On Sun, Jun 25, 2000 at 12:00:05AM -0500, linux-abit-digest wrote:
> I've just installed kernel 2.4.0-test-ac22 on my BP6 (dual celeron 466
> clocked to 525), and although it runs just fine (it seems like it
> doesn't crash as often as the stables kernels did),
I never had a crash under Linux with the stable series
(except when playing with OpenGL where the drivers are
in very alpha-state).
Did you try not to overclock?
> I'm getting these meesages in my dmesg:
>
> APIC error interrupt on CPU#0, should never happen.
> ... APIC ESR0: 0000000a
> ... APIC ESR1: 0000000a
> ... bit 1: APIC Receive CS Error (hw problem).
> ... bit 3: APIC Receive Accept Error.
>
>
> I've heard that this is quite normal on BP6 boards,
Yes. You know, dual configuration of the BP6 is
"designed for testing only"...
> but does anybody know what this means to my computer
It means that the bus where the processors and interrupt
controller talk about who does what gets bit errors.
Single-bit errors are detected and retried (and produce
the message in kernels > 2.3.something), double-bit
error can do anything.
It's a probability thing - if you get a few of them daily
and your machine is stable, ignore it. If you get tons
of them, there is a chance you'll get problems later.
> and if it can be solved?
Not overclocking reduces the incidence of these errors
dramatically, but if you see them at 75 MHz bus clock,
I doubt they will go completely away.
You can try to _lower_ CPU voltage a notch - when I did
some tests this seemed to have some effect.
BTW, the kernel developers have probably touched something
that triggers the errors. With 2.3.something there were
several errors daily in the log, with 2.4.0-test1
and -test2 I encounter one in a few days (433 @ 507).
Regards
--
Stano
--
=- To unsubscribe, email [EMAIL PROTECTED] with the -=
=- body of "unsubscribe linux-abit". -=
