From owner-freebsd-current Tue May 12 18:51:49 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id SAA18563 for freebsd-current-outgoing; Tue, 12 May 1998 18:51:49 -0700 (PDT) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from smtp03.primenet.com (daemon@smtp03.primenet.com [206.165.6.133]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id SAA18556 for ; Tue, 12 May 1998 18:51:44 -0700 (PDT) (envelope-from tlambert@usr09.primenet.com) Received: (from daemon@localhost) by smtp03.primenet.com (8.8.8/8.8.8) id SAA04464; Tue, 12 May 1998 18:51:46 -0700 (MST) Received: from usr09.primenet.com(206.165.6.209) via SMTP by smtp03.primenet.com, id smtpd004402; Tue May 12 18:51:38 1998 Received: (from tlambert@localhost) by usr09.primenet.com (8.8.5/8.8.5) id SAA26286; Tue, 12 May 1998 18:51:29 -0700 (MST) From: Terry Lambert Message-Id: <199805130151.SAA26286@usr09.primenet.com> Subject: Re: Intel Etherexpress PRO/100+ PCI To: dg@root.com Date: Wed, 13 May 1998 01:51:29 +0000 (GMT) Cc: tlambert@primenet.com, steve@visint.co.uk, freebsd-current@FreeBSD.ORG In-Reply-To: <199805122233.PAA28496@implode.root.com> from "David Greenman" at May 12, 98 03:33:15 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > >Uh, that's what I said. > > No, you said that you thought the reporting was wrong because one card > should report using on INT A while another should report using INT B, etc. [ ... ] > > ...and here you say this again. INT A is a PCI bus pin, it is not a > specific interrupt. All of the PCI cards that need to interrupt use the > "INT A" line, but since that is different for each slot, a different physical > interrupt is usually used. A card would use INT B usually only after also > using INT A (i.e. it needs two interrupts). Which "INT" line the card uses > is a function of how the card is wired and has nothing to do with the BIOS > or FreeBSD. Right. > So, the "INT" letter is irrelevant. What you should be paying attention > to is the irq number that is reported, and you want that to be unique if > possible. The int that is reported by FreeBSD should match the PCI mapping. For example: ... int a irq 10 ... ... int a irq 11 ... ... int a irq 12 ... ... int a irq 13 ... Should be reported based on the PCI mapping of PCI bridge pins to irq's. That is, reporting "int a" is incorrect, since "a" is not the PCI bridge pin, it's the slot pin. The slot pin is not useful, unless there was additional reporting of the card by the slot it was in, and whether or not the PCI interrupts were cascaded or not by the PCI bridge used. In other words, if it's always going to be "int a", why report the "int" at all? If I want to know if interrupts are being shared, which is the primary thing I'm interested in, the int pin(s) utilized by the card *from the point of view of the PCI bridge* is what's useful. This, for example, is not very useful, except for making the dmesg more noisy than it needs to be: ] vga0 rev 6 int a irq 12 on pci0:10 ] de0 rev 32 int a irq 10 on pci0:11 ] ncr0 rev 18 int a irq 11 on pci0:12 In the other hand, if I had multiple bridges, knowing the int->irq mapping could be *very* useful. As it is, it's possible for me to get the hierarchical mapping only by looking at the "pciX:YY". Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message