Re: 2.6.1 IO lockup on SMP systems

[Arjan van de Ven]

On Thu, 2004-02-26 at 15:37, Dave Jones wrote:
> On Thu, Feb 26, 2004 at 05:11:35AM -0800, Andrew Morton wrote:
>  > Andrew Morton <akpm@xxxxxxxx> wrote:
>  > >
>  > > Not sure why the oom-killer didn't do anything.
>  > 
>  > There's still free swap space.  The oom-killer has problems.
> 
> That sounds odd. Surely if we have free swap, we don't
> want the oom-killer to do anything ?

with highmem it's not so easy :)
the lowzone can be entirely pinned by pagetables and such and the
highmem zone can be free... and still you want to oomkill.

Attachment: signature.asc
Description: This is a digitally signed message part