On Tue 04-12-12 12:32:17, Glauber Costa wrote: > On 12/04/2012 12:17 PM, Michal Hocko wrote: > >> But it should be extremely easy to protect against this. It is just a > >> > matter of not returning online css in the iterator: then we'll never see > >> > them until they are online. This also sounds a lot more correct than > >> > returning allocated css. > > Yes but... Look at your other patch which relies on iterator when counting > > children to find out if there is any available. > > > And what is the problem with it ? Bahh. Right you are because the value is copied only at the css_online time. So even if mem_cgroup_hierarchy_write wouldn't see any child (because they are still offline) and managed to set use_hierarchy=1 with some children linked in all would be fixed in mem_cgroup_css_online. P.S. Hey mhocko stop saying crap. -- 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>