Messages in this thread | | | Date | Wed, 10 Dec 2008 10:26:28 -0800 | Subject | Re: [RFC][PATCH 1/6] memcg: fix pre_destory handler | From | Paul Menage <> |
| |
On Wed, Dec 10, 2008 at 5:47 AM, Daisuke Nishimura <d-nishimura@mtf.biglobe.ne.jp> wrote: > Hmm.. but doesn't per-hierarchy-mutex solve the problem if memory and cpuset > mounted on the same hierarchy ? >
It's not a per-hierarchy mutex - it's a per-subsystem lock against changes on that subsystem's hierarchy. So each subsystem just has to take its own lock, rather than a global or per-hierarchy lock. The cgroups code takes care of acquiring the multiple locks in a safe order when necessary.
Paul
| |