[PATCH, bfd] Fix incorrect initialization of function pointer variable with -Bsymbolic-functions

[Thomas Preud'homme]

> From: Marcus Shawcroft [mailto:marcus.shawcroft@gmail.com]
> Sent: Tuesday, January 13, 2015 7:07 PM
> 

Hi Marcus,

> 
> On 12 January 2015 at 14:47, Thomas Preud'homme
> <thomas.preudhomme@arm.com> wrote:
> 
> > bfd/ChangeLog
> >
> > 2015-01-06 Thomas Preud'homme thomas.preudhomme@arm.com
> >
> >     * elfnn-aarch64.c (elfNN_aarch64_final_link_relocate): Use
> >     SYMBOLIC_BIND to check if a symbol should be bound symbolically.
> >
> >
> > ld/testsuite/ChangeLog
> >
> > 2015-01-09 Thomas Preud'homme thomas.preudhomme@arm.com
> >
> >     * ld-aarch64/aarch64-elf.exp: Added relocs-257-symbolic-func test.
> >     * ld-aarch64/relocs-257-symbolic-func.d: New file.
> >     * ld-aarch64/relocs-257-symbolic-func.s: Likewise.
> 
> OK /Marcus

I just tried a backport to 2.25 and 2.24 branches and this patch applies
cleanly without any modification. There is no regression when running
the testsuite for either 2.25 and 2.24 neither does gcc testsuite (using
GCC trunk r219925 for testing binutils 2.25 and GCC 4.8 r219923 for
testing binutils 2.24).

Proposed ChangeLog entries:

*** bfd/ChangeLog ***

2015-01-21  Thomas Preud'homme  <thomas.preudhomme@arm.com>

	Backport from mainline
	2015-01-13  Thomas Preud'homme <thomas.preudhomme@arm.com>

	* elfnn-aarch64.c (elfNN_aarch64_final_link_relocate): Use SYMBOLIC_BIND
	to check if a symbol should be bound symbolically.

*** ld/testsuite/ChangeLog ***

2015-01-21  Thomas Preud'homme <thomas.preudhomme@arm.com>

	Backport from mainline
	2015-01-13  Thomas Preud'homme <thomas.preudhomme@arm.com>

	* ld-aarch64/aarch64-elf.exp: Added relocs-257-symbolic-func test.
	* ld-aarch64/relocs-257-symbolic-func.d: New file.
	* ld-aarch64/relocs-257-symbolic-func.s: Likewise.

Is this ok for backport?

Best regards,

Thomas