Various cleanup:

 - Got rid of dummy proc[0].  Now proc[0] is init.
 - Added initcode.S to exec /init, so that /init is
   just a regular binary.
 - Moved exec out of sysfile to exec.c
 - Moved code dealing with fs guts (like struct inode)
   from sysfile.c to fs.c.  Code dealing with system call
   arguments stays in sysfile.c
 - Refactored directory routines in fs.c; should be simpler.
 - Changed iget to return *unlocked* inode structure.
   This solves the lookup-then-use race in namei
   without introducing deadlocks.
   It also enabled getting rid of the dummy proc[0].

Gcc expects to be able to pick up the return
address off the stack, so put one there for it.
(Bug was hidden by bad segment limits.)

Because limit gives the address of the last addressable
byte in the segment, the limit argument to SEG16 and SEG
needs to have 1 subtracted from it.

Cannot use "g" as an arg constraint because
some memory references aren't actually valid
and gcc is a bit too smart in parsing (%0) and
a bit too dumb in knowing which memory refs
are not valid.  Using "r" seems to constrain gcc
to use a register as in (%eax) and not substitute
a memory reference like (-36(%ebp)).

No one really understands these things.