Re: IO-APIC on MS-6163 (2.4.23). ACPI?

[Ryan Underwood <nemesis-lists-RBwgCYbvQ+vk1uMJSBkQmQ@public.gmane.org>]

[-- Attachment #1: Type: text/plain, Size: 1703 bytes --]


On Wed, Dec 10, 2003 at 02:40:25PM +0100, Mikael Pettersson wrote:
> Ryan Underwood writes:
>  > 
>  > Attached dmesg and interrupts with acpi=off.  Also attached kernel
>  > config.
>  > 
>  > As you see, it acknowledges a local APIC but mentions nothing about
>  > IO-APIC.  So I guess this isn't ACPI issue after all.  But, I wonder why
>  > the IO-APIC isn't being used on this board anymore.  There was a
>  > blacklist entry for it, but it was removed after it was discovered to be
>  > in error.
> 
> Your dmesg log doesn't mention any MP-table. With neither ACPI MADT
> nor an MP-table the kernel has no way of knowing about any I/O-APIC.
> 
> You do have UP_IOAPIC enabled, so the kernel should look for it.
> Please check your BIOS settings, and try with no ACPI at all in the kernel
> (just to verify that acpi=off doesn't prevent MP-table parsing).

I went back and found the messages corresponding to our previous
discussions:

http://www.cs.helsinki.fi/linux/linux-kernel/2003-23/1675.html

Looks like again I've confused the local APIC with the IO-APIC regarding
that blacklist entry.  However, I thought 440BX had a IO-APIC built-in
to the southbridge.  Again, that was proved to be wrong by examining the
integrator's guide:
ftp://download.intel.com/design/chipsets/designex/29063401.pdf

page 99 of the PDF says:
"An I/O APIC is required for a DP system and optional for a UP system."

So it looks like I've never had an I/O-APIC on this board at all, only a
local APIC, and this thread is thus a waste of time. :)

In any case, thanks for leading me to the right answer!

-- 
Ryan Underwood, <nemesis-RBwgCYbvQ+vk1uMJSBkQmQ@public.gmane.org>

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]