From:	sandmann AT clio DOT rice DOT edu (Charles Sandmann)
Message-Id:	<9608021343.AA10583@clio.rice.edu>
Subject:	Re: NextGen 586, Floating point woes
To:	demmer AT LSTM DOT Ruhr-UNI-Bochum DOT De (Thomas Demmer)
Date:	Fri, 2 Aug 1996 08:43:31 -0600 (CDT)
Cc:	djgpp AT delorie DOT com (DJGPP Mailing List)
In-Reply-To:	<3201D46D.52BF@LSTM.Ruhr-UNI-Bochum.De> from "Thomas Demmer" at Aug 2, 96 11:11:57 am

First suggestion - are you using CWSDPMI r2?  I was sent a patch for r2
to allow FPU stuff to work with the NextGen.

> > they fail (read:  crash with a floating point exception) on the
> > NextGen 586 processor.  Why?

If the above fix doesn't help, the exact error message, EIP, registers, etc
would be helpful (using the debugger to pinpoint the asm code of the line
causing the problem).

> The function is called by npxsetup.c, same directory. It checks
> the environment variable 387 to see if you set it to "N", but
> unfortunately not, if you set it to "Y". Charles, is that not something
> to change?

Setting it to Y causes hard crashes (taking out Windows, for example) on
boxes I tried without FPUs for a couple of reasons.  We need to find out
why code directly out of the intel manual doesn't work on this environment,
if we need to set something else first, for example.

In any case, none of these things get fixed unless someone wants to spend
the time to work through them - I'm glad to help but it works on everything
I have access to.

- Raw text -