S/390 port modernization patches

[Ulrich Weigand <weigand@i1.informatik.uni-erlangen.de>]

Hello,

the S/390 backend has unfortunately fallen a bit behind recent
gdb development.  However, I've finally finished a (hopefully)
thorough modernisation of the port that should bring it up to
date.  I'll try to keep in closer contact with ongoing gdb
development in the future to avoid this happening again ...

Thanks to everyone who has helped keeping the s390 port working,
in particular to Jim Blandy!

[ B.t.w. the legal situation as to IBM copyright assignments
  is still unchanged - we need separate 'Software Letters'
  for each code contribution.  I'm posting the patches to
  this list for review; if and when they are approved, I'll
  get a new Software Letter for them sent. ]


This set of patches does:

- eliminate every use of deprecated interfaces and features
  in the s390 backend; in particular implements all changes
  mentioned in
   http://sources.redhat.com/ml/gdb/2003-04/msg00279.html

- simplify the build process a bit (merges s390 and s390x
  into a single configure.tgt target, removes tm-s390.h)

- perform various coding style and other cleanups

- activate DWARF-2 CFI based frame handling on s390; this needs
  some changes to common code support

- activate bi-arch mode (debugging 32-bit applications with
  a 64-bit gdb)

- fix various bugs, some in common code

In total, test suite results improve from 121 to 25 unexpected
failures on s390, and from 176 to 28 unexpected failures on s390x.


I've split the changes into 8 separate patches; while each builds
upon the previous ones (usually), they should be self-contained.
 
The patches are organized like this:

 1: modernization 1/4: register handling and build process
 2: modernization 2/4: watch point handling
 3: modernization 3/4: ABI related (inferior calls, return values)
 4: modernization 4/4: new-style frame handling

 5: DWARF-2 CFI based frame handling (incl. common code)
 6: activate bi-arch mode

 7: fix some 64-bit Objective-C bugs
 8: fix some Java test case problems

(The last two are not s390 related; they just fell out from
trying to fix as many test suite failures as possible ...)
 

Bye,
Ulrich

-- 
  Dr. Ulrich Weigand
  weigand@informatik.uni-erlangen.de