This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: PATCH: Clear non_elf on _GLOBAL_OFFSET_TABLE_


On Thu, Sep 30, 2010 at 8:20 PM, Alan Modra <amodra@gmail.com> wrote:
> On Thu, Sep 30, 2010 at 07:02:58PM -0700, H.J. Lu wrote:
>> On Fri, Sep 24, 2010 at 5:28 AM, Alan Modra <amodra@gmail.com> wrote:
>> > On Thu, Sep 23, 2010 at 10:27:36PM -0700, H.J. Lu wrote:
>> >> This patch clears non_elf on _GLOBAL_OFFSET_TABLE_
>> >
>> > To stop _bfd_elf_fix_symbol_flags non_elf code?
>> >
>> Yes, ?it has
> [snip]
>
> OK, then you should probably do this in _bfd_elf_define_linkage_sym.
> That way you'll treat _DYNAMIC and _PROCEDURE_LINKAGE_TABLE_ the same
> as _GLOBAL_OFFSET_TABLE_, and you'll also fix _GLOBAL_OFFSET_TABLE_
> for other targets. ?Please do run the testsuite for affected targets
> (grep _bfd_elf_define_linkage_sym bfd/*.c) to check for regressions
> before committing this change.
>

This is the patch I checked in. I tested it for

alpha-linux
bfin-elf
cr16-elf
frv-linux
i386-solaris2
mips64-linux
mips-linux
mn10300-elf
x86_64-solaris2

There are no regressions.

Thanks.


-- 
H.J.
---
2010-09-30  H.J. Lu  <hongjiu.lu@intel.com>

	PR ld/11812
	* elf32-i386.c (elf_i386_size_dynamic_sections): Don't
	 allocate .got.plt section if there are no GOT nor PLT
	 entries and there is no refeence to _GLOBAL_OFFSET_TABLE_.
	 * elf64-x86-64.c (elf64_x86_64_size_dynamic_sections): Li.kewise.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]