Skip to contentSkip to conversation

PROBLEM: kernel crashes when recording audio

DISCUSSION Mikael Pettersson 2 participants 2 messages Sep 29, 2002

lore
Share & tools
lore
#0 lore raw
Quoted 10 lines · Malte Cornils

Long description:
When trying to capture audio/video from my bttv card (see lspci output below), my system's reaction ranges from hard freezes to instant reboots to segfaults - the latter variant enabled me to capture a kernel oops. I'm using avicap from avifile project, with XviD capture codec. Video capture works flawlessly. When I enable audio capture in addition (48 kHz, 16bit, Mono) is when the problem starts (not after some time, but at the moment I start the capture process). I do not have many ideas on why smp_apic_timer_interrupt is the problem, since this is a uniprocessor system.

smp_apic_timer_interrupt is also used in UP local APIC kernels. You're either running an SMP kernel on a UP box, or a local APIC enabled kernel.

Quoted 9 lines
Unable to handle kernel paging request at virtual address ffffe0b0
c0111f6b
*pde = 00001063
Oops: 0002
...
Code;  c0111f6b <smp_apic_timer_interrupt+b/d0>
00000000 <_EIP>:
Code;  c0111f6b <smp_apic_timer_interrupt+b/d0>   <=====
  0:   c7 05 b0 e0 ff ff 00      movl   $0x0,0xffffe0b0   <=====
Quoted 8 lines
Code;  c0111f72 <smp_apic_timer_interrupt+12/d0>
  7:   00 00 00 
Code;  c0111f75 <smp_apic_timer_interrupt+15/d0>
  a:   ff 05 84 53 25 c0         incl   0xc0255384
Code;  c0111f7b <smp_apic_timer_interrupt+1b/d0>
 10:   31 d2                     xor    %edx,%edx
Code;  c0111f7d <smp_apic_timer_interrupt+1d/d0>
 12:   f6 40 00 00               testb  $0x0,0x0(%eax)

You got a page fault in ack_APIC_irq(). This should never happen.

#1 lore raw

Hello,

Quoted 3 lines · Mikael Pettersson

smp_apic_timer_interrupt is also used in UP local APIC kernels. You're either running an SMP kernel on a UP box, or a local APIC enabled kernel.

CONFIG_SMP is not set, CONFIG_X86_LOCAL_APIC is enabled.

Quoted 1 line

You got a page fault in ack_APIC_irq(). This should never happen.

That's quite cool :-) How can I prevent this from happening then? Hmm... BTW, I got the same error on 2.4.19, with ALSA it works fine.

-Malte #8-)