1. 04 Jun, 2009 7 commits
    • Peter Zijlstra's avatar
      perf report: Simplify symbol output · 95ed6fd0
      Peter Zijlstra authored
      The DSO can be printed already - no need to repeat it in the
      symbol field.
      Signed-off-by: default avatarPeter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      95ed6fd0
    • Ingo Molnar's avatar
      perf_counter tools: Build with native optimization · af794b94
      Ingo Molnar authored
      Build the tools with -march=native by default.
      
      No measurable difference in speed though, compared to the
      default, on a Nehalem testbox.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      af794b94
    • Ingo Molnar's avatar
      perf record/report: Fix PID/COMM handling · df97992c
      Ingo Molnar authored
      Fix two bugs causing lost comm mappings:
      
       - initial PID is not 0 but getpid()
      
       - when we are unable to handle an mmap event, dont assume the event
         itself is broken - try to parse the stream. This way we wont lose
         comm events.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      df97992c
    • Yong Wang's avatar
      perf_counter: Documentation update · 3aff27ca
      Yong Wang authored
      The 'nmi' bit is no longer there.
      Signed-off-by: default avatarYong Wang <yong.y.wang@intel.com>
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Arjan van de Ven <arjan@linux.intel.com>
      LKML-Reference: <20090603084225.GA6553@ywang-moblin2.bj.intel.com>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      3aff27ca
    • Paul Mackerras's avatar
      perf_counter: powerpc: Use new identifier names in powerpc-specific code · 1b58c251
      Paul Mackerras authored
      Commit b23f3325 ("perf_counter: Rename various fields") fixed up
      most of the uses of the renamed fields, but missed one instance
      of "record_type" in powerpc-specific code which needs to be changed
      to "sample_type", and a "PERF_RECORD_ADDR" in the same statement that
      needs to be changed to "PERF_SAMPLE_ADDR", causing compilation
      errors on powerpc.  This fixes it.
      Signed-off-by: default avatarPaul Mackerras <paulus@samba.org>
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      LKML-Reference: <18983.3111.770392.800486@cargo.ozlabs.ibm.com>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      1b58c251
    • Ingo Molnar's avatar
      perf top: Reduce default filter threshold · 6e53cdf1
      Ingo Molnar authored
      On idle systems 'perf top' comes up empty by default, because the event
      count filter is set to 100.
      
      Reduce it to 5 instead.
      
      Also add an option to limit the number of functions displayed.
      Reported-by: default avatarSteven Rostedt <rostedt@goodmis.org>
      Acked-by: default avatarPeter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      6e53cdf1
    • Arnaldo Carvalho de Melo's avatar
      perf report: Fix rbtree bug · a4c43bea
      Arnaldo Carvalho de Melo authored
      Ingo Molnar reported:
      
      > FYI, i just got this crash (segfault) in perf report after
      > collecting a long profile from Xorg:
      >
      > Starting program: /home/mingo/tip/Documentation/perf_counter/perf report
      > [Thread debugging using libthread_db enabled]
      > Detaching after fork from child process 20008.
      > [New Thread 0x7f92fd62a6f0 (LWP 20005)]
      >
      > Program received signal SIGSEGV, Segmentation fault.
      > 0x000000000041031a in __rb_erase_color (node=0x142c0901, parent=0x0,
      > root=0x881918)
      >     at util/rbtree.c:143
      > 143			if (parent->rb_left == node)
      
      It was a problem introduced in this cset:
      
       perf report: Fix comm sorting - 8229289b
      
      This patch should fix it.
      Signed-off-by: default avatarArnaldo Carvalho de Melo <acme@redhat.com>
      Acked-by: default avatarPeter Zijlstra <peterz@infradead.org>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: Stephane Eranian <eranian@googlemail.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      a4c43bea
  2. 03 Jun, 2009 24 commits
    • Ingo Molnar's avatar
      perf report: Handle all known event types · d11444df
      Ingo Molnar authored
      We have munmap, throttle/unthrottle and period events as well,
      process them - otherwise they are considered broke events and
      we mis-parse the next few events.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      d11444df
    • Ingo Molnar's avatar
      perf report: Split out event processing helpers · 75051724
      Ingo Molnar authored
      - Introduce per event helper functions
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      75051724
    • Ingo Molnar's avatar
      perf report: Clean up event processing · d80d338d
      Ingo Molnar authored
      - Split out event processig into process_events() helper.
      
      - Untangle the cwd parameters - it's constant so can be a static.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      d80d338d
    • Ingo Molnar's avatar
      perf_counter: Fix throttling lock-up · 128f048f
      Ingo Molnar authored
      Throttling logic is broken and we can lock up with too small
      hw sampling intervals.
      
      Make the throttling code more robust: disable counters even
      if we already disabled them.
      
      ( Also clean up whitespace damage i noticed while reading
        various pieces of code related to throttling. )
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      128f048f
    • Ingo Molnar's avatar
      perf_counter tools: Work around warnings in older GCCs · 233f0b95
      Ingo Molnar authored
      GCC 4.1.2 produces:
      
       util/parse-options.c: In function ‘get_value’:
       util/parse-options.c:36: warning: ‘arg’ may be used uninitialized in this function
      
       builtin-top.c: In function ‘display_thread’:
       builtin-top.c:178: warning: ‘printed’ may be used uninitialized in this function
      
      Annotate them away by initializing these variables to 0.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      233f0b95
    • Arnaldo Carvalho de Melo's avatar
      perf_counter tools: Fix off-by-one bug in symbol__new · 18374ab7
      Arnaldo Carvalho de Melo authored
      The end is really (start + len - 1). Noticed when synthesizing
      the PLT symbols, that are small (16 bytes), and hot on the
      start RIP.
      Signed-off-by: default avatarArnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Peter Zijlstra <peterz@infradead.org>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <20090603174921.GG7805@ghostprotocols.net>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      18374ab7
    • Ingo Molnar's avatar
      perf_counter tools: Optimize harder · 095b3a6a
      Ingo Molnar authored
      Use -O6 to build the tools.
      
      Before:
      
          12387507370  instructions         #    3121.653 M/sec
      
      After:
      
           6244894971  instructions         #    3458.437 M/sec
      
      Almost twice as fast!
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      095b3a6a
    • Ingo Molnar's avatar
      perf help: Fix bug when there's no perf-* command around · 051cdc3c
      Ingo Molnar authored
      main_cmds can be empty - fix util/help.c to handle this case
      without segfaulting.
      Reported-by: default avatarSteven Rostedt <rostedt@goodmis.org>
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      051cdc3c
    • Ingo Molnar's avatar
      perf report: Add front-entry cache for lookups · eed4dcd4
      Ingo Molnar authored
      Before:
      
       Performance counter stats for './perf report -i perf.data.big':
      
           12453988058  instructions
      
       Performance counter stats for './perf report -i perf.data.big':
      
           12379566017  instructions
      
      0.60% reduction.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      eed4dcd4
    • Ingo Molnar's avatar
      perf stat: Print out all arguments · 44db76c8
      Ingo Molnar authored
      Before:
      
       Performance counter stats for '/home/mingo/hackbench':
      
      After:
      
       Performance counter stats for '/home/mingo/hackbench 10':
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      44db76c8
    • Ingo Molnar's avatar
      perf report: Display 100% correctly · e98e96fe
      Ingo Molnar authored
      Needs to be 6.2 not 5.2, for 100.00% to be aligned properly.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      e98e96fe
    • Ingo Molnar's avatar
      perf record: Refine capture printout · 021e9f47
      Ingo Molnar authored
      Print out the number of bytes captured, and the (estimated) number of
      events the output file contains.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      021e9f47
    • Ingo Molnar's avatar
      perf_counter tools: Clean up old kerneltop references · f2521b6e
      Ingo Molnar authored
      kerneltop has been replaced with perf top - so fix up a few
      remaining references to it in display text and error messages.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      f2521b6e
    • Peter Zijlstra's avatar
      perf_counter: Fix race in counter initialization · a96bbc16
      Peter Zijlstra authored
      We need the PID namespace and counter ID available when the
      counter overflows and we need to generate a sample event.
      
      [ Impact: fix kernel crash with high-frequency sampling ]
      Reported-by: default avatarIngo Molnar <mingo@elte.hu>
      Signed-off-by: default avatarPeter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <new-submission>
      [ fixed a further crash and cleaned up the initialization a bit ]
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      a96bbc16
    • Peter Zijlstra's avatar
      perf report: Fix comm sorting · 8229289b
      Peter Zijlstra authored
      Since we can (and do) change comm strings during the collection
      phase, we cannot actually sort on them to build the histogram.
      Therefore add an (optional) third sorting phase to collapse the
      histrogram.
      
      Comm sorting now builds the histrogram on threads and then in
      the collapse phase collects all threads with the same comm.
      
      This collapsed histogram is then reversed and sorted on events.
      Signed-off-by: default avatarPeter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      8229289b
    • Peter Zijlstra's avatar
      perf record: Use long arg for counter period · e61078a0
      Peter Zijlstra authored
      I wrote this to test the extended period emulation, we might as
      well merge it.
      Signed-off-by: default avatarPeter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      e61078a0
    • Peter Zijlstra's avatar
      perf_counter: Add a comm hook for pure fork()s · 226f62fd
      Peter Zijlstra authored
      I noticed missing COMM events and found that we missed
      reporting them for pure forks.
      Signed-off-by: default avatarPeter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      226f62fd
    • Paul Mackerras's avatar
      perf_counter: powerpc: Fix race causing "oops trying to read PMC0" errors · dcd945e0
      Paul Mackerras authored
      When using interrupting counters and limited (non-interrupting)
      counters at the same time, it's possible that we get an
      interrupt in write_mmcr0() after writing MMCR0 but before we
      have set up the counters using limited PMCs.  What happens then
      is that we get into perf_counter_interrupt() with
      counter->hw.idx = 0 for the limited counters, leading to the
      "oops trying to read PMC0" error message being printed.
      
      This fixes the problem by making perf_counter_interrupt()
      robust against counter->hw.idx being zero (the counter is just
      ignored in that case) and also by changing write_mmcr0() to
      write MMCR0 initially with the counter overflow interrupt
      enable bits masked (set to 0).  If the MMCR0 value requested by
      the caller has either of those bits set, we write MMCR0 again
      with the requested value of those bits after setting up the
      limited counters properly.
      Signed-off-by: default avatarPaul Mackerras <paulus@samba.org>
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      Cc: Stephane Eranian <eranian@googlemail.com>
      LKML-Reference: <18982.17684.138182.954599@cargo.ozlabs.ibm.com>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      dcd945e0
    • Paul Mackerras's avatar
      perf_counter: powerpc: Fix event alternative code generation on POWER5/5+ · 6984efb6
      Paul Mackerras authored
      Commit ef923214 ("perf_counter: powerpc: use u64 for event
      codes internally") introduced a bug where the return value from
      function find_alternative_bdecode gets put into a u64 variable
      and later tested to see if it is < 0.  The effect is that we
      get extra, bogus event code alternatives on POWER5 and POWER5+,
      leading to error messages such as "oops compute_mmcr failed"
      being printed and counters not counting properly.
      
      This fixes it by using s64 for the return type of
      find_alternative_bdecode and for the local variable that the
      caller puts the value in.  It also makes the event argument a
      u64 on POWER5+ for consistency.
      Signed-off-by: default avatarPaul Mackerras <paulus@samba.org>
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      Cc: Stephane Eranian <eranian@googlemail.com>
      LKML-Reference: <18982.17586.666132.90983@cargo.ozlabs.ibm.com>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      6984efb6
    • Ingo Molnar's avatar
      perf report: Handle vDSO symbols properly · ed966aac
      Ingo Molnar authored
      We were not looking up vDSO symbols properly, because they
      are in the kallsyms but are user-mode entries.
      
      Pass negative addresses to the kernel dso object, this
      way we resolve them properly:
      
           0.05%  [kernel]: vread_tsc
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      ed966aac
    • Ingo Molnar's avatar
      perf report: Improve sort key recognition · 5352f35d
      Ingo Molnar authored
       - allow case-insensitive tokens - such as --sort Comm,Symbol
       - allow substring shortcuts: --sort sym
       - detect invalid tokens and bail out
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      5352f35d
    • Ingo Molnar's avatar
      perf report: Print -D to stdout · 3502973d
      Ingo Molnar authored
      -D prints to stderr - which is a bit confusing - print to stdout
      instead.
      
      Also clean up the if (dump_trace) patterns via a dprintf helper.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      3502973d
    • Arnaldo Carvalho de Melo's avatar
      perf_counter tools: Cover PLT symbols too · 8ce998d6
      Arnaldo Carvalho de Melo authored
      PLT, the Program Linking Table, is used with the dynamic linker to
      allow PIC code in executables and shared objects to figure out
      where functions are in other shared objects.
      
      It is one of the sources of unknown/unresolved symbols - this patch
      does what binutils figures out when you ask it to disassembly.
      (objdump -S)
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: John Kacur <jkacur@redhat.com>
      Cc: Stephane Eranian <eranian@googlemail.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      8ce998d6
    • Yong Wang's avatar
      perf_counter/x86: Remove the IRQ (non-NMI) handling bits · a3288106
      Yong Wang authored
      Remove the IRQ (non-NMI) handling bits as NMI will be used always.
      Signed-off-by: default avatarYong Wang <yong.y.wang@intel.com>
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <20090603051255.GA2791@ywang-moblin2.bj.intel.com>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      a3288106
  3. 02 Jun, 2009 9 commits
    • Ingo Molnar's avatar
      perf record: Print out the number of events captured · addc2785
      Ingo Molnar authored
      It makes sense to inform the user about how many events
      perf record has written - so that the sufficiency of
      profiling coverage and intensity can be determined at
      a glance.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      addc2785
    • Ingo Molnar's avatar
      perf_counter tools: Make source code headers more coherent · bf9e1876
      Ingo Molnar authored
      The perf commands had different ways of describing themselves,
      introduce a coherent command-file-header format taken from the
      Git project.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      bf9e1876
    • Ingo Molnar's avatar
      perf report: Print more info instead of <unknown> entries · 0a520c63
      Ingo Molnar authored
      Sometimes we still fail to find a DSO or look up a symbol,
      print out the raw information in this case (which an help
      debug the problem), instead of a not very helpful <unknown>
      string.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      0a520c63
    • Ingo Molnar's avatar
      perf record: Increase mmap buffering default · 3cf165fc
      Ingo Molnar authored
      I've run into mmap overruns with the current 16 pages default,
      increase it to 128 pages.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      3cf165fc
    • Ingo Molnar's avatar
      perf record: Add --append option · abaff32a
      Ingo Molnar authored
      Allow incremental profiling via 'perf record -A' - this will append
      to an existing perf.data.
      
      Also reorder perf record options by utility / likelyhood of usage.
      
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      abaff32a
    • Ingo Molnar's avatar
      perf report: Fix column width/alignment of dsos · cf25c63c
      Ingo Molnar authored
      Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      cf25c63c
    • Peter Zijlstra's avatar
      perf report: Separate out idle threads · 436224a6
      Peter Zijlstra authored
      Introduce the special comm name [idle] for idle theads.
      Signed-off-by: default avatarPeter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      436224a6
    • Peter Zijlstra's avatar
      perf_counter tools: Fix up the ABI shakeup · c70975bc
      Peter Zijlstra authored
      Signed-off-by: default avatarPeter Zijlstra <a.p.zijlstra@chello.nl>
      LKML-Reference: <new-submission>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      Cc: Stephane Eranian <eranian@googlemail.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      c70975bc
    • Peter Zijlstra's avatar
      perf_counter: Rename perf_counter_hw_event => perf_counter_attr · 0d48696f
      Peter Zijlstra authored
      The structure isn't hw only and when I read event, I think about those
      things that fall out the other end. Rename the thing.
      Signed-off-by: default avatarPeter Zijlstra <a.p.zijlstra@chello.nl>
      Cc: Mike Galbraith <efault@gmx.de>
      Cc: Paul Mackerras <paulus@samba.org>
      Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
      Cc: Marcelo Tosatti <mtosatti@redhat.com>
      Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: John Kacur <jkacur@redhat.com>
      Cc: Stephane Eranian <eranian@googlemail.com>
      LKML-Reference: <new-submission>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      0d48696f