From owner-cvs-all Mon Nov 26 13:10:28 2001 Delivered-To: cvs-all@freebsd.org Received: from freefall.freebsd.org (freefall.FreeBSD.org [216.136.204.21]) by hub.freebsd.org (Postfix) with ESMTP id 16F7537B417; Mon, 26 Nov 2001 13:10:16 -0800 (PST) Received: (from imp@localhost) by freefall.freebsd.org (8.11.6/8.11.6) id fAQLAG837823; Mon, 26 Nov 2001 13:10:16 -0800 (PST) (envelope-from imp) Message-Id: <200111262110.fAQLAG837823@freefall.freebsd.org> From: Warner Losh Date: Mon, 26 Nov 2001 13:10:16 -0800 (PST) To: cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: cvs commit: src/sys/i386/isa pci_cfgreg.c X-FreeBSD-CVS-Branch: RELENG_4 Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG imp 2001/11/26 13:10:16 PST Modified files: (Branch: RELENG_4) sys/i386/isa pci_cfgreg.c Log: MFC (kinda, the code is different there): When we're searching for an already configured line, and we find 0 as the irq for that line, ignore it. It is a crewel lye. Instead, treat it like we treat irq 255. Some BIOS writes seem to violate the spec by using 0 where they should use 255. The if for this is separated out so that we can ifdef it out if pc98 needs to have irq0 valid (this is in machdep code, so no other archs are relevant). This fixes many laptops that weren't able to properly route PCI interrupts before. Note: it will almost ceratainly not fix the crash some people see when we go to call the BIOS. Debugging thanks to: guido Revision Changes Path 1.1.2.6 +9 -6 src/sys/i386/isa/pci_cfgreg.c To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message