Xref: news2.mv.net comp.os.msdos.djgpp:5517
From: tibors AT opera DOT iinet DOT net DOT au (Tibor Sashegyi)
Newsgroups: comp.os.msdos.djgpp
Subject: Re: GDB remote serial debugging
Date: 28 Jun 1996 18:30:23 +0800
Organization: iiNet Technologies
Lines: 30
Message-ID: <4r0c7v$gfe@opera.iinet.net.au>
References: <31D28B43 DOT 38C0 AT info DOT vub DOT ac DOT be>
NNTP-Posting-Host: opera.iinet.net.au
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp

In comp.os.msdos.djgpp you write:
>I'm trying to setup gdb for remote debugging. I've found the TSR program
>asynctsr.com in the archive aeasy102.zip. This archive also contains a
>test program gtest, that sets up a tty session with a remote host and
>works just fine.
>But when I recompile gtest, it crashes when I run it with the error:
>Segmentation violation in pointer 0x00000018

>This happens when it tries to read the address at interrupt 12 (where
>the TSR program is installed). It reads the address as:
>	*((unsigned short*)(0xE0000000 + 12*4))
>This produces a general protection error. As said, the compiled program
>I found in the archive works perfectly. It's just the program that I
>compile with gcc v. 2.7.2 that produces this problem. Since I have to
>recompile gdb to set it up for remote debugging, gdb suffers from the
>same problem.

DJGPP's DPMI stuff has changed between version 1.0 and 2.0 so that the
access to low core doesn't work the same way anymore.

I posted an article with this and more questions, but got no answers. At the
moment I have changed GDB to allow remote debugging via COM1 or COM2 and
I am also working on the target specific code to support remote debugging
via a serial port. The target is a single board PC based on a AMD 386.

Let me know if you are interested in cooperating on this project.

Tibor

Email: tibors AT iores DOT com DOT au