[Bug fortran/97320] New: False positive "Array reference out of bounds in loop" in a protecting if block

[Bug fortran/97320] New: False positive "Array reference out of bounds in loop" in a protecting if block

[trnka at scm dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97320

            Bug ID: 97320
           Summary: False positive "Array reference out of bounds in loop"
                    in a protecting if block
           Product: gcc
           Version: 10.2.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: trnka at scm dot com
  Target Milestone: ---

Compiling the following testcase with "gfortran -c -Wdo-subscript
do-subscript-test.f90" leads to a bunch of warnings like the following:
do-subscript-test.f90:13:26:

    7 |    i_: do i = 1, 10
      |                   2       
......
   13 |                a(i-4) > a(i-5)) &
      |                          1
Warning: Array reference at (1) out of bounds (-4 < 1) in loop beginning at (2)
[-Wdo-subscript]

Given that all these accesses are in an "if (i > 5)" block, the warnings are
false positives and the out-of-bounds accesses cannot really occur. Inspecting
the .optimized output confirms this.

I can reproduce this using GCC 8 through 10 (GNU Fortran (GCC) 10.2.1 20200827
(Red Hat 10.2.1-3)), with or without -O2.

===== do-subscript-test.f90 =====
subroutine do_subscript_test()
   implicit none

   integer :: i
   real :: a(10)

   i_: do i = 1, 10
      if (i > 5) then
         if (a(i)   > a(i-1) .and. &
               a(i-1) > a(i-2) .and. &
               a(i-2) > a(i-3) .and. &
               a(i-3) > a(i-4) .and. &
               a(i-4) > a(i-5)) &
            exit i_
      end if
   end do i_
end subroutine

[Bug fortran/97320] False positive "Array reference out of bounds in loop" in a protecting if block

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97320

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |WAITING
   Last reconfirmed|                            |2020-10-07

--- Comment #1 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
The manual says

Warn if an array subscript inside a DO loop could lead to an out-of-bounds
access 
even if the compiler cannot prove that the statement is actually executed, 
in cases like

or don't use -Wdo-subscript if there is an IF in the DO block.

[Bug fortran/97320] False positive "Array reference out of bounds in loop" in a protecting if block

[Jacques.Lebourlot at obspm dot fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97320

Jacques Le Bourlot <Jacques.Lebourlot at obspm dot fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |Jacques.Lebourlot at obspm dot fr

--- Comment #2 from Jacques Le Bourlot <Jacques.Lebourlot at obspm dot fr> ---
Same here.

This is annoying because *some* warnings ARE relevant, and I need them to clean
a huge (and old) code. But they are lost in too many false positives.

Here is the test case I can provide (similar to Tomas')

-----------------------------
PROGRAM Warn_Bug

   INTEGER, PARAMETER      :: N = 2
   INTEGER, DIMENSION(0:N) :: a
   INTEGER                 :: i

   a = [ (i, i = 0, N) ]

   DO i = 0, N
      IF (i == 0) THEN
         PRINT *, " Do nothing"
      ELSE IF (i >= 1) THEN
         PRINT *, " OK:", a(i) - a(i-1)
      ENDIF
   ENDDO

END PROGRAM Warn_Bug
-----------------------------

The compiler should see that "IF (i >= 1)" THEN "i-1 >= 0". Quoting the manual
is not an adequate answer. An upgrade would be most welcomed.

Thank you.

Jacques

[Bug fortran/97320] False positive "Array reference out of bounds in loop" in a protecting if block

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97320

kargl at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |kargl at gcc dot gnu.org

--- Comment #3 from kargl at gcc dot gnu.org ---
(In reply to Jacques Le Bourlot from comment #2)
> 
> The compiler should see that "IF (i >= 1)" THEN "i-1 >= 0". Quoting the
> manual is not an adequate answer. An upgrade would be most welcomed.
> 

gfortran is mostly maintained by a diminishing number of contributors.
Messages telling the few remaining contributors what they ought to
do are likely to fall on deaf ears.  Patches from new contributors
are welcomed; otherwise, sometime in the future, perhaps a decade 
or so, someone might care enough to try to fix this annoyance.  In the 
meantime, add -Wno-do-subscript to your options.

[Bug fortran/97320] False positive "Array reference out of bounds in loop" in a protecting if block

[Jacques.Lebourlot at obspm dot fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97320

--- Comment #4 from Jacques Le Bourlot <Jacques.Lebourlot at obspm dot fr> ---
(In reply to kargl from comment #3)
> (In reply to Jacques Le Bourlot from comment #2)
> > 
> > The compiler should see that "IF (i >= 1)" THEN "i-1 >= 0". Quoting the
> > manual is not an adequate answer. An upgrade would be most welcomed.
> > 
> 
> gfortran is mostly maintained by a diminishing number of contributors.
> Messages telling the few remaining contributors what they ought to
> do are likely to fall on deaf ears.  Patches from new contributors
> are welcomed; otherwise, sometime in the future, perhaps a decade 
> or so, someone might care enough to try to fix this annoyance.  In the 
> meantime, add -Wno-do-subscript to your options.

My apologies. I did not mean to be rude or to complain. And I am not skilled
enough to contribute. Sorry.

But I will keep -Wdo-subscript, because I need the few cases that are *real*
positives.

Jacques

[Bug fortran/97320] False positive "Array reference out of bounds in loop" in a protecting if block

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97320

anlauf at gcc dot gnu.org changed:

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

--- Comment #5 from anlauf at gcc dot gnu.org ---
See PR 94978.

*** This bug has been marked as a duplicate of bug 94978 ***

[Bug fortran/97320] False positive "Array reference out of bounds in loop" in a protecting if block

[tkoenig at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97320

Thomas Koenig <tkoenig at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tkoenig at gcc dot gnu.org
             Status|RESOLVED                    |NEW
         Resolution|DUPLICATE                   |---
           Severity|normal                      |enhancement
         Depends on|                            |90302

--- Comment #6 from Thomas Koenig <tkoenig at gcc dot gnu.org> ---
It's not an exact duplicate of PR 94978; that bug is about
a false positive without -Wdo-subscript, whereas this one is
about a false positive with -Wdo-subscript.

The reason why this is rather difficult to resolve is one
of translation phases.

In the gfortran front end, we create a syntax tree from the
Fortran source code.  On the basis of that syntax tree (where
we still know a lot about the langauge) we issue that warning.

The next step is conversion to an intermediate language, which
gets handed to the main part of gcc for further processing
(known as the "middle end").

It is the middle which is does most of the optimizations, and
which has the tools to do so.  In this particular instance, we
would need "range propagation" (where the compiler can infer the
range of variables from previous statements).  We don't do that
in the front end, because a) it would be a major piece of work, and
b) it would duplicate a lot of what the middle end already does.

The most elegant solution would be support from the middle and
back end to put in a pseudo statement, like a __bulitin_warning
"function".

Code like

    integer :: a(12)
    do i=1,10
       a(i-1) = 1

could then be annotated like

   do i=1,10
     if (0 < lbound(a)) call __builtin_warning ("index out of bounds")
     if (9 > ubound(a)) call __builtin_warning ("index out of bounds")
     a(i-1) = 1

and if the compiler could not prove that these statements get removed
by dead code elimination, it would issue the warning in the final phase of
translation.

This would pretty much eliminate false positives, and would be
far superior than what we currently do.

Unfortunately, this is a part of a compiler with which I am almost
totally unfamiliar, so I cannot help there. Some preliminary work
has been done (see PR 90302), but I don't know how far it has
progressed in the meantime.


Referenced Bugs:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90302
[Bug 90302] Implement __builtin_warning