Subject: Re: [linux-audio-dev] VP kernels
From: Ingo Molnar (mingo_AT_elte.hu)
Date: Thu Sep 23 2004 - 23:32:08 EEST
* Ingo Molnar <mingo_AT_elte.hu> wrote:
> ok, some progress: managed to reproduce the jackd instant lockup on my
> P4 HT SMP box. (hacked my kernel to ignore the second CPU's ACPI
> entries so that i got a single-physical-CPU HT setup. The box has the
> same i8x0 sound stuff as your box.)
>
> not much debug info yet, but it's not a classic hard lockup. It rather
> seems to be jackd (or the kernel) looping indefinitely. Keyboard
> doesnt work because it's threaded by default ...
jackd is looping in the kernel:
Pid: 2305, comm: jackd
EIP: 0060:[<c0125772>] CPU: 1
EIP is at del_timer_sync+0xe2/0x100
EFLAGS: 00000246 Not tainted (2.6.9-rc2-mm3-VP-S5)
EAX: 00000000 EBX: f68f4000 ECX: ffffffff EDX: c1815f20
ESI: c0464828 EDI: c1816920 EBP: f68f4f50 DS: 007b ES: 007b
CR0: 8005003b CR2: b7fe49a0 CR3: 37413900 CR4: 000006b0
[<c01257b8>] del_singleshot_timer_sync+0x28/0x40
[<c0372096>] schedule_timeout+0xa6/0xf0
[<c0126120>] process_timeout+0x0/0x10
[<c012621e>] sys_nanosleep+0xde/0x170
[<c01044c9>] sysenter_past_esp+0x52/0x71
in del_singleshot_timer_sync() - a classic deadlock scenario. I dont yet
know what the other CPU is doing, but del_singleshot_timer_sync() is not
very preemption-friendly ... My bet is that jackd is starving softirqd
which wants to execute the timer function that jackd is trying to delete
=> deadlock.
Ingo
This archive was generated by hypermail 2b28 : Thu Sep 23 2004 - 23:34:29 EEST