Harden writer of munged bitcode for fuzzing

unittests/Bitcode/NaClMungeWriteErrorTests.cpp

Index: unittests/Bitcode/NaClMungeWriteErrorTests.cpp
diff --git a/unittests/Bitcode/NaClMungeWriteErrorTests.cpp b/unittests/Bitcode/NaClMungeWriteErrorTests.cpp
new file mode 100644
index 0000000000000000000000000000000000000000..3832fd6867c7a0780697210dfc1cea16c497e9c8
--- /dev/null
+++ b/unittests/Bitcode/NaClMungeWriteErrorTests.cpp
@@ -0,0 +1,327 @@
+//===- llvm/unittest/Bitcode/NaClMungeWriteErrorTests.cpp -----------------===//
+//     Tests parser for PNaCl bitcode instructions.
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+// Tests write errors for munged bitcode.
+
+#include "llvm/ADT/STLExtras.h"
+#include "llvm/Bitcode/NaCl/NaClBitcodeMunge.h"
+#include "llvm/Bitcode/NaCl/NaClBitcodeParser.h"
+#include "llvm/Bitcode/NaCl/NaClLLVMBitCodes.h"
+
+#include "gtest/gtest.h"
+
+using namespace llvm;
+
+namespace {
+
+// Test list of bitcode records.
+static const uint64_t Terminator = 0x5768798008978675LL;
+  const uint64_t BitcodeRecords[] = {
+    1, naclbitc::BLK_CODE_ENTER, naclbitc::MODULE_BLOCK_ID, 2, Terminator,
+    1, naclbitc::BLK_CODE_ENTER, naclbitc::TYPE_BLOCK_ID_NEW, 3, Terminator,
+    3, naclbitc::TYPE_CODE_NUMENTRY, 2, Terminator,
+    3, naclbitc::TYPE_CODE_VOID, Terminator,
+    3, naclbitc::TYPE_CODE_FUNCTION, 0, 0, Terminator,
+    0, naclbitc::BLK_CODE_EXIT, Terminator,
+    3, naclbitc::MODULE_CODE_FUNCTION, 1, 0, 0, 0, Terminator,
+    1, naclbitc::BLK_CODE_ENTER, naclbitc::FUNCTION_BLOCK_ID, 2, Terminator,
+    3, naclbitc::FUNC_CODE_DECLAREBLOCKS, 1, Terminator,
+    3, naclbitc::FUNC_CODE_INST_RET, Terminator,
+    0, naclbitc::BLK_CODE_EXIT, Terminator,
+    0, naclbitc::BLK_CODE_EXIT, Terminator
+  };
+
+// Expected output when bitcode records are dumped.
+const char* ExpectedDump =
+    "       0:0|<65532, 80, 69, 88, 69, 1, 0,|Magic Number: 'PEXE' (80, 69, "
+    "88, 69)\n"
+    "          | 8, 0, 17, 0, 4, 0, 2, 0, 0, |PNaCl Version: 2\n"
+    "          | 0>                          |\n"
+    "      16:0|1: <65535, 8, 2>             |module {  // BlockID = 8\n"
+    "      24:0|  1: <65535, 17, 3>          |  types {  // BlockID = 17\n"
+    "      32:0|    3: <1, 2>                |    count 2;\n"
+    "      34:5|    3: <2>                   |    @t0 = void;\n"
+    "      36:4|    3: <21, 0, 0>            |    @t1 = void ();\n"
+    "      39:7|  0: <65534>                 |  }\n"
+    "      44:0|  3: <8, 1, 0, 0, 0>         |  define external void @f0();\n"
+    "      48:6|  1: <65535, 12, 2>          |  function void @f0() {  \n"
+    "          |                             |                   // BlockID "
+    "= 12\n"
+    "      56:0|    3: <1, 1>                |    blocks 1;\n"
+    "          |                             |  %b0:\n"
+    "      58:4|    3: <10>                  |    ret void;\n"
+    "      60:2|  0: <65534>                 |  }\n"
+    "      64:0|0: <65534>                   |}\n"
+    ;
+
+// Edit to change void type with an illegal abbreviation index.
+const uint64_t VoidTypeIndex = 3; // Index for "@t0 = void".
+const uint64_t AbbrevIndex4VoidTypeEdit[] = {
+  VoidTypeIndex, NaClMungedBitcode::Replace,
+  4, naclbitc::TYPE_CODE_VOID, Terminator,
+};
+
+// Edit to add local abbreviation for "ret void", and then use on that
+// instruction.
+const uint64_t RetVoidIndex = 9; // return void;
+const uint64_t UseLocalRetVoidAbbrevEdits[] = {
+  RetVoidIndex, NaClMungedBitcode::AddBefore,
+    2, naclbitc::BLK_CODE_DEFINE_ABBREV, 1, 1,
+       naclbitc::FUNC_CODE_INST_RET, Terminator,
+  RetVoidIndex, NaClMungedBitcode::Replace,
+    4, naclbitc::FUNC_CODE_INST_RET, Terminator
+};
+
+#define ARRAY_ARGS(Records) Records, array_lengthof(Records)
+
+#define ARRAY_ARGS_TERM(Records) ARRAY_ARGS(Records), Terminator
+
+std::string stringify(NaClBitcodeMunger &Munger) {
+  std::string Buffer;
+  raw_string_ostream StrBuf(Buffer);
+  Munger.getMungedBitcode().print(StrBuf);
+  return StrBuf.str();
+}
+
+// Show that we can dump the bitcode records
+TEST(NaClMungeWriteErrorTests, DumpBitcodeRecords) {
+  NaClObjDumpMunger Munger(ARRAY_ARGS_TERM(BitcodeRecords));
+  EXPECT_TRUE(Munger.runTest("DumpBitcodeRecords"));
+  EXPECT_EQ(ExpectedDump, Munger.getTestResults());
+}
+
+// Show that by default, one can't write a bad abbreviation index.
+TEST(NaClMungeWriteErrorTests, CantWriteBadAbbrevIndex) {
+  NaClWriteMunger Munger(ARRAY_ARGS_TERM(BitcodeRecords));
+  EXPECT_FALSE(Munger.runTest("CantWriteBadAbbrevIndex",
+                              ARRAY_ARGS(AbbrevIndex4VoidTypeEdit)));
+  EXPECT_EQ(
+      "Error (Block 17): Uses illegal abbreviation index: 4: [2]\n"
+      "Error (Block 17): Missing close block.\n"
+      "Error (Block 8): Missing close block.\n"
+      "Error: Unable to generate bitcode file due to write errors\n",
+      Munger.getTestResults());
+}
+
+// Show that we can't write more local abbreviations than specified in
+// the corresponding enclosing block.
+TEST(NaClMungeWriteErrorTests, CantDumpTooManyLocalAbbreviations) {

[jvoung (off chromium), 2015/05/13 00:58:17]

Maybe "define" instead of "dump"? I think of the object dumps / ExpectedDump,
etc.., instead with "dump".

[Karl, 2015/05/13 21:39:25]

Changing CantWriteTooManyLocallAbbreviations.

+  NaClWriteMunger Munger(ARRAY_ARGS_TERM(BitcodeRecords));
+  Munger.munge(ARRAY_ARGS(UseLocalRetVoidAbbrevEdits));
+  EXPECT_EQ(
+      "       1: [65535, 8, 2]\n"
+      "         1: [65535, 17, 3]\n"
+      "           3: [1, 2]\n"
+      "           3: [2]\n"
+      "           3: [21, 0, 0]\n"
+      "         0: [65534]\n"
+      "         3: [8, 1, 0, 0, 0]\n"
+      "         1: [65535, 12, 2]\n"
+      "           3: [1, 1]\n"
+      "           2: [65533, 1, 1, 10]\n"
+      "           4: [10]\n"
+      "         0: [65534]\n"
+      "       0: [65534]\n",
+      stringify(Munger));
+
+  EXPECT_FALSE(Munger.runTest("CantDumpTooManyLocalAbbreviations"));
+  EXPECT_EQ(
+      "Error (Block 12): Exceeds abbreviation index limit of 3: 2:"
+      " [65533, 1, 1, 10]\n"
+      "Error (Block 12): Missing close block.\n"
+      "Error (Block 8): Missing close block.\n"
+      "Error: Unable to generate bitcode file due to write errors\n",
+      Munger.getTestResults());
+}
+
+// Show what happens when there are more enter blocks then exit blocks.
+TEST(NaClMungeWriteErrorTests, TooManyEnterBlocks) {
+  NaClWriteMunger Munger(ARRAY_ARGS_TERM(BitcodeRecords));
+  // Remove all but first two records (i.e. two enter blocks).
+  NaClMungedBitcode &MungedBitcode = Munger.getMungedBitcode();
+  for (size_t i = 2; i < MungedBitcode.getBaseRecords().size(); ++i) {
+    MungedBitcode.remove(i);
+  }
+
+  EXPECT_FALSE(Munger.runTest("TooManyEnterBlocks"));
+  EXPECT_EQ(
+      "Error (Block 17): Missing close block.\n"
+      "Error (Block 8): Missing close block.\n"
+      "Error: Unable to generate bitcode file due to write errors\n",
+      Munger.getTestResults());
+}
+
+// Show what happens when there are fewer enter blocks than exit
+// blocks.
+TEST(NaClMungeWriteErrorTests, TooFewExitBlocks) {

[jvoung (off chromium), 2015/05/13 00:58:17]

re: TooFewExitBlocks, in this case there are too few enter blocks? Or too many
exit blocks.

[Karl, 2015/05/13 21:39:24]

Fixing to TooMayExitBlocks.

+  NaClWriteMunger Munger(ARRAY_ARGS_TERM(BitcodeRecords));
+  // Add two exit blocks.
+  NaClMungedBitcode &MungedBitcode = Munger.getMungedBitcode();
+  NaClRecordVector Values;
+  NaClBitcodeAbbrevRecord Record(0, naclbitc::BLK_CODE_EXIT, Values);
+  for (size_t i = 0; i < 2; ++i)
+    MungedBitcode.addAfter(MungedBitcode.getBaseRecords().size() - 1, Record);
+
+  EXPECT_FALSE(Munger.runTest("TooFewExitBlocks"));

[jvoung (off chromium), 2015/05/13 00:58:17]

"TooFewExitBlocks" -> ?

[Karl, 2015/05/13 21:39:24]

Done.

+  EXPECT_EQ(
+      "Error (Block unknown): Extraneous exit block: 0: [65534]\n"
+      "Error: Unable to generate bitcode file due to write errors\n",
+      Munger.getTestResults());
+}
+
+// Show that an error occurs when writing a bitcode record that isn't
+// in any block.
+TEST(NaClMungeWriteErrorTests, CantWriteRecordOutsideBlock) {
+  NaClWriteMunger Munger(ARRAY_ARGS_TERM(BitcodeRecords));
+  NaClMungedBitcode &MungedBitcode = Munger.getMungedBitcode();
+  NaClRecordVector Values;
+  Values.push_back(4);
+  NaClBitcodeAbbrevRecord Record(naclbitc::UNABBREV_RECORD,
+                                 naclbitc::MODULE_CODE_VERSION,
+                                 Values);
+
+  MungedBitcode.addAfter(MungedBitcode.getBaseRecords().size() - 1, Record);
+  EXPECT_FALSE(Munger.runTest("CantWriteRecordOutsideBlock"));
+  EXPECT_EQ(
+      "Error (Block unknown): Record outside block: 3: [1, 4]\n"
+      "Error: Unable to generate bitcode file due to write errors\n",
+      Munger.getTestResults());
+}
+
+// Show that error recovery works when parsing bitcode with a bad
+// abbreviation index.

[jvoung (off chromium), 2015/05/13 00:58:17]

So to be clear -- the error recovery is in the writer, which writes out
parseable bitcode as the recover step. The reader doesn't recover and doesn't
run into an error.

[Karl, 2015/05/13 21:39:24]

Yes.

Rewriting error recovery comments to be clearer about this.

+TEST(NaClMungeWriteErrorTests, RecoverWhenParsingBadAbbrevIndex) {
+  NaClParseBitcodeMunger Munger(ARRAY_ARGS_TERM(BitcodeRecords));
+  Munger.setTryToRecoverOnWrite(true);
+  EXPECT_TRUE(
+      Munger.runTest("RecoverWhenParsingBadAbbrevIndex",
+                     ARRAY_ARGS(AbbrevIndex4VoidTypeEdit), true));
+  EXPECT_EQ(
+      "Error (Block 17): Uses illegal abbreviation index: 4: [2]\n"
+      "Successful parse!\n",
+      Munger.getTestResults());
+}
+
+// Show that error recovery works when dumping bitcode with a bad
+// abbreviation index.
+TEST(NaClMungeWriteErrorTests, RecoverWhenParsingBadAbbreviationIndex) {
+  NaClObjDumpMunger Munger(ARRAY_ARGS_TERM(BitcodeRecords));
+  Munger.setTryToRecoverOnWrite(true);
+  EXPECT_TRUE(Munger.runTest("RecoverWhenParsingBadAbbreviationIndex",
+                             ARRAY_ARGS(AbbrevIndex4VoidTypeEdit)));
+  std::string Results(
+      "Error (Block 17): Uses illegal abbreviation index: 4: [2]\n");
+  Results.append(ExpectedDump);
+  EXPECT_EQ(Results, Munger.getTestResults());
+}
+
+// Show that error recovery works when there is too many locally
+// defined abbreviations for the corresponding number of bits defined
+// in the corresponding enter block.
+TEST(NaClMungeWriteErrorTests, RecoverTooManyLocalAbbreviations) {
+  NaClObjDumpMunger Munger(ARRAY_ARGS_TERM(BitcodeRecords));
+  Munger.setTryToRecoverOnWrite(true);
+  Munger.munge(ARRAY_ARGS(UseLocalRetVoidAbbrevEdits));
+
+  EXPECT_TRUE(Munger.runTest("RecoverTooManyLocalAbbreviations"));
+  std::string Results(
+      "Error (Block 12): Exceeds abbreviation index limit of 3: 2:"
+      " [65533, 1, 1, 10]\n"
+      "Error (Block 12): Uses illegal abbreviation index: 4: [10]\n");
+  Results.append(ExpectedDump);
+  EXPECT_EQ(
+      Results,
+      Munger.getTestResults());
+}
+
+// Show that error recovery works when there are more enter blocks than
+// exit blocks.
+TEST(NaClMungeWriteErrorTests, RecoverTooManyEnterBlocks) {
+  NaClObjDumpMunger Munger(ARRAY_ARGS_TERM(BitcodeRecords));
+  // Remove all but first two records (i.e. two enter blocks).
+  NaClMungedBitcode &MungedBitcode = Munger.getMungedBitcode();
+  for (size_t i = 2; i < MungedBitcode.getBaseRecords().size(); ++i) {
+    MungedBitcode.remove(i);
+  }
+
+  Munger.setTryToRecoverOnWrite(true);
+  EXPECT_TRUE(Munger.runTest("RecoverTooManyEnterBlocks"));
+  EXPECT_EQ(
+      "Error (Block 17): Missing close block.\n"
+      "Error (Block 8): Missing close block.\n"
+      "       0:0|<65532, 80, 69, 88, 69, 1, 0,|Magic Number: 'PEXE' (80, 69,"
+      " 88, 69)\n"
+      "          | 8, 0, 17, 0, 4, 0, 2, 0, 0, |PNaCl Version: 2\n"
+      "          | 0>                          |\n"
+      "      16:0|1: <65535, 8, 2>             |module {  // BlockID = 8\n"
+      "      24:0|  1: <65535, 17, 3>          |  types {  // BlockID = 17\n"
+      "      32:0|  0: <65534>                 |  }\n"
+      "      36:0|0: <65534>                   |}\n",
+      Munger.getTestResults());
+}
+
+// Show that error recovery works when there are fewer enter blocks
+// than exit blocks.

[jvoung (off chromium), 2015/05/13 00:58:17]

Similar comment to TooFewExitBlocks

+TEST(NaClMungeWriteErrorTests, RecoverTooFewExitBlocks) {
+  NaClObjDumpMunger Munger(ARRAY_ARGS_TERM(BitcodeRecords));
+  // Add two exit blocks.
+  NaClMungedBitcode &MungedBitcode = Munger.getMungedBitcode();
+  NaClRecordVector Values;
+  NaClBitcodeAbbrevRecord Record(0, naclbitc::BLK_CODE_EXIT, Values);
+  for (size_t i = 0; i < 2; ++i)
+    MungedBitcode.addAfter(MungedBitcode.getBaseRecords().size() - 1, Record);
+
+  Munger.setTryToRecoverOnWrite(true);
+  EXPECT_TRUE(Munger.runTest("RecoverTooFewExitBlocks"));
+  std::string Results(
+      "Error (Block unknown): Extraneous exit block: 0: [65534]\n"
+      "Error (Block unknown): Extraneous exit block: 0: [65534]\n");
+  Results.append(ExpectedDump);
+  EXPECT_EQ(
+      Results,
+      Munger.getTestResults());
+}
+
+// Show that error recovery works when writing a bitcode record that
+// isn't in any block.
+TEST(NaClMungeWriteErrorTests, RecoverWriteRecordOutsideBlock) {
+  NaClWriteMunger Munger(ARRAY_ARGS_TERM(BitcodeRecords));
+  NaClMungedBitcode &MungedBitcode = Munger.getMungedBitcode();
+  NaClRecordVector Values;
+  Values.push_back(4);
+  NaClBitcodeAbbrevRecord Record(naclbitc::UNABBREV_RECORD,
+                                 naclbitc::MODULE_CODE_VERSION,
+                                 Values);
+  MungedBitcode.addAfter(MungedBitcode.getBaseRecords().size() - 1, Record);
+
+  Munger.setTryToRecoverOnWrite(true);
+  EXPECT_TRUE(Munger.runTest("RecoverWriteRecordOutsideBlock"));
+  EXPECT_EQ(
+      "Error (Block unknown): Record outside block: 3: [1, 4]\n"
+      "Error (Block unknown): Missing close block.\n"
+      "       1: [65535, 8, 2]\n"
+      "         1: [65535, 17, 3]\n"
+      "           3: [1, 2]\n"
+      "           3: [2]\n"
+      "           3: [21, 0, 0]\n"
+      "         0: [65534]\n"
+      "         3: [8, 1, 0, 0, 0]\n"
+      "         1: [65535, 12, 2]\n"
+      "           3: [1, 1]\n"
+      "           3: [10]\n"
+      "         0: [65534]\n"
+      "       0: [65534]\n"
+      "       1: [65535, 4294967295, 3]\n"
+      "         3: [1, 4]\n"
+      "       0: [65534]\n",
+      Munger.getTestResults());
+}
+
+} // end of anonymous namespace.