在 2022/8/17 21:57, Guo Zhi 写道:
Follow VIRTIO 1.1 spec, only writing out a single used ring for a batch
of descriptors.
Signed-off-by: Guo Zhi
---
drivers/vhost/vhost.c | 14 --
drivers/vhost/vhost.h | 1 +
2 files changed, 13 insertions(+), 2 deletions(-)
diff --git
在 2022/8/17 21:57, Guo Zhi 写道:
If in order feature negotiated, we can skip the used ring to get
buffer's desc id sequentially.
Signed-off-by: Guo Zhi
---
drivers/virtio/virtio_ring.c | 53 ++--
1 file changed, 45 insertions(+), 8 deletions(-)
diff --git
On Wed, Aug 24, 2022 at 10:48:30PM +0300, Alvaro Karsz wrote:
> > How about calculating the minimum of the limits?
>
> Sounds reasonable.
> Should I add it to this patch (as v2)?
> Or maybe it can be a follow up patch, and it can include the write
> zeros command as well.
If you can include it
在 2022/8/17 21:57, Guo Zhi 写道:
We didn't unmask F_NEXT flag in desc_extra in the end of a chain,
unmask it so that we can access desc_extra to get next information.
Signed-off-by: Guo Zhi
I post a similar patch in the past.
Please share the perf numbers (e.g pps via pktgen).
Thanks
在 2022/8/17 21:57, Guo Zhi 写道:
Only add to used ring when a batch of buffer have all been used. And if
in order feature negotiated, only add the last used descriptor for a
batch of buffer.
Signed-off-by: Guo Zhi
---
drivers/vhost/test.c | 8 +++-
1 file changed, 7 insertions(+), 1
On Fri, Aug 12, 2022 at 03:40:00PM -0700, Rob Clark wrote:
> From: Rob Clark
>
> When VIRTGPU_EXECBUF_RING_IDX is used, we should be considering the
> timeline that the EB if running on rather than the global driver fence
> context.
>
> Fixes: 85c83ea915ed ("drm/virtio: implement context init:
在 2022/8/17 21:57, Guo Zhi 写道:
Vsock uses buffers in order, and for tx driver doesn't have to
know the length of the buffer. So we can do a batch for vsock if
in order negotiated, only write one used ring for a batch of buffers
Signed-off-by: Guo Zhi
---
drivers/vhost/vsock.c | 9 -
Commit e8ae0e140c05 ("vfio: Require that devices support DMA cache
coherence") requires IOMMU drivers to advertise
IOMMU_CAP_CACHE_COHERENCY, in order to be used by VFIO. Since VFIO does
not provide to userspace the ability to maintain coherency through cache
invalidations, it requires hardware
On 2022-08-25 16:46, Jean-Philippe Brucker wrote:
Commit e8ae0e140c05 ("vfio: Require that devices support DMA cache
coherence") requires IOMMU drivers to advertise
IOMMU_CAP_CACHE_COHERENCY, in order to be used by VFIO. Since VFIO does
not provide to userspace the ability to maintain coherency