"change_to_user" in process_blocking_lock_queue()

[Jeremy Allison]

On Tue, Nov 04, 2008 at 05:40:48PM +0100, Volker Lendecke wrote:
> Hi, Jeremy!
> 
> Right now I wonder why we do the change_to_user() in
> process_blocking_lock_queue(). As far as I can see we only
> do operations on already opened file descriptors or the
> brlock.tdb which also is already open. Neither in blocking.c
> nor in locking/*.c we look at current_user, so a user-level
> access check is also not done.
> 
> Do you remember why that change_to_user() is there? Am I
> missing something?

No, it's probably not needed. Added when I originally
added the blocking lock code to make sure it kept the
original user context from the first processing of
the SMB. Almost certainly paranoia inherited from
the first implementation. Junk it :-).

Thanks for looking closely :-).

Jeremy.