Reland "[regexp] Hard-crash on invalid offsets in AdvanceCurrentPosition"

schuay · Commit Bot · commit ff8d0f92d423 · 2021-01-14T07:37:27.000Z
This is a reland of 164cf80 The reland fixes UB (left-shift of negative integer type) with a static_cast<uint32_t>. Original change's description: > [regexp] Hard-crash on invalid offsets in AdvanceCurrentPosition > > Drive-by: Range checks in `Emit(byte, twenty_four_bits)` to ensure the > given packed bits actually fit into 24 bits. > > Bug: chromium:1166138 > Change-Id: I2e711e6466bb48d7b9897f68dfe621d12bd92508 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2625877 > Commit-Queue: Jakob Gruber <jgruber@chromium.org> > Commit-Queue: Leszek Swirski <leszeks@chromium.org> > Auto-Submit: Jakob Gruber <jgruber@chromium.org> > Reviewed-by: Leszek Swirski <leszeks@chromium.org> > Cr-Commit-Position: refs/heads/master@{#72064} Tbr: leszeks@chromium.org Bug: chromium:1166138 Change-Id: I514495e14bb99dfc9588fdb4a9f35d67d8d64acb Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2626663 Reviewed-by: Jakob Gruber <jgruber@chromium.org> Commit-Queue: Jakob Gruber <jgruber@chromium.org> Cr-Commit-Position: refs/heads/master@{#72088}
diff --git a/src/regexp/regexp-bytecode-generator-inl.h b/src/regexp/regexp-bytecode-generator-inl.h
@@ -14,13 +14,13 @@ namespace v8 {
 namespace internal {
 
 void RegExpBytecodeGenerator::Emit(uint32_t byte, uint32_t twenty_four_bits) {
-  uint32_t word = ((twenty_four_bits << BYTECODE_SHIFT) | byte);
-  DCHECK(pc_ <= buffer_.length());
-  if (pc_ + 3 >= buffer_.length()) {
-    Expand();
-  }
-  *reinterpret_cast<uint32_t*>(buffer_.begin() + pc_) = word;
-  pc_ += 4;
+  DCHECK(is_uint24(twenty_four_bits));
+  Emit32((twenty_four_bits << BYTECODE_SHIFT) | byte);
+}
+
+void RegExpBytecodeGenerator::Emit(uint32_t byte, int32_t twenty_four_bits) {
+  DCHECK(is_int24(twenty_four_bits));
+  Emit32((static_cast<uint32_t>(twenty_four_bits) << BYTECODE_SHIFT) | byte);
 }
 
 void RegExpBytecodeGenerator::Emit16(uint32_t word) {
diff --git a/src/regexp/regexp-bytecode-generator.cc b/src/regexp/regexp-bytecode-generator.cc
@@ -165,8 +165,10 @@ bool RegExpBytecodeGenerator::Succeed() {
 void RegExpBytecodeGenerator::Fail() { Emit(BC_FAIL, 0); }
 
 void RegExpBytecodeGenerator::AdvanceCurrentPosition(int by) {
-  DCHECK_LE(kMinCPOffset, by);
-  DCHECK_GE(kMaxCPOffset, by);
+  // TODO(chromium:1166138): Turn back into DCHECKs once the underlying issue
+  // is fixed.
+  CHECK_LE(kMinCPOffset, by);
+  CHECK_GE(kMaxCPOffset, by);
   advance_current_start_ = pc_;
   advance_current_offset_ = by;
   Emit(BC_ADVANCE_CP, by);
diff --git a/src/regexp/regexp-bytecode-generator.h b/src/regexp/regexp-bytecode-generator.h
@@ -86,6 +86,7 @@ class V8_EXPORT_PRIVATE RegExpBytecodeGenerator : public RegExpMacroAssembler {
   inline void Emit16(uint32_t x);
   inline void Emit8(uint32_t x);
   inline void Emit(uint32_t bc, uint32_t arg);
+  inline void Emit(uint32_t bc, int32_t arg);
   // Bytecode buffer.
   int length();
   void Copy(byte* a);
diff --git a/test/mjsunit/mjsunit.status b/test/mjsunit/mjsunit.status
@@ -73,6 +73,9 @@
   # https://crbug.com/1129854
   'tools/log': ['arch == arm or arch == arm64', SKIP],
 
+  # https://crbug.com/1166138
+  'regress/regress-1166138': SKIP,
+
   ##############################################################################
   # Tests where variants make no sense.
   'd8/enable-tracing': [PASS, NO_VARIANTS],
diff --git a/test/mjsunit/regress/regress-1166138.js b/test/mjsunit/regress/regress-1166138.js
@@ -0,0 +1,7 @@
+// Copyright 2020 the V8 project authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+let badregexp =  "(?:" +  " ".repeat(32768*2)+  ")*";
+reg = RegExp(badregexp);
+reg.test()
