>> Someone at this forum said that this parser can be avoided via a (less >> powerfull) MMIO method, thus providing 3D support for ATI cards in the >> linux-libre kernel, which would be great. I'll ask this person about it. > > Hmm, interesting, I hadn't realized that. This would indeed be awesome. >
I've got a response from John Bridgman, AMD's responsible for GNU/Linux drivers, and it seems that whithout the microcode we can only achieve 2D (and partial 3D) acceleration, but just for old cards. New cards need the firmware for both 2D and 3D routines, and even Xv extensions. I'm not sure if it's worth maintaining a deblobed driver, as those cards are usable without DRM anyway (with no acceleration). An additional response from Alex Deucher (xorg.org developer) confirms the technical difficulties. Mr Bridgman also ask's me why embedded firmware is ok for us while the same binary in a file is not. I think a correct answer is this: If the firmware is invisibly burned inside the hardware in a way I cannot access to it, then I don't need a software license to use my video card. If I need to get and use a file from a software provider (the card vendor), then it's fine as long as its license respects my freedom. As my response can be taken as GNU's one, I prefer to ask you before replying to him. I think it would be nice to convince them about this issue, is a tiny piece of code that will render the driver non free, thus forcing us to give AMD not credit for their freedom efforts. _______________________________________________ linux-libre mailing list [email protected] http://www.fsfla.org/cgi-bin/mailman/listinfo/linux-libre
