[PATCH net-next v7] vhost_net: avoid tx queue stuck when sendmsg fails

[PATCH net-next v7] vhost_net: avoid tx queue stuck when sendmsg fails

[wangyunjian]

From: Yunjian Wang <wangyunjian@huawei.com>

Currently the driver doesn't drop a packet which can't be sent by tun
(e.g bad packet). In this case, the driver will always process the
same packet lead to the tx queue stuck.

To fix this issue:
1. in the case of persistent failure (e.g bad packet), the driver
   can skip this descriptor by ignoring the error.
2. in the case of transient failure (e.g -ENOBUFS, -EAGAIN and -ENOMEM),
   the driver schedules the worker to try again.

Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
---
v7:
   * code rebase
v6:
   * update code styles and commit log
---
 drivers/vhost/net.c | 26 ++++++++++++++------------
 1 file changed, 14 insertions(+), 12 deletions(-)

diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c
index 3b744031ec8f..df82b124170e 100644
--- a/drivers/vhost/net.c
+++ b/drivers/vhost/net.c
@@ -828,14 +828,15 @@ static void handle_tx_copy(struct vhost_net *net, struct socket *sock)
 				msg.msg_flags &= ~MSG_MORE;
 		}
 
-		/* TODO: Check specific error and bomb out unless ENOBUFS? */
 		err = sock->ops->sendmsg(sock, &msg, len);
 		if (unlikely(err < 0)) {
-			vhost_discard_vq_desc(vq, 1);
-			vhost_net_enable_vq(net, vq);
-			break;
-		}
-		if (err != len)
+			if (err == -EAGAIN || err == -ENOMEM || err == -ENOBUFS) {
+				vhost_discard_vq_desc(vq, 1);
+				vhost_net_enable_vq(net, vq);
+				break;
+			}
+			pr_debug("Fail to send packet: err %d", err);
+		} else if (unlikely(err != len))
 			pr_debug("Truncated TX packet: len %d != %zd\n",
 				 err, len);
 done:
@@ -924,7 +925,6 @@ static void handle_tx_zerocopy(struct vhost_net *net, struct socket *sock)
 			msg.msg_flags &= ~MSG_MORE;
 		}
 
-		/* TODO: Check specific error and bomb out unless ENOBUFS? */
 		err = sock->ops->sendmsg(sock, &msg, len);
 		if (unlikely(err < 0)) {
 			if (zcopy_used) {
@@ -933,11 +933,13 @@ static void handle_tx_zerocopy(struct vhost_net *net, struct socket *sock)
 				nvq->upend_idx = ((unsigned)nvq->upend_idx - 1)
 					% UIO_MAXIOV;
 			}
-			vhost_discard_vq_desc(vq, 1);
-			vhost_net_enable_vq(net, vq);
-			break;
-		}
-		if (err != len)
+			if (err == -EAGAIN || err == -ENOMEM || err == -ENOBUFS) {
+				vhost_discard_vq_desc(vq, 1);
+				vhost_net_enable_vq(net, vq);
+				break;
+			}
+			pr_debug("Fail to send packet: err %d", err);
+		} else if (unlikely(err != len))
 			pr_debug("Truncated TX packet: "
 				 " len %d != %zd\n", err, len);
 		if (!zcopy_used)
-- 
2.23.0

Re: [PATCH net-next v7] vhost_net: avoid tx queue stuck when sendmsg fails

[Jason Wang]

On 2021/1/15 下午12:46, wangyunjian wrote:
> From: Yunjian Wang <wangyunjian@huawei.com>
>
> Currently the driver doesn't drop a packet which can't be sent by tun
> (e.g bad packet). In this case, the driver will always process the
> same packet lead to the tx queue stuck.
>
> To fix this issue:
> 1. in the case of persistent failure (e.g bad packet), the driver
>     can skip this descriptor by ignoring the error.
> 2. in the case of transient failure (e.g -ENOBUFS, -EAGAIN and -ENOMEM),
>     the driver schedules the worker to try again.
>
> Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>


Acked-by: Jason Wang <jasowang@redhat.com>


> ---
> v7:
>     * code rebase
> v6:
>     * update code styles and commit log
> ---
>   drivers/vhost/net.c | 26 ++++++++++++++------------
>   1 file changed, 14 insertions(+), 12 deletions(-)
>
> diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c
> index 3b744031ec8f..df82b124170e 100644
> --- a/drivers/vhost/net.c
> +++ b/drivers/vhost/net.c
> @@ -828,14 +828,15 @@ static void handle_tx_copy(struct vhost_net *net, struct socket *sock)
>   				msg.msg_flags &= ~MSG_MORE;
>   		}
>   
> -		/* TODO: Check specific error and bomb out unless ENOBUFS? */
>   		err = sock->ops->sendmsg(sock, &msg, len);
>   		if (unlikely(err < 0)) {
> -			vhost_discard_vq_desc(vq, 1);
> -			vhost_net_enable_vq(net, vq);
> -			break;
> -		}
> -		if (err != len)
> +			if (err == -EAGAIN || err == -ENOMEM || err == -ENOBUFS) {
> +				vhost_discard_vq_desc(vq, 1);
> +				vhost_net_enable_vq(net, vq);
> +				break;
> +			}
> +			pr_debug("Fail to send packet: err %d", err);
> +		} else if (unlikely(err != len))
>   			pr_debug("Truncated TX packet: len %d != %zd\n",
>   				 err, len);
>   done:
> @@ -924,7 +925,6 @@ static void handle_tx_zerocopy(struct vhost_net *net, struct socket *sock)
>   			msg.msg_flags &= ~MSG_MORE;
>   		}
>   
> -		/* TODO: Check specific error and bomb out unless ENOBUFS? */
>   		err = sock->ops->sendmsg(sock, &msg, len);
>   		if (unlikely(err < 0)) {
>   			if (zcopy_used) {
> @@ -933,11 +933,13 @@ static void handle_tx_zerocopy(struct vhost_net *net, struct socket *sock)
>   				nvq->upend_idx = ((unsigned)nvq->upend_idx - 1)
>   					% UIO_MAXIOV;
>   			}
> -			vhost_discard_vq_desc(vq, 1);
> -			vhost_net_enable_vq(net, vq);
> -			break;
> -		}
> -		if (err != len)
> +			if (err == -EAGAIN || err == -ENOMEM || err == -ENOBUFS) {
> +				vhost_discard_vq_desc(vq, 1);
> +				vhost_net_enable_vq(net, vq);
> +				break;
> +			}
> +			pr_debug("Fail to send packet: err %d", err);
> +		} else if (unlikely(err != len))
>   			pr_debug("Truncated TX packet: "
>   				 " len %d != %zd\n", err, len);
>   		if (!zcopy_used)

Re: [PATCH net-next v7] vhost_net: avoid tx queue stuck when sendmsg fails

[Willem de Bruijn]

On Fri, Jan 15, 2021 at 1:12 AM Jason Wang <jasowang@redhat.com> wrote:
>
>
> On 2021/1/15 下午12:46, wangyunjian wrote:
> > From: Yunjian Wang <wangyunjian@huawei.com>
> >
> > Currently the driver doesn't drop a packet which can't be sent by tun
> > (e.g bad packet). In this case, the driver will always process the
> > same packet lead to the tx queue stuck.
> >
> > To fix this issue:
> > 1. in the case of persistent failure (e.g bad packet), the driver
> >     can skip this descriptor by ignoring the error.
> > 2. in the case of transient failure (e.g -ENOBUFS, -EAGAIN and -ENOMEM),
> >     the driver schedules the worker to try again.
> >
> > Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
>
>
> Acked-by: Jason Wang <jasowang@redhat.com>

Acked-by: Willem de Bruijn <willemb@google.com>

Re: [PATCH net-next v7] vhost_net: avoid tx queue stuck when sendmsg fails

[Jakub Kicinski]

On Fri, 15 Jan 2021 12:46:20 +0800 wangyunjian wrote:
> From: Yunjian Wang <wangyunjian@huawei.com>
> 
> Currently the driver doesn't drop a packet which can't be sent by tun
> (e.g bad packet). In this case, the driver will always process the
> same packet lead to the tx queue stuck.
> 
> To fix this issue:
> 1. in the case of persistent failure (e.g bad packet), the driver
>    can skip this descriptor by ignoring the error.
> 2. in the case of transient failure (e.g -ENOBUFS, -EAGAIN and -ENOMEM),
>    the driver schedules the worker to try again.
> 
> Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>

Michael, LMK if you want to have a closer look otherwise I'll apply
tomorrow.

Re: [PATCH net-next v7] vhost_net: avoid tx queue stuck when sendmsg fails

[Michael S. Tsirkin]

On Mon, Jan 18, 2021 at 02:33:29PM -0800, Jakub Kicinski wrote:
> On Fri, 15 Jan 2021 12:46:20 +0800 wangyunjian wrote:
> > From: Yunjian Wang <wangyunjian@huawei.com>
> > 
> > Currently the driver doesn't drop a packet which can't be sent by tun
> > (e.g bad packet). In this case, the driver will always process the
> > same packet lead to the tx queue stuck.
> > 
> > To fix this issue:
> > 1. in the case of persistent failure (e.g bad packet), the driver
> >    can skip this descriptor by ignoring the error.
> > 2. in the case of transient failure (e.g -ENOBUFS, -EAGAIN and -ENOMEM),
> >    the driver schedules the worker to try again.
> > 
> > Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
> 
> Michael, LMK if you want to have a closer look otherwise I'll apply
> tomorrow.

Thanks for the reminder. Acked.

-- 
MST

Re: [PATCH net-next v7] vhost_net: avoid tx queue stuck when sendmsg fails

[Jakub Kicinski]

On Tue, 19 Jan 2021 04:56:59 -0500 Michael S. Tsirkin wrote:
> On Mon, Jan 18, 2021 at 02:33:29PM -0800, Jakub Kicinski wrote:
> > On Fri, 15 Jan 2021 12:46:20 +0800 wangyunjian wrote:  
> > > From: Yunjian Wang <wangyunjian@huawei.com>
> > > 
> > > Currently the driver doesn't drop a packet which can't be sent by tun
> > > (e.g bad packet). In this case, the driver will always process the
> > > same packet lead to the tx queue stuck.
> > > 
> > > To fix this issue:
> > > 1. in the case of persistent failure (e.g bad packet), the driver
> > >    can skip this descriptor by ignoring the error.
> > > 2. in the case of transient failure (e.g -ENOBUFS, -EAGAIN and -ENOMEM),
> > >    the driver schedules the worker to try again.
> > > 
> > > Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>  
> > 
> > Michael, LMK if you want to have a closer look otherwise I'll apply
> > tomorrow.  
> 
> Thanks for the reminder. Acked.

Applied, thanks everyone!

Re: [PATCH net-next v7] vhost_net: avoid tx queue stuck when sendmsg fails

[Michael S. Tsirkin]

On Fri, Jan 15, 2021 at 12:46:20PM +0800, wangyunjian wrote:
> From: Yunjian Wang <wangyunjian@huawei.com>
> 
> Currently the driver doesn't drop a packet which can't be sent by tun
> (e.g bad packet). In this case, the driver will always process the
> same packet lead to the tx queue stuck.
> 
> To fix this issue:
> 1. in the case of persistent failure (e.g bad packet), the driver
>    can skip this descriptor by ignoring the error.
> 2. in the case of transient failure (e.g -ENOBUFS, -EAGAIN and -ENOMEM),
>    the driver schedules the worker to try again.
> 
> Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>

Acked-by: Michael S. Tsirkin <mst@redhat.com>

> ---
> v7:
>    * code rebase
> v6:
>    * update code styles and commit log
> ---
>  drivers/vhost/net.c | 26 ++++++++++++++------------
>  1 file changed, 14 insertions(+), 12 deletions(-)
> 
> diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c
> index 3b744031ec8f..df82b124170e 100644
> --- a/drivers/vhost/net.c
> +++ b/drivers/vhost/net.c
> @@ -828,14 +828,15 @@ static void handle_tx_copy(struct vhost_net *net, struct socket *sock)
>  				msg.msg_flags &= ~MSG_MORE;
>  		}
>  
> -		/* TODO: Check specific error and bomb out unless ENOBUFS? */
>  		err = sock->ops->sendmsg(sock, &msg, len);
>  		if (unlikely(err < 0)) {
> -			vhost_discard_vq_desc(vq, 1);
> -			vhost_net_enable_vq(net, vq);
> -			break;
> -		}
> -		if (err != len)
> +			if (err == -EAGAIN || err == -ENOMEM || err == -ENOBUFS) {
> +				vhost_discard_vq_desc(vq, 1);
> +				vhost_net_enable_vq(net, vq);
> +				break;
> +			}
> +			pr_debug("Fail to send packet: err %d", err);
> +		} else if (unlikely(err != len))
>  			pr_debug("Truncated TX packet: len %d != %zd\n",
>  				 err, len);
>  done:
> @@ -924,7 +925,6 @@ static void handle_tx_zerocopy(struct vhost_net *net, struct socket *sock)
>  			msg.msg_flags &= ~MSG_MORE;
>  		}
>  
> -		/* TODO: Check specific error and bomb out unless ENOBUFS? */
>  		err = sock->ops->sendmsg(sock, &msg, len);
>  		if (unlikely(err < 0)) {
>  			if (zcopy_used) {
> @@ -933,11 +933,13 @@ static void handle_tx_zerocopy(struct vhost_net *net, struct socket *sock)
>  				nvq->upend_idx = ((unsigned)nvq->upend_idx - 1)
>  					% UIO_MAXIOV;
>  			}
> -			vhost_discard_vq_desc(vq, 1);
> -			vhost_net_enable_vq(net, vq);
> -			break;
> -		}
> -		if (err != len)
> +			if (err == -EAGAIN || err == -ENOMEM || err == -ENOBUFS) {
> +				vhost_discard_vq_desc(vq, 1);
> +				vhost_net_enable_vq(net, vq);
> +				break;
> +			}
> +			pr_debug("Fail to send packet: err %d", err);
> +		} else if (unlikely(err != len))
>  			pr_debug("Truncated TX packet: "
>  				 " len %d != %zd\n", err, len);
>  		if (!zcopy_used)
> -- 
> 2.23.0