Kernels that do not have f6b10209b90d ("virtio-net: switch to use build_skb() for small buffer") will have an extra call to skb_to_sgvec that is not handled by e2fcad58fd23 ("virtio_net: check return value of skb_to_sgvec always"). Since the former does not appear to be stable material, just fix the call up directly.
Cc: Jason A. Donenfeld Jason@zx2c4.com Cc: Sergei Shtylyov sergei.shtylyov@cogentembedded.com Cc: "Michael S. Tsirkin" mst@redhat.com Cc: Jason Wang jasowang@redhat.com Cc: David S. Miller davem@davemloft.net Signed-off-by: Nathan Chancellor natechancellor@gmail.com ---
Greg, if you feel this should be squashed into the previous commit, by all means, feel free. I am not entirely sure if the dev_kfree_skb is necessary here, I was hoping to get some review.
--- drivers/net/virtio_net.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c index ede40841e962..1da54735681b 100644 --- a/drivers/net/virtio_net.c +++ b/drivers/net/virtio_net.c @@ -551,7 +551,12 @@ static int add_recvbuf_small(struct receive_queue *rq, gfp_t gfp) hdr = skb_vnet_hdr(skb); sg_init_table(rq->sg, MAX_SKB_FRAGS + 2); sg_set_buf(rq->sg, &hdr->hdr, sizeof hdr->hdr); - skb_to_sgvec(skb, rq->sg + 1, 0, skb->len); + + err = skb_to_sgvec(skb, rq->sg + 1, 0, skb->len); + if (unlikely(err < 0)) { + dev_kfree_skb(skb); + return err; + }
err = virtqueue_add_inbuf(rq->vq, rq->sg, 2, skb, gfp); if (err < 0)