[Bug c/55770] New: when compiling code with -fwhole-program and -fipa-matrix-reorg,compiler will produce wrong code

[Bug c/55770] New: when compiling code with -fwhole-program and -fipa-matrix-reorg,compiler will produce wrong code

[hcj20080469 at 163 dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55770

             Bug #: 55770
           Summary: when compiling code with -fwhole-program and
                    -fipa-matrix-reorg,compiler will produce wrong code
    Classification: Unclassified
           Product: gcc
           Version: 4.7.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: hcj20080469@163.com


compiler version :gcc 4.7.1
--build=i586-suse-linux 
--host=i586-suse-linux
--target=i586-target-linux-gnu

test case test.c:

#include <string.h>

void __attribute__((noinline)) mem_init (void);
int **vel;
int *p;
int main (int argc, char **argv)
{
  int i, j, k;
  mem_init();
  p=malloc(4);
  *p=123;
 printf("p1:%d\n",*p);
  return 0;
}

void __attribute__((noinline))
mem_init (void)
{

  int i, j, k,d;

  d = 0;
  vel = (int **) malloc(2*sizeof(int *));
  for(i=0;i<2;i++)
    vel[i]=(int *)malloc(4*sizeof(int));
  for(i=0;i<2;i++)
        for(j=0;j<4;j++)
      vel[i][j]=i+j;

}


with command compile code
i586-target-linux-gnu-gcc test.c   -fno-ipa-matrix-reorg -fdump-tree-optimized 
-O3 -fwhole-program -fno-tree-fre -lm -S -o test-2.s

i586-target-linux-gnu-gcc test-2.s -o test-2.exe


assembling file test-2.s content following:

mem_init:
.LFB34:
    subl    $28, %esp
    movl    $8, (%esp)
    movl    $16, T.6
    movl    $32, T.7
    call    malloc # get 8 byte in the heap space 
    movl    %eax, vel
    movl    $0, (%eax)
    movl    $1, 4(%eax)
    movl    $2, 8(%eax)
    movl    $3, 12(%eax)
    movl    $1, 16(%eax)
    movl    $2, 20(%eax)
    movl    $3, 24(%eax)
    movl    $4, 28(%eax)# wring 32 byte date to the heap space
    addl    $28, %esp
    .cfi_def_cfa_offset 4
    ret
    .cfi_endproc

run the test-2.exe , give error  infomation
test-1.exe: malloc.c:3096: sYSMALLOc: Assertion `(old_top == (((mbinptr)
(((char *) &((av)->bins[((1) - 1) * 2])) - _0x1) && ((unsigned long)old_end &
pagemask) == 0)' failed.
Aborted

[Bug tree-optimization/55770] when compiling code with -fwhole-program and -fipa-matrix-reorg,compiler will produce wrong code

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55770

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
          Component|c                           |tree-optimization
         Resolution|                            |WONTFIX
   Target Milestone|---                         |4.8.0

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> 2012-12-21 02:02:27 UTC ---
The option has been disabled for 4.8 and above:
r190298 | rguenth | 2012-08-10 07:19:09 -0700 (Fri, 10 Aug 2012) | 16 lines

2012-08-10  Richard Guenther  <rguenther@suse.de>

        * Makefile.in (OBJS): Remove matrix-reorg.o.
        (matrix-reorg.o): Remove dependence rule.
        (GTFILES): Remove matrix-reorg.c.
        * matrix-reorg.c: Remove.
        * passes.c (init_optimization_passes): Do not schedule
        pass_ipa_matrix_reorg.
        * tree-pass.h (pass_ipa_matrix_reorg): Remove.
        * common.opt (fipa-matrix-reorg): Stub out.
        * doc/invoke.texi (fipa-matrix-reorg): Remove documentation.

        * gcc.dg/matrix/*.c: Adjust and move ...
        * gcc.dg/torture/: ... here.
        * gcc.dg/matrix: Remove directory.


 So closing as won't fix.

[Bug tree-optimization/55770] when compiling code with -fwhole-program and -fipa-matrix-reorg,compiler will produce wrong code

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55770

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> 2012-12-21 10:12:24 UTC ---
Yeah, it's very broken - don't use it.