IOAPIC on Via KT266a

IOAPIC on Via KT266a

[fauxpas]

[-- Attachment #1: Type: text/plain, Size: 716 bytes --]

I have a UP system with Via kt266a chipset.  When I enable APIC in
the BIOS and in Linux, the system boots mostly normally, but I get
a fairly constant stream of

APIC error on CPU0: 02(02)

with a smattering of other numeric codes from time to time.  Most
things still work, but there are a number of oddities and instabilities,
most notably my integrated uhci USB controllers give usb bulk-msg
timeouts on every device.

This happens with 2.4.20 vanilla, 2.4.21-pre4-ac6, Debian's 2.4.18-bf24
and every other 2.4 series kernel I've tried... can somebody point me 
at a faq/patch/person to help get me started resolving this issue ?

Thanks in advance.

-- 
Josh Litherland (fauxpas@temp123.org)

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

Re: IOAPIC on Via KT266a

[Mikael Pettersson]

fauxpas@temp123.org writes:
 > I have a UP system with Via kt266a chipset.  When I enable APIC in
 > the BIOS and in Linux, the system boots mostly normally, but I get
 > a fairly constant stream of
 > 
 > APIC error on CPU0: 02(02)
 > 
 > with a smattering of other numeric codes from time to time.  Most
 > things still work, but there are a number of oddities and instabilities,
 > most notably my integrated uhci USB controllers give usb bulk-msg
 > timeouts on every device.

02 is "Receive Checksum Error", and it's explained both in apic.c
and in Intel's IA32 Volume 3 manual.

This is almost certainly a hardware problem: your machine's APIC bus
is corrupting messages, or some other agent than the CPU is creating
corrupt messages. This isn't exactly unheard of for non-Intel chipsets.

Disable IO_APIC support and you should be fine. If you still get these
errors, you'll have to either live with them or disable UP local APIC
support as well.

/Mikael

Re: IOAPIC on Via KT266a

[fauxpas]

[-- Attachment #1: Type: text/plain, Size: 1352 bytes --]

On Fri, Feb 28, 2003 at 03:19:01PM +0100, Mikael Pettersson wrote:

> This is almost certainly a hardware problem: your machine's APIC bus
> is corrupting messages, or some other agent than the CPU is creating
> corrupt messages. This isn't exactly unheard of for non-Intel chipsets.

Hmmm... windows seems to have a workaround for at least part of the
problem, I spent the last day trying to confuse the APIC unsuccessfully
on that system.  Could it be a problem with the uhci driver perhaps?
The only two persistant symptoms are the usb failure and the slew of 
messages.

The APIC errors I receive are mostly 02(02) but a good deal of
02(01), 01(02) and 01(01).  Checksum errors galore.

This only happens with IO-APIC is enabled throughout:

hub.c: new USB device 00:11.2-1, assigned address 2
usb_control/bulk_msg: timeout
usb.c: USB device not accepting new address=2 (error=-110)
hub.c: new USB device 00:11.2-1, assigned address 3
b.c: USB device not accepting new address=3 (error=-110)
hub.c: new USB device 00:11.2-2, assigned address 4
usb_control/bulk_msg: timeout
usb.c: USB device not accepting new address=4 (error=-110)
hub.c: new USB device 00:11.2-2, assigned address 5
usb_control/bulk_msg: timeout
usb.c: USB device not accepting new address=5 (error=-110)

-- 
Josh Litherland (fauxpas@temp123.org)

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