[Bug fortran/107576] New: [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[Bug fortran/107576] New: [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

            Bug ID: 107576
           Summary: [10/11/12/13 Regression] ICE in
                    gfc_conv_procedure_call, at fortran/trans-expr.cc:6193
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gscfq@t-online.de
  Target Milestone: ---

Started with r7 :


$ cat z1.f90
program p
   integer, pointer :: z(:) => null()
   call s(z)
   call s(null(z))
end


$ gfortran-6 -c z1.f90 -std=legacy
$
$ gfortran-13-20221106 -c z1.f90 -std=legacy
z1.f90:4:18:

    4 |    call s(null(z))
      |                  1
Warning: Unexpected NULL() intrinsic at (1) to dummy '_formal_0'
z1.f90:4:18: internal compiler error: in gfc_conv_procedure_call, at
fortran/trans-expr.cc:6193
0x8afc1a gfc_conv_procedure_call(gfc_se*, gfc_symbol*, gfc_actual_arglist*,
gfc_expr*, vec<tree_node*, va_gc, vl_embed>*)
        ../../gcc/fortran/trans-expr.cc:6191
0x8ed790 gfc_trans_call(gfc_code*, bool, tree_node*, tree_node*, bool)
        ../../gcc/fortran/trans-stmt.cc:422
0x8740f8 trans_code
        ../../gcc/fortran/trans.cc:2018
0x89d4ce gfc_generate_function_code(gfc_namespace*)
        ../../gcc/fortran/trans-decl.cc:7659
0x81f57e translate_all_program_units
        ../../gcc/fortran/parse.cc:6696
0x81f57e gfc_parse_file()
        ../../gcc/fortran/parse.cc:7002
0x86ccbf gfc_be_parse_file
        ../../gcc/fortran/f95-lang.cc:229

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

G. Steinmetz <gscfq@t-online.de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |accepts-invalid,
                   |                            |ice-on-invalid-code

--- Comment #1 from G. Steinmetz <gscfq@t-online.de> ---

Variant z2.f90 with reversed order always compiles,
for every tested release, with and without -std=legacy :
(but no explicit interface, so error expected)


$ cat z2.f90
program p
   integer, pointer :: z(:) => null()
   call s(null(z))
   call s(z)
end

$ gfortran-13-20221106 -c z2.f90
$ gfortran-13-20221106 -c z2.f90 -std=legacy


Test cases with an explicit interface and specified pointer dummy (ok) :

$ cat z1c.f90
program p
   integer, pointer :: z(:) => null()
   call s(z)
   call s(null(z))
contains
   subroutine s(x)
      integer, pointer :: x(:)
   end
end

$ cat z2c.f90
program p
   integer, pointer :: z(:) => null()
   call s(null(z))
   call s(z)
contains
   subroutine s(x)
      integer, pointer :: x(:)
   end
end

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P4
   Target Milestone|---                         |10.5

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

anlauf at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2022-11-08

--- Comment #2 from anlauf at gcc dot gnu.org ---
Might be related to interface checking.

The ICE is reproduced by using -fallow-argument-mismatch instead of -std=legacy

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

--- Comment #3 from anlauf at gcc dot gnu.org ---
I am still wondering if this is really invalid code.

null(z) has the same characteristics as z, see F2018:16.9.144
So is there really a mismatch of arguments?

Second, just one observation: since the permutation of call s() works,
I played and deactivated the if-branch at trans-expr.cc:6191, which avoids
the assert, and I got a tree-dump that corresponds to my expectation.

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

--- Comment #4 from anlauf at gcc dot gnu.org ---
Partial fix:

diff --git a/gcc/fortran/trans-expr.cc b/gcc/fortran/trans-expr.cc
index b95c5cf2f96..f09fad337bc 100644
--- a/gcc/fortran/trans-expr.cc
+++ b/gcc/fortran/trans-expr.cc
@@ -6193,6 +6193,7 @@ gfc_conv_procedure_call (gfc_se * se, gfc_symbol * sym,
        }
       else if (arg->expr->expr_type == EXPR_NULL
               && fsym && !fsym->attr.pointer
+              && fsym->attr.optional
               && (fsym->ts.type != BT_CLASS
                   || !CLASS_DATA (fsym)->attr.class_pointer))
        {

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

--- Comment #5 from G. Steinmetz <gscfq@t-online.de> ---

I had in mind Fortran 2018, 15.4.2.2 Explicit interface,
and therefore had added the examples z1c.f90 and z2c.f90.

Some other compilers (via Compiler Explorer) show an error 
for both z1.f90 and z2.f90 :

latest flang-new :
error: Semantic errors in /app/example.f90
error: Null pointer argument requires an explicit interface
     call s(null(z))

latest ifort/ifx :
error #8615: Intrinsic NULL() is passed as actual argument. An explicit
interface specifying that the corresponding dummy is pointer, allocatable or
optional is required.
   call s(null(z))

Would be interesting what e.g. NAG and Cray say about it.

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

anlauf at gcc dot gnu.org changed:

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

--- Comment #6 from anlauf at gcc dot gnu.org ---
(In reply to G. Steinmetz from comment #5)
> I had in mind Fortran 2018, 15.4.2.2 Explicit interface,
> and therefore had added the examples z1c.f90 and z2c.f90.

Can you point to the relevant clause?  I just don't see it...
(Although I agree one should need an explicit interface here.)

15.4.2.2(3) talks about dummy arguments, not about procedure references.
Am I missing something?

And call s(z) should be valid, although z has the POINTER attribute,
as long as z is associated.

> latest ifort/ifx :
> error #8615: Intrinsic NULL() is passed as actual argument. An explicit
> interface specifying that the corresponding dummy is pointer, allocatable or
> optional is required.
>    call s(null(z))

Yeah, that makes sense.

> Would be interesting what e.g. NAG and Cray say about it.

NAG Fortran Compiler Release 7.1(Hanzomon) Build 7101
Error: pr107576-z1.f90, line 4: Actual argument 1 to S is a disassociated
pointer

On the other hand,

Cray Fortran : Version 14.0.0  Mon Nov 14, 2022  18:53:42

is happily compiling the code, but I don't know what it does...

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

--- Comment #7 from G. Steinmetz <gscfq@t-online.de> ---

Yes, call s(z) is valid in z1, but is necessary to trigger
the ICE somehow with -std=legacy and null().

Yes, 15.4.2.2(3) was it, explicit interface of the callee with
all the characteristics of that procedure(subr/func) and dummies.

Two things are mixed here.. The following is e.g. handled correctly,
with a different error, but rejected without an explicit interface:

$ cat z3.f90
program p
   call s(null())
end

$ cat z3c.f90
program p
   call s(null())
contains
   subroutine s(x)
      integer, pointer :: x
   end
end

$ cat z3i.f90
program p
   interface
      subroutine s(x)
         integer, pointer :: x
      end
   end interface
   call s(null())
end

$ gfortran -c z3c.f90
$ gfortran -c z3i.f90
$ gfortran -c z3.f90
z3.f90:2:10:

    2 |    call s(null())
      |          1
Error: MOLD argument to NULL required at (1)

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

--- Comment #8 from anlauf at gcc dot gnu.org ---
(In reply to G. Steinmetz from comment #7)
> $ gfortran -c z3.f90
> z3.f90:2:10:
> 
>     2 |    call s(null())
>       |          1
> Error: MOLD argument to NULL required at (1)

I think that error message might be fine.  It seems to be related to
16.9.144 NULL() and table 16.5: Characteristics of the result of NULL ( ).

This is why z3c/z2i give no error: there is an explicit interface,
and 16.9.144(5) explains how this is resolved.

With gfortran -fallow-argument-mismatch, we have the bad situation that the
user explicitly wants to leave the space of standard conformance, and we are
entering extension space, with all its problems...

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

--- Comment #9 from anlauf at gcc dot gnu.org ---
Something like the following rejects NULL when there is no interface:

diff --git a/gcc/fortran/interface.cc b/gcc/fortran/interface.cc
index 49dbd1d886c..62b325b726d 100644
--- a/gcc/fortran/interface.cc
+++ b/gcc/fortran/interface.cc
@@ -4160,6 +4180,14 @@ gfc_procedure_use (gfc_symbol *sym, gfc_actual_arglist
**ap, locus *where)
              return false;
            }

+         if (a->expr && a->expr->expr_type == EXPR_NULL)
+           {
+             gfc_error ("Passing intrinsic NULL as actual argument at %L "
+                        "requires an explicit interface", &a->expr->where);
+             a->expr->error = 1;
+             return false;
+           }
+
          /* TS 29113, C407b.  */
          if (a->expr && a->expr->expr_type == EXPR_VARIABLE
              && symbol_rank (a->expr->symtree->n.sym) == -1)

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

anlauf at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |anlauf at gcc dot gnu.org
             Status|NEW                         |ASSIGNED

--- Comment #10 from anlauf at gcc dot gnu.org ---
(In reply to anlauf from comment #9)
> Something like the following rejects NULL when there is no interface:

Submitted: https://gcc.gnu.org/pipermail/fortran/2022-November/058531.html

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

--- Comment #11 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Harald Anlauf <anlauf@gcc.gnu.org>:

https://gcc.gnu.org/g:820c25c83561085f54268bd536f9d216d03c3e18

commit r13-4147-g820c25c83561085f54268bd536f9d216d03c3e18
Author: Harald Anlauf <anlauf@gmx.de>
Date:   Thu Nov 17 21:36:49 2022 +0100

    Fortran: reject NULL actual argument without explicit interface [PR107576]

    gcc/fortran/ChangeLog:

            PR fortran/107576
            * interface.cc (gfc_procedure_use): Reject NULL as actual argument
            when there is no explicit procedure interface.

    gcc/testsuite/ChangeLog:

            PR fortran/107576
            * gfortran.dg/null_actual_3.f90: New test.

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

--- Comment #12 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Harald Anlauf
<anlauf@gcc.gnu.org>:

https://gcc.gnu.org/g:2a7a103dd46cc079252302be69cfef188caee74d

commit r12-8942-g2a7a103dd46cc079252302be69cfef188caee74d
Author: Harald Anlauf <anlauf@gmx.de>
Date:   Thu Nov 17 21:36:49 2022 +0100

    Fortran: reject NULL actual argument without explicit interface [PR107576]

    gcc/fortran/ChangeLog:

            PR fortran/107576
            * interface.cc (gfc_procedure_use): Reject NULL as actual argument
            when there is no explicit procedure interface.

    gcc/testsuite/ChangeLog:

            PR fortran/107576
            * gfortran.dg/null_actual_3.f90: New test.

    (cherry picked from commit 820c25c83561085f54268bd536f9d216d03c3e18)

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

--- Comment #13 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-11 branch has been updated by Harald Anlauf
<anlauf@gcc.gnu.org>:

https://gcc.gnu.org/g:55e5ef55a0eee1cb329754f30889d93567daae18

commit r11-10394-g55e5ef55a0eee1cb329754f30889d93567daae18
Author: Harald Anlauf <anlauf@gmx.de>
Date:   Thu Nov 17 21:36:49 2022 +0100

    Fortran: reject NULL actual argument without explicit interface [PR107576]

    gcc/fortran/ChangeLog:

            PR fortran/107576
            * interface.c (gfc_procedure_use): Reject NULL as actual argument
            when there is no explicit procedure interface.

    gcc/testsuite/ChangeLog:

            PR fortran/107576
            * gfortran.dg/null_actual_3.f90: New test.

    (cherry picked from commit 820c25c83561085f54268bd536f9d216d03c3e18)

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

--- Comment #14 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-10 branch has been updated by Harald Anlauf
<anlauf@gcc.gnu.org>:

https://gcc.gnu.org/g:b3b48f883d31910f2672640de2f5ff59e765488c

commit r10-11095-gb3b48f883d31910f2672640de2f5ff59e765488c
Author: Harald Anlauf <anlauf@gmx.de>
Date:   Thu Nov 17 21:36:49 2022 +0100

    Fortran: reject NULL actual argument without explicit interface [PR107576]

    gcc/fortran/ChangeLog:

            PR fortran/107576
            * interface.c (gfc_procedure_use): Reject NULL as actual argument
            when there is no explicit procedure interface.

    gcc/testsuite/ChangeLog:

            PR fortran/107576
            * gfortran.dg/null_actual_3.f90: New test.

    (cherry picked from commit 820c25c83561085f54268bd536f9d216d03c3e18)

[Bug fortran/107576] [10/11/12/13 Regression] ICE in gfc_conv_procedure_call, at fortran/trans-expr.cc:6193

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107576

anlauf at gcc dot gnu.org changed:

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

--- Comment #15 from anlauf at gcc dot gnu.org ---
Fixed on all open branches.  Closing.

Thanks for the report!