Re: [PATCH v2] virtio: Work around frames incorrectly marked as gso

On 24/02/2020 12:46, Michael S. Tsirkin wrote:

On Mon, Feb 24, 2020 at 10:19:12AM +0000, anton.ivanov@cambridgegreys.com wrote:

quoted

From: Anton Ivanov <anton.ivanov@cambridgegreys.com>

Some of the locally generated frames marked as GSO which
arrive at virtio_net_hdr_from_skb() have no GSO_TYPE, no
fragments (data_len = 0) and length significantly shorter
than the MTU (752 in my experiments).

This is observed on raw sockets reading off vEth interfaces
in all 4.x and 5.x kernels I tested.

A bit more info on how to reproduce couldn't hurt here.

Will do, a v3 will follow shortly.

quoted

These frames are reported as invalid while they are in fact
gso-less frames.

This patch marks the vnet header as no-GSO for them instead
of reporting it as invalid.

Signed-off-by: Anton Ivanov <anton.ivanov@cambridgegreys.com>

Acked-by: Michael S. Tsirkin <mst@redhat.com>

Eric - as you looked at this in the past, would you mind acking please?

quoted

---
  include/linux/virtio_net.h | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/include/linux/virtio_net.h b/include/linux/virtio_net.h
index 0d1fe9297ac6..94fb78c3a2ab 100644
--- a/include/linux/virtio_net.h
+++ b/include/linux/virtio_net.h

@@ -100,8 +100,8 @@ static inline int virtio_net_hdr_from_skb(const struct sk_buff *skb,
  {
  	memset(hdr, 0, sizeof(*hdr));   /* no info leak */
  
-	if (skb_is_gso(skb)) {
-		struct skb_shared_info *sinfo = skb_shinfo(skb);
+	struct skb_shared_info *sinfo = skb_shinfo(skb);

I need to move this a few lines up - the kernel build robot is quite rightfully complaining.

quoted

+	if (skb_is_gso(skb) && sinfo->gso_type) {
  
  		/* This is a hint as to how much should be linear. */
  		hdr->hdr_len = __cpu_to_virtio16(little_endian,
-- 
2.20.1

-- 
Anton R. Ivanov
Cambridgegreys Limited. Registered in England. Company Number 10273661
https://www.cambridgegreys.com/