[flang] [cuda] Fix CUDA implicit data transfer entity creation (llvm#139414)

Fixed an issue in `genCUDAImplicitDataTransfer` where creating an
`hlfir::Entity` from a symbol address could fail when the address comes
from a `hlfir.declare` operation. Fix is to check if the address comes
from a `hlfir.declare` operation. If so, use the base value from the
declare op when available. Falling back to the original address
otherwise.

Author: wangzpgi

flang/lib/Lower/Bridge.cpp

@@ -4778,7 +4778,14 @@ class FirConverter : public Fortran::lower::AbstractConverter {
               nbDeviceResidentObject <= 1 &&
               "Only one reference to the device resident object is supported");
           auto addr = getSymbolAddress(sym);
-          hlfir::Entity entity{addr};
+          mlir::Value baseValue;
+          if (auto declareOp =
+                  llvm::dyn_cast<hlfir::DeclareOp>(addr.getDefiningOp()))
+            baseValue = declareOp.getBase();
+          else
+            baseValue = addr;
+
+          hlfir::Entity entity{baseValue};
           auto [temp, cleanup] =
               hlfir::createTempFromMold(loc, builder, entity);
           auto needCleanup = fir::getIntIfConstant(cleanup);

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

@@ -0,0 +1,27 @@
+! RUN: bbc -emit-hlfir -fcuda %s -o - | FileCheck %s
+
+subroutine testr2(N1,N2)
+  real(4), managed :: ai4(N1,N2)
+  real(4), allocatable :: bRefi4(:)
+
+  integer :: i1, i2
+
+  do i2 = 1, N2
+    do i1 = 1, N1
+      ai4(i1,i2) = i1 + N1*(i2-1)
+    enddo
+  enddo
+
+  allocate(bRefi4 (N1))
+  do i1 = 1, N1
+    bRefi4(i1) = (ai4(i1,1)+ai4(i1,N2))*N2/2
+  enddo
+  deallocate(bRefi4)
+
+end subroutine
+
+!CHECK-LABEL: func.func @_QPtestr2
+!CHECK: %[[ALLOC:.*]] = cuf.alloc !fir.array<?x?xf32>, %{{.*}}, %{{.*}} : index, index {bindc_name = "ai4", data_attr = #cuf.cuda<managed>, uniq_name = "_QFtestr2Eai4"} -> !fir.ref<!fir.array<?x?xf32>>
+!CHECK: %[[DECLARE:.*]]:2 = hlfir.declare %[[ALLOC]](%{{.*}}) {data_attr = #cuf.cuda<managed>, uniq_name = "_QFtestr2Eai4"} : (!fir.ref<!fir.array<?x?xf32>>, !fir.shape<2>) -> (!fir.box<!fir.array<?x?xf32>>, !fir.ref<!fir.array<?x?xf32>>)
+!CHECK: %[[DEST:.*]] = hlfir.designate %[[DECLARE]]#0 (%{{.*}}, %{{.*}}) : (!fir.box<!fir.array<?x?xf32>>, i64, i64) -> !fir.ref<f32>
+!CHECK: cuf.data_transfer %{{.*}}#0 to %[[DEST]] {transfer_kind = #cuf.cuda_transfer<host_device>} : !fir.ref<f32>, !fir.ref<f32>