Re: [RFC PATCH] virtio_blk: Checking "private_data" to avoid kernel panic when hotplugging

["Michael S. Tsirkin" <mst@redhat.com>]

On Wed, Apr 11, 2012 at 11:17:16AM +0800, Ren Mingxin wrote:
>  On 04/09/2012 03:53 PM, Michael S. Tsirkin wrote:
> >On Wed, Mar 28, 2012 at 11:40:33AM +0800, Ren Mingxin wrote:
> >>On guest with upstream's kernel(3.3.0-rc7), I
> >>mounted virtblk as:
> >>   a) # mkfs /dev/vda
> >>   b) # mount /dev/vda /mnt
> >>   c) # cd /mnt
> >>
> >>Then I did hotplug for virtblk via virsh on host as:
> >>   a) # sudo virsh detach-disk guest vda
> >>   b) # sudo virsh attach-disk guest /media/data/test.img vda
> >>
> >>I encountered guest's kernel panic (*probability*
> >>*event*)whose backtrace liked this:
> >Any news here? Managed to trace?
> 
> Sorry, I had a leave last week.
> 
> >Does this still happen with 3.4-rc2?
> 
> I retested this with 3.4-rc2, but the panic couldn't be reproduced.
> (Last time, It seemed that the upstream's kernel was 3.3-rc7)
> 
> >There's a chance you are hitting a race fixed by
> >4678d6f970c2f7c0cbfefc0cc666432d153b321b.
> >
> >If it's still not fixed it might make sense to enable slab debugging -
> >we might have a use after free here.
> >
> 
> I don't think the below commit has fixed this panic issue, for it is
> included
> in my last test environment(3.3-rc7).
> 4678d6f970c2f7c0cbfefc0cc666432d153b321b
> 
> Now, I'd find which commit makes sense.

Great, thanks very much. I do suggest slab debugging if
the race reproduces for you with it enabled - reducing
the chance random struct rearrangements hide the bug.

> -- 
> Thanks,
> Ren