This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH, x86_64]: Use soft-fp for TFmode (128bit) FP.
- From: "Joseph S. Myers" <joseph at codesourcery dot com>
- To: Uros Bizjak <ubizjak at gmail dot com>
- Cc: GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: Thu, 12 Apr 2007 20:48:17 +0000 (UTC)
- Subject: Re: [PATCH, x86_64]: Use soft-fp for TFmode (128bit) FP.
- References: <461E96A4.4090407@gmail.com>
On Thu, 12 Apr 2007, Uros Bizjak wrote:
> This patch implements TFmode (128bit) floating point by using soft-fp for
> x86_64 targets. To support TImode FP-int conversions, a bunch of support files
> is added to soft-fp directory (+ a couple of small bugfixes, as can be seen in
> attached soft-fp diff).
The changes to existing soft-fp files certainly need to go in FSF glibc
first.
Whether the new files need to go there is less clear; however, I would
presume that they do need to do so unless RMS says otherwise. They
wouldn't be used in glibc by any existing glibc configuration, but since
the code is shared with GCC this should not be fatal to their inclusion in
glibc, and they could be useful in glibc for MIPS64 soft-float in future.
Is there any particular reason you didn't include conversions to/from
XFmode (apart from my notes at
<http://gcc.gnu.org/wiki/Software_floating_point> that you'd need some
changes to the core soft-fp conversion macros to handle conversions
between types with same normal exponent range but different mantissa
size)? What happens with such conversions with this patch?
If possible I'd say you should define FP_INIT_ROUNDMODE and
FP_HANDLE_EXCEPTIONS in sfp-machine.h to use the hardware rounding mode
and exceptions so that this type works like the other floating-point
types.
--
Joseph S. Myers
joseph@codesourcery.com