On 1/9/26 17:18, Stefano Garzarella wrote:
> On Thu, Jan 08, 2026 at 10:54:54AM +0100, Michal Luczaj wrote:
...
>> @@ -1375,7 +1375,8 @@ virtio_transport_recv_enqueue(struct vsock_sock *vsk,
>> * of a new message.
>> */
>> if (skb->len < skb_tailroom(last_skb) &&
>> - !(le32_to_cpu(last_hdr->flags) & VIRTIO_VSOCK_SEQ_EOM)) {
>> + !(le32_to_cpu(last_hdr->flags) & VIRTIO_VSOCK_SEQ_EOM) &&
>> + !skb_is_nonlinear(skb)) {
>
> Why here? I mean we can do the check even early, something like this:
>
> --- a/net/vmw_vsock/virtio_transport_common.c
> +++ b/net/vmw_vsock/virtio_transport_common.c
> @@ -1361,7 +1361,8 @@ virtio_transport_recv_enqueue(struct vsock_sock *vsk,
> * to avoid wasting memory queueing the entire buffer with a small
> * payload.
> */
> - if (len <= GOOD_COPY_LEN && !skb_queue_empty(&vvs->rx_queue)) {
> + if (len <= GOOD_COPY_LEN && !skb_queue_empty(&vvs->rx_queue) &&
> + !skb_is_nonlinear(skb)) {
> struct virtio_vsock_hdr *last_hdr;
> struct sk_buff *last_skb;
Right, can do. I've assumed skb being non-linear is the least likely in
this context.
> I would also add the reason in the comment before that to make it clear.
OK, sure.
