[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.

flang/runtime/type-code.cpp

@@ -10,7 +10,9 @@
 
 namespace Fortran::runtime {
 
-TypeCode::TypeCode(TypeCategory f, int kind) {
+RT_OFFLOAD_API_GROUP_BEGIN
+
+RT_API_ATTRS TypeCode::TypeCode(TypeCategory f, int kind) {
   switch (f) {
   case TypeCategory::Integer:
     switch (kind) {
@@ -110,7 +112,7 @@ TypeCode::TypeCode(TypeCategory f, int kind) {
   }
 }
 
-std::optional<std::pair<TypeCategory, int>>
+RT_API_ATTRS std::optional<std::pair<TypeCategory, int>>
 TypeCode::GetCategoryAndKind() const {
   switch (raw_) {
   case CFI_type_signed_char:
@@ -205,4 +207,7 @@ TypeCode::GetCategoryAndKind() const {
     return std::nullopt;
   }
 }
+
+RT_OFFLOAD_API_GROUP_END
+
 } // namespace Fortran::runtime