Re: [PATCHSET 0/8] perf tools: Fix scalability problem on callchain merging (v5)

[Namhyung Kim <namhyung@kernel.org>]

Hi Jiri,

On Sun, 13 Oct 2013 14:34:44 +0200, Jiri Olsa wrote:
> On Fri, Oct 11, 2013 at 02:15:35PM +0900, Namhyung Kim wrote:
>> Hello,
>> 
>> This is a new version of callchain improvement patchset.  Basically
>> it's almost same as v4 but rebased on current acme/perf/core and some
>> functions are renamed as Frederic requested.
>> 
>> Now I'm hunting down a bug in 'perf report -s sym' which was found
>> during the test, but I think it's not related to this change as it can
>> be reproduced in earlier versions too.
>> 
>> I put this series on 'perf/callchain-v5' branch in my tree
>> 
>>   git://git.kernel.org/pub/scm/linux/kernel/git/namhyung/linux-perf.git
> hi,
> while hunting another issue, I got segfault in callchains code:
>
> Press '?' for help on key bindings
> Program received signal SIGSEGV, Segmentation fault.
> 0x00000000004dcd89 in callchain_node__init_have_children_rb_tree (node=0x21a0330) at ui/browsers/hists.c:158
> 158                                     chain->ms.has_children = chain->list.next == &child->val &&
> Missing separate debuginfos, use: debuginfo-install
> audit-libs-2.2.1-1.fc16.x86_64 elfutils-libelf-0.154-2.fc16.x86_64
> elfutils-libs-0.154-2.fc16.x86_64 glibc-2.14.90-24.fc16.9.x86_64
> libunwind-0.99-2.20110424git1e10c293.fc16.x86_64
> nss-softokn-freebl-3.14.1-3.fc16.x86_64 numactl-2.0.7-2.fc16.x86_64
> perl-libs-5.14.3-205.fc16.x86_64 python-libs-2.7.3-4.fc16.x86_64
> slang-2.2.4-1.fc16.x86_64 xz-libs-5.1.1-1alpha.fc16.x86_64
> zlib-1.2.5-7.fc16.x86_64
> (gdb) bt
> #0  0x00000000004dcd89 in callchain_node__init_have_children_rb_tree (node=0x21a0330) at ui/browsers/hists.c:158
> #1  0x00000000004dcdf9 in callchain_node__init_have_children_rb_tree (node=0x1fa90a0) at ui/browsers/hists.c:162
> #2  0x00000000004dcead in callchain_node__init_have_children (node=0x1fa90a0) at ui/browsers/hists.c:173
> #3  0x00000000004dcf10 in callchain__init_have_children (root=0x1fa8ff8) at ui/browsers/hists.c:182
> #4  0x00000000004dcf97 in hist_entry__init_have_children (he=0x1fa8f00) at ui/browsers/hists.c:190
> #5  0x00000000004debf3 in hist_browser__show_entry (browser=0xa27010, entry=0x1fa8f00, row=66) at ui/browsers/hists.c:739
> #6  0x00000000004df062 in hist_browser__refresh (browser=0xa27010) at ui/browsers/hists.c:823
> #7  0x00000000004d8487 in __ui_browser__refresh (browser=0xa27010) at ui/browser.c:307
> #8  0x00000000004d8681 in ui_browser__run (browser=0xa27010, delay_secs=0) at ui/browser.c:362
> #9  0x00000000004dd6a7 in hist_browser__run (browser=0xa27010, ev_name=0xa26ff0 "cycles:pp", hbt=0x0) at ui/browsers/hists.c:335
> #10 0x00000000004e0d9b in perf_evsel__hists_browse (evsel=0xa24b60, nr_events=1, 
>     helpline=0x58e588 "For a higher level overview, try: perf report --sort comm,dso", ev_name=0xa26ff0 "cycles:pp", left_exits=false, 
>     hbt=0x0, min_pcnt=0, env=0xa23f90) at ui/browsers/hists.c:1430
> #11 0x00000000004e2605 in perf_evlist__tui_browse_hists (evlist=0xa242e0, 
>     help=0x58e588 "For a higher level overview, try: perf report --sort comm,dso", hbt=0x0, min_pcnt=0, env=0xa23f90)
>     at ui/browsers/hists.c:1950
> #12 0x000000000042c3cb in __cmd_report (rep=0x7fffffffdf40) at builtin-report.c:590
> #13 0x000000000042d84a in cmd_report (argc=0, argv=0x7fffffffe3a0, prefix=0x0) at builtin-report.c:988
> #14 0x0000000000418fed in run_builtin (p=0x803c28, argc=5, argv=0x7fffffffe3a0) at perf.c:319
> #15 0x0000000000419225 in handle_internal_command (argc=5, argv=0x7fffffffe3a0) at perf.c:376
> #16 0x0000000000419371 in run_argv (argcp=0x7fffffffe27c, argv=0x7fffffffe270) at perf.c:420
> #17 0x0000000000419658 in main (argc=5, argv=0x7fffffffe3a0) at perf.c:529
>
>
> I got it on the original code, so I've tried your's
> perf/callchain-v5 with same result.
>
> I put perf archive and data output in here if you're interested:
> http://people.redhat.com/jolsa/cc/

Thanks, I'll take a look at this today.
Namhyung