[Bug lto/65844] New: [5 Regression] ICE (verify_cgraph_node failed) on i686-linux-gnu

[Bug lto/65844] New: [5 Regression] ICE (verify_cgraph_node failed) on i686-linux-gnu

[doko at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65844

            Bug ID: 65844
           Summary: [5 Regression] ICE (verify_cgraph_node failed) on
                    i686-linux-gnu
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: lto
          Assignee: unassigned at gcc dot gnu.org
          Reporter: doko at gcc dot gnu.org

seen with 5.1 rc2 on i686-linux-gnu, works on x86_64, arm, ppc64le, arm64,
powerpc.

building python 3.5 alpha 4 with lto and pgo (using the testsuite to collect
the perf data):

lto1: error: indirect call frequency 1001 does not match BB frequency 1000
bytes_dealloc/5157 (bytes_dealloc) @0xf6b660e4
  Type: function definition analyzed
  Visibility: prevailing_def_ironly
  previous sharing asm name: 57969
  Address is taken.
  References: 
  Referring: PyBytes_Type/5273 (addr)_textiowrapper_writeflush/54957 (addr)
(speculative)code_dealloc/6404 (addr)
(speculative)_io_FileIO___init___impl/53523 (addr)
(speculative)match_dealloc/48926 (addr) (speculative)os_stat/46510 (addr)
(speculative)PyBuffer_Release/59115 (addr) (speculative)bytesio_dealloc/59792
(addr) (speculative)_Pickler_ClearBuffer/39048 (addr)
(speculative)s_dealloc/37732 (addr) (speculative)assemble_free/22984 (addr)
(speculative)code_dealloc/6404 (addr) (speculative)decode_unicode.isra.8/21159
(addr) (speculative)
  Availability: available
  Profile id: 1495503281
  First run: 402
  Function flags: body icf_merged hot
  Called by: 
  Calls: 
   Indirect call(1.00 per call) 
lto1: internal compiler error: verify_cgraph_node failed
0x82a24dc cgraph_node::verify_node()
    ../../src/gcc/cgraph.c:3151
0x8296e88 symtab_node::verify()
    ../../src/gcc/symtab.c:1103
0x82982a4 symtab_node::verify_symtab_nodes()
    ../../src/gcc/symtab.c:1123
0x84cd553 symbol_table::remove_unreachable_nodes(_IO_FILE*)
    ../../src/gcc/ipa.c:686
0x8596895 execute_todo
    ../../src/gcc/passes.c:2025
Please submit a full bug report,
with preprocessed source if appropriate.

[Bug lto/65844] [5/6 Regression] ICE (verify_cgraph_node failed) on i686-linux-gnu

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65844

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-checking
                 CC|                            |hubicka at gcc dot gnu.org
            Version|5.0                         |5.1.0
   Target Milestone|---                         |5.2
            Summary|[5 Regression] ICE          |[5/6 Regression] ICE
                   |(verify_cgraph_node failed) |(verify_cgraph_node failed)
                   |on i686-linux-gnu           |on i686-linux-gnu

[Bug lto/65844] [5/6 Regression] ICE (verify_cgraph_node failed) on i686-linux-gnu

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65844

Martin Liška <marxin at gcc dot gnu.org> changed:

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

--- Comment #1 from Martin Liška <marxin at gcc dot gnu.org> ---
Hello.

Is it possible to reproduce the ICE on x86_64 by adding '-m32' to {C,LD}FLAGS.

I face following SEGFAULT (with -fprofile-generate -lto -m32) in:

gdb --args ./Programs/_freeze_importlib ./Lib/importlib/_bootstrap.py
Python/importlib.h


(gdb) r
Starting program:
/home/marxin/Programming/Python-3.5.0a4/Programs/_freeze_importlib
./Lib/importlib/_bootstrap.py Python/importlib.h
Got object file from memory but can't read symbols: File truncated.
Missing separate debuginfos, use: zypper install
glibc-32bit-debuginfo-2.19-16.9.1.x86_64
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".

Program received signal SIGSEGV, Segmentation fault.
va_build_value (flags=0, va=0xffffcb34 ":\321\071\bt_9\bt_9\b\340\215T\b",
format=0x839d142 "(szz)") at Python/modsupport.c:455
455            Py_VA_COPY(lva, va);
Missing separate debuginfos, use: zypper install
libgcc_s1-32bit-debuginfo-4.8.3+r212056-2.2.4.x86_64
(gdb) bt
#0  va_build_value (flags=0, va=0xffffcb34 ":\321\071\bt_9\bt_9\b\340\215T\b",
format=0x839d142 "(szz)") at Python/modsupport.c:455
#1  Py_BuildValue (format=0x839d142 "(szz)") at Python/modsupport.c:420
#2  0x0814c1d7 in _PySys_Init () at ./Python/sysmodule.c:1713
#3  0x0819dad0 in _Py_InitializeEx_Private (install_sigs=1,
install_importlib=0) at Python/pylifecycle.c:377
#4  0x08365fee in main (argc=3, argv=0xffffcd74) at
Programs/_freeze_importlib.c:83

Thanks,
Martin
>From gcc-bugs-return-484440-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Thu Apr 23 11:06:26 2015
Return-Path: <gcc-bugs-return-484440-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 117794 invoked by alias); 23 Apr 2015 11:06:26 -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 117471 invoked by uid 48); 23 Apr 2015 11:06:22 -0000
From: "ubizjak at gmail dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug rtl-optimization/65857] New: combine won't generate zero-extend from HImode memory
Date: Thu, 23 Apr 2015 11:06:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: new
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: rtl-optimization
X-Bugzilla-Version: 6.0
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: ubizjak at gmail 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: bug_id short_desc product version bug_status bug_severity priority component assigned_to reporter cf_gcctarget
Message-ID: <bug-65857-4@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: 2015-04/txt/msg01992.txt.bz2
Content-length: 1824

https://gcc.gnu.org/bugzilla/show_bug.cgi?ide857

            Bug ID: 65857
           Summary: combine won't generate zero-extend from HImode memory
           Product: gcc
           Version: 6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: rtl-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: ubizjak at gmail dot com
            Target: x86

Following testcase:

--cut here--
extern unsigned char *a;

int test (void)
{
  return (unsigned char) a[0] | (unsigned char) a[1] << 8;
}
--cut here--

compiles (-O2 )nicely to:

        movq    a(%rip), %rax
        movzwl  (%rax), %eax
        ret

However, combine will not combine HImode load with zero-extend:

(insn 5 2 6 2 (set (reg/f:DI 92 [ a ])
        (mem/f/c:DI (symbol_ref:DI ("a") [flags 0x40]  <var_decl 0x2b006b8dfc60
a>) [1 a+0 S8 A64])) 85 {*movdi_internal}
     (nil))
(insn 6 5 11 2 (set (reg:SI 91 [ D.1849 ])
        (zero_extend:SI (mem:HI (reg/f:DI 92 [ a ]) [0 MEM[(unsigned char
*)_2]+0 S2 A8]))) ins.c:5 135 {*zero_extendhisi2}
     (expr_list:REG_DEAD (reg/f:DI 92 [ a ])
        (nil)))

It just chickens out with:

Trying 5 -> 6:
Failed to match this instruction:
(set (reg:SI 91 [ D.1849 ])
    (zero_extend:SI (mem:HI (mem/f/c:DI (symbol_ref:DI ("a") [flags 0x40]
<var_decl 0x2b006b8dfc60 a>) [1 a+0 S8 A64]) [0 MEM[(unsigned char *)_2]+0 S2 A
8])))

Come on, combine ... the pattern is defined as:

(define_insn "*zero_extend<mode>si2"
  [(set (match_operand:SI 0 "register_operand" "=r")
    (zero_extend:SI
      (match_operand:SWI12 1 "nonimmediate_operand" "<r>m")))]
  "!(TARGET_ZERO_EXTEND_WITH_AND && optimize_function_for_speed_p (cfun))"
  "movz{<imodesuffix>l|x}\t{%1, %0|%0, %1}"
  [(set_attr "type" "imovx")
   (set_attr "mode" "SI")])

[Bug ipa/65844] [5/6 Regression] ICE (verify_cgraph_node failed) on i686-linux-gnu

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65844

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |lto
             Status|UNCONFIRMED                 |WAITING
   Last reconfirmed|                            |2015-06-30
          Component|lto                         |ipa
     Ever confirmed|0                           |1

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
Waiting for a testcase.

[Bug ipa/65844] [5/6 Regression] ICE (verify_cgraph_node failed) on i686-linux-gnu

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65844

--- Comment #3 from Martin Liška <marxin at gcc dot gnu.org> ---
Even with fixed configure script, I am unable to reproduce the issue on
x86_64-linux-pc with -m32 option added.

Martin
>From gcc-bugs-return-491142-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Wed Jul 01 10:41:12 2015
Return-Path: <gcc-bugs-return-491142-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 121621 invoked by alias); 1 Jul 2015 10:41:12 -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 121308 invoked by uid 48); 1 Jul 2015 10:41:08 -0000
From: "charlie at charliedyson dot net" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/66717] In variable declaration, decltype incorrectly deduces return type of function returning const reference to int
Date: Wed, 01 Jul 2015 10:41:00 -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: 5.1.0
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: charlie at charliedyson dot net
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution:
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:
Message-ID: <bug-66717-4-OlBjByZZ2U@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-66717-4@http.gcc.gnu.org/bugzilla/>
References: <bug-66717-4@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: 2015-07/txt/msg00032.txt.bz2
Content-length: 367

https://gcc.gnu.org/bugzilla/show_bug.cgi?idf717

--- Comment #2 from Charlie <charlie at charliedyson dot net> ---
I think you may be right that it's only the diagnostic that's wrong. The
following compiles fine https://goo.gl/eyuv17

  static int foo = 1;
  decltype (const_ref_to_int ()) x = foo;
  static_assert (std::is_same<decltype (x), const int&>(), "");

[Bug ipa/65844] [5/6 Regression] ICE (verify_cgraph_node failed) on i686-linux-gnu

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65844

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|5.2                         |5.3

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 5.2 is being released, adjusting target milestone to 5.3.