On Sun, 5 Oct 2008 20:09:22 +0200 (CEST)
Thomas Gleixner [off-list ref] wrote:

The cat /debug/tracing/tracing_enabled == 1 might be just a user space
interface thingy.

Thanks,

	Thomas, Ilpo etc, I got a trace exactly after the patch
provided by Thomas detected the stall:

http://www.abusar.org/ftrace/trace-final.txt.bz2

	Thomas, since you didn't reply to the previous messages, I
assume the previous traces were invalid. But could you please check
this one?

	It would be wise to search for "monit" in the trace, since
monit activated the script "stop-tracing.sh" which stops the trace and
dump the trace output to the file. So what probably interest you is
what comes *before* "monit".

	"monit" appears for the first time in line 16921.

	To make it more clear, I configured monit with the following:

check file syslog with path /var/adm/messages
	if match "Tracer stopped" then exec "/home/fraga/stop-tracing.sh"

	And the stop-tracing.sh script is just that:

#!/bin/sh
/usr/local/bin/echo 0 > /debug/tracing/tracing_enabled
/usr/local/bin/cat /debug/tracing/trace > /home/fraga/trace.txt

	So I expect that this trace is right now.

	If monit was too slow in detecting the "Tracer stopped" log entry, I'll 
use a Perl script which is faster.

	Thanks, I wait for your reply.

	Ps: flood_mail.pl is just a perl script I use to prevent people from flooding
my smtp server with mail.

	Ps2: I just don't know if it would be better to stop the tracing *before*
dump the trace or it would be better to dump the trace without stopping it... I noticed
that even the "echo" I use to stop the tracing, generated a lot of useless tracing information.

--