Date:	Wed, 16 Sep 1998 19:14:37 +0300 (IDT)
From:	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To:	jtgalkowski AT alum DOT mit DOT edu
cc:	djgpp AT delorie DOT com
Subject:	Re: Recently installed version of DJGPP's gcc hangs
In-Reply-To:	<6to9i8$of7$1@nnrp1.dejanews.com>
Message-ID:	<Pine.SUN.3.91.980916190638.8080A-100000@is>
MIME-Version:	1.0

On Wed, 16 Sep 1998 jtgalkowski AT alum DOT mit DOT edu wrote:

> It turns out that the error described below is an RMCB
> page fault of the kind which has been discussed in this
> group before, notably when handling interrupts.  There was
> a contribution from Mr Church about how memory needs to be
> locked because of the dynamic disposition of virtual memory
> pages.

This is irrelevant to the case in point.  Memory only needs to be locked 
if the code that runs is called from a hardware interrupt handler.  GCC 
doesn't use such a code, so this is not your case.

> as I promised in my original post -- I did
> a "go32-v2" before trying this.

Running go32-v2 before the compilation can neither help nor hurt.

> Page Fault cr2=10020b5c in RMCB at eip=10bc; flags=3002
> 
> eax=00000000 ebx=000010b6 ecx=00000019 edx=00000005
> esi=000010b6 edi=00020b5c ebp=000623b8 esp=000020a4
> cs=2b ds=3b fs=33 gs=bf _s=33 error=0006

The first thing I would suspect is that CWSDPMI, the DPMI server used on
plain DOS, runs out of its internal heap, so I would suggest to bump up
its heap size and try again.  To this end, use the CWSPARAM.EXE program
and change the "Paragraphs of memory for extra CWSDPMI internal heap" 
setting (see the file CWSPARAM.DOC for more details).  More info about
this is available in the DJGPP FAQ list (v2/faq210b.zip, from the same
place you get DJGPP), section 6.4.

- Raw text -