[Qemu-devel] [PATCH for 2.3 V2] virtio-net: validate backend queue numbers against bus limitation

[Qemu-devel] [PATCH for 2.3 V2] virtio-net: validate backend queue numbers against bus limitation

[Jason Wang]

We don't validate the backend queue numbers against bus limitation,
this will easily crash qemu if it exceeds the limitation which will
hit the abort() in virtio_del_queue(). An example is trying to
starting a virtio-net device with 256 queues. E.g:

./qemu-system-x86_64 -netdev tap,id=hn0,queues=256 -device
virtio-net-pci,netdev=hn0

Fixing this by doing the validation and fail early.

Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: qemu-stable <qemu-stable@nongnu.org>
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
Changes from V1:
Tweak the commit log.
---
 hw/net/virtio-net.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/hw/net/virtio-net.c b/hw/net/virtio-net.c
index 27adcc5..59f76bc 100644
--- a/hw/net/virtio-net.c
+++ b/hw/net/virtio-net.c
@@ -1588,6 +1588,13 @@ static void virtio_net_device_realize(DeviceState *dev, Error **errp)
     virtio_init(vdev, "virtio-net", VIRTIO_ID_NET, n->config_size);
 
     n->max_queues = MAX(n->nic_conf.peers.queues, 1);
+    if (n->max_queues * 2 + 1 > VIRTIO_PCI_QUEUE_MAX) {
+        error_setg(errp, "Invalid number of queues (= %" PRIu32 "), "
+                   "must be a postive integer less than %d.",
+                   n->max_queues, (VIRTIO_PCI_QUEUE_MAX - 1) / 2);
+        virtio_cleanup(vdev);
+        return;
+    }
     n->vqs = g_malloc0(sizeof(VirtIONetQueue) * n->max_queues);
     n->vqs[0].rx_vq = virtio_add_queue(vdev, 256, virtio_net_handle_rx);
     n->curr_queues = 1;
-- 
2.1.0

Re: [Qemu-devel] [PATCH for 2.3 V2] virtio-net: validate backend queue numbers against bus limitation

[Stefan Hajnoczi]

[-- Attachment #1: Type: text/plain, Size: 1986 bytes --]

On Fri, Mar 20, 2015 at 02:07:50PM +0800, Jason Wang wrote:
> We don't validate the backend queue numbers against bus limitation,
> this will easily crash qemu if it exceeds the limitation which will
> hit the abort() in virtio_del_queue(). An example is trying to
> starting a virtio-net device with 256 queues. E.g:
> 
> ./qemu-system-x86_64 -netdev tap,id=hn0,queues=256 -device
> virtio-net-pci,netdev=hn0
> 
> Fixing this by doing the validation and fail early.
> 
> Cc: Michael S. Tsirkin <mst@redhat.com>
> Cc: qemu-stable <qemu-stable@nongnu.org>
> Signed-off-by: Jason Wang <jasowang@redhat.com>
> ---
> Changes from V1:
> Tweak the commit log.
> ---
>  hw/net/virtio-net.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/hw/net/virtio-net.c b/hw/net/virtio-net.c
> index 27adcc5..59f76bc 100644
> --- a/hw/net/virtio-net.c
> +++ b/hw/net/virtio-net.c
> @@ -1588,6 +1588,13 @@ static void virtio_net_device_realize(DeviceState *dev, Error **errp)
>      virtio_init(vdev, "virtio-net", VIRTIO_ID_NET, n->config_size);
>  
>      n->max_queues = MAX(n->nic_conf.peers.queues, 1);
> +    if (n->max_queues * 2 + 1 > VIRTIO_PCI_QUEUE_MAX) {
> +        error_setg(errp, "Invalid number of queues (= %" PRIu32 "), "
> +                   "must be a postive integer less than %d.",
> +                   n->max_queues, (VIRTIO_PCI_QUEUE_MAX - 1) / 2);
> +        virtio_cleanup(vdev);
> +        return;
> +    }

If VIRTIO_PCI_QUEUE_MAX is really PCI-only then this is a layering
violation.  Virtio devices should not assume a particular transport
(PCI, CCW, etc).

Either the constant should be renamed if it truly applies to all
transports, or there should be a virtio_get_max_queues() function that
lets the transport pick a value.

That said, virtio-scsi.c already checks VIRTIO_PCI_QUEUE_MAX so maybe
this patch is fine and a cleanup should be written later:

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>

[-- Attachment #2: Type: application/pgp-signature, Size: 473 bytes --]

Re: [Qemu-devel] [PATCH for 2.3 V2] virtio-net: validate backend queue numbers against bus limitation

[Michael S. Tsirkin]

On Mon, Mar 23, 2015 at 01:20:28PM +0000, Stefan Hajnoczi wrote:
> On Fri, Mar 20, 2015 at 02:07:50PM +0800, Jason Wang wrote:
> > We don't validate the backend queue numbers against bus limitation,
> > this will easily crash qemu if it exceeds the limitation which will
> > hit the abort() in virtio_del_queue(). An example is trying to
> > starting a virtio-net device with 256 queues. E.g:
> > 
> > ./qemu-system-x86_64 -netdev tap,id=hn0,queues=256 -device
> > virtio-net-pci,netdev=hn0
> > 
> > Fixing this by doing the validation and fail early.
> > 
> > Cc: Michael S. Tsirkin <mst@redhat.com>
> > Cc: qemu-stable <qemu-stable@nongnu.org>
> > Signed-off-by: Jason Wang <jasowang@redhat.com>
> > ---
> > Changes from V1:
> > Tweak the commit log.
> > ---
> >  hw/net/virtio-net.c | 7 +++++++
> >  1 file changed, 7 insertions(+)
> > 
> > diff --git a/hw/net/virtio-net.c b/hw/net/virtio-net.c
> > index 27adcc5..59f76bc 100644
> > --- a/hw/net/virtio-net.c
> > +++ b/hw/net/virtio-net.c
> > @@ -1588,6 +1588,13 @@ static void virtio_net_device_realize(DeviceState *dev, Error **errp)
> >      virtio_init(vdev, "virtio-net", VIRTIO_ID_NET, n->config_size);
> >  
> >      n->max_queues = MAX(n->nic_conf.peers.queues, 1);
> > +    if (n->max_queues * 2 + 1 > VIRTIO_PCI_QUEUE_MAX) {
> > +        error_setg(errp, "Invalid number of queues (= %" PRIu32 "), "
> > +                   "must be a postive integer less than %d.",
> > +                   n->max_queues, (VIRTIO_PCI_QUEUE_MAX - 1) / 2);
> > +        virtio_cleanup(vdev);
> > +        return;
> > +    }
> 
> If VIRTIO_PCI_QUEUE_MAX is really PCI-only then this is a layering
> violation.  Virtio devices should not assume a particular transport
> (PCI, CCW, etc).
> 
> Either the constant should be renamed if it truly applies to all
> transports, or there should be a virtio_get_max_queues() function that
> lets the transport pick a value.
> 
> That said, virtio-scsi.c already checks VIRTIO_PCI_QUEUE_MAX so maybe
> this patch is fine and a cleanup should be written later:
> 
> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>

Yes, we need to rename this constant, no problem here.