On Thu, Aug 06, 2020 at 03:27:38PM +0800, Jason Wang wrote:
On 2020/8/6 下午1:53, Michael S. Tsirkin wrote:So that's a virtio_vdpa limitation.
On Thu, Aug 06, 2020 at 11:23:05AM +0800, Jason Wang wrote:
On 2020/8/5 下午7:40, Michael S. Tsirkin wrote:Hmm I don't really see why. Assume host maps guest memory properly,
On Wed, Aug 05, 2020 at 02:14:07PM +0800, Jason Wang wrote:My understanding is that, IOMMU_PLATFORM is mandatory for hardware vDPA to
On 2020/8/4 上午5:00, Michael S. Tsirkin wrote:Well hardware vendors are I think interested in supporting legacy
Some legacy guests just assume features are 0 after reset.I wonder whether it's easier to just support modern device?
We detect that config space is accessed before features are
set and set features to 0 automatically.
Note: some legacy guests might not even access config space, if this is
reported in the field we might need to catch a kick to handle these.
guests. Limiting vdpa to modern only would make it uncompetitive.
VM does not have an IOMMU, legacy guest can just work.
Yes, guest may not set IOMMU_PLATFORM.
Care explaining what's wrong with this picture?
The problem is virtio_vdpa, without IOMMU_PLATFORM it uses PA which can not
work if IOMMU is enabled.
In the same way, if a device
does not have an on-device iommu *and* is not behind an iommu,
then vdpa can't bind to it.
But this virtio_vdpa specific hack does not belong in a generic vdpa code.
So it can only work for modern device ...