Re: [PATCH net-next v14 2/5] tun/tap: add ptr_ring consume helper with netdev queue wakeup

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] vhost-net bypasses `tun_ring_consume()`, causing permanent TX queue stalls when `IFF_BACKPRESSURE` is used.
--

--- Patch [2]: [PATCH net-next v14 2/5] tun/tap: add ptr_ring consume helper with netdev queue wakeup ---
commit 751c5819e676787b1b9d745ec0ad8353ee77415e
Author: Simon Schippers <simon.schippers@tu-dortmund.de>

tun/tap: add ptr_ring consume helper with netdev queue wakeup

This commit introduces tun_ring_consume() that wraps ptr_ring_consume() and
calls __tun_wake_queue(). The queue is woken up when the ring is half empty
or entirely empty to support traffic shaping with IFF_BACKPRESSURE.

> diff --git a/drivers/net/tun.c b/drivers/net/tun.c
> index 51e80000bd0ed..d49b6bfd104db 100644
> --- a/drivers/net/tun.c
> +++ b/drivers/net/tun.c
[ ... ]
> @@ -2116,13 +2171,61 @@ static ssize_t tun_put_user(struct tun_struct *tun,
[ ... ]
> +static void *tun_ring_consume(struct tun_struct *tun, struct tun_file *tfile)
> +{
> +	void *ptr;
> +
> +	spin_lock(&tfile->tx_ring.consumer_lock);
> +	ptr = __ptr_ring_consume(&tfile->tx_ring);
> +	if (ptr)
> +		__tun_wake_queue(tun, tfile, 1);
> +
> +	spin_unlock(&tfile->tx_ring.consumer_lock);
> +	return ptr;
> +}

[Severity: Medium]
Will vhost-net bypass this wrapper and cause permanent TX queue stalls when
IFF_BACKPRESSURE is enabled?

Looking at vhost_net_buf_produce() in drivers/vhost/net.c, vhost-net consumes
packets directly from the tun ring:

drivers/vhost/net.c:vhost_net_buf_produce() {
    ...
	rxq->tail = ptr_ring_consume_batched(nvq->rx_ring, rxq->queue,
					      VHOST_NET_BATCH);
    ...
}

Because vhost-net uses ptr_ring_consume_batched() directly, it bypasses
tun_ring_consume() and prevents __tun_wake_queue() from being called. This
leaves the stopped netdev queue permanently stalled.

While IFF_BACKPRESSURE is currently masked out of TUN_FEATURES in this commit
(meaning it cannot be triggered by userspace today), this structural flaw in
the backpressure design will trigger unconditionally once the feature flag is
fully enabled.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260803183641.96882-1-simon.schippers@tu-dortmund.de?part=2