Re: [PATCH vhost v4 03/15] vdpa: Add VHOST_BACKEND_F_CHANGEABLE_VQ_STATE_IN_SUSPEND flag

[Eugenio Perez Martin]

On Tue, Dec 19, 2023 at 7:09 PM Dragos Tatulea <dtatulea@xxxxxxxxxx> wrote:
>
> The virtio spec doesn't allow changing virtqueue state after
> DRIVER_OK. Some devices do support this operation when the device is
> suspended. The VHOST_BACKEND_F_CHANGEABLE_VQ_STATE_IN_SUSPEND flag
> advertises this support as a backend features.
>
> Signed-off-by: Dragos Tatulea <dtatulea@xxxxxxxxxx>
> Suggested-by: Eugenio Pérez <eperezma@xxxxxxxxxx>
> ---
>  include/uapi/linux/vhost_types.h | 4 ++++
>  1 file changed, 4 insertions(+)
>
> diff --git a/include/uapi/linux/vhost_types.h b/include/uapi/linux/vhost_types.h
> index aacd067afc89..848dadc95ca1 100644
> --- a/include/uapi/linux/vhost_types.h
> +++ b/include/uapi/linux/vhost_types.h
> @@ -196,5 +196,9 @@ struct vhost_vdpa_iova_range {
>   * and is in state DRIVER_OK.
>   */
>  #define VHOST_BACKEND_F_CHANGEABLE_VQ_ADDR_IN_SUSPEND  0x9
> +/* Device supports changing virtqueue state when device is suspended
> + * and is in state DRIVER_OK.
> + */
> +#define VHOST_BACKEND_F_CHANGEABLE_VQ_STATE_IN_SUSPEND  0x10

Acked-by: Eugenio Pérez <eperezma@xxxxxxxxxx>

It would be great to find a shorter ID but I'm very bad at naming :(.
Maybe it is ok to drop the _BACKEND_ word? I'm ok with carrying the
acked-by if so.

Thanks!

>
>  #endif
> --
> 2.43.0
>