Re: [Patch]: Fix ld pr11138 FAILures on mips*.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 12/08/2011 12:34 PM, Richard Sandiford wrote:
David Daney<david.daney@xxxxxxxxxx>  writes:
There again, is this symbol (as opposed to the DT_MIPS_RLD_MAP tag)
actually part of the ABI?  I can't find any reference to it in the
original psABI, the SGI ELF64 spec, gdb or glibc.  If it's just an
internal thing, maybe we could get rid of it altogether, or at least
make it bind locally rather than globally.


That is an option too I suppose.  I would say that it is part of a de
facto ABI if nothing else.  The question of weather anybody uses it it a
different question.  I thought boehm-gc may have used it, but I cannot
find it there now.

Yeah, good point.  It occured to me rather belatedly that if wasn't
part of the de facto ABI, it wouldn't have two distinct names...

So if we were just doing this for the tag, the simplest way would have
been to find the .rld_map section in _bfd_mips_elf_finish_dynamic_sections.
But I agree that we should keep the symbol Just In Case.  So...

It might be possible to #define elf_backend_output_arch_local_syms and
then handle calculation of the rld_value value there instead.

If this seems like a good approach, I can prepare and test a patch that
does that.

...how about caching the hash table entry for __rld_map/__RLD_MAP/
__rld_obj_head in mips_elf_link_hash_entry, instead of rld_value.
(One field shared by all three should be enough.)  We can then use
that in _bfd_mips_elf_finish_dynamic_sections.

The code to set sym->st_value in _bfd_mips_elf_finish_dynamic_symbol
should already be redundant: the symbol is defined as being at the
start of .rld_map by:

	  s = bfd_get_section_by_name (abfd, ".rld_map");
	  BFD_ASSERT (s != NULL);

	  name = SGI_COMPAT (abfd) ? "__rld_map" : "__RLD_MAP";
	  bh = NULL;
	  if (!(_bfd_generic_link_add_one_symbol
		(info, abfd, name, BSF_GLOBAL, s, 0, NULL, FALSE,
		 get_elf_backend_data (abfd)->collect,&bh)))
	    return FALSE;

And the code to clear the first word should be redundant too, since
_bfd_mips_elf_size_dynamic_sections uses bfd_zalloc to allocate the
section's contents.

That's right, also elsewhere we set the size of the section to 4, which is not correct for ELF64 targets.

 So if we do cache the hash table entry, the whole
_bfd_mips_elf_finish_dynamic_symbol handling should become dead.

Richard

Thanks Richard,

Those are all good ideas, so I think I will hack up a new patch to try and implement this plan. Some care also needs to be taken for the SGI case as there the symbol seems to come from one of the input objects rather than being synthesized by the linker.

David Daney





[Index of Archives]     [Linux MIPS Home]     [LKML Archive]     [Linux ARM Kernel]     [Linux ARM]     [Linux]     [Git]     [Yosemite News]     [Linux SCSI]     [Linux Hams]

  Powered by Linux