Re: [RFC PATCH v2] virtio/vsock: allocate multiple skbuffs on tx

[Stefano Garzarella]

On Sun, Mar 19, 2023 at 09:46:10PM +0300, Arseniy Krasnov wrote:
> This adds small optimization for tx path: instead of allocating single
> skbuff on every call to transport, allocate multiple skbuff's until
> credit space allows, thus trying to send as much as possible data without
> return to af_vsock.c.
>
> Signed-off-by: Arseniy Krasnov <AVKrasnov@xxxxxxxxxxxxxx>
> ---
> Link to v1:
> https://lore.kernel.org/netdev/2c52aa26-8181-d37a-bccd-a86bd3cbc6e1@xxxxxxxxxxxxxx/
>
> Changelog:
> v1 -> v2:
> - If sent something, return number of bytes sent (even in
>   case of error). Return error only if failed to sent first
>   skbuff.
>
> net/vmw_vsock/virtio_transport_common.c | 53 ++++++++++++++++++-------
> 1 file changed, 39 insertions(+), 14 deletions(-)
>
> diff --git a/net/vmw_vsock/virtio_transport_common.c b/net/vmw_vsock/virtio_transport_common.c
> index 6564192e7f20..3fdf1433ec28 100644
> --- a/net/vmw_vsock/virtio_transport_common.c
> +++ b/net/vmw_vsock/virtio_transport_common.c
> @@ -196,7 +196,8 @@ static int virtio_transport_send_pkt_info(struct vsock_sock *vsk,
> 	const struct virtio_transport *t_ops;
> 	struct virtio_vsock_sock *vvs;
> 	u32 pkt_len = info->pkt_len;
> -	struct sk_buff *skb;
> +	u32 rest_len;
> +	int ret;
>
> 	info->type = virtio_transport_get_type(sk_vsock(vsk));
>
> @@ -216,10 +217,6 @@ static int virtio_transport_send_pkt_info(struct vsock_sock *vsk,
>
> 	vvs = vsk->trans;
>
> -	/* we can send less than pkt_len bytes */
> -	if (pkt_len > VIRTIO_VSOCK_MAX_PKT_BUF_SIZE)
> -		pkt_len = VIRTIO_VSOCK_MAX_PKT_BUF_SIZE;
> -
> 	/* virtio_transport_get_credit might return less than pkt_len credit */
> 	pkt_len = virtio_transport_get_credit(vvs, pkt_len);
>
> @@ -227,17 +224,45 @@ static int virtio_transport_send_pkt_info(struct vsock_sock *vsk,
> 	if (pkt_len == 0 && info->op == VIRTIO_VSOCK_OP_RW)
> 		return pkt_len;
>
> -	skb = virtio_transport_alloc_skb(info, pkt_len,
> -					 src_cid, src_port,
> -					 dst_cid, dst_port);
> -	if (!skb) {
> -		virtio_transport_put_credit(vvs, pkt_len);
> -		return -ENOMEM;
> -	}
> +	ret = 0;
> +	rest_len = pkt_len;
> +
> +	do {
> +		struct sk_buff *skb;
> +		size_t skb_len;
> +
> +		skb_len = min_t(u32, VIRTIO_VSOCK_MAX_PKT_BUF_SIZE, rest_len);
> +
> +		skb = virtio_transport_alloc_skb(info, skb_len,
> +						 src_cid, src_port,
> +						 dst_cid, dst_port);
> +		if (!skb) {
> +			ret = -ENOMEM;
> +			break;
> +		}
> +
> +		virtio_transport_inc_tx_pkt(vvs, skb);
> +
> +		ret = t_ops->send_pkt(skb);
> +
> +		if (ret < 0)
> +			break;
>
> -	virtio_transport_inc_tx_pkt(vvs, skb);
> +		rest_len -= skb_len;

t_ops->send_pkt() is returning the number of bytes sent. Current
implementations always return `skb_len`, so there should be no problem,
but it would be better to put a comment here, or we should handle the
case where ret != skb_len to avoid future issues.

> +	} while (rest_len);
>
> -	return t_ops->send_pkt(skb);
> +	/* Don't call this function with zero as argument:
> +	 * it tries to acquire spinlock and such argument
> +	 * makes this call useless.

Good point, can we do the same also for virtio_transport_get_credit()?
(Maybe in a separate patch)

I'm thinking if may be better to do it directly inside the functions,
but I don't have a strong opinion on that since we only call them here.

Thanks,
Stefano

> +	 */
> +	if (rest_len)
> +		virtio_transport_put_credit(vvs, rest_len);
> +
> +	/* Return number of bytes, if any data has been sent. */
> +	if (rest_len != pkt_len)
> +		ret = pkt_len - rest_len;
> +
> +	return ret;
> }
>
> static bool virtio_transport_inc_rx_pkt(struct virtio_vsock_sock *vvs,
> --
> 2.25.1