On Fri, Apr 01, 2016 at 04:30:44PM +0200, Cornelia Huck wrote: > On Fri, 1 Apr 2016 16:14:22 +0200 > Christian Borntraeger <borntrae...@de.ibm.com> wrote: > > > On 04/01/2016 03:19 PM, Paolo Bonzini wrote: > > > Reentrancy cannot happen while the BQL is being held. > > > > > > Reviewed-by: Cornelia Huck <cornelia.h...@de.ibm.com> > > > Signed-off-by: Paolo Bonzini <pbonz...@redhat.com> > > > > Reverting this patch makes the segfaults go away. > > > > > > > > > > > > > --- > > > hw/block/dataplane/virtio-blk.c | 12 ++---------- > > > hw/scsi/virtio-scsi-dataplane.c | 9 +-------- > > > include/hw/virtio/virtio-scsi.h | 2 -- > > > 3 files changed, 3 insertions(+), 20 deletions(-) > > :( > > On the one hand, I'm wondering what we're missing here. > > On the other hand, let's just skip the cleanup patches and get the bug > fixed?
For 2.6, absolutely. I would also drop 8/9. For debugging (that we can keep up for now), instead of dropping the checks, let's replace them with asserts. -- MST
