NetBSD/libexec/ld.elf_so
tv 7cfc131b50 Don't bomb if the ${DPADD} files are missing -- we want at least one of
them, where the other may not exist, and the linker will fail for us if
both don't exist.
2002-03-01 20:15:16 +00:00
..
arch
Makefile Don't bomb if the ${DPADD} files are missing -- we want at least one of 2002-03-01 20:15:16 +00:00
README
debug.c
debug.h
headers.c
ld.elf_so.1
load.c
malloc.c
map_object.c
paths.c
reloc.c
rtld.c
rtld.h
rtldenv.h
search.c
symbol.c
sysident.h
xmalloc.c
xprintf.c

README

$NetBSD: README,v 1.8 2001/12/20 06:49:03 thorpej Exp $

BUGS/PROBLEMS:

* rtld.c around line 184:

	#iffdef __alpha__
	/* XXX XXX XXX */
	_rtld_objself.pltgot = NULL;
	#endif

  It's a workaround for a linker bug, and will go away when the linker is
  fixed.

* The method used to relocate ld.so is shoddy and fragile.  Currently it's
  something like:

  alpha, powerpc:
	(1) assembly code relocates RELATIVE relocations in the GOT
	    (on some ports, all relative relocations; this will probably
	    have to be #ifdef'd in reloc.c), assuming that everything
	    in the GOT is a RELATIVE relocation.

	(2) alpha only: C code relocates the rest of the loader,
	    excluding those RELATIVE relocations already done (via an ...
	    iffy heuristic).

  arm, i386, m68k, sparc:
	(1) C code relocates the loader in _rtld_init() carefully so that
	    it does not touch any globals.

  mips:
	(1) loads ld.elf_so at 0, so no relocating is necessary.

  Something better (like that the glibc ELF ld.so does) should be done
  eventually, where a single set of C code is compiled twice, once for
  normal use, and once to be used to relocate the shared linker.  No
  relocations should be processed in assembly language.

* On the Alpha, this code Really needs to be checked to be sure that
  it doesn't make calls that could use FP, while doing lazy binding.

TO DO:

* Support for coexistance of 32-bit and 64-bit ELF on platforms that can
  do that.

* Support for relocating ld.elf_so on mips