Re: WITNESS whine...

On Sunday 29 February 2004 11:09 am, Poul-Henning Kamp wrote:
> Mismatched locks to msleep(0xc1247d20, vmpfw):
>   old 0xc073af20 (vm page queue mutex), new 0 (null)
> Stack backtrace:
> backtrace(1,c657c7e0,0,d9c14bcc,c0509a06) at 0xc0503b5a = backtrace+0x12
> sleepq_add(c62bad00,c1247d20,0,c06b3fc5,0) at 0xc0521ccf = sleepq_add+0x1bf
> msleep(c1247d20,c073af20,244,c06b3fc5,0) at 0xc0509a06 = msleep+0x29e
> vm_page_sleep_if_busy(c1247d20,1,c06b3fc5) at 0xc05f6a55 =
> vm_page_sleep_if_busy +0xa5
> vm_fault(c633ddd4,8054000,2,8,c657c7e0) at 0xc05eb5f4 = vm_fault+0x31c
> trap_pfault(d9c14d48,1,8054000,8054000,0) at 0xc06503ab = trap_pfault+0xe3
> trap(2f,2f,2f,8054000,4) at 0xc064ff5a = trap+0x1e2
> calltrap() at 0xc063f558 = calltrap+0x5
> --- trap 0xc, eip = 0x28127eeb, esp = 0xbfbfebe0, ebp = 0xbfbfec18 ---
> Mismatched locks to msleep(0xc13005b0, vmpfw):
>   old 0xc073af20 (vm page queue mutex), new 0 (null)
> Stack backtrace:
> backtrace(1,c657c7e0,0,d9c14bcc,c0509a06) at 0xc0503b5a = backtrace+0x12
> sleepq_add(c62ba5c0,c13005b0,0,c06b3fc5,0) at 0xc0521ccf = sleepq_add+0x1bf
> msleep(c13005b0,c073af20,244,c06b3fc5,0) at 0xc0509a06 = msleep+0x29e
> vm_page_sleep_if_busy(c13005b0,1,c06b3fc5) at 0xc05f6a55 =
> vm_page_sleep_if_busy +0xa5
> vm_fault(c633ddd4,8056000,2,8,c657c7e0) at 0xc05eb5f4 = vm_fault+0x31c
> trap_pfault(d9c14d48,1,8056000,8056000,0) at 0xc06503ab = trap_pfault+0xe3
> trap(2f,2f,2f,8056000,4) at 0xc064ff5a = trap+0x1e2
> calltrap() at 0xc063f558 = calltrap+0x5
> --- trap 0xc, eip = 0x28127eeb, esp = 0xbfbfebe0, ebp = 0xbfbfec18 ---

Not, witness, sleepq(9). :)  It's a bug in msleep() though and I'm testing a 
fix for it.

-- 
John Baldwin <jhb_at_FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve"  =  http://www.FreeBSD.org