From owner-freebsd-current  Wed Apr 17 23:23:56 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id XAA17067
          for current-outgoing; Wed, 17 Apr 1996 23:23:56 -0700 (PDT)
Received: from nixpbe.pdb.sni.de (mail.sni.de [192.109.2.33])
          by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id XAA17062
          for <current@freefall.freebsd.org>; Wed, 17 Apr 1996 23:23:47 -0700 (PDT)
Received: (from nerv@localhost) by nixpbe.pdb.sni.de (8.6.12/8.6.12) id IAA20609 for current@freefall.freebsd.org; Thu, 18 Apr 1996 08:24:07 +0200
Message-Id: <199604180624.IAA20609@nixpbe.pdb.sni.de>
Subject: Re: EtherExpress16 problems
To: dwhite@resnet.uoregon.edu
Date: Thu, 18 Apr 96 8:23:30 MDT
From: Greg Lehey <lehey.pad@sni.de>
Cc: alk@Think.COM, current@freefall.freebsd.org, questions@freebsd.org
In-Reply-To: <Pine.BSF.3.91.960417110844.5027L-100000@riley-net170-164.uoregon.edu>; from "Doug White" at Apr 17, 96 11:11 am
X-Mailer: xmail 2.4 (based on ELM 2.2 PL16)
Sender: owner-current@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

>
> On Wed, 17 Apr 1996, Tony Kimball wrote:
>> (Incidentally, I do not understand about IRQs 2 and 9.  Is there a
>> document I might read?  Are they available for assignment to devices,
>> generally?)
>
> The book "Running FreeBSD" has a good discussion on it.

Thank you.

> The 2/9 is a
> hack to support 16 bit interrupts.  Back in the old days of XTs they only
> had 8 bit interrupts, and when 16 bit came they decided to cascase IRQ 2
> to 9.  Don't ask me why.  :)

The interrupt controller chip used in the XT and AT was the Intel
8259A.  It handled 8 interrupts, but had a so-called "cascade mode",
where you could designate each of the primary interrupts as
representing another group of 8.  This is what they did in the AT:
they took IRQ2, which up to this point had hardly been used, and used
it for the interrupt input from the second 8259A.  The *real* IRQ2
thus represents all of the IRQs from 8 to 15.  This is (almost)
completely transparent to the software, the only exception being the
interrupt controller initialisation routines.

There is also an ISA bus line called IRQ2 (B4: on the rear of the
connector, 4th from left when you have the connector pointing down)
It used to be connected to IRQ2 on the XT, but it now homeless, since
it couldn't be connected to IRQ2.  Instead, they connected it to
IRQ9.  Unfortunately, the line is (AFAIK) still called IRQ2.  Thus the
confusion.

Bottom line: it's really IRQ9, but it's the only one of the IRQs 8-15
which you can use with an 8 bit board.

Is this clear?  Is it interesting?  Shall I put it in the next edition
of the book?

Greg