Hi,
I'm having trouble booting any of the more recent kernels on one particular machine. The standard redhat 7.1 kernel (2.4.2-2) boots just fine, but booting a more recent kernel crashes the machine, somewhere just after or just before the hard disk gets mounted.
The machine has a MSI 6378 motherboard (VIA VT8361A & VT82C686A chipset) with an AMD Duron 750.
I've tried 2.4.12, 2.4.13, 2.4.14, and 2.4.16-pre1. I've attached the ksymoops output for the 2.4.14 crash. Note that this is the last oops from several that appeared on the screen, and I've manually copied it to paper and back, so there may be errors in it.
I have compiled a 2.4.16-pre1 with the latest kmsgdump, but then discovered this machine doesn't even have a floppy drive. So I'm getting one, I will post the processed oops later today.
The .config used for the 2.4.14 kernel is also attached. The only other patch to this kernel was the one to fix compilation with loopback support. I'll gladly provide more info if needed.
Please cc me in any replies as I'm too bandwidth-crippled right now to subscribe to LK.
Bye for now,
Ward Vandewege.
--
Pong.be -( "Microsoft isn't evil, they just make really crappy )-
Virtual hosting -( operating systems." -- Linus )-
http://pong.be -( )-
Ward Vandewege wrote:
> On Mon, Nov 26, 2001 at 02:07:00PM +0100, Fran?ois Cami wrote:
>
>>Maybe that will sound funny but... Could you try
>>compiling the kernel for an i386 processor and
>>report the results ?
>>
>
> I've compiled 2.4.16-pre1 optimized for i386 instead of Athlon/Duron, and had no problem whatsoever booting. I'm now beating it a bit, compiling a kernel etc, but it appears to be rock solid.
>
> This seems to suggest there is a problem in the Athlon/Duron optimization code.
As I thought.
> Thanks Fran?ois!
You're welcome.
Fran?ois
On Mon, Nov 26, 2001 at 02:07:00PM +0100, Fran?ois Cami wrote:
>
> Maybe that will sound funny but... Could you try
> compiling the kernel for an i386 processor and
> report the results ?
I've compiled 2.4.16-pre1 optimized for i386 instead of Athlon/Duron, and had no problem whatsoever booting. I'm now beating it a bit, compiling a kernel etc, but it appears to be rock solid.
This seems to suggest there is a problem in the Athlon/Duron optimization code.
Thanks Fran?ois!
Bye for now,
Ward Vandewege.
--
Pong.be -( "Fools ignore complexity. Pragmatists suffer it. Some )-
Virtual hosting -( can avoid it. Geniuses remove it." -- Perlis's )-
http://pong.be -( Programming Proverb #58, SIGPLAN Notices, Sept. 1982 )-
On Mon, Nov 26, 2001 at 02:58:48PM +0100, Fran?ois Cami wrote:
> Ward Vandewege wrote:
>
> > This seems to suggest there is a problem in the Athlon/Duron optimization code.
>
> As a matter of fact the problem seems to be more with the KT133/KT133A
> chipset, than with the Athlon/Duron optimization code.
>
> I'm nearly sure that compiling for K6 or i586 (the MMX one)
> would work fine, too, and it will beat a 386 kernel hands down.
Confirmed. 2.4.16-pre1 works just fine on my Duron, when compiled for K6.
I'm now compiling 2.4.16 - it's a busy day for gcc :-)
Bye for now,
Ward.
--
Pong.be -( "If you think penguins are fat and waddle, you have )-
Virtual hosting -( never been attacked by one running at you in excess of )-
http://pong.be -( 100 MPH." -- Linus )-