perf script: Display off-cpu samples correctly

No PERF_SAMPLE_CALLCHAIN in sample_type, but 'perf script' needs to
display a callchain, have to specify manually.

Also, prefer displaying a callchain:

 gvfs-afc-volume    2267 [001] 3829232.955656: 1001115340 offcpu-time:
            77f05292603f __pselect+0xbf (/usr/lib/x86_64-linux-gnu/libc.so.6)
            77f052a1801c [unknown] (/usr/lib/x86_64-linux-gnu/libusbmuxd-2.0.so.6.0.0)
            77f052a18d45 [unknown] (/usr/lib/x86_64-linux-gnu/libusbmuxd-2.0.so.6.0.0)
            77f05289ca94 start_thread+0x384 (/usr/lib/x86_64-linux-gnu/libc.so.6)
            77f052929c3c clone3+0x2c (/usr/lib/x86_64-linux-gnu/libc.so.6)

to a raw binary BPF output:

  BPF output: 0000: dd 08 00 00 db 08 00 00  <DD>...<DB>...
  0008: cc ce ab 3b 00 00 00 00  <CC>Ϋ;....
  0010: 06 00 00 00 00 00 00 00  ........
  0018: 00 fe ff ff ff ff ff ff  .<FE><FF><FF><FF><FF><FF><FF>
  0020: 3f 60 92 52 f0 77 00 00  ?`.R<F0>w..
  0028: 1c 80 a1 52 f0 77 00 00  ..<A1>R<F0>w..
  0030: 45 8d a1 52 f0 77 00 00  E.<A1>R<F0>w..
  0038: 94 ca 89 52 f0 77 00 00  .<CA>.R<F0>w..
  0040: 3c 9c 92 52 f0 77 00 00  <..R<F0>w..
  0048: 00 00 00 00 00 00 00 00  ........
  0050: 00 00 00 00              ....

Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Howard Chu <howardchu95@gmail.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: Gautam Menghani <gautam@linux.ibm.com>
Tested-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20241108204137.2444151-9-howardchu95@gmail.com
Link: https://lore.kernel.org/r/20250501022809.449767-8-howardchu95@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf record --off-cpu: Disable perf_event's callchain collection

There is a check in evsel.c that does this:

if (evsel__is_offcpu_event(evsel))
evsel->core.attr.sample_type &= OFFCPU_SAMPLE_TYPES;

This along with:

 #define OFFCPU_SAMPLE_TYPES  (PERF_SAMPLE_IDENTIFIER | PERF_SAMPLE_IP | \
      PERF_SAMPLE_TID | PERF_SAMPLE_TIME | \
      PERF_SAMPLE_ID | PERF_SAMPLE_CPU | \
      PERF_SAMPLE_PERIOD | PERF_SAMPLE_CALLCHAIN | \
      PERF_SAMPLE_CGROUP)

will tell perf_event to collect callchain.

We don't need the callchain from perf_event when collecting off-cpu
samples, because it's prev's callchain, not next's callchain.

   (perf_event)     (task_storage) (needed)
   prev             next
   |                  |
   ---sched_switch---->

Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Howard Chu <howardchu95@gmail.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: Gautam Menghani <gautam@linux.ibm.com>
Tested-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20241108204137.2444151-8-howardchu95@gmail.com
Link: https://lore.kernel.org/r/20250501022809.449767-7-howardchu95@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf evsel: Assemble off-cpu samples

Use the data in bpf-output samples, to assemble off-cpu samples.

In evsel__is_offcpu_event(), check if sample_type is PERF_SAMPLE_RAW to
support off-cpu sample data created by an older version of perf.

Testing compatibility on off-cpu samples collected by perf before this patch series:

See below, the sample_type still uses PERF_SAMPLE_CALLCHAIN

$ perf script --header -i ./perf.data.ptn | grep "event : name = offcpu-time"
 # event : name = offcpu-time, , id = { 237917, 237918, 237919, 237920 }, type = 1 (software), size = 136, config = 0xa (PERF_COUNT_SW_BPF_OUTPUT), { sample_period, sample_freq } = 1, sample_type = IP|TID|TIME|CALLCHAIN|CPU|PERIOD|IDENTIFIER, read_format = ID|LOST, disabled = 1, freq = 1, sample_id_all = 1

The output is correct.

  $ perf script -i ./perf.data.ptn | grep offcpu-time
  gmain    2173 [000] 18446744069.414584:  100102015 offcpu-time:
  NetworkManager     901 [000] 18446744069.414584:    5603579 offcpu-time:
  Web Content 1183550 [000] 18446744069.414584:      46278 offcpu-time:
  gnome-control-c 2200559 [000] 18446744069.414584: 11998247014 offcpu-time:
  <SNIP>
  $

And after this patch series:

  $ perf script --header -i ./perf.data.off-cpu-v9 | grep "event : name = offcpu-time"
   # event : name = offcpu-time, , id = { 237959, 237960, 237961, 237962 }, type = 1 (software), size = 136, config = 0xa (PERF_COUNT_SW_BPF_OUTPUT), { sample_period, sample_freq } = 1, sample_type = IP|TID|TIME|CPU|PERIOD|RAW|IDENTIFIER, read_format = ID|LOST, disabled = 1, freq = 1, sample_id_all = 1

  $ ./perf script -i ./perf.data.off-cpu-v9 | grep offcpu-time
  gnome-shell    1875 [001] 4789616.361225:  100097057 offcpu-time:
  gnome-shell    1875 [001] 4789616.461419:  100107463 offcpu-time:
      firefox 2206821 [002] 4789616.475690:  255257245 offcpu-time:
  $

Committer testing:

The command to record those samples:

  root@number:~# perf record --off-cpu -a sleep 1
  [ perf record: Woken up 1 times to write data ]
  [ perf record: Captured and wrote 2.092 MB perf.data (1552 samples) ]
  root@number:~#

Then, before this patch series, the sample_type for the "offcpu-time" event is:

  root@number:~# perf evlist -v | grep offcpu-time
  offcpu-time: type: 1 (PERF_TYPE_SOFTWARE), size: 136, config: 0xa (PERF_COUNT_SW_BPF_OUTPUT), { sample_period, sample_freq }: 1, sample_type: IP|TID|TIME|CALLCHAIN|CPU|PERIOD|IDENTIFIER, read_format: ID|LOST, disabled: 1, freq: 1, sample_id_all: 1
  root@number:~#

And after it, after recording it again:

  root@number:~# perf record --off-cpu -a sleep 1 ; perf evlist -v | grep offcpu-time
  [ perf record: Woken up 1 times to write data ]
  [ perf record: Captured and wrote 2.151 MB perf.data (2843 samples) ]
  offcpu-time: type: 1 (PERF_TYPE_SOFTWARE), size: 136, config: 0xa (PERF_COUNT_SW_BPF_OUTPUT), { sample_period, sample_freq }: 1, sample_type: IP|TID|TIME|CPU|PERIOD|IDENTIFIER, read_format: ID|LOST, disabled: 1, sample_id_all: 1
  root@number:~#

Suggested-by: Namhyung Kim <namhyung@kernel.org>
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Howard Chu <howardchu95@gmail.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: Gautam Menghani <gautam@linux.ibm.com>
Tested-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20241108204137.2444151-7-howardchu95@gmail.com
Link: https://lore.kernel.org/r/20250501022809.449767-6-howardchu95@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf record --off-cpu: Dump off-cpu samples in BPF

Collect tid, period, callchain, and cgroup id and dump them when off-cpu
time threshold is reached.

We don't collect the off-cpu time twice (the delta), it's either in
direct samples, or accumulated samples that are dumped at the end of
perf.data.

Suggested-by: Namhyung Kim <namhyung@kernel.org>
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Howard Chu <howardchu95@gmail.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: Gautam Menghani <gautam@linux.ibm.com>
Tested-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20241108204137.2444151-6-howardchu95@gmail.com
Link: https://lore.kernel.org/r/20250501022809.449767-5-howardchu95@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf record --off-cpu: Preparation of off-cpu BPF program

Set the perf_event map in BPF for dumping off-cpu samples, and set the
offcpu_thresh to specify the threshold.

Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Howard Chu <howardchu95@gmail.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: Gautam Menghani <gautam@linux.ibm.com>
Tested-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20241108204137.2444151-5-howardchu95@gmail.com
Link: https://lore.kernel.org/r/20250501022809.449767-4-howardchu95@gmail.com
[ Added some missing iteration variables to off_cpu_config() and fixed up
  a manually edited patch hunk line boundary line ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf record --off-cpu: Parse off-cpu event

Parse the off-cpu event using parse_event(), as bpf-output.

Call evlist__enable_evsel() on off-cpu event. This fixes the inability
to collect direct off-cpu samples on a workload, as reported by Arnaldo
Carvalho de Melo <acme@redhat.com>.

The reason being, workload sets enable_on_exec instead of calling
evlist__enable(), but off-cpu event does not attach to an executable and
execve won't be called, so the fds from perf_event_open() are not
enabled.

no-inherit should be set to 1, here's the reason:

We update the BPF perf_event map for direct off-cpu sample dumping (in
following patches), it executes as follows:

bpf_map_update_value()
 bpf_fd_array_map_update_elem()
  perf_event_fd_array_get_ptr()
   perf_event_read_local()

In perf_event_read_local(), there is:

int perf_event_read_local(struct perf_event *event, u64 *value,
  u64 *enabled, u64 *running)
{
...
/*
 * It must not be an event with inherit set, we cannot read
 * all child counters from atomic context.
 */
if (event->attr.inherit) {
ret = -EOPNOTSUPP;
goto out;
}

Which means no-inherit has to be true for updating the BPF perf_event
map.

Moreover, for bpf-output events, we primarily want a system-wide event
instead of a per-task event.

The reason is that in BPF's bpf_perf_event_output(), BPF uses the CPU
index to retrieve the perf_event file descriptor it outputs to.

Making a bpf-output event system-wide naturally satisfies this
requirement by mapping CPU appropriately.

Suggested-by: Namhyung Kim <namhyung@kernel.org>
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Howard Chu <howardchu95@gmail.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: Gautam Menghani <gautam@linux.ibm.com>
Tested-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20241108204137.2444151-4-howardchu95@gmail.com
Link: https://lore.kernel.org/r/20250501022809.449767-3-howardchu95@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf evsel: Expose evsel__is_offcpu_event() for future use

Expose evsel__is_offcpu_event() so it can be used in off_cpu_config(),
evsel__parse_sample() and 'perf script'.

Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Howard Chu <howardchu95@gmail.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: Gautam Menghani <gautam@linux.ibm.com>
Tested-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20241108204137.2444151-3-howardchu95@gmail.com
Link: https://lore.kernel.org/r/20250501022809.449767-2-howardchu95@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf symbol-minimal: Fix double free in filename__read_build_id

Running the "perf script task-analyzer tests" with address sanitizer
showed a double free:
```
FAIL: "test_csv_extended_times" Error message: "Failed to find required string:'Out-Out;'."
=================================================================
==19190==ERROR: AddressSanitizer: attempting double-free on 0x50b000017b10 in thread T0:
    #0 0x55da9601c78a in free (perf+0x26078a) (BuildId: e7ef50e08970f017a96fde6101c5e2491acc674a)
    #1 0x55da96640c63 in filename__read_build_id tools/perf/util/symbol-minimal.c:221:2

0x50b000017b10 is located 0 bytes inside of 112-byte region [0x50b000017b10,0x50b000017b80)
freed by thread T0 here:
    #0 0x55da9601ce40 in realloc (perf+0x260e40) (BuildId: e7ef50e08970f017a96fde6101c5e2491acc674a)
    #1 0x55da96640ad6 in filename__read_build_id tools/perf/util/symbol-minimal.c:204:10

previously allocated by thread T0 here:
    #0 0x55da9601ca23 in malloc (perf+0x260a23) (BuildId: e7ef50e08970f017a96fde6101c5e2491acc674a)
    #1 0x55da966407e7 in filename__read_build_id tools/perf/util/symbol-minimal.c:181:9

SUMMARY: AddressSanitizer: double-free (perf+0x26078a) (BuildId: e7ef50e08970f017a96fde6101c5e2491acc674a) in free
==19190==ABORTING
FAIL: "invocation of perf script report task-analyzer --csv-summary csvsummary --summary-extended command failed" Error message: ""
FAIL: "test_csvsummary_extended" Error message: "Failed to find required string:'Out-Out;'."
---- end(-1) ----
132: perf script task-analyzer tests                                 : FAILED!
```

The buf_size if always set to phdr->p_filesz, but that may be 0
causing a free and realloc to return NULL. This is treated in
filename__read_build_id like a failure and the buffer is freed again.

To avoid this problem only grow buf, meaning the buf_size will never
be 0. This also reduces the number of memory (re)allocations.

Fixes: b691f64360ecec49 ("perf symbols: Implement poor man's ELF parser")
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20250501070003.22251-1-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf mem: Add 'dtlb' output field

This is a breakdown of perf_mem_data_src.mem_dtlb values.  It assumes
PMU drivers would set PERF_MEM_TLB_HIT bit with an appropriate level.

And having PERF_MEM_TLB_MISS means that it failed to find one in any
levels of TLB.  For now, it doesn't use PERF_MEM_TLB_{WK,OS} bits.

Also it seems Intel machines don't distinguish L1 or L2 precisely.  So I
added ANY_HIT (printed as "L?-Hit") to handle the case.

  $ perf mem report -F overhead,dtlb,dso --stdio
  ...
  #           --- D-TLB ----
  # Overhead   L?-Hit   Miss  Shared Object
  # ........  ..............  .................
  #
      67.03%    99.5%   0.5%  [unknown]
      31.23%    99.2%   0.8%  [kernel.kallsyms]
       1.08%    97.8%   2.2%  [i915]
       0.36%   100.0%   0.0%  [JIT] tid 6853
       0.12%   100.0%   0.0%  [drm]
       0.05%   100.0%   0.0%  [drm_kms_helper]
       0.05%   100.0%   0.0%  [ext4]
       0.02%   100.0%   0.0%  [aesni_intel]
       0.02%   100.0%   0.0%  [crc32c_intel]
       0.02%   100.0%   0.0%  [dm_crypt]
       ...

Committer testing:

  # perf report --header | grep cpudesc
  # cpudesc : AMD Ryzen 9 9950X3D 16-Core Processor
  # perf mem report -F overhead,dtlb,dso --stdio | head -20
  # To display the perf.data header info, please use --header/--header-only options.
  #
  #
  # Total Lost Samples: 0
  #
  # Samples: 2K of event 'cycles:P'
  # Total weight : 2637
  # Sort order   : local_weight,mem,sym,dso,symbol_daddr,dso_daddr,snoop,tlb,locked,blocked,local_ins_lat,local_p_stage_cyc
  #
  #           ---------- D-TLB -----------
  # Overhead   L1-Hit L2-Hit   Miss  Other  Shared Object
  # ........  ............................  .................................
  #
      77.47%    18.4%   0.1%   0.6%  80.9%  [kernel.kallsyms]
       5.61%    36.5%   0.7%   1.4%  61.5%  libxul.so
       2.77%    39.7%   0.0%  12.3%  47.9%  libc.so.6
       2.01%    34.0%   1.9%   1.9%  62.3%  libglib-2.0.so.0.8400.1
       1.93%    31.4%   2.0%   2.0%  64.7%  [amdgpu]
       1.63%    48.8%   0.0%   0.0%  51.2%  [JIT] tid 60168
       1.14%     3.3%   0.0%   0.0%  96.7%  [vdso]
  #

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Link: https://lore.kernel.org/r/20250430205548.789750-12-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf mem: Add 'snoop' output field

This is a breakdown of perf_mem_data_src.mem_snoop values.  For now, it
doesn't use mem_snoopx values like FWD and PEER.

  $ perf mem report -F overhead,snoop,comm --stdio
  ...
  #           ---------- Snoop -----------
  # Overhead      Hit   HitM   Miss  Other  Command
  # ........  ............................  ...............
  #
      34.24%     0.6%   0.0%   0.0%  99.4%  gnome-shell
      12.02%     1.0%   0.0%   0.0%  99.0%  chrome
       9.32%     1.0%   0.0%   0.3%  98.7%  Isolated Web Co
       6.85%     1.0%   0.3%   0.0%  98.6%  swapper
       6.30%     0.8%   0.8%   0.0%  98.5%  Xorg
       3.02%     2.4%   0.0%   0.0%  97.6%  VizCompositorTh
       2.35%     0.0%   0.0%   0.0% 100.0%  firefox-esr
       2.04%     0.0%   0.0%   0.0% 100.0%  JS Helper
       1.51%     3.2%   0.0%   0.0%  96.8%  threaded-ml
       1.44%     0.0%   0.0%   0.0% 100.0%  AudioIP~allback
       ...

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Link: https://lore.kernel.org/r/20250430205548.789750-11-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf mem: Add 'cache' and 'memory' output fields

This is a breakdown of perf_mem_data_src.mem_lvl_num.  But it's also
divided into two parts because the combination is bigger than 8.

Since there are many entries for different cache levels, 'cache' field
focuses on them.  I generalized buffers like LFB, MAB and MHB to L1-buf
and L2-buf.

The rest goes to 'memory' field which can be RAM, CXL, PMEM, IO, etc.

  $ perf mem report -F cache,mem,dso --stdio
  ...
  #
  # -------------- Cache --------------  --- Memory ---
  #      L1     L2     L3 L1-buf  Other      RAM  Other  Shared Object
  # ...................................  ..............  ....................................
  #
      53.9%   3.6%  16.2%  21.6%   4.8%     4.8%  95.2%  [kernel.kallsyms]
      64.7%   1.7%   3.5%  17.4%  12.8%    12.8%  87.2%  chrome (deleted)
      78.3%   2.8%   0.0%   1.0%  17.9%    17.9%  82.1%  libc.so.6
      39.6%   1.5%   0.0%   5.7%  53.2%    53.2%  46.8%  libxul.so
      26.2%   0.0%   0.0%   0.0%  73.8%    73.8%  26.2%  [unknown]
      85.5%   0.0%   0.0%  14.5%   0.0%     0.0% 100.0%  libspa-audioconvert.so
      66.3%   4.4%   0.0%  29.4%   0.0%     0.0% 100.0%  libglib-2.0.so.0.8200.1 (deleted)
       1.9%   0.0%   0.0%   0.0%  98.1%    98.1%   1.9%  libmutter-cogl-15.so.0.0.0 (deleted)
      10.6%   0.0%   0.0%  89.4%   0.0%     0.0% 100.0%  libpulsecommon-16.1.so
       0.0%   0.0%   0.0% 100.0%   0.0%     0.0% 100.0%  libfreeblpriv3.so (deleted)
       ...

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Link: https://lore.kernel.org/r/20250430205548.789750-10-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hist: Hide unused mem stat columns

Some mem_stat types don't use all 8 columns.  And there are cases only
samples in certain kinds of mem_stat types are available only.  For that
case hide columns which has no samples.

The new output for the previous data would be:

  $ perf mem report -F overhead,op,comm --stdio
  ...
  #           ------ Mem Op -------
  # Overhead     Load  Store  Other  Command
  # ........  .....................  ...............
  #
      44.85%    21.1%  30.7%  48.3%  swapper
      26.82%    98.8%   0.3%   0.9%  netsli-prober
       7.19%    51.7%  13.7%  34.6%  perf
       5.81%    89.7%   2.2%   8.1%  qemu-system-ppc
       4.77%   100.0%   0.0%   0.0%  notifications_c
       1.77%    95.9%   1.2%   3.0%  MemoryReleaser
       0.77%    71.6%   4.1%  24.3%  DefaultEventMan
       0.19%    66.7%  22.2%  11.1%  gnome-shell
       ...

On Intel machines, the event is only for loads or stores so it'll have
only one column:

  #            Mem Op
  # Overhead     Load  Command
  # ........  .......  ...............
  #
      20.55%   100.0%  swapper
      17.13%   100.0%  chrome
       9.02%   100.0%  data-loop.0
       6.26%   100.0%  pipewire-pulse
       5.63%   100.0%  threaded-ml
       5.47%   100.0%  GraphRunner
       5.37%   100.0%  AudioIP~allback
       5.30%   100.0%  Chrome_ChildIOT
       3.17%   100.0%  Isolated Web Co
       ...

Committer testing:

  # grep "model name" -m1 /proc/cpuinfo
  model name : AMD Ryzen 9 9950X3D 16-Core Processo
  # perf mem report -F overhead,op,comm --stdio
  # Total Lost Samples: 0
  #
  # Samples: 2K of event 'cycles:P'
  # Total weight : 2637
  # Sort order   : local_weight,mem,sym,dso,symbol_daddr,dso_daddr,snoop,tlb,locked,blocked,local_ins_lat,local_p_stage_cyc
  #
  #           ------ Mem Op -------
  # Overhead     Load  Store  Other  Command
  # ........  .....................  ...............
  #
      61.02%    14.4%  25.5%  60.1%  swapper
       5.61%    26.4%  13.5%  60.1%  Isolated Web Co
       5.50%    21.4%  29.7%  49.0%  perf
       4.74%    27.2%  15.2%  57.6%  gnome-shell
       4.63%    33.6%  11.5%  54.9%  mdns_service
       4.29%    28.3%  12.4%  59.3%  ptyxis
       2.16%    24.6%  19.3%  56.1%  DOM Worker
       0.99%    23.1%  34.6%  42.3%  firefox
       0.72%    26.3%  15.8%  57.9%  IPC I/O Parent
       0.61%    12.5%  12.5%  75.0%  kworker/u130:20
       0.61%    37.5%  18.8%  43.8%  podman
       0.57%    33.3%   6.7%  60.0%  Timer
       0.53%    14.3%   7.1%  78.6%  KMS thread
       0.49%    30.8%   7.7%  61.5%  kworker/u130:3-
       0.46%    41.7%  33.3%  25.0%  IPDL Background

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Link: https://lore.kernel.org/r/20250430205548.789750-9-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf mem: Add 'op' output field

This is an actual example of the he_mem_stat based sample breakdown.  It
uses 'mem_op' field of union perf_mem_data_src which means memory
operations.

It'd have basically 'load' or 'store' which can be useful if PMU doesn't
have separate events for them like IBS or SPE.  In addition, there's an
entry in case load and store happen at the same time.  Also adds entries
for prefetching and execution.

  $ perf mem report -F +op -s comm --stdio
  # To display the perf.data header info, please use --header/--header-only options.
  #
  #
  # Total Lost Samples: 0
  #
  # Samples: 4K of event 'ibs_op//'
  # Total weight : 9559
  # Sort order   : comm
  #
  #                   --------------------- Mem Op ----------------------
  # Overhead  Samples   Load  Store Ld+St Pfetch  Exec  Other   N/A   N/A  Command
  # ........  ....... ...................................................  ...............
  #
      44.85%     4077  21.1%  30.7%  0.0%   0.0%  0.0%  48.3%  0.0%  0.0%  swapper
      26.82%       45  98.8%   0.3%  0.0%   0.0%  0.0%   0.9%  0.0%  0.0%  netsli-prober
       7.19%      442  51.7%  13.7%  0.0%   0.0%  0.0%  34.6%  0.0%  0.0%  perf
       5.81%       75  89.7%   2.2%  0.0%   0.0%  0.0%   8.1%  0.0%  0.0%  qemu-system-ppc
       4.77%        1 100.0%   0.0%  0.0%   0.0%  0.0%   0.0%  0.0%  0.0%  notifications_c
       1.77%       10  95.9%   1.2%  0.0%   0.0%  0.0%   3.0%  0.0%  0.0%  MemoryReleaser
       0.77%       32  71.6%   4.1%  0.0%   0.0%  0.0%  24.3%  0.0%  0.0%  DefaultEventMan
       0.19%       10  66.7%  22.2%  0.0%   0.0%  0.0%  11.1%  0.0%  0.0%  gnome-shell

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Link: https://lore.kernel.org/r/20250430205548.789750-8-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hist: Implement output fields for mem stats

This is a preparation for later changes to support mem_stat output.  The
new fields will need two lines for the header - the first line will show
type of mem stat and the second line will show the name of each item
which is returned by mem_stat_name().

Each element in the mem_stat array will be printed in percentage for the
hist_entry and their sum would be 100%.

Add new output field dimension only for SORT_MODE__MEM using mem_stat.

To handle possible name conflict with existing sort keys, move the order
of checking output field dimensions after the sort dimensions when it
looks for sort keys.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Link: https://lore.kernel.org/r/20250430205548.789750-7-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hist: Basic support for mem_stat accounting

Add a logic to account he->mem_stat based on mem_stat_type in hists.

Each mem_stat entry will have different meaning based on the type so the
index in the array is calculated at runtime using the corresponding
value in the sample.data_src.

Still hists has no mem_stat_types yet so this code won't work for now.

Later hists->mem_stat_types will be allocated based on what users want
in the output actually.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Link: https://lore.kernel.org/r/20250430205548.789750-6-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hist: Add struct he_mem_stat

The 'struct he_mem_stat' is to save detailed information about memory
instruction.  It'll be used to show breakdown of various data from
PERF_SAMPLE_DATA_SRC.  Note that this structure is generic and the
contents will be different depending on actual data it'll use later.

The information about the actual data will be saved in 'struct hists'
and its length is in nr_mem_stats.  This commit just adds ground works
and does nothing since hists->nr_mem_stats is 0 for now.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Link: https://lore.kernel.org/r/20250430205548.789750-5-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hist: Support multi-line header

This is a preparation to support multi-line headers in 'perf mem report'.

Normal sort keys and output fields that don't have contents for multi-
line will print the header string at the last line only.

As we don't use multi-line headers normally, it should not have any
changes in the output.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Link: https://lore.kernel.org/r/20250430205548.789750-4-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf record: Add --sample-mem-info option

There's no way to enable PERF_SAMPLE_DATA_SRC without PERF_SAMPLE_ADDR
which brings a lot of overhead due to the number of MMAP[2] records.

Let's add a new option to enable this information separately.

Committer testing:

  # perf record -a --sample-mem-info
  ^C[ perf record: Woken up 1 times to write data ]
  [ perf record: Captured and wrote 1.815 MB perf.data (2637 samples) ]
  #
  # perf evlist -v
  cycles:P: type: 0 (PERF_TYPE_HARDWARE), size: 136, config: 0 (PERF_COUNT_HW_CPU_CYCLES), { sample_period, sample_freq }: 4000, sample_type: IP|TID|TIME|CPU|PERIOD|IDENTIFIER|DATA_SRC, read_format: ID|LOST, disabled: 1, freq: 1, precise_ip: 2, sample_id_all: 1
  dummy:u: type: 1 (PERF_TYPE_SOFTWARE), size: 136, config: 0x9 (PERF_COUNT_SW_DUMMY), { sample_period, sample_freq }: 1, sample_type: IP|TID|TIME|CPU|IDENTIFIER|DATA_SRC, read_format: ID|LOST, exclude_kernel: 1, exclude_hv: 1, mmap: 1, comm: 1, task: 1, sample_id_all: 1, exclude_guest: 1, mmap2: 1, comm_exec: 1, ksymbol: 1, bpf_event: 1
  #
  # perf report -D |& grep -w PERF_RECORD_SAMPLE -A3 -m1
  0 44675164447282 0x1a7590 [0x40]: PERF_RECORD_SAMPLE(IP, 0x4001): 107299/107299: 0xffffffffac4a5e11 period: 144 addr: 0
   . data_src: 0x229080142
   ... thread: perf:107299
   ...... dso: /lib/modules/6.15.0-rc4+/build/vmlinux
  #

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Link: https://lore.kernel.org/r/20250430205548.789750-3-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hist: Remove output field from sort-list properly

When it removes an output format for cancelled children or latency, it
should delete itself from the sort list as well.  Otherwise assertion
in fmt_free() will fire.

  $ perf report -H --stdio
  perf: ui/hist.c:603: fmt_free: Assertion `!(!list_empty(&fmt->sort_list))' failed.
  Aborted (core dumped)

Also convert to perf_hpp__column_unregister() for the same open codes.

Committer notes:

Before this patch:

  # perf test hierarchy
   83: perf report --hierarchy                                         : FAILED!
  # perf test -v hierarchy
  --- start ---
  test child forked, pid 102242
  perf report --hierarchy
  Linux
  [ perf record: Woken up 1 times to write data ]
  [ perf record: Captured and wrote 0.025 MB /tmp/perf-test-report.HX0N85TlPq/perf-report-hierarchy-perf.data (6 samples) ]
  perf: ui/hist.c:603: fmt_free: Assertion `!(!list_empty(&fmt->sort_list))' failed.
  /home/acme/libexec/perf-core/tests/shell/perf-report-hierarchy.sh: line 34: 102250 Aborted                 (core dumped) perf report --hierarchy > /dev/null
  --- Cleaning up ---
  ---- end(-1) ----
   83: perf report --hierarchy                                         : FAILED!
  #

After:

  # perf test hierarchy
   83: perf report --hierarchy                                         : Ok
  #

Fixes: dbd11b6bdab12f60 ("perf hist: Remove formats in hierarchy when cancel children")
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20250430180321.736939-1-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf test perf-report-hierarchy: Add new test

Super simple test to check that at least we're not segfaulting when
trying to use 'perf report --hierarchy', more subtests should be added
to make sure the output is the expected one.

This is being merged right before a fix for that that this test detects:

  # perf test hierarchy
   83: perf report --hierarchy                                         : FAILED!
  # perf test -v hierarchy
  --- start ---
  test child forked, pid 102242
  perf report --hierarchy
  Linux
  [ perf record: Woken up 1 times to write data ]
  [ perf record: Captured and wrote 0.025 MB /tmp/perf-test-report.HX0N85TlPq/perf-report-hierarchy-perf.data (6 samples) ]
  perf: ui/hist.c:603: fmt_free: Assertion `!(!list_empty(&fmt->sort_list))' failed.
  /home/acme/libexec/perf-core/tests/shell/perf-report-hierarchy.sh: line 34: 102250 Aborted                 (core dumped) perf report --hierarchy > /dev/null
  --- Cleaning up ---
  ---- end(-1) ----
   83: perf report --hierarchy                                         : FAILED!
  #

Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Link: https://lore.kernel.org/lkml/20250430180321.736939-1-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf test amd ibs: Add sample period unit test

IBS Fetch and IBS Op PMUs has various constraints on supported sample
periods. Add perf unit tests to test those.

Running it in parallel with other tests causes intermittent failures.
Mark it exclusive to force it to run sequentially. Sample output on a
Zen5 machine:

Without kernel fixes:

  $ sudo ./perf test -vv 112
  112: AMD IBS sample period:
  --- start ---
  test child forked, pid 8774
  Using CPUID AuthenticAMD-26-2-1

  IBS config tests:
  -----------------
  Fetch PMU tests:
  0xffff            : Ok   (nr samples: 1078)
  0x1000            : Ok   (nr samples: 17030)
  0xff              : Ok   (nr samples: 41068)
  0x1               : Ok   (nr samples: 40543)
  0x0               : Ok
  0x10000           : Ok
  Op PMU tests:
  0x0               : Ok
  0x1               : Fail
  0x8               : Fail
  0x9               : Ok   (nr samples: 40543)
  0xf               : Ok   (nr samples: 40543)
  0x1000            : Ok   (nr samples: 18736)
  0xffff            : Ok   (nr samples: 1168)
  0x10000           : Ok
  0x100000          : Fail (nr samples: 14)
  0xf00000          : Fail (nr samples: 1)
  0xf0ffff          : Fail (nr samples: 1)
  0x1f0ffff         : Fail (nr samples: 1)
  0x7f0ffff         : Fail (nr samples: 0)
  0x8f0ffff         : Ok
  0x17f0ffff        : Ok

  IBS sample period constraint tests:
  -----------------------------------
  Fetch PMU test:
  freq 0, sample_freq         0: Ok
  freq 0, sample_freq         1: Fail
  freq 0, sample_freq        15: Fail
  freq 0, sample_freq        16: Ok   (nr samples: 1604)
  freq 0, sample_freq        17: Ok   (nr samples: 1604)
  freq 0, sample_freq       143: Ok   (nr samples: 1604)
  freq 0, sample_freq       144: Ok   (nr samples: 1604)
  freq 0, sample_freq       145: Ok   (nr samples: 1604)
  freq 0, sample_freq      1234: Ok   (nr samples: 1566)
  freq 0, sample_freq      4103: Ok   (nr samples: 1119)
  freq 0, sample_freq     65520: Ok   (nr samples: 2264)
  freq 0, sample_freq     65535: Ok   (nr samples: 2263)
  freq 0, sample_freq     65552: Ok   (nr samples: 1166)
  freq 0, sample_freq   8388607: Ok   (nr samples: 268)
  freq 0, sample_freq 268435455: Ok   (nr samples: 8)
  freq 1, sample_freq         0: Ok
  freq 1, sample_freq         1: Ok   (nr samples: 4)
  freq 1, sample_freq        15: Ok   (nr samples: 4)
  freq 1, sample_freq        16: Ok   (nr samples: 4)
  freq 1, sample_freq        17: Ok   (nr samples: 4)
  freq 1, sample_freq       143: Ok   (nr samples: 5)
  freq 1, sample_freq       144: Ok   (nr samples: 5)
  freq 1, sample_freq       145: Ok   (nr samples: 5)
  freq 1, sample_freq      1234: Ok   (nr samples: 7)
  freq 1, sample_freq      4103: Ok   (nr samples: 35)
  freq 1, sample_freq     65520: Ok   (nr samples: 642)
  freq 1, sample_freq     65535: Ok   (nr samples: 636)
  freq 1, sample_freq     65552: Ok   (nr samples: 651)
  freq 1, sample_freq   8388607: Ok
  Op PMU test:
  freq 0, sample_freq         0: Ok
  freq 0, sample_freq         1: Fail
  freq 0, sample_freq        15: Fail
  freq 0, sample_freq        16: Fail
  freq 0, sample_freq        17: Fail
  freq 0, sample_freq       143: Fail
  freq 0, sample_freq       144: Ok   (nr samples: 1604)
  freq 0, sample_freq       145: Ok   (nr samples: 1604)
  freq 0, sample_freq      1234: Ok   (nr samples: 1604)
  freq 0, sample_freq      4103: Ok   (nr samples: 1604)
  freq 0, sample_freq     65520: Ok   (nr samples: 2227)
  freq 0, sample_freq     65535: Ok   (nr samples: 2296)
  freq 0, sample_freq     65552: Ok   (nr samples: 2213)
  freq 0, sample_freq   8388607: Ok   (nr samples: 250)
  freq 0, sample_freq 268435455: Ok   (nr samples: 8)
  freq 1, sample_freq         0: Ok
  freq 1, sample_freq         1: Fail (nr samples: 4)
  freq 1, sample_freq        15: Fail (nr samples: 4)
  freq 1, sample_freq        16: Fail (nr samples: 4)
  freq 1, sample_freq        17: Fail (nr samples: 4)
  freq 1, sample_freq       143: Fail (nr samples: 5)
  freq 1, sample_freq       144: Fail (nr samples: 5)
  freq 1, sample_freq       145: Fail (nr samples: 5)
  freq 1, sample_freq      1234: Fail (nr samples: 8)
  freq 1, sample_freq      4103: Fail (nr samples: 33)
  freq 1, sample_freq     65520: Fail (nr samples: 546)
  freq 1, sample_freq     65535: Fail (nr samples: 544)
  freq 1, sample_freq     65552: Fail (nr samples: 555)
  freq 1, sample_freq   8388607: Ok

  IBS ioctl() tests:
  ------------------
  Fetch PMU tests
  ioctl(period = 0x0      ): Ok
  ioctl(period = 0x1      ): Fail
  ioctl(period = 0xf      ): Fail
  ioctl(period = 0x10     ): Ok
  ioctl(period = 0x11     ): Fail
  ioctl(period = 0x1f     ): Fail
  ioctl(period = 0x20     ): Ok
  ioctl(period = 0x80     ): Ok
  ioctl(period = 0x8f     ): Fail
  ioctl(period = 0x90     ): Ok
  ioctl(period = 0x91     ): Fail
  ioctl(period = 0x100    ): Ok
  ioctl(period = 0xfff0   ): Ok
  ioctl(period = 0xffff   ): Fail
  ioctl(period = 0x10000  ): Ok
  ioctl(period = 0x1fff0  ): Ok
  ioctl(period = 0x1fff5  ): Fail
  ioctl(freq   = 0x0      ): Ok
  ioctl(freq   = 0x1      ): Ok
  ioctl(freq   = 0xf      ): Ok
  ioctl(freq   = 0x10     ): Ok
  ioctl(freq   = 0x11     ): Ok
  ioctl(freq   = 0x1f     ): Ok
  ioctl(freq   = 0x20     ): Ok
  ioctl(freq   = 0x80     ): Ok
  ioctl(freq   = 0x8f     ): Ok
  ioctl(freq   = 0x90     ): Ok
  ioctl(freq   = 0x91     ): Ok
  ioctl(freq   = 0x100    ): Ok
  Op PMU tests
  ioctl(period = 0x0      ): Ok
  ioctl(period = 0x1      ): Fail
  ioctl(period = 0xf      ): Fail
  ioctl(period = 0x10     ): Fail
  ioctl(period = 0x11     ): Fail
  ioctl(period = 0x1f     ): Fail
  ioctl(period = 0x20     ): Fail
  ioctl(period = 0x80     ): Fail
  ioctl(period = 0x8f     ): Fail
  ioctl(period = 0x90     ): Ok
  ioctl(period = 0x91     ): Fail
  ioctl(period = 0x100    ): Ok
  ioctl(period = 0xfff0   ): Ok
  ioctl(period = 0xffff   ): Fail
  ioctl(period = 0x10000  ): Ok
  ioctl(period = 0x1fff0  ): Ok
  ioctl(period = 0x1fff5  ): Fail
  ioctl(freq   = 0x0      ): Ok
  ioctl(freq   = 0x1      ): Ok
  ioctl(freq   = 0xf      ): Ok
  ioctl(freq   = 0x10     ): Ok
  ioctl(freq   = 0x11     ): Ok
  ioctl(freq   = 0x1f     ): Ok
  ioctl(freq   = 0x20     ): Ok
  ioctl(freq   = 0x80     ): Ok
  ioctl(freq   = 0x8f     ): Ok
  ioctl(freq   = 0x90     ): Ok
  ioctl(freq   = 0x91     ): Ok
  ioctl(freq   = 0x100    ): Ok

  IBS freq (negative) tests:
  --------------------------
  freq 1, sample_freq 200000: Fail

  IBS L3MissOnly test: (takes a while)
  --------------------
  Fetch L3MissOnly: Fail (nr_samples: 1213)
  Op L3MissOnly:    Ok   (nr_samples: 1193)
  ---- end(-1) ----
  112: AMD IBS sample period                                           : FAILED!

With kernel fixes:

  $ sudo ./perf test -vv 112
  112: AMD IBS sample period:
  --- start ---
  test child forked, pid 6939
  Using CPUID AuthenticAMD-26-2-1

  IBS config tests:
  -----------------
  Fetch PMU tests:
  0xffff            : Ok   (nr samples: 969)
  0x1000            : Ok   (nr samples: 15540)
  0xff              : Ok   (nr samples: 40555)
  0x1               : Ok   (nr samples: 40543)
  0x0               : Ok
  0x10000           : Ok
  Op PMU tests:
  0x0               : Ok
  0x1               : Ok
  0x8               : Ok
  0x9               : Ok   (nr samples: 40543)
  0xf               : Ok   (nr samples: 40543)
  0x1000            : Ok   (nr samples: 19156)
  0xffff            : Ok   (nr samples: 1169)
  0x10000           : Ok
  0x100000          : Ok   (nr samples: 1151)
  0xf00000          : Ok   (nr samples: 76)
  0xf0ffff          : Ok   (nr samples: 73)
  0x1f0ffff         : Ok   (nr samples: 33)
  0x7f0ffff         : Ok   (nr samples: 10)
  0x8f0ffff         : Ok
  0x17f0ffff        : Ok

  IBS sample period constraint tests:
  -----------------------------------
  Fetch PMU test:
  freq 0, sample_freq         0: Ok
  freq 0, sample_freq         1: Ok
  freq 0, sample_freq        15: Ok
  freq 0, sample_freq        16: Ok   (nr samples: 1203)
  freq 0, sample_freq        17: Ok   (nr samples: 1604)
  freq 0, sample_freq       143: Ok   (nr samples: 1604)
  freq 0, sample_freq       144: Ok   (nr samples: 1604)
  freq 0, sample_freq       145: Ok   (nr samples: 1604)
  freq 0, sample_freq      1234: Ok   (nr samples: 1604)
  freq 0, sample_freq      4103: Ok   (nr samples: 1343)
  freq 0, sample_freq     65520: Ok   (nr samples: 2254)
  freq 0, sample_freq     65535: Ok   (nr samples: 2136)
  freq 0, sample_freq     65552: Ok   (nr samples: 1158)
  freq 0, sample_freq   8388607: Ok   (nr samples: 257)
  freq 0, sample_freq 268435455: Ok   (nr samples: 8)
  freq 1, sample_freq         0: Ok
  freq 1, sample_freq         1: Ok   (nr samples: 4)
  freq 1, sample_freq        15: Ok   (nr samples: 4)
  freq 1, sample_freq        16: Ok   (nr samples: 4)
  freq 1, sample_freq        17: Ok   (nr samples: 4)
  freq 1, sample_freq       143: Ok   (nr samples: 5)
  freq 1, sample_freq       144: Ok   (nr samples: 5)
  freq 1, sample_freq       145: Ok   (nr samples: 5)
  freq 1, sample_freq      1234: Ok   (nr samples: 8)
  freq 1, sample_freq      4103: Ok   (nr samples: 34)
  freq 1, sample_freq     65520: Ok   (nr samples: 458)
  freq 1, sample_freq     65535: Ok   (nr samples: 628)
  freq 1, sample_freq     65552: Ok   (nr samples: 396)
  freq 1, sample_freq   8388607: Ok
  Op PMU test:
  freq 0, sample_freq         0: Ok
  freq 0, sample_freq         1: Ok
  freq 0, sample_freq        15: Ok
  freq 0, sample_freq        16: Ok
  freq 0, sample_freq        17: Ok
  freq 0, sample_freq       143: Ok
  freq 0, sample_freq       144: Ok   (nr samples: 1604)
  freq 0, sample_freq       145: Ok   (nr samples: 1604)
  freq 0, sample_freq      1234: Ok   (nr samples: 1604)
  freq 0, sample_freq      4103: Ok   (nr samples: 1604)
  freq 0, sample_freq     65520: Ok   (nr samples: 2250)
  freq 0, sample_freq     65535: Ok   (nr samples: 2158)
  freq 0, sample_freq     65552: Ok   (nr samples: 2296)
  freq 0, sample_freq   8388607: Ok   (nr samples: 243)
  freq 0, sample_freq 268435455: Ok   (nr samples: 6)
  freq 1, sample_freq         0: Ok
  freq 1, sample_freq         1: Ok   (nr samples: 4)
  freq 1, sample_freq        15: Ok   (nr samples: 4)
  freq 1, sample_freq        16: Ok   (nr samples: 4)
  freq 1, sample_freq        17: Ok   (nr samples: 4)
  freq 1, sample_freq       143: Ok   (nr samples: 4)
  freq 1, sample_freq       144: Ok   (nr samples: 5)
  freq 1, sample_freq       145: Ok   (nr samples: 4)
  freq 1, sample_freq      1234: Ok   (nr samples: 6)
  freq 1, sample_freq      4103: Ok   (nr samples: 27)
  freq 1, sample_freq     65520: Ok   (nr samples: 542)
  freq 1, sample_freq     65535: Ok   (nr samples: 550)
  freq 1, sample_freq     65552: Ok   (nr samples: 552)
  freq 1, sample_freq   8388607: Ok

  IBS ioctl() tests:
  ------------------
  Fetch PMU tests
  ioctl(period = 0x0      ): Ok
  ioctl(period = 0x1      ): Ok
  ioctl(period = 0xf      ): Ok
  ioctl(period = 0x10     ): Ok
  ioctl(period = 0x11     ): Ok
  ioctl(period = 0x1f     ): Ok
  ioctl(period = 0x20     ): Ok
  ioctl(period = 0x80     ): Ok
  ioctl(period = 0x8f     ): Ok
  ioctl(period = 0x90     ): Ok
  ioctl(period = 0x91     ): Ok
  ioctl(period = 0x100    ): Ok
  ioctl(period = 0xfff0   ): Ok
  ioctl(period = 0xffff   ): Ok
  ioctl(period = 0x10000  ): Ok
  ioctl(period = 0x1fff0  ): Ok
  ioctl(period = 0x1fff5  ): Ok
  ioctl(freq   = 0x0      ): Ok
  ioctl(freq   = 0x1      ): Ok
  ioctl(freq   = 0xf      ): Ok
  ioctl(freq   = 0x10     ): Ok
  ioctl(freq   = 0x11     ): Ok
  ioctl(freq   = 0x1f     ): Ok
  ioctl(freq   = 0x20     ): Ok
  ioctl(freq   = 0x80     ): Ok
  ioctl(freq   = 0x8f     ): Ok
  ioctl(freq   = 0x90     ): Ok
  ioctl(freq   = 0x91     ): Ok
  ioctl(freq   = 0x100    ): Ok
  Op PMU tests
  ioctl(period = 0x0      ): Ok
  ioctl(period = 0x1      ): Ok
  ioctl(period = 0xf      ): Ok
  ioctl(period = 0x10     ): Ok
  ioctl(period = 0x11     ): Ok
  ioctl(period = 0x1f     ): Ok
  ioctl(period = 0x20     ): Ok
  ioctl(period = 0x80     ): Ok
  ioctl(period = 0x8f     ): Ok
  ioctl(period = 0x90     ): Ok
  ioctl(period = 0x91     ): Ok
  ioctl(period = 0x100    ): Ok
  ioctl(period = 0xfff0   ): Ok
  ioctl(period = 0xffff   ): Ok
  ioctl(period = 0x10000  ): Ok
  ioctl(period = 0x1fff0  ): Ok
  ioctl(period = 0x1fff5  ): Ok
  ioctl(freq   = 0x0      ): Ok
  ioctl(freq   = 0x1      ): Ok
  ioctl(freq   = 0xf      ): Ok
  ioctl(freq   = 0x10     ): Ok
  ioctl(freq   = 0x11     ): Ok
  ioctl(freq   = 0x1f     ): Ok
  ioctl(freq   = 0x20     ): Ok
  ioctl(freq   = 0x80     ): Ok
  ioctl(freq   = 0x8f     ): Ok
  ioctl(freq   = 0x90     ): Ok
  ioctl(freq   = 0x91     ): Ok
  ioctl(freq   = 0x100    ): Ok

  IBS freq (negative) tests:
  --------------------------
  freq 1, sample_freq 200000: Ok

  IBS L3MissOnly test: (takes a while)
  --------------------
  Fetch L3MissOnly: Ok   (nr_samples: 1301)
  Op L3MissOnly:    Ok   (nr_samples: 1590)
  ---- end(0) ----
  112: AMD IBS sample period                                           : Ok

Committer notes:

Avoid using PAGE_SIZE as that define is also in sys/user.h

Make it a variable not to call sysconf() multiple times.

Also cast func to void * when passing it as the first arg to memcpy to
avoid this with some versions of clang:

  arch/x86/tests/amd-ibs-period.c:81:3: error: no matching function for call to 'memcpy'
                  memcpy(func, insn1, sizeof(insn1));
                  ^~~~~~
  /usr/include/string.h:27:7: note: candidate function not viable: no known conversion from 'int (*)(void)' to 'void *' for 1st argument
  void *memcpy (void *__restrict, const void *__restrict, size_t);
        ^
  /usr/include/fortify/string.h:40:27: note: candidate function not viable: no known conversion from 'int (*)(void)' to 'void *const' for 1st argument
  _FORTIFY_FN(memcpy) void *memcpy(void * _FORTIFY_POS0 __od,
                            ^
  arch/x86/tests/amd-ibs-period.c:87:3: error: no matching function for call to 'memcpy'

This one, for instance:

  Alpine clang version 19.1.4
  Target: x86_64-alpine-linux-musl
  Thread model: posix
  InstalledDir: /usr/lib/llvm19/bin
  Configuration file: /etc/clang19/x86_64-alpine-linux-musl.cfg
  System configuration file directory: /etc/clang19

Signed-off-by: Ravi Bangoria <ravi.bangoria@amd.com>
Cc: Ananth Narayan <ananth.narayan@amd.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Joe Mario <jmario@redhat.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sandipan Das <sandipan.das@amd.com>
Cc: Santosh Shukla <santosh.shukla@amd.com>
Cc: Stephane Eranian <eranian@google.com>
Link: https://lore.kernel.org/r/20250429035938.1301-5-ravi.bangoria@amd.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf mem/c2c amd: Add ldlat support

'perf mem/c2c' uses IBS Op PMU on AMD platforms.

IBS Op PMU on Zen5 uarch has added support for Load Latency filtering.

Implement 'perf mem/c2c' --ldlat using IBS Op Load Latency filtering
capability.

Some subtle differences between AMD and other arch:

o --ldlat is disabled by default on AMD

o Supported values are 128 to 2048.

Signed-off-by: Ravi Bangoria <ravi.bangoria@amd.com>
Cc: Ananth Narayan <ananth.narayan@amd.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Joe Mario <jmario@redhat.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sandipan Das <sandipan.das@amd.com>
Cc: Santosh Shukla <santosh.shukla@amd.com>
Cc: Stephane Eranian <eranian@google.com>
Link: https://lore.kernel.org/r/20250429035938.1301-4-ravi.bangoria@amd.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf amd ibs: Incorporate Zen5 DTLB and PageSize information

IBS Op PMU on Zen5 reports DTLB and page size information differently
compared to prior generation.

  IBS_OP_DATA3     Zen3/4                 Zen5
  ----------------------------------------------------------------
  19               IbsDcL2TlbHit1G        Reserved
  ----------------------------------------------------------------
   6               IbsDcL2tlbHit2M        Reserved
  ----------------------------------------------------------------
   5               IbsDcL1TlbHit1G        PageSize:
   4               IbsDcL1TlbHit2M          0 - 4K
                                            1 - 2M
                                            2 - 1G
                                            3 - Reserved
                                          Valid only if
                                            IbsDcPhyAddrValid = 1
  ----------------------------------------------------------------
   3               IbsDcL2TlbMiss         IbsDcL2TlbMiss
                                          Valid only if
                                            IbsDcPhyAddrValid = 1
  ----------------------------------------------------------------
   2               IbsDcL1tlbMiss         IbsDcL1tlbMiss
                                          Valid only if
                                            IbsDcPhyAddrValid = 1
  ----------------------------------------------------------------

Kernel expose this change as "dtlb_pgsize" capability in PMU sysfs.

Change IBS register raw-dump logic according to new bit definitions.

Signed-off-by: Ravi Bangoria <ravi.bangoria@amd.com>
Cc: Ananth Narayan <ananth.narayan@amd.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Joe Mario <jmario@redhat.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sandipan Das <sandipan.das@amd.com>
Cc: Santosh Shukla <santosh.shukla@amd.com>
Cc: Stephane Eranian <eranian@google.com>
Link: https://lore.kernel.org/r/20250429035938.1301-3-ravi.bangoria@amd.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf amd ibs: Add Load Latency bits in raw dump

IBS OP PMU on Zen5 supports Load Latency filtering. Decode and dump Load
Latency filtering related bits into perf script raw dump.

Also add oneliner example in the perf-amd-ibs man page.

Signed-off-by: Ravi Bangoria <ravi.bangoria@amd.com>
Cc: Ananth Narayan <ananth.narayan@amd.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Joe Mario <jmario@redhat.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sandipan Das <sandipan.das@amd.com>
Cc: Santosh Shukla <santosh.shukla@amd.com>
Cc: Stephane Eranian <eranian@google.com>
Link: https://lore.kernel.org/r/20250429035938.1301-2-ravi.bangoria@amd.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf symbols: Handle 'u' and 'l' symbols in /proc/kallsyms

I started seeing this in recent Fedora 42 kernels:

  # uname -a
  Linux number 6.14.3-300.fc42.x86_64 #1 SMP PREEMPT_DYNAMIC Sun Apr 20 16:08:39 UTC 2025 x86_64 GNU/Linux
  #

  # perf test vmlinux
    1: vmlinux symtab matches kallsyms                                 : FAILED!
  #

Where we have Rust enabled:

  # grep CONFIG_RUST /boot/config-6.14.3-300.fc42.x86_64
  CONFIG_RUSTC_VERSION=108600
  CONFIG_RUST_IS_AVAILABLE=y
  CONFIG_RUSTC_LLVM_VERSION=200101
  CONFIG_RUSTC_HAS_COERCE_POINTEE=y
  CONFIG_RUST=y
  CONFIG_RUSTC_VERSION_TEXT="rustc 1.86.0 (05f9846f8 2025-03-31) (Fedora 1.86.0-1.fc42)"
  CONFIG_RUST_FW_LOADER_ABSTRACTIONS=y
  CONFIG_RUST_PHYLIB_ABSTRACTIONS=y
  # CONFIG_RUST_DEBUG_ASSERTIONS is not set
  CONFIG_RUST_OVERFLOW_CHECKS=y
  # CONFIG_RUST_BUILD_ASSERT_ALLOW is not set
  #

Looking at the reason for the failure:

  # perf test -v vmlinux |& grep ^ERR
  ERR : 0xffffffff99efc7d0: __pfx__RNCINvNtNtNtCsf5tcb0XGUW4_4core4iter8adapters3map12map_try_foldjNtCsagR6JbSOIa9_12drm_panic_qr7VersionuINtNtNtBa_3ops12control_flow11ControlFlowB10_ENcB10_0NCINvNvNtNtNtB8_6traits8iterator8Iterator4find5checkB10_NCNvMB12_B10_13from_segments0E0E0B12_ not on kallsyms
  ERR : 0xffffffff99efc7e0: _RNCINvNtNtNtCsf5tcb0XGUW4_4core4iter8adapters3map12map_try_foldjNtCsagR6JbSOIa9_12drm_panic_qr7VersionuINtNtNtBa_3ops12control_flow11ControlFlowB10_ENcB10_0NCINvNvNtNtNtB8_6traits8iterator8Iterator4find5checkB10_NCNvMB12_B10_13from_segments0E0E0B12_ not on kallsyms
  #

But:

  # grep -w u /proc/kallsyms
  ffffffff99efc7d0 u __pfx__RNCINvNtNtNtCsf5tcb0XGUW4_4core4iter8adapters3map12map_try_foldjNtCsagR6JbSOIa9_12drm_panic_qr7VersionuINtNtNtBa_3ops12control_flow11ControlFlowB10_ENcB10_0NCINvNvNtNtNtB8_6traits8iterator8Iterator4find5checkB10_NCNvMB12_B10_13from_segments0E0E0B12_
  ffffffff99efc7e0 u _RNCINvNtNtNtCsf5tcb0XGUW4_4core4iter8adapters3map12map_try_foldjNtCsagR6JbSOIa9_12drm_panic_qr7VersionuINtNtNtBa_3ops12control_flow11ControlFlowB10_ENcB10_0NCINvNvNtNtNtB8_6traits8iterator8Iterator4find5checkB10_NCNvMB12_B10_13from_segments0E0E0B12_
  #

The test checks that "vmlinux symtab matches kallsyms", so it finds those two
symbols in vmlinux:

  # pahole --running_kernel_vmlinux
  /usr/lib/debug/lib/modules/6.14.3-300.fc42.x86_64/vmlinux
  #

  # readelf -sW /usr/lib/debug/lib/modules/6.14.3-300.fc42.x86_64/vmlinux | grep -Ew '(__pfx__RNCINvNtNtNtCsf5tcb0XGUW4_4core4iter8adapters3map12map_try_foldjNtCsagR6JbSOIa9_12drm_panic_qr7VersionuINtNtNtBa_3ops12control_flow11ControlFlowB10_ENcB10_0NCINvNvNtNtNtB8_6traits8iterator8Iterator4find5checkB10_NCNvMB12_B10_13from_segments0E0E0B12_|_RNCINvNtNtNtCsf5tcb0XGUW4_4core4iter8adapters3map12map_try_foldjNtCsagR6JbSOIa9_12drm_panic_qr7VersionuINtNtNtBa_3ops12control_flow11ControlFlowB10_ENcB10_0NCINvNvNtNtNtB8_6traits8iterator8Iterator4find5checkB10_NCNvMB12_B10_13from_segments0E0E0B12_)'
 81844: ffffffff81efc7e0   524 FUNC    LOCAL  DEFAULT    1 _RNCINvNtNtNtCsf5tcb0XGUW4_4core4iter8adapters3map12map_try_foldjNtCsagR6JbSOIa9_12drm_panic_qr7VersionuINtNtNtBa_3ops12control_flow11ControlFlowB10_ENcB10_0NCINvNvNtNtNtB8_6traits8iterator8Iterator4find5checkB10_NCNvMB12_B10_13from_segments0E0E0B12_
144259: ffffffff81efc7d0    16 FUNC    LOCAL  DEFAULT    1 __pfx__RNCINvNtNtNtCsf5tcb0XGUW4_4core4iter8adapters3map12map_try_foldjNtCsagR6JbSOIa9_12drm_panic_qr7VersionuINtNtNtBa_3ops12control_flow11ControlFlowB10_ENcB10_0NCINvNvNtNtNtB8_6traits8iterator8Iterator4find5checkB10_NCNvMB12_B10_13from_segments0E0E0B12_
  #

It is there.

From the nm documentation we can see that:

           "U" The symbol is undefined.

           "u" The symbol is a unique global symbol.  This is a GNU extension to the
       standard set of ELF symbol bindings.  For such a symbol the dynamic
       linker will make sure that in the entire process there is just one
       symbol with this name and type in use.

So lets consider 'u' symbols in /proc/kallsyms when loading it to cover this case.

Fedora:40 shows this as a 'l' symbol, so consider that as well.

With this patch 'perf test 1' is happy again:

  # perf test vmlinux
    1: vmlinux symtab matches kallsyms                                 : Ok
  #

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Masahiro Yamada <masahiroy@kernel.org>
Cc: Miguel Ojeda <ojeda@kernel.org>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/aBE_n0PGl3g6h-cS@x1
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf test probe_vfs_getname: Skip if no suitable line detected

In some cases when calling function add_probe_vfs_getname, line number
can't be detected by 'perf probe -L getname_flags':

  78         atomic_set(&result->refcnt, 1);

     // one of the following lines should have line number
     // but sometimes it does not because of optimization
     result->uptr = filename;
             result->aname = NULL;

  81         audit_getname(result);

To prevent false failures, skip the affected tests if no suitable line
numbers can be detected.

Signed-off-by: Jakub Brnak <jbrnak@redhat.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Michael Petlan <mpetlan@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Tomas Glozar <tglozar@redhat.com>
Link: https://lore.kernel.org/r/20250324144523.597557-1-jbrnak@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf lock contention: Symbolize zone->lock using BTF

The struct zone is embedded in struct pglist_data which can be allocated
for each NUMA node early in the boot process.  As it's not a slab object
nor a global lock, this was not symbolized.

Since the zone->lock is often contended, it'd be nice if we can
symbolize it.  On NUMA systems, node_data array will have pointers for
struct pglist_data.  By following the pointer, it can calculate the
address of each zone and its lock using BTF.  On UMA, it can just use
contig_page_data and its zones.

The following example shows the zone lock contention at the end.

  $ sudo ./perf lock con -abl -E 5 -- ./perf bench sched messaging
  # Running 'sched/messaging' benchmark:
  # 20 sender and receiver processes per group
  # 10 groups == 400 processes run

       Total time: 0.038 [sec]
   contended   total wait     max wait     avg wait            address   symbol

        5167     18.17 ms     10.27 us      3.52 us   ffff953340052d00   &kmem_cache_node (spinlock)
          38     11.75 ms    465.49 us    309.13 us   ffff95334060c480   &sock_inode_cache (spinlock)
        3916     10.13 ms     10.43 us      2.59 us   ffff953342aecb40   &kmem_cache_node (spinlock)
        2963     10.02 ms     13.75 us      3.38 us   ffff9533d2344098   &kmalloc-rnd-08-2k (spinlock)
         216      5.05 ms     99.49 us     23.39 us   ffff9542bf7d65d0   zone_lock (spinlock)

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Song Liu <song@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: bpf@vger.kernel.org
Cc: linux-mm@kvack.org
Link: https://lore.kernel.org/r/20250401063055.7431-1-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf test: Add perf trace summary test

  $ sudo ./perf test -vv 'trace summary'
  109: perf trace summary:
  --- start ---
  test child forked, pid 3501572
  testing: perf trace -s -- true
  testing: perf trace -S -- true
  testing: perf trace -s --summary-mode=thread -- true
  testing: perf trace -S --summary-mode=total -- true
  testing: perf trace -as --summary-mode=thread --no-bpf-summary -- true
  testing: perf trace -as --summary-mode=total --no-bpf-summary -- true
  testing: perf trace -as --summary-mode=thread --bpf-summary -- true
  testing: perf trace -as --summary-mode=total --bpf-summary -- true
  testing: perf trace -aS --summary-mode=total --bpf-summary -- true
  ---- end(0) ----
  109: perf trace summary                                              : Ok

Reviewed-by: Howard Chu <howardchu95@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Howard Chu <howardchu95@gmail.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Song Liu <song@kernel.org>
Cc: bpf@vger.kernel.org
Link: https://lore.kernel.org/r/20250326044001.3503432-2-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf trace: Implement syscall summary in BPF

When -s/--summary option is used, it doesn't need (augmented) arguments
of syscalls.  Let's skip the augmentation and load another small BPF
program to collect the statistics in the kernel instead of copying the
data to the ring-buffer to calculate the stats in userspace.  This will
be much more light-weight than the existing approach and remove any lost
events.

Let's add a new option --bpf-summary to control this behavior.  I cannot
make it default because there's no way to get e_machine in the BPF which
is needed for detecting different ABIs like 32-bit compat mode.

No functional changes intended except for no more LOST events. :)

  $ sudo ./perf trace -as --summary-mode=total --bpf-summary sleep 1

   Summary of events:

   total, 6194 events

     syscall            calls  errors  total       min       avg       max       stddev
                                       (msec)    (msec)    (msec)    (msec)        (%)
     --------------- --------  ------ -------- --------- --------- ---------     ------
     epoll_wait           561      0  4530.843     0.000     8.076   520.941     18.75%
     futex                693     45  4317.231     0.000     6.230   500.077     21.98%
     poll                 300      0  1040.109     0.000     3.467   120.928     17.02%
     clock_nanosleep        1      0  1000.172  1000.172  1000.172  1000.172      0.00%
     ppoll                360      0   872.386     0.001     2.423   253.275     41.91%
     epoll_pwait           14      0   384.349     0.001    27.453   380.002     98.79%
     pselect6              14      0   108.130     7.198     7.724     8.206      0.85%
     nanosleep             39      0    43.378     0.069     1.112    10.084     44.23%
     ...

Reviewed-by: Howard Chu <howardchu95@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Howard Chu <howardchu95@gmail.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Song Liu <song@kernel.org>
Link: https://lore.kernel.org/r/20250326044001.3503432-1-namhyung@kernel.org
[ Added fixup sent from Namhyung in response to my report to make it also dependent on CONFIG_TRACE ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

MAINTAINERS: Add hisilicon PMU JSON events under its entry

The all hisilicon PMU JSON events were missing to be listed there.

Reviewed-by: James Clark <james.clark@linaro.org>
Signed-off-by: Junhao He <hejunhao3@huawei.com>
Acked-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Acked-by: Yicong Yang <yangyicong@hisilicon.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20250418070812.3771441-4-hejunhao3@huawei.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events arm64: Drop hip08 PublicDescription if same as BriefDescription

If BriefDescription and PublicDescription are the same, only
BriefDescription is needed. It will be used for both long and short
format outputs.

Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Yicong Yang <yangyicong@hisilicon.com>
Signed-off-by: Junhao He <hejunhao3@huawei.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20250418070812.3771441-3-hejunhao3@huawei.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events arm64: Fill up Desc field for Hisi hip08 hha pmu

In the same PMU, when some JSON events have the "BriefDescription" field
populated while others do not, the cmp_sevent() function will split these
two types of events into separate groups. As a result, when using perf
list to display events, the two types of events cannot be grouped together
in the output.

before patch:
 $ perf list pmu
 ...
 uncore hha:
   hisi_sccl1_hha2/sdir-hit/
   hisi_sccl1_hha2/sdir-lookup/
 ...
 uncore hha:
   edir-hit
      [Count of The number of HHA E-Dir hit operations. Unit: hisi_sccl1_hha2]
 ...

after patch:
 $ perf list pmu
 ...
 uncore hha:
   edir-hit
      [Count of The number of HHA E-Dir hit operations. Unit: hisi_sccl1_hha2]
   sdir-hit
      [Count of The number of HHA S-Dir hit operations. Unit: hisi_sccl1_hha2]
   sdir-lookup
      [Count of the number of HHA S-Dir lookup operations. Unit: hisi_sccl1_hha2]
 ...

Reviewed-by: James Clark <james.clark@linaro.org>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Yicong Yang <yangyicong@hisilicon.com>
Signed-off-by: Junhao He <hejunhao3@huawei.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Cc: Leo Yan <leo.yan@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20250418070812.3771441-2-hejunhao3@huawei.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf bench evlist-open-close: Reduce scope of 2 variables

Make 2 global variables local. Reduces ELF binary size by removing
relocations. For a no flags build, the perf binary size is reduced by
4,144 bytes on x86-64.

Reviewed-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Dapeng Mi <dapeng1.mi@linux.intel.com>
Cc: Dominique Martinet <asmadeus@codewreck.org>
Cc: Dr. David Alan Gilbert <linux@treblig.org>
Cc: Hao Ge <gehao@kylinos.cn>
Cc: Howard Chu <howardchu95@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@linaro.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Levi Yun <yeoreum.yun@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Tengda Wu <wutengda@huaweicloud.com>
Cc: Thomas Richter <tmricht@linux.ibm.com>
Cc: Veronika Molnarova <vmolnaro@redhat.com>
Cc: Weilin Wang <weilin.wang@intel.com>
Cc: Xu Yang <xu.yang_2@nxp.com>
Link: https://lore.kernel.org/r/20250410173631.1713627-3-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf tests record: Cleanup improvements

Remove the script output file. Add a trap debug message. Minor style
consistency changes.

Signed-off-by: Ian Rogers <irogers@google.com>
Link: https://lore.kernel.org/r/20250410173631.1713627-2-irogers@google.com
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Levi Yun <yeoreum.yun@arm.com>
Cc: Dominique Martinet <asmadeus@codewreck.org>
Cc: Howard Chu <howardchu95@gmail.com>
Cc: Tengda Wu <wutengda@huaweicloud.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Weilin Wang <weilin.wang@intel.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Hao Ge <gehao@kylinos.cn>
Cc: James Clark <james.clark@linaro.org>
Cc: Thomas Richter <tmricht@linux.ibm.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Dapeng Mi <dapeng1.mi@linux.intel.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Xu Yang <xu.yang_2@nxp.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Veronika Molnarova <vmolnaro@redhat.com>
Cc: Dr. David Alan Gilbert <linux@treblig.org>
Cc: bpf@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: linux-perf-users@vger.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf tests metric-only perf stat: Fix tests 84 and 86 s390

On s390x KVM and z/VM machines the CPU Measurement Facility is not
available. Events cycles and instructions do not exist.  Running above
tests on s390 KVM and z/VM guests always fail with this error:

  # ./perf test 84 86
  84: perf stat JSON output linter          : FAILED!
  86: perf stat STD output linter           : FAILED!
  #

Root cause is command:

  # perf stat -j --metric-only -e instructions,cycles -- true
  {"metric-value" : "none"}
  #

Which fails due to unsupported events and returns "none".
Do not execute this test case on s390 KVM and z/VM machines.

Output after:
  # ./perf test 84 86
  84: perf stat JSON output linter          : Ok
  86: perf stat STD output linter           : Ok
  #

Fixes: 45a86d017adf4d6c ("perf test: Add --metric-only to perf stat output tests")
Suggested-by: Heiko Carstens <hca@linux.ibm.com>
Suggested-by: Sumanth Korikkar <sumanthk@linux.ibm.com>
Reviewed-by: Sumanth Korikkar <sumanthk@linux.ibm.com>
Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Cc: Alexander Gordeev <agordeev@linux.ibm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Link: https://lore.kernel.org/r/20250424133310.37452-1-tmricht@linux.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf tool_pmu: Fix aggregation on duration_time

evsel__count_has_error() fails counters when the enabled or running time
are 0. The duration_time event reads 0 when the cpu_map_idx != 0 to
avoid aggregating time over CPUs. Change the enable and running time
to always have a ratio of 100% so that evsel__count_has_error won't
fail.

Before:
```
$ sudo /tmp/perf/perf stat --per-core -a -M UNCORE_FREQ sleep 1

 Performance counter stats for 'system wide':

S0-D0-C0              1      2,615,819,485      UNC_CLOCK.SOCKET                 #     2.61 UNCORE_FREQ
S0-D0-C0              2      <not counted>      duration_time

       1.002111784 seconds time elapsed
```

After:
```
$ perf stat --per-core -a -M UNCORE_FREQ sleep 1

 Performance counter stats for 'system wide':

S0-D0-C0              1        758,160,296      UNC_CLOCK.SOCKET                 #     0.76 UNCORE_FREQ
S0-D0-C0              2      1,003,438,246      duration_time

       1.002486017 seconds time elapsed
```

Note: the metric reads the value a different way and isn't impacted.

Fixes: 240505b2d0adcdc8 ("perf tool_pmu: Factor tool events into their own PMU")
Reported-by: Stephane Eranian <eranian@google.com>
Reviewed-by: James Clark <james.clark@linaro.org>
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Richter <tmricht@linux.ibm.com>
Link: https://lore.kernel.org/r/20250423050358.94310-1-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf session: Skip unsupported new event types

`perf report` currently halts with an error when encountering
unsupported new event types (`event.type >= PERF_RECORD_HEADER_MAX`).

This patch modifies the behavior to skip these samples and continue
processing the remaining events.

Additionally, stops reporting if the new event size is not 8-byte
aligned.

Suggested-by: Arnaldo Carvalho de Melo <acme@kernel.org>
Suggested-by: Namhyung Kim <namhyung@kernel.org>
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Chun-Tse Shao <ctshao@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ben Gainey <ben.gainey@arm.com>
Cc: Dmitriy Vyukov <dvyukov@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20250414173921.2905822-1-ctshao@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hist: Allow custom output fields in hierarchy mode

Now it can handle multiple output fields and sort keys in separate
levels, so it should be ok to use it in the hierarchy mode.  This
allows fully customized output format.

  $ perf report -F latency,comm,parallelism -H --stdio
  ...
  #     Latency  Command / Parallelism
  # ...........  .....................
  #
      31.84%     cc1
         29.96%     5
          1.24%     4
          0.37%     6
          0.26%     3
          0.02%     2
      24.68%     as
         22.39%     5
          1.12%     2
          0.98%     4
          0.12%     3
          0.07%     6
          ...

Committer testing:

Before:

  $ perf report -F latency,comm,parallelism -H --stdio
  Error: --hierarchy and --fields options cannot be used together

   Usage: perf report [<options>]

      -F, --fields <key[,keys...]>
                            output field(s): overhead latency overhead_sys overhead_us
     overhead_guest_sys overhead_guest_us overhead_children
     latency_children sample period weight1 weight2 weight3
  <SNIP>
      -H, --hierarchy       Show entries in a hierarchy
  $

After:

  $ perf report -F latency,comm,parallelism -H --stdio
  # Total Lost Samples: 0
  #
  # Samples: 1K of event 'cycles:Pu'
  # Event count (approx.): 1581450138
  #
  #     Latency  Command / Parallelism
  # ...........  .....................
  #
      97.66%     git
         96.95%     1
          0.55%     2
          0.04%     5
          0.03%     8
          0.03%     4
          0.02%     3
          0.01%     9
          0.01%     7
          0.01%     6
          0.01%     10
          0.00%     12
       2.34%     git-remote-http
          2.24%     1
          0.07%     5
          0.02%     2
          0.00%     4

  #
  # (Tip: To analyze particular parallelism levels, try: perf report --latency --parallelism=32-64)

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20250331073722.4695-5-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hist: Set levels in output_field_add()

It turns out that the output fields didn't consider the hierarchy mode
and put all the fields in the same level.  To support hierarchy, each
non-output field should be in a separate level.

Pass a pointer to level to output_field_add() and make it increase the
level when it sees non-output fields.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20250331073722.4695-4-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hist: Remove formats in hierarchy when cancel latency

Likewise, it should remove latency output fields in hierarchy list.
Pass evlist to perf_hpp__cancel_latency() to handle them properly.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20250331073722.4695-3-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf hist: Remove formats in hierarchy when cancel children

This is to support hierarchy options with custom output fields.
Currently perf_hpp__cancel_cumulate() only removes accumulated
overhead and latency fields from the global perf_hpp_list.

This is not used in the hierarchy mode because each evsel's hist
has its own separate hpp_list.  So it needs to remove the fields
from the lists too.  Pass evlist to the function so that it can
iterate the evsels.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20250331073722.4695-2-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf record: Retirement latency cleanup in evsel__config

'perf record' will fail with retirement latency events as the open
doesn't do a perf_event_open system call.

Use evsel__config() to set up such events for recording by removing the
flag and enabling sample weights - the sample weights containing the
retirement latency.

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-17-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf pmu-events: Add retirement latency to JSON events inside of perf

The updated Intel vendor events add retirement latency for
graniterapids:

  https://lore.kernel.org/lkml/20250322063403.364981-14-irogers@google.com/

This change makes those values available within an alias/event within a
PMU and saves them into the evsel at event parse time.

When no TPEBS data is available the default values are substituted in
for TMA metrics that are using retirement latency events - currently
just those on graniterapids.

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-16-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf stat: Add mean, min, max and last --tpebs-mode options

Add command line configuration option for how retirement latency
events are combined.

The default "mean" gives the average of retirement latency.

"min" or "max" give the smallest or largest retirment latency times
respectively.

"last" uses the last retirment latency sample's time.

Committer notes:

Enclose parse_tpebs_mode() under HAVE_ARCH_X86_64_SUPPORT to match the
ifdef block where it is used, fixing the build in systems like:

  20     5.60 debian:experimental-x-mips    : FAIL gcc version 14.2.0 (Debian 14.2.0-1)
    builtin-stat.c:2330:12: error: 'parse_tpebs_mode' defined but not used [-Werror=unused-function]
     2330 | static int parse_tpebs_mode(const struct option *opt, const char *str,
          |            ^~~~~~~~~~~~~~~~

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-15-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf intel-tpebs: Use stats for retirement latency statistics

struct stats provides access to mean, min and max.

It also provides uniformity with statistics code used elsewhere in perf.

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-14-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf intel-tpebs: Don't close record on read

Factor sending record control fd code into its own function.

Rather than killing the record process send it a ping when reading.

Timeouts were witnessed if done too frequently, so only ping for the
first tpebs events.

Don't kill the record command send it a stop command.

As close isn't reliably called also close on evsel__exit.

Add extra checks on the record being terminated to avoid warnings.

Adjust the locking as needed and incorporate extra -Wthread-safety
checks.

Check to do six 500ms poll timeouts when sending commands, rather than
the larger 3000ms, to allow the record process terminating to be better
witnessed.

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-13-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf intel-tpebs: Add mutex for tpebs_results

Ensure sample reader isn't racing with events being added/removed.

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-12-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf intel-tpebs: Add support for updating counts in evsel__tpebs_read

Rename to reflect evsel argument and for consistency with other tpebs
functions.

Update count from prev_raw_counts when available.

Eventually this will allow inteval mode support.

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-11-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf intel-tpebs: Refactor tpebs_results list

evsel names and metric-ids are used for matching but this can be
problematic, for example, multiple occurrences of the same retirement
latency event become a single event for the record.

Change the name of the record events so they are unique and reflect the
evsel of the retirement latency event that opens them (the retirement
latency event's evsel address is embedded within them).

This allows an evsel based close to close the event when the retirement
latency event is closed.

This is important as 'perf stat' has an evlist and the session listen to
the record events has an evlist, knowing which event should remove the
tpebs_retire_lat can't be tied to an evlist list as there is more than
1, so closing which evlist should cause the tpebs to stop?

Using the evsel and the last one out doing the tpebs_stop is cleaner.

Committer notes:

Fix the build on 32-bit systems by using unsigned long when converting
pointers to integers instead of uint64_t. Fixes:

  20     4.97 debian:experimental-x-mips    : FAIL gcc version 14.2.0 (Debian 14.2.0-13)
    util/intel-tpebs.c: In function 'tpebs_retire_lat__find':
    util/intel-tpebs.c:377:21: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast]
      377 |                 if ((uint64_t)t->evsel == num)
          |                     ^
    cc1: all warnings being treated as errors

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-10-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf intel-tpebs: Ensure events are opened, factor out finding

Factor out finding an tpebs_retire_lat from an evsel.

Don't blindly return when ignoring an open request, which happens after
the first open request, ensure the event was started on a fork of perf
record.

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-9-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf intel-tpebs: Inline get_perf_record_args

Code is short enough to be inlined and there are no error cases when
made inline.

Make the implicit NULL pointer at the end of the argv explicit.

Move the fixed number of arguments before the variable number of
arguments.

Correctly size the argv allocation and zero when feeing to avoid a
dangling pointer.

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-8-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf intel-tpebs: Reduce scope of the tpebs_events_size variable

Moved to record argument computation rather than being global.

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-7-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf intel-tpebs: Move the cpumap_buf variable out of evsel__tpebs_open()

The buffer holds the cpumap to pass to the 'perf record' command, so
move it down to the 'perf record' function.

Make this function an evsel function given the need for the evsel for
the cpumap.

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-6-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf intel-tpebs: Separate evsel__tpebs_prepare() out of evsel__tpebs_open()

Separate the creation of the tpebs_retire_lat result out of the opening
step.

This is in preparation for adding a prepare operation for evlists.

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-5-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf intel-tpebs: Rename tpebs_start to evsel__tpebs_open

Try to add more consistency to evsel by having tpebs_start renamed to
evsel__tpebs_open, passing the evsel that is being opened. The unusual
behavior of evsel__tpebs_open opening all events on the evlist is kept
and will be cleaned up further in later patches. The comments are
cleaned up as tpebs_start isn't called from evlist.

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-4-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf intel-tpebs: Simplify tpebs_cmd

No need to dynamically allocate when there is 1. tpebs_pid duplicates
tpebs_cmd.pid, so remove. Use 0 as the uninitialized value (PID == 0
is reserved for the kernel) rather than -1.

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-3-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf intel-tpebs: Cleanup header

Remove arch conditional compilation. Arch conditional compilation
belongs in the arch/ directory.

Tidy header guards to match other files. Remove unneeded includes and
switch to forward declarations when necesary.

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Weilin Wang <weilin.wang@intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Falcon <thomas.falcon@intel.com>
Link: https://lore.kernel.org/r/20250414174134.3095492-2-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update westmereep-dp events

Update event topic moving other topic events to cache and virtual
memory.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-36-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update westmereep-dp events

Update event topic moving other topic events to cache and virtual
memory.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-35-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update westmereep-dp events

Update event topic moving other topic events to cache and virtual
memory.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-34-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update tigerlake metrics

Switch to metrics generated from the TMA spreadsheet. Minor threshold
simplification.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-33-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update snowridgex events

Update event topic moving other topic events to cache and memory. Add
PDIST counter into descriptions.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-32-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update skylakex events/metrics

Update event topics, metrics to be generated from the TMA spreadsheet
and other small clean ups.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-31-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update skylake metrics

Switch to metrics generated from the TMA spreadsheet. Minor threshold
simplification.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-30-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update sierraforest events/metrics

Update events from v1.08 to v1.09.

Update event topics, addition of PDIST counter into descriptions,
metrics to be generated from the TMA spreadsheet and other small clean
ups. The use of the spreadsheet for conversion has added thresholds.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-29-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update sapphirerapids events/metrics

Update event topics, addition of PDIST counter into descriptions,
metrics to be generated from the TMA spreadsheet and other small clean
ups.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-28-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update sandybridge metrics

Update TMA metrics from 4.8 to 5.02. Move INSTS_WRITTEN_TO_IQ.INSTS to
the frontend topic.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-27-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update rocketlake events/metrics

Update event topics, metrics to be generated from the TMA spreadsheet
and other small clean ups.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-26-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update nehalemex events

Update event topic moving other topic events to cache and virtual
memory.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-25-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update nehalemep events

Update event topic moving other topic events to cache and virtual
memory.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-24-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update meteorlake events/metrics

Update events from v1.12 to v1.13.
Update event topics, addition of PDIST counter into descriptions,
metrics to be generated from the TMA spreadsheet and other small clean
ups.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-23-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update lunarlake events/metrics

Update event topics, addition of PDIST counter into descriptions,
metrics to be generated from the TMA spreadsheet and other small clean
ups.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-22-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update jaketown metrics

Update TMA metrics from 4.8 to 5.02. Move INSTS_WRITTEN_TO_IQ.INSTS to
the frontend topic.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-21-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update ivytown metrics

Update TMA metrics from 4.8 to 5.02.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-20-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update ivybridge metrics

Update TMA metrics from 4.8 to 5.02.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-19-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update icelakex events/metrics

Update event topics, metrics to be generated from the TMA spreadsheet
and other small clean ups.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-18-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update icelake events/metrics

Update event topics, metrics to be generated from the TMA spreadsheet
and other small clean ups.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-17-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update haswellx metrics

Switch to metrics generated from the TMA spreadsheet. Minor threshold
simplification.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-16-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update haswell metrics

Switch to metrics generated from the TMA spreadsheet. Minor threshold
simplification.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-15-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Add graniterapids retirement latencies

Add retirement latencies for use in place of retirement latency events.
Update events from v1.06 to v1.08.
Update event topics, addition of PDIST counter into descriptions,
metrics to be generated from the TMA spreadsheet and other small clean
ups.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-14-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update grandridge events/metrics

Update events from v1.05 to v1.07. Update event topics, addition of
PDIST counter into descriptions, metrics to be generated from the TMA
spreadsheet and other small clean ups.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-13-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update emeraldrapids events/metrics

Update event topics, addition of PDIST counter into descriptions,
metrics to be generated from the TMA spreadsheet and other small clean
ups.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-12-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update elkhartlake events

Update event topic moving other topic events to cache and memory. Add
PDIST counter into descriptions.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-11-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update clearwaterforest events

Update event topic of OCR.DEMAND_DATA_RD.ANY_RESPONSE and
OCR.DEMAND_RFO.ANY_RESPONSE to be cache. Add PDIST counter into
descriptions.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-10-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update cascadelakex events/metrics

Update event topics, metrics to be generated from the TMA spreadsheet
and other small clean ups.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-9-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update broadwellx metrics

Switch to metrics generated from the TMA spreadsheet. Minor threshold
simplification.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-8-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update broadwellde metrics

Switch to metrics generated from the TMA spreadsheet. Minor threshold
simplification.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-7-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update broadwell metrics

Switch to metrics generated from the TMA spreadsheet. Minor threshold
simplification.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-6-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update bonnell events

Move DISPATCH_BLOCKED.ANY to the pipeline topic.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-5-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update arrowlake events/metrics

Update events from v1.07 to v1.08. Update event topics, addition of
PIST counter into descriptions, metrics to be generated from the TMA
spreadsheet and other small clean ups.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-4-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update AlderlakeN events/metrics

Update events from v1.28 to v1.29. Update event topics, addition of
PDIST counter into descriptions, metrics to be generated from the TMA
spreadsheet and other small clean ups.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-3-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf vendor events: Update alderlake events/metrics

Update events from v1.28 to v1.29. Update event topics, addition of
PDIST counter into descriptions, metrics to be generated from the TMA
spreadsheet and other small clean ups.

Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Thomas Falcon <thomas.falcon@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Andreas Färber <afaerber@suse.de>
Cc: Caleb Biggers <caleb.biggers@intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Perry Taylor <perry.taylor@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Weilin Wang <weilin.wang@intel.com>
Link: https://lore.kernel.org/r/20250328175006.43110-2-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

Merge remote-tracking branch 'torvalds/master' into perf-tools-next

Sync with upstream to pick up the perf-tools patches that updates the
header files copies to address the check_header.sh warnings. There are
also some libbpf updates, better pick those to be on the same page with
libbpf since perf uses it in various places.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

Merge tag 'edac_urgent_for_v6.15_rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/ras/ras

Pull EDAC fixes from Borislav Petkov:
 "Two fixes to the AMD translation library for the MI300 side of things:

   - Use the row[13] bit when calculating the memory row to retire

   - Mask the physical row address in order to avoid creating duplicate
     error records"

* tag 'edac_urgent_for_v6.15_rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/ras/ras:
  RAS/AMD/FMPM: Get masked address
  RAS/AMD/ATL: Include row[13] bit in row retirement

Merge tag 'fs_for_v6.15-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs

Pull isofs fix from Jan Kara:
 "Fix a case where isofs could be reading beyond end of the passed
  file handle if its type was incorrectly set"

* tag 'fs_for_v6.15-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
  isofs: Prevent the use of too small fid

Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma

Pull rdma fixes from Jason Gunthorpe:

 - Fix hang in bnxt_re due to miscomputing the budget

 - Avoid a -Wformat-security message in dev_set_name()

 - Avoid an unused definition warning in fs.c with some kconfigs

 - Fix error handling in usnic and remove IS_ERR_OR_NULL() usage

 - Regression in RXE support foudn by blktests due to missing ODP
   exclusions

 - Set the dma_segment_size on HNS so it doesn't corrupt DMA when using
   very large IOs

 - Move a INIT_WORK to near when the work is allocated in cm.c to fix a
   racey crash where work in progress was being init'd

 - Use __GFP_NOWARN to not dump in kvcalloc() if userspace requests a
   very big MR

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
  RDMA/bnxt_re: Remove unusable nq variable
  RDMA/core: Silence oversized kvmalloc() warning
  RDMA/cma: Fix workqueue crash in cma_netevent_work_handler
  RDMA/hns: Fix wrong maximum DMA segment size
  RDMA/rxe: Fix null pointer dereference in ODP MR check
  RDMA/mlx5: Fix compilation warning when USER_ACCESS isn't set
  RDMA/usnic: Fix passing zero to PTR_ERR in usnic_ib_pci_probe()
  RDMA/ucaps: Avoid format-security warning
  RDMA/bnxt_re: Fix budget handling of notification queue

Merge tag 'vfs-6.15-rc3.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs

Pull vfs fixes from Christian Brauner:

 - Fix NULL pointer dereference in virtiofs

 - Fix slab OOB access in hfs/hfsplus

 - Only create /proc/fs/netfs when CONFIG_PROC_FS is set

 - Fix getname_flags() to initialize pointer correctly

 - Convert dentry flags to enum

 - Don't allow datadir without lowerdir in overlayfs

 - Use namespace_{lock,unlock} helpers in dissolve_on_fput() instead of
   plain namespace_sem so unmounted mounts are properly cleaned up

 - Skip unnecessary ifs_block_is_uptodate check in iomap

 - Remove an unused forward declaration in overlayfs

 - Fix devpts uid/gid handling after converting to the new mount api

 - Fix afs_dynroot_readdir() to not use the RCU read lock

 - Fix mount_setattr() and open_tree_attr() to not pointlessly do path
   lookup or walk the mount tree if no mount option change has been
   requested

* tag 'vfs-6.15-rc3.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
  fs: use namespace_{lock,unlock} in dissolve_on_fput()
  iomap: skip unnecessary ifs_block_is_uptodate check
  fs: Fix filename init after recent refactoring
  netfs: Only create /proc/fs/netfs with CONFIG_PROC_FS
  mount: ensure we don't pointlessly walk the mount tree
  dcache: convert dentry flag macros to enum
  afs: Fix afs_dynroot_readdir() to not use the RCU read lock
  hfs/hfsplus: fix slab-out-of-bounds in hfs_bnode_read_key
  virtiofs: add filesystem context source name check
  devpts: Fix type for uid and gid params
  ovl: remove unused forward declaration
  ovl: don't allow datadir only

Merge tag 'perf-tools-fixes-for-v6.15-2025-04-13' of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools

Pull perf tools fixes from Namhyung Kim:
 "A couple of fixes and the usual tooling header updates:

   - fix a build error on ARM64 when libunwind is requested

   - fix an infinite loop with branch stack on AMD Zen3

   - sync tooling headers with the kernel source"

* tag 'perf-tools-fixes-for-v6.15-2025-04-13' of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools:
  perf tools: Remove evsel__handle_error_quirks()
  perf libunwind arm64: Fix missing close parens in an if statement
  tools headers: Update the arch/x86/lib/memset_64.S copy with the kernel sources
  tools headers: Update the x86 headers with the kernel sources
  tools headers: Update the linux/unaligned.h copy with the kernel sources
  tools headers: Update the uapi/asm-generic/mman-common.h copy with the kernel sources
  tools headers: Update the uapi/linux/prctl.h copy with the kernel sources
  tools headers: Update the syscall table with the kernel sources
  tools headers: Update the VFS headers with the kernel sources
  tools headers: Update the uapi/linux/perf_event.h copy with the kernel sources
  tools headers: Update the socket headers with the kernel sources
  tools headers: Update the KVM headers with the kernel sources

Linux 6.15-rc2

Merge tag 'erofs-for-6.15-rc2-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs

Pull erofs fixes from Gao Xiang:

 - Properly handle errors when file-backed I/O fails

 - Fix compilation issues on ARM platform (arm-linux-gnueabi)

 - Fix parsing of encoded extents

 - Minor cleanup

* tag 'erofs-for-6.15-rc2-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs:
  erofs: remove duplicate code
  erofs: fix encoded extents handling
  erofs: add __packed annotation to union(__le16..)
  erofs: set error to bio if file-backed IO fails