- Fix SMP f_list corruption problem.

   This was rather subtle.  It turns out that gcc was doing a small
   amount of reordering around the file_lock because it doesn't see our
   spinlock implementation as being a barrier.  To fix this I

   - Added the appropriate barriers to all the spinlocks.
   - Removed the atomic opencoded spinlock and redid it as a proper one.

   SMP now seems stable on a 2xA500 and has survived a 10 hour 35 loop
   make -j 4 kernel compile without showing any problems (previously, it
   usually fell over in the first loop).

 - Since we have one or two users who don't want SMP, make the
   code base compile for UP again ...

 - missing parenthesis for pointer dereferencing
Committed-by: James Bottomley <jejb@parisc-linux.org>
Committed-by: Thibaut Varene <varenet@parisc-linux.org>