linux-kernel - Re: [PATCH tip/core/rcu 09/12] rcu: Make call_rcu_tasks() tolerate first call with irqs disabled

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]

Message-Id: <20160615225420.GB3923@linux.vnet.ibm.com>
Date:	Wed, 15 Jun 2016 15:54:20 -0700
From:	"Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>
To:	Peter Zijlstra <peterz@...radead.org>
Cc:	linux-kernel@...r.kernel.org, mingo@...nel.org,
	jiangshanlai@...il.com, dipankar@...ibm.com,
	akpm@...ux-foundation.org, mathieu.desnoyers@...icios.com,
	josh@...htriplett.org, tglx@...utronix.de, rostedt@...dmis.org,
	dhowells@...hat.com, edumazet@...gle.com, dvhart@...ux.intel.com,
	fweisbec@...il.com, oleg@...hat.com, bobby.prani@...il.com
Subject: Re: [PATCH tip/core/rcu 09/12] rcu: Make call_rcu_tasks() tolerate
 first call with irqs disabled

On Thu, Jun 16, 2016 at 12:15:14AM +0200, Peter Zijlstra wrote:
> On Wed, Jun 15, 2016 at 02:46:10PM -0700, Paul E. McKenney wrote:
> > Currently, if the very first call to call_rcu_tasks() has irqs disabled,
> > it will create the rcu_tasks_kthread with irqs disabled, which will
> > result in a splat in the memory allocator, which kthread_run() invokes
> > with the expectation that irqs are enabled.
> > 
> > This commit fixes this problem by deferring kthread creation if called
> > with irqs disabled.  The first call to call_rcu_tasks() that has irqs
> > enabled will create the kthread.
> > 
> > This bug was detected by rcutorture changes that were motivated by
> > Iftekhar Ahmed's mutation-testing efforts.
> > 
> > Signed-off-by: Paul E. McKenney <paulmck@...ux.vnet.ibm.com>
> 
> > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > index 218f8e83db73..4a3b279beb42 100644
> > --- a/kernel/sched/fair.c
> > +++ b/kernel/sched/fair.c
> > @@ -2175,7 +2175,7 @@ void task_numa_free(struct task_struct *p)
> >  
> >  		grp->nr_tasks--;
> >  		spin_unlock_irqrestore(&grp->lock, flags);
> > -		RCU_INIT_POINTER(p->numa_group, NULL);
> > +		rcu_assign_pointer(p->numa_group, NULL);
> >  		put_numa_group(grp);
> >  	}
> 
> This seems entirely unrelated; albeit desired given that other patch.

Yikes!

As you probably guessed, this was my test case for rcu_assign_pointer(NULL),
and I clearly failed to clean up after myself.  It turns out that more than
30 few rcu_assign_pointer(NULL) instances have been added in the meantime,
several of which look to be in popular core code.  So some testing will
happen.

But if you would like to have your code to also participate in this testing
effort, here is that patch standalone.

							Thanx, Paul

------------------------------------------------------------------------

commit 0f11d148dfdb67b55efdc72a1a959c8e44c5d54c
Author: Paul E. McKenney <paulmck@...ux.vnet.ibm.com>
Date:   Wed Jun 15 15:52:25 2016 -0700

    sched: Switch from RCU_INIT_POINTER() to rcu_assign_pointer()
    
    Given that rcu_assign_pointer() now avoids providing memory ordering
    when the value assigned is the constant NULL, this commit switches
    task_numa_free() from RCU_INIT_POINTER() to rcu_assign_pointer().
    
    Signed-off-by: Paul E. McKenney <paulmck@...ux.vnet.ibm.com>

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 218f8e83db73..4a3b279beb42 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -2175,7 +2175,7 @@ void task_numa_free(struct task_struct *p)
 
 		grp->nr_tasks--;
 		spin_unlock_irqrestore(&grp->lock, flags);
-		RCU_INIT_POINTER(p->numa_group, NULL);
+		rcu_assign_pointer(p->numa_group, NULL);
 		put_numa_group(grp);
 	}