This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Another patch to the mips64vr-elf config.


Eric Christopher <echristo@redhat.com> writes:
> This really won't do either if someone picks mips64vr-linux-gnu...

That isn't a support config though.  I doubt it would use vr.h anyway.
All vr.h really does is make vr4100 the default architecture (which you
could do just as easily in config.gcc) and set up stuff for the 6-ABI
multilibs.  The multilibs for mips64vr-linux-gnu wouldn't be the same
as the *-elf ones.

I guess an alternative would be to change the original patch to
something like:

  if ((target_flags_explicit & MASK_SOFT_FLOAT) == 0
      && MIPS_MARCH_CONTROLS_SOFT_FLOAT)
    {
      ...
    }

where MIPS_MARCH_CONTROLS_SOFT_FLOAT is 0 by default but is set to
1 for some targets.  I don't like the idea of turning it on for
all *-elf targets, though.  If it isn't general enough to be
applied for all configurations (and I agree it isn't, sorry for not
realising earlier) then I'd rather just enable it for mips64vr-elf.

Would the MIPS_MARCH_CONTROLS_SOFT_FLOAT thing be OK?

Richard


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]