Re: DragonFly-2.3.1.679.g4d28e master sys/bus/isa isa_common.c sys/bus/pci/amd64 pci_cfgreg.c pci_cfgreg.h pcibus.c pcibus.h sys/bus/pci dc21040reg.h eisa_pci.c fixup_pci.c hostb_pci.c sys/bus/pci/i386 legacy.c pci_bus.c pci_cfgreg.c pci_cfgreg.h pci

[Sepherosa Ziehau <sepherosa@xxxxxxxxx>]

On Thu, Jul 16, 2009 at 6:28 AM, YONETANI Tomokazu<qhwt+dfly@les.ath.cx> wrote:
> On Wed, Jul 08, 2009 at 03:59:07AM -0700, Sepherosa Ziehau wrote:
>> commit 4d28e78f7e70287ae03930ce9e16ba027ca69477
>> Author: Sepherosa Ziehau <sephe@dragonflybsd.org>
>> Date:   Sat Jul 4 11:32:50 2009 +0800
>>
>>     Sync PCI code with FreeBSD 7.2
>                :
>
>> http://gitweb.dragonflybsd.org/dragonfly.git/commitdiff/4d28e78f7e70287ae03930ce9e16ba027ca69477
>
> After this commit (with or without a few other `put back' commits),
> one of my PC won't boot because it can't find the root device, with or
> without ACPI enabled.  The kernel built from 4d28e78f7^ (that is, 061e48458)
> boots without a problem.  The most obvious change before and after this commit
> was that IRQ for atapci has changed from 23 to 10.  The following is

When APIC_IO is used, 'irq' only has software meaning.  It could be
changed, if intline reading order is changed; since we brought in
freebsd's PIR enumeration code, it changes the original intline
reading order.  But I do suspect that will be the cause.  Any chance
to pull all commits (inclusively) ->
a0f63ddc0da202edb4f9107d22aa5a54ec8d20cd, plus
0ba04f16a7a3f73835dfba74b6ab409214ec58ef into your local tree and test
again?

Best Regards,
sephe

-- 
Live Free or Die