[Bug tree-optimization/61518] New: wrong code (by tree vectorizer) at -O3 on x86_64-linux-gnu

[Bug tree-optimization/61518] New: wrong code (by tree vectorizer) at -O3 on x86_64-linux-gnu

[su at cs dot ucdavis.edu]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61518

            Bug ID: 61518
           Summary: wrong code (by tree vectorizer) at -O3 on
                    x86_64-linux-gnu
           Product: gcc
           Version: 4.10.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: su at cs dot ucdavis.edu

The current gcc trunk miscompiles the following testcase on x86_64-linux at -O3
in both 32-bit and 64-bit modes.  

This is a regression from 4.9.x.  

This appears to be a tree vectorizer bug as it goes away with
-fno-tree-vectorize. 


$ gcc-trunk -v
Using built-in specs.
COLLECT_GCC=gcc-trunk
COLLECT_LTO_WRAPPER=/usr/local/gcc-trunk/libexec/gcc/x86_64-unknown-linux-gnu/4.10.0/lto-wrapper
Target: x86_64-unknown-linux-gnu
Configured with: ../gcc-trunk/configure --prefix=/usr/local/gcc-trunk
--enable-languages=c,c++ --disable-werror --enable-multilib
Thread model: posix
gcc version 4.10.0 20140615 (experimental) [trunk revision 211685] (GCC) 
$ 
$ gcc-trunk -O2 small.c; a.out
$ gcc-4.9 -O3 small.c; a.out
$ gcc-trunk -O3 -fno-tree-vectorize small.c; a.out
$ 
$ gcc-trunk -O3 small.c
$ a.out
Aborted (core dumped)
$ 


---------------------------


int a, b, c[1], d, e, f;

void
fn1 ()
{
  for (; d < 1; d++)
    {
      if (b) 
    {
      a = e++ && f; 
      b = f; 
    }
      c[b] = 0;
    }
}

int
main ()
{
  fn1 ();

  if (e != 0) 
    __builtin_abort (); 

  return 0;
}

[Bug tree-optimization/61518] [4.10 Regression] wrong code (by tree vectorizer) at -O3 on x86_64-linux-gnu

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61518

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-06-16
                 CC|                            |jakub at gcc dot gnu.org,
                   |                            |kyukhin at gcc dot gnu.org
   Target Milestone|---                         |4.10.0
            Summary|wrong code (by tree         |[4.10 Regression] wrong
                   |vectorizer) at -O3 on       |code (by tree vectorizer)
                   |x86_64-linux-gnu            |at -O3 on x86_64-linux-gnu
     Ever confirmed|0                           |1

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Started with r209972.

[Bug tree-optimization/61518] [4.10 Regression] wrong code (by tree vectorizer) at -O3 on x86_64-linux-gnu

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61518

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |wrong-code
           Priority|P3                          |P1

[Bug tree-optimization/61518] [4.10 Regression] wrong code (by tree vectorizer) at -O3 on x86_64-linux-gnu

[ysrumyan at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61518

Yuri Rumyantsev <ysrumyan at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ysrumyan at gmail dot com

--- Comment #2 from Yuri Rumyantsev <ysrumyan at gmail dot com> ---
This is my fault - forgot to check that reduction variable is used in reduction
statement or in phi-function only. I will provide a patch after required
testing completion.

[Bug tree-optimization/61518] [4.10 Regression] wrong code (by tree vectorizer) at -O3 on x86_64-linux-gnu

[kyukhin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61518

--- Comment #3 from Kirill Yukhin <kyukhin at gcc dot gnu.org> ---
Author: kyukhin
Date: Wed Jun 18 11:40:59 2014
New Revision: 211780

URL: https://gcc.gnu.org/viewcvs?rev=211780&root=gcc&view=rev
Log:
        PR tree-optimization/61518
gcc/
        * tree-if-conv.c (is_cond_scalar_reduction): Add missed check that
        reduction var is used in reduction stmt or phi-function only.

gcc/testsuite/
        * gcc.dg/torture/pr61518.c: New test.


Added:
    trunk/gcc/testsuite/gcc.dg/torture/pr61518.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/tree-if-conv.c

[Bug tree-optimization/61518] [4.10 Regression] wrong code (by tree vectorizer) at -O3 on x86_64-linux-gnu

[mikpelinux at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61518

--- Comment #4 from Mikael Pettersson <mikpelinux at gmail dot com> ---
Isn't this fixed now?

[Bug tree-optimization/61518] [5 Regression] wrong code (by tree vectorizer) at -O3 on x86_64-linux-gnu

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61518

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

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

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Assuming yes, the testcase is in the testsuite after all.