2 _lock call can return one of
3 0 - all required reservations successfully made
4 -ENOSPC - filesystem is too full. This can only apply
5 when trying to lock a block was didn't
7 1 - No free space, but this is an over-write so
8 we may not fail. A checkpoint has been requested,
9 wait for it to complete.
11 These get called under the checkpoint lock.
12 If a '1' is returned, then the checkpoint should be unlocked
13 with lafs_checkpoint_unlock_wait(fs) which will unlock and wait
14 for the next checkpoint to fail. It will BUG if a checkpoint
15 had not been requested.