]> git.baikalelectronics.ru Git - kernel.git/commit
vdpa/mlx5: should verify CTRL_VQ feature exists for MQ
authorSi-Wei Liu <si-wei.liu@oracle.com>
Sat, 15 Jan 2022 00:28:00 +0000 (19:28 -0500)
committerMichael S. Tsirkin <mst@redhat.com>
Fri, 4 Mar 2022 16:56:33 +0000 (11:56 -0500)
commit5fb106c1197584ce42ac4856e5edcddfb0a90515
treec4014bc161063281201ea78b1cfb52d4fbc783d6
parent83f317e5a8c9ae3413f88c806e5f2cf49871c12d
vdpa/mlx5: should verify CTRL_VQ feature exists for MQ

Per VIRTIO v1.1 specification, section 5.1.3.1 Feature bit requirements:
"VIRTIO_NET_F_MQ Requires VIRTIO_NET_F_CTRL_VQ".

There's assumption in the mlx5_vdpa multiqueue code that MQ must come
together with CTRL_VQ. However, there's nowhere in the upper layer to
guarantee this assumption would hold. Were there an untrusted driver
sending down MQ without CTRL_VQ, it would compromise various spots for
e.g. is_index_valid() and is_ctrl_vq_idx(). Although this doesn't end
up with immediate panic or security loophole as of today's code, the
chance for this to be taken advantage of due to future code change is
not zero.

Harden the crispy assumption by failing the set_driver_features() call
when seeing (MQ && !CTRL_VQ). For that end, verify_min_features() is
renamed to verify_driver_features() to reflect the fact that it now does
more than just validate the minimum features. verify_driver_features()
is now used to accommodate various checks against the driver features
for set_driver_features().

Signed-off-by: Si-Wei Liu <si-wei.liu@oracle.com>
Link: https://lore.kernel.org/r/1642206481-30721-3-git-send-email-si-wei.liu@oracle.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eli Cohen <elic@nvidia.com>
Acked-by: Jason Wang <jasowang@redhat.com>
drivers/vdpa/mlx5/net/mlx5_vnet.c