null pointer bug (from ralph)

have to reacquire lock before continuing

delete read-only checking; it must be done by the filesystem layers

delete read-only checking; it must be done by the filesystem layers

delete read-only checking; it must be done by the filesystem layers

do not allow time to go backwards when in secure mode

vnode_free_list_slock has to be exported for LFS

new system call argument convention

make provision for shared access locks on vnodes not otherwise
doing underlying locking

more precise output from lockmgr_printinfo

get rid of dead code

merge vfs_lock and vfs_busy; interlock mount points properly

allow drained locks to be reenabled

attempt to avoid spurious simple_lock complaints from interrupt level

sysctl_vnode passes a proc pointer

hack to try to avoid improperly detected recursive locks

need to pass proc to vinvalbuf

accounting for simple locks and unlocks

track number of simple locks

convert VOP_UNLOCK and vrele into vput's; add proc parameter to union_dircache

delete reference if vget lock attempt fails

if mountroot != NULL, then just call it

add back trace to simple lock diagnostic information (if available)

accounting for locks and unlocks;
give file and line in simple lock error diagnostics

get rid of attempt to detect hanging