From: Ingo Molnar <mingo@elte.hu>
To: Frederic Weisbecker <fweisbec@gmail.com>,
Oleg Nesterov <oleg@redhat.com>,
Andrew Morton <akpm@linux-foundation.org>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>,
Zhaolei <zhaolei@cn.fujitsu.com>,
Steven Rostedt <rostedt@goodmis.org>,
Tom Zanussi <tzanussi@gmail.com>,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 3/4] ftrace: add max execution time mesurement to workqueue tracer
Date: Mon, 13 Apr 2009 23:21:59 +0200 [thread overview]
Message-ID: <20090413212159.GA8514@elte.hu> (raw)
In-Reply-To: <20090413161649.GH5977@nowhere>
(Oleg, Andrew: it's about workqueue tracing design.)
* Frederic Weisbecker <fweisbec@gmail.com> wrote:
> > if (tsk) {
> > - seq_printf(s, "%3d %6d %6u %s\n", cws->cpu,
> > + seq_printf(s, "%3d %6d %6u %5lu.%06lu"
> > + " %s\n",
> > + cws->cpu,
> > atomic_read(&cws->inserted), cws->executed,
> > + exec_secs, exec_usec_rem,
>
>
> You are measuring the latency from a workqueue thread point of
> view. While I find the work latency measurement very interesting,
> I think this patch does it in the wrong place. The _work_ latency
> point of view seems to me much more rich as an information source.
>
> There are several reasons for that.
>
> Indeed this patch is useful for workqueues that receive always the
> same work to perform so that you can find very easily the guilty
> worklet. But the sense of this design is lost once we consider the
> workqueue threads that receive random works. Of course the best
> example is events/%d One will observe the max latency that
> happened on event/0 as an exemple but he will only be able to feel
> a silent FUD because he has no way to find which work caused this
> max latency.
Expanding the trace view in a per worklet fashion is also useful for
debugging: sometimes inefficiencies (or hangs) are related to the
mixing of high-speed worklets with blocking worklets. This is not
exposed if we stay at the workqueue level only.
> Especially the events/%d latency measurement seems to me very
> important because a single work from a random driver can propagate
> its latency all over the system.
>
> A single work that consumes too much cpu time, waits for long
> coming events, sleeps too much, tries to take too often contended
> locks, or whatever... such single work may delay all pending works
> in the queue and the only max latency for a given workqueue is not
> helpful to find these culprits.
>
> Having this max latency snapshot per work and not per workqueue
> thread would be useful for every kind of workqueue latency
> instrumentation:
>
> - workqueues with single works
> - workqueue with random works
>
> A developer will also be able to measure its own worklet action
> and find if it takes too much time, even if it isn't the worst
> worklet in the workqueue to cause latencies.
>
> The end result would be to have a descending latency sort of works
> per cpu workqueue threads (or better: per workqueue group).
>
> What do you think?
Sounds like a good idea to me. It would also allow histograms based
on worklet identity, etc. Often the most active kevents worklet
should be considered to be split out as a new workqueue.
And if we have a per worklet tracepoint it would also allow a trace
filter to only trace a given type of worklet.
Ingo
next prev parent reply other threads:[~2009-04-13 21:22 UTC|newest]
Thread overview: 94+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-04-13 0:53 [PATCH 0/1] tracing, workqueuetrace: Make workqueue tracepoints use TRACE_EVENT macro Zhaolei
2009-04-13 0:54 ` [PATCH 1/1] " Zhaolei
2009-04-13 1:44 ` KOSAKI Motohiro
2009-04-13 3:25 ` KOSAKI Motohiro
2009-04-13 3:49 ` Zhaolei
2009-04-13 3:57 ` Ingo Molnar
2009-04-13 4:07 ` Re: [PATCH 1/1] tracing, workqueuetrace: Make workqueuetracepoints " Zhaolei
2009-04-13 3:57 ` [PATCH 1/1] tracing, workqueuetrace: Make workqueue tracepoints " KOSAKI Motohiro
2009-04-13 5:52 ` [PATCH v2 1/4] ftrace, " KOSAKI Motohiro
2009-04-13 5:53 ` [PATCH v2 2/4] ftrace: introduce workqueue_handler_exit tracepoint and rename workqueue_execution to workqueue_handler_entry KOSAKI Motohiro
2009-04-13 16:25 ` Frederic Weisbecker
2009-04-15 10:22 ` Oleg Nesterov
2009-04-15 11:09 ` Ingo Molnar
2009-04-15 11:33 ` Oleg Nesterov
2009-04-15 11:45 ` Ingo Molnar
2009-04-15 12:05 ` Oleg Nesterov
2009-04-13 5:53 ` [PATCH v2 3/4] ftrace: add max execution time mesurement to workqueue tracer KOSAKI Motohiro
2009-04-13 16:16 ` Frederic Weisbecker
2009-04-13 21:21 ` Ingo Molnar [this message]
2009-04-14 1:43 ` KOSAKI Motohiro
2009-04-14 11:40 ` Frederic Weisbecker
2009-04-15 0:31 ` KOSAKI Motohiro
2009-04-13 5:55 ` [PATCH v2 4/4] ftrace: add latecy mesurement feature " KOSAKI Motohiro
2009-04-14 2:50 ` KOSAKI Motohiro
2009-04-13 15:24 ` [PATCH v2 1/4] ftrace, workqueuetrace: Make workqueue tracepoints use TRACE_EVENT macro Frederic Weisbecker
2009-04-14 4:03 ` [PATCH v2 5/4] ftrace, workqueuetrace: display work name KOSAKI Motohiro
2009-04-14 21:16 ` Frederic Weisbecker
2009-04-14 23:55 ` KOSAKI Motohiro
2009-04-15 1:15 ` Frederic Weisbecker
2009-04-15 6:13 ` KOSAKI Motohiro
2009-04-15 6:17 ` Zhaolei
2009-04-15 9:44 ` Ingo Molnar
2009-04-15 16:23 ` Frederic Weisbecker
2009-04-17 6:56 ` [PATCH v3 0/1] ftrace, workqueuetrace: Make workqueue tracepoints use TRACE_EVENT macro Zhaolei
2009-04-17 7:15 ` [PATCH v3 1/1] " Zhaolei
2009-04-17 13:45 ` Ingo Molnar
2009-04-20 1:30 ` [PATCH v3 1/1] ftrace, workqueuetrace: Make workqueuetracepoints " Zhaolei
2009-04-20 1:38 ` KOSAKI Motohiro
2009-04-20 1:43 ` Zhaolei
2009-04-20 1:49 ` KOSAKI Motohiro
2009-04-20 8:46 ` Ingo Molnar
2009-04-20 22:25 ` Oleg Nesterov
2009-04-20 23:48 ` Frederic Weisbecker
2009-04-21 15:28 ` Oleg Nesterov
2009-04-21 15:50 ` Oleg Nesterov
2009-04-21 18:33 ` Frederic Weisbecker
2009-04-21 18:28 ` Frederic Weisbecker
2009-04-21 19:37 ` Oleg Nesterov
2009-04-24 11:42 ` [PATCH 0/4] workqueue_tracepoint: Add worklet tracepoints for worklet lifecycle tracing Zhaolei
2009-04-24 11:43 ` [PATCH 1/4] workqueue_tracepoint: introduce workqueue_handler_exit tracepoint and rename workqueue_execution to workqueue_handler_entry Zhaolei
2009-04-24 11:45 ` [PATCH 2/4] workqueue_tracepoint: Add workqueue_flush and worklet_cancel tracepoint Zhaolei
2009-04-24 11:45 ` [PATCH 3/4] workqueue_tracepoint: Change tracepoint name to fit worklet and workqueue lifecycle Zhaolei
2009-04-24 11:46 ` [PATCH 4/4] workqueue_trace: Separate worklet_insertion into worklet_enqueue and worklet_enqueue_delayed Zhaolei
2009-04-24 20:06 ` [PATCH 0/4] workqueue_tracepoint: Add worklet tracepoints for worklet lifecycle tracing Andrew Morton
2009-04-24 22:59 ` Frederic Weisbecker
2009-04-24 23:20 ` Andrew Morton
2009-04-25 0:37 ` Frederic Weisbecker
2009-04-25 1:28 ` Andrew Morton
2009-04-25 2:00 ` Steven Rostedt
2009-04-25 2:24 ` Andrew Morton
2009-04-25 2:51 ` Steven Rostedt
2009-04-25 3:10 ` Andrew Morton
2009-04-25 3:32 ` Steven Rostedt
2009-04-26 10:47 ` Ingo Molnar
2009-04-27 5:44 ` Andrew Morton
2009-04-27 15:02 ` Oleg Nesterov
2009-04-27 15:43 ` Ingo Molnar
2009-04-27 19:09 ` Oleg Nesterov
2009-04-28 13:42 ` Frederic Weisbecker
2009-04-28 16:43 ` Oleg Nesterov
2009-04-28 17:49 ` Frederic Weisbecker
2009-04-24 23:27 ` Frederic Weisbecker
2009-04-28 17:24 ` Frank Ch. Eigler
2009-04-28 18:48 ` Andrew Morton
2009-04-28 20:51 ` Frank Ch. Eigler
2009-04-29 4:03 ` KOSAKI Motohiro
2009-04-29 4:29 ` Andrew Morton
2009-04-29 5:21 ` KOSAKI Motohiro
2009-04-20 17:11 ` [PATCH v3 1/1] ftrace, workqueuetrace: Make workqueue tracepoints use TRACE_EVENT macro Frederic Weisbecker
2009-04-21 1:20 ` KOSAKI Motohiro
2009-04-20 6:42 ` [PATCH 0/4] ftrace, workqueuetrace: Add worklet informationame Zhaolei
2009-04-20 6:58 ` [PATCH 1/4] trace_workqueue: use list_for_each_entry() instead of list_for_each_entry_safe() Zhaolei
2009-04-20 7:26 ` Frederic Weisbecker
2009-04-20 6:59 ` [PATCH 2/4] trace_workqueue: Remove cpu_workqueue_stats->first_entry Zhaolei
2009-04-20 7:02 ` [PATCH 3/4] trace_workqueue: Remove blank line between each cpu Zhaolei
2009-04-20 7:09 ` [PATCH 4/4] trace_workqueue: Add worklet information Zhaolei
2009-04-20 11:36 ` Frederic Weisbecker
2009-04-21 1:57 ` Zhaolei
2009-04-21 23:39 ` Frederic Weisbecker
2009-04-20 7:23 ` [PATCH 0/4] ftrace, workqueuetrace: Add worklet informationame Frederic Weisbecker
2009-04-13 14:34 ` [PATCH 1/1] tracing, workqueuetrace: Make workqueue tracepoints use TRACE_EVENT macro Frederic Weisbecker
2009-04-13 14:13 ` Frederic Weisbecker
2009-04-15 6:47 ` KOSAKI Motohiro
2009-04-13 13:57 ` [PATCH 0/1] " Frederic Weisbecker
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=20090413212159.GA8514@elte.hu \
--to=mingo@elte.hu \
--cc=akpm@linux-foundation.org \
--cc=fweisbec@gmail.com \
--cc=kosaki.motohiro@jp.fujitsu.com \
--cc=linux-kernel@vger.kernel.org \
--cc=oleg@redhat.com \
--cc=rostedt@goodmis.org \
--cc=tzanussi@gmail.com \
--cc=zhaolei@cn.fujitsu.com \
/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
Powered by JetHome