linux.git/tools/perf, branch v6.2-rc4

perf auxtrace: Fix address filter duplicate symbol selection

2023-01-11T17:03:44+00:00

When a match has been made to the nth duplicate symbol, return
success not error.

Example:

  Before:

    $ cat file.c
    cat: file.c: No such file or directory
    $ cat file1.c
    #include 

    static void func(void)
    {
            printf("First func\n");
    }

    void other(void);

    int main()
    {
            func();
            other();
            return 0;
    }
    $ cat file2.c
    #include 

    static void func(void)
    {
            printf("Second func\n");
    }

    void other(void)
    {
            func();
    }

    $ gcc -Wall -Wextra -o test file1.c file2.c
    $ perf record -e intel_pt//u --filter 'filter func @ ./test' -- ./test
    Multiple symbols with name 'func'
    #1      0x1149  l       func
                    which is near           main
    #2      0x1179  l       func
                    which is near           other
    Disambiguate symbol name by inserting #n after the name e.g. func #2
    Or select a global symbol by inserting #0 or #g or #G
    Failed to parse address filter: 'filter func @ ./test'
    Filter format is: filter|start|stop|tracestop  [/ ] [@]
    Where multiple filters are separated by space or comma.
    $ perf record -e intel_pt//u --filter 'filter func #2 @ ./test' -- ./test
    Failed to parse address filter: 'filter func #2 @ ./test'
    Filter format is: filter|start|stop|tracestop  [/ ] [@]
    Where multiple filters are separated by space or comma.

  After:

    $ perf record -e intel_pt//u --filter 'filter func #2 @ ./test' -- ./test
    First func
    Second func
    [ perf record: Woken up 1 times to write data ]
    [ perf record: Captured and wrote 0.016 MB perf.data ]
    $ perf script --itrace=b -Ftime,flags,ip,sym,addr --ns
    1231062.526977619:   tr strt                               0 [unknown] =>     558495708179 func
    1231062.526977619:   tr end  call               558495708188 func =>     558495708050 _init
    1231062.526979286:   tr strt                               0 [unknown] =>     55849570818d func
    1231062.526979286:   tr end  return             55849570818f func =>     55849570819d other

Fixes: 1b36c03e356936d6 ("perf record: Add support for using symbols in address filters")
Reported-by: Dmitrii Dolgov <9erthalion6@gmail.com>
Signed-off-by: Adrian Hunter 
Tested-by: Dmitry Dolgov <9erthalion6@gmail.com>
Cc: Adrian Hunter 
Cc: Ian Rogers 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20230110185659.15979-1-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo

perf bpf: Avoid build breakage with libbpf < 0.8.0 + LIBBPF_DYNAMIC=1

2023-01-10T15:11:09+00:00

In 746bd29e348f99b4 ("perf build: Use tools/lib headers from install
path") we stopped having the tools/lib/ directory from the kernel
sources in the header include path unconditionally, which breaks the
build on systems with older versions of libbpf-devel, in this case 0.7.0
as some of the structures and function declarations present in the newer
version of libbpf included in the kernel sources (tools/lib/bpf) are not
anymore used, just the ones in the system libbpf.

So instead of trying to provide alternative functions when the
libbpf-bpf_program__set_insns feature test fails, fail a
LIBBPF_DYNAMIC=1 build (requesting the use of the system's libbpf) and
emit this build error message:

  $ make LIBBPF_DYNAMIC=1 -C tools/perf
  Makefile.config:593: *** Error: libbpf devel library needs to be >= 0.8.0 to build with LIBBPF_DYNAMIC, update or build statically with the version that comes with the kernel sources.  Stop.
  $

For v6.3 these tests will be revamped and we'll require libbpf 1.0 as a
minimal version for using LIBBPF_DYNAMIC=1, most distros should have it
by now or at v6.3 time.

Fixes: 746bd29e348f99b4 ("perf build: Use tools/lib headers from install path")
Cc: Alexander Shishkin 
Cc: Ian Rogers 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Mark Rutland 
Cc: Mike Leach 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Link: http://lore.kernel.org/lkml/CAP-5=fVa51_URGsdDFVTzpyGmdDRj_Dj2EKPuDHNQ0BYgMSzUA@mail.gmail.com
Signed-off-by: Arnaldo Carvalho de Melo

perf build: Fix build error when NO_LIBBPF=1

2023-01-10T13:54:35+00:00

The $(LIBBPF) target should only be a dependency of prepare if the
static version of libbpf is needed. Add a new LIBBPF_STATIC variable
that is set by Makefile.config. Use LIBBPF_STATIC to determine whether
the CFLAGS, etc. need updating and for adding $(LIBBPF) as a prepare
dependency.

As Makefile.config isn't loaded for "clean" as a target, always set
LIBBPF_OUTPUT regardless of whether it is needed for $(LIBBPF). This
is done to minimize conditional logic for $(LIBBPF)-clean.

This issue and an original fix was reported by Mike Leach in:
https://lore.kernel.org/lkml/20230105172243.7238-1-mike.leach@linaro.org/

Fixes: 746bd29e348f99b4 ("perf build: Use tools/lib headers from install path")
Reported-by: Mike Leach 
Signed-off-by: Ian Rogers 
Tested-by: Jiri Olsa 
Cc: Alexander Shishkin 
Cc: bpf@vger.kernel.org
Cc: Ian Rogers 
Cc: Mark Rutland 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Link: http://lore.kernel.org/lkml/20230106151320.619514-2-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo

perf tools: Don't install libtraceevent plugins as its not anymore in the kernel sources

2023-01-10T13:52:49+00:00

While doing 'make -C tools/perf build-test' one can notice error
messages while trying to install libtraceevent plugins, stop doing that
as libtraceevent isn't anymore a homie.

These are the warnings dealt with:

   make_install_prefix_slash_O: make install prefix=/tmp/krava/
    failed to find: /tmp/krava/etc/bash_completion.d/perf
    failed to find: /tmp/krava/lib64/traceevent/plugins/plugin_cfg80211.so
    failed to find: /tmp/krava/lib64/traceevent/plugins/plugin_scsi.so
    failed to find: /tmp/krava/lib64/traceevent/plugins/plugin_xen.so
    failed to find: /tmp/krava/lib64/traceevent/plugins/plugin_function.so
    failed to find: /tmp/krava/lib64/traceevent/plugins/plugin_sched_switch.so
    failed to find: /tmp/krava/lib64/traceevent/plugins/plugin_mac80211.so
    failed to find: /tmp/krava/lib64/traceevent/plugins/plugin_kvm.so
    failed to find: /tmp/krava/lib64/traceevent/plugins/plugin_kmem.so
    failed to find: /tmp/krava/lib64/traceevent/plugins/plugin_hrtimer.so
    failed to find: /tmp/krava/lib64/traceevent/plugins/plugin_jbd2.so

Fixes: 4171925aa9f3f7bf ("tools lib traceevent: Remove libtraceevent")
Acked-by: Ian Rogers 
Cc: Adrian Hunter 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Steven Rostedt 
Link: http://lore.kernel.org/lkml/Y7xXz+TSpiCbQGjw@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo

perf kmem: Support field "node" in evsel__process_alloc_event() coping with recent tracepoint restructuring

2023-01-10T13:52:49+00:00

Commit 11e9734bcb6a7361 ("mm/slab_common: unify NUMA and UMA version of
tracepoints") adds the field "node" into the tracepoints 'kmalloc' and
'kmem_cache_alloc', so this patch modifies the event process function to
support the field "node".

If field "node" is detected by checking function evsel__field(), it
stats the cross allocation.

When the "node" value is NUMA_NO_NODE (-1), it means the memory can be
allocated from any memory node, in this case, we don't account it as a
cross allocation.

Fixes: 11e9734bcb6a7361 ("mm/slab_common: unify NUMA and UMA version of tracepoints")
Reported-by: Ravi Bangoria 
Reviewed-by: James Clark 
Signed-off-by: Leo Yan 
Cc: Alexander Shishkin 
Cc: Hyeonggon Yoo <42.hyeyoo@gmail.com>
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Mark Rutland 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Vlastimil Babka 
Link: https://lore.kernel.org/r/20230108062400.250690-2-leo.yan@linaro.org
Signed-off-by: Arnaldo Carvalho de Melo

perf kmem: Support legacy tracepoints

2023-01-10T13:52:49+00:00

Commit 11e9734bcb6a7361 ("mm/slab_common: unify NUMA and UMA version of
tracepoints") removed tracepoints 'kmalloc_node' and
'kmem_cache_alloc_node', we need to consider the tool should be backward
compatible.

If it detect the tracepoint "kmem:kmalloc_node", this patch enables the
legacy tracepoints, otherwise, it will ignore them.

Fixes: 11e9734bcb6a7361 ("mm/slab_common: unify NUMA and UMA version of tracepoints")
Reported-by: Ravi Bangoria 
Reviewed-by: James Clark 
Signed-off-by: Leo Yan 
Cc: Alexander Shishkin 
Cc: Hyeonggon Yoo <42.hyeyoo@gmail.com>
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Mark Rutland 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Vlastimil Babka 
Link: https://lore.kernel.org/r/20230108062400.250690-1-leo.yan@linaro.org
Signed-off-by: Arnaldo Carvalho de Melo

perf build: Properly guard libbpf includes

2023-01-10T13:51:39+00:00

Including libbpf header files should be guarded by HAVE_LIBBPF_SUPPORT.
In bpf_counter.h, move the skeleton utilities under HAVE_BPF_SKEL.

Fixes: d6a735ef3277c45f ("perf bpf_counter: Move common functions to bpf_counter.h")
Reported-by: Mike Leach 
Signed-off-by: Ian Rogers 
Tested-by: Arnaldo Carvalho de Melo 
Tested-by: Jiri Olsa 
Tested-by: Mike Leach 
Cc: Alexander Shishkin 
Cc: Mark Rutland 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Link: http://lore.kernel.org/lkml/20230105172243.7238-1-mike.leach@linaro.org
Signed-off-by: Arnaldo Carvalho de Melo

perf tests bpf prologue: Fix bpf-script-test-prologue test compile issue with clang

2023-01-09T14:57:47+00:00

While running 'perf test' for bpf, observed that "BPF prologue
generation" test case fails to compile with clang. Logs below from
powerpc:

  :33:2: error: use of undeclared identifier 'fmode_t'
          fmode_t f_mode = (fmode_t)_f_mode;
          ^
  :37:6: error: use of undeclared identifier 'f_mode'; did you mean '_f_mode'?
          if (f_mode & FMODE_WRITE)
              ^~~~~~
              _f_mode
  :30:60: note: '_f_mode' declared here
  int bpf_func__null_lseek(void *ctx, int err, unsigned long _f_mode,
                                                             ^
  2 errors generated.

The test code tests/bpf-script-test-prologue.c uses fmode_t.  And the
error above is for "fmode_t" which is defined in include/linux/types.h
as part of kernel build directory: "/lib/modules//build"
that comes from kernel devel [ soft link to /usr/src/ ].

Clang picks this header file from "-working-directory" build option that
specifies this build folder.

But the commit 14e4b9f4289aed2c ("perf trace: Raw augmented syscalls fix
libbpf 1.0+ compatibility") changed the include directory to use:
"/usr/include".

Post this change, types.h from /usr/include/ is getting picked upwhich
doesn’t contain definition of "fmode_t" and hence fails to compile.

Compilation command before this commit:

  /usr/bin/clang -D__KERNEL__ -D__NR_CPUS__=72 -DLINUX_VERSION_CODE=0x50e00 -xc  -I/root/lib/perf/include/bpf -nostdinc -I./arch/powerpc/include -I./arch/powerpc/include/generated  -I./include -I./arch/powerpc/include/uapi -I./arch/powerpc/include/generated/uapi -I./include/uapi -I./include/generated/uapi -include ./include/linux/compiler-version.h -include ./include/linux/kconfig.h  -Wno-unused-value -Wno-pointer-sign -working-directory /lib/modules//build -c - -target bpf  -g -O2 -o -

Compilation command after this commit:

  /usr/bin/clang -D__KERNEL__ -D__NR_CPUS__=72 -DLINUX_VERSION_CODE=0x50e00 -xc  -I/usr/include/ -nostdinc -I./arch/powerpc/include -I./arch/powerpc/include/generated  -I./include -I./arch/powerpc/include/uapi -I./arch/powerpc/include/generated/uapi -I./include/uapi -I./include/generated/uapi -include ./include/linux/compiler-version.h -include ./include/linux/kconfig.h  -Wno-unused-value -Wno-pointer-sign -working-directory /lib/modules//build -c - -target bpf  -g -O2 -o -

The difference is addition of -I/usr/include/  in the first line which
is causing the error. Fix this by adding typedef for "fmode_t" in the
testcase to solve the compile issue.

Fixes: 14e4b9f4289aed2c ("perf trace: Raw augmented syscalls fix libbpf 1.0+ compatibility")
Signed-off-by: Athira Jajeev 
Tested-by: Arnaldo Carvalho de Melo 
Cc: Andi Kleen 
Cc: Disha Goel 
Cc: Ian Rogers 
Cc: James Clark 
Cc: Jiri Olsa 
Cc: Kajol Jain 
Cc: linuxppc-dev@lists.ozlabs.org
Cc: Madhavan Srinivasan 
Cc: Michael Ellerman 
Cc: Nageswara R Sastry 
Cc: Namhyung Kim 
Link: https://lore.kernel.org/linux-perf-users/20230105120436.92051-1-atrajeev@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo

perf tools: Fix build on uClibc systems by adding missing sys/types.h include

2023-01-04T19:44:01+00:00

Not all libc implementations define ssize_t as part of stdio.h like
glibc does since the standard only requires this type to be defined by
unistd.h and sys/types.h. For this reason the perf build is currently
broken for toolchains based on uClibc, for instance.

Include sys/types.h explicitly to fix that.

Committer notes:

In addition, in the past this worked in uClibc test systems as there was
another way to get to sys/types.h that got removed in that cset:

  tools/perf/util/trace-event.h
    /usr/include/traceevent/event_parse.h # This got removed from util/trace-event.h in 378ef0f5d9d7f465
      /usr/include/regex.h
        /usr/include/sys/types.h
          typedef __ssize_t ssize_t;

So the size_t that is used in tools/perf/util/trace-event.h was being
obtained indirectly, by chance.

Fixes: 378ef0f5d9d7f465 ("perf build: Use libtraceevent from the system")
Signed-off-by: Jesus Sanchez-Palencia 
Cc: Alexander Shishkin 
Cc: Ian Rogers 
Cc: Jiri Olsa 
Cc: Mark Rutland 
Cc: Namhyung Kim 
Link: http://lore.kernel.org/lkml/20230104193414.606905-1-jesussanp@google.com
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Fix handling of --for-each-cgroup with --bpf-counters to match non BPF mode

2023-01-04T14:11:41+00:00

The --for-each-cgroup can have the same cgroup multiple times, but this
confuses BPF counters (since they have the same cgroup id), making only
the last cgroup events to be counted.

Let's check the cgroup name before adding a new entry to the cgroups
list.

Before:

  $ sudo ./perf stat -a --bpf-counters --for-each-cgroup /,/ sleep 1

   Performance counter stats for 'system wide':

        msec cpu-clock                        /
             context-switches                 /
             cpu-migrations                   /
             page-faults                      /
             cycles                           /
             instructions                     /
             branches                         /
             branch-misses                    /
            8,016.04 msec cpu-clock                        /                #    7.998 CPUs utilized
               6,152      context-switches                 /                #  767.461 /sec
                 250      cpu-migrations                   /                #   31.187 /sec
                 442      page-faults                      /                #   55.139 /sec
         613,111,487      cycles                           /                #    0.076 GHz
         280,599,604      instructions                     /                #    0.46  insn per cycle
          57,692,724      branches                         /                #    7.197 M/sec
           3,385,168      branch-misses                    /                #    5.87% of all branches

         1.002220125 seconds time elapsed

After it becomes similar to the non-BPF mode:

  $ sudo ./perf stat -a --bpf-counters --for-each-cgroup /,/  sleep 1

   Performance counter stats for 'system wide':

            8,013.38 msec cpu-clock                        /                #    7.998 CPUs utilized
               6,859      context-switches                 /                #  855.944 /sec
                 334      cpu-migrations                   /                #   41.680 /sec
                 345      page-faults                      /                #   43.053 /sec
         782,326,119      cycles                           /                #    0.098 GHz
         471,645,724      instructions                     /                #    0.60  insn per cycle
          94,963,430      branches                         /                #   11.851 M/sec
           3,685,511      branch-misses                    /                #    3.88% of all branches

         1.001864539 seconds time elapsed

Committer notes:

As a reminder, to test with BPF counters one has to use BUILD_BPF_SKEL=1
in the make command line and have clang/llvm installed when building
perf, otherwise the --bpf-counters option will not be available:

  # perf stat -a --bpf-counters --for-each-cgroup /,/ sleep 1
  Error: unknown option `bpf-counters'

   Usage: perf stat [] []

      -a, --all-cpus        system-wide collection from all CPUs
  
  #

Fixes: bb1c15b60b981d10 ("perf stat: Support regex pattern in --for-each-cgroup")
Signed-off-by: Namhyung Kim 
Tested-by: Arnaldo Carvalho de Melo 
Cc: Adrian Hunter 
Cc: bpf@vger.kernel.org
Cc: Ian Rogers 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: Song Liu 
Link: https://lore.kernel.org/r/20230104064402.1551516-5-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo