[Bug c++/58864] New: [4.8/4.9 regression] ICE in connect_traces, at dwarf2cfi.c:NNNN

[Bug c++/58864] New: [4.8/4.9 regression] ICE in connect_traces, at dwarf2cfi.c:NNNN

[ppluzhnikov at google dot com]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="UTF-8", Size: 3096 bytes --]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58864

            Bug ID: 58864
           Summary: [4.8/4.9 regression] ICE in connect_traces, at
                    dwarf2cfi.c:NNNN
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: ppluzhnikov at google dot com

Google ref: b/11306585

Attached test case causes ICE with current trunk (g++ (GCC) 4.9.0 20131016
(experimental))

As far as I can tell, 4.7 is fine, but 4.8 and above is broken.

g++ -Os -m32 -c t.cc

t.cc: In member function ‘void L::compute()’:
t.cc:69:1: internal compiler error: in connect_traces, at dwarf2cfi.c:2686
 }
 ^
0x84d842 connect_traces
        ../../gcc/dwarf2cfi.c:2686
0x84d842 execute_dwarf2_frame
        ../../gcc/dwarf2cfi.c:2906
0x84d842 execute
        ../../gcc/dwarf2cfi.c:3401
Please submit a full bug report,
with preprocessed source if appropriate.
>From gcc-bugs-return-432635-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Thu Oct 24 17:16:09 2013
Return-Path: <gcc-bugs-return-432635-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 17445 invoked by alias); 24 Oct 2013 17:16:09 -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 17417 invoked by uid 48); 24 Oct 2013 17:16:05 -0000
From: "ppluzhnikov at google dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/58864] [4.8/4.9 regression] ICE in connect_traces, at dwarf2cfi.c:NNNN
Date: Thu, 24 Oct 2013 17:16:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Version: unknown
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: ppluzhnikov at google dot com
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields: attachments.created
Message-ID: <bug-58864-4-Mwz2exl89Z@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-58864-4@http.gcc.gnu.org/bugzilla/>
References: <bug-58864-4@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: 2013-10/txt/msg01779.txt.bz2
Content-length: 229

http://gcc.gnu.org/bugzilla/show_bug.cgi?idX864

--- Comment #1 from Paul Pluzhnikov <ppluzhnikov at google dot com> ---
Created attachment 31084
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id1084&actioníit
test case

[Bug target/58864] [4.8/4.9 regression] ICE in connect_traces, at dwarf2cfi.c:NNNN

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58864

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.8.3

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
I cannot reproduce this with 4.8 or trunk on a x86_64 host.

[Bug target/58864] [4.8/4.9 regression] ICE in connect_traces, at dwarf2cfi.c:NNNN

[ppluzhnikov at google dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58864

--- Comment #3 from Paul Pluzhnikov <ppluzhnikov at google dot com> ---
> I cannot reproduce this with 4.8 or trunk on a x86_64 host.

Strange: the crash reproduces for me using multiple independent builds of trunk
on x86_64, configured as:

../configure --prefix /tmp/gcc-svn-install --enable-linker-build-id
--disable-lto --with-linker-hash-style=gnu --disable-libstdcxx-pch
--enable-languages=c,c++

Also reproduced using a clean gcc-4_8 branch as of 2 weeks ago:
gcc (GCC) 4.8.2 20131012 (prerelease), configured similarly:
../configure --prefix /tmp/gcc-svn-4.8-install --enable-languages=c,c++
--enable-linker-build-id --with-linker-hash-style=gnu --disable-lto

t.cc: In member function 'void L::compute()':
t.cc:69:1: internal compiler error: in connect_traces, at dwarf2cfi.c:2676
 }
 ^
0x6a5ee7 connect_traces
    ../../gcc/dwarf2cfi.c:2676
0x6a5ee7 execute_dwarf2_frame
    ../../gcc/dwarf2cfi.c:2898



I just repeated a build on a clean gcc-4.8:
  URL: svn+ssh://ppluzhnikov@gcc.gnu.org/svn/gcc/branches/gcc-4_8-branch
  Revision: 204069
with exactly the same outcome.


You are using '-Os -m32', right?

[Bug target/58864] [4.8/4.9 regression] ICE in connect_traces, at dwarf2cfi.c:NNNN

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58864

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|i?86-*-linux-gnu            |i686-*-linux-gnu
           Priority|P3                          |P2
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2013-11-19
     Ever confirmed|0                           |1

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to Paul Pluzhnikov from comment #3)
> > I cannot reproduce this with 4.8 or trunk on a x86_64 host.
> 
> Strange: the crash reproduces for me using multiple independent builds of
> trunk on x86_64, configured as:
> 
> ../configure --prefix /tmp/gcc-svn-install --enable-linker-build-id
> --disable-lto --with-linker-hash-style=gnu --disable-libstdcxx-pch
> --enable-languages=c,c++
> 
> Also reproduced using a clean gcc-4_8 branch as of 2 weeks ago:
> gcc (GCC) 4.8.2 20131012 (prerelease), configured similarly:
> ../configure --prefix /tmp/gcc-svn-4.8-install --enable-languages=c,c++
> --enable-linker-build-id --with-linker-hash-style=gnu --disable-lto
> 
> t.cc: In member function 'void L::compute()':
> t.cc:69:1: internal compiler error: in connect_traces, at dwarf2cfi.c:2676
>  }
>  ^
> 0x6a5ee7 connect_traces
> 	../../gcc/dwarf2cfi.c:2676
> 0x6a5ee7 execute_dwarf2_frame
> 	../../gcc/dwarf2cfi.c:2898
> 
> 
> 
> I just repeated a build on a clean gcc-4.8:
>   URL: svn+ssh://ppluzhnikov@gcc.gnu.org/svn/gcc/branches/gcc-4_8-branch
>   Revision: 204069
> with exactly the same outcome.
> 
> 
> You are using '-Os -m32', right?

Yes.  Ah, it breaks with -march=i686 but works with -march=i586.

Confirmed.
>From gcc-bugs-return-435024-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Tue Nov 19 10:29:48 2013
Return-Path: <gcc-bugs-return-435024-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 12315 invoked by alias); 19 Nov 2013 10:29:47 -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 12208 invoked by uid 48); 19 Nov 2013 10:29:44 -0000
From: "rguenth at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/58874] [4.9 Regression] [c++11] ICE with OpenMP reduction declaration and -std=c++11
Date: Tue, 19 Nov 2013 10:29:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Version: 4.9.0
X-Bugzilla-Keywords: ice-on-valid-code, openmp
X-Bugzilla-Severity: normal
X-Bugzilla-Who: rguenth at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Priority: P1
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 4.9.0
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields: priority
Message-ID: <bug-58874-4-j65FQKaXIe@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-58874-4@http.gcc.gnu.org/bugzilla/>
References: <bug-58874-4@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: 2013-11/txt/msg01801.txt.bz2
Content-length: 290

http://gcc.gnu.org/bugzilla/show_bug.cgi?idX874

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1

[Bug target/58864] [4.8/4.9 regression] ICE in connect_traces, at dwarf2cfi.c:NNNN

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58864

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
                 CC|                            |jakub at gcc dot gnu.org
           Assignee|unassigned at gcc dot gnu.org      |jakub at gcc dot gnu.org

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 31326
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31326&action=edit
gcc49-pr58864.patch

Untested fix.  Note looking around at other callers of emit_conditional_move, I
think there are other suspect places, in particular emit_store_flag and
expand_doubleword_shift.  ifcvt.c should be fine, because no stack adjustments
should be pending after expansion.

[Bug target/58864] [4.8/4.9 regression] ICE in connect_traces, at dwarf2cfi.c:NNNN

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58864

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 31328
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31328&action=edit
gcc49-pr58864.patch

Alternate patch.  Rather than doing do_pending_stack_adjust () everywhere just
in case emit_conditional_move could be called and it's insns removed, this
allows remember current pending_stack_adjust state and restore it if we decide
to delete the insns.

[Bug target/58864] [4.8/4.9 regression] ICE in connect_traces, at dwarf2cfi.c:NNNN

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58864

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Author: jakub
Date: Tue Dec  3 07:29:20 2013
New Revision: 205618

URL: http://gcc.gnu.org/viewcvs?rev=205618&root=gcc&view=rev
Log:
    PR target/58864
    * dojump.c (save_pending_stack_adjust, restore_pending_stack_adjust):
    New functions.
    * expr.h (struct saved_pending_stack_adjust): New type.
    (save_pending_stack_adjust, restore_pending_stack_adjust): New
    prototypes.
    * optabs.c (emit_conditional_move): Call save_pending_stack_adjust
    and get_last_insn before do_pending_stack_adjust, call
    restore_pending_stack_adjust after delete_insns_since.
    * expr.c (expand_expr_real_2): Don't call do_pending_stack_adjust
    before calling emit_conditional_move.
    * expmed.c (expand_sdiv_pow2): Likewise.
    * calls.c (expand_call): Use {save,restore}_pending_stack_adjust.

    * g++.dg/opt/pr58864.C: New test.

Added:
    trunk/gcc/testsuite/g++.dg/opt/pr58864.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/calls.c
    trunk/gcc/dojump.c
    trunk/gcc/expmed.c
    trunk/gcc/expr.c
    trunk/gcc/expr.h
    trunk/gcc/optabs.c
    trunk/gcc/testsuite/ChangeLog

[Bug target/58864] [4.8/4.9 regression] ICE in connect_traces, at dwarf2cfi.c:NNNN

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58864

--- Comment #8 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Author: jakub
Date: Tue Dec  3 07:42:05 2013
New Revision: 205620

URL: http://gcc.gnu.org/viewcvs?rev=205620&root=gcc&view=rev
Log:
    PR target/58864
    * optabs.c (emit_conditional_move): Save and restore
    pending_stack_adjust and stack_pointer_delta if cmove can't be used.

    * g++.dg/opt/pr58864.C: New test.

Modified:
    branches/gcc-4_8-branch/gcc/ChangeLog
    branches/gcc-4_8-branch/gcc/optabs.c
    branches/gcc-4_8-branch/gcc/testsuite/ChangeLog

[Bug target/58864] [4.8/4.9 regression] ICE in connect_traces, at dwarf2cfi.c:NNNN

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58864

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

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

--- Comment #9 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed.

[Bug target/58864] [4.8/4.9 regression] ICE in connect_traces, at dwarf2cfi.c:NNNN

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58864

--- Comment #10 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Author: jakub
Date: Sat Dec  7 15:43:48 2013
New Revision: 205784

URL: http://gcc.gnu.org/viewcvs?rev=205784&root=gcc&view=rev
Log:
    PR target/58864
    * optabs.c (emit_conditional_move): Save and restore
    pending_stack_adjust and stack_pointer_delta if cmove can't be used.

    * g++.dg/opt/pr58864.C: New test.

Added:
    branches/gcc-4_8-branch/gcc/testsuite/g++.dg/opt/pr58864.C

[Bug target/58864] [4.8/4.9 regression] ICE in connect_traces, at dwarf2cfi.c:NNNN

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58864

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pinskia at gcc dot gnu.org

--- Comment #11 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
This was introduced by my patch to use emit_conditional_move in expr.c some
more.  emit_conditional_add has a similar bug and I will submit patch too.

[Bug target/58864] [4.8/4.9 regression] ICE in connect_traces, at dwarf2cfi.c:NNNN

[d.glazachev at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58864

Denis <d.glazachev at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |d.glazachev at gmail dot com

--- Comment #12 from Denis <d.glazachev at gmail dot com> ---
This or similar bug still exists in 4.9.3
Can be worked around by -fno-defer-pop