[Mesa-dev] [PATCH 1/2] radeon/uvd, st/vdpau: fix video format reporting

Grigori Goronzy greg at chown.ath.cx
Mon Oct 7 02:37:38 PDT 2013


On 07.10.2013 11:25, Christian König wrote:
> Am 01.10.2013 21:12, schrieb Ilia Mirkin:
>> On Tue, Oct 1, 2013 at 3:06 PM, Grigori Goronzy <greg at chown.ath.cx>
>> wrote:
>>> UVD can only support NV12 in the case of hardware decoding, but we
>>> can still use all other formats for software decoding. Use the UNKNOWN
>>> entrypoint to signal that we're not interesting in hardware decoding.
>> FYI, in nouveau, we use profile == UNKNOWN to determine this.
>
> Which is probably the better way of doing this.
>

Yup, I'll change this.

> Apart from that you also need to fix the case where the application
> wants to read the surface in a different format than the decoder decoded
> it, e.g. format conversion while reading ti.
>

Yes, but this is a separate issue I think. We fail safely, and it 
doesn't matter for the common path when only PutBits is used.

That said it shouldn't be too hard to convert between YV12 and NV12 or 
YUYV and UYVY. The chroma type is fixed for a given VdpVideoSurface so 
it doesn't make sense to request a 4:2:2 format for a 4:2:0 surface or 
vice versa.

Best regards
Grigori

> Christian.
>
>>> ---
>>>   src/gallium/drivers/radeon/radeon_uvd.c  | 7 +++++--
>>>   src/gallium/state_trackers/vdpau/query.c | 4 ++--
>>>   2 files changed, 7 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/src/gallium/drivers/radeon/radeon_uvd.c
>>> b/src/gallium/drivers/radeon/radeon_uvd.c
>>> index 382b410..161dee8 100644
>>> --- a/src/gallium/drivers/radeon/radeon_uvd.c
>>> +++ b/src/gallium/drivers/radeon/radeon_uvd.c
>>> @@ -1154,6 +1154,9 @@ boolean ruvd_is_format_supported(struct
>>> pipe_screen *screen,
>>>                                   enum pipe_video_profile profile,
>>>                                   enum pipe_video_entrypoint entrypoint)
>>>   {
>>> -       /* we can only handle this one anyway */
>>> -       return format == PIPE_FORMAT_NV12;
>>> +       /* we can only handle this one with UVD */
>>> +       if (entrypoint != PIPE_VIDEO_ENTRYPOINT_UNKNOWN)
>>> +               return format == PIPE_FORMAT_NV12;
>>> +
>>> +       return vl_video_buffer_is_format_supported(screen, format,
>>> profile, entrypoint);
>>>   }
>>> diff --git a/src/gallium/state_trackers/vdpau/query.c
>>> b/src/gallium/state_trackers/vdpau/query.c
>>> index 1d35252..e7486b3 100644
>>> --- a/src/gallium/state_trackers/vdpau/query.c
>>> +++ b/src/gallium/state_trackers/vdpau/query.c
>>> @@ -147,7 +147,7 @@
>>> vlVdpVideoSurfaceQueryGetPutBitsYCbCrCapabilities(VdpDevice device,
>>> VdpChromaTyp
>>>         pscreen,
>>>         FormatYCBCRToPipe(bits_ycbcr_format),
>>>         PIPE_VIDEO_PROFILE_UNKNOWN,
>>> -      PIPE_VIDEO_ENTRYPOINT_BITSTREAM
>>> +      PIPE_VIDEO_ENTRYPOINT_UNKNOWN
>>>      );
>>>      pipe_mutex_unlock(dev->mutex);
>>>
>>> @@ -398,7 +398,7 @@
>>> vlVdpOutputSurfaceQueryPutBitsYCbCrCapabilities(VdpDevice device,
>>> VdpRGBAFormat
>>>      (
>>>         pscreen, ycbcr_format,
>>>         PIPE_VIDEO_PROFILE_UNKNOWN,
>>> -      PIPE_VIDEO_ENTRYPOINT_BITSTREAM
>>> +      PIPE_VIDEO_ENTRYPOINT_UNKNOWN
>>>      );
>>>      pipe_mutex_unlock(dev->mutex);
>>>
>>> --
>>> 1.8.1.2
>>>
>>> _______________________________________________
>>> mesa-dev mailing list
>>> mesa-dev at lists.freedesktop.org
>>> http://lists.freedesktop.org/mailman/listinfo/mesa-dev
>> _______________________________________________
>> mesa-dev mailing list
>> mesa-dev at lists.freedesktop.org
>> http://lists.freedesktop.org/mailman/listinfo/mesa-dev
>



More information about the mesa-dev mailing list