Date: Sat, 15 Oct 2005 12:54:19 -0600 (MDT) From: "M. Warner Losh" <imp@bsdimp.com> To: freebsd@psam.se Cc: freebsd-current@freebsd.org Subject: Re: Thoshiba Tecra 8000 with 3com 3CXFE575CT Message-ID: <20051015.125419.47698820.imp@bsdimp.com> In-Reply-To: <4350C749.9070704@psam.se> References: <20051013.213943.05622321.imp@bsdimp.com> <43503EB8.7040509@psam.se> <4350C749.9070704@psam.se>
next in thread | previous in thread | raw e-mail | index | archive | help
Please try this patch. I think we've had a minor regression in our PCI BIOS routed interrupts. Also, please try w/o ACPI enabled to see if that changes things. With acpi enabled, I'm not sure this patch will do anything for you. If it does with acpi disabled, it gives a real strong clue what to look for in the acpi code. We used to always route the interrupt, even when the bios said it had one, now it looks like we do that less agressively. Warner Index: pci_pir.c =================================================================== RCS file: /home/ncvs/src/sys/i386/pci/pci_pir.c,v retrieving revision 1.119 diff -u -r1.119 pci_pir.c --- pci_pir.c 8 Sep 2005 17:07:12 -0000 1.119 +++ pci_pir.c 15 Oct 2005 18:51:07 -0000 @@ -348,6 +348,9 @@ irq, entry->pe_bus, entry->pe_device, pin + 'A', pci_link->pl_id); pci_link->pl_irq = irq; + if (!pci_link->pl_routed) + pci_pir_biosroute(entry->pe_bus, entry->pe_device, 0, + pin - 1, pci_link->pl_irq); pci_link->pl_routed = 1; return; }
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20051015.125419.47698820.imp>