On Fri, 2005-01-14 at 19:51 +0100, Jerome Glisse wrote: > > > > I will try your patch but i bet it works. > > > I am agree too that this should not be in driver specifique but how to > > > handle the r300 case properly on ppc ? > > > > There's no 'r300 case'. This will be the same for all Radeons and > > probably for most if not all other hardware as well. > > What i mean is that r200 already work on ppc and they never needed > to have the endian swapping ?
It probably doesn't work (out of the Mesa tree - in the X tree, it uses the working X register access macros) right now. I've been using this patch but never got around to cleaning it up and unfortunately failed to prod Ian into doing it either when he worked around the issue in the r128 driver. ;) > So if we have to endian swapping for all radeon maybe we could change > radeon_macros.h and use the MESA_BIG_ENDIAN define ? Yes, and/or use optimized code for certain architectures, as my patch does. > By the way i tested your patch and it works properly over here. Cool, thanks for testing. > Still another endian question, do you have any idea where i may look > for properly send texture when i am on ppc. I have looked at r200 > and it doesn't seems that there is any specifics things to do. > But here with r300 we got the component switched :( Textures are transferred via hostdata blits, it's the same problem as for 2D (but in the DRM). -- Earthling Michel DÃnzer | Debian (powerpc), X and DRI developer Libre software enthusiast | http://svcs.affero.net/rm.php?r=daenzer ------------------------------------------------------- The SF.Net email is sponsored by: Beat the post-holiday blues Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek. It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt -- _______________________________________________ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
