>Number:         35886
>Category:       kern
>Synopsis:       panic "lockmgr: no context" during process exit
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Mar 01 16:25:00 +0000 2007
>Originator:     Chuck Silvers
>Release:        NetBSD 4.0_BETA2
>Organization:
me
>Environment:
Architecture: i386
Machine: i386
>Description:

I got this panic the other day:

#10 0xc0467714 in panic ()
#11 0xc0440ad2 in lockmgr ()
#12 0xc03fa3ef in uvm_unmap1 ()
#13 0xc03f3ab5 in km_vacache_free ()
#14 0xc0463803 in pool_allocator_free ()
#15 0xc0464184 in pr_pagelist_free ()
#16 0xc0464108 in pool_put ()
#17 0xc0463803 in pool_allocator_free ()
#18 0xc0464184 in pr_pagelist_free ()
#19 0xc0464108 in pool_put ()
#20 0xc044dc00 in sigactsfree ()
#21 0xc043a66c in exit1 ()
#22 0xc043aa4f in sys_exit ()
#23 0xc05068ee in syscall_plain ()
#24 0xc010067c in syscall1 ()

(gdb) x/s panicstr
0xc0903cce <copyright+79022>:    "lockmgr: no context"


lockmgr() is complaining that curlwp is NULL, which is true since
exit1() sets it to NULL before calling sigactsfree() and several
other functions that call pool_put().

it looks like this problem exists in 4.x and 3.x only (and not -current).


>How-To-Repeat:
	have a process exit and be unlucky.
>Fix:
	not provided.