Merge branch 'main' into hgh/libcxx/P2255R2-A_type_trait_to_detect_reference_binding_to_temporary

Author: H-G-Hristov

flang/test/Driver/config-file.f90

@@ -77,13 +77,13 @@
 ! CHECK-LINKING-LIBOMP-GOES-AFTER: "--as-needed" "{{.*}}-{{.*}}.o" "-lmylib" "foo.a" "-lm" "-Bstatic" "-lhappy" "-Bdynamic" {{.*}}"-lomp"
 ! CHECK-NOLINKING: Configuration file: {{.*}}Inputs{{.}}config-l.cfg
 ! CHECK-NOLINKING: "-ffast-math"
-! CHECK-NOLINKING-NO: "-lm" "-Bstatic" "-lhappy" "-Bdynamic"
+! CHECK-NOLINKING-NOT: "-lm" "-Bstatic" "-lhappy" "-Bdynamic"
 ! CHECK-NOLINKING-OPENMP: Configuration file: {{.*}}Inputs{{.}}config-l.cfg
 ! CHECK-NOLINKING-OPENMP: "-ffast-math" {{.*}}"-fopenmp"
-! CHECK-NOLINKING-OPENMP-NO: "-lm" "-Bstatic" "-lhappy" "-Bdynamic" {{.}}"-lomp"
+! CHECK-NOLINKING-OPENMP-NOT: "-lm" "-Bstatic" "-lhappy" "-Bdynamic" {{.}}"-lomp"
 ! CHECK-LINKING-MSVC: Configuration file: {{.*}}Inputs{{.}}config-l.cfg
 ! CHECK-LINKING-MSVC: "-ffast-math"
 ! CHECK-LINKING-MSVC: "--as-needed" "{{.*}}-{{.*}}.o" "mylib.lib" "foo.lib" "m.lib" "-Bstatic" "happy.lib" "-Bdynamic"
 ! CHECK-NOLINKING-MSVC: Configuration file: {{.*}}Inputs{{.}}config-l.cfg
 ! CHECK-NOLINKING-MSVC: "-ffast-math"
-! CHECK-NOLINKING-MSVC-NO: "m.lib" "-Bstatic" "happy.lib" "-Bdynamic"
+! CHECK-NOLINKING-MSVC-NOT: "m.lib" "-Bstatic" "happy.lib" "-Bdynamic"

flang/test/Lower/CUDA/cuda-data-transfer.cuf

@@ -393,4 +393,4 @@ end subroutine
 ! CHECK: %[[ALLOC_TMP:.*]] = fir.allocmem !fir.array<10xi32> {bindc_name = ".tmp", uniq_name = ""}
 ! CHECK: %[[TMP:.*]]:2 = hlfir.declare %[[ALLOC_TMP]](%{{.*}}) {uniq_name = ".tmp"} : (!fir.heap<!fir.array<10xi32>>, !fir.shape<1>) -> (!fir.heap<!fir.array<10xi32>>, !fir.heap<!fir.array<10xi32>>)
 ! CHECK: cuf.data_transfer %[[ADEV_DECL]]#1 to %[[TMP]]#0 {transfer_kind = #cuf.cuda_transfer<device_host>} : !fir.ref<!fir.array<10xi32>>, !fir.heap<!fir.array<10xi32>>
-! CHECL: hlfir.assign 
+! CHECK: hlfir.assign

flang/test/Lower/HLFIR/type-info-components.f90

@@ -17,7 +17,7 @@ subroutine test_1(x)
   type(sometype) :: x
 end subroutine
 ! CHECK-LABEL:   fir.type_info @_QFtest_1Tsometype
-! CHECK-SAME     component_info {
+! CHECK-SAME:    component_info {
 ! CHECK:           fir.dt_component "i" lbs [-1] init @_QFtest_1E.di.sometype.i
 ! CHECK-NOT:       fir.dt_component "j"
 ! CHECK:           fir.dt_component "p" init @_QFtest_1E.di.sometype.p
@@ -35,7 +35,7 @@ subroutine test_nesting(x)
   type(sometype2) :: x
 end subroutine
 ! CHECK-LABEL:   fir.type_info @_QFtest_nestingTsome_sub_type
-! CHECK-SAME     component_info {
+! CHECK-SAME:    component_info {
 ! CHECK:           fir.dt_component "i" init @_QFtest_nestingE.di.some_sub_type.i
 ! CHECK:         }
 
@@ -50,6 +50,6 @@ subroutine data_like(x)
   type(sometype3) :: x
 end subroutine
 ! CHECK-LABEL:   fir.type_info @_QFdata_likeTsometype3
-! CHECK-SAME     component_info {
+! CHECK-SAME:    component_info {
 ! CHECK:           fir.dt_component "i" init @_QFdata_likeE.di.sometype3.i
 ! CHECK:         }

flang/test/Lower/OpenMP/DelayedPrivatization/target-private-multiple-variables.f90

@@ -148,9 +148,9 @@ end subroutine target_allocatable
 ! CHECK-SAME:     map_entries(
 ! CHECK-SAME:       %[[MAPPED_MI0]] -> %[[MAPPED_ARG0:[^,]+]],
 ! CHECK-SAME:       %[[ALLOC_VAR_MAP]] -> %[[MAPPED_ARG1:[^,]+]]
-! CHECK-SAME        %[[REAL_ARR_DESC_MAP]] -> %[[MAPPED_ARG2:[^,]+]]
-! CHECK_SAME        %[[CHAR_VAR_DESC_MAP]] -> %[[MAPPED_ARG3:.[^,]+]] :
-! CHECK-SAME        !fir.ref<i32>, !fir.ref<!fir.box<!fir.heap<i32>>>, !fir.ref<!fir.box<!fir.array<?xf32>>>, !fir.ref<!fir.boxchar<1>>)
+! CHECK-SAME:       %[[REAL_ARR_DESC_MAP]] -> %[[MAPPED_ARG2:[^,]+]]
+! CHECK-SAME:       %[[CHAR_VAR_DESC_MAP]] -> %[[MAPPED_ARG3:.[^,]+]] :
+! CHECK-SAME:       !fir.ref<i32>, !fir.ref<!fir.box<!fir.heap<i32>>>, !fir.ref<!fir.box<!fir.array<?xf32>>>, !fir.ref<!fir.boxchar<1>>)
 ! CHECK-SAME:     private(
 ! CHECK-SAME:       @[[ALLOC_PRIVATIZER_SYM]] %{{[^[:space:]]+}}#0 -> %[[ALLOC_ARG:[^,]+]] [map_idx=1],
 ! CHECK-SAME:       @[[REAL_PRIVATIZER_SYM]] %{{[^[:space:]]+}}#0 -> %[[REAL_ARG:[^,]+]],

flang/test/Lower/OpenMP/copyprivate2.f90

@@ -43,7 +43,7 @@
 !CHECK:           omp.single copyprivate(
 !CHECK-SAME:        %[[A]]#0 -> @_copy_box_heap_Uxi32 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xi32>>>>,
 !CHECK-SAME:        %[[P]]#0 -> @_copy_box_ptr_i32 : !fir.ref<!fir.box<!fir.ptr<i32>>>)
-!CHEK:          }
+!CHECK:         }
 subroutine test_alloc_ptr()
   integer, allocatable :: a(:)
   integer, pointer :: p

flang/test/Lower/OpenMP/wsloop-reduction-mul-byref.f90

@@ -27,7 +27,7 @@
 ! CHECK-SAME:    alloc {
 ! CHECK:           %[[REF:.*]] = fir.alloca i64
 ! CHECK:           omp.yield(%[[REF]] : !fir.ref<i64>)
-! CHECK-LABE:    } init {
+! CHECK-LABEL:   } init {
 ! CHECK:         ^bb0(%[[VAL_0:.*]]: !fir.ref<i64>, %[[ALLOC:.*]]: !fir.ref<i64>):
 ! CHECK:           %[[VAL_1:.*]] = arith.constant 1 : i64
 ! CHECK:           fir.store %[[VAL_1]] to %[[ALLOC]] : !fir.ref<i64>

llvm/lib/Target/AMDGPU/AMDGPUAtomicOptimizer.cpp

@@ -151,23 +151,18 @@ PreservedAnalyses AMDGPUAtomicOptimizerPass::run(Function &F,
 }
 
 bool AMDGPUAtomicOptimizerImpl::run() {
-
   // Scan option None disables the Pass
-  if (ScanImpl == ScanOptions::None) {
+  if (ScanImpl == ScanOptions::None)
     return false;
-  }
 
   visit(F);
+  if (ToReplace.empty())
+    return false;
 
-  const bool Changed = !ToReplace.empty();
-
-  for (ReplacementInfo &Info : ToReplace) {
-    optimizeAtomic(*Info.I, Info.Op, Info.ValIdx, Info.ValDivergent);
-  }
-
+  for (auto &[I, Op, ValIdx, ValDivergent] : ToReplace)
+    optimizeAtomic(*I, Op, ValIdx, ValDivergent);
   ToReplace.clear();
-
-  return Changed;
+  return true;
 }
 
 static bool isLegalCrossLaneType(Type *Ty) {
@@ -247,9 +242,7 @@ void AMDGPUAtomicOptimizerImpl::visitAtomicRMWInst(AtomicRMWInst &I) {
   // If we get here, we can optimize the atomic using a single wavefront-wide
   // atomic operation to do the calculation for the entire wavefront, so
   // remember the instruction so we can come back to it.
-  const ReplacementInfo Info = {&I, Op, ValIdx, ValDivergent};
-
-  ToReplace.push_back(Info);
+  ToReplace.push_back({&I, Op, ValIdx, ValDivergent});
 }
 
 void AMDGPUAtomicOptimizerImpl::visitIntrinsicInst(IntrinsicInst &I) {
@@ -333,17 +326,14 @@ void AMDGPUAtomicOptimizerImpl::visitIntrinsicInst(IntrinsicInst &I) {
   // If any of the other arguments to the intrinsic are divergent, we can't
   // optimize the operation.
   for (unsigned Idx = 1; Idx < I.getNumOperands(); Idx++) {
-    if (UA.isDivergentUse(I.getOperandUse(Idx))) {
+    if (UA.isDivergentUse(I.getOperandUse(Idx)))
       return;
-    }
   }
 
   // If we get here, we can optimize the atomic using a single wavefront-wide
   // atomic operation to do the calculation for the entire wavefront, so
   // remember the instruction so we can come back to it.
-  const ReplacementInfo Info = {&I, Op, ValIdx, ValDivergent};
-
-  ToReplace.push_back(Info);
+  ToReplace.push_back({&I, Op, ValIdx, ValDivergent});
 }
 
 // Use the builder to create the non-atomic counterpart of the specified