[0/2] vdpa: Block vq property change in DRIVER_OK

Message

Dragos Tatulea Dec. 25, 2023, 1:42 p.m. UTC

This series prevents the change of virtqueue address or state when a
device is in DRIVER_OK and not suspended. The virtio spec doesn't
allow changing virtqueue addresses and state in DRIVER_OK, but some devices
do support this operation when the device is suspended. The series
leaves the door open for these devices.

The series was suggested while discussing the addition of resumable
virtuque support in the mlx5_vdpa driver [0].

[0] https://lore.kernel.org/virtualization/[e-mail]/T/#m044ddf540d98a6b025f81bffa058ca647a3d013e

Dragos Tatulea (2):
  vdpa: Track device suspended state
  vdpa: Block vq property changes in DRIVER_OK

 drivers/vhost/vdpa.c | 23 +++++++++++++++++++++--
 1 file changed, 21 insertions(+), 2 deletions(-)

Comments

Michael S. Tsirkin Dec. 25, 2023, 2:40 p.m. UTC | #1

On Mon, Dec 25, 2023 at 03:42:08PM +0200, Dragos Tatulea wrote:
> This series prevents the change of virtqueue address or state when a
> device is in DRIVER_OK and not suspended. The virtio spec doesn't
> allow changing virtqueue addresses and state in DRIVER_OK, but some devices
> do support this operation when the device is suspended. The series
> leaves the door open for these devices.
> 
> The series was suggested while discussing the addition of resumable
> virtuque support in the mlx5_vdpa driver [0].


I am confused. Isn't this also included in
 vdpa/mlx5: Add support for resumable vqs

do you now want this merged separately?

> [0] https://lore.kernel.org/virtualization/[e-mail]/T/#m044ddf540d98a6b025f81bffa058ca647a3d013e
> 
> Dragos Tatulea (2):
>   vdpa: Track device suspended state
>   vdpa: Block vq property changes in DRIVER_OK
> 
>  drivers/vhost/vdpa.c | 23 +++++++++++++++++++++--
>  1 file changed, 21 insertions(+), 2 deletions(-)
> 
> -- 
> 2.43.0

Dragos Tatulea Dec. 25, 2023, 3:09 p.m. UTC | #2

On Mon, 2023-12-25 at 09:40 -0500, Michael S. Tsirkin wrote:
> On Mon, Dec 25, 2023 at 03:42:08PM +0200, Dragos Tatulea wrote:
> > This series prevents the change of virtqueue address or state when a
> > device is in DRIVER_OK and not suspended. The virtio spec doesn't
> > allow changing virtqueue addresses and state in DRIVER_OK, but some devices
> > do support this operation when the device is suspended. The series
> > leaves the door open for these devices.
> > 
> > The series was suggested while discussing the addition of resumable
> > virtuque support in the mlx5_vdpa driver [0].
> 
> 
> I am confused. Isn't this also included in
>  vdpa/mlx5: Add support for resumable vqs
> 
> do you now want this merged separately?
The discussion in the linked thread lead to having 2 series that are
independent: this series and the original v2 of of the "vdpa/mlx5: Add support
for resumable vqs" series (for which I will send a v5 now that is same as v2 +
Acked-by tags).

> 
> > [0] https://lore.kernel.org/virtualization/[e-mail]/T/#m044ddf540d98a6b025f81bffa058ca647a3d013e
> > 
> > Dragos Tatulea (2):
> >   vdpa: Track device suspended state
> >   vdpa: Block vq property changes in DRIVER_OK
> > 
> >  drivers/vhost/vdpa.c | 23 +++++++++++++++++++++--
> >  1 file changed, 21 insertions(+), 2 deletions(-)
> > 
> > -- 
> > 2.43.0
>