From: Charles Sandmann <sandmann AT praline DOT no DOT NeoSoft DOT com>
Subject: Re: Interrupt conflict
To: mmcmpope AT inet DOT uni-c DOT dk (MouseHouse)
Date: Fri, 17 Mar 1995 11:03:27 -0600 (CST)
Cc: djgpp AT sun DOT soe DOT clarkson DOT edu

> >I am having problems with an interrupt conflict.
> > ...
> >DPMI is not being used.

I talked to Antonio about this.  When he runs his tests under DPMI they
work fine, which says this is a GO32 problem.  His problem is he is
also using graphics, so a test version of GRX 2.0 is required to both
at the same time.

To get stable interrupt handling you must run under DPMI, or re-write
sections of GO32.  If you are very careful (and lucky) you may get GO32
to work in some cases, but this is really limited to very simple operations.

> Me too. I've completely failed to get any interrupt servicing to work for
> more than 30 seconds. According to the FAQ, all I should do was post my
> problem on this list.
> 
> I know you may have been through this discussion before, but is there any
> way to make soundcard and serial interrupts work in DJGPP? Any way?
> Everything I've seen, including the serial routines from FSDB and the
> soundblaster SB02.zip example crash.

Did you try running in DPMI mode?  Are you running with lots of memory?

You can get stable interrupts under DJGPP, it just isn't easy yet.  I
have test codes which generate millions of interrupts and run for hours
on end (under DPMI) without any problems at all.