[PATCH] Add support for MIPS r3 and r5

[Andrew Bennett]

> -----Original Message-----
> From: Richard Sandiford [mailto:rdsandiford@googlemail.com]
> Sent: 08 May 2014 20:28
> To: Andrew Bennett
> Cc: gcc-patches@gcc.gnu.org; Matthew Fortune; Saeed Ghazanfar; Rich Fuhler
> Subject: Re: [PATCH] Add support for MIPS r3 and r5
> 
> Andrew Bennett <Andrew.Bennett@imgtec.com> writes:
> > diff --git a/gcc/config/mips/mips-cpus.def b/gcc/config/mips/mips-cpus.def
> > index 07fbf9c..f2e23c6 100644
> > --- a/gcc/config/mips/mips-cpus.def
> > +++ b/gcc/config/mips/mips-cpus.def
> > @@ -44,9 +44,13 @@ MIPS_CPU ("mips4", PROCESSOR_R8000, 4, 0)
> >     isn't tuned to a specific processor.  */
> >  MIPS_CPU ("mips32", PROCESSOR_4KC, 32, PTF_AVOID_BRANCHLIKELY)
> >  MIPS_CPU ("mips32r2", PROCESSOR_74KF2_1, 33, PTF_AVOID_BRANCHLIKELY)
> > +MIPS_CPU ("mips32r3", PROCESSOR_M4K, 34, PTF_AVOID_BRANCHLIKELY)
> > +MIPS_CPU ("mips32r5", PROCESSOR_74KF2_1, 36, PTF_AVOID_BRANCHLIKELY)
> 
> Looks odd for mips32r2 and mips32r5 to have the same processor tuning
> but mips32r3 to be different.  I assume 74KF2_1 is just a reasonable
> default, given the lack of tuning for a real r5 CPU?  That's fine if so,
> but probably deserves a comment.

mips32r3 is basically micromips, so I have just used the same processor 
definition that the micromips entry uses.  On the processor for mips32r5
we should probably think of adding a processor definition for the P5600, 
but using 74KF2_1 for the moment will be fine (and I will add a comment 
to explain this).
 
> > @@ -141,7 +151,8 @@ along with GCC; see the file COPYING3.  If not see
> >    "%{EL:-m elf32lmip} \
> >     %{EB:-m elf32bmip} \
> >     %(endian_spec) \
> > -   %{G*} %{mips1} %{mips2} %{mips3} %{mips4} %{mips32} %{mips32r2}
> %{mips64} \
> > +   %{G*} %{mips1} %{mips2} %{mips3} %{mips4} %{mips32} %{mips32r2}
> \
> > +   %{mips32r3} %{mips32r5} %{mips64} \
> >     %(netbsd_link_spec)"
> >
> >  #define NETBSD_ENTRY_POINT "__start"
> 
> Not sure the omission of mips64r2 was deliberate here, or in vxworks.h.

Yes it was deliberate: mips64r2 was already missing from the original ASM_SPEC 
in vxworks.h and the LINK_SPEC in netbsfd.h.  

> 
> As Joseph said, the .po stuff should be left alone.  The .pot file is
> regenerated near to a release so that the translators can update the
> .po files.

Thats fine, (and many thanks for both of you for clarifying this).  I will 
remove the changes from the patch.

> Looks good otherwise, thanks.

I will prepare the revised patch, and will commit it once the copyright 
assignment issues have been resolved.
 

Many thanks,


Andrew