On Tue 24-05-16 13:37:46, roy.qing.li@xxxxxxxxx wrote: > From: Li RongQing <roy.qing.li@xxxxxxxxx> > > The margin may be set to the difference value between memory limit and > memory count firstly. which maybe returned wrongly if memsw.count excess > memsw.limit, because try_charge forces charging __GFP_NOFAIL allocations, > which may result in memsw.limit excess. If we are below memory.limit > and there's nothing to reclaim to reduce memsw.usage, might end up > looping in try_charge forever. This is quite hard for me to grasp. What would you say about the following: " mem_cgroup_margin might return memory.limit - memory_count when the memsw.limit is in excess. This doesn't happen usually because we do not allow excess on hard limits and memory.limit <= memsw.limit but __GFP_NOFAIL charges can force the charge and cause the excess when no memory is really swapable (swap is full or no anonymous memory is left). " > > Signed-off-by: Li RongQing <roy.qing.li@xxxxxxxxx> > Acked-by: Vladimir Davydov <vdavydov@xxxxxxxxxxxxx> > Cc: Michal Hocko <mhocko@xxxxxxxx> Acked-by: Michal Hocko <mhocko@xxxxxxxx> > --- > mm/memcontrol.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/mm/memcontrol.c b/mm/memcontrol.c > index 00981d2..12aaadd 100644 > --- a/mm/memcontrol.c > +++ b/mm/memcontrol.c > @@ -1090,6 +1090,8 @@ static unsigned long mem_cgroup_margin(struct mem_cgroup *memcg) > limit = READ_ONCE(memcg->memsw.limit); > if (count <= limit) > margin = min(margin, limit - count); > + else > + margin = 0; > } > > return margin; > -- > 2.1.4 -- 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
