Spurious R_386_NONE relocations

[Alan Modra]

On Wed, Aug 25, 2004 at 11:02:09AM -0400, Jeff Baker wrote:
> This patch causes the appropriate data to be nulled,

Good, probably.  I say probably, because zero might not be correct for
R_386_PC32.  I'm applying the following:

	* elf32-i386.c (elf_i386_relocate_section): Tweak last change so
	that pcrel correction is applied for R_386_PC32.

Index: bfd/elf32-i386.c
===================================================================
RCS file: /cvs/src/src/bfd/elf32-i386.c,v
retrieving revision 1.126
diff -u -p -r1.126 elf32-i386.c
--- bfd/elf32-i386.c	25 Aug 2004 14:35:58 -0000	1.126
+++ bfd/elf32-i386.c	25 Aug 2004 23:00:19 -0000
@@ -2328,7 +2328,7 @@ elf_i386_relocate_section (bfd *output_b
 		 This is a hint to unwinders and other consumers of
 		 exception handling info that the FDE is invalid.  */
 	      bfd_put_32 (input_bfd, 0, contents + rel->r_offset);
-	      continue;
+	      break;
 	    }
 
 	  if ((input_section->flags & SEC_ALLOC) == 0)

> but many NONE 
> relocations are still produced.

I'm not particularly concerned about this, and certainly not motivated
personally to write another version of elf-eh-frame.c just to edit
.eh_frame produced by an old gcc.

>  Using a loader set to ignore NONE 
> relocations causes the program to SEGV calling the global constructor 
> for the class contained in the shared object.

If the above change doesn't fix it, I'm afraid you're going to have to
figure out why yourself..

-- 
Alan Modra
IBM OzLabs - Linux Technology Centre