GH-48335: [C++][Parquet] Fuzz encrypted files

Author: pitrou

cpp/src/arrow/util/fuzz_internal.cc

@@ -49,7 +49,7 @@ void LogFuzzStatus(const Status& st, const uint8_t* data, int64_t size) {
     return value;
   }();
 
-  if (kVerbosity >= 1) {
+  if (!st.ok() && kVerbosity >= 1) {
     ARROW_LOG(WARNING) << "Fuzzing input with size=" << size
                        << " failed: " << st.ToString();
   } else if (st.IsOutOfMemory()) {

cpp/src/parquet/CMakeLists.txt

@@ -151,6 +151,7 @@ endif()
 # Library config
 
 set(PARQUET_SRCS
+    arrow/fuzz_internal.cc
     arrow/path_internal.cc
     arrow/reader.cc
     arrow/reader_internal.cc

cpp/src/parquet/arrow/arrow_reader_writer_test.cc

@@ -51,7 +51,7 @@
 #include "arrow/type_fwd.h"
 #include "arrow/type_traits.h"
 #include "arrow/util/checked_cast.h"
-#include "arrow/util/config.h"  // for ARROW_CSV definition
+#include "arrow/util/config.h"  // for ARROW_CSV and PARQUET_REQUIRE_ENCRYPTION
 #include "arrow/util/decimal.h"
 #include "arrow/util/future.h"
 #include "arrow/util/key_value_metadata.h"
@@ -65,6 +65,7 @@
 #include "parquet/api/reader.h"
 #include "parquet/api/writer.h"
 
+#include "parquet/arrow/fuzz_internal.h"
 #include "parquet/arrow/reader.h"
 #include "parquet/arrow/reader_internal.h"
 #include "parquet/arrow/schema.h"
@@ -5830,29 +5831,53 @@ TEST(TestArrowReadWrite, MultithreadedWrite) {
 }
 
 TEST(TestArrowReadWrite, FuzzReader) {
+  using ::parquet::fuzzing::internal::FuzzReader;
+
   constexpr size_t kMaxFileSize = 1024 * 1024 * 1;
+
   auto check_bad_file = [&](const std::string& file_name) {
     SCOPED_TRACE(file_name);
     auto path = test::get_data_file(file_name, /*is_good=*/false);
     PARQUET_ASSIGN_OR_THROW(auto source, ::arrow::io::MemoryMappedFile::Open(
                                              path, ::arrow::io::FileMode::READ));
     PARQUET_ASSIGN_OR_THROW(auto buffer, source->Read(kMaxFileSize));
-    auto s = internal::FuzzReader(buffer->data(), buffer->size());
+    auto s = FuzzReader(buffer->data(), buffer->size());
     ASSERT_NOT_OK(s);
   };
+
+  auto check_good_file = [&](const std::string& file_name, bool expect_error = false) {
+    SCOPED_TRACE(file_name);
+    auto path = test::get_data_file(file_name, /*is_good=*/true);
+    PARQUET_ASSIGN_OR_THROW(auto source, ::arrow::io::MemoryMappedFile::Open(
+                                             path, ::arrow::io::FileMode::READ));
+    PARQUET_ASSIGN_OR_THROW(auto buffer, source->Read(kMaxFileSize));
+    auto s = FuzzReader(buffer->data(), buffer->size());
+    if (expect_error) {
+      ASSERT_NOT_OK(s);
+    } else {
+      ASSERT_OK(s);
+    }
+  };
+
   check_bad_file("PARQUET-1481.parquet");
   check_bad_file("ARROW-GH-41317.parquet");
   check_bad_file("ARROW-GH-41321.parquet");
   check_bad_file("ARROW-RS-GH-6229-LEVELS.parquet");
   check_bad_file("ARROW-RS-GH-6229-DICTHEADER.parquet");
-  {
-    auto path = test::get_data_file("alltypes_plain.parquet", /*is_good=*/true);
-    PARQUET_ASSIGN_OR_THROW(auto source, ::arrow::io::MemoryMappedFile::Open(
-                                             path, ::arrow::io::FileMode::READ));
-    PARQUET_ASSIGN_OR_THROW(auto buffer, source->Read(kMaxFileSize));
-    auto s = internal::FuzzReader(buffer->data(), buffer->size());
-    ASSERT_OK(s);
-  }
+
+  check_good_file("alltypes_plain.parquet");
+  check_good_file("data_index_bloom_encoding_stats.parquet");
+  check_good_file("data_index_bloom_encoding_with_length.parquet");
+#ifdef PARQUET_REQUIRE_ENCRYPTION
+  // Encrypted files in the testing repo should be ok, except those
+  // that require external key material or an explicitly-supplied AAD.
+  check_good_file("uniform_encryption.parquet.encrypted");
+  check_good_file("encrypt_columns_and_footer_aad.parquet.encrypted");
+  check_good_file("encrypt_columns_and_footer.parquet.encrypted");
+  check_good_file("encrypt_columns_plaintext_footer.parquet.encrypted");
+#else
+  check_good_file("uniform_encryption.parquet.encrypted", /*expect_error=*/true);
+#endif
 }
 
 // Test writing table with a closed writer, should not segfault (GH-37969).

cpp/src/parquet/arrow/fuzz.cc

@@ -17,10 +17,10 @@
 
 #include "arrow/status.h"
 #include "arrow/util/fuzz_internal.h"
-#include "parquet/arrow/reader.h"
+#include "parquet/arrow/fuzz_internal.h"
 
 extern "C" int LLVMFuzzerTestOneInput(const uint8_t* data, size_t size) {
-  auto status = parquet::arrow::internal::FuzzReader(data, static_cast<int64_t>(size));
+  auto status = parquet::fuzzing::internal::FuzzReader(data, static_cast<int64_t>(size));
   arrow::internal::LogFuzzStatus(status, data, static_cast<int64_t>(size));
   return 0;
 }