[flang] [llvm] [flang][llvm][OpenMP][OpenACC] Add implicit casts to omp.atomic and acc.atomic (PR #131603)

[Tom Eccles via llvm-commits]

================
@@ -386,6 +442,7 @@ void genOmpAccAtomicRead(Fortran::lower::AbstractConverter &converter,
       fir::getBase(converter.genExprAddr(fromExpr, stmtCtx));
   mlir::Value toAddress = fir::getBase(converter.genExprAddr(
       *Fortran::semantics::GetExpr(assignmentStmtVariable), stmtCtx));
+  emitImplicitCast(converter, loc, fromAddress, toAddress, elementType);
----------------
tblah wrote:

I don't think that would be useful (if I understand correctly). If flang wants to generate an error and terminate compilation it can already do that. If malformed OpenMP dialect IR without that attribute is still valid, then the attribute is not serving to guard against any misuse.

Please do correct me if I misunderstood. 

On second thought, it is quite easy to convert between different fir pointer types (!fir.ref<>, !fir.heap<>, etd) so we might as well just require that input and output types of omp atomic operations are always exactly the same.

https://github.com/llvm/llvm-project/pull/131603