Path: utzoo!attcan!uunet!cs.utexas.edu!csd4.csd.uwm.edu!bionet!ames!amdahl!amdcad!diablo!phil
From: phil@diablo.amd.com (Phil Ngai)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Interupt conflicts
Message-ID: <26953@amdcad.AMD.COM>
Date: 25 Aug 89 01:27:31 GMT
References: <487@v7fs1.UUCP>
Sender: news@amdcad.AMD.COM
Reply-To: phil@diablo.AMD.COM (Phil Ngai)
Organization: Advanced Micro Devices, Inc. Sunnyvale CA
Lines: 28

In article <487@v7fs1.UUCP> sv@v7fs1.UUCP (Steve Verity) writes:
|	...So, Do you think that an I/O channel board is allowed to
|grab the IRQ line, and yank it low, until it needs to generate an
|interupt, at which point, it drives the line high???

Looking at the schematics on the IBM serial card, I see that it drives
the INT line with a 74125 tri-state driver. The 74125's output enable
is controlled by a register set by software. The 74125's input is
connected to the serial chip's interrupt line. When an interrupt
happens, the 74125 drives the INT line high. 

So, yes, that does seem to be the way it happens.

|	Yes, the interupts on the AT are active high. When a board
|wants an interrupt serviced, it must first pull the line low *WITH AN
|OPEN COLLECTOR DEVICE* and then let it *FLOAT* high. (these lines are
|pulled up on the I/O chanel.)

Wrong, bozo. The 74125 is not an open collector device. Nor does the
INT line float high when the interrupt arrives. It is driven high by
the 74125. And in the case that started all this, it is quite possible
for two cards which are mistakenly configured for the same interrupt
to end up with two INT line drivers fighting each other.

You seem to think you know a lot, unfortunately, you don't.
--
Phil Ngai, phil@diablo.amd.com		{uunet,decwrl,ucbvax}!amdcad!phil
"Today surgeons are highly respected but they were once just grave robbers."