2000-11-01 08:59:13

by john slee

[permalink] [raw]
Subject: test10 dies very early in boot

hardware:
* abit be6-2 mainboard
* 533 celeron (not overclocked)
* 192mb sdram
* seagate 20gb ide disk (not on ata66 port)

compiler: gcc version 2.95.2 20000220 (Debian GNU/Linux)

it gets as far as uncompressing the kernel and trying to boot it. no
further. (doesn't get as far as displaying the 'Linux version ...'
message). sysrq doesn't work (not suprising), ctrl-alt-delete doesn't
work either. reset button does work :-)

i did have an amateurish poke around comparing it to t10p7 (which boots
and runs fine on a near identical machine here at home), and all i could
turn up that seemed relevant was an a20 change in
arch/i386/boot/setup.S... and reversing that change didn't let it boot.
i have undoubtedly missed something.

i guess it had to happen eventually. this is the first kernel i've had
trouble booting in 3 years or so.

that was with bzimage.. will try with zimage tomorrow.

2.4.x is extremely impressive.

j.


2000-11-02 02:12:51

by David Konerding

[permalink] [raw]
Subject: Re: test10 dies very early in boot

john slee writes:
>hardware:
> * abit be6-2 mainboard
> * 533 celeron (not overclocked)
> * 192mb sdram
> * seagate 20gb ide disk (not on ata66 port)
>
>compiler: gcc version 2.95.2 20000220 (Debian GNU/Linux)
>
>it gets as far as uncompressing the kernel and trying to boot it. no
>further. (doesn't get as far as displaying the 'Linux version ...'
>message). sysrq doesn't work (not suprising), ctrl-alt-delete doesn't
>work either. reset button does work :-)


Um, I had the same problem with the kernel when I first downloaded it to try it out.
I was stumped for a while, turning of a number of features in the kernel,
before I figured it out. You need to switch the processor type the kernel is compiled
for from "P-III" to whatever is apropriate for your Celeron (I used a P-II, which
is basically the same instruction set as Celeron in Linux's view).

I didn't look at the source, but I bet early on the kernel tries to execute a P-III
instruction when compiled for P-III, and the P-II or Celeron gets confused and shuts down.

Once I compiled with P-II, kernel booted A-OK to multiuser :-)

Dave