:
* X / Y = (X * mult) >> HIST_DIV_SHIFT
*
* As long, as X < Z the results will not be off by more than 1.
*/
if (val1 < (1 << HIST_DIV_SHIFT)) {
u64 mult = operand2->div_multiplier;
return (val1 * mult + ((1 << HIST_DIV_SHIFT) - 1)) >> HIST_DIV_SHIFT;
}
return div64_u64(val1, operand2->constant);
}
static u64 hist_field_mult(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
struct hist_field *operand1 = hist_field->operands[0];
struct hist_field *operand2 = hist_field->operands[1];
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
u64 val1 = hist_fn_call(operand1, elt, buffer, rbe, event);
u64 val2 = hist_fn_call(operand2, elt, buffer, rbe, event);
return val1 * val2;
}
static u64 hist_field_unary_minus(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
struct hist_field *operand = hist_field->operands[0];
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
s64 sval = (s64)hist_fn_call(operand, elt, buffer, rbe, event);
u64 val = (u64)-sval;
return val;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
#define DEFINE_HIST_FIELD_FN(type) \
static u64 hist_field_##type(struct hist_field *hist_field, \
struct tracing_map_elt *elt, \
struct trace_buffer *buffer, \
struct ring_buffer_event *rbe, \
void *event) \
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
{ \
type *addr = (type *)(event + hist_field->field->offset); \
\
return (u64)(unsigned long)*addr; \
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
DEFINE_HIST_FIELD_FN(s64);
DEFINE_HIST_FIELD_FN(u64);
DEFINE_HIST_FIELD_FN(s32);
DEFINE_HIST_FIELD_FN(u32);
DEFINE_HIST_FIELD_FN(s16);
DEFINE_HIST_FIELD_FN(u16);
DEFINE_HIST_FIELD_FN(s8);
DEFINE_HIST_FIELD_FN(u8);
#define for_each_hist_field(i, hist_data) \
for ((i) = 0; (i) < (hist_data)->n_fields; (i)++)
#define for_each_hist_val_field(i, hist_data) \
for ((i) = 0; (i) < (hist_data)->n_vals; (i)++)
#define for_each_hist_key_field(i, hist_data) \
for ((i) = (hist_data)->n_vals; (i) < (hist_data)->n_fields; (i)++)
#define HITCOUNT_IDX 0
#define HIST_KEY_SIZE_MAX (MAX_FILTER_STR_VAL + HIST_STACKTRACE_SIZE)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
enum hist_field_flags {
HIST_FIELD_FL_HITCOUNT = 1 << 0,
HIST_FIELD_FL_KEY = 1 << 1,
HIST_FIELD_FL_STRING = 1 << 2,
HIST_FIELD_FL_HEX = 1 << 3,
HIST_FIELD_FL_SYM = 1 << 4,
HIST_FIELD_FL_SYM_OFFSET = 1 << 5,
HIST_FIELD_FL_EXECNAME = 1 << 6,
HIST_FIELD_FL_SYSCALL = 1 << 7,
HIST_FIELD_FL_STACKTRACE = 1 << 8,
HIST_FIELD_FL_LOG2 = 1 << 9,
HIST_FIELD_FL_TIMESTAMP = 1 << 10,
HIST_FIELD_FL_TIMESTAMP_USECS = 1 << 11,
HIST_FIELD_FL_VAR = 1 << 12,
HIST_FIELD_FL_EXPR = 1 << 13,
HIST_FIELD_FL_VAR_REF = 1 << 14,
HIST_FIELD_FL_CPU = 1 << 15,
HIST_FIELD_FL_ALIAS = 1 << 16,
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
HIST_FIELD_FL_BUCKET = 1 << 17,
tracing: Add support for creating hist trigger variables from literal Currently hist trigger expressions don't support the use of numeric literals: e.g. echo 'hist:keys=common_pid:x=$y-1234' --> is not valid expression syntax Having the ability to use numeric constants in hist triggers supports a wider range of expressions for creating variables. Add support for creating trace event histogram variables from numeric literals. e.g. echo 'hist:keys=common_pid:x=1234,y=size-1024' >> event/trigger A negative numeric constant is created, using unary minus operator (parentheses are required). e.g. echo 'hist:keys=common_pid:z=-(2)' >> event/trigger Constants can be used with division/multiplication (added in the next patch in this series) to implement granularity filters for frequent trace events. For instance we can limit emitting the rss_stat trace event to when there is a 512KB cross over in the rss size: # Create a synthetic event to monitor instead of the high frequency # rss_stat event echo 'rss_stat_throttled unsigned int mm_id; unsigned int curr; int member; long size' >> tracing/synthetic_events # Create a hist trigger that emits the synthetic rss_stat_throttled # event only when the rss size crosses a 512KB boundary. echo 'hist:keys=keys=mm_id,member:bucket=size/0x80000:onchange($bucket) .rss_stat_throttled(mm_id,curr,member,size)' >> events/kmem/rss_stat/trigger A use case for using constants with addition/subtraction is not yet known, but for completeness the use of constants are supported for all operators. Link: https://lkml.kernel.org/r/20211025200852.3002369-2-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-10-25 20:08:33 +00:00
HIST_FIELD_FL_CONST = 1 << 18,
HIST_FIELD_FL_PERCENT = 1 << 19,
HIST_FIELD_FL_GRAPH = 1 << 20,
};
struct var_defs {
unsigned int n_vars;
char *name[TRACING_MAP_VARS_MAX];
char *expr[TRACING_MAP_VARS_MAX];
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
};
struct hist_trigger_attrs {
char *keys_str;
char *vals_str;
char *sort_key_str;
char *name;
char *clock;
bool pause;
bool cont;
bool clear;
bool ts_in_usecs;
bool no_hitcount;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
unsigned int map_bits;
char *assignment_str[TRACING_MAP_VARS_MAX];
unsigned int n_assignments;
char *action_str[HIST_ACTIONS_MAX];
unsigned int n_actions;
struct var_defs var_defs;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
};
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
struct field_var {
struct hist_field *var;
struct hist_field *val;
};
struct field_var_hist {
struct hist_trigger_data *hist_data;
char *cmd;
};
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
struct hist_trigger_data {
struct hist_field *fields[HIST_FIELDS_MAX];
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
unsigned int n_vals;
unsigned int n_keys;
unsigned int n_fields;
unsigned int n_vars;
unsigned int n_var_str;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
unsigned int key_size;
struct tracing_map_sort_key sort_keys[TRACING_MAP_SORT_KEYS_MAX];
unsigned int n_sort_keys;
struct trace_event_file *event_file;
struct hist_trigger_attrs *attrs;
struct tracing_map *map;
bool enable_timestamps;
bool remove;
struct hist_field *var_refs[TRACING_MAP_VARS_MAX];
unsigned int n_var_refs;
struct action_data *actions[HIST_ACTIONS_MAX];
unsigned int n_actions;
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
struct field_var *field_vars[SYNTH_FIELDS_MAX];
unsigned int n_field_vars;
unsigned int n_field_var_str;
struct field_var_hist *field_var_hists[SYNTH_FIELDS_MAX];
unsigned int n_field_var_hists;
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
struct field_var *save_vars[SYNTH_FIELDS_MAX];
unsigned int n_save_vars;
unsigned int n_save_var_str;
};
struct action_data;
typedef void (*action_fn_t) (struct hist_trigger_data *hist_data,
struct tracing_map_elt *elt,
struct trace_buffer *buffer, void *rec,
struct ring_buffer_event *rbe, void *key,
struct action_data *data, u64 *var_ref_vals);
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
typedef bool (*check_track_val_fn_t) (u64 track_val, u64 var_val);
enum handler_id {
HANDLER_ONMATCH = 1,
HANDLER_ONMAX,
HANDLER_ONCHANGE,
};
enum action_id {
ACTION_SAVE = 1,
ACTION_TRACE,
ACTION_SNAPSHOT,
};
struct action_data {
enum handler_id handler;
enum action_id action;
char *action_name;
action_fn_t fn;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
unsigned int n_params;
char *params[SYNTH_FIELDS_MAX];
/*
* When a histogram trigger is hit, the values of any
* references to variables, including variables being passed
* as parameters to synthetic events, are collected into a
* var_ref_vals array. This var_ref_idx array is an array of
* indices into the var_ref_vals array, one for each synthetic
* event param, and is passed to the synthetic event
* invocation.
*/
tracing/hist: Fix out-of-bound write on 'action_data.var_ref_idx' When generate a synthetic event with many params and then create a trace action for it [1], kernel panic happened [2]. It is because that in trace_action_create() 'data->n_params' is up to SYNTH_FIELDS_MAX (current value is 64), and array 'data->var_ref_idx' keeps indices into array 'hist_data->var_refs' for each synthetic event param, but the length of 'data->var_ref_idx' is TRACING_MAP_VARS_MAX (current value is 16), so out-of-bound write happened when 'data->n_params' more than 16. In this case, 'data->match_data.event' is overwritten and eventually cause the panic. To solve the issue, adjust the length of 'data->var_ref_idx' to be SYNTH_FIELDS_MAX and add sanity checks to avoid out-of-bound write. [1] # cd /sys/kernel/tracing/ # echo "my_synth_event int v1; int v2; int v3; int v4; int v5; int v6;\ int v7; int v8; int v9; int v10; int v11; int v12; int v13; int v14;\ int v15; int v16; int v17; int v18; int v19; int v20; int v21; int v22;\ int v23; int v24; int v25; int v26; int v27; int v28; int v29; int v30;\ int v31; int v32; int v33; int v34; int v35; int v36; int v37; int v38;\ int v39; int v40; int v41; int v42; int v43; int v44; int v45; int v46;\ int v47; int v48; int v49; int v50; int v51; int v52; int v53; int v54;\ int v55; int v56; int v57; int v58; int v59; int v60; int v61; int v62;\ int v63" >> synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="bash"' >> \ events/sched/sched_waking/trigger # echo "hist:keys=next_pid:onmatch(sched.sched_waking).my_synth_event(\ pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,\ pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,\ pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,\ pid,pid,pid,pid,pid,pid,pid,pid,pid)" >> events/sched/sched_switch/trigger [2] BUG: unable to handle page fault for address: ffff91c900000000 PGD 61001067 P4D 61001067 PUD 0 Oops: 0000 [#1] PREEMPT SMP NOPTI CPU: 2 PID: 322 Comm: bash Tainted: G W 6.1.0-rc8+ #229 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.15.0-0-g2dd4b9b3f840-prebuilt.qemu.org 04/01/2014 RIP: 0010:strcmp+0xc/0x30 Code: 75 f7 31 d2 44 0f b6 04 16 44 88 04 11 48 83 c2 01 45 84 c0 75 ee c3 cc cc cc cc 0f 1f 00 31 c0 eb 08 48 83 c0 01 84 d2 74 13 <0f> b6 14 07 3a 14 06 74 ef 19 c0 83 c8 01 c3 cc cc cc cc 31 c3 RSP: 0018:ffff9b3b00f53c48 EFLAGS: 00000246 RAX: 0000000000000000 RBX: ffffffffba958a68 RCX: 0000000000000000 RDX: 0000000000000010 RSI: ffff91c943d33a90 RDI: ffff91c900000000 RBP: ffff91c900000000 R08: 00000018d604b529 R09: 0000000000000000 R10: ffff91c9483eddb1 R11: ffff91ca483eddab R12: ffff91c946171580 R13: ffff91c9479f0538 R14: ffff91c9457c2848 R15: ffff91c9479f0538 FS: 00007f1d1cfbe740(0000) GS:ffff91c9bdc80000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffff91c900000000 CR3: 0000000006316000 CR4: 00000000000006e0 Call Trace: <TASK> __find_event_file+0x55/0x90 action_create+0x76c/0x1060 event_hist_trigger_parse+0x146d/0x2060 ? event_trigger_write+0x31/0xd0 trigger_process_regex+0xbb/0x110 event_trigger_write+0x6b/0xd0 vfs_write+0xc8/0x3e0 ? alloc_fd+0xc0/0x160 ? preempt_count_add+0x4d/0xa0 ? preempt_count_add+0x70/0xa0 ksys_write+0x5f/0xe0 do_syscall_64+0x3b/0x90 entry_SYSCALL_64_after_hwframe+0x63/0xcd RIP: 0033:0x7f1d1d0cf077 Code: 64 89 02 48 c7 c0 ff ff ff ff eb bb 0f 1f 80 00 00 00 00 f3 0f 1e fa 64 8b 04 25 18 00 00 00 85 c0 75 10 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 51 c3 48 83 ec 28 48 89 54 24 18 48 89 74 RSP: 002b:00007ffcebb0e568 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 RAX: ffffffffffffffda RBX: 0000000000000143 RCX: 00007f1d1d0cf077 RDX: 0000000000000143 RSI: 00005639265aa7e0 RDI: 0000000000000001 RBP: 00005639265aa7e0 R08: 000000000000000a R09: 0000000000000142 R10: 000056392639c017 R11: 0000000000000246 R12: 0000000000000143 R13: 00007f1d1d1ae6a0 R14: 00007f1d1d1aa4a0 R15: 00007f1d1d1a98a0 </TASK> Modules linked in: CR2: ffff91c900000000 ---[ end trace 0000000000000000 ]--- RIP: 0010:strcmp+0xc/0x30 Code: 75 f7 31 d2 44 0f b6 04 16 44 88 04 11 48 83 c2 01 45 84 c0 75 ee c3 cc cc cc cc 0f 1f 00 31 c0 eb 08 48 83 c0 01 84 d2 74 13 <0f> b6 14 07 3a 14 06 74 ef 19 c0 83 c8 01 c3 cc cc cc cc 31 c3 RSP: 0018:ffff9b3b00f53c48 EFLAGS: 00000246 RAX: 0000000000000000 RBX: ffffffffba958a68 RCX: 0000000000000000 RDX: 0000000000000010 RSI: ffff91c943d33a90 RDI: ffff91c900000000 RBP: ffff91c900000000 R08: 00000018d604b529 R09: 0000000000000000 R10: ffff91c9483eddb1 R11: ffff91ca483eddab R12: ffff91c946171580 R13: ffff91c9479f0538 R14: ffff91c9457c2848 R15: ffff91c9479f0538 FS: 00007f1d1cfbe740(0000) GS:ffff91c9bdc80000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffff91c900000000 CR3: 0000000006316000 CR4: 00000000000006e0 Link: https://lore.kernel.org/linux-trace-kernel/20221207035143.2278781-1-zhengyejian1@huawei.com Cc: <mhiramat@kernel.org> Cc: <zanussi@kernel.org> Cc: stable@vger.kernel.org Fixes: d380dcde9a07 ("tracing: Fix now invalid var_ref_vals assumption in trace action") Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-12-07 03:51:43 +00:00
unsigned int var_ref_idx[SYNTH_FIELDS_MAX];
struct synth_event *synth_event;
bool use_trace_keyword;
char *synth_event_name;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
union {
struct {
char *event;
char *event_system;
} match_data;
struct {
/*
* var_str contains the $-unstripped variable
* name referenced by var_ref, and used when
* printing the action. Because var_ref
* creation is deferred to create_actions(),
* we need a per-action way to save it until
* then, thus var_str.
*/
char *var_str;
/*
* var_ref refers to the variable being
* tracked e.g onmax($var).
*/
struct hist_field *var_ref;
/*
* track_var contains the 'invisible' tracking
* variable created to keep the current
* e.g. max value.
*/
struct hist_field *track_var;
check_track_val_fn_t check_val;
action_fn_t save_data;
} track_data;
};
};
struct track_data {
u64 track_val;
bool updated;
unsigned int key_len;
void *key;
struct tracing_map_elt elt;
struct action_data *action_data;
struct hist_trigger_data *hist_data;
};
struct hist_elt_data {
char *comm;
u64 *var_ref_vals;
char **field_var_str;
int n_field_var_str;
};
struct snapshot_context {
struct tracing_map_elt *elt;
void *key;
};
/*
* Returns the specific division function to use if the divisor
* is constant. This avoids extra branches when the trigger is hit.
*/
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
static enum hist_field_fn hist_field_get_div_fn(struct hist_field *divisor)
{
u64 div = divisor->constant;
if (!(div & (div - 1)))
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
return HIST_FIELD_FN_DIV_POWER2;
/* If the divisor is too large, do a regular division */
if (div > (1 << HIST_DIV_SHIFT))
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
return HIST_FIELD_FN_DIV_NOT_POWER2;
divisor->div_multiplier = div64_u64((u64)(1 << HIST_DIV_SHIFT), div);
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
return HIST_FIELD_FN_DIV_MULT_SHIFT;
}
static void track_data_free(struct track_data *track_data)
{
struct hist_elt_data *elt_data;
if (!track_data)
return;
kfree(track_data->key);
elt_data = track_data->elt.private_data;
if (elt_data) {
kfree(elt_data->comm);
kfree(elt_data);
}
kfree(track_data);
}
static struct track_data *track_data_alloc(unsigned int key_len,
struct action_data *action_data,
struct hist_trigger_data *hist_data)
{
struct track_data *data = kzalloc(sizeof(*data), GFP_KERNEL);
struct hist_elt_data *elt_data;
if (!data)
return ERR_PTR(-ENOMEM);
data->key = kzalloc(key_len, GFP_KERNEL);
if (!data->key) {
track_data_free(data);
return ERR_PTR(-ENOMEM);
}
data->key_len = key_len;
data->action_data = action_data;
data->hist_data = hist_data;
elt_data = kzalloc(sizeof(*elt_data), GFP_KERNEL);
if (!elt_data) {
track_data_free(data);
return ERR_PTR(-ENOMEM);
}
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
data->elt.private_data = elt_data;
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
elt_data->comm = kzalloc(TASK_COMM_LEN, GFP_KERNEL);
if (!elt_data->comm) {
track_data_free(data);
return ERR_PTR(-ENOMEM);
}
return data;
}
#define HIST_PREFIX "hist:"
static char *last_cmd;
static char last_cmd_loc[MAX_FILTER_STR_VAL];
static int errpos(char *str)
{
if (!str || !last_cmd)
return 0;
return err_pos(last_cmd, str);
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
}
static void last_cmd_set(struct trace_event_file *file, char *str)
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
{
const char *system = NULL, *name = NULL;
struct trace_event_call *call;
int len;
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
if (!str)
return;
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
/* sizeof() contains the nul byte */
len = sizeof(HIST_PREFIX) + strlen(str);
kfree(last_cmd);
last_cmd = kzalloc(len, GFP_KERNEL);
if (!last_cmd)
return;
strcpy(last_cmd, HIST_PREFIX);
/* Again, sizeof() contains the nul byte */
len -= sizeof(HIST_PREFIX);
strncat(last_cmd, str, len);
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
if (file) {
call = file->event_call;
system = call->class->system;
if (system) {
name = trace_event_name(call);
if (!name)
system = NULL;
}
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
}
if (system)
snprintf(last_cmd_loc, MAX_FILTER_STR_VAL, HIST_PREFIX "%s:%s", system, name);
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
}
static void hist_err(struct trace_array *tr, u8 err_type, u16 err_pos)
{
if (!last_cmd)
return;
tracing_log_err(tr, last_cmd_loc, last_cmd, err_text,
err_type, err_pos);
}
static void hist_err_clear(void)
{
if (last_cmd)
last_cmd[0] = '\0';
last_cmd_loc[0] = '\0';
}
typedef void (*synth_probe_func_t) (void *__data, u64 *var_ref_vals,
unsigned int *var_ref_idx);
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
static inline void trace_synth(struct synth_event *event, u64 *var_ref_vals,
unsigned int *var_ref_idx)
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
{
struct tracepoint *tp = event->tp;
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
if (unlikely(atomic_read(&tp->key.enabled) > 0)) {
struct tracepoint_func *probe_func_ptr;
synth_probe_func_t probe_func;
void *__data;
if (!(cpu_online(raw_smp_processor_id())))
return;
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
probe_func_ptr = rcu_dereference_sched((tp)->funcs);
if (probe_func_ptr) {
do {
probe_func = probe_func_ptr->func;
__data = probe_func_ptr->data;
probe_func(__data, var_ref_vals, var_ref_idx);
} while ((++probe_func_ptr)->func);
}
}
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
}
static void action_trace(struct hist_trigger_data *hist_data,
struct tracing_map_elt *elt,
struct trace_buffer *buffer, void *rec,
struct ring_buffer_event *rbe, void *key,
struct action_data *data, u64 *var_ref_vals)
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
{
struct synth_event *event = data->synth_event;
trace_synth(event, var_ref_vals, data->var_ref_idx);
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
}
struct hist_var_data {
struct list_head list;
struct hist_trigger_data *hist_data;
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
};
static u64 hist_field_timestamp(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
struct hist_trigger_data *hist_data = hist_field->hist_data;
struct trace_array *tr = hist_data->event_file->tr;
u64 ts = ring_buffer_event_time_stamp(buffer, rbe);
if (hist_data->attrs->ts_in_usecs && trace_clock_in_ns(tr))
ts = ns2usecs(ts);
return ts;
}
static u64 hist_field_cpu(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
int cpu = smp_processor_id();
return cpu;
}
/**
* check_field_for_var_ref - Check if a VAR_REF field references a variable
* @hist_field: The VAR_REF field to check
* @var_data: The hist trigger that owns the variable
* @var_idx: The trigger variable identifier
*
* Check the given VAR_REF field to see whether or not it references
* the given variable associated with the given trigger.
*
* Return: The VAR_REF field if it does reference the variable, NULL if not
*/
static struct hist_field *
check_field_for_var_ref(struct hist_field *hist_field,
struct hist_trigger_data *var_data,
unsigned int var_idx)
{
WARN_ON(!(hist_field && hist_field->flags & HIST_FIELD_FL_VAR_REF));
if (hist_field && hist_field->var.idx == var_idx &&
hist_field->var.hist_data == var_data)
return hist_field;
return NULL;
}
/**
* find_var_ref - Check if a trigger has a reference to a trigger variable
* @hist_data: The hist trigger that might have a reference to the variable
* @var_data: The hist trigger that owns the variable
* @var_idx: The trigger variable identifier
*
* Check the list of var_refs[] on the first hist trigger to see
* whether any of them are references to the variable on the second
* trigger.
*
* Return: The VAR_REF field referencing the variable if so, NULL if not
*/
static struct hist_field *find_var_ref(struct hist_trigger_data *hist_data,
struct hist_trigger_data *var_data,
unsigned int var_idx)
{
struct hist_field *hist_field;
unsigned int i;
for (i = 0; i < hist_data->n_var_refs; i++) {
hist_field = hist_data->var_refs[i];
if (check_field_for_var_ref(hist_field, var_data, var_idx))
return hist_field;
}
return NULL;
}
/**
* find_any_var_ref - Check if there is a reference to a given trigger variable
* @hist_data: The hist trigger
* @var_idx: The trigger variable identifier
*
* Check to see whether the given variable is currently referenced by
* any other trigger.
*
* The trigger the variable is defined on is explicitly excluded - the
* assumption being that a self-reference doesn't prevent a trigger
* from being removed.
*
* Return: The VAR_REF field referencing the variable if so, NULL if not
*/
static struct hist_field *find_any_var_ref(struct hist_trigger_data *hist_data,
unsigned int var_idx)
{
struct trace_array *tr = hist_data->event_file->tr;
struct hist_field *found = NULL;
struct hist_var_data *var_data;
list_for_each_entry(var_data, &tr->hist_vars, list) {
if (var_data->hist_data == hist_data)
continue;
found = find_var_ref(var_data->hist_data, hist_data, var_idx);
if (found)
break;
}
return found;
}
/**
* check_var_refs - Check if there is a reference to any of trigger's variables
* @hist_data: The hist trigger
*
* A trigger can define one or more variables. If any one of them is
* currently referenced by any other trigger, this function will
* determine that.
*
* Typically used to determine whether or not a trigger can be removed
* - if there are any references to a trigger's variables, it cannot.
*
* Return: True if there is a reference to any of trigger's variables
*/
static bool check_var_refs(struct hist_trigger_data *hist_data)
{
struct hist_field *field;
bool found = false;
int i;
for_each_hist_field(i, hist_data) {
field = hist_data->fields[i];
if (field && field->flags & HIST_FIELD_FL_VAR) {
if (find_any_var_ref(hist_data, field->var.idx)) {
found = true;
break;
}
}
}
return found;
}
static struct hist_var_data *find_hist_vars(struct hist_trigger_data *hist_data)
{
struct trace_array *tr = hist_data->event_file->tr;
struct hist_var_data *var_data, *found = NULL;
list_for_each_entry(var_data, &tr->hist_vars, list) {
if (var_data->hist_data == hist_data) {
found = var_data;
break;
}
}
return found;
}
static bool field_has_hist_vars(struct hist_field *hist_field,
unsigned int level)
{
int i;
if (level > 3)
return false;
if (!hist_field)
return false;
if (hist_field->flags & HIST_FIELD_FL_VAR ||
hist_field->flags & HIST_FIELD_FL_VAR_REF)
return true;
for (i = 0; i < HIST_FIELD_OPERANDS_MAX; i++) {
struct hist_field *operand;
operand = hist_field->operands[i];
if (field_has_hist_vars(operand, level + 1))
return true;
}
return false;
}
static bool has_hist_vars(struct hist_trigger_data *hist_data)
{
struct hist_field *hist_field;
int i;
for_each_hist_field(i, hist_data) {
hist_field = hist_data->fields[i];
if (field_has_hist_vars(hist_field, 0))
return true;
}
return false;
}
static int save_hist_vars(struct hist_trigger_data *hist_data)
{
struct trace_array *tr = hist_data->event_file->tr;
struct hist_var_data *var_data;
var_data = find_hist_vars(hist_data);
if (var_data)
return 0;
if (tracing_check_open_get_tr(tr))
return -ENODEV;
var_data = kzalloc(sizeof(*var_data), GFP_KERNEL);
if (!var_data) {
trace_array_put(tr);
return -ENOMEM;
}
var_data->hist_data = hist_data;
list_add(&var_data->list, &tr->hist_vars);
return 0;
}
static void remove_hist_vars(struct hist_trigger_data *hist_data)
{
struct trace_array *tr = hist_data->event_file->tr;
struct hist_var_data *var_data;
var_data = find_hist_vars(hist_data);
if (!var_data)
return;
if (WARN_ON(check_var_refs(hist_data)))
return;
list_del(&var_data->list);
kfree(var_data);
trace_array_put(tr);
}
static struct hist_field *find_var_field(struct hist_trigger_data *hist_data,
const char *var_name)
{
struct hist_field *hist_field, *found = NULL;
int i;
for_each_hist_field(i, hist_data) {
hist_field = hist_data->fields[i];
if (hist_field && hist_field->flags & HIST_FIELD_FL_VAR &&
strcmp(hist_field->var.name, var_name) == 0) {
found = hist_field;
break;
}
}
return found;
}
static struct hist_field *find_var(struct hist_trigger_data *hist_data,
struct trace_event_file *file,
const char *var_name)
{
struct hist_trigger_data *test_data;
struct event_trigger_data *test;
struct hist_field *hist_field;
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
lockdep_assert_held(&event_mutex);
hist_field = find_var_field(hist_data, var_name);
if (hist_field)
return hist_field;
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
list_for_each_entry(test, &file->triggers, list) {
if (test->cmd_ops->trigger_type == ETT_EVENT_HIST) {
test_data = test->private_data;
hist_field = find_var_field(test_data, var_name);
if (hist_field)
return hist_field;
}
}
return NULL;
}
static struct trace_event_file *find_var_file(struct trace_array *tr,
char *system,
char *event_name,
char *var_name)
{
struct hist_trigger_data *var_hist_data;
struct hist_var_data *var_data;
struct trace_event_file *file, *found = NULL;
if (system)
return find_event_file(tr, system, event_name);
list_for_each_entry(var_data, &tr->hist_vars, list) {
var_hist_data = var_data->hist_data;
file = var_hist_data->event_file;
if (file == found)
continue;
if (find_var_field(var_hist_data, var_name)) {
if (found) {
hist_err(tr, HIST_ERR_VAR_NOT_UNIQUE, errpos(var_name));
return NULL;
}
found = file;
}
}
return found;
}
static struct hist_field *find_file_var(struct trace_event_file *file,
const char *var_name)
{
struct hist_trigger_data *test_data;
struct event_trigger_data *test;
struct hist_field *hist_field;
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
lockdep_assert_held(&event_mutex);
list_for_each_entry(test, &file->triggers, list) {
if (test->cmd_ops->trigger_type == ETT_EVENT_HIST) {
test_data = test->private_data;
hist_field = find_var_field(test_data, var_name);
if (hist_field)
return hist_field;
}
}
return NULL;
}
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
static struct hist_field *
find_match_var(struct hist_trigger_data *hist_data, char *var_name)
{
struct trace_array *tr = hist_data->event_file->tr;
struct hist_field *hist_field, *found = NULL;
struct trace_event_file *file;
unsigned int i;
for (i = 0; i < hist_data->n_actions; i++) {
struct action_data *data = hist_data->actions[i];
if (data->handler == HANDLER_ONMATCH) {
char *system = data->match_data.event_system;
char *event_name = data->match_data.event;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
file = find_var_file(tr, system, event_name, var_name);
if (!file)
continue;
hist_field = find_file_var(file, var_name);
if (hist_field) {
if (found) {
hist_err(tr, HIST_ERR_VAR_NOT_UNIQUE,
errpos(var_name));
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
return ERR_PTR(-EINVAL);
}
found = hist_field;
}
}
}
return found;
}
static struct hist_field *find_event_var(struct hist_trigger_data *hist_data,
char *system,
char *event_name,
char *var_name)
{
struct trace_array *tr = hist_data->event_file->tr;
struct hist_field *hist_field = NULL;
struct trace_event_file *file;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
if (!system || !event_name) {
hist_field = find_match_var(hist_data, var_name);
if (IS_ERR(hist_field))
return NULL;
if (hist_field)
return hist_field;
}
file = find_var_file(tr, system, event_name, var_name);
if (!file)
return NULL;
hist_field = find_file_var(file, var_name);
return hist_field;
}
static u64 hist_field_var_ref(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
struct hist_elt_data *elt_data;
u64 var_val = 0;
if (WARN_ON_ONCE(!elt))
return var_val;
elt_data = elt->private_data;
var_val = elt_data->var_ref_vals[hist_field->var_ref_idx];
return var_val;
}
static bool resolve_var_refs(struct hist_trigger_data *hist_data, void *key,
u64 *var_ref_vals, bool self)
{
struct hist_trigger_data *var_data;
struct tracing_map_elt *var_elt;
struct hist_field *hist_field;
unsigned int i, var_idx;
bool resolved = true;
u64 var_val = 0;
for (i = 0; i < hist_data->n_var_refs; i++) {
hist_field = hist_data->var_refs[i];
var_idx = hist_field->var.idx;
var_data = hist_field->var.hist_data;
if (var_data == NULL) {
resolved = false;
break;
}
if ((self && var_data != hist_data) ||
(!self && var_data == hist_data))
continue;
var_elt = tracing_map_lookup(var_data->map, key);
if (!var_elt) {
resolved = false;
break;
}
if (!tracing_map_var_set(var_elt, var_idx)) {
resolved = false;
break;
}
if (self || !hist_field->read_once)
var_val = tracing_map_read_var(var_elt, var_idx);
else
var_val = tracing_map_read_var_once(var_elt, var_idx);
var_ref_vals[i] = var_val;
}
return resolved;
}
static const char *hist_field_name(struct hist_field *field,
unsigned int level)
{
const char *field_name = "";
if (WARN_ON_ONCE(!field))
return field_name;
if (level > 1)
return field_name;
if (field->field)
field_name = field->field->name;
else if (field->flags & HIST_FIELD_FL_LOG2 ||
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
field->flags & HIST_FIELD_FL_ALIAS ||
field->flags & HIST_FIELD_FL_BUCKET)
field_name = hist_field_name(field->operands[0], ++level);
else if (field->flags & HIST_FIELD_FL_CPU)
tracing/histogram: Rename "cpu" to "common_cpu" Currently the histogram logic allows the user to write "cpu" in as an event field, and it will record the CPU that the event happened on. The problem with this is that there's a lot of events that have "cpu" as a real field, and using "cpu" as the CPU it ran on, makes it impossible to run histograms on the "cpu" field of events. For example, if I want to have a histogram on the count of the workqueue_queue_work event on its cpu field, running: ># echo 'hist:keys=cpu' > events/workqueue/workqueue_queue_work/trigger Gives a misleading and wrong result. Change the command to "common_cpu" as no event should have "common_*" fields as that's a reserved name for fields used by all events. And this makes sense here as common_cpu would be a field used by all events. Now we can even do: ># echo 'hist:keys=common_cpu,cpu if cpu < 100' > events/workqueue/workqueue_queue_work/trigger ># cat events/workqueue/workqueue_queue_work/hist # event histogram # # trigger info: hist:keys=common_cpu,cpu:vals=hitcount:sort=hitcount:size=2048 if cpu < 100 [active] # { common_cpu: 0, cpu: 2 } hitcount: 1 { common_cpu: 0, cpu: 4 } hitcount: 1 { common_cpu: 7, cpu: 7 } hitcount: 1 { common_cpu: 0, cpu: 7 } hitcount: 1 { common_cpu: 0, cpu: 1 } hitcount: 1 { common_cpu: 0, cpu: 6 } hitcount: 2 { common_cpu: 0, cpu: 5 } hitcount: 2 { common_cpu: 1, cpu: 1 } hitcount: 4 { common_cpu: 6, cpu: 6 } hitcount: 4 { common_cpu: 5, cpu: 5 } hitcount: 14 { common_cpu: 4, cpu: 4 } hitcount: 26 { common_cpu: 0, cpu: 0 } hitcount: 39 { common_cpu: 2, cpu: 2 } hitcount: 184 Now for backward compatibility, I added a trick. If "cpu" is used, and the field is not found, it will fall back to "common_cpu" and work as it did before. This way, it will still work for old programs that use "cpu" to get the actual CPU, but if the event has a "cpu" as a field, it will get that event's "cpu" field, which is probably what it wants anyway. I updated the tracefs/README to include documentation about both the common_timestamp and the common_cpu. This way, if that text is present in the README, then an application can know that common_cpu is supported over just plain "cpu". Link: https://lkml.kernel.org/r/20210721110053.26b4f641@oasis.local.home Cc: Namhyung Kim <namhyung@kernel.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: stable@vger.kernel.org Fixes: 8b7622bf94a44 ("tracing: Add cpu field for hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-21 15:00:53 +00:00
field_name = "common_cpu";
else if (field->flags & HIST_FIELD_FL_EXPR ||
field->flags & HIST_FIELD_FL_VAR_REF) {
if (field->system) {
static char full_name[MAX_FILTER_STR_VAL];
strcat(full_name, field->system);
strcat(full_name, ".");
strcat(full_name, field->event_name);
strcat(full_name, ".");
strcat(full_name, field->name);
field_name = full_name;
} else
field_name = field->name;
} else if (field->flags & HIST_FIELD_FL_TIMESTAMP)
field_name = "common_timestamp";
else if (field->flags & HIST_FIELD_FL_STACKTRACE) {
if (field->field)
field_name = field->field->name;
else
field_name = "common_stacktrace";
} else if (field->flags & HIST_FIELD_FL_HITCOUNT)
field_name = "hitcount";
if (field_name == NULL)
field_name = "";
return field_name;
}
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
static enum hist_field_fn select_value_fn(int field_size, int field_is_signed)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
{
switch (field_size) {
case 8:
if (field_is_signed)
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
return HIST_FIELD_FN_S64;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
else
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
return HIST_FIELD_FN_U64;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
case 4:
if (field_is_signed)
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
return HIST_FIELD_FN_S32;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
else
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
return HIST_FIELD_FN_U32;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
case 2:
if (field_is_signed)
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
return HIST_FIELD_FN_S16;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
else
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
return HIST_FIELD_FN_U16;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
case 1:
if (field_is_signed)
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
return HIST_FIELD_FN_S8;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
else
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
return HIST_FIELD_FN_U8;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
return HIST_FIELD_FN_NOP;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
static int parse_map_size(char *str)
{
unsigned long size, map_bits;
int ret;
ret = kstrtoul(str, 0, &size);
if (ret)
goto out;
map_bits = ilog2(roundup_pow_of_two(size));
if (map_bits < TRACING_MAP_BITS_MIN ||
map_bits > TRACING_MAP_BITS_MAX)
ret = -EINVAL;
else
ret = map_bits;
out:
return ret;
}
static void destroy_hist_trigger_attrs(struct hist_trigger_attrs *attrs)
{
unsigned int i;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (!attrs)
return;
for (i = 0; i < attrs->n_assignments; i++)
kfree(attrs->assignment_str[i]);
for (i = 0; i < attrs->n_actions; i++)
kfree(attrs->action_str[i]);
kfree(attrs->name);
kfree(attrs->sort_key_str);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
kfree(attrs->keys_str);
kfree(attrs->vals_str);
kfree(attrs->clock);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
kfree(attrs);
}
static int parse_action(char *str, struct hist_trigger_attrs *attrs)
{
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
int ret = -EINVAL;
if (attrs->n_actions >= HIST_ACTIONS_MAX)
return ret;
if ((str_has_prefix(str, "onmatch(")) ||
(str_has_prefix(str, "onmax(")) ||
(str_has_prefix(str, "onchange("))) {
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
attrs->action_str[attrs->n_actions] = kstrdup(str, GFP_KERNEL);
if (!attrs->action_str[attrs->n_actions]) {
ret = -ENOMEM;
return ret;
}
attrs->n_actions++;
ret = 0;
}
return ret;
}
static int parse_assignment(struct trace_array *tr,
char *str, struct hist_trigger_attrs *attrs)
{
int len, ret = 0;
if ((len = str_has_prefix(str, "key=")) ||
(len = str_has_prefix(str, "keys="))) {
attrs->keys_str = kstrdup(str + len, GFP_KERNEL);
if (!attrs->keys_str) {
ret = -ENOMEM;
goto out;
}
} else if ((len = str_has_prefix(str, "val=")) ||
(len = str_has_prefix(str, "vals=")) ||
(len = str_has_prefix(str, "values="))) {
attrs->vals_str = kstrdup(str + len, GFP_KERNEL);
if (!attrs->vals_str) {
ret = -ENOMEM;
goto out;
}
} else if ((len = str_has_prefix(str, "sort="))) {
attrs->sort_key_str = kstrdup(str + len, GFP_KERNEL);
if (!attrs->sort_key_str) {
ret = -ENOMEM;
goto out;
}
} else if (str_has_prefix(str, "name=")) {
attrs->name = kstrdup(str, GFP_KERNEL);
if (!attrs->name) {
ret = -ENOMEM;
goto out;
}
} else if ((len = str_has_prefix(str, "clock="))) {
str += len;
str = strstrip(str);
attrs->clock = kstrdup(str, GFP_KERNEL);
if (!attrs->clock) {
ret = -ENOMEM;
goto out;
}
} else if ((len = str_has_prefix(str, "size="))) {
int map_bits = parse_map_size(str + len);
if (map_bits < 0) {
ret = map_bits;
goto out;
}
attrs->map_bits = map_bits;
} else {
char *assignment;
if (attrs->n_assignments == TRACING_MAP_VARS_MAX) {
hist_err(tr, HIST_ERR_TOO_MANY_VARS, errpos(str));
ret = -EINVAL;
goto out;
}
assignment = kstrdup(str, GFP_KERNEL);
if (!assignment) {
ret = -ENOMEM;
goto out;
}
attrs->assignment_str[attrs->n_assignments++] = assignment;
}
out:
return ret;
}
static struct hist_trigger_attrs *
parse_hist_trigger_attrs(struct trace_array *tr, char *trigger_str)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
{
struct hist_trigger_attrs *attrs;
int ret = 0;
attrs = kzalloc(sizeof(*attrs), GFP_KERNEL);
if (!attrs)
return ERR_PTR(-ENOMEM);
while (trigger_str) {
char *str = strsep(&trigger_str, ":");
char *rhs;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
rhs = strchr(str, '=');
if (rhs) {
if (!strlen(++rhs)) {
ret = -EINVAL;
hist_err(tr, HIST_ERR_EMPTY_ASSIGNMENT, errpos(str));
goto free;
}
ret = parse_assignment(tr, str, attrs);
if (ret)
goto free;
} else if (strcmp(str, "nohitcount") == 0 ||
strcmp(str, "NOHC") == 0)
attrs->no_hitcount = true;
else if (strcmp(str, "pause") == 0)
attrs->pause = true;
else if ((strcmp(str, "cont") == 0) ||
(strcmp(str, "continue") == 0))
attrs->cont = true;
else if (strcmp(str, "clear") == 0)
attrs->clear = true;
else {
ret = parse_action(str, attrs);
if (ret)
goto free;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
}
if (!attrs->keys_str) {
ret = -EINVAL;
goto free;
}
if (!attrs->clock) {
attrs->clock = kstrdup("global", GFP_KERNEL);
if (!attrs->clock) {
ret = -ENOMEM;
goto free;
}
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
return attrs;
free:
destroy_hist_trigger_attrs(attrs);
return ERR_PTR(ret);
}
static inline void save_comm(char *comm, struct task_struct *task)
{
if (!task->pid) {
strcpy(comm, "<idle>");
return;
}
if (WARN_ON_ONCE(task->pid < 0)) {
strcpy(comm, "<XXX>");
return;
}
strncpy(comm, task->comm, TASK_COMM_LEN);
}
static void hist_elt_data_free(struct hist_elt_data *elt_data)
{
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
unsigned int i;
for (i = 0; i < elt_data->n_field_var_str; i++)
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
kfree(elt_data->field_var_str[i]);
kfree(elt_data->field_var_str);
kfree(elt_data->comm);
kfree(elt_data);
}
static void hist_trigger_elt_data_free(struct tracing_map_elt *elt)
{
struct hist_elt_data *elt_data = elt->private_data;
hist_elt_data_free(elt_data);
}
static int hist_trigger_elt_data_alloc(struct tracing_map_elt *elt)
{
struct hist_trigger_data *hist_data = elt->map->private_data;
unsigned int size = TASK_COMM_LEN;
struct hist_elt_data *elt_data;
tracing: Allow execnames to be passed as args for synthetic events Allow common_pid.execname to be saved in a variable in one histogram to be passed to another histogram that can pass it as a parameter to a synthetic event. ># echo 'hist:keys=pid:__arg__1=common_timestamp.usecs:arg2=common_pid.execname' \ > events/sched/sched_waking/trigger ># echo 'wakeup_lat s32 pid; u64 delta; char wake_comm[]' > synthetic_events ># echo 'hist:keys=next_pid:pid=next_pid,delta=common_timestamp.usecs-$__arg__1,exec=$arg2'\ ':onmatch(sched.sched_waking).trace(wakeup_lat,$pid,$delta,$exec)' \ > events/sched/sched_switch/trigger The above is a wake up latency synthetic event setup that passes the execname of the common_pid that woke the task to the scheduling of that task, which triggers a synthetic event that passes the original execname as a parameter to display it. ># echo 1 > events/synthetic/enable ># cat trace <idle>-0 [006] d..4 186.863801: wakeup_lat: pid=1306 delta=65 wake_comm=kworker/u16:3 <idle>-0 [000] d..4 186.863858: wakeup_lat: pid=163 delta=27 wake_comm=<idle> <idle>-0 [001] d..4 186.863903: wakeup_lat: pid=1307 delta=36 wake_comm=kworker/u16:4 <idle>-0 [000] d..4 186.863927: wakeup_lat: pid=163 delta=5 wake_comm=<idle> <idle>-0 [006] d..4 186.863957: wakeup_lat: pid=1306 delta=24 wake_comm=kworker/u16:3 sshd-1306 [006] d..4 186.864051: wakeup_lat: pid=61 delta=62 wake_comm=<idle> <idle>-0 [000] d..4 186.965030: wakeup_lat: pid=609 delta=18 wake_comm=<idle> <idle>-0 [006] d..4 186.987582: wakeup_lat: pid=1306 delta=65 wake_comm=kworker/u16:3 <idle>-0 [000] d..4 186.987639: wakeup_lat: pid=163 delta=27 wake_comm=<idle> Link: https://lkml.kernel.org/r/20210722142837.458596338@goodmis.org Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-22 14:27:07 +00:00
struct hist_field *hist_field;
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
unsigned int i, n_str;
elt_data = kzalloc(sizeof(*elt_data), GFP_KERNEL);
if (!elt_data)
return -ENOMEM;
tracing: Allow execnames to be passed as args for synthetic events Allow common_pid.execname to be saved in a variable in one histogram to be passed to another histogram that can pass it as a parameter to a synthetic event. ># echo 'hist:keys=pid:__arg__1=common_timestamp.usecs:arg2=common_pid.execname' \ > events/sched/sched_waking/trigger ># echo 'wakeup_lat s32 pid; u64 delta; char wake_comm[]' > synthetic_events ># echo 'hist:keys=next_pid:pid=next_pid,delta=common_timestamp.usecs-$__arg__1,exec=$arg2'\ ':onmatch(sched.sched_waking).trace(wakeup_lat,$pid,$delta,$exec)' \ > events/sched/sched_switch/trigger The above is a wake up latency synthetic event setup that passes the execname of the common_pid that woke the task to the scheduling of that task, which triggers a synthetic event that passes the original execname as a parameter to display it. ># echo 1 > events/synthetic/enable ># cat trace <idle>-0 [006] d..4 186.863801: wakeup_lat: pid=1306 delta=65 wake_comm=kworker/u16:3 <idle>-0 [000] d..4 186.863858: wakeup_lat: pid=163 delta=27 wake_comm=<idle> <idle>-0 [001] d..4 186.863903: wakeup_lat: pid=1307 delta=36 wake_comm=kworker/u16:4 <idle>-0 [000] d..4 186.863927: wakeup_lat: pid=163 delta=5 wake_comm=<idle> <idle>-0 [006] d..4 186.863957: wakeup_lat: pid=1306 delta=24 wake_comm=kworker/u16:3 sshd-1306 [006] d..4 186.864051: wakeup_lat: pid=61 delta=62 wake_comm=<idle> <idle>-0 [000] d..4 186.965030: wakeup_lat: pid=609 delta=18 wake_comm=<idle> <idle>-0 [006] d..4 186.987582: wakeup_lat: pid=1306 delta=65 wake_comm=kworker/u16:3 <idle>-0 [000] d..4 186.987639: wakeup_lat: pid=163 delta=27 wake_comm=<idle> Link: https://lkml.kernel.org/r/20210722142837.458596338@goodmis.org Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-22 14:27:07 +00:00
for_each_hist_field(i, hist_data) {
hist_field = hist_data->fields[i];
tracing: Allow execnames to be passed as args for synthetic events Allow common_pid.execname to be saved in a variable in one histogram to be passed to another histogram that can pass it as a parameter to a synthetic event. ># echo 'hist:keys=pid:__arg__1=common_timestamp.usecs:arg2=common_pid.execname' \ > events/sched/sched_waking/trigger ># echo 'wakeup_lat s32 pid; u64 delta; char wake_comm[]' > synthetic_events ># echo 'hist:keys=next_pid:pid=next_pid,delta=common_timestamp.usecs-$__arg__1,exec=$arg2'\ ':onmatch(sched.sched_waking).trace(wakeup_lat,$pid,$delta,$exec)' \ > events/sched/sched_switch/trigger The above is a wake up latency synthetic event setup that passes the execname of the common_pid that woke the task to the scheduling of that task, which triggers a synthetic event that passes the original execname as a parameter to display it. ># echo 1 > events/synthetic/enable ># cat trace <idle>-0 [006] d..4 186.863801: wakeup_lat: pid=1306 delta=65 wake_comm=kworker/u16:3 <idle>-0 [000] d..4 186.863858: wakeup_lat: pid=163 delta=27 wake_comm=<idle> <idle>-0 [001] d..4 186.863903: wakeup_lat: pid=1307 delta=36 wake_comm=kworker/u16:4 <idle>-0 [000] d..4 186.863927: wakeup_lat: pid=163 delta=5 wake_comm=<idle> <idle>-0 [006] d..4 186.863957: wakeup_lat: pid=1306 delta=24 wake_comm=kworker/u16:3 sshd-1306 [006] d..4 186.864051: wakeup_lat: pid=61 delta=62 wake_comm=<idle> <idle>-0 [000] d..4 186.965030: wakeup_lat: pid=609 delta=18 wake_comm=<idle> <idle>-0 [006] d..4 186.987582: wakeup_lat: pid=1306 delta=65 wake_comm=kworker/u16:3 <idle>-0 [000] d..4 186.987639: wakeup_lat: pid=163 delta=27 wake_comm=<idle> Link: https://lkml.kernel.org/r/20210722142837.458596338@goodmis.org Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-22 14:27:07 +00:00
if (hist_field->flags & HIST_FIELD_FL_EXECNAME) {
elt_data->comm = kzalloc(size, GFP_KERNEL);
if (!elt_data->comm) {
kfree(elt_data);
return -ENOMEM;
}
break;
}
}
n_str = hist_data->n_field_var_str + hist_data->n_save_var_str +
hist_data->n_var_str;
if (n_str > SYNTH_FIELDS_MAX) {
hist_elt_data_free(elt_data);
return -EINVAL;
}
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
BUILD_BUG_ON(STR_VAR_LEN_MAX & (sizeof(u64) - 1));
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
size = STR_VAR_LEN_MAX;
elt_data->field_var_str = kcalloc(n_str, sizeof(char *), GFP_KERNEL);
if (!elt_data->field_var_str) {
hist_elt_data_free(elt_data);
return -EINVAL;
}
elt_data->n_field_var_str = n_str;
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
for (i = 0; i < n_str; i++) {
elt_data->field_var_str[i] = kzalloc(size, GFP_KERNEL);
if (!elt_data->field_var_str[i]) {
hist_elt_data_free(elt_data);
return -ENOMEM;
}
}
elt->private_data = elt_data;
return 0;
}
static void hist_trigger_elt_data_init(struct tracing_map_elt *elt)
{
struct hist_elt_data *elt_data = elt->private_data;
if (elt_data->comm)
save_comm(elt_data->comm, current);
}
static const struct tracing_map_ops hist_trigger_elt_data_ops = {
.elt_alloc = hist_trigger_elt_data_alloc,
.elt_free = hist_trigger_elt_data_free,
.elt_init = hist_trigger_elt_data_init,
};
static const char *get_hist_field_flags(struct hist_field *hist_field)
{
const char *flags_str = NULL;
if (hist_field->flags & HIST_FIELD_FL_HEX)
flags_str = "hex";
else if (hist_field->flags & HIST_FIELD_FL_SYM)
flags_str = "sym";
else if (hist_field->flags & HIST_FIELD_FL_SYM_OFFSET)
flags_str = "sym-offset";
else if (hist_field->flags & HIST_FIELD_FL_EXECNAME)
flags_str = "execname";
else if (hist_field->flags & HIST_FIELD_FL_SYSCALL)
flags_str = "syscall";
else if (hist_field->flags & HIST_FIELD_FL_LOG2)
flags_str = "log2";
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
else if (hist_field->flags & HIST_FIELD_FL_BUCKET)
flags_str = "buckets";
else if (hist_field->flags & HIST_FIELD_FL_TIMESTAMP_USECS)
flags_str = "usecs";
else if (hist_field->flags & HIST_FIELD_FL_PERCENT)
flags_str = "percent";
else if (hist_field->flags & HIST_FIELD_FL_GRAPH)
flags_str = "graph";
tracing/histogram: Add stacktrace type Now that stacktraces can be part of synthetic events, allow a key to be typed as a stacktrace. # cd /sys/kernel/tracing # echo 's:block_lat u64 delta; unsigned long stack[];' >> dynamic_events # echo 'hist:keys=next_pid:ts=common_timestamp.usecs,st=stacktrace if prev_state == 2' >> events/sched/sched_switch/trigger # echo 'hist:keys=prev_pid:delta=common_timestamp.usecs-$ts,st2=$st:onmatch(sched.sched_switch).trace(block_lat,$delta,$st2)' >> events/sched/sched_switch/trigger # echo 'hist:keys=delta.buckets=100,stack.stacktrace:sort=delta' > events/synthetic/block_lat/trigger # cat events/synthetic/block_lat/hist # event histogram # # trigger info: hist:keys=delta.buckets=100,stacktrace:vals=hitcount:sort=delta.buckets=100:size=2048 [active] # { delta: ~ 0-99, stacktrace: event_hist_trigger+0x464/0x480 event_triggers_call+0x52/0xe0 trace_event_buffer_commit+0x193/0x250 trace_event_raw_event_sched_switch+0xfc/0x150 __traceiter_sched_switch+0x41/0x60 __schedule+0x448/0x7b0 schedule_idle+0x26/0x40 cpu_startup_entry+0x19/0x20 start_secondary+0xed/0xf0 secondary_startup_64_no_verify+0xe0/0xeb } hitcount: 6 { delta: ~ 0-99, stacktrace: event_hist_trigger+0x464/0x480 event_triggers_call+0x52/0xe0 trace_event_buffer_commit+0x193/0x250 trace_event_raw_event_sched_switch+0xfc/0x150 __traceiter_sched_switch+0x41/0x60 __schedule+0x448/0x7b0 schedule_idle+0x26/0x40 cpu_startup_entry+0x19/0x20 __pfx_kernel_init+0x0/0x10 arch_call_rest_init+0xa/0x24 start_kernel+0x964/0x98d secondary_startup_64_no_verify+0xe0/0xeb } hitcount: 3 { delta: ~ 0-99, stacktrace: event_hist_trigger+0x464/0x480 event_triggers_call+0x52/0xe0 trace_event_buffer_commit+0x193/0x250 trace_event_raw_event_sched_switch+0xfc/0x150 __traceiter_sched_switch+0x41/0x60 __schedule+0x448/0x7b0 schedule+0x5a/0xb0 worker_thread+0xaf/0x380 kthread+0xe9/0x110 ret_from_fork+0x2c/0x50 } hitcount: 1 { delta: ~ 100-199, stacktrace: event_hist_trigger+0x464/0x480 event_triggers_call+0x52/0xe0 trace_event_buffer_commit+0x193/0x250 trace_event_raw_event_sched_switch+0xfc/0x150 __traceiter_sched_switch+0x41/0x60 __schedule+0x448/0x7b0 schedule_idle+0x26/0x40 cpu_startup_entry+0x19/0x20 start_secondary+0xed/0xf0 secondary_startup_64_no_verify+0xe0/0xeb } hitcount: 15 [..] { delta: ~ 8500-8599, stacktrace: event_hist_trigger+0x464/0x480 event_triggers_call+0x52/0xe0 trace_event_buffer_commit+0x193/0x250 trace_event_raw_event_sched_switch+0xfc/0x150 __traceiter_sched_switch+0x41/0x60 __schedule+0x448/0x7b0 schedule_idle+0x26/0x40 cpu_startup_entry+0x19/0x20 start_secondary+0xed/0xf0 secondary_startup_64_no_verify+0xe0/0xeb } hitcount: 1 Totals: Hits: 89 Entries: 11 Dropped: 0 Link: https://lkml.kernel.org/r/20230117152236.167046397@goodmis.org Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Tom Zanussi <zanussi@kernel.org> Cc: Ross Zwisler <zwisler@google.com> Cc: Ching-lin Yu <chinglinyu@google.com> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2023-01-17 15:21:29 +00:00
else if (hist_field->flags & HIST_FIELD_FL_STACKTRACE)
flags_str = "stacktrace";
return flags_str;
}
static void expr_field_str(struct hist_field *field, char *expr)
{
if (field->flags & HIST_FIELD_FL_VAR_REF)
strcat(expr, "$");
tracing: Add support for creating hist trigger variables from literal Currently hist trigger expressions don't support the use of numeric literals: e.g. echo 'hist:keys=common_pid:x=$y-1234' --> is not valid expression syntax Having the ability to use numeric constants in hist triggers supports a wider range of expressions for creating variables. Add support for creating trace event histogram variables from numeric literals. e.g. echo 'hist:keys=common_pid:x=1234,y=size-1024' >> event/trigger A negative numeric constant is created, using unary minus operator (parentheses are required). e.g. echo 'hist:keys=common_pid:z=-(2)' >> event/trigger Constants can be used with division/multiplication (added in the next patch in this series) to implement granularity filters for frequent trace events. For instance we can limit emitting the rss_stat trace event to when there is a 512KB cross over in the rss size: # Create a synthetic event to monitor instead of the high frequency # rss_stat event echo 'rss_stat_throttled unsigned int mm_id; unsigned int curr; int member; long size' >> tracing/synthetic_events # Create a hist trigger that emits the synthetic rss_stat_throttled # event only when the rss size crosses a 512KB boundary. echo 'hist:keys=keys=mm_id,member:bucket=size/0x80000:onchange($bucket) .rss_stat_throttled(mm_id,curr,member,size)' >> events/kmem/rss_stat/trigger A use case for using constants with addition/subtraction is not yet known, but for completeness the use of constants are supported for all operators. Link: https://lkml.kernel.org/r/20211025200852.3002369-2-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-10-25 20:08:33 +00:00
else if (field->flags & HIST_FIELD_FL_CONST) {
char str[HIST_CONST_DIGITS_MAX];
snprintf(str, HIST_CONST_DIGITS_MAX, "%llu", field->constant);
strcat(expr, str);
}
strcat(expr, hist_field_name(field, 0));
if (field->flags && !(field->flags & HIST_FIELD_FL_VAR_REF)) {
const char *flags_str = get_hist_field_flags(field);
if (flags_str) {
strcat(expr, ".");
strcat(expr, flags_str);
}
}
}
static char *expr_str(struct hist_field *field, unsigned int level)
{
char *expr;
if (level > 1)
return NULL;
expr = kzalloc(MAX_FILTER_STR_VAL, GFP_KERNEL);
if (!expr)
return NULL;
if (!field->operands[0]) {
expr_field_str(field, expr);
return expr;
}
if (field->operator == FIELD_OP_UNARY_MINUS) {
char *subexpr;
strcat(expr, "-(");
subexpr = expr_str(field->operands[0], ++level);
if (!subexpr) {
kfree(expr);
return NULL;
}
strcat(expr, subexpr);
strcat(expr, ")");
kfree(subexpr);
return expr;
}
expr_field_str(field->operands[0], expr);
switch (field->operator) {
case FIELD_OP_MINUS:
strcat(expr, "-");
break;
case FIELD_OP_PLUS:
strcat(expr, "+");
break;
case FIELD_OP_DIV:
strcat(expr, "/");
break;
case FIELD_OP_MULT:
strcat(expr, "*");
break;
default:
kfree(expr);
return NULL;
}
expr_field_str(field->operands[1], expr);
return expr;
}
/*
* If field_op != FIELD_OP_NONE, *sep points to the root operator
* of the expression tree to be evaluated.
*/
static int contains_operator(char *str, char **sep)
{
enum field_op_id field_op = FIELD_OP_NONE;
char *minus_op, *plus_op, *div_op, *mult_op;
/*
* Report the last occurrence of the operators first, so that the
* expression is evaluated left to right. This is important since
* subtraction and division are not associative.
*
* e.g
* 64/8/4/2 is 1, i.e 64/8/4/2 = ((64/8)/4)/2
* 14-7-5-2 is 0, i.e 14-7-5-2 = ((14-7)-5)-2
*/
/*
* First, find lower precedence addition and subtraction
* since the expression will be evaluated recursively.
*/
minus_op = strrchr(str, '-');
if (minus_op) {
/*
* Unary minus is not supported in sub-expressions. If
* present, it is always the next root operator.
*/
if (minus_op == str) {
field_op = FIELD_OP_UNARY_MINUS;
goto out;
}
field_op = FIELD_OP_MINUS;
}
plus_op = strrchr(str, '+');
if (plus_op || minus_op) {
/*
* For operators of the same precedence use to rightmost as the
* root, so that the expression is evaluated left to right.
*/
if (plus_op > minus_op)
field_op = FIELD_OP_PLUS;
goto out;
}
/*
* Multiplication and division have higher precedence than addition and
* subtraction.
*/
div_op = strrchr(str, '/');
if (div_op)
field_op = FIELD_OP_DIV;
mult_op = strrchr(str, '*');
/*
* For operators of the same precedence use to rightmost as the
* root, so that the expression is evaluated left to right.
*/
if (mult_op > div_op)
field_op = FIELD_OP_MULT;
out:
if (sep) {
switch (field_op) {
case FIELD_OP_UNARY_MINUS:
case FIELD_OP_MINUS:
*sep = minus_op;
break;
case FIELD_OP_PLUS:
*sep = plus_op;
break;
case FIELD_OP_DIV:
*sep = div_op;
break;
case FIELD_OP_MULT:
*sep = mult_op;
break;
case FIELD_OP_NONE:
default:
*sep = NULL;
break;
}
}
return field_op;
}
tracing: Fix histogram code when expression has same var as value While working on a tool to convert SQL syntex into the histogram language of the kernel, I discovered the following bug: # echo 'first u64 start_time u64 end_time pid_t pid u64 delta' >> synthetic_events # echo 'hist:keys=pid:start=common_timestamp' > events/sched/sched_waking/trigger # echo 'hist:keys=next_pid:delta=common_timestamp-$start,start2=$start:onmatch(sched.sched_waking).trace(first,$start2,common_timestamp,next_pid,$delta)' > events/sched/sched_switch/trigger Would not display any histograms in the sched_switch histogram side. But if I were to swap the location of "delta=common_timestamp-$start" with "start2=$start" Such that the last line had: # echo 'hist:keys=next_pid:start2=$start,delta=common_timestamp-$start:onmatch(sched.sched_waking).trace(first,$start2,common_timestamp,next_pid,$delta)' > events/sched/sched_switch/trigger The histogram works as expected. What I found out is that the expressions clear out the value once it is resolved. As the variables are resolved in the order listed, when processing: delta=common_timestamp-$start The $start is cleared. When it gets to "start2=$start", it errors out with "unresolved symbol" (which is silent as this happens at the location of the trace), and the histogram is dropped. When processing the histogram for variable references, instead of adding a new reference for a variable used twice, use the same reference. That way, not only is it more efficient, but the order will no longer matter in processing of the variables. From Tom Zanussi: "Just to clarify some more about what the problem was is that without your patch, we would have two separate references to the same variable, and during resolve_var_refs(), they'd both want to be resolved separately, so in this case, since the first reference to start wasn't part of an expression, it wouldn't get the read-once flag set, so would be read normally, and then the second reference would do the read-once read and also be read but using read-once. So everything worked and you didn't see a problem: from: start2=$start,delta=common_timestamp-$start In the second case, when you switched them around, the first reference would be resolved by doing the read-once, and following that the second reference would try to resolve and see that the variable had already been read, so failed as unset, which caused it to short-circuit out and not do the trigger action to generate the synthetic event: to: delta=common_timestamp-$start,start2=$start With your patch, we only have the single resolution which happens correctly the one time it's resolved, so this can't happen." Link: https://lore.kernel.org/r/20200116154216.58ca08eb@gandalf.local.home Cc: stable@vger.kernel.org Fixes: 067fe038e70f6 ("tracing: Add variable reference handling to hist triggers") Reviewed-by: Tom Zanuss <zanussi@kernel.org> Tested-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2020-01-20 18:07:31 +00:00
static void get_hist_field(struct hist_field *hist_field)
{
hist_field->ref++;
}
static void __destroy_hist_field(struct hist_field *hist_field)
{
tracing: Fix histogram code when expression has same var as value While working on a tool to convert SQL syntex into the histogram language of the kernel, I discovered the following bug: # echo 'first u64 start_time u64 end_time pid_t pid u64 delta' >> synthetic_events # echo 'hist:keys=pid:start=common_timestamp' > events/sched/sched_waking/trigger # echo 'hist:keys=next_pid:delta=common_timestamp-$start,start2=$start:onmatch(sched.sched_waking).trace(first,$start2,common_timestamp,next_pid,$delta)' > events/sched/sched_switch/trigger Would not display any histograms in the sched_switch histogram side. But if I were to swap the location of "delta=common_timestamp-$start" with "start2=$start" Such that the last line had: # echo 'hist:keys=next_pid:start2=$start,delta=common_timestamp-$start:onmatch(sched.sched_waking).trace(first,$start2,common_timestamp,next_pid,$delta)' > events/sched/sched_switch/trigger The histogram works as expected. What I found out is that the expressions clear out the value once it is resolved. As the variables are resolved in the order listed, when processing: delta=common_timestamp-$start The $start is cleared. When it gets to "start2=$start", it errors out with "unresolved symbol" (which is silent as this happens at the location of the trace), and the histogram is dropped. When processing the histogram for variable references, instead of adding a new reference for a variable used twice, use the same reference. That way, not only is it more efficient, but the order will no longer matter in processing of the variables. From Tom Zanussi: "Just to clarify some more about what the problem was is that without your patch, we would have two separate references to the same variable, and during resolve_var_refs(), they'd both want to be resolved separately, so in this case, since the first reference to start wasn't part of an expression, it wouldn't get the read-once flag set, so would be read normally, and then the second reference would do the read-once read and also be read but using read-once. So everything worked and you didn't see a problem: from: start2=$start,delta=common_timestamp-$start In the second case, when you switched them around, the first reference would be resolved by doing the read-once, and following that the second reference would try to resolve and see that the variable had already been read, so failed as unset, which caused it to short-circuit out and not do the trigger action to generate the synthetic event: to: delta=common_timestamp-$start,start2=$start With your patch, we only have the single resolution which happens correctly the one time it's resolved, so this can't happen." Link: https://lore.kernel.org/r/20200116154216.58ca08eb@gandalf.local.home Cc: stable@vger.kernel.org Fixes: 067fe038e70f6 ("tracing: Add variable reference handling to hist triggers") Reviewed-by: Tom Zanuss <zanussi@kernel.org> Tested-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2020-01-20 18:07:31 +00:00
if (--hist_field->ref > 1)
return;
kfree(hist_field->var.name);
kfree(hist_field->name);
/* Can likely be a const */
kfree_const(hist_field->type);
tracing: Fix memory leaks in trace_events_hist.c kmemleak report 1: [<9092c50b>] kmem_cache_alloc_trace+0x138/0x270 [<05a2c9ed>] create_field_var+0xcf/0x180 [<528a2d68>] action_create+0xe2/0xc80 [<63f50b61>] event_hist_trigger_func+0x15b5/0x1920 [<28ea5d3d>] trigger_process_regex+0x7b/0xc0 [<3138e86f>] event_trigger_write+0x4d/0xb0 [<ffd66c19>] __vfs_write+0x30/0x200 [<4f424a0d>] vfs_write+0x96/0x1b0 [<da59a290>] ksys_write+0x53/0xc0 [<3717101a>] __ia32_sys_write+0x15/0x20 [<c5f23497>] do_fast_syscall_32+0x70/0x250 [<46e2629c>] entry_SYSENTER_32+0xaf/0x102 This is because save_vars[] of struct hist_trigger_data are not destroyed kmemleak report 2: [<9092c50b>] kmem_cache_alloc_trace+0x138/0x270 [<6e5e97c5>] create_var+0x3c/0x110 [<de82f1b9>] create_field_var+0xaf/0x180 [<528a2d68>] action_create+0xe2/0xc80 [<63f50b61>] event_hist_trigger_func+0x15b5/0x1920 [<28ea5d3d>] trigger_process_regex+0x7b/0xc0 [<3138e86f>] event_trigger_write+0x4d/0xb0 [<ffd66c19>] __vfs_write+0x30/0x200 [<4f424a0d>] vfs_write+0x96/0x1b0 [<da59a290>] ksys_write+0x53/0xc0 [<3717101a>] __ia32_sys_write+0x15/0x20 [<c5f23497>] do_fast_syscall_32+0x70/0x250 [<46e2629c>] entry_SYSENTER_32+0xaf/0x102 struct hist_field allocated through create_var() do not initialize "ref" field to 1. The code in __destroy_hist_field() does not destroy object if "ref" is initialized to zero, the condition if (--hist_field->ref > 1) always passes since unsigned int wraps. kmemleak report 3: [<f8666fcc>] __kmalloc_track_caller+0x139/0x2b0 [<bb7f80a5>] kstrdup+0x27/0x50 [<39d70006>] init_var_ref+0x58/0xd0 [<8ca76370>] create_var_ref+0x89/0xe0 [<f045fc39>] action_create+0x38f/0xc80 [<7c146821>] event_hist_trigger_func+0x15b5/0x1920 [<07de3f61>] trigger_process_regex+0x7b/0xc0 [<e87daf8f>] event_trigger_write+0x4d/0xb0 [<19bf1512>] __vfs_write+0x30/0x200 [<64ce4d27>] vfs_write+0x96/0x1b0 [<a6f34170>] ksys_write+0x53/0xc0 [<7d4230cd>] __ia32_sys_write+0x15/0x20 [<8eadca00>] do_fast_syscall_32+0x70/0x250 [<235cf985>] entry_SYSENTER_32+0xaf/0x102 hist_fields (system & event_name) are not freed Link: http://lkml.kernel.org/r/20200422061503.GA5151@cosmos Signed-off-by: Vamshi K Sthambamkadi <vamshi.k.sthambamkadi@gmail.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2020-04-22 06:15:06 +00:00
kfree(hist_field->system);
kfree(hist_field->event_name);
kfree(hist_field);
}
static void destroy_hist_field(struct hist_field *hist_field,
unsigned int level)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
{
unsigned int i;
if (level > 3)
return;
if (!hist_field)
return;
if (hist_field->flags & HIST_FIELD_FL_VAR_REF)
return; /* var refs will be destroyed separately */
for (i = 0; i < HIST_FIELD_OPERANDS_MAX; i++)
destroy_hist_field(hist_field->operands[i], level + 1);
__destroy_hist_field(hist_field);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
static struct hist_field *create_hist_field(struct hist_trigger_data *hist_data,
struct ftrace_event_field *field,
unsigned long flags,
char *var_name)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
{
struct hist_field *hist_field;
if (field && is_function_field(field))
return NULL;
hist_field = kzalloc(sizeof(struct hist_field), GFP_KERNEL);
if (!hist_field)
return NULL;
tracing: Fix histogram code when expression has same var as value While working on a tool to convert SQL syntex into the histogram language of the kernel, I discovered the following bug: # echo 'first u64 start_time u64 end_time pid_t pid u64 delta' >> synthetic_events # echo 'hist:keys=pid:start=common_timestamp' > events/sched/sched_waking/trigger # echo 'hist:keys=next_pid:delta=common_timestamp-$start,start2=$start:onmatch(sched.sched_waking).trace(first,$start2,common_timestamp,next_pid,$delta)' > events/sched/sched_switch/trigger Would not display any histograms in the sched_switch histogram side. But if I were to swap the location of "delta=common_timestamp-$start" with "start2=$start" Such that the last line had: # echo 'hist:keys=next_pid:start2=$start,delta=common_timestamp-$start:onmatch(sched.sched_waking).trace(first,$start2,common_timestamp,next_pid,$delta)' > events/sched/sched_switch/trigger The histogram works as expected. What I found out is that the expressions clear out the value once it is resolved. As the variables are resolved in the order listed, when processing: delta=common_timestamp-$start The $start is cleared. When it gets to "start2=$start", it errors out with "unresolved symbol" (which is silent as this happens at the location of the trace), and the histogram is dropped. When processing the histogram for variable references, instead of adding a new reference for a variable used twice, use the same reference. That way, not only is it more efficient, but the order will no longer matter in processing of the variables. From Tom Zanussi: "Just to clarify some more about what the problem was is that without your patch, we would have two separate references to the same variable, and during resolve_var_refs(), they'd both want to be resolved separately, so in this case, since the first reference to start wasn't part of an expression, it wouldn't get the read-once flag set, so would be read normally, and then the second reference would do the read-once read and also be read but using read-once. So everything worked and you didn't see a problem: from: start2=$start,delta=common_timestamp-$start In the second case, when you switched them around, the first reference would be resolved by doing the read-once, and following that the second reference would try to resolve and see that the variable had already been read, so failed as unset, which caused it to short-circuit out and not do the trigger action to generate the synthetic event: to: delta=common_timestamp-$start,start2=$start With your patch, we only have the single resolution which happens correctly the one time it's resolved, so this can't happen." Link: https://lore.kernel.org/r/20200116154216.58ca08eb@gandalf.local.home Cc: stable@vger.kernel.org Fixes: 067fe038e70f6 ("tracing: Add variable reference handling to hist triggers") Reviewed-by: Tom Zanuss <zanussi@kernel.org> Tested-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2020-01-20 18:07:31 +00:00
hist_field->ref = 1;
hist_field->hist_data = hist_data;
if (flags & HIST_FIELD_FL_EXPR || flags & HIST_FIELD_FL_ALIAS)
goto out; /* caller will populate */
if (flags & HIST_FIELD_FL_VAR_REF) {
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
hist_field->fn_num = HIST_FIELD_FN_VAR_REF;
goto out;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (flags & HIST_FIELD_FL_HITCOUNT) {
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
hist_field->fn_num = HIST_FIELD_FN_COUNTER;
hist_field->size = sizeof(u64);
hist_field->type = "u64";
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
goto out;
}
tracing: Add support for creating hist trigger variables from literal Currently hist trigger expressions don't support the use of numeric literals: e.g. echo 'hist:keys=common_pid:x=$y-1234' --> is not valid expression syntax Having the ability to use numeric constants in hist triggers supports a wider range of expressions for creating variables. Add support for creating trace event histogram variables from numeric literals. e.g. echo 'hist:keys=common_pid:x=1234,y=size-1024' >> event/trigger A negative numeric constant is created, using unary minus operator (parentheses are required). e.g. echo 'hist:keys=common_pid:z=-(2)' >> event/trigger Constants can be used with division/multiplication (added in the next patch in this series) to implement granularity filters for frequent trace events. For instance we can limit emitting the rss_stat trace event to when there is a 512KB cross over in the rss size: # Create a synthetic event to monitor instead of the high frequency # rss_stat event echo 'rss_stat_throttled unsigned int mm_id; unsigned int curr; int member; long size' >> tracing/synthetic_events # Create a hist trigger that emits the synthetic rss_stat_throttled # event only when the rss size crosses a 512KB boundary. echo 'hist:keys=keys=mm_id,member:bucket=size/0x80000:onchange($bucket) .rss_stat_throttled(mm_id,curr,member,size)' >> events/kmem/rss_stat/trigger A use case for using constants with addition/subtraction is not yet known, but for completeness the use of constants are supported for all operators. Link: https://lkml.kernel.org/r/20211025200852.3002369-2-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-10-25 20:08:33 +00:00
if (flags & HIST_FIELD_FL_CONST) {
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
hist_field->fn_num = HIST_FIELD_FN_CONST;
tracing: Add support for creating hist trigger variables from literal Currently hist trigger expressions don't support the use of numeric literals: e.g. echo 'hist:keys=common_pid:x=$y-1234' --> is not valid expression syntax Having the ability to use numeric constants in hist triggers supports a wider range of expressions for creating variables. Add support for creating trace event histogram variables from numeric literals. e.g. echo 'hist:keys=common_pid:x=1234,y=size-1024' >> event/trigger A negative numeric constant is created, using unary minus operator (parentheses are required). e.g. echo 'hist:keys=common_pid:z=-(2)' >> event/trigger Constants can be used with division/multiplication (added in the next patch in this series) to implement granularity filters for frequent trace events. For instance we can limit emitting the rss_stat trace event to when there is a 512KB cross over in the rss size: # Create a synthetic event to monitor instead of the high frequency # rss_stat event echo 'rss_stat_throttled unsigned int mm_id; unsigned int curr; int member; long size' >> tracing/synthetic_events # Create a hist trigger that emits the synthetic rss_stat_throttled # event only when the rss size crosses a 512KB boundary. echo 'hist:keys=keys=mm_id,member:bucket=size/0x80000:onchange($bucket) .rss_stat_throttled(mm_id,curr,member,size)' >> events/kmem/rss_stat/trigger A use case for using constants with addition/subtraction is not yet known, but for completeness the use of constants are supported for all operators. Link: https://lkml.kernel.org/r/20211025200852.3002369-2-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-10-25 20:08:33 +00:00
hist_field->size = sizeof(u64);
hist_field->type = kstrdup("u64", GFP_KERNEL);
if (!hist_field->type)
goto free;
goto out;
}
if (flags & HIST_FIELD_FL_STACKTRACE) {
tracing/histogram: Fix stacktrace key The current code will always use the current stacktrace as a key even if a stacktrace contained in a specific event field was specified. For example, we expect to use the 'unsigned long[] stack' field in the below event in the histogram: # echo 's:block_lat pid_t pid; u64 delta; unsigned long[] stack;' > /sys/kernel/debug/tracing/dynamic_events # echo 'hist:keys=delta.buckets=100,stack.stacktrace:sort=delta' > /sys/kernel/debug/tracing/events/synthetic/block_lat/trigger But in fact, when we type out the trigger, we see that it's using the plain old global 'stacktrace' as the key, which is just the stacktrace when the event was hit and not the stacktrace contained in the event, which is what we want: # cat /sys/kernel/debug/tracing/events/synthetic/block_lat/trigger hist:keys=delta.buckets=100,stacktrace:vals=hitcount:sort=delta.buckets=100:size=2048 [active] And in fact, there's no code to actually retrieve it from the event, so we need to add HIST_FIELD_FN_STACK and hist_field_stack() to get it and hook it into the trigger code. For now, since the stack is just using dynamic strings, this could just use the dynamic string function, but it seems cleaner to have a dedicated function an be able to tweak independently as necessary. Link: https://lkml.kernel.org/r/11aa614c82976adbfa4ea763dbe885b5fb01d59c.1676063532.git.zanussi@kernel.org Signed-off-by: Tom Zanussi <zanussi@kernel.org> [ Fixed 32bit build warning reported by kernel test robot <lkp@intel.com> ] Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2023-02-10 21:33:05 +00:00
if (field)
hist_field->fn_num = HIST_FIELD_FN_STACK;
else
hist_field->fn_num = HIST_FIELD_FN_NOP;
hist_field->size = HIST_STACKTRACE_SIZE;
hist_field->type = kstrdup_const("unsigned long[]", GFP_KERNEL);
if (!hist_field->type)
goto free;
goto out;
}
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
if (flags & (HIST_FIELD_FL_LOG2 | HIST_FIELD_FL_BUCKET)) {
unsigned long fl = flags & ~(HIST_FIELD_FL_LOG2 | HIST_FIELD_FL_BUCKET);
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
hist_field->fn_num = flags & HIST_FIELD_FL_LOG2 ? HIST_FIELD_FN_LOG2 :
HIST_FIELD_FN_BUCKET;
hist_field->operands[0] = create_hist_field(hist_data, field, fl, NULL);
if (!hist_field->operands[0])
goto free;
hist_field->size = hist_field->operands[0]->size;
hist_field->type = kstrdup_const(hist_field->operands[0]->type, GFP_KERNEL);
if (!hist_field->type)
goto free;
tracing: Add hist trigger 'log2' modifier Allow users to have numeric fields displayed as log2 values in case value range is very wide by appending '.log2' to field names. For example, # echo 'hist:key=bytes_req' > kmalloc/trigger # cat kmalloc/hist { bytes_req: 504 } hitcount: 1 { bytes_req: 11 } hitcount: 1 { bytes_req: 104 } hitcount: 1 { bytes_req: 48 } hitcount: 1 { bytes_req: 2048 } hitcount: 1 { bytes_req: 4096 } hitcount: 1 { bytes_req: 240 } hitcount: 1 { bytes_req: 392 } hitcount: 1 { bytes_req: 13 } hitcount: 1 { bytes_req: 28 } hitcount: 1 { bytes_req: 12 } hitcount: 1 { bytes_req: 64 } hitcount: 2 { bytes_req: 128 } hitcount: 2 { bytes_req: 32 } hitcount: 2 { bytes_req: 8 } hitcount: 11 { bytes_req: 10 } hitcount: 13 { bytes_req: 24 } hitcount: 25 { bytes_req: 160 } hitcount: 29 { bytes_req: 16 } hitcount: 33 { bytes_req: 80 } hitcount: 36 When using '.log2' modifier, the output looks like: # echo 'hist:key=bytes_req.log2' > kmalloc/trigger # cat kmalloc/hist { bytes_req: ~ 2^12 } hitcount: 1 { bytes_req: ~ 2^11 } hitcount: 1 { bytes_req: ~ 2^9 } hitcount: 2 { bytes_req: ~ 2^6 } hitcount: 3 { bytes_req: ~ 2^3 } hitcount: 13 { bytes_req: ~ 2^5 } hitcount: 19 { bytes_req: ~ 2^8 } hitcount: 49 { bytes_req: ~ 2^7 } hitcount: 57 { bytes_req: ~ 2^4 } hitcount: 74 Link: http://lkml.kernel.org/r/7ff396b246c6a881f46b979735fddf05a0d6c71a.1457029949.git.tom.zanussi@linux.intel.com Cc: Tom Zanussi <tom.zanussi@linux.intel.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Signed-off-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:55:02 +00:00
goto out;
}
if (flags & HIST_FIELD_FL_TIMESTAMP) {
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
hist_field->fn_num = HIST_FIELD_FN_TIMESTAMP;
hist_field->size = sizeof(u64);
hist_field->type = "u64";
goto out;
}
if (flags & HIST_FIELD_FL_CPU) {
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
hist_field->fn_num = HIST_FIELD_FN_CPU;
hist_field->size = sizeof(int);
hist_field->type = "unsigned int";
goto out;
}
if (WARN_ON_ONCE(!field))
goto out;
tracing: Do not reference char * as a string in histograms The histogram logic was allowing events with char * pointers to be used as normal strings. But it was easy to crash the kernel with: # echo 'hist:keys=filename' > events/syscalls/sys_enter_openat/trigger And open some files, and boom! BUG: unable to handle page fault for address: 00007f2ced0c3280 #PF: supervisor read access in kernel mode #PF: error_code(0x0000) - not-present page PGD 1173fa067 P4D 1173fa067 PUD 1171b6067 PMD 1171dd067 PTE 0 Oops: 0000 [#1] PREEMPT SMP CPU: 6 PID: 1810 Comm: cat Not tainted 5.13.0-rc5-test+ #61 Hardware name: Hewlett-Packard HP Compaq Pro 6300 SFF/339A, BIOS K01 v03.03 07/14/2016 RIP: 0010:strlen+0x0/0x20 Code: f6 82 80 2a 0b a9 20 74 11 0f b6 50 01 48 83 c0 01 f6 82 80 2a 0b a9 20 75 ef c3 66 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 40 00 <80> 3f 00 74 10 48 89 f8 48 83 c0 01 80 38 00 75 f7 48 29 f8 c3 RSP: 0018:ffffbdbf81567b50 EFLAGS: 00010246 RAX: 0000000000000003 RBX: ffff93815cdb3800 RCX: ffff9382401a22d0 RDX: 0000000000000100 RSI: 0000000000000000 RDI: 00007f2ced0c3280 RBP: 0000000000000100 R08: ffff9382409ff074 R09: ffffbdbf81567c98 R10: ffff9382409ff074 R11: 0000000000000000 R12: ffff9382409ff074 R13: 0000000000000001 R14: ffff93815a744f00 R15: 00007f2ced0c3280 FS: 00007f2ced0f8580(0000) GS:ffff93825a800000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007f2ced0c3280 CR3: 0000000107069005 CR4: 00000000001706e0 Call Trace: event_hist_trigger+0x463/0x5f0 ? find_held_lock+0x32/0x90 ? sched_clock_cpu+0xe/0xd0 ? lock_release+0x155/0x440 ? kernel_init_free_pages+0x6d/0x90 ? preempt_count_sub+0x9b/0xd0 ? kernel_init_free_pages+0x6d/0x90 ? get_page_from_freelist+0x12c4/0x1680 ? __rb_reserve_next+0xe5/0x460 ? ring_buffer_lock_reserve+0x12a/0x3f0 event_triggers_call+0x52/0xe0 ftrace_syscall_enter+0x264/0x2c0 syscall_trace_enter.constprop.0+0x1ee/0x210 do_syscall_64+0x1c/0x80 entry_SYSCALL_64_after_hwframe+0x44/0xae Where it triggered a fault on strlen(key) where key was the filename. The reason is that filename is a char * to user space, and the histogram code just blindly dereferenced it, with obvious bad results. I originally tried to use strncpy_from_user/kernel_nofault() but found that there's other places that its dereferenced and not worth the effort. Just do not allow "char *" to act like strings. Link: https://lkml.kernel.org/r/20210715000206.025df9d2@rorschach.local.home Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tzvetomir Stoyanov <tz.stoyanov@gmail.com> Cc: stable@vger.kernel.org Acked-by: Namhyung Kim <namhyung@kernel.org> Acked-by: Tom Zanussi <zanussi@kernel.org> Fixes: 79e577cbce4c4 ("tracing: Support string type key properly") Fixes: 5967bd5c4239 ("tracing: Let filter_assign_type() detect FILTER_PTR_STRING") Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-15 04:02:06 +00:00
/* Pointers to strings are just pointers and dangerous to dereference */
if (is_string_field(field) &&
(field->filter_type != FILTER_PTR_STRING)) {
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
flags |= HIST_FIELD_FL_STRING;
hist_field->size = MAX_FILTER_STR_VAL;
hist_field->type = kstrdup_const(field->type, GFP_KERNEL);
if (!hist_field->type)
goto free;
if (field->filter_type == FILTER_STATIC_STRING) {
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
hist_field->fn_num = HIST_FIELD_FN_STRING;
hist_field->size = field->size;
tracing: Support __rel_loc relative dynamic data location attribute Add '__rel_loc' new dynamic data location attribute which encodes the data location from the next to the field itself. The '__data_loc' is used for encoding the dynamic data location on the trace event record. But '__data_loc' is not useful if the writer doesn't know the event header (e.g. user event), because it records the dynamic data offset from the entry of the record, not the field itself. This new '__rel_loc' attribute encodes the data location relatively from the next of the field. For example, when there is a record like below (the number in the parentheses is the size of fields) |header(N)|common(M)|fields(K)|__data_loc(4)|fields(L)|data(G)| In this case, '__data_loc' field will be __data_loc = (G << 16) | (N+M+K+4+L) If '__rel_loc' is used, this will be |header(N)|common(M)|fields(K)|__rel_loc(4)|fields(L)|data(G)| where __rel_loc = (G << 16) | (L) This case shows L bytes after the '__rel_loc' attribute field, if there is no fields after the __rel_loc field, L must be 0. This is relatively easy (and no need to consider the kernel header change) when the event data fields are composed by user who doesn't know header and common fields. Link: https://lkml.kernel.org/r/163757341258.510314.4214431827833229956.stgit@devnote2 Cc: Beau Belgrave <beaub@linux.microsoft.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-11-22 09:30:12 +00:00
} else if (field->filter_type == FILTER_DYN_STRING) {
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
hist_field->fn_num = HIST_FIELD_FN_DYNSTRING;
tracing: Support __rel_loc relative dynamic data location attribute Add '__rel_loc' new dynamic data location attribute which encodes the data location from the next to the field itself. The '__data_loc' is used for encoding the dynamic data location on the trace event record. But '__data_loc' is not useful if the writer doesn't know the event header (e.g. user event), because it records the dynamic data offset from the entry of the record, not the field itself. This new '__rel_loc' attribute encodes the data location relatively from the next of the field. For example, when there is a record like below (the number in the parentheses is the size of fields) |header(N)|common(M)|fields(K)|__data_loc(4)|fields(L)|data(G)| In this case, '__data_loc' field will be __data_loc = (G << 16) | (N+M+K+4+L) If '__rel_loc' is used, this will be |header(N)|common(M)|fields(K)|__rel_loc(4)|fields(L)|data(G)| where __rel_loc = (G << 16) | (L) This case shows L bytes after the '__rel_loc' attribute field, if there is no fields after the __rel_loc field, L must be 0. This is relatively easy (and no need to consider the kernel header change) when the event data fields are composed by user who doesn't know header and common fields. Link: https://lkml.kernel.org/r/163757341258.510314.4214431827833229956.stgit@devnote2 Cc: Beau Belgrave <beaub@linux.microsoft.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-11-22 09:30:12 +00:00
} else if (field->filter_type == FILTER_RDYN_STRING)
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
hist_field->fn_num = HIST_FIELD_FN_RELDYNSTRING;
else
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
hist_field->fn_num = HIST_FIELD_FN_PSTRING;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
} else {
hist_field->size = field->size;
hist_field->is_signed = field->is_signed;
hist_field->type = kstrdup_const(field->type, GFP_KERNEL);
if (!hist_field->type)
goto free;
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
hist_field->fn_num = select_value_fn(field->size,
field->is_signed);
if (hist_field->fn_num == HIST_FIELD_FN_NOP) {
destroy_hist_field(hist_field, 0);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
return NULL;
}
}
out:
hist_field->field = field;
hist_field->flags = flags;
if (var_name) {
hist_field->var.name = kstrdup(var_name, GFP_KERNEL);
if (!hist_field->var.name)
goto free;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
return hist_field;
free:
destroy_hist_field(hist_field, 0);
return NULL;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
static void destroy_hist_fields(struct hist_trigger_data *hist_data)
{
unsigned int i;
for (i = 0; i < HIST_FIELDS_MAX; i++) {
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (hist_data->fields[i]) {
destroy_hist_field(hist_data->fields[i], 0);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
hist_data->fields[i] = NULL;
}
}
for (i = 0; i < hist_data->n_var_refs; i++) {
WARN_ON(!(hist_data->var_refs[i]->flags & HIST_FIELD_FL_VAR_REF));
__destroy_hist_field(hist_data->var_refs[i]);
hist_data->var_refs[i] = NULL;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
static int init_var_ref(struct hist_field *ref_field,
struct hist_field *var_field,
char *system, char *event_name)
{
int err = 0;
ref_field->var.idx = var_field->var.idx;
ref_field->var.hist_data = var_field->hist_data;
ref_field->size = var_field->size;
ref_field->is_signed = var_field->is_signed;
ref_field->flags |= var_field->flags &
(HIST_FIELD_FL_TIMESTAMP | HIST_FIELD_FL_TIMESTAMP_USECS);
if (system) {
ref_field->system = kstrdup(system, GFP_KERNEL);
if (!ref_field->system)
return -ENOMEM;
}
if (event_name) {
ref_field->event_name = kstrdup(event_name, GFP_KERNEL);
if (!ref_field->event_name) {
err = -ENOMEM;
goto free;
}
}
if (var_field->var.name) {
ref_field->name = kstrdup(var_field->var.name, GFP_KERNEL);
if (!ref_field->name) {
err = -ENOMEM;
goto free;
}
} else if (var_field->name) {
ref_field->name = kstrdup(var_field->name, GFP_KERNEL);
if (!ref_field->name) {
err = -ENOMEM;
goto free;
}
}
ref_field->type = kstrdup_const(var_field->type, GFP_KERNEL);
if (!ref_field->type) {
err = -ENOMEM;
goto free;
}
out:
return err;
free:
kfree(ref_field->system);
ref_field->system = NULL;
kfree(ref_field->event_name);
ref_field->event_name = NULL;
kfree(ref_field->name);
ref_field->name = NULL;
goto out;
}
static int find_var_ref_idx(struct hist_trigger_data *hist_data,
struct hist_field *var_field)
{
struct hist_field *ref_field;
int i;
for (i = 0; i < hist_data->n_var_refs; i++) {
ref_field = hist_data->var_refs[i];
if (ref_field->var.idx == var_field->var.idx &&
ref_field->var.hist_data == var_field->hist_data)
return i;
}
return -ENOENT;
}
/**
* create_var_ref - Create a variable reference and attach it to trigger
* @hist_data: The trigger that will be referencing the variable
* @var_field: The VAR field to create a reference to
* @system: The optional system string
* @event_name: The optional event_name string
*
* Given a variable hist_field, create a VAR_REF hist_field that
* represents a reference to it.
*
* This function also adds the reference to the trigger that
* now references the variable.
*
* Return: The VAR_REF field if successful, NULL if not
*/
static struct hist_field *create_var_ref(struct hist_trigger_data *hist_data,
struct hist_field *var_field,
char *system, char *event_name)
{
unsigned long flags = HIST_FIELD_FL_VAR_REF;
struct hist_field *ref_field;
tracing: Fix histogram code when expression has same var as value While working on a tool to convert SQL syntex into the histogram language of the kernel, I discovered the following bug: # echo 'first u64 start_time u64 end_time pid_t pid u64 delta' >> synthetic_events # echo 'hist:keys=pid:start=common_timestamp' > events/sched/sched_waking/trigger # echo 'hist:keys=next_pid:delta=common_timestamp-$start,start2=$start:onmatch(sched.sched_waking).trace(first,$start2,common_timestamp,next_pid,$delta)' > events/sched/sched_switch/trigger Would not display any histograms in the sched_switch histogram side. But if I were to swap the location of "delta=common_timestamp-$start" with "start2=$start" Such that the last line had: # echo 'hist:keys=next_pid:start2=$start,delta=common_timestamp-$start:onmatch(sched.sched_waking).trace(first,$start2,common_timestamp,next_pid,$delta)' > events/sched/sched_switch/trigger The histogram works as expected. What I found out is that the expressions clear out the value once it is resolved. As the variables are resolved in the order listed, when processing: delta=common_timestamp-$start The $start is cleared. When it gets to "start2=$start", it errors out with "unresolved symbol" (which is silent as this happens at the location of the trace), and the histogram is dropped. When processing the histogram for variable references, instead of adding a new reference for a variable used twice, use the same reference. That way, not only is it more efficient, but the order will no longer matter in processing of the variables. From Tom Zanussi: "Just to clarify some more about what the problem was is that without your patch, we would have two separate references to the same variable, and during resolve_var_refs(), they'd both want to be resolved separately, so in this case, since the first reference to start wasn't part of an expression, it wouldn't get the read-once flag set, so would be read normally, and then the second reference would do the read-once read and also be read but using read-once. So everything worked and you didn't see a problem: from: start2=$start,delta=common_timestamp-$start In the second case, when you switched them around, the first reference would be resolved by doing the read-once, and following that the second reference would try to resolve and see that the variable had already been read, so failed as unset, which caused it to short-circuit out and not do the trigger action to generate the synthetic event: to: delta=common_timestamp-$start,start2=$start With your patch, we only have the single resolution which happens correctly the one time it's resolved, so this can't happen." Link: https://lore.kernel.org/r/20200116154216.58ca08eb@gandalf.local.home Cc: stable@vger.kernel.org Fixes: 067fe038e70f6 ("tracing: Add variable reference handling to hist triggers") Reviewed-by: Tom Zanuss <zanussi@kernel.org> Tested-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2020-01-20 18:07:31 +00:00
int i;
/* Check if the variable already exists */
for (i = 0; i < hist_data->n_var_refs; i++) {
ref_field = hist_data->var_refs[i];
if (ref_field->var.idx == var_field->var.idx &&
ref_field->var.hist_data == var_field->hist_data) {
get_hist_field(ref_field);
return ref_field;
}
}
tracing/hist: Fix out-of-bound write on 'action_data.var_ref_idx' When generate a synthetic event with many params and then create a trace action for it [1], kernel panic happened [2]. It is because that in trace_action_create() 'data->n_params' is up to SYNTH_FIELDS_MAX (current value is 64), and array 'data->var_ref_idx' keeps indices into array 'hist_data->var_refs' for each synthetic event param, but the length of 'data->var_ref_idx' is TRACING_MAP_VARS_MAX (current value is 16), so out-of-bound write happened when 'data->n_params' more than 16. In this case, 'data->match_data.event' is overwritten and eventually cause the panic. To solve the issue, adjust the length of 'data->var_ref_idx' to be SYNTH_FIELDS_MAX and add sanity checks to avoid out-of-bound write. [1] # cd /sys/kernel/tracing/ # echo "my_synth_event int v1; int v2; int v3; int v4; int v5; int v6;\ int v7; int v8; int v9; int v10; int v11; int v12; int v13; int v14;\ int v15; int v16; int v17; int v18; int v19; int v20; int v21; int v22;\ int v23; int v24; int v25; int v26; int v27; int v28; int v29; int v30;\ int v31; int v32; int v33; int v34; int v35; int v36; int v37; int v38;\ int v39; int v40; int v41; int v42; int v43; int v44; int v45; int v46;\ int v47; int v48; int v49; int v50; int v51; int v52; int v53; int v54;\ int v55; int v56; int v57; int v58; int v59; int v60; int v61; int v62;\ int v63" >> synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="bash"' >> \ events/sched/sched_waking/trigger # echo "hist:keys=next_pid:onmatch(sched.sched_waking).my_synth_event(\ pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,\ pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,\ pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,\ pid,pid,pid,pid,pid,pid,pid,pid,pid)" >> events/sched/sched_switch/trigger [2] BUG: unable to handle page fault for address: ffff91c900000000 PGD 61001067 P4D 61001067 PUD 0 Oops: 0000 [#1] PREEMPT SMP NOPTI CPU: 2 PID: 322 Comm: bash Tainted: G W 6.1.0-rc8+ #229 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.15.0-0-g2dd4b9b3f840-prebuilt.qemu.org 04/01/2014 RIP: 0010:strcmp+0xc/0x30 Code: 75 f7 31 d2 44 0f b6 04 16 44 88 04 11 48 83 c2 01 45 84 c0 75 ee c3 cc cc cc cc 0f 1f 00 31 c0 eb 08 48 83 c0 01 84 d2 74 13 <0f> b6 14 07 3a 14 06 74 ef 19 c0 83 c8 01 c3 cc cc cc cc 31 c3 RSP: 0018:ffff9b3b00f53c48 EFLAGS: 00000246 RAX: 0000000000000000 RBX: ffffffffba958a68 RCX: 0000000000000000 RDX: 0000000000000010 RSI: ffff91c943d33a90 RDI: ffff91c900000000 RBP: ffff91c900000000 R08: 00000018d604b529 R09: 0000000000000000 R10: ffff91c9483eddb1 R11: ffff91ca483eddab R12: ffff91c946171580 R13: ffff91c9479f0538 R14: ffff91c9457c2848 R15: ffff91c9479f0538 FS: 00007f1d1cfbe740(0000) GS:ffff91c9bdc80000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffff91c900000000 CR3: 0000000006316000 CR4: 00000000000006e0 Call Trace: <TASK> __find_event_file+0x55/0x90 action_create+0x76c/0x1060 event_hist_trigger_parse+0x146d/0x2060 ? event_trigger_write+0x31/0xd0 trigger_process_regex+0xbb/0x110 event_trigger_write+0x6b/0xd0 vfs_write+0xc8/0x3e0 ? alloc_fd+0xc0/0x160 ? preempt_count_add+0x4d/0xa0 ? preempt_count_add+0x70/0xa0 ksys_write+0x5f/0xe0 do_syscall_64+0x3b/0x90 entry_SYSCALL_64_after_hwframe+0x63/0xcd RIP: 0033:0x7f1d1d0cf077 Code: 64 89 02 48 c7 c0 ff ff ff ff eb bb 0f 1f 80 00 00 00 00 f3 0f 1e fa 64 8b 04 25 18 00 00 00 85 c0 75 10 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 51 c3 48 83 ec 28 48 89 54 24 18 48 89 74 RSP: 002b:00007ffcebb0e568 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 RAX: ffffffffffffffda RBX: 0000000000000143 RCX: 00007f1d1d0cf077 RDX: 0000000000000143 RSI: 00005639265aa7e0 RDI: 0000000000000001 RBP: 00005639265aa7e0 R08: 000000000000000a R09: 0000000000000142 R10: 000056392639c017 R11: 0000000000000246 R12: 0000000000000143 R13: 00007f1d1d1ae6a0 R14: 00007f1d1d1aa4a0 R15: 00007f1d1d1a98a0 </TASK> Modules linked in: CR2: ffff91c900000000 ---[ end trace 0000000000000000 ]--- RIP: 0010:strcmp+0xc/0x30 Code: 75 f7 31 d2 44 0f b6 04 16 44 88 04 11 48 83 c2 01 45 84 c0 75 ee c3 cc cc cc cc 0f 1f 00 31 c0 eb 08 48 83 c0 01 84 d2 74 13 <0f> b6 14 07 3a 14 06 74 ef 19 c0 83 c8 01 c3 cc cc cc cc 31 c3 RSP: 0018:ffff9b3b00f53c48 EFLAGS: 00000246 RAX: 0000000000000000 RBX: ffffffffba958a68 RCX: 0000000000000000 RDX: 0000000000000010 RSI: ffff91c943d33a90 RDI: ffff91c900000000 RBP: ffff91c900000000 R08: 00000018d604b529 R09: 0000000000000000 R10: ffff91c9483eddb1 R11: ffff91ca483eddab R12: ffff91c946171580 R13: ffff91c9479f0538 R14: ffff91c9457c2848 R15: ffff91c9479f0538 FS: 00007f1d1cfbe740(0000) GS:ffff91c9bdc80000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffff91c900000000 CR3: 0000000006316000 CR4: 00000000000006e0 Link: https://lore.kernel.org/linux-trace-kernel/20221207035143.2278781-1-zhengyejian1@huawei.com Cc: <mhiramat@kernel.org> Cc: <zanussi@kernel.org> Cc: stable@vger.kernel.org Fixes: d380dcde9a07 ("tracing: Fix now invalid var_ref_vals assumption in trace action") Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-12-07 03:51:43 +00:00
/* Sanity check to avoid out-of-bound write on 'hist_data->var_refs' */
if (hist_data->n_var_refs >= TRACING_MAP_VARS_MAX)
return NULL;
ref_field = create_hist_field(var_field->hist_data, NULL, flags, NULL);
if (ref_field) {
if (init_var_ref(ref_field, var_field, system, event_name)) {
destroy_hist_field(ref_field, 0);
return NULL;
}
hist_data->var_refs[hist_data->n_var_refs] = ref_field;
ref_field->var_ref_idx = hist_data->n_var_refs++;
}
return ref_field;
}
static bool is_var_ref(char *var_name)
{
if (!var_name || strlen(var_name) < 2 || var_name[0] != '$')
return false;
return true;
}
static char *field_name_from_var(struct hist_trigger_data *hist_data,
char *var_name)
{
char *name, *field;
unsigned int i;
for (i = 0; i < hist_data->attrs->var_defs.n_vars; i++) {
name = hist_data->attrs->var_defs.name[i];
if (strcmp(var_name, name) == 0) {
field = hist_data->attrs->var_defs.expr[i];
if (contains_operator(field, NULL) || is_var_ref(field))
continue;
return field;
}
}
return NULL;
}
static char *local_field_var_ref(struct hist_trigger_data *hist_data,
char *system, char *event_name,
char *var_name)
{
struct trace_event_call *call;
if (system && event_name) {
call = hist_data->event_file->event_call;
if (strcmp(system, call->class->system) != 0)
return NULL;
if (strcmp(event_name, trace_event_name(call)) != 0)
return NULL;
}
if (!!system != !!event_name)
return NULL;
if (!is_var_ref(var_name))
return NULL;
var_name++;
return field_name_from_var(hist_data, var_name);
}
static struct hist_field *parse_var_ref(struct hist_trigger_data *hist_data,
char *system, char *event_name,
char *var_name)
{
struct hist_field *var_field = NULL, *ref_field = NULL;
struct trace_array *tr = hist_data->event_file->tr;
if (!is_var_ref(var_name))
return NULL;
var_name++;
var_field = find_event_var(hist_data, system, event_name, var_name);
if (var_field)
ref_field = create_var_ref(hist_data, var_field,
system, event_name);
if (!ref_field)
hist_err(tr, HIST_ERR_VAR_NOT_FOUND, errpos(var_name));
return ref_field;
}
static struct ftrace_event_field *
parse_field(struct hist_trigger_data *hist_data, struct trace_event_file *file,
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
char *field_str, unsigned long *flags, unsigned long *buckets)
{
struct ftrace_event_field *field = NULL;
char *field_name, *modifier, *str;
struct trace_array *tr = file->tr;
modifier = str = kstrdup(field_str, GFP_KERNEL);
if (!modifier)
return ERR_PTR(-ENOMEM);
field_name = strsep(&modifier, ".");
if (modifier) {
if (strcmp(modifier, "hex") == 0)
*flags |= HIST_FIELD_FL_HEX;
else if (strcmp(modifier, "sym") == 0)
*flags |= HIST_FIELD_FL_SYM;
/*
* 'sym-offset' occurrences in the trigger string are modified
* to 'symXoffset' to simplify arithmetic expression parsing.
*/
else if (strcmp(modifier, "symXoffset") == 0)
*flags |= HIST_FIELD_FL_SYM_OFFSET;
else if ((strcmp(modifier, "execname") == 0) &&
(strcmp(field_name, "common_pid") == 0))
*flags |= HIST_FIELD_FL_EXECNAME;
else if (strcmp(modifier, "syscall") == 0)
*flags |= HIST_FIELD_FL_SYSCALL;
tracing/histogram: Add stacktrace type Now that stacktraces can be part of synthetic events, allow a key to be typed as a stacktrace. # cd /sys/kernel/tracing # echo 's:block_lat u64 delta; unsigned long stack[];' >> dynamic_events # echo 'hist:keys=next_pid:ts=common_timestamp.usecs,st=stacktrace if prev_state == 2' >> events/sched/sched_switch/trigger # echo 'hist:keys=prev_pid:delta=common_timestamp.usecs-$ts,st2=$st:onmatch(sched.sched_switch).trace(block_lat,$delta,$st2)' >> events/sched/sched_switch/trigger # echo 'hist:keys=delta.buckets=100,stack.stacktrace:sort=delta' > events/synthetic/block_lat/trigger # cat events/synthetic/block_lat/hist # event histogram # # trigger info: hist:keys=delta.buckets=100,stacktrace:vals=hitcount:sort=delta.buckets=100:size=2048 [active] # { delta: ~ 0-99, stacktrace: event_hist_trigger+0x464/0x480 event_triggers_call+0x52/0xe0 trace_event_buffer_commit+0x193/0x250 trace_event_raw_event_sched_switch+0xfc/0x150 __traceiter_sched_switch+0x41/0x60 __schedule+0x448/0x7b0 schedule_idle+0x26/0x40 cpu_startup_entry+0x19/0x20 start_secondary+0xed/0xf0 secondary_startup_64_no_verify+0xe0/0xeb } hitcount: 6 { delta: ~ 0-99, stacktrace: event_hist_trigger+0x464/0x480 event_triggers_call+0x52/0xe0 trace_event_buffer_commit+0x193/0x250 trace_event_raw_event_sched_switch+0xfc/0x150 __traceiter_sched_switch+0x41/0x60 __schedule+0x448/0x7b0 schedule_idle+0x26/0x40 cpu_startup_entry+0x19/0x20 __pfx_kernel_init+0x0/0x10 arch_call_rest_init+0xa/0x24 start_kernel+0x964/0x98d secondary_startup_64_no_verify+0xe0/0xeb } hitcount: 3 { delta: ~ 0-99, stacktrace: event_hist_trigger+0x464/0x480 event_triggers_call+0x52/0xe0 trace_event_buffer_commit+0x193/0x250 trace_event_raw_event_sched_switch+0xfc/0x150 __traceiter_sched_switch+0x41/0x60 __schedule+0x448/0x7b0 schedule+0x5a/0xb0 worker_thread+0xaf/0x380 kthread+0xe9/0x110 ret_from_fork+0x2c/0x50 } hitcount: 1 { delta: ~ 100-199, stacktrace: event_hist_trigger+0x464/0x480 event_triggers_call+0x52/0xe0 trace_event_buffer_commit+0x193/0x250 trace_event_raw_event_sched_switch+0xfc/0x150 __traceiter_sched_switch+0x41/0x60 __schedule+0x448/0x7b0 schedule_idle+0x26/0x40 cpu_startup_entry+0x19/0x20 start_secondary+0xed/0xf0 secondary_startup_64_no_verify+0xe0/0xeb } hitcount: 15 [..] { delta: ~ 8500-8599, stacktrace: event_hist_trigger+0x464/0x480 event_triggers_call+0x52/0xe0 trace_event_buffer_commit+0x193/0x250 trace_event_raw_event_sched_switch+0xfc/0x150 __traceiter_sched_switch+0x41/0x60 __schedule+0x448/0x7b0 schedule_idle+0x26/0x40 cpu_startup_entry+0x19/0x20 start_secondary+0xed/0xf0 secondary_startup_64_no_verify+0xe0/0xeb } hitcount: 1 Totals: Hits: 89 Entries: 11 Dropped: 0 Link: https://lkml.kernel.org/r/20230117152236.167046397@goodmis.org Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Tom Zanussi <zanussi@kernel.org> Cc: Ross Zwisler <zwisler@google.com> Cc: Ching-lin Yu <chinglinyu@google.com> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2023-01-17 15:21:29 +00:00
else if (strcmp(modifier, "stacktrace") == 0)
*flags |= HIST_FIELD_FL_STACKTRACE;
else if (strcmp(modifier, "log2") == 0)
*flags |= HIST_FIELD_FL_LOG2;
else if (strcmp(modifier, "usecs") == 0)
*flags |= HIST_FIELD_FL_TIMESTAMP_USECS;
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
else if (strncmp(modifier, "bucket", 6) == 0) {
int ret;
modifier += 6;
if (*modifier == 's')
modifier++;
if (*modifier != '=')
goto error;
modifier++;
ret = kstrtoul(modifier, 0, buckets);
if (ret || !(*buckets))
goto error;
*flags |= HIST_FIELD_FL_BUCKET;
} else if (strncmp(modifier, "percent", 7) == 0) {
if (*flags & (HIST_FIELD_FL_VAR | HIST_FIELD_FL_KEY))
goto error;
*flags |= HIST_FIELD_FL_PERCENT;
} else if (strncmp(modifier, "graph", 5) == 0) {
if (*flags & (HIST_FIELD_FL_VAR | HIST_FIELD_FL_KEY))
goto error;
*flags |= HIST_FIELD_FL_GRAPH;
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
} else {
error:
hist_err(tr, HIST_ERR_BAD_FIELD_MODIFIER, errpos(modifier));
field = ERR_PTR(-EINVAL);
goto out;
}
}
if (strcmp(field_name, "common_timestamp") == 0) {
*flags |= HIST_FIELD_FL_TIMESTAMP;
hist_data->enable_timestamps = true;
if (*flags & HIST_FIELD_FL_TIMESTAMP_USECS)
hist_data->attrs->ts_in_usecs = true;
} else if (strcmp(field_name, "common_stacktrace") == 0) {
*flags |= HIST_FIELD_FL_STACKTRACE;
tracing/histogram: Rename "cpu" to "common_cpu" Currently the histogram logic allows the user to write "cpu" in as an event field, and it will record the CPU that the event happened on. The problem with this is that there's a lot of events that have "cpu" as a real field, and using "cpu" as the CPU it ran on, makes it impossible to run histograms on the "cpu" field of events. For example, if I want to have a histogram on the count of the workqueue_queue_work event on its cpu field, running: ># echo 'hist:keys=cpu' > events/workqueue/workqueue_queue_work/trigger Gives a misleading and wrong result. Change the command to "common_cpu" as no event should have "common_*" fields as that's a reserved name for fields used by all events. And this makes sense here as common_cpu would be a field used by all events. Now we can even do: ># echo 'hist:keys=common_cpu,cpu if cpu < 100' > events/workqueue/workqueue_queue_work/trigger ># cat events/workqueue/workqueue_queue_work/hist # event histogram # # trigger info: hist:keys=common_cpu,cpu:vals=hitcount:sort=hitcount:size=2048 if cpu < 100 [active] # { common_cpu: 0, cpu: 2 } hitcount: 1 { common_cpu: 0, cpu: 4 } hitcount: 1 { common_cpu: 7, cpu: 7 } hitcount: 1 { common_cpu: 0, cpu: 7 } hitcount: 1 { common_cpu: 0, cpu: 1 } hitcount: 1 { common_cpu: 0, cpu: 6 } hitcount: 2 { common_cpu: 0, cpu: 5 } hitcount: 2 { common_cpu: 1, cpu: 1 } hitcount: 4 { common_cpu: 6, cpu: 6 } hitcount: 4 { common_cpu: 5, cpu: 5 } hitcount: 14 { common_cpu: 4, cpu: 4 } hitcount: 26 { common_cpu: 0, cpu: 0 } hitcount: 39 { common_cpu: 2, cpu: 2 } hitcount: 184 Now for backward compatibility, I added a trick. If "cpu" is used, and the field is not found, it will fall back to "common_cpu" and work as it did before. This way, it will still work for old programs that use "cpu" to get the actual CPU, but if the event has a "cpu" as a field, it will get that event's "cpu" field, which is probably what it wants anyway. I updated the tracefs/README to include documentation about both the common_timestamp and the common_cpu. This way, if that text is present in the README, then an application can know that common_cpu is supported over just plain "cpu". Link: https://lkml.kernel.org/r/20210721110053.26b4f641@oasis.local.home Cc: Namhyung Kim <namhyung@kernel.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: stable@vger.kernel.org Fixes: 8b7622bf94a44 ("tracing: Add cpu field for hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-21 15:00:53 +00:00
} else if (strcmp(field_name, "common_cpu") == 0)
*flags |= HIST_FIELD_FL_CPU;
else if (strcmp(field_name, "hitcount") == 0)
*flags |= HIST_FIELD_FL_HITCOUNT;
else {
field = trace_find_event_field(file->event_call, field_name);
if (!field || !field->size) {
tracing/histogram: Rename "cpu" to "common_cpu" Currently the histogram logic allows the user to write "cpu" in as an event field, and it will record the CPU that the event happened on. The problem with this is that there's a lot of events that have "cpu" as a real field, and using "cpu" as the CPU it ran on, makes it impossible to run histograms on the "cpu" field of events. For example, if I want to have a histogram on the count of the workqueue_queue_work event on its cpu field, running: ># echo 'hist:keys=cpu' > events/workqueue/workqueue_queue_work/trigger Gives a misleading and wrong result. Change the command to "common_cpu" as no event should have "common_*" fields as that's a reserved name for fields used by all events. And this makes sense here as common_cpu would be a field used by all events. Now we can even do: ># echo 'hist:keys=common_cpu,cpu if cpu < 100' > events/workqueue/workqueue_queue_work/trigger ># cat events/workqueue/workqueue_queue_work/hist # event histogram # # trigger info: hist:keys=common_cpu,cpu:vals=hitcount:sort=hitcount:size=2048 if cpu < 100 [active] # { common_cpu: 0, cpu: 2 } hitcount: 1 { common_cpu: 0, cpu: 4 } hitcount: 1 { common_cpu: 7, cpu: 7 } hitcount: 1 { common_cpu: 0, cpu: 7 } hitcount: 1 { common_cpu: 0, cpu: 1 } hitcount: 1 { common_cpu: 0, cpu: 6 } hitcount: 2 { common_cpu: 0, cpu: 5 } hitcount: 2 { common_cpu: 1, cpu: 1 } hitcount: 4 { common_cpu: 6, cpu: 6 } hitcount: 4 { common_cpu: 5, cpu: 5 } hitcount: 14 { common_cpu: 4, cpu: 4 } hitcount: 26 { common_cpu: 0, cpu: 0 } hitcount: 39 { common_cpu: 2, cpu: 2 } hitcount: 184 Now for backward compatibility, I added a trick. If "cpu" is used, and the field is not found, it will fall back to "common_cpu" and work as it did before. This way, it will still work for old programs that use "cpu" to get the actual CPU, but if the event has a "cpu" as a field, it will get that event's "cpu" field, which is probably what it wants anyway. I updated the tracefs/README to include documentation about both the common_timestamp and the common_cpu. This way, if that text is present in the README, then an application can know that common_cpu is supported over just plain "cpu". Link: https://lkml.kernel.org/r/20210721110053.26b4f641@oasis.local.home Cc: Namhyung Kim <namhyung@kernel.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: stable@vger.kernel.org Fixes: 8b7622bf94a44 ("tracing: Add cpu field for hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-21 15:00:53 +00:00
/*
* For backward compatibility, if field_name
* was "cpu" or "stacktrace", then we treat this
* the same as common_cpu and common_stacktrace
* respectively. This also works for "CPU", and
* "STACKTRACE".
tracing/histogram: Rename "cpu" to "common_cpu" Currently the histogram logic allows the user to write "cpu" in as an event field, and it will record the CPU that the event happened on. The problem with this is that there's a lot of events that have "cpu" as a real field, and using "cpu" as the CPU it ran on, makes it impossible to run histograms on the "cpu" field of events. For example, if I want to have a histogram on the count of the workqueue_queue_work event on its cpu field, running: ># echo 'hist:keys=cpu' > events/workqueue/workqueue_queue_work/trigger Gives a misleading and wrong result. Change the command to "common_cpu" as no event should have "common_*" fields as that's a reserved name for fields used by all events. And this makes sense here as common_cpu would be a field used by all events. Now we can even do: ># echo 'hist:keys=common_cpu,cpu if cpu < 100' > events/workqueue/workqueue_queue_work/trigger ># cat events/workqueue/workqueue_queue_work/hist # event histogram # # trigger info: hist:keys=common_cpu,cpu:vals=hitcount:sort=hitcount:size=2048 if cpu < 100 [active] # { common_cpu: 0, cpu: 2 } hitcount: 1 { common_cpu: 0, cpu: 4 } hitcount: 1 { common_cpu: 7, cpu: 7 } hitcount: 1 { common_cpu: 0, cpu: 7 } hitcount: 1 { common_cpu: 0, cpu: 1 } hitcount: 1 { common_cpu: 0, cpu: 6 } hitcount: 2 { common_cpu: 0, cpu: 5 } hitcount: 2 { common_cpu: 1, cpu: 1 } hitcount: 4 { common_cpu: 6, cpu: 6 } hitcount: 4 { common_cpu: 5, cpu: 5 } hitcount: 14 { common_cpu: 4, cpu: 4 } hitcount: 26 { common_cpu: 0, cpu: 0 } hitcount: 39 { common_cpu: 2, cpu: 2 } hitcount: 184 Now for backward compatibility, I added a trick. If "cpu" is used, and the field is not found, it will fall back to "common_cpu" and work as it did before. This way, it will still work for old programs that use "cpu" to get the actual CPU, but if the event has a "cpu" as a field, it will get that event's "cpu" field, which is probably what it wants anyway. I updated the tracefs/README to include documentation about both the common_timestamp and the common_cpu. This way, if that text is present in the README, then an application can know that common_cpu is supported over just plain "cpu". Link: https://lkml.kernel.org/r/20210721110053.26b4f641@oasis.local.home Cc: Namhyung Kim <namhyung@kernel.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: stable@vger.kernel.org Fixes: 8b7622bf94a44 ("tracing: Add cpu field for hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-21 15:00:53 +00:00
*/
tracing/histogram: Fix sorting on old "cpu" value When trying to add a histogram against an event with the "cpu" field, it was impossible due to "cpu" being a keyword to key off of the running CPU. So to fix this, it was changed to "common_cpu" to match the other generic fields (like "common_pid"). But since some scripts used "cpu" for keying off of the CPU (for events that did not have "cpu" as a field, which is most of them), a backward compatibility trick was added such that if "cpu" was used as a key, and the event did not have "cpu" as a field name, then it would fallback and switch over to "common_cpu". This fix has a couple of subtle bugs. One was that when switching over to "common_cpu", it did not change the field name, it just set a flag. But the code still found a "cpu" field. The "cpu" field is used for filtering and is returned when the event does not have a "cpu" field. This was found by: # cd /sys/kernel/tracing # echo hist:key=cpu,pid:sort=cpu > events/sched/sched_wakeup/trigger # cat events/sched/sched_wakeup/hist Which showed the histogram unsorted: { cpu: 19, pid: 1175 } hitcount: 1 { cpu: 6, pid: 239 } hitcount: 2 { cpu: 23, pid: 1186 } hitcount: 14 { cpu: 12, pid: 249 } hitcount: 2 { cpu: 3, pid: 994 } hitcount: 5 Instead of hard coding the "cpu" checks, take advantage of the fact that trace_event_field_field() returns a special field for "cpu" and "CPU" if the event does not have "cpu" as a field. This special field has the "filter_type" of "FILTER_CPU". Check that to test if the returned field is of the CPU type instead of doing the string compare. Also, fix the sorting bug by testing for the hist_field flag of HIST_FIELD_FL_CPU when setting up the sort routine. Otherwise it will use the special CPU field to know what compare routine to use, and since that special field does not have a size, it returns tracing_map_cmp_none. Cc: stable@vger.kernel.org Fixes: 1e3bac71c505 ("tracing/histogram: Rename "cpu" to "common_cpu"") Reported-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-03-02 03:29:04 +00:00
if (field && field->filter_type == FILTER_CPU) {
tracing/histogram: Rename "cpu" to "common_cpu" Currently the histogram logic allows the user to write "cpu" in as an event field, and it will record the CPU that the event happened on. The problem with this is that there's a lot of events that have "cpu" as a real field, and using "cpu" as the CPU it ran on, makes it impossible to run histograms on the "cpu" field of events. For example, if I want to have a histogram on the count of the workqueue_queue_work event on its cpu field, running: ># echo 'hist:keys=cpu' > events/workqueue/workqueue_queue_work/trigger Gives a misleading and wrong result. Change the command to "common_cpu" as no event should have "common_*" fields as that's a reserved name for fields used by all events. And this makes sense here as common_cpu would be a field used by all events. Now we can even do: ># echo 'hist:keys=common_cpu,cpu if cpu < 100' > events/workqueue/workqueue_queue_work/trigger ># cat events/workqueue/workqueue_queue_work/hist # event histogram # # trigger info: hist:keys=common_cpu,cpu:vals=hitcount:sort=hitcount:size=2048 if cpu < 100 [active] # { common_cpu: 0, cpu: 2 } hitcount: 1 { common_cpu: 0, cpu: 4 } hitcount: 1 { common_cpu: 7, cpu: 7 } hitcount: 1 { common_cpu: 0, cpu: 7 } hitcount: 1 { common_cpu: 0, cpu: 1 } hitcount: 1 { common_cpu: 0, cpu: 6 } hitcount: 2 { common_cpu: 0, cpu: 5 } hitcount: 2 { common_cpu: 1, cpu: 1 } hitcount: 4 { common_cpu: 6, cpu: 6 } hitcount: 4 { common_cpu: 5, cpu: 5 } hitcount: 14 { common_cpu: 4, cpu: 4 } hitcount: 26 { common_cpu: 0, cpu: 0 } hitcount: 39 { common_cpu: 2, cpu: 2 } hitcount: 184 Now for backward compatibility, I added a trick. If "cpu" is used, and the field is not found, it will fall back to "common_cpu" and work as it did before. This way, it will still work for old programs that use "cpu" to get the actual CPU, but if the event has a "cpu" as a field, it will get that event's "cpu" field, which is probably what it wants anyway. I updated the tracefs/README to include documentation about both the common_timestamp and the common_cpu. This way, if that text is present in the README, then an application can know that common_cpu is supported over just plain "cpu". Link: https://lkml.kernel.org/r/20210721110053.26b4f641@oasis.local.home Cc: Namhyung Kim <namhyung@kernel.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: stable@vger.kernel.org Fixes: 8b7622bf94a44 ("tracing: Add cpu field for hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-21 15:00:53 +00:00
*flags |= HIST_FIELD_FL_CPU;
} else if (field && field->filter_type == FILTER_STACKTRACE) {
*flags |= HIST_FIELD_FL_STACKTRACE;
tracing/histogram: Rename "cpu" to "common_cpu" Currently the histogram logic allows the user to write "cpu" in as an event field, and it will record the CPU that the event happened on. The problem with this is that there's a lot of events that have "cpu" as a real field, and using "cpu" as the CPU it ran on, makes it impossible to run histograms on the "cpu" field of events. For example, if I want to have a histogram on the count of the workqueue_queue_work event on its cpu field, running: ># echo 'hist:keys=cpu' > events/workqueue/workqueue_queue_work/trigger Gives a misleading and wrong result. Change the command to "common_cpu" as no event should have "common_*" fields as that's a reserved name for fields used by all events. And this makes sense here as common_cpu would be a field used by all events. Now we can even do: ># echo 'hist:keys=common_cpu,cpu if cpu < 100' > events/workqueue/workqueue_queue_work/trigger ># cat events/workqueue/workqueue_queue_work/hist # event histogram # # trigger info: hist:keys=common_cpu,cpu:vals=hitcount:sort=hitcount:size=2048 if cpu < 100 [active] # { common_cpu: 0, cpu: 2 } hitcount: 1 { common_cpu: 0, cpu: 4 } hitcount: 1 { common_cpu: 7, cpu: 7 } hitcount: 1 { common_cpu: 0, cpu: 7 } hitcount: 1 { common_cpu: 0, cpu: 1 } hitcount: 1 { common_cpu: 0, cpu: 6 } hitcount: 2 { common_cpu: 0, cpu: 5 } hitcount: 2 { common_cpu: 1, cpu: 1 } hitcount: 4 { common_cpu: 6, cpu: 6 } hitcount: 4 { common_cpu: 5, cpu: 5 } hitcount: 14 { common_cpu: 4, cpu: 4 } hitcount: 26 { common_cpu: 0, cpu: 0 } hitcount: 39 { common_cpu: 2, cpu: 2 } hitcount: 184 Now for backward compatibility, I added a trick. If "cpu" is used, and the field is not found, it will fall back to "common_cpu" and work as it did before. This way, it will still work for old programs that use "cpu" to get the actual CPU, but if the event has a "cpu" as a field, it will get that event's "cpu" field, which is probably what it wants anyway. I updated the tracefs/README to include documentation about both the common_timestamp and the common_cpu. This way, if that text is present in the README, then an application can know that common_cpu is supported over just plain "cpu". Link: https://lkml.kernel.org/r/20210721110053.26b4f641@oasis.local.home Cc: Namhyung Kim <namhyung@kernel.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: stable@vger.kernel.org Fixes: 8b7622bf94a44 ("tracing: Add cpu field for hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-21 15:00:53 +00:00
} else {
hist_err(tr, HIST_ERR_FIELD_NOT_FOUND,
errpos(field_name));
field = ERR_PTR(-EINVAL);
goto out;
}
}
}
out:
kfree(str);
return field;
}
static struct hist_field *create_alias(struct hist_trigger_data *hist_data,
struct hist_field *var_ref,
char *var_name)
{
struct hist_field *alias = NULL;
unsigned long flags = HIST_FIELD_FL_ALIAS | HIST_FIELD_FL_VAR;
alias = create_hist_field(hist_data, NULL, flags, var_name);
if (!alias)
return NULL;
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
alias->fn_num = var_ref->fn_num;
alias->operands[0] = var_ref;
if (init_var_ref(alias, var_ref, var_ref->system, var_ref->event_name)) {
destroy_hist_field(alias, 0);
return NULL;
}
tracing: Make sure variable reference alias has correct var_ref_idx Original changelog from Steve Rostedt (except last sentence which explains the problem, and the Fixes: tag): I performed a three way histogram with the following commands: echo 'irq_lat u64 lat pid_t pid' > synthetic_events echo 'wake_lat u64 lat u64 irqlat pid_t pid' >> synthetic_events echo 'hist:keys=common_pid:irqts=common_timestamp.usecs if function == 0xffffffff81200580' > events/timer/hrtimer_start/trigger echo 'hist:keys=common_pid:lat=common_timestamp.usecs-$irqts:onmatch(timer.hrtimer_start).irq_lat($lat,pid) if common_flags & 1' > events/sched/sched_waking/trigger echo 'hist:keys=pid:wakets=common_timestamp.usecs,irqlat=lat' > events/synthetic/irq_lat/trigger echo 'hist:keys=next_pid:lat=common_timestamp.usecs-$wakets,irqlat=$irqlat:onmatch(synthetic.irq_lat).wake_lat($lat,$irqlat,next_pid)' > events/sched/sched_switch/trigger echo 1 > events/synthetic/wake_lat/enable Basically I wanted to see: hrtimer_start (calling function tick_sched_timer) Note: # grep tick_sched_timer /proc/kallsyms ffffffff81200580 t tick_sched_timer And save the time of that, and then record sched_waking if it is called in interrupt context and with the same pid as the hrtimer_start, it will record the latency between that and the waking event. I then look at when the task that is woken is scheduled in, and record the latency between the wakeup and the task running. At the end, the wake_lat synthetic event will show the wakeup to scheduled latency, as well as the irq latency in from hritmer_start to the wakeup. The problem is that I found this: <idle>-0 [007] d... 190.485261: wake_lat: lat=27 irqlat=190485230 pid=698 <idle>-0 [005] d... 190.485283: wake_lat: lat=40 irqlat=190485239 pid=10 <idle>-0 [002] d... 190.488327: wake_lat: lat=56 irqlat=190488266 pid=335 <idle>-0 [005] d... 190.489330: wake_lat: lat=64 irqlat=190489262 pid=10 <idle>-0 [003] d... 190.490312: wake_lat: lat=43 irqlat=190490265 pid=77 <idle>-0 [005] d... 190.493322: wake_lat: lat=54 irqlat=190493262 pid=10 <idle>-0 [005] d... 190.497305: wake_lat: lat=35 irqlat=190497267 pid=10 <idle>-0 [005] d... 190.501319: wake_lat: lat=50 irqlat=190501264 pid=10 The irqlat seemed quite large! Investigating this further, if I had enabled the irq_lat synthetic event, I noticed this: <idle>-0 [002] d.s. 249.429308: irq_lat: lat=164968 pid=335 <idle>-0 [002] d... 249.429369: wake_lat: lat=55 irqlat=249429308 pid=335 Notice that the timestamp of the irq_lat "249.429308" is awfully similar to the reported irqlat variable. In fact, all instances were like this. It appeared that: irqlat=$irqlat Wasn't assigning the old $irqlat to the new irqlat variable, but instead was assigning the $irqts to it. The issue is that assigning the old $irqlat to the new irqlat variable creates a variable reference alias, but the alias creation code forgets to make sure the alias uses the same var_ref_idx to access the reference. Link: http://lkml.kernel.org/r/1567375321.5282.12.camel@kernel.org Cc: Linux Trace Devel <linux-trace-devel@vger.kernel.org> Cc: linux-rt-users <linux-rt-users@vger.kernel.org> Cc: stable@vger.kernel.org Fixes: 7e8b88a30b085 ("tracing: Add hist trigger support for variable reference aliases") Reported-by: Steven Rostedt (VMware) <rostedt@goodmis.org> Signed-off-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-09-01 22:02:01 +00:00
alias->var_ref_idx = var_ref->var_ref_idx;
return alias;
}
tracing: Add support for creating hist trigger variables from literal Currently hist trigger expressions don't support the use of numeric literals: e.g. echo 'hist:keys=common_pid:x=$y-1234' --> is not valid expression syntax Having the ability to use numeric constants in hist triggers supports a wider range of expressions for creating variables. Add support for creating trace event histogram variables from numeric literals. e.g. echo 'hist:keys=common_pid:x=1234,y=size-1024' >> event/trigger A negative numeric constant is created, using unary minus operator (parentheses are required). e.g. echo 'hist:keys=common_pid:z=-(2)' >> event/trigger Constants can be used with division/multiplication (added in the next patch in this series) to implement granularity filters for frequent trace events. For instance we can limit emitting the rss_stat trace event to when there is a 512KB cross over in the rss size: # Create a synthetic event to monitor instead of the high frequency # rss_stat event echo 'rss_stat_throttled unsigned int mm_id; unsigned int curr; int member; long size' >> tracing/synthetic_events # Create a hist trigger that emits the synthetic rss_stat_throttled # event only when the rss size crosses a 512KB boundary. echo 'hist:keys=keys=mm_id,member:bucket=size/0x80000:onchange($bucket) .rss_stat_throttled(mm_id,curr,member,size)' >> events/kmem/rss_stat/trigger A use case for using constants with addition/subtraction is not yet known, but for completeness the use of constants are supported for all operators. Link: https://lkml.kernel.org/r/20211025200852.3002369-2-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-10-25 20:08:33 +00:00
static struct hist_field *parse_const(struct hist_trigger_data *hist_data,
char *str, char *var_name,
unsigned long *flags)
{
struct trace_array *tr = hist_data->event_file->tr;
struct hist_field *field = NULL;
u64 constant;
if (kstrtoull(str, 0, &constant)) {
hist_err(tr, HIST_ERR_EXPECT_NUMBER, errpos(str));
return NULL;
}
*flags |= HIST_FIELD_FL_CONST;
field = create_hist_field(hist_data, NULL, *flags, var_name);
if (!field)
return NULL;
field->constant = constant;
return field;
}
static struct hist_field *parse_atom(struct hist_trigger_data *hist_data,
struct trace_event_file *file, char *str,
unsigned long *flags, char *var_name)
{
char *s, *ref_system = NULL, *ref_event = NULL, *ref_var = str;
struct ftrace_event_field *field = NULL;
struct hist_field *hist_field = NULL;
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
unsigned long buckets = 0;
int ret = 0;
tracing: Add support for creating hist trigger variables from literal Currently hist trigger expressions don't support the use of numeric literals: e.g. echo 'hist:keys=common_pid:x=$y-1234' --> is not valid expression syntax Having the ability to use numeric constants in hist triggers supports a wider range of expressions for creating variables. Add support for creating trace event histogram variables from numeric literals. e.g. echo 'hist:keys=common_pid:x=1234,y=size-1024' >> event/trigger A negative numeric constant is created, using unary minus operator (parentheses are required). e.g. echo 'hist:keys=common_pid:z=-(2)' >> event/trigger Constants can be used with division/multiplication (added in the next patch in this series) to implement granularity filters for frequent trace events. For instance we can limit emitting the rss_stat trace event to when there is a 512KB cross over in the rss size: # Create a synthetic event to monitor instead of the high frequency # rss_stat event echo 'rss_stat_throttled unsigned int mm_id; unsigned int curr; int member; long size' >> tracing/synthetic_events # Create a hist trigger that emits the synthetic rss_stat_throttled # event only when the rss size crosses a 512KB boundary. echo 'hist:keys=keys=mm_id,member:bucket=size/0x80000:onchange($bucket) .rss_stat_throttled(mm_id,curr,member,size)' >> events/kmem/rss_stat/trigger A use case for using constants with addition/subtraction is not yet known, but for completeness the use of constants are supported for all operators. Link: https://lkml.kernel.org/r/20211025200852.3002369-2-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-10-25 20:08:33 +00:00
if (isdigit(str[0])) {
hist_field = parse_const(hist_data, str, var_name, flags);
if (!hist_field) {
ret = -EINVAL;
goto out;
}
return hist_field;
}
s = strchr(str, '.');
if (s) {
s = strchr(++s, '.');
if (s) {
ref_system = strsep(&str, ".");
if (!str) {
ret = -EINVAL;
goto out;
}
ref_event = strsep(&str, ".");
if (!str) {
ret = -EINVAL;
goto out;
}
ref_var = str;
}
}
s = local_field_var_ref(hist_data, ref_system, ref_event, ref_var);
if (!s) {
hist_field = parse_var_ref(hist_data, ref_system,
ref_event, ref_var);
if (hist_field) {
if (var_name) {
hist_field = create_alias(hist_data, hist_field, var_name);
if (!hist_field) {
ret = -ENOMEM;
goto out;
}
}
return hist_field;
}
} else
str = s;
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
field = parse_field(hist_data, file, str, flags, &buckets);
if (IS_ERR(field)) {
ret = PTR_ERR(field);
goto out;
}
hist_field = create_hist_field(hist_data, field, *flags, var_name);
if (!hist_field) {
ret = -ENOMEM;
goto out;
}
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
hist_field->buckets = buckets;
return hist_field;
out:
return ERR_PTR(ret);
}
static struct hist_field *parse_expr(struct hist_trigger_data *hist_data,
struct trace_event_file *file,
char *str, unsigned long flags,
char *var_name, unsigned int *n_subexprs);
static struct hist_field *parse_unary(struct hist_trigger_data *hist_data,
struct trace_event_file *file,
char *str, unsigned long flags,
char *var_name, unsigned int *n_subexprs)
{
struct hist_field *operand1, *expr = NULL;
unsigned long operand_flags;
int ret = 0;
char *s;
/* Unary minus operator, increment n_subexprs */
++*n_subexprs;
/* we support only -(xxx) i.e. explicit parens required */
if (*n_subexprs > 3) {
hist_err(file->tr, HIST_ERR_TOO_MANY_SUBEXPR, errpos(str));
ret = -EINVAL;
goto free;
}
str++; /* skip leading '-' */
s = strchr(str, '(');
if (s)
str++;
else {
ret = -EINVAL;
goto free;
}
s = strrchr(str, ')');
if (s) {
/* unary minus not supported in sub-expressions */
if (*(s+1) != '\0') {
hist_err(file->tr, HIST_ERR_UNARY_MINUS_SUBEXPR,
errpos(str));
ret = -EINVAL;
goto free;
}
*s = '\0';
}
else {
ret = -EINVAL; /* no closing ')' */
goto free;
}
flags |= HIST_FIELD_FL_EXPR;
expr = create_hist_field(hist_data, NULL, flags, var_name);
if (!expr) {
ret = -ENOMEM;
goto free;
}
operand_flags = 0;
operand1 = parse_expr(hist_data, file, str, operand_flags, NULL, n_subexprs);
if (IS_ERR(operand1)) {
ret = PTR_ERR(operand1);
goto free;
}
if (operand1->flags & HIST_FIELD_FL_STRING) {
/* String type can not be the operand of unary operator. */
hist_err(file->tr, HIST_ERR_INVALID_STR_OPERAND, errpos(str));
destroy_hist_field(operand1, 0);
ret = -EINVAL;
goto free;
}
expr->flags |= operand1->flags &
(HIST_FIELD_FL_TIMESTAMP | HIST_FIELD_FL_TIMESTAMP_USECS);
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
expr->fn_num = HIST_FIELD_FN_UMINUS;
expr->operands[0] = operand1;
expr->size = operand1->size;
expr->is_signed = operand1->is_signed;
expr->operator = FIELD_OP_UNARY_MINUS;
expr->name = expr_str(expr, 0);
expr->type = kstrdup_const(operand1->type, GFP_KERNEL);
if (!expr->type) {
ret = -ENOMEM;
goto free;
}
return expr;
free:
destroy_hist_field(expr, 0);
return ERR_PTR(ret);
}
/*
* If the operands are var refs, return pointers the
* variable(s) referenced in var1 and var2, else NULL.
*/
static int check_expr_operands(struct trace_array *tr,
struct hist_field *operand1,
struct hist_field *operand2,
struct hist_field **var1,
struct hist_field **var2)
{
unsigned long operand1_flags = operand1->flags;
unsigned long operand2_flags = operand2->flags;
if ((operand1_flags & HIST_FIELD_FL_VAR_REF) ||
(operand1_flags & HIST_FIELD_FL_ALIAS)) {
struct hist_field *var;
var = find_var_field(operand1->var.hist_data, operand1->name);
if (!var)
return -EINVAL;
operand1_flags = var->flags;
*var1 = var;
}
if ((operand2_flags & HIST_FIELD_FL_VAR_REF) ||
(operand2_flags & HIST_FIELD_FL_ALIAS)) {
struct hist_field *var;
var = find_var_field(operand2->var.hist_data, operand2->name);
if (!var)
return -EINVAL;
operand2_flags = var->flags;
*var2 = var;
}
if ((operand1_flags & HIST_FIELD_FL_TIMESTAMP_USECS) !=
(operand2_flags & HIST_FIELD_FL_TIMESTAMP_USECS)) {
hist_err(tr, HIST_ERR_TIMESTAMP_MISMATCH, 0);
return -EINVAL;
}
return 0;
}
static struct hist_field *parse_expr(struct hist_trigger_data *hist_data,
struct trace_event_file *file,
char *str, unsigned long flags,
char *var_name, unsigned int *n_subexprs)
{
struct hist_field *operand1 = NULL, *operand2 = NULL, *expr = NULL;
struct hist_field *var1 = NULL, *var2 = NULL;
unsigned long operand_flags, operand2_flags;
int field_op, ret = -EINVAL;
char *sep, *operand1_str;
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
enum hist_field_fn op_fn;
bool combine_consts;
if (*n_subexprs > 3) {
hist_err(file->tr, HIST_ERR_TOO_MANY_SUBEXPR, errpos(str));
return ERR_PTR(-EINVAL);
}
field_op = contains_operator(str, &sep);
if (field_op == FIELD_OP_NONE)
return parse_atom(hist_data, file, str, &flags, var_name);
if (field_op == FIELD_OP_UNARY_MINUS)
return parse_unary(hist_data, file, str, flags, var_name, n_subexprs);
/* Binary operator found, increment n_subexprs */
++*n_subexprs;
/* Split the expression string at the root operator */
if (!sep)
return ERR_PTR(-EINVAL);
*sep = '\0';
operand1_str = str;
str = sep+1;
/* Binary operator requires both operands */
if (*operand1_str == '\0' || *str == '\0')
return ERR_PTR(-EINVAL);
operand_flags = 0;
/* LHS of string is an expression e.g. a+b in a+b+c */
operand1 = parse_expr(hist_data, file, operand1_str, operand_flags, NULL, n_subexprs);
if (IS_ERR(operand1))
return ERR_CAST(operand1);
if (operand1->flags & HIST_FIELD_FL_STRING) {
hist_err(file->tr, HIST_ERR_INVALID_STR_OPERAND, errpos(operand1_str));
ret = -EINVAL;
goto free_op1;
}
/* RHS of string is another expression e.g. c in a+b+c */
operand_flags = 0;
operand2 = parse_expr(hist_data, file, str, operand_flags, NULL, n_subexprs);
if (IS_ERR(operand2)) {
ret = PTR_ERR(operand2);
goto free_op1;
}
if (operand2->flags & HIST_FIELD_FL_STRING) {
hist_err(file->tr, HIST_ERR_INVALID_STR_OPERAND, errpos(str));
ret = -EINVAL;
goto free_operands;
}
switch (field_op) {
case FIELD_OP_MINUS:
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
op_fn = HIST_FIELD_FN_MINUS;
break;
case FIELD_OP_PLUS:
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
op_fn = HIST_FIELD_FN_PLUS;
break;
case FIELD_OP_DIV:
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
op_fn = HIST_FIELD_FN_DIV;
break;
case FIELD_OP_MULT:
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
op_fn = HIST_FIELD_FN_MULT;
break;
default:
ret = -EINVAL;
goto free_operands;
}
ret = check_expr_operands(file->tr, operand1, operand2, &var1, &var2);
if (ret)
goto free_operands;
operand_flags = var1 ? var1->flags : operand1->flags;
operand2_flags = var2 ? var2->flags : operand2->flags;
/*
* If both operands are constant, the expression can be
* collapsed to a single constant.
*/
combine_consts = operand_flags & operand2_flags & HIST_FIELD_FL_CONST;
flags |= combine_consts ? HIST_FIELD_FL_CONST : HIST_FIELD_FL_EXPR;
flags |= operand1->flags &
(HIST_FIELD_FL_TIMESTAMP | HIST_FIELD_FL_TIMESTAMP_USECS);
expr = create_hist_field(hist_data, NULL, flags, var_name);
if (!expr) {
ret = -ENOMEM;
goto free_operands;
}
operand1->read_once = true;
operand2->read_once = true;
/* The operands are now owned and free'd by 'expr' */
expr->operands[0] = operand1;
expr->operands[1] = operand2;
tracing / histogram: Give calculation hist_fields a size When working on my user space applications, I found a bug in the synthetic event code where the automated synthetic event field was not matching the event field calculation it was attached to. Looking deeper into it, it was because the calculation hist_field was not given a size. The synthetic event fields are matched to their hist_fields either by having the field have an identical string type, or if that does not match, then the size and signed values are used to match the fields. The problem arose when I tried to match a calculation where the fields were "unsigned int". My tool created a synthetic event of type "u32". But it failed to match. The string was: diff=field1-field2:onmatch(event).trace(synth,$diff) Adding debugging into the kernel, I found that the size of "diff" was 0. And since it was given "unsigned int" as a type, the histogram fallback code used size and signed. The signed matched, but the size of u32 (4) did not match zero, and the event failed to be created. This can be worse if the field you want to match is not one of the acceptable fields for a synthetic event. As event fields can have any type that is supported in Linux, this can cause an issue. For example, if a type is an enum. Then there's no way to use that with any calculations. Have the calculation field simply take on the size of what it is calculating. Link: https://lkml.kernel.org/r/20210730171951.59c7743f@oasis.local.home Cc: Tom Zanussi <zanussi@kernel.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: stable@vger.kernel.org Fixes: 100719dcef447 ("tracing: Add simple expression support to hist triggers") Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-30 21:19:51 +00:00
if (field_op == FIELD_OP_DIV &&
operand2_flags & HIST_FIELD_FL_CONST) {
u64 divisor = var2 ? var2->constant : operand2->constant;
if (!divisor) {
hist_err(file->tr, HIST_ERR_DIVISION_BY_ZERO, errpos(str));
ret = -EDOM;
goto free_expr;
}
/*
* Copy the divisor here so we don't have to look it up
* later if this is a var ref
*/
operand2->constant = divisor;
op_fn = hist_field_get_div_fn(operand2);
}
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
expr->fn_num = op_fn;
if (combine_consts) {
if (var1)
expr->operands[0] = var1;
if (var2)
expr->operands[1] = var2;
tracing / histogram: Give calculation hist_fields a size When working on my user space applications, I found a bug in the synthetic event code where the automated synthetic event field was not matching the event field calculation it was attached to. Looking deeper into it, it was because the calculation hist_field was not given a size. The synthetic event fields are matched to their hist_fields either by having the field have an identical string type, or if that does not match, then the size and signed values are used to match the fields. The problem arose when I tried to match a calculation where the fields were "unsigned int". My tool created a synthetic event of type "u32". But it failed to match. The string was: diff=field1-field2:onmatch(event).trace(synth,$diff) Adding debugging into the kernel, I found that the size of "diff" was 0. And since it was given "unsigned int" as a type, the histogram fallback code used size and signed. The signed matched, but the size of u32 (4) did not match zero, and the event failed to be created. This can be worse if the field you want to match is not one of the acceptable fields for a synthetic event. As event fields can have any type that is supported in Linux, this can cause an issue. For example, if a type is an enum. Then there's no way to use that with any calculations. Have the calculation field simply take on the size of what it is calculating. Link: https://lkml.kernel.org/r/20210730171951.59c7743f@oasis.local.home Cc: Tom Zanussi <zanussi@kernel.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: stable@vger.kernel.org Fixes: 100719dcef447 ("tracing: Add simple expression support to hist triggers") Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-30 21:19:51 +00:00
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
expr->constant = hist_fn_call(expr, NULL, NULL, NULL, NULL);
expr->fn_num = HIST_FIELD_FN_CONST;
expr->operands[0] = NULL;
expr->operands[1] = NULL;
/*
* var refs won't be destroyed immediately
* See: destroy_hist_field()
*/
destroy_hist_field(operand2, 0);
destroy_hist_field(operand1, 0);
expr->name = expr_str(expr, 0);
} else {
/* The operand sizes should be the same, so just pick one */
expr->size = operand1->size;
expr->is_signed = operand1->is_signed;
expr->operator = field_op;
expr->type = kstrdup_const(operand1->type, GFP_KERNEL);
if (!expr->type) {
ret = -ENOMEM;
goto free_expr;
}
expr->name = expr_str(expr, 0);
}
return expr;
free_operands:
destroy_hist_field(operand2, 0);
free_op1:
destroy_hist_field(operand1, 0);
return ERR_PTR(ret);
free_expr:
destroy_hist_field(expr, 0);
return ERR_PTR(ret);
}
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
static char *find_trigger_filter(struct hist_trigger_data *hist_data,
struct trace_event_file *file)
{
struct event_trigger_data *test;
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
lockdep_assert_held(&event_mutex);
list_for_each_entry(test, &file->triggers, list) {
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
if (test->cmd_ops->trigger_type == ETT_EVENT_HIST) {
if (test->private_data == hist_data)
return test->filter_str;
}
}
return NULL;
}
static struct event_command trigger_hist_cmd;
static int event_hist_trigger_parse(struct event_command *cmd_ops,
struct trace_event_file *file,
char *glob, char *cmd,
char *param_and_filter);
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
static bool compatible_keys(struct hist_trigger_data *target_hist_data,
struct hist_trigger_data *hist_data,
unsigned int n_keys)
{
struct hist_field *target_hist_field, *hist_field;
unsigned int n, i, j;
if (hist_data->n_fields - hist_data->n_vals != n_keys)
return false;
i = hist_data->n_vals;
j = target_hist_data->n_vals;
for (n = 0; n < n_keys; n++) {
hist_field = hist_data->fields[i + n];
target_hist_field = target_hist_data->fields[j + n];
if (strcmp(hist_field->type, target_hist_field->type) != 0)
return false;
if (hist_field->size != target_hist_field->size)
return false;
if (hist_field->is_signed != target_hist_field->is_signed)
return false;
}
return true;
}
static struct hist_trigger_data *
find_compatible_hist(struct hist_trigger_data *target_hist_data,
struct trace_event_file *file)
{
struct hist_trigger_data *hist_data;
struct event_trigger_data *test;
unsigned int n_keys;
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
lockdep_assert_held(&event_mutex);
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
n_keys = target_hist_data->n_fields - target_hist_data->n_vals;
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
list_for_each_entry(test, &file->triggers, list) {
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
if (test->cmd_ops->trigger_type == ETT_EVENT_HIST) {
hist_data = test->private_data;
if (compatible_keys(target_hist_data, hist_data, n_keys))
return hist_data;
}
}
return NULL;
}
static struct trace_event_file *event_file(struct trace_array *tr,
char *system, char *event_name)
{
struct trace_event_file *file;
file = __find_event_file(tr, system, event_name);
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
if (!file)
return ERR_PTR(-EINVAL);
return file;
}
static struct hist_field *
find_synthetic_field_var(struct hist_trigger_data *target_hist_data,
char *system, char *event_name, char *field_name)
{
struct hist_field *event_var;
char *synthetic_name;
synthetic_name = kzalloc(MAX_FILTER_STR_VAL, GFP_KERNEL);
if (!synthetic_name)
return ERR_PTR(-ENOMEM);
strcpy(synthetic_name, "synthetic_");
strcat(synthetic_name, field_name);
event_var = find_event_var(target_hist_data, system, event_name, synthetic_name);
kfree(synthetic_name);
return event_var;
}
/**
* create_field_var_hist - Automatically create a histogram and var for a field
* @target_hist_data: The target hist trigger
* @subsys_name: Optional subsystem name
* @event_name: Optional event name
* @field_name: The name of the field (and the resulting variable)
*
* Hist trigger actions fetch data from variables, not directly from
* events. However, for convenience, users are allowed to directly
* specify an event field in an action, which will be automatically
* converted into a variable on their behalf.
*
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
* If a user specifies a field on an event that isn't the event the
* histogram currently being defined (the target event histogram), the
* only way that can be accomplished is if a new hist trigger is
* created and the field variable defined on that.
*
* This function creates a new histogram compatible with the target
* event (meaning a histogram with the same key as the target
* histogram), and creates a variable for the specified field, but
* with 'synthetic_' prepended to the variable name in order to avoid
* collision with normal field variables.
*
* Return: The variable created for the field.
*/
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
static struct hist_field *
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
create_field_var_hist(struct hist_trigger_data *target_hist_data,
char *subsys_name, char *event_name, char *field_name)
{
struct trace_array *tr = target_hist_data->event_file->tr;
struct hist_trigger_data *hist_data;
unsigned int i, n, first = true;
struct field_var_hist *var_hist;
struct trace_event_file *file;
struct hist_field *key_field;
struct hist_field *event_var;
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
char *saved_filter;
char *cmd;
int ret;
if (target_hist_data->n_field_var_hists >= SYNTH_FIELDS_MAX) {
hist_err(tr, HIST_ERR_TOO_MANY_FIELD_VARS, errpos(field_name));
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
return ERR_PTR(-EINVAL);
}
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
file = event_file(tr, subsys_name, event_name);
if (IS_ERR(file)) {
hist_err(tr, HIST_ERR_EVENT_FILE_NOT_FOUND, errpos(field_name));
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
ret = PTR_ERR(file);
return ERR_PTR(ret);
}
/*
* Look for a histogram compatible with target. We'll use the
* found histogram specification to create a new matching
* histogram with our variable on it. target_hist_data is not
* yet a registered histogram so we can't use that.
*/
hist_data = find_compatible_hist(target_hist_data, file);
if (!hist_data) {
hist_err(tr, HIST_ERR_HIST_NOT_FOUND, errpos(field_name));
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
return ERR_PTR(-EINVAL);
}
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
/* See if a synthetic field variable has already been created */
event_var = find_synthetic_field_var(target_hist_data, subsys_name,
event_name, field_name);
if (!IS_ERR_OR_NULL(event_var))
return event_var;
var_hist = kzalloc(sizeof(*var_hist), GFP_KERNEL);
if (!var_hist)
return ERR_PTR(-ENOMEM);
cmd = kzalloc(MAX_FILTER_STR_VAL, GFP_KERNEL);
if (!cmd) {
kfree(var_hist);
return ERR_PTR(-ENOMEM);
}
/* Use the same keys as the compatible histogram */
strcat(cmd, "keys=");
for_each_hist_key_field(i, hist_data) {
key_field = hist_data->fields[i];
if (!first)
strcat(cmd, ",");
strcat(cmd, key_field->field->name);
first = false;
}
/* Create the synthetic field variable specification */
strcat(cmd, ":synthetic_");
strcat(cmd, field_name);
strcat(cmd, "=");
strcat(cmd, field_name);
/* Use the same filter as the compatible histogram */
saved_filter = find_trigger_filter(hist_data, file);
if (saved_filter) {
strcat(cmd, " if ");
strcat(cmd, saved_filter);
}
var_hist->cmd = kstrdup(cmd, GFP_KERNEL);
if (!var_hist->cmd) {
kfree(cmd);
kfree(var_hist);
return ERR_PTR(-ENOMEM);
}
/* Save the compatible histogram information */
var_hist->hist_data = hist_data;
/* Create the new histogram with our variable */
ret = event_hist_trigger_parse(&trigger_hist_cmd, file,
"", "hist", cmd);
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
if (ret) {
kfree(cmd);
kfree(var_hist->cmd);
kfree(var_hist);
hist_err(tr, HIST_ERR_HIST_CREATE_FAIL, errpos(field_name));
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
return ERR_PTR(ret);
}
kfree(cmd);
/* If we can't find the variable, something went wrong */
event_var = find_synthetic_field_var(target_hist_data, subsys_name,
event_name, field_name);
if (IS_ERR_OR_NULL(event_var)) {
kfree(var_hist->cmd);
kfree(var_hist);
hist_err(tr, HIST_ERR_SYNTH_VAR_NOT_FOUND, errpos(field_name));
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
return ERR_PTR(-EINVAL);
}
n = target_hist_data->n_field_var_hists;
target_hist_data->field_var_hists[n] = var_hist;
target_hist_data->n_field_var_hists++;
return event_var;
}
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
static struct hist_field *
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
find_target_event_var(struct hist_trigger_data *hist_data,
char *subsys_name, char *event_name, char *var_name)
{
struct trace_event_file *file = hist_data->event_file;
struct hist_field *hist_field = NULL;
if (subsys_name) {
struct trace_event_call *call;
if (!event_name)
return NULL;
call = file->event_call;
if (strcmp(subsys_name, call->class->system) != 0)
return NULL;
if (strcmp(event_name, trace_event_name(call)) != 0)
return NULL;
}
hist_field = find_var_field(hist_data, var_name);
return hist_field;
}
static inline void __update_field_vars(struct tracing_map_elt *elt,
struct trace_buffer *buffer,
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
struct ring_buffer_event *rbe,
void *rec,
struct field_var **field_vars,
unsigned int n_field_vars,
unsigned int field_var_str_start)
{
struct hist_elt_data *elt_data = elt->private_data;
unsigned int i, j, var_idx;
u64 var_val;
/* Make sure stacktrace can fit in the string variable length */
BUILD_BUG_ON((HIST_STACKTRACE_DEPTH + 1) * sizeof(long) >= STR_VAR_LEN_MAX);
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
for (i = 0, j = field_var_str_start; i < n_field_vars; i++) {
struct field_var *field_var = field_vars[i];
struct hist_field *var = field_var->var;
struct hist_field *val = field_var->val;
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
var_val = hist_fn_call(val, elt, buffer, rbe, rec);
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
var_idx = var->var.idx;
if (val->flags & (HIST_FIELD_FL_STRING |
HIST_FIELD_FL_STACKTRACE)) {
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
char *str = elt_data->field_var_str[j++];
char *val_str = (char *)(uintptr_t)var_val;
unsigned int size;
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
if (val->flags & HIST_FIELD_FL_STRING) {
size = min(val->size, STR_VAR_LEN_MAX);
strscpy(str, val_str, size);
} else {
char *stack_start = str + sizeof(unsigned long);
int e;
e = stack_trace_save((void *)stack_start,
HIST_STACKTRACE_DEPTH,
HIST_STACKTRACE_SKIP);
if (e < HIST_STACKTRACE_DEPTH - 1)
((unsigned long *)stack_start)[e] = 0;
*((unsigned long *)str) = e;
}
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
var_val = (u64)(uintptr_t)str;
}
tracing_map_set_var(elt, var_idx, var_val);
}
}
static void update_field_vars(struct hist_trigger_data *hist_data,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
struct ring_buffer_event *rbe,
void *rec)
{
__update_field_vars(elt, buffer, rbe, rec, hist_data->field_vars,
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
hist_data->n_field_vars, 0);
}
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
static void save_track_data_vars(struct hist_trigger_data *hist_data,
struct tracing_map_elt *elt,
struct trace_buffer *buffer, void *rec,
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
struct ring_buffer_event *rbe, void *key,
struct action_data *data, u64 *var_ref_vals)
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
{
__update_field_vars(elt, buffer, rbe, rec, hist_data->save_vars,
hist_data->n_save_vars, hist_data->n_field_var_str);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
}
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
static struct hist_field *create_var(struct hist_trigger_data *hist_data,
struct trace_event_file *file,
char *name, int size, const char *type)
{
struct hist_field *var;
int idx;
if (find_var(hist_data, file, name) && !hist_data->remove) {
var = ERR_PTR(-EINVAL);
goto out;
}
var = kzalloc(sizeof(struct hist_field), GFP_KERNEL);
if (!var) {
var = ERR_PTR(-ENOMEM);
goto out;
}
idx = tracing_map_add_var(hist_data->map);
if (idx < 0) {
kfree(var);
var = ERR_PTR(-EINVAL);
goto out;
}
tracing: Fix memory leaks in trace_events_hist.c kmemleak report 1: [<9092c50b>] kmem_cache_alloc_trace+0x138/0x270 [<05a2c9ed>] create_field_var+0xcf/0x180 [<528a2d68>] action_create+0xe2/0xc80 [<63f50b61>] event_hist_trigger_func+0x15b5/0x1920 [<28ea5d3d>] trigger_process_regex+0x7b/0xc0 [<3138e86f>] event_trigger_write+0x4d/0xb0 [<ffd66c19>] __vfs_write+0x30/0x200 [<4f424a0d>] vfs_write+0x96/0x1b0 [<da59a290>] ksys_write+0x53/0xc0 [<3717101a>] __ia32_sys_write+0x15/0x20 [<c5f23497>] do_fast_syscall_32+0x70/0x250 [<46e2629c>] entry_SYSENTER_32+0xaf/0x102 This is because save_vars[] of struct hist_trigger_data are not destroyed kmemleak report 2: [<9092c50b>] kmem_cache_alloc_trace+0x138/0x270 [<6e5e97c5>] create_var+0x3c/0x110 [<de82f1b9>] create_field_var+0xaf/0x180 [<528a2d68>] action_create+0xe2/0xc80 [<63f50b61>] event_hist_trigger_func+0x15b5/0x1920 [<28ea5d3d>] trigger_process_regex+0x7b/0xc0 [<3138e86f>] event_trigger_write+0x4d/0xb0 [<ffd66c19>] __vfs_write+0x30/0x200 [<4f424a0d>] vfs_write+0x96/0x1b0 [<da59a290>] ksys_write+0x53/0xc0 [<3717101a>] __ia32_sys_write+0x15/0x20 [<c5f23497>] do_fast_syscall_32+0x70/0x250 [<46e2629c>] entry_SYSENTER_32+0xaf/0x102 struct hist_field allocated through create_var() do not initialize "ref" field to 1. The code in __destroy_hist_field() does not destroy object if "ref" is initialized to zero, the condition if (--hist_field->ref > 1) always passes since unsigned int wraps. kmemleak report 3: [<f8666fcc>] __kmalloc_track_caller+0x139/0x2b0 [<bb7f80a5>] kstrdup+0x27/0x50 [<39d70006>] init_var_ref+0x58/0xd0 [<8ca76370>] create_var_ref+0x89/0xe0 [<f045fc39>] action_create+0x38f/0xc80 [<7c146821>] event_hist_trigger_func+0x15b5/0x1920 [<07de3f61>] trigger_process_regex+0x7b/0xc0 [<e87daf8f>] event_trigger_write+0x4d/0xb0 [<19bf1512>] __vfs_write+0x30/0x200 [<64ce4d27>] vfs_write+0x96/0x1b0 [<a6f34170>] ksys_write+0x53/0xc0 [<7d4230cd>] __ia32_sys_write+0x15/0x20 [<8eadca00>] do_fast_syscall_32+0x70/0x250 [<235cf985>] entry_SYSENTER_32+0xaf/0x102 hist_fields (system & event_name) are not freed Link: http://lkml.kernel.org/r/20200422061503.GA5151@cosmos Signed-off-by: Vamshi K Sthambamkadi <vamshi.k.sthambamkadi@gmail.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2020-04-22 06:15:06 +00:00
var->ref = 1;
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
var->flags = HIST_FIELD_FL_VAR;
var->var.idx = idx;
var->var.hist_data = var->hist_data = hist_data;
var->size = size;
var->var.name = kstrdup(name, GFP_KERNEL);
var->type = kstrdup_const(type, GFP_KERNEL);
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
if (!var->var.name || !var->type) {
kfree_const(var->type);
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
kfree(var->var.name);
kfree(var);
var = ERR_PTR(-ENOMEM);
}
out:
return var;
}
static struct field_var *create_field_var(struct hist_trigger_data *hist_data,
struct trace_event_file *file,
char *field_name)
{
struct hist_field *val = NULL, *var = NULL;
unsigned long flags = HIST_FIELD_FL_VAR;
struct trace_array *tr = file->tr;
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
struct field_var *field_var;
int ret = 0;
if (hist_data->n_field_vars >= SYNTH_FIELDS_MAX) {
hist_err(tr, HIST_ERR_TOO_MANY_FIELD_VARS, errpos(field_name));
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
ret = -EINVAL;
goto err;
}
val = parse_atom(hist_data, file, field_name, &flags, NULL);
if (IS_ERR(val)) {
hist_err(tr, HIST_ERR_FIELD_VAR_PARSE_FAIL, errpos(field_name));
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
ret = PTR_ERR(val);
goto err;
}
var = create_var(hist_data, file, field_name, val->size, val->type);
if (IS_ERR(var)) {
hist_err(tr, HIST_ERR_VAR_CREATE_FIND_FAIL, errpos(field_name));
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
kfree(val);
ret = PTR_ERR(var);
goto err;
}
field_var = kzalloc(sizeof(struct field_var), GFP_KERNEL);
if (!field_var) {
kfree(val);
kfree(var);
ret = -ENOMEM;
goto err;
}
field_var->var = var;
field_var->val = val;
out:
return field_var;
err:
field_var = ERR_PTR(ret);
goto out;
}
/**
* create_target_field_var - Automatically create a variable for a field
* @target_hist_data: The target hist trigger
* @subsys_name: Optional subsystem name
* @event_name: Optional event name
* @var_name: The name of the field (and the resulting variable)
*
* Hist trigger actions fetch data from variables, not directly from
* events. However, for convenience, users are allowed to directly
* specify an event field in an action, which will be automatically
* converted into a variable on their behalf.
*
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
* This function creates a field variable with the name var_name on
* the hist trigger currently being defined on the target event. If
* subsys_name and event_name are specified, this function simply
* verifies that they do in fact match the target event subsystem and
* event name.
*
* Return: The variable created for the field.
*/
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
static struct field_var *
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
create_target_field_var(struct hist_trigger_data *target_hist_data,
char *subsys_name, char *event_name, char *var_name)
{
struct trace_event_file *file = target_hist_data->event_file;
if (subsys_name) {
struct trace_event_call *call;
if (!event_name)
return NULL;
call = file->event_call;
if (strcmp(subsys_name, call->class->system) != 0)
return NULL;
if (strcmp(event_name, trace_event_name(call)) != 0)
return NULL;
}
return create_field_var(target_hist_data, file, var_name);
}
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
static bool check_track_val_max(u64 track_val, u64 var_val)
{
if (var_val <= track_val)
return false;
return true;
}
static bool check_track_val_changed(u64 track_val, u64 var_val)
{
if (var_val == track_val)
return false;
return true;
}
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
static u64 get_track_val(struct hist_trigger_data *hist_data,
struct tracing_map_elt *elt,
struct action_data *data)
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
{
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
unsigned int track_var_idx = data->track_data.track_var->var.idx;
u64 track_val;
track_val = tracing_map_read_var(elt, track_var_idx);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
return track_val;
}
static void save_track_val(struct hist_trigger_data *hist_data,
struct tracing_map_elt *elt,
struct action_data *data, u64 var_val)
{
unsigned int track_var_idx = data->track_data.track_var->var.idx;
tracing_map_set_var(elt, track_var_idx, var_val);
}
static void save_track_data(struct hist_trigger_data *hist_data,
struct tracing_map_elt *elt,
struct trace_buffer *buffer, void *rec,
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
struct ring_buffer_event *rbe, void *key,
struct action_data *data, u64 *var_ref_vals)
{
if (data->track_data.save_data)
data->track_data.save_data(hist_data, elt, buffer, rec, rbe,
key, data, var_ref_vals);
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
}
static bool check_track_val(struct tracing_map_elt *elt,
struct action_data *data,
u64 var_val)
{
struct hist_trigger_data *hist_data;
u64 track_val;
hist_data = data->track_data.track_var->hist_data;
track_val = get_track_val(hist_data, elt, data);
return data->track_data.check_val(track_val, var_val);
}
#ifdef CONFIG_TRACER_SNAPSHOT
static bool cond_snapshot_update(struct trace_array *tr, void *cond_data)
{
/* called with tr->max_lock held */
struct track_data *track_data = tr->cond_snapshot->cond_data;
struct hist_elt_data *elt_data, *track_elt_data;
struct snapshot_context *context = cond_data;
tracing: Add a check_val() check before updating cond_snapshot() track_val Without this check a snapshot is taken whenever a bucket's max is hit, rather than only when the global max is hit, as it should be. Before: In this example, we do a first run of the workload (cyclictest), examine the output, note the max ('triggering value') (347), then do a second run and note the max again. In this case, the max in the second run (39) is below the max in the first run, but since we haven't cleared the histogram, the first max is still in the histogram and is higher than any other max, so it should still be the max for the snapshot. It isn't however - the value should still be 347 after the second run. # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0:onmax($wakeup_lat).save(next_prio,next_comm,prev_pid,prev_prio,prev_comm):onmax($wakeup_lat).snapshot() if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # cyclictest -p 80 -n -s -t 2 -D 2 # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 2143 } hitcount: 199 max: 44 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/4 { next_pid: 2145 } hitcount: 1325 max: 38 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/2 { next_pid: 2144 } hitcount: 1982 max: 347 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 Snapshot taken (see tracing/snapshot). Details: triggering value { onmax($wakeup_lat) }: 347 triggered by event with key: { next_pid: 2144 } # cyclictest -p 80 -n -s -t 2 -D 2 # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 2143 } hitcount: 199 max: 44 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/4 { next_pid: 2148 } hitcount: 199 max: 16 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/1 { next_pid: 2145 } hitcount: 1325 max: 38 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/2 { next_pid: 2150 } hitcount: 1326 max: 39 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/4 { next_pid: 2144 } hitcount: 1982 max: 347 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 { next_pid: 2149 } hitcount: 1983 max: 130 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/0 Snapshot taken (see tracing/snapshot). Details: triggering value { onmax($wakeup_lat) }: 39 triggered by event with key: { next_pid: 2150 } After: In this example, we do a first run of the workload (cyclictest), examine the output, note the max ('triggering value') (375), then do a second run and note the max again. In this case, the max in the second run is still 375, the highest in any bucket, as it should be. # cyclictest -p 80 -n -s -t 2 -D 2 # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 2072 } hitcount: 200 max: 28 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/5 { next_pid: 2074 } hitcount: 1323 max: 375 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/2 { next_pid: 2073 } hitcount: 1980 max: 153 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 Snapshot taken (see tracing/snapshot). Details: triggering value { onmax($wakeup_lat) }: 375 triggered by event with key: { next_pid: 2074 } # cyclictest -p 80 -n -s -t 2 -D 2 # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 2101 } hitcount: 199 max: 49 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 { next_pid: 2072 } hitcount: 200 max: 28 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/5 { next_pid: 2074 } hitcount: 1323 max: 375 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/2 { next_pid: 2103 } hitcount: 1325 max: 74 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/4 { next_pid: 2073 } hitcount: 1980 max: 153 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 { next_pid: 2102 } hitcount: 1981 max: 84 next_prio: 19 next_comm: cyclictest prev_pid: 12 prev_prio: 120 prev_comm: kworker/0:1 Snapshot taken (see tracing/snapshot). Details: triggering value { onmax($wakeup_lat) }: 375 triggered by event with key: { next_pid: 2074 } Link: http://lkml.kernel.org/r/95958351329f129c07504b4d1769c47a97b70d65.1555597045.git.tom.zanussi@linux.intel.com Cc: stable@vger.kernel.org Fixes: a3785b7eca8fd ("tracing: Add hist trigger snapshot() action") Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-04-18 15:18:52 +00:00
struct action_data *action;
u64 track_val;
if (!track_data)
return false;
tracing: Add a check_val() check before updating cond_snapshot() track_val Without this check a snapshot is taken whenever a bucket's max is hit, rather than only when the global max is hit, as it should be. Before: In this example, we do a first run of the workload (cyclictest), examine the output, note the max ('triggering value') (347), then do a second run and note the max again. In this case, the max in the second run (39) is below the max in the first run, but since we haven't cleared the histogram, the first max is still in the histogram and is higher than any other max, so it should still be the max for the snapshot. It isn't however - the value should still be 347 after the second run. # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0:onmax($wakeup_lat).save(next_prio,next_comm,prev_pid,prev_prio,prev_comm):onmax($wakeup_lat).snapshot() if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # cyclictest -p 80 -n -s -t 2 -D 2 # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 2143 } hitcount: 199 max: 44 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/4 { next_pid: 2145 } hitcount: 1325 max: 38 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/2 { next_pid: 2144 } hitcount: 1982 max: 347 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 Snapshot taken (see tracing/snapshot). Details: triggering value { onmax($wakeup_lat) }: 347 triggered by event with key: { next_pid: 2144 } # cyclictest -p 80 -n -s -t 2 -D 2 # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 2143 } hitcount: 199 max: 44 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/4 { next_pid: 2148 } hitcount: 199 max: 16 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/1 { next_pid: 2145 } hitcount: 1325 max: 38 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/2 { next_pid: 2150 } hitcount: 1326 max: 39 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/4 { next_pid: 2144 } hitcount: 1982 max: 347 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 { next_pid: 2149 } hitcount: 1983 max: 130 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/0 Snapshot taken (see tracing/snapshot). Details: triggering value { onmax($wakeup_lat) }: 39 triggered by event with key: { next_pid: 2150 } After: In this example, we do a first run of the workload (cyclictest), examine the output, note the max ('triggering value') (375), then do a second run and note the max again. In this case, the max in the second run is still 375, the highest in any bucket, as it should be. # cyclictest -p 80 -n -s -t 2 -D 2 # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 2072 } hitcount: 200 max: 28 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/5 { next_pid: 2074 } hitcount: 1323 max: 375 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/2 { next_pid: 2073 } hitcount: 1980 max: 153 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 Snapshot taken (see tracing/snapshot). Details: triggering value { onmax($wakeup_lat) }: 375 triggered by event with key: { next_pid: 2074 } # cyclictest -p 80 -n -s -t 2 -D 2 # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 2101 } hitcount: 199 max: 49 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 { next_pid: 2072 } hitcount: 200 max: 28 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/5 { next_pid: 2074 } hitcount: 1323 max: 375 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/2 { next_pid: 2103 } hitcount: 1325 max: 74 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/4 { next_pid: 2073 } hitcount: 1980 max: 153 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 { next_pid: 2102 } hitcount: 1981 max: 84 next_prio: 19 next_comm: cyclictest prev_pid: 12 prev_prio: 120 prev_comm: kworker/0:1 Snapshot taken (see tracing/snapshot). Details: triggering value { onmax($wakeup_lat) }: 375 triggered by event with key: { next_pid: 2074 } Link: http://lkml.kernel.org/r/95958351329f129c07504b4d1769c47a97b70d65.1555597045.git.tom.zanussi@linux.intel.com Cc: stable@vger.kernel.org Fixes: a3785b7eca8fd ("tracing: Add hist trigger snapshot() action") Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-04-18 15:18:52 +00:00
action = track_data->action_data;
track_val = get_track_val(track_data->hist_data, context->elt,
track_data->action_data);
tracing: Add a check_val() check before updating cond_snapshot() track_val Without this check a snapshot is taken whenever a bucket's max is hit, rather than only when the global max is hit, as it should be. Before: In this example, we do a first run of the workload (cyclictest), examine the output, note the max ('triggering value') (347), then do a second run and note the max again. In this case, the max in the second run (39) is below the max in the first run, but since we haven't cleared the histogram, the first max is still in the histogram and is higher than any other max, so it should still be the max for the snapshot. It isn't however - the value should still be 347 after the second run. # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0:onmax($wakeup_lat).save(next_prio,next_comm,prev_pid,prev_prio,prev_comm):onmax($wakeup_lat).snapshot() if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # cyclictest -p 80 -n -s -t 2 -D 2 # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 2143 } hitcount: 199 max: 44 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/4 { next_pid: 2145 } hitcount: 1325 max: 38 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/2 { next_pid: 2144 } hitcount: 1982 max: 347 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 Snapshot taken (see tracing/snapshot). Details: triggering value { onmax($wakeup_lat) }: 347 triggered by event with key: { next_pid: 2144 } # cyclictest -p 80 -n -s -t 2 -D 2 # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 2143 } hitcount: 199 max: 44 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/4 { next_pid: 2148 } hitcount: 199 max: 16 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/1 { next_pid: 2145 } hitcount: 1325 max: 38 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/2 { next_pid: 2150 } hitcount: 1326 max: 39 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/4 { next_pid: 2144 } hitcount: 1982 max: 347 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 { next_pid: 2149 } hitcount: 1983 max: 130 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/0 Snapshot taken (see tracing/snapshot). Details: triggering value { onmax($wakeup_lat) }: 39 triggered by event with key: { next_pid: 2150 } After: In this example, we do a first run of the workload (cyclictest), examine the output, note the max ('triggering value') (375), then do a second run and note the max again. In this case, the max in the second run is still 375, the highest in any bucket, as it should be. # cyclictest -p 80 -n -s -t 2 -D 2 # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 2072 } hitcount: 200 max: 28 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/5 { next_pid: 2074 } hitcount: 1323 max: 375 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/2 { next_pid: 2073 } hitcount: 1980 max: 153 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 Snapshot taken (see tracing/snapshot). Details: triggering value { onmax($wakeup_lat) }: 375 triggered by event with key: { next_pid: 2074 } # cyclictest -p 80 -n -s -t 2 -D 2 # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 2101 } hitcount: 199 max: 49 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 { next_pid: 2072 } hitcount: 200 max: 28 next_prio: 120 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/5 { next_pid: 2074 } hitcount: 1323 max: 375 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/2 { next_pid: 2103 } hitcount: 1325 max: 74 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/4 { next_pid: 2073 } hitcount: 1980 max: 153 next_prio: 19 next_comm: cyclictest prev_pid: 0 prev_prio: 120 prev_comm: swapper/6 { next_pid: 2102 } hitcount: 1981 max: 84 next_prio: 19 next_comm: cyclictest prev_pid: 12 prev_prio: 120 prev_comm: kworker/0:1 Snapshot taken (see tracing/snapshot). Details: triggering value { onmax($wakeup_lat) }: 375 triggered by event with key: { next_pid: 2074 } Link: http://lkml.kernel.org/r/95958351329f129c07504b4d1769c47a97b70d65.1555597045.git.tom.zanussi@linux.intel.com Cc: stable@vger.kernel.org Fixes: a3785b7eca8fd ("tracing: Add hist trigger snapshot() action") Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-04-18 15:18:52 +00:00
if (!action->track_data.check_val(track_data->track_val, track_val))
return false;
track_data->track_val = track_val;
memcpy(track_data->key, context->key, track_data->key_len);
elt_data = context->elt->private_data;
track_elt_data = track_data->elt.private_data;
if (elt_data->comm)
strncpy(track_elt_data->comm, elt_data->comm, TASK_COMM_LEN);
track_data->updated = true;
return true;
}
static void save_track_data_snapshot(struct hist_trigger_data *hist_data,
struct tracing_map_elt *elt,
struct trace_buffer *buffer, void *rec,
struct ring_buffer_event *rbe, void *key,
struct action_data *data,
u64 *var_ref_vals)
{
struct trace_event_file *file = hist_data->event_file;
struct snapshot_context context;
context.elt = elt;
context.key = key;
tracing_snapshot_cond(file->tr, &context);
}
static void hist_trigger_print_key(struct seq_file *m,
struct hist_trigger_data *hist_data,
void *key,
struct tracing_map_elt *elt);
static struct action_data *snapshot_action(struct hist_trigger_data *hist_data)
{
unsigned int i;
if (!hist_data->n_actions)
return NULL;
for (i = 0; i < hist_data->n_actions; i++) {
struct action_data *data = hist_data->actions[i];
if (data->action == ACTION_SNAPSHOT)
return data;
}
return NULL;
}
static void track_data_snapshot_print(struct seq_file *m,
struct hist_trigger_data *hist_data)
{
struct trace_event_file *file = hist_data->event_file;
struct track_data *track_data;
struct action_data *action;
track_data = tracing_cond_snapshot_data(file->tr);
if (!track_data)
return;
if (!track_data->updated)
return;
action = snapshot_action(hist_data);
if (!action)
return;
seq_puts(m, "\nSnapshot taken (see tracing/snapshot). Details:\n");
seq_printf(m, "\ttriggering value { %s(%s) }: %10llu",
action->handler == HANDLER_ONMAX ? "onmax" : "onchange",
action->track_data.var_str, track_data->track_val);
seq_puts(m, "\ttriggered by event with key: ");
hist_trigger_print_key(m, hist_data, track_data->key, &track_data->elt);
seq_putc(m, '\n');
}
#else
static bool cond_snapshot_update(struct trace_array *tr, void *cond_data)
{
return false;
}
static void save_track_data_snapshot(struct hist_trigger_data *hist_data,
struct tracing_map_elt *elt,
struct trace_buffer *buffer, void *rec,
struct ring_buffer_event *rbe, void *key,
struct action_data *data,
u64 *var_ref_vals) {}
static void track_data_snapshot_print(struct seq_file *m,
struct hist_trigger_data *hist_data) {}
#endif /* CONFIG_TRACER_SNAPSHOT */
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
static void track_data_print(struct seq_file *m,
struct hist_trigger_data *hist_data,
struct tracing_map_elt *elt,
struct action_data *data)
{
u64 track_val = get_track_val(hist_data, elt, data);
unsigned int i, save_var_idx;
if (data->handler == HANDLER_ONMAX)
seq_printf(m, "\n\tmax: %10llu", track_val);
else if (data->handler == HANDLER_ONCHANGE)
seq_printf(m, "\n\tchanged: %10llu", track_val);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
if (data->action == ACTION_SNAPSHOT)
return;
for (i = 0; i < hist_data->n_save_vars; i++) {
struct hist_field *save_val = hist_data->save_vars[i]->val;
struct hist_field *save_var = hist_data->save_vars[i]->var;
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
u64 val;
save_var_idx = save_var->var.idx;
val = tracing_map_read_var(elt, save_var_idx);
if (save_val->flags & HIST_FIELD_FL_STRING) {
seq_printf(m, " %s: %-32s", save_var->var.name,
(char *)(uintptr_t)(val));
} else
seq_printf(m, " %s: %10llu", save_var->var.name, val);
}
}
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
static void ontrack_action(struct hist_trigger_data *hist_data,
struct tracing_map_elt *elt,
struct trace_buffer *buffer, void *rec,
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
struct ring_buffer_event *rbe, void *key,
struct action_data *data, u64 *var_ref_vals)
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
{
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
u64 var_val = var_ref_vals[data->track_data.var_ref->var_ref_idx];
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
if (check_track_val(elt, data, var_val)) {
save_track_val(hist_data, elt, data, var_val);
save_track_data(hist_data, elt, buffer, rec, rbe,
key, data, var_ref_vals);
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
}
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
}
static void action_data_destroy(struct action_data *data)
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
{
unsigned int i;
lockdep_assert_held(&event_mutex);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
kfree(data->action_name);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
for (i = 0; i < data->n_params; i++)
kfree(data->params[i]);
if (data->synth_event)
data->synth_event->ref--;
kfree(data->synth_event_name);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
kfree(data);
}
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
static void track_data_destroy(struct hist_trigger_data *hist_data,
struct action_data *data)
{
struct trace_event_file *file = hist_data->event_file;
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
destroy_hist_field(data->track_data.track_var, 0);
if (data->action == ACTION_SNAPSHOT) {
struct track_data *track_data;
track_data = tracing_cond_snapshot_data(file->tr);
if (track_data && track_data->hist_data == hist_data) {
tracing_snapshot_cond_disable(file->tr);
track_data_free(track_data);
}
}
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
kfree(data->track_data.var_str);
action_data_destroy(data);
}
static int action_create(struct hist_trigger_data *hist_data,
struct action_data *data);
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
static int track_data_create(struct hist_trigger_data *hist_data,
struct action_data *data)
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
{
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
struct hist_field *var_field, *ref_field, *track_var = NULL;
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
struct trace_event_file *file = hist_data->event_file;
struct trace_array *tr = file->tr;
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
char *track_data_var_str;
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
int ret = 0;
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
track_data_var_str = data->track_data.var_str;
if (track_data_var_str[0] != '$') {
hist_err(tr, HIST_ERR_ONX_NOT_VAR, errpos(track_data_var_str));
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
return -EINVAL;
}
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
track_data_var_str++;
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
var_field = find_target_event_var(hist_data, NULL, NULL, track_data_var_str);
if (!var_field) {
hist_err(tr, HIST_ERR_ONX_VAR_NOT_FOUND, errpos(track_data_var_str));
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
return -EINVAL;
}
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
ref_field = create_var_ref(hist_data, var_field, NULL, NULL);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
if (!ref_field)
return -ENOMEM;
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
data->track_data.var_ref = ref_field;
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
if (data->handler == HANDLER_ONMAX)
track_var = create_var(hist_data, file, "__max", sizeof(u64), "u64");
if (IS_ERR(track_var)) {
hist_err(tr, HIST_ERR_ONX_VAR_CREATE_FAIL, 0);
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
ret = PTR_ERR(track_var);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
goto out;
}
if (data->handler == HANDLER_ONCHANGE)
track_var = create_var(hist_data, file, "__change", sizeof(u64), "u64");
if (IS_ERR(track_var)) {
hist_err(tr, HIST_ERR_ONX_VAR_CREATE_FAIL, 0);
ret = PTR_ERR(track_var);
goto out;
}
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
data->track_data.track_var = track_var;
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
ret = action_create(hist_data, data);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
out:
return ret;
}
static int parse_action_params(struct trace_array *tr, char *params,
struct action_data *data)
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
{
char *param, *saved_param;
bool first_param = true;
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
int ret = 0;
while (params) {
if (data->n_params >= SYNTH_FIELDS_MAX) {
hist_err(tr, HIST_ERR_TOO_MANY_PARAMS, 0);
ret = -EINVAL;
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
goto out;
}
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
param = strsep(&params, ",");
if (!param) {
hist_err(tr, HIST_ERR_PARAM_NOT_FOUND, 0);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
ret = -EINVAL;
goto out;
}
param = strstrip(param);
if (strlen(param) < 2) {
hist_err(tr, HIST_ERR_INVALID_PARAM, errpos(param));
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
ret = -EINVAL;
goto out;
}
saved_param = kstrdup(param, GFP_KERNEL);
if (!saved_param) {
ret = -ENOMEM;
goto out;
}
if (first_param && data->use_trace_keyword) {
data->synth_event_name = saved_param;
first_param = false;
continue;
}
first_param = false;
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
data->params[data->n_params++] = saved_param;
}
out:
return ret;
}
static int action_parse(struct trace_array *tr, char *str, struct action_data *data,
enum handler_id handler)
{
char *action_name;
int ret = 0;
strsep(&str, ".");
if (!str) {
hist_err(tr, HIST_ERR_ACTION_NOT_FOUND, 0);
ret = -EINVAL;
goto out;
}
action_name = strsep(&str, "(");
if (!action_name || !str) {
hist_err(tr, HIST_ERR_ACTION_NOT_FOUND, 0);
ret = -EINVAL;
goto out;
}
if (str_has_prefix(action_name, "save")) {
char *params = strsep(&str, ")");
if (!params) {
hist_err(tr, HIST_ERR_NO_SAVE_PARAMS, 0);
ret = -EINVAL;
goto out;
}
ret = parse_action_params(tr, params, data);
if (ret)
goto out;
if (handler == HANDLER_ONMAX)
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
data->track_data.check_val = check_track_val_max;
else if (handler == HANDLER_ONCHANGE)
data->track_data.check_val = check_track_val_changed;
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
else {
hist_err(tr, HIST_ERR_ACTION_MISMATCH, errpos(action_name));
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
ret = -EINVAL;
goto out;
}
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
data->track_data.save_data = save_track_data_vars;
data->fn = ontrack_action;
data->action = ACTION_SAVE;
} else if (str_has_prefix(action_name, "snapshot")) {
char *params = strsep(&str, ")");
if (!str) {
hist_err(tr, HIST_ERR_NO_CLOSING_PAREN, errpos(params));
ret = -EINVAL;
goto out;
}
if (handler == HANDLER_ONMAX)
data->track_data.check_val = check_track_val_max;
else if (handler == HANDLER_ONCHANGE)
data->track_data.check_val = check_track_val_changed;
else {
hist_err(tr, HIST_ERR_ACTION_MISMATCH, errpos(action_name));
ret = -EINVAL;
goto out;
}
data->track_data.save_data = save_track_data_snapshot;
data->fn = ontrack_action;
data->action = ACTION_SNAPSHOT;
} else {
char *params = strsep(&str, ")");
if (str_has_prefix(action_name, "trace"))
data->use_trace_keyword = true;
if (params) {
ret = parse_action_params(tr, params, data);
if (ret)
goto out;
}
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
if (handler == HANDLER_ONMAX)
data->track_data.check_val = check_track_val_max;
else if (handler == HANDLER_ONCHANGE)
data->track_data.check_val = check_track_val_changed;
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
if (handler != HANDLER_ONMATCH) {
data->track_data.save_data = action_trace;
data->fn = ontrack_action;
} else
data->fn = action_trace;
data->action = ACTION_TRACE;
}
data->action_name = kstrdup(action_name, GFP_KERNEL);
if (!data->action_name) {
ret = -ENOMEM;
goto out;
}
data->handler = handler;
out:
return ret;
}
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
static struct action_data *track_data_parse(struct hist_trigger_data *hist_data,
char *str, enum handler_id handler)
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
{
struct action_data *data;
int ret = -EINVAL;
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
char *var_str;
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
data = kzalloc(sizeof(*data), GFP_KERNEL);
if (!data)
return ERR_PTR(-ENOMEM);
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
var_str = strsep(&str, ")");
if (!var_str || !str) {
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
ret = -EINVAL;
goto free;
}
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
data->track_data.var_str = kstrdup(var_str, GFP_KERNEL);
if (!data->track_data.var_str) {
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
ret = -ENOMEM;
goto free;
}
ret = action_parse(hist_data->event_file->tr, str, data, handler);
if (ret)
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
goto free;
out:
return data;
free:
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
track_data_destroy(hist_data, data);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
data = ERR_PTR(ret);
goto out;
}
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
static void onmatch_destroy(struct action_data *data)
{
kfree(data->match_data.event);
kfree(data->match_data.event_system);
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
action_data_destroy(data);
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
}
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
static void destroy_field_var(struct field_var *field_var)
{
if (!field_var)
return;
destroy_hist_field(field_var->var, 0);
destroy_hist_field(field_var->val, 0);
kfree(field_var);
}
static void destroy_field_vars(struct hist_trigger_data *hist_data)
{
unsigned int i;
for (i = 0; i < hist_data->n_field_vars; i++)
destroy_field_var(hist_data->field_vars[i]);
tracing: Fix memory leaks in trace_events_hist.c kmemleak report 1: [<9092c50b>] kmem_cache_alloc_trace+0x138/0x270 [<05a2c9ed>] create_field_var+0xcf/0x180 [<528a2d68>] action_create+0xe2/0xc80 [<63f50b61>] event_hist_trigger_func+0x15b5/0x1920 [<28ea5d3d>] trigger_process_regex+0x7b/0xc0 [<3138e86f>] event_trigger_write+0x4d/0xb0 [<ffd66c19>] __vfs_write+0x30/0x200 [<4f424a0d>] vfs_write+0x96/0x1b0 [<da59a290>] ksys_write+0x53/0xc0 [<3717101a>] __ia32_sys_write+0x15/0x20 [<c5f23497>] do_fast_syscall_32+0x70/0x250 [<46e2629c>] entry_SYSENTER_32+0xaf/0x102 This is because save_vars[] of struct hist_trigger_data are not destroyed kmemleak report 2: [<9092c50b>] kmem_cache_alloc_trace+0x138/0x270 [<6e5e97c5>] create_var+0x3c/0x110 [<de82f1b9>] create_field_var+0xaf/0x180 [<528a2d68>] action_create+0xe2/0xc80 [<63f50b61>] event_hist_trigger_func+0x15b5/0x1920 [<28ea5d3d>] trigger_process_regex+0x7b/0xc0 [<3138e86f>] event_trigger_write+0x4d/0xb0 [<ffd66c19>] __vfs_write+0x30/0x200 [<4f424a0d>] vfs_write+0x96/0x1b0 [<da59a290>] ksys_write+0x53/0xc0 [<3717101a>] __ia32_sys_write+0x15/0x20 [<c5f23497>] do_fast_syscall_32+0x70/0x250 [<46e2629c>] entry_SYSENTER_32+0xaf/0x102 struct hist_field allocated through create_var() do not initialize "ref" field to 1. The code in __destroy_hist_field() does not destroy object if "ref" is initialized to zero, the condition if (--hist_field->ref > 1) always passes since unsigned int wraps. kmemleak report 3: [<f8666fcc>] __kmalloc_track_caller+0x139/0x2b0 [<bb7f80a5>] kstrdup+0x27/0x50 [<39d70006>] init_var_ref+0x58/0xd0 [<8ca76370>] create_var_ref+0x89/0xe0 [<f045fc39>] action_create+0x38f/0xc80 [<7c146821>] event_hist_trigger_func+0x15b5/0x1920 [<07de3f61>] trigger_process_regex+0x7b/0xc0 [<e87daf8f>] event_trigger_write+0x4d/0xb0 [<19bf1512>] __vfs_write+0x30/0x200 [<64ce4d27>] vfs_write+0x96/0x1b0 [<a6f34170>] ksys_write+0x53/0xc0 [<7d4230cd>] __ia32_sys_write+0x15/0x20 [<8eadca00>] do_fast_syscall_32+0x70/0x250 [<235cf985>] entry_SYSENTER_32+0xaf/0x102 hist_fields (system & event_name) are not freed Link: http://lkml.kernel.org/r/20200422061503.GA5151@cosmos Signed-off-by: Vamshi K Sthambamkadi <vamshi.k.sthambamkadi@gmail.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2020-04-22 06:15:06 +00:00
for (i = 0; i < hist_data->n_save_vars; i++)
destroy_field_var(hist_data->save_vars[i]);
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
}
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
static void save_field_var(struct hist_trigger_data *hist_data,
struct field_var *field_var)
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
{
hist_data->field_vars[hist_data->n_field_vars++] = field_var;
/* Stack traces are saved in the string storage too */
if (field_var->val->flags & (HIST_FIELD_FL_STRING | HIST_FIELD_FL_STACKTRACE))
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
hist_data->n_field_var_str++;
}
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
static int check_synth_field(struct synth_event *event,
struct hist_field *hist_field,
unsigned int field_pos)
{
struct synth_field *field;
if (field_pos >= event->n_fields)
return -EINVAL;
field = event->fields[field_pos];
tracing: Add support for dynamic strings to synthetic events Currently, sythetic events only support static string fields such as: # echo 'test_latency u64 lat; char somename[32]' > /sys/kernel/debug/tracing/synthetic_events Which is fine, but wastes a lot of space in the event. It also prevents the most commonly-defined strings in the existing trace events e.g. those defined using __string(), from being passed to synthetic events via the trace() action. With this change, synthetic events with dynamic fields can be defined: # echo 'test_latency u64 lat; char somename[]' > /sys/kernel/debug/tracing/synthetic_events And the trace() action can be used to generate events using either dynamic or static strings: # echo 'hist:keys=name:lat=common_timestamp.usecs-$ts0:onmatch(sys.event).test_latency($lat,name)' > /sys/kernel/debug/tracing/events The synthetic event dynamic strings are implemented in the same way as the existing __data_loc strings and appear as such in the format file. [ <rostedt@goodmis.org>: added __set_synth_event_print_fmt() changes: I added the following to make it work with trace-cmd. Dynamic strings must have __get_str() for events in the print_fmt otherwise it can't be parsed correctly. ] Link: https://lore.kernel.org/r/cover.1601588066.git.zanussi@kernel.org Link: https://lkml.kernel.org/r/3ed35b6d0e390f5b94cb4a9ba1cc18f5982ab277.1601848695.git.zanussi@kernel.org Tested-by: Axel Rasmussen <axelrasmussen@google.com> Signed-off-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2020-10-04 22:14:06 +00:00
/*
* A dynamic string synth field can accept static or
* dynamic. A static string synth field can only accept a
* same-sized static string, which is checked for later.
*/
if (strstr(hist_field->type, "char[") && field->is_string
&& field->is_dynamic)
return 0;
tracing: Allow synthetic events to pass around stacktraces Allow a stacktrace from one event to be displayed by the end event of a synthetic event. This is very useful when looking for the longest latency of a sleep or something blocked on I/O. # cd /sys/kernel/tracing/ # echo 's:block_lat pid_t pid; u64 delta; unsigned long[] stack;' > dynamic_events # echo 'hist:keys=next_pid:ts=common_timestamp.usecs,st=stacktrace if prev_state == 1||prev_state == 2' > events/sched/sched_switch/trigger # echo 'hist:keys=prev_pid:delta=common_timestamp.usecs-$ts,s=$st:onmax($delta).trace(block_lat,prev_pid,$delta,$s)' >> events/sched/sched_switch/trigger The above creates a "block_lat" synthetic event that take the stacktrace of when a task schedules out in either the interruptible or uninterruptible states, and on a new per process max $delta (the time it was scheduled out), will print the process id and the stacktrace. # echo 1 > events/synthetic/block_lat/enable # cat trace # TASK-PID CPU# ||||| TIMESTAMP FUNCTION # | | | ||||| | | kworker/u16:0-767 [006] d..4. 560.645045: block_lat: pid=767 delta=66 stack=STACK: => __schedule => schedule => pipe_read => vfs_read => ksys_read => do_syscall_64 => 0x966000aa <idle>-0 [003] d..4. 561.132117: block_lat: pid=0 delta=413787 stack=STACK: => __schedule => schedule => schedule_hrtimeout_range_clock => do_sys_poll => __x64_sys_poll => do_syscall_64 => 0x966000aa <...>-153 [006] d..4. 562.068407: block_lat: pid=153 delta=54 stack=STACK: => __schedule => schedule => io_schedule => rq_qos_wait => wbt_wait => __rq_qos_throttle => blk_mq_submit_bio => submit_bio_noacct_nocheck => ext4_bio_write_page => mpage_submit_page => mpage_process_page_bufs => mpage_prepare_extent_to_map => ext4_do_writepages => ext4_writepages => do_writepages => __writeback_single_inode Link: https://lkml.kernel.org/r/20230117152236.010941267@goodmis.org Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Tom Zanussi <zanussi@kernel.org> Cc: Ross Zwisler <zwisler@google.com> Cc: Ching-lin Yu <chinglinyu@google.com> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2023-01-17 15:21:28 +00:00
if (strstr(hist_field->type, "long[") && field->is_stack)
return 0;
if (strcmp(field->type, hist_field->type) != 0) {
if (field->size != hist_field->size ||
(!field->is_string && field->is_signed != hist_field->is_signed))
return -EINVAL;
}
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
return 0;
}
static struct hist_field *
trace_action_find_var(struct hist_trigger_data *hist_data,
struct action_data *data,
char *system, char *event, char *var)
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
{
struct trace_array *tr = hist_data->event_file->tr;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
struct hist_field *hist_field;
var++; /* skip '$' */
hist_field = find_target_event_var(hist_data, system, event, var);
if (!hist_field) {
if (!system && data->handler == HANDLER_ONMATCH) {
system = data->match_data.event_system;
event = data->match_data.event;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
}
hist_field = find_event_var(hist_data, system, event, var);
}
if (!hist_field)
hist_err(tr, HIST_ERR_PARAM_NOT_FOUND, errpos(var));
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
return hist_field;
}
static struct hist_field *
trace_action_create_field_var(struct hist_trigger_data *hist_data,
struct action_data *data, char *system,
char *event, char *var)
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
{
struct hist_field *hist_field = NULL;
struct field_var *field_var;
/*
* First try to create a field var on the target event (the
* currently being defined). This will create a variable for
* unqualified fields on the target event, or if qualified,
* target fields that have qualified names matching the target.
*/
field_var = create_target_field_var(hist_data, system, event, var);
if (field_var && !IS_ERR(field_var)) {
save_field_var(hist_data, field_var);
hist_field = field_var->var;
} else {
field_var = NULL;
/*
* If no explicit system.event is specified, default to
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
* looking for fields on the onmatch(system.event.xxx)
* event.
*/
if (!system && data->handler == HANDLER_ONMATCH) {
system = data->match_data.event_system;
event = data->match_data.event;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
}
tracing / histogram: Fix NULL pointer dereference on strcmp() on NULL event name The following commands: # echo 'read_max u64 size;' > synthetic_events # echo 'hist:keys=common_pid:count=count:onmax($count).trace(read_max,count)' > events/syscalls/sys_enter_read/trigger Causes: BUG: kernel NULL pointer dereference, address: 0000000000000000 #PF: supervisor read access in kernel mode #PF: error_code(0x0000) - not-present page PGD 0 P4D 0 Oops: 0000 [#1] PREEMPT SMP CPU: 4 PID: 1763 Comm: bash Not tainted 5.14.0-rc2-test+ #155 Hardware name: Hewlett-Packard HP Compaq Pro 6300 SFF/339A, BIOS K01 v03.03 07/14/2016 RIP: 0010:strcmp+0xc/0x20 Code: 75 f7 31 c0 0f b6 0c 06 88 0c 02 48 83 c0 01 84 c9 75 f1 4c 89 c0 c3 0f 1f 80 00 00 00 00 31 c0 eb 08 48 83 c0 01 84 d2 74 0f <0f> b6 14 07 3a 14 06 74 ef 19 c0 83 c8 01 c3 31 c0 c3 66 90 48 89 RSP: 0018:ffffb5fdc0963ca8 EFLAGS: 00010246 RAX: 0000000000000000 RBX: ffffffffb3a4e040 RCX: 0000000000000000 RDX: 0000000000000000 RSI: ffff9714c0d0b640 RDI: 0000000000000000 RBP: 0000000000000000 R08: 00000022986b7cde R09: ffffffffb3a4dff8 R10: 0000000000000000 R11: 0000000000000000 R12: ffff9714c50603c8 R13: 0000000000000000 R14: ffff97143fdf9e48 R15: ffff9714c01a2210 FS: 00007f1fa6785740(0000) GS:ffff9714da400000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000000 CR3: 000000002d863004 CR4: 00000000001706e0 Call Trace: __find_event_file+0x4e/0x80 action_create+0x6b7/0xeb0 ? kstrdup+0x44/0x60 event_hist_trigger_func+0x1a07/0x2130 trigger_process_regex+0xbd/0x110 event_trigger_write+0x71/0xd0 vfs_write+0xe9/0x310 ksys_write+0x68/0xe0 do_syscall_64+0x3b/0x90 entry_SYSCALL_64_after_hwframe+0x44/0xae RIP: 0033:0x7f1fa6879e87 The problem was the "trace(read_max,count)" where the "count" should be "$count" as "onmax()" only handles variables (although it really should be able to figure out that "count" is a field of sys_enter_read). But there's a path that does not find the variable and ends up passing a NULL for the event, which ends up getting passed to "strcmp()". Add a check for NULL to return and error on the command with: # cat error_log hist:syscalls:sys_enter_read: error: Couldn't create or find variable Command: hist:keys=common_pid:count=count:onmax($count).trace(read_max,count) ^ Link: https://lkml.kernel.org/r/20210808003011.4037f8d0@oasis.local.home Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: stable@vger.kernel.org Fixes: 50450603ec9cb tracing: Add 'onmax' hist trigger action support Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-08-08 04:30:11 +00:00
if (!event)
goto free;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
/*
* At this point, we're looking at a field on another
* event. Because we can't modify a hist trigger on
* another event to add a variable for a field, we need
* to create a new trigger on that event and create the
* variable at the same time.
*/
hist_field = create_field_var_hist(hist_data, system, event, var);
if (IS_ERR(hist_field))
goto free;
}
out:
return hist_field;
free:
destroy_field_var(field_var);
hist_field = NULL;
goto out;
}
static int trace_action_create(struct hist_trigger_data *hist_data,
struct action_data *data)
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
{
struct trace_array *tr = hist_data->event_file->tr;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
char *event_name, *param, *system = NULL;
struct hist_field *hist_field, *var_ref;
unsigned int i;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
unsigned int field_pos = 0;
struct synth_event *event;
char *synth_event_name;
int var_ref_idx, ret = 0;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
lockdep_assert_held(&event_mutex);
tracing/hist: Fix out-of-bound write on 'action_data.var_ref_idx' When generate a synthetic event with many params and then create a trace action for it [1], kernel panic happened [2]. It is because that in trace_action_create() 'data->n_params' is up to SYNTH_FIELDS_MAX (current value is 64), and array 'data->var_ref_idx' keeps indices into array 'hist_data->var_refs' for each synthetic event param, but the length of 'data->var_ref_idx' is TRACING_MAP_VARS_MAX (current value is 16), so out-of-bound write happened when 'data->n_params' more than 16. In this case, 'data->match_data.event' is overwritten and eventually cause the panic. To solve the issue, adjust the length of 'data->var_ref_idx' to be SYNTH_FIELDS_MAX and add sanity checks to avoid out-of-bound write. [1] # cd /sys/kernel/tracing/ # echo "my_synth_event int v1; int v2; int v3; int v4; int v5; int v6;\ int v7; int v8; int v9; int v10; int v11; int v12; int v13; int v14;\ int v15; int v16; int v17; int v18; int v19; int v20; int v21; int v22;\ int v23; int v24; int v25; int v26; int v27; int v28; int v29; int v30;\ int v31; int v32; int v33; int v34; int v35; int v36; int v37; int v38;\ int v39; int v40; int v41; int v42; int v43; int v44; int v45; int v46;\ int v47; int v48; int v49; int v50; int v51; int v52; int v53; int v54;\ int v55; int v56; int v57; int v58; int v59; int v60; int v61; int v62;\ int v63" >> synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="bash"' >> \ events/sched/sched_waking/trigger # echo "hist:keys=next_pid:onmatch(sched.sched_waking).my_synth_event(\ pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,\ pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,\ pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,pid,\ pid,pid,pid,pid,pid,pid,pid,pid,pid)" >> events/sched/sched_switch/trigger [2] BUG: unable to handle page fault for address: ffff91c900000000 PGD 61001067 P4D 61001067 PUD 0 Oops: 0000 [#1] PREEMPT SMP NOPTI CPU: 2 PID: 322 Comm: bash Tainted: G W 6.1.0-rc8+ #229 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.15.0-0-g2dd4b9b3f840-prebuilt.qemu.org 04/01/2014 RIP: 0010:strcmp+0xc/0x30 Code: 75 f7 31 d2 44 0f b6 04 16 44 88 04 11 48 83 c2 01 45 84 c0 75 ee c3 cc cc cc cc 0f 1f 00 31 c0 eb 08 48 83 c0 01 84 d2 74 13 <0f> b6 14 07 3a 14 06 74 ef 19 c0 83 c8 01 c3 cc cc cc cc 31 c3 RSP: 0018:ffff9b3b00f53c48 EFLAGS: 00000246 RAX: 0000000000000000 RBX: ffffffffba958a68 RCX: 0000000000000000 RDX: 0000000000000010 RSI: ffff91c943d33a90 RDI: ffff91c900000000 RBP: ffff91c900000000 R08: 00000018d604b529 R09: 0000000000000000 R10: ffff91c9483eddb1 R11: ffff91ca483eddab R12: ffff91c946171580 R13: ffff91c9479f0538 R14: ffff91c9457c2848 R15: ffff91c9479f0538 FS: 00007f1d1cfbe740(0000) GS:ffff91c9bdc80000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffff91c900000000 CR3: 0000000006316000 CR4: 00000000000006e0 Call Trace: <TASK> __find_event_file+0x55/0x90 action_create+0x76c/0x1060 event_hist_trigger_parse+0x146d/0x2060 ? event_trigger_write+0x31/0xd0 trigger_process_regex+0xbb/0x110 event_trigger_write+0x6b/0xd0 vfs_write+0xc8/0x3e0 ? alloc_fd+0xc0/0x160 ? preempt_count_add+0x4d/0xa0 ? preempt_count_add+0x70/0xa0 ksys_write+0x5f/0xe0 do_syscall_64+0x3b/0x90 entry_SYSCALL_64_after_hwframe+0x63/0xcd RIP: 0033:0x7f1d1d0cf077 Code: 64 89 02 48 c7 c0 ff ff ff ff eb bb 0f 1f 80 00 00 00 00 f3 0f 1e fa 64 8b 04 25 18 00 00 00 85 c0 75 10 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 51 c3 48 83 ec 28 48 89 54 24 18 48 89 74 RSP: 002b:00007ffcebb0e568 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 RAX: ffffffffffffffda RBX: 0000000000000143 RCX: 00007f1d1d0cf077 RDX: 0000000000000143 RSI: 00005639265aa7e0 RDI: 0000000000000001 RBP: 00005639265aa7e0 R08: 000000000000000a R09: 0000000000000142 R10: 000056392639c017 R11: 0000000000000246 R12: 0000000000000143 R13: 00007f1d1d1ae6a0 R14: 00007f1d1d1aa4a0 R15: 00007f1d1d1a98a0 </TASK> Modules linked in: CR2: ffff91c900000000 ---[ end trace 0000000000000000 ]--- RIP: 0010:strcmp+0xc/0x30 Code: 75 f7 31 d2 44 0f b6 04 16 44 88 04 11 48 83 c2 01 45 84 c0 75 ee c3 cc cc cc cc 0f 1f 00 31 c0 eb 08 48 83 c0 01 84 d2 74 13 <0f> b6 14 07 3a 14 06 74 ef 19 c0 83 c8 01 c3 cc cc cc cc 31 c3 RSP: 0018:ffff9b3b00f53c48 EFLAGS: 00000246 RAX: 0000000000000000 RBX: ffffffffba958a68 RCX: 0000000000000000 RDX: 0000000000000010 RSI: ffff91c943d33a90 RDI: ffff91c900000000 RBP: ffff91c900000000 R08: 00000018d604b529 R09: 0000000000000000 R10: ffff91c9483eddb1 R11: ffff91ca483eddab R12: ffff91c946171580 R13: ffff91c9479f0538 R14: ffff91c9457c2848 R15: ffff91c9479f0538 FS: 00007f1d1cfbe740(0000) GS:ffff91c9bdc80000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffff91c900000000 CR3: 0000000006316000 CR4: 00000000000006e0 Link: https://lore.kernel.org/linux-trace-kernel/20221207035143.2278781-1-zhengyejian1@huawei.com Cc: <mhiramat@kernel.org> Cc: <zanussi@kernel.org> Cc: stable@vger.kernel.org Fixes: d380dcde9a07 ("tracing: Fix now invalid var_ref_vals assumption in trace action") Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-12-07 03:51:43 +00:00
/* Sanity check to avoid out-of-bound write on 'data->var_ref_idx' */
if (data->n_params > SYNTH_FIELDS_MAX)
return -EINVAL;
if (data->use_trace_keyword)
synth_event_name = data->synth_event_name;
else
synth_event_name = data->action_name;
event = find_synth_event(synth_event_name);
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
if (!event) {
hist_err(tr, HIST_ERR_SYNTH_EVENT_NOT_FOUND, errpos(synth_event_name));
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
return -EINVAL;
}
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
event->ref++;
for (i = 0; i < data->n_params; i++) {
char *p;
p = param = kstrdup(data->params[i], GFP_KERNEL);
if (!param) {
ret = -ENOMEM;
goto err;
}
system = strsep(&param, ".");
if (!param) {
param = (char *)system;
system = event_name = NULL;
} else {
event_name = strsep(&param, ".");
if (!param) {
kfree(p);
ret = -EINVAL;
goto err;
}
}
if (param[0] == '$')
hist_field = trace_action_find_var(hist_data, data,
system, event_name,
param);
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
else
hist_field = trace_action_create_field_var(hist_data,
data,
system,
event_name,
param);
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
if (!hist_field) {
kfree(p);
ret = -EINVAL;
goto err;
}
if (check_synth_field(event, hist_field, field_pos) == 0) {
var_ref = create_var_ref(hist_data, hist_field,
system, event_name);
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
if (!var_ref) {
kfree(p);
ret = -ENOMEM;
goto err;
}
var_ref_idx = find_var_ref_idx(hist_data, var_ref);
if (WARN_ON(var_ref_idx < 0)) {
kfree(p);
ret = var_ref_idx;
goto err;
}
data->var_ref_idx[i] = var_ref_idx;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
field_pos++;
kfree(p);
continue;
}
hist_err(tr, HIST_ERR_SYNTH_TYPE_MISMATCH, errpos(param));
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
kfree(p);
ret = -EINVAL;
goto err;
}
if (field_pos != event->n_fields) {
hist_err(tr, HIST_ERR_SYNTH_COUNT_MISMATCH, errpos(event->name));
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
ret = -EINVAL;
goto err;
}
data->synth_event = event;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
out:
return ret;
err:
event->ref--;
goto out;
}
static int action_create(struct hist_trigger_data *hist_data,
struct action_data *data)
{
struct trace_event_file *file = hist_data->event_file;
struct trace_array *tr = file->tr;
struct track_data *track_data;
struct field_var *field_var;
unsigned int i;
char *param;
int ret = 0;
if (data->action == ACTION_TRACE)
return trace_action_create(hist_data, data);
if (data->action == ACTION_SNAPSHOT) {
track_data = track_data_alloc(hist_data->key_size, data, hist_data);
if (IS_ERR(track_data)) {
ret = PTR_ERR(track_data);
goto out;
}
ret = tracing_snapshot_cond_enable(file->tr, track_data,
cond_snapshot_update);
if (ret)
track_data_free(track_data);
goto out;
}
if (data->action == ACTION_SAVE) {
if (hist_data->n_save_vars) {
ret = -EEXIST;
hist_err(tr, HIST_ERR_TOO_MANY_SAVE_ACTIONS, 0);
goto out;
}
for (i = 0; i < data->n_params; i++) {
param = kstrdup(data->params[i], GFP_KERNEL);
if (!param) {
ret = -ENOMEM;
goto out;
}
field_var = create_target_field_var(hist_data, NULL, NULL, param);
if (IS_ERR(field_var)) {
hist_err(tr, HIST_ERR_FIELD_VAR_CREATE_FAIL,
errpos(param));
ret = PTR_ERR(field_var);
kfree(param);
goto out;
}
hist_data->save_vars[hist_data->n_save_vars++] = field_var;
if (field_var->val->flags &
(HIST_FIELD_FL_STRING | HIST_FIELD_FL_STACKTRACE))
hist_data->n_save_var_str++;
kfree(param);
}
}
out:
return ret;
}
static int onmatch_create(struct hist_trigger_data *hist_data,
struct action_data *data)
{
return action_create(hist_data, data);
}
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
static struct action_data *onmatch_parse(struct trace_array *tr, char *str)
{
char *match_event, *match_event_system;
struct action_data *data;
int ret = -EINVAL;
data = kzalloc(sizeof(*data), GFP_KERNEL);
if (!data)
return ERR_PTR(-ENOMEM);
match_event = strsep(&str, ")");
if (!match_event || !str) {
hist_err(tr, HIST_ERR_NO_CLOSING_PAREN, errpos(match_event));
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
goto free;
}
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
match_event_system = strsep(&match_event, ".");
if (!match_event) {
hist_err(tr, HIST_ERR_SUBSYS_NOT_FOUND, errpos(match_event_system));
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
goto free;
}
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
if (IS_ERR(event_file(tr, match_event_system, match_event))) {
hist_err(tr, HIST_ERR_INVALID_SUBSYS_EVENT, errpos(match_event));
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
goto free;
}
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
data->match_data.event = kstrdup(match_event, GFP_KERNEL);
if (!data->match_data.event) {
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
ret = -ENOMEM;
goto free;
}
data->match_data.event_system = kstrdup(match_event_system, GFP_KERNEL);
if (!data->match_data.event_system) {
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
ret = -ENOMEM;
goto free;
}
ret = action_parse(tr, str, data, HANDLER_ONMATCH);
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
if (ret)
goto free;
out:
return data;
free:
onmatch_destroy(data);
data = ERR_PTR(ret);
goto out;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
static int create_hitcount_val(struct hist_trigger_data *hist_data)
{
hist_data->fields[HITCOUNT_IDX] =
create_hist_field(hist_data, NULL, HIST_FIELD_FL_HITCOUNT, NULL);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (!hist_data->fields[HITCOUNT_IDX])
return -ENOMEM;
hist_data->n_vals++;
hist_data->n_fields++;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (WARN_ON(hist_data->n_vals > TRACING_MAP_VALS_MAX))
return -EINVAL;
return 0;
}
static int __create_val_field(struct hist_trigger_data *hist_data,
unsigned int val_idx,
struct trace_event_file *file,
char *var_name, char *field_str,
unsigned long flags)
{
struct hist_field *hist_field;
int ret = 0, n_subexprs = 0;
hist_field = parse_expr(hist_data, file, field_str, flags, var_name, &n_subexprs);
if (IS_ERR(hist_field)) {
ret = PTR_ERR(hist_field);
goto out;
}
2023-05-24 02:11:08 +00:00
/* values and variables should not have some modifiers */
if (hist_field->flags & HIST_FIELD_FL_VAR) {
/* Variable */
if (hist_field->flags & (HIST_FIELD_FL_GRAPH | HIST_FIELD_FL_PERCENT |
HIST_FIELD_FL_BUCKET | HIST_FIELD_FL_LOG2))
goto err;
} else {
/* Value */
if (hist_field->flags & (HIST_FIELD_FL_GRAPH | HIST_FIELD_FL_PERCENT |
HIST_FIELD_FL_BUCKET | HIST_FIELD_FL_LOG2 |
HIST_FIELD_FL_SYM | HIST_FIELD_FL_SYM_OFFSET |
HIST_FIELD_FL_SYSCALL | HIST_FIELD_FL_STACKTRACE))
goto err;
tracing: Do not let histogram values have some modifiers Histogram values can not be strings, stacktraces, graphs, symbols, syscalls, or grouped in buckets or log. Give an error if a value is set to do so. Note, the histogram code was not prepared to handle these modifiers for histograms and caused a bug. Mark Rutland reported: # echo 'p:copy_to_user __arch_copy_to_user n=$arg2' >> /sys/kernel/tracing/kprobe_events # echo 'hist:keys=n:vals=hitcount.buckets=8:sort=hitcount' > /sys/kernel/tracing/events/kprobes/copy_to_user/trigger # cat /sys/kernel/tracing/events/kprobes/copy_to_user/hist [ 143.694628] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000 [ 143.695190] Mem abort info: [ 143.695362] ESR = 0x0000000096000004 [ 143.695604] EC = 0x25: DABT (current EL), IL = 32 bits [ 143.695889] SET = 0, FnV = 0 [ 143.696077] EA = 0, S1PTW = 0 [ 143.696302] FSC = 0x04: level 0 translation fault [ 143.702381] Data abort info: [ 143.702614] ISV = 0, ISS = 0x00000004 [ 143.702832] CM = 0, WnR = 0 [ 143.703087] user pgtable: 4k pages, 48-bit VAs, pgdp=00000000448f9000 [ 143.703407] [0000000000000000] pgd=0000000000000000, p4d=0000000000000000 [ 143.704137] Internal error: Oops: 0000000096000004 [#1] PREEMPT SMP [ 143.704714] Modules linked in: [ 143.705273] CPU: 0 PID: 133 Comm: cat Not tainted 6.2.0-00003-g6fc512c10a7c #3 [ 143.706138] Hardware name: linux,dummy-virt (DT) [ 143.706723] pstate: 80000005 (Nzcv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--) [ 143.707120] pc : hist_field_name.part.0+0x14/0x140 [ 143.707504] lr : hist_field_name.part.0+0x104/0x140 [ 143.707774] sp : ffff800008333a30 [ 143.707952] x29: ffff800008333a30 x28: 0000000000000001 x27: 0000000000400cc0 [ 143.708429] x26: ffffd7a653b20260 x25: 0000000000000000 x24: ffff10d303ee5800 [ 143.708776] x23: ffffd7a6539b27b0 x22: ffff10d303fb8c00 x21: 0000000000000001 [ 143.709127] x20: ffff10d303ec2000 x19: 0000000000000000 x18: 0000000000000000 [ 143.709478] x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000000 [ 143.709824] x14: 0000000000000000 x13: 203a6f666e692072 x12: 6567676972742023 [ 143.710179] x11: 0a230a6d6172676f x10: 000000000000002c x9 : ffffd7a6521e018c [ 143.710584] x8 : 000000000000002c x7 : 7f7f7f7f7f7f7f7f x6 : 000000000000002c [ 143.710915] x5 : ffff10d303b0103e x4 : ffffd7a653b20261 x3 : 000000000000003d [ 143.711239] x2 : 0000000000020001 x1 : 0000000000000001 x0 : 0000000000000000 [ 143.711746] Call trace: [ 143.712115] hist_field_name.part.0+0x14/0x140 [ 143.712642] hist_field_name.part.0+0x104/0x140 [ 143.712925] hist_field_print+0x28/0x140 [ 143.713125] event_hist_trigger_print+0x174/0x4d0 [ 143.713348] hist_show+0xf8/0x980 [ 143.713521] seq_read_iter+0x1bc/0x4b0 [ 143.713711] seq_read+0x8c/0xc4 [ 143.713876] vfs_read+0xc8/0x2a4 [ 143.714043] ksys_read+0x70/0xfc [ 143.714218] __arm64_sys_read+0x24/0x30 [ 143.714400] invoke_syscall+0x50/0x120 [ 143.714587] el0_svc_common.constprop.0+0x4c/0x100 [ 143.714807] do_el0_svc+0x44/0xd0 [ 143.714970] el0_svc+0x2c/0x84 [ 143.715134] el0t_64_sync_handler+0xbc/0x140 [ 143.715334] el0t_64_sync+0x190/0x194 [ 143.715742] Code: a9bd7bfd 910003fd a90153f3 aa0003f3 (f9400000) [ 143.716510] ---[ end trace 0000000000000000 ]--- Segmentation fault Link: https://lkml.kernel.org/r/20230302020810.559462599@goodmis.org Cc: stable@vger.kernel.org Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Fixes: c6afad49d127f ("tracing: Add hist trigger 'sym' and 'sym-offset' modifiers") Reported-by: Mark Rutland <mark.rutland@arm.com> Tested-by: Mark Rutland <mark.rutland@arm.com> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2023-03-02 01:00:52 +00:00
}
hist_data->fields[val_idx] = hist_field;
++hist_data->n_vals;
++hist_data->n_fields;
if (WARN_ON(hist_data->n_vals > TRACING_MAP_VALS_MAX + TRACING_MAP_VARS_MAX))
ret = -EINVAL;
out:
return ret;
2023-05-24 02:11:08 +00:00
err:
hist_err(file->tr, HIST_ERR_BAD_FIELD_MODIFIER, errpos(field_str));
return -EINVAL;
}
static int create_val_field(struct hist_trigger_data *hist_data,
unsigned int val_idx,
struct trace_event_file *file,
char *field_str)
{
if (WARN_ON(val_idx >= TRACING_MAP_VALS_MAX))
return -EINVAL;
return __create_val_field(hist_data, val_idx, file, NULL, field_str, 0);
}
static const char no_comm[] = "(no comm)";
tracing: Allow execnames to be passed as args for synthetic events Allow common_pid.execname to be saved in a variable in one histogram to be passed to another histogram that can pass it as a parameter to a synthetic event. ># echo 'hist:keys=pid:__arg__1=common_timestamp.usecs:arg2=common_pid.execname' \ > events/sched/sched_waking/trigger ># echo 'wakeup_lat s32 pid; u64 delta; char wake_comm[]' > synthetic_events ># echo 'hist:keys=next_pid:pid=next_pid,delta=common_timestamp.usecs-$__arg__1,exec=$arg2'\ ':onmatch(sched.sched_waking).trace(wakeup_lat,$pid,$delta,$exec)' \ > events/sched/sched_switch/trigger The above is a wake up latency synthetic event setup that passes the execname of the common_pid that woke the task to the scheduling of that task, which triggers a synthetic event that passes the original execname as a parameter to display it. ># echo 1 > events/synthetic/enable ># cat trace <idle>-0 [006] d..4 186.863801: wakeup_lat: pid=1306 delta=65 wake_comm=kworker/u16:3 <idle>-0 [000] d..4 186.863858: wakeup_lat: pid=163 delta=27 wake_comm=<idle> <idle>-0 [001] d..4 186.863903: wakeup_lat: pid=1307 delta=36 wake_comm=kworker/u16:4 <idle>-0 [000] d..4 186.863927: wakeup_lat: pid=163 delta=5 wake_comm=<idle> <idle>-0 [006] d..4 186.863957: wakeup_lat: pid=1306 delta=24 wake_comm=kworker/u16:3 sshd-1306 [006] d..4 186.864051: wakeup_lat: pid=61 delta=62 wake_comm=<idle> <idle>-0 [000] d..4 186.965030: wakeup_lat: pid=609 delta=18 wake_comm=<idle> <idle>-0 [006] d..4 186.987582: wakeup_lat: pid=1306 delta=65 wake_comm=kworker/u16:3 <idle>-0 [000] d..4 186.987639: wakeup_lat: pid=163 delta=27 wake_comm=<idle> Link: https://lkml.kernel.org/r/20210722142837.458596338@goodmis.org Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-22 14:27:07 +00:00
static u64 hist_field_execname(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
struct hist_elt_data *elt_data;
if (WARN_ON_ONCE(!elt))
return (u64)(unsigned long)no_comm;
elt_data = elt->private_data;
if (WARN_ON_ONCE(!elt_data->comm))
return (u64)(unsigned long)no_comm;
return (u64)(unsigned long)(elt_data->comm);
}
tracing/histogram: Fix stacktrace key The current code will always use the current stacktrace as a key even if a stacktrace contained in a specific event field was specified. For example, we expect to use the 'unsigned long[] stack' field in the below event in the histogram: # echo 's:block_lat pid_t pid; u64 delta; unsigned long[] stack;' > /sys/kernel/debug/tracing/dynamic_events # echo 'hist:keys=delta.buckets=100,stack.stacktrace:sort=delta' > /sys/kernel/debug/tracing/events/synthetic/block_lat/trigger But in fact, when we type out the trigger, we see that it's using the plain old global 'stacktrace' as the key, which is just the stacktrace when the event was hit and not the stacktrace contained in the event, which is what we want: # cat /sys/kernel/debug/tracing/events/synthetic/block_lat/trigger hist:keys=delta.buckets=100,stacktrace:vals=hitcount:sort=delta.buckets=100:size=2048 [active] And in fact, there's no code to actually retrieve it from the event, so we need to add HIST_FIELD_FN_STACK and hist_field_stack() to get it and hook it into the trigger code. For now, since the stack is just using dynamic strings, this could just use the dynamic string function, but it seems cleaner to have a dedicated function an be able to tweak independently as necessary. Link: https://lkml.kernel.org/r/11aa614c82976adbfa4ea763dbe885b5fb01d59c.1676063532.git.zanussi@kernel.org Signed-off-by: Tom Zanussi <zanussi@kernel.org> [ Fixed 32bit build warning reported by kernel test robot <lkp@intel.com> ] Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2023-02-10 21:33:05 +00:00
static u64 hist_field_stack(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
u32 str_item = *(u32 *)(event + hist_field->field->offset);
int str_loc = str_item & 0xffff;
char *addr = (char *)(event + str_loc);
return (u64)(unsigned long)addr;
}
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
static u64 hist_fn_call(struct hist_field *hist_field,
struct tracing_map_elt *elt,
struct trace_buffer *buffer,
struct ring_buffer_event *rbe,
void *event)
{
switch (hist_field->fn_num) {
case HIST_FIELD_FN_VAR_REF:
return hist_field_var_ref(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_COUNTER:
return hist_field_counter(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_CONST:
return hist_field_const(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_LOG2:
return hist_field_log2(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_BUCKET:
return hist_field_bucket(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_TIMESTAMP:
return hist_field_timestamp(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_CPU:
return hist_field_cpu(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_STRING:
return hist_field_string(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_DYNSTRING:
return hist_field_dynstring(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_RELDYNSTRING:
return hist_field_reldynstring(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_PSTRING:
return hist_field_pstring(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_S64:
return hist_field_s64(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_U64:
return hist_field_u64(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_S32:
return hist_field_s32(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_U32:
return hist_field_u32(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_S16:
return hist_field_s16(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_U16:
return hist_field_u16(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_S8:
return hist_field_s8(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_U8:
return hist_field_u8(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_UMINUS:
return hist_field_unary_minus(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_MINUS:
return hist_field_minus(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_PLUS:
return hist_field_plus(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_DIV:
return hist_field_div(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_MULT:
return hist_field_mult(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_DIV_POWER2:
return div_by_power_of_two(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_DIV_NOT_POWER2:
return div_by_not_power_of_two(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_DIV_MULT_SHIFT:
return div_by_mult_and_shift(hist_field, elt, buffer, rbe, event);
case HIST_FIELD_FN_EXECNAME:
return hist_field_execname(hist_field, elt, buffer, rbe, event);
tracing/histogram: Fix stacktrace key The current code will always use the current stacktrace as a key even if a stacktrace contained in a specific event field was specified. For example, we expect to use the 'unsigned long[] stack' field in the below event in the histogram: # echo 's:block_lat pid_t pid; u64 delta; unsigned long[] stack;' > /sys/kernel/debug/tracing/dynamic_events # echo 'hist:keys=delta.buckets=100,stack.stacktrace:sort=delta' > /sys/kernel/debug/tracing/events/synthetic/block_lat/trigger But in fact, when we type out the trigger, we see that it's using the plain old global 'stacktrace' as the key, which is just the stacktrace when the event was hit and not the stacktrace contained in the event, which is what we want: # cat /sys/kernel/debug/tracing/events/synthetic/block_lat/trigger hist:keys=delta.buckets=100,stacktrace:vals=hitcount:sort=delta.buckets=100:size=2048 [active] And in fact, there's no code to actually retrieve it from the event, so we need to add HIST_FIELD_FN_STACK and hist_field_stack() to get it and hook it into the trigger code. For now, since the stack is just using dynamic strings, this could just use the dynamic string function, but it seems cleaner to have a dedicated function an be able to tweak independently as necessary. Link: https://lkml.kernel.org/r/11aa614c82976adbfa4ea763dbe885b5fb01d59c.1676063532.git.zanussi@kernel.org Signed-off-by: Tom Zanussi <zanussi@kernel.org> [ Fixed 32bit build warning reported by kernel test robot <lkp@intel.com> ] Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2023-02-10 21:33:05 +00:00
case HIST_FIELD_FN_STACK:
return hist_field_stack(hist_field, elt, buffer, rbe, event);
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
default:
return 0;
}
}
tracing: Allow execnames to be passed as args for synthetic events Allow common_pid.execname to be saved in a variable in one histogram to be passed to another histogram that can pass it as a parameter to a synthetic event. ># echo 'hist:keys=pid:__arg__1=common_timestamp.usecs:arg2=common_pid.execname' \ > events/sched/sched_waking/trigger ># echo 'wakeup_lat s32 pid; u64 delta; char wake_comm[]' > synthetic_events ># echo 'hist:keys=next_pid:pid=next_pid,delta=common_timestamp.usecs-$__arg__1,exec=$arg2'\ ':onmatch(sched.sched_waking).trace(wakeup_lat,$pid,$delta,$exec)' \ > events/sched/sched_switch/trigger The above is a wake up latency synthetic event setup that passes the execname of the common_pid that woke the task to the scheduling of that task, which triggers a synthetic event that passes the original execname as a parameter to display it. ># echo 1 > events/synthetic/enable ># cat trace <idle>-0 [006] d..4 186.863801: wakeup_lat: pid=1306 delta=65 wake_comm=kworker/u16:3 <idle>-0 [000] d..4 186.863858: wakeup_lat: pid=163 delta=27 wake_comm=<idle> <idle>-0 [001] d..4 186.863903: wakeup_lat: pid=1307 delta=36 wake_comm=kworker/u16:4 <idle>-0 [000] d..4 186.863927: wakeup_lat: pid=163 delta=5 wake_comm=<idle> <idle>-0 [006] d..4 186.863957: wakeup_lat: pid=1306 delta=24 wake_comm=kworker/u16:3 sshd-1306 [006] d..4 186.864051: wakeup_lat: pid=61 delta=62 wake_comm=<idle> <idle>-0 [000] d..4 186.965030: wakeup_lat: pid=609 delta=18 wake_comm=<idle> <idle>-0 [006] d..4 186.987582: wakeup_lat: pid=1306 delta=65 wake_comm=kworker/u16:3 <idle>-0 [000] d..4 186.987639: wakeup_lat: pid=163 delta=27 wake_comm=<idle> Link: https://lkml.kernel.org/r/20210722142837.458596338@goodmis.org Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-22 14:27:07 +00:00
/* Convert a var that points to common_pid.execname to a string */
static void update_var_execname(struct hist_field *hist_field)
{
hist_field->flags = HIST_FIELD_FL_STRING | HIST_FIELD_FL_VAR |
HIST_FIELD_FL_EXECNAME;
hist_field->size = MAX_FILTER_STR_VAL;
hist_field->is_signed = 0;
kfree_const(hist_field->type);
hist_field->type = "char[]";
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
hist_field->fn_num = HIST_FIELD_FN_EXECNAME;
tracing: Allow execnames to be passed as args for synthetic events Allow common_pid.execname to be saved in a variable in one histogram to be passed to another histogram that can pass it as a parameter to a synthetic event. ># echo 'hist:keys=pid:__arg__1=common_timestamp.usecs:arg2=common_pid.execname' \ > events/sched/sched_waking/trigger ># echo 'wakeup_lat s32 pid; u64 delta; char wake_comm[]' > synthetic_events ># echo 'hist:keys=next_pid:pid=next_pid,delta=common_timestamp.usecs-$__arg__1,exec=$arg2'\ ':onmatch(sched.sched_waking).trace(wakeup_lat,$pid,$delta,$exec)' \ > events/sched/sched_switch/trigger The above is a wake up latency synthetic event setup that passes the execname of the common_pid that woke the task to the scheduling of that task, which triggers a synthetic event that passes the original execname as a parameter to display it. ># echo 1 > events/synthetic/enable ># cat trace <idle>-0 [006] d..4 186.863801: wakeup_lat: pid=1306 delta=65 wake_comm=kworker/u16:3 <idle>-0 [000] d..4 186.863858: wakeup_lat: pid=163 delta=27 wake_comm=<idle> <idle>-0 [001] d..4 186.863903: wakeup_lat: pid=1307 delta=36 wake_comm=kworker/u16:4 <idle>-0 [000] d..4 186.863927: wakeup_lat: pid=163 delta=5 wake_comm=<idle> <idle>-0 [006] d..4 186.863957: wakeup_lat: pid=1306 delta=24 wake_comm=kworker/u16:3 sshd-1306 [006] d..4 186.864051: wakeup_lat: pid=61 delta=62 wake_comm=<idle> <idle>-0 [000] d..4 186.965030: wakeup_lat: pid=609 delta=18 wake_comm=<idle> <idle>-0 [006] d..4 186.987582: wakeup_lat: pid=1306 delta=65 wake_comm=kworker/u16:3 <idle>-0 [000] d..4 186.987639: wakeup_lat: pid=163 delta=27 wake_comm=<idle> Link: https://lkml.kernel.org/r/20210722142837.458596338@goodmis.org Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-22 14:27:07 +00:00
}
static int create_var_field(struct hist_trigger_data *hist_data,
unsigned int val_idx,
struct trace_event_file *file,
char *var_name, char *expr_str)
{
struct trace_array *tr = hist_data->event_file->tr;
unsigned long flags = 0;
int ret;
if (WARN_ON(val_idx >= TRACING_MAP_VALS_MAX + TRACING_MAP_VARS_MAX))
return -EINVAL;
if (find_var(hist_data, file, var_name) && !hist_data->remove) {
hist_err(tr, HIST_ERR_DUPLICATE_VAR, errpos(var_name));
return -EINVAL;
}
flags |= HIST_FIELD_FL_VAR;
hist_data->n_vars++;
if (WARN_ON(hist_data->n_vars > TRACING_MAP_VARS_MAX))
return -EINVAL;
ret = __create_val_field(hist_data, val_idx, file, var_name, expr_str, flags);
tracing: Allow execnames to be passed as args for synthetic events Allow common_pid.execname to be saved in a variable in one histogram to be passed to another histogram that can pass it as a parameter to a synthetic event. ># echo 'hist:keys=pid:__arg__1=common_timestamp.usecs:arg2=common_pid.execname' \ > events/sched/sched_waking/trigger ># echo 'wakeup_lat s32 pid; u64 delta; char wake_comm[]' > synthetic_events ># echo 'hist:keys=next_pid:pid=next_pid,delta=common_timestamp.usecs-$__arg__1,exec=$arg2'\ ':onmatch(sched.sched_waking).trace(wakeup_lat,$pid,$delta,$exec)' \ > events/sched/sched_switch/trigger The above is a wake up latency synthetic event setup that passes the execname of the common_pid that woke the task to the scheduling of that task, which triggers a synthetic event that passes the original execname as a parameter to display it. ># echo 1 > events/synthetic/enable ># cat trace <idle>-0 [006] d..4 186.863801: wakeup_lat: pid=1306 delta=65 wake_comm=kworker/u16:3 <idle>-0 [000] d..4 186.863858: wakeup_lat: pid=163 delta=27 wake_comm=<idle> <idle>-0 [001] d..4 186.863903: wakeup_lat: pid=1307 delta=36 wake_comm=kworker/u16:4 <idle>-0 [000] d..4 186.863927: wakeup_lat: pid=163 delta=5 wake_comm=<idle> <idle>-0 [006] d..4 186.863957: wakeup_lat: pid=1306 delta=24 wake_comm=kworker/u16:3 sshd-1306 [006] d..4 186.864051: wakeup_lat: pid=61 delta=62 wake_comm=<idle> <idle>-0 [000] d..4 186.965030: wakeup_lat: pid=609 delta=18 wake_comm=<idle> <idle>-0 [006] d..4 186.987582: wakeup_lat: pid=1306 delta=65 wake_comm=kworker/u16:3 <idle>-0 [000] d..4 186.987639: wakeup_lat: pid=163 delta=27 wake_comm=<idle> Link: https://lkml.kernel.org/r/20210722142837.458596338@goodmis.org Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-22 14:27:07 +00:00
if (!ret && hist_data->fields[val_idx]->flags & HIST_FIELD_FL_EXECNAME)
update_var_execname(hist_data->fields[val_idx]);
if (!ret && hist_data->fields[val_idx]->flags &
(HIST_FIELD_FL_STRING | HIST_FIELD_FL_STACKTRACE))
hist_data->fields[val_idx]->var_str_idx = hist_data->n_var_str++;
return ret;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
static int create_val_fields(struct hist_trigger_data *hist_data,
struct trace_event_file *file)
{
unsigned int i, j = 1, n_hitcount = 0;
char *fields_str, *field_str;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
int ret;
ret = create_hitcount_val(hist_data);
if (ret)
goto out;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
fields_str = hist_data->attrs->vals_str;
if (!fields_str)
goto out;
for (i = 0, j = 1; i < TRACING_MAP_VALS_MAX &&
j < TRACING_MAP_VALS_MAX; i++) {
field_str = strsep(&fields_str, ",");
if (!field_str)
break;
if (strcmp(field_str, "hitcount") == 0) {
if (!n_hitcount++)
continue;
}
ret = create_val_field(hist_data, j++, file, field_str);
if (ret)
goto out;
}
if (fields_str && (strcmp(fields_str, "hitcount") != 0))
ret = -EINVAL;
out:
/* There is only raw hitcount but nohitcount suppresses it. */
if (j == 1 && hist_data->attrs->no_hitcount) {
hist_err(hist_data->event_file->tr, HIST_ERR_NEED_NOHC_VAL, 0);
ret = -ENOENT;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
return ret;
}
static int create_key_field(struct hist_trigger_data *hist_data,
unsigned int key_idx,
unsigned int key_offset,
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
struct trace_event_file *file,
char *field_str)
{
struct trace_array *tr = hist_data->event_file->tr;
struct hist_field *hist_field = NULL;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
unsigned long flags = 0;
unsigned int key_size;
int ret = 0, n_subexprs = 0;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (WARN_ON(key_idx >= HIST_FIELDS_MAX))
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
return -EINVAL;
flags |= HIST_FIELD_FL_KEY;
if (strcmp(field_str, "stacktrace") == 0) {
flags |= HIST_FIELD_FL_STACKTRACE;
key_size = sizeof(unsigned long) * HIST_STACKTRACE_DEPTH;
hist_field = create_hist_field(hist_data, NULL, flags, NULL);
} else {
hist_field = parse_expr(hist_data, file, field_str, flags,
NULL, &n_subexprs);
if (IS_ERR(hist_field)) {
ret = PTR_ERR(hist_field);
goto out;
}
if (field_has_hist_vars(hist_field, 0)) {
hist_err(tr, HIST_ERR_INVALID_REF_KEY, errpos(field_str));
destroy_hist_field(hist_field, 0);
ret = -EINVAL;
goto out;
}
key_size = hist_field->size;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
hist_data->fields[key_idx] = hist_field;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
key_size = ALIGN(key_size, sizeof(u64));
hist_data->fields[key_idx]->size = key_size;
hist_data->fields[key_idx]->offset = key_offset;
hist_data->key_size += key_size;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (hist_data->key_size > HIST_KEY_SIZE_MAX) {
ret = -EINVAL;
goto out;
}
hist_data->n_keys++;
hist_data->n_fields++;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (WARN_ON(hist_data->n_keys > TRACING_MAP_KEYS_MAX))
return -EINVAL;
ret = key_size;
out:
return ret;
}
static int create_key_fields(struct hist_trigger_data *hist_data,
struct trace_event_file *file)
{
unsigned int i, key_offset = 0, n_vals = hist_data->n_vals;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
char *fields_str, *field_str;
int ret = -EINVAL;
fields_str = hist_data->attrs->keys_str;
if (!fields_str)
goto out;
for (i = n_vals; i < n_vals + TRACING_MAP_KEYS_MAX; i++) {
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
field_str = strsep(&fields_str, ",");
if (!field_str)
break;
ret = create_key_field(hist_data, i, key_offset,
file, field_str);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (ret < 0)
goto out;
key_offset += ret;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
if (fields_str) {
ret = -EINVAL;
goto out;
}
ret = 0;
out:
return ret;
}
static int create_var_fields(struct hist_trigger_data *hist_data,
struct trace_event_file *file)
{
unsigned int i, j = hist_data->n_vals;
int ret = 0;
unsigned int n_vars = hist_data->attrs->var_defs.n_vars;
for (i = 0; i < n_vars; i++) {
char *var_name = hist_data->attrs->var_defs.name[i];
char *expr = hist_data->attrs->var_defs.expr[i];
ret = create_var_field(hist_data, j++, file, var_name, expr);
if (ret)
goto out;
}
out:
return ret;
}
static void free_var_defs(struct hist_trigger_data *hist_data)
{
unsigned int i;
for (i = 0; i < hist_data->attrs->var_defs.n_vars; i++) {
kfree(hist_data->attrs->var_defs.name[i]);
kfree(hist_data->attrs->var_defs.expr[i]);
}
hist_data->attrs->var_defs.n_vars = 0;
}
static int parse_var_defs(struct hist_trigger_data *hist_data)
{
struct trace_array *tr = hist_data->event_file->tr;
char *s, *str, *var_name, *field_str;
unsigned int i, j, n_vars = 0;
int ret = 0;
for (i = 0; i < hist_data->attrs->n_assignments; i++) {
str = hist_data->attrs->assignment_str[i];
for (j = 0; j < TRACING_MAP_VARS_MAX; j++) {
field_str = strsep(&str, ",");
if (!field_str)
break;
var_name = strsep(&field_str, "=");
if (!var_name || !field_str) {
hist_err(tr, HIST_ERR_MALFORMED_ASSIGNMENT,
errpos(var_name));
ret = -EINVAL;
goto free;
}
if (n_vars == TRACING_MAP_VARS_MAX) {
hist_err(tr, HIST_ERR_TOO_MANY_VARS, errpos(var_name));
ret = -EINVAL;
goto free;
}
s = kstrdup(var_name, GFP_KERNEL);
if (!s) {
ret = -ENOMEM;
goto free;
}
hist_data->attrs->var_defs.name[n_vars] = s;
s = kstrdup(field_str, GFP_KERNEL);
if (!s) {
tracing/histograms: Fix memory leak problem This reverts commit 46bbe5c671e06f070428b9be142cc4ee5cedebac. As commit 46bbe5c671e0 ("tracing: fix double free") said, the "double free" problem reported by clang static analyzer is: > In parse_var_defs() if there is a problem allocating > var_defs.expr, the earlier var_defs.name is freed. > This free is duplicated by free_var_defs() which frees > the rest of the list. However, if there is a problem allocating N-th var_defs.expr: + in parse_var_defs(), the freed 'earlier var_defs.name' is actually the N-th var_defs.name; + then in free_var_defs(), the names from 0th to (N-1)-th are freed; IF ALLOCATING PROBLEM HAPPENED HERE!!! -+ \ | 0th 1th (N-1)-th N-th V +-------------+-------------+-----+-------------+----------- var_defs: | name | expr | name | expr | ... | name | expr | name | /// +-------------+-------------+-----+-------------+----------- These two frees don't act on same name, so there was no "double free" problem before. Conversely, after that commit, we get a "memory leak" problem because the above "N-th var_defs.name" is not freed. If enable CONFIG_DEBUG_KMEMLEAK and inject a fault at where the N-th var_defs.expr allocated, then execute on shell like: $ echo 'hist:key=call_site:val=$v1,$v2:v1=bytes_req,v2=bytes_alloc' > \ /sys/kernel/debug/tracing/events/kmem/kmalloc/trigger Then kmemleak reports: unreferenced object 0xffff8fb100ef3518 (size 8): comm "bash", pid 196, jiffies 4295681690 (age 28.538s) hex dump (first 8 bytes): 76 31 00 00 b1 8f ff ff v1...... backtrace: [<0000000038fe4895>] kstrdup+0x2d/0x60 [<00000000c99c049a>] event_hist_trigger_parse+0x206f/0x20e0 [<00000000ae70d2cc>] trigger_process_regex+0xc0/0x110 [<0000000066737a4c>] event_trigger_write+0x75/0xd0 [<000000007341e40c>] vfs_write+0xbb/0x2a0 [<0000000087fde4c2>] ksys_write+0x59/0xd0 [<00000000581e9cdf>] do_syscall_64+0x3a/0x80 [<00000000cf3b065c>] entry_SYSCALL_64_after_hwframe+0x46/0xb0 Link: https://lkml.kernel.org/r/20220711014731.69520-1-zhengyejian1@huawei.com Cc: stable@vger.kernel.org Fixes: 46bbe5c671e0 ("tracing: fix double free") Reported-by: Hulk Robot <hulkci@huawei.com> Suggested-by: Steven Rostedt <rostedt@goodmis.org> Reviewed-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-07-11 01:47:31 +00:00
kfree(hist_data->attrs->var_defs.name[n_vars]);
hist_data->attrs->var_defs.name[n_vars] = NULL;
ret = -ENOMEM;
goto free;
}
hist_data->attrs->var_defs.expr[n_vars++] = s;
hist_data->attrs->var_defs.n_vars = n_vars;
}
}
return ret;
free:
free_var_defs(hist_data);
return ret;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
static int create_hist_fields(struct hist_trigger_data *hist_data,
struct trace_event_file *file)
{
int ret;
ret = parse_var_defs(hist_data);
if (ret)
return ret;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
ret = create_val_fields(hist_data, file);
if (ret)
goto out;
ret = create_var_fields(hist_data, file);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (ret)
goto out;
ret = create_key_fields(hist_data, file);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
out:
free_var_defs(hist_data);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
return ret;
}
static int is_descending(struct trace_array *tr, const char *str)
{
if (!str)
return 0;
if (strcmp(str, "descending") == 0)
return 1;
if (strcmp(str, "ascending") == 0)
return 0;
hist_err(tr, HIST_ERR_INVALID_SORT_MODIFIER, errpos((char *)str));
return -EINVAL;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
static int create_sort_keys(struct hist_trigger_data *hist_data)
{
struct trace_array *tr = hist_data->event_file->tr;
char *fields_str = hist_data->attrs->sort_key_str;
struct tracing_map_sort_key *sort_key;
int descending, ret = 0;
unsigned int i, j, k;
hist_data->n_sort_keys = 1; /* we always have at least one, hitcount */
if (!fields_str)
goto out;
for (i = 0; i < TRACING_MAP_SORT_KEYS_MAX; i++) {
struct hist_field *hist_field;
char *field_str, *field_name;
const char *test_name;
sort_key = &hist_data->sort_keys[i];
field_str = strsep(&fields_str, ",");
if (!field_str)
break;
if (!*field_str) {
ret = -EINVAL;
hist_err(tr, HIST_ERR_EMPTY_SORT_FIELD, errpos("sort="));
break;
}
if ((i == TRACING_MAP_SORT_KEYS_MAX - 1) && fields_str) {
hist_err(tr, HIST_ERR_TOO_MANY_SORT_FIELDS, errpos("sort="));
ret = -EINVAL;
break;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
field_name = strsep(&field_str, ".");
if (!field_name || !*field_name) {
ret = -EINVAL;
hist_err(tr, HIST_ERR_EMPTY_SORT_FIELD, errpos("sort="));
break;
}
if (strcmp(field_name, "hitcount") == 0) {
descending = is_descending(tr, field_str);
if (descending < 0) {
ret = descending;
break;
}
sort_key->descending = descending;
continue;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
for (j = 1, k = 1; j < hist_data->n_fields; j++) {
unsigned int idx;
hist_field = hist_data->fields[j];
if (hist_field->flags & HIST_FIELD_FL_VAR)
continue;
idx = k++;
test_name = hist_field_name(hist_field, 0);
if (strcmp(field_name, test_name) == 0) {
sort_key->field_idx = idx;
descending = is_descending(tr, field_str);
if (descending < 0) {
ret = descending;
goto out;
}
sort_key->descending = descending;
break;
}
}
if (j == hist_data->n_fields) {
ret = -EINVAL;
hist_err(tr, HIST_ERR_INVALID_SORT_FIELD, errpos(field_name));
break;
}
}
hist_data->n_sort_keys = i;
out:
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
return ret;
}
static void destroy_actions(struct hist_trigger_data *hist_data)
{
unsigned int i;
for (i = 0; i < hist_data->n_actions; i++) {
struct action_data *data = hist_data->actions[i];
if (data->handler == HANDLER_ONMATCH)
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
onmatch_destroy(data);
else if (data->handler == HANDLER_ONMAX ||
data->handler == HANDLER_ONCHANGE)
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
track_data_destroy(hist_data, data);
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
else
kfree(data);
}
}
static int parse_actions(struct hist_trigger_data *hist_data)
{
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
struct trace_array *tr = hist_data->event_file->tr;
struct action_data *data;
unsigned int i;
int ret = 0;
char *str;
int len;
for (i = 0; i < hist_data->attrs->n_actions; i++) {
str = hist_data->attrs->action_str[i];
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
if ((len = str_has_prefix(str, "onmatch("))) {
char *action_str = str + len;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
data = onmatch_parse(tr, action_str);
if (IS_ERR(data)) {
ret = PTR_ERR(data);
break;
}
} else if ((len = str_has_prefix(str, "onmax("))) {
char *action_str = str + len;
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
data = track_data_parse(hist_data, action_str,
HANDLER_ONMAX);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
if (IS_ERR(data)) {
ret = PTR_ERR(data);
break;
}
} else if ((len = str_has_prefix(str, "onchange("))) {
char *action_str = str + len;
data = track_data_parse(hist_data, action_str,
HANDLER_ONCHANGE);
if (IS_ERR(data)) {
ret = PTR_ERR(data);
break;
}
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
} else {
ret = -EINVAL;
break;
}
hist_data->actions[hist_data->n_actions++] = data;
}
return ret;
}
static int create_actions(struct hist_trigger_data *hist_data)
{
struct action_data *data;
unsigned int i;
int ret = 0;
for (i = 0; i < hist_data->attrs->n_actions; i++) {
data = hist_data->actions[i];
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
if (data->handler == HANDLER_ONMATCH) {
ret = onmatch_create(hist_data, data);
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
if (ret)
break;
} else if (data->handler == HANDLER_ONMAX ||
data->handler == HANDLER_ONCHANGE) {
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
ret = track_data_create(hist_data, data);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
if (ret)
break;
} else {
ret = -EINVAL;
break;
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
}
}
return ret;
}
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
static void print_actions(struct seq_file *m,
struct hist_trigger_data *hist_data,
struct tracing_map_elt *elt)
{
unsigned int i;
for (i = 0; i < hist_data->n_actions; i++) {
struct action_data *data = hist_data->actions[i];
if (data->action == ACTION_SNAPSHOT)
continue;
if (data->handler == HANDLER_ONMAX ||
data->handler == HANDLER_ONCHANGE)
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
track_data_print(m, hist_data, elt, data);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
}
}
static void print_action_spec(struct seq_file *m,
struct hist_trigger_data *hist_data,
struct action_data *data)
{
unsigned int i;
if (data->action == ACTION_SAVE) {
for (i = 0; i < hist_data->n_save_vars; i++) {
seq_printf(m, "%s", hist_data->save_vars[i]->var->var.name);
if (i < hist_data->n_save_vars - 1)
seq_puts(m, ",");
}
} else if (data->action == ACTION_TRACE) {
if (data->use_trace_keyword)
seq_printf(m, "%s", data->synth_event_name);
for (i = 0; i < data->n_params; i++) {
if (i || data->use_trace_keyword)
seq_puts(m, ",");
seq_printf(m, "%s", data->params[i]);
}
}
}
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
static void print_track_data_spec(struct seq_file *m,
struct hist_trigger_data *hist_data,
struct action_data *data)
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
{
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
if (data->handler == HANDLER_ONMAX)
seq_puts(m, ":onmax(");
else if (data->handler == HANDLER_ONCHANGE)
seq_puts(m, ":onchange(");
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
seq_printf(m, "%s", data->track_data.var_str);
seq_printf(m, ").%s(", data->action_name);
print_action_spec(m, hist_data, data);
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
seq_puts(m, ")");
}
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
static void print_onmatch_spec(struct seq_file *m,
struct hist_trigger_data *hist_data,
struct action_data *data)
{
seq_printf(m, ":onmatch(%s.%s).", data->match_data.event_system,
data->match_data.event);
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
seq_printf(m, "%s(", data->action_name);
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
print_action_spec(m, hist_data, data);
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
seq_puts(m, ")");
}
tracing: Add action comparisons when testing matching hist triggers Actions also need to be considered when checking for matching triggers - triggers differing only by action should be allowed, but currently aren't because the matching check ignores the action and erroneously returns -EEXIST. Add and call an actions_match() function to address that. Here's an example using onmatch() actions. The first -EEXIST shouldn't occur because the onmatch() is different in the second wakeup_latency() param. The second -EEXIST shouldn't occur because it's a different action (in this case, it doesn't have an action, so shouldn't be seen as being the same and therefore rejected). In the after case, both are correctly accepted (and trying to add one of them again returns -EEXIST as it should). before: # echo 'wakeup_latency u64 lat; pid_t pid' >> /sys/kernel/debug/tracing/synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0 if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,next_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,prev_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger -su: echo: write error: File exists # echo 'hist:keys=next_pid if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger -su: echo: write error: File exists after: # echo 'wakeup_latency u64 lat; pid_t pid' >> /sys/kernel/debug/tracing/synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0 if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,next_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,prev_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/a7fd668b87ec10736c8f016ac4279c8480d50c2b.1522256721.git.tom.zanussi@linux.intel.com Tested-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-03-28 20:10:55 +00:00
static bool actions_match(struct hist_trigger_data *hist_data,
struct hist_trigger_data *hist_data_test)
{
unsigned int i, j;
if (hist_data->n_actions != hist_data_test->n_actions)
return false;
for (i = 0; i < hist_data->n_actions; i++) {
struct action_data *data = hist_data->actions[i];
struct action_data *data_test = hist_data_test->actions[i];
char *action_name, *action_name_test;
tracing: Add action comparisons when testing matching hist triggers Actions also need to be considered when checking for matching triggers - triggers differing only by action should be allowed, but currently aren't because the matching check ignores the action and erroneously returns -EEXIST. Add and call an actions_match() function to address that. Here's an example using onmatch() actions. The first -EEXIST shouldn't occur because the onmatch() is different in the second wakeup_latency() param. The second -EEXIST shouldn't occur because it's a different action (in this case, it doesn't have an action, so shouldn't be seen as being the same and therefore rejected). In the after case, both are correctly accepted (and trying to add one of them again returns -EEXIST as it should). before: # echo 'wakeup_latency u64 lat; pid_t pid' >> /sys/kernel/debug/tracing/synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0 if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,next_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,prev_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger -su: echo: write error: File exists # echo 'hist:keys=next_pid if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger -su: echo: write error: File exists after: # echo 'wakeup_latency u64 lat; pid_t pid' >> /sys/kernel/debug/tracing/synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0 if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,next_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,prev_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/a7fd668b87ec10736c8f016ac4279c8480d50c2b.1522256721.git.tom.zanussi@linux.intel.com Tested-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-03-28 20:10:55 +00:00
if (data->handler != data_test->handler)
return false;
if (data->action != data_test->action)
tracing: Add action comparisons when testing matching hist triggers Actions also need to be considered when checking for matching triggers - triggers differing only by action should be allowed, but currently aren't because the matching check ignores the action and erroneously returns -EEXIST. Add and call an actions_match() function to address that. Here's an example using onmatch() actions. The first -EEXIST shouldn't occur because the onmatch() is different in the second wakeup_latency() param. The second -EEXIST shouldn't occur because it's a different action (in this case, it doesn't have an action, so shouldn't be seen as being the same and therefore rejected). In the after case, both are correctly accepted (and trying to add one of them again returns -EEXIST as it should). before: # echo 'wakeup_latency u64 lat; pid_t pid' >> /sys/kernel/debug/tracing/synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0 if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,next_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,prev_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger -su: echo: write error: File exists # echo 'hist:keys=next_pid if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger -su: echo: write error: File exists after: # echo 'wakeup_latency u64 lat; pid_t pid' >> /sys/kernel/debug/tracing/synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0 if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,next_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,prev_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/a7fd668b87ec10736c8f016ac4279c8480d50c2b.1522256721.git.tom.zanussi@linux.intel.com Tested-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-03-28 20:10:55 +00:00
return false;
if (data->n_params != data_test->n_params)
return false;
for (j = 0; j < data->n_params; j++) {
if (strcmp(data->params[j], data_test->params[j]) != 0)
return false;
}
if (data->use_trace_keyword)
action_name = data->synth_event_name;
else
action_name = data->action_name;
if (data_test->use_trace_keyword)
action_name_test = data_test->synth_event_name;
else
action_name_test = data_test->action_name;
if (strcmp(action_name, action_name_test) != 0)
return false;
if (data->handler == HANDLER_ONMATCH) {
if (strcmp(data->match_data.event_system,
data_test->match_data.event_system) != 0)
tracing: Add action comparisons when testing matching hist triggers Actions also need to be considered when checking for matching triggers - triggers differing only by action should be allowed, but currently aren't because the matching check ignores the action and erroneously returns -EEXIST. Add and call an actions_match() function to address that. Here's an example using onmatch() actions. The first -EEXIST shouldn't occur because the onmatch() is different in the second wakeup_latency() param. The second -EEXIST shouldn't occur because it's a different action (in this case, it doesn't have an action, so shouldn't be seen as being the same and therefore rejected). In the after case, both are correctly accepted (and trying to add one of them again returns -EEXIST as it should). before: # echo 'wakeup_latency u64 lat; pid_t pid' >> /sys/kernel/debug/tracing/synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0 if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,next_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,prev_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger -su: echo: write error: File exists # echo 'hist:keys=next_pid if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger -su: echo: write error: File exists after: # echo 'wakeup_latency u64 lat; pid_t pid' >> /sys/kernel/debug/tracing/synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0 if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,next_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,prev_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/a7fd668b87ec10736c8f016ac4279c8480d50c2b.1522256721.git.tom.zanussi@linux.intel.com Tested-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-03-28 20:10:55 +00:00
return false;
if (strcmp(data->match_data.event,
data_test->match_data.event) != 0)
tracing: Add action comparisons when testing matching hist triggers Actions also need to be considered when checking for matching triggers - triggers differing only by action should be allowed, but currently aren't because the matching check ignores the action and erroneously returns -EEXIST. Add and call an actions_match() function to address that. Here's an example using onmatch() actions. The first -EEXIST shouldn't occur because the onmatch() is different in the second wakeup_latency() param. The second -EEXIST shouldn't occur because it's a different action (in this case, it doesn't have an action, so shouldn't be seen as being the same and therefore rejected). In the after case, both are correctly accepted (and trying to add one of them again returns -EEXIST as it should). before: # echo 'wakeup_latency u64 lat; pid_t pid' >> /sys/kernel/debug/tracing/synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0 if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,next_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,prev_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger -su: echo: write error: File exists # echo 'hist:keys=next_pid if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger -su: echo: write error: File exists after: # echo 'wakeup_latency u64 lat; pid_t pid' >> /sys/kernel/debug/tracing/synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0 if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,next_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,prev_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/a7fd668b87ec10736c8f016ac4279c8480d50c2b.1522256721.git.tom.zanussi@linux.intel.com Tested-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-03-28 20:10:55 +00:00
return false;
} else if (data->handler == HANDLER_ONMAX ||
data->handler == HANDLER_ONCHANGE) {
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
if (strcmp(data->track_data.var_str,
data_test->track_data.var_str) != 0)
tracing: Add action comparisons when testing matching hist triggers Actions also need to be considered when checking for matching triggers - triggers differing only by action should be allowed, but currently aren't because the matching check ignores the action and erroneously returns -EEXIST. Add and call an actions_match() function to address that. Here's an example using onmatch() actions. The first -EEXIST shouldn't occur because the onmatch() is different in the second wakeup_latency() param. The second -EEXIST shouldn't occur because it's a different action (in this case, it doesn't have an action, so shouldn't be seen as being the same and therefore rejected). In the after case, both are correctly accepted (and trying to add one of them again returns -EEXIST as it should). before: # echo 'wakeup_latency u64 lat; pid_t pid' >> /sys/kernel/debug/tracing/synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0 if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,next_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,prev_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger -su: echo: write error: File exists # echo 'hist:keys=next_pid if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger -su: echo: write error: File exists after: # echo 'wakeup_latency u64 lat; pid_t pid' >> /sys/kernel/debug/tracing/synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0 if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,next_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,prev_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/a7fd668b87ec10736c8f016ac4279c8480d50c2b.1522256721.git.tom.zanussi@linux.intel.com Tested-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-03-28 20:10:55 +00:00
return false;
}
}
return true;
}
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
static void print_actions_spec(struct seq_file *m,
struct hist_trigger_data *hist_data)
{
unsigned int i;
for (i = 0; i < hist_data->n_actions; i++) {
struct action_data *data = hist_data->actions[i];
if (data->handler == HANDLER_ONMATCH)
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
print_onmatch_spec(m, hist_data, data);
else if (data->handler == HANDLER_ONMAX ||
data->handler == HANDLER_ONCHANGE)
tracing: Generalize hist trigger onmax and save action The action refactor code allowed actions and handlers to be separated, but the existing onmax handler and save action code is still not flexible enough to handle arbitrary coupling. This change generalizes them and in the process makes additional handlers and actions easier to implement. The onmax action can be broken up and thought of as two separate components - a variable to be tracked (the parameter given to the onmax($var_to_track) function) and an invisible variable created to save the ongoing result of doing something with that variable, such as saving the max value of that variable so far seen. Separating it out like this and renaming it appropriately allows us to use the same code for similar tracking functions such as onchange($var_to_track), which would just track the last value seen rather than the max seen so far, which is useful in some situations. Additionally, because different handlers and actions may want to save and access data differently e.g. save and retrieve tracking values as local variables vs something more global, save_val() and get_val() interface functions are introduced and max-specific implementations are used instead. The same goes for the code that checks whether a maximum has been hit - a generic check_val() interface and max-checking implementation is used instead, which allows future patches to make use of he same code using their own implemetations of similar functionality. Link: http://lkml.kernel.org/r/980ea73dd8e3f36db3d646f99652f8fed42b77d4.1550100284.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-13 23:42:44 +00:00
print_track_data_spec(m, hist_data, data);
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
}
}
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
static void destroy_field_var_hists(struct hist_trigger_data *hist_data)
{
unsigned int i;
for (i = 0; i < hist_data->n_field_var_hists; i++) {
kfree(hist_data->field_var_hists[i]->cmd);
kfree(hist_data->field_var_hists[i]);
}
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
static void destroy_hist_data(struct hist_trigger_data *hist_data)
{
if (!hist_data)
return;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
destroy_hist_trigger_attrs(hist_data->attrs);
destroy_hist_fields(hist_data);
tracing_map_destroy(hist_data->map);
destroy_actions(hist_data);
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
destroy_field_vars(hist_data);
destroy_field_var_hists(hist_data);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
kfree(hist_data);
}
static int create_tracing_map_fields(struct hist_trigger_data *hist_data)
{
struct tracing_map *map = hist_data->map;
struct ftrace_event_field *field;
struct hist_field *hist_field;
int i, idx = 0;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
for_each_hist_field(i, hist_data) {
hist_field = hist_data->fields[i];
if (hist_field->flags & HIST_FIELD_FL_KEY) {
tracing_map_cmp_fn_t cmp_fn;
field = hist_field->field;
if (hist_field->flags & HIST_FIELD_FL_STACKTRACE)
cmp_fn = tracing_map_cmp_none;
tracing/histogram: Fix sorting on old "cpu" value When trying to add a histogram against an event with the "cpu" field, it was impossible due to "cpu" being a keyword to key off of the running CPU. So to fix this, it was changed to "common_cpu" to match the other generic fields (like "common_pid"). But since some scripts used "cpu" for keying off of the CPU (for events that did not have "cpu" as a field, which is most of them), a backward compatibility trick was added such that if "cpu" was used as a key, and the event did not have "cpu" as a field name, then it would fallback and switch over to "common_cpu". This fix has a couple of subtle bugs. One was that when switching over to "common_cpu", it did not change the field name, it just set a flag. But the code still found a "cpu" field. The "cpu" field is used for filtering and is returned when the event does not have a "cpu" field. This was found by: # cd /sys/kernel/tracing # echo hist:key=cpu,pid:sort=cpu > events/sched/sched_wakeup/trigger # cat events/sched/sched_wakeup/hist Which showed the histogram unsorted: { cpu: 19, pid: 1175 } hitcount: 1 { cpu: 6, pid: 239 } hitcount: 2 { cpu: 23, pid: 1186 } hitcount: 14 { cpu: 12, pid: 249 } hitcount: 2 { cpu: 3, pid: 994 } hitcount: 5 Instead of hard coding the "cpu" checks, take advantage of the fact that trace_event_field_field() returns a special field for "cpu" and "CPU" if the event does not have "cpu" as a field. This special field has the "filter_type" of "FILTER_CPU". Check that to test if the returned field is of the CPU type instead of doing the string compare. Also, fix the sorting bug by testing for the hist_field flag of HIST_FIELD_FL_CPU when setting up the sort routine. Otherwise it will use the special CPU field to know what compare routine to use, and since that special field does not have a size, it returns tracing_map_cmp_none. Cc: stable@vger.kernel.org Fixes: 1e3bac71c505 ("tracing/histogram: Rename "cpu" to "common_cpu"") Reported-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-03-02 03:29:04 +00:00
else if (!field || hist_field->flags & HIST_FIELD_FL_CPU)
cmp_fn = tracing_map_cmp_num(hist_field->size,
hist_field->is_signed);
else if (is_string_field(field))
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
cmp_fn = tracing_map_cmp_string;
else
cmp_fn = tracing_map_cmp_num(field->size,
field->is_signed);
idx = tracing_map_add_key_field(map,
hist_field->offset,
cmp_fn);
} else if (!(hist_field->flags & HIST_FIELD_FL_VAR))
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
idx = tracing_map_add_sum_field(map);
if (idx < 0)
return idx;
if (hist_field->flags & HIST_FIELD_FL_VAR) {
idx = tracing_map_add_var(map);
if (idx < 0)
return idx;
hist_field->var.idx = idx;
hist_field->var.hist_data = hist_data;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
return 0;
}
static struct hist_trigger_data *
create_hist_data(unsigned int map_bits,
struct hist_trigger_attrs *attrs,
struct trace_event_file *file,
bool remove)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
{
const struct tracing_map_ops *map_ops = NULL;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
struct hist_trigger_data *hist_data;
int ret = 0;
hist_data = kzalloc(sizeof(*hist_data), GFP_KERNEL);
if (!hist_data)
return ERR_PTR(-ENOMEM);
hist_data->attrs = attrs;
hist_data->remove = remove;
hist_data->event_file = file;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
ret = parse_actions(hist_data);
if (ret)
goto free;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
ret = create_hist_fields(hist_data, file);
if (ret)
goto free;
ret = create_sort_keys(hist_data);
if (ret)
goto free;
map_ops = &hist_trigger_elt_data_ops;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
hist_data->map = tracing_map_create(map_bits, hist_data->key_size,
map_ops, hist_data);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (IS_ERR(hist_data->map)) {
ret = PTR_ERR(hist_data->map);
hist_data->map = NULL;
goto free;
}
ret = create_tracing_map_fields(hist_data);
if (ret)
goto free;
out:
return hist_data;
free:
hist_data->attrs = NULL;
destroy_hist_data(hist_data);
hist_data = ERR_PTR(ret);
goto out;
}
static void hist_trigger_elt_update(struct hist_trigger_data *hist_data,
struct tracing_map_elt *elt,
struct trace_buffer *buffer, void *rec,
struct ring_buffer_event *rbe,
u64 *var_ref_vals)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
{
struct hist_elt_data *elt_data;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
struct hist_field *hist_field;
unsigned int i, var_idx;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
u64 hist_val;
elt_data = elt->private_data;
elt_data->var_ref_vals = var_ref_vals;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
for_each_hist_val_field(i, hist_data) {
hist_field = hist_data->fields[i];
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
hist_val = hist_fn_call(hist_field, elt, buffer, rbe, rec);
if (hist_field->flags & HIST_FIELD_FL_VAR) {
var_idx = hist_field->var.idx;
if (hist_field->flags &
(HIST_FIELD_FL_STRING | HIST_FIELD_FL_STACKTRACE)) {
unsigned int str_start, var_str_idx, idx;
char *str, *val_str;
unsigned int size;
str_start = hist_data->n_field_var_str +
hist_data->n_save_var_str;
var_str_idx = hist_field->var_str_idx;
idx = str_start + var_str_idx;
str = elt_data->field_var_str[idx];
val_str = (char *)(uintptr_t)hist_val;
if (hist_field->flags & HIST_FIELD_FL_STRING) {
size = min(hist_field->size, STR_VAR_LEN_MAX);
strscpy(str, val_str, size);
} else {
char *stack_start = str + sizeof(unsigned long);
int e;
e = stack_trace_save((void *)stack_start,
HIST_STACKTRACE_DEPTH,
HIST_STACKTRACE_SKIP);
if (e < HIST_STACKTRACE_DEPTH - 1)
((unsigned long *)stack_start)[e] = 0;
*((unsigned long *)str) = e;
}
hist_val = (u64)(uintptr_t)str;
}
tracing_map_set_var(elt, var_idx, hist_val);
continue;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
tracing_map_update_sum(elt, i, hist_val);
}
for_each_hist_key_field(i, hist_data) {
hist_field = hist_data->fields[i];
if (hist_field->flags & HIST_FIELD_FL_VAR) {
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
hist_val = hist_fn_call(hist_field, elt, buffer, rbe, rec);
var_idx = hist_field->var.idx;
tracing_map_set_var(elt, var_idx, hist_val);
}
}
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
update_field_vars(hist_data, elt, buffer, rbe, rec);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
static inline void add_to_key(char *compound_key, void *key,
struct hist_field *key_field, void *rec)
{
size_t size = key_field->size;
if (key_field->flags & HIST_FIELD_FL_STRING) {
struct ftrace_event_field *field;
field = key_field->field;
tracing: Support __rel_loc relative dynamic data location attribute Add '__rel_loc' new dynamic data location attribute which encodes the data location from the next to the field itself. The '__data_loc' is used for encoding the dynamic data location on the trace event record. But '__data_loc' is not useful if the writer doesn't know the event header (e.g. user event), because it records the dynamic data offset from the entry of the record, not the field itself. This new '__rel_loc' attribute encodes the data location relatively from the next of the field. For example, when there is a record like below (the number in the parentheses is the size of fields) |header(N)|common(M)|fields(K)|__data_loc(4)|fields(L)|data(G)| In this case, '__data_loc' field will be __data_loc = (G << 16) | (N+M+K+4+L) If '__rel_loc' is used, this will be |header(N)|common(M)|fields(K)|__rel_loc(4)|fields(L)|data(G)| where __rel_loc = (G << 16) | (L) This case shows L bytes after the '__rel_loc' attribute field, if there is no fields after the __rel_loc field, L must be 0. This is relatively easy (and no need to consider the kernel header change) when the event data fields are composed by user who doesn't know header and common fields. Link: https://lkml.kernel.org/r/163757341258.510314.4214431827833229956.stgit@devnote2 Cc: Beau Belgrave <beaub@linux.microsoft.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-11-22 09:30:12 +00:00
if (field->filter_type == FILTER_DYN_STRING ||
field->filter_type == FILTER_RDYN_STRING)
size = *(u32 *)(rec + field->offset) >> 16;
else if (field->filter_type == FILTER_STATIC_STRING)
size = field->size;
/* ensure NULL-termination */
if (size > key_field->size - 1)
size = key_field->size - 1;
tracing: Use strncpy instead of memcpy for string keys in hist triggers Because there may be random garbage beyond a string's null terminator, it's not correct to copy the the complete character array for use as a hist trigger key. This results in multiple histogram entries for the 'same' string key. So, in the case of a string key, use strncpy instead of memcpy to avoid copying in the extra bytes. Before, using the gdbus entries in the following hist trigger as an example: # echo 'hist:key=comm' > /sys/kernel/debug/tracing/events/sched/sched_waking/trigger # cat /sys/kernel/debug/tracing/events/sched/sched_waking/hist ... { comm: ImgDecoder #4 } hitcount: 203 { comm: gmain } hitcount: 213 { comm: gmain } hitcount: 216 { comm: StreamTrans #73 } hitcount: 221 { comm: mozStorage #3 } hitcount: 230 { comm: gdbus } hitcount: 233 { comm: StyleThread#5 } hitcount: 253 { comm: gdbus } hitcount: 256 { comm: gdbus } hitcount: 260 { comm: StyleThread#4 } hitcount: 271 ... # cat /sys/kernel/debug/tracing/events/sched/sched_waking/hist | egrep gdbus | wc -l 51 After: # cat /sys/kernel/debug/tracing/events/sched/sched_waking/hist | egrep gdbus | wc -l 1 Link: http://lkml.kernel.org/r/50c35ae1267d64eee975b8125e151e600071d4dc.1549309756.git.tom.zanussi@linux.intel.com Cc: Namhyung Kim <namhyung@kernel.org> Cc: stable@vger.kernel.org Fixes: 79e577cbce4c4 ("tracing: Support string type key properly") Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-02-04 21:07:24 +00:00
strncpy(compound_key + key_field->offset, (char *)key, size);
} else
memcpy(compound_key + key_field->offset, key, size);
}
static void
hist_trigger_actions(struct hist_trigger_data *hist_data,
struct tracing_map_elt *elt,
struct trace_buffer *buffer, void *rec,
struct ring_buffer_event *rbe, void *key,
u64 *var_ref_vals)
{
struct action_data *data;
unsigned int i;
for (i = 0; i < hist_data->n_actions; i++) {
data = hist_data->actions[i];
data->fn(hist_data, elt, buffer, rec, rbe, key, data, var_ref_vals);
}
}
static void event_hist_trigger(struct event_trigger_data *data,
struct trace_buffer *buffer, void *rec,
struct ring_buffer_event *rbe)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
{
struct hist_trigger_data *hist_data = data->private_data;
bool use_compound_key = (hist_data->n_keys > 1);
unsigned long entries[HIST_STACKTRACE_DEPTH];
u64 var_ref_vals[TRACING_MAP_VARS_MAX];
char compound_key[HIST_KEY_SIZE_MAX];
struct tracing_map_elt *elt = NULL;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
struct hist_field *key_field;
u64 field_contents;
void *key = NULL;
unsigned int i;
if (unlikely(!rbe))
return;
memset(compound_key, 0, hist_data->key_size);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
for_each_hist_key_field(i, hist_data) {
key_field = hist_data->fields[i];
if (key_field->flags & HIST_FIELD_FL_STACKTRACE) {
tracing: Simplify stacktrace retrieval in histograms The indirection through struct stack_trace is not necessary at all. Use the storage array based interface. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Tom Zanussi <tom.zanussi@linux.intel.com> Reviewed-by: Tom Zanussi <tom.zanussi@linux.intel.com> Reviewed-by: Josh Poimboeuf <jpoimboe@redhat.com> Acked-by: Steven Rostedt (VMware) <rostedt@goodmis.org> Cc: Andy Lutomirski <luto@kernel.org> Cc: Alexander Potapenko <glider@google.com> Cc: Alexey Dobriyan <adobriyan@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: linux-mm@kvack.org Cc: David Rientjes <rientjes@google.com> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Dmitry Vyukov <dvyukov@google.com> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: kasan-dev@googlegroups.com Cc: Mike Rapoport <rppt@linux.vnet.ibm.com> Cc: Akinobu Mita <akinobu.mita@gmail.com> Cc: Christoph Hellwig <hch@lst.de> Cc: iommu@lists.linux-foundation.org Cc: Robin Murphy <robin.murphy@arm.com> Cc: Marek Szyprowski <m.szyprowski@samsung.com> Cc: Johannes Thumshirn <jthumshirn@suse.de> Cc: David Sterba <dsterba@suse.com> Cc: Chris Mason <clm@fb.com> Cc: Josef Bacik <josef@toxicpanda.com> Cc: linux-btrfs@vger.kernel.org Cc: dm-devel@redhat.com Cc: Mike Snitzer <snitzer@redhat.com> Cc: Alasdair Kergon <agk@redhat.com> Cc: Daniel Vetter <daniel@ffwll.ch> Cc: intel-gfx@lists.freedesktop.org Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: dri-devel@lists.freedesktop.org Cc: David Airlie <airlied@linux.ie> Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Miroslav Benes <mbenes@suse.cz> Cc: linux-arch@vger.kernel.org Link: https://lkml.kernel.org/r/20190425094802.979089273@linutronix.de
2019-04-25 09:45:13 +00:00
memset(entries, 0, HIST_STACKTRACE_SIZE);
tracing/histogram: Fix stacktrace key The current code will always use the current stacktrace as a key even if a stacktrace contained in a specific event field was specified. For example, we expect to use the 'unsigned long[] stack' field in the below event in the histogram: # echo 's:block_lat pid_t pid; u64 delta; unsigned long[] stack;' > /sys/kernel/debug/tracing/dynamic_events # echo 'hist:keys=delta.buckets=100,stack.stacktrace:sort=delta' > /sys/kernel/debug/tracing/events/synthetic/block_lat/trigger But in fact, when we type out the trigger, we see that it's using the plain old global 'stacktrace' as the key, which is just the stacktrace when the event was hit and not the stacktrace contained in the event, which is what we want: # cat /sys/kernel/debug/tracing/events/synthetic/block_lat/trigger hist:keys=delta.buckets=100,stacktrace:vals=hitcount:sort=delta.buckets=100:size=2048 [active] And in fact, there's no code to actually retrieve it from the event, so we need to add HIST_FIELD_FN_STACK and hist_field_stack() to get it and hook it into the trigger code. For now, since the stack is just using dynamic strings, this could just use the dynamic string function, but it seems cleaner to have a dedicated function an be able to tweak independently as necessary. Link: https://lkml.kernel.org/r/11aa614c82976adbfa4ea763dbe885b5fb01d59c.1676063532.git.zanussi@kernel.org Signed-off-by: Tom Zanussi <zanussi@kernel.org> [ Fixed 32bit build warning reported by kernel test robot <lkp@intel.com> ] Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2023-02-10 21:33:05 +00:00
if (key_field->field) {
unsigned long *stack, n_entries;
field_contents = hist_fn_call(key_field, elt, buffer, rbe, rec);
stack = (unsigned long *)(long)field_contents;
n_entries = *stack;
memcpy(entries, ++stack, n_entries * sizeof(unsigned long));
} else {
stack_trace_save(entries, HIST_STACKTRACE_DEPTH,
HIST_STACKTRACE_SKIP);
}
key = entries;
} else {
tracing/hist: Call hist functions directly via a switch statement Due to retpolines, indirect calls are much more expensive than direct calls. The histograms have a select set of functions it uses for the histograms, instead of using function pointers to call them, create a hist_fn_call() function that uses a switch statement to call the histogram functions directly. This gives a 13% speedup to the histogram logic. Using the histogram benchmark: Before: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 129 } hitcount: 2213 { delta: 130 } hitcount: 285965 { delta: 131 } hitcount: 1146545 { delta: 132 } hitcount: 5185432 { delta: 133 } hitcount: 19896215 { delta: 134 } hitcount: 53118616 { delta: 135 } hitcount: 83816709 { delta: 136 } hitcount: 68329562 { delta: 137 } hitcount: 41859349 { delta: 138 } hitcount: 46257797 { delta: 139 } hitcount: 54400831 { delta: 140 } hitcount: 72875007 { delta: 141 } hitcount: 76193272 { delta: 142 } hitcount: 49504263 { delta: 143 } hitcount: 38821072 { delta: 144 } hitcount: 47702679 { delta: 145 } hitcount: 41357297 { delta: 146 } hitcount: 22058238 { delta: 147 } hitcount: 9720002 { delta: 148 } hitcount: 3193542 { delta: 149 } hitcount: 927030 { delta: 150 } hitcount: 850772 { delta: 151 } hitcount: 1477380 { delta: 152 } hitcount: 2687977 { delta: 153 } hitcount: 2865985 { delta: 154 } hitcount: 1977492 { delta: 155 } hitcount: 2475607 { delta: 156 } hitcount: 3403612 After: # event histogram # # trigger info: hist:keys=delta:vals=hitcount:sort=delta:size=2048 if delta > 0 [active] # { delta: 113 } hitcount: 272 { delta: 114 } hitcount: 840 { delta: 118 } hitcount: 344 { delta: 119 } hitcount: 25428 { delta: 120 } hitcount: 350590 { delta: 121 } hitcount: 1892484 { delta: 122 } hitcount: 6205004 { delta: 123 } hitcount: 11583521 { delta: 124 } hitcount: 37590979 { delta: 125 } hitcount: 108308504 { delta: 126 } hitcount: 131672461 { delta: 127 } hitcount: 88700598 { delta: 128 } hitcount: 65939870 { delta: 129 } hitcount: 45055004 { delta: 130 } hitcount: 33174464 { delta: 131 } hitcount: 31813493 { delta: 132 } hitcount: 29011676 { delta: 133 } hitcount: 22798782 { delta: 134 } hitcount: 22072486 { delta: 135 } hitcount: 17034113 { delta: 136 } hitcount: 8982490 { delta: 137 } hitcount: 2865908 { delta: 138 } hitcount: 980382 { delta: 139 } hitcount: 1651944 { delta: 140 } hitcount: 4112073 { delta: 141 } hitcount: 3963269 { delta: 142 } hitcount: 1712508 { delta: 143 } hitcount: 575941 { delta: 144 } hitcount: 351427 { delta: 145 } hitcount: 218077 { delta: 146 } hitcount: 167297 { delta: 147 } hitcount: 146198 { delta: 148 } hitcount: 116122 { delta: 149 } hitcount: 58993 { delta: 150 } hitcount: 40228 The delta above is in nanoseconds. It brings the fastest time down from 129ns to 113ns, and the peak from 141ns to 126ns. Link: https://lkml.kernel.org/r/20220906225529.411545333@goodmis.org Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Masami Hiramatsu <mhiramat@kernel.org> Cc: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-09-06 22:53:15 +00:00
field_contents = hist_fn_call(key_field, elt, buffer, rbe, rec);
if (key_field->flags & HIST_FIELD_FL_STRING) {
key = (void *)(unsigned long)field_contents;
use_compound_key = true;
} else
key = (void *)&field_contents;
}
if (use_compound_key)
add_to_key(compound_key, key, key_field, rec);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
if (use_compound_key)
key = compound_key;
if (hist_data->n_var_refs &&
!resolve_var_refs(hist_data, key, var_ref_vals, false))
return;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
elt = tracing_map_insert(hist_data->map, key);
if (!elt)
return;
hist_trigger_elt_update(hist_data, elt, buffer, rec, rbe, var_ref_vals);
if (resolve_var_refs(hist_data, key, var_ref_vals, true))
hist_trigger_actions(hist_data, elt, buffer, rec, rbe, key, var_ref_vals);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
static void hist_trigger_stacktrace_print(struct seq_file *m,
unsigned long *stacktrace_entries,
unsigned int max_entries)
{
unsigned int spaces = 8;
unsigned int i;
for (i = 0; i < max_entries; i++) {
if (!stacktrace_entries[i])
return;
seq_printf(m, "%*c", 1 + spaces, ' ');
seq_printf(m, "%pS\n", (void*)stacktrace_entries[i]);
}
}
static void hist_trigger_print_key(struct seq_file *m,
struct hist_trigger_data *hist_data,
void *key,
struct tracing_map_elt *elt)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
{
struct hist_field *key_field;
bool multiline = false;
const char *field_name;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
unsigned int i;
u64 uval;
seq_puts(m, "{ ");
for_each_hist_key_field(i, hist_data) {
key_field = hist_data->fields[i];
if (i > hist_data->n_vals)
seq_puts(m, ", ");
field_name = hist_field_name(key_field, 0);
if (key_field->flags & HIST_FIELD_FL_HEX) {
uval = *(u64 *)(key + key_field->offset);
seq_printf(m, "%s: %llx", field_name, uval);
} else if (key_field->flags & HIST_FIELD_FL_SYM) {
uval = *(u64 *)(key + key_field->offset);
seq_printf(m, "%s: [%llx] %-45ps", field_name,
uval, (void *)(uintptr_t)uval);
} else if (key_field->flags & HIST_FIELD_FL_SYM_OFFSET) {
uval = *(u64 *)(key + key_field->offset);
seq_printf(m, "%s: [%llx] %-55pS", field_name,
uval, (void *)(uintptr_t)uval);
} else if (key_field->flags & HIST_FIELD_FL_EXECNAME) {
struct hist_elt_data *elt_data = elt->private_data;
char *comm;
if (WARN_ON_ONCE(!elt_data))
return;
comm = elt_data->comm;
uval = *(u64 *)(key + key_field->offset);
seq_printf(m, "%s: %-16s[%10llu]", field_name,
comm, uval);
} else if (key_field->flags & HIST_FIELD_FL_SYSCALL) {
const char *syscall_name;
uval = *(u64 *)(key + key_field->offset);
syscall_name = get_syscall_name(uval);
if (!syscall_name)
syscall_name = "unknown_syscall";
seq_printf(m, "%s: %-30s[%3llu]", field_name,
syscall_name, uval);
} else if (key_field->flags & HIST_FIELD_FL_STACKTRACE) {
if (key_field->field)
seq_printf(m, "%s.stacktrace", key_field->field->name);
else
seq_puts(m, "common_stacktrace:\n");
hist_trigger_stacktrace_print(m,
key + key_field->offset,
HIST_STACKTRACE_DEPTH);
multiline = true;
tracing: Add hist trigger 'log2' modifier Allow users to have numeric fields displayed as log2 values in case value range is very wide by appending '.log2' to field names. For example, # echo 'hist:key=bytes_req' > kmalloc/trigger # cat kmalloc/hist { bytes_req: 504 } hitcount: 1 { bytes_req: 11 } hitcount: 1 { bytes_req: 104 } hitcount: 1 { bytes_req: 48 } hitcount: 1 { bytes_req: 2048 } hitcount: 1 { bytes_req: 4096 } hitcount: 1 { bytes_req: 240 } hitcount: 1 { bytes_req: 392 } hitcount: 1 { bytes_req: 13 } hitcount: 1 { bytes_req: 28 } hitcount: 1 { bytes_req: 12 } hitcount: 1 { bytes_req: 64 } hitcount: 2 { bytes_req: 128 } hitcount: 2 { bytes_req: 32 } hitcount: 2 { bytes_req: 8 } hitcount: 11 { bytes_req: 10 } hitcount: 13 { bytes_req: 24 } hitcount: 25 { bytes_req: 160 } hitcount: 29 { bytes_req: 16 } hitcount: 33 { bytes_req: 80 } hitcount: 36 When using '.log2' modifier, the output looks like: # echo 'hist:key=bytes_req.log2' > kmalloc/trigger # cat kmalloc/hist { bytes_req: ~ 2^12 } hitcount: 1 { bytes_req: ~ 2^11 } hitcount: 1 { bytes_req: ~ 2^9 } hitcount: 2 { bytes_req: ~ 2^6 } hitcount: 3 { bytes_req: ~ 2^3 } hitcount: 13 { bytes_req: ~ 2^5 } hitcount: 19 { bytes_req: ~ 2^8 } hitcount: 49 { bytes_req: ~ 2^7 } hitcount: 57 { bytes_req: ~ 2^4 } hitcount: 74 Link: http://lkml.kernel.org/r/7ff396b246c6a881f46b979735fddf05a0d6c71a.1457029949.git.tom.zanussi@linux.intel.com Cc: Tom Zanussi <tom.zanussi@linux.intel.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Signed-off-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:55:02 +00:00
} else if (key_field->flags & HIST_FIELD_FL_LOG2) {
seq_printf(m, "%s: ~ 2^%-2llu", field_name,
tracing: Add hist trigger 'log2' modifier Allow users to have numeric fields displayed as log2 values in case value range is very wide by appending '.log2' to field names. For example, # echo 'hist:key=bytes_req' > kmalloc/trigger # cat kmalloc/hist { bytes_req: 504 } hitcount: 1 { bytes_req: 11 } hitcount: 1 { bytes_req: 104 } hitcount: 1 { bytes_req: 48 } hitcount: 1 { bytes_req: 2048 } hitcount: 1 { bytes_req: 4096 } hitcount: 1 { bytes_req: 240 } hitcount: 1 { bytes_req: 392 } hitcount: 1 { bytes_req: 13 } hitcount: 1 { bytes_req: 28 } hitcount: 1 { bytes_req: 12 } hitcount: 1 { bytes_req: 64 } hitcount: 2 { bytes_req: 128 } hitcount: 2 { bytes_req: 32 } hitcount: 2 { bytes_req: 8 } hitcount: 11 { bytes_req: 10 } hitcount: 13 { bytes_req: 24 } hitcount: 25 { bytes_req: 160 } hitcount: 29 { bytes_req: 16 } hitcount: 33 { bytes_req: 80 } hitcount: 36 When using '.log2' modifier, the output looks like: # echo 'hist:key=bytes_req.log2' > kmalloc/trigger # cat kmalloc/hist { bytes_req: ~ 2^12 } hitcount: 1 { bytes_req: ~ 2^11 } hitcount: 1 { bytes_req: ~ 2^9 } hitcount: 2 { bytes_req: ~ 2^6 } hitcount: 3 { bytes_req: ~ 2^3 } hitcount: 13 { bytes_req: ~ 2^5 } hitcount: 19 { bytes_req: ~ 2^8 } hitcount: 49 { bytes_req: ~ 2^7 } hitcount: 57 { bytes_req: ~ 2^4 } hitcount: 74 Link: http://lkml.kernel.org/r/7ff396b246c6a881f46b979735fddf05a0d6c71a.1457029949.git.tom.zanussi@linux.intel.com Cc: Tom Zanussi <tom.zanussi@linux.intel.com> Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Signed-off-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:55:02 +00:00
*(u64 *)(key + key_field->offset));
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
} else if (key_field->flags & HIST_FIELD_FL_BUCKET) {
unsigned long buckets = key_field->buckets;
uval = *(u64 *)(key + key_field->offset);
seq_printf(m, "%s: ~ %llu-%llu", field_name,
uval, uval + buckets -1);
} else if (key_field->flags & HIST_FIELD_FL_STRING) {
seq_printf(m, "%s: %-50s", field_name,
(char *)(key + key_field->offset));
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
} else {
uval = *(u64 *)(key + key_field->offset);
seq_printf(m, "%s: %10llu", field_name, uval);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
}
if (!multiline)
seq_puts(m, " ");
seq_puts(m, "}");
}
/* Get the 100 times of the percentage of @val in @total */
static inline unsigned int __get_percentage(u64 val, u64 total)
{
if (!total)
goto div0;
if (val < (U64_MAX / 10000))
return (unsigned int)div64_ul(val * 10000, total);
total = div64_u64(total, 10000);
if (!total)
goto div0;
return (unsigned int)div64_ul(val, total);
div0:
return val ? UINT_MAX : 0;
}
#define BAR_CHAR '#'
static inline const char *__fill_bar_str(char *buf, int size, u64 val, u64 max)
{
unsigned int len = __get_percentage(val, max);
int i;
if (len == UINT_MAX) {
snprintf(buf, size, "[ERROR]");
return buf;
}
len = len * size / 10000;
for (i = 0; i < len && i < size; i++)
buf[i] = BAR_CHAR;
while (i < size)
buf[i++] = ' ';
buf[size] = '\0';
return buf;
}
struct hist_val_stat {
u64 max;
u64 total;
};
static void hist_trigger_print_val(struct seq_file *m, unsigned int idx,
const char *field_name, unsigned long flags,
struct hist_val_stat *stats,
struct tracing_map_elt *elt)
{
u64 val = tracing_map_read_sum(elt, idx);
unsigned int pc;
char bar[21];
if (flags & HIST_FIELD_FL_PERCENT) {
pc = __get_percentage(val, stats[idx].total);
if (pc == UINT_MAX)
seq_printf(m, " %s (%%):[ERROR]", field_name);
else
seq_printf(m, " %s (%%): %3u.%02u", field_name,
pc / 100, pc % 100);
} else if (flags & HIST_FIELD_FL_GRAPH) {
seq_printf(m, " %s: %20s", field_name,
__fill_bar_str(bar, 20, val, stats[idx].max));
} else if (flags & HIST_FIELD_FL_HEX) {
seq_printf(m, " %s: %10llx", field_name, val);
} else {
seq_printf(m, " %s: %10llu", field_name, val);
}
}
static void hist_trigger_entry_print(struct seq_file *m,
struct hist_trigger_data *hist_data,
struct hist_val_stat *stats,
void *key,
struct tracing_map_elt *elt)
{
const char *field_name;
unsigned int i = HITCOUNT_IDX;
unsigned long flags;
hist_trigger_print_key(m, hist_data, key, elt);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
/* At first, show the raw hitcount if !nohitcount */
if (!hist_data->attrs->no_hitcount)
hist_trigger_print_val(m, i, "hitcount", 0, stats, elt);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
for (i = 1; i < hist_data->n_vals; i++) {
field_name = hist_field_name(hist_data->fields[i], 0);
flags = hist_data->fields[i]->flags;
if (flags & HIST_FIELD_FL_VAR || flags & HIST_FIELD_FL_EXPR)
continue;
seq_puts(m, " ");
hist_trigger_print_val(m, i, field_name, flags, stats, elt);
}
tracing: Add 'onmax' hist trigger action support Add an 'onmax(var).save(field,...)' hist trigger action which is invoked whenever an event exceeds the current maximum. The end result is that the trace event fields or variables specified as the onmax.save() params will be saved if 'var' exceeds the current maximum for that hist trigger entry. This allows context from the event that exhibited the new maximum to be saved for later reference. When the histogram is displayed, additional fields displaying the saved values will be printed. As an example the below defines a couple of hist triggers, one for sched_wakeup and another for sched_switch, keyed on pid. Whenever a sched_wakeup occurs, the timestamp is saved in the entry corresponding to the current pid, and when the scheduler switches back to that pid, the timestamp difference is calculated. If the resulting latency exceeds the current maximum latency, the specified save() values are saved: # echo 'hist:keys=pid:ts0=common_timestamp.usecs \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:\ wakeup_lat=common_timestamp.usecs-$ts0:\ onmax($wakeup_lat).save(next_comm,prev_pid,prev_prio,prev_comm) \ if next_comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_switch/trigger When the histogram is displayed, the max value and the saved values corresponding to the max are displayed following the rest of the fields: # cat /sys/kernel/debug/tracing/events/sched/sched_switch/hist { next_pid: 3728 } hitcount: 199 \ max: 123 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/3 { next_pid: 3730 } hitcount: 1321 \ max: 15 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/1 { next_pid: 3729 } hitcount: 1973\ max: 25 next_comm: cyclictest prev_pid: 0 \ prev_prio: 120 prev_comm: swapper/0 Totals: Hits: 3493 Entries: 3 Dropped: 0 Link: http://lkml.kernel.org/r/006907f71b1e839bb059337ec3c496f84fcb71de.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:01 +00:00
print_actions(m, hist_data, elt);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
seq_puts(m, "\n");
}
static int print_entries(struct seq_file *m,
struct hist_trigger_data *hist_data)
{
struct tracing_map_sort_entry **sort_entries = NULL;
struct tracing_map *map = hist_data->map;
int i, j, n_entries;
struct hist_val_stat *stats = NULL;
u64 val;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
n_entries = tracing_map_sort_entries(map, hist_data->sort_keys,
hist_data->n_sort_keys,
&sort_entries);
if (n_entries < 0)
return n_entries;
/* Calculate the max and the total for each field if needed. */
for (j = 0; j < hist_data->n_vals; j++) {
if (!(hist_data->fields[j]->flags &
(HIST_FIELD_FL_PERCENT | HIST_FIELD_FL_GRAPH)))
continue;
if (!stats) {
stats = kcalloc(hist_data->n_vals, sizeof(*stats),
GFP_KERNEL);
if (!stats) {
n_entries = -ENOMEM;
goto out;
}
}
for (i = 0; i < n_entries; i++) {
val = tracing_map_read_sum(sort_entries[i]->elt, j);
stats[j].total += val;
if (stats[j].max < val)
stats[j].max = val;
}
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
for (i = 0; i < n_entries; i++)
hist_trigger_entry_print(m, hist_data, stats,
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
sort_entries[i]->key,
sort_entries[i]->elt);
kfree(stats);
out:
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
tracing_map_destroy_sort_entries(sort_entries, n_entries);
return n_entries;
}
static void hist_trigger_show(struct seq_file *m,
struct event_trigger_data *data, int n)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
{
struct hist_trigger_data *hist_data;
int n_entries;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (n > 0)
seq_puts(m, "\n\n");
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
seq_puts(m, "# event histogram\n#\n# trigger info: ");
data->ops->print(m, data);
seq_puts(m, "#\n\n");
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
hist_data = data->private_data;
n_entries = print_entries(m, hist_data);
if (n_entries < 0)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
n_entries = 0;
track_data_snapshot_print(m, hist_data);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
seq_printf(m, "\nTotals:\n Hits: %llu\n Entries: %u\n Dropped: %llu\n",
(u64)atomic64_read(&hist_data->map->hits),
n_entries, (u64)atomic64_read(&hist_data->map->drops));
}
static int hist_show(struct seq_file *m, void *v)
{
struct event_trigger_data *data;
struct trace_event_file *event_file;
int n = 0, ret = 0;
mutex_lock(&event_mutex);
event_file = event_file_data(m->private);
if (unlikely(!event_file)) {
ret = -ENODEV;
goto out_unlock;
}
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
list_for_each_entry(data, &event_file->triggers, list) {
if (data->cmd_ops->trigger_type == ETT_EVENT_HIST)
hist_trigger_show(m, data, n++);
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
out_unlock:
mutex_unlock(&event_mutex);
return ret;
}
static int event_hist_open(struct inode *inode, struct file *file)
{
int ret;
ret = security_locked_down(LOCKDOWN_TRACEFS);
if (ret)
return ret;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
return single_open(file, hist_show, file);
}
const struct file_operations event_hist_fops = {
.open = event_hist_open,
.read = seq_read,
.llseek = seq_lseek,
.release = single_release,
};
#ifdef CONFIG_HIST_TRIGGERS_DEBUG
static void hist_field_debug_show_flags(struct seq_file *m,
unsigned long flags)
{
seq_puts(m, " flags:\n");
if (flags & HIST_FIELD_FL_KEY)
seq_puts(m, " HIST_FIELD_FL_KEY\n");
else if (flags & HIST_FIELD_FL_HITCOUNT)
seq_puts(m, " VAL: HIST_FIELD_FL_HITCOUNT\n");
else if (flags & HIST_FIELD_FL_VAR)
seq_puts(m, " HIST_FIELD_FL_VAR\n");
else if (flags & HIST_FIELD_FL_VAR_REF)
seq_puts(m, " HIST_FIELD_FL_VAR_REF\n");
else
seq_puts(m, " VAL: normal u64 value\n");
if (flags & HIST_FIELD_FL_ALIAS)
seq_puts(m, " HIST_FIELD_FL_ALIAS\n");
tracing: Add support for creating hist trigger variables from literal Currently hist trigger expressions don't support the use of numeric literals: e.g. echo 'hist:keys=common_pid:x=$y-1234' --> is not valid expression syntax Having the ability to use numeric constants in hist triggers supports a wider range of expressions for creating variables. Add support for creating trace event histogram variables from numeric literals. e.g. echo 'hist:keys=common_pid:x=1234,y=size-1024' >> event/trigger A negative numeric constant is created, using unary minus operator (parentheses are required). e.g. echo 'hist:keys=common_pid:z=-(2)' >> event/trigger Constants can be used with division/multiplication (added in the next patch in this series) to implement granularity filters for frequent trace events. For instance we can limit emitting the rss_stat trace event to when there is a 512KB cross over in the rss size: # Create a synthetic event to monitor instead of the high frequency # rss_stat event echo 'rss_stat_throttled unsigned int mm_id; unsigned int curr; int member; long size' >> tracing/synthetic_events # Create a hist trigger that emits the synthetic rss_stat_throttled # event only when the rss size crosses a 512KB boundary. echo 'hist:keys=keys=mm_id,member:bucket=size/0x80000:onchange($bucket) .rss_stat_throttled(mm_id,curr,member,size)' >> events/kmem/rss_stat/trigger A use case for using constants with addition/subtraction is not yet known, but for completeness the use of constants are supported for all operators. Link: https://lkml.kernel.org/r/20211025200852.3002369-2-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-10-25 20:08:33 +00:00
else if (flags & HIST_FIELD_FL_CONST)
seq_puts(m, " HIST_FIELD_FL_CONST\n");
}
static int hist_field_debug_show(struct seq_file *m,
struct hist_field *field, unsigned long flags)
{
if ((field->flags & flags) != flags) {
seq_printf(m, "ERROR: bad flags - %lx\n", flags);
return -EINVAL;
}
hist_field_debug_show_flags(m, field->flags);
if (field->field)
seq_printf(m, " ftrace_event_field name: %s\n",
field->field->name);
if (field->flags & HIST_FIELD_FL_VAR) {
seq_printf(m, " var.name: %s\n", field->var.name);
seq_printf(m, " var.idx (into tracing_map_elt.vars[]): %u\n",
field->var.idx);
}
tracing: Add support for creating hist trigger variables from literal Currently hist trigger expressions don't support the use of numeric literals: e.g. echo 'hist:keys=common_pid:x=$y-1234' --> is not valid expression syntax Having the ability to use numeric constants in hist triggers supports a wider range of expressions for creating variables. Add support for creating trace event histogram variables from numeric literals. e.g. echo 'hist:keys=common_pid:x=1234,y=size-1024' >> event/trigger A negative numeric constant is created, using unary minus operator (parentheses are required). e.g. echo 'hist:keys=common_pid:z=-(2)' >> event/trigger Constants can be used with division/multiplication (added in the next patch in this series) to implement granularity filters for frequent trace events. For instance we can limit emitting the rss_stat trace event to when there is a 512KB cross over in the rss size: # Create a synthetic event to monitor instead of the high frequency # rss_stat event echo 'rss_stat_throttled unsigned int mm_id; unsigned int curr; int member; long size' >> tracing/synthetic_events # Create a hist trigger that emits the synthetic rss_stat_throttled # event only when the rss size crosses a 512KB boundary. echo 'hist:keys=keys=mm_id,member:bucket=size/0x80000:onchange($bucket) .rss_stat_throttled(mm_id,curr,member,size)' >> events/kmem/rss_stat/trigger A use case for using constants with addition/subtraction is not yet known, but for completeness the use of constants are supported for all operators. Link: https://lkml.kernel.org/r/20211025200852.3002369-2-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-10-25 20:08:33 +00:00
if (field->flags & HIST_FIELD_FL_CONST)
seq_printf(m, " constant: %llu\n", field->constant);
if (field->flags & HIST_FIELD_FL_ALIAS)
seq_printf(m, " var_ref_idx (into hist_data->var_refs[]): %u\n",
field->var_ref_idx);
if (field->flags & HIST_FIELD_FL_VAR_REF) {
seq_printf(m, " name: %s\n", field->name);
seq_printf(m, " var.idx (into tracing_map_elt.vars[]): %u\n",
field->var.idx);
seq_printf(m, " var.hist_data: %p\n", field->var.hist_data);
seq_printf(m, " var_ref_idx (into hist_data->var_refs[]): %u\n",
field->var_ref_idx);
if (field->system)
seq_printf(m, " system: %s\n", field->system);
if (field->event_name)
seq_printf(m, " event_name: %s\n", field->event_name);
}
seq_printf(m, " type: %s\n", field->type);
seq_printf(m, " size: %u\n", field->size);
seq_printf(m, " is_signed: %u\n", field->is_signed);
return 0;
}
static int field_var_debug_show(struct seq_file *m,
struct field_var *field_var, unsigned int i,
bool save_vars)
{
const char *vars_name = save_vars ? "save_vars" : "field_vars";
struct hist_field *field;
int ret = 0;
seq_printf(m, "\n hist_data->%s[%d]:\n", vars_name, i);
field = field_var->var;
seq_printf(m, "\n %s[%d].var:\n", vars_name, i);
hist_field_debug_show_flags(m, field->flags);
seq_printf(m, " var.name: %s\n", field->var.name);
seq_printf(m, " var.idx (into tracing_map_elt.vars[]): %u\n",
field->var.idx);
field = field_var->val;
seq_printf(m, "\n %s[%d].val:\n", vars_name, i);
if (field->field)
seq_printf(m, " ftrace_event_field name: %s\n",
field->field->name);
else {
ret = -EINVAL;
goto out;
}
seq_printf(m, " type: %s\n", field->type);
seq_printf(m, " size: %u\n", field->size);
seq_printf(m, " is_signed: %u\n", field->is_signed);
out:
return ret;
}
static int hist_action_debug_show(struct seq_file *m,
struct action_data *data, int i)
{
int ret = 0;
if (data->handler == HANDLER_ONMAX ||
data->handler == HANDLER_ONCHANGE) {
seq_printf(m, "\n hist_data->actions[%d].track_data.var_ref:\n", i);
ret = hist_field_debug_show(m, data->track_data.var_ref,
HIST_FIELD_FL_VAR_REF);
if (ret)
goto out;
seq_printf(m, "\n hist_data->actions[%d].track_data.track_var:\n", i);
ret = hist_field_debug_show(m, data->track_data.track_var,
HIST_FIELD_FL_VAR);
if (ret)
goto out;
}
if (data->handler == HANDLER_ONMATCH) {
seq_printf(m, "\n hist_data->actions[%d].match_data.event_system: %s\n",
i, data->match_data.event_system);
seq_printf(m, " hist_data->actions[%d].match_data.event: %s\n",
i, data->match_data.event);
}
out:
return ret;
}
static int hist_actions_debug_show(struct seq_file *m,
struct hist_trigger_data *hist_data)
{
int i, ret = 0;
if (hist_data->n_actions)
seq_puts(m, "\n action tracking variables (for onmax()/onchange()/onmatch()):\n");
for (i = 0; i < hist_data->n_actions; i++) {
struct action_data *action = hist_data->actions[i];
ret = hist_action_debug_show(m, action, i);
if (ret)
goto out;
}
if (hist_data->n_save_vars)
seq_puts(m, "\n save action variables (save() params):\n");
for (i = 0; i < hist_data->n_save_vars; i++) {
ret = field_var_debug_show(m, hist_data->save_vars[i], i, true);
if (ret)
goto out;
}
out:
return ret;
}
static void hist_trigger_debug_show(struct seq_file *m,
struct event_trigger_data *data, int n)
{
struct hist_trigger_data *hist_data;
int i, ret;
if (n > 0)
seq_puts(m, "\n\n");
seq_puts(m, "# event histogram\n#\n# trigger info: ");
data->ops->print(m, data);
seq_puts(m, "#\n\n");
hist_data = data->private_data;
seq_printf(m, "hist_data: %p\n\n", hist_data);
seq_printf(m, " n_vals: %u\n", hist_data->n_vals);
seq_printf(m, " n_keys: %u\n", hist_data->n_keys);
seq_printf(m, " n_fields: %u\n", hist_data->n_fields);
seq_puts(m, "\n val fields:\n\n");
seq_puts(m, " hist_data->fields[0]:\n");
ret = hist_field_debug_show(m, hist_data->fields[0],
HIST_FIELD_FL_HITCOUNT);
if (ret)
return;
for (i = 1; i < hist_data->n_vals; i++) {
seq_printf(m, "\n hist_data->fields[%d]:\n", i);
ret = hist_field_debug_show(m, hist_data->fields[i], 0);
if (ret)
return;
}
seq_puts(m, "\n key fields:\n");
for (i = hist_data->n_vals; i < hist_data->n_fields; i++) {
seq_printf(m, "\n hist_data->fields[%d]:\n", i);
ret = hist_field_debug_show(m, hist_data->fields[i],
HIST_FIELD_FL_KEY);
if (ret)
return;
}
if (hist_data->n_var_refs)
seq_puts(m, "\n variable reference fields:\n");
for (i = 0; i < hist_data->n_var_refs; i++) {
seq_printf(m, "\n hist_data->var_refs[%d]:\n", i);
ret = hist_field_debug_show(m, hist_data->var_refs[i],
HIST_FIELD_FL_VAR_REF);
if (ret)
return;
}
if (hist_data->n_field_vars)
seq_puts(m, "\n field variables:\n");
for (i = 0; i < hist_data->n_field_vars; i++) {
ret = field_var_debug_show(m, hist_data->field_vars[i], i, false);
if (ret)
return;
}
ret = hist_actions_debug_show(m, hist_data);
if (ret)
return;
}
static int hist_debug_show(struct seq_file *m, void *v)
{
struct event_trigger_data *data;
struct trace_event_file *event_file;
int n = 0, ret = 0;
mutex_lock(&event_mutex);
event_file = event_file_data(m->private);
if (unlikely(!event_file)) {
ret = -ENODEV;
goto out_unlock;
}
list_for_each_entry(data, &event_file->triggers, list) {
if (data->cmd_ops->trigger_type == ETT_EVENT_HIST)
hist_trigger_debug_show(m, data, n++);
}
out_unlock:
mutex_unlock(&event_mutex);
return ret;
}
static int event_hist_debug_open(struct inode *inode, struct file *file)
{
int ret;
ret = security_locked_down(LOCKDOWN_TRACEFS);
if (ret)
return ret;
return single_open(file, hist_debug_show, file);
}
const struct file_operations event_hist_debug_fops = {
.open = event_hist_debug_open,
.read = seq_read,
.llseek = seq_lseek,
.release = single_release,
};
#endif
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
static void hist_field_print(struct seq_file *m, struct hist_field *hist_field)
{
const char *field_name = hist_field_name(hist_field, 0);
if (hist_field->var.name)
seq_printf(m, "%s=", hist_field->var.name);
if (hist_field->flags & HIST_FIELD_FL_CPU)
tracing/histogram: Rename "cpu" to "common_cpu" Currently the histogram logic allows the user to write "cpu" in as an event field, and it will record the CPU that the event happened on. The problem with this is that there's a lot of events that have "cpu" as a real field, and using "cpu" as the CPU it ran on, makes it impossible to run histograms on the "cpu" field of events. For example, if I want to have a histogram on the count of the workqueue_queue_work event on its cpu field, running: ># echo 'hist:keys=cpu' > events/workqueue/workqueue_queue_work/trigger Gives a misleading and wrong result. Change the command to "common_cpu" as no event should have "common_*" fields as that's a reserved name for fields used by all events. And this makes sense here as common_cpu would be a field used by all events. Now we can even do: ># echo 'hist:keys=common_cpu,cpu if cpu < 100' > events/workqueue/workqueue_queue_work/trigger ># cat events/workqueue/workqueue_queue_work/hist # event histogram # # trigger info: hist:keys=common_cpu,cpu:vals=hitcount:sort=hitcount:size=2048 if cpu < 100 [active] # { common_cpu: 0, cpu: 2 } hitcount: 1 { common_cpu: 0, cpu: 4 } hitcount: 1 { common_cpu: 7, cpu: 7 } hitcount: 1 { common_cpu: 0, cpu: 7 } hitcount: 1 { common_cpu: 0, cpu: 1 } hitcount: 1 { common_cpu: 0, cpu: 6 } hitcount: 2 { common_cpu: 0, cpu: 5 } hitcount: 2 { common_cpu: 1, cpu: 1 } hitcount: 4 { common_cpu: 6, cpu: 6 } hitcount: 4 { common_cpu: 5, cpu: 5 } hitcount: 14 { common_cpu: 4, cpu: 4 } hitcount: 26 { common_cpu: 0, cpu: 0 } hitcount: 39 { common_cpu: 2, cpu: 2 } hitcount: 184 Now for backward compatibility, I added a trick. If "cpu" is used, and the field is not found, it will fall back to "common_cpu" and work as it did before. This way, it will still work for old programs that use "cpu" to get the actual CPU, but if the event has a "cpu" as a field, it will get that event's "cpu" field, which is probably what it wants anyway. I updated the tracefs/README to include documentation about both the common_timestamp and the common_cpu. This way, if that text is present in the README, then an application can know that common_cpu is supported over just plain "cpu". Link: https://lkml.kernel.org/r/20210721110053.26b4f641@oasis.local.home Cc: Namhyung Kim <namhyung@kernel.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: stable@vger.kernel.org Fixes: 8b7622bf94a44 ("tracing: Add cpu field for hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-21 15:00:53 +00:00
seq_puts(m, "common_cpu");
tracing: Add support for creating hist trigger variables from literal Currently hist trigger expressions don't support the use of numeric literals: e.g. echo 'hist:keys=common_pid:x=$y-1234' --> is not valid expression syntax Having the ability to use numeric constants in hist triggers supports a wider range of expressions for creating variables. Add support for creating trace event histogram variables from numeric literals. e.g. echo 'hist:keys=common_pid:x=1234,y=size-1024' >> event/trigger A negative numeric constant is created, using unary minus operator (parentheses are required). e.g. echo 'hist:keys=common_pid:z=-(2)' >> event/trigger Constants can be used with division/multiplication (added in the next patch in this series) to implement granularity filters for frequent trace events. For instance we can limit emitting the rss_stat trace event to when there is a 512KB cross over in the rss size: # Create a synthetic event to monitor instead of the high frequency # rss_stat event echo 'rss_stat_throttled unsigned int mm_id; unsigned int curr; int member; long size' >> tracing/synthetic_events # Create a hist trigger that emits the synthetic rss_stat_throttled # event only when the rss size crosses a 512KB boundary. echo 'hist:keys=keys=mm_id,member:bucket=size/0x80000:onchange($bucket) .rss_stat_throttled(mm_id,curr,member,size)' >> events/kmem/rss_stat/trigger A use case for using constants with addition/subtraction is not yet known, but for completeness the use of constants are supported for all operators. Link: https://lkml.kernel.org/r/20211025200852.3002369-2-kaleshsingh@google.com Signed-off-by: Kalesh Singh <kaleshsingh@google.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-10-25 20:08:33 +00:00
else if (hist_field->flags & HIST_FIELD_FL_CONST)
seq_printf(m, "%llu", hist_field->constant);
else if (field_name) {
if (hist_field->flags & HIST_FIELD_FL_VAR_REF ||
hist_field->flags & HIST_FIELD_FL_ALIAS)
seq_putc(m, '$');
seq_printf(m, "%s", field_name);
} else if (hist_field->flags & HIST_FIELD_FL_TIMESTAMP)
seq_puts(m, "common_timestamp");
tracing: Restore proper field flag printing when displaying triggers The flag-printing code used when displaying hist triggers somehow got dropped during refactoring of the inter-event patchset. This restores it. Below are a couple examples - in the first case, .usecs wasn't being displayed properly for common_timestamps and the second illustrates the same for other flags such as .execname. Before: # echo 'hist:key=common_pid.execname:val=count:sort=count' > /sys/kernel/debug/tracing/events/syscalls/sys_enter_read/trigger # cat /sys/kernel/debug/tracing/events/syscalls/sys_enter_read/trigger hist:keys=common_pid:vals=hitcount,count:sort=count:size=2048 [active] # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # cat /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger hist:keys=pid:vals=hitcount:ts0=common_timestamp:sort=hitcount:size=2048:clock=global if comm=="cyclictest" [active] After: # echo 'hist:key=common_pid.execname:val=count:sort=count' > /sys/kernel/debug/tracing/events/syscalls/sys_enter_read/trigger # cat /sys/kernel/debug/tracing/events/syscalls/sys_enter_read/trigger hist:keys=common_pid.execname:vals=hitcount,count:sort=count:size=2048 [active] # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # cat /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger hist:keys=pid:vals=hitcount:ts0=common_timestamp.usecs:sort=hitcount:size=2048:clock=global if comm=="cyclictest" [active] Link: http://lkml.kernel.org/r/492bab42ff21806600af98a8ea901af10efbee0c.1524790601.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-04-27 01:04:47 +00:00
if (hist_field->flags) {
if (!(hist_field->flags & HIST_FIELD_FL_VAR_REF) &&
!(hist_field->flags & HIST_FIELD_FL_EXPR) &&
!(hist_field->flags & HIST_FIELD_FL_STACKTRACE)) {
tracing: Restore proper field flag printing when displaying triggers The flag-printing code used when displaying hist triggers somehow got dropped during refactoring of the inter-event patchset. This restores it. Below are a couple examples - in the first case, .usecs wasn't being displayed properly for common_timestamps and the second illustrates the same for other flags such as .execname. Before: # echo 'hist:key=common_pid.execname:val=count:sort=count' > /sys/kernel/debug/tracing/events/syscalls/sys_enter_read/trigger # cat /sys/kernel/debug/tracing/events/syscalls/sys_enter_read/trigger hist:keys=common_pid:vals=hitcount,count:sort=count:size=2048 [active] # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # cat /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger hist:keys=pid:vals=hitcount:ts0=common_timestamp:sort=hitcount:size=2048:clock=global if comm=="cyclictest" [active] After: # echo 'hist:key=common_pid.execname:val=count:sort=count' > /sys/kernel/debug/tracing/events/syscalls/sys_enter_read/trigger # cat /sys/kernel/debug/tracing/events/syscalls/sys_enter_read/trigger hist:keys=common_pid.execname:vals=hitcount,count:sort=count:size=2048 [active] # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # cat /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger hist:keys=pid:vals=hitcount:ts0=common_timestamp.usecs:sort=hitcount:size=2048:clock=global if comm=="cyclictest" [active] Link: http://lkml.kernel.org/r/492bab42ff21806600af98a8ea901af10efbee0c.1524790601.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-04-27 01:04:47 +00:00
const char *flags = get_hist_field_flags(hist_field);
if (flags)
seq_printf(m, ".%s", flags);
}
}
tracing: Add linear buckets to histogram logic There's been several times I wished the histogram logic had a "grouping" feature for the buckets. Currently, each bucket has a size of one. That is, if you trace the amount of requested allocations, each allocation is its own bucket, even if you are interested in what allocates 100 bytes or less, 100 to 200, 200 to 300, etc. Also, without grouping, it fills up the allocated histogram buckets quickly. If you are tracking latency, and don't care if something is 200 microseconds off, or 201 microseconds off, but want to track them by say 10 microseconds each. This can not currently be done. There is a log2 but that grouping get's too big too fast for a lot of cases. Introduce a "buckets=SIZE" command to each field where it will record in a rounded number. For example: ># echo 'hist:keys=bytes_req.buckets=100:sort=bytes_req' > events/kmem/kmalloc/trigger ># cat events/kmem/kmalloc/hist # event histogram # # trigger info: hist:keys=bytes_req.buckets=100:vals=hitcount:sort=bytes_req.buckets=100:size=2048 [active] # { bytes_req: ~ 0-99 } hitcount: 3149 { bytes_req: ~ 100-199 } hitcount: 1468 { bytes_req: ~ 200-299 } hitcount: 39 { bytes_req: ~ 300-399 } hitcount: 306 { bytes_req: ~ 400-499 } hitcount: 364 { bytes_req: ~ 500-599 } hitcount: 32 { bytes_req: ~ 600-699 } hitcount: 69 { bytes_req: ~ 700-799 } hitcount: 37 { bytes_req: ~ 1200-1299 } hitcount: 16 { bytes_req: ~ 1400-1499 } hitcount: 30 { bytes_req: ~ 2000-2099 } hitcount: 6 { bytes_req: ~ 4000-4099 } hitcount: 2168 { bytes_req: ~ 5000-5099 } hitcount: 6 Totals: Hits: 7690 Entries: 13 Dropped: 0 Link: https://lkml.kernel.org/r/20210707213921.980359719@goodmis.org Acked-by: Namhyung Kim <namhyung@kernel.org> Reviewed-by: Tom Zanussi <zanussi@kernel.org> Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org> Tested-by: Daniel Bristot de Oliveira <bristot@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2021-07-07 21:36:24 +00:00
if (hist_field->buckets)
seq_printf(m, "=%ld", hist_field->buckets);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
static int event_hist_trigger_print(struct seq_file *m,
struct event_trigger_data *data)
{
struct hist_trigger_data *hist_data = data->private_data;
struct hist_field *field;
bool have_var = false;
bool show_val = false;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
unsigned int i;
seq_puts(m, HIST_PREFIX);
if (data->name)
seq_printf(m, "%s:", data->name);
seq_puts(m, "keys=");
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
for_each_hist_key_field(i, hist_data) {
field = hist_data->fields[i];
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (i > hist_data->n_vals)
seq_puts(m, ",");
if (field->flags & HIST_FIELD_FL_STACKTRACE) {
if (field->field)
seq_printf(m, "%s.stacktrace", field->field->name);
else
seq_puts(m, "common_stacktrace");
} else
hist_field_print(m, field);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
seq_puts(m, ":vals=");
for_each_hist_val_field(i, hist_data) {
field = hist_data->fields[i];
if (field->flags & HIST_FIELD_FL_VAR) {
have_var = true;
continue;
}
if (i == HITCOUNT_IDX) {
if (hist_data->attrs->no_hitcount)
continue;
seq_puts(m, "hitcount");
} else {
if (show_val)
seq_puts(m, ",");
hist_field_print(m, field);
}
show_val = true;
}
if (have_var) {
unsigned int n = 0;
seq_puts(m, ":");
for_each_hist_val_field(i, hist_data) {
field = hist_data->fields[i];
if (field->flags & HIST_FIELD_FL_VAR) {
if (n++)
seq_puts(m, ",");
hist_field_print(m, field);
}
}
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
seq_puts(m, ":sort=");
for (i = 0; i < hist_data->n_sort_keys; i++) {
struct tracing_map_sort_key *sort_key;
unsigned int idx, first_key_idx;
/* skip VAR vals */
first_key_idx = hist_data->n_vals - hist_data->n_vars;
sort_key = &hist_data->sort_keys[i];
idx = sort_key->field_idx;
if (WARN_ON(idx >= HIST_FIELDS_MAX))
return -EINVAL;
if (i > 0)
seq_puts(m, ",");
if (idx == HITCOUNT_IDX)
seq_puts(m, "hitcount");
else {
if (idx >= first_key_idx)
idx += hist_data->n_vars;
hist_field_print(m, hist_data->fields[idx]);
}
if (sort_key->descending)
seq_puts(m, ".descending");
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
seq_printf(m, ":size=%u", (1 << hist_data->map->map_bits));
if (hist_data->enable_timestamps)
seq_printf(m, ":clock=%s", hist_data->attrs->clock);
if (hist_data->attrs->no_hitcount)
seq_puts(m, ":nohitcount");
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
tracing: Add 'onmatch' hist trigger action support Add an 'onmatch(matching.event).<synthetic_event_name>(param list)' hist trigger action which is invoked with the set of variables or event fields named in the 'param list'. The result is the generation of a synthetic event that consists of the values contained in those variables and/or fields at the time the invoking event was hit. As an example the below defines a simple synthetic event using a variable defined on the sched_wakeup_new event, and shows the event definition with unresolved fields, since the sched_wakeup_new event with the testpid variable hasn't been defined yet: # echo 'wakeup_new_test pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events # cat /sys/kernel/debug/tracing/synthetic_events wakeup_new_test pid_t pid; int prio The following hist trigger both defines a testpid variable and specifies an onmatch() trace action that uses that variable along with a non-variable field to generate a wakeup_new_test synthetic event whenever a sched_wakeup_new event occurs, which because of the 'if comm == "cyclictest"' filter only happens when the executable is cyclictest: # echo 'hist:testpid=pid:keys=$testpid:\ onmatch(sched.sched_wakeup_new).wakeup_new_test($testpid, prio) \ if comm=="cyclictest"' >> \ /sys/kernel/debug/tracing/events/sched/sched_wakeup_new/trigger Creating and displaying a histogram based on those events is now just a matter of using the fields and new synthetic event in the tracing/events/synthetic directory, as usual: # echo 'hist:keys=pid,prio:sort=pid,prio' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_new_test/trigger Link: http://lkml.kernel.org/r/8c2a574bcb7530c876629c901ecd23911b14afe8.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Rajvi Jingar <rajvi.jingar@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:52:00 +00:00
print_actions_spec(m, hist_data);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (data->filter_str)
seq_printf(m, " if %s", data->filter_str);
if (data->paused)
seq_puts(m, " [paused]");
else
seq_puts(m, " [active]");
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
seq_putc(m, '\n');
return 0;
}
static int event_hist_trigger_init(struct event_trigger_data *data)
{
struct hist_trigger_data *hist_data = data->private_data;
if (!data->ref && hist_data->attrs->name)
save_named_trigger(hist_data->attrs->name, data);
data->ref++;
return 0;
}
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
static void unregister_field_var_hists(struct hist_trigger_data *hist_data)
{
struct trace_event_file *file;
unsigned int i;
char *cmd;
int ret;
for (i = 0; i < hist_data->n_field_var_hists; i++) {
file = hist_data->field_var_hists[i]->hist_data->event_file;
cmd = hist_data->field_var_hists[i]->cmd;
ret = event_hist_trigger_parse(&trigger_hist_cmd, file,
"!hist", "hist", cmd);
WARN_ON_ONCE(ret < 0);
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
}
}
static void event_hist_trigger_free(struct event_trigger_data *data)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
{
struct hist_trigger_data *hist_data = data->private_data;
if (WARN_ON_ONCE(data->ref <= 0))
return;
data->ref--;
if (!data->ref) {
if (data->name)
del_named_trigger(data);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
trigger_data_free(data);
remove_hist_vars(hist_data);
tracing: Add support for 'field variables' Users should be able to directly specify event fields in hist trigger 'actions' rather than being forced to explicitly create a variable for that purpose. Add support allowing fields to be used directly in actions, which essentially does just that - creates 'invisible' variables for each bare field specified in an action. If a bare field refers to a field on another (matching) event, it even creates a special histogram for the purpose (since variables can't be defined on an existing histogram after histogram creation). Here's a simple example that demonstrates both. Basically the onmatch() action creates a list of variables corresponding to the parameters of the synthetic event to be generated, and then uses those values to generate the event. So for the wakeup_latency synthetic event 'call' below the first param, $wakeup_lat, is a variable defined explicitly on sched_switch, where 'next_pid' is just a normal field on sched_switch, and prio is a normal field on sched_waking. Since the mechanism works on variables, those two normal fields just have 'invisible' variables created internally for them. In the case of 'prio', which is on another event, we actually need to create an additional hist trigger and define the invisible variable on that, since once a hist trigger is defined, variables can't be added to it later. echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> /sys/kernel/debug/tracing/synthetic_events echo 'hist:keys=pid:ts0=common_timestamp.usecs >> /sys/kernel/debug/tracing/events/sched/sched_waking/trigger echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0: onmatch(sched.sched_waking).wakeup_latency($wakeup_lat,next_pid,prio) >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/8e8dcdac1ea180ed7a3689e1caeeccede9dc42b3.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:59 +00:00
unregister_field_var_hists(hist_data);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
destroy_hist_data(hist_data);
}
}
static struct event_trigger_ops event_hist_trigger_ops = {
.trigger = event_hist_trigger,
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
.print = event_hist_trigger_print,
.init = event_hist_trigger_init,
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
.free = event_hist_trigger_free,
};
static int event_hist_trigger_named_init(struct event_trigger_data *data)
{
data->ref++;
save_named_trigger(data->named_data->name, data);
event_hist_trigger_init(data->named_data);
return 0;
}
static void event_hist_trigger_named_free(struct event_trigger_data *data)
{
if (WARN_ON_ONCE(data->ref <= 0))
return;
event_hist_trigger_free(data->named_data);
data->ref--;
if (!data->ref) {
del_named_trigger(data);
trigger_data_free(data);
}
}
static struct event_trigger_ops event_hist_trigger_named_ops = {
.trigger = event_hist_trigger,
.print = event_hist_trigger_print,
.init = event_hist_trigger_named_init,
.free = event_hist_trigger_named_free,
};
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
static struct event_trigger_ops *event_hist_get_trigger_ops(char *cmd,
char *param)
{
return &event_hist_trigger_ops;
}
static void hist_clear(struct event_trigger_data *data)
{
struct hist_trigger_data *hist_data = data->private_data;
if (data->name)
pause_named_trigger(data);
tracepoint_synchronize_unregister();
tracing_map_clear(hist_data->map);
if (data->name)
unpause_named_trigger(data);
}
static bool compatible_field(struct ftrace_event_field *field,
struct ftrace_event_field *test_field)
{
if (field == test_field)
return true;
if (field == NULL || test_field == NULL)
return false;
if (strcmp(field->name, test_field->name) != 0)
return false;
if (strcmp(field->type, test_field->type) != 0)
return false;
if (field->size != test_field->size)
return false;
if (field->is_signed != test_field->is_signed)
return false;
return true;
}
static bool hist_trigger_match(struct event_trigger_data *data,
struct event_trigger_data *data_test,
struct event_trigger_data *named_data,
bool ignore_filter)
{
struct tracing_map_sort_key *sort_key, *sort_key_test;
struct hist_trigger_data *hist_data, *hist_data_test;
struct hist_field *key_field, *key_field_test;
unsigned int i;
if (named_data && (named_data != data_test) &&
(named_data != data_test->named_data))
return false;
if (!named_data && is_named_trigger(data_test))
return false;
hist_data = data->private_data;
hist_data_test = data_test->private_data;
if (hist_data->n_vals != hist_data_test->n_vals ||
hist_data->n_fields != hist_data_test->n_fields ||
hist_data->n_sort_keys != hist_data_test->n_sort_keys)
return false;
if (!ignore_filter) {
if ((data->filter_str && !data_test->filter_str) ||
(!data->filter_str && data_test->filter_str))
return false;
}
for_each_hist_field(i, hist_data) {
key_field = hist_data->fields[i];
key_field_test = hist_data_test->fields[i];
if (key_field->flags != key_field_test->flags)
return false;
if (!compatible_field(key_field->field, key_field_test->field))
return false;
if (key_field->offset != key_field_test->offset)
return false;
if (key_field->size != key_field_test->size)
return false;
if (key_field->is_signed != key_field_test->is_signed)
return false;
if (!!key_field->var.name != !!key_field_test->var.name)
return false;
if (key_field->var.name &&
strcmp(key_field->var.name, key_field_test->var.name) != 0)
return false;
}
for (i = 0; i < hist_data->n_sort_keys; i++) {
sort_key = &hist_data->sort_keys[i];
sort_key_test = &hist_data_test->sort_keys[i];
if (sort_key->field_idx != sort_key_test->field_idx ||
sort_key->descending != sort_key_test->descending)
return false;
}
if (!ignore_filter && data->filter_str &&
(strcmp(data->filter_str, data_test->filter_str) != 0))
return false;
tracing: Add action comparisons when testing matching hist triggers Actions also need to be considered when checking for matching triggers - triggers differing only by action should be allowed, but currently aren't because the matching check ignores the action and erroneously returns -EEXIST. Add and call an actions_match() function to address that. Here's an example using onmatch() actions. The first -EEXIST shouldn't occur because the onmatch() is different in the second wakeup_latency() param. The second -EEXIST shouldn't occur because it's a different action (in this case, it doesn't have an action, so shouldn't be seen as being the same and therefore rejected). In the after case, both are correctly accepted (and trying to add one of them again returns -EEXIST as it should). before: # echo 'wakeup_latency u64 lat; pid_t pid' >> /sys/kernel/debug/tracing/synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0 if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,next_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,prev_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger -su: echo: write error: File exists # echo 'hist:keys=next_pid if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger -su: echo: write error: File exists after: # echo 'wakeup_latency u64 lat; pid_t pid' >> /sys/kernel/debug/tracing/synthetic_events # echo 'hist:keys=pid:ts0=common_timestamp.usecs if comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0 if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,next_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid:onmatch(sched.sched_wakeup).wakeup_latency(sched.sched_switch.$wakeup_lat,prev_pid) if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger # echo 'hist:keys=next_pid if next_comm=="cyclictest"' >> /sys/kernel/debug/tracing/events/sched/sched_switch/trigger Link: http://lkml.kernel.org/r/a7fd668b87ec10736c8f016ac4279c8480d50c2b.1522256721.git.tom.zanussi@linux.intel.com Tested-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-03-28 20:10:55 +00:00
if (!actions_match(hist_data, hist_data_test))
return false;
return true;
}
static bool existing_hist_update_only(char *glob,
struct event_trigger_data *data,
struct trace_event_file *file)
{
struct hist_trigger_data *hist_data = data->private_data;
struct event_trigger_data *test, *named_data = NULL;
bool updated = false;
if (!hist_data->attrs->pause && !hist_data->attrs->cont &&
!hist_data->attrs->clear)
goto out;
if (hist_data->attrs->name) {
named_data = find_named_trigger(hist_data->attrs->name);
if (named_data) {
if (!hist_trigger_match(data, named_data, named_data,
true))
goto out;
}
}
if (hist_data->attrs->name && !named_data)
goto out;
list_for_each_entry(test, &file->triggers, list) {
if (test->cmd_ops->trigger_type == ETT_EVENT_HIST) {
if (!hist_trigger_match(data, test, named_data, false))
continue;
if (hist_data->attrs->pause)
test->paused = true;
else if (hist_data->attrs->cont)
test->paused = false;
else if (hist_data->attrs->clear)
hist_clear(test);
updated = true;
goto out;
}
}
out:
return updated;
}
static int hist_register_trigger(char *glob,
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
struct event_trigger_data *data,
struct trace_event_file *file)
{
struct hist_trigger_data *hist_data = data->private_data;
struct event_trigger_data *test, *named_data = NULL;
struct trace_array *tr = file->tr;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
int ret = 0;
if (hist_data->attrs->name) {
named_data = find_named_trigger(hist_data->attrs->name);
if (named_data) {
if (!hist_trigger_match(data, named_data, named_data,
true)) {
hist_err(tr, HIST_ERR_NAMED_MISMATCH, errpos(hist_data->attrs->name));
ret = -EINVAL;
goto out;
}
}
}
if (hist_data->attrs->name && !named_data)
goto new;
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
lockdep_assert_held(&event_mutex);
list_for_each_entry(test, &file->triggers, list) {
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (test->cmd_ops->trigger_type == ETT_EVENT_HIST) {
if (hist_trigger_match(data, test, named_data, false)) {
hist_err(tr, HIST_ERR_TRIGGER_EEXIST, 0);
ret = -EEXIST;
goto out;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
}
}
new:
if (hist_data->attrs->cont || hist_data->attrs->clear) {
hist_err(tr, HIST_ERR_TRIGGER_ENOENT_CLEAR, 0);
ret = -ENOENT;
goto out;
}
if (hist_data->attrs->pause)
data->paused = true;
if (named_data) {
data->private_data = named_data->private_data;
set_named_trigger_data(data, named_data);
data->ops = &event_hist_trigger_named_ops;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (data->ops->init) {
ret = data->ops->init(data);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (ret < 0)
goto out;
}
if (hist_data->enable_timestamps) {
char *clock = hist_data->attrs->clock;
ret = tracing_set_clock(file->tr, hist_data->attrs->clock);
if (ret) {
hist_err(tr, HIST_ERR_SET_CLOCK_FAIL, errpos(clock));
goto out;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
tracing_set_filter_buffering(file->tr, true);
}
if (named_data)
destroy_hist_data(hist_data);
out:
return ret;
}
static int hist_trigger_enable(struct event_trigger_data *data,
struct trace_event_file *file)
{
int ret = 0;
list_add_tail_rcu(&data->list, &file->triggers);
update_cond_flag(file);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (trace_event_trigger_enable_disable(file, 1) < 0) {
list_del_rcu(&data->list);
update_cond_flag(file);
ret--;
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
return ret;
}
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
static bool have_hist_trigger_match(struct event_trigger_data *data,
struct trace_event_file *file)
{
struct hist_trigger_data *hist_data = data->private_data;
struct event_trigger_data *test, *named_data = NULL;
bool match = false;
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
lockdep_assert_held(&event_mutex);
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
if (hist_data->attrs->name)
named_data = find_named_trigger(hist_data->attrs->name);
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
list_for_each_entry(test, &file->triggers, list) {
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
if (test->cmd_ops->trigger_type == ETT_EVENT_HIST) {
if (hist_trigger_match(data, test, named_data, false)) {
match = true;
break;
}
}
}
return match;
}
static bool hist_trigger_check_refs(struct event_trigger_data *data,
struct trace_event_file *file)
{
struct hist_trigger_data *hist_data = data->private_data;
struct event_trigger_data *test, *named_data = NULL;
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
lockdep_assert_held(&event_mutex);
if (hist_data->attrs->name)
named_data = find_named_trigger(hist_data->attrs->name);
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
list_for_each_entry(test, &file->triggers, list) {
if (test->cmd_ops->trigger_type == ETT_EVENT_HIST) {
if (!hist_trigger_match(data, test, named_data, false))
continue;
hist_data = test->private_data;
if (check_var_refs(hist_data))
return true;
break;
}
}
return false;
}
static void hist_unregister_trigger(char *glob,
struct event_trigger_data *data,
struct trace_event_file *file)
{
struct event_trigger_data *test = NULL, *iter, *named_data = NULL;
struct hist_trigger_data *hist_data = data->private_data;
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
lockdep_assert_held(&event_mutex);
if (hist_data->attrs->name)
named_data = find_named_trigger(hist_data->attrs->name);
list_for_each_entry(iter, &file->triggers, list) {
if (iter->cmd_ops->trigger_type == ETT_EVENT_HIST) {
if (!hist_trigger_match(data, iter, named_data, false))
continue;
test = iter;
list_del_rcu(&test->list);
trace_event_trigger_enable_disable(file, 0);
update_cond_flag(file);
break;
}
}
if (test && test->ops->free)
test->ops->free(test);
if (hist_data->enable_timestamps) {
if (!hist_data->remove || test)
tracing_set_filter_buffering(file->tr, false);
}
}
static bool hist_file_check_refs(struct trace_event_file *file)
{
struct hist_trigger_data *hist_data;
struct event_trigger_data *test;
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
lockdep_assert_held(&event_mutex);
list_for_each_entry(test, &file->triggers, list) {
if (test->cmd_ops->trigger_type == ETT_EVENT_HIST) {
hist_data = test->private_data;
if (check_var_refs(hist_data))
return true;
}
}
return false;
}
static void hist_unreg_all(struct trace_event_file *file)
{
struct event_trigger_data *test, *n;
struct hist_trigger_data *hist_data;
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
struct synth_event *se;
const char *se_name;
lockdep_assert_held(&event_mutex);
if (hist_file_check_refs(file))
return;
list_for_each_entry_safe(test, n, &file->triggers, list) {
if (test->cmd_ops->trigger_type == ETT_EVENT_HIST) {
hist_data = test->private_data;
list_del_rcu(&test->list);
trace_event_trigger_enable_disable(file, 0);
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
se_name = trace_event_name(file->event_call);
se = find_synth_event(se_name);
if (se)
se->ref--;
update_cond_flag(file);
if (hist_data->enable_timestamps)
tracing_set_filter_buffering(file->tr, false);
if (test->ops->free)
test->ops->free(test);
}
}
}
static int event_hist_trigger_parse(struct event_command *cmd_ops,
struct trace_event_file *file,
char *glob, char *cmd,
char *param_and_filter)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
{
unsigned int hist_trigger_bits = TRACING_MAP_BITS_DEFAULT;
struct event_trigger_data *trigger_data;
struct hist_trigger_attrs *attrs;
struct hist_trigger_data *hist_data;
char *param, *filter, *p, *start;
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
struct synth_event *se;
const char *se_name;
bool remove;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
int ret = 0;
lockdep_assert_held(&event_mutex);
if (WARN_ON(!glob))
return -EINVAL;
if (glob[0]) {
hist_err_clear();
last_cmd_set(file, param_and_filter);
}
remove = event_trigger_check_remove(glob);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (event_trigger_empty_param(param_and_filter))
return -EINVAL;
/*
* separate the trigger from the filter (k:v [if filter])
* allowing for whitespace in the trigger
*/
p = param = param_and_filter;
do {
p = strstr(p, "if");
if (!p)
break;
if (p == param_and_filter)
return -EINVAL;
if (*(p - 1) != ' ' && *(p - 1) != '\t') {
p++;
continue;
}
if (p >= param_and_filter + strlen(param_and_filter) - (sizeof("if") - 1) - 1)
return -EINVAL;
if (*(p + sizeof("if") - 1) != ' ' && *(p + sizeof("if") - 1) != '\t') {
p++;
continue;
}
break;
} while (1);
if (!p)
filter = NULL;
else {
*(p - 1) = '\0';
filter = strstrip(p);
param = strstrip(param);
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
/*
* To simplify arithmetic expression parsing, replace occurrences of
* '.sym-offset' modifier with '.symXoffset'
*/
start = strstr(param, ".sym-offset");
while (start) {
*(start + 4) = 'X';
start = strstr(start + 11, ".sym-offset");
}
attrs = parse_hist_trigger_attrs(file->tr, param);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (IS_ERR(attrs))
return PTR_ERR(attrs);
if (attrs->map_bits)
hist_trigger_bits = attrs->map_bits;
hist_data = create_hist_data(hist_trigger_bits, attrs, file, remove);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
if (IS_ERR(hist_data)) {
destroy_hist_trigger_attrs(attrs);
return PTR_ERR(hist_data);
}
trigger_data = event_trigger_alloc(cmd_ops, cmd, param, hist_data);
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
if (!trigger_data) {
ret = -ENOMEM;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
goto out_free;
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
}
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
ret = event_trigger_set_filter(cmd_ops, file, filter, trigger_data);
if (ret < 0)
goto out_free;
if (remove) {
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
if (!have_hist_trigger_match(trigger_data, file))
goto out_free;
if (hist_trigger_check_refs(trigger_data, file)) {
ret = -EBUSY;
goto out_free;
}
event_trigger_unregister(cmd_ops, file, glob+1, trigger_data);
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
se_name = trace_event_name(file->event_call);
se = find_synth_event(se_name);
if (se)
se->ref--;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
ret = 0;
goto out_free;
}
if (existing_hist_update_only(glob, trigger_data, file))
goto out_free;
ret = event_trigger_register(cmd_ops, file, glob, trigger_data);
if (ret < 0)
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
goto out_free;
if (get_named_trigger_data(trigger_data))
goto enable;
if (has_hist_vars(hist_data))
save_hist_vars(hist_data);
ret = create_actions(hist_data);
if (ret)
goto out_unreg;
ret = tracing_map_init(hist_data->map);
if (ret)
goto out_unreg;
enable:
ret = hist_trigger_enable(trigger_data, file);
if (ret)
goto out_unreg;
tracing: Add support for 'synthetic' events Synthetic events are user-defined events generated from hist trigger variables saved from one or more other events. To define a synthetic event, the user writes a simple specification consisting of the name of the new event along with one or more variables and their type(s), to the tracing/synthetic_events file. For instance, the following creates a new event named 'wakeup_latency' with 3 fields: lat, pid, and prio: # echo 'wakeup_latency u64 lat; pid_t pid; int prio' >> \ /sys/kernel/debug/tracing/synthetic_events Reading the tracing/synthetic_events file lists all the currently-defined synthetic events, in this case the event we defined above: # cat /sys/kernel/debug/tracing/synthetic_events wakeup_latency u64 lat; pid_t pid; int prio At this point, the synthetic event is ready to use, and a histogram can be defined using it: # echo 'hist:keys=pid,prio,lat.log2:sort=pid,lat' >> \ /sys/kernel/debug/tracing/events/synthetic/wakeup_latency/trigger The new event is created under the tracing/events/synthetic/ directory and looks and behaves just like any other event: # ls /sys/kernel/debug/tracing/events/synthetic/wakeup_latency enable filter format hist id trigger Although a histogram can be defined for it, nothing will happen until an action tracing that event via the trace_synth() function occurs. The trace_synth() function is very similar to all the other trace_* invocations spread throughout the kernel, except in this case the trace_ function and its corresponding tracepoint isn't statically generated but defined by the user at run-time. How this can be automatically hooked up via a hist trigger 'action' is discussed in a subsequent patch. Link: http://lkml.kernel.org/r/c68df2284b7d172669daf9be29db62ad49bbc559.1516069914.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> [fix noderef.cocci warnings, sizeof pointer for kcalloc of event->fields] Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2018-01-16 02:51:58 +00:00
se_name = trace_event_name(file->event_call);
se = find_synth_event(se_name);
if (se)
se->ref++;
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
out:
tracing/hist: Fix issue of losting command info in error_log When input some constructed invalid 'trigger' command, command info in 'error_log' are lost [1]. The root cause is that there is a path that event_hist_trigger_parse() is recursely called once and 'last_cmd' which save origin command is cleared, then later calling of hist_err() will no longer record origin command info: event_hist_trigger_parse() { last_cmd_set() // <1> 'last_cmd' save origin command here at first create_actions() { onmatch_create() { action_create() { trace_action_create() { trace_action_create_field_var() { create_field_var_hist() { event_hist_trigger_parse() { // <2> recursely called once hist_err_clear() // <3> 'last_cmd' is cleared here } hist_err() // <4> No longer find origin command!!! Since 'glob' is empty string while running into the recurse call, we can trickly check it and bypass the call of hist_err_clear() to solve it. [1] # cd /sys/kernel/tracing # echo "my_synth_event int v1; int v2; int v3;" >> synthetic_events # echo 'hist:keys=pid' >> events/sched/sched_waking/trigger # echo "hist:keys=next_pid:onmatch(sched.sched_waking).my_synth_event(\ pid,pid1)" >> events/sched/sched_switch/trigger # cat error_log [ 8.405018] hist:sched:sched_switch: error: Couldn't find synthetic event Command: hist:keys=next_pid:onmatch(sched.sched_waking).my_synth_event(pid,pid1) ^ [ 8.816902] hist:sched:sched_switch: error: Couldn't find field Command: hist:keys=next_pid:onmatch(sched.sched_waking).my_synth_event(pid,pid1) ^ [ 8.816902] hist:sched:sched_switch: error: Couldn't parse field variable Command: hist:keys=next_pid:onmatch(sched.sched_waking).my_synth_event(pid,pid1) ^ [ 8.999880] : error: Couldn't find field Command: ^ [ 8.999880] : error: Couldn't parse field variable Command: ^ [ 8.999880] : error: Couldn't find field Command: ^ [ 8.999880] : error: Couldn't create histogram for field Command: ^ Link: https://lore.kernel.org/linux-trace-kernel/20221207135326.3483216-1-zhengyejian1@huawei.com Cc: <mhiramat@kernel.org> Cc: <zanussi@kernel.org> Fixes: f404da6e1d46 ("tracing: Add 'last error' error facility for hist triggers") Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
2022-12-07 13:53:26 +00:00
if (ret == 0 && glob[0])
hist_err_clear();
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
return ret;
out_unreg:
event_trigger_unregister(cmd_ops, file, glob+1, trigger_data);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
out_free:
event_trigger_reset_filter(cmd_ops, trigger_data);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
remove_hist_vars(hist_data);
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
kfree(trigger_data);
destroy_hist_data(hist_data);
goto out;
}
static struct event_command trigger_hist_cmd = {
.name = "hist",
.trigger_type = ETT_EVENT_HIST,
.flags = EVENT_CMD_FL_NEEDS_REC,
.parse = event_hist_trigger_parse,
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
.reg = hist_register_trigger,
.unreg = hist_unregister_trigger,
.unreg_all = hist_unreg_all,
tracing: Add 'hist' event trigger command 'hist' triggers allow users to continually aggregate trace events, which can then be viewed afterwards by simply reading a 'hist' file containing the aggregation in a human-readable format. The basic idea is very simple and boils down to a mechanism whereby trace events, rather than being exhaustively dumped in raw form and viewed directly, are automatically 'compressed' into meaningful tables completely defined by the user. This is done strictly via single-line command-line commands and without the aid of any kind of programming language or interpreter. A surprising number of typical use cases can be accomplished by users via this simple mechanism. In fact, a large number of the tasks that users typically do using the more complicated script-based tracing tools, at least during the initial stages of an investigation, can be accomplished by simply specifying a set of keys and values to be used in the creation of a hash table. The Linux kernel trace event subsystem happens to provide an extensive list of keys and values ready-made for such a purpose in the form of the event format files associated with each trace event. By simply consulting the format file for field names of interest and by plugging them into the hist trigger command, users can create an endless number of useful aggregations to help with investigating various properties of the system. See Documentation/trace/events.txt for examples. hist triggers are implemented on top of the existing event trigger infrastructure, and as such are consistent with the existing triggers from a user's perspective as well. The basic syntax follows the existing trigger syntax. Users start an aggregation by writing a 'hist' trigger to the event of interest's trigger file: # echo hist:keys=xxx [ if filter] > event/trigger Once a hist trigger has been set up, by default it continually aggregates every matching event into a hash table using the event key and a value field named 'hitcount'. To view the aggregation at any point in time, simply read the 'hist' file in the same directory as the 'trigger' file: # cat event/hist The detailed syntax provides additional options for user control, and is described exhaustively in Documentation/trace/events.txt and in the virtual tracing/README file in the tracing subsystem. Link: http://lkml.kernel.org/r/72d263b5e1853fe9c314953b65833c3aa75479f2.1457029949.git.tom.zanussi@linux.intel.com Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com> Tested-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Namhyung Kim <namhyung@kernel.org> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2016-03-03 18:54:42 +00:00
.get_trigger_ops = event_hist_get_trigger_ops,
.set_filter = set_trigger_filter,
};
__init int register_trigger_hist_cmd(void)
{
int ret;
ret = register_event_command(&trigger_hist_cmd);
WARN_ON(ret < 0);
return ret;
}
static void
hist_enable_trigger(struct event_trigger_data *data,
struct trace_buffer *buffer, void *rec,
struct ring_buffer_event *event)
{
struct enable_trigger_data *enable_data = data->private_data;
struct event_trigger_data *test;
tracing: trigger: Replace unneeded RCU-list traversals With CONFIG_PROVE_RCU_LIST, I had many suspicious RCU warnings when I ran ftracetest trigger testcases. ----- # dmesg -c > /dev/null # ./ftracetest test.d/trigger ... # dmesg | grep "RCU-list traversed" | cut -f 2 -d ] | cut -f 2 -d " " kernel/trace/trace_events_hist.c:6070 kernel/trace/trace_events_hist.c:1760 kernel/trace/trace_events_hist.c:5911 kernel/trace/trace_events_trigger.c:504 kernel/trace/trace_events_hist.c:1810 kernel/trace/trace_events_hist.c:3158 kernel/trace/trace_events_hist.c:3105 kernel/trace/trace_events_hist.c:5518 kernel/trace/trace_events_hist.c:5998 kernel/trace/trace_events_hist.c:6019 kernel/trace/trace_events_hist.c:6044 kernel/trace/trace_events_trigger.c:1500 kernel/trace/trace_events_trigger.c:1540 kernel/trace/trace_events_trigger.c:539 kernel/trace/trace_events_trigger.c:584 ----- I investigated those warnings and found that the RCU-list traversals in event trigger and hist didn't need to use RCU version because those were called only under event_mutex. I also checked other RCU-list traversals related to event trigger list, and found that most of them were called from event_hist_trigger_func() or hist_unregister_trigger() or register/unregister functions except for a few cases. Replace these unneeded RCU-list traversals with normal list traversal macro and lockdep_assert_held() to check the event_mutex is held. Link: http://lkml.kernel.org/r/157680910305.11685.15110237954275915782.stgit@devnote2 Cc: stable@vger.kernel.org Fixes: 30350d65ac567 ("tracing: Add variable support to hist triggers") Reviewed-by: Tom Zanussi <zanussi@kernel.org> Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2019-12-20 02:31:43 +00:00
list_for_each_entry_rcu(test, &enable_data->file->triggers, list,
lockdep_is_held(&event_mutex)) {
if (test->cmd_ops->trigger_type == ETT_EVENT_HIST) {
if (enable_data->enable)
test->paused = false;
else
test->paused = true;
}
}
}
static void
hist_enable_count_trigger(struct event_trigger_data *data,
struct trace_buffer *buffer, void *rec,
struct ring_buffer_event *event)
{
if (!data->count)
return;
if (data->count != -1)
(data->count)--;
hist_enable_trigger(data, buffer, rec, event);
}
static struct event_trigger_ops hist_enable_trigger_ops = {
.trigger = hist_enable_trigger,
.print = event_enable_trigger_print,
.init = event_trigger_init,
.free = event_enable_trigger_free,
};
static struct event_trigger_ops hist_enable_count_trigger_ops = {
.trigger = hist_enable_count_trigger,
.print = event_enable_trigger_print,
.init = event_trigger_init,
.free = event_enable_trigger_free,
};
static struct event_trigger_ops hist_disable_trigger_ops = {
.trigger = hist_enable_trigger,
.print = event_enable_trigger_print,
.init = event_trigger_init,
.free = event_enable_trigger_free,
};
static struct event_trigger_ops hist_disable_count_trigger_ops = {
.trigger = hist_enable_count_trigger,
.print = event_enable_trigger_print,
.init = event_trigger_init,
.free = event_enable_trigger_free,
};
static struct event_trigger_ops *
hist_enable_get_trigger_ops(char *cmd, char *param)
{
struct event_trigger_ops *ops;
bool enable;
enable = (strcmp(cmd, ENABLE_HIST_STR) == 0);
if (enable)
ops = param ? &hist_enable_count_trigger_ops :
&hist_enable_trigger_ops;
else
ops = param ? &hist_disable_count_trigger_ops :
&hist_disable_trigger_ops;
return ops;
}
static void hist_enable_unreg_all(struct trace_event_file *file)
{
struct event_trigger_data *test, *n;
list_for_each_entry_safe(test, n, &file->triggers, list) {
if (test->cmd_ops->trigger_type == ETT_HIST_ENABLE) {
list_del_rcu(&test->list);
update_cond_flag(file);
trace_event_trigger_enable_disable(file, 0);
if (test->ops->free)
test->ops->free(test);
}
}
}
static struct event_command trigger_hist_enable_cmd = {
.name = ENABLE_HIST_STR,
.trigger_type = ETT_HIST_ENABLE,
.parse = event_enable_trigger_parse,
.reg = event_enable_register_trigger,
.unreg = event_enable_unregister_trigger,
.unreg_all = hist_enable_unreg_all,
.get_trigger_ops = hist_enable_get_trigger_ops,
.set_filter = set_trigger_filter,
};
static struct event_command trigger_hist_disable_cmd = {
.name = DISABLE_HIST_STR,
.trigger_type = ETT_HIST_ENABLE,
.parse = event_enable_trigger_parse,
.reg = event_enable_register_trigger,
.unreg = event_enable_unregister_trigger,
.unreg_all = hist_enable_unreg_all,
.get_trigger_ops = hist_enable_get_trigger_ops,
.set_filter = set_trigger_filter,
};
static __init void unregister_trigger_hist_enable_disable_cmds(void)
{
unregister_event_command(&trigger_hist_enable_cmd);
unregister_event_command(&trigger_hist_disable_cmd);
}
__init int register_trigger_hist_enable_disable_cmds(void)
{
int ret;
ret = register_event_command(&trigger_hist_enable_cmd);
if (WARN_ON(ret < 0))
return ret;
ret = register_event_command(&trigger_hist_disable_cmd);
if (WARN_ON(ret < 0))
unregister_trigger_hist_enable_disable_cmds();
return ret;
}