[all-commits] [llvm/llvm-project] 4bdec5: [flang][runtime] Enable more code for offload devi...

[Slava Zakharin via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 4bdec5830bc398cecf6e775cc54d9dd511e6e237
      https://github.com/llvm/llvm-project/commit/4bdec5830bc398cecf6e775cc54d9dd511e6e237
  Author: Slava Zakharin <szakharin at nvidia.com>
  Date:   2023-09-27 (Wed, 27 Sep 2023)

  Changed paths:
    M flang/include/flang/Runtime/api-attrs.h
    M flang/include/flang/Runtime/descriptor.h
    M flang/include/flang/Runtime/memory.h
    M flang/include/flang/Runtime/type-code.h
    M flang/runtime/CMakeLists.txt
    M flang/runtime/ISO_Fortran_util.h
    M flang/runtime/derived.h
    M flang/runtime/descriptor.cpp
    M flang/runtime/terminator.cpp
    M flang/runtime/terminator.h
    M flang/runtime/type-code.cpp
    M flang/runtime/type-info.h

  Log Message:
  -----------
  [flang][runtime] Enable more code for offload device builds. (#67489)

I extended the "closure" of the device code containing the initial
transformational.cpp. The device side of the library should not be
complete at least for some APIs. For example, I tested with C OpenMP
code calling BesselJnX0 with a nullptr descriptor that failed with
a runtime error when executing on a GPU.

I added `--expt-relaxed-constexpr` for NVCC compiler to avoid multiple
warnings about missing `__attribute__((device))` on constexpr methods
coming from C++ header files.