[Bug ld/14189] New: assertion fail ../../bfd/elf32-arm.c:12264

2012-05-31 Thread ronald.wahl at raritan dot com 
http://sourceware.org/bugzilla/show_bug.cgi?id=14189

 Bug #: 14189
   Summary: assertion fail ../../bfd/elf32-arm.c:12264
   Product: binutils
   Version: 2.23 (HEAD)
Status: NEW
  Severity: normal
  Priority: P2
 Component: ld
AssignedTo: unassig...@sourceware.org
ReportedBy: ronald.w...@raritan.com
Classification: Unclassified


Created attachment 6428
  --> http://sourceware.org/bugzilla/attachment.cgi?id=6428
test case

I have an issue were I get an "assertion fail ../../bfd/elf32-arm.c:12264" when
I link a shared library which links itself two static libraries and hides the
symbols using the "--exclude-libs ALL" switch. The sources are build with
-ffunction-sections and -fdata-sections and at link time --gc-section is used.
When removing --gc-sections or --exclude-libs the problem disappears.

See the attached testcase (tar containing 3 short C files and a Makefile which
you need to adapt for the right compiler name since it is an arm cross
compiler)

Actually the output is currently:

$ make
../bin/arm-PP-linux-uclibcgnueabi-gcc -fPIC -ffunction-sections -fdata-sections
-c -o test.o test.c
../bin/arm-PP-linux-uclibcgnueabi-gcc -fPIC -ffunction-sections -fdata-sections
-c -o test1.o test1.c
rm -f libtest1.a
../bin/arm-PP-linux-uclibcgnueabi-ar cru libtest1.a test1.o
../bin/arm-PP-linux-uclibcgnueabi-gcc -fPIC -ffunction-sections -fdata-sections
-c -o test2.o test2.c
rm -f libtest2.a
../bin/arm-PP-linux-uclibcgnueabi-ar cru libtest2.a test2.o
../bin/arm-PP-linux-uclibcgnueabi-gcc -shared -o libtest.so test.o
-Wl,--gc-sections,--exclude-libs,ALL -L. -ltest2 -ltest1
/opt/pp_toolchain/cross/008/bin/../lib/gcc/arm-pp-linux-uclibcgnueabi/4.6.3/../../../../arm-pp-linux-uclibcgnueabi/bin/ld:
BFD (Linux/GNU Binutils) 2.22.52.0.2.20120424 assertion fail
../../bfd/elf32-arm.c:12264

Looking at the code I saw that there was some recent activity where the assert
is generated (PLT ref count book-keeping check). Probably some case was
overlooked?!

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
--- You are receiving this mail because: ---
You are on the CC list for the bug.

___
bug-binutils mailing list
bug-binutils@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-binutils

[Bug ld/14189] assertion fail ../../bfd/elf32-arm.c:12264

2012-05-31 Thread hjl.tools at gmail dot com 
http://sourceware.org/bugzilla/show_bug.cgi?id=14189

H.J. Lu  changed:

   What|Removed |Added

 Status|NEW |WAITING
 CC||hjl.tools at gmail dot com

--- Comment #1 from H.J. Lu  2012-05-31 19:09:50 
UTC ---
Please provide a self-contained testcase with only assembly codes
for cross binutils.  Not all binutils developers have ARM cross
compilers.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
--- You are receiving this mail because: ---
You are on the CC list for the bug.

___
bug-binutils mailing list
bug-binutils@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-binutils

[Bug ld/14189] assertion fail ../../bfd/elf32-arm.c:12264

2012-05-31 Thread ronald.wahl at raritan dot com 
http://sourceware.org/bugzilla/show_bug.cgi?id=14189

Ronald Wahl  changed:

   What|Removed |Added

   Attachment #6428|0   |1
is obsolete||

--- Comment #2 from Ronald Wahl  2012-05-31 
20:45:08 UTC ---
Created attachment 6430
  --> http://sourceware.org/bugzilla/attachment.cgi?id=6430
test case with assembly only

Added test case which uses only assembly so no cross compiler is required.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
--- You are receiving this mail because: ---
You are on the CC list for the bug.

___
bug-binutils mailing list
bug-binutils@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-binutils

[Bug ld/14189] assertion fail ../../bfd/elf32-arm.c:12264

2012-05-31 Thread ronald.wahl at raritan dot com 
http://sourceware.org/bugzilla/show_bug.cgi?id=14189

Ronald Wahl  changed:

   What|Removed |Added

 Status|WAITING |NEW

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
--- You are receiving this mail because: ---
You are on the CC list for the bug.

___
bug-binutils mailing list
bug-binutils@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-binutils

[Bug ld/14189] --gc-sections doesn't work

2012-05-31 Thread hjl.tools at gmail dot com 
http://sourceware.org/bugzilla/show_bug.cgi?id=14189

H.J. Lu  changed:

   What|Removed |Added

 CC||rsandifo at sourceware dot
   ||org
Summary|assertion fail  |--gc-sections doesn't work
   |../../bfd/elf32-arm.c:12264 |

--- Comment #3 from H.J. Lu  2012-05-31 21:08:22 
UTC ---
It was caused by

http://sourceware.org/ml/binutils-cvs/2011-03/msg00080.html

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
--- You are receiving this mail because: ---
You are on the CC list for the bug.

___
bug-binutils mailing list
bug-binutils@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-binutils

[Bug binutils/14190] New: objdump 2.19 [avr 8] using -d, "lpm" without post-increment incorrectly disassembled with "+"

2012-05-31 Thread Xarquon at gmail dot com 
http://sourceware.org/bugzilla/show_bug.cgi?id=14190

 Bug #: 14190
   Summary: objdump 2.19 [avr 8]  using -d, "lpm" without
post-increment incorrectly disassembled with "+"
   Product: binutils
   Version: 2.19
Status: NEW
  Severity: normal
  Priority: P2
 Component: binutils
AssignedTo: unassig...@sourceware.org
ReportedBy: xarq...@gmail.com
Classification: Unclassified


Created attachment 6431
  --> http://sourceware.org/bugzilla/attachment.cgi?id=6431
Assembler test file to show "lpm r2,Z" disassembling incorrectly

Hello,
I believe objdump -d (for AVR-8 architecture, any variant) incorrectly
dis-assembles the "lpm" instruction without post-increment of Z register with a
post-increment symbol ("+").  See the AVR Instruction Set manual LPM
description to see the three opcode variations (i, ii, and iii below).  Opcode
varient (ii) should not post-increment Z.

Repro:
[Sample assembly file showing assembled listing]
> avr-as -aln -mmcu=atmega328p avr-objdump-bug.S -o avr-objdump-bug.o
   1  C895  lpm ; (i)   implied opcode (r0,Z)   
   2 0002 0490  lpm r0,Z; (ii)  explicit without
post-increment
   3 0004 0590  lpm r0,Z+   ; (iii) explicit with
post-increment
   4 0006 2490  lpm r2,Z; (ii)  explicit without
post-increment
   5 0008 2590  lpm r2,Z+   ; (iii) with post-increment

> avr-objdump -d avr-objdump-bug.o

avr-objdump-bug.o: file format elf32-avr

Disassembly of section .text:

 <.text>:
   0:   c8 95   lpm
   2:   04 90   lpm r0, Z+  [INCORRECT]
   4:   05 90   lpm r0, Z+
   6:   24 90   lpm r2, Z+  [INCORRECT]
   8:   25 90   lpm r2, Z+


[NOTE: see how the op-codes differ between 2: and 4: and 6: and 8: (and I
believe are correct), but the instructions are disassembled the same.  I
believe the x4 xx op-codes should not have a post-increment output on the
mnemonic.]

> avr-objdump -v
GNU objdump (WinAVR 20100110) 2.19

Thanks!

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
--- You are receiving this mail because: ---
You are on the CC list for the bug.

___
bug-binutils mailing list
bug-binutils@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-binutils