[Bug target/66648] New: [4.9/5/6 regression] incorrect memcpy expansion with unrolled_loop strategy at -O2

[Bug target/66648] New: [4.9/5/6 regression] incorrect memcpy expansion with unrolled_loop strategy at -O2

[ebotcazou at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66648

            Bug ID: 66648
           Summary: [4.9/5/6 regression] incorrect memcpy expansion with
                    unrolled_loop strategy at -O2
           Product: gcc
           Version: 6.0
            Status: UNCONFIRMED
          Keywords: wrong-code
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: ebotcazou at gcc dot gnu.org
  Target Milestone: ---
            Target: x86_64-*-linux

Created attachment 35840
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=35840&action=edit
Testcase

One of the numerous paths of ix86_expand_set_or_movmem using the unrolled_loop
strategy can generate wrong code (buffer overflow) at -O2, when VRP is able to
compute a minimum copying size.

Compile and run the attached testcase with -O2
-mstringop-strategy=unrolled_loop -mtune=nocona on a x86-64/Linux machine.

[Bug target/66648] [4.9/5/6 regression] incorrect memcpy expansion with unrolled_loop strategy at -O2

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66648

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.9.3

[Bug target/66648] [4.9/5/6 regression] incorrect memcpy expansion with unrolled_loop strategy at -O2

[ubizjak at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66648

Uroš Bizjak <ubizjak at gmail dot com> changed:

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

--- Comment #1 from Uroš Bizjak <ubizjak at gmail dot com> ---
Adding CC.
>From gcc-bugs-return-490065-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Wed Jun 24 12:34:01 2015
Return-Path: <gcc-bugs-return-490065-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 84543 invoked by alias); 24 Jun 2015 12:34:01 -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 84483 invoked by uid 48); 24 Jun 2015 12:33:56 -0000
From: "olegendo at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/66611] SH: ICE on -O2
Date: Wed, 24 Jun 2015 12:34:00 -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: 6.0
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: olegendo at gcc dot gnu.org
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: bug_status cf_reconfirmed_on everconfirmed
Message-ID: <bug-66611-4-Uvuf3Q11Wb@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-66611-4@http.gcc.gnu.org/bugzilla/>
References: <bug-66611-4@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: 2015-06/txt/msg02397.txt.bz2
Content-length: 588

https://gcc.gnu.org/bugzilla/show_bug.cgi?idf611

Oleg Endo <olegendo at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-06-24
     Ever confirmed|0                           |1

--- Comment #2 from Oleg Endo <olegendo at gcc dot gnu.org> ---
Sorry, not the same issue as PR 66358.  It seems more like another bug of PR
65979, which is a buggy peephole optimization.

[Bug target/66648] [4.9/5/6 regression] incorrect memcpy expansion with unrolled_loop strategy at -O2

[ubizjak at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66648

Uroš Bizjak <ubizjak at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-06-26
                 CC|                            |ubizjak at gmail dot com
     Ever confirmed|0                           |1

--- Comment #2 from Uroš Bizjak <ubizjak at gmail dot com> ---
Confirmed, looking into it.
>From gcc-bugs-return-490263-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Fri Jun 26 10:31:14 2015
Return-Path: <gcc-bugs-return-490263-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 111822 invoked by alias); 26 Jun 2015 10:31:13 -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 111715 invoked by uid 55); 26 Jun 2015 10:31:08 -0000
From: "uros at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug rtl-optimization/66412] [5/6 Regression] ICE on valid code at -O2 and -O3 with -g enabled in simplify_subreg, at simplify-rtx.c:5748
Date: Fri, 26 Jun 2015 10:31:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
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: uros at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution:
X-Bugzilla-Priority: P2
X-Bugzilla-Assigned-To: segher at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 5.2
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-66412-4-RO5IrEQWFU@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-66412-4@http.gcc.gnu.org/bugzilla/>
References: <bug-66412-4@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: 2015-06/txt/msg02595.txt.bz2
Content-length: 748

https://gcc.gnu.org/bugzilla/show_bug.cgi?idf412

--- Comment #10 from uros at gcc dot gnu.org ---
Author: uros
Date: Fri Jun 26 10:30:37 2015
New Revision: 225006

URL: https://gcc.gnu.org/viewcvs?rev"5006&root=gcc&view=rev
Log:
        PR target/66412
        * config/i386/i386.md (various splitters): Use shallow_copy_rtx
        before doing PUT_MODE or PUT_CODE on operands to avoid
        in-place RTX modification.

testsuite/ChangeLog:

        PR target/66412
        * gcc.target/i386/pr66412.c: New test.


Added:
    branches/gcc-5-branch/gcc/testsuite/gcc.target/i386/pr66412.c
Modified:
    branches/gcc-5-branch/gcc/ChangeLog
    branches/gcc-5-branch/gcc/config/i386/i386.md
    branches/gcc-5-branch/gcc/testsuite/ChangeLog

[Bug target/66648] [4.9/5/6 regression] incorrect memcpy expansion with unrolled_loop strategy at -O2

[ubizjak at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66648

--- Comment #3 from Uroš Bizjak <ubizjak at gmail dot com> ---
Untested patch:

--cut here--
Index: config/i386/i386.c
===================================================================
--- config/i386/i386.c  (revision 224993)
+++ config/i386/i386.c  (working copy)
@@ -25008,7 +25008,8 @@ ix86_expand_set_or_movmem (rtx dst, rtx src, rtx c
       dst = change_address (dst, BLKmode, destreg);
       set_mem_align (dst, desired_align * BITS_PER_UNIT);
       epilogue_size_needed = 0;
-      if (need_zero_guard && !min_size)
+      if (need_zero_guard
+         && min_size < (unsigned HOST_WIDE_INT) size_needed)
        {
          /* It is possible that we copied enough so the main loop will not
             execute.  */
@@ -25140,7 +25141,7 @@ ix86_expand_set_or_movmem (rtx dst, rtx src, rtx c
          max_size -= align_bytes;
        }
       if (need_zero_guard
-         && !min_size
+         && min_size < (unsigned HOST_WIDE_INT) size_needed
          && (count < (unsigned HOST_WIDE_INT) size_needed
              || (align_bytes == 0
                  && count < ((unsigned HOST_WIDE_INT) size_needed
--cut here--

We have to emit check for main loop execution UNLESS we guarantee that min_size
is at least equal to size_needed. In this testcase, VRP declared minimum
copying size, and after unaligned prologue adjustment, min_size was still
non-zero. The !min_size check didn't account for the case that min_size can be
non-zero, but still less than size_needed.
>From gcc-bugs-return-490272-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Fri Jun 26 12:18:55 2015
Return-Path: <gcc-bugs-return-490272-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 30765 invoked by alias); 26 Jun 2015 12:18:54 -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 30699 invoked by uid 48); 26 Jun 2015 12:18:51 -0000
From: "antonio.poggiali at datalogic dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/66666] ARM wrong copy constructor address on multiple inheritance
Date: Fri, 26 Jun 2015 12:18:00 -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: 4.9.1
X-Bugzilla-Keywords:
X-Bugzilla-Severity: major
X-Bugzilla-Who: antonio.poggiali at datalogic dot com
X-Bugzilla-Status: UNCONFIRMED
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-66666-4-WwIavDszkJ@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-66666-4@http.gcc.gnu.org/bugzilla/>
References: <bug-66666-4@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: 2015-06/txt/msg02604.txt.bz2
Content-length: 189

https://gcc.gnu.org/bugzilla/show_bug.cgi?idf666

--- Comment #5 from Antonio Poggiali <antonio.poggiali at datalogic dot com> ---
Sorry but this comment is incorrect. See the next one.

[Bug target/66648] [4.9/5/6 regression] incorrect memcpy expansion with unrolled_loop strategy at -O2

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66648

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
GCC 4.9.3 has been released.

[Bug target/66648] [4.9/5/6 regression] incorrect memcpy expansion with unrolled_loop strategy at -O2

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66648

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.9.3                       |4.9.4

[Bug target/66648] [4.9/5/6 regression] incorrect memcpy expansion with unrolled_loop strategy at -O2

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66648

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2

[Bug target/66648] [4.9/5/6 regression] incorrect memcpy expansion with unrolled_loop strategy at -O2

[uros at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66648

--- Comment #6 from uros at gcc dot gnu.org ---
Author: uros
Date: Sun Aug 16 17:50:20 2015
New Revision: 226925

URL: https://gcc.gnu.org/viewcvs?rev=226925&root=gcc&view=rev
Log:
        Backport from mainline:
        2015-07-25  Uros Bizjak  <ubizjak@gmail.com>

        PR target/66648
        * config/i386/i386.c (ix86_expand_set_or_movmem): Emit main loop
        execution guard when min_size is less than size_needed.

testsuite/ChangeLog:

        Backport from mainline:
        2015-07-25  Uros Bizjak  <ubizjak@gmail.com>

        PR target/66648
        * gcc.target/i386/pr66648.c: New test.


Added:
    branches/gcc-5-branch/gcc/testsuite/gcc.target/i386/pr66648.c
Modified:
    branches/gcc-5-branch/gcc/ChangeLog
    branches/gcc-5-branch/gcc/config/i386/i386.c
    branches/gcc-5-branch/gcc/testsuite/ChangeLog

[Bug target/66648] [4.9/5/6 regression] incorrect memcpy expansion with unrolled_loop strategy at -O2

[uros at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66648

--- Comment #7 from uros at gcc dot gnu.org ---
Author: uros
Date: Sun Aug 16 17:52:51 2015
New Revision: 226926

URL: https://gcc.gnu.org/viewcvs?rev=226926&root=gcc&view=rev
Log:
        Backport from mainline:
        2015-07-25  Uros Bizjak  <ubizjak@gmail.com>

        PR target/66648
        * config/i386/i386.c (ix86_expand_set_or_movmem): Emit main loop
        execution guard when min_size is less than size_needed.

testsuite/ChangeLog:

        Backport from mainline:
        2015-07-25  Uros Bizjak  <ubizjak@gmail.com>

        PR target/66648
        * gcc.target/i386/pr66648.c: New test.


Added:
    branches/gcc-4_9-branch/gcc/testsuite/gcc.target/i386/pr66648.c
Modified:
    branches/gcc-4_9-branch/gcc/ChangeLog
    branches/gcc-4_9-branch/gcc/config/i386/i386.c
    branches/gcc-4_9-branch/gcc/testsuite/ChangeLog

[Bug target/66648] [4.9/5/6 regression] incorrect memcpy expansion with unrolled_loop strategy at -O2

[ubizjak at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66648

Uroš Bizjak <ubizjak at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|x86_64-*-linux              |x86
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED

--- Comment #8 from Uroš Bizjak <ubizjak at gmail dot com> ---
Fixed everywhere.
>From gcc-bugs-return-494917-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Sun Aug 16 18:01:30 2015
Return-Path: <gcc-bugs-return-494917-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 84379 invoked by alias); 16 Aug 2015 18:01:30 -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 84325 invoked by uid 48); 16 Aug 2015 18:01:26 -0000
From: "Joost.VandeVondele at mat dot ethz.ch" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug fortran/52846] [F2008] Support submodules
Date: Sun, 16 Aug 2015 18:01:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: fortran
X-Bugzilla-Version: 4.8.0
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: Joost.VandeVondele at mat dot ethz.ch
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution:
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: pault at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-52846-4-VX6T7OewPL@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-52846-4@http.gcc.gnu.org/bugzilla/>
References: <bug-52846-4@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: 2015-08/txt/msg01059.txt.bz2
Content-length: 848

https://gcc.gnu.org/bugzilla/show_bug.cgi?idR846

--- Comment #26 from Joost VandeVondele <Joost.VandeVondele at mat dot ethz.ch> ---
One more datapoint on the .smod / .mod issue.

I ran into the first package that fails to build with 6.0 because of this.
libxc (version 2.2.2, Tobias Burnus actually being one of the authors), relies
on a configure macro named ax_cv_f90_modext this macro now leads in the
Makefile to :

[...]
am__untar = $${TAR-tar} xf -
ax_cv_f90_modext = mod
smod
bindir = ${exec_prefix}/bin
[...]

which leads to an obvious:> make -j
Makefile:319: *** missing separator.  Stop.

While arguably this is a bug in the ax_cv_f90_modext macro, it might be
indicative of possible problems 'in the wild'.

However, I do like the idea of having only public info in the .mod, and hence
improving the recompilation cascade problem.

[Bug target/66648] [4.9/5/6 regression] incorrect memcpy expansion with unrolled_loop strategy at -O2

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66648

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |zsojka at seznam dot cz

--- Comment #9 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
*** Bug 63590 has been marked as a duplicate of this bug. ***