From: Pekka Enberg <penberg@kernel.org>
To: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>,
Peter Zijlstra <peterz@infradead.org>,
Steven Rostedt <rostedt@goodmis.org>,
Arjan van de Ven <arjan@linux.intel.com>,
Arnaldo Carvalho de Melo <acme@redhat.com>,
Frederic Weisbecker <fweisbec@gmail.com>,
linux-kernel@vger.kernel.org,
Linus Torvalds <torvalds@linux-foundation.org>,
Andrew Morton <akpm@linux-foundation.org>,
Darren Hart <dvhart@linux.intel.com>,
Arjan van de Ven <arjan@infradead.org>
Subject: Re: [patch] trace: Add user-space event tracing/injection
Date: Wed, 17 Nov 2010 14:17:06 +0200 [thread overview]
Message-ID: <4CE3C7C2.7000200@kernel.org> (raw)
In-Reply-To: <20101117120740.GA24972@elte.hu>
Hi Ingo,
On 11/17/10 2:07 PM, Ingo Molnar wrote:
> * Pekka Enberg<penberg@kernel.org> wrote:
>
>> (Trimming CC for sanity.)
> [ Added back Linus (in case he wants to object to the new prctl()) and lkml. ]
>
>> On 11/16/10 11:04 PM, Thomas Gleixner wrote:
>>> 'trace' is our shot at improving the situation: it aims at providing a
>>> simple to use and straightforward tracing tool based on the perf
>>> infrastructure and on the well-known perf profiling workflow:
>>>
>>> trace record firefox # trace firefox and all child tasks,
>>> put it into trace.data
>>>
>>> trace summary # show quick overview of the trace,
>>> based on trace.data
>>>
>>> trace report # detailed traces, based on trace.data
>> Nice work guys!
>>
>> Does this concept lend itself to tracing latencies in userspace applications that
>> run in virtual machines (e.g. the Java kind)? I'm of course interested in this
>> because of Jato [1] where bunch of interesting things can cause jitter: JIT
>> compilation, GC, kernel, and the actual application doing something (in either
>> native code or JIT'd code). It's important to be able to measure where "slowness"
>> to desktop applications and certain class of server applications comes from to be
>> able to improve things.
> Makes quite a bit of sense.
>
> How about the attached patch? It works fine with the simple testcase included in the
> changelog. There's a common-sense limit on the message size - but otherwise it adds
> support for apps to generate a free-form string trace event.
>
> Thanks,
>
> Ingo
>
> ---------------------------------->
> Subject: trace: Add user-space event tracing/injection
> From: Ingo Molnar<mingo@elte.hu>
> Date: Wed Nov 17 10:11:53 CET 2010
>
> This feature (suggested by Darren Hart and Pekka Engberg) allows user-space
> programs to print trace events in a very simple and self-contained way:
>
> #include<sys/prctl.h>
> #include<string.h>
>
> #define PR_TASK_PERF_USER_TRACE 35
>
> int main(void)
> {
> char *msg = "Hello World!\n";
>
> prctl(PR_TASK_PERF_USER_TRACE, msg, strlen(msg));
>
> return 0;
> }
>
> These show up in 'trace' output as:
>
> $ trace report
> #
> # trace events of 'sleep 1':
> #
> testit/ 6006 ( 0.002 ms):<"Hello World!">
> testit/ 6006 ( 0.002 ms):<"Hello World!">
Wow! This looks really nice!
What does the duration in milliseconds mean there? For things like GC
and JIT, I want something like:
void gc(void)
{
prctl(PR_TASK_PERF_USER_TRACE_START, ...)
collect();
prctl(PR_TASK_PERF_USER_TRACE_END, ...)
}
So that it's clear from the tracing output that the VM was busy doing GC
for n milliseconds. Barring background JIT'ing and pauseless GC, I'd
also be interested in showing how much time the VM was actually
_blocking_ the running application (which can happen with signals too,
btw, for things like accessing data that's lazily initialized).
Pekka
next prev parent reply other threads:[~2010-11-17 12:17 UTC|newest]
Thread overview: 88+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-11-16 21:04 [ANNOUNCE] New utility: 'trace' Thomas Gleixner
2010-11-16 21:27 ` Darren Hart
2010-11-16 21:59 ` Ingo Molnar
2010-11-16 22:03 ` Peter Zijlstra
2010-11-16 22:08 ` Ingo Molnar
2010-11-16 22:09 ` Darren Hart
2010-11-16 22:48 ` Darren Hart
2010-11-16 22:07 ` Darren Hart
2010-11-16 22:17 ` Frederic Weisbecker
2010-11-17 8:30 ` Ingo Molnar
2010-11-17 11:35 ` Peter Zijlstra
2010-11-17 12:53 ` Frederic Weisbecker
2010-11-17 13:02 ` Peter Zijlstra
2010-11-17 13:10 ` Ingo Molnar
2010-11-17 13:36 ` Peter Zijlstra
2010-11-17 13:43 ` Frederic Weisbecker
2010-11-17 13:53 ` Peter Zijlstra
2010-11-17 14:10 ` Frederic Weisbecker
2010-11-17 15:02 ` Steven Rostedt
2010-11-17 18:13 ` Ted Ts'o
2010-11-17 18:29 ` Frederic Weisbecker
2010-11-17 18:30 ` Darren Hart
2010-11-17 16:49 ` Darren Hart
2010-11-18 23:23 ` Joe Perches
2010-11-16 21:35 ` Ingo Molnar
2010-11-17 1:37 ` Ted Ts'o
2010-11-17 1:47 ` Steven Rostedt
2010-11-17 3:16 ` Ted Ts'o
2010-11-17 3:34 ` Steven Rostedt
2010-11-17 19:00 ` Bob Copeland
2010-11-17 13:24 ` Ingo Molnar
2010-11-17 13:38 ` Peter Zijlstra
2010-11-17 14:00 ` Ingo Molnar
2010-11-17 14:11 ` Peter Zijlstra
2010-11-17 14:18 ` Peter Zijlstra
2010-11-17 14:37 ` Thomas Gleixner
2010-11-17 14:41 ` Peter Zijlstra
2010-11-17 15:41 ` Tom Zanussi
2010-11-17 15:55 ` Avi Kivity
2010-11-17 15:58 ` Thomas Gleixner
2010-11-17 16:20 ` Avi Kivity
2010-11-17 15:59 ` Peter Zijlstra
2010-11-17 16:08 ` Avi Kivity
2010-11-17 15:10 ` Steven Rostedt
2010-11-17 15:43 ` Peter Zijlstra
2010-11-17 16:04 ` Steven Rostedt
2010-11-17 16:15 ` Steven Rostedt
2010-11-17 19:40 ` Darren Hart
2010-11-17 18:23 ` Mathieu Desnoyers
2010-11-18 5:58 ` Masami Hiramatsu
2010-11-18 6:00 ` Masami Hiramatsu
2010-11-17 15:33 ` Tom Zanussi
2010-11-17 18:36 ` Mathieu Desnoyers
2010-11-17 18:53 ` Tom Zanussi
2010-11-17 19:02 ` Mathieu Desnoyers
2010-11-17 19:25 ` Tom Zanussi
2010-11-17 19:25 ` Steven Rostedt
2010-11-17 9:49 ` Philipp Marek
[not found] ` <4CE38C53.8090606@kernel.org>
2010-11-17 12:07 ` [patch] trace: Add user-space event tracing/injection Ingo Molnar
2010-11-17 12:17 ` Pekka Enberg [this message]
2010-11-17 12:30 ` Ingo Molnar
2010-11-17 12:33 ` Pekka Enberg
2010-11-17 12:42 ` Peter Zijlstra
2010-11-17 12:58 ` Ingo Molnar
2010-11-17 13:09 ` Peter Zijlstra
2010-11-17 13:36 ` Ingo Molnar
2010-11-17 12:24 ` Ingo Molnar
2010-11-17 12:30 ` Pekka Enberg
2010-11-17 12:29 ` Peter Zijlstra
2010-11-17 12:37 ` Peter Zijlstra
2010-11-17 13:10 ` Frederic Weisbecker
2010-11-17 13:31 ` Peter Zijlstra
2010-11-17 13:33 ` Frederic Weisbecker
2010-11-18 1:18 ` Darren Hart
2010-11-18 8:55 ` Ingo Molnar
2010-11-18 16:25 ` hp
2010-11-18 18:56 ` Ingo Molnar
2010-11-18 19:13 ` AW: " Reichert, Hans-Peter
2010-11-18 17:06 ` Darren Hart
2010-11-17 12:47 ` [ANNOUNCE] New utility: 'trace' Török Edwin
2010-11-17 12:51 ` Peter Zijlstra
2010-11-17 13:05 ` Török Edwin
2010-11-17 13:10 ` Peter Zijlstra
2010-11-17 13:32 ` Frederic Weisbecker
2010-11-18 0:47 ` Ian Munsie
[not found] ` <20101118151141.GA3368@redhat.com>
2010-11-19 2:32 ` Ian Munsie
2010-11-19 15:23 ` Jason Baron
[not found] ` <4CECACF9.3080907@linux.vnet.ibm.com>
[not found] ` <4CFD7182.4060206@linux.vnet.ibm.com>
2010-12-07 4:03 ` PowerPC, ftrace: Add PPC raw syscall tracepoints & ftrace fixes Ian Munsie
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=4CE3C7C2.7000200@kernel.org \
--to=penberg@kernel.org \
--cc=acme@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=arjan@infradead.org \
--cc=arjan@linux.intel.com \
--cc=dvhart@linux.intel.com \
--cc=fweisbec@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
--cc=tglx@linutronix.de \
--cc=torvalds@linux-foundation.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;
as well as URLs for NNTP newsgroup(s).