|
hello everyone,
I have compiled a new kernel for an 486 which was
going to be a filtering firewall. Not for a while I'm afraid, because it won't
boot up anymore. (btw: I started with a clean freshly installed system with
debian 2.2.18pre21)
Lilo works fine; it boots the kernel;
after the kernel has started booting, there is a
step where the ext2 fs root is mounted read-only (as usual);
the next step is my problem: "kernel-panic, no init
found. try to pass init= to the kernel." and there it halts.
No problem I thought, reboot, hold shift at startup
and inform lilo to pass init=/sbin/init to the kernel (which is a stupid thing
to do since that is the default, this part ought to go by itself!) right: that
did'nt work.
Euhm... corrupt init? (all of a sudden??); used
Loadlin with a kernel-image from the debian-1 CD and mounted my freshly
installed system as root (/dev/hda2) (another funny thing: my swap is /dev/hda1,
I seem to have made it a primary partition; is that a bad thing
anyone?)
an annoying modprobe-loop as a result but I can
enter the system (in single mode only) and find a perfectly normal init, and a
perfectly normal /bin/sh and /etc/init (which I also have tried)
Here's my big question: when I boot from a CD
Linux-image, does my system still use the init in /dev/hda2/sbin/ ? (I think so
since there's no init on the CD)
Why then can my new kernel not use this init but
more important: why can't my old kernel too! (the one after first installation
residing on my HDD which is now named vmlinuz.old)
Would be nice if anyone can help. baking a kernel
on this machine takes about half a day...
greetz, Mythiq
|