From: Arnaldo Carvalho de Melo <acme@kernel.org>
To: Jonah Petri <jpetri@izotope.com>
Cc: linux-rt-users@vger.kernel.org
Subject: Re: diagnosing unexpected sched_stat_wait
Date: Fri, 4 Mar 2016 19:05:31 -0300 [thread overview]
Message-ID: <20160304220531.GB3604@kernel.org> (raw)
In-Reply-To: <CAJOrw7was05Absp0UiCVO5mZEoZqXHD-ZNOoDwoG7t1S1DED_A@mail.gmail.com>
Em Fri, Mar 04, 2016 at 02:15:25PM -0500, Jonah Petri escreveu:
> Hello,
>
> My application is doing realtime low-latency audio, and we're seeing
> occasional dropouts, which we've traced (via ftrace) to an unexplained
> sched_stat_wait event on our audio-handling user thread.
>
> We're really not sure what might be causing this, but if I could get
> the user stack trace causing this event, it would let me see exactly
> the problem. Even the top frame of the stack would be invaluable.
>
> Is there a way to get this info? Preferrably with ftrace, but any
> methods or hints would be much appreciated.
[root@jouet ~]# perf list *:sched_stat_wait
List of pre-defined events (to be used in -e):
sched:sched_stat_wait [Tracepoint event]
[root@jouet ~]# perf record -a --call-graph dwarf -e sched:*stat_wait
^C[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 4.157 MB perf.data (763 samples) ]
[root@jouet ~]#
[root@jouet ~]# perf report --stdio | head -53
Failed to open /tmp/perf-7459.map, continuing without symbols
no symbols found in /usr/sbin/rngd, maybe install a debug package?
# To display the perf.data header info, please use --header/--header-only options.
#
#
# Total Lost Samples: 0
#
# Samples: 763 of event 'sched:sched_stat_wait'
# Event count (approx.): 2027097
#
# Children Self Trace output
# ........ ........ ...............................................
#
8.94% 8.94% comm=Xorg pid=2730 delay=181323 [ns]
|
---entry_SYSCALL_64_fastpath
sys_poll
do_sys_poll
poll_schedule_timeout
schedule_hrtimeout_range
schedule_hrtimeout_range_clock
schedule
__schedule
pick_next_task_fair
set_next_entity
update_stats_wait_end
0x563bfb142909
__libc_start_main
main
meta_run
g_main_loop_run
g_main_context_iterate.isra.29
0x7fb1060a9fdd
4.73% 4.73% comm=gnome-terminal- pid=3412 delay=95977 [ns]
|
---entry_SYSCALL_64_fastpath
sys_poll
do_sys_poll
poll_schedule_timeout
schedule_hrtimeout_range
schedule_hrtimeout_range_clock
schedule
__schedule
pick_next_task_fair
set_next_entity
update_stats_wait_end
_start
__libc_start_main
main
g_application_run
g_main_context_iteration
g_main_context_iterate.isra.29
0x7fd2cbd9bfdd
[root@jouet ~]#
Replace 'record -a' with 'record --pid YOUR-PID' or workload, etc, see:
man perf-record
If you get some unresolved symbols, install 'gcc -g' built binaries or
-debuginfo packages or equivalent in !fedora/RHEL distros.
- Arnaldo
next prev parent reply other threads:[~2016-03-04 22:05 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-03-04 19:15 diagnosing unexpected sched_stat_wait Jonah Petri
2016-03-04 22:05 ` Arnaldo Carvalho de Melo [this message]
2016-03-04 22:24 ` Jonah Petri
2016-03-07 14:28 ` Arnaldo Carvalho de Melo
2016-03-07 14:53 ` Jonah Petri
2016-03-09 18:30 ` Jonah Petri
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=20160304220531.GB3604@kernel.org \
--to=acme@kernel.org \
--cc=jpetri@izotope.com \
--cc=linux-rt-users@vger.kernel.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.