[clang][dataflow] Add reverse null checker

clang-tools-extra/clang-tidy/bugprone/BugproneTidyModule.cpp

@@ -48,6 +48,7 @@
 #include "NoEscapeCheck.h"
 #include "NonZeroEnumToBoolConversionCheck.h"
 #include "NotNullTerminatedResultCheck.h"
+#include "NullCheckAfterDereferenceCheck.h"
 #include "OptionalValueConversionCheck.h"
 #include "ParentVirtualCallCheck.h"
 #include "PosixReturnCheck.h"
@@ -180,6 +181,8 @@ class BugproneModule : public ClangTidyModule {
     CheckFactories.registerCheck<PosixReturnCheck>("bugprone-posix-return");
     CheckFactories.registerCheck<ReservedIdentifierCheck>(
         "bugprone-reserved-identifier");
+    CheckFactories.registerCheck<NullCheckAfterDereferenceCheck>(
+        "bugprone-null-check-after-dereference");
     CheckFactories.registerCheck<SharedPtrArrayMismatchCheck>(
         "bugprone-shared-ptr-array-mismatch");
     CheckFactories.registerCheck<SignalHandlerCheck>("bugprone-signal-handler");

clang-tools-extra/clang-tidy/bugprone/CMakeLists.txt

@@ -44,6 +44,7 @@ add_clang_library(clangTidyBugproneModule
   NoEscapeCheck.cpp
   NonZeroEnumToBoolConversionCheck.cpp
   NotNullTerminatedResultCheck.cpp
+  NullCheckAfterDereferenceCheck.cpp
   OptionalValueConversionCheck.cpp
   ParentVirtualCallCheck.cpp
   PosixReturnCheck.cpp

clang-tools-extra/clang-tidy/bugprone/NullCheckAfterDereferenceCheck.cpp

@@ -0,0 +1,175 @@
+//===--- NullCheckAfterDereferenceCheck.cpp - clang-tidy-------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "NullCheckAfterDereferenceCheck.h"
+#include "clang/AST/ASTContext.h"
+#include "clang/AST/DeclCXX.h"
+#include "clang/AST/DeclTemplate.h"
+#include "clang/ASTMatchers/ASTMatchFinder.h"
+#include "clang/ASTMatchers/ASTMatchers.h"
+#include "clang/Analysis/CFG.h"
+#include "clang/Analysis/FlowSensitive/ControlFlowContext.h"
+#include "clang/Analysis/FlowSensitive/DataflowAnalysisContext.h"
+#include "clang/Analysis/FlowSensitive/DataflowEnvironment.h"
+#include "clang/Analysis/FlowSensitive/DataflowLattice.h"
+#include "clang/Analysis/FlowSensitive/Models/NullPointerAnalysisModel.h"
+#include "clang/Analysis/FlowSensitive/WatchedLiteralsSolver.h"
+#include "clang/Basic/SourceLocation.h"
+#include "llvm/ADT/Any.h"
+#include "llvm/ADT/STLExtras.h"
+#include "llvm/Support/Error.h"
+#include <memory>
+#include <vector>
+
+namespace clang::tidy::bugprone {
+
+using ast_matchers::MatchFinder;
+using dataflow::NullPointerAnalysisModel;
+using dataflow::NullCheckAfterDereferenceDiagnoser;
+
+static constexpr llvm::StringLiteral FuncID("fun");
+
+struct ExpandedResult {
+  SourceLocation WarningLoc;
+  std::optional<SourceLocation> DerefLoc;
+};
+
+using ExpandedResultType = std::pair<std::vector<ExpandedResult>,
+                                     std::vector<ExpandedResult>>;
+
+static std::optional<ExpandedResultType> analyzeFunction(
+        const FunctionDecl &FuncDecl) {
+  using dataflow::ControlFlowContext;
+  using dataflow::DataflowAnalysisState;
+  using llvm::Expected;
+
+  ASTContext &ASTCtx = FuncDecl.getASTContext();
+
+  if (FuncDecl.getBody() == nullptr) {
+    return std::nullopt;
+  }
+
+  Expected<ControlFlowContext> Context =
+      ControlFlowContext::build(FuncDecl, *FuncDecl.getBody(), ASTCtx);
+  if (!Context)
+    return std::nullopt;
+
+  dataflow::DataflowAnalysisContext AnalysisContext(
+      std::make_unique<dataflow::WatchedLiteralsSolver>());
+  dataflow::Environment Env(AnalysisContext, FuncDecl);
+  NullPointerAnalysisModel Analysis(ASTCtx);
+  NullCheckAfterDereferenceDiagnoser Diagnoser;
+  NullCheckAfterDereferenceDiagnoser::ResultType Diagnostics;
+
+  using LatticeState = DataflowAnalysisState<NullPointerAnalysisModel::Lattice>;
+  using DetailMaybeLatticeStates = std::vector<std::optional<LatticeState>>;
+
+  auto DiagnoserImpl = [&ASTCtx, &Diagnoser, &Diagnostics](
+              const CFGElement &Elt,
+              const LatticeState &S) mutable -> void {
+            auto EltDiagnostics = Diagnoser.diagnose(ASTCtx, &Elt, S.Env);
+            llvm::move(EltDiagnostics.first,
+                       std::back_inserter(Diagnostics.first));
+            llvm::move(EltDiagnostics.second,
+                       std::back_inserter(Diagnostics.second));
+          };
+
+  Expected<DetailMaybeLatticeStates>
+      BlockToOutputState = dataflow::runDataflowAnalysis(
+          *Context, Analysis, Env, DiagnoserImpl);
+
+  if (llvm::Error E = BlockToOutputState.takeError()) {
+    llvm::dbgs() << "Dataflow analysis failed: " << llvm::toString(std::move(E))
+                 << ".\n";
+    return std::nullopt;
+  }
+
+  ExpandedResultType ExpandedDiagnostics;
+
+  llvm::transform(Diagnostics.first,
+                  std::back_inserter(ExpandedDiagnostics.first),
+                  [&](SourceLocation WarningLoc) -> ExpandedResult {
+                    if (auto Val = Diagnoser.WarningLocToVal[WarningLoc];
+                        auto DerefExpr = Diagnoser.ValToDerefLoc[Val]) {
+                      return {WarningLoc, DerefExpr->getBeginLoc()};
+                    }
+
+                    return {WarningLoc, std::nullopt};
+                  });
+
+  llvm::transform(Diagnostics.second,
+                  std::back_inserter(ExpandedDiagnostics.second),
+                  [&](SourceLocation WarningLoc) -> ExpandedResult {
+                    if (auto Val = Diagnoser.WarningLocToVal[WarningLoc];
+                        auto DerefExpr = Diagnoser.ValToDerefLoc[Val]) {
+                      return {WarningLoc, DerefExpr->getBeginLoc()};
+                    }
+
+                    return {WarningLoc, std::nullopt};
+                  });
+
+  return ExpandedDiagnostics;
+}
+
+void NullCheckAfterDereferenceCheck::registerMatchers(MatchFinder *Finder) {
+  using namespace ast_matchers;
+
+  auto hasPointerValue =
+      hasDescendant(NullPointerAnalysisModel::ptrValueMatcher());
+  Finder->addMatcher(
+      decl(anyOf(functionDecl(unless(isExpansionInSystemHeader()),
+                              // FIXME: Remove the filter below when lambdas are
+                              // well supported by the check.
+                              unless(hasDeclContext(cxxRecordDecl(isLambda()))),
+                              hasBody(hasPointerValue)),
+                 cxxConstructorDecl(hasAnyConstructorInitializer(
+                     withInitializer(hasPointerValue)))))
+          .bind(FuncID),
+      this);
+}
+
+void NullCheckAfterDereferenceCheck::check(
+    const MatchFinder::MatchResult &Result) {
+  if (Result.SourceManager->getDiagnostics().hasUncompilableErrorOccurred())
+    return;
+
+  const auto *FuncDecl = Result.Nodes.getNodeAs<FunctionDecl>(FuncID);
+  assert(FuncDecl && "invalid FuncDecl matcher"); 
+  if (FuncDecl->isTemplated())
+    return;
+
+  if (const auto Diagnostics = analyzeFunction(*FuncDecl)) {
+    const auto& [CheckWhenNullLocations, CheckAfterDereferenceLocations] =
+        *Diagnostics;
+
+    for (const auto [WarningLoc, DerefLoc] : CheckAfterDereferenceLocations) {
+      diag(WarningLoc,
+           "pointer value is checked even though "
+           "it cannot be null at this point");
+
+      if (DerefLoc) {
+        diag(*DerefLoc,
+             "one of the locations where the pointer's value cannot be null",
+             DiagnosticIDs::Note);
+      }
+    }
+
+    for (const auto [WarningLoc, DerefLoc] : CheckWhenNullLocations) {
+      diag(WarningLoc,
+           "pointer value is checked but it can only be null at this point");
+
+      if (DerefLoc) {
+        diag(*DerefLoc,
+             "one of the locations where the pointer's value can only be null",
+             DiagnosticIDs::Note);
+      }
+    }
+  }
+}
+
+} // namespace clang::tidy::bugprone

clang-tools-extra/clang-tidy/bugprone/NullCheckAfterDereferenceCheck.h

@@ -0,0 +1,37 @@
+//===--- NullCheckAfterDereferenceCheck.h - clang-tidy ----------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CLANG_TOOLS_EXTRA_CLANG_TIDY_BUGPRONE_NULLCHECKAFTERDEREFERENCECHECK_H
+#define LLVM_CLANG_TOOLS_EXTRA_CLANG_TIDY_BUGPRONE_NULLCHECKAFTERDEREFERENCECHECK_H
+
+#include "../ClangTidyCheck.h"
+#include "clang/ASTMatchers/ASTMatchFinder.h"
+
+namespace clang::tidy::bugprone {
+
+/// Finds checks for pointer nullability after a pointer has already been
+/// dereferenced, using the data-flow framework.
+///
+/// For the user-facing documentation see:
+/// http://clang.llvm.org/extra/clang-tidy/checks/bugprone/null-check-after-dereference.html
+class NullCheckAfterDereferenceCheck : public ClangTidyCheck {
+public:
+  NullCheckAfterDereferenceCheck(StringRef Name, ClangTidyContext *Context)
+      : ClangTidyCheck(Name, Context) {}
+
+  // The data-flow framework does not support C because of AST differences.
+  bool isLanguageVersionSupported(const LangOptions &LangOpts) const override {
+    return LangOpts.CPlusPlus;
+  }
+  void registerMatchers(ast_matchers::MatchFinder *Finder) override;
+  void check(const ast_matchers::MatchFinder::MatchResult &Result) override;
+};
+
+} // namespace clang::tidy::bugprone
+
+#endif // LLVM_CLANG_TOOLS_EXTRA_CLANG_TIDY_BUGPRONE_NULLCHECKAFTERDEREFERENCECHECK_H

clang-tools-extra/clangd/TidyProvider.cpp

@@ -219,9 +219,10 @@ TidyProvider disableUnusableChecks(llvm::ArrayRef<std::string> ExtraBadChecks) {
       "-bugprone-use-after-move",
       // Alias for bugprone-use-after-move.
       "-hicpp-invalid-access-moved",
-      // Check uses dataflow analysis, which might hang/crash unexpectedly on
+      // Checks use dataflow analysis, which might hang/crash unexpectedly on
       // incomplete code.
       "-bugprone-unchecked-optional-access",
+      "-bugprone-null-check-after-dereference",
 
       // ----- Performance problems -----