Re: [patch] memcg: give current access to memory reserves if it's trying to die

On Thu, 17 Mar 2011 16:53:19 -0700
Andrew Morton [off-list ref] wrote:

On Wed, 9 Mar 2011 13:27:50 -0800 (PST)
David Rientjes [off-list ref] wrote:

quoted

When a memcg is oom and current has already received a SIGKILL, then give
it access to memory reserves with a higher scheduling priority so that it
may quickly exit and free its memory.

This is identical to the global oom killer and is done even before
checking for panic_on_oom: a pending SIGKILL here while panic_on_oom is
selected is guaranteed to have come from userspace; the thread only needs
access to memory reserves to exit and thus we don't unnecessarily panic
the machine until the kernel has no last resort to free memory.

Signed-off-by: David Rientjes <rientjes@google.com>
---
 mm/oom_kill.c |   11 +++++++++++
 1 files changed, 11 insertions(+), 0 deletions(-)

diff --git a/mm/oom_kill.c b/mm/oom_kill.c
--- a/mm/oom_kill.c
+++ b/mm/oom_kill.c

@@ -537,6 +537,17 @@ void mem_cgroup_out_of_memory(struct mem_cgroup *mem, gfp_t gfp_mask)
 	unsigned int points = 0;
 	struct task_struct *p;
 
+	/*
+	 * If current has a pending SIGKILL, then automatically select it.  The
+	 * goal is to allow it to allocate so that it may quickly exit and free
+	 * its memory.
+	 */
+	if (fatal_signal_pending(current)) {
+		set_thread_flag(TIF_MEMDIE);
+		boost_dying_task_prio(current, NULL);
+		return;
+	}
+
 	check_panic_on_oom(CONSTRAINT_MEMCG, gfp_mask, 0, NULL);
 	limit = mem_cgroup_get_limit(mem) >> PAGE_SHIFT;
 	read_lock(&tasklist_lock);

The code duplication seems a bit gratuitous.



Was it deliberate that mem_cgroup_out_of_memory() ignores the oom
notifier callbacks?

I'm not sure for what purpose notifier chain for oom exists.
At a loock, it's for s390/powerpc Collaborative Memory Manager.. ?

About memcg, notifier to userland already exists and I though I don't
need to call CMM callbacks (for now, there is no user with memcg, I guess.)

Thanks,
-kame

(Why does that notifier list exist at all?  Wouldn't it be better to do
this via a vmscan shrinker?  Perhaps altered to be passed the scanning
priority?)

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>