This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: i386-linux cross sh-elf breakage
- To: Clinton Popetz <cpopetz at cygnus dot com>
- Subject: Re: i386-linux cross sh-elf breakage
- From: Richard Henderson <rth at cygnus dot com>
- Date: Fri, 19 May 2000 17:41:08 -0700
- Cc: Toshiyasu Morita <tm at netcom dot com>, gcc-patches at gcc dot gnu dot org
- References: <200005152231.PAA01084@netcom.com> <20000516140146.A17877@cpopetz.com> <20000516175851.C7380@cygnus.com> <20000516212713.B4299@cpopetz.com> <20000518204943.B11020@cygnus.com> <20000519162632.A3003@cpopetz.com>
On Fri, May 19, 2000 at 04:26:32PM -0500, Clinton Popetz wrote:
> (optimize_mode_switching): Use MODE_AT_EXIT instead of
> MODE_USES_IN_EXIT_BLOCK. Change leftover instance of
> FP_MODE_NONE to no_mode.
I don't see the need for MODE_AT_EXIT. It's always going
to be the same as MODE_AT_ENTRY. In any sane system anyway.
You could rename it if you like to be more clear on that fact;
certainly that's not a large change at this point.
> + #ifdef MODE_AT_EXIT
> + if (bb == (n_basic_blocks - 1) && MODE_AT_EXIT (e) != no_mode
> + && last_mode != MODE_AT_EXIT (e))
Please don't assume (n_basic_blocks - 1). Please follow the CFG
edges back from EXIT. It is rare, but not impossible, for a
target to have return insns at this stage.
r~