A colleague at AMD also had a look at this issue, please also provide
the information he requests.
thanks,
Markus
---------- Forwarded message ----------
From: Andreas Herrmann <[EMAIL PROTECTED]>
Date: Fri, 20 Jul 2007 12:45:30 +0200
Subject: Re: Fwd: [Em28xx] BUG REPORT: APIC error with Hauppauge HVR 950
To: Markus Rechberger <[EMAIL PROTECTED]>
On Fri, Jul 20, 2007 at 10:51:41AM +0200, Markus Rechberger wrote:
> ---------- Forwarded message ----------
> From: Lincoln Stein <[EMAIL PROTECTED]>
> Date: Thu, 19 Jul 2007 17:35:21 -0400
> Subject: [Em28xx] BUG REPORT: APIC error with Hauppauge HVR 950
>
> Hi All,
>
> I am using the Hauppauge WinTV-HVR-950 with Linux kernel
> 2.6.20-16-generic(Ubuntu Feisty Fawn); this is on an AMD64 dual-core
> system. Video capture
> from terrestrial digital works fine except that I am intermittently getting
> kernel errors like this while capturing a video or EIT stream:
>
> kernel: [93964.072142] APIC error on CPU0: 00(40)
>
Hmm, would be nice to have complete /var/log/messages
"noapic" can be one solution to this but it has some drawbacks:
- It disables use of I/O APIC which is needed to route IRQs to
multiple CPUs, I.e., IRQs will be bound to CPU0. AFAIK you won't
be able to balance IRQs between your CPUs anymore.
- You just have 16 IRQs.
You might have to deal with other errors when disabling IO-APIC. Please
check for errors in /proc/interrupts when using "noapic".
What is the counter for "ERR"?
Other possible kernel parameters which solve related problems with
broken BIOSs and chipsets are worth a try as well:
enable_timer_pin_1 [i386,x86-64]
Enable PIN 1 of APIC timer
Can be useful to work around chipset bugs
(in particular on some ATI chipsets).
The kernel tries to set a reasonable default.
disable_timer_pin_1 [i386,x86-64]
Disable PIN 1 of APIC timer
Can be useful to work around chipset bugs.
disable_8254_timer
enable_8254_timer
[IA32/X86_64] Disable/Enable interrupt 0 timer routing
over the 8254 in addition to over the IO-APIC. The
kernel tries to set a sensible default.
acpi_skip_timer_override [HW,ACPI]
Recognize and ignore IRQ0/pin2 Interrupt Override.
For broken nForce2 BIOS resulting in XT-PIC timer.
acpi_use_timer_override [HW,ACPI}
Use timer override. For some broken Nvidia NF5 boards
that require a timer override, but don't have
HPET
But w/o more debug and hardware info it is hard to tell what is the proper
solution for your problem.
> After the error message occurs the processing of USB interrupts from the
> mouse slows to a crawl, making the USB mouse unusable. The error
> reproducibly occurs within hours to days of activating the HVR-950,
> depending on how much I am using it. Since yesterday I have booted the
> system up with "noapic" and so far the error has not occurred.
>
> I am using the Mercurial version of the v4l-dvb-experimental source code
> last updated 13 July 2007. Is this a known bug?
>
> Thanks in advance,
>
> Lincoln Stein
BTW, on my private AMD 64 X2 5000+ with Gigabyte mainboard and Nvidia
chipset I have to use acpi_use_timer_override to be able to boot the machine.
With noapic the machine booted as well but was quite unresponsive due to
interrupt delivery errors (as seen in /proc/interrupts).
This BIOS/chipset problems with APIC are a mess and it really depends on
your configuration to decide what is a proper workaround.
And last not least some of the problems can be solved with BIOS updates.
Regards,
Andreas
_______________________________________________
Em28xx mailing list
[email protected]
http://mcentral.de/mailman/listinfo/em28xx
