Linux Virtualization - Nov 2019 - [PATCH v5] vhost: introduce mdev based hardware backend

On Wed, Nov 06, 2019 at 07:59:02AM -0500, Michael S. Tsirkin
wrote:
> On Tue, Nov 05, 2019 at 07:53:32PM +0800, Tiwei Bie wrote:
> > This patch introduces a mdev based hardware vhost backend.
> > This backend is built on top of the same abstraction used
> > in virtio-mdev and provides a generic vhost interface for
> > userspace to accelerate the virtio devices in guest.
> > 
> > This backend is implemented as a mdev device driver on top
> > of the same mdev device ops used in virtio-mdev but using
> > a different mdev class id, and it will register the device
> > as a VFIO device for userspace to use. Userspace can setup
> > the IOMMU with the existing VFIO container/group APIs and
> > then get the device fd with the device name. After getting
> > the device fd, userspace can use vhost ioctls on top of it
> > to setup the backend.
> > 
> > Signed-off-by: Tiwei Bie <tiwei.bie at intel.com>
> 
> So at this point, looks like the only thing missing is IFC, and then all
> these patches can go in.
> But as IFC is still being worked on anyway, it makes sense to
> address the minor comments manwhile so we don't need
> patches on top.
> Right?
Yeah, of course.

Thanks,
Tiwei

On 2019/11/6 ??10:39, Tiwei Bie wrote:
> On Wed, Nov 06, 2019 at 07:59:02AM -0500, Michael S. Tsirkin wrote:
>> On Tue, Nov 05, 2019 at 07:53:32PM +0800, Tiwei Bie wrote:
>>> This patch introduces a mdev based hardware vhost backend.
>>> This backend is built on top of the same abstraction used
>>> in virtio-mdev and provides a generic vhost interface for
>>> userspace to accelerate the virtio devices in guest.
>>>
>>> This backend is implemented as a mdev device driver on top
>>> of the same mdev device ops used in virtio-mdev but using
>>> a different mdev class id, and it will register the device
>>> as a VFIO device for userspace to use. Userspace can setup
>>> the IOMMU with the existing VFIO container/group APIs and
>>> then get the device fd with the device name. After getting
>>> the device fd, userspace can use vhost ioctls on top of it
>>> to setup the backend.
>>>
>>> Signed-off-by: Tiwei Bie <tiwei.bie at intel.com>
>> So at this point, looks like the only thing missing is IFC, and then
all
>> these patches can go in.
>> But as IFC is still being worked on anyway, it makes sense to
>> address the minor comments manwhile so we don't need
>> patches on top.
>> Right?
> Yeah, of course.
>
> Thanks,
> Tiwei

Please send V6 and I will ack there.

Thanks

On Thu, Nov 07, 2019 at 12:16:01PM +0800, Jason Wang
wrote:
> On 2019/11/6 ??10:39, Tiwei Bie wrote:
> > On Wed, Nov 06, 2019 at 07:59:02AM -0500, Michael S. Tsirkin wrote:
> > > On Tue, Nov 05, 2019 at 07:53:32PM +0800, Tiwei Bie wrote:
> > > > This patch introduces a mdev based hardware vhost backend.
> > > > This backend is built on top of the same abstraction used
> > > > in virtio-mdev and provides a generic vhost interface for
> > > > userspace to accelerate the virtio devices in guest.
> > > > 
> > > > This backend is implemented as a mdev device driver on top
> > > > of the same mdev device ops used in virtio-mdev but using
> > > > a different mdev class id, and it will register the device
> > > > as a VFIO device for userspace to use. Userspace can setup
> > > > the IOMMU with the existing VFIO container/group APIs and
> > > > then get the device fd with the device name. After getting
> > > > the device fd, userspace can use vhost ioctls on top of it
> > > > to setup the backend.
> > > > 
> > > > Signed-off-by: Tiwei Bie <tiwei.bie at intel.com>
> > > So at this point, looks like the only thing missing is IFC, and
then all
> > > these patches can go in.
> > > But as IFC is still being worked on anyway, it makes sense to
> > > address the minor comments manwhile so we don't need
> > > patches on top.
> > > Right?
> > Yeah, of course.
> > 
> > Thanks,
> > Tiwei
> 
> 
> Please send V6 and I will ack there.
Got it, I will send it soon.

Thanks!
Tiwei