Re: [PATCH] mm,oom: Don't call schedule_timeout_killable() with oom_lock held.

[Michal Hocko <mhocko@kernel.org>]

On Fri 18-05-18 19:14:12, Tetsuo Handa wrote:
> Michal Hocko wrote:
> > On Sat 12-05-18 23:18:24, Tetsuo Handa wrote:
> > [...]
> > > @@ -4241,6 +4240,12 @@ bool gfp_pfmemalloc_allowed(gfp_t gfp_mask)
> > >  	/* Retry as long as the OOM killer is making progress */
> > >  	if (did_some_progress) {
> > >  		no_progress_loops = 0;
> > > +		/*
> > > +		 * This schedule_timeout_*() serves as a guaranteed sleep for
> > > +		 * PF_WQ_WORKER threads when __zone_watermark_ok() == false.
> > > +		 */
> > > +		if (!tsk_is_oom_victim(current))
> > > +			schedule_timeout_uninterruptible(1);
> > >  		goto retry;
> > 
> > We already do have that sleep for PF_WQ_WORKER in should_reclaim_retry.
> > Why do we need it here as well?
> 
> Because that path depends on __zone_watermark_ok() == true which is not
> guaranteed to be executed.

Is there any reason we cannot do the special cased sleep for
PF_WQ_WORKER in should_reclaim_retry? The current code is complex enough
to make it even more so. If we need a hack for PF_WQ_WORKER case then we
definitely want to have a single place to do so.
-- 
Michal Hocko
SUSE Labs