> Works on the 17" G4 (5,5). I had to add asm/suspend.h manually but I > suspect that's just due to confusion in my repos. > > The patch clashes with swsusp changes to radeon_pm.c (just remove the > if (!state == PM_SUSP_MEM) return 0; at the start of radeon_pci_suspend > before patching), I've disabled swsusp for the first try. A test with > swsusp enabled is next. > > Thanks a whole bunch - I had even added swsusp support to pmud to work > around missing sleep. But sleep is so much faster :-)
Well, I do intend to work on improving swsusp too. It's useful when you know you'll be sleeping a long time or are running out of battery, and there are all those people with nVidia chips ... Hopefully, I'll get the iBook G4 soon, though I'll need some volunteers for early tests, I'll probably get something to test tomorrow, though it will be really early, and since unlike the M10, I don't have a machine to experiment with directly, it may take a fair amount of ping-pong'ing before I get it right. Also, the resume code is still full of cruft that need to be cleaned up, some of the stuff factored between the various chip types (you may have noticed we also support one of the desktop cards in there, in fact, Paulus got his dual G5 "QuickSilver" to sleep & wakeup with a Radeon_QW card :) Ultimately, we may be able to add support for a lot more desktops provided somebody has the courage to parse the ndrver traces and produce proper init code. Many kudos to Paul, btw, he is the one who used my crude tool for spying the MacOS driver and turned that into working code to re-init the chip. All I did was add my knowledge of the beast & specs to turn that into something a bit more readable & generic, so it has a chance to work on more than his single model, and now trying to get the M9+ right. Ben.
