linux-stable.git/tools/lib/perf/include/internal, branch v6.5.2

perf evlist: Propagate user CPU maps intersecting core PMU maps

2023-05-27T12:38:39+00:00

The CPU map for a non-core PMU gives a default CPU value for
perf_event_open. For core PMUs the CPU map lists all CPUs the evsel
may be opened on. If there are >1 core PMU, the CPU maps will list the
CPUs for that core PMU, but the user_requested_cpus may contain CPUs
that are invalid for the PMU and cause perf_event_open to fail. To
avoid this, when propagating the CPU map for core PMUs intersect it
with the CPU map of the PMU (the evsel's "own_cpus").

Add comments to __perf_evlist__propagate_maps to explain its somewhat
complex behavior. Fix the related comments for system_wide in struct
perf_evsel.

Reviewed-by: Kan Liang 
Signed-off-by: Ian Rogers 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Ali Saidi 
Cc: Athira Rajeev 
Cc: Dmitrii Dolgov <9erthalion6@gmail.com>
Cc: Huacai Chen 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jing Zhang 
Cc: Jiri Olsa 
Cc: John Garry 
Cc: Kajol Jain 
Cc: Kang Minchul 
Cc: Leo Yan 
Cc: Madhavan Srinivasan 
Cc: Mark Rutland 
Cc: Mike Leach 
Cc: Ming Wang 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Ravi Bangoria 
Cc: Rob Herring 
Cc: Sandipan Das 
Cc: Sean Christopherson 
Cc: Suzuki Poulouse 
Cc: Thomas Richter 
Cc: Will Deacon 
Cc: Xing Zhengjun 
Cc: coresight@lists.linaro.org
Cc: linux-arm-kernel@lists.infradead.org
Link: https://lore.kernel.org/r/20230527072210.2900565-9-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo

perf evsel: Add is_pmu_core inorder to interpret own_cpus

2023-05-27T12:38:10+00:00

The behaviour of handling cpu maps varies for core and other PMUs. For
core PMUs the cpu map lists all valid CPUs, whereas for other PMUs the
map is the default CPU. Add a flag in the evsel to indicate if a PMU
is core to help with later interpreting of the cpu maps and populate
it when the evsel is created during parsing. When propagating cpu
maps, core PMUs should intersect the cpu map of the PMU with the user
requested one.

Reviewed-by: Kan Liang 
Signed-off-by: Ian Rogers 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Ali Saidi 
Cc: Athira Rajeev 
Cc: Dmitrii Dolgov <9erthalion6@gmail.com>
Cc: Huacai Chen 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jing Zhang 
Cc: Jiri Olsa 
Cc: John Garry 
Cc: Kajol Jain 
Cc: Kang Minchul 
Cc: Leo Yan 
Cc: Madhavan Srinivasan 
Cc: Mark Rutland 
Cc: Mike Leach 
Cc: Ming Wang 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Ravi Bangoria 
Cc: Rob Herring 
Cc: Sandipan Das 
Cc: Sean Christopherson 
Cc: Suzuki Poulouse 
Cc: Thomas Richter 
Cc: Will Deacon 
Cc: Xing Zhengjun 
Cc: coresight@lists.linaro.org
Cc: linux-arm-kernel@lists.infradead.org
Link: https://lore.kernel.org/r/20230527072210.2900565-7-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo

libperf rc_check: Enable implicitly with sanitizers

2023-04-20T18:12:02+00:00

If using leak sanitizer then implicitly enable reference count checking.

Signed-off-by: Ian Rogers 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Mark Rutland 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Link: https://lore.kernel.org/r/20230420171812.561603-1-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo

perf cpumap: Add reference count checking

2023-04-17T19:50:02+00:00

Enabled when REFCNT_CHECKING is defined. The change adds a memory
allocated pointer that is interposed between the reference counted cpu
map at a get and freed by a put. The pointer replaces the original
perf_cpu_map struct, so use of the perf_cpu_map via APIs remains
unchanged. Any use of the cpu map without the API requires two versions,
handled via the RC_CHK_ACCESS macro.

This change is intended to catch:

 - use after put: using a cpumap after you have put it will cause a
   segv.
 - unbalanced puts: two puts for a get will result in a double free
   that can be captured and reported by tools like address sanitizer,
   including with the associated stack traces of allocation and frees.
 - missing puts: if a put is missing then the get turns into a memory
   leak that can be reported by leak sanitizer, including the stack
   trace at the point the get occurs.

Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Alexey Bayduraev 
Cc: Andi Kleen 
Cc: Andrew Morton 
Cc: Andy Shevchenko 
Cc: Darren Hart 
Cc: Davidlohr Bueso 
Cc: Dmitriy Vyukov 
Cc: Eric Dumazet 
Cc: German Gomez 
Cc: Hao Luo 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: John Garry 
Cc: Kajol Jain 
Cc: Kan Liang 
Cc: Leo Yan 
Cc: Madhavan Srinivasan 
Cc: Mark Rutland 
Cc: Masami Hiramatsu 
Cc: Miaoqian Lin 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Riccardo Mancini 
Cc: Shunsuke Nakamura 
Cc: Song Liu 
Cc: Stephen Brennan 
Cc: Steven Rostedt (VMware) 
Cc: Thomas Gleixner 
Cc: Thomas Richter ,
Cc: Yury Norov 
Link: https://lore.kernel.org/lkml/20230407230405.2931830-3-irogers@google.com
[ Extracted from a larger patch ]
Signed-off-by: Arnaldo Carvalho de Melo

libperf: Add reference count checking macros

2023-04-17T18:53:01+00:00

The macros serve as a way to debug use of a reference counted struct.

The macros add a memory allocated pointer that is interposed between
the reference counted original struct at a get and freed by a put.

The pointer replaces the original struct, so use of the struct name
via APIs remains unchanged.

Signed-off-by: Ian Rogers 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Alexey Bayduraev 
Cc: Andi Kleen 
Cc: Andrew Morton 
Cc: Andy Shevchenko 
Cc: Darren Hart 
Cc: Davidlohr Bueso 
Cc: Dmitriy Vyukov 
Cc: Eric Dumazet 
Cc: German Gomez 
Cc: Hao Luo 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: John Garry 
Cc: Kajol Jain 
Cc: Kan Liang 
Cc: Leo Yan 
Cc: Madhavan Srinivasan 
Cc: Mark Rutland 
Cc: Masami Hiramatsu 
Cc: Miaoqian Lin 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Riccardo Mancini 
Cc: Shunsuke Nakamura 
Cc: Song Liu 
Cc: Stephane Eranian 
Cc: Stephen Brennan 
Cc: Steven Rostedt (VMware) 
Cc: Thomas Gleixner 
Cc: Thomas Richter 
Cc: Yury Norov 
Link: http://lore.kernel.org/lkml/20230407230405.2931830-2-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo

libperf: Add perf_cpu_map__refcnt() interanl accessor to use in the maps test

2023-04-17T18:52:36+00:00

To remove one more direct access to 'struct perf_cpu_map' so that we can
intercept accesses to its instantiations and refcount check it to catch
use after free, etc.

Cc: Adrian Hunter 
Cc: Alexey Bayduraev 
Cc: Dmitriy Vyukov 
Cc: Ian Rogers 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Riccardo Mancini 
Cc: Stephane Eranian 
Cc: Stephen Brennan 
Link: https://lore.kernel.org/lkml/ZD1qdYjG+DL6KOfP@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo

libperf: Add a perf_cpu_map__set_nr() available as an internal function for tools/perf to use

2023-04-12T17:53:35+00:00

We'll need to reference count check 'struct perf_cpu_map', so wrap
accesses to its internal state to allow intercepting accesses to its
instances.

Cc: Adrian Hunter 
Cc: Alexey Bayduraev 
Cc: Dmitriy Vyukov 
Cc: Ian Rogers 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Riccardo Mancini 
Cc: Stephane Eranian 
Cc: Stephen Brennan 
Link: https://lore.kernel.org/
Signed-off-by: Arnaldo Carvalho de Melo

libperf: Make perf_cpu_map__alloc() available as an internal function for tools/perf to use

2023-04-12T17:44:24+00:00

We had the open coded equivalent in perf_cpu_map__empty_new(), so reuse
what is in libperf.

Signed-off-by: Ian Rogers 
Cc: Adrian Hunter 
Cc: Alexey Bayduraev 
Cc: Dmitriy Vyukov 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Riccardo Mancini 
Cc: Stephane Eranian 
Cc: Stephen Brennan 
Link: https://lore.kernel.org/lkml/20230407230405.2931830-3-irogers@google.com
[ Split from a larger patch ]
Signed-off-by: Arnaldo Carvalho de Melo

perf evlist: Remove nr_groups

2023-03-13T20:42:27+00:00

Maintaining the number of groups during event parsing is problematic
and since changing to sort/regroup events can only be computed by a
linear pass over the evlist. As the value is generally only used in
tests, rather than hold it in a variable compute it by passing over
the evlist when necessary.

This change highlights that libpfm's counting of groups with a single
entry disagreed with regular event parsing. The libpfm tests are
updated accordingly.

Signed-off-by: Ian Rogers 
Cc: Adrian Hunter 
Cc: Alexander Shishkin 
Cc: Florian Fischer 
Cc: Ingo Molnar 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: John Garry 
Cc: Kajol Jain 
Cc: Kan Liang 
Cc: Kim Phillips 
Cc: Leo Yan 
Cc: Mark Rutland 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Ravi Bangoria 
Cc: Sean Christopherson 
Cc: Steinar H. Gunderson 
Cc: Stephane Eranian 
Cc: Suzuki Poulouse 
Cc: Xing Zhengjun 
Link: https://lore.kernel.org/r/20230312021543.3060328-2-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo

libperf: Propagate maps only if necessary

2022-10-06T11:03:53+00:00

The current code propagate evsel's cpu map settings to evlist when it's
added to an evlist.  But the evlist->all_cpus and each evsel's cpus will
be updated in perf_evlist__set_maps() later.  No need to do it before
evlist's cpus are set actually.

In fact it discards this intermediate all_cpus maps at the beginning
of perf_evlist__set_maps().  Let's not do this.  It's only needed when
an evsel is added after the evlist cpu/thread maps are set.

Reviewed-by: Adrian Hunter 
Signed-off-by: Namhyung Kim 
Cc: Ian Rogers 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Kan Liang 
Cc: Leo Yan 
Cc: Peter Zijlstra 
Link: https://lore.kernel.org/r/20221003204647.1481128-3-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo