Merge branch 'main' into do-not-rely-on-offsets-in-lto-tests

Author: kovdan01

.ci/compute_projects.py

@@ -57,6 +57,13 @@
     ".ci": {"llvm", "clang", "lld", "lldb"},
 }
 
+# This mapping describes runtimes that should be enabled for a specific project,
+# but not necessarily run for testing. The only case of this currently is lldb
+# which needs some runtimes enabled for tests.
+DEPENDENT_RUNTIMES_TO_BUILD = {"lldb": {"libcxx", "libcxxabi", "libunwind"}}
+
+# This mapping describes runtimes that should be tested when the key project is
+# touched.
 DEPENDENT_RUNTIMES_TO_TEST = {"clang": {"libcxx", "libcxxabi", "libunwind"}}
 
 EXCLUDE_LINUX = {
@@ -180,16 +187,20 @@ def _compute_project_check_targets(projects_to_test: Set[str]) -> Set[str]:
 def _compute_runtimes_to_test(projects_to_test: Set[str]) -> Set[str]:
     runtimes_to_test = set()
     for project_to_test in projects_to_test:
-        if project_to_test not in DEPENDENT_RUNTIMES_TO_TEST:
-            continue
-        runtimes_to_test.update(DEPENDENT_RUNTIMES_TO_TEST[project_to_test])
+        if project_to_test in DEPENDENT_RUNTIMES_TO_TEST:
+            runtimes_to_test.update(DEPENDENT_RUNTIMES_TO_TEST[project_to_test])
+        if project_to_test in DEPENDENT_RUNTIMES_TO_BUILD:
+            runtimes_to_test.update(DEPENDENT_RUNTIMES_TO_BUILD[project_to_test])
     return runtimes_to_test
 
 
-def _compute_runtime_check_targets(runtimes_to_test: Set[str]) -> Set[str]:
+def _compute_runtime_check_targets(projects_to_test: Set[str]) -> Set[str]:
     check_targets = set()
-    for runtime_to_test in runtimes_to_test:
-        check_targets.add(PROJECT_CHECK_TARGETS[runtime_to_test])
+    for project_to_test in projects_to_test:
+        if project_to_test not in DEPENDENT_RUNTIMES_TO_TEST:
+            continue
+        for runtime_to_test in DEPENDENT_RUNTIMES_TO_TEST[project_to_test]:
+            check_targets.add(PROJECT_CHECK_TARGETS[runtime_to_test])
     return check_targets
 
 
@@ -216,16 +227,16 @@ def get_env_variables(modified_files: list[str], platform: str) -> Set[str]:
     projects_to_test = _compute_projects_to_test(modified_projects, platform)
     projects_to_build = _compute_projects_to_build(projects_to_test)
     projects_check_targets = _compute_project_check_targets(projects_to_test)
-    runtimes_to_test = _compute_runtimes_to_test(projects_to_test)
-    runtimes_check_targets = _compute_runtime_check_targets(runtimes_to_test)
+    runtimes_to_build = _compute_runtimes_to_test(projects_to_test)
+    runtimes_check_targets = _compute_runtime_check_targets(projects_to_test)
     # We use a semicolon to separate the projects/runtimes as they get passed
     # to the CMake invocation and thus we need to use the CMake list separator
     # (;). We use spaces to separate the check targets as they end up getting
     # passed to ninja.
     return {
         "projects_to_build": ";".join(sorted(projects_to_build)),
         "project_check_targets": " ".join(sorted(projects_check_targets)),
-        "runtimes_to_build": ";".join(sorted(runtimes_to_test)),
+        "runtimes_to_build": ";".join(sorted(runtimes_to_build)),
         "runtimes_check_targets": " ".join(sorted(runtimes_check_targets)),
     }
 

.ci/compute_projects_test.py

@@ -205,6 +205,17 @@ def test_ci(self):
             "check-cxx check-cxxabi check-unwind",
         )
 
+    def test_lldb(self):
+        env_variables = compute_projects.get_env_variables(
+            ["lldb/CMakeLists.txt"], "Linux"
+        )
+        self.assertEqual(env_variables["projects_to_build"], "clang;lldb;llvm")
+        self.assertEqual(env_variables["project_check_targets"], "check-lldb")
+        self.assertEqual(
+            env_variables["runtimes_to_build"], "libcxx;libcxxabi;libunwind"
+        )
+        self.assertEqual(env_variables["runtimes_check_targets"], "")
+
 
 if __name__ == "__main__":
     unittest.main()

.ci/monolithic-linux.sh

@@ -55,6 +55,8 @@ trap at-exit EXIT
 
 projects="${1}"
 targets="${2}"
+runtimes="${3}"
+runtime_targets="${4}"
 
 lit_args="-v --xunit-xml-output ${BUILD_DIR}/test-results.xml --use-unique-output-file-name --timeout=1200 --time-tests"
 
@@ -70,7 +72,7 @@ export LLVM_SYMBOLIZER_PATH=`which llvm-symbolizer`
 # It will not be built unless it is used.
 cmake -S "${MONOREPO_ROOT}"/llvm -B "${BUILD_DIR}" \
       -D LLVM_ENABLE_PROJECTS="${projects}" \
-      -D LLVM_ENABLE_RUNTIMES="libcxx;libcxxabi;libunwind" \
+      -D LLVM_ENABLE_RUNTIMES="${runtimes}" \
       -G Ninja \
       -D CMAKE_PREFIX_PATH="${HOME}/.local" \
       -D CMAKE_BUILD_TYPE=Release \
@@ -91,62 +93,28 @@ echo "--- ninja"
 # Targets are not escaped as they are passed as separate arguments.
 ninja -C "${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 ${RUNTIMES_BUILD_DIR}
-
+if [[ "${runtimes_targets}" != "" ]]; then
   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="${INSTALL_DIR}" \
-      -D LIBCXX_TEST_PARAMS="std=c++26" \
-      -D LIBCXXABI_TEST_PARAMS="std=c++26" \
-      -D LLVM_LIT_ARGS="${lit_args}"
+  cmake \
+    -D LIBCXX_TEST_PARAMS="std=c++26" \
+    -D LIBCXXABI_TEST_PARAMS="std=c++26" \
+    "${BUILD_DIR}"
 
   echo "--- ninja runtimes C++26"
 
-  ninja -vC "${RUNTIMES_BUILD_DIR}" ${runtime_targets}
+  ninja -C "${BUILD_DIR}" ${runtime_targets}
 
   echo "--- cmake runtimes clang modules"
 
-  # We don't need to do a clean build of runtimes, because LIBCXX_TEST_PARAMS
-  # and LIBCXXABI_TEST_PARAMS only affect lit configuration, which successfully
-  # propagates without a clean build. Other that those two variables, builds
-  # are supposed to be the same.
-
-  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" \
-      -D LLVM_LIT_ARGS="${lit_args}"
+  cmake \
+    -D LIBCXX_TEST_PARAMS="enable_modules=clang" \
+    -D LIBCXXABI_TEST_PARAMS="enable_modules=clang" \
+    "${BUILD_DIR}"
 
   echo "--- ninja runtimes clang modules"
 
-  ninja -vC "${RUNTIMES_BUILD_DIR}" ${runtime_targets}
+  ninja -C "${BUILD_DIR}" ${runtime_targets}
 fi

.git-blame-ignore-revs

@@ -100,3 +100,9 @@ d33bf2e9df578ff7e44fd22504d6ad5a122b7ee6
 
 # [lldb][NFC] clang-format MainLoopPosix.cpp
 66bdbfbaa08fa3d8e64a7fe136a8fb717f5cdbb7
+
+# [clang-tidy][NFC] Run clang-format on "clang-tools-extra/clang-tidy"
+65d66625b3e2b8322ed99d82edabecbafcd0885b
+ce46adb8b7ce645353eccaedf31ed9765dab77bb
+68070f908bb7ac5f0b5fa9722caa504ecf723f6b
+5213c57cb1f0d78aad9a253b7f6a2b62ff4c7859

.github/workflows/email-check.yaml

@@ -32,7 +32,8 @@ jobs:
           COMMENT: >-
             ⚠️ We detected that you are using a GitHub private e-mail address to contribute to the repo.<br/>
             Please turn off [Keep my email addresses private](https://github.com/settings/emails) setting in your account.<br/>
-            See [LLVM Discourse](https://discourse.llvm.org/t/hidden-emails-on-github-should-we-do-something-about-it) for more information.
+            See [LLVM Developer Policy](https://llvm.org/docs/DeveloperPolicy.html#email-addresses) and
+            [LLVM Discourse](https://discourse.llvm.org/t/hidden-emails-on-github-should-we-do-something-about-it) for more information.
         run: |
           cat << EOF > comments
           [{"body" : "$COMMENT"}]

.github/workflows/libcxx-build-and-test.yaml

@@ -52,8 +52,8 @@ jobs:
         cxx: [ 'clang++-21' ]
         include:
           - config: 'generic-gcc'
-            cc: 'gcc-14'
-            cxx: 'g++-14'
+            cc: 'gcc-15'
+            cxx: 'g++-15'
     steps:
       - uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
       - name: ${{ matrix.config }}.${{ matrix.cxx }}
@@ -92,8 +92,8 @@ jobs:
         cxx: [ 'clang++-21' ]
         include:
           - config: 'generic-gcc-cxx11'
-            cc: 'gcc-14'
-            cxx: 'g++-14'
+            cc: 'gcc-15'
+            cxx: 'g++-15'
           - config: 'generic-cxx26'
             cc: 'clang-20'
             cxx: 'clang++-20'

bolt/include/bolt/Core/BinaryBasicBlock.h

@@ -52,9 +52,8 @@ class BinaryBasicBlock {
     uint64_t MispredictedCount; /// number of branches mispredicted
 
     bool operator<(const BinaryBranchInfo &Other) const {
-      return (Count < Other.Count) ||
-             (Count == Other.Count &&
-              MispredictedCount < Other.MispredictedCount);
+      return std::tie(Count, MispredictedCount) <
+             std::tie(Other.Count, Other.MispredictedCount);
     }
   };
 

bolt/include/bolt/Core/BinaryFunction.h

@@ -388,6 +388,10 @@ class BinaryFunction {
   /// The profile data for the number of times the function was executed.
   uint64_t ExecutionCount{COUNT_NO_PROFILE};
 
+  /// Profile data for the number of times this function was entered from
+  /// external code (DSO, JIT, etc).
+  uint64_t ExternEntryCount{0};
+
   /// Profile match ratio.
   float ProfileMatchRatio{0.0f};
 
@@ -1877,6 +1881,10 @@ class BinaryFunction {
     return *this;
   }
 
+  /// Set the profile data for the number of times the function was entered from
+  /// external code (DSO/JIT).
+  void setExternEntryCount(uint64_t Count) { ExternEntryCount = Count; }
+
   /// Adjust execution count for the function by a given \p Count. The value
   /// \p Count will be subtracted from the current function count.
   ///
@@ -1904,6 +1912,10 @@ class BinaryFunction {
   /// Return COUNT_NO_PROFILE if there's no profile info.
   uint64_t getExecutionCount() const { return ExecutionCount; }
 
+  /// Return the profile information about the number of times the function was
+  /// entered from external code (DSO/JIT).
+  uint64_t getExternEntryCount() const { return ExternEntryCount; }
+
   /// Return the raw profile information about the number of branch
   /// executions corresponding to this function.
   uint64_t getRawSampleCount() const { return RawSampleCount; }

bolt/include/bolt/Passes/PAuthGadgetScanner.h

@@ -77,9 +77,7 @@ struct MCInstInBFReference {
     return BF == RHS.BF && Offset == RHS.Offset;
   }
   bool operator<(const MCInstInBFReference &RHS) const {
-    if (BF != RHS.BF)
-      return BF < RHS.BF;
-    return Offset < RHS.Offset;
+    return std::tie(BF, Offset) < std::tie(RHS.BF, RHS.Offset);
   }
   operator MCInst &() const {
     assert(BF != nullptr);

bolt/include/bolt/Profile/DataAggregator.h

@@ -78,6 +78,13 @@ class DataAggregator : public DataReader {
   static bool checkPerfDataMagic(StringRef FileName);
 
 private:
+  struct LBREntry {
+    uint64_t From;
+    uint64_t To;
+    bool Mispred;
+  };
+  friend raw_ostream &operator<<(raw_ostream &OS, const LBREntry &);
+
   struct PerfBranchSample {
     SmallVector<LBREntry, 32> LBR;
   };
@@ -92,24 +99,28 @@ class DataAggregator : public DataReader {
     uint64_t Addr;
   };
 
+  /// Container for the unit of branch data.
+  /// Backwards compatible with legacy use for branches and fall-throughs:
+  /// - if \p Branch is FT_ONLY or FT_EXTERNAL_ORIGIN, the trace only
+  ///   contains fall-through data,
+  /// - if \p To is BR_ONLY, the trace only contains branch data.
   struct Trace {
+    static constexpr const uint64_t EXTERNAL = 0ULL;
+    static constexpr const uint64_t BR_ONLY = -1ULL;
+    static constexpr const uint64_t FT_ONLY = -1ULL;
+    static constexpr const uint64_t FT_EXTERNAL_ORIGIN = -2ULL;
+
+    uint64_t Branch;
     uint64_t From;
     uint64_t To;
-    Trace(uint64_t From, uint64_t To) : From(From), To(To) {}
-    bool operator==(const Trace &Other) const {
-      return From == Other.From && To == Other.To;
-    }
+    auto tie() const { return std::tie(Branch, From, To); }
+    bool operator==(const Trace &Other) const { return tie() == Other.tie(); }
+    bool operator<(const Trace &Other) const { return tie() < Other.tie(); }
   };
+  friend raw_ostream &operator<<(raw_ostream &OS, const Trace &);
 
   struct TraceHash {
-    size_t operator()(const Trace &L) const {
-      return std::hash<uint64_t>()(L.From << 32 | L.To);
-    }
-  };
-
-  struct FTInfo {
-    uint64_t InternCount{0};
-    uint64_t ExternCount{0};
+    size_t operator()(const Trace &L) const { return hash_combine(L.tie()); }
   };
 
   struct TakenBranchInfo {
@@ -119,8 +130,8 @@ class DataAggregator : public DataReader {
 
   /// Intermediate storage for profile data. We save the results of parsing
   /// and use them later for processing and assigning profile.
-  std::unordered_map<Trace, TakenBranchInfo, TraceHash> BranchLBRs;
-  std::unordered_map<Trace, FTInfo, TraceHash> FallthroughLBRs;
+  std::unordered_map<Trace, TakenBranchInfo, TraceHash> TraceMap;
+  std::vector<std::pair<Trace, TakenBranchInfo>> Traces;
   std::unordered_map<uint64_t, uint64_t> BasicSamples;
   std::vector<PerfMemSample> MemSamples;
 
@@ -193,8 +204,8 @@ class DataAggregator : public DataReader {
   /// Return a vector of offsets corresponding to a trace in a function
   /// if the trace is valid, std::nullopt otherwise.
   std::optional<SmallVector<std::pair<uint64_t, uint64_t>, 16>>
-  getFallthroughsInTrace(BinaryFunction &BF, const LBREntry &First,
-                         const LBREntry &Second, uint64_t Count = 1) const;
+  getFallthroughsInTrace(BinaryFunction &BF, const Trace &Trace,
+                         uint64_t Count) const;
 
   /// Record external entry into the function \p BF.
   ///
@@ -258,8 +269,7 @@ class DataAggregator : public DataReader {
   bool doBranch(uint64_t From, uint64_t To, uint64_t Count, uint64_t Mispreds);
 
   /// Register a trace between two LBR entries supplied in execution order.
-  bool doTrace(const LBREntry &First, const LBREntry &Second,
-               uint64_t Count = 1);
+  bool doTrace(const Trace &Trace, uint64_t Count);
 
   /// Parser helpers
   /// Return false if we exhausted our parser buffer and finished parsing
@@ -476,7 +486,6 @@ class DataAggregator : public DataReader {
 
   /// Debugging dump methods
   void dump() const;
-  void dump(const LBREntry &LBR) const;
   void dump(const PerfBranchSample &Sample) const;
   void dump(const PerfMemSample &Sample) const;
 
@@ -504,6 +513,27 @@ class DataAggregator : public DataReader {
 
   friend class YAMLProfileWriter;
 };
+
+inline raw_ostream &operator<<(raw_ostream &OS,
+                               const DataAggregator::LBREntry &L) {
+  OS << formatv("{0:x} -> {1:x}/{2}", L.From, L.To, L.Mispred ? 'M' : 'P');
+  return OS;
+}
+
+inline raw_ostream &operator<<(raw_ostream &OS,
+                               const DataAggregator::Trace &T) {
+  switch (T.Branch) {
+  case DataAggregator::Trace::FT_ONLY:
+  case DataAggregator::Trace::FT_EXTERNAL_ORIGIN:
+    break;
+  default:
+    OS << Twine::utohexstr(T.Branch) << " -> ";
+  }
+  OS << Twine::utohexstr(T.From);
+  if (T.To != DataAggregator::Trace::BR_ONLY)
+    OS << " ... " << Twine::utohexstr(T.To);
+  return OS;
+}
 } // namespace bolt
 } // namespace llvm