On Tue, Dec 17, 2019 at 05:13:59PM +0100, Dmitry Sepp wrote:
> Hi,
>
> On Dienstag, 17. Dezember 2019 15:09:16 CET Keiichi Watanabe wrote:
> > Hi,
> >
> > Thanks Tomasz and Gerd for the suggestions and information.
> >
> > On Tue, Dec 17, 2019 at 10:39 PM Gerd Hoffmann wrote:
> > > Hi,
> > >
Hi,
On Dienstag, 17. Dezember 2019 15:09:16 CET Keiichi Watanabe wrote:
> Hi,
>
> Thanks Tomasz and Gerd for the suggestions and information.
>
> On Tue, Dec 17, 2019 at 10:39 PM Gerd Hoffmann wrote:
> > Hi,
> >
> > > On the host side, the encode and decode APIs are different as well, so
Hi,
> +\subsection{Feature bits}\label{sec:Device Types / Sound Device / Feature
> bits}
> +
> +None currently defined.
Flags for hostmem & guestmem here?
This could be an information the driver might want to know before
initializing the virtqueues.
> +\item START
> +\item PAUSE
> +\item
Hi,
Thanks Tomasz and Gerd for the suggestions and information.
On Tue, Dec 17, 2019 at 10:39 PM Gerd Hoffmann wrote:
>
> Hi,
>
> > On the host side, the encode and decode APIs are different as well, so
> > having separate implementation decoder and encoder, possibly just
> > sharing some
Hi,
> On the host side, the encode and decode APIs are different as well, so
> having separate implementation decoder and encoder, possibly just
> sharing some helper code, would make much more sense.
When going down that route I'd suggest to use two device ids (even when
specifying both
On Mon, Dec 16, 2019 at 7:32 PM Gerd Hoffmann wrote:
>
> Hi,
>
> > > Hmm, modern GPUs support both encoding and decoding ...
> >
> > Many SoC architectures have completely separate IP blocks for encoding
> > and decoding. Similarly, in GPUs those are usually completely separate
> > parts of the
> > > Of course only virtio drivers would try step (2), other drivers (when
> > > sharing buffers between intel gvt device and virtio-gpu for example)
> > > would go straight to (3).
> >
> > For virtio-gpu as it is today, it's not clear to me that they're
> > equivalent. As I read it, the