| /* | |
| * Copyright 2016 Google Inc. | |
| * | |
| * Use of this source code is governed by a BSD-style license that can be | |
| * found in the LICENSE file. | |
| */ | |
| #ifndef Fuzz_DEFINED | |
| #define Fuzz_DEFINED | |
| #include "include/core/SkData.h" | |
| #include "include/core/SkImageFilter.h" | |
| #include "include/core/SkRegion.h" | |
| #include "include/core/SkTypes.h" | |
| #include "include/private/base/SkMalloc.h" | |
| #include "include/private/base/SkTFitsIn.h" | |
| #include "tools/Registry.h" | |
| #include <limits> | |
| #include <cmath> | |
| #include <signal.h> | |
| #include <limits> | |
| class Fuzz { | |
| public: | |
| explicit Fuzz(const uint8_t* data, size_t size) : fData(data), fSize(size), fNextByte(0) {} | |
| Fuzz() = delete; | |
| // Make noncopyable | |
| Fuzz(Fuzz&) = delete; | |
| Fuzz& operator=(Fuzz&) = delete; | |
| // Returns the total number of "random" bytes available. | |
| size_t size() const { | |
| return fSize; | |
| } | |
| // Returns if there are no bytes remaining for fuzzing. | |
| bool exhausted() const { | |
| return fSize == fNextByte; | |
| } | |
| void deplete() { | |
| fNextByte = fSize; | |
| } | |
| size_t remainingSize() const { | |
| return fSize - fNextByte; | |
| } | |
| const uint8_t *remainingData() const { | |
| return fData + fNextByte; | |
| } | |
| // next() loads fuzzed bytes into the variable passed in by pointer. | |
| // We use this approach instead of T next() because different compilers | |
| // evaluate function parameters in different orders. If fuzz->next() | |
| // returned 5 and then 7, foo(fuzz->next(), fuzz->next()) would be | |
| // foo(5, 7) when compiled on GCC and foo(7, 5) when compiled on Clang. | |
| // By requiring params to be passed in, we avoid the temptation to call | |
| // next() in a way that does not consume fuzzed bytes in a single | |
| // platform-independent order. | |
| template <typename T> | |
| void next(T* t) { this->nextBytes(t, sizeof(T)); } | |
| // This is a convenient way to initialize more than one argument at a time. | |
| template <typename Arg, typename... Args> | |
| void next(Arg* first, Args... rest); | |
| // nextRange returns values only in [min, max]. | |
| template <typename T, typename Min, typename Max> | |
| void nextRange(T*, Min, Max); | |
| // nextEnum is a wrapper around nextRange for enums. | |
| template <typename T> | |
| void nextEnum(T* ptr, T max); | |
| // nextN loads n * sizeof(T) bytes into ptr | |
| template <typename T> | |
| void nextN(T* ptr, int n); | |
| void signalBug() { | |
| // Tell the fuzzer that these inputs found a bug. | |
| SkDebugf("Signal bug\n"); | |
| raise(SIGSEGV); | |
| } | |
| // Specialized versions for when true random doesn't quite make sense | |
| void next(bool* b); | |
| void next(SkRegion* region); | |
| bool nextBool() { | |
| bool b; | |
| this->next(&b); | |
| return b; | |
| } | |
| void nextRange(float* f, float min, float max); | |
| private: | |
| template <typename T> | |
| T nextT(); | |
| const uint8_t *fData; | |
| size_t fSize; | |
| size_t fNextByte; | |
| friend void fuzz__MakeEncoderCorpus(Fuzz*); | |
| void nextBytes(void* ptr, size_t size); | |
| }; | |
| template <typename Arg, typename... Args> | |
| inline void Fuzz::next(Arg* first, Args... rest) { | |
| this->next(first); | |
| this->next(rest...); | |
| } | |
| template <typename T, typename Min, typename Max> | |
| inline void Fuzz::nextRange(T* value, Min min, Max max) { | |
| // UBSAN worries if we make an enum with out of range values, even temporarily. | |
| using Raw = typename sk_strip_enum<T>::type; | |
| Raw raw; | |
| this->next(&raw); | |
| if (raw < (Raw)min) { raw = (Raw)min; } | |
| if (raw > (Raw)max) { raw = (Raw)max; } | |
| *value = (T)raw; | |
| } | |
| template <typename T> | |
| inline void Fuzz::nextEnum(T* value, T max) { | |
| // This works around the fact that UBSAN will assert if we put an invalid | |
| // value into an enum. We might see issues with enums being represented | |
| // on Windows differently than Linux, but that's not a thing we can fix here. | |
| using U = typename std::underlying_type<T>::type; | |
| U v; | |
| this->next(&v); | |
| if (v < (U)0) { *value = (T)0; return;} | |
| if (v > (U)max) { *value = (T)max; return;} | |
| *value = (T)v; | |
| } | |
| template <typename T> | |
| inline void Fuzz::nextN(T* ptr, int n) { | |
| for (int i = 0; i < n; i++) { | |
| this->next(ptr+i); | |
| } | |
| } | |
| struct Fuzzable { | |
| const char* name; | |
| void (*fn)(Fuzz*); | |
| }; | |
| // Not static so that we can link these into oss-fuzz harnesses if we like. | |
| #define DEF_FUZZ(name, f) \ | |
| void fuzz_##name(Fuzz*); \ | |
| sk_tools::Registry<Fuzzable> register_##name({#name, fuzz_##name}); \ | |
| void fuzz_##name(Fuzz* f) | |
| #endif // Fuzz_DEFINED |