1. 13 5月, 2016 3 次提交
    • A
      perf stat: Fallback to user only counters when perf_event_paranoid > 1 · 42ef8a78
      Arnaldo Carvalho de Melo 提交于
      After 0161028b ("perf/core: Change the default paranoia level to 2")
      'perf stat' fails for users without CAP_SYS_ADMIN, so just use
      'perf_evsel__fallback()' to have the same behaviour as 'perf record',
      i.e. set perf_event_attr.exclude_kernel to 1.
      
      Now:
      
        [acme@jouet linux]$ perf stat usleep 1
      
         Performance counter stats for 'usleep 1':
      
                0.352536      task-clock:u (msec)  #   0.423 CPUs utilized
                       0      context-switches:u   #   0.000 K/sec
                       0      cpu-migrations:u     #   0.000 K/sec
                      49      page-faults:u        #   0.139 M/sec
                 309,407      cycles:u             #   0.878 GHz
                 243,791      instructions:u       #   0.79  insn per cycle
                  49,622      branches:u           # 140.757 M/sec
                   3,884      branch-misses:u      #   7.83% of all branches
      
             0.000834174 seconds time elapsed
      
        [acme@jouet linux]$
      Reported-by: NIngo Molnar <mingo@kernel.org>
      Cc: Adrian Hunter <adrian.hunter@intel.com>
      Cc: Andy Lutomirski <luto@kernel.org>
      Cc: David Ahern <dsahern@gmail.com>
      Cc: Jiri Olsa <jolsa@kernel.org>
      Cc: Kees Cook <keescook@chromium.org>
      Cc: Linus Torvalds <torvalds@linux-foundation.org>
      Cc: Masami Hiramatsu <mhiramat@kernel.org>
      Cc: Namhyung Kim <namhyung@kernel.org>
      Cc: Peter Zijlstra <peterz@infradead.org>
      Cc: Wang Nan <wangnan0@huawei.com>
      Link: http://lkml.kernel.org/n/tip-b20jmx4dxt5hpaa9t2rroi0o@git.kernel.orgSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      42ef8a78
    • A
      perf evsel: Handle EACCESS + perf_event_paranoid=2 in fallback() · 08094828
      Arnaldo Carvalho de Melo 提交于
      Now with the default for the kernel.perf_event_paranoid sysctl being 2 [1]
      we need to fall back to :u, i.e. to set perf_event_attr.exclude_kernel
      to 1.
      
      Before:
      
        [acme@jouet linux]$ perf record usleep 1
        Error:
        You may not have permission to collect stats.
      
        Consider tweaking /proc/sys/kernel/perf_event_paranoid,
        which controls use of the performance events system by
        unprivileged users (without CAP_SYS_ADMIN).
      
        The current value is 2:
      
          -1: Allow use of (almost) all events by all users
        >= 0: Disallow raw tracepoint access by users without CAP_IOC_LOCK
        >= 1: Disallow CPU event access by users without CAP_SYS_ADMIN
        >= 2: Disallow kernel profiling by users without CAP_SYS_ADMIN
        [acme@jouet linux]$
      
      After:
      
        [acme@jouet linux]$ perf record usleep 1
        [ perf record: Woken up 1 times to write data ]
        [ perf record: Captured and wrote 0.016 MB perf.data (7 samples) ]
        [acme@jouet linux]$ perf evlist
        cycles:u
        [acme@jouet linux]$ perf evlist -v
        cycles:u: size: 112, { sample_period, sample_freq }: 4000, sample_type: IP|TID|TIME|PERIOD, disabled: 1, inherit: 1, exclude_kernel: 1, mmap: 1, comm: 1, freq: 1, enable_on_exec: 1, task: 1, sample_id_all: 1, exclude_guest: 1, mmap2: 1, comm_exec: 1
        [acme@jouet linux]$
      
      And if the user turns on verbose mode, an explanation will appear:
      
        [acme@jouet linux]$ perf record -v usleep 1
        Warning:
        kernel.perf_event_paranoid=2, trying to fall back to excluding kernel samples
        mmap size 528384B
        [ perf record: Woken up 1 times to write data ]
        Looking at the vmlinux_path (8 entries long)
        Using /lib/modules/4.6.0-rc7+/build/vmlinux for symbols
        [ perf record: Captured and wrote 0.016 MB perf.data (7 samples) ]
        [acme@jouet linux]$
      
      [1] 0161028b ("perf/core: Change the default paranoia level to 2")
      Reported-by: NIngo Molnar <mingo@kernel.org>
      Cc: Adrian Hunter <adrian.hunter@intel.com>
      Cc: Andy Lutomirski <luto@kernel.org>
      Cc: David Ahern <dsahern@gmail.com>
      Cc: Jiri Olsa <jolsa@kernel.org>
      Cc: Kees Cook <keescook@chromium.org>
      Cc: Linus Torvalds <torvalds@linux-foundation.org>
      Cc: Masami Hiramatsu <mhiramat@kernel.org>
      Cc: Namhyung Kim <namhyung@kernel.org>
      Cc: Peter Zijlstra <peterz@infradead.org>
      Cc: Wang Nan <wangnan0@huawei.com>
      Link: http://lkml.kernel.org/n/tip-b20jmx4dxt5hpaa9t2rroi0o@git.kernel.orgSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      08094828
    • A
      perf evsel: Improve EPERM error handling in open_strerror() · 7d173913
      Arnaldo Carvalho de Melo 提交于
      We were showing a hardcoded default value for the kernel.perf_event_paranoid
      sysctl, now that it became more paranoid (1 -> 2 [1]), this would need to be
      updated, instead show the current value:
      
        [acme@jouet linux]$ perf record ls
        Error:
        You may not have permission to collect stats.
      
        Consider tweaking /proc/sys/kernel/perf_event_paranoid,
        which controls use of the performance events system by
        unprivileged users (without CAP_SYS_ADMIN).
      
        The current value is 2:
      
          -1: Allow use of (almost) all events by all users
        >= 0: Disallow raw tracepoint access by users without CAP_IOC_LOCK
        >= 1: Disallow CPU event access by users without CAP_SYS_ADMIN
        >= 2: Disallow kernel profiling by users without CAP_SYS_ADMIN
        [acme@jouet linux]$
      
      [1] 0161028b ("perf/core: Change the default paranoia level to 2")
      Reported-by: NIngo Molnar <mingo@kernel.org>
      Cc: Adrian Hunter <adrian.hunter@intel.com>
      Cc: Andy Lutomirski <luto@kernel.org>
      Cc: David Ahern <dsahern@gmail.com>
      Cc: Jiri Olsa <jolsa@kernel.org>
      Cc: Kees Cook <keescook@chromium.org>
      Cc: Linus Torvalds <torvalds@linux-foundation.org>
      Cc: Masami Hiramatsu <mhiramat@kernel.org>
      Cc: Namhyung Kim <namhyung@kernel.org>
      Cc: Peter Zijlstra <peterz@infradead.org>
      Cc: Wang Nan <wangnan0@huawei.com>
      Link: http://lkml.kernel.org/n/tip-0gc4rdpg8d025r5not8s8028@git.kernel.orgSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      7d173913
  2. 12 5月, 2016 6 次提交
  3. 11 5月, 2016 1 次提交
    • N
      perf diff: Fix duplicated output column · e9d848cb
      Namhyung Kim 提交于
      The commit b97511c5 ("perf tools: Add overhead/overhead_children
      keys defaults via string") moved initialization of column headers but it
      missed to check the sort__mode.  As 'perf diff' doesn't call
      perf_hpp__init(), the setup_overhead() also should not be called.
      
      Before:
      
        # Baseline    Delta  Children  Overhead  Shared Object        Symbol
        # ........  .......  ........  ........  ...................  .......................
        #
            28.48%  -28.47%    28.48%    28.48%  [kernel.vmlinux ]    [k] intel_idle
            11.51%  -11.47%    11.51%    11.51%  libxul.so            [.] 0x0000000001a360f7
             3.49%   -3.49%     3.49%     3.49%  [kernel.vmlinux]     [k] generic_exec_single
             2.91%   -2.89%     2.91%     2.91%  libdbus-1.so.3.8.11  [.] 0x000000000000cdc2
             2.86%   -2.85%     2.86%     2.86%  libxcb.so.1.1.0      [.] 0x000000000000c890
             2.44%   -2.39%     2.44%     2.44%  [kernel.vmlinux]     [k] perf_event_aux_ctx
      
      After:
      
        # Baseline    Delta  Shared Object        Symbol
        # ........  .......  ...................  .......................
        #
            28.48%  -28.47%  [kernel.vmlinux]     [k] intel_idle
            11.51%  -11.47%  libxul.so            [.] 0x0000000001a360f7
             3.49%   -3.49%  [kernel.vmlinux]     [k] generic_exec_single
             2.91%   -2.89%  libdbus-1.so.3.8.11  [.] 0x000000000000cdc2
             2.86%   -2.85%  libxcb.so.1.1.0      [.] 0x000000000000c890
             2.44%   -2.39%  [kernel.vmlinux]     [k] perf_event_aux_ctx
      Signed-off-by: NNamhyung Kim <namhyung@kernel.org>
      Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      Acked-by: NJiri Olsa <jolsa@kernel.org>
      Cc: <stable@vger.kernel.org> # 4.5+
      Cc: Linus Torvalds <torvalds@linux-foundation.org>
      Cc: Peter Zijlstra <peterz@infradead.org>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Fixes: b97511c5 ("perf tools: Add overhead/overhead_children keys defaults via string")
      Link: http://lkml.kernel.org/r/1462890384-12486-2-git-send-email-acme@kernel.orgSigned-off-by: NIngo Molnar <mingo@kernel.org>
      e9d848cb
  4. 18 4月, 2016 1 次提交
    • A
      perf intel-pt: Fix segfault tracing transactions · 1342e0b7
      Adrian Hunter 提交于
      Tracing a workload that uses transactions gave a seg fault as follows:
      
        perf record -e intel_pt// workload
        perf report
        Program received signal SIGSEGV, Segmentation fault.
        0x000000000054b58c in intel_pt_reset_last_branch_rb (ptq=0x1a36110)
        	at util/intel-pt.c:929
        929 ptq->last_branch_rb->nr = 0;
        (gdb) p ptq->last_branch_rb
        $1 = (struct branch_stack *) 0x0
        (gdb) up
        1148 intel_pt_reset_last_branch_rb(ptq);
        (gdb) l
        1143 if (ret)
        1144 pr_err("Intel Processor Trace: failed to deliver transaction event
        1145 ret);
        1146
        1147 if (pt->synth_opts.callchain)
        1148 intel_pt_reset_last_branch_rb(ptq);
        1149
        1150 return ret;
        1151 }
        1152
        (gdb) p pt->synth_opts.callchain
        $2 = true
        (gdb)
        (gdb) bt
         #0 0x000000000054b58c in intel_pt_reset_last_branch_rb (ptq=0x1a36110)
         #1 0x000000000054c1e0 in intel_pt_synth_transaction_sample (ptq=0x1a36110)
         #2 0x000000000054c5b2 in intel_pt_sample (ptq=0x1a36110)
      
      Caused by checking the 'callchain' flag when it should have been the
      'last_branch' flag.  Fix that.
      Reported-by: NAndi Kleen <ak@linux.intel.com>
      Signed-off-by: NAdrian Hunter <adrian.hunter@intel.com>
      Cc: Jiri Olsa <jolsa@redhat.com>
      Cc: stable@vger.kernel.org # v4.4+
      Fixes: f14445ee ("perf intel-pt: Support generating branch stack")
      Link: http://lkml.kernel.org/r/1460977068-11566-1-git-send-email-adrian.hunter@intel.comSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      1342e0b7
  5. 31 3月, 2016 2 次提交
  6. 30 3月, 2016 1 次提交
    • A
      perf tools: Add missing initialization of perf_sample.cpumode in synthesized samples · 3ea223ad
      Arnaldo Carvalho de Melo 提交于
      In 473398a2 ("perf tools: Add cpumode to struct perf_sample"), I
      missed some places where perf_sample fields are directly initialized in
      addition to what is done in perf_evsel__parse_sample(), namely when
      synthesizing PERF_RECORD_{MMAP*,COMM,FORK,EXIT} for pre-existing threads
      and also in intel_pt and intel_bts when synthesizing events from
      processor trace, the jitdump code also was affected, fix it.
      
      The problem was noticed with running:
      
        # perf record -e intel_pt//u true
        # perf script
      
      Where the samples wouldn't get resolved because perf_sample.cpumode
      would be left as zero, i.e. PERF_RECORD_MISC_CPUMODE_UNKNOWN, not
      resolving as kernel, hypervisor or user cpu modes.
      
      Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
      Cc: Andi Kleen <ak@linux.intel.com>
      Cc: David Ahern <dsahern@gmail.com>
      Cc: Jiri Olsa <jolsa@redhat.com>
      Cc: Namhyung Kim <namhyung@kernel.org>
      Cc: Peter Zijlstra <peterz@infradead.org>
      Cc: Stephane Eranian <eranian@google.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Fixes: 473398a2 ("perf tools: Add cpumode to struct perf_sample")
      Link: http://lkml.kernel.org/n/tip-n5sdauxgk24d5nun8kuuu2mh@git.kernel.orgSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      3ea223ad
  7. 29 3月, 2016 1 次提交
  8. 24 3月, 2016 10 次提交
  9. 23 3月, 2016 8 次提交
  10. 22 3月, 2016 2 次提交
  11. 19 3月, 2016 2 次提交
  12. 16 3月, 2016 2 次提交
    • V
      mm, tracing: unify mm flags handling in tracepoints and printk · 420adbe9
      Vlastimil Babka 提交于
      In tracepoints, it's possible to print gfp flags in a human-friendly
      format through a macro show_gfp_flags(), which defines a translation
      array and passes is to __print_flags().  Since the following patch will
      introduce support for gfp flags printing in printk(), it would be nice
      to reuse the array.  This is not straightforward, since __print_flags()
      can't simply reference an array defined in a .c file such as mm/debug.c
      - it has to be a macro to allow the macro magic to communicate the
      format to userspace tools such as trace-cmd.
      
      The solution is to create a macro __def_gfpflag_names which is used both
      in show_gfp_flags(), and to define the gfpflag_names[] array in
      mm/debug.c.
      
      On the other hand, mm/debug.c also defines translation tables for page
      flags and vma flags, and desire was expressed (but not implemented in
      this series) to use these also from tracepoints.  Thus, this patch also
      renames the events/gfpflags.h file to events/mmflags.h and moves the
      table definitions there, using the same macro approach as for gfpflags.
      This allows translating all three kinds of mm-specific flags both in
      tracepoints and printk.
      Signed-off-by: NVlastimil Babka <vbabka@suse.cz>
      Reviewed-by: NMichal Hocko <mhocko@suse.com>
      Cc: Steven Rostedt <rostedt@goodmis.org>
      Cc: Peter Zijlstra <peterz@infradead.org>
      Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
      Cc: Ingo Molnar <mingo@redhat.com>
      Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk>
      Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
      Cc: Minchan Kim <minchan@kernel.org>
      Cc: Sasha Levin <sasha.levin@oracle.com>
      Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
      Cc: Mel Gorman <mgorman@suse.de>
      Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
      Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      420adbe9
    • V
      tools, perf: make gfp_compact_table up to date · 14e0a214
      Vlastimil Babka 提交于
      When updating tracing's show_gfp_flags() I have noticed that perf's
      gfp_compact_table is also outdated.  Fill in the missing flags and place
      a note in gfp.h to increase chance that future updates are synced.
      Convert the __GFP_X flags from "GFP_X" to "__GFP_X" strings in line with
      the previous patch.
      Signed-off-by: NVlastimil Babka <vbabka@suse.cz>
      Acked-by: NDavid Rientjes <rientjes@google.com>
      Cc: Steven Rostedt <rostedt@goodmis.org>
      Cc: Peter Zijlstra <peterz@infradead.org>
      Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
      Cc: Ingo Molnar <mingo@redhat.com>
      Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk>
      Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
      Cc: Minchan Kim <minchan@kernel.org>
      Cc: Sasha Levin <sasha.levin@oracle.com>
      Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
      Cc: Mel Gorman <mgorman@suse.de>
      Cc: Michal Hocko <mhocko@suse.com>
      Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
      Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      14e0a214
  13. 12 3月, 2016 1 次提交
    • S
      perf test: Remove 'core_id' check in topo test · 4c9d6c18
      Sukadev Bhattiprolu 提交于
      The topology test case of 'perf test' seems to be broken on my x86
      system - due to the comparison of a "core-id" with # of CPUs online.
      
      There are 8 online CPUs:
      
      	$ cat /sys/devices/system/cpu/online
      	0-7
      
      but core-ids are not sequential and some core-ids exceed the number
      of online CPUs.
      
      	$ cat /sys/devices/system/cpu/cpu?/topology/core_id
      	0
      	1
      	9
      	10
      	0
      	1
      	9
      	10
      
      Looks like we can safely remove the check.  Output before:
      
      	$ perf --version
      	perf version 4.4.rc1.g34258a
      
      	$ perf test -v topo
      	36: Test topology in session                                 :
      	--- start ---
      	test child forked, pid 5906
      	templ file: /tmp/perf-test-vCwWG3
      	core_id number is too big.You may need to upgrade the perf tool.
      	test child interrupted
      	---- end ----
      	Test topology in session: FAILED!
      
      and after:
      
      	$ perf test -v topo
      	36: Test topology in session                                 :
      	--- start ---
      	test child forked, pid 6532
      	templ file: /tmp/perf-test-y10wFJ
      	CPU 0, core 0, socket 0
      	CPU 1, core 1, socket 0
      	CPU 2, core 9, socket 0
      	CPU 3, core 10, socket 0
      	CPU 4, core 0, socket 1
      	CPU 5, core 1, socket 1
      	CPU 6, core 9, socket 1
      	CPU 7, core 10, socket 1
      	test child finished with 0
      	---- end ----
      	Test topology in session: Ok
      Signed-off-by: NSukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
      Cc: Jan Stancek <jstancek@redhat.com>
      Cc: Jiri Olsa <jolsa@redhat.com>
      Cc: Kan Liang <kan.liang@intel.com>
      Link: http://lkml.kernel.org/r/20151203233219.GA27696@us.ibm.comSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      4c9d6c18