[preprocessor] When #including something that contributes no tokens at all,

don't recursively continue lexing.

This avoids a stack overflow with a sequence of many empty #includes.
rdar://11988695

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@167801 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/include/clang/Basic/TokenKinds.def b/include/clang/Basic/TokenKinds.def
index 25e8d5a635c7b48ef76ba3fdeb70cbc22ba771ec..704f1d9b32659293cde677dc55132d55b353d26d 100644 (file)
--- a/include/clang/Basic/TokenKinds.def
+++ b/include/clang/Basic/TokenKinds.def
@@ -103,6 +103,8 @@ TOK(unknown)             // Not a token.
 TOK(eof)                 // End of file.
 TOK(eod)                 // End of preprocessing directive (end of line inside a
                          // directive).
+TOK(included_eof)        // End of included file, only if
+                         // PreprocessorLexer::EnableIncludedEOF is true.
 TOK(code_completion)     // Code completion marker
 TOK(cxx_defaultarg_end)  // C++ default argument end marker
 

diff --git a/include/clang/Lex/PreprocessorLexer.h b/include/clang/Lex/PreprocessorLexer.h
index 20fb8a0c48041199fee26c6f28bf9621f7a3c46a..b85729dd08c0543ca85bd2939655a0f590142443 100644 (file)
--- a/include/clang/Lex/PreprocessorLexer.h
+++ b/include/clang/Lex/PreprocessorLexer.h
@@ -61,6 +61,10 @@ protected:
   /// Note that in raw mode that the PP pointer may be null.
   bool LexingRawMode;
 
+  /// \brief When true, if EOF of the current lexer is found, tok::included_eof
+  /// is returned instead of continuing lexing higher in the include stack.
+  bool EnableIncludedEOF;
+
   /// \brief A state machine that detects the \#ifndef-wrapping a file
   /// idiom for the multiple-include optimization.
   MultipleIncludeOpt MIOpt;
@@ -79,7 +83,8 @@ protected:
     : PP(0), InitialNumSLocEntries(0),
       ParsingPreprocessorDirective(false),
       ParsingFilename(false),
-      LexingRawMode(false) {}
+      LexingRawMode(false),
+      EnableIncludedEOF(false) {}
 
   virtual ~PreprocessorLexer() {}
 
@@ -147,6 +152,11 @@ public:
   /// \brief Return true if this lexer is in raw mode or not.
   bool isLexingRawMode() const { return LexingRawMode; }
 
+  /// \brief When true, if EOF of the current lexer is found, tok::included_eof
+  /// is returned instead of continuing lexing higher in the include stack.
+  /// False is the default behavior.
+  void setEnableIncludedEOF(bool Enable) { EnableIncludedEOF = Enable; }
+
   /// \brief Return the preprocessor object for this lexer.
   Preprocessor *getPP() const { return PP; }
 

diff --git a/lib/Lex/Lexer.cpp b/lib/Lex/Lexer.cpp
index e6e7ca5ee17013349fa7433c84bec85f5a74bcff..7983a6b3e8624a51447099642fdcb0d0d4197838 100644 (file)
--- a/lib/Lex/Lexer.cpp
+++ b/lib/Lex/Lexer.cpp
@@ -2595,8 +2595,14 @@ LexNextToken:
       // Read the PP instance variable into an automatic variable, because
       // LexEndOfFile will often delete 'this'.
       Preprocessor *PPCache = PP;
+      bool EnableIncludedEOFCache = EnableIncludedEOF;
       if (LexEndOfFile(Result, CurPtr-1))  // Retreat back into the file.
         return;   // Got a token to return.
+
+      if (EnableIncludedEOFCache) {
+        Result.setKind(tok::included_eof);
+        return;
+      }
       assert(PPCache && "Raw buffer::LexEndOfFile should return a token");
       return PPCache->Lex(Result);
     }
@@ -3234,5 +3240,21 @@ HandleDirective:
     }
     goto LexNextToken;   // GCC isn't tail call eliminating.
   }
+
+  if (PreprocessorLexer *PPLex = PP->getCurrentLexer()) {
+    // If we #include something that contributes no tokens at all, return with
+    // a tok::included_eof instead of recursively continuing lexing.
+    // This avoids a stack overflow with a sequence of many empty #includes.
+    PPLex->setEnableIncludedEOF(true);
+    PP->Lex(Result);
+    if (Result.isNot(tok::included_eof)) {
+      if (Result.isNot(tok::eof) && Result.isNot(tok::eod))
+        PPLex->setEnableIncludedEOF(false);
+      return;
+    }
+    if (PP->isCurrentLexer(this))
+      goto LexNextToken;
+  }
+
   return PP->Lex(Result);
 }

diff --git a/lib/Lex/PreprocessorLexer.cpp b/lib/Lex/PreprocessorLexer.cpp
index a64c84d6bbd3f6c28d8861c8e79f2f5f6edc7913..390d4c4523b0759fad1aaf6a3576ce0d0e267b50 100644 (file)
--- a/lib/Lex/PreprocessorLexer.cpp
+++ b/lib/Lex/PreprocessorLexer.cpp
@@ -22,7 +22,7 @@ void PreprocessorLexer::anchor() { }
 PreprocessorLexer::PreprocessorLexer(Preprocessor *pp, FileID fid)
   : PP(pp), FID(fid), InitialNumSLocEntries(0),
     ParsingPreprocessorDirective(false),
-    ParsingFilename(false), LexingRawMode(false) {
+    ParsingFilename(false), LexingRawMode(false), EnableIncludedEOF(false) {
   if (pp)
     InitialNumSLocEntries = pp->getSourceManager().local_sloc_entry_size();
 }

diff --git a/test/Index/Inputs/empty.h b/test/Index/Inputs/empty.h
new file mode 100644 (file)
index 0000000..e69de29

diff --git a/test/Index/pp-many-includes.c b/test/Index/pp-many-includes.c
new file mode 100644 (file)
index 0000000..8a53b4f
--- /dev/null
+++ b/test/Index/pp-many-includes.c
@@ -0,0 +1,208 @@
+// RUN: env LIBCLANG_THREAD_SMALLSTACK=1 c-index-test -index-file %s | FileCheck %s
+
+// rdar://11988695
+// Check that we don't get stack overflow while including a lot of empty includes.
+
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+#include "Inputs/empty.h"
+
+// CHECK: [indexDeclaration]: kind: function | name: foo
+void foo();

diff --git a/tools/libclang/CIndex.cpp b/tools/libclang/CIndex.cpp
index 3a6c408bff1f470aeb9abc0da818bedef826d9db..bb75db3554b6283e9312e1c21c7da7e0c9fdc5cb 100644 (file)
--- a/tools/libclang/CIndex.cpp
+++ b/tools/libclang/CIndex.cpp
@@ -6119,6 +6119,8 @@ namespace clang {
 bool RunSafely(llvm::CrashRecoveryContext &CRC,
                void (*Fn)(void*), void *UserData,
                unsigned Size) {
+  if (getenv("LIBCLANG_THREAD_SMALLSTACK"))
+    Size = 0x10000; // 64K.
   if (!Size)
     Size = GetSafetyThreadStackSize();
   if (Size)