Re: [PATCH 2/2] tcp_vu, udp_vu: Account for virtio net header in minimum frame size

[Stefano Brivio <sbrivio@redhat.com>]

On Thu, 12 Feb 2026 12:39:32 +0100
Laurent Vivier <lvivier@redhat.com> wrote:

> In the vhost-user paths, the buffers provided by the virtio queue
> include the virtio net header (VNET_HLEN) prepended to the Ethernet
> frame. The minimum size checks using ETH_ZLEN must therefore account
> for this additional header length, otherwise we underestimate the
> minimum buffer size needed.

Oops, thanks for fixing this.

> Use ETH_ZLEN + VNET_HLEN instead of bare ETH_ZLEN in vu_collect()
> calls and the corresponding ASSERT() checks.
> 
> Fixes: 0cb8f9003654 ("tcp, udp: Pad batched frames for vhost-user modes to 60 bytes (802.3 minimum)")
> Signed-off-by: Laurent Vivier <lvivier@redhat.com>
> ---
>  tcp_vu.c | 8 ++++----
>  udp_vu.c | 2 +-
>  2 files changed, 5 insertions(+), 5 deletions(-)
> 
> diff --git a/tcp_vu.c b/tcp_vu.c
> index f7bda4943e43..2d593d534d68 100644
> --- a/tcp_vu.c
> +++ b/tcp_vu.c
> @@ -90,12 +90,12 @@ int tcp_vu_send_flag(const struct ctx *c, struct tcp_tap_conn *conn, int flags)
>  	vu_set_element(&flags_elem[0], NULL, &flags_iov[0]);
>  
>  	elem_cnt = vu_collect(vdev, vq, &flags_elem[0], 1,
> -			      MAX(hdrlen + sizeof(*opts), ETH_ZLEN), NULL);
> +			      MAX(hdrlen + sizeof(*opts), ETH_ZLEN + VNET_HLEN), NULL);

I'm applying this now but, perhaps as a follow-up, should we consider
to clarify further the arguments to vu_collect(), and perhaps related
sizes?

I'm not the most indicated person to do this I guess, as I keep getting
them wrong. :(

In this case, I really thought VNET_HLEN would be added internally,
because of:

 * @size:		Maximum size of the data in the frame

-- 
Stefano