Re: Fix linking of .gpdword on MIPS

[Richard Sandiford <rsandifo at nildram dot co dot uk>]

Daniel Jacobowitz <drow@false.org> writes:
> On Mon, Feb 26, 2007 at 10:47:28PM +0000, Richard Sandiford wrote:
>> Are we really confident that this is the only exception, and that true
>> relocations against symbol zero will not reach here?  If so, I think
>> the patch below is the correct fix.  If not, it seems dangerous to
>> explicitly ignore what I understand to be correct semantics,
>> and treat S + A as zero simply because S is zero.
>
> It's the right thing to do if the comment accurately described the
> check.

By "true relocation against symbol zero", I meant one that was in
an input file.  That's technically possible, isn't it?  (If ELF lets
you do something, it seems something somewhere will.)

Anyway, it's moot now.  I was wanting to move away from an r_symndx
check, and that's exactly what HJ's patch does, so I'm happy.
And sorry for missing HJ's patch and posting one of my own.  I've a
huge backlog of binutils mail that I must go through one of these days.
I'd searched the subjects for MIPS, and that was about all.

> I was as thorough as I could be checking my assumptions, but obviously
> not thorough enough.  Does HJ's patch help?

I'll give it a go.  I'm not sure I can help with:

    http://sourceware.org/ml/binutils/2007-02/msg00110.html

though, which seems to have been the last word on the patch.

Richard