Enhance soft-fp for TFmode<->[DI,SI,TI,XF]mode conversions

[Jakub Jelinek]

On Sun, Apr 22, 2007 at 09:21:01PM +0200, Uros Bizjak wrote:
> This patch enhances soft-fp for TFmode<->[DI,SI,TI,XF]mode conversions. 
> It was tested for x86_64 target with gcc testsuite.

Thanks, I have committed the changes to existing files (with one additional
fix in FP_TRUNC) plus a testcase that showed on sparc some errors
that are fixed by this commit.

>         * soft-fp/floattisf.c: New file.
>         * soft-fp/floattidf.c: New file.
>         * soft-fp/floattitf.c: New file.
>         * soft-fp/floatuntisf.c: New file.
>         * soft-fp/floatuntidf.c: New file.
>         * soft-fp/floatuntitf.c: New file.
>         * soft-fp/fixsfti.c: New file.
>         * soft-fp/fixdfti.c: New file.
>         * soft-fp/fixtfti.c: New file.
>         * soft-fp/fixunssfti.c: New file.
>         * soft-fp/fixunsdfti.c: New file.
>         * soft-fp/fixunstfti.c: New file.
>         * soft-fp/extendxftf.c: New file.
>         * soft-fp/trunctfxf.c: New file.

But I don't think we should keep these files in glibc, nothing in glibc
uses them and glibc will unlikely contain 128-bit integer support
(nor need TF<->XF mode conversion).  We don't have soft-fp/addxf3.c etc.
either, this kind of sources IMHO should be maintained wherever they
are actually used.

	Jakub