Merge remote-tracking branch 'origin/sycl-web' into llvmspirv_pulldown

Author: sys-ce-bb

.ci/generate-buildkite-pipeline-premerge

@@ -85,6 +85,22 @@ function compute-projects-to-test() {
   done
 }
 
+function compute-runtimes-to-test() {
+  projects=${@}
+  for project in ${projects}; do
+    case ${project} in
+    clang)
+      for p in libcxx libcxxabi libunwind; do
+        echo $p
+      done
+    ;;
+    *)
+      # Nothing to do
+    ;;
+    esac
+  done
+}
+
 function add-dependencies() {
   projects=${@}
   for project in ${projects}; do
@@ -178,6 +194,15 @@ function check-targets() {
     cross-project-tests)
       echo "check-cross-project"
     ;;
+    libcxx)
+      echo "check-cxx"
+    ;;
+    libcxxabi)
+      echo "check-cxxabi"
+    ;;
+    libunwind)
+      echo "check-unwind"
+    ;;
     lldb)
       echo "check-all" # TODO: check-lldb may not include all the LLDB tests?
     ;;
@@ -207,17 +232,6 @@ if echo "$modified_dirs" | grep -q -E "^(libcxx|libcxxabi|libunwind|runtimes|cma
 EOF
 fi
 
-# If clang changed.
-if echo "$modified_dirs" | grep -q -E "^(clang)$"; then
-  cat <<EOF
-- trigger: "clang-ci"
-  build:
-    message: "${buildMessage}"
-    commit: "${BUILDKITE_COMMIT}"
-    branch: "${BUILDKITE_BRANCH}"
-EOF
-fi
-
 # Generic pipeline for projects that have not defined custom steps.
 #
 # Individual projects should instead define the pre-commit CI tests that suits their
@@ -231,6 +245,10 @@ linux_projects_to_test=$(exclude-linux $(compute-projects-to-test ${modified_pro
 linux_check_targets=$(check-targets ${linux_projects_to_test} | sort | uniq)
 linux_projects=$(add-dependencies ${linux_projects_to_test} | sort | uniq)
 
+linux_runtimes_to_test=$(compute-runtimes-to-test ${linux_projects_to_test})
+linux_runtime_check_targets=$(check-targets ${linux_runtimes_to_test} | sort | uniq)
+linux_runtimes=$(echo ${linux_runtimes_to_test} | sort | uniq)
+
 windows_projects_to_test=$(exclude-windows $(compute-projects-to-test ${modified_projects}))
 windows_check_targets=$(check-targets ${windows_projects_to_test} | sort | uniq)
 windows_projects=$(add-dependencies ${windows_projects_to_test} | sort | uniq)
@@ -255,7 +273,7 @@ if [[ "${linux_projects}" != "" ]]; then
     CC: 'clang'
     CXX: 'clang++'
   commands:
-  - './.ci/monolithic-linux.sh "$(echo ${linux_projects} | tr ' ' ';')" "$(echo ${linux_check_targets})"'
+  - './.ci/monolithic-linux.sh "$(echo ${linux_projects} | tr ' ' ';')" "$(echo ${linux_check_targets})" "$(echo ${linux_runtimes} | tr ' ' ';')" "$(echo ${linux_runtime_check_targets})"'
 EOF
 fi
 

.ci/monolithic-linux.sh

@@ -18,6 +18,7 @@ set -o pipefail
 
 MONOREPO_ROOT="${MONOREPO_ROOT:="$(git rev-parse --show-toplevel)"}"
 BUILD_DIR="${BUILD_DIR:=${MONOREPO_ROOT}/build}"
+INSTALL_DIR="${BUILD_DIR}/install"
 rm -rf "${BUILD_DIR}"
 
 ccache --zero-stats
@@ -49,8 +50,79 @@ cmake -S "${MONOREPO_ROOT}"/llvm -B "${BUILD_DIR}" \
       -D LLVM_ENABLE_LLD=ON \
       -D CMAKE_CXX_FLAGS=-gmlt \
       -D LLVM_CCACHE_BUILD=ON \
-      -D MLIR_ENABLE_BINDINGS_PYTHON=ON
+      -D MLIR_ENABLE_BINDINGS_PYTHON=ON \
+      -D CMAKE_INSTALL_PREFIX="${INSTALL_DIR}"
 
 echo "--- ninja"
 # Targets are not escaped as they are passed as separate arguments.
 ninja -C "${BUILD_DIR}" -k 0 ${targets}
+
+runtimes="${3}"
+runtime_targets="${4}"
+
+# Compiling runtimes with just-built Clang and running their tests
+# as an additional testing for Clang.
+if [[ "${runtimes}" != "" ]]; then
+  if [[ "${runtime_targets}" == "" ]]; then
+    echo "Runtimes to build are specified, but targets are not."
+    exit 1
+  fi
+
+  echo "--- ninja install-clang"
+
+  ninja -C ${BUILD_DIR} install-clang install-clang-resource-headers
+
+  RUNTIMES_BUILD_DIR="${MONOREPO_ROOT}/build-runtimes"
+  INSTALL_DIR="${BUILD_DIR}/install"
+  mkdir -p ${RUNTIMES_BUILD_DIR}
+
+  echo "--- cmake runtimes C++03"
+
+  cmake -S "${MONOREPO_ROOT}/runtimes" -B "${RUNTIMES_BUILD_DIR}" -GNinja \
+      -D CMAKE_C_COMPILER="${INSTALL_DIR}/bin/clang" \
+      -D CMAKE_CXX_COMPILER="${INSTALL_DIR}/bin/clang++" \
+      -D LLVM_ENABLE_RUNTIMES="${runtimes}" \
+      -D LIBCXX_CXX_ABI=libcxxabi \
+      -D CMAKE_BUILD_TYPE=RelWithDebInfo \
+      -D CMAKE_INSTALL_PREFIX="${INSTALL_DIR}" \
+      -D LIBCXX_TEST_PARAMS="std=c++03" \
+      -D LIBCXXABI_TEST_PARAMS="std=c++03"
+
+  echo "--- ninja runtimes C++03"
+
+  ninja -vC "${RUNTIMES_BUILD_DIR}" ${runtime_targets}
+
+  echo "--- cmake runtimes C++26"
+
+  rm -rf "${RUNTIMES_BUILD_DIR}"
+  cmake -S "${MONOREPO_ROOT}/runtimes" -B "${RUNTIMES_BUILD_DIR}" -GNinja \
+      -D CMAKE_C_COMPILER="${INSTALL_DIR}/bin/clang" \
+      -D CMAKE_CXX_COMPILER="${INSTALL_DIR}/bin/clang++" \
+      -D LLVM_ENABLE_RUNTIMES="${runtimes}" \
+      -D LIBCXX_CXX_ABI=libcxxabi \
+      -D CMAKE_BUILD_TYPE=RelWithDebInfo \
+      -D CMAKE_INSTALL_PREFIX="${INSTALL_DIR}" \
+      -D LIBCXX_TEST_PARAMS="std=c++26" \
+      -D LIBCXXABI_TEST_PARAMS="std=c++26"
+
+  echo "--- ninja runtimes C++26"
+
+  ninja -vC "${RUNTIMES_BUILD_DIR}" ${runtime_targets}
+
+  echo "--- cmake runtimes clang modules"
+
+  rm -rf "${RUNTIMES_BUILD_DIR}"
+  cmake -S "${MONOREPO_ROOT}/runtimes" -B "${RUNTIMES_BUILD_DIR}" -GNinja \
+      -D CMAKE_C_COMPILER="${INSTALL_DIR}/bin/clang" \
+      -D CMAKE_CXX_COMPILER="${INSTALL_DIR}/bin/clang++" \
+      -D LLVM_ENABLE_RUNTIMES="${runtimes}" \
+      -D LIBCXX_CXX_ABI=libcxxabi \
+      -D CMAKE_BUILD_TYPE=RelWithDebInfo \
+      -D CMAKE_INSTALL_PREFIX="${INSTALL_DIR}" \
+      -D LIBCXX_TEST_PARAMS="enable_modules=clang" \
+      -D LIBCXXABI_TEST_PARAMS="enable_modules=clang"
+
+  echo "--- ninja runtimes clang modules"
+  
+  ninja -vC "${RUNTIMES_BUILD_DIR}" ${runtime_targets}
+fi

bolt/CMakeLists.txt

@@ -1,3 +1,5 @@
+set(LLVM_SUBPROJECT_TITLE "BOLT")
+
 include(ExternalProject)
 
 set(BOLT_SOURCE_DIR ${CMAKE_CURRENT_SOURCE_DIR})
@@ -121,7 +123,7 @@ option(BOLT_BUILD_TOOLS
   "Build the BOLT tools. If OFF, just generate build targets." ON)
 
 add_custom_target(bolt)
-set_target_properties(bolt PROPERTIES FOLDER "BOLT")
+set_target_properties(bolt PROPERTIES FOLDER "BOLT/Metatargets")
 add_llvm_install_targets(install-bolt DEPENDS bolt COMPONENT bolt)
 
 include_directories(

bolt/cmake/modules/AddBOLT.cmake

@@ -3,7 +3,6 @@ include(LLVMDistributionSupport)
 
 macro(add_bolt_executable name)
   add_llvm_executable(${name} ${ARGN})
-  set_target_properties(${name} PROPERTIES FOLDER "BOLT")
 endmacro()
 
 macro(add_bolt_tool name)

bolt/docs/BAT.md

@@ -106,9 +106,14 @@ equals output offset.
 `BRANCHENTRY` bit denotes whether a given offset pair is a control flow source
 (branch or call instruction). If not set, it signifies a control flow target
 (basic block offset).
+
 `InputAddr` is omitted for equal offsets in input and output function. In this
 case, `BRANCHENTRY` bits are encoded separately in a `BranchEntries` bitvector.
 
+Deleted basic blocks are emitted as having `OutputOffset` equal to the size of
+the function. They don't affect address translation and only participate in
+input basic block mapping.
+
 ### Secondary Entry Points table
 The table is emitted for hot fragments only. It contains `NumSecEntryPoints`
 offsets denoting secondary entry points, delta encoded, implicitly starting at zero.

bolt/docs/CMakeLists.txt

@@ -79,6 +79,7 @@ if (LLVM_ENABLE_DOXYGEN)
     COMMAND ${DOXYGEN_EXECUTABLE} ${CMAKE_CURRENT_BINARY_DIR}/doxygen.cfg
     WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}
     COMMENT "Generating bolt doxygen documentation." VERBATIM)
+  set_target_properties(doxygen-bolt PROPERTIES FOLDER "BOLT/Docs")
 
   if (LLVM_BUILD_DOCS)
     add_dependencies(doxygen doxygen-bolt)

bolt/include/bolt/Core/BinaryContext.h

@@ -17,6 +17,7 @@
 #include "bolt/Core/BinaryData.h"
 #include "bolt/Core/BinarySection.h"
 #include "bolt/Core/DebugData.h"
+#include "bolt/Core/DynoStats.h"
 #include "bolt/Core/JumpTable.h"
 #include "bolt/Core/MCPlusBuilder.h"
 #include "bolt/RuntimeLibs/RuntimeLibrary.h"
@@ -717,6 +718,9 @@ class BinaryContext {
     uint64_t NumStaleBlocksWithEqualIcount{0};
   } Stats;
 
+  // Original binary execution count stats.
+  DynoStats InitialDynoStats;
+
   // Address of the first allocated segment.
   uint64_t FirstAllocAddress{std::numeric_limits<uint64_t>::max()};
 
@@ -1220,8 +1224,7 @@ class BinaryContext {
 
   /// Return a signed value of \p Size stored at \p Address. The address has
   /// to be a valid statically allocated address for the binary.
-  ErrorOr<uint64_t> getSignedValueAtAddress(uint64_t Address,
-                                            size_t Size) const;
+  ErrorOr<int64_t> getSignedValueAtAddress(uint64_t Address, size_t Size) const;
 
   /// Special case of getUnsignedValueAtAddress() that uses a pointer size.
   ErrorOr<uint64_t> getPointerAtAddress(uint64_t Address) const {

bolt/include/bolt/Passes/BinaryPasses.h

@@ -53,15 +53,31 @@ class BinaryFunctionPass {
   virtual Error runOnFunctions(BinaryContext &BC) = 0;
 };
 
+/// A pass to set initial program-wide dynostats.
+class DynoStatsSetPass : public BinaryFunctionPass {
+public:
+  DynoStatsSetPass() : BinaryFunctionPass(false) {}
+
+  const char *getName() const override {
+    return "set dyno-stats before optimizations";
+  }
+
+  bool shouldPrint(const BinaryFunction &BF) const override { return false; }
+
+  Error runOnFunctions(BinaryContext &BC) override {
+    BC.InitialDynoStats = getDynoStats(BC.getBinaryFunctions(), BC.isAArch64());
+    return Error::success();
+  }
+};
+
 /// A pass to print program-wide dynostats.
 class DynoStatsPrintPass : public BinaryFunctionPass {
 protected:
-  DynoStats PrevDynoStats;
   std::string Title;
 
 public:
-  DynoStatsPrintPass(const DynoStats &PrevDynoStats, const char *Title)
-      : BinaryFunctionPass(false), PrevDynoStats(PrevDynoStats), Title(Title) {}
+  DynoStatsPrintPass(const char *Title)
+      : BinaryFunctionPass(false), Title(Title) {}
 
   const char *getName() const override {
     return "print dyno-stats after optimizations";
@@ -70,6 +86,7 @@ class DynoStatsPrintPass : public BinaryFunctionPass {
   bool shouldPrint(const BinaryFunction &BF) const override { return false; }
 
   Error runOnFunctions(BinaryContext &BC) override {
+    const DynoStats PrevDynoStats = BC.InitialDynoStats;
     const DynoStats NewDynoStats =
         getDynoStats(BC.getBinaryFunctions(), BC.isAArch64());
     const bool Changed = (NewDynoStats != PrevDynoStats);

bolt/include/bolt/Profile/BoltAddressTranslation.h

@@ -70,7 +70,7 @@ class BinaryFunction;
 class BoltAddressTranslation {
 public:
   // In-memory representation of the address translation table
-  using MapTy = std::map<uint32_t, uint32_t>;
+  using MapTy = std::multimap<uint32_t, uint32_t>;
 
   // List of taken fall-throughs
   using FallthroughListTy = SmallVector<std::pair<uint64_t, uint64_t>, 16>;
@@ -218,6 +218,7 @@ class BoltAddressTranslation {
     auto begin() const { return Map.begin(); }
     auto end() const { return Map.end(); }
     auto upper_bound(uint32_t Offset) const { return Map.upper_bound(Offset); }
+    auto size() const { return Map.size(); }
   };
 
   /// Map function output address to its hash and basic blocks hash map.

bolt/lib/Core/BinaryContext.cpp

@@ -142,7 +142,7 @@ BinaryContext::BinaryContext(std::unique_ptr<MCContext> Ctx,
       AsmInfo(std::move(AsmInfo)), MII(std::move(MII)), STI(std::move(STI)),
       InstPrinter(std::move(InstPrinter)), MIA(std::move(MIA)),
       MIB(std::move(MIB)), MRI(std::move(MRI)), DisAsm(std::move(DisAsm)),
-      Logger(Logger) {
+      Logger(Logger), InitialDynoStats(isAArch64()) {
   Relocation::Arch = this->TheTriple->getArch();
   RegularPageSize = isAArch64() ? RegularPageSizeAArch64 : RegularPageSizeX86;
   PageAlign = opts::NoHugePages ? RegularPageSize : HugePageSize;
@@ -2217,8 +2217,8 @@ ErrorOr<uint64_t> BinaryContext::getUnsignedValueAtAddress(uint64_t Address,
   return DE.getUnsigned(&ValueOffset, Size);
 }
 
-ErrorOr<uint64_t> BinaryContext::getSignedValueAtAddress(uint64_t Address,
-                                                         size_t Size) const {
+ErrorOr<int64_t> BinaryContext::getSignedValueAtAddress(uint64_t Address,
+                                                        size_t Size) const {
   const ErrorOr<const BinarySection &> Section = getSectionForAddress(Address);
   if (!Section)
     return std::make_error_code(std::errc::bad_address);

bolt/lib/Core/BinaryFunction.cpp

@@ -851,15 +851,19 @@ BinaryFunction::processIndirectBranch(MCInst &Instruction, unsigned Size,
     return IndirectBranchType::UNKNOWN;
   }
 
-  // RIP-relative addressing should be converted to symbol form by now
-  // in processed instructions (but not in jump).
-  if (DispExpr) {
+  auto getExprValue = [&](const MCExpr *Expr) {
     const MCSymbol *TargetSym;
     uint64_t TargetOffset;
-    std::tie(TargetSym, TargetOffset) = BC.MIB->getTargetSymbolInfo(DispExpr);
+    std::tie(TargetSym, TargetOffset) = BC.MIB->getTargetSymbolInfo(Expr);
     ErrorOr<uint64_t> SymValueOrError = BC.getSymbolValue(*TargetSym);
-    assert(SymValueOrError && "global symbol needs a value");
-    ArrayStart = *SymValueOrError + TargetOffset;
+    assert(SymValueOrError && "Global symbol needs a value");
+    return *SymValueOrError + TargetOffset;
+  };
+
+  // RIP-relative addressing should be converted to symbol form by now
+  // in processed instructions (but not in jump).
+  if (DispExpr) {
+    ArrayStart = getExprValue(DispExpr);
     BaseRegNum = BC.MIB->getNoRegister();
     if (BC.isAArch64()) {
       ArrayStart &= ~0xFFFULL;
@@ -3698,6 +3702,13 @@ BinaryFunction::BasicBlockListType BinaryFunction::dfs() const {
 
 size_t BinaryFunction::computeHash(bool UseDFS, HashFunction HashFunction,
                                    OperandHashFuncTy OperandHashFunc) const {
+  LLVM_DEBUG({
+    dbgs() << "BOLT-DEBUG: computeHash " << getPrintName() << ' '
+           << (UseDFS ? "dfs" : "bin") << " order "
+           << (HashFunction == HashFunction::StdHash ? "std::hash" : "xxh3")
+           << '\n';
+  });
+
   if (size() == 0)
     return 0;
 

bolt/lib/Core/DebugNames.cpp

@@ -112,8 +112,6 @@ void DWARF5AcceleratorTable::addUnit(DWARFUnit &Unit,
 // Returns true if DW_TAG_variable should be included in .debug-names based on
 // section 6.1.1.1 for DWARF5 spec.
 static bool shouldIncludeVariable(const DWARFUnit &Unit, const DIE &Die) {
-  if (Die.findAttribute(dwarf::Attribute::DW_AT_declaration))
-    return false;
   const DIEValue LocAttrInfo =
       Die.findAttribute(dwarf::Attribute::DW_AT_location);
   if (!LocAttrInfo)
@@ -148,6 +146,8 @@ static bool shouldIncludeVariable(const DWARFUnit &Unit, const DIE &Die) {
 
 bool static canProcess(const DWARFUnit &Unit, const DIE &Die,
                        std::string &NameToUse, const bool TagsOnly) {
+  if (Die.findAttribute(dwarf::Attribute::DW_AT_declaration))
+    return false;
   switch (Die.getTag()) {
   case dwarf::DW_TAG_base_type:
   case dwarf::DW_TAG_class_type: