dl580g3/amd64: [ACPI] 6.0-RELEASE will not boot with ACPI enabled

John Baldwin jhb at freebsd.org
Wed Jan 4 10:31:45 PST 2006


On Wednesday 04 January 2006 10:25 am, paketix at bluewin.ch wrote:
> >-- Original-Nachricht --
> >From: John Baldwin <jhb at freebsd.org>
> >To: paketix at bluewin.ch
> >Subject: Re: dl580g3/amd64: [ACPI] 6.0-RELEASE will not boot with ACPI
> > enabled Date: Tue, 3 Jan 2006 13:32:54 -0500
> >Cc: freebsd-acpi at freebsd.org
> >
> >On Friday 30 December 2005 02:38 am, paketix at bluewin.ch wrote:
> >> hi john
> >> thanks for your feedback!
> >> i created a smp kernel from RELENG_6_0 this morning and tested it on the
> >> dl580g3 with acpi enabled - but still the same behaviour...
> >> ...the system boots and hangs somewhere after initializing the nics
> >>
> >> please find the new logs (smp kernel config and boot -v) at
> >> http://mypage.bluewin.ch/poe/
> >
> >So 6.0 doesn't work with ACPI enabled with or without SMP.  Does it work
> >with
> >ACPI disabled?
> >
> >--
> >John Baldwin <jhb at FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
> >"Power Users Use the Power to Serve"  =  http://www.FreeBSD.org
>
> sorry for the missing logs

Does all combinations of 5.4 work ok (i386 vs amd64, UP vs SMP, acpi vs 
noacpi)?

> 6.0 works when acpi is disabled with or without SMP
> i did some further tests today with out of the box RELENG_6_0 kernels
> (GENERIC and SMP) - hope that helps to clearify the problem
> please find the 'boot -v' logs on my homepage as usual
> http://mypage.bluewin.ch/poe/
> a few remarks regarding the logs:
>
> - dl580g3_60release-p1_amd64_acpi_bootv.txt
>   releng_6_0 kernel uniprocessor amd64 with acpi enabled
>   fatal trap 28 after bpf is attached to lo0:

The trap 28 is fixed in releng_6.

> - dl580g3_60release-p1smp_i386_acpi_bootv.txt
>   releng_6_0 kernel multiprocessing i386 with acpi enabled
>   system resets after bpf is attached to lo0:

I would have no idea about why it would reset.

> - dl580g3_60release-p1smp_i386_noacpi_bootv.txt
>   releng_6_0 kernel multiprocessing i386 with acpi disabled
>   system hangs after reinit of ata0:

Very odd.  You are getting stray irq 13's, as if the ATA interrupt is hooked 
up to IRQ 13 rather than IRQ 14 on the APIC.  Does it always work ok if you 
disable APIC?

-- 
John Baldwin <jhb at FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve"  =  http://www.FreeBSD.org


More information about the freebsd-acpi mailing list