[Bug debug/98716] New: [11 Regression] sanitizer regressions by r11-6755

[Bug debug/98716] New: [11 Regression] sanitizer regressions by r11-6755

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98716

            Bug ID: 98716
           Summary: [11 Regression] sanitizer regressions by r11-6755
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: debug
          Assignee: unassigned at gcc dot gnu.org
          Reporter: hjl.tools at gmail dot com
                CC: mark at gcc dot gnu.org
  Target Milestone: ---

On Linux/x86-64 with binutils master branch plus

https://sourceware.org/pipermail/binutils/2021-January/114980.html

r11-6755 caused:

FAIL: c-c++-common/asan/alloca_big_alignment.c   -O2 -flto
-fno-use-linker-plugin -flto-partition=none  output pattern test
FAIL: c-c++-common/asan/alloca_big_alignment.c   -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  output pattern test
FAIL: c-c++-common/asan/alloca_detect_custom_size.c   -O2 -flto
-fno-use-linker-plugin -flto-partition=none  output pattern test
FAIL: c-c++-common/asan/alloca_detect_custom_size.c   -O2 -flto
-fuse-linker-plugin -fno-fat-lto-objects  output pattern test
FAIL: c-c++-common/asan/alloca_overflow_partial.c   -O2 -flto
-fno-use-linker-plugin -flto-partition=none  output pattern test
FAIL: c-c++-common/asan/alloca_overflow_partial.c   -O2 -flto
-fuse-linker-plugin -fno-fat-lto-objects  output pattern test
FAIL: c-c++-common/asan/alloca_overflow_right.c   -O2 -flto
-fno-use-linker-plugin -flto-partition=none  output pattern test
FAIL: c-c++-common/asan/alloca_overflow_right.c   -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  output pattern test
FAIL: c-c++-common/asan/alloca_underflow_left.c   -O2 -flto
-fno-use-linker-plugin -flto-partition=none  output pattern test
FAIL: c-c++-common/asan/alloca_underflow_left.c   -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  output pattern test
FAIL: c-c++-common/asan/global-overflow-1.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  output pattern test
FAIL: c-c++-common/asan/global-overflow-1.c   -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  output pattern test
FAIL: c-c++-common/asan/heap-overflow-1.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  output pattern test
FAIL: c-c++-common/asan/heap-overflow-1.c   -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  output pattern test
FAIL: c-c++-common/asan/null-deref-1.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  output pattern test
FAIL: c-c++-common/asan/null-deref-1.c   -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  output pattern test
FAIL: c-c++-common/asan/sanity-check-pure-c-1.c   -O2 -flto
-fno-use-linker-plugin -flto-partition=none  output pattern test
FAIL: c-c++-common/asan/sanity-check-pure-c-1.c   -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  output pattern test
FAIL: c-c++-common/asan/stack-overflow-1.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  output pattern test
FAIL: c-c++-common/asan/stack-overflow-1.c   -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  output pattern test
FAIL: c-c++-common/asan/strncpy-overflow-1.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  output pattern test
FAIL: c-c++-common/asan/strncpy-overflow-1.c   -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  output pattern test
FAIL: c-c++-common/asan/use-after-free-1.c   -O2 -flto -fno-use-linker-plugin
-flto-partition=none  output pattern test
FAIL: c-c++-common/asan/use-after-free-1.c   -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  output pattern test

[Bug debug/98716] [11 Regression] sanitizer regressions by r11-6755

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98716

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ian at gcc dot gnu.org,
                   |                            |jakub at gcc dot gnu.org

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
And go is also mostly broken.
I think at least libbacktrace needs some DWARF5 work.

[Bug debug/98716] [11 Regression] sanitizer regressions by r11-6755

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98716

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |11.0
           Priority|P3                          |P1

[Bug debug/98716] [11 Regression] sanitizer regressions by r11-6755

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98716

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
The failed asan tests are all with -flto, and the changes in the output are
from:
WRITE of size 1 at 0x7ffeab2b5d8a thread T0
    #0 0x401288 in foo
/home/jakub/src/gcc/gcc/testsuite/c-c++-common/asan/alloca_big_alignment.c:11
    #1 0x4010c4 in main
/home/jakub/src/gcc/gcc/testsuite/c-c++-common/asan/alloca_big_alignment.c:15
    #2 0x7f9c719311a1 in __libc_start_main (/lib64/libc.so.6+0x281a1)
    #3 0x40112d in _start
(/home/jakub/src/gcc/obj66/gcc/testsuite/gcc6/alloca_big_alignment.exe+0x40112d)
like output to:
WRITE of size 1 at 0x7ffe337c4b0a thread T0
    #0 0x401288 in foo gcc/obj2/gcc/testsuite/gcc3/<artificial>:11
    #1 0x4010c4 in main gcc/obj2/gcc/testsuite/gcc3/<artificial>:15
    #2 0x7efe2df801a1 in __libc_start_main (/lib64/libc.so.6+0x281a1)
    #3 0x40112d in _start
(gcc/obj2/gcc/testsuite/gcc3/alloca_big_alignment.exe+0x40112d)

[Bug debug/98716] [11 Regression] sanitizer regressions by r11-6755

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98716

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
So, seems when asan alloca_big_alignment test is built with -gdwarf-4 -flto, we
have:
  Offset:                      0x57
  Length:                      214
  DWARF Version:               4
  Prologue Length:             105
  Minimum Instruction Length:  1
  Maximum Ops per Instruction: 1
  Initial value of 'is_stmt':  1
  Line Base:                   -5
  Line Range:                  14
  Opcode Base:                 13

 Opcodes:
  Opcode 1 has 0 args
  Opcode 2 has 1 arg
  Opcode 3 has 1 arg
  Opcode 4 has 1 arg
  Opcode 5 has 1 arg
  Opcode 6 has 0 args
  Opcode 7 has 0 args
  Opcode 8 has 0 args
  Opcode 9 has 1 arg
  Opcode 10 has 0 args
  Opcode 11 has 0 args
  Opcode 12 has 1 arg

 The Directory Table (offset 0x73):
  1     /usr/src/gcc/gcc/testsuite/c-c++-common/asan

 The File Name Table (offset 0xa1):
  Entry Dir     Time    Size    Name
  1     1       0       0       alloca_big_alignment.c
  2     0       0       0       <built-in>

 Line Number Statements:
  [0x000000ca]  Set column to 56
  [0x000000cc]  Extended opcode 2: set Address to 0x4011f0
  [0x000000d7]  Special opcode 12: advance Address by 0 to 0x4011f0 and Line by
7 to 8
  [0x000000d8]  Set is_stmt to 0
  [0x000000d9]  Copy (view 1)
  [0x000000da]  Set column to 17
  [0x000000dc]  Special opcode 20: advance Address by 1 to 0x4011f1 and Line by
1 to 9
  [0x000000dd]  Special opcode 89: advance Address by 6 to 0x4011f7 and Line by
0 to 9
  [0x000000de]  Set column to 56
  [0x000000e0]  Special opcode 130: advance Address by 9 to 0x401200 and Line
by -1 to 8
  [0x000000e1]  Advance PC by constant 17 to 0x401211
  [0x000000e2]  Special opcode 47: advance Address by 3 to 0x401214 and Line by
0 to 8
  [0x000000e3]  Set column to 3
  [0x000000e5]  Set is_stmt to 1
  [0x000000e6]  Special opcode 90: advance Address by 6 to 0x40121a and Line by
1 to 9
  [0x000000e7]  Set is_stmt to 0

This is for a TU created by LTO, which in .debug_info has:
    <565>   DW_AT_name        : (indirect string, offset: 0x676): <artificial>
    <569>   DW_AT_comp_dir    : (indirect string, offset: 0x652):
/usr/src/gcc/obj2/gcc/testsuite/gcc
Now, with -gdwarf-5 -flto, is still has:
    <566>   DW_AT_name        : (indirect line string, offset: 0x11d):
<artificial>
    <56a>   DW_AT_comp_dir    : (indirect line string, offset: 0x12a):
/usr/src/gcc/obj2/gcc/testsuite/gcc
but the .debug_lines part starts with:
  Offset:                      0x57
  Length:                      162
  DWARF Version:               5
  Address size (bytes):        8
  Segment selector (bytes):    0
  Prologue Length:             51
  Minimum Instruction Length:  1
  Maximum Ops per Instruction: 1
  Initial value of 'is_stmt':  1
  Line Base:                   -5
  Line Range:                  14
  Opcode Base:                 13

 Opcodes:
  Opcode 1 has 0 args
  Opcode 2 has 1 arg
  Opcode 3 has 1 arg
  Opcode 4 has 1 arg
  Opcode 5 has 1 arg
  Opcode 6 has 0 args
  Opcode 7 has 0 args
  Opcode 8 has 0 args
  Opcode 9 has 1 arg
  Opcode 10 has 0 args
  Opcode 11 has 0 args
  Opcode 12 has 1 arg

 The Directory Table (offset 0x79, lines 2, columns 1):
  Entry Name
  0     (indirect line string, offset: 0x12a):
/usr/src/gcc/obj2/gcc/testsuite/gcc
  1     (indirect line string, offset: 0x14e):
/usr/src/gcc/gcc/testsuite/c-c++-common/asan

 The File Name Table (offset 0x87, lines 3, columns 2):
  Entry Dir     Name
  0     0       (indirect line string, offset: 0x11d): <artificial>
  1     1       (indirect line string, offset: 0x1b3): alloca_big_alignment.c
  2     0       (indirect line string, offset: 0x17b): <built-in>

 Line Number Statements:
  [0x00000096]  Set column to 56
  [0x00000098]  Extended opcode 2: set Address to 0x4011f0
  [0x000000a3]  Special opcode 12: advance Address by 0 to 0x4011f0 and Line by
7 to 8
  [0x000000a4]  Set is_stmt to 0
  [0x000000a5]  Copy (view 1)
  [0x000000a6]  Set column to 17
  [0x000000a8]  Special opcode 20: advance Address by 1 to 0x4011f1 and Line by
1 to 9
  [0x000000a9]  Special opcode 89: advance Address by 6 to 0x4011f7 and Line by
0 to 9
  [0x000000aa]  Set column to 56
  [0x000000ac]  Special opcode 130: advance Address by 9 to 0x401200 and Line
by -1 to 8
  [0x000000ad]  Advance PC by constant 17 to 0x401211
  [0x000000ae]  Special opcode 47: advance Address by 3 to 0x401214 and Line by
0 to 8
  [0x000000af]  Set column to 3
  [0x000000b1]  Set is_stmt to 1
  [0x000000b2]  Special opcode 90: advance Address by 6 to 0x40121a and Line by
1 to 9
  [0x000000b3]  Set is_stmt to 0

So the important difference is in that DWARF5 has the 0 entry in the filename
table while DWARF4 does not.

DWARF5 Table 6.4: Line number program initial state says that the
initial file is 1, which means I think read_line_program in libbacktrace is
incorrect for DWARF5.

[Bug debug/98716] [11 Regression] sanitizer regressions by r11-6755

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98716

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |jakub at gcc dot gnu.org
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2021-01-18
     Ever confirmed|0                           |1

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 49993
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=49993&action=edit
gcc11-pr98716.patch

Untested fix.

[Bug debug/98716] [11 Regression] sanitizer regressions by r11-6755

[ian at airs dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98716

Ian Lance Taylor <ian at airs dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ian at airs dot com

--- Comment #5 from Ian Lance Taylor <ian at airs dot com> ---
I'm not seeing any failures in the Go testsuite with GNU binutils 2.35.1. 
Anybody know what changed in newer version of the binutils?

[Bug debug/98716] [11 Regression] sanitizer regressions by r11-6755

[ian at airs dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98716

--- Comment #6 from Ian Lance Taylor <ian at airs dot com> ---
On the other hand the libbacktrace testsuite now fails when using dwz
0.13+20201015-2.  But I guess that is not a GCC problem.

dwz -m b3test_dwz_common.debug b3test_dwz_1 b3test_dwz_2
dwz: b3test_dwz_1: Unknown DWARF DW_FORM_implicit_const
dwz: b3test_dwz_1: Couldn't read abbrev at offset 0x0
dwz: b3test_dwz_2: Unknown DWARF DW_FORM_implicit_const
dwz: b3test_dwz_2: Couldn't read abbrev at offset 0x0
dwz: Too few files for multifile optimization
make[3]: *** [Makefile:2436: b3test_dwz] Error 1

[Bug debug/98716] [11 Regression] sanitizer regressions by r11-6755

[mark at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98716

--- Comment #7 from Mark Wielaard <mark at gcc dot gnu.org> ---
(In reply to Ian Lance Taylor from comment #5)
> I'm not seeing any failures in the Go testsuite with GNU binutils 2.35.1. 
> Anybody know what changed in newer version of the binutils?

The difference is that with binutils 2.36 (not released yet) gas generates when
asked to (with -gdwarf-5) DWARF 5 line tables. While 2.35.1 always generated
DWARF 4 line tables.

[Bug debug/98716] [11 Regression] sanitizer regressions by r11-6755

[mark at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98716

--- Comment #8 from Mark Wielaard <mark at gcc dot gnu.org> ---
(In reply to Ian Lance Taylor from comment #6)
> On the other hand the libbacktrace testsuite now fails when using dwz
> 0.13+20201015-2.  But I guess that is not a GCC problem.
> 
> dwz -m b3test_dwz_common.debug b3test_dwz_1 b3test_dwz_2
> dwz: b3test_dwz_1: Unknown DWARF DW_FORM_implicit_const
> dwz: b3test_dwz_1: Couldn't read abbrev at offset 0x0
> dwz: b3test_dwz_2: Unknown DWARF DW_FORM_implicit_const
> dwz: b3test_dwz_2: Couldn't read abbrev at offset 0x0
> dwz: Too few files for multifile optimization
> make[3]: *** [Makefile:2436: b3test_dwz] Error 1

Yes, you need the following upstream patch (already in Fedora rawhide):
https://sourceware.org/pipermail/dwz/2021q1/000775.html

[Bug debug/98716] [11 Regression] sanitizer regressions by r11-6755

[mark at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98716

--- Comment #9 from Mark Wielaard <mark at gcc dot gnu.org> ---
(In reply to Mark Wielaard from comment #7)
> (In reply to Ian Lance Taylor from comment #5)
> > I'm not seeing any failures in the Go testsuite with GNU binutils 2.35.1. 
> > Anybody know what changed in newer version of the binutils?
> 
> The difference is that with binutils 2.36 (not released yet) gas generates
> when asked to (with -gdwarf-5) DWARF 5 line tables. While 2.35.1 always
> generated DWARF 4 line tables.

Actually 2.35.1 does already have support for -gdwarf-5, but had some bugs
which are detected by gcc configure, so that it is not used. The bugs were
fixed on the 2.35 branch, but there is no 2.35.2 release (yet?). Anyway. The
difference is the gcc configure check for HAVE_AS_WORKING_DWARF_N_FLAG.

[Bug debug/98716] [11 Regression] sanitizer regressions by r11-6755

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98716

--- Comment #10 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Ian Lance Taylor <ian@gcc.gnu.org>:

https://gcc.gnu.org/g:325e70b47c6c321710c7b9c792b8fbee95cecd63

commit r11-6779-g325e70b47c6c321710c7b9c792b8fbee95cecd63
Author: Ian Lance Taylor <iant@golang.org>
Date:   Mon Jan 18 14:38:10 2021 -0800

    libbacktrace: use correct directory/filename for DWARF 5

            PR debug/98716
            * dwarf.c (read_v2_paths): Allocate zero entry for dirs and
            filenames.
            (read_line_program): Remove parameter u, change caller.  Don't
            subtract one from dirs and filenames index.
            (read_function_entry): Don't subtract one from filenames index.

[Bug debug/98716] [11 Regression] sanitizer regressions by r11-6755

[ian at airs dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98716

Ian Lance Taylor <ian at airs dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
             Status|ASSIGNED                    |RESOLVED

--- Comment #11 from Ian Lance Taylor <ian at airs dot com> ---
This is probably fixed.  I couldn't fully test it because of PR 98708.  Let me
know if it's still broken.