The vhost code currently assumes vring are mapped in memory following the
legacy virtio layout. This may cause QEMU to fail with virtio 1 devices
if the used ring is mapped below the descriptor table.
---
Greg Kurz (3):
vhost: adapt vhost_verify_ring_mappings() to virtio 1 ring layout
vhost: drop legacy vring layout bits
virtio: drop virtio_queue_get_ring_{size,addr}()
hw/virtio/vhost.c | 92 +++++++++++++++++++++++++++++---------------
hw/virtio/virtio.c | 11 -----
include/hw/virtio/vhost.h | 7 ++-
include/hw/virtio/virtio.h | 2 -
4 files changed, 64 insertions(+), 48 deletions(-)
--
Greg
