Newly compiled kernel behaves as if it was the old one?
I've just installed debian 3.0, which somehow got me kernel 2.2.20. Of course I didn't like this, to I got the sources & headers for 2.4.18 and compiled this kernel (with USB fs support and stuff). Now when I reboot the computer, everything _seems_ to be going fine, except that somehow the new compiled stuff doesn't work, for example mounting the usbfs tells me that this fs is not supported in the kernel.
Worst thing ever though: modprobe seems not to be able to find the modules compiled for 2.4.18, just the ones for 2.2.20, and it _loads_ them! When I copy some of the 2.4.18 modules into the 2.2.20 directory and try to load them, it gives me a kernel version mismatch, claiming that I'm actually running 2.2.20!
I've recompiled the kernel about 10 times yesterday, yes, I did make mrproper, I didn't forget to make modules and make modules_install. I've even deleted the 2.2.20 vmlinuz, so there isn't even a 2.2.20 kernel on my hard drive.
After I've almost gone ballistic yesterday well after midnight, I've decided I'd better ask for help here before I do something I'd regret and my PC wouldn't live to regret... So, does anybody know where this may come from or what I could do about it?
Thanks a lot for any advice!
you should paste /etc/lilo.conf and a directory listing of /boot and the files in / .
meanwhile check this...
a) have you copied the new kernel image and system.map to the boot directory?
b) have you modify /etc/lilo.conf adding the entry for the new kernel?
c) if you have modified lilo.conf... have you executed /sbin/lilo to make the changes to be operative?
d) have you modified /etc/modules adding the new modules that you want to load at the boot? (and removing the ones you don't want to load)
Yup, if using lilo, you have to execute /sbin/lilo after making sure that the path to the new kernel image is correct inside /etc/lilo.conf
You can type:
to see what version of the kernel you are currently running. Sounds to me like it's still booting the old kernel, but you think it's booting the new one =P
Nope, the System.map is in /boot, map=System.boot is set in lilo.conf,, the right kernel was set to be booted too (I always change the lable when I recompile the kernel, and the right label was shown when I rebooted), so I definitely have run lilo after changing lilo.conf.
uname -r says taht it's 2.2.20_idepci, but this kernel doesn't exist on my ahrd drive anymore, I've deleted it to make fully sure.
It's not the first time I compile a new kernel, but the first time I have this kind of trouble... any ideas?
you should paste /etc/lilo.conf (without commentary and blank lines) and a directory listing of /boot and the files in / .
Without that we can't help you.
Number 1 RULE compiling a new kernel:
DO NOT REMOVE THE OLD KERNEL UNTIL YOU BE SURE THE NEW ONE IS WORKING FINE !! :)
nevertheless, I always keep the last two running kernels for emergency purposes.
|All times are GMT -5. The time now is 12:11 AM.|