From: Yang Jihong <yangjihong1@huawei.com>
To: <peterz@infradead.org>, <mingo@redhat.com>, <acme@kernel.org>,
<namhyung@kernel.org>, <mark.rutland@arm.com>,
<alexander.shishkin@linux.intel.com>, <jolsa@kernel.org>,
<irogers@google.com>, <adrian.hunter@intel.com>,
<linux-perf-users@vger.kernel.org>,
<linux-kernel@vger.kernel.org>
Cc: <yangjihong1@huawei.com>
Subject: [PATCH 0/3] perf record: Fix segfault with '--timestamp-filename' option and pipe mode
Date: Fri, 19 Jan 2024 04:03:01 +0000 [thread overview]
Message-ID: <20240119040304.3708522-1-yangjihong1@huawei.com> (raw)
When perf record uses '--timestamp-filename' option and pipe mode,
will occur segfault:
# perf record --timestamp-filename -o- true 1>/dev/null
[ perf record: Woken up 1 times to write data ]
[ perf record: Dump -.2024011813361681 ]
perf: Segmentation fault
Segmentation fault (core dumped)
Debug the core file by using the gdb:
# gdb perf core.3706841
<SNIP>
[New LWP 3706841]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Core was generated by `perf record --timestamp-filename -o- true'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0 evsel__free_stat_priv (evsel=0x555ea56197af) at util/stat.c:145
145 struct perf_stat_evsel *ps = evsel->stats;
(gdb) bt
#0 evsel__free_stat_priv (evsel=0x555ea56197af) at util/stat.c:145
#1 evlist__free_stats (evlist=evlist@entry=0x555bf0dad1b0) at util/stat.c:215
#2 0x0000555befd425e8 in evlist__delete (evlist=0x555bf0dad1b0) at util/evlist.c:175
#3 0x0000555befd42785 in evlist__delete (evlist=<optimized out>) at util/evlist.c:181
#4 0x0000555befc7f547 in cmd_record (argc=<optimized out>, argv=<optimized out>) at builtin-record.c:4252
#5 0x0000555befd27700 in run_builtin (p=p@entry=0x555bf05acf88 <commands+264>, argc=argc@entry=4, argv=argv@entry=0x7ffc2696a920) at perf.c:349
#6 0x0000555befc68751 in handle_internal_command (argv=0x7ffc2696a920, argc=4) at perf.c:402
#7 run_argv (argv=<synthetic pointer>, argcp=<synthetic pointer>) at perf.c:446
#8 main (argc=4, argv=0x7ffc2696a920) at perf.c:562
It is found that the memory of the evsel is modified.
The reason is that perf_data__switch() not initialized 'new_filename',
as a result, it uses the on-stack variable. It happens that the evsel
address is stored here. 'new_filename' is freed later when uses
'--timestamp-filename' option. Therefore, segfault occurs in the evsel_delete().
1. patch1 fixes this problem.
2. patch2 to optimize the process, check conflict between
'--timestamp-filename' option and the pipe mode before recording to
avoid switch perf data.
3. patch3 fixes the code style problem by the way.
Yang Jihong (3):
perf record: Fix possible incorrect free in record__switch_output()
perf record: Check conflict between '--timestamp-filename' option and
pipe mode before recording
perf data: Minor code style alignment cleanup
tools/perf/builtin-record.c | 14 ++++++++++----
tools/perf/util/data.c | 10 ++++------
tools/perf/util/data.h | 6 +++---
3 files changed, 17 insertions(+), 13 deletions(-)
--
2.34.1
next reply other threads:[~2024-01-19 4:07 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-01-19 4:03 Yang Jihong [this message]
2024-01-19 4:03 ` [PATCH 1/3] perf record: Fix possible incorrect free in record__switch_output() Yang Jihong
2024-01-19 4:03 ` [PATCH 2/3] perf record: Check conflict between '--timestamp-filename' option and pipe mode before recording Yang Jihong
2024-01-19 4:03 ` [PATCH 3/3] perf data: Minor code style alignment cleanup Yang Jihong
2024-01-19 22:06 ` [PATCH 0/3] perf record: Fix segfault with '--timestamp-filename' option and pipe mode Namhyung Kim
2024-01-22 20:38 ` Namhyung Kim
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240119040304.3708522-1-yangjihong1@huawei.com \
--to=yangjihong1@huawei.com \
--cc=acme@kernel.org \
--cc=adrian.hunter@intel.com \
--cc=alexander.shishkin@linux.intel.com \
--cc=irogers@google.com \
--cc=jolsa@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mark.rutland@arm.com \
--cc=mingo@redhat.com \
--cc=namhyung@kernel.org \
--cc=peterz@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®