From owner-freebsd-current Mon Aug 17 11:17:42 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id LAA15578 for freebsd-current-outgoing; Mon, 17 Aug 1998 11:17:42 -0700 (PDT) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from pinhead.parag.codegen.com (ppp-asfm08--171.sirius.net [205.134.241.171]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id LAA15523 for ; Mon, 17 Aug 1998 11:17:25 -0700 (PDT) (envelope-from parag@pinhead.parag.codegen.com) Received: from pinhead.parag.codegen.com (localhost.parag.codegen.com [127.0.0.1]) by pinhead.parag.codegen.com (8.9.1/8.8.8) with ESMTP id LAA08023; Mon, 17 Aug 1998 11:16:29 -0700 (PDT) (envelope-from parag@pinhead.parag.codegen.com) Message-Id: <199808171816.LAA08023@pinhead.parag.codegen.com> X-Mailer: exmh version 2.0.2 2/24/98 To: Alfred Perlstein cc: Paul van der Zwan , current@FreeBSD.ORG Subject: Re: Cards sharing irq's on PCI bus In-reply-to: Your message of "Mon, 17 Aug 1998 10:11:10 CDT." X-Face: =O'Kj74icvU|oS*<7gS/8'\Pbpm}okVj*@UC!IgkmZQAO!W[|iBiMs*|)n*`X ]pW%m>Oz_mK^Gdazsr.Z0/JsFS1uF8gBVIoChGwOy{EK=<6g?aHE`[\S]C]T0Wm X-URL: http://www.codegen.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Mon, 17 Aug 1998 11:16:29 -0700 From: Parag Patel Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG >well, that's unclear, last i heard linux was the only OS to support shared >IRQs and i don't even know how well it handled it. (i think it was more of >a "don't use them at the same time" deal) Um - I've used shared IRQs on non-Intel systems with OpenBSD/MIPS (the arc port) successfully. This was as a first-cut port to a somewhat broken NKK/MIPS eval board that simply connected each PCI IRQ from each slot together instead of "rotating" them like PC motherboards typically do. (It then ran them all into an Intel SIO chip's PC-compatible interrupt handler which would then be initially programmed to raise a single ISA IRQ for any PCI interrupt.) As long as a PCI driver's interrupt handler first checks the card to see if the interrupt is for it, and if not return, shared IRQs shouldn't be a problem, nor slow down performance too much. Some chips may be really slow to query and see if they really generated an interrupt, or some drivers may be assuming that if their i-handlers were called then they must have something to do instead of checking first. (Newer NKK eval boards and our customer's custom design route each PCI interrupt line directly to one of six separate interrupt pins on the MIPS processor, making things a whole lot more sensible as well as quite a bit faster.) -- Parag To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message