Re: [PATCH net-next v3] virtio-net: xsk: support tx wake up

["Michael S. Tsirkin" <mst@redhat.com>]

On Mon, Jun 22, 2026 at 08:27:12PM +0800, Menglong Dong wrote:
> On 2026/6/22 06:31 Michael S. Tsirkin <mst@redhat.com> write:
> > On Tue, Jun 16, 2026 at 07:59:12PM +0800, Menglong Dong wrote:
> [...]
> > >  
> > > +	vring_size = virtqueue_get_vring_size(sq->vq);
> > > +	need_wakeup = xsk_uses_need_wakeup(pool);
> > > +
> > > +	if (need_wakeup && vring_size == sq->vq->num_free)
> > > +		xsk_set_tx_need_wakeup(pool);
> > > +
> > 
> > why are we doing this here?
> > the check after virtnet_xsk_xmit_batch not enough?
> > I vaguely think it's some kind of race we are closing?
> > Pls add a comment to explain.
> 
> Hi, Michael. Thanks for your review.
> 
> Yeah, it's for a race condition between user space and kernel
> space. I added a comment in V2, which is too confusing, and
> I removed it 😢. I'll make it more clear and add it in the V4. The
> origin comment is:
> 
>  * If the sq->vq is empty, and the tx ring is empty, and the user
>  * submit an entry to the tx ring after virtnet_xsk_xmit_batch() and
>  * before xsk_set_tx_need_wakeup(), we will lose the chance to wake
>  * up the tx napi, so we have to set the need_wakeup flag here.
> 
> And the logic is like this:
> 
> Kernel: tx NAPI is waked up from skb_xmit_done() ->
> Kernel: sq->vq and xsk->tx_ring are both empty ->
> Kernel: call virtnet_xsk_xmit_batch()
> 
>     User: submit a entry to the xsk->tx_ring
>     User: check the wakeup flag
>     User: wakeup flag is not set, skip send()
> 
> Kernel: call xsk_set_tx_need_wakeup(), because sq->vq is empty
> 
> If we don't send more data, the data in the xsk->tx_ring will
> not be sent forever.

I'm not 100% sure I understand, but when someone fixes cross-CPU races
with no synchronization or CPU memory barriers just with extra checks,
this always gives me pause.

AI helped write this for me, for example:
  1. Kernel: xsk_set_tx_need_wakeup stores NEED_WAKEUP (sits in store buffer)
  2. Kernel: xsk_tx_peek_release_desc_batch - load, sees empty (reordered before the store is globally visible)
  3. Kernel: peek finds nothing, returns 0
  4. Userspace: stores entry + producer
  5. Userspace: loads flags - doesn't see NEED_WAKEUP yet (still in kernel's store buffer)
  6. Userspeace: skips send()
  7. Kernel: NEED_WAKEUP store finally becomes visible - too late

Seems legit?



> > 
> > >  	sent = virtnet_xsk_xmit_batch(sq, pool, budget, &kicks);
> > >  
> > > +	if (need_wakeup) {
> > > +		if (vring_size == sq->vq->num_free)
> > > +			/* we can't wake up by ourself, and it should be done
> > > +			 * by the user.
> > > +			 */
> > > +			xsk_set_tx_need_wakeup(pool);
> > > +		else
> > > +			/* we can wake up from skb_xmit_done() */
> > > +			xsk_clear_tx_need_wakeup(pool);
> > 
> > But what if we don't have get tx napi so no wakeup in skb_xmit_done?
> 
> Sorry that I'm not sure what "get tx napi" means here ;(
> 
> There are entry in sq->vq, so skb_xmit_done() will be called after
> the entries in the ring is consumed by the HOST, right?
> Then, the corresponding sq->napi will be scheduled, as we ensure
> that tx napi is always enabled, which means napi->weight is not
> zero, in this commit:
> 1df5116a41a8 ("virtio_net: xsk: prevent disable tx napi")

Oh I forgot we did that. But can xsk bind when tx napi has already
been disabled previously?


> Right?
> 
> Thanks!
> Menglong Dong
> 
> > 
> > 
> > > +	}
> > > +
> > >  	if (!is_xdp_raw_buffer_queue(vi, sq - vi->sq))
> > >  		check_sq_full_and_disable(vi, vi->dev, sq);
> > >  
> > > @@ -1470,9 +1488,6 @@ static bool virtnet_xsk_xmit(struct send_queue *sq, struct xsk_buff_pool *pool,
> > >  	u64_stats_add(&sq->stats.xdp_tx,  sent);
> > >  	u64_stats_update_end(&sq->stats.syncp);
> > >  
> > > -	if (xsk_uses_need_wakeup(pool))
> > > -		xsk_set_tx_need_wakeup(pool);
> > > -
> > >  	return sent;
> > >  }
> > >  
> > > -- 
> > > 2.54.0
> > 
> > 
> > 
> 
> 
>