Messages in this thread | | | Date | Mon, 17 Mar 2008 20:45:42 +0530 | From | Balbir Singh <> | Subject | Re: [RFC][0/3] Virtual address space control for cgroups |
| |
Paul Menage wrote: > On Mon, Mar 17, 2008 at 1:08 PM, Balbir Singh <balbir@linux.vnet.ibm.com> wrote: >> I understand the per-mm pointer overhead back to the cgroup. I don't understand >> the part about adding a per-mm pointer back to the "owning" task. We already >> have task->mm. > > Yes, but we don't have mm->owner, which is what I was proposing - > mm->owner would be a pointer typically to the mm's thread group > leader. It would remove the need to have to have pointers for the > various different cgroup subsystems that need to act on an mm rather > than a task_struct, since then you could use > mm->owner->cgroups[subsys_id]. >
Aaahh.. Yes.. mm->owner might be a good idea. The only thing we'll need to handle is when mm->owner dies (I think the thread group is still kept around). The other disadvantage is the double dereferencing, which should not be all that bad.
> But this is kind of orthogonal to whether virtual address space limits > should be a separate cgroup subsystem. >
Yes, sure.
-- Warm Regards, Balbir Singh Linux Technology Center IBM, ISTL
| |