Hi
On Tue, May 14, 2019 at 7:38 AM Gerd Hoffmann <kra...@redhat.com> wrote:
>
> Hi,
>
> > +VhostUserGpuCursorUpdate
> > +^^^^^^^^^^^^^^^^^^^^^^^^
> > +
> > ++-----+-------+-------+--------+
> > +| pos | hot_x | hot_y | cursor |
> > ++-----+-------+-------+--------+
> > +
> > +:pos: a ``VhostUserGpuCursorPos``, the cursor location
> > +
> > +:hot_x/hot_y: ``u32``, the cursor hot location
> > +
> > +:cursor: ``[u32; 64 * 64]``, 64x64 RGBA cursor data
>
> Should clarify here what exactly RGBA is. (PIXMAN_a8r8g8b8 I guess).
ok
>
> > +VhostUserGpuUpdate
> > +^^^^^^^^^^^^^^^^^^
> > +
> > ++------------+---+---+---+---+------+
> > +| scanout-id | x | y | w | h | data |
> > ++------------+---+---+---+---+------+
> > +
> > +:scanout-id: ``u32``, the scanout content to update
> > +
> > +:x/y/w/h: ``u32``, region of the update
> > +
> > +:data: RGBA data (the size is computed based on the region size, and
> > + the request type)
>
> Likewise. Also: alpha channel for the framebuffer?
It is actually PIXMAN_x8r8g8b8, fixed
>
> > +C structure
> > +-----------
> > +
> > +In QEMU the vhost-user-gpu message is implemented with the following
> > struct:
> > +
> > +.. code:: c
> > +
> > + typedef struct VhostUserGpuMsg {
> > + uint32_t request; /* VhostUserGpuRequest */
> > + uint32_t flags;
> > + uint32_t size; /* the following payload size */
>
> uint32_t padding;
>
> > + union {
> > + VhostUserGpuCursorPos cursor_pos;
> > + VhostUserGpuCursorUpdate cursor_update;
> > + VhostUserGpuScanout scanout;
> > + VhostUserGpuUpdate update;
> > + VhostUserGpuDMABUFScanout dmabuf_scanout;
> > + struct virtio_gpu_resp_display_info display_info;
> > + uint64_t u64;
>
> ... so this 64bit value will be aligned.
vhost-user didn't bother. Should we?
--
Marc-André Lureau
