linux-kernel - Re: [PATCH] tracing/ftrace: stop {irqs,preempt}soff tracers when tracing is stopped

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Tue, 17 Mar 2009 23:54:55 +0100
From:	Frederic Weisbecker <fweisbec@...il.com>
To:	Steven Rostedt <rostedt@...dmis.org>, Ingo Molnar <mingo@...e.hu>
Cc:	LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] tracing/ftrace: stop {irqs,preempt}soff tracers when
	tracing is stopped

On Tue, Mar 17, 2009 at 05:59:22PM -0400, Steven Rostedt wrote:
> 
> On Tue, 17 Mar 2009, Frederic Weisbecker wrote:
> > > > +	/*
> > > > +	 * Stop the tracer to avoid a warning subsequent
> > > > +	 * to buffer flipping failure because tracing_stop()
> > > > +	 * disables the tr and max buffers, making flipping impossible
> > > > +	 * in case of parallels max preempt off latencies.
> > > > +	 */
> > > > +	trace->stop(tr);
> > > >  	/* stop the tracing. */
> > > >  	tracing_stop();
> > > 
> > > I'm actually thinking that tracing_stop() should call the current tracer
> > > "stop" function.
> > > 
> > > -- Steve
> > 
> > 
> > Indeed, it could be better.
> > But I think of a combination of stop() callback call plus
> > ring buffer disabled because I'm not sure all tracers implement
> > the stop callback.
> 
> Yes, that's what I meant. The tracing_stop should do a few things, and one
> of them is to call the stop function if it exists.
> 
> > 
> > Should I send a v2 based on the above?
> 
> Sure.
> 
> -- Steve


Ingo,

I tested the call to the start/stop callbacks from tracing_start() and tracing_stop()
but it's not possible for now: the sched_switch tracer register/unregister his
tracepoints from these callbacks, thus allocate some memory. But tracing_start/stop
can be called from atomic.

I discussed about that with Steven on irc, and for now the v1 should be applied
as a temporary solution. We plan to simplify later the start/stop callbacks from
each tracers that need it.

Thanks.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/