lkml.org 
[lkml]   [2008]   [Oct]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
DateFri, 17 Oct 2008 16:29:37 -0700
FromTim Bird <>
SubjectRe: [PATCH 4/5] tracing/fastboot: Add a time field on the sched_switch entry
� wrote:
> 2008/10/12 Fr�d�ric Weisbecker <fweisbec@gmail.com>:
>> 2008/10/12 Steven Rostedt <rostedt@goodmis.org>:
>>> You can tap into the context switch trace_point too, and have that record
>>> a special trace entry for you instead.
>
> Another thing I could do: when a sched switch entry occurs, I could
> send a special timestamp entry that
> fits my needs just before the commit of the sched entry, only if the
> current tracer is the boot tracer. This way I can output
> the appropriate timestamp just before output the sched entry.
>
> What do you think?

I may have lost context here, but what's the purpose of the
extra timestamp again? If it's just to have zero-based
timestamps on the ultimate output, can't you just put a single
special "start" timestamp in the trace log. Then that can
be subtraced from all tracer-generated stamps in post-processing.
I'm not familiar enough with the tracer to know if the
post-processing is done in kernel space. But in any event
you could use something like scripts/show_delta to get
relative timestamps (in user space). Right now, scripts/show_delta
doesn't act like a filter - I'll have to look at that.

Sorry if I'm just adding noise...
-- Tim

=============================
Tim Bird
Architecture Group Chair, CE Linux Forum
Senior Staff Engineer, Sony Corporation of America
=============================
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2008-10-18 01:33    [from the cache]
©2003-2010