[Coverage] Rework !SystemHeadersCoverage (#91446)

- Introduce `LeafExprSet`,
  - Suppress traversing LAnd and LOr expr under system headers.
- Handle LAnd and LOr as instrumented leaves to override
`!isInstrumentedCondition(C)`.
- Replace Loc with FileLoc if it is expanded with system headers.

Fixes #78920

Author: chapuni

clang/lib/CodeGen/CoverageMappingGen.cpp

@@ -17,6 +17,7 @@
 #include "clang/Basic/FileManager.h"
 #include "clang/Frontend/FrontendDiagnostic.h"
 #include "clang/Lex/Lexer.h"
+#include "llvm/ADT/DenseSet.h"
 #include "llvm/ADT/SmallSet.h"
 #include "llvm/ADT/StringExtras.h"
 #include "llvm/ProfileData/Coverage/CoverageMapping.h"
@@ -336,16 +337,26 @@ class CoverageMappingBuilder {
 
     llvm::SmallSet<FileID, 8> Visited;
     SmallVector<std::pair<SourceLocation, unsigned>, 8> FileLocs;
-    for (const auto &Region : SourceRegions) {
+    for (auto &Region : SourceRegions) {
       SourceLocation Loc = Region.getBeginLoc();
+
+      // Replace Loc with FileLoc if it is expanded with system headers.
+      if (!SystemHeadersCoverage && SM.isInSystemMacro(Loc)) {
+        auto BeginLoc = SM.getSpellingLoc(Loc);
+        auto EndLoc = SM.getSpellingLoc(Region.getEndLoc());
+        if (SM.isWrittenInSameFile(BeginLoc, EndLoc)) {
+          Loc = SM.getFileLoc(Loc);
+          Region.setStartLoc(Loc);
+          Region.setEndLoc(SM.getFileLoc(Region.getEndLoc()));
+        }
+      }
+
       FileID File = SM.getFileID(Loc);
       if (!Visited.insert(File).second)
         continue;
 
-      // Do not map FileID's associated with system headers unless collecting
-      // coverage from system headers is explicitly enabled.
-      if (!SystemHeadersCoverage && SM.isInSystemHeader(SM.getSpellingLoc(Loc)))
-        continue;
+      assert(SystemHeadersCoverage ||
+             !SM.isInSystemHeader(SM.getSpellingLoc(Loc)));
 
       unsigned Depth = 0;
       for (SourceLocation Parent = getIncludeOrExpansionLoc(Loc);
@@ -818,6 +829,10 @@ struct CounterCoverageMappingBuilder
   /// A stack of currently live regions.
   llvm::SmallVector<SourceMappingRegion> RegionStack;
 
+  /// Set if the Expr should be handled as a leaf even if it is kind of binary
+  /// logical ops (&&, ||).
+  llvm::DenseSet<const Stmt *> LeafExprSet;
+
   /// An object to manage MCDC regions.
   MCDCCoverageBuilder MCDCBuilder;
 
@@ -1040,7 +1055,10 @@ struct CounterCoverageMappingBuilder
     // region onto RegionStack but immediately pop it (which adds it to the
     // function's SourceRegions) because it doesn't apply to any other source
     // code other than the Condition.
-    if (CodeGenFunction::isInstrumentedCondition(C)) {
+    // With !SystemHeadersCoverage, binary logical ops in system headers may be
+    // treated as instrumentable conditions.
+    if (CodeGenFunction::isInstrumentedCondition(C) ||
+        LeafExprSet.count(CodeGenFunction::stripCond(C))) {
       mcdc::Parameters BranchParams;
       mcdc::ConditionID ID = MCDCBuilder.getCondID(C);
       if (ID >= 0)
@@ -2070,7 +2088,20 @@ struct CounterCoverageMappingBuilder
     createDecisionRegion(E, DecisionParams);
   }
 
+  /// Check if E belongs to system headers.
+  bool isExprInSystemHeader(const BinaryOperator *E) const {
+    return (!SystemHeadersCoverage &&
+            SM.isInSystemHeader(SM.getSpellingLoc(E->getOperatorLoc())) &&
+            SM.isInSystemHeader(SM.getSpellingLoc(E->getBeginLoc())) &&
+            SM.isInSystemHeader(SM.getSpellingLoc(E->getEndLoc())));
+  }
+
   void VisitBinLAnd(const BinaryOperator *E) {
+    if (isExprInSystemHeader(E)) {
+      LeafExprSet.insert(E);
+      return;
+    }
+
     bool IsRootNode = MCDCBuilder.isIdle();
 
     // Keep track of Binary Operator and assign MCDC condition IDs.
@@ -2125,6 +2156,11 @@ struct CounterCoverageMappingBuilder
   }
 
   void VisitBinLOr(const BinaryOperator *E) {
+    if (isExprInSystemHeader(E)) {
+      LeafExprSet.insert(E);
+      return;
+    }
+
     bool IsRootNode = MCDCBuilder.isIdle();
 
     // Keep track of Binary Operator and assign MCDC condition IDs.

clang/test/CoverageMapping/mcdc-system-headers.cpp

@@ -0,0 +1,50 @@
+// RUN: %clang_cc1 -std=c++11 -fprofile-instrument=clang -fcoverage-mapping -dump-coverage-mapping  -fcoverage-mcdc -mllvm -system-headers-coverage -emit-llvm-only -o - %s | FileCheck %s --check-prefixes=CHECK,W_SYS
+// RUN: %clang_cc1 -std=c++11 -fprofile-instrument=clang -fcoverage-mapping -dump-coverage-mapping -fcoverage-mcdc -emit-llvm-only -o - %s | FileCheck %s --check-prefixes=CHECK,X_SYS
+
+#ifdef IS_SYSHEADER
+
+#pragma clang system_header
+#define CONST 42
+#define EXPR1(x) (x)
+#define EXPR2(x) ((x) && (x))
+
+#else
+
+#define IS_SYSHEADER
+#include __FILE__
+
+// CHECK: _Z5func0i:
+int func0(int a) {
+  // CHECK: Decision,File 0, [[@LINE+3]]:11 -> [[@LINE+3]]:21 = M:0, C:2
+  // W_SYS: Expansion,File 0, [[@LINE+2]]:11 -> [[@LINE+2]]:16 = #0 (Expanded file = 1)
+  // X_SYS: Branch,File 0, [[@LINE+1]]:11 -> [[@LINE+1]]:11 = 0, 0 [1,2,0]
+  return (CONST && a);
+  // CHECK: Branch,File 0, [[@LINE-1]]:20 -> [[@LINE-1]]:21 = #2, (#1 - #2) [2,0,0]
+  // W_SYS: Branch,File 1, [[@LINE-16]]:15 -> [[@LINE-16]]:17 = 0, 0 [1,2,0]
+}
+
+// CHECK: _Z5func1ii:
+int func1(int a, int b) {
+  // CHECK: Decision,File 0, [[@LINE+2]]:11 -> [[@LINE+2]]:21 = M:0, C:2
+  // CHECK: Branch,File 0, [[@LINE+1]]:11 -> [[@LINE+1]]:12 = (#0 - #1), #1 [1,0,2]
+  return (a || EXPR1(b));
+  // W_SYS: Expansion,File 0, [[@LINE-1]]:16 -> [[@LINE-1]]:21 = #1 (Expanded file = 1)
+  // W_SYS: Branch,File 1, [[@LINE-24]]:18 -> [[@LINE-24]]:21 = (#1 - #2), #2 [2,0,0]
+  // X_SYS: Branch,File 0, [[@LINE-3]]:16 -> [[@LINE-3]]:16 = (#1 - #2), #2 [2,0,0]
+}
+
+// CHECK: _Z5func2ii:
+int func2(int a, int b) {
+  // W_SYS: Decision,File 0, [[@LINE+5]]:11 -> [[@LINE+5]]:28 = M:0, C:3
+  // X_SYS: Decision,File 0, [[@LINE+4]]:11 -> [[@LINE+4]]:28 = M:0, C:2
+  // W_SYS: Expansion,File 0, [[@LINE+3]]:11 -> [[@LINE+3]]:16 = #0 (Expanded file = 1)
+  // W_SYS: Expansion,File 0, [[@LINE+2]]:23 -> [[@LINE+2]]:28 = #1 (Expanded file = 2)
+  // X_SYS: Branch,File 0, [[@LINE+1]]:11 -> [[@LINE+1]]:11 = #1, (#0 - #1) [1,2,0]
+  return (EXPR2(a) && EXPR1(a));
+  // W_SYS: Branch,File 1, [[@LINE-35]]:19 -> [[@LINE-35]]:22 = #3, (#0 - #3) [1,3,0]
+  // W_SYS: Branch,File 1, [[@LINE-36]]:26 -> [[@LINE-36]]:29 = #4, (#3 - #4) [3,2,0]
+  // W_SYS: Branch,File 2, [[@LINE-38]]:18 -> [[@LINE-38]]:21 = #2, (#1 - #2) [2,0,0]
+  // X_SYS: Branch,File 0, [[@LINE-4]]:23 -> [[@LINE-4]]:23 = #2, (#1 - #2) [2,0,0]
+}
+
+#endif