Luc Verhaegen schreef:

While this will allow most R5xx users to already move away from shadowFB, it is still not as fast as it could be :) For instance, EXA doesn't have accelerated (DMAed) up or download yet, but i hope to make this happen soon.

I would like some PPC users to give both XAA and EXA a thorough try. I expect at least exa to be broken, as the prepare/finishAccess hooks probably need to have their surface offset.

So please give it a spin, and report problems asap :)

Does this require kernel support? I'm using FreeBSD 7.0-RC1, with this in the kernel: device pci device vga options VESA device agp device drm The radeondrm driver only seems to support < r400 cards. Rene -- GPG fingerprint = E738 5471 D185 7013 0EE0 4FC8 3C1D 6F83 12E1 84F6 (subkeys.pgp.net) "It won't fit on the line." -- me, 2001 -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org

Hi all, first of all thanks for the great work on the driver. Without it I would be hardly able to use my notebook for my daily work. System: IBM/Lenovo Thinkpad T60p Mobility FireGL V5200 1600x1200 LCD Panel Ubuntu 8.04/hardy X.Org X Server 1.4.0.90 libxaa 1.2.0 libexa 2.2.0 Hardy radeonhd package: 1.1.0-1 Luc Verhaegen wrote:

While this will allow most R5xx users to already move away from shadowFB, it is still not as fast as it could be :) For instance, EXA doesn't have accelerated (DMAed) up or download yet, but i hope to make this happen soon.

Great news! I have immediately tried to run it and xaa as well as exa work without crashes. Performance-wise I have to say that EXA is _really_ slow, expecially with firefox/thunderbird. XAA is almost as fast as ShadowFB but scrolling in Thunderbird is sluggish. Video playback looked ok though. I rolled back to the hardy driver for now to be able to use ShadowFB; it seems to be impossible to activate after the update. Would it be possible to benchmark the new features for comparison to ShadowFB ? Cheers, Jörg http://wwwhep.physik.uni-freiburg.de/~mechnich/Xorg.0.log.exa http://wwwhep.physik.uni-freiburg.de/~mechnich/Xorg.0.log.sfb http://wwwhep.physik.uni-freiburg.de/~mechnich/Xorg.0.log.xaa -- To unsubscribe, e-mail: radeonhd+unsubscribe@opensuse.org For additional commands, e-mail: radeonhd+help@opensuse.org

Hi, just to report XAA and EXA for my card 0x7280:0x148C:0x2223: PowerColor X1950 Pro AGP 512MB XAA and EXA work, no crashes or anything bad that is immediately noticeable. EXA is dead slow, and I mean very very very slow, perhaps because of that DMA missing thing, it is unusable for me, anyway other than that it runs ok. XAA is almost identical to shadowfb however I noticed some slow scrolling in Thunderbird 2.0.0.6 on the messages list (big big list), all other areas appear to scroll fine, nice and smooth. Firefox didn't noticed much anything. On other applications / windows there were some slow downs here and there but still very usable. I'm using shadowfb for now (faster than current alternatives), but that may change on the next few commits. I attached Xorg.logs for EXA and XAA just to give you some feedback. If a more detailed logs or config files and versions are needed please say so. Best regards and great work, thanks :) David Morgado Luc Verhaegen escreveu:

While this will allow most R5xx users to already move away from shadowFB, it is still not as fast as it could be :) For instance, EXA doesn't have accelerated (DMAed) up or download yet, but i hope to make this happen soon.

I would like some PPC users to give both XAA and EXA a thorough try. I expect at least exa to be broken, as the prepare/finishAccess hooks probably need to have their surface offset.

So please give it a spin, and report problems asap :)

Thanks,

Luc Verhaegen. SUSE/Novell X Driver Developer.