diff --git a/.doozer.json b/.doozer.json index eaea3b02..589dccc3 100644 --- a/.doozer.json +++ b/.doozer.json @@ -1,19 +1,54 @@ { "targets": { - "xenial-amd64": { - "buildenv": "xenial-amd64", + "jessie-i386": { + "buildenv": "jessie-i386", + "builddeps": ["build-essential", "cmake", "clang"], + "buildcmd": ["mkdir cm", "cd cm", "CXX=clang++ cmake ..", "cmake --build .", "ctest --output-on-failure"] + }, + "precise-i386": { + "buildenv": "precise-i386", "builddeps": ["build-essential", "cmake"], - "buildcmd": ["mkdir cm", "cd cm", "cmake ..", "cmake --build .", "ctest"] + "buildcmd": ["mkdir cm", "cd cm", "cmake ..", "cmake --build .", "ctest --output-on-failure"] + }, + "precise-amd64": { + "buildenv": "precise-amd64", + "builddeps": ["build-essential", "cmake"], + "buildcmd": ["mkdir cm", "cd cm", "cmake ..", "cmake --build .", "ctest --output-on-failure"] + }, + "trusty-i386": { + "buildenv": "trusty-i386", + "builddeps": ["build-essential", "cmake", "clang"], + "buildcmd": ["make check CXX=clang++"] + }, + "trusty-amd64": { + "buildenv": "trusty-amd64", + "builddeps": ["build-essential", "cmake", "clang"], + "buildcmd": ["make check CXX=clang++"] }, "xenial-i386": { "buildenv": "xenial-i386", "builddeps": ["build-essential", "cmake"], - "buildcmd": ["mkdir cm", "cd cm", "cmake ..", "cmake --build .", "ctest"] + "buildcmd": ["mkdir cm", "cd cm", "cmake ..", "cmake --build .", "ctest --output-on-failure"] + }, + "xenial-amd64": { + "buildenv": "xenial-amd64", + "builddeps": ["build-essential", "cmake"], + "buildcmd": ["mkdir cm", "cd cm", "cmake ..", "cmake --build .", "ctest --output-on-failure"] + }, + "fedora24-x86_64": { + "buildenv": "fedora24-x86_64", + "builddeps": ["cmake", "make", "clang"], + "buildcmd": ["mkdir cm", "cd cm", "CXX=clang++ cmake ..", "cmake --build .", "ctest --output-on-failure"] + }, + "centos7-x86_64": { + "buildenv": "centos7-x86_64", + "builddeps": ["build-essential", "cmake"], + "buildcmd": ["mkdir cm", "cd cm", "cmake ..", "cmake --build .", "ctest --output-on-failure"] }, "osx": { "buildenv": "osx", - "builddeps": ["build-essential", "cmake"], - "buildcmd": ["mkdir cm", "cd cm", "cmake ..", "cmake --build .", "ctest"] + "builddeps": ["build-essential"], + "buildcmd": ["make check"] } } } diff --git a/.gitignore b/.gitignore index 42a18e82..eedc9641 100644 --- a/.gitignore +++ b/.gitignore @@ -20,3 +20,6 @@ cmake-build-debug test/test-* +.svn + +test/thirdparty/Fuzzer/libFuzzer.a diff --git a/Makefile b/Makefile index fcce453e..e2c120db 100644 --- a/Makefile +++ b/Makefile @@ -10,7 +10,7 @@ all: # clean up clean: - rm -fr json_unit json_benchmarks fuzz fuzz-testing *.dSYM + rm -fr json_unit json_benchmarks fuzz fuzz-testing *.dSYM test/*.dSYM rm -fr benchmarks/files/numbers/*.json $(MAKE) clean -Cdoc $(MAKE) clean -Ctest @@ -49,15 +49,40 @@ doctest: fuzz_testing: rm -fr fuzz-testing mkdir -p fuzz-testing fuzz-testing/testcases fuzz-testing/out - $(MAKE) fuzz CXX=afl-clang++ - mv fuzz fuzz-testing + $(MAKE) parse_afl_fuzzer -C test CXX=afl-clang++ + mv test/parse_afl_fuzzer fuzz-testing/fuzzer find test/data/json_tests -size -5k -name *json | xargs -I{} cp "{}" fuzz-testing/testcases - @echo "Execute: afl-fuzz -i fuzz-testing/testcases -o fuzz-testing/out fuzz-testing/fuzz" + @echo "Execute: afl-fuzz -i fuzz-testing/testcases -o fuzz-testing/out fuzz-testing/fuzzer" -# the fuzzer binary -fuzz: test/src/fuzz.cpp src/json.hpp - $(CXX) -std=c++11 $(CXXFLAGS) $(FLAGS) $(CPPFLAGS) -I src $< $(LDFLAGS) -o $@ +fuzz_testing_cbor: + rm -fr fuzz-testing + mkdir -p fuzz-testing fuzz-testing/testcases fuzz-testing/out + $(MAKE) parse_cbor_fuzzer -C test CXX=afl-clang++ + mv test/parse_cbor_fuzzer fuzz-testing/fuzzer + find test/data -size -5k -name *.cbor | xargs -I{} cp "{}" fuzz-testing/testcases + @echo "Execute: afl-fuzz -i fuzz-testing/testcases -o fuzz-testing/out fuzz-testing/fuzzer" +fuzz_testing_msgpack: + rm -fr fuzz-testing + mkdir -p fuzz-testing fuzz-testing/testcases fuzz-testing/out + $(MAKE) parse_msgpack_fuzzer -C test CXX=afl-clang++ + mv test/parse_msgpack_fuzzer fuzz-testing/fuzzer + find test/data -size -5k -name *.msgpack | xargs -I{} cp "{}" fuzz-testing/testcases + @echo "Execute: afl-fuzz -i fuzz-testing/testcases -o fuzz-testing/out fuzz-testing/fuzzer" + +fuzzing-start: + afl-fuzz -S fuzzer1 -i fuzz-testing/testcases -o fuzz-testing/out fuzz-testing/fuzzer > /dev/null & + afl-fuzz -S fuzzer2 -i fuzz-testing/testcases -o fuzz-testing/out fuzz-testing/fuzzer > /dev/null & + afl-fuzz -S fuzzer3 -i fuzz-testing/testcases -o fuzz-testing/out fuzz-testing/fuzzer > /dev/null & + afl-fuzz -S fuzzer4 -i fuzz-testing/testcases -o fuzz-testing/out fuzz-testing/fuzzer > /dev/null & + afl-fuzz -S fuzzer5 -i fuzz-testing/testcases -o fuzz-testing/out fuzz-testing/fuzzer > /dev/null & + afl-fuzz -S fuzzer6 -i fuzz-testing/testcases -o fuzz-testing/out fuzz-testing/fuzzer > /dev/null & + afl-fuzz -S fuzzer7 -i fuzz-testing/testcases -o fuzz-testing/out fuzz-testing/fuzzer > /dev/null & + afl-fuzz -M fuzzer0 -i fuzz-testing/testcases -o fuzz-testing/out fuzz-testing/fuzzer + +fuzzing-stop: + -killall fuzzer + -killall afl-fuzz ########################################################################## # static analyzer diff --git a/doc/Makefile b/doc/Makefile index 56198923..04a5cc7b 100644 --- a/doc/Makefile +++ b/doc/Makefile @@ -57,6 +57,7 @@ doxygen: create_output create_links $(SED) -i 's@< ObjectType, ArrayType, StringType, BooleanType, NumberIntegerType, NumberFloatType, AllocatorType >@@g' html/*.html $(SED) -i 's@< ObjectType, ArrayType, StringType, BooleanType, NumberIntegerType, NumberUnsignedType, NumberFloatType, AllocatorType >@@g' html/*.html $(SED) -i 's@< ObjectType, ArrayType, StringType, BooleanType, NumberIntegerType, NumberUnsignedType, NumberFloatType, AllocatorType >@@g' html/*.html + $(SED) -i 's@< ObjectType, ArrayType, StringType, BooleanType, NumberIntegerType, NumberUnsignedType, NumberFloatType, AllocatorType >@@g' html/*.html upload: clean doxygen check_output cd html ; ../scripts/git-update-ghpages nlohmann/json diff --git a/src/json.hpp b/src/json.hpp index b5088355..5ad1b4e5 100644 --- a/src/json.hpp +++ b/src/json.hpp @@ -3802,7 +3802,7 @@ class basic_json container `c`, the expression `c.front()` is equivalent to `*c.begin()`. @return In case of a structured type (array or object), a reference to the - first element is returned. In cast of number, string, or boolean values, a + first element is returned. In case of number, string, or boolean values, a reference to the value is returned. @complexity Constant. @@ -3845,7 +3845,7 @@ class basic_json @endcode @return In case of a structured type (array or object), a reference to the - last element is returned. In cast of number, string, or boolean values, a + last element is returned. In case of number, string, or boolean values, a reference to the value is returned. @complexity Constant. @@ -4196,10 +4196,14 @@ class basic_json element is not found or the JSON value is not an object, end() is returned. + @note This method always returns @ref end() when executed on a JSON type + that is not an object. + @param[in] key key value of the element to search for @return Iterator to an element with key equivalent to @a key. If no such - element is found, past-the-end (see end()) iterator is returned. + element is found or the JSON value is not an object, past-the-end (see + @ref end()) iterator is returned. @complexity Logarithmic in the size of the JSON object. @@ -4242,6 +4246,9 @@ class basic_json default `std::map` type, the return value will always be `0` (@a key was not found) or `1` (@a key was found). + @note This method always returns `0` when executed on a JSON type that is + not an object. + @param[in] key key value of the element to count @return Number of elements with key @a key. If the JSON value is not an @@ -4801,9 +4808,6 @@ class basic_json object | `{}` array | `[]` - @note Floating-point numbers are set to `0.0` which will be serialized to - `0`. The vale type remains @ref number_float_t. - @complexity Linear in the size of the JSON value. @liveexample{The example below shows the effect of `clear()` to different @@ -6840,6 +6844,50 @@ class basic_json } } + + /* + @brief checks if given lengths do not exceed the size of a given vector + + To secure the access to the byte vector during CBOR/MessagePack + deserialization, bytes are copied from the vector into buffers. This + function checks if the number of bytes to copy (@a len) does not exceed the + size @s size of the vector. Additionally, an @a offset is given from where + to start reading the bytes. + + This function checks whether reading the bytes is safe; that is, offset is a + valid index in the vector, offset+len + + @param[in] size size of the byte vector + @param[in] len number of bytes to read + @param[in] offset offset where to start reading + + vec: x x x x x X X X X X + ^ ^ ^ + 0 offset len + + @throws out_of_range if `len > v.size()` + */ + static void check_length(const size_t size, const size_t len, const size_t offset) + { + // simple case: requested length is greater than the vector's length + if (len > size or offset > size) + { + throw std::out_of_range("len out of range"); + } + + // second case: adding offset would result in overflow + if ((size > (std::numeric_limits::max() - offset))) + { + throw std::out_of_range("len+offset out of range"); + } + + // last case: reading past the end of the vector + if (len + offset > size) + { + throw std::out_of_range("len+offset out of range"); + } + } + /*! @brief create a JSON value from a given MessagePack vector @@ -6856,6 +6904,9 @@ class basic_json */ static basic_json from_msgpack_internal(const std::vector& v, size_t& idx) { + // make sure reading 1 byte is safe + check_length(v.size(), 1, idx); + // store and increment index const size_t current_idx = idx++; @@ -6891,6 +6942,7 @@ class basic_json const size_t len = v[current_idx] & 0x1f; const size_t offset = current_idx + 1; idx += len; // skip content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } } @@ -6920,6 +6972,7 @@ class basic_json case 0xca: // float 32 { // copy bytes in reverse order into the double variable + check_length(v.size(), sizeof(float), 1); float res; for (size_t byte = 0; byte < sizeof(float); ++byte) { @@ -6932,6 +6985,7 @@ class basic_json case 0xcb: // float 64 { // copy bytes in reverse order into the double variable + check_length(v.size(), sizeof(double), 1); double res; for (size_t byte = 0; byte < sizeof(double); ++byte) { @@ -6994,6 +7048,7 @@ class basic_json const auto len = static_cast(get_from_vector(v, current_idx)); const size_t offset = current_idx + 2; idx += len + 1; // skip size byte + content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7002,6 +7057,7 @@ class basic_json const auto len = static_cast(get_from_vector(v, current_idx)); const size_t offset = current_idx + 3; idx += len + 2; // skip 2 size bytes + content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7010,6 +7066,7 @@ class basic_json const auto len = static_cast(get_from_vector(v, current_idx)); const size_t offset = current_idx + 5; idx += len + 4; // skip 4 size bytes + content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7087,6 +7144,9 @@ class basic_json */ static basic_json from_cbor_internal(const std::vector& v, size_t& idx) { + // make sure reading 1 byte is safe + check_length(v.size(), 1, idx); + // store and increment index const size_t current_idx = idx++; @@ -7228,6 +7288,7 @@ class basic_json const auto len = static_cast(v[current_idx] - 0x60); const size_t offset = current_idx + 1; idx += len; // skip content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7236,6 +7297,7 @@ class basic_json const auto len = static_cast(get_from_vector(v, current_idx)); const size_t offset = current_idx + 2; idx += len + 1; // skip size byte + content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7244,6 +7306,7 @@ class basic_json const auto len = static_cast(get_from_vector(v, current_idx)); const size_t offset = current_idx + 3; idx += len + 2; // skip 2 size bytes + content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7252,6 +7315,7 @@ class basic_json const auto len = static_cast(get_from_vector(v, current_idx)); const size_t offset = current_idx + 5; idx += len + 4; // skip 4 size bytes + content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7260,6 +7324,7 @@ class basic_json const auto len = static_cast(get_from_vector(v, current_idx)); const size_t offset = current_idx + 9; idx += len + 8; // skip 8 size bytes + content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7489,6 +7554,7 @@ class basic_json case 0xf9: // Half-Precision Float (two-byte IEEE 754) { + check_length(v.size(), 2, 1); idx += 2; // skip two content bytes // code from RFC 7049, Appendix D, Figure 3: @@ -7520,6 +7586,7 @@ class basic_json case 0xfa: // Single-Precision Float (four-byte IEEE 754) { // copy bytes in reverse order into the float variable + check_length(v.size(), sizeof(float), 1); float res; for (size_t byte = 0; byte < sizeof(float); ++byte) { @@ -7531,6 +7598,7 @@ class basic_json case 0xfb: // Double-Precision Float (eight-byte IEEE 754) { + check_length(v.size(), sizeof(double), 1); // copy bytes in reverse order into the double variable double res; for (size_t byte = 0; byte < sizeof(double); ++byte) @@ -9004,7 +9072,7 @@ class basic_json // immediately abort if stream is erroneous if (s.fail()) { - throw std::invalid_argument("stream error: " + std::string(strerror(errno))); + throw std::invalid_argument("stream error"); } // fill buffer diff --git a/src/json.hpp.re2c b/src/json.hpp.re2c index e8a52a0f..be1ebf55 100644 --- a/src/json.hpp.re2c +++ b/src/json.hpp.re2c @@ -3802,7 +3802,7 @@ class basic_json container `c`, the expression `c.front()` is equivalent to `*c.begin()`. @return In case of a structured type (array or object), a reference to the - first element is returned. In cast of number, string, or boolean values, a + first element is returned. In case of number, string, or boolean values, a reference to the value is returned. @complexity Constant. @@ -3845,7 +3845,7 @@ class basic_json @endcode @return In case of a structured type (array or object), a reference to the - last element is returned. In cast of number, string, or boolean values, a + last element is returned. In case of number, string, or boolean values, a reference to the value is returned. @complexity Constant. @@ -4196,10 +4196,14 @@ class basic_json element is not found or the JSON value is not an object, end() is returned. + @note This method always returns @ref end() when executed on a JSON type + that is not an object. + @param[in] key key value of the element to search for @return Iterator to an element with key equivalent to @a key. If no such - element is found, past-the-end (see end()) iterator is returned. + element is found or the JSON value is not an object, past-the-end (see + @ref end()) iterator is returned. @complexity Logarithmic in the size of the JSON object. @@ -4242,6 +4246,9 @@ class basic_json default `std::map` type, the return value will always be `0` (@a key was not found) or `1` (@a key was found). + @note This method always returns `0` when executed on a JSON type that is + not an object. + @param[in] key key value of the element to count @return Number of elements with key @a key. If the JSON value is not an @@ -4801,9 +4808,6 @@ class basic_json object | `{}` array | `[]` - @note Floating-point numbers are set to `0.0` which will be serialized to - `0`. The vale type remains @ref number_float_t. - @complexity Linear in the size of the JSON value. @liveexample{The example below shows the effect of `clear()` to different @@ -6840,6 +6844,50 @@ class basic_json } } + + /* + @brief checks if given lengths do not exceed the size of a given vector + + To secure the access to the byte vector during CBOR/MessagePack + deserialization, bytes are copied from the vector into buffers. This + function checks if the number of bytes to copy (@a len) does not exceed the + size @s size of the vector. Additionally, an @a offset is given from where + to start reading the bytes. + + This function checks whether reading the bytes is safe; that is, offset is a + valid index in the vector, offset+len + + @param[in] size size of the byte vector + @param[in] len number of bytes to read + @param[in] offset offset where to start reading + + vec: x x x x x X X X X X + ^ ^ ^ + 0 offset len + + @throws out_of_range if `len > v.size()` + */ + static void check_length(const size_t size, const size_t len, const size_t offset) + { + // simple case: requested length is greater than the vector's length + if (len > size or offset > size) + { + throw std::out_of_range("len out of range"); + } + + // second case: adding offset would result in overflow + if ((size > (std::numeric_limits::max() - offset))) + { + throw std::out_of_range("len+offset out of range"); + } + + // last case: reading past the end of the vector + if (len + offset > size) + { + throw std::out_of_range("len+offset out of range"); + } + } + /*! @brief create a JSON value from a given MessagePack vector @@ -6856,6 +6904,9 @@ class basic_json */ static basic_json from_msgpack_internal(const std::vector& v, size_t& idx) { + // make sure reading 1 byte is safe + check_length(v.size(), 1, idx); + // store and increment index const size_t current_idx = idx++; @@ -6891,6 +6942,7 @@ class basic_json const size_t len = v[current_idx] & 0x1f; const size_t offset = current_idx + 1; idx += len; // skip content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } } @@ -6920,6 +6972,7 @@ class basic_json case 0xca: // float 32 { // copy bytes in reverse order into the double variable + check_length(v.size(), sizeof(float), 1); float res; for (size_t byte = 0; byte < sizeof(float); ++byte) { @@ -6932,6 +6985,7 @@ class basic_json case 0xcb: // float 64 { // copy bytes in reverse order into the double variable + check_length(v.size(), sizeof(double), 1); double res; for (size_t byte = 0; byte < sizeof(double); ++byte) { @@ -6994,6 +7048,7 @@ class basic_json const auto len = static_cast(get_from_vector(v, current_idx)); const size_t offset = current_idx + 2; idx += len + 1; // skip size byte + content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7002,6 +7057,7 @@ class basic_json const auto len = static_cast(get_from_vector(v, current_idx)); const size_t offset = current_idx + 3; idx += len + 2; // skip 2 size bytes + content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7010,6 +7066,7 @@ class basic_json const auto len = static_cast(get_from_vector(v, current_idx)); const size_t offset = current_idx + 5; idx += len + 4; // skip 4 size bytes + content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7087,6 +7144,9 @@ class basic_json */ static basic_json from_cbor_internal(const std::vector& v, size_t& idx) { + // make sure reading 1 byte is safe + check_length(v.size(), 1, idx); + // store and increment index const size_t current_idx = idx++; @@ -7228,6 +7288,7 @@ class basic_json const auto len = static_cast(v[current_idx] - 0x60); const size_t offset = current_idx + 1; idx += len; // skip content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7236,6 +7297,7 @@ class basic_json const auto len = static_cast(get_from_vector(v, current_idx)); const size_t offset = current_idx + 2; idx += len + 1; // skip size byte + content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7244,6 +7306,7 @@ class basic_json const auto len = static_cast(get_from_vector(v, current_idx)); const size_t offset = current_idx + 3; idx += len + 2; // skip 2 size bytes + content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7252,6 +7315,7 @@ class basic_json const auto len = static_cast(get_from_vector(v, current_idx)); const size_t offset = current_idx + 5; idx += len + 4; // skip 4 size bytes + content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7260,6 +7324,7 @@ class basic_json const auto len = static_cast(get_from_vector(v, current_idx)); const size_t offset = current_idx + 9; idx += len + 8; // skip 8 size bytes + content bytes + check_length(v.size(), len, offset); return std::string(reinterpret_cast(v.data()) + offset, len); } @@ -7489,6 +7554,7 @@ class basic_json case 0xf9: // Half-Precision Float (two-byte IEEE 754) { + check_length(v.size(), 2, 1); idx += 2; // skip two content bytes // code from RFC 7049, Appendix D, Figure 3: @@ -7520,6 +7586,7 @@ class basic_json case 0xfa: // Single-Precision Float (four-byte IEEE 754) { // copy bytes in reverse order into the float variable + check_length(v.size(), sizeof(float), 1); float res; for (size_t byte = 0; byte < sizeof(float); ++byte) { @@ -7531,6 +7598,7 @@ class basic_json case 0xfb: // Double-Precision Float (eight-byte IEEE 754) { + check_length(v.size(), sizeof(double), 1); // copy bytes in reverse order into the double variable double res; for (size_t byte = 0; byte < sizeof(double); ++byte) @@ -9004,7 +9072,7 @@ class basic_json // immediately abort if stream is erroneous if (s.fail()) { - throw std::invalid_argument("stream error: " + std::string(strerror(errno))); + throw std::invalid_argument("stream error"); } // fill buffer diff --git a/test/CMakeLists.txt b/test/CMakeLists.txt index f57ddf71..47785aee 100644 --- a/test/CMakeLists.txt +++ b/test/CMakeLists.txt @@ -1,7 +1,7 @@ # The unit test executable. set(JSON_UNITTEST_TARGET_NAME "json_unit") add_executable(${JSON_UNITTEST_TARGET_NAME} - "src/catch.hpp" + "thirdparty/catch/catch.hpp" "src/unit.cpp" "src/unit-algorithms.cpp" "src/unit-allocator.cpp" @@ -44,7 +44,7 @@ set_target_properties(${JSON_UNITTEST_TARGET_NAME} PROPERTIES COMPILE_OPTIONS "$<$:/EHsc;$<$:/Od>>" ) -target_include_directories(${JSON_UNITTEST_TARGET_NAME} PRIVATE "src") +target_include_directories(${JSON_UNITTEST_TARGET_NAME} PRIVATE "src" "thirdparty/catch") target_link_libraries(${JSON_UNITTEST_TARGET_NAME} ${JSON_TARGET_NAME}) add_test(NAME "${JSON_UNITTEST_TARGET_NAME}_default" diff --git a/test/Makefile b/test/Makefile index 488cec89..556ab0d7 100644 --- a/test/Makefile +++ b/test/Makefile @@ -4,7 +4,7 @@ # additional flags CXXFLAGS += -std=c++11 -Wall -Wextra -pedantic -Weffc++ -Wcast-align -Wcast-qual -Wno-ctor-dtor-privacy -Wdisabled-optimization -Wformat=2 -Winit-self -Wmissing-declarations -Wmissing-include-dirs -Wold-style-cast -Woverloaded-virtual -Wredundant-decls -Wshadow -Wsign-conversion -Wsign-promo -Wstrict-overflow=5 -Wswitch -Wundef -Wno-unused -Wnon-virtual-dtor -Wreorder -Wdeprecated -Wno-float-equal -CPPFLAGS += -I ../src -I . +CPPFLAGS += -I ../src -I . -I thirdparty/catch SOURCES = src/unit.cpp \ src/unit-algorithms.cpp \ @@ -57,11 +57,11 @@ clean: # single test file ############################################################################## -json_unit: $(OBJECTS) ../src/json.hpp src/catch.hpp +json_unit: $(OBJECTS) ../src/json.hpp thirdparty/catch/catch.hpp @echo "[CXXLD] $@" @$(CXX) $(CXXFLAGS) $(LDFLAGS) $(OBJECTS) -o $@ -%.o: %.cpp ../src/json.hpp src/catch.hpp +%.o: %.cpp ../src/json.hpp thirdparty/catch/catch.hpp @echo "[CXX] $@" @$(CXX) $(CXXFLAGS) $(CPPFLAGS) -c $< -o $@ @@ -70,7 +70,7 @@ json_unit: $(OBJECTS) ../src/json.hpp src/catch.hpp # individual test cases ############################################################################## -test-%: src/unit-%.cpp ../src/json.hpp src/catch.hpp +test-%: src/unit-%.cpp ../src/json.hpp thirdparty/catch/catch.hpp @echo "[CXXLD] $@" @$(CXX) $(CXXFLAGS) $(CPPFLAGS) $(LDFLAGS) -DCATCH_CONFIG_MAIN $< -o $@ @@ -78,3 +78,17 @@ TEST_PATTERN = "*" TEST_PREFIX = "" check: $(TESTCASES) @cd .. ; for testcase in $(TESTCASES); do echo "Executing $$testcase..."; $(TEST_PREFIX)test/$$testcase $(TEST_PATTERN) || exit 1; done + + +############################################################################## +# fuzzer +############################################################################## + +parse_afl_fuzzer: + $(CXX) $(CXXFLAGS) $(CPPFLAGS) src/fuzzer-driver_afl.cpp src/fuzzer-parse_json.cpp -o $@ + +parse_cbor_fuzzer: + $(CXX) $(CXXFLAGS) $(CPPFLAGS) src/fuzzer-driver_afl.cpp src/fuzzer-parse_cbor.cpp -o $@ + +parse_msgpack_fuzzer: + $(CXX) $(CXXFLAGS) $(CPPFLAGS) src/fuzzer-driver_afl.cpp src/fuzzer-parse_msgpack.cpp -o $@ diff --git a/test/data/cbor_regression/test01 b/test/data/cbor_regression/test01 new file mode 100644 index 00000000..e30ed176 --- /dev/null +++ b/test/data/cbor_regression/test01 @@ -0,0 +1 @@ +{ÿÿÿÿÿÿÿÿÿÿÿÿÿÿ \ No newline at end of file diff --git a/test/data/cbor_regression/test02 b/test/data/cbor_regression/test02 new file mode 100644 index 00000000..8de7c9e5 Binary files /dev/null and b/test/data/cbor_regression/test02 differ diff --git a/test/data/cbor_regression/test03 b/test/data/cbor_regression/test03 new file mode 100644 index 00000000..d2ae80c7 Binary files /dev/null and b/test/data/cbor_regression/test03 differ diff --git a/test/data/cbor_regression/test04 b/test/data/cbor_regression/test04 new file mode 100644 index 00000000..d1ef19b5 Binary files /dev/null and b/test/data/cbor_regression/test04 differ diff --git a/test/data/cbor_regression/test05 b/test/data/cbor_regression/test05 new file mode 100644 index 00000000..42b437e0 Binary files /dev/null and b/test/data/cbor_regression/test05 differ diff --git a/test/data/cbor_regression/test06 b/test/data/cbor_regression/test06 new file mode 100644 index 00000000..82f3520d Binary files /dev/null and b/test/data/cbor_regression/test06 differ diff --git a/test/data/cbor_regression/test07 b/test/data/cbor_regression/test07 new file mode 100644 index 00000000..679a1663 Binary files /dev/null and b/test/data/cbor_regression/test07 differ diff --git a/test/data/cbor_regression/test08 b/test/data/cbor_regression/test08 new file mode 100644 index 00000000..bb32f698 Binary files /dev/null and b/test/data/cbor_regression/test08 differ diff --git a/test/data/cbor_regression/test09 b/test/data/cbor_regression/test09 new file mode 100644 index 00000000..e9485810 Binary files /dev/null and b/test/data/cbor_regression/test09 differ diff --git a/test/data/cbor_regression/test10 b/test/data/cbor_regression/test10 new file mode 100644 index 00000000..92079a80 --- /dev/null +++ b/test/data/cbor_regression/test10 @@ -0,0 +1 @@ +”{ÿÿÿÿÿÿÿÿÿ’ÿÿÿÿÿÿÿÿúúúúúúúúúúúúetú \ No newline at end of file diff --git a/test/data/cbor_regression/test11 b/test/data/cbor_regression/test11 new file mode 100644 index 00000000..0e3f141f Binary files /dev/null and b/test/data/cbor_regression/test11 differ diff --git a/test/data/cbor_regression/test12 b/test/data/cbor_regression/test12 new file mode 100644 index 00000000..0ce68701 Binary files /dev/null and b/test/data/cbor_regression/test12 differ diff --git a/test/data/cbor_regression/test13 b/test/data/cbor_regression/test13 new file mode 100644 index 00000000..35f3cb59 Binary files /dev/null and b/test/data/cbor_regression/test13 differ diff --git a/test/data/cbor_regression/test14 b/test/data/cbor_regression/test14 new file mode 100644 index 00000000..64812d4f Binary files /dev/null and b/test/data/cbor_regression/test14 differ diff --git a/test/data/cbor_regression/test15 b/test/data/cbor_regression/test15 new file mode 100644 index 00000000..5c537f8d Binary files /dev/null and b/test/data/cbor_regression/test15 differ diff --git a/test/data/cbor_regression/test16 b/test/data/cbor_regression/test16 new file mode 100644 index 00000000..e342979e Binary files /dev/null and b/test/data/cbor_regression/test16 differ diff --git a/test/data/cbor_regression/test17 b/test/data/cbor_regression/test17 new file mode 100644 index 00000000..0bde9e5d Binary files /dev/null and b/test/data/cbor_regression/test17 differ diff --git a/test/data/cbor_regression/test18 b/test/data/cbor_regression/test18 new file mode 100644 index 00000000..0a004d9e Binary files /dev/null and b/test/data/cbor_regression/test18 differ diff --git a/test/data/cbor_regression/test19 b/test/data/cbor_regression/test19 new file mode 100644 index 00000000..6ff4823a Binary files /dev/null and b/test/data/cbor_regression/test19 differ diff --git a/test/data/cbor_regression/test20 b/test/data/cbor_regression/test20 new file mode 100644 index 00000000..d32904ab Binary files /dev/null and b/test/data/cbor_regression/test20 differ diff --git a/test/data/cbor_regression/test21 b/test/data/cbor_regression/test21 new file mode 100644 index 00000000..156afe0f Binary files /dev/null and b/test/data/cbor_regression/test21 differ diff --git a/test/src/fuzz.cpp b/test/src/fuzz.cpp deleted file mode 100644 index ef403ea8..00000000 --- a/test/src/fuzz.cpp +++ /dev/null @@ -1,34 +0,0 @@ -/* - __ _____ _____ _____ - __| | __| | | | JSON for Modern C++ (fuzz test support) -| | |__ | | | | | | version 2.0.9 -|_____|_____|_____|_|___| https://github.com/nlohmann/json - -Run "make fuzz_testing" and follow the instructions. - -Licensed under the MIT License . -*/ - -#include - -using json = nlohmann::json; - -int main() -{ -#ifdef __AFL_HAVE_MANUAL_CONTROL - while (__AFL_LOOP(1000)) - { -#endif - try - { - json j(std::cin); - std::cout << j << std::endl; - } - catch (std::invalid_argument& e) - { - std::cout << "Invalid argument in parsing" << e.what() << '\n'; - } -#ifdef __AFL_HAVE_MANUAL_CONTROL - } -#endif -} diff --git a/test/src/fuzzer-driver_afl.cpp b/test/src/fuzzer-driver_afl.cpp new file mode 100644 index 00000000..eddeae41 --- /dev/null +++ b/test/src/fuzzer-driver_afl.cpp @@ -0,0 +1,38 @@ +/* + __ _____ _____ _____ + __| | __| | | | JSON for Modern C++ (fuzz test support) +| | |__ | | | | | | version 2.0.9 +|_____|_____|_____|_|___| https://github.com/nlohmann/json + +This file implements a driver for American Fuzzy Lop (afl-fuzz). It relies on +an implementation of the `LLVMFuzzerTestOneInput` function which processes a +passed byte array. + +Licensed under the MIT License . +*/ + +#include // for vector +#include // for uint8_t +#include // for cin + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t* data, size_t size); + +int main() +{ +#ifdef __AFL_HAVE_MANUAL_CONTROL + while (__AFL_LOOP(1000)) + { +#endif + // copy stdin to byte vector + std::vector vec; + char c; + while (std::cin.get(c)) + { + vec.push_back(static_cast(c)); + } + + LLVMFuzzerTestOneInput(vec.data(), vec.size()); +#ifdef __AFL_HAVE_MANUAL_CONTROL + } +#endif +} diff --git a/test/src/fuzzer-parse_cbor.cpp b/test/src/fuzzer-parse_cbor.cpp new file mode 100644 index 00000000..42ce679c --- /dev/null +++ b/test/src/fuzzer-parse_cbor.cpp @@ -0,0 +1,68 @@ +/* + __ _____ _____ _____ + __| | __| | | | JSON for Modern C++ (fuzz test support) +| | |__ | | | | | | version 2.0.9 +|_____|_____|_____|_|___| https://github.com/nlohmann/json + +This file implements a parser test suitable for fuzz testing. Given a byte +array data, it performs the following steps: + +- j1 = from_cbor(data) +- vec = to_cbor(j1) +- j2 = from_cbor(vec) +- assert(j1 == j2) + +The provided function `LLVMFuzzerTestOneInput` can be used in different fuzzer +drivers. + +Licensed under the MIT License . +*/ + +#include +#include +#include + +using json = nlohmann::json; + +// see http://llvm.org/docs/LibFuzzer.html +extern "C" int LLVMFuzzerTestOneInput(const uint8_t* data, size_t size) +{ + try + { + // step 1: parse input + std::vector vec1(data, data + size); + json j1 = json::from_cbor(vec1); + + try + { + // step 2: round trip + std::vector vec2 = json::to_cbor(j1); + + // parse serialization + json j2 = json::from_cbor(vec2); + + // deserializations must match + assert(j1 == j2); + } + catch (const std::invalid_argument&) + { + // parsing a CBOR serialization must not fail + assert(false); + } + } + catch (const std::invalid_argument&) + { + // parse errors are ok, because input may be random bytes + } + catch (const std::out_of_range&) + { + // parse errors are ok, because input may be random bytes + } + catch (const std::domain_error&) + { + // parse errors are ok, because input may be random bytes + } + + // return 0 - non-zero return values are reserved for future use + return 0; +} diff --git a/test/src/fuzzer-parse_json.cpp b/test/src/fuzzer-parse_json.cpp new file mode 100644 index 00000000..51ac440d --- /dev/null +++ b/test/src/fuzzer-parse_json.cpp @@ -0,0 +1,65 @@ +/* + __ _____ _____ _____ + __| | __| | | | JSON for Modern C++ (fuzz test support) +| | |__ | | | | | | version 2.0.9 +|_____|_____|_____|_|___| https://github.com/nlohmann/json + +This file implements a parser test suitable for fuzz testing. Given a byte +array data, it performs the following steps: + +- j1 = parse(data) +- s1 = serialize(j1) +- j2 = parse(s1) +- s2 = serialize(j2) +- assert(s1 == s2) + +The provided function `LLVMFuzzerTestOneInput` can be used in different fuzzer +drivers. + +Licensed under the MIT License . +*/ + +#include +#include +#include + +using json = nlohmann::json; + +// see http://llvm.org/docs/LibFuzzer.html +extern "C" int LLVMFuzzerTestOneInput(const uint8_t* data, size_t size) +{ + try + { + // step 1: parse input + json j1 = json::parse(data, data + size); + + try + { + // step 2: round trip + + // first serialization + std::string s1 = j1.dump(); + + // parse serialization + json j2 = json::parse(s1); + + // second serialization + std::string s2 = j2.dump(); + + // serializations must match + assert(s1 == s2); + } + catch (const std::invalid_argument&) + { + // parsing a JSON serialization must not fail + assert(false); + } + } + catch (const std::invalid_argument&) + { + // parse errors are ok, because input may be random bytes + } + + // return 0 - non-zero return values are reserved for future use + return 0; +} diff --git a/test/src/fuzzer-parse_msgpack.cpp b/test/src/fuzzer-parse_msgpack.cpp new file mode 100644 index 00000000..992697c2 --- /dev/null +++ b/test/src/fuzzer-parse_msgpack.cpp @@ -0,0 +1,68 @@ +/* + __ _____ _____ _____ + __| | __| | | | JSON for Modern C++ (fuzz test support) +| | |__ | | | | | | version 2.0.9 +|_____|_____|_____|_|___| https://github.com/nlohmann/json + +This file implements a parser test suitable for fuzz testing. Given a byte +array data, it performs the following steps: + +- j1 = from_msgpack(data) +- vec = to_msgpack(j1) +- j2 = from_msgpack(vec) +- assert(j1 == j2) + +The provided function `LLVMFuzzerTestOneInput` can be used in different fuzzer +drivers. + +Licensed under the MIT License . +*/ + +#include +#include +#include + +using json = nlohmann::json; + +// see http://llvm.org/docs/LibFuzzer.html +extern "C" int LLVMFuzzerTestOneInput(const uint8_t* data, size_t size) +{ + try + { + // step 1: parse input + std::vector vec1(data, data + size); + json j1 = json::from_msgpack(vec1); + + try + { + // step 2: round trip + std::vector vec2 = json::to_msgpack(j1); + + // parse serialization + json j2 = json::from_msgpack(vec2); + + // deserializations must match + assert(j1 == j2); + } + catch (const std::invalid_argument&) + { + // parsing a MessagePack serialization must not fail + assert(false); + } + } + catch (const std::invalid_argument&) + { + // parse errors are ok, because input may be random bytes + } + catch (const std::out_of_range&) + { + // parse errors are ok, because input may be random bytes + } + catch (const std::domain_error&) + { + // parse errors are ok, because input may be random bytes + } + + // return 0 - non-zero return values are reserved for future use + return 0; +} diff --git a/test/src/unit-cbor.cpp b/test/src/unit-cbor.cpp index 14944cff..9c41dc2f 100644 --- a/test/src/unit-cbor.cpp +++ b/test/src/unit-cbor.cpp @@ -1186,6 +1186,84 @@ TEST_CASE("single CBOR roundtrip") } } +TEST_CASE("CBOR regressions") +{ + SECTION("fuzz test results") + { + /* + The following test cases were found during a two-day session with + AFL-Fuzz. As a result, empty byte vectors and excessive lengths are + detected. + */ + for (std::string filename : + { + "test/data/cbor_regression/test01", + "test/data/cbor_regression/test02", + "test/data/cbor_regression/test03", + "test/data/cbor_regression/test04", + "test/data/cbor_regression/test05", + "test/data/cbor_regression/test06", + "test/data/cbor_regression/test07", + "test/data/cbor_regression/test08", + "test/data/cbor_regression/test09", + "test/data/cbor_regression/test10", + "test/data/cbor_regression/test11", + "test/data/cbor_regression/test12", + "test/data/cbor_regression/test13", + "test/data/cbor_regression/test14", + "test/data/cbor_regression/test15", + "test/data/cbor_regression/test16", + "test/data/cbor_regression/test17", + "test/data/cbor_regression/test18", + "test/data/cbor_regression/test19", + "test/data/cbor_regression/test20", + "test/data/cbor_regression/test21" + }) + { + CAPTURE(filename); + + try + { + // parse CBOR file + std::ifstream f_cbor(filename, std::ios::binary); + std::vector vec1( + (std::istreambuf_iterator(f_cbor)), + std::istreambuf_iterator()); + json j1 = json::from_cbor(vec1); + + try + { + // step 2: round trip + std::vector vec2 = json::to_cbor(j1); + + // parse serialization + json j2 = json::from_cbor(vec2); + + // deserializations must match + CHECK(j1 == j2); + } + catch (const std::invalid_argument&) + { + // parsing a CBOR serialization must not fail + CHECK(false); + } + } + catch (const std::invalid_argument&) + { + // parse errors are ok, because input may be random bytes + } + catch (const std::out_of_range&) + { + // parse errors are ok, because input may be random bytes + } + catch (const std::domain_error&) + { + // parse errors are ok, because input may be random bytes + } + } + } +} + TEST_CASE("CBOR roundtrips", "[hide]") { SECTION("input from flynn") diff --git a/test/src/unit-regression.cpp b/test/src/unit-regression.cpp index 1e720ddb..421a386c 100644 --- a/test/src/unit-regression.cpp +++ b/test/src/unit-regression.cpp @@ -540,4 +540,74 @@ TEST_CASE("regression tests") CHECK(j.is_number_float()); CHECK(j.dump() == "1.66020696663386e+20"); } + + SECTION("issue #405 - Heap-buffer-overflow (OSS-Fuzz issue 342)") + { + // original test case + std::vector vec {0x65, 0xf5, 0x0a, 0x48, 0x21}; + CHECK_THROWS_AS(json::from_cbor(vec), std::out_of_range); + } + + SECTION("issue #407 - Heap-buffer-overflow (OSS-Fuzz issue 343)") + { + // original test case: incomplete float64 + std::vector vec1 {0xcb, 0x8f, 0x0a}; + CHECK_THROWS_AS(json::from_msgpack(vec1), std::out_of_range); + + // related test case: incomplete float32 + std::vector vec2 {0xca, 0x8f, 0x0a}; + CHECK_THROWS_AS(json::from_msgpack(vec2), std::out_of_range); + + // related test case: incomplete Half-Precision Float (CBOR) + std::vector vec3 {0xf9, 0x8f}; + CHECK_THROWS_AS(json::from_cbor(vec3), std::out_of_range); + + // related test case: incomplete Single-Precision Float (CBOR) + std::vector vec4 {0xfa, 0x8f, 0x0a}; + CHECK_THROWS_AS(json::from_cbor(vec4), std::out_of_range); + + // related test case: incomplete Double-Precision Float (CBOR) + std::vector vec5 {0xfb, 0x8f, 0x0a}; + CHECK_THROWS_AS(json::from_cbor(vec5), std::out_of_range); + } + + SECTION("issue #408 - Heap-buffer-overflow (OSS-Fuzz issue 344)") + { + // original test case + std::vector vec1 {0x87}; + CHECK_THROWS_AS(json::from_msgpack(vec1), std::out_of_range); + + // more test cases for MessagePack + for (uint8_t b : + { + 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, // fixmap + 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, // fixarray + 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, // fixstr + 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf + }) + { + std::vector vec(1, b); + CHECK_THROWS_AS(json::from_msgpack(vec), std::out_of_range); + } + + // more test cases for CBOR + for (uint8_t b : + { + 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, + 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, // UTF-8 string + 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, + 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, // array + 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, + 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7 // map + }) + { + std::vector vec(1, b); + CHECK_THROWS_AS(json::from_cbor(vec), std::out_of_range); + } + + // special case: empty input + std::vector vec2; + CHECK_THROWS_AS(json::from_cbor(vec2), std::out_of_range); + CHECK_THROWS_AS(json::from_msgpack(vec2), std::out_of_range); + } } diff --git a/test/thirdparty/Fuzzer/CMakeLists.txt b/test/thirdparty/Fuzzer/CMakeLists.txt new file mode 100644 index 00000000..70bd017b --- /dev/null +++ b/test/thirdparty/Fuzzer/CMakeLists.txt @@ -0,0 +1,45 @@ +set(LIBFUZZER_FLAGS_BASE "${CMAKE_CXX_FLAGS}") +# Disable the coverage and sanitizer instrumentation for the fuzzer itself. +set(CMAKE_CXX_FLAGS "${LIBFUZZER_FLAGS_BASE} -fno-sanitize-coverage=trace-pc-guard,edge,trace-cmp,indirect-calls,8bit-counters -Werror") +if( LLVM_USE_SANITIZE_COVERAGE ) + if(NOT "${LLVM_USE_SANITIZER}" STREQUAL "Address") + message(FATAL_ERROR + "LibFuzzer and its tests require LLVM_USE_SANITIZER=Address and " + "LLVM_USE_SANITIZE_COVERAGE=YES to be set." + ) + endif() + add_library(LLVMFuzzerNoMainObjects OBJECT + FuzzerCrossOver.cpp + FuzzerDriver.cpp + FuzzerExtFunctionsDlsym.cpp + FuzzerExtFunctionsWeak.cpp + FuzzerExtFunctionsWeakAlias.cpp + FuzzerIO.cpp + FuzzerIOPosix.cpp + FuzzerIOWindows.cpp + FuzzerLoop.cpp + FuzzerMerge.cpp + FuzzerMutate.cpp + FuzzerSHA1.cpp + FuzzerTracePC.cpp + FuzzerTraceState.cpp + FuzzerUtil.cpp + FuzzerUtilDarwin.cpp + FuzzerUtilLinux.cpp + FuzzerUtilPosix.cpp + FuzzerUtilWindows.cpp + ) + add_library(LLVMFuzzerNoMain STATIC + $ + ) + target_link_libraries(LLVMFuzzerNoMain ${PTHREAD_LIB}) + add_library(LLVMFuzzer STATIC + FuzzerMain.cpp + $ + ) + target_link_libraries(LLVMFuzzer ${PTHREAD_LIB}) + + if( LLVM_INCLUDE_TESTS ) + add_subdirectory(test) + endif() +endif() diff --git a/test/thirdparty/Fuzzer/FuzzerCorpus.h b/test/thirdparty/Fuzzer/FuzzerCorpus.h new file mode 100644 index 00000000..663c5854 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerCorpus.h @@ -0,0 +1,217 @@ +//===- FuzzerCorpus.h - Internal header for the Fuzzer ----------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// fuzzer::InputCorpus +//===----------------------------------------------------------------------===// + +#ifndef LLVM_FUZZER_CORPUS +#define LLVM_FUZZER_CORPUS + +#include "FuzzerDefs.h" +#include "FuzzerIO.h" +#include "FuzzerRandom.h" +#include "FuzzerSHA1.h" +#include "FuzzerTracePC.h" +#include +#include +#include + +namespace fuzzer { + +struct InputInfo { + Unit U; // The actual input data. + uint8_t Sha1[kSHA1NumBytes]; // Checksum. + // Number of features that this input has and no smaller input has. + size_t NumFeatures = 0; + size_t Tmp = 0; // Used by ValidateFeatureSet. + // Stats. + size_t NumExecutedMutations = 0; + size_t NumSuccessfullMutations = 0; + bool MayDeleteFile = false; +}; + +class InputCorpus { + public: + static const size_t kFeatureSetSize = 1 << 16; + InputCorpus(const std::string &OutputCorpus) : OutputCorpus(OutputCorpus) { + memset(InputSizesPerFeature, 0, sizeof(InputSizesPerFeature)); + memset(SmallestElementPerFeature, 0, sizeof(SmallestElementPerFeature)); + } + ~InputCorpus() { + for (auto II : Inputs) + delete II; + } + size_t size() const { return Inputs.size(); } + size_t SizeInBytes() const { + size_t Res = 0; + for (auto II : Inputs) + Res += II->U.size(); + return Res; + } + size_t NumActiveUnits() const { + size_t Res = 0; + for (auto II : Inputs) + Res += !II->U.empty(); + return Res; + } + bool empty() const { return Inputs.empty(); } + const Unit &operator[] (size_t Idx) const { return Inputs[Idx]->U; } + void AddToCorpus(const Unit &U, size_t NumFeatures, bool MayDeleteFile = false) { + assert(!U.empty()); + uint8_t Hash[kSHA1NumBytes]; + if (FeatureDebug) + Printf("ADD_TO_CORPUS %zd NF %zd\n", Inputs.size(), NumFeatures); + ComputeSHA1(U.data(), U.size(), Hash); + Hashes.insert(Sha1ToString(Hash)); + Inputs.push_back(new InputInfo()); + InputInfo &II = *Inputs.back(); + II.U = U; + II.NumFeatures = NumFeatures; + II.MayDeleteFile = MayDeleteFile; + memcpy(II.Sha1, Hash, kSHA1NumBytes); + UpdateCorpusDistribution(); + ValidateFeatureSet(); + } + + bool HasUnit(const Unit &U) { return Hashes.count(Hash(U)); } + bool HasUnit(const std::string &H) { return Hashes.count(H); } + InputInfo &ChooseUnitToMutate(Random &Rand) { + InputInfo &II = *Inputs[ChooseUnitIdxToMutate(Rand)]; + assert(!II.U.empty()); + return II; + }; + + // Returns an index of random unit from the corpus to mutate. + // Hypothesis: units added to the corpus last are more likely to be + // interesting. This function gives more weight to the more recent units. + size_t ChooseUnitIdxToMutate(Random &Rand) { + size_t Idx = static_cast(CorpusDistribution(Rand.Get_mt19937())); + assert(Idx < Inputs.size()); + return Idx; + } + + void PrintStats() { + for (size_t i = 0; i < Inputs.size(); i++) { + const auto &II = *Inputs[i]; + Printf(" [%zd %s]\tsz: %zd\truns: %zd\tsucc: %zd\n", i, + Sha1ToString(II.Sha1).c_str(), II.U.size(), + II.NumExecutedMutations, II.NumSuccessfullMutations); + } + } + + void PrintFeatureSet() { + for (size_t i = 0; i < kFeatureSetSize; i++) { + if(size_t Sz = GetFeature(i)) + Printf("[%zd: id %zd sz%zd] ", i, SmallestElementPerFeature[i], Sz); + } + Printf("\n\t"); + for (size_t i = 0; i < Inputs.size(); i++) + if (size_t N = Inputs[i]->NumFeatures) + Printf(" %zd=>%zd ", i, N); + Printf("\n"); + } + + void DeleteInput(size_t Idx) { + InputInfo &II = *Inputs[Idx]; + if (!OutputCorpus.empty() && II.MayDeleteFile) + RemoveFile(DirPlusFile(OutputCorpus, Sha1ToString(II.Sha1))); + Unit().swap(II.U); + if (FeatureDebug) + Printf("EVICTED %zd\n", Idx); + } + + bool AddFeature(size_t Idx, uint32_t NewSize, bool Shrink) { + assert(NewSize); + Idx = Idx % kFeatureSetSize; + uint32_t OldSize = GetFeature(Idx); + if (OldSize == 0 || (Shrink && OldSize > NewSize)) { + if (OldSize > 0) { + size_t OldIdx = SmallestElementPerFeature[Idx]; + InputInfo &II = *Inputs[OldIdx]; + assert(II.NumFeatures > 0); + II.NumFeatures--; + if (II.NumFeatures == 0) + DeleteInput(OldIdx); + } + if (FeatureDebug) + Printf("ADD FEATURE %zd sz %d\n", Idx, NewSize); + SmallestElementPerFeature[Idx] = Inputs.size(); + InputSizesPerFeature[Idx] = NewSize; + CountingFeatures = true; + return true; + } + return false; + } + + size_t NumFeatures() const { + size_t Res = 0; + for (size_t i = 0; i < kFeatureSetSize; i++) + Res += GetFeature(i) != 0; + return Res; + } + + void ResetFeatureSet() { + assert(Inputs.empty()); + memset(InputSizesPerFeature, 0, sizeof(InputSizesPerFeature)); + memset(SmallestElementPerFeature, 0, sizeof(SmallestElementPerFeature)); + } + +private: + + static const bool FeatureDebug = false; + + size_t GetFeature(size_t Idx) const { return InputSizesPerFeature[Idx]; } + + void ValidateFeatureSet() { + if (!CountingFeatures) return; + if (FeatureDebug) + PrintFeatureSet(); + for (size_t Idx = 0; Idx < kFeatureSetSize; Idx++) + if (GetFeature(Idx)) + Inputs[SmallestElementPerFeature[Idx]]->Tmp++; + for (auto II: Inputs) { + if (II->Tmp != II->NumFeatures) + Printf("ZZZ %zd %zd\n", II->Tmp, II->NumFeatures); + assert(II->Tmp == II->NumFeatures); + II->Tmp = 0; + } + } + + // Updates the probability distribution for the units in the corpus. + // Must be called whenever the corpus or unit weights are changed. + void UpdateCorpusDistribution() { + size_t N = Inputs.size(); + Intervals.resize(N + 1); + Weights.resize(N); + std::iota(Intervals.begin(), Intervals.end(), 0); + if (CountingFeatures) + for (size_t i = 0; i < N; i++) + Weights[i] = Inputs[i]->NumFeatures * (i + 1); + else + std::iota(Weights.begin(), Weights.end(), 1); + CorpusDistribution = std::piecewise_constant_distribution( + Intervals.begin(), Intervals.end(), Weights.begin()); + } + std::piecewise_constant_distribution CorpusDistribution; + + std::vector Intervals; + std::vector Weights; + + std::unordered_set Hashes; + std::vector Inputs; + + bool CountingFeatures = false; + uint32_t InputSizesPerFeature[kFeatureSetSize]; + uint32_t SmallestElementPerFeature[kFeatureSetSize]; + + std::string OutputCorpus; +}; + +} // namespace fuzzer + +#endif // LLVM_FUZZER_CORPUS diff --git a/test/thirdparty/Fuzzer/FuzzerCrossOver.cpp b/test/thirdparty/Fuzzer/FuzzerCrossOver.cpp new file mode 100644 index 00000000..8b0fd7d5 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerCrossOver.cpp @@ -0,0 +1,52 @@ +//===- FuzzerCrossOver.cpp - Cross over two test inputs -------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Cross over test inputs. +//===----------------------------------------------------------------------===// + +#include "FuzzerDefs.h" +#include "FuzzerMutate.h" +#include "FuzzerRandom.h" +#include + +namespace fuzzer { + +// Cross Data1 and Data2, store the result (up to MaxOutSize bytes) in Out. +size_t MutationDispatcher::CrossOver(const uint8_t *Data1, size_t Size1, + const uint8_t *Data2, size_t Size2, + uint8_t *Out, size_t MaxOutSize) { + assert(Size1 || Size2); + MaxOutSize = Rand(MaxOutSize) + 1; + size_t OutPos = 0; + size_t Pos1 = 0; + size_t Pos2 = 0; + size_t *InPos = &Pos1; + size_t InSize = Size1; + const uint8_t *Data = Data1; + bool CurrentlyUsingFirstData = true; + while (OutPos < MaxOutSize && (Pos1 < Size1 || Pos2 < Size2)) { + // Merge a part of Data into Out. + size_t OutSizeLeft = MaxOutSize - OutPos; + if (*InPos < InSize) { + size_t InSizeLeft = InSize - *InPos; + size_t MaxExtraSize = std::min(OutSizeLeft, InSizeLeft); + size_t ExtraSize = Rand(MaxExtraSize) + 1; + memcpy(Out + OutPos, Data + *InPos, ExtraSize); + OutPos += ExtraSize; + (*InPos) += ExtraSize; + } + // Use the other input data on the next iteration. + InPos = CurrentlyUsingFirstData ? &Pos2 : &Pos1; + InSize = CurrentlyUsingFirstData ? Size2 : Size1; + Data = CurrentlyUsingFirstData ? Data2 : Data1; + CurrentlyUsingFirstData = !CurrentlyUsingFirstData; + } + return OutPos; +} + +} // namespace fuzzer diff --git a/test/thirdparty/Fuzzer/FuzzerDefs.h b/test/thirdparty/Fuzzer/FuzzerDefs.h new file mode 100644 index 00000000..0f5b8a7c --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerDefs.h @@ -0,0 +1,89 @@ +//===- FuzzerDefs.h - Internal header for the Fuzzer ------------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Basic definitions. +//===----------------------------------------------------------------------===// + +#ifndef LLVM_FUZZER_DEFS_H +#define LLVM_FUZZER_DEFS_H + +#include +#include +#include +#include +#include +#include + +// Platform detection. +#ifdef __linux__ +#define LIBFUZZER_APPLE 0 +#define LIBFUZZER_LINUX 1 +#define LIBFUZZER_WINDOWS 0 +#elif __APPLE__ +#define LIBFUZZER_APPLE 1 +#define LIBFUZZER_LINUX 0 +#define LIBFUZZER_WINDOWS 0 +#elif _WIN32 +#define LIBFUZZER_APPLE 0 +#define LIBFUZZER_LINUX 0 +#define LIBFUZZER_WINDOWS 1 +#else +#error "Support for your platform has not been implemented" +#endif + +#define LIBFUZZER_POSIX LIBFUZZER_APPLE || LIBFUZZER_LINUX + +#ifdef __x86_64 +#define ATTRIBUTE_TARGET_POPCNT __attribute__((target("popcnt"))) +#else +#define ATTRIBUTE_TARGET_POPCNT +#endif + + +#ifdef __clang__ // avoid gcc warning. +# define ATTRIBUTE_NO_SANITIZE_MEMORY __attribute__((no_sanitize("memory"))) +#else +# define ATTRIBUTE_NO_SANITIZE_MEMORY +#endif + +namespace fuzzer { + +template T Min(T a, T b) { return a < b ? a : b; } +template T Max(T a, T b) { return a > b ? a : b; } + +class Random; +class Dictionary; +class DictionaryEntry; +class MutationDispatcher; +struct FuzzingOptions; +class InputCorpus; +struct InputInfo; +struct ExternalFunctions; + +// Global interface to functions that may or may not be available. +extern ExternalFunctions *EF; + +typedef std::vector Unit; +typedef std::vector UnitVector; +typedef int (*UserCallback)(const uint8_t *Data, size_t Size); + +int FuzzerDriver(int *argc, char ***argv, UserCallback Callback); + +struct ScopedDoingMyOwnMemmem { + ScopedDoingMyOwnMemmem(); + ~ScopedDoingMyOwnMemmem(); +}; + +inline uint8_t Bswap(uint8_t x) { return x; } +inline uint16_t Bswap(uint16_t x) { return __builtin_bswap16(x); } +inline uint32_t Bswap(uint32_t x) { return __builtin_bswap32(x); } +inline uint64_t Bswap(uint64_t x) { return __builtin_bswap64(x); } + +} // namespace fuzzer + +#endif // LLVM_FUZZER_DEFS_H diff --git a/test/thirdparty/Fuzzer/FuzzerDictionary.h b/test/thirdparty/Fuzzer/FuzzerDictionary.h new file mode 100644 index 00000000..eba0eabb --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerDictionary.h @@ -0,0 +1,124 @@ +//===- FuzzerDictionary.h - Internal header for the Fuzzer ------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// fuzzer::Dictionary +//===----------------------------------------------------------------------===// + +#ifndef LLVM_FUZZER_DICTIONARY_H +#define LLVM_FUZZER_DICTIONARY_H + +#include "FuzzerDefs.h" +#include "FuzzerIO.h" +#include "FuzzerUtil.h" +#include +#include + +namespace fuzzer { +// A simple POD sized array of bytes. +template class FixedWord { +public: + FixedWord() {} + FixedWord(const uint8_t *B, uint8_t S) { Set(B, S); } + + void Set(const uint8_t *B, uint8_t S) { + assert(S <= kMaxSize); + memcpy(Data, B, S); + Size = S; + } + + bool operator==(const FixedWord &w) const { + return Size == w.Size && 0 == memcmp(Data, w.Data, Size); + } + + bool operator<(const FixedWord &w) const { + if (Size != w.Size) + return Size < w.Size; + return memcmp(Data, w.Data, Size) < 0; + } + + static size_t GetMaxSize() { return kMaxSize; } + const uint8_t *data() const { return Data; } + uint8_t size() const { return Size; } + +private: + uint8_t Size = 0; + uint8_t Data[kMaxSize]; +}; + +typedef FixedWord<27> Word; // 28 bytes. + +class DictionaryEntry { + public: + DictionaryEntry() {} + DictionaryEntry(Word W) : W(W) {} + DictionaryEntry(Word W, size_t PositionHint) : W(W), PositionHint(PositionHint) {} + const Word &GetW() const { return W; } + + bool HasPositionHint() const { return PositionHint != std::numeric_limits::max(); } + size_t GetPositionHint() const { + assert(HasPositionHint()); + return PositionHint; + } + void IncUseCount() { UseCount++; } + void IncSuccessCount() { SuccessCount++; } + size_t GetUseCount() const { return UseCount; } + size_t GetSuccessCount() const {return SuccessCount; } + + void Print(const char *PrintAfter = "\n") { + PrintASCII(W.data(), W.size()); + if (HasPositionHint()) + Printf("@%zd", GetPositionHint()); + Printf("%s", PrintAfter); + } + +private: + Word W; + size_t PositionHint = std::numeric_limits::max(); + size_t UseCount = 0; + size_t SuccessCount = 0; +}; + +class Dictionary { + public: + static const size_t kMaxDictSize = 1 << 14; + + bool ContainsWord(const Word &W) const { + return std::any_of(begin(), end(), [&](const DictionaryEntry &DE) { + return DE.GetW() == W; + }); + } + const DictionaryEntry *begin() const { return &DE[0]; } + const DictionaryEntry *end() const { return begin() + Size; } + DictionaryEntry & operator[] (size_t Idx) { + assert(Idx < Size); + return DE[Idx]; + } + void push_back(DictionaryEntry DE) { + if (Size < kMaxDictSize) + this->DE[Size++] = DE; + } + void clear() { Size = 0; } + bool empty() const { return Size == 0; } + size_t size() const { return Size; } + +private: + DictionaryEntry DE[kMaxDictSize]; + size_t Size = 0; +}; + +// Parses one dictionary entry. +// If successfull, write the enty to Unit and returns true, +// otherwise returns false. +bool ParseOneDictionaryEntry(const std::string &Str, Unit *U); +// Parses the dictionary file, fills Units, returns true iff all lines +// were parsed succesfully. +bool ParseDictionaryFile(const std::string &Text, std::vector *Units); + +} // namespace fuzzer + +#endif // LLVM_FUZZER_DICTIONARY_H diff --git a/test/thirdparty/Fuzzer/FuzzerDriver.cpp b/test/thirdparty/Fuzzer/FuzzerDriver.cpp new file mode 100644 index 00000000..95b0721c --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerDriver.cpp @@ -0,0 +1,545 @@ +//===- FuzzerDriver.cpp - FuzzerDriver function and flags -----------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// FuzzerDriver and flag parsing. +//===----------------------------------------------------------------------===// + +#include "FuzzerCorpus.h" +#include "FuzzerInterface.h" +#include "FuzzerInternal.h" +#include "FuzzerIO.h" +#include "FuzzerMutate.h" +#include "FuzzerRandom.h" +#include "FuzzerTracePC.h" +#include +#include +#include +#include +#include +#include +#include + +// This function should be present in the libFuzzer so that the client +// binary can test for its existence. +extern "C" __attribute__((used)) void __libfuzzer_is_present() {} + +namespace fuzzer { + +// Program arguments. +struct FlagDescription { + const char *Name; + const char *Description; + int Default; + int *IntFlag; + const char **StrFlag; + unsigned int *UIntFlag; +}; + +struct { +#define FUZZER_DEPRECATED_FLAG(Name) +#define FUZZER_FLAG_INT(Name, Default, Description) int Name; +#define FUZZER_FLAG_UNSIGNED(Name, Default, Description) unsigned int Name; +#define FUZZER_FLAG_STRING(Name, Description) const char *Name; +#include "FuzzerFlags.def" +#undef FUZZER_DEPRECATED_FLAG +#undef FUZZER_FLAG_INT +#undef FUZZER_FLAG_UNSIGNED +#undef FUZZER_FLAG_STRING +} Flags; + +static const FlagDescription FlagDescriptions [] { +#define FUZZER_DEPRECATED_FLAG(Name) \ + {#Name, "Deprecated; don't use", 0, nullptr, nullptr, nullptr}, +#define FUZZER_FLAG_INT(Name, Default, Description) \ + {#Name, Description, Default, &Flags.Name, nullptr, nullptr}, +#define FUZZER_FLAG_UNSIGNED(Name, Default, Description) \ + {#Name, Description, static_cast(Default), \ + nullptr, nullptr, &Flags.Name}, +#define FUZZER_FLAG_STRING(Name, Description) \ + {#Name, Description, 0, nullptr, &Flags.Name, nullptr}, +#include "FuzzerFlags.def" +#undef FUZZER_DEPRECATED_FLAG +#undef FUZZER_FLAG_INT +#undef FUZZER_FLAG_UNSIGNED +#undef FUZZER_FLAG_STRING +}; + +static const size_t kNumFlags = + sizeof(FlagDescriptions) / sizeof(FlagDescriptions[0]); + +static std::vector *Inputs; +static std::string *ProgName; + +static void PrintHelp() { + Printf("Usage:\n"); + auto Prog = ProgName->c_str(); + Printf("\nTo run fuzzing pass 0 or more directories.\n"); + Printf("%s [-flag1=val1 [-flag2=val2 ...] ] [dir1 [dir2 ...] ]\n", Prog); + + Printf("\nTo run individual tests without fuzzing pass 1 or more files:\n"); + Printf("%s [-flag1=val1 [-flag2=val2 ...] ] file1 [file2 ...]\n", Prog); + + Printf("\nFlags: (strictly in form -flag=value)\n"); + size_t MaxFlagLen = 0; + for (size_t F = 0; F < kNumFlags; F++) + MaxFlagLen = std::max(strlen(FlagDescriptions[F].Name), MaxFlagLen); + + for (size_t F = 0; F < kNumFlags; F++) { + const auto &D = FlagDescriptions[F]; + if (strstr(D.Description, "internal flag") == D.Description) continue; + Printf(" %s", D.Name); + for (size_t i = 0, n = MaxFlagLen - strlen(D.Name); i < n; i++) + Printf(" "); + Printf("\t"); + Printf("%d\t%s\n", D.Default, D.Description); + } + Printf("\nFlags starting with '--' will be ignored and " + "will be passed verbatim to subprocesses.\n"); +} + +static const char *FlagValue(const char *Param, const char *Name) { + size_t Len = strlen(Name); + if (Param[0] == '-' && strstr(Param + 1, Name) == Param + 1 && + Param[Len + 1] == '=') + return &Param[Len + 2]; + return nullptr; +} + +// Avoid calling stol as it triggers a bug in clang/glibc build. +static long MyStol(const char *Str) { + long Res = 0; + long Sign = 1; + if (*Str == '-') { + Str++; + Sign = -1; + } + for (size_t i = 0; Str[i]; i++) { + char Ch = Str[i]; + if (Ch < '0' || Ch > '9') + return Res; + Res = Res * 10 + (Ch - '0'); + } + return Res * Sign; +} + +static bool ParseOneFlag(const char *Param) { + if (Param[0] != '-') return false; + if (Param[1] == '-') { + static bool PrintedWarning = false; + if (!PrintedWarning) { + PrintedWarning = true; + Printf("INFO: libFuzzer ignores flags that start with '--'\n"); + } + for (size_t F = 0; F < kNumFlags; F++) + if (FlagValue(Param + 1, FlagDescriptions[F].Name)) + Printf("WARNING: did you mean '%s' (single dash)?\n", Param + 1); + return true; + } + for (size_t F = 0; F < kNumFlags; F++) { + const char *Name = FlagDescriptions[F].Name; + const char *Str = FlagValue(Param, Name); + if (Str) { + if (FlagDescriptions[F].IntFlag) { + int Val = MyStol(Str); + *FlagDescriptions[F].IntFlag = Val; + if (Flags.verbosity >= 2) + Printf("Flag: %s %d\n", Name, Val);; + return true; + } else if (FlagDescriptions[F].UIntFlag) { + unsigned int Val = std::stoul(Str); + *FlagDescriptions[F].UIntFlag = Val; + if (Flags.verbosity >= 2) + Printf("Flag: %s %u\n", Name, Val); + return true; + } else if (FlagDescriptions[F].StrFlag) { + *FlagDescriptions[F].StrFlag = Str; + if (Flags.verbosity >= 2) + Printf("Flag: %s %s\n", Name, Str); + return true; + } else { // Deprecated flag. + Printf("Flag: %s: deprecated, don't use\n", Name); + return true; + } + } + } + Printf("\n\nWARNING: unrecognized flag '%s'; " + "use -help=1 to list all flags\n\n", Param); + return true; +} + +// We don't use any library to minimize dependencies. +static void ParseFlags(const std::vector &Args) { + for (size_t F = 0; F < kNumFlags; F++) { + if (FlagDescriptions[F].IntFlag) + *FlagDescriptions[F].IntFlag = FlagDescriptions[F].Default; + if (FlagDescriptions[F].UIntFlag) + *FlagDescriptions[F].UIntFlag = + static_cast(FlagDescriptions[F].Default); + if (FlagDescriptions[F].StrFlag) + *FlagDescriptions[F].StrFlag = nullptr; + } + Inputs = new std::vector; + for (size_t A = 1; A < Args.size(); A++) { + if (ParseOneFlag(Args[A].c_str())) continue; + Inputs->push_back(Args[A]); + } +} + +static std::mutex Mu; + +static void PulseThread() { + while (true) { + SleepSeconds(600); + std::lock_guard Lock(Mu); + Printf("pulse...\n"); + } +} + +static void WorkerThread(const std::string &Cmd, std::atomic *Counter, + unsigned NumJobs, std::atomic *HasErrors) { + while (true) { + unsigned C = (*Counter)++; + if (C >= NumJobs) break; + std::string Log = "fuzz-" + std::to_string(C) + ".log"; + std::string ToRun = Cmd + " > " + Log + " 2>&1\n"; + if (Flags.verbosity) + Printf("%s", ToRun.c_str()); + int ExitCode = ExecuteCommand(ToRun); + if (ExitCode != 0) + *HasErrors = true; + std::lock_guard Lock(Mu); + Printf("================== Job %u exited with exit code %d ============\n", + C, ExitCode); + fuzzer::CopyFileToErr(Log); + } +} + +std::string CloneArgsWithoutX(const std::vector &Args, + const char *X1, const char *X2) { + std::string Cmd; + for (auto &S : Args) { + if (FlagValue(S.c_str(), X1) || FlagValue(S.c_str(), X2)) + continue; + Cmd += S + " "; + } + return Cmd; +} + +static int RunInMultipleProcesses(const std::vector &Args, + unsigned NumWorkers, unsigned NumJobs) { + std::atomic Counter(0); + std::atomic HasErrors(false); + std::string Cmd = CloneArgsWithoutX(Args, "jobs", "workers"); + std::vector V; + std::thread Pulse(PulseThread); + Pulse.detach(); + for (unsigned i = 0; i < NumWorkers; i++) + V.push_back(std::thread(WorkerThread, Cmd, &Counter, NumJobs, &HasErrors)); + for (auto &T : V) + T.join(); + return HasErrors ? 1 : 0; +} + +static void RssThread(Fuzzer *F, size_t RssLimitMb) { + while (true) { + SleepSeconds(1); + size_t Peak = GetPeakRSSMb(); + if (Peak > RssLimitMb) + F->RssLimitCallback(); + } +} + +static void StartRssThread(Fuzzer *F, size_t RssLimitMb) { + if (!RssLimitMb) return; + std::thread T(RssThread, F, RssLimitMb); + T.detach(); +} + +int RunOneTest(Fuzzer *F, const char *InputFilePath, size_t MaxLen) { + Unit U = FileToVector(InputFilePath); + if (MaxLen && MaxLen < U.size()) + U.resize(MaxLen); + F->RunOne(U.data(), U.size()); + F->TryDetectingAMemoryLeak(U.data(), U.size(), true); + return 0; +} + +static bool AllInputsAreFiles() { + if (Inputs->empty()) return false; + for (auto &Path : *Inputs) + if (!IsFile(Path)) + return false; + return true; +} + +int MinimizeCrashInput(const std::vector &Args) { + if (Inputs->size() != 1) { + Printf("ERROR: -minimize_crash should be given one input file\n"); + exit(1); + } + std::string InputFilePath = Inputs->at(0); + std::string BaseCmd = + CloneArgsWithoutX(Args, "minimize_crash", "exact_artifact_path"); + auto InputPos = BaseCmd.find(" " + InputFilePath + " "); + assert(InputPos != std::string::npos); + BaseCmd.erase(InputPos, InputFilePath.size() + 1); + if (Flags.runs <= 0 && Flags.max_total_time == 0) { + Printf("INFO: you need to specify -runs=N or " + "-max_total_time=N with -minimize_crash=1\n" + "INFO: defaulting to -max_total_time=600\n"); + BaseCmd += " -max_total_time=600"; + } + // BaseCmd += " > /dev/null 2>&1 "; + + std::string CurrentFilePath = InputFilePath; + while (true) { + Unit U = FileToVector(CurrentFilePath); + if (U.size() < 2) { + Printf("CRASH_MIN: '%s' is small enough\n", CurrentFilePath.c_str()); + return 0; + } + Printf("CRASH_MIN: minimizing crash input: '%s' (%zd bytes)\n", + CurrentFilePath.c_str(), U.size()); + + auto Cmd = BaseCmd + " " + CurrentFilePath; + + Printf("CRASH_MIN: executing: %s\n", Cmd.c_str()); + int ExitCode = ExecuteCommand(Cmd); + if (ExitCode == 0) { + Printf("ERROR: the input %s did not crash\n", CurrentFilePath.c_str()); + exit(1); + } + Printf("CRASH_MIN: '%s' (%zd bytes) caused a crash. Will try to minimize " + "it further\n", + CurrentFilePath.c_str(), U.size()); + + std::string ArtifactPath = "minimized-from-" + Hash(U); + Cmd += " -minimize_crash_internal_step=1 -exact_artifact_path=" + + ArtifactPath; + Printf("CRASH_MIN: executing: %s\n", Cmd.c_str()); + ExitCode = ExecuteCommand(Cmd); + if (ExitCode == 0) { + if (Flags.exact_artifact_path) { + CurrentFilePath = Flags.exact_artifact_path; + WriteToFile(U, CurrentFilePath); + } + Printf("CRASH_MIN: failed to minimize beyond %s (%d bytes), exiting\n", + CurrentFilePath.c_str(), U.size()); + return 0; + } + CurrentFilePath = ArtifactPath; + Printf("\n\n\n\n\n\n*********************************\n"); + } + return 0; +} + +int MinimizeCrashInputInternalStep(Fuzzer *F, InputCorpus *Corpus) { + assert(Inputs->size() == 1); + std::string InputFilePath = Inputs->at(0); + Unit U = FileToVector(InputFilePath); + assert(U.size() > 2); + Printf("INFO: Starting MinimizeCrashInputInternalStep: %zd\n", U.size()); + Corpus->AddToCorpus(U, 0); + F->SetMaxInputLen(U.size()); + F->SetMaxMutationLen(U.size() - 1); + F->MinimizeCrashLoop(U); + Printf("INFO: Done MinimizeCrashInputInternalStep, no crashes found\n"); + exit(0); + return 0; +} + +int FuzzerDriver(int *argc, char ***argv, UserCallback Callback) { + using namespace fuzzer; + assert(argc && argv && "Argument pointers cannot be nullptr"); + EF = new ExternalFunctions(); + if (EF->LLVMFuzzerInitialize) + EF->LLVMFuzzerInitialize(argc, argv); + const std::vector Args(*argv, *argv + *argc); + assert(!Args.empty()); + ProgName = new std::string(Args[0]); + ParseFlags(Args); + if (Flags.help) { + PrintHelp(); + return 0; + } + + if (Flags.minimize_crash) + return MinimizeCrashInput(Args); + + if (Flags.close_fd_mask & 2) + DupAndCloseStderr(); + if (Flags.close_fd_mask & 1) + CloseStdout(); + + if (Flags.jobs > 0 && Flags.workers == 0) { + Flags.workers = std::min(NumberOfCpuCores() / 2, Flags.jobs); + if (Flags.workers > 1) + Printf("Running %u workers\n", Flags.workers); + } + + if (Flags.workers > 0 && Flags.jobs > 0) + return RunInMultipleProcesses(Args, Flags.workers, Flags.jobs); + + const size_t kMaxSaneLen = 1 << 20; + const size_t kMinDefaultLen = 64; + FuzzingOptions Options; + Options.Verbosity = Flags.verbosity; + Options.MaxLen = Flags.max_len; + Options.UnitTimeoutSec = Flags.timeout; + Options.ErrorExitCode = Flags.error_exitcode; + Options.TimeoutExitCode = Flags.timeout_exitcode; + Options.MaxTotalTimeSec = Flags.max_total_time; + Options.DoCrossOver = Flags.cross_over; + Options.MutateDepth = Flags.mutate_depth; + Options.UseCounters = Flags.use_counters; + Options.UseIndirCalls = Flags.use_indir_calls; + Options.UseMemcmp = Flags.use_memcmp; + Options.UseMemmem = Flags.use_memmem; + Options.UseCmp = Flags.use_cmp; + Options.UseValueProfile = Flags.use_value_profile; + Options.Shrink = Flags.shrink; + Options.ShuffleAtStartUp = Flags.shuffle; + Options.PreferSmall = Flags.prefer_small; + Options.ReloadIntervalSec = Flags.reload; + Options.OnlyASCII = Flags.only_ascii; + Options.OutputCSV = Flags.output_csv; + Options.DetectLeaks = Flags.detect_leaks; + Options.TraceMalloc = Flags.trace_malloc; + Options.RssLimitMb = Flags.rss_limit_mb; + if (Flags.runs >= 0) + Options.MaxNumberOfRuns = Flags.runs; + if (!Inputs->empty() && !Flags.minimize_crash_internal_step) + Options.OutputCorpus = (*Inputs)[0]; + Options.ReportSlowUnits = Flags.report_slow_units; + if (Flags.artifact_prefix) + Options.ArtifactPrefix = Flags.artifact_prefix; + if (Flags.exact_artifact_path) + Options.ExactArtifactPath = Flags.exact_artifact_path; + std::vector Dictionary; + if (Flags.dict) + if (!ParseDictionaryFile(FileToString(Flags.dict), &Dictionary)) + return 1; + if (Flags.verbosity > 0 && !Dictionary.empty()) + Printf("Dictionary: %zd entries\n", Dictionary.size()); + bool DoPlainRun = AllInputsAreFiles(); + Options.SaveArtifacts = + !DoPlainRun || Flags.minimize_crash_internal_step; + Options.PrintNewCovPcs = Flags.print_pcs; + Options.PrintFinalStats = Flags.print_final_stats; + Options.PrintCorpusStats = Flags.print_corpus_stats; + Options.PrintCoverage = Flags.print_coverage; + Options.DumpCoverage = Flags.dump_coverage; + if (Flags.exit_on_src_pos) + Options.ExitOnSrcPos = Flags.exit_on_src_pos; + if (Flags.exit_on_item) + Options.ExitOnItem = Flags.exit_on_item; + + unsigned Seed = Flags.seed; + // Initialize Seed. + if (Seed == 0) + Seed = (std::chrono::system_clock::now().time_since_epoch().count() << 10) + + GetPid(); + if (Flags.verbosity) + Printf("INFO: Seed: %u\n", Seed); + + Random Rand(Seed); + auto *MD = new MutationDispatcher(Rand, Options); + auto *Corpus = new InputCorpus(Options.OutputCorpus); + auto *F = new Fuzzer(Callback, *Corpus, *MD, Options); + + for (auto &U: Dictionary) + if (U.size() <= Word::GetMaxSize()) + MD->AddWordToManualDictionary(Word(U.data(), U.size())); + + StartRssThread(F, Flags.rss_limit_mb); + + Options.HandleAbrt = Flags.handle_abrt; + Options.HandleBus = Flags.handle_bus; + Options.HandleFpe = Flags.handle_fpe; + Options.HandleIll = Flags.handle_ill; + Options.HandleInt = Flags.handle_int; + Options.HandleSegv = Flags.handle_segv; + Options.HandleTerm = Flags.handle_term; + SetSignalHandler(Options); + + if (Flags.minimize_crash_internal_step) + return MinimizeCrashInputInternalStep(F, Corpus); + + if (DoPlainRun) { + Options.SaveArtifacts = false; + int Runs = std::max(1, Flags.runs); + Printf("%s: Running %zd inputs %d time(s) each.\n", ProgName->c_str(), + Inputs->size(), Runs); + for (auto &Path : *Inputs) { + auto StartTime = system_clock::now(); + Printf("Running: %s\n", Path.c_str()); + for (int Iter = 0; Iter < Runs; Iter++) + RunOneTest(F, Path.c_str(), Options.MaxLen); + auto StopTime = system_clock::now(); + auto MS = duration_cast(StopTime - StartTime).count(); + Printf("Executed %s in %zd ms\n", Path.c_str(), (long)MS); + } + Printf("***\n" + "*** NOTE: fuzzing was not performed, you have only\n" + "*** executed the target code on a fixed set of inputs.\n" + "***\n"); + F->PrintFinalStats(); + exit(0); + } + + if (Flags.merge) { + if (Options.MaxLen == 0) + F->SetMaxInputLen(kMaxSaneLen); + if (TPC.UsingTracePcGuard()) { + if (Flags.merge_control_file) + F->CrashResistantMergeInternalStep(Flags.merge_control_file); + else + F->CrashResistantMerge(Args, *Inputs); + } else { + F->Merge(*Inputs); + } + exit(0); + } + + size_t TemporaryMaxLen = Options.MaxLen ? Options.MaxLen : kMaxSaneLen; + + UnitVector InitialCorpus; + for (auto &Inp : *Inputs) { + Printf("Loading corpus dir: %s\n", Inp.c_str()); + ReadDirToVectorOfUnits(Inp.c_str(), &InitialCorpus, nullptr, + TemporaryMaxLen, /*ExitOnError=*/false); + } + + if (Options.MaxLen == 0) { + size_t MaxLen = 0; + for (auto &U : InitialCorpus) + MaxLen = std::max(U.size(), MaxLen); + F->SetMaxInputLen(std::min(std::max(kMinDefaultLen, MaxLen), kMaxSaneLen)); + } + + if (InitialCorpus.empty()) { + InitialCorpus.push_back(Unit({'\n'})); // Valid ASCII input. + if (Options.Verbosity) + Printf("INFO: A corpus is not provided, starting from an empty corpus\n"); + } + F->ShuffleAndMinimize(&InitialCorpus); + InitialCorpus.clear(); // Don't need this memory any more. + F->Loop(); + + if (Flags.verbosity) + Printf("Done %d runs in %zd second(s)\n", F->getTotalNumberOfRuns(), + F->secondsSinceProcessStartUp()); + F->PrintFinalStats(); + + exit(0); // Don't let F destroy itself. +} + +// Storage for global ExternalFunctions object. +ExternalFunctions *EF = nullptr; + +} // namespace fuzzer diff --git a/test/thirdparty/Fuzzer/FuzzerExtFunctions.def b/test/thirdparty/Fuzzer/FuzzerExtFunctions.def new file mode 100644 index 00000000..61c72e4a --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerExtFunctions.def @@ -0,0 +1,50 @@ +//===- FuzzerExtFunctions.def - External functions --------------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// This defines the external function pointers that +// ``fuzzer::ExternalFunctions`` should contain and try to initialize. The +// EXT_FUNC macro must be defined at the point of inclusion. The signature of +// the macro is: +// +// EXT_FUNC(, , , ) +//===----------------------------------------------------------------------===// + +// Optional user functions +EXT_FUNC(LLVMFuzzerInitialize, int, (int *argc, char ***argv), false); +EXT_FUNC(LLVMFuzzerCustomMutator, size_t, + (uint8_t * Data, size_t Size, size_t MaxSize, unsigned int Seed), + false); +EXT_FUNC(LLVMFuzzerCustomCrossOver, size_t, + (const uint8_t * Data1, size_t Size1, + const uint8_t * Data2, size_t Size2, + uint8_t * Out, size_t MaxOutSize, unsigned int Seed), + false); + +// Sanitizer functions +EXT_FUNC(__lsan_enable, void, (), false); +EXT_FUNC(__lsan_disable, void, (), false); +EXT_FUNC(__lsan_do_recoverable_leak_check, int, (), false); +EXT_FUNC(__sanitizer_get_number_of_counters, size_t, (), false); +EXT_FUNC(__sanitizer_install_malloc_and_free_hooks, int, + (void (*malloc_hook)(const volatile void *, size_t), + void (*free_hook)(const volatile void *)), + false); +EXT_FUNC(__sanitizer_get_total_unique_caller_callee_pairs, size_t, (), false); +EXT_FUNC(__sanitizer_get_total_unique_coverage, size_t, (), true); +EXT_FUNC(__sanitizer_print_memory_profile, int, (size_t), false); +EXT_FUNC(__sanitizer_print_stack_trace, void, (), true); +EXT_FUNC(__sanitizer_symbolize_pc, void, + (void *, const char *fmt, char *out_buf, size_t out_buf_size), false); +EXT_FUNC(__sanitizer_get_module_and_offset_for_pc, int, + (void *pc, char *module_path, + size_t module_path_len,void **pc_offset), false); +EXT_FUNC(__sanitizer_reset_coverage, void, (), true); +EXT_FUNC(__sanitizer_set_death_callback, void, (void (*)(void)), true); +EXT_FUNC(__sanitizer_set_report_fd, void, (void*), false); +EXT_FUNC(__sanitizer_update_counter_bitset_and_clear_counters, uintptr_t, + (uint8_t*), false); diff --git a/test/thirdparty/Fuzzer/FuzzerExtFunctions.h b/test/thirdparty/Fuzzer/FuzzerExtFunctions.h new file mode 100644 index 00000000..2672a385 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerExtFunctions.h @@ -0,0 +1,35 @@ +//===- FuzzerExtFunctions.h - Interface to external functions ---*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Defines an interface to (possibly optional) functions. +//===----------------------------------------------------------------------===// + +#ifndef LLVM_FUZZER_EXT_FUNCTIONS_H +#define LLVM_FUZZER_EXT_FUNCTIONS_H + +#include +#include + +namespace fuzzer { + +struct ExternalFunctions { + // Initialize function pointers. Functions that are not available will be set + // to nullptr. Do not call this constructor before ``main()`` has been + // entered. + ExternalFunctions(); + +#define EXT_FUNC(NAME, RETURN_TYPE, FUNC_SIG, WARN) \ + RETURN_TYPE(*NAME) FUNC_SIG = nullptr + +#include "FuzzerExtFunctions.def" + +#undef EXT_FUNC +}; +} // namespace fuzzer + +#endif diff --git a/test/thirdparty/Fuzzer/FuzzerExtFunctionsDlsym.cpp b/test/thirdparty/Fuzzer/FuzzerExtFunctionsDlsym.cpp new file mode 100644 index 00000000..06bddd5d --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerExtFunctionsDlsym.cpp @@ -0,0 +1,52 @@ +//===- FuzzerExtFunctionsDlsym.cpp - Interface to external functions ------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Implementation for operating systems that support dlsym(). We only use it on +// Apple platforms for now. We don't use this approach on Linux because it +// requires that clients of LibFuzzer pass ``--export-dynamic`` to the linker. +// That is a complication we don't wish to expose to clients right now. +//===----------------------------------------------------------------------===// +#include "FuzzerDefs.h" +#if LIBFUZZER_APPLE + +#include "FuzzerExtFunctions.h" +#include "FuzzerIO.h" +#include + +using namespace fuzzer; + +template +static T GetFnPtr(const char *FnName, bool WarnIfMissing) { + dlerror(); // Clear any previous errors. + void *Fn = dlsym(RTLD_DEFAULT, FnName); + if (Fn == nullptr) { + if (WarnIfMissing) { + const char *ErrorMsg = dlerror(); + Printf("WARNING: Failed to find function \"%s\".", FnName); + if (ErrorMsg) + Printf(" Reason %s.", ErrorMsg); + Printf("\n"); + } + } + return reinterpret_cast(Fn); +} + +namespace fuzzer { + +ExternalFunctions::ExternalFunctions() { +#define EXT_FUNC(NAME, RETURN_TYPE, FUNC_SIG, WARN) \ + this->NAME = GetFnPtr(#NAME, WARN) + +#include "FuzzerExtFunctions.def" + +#undef EXT_FUNC +} + +} // namespace fuzzer + +#endif // LIBFUZZER_APPLE diff --git a/test/thirdparty/Fuzzer/FuzzerExtFunctionsWeak.cpp b/test/thirdparty/Fuzzer/FuzzerExtFunctionsWeak.cpp new file mode 100644 index 00000000..7b02b6f0 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerExtFunctionsWeak.cpp @@ -0,0 +1,53 @@ +//===- FuzzerExtFunctionsWeak.cpp - Interface to external functions -------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Implementation for Linux. This relies on the linker's support for weak +// symbols. We don't use this approach on Apple platforms because it requires +// clients of LibFuzzer to pass ``-U _`` to the linker to allow +// weak symbols to be undefined. That is a complication we don't want to expose +// to clients right now. +//===----------------------------------------------------------------------===// +#include "FuzzerDefs.h" +#if LIBFUZZER_LINUX + +#include "FuzzerExtFunctions.h" +#include "FuzzerIO.h" + +extern "C" { +// Declare these symbols as weak to allow them to be optionally defined. +#define EXT_FUNC(NAME, RETURN_TYPE, FUNC_SIG, WARN) \ + __attribute__((weak)) RETURN_TYPE NAME FUNC_SIG + +#include "FuzzerExtFunctions.def" + +#undef EXT_FUNC +} + +using namespace fuzzer; + +static void CheckFnPtr(void *FnPtr, const char *FnName, bool WarnIfMissing) { + if (FnPtr == nullptr && WarnIfMissing) { + Printf("WARNING: Failed to find function \"%s\".\n", FnName); + } +} + +namespace fuzzer { + +ExternalFunctions::ExternalFunctions() { +#define EXT_FUNC(NAME, RETURN_TYPE, FUNC_SIG, WARN) \ + this->NAME = ::NAME; \ + CheckFnPtr((void *)::NAME, #NAME, WARN); + +#include "FuzzerExtFunctions.def" + +#undef EXT_FUNC +} + +} // namespace fuzzer + +#endif // LIBFUZZER_LINUX diff --git a/test/thirdparty/Fuzzer/FuzzerExtFunctionsWeakAlias.cpp b/test/thirdparty/Fuzzer/FuzzerExtFunctionsWeakAlias.cpp new file mode 100644 index 00000000..e10f7b4d --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerExtFunctionsWeakAlias.cpp @@ -0,0 +1,56 @@ +//===- FuzzerExtFunctionsWeakAlias.cpp - Interface to external functions --===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Implementation using weak aliases. Works for Windows. +//===----------------------------------------------------------------------===// +#include "FuzzerDefs.h" +#if LIBFUZZER_WINDOWS + +#include "FuzzerExtFunctions.h" +#include "FuzzerIO.h" + +using namespace fuzzer; + +extern "C" { +// Declare these symbols as weak to allow them to be optionally defined. +#define EXT_FUNC(NAME, RETURN_TYPE, FUNC_SIG, WARN) \ + RETURN_TYPE NAME##Def FUNC_SIG { \ + Printf("ERROR: Function \"%s\" not defined.\n", #NAME); \ + exit(1); \ + } \ + RETURN_TYPE NAME FUNC_SIG __attribute__((weak, alias(#NAME "Def"))); + +#include "FuzzerExtFunctions.def" + +#undef EXT_FUNC +} + +template +static T *GetFnPtr(T *Fun, T *FunDef, const char *FnName, bool WarnIfMissing) { + if (Fun == FunDef) { + if (WarnIfMissing) + Printf("WARNING: Failed to find function \"%s\".\n", FnName); + return nullptr; + } + return Fun; +} + +namespace fuzzer { + +ExternalFunctions::ExternalFunctions() { +#define EXT_FUNC(NAME, RETURN_TYPE, FUNC_SIG, WARN) \ + this->NAME = GetFnPtr(::NAME, ::NAME##Def, #NAME, WARN); + +#include "FuzzerExtFunctions.def" + +#undef EXT_FUNC +} + +} // namespace fuzzer + +#endif // LIBFUZZER_WINDOWS diff --git a/test/thirdparty/Fuzzer/FuzzerFlags.def b/test/thirdparty/Fuzzer/FuzzerFlags.def new file mode 100644 index 00000000..25ef1741 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerFlags.def @@ -0,0 +1,115 @@ +//===- FuzzerFlags.def - Run-time flags -------------------------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Flags. FUZZER_FLAG_INT/FUZZER_FLAG_STRING macros should be defined at the +// point of inclusion. We are not using any flag parsing library for better +// portability and independence. +//===----------------------------------------------------------------------===// +FUZZER_FLAG_INT(verbosity, 1, "Verbosity level.") +FUZZER_FLAG_UNSIGNED(seed, 0, "Random seed. If 0, seed is generated.") +FUZZER_FLAG_INT(runs, -1, + "Number of individual test runs (-1 for infinite runs).") +FUZZER_FLAG_INT(max_len, 0, "Maximum length of the test input. " + "If 0, libFuzzer tries to guess a good value based on the corpus " + "and reports it. ") +FUZZER_FLAG_INT(cross_over, 1, "If 1, cross over inputs.") +FUZZER_FLAG_INT(mutate_depth, 5, + "Apply this number of consecutive mutations to each input.") +FUZZER_FLAG_INT(shuffle, 1, "Shuffle inputs at startup") +FUZZER_FLAG_INT(prefer_small, 1, + "If 1, always prefer smaller inputs during the corpus shuffle.") +FUZZER_FLAG_INT( + timeout, 1200, + "Timeout in seconds (if positive). " + "If one unit runs more than this number of seconds the process will abort.") +FUZZER_FLAG_INT(error_exitcode, 77, "When libFuzzer itself reports a bug " + "this exit code will be used.") +FUZZER_FLAG_INT(timeout_exitcode, 77, "When libFuzzer reports a timeout " + "this exit code will be used.") +FUZZER_FLAG_INT(max_total_time, 0, "If positive, indicates the maximal total " + "time in seconds to run the fuzzer.") +FUZZER_FLAG_INT(help, 0, "Print help.") +FUZZER_FLAG_INT(merge, 0, "If 1, the 2-nd, 3-rd, etc corpora will be " + "merged into the 1-st corpus. Only interesting units will be taken. " + "This flag can be used to minimize a corpus.") +FUZZER_FLAG_STRING(merge_control_file, "internal flag") +FUZZER_FLAG_INT(minimize_crash, 0, "If 1, minimizes the provided" + " crash input. Use with -runs=N or -max_total_time=N to limit " + "the number attempts") +FUZZER_FLAG_INT(minimize_crash_internal_step, 0, "internal flag") +FUZZER_FLAG_INT(use_counters, 1, "Use coverage counters") +FUZZER_FLAG_INT(use_indir_calls, 1, "Use indirect caller-callee counters") +FUZZER_FLAG_INT(use_memcmp, 1, + "Use hints from intercepting memcmp, strcmp, etc") +FUZZER_FLAG_INT(use_memmem, 1, + "Use hints from intercepting memmem, strstr, etc") +FUZZER_FLAG_INT(use_value_profile, 0, + "Experimental. Use value profile to guide fuzzing.") +FUZZER_FLAG_INT(use_cmp, 1, "Use CMP traces to guide mutations") +FUZZER_FLAG_INT(shrink, 0, "Experimental. Try to shrink corpus elements.") +FUZZER_FLAG_UNSIGNED(jobs, 0, "Number of jobs to run. If jobs >= 1 we spawn" + " this number of jobs in separate worker processes" + " with stdout/stderr redirected to fuzz-JOB.log.") +FUZZER_FLAG_UNSIGNED(workers, 0, + "Number of simultaneous worker processes to run the jobs." + " If zero, \"min(jobs,NumberOfCpuCores()/2)\" is used.") +FUZZER_FLAG_INT(reload, 1, + "Reload the main corpus every seconds to get new units" + " discovered by other processes. If 0, disabled") +FUZZER_FLAG_INT(report_slow_units, 10, + "Report slowest units if they run for more than this number of seconds.") +FUZZER_FLAG_INT(only_ascii, 0, + "If 1, generate only ASCII (isprint+isspace) inputs.") +FUZZER_FLAG_STRING(dict, "Experimental. Use the dictionary file.") +FUZZER_FLAG_STRING(artifact_prefix, "Write fuzzing artifacts (crash, " + "timeout, or slow inputs) as " + "$(artifact_prefix)file") +FUZZER_FLAG_STRING(exact_artifact_path, + "Write the single artifact on failure (crash, timeout) " + "as $(exact_artifact_path). This overrides -artifact_prefix " + "and will not use checksum in the file name. Do not " + "use the same path for several parallel processes.") +FUZZER_FLAG_INT(output_csv, 0, "Enable pulse output in CSV format.") +FUZZER_FLAG_INT(print_pcs, 0, "If 1, print out newly covered PCs.") +FUZZER_FLAG_INT(print_final_stats, 0, "If 1, print statistics at exit.") +FUZZER_FLAG_INT(print_corpus_stats, 0, + "If 1, print statistics on corpus elements at exit.") +FUZZER_FLAG_INT(print_coverage, 0, "If 1, print coverage information at exit." + " Experimental, only with trace-pc-guard") +FUZZER_FLAG_INT(dump_coverage, 0, "If 1, dump coverage information at exit." + " Experimental, only with trace-pc-guard") +FUZZER_FLAG_INT(handle_segv, 1, "If 1, try to intercept SIGSEGV.") +FUZZER_FLAG_INT(handle_bus, 1, "If 1, try to intercept SIGSEGV.") +FUZZER_FLAG_INT(handle_abrt, 1, "If 1, try to intercept SIGABRT.") +FUZZER_FLAG_INT(handle_ill, 1, "If 1, try to intercept SIGILL.") +FUZZER_FLAG_INT(handle_fpe, 1, "If 1, try to intercept SIGFPE.") +FUZZER_FLAG_INT(handle_int, 1, "If 1, try to intercept SIGINT.") +FUZZER_FLAG_INT(handle_term, 1, "If 1, try to intercept SIGTERM.") +FUZZER_FLAG_INT(close_fd_mask, 0, "If 1, close stdout at startup; " + "if 2, close stderr; if 3, close both. " + "Be careful, this will also close e.g. asan's stderr/stdout.") +FUZZER_FLAG_INT(detect_leaks, 1, "If 1, and if LeakSanitizer is enabled " + "try to detect memory leaks during fuzzing (i.e. not only at shut down).") +FUZZER_FLAG_INT(trace_malloc, 0, "If >= 1 will print all mallocs/frees. " + "If >= 2 will also print stack traces.") +FUZZER_FLAG_INT(rss_limit_mb, 2048, "If non-zero, the fuzzer will exit upon" + "reaching this limit of RSS memory usage.") +FUZZER_FLAG_STRING(exit_on_src_pos, "Exit if a newly found PC originates" + " from the given source location. Example: -exit_on_src_pos=foo.cc:123. " + "Used primarily for testing libFuzzer itself.") +FUZZER_FLAG_STRING(exit_on_item, "Exit if an item with a given sha1 sum" + " was added to the corpus. " + "Used primarily for testing libFuzzer itself.") + +FUZZER_DEPRECATED_FLAG(exit_on_first) +FUZZER_DEPRECATED_FLAG(save_minimized_corpus) +FUZZER_DEPRECATED_FLAG(sync_command) +FUZZER_DEPRECATED_FLAG(sync_timeout) +FUZZER_DEPRECATED_FLAG(test_single_input) +FUZZER_DEPRECATED_FLAG(drill) +FUZZER_DEPRECATED_FLAG(truncate_units) diff --git a/test/thirdparty/Fuzzer/FuzzerIO.cpp b/test/thirdparty/Fuzzer/FuzzerIO.cpp new file mode 100644 index 00000000..eda8e877 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerIO.cpp @@ -0,0 +1,117 @@ +//===- FuzzerIO.cpp - IO utils. -------------------------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// IO functions. +//===----------------------------------------------------------------------===// + +#include "FuzzerIO.h" +#include "FuzzerDefs.h" +#include "FuzzerExtFunctions.h" +#include +#include +#include +#include +#include +#include + +namespace fuzzer { + +static FILE *OutputFile = stderr; + +long GetEpoch(const std::string &Path) { + struct stat St; + if (stat(Path.c_str(), &St)) + return 0; // Can't stat, be conservative. + return St.st_mtime; +} + +Unit FileToVector(const std::string &Path, size_t MaxSize, bool ExitOnError) { + std::ifstream T(Path); + if (ExitOnError && !T) { + Printf("No such directory: %s; exiting\n", Path.c_str()); + exit(1); + } + + T.seekg(0, T.end); + size_t FileLen = T.tellg(); + if (MaxSize) + FileLen = std::min(FileLen, MaxSize); + + T.seekg(0, T.beg); + Unit Res(FileLen); + T.read(reinterpret_cast(Res.data()), FileLen); + return Res; +} + +std::string FileToString(const std::string &Path) { + std::ifstream T(Path); + return std::string((std::istreambuf_iterator(T)), + std::istreambuf_iterator()); +} + +void CopyFileToErr(const std::string &Path) { + Printf("%s", FileToString(Path).c_str()); +} + +void WriteToFile(const Unit &U, const std::string &Path) { + // Use raw C interface because this function may be called from a sig handler. + FILE *Out = fopen(Path.c_str(), "w"); + if (!Out) return; + fwrite(U.data(), sizeof(U[0]), U.size(), Out); + fclose(Out); +} + +void ReadDirToVectorOfUnits(const char *Path, std::vector *V, + long *Epoch, size_t MaxSize, bool ExitOnError) { + long E = Epoch ? *Epoch : 0; + std::vector Files; + ListFilesInDirRecursive(Path, Epoch, &Files, /*TopDir*/true); + size_t NumLoaded = 0; + for (size_t i = 0; i < Files.size(); i++) { + auto &X = Files[i]; + if (Epoch && GetEpoch(X) < E) continue; + NumLoaded++; + if ((NumLoaded & (NumLoaded - 1)) == 0 && NumLoaded >= 1024) + Printf("Loaded %zd/%zd files from %s\n", NumLoaded, Files.size(), Path); + auto S = FileToVector(X, MaxSize, ExitOnError); + if (!S.empty()) + V->push_back(S); + } +} + +std::string DirPlusFile(const std::string &DirPath, + const std::string &FileName) { + return DirPath + GetSeparator() + FileName; +} + +void DupAndCloseStderr() { + int OutputFd = DuplicateFile(2); + if (OutputFd > 0) { + FILE *NewOutputFile = OpenFile(OutputFd, "w"); + if (NewOutputFile) { + OutputFile = NewOutputFile; + if (EF->__sanitizer_set_report_fd) + EF->__sanitizer_set_report_fd(reinterpret_cast(OutputFd)); + CloseFile(2); + } + } +} + +void CloseStdout() { + CloseFile(1); +} + +void Printf(const char *Fmt, ...) { + va_list ap; + va_start(ap, Fmt); + vfprintf(OutputFile, Fmt, ap); + va_end(ap); + fflush(OutputFile); +} + +} // namespace fuzzer diff --git a/test/thirdparty/Fuzzer/FuzzerIO.h b/test/thirdparty/Fuzzer/FuzzerIO.h new file mode 100644 index 00000000..741fecf4 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerIO.h @@ -0,0 +1,64 @@ +//===- FuzzerIO.h - Internal header for IO utils ----------------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// IO interface. +//===----------------------------------------------------------------------===// + +#ifndef LLVM_FUZZER_IO_H +#define LLVM_FUZZER_IO_H + +#include "FuzzerDefs.h" + +namespace fuzzer { + +long GetEpoch(const std::string &Path); + +Unit FileToVector(const std::string &Path, size_t MaxSize = 0, + bool ExitOnError = true); + +std::string FileToString(const std::string &Path); + +void CopyFileToErr(const std::string &Path); + +void WriteToFile(const Unit &U, const std::string &Path); + +void ReadDirToVectorOfUnits(const char *Path, std::vector *V, + long *Epoch, size_t MaxSize, bool ExitOnError); + +// Returns "Dir/FileName" or equivalent for the current OS. +std::string DirPlusFile(const std::string &DirPath, + const std::string &FileName); + +// Returns the name of the dir, similar to the 'dirname' utility. +std::string DirName(const std::string &FileName); + +void DupAndCloseStderr(); + +void CloseStdout(); + +void Printf(const char *Fmt, ...); + +// Platform specific functions: +bool IsFile(const std::string &Path); + +void ListFilesInDirRecursive(const std::string &Dir, long *Epoch, + std::vector *V, bool TopDir); + +char GetSeparator(); + +FILE* OpenFile(int Fd, const char *Mode); + +int CloseFile(int Fd); + +int DuplicateFile(int Fd); + +void RemoveFile(const std::string &Path); + +} // namespace fuzzer + +#endif // LLVM_FUZZER_IO_H diff --git a/test/thirdparty/Fuzzer/FuzzerIOPosix.cpp b/test/thirdparty/Fuzzer/FuzzerIOPosix.cpp new file mode 100644 index 00000000..720bc130 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerIOPosix.cpp @@ -0,0 +1,88 @@ +//===- FuzzerIOPosix.cpp - IO utils for Posix. ----------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// IO functions implementation using Posix API. +//===----------------------------------------------------------------------===// +#include "FuzzerDefs.h" +#if LIBFUZZER_POSIX + +#include "FuzzerExtFunctions.h" +#include "FuzzerIO.h" +#include +#include +#include +#include +#include +#include +#include +#include +#include + +namespace fuzzer { + +bool IsFile(const std::string &Path) { + struct stat St; + if (stat(Path.c_str(), &St)) + return false; + return S_ISREG(St.st_mode); +} + +void ListFilesInDirRecursive(const std::string &Dir, long *Epoch, + std::vector *V, bool TopDir) { + auto E = GetEpoch(Dir); + if (Epoch) + if (E && *Epoch >= E) return; + + DIR *D = opendir(Dir.c_str()); + if (!D) { + Printf("No such directory: %s; exiting\n", Dir.c_str()); + exit(1); + } + while (auto E = readdir(D)) { + std::string Path = DirPlusFile(Dir, E->d_name); + if (E->d_type == DT_REG || E->d_type == DT_LNK) + V->push_back(Path); + else if (E->d_type == DT_DIR && *E->d_name != '.') + ListFilesInDirRecursive(Path, Epoch, V, false); + } + closedir(D); + if (Epoch && TopDir) + *Epoch = E; +} + +char GetSeparator() { + return '/'; +} + +FILE* OpenFile(int Fd, const char* Mode) { + return fdopen(Fd, Mode); +} + +int CloseFile(int fd) { + return close(fd); +} + +int DuplicateFile(int Fd) { + return dup(Fd); +} + +void RemoveFile(const std::string &Path) { + unlink(Path.c_str()); +} + +std::string DirName(const std::string &FileName) { + char *Tmp = new char[FileName.size() + 1]; + memcpy(Tmp, FileName.c_str(), FileName.size() + 1); + std::string Res = dirname(Tmp); + delete [] Tmp; + return Res; +} + +} // namespace fuzzer + +#endif // LIBFUZZER_POSIX diff --git a/test/thirdparty/Fuzzer/FuzzerIOWindows.cpp b/test/thirdparty/Fuzzer/FuzzerIOWindows.cpp new file mode 100644 index 00000000..a4738eb9 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerIOWindows.cpp @@ -0,0 +1,282 @@ +//===- FuzzerIOWindows.cpp - IO utils for Windows. ------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// IO functions implementation for Windows. +//===----------------------------------------------------------------------===// +#include "FuzzerDefs.h" +#if LIBFUZZER_WINDOWS + +#include "FuzzerExtFunctions.h" +#include "FuzzerIO.h" +#include +#include +#include +#include +#include +#include +#include +#include + +namespace fuzzer { + +static bool IsFile(const std::string &Path, const DWORD &FileAttributes) { + + if (FileAttributes & FILE_ATTRIBUTE_NORMAL) + return true; + + if (FileAttributes & FILE_ATTRIBUTE_DIRECTORY) + return false; + + HANDLE FileHandle( + CreateFileA(Path.c_str(), 0, FILE_SHARE_READ, NULL, OPEN_EXISTING, + FILE_FLAG_BACKUP_SEMANTICS, 0)); + + if (FileHandle == INVALID_HANDLE_VALUE) { + Printf("CreateFileA() failed for \"%s\" (Error code: %lu).\n", Path.c_str(), + GetLastError()); + return false; + } + + DWORD FileType = GetFileType(FileHandle); + + if (FileType == FILE_TYPE_UNKNOWN) { + Printf("GetFileType() failed for \"%s\" (Error code: %lu).\n", Path.c_str(), + GetLastError()); + CloseHandle(FileHandle); + return false; + } + + if (FileType != FILE_TYPE_DISK) { + CloseHandle(FileHandle); + return false; + } + + CloseHandle(FileHandle); + return true; +} + +bool IsFile(const std::string &Path) { + DWORD Att = GetFileAttributesA(Path.c_str()); + + if (Att == INVALID_FILE_ATTRIBUTES) { + Printf("GetFileAttributesA() failed for \"%s\" (Error code: %lu).\n", + Path.c_str(), GetLastError()); + return false; + } + + return IsFile(Path, Att); +} + +void ListFilesInDirRecursive(const std::string &Dir, long *Epoch, + std::vector *V, bool TopDir) { + auto E = GetEpoch(Dir); + if (Epoch) + if (E && *Epoch >= E) return; + + std::string Path(Dir); + assert(!Path.empty()); + if (Path.back() != '\\') + Path.push_back('\\'); + Path.push_back('*'); + + // Get the first directory entry. + WIN32_FIND_DATAA FindInfo; + HANDLE FindHandle(FindFirstFileA(Path.c_str(), &FindInfo)); + if (FindHandle == INVALID_HANDLE_VALUE) + { + Printf("No file found in: %s.\n", Dir.c_str()); + return; + } + + do { + std::string FileName = DirPlusFile(Dir, FindInfo.cFileName); + + if (FindInfo.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) { + size_t FilenameLen = strlen(FindInfo.cFileName); + if ((FilenameLen == 1 && FindInfo.cFileName[0] == '.') || + (FilenameLen == 2 && FindInfo.cFileName[0] == '.' && + FindInfo.cFileName[1] == '.')) + continue; + + ListFilesInDirRecursive(FileName, Epoch, V, false); + } + else if (IsFile(FileName, FindInfo.dwFileAttributes)) + V->push_back(FileName); + } while (FindNextFileA(FindHandle, &FindInfo)); + + DWORD LastError = GetLastError(); + if (LastError != ERROR_NO_MORE_FILES) + Printf("FindNextFileA failed (Error code: %lu).\n", LastError); + + FindClose(FindHandle); + + if (Epoch && TopDir) + *Epoch = E; +} + +char GetSeparator() { + return '\\'; +} + +FILE* OpenFile(int Fd, const char* Mode) { + return _fdopen(Fd, Mode); +} + +int CloseFile(int Fd) { + return _close(Fd); +} + +int DuplicateFile(int Fd) { + return _dup(Fd); +} + +void RemoveFile(const std::string &Path) { + _unlink(Path.c_str()); +} + +static bool IsSeparator(char C) { + return C == '\\' || C == '/'; +} + +// Parse disk designators, like "C:\". If Relative == true, also accepts: "C:". +// Returns number of characters considered if successful. +static size_t ParseDrive(const std::string &FileName, const size_t Offset, + bool Relative = true) { + if (Offset + 1 >= FileName.size() || FileName[Offset + 1] != ':') + return 0; + if (Offset + 2 >= FileName.size() || !IsSeparator(FileName[Offset + 2])) { + if (!Relative) // Accept relative path? + return 0; + else + return 2; + } + return 3; +} + +// Parse a file name, like: SomeFile.txt +// Returns number of characters considered if successful. +static size_t ParseFileName(const std::string &FileName, const size_t Offset) { + size_t Pos = Offset; + const size_t End = FileName.size(); + for(; Pos < End && !IsSeparator(FileName[Pos]); ++Pos) + ; + return Pos - Offset; +} + +// Parse a directory ending in separator, like: SomeDir\ +// Returns number of characters considered if successful. +static size_t ParseDir(const std::string &FileName, const size_t Offset) { + size_t Pos = Offset; + const size_t End = FileName.size(); + if (Pos >= End || IsSeparator(FileName[Pos])) + return 0; + for(; Pos < End && !IsSeparator(FileName[Pos]); ++Pos) + ; + if (Pos >= End) + return 0; + ++Pos; // Include separator. + return Pos - Offset; +} + +// Parse a servername and share, like: SomeServer\SomeShare\ +// Returns number of characters considered if successful. +static size_t ParseServerAndShare(const std::string &FileName, + const size_t Offset) { + size_t Pos = Offset, Res; + if (!(Res = ParseDir(FileName, Pos))) + return 0; + Pos += Res; + if (!(Res = ParseDir(FileName, Pos))) + return 0; + Pos += Res; + return Pos - Offset; +} + +// Parse the given Ref string from the position Offset, to exactly match the given +// string Patt. +// Returns number of characters considered if successful. +static size_t ParseCustomString(const std::string &Ref, size_t Offset, + const char *Patt) { + size_t Len = strlen(Patt); + if (Offset + Len > Ref.size()) + return 0; + return Ref.compare(Offset, Len, Patt) == 0 ? Len : 0; +} + +// Parse a location, like: +// \\?\UNC\Server\Share\ \\?\C:\ \\Server\Share\ \ C:\ C: +// Returns number of characters considered if successful. +static size_t ParseLocation(const std::string &FileName) { + size_t Pos = 0, Res; + + if ((Res = ParseCustomString(FileName, Pos, R"(\\?\)"))) { + Pos += Res; + if ((Res = ParseCustomString(FileName, Pos, R"(UNC\)"))) { + Pos += Res; + if ((Res = ParseServerAndShare(FileName, Pos))) + return Pos + Res; + return 0; + } + if ((Res = ParseDrive(FileName, Pos, false))) + return Pos + Res; + return 0; + } + + if (Pos < FileName.size() && IsSeparator(FileName[Pos])) { + ++Pos; + if (Pos < FileName.size() && IsSeparator(FileName[Pos])) { + ++Pos; + if ((Res = ParseServerAndShare(FileName, Pos))) + return Pos + Res; + return 0; + } + return Pos; + } + + if ((Res = ParseDrive(FileName, Pos))) + return Pos + Res; + + return Pos; +} + +std::string DirName(const std::string &FileName) { + size_t LocationLen = ParseLocation(FileName); + size_t DirLen = 0, Res; + while ((Res = ParseDir(FileName, LocationLen + DirLen))) + DirLen += Res; + size_t FileLen = ParseFileName(FileName, LocationLen + DirLen); + + if (LocationLen + DirLen + FileLen != FileName.size()) { + Printf("DirName() failed for \"%s\", invalid path.\n", FileName.c_str()); + exit(1); + } + + if (DirLen) { + --DirLen; // Remove trailing separator. + if (!FileLen) { // Path ended in separator. + assert(DirLen); + // Remove file name from Dir. + while (DirLen && !IsSeparator(FileName[LocationLen + DirLen - 1])) + --DirLen; + if (DirLen) // Remove trailing separator. + --DirLen; + } + } + + if (!LocationLen) { // Relative path. + if (!DirLen) + return "."; + return std::string(".\\").append(FileName, 0, DirLen); + } + + return FileName.substr(0, LocationLen + DirLen); +} + +} // namespace fuzzer + +#endif // LIBFUZZER_WINDOWS diff --git a/test/thirdparty/Fuzzer/FuzzerInterface.h b/test/thirdparty/Fuzzer/FuzzerInterface.h new file mode 100644 index 00000000..d47e20e3 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerInterface.h @@ -0,0 +1,67 @@ +//===- FuzzerInterface.h - Interface header for the Fuzzer ------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Define the interface between libFuzzer and the library being tested. +//===----------------------------------------------------------------------===// + +// NOTE: the libFuzzer interface is thin and in the majority of cases +// you should not include this file into your target. In 95% of cases +// all you need is to define the following function in your file: +// extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size); + +// WARNING: keep the interface in C. + +#ifndef LLVM_FUZZER_INTERFACE_H +#define LLVM_FUZZER_INTERFACE_H + +#include +#include + +#ifdef __cplusplus +extern "C" { +#endif // __cplusplus + +// Mandatory user-provided target function. +// Executes the code under test with [Data, Data+Size) as the input. +// libFuzzer will invoke this function *many* times with different inputs. +// Must return 0. +int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size); + +// Optional user-provided initialization function. +// If provided, this function will be called by libFuzzer once at startup. +// It may read and modify argc/argv. +// Must return 0. +int LLVMFuzzerInitialize(int *argc, char ***argv); + +// Optional user-provided custom mutator. +// Mutates raw data in [Data, Data+Size) inplace. +// Returns the new size, which is not greater than MaxSize. +// Given the same Seed produces the same mutation. +size_t LLVMFuzzerCustomMutator(uint8_t *Data, size_t Size, size_t MaxSize, + unsigned int Seed); + +// Optional user-provided custom cross-over function. +// Combines pieces of Data1 & Data2 together into Out. +// Returns the new size, which is not greater than MaxOutSize. +// Should produce the same mutation given the same Seed. +size_t LLVMFuzzerCustomCrossOver(const uint8_t *Data1, size_t Size1, + const uint8_t *Data2, size_t Size2, + uint8_t *Out, size_t MaxOutSize, + unsigned int Seed); + +// Experimental, may go away in future. +// libFuzzer-provided function to be used inside LLVMFuzzerTestOneInput. +// Mutates raw data in [Data, Data+Size) inplace. +// Returns the new size, which is not greater than MaxSize. +size_t LLVMFuzzerMutate(uint8_t *Data, size_t Size, size_t MaxSize); + +#ifdef __cplusplus +} // extern "C" +#endif // __cplusplus + +#endif // LLVM_FUZZER_INTERFACE_H diff --git a/test/thirdparty/Fuzzer/FuzzerInternal.h b/test/thirdparty/Fuzzer/FuzzerInternal.h new file mode 100644 index 00000000..c0417060 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerInternal.h @@ -0,0 +1,182 @@ +//===- FuzzerInternal.h - Internal header for the Fuzzer --------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Define the main class fuzzer::Fuzzer and most functions. +//===----------------------------------------------------------------------===// + +#ifndef LLVM_FUZZER_INTERNAL_H +#define LLVM_FUZZER_INTERNAL_H + +#include "FuzzerDefs.h" +#include "FuzzerExtFunctions.h" +#include "FuzzerInterface.h" +#include "FuzzerOptions.h" +#include "FuzzerSHA1.h" +#include "FuzzerValueBitMap.h" +#include +#include +#include +#include +#include +#include + +namespace fuzzer { + +using namespace std::chrono; + +class Fuzzer { +public: + + // Aggregates all available coverage measurements. + struct Coverage { + Coverage() { Reset(); } + + void Reset() { + BlockCoverage = 0; + CallerCalleeCoverage = 0; + CounterBitmapBits = 0; + CounterBitmap.clear(); + VPMap.Reset(); + } + + size_t BlockCoverage; + size_t CallerCalleeCoverage; + // Precalculated number of bits in CounterBitmap. + size_t CounterBitmapBits; + std::vector CounterBitmap; + ValueBitMap VPMap; + }; + + Fuzzer(UserCallback CB, InputCorpus &Corpus, MutationDispatcher &MD, + FuzzingOptions Options); + ~Fuzzer(); + void Loop(); + void MinimizeCrashLoop(const Unit &U); + void ShuffleAndMinimize(UnitVector *V); + void InitializeTraceState(); + void RereadOutputCorpus(size_t MaxSize); + + size_t secondsSinceProcessStartUp() { + return duration_cast(system_clock::now() - ProcessStartTime) + .count(); + } + + bool TimedOut() { + return Options.MaxTotalTimeSec > 0 && + secondsSinceProcessStartUp() > + static_cast(Options.MaxTotalTimeSec); + } + + size_t execPerSec() { + size_t Seconds = secondsSinceProcessStartUp(); + return Seconds ? TotalNumberOfRuns / Seconds : 0; + } + + size_t getTotalNumberOfRuns() { return TotalNumberOfRuns; } + + static void StaticAlarmCallback(); + static void StaticCrashSignalCallback(); + static void StaticInterruptCallback(); + + void ExecuteCallback(const uint8_t *Data, size_t Size); + size_t RunOne(const uint8_t *Data, size_t Size); + + // Merge Corpora[1:] into Corpora[0]. + void Merge(const std::vector &Corpora); + void CrashResistantMerge(const std::vector &Args, + const std::vector &Corpora); + void CrashResistantMergeInternalStep(const std::string &ControlFilePath); + // Returns a subset of 'Extra' that adds coverage to 'Initial'. + UnitVector FindExtraUnits(const UnitVector &Initial, const UnitVector &Extra); + MutationDispatcher &GetMD() { return MD; } + void PrintFinalStats(); + void SetMaxInputLen(size_t MaxInputLen); + void SetMaxMutationLen(size_t MaxMutationLen); + void RssLimitCallback(); + + // Public for tests. + void ResetCoverage(); + + bool InFuzzingThread() const { return IsMyThread; } + size_t GetCurrentUnitInFuzzingThead(const uint8_t **Data) const; + void TryDetectingAMemoryLeak(const uint8_t *Data, size_t Size, + bool DuringInitialCorpusExecution); + + void HandleMalloc(size_t Size); + +private: + void AlarmCallback(); + void CrashCallback(); + void InterruptCallback(); + void MutateAndTestOne(); + void ReportNewCoverage(InputInfo *II, const Unit &U); + size_t RunOne(const Unit &U) { return RunOne(U.data(), U.size()); } + void WriteToOutputCorpus(const Unit &U); + void WriteUnitToFileWithPrefix(const Unit &U, const char *Prefix); + void PrintStats(const char *Where, const char *End = "\n", size_t Units = 0); + void PrintStatusForNewUnit(const Unit &U); + void ShuffleCorpus(UnitVector *V); + void AddToCorpus(const Unit &U); + void CheckExitOnSrcPosOrItem(); + + // Trace-based fuzzing: we run a unit with some kind of tracing + // enabled and record potentially useful mutations. Then + // We apply these mutations one by one to the unit and run it again. + + // Start tracing; forget all previously proposed mutations. + void StartTraceRecording(); + // Stop tracing. + void StopTraceRecording(); + + void SetDeathCallback(); + static void StaticDeathCallback(); + void DumpCurrentUnit(const char *Prefix); + void DeathCallback(); + + void ResetEdgeCoverage(); + void ResetCounters(); + void PrepareCounters(Fuzzer::Coverage *C); + bool RecordMaxCoverage(Fuzzer::Coverage *C); + + void AllocateCurrentUnitData(); + uint8_t *CurrentUnitData = nullptr; + std::atomic CurrentUnitSize; + uint8_t BaseSha1[kSHA1NumBytes]; // Checksum of the base unit. + bool RunningCB = false; + + size_t TotalNumberOfRuns = 0; + size_t NumberOfNewUnitsAdded = 0; + + bool HasMoreMallocsThanFrees = false; + size_t NumberOfLeakDetectionAttempts = 0; + + UserCallback CB; + InputCorpus &Corpus; + MutationDispatcher &MD; + FuzzingOptions Options; + + system_clock::time_point ProcessStartTime = system_clock::now(); + system_clock::time_point UnitStartTime, UnitStopTime; + long TimeOfLongestUnitInSeconds = 0; + long EpochOfLastReadOfOutputCorpus = 0; + + // Maximum recorded coverage. + Coverage MaxCoverage; + + size_t MaxInputLen = 0; + size_t MaxMutationLen = 0; + + // Need to know our own thread. + static thread_local bool IsMyThread; + + bool InMergeMode = false; +}; + +}; // namespace fuzzer + +#endif // LLVM_FUZZER_INTERNAL_H diff --git a/test/thirdparty/Fuzzer/FuzzerLoop.cpp b/test/thirdparty/Fuzzer/FuzzerLoop.cpp new file mode 100644 index 00000000..73e058ff --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerLoop.cpp @@ -0,0 +1,792 @@ +//===- FuzzerLoop.cpp - Fuzzer's main loop --------------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Fuzzer's main loop. +//===----------------------------------------------------------------------===// + +#include "FuzzerCorpus.h" +#include "FuzzerInternal.h" +#include "FuzzerIO.h" +#include "FuzzerMutate.h" +#include "FuzzerRandom.h" +#include "FuzzerTracePC.h" +#include +#include +#include +#include + +#if defined(__has_include) +#if __has_include() +#include +#endif +#if __has_include() +#include +#endif +#endif + +#define NO_SANITIZE_MEMORY +#if defined(__has_feature) +#if __has_feature(memory_sanitizer) +#undef NO_SANITIZE_MEMORY +#define NO_SANITIZE_MEMORY __attribute__((no_sanitize_memory)) +#endif +#endif + +namespace fuzzer { +static const size_t kMaxUnitSizeToPrint = 256; + +thread_local bool Fuzzer::IsMyThread; + +static void MissingExternalApiFunction(const char *FnName) { + Printf("ERROR: %s is not defined. Exiting.\n" + "Did you use -fsanitize-coverage=... to build your code?\n", + FnName); + exit(1); +} + +#define CHECK_EXTERNAL_FUNCTION(fn) \ + do { \ + if (!(EF->fn)) \ + MissingExternalApiFunction(#fn); \ + } while (false) + +// Only one Fuzzer per process. +static Fuzzer *F; + +void Fuzzer::ResetEdgeCoverage() { + CHECK_EXTERNAL_FUNCTION(__sanitizer_reset_coverage); + EF->__sanitizer_reset_coverage(); +} + +void Fuzzer::ResetCounters() { + if (Options.UseCounters) + EF->__sanitizer_update_counter_bitset_and_clear_counters(0); +} + +void Fuzzer::PrepareCounters(Fuzzer::Coverage *C) { + if (Options.UseCounters) { + size_t NumCounters = EF->__sanitizer_get_number_of_counters(); + C->CounterBitmap.resize(NumCounters); + } +} + +// Records data to a maximum coverage tracker. Returns true if additional +// coverage was discovered. +bool Fuzzer::RecordMaxCoverage(Fuzzer::Coverage *C) { + bool Res = false; + + uint64_t NewBlockCoverage = EF->__sanitizer_get_total_unique_coverage(); + if (NewBlockCoverage > C->BlockCoverage) { + Res = true; + C->BlockCoverage = NewBlockCoverage; + } + + if (Options.UseIndirCalls && + EF->__sanitizer_get_total_unique_caller_callee_pairs) { + uint64_t NewCallerCalleeCoverage = + EF->__sanitizer_get_total_unique_caller_callee_pairs(); + if (NewCallerCalleeCoverage > C->CallerCalleeCoverage) { + Res = true; + C->CallerCalleeCoverage = NewCallerCalleeCoverage; + } + } + + if (Options.UseCounters) { + uint64_t CounterDelta = + EF->__sanitizer_update_counter_bitset_and_clear_counters( + C->CounterBitmap.data()); + if (CounterDelta > 0) { + Res = true; + C->CounterBitmapBits += CounterDelta; + } + } + + return Res; +} + +// Leak detection is expensive, so we first check if there were more mallocs +// than frees (using the sanitizer malloc hooks) and only then try to call lsan. +struct MallocFreeTracer { + void Start(int TraceLevel) { + this->TraceLevel = TraceLevel; + if (TraceLevel) + Printf("MallocFreeTracer: START\n"); + Mallocs = 0; + Frees = 0; + } + // Returns true if there were more mallocs than frees. + bool Stop() { + if (TraceLevel) + Printf("MallocFreeTracer: STOP %zd %zd (%s)\n", Mallocs.load(), + Frees.load(), Mallocs == Frees ? "same" : "DIFFERENT"); + bool Result = Mallocs > Frees; + Mallocs = 0; + Frees = 0; + TraceLevel = 0; + return Result; + } + std::atomic Mallocs; + std::atomic Frees; + int TraceLevel = 0; +}; + +static MallocFreeTracer AllocTracer; + +ATTRIBUTE_NO_SANITIZE_MEMORY +void MallocHook(const volatile void *ptr, size_t size) { + size_t N = AllocTracer.Mallocs++; + F->HandleMalloc(size); + if (int TraceLevel = AllocTracer.TraceLevel) { + Printf("MALLOC[%zd] %p %zd\n", N, ptr, size); + if (TraceLevel >= 2 && EF) + EF->__sanitizer_print_stack_trace(); + } +} + +ATTRIBUTE_NO_SANITIZE_MEMORY +void FreeHook(const volatile void *ptr) { + size_t N = AllocTracer.Frees++; + if (int TraceLevel = AllocTracer.TraceLevel) { + Printf("FREE[%zd] %p\n", N, ptr); + if (TraceLevel >= 2 && EF) + EF->__sanitizer_print_stack_trace(); + } +} + +// Crash on a single malloc that exceeds the rss limit. +void Fuzzer::HandleMalloc(size_t Size) { + if (!Options.RssLimitMb || (Size >> 20) < (size_t)Options.RssLimitMb) + return; + Printf("==%d== ERROR: libFuzzer: out-of-memory (malloc(%zd))\n", GetPid(), + Size); + Printf(" To change the out-of-memory limit use -rss_limit_mb=\n\n"); + if (EF->__sanitizer_print_stack_trace) + EF->__sanitizer_print_stack_trace(); + DumpCurrentUnit("oom-"); + Printf("SUMMARY: libFuzzer: out-of-memory\n"); + PrintFinalStats(); + _Exit(Options.ErrorExitCode); // Stop right now. +} + +Fuzzer::Fuzzer(UserCallback CB, InputCorpus &Corpus, MutationDispatcher &MD, + FuzzingOptions Options) + : CB(CB), Corpus(Corpus), MD(MD), Options(Options) { + SetDeathCallback(); + InitializeTraceState(); + assert(!F); + F = this; + TPC.ResetMaps(); + ResetCoverage(); + IsMyThread = true; + if (Options.DetectLeaks && EF->__sanitizer_install_malloc_and_free_hooks) + EF->__sanitizer_install_malloc_and_free_hooks(MallocHook, FreeHook); + TPC.SetUseCounters(Options.UseCounters); + TPC.SetUseValueProfile(Options.UseValueProfile); + TPC.SetPrintNewPCs(Options.PrintNewCovPcs); + + if (Options.Verbosity) + TPC.PrintModuleInfo(); + if (!Options.OutputCorpus.empty() && Options.ReloadIntervalSec) + EpochOfLastReadOfOutputCorpus = GetEpoch(Options.OutputCorpus); + MaxInputLen = MaxMutationLen = Options.MaxLen; + AllocateCurrentUnitData(); + CurrentUnitSize = 0; + memset(BaseSha1, 0, sizeof(BaseSha1)); +} + +Fuzzer::~Fuzzer() { } + +void Fuzzer::AllocateCurrentUnitData() { + if (CurrentUnitData || MaxInputLen == 0) return; + CurrentUnitData = new uint8_t[MaxInputLen]; +} + +void Fuzzer::SetDeathCallback() { + CHECK_EXTERNAL_FUNCTION(__sanitizer_set_death_callback); + EF->__sanitizer_set_death_callback(StaticDeathCallback); +} + +void Fuzzer::StaticDeathCallback() { + assert(F); + F->DeathCallback(); +} + +static void WarnOnUnsuccessfullMerge(bool DoWarn) { + if (!DoWarn) return; + Printf( + "***\n" + "***\n" + "***\n" + "*** NOTE: merge did not succeed due to a failure on one of the inputs.\n" + "*** You will need to filter out crashes from the corpus, e.g. like this:\n" + "*** for f in WITH_CRASHES/*; do ./fuzzer $f && cp $f NO_CRASHES; done\n" + "*** Future versions may have crash-resistant merge, stay tuned.\n" + "***\n" + "***\n" + "***\n"); +} + +void Fuzzer::DumpCurrentUnit(const char *Prefix) { + WarnOnUnsuccessfullMerge(InMergeMode); + if (!CurrentUnitData) return; // Happens when running individual inputs. + MD.PrintMutationSequence(); + Printf("; base unit: %s\n", Sha1ToString(BaseSha1).c_str()); + size_t UnitSize = CurrentUnitSize; + if (UnitSize <= kMaxUnitSizeToPrint) { + PrintHexArray(CurrentUnitData, UnitSize, "\n"); + PrintASCII(CurrentUnitData, UnitSize, "\n"); + } + WriteUnitToFileWithPrefix({CurrentUnitData, CurrentUnitData + UnitSize}, + Prefix); +} + +NO_SANITIZE_MEMORY +void Fuzzer::DeathCallback() { + DumpCurrentUnit("crash-"); + PrintFinalStats(); +} + +void Fuzzer::StaticAlarmCallback() { + assert(F); + F->AlarmCallback(); +} + +void Fuzzer::StaticCrashSignalCallback() { + assert(F); + F->CrashCallback(); +} + +void Fuzzer::StaticInterruptCallback() { + assert(F); + F->InterruptCallback(); +} + +void Fuzzer::CrashCallback() { + Printf("==%lu== ERROR: libFuzzer: deadly signal\n", GetPid()); + if (EF->__sanitizer_print_stack_trace) + EF->__sanitizer_print_stack_trace(); + Printf("NOTE: libFuzzer has rudimentary signal handlers.\n" + " Combine libFuzzer with AddressSanitizer or similar for better " + "crash reports.\n"); + Printf("SUMMARY: libFuzzer: deadly signal\n"); + DumpCurrentUnit("crash-"); + PrintFinalStats(); + exit(Options.ErrorExitCode); +} + +void Fuzzer::InterruptCallback() { + Printf("==%lu== libFuzzer: run interrupted; exiting\n", GetPid()); + PrintFinalStats(); + _Exit(0); // Stop right now, don't perform any at-exit actions. +} + +NO_SANITIZE_MEMORY +void Fuzzer::AlarmCallback() { + assert(Options.UnitTimeoutSec > 0); + if (!InFuzzingThread()) return; + if (!RunningCB) + return; // We have not started running units yet. + size_t Seconds = + duration_cast(system_clock::now() - UnitStartTime).count(); + if (Seconds == 0) + return; + if (Options.Verbosity >= 2) + Printf("AlarmCallback %zd\n", Seconds); + if (Seconds >= (size_t)Options.UnitTimeoutSec) { + Printf("ALARM: working on the last Unit for %zd seconds\n", Seconds); + Printf(" and the timeout value is %d (use -timeout=N to change)\n", + Options.UnitTimeoutSec); + DumpCurrentUnit("timeout-"); + Printf("==%lu== ERROR: libFuzzer: timeout after %d seconds\n", GetPid(), + Seconds); + if (EF->__sanitizer_print_stack_trace) + EF->__sanitizer_print_stack_trace(); + Printf("SUMMARY: libFuzzer: timeout\n"); + PrintFinalStats(); + _Exit(Options.TimeoutExitCode); // Stop right now. + } +} + +void Fuzzer::RssLimitCallback() { + Printf( + "==%lu== ERROR: libFuzzer: out-of-memory (used: %zdMb; limit: %zdMb)\n", + GetPid(), GetPeakRSSMb(), Options.RssLimitMb); + Printf(" To change the out-of-memory limit use -rss_limit_mb=\n\n"); + if (EF->__sanitizer_print_memory_profile) + EF->__sanitizer_print_memory_profile(95); + DumpCurrentUnit("oom-"); + Printf("SUMMARY: libFuzzer: out-of-memory\n"); + PrintFinalStats(); + _Exit(Options.ErrorExitCode); // Stop right now. +} + +void Fuzzer::PrintStats(const char *Where, const char *End, size_t Units) { + size_t ExecPerSec = execPerSec(); + if (Options.OutputCSV) { + static bool csvHeaderPrinted = false; + if (!csvHeaderPrinted) { + csvHeaderPrinted = true; + Printf("runs,block_cov,bits,cc_cov,corpus,execs_per_sec,tbms,reason\n"); + } + Printf("%zd,%zd,%zd,%zd,%zd,%zd,%s\n", TotalNumberOfRuns, + MaxCoverage.BlockCoverage, MaxCoverage.CounterBitmapBits, + MaxCoverage.CallerCalleeCoverage, Corpus.size(), ExecPerSec, Where); + } + + if (!Options.Verbosity) + return; + Printf("#%zd\t%s", TotalNumberOfRuns, Where); + if (MaxCoverage.BlockCoverage) + Printf(" cov: %zd", MaxCoverage.BlockCoverage); + if (size_t N = MaxCoverage.VPMap.GetNumBitsSinceLastMerge()) + Printf(" vp: %zd", N); + if (size_t N = TPC.GetTotalPCCoverage()) + Printf(" cov: %zd", N); + if (auto TB = MaxCoverage.CounterBitmapBits) + Printf(" bits: %zd", TB); + if (size_t N = Corpus.NumFeatures()) + Printf( " ft: %zd", N); + if (MaxCoverage.CallerCalleeCoverage) + Printf(" indir: %zd", MaxCoverage.CallerCalleeCoverage); + if (!Corpus.empty()) { + Printf(" corp: %zd", Corpus.NumActiveUnits()); + if (size_t N = Corpus.SizeInBytes()) { + if (N < (1<<14)) + Printf("/%zdb", N); + else if (N < (1 << 24)) + Printf("/%zdKb", N >> 10); + else + Printf("/%zdMb", N >> 20); + } + } + if (Units) + Printf(" units: %zd", Units); + + Printf(" exec/s: %zd", ExecPerSec); + Printf(" rss: %zdMb", GetPeakRSSMb()); + Printf("%s", End); +} + +void Fuzzer::PrintFinalStats() { + if (Options.PrintCoverage) + TPC.PrintCoverage(); + if (Options.DumpCoverage) + TPC.DumpCoverage(); + if (Options.PrintCorpusStats) + Corpus.PrintStats(); + if (!Options.PrintFinalStats) return; + size_t ExecPerSec = execPerSec(); + Printf("stat::number_of_executed_units: %zd\n", TotalNumberOfRuns); + Printf("stat::average_exec_per_sec: %zd\n", ExecPerSec); + Printf("stat::new_units_added: %zd\n", NumberOfNewUnitsAdded); + Printf("stat::slowest_unit_time_sec: %zd\n", TimeOfLongestUnitInSeconds); + Printf("stat::peak_rss_mb: %zd\n", GetPeakRSSMb()); +} + +void Fuzzer::SetMaxInputLen(size_t MaxInputLen) { + assert(this->MaxInputLen == 0); // Can only reset MaxInputLen from 0 to non-0. + assert(MaxInputLen); + this->MaxInputLen = MaxInputLen; + this->MaxMutationLen = MaxInputLen; + AllocateCurrentUnitData(); + Printf("INFO: -max_len is not provided, using %zd\n", MaxInputLen); +} + +void Fuzzer::SetMaxMutationLen(size_t MaxMutationLen) { + assert(MaxMutationLen && MaxMutationLen <= MaxInputLen); + this->MaxMutationLen = MaxMutationLen; +} + +void Fuzzer::CheckExitOnSrcPosOrItem() { + if (!Options.ExitOnSrcPos.empty()) { + static auto *PCsSet = new std::set; + for (size_t i = 1, N = TPC.GetNumPCs(); i < N; i++) { + uintptr_t PC = TPC.GetPC(i); + if (!PC) continue; + if (!PCsSet->insert(PC).second) continue; + std::string Descr = DescribePC("%L", PC); + if (Descr.find(Options.ExitOnSrcPos) != std::string::npos) { + Printf("INFO: found line matching '%s', exiting.\n", + Options.ExitOnSrcPos.c_str()); + _Exit(0); + } + } + } + if (!Options.ExitOnItem.empty()) { + if (Corpus.HasUnit(Options.ExitOnItem)) { + Printf("INFO: found item with checksum '%s', exiting.\n", + Options.ExitOnItem.c_str()); + _Exit(0); + } + } +} + +void Fuzzer::RereadOutputCorpus(size_t MaxSize) { + if (Options.OutputCorpus.empty() || !Options.ReloadIntervalSec) return; + std::vector AdditionalCorpus; + ReadDirToVectorOfUnits(Options.OutputCorpus.c_str(), &AdditionalCorpus, + &EpochOfLastReadOfOutputCorpus, MaxSize, + /*ExitOnError*/ false); + if (Options.Verbosity >= 2) + Printf("Reload: read %zd new units.\n", AdditionalCorpus.size()); + bool Reloaded = false; + for (auto &U : AdditionalCorpus) { + if (U.size() > MaxSize) + U.resize(MaxSize); + if (!Corpus.HasUnit(U)) { + if (size_t NumFeatures = RunOne(U)) { + CheckExitOnSrcPosOrItem(); + Corpus.AddToCorpus(U, NumFeatures); + Reloaded = true; + } + } + } + if (Reloaded) + PrintStats("RELOAD"); +} + +void Fuzzer::ShuffleCorpus(UnitVector *V) { + std::random_shuffle(V->begin(), V->end(), MD.GetRand()); + if (Options.PreferSmall) + std::stable_sort(V->begin(), V->end(), [](const Unit &A, const Unit &B) { + return A.size() < B.size(); + }); +} + +void Fuzzer::ShuffleAndMinimize(UnitVector *InitialCorpus) { + Printf("#0\tREAD units: %zd\n", InitialCorpus->size()); + if (Options.ShuffleAtStartUp) + ShuffleCorpus(InitialCorpus); + + // Test the callback with empty input and never try it again. + uint8_t dummy; + ExecuteCallback(&dummy, 0); + + for (const auto &U : *InitialCorpus) { + if (size_t NumFeatures = RunOne(U)) { + CheckExitOnSrcPosOrItem(); + Corpus.AddToCorpus(U, NumFeatures); + if (Options.Verbosity >= 2) + Printf("NEW0: %zd L %zd\n", MaxCoverage.BlockCoverage, U.size()); + } + TryDetectingAMemoryLeak(U.data(), U.size(), + /*DuringInitialCorpusExecution*/ true); + } + PrintStats("INITED"); + if (Corpus.empty()) { + Printf("ERROR: no interesting inputs were found. " + "Is the code instrumented for coverage? Exiting.\n"); + exit(1); + } +} + +size_t Fuzzer::RunOne(const uint8_t *Data, size_t Size) { + if (!Size) return 0; + TotalNumberOfRuns++; + + ExecuteCallback(Data, Size); + + size_t Res = 0; + if (size_t NumFeatures = TPC.CollectFeatures([&](size_t Feature) -> bool { + return Corpus.AddFeature(Feature, Size, Options.Shrink); + })) + Res = NumFeatures; + + if (!TPC.UsingTracePcGuard()) { + if (TPC.UpdateValueProfileMap(&MaxCoverage.VPMap)) + Res = 1; + if (!Res && RecordMaxCoverage(&MaxCoverage)) + Res = 1; + } + + auto TimeOfUnit = + duration_cast(UnitStopTime - UnitStartTime).count(); + if (!(TotalNumberOfRuns & (TotalNumberOfRuns - 1)) && + secondsSinceProcessStartUp() >= 2) + PrintStats("pulse "); + if (TimeOfUnit > TimeOfLongestUnitInSeconds * 1.1 && + TimeOfUnit >= Options.ReportSlowUnits) { + TimeOfLongestUnitInSeconds = TimeOfUnit; + Printf("Slowest unit: %zd s:\n", TimeOfLongestUnitInSeconds); + WriteUnitToFileWithPrefix({Data, Data + Size}, "slow-unit-"); + } + return Res; +} + +size_t Fuzzer::GetCurrentUnitInFuzzingThead(const uint8_t **Data) const { + assert(InFuzzingThread()); + *Data = CurrentUnitData; + return CurrentUnitSize; +} + +void Fuzzer::ExecuteCallback(const uint8_t *Data, size_t Size) { + assert(InFuzzingThread()); + // We copy the contents of Unit into a separate heap buffer + // so that we reliably find buffer overflows in it. + uint8_t *DataCopy = new uint8_t[Size]; + memcpy(DataCopy, Data, Size); + if (CurrentUnitData && CurrentUnitData != Data) + memcpy(CurrentUnitData, Data, Size); + CurrentUnitSize = Size; + AllocTracer.Start(Options.TraceMalloc); + UnitStartTime = system_clock::now(); + ResetCounters(); // Reset coverage right before the callback. + TPC.ResetMaps(); + RunningCB = true; + int Res = CB(DataCopy, Size); + RunningCB = false; + UnitStopTime = system_clock::now(); + (void)Res; + assert(Res == 0); + HasMoreMallocsThanFrees = AllocTracer.Stop(); + CurrentUnitSize = 0; + delete[] DataCopy; +} + +void Fuzzer::WriteToOutputCorpus(const Unit &U) { + if (Options.OnlyASCII) + assert(IsASCII(U)); + if (Options.OutputCorpus.empty()) + return; + std::string Path = DirPlusFile(Options.OutputCorpus, Hash(U)); + WriteToFile(U, Path); + if (Options.Verbosity >= 2) + Printf("Written to %s\n", Path.c_str()); +} + +void Fuzzer::WriteUnitToFileWithPrefix(const Unit &U, const char *Prefix) { + if (!Options.SaveArtifacts) + return; + std::string Path = Options.ArtifactPrefix + Prefix + Hash(U); + if (!Options.ExactArtifactPath.empty()) + Path = Options.ExactArtifactPath; // Overrides ArtifactPrefix. + WriteToFile(U, Path); + Printf("artifact_prefix='%s'; Test unit written to %s\n", + Options.ArtifactPrefix.c_str(), Path.c_str()); + if (U.size() <= kMaxUnitSizeToPrint) + Printf("Base64: %s\n", Base64(U).c_str()); +} + +void Fuzzer::PrintStatusForNewUnit(const Unit &U) { + if (!Options.PrintNEW) + return; + PrintStats("NEW ", ""); + if (Options.Verbosity) { + Printf(" L: %zd ", U.size()); + MD.PrintMutationSequence(); + Printf("\n"); + } +} + +void Fuzzer::ReportNewCoverage(InputInfo *II, const Unit &U) { + II->NumSuccessfullMutations++; + MD.RecordSuccessfulMutationSequence(); + PrintStatusForNewUnit(U); + WriteToOutputCorpus(U); + NumberOfNewUnitsAdded++; + TPC.PrintNewPCs(); +} + +// Finds minimal number of units in 'Extra' that add coverage to 'Initial'. +// We do it by actually executing the units, sometimes more than once, +// because we may be using different coverage-like signals and the only +// common thing between them is that we can say "this unit found new stuff". +UnitVector Fuzzer::FindExtraUnits(const UnitVector &Initial, + const UnitVector &Extra) { + UnitVector Res = Extra; + UnitVector Tmp; + size_t OldSize = Res.size(); + for (int Iter = 0; Iter < 10; Iter++) { + ShuffleCorpus(&Res); + TPC.ResetMaps(); + Corpus.ResetFeatureSet(); + ResetCoverage(); + + for (auto &U : Initial) { + TPC.ResetMaps(); + RunOne(U); + } + + Tmp.clear(); + for (auto &U : Res) { + TPC.ResetMaps(); + if (RunOne(U)) + Tmp.push_back(U); + } + + char Stat[7] = "MIN "; + Stat[3] = '0' + Iter; + PrintStats(Stat, "\n", Tmp.size()); + + size_t NewSize = Tmp.size(); + assert(NewSize <= OldSize); + Res.swap(Tmp); + + if (NewSize + 5 >= OldSize) + break; + OldSize = NewSize; + } + return Res; +} + +void Fuzzer::Merge(const std::vector &Corpora) { + if (Corpora.size() <= 1) { + Printf("Merge requires two or more corpus dirs\n"); + return; + } + InMergeMode = true; + std::vector ExtraCorpora(Corpora.begin() + 1, Corpora.end()); + + assert(MaxInputLen > 0); + UnitVector Initial, Extra; + ReadDirToVectorOfUnits(Corpora[0].c_str(), &Initial, nullptr, MaxInputLen, + true); + for (auto &C : ExtraCorpora) + ReadDirToVectorOfUnits(C.c_str(), &Extra, nullptr, MaxInputLen, true); + + if (!Initial.empty()) { + Printf("=== Minimizing the initial corpus of %zd units\n", Initial.size()); + Initial = FindExtraUnits({}, Initial); + } + + Printf("=== Merging extra %zd units\n", Extra.size()); + auto Res = FindExtraUnits(Initial, Extra); + + for (auto &U: Res) + WriteToOutputCorpus(U); + + Printf("=== Merge: written %zd units\n", Res.size()); +} + +// Tries detecting a memory leak on the particular input that we have just +// executed before calling this function. +void Fuzzer::TryDetectingAMemoryLeak(const uint8_t *Data, size_t Size, + bool DuringInitialCorpusExecution) { + if (!HasMoreMallocsThanFrees) return; // mallocs==frees, a leak is unlikely. + if (!Options.DetectLeaks) return; + if (!&(EF->__lsan_enable) || !&(EF->__lsan_disable) || + !(EF->__lsan_do_recoverable_leak_check)) + return; // No lsan. + // Run the target once again, but with lsan disabled so that if there is + // a real leak we do not report it twice. + EF->__lsan_disable(); + ExecuteCallback(Data, Size); + EF->__lsan_enable(); + if (!HasMoreMallocsThanFrees) return; // a leak is unlikely. + if (NumberOfLeakDetectionAttempts++ > 1000) { + Options.DetectLeaks = false; + Printf("INFO: libFuzzer disabled leak detection after every mutation.\n" + " Most likely the target function accumulates allocated\n" + " memory in a global state w/o actually leaking it.\n" + " You may try running this binary with -trace_malloc=[12]" + " to get a trace of mallocs and frees.\n" + " If LeakSanitizer is enabled in this process it will still\n" + " run on the process shutdown.\n"); + return; + } + // Now perform the actual lsan pass. This is expensive and we must ensure + // we don't call it too often. + if (EF->__lsan_do_recoverable_leak_check()) { // Leak is found, report it. + if (DuringInitialCorpusExecution) + Printf("\nINFO: a leak has been found in the initial corpus.\n\n"); + Printf("INFO: to ignore leaks on libFuzzer side use -detect_leaks=0.\n\n"); + CurrentUnitSize = Size; + DumpCurrentUnit("leak-"); + PrintFinalStats(); + _Exit(Options.ErrorExitCode); // not exit() to disable lsan further on. + } +} + +void Fuzzer::MutateAndTestOne() { + MD.StartMutationSequence(); + + auto &II = Corpus.ChooseUnitToMutate(MD.GetRand()); + const auto &U = II.U; + memcpy(BaseSha1, II.Sha1, sizeof(BaseSha1)); + assert(CurrentUnitData); + size_t Size = U.size(); + assert(Size <= MaxInputLen && "Oversized Unit"); + memcpy(CurrentUnitData, U.data(), Size); + + assert(MaxMutationLen > 0); + + for (int i = 0; i < Options.MutateDepth; i++) { + if (TotalNumberOfRuns >= Options.MaxNumberOfRuns) + break; + size_t NewSize = 0; + NewSize = MD.Mutate(CurrentUnitData, Size, MaxMutationLen); + assert(NewSize > 0 && "Mutator returned empty unit"); + assert(NewSize <= MaxMutationLen && "Mutator return overisized unit"); + Size = NewSize; + if (i == 0) + StartTraceRecording(); + II.NumExecutedMutations++; + if (size_t NumFeatures = RunOne(CurrentUnitData, Size)) { + Corpus.AddToCorpus({CurrentUnitData, CurrentUnitData + Size}, NumFeatures, + /*MayDeleteFile=*/true); + ReportNewCoverage(&II, {CurrentUnitData, CurrentUnitData + Size}); + CheckExitOnSrcPosOrItem(); + } + StopTraceRecording(); + TryDetectingAMemoryLeak(CurrentUnitData, Size, + /*DuringInitialCorpusExecution*/ false); + } +} + +void Fuzzer::ResetCoverage() { + ResetEdgeCoverage(); + MaxCoverage.Reset(); + PrepareCounters(&MaxCoverage); +} + +void Fuzzer::Loop() { + system_clock::time_point LastCorpusReload = system_clock::now(); + if (Options.DoCrossOver) + MD.SetCorpus(&Corpus); + while (true) { + auto Now = system_clock::now(); + if (duration_cast(Now - LastCorpusReload).count() >= + Options.ReloadIntervalSec) { + RereadOutputCorpus(MaxInputLen); + LastCorpusReload = system_clock::now(); + } + if (TotalNumberOfRuns >= Options.MaxNumberOfRuns) + break; + if (TimedOut()) break; + // Perform several mutations and runs. + MutateAndTestOne(); + } + + PrintStats("DONE ", "\n"); + MD.PrintRecommendedDictionary(); +} + +void Fuzzer::MinimizeCrashLoop(const Unit &U) { + if (U.size() <= 2) return; + while (!TimedOut() && TotalNumberOfRuns < Options.MaxNumberOfRuns) { + MD.StartMutationSequence(); + memcpy(CurrentUnitData, U.data(), U.size()); + for (int i = 0; i < Options.MutateDepth; i++) { + size_t NewSize = MD.Mutate(CurrentUnitData, U.size(), MaxMutationLen); + assert(NewSize > 0 && NewSize <= MaxMutationLen); + RunOne(CurrentUnitData, NewSize); + TryDetectingAMemoryLeak(CurrentUnitData, NewSize, + /*DuringInitialCorpusExecution*/ false); + } + } +} + +} // namespace fuzzer + +extern "C" { + +size_t LLVMFuzzerMutate(uint8_t *Data, size_t Size, size_t MaxSize) { + assert(fuzzer::F); + return fuzzer::F->GetMD().DefaultMutate(Data, Size, MaxSize); +} +} // extern "C" diff --git a/test/thirdparty/Fuzzer/FuzzerMain.cpp b/test/thirdparty/Fuzzer/FuzzerMain.cpp new file mode 100644 index 00000000..af865720 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerMain.cpp @@ -0,0 +1,21 @@ +//===- FuzzerMain.cpp - main() function and flags -------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// main() and flags. +//===----------------------------------------------------------------------===// + +#include "FuzzerDefs.h" + +extern "C" { +// This function should be defined by the user. +int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size); +} // extern "C" + +int main(int argc, char **argv) { + return fuzzer::FuzzerDriver(&argc, &argv, LLVMFuzzerTestOneInput); +} diff --git a/test/thirdparty/Fuzzer/FuzzerMerge.cpp b/test/thirdparty/Fuzzer/FuzzerMerge.cpp new file mode 100644 index 00000000..84660e0f --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerMerge.cpp @@ -0,0 +1,261 @@ +//===- FuzzerMerge.cpp - merging corpora ----------------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Merging corpora. +//===----------------------------------------------------------------------===// + +#include "FuzzerInternal.h" +#include "FuzzerIO.h" +#include "FuzzerMerge.h" +#include "FuzzerTracePC.h" +#include "FuzzerUtil.h" + +#include +#include +#include + +namespace fuzzer { + +bool Merger::Parse(const std::string &Str, bool ParseCoverage) { + std::istringstream SS(Str); + return Parse(SS, ParseCoverage); +} + +void Merger::ParseOrExit(std::istream &IS, bool ParseCoverage) { + if (!Parse(IS, ParseCoverage)) { + Printf("MERGE: failed to parse the control file (unexpected error)\n"); + exit(1); + } +} + +// The control file example: +// +// 3 # The number of inputs +// 1 # The number of inputs in the first corpus, <= the previous number +// file0 +// file1 +// file2 # One file name per line. +// STARTED 0 123 # FileID, file size +// DONE 0 1 4 6 8 # FileID COV1 COV2 ... +// STARTED 1 456 # If DONE is missing, the input crashed while processing. +// STARTED 2 567 +// DONE 2 8 9 +bool Merger::Parse(std::istream &IS, bool ParseCoverage) { + LastFailure.clear(); + std::string Line; + + // Parse NumFiles. + if (!std::getline(IS, Line, '\n')) return false; + std::istringstream L1(Line); + size_t NumFiles = 0; + L1 >> NumFiles; + if (NumFiles == 0 || NumFiles > 10000000) return false; + + // Parse NumFilesInFirstCorpus. + if (!std::getline(IS, Line, '\n')) return false; + std::istringstream L2(Line); + NumFilesInFirstCorpus = NumFiles + 1; + L2 >> NumFilesInFirstCorpus; + if (NumFilesInFirstCorpus > NumFiles) return false; + + // Parse file names. + Files.resize(NumFiles); + for (size_t i = 0; i < NumFiles; i++) + if (!std::getline(IS, Files[i].Name, '\n')) + return false; + + // Parse STARTED and DONE lines. + size_t ExpectedStartMarker = 0; + const size_t kInvalidStartMarker = -1; + size_t LastSeenStartMarker = kInvalidStartMarker; + while (std::getline(IS, Line, '\n')) { + std::istringstream ISS1(Line); + std::string Marker; + size_t N; + ISS1 >> Marker; + ISS1 >> N; + if (Marker == "STARTED") { + // STARTED FILE_ID FILE_SIZE + if (ExpectedStartMarker != N) + return false; + ISS1 >> Files[ExpectedStartMarker].Size; + LastSeenStartMarker = ExpectedStartMarker; + assert(ExpectedStartMarker < Files.size()); + ExpectedStartMarker++; + } else if (Marker == "DONE") { + // DONE FILE_SIZE COV1 COV2 COV3 ... + size_t CurrentFileIdx = N; + if (CurrentFileIdx != LastSeenStartMarker) + return false; + LastSeenStartMarker = kInvalidStartMarker; + if (ParseCoverage) { + auto &V = Files[CurrentFileIdx].Features; + V.clear(); + while (ISS1 >> std::hex >> N) + V.push_back(N); + std::sort(V.begin(), V.end()); + } + } else { + return false; + } + } + if (LastSeenStartMarker != kInvalidStartMarker) + LastFailure = Files[LastSeenStartMarker].Name; + + FirstNotProcessedFile = ExpectedStartMarker; + return true; +} + +// Decides which files need to be merged (add thost to NewFiles). +// Returns the number of new features added. +size_t Merger::Merge(std::vector *NewFiles) { + NewFiles->clear(); + assert(NumFilesInFirstCorpus <= Files.size()); + std::set AllFeatures; + + // What features are in the initial corpus? + for (size_t i = 0; i < NumFilesInFirstCorpus; i++) { + auto &Cur = Files[i].Features; + AllFeatures.insert(Cur.begin(), Cur.end()); + } + size_t InitialNumFeatures = AllFeatures.size(); + + // Remove all features that we already know from all other inputs. + for (size_t i = NumFilesInFirstCorpus; i < Files.size(); i++) { + auto &Cur = Files[i].Features; + std::vector Tmp; + std::set_difference(Cur.begin(), Cur.end(), AllFeatures.begin(), + AllFeatures.end(), std::inserter(Tmp, Tmp.begin())); + Cur.swap(Tmp); + } + + // Sort. Give preference to + // * smaller files + // * files with more features. + std::sort(Files.begin() + NumFilesInFirstCorpus, Files.end(), + [&](const MergeFileInfo &a, const MergeFileInfo &b) -> bool { + if (a.Size != b.Size) + return a.Size < b.Size; + return a.Features.size() > b.Features.size(); + }); + + // One greedy pass: add the file's features to AllFeatures. + // If new features were added, add this file to NewFiles. + for (size_t i = NumFilesInFirstCorpus; i < Files.size(); i++) { + auto &Cur = Files[i].Features; + // Printf("%s -> sz %zd ft %zd\n", Files[i].Name.c_str(), + // Files[i].Size, Cur.size()); + size_t OldSize = AllFeatures.size(); + AllFeatures.insert(Cur.begin(), Cur.end()); + if (AllFeatures.size() > OldSize) + NewFiles->push_back(Files[i].Name); + } + return AllFeatures.size() - InitialNumFeatures; +} + +// Inner process. May crash if the target crashes. +void Fuzzer::CrashResistantMergeInternalStep(const std::string &CFPath) { + Printf("MERGE-INNER: using the control file '%s'\n", CFPath.c_str()); + Merger M; + std::ifstream IF(CFPath); + M.ParseOrExit(IF, false); + IF.close(); + if (!M.LastFailure.empty()) + Printf("MERGE-INNER: '%s' caused a failure at the previous merge step\n", + M.LastFailure.c_str()); + + Printf("MERGE-INNER: %zd total files;" + " %zd processed earlier; will process %zd files now\n", + M.Files.size(), M.FirstNotProcessedFile, + M.Files.size() - M.FirstNotProcessedFile); + + std::ofstream OF(CFPath, std::ofstream::out | std::ofstream::app); + for (size_t i = M.FirstNotProcessedFile; i < M.Files.size(); i++) { + auto U = FileToVector(M.Files[i].Name); + if (U.size() > MaxInputLen) { + U.resize(MaxInputLen); + U.shrink_to_fit(); + } + std::ostringstream StartedLine; + // Write the pre-run marker. + OF << "STARTED " << std::dec << i << " " << U.size() << "\n"; + OF.flush(); // Flush is important since ExecuteCommand may crash. + // Run. + TPC.ResetMaps(); + ExecuteCallback(U.data(), U.size()); + // Collect coverage. + std::set Features; + TPC.CollectFeatures([&](size_t Feature) -> bool { + Features.insert(Feature); + return true; + }); + // Show stats. + TotalNumberOfRuns++; + if (!(TotalNumberOfRuns & (TotalNumberOfRuns - 1))) + PrintStats("pulse "); + // Write the post-run marker and the coverage. + OF << "DONE " << i; + for (size_t F : Features) + OF << " " << std::hex << F; + OF << "\n"; + } +} + +// Outer process. Does not call the target code and thus sohuld not fail. +void Fuzzer::CrashResistantMerge(const std::vector &Args, + const std::vector &Corpora) { + if (Corpora.size() <= 1) { + Printf("Merge requires two or more corpus dirs\n"); + return; + } + std::vector AllFiles; + ListFilesInDirRecursive(Corpora[0], nullptr, &AllFiles, /*TopDir*/true); + size_t NumFilesInFirstCorpus = AllFiles.size(); + for (size_t i = 1; i < Corpora.size(); i++) + ListFilesInDirRecursive(Corpora[i], nullptr, &AllFiles, /*TopDir*/true); + Printf("MERGE-OUTER: %zd files, %zd in the initial corpus\n", + AllFiles.size(), NumFilesInFirstCorpus); + std::string CFPath = + "libFuzzerTemp." + std::to_string(GetPid()) + ".txt"; + // Write the control file. + RemoveFile(CFPath); + std::ofstream ControlFile(CFPath); + ControlFile << AllFiles.size() << "\n"; + ControlFile << NumFilesInFirstCorpus << "\n"; + for (auto &Path: AllFiles) + ControlFile << Path << "\n"; + ControlFile.close(); + + // Execute the inner process untill it passes. + // Every inner process should execute at least one input. + std::string BaseCmd = CloneArgsWithoutX(Args, "keep-all-flags"); + for (size_t i = 1; i <= AllFiles.size(); i++) { + Printf("MERGE-OUTER: attempt %zd\n", i); + auto ExitCode = + ExecuteCommand(BaseCmd + " -merge_control_file=" + CFPath); + if (!ExitCode) { + Printf("MERGE-OUTER: succesfull in %zd attempt(s)\n", i); + break; + } + } + // Read the control file and do the merge. + Merger M; + std::ifstream IF(CFPath); + M.ParseOrExit(IF, true); + IF.close(); + std::vector NewFiles; + size_t NumNewFeatures = M.Merge(&NewFiles); + Printf("MERGE-OUTER: %zd new files with %zd new features added\n", + NewFiles.size(), NumNewFeatures); + for (auto &F: NewFiles) + WriteToOutputCorpus(FileToVector(F)); + // We are done, delete the control file. + RemoveFile(CFPath); +} + +} // namespace fuzzer diff --git a/test/thirdparty/Fuzzer/FuzzerMerge.h b/test/thirdparty/Fuzzer/FuzzerMerge.h new file mode 100644 index 00000000..8a2fe5d7 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerMerge.h @@ -0,0 +1,70 @@ +//===- FuzzerMerge.h - merging corpa ----------------------------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Merging Corpora. +// +// The task: +// Take the existing corpus (possibly empty) and merge new inputs into +// it so that only inputs with new coverage ('features') are added. +// The process should tolerate the crashes, OOMs, leaks, etc. +// +// Algorithm: +// The outter process collects the set of files and writes their names +// into a temporary "control" file, then repeatedly launches the inner +// process until all inputs are processed. +// The outer process does not actually execute the target code. +// +// The inner process reads the control file and sees a) list of all the inputs +// and b) the last processed input. Then it starts processing the inputs one +// by one. Before processing every input it writes one line to control file: +// STARTED INPUT_ID INPUT_SIZE +// After processing an input it write another line: +// DONE INPUT_ID Feature1 Feature2 Feature3 ... +// If a crash happens while processing an input the last line in the control +// file will be "STARTED INPUT_ID" and so the next process will know +// where to resume. +// +// Once all inputs are processed by the innner process(es) the outer process +// reads the control files and does the merge based entirely on the contents +// of control file. +// It uses a single pass greedy algorithm choosing first the smallest inputs +// within the same size the inputs that have more new features. +// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_FUZZER_MERGE_H +#define LLVM_FUZZER_MERGE_H + +#include "FuzzerDefs.h" + +#include +#include + +namespace fuzzer { + +struct MergeFileInfo { + std::string Name; + size_t Size = 0; + std::vector Features; +}; + +struct Merger { + std::vector Files; + size_t NumFilesInFirstCorpus = 0; + size_t FirstNotProcessedFile = 0; + std::string LastFailure; + + bool Parse(std::istream &IS, bool ParseCoverage); + bool Parse(const std::string &Str, bool ParseCoverage); + void ParseOrExit(std::istream &IS, bool ParseCoverage); + size_t Merge(std::vector *NewFiles); +}; + +} // namespace fuzzer + +#endif // LLVM_FUZZER_MERGE_H diff --git a/test/thirdparty/Fuzzer/FuzzerMutate.cpp b/test/thirdparty/Fuzzer/FuzzerMutate.cpp new file mode 100644 index 00000000..80b7a5f9 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerMutate.cpp @@ -0,0 +1,527 @@ +//===- FuzzerMutate.cpp - Mutate a test input -----------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Mutate a test input. +//===----------------------------------------------------------------------===// + +#include "FuzzerCorpus.h" +#include "FuzzerDefs.h" +#include "FuzzerExtFunctions.h" +#include "FuzzerIO.h" +#include "FuzzerMutate.h" +#include "FuzzerOptions.h" + +namespace fuzzer { + +const size_t Dictionary::kMaxDictSize; + +static void PrintASCII(const Word &W, const char *PrintAfter) { + PrintASCII(W.data(), W.size(), PrintAfter); +} + +MutationDispatcher::MutationDispatcher(Random &Rand, + const FuzzingOptions &Options) + : Rand(Rand), Options(Options) { + DefaultMutators.insert( + DefaultMutators.begin(), + { + {&MutationDispatcher::Mutate_EraseBytes, "EraseBytes"}, + {&MutationDispatcher::Mutate_InsertByte, "InsertByte"}, + {&MutationDispatcher::Mutate_InsertRepeatedBytes, + "InsertRepeatedBytes"}, + {&MutationDispatcher::Mutate_ChangeByte, "ChangeByte"}, + {&MutationDispatcher::Mutate_ChangeBit, "ChangeBit"}, + {&MutationDispatcher::Mutate_ShuffleBytes, "ShuffleBytes"}, + {&MutationDispatcher::Mutate_ChangeASCIIInteger, "ChangeASCIIInt"}, + {&MutationDispatcher::Mutate_ChangeBinaryInteger, "ChangeBinInt"}, + {&MutationDispatcher::Mutate_CopyPart, "CopyPart"}, + {&MutationDispatcher::Mutate_CrossOver, "CrossOver"}, + {&MutationDispatcher::Mutate_AddWordFromManualDictionary, + "ManualDict"}, + {&MutationDispatcher::Mutate_AddWordFromTemporaryAutoDictionary, + "TempAutoDict"}, + {&MutationDispatcher::Mutate_AddWordFromPersistentAutoDictionary, + "PersAutoDict"}, + }); + if(Options.UseCmp) + DefaultMutators.push_back( + {&MutationDispatcher::Mutate_AddWordFromTORC, "CMP"}); + + if (EF->LLVMFuzzerCustomMutator) + Mutators.push_back({&MutationDispatcher::Mutate_Custom, "Custom"}); + else + Mutators = DefaultMutators; + + if (EF->LLVMFuzzerCustomCrossOver) + Mutators.push_back( + {&MutationDispatcher::Mutate_CustomCrossOver, "CustomCrossOver"}); +} + +static char RandCh(Random &Rand) { + if (Rand.RandBool()) return Rand(256); + const char *Special = "!*'();:@&=+$,/?%#[]012Az-`~.\xff\x00"; + return Special[Rand(sizeof(Special) - 1)]; +} + +size_t MutationDispatcher::Mutate_Custom(uint8_t *Data, size_t Size, + size_t MaxSize) { + return EF->LLVMFuzzerCustomMutator(Data, Size, MaxSize, Rand.Rand()); +} + +size_t MutationDispatcher::Mutate_CustomCrossOver(uint8_t *Data, size_t Size, + size_t MaxSize) { + if (!Corpus || Corpus->size() < 2 || Size == 0) + return 0; + size_t Idx = Rand(Corpus->size()); + const Unit &Other = (*Corpus)[Idx]; + if (Other.empty()) + return 0; + MutateInPlaceHere.resize(MaxSize); + auto &U = MutateInPlaceHere; + size_t NewSize = EF->LLVMFuzzerCustomCrossOver( + Data, Size, Other.data(), Other.size(), U.data(), U.size(), Rand.Rand()); + if (!NewSize) + return 0; + assert(NewSize <= MaxSize && "CustomCrossOver returned overisized unit"); + memcpy(Data, U.data(), NewSize); + return NewSize; +} + +size_t MutationDispatcher::Mutate_ShuffleBytes(uint8_t *Data, size_t Size, + size_t MaxSize) { + if (Size > MaxSize) return 0; + assert(Size); + size_t ShuffleAmount = + Rand(std::min(Size, (size_t)8)) + 1; // [1,8] and <= Size. + size_t ShuffleStart = Rand(Size - ShuffleAmount); + assert(ShuffleStart + ShuffleAmount <= Size); + std::random_shuffle(Data + ShuffleStart, Data + ShuffleStart + ShuffleAmount, + Rand); + return Size; +} + +size_t MutationDispatcher::Mutate_EraseBytes(uint8_t *Data, size_t Size, + size_t MaxSize) { + assert(Size); + if (Size == 1) return 0; + size_t N = Rand(Size / 2) + 1; + assert(N < Size); + size_t Idx = Rand(Size - N + 1); + // Erase Data[Idx:Idx+N]. + memmove(Data + Idx, Data + Idx + N, Size - Idx - N); + // Printf("Erase: %zd %zd => %zd; Idx %zd\n", N, Size, Size - N, Idx); + return Size - N; +} + +size_t MutationDispatcher::Mutate_InsertByte(uint8_t *Data, size_t Size, + size_t MaxSize) { + if (Size >= MaxSize) return 0; + size_t Idx = Rand(Size + 1); + // Insert new value at Data[Idx]. + memmove(Data + Idx + 1, Data + Idx, Size - Idx); + Data[Idx] = RandCh(Rand); + return Size + 1; +} + +size_t MutationDispatcher::Mutate_InsertRepeatedBytes(uint8_t *Data, + size_t Size, + size_t MaxSize) { + const size_t kMinBytesToInsert = 3; + if (Size + kMinBytesToInsert >= MaxSize) return 0; + size_t MaxBytesToInsert = std::min(MaxSize - Size, (size_t)128); + size_t N = Rand(MaxBytesToInsert - kMinBytesToInsert + 1) + kMinBytesToInsert; + assert(Size + N <= MaxSize && N); + size_t Idx = Rand(Size + 1); + // Insert new values at Data[Idx]. + memmove(Data + Idx + N, Data + Idx, Size - Idx); + // Give preference to 0x00 and 0xff. + uint8_t Byte = Rand.RandBool() ? Rand(256) : (Rand.RandBool() ? 0 : 255); + for (size_t i = 0; i < N; i++) + Data[Idx + i] = Byte; + return Size + N; +} + +size_t MutationDispatcher::Mutate_ChangeByte(uint8_t *Data, size_t Size, + size_t MaxSize) { + if (Size > MaxSize) return 0; + size_t Idx = Rand(Size); + Data[Idx] = RandCh(Rand); + return Size; +} + +size_t MutationDispatcher::Mutate_ChangeBit(uint8_t *Data, size_t Size, + size_t MaxSize) { + if (Size > MaxSize) return 0; + size_t Idx = Rand(Size); + Data[Idx] ^= 1 << Rand(8); + return Size; +} + +size_t MutationDispatcher::Mutate_AddWordFromManualDictionary(uint8_t *Data, + size_t Size, + size_t MaxSize) { + return AddWordFromDictionary(ManualDictionary, Data, Size, MaxSize); +} + +size_t MutationDispatcher::Mutate_AddWordFromTemporaryAutoDictionary( + uint8_t *Data, size_t Size, size_t MaxSize) { + return AddWordFromDictionary(TempAutoDictionary, Data, Size, MaxSize); +} + +size_t MutationDispatcher::ApplyDictionaryEntry(uint8_t *Data, size_t Size, + size_t MaxSize, + DictionaryEntry &DE) { + const Word &W = DE.GetW(); + bool UsePositionHint = DE.HasPositionHint() && + DE.GetPositionHint() + W.size() < Size && + Rand.RandBool(); + if (Rand.RandBool()) { // Insert W. + if (Size + W.size() > MaxSize) return 0; + size_t Idx = UsePositionHint ? DE.GetPositionHint() : Rand(Size + 1); + memmove(Data + Idx + W.size(), Data + Idx, Size - Idx); + memcpy(Data + Idx, W.data(), W.size()); + Size += W.size(); + } else { // Overwrite some bytes with W. + if (W.size() > Size) return 0; + size_t Idx = UsePositionHint ? DE.GetPositionHint() : Rand(Size - W.size()); + memcpy(Data + Idx, W.data(), W.size()); + } + return Size; +} + +// Somewhere in the past we have observed a comparison instructions +// with arguments Arg1 Arg2. This function tries to guess a dictionary +// entry that will satisfy that comparison. +// It first tries to find one of the arguments (possibly swapped) in the +// input and if it succeeds it creates a DE with a position hint. +// Otherwise it creates a DE with one of the arguments w/o a position hint. +template +DictionaryEntry MutationDispatcher::MakeDictionaryEntryFromCMP( + T Arg1, T Arg2, const uint8_t *Data, size_t Size) { + ScopedDoingMyOwnMemmem scoped_doing_my_own_memmem; + bool HandleFirst = Rand.RandBool(); + T ExistingBytes, DesiredBytes; + Word W; + const uint8_t *End = Data + Size; + for (int Arg = 0; Arg < 2; Arg++) { + ExistingBytes = HandleFirst ? Arg1 : Arg2; + DesiredBytes = HandleFirst ? Arg2 : Arg1; + DesiredBytes += Rand(-1, 1); + if (Rand.RandBool()) ExistingBytes = Bswap(ExistingBytes); + if (Rand.RandBool()) DesiredBytes = Bswap(DesiredBytes); + HandleFirst = !HandleFirst; + W.Set(reinterpret_cast(&DesiredBytes), sizeof(T)); + const size_t kMaxNumPositions = 8; + size_t Positions[kMaxNumPositions]; + size_t NumPositions = 0; + for (const uint8_t *Cur = Data; + Cur < End && NumPositions < kMaxNumPositions; Cur++) { + Cur = (uint8_t *)SearchMemory(Cur, End - Cur, &ExistingBytes, sizeof(T)); + if (!Cur) break; + Positions[NumPositions++] = Cur - Data; + } + if (!NumPositions) break; + return DictionaryEntry(W, Positions[Rand(NumPositions)]); + } + DictionaryEntry DE(W); + return DE; +} + +size_t MutationDispatcher::Mutate_AddWordFromTORC( + uint8_t *Data, size_t Size, size_t MaxSize) { + Word W; + DictionaryEntry DE; + if (Rand.RandBool()) { + auto X = TPC.TORC8.Get(Rand.Rand()); + DE = MakeDictionaryEntryFromCMP(X.A, X.B, Data, Size); + } else { + auto X = TPC.TORC4.Get(Rand.Rand()); + if ((X.A >> 16) == 0 && (X.B >> 16) == 0 && Rand.RandBool()) + DE = MakeDictionaryEntryFromCMP((uint16_t)X.A, (uint16_t)X.B, Data, + Size); + else + DE = MakeDictionaryEntryFromCMP(X.A, X.B, Data, Size); + } + Size = ApplyDictionaryEntry(Data, Size, MaxSize, DE); + if (!Size) return 0; + DictionaryEntry &DERef = + CmpDictionaryEntriesDeque[CmpDictionaryEntriesDequeIdx++ % + kCmpDictionaryEntriesDequeSize]; + DERef = DE; + CurrentDictionaryEntrySequence.push_back(&DERef); + return Size; +} + +size_t MutationDispatcher::Mutate_AddWordFromPersistentAutoDictionary( + uint8_t *Data, size_t Size, size_t MaxSize) { + return AddWordFromDictionary(PersistentAutoDictionary, Data, Size, MaxSize); +} + +size_t MutationDispatcher::AddWordFromDictionary(Dictionary &D, uint8_t *Data, + size_t Size, size_t MaxSize) { + if (Size > MaxSize) return 0; + if (D.empty()) return 0; + DictionaryEntry &DE = D[Rand(D.size())]; + Size = ApplyDictionaryEntry(Data, Size, MaxSize, DE); + if (!Size) return 0; + DE.IncUseCount(); + CurrentDictionaryEntrySequence.push_back(&DE); + return Size; +} + +// Overwrites part of To[0,ToSize) with a part of From[0,FromSize). +// Returns ToSize. +size_t MutationDispatcher::CopyPartOf(const uint8_t *From, size_t FromSize, + uint8_t *To, size_t ToSize) { + // Copy From[FromBeg, FromBeg + CopySize) into To[ToBeg, ToBeg + CopySize). + size_t ToBeg = Rand(ToSize); + size_t CopySize = Rand(ToSize - ToBeg) + 1; + assert(ToBeg + CopySize <= ToSize); + CopySize = std::min(CopySize, FromSize); + size_t FromBeg = Rand(FromSize - CopySize + 1); + assert(FromBeg + CopySize <= FromSize); + memmove(To + ToBeg, From + FromBeg, CopySize); + return ToSize; +} + +// Inserts part of From[0,ToSize) into To. +// Returns new size of To on success or 0 on failure. +size_t MutationDispatcher::InsertPartOf(const uint8_t *From, size_t FromSize, + uint8_t *To, size_t ToSize, + size_t MaxToSize) { + if (ToSize >= MaxToSize) return 0; + size_t AvailableSpace = MaxToSize - ToSize; + size_t MaxCopySize = std::min(AvailableSpace, FromSize); + size_t CopySize = Rand(MaxCopySize) + 1; + size_t FromBeg = Rand(FromSize - CopySize + 1); + assert(FromBeg + CopySize <= FromSize); + size_t ToInsertPos = Rand(ToSize + 1); + assert(ToInsertPos + CopySize <= MaxToSize); + size_t TailSize = ToSize - ToInsertPos; + if (To == From) { + MutateInPlaceHere.resize(MaxToSize); + memcpy(MutateInPlaceHere.data(), From + FromBeg, CopySize); + memmove(To + ToInsertPos + CopySize, To + ToInsertPos, TailSize); + memmove(To + ToInsertPos, MutateInPlaceHere.data(), CopySize); + } else { + memmove(To + ToInsertPos + CopySize, To + ToInsertPos, TailSize); + memmove(To + ToInsertPos, From + FromBeg, CopySize); + } + return ToSize + CopySize; +} + +size_t MutationDispatcher::Mutate_CopyPart(uint8_t *Data, size_t Size, + size_t MaxSize) { + if (Size > MaxSize) return 0; + if (Rand.RandBool()) + return CopyPartOf(Data, Size, Data, Size); + else + return InsertPartOf(Data, Size, Data, Size, MaxSize); +} + +size_t MutationDispatcher::Mutate_ChangeASCIIInteger(uint8_t *Data, size_t Size, + size_t MaxSize) { + if (Size > MaxSize) return 0; + size_t B = Rand(Size); + while (B < Size && !isdigit(Data[B])) B++; + if (B == Size) return 0; + size_t E = B; + while (E < Size && isdigit(Data[E])) E++; + assert(B < E); + // now we have digits in [B, E). + // strtol and friends don't accept non-zero-teminated data, parse it manually. + uint64_t Val = Data[B] - '0'; + for (size_t i = B + 1; i < E; i++) + Val = Val * 10 + Data[i] - '0'; + + // Mutate the integer value. + switch(Rand(5)) { + case 0: Val++; break; + case 1: Val--; break; + case 2: Val /= 2; break; + case 3: Val *= 2; break; + case 4: Val = Rand(Val * Val); break; + default: assert(0); + } + // Just replace the bytes with the new ones, don't bother moving bytes. + for (size_t i = B; i < E; i++) { + size_t Idx = E + B - i - 1; + assert(Idx >= B && Idx < E); + Data[Idx] = (Val % 10) + '0'; + Val /= 10; + } + return Size; +} + +template +size_t ChangeBinaryInteger(uint8_t *Data, size_t Size, Random &Rand) { + if (Size < sizeof(T)) return 0; + size_t Off = Rand(Size - sizeof(T) + 1); + assert(Off + sizeof(T) <= Size); + T Val; + if (Off < 64 && !Rand(4)) { + Val = Size; + if (Rand.RandBool()) + Val = Bswap(Val); + } else { + memcpy(&Val, Data + Off, sizeof(Val)); + T Add = Rand(21); + Add -= 10; + if (Rand.RandBool()) + Val = Bswap(T(Bswap(Val) + Add)); // Add assuming different endiannes. + else + Val = Val + Add; // Add assuming current endiannes. + if (Add == 0 || Rand.RandBool()) // Maybe negate. + Val = -Val; + } + memcpy(Data + Off, &Val, sizeof(Val)); + return Size; +} + +size_t MutationDispatcher::Mutate_ChangeBinaryInteger(uint8_t *Data, + size_t Size, + size_t MaxSize) { + if (Size > MaxSize) return 0; + switch (Rand(4)) { + case 3: return ChangeBinaryInteger(Data, Size, Rand); + case 2: return ChangeBinaryInteger(Data, Size, Rand); + case 1: return ChangeBinaryInteger(Data, Size, Rand); + case 0: return ChangeBinaryInteger(Data, Size, Rand); + default: assert(0); + } + return 0; +} + +size_t MutationDispatcher::Mutate_CrossOver(uint8_t *Data, size_t Size, + size_t MaxSize) { + if (Size > MaxSize) return 0; + if (!Corpus || Corpus->size() < 2 || Size == 0) return 0; + size_t Idx = Rand(Corpus->size()); + const Unit &O = (*Corpus)[Idx]; + if (O.empty()) return 0; + MutateInPlaceHere.resize(MaxSize); + auto &U = MutateInPlaceHere; + size_t NewSize = 0; + switch(Rand(3)) { + case 0: + NewSize = CrossOver(Data, Size, O.data(), O.size(), U.data(), U.size()); + break; + case 1: + NewSize = InsertPartOf(O.data(), O.size(), U.data(), U.size(), MaxSize); + if (NewSize) + break; + // LLVM_FALLTHROUGH; + case 2: + NewSize = CopyPartOf(O.data(), O.size(), U.data(), U.size()); + break; + default: assert(0); + } + assert(NewSize > 0 && "CrossOver returned empty unit"); + assert(NewSize <= MaxSize && "CrossOver returned overisized unit"); + memcpy(Data, U.data(), NewSize); + return NewSize; +} + +void MutationDispatcher::StartMutationSequence() { + CurrentMutatorSequence.clear(); + CurrentDictionaryEntrySequence.clear(); +} + +// Copy successful dictionary entries to PersistentAutoDictionary. +void MutationDispatcher::RecordSuccessfulMutationSequence() { + for (auto DE : CurrentDictionaryEntrySequence) { + // PersistentAutoDictionary.AddWithSuccessCountOne(DE); + DE->IncSuccessCount(); + // Linear search is fine here as this happens seldom. + if (!PersistentAutoDictionary.ContainsWord(DE->GetW())) + PersistentAutoDictionary.push_back({DE->GetW(), 1}); + } +} + +void MutationDispatcher::PrintRecommendedDictionary() { + std::vector V; + for (auto &DE : PersistentAutoDictionary) + if (!ManualDictionary.ContainsWord(DE.GetW())) + V.push_back(DE); + if (V.empty()) return; + Printf("###### Recommended dictionary. ######\n"); + for (auto &DE: V) { + Printf("\""); + PrintASCII(DE.GetW(), "\""); + Printf(" # Uses: %zd\n", DE.GetUseCount()); + } + Printf("###### End of recommended dictionary. ######\n"); +} + +void MutationDispatcher::PrintMutationSequence() { + Printf("MS: %zd ", CurrentMutatorSequence.size()); + for (auto M : CurrentMutatorSequence) + Printf("%s-", M.Name); + if (!CurrentDictionaryEntrySequence.empty()) { + Printf(" DE: "); + for (auto DE : CurrentDictionaryEntrySequence) { + Printf("\""); + PrintASCII(DE->GetW(), "\"-"); + } + } +} + +size_t MutationDispatcher::Mutate(uint8_t *Data, size_t Size, size_t MaxSize) { + return MutateImpl(Data, Size, MaxSize, Mutators); +} + +size_t MutationDispatcher::DefaultMutate(uint8_t *Data, size_t Size, + size_t MaxSize) { + return MutateImpl(Data, Size, MaxSize, DefaultMutators); +} + +// Mutates Data in place, returns new size. +size_t MutationDispatcher::MutateImpl(uint8_t *Data, size_t Size, + size_t MaxSize, + const std::vector &Mutators) { + assert(MaxSize > 0); + if (Size == 0) { + for (size_t i = 0; i < MaxSize; i++) + Data[i] = RandCh(Rand); + if (Options.OnlyASCII) + ToASCII(Data, MaxSize); + return MaxSize; + } + assert(Size > 0); + // Some mutations may fail (e.g. can't insert more bytes if Size == MaxSize), + // in which case they will return 0. + // Try several times before returning un-mutated data. + for (int Iter = 0; Iter < 100; Iter++) { + auto M = Mutators[Rand(Mutators.size())]; + size_t NewSize = (this->*(M.Fn))(Data, Size, MaxSize); + if (NewSize && NewSize <= MaxSize) { + if (Options.OnlyASCII) + ToASCII(Data, NewSize); + CurrentMutatorSequence.push_back(M); + return NewSize; + } + } + return std::min(Size, MaxSize); +} + +void MutationDispatcher::AddWordToManualDictionary(const Word &W) { + ManualDictionary.push_back( + {W, std::numeric_limits::max()}); +} + +void MutationDispatcher::AddWordToAutoDictionary(DictionaryEntry DE) { + static const size_t kMaxAutoDictSize = 1 << 14; + if (TempAutoDictionary.size() >= kMaxAutoDictSize) return; + TempAutoDictionary.push_back(DE); +} + +void MutationDispatcher::ClearAutoDictionary() { + TempAutoDictionary.clear(); +} + +} // namespace fuzzer diff --git a/test/thirdparty/Fuzzer/FuzzerMutate.h b/test/thirdparty/Fuzzer/FuzzerMutate.h new file mode 100644 index 00000000..d3c0b001 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerMutate.h @@ -0,0 +1,145 @@ +//===- FuzzerMutate.h - Internal header for the Fuzzer ----------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// fuzzer::MutationDispatcher +//===----------------------------------------------------------------------===// + +#ifndef LLVM_FUZZER_MUTATE_H +#define LLVM_FUZZER_MUTATE_H + +#include "FuzzerDefs.h" +#include "FuzzerDictionary.h" +#include "FuzzerRandom.h" + +namespace fuzzer { + +class MutationDispatcher { +public: + MutationDispatcher(Random &Rand, const FuzzingOptions &Options); + ~MutationDispatcher() {} + /// Indicate that we are about to start a new sequence of mutations. + void StartMutationSequence(); + /// Print the current sequence of mutations. + void PrintMutationSequence(); + /// Indicate that the current sequence of mutations was successfull. + void RecordSuccessfulMutationSequence(); + /// Mutates data by invoking user-provided mutator. + size_t Mutate_Custom(uint8_t *Data, size_t Size, size_t MaxSize); + /// Mutates data by invoking user-provided crossover. + size_t Mutate_CustomCrossOver(uint8_t *Data, size_t Size, size_t MaxSize); + /// Mutates data by shuffling bytes. + size_t Mutate_ShuffleBytes(uint8_t *Data, size_t Size, size_t MaxSize); + /// Mutates data by erasing bytes. + size_t Mutate_EraseBytes(uint8_t *Data, size_t Size, size_t MaxSize); + /// Mutates data by inserting a byte. + size_t Mutate_InsertByte(uint8_t *Data, size_t Size, size_t MaxSize); + /// Mutates data by inserting several repeated bytes. + size_t Mutate_InsertRepeatedBytes(uint8_t *Data, size_t Size, size_t MaxSize); + /// Mutates data by chanding one byte. + size_t Mutate_ChangeByte(uint8_t *Data, size_t Size, size_t MaxSize); + /// Mutates data by chanding one bit. + size_t Mutate_ChangeBit(uint8_t *Data, size_t Size, size_t MaxSize); + /// Mutates data by copying/inserting a part of data into a different place. + size_t Mutate_CopyPart(uint8_t *Data, size_t Size, size_t MaxSize); + + /// Mutates data by adding a word from the manual dictionary. + size_t Mutate_AddWordFromManualDictionary(uint8_t *Data, size_t Size, + size_t MaxSize); + + /// Mutates data by adding a word from the temporary automatic dictionary. + size_t Mutate_AddWordFromTemporaryAutoDictionary(uint8_t *Data, size_t Size, + size_t MaxSize); + + /// Mutates data by adding a word from the TORC. + size_t Mutate_AddWordFromTORC(uint8_t *Data, size_t Size, size_t MaxSize); + + /// Mutates data by adding a word from the persistent automatic dictionary. + size_t Mutate_AddWordFromPersistentAutoDictionary(uint8_t *Data, size_t Size, + size_t MaxSize); + + /// Tries to find an ASCII integer in Data, changes it to another ASCII int. + size_t Mutate_ChangeASCIIInteger(uint8_t *Data, size_t Size, size_t MaxSize); + /// Change a 1-, 2-, 4-, or 8-byte integer in interesting ways. + size_t Mutate_ChangeBinaryInteger(uint8_t *Data, size_t Size, size_t MaxSize); + + /// CrossOver Data with some other element of the corpus. + size_t Mutate_CrossOver(uint8_t *Data, size_t Size, size_t MaxSize); + + /// Applies one of the configured mutations. + /// Returns the new size of data which could be up to MaxSize. + size_t Mutate(uint8_t *Data, size_t Size, size_t MaxSize); + /// Applies one of the default mutations. Provided as a service + /// to mutation authors. + size_t DefaultMutate(uint8_t *Data, size_t Size, size_t MaxSize); + + /// Creates a cross-over of two pieces of Data, returns its size. + size_t CrossOver(const uint8_t *Data1, size_t Size1, const uint8_t *Data2, + size_t Size2, uint8_t *Out, size_t MaxOutSize); + + void AddWordToManualDictionary(const Word &W); + + void AddWordToAutoDictionary(DictionaryEntry DE); + void ClearAutoDictionary(); + void PrintRecommendedDictionary(); + + void SetCorpus(const InputCorpus *Corpus) { this->Corpus = Corpus; } + + Random &GetRand() { return Rand; } + +private: + + struct Mutator { + size_t (MutationDispatcher::*Fn)(uint8_t *Data, size_t Size, size_t Max); + const char *Name; + }; + + size_t AddWordFromDictionary(Dictionary &D, uint8_t *Data, size_t Size, + size_t MaxSize); + size_t MutateImpl(uint8_t *Data, size_t Size, size_t MaxSize, + const std::vector &Mutators); + + size_t InsertPartOf(const uint8_t *From, size_t FromSize, uint8_t *To, + size_t ToSize, size_t MaxToSize); + size_t CopyPartOf(const uint8_t *From, size_t FromSize, uint8_t *To, + size_t ToSize); + size_t ApplyDictionaryEntry(uint8_t *Data, size_t Size, size_t MaxSize, + DictionaryEntry &DE); + + template + DictionaryEntry MakeDictionaryEntryFromCMP(T Arg1, T Arg2, + const uint8_t *Data, size_t Size); + + Random &Rand; + const FuzzingOptions &Options; + + // Dictionary provided by the user via -dict=DICT_FILE. + Dictionary ManualDictionary; + // Temporary dictionary modified by the fuzzer itself, + // recreated periodically. + Dictionary TempAutoDictionary; + // Persistent dictionary modified by the fuzzer, consists of + // entries that led to successfull discoveries in the past mutations. + Dictionary PersistentAutoDictionary; + + std::vector CurrentMutatorSequence; + std::vector CurrentDictionaryEntrySequence; + + static const size_t kCmpDictionaryEntriesDequeSize = 16; + DictionaryEntry CmpDictionaryEntriesDeque[kCmpDictionaryEntriesDequeSize]; + size_t CmpDictionaryEntriesDequeIdx = 0; + + const InputCorpus *Corpus = nullptr; + std::vector MutateInPlaceHere; + + std::vector Mutators; + std::vector DefaultMutators; +}; + +} // namespace fuzzer + +#endif // LLVM_FUZZER_MUTATE_H diff --git a/test/thirdparty/Fuzzer/FuzzerOptions.h b/test/thirdparty/Fuzzer/FuzzerOptions.h new file mode 100644 index 00000000..34e93fc3 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerOptions.h @@ -0,0 +1,68 @@ +//===- FuzzerOptions.h - Internal header for the Fuzzer ---------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// fuzzer::FuzzingOptions +//===----------------------------------------------------------------------===// + +#ifndef LLVM_FUZZER_OPTIONS_H +#define LLVM_FUZZER_OPTIONS_H + +#include "FuzzerDefs.h" + +namespace fuzzer { + +struct FuzzingOptions { + int Verbosity = 1; + size_t MaxLen = 0; + int UnitTimeoutSec = 300; + int TimeoutExitCode = 77; + int ErrorExitCode = 77; + int MaxTotalTimeSec = 0; + int RssLimitMb = 0; + bool DoCrossOver = true; + int MutateDepth = 5; + bool UseCounters = false; + bool UseIndirCalls = true; + bool UseMemcmp = true; + bool UseMemmem = true; + bool UseCmp = false; + bool UseValueProfile = false; + bool Shrink = false; + int ReloadIntervalSec = 1; + bool ShuffleAtStartUp = true; + bool PreferSmall = true; + size_t MaxNumberOfRuns = -1L; + int ReportSlowUnits = 10; + bool OnlyASCII = false; + std::string OutputCorpus; + std::string ArtifactPrefix = "./"; + std::string ExactArtifactPath; + std::string ExitOnSrcPos; + std::string ExitOnItem; + bool SaveArtifacts = true; + bool PrintNEW = true; // Print a status line when new units are found; + bool OutputCSV = false; + bool PrintNewCovPcs = false; + bool PrintFinalStats = false; + bool PrintCorpusStats = false; + bool PrintCoverage = false; + bool DumpCoverage = false; + bool DetectLeaks = true; + int TraceMalloc = 0; + bool HandleAbrt = false; + bool HandleBus = false; + bool HandleFpe = false; + bool HandleIll = false; + bool HandleInt = false; + bool HandleSegv = false; + bool HandleTerm = false; +}; + +} // namespace fuzzer + +#endif // LLVM_FUZZER_OPTIONS_H diff --git a/test/thirdparty/Fuzzer/FuzzerRandom.h b/test/thirdparty/Fuzzer/FuzzerRandom.h new file mode 100644 index 00000000..b1be0bb9 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerRandom.h @@ -0,0 +1,36 @@ +//===- FuzzerRandom.h - Internal header for the Fuzzer ----------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// fuzzer::Random +//===----------------------------------------------------------------------===// + +#ifndef LLVM_FUZZER_RANDOM_H +#define LLVM_FUZZER_RANDOM_H + +#include + +namespace fuzzer { +class Random { + public: + Random(unsigned int seed) : R(seed) {} + size_t Rand() { return R(); } + size_t RandBool() { return Rand() % 2; } + size_t operator()(size_t n) { return n ? Rand() % n : 0; } + intptr_t operator()(intptr_t From, intptr_t To) { + assert(From < To); + intptr_t RangeSize = To - From + 1; + return operator()(RangeSize) + From; + } + std::mt19937 &Get_mt19937() { return R; } + private: + std::mt19937 R; +}; + +} // namespace fuzzer + +#endif // LLVM_FUZZER_RANDOM_H diff --git a/test/thirdparty/Fuzzer/FuzzerSHA1.cpp b/test/thirdparty/Fuzzer/FuzzerSHA1.cpp new file mode 100644 index 00000000..d2f8e811 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerSHA1.cpp @@ -0,0 +1,222 @@ +//===- FuzzerSHA1.h - Private copy of the SHA1 implementation ---*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// This code is taken from public domain +// (http://oauth.googlecode.com/svn/code/c/liboauth/src/sha1.c) +// and modified by adding anonymous namespace, adding an interface +// function fuzzer::ComputeSHA1() and removing unnecessary code. +// +// lib/Fuzzer can not use SHA1 implementation from openssl because +// openssl may not be available and because we may be fuzzing openssl itself. +// For the same reason we do not want to depend on SHA1 from LLVM tree. +//===----------------------------------------------------------------------===// + +#include "FuzzerSHA1.h" +#include "FuzzerDefs.h" + +/* This code is public-domain - it is based on libcrypt + * placed in the public domain by Wei Dai and other contributors. + */ + +#include +#include +#include +#include + +namespace { // Added for LibFuzzer + +#ifdef __BIG_ENDIAN__ +# define SHA_BIG_ENDIAN +#elif defined __LITTLE_ENDIAN__ +/* override */ +#elif defined __BYTE_ORDER +# if __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__ +# define SHA_BIG_ENDIAN +# endif +#else // ! defined __LITTLE_ENDIAN__ +# include // machine/endian.h +# if __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__ +# define SHA_BIG_ENDIAN +# endif +#endif + + +/* header */ + +#define HASH_LENGTH 20 +#define BLOCK_LENGTH 64 + +typedef struct sha1nfo { + uint32_t buffer[BLOCK_LENGTH/4]; + uint32_t state[HASH_LENGTH/4]; + uint32_t byteCount; + uint8_t bufferOffset; + uint8_t keyBuffer[BLOCK_LENGTH]; + uint8_t innerHash[HASH_LENGTH]; +} sha1nfo; + +/* public API - prototypes - TODO: doxygen*/ + +/** + */ +void sha1_init(sha1nfo *s); +/** + */ +void sha1_writebyte(sha1nfo *s, uint8_t data); +/** + */ +void sha1_write(sha1nfo *s, const char *data, size_t len); +/** + */ +uint8_t* sha1_result(sha1nfo *s); + + +/* code */ +#define SHA1_K0 0x5a827999 +#define SHA1_K20 0x6ed9eba1 +#define SHA1_K40 0x8f1bbcdc +#define SHA1_K60 0xca62c1d6 + +void sha1_init(sha1nfo *s) { + s->state[0] = 0x67452301; + s->state[1] = 0xefcdab89; + s->state[2] = 0x98badcfe; + s->state[3] = 0x10325476; + s->state[4] = 0xc3d2e1f0; + s->byteCount = 0; + s->bufferOffset = 0; +} + +uint32_t sha1_rol32(uint32_t number, uint8_t bits) { + return ((number << bits) | (number >> (32-bits))); +} + +void sha1_hashBlock(sha1nfo *s) { + uint8_t i; + uint32_t a,b,c,d,e,t; + + a=s->state[0]; + b=s->state[1]; + c=s->state[2]; + d=s->state[3]; + e=s->state[4]; + for (i=0; i<80; i++) { + if (i>=16) { + t = s->buffer[(i+13)&15] ^ s->buffer[(i+8)&15] ^ s->buffer[(i+2)&15] ^ s->buffer[i&15]; + s->buffer[i&15] = sha1_rol32(t,1); + } + if (i<20) { + t = (d ^ (b & (c ^ d))) + SHA1_K0; + } else if (i<40) { + t = (b ^ c ^ d) + SHA1_K20; + } else if (i<60) { + t = ((b & c) | (d & (b | c))) + SHA1_K40; + } else { + t = (b ^ c ^ d) + SHA1_K60; + } + t+=sha1_rol32(a,5) + e + s->buffer[i&15]; + e=d; + d=c; + c=sha1_rol32(b,30); + b=a; + a=t; + } + s->state[0] += a; + s->state[1] += b; + s->state[2] += c; + s->state[3] += d; + s->state[4] += e; +} + +void sha1_addUncounted(sha1nfo *s, uint8_t data) { + uint8_t * const b = (uint8_t*) s->buffer; +#ifdef SHA_BIG_ENDIAN + b[s->bufferOffset] = data; +#else + b[s->bufferOffset ^ 3] = data; +#endif + s->bufferOffset++; + if (s->bufferOffset == BLOCK_LENGTH) { + sha1_hashBlock(s); + s->bufferOffset = 0; + } +} + +void sha1_writebyte(sha1nfo *s, uint8_t data) { + ++s->byteCount; + sha1_addUncounted(s, data); +} + +void sha1_write(sha1nfo *s, const char *data, size_t len) { + for (;len--;) sha1_writebyte(s, (uint8_t) *data++); +} + +void sha1_pad(sha1nfo *s) { + // Implement SHA-1 padding (fips180-2 §5.1.1) + + // Pad with 0x80 followed by 0x00 until the end of the block + sha1_addUncounted(s, 0x80); + while (s->bufferOffset != 56) sha1_addUncounted(s, 0x00); + + // Append length in the last 8 bytes + sha1_addUncounted(s, 0); // We're only using 32 bit lengths + sha1_addUncounted(s, 0); // But SHA-1 supports 64 bit lengths + sha1_addUncounted(s, 0); // So zero pad the top bits + sha1_addUncounted(s, s->byteCount >> 29); // Shifting to multiply by 8 + sha1_addUncounted(s, s->byteCount >> 21); // as SHA-1 supports bitstreams as well as + sha1_addUncounted(s, s->byteCount >> 13); // byte. + sha1_addUncounted(s, s->byteCount >> 5); + sha1_addUncounted(s, s->byteCount << 3); +} + +uint8_t* sha1_result(sha1nfo *s) { + // Pad to complete the last block + sha1_pad(s); + +#ifndef SHA_BIG_ENDIAN + // Swap byte order back + int i; + for (i=0; i<5; i++) { + s->state[i]= + (((s->state[i])<<24)& 0xff000000) + | (((s->state[i])<<8) & 0x00ff0000) + | (((s->state[i])>>8) & 0x0000ff00) + | (((s->state[i])>>24)& 0x000000ff); + } +#endif + + // Return pointer to hash (20 characters) + return (uint8_t*) s->state; +} + +} // namespace; Added for LibFuzzer + +namespace fuzzer { + +// The rest is added for LibFuzzer +void ComputeSHA1(const uint8_t *Data, size_t Len, uint8_t *Out) { + sha1nfo s; + sha1_init(&s); + sha1_write(&s, (const char*)Data, Len); + memcpy(Out, sha1_result(&s), HASH_LENGTH); +} + +std::string Sha1ToString(const uint8_t Sha1[kSHA1NumBytes]) { + std::stringstream SS; + for (int i = 0; i < kSHA1NumBytes; i++) + SS << std::hex << std::setfill('0') << std::setw(2) << (unsigned)Sha1[i]; + return SS.str(); +} + +std::string Hash(const Unit &U) { + uint8_t Hash[kSHA1NumBytes]; + ComputeSHA1(U.data(), U.size(), Hash); + return Sha1ToString(Hash); +} + +} diff --git a/test/thirdparty/Fuzzer/FuzzerSHA1.h b/test/thirdparty/Fuzzer/FuzzerSHA1.h new file mode 100644 index 00000000..3b5e6e80 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerSHA1.h @@ -0,0 +1,33 @@ +//===- FuzzerSHA1.h - Internal header for the SHA1 utils --------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// SHA1 utils. +//===----------------------------------------------------------------------===// + +#ifndef LLVM_FUZZER_SHA1_H +#define LLVM_FUZZER_SHA1_H + +#include "FuzzerDefs.h" +#include +#include + +namespace fuzzer { + +// Private copy of SHA1 implementation. +static const int kSHA1NumBytes = 20; + +// Computes SHA1 hash of 'Len' bytes in 'Data', writes kSHA1NumBytes to 'Out'. +void ComputeSHA1(const uint8_t *Data, size_t Len, uint8_t *Out); + +std::string Sha1ToString(const uint8_t Sha1[kSHA1NumBytes]); + +std::string Hash(const Unit &U); + +} // namespace fuzzer + +#endif // LLVM_FUZZER_SHA1_H diff --git a/test/thirdparty/Fuzzer/FuzzerTracePC.cpp b/test/thirdparty/Fuzzer/FuzzerTracePC.cpp new file mode 100644 index 00000000..81c084b5 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerTracePC.cpp @@ -0,0 +1,339 @@ +//===- FuzzerTracePC.cpp - PC tracing--------------------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Trace PCs. +// This module implements __sanitizer_cov_trace_pc_guard[_init], +// the callback required for -fsanitize-coverage=trace-pc-guard instrumentation. +// +//===----------------------------------------------------------------------===// + +#include "FuzzerCorpus.h" +#include "FuzzerDefs.h" +#include "FuzzerDictionary.h" +#include "FuzzerExtFunctions.h" +#include "FuzzerIO.h" +#include "FuzzerTracePC.h" +#include "FuzzerValueBitMap.h" +#include +#include +#include +#include + +namespace fuzzer { + +TracePC TPC; + +void TracePC::HandleTrace(uint32_t *Guard, uintptr_t PC) { + uint32_t Idx = *Guard; + if (!Idx) return; + PCs[Idx % kNumPCs] = PC; + Counters[Idx % kNumCounters]++; +} + +size_t TracePC::GetTotalPCCoverage() { + size_t Res = 0; + for (size_t i = 1; i < GetNumPCs(); i++) + if (PCs[i]) + Res++; + return Res; +} + +void TracePC::HandleInit(uint32_t *Start, uint32_t *Stop) { + if (Start == Stop || *Start) return; + assert(NumModules < sizeof(Modules) / sizeof(Modules[0])); + for (uint32_t *P = Start; P < Stop; P++) + *P = ++NumGuards; + Modules[NumModules].Start = Start; + Modules[NumModules].Stop = Stop; + NumModules++; +} + +void TracePC::PrintModuleInfo() { + Printf("INFO: Loaded %zd modules (%zd guards): ", NumModules, NumGuards); + for (size_t i = 0; i < NumModules; i++) + Printf("[%p, %p), ", Modules[i].Start, Modules[i].Stop); + Printf("\n"); +} + +void TracePC::HandleCallerCallee(uintptr_t Caller, uintptr_t Callee) { + const uintptr_t kBits = 12; + const uintptr_t kMask = (1 << kBits) - 1; + uintptr_t Idx = (Caller & kMask) | ((Callee & kMask) << kBits); + HandleValueProfile(Idx); +} + +static bool IsInterestingCoverageFile(std::string &File) { + if (File.find("compiler-rt/lib/") != std::string::npos) + return false; // sanitizer internal. + if (File.find("/usr/lib/") != std::string::npos) + return false; + if (File.find("/usr/include/") != std::string::npos) + return false; + if (File == "") + return false; + return true; +} + +void TracePC::PrintNewPCs() { + if (DoPrintNewPCs) { + if (!PrintedPCs) + PrintedPCs = new std::set; + for (size_t i = 1; i < GetNumPCs(); i++) + if (PCs[i] && PrintedPCs->insert(PCs[i]).second) + PrintPC("\tNEW_PC: %p %F %L\n", "\tNEW_PC: %p\n", PCs[i]); + } +} + +void TracePC::PrintCoverage() { + if (!EF->__sanitizer_symbolize_pc || + !EF->__sanitizer_get_module_and_offset_for_pc) { + Printf("INFO: __sanitizer_symbolize_pc or " + "__sanitizer_get_module_and_offset_for_pc is not available," + " not printing coverage\n"); + return; + } + std::map> CoveredPCsPerModule; + std::map ModuleOffsets; + std::set CoveredDirs, CoveredFiles, CoveredFunctions, + CoveredLines; + Printf("COVERAGE:\n"); + for (size_t i = 1; i < GetNumPCs(); i++) { + if (!PCs[i]) continue; + std::string FileStr = DescribePC("%s", PCs[i]); + if (!IsInterestingCoverageFile(FileStr)) continue; + std::string FixedPCStr = DescribePC("%p", PCs[i]); + std::string FunctionStr = DescribePC("%F", PCs[i]); + std::string LineStr = DescribePC("%l", PCs[i]); + char ModulePathRaw[4096] = ""; // What's PATH_MAX in portable C++? + void *OffsetRaw = nullptr; + if (!EF->__sanitizer_get_module_and_offset_for_pc( + reinterpret_cast(PCs[i]), ModulePathRaw, + sizeof(ModulePathRaw), &OffsetRaw)) + continue; + std::string Module = ModulePathRaw; + uintptr_t FixedPC = std::stol(FixedPCStr, 0, 16); + uintptr_t PcOffset = reinterpret_cast(OffsetRaw); + ModuleOffsets[Module] = FixedPC - PcOffset; + CoveredPCsPerModule[Module].push_back(PcOffset); + CoveredFunctions.insert(FunctionStr); + CoveredFiles.insert(FileStr); + CoveredDirs.insert(DirName(FileStr)); + if (!CoveredLines.insert(FileStr + ":" + LineStr).second) + continue; + Printf("COVERED: %s %s:%s\n", FunctionStr.c_str(), + FileStr.c_str(), LineStr.c_str()); + } + + std::string CoveredDirsStr; + for (auto &Dir : CoveredDirs) { + if (!CoveredDirsStr.empty()) + CoveredDirsStr += ","; + CoveredDirsStr += Dir; + } + Printf("COVERED_DIRS: %s\n", CoveredDirsStr.c_str()); + + for (auto &M : CoveredPCsPerModule) { + std::set UncoveredFiles, UncoveredFunctions; + std::map > UncoveredLines; // Func+File => lines + auto &ModuleName = M.first; + auto &CoveredOffsets = M.second; + uintptr_t ModuleOffset = ModuleOffsets[ModuleName]; + std::sort(CoveredOffsets.begin(), CoveredOffsets.end()); + Printf("MODULE_WITH_COVERAGE: %s\n", ModuleName.c_str()); + // sancov does not yet fully support DSOs. + // std::string Cmd = "sancov -print-coverage-pcs " + ModuleName; + std::string Cmd = "objdump -d " + ModuleName + + " | grep 'call.*__sanitizer_cov_trace_pc_guard' | awk -F: '{print $1}'"; + std::string SanCovOutput; + if (!ExecuteCommandAndReadOutput(Cmd, &SanCovOutput)) { + Printf("INFO: Command failed: %s\n", Cmd.c_str()); + continue; + } + std::istringstream ISS(SanCovOutput); + std::string S; + while (std::getline(ISS, S, '\n')) { + uintptr_t PcOffset = std::stol(S, 0, 16); + if (!std::binary_search(CoveredOffsets.begin(), CoveredOffsets.end(), + PcOffset)) { + uintptr_t PC = ModuleOffset + PcOffset; + auto FileStr = DescribePC("%s", PC); + if (!IsInterestingCoverageFile(FileStr)) continue; + if (CoveredFiles.count(FileStr) == 0) { + UncoveredFiles.insert(FileStr); + continue; + } + auto FunctionStr = DescribePC("%F", PC); + if (CoveredFunctions.count(FunctionStr) == 0) { + UncoveredFunctions.insert(FunctionStr); + continue; + } + std::string LineStr = DescribePC("%l", PC); + uintptr_t Line = std::stoi(LineStr); + std::string FileLineStr = FileStr + ":" + LineStr; + if (CoveredLines.count(FileLineStr) == 0) + UncoveredLines[FunctionStr + " " + FileStr].insert(Line); + } + } + for (auto &FileLine: UncoveredLines) + for (int Line : FileLine.second) + Printf("UNCOVERED_LINE: %s:%d\n", FileLine.first.c_str(), Line); + for (auto &Func : UncoveredFunctions) + Printf("UNCOVERED_FUNC: %s\n", Func.c_str()); + for (auto &File : UncoveredFiles) + Printf("UNCOVERED_FILE: %s\n", File.c_str()); + } +} + +void TracePC::DumpCoverage() { + __sanitizer_dump_coverage(PCs, GetNumPCs()); +} + +// Value profile. +// We keep track of various values that affect control flow. +// These values are inserted into a bit-set-based hash map. +// Every new bit in the map is treated as a new coverage. +// +// For memcmp/strcmp/etc the interesting value is the length of the common +// prefix of the parameters. +// For cmp instructions the interesting value is a XOR of the parameters. +// The interesting value is mixed up with the PC and is then added to the map. + +ATTRIBUTE_NO_SANITIZE_MEMORY +void TracePC::AddValueForMemcmp(void *caller_pc, const void *s1, const void *s2, + size_t n) { + if (!n) return; + size_t Len = std::min(n, (size_t)32); + const uint8_t *A1 = reinterpret_cast(s1); + const uint8_t *A2 = reinterpret_cast(s2); + size_t I = 0; + for (; I < Len; I++) + if (A1[I] != A2[I]) + break; + size_t PC = reinterpret_cast(caller_pc); + size_t Idx = I; + // if (I < Len) + // Idx += __builtin_popcountl((A1[I] ^ A2[I])) - 1; + TPC.HandleValueProfile((PC & 4095) | (Idx << 12)); +} + +ATTRIBUTE_NO_SANITIZE_MEMORY +void TracePC::AddValueForStrcmp(void *caller_pc, const char *s1, const char *s2, + size_t n) { + if (!n) return; + size_t Len = std::min(n, (size_t)32); + const uint8_t *A1 = reinterpret_cast(s1); + const uint8_t *A2 = reinterpret_cast(s2); + size_t I = 0; + for (; I < Len; I++) + if (A1[I] != A2[I] || A1[I] == 0) + break; + size_t PC = reinterpret_cast(caller_pc); + size_t Idx = I; + // if (I < Len && A1[I]) + // Idx += __builtin_popcountl((A1[I] ^ A2[I])) - 1; + TPC.HandleValueProfile((PC & 4095) | (Idx << 12)); +} + +template +ATTRIBUTE_TARGET_POPCNT +#ifdef __clang__ // g++ can't handle this __attribute__ here :( +__attribute__((always_inline)) +#endif // __clang__ +void TracePC::HandleCmp(void *PC, T Arg1, T Arg2) { + uintptr_t PCuint = reinterpret_cast(PC); + uint64_t ArgXor = Arg1 ^ Arg2; + uint64_t ArgDistance = __builtin_popcountl(ArgXor) + 1; // [1,65] + uintptr_t Idx = ((PCuint & 4095) + 1) * ArgDistance; + if (sizeof(T) == 4) + TORC4.Insert(ArgXor, Arg1, Arg2); + else if (sizeof(T) == 8) + TORC8.Insert(ArgXor, Arg1, Arg2); + HandleValueProfile(Idx); +} + +} // namespace fuzzer + +extern "C" { +__attribute__((visibility("default"))) +void __sanitizer_cov_trace_pc_guard(uint32_t *Guard) { + uintptr_t PC = (uintptr_t)__builtin_return_address(0); + fuzzer::TPC.HandleTrace(Guard, PC); +} + +__attribute__((visibility("default"))) +void __sanitizer_cov_trace_pc_guard_init(uint32_t *Start, uint32_t *Stop) { + fuzzer::TPC.HandleInit(Start, Stop); +} + +__attribute__((visibility("default"))) +void __sanitizer_cov_trace_pc_indir(uintptr_t Callee) { + uintptr_t PC = (uintptr_t)__builtin_return_address(0); + fuzzer::TPC.HandleCallerCallee(PC, Callee); +} + +__attribute__((visibility("default"))) +void __sanitizer_cov_trace_cmp8(uint64_t Arg1, uint64_t Arg2) { + fuzzer::TPC.HandleCmp(__builtin_return_address(0), Arg1, Arg2); +} +__attribute__((visibility("default"))) +void __sanitizer_cov_trace_cmp4(uint32_t Arg1, uint32_t Arg2) { + fuzzer::TPC.HandleCmp(__builtin_return_address(0), Arg1, Arg2); +} +__attribute__((visibility("default"))) +void __sanitizer_cov_trace_cmp2(uint16_t Arg1, uint16_t Arg2) { + fuzzer::TPC.HandleCmp(__builtin_return_address(0), Arg1, Arg2); +} +__attribute__((visibility("default"))) +void __sanitizer_cov_trace_cmp1(uint8_t Arg1, uint8_t Arg2) { + fuzzer::TPC.HandleCmp(__builtin_return_address(0), Arg1, Arg2); +} + +__attribute__((visibility("default"))) +void __sanitizer_cov_trace_switch(uint64_t Val, uint64_t *Cases) { + // Updates the value profile based on the relative position of Val and Cases. + // We want to handle one random case at every call (handling all is slow). + // Since none of the arguments contain any random bits we use a thread-local + // counter to choose the random case to handle. + static thread_local size_t Counter; + Counter++; + uint64_t N = Cases[0]; + uint64_t *Vals = Cases + 2; + char *PC = (char*)__builtin_return_address(0); + // We need a random number < N using Counter as a seed. But w/o DIV. + // * find a power of two >= N + // * mask Counter with this power of two. + // * maybe subtract N. + size_t Nlog = sizeof(long) * 8 - __builtin_clzl((long)N); + size_t PowerOfTwoGeN = 1U << Nlog; + assert(PowerOfTwoGeN >= N); + size_t Idx = Counter & (PowerOfTwoGeN - 1); + if (Idx >= N) + Idx -= N; + assert(Idx < N); + uint64_t TwoIn32 = 1ULL << 32; + if ((Val | Vals[Idx]) < TwoIn32) + fuzzer::TPC.HandleCmp(PC + Idx, static_cast(Val), + static_cast(Vals[Idx])); + else + fuzzer::TPC.HandleCmp(PC + Idx, Val, Vals[Idx]); +} + +__attribute__((visibility("default"))) +void __sanitizer_cov_trace_div4(uint32_t Val) { + fuzzer::TPC.HandleCmp(__builtin_return_address(0), Val, (uint32_t)0); +} +__attribute__((visibility("default"))) +void __sanitizer_cov_trace_div8(uint64_t Val) { + fuzzer::TPC.HandleCmp(__builtin_return_address(0), Val, (uint64_t)0); +} +__attribute__((visibility("default"))) +void __sanitizer_cov_trace_gep(uintptr_t Idx) { + fuzzer::TPC.HandleCmp(__builtin_return_address(0), Idx, (uintptr_t)0); +} + +} // extern "C" diff --git a/test/thirdparty/Fuzzer/FuzzerTracePC.h b/test/thirdparty/Fuzzer/FuzzerTracePC.h new file mode 100644 index 00000000..68827f80 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerTracePC.h @@ -0,0 +1,158 @@ +//===- FuzzerTracePC.h - Internal header for the Fuzzer ---------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// fuzzer::TracePC +//===----------------------------------------------------------------------===// + +#ifndef LLVM_FUZZER_TRACE_PC +#define LLVM_FUZZER_TRACE_PC + +#include "FuzzerDefs.h" +#include "FuzzerValueBitMap.h" +#include + +namespace fuzzer { + +// TableOfRecentCompares (TORC) remembers the most recently performed +// comparisons of type T. +// We record the arguments of CMP instructions in this table unconditionally +// because it seems cheaper this way than to compute some expensive +// conditions inside __sanitizer_cov_trace_cmp*. +// After the unit has been executed we may decide to use the contents of +// this table to populate a Dictionary. +template +struct TableOfRecentCompares { + static const size_t kSize = kSizeT; + struct Pair { + T A, B; + }; + void Insert(size_t Idx, T Arg1, T Arg2) { + Idx = Idx % kSize; + Table[Idx].A = Arg1; + Table[Idx].B = Arg2; + } + + Pair Get(size_t I) { return Table[I % kSize]; } + + Pair Table[kSize]; +}; + +class TracePC { + public: + static const size_t kFeatureSetSize = ValueBitMap::kNumberOfItems; + + void HandleTrace(uint32_t *guard, uintptr_t PC); + void HandleInit(uint32_t *start, uint32_t *stop); + void HandleCallerCallee(uintptr_t Caller, uintptr_t Callee); + void HandleValueProfile(size_t Value) { ValueProfileMap.AddValue(Value); } + template void HandleCmp(void *PC, T Arg1, T Arg2); + size_t GetTotalPCCoverage(); + void SetUseCounters(bool UC) { UseCounters = UC; } + void SetUseValueProfile(bool VP) { UseValueProfile = VP; } + void SetPrintNewPCs(bool P) { DoPrintNewPCs = P; } + template size_t CollectFeatures(Callback CB); + bool UpdateValueProfileMap(ValueBitMap *MaxValueProfileMap) { + return UseValueProfile && MaxValueProfileMap->MergeFrom(ValueProfileMap); + } + + void ResetMaps() { + ValueProfileMap.Reset(); + memset(Counters, 0, sizeof(Counters)); + } + + void UpdateFeatureSet(size_t CurrentElementIdx, size_t CurrentElementSize); + void PrintFeatureSet(); + + void PrintModuleInfo(); + + void PrintCoverage(); + void DumpCoverage(); + + void AddValueForMemcmp(void *caller_pc, const void *s1, const void *s2, + size_t n); + void AddValueForStrcmp(void *caller_pc, const char *s1, const char *s2, + size_t n); + + bool UsingTracePcGuard() const {return NumModules; } + + static const size_t kTORCSize = 1 << 5; + TableOfRecentCompares TORC4; + TableOfRecentCompares TORC8; + + void PrintNewPCs(); + size_t GetNumPCs() const { return Min(kNumPCs, NumGuards + 1); } + uintptr_t GetPC(size_t Idx) { + assert(Idx < GetNumPCs()); + return PCs[Idx]; + } + +private: + bool UseCounters = false; + bool UseValueProfile = false; + bool DoPrintNewPCs = false; + + struct Module { + uint32_t *Start, *Stop; + }; + + Module Modules[4096]; + size_t NumModules; // linker-initialized. + size_t NumGuards; // linker-initialized. + + static const size_t kNumCounters = 1 << 14; + alignas(8) uint8_t Counters[kNumCounters]; + + static const size_t kNumPCs = 1 << 24; + uintptr_t PCs[kNumPCs]; + + std::set *PrintedPCs; + + ValueBitMap ValueProfileMap; +}; + +template +size_t TracePC::CollectFeatures(Callback CB) { + if (!UsingTracePcGuard()) return 0; + size_t Res = 0; + const size_t Step = 8; + assert(reinterpret_cast(Counters) % Step == 0); + size_t N = Min(kNumCounters, NumGuards + 1); + N = (N + Step - 1) & ~(Step - 1); // Round up. + for (size_t Idx = 0; Idx < N; Idx += Step) { + uint64_t Bundle = *reinterpret_cast(&Counters[Idx]); + if (!Bundle) continue; + for (size_t i = Idx; i < Idx + Step; i++) { + uint8_t Counter = (Bundle >> ((i - Idx) * 8)) & 0xff; + if (!Counter) continue; + Counters[i] = 0; + unsigned Bit = 0; + /**/ if (Counter >= 128) Bit = 7; + else if (Counter >= 32) Bit = 6; + else if (Counter >= 16) Bit = 5; + else if (Counter >= 8) Bit = 4; + else if (Counter >= 4) Bit = 3; + else if (Counter >= 3) Bit = 2; + else if (Counter >= 2) Bit = 1; + size_t Feature = (i * 8 + Bit); + if (CB(Feature)) + Res++; + } + } + if (UseValueProfile) + ValueProfileMap.ForEach([&](size_t Idx) { + if (CB(NumGuards * 8 + Idx)) + Res++; + }); + return Res; +} + +extern TracePC TPC; + +} // namespace fuzzer + +#endif // LLVM_FUZZER_TRACE_PC diff --git a/test/thirdparty/Fuzzer/FuzzerTraceState.cpp b/test/thirdparty/Fuzzer/FuzzerTraceState.cpp new file mode 100644 index 00000000..be62a662 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerTraceState.cpp @@ -0,0 +1,325 @@ +//===- FuzzerTraceState.cpp - Trace-based fuzzer mutator ------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Data tracing. +//===----------------------------------------------------------------------===// + +#include "FuzzerDictionary.h" +#include "FuzzerInternal.h" +#include "FuzzerIO.h" +#include "FuzzerMutate.h" +#include "FuzzerRandom.h" +#include "FuzzerTracePC.h" +#include +#include +#include +#include +#include + +namespace fuzzer { + +// For now, very simple: put Size bytes of Data at position Pos. +struct TraceBasedMutation { + uint32_t Pos; + Word W; +}; + +// Declared as static globals for faster checks inside the hooks. +static bool RecordingMemcmp = false; +static bool RecordingMemmem = false; +static bool DoingMyOwnMemmem = false; + +ScopedDoingMyOwnMemmem::ScopedDoingMyOwnMemmem() { DoingMyOwnMemmem = true; } +ScopedDoingMyOwnMemmem::~ScopedDoingMyOwnMemmem() { DoingMyOwnMemmem = false; } + +class TraceState { +public: + TraceState(MutationDispatcher &MD, const FuzzingOptions &Options, + const Fuzzer *F) + : MD(MD), Options(Options), F(F) {} + + void TraceMemcmpCallback(size_t CmpSize, const uint8_t *Data1, + const uint8_t *Data2); + + void TraceSwitchCallback(uintptr_t PC, size_t ValSizeInBits, uint64_t Val, + size_t NumCases, uint64_t *Cases); + int TryToAddDesiredData(uint64_t PresentData, uint64_t DesiredData, + size_t DataSize); + int TryToAddDesiredData(const uint8_t *PresentData, + const uint8_t *DesiredData, size_t DataSize); + + void StartTraceRecording() { + if (!Options.UseMemcmp) + return; + RecordingMemcmp = Options.UseMemcmp; + RecordingMemmem = Options.UseMemmem; + NumMutations = 0; + InterestingWords.clear(); + MD.ClearAutoDictionary(); + } + + void StopTraceRecording() { + if (!RecordingMemcmp) + return; + RecordingMemcmp = false; + for (size_t i = 0; i < NumMutations; i++) { + auto &M = Mutations[i]; + if (Options.Verbosity >= 2) { + AutoDictUnitCounts[M.W]++; + AutoDictAdds++; + if ((AutoDictAdds & (AutoDictAdds - 1)) == 0) { + typedef std::pair CU; + std::vector CountedUnits; + for (auto &I : AutoDictUnitCounts) + CountedUnits.push_back(std::make_pair(I.second, I.first)); + std::sort(CountedUnits.begin(), CountedUnits.end(), + [](const CU &a, const CU &b) { return a.first > b.first; }); + Printf("AutoDict:\n"); + for (auto &I : CountedUnits) { + Printf(" %zd ", I.first); + PrintASCII(I.second.data(), I.second.size()); + Printf("\n"); + } + } + } + MD.AddWordToAutoDictionary({M.W, M.Pos}); + } + for (auto &W : InterestingWords) + MD.AddWordToAutoDictionary({W}); + } + + void AddMutation(uint32_t Pos, uint32_t Size, const uint8_t *Data) { + if (NumMutations >= kMaxMutations) return; + auto &M = Mutations[NumMutations++]; + M.Pos = Pos; + M.W.Set(Data, Size); + } + + void AddMutation(uint32_t Pos, uint32_t Size, uint64_t Data) { + assert(Size <= sizeof(Data)); + AddMutation(Pos, Size, reinterpret_cast(&Data)); + } + + void AddInterestingWord(const uint8_t *Data, size_t Size) { + if (!RecordingMemmem || !F->InFuzzingThread()) return; + if (Size <= 1) return; + Size = std::min(Size, Word::GetMaxSize()); + Word W(Data, Size); + InterestingWords.insert(W); + } + + private: + bool IsTwoByteData(uint64_t Data) { + int64_t Signed = static_cast(Data); + Signed >>= 16; + return Signed == 0 || Signed == -1L; + } + + // We don't want to create too many trace-based mutations as it is both + // expensive and useless. So after some number of mutations is collected, + // start rejecting some of them. The more there are mutations the more we + // reject. + bool WantToHandleOneMoreMutation() { + const size_t FirstN = 64; + // Gladly handle first N mutations. + if (NumMutations <= FirstN) return true; + size_t Diff = NumMutations - FirstN; + size_t DiffLog = sizeof(long) * 8 - __builtin_clzl((long)Diff); + assert(DiffLog > 0 && DiffLog < 64); + bool WantThisOne = MD.GetRand()(1 << DiffLog) == 0; // 1 out of DiffLog. + return WantThisOne; + } + + static const size_t kMaxMutations = 1 << 16; + size_t NumMutations; + TraceBasedMutation Mutations[kMaxMutations]; + // TODO: std::set is too inefficient, need to have a custom DS here. + std::set InterestingWords; + MutationDispatcher &MD; + const FuzzingOptions Options; + const Fuzzer *F; + std::map AutoDictUnitCounts; + size_t AutoDictAdds = 0; +}; + +int TraceState::TryToAddDesiredData(uint64_t PresentData, uint64_t DesiredData, + size_t DataSize) { + if (NumMutations >= kMaxMutations || !WantToHandleOneMoreMutation()) return 0; + ScopedDoingMyOwnMemmem scoped_doing_my_own_memmem; + const uint8_t *UnitData; + auto UnitSize = F->GetCurrentUnitInFuzzingThead(&UnitData); + int Res = 0; + const uint8_t *Beg = UnitData; + const uint8_t *End = Beg + UnitSize; + for (const uint8_t *Cur = Beg; Cur < End; Cur++) { + Cur = (uint8_t *)SearchMemory(Cur, End - Cur, &PresentData, DataSize); + if (!Cur) + break; + size_t Pos = Cur - Beg; + assert(Pos < UnitSize); + AddMutation(Pos, DataSize, DesiredData); + AddMutation(Pos, DataSize, DesiredData + 1); + AddMutation(Pos, DataSize, DesiredData - 1); + Res++; + } + return Res; +} + +int TraceState::TryToAddDesiredData(const uint8_t *PresentData, + const uint8_t *DesiredData, + size_t DataSize) { + if (NumMutations >= kMaxMutations || !WantToHandleOneMoreMutation()) return 0; + ScopedDoingMyOwnMemmem scoped_doing_my_own_memmem; + const uint8_t *UnitData; + auto UnitSize = F->GetCurrentUnitInFuzzingThead(&UnitData); + int Res = 0; + const uint8_t *Beg = UnitData; + const uint8_t *End = Beg + UnitSize; + for (const uint8_t *Cur = Beg; Cur < End; Cur++) { + Cur = (uint8_t *)SearchMemory(Cur, End - Cur, PresentData, DataSize); + if (!Cur) + break; + size_t Pos = Cur - Beg; + assert(Pos < UnitSize); + AddMutation(Pos, DataSize, DesiredData); + Res++; + } + return Res; +} + +void TraceState::TraceMemcmpCallback(size_t CmpSize, const uint8_t *Data1, + const uint8_t *Data2) { + if (!RecordingMemcmp || !F->InFuzzingThread()) return; + CmpSize = std::min(CmpSize, Word::GetMaxSize()); + int Added2 = TryToAddDesiredData(Data1, Data2, CmpSize); + int Added1 = TryToAddDesiredData(Data2, Data1, CmpSize); + if ((Added1 || Added2) && Options.Verbosity >= 3) { + Printf("MemCmp Added %d%d: ", Added1, Added2); + if (Added1) PrintASCII(Data1, CmpSize); + if (Added2) PrintASCII(Data2, CmpSize); + Printf("\n"); + } +} + +void TraceState::TraceSwitchCallback(uintptr_t PC, size_t ValSizeInBits, + uint64_t Val, size_t NumCases, + uint64_t *Cases) { + if (F->InFuzzingThread()) return; + size_t ValSize = ValSizeInBits / 8; + bool TryShort = IsTwoByteData(Val); + for (size_t i = 0; i < NumCases; i++) + TryShort &= IsTwoByteData(Cases[i]); + + if (Options.Verbosity >= 3) + Printf("TraceSwitch: %p %zd # %zd; TryShort %d\n", PC, Val, NumCases, + TryShort); + + for (size_t i = 0; i < NumCases; i++) { + TryToAddDesiredData(Val, Cases[i], ValSize); + if (TryShort) + TryToAddDesiredData(Val, Cases[i], 2); + } +} + +static TraceState *TS; + +void Fuzzer::StartTraceRecording() { + if (!TS) return; + TS->StartTraceRecording(); +} + +void Fuzzer::StopTraceRecording() { + if (!TS) return; + TS->StopTraceRecording(); +} + +void Fuzzer::InitializeTraceState() { + if (!Options.UseMemcmp) return; + TS = new TraceState(MD, Options, this); +} + +static size_t InternalStrnlen(const char *S, size_t MaxLen) { + size_t Len = 0; + for (; Len < MaxLen && S[Len]; Len++) {} + return Len; +} + +} // namespace fuzzer + +using fuzzer::TS; +using fuzzer::RecordingMemcmp; + +extern "C" { + +// We may need to avoid defining weak hooks to stay compatible with older clang. +#ifndef LLVM_FUZZER_DEFINES_SANITIZER_WEAK_HOOOKS +# define LLVM_FUZZER_DEFINES_SANITIZER_WEAK_HOOOKS 1 +#endif + +#if LLVM_FUZZER_DEFINES_SANITIZER_WEAK_HOOOKS +void __sanitizer_weak_hook_memcmp(void *caller_pc, const void *s1, + const void *s2, size_t n, int result) { + fuzzer::TPC.AddValueForMemcmp(caller_pc, s1, s2, n); + if (!RecordingMemcmp) return; + if (result == 0) return; // No reason to mutate. + if (n <= 1) return; // Not interesting. + TS->TraceMemcmpCallback(n, reinterpret_cast(s1), + reinterpret_cast(s2)); +} + +void __sanitizer_weak_hook_strncmp(void *caller_pc, const char *s1, + const char *s2, size_t n, int result) { + fuzzer::TPC.AddValueForStrcmp(caller_pc, s1, s2, n); + if (!RecordingMemcmp) return; + if (result == 0) return; // No reason to mutate. + size_t Len1 = fuzzer::InternalStrnlen(s1, n); + size_t Len2 = fuzzer::InternalStrnlen(s2, n); + n = std::min(n, Len1); + n = std::min(n, Len2); + if (n <= 1) return; // Not interesting. + TS->TraceMemcmpCallback(n, reinterpret_cast(s1), + reinterpret_cast(s2)); +} + +void __sanitizer_weak_hook_strcmp(void *caller_pc, const char *s1, + const char *s2, int result) { + fuzzer::TPC.AddValueForStrcmp(caller_pc, s1, s2, 64); + if (!RecordingMemcmp) return; + if (result == 0) return; // No reason to mutate. + size_t Len1 = strlen(s1); + size_t Len2 = strlen(s2); + size_t N = std::min(Len1, Len2); + if (N <= 1) return; // Not interesting. + TS->TraceMemcmpCallback(N, reinterpret_cast(s1), + reinterpret_cast(s2)); +} + +void __sanitizer_weak_hook_strncasecmp(void *called_pc, const char *s1, + const char *s2, size_t n, int result) { + return __sanitizer_weak_hook_strncmp(called_pc, s1, s2, n, result); +} +void __sanitizer_weak_hook_strcasecmp(void *called_pc, const char *s1, + const char *s2, int result) { + return __sanitizer_weak_hook_strcmp(called_pc, s1, s2, result); +} +void __sanitizer_weak_hook_strstr(void *called_pc, const char *s1, + const char *s2, char *result) { + TS->AddInterestingWord(reinterpret_cast(s2), strlen(s2)); +} +void __sanitizer_weak_hook_strcasestr(void *called_pc, const char *s1, + const char *s2, char *result) { + TS->AddInterestingWord(reinterpret_cast(s2), strlen(s2)); +} +void __sanitizer_weak_hook_memmem(void *called_pc, const void *s1, size_t len1, + const void *s2, size_t len2, void *result) { + if (fuzzer::DoingMyOwnMemmem) return; + TS->AddInterestingWord(reinterpret_cast(s2), len2); +} + +#endif // LLVM_FUZZER_DEFINES_SANITIZER_WEAK_HOOOKS +} // extern "C" diff --git a/test/thirdparty/Fuzzer/FuzzerUtil.cpp b/test/thirdparty/Fuzzer/FuzzerUtil.cpp new file mode 100644 index 00000000..f5fd3a85 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerUtil.cpp @@ -0,0 +1,218 @@ +//===- FuzzerUtil.cpp - Misc utils ----------------------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Misc utils. +//===----------------------------------------------------------------------===// + +#include "FuzzerUtil.h" +#include "FuzzerIO.h" +#include "FuzzerInternal.h" +#include +#include +#include +#include +#include +#include +#include +#include +#include + +namespace fuzzer { + +void PrintHexArray(const uint8_t *Data, size_t Size, + const char *PrintAfter) { + for (size_t i = 0; i < Size; i++) + Printf("0x%x,", (unsigned)Data[i]); + Printf("%s", PrintAfter); +} + +void Print(const Unit &v, const char *PrintAfter) { + PrintHexArray(v.data(), v.size(), PrintAfter); +} + +void PrintASCIIByte(uint8_t Byte) { + if (Byte == '\\') + Printf("\\\\"); + else if (Byte == '"') + Printf("\\\""); + else if (Byte >= 32 && Byte < 127) + Printf("%c", Byte); + else + Printf("\\x%02x", Byte); +} + +void PrintASCII(const uint8_t *Data, size_t Size, const char *PrintAfter) { + for (size_t i = 0; i < Size; i++) + PrintASCIIByte(Data[i]); + Printf("%s", PrintAfter); +} + +void PrintASCII(const Unit &U, const char *PrintAfter) { + PrintASCII(U.data(), U.size(), PrintAfter); +} + +bool ToASCII(uint8_t *Data, size_t Size) { + bool Changed = false; + for (size_t i = 0; i < Size; i++) { + uint8_t &X = Data[i]; + auto NewX = X; + NewX &= 127; + if (!isspace(NewX) && !isprint(NewX)) + NewX = ' '; + Changed |= NewX != X; + X = NewX; + } + return Changed; +} + +bool IsASCII(const Unit &U) { return IsASCII(U.data(), U.size()); } + +bool IsASCII(const uint8_t *Data, size_t Size) { + for (size_t i = 0; i < Size; i++) + if (!(isprint(Data[i]) || isspace(Data[i]))) return false; + return true; +} + +bool ParseOneDictionaryEntry(const std::string &Str, Unit *U) { + U->clear(); + if (Str.empty()) return false; + size_t L = 0, R = Str.size() - 1; // We are parsing the range [L,R]. + // Skip spaces from both sides. + while (L < R && isspace(Str[L])) L++; + while (R > L && isspace(Str[R])) R--; + if (R - L < 2) return false; + // Check the closing " + if (Str[R] != '"') return false; + R--; + // Find the opening " + while (L < R && Str[L] != '"') L++; + if (L >= R) return false; + assert(Str[L] == '\"'); + L++; + assert(L <= R); + for (size_t Pos = L; Pos <= R; Pos++) { + uint8_t V = (uint8_t)Str[Pos]; + if (!isprint(V) && !isspace(V)) return false; + if (V =='\\') { + // Handle '\\' + if (Pos + 1 <= R && (Str[Pos + 1] == '\\' || Str[Pos + 1] == '"')) { + U->push_back(Str[Pos + 1]); + Pos++; + continue; + } + // Handle '\xAB' + if (Pos + 3 <= R && Str[Pos + 1] == 'x' + && isxdigit(Str[Pos + 2]) && isxdigit(Str[Pos + 3])) { + char Hex[] = "0xAA"; + Hex[2] = Str[Pos + 2]; + Hex[3] = Str[Pos + 3]; + U->push_back(strtol(Hex, nullptr, 16)); + Pos += 3; + continue; + } + return false; // Invalid escape. + } else { + // Any other character. + U->push_back(V); + } + } + return true; +} + +bool ParseDictionaryFile(const std::string &Text, std::vector *Units) { + if (Text.empty()) { + Printf("ParseDictionaryFile: file does not exist or is empty\n"); + return false; + } + std::istringstream ISS(Text); + Units->clear(); + Unit U; + int LineNo = 0; + std::string S; + while (std::getline(ISS, S, '\n')) { + LineNo++; + size_t Pos = 0; + while (Pos < S.size() && isspace(S[Pos])) Pos++; // Skip spaces. + if (Pos == S.size()) continue; // Empty line. + if (S[Pos] == '#') continue; // Comment line. + if (ParseOneDictionaryEntry(S, &U)) { + Units->push_back(U); + } else { + Printf("ParseDictionaryFile: error in line %d\n\t\t%s\n", LineNo, + S.c_str()); + return false; + } + } + return true; +} + +std::string Base64(const Unit &U) { + static const char Table[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZ" + "abcdefghijklmnopqrstuvwxyz" + "0123456789+/"; + std::string Res; + size_t i; + for (i = 0; i + 2 < U.size(); i += 3) { + uint32_t x = (U[i] << 16) + (U[i + 1] << 8) + U[i + 2]; + Res += Table[(x >> 18) & 63]; + Res += Table[(x >> 12) & 63]; + Res += Table[(x >> 6) & 63]; + Res += Table[x & 63]; + } + if (i + 1 == U.size()) { + uint32_t x = (U[i] << 16); + Res += Table[(x >> 18) & 63]; + Res += Table[(x >> 12) & 63]; + Res += "=="; + } else if (i + 2 == U.size()) { + uint32_t x = (U[i] << 16) + (U[i + 1] << 8); + Res += Table[(x >> 18) & 63]; + Res += Table[(x >> 12) & 63]; + Res += Table[(x >> 6) & 63]; + Res += "="; + } + return Res; +} + +std::string DescribePC(const char *SymbolizedFMT, uintptr_t PC) { + if (!EF->__sanitizer_symbolize_pc) return ""; + char PcDescr[1024]; + EF->__sanitizer_symbolize_pc(reinterpret_cast(PC), + SymbolizedFMT, PcDescr, sizeof(PcDescr)); + PcDescr[sizeof(PcDescr) - 1] = 0; // Just in case. + return PcDescr; +} + +void PrintPC(const char *SymbolizedFMT, const char *FallbackFMT, uintptr_t PC) { + if (EF->__sanitizer_symbolize_pc) + Printf("%s", DescribePC(SymbolizedFMT, PC).c_str()); + else + Printf(FallbackFMT, PC); +} + +unsigned NumberOfCpuCores() { + unsigned N = std::thread::hardware_concurrency(); + if (!N) { + Printf("WARNING: std::thread::hardware_concurrency not well defined for " + "your platform. Assuming CPU count of 1.\n"); + N = 1; + } + return N; +} + +bool ExecuteCommandAndReadOutput(const std::string &Command, std::string *Out) { + FILE *Pipe = OpenProcessPipe(Command.c_str(), "r"); + if (!Pipe) return false; + char Buff[1024]; + size_t N; + while ((N = fread(Buff, 1, sizeof(Buff), Pipe)) > 0) + Out->append(Buff, N); + return true; +} + +} // namespace fuzzer diff --git a/test/thirdparty/Fuzzer/FuzzerUtil.h b/test/thirdparty/Fuzzer/FuzzerUtil.h new file mode 100644 index 00000000..08058c56 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerUtil.h @@ -0,0 +1,72 @@ +//===- FuzzerUtil.h - Internal header for the Fuzzer Utils ------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Util functions. +//===----------------------------------------------------------------------===// + +#ifndef LLVM_FUZZER_UTIL_H +#define LLVM_FUZZER_UTIL_H + +#include "FuzzerDefs.h" + +namespace fuzzer { + +void PrintHexArray(const Unit &U, const char *PrintAfter = ""); + +void PrintHexArray(const uint8_t *Data, size_t Size, + const char *PrintAfter = ""); + +void PrintASCII(const uint8_t *Data, size_t Size, const char *PrintAfter = ""); + +void PrintASCII(const Unit &U, const char *PrintAfter = ""); + +// Changes U to contain only ASCII (isprint+isspace) characters. +// Returns true iff U has been changed. +bool ToASCII(uint8_t *Data, size_t Size); + +bool IsASCII(const Unit &U); + +bool IsASCII(const uint8_t *Data, size_t Size); + +std::string Base64(const Unit &U); + +void PrintPC(const char *SymbolizedFMT, const char *FallbackFMT, uintptr_t PC); + +std::string DescribePC(const char *SymbolizedFMT, uintptr_t PC); + +unsigned NumberOfCpuCores(); + +bool ExecuteCommandAndReadOutput(const std::string &Command, std::string *Out); + +// Platform specific functions. +void SetSignalHandler(const FuzzingOptions& Options); + +void SleepSeconds(int Seconds); + +unsigned long GetPid(); + +size_t GetPeakRSSMb(); + +int ExecuteCommand(const std::string &Command); + +FILE *OpenProcessPipe(const char *Command, const char *Mode); + +const void *SearchMemory(const void *haystack, size_t haystacklen, + const void *needle, size_t needlelen); + +std::string CloneArgsWithoutX(const std::vector &Args, + const char *X1, const char *X2); + +inline std::string CloneArgsWithoutX(const std::vector &Args, + const char *X) { + return CloneArgsWithoutX(Args, X, X); +} + +} // namespace fuzzer + +#endif // LLVM_FUZZER_UTIL_H diff --git a/test/thirdparty/Fuzzer/FuzzerUtilDarwin.cpp b/test/thirdparty/Fuzzer/FuzzerUtilDarwin.cpp new file mode 100644 index 00000000..9674368c --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerUtilDarwin.cpp @@ -0,0 +1,152 @@ +//===- FuzzerUtilDarwin.cpp - Misc utils ----------------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Misc utils for Darwin. +//===----------------------------------------------------------------------===// +#include "FuzzerDefs.h" +#if LIBFUZZER_APPLE + +#include "FuzzerIO.h" +#include +#include +#include +#include + +// There is no header for this on macOS so declare here +extern "C" char **environ; + +namespace fuzzer { + +static std::mutex SignalMutex; +// Global variables used to keep track of how signal handling should be +// restored. They should **not** be accessed without holding `SignalMutex`. +static int ActiveThreadCount = 0; +static struct sigaction OldSigIntAction; +static struct sigaction OldSigQuitAction; +static sigset_t OldBlockedSignalsSet; + +// This is a reimplementation of Libc's `system()`. On Darwin the Libc +// implementation contains a mutex which prevents it from being used +// concurrently. This implementation **can** be used concurrently. It sets the +// signal handlers when the first thread enters and restores them when the last +// thread finishes execution of the function and ensures this is not racey by +// using a mutex. +int ExecuteCommand(const std::string &Command) { + posix_spawnattr_t SpawnAttributes; + if (posix_spawnattr_init(&SpawnAttributes)) + return -1; + // Block and ignore signals of the current process when the first thread + // enters. + { + std::lock_guard Lock(SignalMutex); + if (ActiveThreadCount == 0) { + static struct sigaction IgnoreSignalAction; + sigset_t BlockedSignalsSet; + memset(&IgnoreSignalAction, 0, sizeof(IgnoreSignalAction)); + IgnoreSignalAction.sa_handler = SIG_IGN; + + if (sigaction(SIGINT, &IgnoreSignalAction, &OldSigIntAction) == -1) { + Printf("Failed to ignore SIGINT\n"); + (void)posix_spawnattr_destroy(&SpawnAttributes); + return -1; + } + if (sigaction(SIGQUIT, &IgnoreSignalAction, &OldSigQuitAction) == -1) { + Printf("Failed to ignore SIGQUIT\n"); + // Try our best to restore the signal handlers. + (void)sigaction(SIGINT, &OldSigIntAction, NULL); + (void)posix_spawnattr_destroy(&SpawnAttributes); + return -1; + } + + (void)sigemptyset(&BlockedSignalsSet); + (void)sigaddset(&BlockedSignalsSet, SIGCHLD); + if (sigprocmask(SIG_BLOCK, &BlockedSignalsSet, &OldBlockedSignalsSet) == + -1) { + Printf("Failed to block SIGCHLD\n"); + // Try our best to restore the signal handlers. + (void)sigaction(SIGQUIT, &OldSigQuitAction, NULL); + (void)sigaction(SIGINT, &OldSigIntAction, NULL); + (void)posix_spawnattr_destroy(&SpawnAttributes); + return -1; + } + } + ++ActiveThreadCount; + } + + // NOTE: Do not introduce any new `return` statements past this + // point. It is important that `ActiveThreadCount` always be decremented + // when leaving this function. + + // Make sure the child process uses the default handlers for the + // following signals rather than inheriting what the parent has. + sigset_t DefaultSigSet; + (void)sigemptyset(&DefaultSigSet); + (void)sigaddset(&DefaultSigSet, SIGQUIT); + (void)sigaddset(&DefaultSigSet, SIGINT); + (void)posix_spawnattr_setsigdefault(&SpawnAttributes, &DefaultSigSet); + // Make sure the child process doesn't block SIGCHLD + (void)posix_spawnattr_setsigmask(&SpawnAttributes, &OldBlockedSignalsSet); + short SpawnFlags = POSIX_SPAWN_SETSIGDEF | POSIX_SPAWN_SETSIGMASK; + (void)posix_spawnattr_setflags(&SpawnAttributes, SpawnFlags); + + pid_t Pid; + char **Environ = environ; // Read from global + const char *CommandCStr = Command.c_str(); + const char *Argv[] = {"sh", "-c", CommandCStr, NULL}; + int ErrorCode = 0, ProcessStatus = 0; + // FIXME: We probably shouldn't hardcode the shell path. + ErrorCode = posix_spawn(&Pid, "/bin/sh", NULL, &SpawnAttributes, + (char *const *)Argv, Environ); + (void)posix_spawnattr_destroy(&SpawnAttributes); + if (!ErrorCode) { + pid_t SavedPid = Pid; + do { + // Repeat until call completes uninterrupted. + Pid = waitpid(SavedPid, &ProcessStatus, /*options=*/0); + } while (Pid == -1 && errno == EINTR); + if (Pid == -1) { + // Fail for some other reason. + ProcessStatus = -1; + } + } else if (ErrorCode == ENOMEM || ErrorCode == EAGAIN) { + // Fork failure. + ProcessStatus = -1; + } else { + // Shell execution failure. + ProcessStatus = W_EXITCODE(127, 0); + } + + // Restore the signal handlers of the current process when the last thread + // using this function finishes. + { + std::lock_guard Lock(SignalMutex); + --ActiveThreadCount; + if (ActiveThreadCount == 0) { + bool FailedRestore = false; + if (sigaction(SIGINT, &OldSigIntAction, NULL) == -1) { + Printf("Failed to restore SIGINT handling\n"); + FailedRestore = true; + } + if (sigaction(SIGQUIT, &OldSigQuitAction, NULL) == -1) { + Printf("Failed to restore SIGQUIT handling\n"); + FailedRestore = true; + } + if (sigprocmask(SIG_BLOCK, &OldBlockedSignalsSet, NULL) == -1) { + Printf("Failed to unblock SIGCHLD\n"); + FailedRestore = true; + } + if (FailedRestore) + ProcessStatus = -1; + } + } + return ProcessStatus; +} + +} // namespace fuzzer + +#endif // LIBFUZZER_APPLE diff --git a/test/thirdparty/Fuzzer/FuzzerUtilLinux.cpp b/test/thirdparty/Fuzzer/FuzzerUtilLinux.cpp new file mode 100644 index 00000000..dfe7e6f4 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerUtilLinux.cpp @@ -0,0 +1,24 @@ +//===- FuzzerUtilLinux.cpp - Misc utils for Linux. ------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Misc utils for Linux. +//===----------------------------------------------------------------------===// +#include "FuzzerDefs.h" +#if LIBFUZZER_LINUX + +#include + +namespace fuzzer { + +int ExecuteCommand(const std::string &Command) { + return system(Command.c_str()); +} + +} // namespace fuzzer + +#endif // LIBFUZZER_LINUX diff --git a/test/thirdparty/Fuzzer/FuzzerUtilPosix.cpp b/test/thirdparty/Fuzzer/FuzzerUtilPosix.cpp new file mode 100644 index 00000000..8b484b8e --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerUtilPosix.cpp @@ -0,0 +1,117 @@ +//===- FuzzerUtilPosix.cpp - Misc utils for Posix. ------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Misc utils implementation using Posix API. +//===----------------------------------------------------------------------===// +#include "FuzzerDefs.h" +#if LIBFUZZER_POSIX +#include "FuzzerIO.h" +#include "FuzzerInternal.h" +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +namespace fuzzer { + +static void AlarmHandler(int, siginfo_t *, void *) { + Fuzzer::StaticAlarmCallback(); +} + +static void CrashHandler(int, siginfo_t *, void *) { + Fuzzer::StaticCrashSignalCallback(); +} + +static void InterruptHandler(int, siginfo_t *, void *) { + Fuzzer::StaticInterruptCallback(); +} + +static void SetSigaction(int signum, + void (*callback)(int, siginfo_t *, void *)) { + struct sigaction sigact; + memset(&sigact, 0, sizeof(sigact)); + sigact.sa_sigaction = callback; + if (sigaction(signum, &sigact, 0)) { + Printf("libFuzzer: sigaction failed with %d\n", errno); + exit(1); + } +} + +void SetTimer(int Seconds) { + struct itimerval T { + {Seconds, 0}, { Seconds, 0 } + }; + if (setitimer(ITIMER_REAL, &T, nullptr)) { + Printf("libFuzzer: setitimer failed with %d\n", errno); + exit(1); + } + SetSigaction(SIGALRM, AlarmHandler); +} + +void SetSignalHandler(const FuzzingOptions& Options) { + if (Options.UnitTimeoutSec > 0) + SetTimer(Options.UnitTimeoutSec / 2 + 1); + if (Options.HandleInt) + SetSigaction(SIGINT, InterruptHandler); + if (Options.HandleTerm) + SetSigaction(SIGTERM, InterruptHandler); + if (Options.HandleSegv) + SetSigaction(SIGSEGV, CrashHandler); + if (Options.HandleBus) + SetSigaction(SIGBUS, CrashHandler); + if (Options.HandleAbrt) + SetSigaction(SIGABRT, CrashHandler); + if (Options.HandleIll) + SetSigaction(SIGILL, CrashHandler); + if (Options.HandleFpe) + SetSigaction(SIGFPE, CrashHandler); +} + +void SleepSeconds(int Seconds) { + sleep(Seconds); // Use C API to avoid coverage from instrumented libc++. +} + +unsigned long GetPid() { return (unsigned long)getpid(); } + +size_t GetPeakRSSMb() { + struct rusage usage; + if (getrusage(RUSAGE_SELF, &usage)) + return 0; + if (LIBFUZZER_LINUX) { + // ru_maxrss is in KiB + return usage.ru_maxrss >> 10; + } else if (LIBFUZZER_APPLE) { + // ru_maxrss is in bytes + return usage.ru_maxrss >> 20; + } + assert(0 && "GetPeakRSSMb() is not implemented for your platform"); + return 0; +} + +FILE *OpenProcessPipe(const char *Command, const char *Mode) { + return popen(Command, Mode); +} + +const void *SearchMemory(const void *Data, size_t DataLen, const void *Patt, + size_t PattLen) { + return memmem(Data, DataLen, Patt, PattLen); +} + +} // namespace fuzzer + +#endif // LIBFUZZER_POSIX diff --git a/test/thirdparty/Fuzzer/FuzzerUtilWindows.cpp b/test/thirdparty/Fuzzer/FuzzerUtilWindows.cpp new file mode 100644 index 00000000..64adb7cd --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerUtilWindows.cpp @@ -0,0 +1,182 @@ +//===- FuzzerUtilWindows.cpp - Misc utils for Windows. --------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// Misc utils implementation for Windows. +//===----------------------------------------------------------------------===// +#include "FuzzerDefs.h" +#if LIBFUZZER_WINDOWS +#include "FuzzerIO.h" +#include "FuzzerInternal.h" +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +namespace fuzzer { + +static const FuzzingOptions* HandlerOpt = nullptr; + +LONG CALLBACK ExceptionHandler(PEXCEPTION_POINTERS ExceptionInfo) { + switch (ExceptionInfo->ExceptionRecord->ExceptionCode) { + case EXCEPTION_ACCESS_VIOLATION: + case EXCEPTION_ARRAY_BOUNDS_EXCEEDED: + case EXCEPTION_STACK_OVERFLOW: + if (HandlerOpt->HandleSegv) + Fuzzer::StaticCrashSignalCallback(); + break; + case EXCEPTION_DATATYPE_MISALIGNMENT: + case EXCEPTION_IN_PAGE_ERROR: + if (HandlerOpt->HandleBus) + Fuzzer::StaticCrashSignalCallback(); + break; + case EXCEPTION_ILLEGAL_INSTRUCTION: + case EXCEPTION_PRIV_INSTRUCTION: + if (HandlerOpt->HandleIll) + Fuzzer::StaticCrashSignalCallback(); + break; + case EXCEPTION_FLT_DENORMAL_OPERAND: + case EXCEPTION_FLT_DIVIDE_BY_ZERO: + case EXCEPTION_FLT_INEXACT_RESULT: + case EXCEPTION_FLT_INVALID_OPERATION: + case EXCEPTION_FLT_OVERFLOW: + case EXCEPTION_FLT_STACK_CHECK: + case EXCEPTION_FLT_UNDERFLOW: + case EXCEPTION_INT_DIVIDE_BY_ZERO: + case EXCEPTION_INT_OVERFLOW: + if (HandlerOpt->HandleFpe) + Fuzzer::StaticCrashSignalCallback(); + break; + } + return EXCEPTION_CONTINUE_SEARCH; +} + +BOOL WINAPI CtrlHandler(DWORD dwCtrlType) { + switch (dwCtrlType) { + case CTRL_C_EVENT: + if (HandlerOpt->HandleInt) + Fuzzer::StaticInterruptCallback(); + return TRUE; + case CTRL_BREAK_EVENT: + if (HandlerOpt->HandleTerm) + Fuzzer::StaticInterruptCallback(); + return TRUE; + } + return FALSE; +} + +void CALLBACK AlarmHandler(PVOID, BOOLEAN) { + Fuzzer::StaticAlarmCallback(); +} + +class TimerQ { + HANDLE TimerQueue; + public: + TimerQ() : TimerQueue(NULL) {}; + ~TimerQ() { + if (TimerQueue) + DeleteTimerQueueEx(TimerQueue, NULL); + }; + void SetTimer(int Seconds) { + if (!TimerQueue) { + TimerQueue = CreateTimerQueue(); + if (!TimerQueue) { + Printf("libFuzzer: CreateTimerQueue failed.\n"); + exit(1); + } + } + HANDLE Timer; + if (!CreateTimerQueueTimer(&Timer, TimerQueue, AlarmHandler, NULL, + Seconds*1000, Seconds*1000, 0)) { + Printf("libFuzzer: CreateTimerQueueTimer failed.\n"); + exit(1); + } + }; +}; + +static TimerQ Timer; + +static void CrashHandler(int) { Fuzzer::StaticCrashSignalCallback(); } + +void SetSignalHandler(const FuzzingOptions& Options) { + HandlerOpt = &Options; + + if (Options.UnitTimeoutSec > 0) + Timer.SetTimer(Options.UnitTimeoutSec / 2 + 1); + + if (Options.HandleInt || Options.HandleTerm) + if (!SetConsoleCtrlHandler(CtrlHandler, TRUE)) { + DWORD LastError = GetLastError(); + Printf("libFuzzer: SetConsoleCtrlHandler failed (Error code: %lu).\n", + LastError); + exit(1); + } + + if (Options.HandleSegv || Options.HandleBus || Options.HandleIll || + Options.HandleFpe) + if (!AddVectoredExceptionHandler(1, ExceptionHandler)) { + Printf("libFuzzer: AddVectoredExceptionHandler failed.\n"); + exit(1); + } + + if (Options.HandleAbrt) + if (SIG_ERR == signal(SIGABRT, CrashHandler)) { + Printf("libFuzzer: signal failed with %d\n", errno); + exit(1); + } +} + +void SleepSeconds(int Seconds) { Sleep(Seconds * 1000); } + +unsigned long GetPid() { return GetCurrentProcessId(); } + +size_t GetPeakRSSMb() { + PROCESS_MEMORY_COUNTERS info; + if (!GetProcessMemoryInfo(GetCurrentProcess(), &info, sizeof(info))) + return 0; + return info.PeakWorkingSetSize >> 20; +} + +FILE *OpenProcessPipe(const char *Command, const char *Mode) { + return _popen(Command, Mode); +} + +int ExecuteCommand(const std::string &Command) { + return system(Command.c_str()); +} + +const void *SearchMemory(const void *Data, size_t DataLen, const void *Patt, + size_t PattLen) { + // TODO: make this implementation more efficient. + const char *Cdata = (const char *)Data; + const char *Cpatt = (const char *)Patt; + + if (!Data || !Patt || DataLen == 0 || PattLen == 0 || DataLen < PattLen) + return NULL; + + if (PattLen == 1) + return memchr(Data, *Cpatt, DataLen); + + const char *End = Cdata + DataLen - PattLen + 1; + + for (const char *It = Cdata; It < End; ++It) + if (It[0] == Cpatt[0] && memcmp(It, Cpatt, PattLen) == 0) + return It; + + return NULL; +} + +} // namespace fuzzer + +#endif // LIBFUZZER_WINDOWS diff --git a/test/thirdparty/Fuzzer/FuzzerValueBitMap.h b/test/thirdparty/Fuzzer/FuzzerValueBitMap.h new file mode 100644 index 00000000..0692acd1 --- /dev/null +++ b/test/thirdparty/Fuzzer/FuzzerValueBitMap.h @@ -0,0 +1,87 @@ +//===- FuzzerValueBitMap.h - INTERNAL - Bit map -----------------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// ValueBitMap. +//===----------------------------------------------------------------------===// + +#ifndef LLVM_FUZZER_VALUE_BIT_MAP_H +#define LLVM_FUZZER_VALUE_BIT_MAP_H + +#include "FuzzerDefs.h" + +namespace fuzzer { + +// A bit map containing kMapSizeInWords bits. +struct ValueBitMap { + static const size_t kMapSizeInBits = 65371; // Prime. + static const size_t kMapSizeInBitsAligned = 65536; // 2^16 + static const size_t kBitsInWord = (sizeof(uintptr_t) * 8); + static const size_t kMapSizeInWords = kMapSizeInBitsAligned / kBitsInWord; + public: + static const size_t kNumberOfItems = kMapSizeInBits; + // Clears all bits. + void Reset() { memset(Map, 0, sizeof(Map)); } + + // Computes a hash function of Value and sets the corresponding bit. + // Returns true if the bit was changed from 0 to 1. + inline bool AddValue(uintptr_t Value) { + uintptr_t Idx = Value < kMapSizeInBits ? Value : Value % kMapSizeInBits; + uintptr_t WordIdx = Idx / kBitsInWord; + uintptr_t BitIdx = Idx % kBitsInWord; + uintptr_t Old = Map[WordIdx]; + uintptr_t New = Old | (1UL << BitIdx); + Map[WordIdx] = New; + return New != Old; + } + + inline bool Get(uintptr_t Idx) { + assert(Idx < kMapSizeInBits); + uintptr_t WordIdx = Idx / kBitsInWord; + uintptr_t BitIdx = Idx % kBitsInWord; + return Map[WordIdx] & (1UL << BitIdx); + } + + size_t GetNumBitsSinceLastMerge() const { return NumBits; } + + // Merges 'Other' into 'this', clears 'Other', updates NumBits, + // returns true if new bits were added. + ATTRIBUTE_TARGET_POPCNT + bool MergeFrom(ValueBitMap &Other) { + uintptr_t Res = 0; + size_t OldNumBits = NumBits; + for (size_t i = 0; i < kMapSizeInWords; i++) { + auto O = Other.Map[i]; + auto M = Map[i]; + if (O) { + Map[i] = (M |= O); + Other.Map[i] = 0; + } + if (M) + Res += __builtin_popcountl(M); + } + NumBits = Res; + return OldNumBits < NumBits; + } + + template + void ForEach(Callback CB) { + for (size_t i = 0; i < kMapSizeInWords; i++) + if (uintptr_t M = Map[i]) + for (size_t j = 0; j < sizeof(M) * 8; j++) + if (M & ((uintptr_t)1 << j)) + CB(i * sizeof(M) * 8 + j); + } + + private: + size_t NumBits = 0; + uintptr_t Map[kMapSizeInWords] __attribute__((aligned(512))); +}; + +} // namespace fuzzer + +#endif // LLVM_FUZZER_VALUE_BIT_MAP_H diff --git a/test/thirdparty/Fuzzer/README.txt b/test/thirdparty/Fuzzer/README.txt new file mode 100644 index 00000000..79f49b55 --- /dev/null +++ b/test/thirdparty/Fuzzer/README.txt @@ -0,0 +1,2 @@ +Move to http://llvm.org/docs/LibFuzzer.html + diff --git a/test/thirdparty/Fuzzer/afl/afl_driver.cpp b/test/thirdparty/Fuzzer/afl/afl_driver.cpp new file mode 100644 index 00000000..fc958955 --- /dev/null +++ b/test/thirdparty/Fuzzer/afl/afl_driver.cpp @@ -0,0 +1,295 @@ +//===- afl_driver.cpp - a glue between AFL and libFuzzer --------*- C++ -* ===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +//===----------------------------------------------------------------------===// + +/* This file allows to fuzz libFuzzer-style target functions + (LLVMFuzzerTestOneInput) with AFL using AFL's persistent (in-process) mode. + +Usage: +################################################################################ +cat << EOF > test_fuzzer.cc +#include +#include +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *data, size_t size) { + if (size > 0 && data[0] == 'H') + if (size > 1 && data[1] == 'I') + if (size > 2 && data[2] == '!') + __builtin_trap(); + return 0; +} +EOF +# Build your target with -fsanitize-coverage=trace-pc using fresh clang. +clang -g -fsanitize-coverage=trace-pc test_fuzzer.cc -c +# Build afl-llvm-rt.o.c from the AFL distribution. +clang -c -w $AFL_HOME/llvm_mode/afl-llvm-rt.o.c +# Build this file, link it with afl-llvm-rt.o.o and the target code. +clang++ afl_driver.cpp test_fuzzer.o afl-llvm-rt.o.o +# Run AFL: +rm -rf IN OUT; mkdir IN OUT; echo z > IN/z; +$AFL_HOME/afl-fuzz -i IN -o OUT ./a.out +################################################################################ +Environment Variables: +There are a few environment variables that can be set to use features that +afl-fuzz doesn't have. + +AFL_DRIVER_STDERR_DUPLICATE_FILENAME: Setting this *appends* stderr to the file +specified. If the file does not exist, it is created. This is useful for getting +stack traces (when using ASAN for example) or original error messages on hard to +reproduce bugs. + +AFL_DRIVER_EXTRA_STATS_FILENAME: Setting this causes afl_driver to write extra +statistics to the file specified. Currently these are peak_rss_mb +(the peak amount of virtual memory used in MB) and slowest_unit_time_secs. If +the file does not exist it is created. If the file does exist then +afl_driver assumes it was restarted by afl-fuzz and will try to read old +statistics from the file. If that fails then the process will quit. + +*/ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +// Platform detection. Copied from FuzzerInternal.h +#ifdef __linux__ +#define LIBFUZZER_LINUX 1 +#define LIBFUZZER_APPLE 0 +#elif __APPLE__ +#define LIBFUZZER_LINUX 0 +#define LIBFUZZER_APPLE 1 +#else +#error "Support for your platform has not been implemented" +#endif + +// Used to avoid repeating error checking boilerplate. If cond is false, a +// fatal error has occured in the program. In this event print error_message +// to stderr and abort(). Otherwise do nothing. Note that setting +// AFL_DRIVER_STDERR_DUPLICATE_FILENAME may cause error_message to be appended +// to the file as well, if the error occurs after the duplication is performed. +#define CHECK_ERROR(cond, error_message) \ + if (!(cond)) { \ + fprintf(stderr, (error_message)); \ + abort(); \ + } + +// libFuzzer interface is thin, so we don't include any libFuzzer headers. +extern "C" { +int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size); +__attribute__((weak)) int LLVMFuzzerInitialize(int *argc, char ***argv); +} + +// Notify AFL about persistent mode. +static volatile char AFL_PERSISTENT[] = "##SIG_AFL_PERSISTENT##"; +extern "C" int __afl_persistent_loop(unsigned int); +static volatile char suppress_warning2 = AFL_PERSISTENT[0]; + +// Notify AFL about deferred forkserver. +static volatile char AFL_DEFER_FORKSVR[] = "##SIG_AFL_DEFER_FORKSRV##"; +extern "C" void __afl_manual_init(); +static volatile char suppress_warning1 = AFL_DEFER_FORKSVR[0]; + +// Input buffer. +static const size_t kMaxAflInputSize = 1 << 20; +static uint8_t AflInputBuf[kMaxAflInputSize]; + +// Variables we need for writing to the extra stats file. +static FILE *extra_stats_file = NULL; +static uint32_t previous_peak_rss = 0; +static time_t slowest_unit_time_secs = 0; +static const int kNumExtraStats = 2; +static const char *kExtraStatsFormatString = "peak_rss_mb : %u\n" + "slowest_unit_time_sec : %u\n"; + +// Copied from FuzzerUtil.cpp. +size_t GetPeakRSSMb() { + struct rusage usage; + if (getrusage(RUSAGE_SELF, &usage)) + return 0; + if (LIBFUZZER_LINUX) { + // ru_maxrss is in KiB + return usage.ru_maxrss >> 10; + } else if (LIBFUZZER_APPLE) { + // ru_maxrss is in bytes + return usage.ru_maxrss >> 20; + } + assert(0 && "GetPeakRSSMb() is not implemented for your platform"); + return 0; +} + +// Based on SetSigaction in FuzzerUtil.cpp +static void SetSigaction(int signum, + void (*callback)(int, siginfo_t *, void *)) { + struct sigaction sigact; + memset(&sigact, 0, sizeof(sigact)); + sigact.sa_sigaction = callback; + if (sigaction(signum, &sigact, 0)) { + fprintf(stderr, "libFuzzer: sigaction failed with %d\n", errno); + exit(1); + } +} + +// Write extra stats to the file specified by the user. If none is specified +// this function will never be called. +static void write_extra_stats() { + uint32_t peak_rss = GetPeakRSSMb(); + + if (peak_rss < previous_peak_rss) + peak_rss = previous_peak_rss; + + int chars_printed = fprintf(extra_stats_file, kExtraStatsFormatString, + peak_rss, slowest_unit_time_secs); + + CHECK_ERROR(chars_printed != 0, "Failed to write extra_stats_file"); + + CHECK_ERROR(fclose(extra_stats_file) == 0, + "Failed to close extra_stats_file"); +} + +// Call write_extra_stats before we exit. +static void crash_handler(int, siginfo_t *, void *) { + // Make sure we don't try calling write_extra_stats again if we crashed while + // trying to call it. + static bool first_crash = true; + CHECK_ERROR(first_crash, + "Crashed in crash signal handler. This is a bug in the fuzzer."); + + first_crash = false; + write_extra_stats(); +} + +// If the user has specified an extra_stats_file through the environment +// variable AFL_DRIVER_EXTRA_STATS_FILENAME, then perform necessary set up +// to write stats to it on exit. If no file is specified, do nothing. Otherwise +// install signal and exit handlers to write to the file when the process exits. +// Then if the file doesn't exist create it and set extra stats to 0. But if it +// does exist then read the initial values of the extra stats from the file +// and check that the file is writable. +static void maybe_initialize_extra_stats() { + // If AFL_DRIVER_EXTRA_STATS_FILENAME isn't set then we have nothing to do. + char *extra_stats_filename = getenv("AFL_DRIVER_EXTRA_STATS_FILENAME"); + if (!extra_stats_filename) + return; + + // Open the file and find the previous peak_rss_mb value. + // This is necessary because the fuzzing process is restarted after N + // iterations are completed. So we may need to get this value from a previous + // process to be accurate. + extra_stats_file = fopen(extra_stats_filename, "r"); + + // If extra_stats_file already exists: read old stats from it. + if (extra_stats_file) { + int matches = fscanf(extra_stats_file, kExtraStatsFormatString, + &previous_peak_rss, &slowest_unit_time_secs); + + // Make sure we have read a real extra stats file and that we have used it + // to set slowest_unit_time_secs and previous_peak_rss. + CHECK_ERROR(matches == kNumExtraStats, "Extra stats file is corrupt"); + + CHECK_ERROR(fclose(extra_stats_file) == 0, "Failed to close file"); + + // Now open the file for writing. + extra_stats_file = fopen(extra_stats_filename, "w"); + CHECK_ERROR(extra_stats_file, + "Failed to open extra stats file for writing"); + } else { + // Looks like this is the first time in a fuzzing job this is being called. + extra_stats_file = fopen(extra_stats_filename, "w+"); + CHECK_ERROR(extra_stats_file, "failed to create extra stats file"); + } + + // Make sure that crash_handler gets called on any kind of fatal error. + int crash_signals[] = {SIGSEGV, SIGBUS, SIGABRT, SIGILL, SIGFPE, SIGINT, + SIGTERM}; + + const size_t num_signals = sizeof(crash_signals) / sizeof(crash_signals[0]); + + for (size_t idx = 0; idx < num_signals; idx++) + SetSigaction(crash_signals[idx], crash_handler); + + // Make sure it gets called on other kinds of exits. + atexit(write_extra_stats); +} + +// If the user asks us to duplicate stderr, then do it. +static void maybe_duplicate_stderr() { + char* stderr_duplicate_filename = + getenv("AFL_DRIVER_STDERR_DUPLICATE_FILENAME"); + + if (!stderr_duplicate_filename) + return; + + FILE* stderr_duplicate_stream = + freopen(stderr_duplicate_filename, "a+", stderr); + + if (!stderr_duplicate_stream) { + fprintf( + stderr, + "Failed to duplicate stderr to AFL_DRIVER_STDERR_DUPLICATE_FILENAME"); + abort(); + } +} + +int main(int argc, char **argv) { + fprintf(stderr, "======================= INFO =========================\n" + "This binary is built for AFL-fuzz.\n" + "To run the target function on a single input execute this:\n" + " %s < INPUT_FILE\n" + "To run the fuzzing execute this:\n" + " afl-fuzz [afl-flags] %s [N] " + "-- run N fuzzing iterations before " + "re-spawning the process (default: 1000)\n" + "======================================================\n", + argv[0], argv[0]); + if (LLVMFuzzerInitialize) + LLVMFuzzerInitialize(&argc, &argv); + // Do any other expensive one-time initialization here. + + maybe_duplicate_stderr(); + maybe_initialize_extra_stats(); + + __afl_manual_init(); + + int N = 1000; + if (argc >= 2) + N = atoi(argv[1]); + assert(N > 0); + time_t unit_time_secs; + int num_runs = 0; + while (__afl_persistent_loop(N)) { + ssize_t n_read = read(0, AflInputBuf, kMaxAflInputSize); + if (n_read > 0) { + // Copy AflInputBuf into a separate buffer to let asan find buffer + // overflows. Don't use unique_ptr/etc to avoid extra dependencies. + uint8_t *copy = new uint8_t[n_read]; + memcpy(copy, AflInputBuf, n_read); + + struct timeval unit_start_time; + CHECK_ERROR(gettimeofday(&unit_start_time, NULL) == 0, + "Calling gettimeofday failed"); + + num_runs++; + LLVMFuzzerTestOneInput(copy, n_read); + + struct timeval unit_stop_time; + CHECK_ERROR(gettimeofday(&unit_stop_time, NULL) == 0, + "Calling gettimeofday failed"); + + // Update slowest_unit_time_secs if we see a new max. + unit_time_secs = unit_stop_time.tv_sec - unit_start_time.tv_sec; + if (slowest_unit_time_secs < unit_time_secs) + slowest_unit_time_secs = unit_time_secs; + + delete[] copy; + } + } + fprintf(stderr, "%s: successfully executed %d input(s)\n", argv[0], num_runs); +} diff --git a/test/thirdparty/Fuzzer/build.sh b/test/thirdparty/Fuzzer/build.sh new file mode 100755 index 00000000..27c148ad --- /dev/null +++ b/test/thirdparty/Fuzzer/build.sh @@ -0,0 +1,10 @@ +#!/bin/bash +LIBFUZZER_SRC_DIR=$(dirname $0) +for f in $LIBFUZZER_SRC_DIR/*.cpp; do + clang -g -O2 -fno-omit-frame-pointer -std=c++11 $f -c & +done +wait +rm -f libFuzzer.a +ar ru libFuzzer.a Fuzzer*.o +rm -f Fuzzer*.o + diff --git a/test/thirdparty/Fuzzer/cxx.dict b/test/thirdparty/Fuzzer/cxx.dict new file mode 100644 index 00000000..41350f47 --- /dev/null +++ b/test/thirdparty/Fuzzer/cxx.dict @@ -0,0 +1,122 @@ +"++" +"--" +"<<" +">>" +"+=" +"-=" +"*=" +"/=" +">>=" +"<<=" +"&=" +"|=" +"^=" +"%=" +"!=" +"&&" +"||" +"==" +">=" +"<=" +"->" +"alignas" +"alignof" +"and" +"and_eq" +"asm" +"auto" +"bitand" +"bitor" +"bool" +"break" +"case" +"catch" +"char" +"char16_t" +"char32_t" +"class" +"compl" +"concept" +"const" +"constexpr" +"const_cast" +"continue" +"decltype" +"default" +"delete" +"do" +"double" +"dynamic_cast" +"else" +"enum" +"explicit" +"export" +"extern" +"false" +"float" +"for" +"friend" +"goto" +"if" +"inline" +"int" +"long" +"mutable" +"namespace" +"new" +"noexcept" +"not" +"not_eq" +"nullptr" +"operator" +"or" +"or_eq" +"private" +"protected" +"public" +"register" +"reinterpret_cast" +"requires" +"return" +"short" +"signed" +"sizeof" +"static" +"static_assert" +"static_cast" +"struct" +"switch" +"template" +"this" +"thread_local" +"throw" +"true" +"try" +"typedef" +"typeid" +"typename" +"union" +"unsigned" +"using" +"virtual" +"void" +"volatile" +"wchar_t" +"while" +"xor" +"xor_eq" +"if" +"elif" +"else" +"endif" +"defined" +"ifdef" +"ifndef" +"define" +"undef" +"include" +"line" +"error" +"pragma" +"override" +"final" diff --git a/test/thirdparty/Fuzzer/standalone/StandaloneFuzzTargetMain.c b/test/thirdparty/Fuzzer/standalone/StandaloneFuzzTargetMain.c new file mode 100644 index 00000000..0d76ea49 --- /dev/null +++ b/test/thirdparty/Fuzzer/standalone/StandaloneFuzzTargetMain.c @@ -0,0 +1,41 @@ +/*===- StandaloneFuzzTargetMain.c - standalone main() for fuzz targets. ---===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// This main() function can be linked to a fuzz target (i.e. a library +// that exports LLVMFuzzerTestOneInput() and possibly LLVMFuzzerInitialize()) +// instead of libFuzzer. This main() function will not perform any fuzzing +// but will simply feed all input files one by one to the fuzz target. +// +// Use this file to provide reproducers for bugs when linking against libFuzzer +// or other fuzzing engine is undesirable. +//===----------------------------------------------------------------------===*/ +#include +#include +#include + +extern int LLVMFuzzerTestOneInput(const unsigned char *data, size_t size); +__attribute__((weak)) extern int LLVMFuzzerInitialize(int *argc, char ***argv); +int main(int argc, char **argv) { + fprintf(stderr, "StandaloneFuzzTargetMain: running %d inputs\n", argc - 1); + if (LLVMFuzzerInitialize) + LLVMFuzzerInitialize(&argc, &argv); + for (int i = 1; i < argc; i++) { + fprintf(stderr, "Running: %s\n", argv[i]); + FILE *f = fopen(argv[i], "r"); + assert(f); + fseek(f, 0, SEEK_END); + size_t len = ftell(f); + fseek(f, 0, SEEK_SET); + unsigned char *buf = (unsigned char*)malloc(len); + size_t n_read = fread(buf, 1, len, f); + assert(n_read == len); + LLVMFuzzerTestOneInput(buf, len); + free(buf); + fprintf(stderr, "Done: %s: (%zd bytes)\n", argv[i], n_read); + } +} diff --git a/test/thirdparty/Fuzzer/test/AFLDriverTest.cpp b/test/thirdparty/Fuzzer/test/AFLDriverTest.cpp new file mode 100644 index 00000000..3dd0b611 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/AFLDriverTest.cpp @@ -0,0 +1,22 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Contains dummy functions used to avoid dependency on AFL. +#include +#include + +extern "C" void __afl_manual_init() {} + +extern "C" int __afl_persistent_loop(unsigned int) { + return 0; +} + +// This declaration exists to prevent the Darwin linker +// from complaining about this being a missing weak symbol. +extern "C" int LLVMFuzzerInitialize(int *argc, char ***argv) { + return 0; +} + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/AbsNegAndConstant64Test.cpp b/test/thirdparty/Fuzzer/test/AbsNegAndConstant64Test.cpp new file mode 100644 index 00000000..57748143 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/AbsNegAndConstant64Test.cpp @@ -0,0 +1,23 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// abs(x) < 0 and y == Const puzzle, 64-bit variant. +#include +#include +#include +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size < 16) return 0; + int64_t x; + uint64_t y; + memcpy(&x, Data, sizeof(x)); + memcpy(&y, Data + sizeof(x), sizeof(y)); + if (labs(x) < 0 && y == 0xbaddcafedeadbeefUL) { + printf("BINGO; Found the target, exiting; x = 0x%lx y 0x%lx\n", x, y); + exit(1); + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/AbsNegAndConstantTest.cpp b/test/thirdparty/Fuzzer/test/AbsNegAndConstantTest.cpp new file mode 100644 index 00000000..69075a45 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/AbsNegAndConstantTest.cpp @@ -0,0 +1,23 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// abs(x) < 0 and y == Const puzzle. +#include +#include +#include +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size < 8) return 0; + int x; + unsigned y; + memcpy(&x, Data, sizeof(x)); + memcpy(&y, Data + sizeof(x), sizeof(y)); + if (abs(x) < 0 && y == 0xbaddcafe) { + printf("BINGO; Found the target, exiting; x = 0x%x y 0x%x\n", x, y); + exit(1); + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/AccumulateAllocationsTest.cpp b/test/thirdparty/Fuzzer/test/AccumulateAllocationsTest.cpp new file mode 100644 index 00000000..604d8fa2 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/AccumulateAllocationsTest.cpp @@ -0,0 +1,17 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Test with a more mallocs than frees, but no leak. +#include +#include + +const int kAllocatedPointersSize = 10000; +int NumAllocatedPointers = 0; +int *AllocatedPointers[kAllocatedPointersSize]; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (NumAllocatedPointers < kAllocatedPointersSize) + AllocatedPointers[NumAllocatedPointers++] = new int; + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/BufferOverflowOnInput.cpp b/test/thirdparty/Fuzzer/test/BufferOverflowOnInput.cpp new file mode 100644 index 00000000..b9d14052 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/BufferOverflowOnInput.cpp @@ -0,0 +1,23 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find the string "Hi!". +#include +#include +#include +#include +#include + +static volatile bool SeedLargeBuffer; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + assert(Data); + if (Size >= 4) + SeedLargeBuffer = true; + if (Size == 3 && SeedLargeBuffer && Data[3]) { + std::cout << "Woops, reading Data[3] w/o crashing\n"; + exit(1); + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/CMakeLists.txt b/test/thirdparty/Fuzzer/test/CMakeLists.txt new file mode 100644 index 00000000..c0457746 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/CMakeLists.txt @@ -0,0 +1,217 @@ +# Build all these tests with -O0, otherwise optimizations may merge some +# basic blocks and we'll fail to discover the targets. +# We change the flags for every build type because we might be doing +# a multi-configuration build (e.g. Xcode) where CMAKE_BUILD_TYPE doesn't +# mean anything. +set(variables_to_filter + CMAKE_CXX_FLAGS_RELEASE + CMAKE_CXX_FLAGS_DEBUG + CMAKE_CXX_FLAGS_RELWITHDEBINFO + CMAKE_CXX_FLAGS_MINSIZEREL + LIBFUZZER_FLAGS_BASE + ) +foreach (VARNAME ${variables_to_filter}) + string(REPLACE " " ";" BUILD_FLAGS_AS_LIST "${${VARNAME}}") + set(new_flags "") + foreach (flag ${BUILD_FLAGS_AS_LIST}) + # NOTE: Use of XX here is to avoid a CMake warning due to CMP0054 + if (NOT ("XX${flag}" MATCHES "XX-O[0123s]")) + set(new_flags "${new_flags} ${flag}") + else() + set(new_flags "${new_flags} -O0") + endif() + endforeach() + set(${VARNAME} "${new_flags}") +endforeach() + +# Enable the coverage instrumentation (it is disabled for the Fuzzer lib). +set(CMAKE_CXX_FLAGS "${LIBFUZZER_FLAGS_BASE} -fsanitize-coverage=trace-pc-guard,indirect-calls,trace-cmp,trace-div,trace-gep -g") + +# add_libfuzzer_test( +# SOURCES source0.cpp [source1.cpp ...] +# ) +# +# Declares a LibFuzzer test executable with target name LLVMFuzzer-. +# +# One or more source files to be compiled into the binary must be declared +# after the SOURCES keyword. +function(add_libfuzzer_test name) + set(multi_arg_options "SOURCES") + cmake_parse_arguments( + "add_libfuzzer_test" "" "" "${multi_arg_options}" ${ARGN}) + if ("${add_libfuzzer_test_SOURCES}" STREQUAL "") + message(FATAL_ERROR "Source files must be specified") + endif() + add_executable(LLVMFuzzer-${name} + ${add_libfuzzer_test_SOURCES} + ) + target_link_libraries(LLVMFuzzer-${name} LLVMFuzzer) + # Place binary where llvm-lit expects to find it + set_target_properties(LLVMFuzzer-${name} + PROPERTIES RUNTIME_OUTPUT_DIRECTORY + "${CMAKE_BINARY_DIR}/lib/Fuzzer/test" + ) + set(TestBinaries ${TestBinaries} LLVMFuzzer-${name} PARENT_SCOPE) +endfunction() + +# Variable to keep track of all test targets +set(TestBinaries) + +############################################################################### +# Basic tests +############################################################################### + +set(Tests + AbsNegAndConstantTest + AbsNegAndConstant64Test + AccumulateAllocationsTest + BufferOverflowOnInput + CallerCalleeTest + CounterTest + CustomCrossOverTest + CustomMutatorTest + DivTest + EmptyTest + FourIndependentBranchesTest + FullCoverageSetTest + InitializeTest + MemcmpTest + LeakTest + LeakTimeoutTest + LoadTest + NullDerefTest + NullDerefOnEmptyTest + NthRunCrashTest + OneHugeAllocTest + OutOfMemoryTest + OutOfMemorySingleLargeMallocTest + RepeatedMemcmp + RepeatedBytesTest + SimpleCmpTest + SimpleDictionaryTest + SimpleHashTest + SimpleTest + SimpleThreadedTest + SingleMemcmpTest + SingleStrcmpTest + SingleStrncmpTest + SpamyTest + ShrinkControlFlowTest + ShrinkValueProfileTest + StrcmpTest + StrncmpOOBTest + StrncmpTest + StrstrTest + SwapCmpTest + SwitchTest + Switch2Test + ThreadedLeakTest + ThreadedTest + TimeoutTest + TimeoutEmptyTest + TraceMallocTest + ) + +if(APPLE) + # LeakSanitizer is not supported on OSX right now + set(HAS_LSAN 0) + message(WARNING "LeakSanitizer is not supported on Apple platforms." + " Building and running LibFuzzer LeakSanitizer tests is disabled." + ) +else() + set(HAS_LSAN 1) +endif() + +foreach(Test ${Tests}) + add_libfuzzer_test(${Test} SOURCES ${Test}.cpp) +endforeach() + + +############################################################################### +# Unit tests +############################################################################### + +add_executable(LLVMFuzzer-Unittest + FuzzerUnittest.cpp + ) + +add_executable(LLVMFuzzer-StandaloneInitializeTest + InitializeTest.cpp + ../standalone/StandaloneFuzzTargetMain.c + ) + +target_link_libraries(LLVMFuzzer-Unittest + gtest + gtest_main + LLVMFuzzerNoMain + ) + +target_include_directories(LLVMFuzzer-Unittest PRIVATE + "${LLVM_MAIN_SRC_DIR}/utils/unittest/googletest/include" + ) + +set(TestBinaries ${TestBinaries} LLVMFuzzer-Unittest) +set_target_properties(LLVMFuzzer-Unittest + PROPERTIES RUNTIME_OUTPUT_DIRECTORY + "${CMAKE_CURRENT_BINARY_DIR}" +) + +set(TestBinaries ${TestBinaries} LLVMFuzzer-StandaloneInitializeTest) +set_target_properties(LLVMFuzzer-StandaloneInitializeTest + PROPERTIES RUNTIME_OUTPUT_DIRECTORY + "${CMAKE_CURRENT_BINARY_DIR}" +) + +############################################################################### +# Additional tests +############################################################################### + +include_directories(..) + +# add_subdirectory(uninstrumented) +add_subdirectory(no-coverage) +add_subdirectory(ubsan) + +add_library(LLVMFuzzer-DSO1 SHARED DSO1.cpp) +add_library(LLVMFuzzer-DSO2 SHARED DSO2.cpp) + +add_executable(LLVMFuzzer-DSOTest + DSOTestMain.cpp + DSOTestExtra.cpp) + +target_link_libraries(LLVMFuzzer-DSOTest + LLVMFuzzer-DSO1 + LLVMFuzzer-DSO2 + LLVMFuzzer + ) + +set_target_properties(LLVMFuzzer-DSOTest PROPERTIES RUNTIME_OUTPUT_DIRECTORY + "${CMAKE_BINARY_DIR}/lib/Fuzzer/test") +set_target_properties(LLVMFuzzer-DSO1 PROPERTIES LIBRARY_OUTPUT_DIRECTORY + "${CMAKE_BINARY_DIR}/lib/Fuzzer/lib") +set_target_properties(LLVMFuzzer-DSO2 PROPERTIES LIBRARY_OUTPUT_DIRECTORY + "${CMAKE_BINARY_DIR}/lib/Fuzzer/lib") + +set(TestBinaries ${TestBinaries} LLVMFuzzer-DSOTest) + +############################################################################### +# Configure lit to run the tests +# +# Note this is done after declaring all tests so we can inform lit if any tests +# need to be disabled. +############################################################################### + +configure_lit_site_cfg( + ${CMAKE_CURRENT_SOURCE_DIR}/lit.site.cfg.in + ${CMAKE_CURRENT_BINARY_DIR}/lit.site.cfg + ) + +configure_lit_site_cfg( + ${CMAKE_CURRENT_SOURCE_DIR}/unit/lit.site.cfg.in + ${CMAKE_CURRENT_BINARY_DIR}/unit/lit.site.cfg + ) + +add_lit_testsuite(check-fuzzer "Running Fuzzer tests" + ${CMAKE_CURRENT_BINARY_DIR} + DEPENDS ${TestBinaries} FileCheck not + ) diff --git a/test/thirdparty/Fuzzer/test/CallerCalleeTest.cpp b/test/thirdparty/Fuzzer/test/CallerCalleeTest.cpp new file mode 100644 index 00000000..3ec025d0 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/CallerCalleeTest.cpp @@ -0,0 +1,59 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. +// Try to find the target using the indirect caller-callee pairs. +#include +#include +#include +#include +#include + +typedef void (*F)(); +static F t[256]; + +void f34() { + std::cerr << "BINGO\n"; + exit(1); +} +void f23() { t[(unsigned)'d'] = f34;} +void f12() { t[(unsigned)'c'] = f23;} +void f01() { t[(unsigned)'b'] = f12;} +void f00() {} + +static F t0[256] = { + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, + f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, f00, +}; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size < 4) return 0; + // Spoof the counters. + for (int i = 0; i < 200; i++) { + f23(); + f12(); + f01(); + } + memcpy(t, t0, sizeof(t)); + t[(unsigned)'a'] = f01; + t[Data[0]](); + t[Data[1]](); + t[Data[2]](); + t[Data[3]](); + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/CounterTest.cpp b/test/thirdparty/Fuzzer/test/CounterTest.cpp new file mode 100644 index 00000000..4917934c --- /dev/null +++ b/test/thirdparty/Fuzzer/test/CounterTest.cpp @@ -0,0 +1,18 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Test for a fuzzer: must find the case where a particular basic block is +// executed many times. +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + int Num = 0; + for (size_t i = 0; i < Size; i++) + if (Data[i] == 'A' + i) + Num++; + if (Num >= 4) { + std::cerr << "BINGO!\n"; + exit(1); + } + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/CustomCrossOverTest.cpp b/test/thirdparty/Fuzzer/test/CustomCrossOverTest.cpp new file mode 100644 index 00000000..b624088b --- /dev/null +++ b/test/thirdparty/Fuzzer/test/CustomCrossOverTest.cpp @@ -0,0 +1,63 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a cutom mutator. +#include +#include +#include +#include +#include +#include +#include + +#include "FuzzerInterface.h" + +static const char *Separator = "-_^_-"; +static const char *Target = "012-_^_-abc"; + +static volatile int sink; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + assert(Data); + std::string Str(reinterpret_cast(Data), Size); + + // Ensure that two different elements exist in the corpus. + if (Size && Data[0] == '0') sink++; + if (Size && Data[0] == 'a') sink--; + + if (Str.find(Target) != std::string::npos) { + std::cout << "BINGO; Found the target, exiting\n"; + exit(1); + } + return 0; +} + +extern "C" size_t LLVMFuzzerCustomCrossOver(const uint8_t *Data1, size_t Size1, + const uint8_t *Data2, size_t Size2, + uint8_t *Out, size_t MaxOutSize, + unsigned int Seed) { + static bool Printed; + static size_t SeparatorLen = strlen(Separator); + + if (!Printed) { + std::cerr << "In LLVMFuzzerCustomCrossover\n"; + Printed = true; + } + + std::mt19937 R(Seed); + + size_t Offset1 = 0; + size_t Len1 = R() % (Size1 - Offset1); + size_t Offset2 = 0; + size_t Len2 = R() % (Size2 - Offset2); + size_t Size = Len1 + Len2 + SeparatorLen; + + if (Size > MaxOutSize) + return 0; + + memcpy(Out, Data1 + Offset1, Len1); + memcpy(Out + Len1, Separator, SeparatorLen); + memcpy(Out + Len1 + SeparatorLen, Data2 + Offset2, Len2); + + return Len1 + Len2 + SeparatorLen; +} diff --git a/test/thirdparty/Fuzzer/test/CustomMutatorTest.cpp b/test/thirdparty/Fuzzer/test/CustomMutatorTest.cpp new file mode 100644 index 00000000..4f84519a --- /dev/null +++ b/test/thirdparty/Fuzzer/test/CustomMutatorTest.cpp @@ -0,0 +1,38 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a cutom mutator. +#include +#include +#include +#include +#include + +#include "FuzzerInterface.h" + +static volatile int Sink; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + assert(Data); + if (Size > 0 && Data[0] == 'H') { + Sink = 1; + if (Size > 1 && Data[1] == 'i') { + Sink = 2; + if (Size > 2 && Data[2] == '!') { + std::cout << "BINGO; Found the target, exiting\n"; + exit(1); + } + } + } + return 0; +} + +extern "C" size_t LLVMFuzzerCustomMutator(uint8_t *Data, size_t Size, + size_t MaxSize, unsigned int Seed) { + static bool Printed; + if (!Printed) { + std::cerr << "In LLVMFuzzerCustomMutator\n"; + Printed = true; + } + return LLVMFuzzerMutate(Data, Size, MaxSize); +} diff --git a/test/thirdparty/Fuzzer/test/DSO1.cpp b/test/thirdparty/Fuzzer/test/DSO1.cpp new file mode 100644 index 00000000..4a293890 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/DSO1.cpp @@ -0,0 +1,12 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Source code for a simple DSO. + +int DSO1(int a) { + if (a < 123456) + return 0; + return 1; +} + +void Uncovered1() { } diff --git a/test/thirdparty/Fuzzer/test/DSO2.cpp b/test/thirdparty/Fuzzer/test/DSO2.cpp new file mode 100644 index 00000000..04b308d1 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/DSO2.cpp @@ -0,0 +1,12 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Source code for a simple DSO. + +int DSO2(int a) { + if (a < 3598235) + return 0; + return 1; +} + +void Uncovered2() {} diff --git a/test/thirdparty/Fuzzer/test/DSOTestExtra.cpp b/test/thirdparty/Fuzzer/test/DSOTestExtra.cpp new file mode 100644 index 00000000..a2274d07 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/DSOTestExtra.cpp @@ -0,0 +1,11 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Source code for a simple DSO. + +int DSOTestExtra(int a) { + if (a < 452345) + return 0; + return 1; +} + diff --git a/test/thirdparty/Fuzzer/test/DSOTestMain.cpp b/test/thirdparty/Fuzzer/test/DSOTestMain.cpp new file mode 100644 index 00000000..3e225d88 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/DSOTestMain.cpp @@ -0,0 +1,31 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Source code for a simple DSO. + +#include +#include +#include +#include +extern int DSO1(int a); +extern int DSO2(int a); +extern int DSOTestExtra(int a); + +static volatile int *nil = 0; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + int x, y, z; + if (Size < sizeof(int) * 3) { + x = y = z = 0; + } else { + memcpy(&x, Data + 0 * sizeof(int), sizeof(int)); + memcpy(&y, Data + 1 * sizeof(int), sizeof(int)); + memcpy(&z, Data + 2 * sizeof(int), sizeof(int)); + } + int sum = DSO1(x) + DSO2(y) + (z ? DSOTestExtra(z) : 0); + if (sum == 3) { + fprintf(stderr, "BINGO %d %d %d\n", x, y, z); + *nil = 0; + } + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/DivTest.cpp b/test/thirdparty/Fuzzer/test/DivTest.cpp new file mode 100644 index 00000000..63f6960f --- /dev/null +++ b/test/thirdparty/Fuzzer/test/DivTest.cpp @@ -0,0 +1,20 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer: find the interesting argument for div. +#include +#include +#include +#include +#include + +static volatile int Sink; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size < 4) return 0; + int a; + memcpy(&a, Data, 4); + Sink = 12345678 / (987654 - a); + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/EmptyTest.cpp b/test/thirdparty/Fuzzer/test/EmptyTest.cpp new file mode 100644 index 00000000..5e843308 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/EmptyTest.cpp @@ -0,0 +1,11 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +// A fuzzer with empty target function. + +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/FourIndependentBranchesTest.cpp b/test/thirdparty/Fuzzer/test/FourIndependentBranchesTest.cpp new file mode 100644 index 00000000..62b3be76 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/FourIndependentBranchesTest.cpp @@ -0,0 +1,22 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find the string "FUZZ". +#include +#include +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + int bits = 0; + if (Size > 0 && Data[0] == 'F') bits |= 1; + if (Size > 1 && Data[1] == 'U') bits |= 2; + if (Size > 2 && Data[2] == 'Z') bits |= 4; + if (Size > 3 && Data[3] == 'Z') bits |= 8; + if (bits == 15) { + std::cerr << "BINGO!\n"; + exit(1); + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/FullCoverageSetTest.cpp b/test/thirdparty/Fuzzer/test/FullCoverageSetTest.cpp new file mode 100644 index 00000000..415e0b47 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/FullCoverageSetTest.cpp @@ -0,0 +1,24 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find the string "FUZZER". +#include +#include +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + int bits = 0; + if (Size > 0 && Data[0] == 'F') bits |= 1; + if (Size > 1 && Data[1] == 'U') bits |= 2; + if (Size > 2 && Data[2] == 'Z') bits |= 4; + if (Size > 3 && Data[3] == 'Z') bits |= 8; + if (Size > 4 && Data[4] == 'E') bits |= 16; + if (Size > 5 && Data[5] == 'R') bits |= 32; + if (bits == 63) { + std::cerr << "BINGO!\n"; + exit(1); + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/FuzzerUnittest.cpp b/test/thirdparty/Fuzzer/test/FuzzerUnittest.cpp new file mode 100644 index 00000000..4992ef57 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/FuzzerUnittest.cpp @@ -0,0 +1,738 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Avoid ODR violations (LibFuzzer is built without ASan and this test is built +// with ASan) involving C++ standard library types when using libcxx. +#define _LIBCPP_HAS_NO_ASAN + +#include "FuzzerCorpus.h" +#include "FuzzerInternal.h" +#include "FuzzerDictionary.h" +#include "FuzzerMerge.h" +#include "FuzzerMutate.h" +#include "FuzzerRandom.h" +#include "gtest/gtest.h" +#include +#include + +using namespace fuzzer; + +// For now, have LLVMFuzzerTestOneInput just to make it link. +// Later we may want to make unittests that actually call LLVMFuzzerTestOneInput. +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + abort(); +} + +TEST(Fuzzer, CrossOver) { + std::unique_ptr t(new ExternalFunctions()); + fuzzer::EF = t.get(); + Random Rand(0); + MutationDispatcher MD(Rand, {}); + Unit A({0, 1, 2}), B({5, 6, 7}); + Unit C; + Unit Expected[] = { + { 0 }, + { 0, 1 }, + { 0, 5 }, + { 0, 1, 2 }, + { 0, 1, 5 }, + { 0, 5, 1 }, + { 0, 5, 6 }, + { 0, 1, 2, 5 }, + { 0, 1, 5, 2 }, + { 0, 1, 5, 6 }, + { 0, 5, 1, 2 }, + { 0, 5, 1, 6 }, + { 0, 5, 6, 1 }, + { 0, 5, 6, 7 }, + { 0, 1, 2, 5, 6 }, + { 0, 1, 5, 2, 6 }, + { 0, 1, 5, 6, 2 }, + { 0, 1, 5, 6, 7 }, + { 0, 5, 1, 2, 6 }, + { 0, 5, 1, 6, 2 }, + { 0, 5, 1, 6, 7 }, + { 0, 5, 6, 1, 2 }, + { 0, 5, 6, 1, 7 }, + { 0, 5, 6, 7, 1 }, + { 0, 1, 2, 5, 6, 7 }, + { 0, 1, 5, 2, 6, 7 }, + { 0, 1, 5, 6, 2, 7 }, + { 0, 1, 5, 6, 7, 2 }, + { 0, 5, 1, 2, 6, 7 }, + { 0, 5, 1, 6, 2, 7 }, + { 0, 5, 1, 6, 7, 2 }, + { 0, 5, 6, 1, 2, 7 }, + { 0, 5, 6, 1, 7, 2 }, + { 0, 5, 6, 7, 1, 2 } + }; + for (size_t Len = 1; Len < 8; Len++) { + std::set FoundUnits, ExpectedUnitsWitThisLength; + for (int Iter = 0; Iter < 3000; Iter++) { + C.resize(Len); + size_t NewSize = MD.CrossOver(A.data(), A.size(), B.data(), B.size(), + C.data(), C.size()); + C.resize(NewSize); + FoundUnits.insert(C); + } + for (const Unit &U : Expected) + if (U.size() <= Len) + ExpectedUnitsWitThisLength.insert(U); + EXPECT_EQ(ExpectedUnitsWitThisLength, FoundUnits); + } +} + +TEST(Fuzzer, Hash) { + uint8_t A[] = {'a', 'b', 'c'}; + fuzzer::Unit U(A, A + sizeof(A)); + EXPECT_EQ("a9993e364706816aba3e25717850c26c9cd0d89d", fuzzer::Hash(U)); + U.push_back('d'); + EXPECT_EQ("81fe8bfe87576c3ecb22426f8e57847382917acf", fuzzer::Hash(U)); +} + +typedef size_t (MutationDispatcher::*Mutator)(uint8_t *Data, size_t Size, + size_t MaxSize); + +void TestEraseBytes(Mutator M, int NumIter) { + std::unique_ptr t(new ExternalFunctions()); + fuzzer::EF = t.get(); + uint8_t REM0[8] = {0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77}; + uint8_t REM1[8] = {0x00, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77}; + uint8_t REM2[8] = {0x00, 0x11, 0x33, 0x44, 0x55, 0x66, 0x77}; + uint8_t REM3[8] = {0x00, 0x11, 0x22, 0x44, 0x55, 0x66, 0x77}; + uint8_t REM4[8] = {0x00, 0x11, 0x22, 0x33, 0x55, 0x66, 0x77}; + uint8_t REM5[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x66, 0x77}; + uint8_t REM6[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x77}; + uint8_t REM7[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66}; + + uint8_t REM8[6] = {0x22, 0x33, 0x44, 0x55, 0x66, 0x77}; + uint8_t REM9[6] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55}; + uint8_t REM10[6] = {0x00, 0x11, 0x22, 0x55, 0x66, 0x77}; + + uint8_t REM11[5] = {0x33, 0x44, 0x55, 0x66, 0x77}; + uint8_t REM12[5] = {0x00, 0x11, 0x22, 0x33, 0x44}; + uint8_t REM13[5] = {0x00, 0x44, 0x55, 0x66, 0x77}; + + + Random Rand(0); + MutationDispatcher MD(Rand, {}); + int FoundMask = 0; + for (int i = 0; i < NumIter; i++) { + uint8_t T[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77}; + size_t NewSize = (MD.*M)(T, sizeof(T), sizeof(T)); + if (NewSize == 7 && !memcmp(REM0, T, 7)) FoundMask |= 1 << 0; + if (NewSize == 7 && !memcmp(REM1, T, 7)) FoundMask |= 1 << 1; + if (NewSize == 7 && !memcmp(REM2, T, 7)) FoundMask |= 1 << 2; + if (NewSize == 7 && !memcmp(REM3, T, 7)) FoundMask |= 1 << 3; + if (NewSize == 7 && !memcmp(REM4, T, 7)) FoundMask |= 1 << 4; + if (NewSize == 7 && !memcmp(REM5, T, 7)) FoundMask |= 1 << 5; + if (NewSize == 7 && !memcmp(REM6, T, 7)) FoundMask |= 1 << 6; + if (NewSize == 7 && !memcmp(REM7, T, 7)) FoundMask |= 1 << 7; + + if (NewSize == 6 && !memcmp(REM8, T, 6)) FoundMask |= 1 << 8; + if (NewSize == 6 && !memcmp(REM9, T, 6)) FoundMask |= 1 << 9; + if (NewSize == 6 && !memcmp(REM10, T, 6)) FoundMask |= 1 << 10; + + if (NewSize == 5 && !memcmp(REM11, T, 5)) FoundMask |= 1 << 11; + if (NewSize == 5 && !memcmp(REM12, T, 5)) FoundMask |= 1 << 12; + if (NewSize == 5 && !memcmp(REM13, T, 5)) FoundMask |= 1 << 13; + } + EXPECT_EQ(FoundMask, (1 << 14) - 1); +} + +TEST(FuzzerMutate, EraseBytes1) { + TestEraseBytes(&MutationDispatcher::Mutate_EraseBytes, 200); +} +TEST(FuzzerMutate, EraseBytes2) { + TestEraseBytes(&MutationDispatcher::Mutate, 2000); +} + +void TestInsertByte(Mutator M, int NumIter) { + std::unique_ptr t(new ExternalFunctions()); + fuzzer::EF = t.get(); + Random Rand(0); + MutationDispatcher MD(Rand, {}); + int FoundMask = 0; + uint8_t INS0[8] = {0xF1, 0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66}; + uint8_t INS1[8] = {0x00, 0xF2, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66}; + uint8_t INS2[8] = {0x00, 0x11, 0xF3, 0x22, 0x33, 0x44, 0x55, 0x66}; + uint8_t INS3[8] = {0x00, 0x11, 0x22, 0xF4, 0x33, 0x44, 0x55, 0x66}; + uint8_t INS4[8] = {0x00, 0x11, 0x22, 0x33, 0xF5, 0x44, 0x55, 0x66}; + uint8_t INS5[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0xF6, 0x55, 0x66}; + uint8_t INS6[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0xF7, 0x66}; + uint8_t INS7[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0xF8}; + for (int i = 0; i < NumIter; i++) { + uint8_t T[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66}; + size_t NewSize = (MD.*M)(T, 7, 8); + if (NewSize == 8 && !memcmp(INS0, T, 8)) FoundMask |= 1 << 0; + if (NewSize == 8 && !memcmp(INS1, T, 8)) FoundMask |= 1 << 1; + if (NewSize == 8 && !memcmp(INS2, T, 8)) FoundMask |= 1 << 2; + if (NewSize == 8 && !memcmp(INS3, T, 8)) FoundMask |= 1 << 3; + if (NewSize == 8 && !memcmp(INS4, T, 8)) FoundMask |= 1 << 4; + if (NewSize == 8 && !memcmp(INS5, T, 8)) FoundMask |= 1 << 5; + if (NewSize == 8 && !memcmp(INS6, T, 8)) FoundMask |= 1 << 6; + if (NewSize == 8 && !memcmp(INS7, T, 8)) FoundMask |= 1 << 7; + } + EXPECT_EQ(FoundMask, 255); +} + +TEST(FuzzerMutate, InsertByte1) { + TestInsertByte(&MutationDispatcher::Mutate_InsertByte, 1 << 15); +} +TEST(FuzzerMutate, InsertByte2) { + TestInsertByte(&MutationDispatcher::Mutate, 1 << 17); +} + +void TestInsertRepeatedBytes(Mutator M, int NumIter) { + std::unique_ptr t(new ExternalFunctions()); + fuzzer::EF = t.get(); + Random Rand(0); + MutationDispatcher MD(Rand, {}); + int FoundMask = 0; + uint8_t INS0[7] = {0x00, 0x11, 0x22, 0x33, 'a', 'a', 'a'}; + uint8_t INS1[7] = {0x00, 0x11, 0x22, 'a', 'a', 'a', 0x33}; + uint8_t INS2[7] = {0x00, 0x11, 'a', 'a', 'a', 0x22, 0x33}; + uint8_t INS3[7] = {0x00, 'a', 'a', 'a', 0x11, 0x22, 0x33}; + uint8_t INS4[7] = {'a', 'a', 'a', 0x00, 0x11, 0x22, 0x33}; + + uint8_t INS5[8] = {0x00, 0x11, 0x22, 0x33, 'b', 'b', 'b', 'b'}; + uint8_t INS6[8] = {0x00, 0x11, 0x22, 'b', 'b', 'b', 'b', 0x33}; + uint8_t INS7[8] = {0x00, 0x11, 'b', 'b', 'b', 'b', 0x22, 0x33}; + uint8_t INS8[8] = {0x00, 'b', 'b', 'b', 'b', 0x11, 0x22, 0x33}; + uint8_t INS9[8] = {'b', 'b', 'b', 'b', 0x00, 0x11, 0x22, 0x33}; + + for (int i = 0; i < NumIter; i++) { + uint8_t T[8] = {0x00, 0x11, 0x22, 0x33}; + size_t NewSize = (MD.*M)(T, 4, 8); + if (NewSize == 7 && !memcmp(INS0, T, 7)) FoundMask |= 1 << 0; + if (NewSize == 7 && !memcmp(INS1, T, 7)) FoundMask |= 1 << 1; + if (NewSize == 7 && !memcmp(INS2, T, 7)) FoundMask |= 1 << 2; + if (NewSize == 7 && !memcmp(INS3, T, 7)) FoundMask |= 1 << 3; + if (NewSize == 7 && !memcmp(INS4, T, 7)) FoundMask |= 1 << 4; + + if (NewSize == 8 && !memcmp(INS5, T, 8)) FoundMask |= 1 << 5; + if (NewSize == 8 && !memcmp(INS6, T, 8)) FoundMask |= 1 << 6; + if (NewSize == 8 && !memcmp(INS7, T, 8)) FoundMask |= 1 << 7; + if (NewSize == 8 && !memcmp(INS8, T, 8)) FoundMask |= 1 << 8; + if (NewSize == 8 && !memcmp(INS9, T, 8)) FoundMask |= 1 << 9; + + } + EXPECT_EQ(FoundMask, (1 << 10) - 1); +} + +TEST(FuzzerMutate, InsertRepeatedBytes1) { + TestInsertRepeatedBytes(&MutationDispatcher::Mutate_InsertRepeatedBytes, 10000); +} +TEST(FuzzerMutate, InsertRepeatedBytes2) { + TestInsertRepeatedBytes(&MutationDispatcher::Mutate, 300000); +} + +void TestChangeByte(Mutator M, int NumIter) { + std::unique_ptr t(new ExternalFunctions()); + fuzzer::EF = t.get(); + Random Rand(0); + MutationDispatcher MD(Rand, {}); + int FoundMask = 0; + uint8_t CH0[8] = {0xF0, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77}; + uint8_t CH1[8] = {0x00, 0xF1, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77}; + uint8_t CH2[8] = {0x00, 0x11, 0xF2, 0x33, 0x44, 0x55, 0x66, 0x77}; + uint8_t CH3[8] = {0x00, 0x11, 0x22, 0xF3, 0x44, 0x55, 0x66, 0x77}; + uint8_t CH4[8] = {0x00, 0x11, 0x22, 0x33, 0xF4, 0x55, 0x66, 0x77}; + uint8_t CH5[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0xF5, 0x66, 0x77}; + uint8_t CH6[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0xF5, 0x77}; + uint8_t CH7[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0xF7}; + for (int i = 0; i < NumIter; i++) { + uint8_t T[9] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77}; + size_t NewSize = (MD.*M)(T, 8, 9); + if (NewSize == 8 && !memcmp(CH0, T, 8)) FoundMask |= 1 << 0; + if (NewSize == 8 && !memcmp(CH1, T, 8)) FoundMask |= 1 << 1; + if (NewSize == 8 && !memcmp(CH2, T, 8)) FoundMask |= 1 << 2; + if (NewSize == 8 && !memcmp(CH3, T, 8)) FoundMask |= 1 << 3; + if (NewSize == 8 && !memcmp(CH4, T, 8)) FoundMask |= 1 << 4; + if (NewSize == 8 && !memcmp(CH5, T, 8)) FoundMask |= 1 << 5; + if (NewSize == 8 && !memcmp(CH6, T, 8)) FoundMask |= 1 << 6; + if (NewSize == 8 && !memcmp(CH7, T, 8)) FoundMask |= 1 << 7; + } + EXPECT_EQ(FoundMask, 255); +} + +TEST(FuzzerMutate, ChangeByte1) { + TestChangeByte(&MutationDispatcher::Mutate_ChangeByte, 1 << 15); +} +TEST(FuzzerMutate, ChangeByte2) { + TestChangeByte(&MutationDispatcher::Mutate, 1 << 17); +} + +void TestChangeBit(Mutator M, int NumIter) { + std::unique_ptr t(new ExternalFunctions()); + fuzzer::EF = t.get(); + Random Rand(0); + MutationDispatcher MD(Rand, {}); + int FoundMask = 0; + uint8_t CH0[8] = {0x01, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77}; + uint8_t CH1[8] = {0x00, 0x13, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77}; + uint8_t CH2[8] = {0x00, 0x11, 0x02, 0x33, 0x44, 0x55, 0x66, 0x77}; + uint8_t CH3[8] = {0x00, 0x11, 0x22, 0x37, 0x44, 0x55, 0x66, 0x77}; + uint8_t CH4[8] = {0x00, 0x11, 0x22, 0x33, 0x54, 0x55, 0x66, 0x77}; + uint8_t CH5[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x54, 0x66, 0x77}; + uint8_t CH6[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x76, 0x77}; + uint8_t CH7[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0xF7}; + for (int i = 0; i < NumIter; i++) { + uint8_t T[9] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77}; + size_t NewSize = (MD.*M)(T, 8, 9); + if (NewSize == 8 && !memcmp(CH0, T, 8)) FoundMask |= 1 << 0; + if (NewSize == 8 && !memcmp(CH1, T, 8)) FoundMask |= 1 << 1; + if (NewSize == 8 && !memcmp(CH2, T, 8)) FoundMask |= 1 << 2; + if (NewSize == 8 && !memcmp(CH3, T, 8)) FoundMask |= 1 << 3; + if (NewSize == 8 && !memcmp(CH4, T, 8)) FoundMask |= 1 << 4; + if (NewSize == 8 && !memcmp(CH5, T, 8)) FoundMask |= 1 << 5; + if (NewSize == 8 && !memcmp(CH6, T, 8)) FoundMask |= 1 << 6; + if (NewSize == 8 && !memcmp(CH7, T, 8)) FoundMask |= 1 << 7; + } + EXPECT_EQ(FoundMask, 255); +} + +TEST(FuzzerMutate, ChangeBit1) { + TestChangeBit(&MutationDispatcher::Mutate_ChangeBit, 1 << 16); +} +TEST(FuzzerMutate, ChangeBit2) { + TestChangeBit(&MutationDispatcher::Mutate, 1 << 18); +} + +void TestShuffleBytes(Mutator M, int NumIter) { + std::unique_ptr t(new ExternalFunctions()); + fuzzer::EF = t.get(); + Random Rand(0); + MutationDispatcher MD(Rand, {}); + int FoundMask = 0; + uint8_t CH0[7] = {0x00, 0x22, 0x11, 0x33, 0x44, 0x55, 0x66}; + uint8_t CH1[7] = {0x11, 0x00, 0x33, 0x22, 0x44, 0x55, 0x66}; + uint8_t CH2[7] = {0x00, 0x33, 0x11, 0x22, 0x44, 0x55, 0x66}; + uint8_t CH3[7] = {0x00, 0x11, 0x22, 0x44, 0x55, 0x66, 0x33}; + uint8_t CH4[7] = {0x00, 0x11, 0x22, 0x33, 0x55, 0x44, 0x66}; + for (int i = 0; i < NumIter; i++) { + uint8_t T[7] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66}; + size_t NewSize = (MD.*M)(T, 7, 7); + if (NewSize == 7 && !memcmp(CH0, T, 7)) FoundMask |= 1 << 0; + if (NewSize == 7 && !memcmp(CH1, T, 7)) FoundMask |= 1 << 1; + if (NewSize == 7 && !memcmp(CH2, T, 7)) FoundMask |= 1 << 2; + if (NewSize == 7 && !memcmp(CH3, T, 7)) FoundMask |= 1 << 3; + if (NewSize == 7 && !memcmp(CH4, T, 7)) FoundMask |= 1 << 4; + } + EXPECT_EQ(FoundMask, 31); +} + +TEST(FuzzerMutate, ShuffleBytes1) { + TestShuffleBytes(&MutationDispatcher::Mutate_ShuffleBytes, 1 << 16); +} +TEST(FuzzerMutate, ShuffleBytes2) { + TestShuffleBytes(&MutationDispatcher::Mutate, 1 << 20); +} + +void TestCopyPart(Mutator M, int NumIter) { + std::unique_ptr t(new ExternalFunctions()); + fuzzer::EF = t.get(); + Random Rand(0); + MutationDispatcher MD(Rand, {}); + int FoundMask = 0; + uint8_t CH0[7] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x00, 0x11}; + uint8_t CH1[7] = {0x55, 0x66, 0x22, 0x33, 0x44, 0x55, 0x66}; + uint8_t CH2[7] = {0x00, 0x55, 0x66, 0x33, 0x44, 0x55, 0x66}; + uint8_t CH3[7] = {0x00, 0x11, 0x22, 0x00, 0x11, 0x22, 0x66}; + uint8_t CH4[7] = {0x00, 0x11, 0x11, 0x22, 0x33, 0x55, 0x66}; + + for (int i = 0; i < NumIter; i++) { + uint8_t T[7] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66}; + size_t NewSize = (MD.*M)(T, 7, 7); + if (NewSize == 7 && !memcmp(CH0, T, 7)) FoundMask |= 1 << 0; + if (NewSize == 7 && !memcmp(CH1, T, 7)) FoundMask |= 1 << 1; + if (NewSize == 7 && !memcmp(CH2, T, 7)) FoundMask |= 1 << 2; + if (NewSize == 7 && !memcmp(CH3, T, 7)) FoundMask |= 1 << 3; + if (NewSize == 7 && !memcmp(CH4, T, 7)) FoundMask |= 1 << 4; + } + + uint8_t CH5[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x00, 0x11, 0x22}; + uint8_t CH6[8] = {0x22, 0x33, 0x44, 0x00, 0x11, 0x22, 0x33, 0x44}; + uint8_t CH7[8] = {0x00, 0x11, 0x22, 0x00, 0x11, 0x22, 0x33, 0x44}; + uint8_t CH8[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x22, 0x33, 0x44}; + uint8_t CH9[8] = {0x00, 0x11, 0x22, 0x22, 0x33, 0x44, 0x33, 0x44}; + + for (int i = 0; i < NumIter; i++) { + uint8_t T[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77}; + size_t NewSize = (MD.*M)(T, 5, 8); + if (NewSize == 8 && !memcmp(CH5, T, 8)) FoundMask |= 1 << 5; + if (NewSize == 8 && !memcmp(CH6, T, 8)) FoundMask |= 1 << 6; + if (NewSize == 8 && !memcmp(CH7, T, 8)) FoundMask |= 1 << 7; + if (NewSize == 8 && !memcmp(CH8, T, 8)) FoundMask |= 1 << 8; + if (NewSize == 8 && !memcmp(CH9, T, 8)) FoundMask |= 1 << 9; + } + + EXPECT_EQ(FoundMask, 1023); +} + +TEST(FuzzerMutate, CopyPart1) { + TestCopyPart(&MutationDispatcher::Mutate_CopyPart, 1 << 10); +} +TEST(FuzzerMutate, CopyPart2) { + TestCopyPart(&MutationDispatcher::Mutate, 1 << 13); +} + +void TestAddWordFromDictionary(Mutator M, int NumIter) { + std::unique_ptr t(new ExternalFunctions()); + fuzzer::EF = t.get(); + Random Rand(0); + MutationDispatcher MD(Rand, {}); + uint8_t Word1[4] = {0xAA, 0xBB, 0xCC, 0xDD}; + uint8_t Word2[3] = {0xFF, 0xEE, 0xEF}; + MD.AddWordToManualDictionary(Word(Word1, sizeof(Word1))); + MD.AddWordToManualDictionary(Word(Word2, sizeof(Word2))); + int FoundMask = 0; + uint8_t CH0[7] = {0x00, 0x11, 0x22, 0xAA, 0xBB, 0xCC, 0xDD}; + uint8_t CH1[7] = {0x00, 0x11, 0xAA, 0xBB, 0xCC, 0xDD, 0x22}; + uint8_t CH2[7] = {0x00, 0xAA, 0xBB, 0xCC, 0xDD, 0x11, 0x22}; + uint8_t CH3[7] = {0xAA, 0xBB, 0xCC, 0xDD, 0x00, 0x11, 0x22}; + uint8_t CH4[6] = {0x00, 0x11, 0x22, 0xFF, 0xEE, 0xEF}; + uint8_t CH5[6] = {0x00, 0x11, 0xFF, 0xEE, 0xEF, 0x22}; + uint8_t CH6[6] = {0x00, 0xFF, 0xEE, 0xEF, 0x11, 0x22}; + uint8_t CH7[6] = {0xFF, 0xEE, 0xEF, 0x00, 0x11, 0x22}; + for (int i = 0; i < NumIter; i++) { + uint8_t T[7] = {0x00, 0x11, 0x22}; + size_t NewSize = (MD.*M)(T, 3, 7); + if (NewSize == 7 && !memcmp(CH0, T, 7)) FoundMask |= 1 << 0; + if (NewSize == 7 && !memcmp(CH1, T, 7)) FoundMask |= 1 << 1; + if (NewSize == 7 && !memcmp(CH2, T, 7)) FoundMask |= 1 << 2; + if (NewSize == 7 && !memcmp(CH3, T, 7)) FoundMask |= 1 << 3; + if (NewSize == 6 && !memcmp(CH4, T, 6)) FoundMask |= 1 << 4; + if (NewSize == 6 && !memcmp(CH5, T, 6)) FoundMask |= 1 << 5; + if (NewSize == 6 && !memcmp(CH6, T, 6)) FoundMask |= 1 << 6; + if (NewSize == 6 && !memcmp(CH7, T, 6)) FoundMask |= 1 << 7; + } + EXPECT_EQ(FoundMask, 255); +} + +TEST(FuzzerMutate, AddWordFromDictionary1) { + TestAddWordFromDictionary( + &MutationDispatcher::Mutate_AddWordFromManualDictionary, 1 << 15); +} + +TEST(FuzzerMutate, AddWordFromDictionary2) { + TestAddWordFromDictionary(&MutationDispatcher::Mutate, 1 << 15); +} + +void TestAddWordFromDictionaryWithHint(Mutator M, int NumIter) { + std::unique_ptr t(new ExternalFunctions()); + fuzzer::EF = t.get(); + Random Rand(0); + MutationDispatcher MD(Rand, {}); + uint8_t W[] = {0xAA, 0xBB, 0xCC, 0xDD, 0xFF, 0xEE, 0xEF}; + size_t PosHint = 7777; + MD.AddWordToAutoDictionary({Word(W, sizeof(W)), PosHint}); + int FoundMask = 0; + for (int i = 0; i < NumIter; i++) { + uint8_t T[10000]; + memset(T, 0, sizeof(T)); + size_t NewSize = (MD.*M)(T, 9000, 10000); + if (NewSize >= PosHint + sizeof(W) && + !memcmp(W, T + PosHint, sizeof(W))) + FoundMask = 1; + } + EXPECT_EQ(FoundMask, 1); +} + +TEST(FuzzerMutate, AddWordFromDictionaryWithHint1) { + TestAddWordFromDictionaryWithHint( + &MutationDispatcher::Mutate_AddWordFromTemporaryAutoDictionary, 1 << 5); +} + +TEST(FuzzerMutate, AddWordFromDictionaryWithHint2) { + TestAddWordFromDictionaryWithHint(&MutationDispatcher::Mutate, 1 << 10); +} + +void TestChangeASCIIInteger(Mutator M, int NumIter) { + std::unique_ptr t(new ExternalFunctions()); + fuzzer::EF = t.get(); + Random Rand(0); + MutationDispatcher MD(Rand, {}); + + uint8_t CH0[8] = {'1', '2', '3', '4', '5', '6', '7', '7'}; + uint8_t CH1[8] = {'1', '2', '3', '4', '5', '6', '7', '9'}; + uint8_t CH2[8] = {'2', '4', '6', '9', '1', '3', '5', '6'}; + uint8_t CH3[8] = {'0', '6', '1', '7', '2', '8', '3', '9'}; + int FoundMask = 0; + for (int i = 0; i < NumIter; i++) { + uint8_t T[8] = {'1', '2', '3', '4', '5', '6', '7', '8'}; + size_t NewSize = (MD.*M)(T, 8, 8); + /**/ if (NewSize == 8 && !memcmp(CH0, T, 8)) FoundMask |= 1 << 0; + else if (NewSize == 8 && !memcmp(CH1, T, 8)) FoundMask |= 1 << 1; + else if (NewSize == 8 && !memcmp(CH2, T, 8)) FoundMask |= 1 << 2; + else if (NewSize == 8 && !memcmp(CH3, T, 8)) FoundMask |= 1 << 3; + else if (NewSize == 8) FoundMask |= 1 << 4; + } + EXPECT_EQ(FoundMask, 31); +} + +TEST(FuzzerMutate, ChangeASCIIInteger1) { + TestChangeASCIIInteger(&MutationDispatcher::Mutate_ChangeASCIIInteger, + 1 << 15); +} + +TEST(FuzzerMutate, ChangeASCIIInteger2) { + TestChangeASCIIInteger(&MutationDispatcher::Mutate, 1 << 15); +} + +void TestChangeBinaryInteger(Mutator M, int NumIter) { + std::unique_ptr t(new ExternalFunctions()); + fuzzer::EF = t.get(); + Random Rand(0); + MutationDispatcher MD(Rand, {}); + + uint8_t CH0[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x79}; + uint8_t CH1[8] = {0x00, 0x11, 0x22, 0x31, 0x44, 0x55, 0x66, 0x77}; + uint8_t CH2[8] = {0xff, 0x10, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77}; + uint8_t CH3[8] = {0x00, 0x11, 0x2a, 0x33, 0x44, 0x55, 0x66, 0x77}; + uint8_t CH4[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x4f, 0x66, 0x77}; + uint8_t CH5[8] = {0xff, 0xee, 0xdd, 0xcc, 0xbb, 0xaa, 0x99, 0x88}; + uint8_t CH6[8] = {0x00, 0x11, 0x22, 0x00, 0x00, 0x00, 0x08, 0x77}; // Size + uint8_t CH7[8] = {0x00, 0x08, 0x00, 0x33, 0x44, 0x55, 0x66, 0x77}; // Sw(Size) + + int FoundMask = 0; + for (int i = 0; i < NumIter; i++) { + uint8_t T[8] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77}; + size_t NewSize = (MD.*M)(T, 8, 8); + /**/ if (NewSize == 8 && !memcmp(CH0, T, 8)) FoundMask |= 1 << 0; + else if (NewSize == 8 && !memcmp(CH1, T, 8)) FoundMask |= 1 << 1; + else if (NewSize == 8 && !memcmp(CH2, T, 8)) FoundMask |= 1 << 2; + else if (NewSize == 8 && !memcmp(CH3, T, 8)) FoundMask |= 1 << 3; + else if (NewSize == 8 && !memcmp(CH4, T, 8)) FoundMask |= 1 << 4; + else if (NewSize == 8 && !memcmp(CH5, T, 8)) FoundMask |= 1 << 5; + else if (NewSize == 8 && !memcmp(CH6, T, 8)) FoundMask |= 1 << 6; + else if (NewSize == 8 && !memcmp(CH7, T, 8)) FoundMask |= 1 << 7; + } + EXPECT_EQ(FoundMask, 255); +} + +TEST(FuzzerMutate, ChangeBinaryInteger1) { + TestChangeBinaryInteger(&MutationDispatcher::Mutate_ChangeBinaryInteger, + 1 << 12); +} + +TEST(FuzzerMutate, ChangeBinaryInteger2) { + TestChangeBinaryInteger(&MutationDispatcher::Mutate, 1 << 15); +} + + +TEST(FuzzerDictionary, ParseOneDictionaryEntry) { + Unit U; + EXPECT_FALSE(ParseOneDictionaryEntry("", &U)); + EXPECT_FALSE(ParseOneDictionaryEntry(" ", &U)); + EXPECT_FALSE(ParseOneDictionaryEntry("\t ", &U)); + EXPECT_FALSE(ParseOneDictionaryEntry(" \" ", &U)); + EXPECT_FALSE(ParseOneDictionaryEntry(" zz\" ", &U)); + EXPECT_FALSE(ParseOneDictionaryEntry(" \"zz ", &U)); + EXPECT_FALSE(ParseOneDictionaryEntry(" \"\" ", &U)); + EXPECT_TRUE(ParseOneDictionaryEntry("\"a\"", &U)); + EXPECT_EQ(U, Unit({'a'})); + EXPECT_TRUE(ParseOneDictionaryEntry("\"abc\"", &U)); + EXPECT_EQ(U, Unit({'a', 'b', 'c'})); + EXPECT_TRUE(ParseOneDictionaryEntry("abc=\"abc\"", &U)); + EXPECT_EQ(U, Unit({'a', 'b', 'c'})); + EXPECT_FALSE(ParseOneDictionaryEntry("\"\\\"", &U)); + EXPECT_TRUE(ParseOneDictionaryEntry("\"\\\\\"", &U)); + EXPECT_EQ(U, Unit({'\\'})); + EXPECT_TRUE(ParseOneDictionaryEntry("\"\\xAB\"", &U)); + EXPECT_EQ(U, Unit({0xAB})); + EXPECT_TRUE(ParseOneDictionaryEntry("\"\\xABz\\xDE\"", &U)); + EXPECT_EQ(U, Unit({0xAB, 'z', 0xDE})); + EXPECT_TRUE(ParseOneDictionaryEntry("\"#\"", &U)); + EXPECT_EQ(U, Unit({'#'})); + EXPECT_TRUE(ParseOneDictionaryEntry("\"\\\"\"", &U)); + EXPECT_EQ(U, Unit({'"'})); +} + +TEST(FuzzerDictionary, ParseDictionaryFile) { + std::vector Units; + EXPECT_FALSE(ParseDictionaryFile("zzz\n", &Units)); + EXPECT_FALSE(ParseDictionaryFile("", &Units)); + EXPECT_TRUE(ParseDictionaryFile("\n", &Units)); + EXPECT_EQ(Units.size(), 0U); + EXPECT_TRUE(ParseDictionaryFile("#zzzz a b c d\n", &Units)); + EXPECT_EQ(Units.size(), 0U); + EXPECT_TRUE(ParseDictionaryFile(" #zzzz\n", &Units)); + EXPECT_EQ(Units.size(), 0U); + EXPECT_TRUE(ParseDictionaryFile(" #zzzz\n", &Units)); + EXPECT_EQ(Units.size(), 0U); + EXPECT_TRUE(ParseDictionaryFile(" #zzzz\naaa=\"aa\"", &Units)); + EXPECT_EQ(Units, std::vector({Unit({'a', 'a'})})); + EXPECT_TRUE( + ParseDictionaryFile(" #zzzz\naaa=\"aa\"\n\nabc=\"abc\"", &Units)); + EXPECT_EQ(Units, + std::vector({Unit({'a', 'a'}), Unit({'a', 'b', 'c'})})); +} + +TEST(FuzzerUtil, Base64) { + EXPECT_EQ("", Base64({})); + EXPECT_EQ("YQ==", Base64({'a'})); + EXPECT_EQ("eA==", Base64({'x'})); + EXPECT_EQ("YWI=", Base64({'a', 'b'})); + EXPECT_EQ("eHk=", Base64({'x', 'y'})); + EXPECT_EQ("YWJj", Base64({'a', 'b', 'c'})); + EXPECT_EQ("eHl6", Base64({'x', 'y', 'z'})); + EXPECT_EQ("YWJjeA==", Base64({'a', 'b', 'c', 'x'})); + EXPECT_EQ("YWJjeHk=", Base64({'a', 'b', 'c', 'x', 'y'})); + EXPECT_EQ("YWJjeHl6", Base64({'a', 'b', 'c', 'x', 'y', 'z'})); +} + +TEST(Corpus, Distribution) { + Random Rand(0); + InputCorpus C(""); + size_t N = 10; + size_t TriesPerUnit = 1<<16; + for (size_t i = 0; i < N; i++) + C.AddToCorpus(Unit{ static_cast(i) }, 0); + + std::vector Hist(N); + for (size_t i = 0; i < N * TriesPerUnit; i++) { + Hist[C.ChooseUnitIdxToMutate(Rand)]++; + } + for (size_t i = 0; i < N; i++) { + // A weak sanity check that every unit gets invoked. + EXPECT_GT(Hist[i], TriesPerUnit / N / 3); + } +} + +TEST(Merge, Bad) { + const char *kInvalidInputs[] = { + "", + "x", + "3\nx", + "2\n3", + "2\n2", + "2\n2\nA\n", + "2\n2\nA\nB\nC\n", + "0\n0\n", + "1\n1\nA\nDONE 0", + "1\n1\nA\nSTARTED 1", + }; + Merger M; + for (auto S : kInvalidInputs) { + // fprintf(stderr, "TESTING:\n%s\n", S); + EXPECT_FALSE(M.Parse(S, false)); + } +} + +void EQ(const std::vector &A, const std::vector &B) { + EXPECT_EQ(A, B); +} + +void EQ(const std::vector &A, const std::vector &B) { + std::set a(A.begin(), A.end()); + std::set b(B.begin(), B.end()); + EXPECT_EQ(a, b); +} + +static void Merge(const std::string &Input, + const std::vector Result, + size_t NumNewFeatures) { + Merger M; + std::vector NewFiles; + EXPECT_TRUE(M.Parse(Input, true)); + EXPECT_EQ(NumNewFeatures, M.Merge(&NewFiles)); + EQ(NewFiles, Result); +} + +TEST(Merge, Good) { + Merger M; + + EXPECT_TRUE(M.Parse("1\n0\nAA\n", false)); + EXPECT_EQ(M.Files.size(), 1U); + EXPECT_EQ(M.NumFilesInFirstCorpus, 0U); + EXPECT_EQ(M.Files[0].Name, "AA"); + EXPECT_TRUE(M.LastFailure.empty()); + EXPECT_EQ(M.FirstNotProcessedFile, 0U); + + EXPECT_TRUE(M.Parse("2\n1\nAA\nBB\nSTARTED 0 42\n", false)); + EXPECT_EQ(M.Files.size(), 2U); + EXPECT_EQ(M.NumFilesInFirstCorpus, 1U); + EXPECT_EQ(M.Files[0].Name, "AA"); + EXPECT_EQ(M.Files[1].Name, "BB"); + EXPECT_EQ(M.LastFailure, "AA"); + EXPECT_EQ(M.FirstNotProcessedFile, 1U); + + EXPECT_TRUE(M.Parse("3\n1\nAA\nBB\nC\n" + "STARTED 0 1000\n" + "DONE 0 1 2 3\n" + "STARTED 1 1001\n" + "DONE 1 4 5 6 \n" + "STARTED 2 1002\n" + "", true)); + EXPECT_EQ(M.Files.size(), 3U); + EXPECT_EQ(M.NumFilesInFirstCorpus, 1U); + EXPECT_EQ(M.Files[0].Name, "AA"); + EXPECT_EQ(M.Files[0].Size, 1000U); + EXPECT_EQ(M.Files[1].Name, "BB"); + EXPECT_EQ(M.Files[1].Size, 1001U); + EXPECT_EQ(M.Files[2].Name, "C"); + EXPECT_EQ(M.Files[2].Size, 1002U); + EXPECT_EQ(M.LastFailure, "C"); + EXPECT_EQ(M.FirstNotProcessedFile, 3U); + EQ(M.Files[0].Features, {1, 2, 3}); + EQ(M.Files[1].Features, {4, 5, 6}); + + + std::vector NewFiles; + + EXPECT_TRUE(M.Parse("3\n2\nAA\nBB\nC\n" + "STARTED 0 1000\nDONE 0 1 2 3\n" + "STARTED 1 1001\nDONE 1 4 5 6 \n" + "STARTED 2 1002\nDONE 2 6 1 3 \n" + "", true)); + EXPECT_EQ(M.Files.size(), 3U); + EXPECT_EQ(M.NumFilesInFirstCorpus, 2U); + EXPECT_TRUE(M.LastFailure.empty()); + EXPECT_EQ(M.FirstNotProcessedFile, 3U); + EQ(M.Files[0].Features, {1, 2, 3}); + EQ(M.Files[1].Features, {4, 5, 6}); + EQ(M.Files[2].Features, {1, 3, 6}); + EXPECT_EQ(0U, M.Merge(&NewFiles)); + EQ(NewFiles, {}); + + EXPECT_TRUE(M.Parse("3\n1\nA\nB\nC\n" + "STARTED 0 1000\nDONE 0 1 2 3\n" + "STARTED 1 1001\nDONE 1 4 5 6 \n" + "STARTED 2 1002\nDONE 2 6 1 3\n" + "", true)); + EQ(M.Files[0].Features, {1, 2, 3}); + EQ(M.Files[1].Features, {4, 5, 6}); + EQ(M.Files[2].Features, {1, 3, 6}); + EXPECT_EQ(3U, M.Merge(&NewFiles)); + EQ(NewFiles, {"B"}); +} + +TEST(Merge, Merge) { + + Merge("3\n1\nA\nB\nC\n" + "STARTED 0 1000\nDONE 0 1 2 3\n" + "STARTED 1 1001\nDONE 1 4 5 6 \n" + "STARTED 2 1002\nDONE 2 6 1 3 \n", + {"B"}, 3); + + Merge("3\n0\nA\nB\nC\n" + "STARTED 0 2000\nDONE 0 1 2 3\n" + "STARTED 1 1001\nDONE 1 4 5 6 \n" + "STARTED 2 1002\nDONE 2 6 1 3 \n", + {"A", "B", "C"}, 6); + + Merge("4\n0\nA\nB\nC\nD\n" + "STARTED 0 2000\nDONE 0 1 2 3\n" + "STARTED 1 1101\nDONE 1 4 5 6 \n" + "STARTED 2 1102\nDONE 2 6 1 3 100 \n" + "STARTED 3 1000\nDONE 3 1 \n", + {"A", "B", "C", "D"}, 7); + + Merge("4\n1\nA\nB\nC\nD\n" + "STARTED 0 2000\nDONE 0 4 5 6 7 8\n" + "STARTED 1 1100\nDONE 1 1 2 3 \n" + "STARTED 2 1100\nDONE 2 2 3 \n" + "STARTED 3 1000\nDONE 3 1 \n", + {"B", "D"}, 3); +} diff --git a/test/thirdparty/Fuzzer/test/InitializeTest.cpp b/test/thirdparty/Fuzzer/test/InitializeTest.cpp new file mode 100644 index 00000000..0d6a0fda --- /dev/null +++ b/test/thirdparty/Fuzzer/test/InitializeTest.cpp @@ -0,0 +1,28 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Make sure LLVMFuzzerInitialize is called. +#include +#include +#include +#include +#include +#include + +static char *argv0; + +extern "C" int LLVMFuzzerInitialize(int *argc, char ***argv) { + assert(*argc > 0); + argv0 = **argv; + fprintf(stderr, "LLVMFuzzerInitialize: %s\n", argv0); + return 0; +} + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size == strlen(argv0) && + !strncmp(reinterpret_cast(Data), argv0, Size)) { + fprintf(stderr, "BINGO %s\n", argv0); + exit(1); + } + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/LeakTest.cpp b/test/thirdparty/Fuzzer/test/LeakTest.cpp new file mode 100644 index 00000000..22e51640 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/LeakTest.cpp @@ -0,0 +1,17 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Test with a leak. +#include +#include + +static volatile void *Sink; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size > 0 && *Data == 'H') { + Sink = new int; + Sink = nullptr; + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/LeakTimeoutTest.cpp b/test/thirdparty/Fuzzer/test/LeakTimeoutTest.cpp new file mode 100644 index 00000000..4f31b3e5 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/LeakTimeoutTest.cpp @@ -0,0 +1,17 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Test with a leak. +#include +#include + +static volatile int *Sink; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (!Size) return 0; + Sink = new int; + Sink = new int; + while (Sink) *Sink = 0; // Infinite loop. + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/LoadTest.cpp b/test/thirdparty/Fuzzer/test/LoadTest.cpp new file mode 100644 index 00000000..c1780d5c --- /dev/null +++ b/test/thirdparty/Fuzzer/test/LoadTest.cpp @@ -0,0 +1,22 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer: find interesting value of array index. +#include +#include +#include +#include +#include + +static volatile int Sink; +const int kArraySize = 1234567; +int array[kArraySize]; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size < 8) return 0; + size_t a = 0; + memcpy(&a, Data, 8); + Sink = array[a % (kArraySize + 1)]; + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/MemcmpTest.cpp b/test/thirdparty/Fuzzer/test/MemcmpTest.cpp new file mode 100644 index 00000000..fdbf9468 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/MemcmpTest.cpp @@ -0,0 +1,31 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find a particular string. +#include +#include +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + // TODO: check other sizes. + if (Size >= 8 && memcmp(Data, "01234567", 8) == 0) { + if (Size >= 12 && memcmp(Data + 8, "ABCD", 4) == 0) { + if (Size >= 14 && memcmp(Data + 12, "XY", 2) == 0) { + if (Size >= 17 && memcmp(Data + 14, "KLM", 3) == 0) { + if (Size >= 27 && memcmp(Data + 17, "ABCDE-GHIJ", 10) == 0){ + fprintf(stderr, "BINGO %zd\n", Size); + for (size_t i = 0; i < Size; i++) { + uint8_t C = Data[i]; + if (C >= 32 && C < 127) + fprintf(stderr, "%c", C); + } + fprintf(stderr, "\n"); + exit(1); + } + } + } + } + } + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/NthRunCrashTest.cpp b/test/thirdparty/Fuzzer/test/NthRunCrashTest.cpp new file mode 100644 index 00000000..b43e69e5 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/NthRunCrashTest.cpp @@ -0,0 +1,18 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Crash on the N-th execution. +#include +#include +#include + +static int Counter; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Counter++ == 1000) { + std::cout << "BINGO; Found the target, exiting\n"; + exit(1); + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/NullDerefOnEmptyTest.cpp b/test/thirdparty/Fuzzer/test/NullDerefOnEmptyTest.cpp new file mode 100644 index 00000000..15371092 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/NullDerefOnEmptyTest.cpp @@ -0,0 +1,19 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find the empty string. +#include +#include +#include +#include + +static volatile int *Null = 0; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size == 0) { + std::cout << "Found the target, dereferencing NULL\n"; + *Null = 1; + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/NullDerefTest.cpp b/test/thirdparty/Fuzzer/test/NullDerefTest.cpp new file mode 100644 index 00000000..3f03d249 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/NullDerefTest.cpp @@ -0,0 +1,26 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find the string "Hi!". +#include +#include +#include +#include + +static volatile int Sink; +static volatile int *Null = 0; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size > 0 && Data[0] == 'H') { + Sink = 1; + if (Size > 1 && Data[1] == 'i') { + Sink = 2; + if (Size > 2 && Data[2] == '!') { + std::cout << "Found the target, dereferencing NULL\n"; + *Null = 1; + } + } + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/OneHugeAllocTest.cpp b/test/thirdparty/Fuzzer/test/OneHugeAllocTest.cpp new file mode 100644 index 00000000..8d3d1d6d --- /dev/null +++ b/test/thirdparty/Fuzzer/test/OneHugeAllocTest.cpp @@ -0,0 +1,28 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Tests OOM handling when there is a single large allocation. +#include +#include +#include +#include +#include +#include + +static volatile char *SinkPtr; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size > 0 && Data[0] == 'H') { + if (Size > 1 && Data[1] == 'i') { + if (Size > 2 && Data[2] == '!') { + size_t kSize = (size_t)1 << 31; + char *p = new char[kSize]; + memset(p, 0, kSize); + SinkPtr = p; + delete [] p; + } + } + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/OutOfMemorySingleLargeMallocTest.cpp b/test/thirdparty/Fuzzer/test/OutOfMemorySingleLargeMallocTest.cpp new file mode 100644 index 00000000..ea23a601 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/OutOfMemorySingleLargeMallocTest.cpp @@ -0,0 +1,27 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Tests OOM handling. +#include +#include +#include +#include +#include +#include + +static volatile char *SinkPtr; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size > 0 && Data[0] == 'H') { + if (Size > 1 && Data[1] == 'i') { + if (Size > 2 && Data[2] == '!') { + size_t kSize = 0xff000000U; + char *p = new char[kSize]; + SinkPtr = p; + delete [] p; + } + } + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/OutOfMemoryTest.cpp b/test/thirdparty/Fuzzer/test/OutOfMemoryTest.cpp new file mode 100644 index 00000000..078a39ee --- /dev/null +++ b/test/thirdparty/Fuzzer/test/OutOfMemoryTest.cpp @@ -0,0 +1,31 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Tests OOM handling. +#include +#include +#include +#include +#include +#include +#include + +static volatile char *SinkPtr; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size > 0 && Data[0] == 'H') { + if (Size > 1 && Data[1] == 'i') { + if (Size > 2 && Data[2] == '!') { + while (true) { + size_t kSize = 1 << 28; + char *p = new char[kSize]; + memset(p, 0, kSize); + SinkPtr = p; + std::this_thread::sleep_for(std::chrono::seconds(1)); + } + } + } + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/RepeatedBytesTest.cpp b/test/thirdparty/Fuzzer/test/RepeatedBytesTest.cpp new file mode 100644 index 00000000..2fa6c78c --- /dev/null +++ b/test/thirdparty/Fuzzer/test/RepeatedBytesTest.cpp @@ -0,0 +1,29 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find repeated bytes. +#include +#include +#include +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + assert(Data); + // Looking for AAAAAAAAAAAAAAAAAAAAAA or some such. + size_t CurA = 0, MaxA = 0; + for (size_t i = 0; i < Size; i++) { + // Make sure there are no conditionals in the loop so that + // coverage can't help the fuzzer. + int EQ = Data[i] == 'A'; + CurA = EQ * (CurA + 1); + int GT = CurA > MaxA; + MaxA = GT * CurA + (!GT) * MaxA; + } + if (MaxA >= 20) { + std::cout << "BINGO; Found the target (Max: " << MaxA << "), exiting\n"; + exit(0); + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/RepeatedMemcmp.cpp b/test/thirdparty/Fuzzer/test/RepeatedMemcmp.cpp new file mode 100644 index 00000000..a327bbee --- /dev/null +++ b/test/thirdparty/Fuzzer/test/RepeatedMemcmp.cpp @@ -0,0 +1,22 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + + +#include +#include +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + int Matches = 0; + for (size_t i = 0; i + 2 < Size; i += 3) { + const char *Pat = i % 2 ? "foo" : "bar"; + if (!memcmp(Data + i, Pat, 3)) + Matches++; + } + if (Matches > 20) { + fprintf(stderr, "BINGO!\n"); + exit(1); + } + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/ShrinkControlFlowTest.cpp b/test/thirdparty/Fuzzer/test/ShrinkControlFlowTest.cpp new file mode 100644 index 00000000..0fd7c5e9 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/ShrinkControlFlowTest.cpp @@ -0,0 +1,28 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Test that we can find the minimal item in the corpus (3 bytes: "FUZ"). +#include +#include +#include +#include +#include + +static volatile int Sink; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + int8_t Ids[256]; + memset(Ids, -1, sizeof(Ids)); + for (size_t i = 0; i < Size; i++) + if (Ids[Data[i]] == -1) + Ids[Data[i]] = i; + int F = Ids[(unsigned char)'F']; + int U = Ids[(unsigned char)'U']; + int Z = Ids[(unsigned char)'Z']; + if (F >= 0 && U > F && Z > U) { + Sink++; + //fprintf(stderr, "IDS: %d %d %d\n", F, U, Z); + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/ShrinkValueProfileTest.cpp b/test/thirdparty/Fuzzer/test/ShrinkValueProfileTest.cpp new file mode 100644 index 00000000..026b8ce2 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/ShrinkValueProfileTest.cpp @@ -0,0 +1,22 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Test that we can find the minimal item in the corpus (3 bytes: "FUZ"). +#include +#include +#include +#include +#include + +static volatile uint32_t Sink; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size < sizeof(uint32_t)) return 0; + uint32_t X, Y; + size_t Offset = Size < 8 ? 0 : Size / 2; + memcpy(&X, Data + Offset, sizeof(uint32_t)); + memcpy(&Y, "FUZZ", sizeof(uint32_t)); + Sink = X == Y; + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/SignedIntOverflowTest.cpp b/test/thirdparty/Fuzzer/test/SignedIntOverflowTest.cpp new file mode 100644 index 00000000..7df32ad5 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/SignedIntOverflowTest.cpp @@ -0,0 +1,28 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Test for signed-integer-overflow. +#include +#include +#include +#include +#include +#include + +static volatile int Sink; +static int Large = INT_MAX; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + assert(Data); + if (Size > 0 && Data[0] == 'H') { + Sink = 1; + if (Size > 1 && Data[1] == 'i') { + Sink = 2; + if (Size > 2 && Data[2] == '!') { + Large++; // int overflow. + } + } + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/SimpleCmpTest.cpp b/test/thirdparty/Fuzzer/test/SimpleCmpTest.cpp new file mode 100644 index 00000000..0220c30f --- /dev/null +++ b/test/thirdparty/Fuzzer/test/SimpleCmpTest.cpp @@ -0,0 +1,46 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find several narrow ranges. +#include +#include +#include +#include + +extern int AllLines[]; + +bool PrintOnce(int Line) { + if (!AllLines[Line]) + fprintf(stderr, "Seen line %d\n", Line); + AllLines[Line] = 1; + return true; +} + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size != 22) return 0; + uint64_t x = 0; + int64_t y = 0; + int32_t z = 0; + uint16_t a = 0; + memcpy(&x, Data, 8); // 8 + memcpy(&y, Data + 8, 8); // 16 + memcpy(&z, Data + 16, sizeof(z)); // 20 + memcpy(&a, Data + 20, sizeof(a)); // 22 + + if (x > 1234567890 && PrintOnce(__LINE__) && + x < 1234567895 && PrintOnce(__LINE__) && + a == 0x4242 && PrintOnce(__LINE__) && + y >= 987654321 && PrintOnce(__LINE__) && + y <= 987654325 && PrintOnce(__LINE__) && + z < -10000 && PrintOnce(__LINE__) && + z >= -10005 && PrintOnce(__LINE__) && + z != -10003 && PrintOnce(__LINE__) && + true) { + fprintf(stderr, "BINGO; Found the target: size %zd (%zd, %zd, %d, %d), exiting.\n", + Size, x, y, z, a); + exit(1); + } + return 0; +} + +int AllLines[__LINE__ + 1]; // Must be the last line. diff --git a/test/thirdparty/Fuzzer/test/SimpleDictionaryTest.cpp b/test/thirdparty/Fuzzer/test/SimpleDictionaryTest.cpp new file mode 100644 index 00000000..cd7292bd --- /dev/null +++ b/test/thirdparty/Fuzzer/test/SimpleDictionaryTest.cpp @@ -0,0 +1,29 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. +// The fuzzer must find a string based on dictionary words: +// "Elvis" +// "Presley" +#include +#include +#include +#include +#include + +static volatile int Zero = 0; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + const char *Expected = "ElvisPresley"; + if (Size < strlen(Expected)) return 0; + size_t Match = 0; + for (size_t i = 0; Expected[i]; i++) + if (Expected[i] + Zero == Data[i]) + Match++; + if (Match == strlen(Expected)) { + std::cout << "BINGO; Found the target, exiting\n"; + exit(1); + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/SimpleHashTest.cpp b/test/thirdparty/Fuzzer/test/SimpleHashTest.cpp new file mode 100644 index 00000000..00599de7 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/SimpleHashTest.cpp @@ -0,0 +1,40 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// This test computes a checksum of the data (all but the last 4 bytes), +// and then compares the last 4 bytes with the computed value. +// A fuzzer with cmp traces is expected to defeat this check. +#include +#include +#include +#include + +// A modified jenkins_one_at_a_time_hash initialized by non-zero, +// so that simple_hash(0) != 0. See also +// https://en.wikipedia.org/wiki/Jenkins_hash_function +static uint32_t simple_hash(const uint8_t *Data, size_t Size) { + uint32_t Hash = 0x12039854; + for (uint32_t i = 0; i < Size; i++) { + Hash += Data[i]; + Hash += (Hash << 10); + Hash ^= (Hash >> 6); + } + Hash += (Hash << 3); + Hash ^= (Hash >> 11); + Hash += (Hash << 15); + return Hash; +} + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size < 14) + return 0; + + uint32_t Hash = simple_hash(&Data[0], Size - 4); + uint32_t Want = reinterpret_cast(&Data[Size - 4])[0]; + if (Hash != Want) + return 0; + fprintf(stderr, "BINGO; simple_hash defeated: %x == %x\n", (unsigned int)Hash, + (unsigned int)Want); + exit(1); + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/SimpleTest.cpp b/test/thirdparty/Fuzzer/test/SimpleTest.cpp new file mode 100644 index 00000000..e53ea160 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/SimpleTest.cpp @@ -0,0 +1,27 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find the string "Hi!". +#include +#include +#include +#include +#include + +static volatile int Sink; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + assert(Data); + if (Size > 0 && Data[0] == 'H') { + Sink = 1; + if (Size > 1 && Data[1] == 'i') { + Sink = 2; + if (Size > 2 && Data[2] == '!') { + std::cout << "BINGO; Found the target, exiting\n"; + exit(0); + } + } + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/SimpleThreadedTest.cpp b/test/thirdparty/Fuzzer/test/SimpleThreadedTest.cpp new file mode 100644 index 00000000..5f02d3f8 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/SimpleThreadedTest.cpp @@ -0,0 +1,25 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Threaded test for a fuzzer. The fuzzer should find "H" +#include +#include +#include +#include +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + auto C = [&] { + if (Size >= 2 && Data[0] == 'H') { + std::cout << "BINGO; Found the target, exiting\n"; + abort(); + } + }; + std::thread T[] = {std::thread(C), std::thread(C), std::thread(C), + std::thread(C), std::thread(C), std::thread(C)}; + for (auto &X : T) + X.join(); + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/SingleMemcmpTest.cpp b/test/thirdparty/Fuzzer/test/SingleMemcmpTest.cpp new file mode 100644 index 00000000..c73f68a7 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/SingleMemcmpTest.cpp @@ -0,0 +1,17 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find a particular string. +#include +#include +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + char *S = (char*)Data; + if (Size >= 6 && !memcmp(S, "qwerty", 6)) { + fprintf(stderr, "BINGO\n"); + exit(1); + } + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/SingleStrcmpTest.cpp b/test/thirdparty/Fuzzer/test/SingleStrcmpTest.cpp new file mode 100644 index 00000000..73470b52 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/SingleStrcmpTest.cpp @@ -0,0 +1,17 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find a particular string. +#include +#include +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + char *S = (char*)Data; + if (Size >= 7 && !strcmp(S, "qwerty")) { + fprintf(stderr, "BINGO\n"); + exit(1); + } + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/SingleStrncmpTest.cpp b/test/thirdparty/Fuzzer/test/SingleStrncmpTest.cpp new file mode 100644 index 00000000..dbcc464b --- /dev/null +++ b/test/thirdparty/Fuzzer/test/SingleStrncmpTest.cpp @@ -0,0 +1,17 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find a particular string. +#include +#include +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + char *S = (char*)Data; + if (Size >= 6 && !strncmp(S, "qwerty", 6)) { + fprintf(stderr, "BINGO\n"); + exit(1); + } + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/SpamyTest.cpp b/test/thirdparty/Fuzzer/test/SpamyTest.cpp new file mode 100644 index 00000000..d294d4dc --- /dev/null +++ b/test/thirdparty/Fuzzer/test/SpamyTest.cpp @@ -0,0 +1,21 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// The test spams to stderr and stdout. +#include +#include +#include +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + assert(Data); + printf("PRINTF_STDOUT\n"); + fflush(stdout); + fprintf(stderr, "PRINTF_STDERR\n"); + std::cout << "STREAM_COUT\n"; + std::cout.flush(); + std::cerr << "STREAM_CERR\n"; + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/StrcmpTest.cpp b/test/thirdparty/Fuzzer/test/StrcmpTest.cpp new file mode 100644 index 00000000..cd91dda7 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/StrcmpTest.cpp @@ -0,0 +1,32 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Break through a series of strcmp. +#include +#include +#include +#include +#include + +bool Eq(const uint8_t *Data, size_t Size, const char *Str) { + char Buff[1024]; + size_t Len = strlen(Str); + if (Size < Len) return false; + if (Len >= sizeof(Buff)) return false; + memcpy(Buff, (char*)Data, Len); + Buff[Len] = 0; + int res = strcmp(Buff, Str); + return res == 0; +} + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Eq(Data, Size, "ABC") && + Size >= 3 && Eq(Data + 3, Size - 3, "QWER") && + Size >= 7 && Eq(Data + 7, Size - 7, "ZXCVN") && + Size >= 14 && Data[13] == 42 + ) { + fprintf(stderr, "BINGO\n"); + exit(1); + } + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/StrncmpOOBTest.cpp b/test/thirdparty/Fuzzer/test/StrncmpOOBTest.cpp new file mode 100644 index 00000000..f70b003a --- /dev/null +++ b/test/thirdparty/Fuzzer/test/StrncmpOOBTest.cpp @@ -0,0 +1,21 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Test that libFuzzer itself does not read out of bounds. +#include +#include +#include +#include +#include +#include + +static volatile int Sink; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size < 5) return 0; + const char *Ch = reinterpret_cast(Data); + if (Ch[Size - 3] == 'a') + Sink = strncmp(Ch + Size - 3, "abcdefg", 6); + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/StrncmpTest.cpp b/test/thirdparty/Fuzzer/test/StrncmpTest.cpp new file mode 100644 index 00000000..5ffd011d --- /dev/null +++ b/test/thirdparty/Fuzzer/test/StrncmpTest.cpp @@ -0,0 +1,28 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find a particular string. +#include +#include +#include +#include + +static volatile int sink; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + // TODO: check other sizes. + char *S = (char*)Data; + if (Size >= 8 && strncmp(S, "123", 8)) + sink = 1; + if (Size >= 8 && strncmp(S, "01234567", 8) == 0) { + if (Size >= 12 && strncmp(S + 8, "ABCD", 4) == 0) { + if (Size >= 14 && strncmp(S + 12, "XY", 2) == 0) { + if (Size >= 17 && strncmp(S + 14, "KLM", 3) == 0) { + fprintf(stderr, "BINGO\n"); + exit(1); + } + } + } + } + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/StrstrTest.cpp b/test/thirdparty/Fuzzer/test/StrstrTest.cpp new file mode 100644 index 00000000..f021e75e --- /dev/null +++ b/test/thirdparty/Fuzzer/test/StrstrTest.cpp @@ -0,0 +1,28 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Test strstr and strcasestr hooks. +#include +#include +#include +#include +#include + +// Windows does not have strcasestr and memmem, so we are not testing them. +#ifdef _WIN32 +#define strcasestr strstr +#define memmem(a, b, c, d) true +#endif + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size < 4) return 0; + std::string s(reinterpret_cast(Data), Size); + if (strstr(s.c_str(), "FUZZ") && + strcasestr(s.c_str(), "aBcD") && + memmem(s.data(), s.size(), "kuku", 4) + ) { + fprintf(stderr, "BINGO\n"); + exit(1); + } + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/SwapCmpTest.cpp b/test/thirdparty/Fuzzer/test/SwapCmpTest.cpp new file mode 100644 index 00000000..f79db4cc --- /dev/null +++ b/test/thirdparty/Fuzzer/test/SwapCmpTest.cpp @@ -0,0 +1,34 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// The fuzzer must find several constants with swapped bytes. +#include +#include +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size < 14) return 0; + uint64_t x = 0; + uint32_t y = 0; + uint16_t z = 0; + memcpy(&x, Data, sizeof(x)); + memcpy(&y, Data + Size / 2, sizeof(y)); + memcpy(&z, Data + Size - sizeof(z), sizeof(z)); + + x = __builtin_bswap64(x); + y = __builtin_bswap32(y); + z = __builtin_bswap16(z); + + if (x == 0x46555A5A5A5A5546ULL && + z == 0x4F4B && + y == 0x66757A7A && + true + ) { + if (Data[Size - 3] == 'z') { + fprintf(stderr, "BINGO; Found the target\n"); + exit(1); + } + } + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/Switch2Test.cpp b/test/thirdparty/Fuzzer/test/Switch2Test.cpp new file mode 100644 index 00000000..3c6a3004 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/Switch2Test.cpp @@ -0,0 +1,35 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find the interesting switch value. +#include +#include +#include +#include +#include + +int Switch(int a) { + switch(a) { + case 100001: return 1; + case 100002: return 2; + case 100003: return 4; + } + return 0; +} + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + const int N = 3; + if (Size < N * sizeof(int)) return 0; + int Res = 0; + for (int i = 0; i < N; i++) { + int X; + memcpy(&X, Data + i * sizeof(int), sizeof(int)); + Res += Switch(X); + } + if (Res == 5 || Res == 3 || Res == 6 || Res == 7) { + fprintf(stderr, "BINGO; Found the target, exiting; Res=%d\n", Res); + exit(1); + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/SwitchTest.cpp b/test/thirdparty/Fuzzer/test/SwitchTest.cpp new file mode 100644 index 00000000..74e86c06 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/SwitchTest.cpp @@ -0,0 +1,58 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find the interesting switch value. +#include +#include +#include +#include +#include + +static volatile int Sink; + +template +bool Switch(const uint8_t *Data, size_t Size) { + T X; + if (Size < sizeof(X)) return false; + memcpy(&X, Data, sizeof(X)); + switch (X) { + case 1: Sink = __LINE__; break; + case 101: Sink = __LINE__; break; + case 1001: Sink = __LINE__; break; + case 10001: Sink = __LINE__; break; +// case 100001: Sink = __LINE__; break; +// case 1000001: Sink = __LINE__; break; + case 10000001: Sink = __LINE__; break; + case 100000001: return true; + } + return false; +} + +bool ShortSwitch(const uint8_t *Data, size_t Size) { + short X; + if (Size < sizeof(short)) return false; + memcpy(&X, Data, sizeof(short)); + switch(X) { + case 42: Sink = __LINE__; break; + case 402: Sink = __LINE__; break; + case 4002: Sink = __LINE__; break; + case 5002: Sink = __LINE__; break; + case 7002: Sink = __LINE__; break; + case 9002: Sink = __LINE__; break; + case 14002: Sink = __LINE__; break; + case 21402: return true; + } + return false; +} + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size >= 4 && Switch(Data, Size) && + Size >= 12 && Switch(Data + 4, Size - 4) && + Size >= 14 && ShortSwitch(Data + 12, 2) + ) { + fprintf(stderr, "BINGO; Found the target, exiting\n"); + exit(1); + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/ThreadedLeakTest.cpp b/test/thirdparty/Fuzzer/test/ThreadedLeakTest.cpp new file mode 100644 index 00000000..75110711 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/ThreadedLeakTest.cpp @@ -0,0 +1,18 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// The fuzzer should find a leak in a non-main thread. +#include +#include +#include + +static volatile int *Sink; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size == 0) return 0; + if (Data[0] != 'F') return 0; + std::thread T([&] { Sink = new int; }); + T.join(); + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/ThreadedTest.cpp b/test/thirdparty/Fuzzer/test/ThreadedTest.cpp new file mode 100644 index 00000000..09137a9a --- /dev/null +++ b/test/thirdparty/Fuzzer/test/ThreadedTest.cpp @@ -0,0 +1,26 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Threaded test for a fuzzer. The fuzzer should not crash. +#include +#include +#include +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size < 8) return 0; + assert(Data); + auto C = [&] { + size_t Res = 0; + for (size_t i = 0; i < Size / 2; i++) + Res += memcmp(Data, Data + Size / 2, 4); + return Res; + }; + std::thread T[] = {std::thread(C), std::thread(C), std::thread(C), + std::thread(C), std::thread(C), std::thread(C)}; + for (auto &X : T) + X.join(); + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/TimeoutEmptyTest.cpp b/test/thirdparty/Fuzzer/test/TimeoutEmptyTest.cpp new file mode 100644 index 00000000..8066f480 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/TimeoutEmptyTest.cpp @@ -0,0 +1,14 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find the empty string. +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + static volatile int Zero = 0; + if (!Size) + while(!Zero) + ; + return 0; +} diff --git a/test/thirdparty/Fuzzer/test/TimeoutTest.cpp b/test/thirdparty/Fuzzer/test/TimeoutTest.cpp new file mode 100644 index 00000000..f8107012 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/TimeoutTest.cpp @@ -0,0 +1,26 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Simple test for a fuzzer. The fuzzer must find the string "Hi!". +#include +#include +#include +#include + +static volatile int Sink; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (Size > 0 && Data[0] == 'H') { + Sink = 1; + if (Size > 1 && Data[1] == 'i') { + Sink = 2; + if (Size > 2 && Data[2] == '!') { + Sink = 2; + while (Sink) + ; + } + } + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/TraceMallocTest.cpp b/test/thirdparty/Fuzzer/test/TraceMallocTest.cpp new file mode 100644 index 00000000..43e6950e --- /dev/null +++ b/test/thirdparty/Fuzzer/test/TraceMallocTest.cpp @@ -0,0 +1,27 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// Tests -trace_malloc +#include +#include +#include +#include +#include + +int *Ptr; + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + if (!Size) return 0; + if (*Data == 1) { + delete Ptr; + Ptr = nullptr; + } else if (*Data == 2) { + delete Ptr; + Ptr = new int; + } else if (*Data == 3) { + if (!Ptr) + Ptr = new int; + } + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/UninstrumentedTest.cpp b/test/thirdparty/Fuzzer/test/UninstrumentedTest.cpp new file mode 100644 index 00000000..ffe952c7 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/UninstrumentedTest.cpp @@ -0,0 +1,11 @@ +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. + +// This test should not be instrumented. +#include +#include + +extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + return 0; +} + diff --git a/test/thirdparty/Fuzzer/test/afl-driver-extra-stats.test b/test/thirdparty/Fuzzer/test/afl-driver-extra-stats.test new file mode 100644 index 00000000..81e384e7 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/afl-driver-extra-stats.test @@ -0,0 +1,28 @@ +; Test that not specifying an extra stats file isn't broken. +RUN: unset AFL_DRIVER_EXTRA_STATS_FILENAME +RUN: AFLDriverTest + +; Test that specifying an invalid extra stats file causes a crash. +RUN: ASAN_OPTIONS= AFL_DRIVER_EXTRA_STATS_FILENAME=%T not --crash AFLDriverTest + +; Test that specifying a corrupted stats file causes a crash. +echo "peak_rss_mb :0" > %t +ASAN_OPTIONS= AFL_DRIVER_EXTRA_STATS_FILENAME=%t not --crash AFLDriverTest + +; Test that specifying a valid nonexistent stats file works. +RUN: rm -f %t +RUN: AFL_DRIVER_EXTRA_STATS_FILENAME=%t AFLDriverTest +RUN: [[ $(grep "peak_rss_mb\|slowest_unit_time_sec" %t | wc -l) -eq 2 ]] + +; Test that specifying a valid preexisting stats file works. +RUN: printf "peak_rss_mb : 0\nslowest_unit_time_sec: 0\n" > %t +RUN: AFL_DRIVER_EXTRA_STATS_FILENAME=%t AFLDriverTest +; Check that both lines were printed. +RUN: [[ $(grep "peak_rss_mb\|slowest_unit_time_sec" %t | wc -l) -eq 2 ]] + +; Test that peak_rss_mb and slowest_unit_time_in_secs are only updated when necessary. +; Check that both lines have 9999 since there's no way we have exceeded that +; amount of time or virtual memory. +RUN: printf "peak_rss_mb : 9999\nslowest_unit_time_sec: 9999\n" > %t +RUN: AFL_DRIVER_EXTRA_STATS_FILENAME=%t AFLDriverTest +RUN: [[ $(grep "9999" %t | wc -l) -eq 2 ]] diff --git a/test/thirdparty/Fuzzer/test/afl-driver-stderr.test b/test/thirdparty/Fuzzer/test/afl-driver-stderr.test new file mode 100644 index 00000000..c0f9c839 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/afl-driver-stderr.test @@ -0,0 +1,10 @@ +; Test that not specifying a stderr file isn't broken. +RUN: unset AFL_DRIVER_STDERR_DUPLICATE_FILENAME +RUN: AFLDriverTest + +; Test that specifying an invalid file causes a crash. +RUN: ASAN_OPTIONS= AFL_DRIVER_STDERR_DUPLICATE_FILENAME="%T" not --crash AFLDriverTest + +; Test that a file is created when specified as the duplicate stderr. +RUN: AFL_DRIVER_STDERR_DUPLICATE_FILENAME=%t AFLDriverTest +RUN: stat %t diff --git a/test/thirdparty/Fuzzer/test/caller-callee.test b/test/thirdparty/Fuzzer/test/caller-callee.test new file mode 100644 index 00000000..76a951c5 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/caller-callee.test @@ -0,0 +1,2 @@ +CHECK: BINGO +RUN: not LLVMFuzzer-CallerCalleeTest -use_value_profile=1 -cross_over=0 -max_len=6 -seed=1 -runs=10000000 2>&1 | FileCheck %s diff --git a/test/thirdparty/Fuzzer/test/coverage.test b/test/thirdparty/Fuzzer/test/coverage.test new file mode 100644 index 00000000..fa11be50 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/coverage.test @@ -0,0 +1,19 @@ +CHECK: COVERAGE: +CHECK-DAG: COVERED: {{.*}}in LLVMFuzzerTestOneInput {{.*}}NullDerefTest.cpp:13 +CHECK-DAG: COVERED: {{.*}}in LLVMFuzzerTestOneInput {{.*}}NullDerefTest.cpp:14 +CHECK-DAG: COVERED: {{.*}}in LLVMFuzzerTestOneInput {{.*}}NullDerefTest.cpp:16 +CHECK-DAG: COVERED: {{.*}}in LLVMFuzzerTestOneInput {{.*}}NullDerefTest.cpp:19 +CHECK: COVERED_DIRS: {{.*}}lib/Fuzzer/test +RUN: not LLVMFuzzer-NullDerefTest -print_coverage=1 2>&1 | FileCheck %s + +RUN: LLVMFuzzer-DSOTest -print_coverage=1 -runs=0 2>&1 | FileCheck %s --check-prefix=DSO +DSO: COVERAGE: +DSO-DAG: COVERED:{{.*}}DSO1{{.*}}DSO1.cpp +DSO-DAG: COVERED:{{.*}}DSO2{{.*}}DSO2.cpp +DSO-DAG: COVERED:{{.*}}LLVMFuzzerTestOneInput{{.*}}DSOTestMain +DSO-DAG: UNCOVERED_LINE:{{.*}}DSO1{{.*}}DSO1.cpp +DSO-DAG: UNCOVERED_LINE:{{.*}}DSO2{{.*}}DSO2.cpp +DSO-DAG: UNCOVERED_FUNC: in Uncovered1 +DSO-DAG: UNCOVERED_FUNC: in Uncovered2 +DSO-DAG: UNCOVERED_LINE: in LLVMFuzzerTestOneInput +DSO-DAG: UNCOVERED_FILE:{{.*}}DSOTestExtra.cpp diff --git a/test/thirdparty/Fuzzer/test/dict1.txt b/test/thirdparty/Fuzzer/test/dict1.txt new file mode 100644 index 00000000..520d0cc7 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/dict1.txt @@ -0,0 +1,4 @@ +# Dictionary for SimpleDictionaryTest + +a="Elvis" +b="Presley" diff --git a/test/thirdparty/Fuzzer/test/dump_coverage.test b/test/thirdparty/Fuzzer/test/dump_coverage.test new file mode 100644 index 00000000..9bd98daa --- /dev/null +++ b/test/thirdparty/Fuzzer/test/dump_coverage.test @@ -0,0 +1,16 @@ +RUN: DIR=%t_workdir +RUN: BUILD_DIR=$(pwd) +RUN: rm -rf $DIR && mkdir -p $DIR && cd $DIR +RUN: not $BUILD_DIR/LLVMFuzzer-NullDerefTest -dump_coverage=1 2>&1 | FileCheck %s +RUN: $BUILD_DIR/LLVMFuzzer-DSOTest -dump_coverage=1 -runs=0 2>&1 | FileCheck %s --check-prefix=DSO +RUN: not $BUILD_DIR/LLVMFuzzer-NullDerefTest -dump_coverage=0 2>&1 | FileCheck %s --check-prefix=NOCOV +RUN: rm -rf $DIR + + +CHECK: SanitizerCoverage: ./LLVMFuzzer-NullDerefTest.{{.*}}.sancov {{.*}} PCs written + +DSO: SanitizerCoverage: ./LLVMFuzzer-DSOTest.{{.*}}.sancov {{.*}} PCs written +DSO-DAG: SanitizerCoverage: ./libLLVMFuzzer-DSO1.{{.*}}.sancov {{.*}} PCs written +DSO-DAG: SanitizerCoverage: ./libLLVMFuzzer-DSO2.{{.*}}.sancov {{.*}} PCs written + +NOCOV-NOT: SanitizerCoverage: {{.*}} PCs written diff --git a/test/thirdparty/Fuzzer/test/fuzzer-customcrossover.test b/test/thirdparty/Fuzzer/test/fuzzer-customcrossover.test new file mode 100644 index 00000000..28d39ce3 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-customcrossover.test @@ -0,0 +1,10 @@ +RUN: rm -rf %t/CustomCrossover +RUN: mkdir -p %t/CustomCrossover +RUN: echo "0123456789" > %t/CustomCrossover/digits +RUN: echo "abcdefghij" > %t/CustomCrossover/chars +RUN: not LLVMFuzzer-CustomCrossOverTest -seed=1 -use_memcmp=0 -runs=100000 %t/CustomCrossover 2>&1 | FileCheck %s --check-prefix=LLVMFuzzerCustomCrossover +RUN: rm -rf %t/CustomCrossover + +LLVMFuzzerCustomCrossover: In LLVMFuzzerCustomCrossover +LLVMFuzzerCustomCrossover: BINGO + diff --git a/test/thirdparty/Fuzzer/test/fuzzer-custommutator.test b/test/thirdparty/Fuzzer/test/fuzzer-custommutator.test new file mode 100644 index 00000000..fcd740bf --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-custommutator.test @@ -0,0 +1,4 @@ +RUN: not LLVMFuzzer-CustomMutatorTest 2>&1 | FileCheck %s --check-prefix=LLVMFuzzerCustomMutator +LLVMFuzzerCustomMutator: In LLVMFuzzerCustomMutator +LLVMFuzzerCustomMutator: BINGO + diff --git a/test/thirdparty/Fuzzer/test/fuzzer-dict.test b/test/thirdparty/Fuzzer/test/fuzzer-dict.test new file mode 100644 index 00000000..dec002f6 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-dict.test @@ -0,0 +1,6 @@ +CHECK: BINGO +Done1000000: Done 1000000 runs in + +RUN: not LLVMFuzzer-SimpleDictionaryTest -dict=%S/dict1.txt -seed=1 -runs=1000003 2>&1 | FileCheck %s +RUN: LLVMFuzzer-SimpleDictionaryTest -seed=1 -runs=1000000 2>&1 | FileCheck %s --check-prefix=Done1000000 + diff --git a/test/thirdparty/Fuzzer/test/fuzzer-dirs.test b/test/thirdparty/Fuzzer/test/fuzzer-dirs.test new file mode 100644 index 00000000..63afe8df --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-dirs.test @@ -0,0 +1,15 @@ +RUN: rm -rf %t/SUB1 +RUN: mkdir -p %t/SUB1/SUB2/SUB3 +RUN: echo a > %t/SUB1/a +RUN: echo b > %t/SUB1/SUB2/b +RUN: echo c > %t/SUB1/SUB2/SUB3/c +RUN: LLVMFuzzer-SimpleTest %t/SUB1 -runs=0 2>&1 | FileCheck %s --check-prefix=SUBDIRS +SUBDIRS: READ units: 3 +RUN: echo zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz > %t/SUB1/long +RUN: LLVMFuzzer-SimpleTest %t/SUB1 -runs=0 2>&1 | FileCheck %s --check-prefix=LONG +LONG: INFO: -max_len is not provided, using 94 +RUN: rm -rf %t/SUB1 + +RUN: not LLVMFuzzer-SimpleTest NONEXISTENT_DIR 2>&1 | FileCheck %s --check-prefix=NONEXISTENT_DIR +NONEXISTENT_DIR: No such directory: NONEXISTENT_DIR; exiting + diff --git a/test/thirdparty/Fuzzer/test/fuzzer-fdmask.test b/test/thirdparty/Fuzzer/test/fuzzer-fdmask.test new file mode 100644 index 00000000..abbc4bd6 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-fdmask.test @@ -0,0 +1,30 @@ +RUN: LLVMFuzzer-SpamyTest -runs=1 2>&1 | FileCheck %s --check-prefix=FD_MASK_0 +RUN: LLVMFuzzer-SpamyTest -runs=1 -close_fd_mask=0 2>&1 | FileCheck %s --check-prefix=FD_MASK_0 +RUN: LLVMFuzzer-SpamyTest -runs=1 -close_fd_mask=1 2>&1 | FileCheck %s --check-prefix=FD_MASK_1 +RUN: LLVMFuzzer-SpamyTest -runs=1 -close_fd_mask=2 2>&1 | FileCheck %s --check-prefix=FD_MASK_2 +RUN: LLVMFuzzer-SpamyTest -runs=1 -close_fd_mask=3 2>&1 | FileCheck %s --check-prefix=FD_MASK_3 + +FD_MASK_0: PRINTF_STDOUT +FD_MASK_0: PRINTF_STDERR +FD_MASK_0: STREAM_COUT +FD_MASK_0: STREAM_CERR +FD_MASK_0: INITED + +FD_MASK_1-NOT: PRINTF_STDOUT +FD_MASK_1: PRINTF_STDERR +FD_MASK_1-NOT: STREAM_COUT +FD_MASK_1: STREAM_CERR +FD_MASK_1: INITED + +FD_MASK_2: PRINTF_STDOUT +FD_MASK_2-NOT: PRINTF_STDERR +FD_MASK_2: STREAM_COUT +FD_MASK_2-NOTE: STREAM_CERR +FD_MASK_2: INITED + +FD_MASK_3-NOT: PRINTF_STDOUT +FD_MASK_3-NOT: PRINTF_STDERR +FD_MASK_3-NOT: STREAM_COUT +FD_MASK_3-NOT: STREAM_CERR +FD_MASK_3: INITED + diff --git a/test/thirdparty/Fuzzer/test/fuzzer-finalstats.test b/test/thirdparty/Fuzzer/test/fuzzer-finalstats.test new file mode 100644 index 00000000..1cbcd10f --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-finalstats.test @@ -0,0 +1,11 @@ +RUN: LLVMFuzzer-SimpleTest -seed=1 -runs=77 -print_final_stats=1 2>&1 | FileCheck %s --check-prefix=FINAL_STATS +FINAL_STATS: stat::number_of_executed_units: 77 +FINAL_STATS: stat::average_exec_per_sec: 0 +FINAL_STATS: stat::new_units_added: +FINAL_STATS: stat::slowest_unit_time_sec: 0 +FINAL_STATS: stat::peak_rss_mb: + +RUN: LLVMFuzzer-SimpleTest %S/dict1.txt -runs=33 -print_final_stats=1 2>&1 | FileCheck %s --check-prefix=FINAL_STATS1 +FINAL_STATS1: stat::number_of_executed_units: 33 +FINAL_STATS1: stat::peak_rss_mb: + diff --git a/test/thirdparty/Fuzzer/test/fuzzer-flags.test b/test/thirdparty/Fuzzer/test/fuzzer-flags.test new file mode 100644 index 00000000..76ea2770 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-flags.test @@ -0,0 +1,10 @@ +RUN: LLVMFuzzer-SimpleTest -foo_bar=1 2>&1 | FileCheck %s --check-prefix=FOO_BAR +FOO_BAR: WARNING: unrecognized flag '-foo_bar=1'; use -help=1 to list all flags +FOO_BAR: BINGO + +RUN: LLVMFuzzer-SimpleTest -runs=10 --max_len=100 2>&1 | FileCheck %s --check-prefix=DASH_DASH +DASH_DASH: WARNING: did you mean '-max_len=100' (single dash)? +DASH_DASH: INFO: A corpus is not provided, starting from an empty corpus + +RUN: LLVMFuzzer-SimpleTest -help=1 2>&1 | FileCheck %s --check-prefix=NO_INTERNAL +NO_INTERNAL-NOT: internal flag diff --git a/test/thirdparty/Fuzzer/test/fuzzer-jobs.test b/test/thirdparty/Fuzzer/test/fuzzer-jobs.test new file mode 100644 index 00000000..5bf8cfad --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-jobs.test @@ -0,0 +1,29 @@ +RUN: rm -rf %tmp +RUN: mkdir %tmp && cd %tmp +# Create a shared corpus directory +RUN: rm -rf FuzzerJobsTestCORPUS +RUN: mkdir FuzzerJobsTestCORPUS +RUN: rm -f fuzz-{0,1}.log +# Start fuzzer and in parallel check that the output files +# that should be created exist. +RUN: LLVMFuzzer-EmptyTest -max_total_time=4 -jobs=2 -workers=2 FuzzerJobsTestCORPUS > %t-fuzzer-jobs-test.log 2>&1 & export FUZZER_PID=$! +# Wait a short while to give time for the child processes +# to start fuzzing +RUN: sleep 2 +# If the instances are running in parallel they should have created their log +# files by now. +RUN: ls fuzz-0.log +RUN: ls fuzz-1.log +# Wait for libfuzzer to finish. +# This probably isn't portable but we need a way to block until +# the fuzzer is done otherwise we might remove the files while +# they are being used. +RUN: while kill -0 ${FUZZER_PID}; do : ; done +RUN: rm -f fuzz-{0,1}.log +RUN: rm -rf FuzzerJobsTestCORPUS +RUN: FileCheck -input-file=%t-fuzzer-jobs-test.log %s +RUN: rm %t-fuzzer-jobs-test.log +RUN: cd ../ + +CHECK-DAG: Job 0 exited with exit code 0 +CHECK-DAG: Job 1 exited with exit code 0 diff --git a/test/thirdparty/Fuzzer/test/fuzzer-leak.test b/test/thirdparty/Fuzzer/test/fuzzer-leak.test new file mode 100644 index 00000000..9cf5c743 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-leak.test @@ -0,0 +1,35 @@ +REQUIRES: lsan +RUN: not LLVMFuzzer-LeakTest -runs=100000 -detect_leaks=1 2>&1 | FileCheck %s --check-prefix=LEAK_DURING +LEAK_DURING: ERROR: LeakSanitizer: detected memory leaks +LEAK_DURING: Direct leak of 4 byte(s) in 1 object(s) allocated from: +LEAK_DURING: INFO: to ignore leaks on libFuzzer side use -detect_leaks=0 +LEAK_DURING: Test unit written to ./leak- +LEAK_DURING-NOT: DONE +LEAK_DURING-NOT: Done + +RUN: not LLVMFuzzer-LeakTest -runs=0 -detect_leaks=1 %S 2>&1 | FileCheck %s --check-prefix=LEAK_IN_CORPUS +LEAK_IN_CORPUS: ERROR: LeakSanitizer: detected memory leaks +LEAK_IN_CORPUS: INFO: a leak has been found in the initial corpus. + +RUN: not LLVMFuzzer-LeakTest -runs=100000000 %S/hi.txt 2>&1 | FileCheck %s --check-prefix=MULTI_RUN_LEAK +MULTI_RUN_LEAK-NOT: pulse +MULTI_RUN_LEAK: LeakSanitizer: detected memory leaks + +RUN: not LLVMFuzzer-LeakTest -runs=100000 -detect_leaks=0 2>&1 | FileCheck %s --check-prefix=LEAK_AFTER +RUN: not LLVMFuzzer-LeakTest -runs=100000 2>&1 | FileCheck %s --check-prefix=LEAK_DURING +RUN: not LLVMFuzzer-ThreadedLeakTest -runs=100000 -detect_leaks=0 2>&1 | FileCheck %s --check-prefix=LEAK_AFTER +RUN: not LLVMFuzzer-ThreadedLeakTest -runs=100000 2>&1 | FileCheck %s --check-prefix=LEAK_DURING +LEAK_AFTER: Done 100000 runs in +LEAK_AFTER: ERROR: LeakSanitizer: detected memory leaks + +RUN: not LLVMFuzzer-LeakTest -runs=100000 -max_len=1 2>&1 | FileCheck %s --check-prefix=MAX_LEN_1 +MAX_LEN_1: Test unit written to ./leak-7cf184f4c67ad58283ecb19349720b0cae756829 + +RUN: not LLVMFuzzer-LeakTimeoutTest -timeout=1 2>&1 | FileCheck %s --check-prefix=LEAK_TIMEOUT +LEAK_TIMEOUT: ERROR: libFuzzer: timeout after +LEAK_TIMEOUT-NOT: LeakSanitizer + +RUN: LLVMFuzzer-AccumulateAllocationsTest -detect_leaks=1 -runs=100000 2>&1 | FileCheck %s --check-prefix=ACCUMULATE_ALLOCS +ACCUMULATE_ALLOCS: INFO: libFuzzer disabled leak detection after every mutation + +RUN: LLVMFuzzer-LeakTest -error_exitcode=0 diff --git a/test/thirdparty/Fuzzer/test/fuzzer-oom-with-profile.test b/test/thirdparty/Fuzzer/test/fuzzer-oom-with-profile.test new file mode 100644 index 00000000..2b2b0b9d --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-oom-with-profile.test @@ -0,0 +1,6 @@ +REQUIRES: linux +RUN: not LLVMFuzzer-OutOfMemoryTest -rss_limit_mb=300 2>&1 | FileCheck %s +CHECK: ERROR: libFuzzer: out-of-memory (used: {{.*}}; limit: 300Mb) +CHECK: Live Heap Allocations +CHECK: Test unit written to ./oom- +SUMMARY: libFuzzer: out-of-memory diff --git a/test/thirdparty/Fuzzer/test/fuzzer-oom.test b/test/thirdparty/Fuzzer/test/fuzzer-oom.test new file mode 100644 index 00000000..8caf649e --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-oom.test @@ -0,0 +1,11 @@ +RUN: not LLVMFuzzer-OutOfMemoryTest -rss_limit_mb=300 2>&1 | FileCheck %s +CHECK: ERROR: libFuzzer: out-of-memory (used: {{.*}}; limit: 300Mb) +CHECK: Test unit written to ./oom- +SUMMARY: libFuzzer: out-of-memory + +RUN: not LLVMFuzzer-OutOfMemorySingleLargeMallocTest 2>&1 | FileCheck %s --check-prefix=SINGLE_LARGE_MALLOC +SINGLE_LARGE_MALLOC: libFuzzer: out-of-memory (malloc(42{{.*}})) +SINGLE_LARGE_MALLOC: in LLVMFuzzerTestOneInput + +# Check that -rss_limit_mb=0 means no limit. +RUN: LLVMFuzzer-AccumulateAllocationsTest -runs=1000 -rss_limit_mb=0 diff --git a/test/thirdparty/Fuzzer/test/fuzzer-printcovpcs.test b/test/thirdparty/Fuzzer/test/fuzzer-printcovpcs.test new file mode 100644 index 00000000..e4c6f0ed --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-printcovpcs.test @@ -0,0 +1,8 @@ +RUN: LLVMFuzzer-SimpleTest -print_pcs=1 -seed=1 2>&1 | FileCheck %s --check-prefix=PCS +PCS-NOT: NEW_PC +PCS:INITED +PCS:NEW_PC: {{0x[a-f0-9]+}} +PCS:NEW_PC: {{0x[a-f0-9]+}} +PCS:NEW +PCS:BINGO + diff --git a/test/thirdparty/Fuzzer/test/fuzzer-runs.test b/test/thirdparty/Fuzzer/test/fuzzer-runs.test new file mode 100644 index 00000000..056c4478 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-runs.test @@ -0,0 +1,8 @@ +RUN: mkdir -p %t +RUN: echo abcd > %t/NthRunCrashTest.in +RUN: LLVMFuzzer-NthRunCrashTest %t/NthRunCrashTest.in +RUN: LLVMFuzzer-NthRunCrashTest %t/NthRunCrashTest.in -runs=10 +RUN: not LLVMFuzzer-NthRunCrashTest %t/NthRunCrashTest.in -runs=10000 2>&1 | FileCheck %s +RUN: rm %t/NthRunCrashTest.in +CHECK: BINGO + diff --git a/test/thirdparty/Fuzzer/test/fuzzer-seed.test b/test/thirdparty/Fuzzer/test/fuzzer-seed.test new file mode 100644 index 00000000..f1bdf9e4 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-seed.test @@ -0,0 +1,3 @@ +RUN: LLVMFuzzer-SimpleCmpTest -seed=-1 -runs=0 2>&1 | FileCheck %s --check-prefix=CHECK_SEED_MINUS_ONE +CHECK_SEED_MINUS_ONE: Seed: 4294967295 + diff --git a/test/thirdparty/Fuzzer/test/fuzzer-segv.test b/test/thirdparty/Fuzzer/test/fuzzer-segv.test new file mode 100644 index 00000000..330f03bc --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-segv.test @@ -0,0 +1,5 @@ +RUN: ASAN_OPTIONS=handle_segv=0 not LLVMFuzzer-NullDerefTest 2>&1 | FileCheck %s --check-prefix=LIBFUZZER_OWN_SEGV_HANDLER +LIBFUZZER_OWN_SEGV_HANDLER: == ERROR: libFuzzer: deadly signal +LIBFUZZER_OWN_SEGV_HANDLER: SUMMARY: libFuzzer: deadly signal +LIBFUZZER_OWN_SEGV_HANDLER: Test unit written to ./crash- + diff --git a/test/thirdparty/Fuzzer/test/fuzzer-singleinputs.test b/test/thirdparty/Fuzzer/test/fuzzer-singleinputs.test new file mode 100644 index 00000000..ca8403bf --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-singleinputs.test @@ -0,0 +1,16 @@ +RUN: not LLVMFuzzer-NullDerefTest %S/hi.txt 2>&1 | FileCheck %s --check-prefix=SingleInput +SingleInput-NOT: Test unit written to ./crash- + +RUN: rm -rf %tmp/SINGLE_INPUTS +RUN: mkdir -p %tmp/SINGLE_INPUTS +RUN: echo aaa > %tmp/SINGLE_INPUTS/aaa +RUN: echo bbb > %tmp/SINGLE_INPUTS/bbb +RUN: LLVMFuzzer-SimpleTest %tmp/SINGLE_INPUTS/aaa %tmp/SINGLE_INPUTS/bbb 2>&1 | FileCheck %s --check-prefix=SINGLE_INPUTS +RUN: LLVMFuzzer-SimpleTest -max_len=2 %tmp/SINGLE_INPUTS/aaa %tmp/SINGLE_INPUTS/bbb 2>&1 | FileCheck %s --check-prefix=SINGLE_INPUTS +RUN: rm -rf %tmp/SINGLE_INPUTS +SINGLE_INPUTS: LLVMFuzzer-SimpleTest: Running 2 inputs 1 time(s) each. +SINGLE_INPUTS: aaa in +SINGLE_INPUTS: bbb in +SINGLE_INPUTS: NOTE: fuzzing was not performed, you have only +SINGLE_INPUTS: executed the target code on a fixed set of inputs. + diff --git a/test/thirdparty/Fuzzer/test/fuzzer-threaded.test b/test/thirdparty/Fuzzer/test/fuzzer-threaded.test new file mode 100644 index 00000000..c58a3345 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-threaded.test @@ -0,0 +1,7 @@ +CHECK: Done 1000 runs in + +RUN: LLVMFuzzer-ThreadedTest -use_traces=1 -runs=1000 2>&1 | FileCheck %s +RUN: LLVMFuzzer-ThreadedTest -use_traces=1 -runs=1000 2>&1 | FileCheck %s +RUN: LLVMFuzzer-ThreadedTest -use_traces=1 -runs=1000 2>&1 | FileCheck %s +RUN: LLVMFuzzer-ThreadedTest -use_traces=1 -runs=1000 2>&1 | FileCheck %s + diff --git a/test/thirdparty/Fuzzer/test/fuzzer-timeout.test b/test/thirdparty/Fuzzer/test/fuzzer-timeout.test new file mode 100644 index 00000000..beb08671 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-timeout.test @@ -0,0 +1,19 @@ +RUN: not LLVMFuzzer-TimeoutTest -timeout=1 2>&1 | FileCheck %s --check-prefix=TimeoutTest +TimeoutTest: ALARM: working on the last Unit for +TimeoutTest: Test unit written to ./timeout- +TimeoutTest: == ERROR: libFuzzer: timeout after +TimeoutTest: #0 +TimeoutTest: #1 +TimeoutTest: #2 +TimeoutTest: SUMMARY: libFuzzer: timeout + +RUN: not LLVMFuzzer-TimeoutTest -timeout=1 %S/hi.txt 2>&1 | FileCheck %s --check-prefix=SingleInputTimeoutTest +SingleInputTimeoutTest: ALARM: working on the last Unit for {{[1-3]}} seconds +SingleInputTimeoutTest-NOT: Test unit written to ./timeout- + +RUN: LLVMFuzzer-TimeoutTest -timeout=1 -timeout_exitcode=0 + +RUN: not LLVMFuzzer-TimeoutEmptyTest -timeout=1 2>&1 | FileCheck %s --check-prefix=TimeoutEmptyTest +TimeoutEmptyTest: ALARM: working on the last Unit for +TimeoutEmptyTest: == ERROR: libFuzzer: timeout after +TimeoutEmptyTest: SUMMARY: libFuzzer: timeout diff --git a/test/thirdparty/Fuzzer/test/fuzzer-traces-hooks.test b/test/thirdparty/Fuzzer/test/fuzzer-traces-hooks.test new file mode 100644 index 00000000..71fe6f2d --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-traces-hooks.test @@ -0,0 +1,25 @@ +// FIXME: Support sanitizer hooks for memcmp and strcmp need +// to be implemented in the sanitizer runtime for platforms other +// than linux +REQUIRES: linux +CHECK: BINGO +Done1000000: Done 1000000 runs in + +RUN: not LLVMFuzzer-MemcmpTest -seed=4294967295 -runs=100000 2>&1 | FileCheck %s +RUN: LLVMFuzzer-MemcmpTest -use_memcmp=0 -seed=4294967295 -runs=1000000 2>&1 | FileCheck %s --check-prefix=Done1000000 + +RUN: not LLVMFuzzer-StrncmpTest -seed=2 -runs=100000 2>&1 | FileCheck %s +RUN: LLVMFuzzer-StrncmpTest -use_memcmp=0 -seed=3 -runs=1000000 2>&1 | FileCheck %s --check-prefix=Done1000000 + +RUN: not LLVMFuzzer-StrcmpTest -seed=4 -runs=200000 2>&1 | FileCheck %s +RUN: LLVMFuzzer-StrcmpTest -use_memcmp=0 -seed=5 -runs=1000000 2>&1 | FileCheck %s --check-prefix=Done1000000 + +RUN: not LLVMFuzzer-StrstrTest -seed=6 -runs=200000 2>&1 | FileCheck %s +RUN: LLVMFuzzer-StrstrTest -use_memmem=0 -seed=7 -runs=1000000 2>&1 | FileCheck %s --check-prefix=Done1000000 + +RUN: LLVMFuzzer-RepeatedMemcmp -seed=10 -runs=100000 2>&1 | FileCheck %s --check-prefix=RECOMMENDED_DICT +RECOMMENDED_DICT:###### Recommended dictionary. ###### +RECOMMENDED_DICT-DAG: "foo" +RECOMMENDED_DICT-DAG: "bar" +RECOMMENDED_DICT:###### End of recommended dictionary. ###### + diff --git a/test/thirdparty/Fuzzer/test/fuzzer-ubsan.test b/test/thirdparty/Fuzzer/test/fuzzer-ubsan.test new file mode 100644 index 00000000..0e8ad6c9 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer-ubsan.test @@ -0,0 +1,4 @@ +RUN: not LLVMFuzzer-SignedIntOverflowTest-Ubsan 2>&1 | FileCheck %s +CHECK: runtime error: signed integer overflow: 2147483647 + 1 cannot be represented in type 'int' +CHECK: Test unit written to ./crash- + diff --git a/test/thirdparty/Fuzzer/test/fuzzer.test b/test/thirdparty/Fuzzer/test/fuzzer.test new file mode 100644 index 00000000..2f91c219 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/fuzzer.test @@ -0,0 +1,57 @@ +CHECK: BINGO +Done1000000: Done 1000000 runs in + +RUN: LLVMFuzzer-SimpleTest 2>&1 | FileCheck %s + +# only_ascii mode. Will perform some minimal self-validation. +RUN: LLVMFuzzer-SimpleTest -only_ascii=1 2>&1 + +RUN: LLVMFuzzer-SimpleCmpTest -max_total_time=1 -use_cmp=0 2>&1 | FileCheck %s --check-prefix=MaxTotalTime +MaxTotalTime: Done {{.*}} runs in {{.}} second(s) + +RUN: not LLVMFuzzer-NullDerefTest 2>&1 | FileCheck %s --check-prefix=NullDerefTest +RUN: not LLVMFuzzer-NullDerefTest -close_fd_mask=3 2>&1 | FileCheck %s --check-prefix=NullDerefTest +NullDerefTest: ERROR: AddressSanitizer: SEGV on unknown address +NullDerefTest: Test unit written to ./crash- +RUN: not LLVMFuzzer-NullDerefTest -artifact_prefix=ZZZ 2>&1 | FileCheck %s --check-prefix=NullDerefTestPrefix +NullDerefTestPrefix: Test unit written to ZZZcrash- +RUN: not LLVMFuzzer-NullDerefTest -artifact_prefix=ZZZ -exact_artifact_path=FOOBAR 2>&1 | FileCheck %s --check-prefix=NullDerefTestExactPath +NullDerefTestExactPath: Test unit written to FOOBAR + +RUN: not LLVMFuzzer-NullDerefOnEmptyTest -print_final_stats=1 2>&1 | FileCheck %s --check-prefix=NULL_DEREF_ON_EMPTY +NULL_DEREF_ON_EMPTY: stat::number_of_executed_units: + +#not LLVMFuzzer-FullCoverageSetTest -timeout=15 -seed=1 -mutate_depth=2 -use_full_coverage_set=1 2>&1 | FileCheck %s + +RUN: not LLVMFuzzer-CounterTest -max_len=6 -seed=1 -timeout=15 2>&1 | FileCheck %s --check-prefix=COUNTERS + +COUNTERS: INITED {{.*}} {{bits:|ft:}} +COUNTERS: NEW {{.*}} {{bits:|ft:}} {{[1-9]*}} +COUNTERS: NEW {{.*}} {{bits:|ft:}} {{[1-9]*}} +COUNTERS: BINGO + +# Don't run UninstrumentedTest for now since we build libFuzzer itself with asan. +DISABLED: not LLVMFuzzer-UninstrumentedTest-Uninstrumented 2>&1 | FileCheck %s --check-prefix=UNINSTRUMENTED +UNINSTRUMENTED: ERROR: __sanitizer_set_death_callback is not defined. Exiting. + +RUN: not LLVMFuzzer-UninstrumentedTest-NoCoverage 2>&1 | FileCheck %s --check-prefix=NO_COVERAGE +NO_COVERAGE: ERROR: no interesting inputs were found. Is the code instrumented for coverage? Exiting + +RUN: not LLVMFuzzer-BufferOverflowOnInput 2>&1 | FileCheck %s --check-prefix=OOB +OOB: AddressSanitizer: heap-buffer-overflow +OOB: is located 0 bytes to the right of 3-byte region + +RUN: not LLVMFuzzer-InitializeTest -use_value_profile=1 2>&1 | FileCheck %s + +RUN: not LLVMFuzzer-DSOTest 2>&1 | FileCheck %s --check-prefix=DSO +DSO: INFO: Loaded 3 modules +DSO: BINGO + +RUN: LLVMFuzzer-SimpleTest -exit_on_src_pos=SimpleTest.cpp:17 2>&1 | FileCheck %s --check-prefix=EXIT_ON_SRC_POS +RUN: LLVMFuzzer-ShrinkControlFlowTest -exit_on_src_pos=ShrinkControlFlowTest.cpp:23 2>&1 | FileCheck %s --check-prefix=EXIT_ON_SRC_POS +EXIT_ON_SRC_POS: INFO: found line matching '{{.*}}', exiting. + +RUN: ASAN_OPTIONS=strict_string_checks=1 not LLVMFuzzer-StrncmpOOBTest -seed=1 -runs=1000000 2>&1 | FileCheck %s --check-prefix=STRNCMP +STRNCMP: AddressSanitizer: heap-buffer-overflow +STRNCMP-NOT: __sanitizer_weak_hook_strncmp +STRNCMP: in LLVMFuzzerTestOneInput diff --git a/test/thirdparty/Fuzzer/test/hi.txt b/test/thirdparty/Fuzzer/test/hi.txt new file mode 100644 index 00000000..2f9031f0 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/hi.txt @@ -0,0 +1 @@ +Hi! \ No newline at end of file diff --git a/test/thirdparty/Fuzzer/test/lit.cfg b/test/thirdparty/Fuzzer/test/lit.cfg new file mode 100644 index 00000000..745af0c3 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/lit.cfg @@ -0,0 +1,29 @@ +import lit.formats +import sys + +config.name = "LLVMFuzzer" +config.test_format = lit.formats.ShTest(True) +config.suffixes = ['.test'] +config.test_source_root = os.path.dirname(__file__) + +# Tweak PATH to include llvm tools dir and current exec dir. +llvm_tools_dir = getattr(config, 'llvm_tools_dir', None) +if (not llvm_tools_dir) or (not os.path.exists(llvm_tools_dir)): + lit_config.fatal("Invalid llvm_tools_dir config attribute: %r" % llvm_tools_dir) +path = os.path.pathsep.join((llvm_tools_dir, config.test_exec_root, + config.environment['PATH'])) +config.environment['PATH'] = path + +if config.has_lsan: + lit_config.note('lsan feature available') + config.available_features.add('lsan') +else: + lit_config.note('lsan feature unavailable') + +if sys.platform.startswith('linux'): + # Note the value of ``sys.platform`` is not consistent + # between python 2 and 3, hence the use of ``.startswith()``. + lit_config.note('linux feature available') + config.available_features.add('linux') +else: + lit_config.note('linux feature unavailable') diff --git a/test/thirdparty/Fuzzer/test/lit.site.cfg.in b/test/thirdparty/Fuzzer/test/lit.site.cfg.in new file mode 100644 index 00000000..03e86c48 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/lit.site.cfg.in @@ -0,0 +1,4 @@ +config.test_exec_root = "@CMAKE_CURRENT_BINARY_DIR@" +config.llvm_tools_dir = "@LLVM_TOOLS_DIR@" +config.has_lsan = True if @HAS_LSAN@ == 1 else False +lit_config.load_config(config, "@CMAKE_CURRENT_SOURCE_DIR@/lit.cfg") diff --git a/test/thirdparty/Fuzzer/test/merge.test b/test/thirdparty/Fuzzer/test/merge.test new file mode 100644 index 00000000..1f1810eb --- /dev/null +++ b/test/thirdparty/Fuzzer/test/merge.test @@ -0,0 +1,46 @@ +CHECK: BINGO + +RUN: rm -rf %tmp/T1 %tmp/T2 +RUN: mkdir -p %tmp/T1 %tmp/T2 +RUN: echo F..... > %tmp/T1/1 +RUN: echo .U.... > %tmp/T1/2 +RUN: echo ..Z... > %tmp/T1/3 + +# T1 has 3 elements, T2 is empty. +RUN: LLVMFuzzer-FullCoverageSetTest -merge=1 %tmp/T1 %tmp/T2 2>&1 | FileCheck %s --check-prefix=CHECK1 +CHECK1: MERGE-OUTER: 3 files, 3 in the initial corpus +CHECK1: MERGE-OUTER: 0 new files with 0 new features added + +RUN: echo ...Z.. > %tmp/T2/1 +RUN: echo ....E. > %tmp/T2/2 +RUN: echo .....R > %tmp/T2/3 +RUN: echo F..... > %tmp/T2/a +RUN: echo .U.... > %tmp/T2/b +RUN: echo ..Z... > %tmp/T2/c + +# T1 has 3 elements, T2 has 6 elements, only 3 are new. +RUN: LLVMFuzzer-FullCoverageSetTest -merge=1 %tmp/T1 %tmp/T2 2>&1 | FileCheck %s --check-prefix=CHECK2 +CHECK2: MERGE-OUTER: 9 files, 3 in the initial corpus +CHECK2: MERGE-OUTER: 3 new files with 3 new features added + +# Now, T1 has 6 units and T2 has no new interesting units. +RUN: LLVMFuzzer-FullCoverageSetTest -merge=1 %tmp/T1 %tmp/T2 2>&1 | FileCheck %s --check-prefix=CHECK3 +CHECK3: MERGE-OUTER: 12 files, 6 in the initial corpus +CHECK3: MERGE-OUTER: 0 new files with 0 new features added + +# Check that we respect max_len during the merge and don't crash. +RUN: rm %tmp/T1/??* +RUN: echo looooooooong > %tmp/T2/looooooooong +RUN: LLVMFuzzer-FullCoverageSetTest -merge=1 %tmp/T1 %tmp/T2 -max_len=6 2>&1 | FileCheck %s --check-prefix=MAX_LEN +MAX_LEN: MERGE-OUTER: 3 new files + +# Check that merge tolerates failures. +RUN: rm %tmp/T1/??* +RUN: echo 'FUZZER' > %tmp/T2/FUZZER +RUN: LLVMFuzzer-FullCoverageSetTest -merge=1 %tmp/T1 %tmp/T2 2>&1 | FileCheck %s --check-prefix=MERGE_WITH_CRASH +MERGE_WITH_CRASH: MERGE-OUTER: succesfull in 2 attempt(s) +MERGE_WITH_CRASH: MERGE-OUTER: 3 new files + +# Check that we actually limit the size with max_len +RUN: LLVMFuzzer-FullCoverageSetTest -merge=1 %tmp/T1 %tmp/T2 -max_len=5 2>&1 | FileCheck %s --check-prefix=MERGE_LEN5 +MERGE_LEN5: MERGE-OUTER: succesfull in 1 attempt(s) diff --git a/test/thirdparty/Fuzzer/test/minimize_crash.test b/test/thirdparty/Fuzzer/test/minimize_crash.test new file mode 100644 index 00000000..7e540659 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/minimize_crash.test @@ -0,0 +1,6 @@ +RUN: echo 'Hi!rv349f34t3gg' > not_minimal_crash +RUN: LLVMFuzzer-NullDerefTest -minimize_crash=1 not_minimal_crash -max_total_time=2 2>&1 | FileCheck %s +CHECK: CRASH_MIN: failed to minimize beyond minimized-from-{{.*}} (3 bytes), exiting +RUN: LLVMFuzzer-NullDerefTest -minimize_crash=1 not_minimal_crash -max_total_time=2 -exact_artifact_path=exact_minimized_path 2>&1 | FileCheck %s --check-prefix=CHECK_EXACT +CHECK_EXACT: CRASH_MIN: failed to minimize beyond exact_minimized_path (3 bytes), exiting +RUN: rm not_minimal_crash minimized-from-* exact_minimized_path diff --git a/test/thirdparty/Fuzzer/test/no-coverage/CMakeLists.txt b/test/thirdparty/Fuzzer/test/no-coverage/CMakeLists.txt new file mode 100644 index 00000000..d2f6f438 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/no-coverage/CMakeLists.txt @@ -0,0 +1,29 @@ +# These tests are not instrumented with coverage, +# but have coverage rt in the binary. + +set(CMAKE_CXX_FLAGS + "${LIBFUZZER_FLAGS_BASE} -fno-sanitize-coverage=edge,trace-cmp,indirect-calls,8bit-counters,trace-pc-guard") + +set(NoCoverageTests + UninstrumentedTest + ) + +foreach(Test ${NoCoverageTests}) + add_libfuzzer_test(${Test}-NoCoverage SOURCES ../${Test}.cpp) +endforeach() + + +############################################################################### +# AFL Driver test +############################################################################### + +add_executable(AFLDriverTest + ../AFLDriverTest.cpp ../../afl/afl_driver.cpp) + +set_target_properties(AFLDriverTest + PROPERTIES RUNTIME_OUTPUT_DIRECTORY + "${CMAKE_BINARY_DIR}/lib/Fuzzer/test" + ) + +# Propagate value into parent directory +set(TestBinaries ${TestBinaries} AFLDriverTest PARENT_SCOPE) diff --git a/test/thirdparty/Fuzzer/test/repeated-bytes.test b/test/thirdparty/Fuzzer/test/repeated-bytes.test new file mode 100644 index 00000000..71394087 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/repeated-bytes.test @@ -0,0 +1,2 @@ +CHECK: BINGO +RUN: LLVMFuzzer-RepeatedBytesTest -seed=1 -runs=1000000 2>&1 | FileCheck %s diff --git a/test/thirdparty/Fuzzer/test/shrink.test b/test/thirdparty/Fuzzer/test/shrink.test new file mode 100644 index 00000000..edb86cb1 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/shrink.test @@ -0,0 +1,7 @@ +RUN: LLVMFuzzer-ShrinkControlFlowTest -seed=1 -exit_on_item=0eb8e4ed029b774d80f2b66408203801cb982a60 -runs=1000000 -shrink=1 2>&1 | FileCheck %s --check-prefix=SHRINK1 +RUN: LLVMFuzzer-ShrinkControlFlowTest -seed=1 -exit_on_item=0eb8e4ed029b774d80f2b66408203801cb982a60 -runs=1000000 -shrink=0 2>&1 | FileCheck %s --check-prefix=SHRINK0 +RUN: LLVMFuzzer-ShrinkValueProfileTest -seed=1 -exit_on_item=aea2e3923af219a8956f626558ef32f30a914ebc -runs=100000 -shrink=1 -use_value_profile=1 2>&1 | FileCheck %s --check-prefix=SHRINK1_VP + +SHRINK0: Done 1000000 runs in +SHRINK1: INFO: found item with checksum '0eb8e4ed029b774d80f2b66408203801cb982a60', exiting. +SHRINK1_VP: INFO: found item with checksum 'aea2e3923af219a8956f626558ef32f30a914ebc', exiting diff --git a/test/thirdparty/Fuzzer/test/simple-cmp.test b/test/thirdparty/Fuzzer/test/simple-cmp.test new file mode 100644 index 00000000..145a0366 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/simple-cmp.test @@ -0,0 +1,2 @@ +CHECK: BINGO +RUN: not LLVMFuzzer-SimpleCmpTest -seed=1 -runs=100000000 2>&1 | FileCheck %s diff --git a/test/thirdparty/Fuzzer/test/standalone.test b/test/thirdparty/Fuzzer/test/standalone.test new file mode 100644 index 00000000..3097b3ff --- /dev/null +++ b/test/thirdparty/Fuzzer/test/standalone.test @@ -0,0 +1,4 @@ +RUN: LLVMFuzzer-StandaloneInitializeTest %S/hi.txt %S/dict1.txt 2>&1 | FileCheck %s +CHECK: StandaloneFuzzTargetMain: running 2 inputs +CHECK: Done: {{.*}}hi.txt: (3 bytes) +CHECK: Done: {{.*}}dict1.txt: (61 bytes) diff --git a/test/thirdparty/Fuzzer/test/swap-cmp.test b/test/thirdparty/Fuzzer/test/swap-cmp.test new file mode 100644 index 00000000..908b7986 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/swap-cmp.test @@ -0,0 +1,2 @@ +CHECK: BINGO +RUN: not LLVMFuzzer-SwapCmpTest -seed=1 -runs=10000000 2>&1 | FileCheck %s diff --git a/test/thirdparty/Fuzzer/test/trace-malloc.test b/test/thirdparty/Fuzzer/test/trace-malloc.test new file mode 100644 index 00000000..c9514790 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/trace-malloc.test @@ -0,0 +1,10 @@ +RUN: LLVMFuzzer-TraceMallocTest -seed=1 -trace_malloc=1 -runs=10000 2>&1 | FileCheck %s +CHECK-DAG: MallocFreeTracer: STOP 0 0 (same) +CHECK-DAG: MallocFreeTracer: STOP 0 1 (DIFFERENT) +CHECK-DAG: MallocFreeTracer: STOP 1 0 (DIFFERENT) +CHECK-DAG: MallocFreeTracer: STOP 1 1 (same) + +RUN: LLVMFuzzer-TraceMallocTest -seed=1 -trace_malloc=2 -runs=1000 2>&1 | FileCheck %s --check-prefix=TRACE2 +TRACE2-DAG: FREE[0] +TRACE2-DAG: MALLOC[0] +TRACE2-DAG: in LLVMFuzzerTestOneInput diff --git a/test/thirdparty/Fuzzer/test/ubsan/CMakeLists.txt b/test/thirdparty/Fuzzer/test/ubsan/CMakeLists.txt new file mode 100644 index 00000000..7a9eacdb --- /dev/null +++ b/test/thirdparty/Fuzzer/test/ubsan/CMakeLists.txt @@ -0,0 +1,15 @@ +# These tests are instrumented with ubsan in non-recovery mode. + +set(CMAKE_CXX_FLAGS + "${LIBFUZZER_FLAGS_BASE} -fsanitize=undefined -fno-sanitize-recover=all") + +set(UbsanTests + SignedIntOverflowTest + ) + +foreach(Test ${UbsanTests}) + add_libfuzzer_test(${Test}-Ubsan SOURCES ../${Test}.cpp) +endforeach() + +# Propagate value into parent directory +set(TestBinaries ${TestBinaries} PARENT_SCOPE) diff --git a/test/thirdparty/Fuzzer/test/ulimit.test b/test/thirdparty/Fuzzer/test/ulimit.test new file mode 100644 index 00000000..a60636c3 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/ulimit.test @@ -0,0 +1,2 @@ +RUN: ulimit -s 1000 +RUN: LLVMFuzzer-SimpleTest diff --git a/test/thirdparty/Fuzzer/test/uninstrumented/CMakeLists.txt b/test/thirdparty/Fuzzer/test/uninstrumented/CMakeLists.txt new file mode 100644 index 00000000..29b66e6e --- /dev/null +++ b/test/thirdparty/Fuzzer/test/uninstrumented/CMakeLists.txt @@ -0,0 +1,16 @@ +# These tests are not instrumented with coverage and don't +# have coverage rt in the binary. + +set(CMAKE_CXX_FLAGS + "${LIBFUZZER_FLAGS_BASE} -fno-sanitize=all -fno-sanitize-coverage=edge,trace-cmp,indirect-calls,8bit-counters,trace-pc-guard") + +set(UninstrumentedTests + UninstrumentedTest + ) + +foreach(Test ${UninstrumentedTests}) + add_libfuzzer_test(${Test}-Uninstrumented SOURCES ../${Test}.cpp) +endforeach() + +# Propagate value into parent directory +set(TestBinaries ${TestBinaries} PARENT_SCOPE) diff --git a/test/thirdparty/Fuzzer/test/unit/lit.cfg b/test/thirdparty/Fuzzer/test/unit/lit.cfg new file mode 100644 index 00000000..0cc31939 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/unit/lit.cfg @@ -0,0 +1,7 @@ +import lit.formats + +config.name = "LLVMFuzzer-Unittest" +print config.test_exec_root +config.test_format = lit.formats.GoogleTest(".", "Unittest") +config.suffixes = [] +config.test_source_root = config.test_exec_root diff --git a/test/thirdparty/Fuzzer/test/unit/lit.site.cfg.in b/test/thirdparty/Fuzzer/test/unit/lit.site.cfg.in new file mode 100644 index 00000000..114daf47 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/unit/lit.site.cfg.in @@ -0,0 +1,2 @@ +config.test_exec_root = "@CMAKE_CURRENT_BINARY_DIR@" +lit_config.load_config(config, "@CMAKE_CURRENT_SOURCE_DIR@/unit/lit.cfg") diff --git a/test/thirdparty/Fuzzer/test/value-profile-cmp.test b/test/thirdparty/Fuzzer/test/value-profile-cmp.test new file mode 100644 index 00000000..48edba4f --- /dev/null +++ b/test/thirdparty/Fuzzer/test/value-profile-cmp.test @@ -0,0 +1,2 @@ +CHECK: BINGO +RUN: not LLVMFuzzer-SimpleCmpTest -seed=1 -use_cmp=0 -use_value_profile=1 -runs=100000000 2>&1 | FileCheck %s diff --git a/test/thirdparty/Fuzzer/test/value-profile-cmp2.test b/test/thirdparty/Fuzzer/test/value-profile-cmp2.test new file mode 100644 index 00000000..43d62400 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/value-profile-cmp2.test @@ -0,0 +1,2 @@ +CHECK: BINGO +RUN: not LLVMFuzzer-SimpleHashTest -seed=1 -use_cmp=0 -use_value_profile=1 -runs=100000000 2>&1 | FileCheck %s diff --git a/test/thirdparty/Fuzzer/test/value-profile-cmp3.test b/test/thirdparty/Fuzzer/test/value-profile-cmp3.test new file mode 100644 index 00000000..8a962763 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/value-profile-cmp3.test @@ -0,0 +1,2 @@ +CHECK: BINGO +RUN: not LLVMFuzzer-AbsNegAndConstantTest -seed=1 -use_cmp=0 -use_value_profile=1 -runs=100000000 2>&1 | FileCheck %s diff --git a/test/thirdparty/Fuzzer/test/value-profile-cmp4.test b/test/thirdparty/Fuzzer/test/value-profile-cmp4.test new file mode 100644 index 00000000..1e7131e5 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/value-profile-cmp4.test @@ -0,0 +1,2 @@ +CHECK: BINGO +RUN: not LLVMFuzzer-AbsNegAndConstant64Test -seed=1 -use_cmp=0 -use_value_profile=1 -runs=100000000 2>&1 | FileCheck %s diff --git a/test/thirdparty/Fuzzer/test/value-profile-div.test b/test/thirdparty/Fuzzer/test/value-profile-div.test new file mode 100644 index 00000000..ba45e412 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/value-profile-div.test @@ -0,0 +1,3 @@ +CHECK: AddressSanitizer: FPE +RUN: not LLVMFuzzer-DivTest -seed=1 -use_value_profile=1 -runs=10000000 2>&1 | FileCheck %s + diff --git a/test/thirdparty/Fuzzer/test/value-profile-load.test b/test/thirdparty/Fuzzer/test/value-profile-load.test new file mode 100644 index 00000000..14d3109a --- /dev/null +++ b/test/thirdparty/Fuzzer/test/value-profile-load.test @@ -0,0 +1,3 @@ +CHECK: AddressSanitizer: global-buffer-overflow +RUN: not LLVMFuzzer-LoadTest -seed=1 -use_cmp=0 -use_value_profile=1 -runs=10000000 2>&1 | FileCheck %s + diff --git a/test/thirdparty/Fuzzer/test/value-profile-mem.test b/test/thirdparty/Fuzzer/test/value-profile-mem.test new file mode 100644 index 00000000..09d737db --- /dev/null +++ b/test/thirdparty/Fuzzer/test/value-profile-mem.test @@ -0,0 +1,2 @@ +CHECK: BINGO +RUN: not LLVMFuzzer-SingleMemcmpTest -seed=1 -use_cmp=0 -use_memcmp=0 -use_value_profile=1 -runs=10000000 2>&1 | FileCheck %s diff --git a/test/thirdparty/Fuzzer/test/value-profile-set.test b/test/thirdparty/Fuzzer/test/value-profile-set.test new file mode 100644 index 00000000..9d06c365 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/value-profile-set.test @@ -0,0 +1,3 @@ +CHECK: BINGO +RUN: not LLVMFuzzer-FourIndependentBranchesTest -seed=1 -use_cmp=0 -use_value_profile=1 -runs=100000000 2>&1 | FileCheck %s + diff --git a/test/thirdparty/Fuzzer/test/value-profile-strcmp.test b/test/thirdparty/Fuzzer/test/value-profile-strcmp.test new file mode 100644 index 00000000..1e7ef9b4 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/value-profile-strcmp.test @@ -0,0 +1,2 @@ +CHECK: BINGO +RUN: not LLVMFuzzer-SingleStrcmpTest -seed=1 -use_cmp=0 -use_memcmp=0 -use_value_profile=1 -runs=10000000 2>&1 | FileCheck %s diff --git a/test/thirdparty/Fuzzer/test/value-profile-strncmp.test b/test/thirdparty/Fuzzer/test/value-profile-strncmp.test new file mode 100644 index 00000000..65097318 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/value-profile-strncmp.test @@ -0,0 +1,2 @@ +CHECK: BINGO +RUN: not LLVMFuzzer-SingleStrncmpTest -seed=1 -use_cmp=0 -use_memcmp=0 -use_value_profile=1 -runs=10000000 2>&1 | FileCheck %s diff --git a/test/thirdparty/Fuzzer/test/value-profile-switch.test b/test/thirdparty/Fuzzer/test/value-profile-switch.test new file mode 100644 index 00000000..1947f568 --- /dev/null +++ b/test/thirdparty/Fuzzer/test/value-profile-switch.test @@ -0,0 +1,3 @@ +CHECK: BINGO +RUN: not LLVMFuzzer-SwitchTest -use_cmp=0 -use_value_profile=1 -runs=100000000 -seed=1 2>&1 | FileCheck %s +RUN: not LLVMFuzzer-Switch2Test -use_cmp=0 -use_value_profile=1 -runs=100000000 -seed=1 2>&1 | FileCheck %s diff --git a/test/thirdparty/catch/LICENSE_1_0.txt b/test/thirdparty/catch/LICENSE_1_0.txt new file mode 100755 index 00000000..36b7cd93 --- /dev/null +++ b/test/thirdparty/catch/LICENSE_1_0.txt @@ -0,0 +1,23 @@ +Boost Software License - Version 1.0 - August 17th, 2003 + +Permission is hereby granted, free of charge, to any person or organization +obtaining a copy of the software and accompanying documentation covered by +this license (the "Software") to use, reproduce, display, distribute, +execute, and transmit the Software, and to prepare derivative works of the +Software, and to permit third-parties to whom the Software is furnished to +do so, all subject to the following: + +The copyright notices in the Software and this entire statement, including +the above license grant, this restriction and the following disclaimer, +must be included in all copies of the Software, in whole or in part, and +all derivative works of the Software, unless such copies or derivative +works are solely in the form of machine-executable object code generated by +a source language processor. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT. IN NO EVENT +SHALL THE COPYRIGHT HOLDERS OR ANYONE DISTRIBUTING THE SOFTWARE BE LIABLE +FOR ANY DAMAGES OR OTHER LIABILITY, WHETHER IN CONTRACT, TORT OR OTHERWISE, +ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER +DEALINGS IN THE SOFTWARE. diff --git a/test/src/catch.hpp b/test/thirdparty/catch/catch.hpp old mode 100644 new mode 100755 similarity index 99% rename from test/src/catch.hpp rename to test/thirdparty/catch/catch.hpp index 2e6fe8d3..3d18eadb --- a/test/src/catch.hpp +++ b/test/thirdparty/catch/catch.hpp @@ -1,6 +1,6 @@ /* - * Catch v1.5.8 - * Generated: 2016-10-26 12:07:30.938259 + * Catch v1.5.9 + * Generated: 2016-11-29 12:14:38.049276 * ---------------------------------------------------------- * This file has been merged from multiple headers. Please don't edit it directly * Copyright (c) 2012 Two Blue Cubes Ltd. All rights reserved. @@ -3428,6 +3428,7 @@ namespace Catch { #include #include #include +#include namespace Catch { @@ -3995,9 +3996,12 @@ namespace Clara { inline void convertInto( std::string const& _source, std::string& _dest ) { _dest = _source; } + char toLowerCh(char c) { + return static_cast( ::tolower( c ) ); + } inline void convertInto( std::string const& _source, bool& _dest ) { std::string sourceLC = _source; - std::transform( sourceLC.begin(), sourceLC.end(), sourceLC.begin(), ::tolower ); + std::transform( sourceLC.begin(), sourceLC.end(), sourceLC.begin(), toLowerCh ); if( sourceLC == "y" || sourceLC == "1" || sourceLC == "true" || sourceLC == "yes" || sourceLC == "on" ) _dest = true; else if( sourceLC == "n" || sourceLC == "0" || sourceLC == "false" || sourceLC == "no" || sourceLC == "off" ) @@ -7578,7 +7582,7 @@ namespace Catch { return os; } - Version libraryVersion( 1, 5, 8, "", 0 ); + Version libraryVersion( 1, 5, 9, "", 0 ); } @@ -9166,6 +9170,7 @@ namespace Catch { public: XmlReporter( ReporterConfig const& _config ) : StreamingReporterBase( _config ), + m_xml(_config.stream()), m_sectionDepth( 0 ) { m_reporterPrefs.shouldRedirectStdOut = true; @@ -9185,7 +9190,6 @@ namespace Catch { virtual void testRunStarting( TestRunInfo const& testInfo ) CATCH_OVERRIDE { StreamingReporterBase::testRunStarting( testInfo ); - m_xml.setStream( stream ); m_xml.startElement( "Catch" ); if( !m_config->name().empty() ) m_xml.writeAttribute( "name", m_config->name() );