Re: PATCH for 64-bit MIPS ELF buglets

[Mark Mitchell <mark@codesourcery.com>]

>>>>> "Ian" == Ian Lance Taylor <ian@zembu.com> writes:

    Ian> I'm afraid that there isn't any documentation on those
    Ian> relocations.  I just wrote them so that they worked.  I can
    Ian> probably answer any questions you have.

OK, I'll try to make this easy on myself. :-)

Can you describe, in the same sort of way used in the "official" MIPS
ABI documentation what R_MIPS16_26 and R_MIPS16_GPREL do?

In other words, what bits are relocated, and what calculation is
performed to calculate their values?  Once I know that, it should be
easy to do the right thing in mips_elf_calculate_relocation (and, if
truly odd things happen, in mips_elf_perform_relocation).

I have to admit that I find the code that was there before pretty much
impenetrable.  That goes for the 16/32-bit stub stuff as well.  I
understand what the point is (16-bit code and 32-bit code needs stubs
when calling each other), but I don't understand what the code is
doing.

Bottom line: the best way to get this fixed is going to be if you can
give me a complete roadmap.  Otherwise, I'm going to be bumbling about
in the dark, which will probably lots of both of our time, while we
iterate until we get it right.

Thanks,

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com