Re: RFC: using ld 2.14 on irix

[Richard Sandiford <rsandifo at redhat dot com>]

Ralf Baechle <ralf@linux-mips.org> writes:
> On Sun, Jun 15, 2003 at 10:37:35AM -0400, Daniel Jacobowitz wrote:
> 
> > The last time I spoke to you about this, I got the impression that the
> > only problem you'd found was the one we solved by switching relocs
> > against local symbols to be against symbol 0.  Now are you saying that
> > you found this problem with global symbols too?
> 
> Just to make the scenario clear, this is the testcase I had extracted and
> which essentially is the same as Richard Sandiford's gromit.c.
> 
> [ralf@dea tmp]$ cat s.s
>         .data
> bar:    .word   bar + 0x1234

Hmm, I'm not sure whether we're talking about the same thing.  This test
case seems to be the same as the one in the message that Maciej pointed
to earlier.  That is, "bar" is local to s.o.  I thought Alex's patch
(which is in 2.14) fixes that case.  And FWIW, I haven't seen it cause
any problems on irix.

The problems I'm seeing are with global symbols.  I.e. with:

        .data
        .globl  bar
bar:    .word   bar + 0x1234

or with:

        .data
foo:    .word   bar + 0x1234

That's what gromit.c was testing.  "y" was initially pointing
to a global symbol "x".

Richard