netdev - Re: [PATCH net 3/3] virtio-net: schedule the pending refill work after being enabled

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20251224051636-mutt-send-email-mst@kernel.org>
Date: Wed, 24 Dec 2025 05:17:30 -0500
From: "Michael S. Tsirkin" <mst@...hat.com>
To: Bui Quang Minh <minhquangbui99@...il.com>
Cc: netdev@...r.kernel.org, Jason Wang <jasowang@...hat.com>,
	Xuan Zhuo <xuanzhuo@...ux.alibaba.com>,
	Eugenio Pérez <eperezma@...hat.com>,
	Andrew Lunn <andrew+netdev@...n.ch>,
	"David S. Miller" <davem@...emloft.net>,
	Eric Dumazet <edumazet@...gle.com>,
	Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
	Alexei Starovoitov <ast@...nel.org>,
	Daniel Borkmann <daniel@...earbox.net>,
	Jesper Dangaard Brouer <hawk@...nel.org>,
	John Fastabend <john.fastabend@...il.com>,
	Stanislav Fomichev <sdf@...ichev.me>,
	virtualization@...ts.linux.dev, linux-kernel@...r.kernel.org,
	bpf@...r.kernel.org
Subject: Re: [PATCH net 3/3] virtio-net: schedule the pending refill work
 after being enabled

On Tue, Dec 23, 2025 at 10:25:33PM +0700, Bui Quang Minh wrote:
> As we need to move the enable_delayed_refill after napi_enable, it's
> possible that a refill work needs to be scheduled in virtnet_receive but
> it cannot. This can make the receive side stuck because if we don't have
> any receive buffers, there will be nothing trigger the refill logic. So
> in case it happens, in virtnet_receive, set the rx queue's
> refill_pending, then when the refill work is enabled again, a refill
> work will be scheduled.
> 
> Signed-off-by: Bui Quang Minh <minhquangbui99@...il.com>

Sounds like this fixes a bug previous patch introduces?
The thing to do is to reorder these two patches then.

> ---
>  drivers/net/virtio_net.c | 21 ++++++++++++---------
>  1 file changed, 12 insertions(+), 9 deletions(-)
> 
> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> index 8016d2b378cf..ddc62dab2f9a 100644
> --- a/drivers/net/virtio_net.c
> +++ b/drivers/net/virtio_net.c
> @@ -383,6 +383,9 @@ struct receive_queue {
>  	/* Is delayed refill enabled? */
>  	bool refill_enabled;
>  
> +	/* A refill work needs to be scheduled when delayed refill is enabled */
> +	bool refill_pending;
> +
>  	/* The lock to synchronize the access to refill_enabled */
>  	spinlock_t refill_lock;
>  
> @@ -720,10 +723,13 @@ static void virtnet_rq_free_buf(struct virtnet_info *vi,
>  		put_page(virt_to_head_page(buf));
>  }
>  
> -static void enable_delayed_refill(struct receive_queue *rq)
> +static void enable_delayed_refill(struct receive_queue *rq,
> +				  bool schedule_refill)
>  {
>  	spin_lock_bh(&rq->refill_lock);
>  	rq->refill_enabled = true;
> +	if (rq->refill_pending || schedule_refill)
> +		schedule_delayed_work(&rq->refill, 0);
>  	spin_unlock_bh(&rq->refill_lock);
>  }
>  
> @@ -3032,6 +3038,8 @@ static int virtnet_receive(struct receive_queue *rq, int budget,
>  			spin_lock(&rq->refill_lock);
>  			if (rq->refill_enabled)
>  				schedule_delayed_work(&rq->refill, 0);
> +			else
> +				rq->refill_pending = true;
>  			spin_unlock(&rq->refill_lock);
>  		}
>  	}
> @@ -3228,11 +3236,8 @@ static int virtnet_open(struct net_device *dev)
>  		if (err < 0)
>  			goto err_enable_qp;
>  
> -		if (i < vi->curr_queue_pairs) {
> -			enable_delayed_refill(&vi->rq[i]);
> -			if (schedule_refill)
> -				schedule_delayed_work(&vi->rq[i].refill, 0);
> -		}
> +		if (i < vi->curr_queue_pairs)
> +			enable_delayed_refill(&vi->rq[i], schedule_refill);
>  	}
>  
>  	if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_STATUS)) {
> @@ -3480,9 +3485,7 @@ static void __virtnet_rx_resume(struct virtnet_info *vi,
>  	if (running)
>  		virtnet_napi_enable(rq);
>  
> -	enable_delayed_refill(rq);
> -	if (schedule_refill)
> -		schedule_delayed_work(&rq->refill, 0);
> +	enable_delayed_refill(rq, schedule_refill);
>  }
>  
>  static void virtnet_rx_resume_all(struct virtnet_info *vi)
> -- 
> 2.43.0