[binutils-gdb] Always clear h->verinfo.verdef when overriding a dynamic definition

[sergiodj+buildbot at sergiodj dot net]

*** TEST RESULTS FOR COMMIT 48e30f5238c70e738f44474d595c476ef4e4ec38 ***

Author: H.J. Lu <hjl.tools@gmail.com>
Branch: master
Commit: 48e30f5238c70e738f44474d595c476ef4e4ec38

Always clear h->verinfo.verdef when overriding a dynamic definition

When linker defines a symbol to override a dynamic definition, it should
always clear h->verinfo.verdef so that the symbol won't be associated
with the version information from the dynamic object.  This happened to
the symbol "_edata" when creating an unversioned dynamic object linking
against:

1. libKF5ConfigCore.so.5.49.0
2. libKF5CoreAddons.so.5.49.0
3. libKF5I18n.so.5.49.0
4. libKF5DBusAddons.so.5.49.0
5. libQt5Xml.so.5.11.1
6. libQt5DBus.so.5.11.1
7. libQt5Core.so.5.11.1

Among them

libQt5Xml.so.5.11.1
   299: 000000000003e000     0 NOTYPE  GLOBAL DEFAULT   18 _edata@@Qt_5
libQt5DBus.so.5.11.1
   597: 0000000000092018     0 NOTYPE  GLOBAL DEFAULT   18 _edata@@Qt_5
libQt5Core.so.5.11.1
  2292: 00000000004df640     0 NOTYPE  GLOBAL DEFAULT   21 _edata@Qt_5
  2293: 00000000004df640     0 NOTYPE  GLOBAL DEFAULT   21 _edata@Qt_5

The problem is triggered by 2 duplicated entries of _edata@Qt_5 in
libQt5Core.so.5.11.1 which was created by gold.  Before this commit,
ld created the dynamic object with "_edata" in its dynamic symbol table
which was linker defined and associated with the version information
from libQt5Core.so.5.11.1.  The code in question was there when the
binutils source was imported to sourceware.org.  When such a dynamic
object was used later, we got:

/usr/bin/ld: bin/libKF5Service.so.5.49.0: _edata: invalid version 21 (max 0)
/usr/bin/ld: bin/libKF5Service.so.5.49.0: error adding symbols: bad value

Tested with many ELF targets.

	PR ld/23499
	* elflink.c (bfd_elf_record_link_assignment): Always clear
	h->verinfo.verdef when overriding a dynamic definition.