On Thu 08-05-14 14:22:24, Johannes Weiner wrote: > On Wed, May 07, 2014 at 04:33:34PM +0200, Michal Hocko wrote: > > On Wed 30-04-14 16:25:36, Johannes Weiner wrote: [...] > > > - if (unlikely(task_in_memcg_oom(current))) > > > - goto nomem; > > > - > > > - if (gfp_mask & __GFP_NOFAIL) > > > - oom = false; > > > retry: > > > if (consume_stock(memcg, nr_pages)) > > > goto done; > > [...] > > > @@ -2662,6 +2660,9 @@ retry: > > > if (mem_cgroup_wait_acct_move(mem_over_limit)) > > > goto retry; > > > > > > + if (gfp_mask & __GFP_NOFAIL) > > > + goto bypass; > > > + > > > > This is a behavior change because we have returned ENOMEM previously > > __GFP_NOFAIL must never return -ENOMEM, or we'd have to rename it ;-) > It just looks like this in the patch, but this is the label code: > > nomem: > if (!(gfp_mask & __GFP_NOFAIL)) > return -ENOMEM; > bypass: > ... Ouch. Brain fart. Sorry... -- Michal Hocko SUSE Labs -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>