This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
RE: [PATCH RFA MIPS] Prohibit vector modes in accumulators
- From: "Moore, Catherine" <Catherine_Moore at mentor dot com>
- To: Matthew Fortune <Matthew dot Fortune at imgtec dot com>, Robert Suchanek <Robert dot Suchanek at imgtec dot com>, "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>
- Date: Fri, 23 Jan 2015 21:19:52 +0000
- Subject: RE: [PATCH RFA MIPS] Prohibit vector modes in accumulators
- Authentication-results: sourceware.org; auth=none
- References: <B5E67142681B53468FAF6B7C3135656244118FCB at hhmail02 dot hh dot imgtec dot org> <6D39441BF12EF246A7ABCE6654B0235320FB27F5 at LEMAIL01 dot le dot imgtec dot org>
> -----Original Message-----
> From: Matthew Fortune [mailto:Matthew.Fortune@imgtec.com]
> Sent: Friday, January 23, 2015 2:51 PM
> To: Robert Suchanek; gcc-patches@gcc.gnu.org
> Cc: Moore, Catherine
> Subject: RE: [PATCH RFA MIPS] Prohibit vector modes in accumulators
>
> > 2015-01-23 Robert Suchanek <robert.suchanek@imgtec.com>
> >
> > * config/mips/mips.c (mips_hard_regno_mode_ok_p): Prohibit
> > accumulators
> > for all vector modes.
>
> This seems like a genuine bug and although it can only be triggered by
> loongson or paired-single support it probably qualifies for fixing.
> My suspicion is that the switch to LRA since GCC 4.9 may be the reason this
> hasn't been noticed before. Reload seemed better in some cases at
> eliminating bad decisions from IRA so this may have simply never made it
> through reload by fluke.
>
> I'd like Catherine to review too since we are in stage4 without a reproducible
> test case.
>
The patch looks reasonable, but I'd like to see a test case that fails before we agree to include for GCC 5.0.