[flang][cuda] Fix detection of assumed size arrays in shared memory offset

[clementval]

The assumed size arrays representation has changed from -1 to a specific op fir.assumed_size_extent : index (#164452). We used to detect assumed size arrays in the shared memory offset computation with the -1 but need to use the specific op now.
Assumed sized notation is used with CUDA Fortran dynamic shared memory and should be a view on the shared memory at offset 0.

[llvmbot]

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

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

Changes

The assumed size arrays representation has changed from -1 to a specific op fir.assumed_size_extent : index.

---

Full diff: https://github.com/llvm/llvm-project/pull/167231.diff

2 Files Affected:

• (modified) flang/lib/Optimizer/Transforms/CUFComputeSharedMemoryOffsetsAndSize.cpp (+1-2)
• (modified) flang/test/Fir/CUDA/cuda-shared-offset.mlir (+14-14)

diff --git a/flang/lib/Optimizer/Transforms/CUFComputeSharedMemoryOffsetsAndSize.cpp b/flang/lib/Optimizer/Transforms/CUFComputeSharedMemoryOffsetsAndSize.cpp
index 09126e047d382..a64494510d847 100644
--- a/flang/lib/Optimizer/Transforms/CUFComputeSharedMemoryOffsetsAndSize.cpp
+++ b/flang/lib/Optimizer/Transforms/CUFComputeSharedMemoryOffsetsAndSize.cpp
@@ -41,8 +41,7 @@ namespace {
 static bool isAssumedSize(mlir::ValueRange shape) {
   if (shape.size() != 1)
     return false;
-  std::optional<std::int64_t> val = fir::getIntIfConstant(shape[0]);
-  if (val && *val == -1)
+  if (llvm::isa_and_nonnull<fir::AssumedSizeExtentOp>(shape[0].getDefiningOp()))
     return true;
   return false;
 }
diff --git a/flang/test/Fir/CUDA/cuda-shared-offset.mlir b/flang/test/Fir/CUDA/cuda-shared-offset.mlir
index 9c057d024426a..37b36b2bd050e 100644
--- a/flang/test/Fir/CUDA/cuda-shared-offset.mlir
+++ b/flang/test/Fir/CUDA/cuda-shared-offset.mlir
@@ -3,9 +3,9 @@
 module attributes {dlti.dl_spec = #dlti.dl_spec<#dlti.dl_entry<!llvm.ptr, 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<f80, dense<128> : 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<i128, dense<128> : vector<2xi64>>, #dlti.dl_entry<i8, dense<8> : vector<2xi64>>, #dlti.dl_entry<!llvm.ptr<272>, dense<64> : vector<4xi64>>, #dlti.dl_entry<i64, dense<64> : vector<2xi64>>, #dlti.dl_entry<i1, dense<8> : vector<2xi64>>, #dlti.dl_entry<"dlti.endianness", "little">, #dlti.dl_entry<"dlti.stack_alignment", 128 : i64>>, fir.defaultkind = "a1c4d8i4l4r4", fir.kindmap = "", gpu.container_module, llvm.data_layout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128", llvm.ident = "flang version 20.0.0 (https://github.com/llvm/llvm-project.git cae351f3453a0a26ec8eb2ddaf773c24a29d929e)", llvm.target_triple = "x86_64-unknown-linux-gnu"} {
   gpu.module @cuda_device_mod {
     gpu.func @_QPdynshared() kernel {
-      %c-1 = arith.constant -1 : index
-      %6 = cuf.shared_memory !fir.array<?xf32>, %c-1 : index {bindc_name = "r", uniq_name = "_QFdynsharedEr"} -> !fir.ref<!fir.array<?xf32>>
-      %7 = fir.shape %c-1 : (index) -> !fir.shape<1>
+      %0 = fir.assumed_size_extent : index
+      %6 = cuf.shared_memory !fir.array<?xf32>, %0 : index {bindc_name = "r", uniq_name = "_QFdynsharedEr"} -> !fir.ref<!fir.array<?xf32>>
+      %7 = fir.shape %0 : (index) -> !fir.shape<1>
       %8 = fir.declare %6(%7) {data_attr = #cuf.cuda<shared>, uniq_name = "_QFdynsharedEr"} : (!fir.ref<!fir.array<?xf32>>, !fir.shape<1>) -> !fir.ref<!fir.array<?xf32>>
       gpu.return
     }
@@ -14,7 +14,7 @@ module attributes {dlti.dl_spec = #dlti.dl_spec<#dlti.dl_entry<!llvm.ptr, dense<
 
 // CHECK-LABEL: gpu.module @cuda_device_mod
 // CHECK: gpu.func @_QPdynshared()
-// CHECK: %{{.*}} = cuf.shared_memory[%c0{{.*}} : i32] !fir.array<?xf32>, %c-1 : index {bindc_name = "r", uniq_name = "_QFdynsharedEr"} -> !fir.ref<!fir.array<?xf32>>       
+// CHECK: %{{.*}} = cuf.shared_memory[%c0{{.*}} : i32] !fir.array<?xf32>, %{{.*}} : index {bindc_name = "r", uniq_name = "_QFdynsharedEr"} -> !fir.ref<!fir.array<?xf32>>       
 // CHECK: gpu.return
 // CHECK: }
 // CHECK: fir.global external @_QPdynshared__shared_mem {alignment = 4 : i64, data_attr = #cuf.cuda<shared>} : !fir.array<0xi8>
@@ -127,16 +127,16 @@ module attributes {dlti.dl_spec = #dlti.dl_spec<#dlti.dl_entry<!llvm.ptr, dense<
   gpu.module @cuda_device_mod {
     gpu.func @_QMmtestsPtestany(%arg0: !fir.ref<!fir.array<?xf32>> {cuf.data_attr = #cuf.cuda<device>, fir.bindc_name = "a"}) attributes {cuf.proc_attr = #cuf.cuda_proc<global>} {
       %0 = fir.dummy_scope : !fir.dscope
-      %c-1 = arith.constant -1 : index
-      %1 = fir.shape %c-1 : (index) -> !fir.shape<1>
+      %a0 = fir.assumed_size_extent : index
+      %1 = fir.shape %a0 : (index) -> !fir.shape<1>
       %2:2 = hlfir.declare %arg0(%1) dummy_scope %0 {data_attr = #cuf.cuda<device>, uniq_name = "_QMmtestsFtestanyEa"} : (!fir.ref<!fir.array<?xf32>>, !fir.shape<1>, !fir.dscope) -> (!fir.box<!fir.array<?xf32>>, !fir.ref<!fir.array<?xf32>>)
       %3 = fir.address_of(@_QM__fortran_builtinsE__builtin_blockdim) : !fir.ref<!fir.type<_QM__fortran_builtinsT__builtin_dim3{x:i32,y:i32,z:i32}>>
       %4:2 = hlfir.declare %3 {uniq_name = "_QM__fortran_builtinsE__builtin_blockdim"} : (!fir.ref<!fir.type<_QM__fortran_builtinsT__builtin_dim3{x:i32,y:i32,z:i32}>>) -> (!fir.ref<!fir.type<_QM__fortran_builtinsT__builtin_dim3{x:i32,y:i32,z:i32}>>, !fir.ref<!fir.type<_QM__fortran_builtinsT__builtin_dim3{x:i32,y:i32,z:i32}>>)
       %5 = fir.address_of(@_QM__fortran_builtinsE__builtin_blockidx) : !fir.ref<!fir.type<_QM__fortran_builtinsT__builtin_dim3{x:i32,y:i32,z:i32}>>
       %6:2 = hlfir.declare %5 {uniq_name = "_QM__fortran_builtinsE__builtin_blockidx"} : (!fir.ref<!fir.type<_QM__fortran_builtinsT__builtin_dim3{x:i32,y:i32,z:i32}>>) -> (!fir.ref<!fir.type<_QM__fortran_builtinsT__builtin_dim3{x:i32,y:i32,z:i32}>>, !fir.ref<!fir.type<_QM__fortran_builtinsT__builtin_dim3{x:i32,y:i32,z:i32}>>)
-      %c-1_0 = arith.constant -1 : index
-      %7 = cuf.shared_memory !fir.array<?xf64>, %c-1_0 : index {bindc_name = "dmasks", uniq_name = "_QMmtestsFtestanyEdmasks"} -> !fir.ref<!fir.array<?xf64>>
-      %8 = fir.shape %c-1_0 : (index) -> !fir.shape<1>
+      %a2 = fir.assumed_size_extent : index
+      %7 = cuf.shared_memory !fir.array<?xf64>, %a2 : index {bindc_name = "dmasks", uniq_name = "_QMmtestsFtestanyEdmasks"} -> !fir.ref<!fir.array<?xf64>>
+      %8 = fir.shape %a2 : (index) -> !fir.shape<1>
       %9:2 = hlfir.declare %7(%8) {data_attr = #cuf.cuda<shared>, uniq_name = "_QMmtestsFtestanyEdmasks"} : (!fir.ref<!fir.array<?xf64>>, !fir.shape<1>) -> (!fir.box<!fir.array<?xf64>>, !fir.ref<!fir.array<?xf64>>)
       %10 = fir.address_of(@_QM__fortran_builtinsE__builtin_griddim) : !fir.ref<!fir.type<_QM__fortran_builtinsT__builtin_dim3{x:i32,y:i32,z:i32}>>
       %11:2 = hlfir.declare %10 {uniq_name = "_QM__fortran_builtinsE__builtin_griddim"} : (!fir.ref<!fir.type<_QM__fortran_builtinsT__builtin_dim3{x:i32,y:i32,z:i32}>>) -> (!fir.ref<!fir.type<_QM__fortran_builtinsT__builtin_dim3{x:i32,y:i32,z:i32}>>, !fir.ref<!fir.type<_QM__fortran_builtinsT__builtin_dim3{x:i32,y:i32,z:i32}>>)
@@ -146,9 +146,9 @@ module attributes {dlti.dl_spec = #dlti.dl_spec<#dlti.dl_entry<!llvm.ptr, dense<
       %15:2 = hlfir.declare %14 {uniq_name = "_QMmtestsFtestanyEiam"} : (!fir.ref<i32>) -> (!fir.ref<i32>, !fir.ref<i32>)
       %16 = fir.alloca i32 {bindc_name = "j", uniq_name = "_QMmtestsFtestanyEj"}
       %17:2 = hlfir.declare %16 {uniq_name = "_QMmtestsFtestanyEj"} : (!fir.ref<i32>) -> (!fir.ref<i32>, !fir.ref<i32>)
-      %c-1_1 = arith.constant -1 : index
-      %18 = cuf.shared_memory !fir.array<?xf32>, %c-1_1 : index {bindc_name = "smasks", uniq_name = "_QMmtestsFtestanyEsmasks"} -> !fir.ref<!fir.array<?xf32>>
-      %19 = fir.shape %c-1_1 : (index) -> !fir.shape<1>
+      %a3 = fir.assumed_size_extent : index
+      %18 = cuf.shared_memory !fir.array<?xf32>, %a3 : index {bindc_name = "smasks", uniq_name = "_QMmtestsFtestanyEsmasks"} -> !fir.ref<!fir.array<?xf32>>
+      %19 = fir.shape %a3 : (index) -> !fir.shape<1>
       %20:2 = hlfir.declare %18(%19) {data_attr = #cuf.cuda<shared>, uniq_name = "_QMmtestsFtestanyEsmasks"} : (!fir.ref<!fir.array<?xf32>>, !fir.shape<1>) -> (!fir.box<!fir.array<?xf32>>, !fir.ref<!fir.array<?xf32>>)
       gpu.return
     }
@@ -156,7 +156,7 @@ module attributes {dlti.dl_spec = #dlti.dl_spec<#dlti.dl_entry<!llvm.ptr, dense<
 }
 
 // CHECK-LABEL: gpu.func @_QMmtestsPtestany
-// CHECK: %{{.*}} = cuf.shared_memory[%c0{{.*}} : i32] !fir.array<?xf64>, %c-1{{.*}} : index {bindc_name = "dmasks", uniq_name = "_QMmtestsFtestanyEdmasks"} -> !fir.ref<!fir.array<?xf64>>
-// CHECK: %{{.*}} = cuf.shared_memory[%c0{{.*}} : i32] !fir.array<?xf32>, %c-1{{.*}} : index {bindc_name = "smasks", uniq_name = "_QMmtestsFtestanyEsmasks"} -> !fir.ref<!fir.array<?xf32>>
+// CHECK: %{{.*}} = cuf.shared_memory[%c0{{.*}} : i32] !fir.array<?xf64>, %{{.*}} : index {bindc_name = "dmasks", uniq_name = "_QMmtestsFtestanyEdmasks"} -> !fir.ref<!fir.array<?xf64>>
+// CHECK: %{{.*}} = cuf.shared_memory[%c0{{.*}} : i32] !fir.array<?xf32>, %{{.*}} : index {bindc_name = "smasks", uniq_name = "_QMmtestsFtestanyEsmasks"} -> !fir.ref<!fir.array<?xf32>>
 
 // CHECK: fir.global external @_QMmtestsPtestany__shared_mem {alignment = 8 : i64, data_attr = #cuf.cuda<shared>} : !fir.array<0xi8>

[jeanPerier]

Sorry for missing this case, LGTM.