This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: Why does mips define elf_backend_sign_extend_vma to true?
hjl@lucon.org ("H . J . Lu") writes:
> [ ... address sign extension ... ]
> Show me where it is documemnted in the 32bit SVR4 MIPS ABI.
Show me where the 32bit SVR4 MIPS ABI says _anything_ about 64-bit
addressing. 8-)
more seriously, look in the MIPS64 ISA documentation off the
www.mips.com publications page. you'll find:
* for addressing, 32-bit compatibility mode requies addr[63:32]
matching addr[31]. if addr[63:32] doesn't satisfy that constraint,
it's a full 64-bit addr (and not valid unless 64-bit addressing is
enabled).
* for most 32-bit operations on values in registers, you have to have
(on input) gpr[63:32] matching gpr[31] else you get
implementation-defined behaviour. (similarly, 32-bit operations
produce sign-extended results.)
cgd