[PATCH] Allow -msoft-float -mlong-double-128 without a warning

[David Edelsohn]

>>>>> Mark Mitchell writes:

Mark> I'm sympathetic to the fact that we're not issuing an error, in view of
Mark> the fact that we don't have -mno-implicit-fp, and, thus issuing an error
Mark> would prevent compliation of GLIBC, but why shouldn't we issue a
Mark> warning, even when TARGET_LONG_DOUBLE_128 is the default?

	I would be happy with always issuing a warning and not implicitly
changing long_double_type_size to 64.

David