Re: [PATCH v2 1/3] virtio_net: Stop doing DMA from the stack

[Michael S. Tsirkin]

On Tue, Oct 27, 2015 at 10:30:19PM -0700, Andy Lutomirski wrote:
> From: Andy Lutomirski <luto@xxxxxxxxxxxxxx>
> 
> Once virtio starts using the DMA API, we won't be able to safely DMA
> from the stack.  virtio-net does a couple of config DMA requests
> from small stack buffers -- switch to using dynamically-allocated
> memory.
> 
> This should have no effect on any performance-critical code paths.
> 
> Cc: netdev@xxxxxxxxxxxxxxx
> Cc: "Michael S. Tsirkin" <mst@xxxxxxxxxx>
> Cc: virtualization@xxxxxxxxxxxxxxxxxxxxxxxxxx
> Reviewed-by: Joerg Roedel <jroedel@xxxxxxx>
> Signed-off-by: Andy Lutomirski <luto@xxxxxxxxxx>
> ---
> 
> Hi Michael and DaveM-
> 
> This is a prerequisite for the virtio DMA fixing project.  It works
> as a standalone patch, though.  Would it make sense to apply it to
> an appropriate networking tree now?
> 
>  drivers/net/virtio_net.c | 53 ++++++++++++++++++++++++++++++++----------------
>  1 file changed, 36 insertions(+), 17 deletions(-)
> 
> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> index d8838dedb7a4..4f10f8a58811 100644
> --- a/drivers/net/virtio_net.c
> +++ b/drivers/net/virtio_net.c
> @@ -976,31 +976,43 @@ static bool virtnet_send_command(struct virtnet_info *vi, u8 class, u8 cmd,
>  				 struct scatterlist *out)
>  {
>  	struct scatterlist *sgs[4], hdr, stat;
> -	struct virtio_net_ctrl_hdr ctrl;
> -	virtio_net_ctrl_ack status = ~0;
> +
> +	struct {
> +		struct virtio_net_ctrl_hdr ctrl;
> +		virtio_net_ctrl_ack status;
> +	} *buf;
> +
>  	unsigned out_num = 0, tmp;
> +	bool ret;
>  
>  	/* Caller should know better */
>  	BUG_ON(!virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ));
>  
> -	ctrl.class = class;
> -	ctrl.cmd = cmd;
> +	buf = kmalloc(sizeof(*buf), GFP_ATOMIC);
> +	if (!buf)
> +		return false;

This is problematic. The command is never retried, the error
is propagated to userspace.

> +	buf->status = ~0;
> +
> +	buf->ctrl.class = class;
> +	buf->ctrl.cmd = cmd;
>  	/* Add header */
> -	sg_init_one(&hdr, &ctrl, sizeof(ctrl));
> +	sg_init_one(&hdr, &buf->ctrl, sizeof(buf->ctrl));
>  	sgs[out_num++] = &hdr;
>  
>  	if (out)
>  		sgs[out_num++] = out;
>  
>  	/* Add return status. */
> -	sg_init_one(&stat, &status, sizeof(status));
> +	sg_init_one(&stat, &buf->status, sizeof(buf->status));
>  	sgs[out_num] = &stat;
>  
>  	BUG_ON(out_num + 1 > ARRAY_SIZE(sgs));
>  	virtqueue_add_sgs(vi->cvq, sgs, out_num, 1, vi, GFP_ATOMIC);
>  
> -	if (unlikely(!virtqueue_kick(vi->cvq)))
> -		return status == VIRTIO_NET_OK;
> +	if (unlikely(!virtqueue_kick(vi->cvq))) {
> +		ret = (buf->status == VIRTIO_NET_OK);
> +		goto out;
> +	}
>  
>  	/* Spin for a response, the kick causes an ioport write, trapping
>  	 * into the hypervisor, so the request should be handled immediately.
> @@ -1009,7 +1021,11 @@ static bool virtnet_send_command(struct virtnet_info *vi, u8 class, u8 cmd,
>  	       !virtqueue_is_broken(vi->cvq))
>  		cpu_relax();
>  
> -	return status == VIRTIO_NET_OK;
> +	ret = (buf->status == VIRTIO_NET_OK);
> +
> +out:
> +	kfree(buf);
> +	return ret;
>  }
>  
>  static int virtnet_set_mac_address(struct net_device *dev, void *p)
> @@ -1151,7 +1167,7 @@ static void virtnet_set_rx_mode(struct net_device *dev)
>  {
>  	struct virtnet_info *vi = netdev_priv(dev);
>  	struct scatterlist sg[2];
> -	u8 promisc, allmulti;
> +	u8 *cmdbyte;
>  	struct virtio_net_ctrl_mac *mac_data;
>  	struct netdev_hw_addr *ha;
>  	int uc_count;
> @@ -1163,22 +1179,25 @@ static void virtnet_set_rx_mode(struct net_device *dev)
>  	if (!virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_RX))
>  		return;
>  
> -	promisc = ((dev->flags & IFF_PROMISC) != 0);
> -	allmulti = ((dev->flags & IFF_ALLMULTI) != 0);
> +	cmdbyte = kmalloc(sizeof(*cmdbyte), GFP_ATOMIC);
> +	if (!cmdbyte)
> +		return;

Here the error is ignored, rx mode will be incorrect.
OTOH it looks like that's already the case.

>  
> -	sg_init_one(sg, &promisc, sizeof(promisc));
> +	sg_init_one(sg, cmdbyte, sizeof(*cmdbyte));
>  
> +	*cmdbyte = ((dev->flags & IFF_PROMISC) != 0);
>  	if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RX,
>  				  VIRTIO_NET_CTRL_RX_PROMISC, sg))
>  		dev_warn(&dev->dev, "Failed to %sable promisc mode.\n",
> -			 promisc ? "en" : "dis");
> -
> -	sg_init_one(sg, &allmulti, sizeof(allmulti));
> +			 *cmdbyte ? "en" : "dis");
>  
> +	*cmdbyte = ((dev->flags & IFF_ALLMULTI) != 0);
>  	if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RX,
>  				  VIRTIO_NET_CTRL_RX_ALLMULTI, sg))
>  		dev_warn(&dev->dev, "Failed to %sable allmulti mode.\n",
> -			 allmulti ? "en" : "dis");
> +			 *cmdbyte ? "en" : "dis");
> +
> +	kfree(cmdbyte);
>  
>  	uc_count = netdev_uc_count(dev);
>  	mc_count = netdev_mc_count(dev);

How about this instead? Less code, more robust.

Warning: untested.  If you do like this approach, Tested-by would be
appreciated.

Signed-off-by: Michael S. Tsirkin <mst@xxxxxxxxxx>

---

diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index d8838ded..f94ab78 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -140,6 +140,12 @@ struct virtnet_info {
 
 	/* CPU hot plug notifier */
 	struct notifier_block nb;
+
+	/* Control VQ buffers: protected by the rtnl lock */
+	struct virtio_net_ctrl_hdr ctrl_hdr;
+	virtio_net_ctrl_ack ctrl_status;
+	u8 ctrl_promisc;
+	u8 ctrl_allmulti;
 };
 
 struct padded_vnet_hdr {
@@ -976,31 +982,30 @@ static bool virtnet_send_command(struct virtnet_info *vi, u8 class, u8 cmd,
 				 struct scatterlist *out)
 {
 	struct scatterlist *sgs[4], hdr, stat;
-	struct virtio_net_ctrl_hdr ctrl;
-	virtio_net_ctrl_ack status = ~0;
 	unsigned out_num = 0, tmp;
 
 	/* Caller should know better */
 	BUG_ON(!virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ));
 
-	ctrl.class = class;
-	ctrl.cmd = cmd;
+	vi->ctrl_status = ~0;
+	vi->ctrl_hdr.class = class;
+	vi->ctrl_hdr.cmd = cmd;
 	/* Add header */
-	sg_init_one(&hdr, &ctrl, sizeof(ctrl));
+	sg_init_one(&hdr, &vi->ctrl_hdr, sizeof(vi->ctrl_hdr));
 	sgs[out_num++] = &hdr;
 
 	if (out)
 		sgs[out_num++] = out;
 
 	/* Add return status. */
-	sg_init_one(&stat, &status, sizeof(status));
+	sg_init_one(&stat, &vi->ctrl_status, sizeof(vi->ctrl_status));
 	sgs[out_num] = &stat;
 
 	BUG_ON(out_num + 1 > ARRAY_SIZE(sgs));
 	virtqueue_add_sgs(vi->cvq, sgs, out_num, 1, vi, GFP_ATOMIC);
 
 	if (unlikely(!virtqueue_kick(vi->cvq)))
-		return status == VIRTIO_NET_OK;
+		return vi->ctrl_status == VIRTIO_NET_OK;
 
 	/* Spin for a response, the kick causes an ioport write, trapping
 	 * into the hypervisor, so the request should be handled immediately.
@@ -1009,7 +1014,7 @@ static bool virtnet_send_command(struct virtnet_info *vi, u8 class, u8 cmd,
 	       !virtqueue_is_broken(vi->cvq))
 		cpu_relax();
 
-	return status == VIRTIO_NET_OK;
+	return vi->ctrl_status == VIRTIO_NET_OK;
 }
 
 static int virtnet_set_mac_address(struct net_device *dev, void *p)
@@ -1151,7 +1156,6 @@ static void virtnet_set_rx_mode(struct net_device *dev)
 {
 	struct virtnet_info *vi = netdev_priv(dev);
 	struct scatterlist sg[2];
-	u8 promisc, allmulti;
 	struct virtio_net_ctrl_mac *mac_data;
 	struct netdev_hw_addr *ha;
 	int uc_count;
@@ -1163,22 +1167,22 @@ static void virtnet_set_rx_mode(struct net_device *dev)
 	if (!virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_RX))
 		return;
 
-	promisc = ((dev->flags & IFF_PROMISC) != 0);
-	allmulti = ((dev->flags & IFF_ALLMULTI) != 0);
+	vi->ctrl_promisc = ((dev->flags & IFF_PROMISC) != 0);
+	vi->ctrl_allmulti = ((dev->flags & IFF_ALLMULTI) != 0);
 
-	sg_init_one(sg, &promisc, sizeof(promisc));
+	sg_init_one(sg, &vi->ctrl_promisc, sizeof(vi->ctrl_promisc));
 
 	if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RX,
 				  VIRTIO_NET_CTRL_RX_PROMISC, sg))
 		dev_warn(&dev->dev, "Failed to %sable promisc mode.\n",
-			 promisc ? "en" : "dis");
+			 vi->ctrl_promisc ? "en" : "dis");
 
-	sg_init_one(sg, &allmulti, sizeof(allmulti));
+	sg_init_one(sg, &vi->ctrl_allmulti, sizeof(vi->ctrl_allmulti));
 
 	if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RX,
 				  VIRTIO_NET_CTRL_RX_ALLMULTI, sg))
 		dev_warn(&dev->dev, "Failed to %sable allmulti mode.\n",
-			 allmulti ? "en" : "dis");
+			 vi->ctrl_allmulti ? "en" : "dis");
 
 	uc_count = netdev_uc_count(dev);
 	mc_count = netdev_mc_count(dev);
> -- 
> 2.4.3
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/