[Bug ipa/60120] New: wrong code (for code with the optimize attribute) at -Os on x86_64-linux-gnu in 32-bit mode

[Bug ipa/60120] New: wrong code (for code with the optimize attribute) at -Os on x86_64-linux-gnu in 32-bit mode

[su at cs dot ucdavis.edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60120

            Bug ID: 60120
           Summary: wrong code (for code with the optimize attribute) at
                    -Os on x86_64-linux-gnu in 32-bit mode
           Product: gcc
           Version: 4.8.3
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: ipa
          Assignee: unassigned at gcc dot gnu.org
          Reporter: su at cs dot ucdavis.edu

The current gcc 4.8 branch miscompiles the following code that uses the
optimize attribute on x86_64-linux at -Os in 32-bit mode (but not 64-bit). 

This is a regression from gcc 4.7.x and doesn't affect the current gcc trunk. 

$ gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/home/su/software/local/gcc-4.8/bin/../libexec/gcc/x86_64-unknown-linux-gnu/4.8.3/lto-wrapper
Target: x86_64-unknown-linux-gnu
Configured with: ../gcc-4.8/configure --enable-languages=c,c++
Thread model: posix
gcc version 4.8.3 20140208 (prerelease) [gcc-4_8-branch revision 207632] (GCC) 
$
$ gcc -m32 -O1 small.c; a.out
2
$ gcc -m64 -Os small.c; a.out
2
$ gcc -m32 -Os small.c; a.out
1
$
$ gcc-4.7.3 -m32 -Os small.c; a.out
2
$ gcc-trunk -m32 -Os small.c; a.out
2
$ 


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


int printf (const char *, ...);

int a;

static int
fn1 (int p)
{
  return p + 1;
}

static 
__attribute__ ((optimize (1)))
int fn2 ()
{
  for (; a < 1; a = fn1 (1))
    ;
  return 0;
}

__attribute__ ((optimize (0)))
int fn3 ()
{
  fn1 (0);
  fn2 ();
  return 0;
}

int
main ()
{
  fn3 ();
  printf ("%d\n", a);
  return 0;
}

[Bug ipa/60120] wrong code (for code with the optimize attribute) at -Os on x86_64-linux-gnu in 32-bit mode

[su at cs dot ucdavis.edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60120

--- Comment #1 from Zhendong Su <su at cs dot ucdavis.edu> ---
Below is another testcase that very likely points to the same root cause, so I
also include it here.  Hope it's useful to help diagnose the issue. 

$ gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/home/su/software/local/gcc-4.8/bin/../libexec/gcc/x86_64-unknown-linux-gnu/4.8.3/lto-wrapper
Target: x86_64-unknown-linux-gnu
Configured with: ../gcc-4.8/configure --enable-languages=c,c++
Thread model: posix
gcc version 4.8.3 20140208 (prerelease) [gcc-4_8-branch revision 207632] (GCC) 
$ 
$ gcc -m32 -O1 small.c; a.out
$ gcc -m64 -Os small.c; a.out
$
$ gcc -m32 -Os small.c                       
$ timeout -s 9 10 a.out
Killed
$
$ gcc-4.7.3 -m32 -Os small.c; a.out
$ gcc-trunk -m32 -Os small.c; a.out
$ 


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


int a = 1;

static int
fn1 (int p1, short p2)
{
  return p2 == 0 ? 0 : p1 / p2;
}

int
fn2 (int p)
{
  return 0 ? 0 : p;
}

__attribute__ ((optimize (0)))
int fn3 ()
{
  if (fn1 (fn2 (72260), a))
    ;
  else
    for (;;)
      ;
  return 0;
}

int
main ()
{
  fn3 ();
  return 0;
}

[Bug ipa/60120] wrong code (for code with the optimize attribute) at -Os on x86_64-linux-gnu in 32-bit mode

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60120

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-02-10
     Ever confirmed|0                           |1

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
This looks like a dup of PR60062 to me - we call fn1 with the argument on the
stack but fn1 expects it in %eax.  Fixed on trunk and thus doesn't reproduce
there.

[Bug ipa/60120] wrong code (for code with the optimize attribute) at -Os on x86_64-linux-gnu in 32-bit mode

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60120

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

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

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
dup.

*** This bug has been marked as a duplicate of bug 60062 ***