2.4.0-test4 doesn't boot

Hi,

booting rsync/bitkeeper kernel 2.4.0-test4 gets me into an infinite
loop: At the end of the kernel boot proces, right after "freeing unused
kernel memory", I get a couple of

kmod: runaway modprobe loop assumed and stopped

and then a never ending loop of

kmod: failed to exec /sbin/modprobe -s -k binfmt-4c46, errno=8

Has anyone else seen this or knows what it means?

I have spent many hours changing all kinds of config options, modutils
version, kernel arguments, but I am giving up now. 2.4.0-test3 from 2
weeks ago worked OK.

The number 4c46 exists as PCI_DEVICE_ID_ATI_RAGE128_LF in
include/linux/pci_ids.h, but whether I enable aty128 support or not
makes no difference (I don't have an ATI card).

--
Martin

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Martin Costabel wrote:
[]

and then a never ending loop of

kmod: failed to exec /sbin/modprobe -s -k binfmt-4c46, errno=8

Sorry to answer my own question, but I don't want to leave loose ends in
the list archives.

Took me a long time to hunt this down to a recent patch in
include/asm-ppc/elf.h which wasn't delivered to me by rsync.

The result was that the kernel didn't understand the ELF bin-format of
/sbin/init, then asked modprobe to load a module for the unknown format,
didn't understand the bin-format of /sbin/modprobe, etc ad infinitum.

The reason why this file wasn't sent to me by rsync was related to the
recent clock changes, which made me run sometimes with a wrong system
time. So what I learned from this story (apart from a lot about the
inner workings of kmod, execve, binfmt_elf, and the boot process) is:

Never use "rsync -au" while your system time is not correct!

--
Martin

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/