Benjamin Herrenschmidt wrote:
first thank you for this nice work. Mostly it works great for me on my
iBook G4 933 (first generation).
I have still some problems with #4.
When i unplug or plug in an USB-Mouse while in sleep the iBook crashes
on resume. The sleep-led turns off, but the display remains dark and i
can't ping the iBook (with #2 one time after resume the display remained
blank too, but the sound (xmms and dmasound_pmac) started again and i
was able to ping the iBook over the net).
An other problem is that the kernel hangs sometimes on boot. With patch
#2 and agpgart compiled into the kernel, it locked up every time after
the agpgart messages. Without agpgart or agpgart compiled as module, i
can boot without problems most times. Loading agpgart as module (hotplug
loads it) make no problems. Btw. what is agpgart needed for? I have not
tested #4 with agpgart compiled into kernel.
Some times (with patch #2 more often as with #4) the kernel hangs
somewhere after the ati messages (i will write down the exact message
down when i see it the next time). I had this two times with vanilla
2.6.9 too, but with the suspend patch it happens quiet more frequent.
One thing i noticed in this context (with vanilla 2.6.9 and 2.6.8.1 and
with the sleep patches too) is an messages "radeonfb (...): Invalid ROM
signature 0 should be0xaa55". (The space between the "be" and "0xaa55"
is really missing!)
An other point where it hangs sometimes is just after the kernel message
"Macintosh non-volatile memory driver v1.1". One time i was only able to
start Linux again after booting MacOS.
If you need some more information, i would be happy to give them to you.
Best Regards,
Sascha
