[Bug fortran/68009] New: [6 Regression] prototype for gfortran_runtime_error with inline matmul

[tkoenig at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68009

            Bug ID: 68009
           Summary: [6 Regression] prototype for gfortran_runtime_error
                    with inline matmul
           Product: gcc
           Version: 6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: tkoenig at gcc dot gnu.org
  Target Milestone: ---

A discussion from the gfortran mailing list that never made it into Bugzilla...

See the thread at https://gcc.gnu.org/ml/fortran/2015-07/msg00079.html .

The declaration of gfortran_runtime_error as emitted by the FE is incorrect,
it does not take into account that it is variadic. This is incorrect,
and causes actual problems on nvptx.

A simplistic solution which sets resolved_isym on the symbol introduces even
stranger behavior: The function call is optimized away completele
(see https://gcc.gnu.org/ml/fortran/2015-07/msg00127.html ).