[flang][cuda] Update c_loc with device variable to get host address (llvm#164317)

Bypass the declare op because it is rewritten in CUFOpConversion and
will only provide the device address. c_loc is expected to have the host
address of a device address to be used in API like `cudaMemcpyToSymbol`
so we need to provide the address of op directly.

Author: clementval

flang/lib/Optimizer/Builder/IntrinsicCall.cpp

@@ -3516,11 +3516,23 @@ static mlir::Value getAddrFromBox(fir::FirOpBuilder &builder,
   return addr;
 }
 
+static void clocDeviceArgRewrite(fir::ExtendedValue arg) {
+  // Special case for device address in c_loc.
+  if (auto emboxOp = mlir::dyn_cast_or_null<fir::EmboxOp>(
+          fir::getBase(arg).getDefiningOp()))
+    if (auto declareOp = mlir::dyn_cast_or_null<hlfir::DeclareOp>(
+            emboxOp.getMemref().getDefiningOp()))
+      if (declareOp.getDataAttr() &&
+          declareOp.getDataAttr() == cuf::DataAttribute::Device)
+        emboxOp.getMemrefMutable().assign(declareOp.getMemref());
+}
+
 static fir::ExtendedValue
 genCLocOrCFunLoc(fir::FirOpBuilder &builder, mlir::Location loc,
                  mlir::Type resultType, llvm::ArrayRef<fir::ExtendedValue> args,
                  bool isFunc = false, bool isDevLoc = false) {
   assert(args.size() == 1);
+  clocDeviceArgRewrite(args[0]);
   mlir::Value res = fir::AllocaOp::create(builder, loc, resultType);
   mlir::Value resAddr;
   if (isDevLoc)

flang/test/Lower/CUDA/cuda-cloc.cuf

@@ -0,0 +1,19 @@
+! RUN: bbc -emit-hlfir -fcuda %s -o - | FileCheck %s
+
+module symbols
+  integer(4), device, target :: sdev(100)
+end module
+
+subroutine sub1
+  use iso_c_binding
+  use symbols
+  print*, c_loc(sdev)
+end subroutine
+
+! CHECK-LABEL: func.func @_QPsub1()
+! CHECK: %[[ADDR:.*]] = fir.address_of(@_QMsymbolsEsdev) : !fir.ref<!fir.array<100xi32>>
+! CHECK: %[[EMBOX:.*]] = fir.embox %[[ADDR]](%{{.*}}) : (!fir.ref<!fir.array<100xi32>>, !fir.shape<1>) -> !fir.box<!fir.array<100xi32>>
+! CHECK: %[[__ADDRESS:.*]] = fir.coordinate_of %{{.*}}, __address : (!fir.ref<!fir.type<_QM__fortran_builtinsT__builtin_c_ptr{__address:i64}>>) -> !fir.ref<i64>
+! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[EMBOX]] : (!fir.box<!fir.array<100xi32>>) -> !fir.ref<!fir.array<100xi32>>
+! CHECK: %[[CONV:.*]] = fir.convert %[[BOX_ADDR]] : (!fir.ref<!fir.array<100xi32>>) -> i64
+! CHECK: fir.store %[[CONV]] to %[[__ADDRESS]] : !fir.ref<i64>