[mythtv-users] XvMC_VLD on EPIA M10K

Bruce Taber brutab at verizon.net
Sat Feb 23 12:50:59 UTC 2008


Paul Bender wrote:
> Axel Thimm wrote:
>   
>> On Thu, Feb 21, 2008 at 08:00:41PM -0500, Bruce Taber wrote:
>>     
>>> An update on this thread: Evidently, the SVN version 6133 fixes the XVMC 
>>> on the Via Epia 10K. Everything seems to be back in working order using 
>>> this version. SVN 6131 on atrpms-bleeding does not have the capability 
>>> to use XVMC on the unit, however, running with minimyth has restored XVMC.
>>>       
>> Indeed, the configure step changed to
>>
>>  XvMC support              yes
>> -XvMC VLD support          yes
>> -XvMC pro support          yes
>> +XvMC VLD support          no
>> +XvMC pro support          no
>>  XvMC OpenGL sup.          no
>>
>> So you're saying that 16133 will restore it?
>>     
>
> No it does not.
>
> Looking into the problem, it appears that the XvMC-VLD detection logic 
> was changed with changeset 15555 
> <http://svn.mythtv.org/trac/changeset/15555/trunk/mythtv/configure>. 
> Now, the logic will disable xvmc_vld and xvmc_pro when libviaXvMC and 
> libXvMCPro are not on the build system. In my opinion, this is a bug 
> since you they are not needed to build MythTV with XvMC-VLD support.
>
> When this change was introduced, it broke MiniMyth because I use the 
> newer Openchrome drivers in MiniMyth. The newer Openchrome drivers 
> rename libviaXvMC -> libchromeXvMC and libviaXvMCPro -> 
> libchromeXvMCPro. As a result, the configure script did not find them 
> and disabled XvMC-VLD support.
>
> When I discovered this, I patched the configure script to look for 
> libchromeXvMC and libchromeXvMCPro as well. This fixed the problem for 
> MiniMyth since the libraries are present in MiniMyth's build environment.
>
> In addition, I reported the bug and provided my patch My patch was added 
> to SVN.
>
> Unfortunately, while my patch fixes it for MiniMyth, it is not correct. 
> I believe the correct fix is to not disable XvMC-VLD based on the libraries.
>
> I believe that the issue results from the overloading of xvmc_vld. At 
> times the configure script uses xvmc_vld to mean the libviaXvMC library 
> just as xvmc_pro means the libviaXvMCPro driver. However, other times 
> the configure script uses it to mean enable xvmc_vld just as xvmc means 
> enable xvmc. Since XvMC-VLD can be supported through the libXVMCW 
> library, there should be separate variables to indicate the libviaXvMC 
> library and the XvMC-VLD functionality.
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users at mythtv.org
> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
>
>   
Yeah, I was fairly certain there were other factors related to this. 
That was the reason for the original post. I did try the openchrome 
drivers on the myth from atrpms-bleeding versions 15741 & 16131 and 
needed some information from the previous version, which, I no longer 
had. Thanks for the explanation.


More information about the mythtv-users mailing list