On Tue, Apr 24, 2012 at 08:37:22AM +0200, Paolo Bonzini wrote:
> Il 24/04/2012 07:02, David Gibson ha scritto:
> > So the patch below fixes my assertion failure, but again, I can't say
> > I understand this well enough to be confident - it might be leaking
> > scsi reqs instead. But if this isn't
Il 24/04/2012 07:02, David Gibson ha scritto:
> So the patch below fixes my assertion failure, but again, I can't say
> I understand this well enough to be confident - it might be leaking
> scsi reqs instead. But if this isn't the right fix, I hope one of you
> can help me find where the real refc
Paolo, Kevin,
I recently hit an assertion failure in the SCSI code (pseries
machine), which I pinned down to a use-after-free of the
BlockAcctCookie within a SCSIDiskReq.
I can't say I understand the refcounting well enough to be confident
of this, but when attempting to debug I noticed that the