Re: [PATCH] virtio_balloon: check virtqueue_add_outbuf() return value

Re: [PATCH] virtio_balloon: check virtqueue_add_outbuf() return value

[Michael S. Tsirkin]

On Thu, May 26, 2022 at 09:34:01PM -0400, Bo Liu wrote:
> virtqueue_add_outbuf() can fail, when it fails, there is no need
> to call the funciont virtqueue_kick().
> Signed-off-by: Bo Liu <liubo03@inspur.com>

Sorry don't get it.
What's the point of this patch really?

> ---
>  drivers/virtio/virtio_balloon.c | 18 +++++++++++-------
>  1 file changed, 11 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/virtio/virtio_balloon.c b/drivers/virtio/virtio_balloon.c
> index b9737da6c4dd..0d7da4d95e1e 100644
> --- a/drivers/virtio/virtio_balloon.c
> +++ b/drivers/virtio/virtio_balloon.c
> @@ -153,16 +153,18 @@ static void tell_host(struct virtio_balloon *vb, struct virtqueue *vq)
>  {
>  	struct scatterlist sg;
>  	unsigned int len;
> +	int err;
>  
>  	sg_init_one(&sg, vb->pfns, sizeof(vb->pfns[0]) * vb->num_pfns);
>  
>  	/* We should always be able to add one buffer to an empty queue. */
> -	virtqueue_add_outbuf(vq, &sg, 1, vb, GFP_KERNEL);
> -	virtqueue_kick(vq);
> -
> -	/* When host has read buffer, this completes via balloon_ack */
> -	wait_event(vb->acked, virtqueue_get_buf(vq, &len));
> +	err = virtqueue_add_outbuf(vq, &sg, 1, vb, GFP_KERNEL);
> +	if (!err) {
> +		virtqueue_kick(vq);
>  
> +		/* When host has read buffer, this completes via balloon_ack */
> +		wait_event(vb->acked, virtqueue_get_buf(vq, &len));
> +	}
>  }
>  
>  static int virtballoon_free_page_report(struct page_reporting_dev_info *pr_dev_info,
> @@ -382,6 +384,7 @@ static void stats_handle_request(struct virtio_balloon *vb)
>  	struct virtqueue *vq;
>  	struct scatterlist sg;
>  	unsigned int len, num_stats;
> +	int err;
>  
>  	num_stats = update_balloon_stats(vb);
>  
> @@ -389,8 +392,9 @@ static void stats_handle_request(struct virtio_balloon *vb)
>  	if (!virtqueue_get_buf(vq, &len))
>  		return;
>  	sg_init_one(&sg, vb->stats, sizeof(vb->stats[0]) * num_stats);
> -	virtqueue_add_outbuf(vq, &sg, 1, vb, GFP_KERNEL);
> -	virtqueue_kick(vq);
> +	err = virtqueue_add_outbuf(vq, &sg, 1, vb, GFP_KERNEL);
> +	if (!err)
> +		virtqueue_kick(vq);
>  }
>  
>  static inline s64 towards_target(struct virtio_balloon *vb)
> -- 
> 2.27.0

_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH] virtio_balloon: check virtqueue_add_outbuf() return value

[Michael S. Tsirkin]

On Fri, May 27, 2022 at 10:43:30AM +0000, Bo Liu (刘波)-浪潮信息 wrote:
> The purpose of this patch is, virtqueue_add_outbuf() adding data to
> virtqueue may succeed or fail. If successful, it will call virtqueue_kick()
> to notifie the host backend driver, but if the virtqueue_add_outbuf() fails,
> the data is not successfully added to the virtqueue, and the host back-end
> driver does not need to be notified for processing.
> 
> Thanks.

Why is it a problem if we do?

-- 
MST

_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH] virtio_balloon: check virtqueue_add_outbuf() return value

[David Hildenbrand]

On 30.05.22 04:16, Bo Liu (刘波)-浪潮信息 wrote:
> Adding this patch can avoid unnecessary VM exits and reduce the number of VM exits
> 

... in corner cases where virtqueue_add_outbuf() fails? Why do we care
about that corner case?

Looks like unnecessary code churn to me, unless I am missing something
important.

-- 
Thanks,

David / dhildenb

_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization