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

[David Rientjes]

On Tue, 21 Jul 2009, Paul Menage wrote:

> No, Rik's patch just fixes the lack of inheritability that David
> introduced originally. It doesn't address the problem that the
> intention of the patches is to disallow separate processes sharing the
> same VM from having different oom_adj scores, which breaks the
> previous ability to vfork() or clone(CLONE_VM) a child and set its
> oom_adj to a non-disabled value prior to execve().
> 

That's exactly the scenario my patches were addressing, actually.  It was 
possible to get an oom killer livelock if a thread was constantly chosen 
when sharing memory with an OOM_DISABLE task.  So your example is a recipe 
for livelock without my patches since the parent is OOM_DISABLE and the 
child is not, yet they share an ->mm so neither can be killed.  If the 
child is repeatedly chosen prior to execve, no memory freeing is possible 
unless another task happens to exceed the badness score of the child.
--
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/