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

[Mark Mitchell <mark at codesourcery dot com>]

Jakub Jelinek wrote:

> Since today glibc uses
> $(sort $(foreach n,40 41 50 51 60 61 62 63 \
>                                       $(foreach m,2 3 4 5 6 7 8 9, \
>                                                   3$m 4$m 5$m),\
>                                     -ffixed-$n)) \
>                    $(sort $(foreach n,$(foreach m,0 1 2 3 4 5 6 7 8 9,\
>                                                 $m 1$m 2$m) 30 31,\
>                                     -ffixed-v$n)) \
>                    -ffixed-vrsave -ffixed-vscr
> instead of -msoft-float where it wants to avoid any modification
> to FP and Altivec registers (apparently -ffixed-f0 etc. doesn't work
> on ppc*), so from glibc POV it really doesn't matter what sysv4.h
> does if TARGET_SOFT_FLOAT.

So, if I understand correctly, there's no reason not to make it an error
to combinE -msoft-float with TARGET_LONG_DOUBLE_128, whether implicitly
or explicitly.  Is that correct?  If so, David, would you object to that
change?

Thanks,

-- 
Mark Mitchell
CodeSourcery
mark@codesourcery.com
(650) 331-3385 x713