From: Andreas Busse (andy@resi.waldorf-gmbh.de)
Date: 06/14/93


From: andy@resi.waldorf-gmbh.de (Andreas Busse)
Subject: Re: 99p9/99p10 crashes my system
Date: 14 Jun 1993 12:28:35 GMT

In article <1993Jun13.115114.3885@mo.hobby.nl>, hans@mo.hobby.nl (Hans Oey) writes:
|> te@ipc1.rrzn.uni-hannover.de (Thomas Esser) writes:
|>
|> >Since, my linux crashes down unexpectedly after 5 - 30 min.
|> >of working. The system completely stands. (no core)
|>
|> >After such a crash, it takes 1 - 3 attempts to get the machine
|> >to reboot correctly. After power off/on, the machine (oftenly)
|> >stopps booting, and the BIOS sends a non-stopping sequence of
|> >3 beeps. This means "basic 64 k RAM error". I get into this
|> >problem after a systemcrash. If occurres 0 - 2 times until
|> >the reboot succedes.
|>
|> >This looks like a Hardware-Problem, but it did not occur with
|> >previous versions of linux (99p7) and it does not occur when running
|> >DOS. The update of the kernel to 99p10 did not help.
|>
|> >Please reply, if you have some hints.
|>
|> The same happens on my 486DX25 machine.
|> Try disabling the external cache. It gets terrible slow,
|> but doesn't freeze anymore. Hopefully 0.99.11 will solve the problem.
|> --
|> Hans Oey
|> hans@mo.hobby.nl

I have no problem with 0.99.9 on my i486/33 MHz. Perhaps there's something
wrong with the cache-rams. Usually they are so slow that they should never
work. But rams often are a bit faster than specified so that at least DOS will work.
Perhaps Linux has a code/data flow combination which scrambles the cache-contents.
If you have an AMI bios you can set the number of waitstates for the external caches
to a different value.

I had a strange problem on a i386 box w/o cache. DOS worked fine no matter how
the rams were configured (0 or 1 wait), but Linux ran only with 0 wait...
Don't ask me why.

Andy