]> git.openfabrics.org - ~shefty/rdma-dev.git/commit
oom: add per-mm oom disable count
authorYing Han <yinghan@google.com>
Tue, 26 Oct 2010 21:21:23 +0000 (14:21 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 26 Oct 2010 23:52:05 +0000 (16:52 -0700)
commit3d5992d2ac7dc09aed8ab537cba074589f0f0a52
tree0fbb51026e8aa13ddd7a6516c04471ceb97bc277
parent0f4d208f1975f16f269134cee5f44c1f048581da
oom: add per-mm oom disable count

It's pointless to kill a task if another thread sharing its mm cannot be
killed to allow future memory freeing.  A subsequent patch will prevent
kills in such cases, but first it's necessary to have a way to flag a task
that shares memory with an OOM_DISABLE task that doesn't incur an
additional tasklist scan, which would make select_bad_process() an O(n^2)
function.

This patch adds an atomic counter to struct mm_struct that follows how
many threads attached to it have an oom_score_adj of OOM_SCORE_ADJ_MIN.
They cannot be killed by the kernel, so their memory cannot be freed in
oom conditions.

This only requires task_lock() on the task that we're operating on, it
does not require mm->mmap_sem since task_lock() pins the mm and the
operation is atomic.

[rientjes@google.com: changelog and sys_unshare() code]
[rientjes@google.com: protect oom_disable_count with task_lock in fork]
[rientjes@google.com: use old_mm for oom_disable_count in exec]
Signed-off-by: Ying Han <yinghan@google.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fs/exec.c
fs/proc/base.c
include/linux/mm_types.h
kernel/exit.c
kernel/fork.c