Re: [tip:sched/core] sched: Fix cgroup movement of forking process

[Frederic Weisbecker <fweisbec@gmail.com>]

On Wed, Dec 21, 2011 at 09:37:33AM -0800, Tejun Heo wrote:
> (cc'ing Frederic)
> 
> On Wed, Dec 21, 2011 at 09:26:32AM -0800, Tejun Heo wrote:
> > Hello, guys.
> > 
> > On Wed, Dec 21, 2011 at 03:44:14AM -0800, tip-bot for Daisuke Nishimura wrote:
> > > sched: Fix cgroup movement of forking process
> > > 
> > > There is a small race between task_fork_fair() and sched_move_task(),
> > > which is trying to move the parent.
> > > 
> > >         task_fork_fair()                 sched_move_task()
> > > --------------------------------+---------------------------------
> > >   cfs_rq = task_cfs_rq(current)
> > >     -> cfs_rq is the "old" one.
> > >   curr = cfs_rq->curr
> > >     -> curr is set to the parent.
> > >                                     task_rq_lock()
> > >                                     dequeue_task()
> > >                                       ->parent.se.vruntime -= (old)cfs_rq->min_vruntime
> > >                                     enqueue_task()
> > >                                       ->parent.se.vruntime += (new)cfs_rq->min_vruntime
> > >                                     task_rq_unlock()
> > >   raw_spin_lock_irqsave(rq->lock)
> > >   se->vruntime = curr->vruntime
> > >     -> vruntime of the child is set to that of the parent
> > >        which has already been updated by sched_move_task().
> > >   se->vruntime -= (old)cfs_rq->min_vruntime.
> > >   raw_spin_unlock_irqrestore(rq->lock)
> > > 
> > > As a result, vruntime of the child becomes far bigger than expected,
> > > if (new)cfs_rq->min_vruntime >> (old)cfs_rq->min_vruntime.
> > > 
> > > This patch fixes this problem by setting "cfs_rq" and "curr" after
> > > holding the rq->lock.
> > 
> > The race shouldn't happen with threadgroup locking scheduled to be
> > merged for the coming merge window.  sched_fork() and cgroup migration
> > become exclusive and won't happen concurrently.  Would still make
> > sense for -stable tho.
> 
> I retract that.  sched_move_task() can also be called from
> cgroup_exit() which is outside of threadgroup locking.
> 
> Frederic, so, it seems we actually have race conditions here.  I
> really wish cgroup made sure that things like this can't happen even
> if we pay a bit of overhead in relatively cold paths.  I could be
> being unrealistic tho.  Any ideas?

Hmm, I'm a bit confused about the issue. But doesn't this patch fix the issue?

Also the parent can't be calling sched_fork() and cgroup_exit() at
the same time.

Or am I missing something?