From:	Vinzent Hoefler <JeLlyFish DOT software AT gmx DOT net>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: Libnet serial port problem
Date:	Wed, 15 Jan 2003 10:07:01 -0500
Organization:	JeLlyFish software
Lines:	20
Message-ID:	<let30b.la2.ln@jellix.jlfencey.com>
References:	<f1a1bbde DOT 0301132122 DOT 7672b79f AT posting DOT google DOT com> <3e23e78b$1 AT news DOT infonet DOT ee> <kc410b DOT 4u DOT ln AT jellix DOT jlfencey DOT com> <3E251E0C DOT 3144A6A8 AT invalid DOT com>
NNTP-Posting-Host:	firewall.mdc-dayton.com (12.161.103.180)
Mime-Version:	1.0
X-Trace:	fu-berlin.de 1042643292 22319457 12.161.103.180 (16 [175126])
X-Orig-Path:	jellix.jlfencey.com!nobody
X-Newsreader:	KCNews v0.19b (CAOS 4.1)
X-Phone:	+1-937-271-7158
X-Homepage:	http://jlfencey.com
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp
Reply-To:	djgpp AT delorie DOT com

Ismo Salonen wrote:

> Actually the problem is hardware related. On older computers the IRQ
> line is
> high level active. Each uart drives the irq line it is connected to
> via tristate buffer. The buffer is enabled with a bit in the uart. If
> two uarts drive the same line then results are unpredictable.

So this is bad hardware design. As always.

But anyway, the OP talked of COM1/2. Those ports usually do not share 
the same IRQ.


Vinzent.

-- 
Ten years of rejection slips is nature's way of telling you to stop
writing.
                -- R. Geis

- Raw text -