[Bug c/23944] New: segv in fold_indirect_ref_1 in fold-const.c:11505

[Bug c/23944] New: segv in fold_indirect_ref_1 in fold-const.c:11505

[marcus at jet dot franken dot de]

testcase below (extract from blender) causes a segfault:  
  
LANG=C projects/gcc/BIN/bin/gcc -c constraint.i   
constraint.i: In function 'f':  
constraint.i:4: internal compiler error: Segmentation fault  
 
gdb shows:  
#0  fold_indirect_ref_1 (type=0x2aaaaaea0e70, op0=0x0) 
at /home/marcus/projects/gcc/gcc/fold-const.c:11505 
11505     STRIP_NOPS (sub);

-- 
           Summary: segv in fold_indirect_ref_1 in fold-const.c:11505
           Product: gcc
           Version: 4.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: marcus at jet dot franken dot de
                CC: gcc-bugs at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23944

[Bug c/23944] segv in fold_indirect_ref_1 in fold-const.c:11505

[marcus at jet dot franken dot de]

------- Additional Comments From marcus at jet dot franken dot de  2005-09-18 11:41 -------
Created an attachment (id=9760)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=9760&action=view)
constraint.i

gcc -c constraint.i

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23944

[Bug c/23944] [4.1 Regression] segv in fold_indirect_ref_1 in fold-const.c:11505

[marcus at jet dot franken dot de]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|segv in fold_indirect_ref_1 |[4.1 Regression] segv in
                   |in fold-const.c:11505       |fold_indirect_ref_1 in fold-
                   |                            |const.c:11505


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23944

[Bug c/23944] [4.1 Regression] segv in fold_indirect_ref_1 in fold-const.c:11505

[rguenth at gcc dot gnu dot org]

------- Additional Comments From rguenth at gcc dot gnu dot org  2005-09-18 12:09 -------
The testcase looks like

float f(float src[][4])
{
 return *(src[3]);
}

and the C frontend produces

 <return_expr 0x40191240
    type <void_type 0x4019c8a0 void VOID
        align 8 symtab 0 alias set -1
        pointer_to_this <pointer_type 0x4019c8fc>>
    side-effects
    arg 0 <modify_expr 0x40193168
        type <real_type 0x4019ca10 float SF
            size <integer_cst 0x4018c408 constant invariant 32>
            unit size <integer_cst 0x4018c198 constant invariant 4>
            align 32 symtab 0 alias set -1 precision 32
            pointer_to_this <pointer_type 0x4019cb24>>
        side-effects
        arg 0 <var_decl 0x40198108 D.1279 type <real_type 0x4019ca10 float>
            used ignored SF file /suse/rguenther/src/tests/pr23944.c line 3 size
<integer_cst 0x4018c408 32> unit size <integer_cst 0x4018c198 4>
            align 32 context <function_decl 0x4020d100 f>>
        arg 1 <array_ref 0x4020ea28 type <real_type 0x4019ca10 float>
           
            arg 0 <indirect_ref 0x40191260 type <array_type 0x4020c398>
               >
            arg 1 <integer_cst 0x4018c1b0 constant invariant 0>>>
    /suse/rguenther/src/tests/pr23944.c:3>

Closer look at the INDIRECT_REF node reveals:

 <indirect_ref 0x40191260
    type <array_type 0x4020c398
        type <real_type 0x4019ca10 float SF
            size <integer_cst 0x4018c408 constant invariant 32>
            unit size <integer_cst 0x4018c198 constant invariant 4>
            align 32 symtab 0 alias set -1 precision 32
            pointer_to_this <pointer_type 0x4019cb24>>
        BLK
        size <integer_cst 0x4018c678 constant invariant 128>
        unit size <integer_cst 0x4018c690 constant invariant 16>
        align 32 symtab 0 alias set -1
        domain <integer_type 0x4020c33c type <integer_type 0x4019c000 unsigned int>
            SI size <integer_cst 0x4018c408 32> unit size <integer_cst 0x4018c198 4>
            align 32 symtab 0 alias set -1 precision 32 min <integer_cst
0x4018c1b0 0> max <integer_cst 0x4018c378 3>>
        pointer_to_this <pointer_type 0x4020c450>>
   >

well - there's something missing here.  Frontend bug.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-on-valid-code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23944

[Bug c/23944] [4.1 Regression] segv in fold_indirect_ref_1 in fold-const.c:11505

[rguenth at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2005-09-18 12:09:23
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23944

[Bug middle-end/23944] [4.1 Regression] segv in fold_indirect_ref_1 in fold-const.c:11505

[rguenth at gcc dot gnu dot org]

------- Additional Comments From rguenth at gcc dot gnu dot org  2005-09-18 12:17 -------
But this one builds the corrupt node:

#0  build1_stat (code=INDIRECT_REF, type=0x4020c398, node=0x0, 
    _loc_name=0x86dbfd0 "../../../src/gcc-unpatched/gcc/gimplify.c", 
    _loc_line=2955, _loc_function=0x86dc518 "fold_indirect_ref_rhs")
    at tree.c:2705
#1  0x0811b85c in fold_indirect_ref_rhs (t=0x40191200) at gimplify.c:2955
#2  0x0811baaf in gimplify_modify_expr_rhs (expr_p=0xbff365a0, 
    from_p=0x40193188, to_p=0x40193184, pre_p=0xbff362c4, post_p=0xbff362c0, 
    want_value=0 '\0') at gimplify.c:2988

I'll look at this.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |rguenth at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED
          Component|c                           |middle-end
   Last reconfirmed|2005-09-18 12:09:23         |2005-09-18 12:17:53
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23944

[Bug middle-end/23944] [4.1 Regression] segv in fold_indirect_ref_1 in fold-const.c:11505

[rguenth at gcc dot gnu dot org]

------- Additional Comments From rguenth at gcc dot gnu dot org  2005-09-18 12:24 -------
I have a patch for which after gimplification we end up with

f (src)
{
  float D.1279;
  float[4] * D.1280;

  D.1280 = src + 48B;
  D.1279 = (*D.1280)[0];
  return D.1279;
}


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23944

[Bug middle-end/23944] [4.1 Regression] segv in fold_indirect_ref_1 in fold-const.c:11505

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-09-18 14:15 -------
Subject: Bug 23944

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	rguenth@gcc.gnu.org	2005-09-18 14:15:20

Modified files:
	gcc            : ChangeLog gimplify.c 
	gcc/testsuite  : ChangeLog 
Added files:
	gcc/testsuite/gcc.c-torture/compile: pr23944.c 

Log message:
	2005-09-18  Richard Guenther  <rguenther@suse.de>
	
	PR middle-end/23944
	* gimplify.c (fold_indirect_ref_rhs): Fix thinko in
	fallback.
	
	* gcc.c-torture/compile/pr23944.c: New testcase.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.9973&r2=2.9974
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/gimplify.c.diff?cvsroot=gcc&r1=2.152&r2=2.153
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&r1=1.6076&r2=1.6077
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.c-torture/compile/pr23944.c.diff?cvsroot=gcc&r1=NONE&r2=1.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23944

[Bug middle-end/23944] [4.1 Regression] segv in fold_indirect_ref_1 in fold-const.c:11505

[rguenth at gcc dot gnu dot org]

------- Additional Comments From rguenth at gcc dot gnu dot org  2005-09-18 14:16 -------
Fixed.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23944

[Bug middle-end/23944] [4.1 Regression] segv in fold_indirect_ref_1 in fold-const.c:11505

[rguenth at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.1.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23944