linux-toradex.git/tools/perf, branch v6.2-rc2

Merge tag 'perf-tools-for-v6.2-2-2022-12-22' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux

2022-12-22T19:07:29+00:00

Pull more perf tools updates from Arnaldo Carvalho de Melo:
 "perf tools fixes and improvements:

   - Don't stop building perf if python setuptools isn't installed, just
     disable the affected perf feature.

   - Remove explicit reference to python 2.x devel files, that warning
     is about python-devel, no matter what version, being unavailable
     and thus disabling the linking with libpython.

   - Don't use -Werror=switch-enum when building the python support that
     handles libtraceevent enumerations, as there is no good way to test
     if some specific enum entry is available with the libtraceevent
     installed on the system.

   - Introduce 'perf lock contention' --type-filter and --lock-filter,
     to filter by lock type and lock name:

        $ sudo ./perf lock record -a -- ./perf bench sched messaging

        $ sudo ./perf lock contention -E 5 -Y spinlock
         contended  total wait   max wait  avg wait      type  caller

               802     1.26 ms   11.73 us   1.58 us  spinlock  __wake_up_common_lock+0x62
                13   787.16 us  105.44 us  60.55 us  spinlock  remove_wait_queue+0x14
                12   612.96 us   78.70 us  51.08 us  spinlock  prepare_to_wait+0x27
               114   340.68 us   12.61 us   2.99 us  spinlock  try_to_wake_up+0x1f5
                83   226.38 us    9.15 us   2.73 us  spinlock  folio_lruvec_lock_irqsave+0x5e

        $ sudo ./perf lock contention -l
         contended  total wait  max wait  avg wait           address  symbol

                57     1.11 ms  42.83 us  19.54 us  ffff9f4140059000
                15   280.88 us  23.51 us  18.73 us  ffffffff9d007a40  jiffies_lock
                 1    20.49 us  20.49 us  20.49 us  ffffffff9d0d50c0  rcu_state
                 1     9.02 us   9.02 us   9.02 us  ffff9f41759e9ba0

        $ sudo ./perf lock contention -L jiffies_lock,rcu_state
         contended  total wait  max wait  avg wait      type  caller

                15   280.88 us  23.51 us  18.73 us  spinlock  tick_sched_do_timer+0x93
                 1    20.49 us  20.49 us  20.49 us  spinlock  __softirqentry_text_start+0xeb

        $ sudo ./perf lock contention -L ffff9f4140059000
         contended  total wait  max wait  avg wait      type  caller

                38   779.40 us  42.83 us  20.51 us  spinlock  worker_thread+0x50
                11   216.30 us  39.87 us  19.66 us  spinlock  queue_work_on+0x39
                 8   118.13 us  20.51 us  14.77 us  spinlock  kthread+0xe5

   - Fix splitting CC into compiler and options when checking if a
     option is present in clang to build the python binding, needed in
     systems such as yocto that set CC to, e.g.: "gcc --sysroot=/a/b/c".

   - Refresh metris and events for Intel systems: alderlake.
     alderlake-n, bonnell, broadwell, broadwellde, broadwellx,
     cascadelakex, elkhartlake, goldmont, goldmontplus, haswell,
     haswellx, icelake, icelakex, ivybridge, ivytown, jaketown,
     knightslanding, meteorlake, nehalemep, nehalemex, sandybridge,
     sapphirerapids, silvermont, skylake, skylakex, snowridgex,
     tigerlake, westmereep-dp, westmereep-sp, westmereex.

   - Add vendor events files (JSON) for AMD Zen 4, from sections
     2.1.15.4 "Core Performance Monitor Counters", 2.1.15.5 "L3 Cache
     Performance Monitor Counter"s and Section 7.1 "Fabric Performance
     Monitor Counter (PMC) Events" in the Processor Programming
     Reference (PPR) for AMD Family 19h Model 11h Revision B1
     processors.

     This constitutes events which capture op dispatch, execution and
     retirement, branch prediction, L1 and L2 cache activity, TLB
     activity, L3 cache activity and data bandwidth for various links
     and interfaces in the Data Fabric.

   - Also, from the same PPR are metrics taken from Section 2.1.15.2
     "Performance Measurement", including pipeline utilization, which
     are new to Zen 4 processors and useful for finding performance
     bottlenecks by analyzing activity at different stages of the
     pipeline.

   - Greatly improve the 'srcline', 'srcline_from', 'srcline_to' and
     'srcfile' sort keys performance by postponing calling the external
     addr2line utility to the collapse phase of histogram bucketing.

   - Fix 'perf test' "all PMU test" to skip parametrized events, that
     requires setting up and are not supported by this test.

   - Update tools/ copies of kernel headers: features,
     disabled-features, fscrypt.h, i915_drm.h, msr-index.h, power pc
     syscall table and kvm.h.

   - Add .DELETE_ON_ERROR special Makefile target to clean up partially
     updated files on error.

   - Simplify the mksyscalltbl script for arm64 by avoiding to run the
     host compiler to create the syscall table, do it all just with the
     shell script.

   - Further fixes to honour quiet mode (-q)"

* tag 'perf-tools-for-v6.2-2-2022-12-22' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux: (67 commits)
  perf python: Fix splitting CC into compiler and options
  perf scripting python: Don't be strict at handling libtraceevent enumerations
  perf arm64: Simplify mksyscalltbl
  perf build: Remove explicit reference to python 2.x devel files
  perf vendor events amd: Add Zen 4 mapping
  perf vendor events amd: Add Zen 4 metrics
  perf vendor events amd: Add Zen 4 uncore events
  perf vendor events amd: Add Zen 4 core events
  perf vendor events intel: Refresh westmereex events
  perf vendor events intel: Refresh westmereep-sp events
  perf vendor events intel: Refresh westmereep-dp events
  perf vendor events intel: Refresh tigerlake metrics and events
  perf vendor events intel: Refresh snowridgex events
  perf vendor events intel: Refresh skylakex metrics and events
  perf vendor events intel: Refresh skylake metrics and events
  perf vendor events intel: Refresh silvermont events
  perf vendor events intel: Refresh sapphirerapids metrics and events
  perf vendor events intel: Refresh sandybridge metrics and events
  perf vendor events intel: Refresh nehalemex events
  perf vendor events intel: Refresh nehalemep events
  ...

perf python: Fix splitting CC into compiler and options

2022-12-22T14:34:30+00:00

Noticed this build failure on archlinux:base when building with clang:

  clang-14: error: optimization flag '-ffat-lto-objects' is not supported [-Werror,-Wignored-optimization-argument]

In tools/perf/util/setup.py we check if clang supports that option, but
since commit 3cad53a6f9cdbafa ("perf python: Account for multiple words
in CC") this got broken as in the common case where CC="clang":

  >>> cc="clang"
  >>> print(cc.split()[0])
  clang
  >>> option="-ffat-lto-objects"
  >>> print(str(cc.split()[1:]) + option)
  []-ffat-lto-objects
  >>>

And then the Popen will call clang with that bogus option name that in
turn will not produce the b"unknown argument" or b"is not supported"
that this function uses to detect if the option is not available and
thus later on clang will be called with an unknown/unsupported option.

Fix it by looking if really there are options in the provided CC
variable, and if so override 'cc' with the first token and append the
options to the 'option' variable.

Fixes: 3cad53a6f9cdbafa ("perf python: Account for multiple words in CC")
Cc: Adrian Hunter 
Cc: Fangrui Song 
Cc: Florian Fainelli 
Cc: Ian Rogers 
Cc: Jiri Olsa 
Cc: John Keeping 
Cc: Khem Raj 
Cc: Leo Yan 
Cc: Michael Petlan 
Cc: Namhyung Kim 
Cc: Nathan Chancellor 
Cc: Nick Desaulniers 
Cc: Sedat Dilek 
Link: http://lore.kernel.org/lkml/Y6Rq5F5NI0v1QQHM@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo

perf scripting python: Don't be strict at handling libtraceevent enumerations

2022-12-21T20:30:38+00:00

The build was failing on archlinux because it has a newer libtraceevent
that added a new entry to the tep_print_arg_type enum:

    19.72 archlinux:base                : FAIL gcc version 12.2.0 (GCC)
    util/scripting-engines/trace-event-python.c: In function ‘define_event_symbols’:
    util/scripting-engines/trace-event-python.c:281:9: error: enumeration value ‘TEP_PRINT_CPUMASK’ not handled in switch [-Werror=switch-enum]
      281 |         switch (args->type) {
          |         ^~~~~~
    cc1: all warnings being treated as errors

Since we build with distros that have different versions of
libtraceevent and there is no way to easily test if these enum entries
are available, just disable -Werror=switch-enum for that specific
object.

Signed-off-by: Arnaldo Carvalho de Melo

perf arm64: Simplify mksyscalltbl

2022-12-21T20:13:31+00:00

This patch isn't intended to have any effect on the compiled code.  It
just removes one level of indirection: calling the *host* compiler to
build and then run a program that just printf:s the numerical entries of
the syscall-table.  In other words, the generated syscalls.c changes
from:

	[46] = "ftruncate",
to:
	[__NR3264_ftruncate] = "ftruncate",

The latter is as good as the former to the user of perf, and this can be
done directly by the shell-script.  The syscalls defined as non-literal
values (like "#define __NR_ftruncate __NR3264_ftruncate") are trivially
resolved at compile-time without namespace-leaking and/or collision for
its sole user, perf/util/syscalltbl.c, that just #includes the generated
file.  A future "-mabi=32" support would probably have to handle this
differently, but that is a pre-existing problem not affected by this
simplification.

Calling the *host* compiler only complicates things and accidentally can
get a completely wrong set of files and syscall numbers, see earlier
commits.  Note that the script parameter hostcc is now unused.

At the time of this patch, powerpc (the origin, see comments), and also
e.g. x86 has moved on, from filtering "gcc -dM -E" output to reading
separate specific text-file, a table of syscall numbers.  IMHO should
arm64 consider adopting this.

Signed-off-by: Hans-Peter Nilsson 

Cc: Alexander Shishkin 
Reviewed-by: Leo Yan 
Tested-by: Leo Yan 
Acked-by: Arnd Bergmann 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: John Garry 
Cc: Kim Phillips 
Cc: Leo Yan 
Cc: Mark Rutland 
Cc: Mathieu Poirier 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Will Deacon 
Cc: linux-arm-kernel@lists.infradead.org
Link: https://lore.kernel.org/r/20201228024159.2BB66203B5@pchp3.se.axis.com
Signed-off-by: Arnaldo Carvalho de Melo

perf build: Remove explicit reference to python 2.x devel files

2022-12-21T17:52:42+00:00

If the libpython feature test (tools/build/feature/test-libpython.c)
fails, then the python-devel is missing, it doesn't mattere if it is for
python2 or 3, remove that explicit 2.x reference.

Reported-by: Linus Torvalds 
Signed-off-by: Arnaldo Carvalho de Melo

perf vendor events amd: Add Zen 4 mapping

2022-12-21T17:52:42+00:00

Add a regular expression in the map file so that appropriate JSON event
files are used for AMD Zen 4 processors. Restrict the regular expression
for AMD Zen 3 processors to known model ranges since they also belong to
Family 19h.

Signed-off-by: Sandipan Das 
Acked-by: Ian Rogers 
Cc: Alexander Shishkin 
Cc: Ananth Narayan 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Jirka Hladky 
Cc: Mark Rutland 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Ravi Bangoria 
Cc: Stephane Eranian 
Link: https://lore.kernel.org/r/20221214082652.419965-5-sandipan.das@amd.com
Signed-off-by: Arnaldo Carvalho de Melo

perf vendor events amd: Add Zen 4 metrics

2022-12-21T17:52:42+00:00

Add metrics taken from Section 2.1.15.2 "Performance Measurement" in
the Processor Programming Reference (PPR) for AMD Family 19h Model 11h
Revision B1 processors.

The recommended metrics are sourced from Table 27 "Guidance for Common
Performance Statistics with Complex Event Selects".

The pipeline utilization metrics are sourced from Table 28 "Guidance
for Pipeline Utilization Analysis Statistics". These are new to Zen 4
processors and useful for finding performance bottlenecks by analyzing
activity at different stages of the pipeline. Metric groups have been
added for Level 1 and Level 2 analysis.

Signed-off-by: Sandipan Das 
Acked-by: Ian Rogers 
Cc: Alexander Shishkin 
Cc: Ananth Narayan 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Jirka Hladky 
Cc: Mark Rutland 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Ravi Bangoria 
Cc: Stephane Eranian 
Link: https://lore.kernel.org/r/20221214082652.419965-4-sandipan.das@amd.com
Signed-off-by: Arnaldo Carvalho de Melo

perf vendor events amd: Add Zen 4 uncore events

2022-12-21T17:52:42+00:00

Add uncore events taken from Section 2.1.15.5 "L3 Cache Performance
Monitor Counter"s and Section 7.1 "Fabric Performance Monitor Counter
(PMC) Events" in the Processor Programming Reference (PPR) for AMD
Family 19h Model 11h Revision B1 processors. This constitutes events
which capture L3 cache activity and data bandwidth for various links
and interfaces in the Data Fabric.

Signed-off-by: Sandipan Das 
Acked-by: Ian Rogers 
Cc: Alexander Shishkin 
Cc: Ananth Narayan 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Jirka Hladky 
Cc: Mark Rutland 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Ravi Bangoria 
Cc: Stephane Eranian 
Link: https://lore.kernel.org/r/20221214082652.419965-3-sandipan.das@amd.com
Signed-off-by: Arnaldo Carvalho de Melo

perf vendor events amd: Add Zen 4 core events

2022-12-21T17:52:42+00:00

Add core events taken from Section 2.1.15.4 "Core Performance Monitor
Counters" in the Processor Programming Reference (PPR) for AMD Family
19h Model 11h Revision B1 processors. This constitutes events which
capture op dispatch, execution and retirement, branch prediction, L1
and L2 cache activity, TLB activity, etc.

Signed-off-by: Sandipan Das 
Acked-by: Ian Rogers 
Cc: Alexander Shishkin 
Cc: Ananth Narayan 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Jirka Hladky 
Cc: Mark Rutland 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Ravi Bangoria 
Cc: Stephane Eranian 
Link: https://lore.kernel.org/r/20221214082652.419965-2-sandipan.das@amd.com
Signed-off-by: Arnaldo Carvalho de Melo

perf vendor events intel: Refresh westmereex events

2022-12-21T17:52:42+00:00

Update the westmereex events using the new tooling from:

  https://github.com/intel/perfmon

The events are unchanged but unused json values are removed. This
increases consistency across the json files.

Signed-off-by: Ian Rogers 
Acked-by: Kan Liang 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Caleb Biggers 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: John Garry 
Cc: Mark Rutland 
Cc: Namhyung Kim 
Cc: Perry Taylor 
Cc: Peter Zijlstra 
Cc: Stephane Eranian 
Cc: Xing Zhengjun 
Link: https://lore.kernel.org/r/20221215065510.1621979-24-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo