[𝘀𝗽𝗿] changes introduced through rebaseupstream/users/vitalybuka/spr/main.testhwasan-implement-sanitizer_specific-for-hwasan

Created using spr 1.3.4

[skip ci]

11 files changed, 91 insertions, 183 deletions

diff --git a/clang/tools/clang-format/git-clang-format.bat b/clang/tools/clang-format/git-clang-format.bat
index d4bc5172989c..9965cd4312fe 100644
--- a/clang/tools/clang-format/git-clang-format.bat
+++ b/clang/tools/clang-format/git-clang-format.bat
@@ -1 +1 @@
-py -3 git-clang-format %*
+py -3 %~pn0 %*
diff --git a/llvm/include/llvm/CodeGen/GlobalISel/GIMatchTableExecutor.h b/llvm/include/llvm/CodeGen/GlobalISel/GIMatchTableExecutor.h
index 73308925e914..694d3d8004af 100644
--- a/llvm/include/llvm/CodeGen/GlobalISel/GIMatchTableExecutor.h
+++ b/llvm/include/llvm/CodeGen/GlobalISel/GIMatchTableExecutor.h
@@ -690,7 +690,9 @@ protected:
   bool isObviouslySafeToFold(MachineInstr &MI, MachineInstr &IntoMI) const;
 
   template <typename Ty> static Ty readBytesAs(const uint8_t *MatchTable) {
-    return *reinterpret_cast<const Ty *>(MatchTable);
+    Ty Ret;
+    memcpy(&Ret, MatchTable, sizeof(Ret));
+    return Ret;
   }
 };
 
diff --git a/llvm/lib/Target/RISCV/RISCVISelLowering.cpp b/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
index c0462dc4b0bc..a75bbb49beda 100644
--- a/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
+++ b/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
@@ -2624,6 +2624,25 @@ SDValue RISCVTargetLowering::computeVLMax(MVT VecVT, const SDLoc &DL,
                              VecVT.getVectorElementCount());
 }
 
+std::pair<unsigned, unsigned>
+RISCVTargetLowering::computeVLMAXBounds(MVT VecVT,
+                                        const RISCVSubtarget &Subtarget) {
+  assert(VecVT.isScalableVector() && "Expected scalable vector");
+
+  unsigned EltSize = VecVT.getScalarSizeInBits();
+  unsigned MinSize = VecVT.getSizeInBits().getKnownMinValue();
+
+  unsigned VectorBitsMax = Subtarget.getRealMaxVLen();
+  unsigned MaxVLMAX =
+      RISCVTargetLowering::computeVLMAX(VectorBitsMax, EltSize, MinSize);
+
+  unsigned VectorBitsMin = Subtarget.getRealMinVLen();
+  unsigned MinVLMAX =
+      RISCVTargetLowering::computeVLMAX(VectorBitsMin, EltSize, MinSize);
+
+  return std::make_pair(MinVLMAX, MaxVLMAX);
+}
+
 // The state of RVV BUILD_VECTOR and VECTOR_SHUFFLE lowering is that very few
 // of either is (currently) supported. This can get us into an infinite loop
 // where we try to lower a BUILD_VECTOR as a VECTOR_SHUFFLE as a BUILD_VECTOR
@@ -8123,16 +8142,8 @@ static SDValue lowerVectorIntrinsicScalars(SDValue Op, SelectionDAG &DAG,
 
     // Optimize for constant AVL
     if (isa<ConstantSDNode>(AVL)) {
-      unsigned EltSize = VT.getScalarSizeInBits();
-      unsigned MinSize = VT.getSizeInBits().getKnownMinValue();
-
-      unsigned VectorBitsMax = Subtarget.getRealMaxVLen();
-      unsigned MaxVLMAX =
-          RISCVTargetLowering::computeVLMAX(VectorBitsMax, EltSize, MinSize);
-
-      unsigned VectorBitsMin = Subtarget.getRealMinVLen();
-      unsigned MinVLMAX =
-          RISCVTargetLowering::computeVLMAX(VectorBitsMin, EltSize, MinSize);
+      const auto [MinVLMAX, MaxVLMAX] =
+          RISCVTargetLowering::computeVLMAXBounds(VT, Subtarget);
 
       uint64_t AVLInt = cast<ConstantSDNode>(AVL)->getZExtValue();
       if (AVLInt <= MinVLMAX) {
diff --git a/llvm/lib/Target/RISCV/RISCVISelLowering.h b/llvm/lib/Target/RISCV/RISCVISelLowering.h
index ae798cc47bf8..41a2dc5771c8 100644
--- a/llvm/lib/Target/RISCV/RISCVISelLowering.h
+++ b/llvm/lib/Target/RISCV/RISCVISelLowering.h
@@ -744,7 +744,13 @@ public:
     // The following equations have been reordered to prevent loss of precision
     // when calculating fractional LMUL.
     return ((VectorBits / EltSize) * MinSize) / RISCV::RVVBitsPerBlock;
-  };
+  }
+
+  // Return inclusive (low, high) bounds on the value of VLMAX for the
+  // given scalable container type given known bounds on VLEN.
+  static std::pair<unsigned, unsigned>
+  computeVLMAXBounds(MVT ContainerVT, const RISCVSubtarget &Subtarget);
+
   static unsigned getRegClassIDForLMUL(RISCVII::VLMUL LMul);
   static unsigned getSubregIndexByMVT(MVT VT, unsigned Index);
   static unsigned getRegClassIDForVecVT(MVT VT);
diff --git a/llvm/lib/Target/X86/X86FastISel.cpp b/llvm/lib/Target/X86/X86FastISel.cpp
index bdc9d1d42dd1..425c52dbe7b1 100644
--- a/llvm/lib/Target/X86/X86FastISel.cpp
+++ b/llvm/lib/Target/X86/X86FastISel.cpp
@@ -711,7 +711,8 @@ bool X86FastISel::handleConstantAddresses(const Value *V, X86AddressMode &AM) {
   // Handle constant address.
   if (const GlobalValue *GV = dyn_cast<GlobalValue>(V)) {
     // Can't handle alternate code models yet.
-    if (TM.getCodeModel() != CodeModel::Small)
+    if (TM.getCodeModel() != CodeModel::Small &&
+        TM.getCodeModel() != CodeModel::Medium)
       return false;
 
     // Can't handle large objects yet.
@@ -1050,7 +1051,8 @@ bool X86FastISel::X86SelectCallAddress(const Value *V, X86AddressMode &AM) {
   // Handle constant address.
   if (const GlobalValue *GV = dyn_cast<GlobalValue>(V)) {
     // Can't handle alternate code models yet.
-    if (TM.getCodeModel() != CodeModel::Small)
+    if (TM.getCodeModel() != CodeModel::Small &&
+        TM.getCodeModel() != CodeModel::Medium)
       return false;
 
     // RIP-relative addresses can't have additional register operands.
@@ -3774,7 +3776,8 @@ unsigned X86FastISel::X86MaterializeFP(const ConstantFP *CFP, MVT VT) {
 
   // Can't handle alternate code models yet.
   CodeModel::Model CM = TM.getCodeModel();
-  if (CM != CodeModel::Small && CM != CodeModel::Large)
+  if (CM != CodeModel::Small && CM != CodeModel::Medium &&
+      CM != CodeModel::Large)
     return 0;
 
   // Get opcode and regclass of the output for the given load instruction.
@@ -3812,7 +3815,7 @@ unsigned X86FastISel::X86MaterializeFP(const ConstantFP *CFP, MVT VT) {
     PICBase = getInstrInfo()->getGlobalBaseReg(FuncInfo.MF);
   else if (OpFlag == X86II::MO_GOTOFF)
     PICBase = getInstrInfo()->getGlobalBaseReg(FuncInfo.MF);
-  else if (Subtarget->is64Bit() && TM.getCodeModel() == CodeModel::Small)
+  else if (Subtarget->is64Bit() && TM.getCodeModel() != CodeModel::Large)
     PICBase = X86::RIP;
 
   // Create the load from the constant pool.
@@ -3842,8 +3845,11 @@ unsigned X86FastISel::X86MaterializeFP(const ConstantFP *CFP, MVT VT) {
 }
 
 unsigned X86FastISel::X86MaterializeGV(const GlobalValue *GV, MVT VT) {
-  // Can't handle alternate code models yet.
-  if (TM.getCodeModel() != CodeModel::Small)
+  // Can't handle large GlobalValues yet.
+  if (TM.getCodeModel() != CodeModel::Small &&
+      TM.getCodeModel() != CodeModel::Medium)
+    return 0;
+  if (!isa<GlobalObject>(GV) || TM.isLargeGlobalObject(cast<GlobalObject>(GV)))
     return 0;
 
   // Materialize addresses with LEA/MOV instructions.
diff --git a/llvm/test/CodeGen/X86/fast-isel-constpool.ll b/llvm/test/CodeGen/X86/fast-isel-constpool.ll
index 9f70fd559073..9e4cbb61308e 100644
--- a/llvm/test/CodeGen/X86/fast-isel-constpool.ll
+++ b/llvm/test/CodeGen/X86/fast-isel-constpool.ll
@@ -1,5 +1,6 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -mtriple=x86_64-apple-darwin -fast-isel -code-model=small < %s | FileCheck %s
+; RUN: llc -mtriple=x86_64-apple-darwin -fast-isel -code-model=medium < %s | FileCheck %s
 ; RUN: llc -mtriple=x86_64-apple-darwin -fast-isel -code-model=large < %s | FileCheck %s --check-prefix=LARGE
 ; RUN: llc -mtriple=x86_64 -fast-isel -code-model=large -relocation-model=pic < %s | FileCheck %s --check-prefix=LARGE_PIC
 ; RUN: llc -mtriple=x86_64-apple-darwin -fast-isel -code-model=small -mattr=avx < %s | FileCheck %s --check-prefix=AVX
diff --git a/llvm/test/CodeGen/X86/fast-isel-medium-code-model.ll b/llvm/test/CodeGen/X86/fast-isel-medium-code-model.ll
new file mode 100644
index 000000000000..4aa230a209e2
--- /dev/null
+++ b/llvm/test/CodeGen/X86/fast-isel-medium-code-model.ll
@@ -0,0 +1,44 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=x86_64-linux-gnu -fast-isel -fast-isel-abort=3 -code-model=medium -large-data-threshold=5 < %s | FileCheck %s
+; RUN: llc -mtriple=x86_64-linux-gnu -fast-isel -code-model=medium -large-data-threshold=3 < %s -o /dev/null \
+; RUN:   -pass-remarks-output=- -pass-remarks-filter=sdagisel | FileCheck %s --check-prefix=FALLBACK --implicit-check-not=missed
+
+declare void @foo()
+
+define void @call_foo() {
+; CHECK-LABEL: call_foo:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    pushq %rax
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    callq foo@PLT
+; CHECK-NEXT:    popq %rax
+; CHECK-NEXT:    .cfi_def_cfa_offset 8
+; CHECK-NEXT:    retq
+  call void @foo()
+  ret void
+}
+
+@g = internal global i32 42
+
+; FALLBACK: FastISel missed terminator
+; FALLBACK: in function: g_addr
+
+define ptr @g_addr() {
+; CHECK-LABEL: g_addr:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    movabsq $g, %rax
+; CHECK-NEXT:    retq
+  ret ptr @g
+}
+
+; FALLBACK: FastISel missed
+; FALLBACK: in function: load_g
+
+define i32 @load_g() {
+; CHECK-LABEL: load_g:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    movl g, %eax
+; CHECK-NEXT:    retq
+  %i = load i32, ptr @g
+  ret i32 %i
+}
diff --git a/mlir/include/mlir/Transforms/FoldUtils.h b/mlir/include/mlir/Transforms/FoldUtils.h
index 28fa18cf942d..2600da361496 100644
--- a/mlir/include/mlir/Transforms/FoldUtils.h
+++ b/mlir/include/mlir/Transforms/FoldUtils.h
@@ -33,8 +33,7 @@ class Value;
 class OperationFolder {
 public:
   OperationFolder(MLIRContext *ctx, OpBuilder::Listener *listener = nullptr)
-      : fusedLocationTag(StringAttr::get(ctx, "CSE")), interfaces(ctx),
-        rewriter(ctx, listener) {}
+      : interfaces(ctx), rewriter(ctx, listener) {}
 
   /// Tries to perform folding on the given `op`, including unifying
   /// deduplicated constants. If successful, replaces `op`'s uses with
@@ -96,15 +95,6 @@ private:
                                     Dialect *dialect, Attribute value,
                                     Type type, Location loc);
 
-  // Fuse `foldedLocation` into the Location of `retainedOp`. This will result
-  // in `retainedOp` having a FusedLoc with `fusedLocationTag` to help trace the
-  // source of the fusion. If `retainedOp` already had a FusedLoc with the same
-  // tag, `foldedLocation` will simply be appended to it.
-  void appendFoldedLocation(Operation *retainedOp, Location foldedLocation);
-
-  /// Tag for annotating fused locations as a result of merging constants.
-  StringAttr fusedLocationTag;
-
   /// A mapping between an insertion region and the constants that have been
   /// created within it.
   DenseMap<Region *, ConstantMap> foldScopes;
diff --git a/mlir/lib/Transforms/Utils/FoldUtils.cpp b/mlir/lib/Transforms/Utils/FoldUtils.cpp
index 136c4d2216b8..90ee5ba51de3 100644
--- a/mlir/lib/Transforms/Utils/FoldUtils.cpp
+++ b/mlir/lib/Transforms/Utils/FoldUtils.cpp
@@ -141,7 +141,6 @@ bool OperationFolder::insertKnownConstant(Operation *op, Attribute constValue) {
   // If there is an existing constant, replace `op`.
   if (folderConstOp) {
     notifyRemoval(op);
-    appendFoldedLocation(folderConstOp, op->getLoc());
     rewriter.replaceOp(op, folderConstOp->getResults());
     return false;
   }
@@ -295,10 +294,8 @@ OperationFolder::tryGetOrCreateConstant(ConstantMap &uniquedConstants,
   // Check if an existing mapping already exists.
   auto constKey = std::make_tuple(dialect, value, type);
   Operation *&constOp = uniquedConstants[constKey];
-  if (constOp) {
-    appendFoldedLocation(constOp, loc);
+  if (constOp)
     return constOp;
-  }
 
   // If one doesn't exist, try to materialize one.
   if (!(constOp = materializeConstant(dialect, rewriter, value, type, loc)))
@@ -319,7 +316,6 @@ OperationFolder::tryGetOrCreateConstant(ConstantMap &uniquedConstants,
   // materialized operation in favor of the existing one.
   if (auto *existingOp = uniquedConstants.lookup(newKey)) {
     notifyRemoval(constOp);
-    appendFoldedLocation(existingOp, constOp->getLoc());
     rewriter.eraseOp(constOp);
     referencedDialects[existingOp].push_back(dialect);
     return constOp = existingOp;
@@ -330,76 +326,3 @@ OperationFolder::tryGetOrCreateConstant(ConstantMap &uniquedConstants,
   auto newIt = uniquedConstants.insert({newKey, constOp});
   return newIt.first->second;
 }
-
-/// Helper that flattens nested fused locations to a single fused location.
-/// Fused locations nested under non-fused locations are not flattened, and
-/// calling this on non-fused locations is a no-op as a result.
-///
-/// Fused locations are only flattened into parent fused locations if the
-/// child fused location has no metadata, or if the metadata of the parent and
-/// child fused locations are the same---this to avoid breaking cases where
-/// metadata matter.
-static Location FlattenFusedLocationRecursively(const Location loc) {
-  auto fusedLoc = dyn_cast<FusedLoc>(loc);
-  if (!fusedLoc)
-    return loc;
-
-  SetVector<Location> flattenedLocs;
-  Attribute metadata = fusedLoc.getMetadata();
-  ArrayRef<Location> unflattenedLocs = fusedLoc.getLocations();
-  bool hasAnyNestedLocChanged = false;
-
-  for (const Location &unflattenedLoc : unflattenedLocs) {
-    Location flattenedLoc = FlattenFusedLocationRecursively(unflattenedLoc);
-
-    auto flattenedFusedLoc = dyn_cast<FusedLoc>(flattenedLoc);
-    if (flattenedFusedLoc && (!flattenedFusedLoc.getMetadata() ||
-                              flattenedFusedLoc.getMetadata() == metadata)) {
-      hasAnyNestedLocChanged = true;
-      ArrayRef<Location> nestedLocations = flattenedFusedLoc.getLocations();
-      flattenedLocs.insert(nestedLocations.begin(), nestedLocations.end());
-    } else {
-      if (flattenedLoc != unflattenedLoc)
-        hasAnyNestedLocChanged = true;
-
-      flattenedLocs.insert(flattenedLoc);
-    }
-  }
-
-  if (!hasAnyNestedLocChanged &&
-      unflattenedLocs.size() == flattenedLocs.size()) {
-    return loc;
-  }
-
-  return FusedLoc::get(loc->getContext(), flattenedLocs.takeVector(),
-                       fusedLoc.getMetadata());
-}
-
-void OperationFolder::appendFoldedLocation(Operation *retainedOp,
-                                           Location foldedLocation) {
-  // Append into existing fused location if it has the same tag.
-  if (auto existingFusedLoc =
-          dyn_cast<FusedLocWith<StringAttr>>(retainedOp->getLoc())) {
-    StringAttr existingMetadata = existingFusedLoc.getMetadata();
-    if (existingMetadata == fusedLocationTag) {
-      ArrayRef<Location> existingLocations = existingFusedLoc.getLocations();
-      SetVector<Location> locations(existingLocations.begin(),
-                                    existingLocations.end());
-      locations.insert(foldedLocation);
-      Location newFusedLoc = FusedLoc::get(
-          retainedOp->getContext(), locations.takeVector(), existingMetadata);
-      retainedOp->setLoc(FlattenFusedLocationRecursively(newFusedLoc));
-      return;
-    }
-  }
-
-  // Create a new fusedloc with retainedOp's loc and foldedLocation.
-  // If they're already equal, no need to fuse.
-  if (retainedOp->getLoc() == foldedLocation)
-    return;
-
-  Location newFusedLoc =
-      FusedLoc::get(retainedOp->getContext(),
-                    {retainedOp->getLoc(), foldedLocation}, fusedLocationTag);
-  retainedOp->setLoc(FlattenFusedLocationRecursively(newFusedLoc));
-}
diff --git a/mlir/test/Transforms/canonicalize-debuginfo.mlir b/mlir/test/Transforms/canonicalize-debuginfo.mlir
deleted file mode 100644
index 217cc29c0095..000000000000
--- a/mlir/test/Transforms/canonicalize-debuginfo.mlir
+++ /dev/null
@@ -1,41 +0,0 @@
-// RUN: mlir-opt %s -pass-pipeline='builtin.module(func.func(canonicalize{test-convergence}))' -split-input-file -mlir-print-debuginfo | FileCheck %s
-
-// CHECK-LABEL: func @merge_constants
-func.func @merge_constants() -> (index, index, index, index, index, index, index) {
-  // CHECK-NEXT: arith.constant 42 : index loc(#[[FusedLoc:.*]])
-  %0 = arith.constant 42 : index loc("merge_constants":0:0)
-  %1 = arith.constant 42 : index loc("merge_constants":1:0)
-  %2 = arith.constant 42 : index loc("merge_constants":2:0)
-  %3 = arith.constant 42 : index loc("merge_constants":2:0) // repeated loc
-  %4 = arith.constant 43 : index loc(fused<"some_label">["merge_constants":3:0])
-  %5 = arith.constant 43 : index loc(fused<"some_label">["merge_constants":3:0])
-  %6 = arith.constant 43 : index loc(fused<"some_other_label">["merge_constants":3:0])
-  return %0, %1, %2, %3, %4, %5, %6 : index, index, index, index, index, index, index
-}
-
-// CHECK-DAG: #[[LocConst0:.*]] = loc("merge_constants":0:0)
-// CHECK-DAG: #[[LocConst1:.*]] = loc("merge_constants":1:0)
-// CHECK-DAG: #[[LocConst2:.*]] = loc("merge_constants":2:0)
-// CHECK-DAG: #[[LocConst3:.*]] = loc("merge_constants":3:0)
-// CHECK-DAG: #[[FusedLoc_CSE_1:.*]] = loc(fused<"CSE">[#[[LocConst0]], #[[LocConst1]], #[[LocConst2]]])
-// CHECK-DAG: #[[FusedLoc_Some_Label:.*]] = loc(fused<"some_label">[#[[LocConst3]]])
-// CHECK-DAG: #[[FusedLoc_Some_Other_Label:.*]] = loc(fused<"some_other_label">[#[[LocConst3]]])
-// CHECK-DAG: #[[FusedLoc_CSE_2:.*]] = loc(fused<"CSE">[#[[FusedLoc_Some_Label]], #[[FusedLoc_Some_Other_Label]]])
-
-// -----
-
-// CHECK-LABEL: func @hoist_constant
-func.func @hoist_constant(%arg0: memref<8xi32>) {
-  // CHECK-NEXT: arith.constant 42 : i32 loc(#[[FusedLoc:.*]])
-  affine.for %arg1 = 0 to 8 {
-    %0 = arith.constant 42 : i32 loc("hoist_constant":0:0)
-    %1 = arith.constant 42 : i32 loc("hoist_constant":1:0)
-    memref.store %0, %arg0[%arg1] : memref<8xi32>
-    memref.store %1, %arg0[%arg1] : memref<8xi32>
-  }
-  return
-}
-
-// CHECK-DAG: #[[LocConst0:.*]] = loc("hoist_constant":0:0)
-// CHECK-DAG: #[[LocConst1:.*]] = loc("hoist_constant":1:0)
-// CHECK: #[[FusedLoc]] = loc(fused<"CSE">[#[[LocConst0]], #[[LocConst1]]])
diff --git a/mlir/test/Transforms/constant-fold-debuginfo.mlir b/mlir/test/Transforms/constant-fold-debuginfo.mlir
deleted file mode 100644
index 79a25f860a48..000000000000
--- a/mlir/test/Transforms/constant-fold-debuginfo.mlir
+++ /dev/null
@@ -1,34 +0,0 @@
-// RUN: mlir-opt %s -split-input-file -test-constant-fold -mlir-print-debuginfo | FileCheck %s
-
-// CHECK-LABEL: func @fold_and_merge
-func.func @fold_and_merge() -> (i32, i32) {
-  %0 = arith.constant 1 : i32
-  %1 = arith.constant 5 : i32
-
-  // CHECK-NEXT: [[C:%.+]] = arith.constant 6 : i32 loc(#[[FusedLoc:.*]])
-  %2 = arith.addi %0, %1 : i32 loc("fold_and_merge":0:0)
-
-  %3 = arith.constant 6 : i32 loc("fold_and_merge":1:0)
-
-  return %2, %3: i32, i32
-}
-
-// CHECK-DAG: #[[LocConst0:.*]] = loc("fold_and_merge":0:0)
-// CHECK-DAG: #[[LocConst1:.*]] = loc("fold_and_merge":1:0)
-// CHECK: #[[FusedLoc]] = loc(fused<"CSE">[#[[LocConst1]], #[[LocConst0]]])
-
-// -----
-
-// CHECK-LABEL: func @materialize_different_dialect
-func.func @materialize_different_dialect() -> (f32, f32) {
-  // CHECK: arith.constant 1.{{0*}}e+00 : f32 loc(#[[FusedLoc:.*]])
-  %0 = arith.constant -1.0 : f32
-  %1 = math.absf %0 : f32 loc("materialize_different_dialect":0:0)
-  %2 = arith.constant 1.0 : f32 loc("materialize_different_dialect":1:0)
-
-  return %1, %2: f32, f32
-}
-
-// CHECK-DAG: #[[LocConst0:.*]] = loc("materialize_different_dialect":0:0)
-// CHECK-DAG: #[[LocConst1:.*]] = loc("materialize_different_dialect":1:0)
-// CHECK: #[[FusedLoc]] = loc(fused<"CSE">[#[[LocConst1]], #[[LocConst0]]])