fcntl F_SETLKW64 failing on Solaris

[Jeremy Allison]

On Wed, Jan 16, 2002 at 01:33:03PM -0000, Mayers, Philip J wrote:
> Interesting, I did notice that constant - I thought that the tdb traversal
> of several (e.g. the locking database) was intentional. Won't this cause
> dead PID TDB records to build up?

No, the code was changed to clean dead records on a dev/inode
pair on every fetch of that dev/inode record.

Jeremy.