[CC Andrew - the patch was posted here http://lkml.kernel.org/r/1442318757-7141-1-git-send-email-jmarchan%40redhat.com] On Tue 15-09-15 15:56:23, Michal Hocko wrote: > On Tue 15-09-15 14:05:57, Jerome Marchand wrote: > > Since commit <6539cc05386> (mm: memcontrol: fold mem_cgroup_do_charge()), > > the order to pass to mem_cgroup_oom() is calculated by passing the number > > of pages to get_order() instead of the expected size in bytes. AFAICT, > > it only affects the value displayed in the oom warning message. > > This patch fix this. > > We haven't noticed that just because the OOM is enabled only for page > faults of order-0 (single page) and get_order work just fine. Thanks for > noticing this. If we ever start triggering OOM on different orders this > would be broken. > > > Signed-off-by: Jerome Marchand <jmarchan@xxxxxxxxxx> > > Acked-by: Michal Hocko <mhocko@xxxxxxxx> > > Btw. a quick git grep shows that at least gart_iommu_init is using > number of pages as well. I haven't checked it does that intentionally, > though. > > Thanks! > > > --- > > mm/memcontrol.c | 3 ++- > > 1 file changed, 2 insertions(+), 1 deletion(-) > > > > diff --git a/mm/memcontrol.c b/mm/memcontrol.c > > index 1742a2d..91bf094 100644 > > --- a/mm/memcontrol.c > > +++ b/mm/memcontrol.c > > @@ -2032,7 +2032,8 @@ retry: > > > > mem_cgroup_events(mem_over_limit, MEMCG_OOM, 1); > > > > - mem_cgroup_oom(mem_over_limit, gfp_mask, get_order(nr_pages)); > > + mem_cgroup_oom(mem_over_limit, gfp_mask, > > + get_order(nr_pages * PAGE_SIZE)); > > nomem: > > if (!(gfp_mask & __GFP_NOFAIL)) > > return -ENOMEM; > > -- > > 1.9.3 > > -- > Michal Hocko > SUSE Labs -- Michal Hocko SUSE Labs -- To unsubscribe from this list: send the line "unsubscribe cgroups" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
