This is the mail archive of the
mailing list for the GCC project.
Re: RESUBMIT: PowerPC -- support single-precision FPU
- From: "Joseph S. Myers" <joseph at codesourcery dot com>
- To: Michael Eager <eager at eagercon dot com>
- Cc: janis187 at us dot ibm dot com, Andreas Tobler <andreast-list at fgznet dot ch>, David Edelsohn <dje dot gcc at gmail dot com>, GCC Patches <gcc-patches at gcc dot gnu dot org>, Andrew Pinski <pinskia at gmail dot com>
- Date: Thu, 2 Oct 2008 21:28:52 +0000 (UTC)
- Subject: Re: RESUBMIT: PowerPC -- support single-precision FPU
- References: <firstname.lastname@example.org> <48D91B67.email@example.com> <48DCFDA7.firstname.lastname@example.org> <1222449719.5930.19.camel@janis-laptop> <48DD3108.email@example.com> <48E10AE2.firstname.lastname@example.org> <email@example.com> <48E50FBB.firstname.lastname@example.org> <48E51A16.email@example.com> <1222978334.6065.17.camel@janis-laptop> <48E52EEA.firstname.lastname@example.org>
On Thu, 2 Oct 2008, Michael Eager wrote:
> Janis Johnson wrote:
> > On Thu, 2008-10-02 at 11:59 -0700, Michael Eager wrote:
> > > Other tests have explicit checks for TFmode, which are missing.
> > > This would probably be better:
> > Decimal float modes SDmode, DDmode, and TDmode also use
> > floating point registers.
> Perhaps we need a macro which tests for 32-bit modes (SFmode
> and SDmode), 64-bit modes (DFmode and DDmode), and a third
> for 128-bit modes (TFmode and TDmode).
Then you have TFmode in the TARGET_IEEEQUAD case, passed by reference -
but I believe that case is long bit-rotten, with lots of TFmode references
that don't try to check for TARGET_IEEEQUAD, and could safely be removed.
Joseph S. Myers