Re: bfd failure (Was: address@hidden: Re: EH notes and split instructions])

[Jakub Jelinek]

On Wed, Feb 27, 2002 at 08:37:35AM -0800, Anthony Green wrote:
> On Tue, 2002-02-26 at 10:51, Jan Hubicka wrote:
> > the guilty is somewhat crazy ordering of purge_dead_edges, that first
> > handles jumps and returns and then does care exception handling.
> > The patch reorders it.  Bootstrapped i386 on mainline, bootstrapping
> > branch. OK (mainline/branch)?
> 
> Thanks Jan -- unfortunately I'm having a hard time testing this.  During
> the libgcj link...
> 
> /home/green/trunk/i/i686-pc-linux-gnu/bin/ld: section .debug_str
> [00000000 -> 00172d57] overlaps section .hash [00000094 -> 00017a7f]
> /home/green/trunk/i/i686-pc-linux-gnu/bin/ld: BFD 2.12.90 20020223
> internal error, aborting at /home/green/trunk/src/bfd/merge.c line 944
> in _bfd_merged_section_offset
> 
> Hmmm....

This looks like .debug_str was for some reason marked as
loadable (SHF_ALLOC?) while it shouldn't.
Can you see if .debug_str in all input objects is:
  [21] .debug_str        PROGBITS        00000000 00f4b3 00866c 01  MS  0  0  1

(important is MS in flags, but not A, W or X).

        Jakub