Messages in this thread | | | Date | Thu, 21 Jun 2012 02:01:44 -0700 (PDT) | From | David Rientjes <> | Subject | Re: [patch] mm, thp: abort compaction if migration page cannot be charged to memcg |
| |
On Thu, 21 Jun 2012, David Rientjes wrote:
> It's possible that subsequent pageblocks would contain memory allocated > from solely non-oom memcgs, but it's certainly not a guarantee and results > in terrible performance as exhibited above. Is there another good > criteria to use when deciding when to stop isolating and attempting to > migrate all of these pageblocks? > > Other ideas? >
The only other alternative that I can think of is to check mem_cgroup_margin() in isolate_migratepages_range() and return a NULL lruvec that would break that pageblock and return, and then set a bit in struct mem_cgroup that labels it as oom so we can check for it on subsequent pageblocks without incurring the locking to do mem_cgroup_margin() in res_counter, and then clear that bit on every uncharge to a memcg, but this still seems like a tremendous waste of cpu (especially if /sys/kernel/mm/transparent_hugepage/defrag == always) if most pageblocks contain pages from an oom memcg.
| |