APIC on AMD Athlon 2500+ Broken?

[Tim Writer]

Kareem Shehata <kareem-d+8TeBu5bOew5LPnMra/2Q at public.gmane.org> writes:

> On Mon, 2004-07-05 at 11:13, Lennart Sorensen wrote:
> > In my experience, your best bet on the nforce2 chipset is to turn off
> > the APIC in the bios, no matter what OS you run (it causes sound
> > choppiness in XP for example, and yeah linux can lock up with it too).
> > It seems not to matter which board maker it is, it seems to just be a
> > not too reliable APIC implementation in an otherwise excelent chipset.
> > Not as if the APIC is a big deal to have on a single cpu system.
> > 
> > I have an A7N8X-E-DX at work, which is running great with 2.6 kernels,
> > and I am 99% sure I turned of the APIC in the bios when we got the
> > machine because I know what it did to XP on the A7N8X-DX boards.
> > 
> > Lennart Sorensen
> 
> I think I once turned off APIC in the bios but left it compiled in the
> kernel, and it still locked up.  Strange.  I think I'll stick with my
> current kernel build since it seems to have everything working, and as
> you said APIC isn't that big a deal.

The BIOS documentation for my home system (single CPU Gigabyte P4) states
that the BIOS and OS APIC settings must match.  IOW, either APIC is disabled
in the BIOS _and_ in the OS or it's enabled in the BIOS _and_ enabled in the
OS.

In my experience, APIC can be a big deal for single CPU systems.  For
example, the on-board Promise RAID controller (which I use as additional IDE
channels) on my home system conflicts with USB unless APIC is enabled in the
BIOS and kernel.  Without APIC, I'm unable to use the RAID controller at all.

-- 
tim writer <tim-s/rLXaiAEBtBDgjK7y7TUQ at public.gmane.org>                                  starnix inc.
905.771.0017 ext. 225                           thornhill, ontario, canada
http://www.starnix.com              professional linux services & products
--
The Toronto Linux Users Group.      Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml