Objects created by "ld -r" can contain merge sections that have duplicate
entries. Linking such an object can result in a bad setup of the GOT
section, if those entries are referenced both by GOT and non-GOT relocations.
$ cat ltoff.s
.section .rodata.str1.8,"aMS", 1
.LC1: .string "foo"
.LC2: .string "foo"
.section .data.rel.local,"aw"
.quad .LC2
.section .rodata,"a"
.LC3: .string "bar"
.balign 8
.space 0x400000
.text
addl [EMAIL PROTECTED](.LC1),r1 ;;
addl [EMAIL PROTECTED](.LC3),r1 ;;
$ gcc -c ltoff.s
$ ld -shared ltoff.o
$ objdump -d
a.out: file format elf64-ia64-little
Disassembly of section .text:
0000000000000210 <.text>:
210: 0b 60 60 02 00 24 [MMI] addl r12=24,r1;;
216: c0 c0 04 00 48 00 addl r12=24,r1
21c: 00 00 04 00 nop.i 0x0;;
$ objdump -R
a.out: file format elf64-ia64-little
DYNAMIC RELOCATION RECORDS
OFFSET TYPE VALUE
0000000000410340 REL64LSB *ABS*+0x0000000000000220
0000000000410360 REL64LSB *ABS*+0x0000000000000220
0000000000410360 REL64LSB *ABS*+0x0000000000000228
Note that there are two relocations for the same GOT entry that resolve to
different addresses (the strings "foo" and "bar", resp.) The insn at
address 216 should have used its own GOT entry.
The problem is in elf64_ia64_relocate_section, adjusting the addends of
the local sym hash for the merge section (through
_bfd_merged_section_offset) creates multiple entries for the same addend.
In the testcase this results in two entries for addend 0 where one wants a
GOT entry, the other one doesn't.
--
Summary: String merging breaks ia64 linker
Product: binutils
Version: 2.18 (HEAD)
Status: NEW
Severity: normal
Priority: P2
Component: ld
AssignedTo: unassigned at sources dot redhat dot com
ReportedBy: schwab at suse dot de
CC: bug-binutils at gnu dot org,hjl at lucon dot org
GCC target triplet: ia64-*-*
http://sourceware.org/bugzilla/show_bug.cgi?id=4590
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
_______________________________________________
bug-binutils mailing list
bug-binutils@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-binutils
