Re: [PATCH] copy over oom_adj value at fork time

[David Rientjes]

On Tue, 28 Jul 2009, KOSAKI Motohiro wrote:

> Actually, if we assume the administrator is really stupid, he can mark
> all processes as OOM_DISABLE. it makes livelock anyway.

Wrong, the oom killer will panic the machine if there are no eligible 
tasks to kill.  That's actually a more desirable result in most 
circumstances over livelocking.

> ITOH, we never seen this livelock on vfork()ed application.
> 

I have, which is why I wrote my patches.

> More important thing is: Documentation/filesysmtem/proc/txt says
> oom_adj is process property and vfork()ed parent and child are definitelly
> another process.
> 

I've proposed adding /proc/pid/oom_adj_child so that newly initialized 
mm's start with a default oom_adj value.  Paul suggested that it be a 
per-task characteristic, so that's what I'm looking at implementing.

Thanks.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/