The opcode list was a member of irprog before which wasn't really
handled consistently. Also make sure opcodes are properly freed at
ir_finish().

This should fix problems with uninitialized backends when
be_get_backend_param() wasn't called also calls be_lower_for_target if
the user hasn't done so yet.

get_backend_irg_list - none of our backends uses it and in the future
                       the backend should drive the compilation and not
                       give hints to a generic driver
reg_class_alignment  - unused

This also means that dependent modules like bedbgout should be
initialized by begnuas.

So we are not forced to use the generic be_Spill, be_Reload functions
which don't work correctly with double-width register requirements.
This should really fix sparc float spill/reload this time.

If req2 was limited and req1 wasn't, then it incorrectly returned true.

This tries to get the names in a consistent format. We basically have 2
views on register allocation/constraints now:

1) Register Requirements and Assigments are per-instruction. Each
   instruction has requirements on its inputs and outputs, register get
   assigned for outputs of an instruction (assignment is an attribute of
   an instruction, not of the Proj-nodes).
   The internal datastructures model this form!

   The functions here have the form:
     arch_get_irn_register_req_in(node, input_nr)
     arch_get_irn_register_in(node, input_nr)
     arch_get_irn_register_req_out(node, output_nr)
     arch_set_irn_register_out(node, output_nr, register)

2) Register Requirements and Assignments are on firm-values. This view
   allows to query/assign requirements and registers at the Proj nodes
   instead of the repsective predecessors.
   This is a convenience interface modeled on top of the other!

   The functions have the form:
     arch_get_irn_register_req(node)
     arch_get_irn_register(node)
     arch_set_irn_register(node, register)

The callee/caller saved information is not constant accross different
ABI's, so don't make it constant. Instead, all BE that still use beabi
provide a callback now.
This allows to implement support for x64_64/Win32 and is a necessary step
for the combined x86 BE.

[r28272]

Only makes it hard to search for arch_register_type_* flags.

[r28185]

[r28172]

[r28145]

[r28127]

[r28071]

[r28039]

[r28038]

[r28010]

eliminate the unnecessary and especially confusing concept of a 'code_generator' an isa-interface is enough

[r28009]

[r28008]

[r27995]