On Fri, 2005-01-14 at 20:04 -0500, Vladimir Dergachev wrote: > > One idea/suggestion that I had for a long while is to not correct for > endianness in software but rather use a feature of Radeon cards where > several apertures are available. > > For example, there are two copies of register aperture one with low-endian > translation and one with big-endian. > > One could map the appropriate one for plain reads,
I don't see the point as the little endian load/store instructions don't involve any overhead. > and use properly modified register addresses for HOSTDATA_BLT - though > I am not 100% sure about how this will work. I doubt it would help for the problems we're seeing with the CP at all. > It might be that ATI got rid of all the friendly endianness swapping bits > in favour of this approach. I'm not aware of anything like that, but the big endian register aperture may indeed go away in future chips. -- 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 [email protected] https://lists.sourceforge.net/lists/listinfo/dri-devel
