strange problem(s) - hardware related?

Ken Moffat ken at kenmoffat.uklinux.net
Sat Jan 11 05:43:10 PST 2003


On Sat, 11 Jan 2003, Dario Birtic wrote:

> 	Hello list,
> 
> 	I made my first LFS almost two years ago. Since then, I made many LFS
> builds on various hardware. I never had any unsolvable problem, at least
> not solvable by help of LFS-news archives :) But, my last build is
> choking, and I am out of luck, debugging it. I tried many "solutions"
> without success. Including but not limiting to kernel changes, gcc
> changes, new build, CFLAGS bashing, hardware changes ... you name it.
> The problem is as follows: box runs fine on itself, idles for days
> without any problems, but, when I do some IDE intensive job, *WITH*
> compiling (kernel-tree would do, mozilla, or anything big), it locks up
> solid, after 2-5 minutes. No SSH, no serial login, nothing. Logs are
> empty, also. Just plain lockup.

 Two suggestions - if you haven't already, add Magic SysRq key (in the
kernel hacking) part of the config. This might allow you to do
alt-printscreen-s to sync the disks before you reboot, giving a small
possibility that something can get logged. Also, try 2.4.20 or
2.4.21-pre? (latest that I know of is pre3) - patch for this is in
2.4/testing on the kernel.org mirrors. It might be a chipset bug that
someone has already fixed.

 Have you ever had a stable linux on this box, e.g. the host system you
used to build LFS ? If that *was* stable while building LFS, it sounds
unfortunately like either a toolchain error in your build, or something
in the host kernel that you either don't have in your current kernel, or
didn't enable.

 Perhaps, running top in another tty while you are compiling might show
something. Sounds like this will be a bear to debug, so good luck!

Ken
-- 
 Out of the darkness a voice spake unto me, saying "smile, things could be
worse". So I smiled, and lo, things became worse.



-- 
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-support' in the subject header of the message



More information about the lfs-support mailing list