Re: 2.6.39-rc4+: oom-killer busy killing tasks

[Christian Kujau <lists@xxxxxxxxxxxxxxx>]

On Wed, 4 May 2011 at 21:12, Dave Chinner wrote:
>> Ok, that will be helpful. Also helpful is that I've (FINALLY!)
>> reproduced this myself, and i think i can now reproduce it at will
>> on a highmem i686 machine. I'll look into it more later tonight....

I've tried to reproduce it on an SMP i686 with 1024MB RAM, but there was 
no OOM there.

> And here's a patch for you to try. It fixes the problem on my test
> machine.....

Excellent! After one run with that patch, the machine does not go OOM any 
more when running du(1) over this XFS filesystem.

I've run du(1) login with my oom-debug script to gather sysrq-w and 
slabinfo, see https://trent.utfs.org/p/bits/2.6.39-rc4/oom/trace/16-* for 
the details.

I'll poke it a bit more over the day, but your patch (one line only! Wow!) 
seems to help.

Thanks so much,
Christian.
-- 
BOFH excuse #160:

non-redundant fan failure