UVD fails to init on rv790

[Alex Deucher]

On Wed, Apr 3, 2013 at 4:45 PM, Dieter Nützel <dieter at nuetzel-hh.de> wrote:
> Am 2013-04-03 15:21, schrieb Alex Deucher:
>
>> On Wed, Apr 3, 2013 at 7:29 AM, Christian König <deathsimple at vodafone.de>
>> wrote:
>>>
>>> Hi Andy,
>>>
>>> crap! I feared that something like this would happen. IIRC we never
>>> tested
>>> UVD on an rv790, and this hardware isn't easy to get any more.
>>>
>>> RV770/RV790 have a separate UVD hardware generation (that's why they have
>>> their own firmware) and there possible is some bug or something like this
>>> that we haven't implemented. You couldn't give me SSH access to that
>>> system?
>>
>>
>> I tested on an RV770 earlier in development and it was working ok at
>> the time I think.  I'll give it another try today.
>>
>> Alex
>
>
>
> Hello Alex,
>
> time goes by so fast --- wife, home, 2 children...
>
> but what do I need:
> I'll test it on my 2 HD4650/RV730 AGP (!) low speed Duron and dual Athlon
> MP.
>
> git:// drm-fixes kernel
> git:// mesa
> patches
>
> R700_rlc.bin
> RV710_uvd.bin
>
> BTW Any hints for libvdpau_r600.so.1.0.0?
> I get wrong colors (red is like blue, blue is like yellow, etc.)
>

What are you testing with?  The adobe flash vdpau implementation
reverses the channels.  IIRC, newer versions of libvdpau have a
workaround for it.  I've generally been testing with mplayer.  mplayer
-vo vdpau -vc ffh264vdpau <file>

Alex