RE: PATCH: Fix hidden vs. versioned symbol

RE: PATCH: Fix hidden vs. versioned symbol

[Harris, Jeff]

The patch is working against my ppc-specific soft-float test with libc and
libgcc.

Thanks for all your help,
Jeff

Re: PATCH: Fix hidden vs. versioned symbol

[H. J. Lu]

On Thu, May 08, 2003 at 12:23:34PM -0700, H. J. Lu wrote:
> 
> Here is the patch I will check in. I will check in a testcase first.
> 

This is the testcase I checked in.

H.J.
--
2003-05-08  H.J. Lu <hongjiu.lu@intel.com>

	* ld-elfvers/vers.exp: Add vers27d4 and vers27d5 to test
	versioned reference for hidden symbol.

	* ld-elfvers/vers27d4.dsym: New file.
	* ld-elfvers/vers27d4.ver: Likewise.

--- ld/testsuite/ld-elfvers/vers.exp.skip	2003-05-08 12:24:01.000000000 -0700
+++ ld/testsuite/ld-elfvers/vers.exp	2003-05-08 12:43:22.000000000 -0700
@@ -937,3 +937,5 @@ build_vers_lib_no_pic "vers27c2" vers27c
 build_vers_lib_pic "vers27d1" vers27d1.c vers27d1 "" vers27a.map vers27d.ver vers27d.dsym vers27d.sym
 build_vers_lib_pic "vers27d2" vers27d2.c vers27d2 "" "" vers27b.ver vers27b.dsym ""
 build_executable "vers27d3" vers27d3.c vers27d3 "vers27b.o vers27d2.so vers27d1.so" "" vers27b.ver vers27b.dsym ""
+build_vers_lib_pic "vers27d4" vers27d2.c vers27d4 "vers27a.so" "" vers27d4.ver vers27d4.dsym ""
+build_executable "vers27d5" vers27d3.c vers27d5 "vers27d4.so vers27b.o vers27a.so" "" vers27b.ver vers27b.dsym ""
--- ld/testsuite/ld-elfvers/vers27d4.dsym.skip	2003-05-08 12:45:40.000000000 -0700
+++ ld/testsuite/ld-elfvers/vers27d4.dsym	2003-05-08 12:39:18.000000000 -0700
@@ -0,0 +1 @@
+[0-9a-f]*      DF \*UND\*	[0-9a-f]*  VERS.0      (0x[0-9a-f][0-9a-f] )?foo
--- ld/testsuite/ld-elfvers/vers27d4.ver.skip	2003-05-08 12:45:35.000000000 -0700
+++ ld/testsuite/ld-elfvers/vers27d4.ver	2003-05-08 12:38:44.000000000 -0700
@@ -0,0 +1,3 @@
+Version References:
+  required from tmpdir/vers27a.so:
+    0x05aa7610 0x00 02 VERS.0

PATCH: Fix hidden vs. versioned symbol

[H. J. Lu]

[-- Attachment #1: Type: text/plain, Size: 706 bytes --]

On Thu, May 08, 2003 at 11:51:22AM -0700, H. J. Lu wrote:
> On Thu, May 08, 2003 at 09:55:12AM -0400, Harris, Jeff wrote:
> > I am having difficulties trying to create a test case other than the -lc
> > linking on ppc.
> > 
> > I do have more information regarding the symbols that were generated which
> > may shed some more light.  The output of nm on the shared library
> > (optionally linked against libc) and the exe (linked against the shared
> > library) show different symbol types for the __muldf3 symbol.  The four
> > cases I have are old/new binutils and with/without linking to -lc.
> > 
> 
> I found a testcase:
> 

Here is the patch I will check in. I will check in a testcase first.


H.J.

[-- Attachment #2: bfd-skip.patch --]
[-- Type: text/plain, Size: 709 bytes --]

2003-05-08  H.J. Lu <hongjiu.lu@intel.com>

	* elflink.h (elf_add_default_symbol): After skipping the
	unversioned symbol, go to non-default one.

--- bfd/elflink.h.hjl	2003-05-07 22:19:10.000000000 -0700
+++ bfd/elflink.h	2003-05-08 12:17:30.000000000 -0700
@@ -1108,7 +1108,7 @@ elf_add_default_symbol (abfd, info, h, n
     return FALSE;
 
   if (skip)
-    return TRUE;
+    goto nondefault;
 
   if (! override)
     {
@@ -1203,6 +1203,7 @@ elf_add_default_symbol (abfd, info, h, n
   /* We also need to define an indirection from the nondefault version
      of the symbol.  */
 
+nondefault:
   len = strlen (name);
   shortname = bfd_hash_allocate (&info->hash->table, len);
   if (shortname == NULL)