Re: Definitive Cyrix patch #3

Andrew Walker (andy@lysaker.kvaerner.no)
Fri, 6 Dec 1996 10:53:51 +0100 (MET)

Mike Jagdis wrote:
>
>
>
> This is version 3 of the Cyrix patch. It is against 2.1.7 (or so)
> but should patch cleanly against later 2.1s from reading the
> patches (other than Documentation/Configure.help).
>
> This adds some optional tweaking of performance features on
> the 6x86 and, completely new in this patch, the 5x86. If you
> enable the VSPM option for the 6x86 it works around the known
> wierdness.
>
> I have a couple of reports that a revision 4 6x86 sometimes
> freezes. I suspect this a problem with suspend on halt but
> have not been able to confirm it yet. If you have this problem
> try deleting the half dozen lines in arch/i386/kernel/head.S
> that set SUSP_HLT and let me know if it works.
>
> All other bugs are currently unknown so tell me about them :-).
>
> Mike

Okay, here's a point of reference

Motherboard: SOYP 5TF5 (Triton 430HX, 512kb pipeline cache)
Award BIOS
Processor: Cyrix 6x86 166+, rev 1, step 7
Ram: 2 x 16MB 60ns EDO

I've been recompiling a few kernels and getting trouble with
the dreaded gcc fatal errors, but with signal 4 (SIGILL) rather
than 11 (SIGSEGV). The Cyrix patch seemed to aggravate the
condition (died 4 or 5 times in a compile rather than once).
Also, with the Cyrix patch fully enabled I can't reboot into
Win95 (no flames please :), it just gets into a loop rebooting
itself and I have to switch the machine off.

I de-tuned the BIOS from 60ns RAM settings to 70ns RAM settings
and diasabled all but the processor identification and things
look fine at the moment.

-Andy

-- 
Andy Walker                              Kvaerner Engineering a.s.
Andrew.Walker@lysaker.kvaerner.no        P.O. Box 222, N-1324 Lysaker, Norway


   ......if the answer isn't violence, neither is your silence......


(pwei barmy army - oslo "filial")