public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug fortran/28439]  New: Multiple evaluations of arithmetic if condition
@ 2006-07-19 16:59 mfpasik at sandia dot gov
  2006-07-19 20:33 ` [Bug fortran/28439] " kargl at gcc dot gnu dot org
                   ` (12 more replies)
  0 siblings, 13 replies; 14+ messages in thread
From: mfpasik at sandia dot gov @ 2006-07-19 16:59 UTC (permalink / raw)
  To: gcc-bugs

The expression/condition in an arithmetic if statement appears to be evaluated
multiple times when it is zero or negative. This is particularly bad when the
expression is a function that depends on input or has state associated with it.

Example:

      program lll

      integer myfunc

      do 99 i=1,10
      if (myfunc(0)) 10,20,30

 10   write (6,*) 'negative'
      goto 99

 20   write (6,*) 'zero'
      goto 99

 30   write (6,*) 'positive'
      goto 99

 99   continue
      end


      integer function myfunc(int)
      integer value
      read (5,*) value
      myfunc = value
      return
      end

Sample input/output:

1
 positive
0
3
 zero
-1
3
 zero
4
 positive
0
0
 zero

Workaround:

Introduce a temporary expression for the if test condition/expression or
convert obsolete arithmetic if to an if-then-else statement.


-- 
           Summary: Multiple evaluations of arithmetic if condition
           Product: gcc
           Version: 4.1.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: mfpasik at sandia dot gov
 GCC build triplet: i686-pc-linux-gnu
  GCC host triplet: i686-pc-linux-gnu
GCC target triplet: i686-pc-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28439


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [Bug fortran/28439] Multiple evaluations of arithmetic if condition
  2006-07-19 16:59 [Bug fortran/28439] New: Multiple evaluations of arithmetic if condition mfpasik at sandia dot gov
@ 2006-07-19 20:33 ` kargl at gcc dot gnu dot org
  2006-07-19 23:03 ` steven at gcc dot gnu dot org
                   ` (11 subsequent siblings)
  12 siblings, 0 replies; 14+ messages in thread
From: kargl at gcc dot gnu dot org @ 2006-07-19 20:33 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #1 from kargl at gcc dot gnu dot org  2006-07-19 20:33 -------
Yep.  The dump-tree-original contains


     if (myfunc (&C.1243) <= 0)
       {
         if (myfunc (&C.1243) < 0) goto __label_000010;
         else goto __label_000020;
       }
     else
       {
          goto __label_000030;
       }


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28439


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [Bug fortran/28439] Multiple evaluations of arithmetic if condition
  2006-07-19 16:59 [Bug fortran/28439] New: Multiple evaluations of arithmetic if condition mfpasik at sandia dot gov
  2006-07-19 20:33 ` [Bug fortran/28439] " kargl at gcc dot gnu dot org
@ 2006-07-19 23:03 ` steven at gcc dot gnu dot org
  2006-07-19 23:03 ` steven at gcc dot gnu dot org
                   ` (10 subsequent siblings)
  12 siblings, 0 replies; 14+ messages in thread
From: steven at gcc dot gnu dot org @ 2006-07-19 23:03 UTC (permalink / raw)
  To: gcc-bugs



-- 

steven at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2006-07-19 23:03:44
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28439


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [Bug fortran/28439] Multiple evaluations of arithmetic if condition
  2006-07-19 16:59 [Bug fortran/28439] New: Multiple evaluations of arithmetic if condition mfpasik at sandia dot gov
  2006-07-19 20:33 ` [Bug fortran/28439] " kargl at gcc dot gnu dot org
  2006-07-19 23:03 ` steven at gcc dot gnu dot org
@ 2006-07-19 23:03 ` steven at gcc dot gnu dot org
  2006-07-19 23:06 ` pinskia at gcc dot gnu dot org
                   ` (9 subsequent siblings)
  12 siblings, 0 replies; 14+ messages in thread
From: steven at gcc dot gnu dot org @ 2006-07-19 23:03 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #2 from steven at gcc dot gnu dot org  2006-07-19 23:03 -------
Wrong code. Bad.


-- 

steven at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |major
           Keywords|                            |wrong-code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28439


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [Bug fortran/28439] Multiple evaluations of arithmetic if condition
  2006-07-19 16:59 [Bug fortran/28439] New: Multiple evaluations of arithmetic if condition mfpasik at sandia dot gov
                   ` (2 preceding siblings ...)
  2006-07-19 23:03 ` steven at gcc dot gnu dot org
@ 2006-07-19 23:06 ` pinskia at gcc dot gnu dot org
  2006-07-19 23:08 ` steven at gcc dot gnu dot org
                   ` (8 subsequent siblings)
  12 siblings, 0 replies; 14+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2006-07-19 23:06 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #3 from pinskia at gcc dot gnu dot org  2006-07-19 23:06 -------
The easiest way to fix this is most likely wrap the function call in a
SAVE_EXPR but I could be wrong.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28439


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [Bug fortran/28439] Multiple evaluations of arithmetic if condition
  2006-07-19 16:59 [Bug fortran/28439] New: Multiple evaluations of arithmetic if condition mfpasik at sandia dot gov
                   ` (3 preceding siblings ...)
  2006-07-19 23:06 ` pinskia at gcc dot gnu dot org
@ 2006-07-19 23:08 ` steven at gcc dot gnu dot org
  2006-07-20  4:37 ` kargl at gcc dot gnu dot org
                   ` (7 subsequent siblings)
  12 siblings, 0 replies; 14+ messages in thread
From: steven at gcc dot gnu dot org @ 2006-07-19 23:08 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #4 from steven at gcc dot gnu dot org  2006-07-19 23:08 -------
Andrew, you're wrong.  :-P

My gfortran skills are not what they used to be, but I believe this is the fix:

Index: trans-stmt.c
===================================================================
--- trans-stmt.c        (revision 115467)
+++ trans-stmt.c        (working copy)
@@ -663,6 +663,7 @@ gfc_trans_arithmetic_if (gfc_code * code

   /* Pre-evaluate COND.  */
   gfc_conv_expr_val (&se, code->expr);
+  se.expr = gfc_evaluate_now (se.expr, &se.pre);

   /* Build something to compare with.  */
   zero = gfc_build_const (TREE_TYPE (se.expr), integer_zero_node);


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28439


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [Bug fortran/28439] Multiple evaluations of arithmetic if condition
  2006-07-19 16:59 [Bug fortran/28439] New: Multiple evaluations of arithmetic if condition mfpasik at sandia dot gov
                   ` (4 preceding siblings ...)
  2006-07-19 23:08 ` steven at gcc dot gnu dot org
@ 2006-07-20  4:37 ` kargl at gcc dot gnu dot org
  2006-07-21 23:39 ` steven at gcc dot gnu dot org
                   ` (6 subsequent siblings)
  12 siblings, 0 replies; 14+ messages in thread
From: kargl at gcc dot gnu dot org @ 2006-07-20  4:37 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #5 from kargl at gcc dot gnu dot org  2006-07-20 04:37 -------
(In reply to comment #4) 
> My gfortran skills are not what they used to be,

You're always welcomed to spend some quality time with the
gfortran source code :-)

> but I believe this is the fix:
> 
> Index: trans-stmt.c
> ===================================================================
> --- trans-stmt.c        (revision 115467)
> +++ trans-stmt.c        (working copy)
> @@ -663,6 +663,7 @@ gfc_trans_arithmetic_if (gfc_code * code
> 
>    /* Pre-evaluate COND.  */
>    gfc_conv_expr_val (&se, code->expr);
> +  se.expr = gfc_evaluate_now (se.expr, &se.pre);
> 

I think you're on the right track, but shouldn't it be

se.expr = gfc_evaluate_now (code->expr, &se.pre);

I'm not sure of the &se.pre part, yet.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28439


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [Bug fortran/28439] Multiple evaluations of arithmetic if condition
  2006-07-19 16:59 [Bug fortran/28439] New: Multiple evaluations of arithmetic if condition mfpasik at sandia dot gov
                   ` (5 preceding siblings ...)
  2006-07-20  4:37 ` kargl at gcc dot gnu dot org
@ 2006-07-21 23:39 ` steven at gcc dot gnu dot org
  2006-07-22  2:01 ` kargl at gcc dot gnu dot org
                   ` (5 subsequent siblings)
  12 siblings, 0 replies; 14+ messages in thread
From: steven at gcc dot gnu dot org @ 2006-07-21 23:39 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #6 from steven at gcc dot gnu dot org  2006-07-21 23:39 -------
The definition of gfc_evaluate_now is:
trans.h:tree gfc_evaluate_now (tree, stmtblock_t *);

And the effect of adding &se.pre is to add the code to the block. You can see
this going on everywhere in gfortran with a simple grep in the source code.

So the use "se.expr = gfc_evaluate_now (code->expr, &se.pre);" is correct. I
actually even tested it, you know ;-)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28439


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [Bug fortran/28439] Multiple evaluations of arithmetic if condition
  2006-07-19 16:59 [Bug fortran/28439] New: Multiple evaluations of arithmetic if condition mfpasik at sandia dot gov
                   ` (6 preceding siblings ...)
  2006-07-21 23:39 ` steven at gcc dot gnu dot org
@ 2006-07-22  2:01 ` kargl at gcc dot gnu dot org
  2006-07-22 17:55 ` kargl at gcc dot gnu dot org
                   ` (4 subsequent siblings)
  12 siblings, 0 replies; 14+ messages in thread
From: kargl at gcc dot gnu dot org @ 2006-07-22  2:01 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #7 from kargl at gcc dot gnu dot org  2006-07-22 02:01 -------
Steven,

I believe you're correct.  I misread your patch where I thought
you were replacing the line above the one you added.  Sorry about
the noise.  I'll test and commit the patch.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28439


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [Bug fortran/28439] Multiple evaluations of arithmetic if condition
  2006-07-19 16:59 [Bug fortran/28439] New: Multiple evaluations of arithmetic if condition mfpasik at sandia dot gov
                   ` (7 preceding siblings ...)
  2006-07-22  2:01 ` kargl at gcc dot gnu dot org
@ 2006-07-22 17:55 ` kargl at gcc dot gnu dot org
  2006-07-22 17:58 ` [Bug fortran/28439] [4.1 only] " kargl at gcc dot gnu dot org
                   ` (3 subsequent siblings)
  12 siblings, 0 replies; 14+ messages in thread
From: kargl at gcc dot gnu dot org @ 2006-07-22 17:55 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #8 from kargl at gcc dot gnu dot org  2006-07-22 17:55 -------
Subject: Bug 28439

Author: kargl
Date: Sat Jul 22 17:55:35 2006
New Revision: 115677

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=115677
Log:
2006-07-22  Steven Bosscher  <steven@gcc.gnu.org> 

        PR fortran/28439
        * trans-stmt.c (gfc_trans_arithmetic_if): Evaluate the condition once.


Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/trans-stmt.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28439


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [Bug fortran/28439] [4.1 only] Multiple evaluations of arithmetic if condition
  2006-07-19 16:59 [Bug fortran/28439] New: Multiple evaluations of arithmetic if condition mfpasik at sandia dot gov
                   ` (8 preceding siblings ...)
  2006-07-22 17:55 ` kargl at gcc dot gnu dot org
@ 2006-07-22 17:58 ` kargl at gcc dot gnu dot org
  2006-07-24 16:54 ` kargl at gcc dot gnu dot org
                   ` (2 subsequent siblings)
  12 siblings, 0 replies; 14+ messages in thread
From: kargl at gcc dot gnu dot org @ 2006-07-22 17:58 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #9 from kargl at gcc dot gnu dot org  2006-07-22 17:58 -------
Fixed on trunk.


-- 

kargl at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|Multiple evaluations of     |[4.1 only] Multiple
                   |arithmetic if condition     |evaluations of arithmetic if
                   |                            |condition


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28439


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [Bug fortran/28439] [4.1 only] Multiple evaluations of arithmetic if condition
  2006-07-19 16:59 [Bug fortran/28439] New: Multiple evaluations of arithmetic if condition mfpasik at sandia dot gov
                   ` (9 preceding siblings ...)
  2006-07-22 17:58 ` [Bug fortran/28439] [4.1 only] " kargl at gcc dot gnu dot org
@ 2006-07-24 16:54 ` kargl at gcc dot gnu dot org
  2006-07-25 23:31 ` kargl at gcc dot gnu dot org
  2006-07-25 23:32 ` kargl at gcc dot gnu dot org
  12 siblings, 0 replies; 14+ messages in thread
From: kargl at gcc dot gnu dot org @ 2006-07-24 16:54 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #10 from kargl at gcc dot gnu dot org  2006-07-24 16:54 -------
Subject: Bug 28439

Author: kargl
Date: Mon Jul 24 16:54:01 2006
New Revision: 115718

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=115718
Log:
2006-07-24  Steven G. Kargl  <kargls@comcast.net>

        PR fortran/28439
        * gfortran.dg/arithmetic_if.f90:  New test.


Added:
    trunk/gcc/testsuite/gfortran.dg/arithmetic_if.f90
Modified:
    trunk/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28439


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [Bug fortran/28439] [4.1 only] Multiple evaluations of arithmetic if condition
  2006-07-19 16:59 [Bug fortran/28439] New: Multiple evaluations of arithmetic if condition mfpasik at sandia dot gov
                   ` (10 preceding siblings ...)
  2006-07-24 16:54 ` kargl at gcc dot gnu dot org
@ 2006-07-25 23:31 ` kargl at gcc dot gnu dot org
  2006-07-25 23:32 ` kargl at gcc dot gnu dot org
  12 siblings, 0 replies; 14+ messages in thread
From: kargl at gcc dot gnu dot org @ 2006-07-25 23:31 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #11 from kargl at gcc dot gnu dot org  2006-07-25 23:31 -------
Subject: Bug 28439

Author: kargl
Date: Tue Jul 25 23:31:16 2006
New Revision: 115743

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=115743
Log:
2006-07-25  Steven G. Kargl  <kargls@comcast.net>

        PR fortran/28439
        *trans-stmt.c (gfc_trans_arithmetic_if):  Evaluate the condition once.

                gfortran.dg/arithmetic_if.f90: New test.


Added:
    branches/gcc-4_1-branch/gcc/testsuite/gfortran.dg/arithmetic_if.f90
Modified:
    branches/gcc-4_1-branch/gcc/fortran/ChangeLog
    branches/gcc-4_1-branch/gcc/fortran/trans-stmt.c
    branches/gcc-4_1-branch/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28439


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [Bug fortran/28439] [4.1 only] Multiple evaluations of arithmetic if condition
  2006-07-19 16:59 [Bug fortran/28439] New: Multiple evaluations of arithmetic if condition mfpasik at sandia dot gov
                   ` (11 preceding siblings ...)
  2006-07-25 23:31 ` kargl at gcc dot gnu dot org
@ 2006-07-25 23:32 ` kargl at gcc dot gnu dot org
  12 siblings, 0 replies; 14+ messages in thread
From: kargl at gcc dot gnu dot org @ 2006-07-25 23:32 UTC (permalink / raw)
  To: gcc-bugs



------- Comment #12 from kargl at gcc dot gnu dot org  2006-07-25 23:32 -------
Fixed in 4.1


-- 

kargl at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|---                         |4.1.2


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28439


^ permalink raw reply	[flat|nested] 14+ messages in thread

end of thread, other threads:[~2006-07-25 23:32 UTC | newest]

Thread overview: 14+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2006-07-19 16:59 [Bug fortran/28439] New: Multiple evaluations of arithmetic if condition mfpasik at sandia dot gov
2006-07-19 20:33 ` [Bug fortran/28439] " kargl at gcc dot gnu dot org
2006-07-19 23:03 ` steven at gcc dot gnu dot org
2006-07-19 23:03 ` steven at gcc dot gnu dot org
2006-07-19 23:06 ` pinskia at gcc dot gnu dot org
2006-07-19 23:08 ` steven at gcc dot gnu dot org
2006-07-20  4:37 ` kargl at gcc dot gnu dot org
2006-07-21 23:39 ` steven at gcc dot gnu dot org
2006-07-22  2:01 ` kargl at gcc dot gnu dot org
2006-07-22 17:55 ` kargl at gcc dot gnu dot org
2006-07-22 17:58 ` [Bug fortran/28439] [4.1 only] " kargl at gcc dot gnu dot org
2006-07-24 16:54 ` kargl at gcc dot gnu dot org
2006-07-25 23:31 ` kargl at gcc dot gnu dot org
2006-07-25 23:32 ` kargl at gcc dot gnu dot org

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).