Patchwork [V2,2/6] vhost_net: use vhost_add_used_and_signal_n() in vhost_zerocopy_signal_used()

login
register
mail settings
Submitter Jason Wang
Date Aug. 30, 2013, 4:29 a.m.
Message ID <1377836962-49780-3-git-send-email-jasowang@redhat.com>
Download mbox | patch
Permalink /patch/271104/
State Superseded
Delegated to: David Miller
Headers show

Comments

Jason Wang - Aug. 30, 2013, 4:29 a.m.
We tend to batch the used adding and signaling in vhost_zerocopy_callback()
which may result more than 100 used buffers to be updated in
vhost_zerocopy_signal_used() in some cases. So wwitch to use
vhost_add_used_and_signal_n() to avoid multiple calls to
vhost_add_used_and_signal(). Which means much more less times of used index
updating and memory barriers.

Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 drivers/vhost/net.c |   13 ++++++++-----
 1 files changed, 8 insertions(+), 5 deletions(-)
Michael S. Tsirkin - Sept. 2, 2013, 5:50 a.m.
On Fri, Aug 30, 2013 at 12:29:18PM +0800, Jason Wang wrote:
> We tend to batch the used adding and signaling in vhost_zerocopy_callback()
> which may result more than 100 used buffers to be updated in
> vhost_zerocopy_signal_used() in some cases. So wwitch to use

switch

> vhost_add_used_and_signal_n() to avoid multiple calls to
> vhost_add_used_and_signal(). Which means much more less times of used index
> updating and memory barriers.

pls put info on perf gain in commit log too

> 
> Signed-off-by: Jason Wang <jasowang@redhat.com>
> ---
>  drivers/vhost/net.c |   13 ++++++++-----
>  1 files changed, 8 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c
> index 280ee66..8a6dd0d 100644
> --- a/drivers/vhost/net.c
> +++ b/drivers/vhost/net.c
> @@ -281,7 +281,7 @@ static void vhost_zerocopy_signal_used(struct vhost_net *net,
>  {
>  	struct vhost_net_virtqueue *nvq =
>  		container_of(vq, struct vhost_net_virtqueue, vq);
> -	int i;
> +	int i, add;
>  	int j = 0;
>  
>  	for (i = nvq->done_idx; i != nvq->upend_idx; i = (i + 1) % UIO_MAXIOV) {
> @@ -289,14 +289,17 @@ static void vhost_zerocopy_signal_used(struct vhost_net *net,
>  			vhost_net_tx_err(net);
>  		if (VHOST_DMA_IS_DONE(vq->heads[i].len)) {
>  			vq->heads[i].len = VHOST_DMA_CLEAR_LEN;
> -			vhost_add_used_and_signal(vq->dev, vq,
> -						  vq->heads[i].id, 0);
>  			++j;
>  		} else
>  			break;
>  	}
> -	if (j)
> -		nvq->done_idx = i;
> +	while (j) {
> +		add = min(UIO_MAXIOV - nvq->done_idx, j);
> +		vhost_add_used_and_signal_n(vq->dev, vq,
> +					    &vq->heads[nvq->done_idx], add);
> +		nvq->done_idx = (nvq->done_idx + add) % UIO_MAXIOV;
> +		j -= add;
> +	}
>  }
>  
>  static void vhost_zerocopy_callback(struct ubuf_info *ubuf, bool success)
> -- 
> 1.7.1
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Jason Wang - Sept. 2, 2013, 6:28 a.m.
On 09/02/2013 01:50 PM, Michael S. Tsirkin wrote:
> On Fri, Aug 30, 2013 at 12:29:18PM +0800, Jason Wang wrote:
>> > We tend to batch the used adding and signaling in vhost_zerocopy_callback()
>> > which may result more than 100 used buffers to be updated in
>> > vhost_zerocopy_signal_used() in some cases. So wwitch to use
> switch

Ok.
>> > vhost_add_used_and_signal_n() to avoid multiple calls to
>> > vhost_add_used_and_signal(). Which means much more less times of used index
>> > updating and memory barriers.
> pls put info on perf gain in commit log too
>

Sure.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c
index 280ee66..8a6dd0d 100644
--- a/drivers/vhost/net.c
+++ b/drivers/vhost/net.c
@@ -281,7 +281,7 @@  static void vhost_zerocopy_signal_used(struct vhost_net *net,
 {
 	struct vhost_net_virtqueue *nvq =
 		container_of(vq, struct vhost_net_virtqueue, vq);
-	int i;
+	int i, add;
 	int j = 0;
 
 	for (i = nvq->done_idx; i != nvq->upend_idx; i = (i + 1) % UIO_MAXIOV) {
@@ -289,14 +289,17 @@  static void vhost_zerocopy_signal_used(struct vhost_net *net,
 			vhost_net_tx_err(net);
 		if (VHOST_DMA_IS_DONE(vq->heads[i].len)) {
 			vq->heads[i].len = VHOST_DMA_CLEAR_LEN;
-			vhost_add_used_and_signal(vq->dev, vq,
-						  vq->heads[i].id, 0);
 			++j;
 		} else
 			break;
 	}
-	if (j)
-		nvq->done_idx = i;
+	while (j) {
+		add = min(UIO_MAXIOV - nvq->done_idx, j);
+		vhost_add_used_and_signal_n(vq->dev, vq,
+					    &vq->heads[nvq->done_idx], add);
+		nvq->done_idx = (nvq->done_idx + add) % UIO_MAXIOV;
+		j -= add;
+	}
 }
 
 static void vhost_zerocopy_callback(struct ubuf_info *ubuf, bool success)