RE: [PATCH, MIPS] Scheduling for M51xx core family

[Robert Suchanek <Robert.Suchanek@imgtec.com>]

Hi Matthew,

> > gcc/
> >
> > 	* config/mips/m5100.md: New file.
> > 	* config/mips/mips-cpus.def (m5100, m5101): Define.
> > 	* config/mips/mips-tables.opt: Regenerate.
> > 	* config/mips/mips.c (mips_rtx_cost_data): Add costs for m5100.
> > 	* config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Map -march=m5100 and
> > 	-march=m5101 to -mips32r5.
> > 	(MIPS_ARCH_FLOAT_SPEC): Map -m5101 to -msoft-float.
> > 	(MIPS_ISA_NAN2008_SPEC): Map -march=m51* to -mnan=2008 if
> > 	!-msoft-float.
> > 	* config/mips/mips.md: Include m5100.md.
> > 	(processor): Add m5100.
> > 	* doc/invoke.texi (-march=@var{arch}): Add m5100, m5101.
> 
> OK, this looks fine.

The patch committed as r226065.

> I did realise while reading through this that the MIPS_ARCH_FLOAT_SPEC
> is not used for and ordinary MIPS Linux compiler which seems odd but
> I presume this is to make it possible to use one hard-float sysroot
> for any core and emulate the FPU when not present.
> 
> I think it is probably a mistake to have put MIPS_ARCH_FLOAT_SPEC in
> the mti-linux.h and android.h DRIVER_SELF_SPECS so I think they need
> removing. Although we support building soft-float multilibs I don't
> think they actually get used very much so leaving the selection of
> soft-float down to the end user in Linux seems wise.
> 
> With the i6400 scheduler committed then we can also get rid of the w32
> and w64 placeholders that were there solely to provide an R6 processor
> to use as the default processor for the generic arch options.

I'll send a follow-up patch to remove the redundant bits.

Regards,
Robert