On Fri 16-12-16 12:31:51, Johannes Weiner wrote: > On Fri, Dec 16, 2016 at 04:58:08PM +0100, Michal Hocko wrote: > > @@ -1013,7 +1013,7 @@ bool out_of_memory(struct oom_control *oc) > > * make sure exclude 0 mask - all other users should have at least > > * ___GFP_DIRECT_RECLAIM to get here. > > */ > > - if (oc->gfp_mask && !(oc->gfp_mask & (__GFP_FS|__GFP_NOFAIL))) > > + if (oc->gfp_mask && !(oc->gfp_mask & __GFP_FS)) > > return true; > > This makes sense, we should go back to what we had here. Because it's > not that the reported OOMs are premature - there is genuinely no more > memory reclaimable from the allocating context - but that this class > of allocations should never invoke the OOM killer in the first place. agreed, at least not with the current implementtion. If we had a proper accounting where we know that the memory pinned by the fs is not really there then we could invoke the oom killer and be safe > > @@ -3737,6 +3752,16 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order, > > */ > > WARN_ON_ONCE(order > PAGE_ALLOC_COSTLY_ORDER); > > > > + /* > > + * Help non-failing allocations by giving them access to memory > > + * reserves but do not use ALLOC_NO_WATERMARKS because this > > + * could deplete whole memory reserves which would just make > > + * the situation worse > > + */ > > + page = __alloc_pages_cpuset_fallback(gfp_mask, order, ALLOC_HARDER, ac); > > + if (page) > > + goto got_pg; > > + > > But this should be a separate patch, IMO. > > Do we observe GFP_NOFS lockups when we don't do this? this is hard to tell but considering users like grow_dev_page we can get stuck with a very slow progress I believe. Those allocations could see some help. > Don't we risk > premature exhaustion of the memory reserves, and it's better to wait > for other reclaimers to make some progress instead? waiting for other reclaimers would be preferable but we should at least give these some priority, which is what ALLOC_HARDER should help with. > Should we give > reserve access to all GFP_NOFS allocations, or just the ones from a > reclaim/cleaning context? I would focus only for those which are important enough. Which are those is a harder question. But certainly those with GFP_NOFAIL are important enough. > All that should go into the changelog of a separate allocation booster > patch, I think. The reason I did both in the same patch is to address the concern about potential lockups when NOFS|NOFAIL cannot make any progress. I've chosen ALLOC_HARDER to give the minimum portion of the reserves so that we do not risk other high priority users to be blocked out but still help a bit at least and prevent from starvation when other reclaimers are faster to consume the reclaimed memory. I can extend the changelog of course but I believe that having both changes together makes some sense. NOFS|NOFAIL allocations are not all that rare and sometimes we really depend on them making a further progress. -- 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>