On Sat, 2012-08-18 at 20:52 +0200, Paolo Bonzini wrote: > Il 14/08/2012 22:20, Nicholas A. Bellinger ha scritto: > >>> > > Since virtio_scsi currently assumes a single vqs for data, this patch > >>> > > simply changes ->cmd_vqs[1] to handle the single VirtQueue. > > Wrong, multiqueue works just fine. :) It's just the kernel driver that > doesn't support it yet. >
<nod> > >>> > > Cc: Paolo Bonzini <pbonz...@redhat.com> > >>> > > Cc: Stefan Hajnoczi <stefa...@linux.vnet.ibm.com> > >>> > > Cc: Michael S. Tsirkin <m...@redhat.com> > >>> > > Signed-off-by: Nicholas Bellinger <n...@linux-iscsi.org> > >> > > >> > This is a bugfix we need even without vhost, right? > >> > > > I believe so, as it appears to be stomping past the end of memory for > > every virtio-scsi initialization regardless of vhost usage.. > > You just did a wrong merge. When commit d2ad7dd (virtio-scsi: add > multiqueue capability, 2012-04-06) changed cmd_vq from pointer to array > of pointers, you should have moved the following fields to the middle of > the struct, just like that commit did. Ahh, I see how virtio_scsi_init() -> virtio_common_init() are setting up the memory now.. Apologies, my mistake. So moving the vhost-scsi related structure members ahead of the VirtQueue releated definitions for RFC-v3, and dropping this patch. Thanks Paolo! --nab
