On Wed, 2004-10-27 at 08:47 +0200, Colin Leroy wrote: > On 27 Oct 2004 at 15h10, Benjamin Herrenschmidt wrote: > > Hi Ben, > > > This is an experimental patch against 2.6.9. It concerns the ATI based > > Aluminium PowerBook. The nVidia based ones aren't concerned at all, and > > the iBook G4 will have to wait a bit more for me to get the video wakeup > > code right. > > Aww, can't wait :) > > Btw, looks like the patch you sent contains changes that have been > integrated post-2.6.9 (2.6.10-rc1) ? I tried 2.6.10-rc1 on my iBook > G4 yesterday, and had video-related problems: when darkening the > screen (using the Fn keys or by leaving the laptop alone enough time), > the screen displays garbage at level 0 (instead of going black as > previously), the sort of thing that happened to me when I messed with > mach64 registers when working on mplayer. You know, screen goes > progressively white, with beautiful color patterns expanding slowly > all over the screen. (can't find an image on google, sorry). > When re-enabling the screen, it's ok. > > I also got lots of gcc internal errors when compiling linux using 2.6.10-rc1, > but that must be unrelated :)
Well, I'd suggest not running that kernel until paul's signal fix gets in, there is a FP register corruption in there... There are other issues with 2.6.10 currently, so it's not really a good kernel to play with for now. I'm surprised by your backlight problems. I've simplified the backlight & blanking code a bit, it should be better now instead of worse ... Can you tweak around maybe and tell me if you find a precise register bit that I'm not toggling properly ? (Use my 2.6.9 sleep patch to get the up-to-date radeonfb). Ben.
