This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: zero vector usagee (Was: Re: RFA: gcc.dg/simd-1.c fix for x86)
- From: Aldy Hernandez <aldyh at redhat dot com>
- To: Joern Rennecke <joern dot rennecke at superh dot com>
- Cc: Daniel Egger <degger at fhm dot edu>, GCC Patches <gcc-patches at gcc dot gnu dot org>, gcc at gcc dot gnu dot org
- Date: Mon, 8 Jul 2002 10:32:10 -0700
- Subject: Re: zero vector usagee (Was: Re: RFA: gcc.dg/simd-1.c fix for x86)
- References: <3D287F6F.E25C9323@superh.com> <1026066347.6648.2.camel@sonja.de.interearth.com> <3D296C3F.83381D5C@superh.com> <m2y9cmw6l2.fsf@culebra.cygnus.com> <3D29C2EA.E79EBA1@superh.com>
> Well, so is it the responsibility of gen_rtx_CONST_VECTOR
> to test for zero, or does every caller have to do that?
hmm, this is all new.
before your patches, the only place where we generated CONST_VECTORs
was in store_constructor() by merit of calling clear_storage(), so we only
generated zeros.
if we are now generating other constants, then we probably need a wrapper
function that generates CONST0_RTX when all elements are zero.
cheers
aldy