Re: LILO bugs #146350 and others

[elf@buici.com]

I need a bit of guidance.

I have no problem using the bf2.4 boot images on any of my computers
including a k6-2/500.

Looking at the code, it appears that the trouble is with a new feature
in the 22.2 lilo that determines the ramdisk loads address using newer
bios calls.  It uses the Int15,0xe820 call to get a memory map.  Lilo
v21 only uses Int15,0x88.  Oddly, though, both methods constrain the
ramdisk to be beneath the 16MB boundary.  My first guess would be that
the fallback code for when e820 isn't available is at fault, but the
reporter has shown that the call works.

So, I could build a boot floppy that 
 a) Remove this feature entirely.  There is a conditional that
    eliminates all of this and falls back to the lilo21 behavior.
 b) Eliminate the 0xe820 call and let it perform that fallback.
 c) Look to the machines that fail and see if there is another common
    factor.

While the new code seems to be correct on the face of it, I'm not sure
why it is advantageous to use these BIOS entry points.  The int15,0x88
call seemed to be adequate for finding the ramdisk loading position.
Here's the note from the upstream CHANGES :

  Changes from version 22.1 to 22.2 (05-Feb-2002) John Coffman

    Boot Loader
    -----------
      - Vary placement of initrd:  (memory <= 64M) ==> load below 15M,
        whereas (memory > 64M) load just above 64M; use int 15h, fn's
        E820 & E801 to determine memory size > 64M


On Fri, May 10, 2002 at 07:51:36PM +1000, Russell Coker wrote:
> Currently I have no time and no hardware to try and fix LILO bugs.
> 
> I would appreciate it if someone else could investigate this matter for me 
> and fix it if it's determined to be serious.
> 
> People have told me that LILO in woody can't boot the kernel-image packages 
> that are in woody.  I expect that this only applies to certain hardware 
> configurations as I wouldn't expect anyone to release an i386 kernel that 
> can't be booted with the latest lilo.
> 
> -- 
> If you send email to me or to a mailing list that I use which has >4 lines
> of legalistic junk at the end then you are specifically authorizing me to do
> whatever I wish with the message and all other messages from your domain, by
> posting the message you agree that your long legalistic sig is void.
> 
> 
> -- 
> To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org


-- 
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org