linux-stable/arch/powerpc/perf/power7-pmu.c

459 lines
11 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Performance counter support for POWER7 processors.
*
* Copyright 2009 Paul Mackerras, IBM Corporation.
*/
#include <linux/kernel.h>
perf: Do the big rename: Performance Counters -> Performance Events Bye-bye Performance Counters, welcome Performance Events! In the past few months the perfcounters subsystem has grown out its initial role of counting hardware events, and has become (and is becoming) a much broader generic event enumeration, reporting, logging, monitoring, analysis facility. Naming its core object 'perf_counter' and naming the subsystem 'perfcounters' has become more and more of a misnomer. With pending code like hw-breakpoints support the 'counter' name is less and less appropriate. All in one, we've decided to rename the subsystem to 'performance events' and to propagate this rename through all fields, variables and API names. (in an ABI compatible fashion) The word 'event' is also a bit shorter than 'counter' - which makes it slightly more convenient to write/handle as well. Thanks goes to Stephane Eranian who first observed this misnomer and suggested a rename. User-space tooling and ABI compatibility is not affected - this patch should be function-invariant. (Also, defconfigs were not touched to keep the size down.) This patch has been generated via the following script: FILES=$(find * -type f | grep -vE 'oprofile|[^K]config') sed -i \ -e 's/PERF_EVENT_/PERF_RECORD_/g' \ -e 's/PERF_COUNTER/PERF_EVENT/g' \ -e 's/perf_counter/perf_event/g' \ -e 's/nb_counters/nb_events/g' \ -e 's/swcounter/swevent/g' \ -e 's/tpcounter_event/tp_event/g' \ $FILES for N in $(find . -name perf_counter.[ch]); do M=$(echo $N | sed 's/perf_counter/perf_event/g') mv $N $M done FILES=$(find . -name perf_event.*) sed -i \ -e 's/COUNTER_MASK/REG_MASK/g' \ -e 's/COUNTER/EVENT/g' \ -e 's/\<event\>/event_id/g' \ -e 's/counter/event/g' \ -e 's/Counter/Event/g' \ $FILES ... to keep it as correct as possible. This script can also be used by anyone who has pending perfcounters patches - it converts a Linux kernel tree over to the new naming. We tried to time this change to the point in time where the amount of pending patches is the smallest: the end of the merge window. Namespace clashes were fixed up in a preparatory patch - and some stylistic fallout will be fixed up in a subsequent patch. ( NOTE: 'counters' are still the proper terminology when we deal with hardware registers - and these sed scripts are a bit over-eager in renaming them. I've undone some of that, but in case there's something left where 'counter' would be better than 'event' we can undo that on an individual basis instead of touching an otherwise nicely automated patch. ) Suggested-by: Stephane Eranian <eranian@google.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Paul Mackerras <paulus@samba.org> Reviewed-by: Arjan van de Ven <arjan@linux.intel.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: David Howells <dhowells@redhat.com> Cc: Kyle McMartin <kyle@mcmartin.ca> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: "David S. Miller" <davem@davemloft.net> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: <linux-arch@vger.kernel.org> LKML-Reference: <new-submission> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-09-21 10:02:48 +00:00
#include <linux/perf_event.h>
#include <linux/string.h>
#include <asm/reg.h>
#include <asm/cputable.h>
#include "internal.h"
/*
* Bits in event code for POWER7
*/
#define PM_PMC_SH 16 /* PMC number (1-based) for direct events */
#define PM_PMC_MSK 0xf
#define PM_PMC_MSKS (PM_PMC_MSK << PM_PMC_SH)
#define PM_UNIT_SH 12 /* TTMMUX number and setting - unit select */
#define PM_UNIT_MSK 0xf
#define PM_COMBINE_SH 11 /* Combined event bit */
#define PM_COMBINE_MSK 1
#define PM_COMBINE_MSKS 0x800
#define PM_L2SEL_SH 8 /* L2 event select */
#define PM_L2SEL_MSK 7
#define PM_PMCSEL_MSK 0xff
/*
* Bits in MMCR1 for POWER7
*/
#define MMCR1_TTM0SEL_SH 60
#define MMCR1_TTM1SEL_SH 56
#define MMCR1_TTM2SEL_SH 52
#define MMCR1_TTM3SEL_SH 48
#define MMCR1_TTMSEL_MSK 0xf
#define MMCR1_L2SEL_SH 45
#define MMCR1_L2SEL_MSK 7
#define MMCR1_PMC1_COMBINE_SH 35
#define MMCR1_PMC2_COMBINE_SH 34
#define MMCR1_PMC3_COMBINE_SH 33
#define MMCR1_PMC4_COMBINE_SH 32
#define MMCR1_PMC1SEL_SH 24
#define MMCR1_PMC2SEL_SH 16
#define MMCR1_PMC3SEL_SH 8
#define MMCR1_PMC4SEL_SH 0
#define MMCR1_PMCSEL_SH(n) (MMCR1_PMC1SEL_SH - (n) * 8)
#define MMCR1_PMCSEL_MSK 0xff
/*
* Power7 event codes.
*/
perf tools: Make Power7 events available for perf Power7 supports over 530 different perf events but only a small subset of these can be specified by name, for the remaining events, we must specify them by their raw code: perf stat -e r2003c <application> This patch makes all the POWER7 events available in sysfs. So we can instead specify these as: perf stat -e 'cpu/PM_CMPLU_STALL_DFU/' <application> where PM_CMPLU_STALL_DFU is the r2003c in previous example. Before this patch is applied, the size of power7-pmu.o is: $ size arch/powerpc/perf/power7-pmu.o text data bss dec hex filename 3073 2720 0 5793 16a1 arch/powerpc/perf/power7-pmu.o and after the patch is applied, it is: $ size arch/powerpc/perf/power7-pmu.o text data bss dec hex filename 15950 31112 0 47062 b7d6 arch/powerpc/perf/power7-pmu.o For the run time overhead, I use two scripts, one is "event_name.sh", which contains 50 event names, it looks like: # ./perf record -e 'cpu/PM_CMPLU_STALL_DFU/' -e ..... /bin/sleep 1 the other one is named "event_code.sh" which use corresponding events raw code instead of events names, it looks like: # ./perf record -e r2003c -e ...... /bin/sleep 1 below is the result. Using events name: [root@localhost perf]# time ./event_name.sh [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.002 MB perf.data (~102 samples) ] real 0m1.192s user 0m0.028s sys 0m0.106s Using events raw code: [root@localhost perf]# time ./event_code.sh [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.003 MB perf.data (~112 samples) ] real 0m1.198s user 0m0.028s sys 0m0.105s Signed-off-by: Runzhen Wang <runzhen@linux.vnet.ibm.com> Acked-by: Michael Ellerman <michael@ellerman.id.au> Cc: icycoder@gmail.com Cc: linuxppc-dev@lists.ozlabs.org Cc: Michael Ellerman <michael@ellerman.id.au> Cc: Paul Mackerras <paulus@samba.org> Cc: Runzhen Wang <runzhew@clemson.edu> Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com> Cc: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com> Link: http://lkml.kernel.org/r/1372407297-6996-3-git-send-email-runzhen@linux.vnet.ibm.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-06-28 08:14:57 +00:00
#define EVENT(_name, _code) \
_name = _code,
perf tools: Make Power7 events available for perf Power7 supports over 530 different perf events but only a small subset of these can be specified by name, for the remaining events, we must specify them by their raw code: perf stat -e r2003c <application> This patch makes all the POWER7 events available in sysfs. So we can instead specify these as: perf stat -e 'cpu/PM_CMPLU_STALL_DFU/' <application> where PM_CMPLU_STALL_DFU is the r2003c in previous example. Before this patch is applied, the size of power7-pmu.o is: $ size arch/powerpc/perf/power7-pmu.o text data bss dec hex filename 3073 2720 0 5793 16a1 arch/powerpc/perf/power7-pmu.o and after the patch is applied, it is: $ size arch/powerpc/perf/power7-pmu.o text data bss dec hex filename 15950 31112 0 47062 b7d6 arch/powerpc/perf/power7-pmu.o For the run time overhead, I use two scripts, one is "event_name.sh", which contains 50 event names, it looks like: # ./perf record -e 'cpu/PM_CMPLU_STALL_DFU/' -e ..... /bin/sleep 1 the other one is named "event_code.sh" which use corresponding events raw code instead of events names, it looks like: # ./perf record -e r2003c -e ...... /bin/sleep 1 below is the result. Using events name: [root@localhost perf]# time ./event_name.sh [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.002 MB perf.data (~102 samples) ] real 0m1.192s user 0m0.028s sys 0m0.106s Using events raw code: [root@localhost perf]# time ./event_code.sh [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.003 MB perf.data (~112 samples) ] real 0m1.198s user 0m0.028s sys 0m0.105s Signed-off-by: Runzhen Wang <runzhen@linux.vnet.ibm.com> Acked-by: Michael Ellerman <michael@ellerman.id.au> Cc: icycoder@gmail.com Cc: linuxppc-dev@lists.ozlabs.org Cc: Michael Ellerman <michael@ellerman.id.au> Cc: Paul Mackerras <paulus@samba.org> Cc: Runzhen Wang <runzhew@clemson.edu> Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com> Cc: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com> Link: http://lkml.kernel.org/r/1372407297-6996-3-git-send-email-runzhen@linux.vnet.ibm.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-06-28 08:14:57 +00:00
enum {
#include "power7-events-list.h"
};
#undef EVENT
/*
* Layout of constraint bits:
* 6666555555555544444444443333333333222222222211111111110000000000
* 3210987654321098765432109876543210987654321098765432109876543210
* < >< ><><><><><><>
* L2 NC P6P5P4P3P2P1
*
* L2 - 16-18 - Required L2SEL value (select field)
*
* NC - number of counters
* 15: NC error 0x8000
* 12-14: number of events needing PMC1-4 0x7000
*
* P6
* 11: P6 error 0x800
* 10-11: Count of events needing PMC6
*
* P1..P5
* 0-9: Count of events needing PMC1..PMC5
*/
static int power7_get_constraint(u64 event, unsigned long *maskp,
powerpc/perf: Adds support for programming of Thresholding in P10 Thresholding, a performance monitoring unit feature, can be used to identify marked instructions which take more than expected cycles between start event and end event. Threshold compare (thresh_cmp) bits are programmed in MMCRA register. In Power9, thresh_cmp bits were part of the event code. But in case of P10, thresh_cmp are not part of event code due to inclusion of MMCR3 bits. Patch here adds an option to use attr.config1 variable to be used to pass thresh_cmp value to be programmed in MMCRA register. A new ppmu flag called PPMU_HAS_ATTR_CONFIG1 has been added and this flag is used to notify the use of attr.config1 variable. Patch has extended the parameter list of 'compute_mmcr', to include power_pmu's 'flags' element and parameter list of get_constraint to include attr.config1 value. It also extend parameter list of power_check_constraints inorder to pass perf_event list. As stated by commit ef0e3b650f8d ("powerpc/perf: Fix Threshold Event Counter Multiplier width for P10"), constraint bits for thresh_cmp is also needed to be increased to 11 bits, which is handled as part of this patch. We added bit number 53 as part of constraint bits of thresh_cmp for power10 to make it an 11 bit field. Updated layout for p10: /* * Layout of constraint bits: * * 60 56 52 48 44 40 36 32 * | - - - - | - - - - | - - - - | - - - - | - - - - | - - - - | - - - - | - - - - | * [ fab_match ] [ thresh_cmp ] [ thresh_ctl ] [ ] * | | * [ thresh_cmp bits for p10] thresh_sel -* * * 28 24 20 16 12 8 4 0 * | - - - - | - - - - | - - - - | - - - - | - - - - | - - - - | - - - - | - - - - | * [ ] | [ ] | [ sample ] [ ] [6] [5] [4] [3] [2] [1] * | | | | | * BHRB IFM -* | | |*radix_scope | Count of events for each PMC. * EBB -* | | p1, p2, p3, p4, p5, p6. * L1 I/D qualifier -* | * nc - number of counters -* * * The PMC fields P1..P6, and NC, are adder fields. As we accumulate constraints * we want the low bit of each field to be added to any existing value. * * Everything else is a value field. */ Result: command#: cat /sys/devices/cpu/format/thresh_cmp config1:0-17 ex. usage: command#: perf record -I --weight -d -e cpu/event=0x67340101EC,thresh_cmp=500/ ./ebizzy -S 2 -t 1 -s 4096 1826636 records/s real 2.00 s user 2.00 s sys 0.00 s [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.038 MB perf.data (61 samples) ] Signed-off-by: Kajol Jain <kjain@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210209095234.837356-1-kjain@linux.ibm.com
2021-02-09 09:52:34 +00:00
unsigned long *valp, u64 event_config1 __maybe_unused)
{
int pmc, sh, unit;
unsigned long mask = 0, value = 0;
pmc = (event >> PM_PMC_SH) & PM_PMC_MSK;
if (pmc) {
if (pmc > 6)
return -1;
sh = (pmc - 1) * 2;
mask |= 2 << sh;
value |= 1 << sh;
if (pmc >= 5 && !(event == 0x500fa || event == 0x600f4))
return -1;
}
if (pmc < 5) {
/* need a counter from PMC1-4 set */
mask |= 0x8000;
value |= 0x1000;
}
unit = (event >> PM_UNIT_SH) & PM_UNIT_MSK;
if (unit == 6) {
/* L2SEL must be identical across events */
int l2sel = (event >> PM_L2SEL_SH) & PM_L2SEL_MSK;
mask |= 0x7 << 16;
value |= l2sel << 16;
}
*maskp = mask;
*valp = value;
return 0;
}
#define MAX_ALT 2 /* at most 2 alternatives for any event */
static const unsigned int event_alternatives[][MAX_ALT] = {
{ 0x200f2, 0x300f2 }, /* PM_INST_DISP */
{ 0x200f4, 0x600f4 }, /* PM_RUN_CYC */
{ 0x400fa, 0x500fa }, /* PM_RUN_INST_CMPL */
};
/*
* Scan the alternatives table for a match and return the
* index into the alternatives table if found, else -1.
*/
static int find_alternative(u64 event)
{
int i, j;
for (i = 0; i < ARRAY_SIZE(event_alternatives); ++i) {
if (event < event_alternatives[i][0])
break;
for (j = 0; j < MAX_ALT && event_alternatives[i][j]; ++j)
if (event == event_alternatives[i][j])
return i;
}
return -1;
}
static s64 find_alternative_decode(u64 event)
{
int pmc, psel;
/* this only handles the 4x decode events */
pmc = (event >> PM_PMC_SH) & PM_PMC_MSK;
psel = event & PM_PMCSEL_MSK;
if ((pmc == 2 || pmc == 4) && (psel & ~7) == 0x40)
return event - (1 << PM_PMC_SH) + 8;
if ((pmc == 1 || pmc == 3) && (psel & ~7) == 0x48)
return event + (1 << PM_PMC_SH) - 8;
return -1;
}
static int power7_get_alternatives(u64 event, unsigned int flags, u64 alt[])
{
int i, j, nalt = 1;
s64 ae;
alt[0] = event;
nalt = 1;
i = find_alternative(event);
if (i >= 0) {
for (j = 0; j < MAX_ALT; ++j) {
ae = event_alternatives[i][j];
if (ae && ae != event)
alt[nalt++] = ae;
}
} else {
ae = find_alternative_decode(event);
if (ae > 0)
alt[nalt++] = ae;
}
if (flags & PPMU_ONLY_COUNT_RUN) {
/*
* We're only counting in RUN state,
* so PM_CYC is equivalent to PM_RUN_CYC
* and PM_INST_CMPL === PM_RUN_INST_CMPL.
* This doesn't include alternatives that don't provide
* any extra flexibility in assigning PMCs.
*/
j = nalt;
for (i = 0; i < nalt; ++i) {
switch (alt[i]) {
case 0x1e: /* PM_CYC */
alt[j++] = 0x600f4; /* PM_RUN_CYC */
break;
case 0x600f4: /* PM_RUN_CYC */
alt[j++] = 0x1e;
break;
case 0x2: /* PM_PPC_CMPL */
alt[j++] = 0x500fa; /* PM_RUN_INST_CMPL */
break;
case 0x500fa: /* PM_RUN_INST_CMPL */
alt[j++] = 0x2; /* PM_PPC_CMPL */
break;
}
}
nalt = j;
}
return nalt;
}
/*
* Returns 1 if event counts things relating to marked instructions
* and thus needs the MMCRA_SAMPLE_ENABLE bit set, or 0 if not.
*/
static int power7_marked_instr_event(u64 event)
{
int pmc, psel;
int unit;
pmc = (event >> PM_PMC_SH) & PM_PMC_MSK;
unit = (event >> PM_UNIT_SH) & PM_UNIT_MSK;
psel = event & PM_PMCSEL_MSK & ~1; /* trim off edge/level bit */
if (pmc >= 5)
return 0;
switch (psel >> 4) {
case 2:
return pmc == 2 || pmc == 4;
case 3:
if (psel == 0x3c)
return pmc == 1;
if (psel == 0x3e)
return pmc != 2;
return 1;
case 4:
case 5:
return unit == 0xd;
case 6:
if (psel == 0x64)
return pmc >= 3;
powerpc/perf: Add missing break in power7_marked_instr_event() In power7_marked_instr_event() there is a switch case that is missing a break or an explicit fallthrough, it's not immediately clear which it should be. The function determines based on the PMU event code, whether the event is a "marked" event (which then requires us to configure the PMU in a certain way). On Power7 there is no specific bit(s) in the event to tell us that, we just have to know. Rather than having a full list of every event and whether they are marked, we pull apart the event code and for events with certain values of certain fields we can say that those are all marked events. We take the psel (bits 0-7) of the event, and look at bits 4-7. For a value of 6 we say that if the entire psel == 0x64 then if the pmc == 3 the event is marked, else not, and otherwise we continue. It is then that we fallthrough to the 8 case, where we return true if the unit == 0xd. The question is should the 6 case also fallthrough and check for unit == 0xd, or should it return. Looking at the full list of events we see that there are zero events where (psel >> 4) == 0x6 and unit == 0xd. So the answer is it doesn't really matter, there are no valid event codes that will return a different result whether we fallthrough or break. But equally, testing the 6 case events against unit == 0xd is slightly bogus, as there are no such events. So to make the code clearer, and avoid any future confusion, have the 6 case break rather than falling through. Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Reviewed-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
2018-09-20 09:41:11 +00:00
break;
case 8:
return unit == 0xd;
}
return 0;
}
static int power7_compute_mmcr(u64 event[], int n_ev,
unsigned int hwc[], struct mmcr_regs *mmcr,
powerpc/perf: Adds support for programming of Thresholding in P10 Thresholding, a performance monitoring unit feature, can be used to identify marked instructions which take more than expected cycles between start event and end event. Threshold compare (thresh_cmp) bits are programmed in MMCRA register. In Power9, thresh_cmp bits were part of the event code. But in case of P10, thresh_cmp are not part of event code due to inclusion of MMCR3 bits. Patch here adds an option to use attr.config1 variable to be used to pass thresh_cmp value to be programmed in MMCRA register. A new ppmu flag called PPMU_HAS_ATTR_CONFIG1 has been added and this flag is used to notify the use of attr.config1 variable. Patch has extended the parameter list of 'compute_mmcr', to include power_pmu's 'flags' element and parameter list of get_constraint to include attr.config1 value. It also extend parameter list of power_check_constraints inorder to pass perf_event list. As stated by commit ef0e3b650f8d ("powerpc/perf: Fix Threshold Event Counter Multiplier width for P10"), constraint bits for thresh_cmp is also needed to be increased to 11 bits, which is handled as part of this patch. We added bit number 53 as part of constraint bits of thresh_cmp for power10 to make it an 11 bit field. Updated layout for p10: /* * Layout of constraint bits: * * 60 56 52 48 44 40 36 32 * | - - - - | - - - - | - - - - | - - - - | - - - - | - - - - | - - - - | - - - - | * [ fab_match ] [ thresh_cmp ] [ thresh_ctl ] [ ] * | | * [ thresh_cmp bits for p10] thresh_sel -* * * 28 24 20 16 12 8 4 0 * | - - - - | - - - - | - - - - | - - - - | - - - - | - - - - | - - - - | - - - - | * [ ] | [ ] | [ sample ] [ ] [6] [5] [4] [3] [2] [1] * | | | | | * BHRB IFM -* | | |*radix_scope | Count of events for each PMC. * EBB -* | | p1, p2, p3, p4, p5, p6. * L1 I/D qualifier -* | * nc - number of counters -* * * The PMC fields P1..P6, and NC, are adder fields. As we accumulate constraints * we want the low bit of each field to be added to any existing value. * * Everything else is a value field. */ Result: command#: cat /sys/devices/cpu/format/thresh_cmp config1:0-17 ex. usage: command#: perf record -I --weight -d -e cpu/event=0x67340101EC,thresh_cmp=500/ ./ebizzy -S 2 -t 1 -s 4096 1826636 records/s real 2.00 s user 2.00 s sys 0.00 s [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.038 MB perf.data (61 samples) ] Signed-off-by: Kajol Jain <kjain@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20210209095234.837356-1-kjain@linux.ibm.com
2021-02-09 09:52:34 +00:00
struct perf_event *pevents[],
u32 flags __maybe_unused)
{
unsigned long mmcr1 = 0;
unsigned long mmcra = MMCRA_SDAR_DCACHE_MISS | MMCRA_SDAR_ERAT_MISS;
unsigned int pmc, unit, combine, l2sel, psel;
unsigned int pmc_inuse = 0;
int i;
/* First pass to count resource use */
for (i = 0; i < n_ev; ++i) {
pmc = (event[i] >> PM_PMC_SH) & PM_PMC_MSK;
if (pmc) {
if (pmc > 6)
return -1;
if (pmc_inuse & (1 << (pmc - 1)))
return -1;
pmc_inuse |= 1 << (pmc - 1);
}
}
/* Second pass: assign PMCs, set all MMCR1 fields */
for (i = 0; i < n_ev; ++i) {
pmc = (event[i] >> PM_PMC_SH) & PM_PMC_MSK;
unit = (event[i] >> PM_UNIT_SH) & PM_UNIT_MSK;
combine = (event[i] >> PM_COMBINE_SH) & PM_COMBINE_MSK;
l2sel = (event[i] >> PM_L2SEL_SH) & PM_L2SEL_MSK;
psel = event[i] & PM_PMCSEL_MSK;
if (!pmc) {
/* Bus event or any-PMC direct event */
for (pmc = 0; pmc < 4; ++pmc) {
if (!(pmc_inuse & (1 << pmc)))
break;
}
if (pmc >= 4)
return -1;
pmc_inuse |= 1 << pmc;
} else {
/* Direct or decoded event */
--pmc;
}
if (pmc <= 3) {
mmcr1 |= (unsigned long) unit
<< (MMCR1_TTM0SEL_SH - 4 * pmc);
mmcr1 |= (unsigned long) combine
<< (MMCR1_PMC1_COMBINE_SH - pmc);
mmcr1 |= psel << MMCR1_PMCSEL_SH(pmc);
if (unit == 6) /* L2 events */
mmcr1 |= (unsigned long) l2sel
<< MMCR1_L2SEL_SH;
}
if (power7_marked_instr_event(event[i]))
mmcra |= MMCRA_SAMPLE_ENABLE;
hwc[i] = pmc;
}
/* Return MMCRx values */
mmcr->mmcr0 = 0;
if (pmc_inuse & 1)
mmcr->mmcr0 = MMCR0_PMC1CE;
if (pmc_inuse & 0x3e)
mmcr->mmcr0 |= MMCR0_PMCjCE;
mmcr->mmcr1 = mmcr1;
mmcr->mmcra = mmcra;
return 0;
}
static void power7_disable_pmc(unsigned int pmc, struct mmcr_regs *mmcr)
{
if (pmc <= 3)
mmcr->mmcr1 &= ~(0xffUL << MMCR1_PMCSEL_SH(pmc));
}
static int power7_generic_events[] = {
[PERF_COUNT_HW_CPU_CYCLES] = PM_CYC,
[PERF_COUNT_HW_STALLED_CYCLES_FRONTEND] = PM_GCT_NOSLOT_CYC,
[PERF_COUNT_HW_STALLED_CYCLES_BACKEND] = PM_CMPLU_STALL,
[PERF_COUNT_HW_INSTRUCTIONS] = PM_INST_CMPL,
[PERF_COUNT_HW_CACHE_REFERENCES] = PM_LD_REF_L1,
[PERF_COUNT_HW_CACHE_MISSES] = PM_LD_MISS_L1,
[PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = PM_BRU_FIN,
[PERF_COUNT_HW_BRANCH_MISSES] = PM_BR_MPRED,
};
#define C(x) PERF_COUNT_HW_CACHE_##x
/*
* Table of generalized cache-related events.
* 0 means not supported, -1 means nonsensical, other values
* are event codes.
*/
static u64 power7_cache_events[C(MAX)][C(OP_MAX)][C(RESULT_MAX)] = {
[C(L1D)] = { /* RESULT_ACCESS RESULT_MISS */
[C(OP_READ)] = { 0xc880, 0x400f0 },
[C(OP_WRITE)] = { 0, 0x300f0 },
[C(OP_PREFETCH)] = { 0xd8b8, 0 },
},
[C(L1I)] = { /* RESULT_ACCESS RESULT_MISS */
[C(OP_READ)] = { 0, 0x200fc },
[C(OP_WRITE)] = { -1, -1 },
[C(OP_PREFETCH)] = { 0x408a, 0 },
},
[C(LL)] = { /* RESULT_ACCESS RESULT_MISS */
[C(OP_READ)] = { 0x16080, 0x26080 },
[C(OP_WRITE)] = { 0x16082, 0x26082 },
[C(OP_PREFETCH)] = { 0, 0 },
},
[C(DTLB)] = { /* RESULT_ACCESS RESULT_MISS */
[C(OP_READ)] = { 0, 0x300fc },
[C(OP_WRITE)] = { -1, -1 },
[C(OP_PREFETCH)] = { -1, -1 },
},
[C(ITLB)] = { /* RESULT_ACCESS RESULT_MISS */
[C(OP_READ)] = { 0, 0x400fc },
[C(OP_WRITE)] = { -1, -1 },
[C(OP_PREFETCH)] = { -1, -1 },
},
[C(BPU)] = { /* RESULT_ACCESS RESULT_MISS */
[C(OP_READ)] = { 0x10068, 0x400f6 },
[C(OP_WRITE)] = { -1, -1 },
[C(OP_PREFETCH)] = { -1, -1 },
},
[C(NODE)] = { /* RESULT_ACCESS RESULT_MISS */
[C(OP_READ)] = { -1, -1 },
[C(OP_WRITE)] = { -1, -1 },
[C(OP_PREFETCH)] = { -1, -1 },
},
};
perf tools: Make Power7 events available for perf Power7 supports over 530 different perf events but only a small subset of these can be specified by name, for the remaining events, we must specify them by their raw code: perf stat -e r2003c <application> This patch makes all the POWER7 events available in sysfs. So we can instead specify these as: perf stat -e 'cpu/PM_CMPLU_STALL_DFU/' <application> where PM_CMPLU_STALL_DFU is the r2003c in previous example. Before this patch is applied, the size of power7-pmu.o is: $ size arch/powerpc/perf/power7-pmu.o text data bss dec hex filename 3073 2720 0 5793 16a1 arch/powerpc/perf/power7-pmu.o and after the patch is applied, it is: $ size arch/powerpc/perf/power7-pmu.o text data bss dec hex filename 15950 31112 0 47062 b7d6 arch/powerpc/perf/power7-pmu.o For the run time overhead, I use two scripts, one is "event_name.sh", which contains 50 event names, it looks like: # ./perf record -e 'cpu/PM_CMPLU_STALL_DFU/' -e ..... /bin/sleep 1 the other one is named "event_code.sh" which use corresponding events raw code instead of events names, it looks like: # ./perf record -e r2003c -e ...... /bin/sleep 1 below is the result. Using events name: [root@localhost perf]# time ./event_name.sh [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.002 MB perf.data (~102 samples) ] real 0m1.192s user 0m0.028s sys 0m0.106s Using events raw code: [root@localhost perf]# time ./event_code.sh [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.003 MB perf.data (~112 samples) ] real 0m1.198s user 0m0.028s sys 0m0.105s Signed-off-by: Runzhen Wang <runzhen@linux.vnet.ibm.com> Acked-by: Michael Ellerman <michael@ellerman.id.au> Cc: icycoder@gmail.com Cc: linuxppc-dev@lists.ozlabs.org Cc: Michael Ellerman <michael@ellerman.id.au> Cc: Paul Mackerras <paulus@samba.org> Cc: Runzhen Wang <runzhew@clemson.edu> Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com> Cc: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com> Link: http://lkml.kernel.org/r/1372407297-6996-3-git-send-email-runzhen@linux.vnet.ibm.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-06-28 08:14:57 +00:00
GENERIC_EVENT_ATTR(cpu-cycles, PM_CYC);
GENERIC_EVENT_ATTR(stalled-cycles-frontend, PM_GCT_NOSLOT_CYC);
GENERIC_EVENT_ATTR(stalled-cycles-backend, PM_CMPLU_STALL);
GENERIC_EVENT_ATTR(instructions, PM_INST_CMPL);
GENERIC_EVENT_ATTR(cache-references, PM_LD_REF_L1);
GENERIC_EVENT_ATTR(cache-misses, PM_LD_MISS_L1);
GENERIC_EVENT_ATTR(branch-instructions, PM_BRU_FIN);
GENERIC_EVENT_ATTR(branch-misses, PM_BR_MPRED);
#define EVENT(_name, _code) POWER_EVENT_ATTR(_name, _name);
#include "power7-events-list.h"
#undef EVENT
#define EVENT(_name, _code) POWER_EVENT_PTR(_name),
static struct attribute *power7_events_attr[] = {
perf tools: Make Power7 events available for perf Power7 supports over 530 different perf events but only a small subset of these can be specified by name, for the remaining events, we must specify them by their raw code: perf stat -e r2003c <application> This patch makes all the POWER7 events available in sysfs. So we can instead specify these as: perf stat -e 'cpu/PM_CMPLU_STALL_DFU/' <application> where PM_CMPLU_STALL_DFU is the r2003c in previous example. Before this patch is applied, the size of power7-pmu.o is: $ size arch/powerpc/perf/power7-pmu.o text data bss dec hex filename 3073 2720 0 5793 16a1 arch/powerpc/perf/power7-pmu.o and after the patch is applied, it is: $ size arch/powerpc/perf/power7-pmu.o text data bss dec hex filename 15950 31112 0 47062 b7d6 arch/powerpc/perf/power7-pmu.o For the run time overhead, I use two scripts, one is "event_name.sh", which contains 50 event names, it looks like: # ./perf record -e 'cpu/PM_CMPLU_STALL_DFU/' -e ..... /bin/sleep 1 the other one is named "event_code.sh" which use corresponding events raw code instead of events names, it looks like: # ./perf record -e r2003c -e ...... /bin/sleep 1 below is the result. Using events name: [root@localhost perf]# time ./event_name.sh [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.002 MB perf.data (~102 samples) ] real 0m1.192s user 0m0.028s sys 0m0.106s Using events raw code: [root@localhost perf]# time ./event_code.sh [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.003 MB perf.data (~112 samples) ] real 0m1.198s user 0m0.028s sys 0m0.105s Signed-off-by: Runzhen Wang <runzhen@linux.vnet.ibm.com> Acked-by: Michael Ellerman <michael@ellerman.id.au> Cc: icycoder@gmail.com Cc: linuxppc-dev@lists.ozlabs.org Cc: Michael Ellerman <michael@ellerman.id.au> Cc: Paul Mackerras <paulus@samba.org> Cc: Runzhen Wang <runzhew@clemson.edu> Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com> Cc: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com> Link: http://lkml.kernel.org/r/1372407297-6996-3-git-send-email-runzhen@linux.vnet.ibm.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2013-06-28 08:14:57 +00:00
GENERIC_EVENT_PTR(PM_CYC),
GENERIC_EVENT_PTR(PM_GCT_NOSLOT_CYC),
GENERIC_EVENT_PTR(PM_CMPLU_STALL),
GENERIC_EVENT_PTR(PM_INST_CMPL),
GENERIC_EVENT_PTR(PM_LD_REF_L1),
GENERIC_EVENT_PTR(PM_LD_MISS_L1),
GENERIC_EVENT_PTR(PM_BRU_FIN),
GENERIC_EVENT_PTR(PM_BR_MPRED),
#include "power7-events-list.h"
#undef EVENT
NULL
};
static struct attribute_group power7_pmu_events_group = {
.name = "events",
.attrs = power7_events_attr,
};
PMU_FORMAT_ATTR(event, "config:0-19");
static struct attribute *power7_pmu_format_attr[] = {
&format_attr_event.attr,
NULL,
};
static struct attribute_group power7_pmu_format_group = {
.name = "format",
.attrs = power7_pmu_format_attr,
};
static const struct attribute_group *power7_pmu_attr_groups[] = {
&power7_pmu_format_group,
&power7_pmu_events_group,
NULL,
};
static struct power_pmu power7_pmu = {
.name = "POWER7",
.n_counter = 6,
.max_alternatives = MAX_ALT + 1,
.add_fields = 0x1555ul,
.test_adder = 0x3000ul,
.compute_mmcr = power7_compute_mmcr,
.get_constraint = power7_get_constraint,
.get_alternatives = power7_get_alternatives,
.disable_pmc = power7_disable_pmc,
.flags = PPMU_ALT_SIPR,
.attr_groups = power7_pmu_attr_groups,
.n_generic = ARRAY_SIZE(power7_generic_events),
.generic_events = power7_generic_events,
.cache_events = &power7_cache_events,
};
int init_power7_pmu(void)
{
if (!cur_cpu_spec->oprofile_cpu_type ||
strcmp(cur_cpu_spec->oprofile_cpu_type, "ppc64/power7"))
return -ENODEV;
if (pvr_version_is(PVR_POWER7p))
power7_pmu.flags |= PPMU_SIAR_VALID;
return register_power_pmu(&power7_pmu);
}