[flang][cuda] Do not convert fir.declare op with no user

[clementval]

No description provided.

[llvmbot]

@llvm/pr-subscribers-flang-fir-hlfir

Author: Valentin Clement (バレンタイン クレメン) (clementval)

Changes

---

Full diff: https://github.com/llvm/llvm-project/pull/164492.diff

2 Files Affected:

• (modified) flang/lib/Optimizer/Transforms/CUFOpConversion.cpp (+4)
• (modified) flang/test/Fir/CUDA/cuda-global-addr.mlir (+20)

diff --git a/flang/lib/Optimizer/Transforms/CUFOpConversion.cpp b/flang/lib/Optimizer/Transforms/CUFOpConversion.cpp
index 759e3a65dd24f..8d00272b09f42 100644
--- a/flang/lib/Optimizer/Transforms/CUFOpConversion.cpp
+++ b/flang/lib/Optimizer/Transforms/CUFOpConversion.cpp
@@ -454,6 +454,8 @@ struct DeclareOpConversion : public mlir::OpRewritePattern<fir::DeclareOp> {
   mlir::LogicalResult
   matchAndRewrite(fir::DeclareOp op,
                   mlir::PatternRewriter &rewriter) const override {
+    if (op.getResult().getUsers().empty())
+      return success();
     if (auto addrOfOp = op.getMemref().getDefiningOp<fir::AddrOfOp>()) {
       if (auto global = symTab.lookup<fir::GlobalOp>(
               addrOfOp.getSymbol().getRootReference().getValue())) {
@@ -963,6 +965,8 @@ class CUFOpConversion : public fir::impl::CUFOpConversionBase<CUFOpConversion> {
     }
 
     target.addDynamicallyLegalOp<fir::DeclareOp>([&](fir::DeclareOp op) {
+      if (op.getResult().getUsers().empty())
+        return true;
       if (inDeviceContext(op))
         return true;
       if (auto addrOfOp = op.getMemref().getDefiningOp<fir::AddrOfOp>()) {
diff --git a/flang/test/Fir/CUDA/cuda-global-addr.mlir b/flang/test/Fir/CUDA/cuda-global-addr.mlir
index 3e50c7a51f49c..6f7816c9163cb 100644
--- a/flang/test/Fir/CUDA/cuda-global-addr.mlir
+++ b/flang/test/Fir/CUDA/cuda-global-addr.mlir
@@ -63,6 +63,7 @@ module attributes {dlti.dl_spec = #dlti.dl_spec<#dlti.dl_entry<f80, dense<128> :
 
 // We cannot call _FortranACUFGetDeviceAddress on a constant global. 
 // There is no symbol for it and the call would result into an unresolved reference.
+// CHECK-LABEL: func.func @_QQmain() attributes {fir.bindc_name = "arraysize"}
 // CHECK-NOT: fir.call {{.*}}GetDeviceAddress
 
 // -----
@@ -90,3 +91,22 @@ func.func @_QQmain() attributes {fir.bindc_name = "test"} {
 // CHECK-NOT: fir.call {{.*}}GetDeviceAddress
 
 }
+
+// -----
+
+// Check that we do not introduce call to _FortranACUFGetDeviceAddress when the
+// value has no user.
+
+module attributes {dlti.dl_spec = #dlti.dl_spec<#dlti.dl_entry<f80, dense<128> : vector<2xi64>>, #dlti.dl_entry<i128, dense<128> : vector<2xi64>>, #dlti.dl_entry<i64, dense<64> : vector<2xi64>>, #dlti.dl_entry<!llvm.ptr<272>, dense<64> : vector<4xi64>>, #dlti.dl_entry<!llvm.ptr<271>, dense<32> : vector<4xi64>>, #dlti.dl_entry<!llvm.ptr<270>, dense<32> : vector<4xi64>>, #dlti.dl_entry<f128, dense<128> : vector<2xi64>>, #dlti.dl_entry<f64, dense<64> : vector<2xi64>>, #dlti.dl_entry<f16, dense<16> : vector<2xi64>>, #dlti.dl_entry<i32, dense<32> : vector<2xi64>>, #dlti.dl_entry<i16, dense<16> : vector<2xi64>>, #dlti.dl_entry<i8, dense<8> : vector<2xi64>>, #dlti.dl_entry<i1, dense<8> : vector<2xi64>>, #dlti.dl_entry<!llvm.ptr, dense<64> : vector<4xi64>>, #dlti.dl_entry<"dlti.endianness", "little">, #dlti.dl_entry<"dlti.stack_alignment", 128 : i64>>} {
+  func.func @_QQmain() attributes {fir.bindc_name = "T"} {
+    %0 = fir.dummy_scope : !fir.dscope
+    %1 = fir.address_of(@_QMcon2Ezzz) : !fir.ref<i32>
+    %2 = fir.declare %1 {data_attr = #cuf.cuda<constant>, uniq_name = "_QMcon2Ezzz"} : (!fir.ref<i32>) -> !fir.ref<i32>
+    return
+  }
+  fir.global @_QMcon2Ezzz {data_attr = #cuf.cuda<constant>} : i32
+}
+
+// CHECK-LABEL:  func.func @_QQmain()
+// CHECK: fir.address_of(@_QMcon2Ezzz) : !fir.ref<i32>
+// CHECK-NOT: fir.call {{.*}}GetDeviceAddress