linux-kernel - Re: [PATCH] perf/core: fix RCU issues with cgroup monitoring mode

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:	Mon, 19 Oct 2015 00:58:47 -0700
From:	Stephane Eranian <eranian@...gle.com>
To:	Peter Zijlstra <peterz@...radead.org>
Cc:	LKML <linux-kernel@...r.kernel.org>,
	"mingo@...e.hu" <mingo@...e.hu>,
	"ak@...ux.intel.com" <ak@...ux.intel.com>,
	Eric Dumazet <edumazet@...gle.com>,
	Arnaldo Carvalho de Melo <acme@...hat.com>
Subject: Re: [PATCH] perf/core: fix RCU issues with cgroup monitoring mode

Peter,

On Sat, Oct 17, 2015 at 2:56 AM, Peter Zijlstra <peterz@...radead.org> wrote:
>
> On Sat, Oct 17, 2015 at 03:28:11AM +0200, Stephane Eranian wrote:
> >
> > This patch eliminates all known RCU violations detected
> > by the RCU  checker (PROVE_RCU). The impact code paths
> > were all related to cgroup mode monitoring and involved
> > access a task's cgrp.
>
> But were they right? This patch provides no clues.
>
I am assuming that is the checker detects something suspicious there is likely
a problem.

Take for instance:
 perf_cgroup_sched_out()->perf_cgroup_from_task() ->task_subsys_state()

That one fires the checker. I think because we are accessing the css
state without
protection.

The other places are similar.

I am also trying to avoid double taking the rcu-read-lock(), so I have moved
the rcu_read_lock() out of some functions to place them in the caller. The
perf_cgroup_switch() is a performance sensitive path.

If this is not the way to fix this, let me know and I'll rewrite the patch.
Thanks.




> > Signed-off-by: Stephane Eranian <eranian@...gle.com>
> > ---
>
> > @@ -2094,6 +2104,7 @@ static int  __perf_install_in_context(void *info)
> >               cpuctx->task_ctx = task_ctx;
> >               task = task_ctx->task;
> >       }
> > +     rcu_read_lock();
> >
> >       cpu_ctx_sched_out(cpuctx, EVENT_ALL);
> >
> > @@ -2112,6 +2123,8 @@ static int  __perf_install_in_context(void *info)
> >        */
> >       perf_event_sched_in(cpuctx, task_ctx, task);
> >
> > +     rcu_read_unlock();
> > +
> >       perf_pmu_enable(cpuctx->ctx.pmu);
> >       perf_ctx_unlock(cpuctx, task_ctx);
> >
> > @@ -2398,7 +2411,9 @@ static void ctx_sched_out(struct perf_event_context *ctx,
> >               return;
> >
> >       update_context_time(ctx);
> > +     rcu_read_lock();
> >       update_cgrp_time_from_cpuctx(cpuctx);
> > +     rcu_read_unlock();
> >       if (!ctx->nr_active)
> >               return;
> >
>
> And these impact !cgroup code, and the last looks like it should be
> inside update_cgrp_time_from_cpuctx() if at all.
--
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/