[patch, binutils] Patch elf/mips.h for -mfp64 support.

[Joseph S. Myers]

On Mon, 9 Sep 2013, Doug Gilmore wrote:

> We would like to enable Linux processes to use 64-bit FP registers (as
> opposed to using register pairs for DP values) on 32-bit MIPS
> processors.

Will you be posting corresponding patches to glibc (setjmp / longjmp / 
*context functions, at least; preferably also dynamic linker sanity checks 
using this new ELF flag to ensure that classic hard-float o32 binaries 
don't use -mfp64 o32 shared libraries, and vice versa) and GDB/gdbserver 
(both register handling and function call/return support for the new ABI 
if not already present)?  What about an ABI supplement explaining how 
-mfp64 affects the hard-float o32 function call / return ABI, if there 
isn't already such a document?

-- 
Joseph S. Myers
joseph@codesourcery.com