This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: weak-dyn test fails on PowerPC
- From: "H.J. Lu" <hjl at lucon dot org>
- To: binutils at sourceware dot org
- Date: Thu, 30 Aug 2007 06:17:27 -0700
- Subject: Re: weak-dyn test fails on PowerPC
- References: <20070829213015.GA18101@caradoc.them.org>
On Wed, Aug 29, 2007 at 05:30:15PM -0400, Daniel Jacobowitz wrote:
> The test has:
>
> SECTIONS
> {
> . = 0x800000;
> PROVIDE (bar = .);
> .data : {
> *(.data)
> }
> }
>
> But the PowerPC linker is placing orphans after the assignment to dot.
> So .hash and others end up between bar and .data. The test fails
> because the dump file checks r_offset precisely.
>
> I think that r_offset does not matter to what we're trying to test
> here and we could allow any offset as long as the relocation is
> against foo. Is that right?
Where is the offset check? I only saw
---
#...
.* foo.*
#pass
---
in weak-dyn-1.rd.
H.J.