Re: [PATCH virtio v2 10/13] virtio_pci_modern: create admin queue of queried size

["Michael S. Tsirkin" <mst@redhat.com>]

On Wed, Jul 10, 2024 at 08:35:58AM +0200, Jiri Pirko wrote:
> From: Jiri Pirko <jiri@nvidia.com>
> 
> Don't limit the admin queue size to VIRTIO_AVQ_SGS_MAX and rather rely
> on the queried queue size.
> 
> Signed-off-by: Jiri Pirko <jiri@nvidia.com>

I have some doubts about this one. Can't we limit the size to something
sensible? E.g. max number of CPUs? Number of VFs? I don't see why we
should just follow what device did blindly, the device has no idea about
use so would tend to over-provision.

> ---
>  drivers/virtio/virtio_pci_modern.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/drivers/virtio/virtio_pci_modern.c b/drivers/virtio/virtio_pci_modern.c
> index 5ceb4b2c18df..a649c9dc436d 100644
> --- a/drivers/virtio/virtio_pci_modern.c
> +++ b/drivers/virtio/virtio_pci_modern.c
> @@ -550,8 +550,7 @@ static struct virtqueue *setup_vq(struct virtio_pci_device *vp_dev,
>  	if (index >= vp_modern_get_num_queues(mdev) && !is_avq)
>  		return ERR_PTR(-EINVAL);
>  
> -	num = is_avq ?
> -		VIRTIO_AVQ_SGS_MAX : vp_modern_get_queue_size(mdev, index);
> +	num = vp_modern_get_queue_size(mdev, index);
>  	/* Check if queue is either not available or already active. */
>  	if (!num || vp_modern_get_queue_enable(mdev, index))
>  		return ERR_PTR(-ENOENT);
> -- 
> 2.45.2