This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: g77 pic case vs. binutils & readelf
- To: Nick Clifton <nickc at cygnus dot com>
- Subject: Re: g77 pic case vs. binutils & readelf
- From: Robert Lipe <robertl at sco dot com>
- Date: Fri, 4 Jun 1999 14:22:21 -0500
- Cc: egcs at egcs dot cygnus dot com
- References: <199906041838.LAA02603@elmo.cygnus.com>
> Could it be a case of REL vs RELA somewhere ?
I don't really know what that means. dump_relocations() in readelf.c is
called with "is_rela" = 0.
> In case it helps I have just checked into sourceware a patched version
> of readelf which should detect bad symbol indcies and print a warning
> message (and then carry on to display the rest of the relocs) rather
> than just core dumping.
(I love being able to just blast these things around in CVS!)
Now the EGCS-compiled g77 case linked against the PIC libg2c.a that is
tormenting the linker displays as:
Relocation section '.rel.got' at offset 0x7b4 contains 4 entries:
Offset Info Type Symbol's Value Symbol's Name
0805582c 02206 R_386_GLOB_DAT 00000000 errno
0805586c 02b06 R_386_GLOB_DAT 08048ab4 ungetc
00000000 00000 R_386_NONE
00000000 452b28 unrecognised: 28 bad symbol index: 0000452b
Is there any way can we find the file that delivered that symbol to us?
RJL