This is the mail archive of the binutils@sources.redhat.com mailing list for the binutils 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: how close to binutils-2.12 previews


Maciej W. Rozycki wrote:
[snip]
> > It's valid to use 64bit opcodes in 32bit object format. Warnings
> > would be misleading in this case. I had this code in my private
> > tree for a while but didn't contribute it due to testsuite failures
> > caused by it.
> 
>  But you can't express all relocations that are needed by "dla".

Dream a moment of n32 ABI, where the relocations are available. :-)

> Also what would you need 64-bit addressing for in a 32-bit object?

Current MIPS64/Linux Kernels are exactly that.

> None of
> sections can lie outside the 32-bit address space, so you'll never get a
> symbol that needs "dla" to take address of (IOW, a "dla" will always do
> the same that a "la" but with more instructions).

What if you have an ARCS64 firmware bootvector outside the 32bit space...

> > >  #ifdef BFD64
> > >    mips64*el-*-linux*)
> > > -    targ_defvec=bfd_elf32_tradlittlemips_vec
> > > -    targ_selvecs="bfd_elf32_tradbigmips_vec bfd_elf64_tradlittlemips_vec bfd_elf64_tradbigmips_vec"
> > > +    targ_defvec=bfd_elf64_tradlittlemips_vec
> > > +    targ_selvecs="bfd_elf64_tradbigmips_vec bfd_elf32_tradlittlemips_vec bfd_elf32_tradbigmips_vec"
> > >      ;;
> > >    mips64*-*-linux*)
> > > -    targ_defvec=bfd_elf32_tradbigmips_vec
> > > -    targ_selvecs="bfd_elf32_tradlittlemips_vec bfd_elf64_tradbigmips_vec bfd_elf64_tradlittlemips_vec"
> > > +    targ_defvec=bfd_elf64_tradbigmips_vec
> > > +    targ_selvecs="bfd_elf64_tradlittlemips_vec bfd_elf32_tradbigmips_vec bfd_elf32_tradlittlemips_vec"
> > >      ;;
> > >  #endif
> > 
> > Default for mips64 is to generate 32bit code, so targ_defvec
> > shouldn't be changed like this.
> 
>  This being because the 64-bit userland is not ready?  OK -- this may be
> deferred until the default changes.

Even the native SGI toolchain uses n32 as default: Better performance
and still enough address space for most applications. Btw, I intend
to change the default to a newly invented n32 vector RSN.


Thiemo


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