checkpoint

checkpoint, but not because it's working

checkpoint before lfs_iupdat -> ITIMES; reads almost work...

check ufs lock count on exit from syscalls

more precise comments

add proc pointers to appropriate vnode operations

first cut for new proc & user structs (still need to put in new vnode calling
convention): get credentials, etc, via proc; u.u_nd, u.u_spare are gone;
purge user.h; remove ../ufs/ from includes; cons

first cut for new proc & user structs (still need to put in new vnode calling
convention): get credentials, etc, via proc; u.u_nd, u.u_spare are gone;
purge user.h; remove ../ufs/ from includes; consolidate *nullops();
chown/fchmod use proc's credentials, not file's

show more ...

add POSIX-style byte-level record locking

update for new VM

brealloc => allocbuf

reincarnated clear of i_rdev when freeing an inode, otherwise it
appears inode has zero size but one disk block. (bostic+trent)

new copyright notice

mount struct prefixes go from m_ to mnt_ and M_ to MNT_

have to take off the hash list when bread of inode fails

iaccess moves into ufs_access; changes for the new quota implementation

reassign vnode ops for FIFO's

optimization from Noemi Paciorek <noemi@osf.org>

initialize i_diroff; proper association of indirect blocks

i_lastr => v_lastr; v_count => v_usecount

debugging, put in accounting for process id holding and waiting for lock

have to update the correct inode (from mike@cs.utah.edu)

overhaul itrunc; cleanups to inactive; use vprintf

checkalias now sets rdev

syncip takes a waitfor flag

suppress printing of pushes of active inodes