Reproducer from https://gitlab.com/qemu-project/qemu/-/issues/1451
creates small packet (1 segment, len = 10 == n->guest_hdr_len),
then destroys queue.

"if (n->host_hdr_len != n->guest_hdr_len)" is triggered, if body creates
zero length/zero segment packet as there is nothing after guest header.

qemu_sendv_packet_async() tries to send it.

slirp discards it because it is smaller than Ethernet header,
but returns 0 because tx hooks are supposed to return total length of data.

0 is propagated upwards and is interpreted as "packet has been sent"
which is terrible because queue is being destroyed, nobody is waiting for TX
to complete and assert it triggered.

Fix is discard such empty packets instead of sending them.

Length 1 packets will go via different codepath:

	virtqueue_push(q->tx_vq, elem, 0);
	virtio_notify(vdev, q->tx_vq);
	g_free(elem);

and aren't problematic.

Signed-off-by: Alexey Dobriyan <adobriyan@yandex-team.ru>
---

	hopefully better changelog.
	use "if (out_num < 1)" so that discard doesn't calculate iov length

 hw/net/virtio-net.c | 18 ++++++++++++------
 1 file changed, 12 insertions(+), 6 deletions(-)

diff --git a/hw/net/virtio-net.c b/hw/net/virtio-net.c
index 24e5e7d347..3644bfd91b 100644
--- a/hw/net/virtio-net.c
+++ b/hw/net/virtio-net.c
@@ -2749,18 +2749,14 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q)
         out_sg = elem->out_sg;
         if (out_num < 1) {
             virtio_error(vdev, "virtio-net header not in first element");
-            virtqueue_detach_element(q->tx_vq, elem, 0);
-            g_free(elem);
-            return -EINVAL;
+            goto detach;
         }
 
         if (n->has_vnet_hdr) {
             if (iov_to_buf(out_sg, out_num, 0, &vhdr, n->guest_hdr_len) <
                 n->guest_hdr_len) {
                 virtio_error(vdev, "virtio-net header incorrect");
-                virtqueue_detach_element(q->tx_vq, elem, 0);
-                g_free(elem);
-                return -EINVAL;
+                goto detach;
             }
             if (n->needs_vnet_hdr_swap) {
                 virtio_net_hdr_swap(vdev, (void *) &vhdr);
@@ -2791,6 +2787,11 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q)
                              n->guest_hdr_len, -1);
             out_num = sg_num;
             out_sg = sg;
+
+            if (out_num < 1) {
+                virtio_error(vdev, "virtio-net nothing to send");
+                goto detach;
+            }
         }
 
         ret = qemu_sendv_packet_async(qemu_get_subqueue(n->nic, queue_index),
@@ -2811,6 +2812,11 @@ drop:
         }
     }
     return num_packets;
+
+detach:
+    virtqueue_detach_element(q->tx_vq, elem, 0);
+    g_free(elem);
+    return -EINVAL;
 }
 
 static void virtio_net_tx_timer(void *opaque);
-- 
2.34.1

On Tue, Apr 30, 2024 at 6:54 PM Alexey Dobriyan
<adobriyan@yandex-team.ru> wrote:
>
> Reproducer from https://gitlab.com/qemu-project/qemu/-/issues/1451
> creates small packet (1 segment, len = 10 == n->guest_hdr_len),
> then destroys queue.
>
> "if (n->host_hdr_len != n->guest_hdr_len)" is triggered, if body creates
> zero length/zero segment packet as there is nothing after guest header.
>
> qemu_sendv_packet_async() tries to send it.
>
> slirp discards it because it is smaller than Ethernet header,
> but returns 0 because tx hooks are supposed to return total length of data.
>
> 0 is propagated upwards and is interpreted as "packet has been sent"
> which is terrible because queue is being destroyed, nobody is waiting for TX
> to complete and assert it triggered.
>
> Fix is discard such empty packets instead of sending them.
>
> Length 1 packets will go via different codepath:
>
>         virtqueue_push(q->tx_vq, elem, 0);
>         virtio_notify(vdev, q->tx_vq);
>         g_free(elem);
>
> and aren't problematic.
>
> Signed-off-by: Alexey Dobriyan <adobriyan@yandex-team.ru>
> ---
>
>         hopefully better changelog.
>         use "if (out_num < 1)" so that discard doesn't calculate iov length
>
>  hw/net/virtio-net.c | 18 ++++++++++++------
>  1 file changed, 12 insertions(+), 6 deletions(-)
>

I tweak the title to "drop too short packets early".

And queued.

Thanks