Re: [Patch] Fortran: %re/%im fixes for OpenMP/OpenACC + gfc_is_simplify_contiguous

[Thomas Schwinge <thomas@codesourcery.com>]

Hi Tobias!

On 2021-02-09T12:41:48+0100, Tobias Burnus <tobias@codesourcery.com> wrote:
> Updated patch. Changes: Testcases split + updated/extended.
> OK for mainline?

I don't have any further comments on the patch itself.

> On 09.02.21 10:45, Thomas Schwinge wrote:
>> Thanks for filing/locating these discussion items for OpenACC/OpenMP
>> upstream.  May also put these references into the testcases, so that once
>> these get addressed, we have something to 'grep' for in GCC?
>
> Actually, they are already in the file.

Heh, thanks, I see -- evidently I haven't been reading carefully...

> Alternative is to add them as
> link, but I am not sure that's better. (I moved them now to the top.)

ACK.  As no doubt you've noticed, I like to put in direct/explicit links
-- but that then occasionally has lead to people external to
OpenACC/OpenMP GitHub reporting that they opened these and got a 404
"Page not found", confusing.

>> I note that 'zz' variants (see below) are not being checked for OpenMP
>
> I have now added them; I had them before but as many checks triggered,
> I thought the tests were not really worthwhile.

(Your call, I suppose -- I just noted the inconistency.)

>>> +!$acc update self(zz%re)
>>> +!$acc update self(zz%im)
>>> +end

>> And for OpenACC, the 'zz' variants do not emit this error message here.
>> (That's not immediately obvious to me.)
>
> Answer: 'git add' missing.

Heh.  ;-)

> The reason is that with and without
> Julian's patch, the error message is different. Without his patch,
> the error is:
> !$acc update self(zz%re) ! { dg-error "not a proper array section" }

Meaning that he'll then to updated these with his patch.  (Julian CCed.)

>> I can see how data mapping of '[...]%re' etc. are problematic (we're
>> constructing an "incomplete object"?), but 'update' etc. I'd have
>> expected to work: would just copy the respective "part".

> Granted. The array(:)%re access might update too much, but that's not
> different to array with strides or with contiguous arrays sections
> which contain component reference (and more than one component).

(Is that indeed allowed to "update too much"?)

> But that's more a question for the spec committee – if it is supposed
> to work, the code needs to be updated.

ACK.


Grüße
 Thomas


> Fortran: %re/%im fixes for OpenMP/OpenACC + gfc_is_simplify_contiguous
>
> gcc/fortran/ChangeLog:
>
>       * expr.c (gfc_is_simplify_contiguous): Handle REF_INQUIRY, i.e.
>       %im and %re which are EXPR_VARIABLE.
>       * openmp.c (resolve_omp_clauses): Diagnose %re/%im explicitly.
>
> gcc/testsuite/ChangeLog:
>
>       * gfortran.dg/goacc/ref_inquiry.f90: New test.
>       * gfortran.dg/gomp/ref_inquiry.f90: New test.
>
>  gcc/fortran/expr.c                              |  2 +
>  gcc/fortran/openmp.c                            |  8 ++++
>  gcc/testsuite/gfortran.dg/goacc/ref_inquiry.f90 | 52 +++++++++++++++++++++++++
>  gcc/testsuite/gfortran.dg/gomp/ref_inquiry.f90  | 39 +++++++++++++++++++
>  4 files changed, 101 insertions(+)
>
> diff --git a/gcc/fortran/expr.c b/gcc/fortran/expr.c
> index 4f456fc629a..92a6700568d 100644
> --- a/gcc/fortran/expr.c
> +++ b/gcc/fortran/expr.c
> @@ -5854,6 +5854,8 @@ gfc_is_simply_contiguous (gfc_expr *expr, bool strict, bool permit_element)
>       part_ref  = ref;
>        else if (ref->type == REF_SUBSTRING)
>       return false;
> +      else if (ref->type == REF_INQUIRY)
> +     return false;
>        else if (ref->u.ar.type != AR_ELEMENT)
>       ar = &ref->u.ar;
>      }
> diff --git a/gcc/fortran/openmp.c b/gcc/fortran/openmp.c
> index 797f6c86b62..f0307c801a5 100644
> --- a/gcc/fortran/openmp.c
> +++ b/gcc/fortran/openmp.c
> @@ -5219,6 +5219,14 @@ resolve_omp_clauses (gfc_code *code, gfc_omp_clauses *omp_clauses,
>                               && array_ref->next->type == REF_SUBSTRING)))
>                     gfc_error ("Unexpected substring reference in %s clause "
>                                "at %L", name, &n->where);
> +                 else if (array_ref && array_ref->type == REF_INQUIRY)
> +                   {
> +                     gcc_assert (array_ref->u.i == INQUIRY_RE
> +                                 || array_ref->u.i == INQUIRY_IM);
> +                     gfc_error ("Unexpected complex-parts designator "
> +                                "reference in %s clause at %L",
> +                                name, &n->where);
> +                   }
>                   else if (!resolved
>                       || n->expr->expr_type != EXPR_VARIABLE
>                       || array_ref->next
> diff --git a/gcc/testsuite/gfortran.dg/goacc/ref_inquiry.f90 b/gcc/testsuite/gfortran.dg/goacc/ref_inquiry.f90
> new file mode 100644
> index 00000000000..274065dae94
> --- /dev/null
> +++ b/gcc/testsuite/gfortran.dg/goacc/ref_inquiry.f90
> @@ -0,0 +1,52 @@
> +! Check for <var>%re, ...%im, ...%kind, ...%len
> +! Cf. also OpenMP's ../gomp/ref_inquiry.f90
> +! Cf. OpenACC spec issue 346
> +!
> +implicit none
> +type t
> +  integer :: i
> +  character :: c
> +  complex :: z
> +  complex :: zz(5)
> +end type t
> +
> +integer :: i
> +character(kind=4, len=5) :: c
> +complex :: z, zz(5)
> +type(t) :: x
> +
> +print *, is_contiguous(zz(:)%re)
> +
> +! inquiry function; expr_type != EXPR_VARIABLE:
> +!$acc enter data copyin(i%kind, c%len)     ! { dg-error "not a proper array section" }
> +!$acc enter data copyin(x%i%kind)          ! { dg-error "not a proper array section" }
> +!$acc enter data copyin(x%c%len)           ! { dg-error "not a proper array section" }
> +!$acc update self(i%kind, c%len)           ! { dg-error "not a proper array section" }
> +!$acc update self(x%i%kind)                ! { dg-error "not a proper array section" }
> +!$acc update self(x%c%len)                 ! { dg-error "not a proper array section" }
> +
> +! EXPR_VARIABLE
> +!$acc enter data copyin(z%re)    ! { dg-error "Unexpected complex-parts designator" }
> +!$acc enter data copyin(z%im)    ! { dg-error "Unexpected complex-parts designator" }
> +!$acc enter data copyin(zz%re)   ! { dg-error "not a proper array section" }
> +                                 ! { dg-error "Array is not contiguous" "" { target *-*-* } .-1 }
> +!$acc enter data copyin(zz%im)   ! { dg-error "not a proper array section" }
> +                                 ! { dg-error "Array is not contiguous" "" { target *-*-* } .-1 }
> +
> +!$acc enter data copyin(x%z%re)  ! { dg-error "Unexpected complex-parts designator" }
> +!$acc enter data copyin(x%z%im)  ! { dg-error "Unexpected complex-parts designator" }
> +!$acc enter data copyin(x%zz%re) ! { dg-error "not a proper array section" }
> +                                 ! { dg-error "Array is not contiguous" "" { target *-*-* } .-1 }
> +!$acc enter data copyin(x%zz%im) ! { dg-error "not a proper array section" }
> +                                 ! { dg-error "Array is not contiguous" "" { target *-*-* } .-1 }
> +
> +!$acc update self(z%re)         ! { dg-error "Unexpected complex-parts designator" }
> +!$acc update self(z%im)         ! { dg-error "Unexpected complex-parts designator" }
> +!$acc update self(zz%re)        ! { dg-error "not a proper array section" }
> +!$acc update self(zz%im)        ! { dg-error "not a proper array section" }
> +
> +!$acc update self(x%z%re)       ! { dg-error "Unexpected complex-parts designator" }
> +!$acc update self(x%z%im)       ! { dg-error "Unexpected complex-parts designator" }
> +!$acc update self(x%zz%re)      ! { dg-error "is not a proper array section" }
> +!$acc update self(x%zz%im)      ! { dg-error "is not a proper array section" }
> +end
> diff --git a/gcc/testsuite/gfortran.dg/gomp/ref_inquiry.f90 b/gcc/testsuite/gfortran.dg/gomp/ref_inquiry.f90
> new file mode 100644
> index 00000000000..37461040560
> --- /dev/null
> +++ b/gcc/testsuite/gfortran.dg/gomp/ref_inquiry.f90
> @@ -0,0 +1,39 @@
> +! Check for <var>%re, ...%im, ...%kind, ...%len
> +! Cf. also OpenACC's ../goacc/ref_inquiry.f90
> +! Cf. also OpenMP spec issue 2661
> +implicit none
> +type t
> +  integer :: i
> +  character :: c
> +  complex :: z
> +  complex :: zz(5)
> +end type t
> +
> +integer :: i
> +character(kind=4, len=5) :: c
> +complex :: z, zz(5)
> +type(t) :: x
> +
> +print *, is_contiguous(zz(:)%re)
> +
> +! inquiry function; expr_type != EXPR_VARIABLE:
> +!$omp target enter data map(to: i%kind, c%len)     ! { dg-error "not a proper array section" }
> +!$omp target enter data map(to: x%i%kind)          ! { dg-error "not a proper array section" }
> +!$omp target enter data map(to: x%c%len)           ! { dg-error "not a proper array section" }
> +
> +! EXPR_VARIABLE
> +!$omp target enter data map(to: z%re)    ! { dg-error "Unexpected complex-parts designator" }
> +!$omp target enter data map(to: z%im)    ! { dg-error "Unexpected complex-parts designator" }
> +!$omp target enter data map(to: zz%re)   ! { dg-error "not a proper array section" }
> +                                         ! { dg-error "Array is not contiguous" "" { target *-*-* } .-1 }
> +!$omp target enter data map(to: zz%im)   ! { dg-error "not a proper array section" }
> +                                         ! { dg-error "Array is not contiguous" "" { target *-*-* } .-1 }
> +
> +!$omp target enter data map(to: x%z%re)  ! { dg-error "Unexpected complex-parts designator" }
> +!$omp target enter data map(to: x%z%im)  ! { dg-error "Unexpected complex-parts designator" }
> +!$omp target enter data map(to: x%zz%re) ! { dg-error "not a proper array section" }
> +                                         ! { dg-error "Array is not contiguous" "" { target *-*-* } .-1 }
> +!$omp target enter data map(to: x%zz%im) ! { dg-error "not a proper array section" }
> +                                         ! { dg-error "Array is not contiguous" "" { target *-*-* } .-1 }
> +
> +end
-----------------
Mentor Graphics (Deutschland) GmbH, Arnulfstrasse 201, 80634 München Registergericht München HRB 106955, Geschäftsführer: Thomas Heurung, Frank Thürauf