[Gllug] New kernel compile probs

David Irvine co2cool at yahoo.com
Sun Sep 9 04:38:12 UTC 2001


Ian Northeast wrote:

>gllug at codex.net wrote:
>
>>IIRC when lilo interprets the lilo.conf file, it has a look to see at what
>>disk offset the various kernel images are located at.  this info is then
>>written into the bootloader stuff.  this allows lilo to execute a
>>'program' without having to know about filesystems, it can just execute
>>from block xxx.  if you dont re run lilo, the data is for the locations of
>>the previous files, and hence they start to get executed.
>>
>>not quite sure why it doesnt just continue booting the old image.  of
>>course the above could be somewhat more complicated than my explanation.
>>
>
>It will, as long as all the blocks comprising the old image have not
>been reused for something. Which, in the case in point, they presumably
>have, leading to a kernel with large holes in it, which understandably
>won't boot too well.
>
>PS keeping the old kernel around has already been mentioned - I would
>just add that keeping a kernel boot floppy (or two - floppies go off
>quite a lot) around is also a good idea, for emergencies.
>
>Regards, Ian
>
I find that adding a test entry to lilo works quite well, you just add 
an entry with image=/boot/testkernel and then just cp the kernal image 
on top of that file every time you recompile,  run lilo and reboot if it 
works you can install it properly, if not you just boot the normal image.

The only time you can have problems is if you break it when you install 
it for good, but it works well for testing purposes.

David




-- 
Gllug mailing list  -  Gllug at linux.co.uk
http://list.ftech.net/mailman/listinfo/gllug




More information about the GLLUG mailing list