[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20150514130823.4502e151e0e31321a114edcd@linux-foundation.org>
Date: Thu, 14 May 2015 13:08:23 -0700
From: Andrew Morton <akpm@...ux-foundation.org>
To: Shreyas B Prabhu <shreyas@...ux.vnet.ibm.com>
Cc: Steven Rostedt <rostedt@...dmis.org>, linux-kernel@...r.kernel.org,
mingo@...hat.com, aneesh.kumar@...ux.vnet.ibm.com,
paulmck@...ux.vnet.ibm.com, preeti@...ux.vnet.ibm.com,
mgorman@...e.de, namhyung@...nel.org, lizf@...fujitsu.com,
acme@...hat.com
Subject: Re: [PATCH RESEND 0/3] tracing/mm: Fix suspicious
rcu_dereference_check() usage warnings
On Wed, 13 May 2015 13:17:06 +0530 Shreyas B Prabhu <shreyas@...ux.vnet.ibm.com> wrote:
> >>> + TP_CONDITION(cpu_online(smp_processor_id())),
> >>
> >> Are we sure these can't generate check_preemption_disabled() warnings?
> >> Is there some reason why all these calls always occur with preemption
> >> disabled?
> >
> > Good catch. I don't think the code does.
> >
> > Now, I'm not sure if we should just add a raw_smp_processor_id(). The
> > idea is just to make sure that the CPU we are running on is online,
> > because it is possible to call theses trace points when the CPU is
> > going offline. If that happens, then there's a race with RCU.
> >
> > Since a task can not be migrated to an offline CPU, we don't need to
> > worry about the cpu_online(smp_processor_id()) returning a false
> > positive. A false negative would just skip a tracepoint, but I'm not
> > sure that is possible either.
> >
> > In any case, comments should also be added to why the condition is
> > there.
> >
> I'll send a patch adding the comments.
Please also confirm that these tracepoints cannot be called
from preemptible contexts. Or switch to raw_smp_processor_id().
--
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/
Powered by blists - more mailing lists