[Bug c++/62091] New: ice in before_dom_children

[Bug c++/62091] New: ice in before_dom_children

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62091

            Bug ID: 62091
           Summary: ice in before_dom_children
           Product: gcc
           Version: 4.9.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dcb314 at hotmail dot com

Created attachment 33291
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33291&action=edit
C++ source code

For the attached code, compiled by gcc trunk dated 20140810,
the following crash message is produced

$ ../results/bin/gcc -c -O2 bug159.cc
snmpmsg.cpp: In member function ‘int
Snmp_pp::SnmpMessage::unload(Snmp_pp::Pdu&, Snmp_pp::OctetStr&,
Snmp_pp::snmp_version&, Snmp_pp::OctetStr*, Snmp_pp::OctetStr*, long int*,
Snmp_pp::UdpAddress*, Snmp_pp::Snmp*)’:
snmpmsg.cpp:540:5: internal compiler error: in before_dom_children, at
tree-ssa-pre.c:4411
0xbd423c eliminate_dom_walker::before_dom_children(basic_block_def*)
    ../../src/trunk/gcc/tree-ssa-pre.c:4411
0xf59987 dom_walker::walk(basic_block_def*)
    ../../src/trunk/gcc/domwalk.c:177
0xbd1332 eliminate
    ../../src/trunk/gcc/tree-ssa-pre.c:4526
0xbd253c execute
    ../../src/trunk/gcc/tree-ssa-pre.c:4941
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.
$  

Compiler flag -O2 required.
>From gcc-bugs-return-458173-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Mon Aug 11 16:16:40 2014
Return-Path: <gcc-bugs-return-458173-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 6109 invoked by alias); 11 Aug 2014 16:16:40 -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 6062 invoked by uid 48); 11 Aug 2014 16:16:36 -0000
From: "trippels at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/62091] ice in before_dom_children
Date: Mon, 11 Aug 2014 16:16:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: tree-optimization
X-Bugzilla-Version: 4.10.0
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: trippels at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 4.10.0
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields: bug_status cf_reconfirmed_on cc component version target_milestone everconfirmed
Message-ID: <bug-62091-4-hHkUh0hghD@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-62091-4@http.gcc.gnu.org/bugzilla/>
References: <bug-62091-4@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: 2014-08/txt/msg00670.txt.bz2
Content-length: 1286

https://gcc.gnu.org/bugzilla/show_bug.cgi?idb091

Markus Trippelsdorf <trippels at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-08-11
                 CC|                            |hubicka at gcc dot gnu.org,
                   |                            |trippels at gcc dot gnu.org
          Component|c++                         |tree-optimization
            Version|4.9.2                       |4.10.0
   Target Milestone|---                         |4.10.0
     Ever confirmed|0                           |1

--- Comment #1 from Markus Trippelsdorf <trippels at gcc dot gnu.org> ---
Started with r213739.

markus@x4 tmp % cat bug159.cc
class SnmpSyntax
{
public:
  virtual SnmpSyntax *m_fn1 () const;
  ~SnmpSyntax () {}
  virtual SnmpSyntax &operator=(const SnmpSyntax &);
};

class A : public SnmpSyntax
{
public:
  A (int);
  SnmpSyntax *m_fn1 () const {}
  SnmpSyntax &operator=(const SnmpSyntax &);
};
int a;
void fn1 ()
{
  for (;; a++)
    switch (0)
    case 0:
      {
        A b (0);
        SnmpSyntax &c = b;
        c.m_fn1 ();
      }
}

[Bug tree-optimization/62091] [4.10 Regression] ice in before_dom_children

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62091

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|ice in before_dom_children  |[4.10 Regression] ice in
                   |                            |before_dom_children

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
4432                  else
4433                    gcc_assert (!ipa_intraprocedural_devirtualization
(stmt));

Honza - these should always be gcc_checking_assert ()s - there is no need to
ICE (and in fact it shouldn't even be a gcc_checking_assert in any tree but
yours used for testing).

[Bug tree-optimization/62091] [5 Regression] ice in before_dom_children

[hubicka at ucw dot cz]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62091

--- Comment #3 from Jan Hubicka <hubicka at ucw dot cz> ---
> Honza - these should always be gcc_checking_assert ()s - there is no need to
> ICE (and in fact it shouldn't even be a gcc_checking_assert in any tree but
> yours used for testing).

Yep, I plan to remove the old intraprocedural code very soon - I am basically
looking for testcases like these.  I went through firefox/testsuite/Qt builds
myself, but wanted to know if there are more side cases.

Honza

[Bug tree-optimization/62091] [5 Regression] ice in before_dom_children

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62091

--- Comment #6 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Author: hubicka
Date: Thu Aug 14 23:36:22 2014
New Revision: 213991

URL: https://gcc.gnu.org/viewcvs?rev=213991&root=gcc&view=rev
Log:

    PR tree-optimization/62091
    * g++.dg/ipa/devirt-38.C: New testcase.
    * tree-ssa-alias.c (walk_aliased_vdefs_1): Do not clear
    function_entry_reached.
    (walk_aliased_vdefs): Clear it here.
    * ipa-devirt.c (check_stmt_for_type_change): Handle static storage.

Added:
    trunk/gcc/testsuite/g++.dg/ipa/devirt-38.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/ipa-devirt.c
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/tree-ssa-alias.c

[Bug tree-optimization/62091] [5 Regression] ice in before_dom_children

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62091

--- Comment #7 from David Binderman <dcb314 at hotmail dot com> ---
I am not sure this is fixed. Attached code, when
compiled by 20140817, with flags -std=gnu++11 -O2,
does the following

$ ../results/bin/gcc -c -std=gnu++11 -O2 bug161.cc
listformatter.cpp: In member function ‘icu_53::UnicodeString&
icu_53::ListFormatter::format(const icu_53::UnicodeString*, int32_t,
icu_53::UnicodeString&, int32_t, int32_t&, UErrorCode&) const’:
listformatter.cpp:304:16: internal compiler error: in before_dom_children, at
tree-ssa-pre.c:4410
0xbecf2c eliminate_dom_walker::before_dom_children(basic_block_def*)
    ../../src/trunk/gcc/tree-ssa-pre.c:4410
0xf72c67 dom_walker::walk(basic_block_def*)
    ../../src/trunk/gcc/domwalk.c:177
0xbe9fd2 eliminate
    ../../src/trunk/gcc/tree-ssa-pre.c:4525
0xbeb1dc execute
    ../../src/trunk/gcc/tree-ssa-pre.c:4940
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-458786-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Tue Aug 19 15:21:53 2014
Return-Path: <gcc-bugs-return-458786-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 10983 invoked by alias); 19 Aug 2014 15:21:52 -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 10913 invoked by uid 48); 19 Aug 2014 15:21:46 -0000
From: "vp at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/55334] [4.8/4.9/5 Regression] mgrid regression (ipa-cp disables vectorization)
Date: Tue, 19 Aug 2014 15:21:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: tree-optimization
X-Bugzilla-Version: 4.8.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: vp at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Priority: P2
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 4.9.2
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields: cc
Message-ID: <bug-55334-4-RyRL9J9xOt@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-55334-4@http.gcc.gnu.org/bugzilla/>
References: <bug-55334-4@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: 2014-08/txt/msg01283.txt.bz2
Content-length: 889

https://gcc.gnu.org/bugzilla/show_bug.cgi?idU334

vp at gcc dot gnu.org changed:

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

--- Comment #37 from vp at gcc dot gnu.org ---
I see this issue in AArch64 as well on 4.9 branch and on the trunk causing
performance regression.

Since MG3P is inlined at MAIN__, all the four calls to RESID are in MAIN__. Two
of which are to the clone resid_.constprop.1 (not vectorized) and one to resid_
(vectorized) and the last one inlined (not vectorized and it looks very similar
to the resid_.constprop.1 version).

I could however, improve by passing -fno-inline-functions and -fno-ipa-cp at
-Ofast (so that all the calls are to resid_ which gets vectorized).

Anyone working on this issue?

[Bug tree-optimization/62091] [5 Regression] ice in before_dom_children

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62091

--- Comment #8 from David Binderman <dcb314 at hotmail dot com> ---
Created attachment 33362
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33362&action=edit
C++ source code

[Bug tree-optimization/62091] [5 Regression] ice in before_dom_children

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62091

--- Comment #9 from Jason Merrill <jason at gcc dot gnu.org> ---
Author: jason
Date: Tue Aug 19 18:25:45 2014
New Revision: 214181

URL: https://gcc.gnu.org/viewcvs?rev=214181&root=gcc&view=rev
Log:
    PR tree-optimization/62091
    * decl2.c (decl_needed_p): Return true for virtual functions when
    devirtualizing.

Modified:
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/decl2.c
    trunk/gcc/testsuite/g++.dg/ipa/devirt-38.C

[Bug tree-optimization/62091] [5 Regression] ice in before_dom_children

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62091

Jason Merrill <jason at gcc dot gnu.org> changed:

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

--- Comment #10 from Jason Merrill <jason at gcc dot gnu.org> ---
My patch doesn't fix the ICE in comment 7, just devirtualization on
devirt-38.C.

[Bug tree-optimization/62091] [5 Regression] ice in before_dom_children

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62091

--- Comment #11 from Jason Merrill <jason at gcc dot gnu.org> ---
Author: jason
Date: Wed Aug 20 01:54:40 2014
New Revision: 214208

URL: https://gcc.gnu.org/viewcvs?rev=214208&root=gcc&view=rev
Log:
    PR c++/61214
    PR tree-optimization/62091
    * decl2.c (decl_needed_p): Return true for virtual functions when
    devirtualizing.

Added:
    branches/gcc-4_9-branch/gcc/testsuite/g++.dg/ipa/devirt-39.C
Modified:
    branches/gcc-4_9-branch/gcc/cp/ChangeLog
    branches/gcc-4_9-branch/gcc/cp/decl2.c

[Bug tree-optimization/62091] [5 Regression] ice in before_dom_children

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62091

--- Comment #12 from Jason Merrill <jason at gcc dot gnu.org> ---
Author: jason
Date: Wed Aug 20 02:03:56 2014
New Revision: 214209

URL: https://gcc.gnu.org/viewcvs?rev=214209&root=gcc&view=rev
Log:
    PR tree-optimization/62091
    * g++.dg/ipa/devirt-39.C: New.

Added:
    trunk/gcc/testsuite/g++.dg/ipa/devirt-39.C

[Bug tree-optimization/62091] [5 Regression] ice in before_dom_children

[hubicka at ucw dot cz]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62091

--- Comment #13 from Jan Hubicka <hubicka at ucw dot cz> ---
Thanks, Jason!
This is patch I am testing that cures the testcase from Comment #7. I would
apprechiate if someone could reduce it - my simple attempts has failed and I am
bit in hurry this week.

The problem here is that get_dynamic_class gets into placement new paranoia
because
it manages to mismatch polymorphic_type[2] with polymorphic_type.

Honza

Index: ipa-devirt.c
===================================================================
--- ipa-devirt.c    (revision 214225)
+++ ipa-devirt.c    (working copy)
@@ -2846,6 +2891,18 @@ ipa_polymorphic_call_context::get_dynami
   bool function_entry_reached = false;
   tree instance_ref = NULL;
   gimple stmt = call;
+  /* Remember OFFSET before it is modified by restrict_to_inner_class.
+     This is because we do not update INSTANCE when walking inwards.  */
+  HOST_WIDE_INT instance_offset = offset;
+
+  /* Walk into inner type. This may clear maybe_derived_type and save us
+     from useless work.  It also makes later comparsions with static type
+     easier.  */
+  if (outer_type)
+    {
+      if (!restrict_to_inner_class (otr_type))
+        return false;
+    }

   if (!maybe_in_construction && !maybe_derived_type)
     return false;

[Bug tree-optimization/62091] [5 Regression] ice in before_dom_children

[trippels at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62091

--- Comment #14 from Markus Trippelsdorf <trippels at gcc dot gnu.org> ---
Reduced testcase for  Comment #7:

typedef enum
{
} UErrorCode;
class UnicodeString
{
public:
  UnicodeString ();
  virtual ~UnicodeString ();
};
class A
{
  UnicodeString &m_fn1 (UnicodeString &, int &p2, UErrorCode &) const;
};
UnicodeString::UnicodeString () {}
UnicodeString &
A::m_fn1 (UnicodeString &, int &p2, UErrorCode &) const
{
  UnicodeString a[2];
}

[Bug tree-optimization/62091] [5 Regression] ice in before_dom_children

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62091

--- Comment #15 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Author: hubicka
Date: Thu Aug 21 13:49:07 2014
New Revision: 214271

URL: https://gcc.gnu.org/viewcvs?rev=214271&root=gcc&view=rev
Log:

    PR tree-optimization/62091
    * g++.dg/ipa/devirt-37.C: Update template.
    * g++.dg/ipa/devirt-40.C: New testcase.
    * ipa-devirt.c (ipa_polymorphic_call_context::restrict_to_inner_type):
    handle correctly arrays.
    (extr_type_from_vtbl_ptr_store): Add debug output; handle multiple
    inheritance binfos.
    (record_known_type): Walk into inner type.
    (ipa_polymorphic_call_context::get_dynamic_type): Likewise; strenghten
    condition on no type changes.

Added:
    trunk/gcc/testsuite/g++.dg/ipa/devirt-40.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/ipa-devirt.c
    trunk/gcc/lto/lto.c
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/g++.dg/ipa/devirt-37.C

[Bug tree-optimization/62091] [5 Regression] ice in before_dom_children

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62091

Jan Hubicka <hubicka at gcc dot gnu.org> changed:

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

--- Comment #16 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Fixed.