- do not record properties on irgs anymore, always do it on the irgs
  entity; entity properties have to be a superset of the entities method
  type properties.
- Remove special irg_inline_property and use mtp_additional_properties
  instead

Also removes some debugging output to simplify the code

This also fixes backend/sign_bit.c

When hard-coding the *, invalid assembler gets emitted if an immediate ends up as ia32_IJmp operand.

This fixes gcc/pr37285.c.

Added some additional functions that allow direct access to some fields
of the backend params struct.  These functions avoid struct type,
thereby making interfacing with the jFirm library easier.

Reusing the same dummy struct as a hidden parameter for multiple
calls is not always correct (see opt/structcall2.c for an example).
For now, just use multiple dummy structs.

We can't handle 128bit long double in softfloat yet, but most libraries
are compiled like this. It's better to have the compiler fail than to
produce code with a mismatching ABI in the default setting.

Conflicts:
	ir/be/beverify.c

This reverts commit 2a018b81.

Unreachable loops always have a loop entry edge and thus a Phi node.

Both mature and immature blocks were dumped in ird_color_block_background.

In particular this allows ifconv when a PhiM is present, but nothing happens with the memory on the two paths.

This is a workaround for keepalive edges which are probably broken
(adding the stuff to the End node is wrong, it should get added to the
 endless loop somehow)

SymConst_addr_ent are never zero for example, also there's no need to
give up after seeing a confirm

I aligning them to a multiple of 100 now, so that when code generated in
an earlier functions only changes lightly it is very likely that the
block numbers in later functions stay the same.