This is the mail archive of the
mailing list for the GCC project.
Re: [patch] rtl.h: Include insn-config.h
- From: Kazu Hirata <kazu at cs dot umass dot edu>
- To: zack at codesourcery dot com
- Cc: roger at www dot eyesopen dot com, gcc-patches at gcc dot gnu dot org
- Date: Sun, 02 Feb 2003 22:48:29 -0500 (EST)
- Subject: Re: [patch] rtl.h: Include insn-config.h
- References: <Pine.LNX.firstname.lastname@example.org><email@example.com>
> I have an alternate suggestion: simply use the definition
> #define CC0_P(X) ((X) == cc0_rtx)
> always. If you look at emit-rtl.c (init_emit_once), cc0_rtx is always
> available, even on non-cc0 targets. On those targets it should still
> be found always to be false.
If people are happy with your suggestion, I am certainly OK with it.
I thought people might want to prevent even tiny slow down that comes
from always defining CC0_P like "#define CC0_P(X) ((X) == cc0_rtx)",
but I guess the difference is too small to see.