[Bug tree-optimization/23234] New: ICE in verify_flow_info()

[Bug tree-optimization/23234] New: ICE in verify_flow_info()

[uttamp at us dot ibm dot com]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1462 bytes --]

Following test program causes an internal compiler error,
$ cat y.c

double func ( double a, double b, double c)
{
  double x0, x1, y0, y1;
  int type = 0;
  double ar;

  if(b==0.0){
      x0 = -c/a;
      return( x0 > 1.0? 1.0 : ( x0 < 0.0 ? 0.0: x0));
    }
  if(a==0.0) {
      y0 = -c/b;
      return( y0 > 1.0? 1.0 : ( y0 < 0.0 ? 0.0: y0));
    }


  x0 = -c/a;
  y0 = -c/b;

  type = ((y0>=1.0)&&(y1>=1.0) ? 16 : type);

  switch(type){
  case 3: ar = (0.5*(y0+y1)); break;
  }

   return ar;

}

$ /opt/mline-20050802/bin/gcc -c -ffast-math -O2 y.c
$ y.c: In function âfuncâ:
y.c:2: error: control flow in the middle of basic block 0
y.c:2: internal compiler error: verify_flow_info failed
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.

This test case works with gcc_cvs code till 20050801. This is also a reason to 
fail eon benchmark.

-- 
           Summary: ICE in verify_flow_info()
           Product: gcc
           Version: 4.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: tree-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: uttamp at us dot ibm dot com
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: powerpc-linux
  GCC host triplet: powerpc-linux
GCC target triplet: powerpc-linu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23234

[Bug tree-optimization/23234] [4.1 Regression] ICE in verify_flow_info()

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-08-04 17:55 -------
Confirmed, reduced testcase:
double func ( double a, double b, double c)
{
  double y0;
  if(a==0.0) {
      y0 = -c/b;
      return y0;
    }
  y0 = -c/b;
  return y0;
}

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pinskia at gcc dot gnu dot
                   |                            |org
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
  GCC build triplet|powerpc-linux               |
   GCC host triplet|powerpc-linux               |
 GCC target triplet|powerpc-linu                |
           Keywords|                            |ice-on-valid-code
      Known to fail|                            |4.1.0
      Known to work|                            |4.0.0
   Last reconfirmed|0000-00-00 00:00:00         |2005-08-04 17:55:39
               date|                            |
            Summary|ICE in verify_flow_info()   |[4.1 Regression] ICE in
                   |                            |verify_flow_info()


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23234

[Bug tree-optimization/23234] [4.1 Regression] ICE in verify_flow_info()

[uttamp at us dot ibm dot com]

------- Additional Comments From uttamp at us dot ibm dot com  2005-08-04 17:57 -------
(In reply to comment #1)
> Confirmed, reduced testcase:
> double func ( double a, double b, double c)
> {
>   double y0;
>   if(a==0.0) {
>       y0 = -c/b;
>       return y0;
>     }
>   y0 = -c/b;
>   return y0;
> }

nice. ;)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23234

[Bug tree-optimization/23234] [4.1 Regression] ICE in verify_flow_info()

[pinskia at gcc dot gnu dot org]

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23234

[Bug tree-optimization/23234] [4.1 Regression] ICE in verify_flow_info()

[steven at gcc dot gnu dot org]

------- Additional Comments From steven at gcc dot gnu dot org  2005-08-07 15:19 -------
I'll take a look at this. 
 

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |steven at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED
   Last reconfirmed|2005-08-04 17:55:39         |2005-08-07 15:19:14
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23234

[Bug tree-optimization/23234] [4.1 Regression] ICE in verify_flow_info()

[steven at gcc dot gnu dot org]

------- Additional Comments From steven at gcc dot gnu dot org  2005-08-07 15:52 -------
Index: tree-ssa-math-opts.c 
=================================================================== 
RCS file: /cvs/gcc/gcc/gcc/tree-ssa-math-opts.c,v 
retrieving revision 2.4 
diff -u -p -r2.4 tree-ssa-math-opts.c 
--- tree-ssa-math-opts.c        1 Aug 2005 08:58:25 -0000       2.4 
+++ tree-ssa-math-opts.c        7 Aug 2005 15:52:07 -0000 
@@ -56,14 +56,15 @@ gate_cse_reciprocals (void) 
 
 /* Check if DEF's uses include more than one floating-point division, 
    and if so replace them by multiplications with the reciprocal.  If 
-   PHI is true, insert the reciprocal calculation before BSI, otherwise 
-   insert it after and move BSI to the new statement. 
+   BEFORE_BSI is true, insert the reciprocal calculation before BSI, 
+   otherwise insert it after and move BSI to the new statement. 
 
    Does not check the type of DEF, nor that DEF is a GIMPLE register. 
    This is done in the caller for speed, because otherwise this routine 
    would be called for every definition and phi node.  */ 
 static void 
-execute_cse_reciprocals_1 (block_stmt_iterator *bsi, tree def, bool phi) 
+execute_cse_reciprocals_1 (block_stmt_iterator *bsi, 
+                          tree def, bool before_bsi) 
 { 
   use_operand_p use_p; 
   imm_use_iterator use_iter; 
@@ -99,7 +100,7 @@ execute_cse_reciprocals_1 (block_stmt_it 
                     fold_build2 (RDIV_EXPR, type, build_real (type, dconst1), 
                                  def)); 
 
-  if (phi) 
+  if (before_bsi) 
     bsi_insert_before (bsi, new_stmt, BSI_SAME_STMT); 
   else 
     bsi_insert_after (bsi, new_stmt, BSI_NEW_STMT); 
@@ -133,7 +134,7 @@ execute_cse_reciprocals (void) 
        { 
          block_stmt_iterator bsi; 
          bsi = bsi_start (single_succ (ENTRY_BLOCK_PTR)); 
-          execute_cse_reciprocals_1 (&bsi, default_def (arg), false); 
+          execute_cse_reciprocals_1 (&bsi, default_def (arg), true); 
        } 
 
   FOR_EACH_BB (bb) 
 

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23234

[Bug tree-optimization/23234] [4.1 Regression] ICE in verify_flow_info()

[steven at gcc dot gnu dot org]

------- Additional Comments From steven at gcc dot gnu dot org  2005-08-07 16:13 -------
That patch from comment #4 is also not right.  See the URL for something 
that I expect to be correct (fingers crossed while it's trying to pass 
testing... ;-) 
 
 

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                URL|                            |http://gcc.gnu.org/ml/gcc-
                   |                            |patches/2005-
                   |                            |08/msg00402.html


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23234

[Bug tree-optimization/23234] [4.1 Regression] ICE in verify_flow_info()

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-08-09 03:28 -------
Subject: Bug 23234

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	steven@gcc.gnu.org	2005-08-09 03:28:38

Modified files:
	gcc            : ChangeLog tree-ssa-math-opts.c 
	gcc/testsuite  : ChangeLog 
Added files:
	gcc/testsuite/gcc.dg/tree-ssa: pr23234.c 

Log message:
	gcc/
	PR tree-optimization/23234
	* tree-ssa-math-opts.c (place_reciprocal): New enum.
	(execute_cse_reciprocals_1): Replace the 'phi' argument with an
	argument of the new enum.
	(execute_cse_reciprocals): Add reciprocals for function arguments
	on the unique successor edge of the entry block.  Update other calls
	to execute_cse_reciprocals_1.
	
	testsuite/
	PR tree-optimization/23234
	* gcc.dg/tree-ssa/pr23234.c: New test.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&r1=2.9683&r2=2.9684
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/tree-ssa-math-opts.c.diff?cvsroot=gcc&r1=2.4&r2=2.5
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&r1=1.5895&r2=1.5896
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.dg/tree-ssa/pr23234.c.diff?cvsroot=gcc&r1=NONE&r2=1.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23234

[Bug tree-optimization/23234] [4.1 Regression] ICE in verify_flow_info()

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-08-09 04:36 -------
Fixed.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23234