2017-05-14 12:52:01 +00:00
|
|
|
===========================================
|
|
|
|
|
How CPU topology info is exported via sysfs
|
|
|
|
|
===========================================
|
[PATCH] Export cpu topology in sysfs
The patch implements cpu topology exportation by sysfs.
Items (attributes) are similar to /proc/cpuinfo.
1) /sys/devices/system/cpu/cpuX/topology/physical_package_id:
represent the physical package id of cpu X;
2) /sys/devices/system/cpu/cpuX/topology/core_id:
represent the cpu core id to cpu X;
3) /sys/devices/system/cpu/cpuX/topology/thread_siblings:
represent the thread siblings to cpu X in the same core;
4) /sys/devices/system/cpu/cpuX/topology/core_siblings:
represent the thread siblings to cpu X in the same physical package;
To implement it in an architecture-neutral way, a new source file,
driver/base/topology.c, is to export the 5 attributes.
If one architecture wants to support this feature, it just needs to
implement 4 defines, typically in file include/asm-XXX/topology.h.
The 4 defines are:
#define topology_physical_package_id(cpu)
#define topology_core_id(cpu)
#define topology_thread_siblings(cpu)
#define topology_core_siblings(cpu)
The type of **_id is int.
The type of siblings is cpumask_t.
To be consistent on all architectures, the 4 attributes should have
deafult values if their values are unavailable. Below is the rule.
1) physical_package_id: If cpu has no physical package id, -1 is the
default value.
2) core_id: If cpu doesn't support multi-core, its core id is 0.
3) thread_siblings: Just include itself, if the cpu doesn't support
HT/multi-thread.
4) core_siblings: Just include itself, if the cpu doesn't support
multi-core and HT/Multi-thread.
So be careful when declaring the 4 defines in include/asm-XXX/topology.h.
If an attribute isn't defined on an architecture, it won't be exported.
Thank Nathan, Greg, Andi, Paul and Venki.
The patch provides defines for i386/x86_64/ia64.
Signed-off-by: Zhang, Yanmin <yanmin.zhang@intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-02-03 11:04:36 +00:00
|
|
|
|
2021-06-11 05:22:49 +00:00
|
|
|
CPU topology info is exported via sysfs. Items (attributes) are similar
|
|
|
|
|
to /proc/cpuinfo output of some architectures. They reside in
|
|
|
|
|
/sys/devices/system/cpu/cpuX/topology/. Please refer to the ABI file:
|
|
|
|
|
Documentation/ABI/stable/sysfs-devices-system-cpu.
|
2016-05-21 09:10:13 +00:00
|
|
|
|
2019-02-26 06:20:00 +00:00
|
|
|
Architecture-neutral, drivers/base/topology.c, exports these attributes.
|
2021-11-29 13:03:09 +00:00
|
|
|
However the die, cluster, book, and drawer hierarchy related sysfs files will
|
|
|
|
|
only be created if an architecture provides the related macros as described
|
|
|
|
|
below.
|
[PATCH] Export cpu topology in sysfs
The patch implements cpu topology exportation by sysfs.
Items (attributes) are similar to /proc/cpuinfo.
1) /sys/devices/system/cpu/cpuX/topology/physical_package_id:
represent the physical package id of cpu X;
2) /sys/devices/system/cpu/cpuX/topology/core_id:
represent the cpu core id to cpu X;
3) /sys/devices/system/cpu/cpuX/topology/thread_siblings:
represent the thread siblings to cpu X in the same core;
4) /sys/devices/system/cpu/cpuX/topology/core_siblings:
represent the thread siblings to cpu X in the same physical package;
To implement it in an architecture-neutral way, a new source file,
driver/base/topology.c, is to export the 5 attributes.
If one architecture wants to support this feature, it just needs to
implement 4 defines, typically in file include/asm-XXX/topology.h.
The 4 defines are:
#define topology_physical_package_id(cpu)
#define topology_core_id(cpu)
#define topology_thread_siblings(cpu)
#define topology_core_siblings(cpu)
The type of **_id is int.
The type of siblings is cpumask_t.
To be consistent on all architectures, the 4 attributes should have
deafult values if their values are unavailable. Below is the rule.
1) physical_package_id: If cpu has no physical package id, -1 is the
default value.
2) core_id: If cpu doesn't support multi-core, its core id is 0.
3) thread_siblings: Just include itself, if the cpu doesn't support
HT/multi-thread.
4) core_siblings: Just include itself, if the cpu doesn't support
multi-core and HT/Multi-thread.
So be careful when declaring the 4 defines in include/asm-XXX/topology.h.
If an attribute isn't defined on an architecture, it won't be exported.
Thank Nathan, Greg, Andi, Paul and Venki.
The patch provides defines for i386/x86_64/ia64.
Signed-off-by: Zhang, Yanmin <yanmin.zhang@intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-02-03 11:04:36 +00:00
|
|
|
|
2008-06-05 04:47:29 +00:00
|
|
|
For an architecture to support this feature, it must define some of
|
2017-05-14 12:52:01 +00:00
|
|
|
these macros in include/asm-XXX/topology.h::
|
|
|
|
|
|
|
|
|
|
#define topology_physical_package_id(cpu)
|
2019-05-13 17:58:47 +00:00
|
|
|
#define topology_die_id(cpu)
|
2021-09-24 08:51:02 +00:00
|
|
|
#define topology_cluster_id(cpu)
|
2017-05-14 12:52:01 +00:00
|
|
|
#define topology_core_id(cpu)
|
|
|
|
|
#define topology_book_id(cpu)
|
|
|
|
|
#define topology_drawer_id(cpu)
|
|
|
|
|
#define topology_sibling_cpumask(cpu)
|
|
|
|
|
#define topology_core_cpumask(cpu)
|
2021-09-24 08:51:02 +00:00
|
|
|
#define topology_cluster_cpumask(cpu)
|
2019-05-13 17:58:56 +00:00
|
|
|
#define topology_die_cpumask(cpu)
|
2017-05-14 12:52:01 +00:00
|
|
|
#define topology_book_cpumask(cpu)
|
|
|
|
|
#define topology_drawer_cpumask(cpu)
|
|
|
|
|
|
|
|
|
|
The type of ``**_id macros`` is int.
|
|
|
|
|
The type of ``**_cpumask macros`` is ``(const) struct cpumask *``. The latter
|
|
|
|
|
correspond with appropriate ``**_siblings`` sysfs attributes (except for
|
2015-05-26 13:11:29 +00:00
|
|
|
topology_sibling_cpumask() which corresponds with thread_siblings).
|
[PATCH] Export cpu topology in sysfs
The patch implements cpu topology exportation by sysfs.
Items (attributes) are similar to /proc/cpuinfo.
1) /sys/devices/system/cpu/cpuX/topology/physical_package_id:
represent the physical package id of cpu X;
2) /sys/devices/system/cpu/cpuX/topology/core_id:
represent the cpu core id to cpu X;
3) /sys/devices/system/cpu/cpuX/topology/thread_siblings:
represent the thread siblings to cpu X in the same core;
4) /sys/devices/system/cpu/cpuX/topology/core_siblings:
represent the thread siblings to cpu X in the same physical package;
To implement it in an architecture-neutral way, a new source file,
driver/base/topology.c, is to export the 5 attributes.
If one architecture wants to support this feature, it just needs to
implement 4 defines, typically in file include/asm-XXX/topology.h.
The 4 defines are:
#define topology_physical_package_id(cpu)
#define topology_core_id(cpu)
#define topology_thread_siblings(cpu)
#define topology_core_siblings(cpu)
The type of **_id is int.
The type of siblings is cpumask_t.
To be consistent on all architectures, the 4 attributes should have
deafult values if their values are unavailable. Below is the rule.
1) physical_package_id: If cpu has no physical package id, -1 is the
default value.
2) core_id: If cpu doesn't support multi-core, its core id is 0.
3) thread_siblings: Just include itself, if the cpu doesn't support
HT/multi-thread.
4) core_siblings: Just include itself, if the cpu doesn't support
multi-core and HT/Multi-thread.
So be careful when declaring the 4 defines in include/asm-XXX/topology.h.
If an attribute isn't defined on an architecture, it won't be exported.
Thank Nathan, Greg, Andi, Paul and Venki.
The patch provides defines for i386/x86_64/ia64.
Signed-off-by: Zhang, Yanmin <yanmin.zhang@intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-02-03 11:04:36 +00:00
|
|
|
|
2008-06-05 04:47:29 +00:00
|
|
|
To be consistent on all architectures, include/linux/topology.h
|
|
|
|
|
provides default definitions for any of the above macros that are
|
|
|
|
|
not defined by include/asm-XXX/topology.h:
|
2017-05-14 12:52:01 +00:00
|
|
|
|
2021-12-06 12:53:11 +00:00
|
|
|
1) topology_physical_package_id: -1
|
|
|
|
|
2) topology_die_id: -1
|
|
|
|
|
3) topology_cluster_id: -1
|
|
|
|
|
4) topology_core_id: 0
|
|
|
|
|
5) topology_book_id: -1
|
|
|
|
|
6) topology_drawer_id: -1
|
|
|
|
|
7) topology_sibling_cpumask: just the given CPU
|
|
|
|
|
8) topology_core_cpumask: just the given CPU
|
|
|
|
|
9) topology_cluster_cpumask: just the given CPU
|
2021-11-29 13:03:09 +00:00
|
|
|
10) topology_die_cpumask: just the given CPU
|
|
|
|
|
11) topology_book_cpumask: just the given CPU
|
|
|
|
|
12) topology_drawer_cpumask: just the given CPU
|
2010-08-31 08:28:17 +00:00
|
|
|
|
2009-10-22 03:45:31 +00:00
|
|
|
Additionally, CPU topology information is provided under
|
2008-12-17 22:14:30 +00:00
|
|
|
/sys/devices/system/cpu and includes these files. The internal
|
|
|
|
|
source for the output is in brackets ("[]").
|
|
|
|
|
|
2017-05-14 12:52:01 +00:00
|
|
|
=========== ==========================================================
|
2009-10-22 03:45:31 +00:00
|
|
|
kernel_max: the maximum CPU index allowed by the kernel configuration.
|
2008-12-17 22:14:30 +00:00
|
|
|
[NR_CPUS-1]
|
|
|
|
|
|
2009-10-22 03:45:31 +00:00
|
|
|
offline: CPUs that are not online because they have been
|
2021-06-23 01:37:48 +00:00
|
|
|
HOTPLUGGED off or exceed the limit of CPUs allowed by the
|
|
|
|
|
kernel configuration (kernel_max above).
|
|
|
|
|
[~cpu_online_mask + cpus >= NR_CPUS]
|
2008-12-17 22:14:30 +00:00
|
|
|
|
2009-10-22 03:45:31 +00:00
|
|
|
online: CPUs that are online and being scheduled [cpu_online_mask]
|
2008-12-17 22:14:30 +00:00
|
|
|
|
2009-10-22 03:45:31 +00:00
|
|
|
possible: CPUs that have been allocated resources and can be
|
2008-12-17 22:14:30 +00:00
|
|
|
brought online if they are present. [cpu_possible_mask]
|
|
|
|
|
|
2009-10-22 03:45:31 +00:00
|
|
|
present: CPUs that have been identified as being present in the
|
2008-12-17 22:14:30 +00:00
|
|
|
system. [cpu_present_mask]
|
2017-05-14 12:52:01 +00:00
|
|
|
=========== ==========================================================
|
2008-12-17 22:14:30 +00:00
|
|
|
|
|
|
|
|
The format for the above output is compatible with cpulist_parse()
|
|
|
|
|
[see <linux/cpumask.h>]. Some examples follow.
|
|
|
|
|
|
2009-10-22 03:45:31 +00:00
|
|
|
In this example, there are 64 CPUs in the system but cpus 32-63 exceed
|
2008-12-17 22:14:30 +00:00
|
|
|
the kernel max which is limited to 0..31 by the NR_CPUS config option
|
2009-10-22 03:45:31 +00:00
|
|
|
being 32. Note also that CPUs 2 and 4-31 are not online but could be
|
2017-05-14 12:52:01 +00:00
|
|
|
brought online as they are both present and possible::
|
2008-12-17 22:14:30 +00:00
|
|
|
|
|
|
|
|
kernel_max: 31
|
|
|
|
|
offline: 2,4-31,32-63
|
|
|
|
|
online: 0-1,3
|
|
|
|
|
possible: 0-31
|
|
|
|
|
present: 0-31
|
|
|
|
|
|
|
|
|
|
In this example, the NR_CPUS config option is 128, but the kernel was
|
2009-10-22 03:45:31 +00:00
|
|
|
started with possible_cpus=144. There are 4 CPUs in the system and cpu2
|
|
|
|
|
was manually taken offline (and is the only CPU that can be brought
|
2017-05-14 12:52:01 +00:00
|
|
|
online.)::
|
2008-12-17 22:14:30 +00:00
|
|
|
|
|
|
|
|
kernel_max: 127
|
|
|
|
|
offline: 2,4-127,128-143
|
|
|
|
|
online: 0-1,3
|
|
|
|
|
possible: 0-127
|
|
|
|
|
present: 0-3
|
|
|
|
|
|
2021-06-23 01:37:48 +00:00
|
|
|
See Documentation/core-api/cpu_hotplug.rst for the possible_cpus=NUM
|
|
|
|
|
kernel start parameter as well as more information on the various cpumasks.
|