On Thu 06-12-12 19:39:41, Ying Han wrote: [...] > Michal, > > I got some trouble while running this patch with my test. The test > creates hundreds of memcgs which each runs some workload to generate > global pressure. At the last, it removes all the memcgs by rmdir. Then > the cmd "ls /dev/cgroup/memory/" hangs afterwards. > > I studied a bit of the patch, but not spending too much time on it > yet. Looks like that the v2 has something different from your last > post, where you replaces the mem_cgroup_get() with css_get() on the > iter->last_visited. Didn't follow why we made that change, but after > restoring the behavior a bit seems passed my test. Hmm, strange. css reference counting should be stronger than mem_cgroup one because it pins css thus cgroup which in turn keeps memcg alive. > Here is the patch I applied on top of this one: [...] -- 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>