Re: 5.2-RC: vnode_pager_alloc: 0xc6bc66f0 is not locked but should be

On 10 Dec, Jun Kuriyama wrote:
> 
> I updated my box to today's RELENG_5_2.  It was okay when 5.2-BETA,
> but I got this panic now.
> 
> 
> ----- console
> ...
> da0 at iir0 bus 2 target 0 lun 0
> da0: <Intel Host Drive   #00 > Fixed Direct Access SCSI-2 device
> da0: Tagged Queueing Enabled
> da0: 114431MB (234356220 512 byte sectors: 255H 63S/T 14588C)
> SMP: AP CPU #1 Launched!
> Mounting root from ufs:/dev/da0s1a
> vnode_pager_alloc: 0xc6bc66f0 is not locked but should be
> Debugger("Lock violation.
> ")
> Stopped at      Debugger+0x55:  xchgl   %ebx,in_Debugger.0
> db> trace
> Debugger(c0754805,c076797b,c6bc66f0,c0754846,e1fc8a70) at Debugger+0x55
> vfs_badlock(c0754846,c076797b,c6bc66f0,c07a8a80,c6bc66f0) at vfs_badlock+0x45
> assert_vop_locked(c6bc66f0,c076797b,c0581630,c07efbf8,c05ac758) at assert_vop_locked+0x62
> vnode_pager_alloc(c6bc66f0,da000,0,5,0) at vnode_pager_alloc+0x37
> vm_pager_allocate(2,c6bc66f0,da000,0,5) at vm_pager_allocate+0x57
> vm_mmap(c29acd68,e1fc8c2c,da000,5,7) at vm_mmap+0x3df
> mmap(c68dfc80,e1fc8d14,c076c639,3ee,8) at mmap+0x6e2
> syscall(2f,2f,2f,0,20002) at syscall+0x2c0
> Xint0x80_syscall() at Xint0x80_syscall+0x1d
> --- syscall (198, FreeBSD ELF32, nosys), eip = 0x2805acab, esp = 0xbfbfecc8, ebp = 0xbfbfed04 ---

I ran into this problem in 5.2-CURRENT.  It has sinced been fixed there,
in sys/vm/mmap.c 1.175.  I believe this is under consideration for
mergeing into RELENG_5_2.  In the mean time, you can get under way again
by setting vfs_badlock_panic and vfs_badlock_print to 0 in the ddb and
continuing.  This panic will also go away if you rebuild your kernel
with vnode lock checking disabled.