[Reland] increase clang-tidy coverage in torch/csrc (pytorch#108875)

Reland  PR pytorch#103058 since there was a time gap between this PR and other PRs in terms of torch/csrc modifications

Pull Request resolved: pytorch#108875
Approved by: https://github.com/Skylion007

.clang-tidy

@@ -44,7 +44,7 @@ modernize-*,
 performance-*,
 readability-container-size-empty,
 '
-HeaderFilterRegex: '^(c10/(?!test)|torch/csrc/(?!deploy/interpreter/cpython)).*$'
+HeaderFilterRegex: '^(c10/(?!test)|torch/csrc/).*$'
 AnalyzeTemporaryDtors: false
 WarningsAsErrors: '*'
 ...

.github/workflows/lint.yml

@@ -17,6 +17,7 @@ jobs:
   lintrunner:
     uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
     with:
+      timeout: 120
       runner: linux.2xlarge
       docker-image: pytorch-linux-focal-linter
       fetch-depth: 0

.lintrunner.toml

@@ -277,33 +277,33 @@ command = [
 code = 'CLANGTIDY'
 include_patterns = [
     'c10/**/*.cpp',
-    'torch/csrc/fx/**/*.cpp',
-    'torch/csrc/generic/**/*.cpp',
-    'torch/csrc/onnx/**/*.cpp',
-    'torch/csrc/tensor/**/*.cpp',
-    'torch/csrc/utils/**/*.cpp',
+    'torch/csrc/**/*.cpp',
 ]
 exclude_patterns = [
     # The negative filters below are to exclude files that include onnx_pb.h or
     # caffe2_pb.h, otherwise we'd have to build protos as part of this CI job.
-    # FunctionsManual.cpp is excluded to keep this diff clean. It will be fixed
-    # in a follow up PR.
-    # that are not easily converted to accepted c++
-    'c10/cuda/**/*.cpp',
-    'c10/test/**/*.cpp',
+    # CUDA files are also excluded.
     '**/fb/**',
-    'torch/csrc/jit/passes/onnx/helper.cpp',
-    'torch/csrc/jit/passes/onnx/shape_type_inference.cpp',
-    'torch/csrc/jit/serialization/onnx.cpp',
-    'torch/csrc/jit/serialization/export.cpp',
-    'torch/csrc/jit/serialization/import.cpp',
+    '**/*pb.h',
+    '**/*CUDA*',
+    '**/cuda/*pp',
+    'c10/test/**',
+    'third_party/**/*',
+    'torch/csrc/api/**',
+    'torch/csrc/autograd/**',
+    'torch/csrc/CudaIPCTypes.cpp',
+    'torch/csrc/cuda/**',
+    'torch/csrc/dynamo/*',
+    'torch/csrc/distributed/**/*',
+    'torch/csrc/inductor/**/*',
+    'torch/csrc/jit/**/*',
     'torch/csrc/jit/serialization/import_legacy.cpp',
+    'torch/csrc/jit/serialization/export.cpp',
+    'torch/csrc/lazy/**/*',
     'torch/csrc/onnx/init.cpp',
-    'torch/csrc/cuda/nccl.*',
-    'torch/csrc/cuda/python_nccl.cpp',
-    'torch/csrc/autograd/FunctionsManual.cpp',
-    'torch/csrc/jit/codegen/cuda/runtime/*',
-    'torch/csrc/utils/disable_torch_function.cpp',
+    'torch/csrc/profiler/**/*',
+    'torch/csrc/quantized/**/*',
+    'torch/csrc/mps/**/*',
 ]
 init_command = [
     'python3',

c10/core/SafePyObject.h

@@ -22,7 +22,7 @@ struct C10_API SafePyObject {
   // Steals a reference to data
   SafePyObject(PyObject* data, c10::impl::PyInterpreter* pyinterpreter)
       : data_(data), pyinterpreter_(pyinterpreter) {}
-  SafePyObject(SafePyObject&& other)
+  SafePyObject(SafePyObject&& other) noexcept
       : data_(std::exchange(other.data_, nullptr)),
         pyinterpreter_(other.pyinterpreter_) {}
 

c10/core/TensorImpl.h

@@ -75,7 +75,7 @@ inline int64_t size_from_dim_(int k, IntArrayRef dims) {
 
 // Product of all dims up to k (not including dims[k])
 inline int64_t size_to_dim_(int k, IntArrayRef dims) {
-  TORCH_CHECK((unsigned)k <= dims.size());
+  TORCH_CHECK(k >= 0 && static_cast<size_t>(k) <= dims.size());
   int64_t r = 1;
   for (const auto i : c10::irange(k)) {
     r *= dims[i];

c10/util/OptionalArrayRef.h

@@ -27,7 +27,7 @@ class OptionalArrayRef final {
 
   OptionalArrayRef(const OptionalArrayRef& other) = default;
 
-  OptionalArrayRef(OptionalArrayRef&& other) = default;
+  OptionalArrayRef(OptionalArrayRef&& other) noexcept = default;
 
   constexpr OptionalArrayRef(const optional<ArrayRef<T>>& other) noexcept
       : wrapped_opt_array_ref(other) {}
@@ -90,7 +90,7 @@ class OptionalArrayRef final {
 
   OptionalArrayRef& operator=(const OptionalArrayRef& other) = default;
 
-  OptionalArrayRef& operator=(OptionalArrayRef&& other) = default;
+  OptionalArrayRef& operator=(OptionalArrayRef&& other) noexcept = default;
 
   constexpr OptionalArrayRef& operator=(
       const optional<ArrayRef<T>>& other) noexcept {

c10/util/either.h

@@ -78,7 +78,7 @@ class either final {
     return *this;
   }
 
-  either<Left, Right>& operator=(either<Left, Right>&& rhs) {
+  either<Left, Right>& operator=(either<Left, Right>&& rhs) noexcept {
     _destruct();
     _side = rhs._side;
     if (_side == Side::left) {

torch/csrc/CudaIPCTypes.cpp

@@ -139,8 +139,8 @@ void ReturnRefCounter(const std::string& handle, uint64_t offset /* unused */) {
 
 CudaIPCSentData::CudaIPCSentData(
     std::string handle,
-    int64_t offset,
-    int64_t* counter_ptr,
+    uint64_t offset,
+    uint64_t* counter_ptr,
     at::Device device)
     : handle_(std::move(handle)),
       offset_(offset),
@@ -206,7 +206,7 @@ CudaIPCSentData::~CudaIPCSentData() {
 #endif
 }
 
-int64_t CudaIPCSentData::counter_value() {
+uint64_t CudaIPCSentData::counter_value() {
   return *counter_ptr_;
 }
 

torch/csrc/CudaIPCTypes.h

@@ -22,25 +22,25 @@ struct CudaIPCReceivedData final {
 
 struct CudaIPCSentData final {
   std::string handle_;
-  int64_t offset_;
-  int64_t* counter_ptr_; // Reference counter shared memory block
+  uint64_t offset_;
+  uint64_t* counter_ptr_; // Reference counter shared memory block
   at::DataPtr original_ptr_; // Original mem allocation
   cudaEvent_t event_; // Sync cuEventDestroy
   bool event_sync_required_;
   at::Device device_;
 
   CudaIPCSentData(
       std::string handle,
-      int64_t offset,
-      int64_t* counter_ptr,
+      uint64_t offset,
+      uint64_t* counter_ptr,
       at::Device device);
   ~CudaIPCSentData();
 
-  int64_t counter_value();
+  uint64_t counter_value();
   std::string handle() {
     return handle_;
   }
-  int64_t offset() {
+  uint64_t offset() {
     return offset_;
   }
   void set_original_ptr(at::DataPtr data_ptr) {
@@ -87,8 +87,8 @@ struct CudaIPCRefCountersFile final {
         handle_(std::move(handle)),
         refcounted_shared_mem_(std::move(data_ptr)) {}
 
-  int64_t* counter_ptr() {
-    return static_cast<int64_t*>(refcounted_shared_mem_.get()) + next_offset_;
+  uint64_t* counter_ptr() {
+    return static_cast<uint64_t*>(refcounted_shared_mem_.get()) + next_offset_;
   }
 
   void set_counter(uint64_t value) {
@@ -103,7 +103,7 @@ struct CudaIPCRefCountersFile final {
     return used_slots_;
   }
 
-  int64_t get_offset() {
+  uint64_t get_offset() {
     return next_offset_;
   }
 

torch/csrc/Storage.cpp

@@ -134,8 +134,8 @@ static PyObject* THPStorage_pynew(
   torch::ParsedArgs<3> parsed_args;
   auto r = parser.parse(args, kwargs, parsed_args);
 
-  int64_t allocator_arg_idx = 0;
-  int64_t device_arg_idx = 1;
+  int allocator_arg_idx = 0;
+  int device_arg_idx = 1;
 
   if (r.idx > 0) {
     allocator_arg_idx = 1;

torch/csrc/jit/passes/onnx/helper.cpp

@@ -186,7 +186,7 @@ Node* createONNXConstant(
     at::Tensor value) {
   Node* constant_node = graph->create(onnx::Constant, 1);
   constant_node->insertBefore(n_to_insert_before);
-  constant_node->t_(attr::value, value);
+  constant_node->t_(attr::value, std::move(value));
   return constant_node;
 }