2023-03-09 07:52:37

by Angus Chen

Subject: [PATCH] virtio_net: Use NETDEV_TX_BUSY when has no buf to send

Don't consume skb if virtqueue_add return -ENOSPC.

Signed-off-by: Angus Chen <[email protected]>
---
drivers/net/virtio_net.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index fb5e68ed3ec2..4096ea3d2eb6 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -1980,7 +1980,7 @@ static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
qnum, err);
dev->stats.tx_dropped++;
dev_kfree_skb_any(skb);
- return NETDEV_TX_OK;
+ return (err == -ENOSPC) ? NETDEV_TX_BUSY : NETDEV_TX_OK;
}

/* Don't wait up for transmitted skbs to be freed. */
--
2.25.1

2023-03-09 10:52:43

by Yunsheng Lin

Subject: Re: [PATCH] virtio_net: Use NETDEV_TX_BUSY when has no buf to send

On 2023/3/9 15:49, Angus Chen wrote:
> Don't consume skb if virtqueue_add return -ENOSPC.

Is this fixing the same out of space problem caused by
xdp xmit as Xuan Zhuo is fixing, or it is fixing a different
case?

https://lore.kernel.org/netdev/[email protected]/T/#mc4c5766a59fb8be988bb6a4dfa48f49e58df3ea6

>
> Signed-off-by: Angus Chen <[email protected]>
> ---
> drivers/net/virtio_net.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> index fb5e68ed3ec2..4096ea3d2eb6 100644
> --- a/drivers/net/virtio_net.c
> +++ b/drivers/net/virtio_net.c
> @@ -1980,7 +1980,7 @@ static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
> qnum, err);
> dev->stats.tx_dropped++;
> dev_kfree_skb_any(skb);

Returning NETDEV_TX_BUSY will caused stack to requeue the skb
and send it again when space is available, but you have freed the skb here,
isn't this cause use-after-free problem?

> - return NETDEV_TX_OK;
> + return (err == -ENOSPC) ? NETDEV_TX_BUSY : NETDEV_TX_OK;
> }
>
> /* Don't wait up for transmitted skbs to be freed. */
>

2023-03-09 12:26:22

by Angus Chen

Subject: RE: [PATCH] virtio_net: Use NETDEV_TX_BUSY when has no buf to send

Hi

> -----Original Message-----
> From: Yunsheng Lin <[email protected]>
> Sent: Thursday, March 9, 2023 6:53 PM
> To: Angus Chen <[email protected]>; [email protected];
> [email protected]; [email protected]; [email protected];
> [email protected]; [email protected]; [email protected]; [email protected];
> [email protected]; [email protected]
> Cc: [email protected]; [email protected];
> [email protected]; [email protected]; Xuan Zhuo
> <[email protected]>
> Subject: Re: [PATCH] virtio_net: Use NETDEV_TX_BUSY when has no buf to send
>
> On 2023/3/9 15:49, Angus Chen wrote:
> > Don't consume skb if virtqueue_add return -ENOSPC.
>
> Is this fixing the same out of space problem caused by
> xdp xmit as Xuan Zhuo is fixing, or it is fixing a different
> case?
Yes, the same print info.
>
> https://lore.kernel.org/netdev/20230308071921-mutt-send-email-mst@kernel.
> org/T/#mc4c5766a59fb8be988bb6a4dfa48f49e58df3ea6
>
En,I haven't noticed this patch before.
> >
> > Signed-off-by: Angus Chen <[email protected]>
> > ---
> > drivers/net/virtio_net.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> > index fb5e68ed3ec2..4096ea3d2eb6 100644
> > --- a/drivers/net/virtio_net.c
> > +++ b/drivers/net/virtio_net.c
> > @@ -1980,7 +1980,7 @@ static netdev_tx_t start_xmit(struct sk_buff *skb,
> struct net_device *dev)
> > qnum, err);
> > dev->stats.tx_dropped++;
> > dev_kfree_skb_any(skb);
>
> Returning NETDEV_TX_BUSY will caused stack to requeue the skb
> and send it again when space is available, but you have freed the skb here,
> isn't this cause use-after-free problem?
Yes，you are right, would be crash after I tested more times .I will drop it,thank you.
>
> > - return NETDEV_TX_OK;
> > + return (err == -ENOSPC) ? NETDEV_TX_BUSY : NETDEV_TX_OK;
> > }
> >
> > /* Don't wait up for transmitted skbs to be freed. */
> >