Merge branch 'upstream' into x86-insert-shuffle-combine

Author: RKSimon

clang/lib/AST/ByteCode/Compiler.cpp

@@ -5167,6 +5167,12 @@ bool Compiler<Emitter>::visitReturnStmt(const ReturnStmt *RS) {
 }
 
 template <class Emitter> bool Compiler<Emitter>::visitIfStmt(const IfStmt *IS) {
+  auto visitChildStmt = [&](const Stmt *S) -> bool {
+    LocalScope<Emitter> SScope(this);
+    if (!visitStmt(S))
+      return false;
+    return SScope.destroyLocals();
+  };
   if (auto *CondInit = IS->getInit())
     if (!visitStmt(CondInit))
       return false;
@@ -5175,7 +5181,22 @@ template <class Emitter> bool Compiler<Emitter>::visitIfStmt(const IfStmt *IS) {
     if (!visitDeclStmt(CondDecl))
       return false;
 
-  // Compile condition.
+  // Save ourselves compiling some code and the jumps, etc. if the condition is
+  // stataically known to be either true or false. We could look at more cases
+  // here, but I think all the ones that actually happen are using a
+  // ConstantExpr.
+  if (const auto *CE = dyn_cast_if_present<ConstantExpr>(IS->getCond());
+      CE && CE->hasAPValueResult() &&
+      CE->getResultAPValueKind() == APValue::ValueKind::Int) {
+    APSInt Value = CE->getResultAsAPSInt();
+    if (Value.getBoolValue())
+      return visitChildStmt(IS->getThen());
+    else if (const Stmt *Else = IS->getElse())
+      return visitChildStmt(Else);
+    return true;
+  }
+
+  // Otherwise, compile the condition.
   if (IS->isNonNegatedConsteval()) {
     if (!this->emitIsConstantContext(IS))
       return false;
@@ -5194,35 +5215,20 @@ template <class Emitter> bool Compiler<Emitter>::visitIfStmt(const IfStmt *IS) {
     LabelTy LabelEnd = this->getLabel();
     if (!this->jumpFalse(LabelElse))
       return false;
-    {
-      LocalScope<Emitter> ThenScope(this);
-      if (!visitStmt(IS->getThen()))
-        return false;
-      if (!ThenScope.destroyLocals())
-        return false;
-    }
+    if (!visitChildStmt(IS->getThen()))
+      return false;
     if (!this->jump(LabelEnd))
       return false;
     this->emitLabel(LabelElse);
-    {
-      LocalScope<Emitter> ElseScope(this);
-      if (!visitStmt(Else))
-        return false;
-      if (!ElseScope.destroyLocals())
-        return false;
-    }
+    if (!visitChildStmt(Else))
+      return false;
     this->emitLabel(LabelEnd);
   } else {
     LabelTy LabelEnd = this->getLabel();
     if (!this->jumpFalse(LabelEnd))
       return false;
-    {
-      LocalScope<Emitter> ThenScope(this);
-      if (!visitStmt(IS->getThen()))
-        return false;
-      if (!ThenScope.destroyLocals())
-        return false;
-    }
+    if (!visitChildStmt(IS->getThen()))
+      return false;
     this->emitLabel(LabelEnd);
   }
 

clang/lib/AST/TypePrinter.cpp

@@ -2093,6 +2093,9 @@ void TypePrinter::printAttributedAfter(const AttributedType *T,
   case attr::ArmMveStrictPolymorphism:
     OS << "__clang_arm_mve_strict_polymorphism";
     break;
+  case attr::ExtVectorType:
+    OS << "ext_vector_type";
+    break;
   }
   OS << "))";
 }

clang/test/Driver/hip-partial-link.hip

@@ -62,11 +62,11 @@
 // STATIC: Found undefined HIP fatbin symbol: __hip_fatbin_[[ID2:[0-9a-f]+]]
 // STATIC: Found undefined HIP gpubin handle symbol: __hip_gpubin_handle_[[ID1]]
 // STATIC: Found undefined HIP gpubin handle symbol: __hip_gpubin_handle_[[ID2]]
-// STATIC: "{{.*}}/clang-offload-bundler" {{.*}}-unbundle
-// STATIC: "{{.*}}/lld" -flavor gnu -m elf64_amdgpu
-// STATIC: "{{.*}}/clang-offload-bundler"
-// STATIC: "{{.*}}/clang{{.*}}" -target x86_64-unknown-linux-gnu
-// STATIC: "{{.*}}/llvm-ar"
+// STATIC: "{{.*[/\\]}}clang-offload-bundler" {{.*}}-unbundle
+// STATIC: "{{.*[/\\]}}lld" -flavor gnu -m elf64_amdgpu
+// STATIC: "{{.*[/\\]}}clang-offload-bundler"
+// STATIC: "{{.*[/\\]}}clang{{.*}}" -target x86_64-unknown-linux-gnu
+// STATIC: "{{.*[/\\]}}llvm-ar"
 
 // RUN: %clang -v --target=x86_64-unknown-linux-gnu --no-offload-new-driver \
 // RUN:   --hip-link -no-hip-rt -fgpu-rdc --offload-arch=gfx906 \

clang/test/Driver/linker-wrapper.c

@@ -119,7 +119,7 @@ __attribute__((visibility("protected"), used)) int x;
 
 // HIP: clang{{.*}} -o [[IMG_GFX90A:.+]] --target=amdgcn-amd-amdhsa -mcpu=gfx90a
 // HIP: clang{{.*}} -o [[IMG_GFX908:.+]] --target=amdgcn-amd-amdhsa -mcpu=gfx908
-// HIP: clang-offload-bundler{{.*}}-type=o -bundle-align=4096 -compress -compression-level=6 -targets=host-x86_64-unknown-linux-gnu,hip-amdgcn-amd-amdhsa--gfx90a,hip-amdgcn-amd-amdhsa--gfx908 -input=/dev/null -input=[[IMG_GFX90A]] -input=[[IMG_GFX908]] -output={{.*}}.hipfb
+// HIP: clang-offload-bundler{{.*}}-type=o -bundle-align=4096 -compress -compression-level=6 -targets=host-x86_64-unknown-linux-gnu,hip-amdgcn-amd-amdhsa--gfx90a,hip-amdgcn-amd-amdhsa--gfx908 -input={{/dev/null|NUL}} -input=[[IMG_GFX90A]] -input=[[IMG_GFX908]] -output={{.*}}.hipfb
 
 // RUN: clang-offload-packager -o %t.out \
 // RUN:   --image=file=%t.elf.o,kind=openmp,triple=amdgcn-amd-amdhsa,arch=gfx908 \
@@ -210,7 +210,7 @@ __attribute__((visibility("protected"), used)) int x;
 // RUN:   %t.o -o a.out 2>&1 | FileCheck %s --check-prefix=RELOCATABLE-LINK-HIP
 
 // RELOCATABLE-LINK-HIP: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa
-// RELOCATABLE-LINK-HIP: clang-offload-bundler{{.*}} -type=o -bundle-align=4096 -targets=host-x86_64-unknown-linux-gnu,hip-amdgcn-amd-amdhsa--gfx90a -input=/dev/null -input={{.*}} -output={{.*}}
+// RELOCATABLE-LINK-HIP: clang-offload-bundler{{.*}} -type=o -bundle-align=4096 -targets=host-x86_64-unknown-linux-gnu,hip-amdgcn-amd-amdhsa--gfx90a -input={{/dev/null|NUL}} -input={{.*}} -output={{.*}}
 // RELOCATABLE-LINK-HIP: /usr/bin/ld.lld{{.*}}-r
 // RELOCATABLE-LINK-HIP: llvm-objcopy{{.*}}a.out --remove-section .llvm.offloading
 

clang/tools/clang-linker-wrapper/ClangLinkerWrapper.cpp

@@ -603,9 +603,11 @@ Expected<StringRef> writeOffloadFile(const OffloadFile &File) {
 
   StringRef Prefix =
       sys::path::stem(Binary.getMemoryBufferRef().getBufferIdentifier());
-
-  auto TempFileOrErr = createOutputFile(
-      Prefix + "-" + Binary.getTriple() + "-" + Binary.getArch(), "o");
+  SmallString<128> Filename;
+  (Prefix + "-" + Binary.getTriple() + "-" + Binary.getArch())
+      .toVector(Filename);
+  llvm::replace(Filename, ':', '-');
+  auto TempFileOrErr = createOutputFile(Filename, "o");
   if (!TempFileOrErr)
     return TempFileOrErr.takeError();
 

flang/lib/Optimizer/Analysis/AliasAnalysis.cpp

@@ -51,40 +51,6 @@ static bool hasGlobalOpTargetAttr(mlir::Value v, fir::AddrOfOp op) {
       v, fir::GlobalOp::getTargetAttrName(globalOpName));
 }
 
-static mlir::Value
-getOriginalDef(mlir::Value v,
-               fir::AliasAnalysis::Source::Attributes &attributes,
-               bool &isCapturedInInternalProcedure, bool &approximateSource) {
-  mlir::Operation *defOp;
-  bool breakFromLoop = false;
-  while (!breakFromLoop && (defOp = v.getDefiningOp())) {
-    mlir::Type ty = defOp->getResultTypes()[0];
-    llvm::TypeSwitch<Operation *>(defOp)
-        .Case<fir::ConvertOp>([&](fir::ConvertOp op) { v = op.getValue(); })
-        .Case<fir::DeclareOp, hlfir::DeclareOp>([&](auto op) {
-          v = op.getMemref();
-          auto varIf = llvm::cast<fir::FortranVariableOpInterface>(defOp);
-          attributes |= getAttrsFromVariable(varIf);
-          isCapturedInInternalProcedure |=
-              varIf.isCapturedInInternalProcedure();
-        })
-        .Case<fir::CoordinateOp>([&](auto op) {
-          if (fir::AliasAnalysis::isPointerReference(ty))
-            attributes.set(fir::AliasAnalysis::Attribute::Pointer);
-          v = op->getOperand(0);
-          approximateSource = true;
-        })
-        .Case<hlfir::DesignateOp>([&](hlfir::DesignateOp op) {
-          auto varIf = llvm::cast<fir::FortranVariableOpInterface>(defOp);
-          attributes |= getAttrsFromVariable(varIf);
-          v = op.getMemref();
-          approximateSource = true;
-        })
-        .Default([&](auto op) { breakFromLoop = true; });
-  }
-  return v;
-}
-
 static bool isEvaluateInMemoryBlockArg(mlir::Value v) {
   if (auto evalInMem = llvm::dyn_cast_or_null<hlfir::EvaluateInMemoryOp>(
           v.getParentRegion()->getParentOp()))
@@ -621,38 +587,34 @@ AliasAnalysis::Source AliasAnalysis::getSource(mlir::Value v,
             if (mlir::isa<fir::PointerType>(boxTy.getEleTy()))
               attributes.set(Attribute::Pointer);
 
-            auto def = getOriginalDef(op.getMemref(), attributes,
-                                      isCapturedInInternalProcedure,
-                                      approximateSource);
-            if (auto addrOfOp = def.template getDefiningOp<fir::AddrOfOp>()) {
-              global = addrOfOp.getSymbol();
-
-              if (hasGlobalOpTargetAttr(def, addrOfOp))
-                attributes.set(Attribute::Target);
+            auto boxSrc = getSource(op.getMemref());
+            attributes |= boxSrc.attributes;
+            approximateSource |= boxSrc.approximateSource;
+            isCapturedInInternalProcedure |=
+                boxSrc.isCapturedInInternalProcedure;
 
+            global = llvm::dyn_cast<mlir::SymbolRefAttr>(boxSrc.origin.u);
+            if (global) {
               type = SourceKind::Global;
-            }
-            // TODO: Add support to fir.allocmem
-            else if (auto allocOp =
-                         def.template getDefiningOp<fir::AllocaOp>()) {
-              v = def;
-              defOp = v.getDefiningOp();
-              type = SourceKind::Allocate;
-            } else if (isDummyArgument(def)) {
-              defOp = nullptr;
-              v = def;
             } else {
-              type = SourceKind::Indirect;
+              auto def = llvm::cast<mlir::Value>(boxSrc.origin.u);
+              // TODO: Add support to fir.allocmem
+              if (auto allocOp = def.template getDefiningOp<fir::AllocaOp>()) {
+                v = def;
+                defOp = v.getDefiningOp();
+                type = SourceKind::Allocate;
+              } else if (isDummyArgument(def)) {
+                defOp = nullptr;
+                v = def;
+              } else {
+                type = SourceKind::Indirect;
+              }
             }
-            // TODO: This assignment is redundant but somehow works around an
-            // apparent MSVC bug reporting "undeclared identifier" at the next
-            // "breakFromLoop = true;".  See
-            // <https://github.com/llvm/llvm-project/pull/127845#issuecomment-2669829610>.
             breakFromLoop = true;
-          } else {
-            // No further tracking for addresses loaded from memory for now.
-            type = SourceKind::Indirect;
+            return;
           }
+          // No further tracking for addresses loaded from memory for now.
+          type = SourceKind::Indirect;
           breakFromLoop = true;
         })
         .Case<fir::AddrOfOp>([&](auto op) {

flang/test/Analysis/AliasAnalysis/source-kind.fir

@@ -0,0 +1,25 @@
+// RUN: fir-opt %s -pass-pipeline='builtin.module(func.func(test-fir-alias-analysis))' -debug-only=fir-alias-analysis 2>&1 | FileCheck %s
+
+// CHECK-LABEL: Testing : "_QFPtest"
+
+// Checking that the source kind of a load of a load is SourceKind::Indirect
+// CHECK: {test.ptr = "load_load"}
+// CHECK-NEXT: SourceKind: Indirect
+
+// Checking that the source kind of a load of an arg is SourceKind::Argument
+// CHECK: {test.ptr = "load_arg"}
+// CHECK-NEXT: SourceKind: Argument
+
+func.func @_QFPtest(%arg0: !fir.ref<!fir.box<!fir.ptr<f32>>> ) attributes {test.ptr = "func"} {
+  
+  %0 = fir.alloca !fir.llvm_ptr<!fir.box<!fir.ptr<f32>>>
+  %1 = fir.convert %arg0 : (!fir.ref<!fir.box<!fir.ptr<f32>>>) -> !fir.llvm_ptr<!fir.box<!fir.ptr<f32>>>
+  fir.store %1 to %0 : !fir.ref<!fir.llvm_ptr<!fir.box<!fir.ptr<f32>>>>
+  %2 = fir.load %0 : !fir.ref<!fir.llvm_ptr<!fir.box<!fir.ptr<f32>>>>
+  %3 = fir.convert %2 : (!fir.llvm_ptr<!fir.box<!fir.ptr<f32>>>) -> !fir.ref<!fir.box<!fir.ptr<f32>>>
+  %15 = fir.load %3 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+  %16 = fir.box_addr %15 {test.ptr = "load_load"} : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
+  %17 = fir.load %arg0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+  %18 = fir.box_addr %17 {test.ptr = "load_arg"} : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
+  return
+}

lld/COFF/Chunks.h

@@ -59,7 +59,8 @@ class Chunk {
     SectionKind,
     SectionECKind,
     OtherKind,
-    ImportThunkKind
+    ImportThunkKind,
+    ECExportThunkKind
   };
   Kind kind() const { return chunkKind; }
 
@@ -827,7 +828,10 @@ static const uint8_t ECExportThunkCode[] = {
 
 class ECExportThunkChunk : public NonSectionCodeChunk {
 public:
-  explicit ECExportThunkChunk(Defined *targetSym) : target(targetSym) {}
+  explicit ECExportThunkChunk(Defined *targetSym)
+      : NonSectionCodeChunk(ECExportThunkKind), target(targetSym) {}
+  static bool classof(const Chunk *c) { return c->kind() == ECExportThunkKind; }
+
   size_t getSize() const override { return sizeof(ECExportThunkCode); };
   void writeTo(uint8_t *buf) const override;
   MachineTypes getMachine() const override { return AMD64; }

lld/test/COFF/build-id-sym.s

@@ -1,6 +1,6 @@
 # REQUIRES: x86
 # RUN: llvm-mc -triple=x86_64-windows-msvc -filetype=obj -o %t.obj %s
-# RUN: lld-link -debug:symtab -entry:main %t.obj -build-id -Brepro -out:%t.exe
+# RUN: lld-link -debug:symtab -entry:main %t.obj -build-id -Brepro -out:%t.exe -guard:cf
 # RUN: llvm-objdump -s -t %t.exe | FileCheck %s
 
 # Check __buildid points to 0x14000203c which is after the signature RSDS.
@@ -21,3 +21,6 @@ main:
 .section .bss,"bw",discard,__buildid
 .global __buildid
 __buildid:
+
+.data
+.quad __buildid

llvm/include/llvm/ExecutionEngine/Orc/MemoryMapper.h

@@ -157,6 +157,7 @@ class SharedMemoryMapper final : public MemoryMapper {
   struct Reservation {
     void *LocalAddr;
     size_t Size;
+    int SharedMemoryId;
   };
 
   ExecutorProcessControl &EPC;