... /linux/memcontrol.h~memory-controller-add-mm-owner include/linux/memcontrol ... 2.6.25-rc5-balbir/mm/memcontrol.c 2008-03-24 ... css); > - rcu_read_unlock(); > > while (res_counter_charge(&mem->res, PAGE_SIZE)) { > if (!(gfp_mask & __GFP_WAIT)) > @@ -988 ...

linux-kernel - Li Zefan - Mar 24 2008 - 21:26

... | 5 ++++- > > kernel/fork.c | 4 ++++ > > mm/memcontrol.c | 42 ++++++++++++++++++++++++++++++++++-------- > > 4 files changed, ... include/linux/memcontrol.h~memory-controller-add-mm-owner include/linux/memcontrol.h ...

linux-kernel - Balbir Singh - Mar 25 2008 - 11:48

... /linux/mm_types.h~memory-controller-add-mm-owner 2008-03-20 ... should be > independent of the control groups, but should be part ... out; >> >> css_get(&mem->css); >> - rcu_assign_pointer(mm->mem_cgroup, mem); >> css_put(&old_mem->css ...

linux-kernel - Balbir Singh - Mar 24 2008 - 12:21

... /linux/mm_types.h~memory-controller-add-mm-owner 2008-03-20 ... should be independent of the control groups, but should be part ... out; > > css_get(&mem->css); > - rcu_assign_pointer(mm->mem_cgroup, mem); > css_put(&old_mem->css) ...

linux-kernel - Paul Menage - Mar 24 2008 - 11:03

... >> the exit.c code will add its counts into these totals." ... charging/uncharging anything to that mm or the cgroup to which ... ? I'll check for all >> mm->owner uses. >> > > Yes - cgroup_attach() uses ... won't > change if you're just using RCU, you can' ...

linux-kernel - Balbir Singh - Mar 24 2008 - 13:33

... of the other users of the mm (by looking > > first in the thread ... > the exit.c code will add its counts into these totals." ... we still need the rcu_read_lock(), since mm->owner can move > > cgroups any time ... won't change if you're just using RCU, you can' ...

linux-kernel - Paul Menage - Mar 24 2008 - 12:34

... For the case you mentioned, mm->owner is a moving target ... an efficient way to find mm->owner in that case. >>> >> But: ... will be a suitable alternate owner >> >> - in the worst case, it' ... share the mm_struct. This makes mm->owner struct approach hard to ...

linux-kernel - Balbir Singh - Mar 26 2008 - 06:29

... ) return; task_lock(task); if (task->mm == mm) { mm->owner = task; } task_unlock(task); } struct ... /* Then try my siblings */ if (need_new_owner(mm)) { for_each_child(current->real_parent, c) { try_give_mm_ownership(c); ...

linux-kernel - Paul Menage - Mar 26 2008 - 07:20

... . > > For the case you mentioned, mm->owner is a moving target and we ... high. I wonder if > there is an efficient way to find mm->owner in that case. > But: - ... your siblings will be a suitable alternate owner - in the worst case, it's not ...

linux-kernel - Paul Menage - Mar 24 2008 - 13:46

... > >> For the case you mentioned, mm->owner is a moving target and we ... high. I wonder if >> there is an efficient way to find mm->owner in that case. >> > > But: > ... siblings > will be a suitable alternate owner > > - in the worst case, it's not ...

linux-kernel - Balbir Singh - Mar 25 2008 - 07:41

... is find some other > process p where p->mm == current->mm and make it the new owner. > Exactly what sequence of clone() calls was used to ... > cases where more than one process is using an mm. > Hmmm.. the 99.9% of the time is just ...

linux-kernel - Balbir Singh - Mar 26 2008 - 07:41

On Wed, Mar 26, 2008 at 4:41 AM, Balbir Singh wrote: > > Hmmm.. the 99.9% of the time is just guess work (not measured, could be possibly > true). I see and understand your code below. But before I try and implement >

linux-kernel - Paul Menage - Mar 26 2008 - 11:21