[pve-devel] [Qemu-devel] segfault in aio_bh_poll async.c:80 WAS: Re: kvm process disappears

Paolo Bonzini pbonzini at redhat.com
Wed May 22 10:41:15 CEST 2013


Il 22/05/2013 08:26, Stefan Priebe - Profihost AG ha scritto:
>> Hi,
>>
>> as i can't reproduce no ;-( i just saw the kernel segfault message and
>> used addr2line and a qemu dbg package to get the code line.
> 
> I've now seen this again for two or three times. It always happens
> when we do an fstrim inside the guest.


> And I've seen this first since josh async rbd patch.

This one?

commit dc7588c1eb3008bda53dde1d6b890cd299758155
Author: Josh Durgin <josh.durgin at inktank.com>
Date:   Fri Mar 29 13:03:23 2013 -0700

    rbd: add an asynchronous flush
    
    The existing bdrv_co_flush_to_disk implementation uses rbd_flush(),
    which is sychronous and causes the main qemu thread to block until it
    is complete. This results in unresponsiveness and extra latency for
    the guest.
    
    Fix this by using an asynchronous version of flush.  This was added to
    librbd with a special #define to indicate its presence, since it will
    be backported to stable versions. Thus, there is no need to check the
    version of librbd.
    
    Implement this as bdrv_aio_flush, since it matches other aio functions
    in the rbd block driver, and leave out bdrv_co_flush_to_disk when the
    asynchronous version is available.
    
    Reported-by: Oliver Francke <oliver at filoo.de>
    Signed-off-by: Josh Durgin <josh.durgin at inktank.com>
    Signed-off-by: Stefan Hajnoczi <stefanha at redhat.com>


Do you see it even with "-drive discard=off"?

Paolo


More information about the pve-devel mailing list