linux.git/tools/perf/tests, branch v6.5-rc2

perf test: Fix event parsing test when PERF_PMU_CAP_EXTENDED_HW_TYPE isn't supported.

2023-07-05T19:23:00+00:00

Arm has multiple PMU types for heterogeneous systems, but doesn't
currently support PERF_PMU_CAP_EXTENDED_HW_TYPE. Make the tests
support both scenarios so that they pass on Arm, and will still pass
once PERF_PMU_CAP_EXTENDED_HW_TYPE support is added.

Fixes: 27c9fcfc1e14 ("perf test: Update parse-events expectations to test for multiple events")
Acked-by: Ian Rogers 
Signed-off-by: James Clark 
Cc: Mark Rutland 
Cc: Peter Zijlstra 
Cc: Adrian Hunter 
Cc: Arnaldo Carvalho de Melo 
Cc: Jiri Olsa 
Cc: Alexander Shishkin 
Cc: Ingo Molnar 
Link: https://lore.kernel.org/r/20230705082653.23566-3-james.clark@arm.com
Signed-off-by: Namhyung Kim

perf test: Fix event parsing test on Arm

2023-07-05T19:22:46+00:00

The test looks for a PMU from sysfs with type = PERF_TYPE_RAW when
opening a raw event. Arm doesn't have a real raw PMU, only core PMUs
with unique types other than raw.

Instead of looking for a matching PMU, just test that the event type
was parsed as raw and skip the PMU search on Arm. The raw event type
test should also apply to all platforms so add it outside of the ifdef.

Fixes: aefde50a446b ("perf test: Fix parse-events tests for >1 core PMU")
Acked-by: Ian Rogers 
Signed-off-by: James Clark 
Cc: Mark Rutland 
Cc: Peter Zijlstra 
Cc: Adrian Hunter 
Cc: Arnaldo Carvalho de Melo 
Cc: Jiri Olsa 
Cc: Alexander Shishkin 
Cc: Ingo Molnar 
Link: https://lore.kernel.org/r/20230705082653.23566-2-james.clark@arm.com
Signed-off-by: Namhyung Kim

perf test: Test perf lock contention CSV output

2023-07-01T17:49:03+00:00

To verify CSV output, just check the number of separators (",") using
the tr and wc commands like this.

  grep -v "^#" ${result} | tr -d -c | wc -c

Now it expects 6 columns (and 5 separators) in the output, but it may
be changed later so count the field in the header first and compare it
to the actual output lines.

  $ cat ${result}
  # output: contended, total wait, max wait, avg wait, type, caller
  1, 28787, 28787, 28787, spinlock, raw_spin_rq_lock_nested+0x1b

The test looks like below now:

  $ sudo ./perf test -v contention
   86: kernel lock contention analysis test                            :
  --- start ---
  test child forked, pid 2705822
  Testing perf lock record and perf lock contention
  Testing perf lock contention --use-bpf
  Testing perf lock record and perf lock contention at the same time
  Testing perf lock contention --threads
  Testing perf lock contention --lock-addr
  Testing perf lock contention --type-filter (w/ spinlock)
  Testing perf lock contention --lock-filter (w/ tasklist_lock)
  Testing perf lock contention --callstack-filter (w/ unix_stream)
  Testing perf lock contention --callstack-filter with task aggregation
  Testing perf lock contention CSV output
  test child finished with 0
  ---- end ----
  kernel lock contention analysis test: Ok

Acked-by: Ian Rogers 
Cc: Hao Luo 
Cc: Peter Zijlstra 
Cc: Adrian Hunter 
Cc: Arnaldo Carvalho de Melo 
Cc: Jiri Olsa 
Cc: Ingo Molnar 
Cc: Song Liu 
Link: https://lore.kernel.org/r/20230628200141.2739587-5-namhyung@kernel.org
Signed-off-by: Namhyung Kim

perf expr: Add has_event function

2023-06-30T05:13:15+00:00

Some events are dependent on firmware/kernel enablement. Allow such
events to be detected when the metric is parsed so that the metric's
event parsing doesn't fail.

Signed-off-by: Ian Rogers 
Tested-by: Namhyung Kim 
Cc: Mark Rutland 
Cc: Eduard Zingerman 
Cc: Sohom Datta 
Cc: Peter Zijlstra 
Cc: Adrian Hunter 
Cc: Caleb Biggers 
Cc: Edward Baker 
Cc: Perry Taylor 
Cc: Samantha Alt 
Cc: Weilin Wang 
Cc: Arnaldo Carvalho de Melo 
Cc: Andrii Nakryiko 
Cc: Jiri Olsa 
Cc: Jing Zhang 
Cc: Kajol Jain 
Cc: Alexander Shishkin 
Cc: Kan Liang 
Cc: Zhengjun Xing 
Cc: John Garry 
Cc: Ingo Molnar 
Link: https://lore.kernel.org/r/20230623151016.4193660-2-irogers@google.com
Signed-off-by: Namhyung Kim

perf test: Fix a compile error on pe-file-parsing.c

2023-06-27T19:14:38+00:00

The dso__find_symbol_by_name() should be have idx pointer argument.
Found during the build-test.

  $ make build-test
  ...
    CC      /tmp/tmp.6JwPK1xbWG/tests/pe-file-parsing.o
  tests/pe-file-parsing.c: In function ‘run_dir’:
  tests/pe-file-parsing.c:64:15: error: too few arguments to function ‘dso__find_symbol_by_name’
     64 |         sym = dso__find_symbol_by_name(dso, "main");
        |               ^~~~~~~~~~~~~~~~~~~~~~~~
  In file included from tests/pe-file-parsing.c:16:
  /usr/local/google/home/namhyung/project/linux/tools/perf/util/symbol.h:135:16: note: declared here
    135 | struct symbol *dso__find_symbol_by_name(struct dso *dso, const char *name, size_t *idx);
        |                ^~~~~~~~~~~~~~~~~~~~~~~~

Fixes: 259dce914e93 ("perf symbol: Remove symbol_name_rb_node")
Acked-by: Ian Rogers 
Cc: Peter Zijlstra 
Cc: Adrian Hunter 
Cc: Arnaldo Carvalho de Melo 
Cc: Jiri Olsa 
Cc: Ingo Molnar 
Link: https://lore.kernel.org/r/20230627063257.549005-1-namhyung@kernel.org
Signed-off-by: Namhyung Kim

perf symbol: Remove now unused symbol_conf.sort_by_name

2023-06-24T04:48:49+00:00

Previously used to specify symbol_name_rb_node was in use.

Signed-off-by: Ian Rogers 
Acked-by: Namhyung Kim 
Cc: Carsten Haitzler 
Cc: Mark Rutland 
Cc: Jason Wang 
Cc: Changbin Du 
Cc: Yang Jihong 
Cc: Peter Zijlstra 
Cc: Adrian Hunter 
Cc: Arnaldo Carvalho de Melo 
Cc: Jiri Olsa 
Cc: Alexander Shishkin 
Cc: Kan Liang 
Cc: Athira Rajeev 
Cc: Ingo Molnar 
Cc: Christophe JAILLET 
Link: https://lore.kernel.org/r/20230623054520.4118442-4-irogers@google.com
Signed-off-by: Namhyung Kim

perf test: Add build tests for BUILD_BPF_SKEL

2023-06-24T04:35:46+00:00

Add tests with and without generating vmlinux.h.

Signed-off-by: Ian Rogers 
Acked-by: Andrii Nakryiko 
Acked-by: Namhyung Kim 
Acked-by: Jiri Olsa 
Cc: James Clark 
Cc: Mark Rutland 
Cc: Yang Jihong 
Cc: Peter Zijlstra 
Cc: Adrian Hunter 
Cc: Arnaldo Carvalho de Melo 
Cc: Alexander Shishkin 
Cc: Tiezhu Yang 
Cc: Ingo Molnar 
Cc: bpf@vger.kernel.org
Link: https://lore.kernel.org/r/20230623041405.4039475-4-irogers@google.com
Signed-off-by: Namhyung Kim

perf test: Skip metrics w/o event name in stat STD output linter

2023-06-24T04:35:45+00:00

This test checks if the output of perf stat to match event names and
metrics.  So it wants the output lines to have both event name and
metric.  Otherwise it should skip the line.

On AMD machines, the instruction event has two metrics and they are printed
in separate lines.  It makes the line without event name like below:

  # perf stat -a sleep 1

   Performance counter stats for 'system wide':

           64,383.34 msec cpu-clock                  #   64.048 CPUs utilized
              14,526      context-switches           #  225.617 /sec
                 112      cpu-migrations             #    1.740 /sec
                 190      page-faults                #    2.951 /sec
         807,558,652      cycles                     #    0.013 GHz                         (83.30%)
          69,809,799      stalled-cycles-frontend    #    8.64% frontend cycles idle        (83.30%)
         196,983,266      stalled-cycles-backend     #   24.39% backend cycles idle         (83.30%)
         424,876,008      instructions               #    0.53  insn per cycle
 (here) --->                                  #    0.46  stalled cycles per insn     (83.30%)
          97,788,321      branches                   #    1.519 M/sec                       (83.34%)
           4,147,377      branch-misses              #    4.24% of all branches             (83.46%)

         1.005241409 seconds time elapsed

Also modern Intel machines have TopDown metrics which also don't have
event names.

  # perf stat -a sleep 1

   Performance counter stats for 'system wide':

            8,015.39 msec cpu-clock                        #    7.996 CPUs utilized
               5,823      context-switches                 #  726.477 /sec
                 189      cpu-migrations                   #   23.580 /sec
                 139      page-faults                      #   17.342 /sec
         435,139,308      cycles                           #    0.054 GHz
         193,891,345      instructions                     #    0.45  insn per cycle
          42,773,028      branches                         #    5.336 M/sec
           2,298,113      branch-misses                    #    5.37% of all branches
                          TopdownL1                 #     25.5 %  tma_backend_bound
              /-->                                  #      7.9 %  tma_bad_speculation
    (here) --+                                      #     55.7 %  tma_frontend_bound
              \-->                                  #     10.9 %  tma_retiring

         1.002395924 seconds time elapsed

There is a check to skip TopdownL1 and TopdownL2 specifically but it
does not cover every affected lines.

So there is another check to skip the line if it has nothing on the left
side of # sign.  Well.. it seems ok but that's not enough too.

When aggregation mode (like --per-socket or --per-thread) is used, it
adds some prefix (e.g. CPU socket, task name and PID) in the output
line.  So the test code ignores them to normalize result.

A problem can happen for per-thread mode when task name contains one or
more spaces.  It'd only ignore the first part of the task name, and it
thinks there's something more in the line so it would not skip.

  # perf stat -a --perf-thread sleep 1
  ...
            perf-21276                  #     70.2 %  tma_backend_bound
            perf-21276                  #      3.9 %  tma_bad_speculation
            perf-21276                  #     10.5 %  tma_frontend_bound
            perf-21276                  #     15.3 %  tma_retiring
	    ^^^^^^^^^^
	    (ignored)

         my task-21328                  #     70.2 %  tma_backend_bound
         my task-21328                  #      3.9 %  tma_bad_speculation
         my task-21328                  #     10.5 %  tma_frontend_bound
         my task-21328                  #     15.3 %  tma_retiring
	 ^^
     (ignored)

So I think it should look at the metric names instead.  Add skip_metric
to hold the list of names to skip.  It would contain 'stalled cycles per
insn' and metrics started by 'tma_'.

Fixes: 99a04a48f225 ("perf test: Add test case for the standard 'perf stat' output")
Acked-by: Ian Rogers 
Cc: Kan Liang 
Cc: Peter Zijlstra 
Cc: Adrian Hunter 
Cc: Arnaldo Carvalho de Melo 
Cc: Jiri Olsa 
Cc: Ingo Molnar 
Link: https://lore.kernel.org/r/20230623230139.985594-2-namhyung@kernel.org
Signed-off-by: Namhyung Kim

perf test: Reorder event name checks in stat STD output linter

2023-06-24T04:35:45+00:00

On AMD machines, the perf stat STD output test failed like below:

  $ sudo ./perf test -v 98
   98: perf stat STD output linter                                     :
  --- start ---
  test child forked, pid 1841901
  Checking STD output: no argswrong event metric.
    expected 'GHz' in 108,121 stalled-cycles-frontend  # 10.88% frontend cycles idle
  test child finished with -1
  ---- end ----
  perf stat STD output linter: FAILED!

This is because there are stalled-cycles-{frontend,backend} events are
used by default.  The current logic checks the event_name array to find
which event it's running.  But 'cycles' event comes before those stalled
cycles event and it matches first.  So it tries to find 'GHz' metric
in the output (which is for the 'cycles') and fails.

Move the stalled-cycles-{frontend,backend} events before 'cycles' so
that it can find the stalled cycles events first.

Also add a space after 'no args' test name for consistency.

Fixes: 99a04a48f225 ("perf test: Add test case for the standard 'perf stat' output")
Acked-by: Ian Rogers 
Cc: Kan Liang 
Cc: Peter Zijlstra 
Cc: Adrian Hunter 
Cc: Arnaldo Carvalho de Melo 
Cc: Jiri Olsa 
Cc: Ingo Molnar 
Link: https://lore.kernel.org/r/20230623230139.985594-1-namhyung@kernel.org
Signed-off-by: Namhyung Kim

perf test: Set PERF_EXEC_PATH for script execution

2023-06-23T05:11:13+00:00

The task-analyzer.py script (actually every other scripts too) requires
PERF_EXEC_PATH env to find dependent libraries and scripts. For scripts
test to run correctly, it needs to set PERF_EXEC_PATH to the perf tool
source directory.

Instead of blindly update the env, let's check the directory structure
to make sure it points to the correct location.

Fixes: e8478b84d6ba ("perf test: add new task-analyzer tests")
Cc: Petar Gligoric 
Cc: Hagen Paul Pfeifer 
Cc: Aditya Gupta 
Cc: Peter Zijlstra 
Cc: Adrian Hunter 
Cc: Arnaldo Carvalho de Melo 
Cc: Jiri Olsa 
Cc: Ingo Molnar 
Acked-by: Ian Rogers 
Signed-off-by: Namhyung Kim