[Flang][OpenMP] verification of lowering to FIR failed with "the type of the operand must be a pointer type whose element type is the same as that of the region argument"

[k-arrows]

Reproducible on Godbolt
https://godbolt.org/z/P8qY5PqMd

Reproducer (reduced from https://github.com/gcc-mirror/gcc/blob/master/libgomp/testsuite/libgomp.fortran/omp_atomic4.f90)

$ cat omp_atomic4_reduced.f90
    real :: c
    double precision :: c2
!$omp atomic capture
    c2 = c
    c = 2.0 * c
!$omp end atomic
end
$ flang-new -fopenmp omp_atomic4_reduced.f90
error: loc("/tmp/omp_atomic4_reduced.f90":1:5): the type of the operand must be a pointer type whose element type is the same as that of the region argument
error: verification of lowering to FIR failed

If I change c = 2.0 * c in the program into c = 2.0, the following is obtained. So I think this is related to (or essentially identical to) #83144.

error: address must dereference to value type
error: verification of lowering to FIR failed

[llvmbot]

@llvm/issue-subscribers-flang-ir

Author: None (k-arrows)

Reproducible on Godbolt https://godbolt.org/z/P8qY5PqMd

Reproducer (reduced from https://github.com/gcc-mirror/gcc/blob/master/libgomp/testsuite/libgomp.fortran/omp_atomic4.f90)

$ cat omp_atomic4_reduced.f90
    real :: c
    double precision :: c2
!$omp atomic capture
    c2 = c
    c = 2.0 * c
!$omp end atomic
end
$ flang-new -fopenmp omp_atomic4_reduced.f90
error: loc("/tmp/omp_atomic4_reduced.f90":1:5): the type of the operand must be a pointer type whose element type is the same as that of the region argument
error: verification of lowering to FIR failed

If I change c = 2.0 * c in the program into c = 2.0, the following is obtained. So I think this is related to (or essentially identical to) #83144.

error: address must dereference to value type
error: verification of lowering to FIR failed

[tblah]

The reproducer in this ticket builds okay for me with a recent flang, but the omp_atomic4.f90 test stops with code 1 so I think there is a second bug here.

[k-arrows]

The reproducer in this ticket builds okay for me with a recent flang, but the omp_atomic4.f90 test stops with code 1 so I think there is a second bug here.

@tblah
Thank you for your comment. Does the following reduced program work as a reproducer?

    real :: c
    double precision :: c2
!$omp atomic write
    c = 3
!$omp atomic capture
    c2 = c
    c = 2.0 * c
!$omp end atomic
    if (c2 .ne. 3) stop 1
end

$ gfortran -fopenmp omp_atomic4_reduced2.f90
$ ./a.out
$ flang -fopenmp omp_atomic4_reduced2.f90
flang-20: warning: OpenMP support in flang is still experimental [-Wexperimental-option]
$ ./a.out
Fortran STOP: code 1

[tblah]

@tblah Thank you for your comment. Does the following reduced program work as a reproducer?

It does. Thanks!

[k-arrows]

For me with a recent flang, the omp_atomic4.f90 test works. I will close this issue.