Re: During VM oom condition, kill all threads in process group

Re: During VM oom condition, kill all threads in process group

[Nick Piggin]

Hi,

I'm interested about this patch dcca2bde4f86a14d3291660bede8f1844fe2b3df

I don't actually have a problem with what was merged, but I still think
that we should be calling the oom killer from this point. The oom killer
knows about what tasks to oom and what not to, whether to panic on oom,
etc.

I have a patch for this, but wasn't really pushing it hard because it's
pretty unlikely for x86 and standard filesystems to oom from here...

What architecture, filesystems, and workload did you observe problems with?
Did you discover which allocation was failing?

Thanks,
Nick

--
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/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: During VM oom condition, kill all threads in process group

[Will Schmidt]

On Wed, 2007-10-17 at 00:41 +0200, Nick Piggin wrote:
> Hi,
> 
Hi Nick, 

> What architecture, filesystems, and workload did you observe problems with?
> Did you discover which allocation was failing?

Powerpc architecture, running GPFS.   Sadly, I dont have any real
details on the workload that actually generated the oom condition, so
dont know what allocation was failing, etc.

> I have a patch for this, but wasn't really pushing it hard because it's
> pretty unlikely for x86 and standard filesystems to oom from here...

Yeah, I suspect pretty unlikely for any architecture to get there.
Powerpc in addition to x86 anyway..  :-)   For my test purposes I added
debug code to force taking that code path.

Thanks,
-Will

> Thanks,
> Nick
> 

--
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/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>