Re: Old PPC patch - 405F, 440 support

[David Edelsohn <dje at watson dot ibm dot com>]

> At this time I am looking for comments, not
> approval - anything from "it's not necessary to recognize these as
> separate CPUs" to "you need to patch here, here, and here too" would
> be helpful.

	1) Why is the "405" initializer in rs6000_override_options changed
with incorrect indentation, and adding "405f" with the same incorrect
indentation?

	2) The ASM_CPU_SPEC for mcpu=405f presumably should be -m405, not
-mppc.

	3) The ppc405f has the same scheduling characteristics as the
PPC405, so there is no reason to add a PROCESSOR_405F in the processor
type enum.  It is even more useless when that processor type is not
referenced in rs6000_override_options initializer.

	4) The patch adds ppc405f option, but does not define the floating
point scheduling information for the ppc405f.

	5) What is the point of adding a ppc440 option most of the
defintions behave exactly like the ppc405 option (and bizarrely refers to
PROCESSOR_PPC403 in rs6000_override_options)?  The author of the patch
obviously does not know anything about the scheduling characteristics of
the ppc440.


	Other than providing a ppc405fp entry in the
processor_target_table for a ppc405 with floating point support, the rest
of the patch looks like someone creating dummy entries without any
information so they can say "gcc -mcpu=440" and feel good.

David