On Mon, May 28, 2012 at 10:58:29AM +0200, Pawe?? Sikora wrote:
> Hi,
> i'm pinging this issue one more time...
we did look into it last time, and we didn't find anything
problematic, we also cannot recreate the issue from the
description given so far ...
> as you known there's a problem with interaction of
> mnt_is_reachable() and vfs elements (especially with autofs
> with short 1 sec unmount timeout) which leads to deadlock on
> all smp cores during heavy i/o workload.
> i've uploaded few screenshots with random oopses from kernel-3.4.0
> with the latest vs2.3.3.4 patch: http://pluto.agmk.net/kernel/
it would really help if you would:
a) record the entire kernel log (e.g. via serial console)
so that the dumps/traces can be read from the beginning
b) run a kernel with necessary debug information so that
the traces can be pinpointed
c) describe the setup and the trigger to recreate this
issue on a test system (preferably kvm because of the
high degree of recreateability)
> could you please investigate this problem?
as soon as we have something to investigate we'll gladly
do so ...
thanks in advance,
Herbert
> BR,
> Pawe??.
Received on Tue May 29 03:36:31 2012