[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <9f8f8acc-6b0b-4edb-bf69-25d7bbd9c147@intel.com>
Date: Thu, 16 Oct 2025 00:49:06 +0800
From: "Li, Tianyou" <tianyou.li@...el.com>
To: James Clark <james.clark@...aro.org>, Namhyung Kim <namhyung@...nel.org>
CC: Mark Rutland <mark.rutland@....com>, Alexander Shishkin
<alexander.shishkin@...ux.intel.com>, Jiri Olsa <jolsa@...nel.org>, "Ian
Rogers" <irogers@...gle.com>, Adrian Hunter <adrian.hunter@...el.com>, "Kan
Liang" <kan.liang@...ux.intel.com>, Ravi Bangoria <ravi.bangoria@....com>,
<wangyang.guo@...el.com>, <pan.deng@...el.com>, <zhiguo.zhou@...el.com>,
<jiebin.sun@...el.com>, <thomas.falcon@...el.com>, <dapeng1.mi@...el.com>,
<linux-perf-users@...r.kernel.org>, <linux-kernel@...r.kernel.org>, "Peter
Zijlstra" <peterz@...radead.org>, Ingo Molnar <mingo@...hat.com>, "Arnaldo
Carvalho de Melo" <acme@...nel.org>
Subject: Re: [PATCH] perf tools annotate: fix a crash when annotate the same
symbol with 's' and 'T'
Hi James,
Appreciated for your review comments. All great suggestions. Thanks.
On 10/15/2025 8:30 PM, James Clark wrote:
>
>
> On 13/10/2025 5:10 pm, Tianyou Li wrote:
>> When perf report with annotation for a symbol, press 's' and 'T',
>> then exit
>> the annotate browser. Once annoate the same symbol, the annoate browser
>> will crash. Stack trace as below:
>>
>
> Hi Tianyou,
>
> Can you explain why this only happens the second time and what you did
> to fix it in the commit message. It took me a minute to work out.
>
Sure, will update the commit message and fix typos in patch v2.
>> Perf: Segmentation fault
>> -------- backtrace --------
>> #0 0x55d365 in ui__signal_backtrace setup.c:0
>> #1 0x7f5ff1a3e930 in __restore_rt libc.so.6[3e930]
>> #2 0x570f08 in arch__is perf[570f08]
>> #3 0x562186 in annotate_get_insn_location perf[562186]
>> #4 0x562626 in __hist_entry__get_data_type annotate.c:0
>> #5 0x56476d in annotation_line__write perf[56476d]
>> #6 0x54e2db in annotate_browser__write annotate.c:0
>> #7 0x54d061 in ui_browser__list_head_refresh perf[54d061]
>> #8 0x54dc9e in annotate_browser__refresh annotate.c:0
>> #9 0x54c03d in __ui_browser__refresh browser.c:0
>> #10 0x54ccf8 in ui_browser__run perf[54ccf8]
>> #11 0x54eb92 in __hist_entry__tui_annotate perf[54eb92]
>> #12 0x552293 in do_annotate hists.c:0
>> #13 0x55941c in evsel__hists_browse hists.c:0
>> #14 0x55b00f in evlist__tui_browse_hists perf[55b00f]
>> #15 0x42ff02 in cmd_report perf[42ff02]
>> #16 0x494008 in run_builtin perf.c:0
>> #17 0x494305 in handle_internal_command perf.c:0
>> #18 0x410547 in main perf[410547]
>> #19 0x7f5ff1a295d0 in __libc_start_call_main libc.so.6[295d0]
>> #20 0x7f5ff1a29680 in __libc_start_main@@GLIBC_2.34
>> libc.so.6[29680]
>> #21 0x410b75 in _start perf[410b75]
>>
>
> This should have a fixes tag. Maybe commit 1d4374afd000 ("perf
> annotate: Add 'T' hot key to toggle data type display")? Or earlier if
> only 's' and not 's' and 'T' are required to reproduce.
>
My apology. Will add the fixes tag in patch v2.
>> Signed-off-by: Tianyou Li <tianyou.li@...el.com>
>> ---
>> tools/perf/ui/browsers/annotate.c | 3 +++
>> tools/perf/util/annotate.c | 2 +-
>> tools/perf/util/annotate.h | 2 ++
>> 3 files changed, 6 insertions(+), 1 deletion(-)
>>
>> diff --git a/tools/perf/ui/browsers/annotate.c
>> b/tools/perf/ui/browsers/annotate.c
>> index 8fe699f98542..1e0873194217 100644
>> --- a/tools/perf/ui/browsers/annotate.c
>> +++ b/tools/perf/ui/browsers/annotate.c
>> @@ -1163,6 +1163,9 @@ int __hist_entry__tui_annotate(struct
>> hist_entry *he, struct map_symbol *ms,
>> }
>> }
>> + if (browser.arch == NULL)
>> + evsel__get_arch(evsel, &browser.arch);
>> +
>
> This technically only needs to be called if symbol_annotate2() doesn't
> get called. So it could be on an "else" after "if (not_annotated ||
> !sym->annotate2)".
>
I have a local version use the 'else' but I thought it might be clearer
by make a NULL check then call the evsel__get_arch. Your suggestion
definitely make sense especially the if/else statement could serve the
whole purpose of initializing the necessary data structures. I will
update the code in patch v2.
> I looked to see if it's better to save the arch in *notes, seeing as
> that's where the annotation is cached, but it doesn't feel any nicer
> than the way you've done it.
>
> You don't need to check the return value for errors because we only
> get here if it worked previously and dso__annotate_warned() is false.
> But it might be worth a comment or checking it anyway to avoid doubt.
>
Yes, I do need to check the return value. Appreciated for your kind
words. I will update the code in patch v2. Thanks.
> Other than that, looks good.
>
> Thanks
> James
>
>> /* Copy necessary information when it's called from perf top */
>> if (hbt != NULL && he != &annotate_he) {
>> annotate_he.hists = he->hists;
>> diff --git a/tools/perf/util/annotate.c b/tools/perf/util/annotate.c
>> index a2e34f149a07..39d6594850f1 100644
>> --- a/tools/perf/util/annotate.c
>> +++ b/tools/perf/util/annotate.c
>> @@ -980,7 +980,7 @@ void symbol__calc_percent(struct symbol *sym,
>> struct evsel *evsel)
>> annotation__calc_percent(notes, evsel, symbol__size(sym));
>> }
>> -static int evsel__get_arch(struct evsel *evsel, struct arch **parch)
>> +int evsel__get_arch(struct evsel *evsel, struct arch **parch)
>> {
>> struct perf_env *env = evsel__env(evsel);
>> const char *arch_name = perf_env__arch(env);
>> diff --git a/tools/perf/util/annotate.h b/tools/perf/util/annotate.h
>> index eaf6c8aa7f47..d4990bff29a7 100644
>> --- a/tools/perf/util/annotate.h
>> +++ b/tools/perf/util/annotate.h
>> @@ -585,4 +585,6 @@ void debuginfo_cache__delete(void);
>> int annotation_br_cntr_entry(char **str, int br_cntr_nr, u64 *br_cntr,
>> int num_aggr, struct evsel *evsel);
>> int annotation_br_cntr_abbr_list(char **str, struct evsel *evsel,
>> bool header);
>> +
>> +int evsel__get_arch(struct evsel *evsel, struct arch **parch);
>> #endif /* __PERF_ANNOTATE_H */
>
>
Powered by blists - more mailing lists