[Bug c++/67989] New: [REGRESSION] g++ ICE on armel valid code

[Bug c++/67989] New: [REGRESSION] g++ ICE on armel valid code

[costamagnagianfranco at yahoo dot it]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

            Bug ID: 67989
           Summary: [REGRESSION] g++ ICE on armel valid code
           Product: gcc
           Version: 5.2.1
            Status: UNCONFIRMED
          Severity: major
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: costamagnagianfranco at yahoo dot it
  Target Milestone: ---

Hi, some regression has happened between
   * Update to SVN 20150903 (r227431, 5.2.1) from the gcc-5-branch.
and
   * Update to SVN 20151003 (r228449, 5.2.1) from the gcc-5-branch.

because with Debian gcc-5.2.1-16 llvm was building fine on armel, while with
gcc5.2.1-21 it is FTBFS with an ICE.

you can see the Debian bug [1] the build logs [2], a valid log [3], the same
source but with the ICE

[1] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=801862
[2] https://buildd.debian.org/status/logs.php?pkg=llvm-toolchain-3.4&arch=armel
[3]
https://buildd.debian.org/status/fetch.php?pkg=llvm-toolchain-3.4&arch=armel&ver=1%3A3.4.2-15&stamp=1439243604
[4]
https://buildd.debian.org/status/fetch.php?pkg=llvm-toolchain-3.4&arch=armel&ver=1%3A3.4.2-15%2Bb1&stamp=1444483187

I'm attaching here the preprocessed source.

I'm not sure about providing everything necessary to look at the issue, please
let me know if you need anything more.

(the same failure is happening e.g. with llvm-3.5)

thanks

G.

[Bug c++/67989] [REGRESSION] g++ ICE on armel valid code

[costamagnagianfranco at yahoo dot it]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

--- Comment #1 from Gianfranco <costamagnagianfranco at yahoo dot it> ---
Created attachment 36526
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36526&action=edit
crash log

[Bug c++/67989] [REGRESSION] g++ ICE on armel valid code

[paolo.carlini at oracle dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

Paolo Carlini <paolo.carlini at oracle dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |paolo.carlini at oracle dot com
           Severity|major                       |normal

[Bug c++/67989] [REGRESSION] g++ ICE on armel valid code

[sylvestre at debian dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

Sylvestre Ledru <sylvestre at debian dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sylvestre at debian dot org

--- Comment #2 from Sylvestre Ledru <sylvestre at debian dot org> ---
The actual error is:

// /usr/include/c++/5/bits/atomic_base.h: In member function 'bool
lldb::SBAddress::OffsetAddress(lldb::addr_t)':
// /usr/include/c++/5/bits/atomic_base.h:396:35: internal compiler error: in
emit_move_insn, at expr.c:3601
//   return __atomic_load_n(&_M_i, __m);
//                                    ^

[Bug target/67989] [5/6 Regression] g++ ICE on armel valid code

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|                            |arm*
           Priority|P3                          |P1
      Known to work|                            |5.2.0
          Component|c++                         |target
                 CC|                            |kyrylo.tkachov at arm dot com
            Summary|[REGRESSION] g++ ICE on     |[5/6 Regression] g++ ICE on
                   |armel valid code            |armel valid code
   Target Milestone|---                         |5.3
      Known to fail|                            |5.2.1

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
Looks like a target issue to me, possibly caused by

2015-10-01  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>

        Backport from mainline
        2015-06-09  Shiva Chen  <shiva0217@gmail.com>

        * sync.md (atomic_load<mode>): Add conditional code for lda/ldr
        (atomic_store<mode>): Likewise.

testcase needs reduction, working on trunk needs to be verified and the
regression bisected (in case it's not obvious).

[Bug target/67989] [5/6 Regression] g++ ICE on armel valid code

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

ktkachov at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-10-22
                 CC|                            |ktkachov at gcc dot gnu.org
     Ever confirmed|0                           |1
      Known to fail|                            |6.0

--- Comment #5 from ktkachov at gcc dot gnu.org ---
The preprocessed source in the log was interspersed with a number of
warnings/debug messages from other components of the parallel build, so it was
not compilable due to the syntax errors.

Once I removed them I reproduced the ICE on GCC 5 and trunk.
Now reducing

[Bug target/67989] [5/6 Regression] g++ ICE on armel valid code

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

--- Comment #6 from ktkachov at gcc dot gnu.org ---
Created attachment 36561
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36561&action=edit
Reduced testcase

Attaching reduced testcase, here's the full ICE backtrace.

Can be reproduced with -O2 -std=c++11 -march=armv4t -marm


besttry.cpp: In member function ‘void
lldb::SBAddress::SetAddress(lldb::SBSection, lldb::addr_t)’:
besttry.cpp:27:23: internal compiler error: in emit_move_insn, at expr.c:3553
         __atomic_load (&_M_i, &tmp, __m);
                       ^
0x8135af emit_move_insn(rtx_def*, rtx_def*)
        $SRC/gcc/expr.c:3552
0x9763c1 expand_atomic_compare_and_swap(rtx_def**, rtx_def**, rtx_def*,
rtx_def*, rtx_def*, bool, memmodel, memmodel)
        $SRC/gcc/optabs.c:5907
0x976a22 expand_atomic_load(rtx_def*, rtx_def*, memmodel)
        $SRC/gcc/optabs.c:6019
0x71e9cf expand_builtin(tree_node*, rtx_def*, rtx_def*, machine_mode, int)
        $SRC/gcc/builtins.c:6738
0x810a85 expand_expr_real_1(tree_node*, rtx_def*, machine_mode,
expand_modifier, rtx_def**, bool)
        $SRC/gcc/expr.c:10474
0x73bded expand_expr
        $SRC/gcc/expr.h:255
0x73bded expand_call_stmt
        $SRC/gcc/cfgexpand.c:2647
0x73bded expand_gimple_stmt_1
        $SRC/gcc/cfgexpand.c:3534
0x73bded expand_gimple_stmt
        $SRC/gcc/cfgexpand.c:3697
0x73cd89 expand_gimple_tailcall
        $SRC/gcc/cfgexpand.c:3744
0x73cd89 expand_gimple_basic_block
        $SRC/gcc/cfgexpand.c:5677
0x741dde execute
        $SRC/gcc/cfgexpand.c:6286
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <http://gcc.gnu.org/bugs.html> for instructions.
>From gcc-bugs-return-500289-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Thu Oct 22 12:07:58 2015
Return-Path: <gcc-bugs-return-500289-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 48707 invoked by alias); 22 Oct 2015 12:07:58 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 48670 invoked by uid 48); 22 Oct 2015 12:07:53 -0000
From: "ktkachov at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/67989] [4.9/5/6 Regression] g++ ICE on armel valid code
Date: Thu, 22 Oct 2015 12:07:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: target
X-Bugzilla-Version: 5.2.1
X-Bugzilla-Keywords: ice-on-valid-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: ktkachov at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution:
X-Bugzilla-Priority: P1
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 5.3
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields: keywords short_desc cf_known_to_fail
Message-ID: <bug-67989-4-JJPdFaqsFW@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-67989-4@http.gcc.gnu.org/bugzilla/>
References: <bug-67989-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2015-10/txt/msg01844.txt.bz2
Content-length: 579

https://gcc.gnu.org/bugzilla/show_bug.cgi?idg989

ktkachov at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-on-valid-code
            Summary|[5/6 Regression] g++ ICE on |[4.9/5/6 Regression] g++
                   |armel valid code            |ICE on armel valid code
      Known to fail|                            |4.9.4

--- Comment #7 from ktkachov at gcc dot gnu.org ---
Btw, I see the ICE on 4.9 as well

[Bug target/67989] [4.9/5/6 Regression] g++ ICE on armel valid code

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

--- Comment #8 from ktkachov at gcc dot gnu.org ---
Also, fails only for the arm-linux targets. arm-none-eabi compiles fine.
Also, needs an -march option of armv6z or lower. armv6kz, armv7-a and higher
don't ICE

[Bug middle-end/67989] [4.9/5/6 Regression] g++ ICE on armel valid code

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

ktkachov at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
          Component|target                      |middle-end
           Assignee|unassigned at gcc dot gnu.org      |ktkachov at gcc dot gnu.org

--- Comment #9 from ktkachov at gcc dot gnu.org ---
I think this is a middle-end bug. I have a patch in testing

[Bug middle-end/67989] [4.9/5/6 Regression] g++ ICE on armel valid code

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

--- Comment #11 from ktkachov at gcc dot gnu.org ---
Author: ktkachov
Date: Mon Oct 26 10:28:12 2015
New Revision: 229317

URL: https://gcc.gnu.org/viewcvs?rev=229317&root=gcc&view=rev
Log:
[optabs.c] Fix PR 67989: Handle const0_rtx target in
expand_atomic_compare_and_swap

        PR middle-end/67989
        * optabs.c (expand_atomic_compare_and_swap): Handle case when
        ptarget_oval or ptarget_bool are const0_rtx.

        * g++.dg/pr67989.C: New test.


Added:
    trunk/gcc/testsuite/g++.dg/pr67989.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/optabs.c
    trunk/gcc/testsuite/ChangeLog

[Bug middle-end/67989] [4.9/5/6 Regression] g++ ICE on armel valid code

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

--- Comment #12 from ktkachov at gcc dot gnu.org ---
Fixed on trunk. Now testing backports to GCC 5 and 4.9

[Bug middle-end/67989] [4.9/5/6 Regression] g++ ICE on armel valid code

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

--- Comment #13 from ktkachov at gcc dot gnu.org ---
Author: ktkachov
Date: Tue Oct 27 14:03:27 2015
New Revision: 229442

URL: https://gcc.gnu.org/viewcvs?rev=229442&root=gcc&view=rev
Log:
[optabs.c] Fix PR 67989: Handle const0_rtx target in
expand_atomic_compare_and_swap

        Backport from mainline
        2015-10-26  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>

        PR middle-end/67989
        * optabs.c (expand_atomic_compare_and_swap): Handle case when
        ptarget_oval or ptarget_bool are const0_rtx.

        * g++.dg/pr67989.C: New test.


Added:
    branches/gcc-5-branch/gcc/testsuite/g++.dg/pr67989.C
Modified:
    branches/gcc-5-branch/gcc/ChangeLog
    branches/gcc-5-branch/gcc/optabs.c
    branches/gcc-5-branch/gcc/testsuite/ChangeLog

[Bug middle-end/67989] [4.9/5/6 Regression] g++ ICE on armel valid code

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

--- Comment #14 from ktkachov at gcc dot gnu.org ---
Author: ktkachov
Date: Tue Oct 27 14:07:04 2015
New Revision: 229443

URL: https://gcc.gnu.org/viewcvs?rev=229443&root=gcc&view=rev
Log:
[optabs.c] Fix PR 67989: Handle const0_rtx target in
expand_atomic_compare_and_swap

        Backport from mainline
        2015-10-26  Kyrylo Tkachov  <kyrylo.tkachov@arm.com>

        PR middle-end/67989
        * optabs.c (expand_atomic_compare_and_swap): Handle case when
        ptarget_oval or ptarget_bool are const0_rtx.

        * g++.dg/pr67989.C: New test.


Added:
    branches/gcc-4_9-branch/gcc/testsuite/g++.dg/pr67989.C
Modified:
    branches/gcc-4_9-branch/gcc/ChangeLog
    branches/gcc-4_9-branch/gcc/optabs.c
    branches/gcc-4_9-branch/gcc/testsuite/ChangeLog

[Bug middle-end/67989] [4.9/5/6 Regression] g++ ICE on armel valid code

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

ktkachov at gcc dot gnu.org changed:

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

--- Comment #15 from ktkachov at gcc dot gnu.org ---
Fixed on all release branches.

[Bug middle-end/67989] [4.9/5/6 Regression] g++ ICE on armel valid code

[costamagnagianfranco at yahoo dot it]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67989

--- Comment #16 from Gianfranco <costamagnagianfranco at yahoo dot it> ---
thanks a lot for the help! with this we will finally finish ocaml and llvm
Debian transitions :)