kernel 2.6.32 doesn't work for PPC?

[Madjic]

Has anyone else encounterd problems with the new kernel. I have 2 machines running kernel-2.6.31-gentoo-r6, a iMac G3 and a G4 Powermac. Both failed booting the new kernel (using the old config files with minor/no changes).
They just hang in the white OF screen and do nothing.
As they are relatively slow, trial and error is not really an option (takes about 1.5h to compile 1 kernel on both machines), and as there are different kernel settings (workstation with sound, X, etc…/server) i'll have to try twice.

so does anyone know what has to be changed?

[acrux]

Code: Select all

make oldconfig

[aguettinger]

I had some similar problems recently when I compiled XFS into the kernel instead of as modul, yaboot refused to load my kernel, the kernel was to big. Apparently yaboot only supports kernel that don't exceed at certain size (<14.X MB). Thus my first guess would be your kernel has become to big. Did the size of your kernels increase compared to 2.6.31? Did you activate 'Optimize for size'?

[Madjic]

I had some similar problems recently when I compiled XFS into the kernel instead of as modul, yaboot refused to load my kernel, the kernel was to big. Apparently yaboot only supports kernel that don't exceed at certain size (<14.X MB). Thus my first guess would be your kernel has become to big. Did the size of your kernels increase compared to 2.6.31? Did you activate 'Optimize for size'?

Ah, that was indeed the problem - although I experienced the maximum size for the kernel image to <13MB, but it's somewhere around 11.XMB and 12.XMB.
On my other machine the kernel is bigger (somewhat around 15 MB, but that's an G4, so it might depend on Processor type, memory or even something else)

now I have a very unpleasent problem.
hwclock doesn't work. I can neither read the hwclock's status, nor can I reset it with the .32 kernel
result is, the machine stopps booting when trying fsck, because "superblock last mount time is in the future"
my knowledge about C and the kernel is poor, so I have absolutely no idea, how to patch it - or if i made some mistake when disableing options in the Device drivers section of the kernel, or if this has even to do something with Optimize for size.

Ideas anyone?

[JoseJX]

Set your hardware clock in open firmware: http://www.gentoo.org/doc/en/gentoo-ppc ... tupidclock

You should then be able to fsck your disk and continue.

[Madjic]

oh, i fixed it with
hwclock --systohc --utc --noadjfile --directisa

fixing it with openfirmware didn't work by the way