Check allocation and array sizes in minidump.cc (#12). r=bryner

http://groups.google.com/group/google-breakpad-dev/browse_thread/thread/7258b34d26ffc890


git-svn-id: http://google-breakpad.googlecode.com/svn/trunk@182 4c0a9323-5329-0410-9bdc-e9ce6186880e

2 files changed, 152 insertions, 17 deletions

diff --git a/src/google_breakpad/processor/minidump.h b/src/google_breakpad/processor/minidump.h
index b632654d..f40f4c06 100644
--- a/src/google_breakpad/processor/minidump.h
+++ b/src/google_breakpad/processor/minidump.h
@@ -221,6 +221,9 @@ class MinidumpMemoryRegion : public MinidumpObject,
  public:
   virtual ~MinidumpMemoryRegion();
 
+  static void set_max_bytes(u_int32_t max_bytes) { max_bytes_ = max_bytes; }
+  static u_int32_t max_bytes() { return max_bytes_; }
+
   // Returns a pointer to the base of the memory region.  Returns the
   // cached value if available, otherwise, reads the minidump file and
   // caches the memory region.
@@ -258,6 +261,10 @@ class MinidumpMemoryRegion : public MinidumpObject,
   template<typename T> bool GetMemoryAtAddressInternal(u_int64_t address,
                                                        T*        value);
 
+  // The largest memory region that will be read from a minidump.  The
+  // default is 1MB.
+  static u_int32_t max_bytes_;
+
   // Base address and size of the memory region, and its position in the
   // minidump file.
   MDMemoryDescriptor* descriptor_;
@@ -312,6 +319,11 @@ class MinidumpThreadList : public MinidumpStream {
  public:
   virtual ~MinidumpThreadList();
 
+  static void set_max_threads(u_int32_t max_threads) {
+    max_threads_ = max_threads;
+  }
+  static u_int32_t max_threads() { return max_threads_; }
+
   unsigned int thread_count() const { return valid_ ? thread_count_ : 0; }
 
   // Sequential access to threads.
@@ -335,6 +347,10 @@ class MinidumpThreadList : public MinidumpStream {
 
   bool Read(u_int32_t aExpectedSize);
 
+  // The largest number of threads that will be read from a minidump.  The
+  // default is 256.
+  static u_int32_t max_threads_;
+
   // Access to threads using the thread ID as the key.
   IDToThreadMap    id_to_thread_map_;
 
@@ -353,6 +369,16 @@ class MinidumpModule : public MinidumpObject,
  public:
   virtual ~MinidumpModule();
 
+  static void set_max_cv_bytes(u_int32_t max_cv_bytes) {
+    max_cv_bytes_ = max_cv_bytes;
+  }
+  static u_int32_t max_cv_bytes() { return max_cv_bytes_; }
+
+  static void set_max_misc_bytes(u_int32_t max_misc_bytes) {
+    max_misc_bytes_ = max_misc_bytes;
+  }
+  static u_int32_t max_misc_bytes() { return max_misc_bytes_; }
+
   const MDRawModule* module() const { return valid_ ? &module_ : NULL; }
 
   // CodeModule implementation
@@ -408,6 +434,12 @@ class MinidumpModule : public MinidumpObject,
   // allow the CodeModule getters to be const methods.
   bool ReadAuxiliaryData();
 
+  // The largest number of bytes that will be read from a minidump for a
+  // CodeView record or miscellaneous debugging record, respectively.  The
+  // default for each is 1024.
+  static u_int32_t max_cv_bytes_;
+  static u_int32_t max_misc_bytes_;
+
   // True after a successful Read.  This is different from valid_, which is
   // not set true until ReadAuxiliaryData also completes successfully.
   // module_valid_ is only used by ReadAuxiliaryData and the functions it
@@ -447,6 +479,11 @@ class MinidumpModuleList : public MinidumpStream,
  public:
   virtual ~MinidumpModuleList();
 
+  static void set_max_modules(u_int32_t max_modules) {
+    max_modules_ = max_modules;
+  }
+  static u_int32_t max_modules() { return max_modules_; }
+
   // CodeModules implementation.
   virtual unsigned int module_count() const {
     return valid_ ? module_count_ : 0;
@@ -472,6 +509,10 @@ class MinidumpModuleList : public MinidumpStream,
 
   bool Read(u_int32_t expected_size);
 
+  // The largest number of modules that will be read from a minidump.  The
+  // default is 1024.
+  static u_int32_t max_modules_;
+
   // Access to modules using addresses as the key.
   RangeMap<u_int64_t, unsigned int> *range_map_;
 
@@ -493,6 +534,11 @@ class MinidumpMemoryList : public MinidumpStream {
  public:
   virtual ~MinidumpMemoryList();
 
+  static void set_max_regions(u_int32_t max_regions) {
+    max_regions_ = max_regions;
+  }
+  static u_int32_t max_regions() { return max_regions_; }
+
   unsigned int region_count() const { return valid_ ? region_count_ : 0; }
 
   // Sequential access to memory regions.
@@ -517,6 +563,10 @@ class MinidumpMemoryList : public MinidumpStream {
 
   bool Read(u_int32_t expected_size);
 
+  // The largest number of memory regions that will be read from a minidump.
+  // The default is 256.
+  static u_int32_t max_regions_;
+
   // Access to memory regions using addresses as the key.
   RangeMap<u_int64_t, unsigned int> *range_map_;
 
@@ -690,6 +740,16 @@ class Minidump {
 
   ~Minidump();
 
+  static void set_max_streams(u_int32_t max_streams) {
+    max_streams_ = max_streams;
+  }
+  static u_int32_t max_streams() { return max_streams_; }
+
+  static void set_max_string_length(u_int32_t max_string_length) {
+    max_string_length_ = max_string_length;
+  }
+  static u_int32_t max_string_length() { return max_string_length_; }
+
   const MDRawHeader* header() const { return valid_ ? &header_ : NULL; }
 
   // Reads the minidump file's header and top-level stream directory.
@@ -779,6 +839,17 @@ class Minidump {
   // Opens the minidump file, or if already open, seeks to the beginning.
   bool Open();
 
+  // The largest number of top-level streams that will be read from a minidump.
+  // Note that streams are only read (and only consume memory) as needed,
+  // when directed by the caller.  The default is 128.
+  static u_int32_t max_streams_;
+
+  // The maximum length of a UTF-16 string that will be read from a minidump
+  // in 16-bit words.  The default is 1024.  UTF-16 strings are converted
+  // to UTF-8 when stored in memory, and each UTF-16 word will be represented
+  // by as many as 3 bytes in UTF-8.
+  static unsigned int max_string_length_;
+
   MDRawHeader               header_;
 
   // The list of streams.
diff --git a/src/processor/minidump.cc b/src/processor/minidump.cc
index e1828a6b..ed0ba365 100644
--- a/src/processor/minidump.cc
+++ b/src/processor/minidump.cc
@@ -686,6 +686,9 @@ void MinidumpContext::Print() {
 //
 
 
+u_int32_t MinidumpMemoryRegion::max_bytes_ = 1024 * 1024;  // 1MB
+
+
 MinidumpMemoryRegion::MinidumpMemoryRegion(Minidump* minidump)
     : MinidumpObject(minidump),
       descriptor_(NULL),
@@ -724,7 +727,13 @@ const u_int8_t* MinidumpMemoryRegion::GetMemory() {
       return NULL;
     }
 
-    // TODO(mmentovai): verify rational size!
+    if (descriptor_->memory.data_size > max_bytes_) {
+      BPLOG(ERROR) << "MinidumpMemoryRegion size " <<
+                      descriptor_->memory.data_size << " exceeds maximum " <<
+                      max_bytes_;
+      return NULL;
+    }
+
     scoped_ptr< vector<u_int8_t> > memory(
         new vector<u_int8_t>(descriptor_->memory.data_size));
 
@@ -1014,6 +1023,9 @@ void MinidumpThread::Print() {
 //
 
 
+u_int32_t MinidumpThreadList::max_threads_ = 256;
+
+
 MinidumpThreadList::MinidumpThreadList(Minidump* minidump)
     : MinidumpStream(minidump),
       id_to_thread_map_(),
@@ -1064,8 +1076,13 @@ bool MinidumpThreadList::Read(u_int32_t expected_size) {
     return false;
   }
 
-  if (thread_count) {
-    // TODO(mmentovai): verify rational size!
+  if (thread_count > max_threads_) {
+    BPLOG(ERROR) << "MinidumpThreadList count " << thread_count <<
+                    " exceeds maximum " << max_threads_;
+    return false;
+  }
+
+  if (thread_count != 0) {
     scoped_ptr<MinidumpThreads> threads(
         new MinidumpThreads(thread_count, MinidumpThread(minidump_)));
 
@@ -1157,6 +1174,10 @@ void MinidumpThreadList::Print() {
 //
 
 
+u_int32_t MinidumpModule::max_cv_bytes_ = 1024;
+u_int32_t MinidumpModule::max_misc_bytes_ = 1024;
+
+
 MinidumpModule::MinidumpModule(Minidump* minidump)
     : MinidumpObject(minidump),
       module_valid_(false),
@@ -1526,7 +1547,12 @@ const u_int8_t* MinidumpModule::GetCVRecord(u_int32_t* size) {
       return NULL;
     }
 
-    // TODO(mmentovai): verify rational size!
+    if (module_.cv_record.data_size > max_cv_bytes_) {
+      BPLOG(ERROR) << "MinidumpModule CodeView record size " <<
+                      module_.cv_record.data_size << " exceeds maximum " <<
+                      max_cv_bytes_;
+      return NULL;
+    }
 
     // Allocating something that will be accessed as MDCVInfoPDB70 or
     // MDCVInfoPDB20 but is allocated as u_int8_t[] can cause alignment
@@ -1649,7 +1675,12 @@ const MDImageDebugMisc* MinidumpModule::GetMiscRecord(u_int32_t* size) {
       return NULL;
     }
 
-    // TODO(mmentovai): verify rational size!
+    if (module_.misc_record.data_size > max_misc_bytes_) {
+      BPLOG(ERROR) << "MinidumpModule miscellaneous debugging record size " <<
+                      module_.misc_record.data_size << " exceeds maximum " <<
+                      max_misc_bytes_;
+      return NULL;
+    }
 
     // Allocating something that will be accessed as MDImageDebugMisc but
     // is allocated as u_int8_t[] can cause alignment problems.  x86 and
@@ -1847,6 +1878,9 @@ void MinidumpModule::Print() {
 //
 
 
+u_int32_t MinidumpModuleList::max_modules_ = 1024;
+
+
 MinidumpModuleList::MinidumpModuleList(Minidump* minidump)
     : MinidumpStream(minidump),
       range_map_(new RangeMap<u_int64_t, unsigned int>()),
@@ -1898,8 +1932,13 @@ bool MinidumpModuleList::Read(u_int32_t expected_size) {
     return false;
   }
 
-  if (module_count) {
-    // TODO(mmentovai): verify rational size!
+  if (module_count > max_modules_) {
+    BPLOG(ERROR) << "MinidumpModuleList count " << module_count_ <<
+                    " exceeds maximum " << max_modules_;
+    return false;
+  }
+
+  if (module_count != 0) {
     scoped_ptr<MinidumpModules> modules(
         new MinidumpModules(module_count, MinidumpModule(minidump_)));
 
@@ -2065,6 +2104,9 @@ void MinidumpModuleList::Print() {
 //
 
 
+u_int32_t MinidumpMemoryList::max_regions_ = 256;
+
+
 MinidumpMemoryList::MinidumpMemoryList(Minidump* minidump)
     : MinidumpStream(minidump),
       range_map_(new RangeMap<u_int64_t, unsigned int>()),
@@ -2120,8 +2162,13 @@ bool MinidumpMemoryList::Read(u_int32_t expected_size) {
     return false;
   }
 
-  if (region_count) {
-    // TODO(mmentovai): verify rational size!
+  if (region_count > max_regions_) {
+    BPLOG(ERROR) << "MinidumpMemoryList count " << region_count <<
+                    " exceeds maximum " << max_regions_;
+    return false;
+  }
+
+  if (region_count != 0) {
     scoped_ptr<MemoryDescriptors> descriptors(
         new MemoryDescriptors(region_count));
 
@@ -2821,6 +2868,10 @@ void MinidumpBreakpadInfo::Print() {
 //
 
 
+u_int32_t Minidump::max_streams_ = 128;
+unsigned int Minidump::max_string_length_ = 1024;
+
+
 Minidump::Minidump(const string& path)
     : header_(),
       directory_(NULL),
@@ -2934,8 +2985,13 @@ bool Minidump::Read() {
     return false;
   }
 
-  if (header_.stream_count) {
-    // TODO(mmentovai): verify rational size!
+  if (header_.stream_count > max_streams_) {
+    BPLOG(ERROR) << "Minidump stream count " << header_.stream_count <<
+                    " exceeds maximum " << max_streams_;
+    return false;
+  }
+
+  if (header_.stream_count != 0) {
     scoped_ptr<MinidumpDirectoryEntries> directory(
         new MinidumpDirectoryEntries(header_.stream_count));
 
@@ -3146,31 +3202,39 @@ string* Minidump::ReadString(off_t offset) {
     return NULL;
   }
   if (!SeekSet(offset)) {
-    BPLOG(ERROR) << "ReadString could not seek to string";
+    BPLOG(ERROR) << "ReadString could not seek to string at offset " << offset;
     return NULL;
   }
 
   u_int32_t bytes;
   if (!ReadBytes(&bytes, sizeof(bytes))) {
-    BPLOG(ERROR) << "ReadString could not read string size";
+    BPLOG(ERROR) << "ReadString could not read string size at offset " <<
+                    offset;
     return NULL;
   }
   if (swap_)
     Swap(&bytes);
 
   if (bytes % 2 != 0) {
-    BPLOG(ERROR) << "ReadString found odd-sized string of " << bytes <<
-                    " bytes at offset " << offset;
+    BPLOG(ERROR) << "ReadString found odd-sized " << bytes <<
+                    "-byte string at offset " << offset;
     return NULL;
   }
   unsigned int utf16_words = bytes / 2;
 
-  // TODO(mmentovai): verify rational size!
+  if (utf16_words > max_string_length_) {
+    BPLOG(ERROR) << "ReadString string length " << utf16_words <<
+                    " exceeds maximum " << max_string_length_ <<
+                    " at offset " << offset;
+    return NULL;
+  }
+
   vector<u_int16_t> string_utf16(utf16_words);
 
   if (utf16_words) {
     if (!ReadBytes(&string_utf16[0], bytes)) {
-      BPLOG(ERROR) << "ReadString could not read string";
+      BPLOG(ERROR) << "ReadString could not read " << bytes <<
+                      "-byte string at offset " << offset;
       return NULL;
     }
   }