Pull x86 cache quality monitoring update from Thomas Gleixner:
"This update provides a complete rewrite of the Cache Quality
Monitoring (CQM) facility.
The existing CQM support was duct taped into perf with a lot of issues
and the attempts to fix those turned out to be incomplete and
horrible.
After lengthy discussions it was decided to integrate the CQM support
into the Resource Director Technology (RDT) facility, which is the
obvious choise as in hardware CQM is part of RDT. This allowed to add
Memory Bandwidth Monitoring support on top.
As a result the mechanisms for allocating cache/memory bandwidth and
the corresponding monitoring mechanisms are integrated into a single
management facility with a consistent user interface"
* 'x86-cache-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (37 commits)
x86/intel_rdt: Turn off most RDT features on Skylake
x86/intel_rdt: Add command line options for resource director technology
x86/intel_rdt: Move special case code for Haswell to a quirk function
x86/intel_rdt: Remove redundant ternary operator on return
x86/intel_rdt/cqm: Improve limbo list processing
x86/intel_rdt/mbm: Fix MBM overflow handler during CPU hotplug
x86/intel_rdt: Modify the intel_pqr_state for better performance
x86/intel_rdt/cqm: Clear the default RMID during hotcpu
x86/intel_rdt: Show bitmask of shareable resource with other executing units
x86/intel_rdt/mbm: Handle counter overflow
x86/intel_rdt/mbm: Add mbm counter initialization
x86/intel_rdt/mbm: Basic counting of MBM events (total and local)
x86/intel_rdt/cqm: Add CPU hotplug support
x86/intel_rdt/cqm: Add sched_in support
x86/intel_rdt: Introduce rdt_enable_key for scheduling
x86/intel_rdt/cqm: Add mount,umount support
x86/intel_rdt/cqm: Add rmdir support
x86/intel_rdt: Separate the ctrl bits from rmdir
x86/intel_rdt/cqm: Add mon_data
x86/intel_rdt: Prepare for RDT monitor data support
...
Move the 'max_precise' capability into generic x86 code where it
belongs. This fixes a sysfs splat on !Intel systems where we fail to set
x86_pmu_caps_group.atts.
Reported-and-tested-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: hpa@zytor.com
Fixes: 22688d1c20f5 ("x86/perf: Export some PMU attributes in caps/ directory")
Link: http://lkml.kernel.org/r/20170828104650.2u3rsim4jafyjzv2@hirez.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
For understanding how the workload maps to memory channels and hardware
behavior, it's very important to collect address maps with physical
addresses. For example, 3D XPoint access can only be found by filtering
the physical address.
Add a new sample type for physical address.
perf already has a facility to collect data virtual address. This patch
introduces a function to convert the virtual address to physical address.
The function is quite generic and can be extended to any architecture as
long as a virtual address is provided.
- For kernel direct mapping addresses, virt_to_phys is used to convert
the virtual addresses to physical address.
- For user virtual addresses, __get_user_pages_fast is used to walk the
pages tables for user physical address.
- This does not work for vmalloc addresses right now. These are not
resolved, but code to do that could be added.
The new sample type requires collecting the virtual address. The
virtual address will not be output unless SAMPLE_ADDR is applied.
For security, the physical address can only be exposed to root or
privileged user.
Tested-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Kan Liang <kan.liang@intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: acme@kernel.org
Cc: mpe@ellerman.id.au
Link: http://lkml.kernel.org/r/1503967969-48278-1-git-send-email-kan.liang@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
I just noticed that hw.itrace_started and hw.config are aliased to the
same location. Now, the PT driver happens to use both, which works out
fine by sheer luck:
- STORE(hw.itrace_start) is ordered before STORE(hw.config), in the
program order, although there are no compiler barriers to ensure that,
- to the perf_log_itrace_start() hw.itrace_start looks set at the same
time as when it is intended to be set because both stores happen in the
same path,
- hw.config is never reset to zero in the PT driver.
Now, the use of hw.config by the PT driver makes more sense (it being a
HW PMU) than messing around with itrace_started, which is an awkward API
to begin with.
This patch replaces hw.itrace_started with an attach_state bit and an
API call for the PMU drivers to use to communicate the condition.
Signed-off-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: vince@deater.net
Link: http://lkml.kernel.org/r/20170330153956.25994-1-alexander.shishkin@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
It can be difficult to figure out for user programs what features
the x86 CPU PMU driver actually supports. Currently it requires
grepping in dmesg, but dmesg is not always available.
This adds a caps directory to /sys/bus/event_source/devices/cpu/,
similar to the caps already used on intel_pt, which can be used to
discover the available capabilities cleanly.
Three capabilities are defined:
- pmu_name: Underlying CPU name known to the driver
- max_precise: Max precise level supported
- branches: Known depth of LBR.
Example:
% grep . /sys/bus/event_source/devices/cpu/caps/*
/sys/bus/event_source/devices/cpu/caps/branches:32
/sys/bus/event_source/devices/cpu/caps/max_precise:3
/sys/bus/event_source/devices/cpu/caps/pmu_name:skylake
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20170822185201.9261-3-andi@firstfloor.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Only show the Intel format attributes in sysfs when the feature is actually
supported with the current model numbers. This allows programs to probe
what format attributes are available, and give a sensible error message
to users if they are not.
This handles near all cases for intel attributes since Nehalem,
except the (obscure) case when the model number if known, but PEBS
is disabled in PERF_CAPABILITIES.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20170822185201.9261-2-andi@firstfloor.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Skylake changed the encoding of the PEBS data source field.
Some combinations are not available anymore, but some new cases
e.g. for L4 cache hit are added.
Fix up the conversion table for Skylake, similar as had been done
for Nehalem.
On Skylake server the encoding for L4 actually means persistent
memory. Handle this case too.
To properly describe it in the abstracted perf format I had to add
some new fields. Since a hit can have only one level add a new
field that is an enumeration, not a bit field to describe
the level. It can describe any level. Some numbers are also
used to describe PMEM and LFB.
Also add a new generic remote flag that can be combined with
the generic level to signify a remote cache.
And there is an extension field for the snoop indication to handle
the Forward state.
I didn't add a generic flag for hops because it's not needed
for Skylake.
I changed the existing encodings for older CPUs to also fill in the
new level and remote fields.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: acme@kernel.org
Cc: jolsa@kernel.org
Link: http://lkml.kernel.org/r/20170816222156.19953-3-andi@firstfloor.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Minor cleanup: use an explicit x86_pmu flag to handle the
missing Lock / TLB information on Nehalem, instead of always
checking the model number for each PEBS sample.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: acme@kernel.org
Cc: jolsa@kernel.org
Link: http://lkml.kernel.org/r/20170816222156.19953-2-andi@firstfloor.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Mark a couple of structures and functions as 'static', pointed out by Sparse:
warning: symbol 'bts_pmu' was not declared. Should it be static?
warning: symbol 'p4_event_aliases' was not declared. Should it be static?
warning: symbol 'rapl_attr_groups' was not declared. Should it be static?
symbol 'process_uv2_message' was not declared. Should it be static?
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Andrew Banman <abanman@hpe.com> # for the UV change
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will.deacon@arm.com>
Cc: kernel-janitors@vger.kernel.org
Link: http://lkml.kernel.org/r/20170810155709.7094-1-colin.king@canonical.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
'perf cqm' never worked due to the incompatibility between perf
infrastructure and cqm hardware support. The hardware uses RMIDs to
track the llc occupancy of tasks and these RMIDs are per package. This
makes monitoring a hierarchy like cgroup along with monitoring of tasks
separately difficult and several patches sent to lkml to fix them were
NACKed. Further more, the following issues in the current perf cqm make
it almost unusable:
1. No support to monitor the same group of tasks for which we do
allocation using resctrl.
2. It gives random and inaccurate data (mostly 0s) once we run out
of RMIDs due to issues in Recycling.
3. Recycling results in inaccuracy of data because we cannot
guarantee that the RMID was stolen from a task when it was not
pulling data into cache or even when it pulled the least data. Also
for monitoring llc_occupancy, if we stop using an RMID_x and then
start using an RMID_y after we reclaim an RMID from an other event,
we miss accounting all the occupancy that was tagged to RMID_x at a
later perf_count.
2. Recycling code makes the monitoring code complex including
scheduling because the event can lose RMID any time. Since MBM
counters count bandwidth for a period of time by taking snap shot of
total bytes at two different times, recycling complicates the way we
count MBM in a hierarchy. Also we need a spin lock while we do the
processing to account for MBM counter overflow. We also currently
use a spin lock in scheduling to prevent the RMID from being taken
away.
4. Lack of support when we run different kind of event like task,
system-wide and cgroup events together. Data mostly prints 0s. This
is also because we can have only one RMID tied to a cpu as defined
by the cqm hardware but a perf can at the same time tie multiple
events during one sched_in.
5. No support of monitoring a group of tasks. There is partial support
for cgroup but it does not work once there is a hierarchy of cgroups
or if we want to monitor a task in a cgroup and the cgroup itself.
6. No support for monitoring tasks for the lifetime without perf
overhead.
7. It reported the aggregate cache occupancy or memory bandwidth over
all sockets. But most cloud and VMM based use cases want to know the
individual per-socket usage.
Signed-off-by: Vikas Shivappa <vikas.shivappa@linux.intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: ravi.v.shankar@intel.com
Cc: tony.luck@intel.com
Cc: fenghua.yu@intel.com
Cc: peterz@infradead.org
Cc: eranian@google.com
Cc: vikas.shivappa@intel.com
Cc: ak@linux.intel.com
Cc: davidcc@google.com
Cc: reinette.chatre@intel.com
Link: http://lkml.kernel.org/r/1501017287-28083-2-git-send-email-vikas.shivappa@linux.intel.com
This patch adds two missing event extra regs for Skylake Server CHA PMU:
- TOR_INSERTS
- TOR_OCCUPANCY
Were missing support for all the filters, including opcode matchers.
Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Kan Liang <kan.liang@intel.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Link: http://lkml.kernel.org/r/1499967350-10385-6-git-send-email-kan.liang@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
There is no field c6 and link for CHA BOX FILTER.
Signed-off-by: Kan Liang <kan.liang@intel.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Link: http://lkml.kernel.org/r/1499967350-10385-5-git-send-email-kan.liang@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
PCU event format for SKX are different from snbep. Introduce a new
format group for SKX PCU.
Signed-off-by: Kan Liang <kan.liang@intel.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Link: http://lkml.kernel.org/r/1499967350-10385-3-git-send-email-kan.liang@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch fixes the event_mask and event_ext_mask for the Intel Skylake
Server UPI PMU. Bit 21 is not used as a filter. The extended umask is
from bit 32 to bit 55. Correct both umasks.
Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Kan Liang <kan.liang@intel.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Link: http://lkml.kernel.org/r/1499967350-10385-2-git-send-email-kan.liang@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
We have 2 functions using the same sched_task callback:
- PEBS drain for free running counters
- LBR save/store
Both of them are called from intel_pmu_sched_task() and
either of them can be unwillingly triggered when the
other one is configured to run.
Let's say there's PEBS drain configured in sched_task
callback for the event, but in the callback itself
(intel_pmu_sched_task()) we will also run the code for
LBR save/restore, which we did not ask for, but the
code in intel_pmu_sched_task() does not check for that.
This can lead to extra cycles in some perf monitoring,
like when we monitor PEBS event without LBR data.
# perf record --no-timestamp -c 10000 -e cycles:p ./perf bench sched pipe -l 1000000
(We need PEBS, non freq/non timestamp event to enable
the sched_task callback)
The perf stat of cycles and msr:write_msr for above
command before the change:
...
Performance counter stats for './perf record --no-timestamp -c 10000 -e cycles:p \
./perf bench sched pipe -l 1000000' (5 runs):
18,519,557,441 cycles:k
91,195,527 msr:write_msr
29.334476406 seconds time elapsed
And after the change:
...
Performance counter stats for './perf record --no-timestamp -c 10000 -e cycles:p \
./perf bench sched pipe -l 1000000' (5 runs):
18,704,973,540 cycles:k
27,184,720 msr:write_msr
16.977875900 seconds time elapsed
There's no affect on cycles:k because the sched_task happens
with events switched off, however the msr:write_msr tracepoint
counter together with almost 50% of time speedup show the
improvement.
Monitoring LBR event and having extra PEBS drain processing
in sched_task callback showed just a little speedup, because
the drain function does not do much extra work in case there
is no PEBS data.
Adding conditions to recognize the configured work that needs
to be done in the x86_pmu's sched_task callback.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Kan Liang <kan.liang@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Link: http://lkml.kernel.org/r/20170719075247.GA27506@krava
Signed-off-by: Ingo Molnar <mingo@kernel.org>
User visible:
. Initial support for namespaces, using setns to access files in
namespaces, grabbing their build-ids, etc. We still need to work
more to deal with namespaces that vanish before we can get the
needed data to do analysis, but this should be as good as what is
in bcc now (Krister Johansen)
. Add header record types to pipe-mode, now this command:
$ perf record -o - -e cycles sleep 1 | perf report --stdio --header
Will show the same as in non-pipe mode, i.e. involving a perf.data
file (David Carrillo-Cisneros)
. Implement a visual marker for fused x86 instructions in the annotate
TUI browser, available now in 'perf report', more work needed to have
it available as well in 'perf top' (Jin Yao)
Further explanation from one of Jin's patches:
│ ┌──cmpl $0x0,argp_program_version_hook
81.93 │ ├──je 20
│ │ lock cmpxchg %esi,0x38a9a4(%rip)
│ │↓ jne 29
│ │↓ jmp 43
11.47 │20:└─→cmpxch %esi,0x38a999(%rip)
That means the cmpl+je is a fused instruction pair and they should be
considered together.
. Record the branch type and then show statistics and info about
in callchain entries (Jin Yao)
Example from one of Jin's patches:
# perf record -g -j any,save_type
# perf report --branch-history --stdio --no-children
38.50% div.c:45 [.] main div
|
---main div.c:42 (RET CROSS_2M cycles:2)
compute_flag div.c:28 (cycles:2)
compute_flag div.c:27 (RET CROSS_2M cycles:1)
rand rand.c:28 (cycles:1)
rand rand.c:28 (RET CROSS_2M cycles:1)
__random random.c:298 (cycles:1)
__random random.c:297 (COND_BWD CROSS_2M cycles:1)
__random random.c:295 (cycles:1)
__random random.c:295 (COND_BWD CROSS_2M cycles:1)
__random random.c:295 (cycles:1)
__random random.c:295 (RET CROSS_2M cycles:9)
. Beautify the fcntl syscall, which is an interesting one in the sense
that infrastructure had to be put in place to change the formatters of
some arguments according to the value in a previous one, i.e. cmd
dictates how arg and the syscall return will be formatted.
(Arnaldo Carvalho de Melo
Infrastructure:
. 'perf test attr' fixes (Jiri Olsa)
Vendor events:
- Add POWER9 PMU events Sukadev (Bhattiprolu)
- Support additional POWER8+ PVR in PMU mapfile (Shriya)
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2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=eFUO
-----END PGP SIGNATURE-----
Merge tag 'perf-core-for-mingo-4.13-20170718' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core
Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo:
User visible changes:
- Initial support for namespaces, using setns to access files in
namespaces, grabbing their build-ids, etc. We still need to work
more to deal with namespaces that vanish before we can get the
needed data to do analysis, but this should be as good as what is
in bcc now (Krister Johansen)
- Add header record types to pipe-mode, now this command:
$ perf record -o - -e cycles sleep 1 | perf report --stdio --header
Will show the same as in non-pipe mode, i.e. involving a perf.data
file (David Carrillo-Cisneros)
- Implement a visual marker for fused x86 instructions in the annotate
TUI browser, available now in 'perf report', more work needed to have
it available as well in 'perf top' (Jin Yao)
Further explanation from one of Jin's patches:
│ ┌──cmpl $0x0,argp_program_version_hook
81.93 │ ├──je 20
│ │ lock cmpxchg %esi,0x38a9a4(%rip)
│ │↓ jne 29
│ │↓ jmp 43
11.47 │20:└─→cmpxch %esi,0x38a999(%rip)
That means the cmpl+je is a fused instruction pair and they should be
considered together.
- Record the branch type and then show statistics and info about
in callchain entries (Jin Yao)
Example from one of Jin's patches:
# perf record -g -j any,save_type
# perf report --branch-history --stdio --no-children
38.50% div.c:45 [.] main div
|
---main div.c:42 (RET CROSS_2M cycles:2)
compute_flag div.c:28 (cycles:2)
compute_flag div.c:27 (RET CROSS_2M cycles:1)
rand rand.c:28 (cycles:1)
rand rand.c:28 (RET CROSS_2M cycles:1)
__random random.c:298 (cycles:1)
__random random.c:297 (COND_BWD CROSS_2M cycles:1)
__random random.c:295 (cycles:1)
__random random.c:295 (COND_BWD CROSS_2M cycles:1)
__random random.c:295 (cycles:1)
__random random.c:295 (RET CROSS_2M cycles:9)
- Beautify the fcntl syscall, which is an interesting one in the sense
that infrastructure had to be put in place to change the formatters of
some arguments according to the value in a previous one, i.e. cmd
dictates how arg and the syscall return will be formatted.
(Arnaldo Carvalho de Melo
Infrastructure changes:
- 'perf test attr' fixes (Jiri Olsa)
Vendor events changes:
- Add POWER9 PMU events Sukadev (Bhattiprolu)
- Support additional POWER8+ PVR in PMU mapfile (Shriya)
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Perf already has support for disassembling the branch instruction
and using the branch type for filtering. The patch just records
the branch type in perf_branch_entry.
Before recording, the patch converts the x86 branch type to
common branch type.
Change log:
v10: Set the branch_map array to be static. The previous version
has it on stack then makes the compiler to create it every
time when the function gets called.
v9: Use __ffs() to find first bit in type in common_branch_type().
It lets the code be clear.
v8: Change PERF_BR_NONE to PERF_BR_UNKNOWN.
v7: Just convert following x86 branch types to common branch types.
X86_BR_CALL -> PERF_BR_CALL
X86_BR_RET -> PERF_BR_RET
X86_BR_JCC -> PERF_BR_COND
X86_BR_JMP -> PERF_BR_UNCOND
X86_BR_IND_CALL -> PERF_BR_IND_CALL
X86_BR_ZERO_CALL -> PERF_BR_CALL
X86_BR_IND_JMP -> PERF_BR_IND
X86_BR_SYSCALL -> PERF_BR_SYSCALL
X86_BR_SYSRET -> PERF_BR_SYSRET
Others are set to PERF_BR_NONE
v6: Not changed.
v5: Just fix the merge error. No other update.
v4: Comparing to previous version, the major changes are:
1. Uses a lookup table to convert x86 branch type to common branch
type.
2. Move the JCC forward/JCC backward and cross page computing to
user space.
3. Initialize branch type to 0 in intel_pmu_lbr_read_32 and
intel_pmu_lbr_read_64
Signed-off-by: Yao Jin <yao.jin@linux.intel.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Link: http://lkml.kernel.org/r/1500379995-6449-3-git-send-email-yao.jin@linux.intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
There's a bug in PEBs event enabling code, that prevents PEBS
freq events to work properly after non freq PEBS event was run.
freq events - perf_event_attr::freq set
-F <freq> option of perf record
PEBS events - perf_event_attr::precise_ip > 0
default for perf record
Like in following example with CPU 0 busy, we expect ~10000 samples
for following perf tool run:
# perf record -F 10000 -C 0 sleep 1
[ perf record: Woken up 2 times to write data ]
[ perf record: Captured and wrote 0.640 MB perf.data (10031 samples) ]
Everything's fine, but once we run non freq PEBS event like:
# perf record -c 10000 -C 0 sleep 1
[ perf record: Woken up 4 times to write data ]
[ perf record: Captured and wrote 1.053 MB perf.data (20061 samples) ]
the freq events start to fail like this:
# perf record -F 10000 -C 0 sleep 1
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.185 MB perf.data (40 samples) ]
The issue is in non freq PEBs event initialization of debug_store reset
field, which value is used to auto-reload the counter value after PEBS
event drain. This value is not being used for PEBS freq events, but once
we run non freq event it stays in debug_store data and screws the
sample_freq counting for PEBS freq events.
Setting the reset field to 0 for freq events.
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20170714163551.19459-1-jolsa@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Add perf core PMU support for Intel Goldmont Plus CPU cores:
- The init code is based on Goldmont.
- There is a new cache event list, based on the Goldmont cache event
list.
- All four general-purpose performance counters support PEBS.
- The first general-purpose performance counter is for reduced skid
PEBS mechanism. Using :ppp to indicate the event which want to do
reduced skid PEBS.
- Goldmont Plus has 4-wide pipeline for Topdown
Signed-off-by: Kan Liang <kan.liang@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: acme@kernel.org
Link: http://lkml.kernel.org/r/20170712134423.17766-1-kan.liang@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Goldmont microarchitecture supports C1/C3/C6, PC2/PC3/PC6/PC10 state
residency counters, the patch enables them for Apollo Lake platform.
The MSR information is based on Intel Software Developers' Manual,
Vol. 4, Order No. 335592, Table 2-6 and 2-12.
Signed-off-by: Harry Pan <harry.pan@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: bp@suse.de
Cc: davidcc@google.com
Cc: gs0622@gmail.com
Cc: lukasz.odzioba@intel.com
Cc: piotr.luc@intel.com
Cc: srinivas.pandruvada@linux.intel.com
Link: http://lkml.kernel.org/r/20170717103749.24337-1-harry.pan@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull SMP hotplug updates from Thomas Gleixner:
"This update is primarily a cleanup of the CPU hotplug locking code.
The hotplug locking mechanism is an open coded RWSEM, which allows
recursive locking. The main problem with that is the recursive nature
as it evades the full lockdep coverage and hides potential deadlocks.
The rework replaces the open coded RWSEM with a percpu RWSEM and
establishes full lockdep coverage that way.
The bulk of the changes fix up recursive locking issues and address
the now fully reported potential deadlocks all over the place. Some of
these deadlocks have been observed in the RT tree, but on mainline the
probability was low enough to hide them away."
* 'smp-hotplug-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (37 commits)
cpu/hotplug: Constify attribute_group structures
powerpc: Only obtain cpu_hotplug_lock if called by rtasd
ARM/hw_breakpoint: Fix possible recursive locking for arch_hw_breakpoint_init
cpu/hotplug: Remove unused check_for_tasks() function
perf/core: Don't release cred_guard_mutex if not taken
cpuhotplug: Link lock stacks for hotplug callbacks
acpi/processor: Prevent cpu hotplug deadlock
sched: Provide is_percpu_thread() helper
cpu/hotplug: Convert hotplug locking to percpu rwsem
s390: Prevent hotplug rwsem recursion
arm: Prevent hotplug rwsem recursion
arm64: Prevent cpu hotplug rwsem recursion
kprobes: Cure hotplug lock ordering issues
jump_label: Reorder hotplug lock and jump_label_lock
perf/tracing/cpuhotplug: Fix locking order
ACPI/processor: Use cpu_hotplug_disable() instead of get_online_cpus()
PCI: Replace the racy recursion prevention
PCI: Use cpu_hotplug_disable() instead of get_online_cpus()
perf/x86/intel: Drop get_online_cpus() in intel_snb_check_microcode()
x86/perf: Drop EXPORT of perf_check_microcode
...
Pull perf updates from Ingo Molnar:
"Most of the changes are for tooling, the main changes in this cycle were:
- Improve Intel-PT hardware tracing support, both on the kernel and
on the tooling side: PTWRITE instruction support, power events for
C-state tracing, etc. (Adrian Hunter)
- Add support to measure SMI cost to the x86 architecture, with
tooling support in 'perf stat' (Kan Liang)
- Support function filtering in 'perf ftrace', plus related
improvements (Namhyung Kim)
- Allow adding and removing fields to the default 'perf script'
columns, using + or - as field prefixes to do so (Andi Kleen)
- Allow resolving the DSO name with 'perf script -F brstack{sym,off},dso'
(Mark Santaniello)
- Add perf tooling unwind support for PowerPC (Paolo Bonzini)
- ... and various other improvements as well"
* 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (84 commits)
perf auxtrace: Add CPU filter support
perf intel-pt: Do not use TSC packets for calculating CPU cycles to TSC
perf intel-pt: Update documentation to include new ptwrite and power events
perf intel-pt: Add example script for power events and PTWRITE
perf intel-pt: Synthesize new power and "ptwrite" events
perf intel-pt: Move code in intel_pt_synth_events() to simplify attr setting
perf intel-pt: Factor out intel_pt_set_event_name()
perf intel-pt: Tidy messages into called function intel_pt_synth_event()
perf intel-pt: Tidy Intel PT evsel lookup into separate function
perf intel-pt: Join needlessly wrapped lines
perf intel-pt: Remove unused instructions_sample_period
perf intel-pt: Factor out common code synthesizing event samples
perf script: Add synthesized Intel PT power and ptwrite events
perf/x86/intel: Constify the 'lbr_desc[]' array and make a function static
perf script: Add 'synth' field for synthesized event payloads
perf auxtrace: Add itrace option to output power events
perf auxtrace: Add itrace option to output ptwrite events
tools include: Add byte-swapping macros to kernel.h
perf script: Add 'synth' event type for synthesized events
x86/insn: perf tools: Add new ptwrite instruction
...
A few minor clean-ups: constify the lbr_desc[] array and make
local function lbr_from_signext_quirk_rd() static to fix a sparse warning:
"symbol 'lbr_from_signext_quirk_rd' was not declared. Should it be static?"
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Cc: Dan Carpenter <dan.carpenter@oracle.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: kernel-janitors@vger.kernel.org
Link: http://lkml.kernel.org/r/20170629091406.9870-1-colin.king@canonical.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Should not init a NULL box. It will cause system crash.
The issue looks like caused by a typo.
This was not noticed because there is no NULL box. Also, for most
boxes, they are enabled by default. The init code is not critical.
Fixes: fff4b87e59 ("perf/x86/intel/uncore: Make package handling more robust")
Signed-off-by: Kan Liang <kan.liang@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Link: http://lkml.kernel.org/r/20170629190926.2456-1-kan.liang@intel.com
Current DTLB load/store miss events (0x608/0x649) only counts 4K,2M and
4M page size.
Need to extend the events to support any page size (4K/2M/4M/1G).
The complete DTLB load/store miss events are:
DTLB_LOAD_MISSES.WALK_COMPLETED 0xe08
DTLB_STORE_MISSES.WALK_COMPLETED 0xe49
Signed-off-by: Kan Liang <Kan.liang@intel.com>
Cc: <stable@vger.kernel.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: eranian@google.com
Link: http://lkml.kernel.org/r/20170619142609.11058-1-kan.liang@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
If intel_snb_check_microcode() is invoked via
microcode_init -> perf_check_microcode -> intel_snb_check_microcode
then get_online_cpus() is invoked nested. This works with the current
implementation of get_online_cpus() but prevents converting it to a percpu
rwsem.
intel_snb_check_microcode() is also invoked from intel_sandybridge_quirk()
unprotected.
Drop get_online_cpus() from intel_snb_check_microcode() and add it to
intel_sandybridge_quirk() so both call sites are protected.
Convert *_online_cpus() to the new interfaces while at it.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Borislav Petkov <bp@suse.de>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Borislav Petkov <bp@alien8.de>
Link: http://lkml.kernel.org/r/20170524081548.594862191@linutronix.de
intel_cqm_init() holds get_online_cpus() while registerring the hotplug
callbacks.
cpuhp_setup_state() invokes get_online_cpus() as well. This is correct, but
prevents the conversion of the hotplug locking to a percpu rwsem.
Use cpuhp_setup_state_cpuslocked() to avoid the nested call. Convert
*_online_cpus() to the new interfaces while at it.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Ingo Molnar <mingo@kernel.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/20170524081548.075604046@linutronix.de
Currently, the SMIs are visible to all performance counters, because
many users want to measure everything including SMIs. But in some
cases, the SMI cycles should not be counted - for example, to calculate
the cost of an SMI itself. So a knob is needed.
When setting FREEZE_WHILE_SMM bit in IA32_DEBUGCTL, all performance
counters will be effected. There is no way to do per-counter freeze
on SMI. So it should not use the per-event interface (e.g. ioctl or
event attribute) to set FREEZE_WHILE_SMM bit.
Adds sysfs entry /sys/device/cpu/freeze_on_smi to set FREEZE_WHILE_SMM
bit in IA32_DEBUGCTL. When set, freezes perfmon and trace messages
while in SMM.
Value has to be 0 or 1. It will be applied to all processors.
Also serialize the entire setting so we don't get multiple concurrent
threads trying to update to different values.
Signed-off-by: Kan Liang <Kan.liang@intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: acme@kernel.org
Cc: bp@alien8.de
Cc: jolsa@kernel.org
Link: http://lkml.kernel.org/r/1494600673-244667-1-git-send-email-kan.liang@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Fixes:
- Support setting probes in versioned user space symbols, such as
pthread_create@@GLIBC_2.1, picking the default one, more work
needed to make it possible to set it on the other versions, as
the 'perf probe' syntax already uses @ for other purposes.
(Paul Clarke)
- Do not special case address zero as an error for routines that
return addresses (symbol lookup), instead use the return as the
success/error indication and pass a pointer to return the address,
fixing 'perf test vmlinux' (the one that compares address between
vmlinux and kallsyms) on s/390, where the '_text' address is equal
to zero (Arnaldo Carvalho de Melo)
Infrastructure:
- More header sanitization, moving stuff out of util.h into
more appropriate headers and objects and sometimes creating
new ones (Arnaldo Carvalho de Melo)
- Refactor a duplicated code for obtaining config file name (Taeung Song)
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2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=gqm+
-----END PGP SIGNATURE-----
Merge tag 'perf-core-for-mingo-4.12-20170503' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/urgent
Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo:
Fixes:
- Support setting probes in versioned user space symbols, such as
pthread_create@@GLIBC_2.1, picking the default one, more work
needed to make it possible to set it on the other versions, as
the 'perf probe' syntax already uses @ for other purposes.
(Paul Clarke)
- Do not special case address zero as an error for routines that
return addresses (symbol lookup), instead use the return as the
success/error indication and pass a pointer to return the address,
fixing 'perf test vmlinux' (the one that compares address between
vmlinux and kallsyms) on s/390, where the '_text' address is equal
to zero (Arnaldo Carvalho de Melo)
Infrastructure changes:
- More header sanitization, moving stuff out of util.h into
more appropriate headers and objects and sometimes creating
new ones (Arnaldo Carvalho de Melo)
- Refactor a duplicated code for obtaining config file name (Taeung Song)
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
It appears as though the Broadwell-EP DRAM units share the special
units quirk with Haswell-EP/KNL.
Without this patch, you get really high results (a single DRAM using 20W
of power).
The powercap driver in drivers/powercap/intel_rapl.c already has this
change.
Signed-off-by: Vince Weaver <vincent.weaver@maine.edu>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@gmail.com>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Cc: <stable@vger.kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Spurious NMIs will be observed with the following command:
while :; do
perf record -bae "cpu/umask=0x01,event=0xcd,ldlat=0x80/pp"
-e "cpu/umask=0x03,event=0x0/"
-e "cpu/umask=0x02,event=0x0/"
-e cycles,branches,cache-misses
-e cache-references -- sleep 10
done
The bug was introduced by commit:
8077eca079 ("perf/x86/pebs: Add workaround for broken OVFL status on HSW+")
That commit clears the status bits for the counters used for PEBS
events, by masking the whole 64 bits pebs_enabled. However, only the
low 32 bits of both status and pebs_enabled are reserved for PEBS-able
counters.
For status bits 32-34 are fixed counter overflow bits. For
pebs_enabled bits 32-34 are for PEBS Load Latency.
In the test case, the PEBS Load Latency event and fixed counter event
could overflow at the same time. The fixed counter overflow bit will
be cleared by mistake. Once it is cleared, the fixed counter overflow
never be processed, which finally trigger spurious NMI.
Correct the PEBS enabled mask by ignoring the non-PEBS bits.
Signed-off-by: Kan Liang <kan.liang@intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Fixes: 8077eca079 ("perf/x86/pebs: Add workaround for broken OVFL status on HSW+")
Link: http://lkml.kernel.org/r/1491333246-3965-1-git-send-email-kan.liang@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Now that Intel PT supports more types of trace content than just branch
tracing, it may be useful to allow the user to disable branch tracing
when it is not needed.
The special case is BDW, where not setting BranchEn is not supported.
This is slightly trickier than necessary, because up to this moment
the driver has been setting BranchEn automatically and the userspace
assumes as much. Instead of reversing the semantics of BranchEn, we
introduce a 'passthrough' bit, which will forego the default and allow
the user to set BranchEn to their heart's content.
Signed-off-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: vince@deater.net
Link: http://lkml.kernel.org/r/20170206144140.14402-1-alexander.shishkin@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Since commit:
1c5ac21a0e ("perf/x86/intel/pt: Don't die on VMXON")
... PT events depend on re-scheduling to get enabled after a VMX session
has taken place. This is, in particular, a problem for CPU context events,
which don't normally get re-scheduled, unless there is a reason.
This patch changes the VMX handling so that PT event gets re-enabled
when VMX root mode exits.
Also, notify the user when there's a gap in PT data due to VMX root
mode by flagging AUX records as partial.
In combination with vmm_exclusive=0 parameter of the kvm_intel driver,
this will result in trace gaps only for the duration of the guest's
timeslices.
Signed-off-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: vince@deater.net
Link: http://lkml.kernel.org/r/20170220133352.17995-5-alexander.shishkin@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
In preparation for adding more flags to perf AUX records, introduce a
separate API for setting the flags for a session, rather than appending
more bool arguments to perf_aux_output_end. This allows to set each
flag at the time a corresponding condition is detected, instead of
tracking it in each driver's private state.
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: vince@deater.net
Link: http://lkml.kernel.org/r/20170220133352.17995-3-alexander.shishkin@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Goldmont supports full Top Down level 1 metrics (FrontendBound,
Retiring, Backend Bound and Bad Speculation).
It has 3 wide pipeline.
Signed-off-by: Kan Liang <kan.liang@intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Link: http://lkml.kernel.org/r/1486711438-80058-1-git-send-email-kan.liang@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Update to the new file paths, remove them from introductory comments.
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20170218113140.8051-1-bp@alien8.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Commit:
8ee83b2ab3 ("perf/x86/intel/pt: Add support for PTWRITE and power event tracing")
forgot to add format strings to the PT driver. So one could enable these features
by setting corresponding bits in the event config, but not by their mnemonic names.
This patch adds the format strings.
Signed-off-by: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: vince@deater.net
Fixes: 8ee83b2ab3 ("perf/x86/intel/pt: Add support for PTWRITE...")
Link: http://lkml.kernel.org/r/20170127151644.8585-2-alexander.shishkin@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The package management code in uncore relies on package mapping being
available before a CPU is started. This changed with:
9d85eb9119 ("x86/smpboot: Make logical package management more robust")
because the ACPI/BIOS information turned out to be unreliable, but that
left uncore in broken state. This was not noticed because on a regular boot
all CPUs are online before uncore is initialized.
Move the allocation to the CPU online callback and simplify the hotplug
handling. At this point the package mapping is established and correct.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sebastian Siewior <bigeasy@linutronix.de>
Cc: Stephane Eranian <eranian@google.com>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: Yasuaki Ishimatsu <yasu.isimatu@gmail.com>
Fixes: 9d85eb9119 ("x86/smpboot: Make logical package management more robust")
Link: http://lkml.kernel.org/r/20170131230141.377156255@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The recent conversion to the hotplug state machine kept two mechanisms from
the original code:
1) The first_init logic which adds the number of online CPUs in a package
to the refcount. That's wrong because the callbacks are executed for
all online CPUs.
Remove it so the refcounting is correct.
2) The on_each_cpu() call to undo box->init() in the error handling
path. That's bogus because when the prepare callback fails no box has
been initialized yet.
Remove it.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sebastian Siewior <bigeasy@linutronix.de>
Cc: Stephane Eranian <eranian@google.com>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: Yasuaki Ishimatsu <yasu.isimatu@gmail.com>
Fixes: 1a246b9f58 ("perf/x86/intel/uncore: Convert to hotplug state machine")
Link: http://lkml.kernel.org/r/20170131230141.298032324@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The package management code in RAPL relies on package mapping being
available before a CPU is started. This changed with:
9d85eb9119 ("x86/smpboot: Make logical package management more robust")
because the ACPI/BIOS information turned out to be unreliable, but that
left RAPL in broken state. This was not noticed because on a regular boot
all CPUs are online before RAPL is initialized.
A possible fix would be to reintroduce the mess which allocates a package
data structure in CPU prepare and when it turns out to already exist in
starting throw it away later in the CPU online callback. But that's a
horrible hack and not required at all because RAPL becomes functional for
perf only in the CPU online callback. That's correct because user space is
not yet informed about the CPU being onlined, so nothing caan rely on RAPL
being available on that particular CPU.
Move the allocation to the CPU online callback and simplify the hotplug
handling. At this point the package mapping is established and correct.
This also adds a missing check for available package data in the
event_init() function.
Reported-by: Yasuaki Ishimatsu <yasu.isimatu@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sebastian Siewior <bigeasy@linutronix.de>
Cc: Stephane Eranian <eranian@google.com>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Fixes: 9d85eb9119 ("x86/smpboot: Make logical package management more robust")
Link: http://lkml.kernel.org/r/20170131230141.212593966@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The CPU hotplug function intel_pmu_cpu_starting() sets
cpu_hw_events.excl_thread_id unconditionally to 1 when the shared exclusive
counters data structure is already availabe for the sibling thread.
This works during the boot process because the first sibling gets threadid
0 assigned and the second sibling which shares the data structure gets 1.
But when the first thread of the core is offlined and onlined again it
shares the data structure with the second thread and gets exclusive thread
id 1 assigned as well.
Prevent this by checking the threadid of the already online thread.
[ tglx: Rewrote changelog ]
Signed-off-by: Zhou Chengming <zhouchengming1@huawei.com>
Cc: NuoHan Qiao <qiaonuohan@huawei.com>
Cc: ak@linux.intel.com
Cc: peterz@infradead.org
Cc: kan.liang@intel.com
Cc: dave.hansen@linux.intel.com
Cc: eranian@google.com
Cc: qiaonuohan@huawei.com
Cc: davidcc@google.com
Cc: guohanjun@huawei.com
Link: http://lkml.kernel.org/r/1484536871-3131-1-git-send-email-zhouchengming1@huawei.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
--- ---
arch/x86/events/intel/core.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
Linus Torvalds