- Original Message -
> From: "jasowang"
> To: "Guo Zhi"
> Cc: "eperezma" , "sgarzare" ,
> "Michael Tsirkin" ,
> "qemu-devel@nongnu.org Developers"
> Sent: Thursday, August 25, 2022 2:06:11 PM
> Subje
l@nongnu.org Developers"
> > Sent: Monday, August 22, 2022 10:08:32 PM
> > Subject: Re: [RFC 1/2] virtio: expose used buffers
>
> > On Thu, Aug 18, 2022 at 5:13 PM Guo Zhi wrote:
> >>
> >> Follow VIRTIO 1.1 spec, we can only writing out a single used ring
- Original Message -
> From: "eperezma"
> To: "Guo Zhi"
> Cc: "jasowang" , "sgarzare" ,
> "Michael Tsirkin" ,
> "qemu-devel@nongnu.org Developers"
> Sent: Monday, August 22, 2022 10:08:32 PM
> Subject:
On Thu, Aug 18, 2022 at 11:13 PM Guo Zhi wrote:
>
> Follow VIRTIO 1.1 spec, we can only writing out a single used ring for a
> batch of descriptors, and only notify guest when the batch of
> descriptors have all been used.
Yes, but I don't see anything that is related to the "exposing used
buffer
On Thu, Aug 18, 2022 at 5:13 PM Guo Zhi wrote:
>
> Follow VIRTIO 1.1 spec, we can only writing out a single used ring for a
> batch of descriptors, and only notify guest when the batch of
> descriptors have all been used.
>
> We do that batch for tx, because the driver doesn't need to know the
> l
Follow VIRTIO 1.1 spec, we can only writing out a single used ring for a
batch of descriptors, and only notify guest when the batch of
descriptors have all been used.
We do that batch for tx, because the driver doesn't need to know the
length of tx buffer, while for rx, we don't apply the batch st
