PR24511, nm should not mark symbols in .init_array as "t"

[Hans-Peter Nilsson]

On Thu, 27 Feb 2020, Alan Modra wrote:

> On Thu, Feb 27, 2020 at 05:26:10PM +1030, Alan Modra wrote:
> > 	* syms.c (stt): Trim off all but 'e', 'i' and 'p' entries.
> > 	(coff_section_type): Adjust comment.
> > 	(decode_section_type): Likewise.  Call coff_section_type before
> > 	decode_section_type.
> > 	(bfd_decode_symclass): Use 'c' for common sections other than
> > 	the standard one.
>
> With the bfd_decode_symclass change, ld SIZEOF test fails due to
> data_end being reported as type '?'.
>
> Does this look a reasonable solution, Hans-Peter?
>
> bfd/
> 	* mmo.c (mmo_scan): Set SEC_DATA for .data.
> ld/
> 	* testsuite/ld-mmix/b-fixo2.d: Adjust for .data change.
> 	* testsuite/ld-mmix/sec-2.d: Likewise.
> 	* testsuite/ld-mmix/sec-3.d: Likewise.
> 	* testsuite/ld-mmix/sec-4.d: Likewise.
> 	* testsuite/ld-mmix/spec802.d: Likewise.
> 	* testsuite/ld-mmix/spec803.d: Likewise.
> 	* testsuite/ld-mmix/spec804.d: Likewise.
> 	* testsuite/ld-mmix/spec805.d: Likewise.
> 	* testsuite/ld-mmix/spec806.d: Likewise.
> 	* testsuite/ld-mmix/spec807.d: Likewise.
> 	* testsuite/ld-mmix/spec808.d: Likewise.


Correcting data sections to be classified as DATA?
Certainly LGTM! :)  Thanks.

brgds, H-P