Merge pull request #53 from wheremyfoodat/GamingProcessingUnit

GPU stuff (Fixed TEV, optional shader->x64 recompilation, vertex cache and more)

.gitignore

@@ -8,6 +8,7 @@
 
 #Clion Files
 .idea/
+cmake-build-*/
 
 Debug/
 Release/

.gitmodules

@@ -13,3 +13,6 @@
 [submodule "third_party/cryptopp/cryptopp"]
 	path = third_party/cryptopp/cryptopp
 	url = https://github.com/weidai11/cryptopp
+[submodule "third_party/xbyak"]
+	path = third_party/xbyak
+	url = https://github.com/herumi/xbyak

CMakeLists.txt

@@ -14,6 +14,7 @@ project(Alber)
 set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR})
 
 option(GPU_DEBUG_INFO "Enable additional GPU debugging info" OFF)
+option(ENABLE_LTO "Enable link-time optimization" OFF)
 
 include_directories(${PROJECT_SOURCE_DIR}/include/)
 include_directories(${PROJECT_SOURCE_DIR}/include/kernel)
@@ -23,7 +24,9 @@ include_directories(third_party/elfio/)
 include_directories(third_party/imgui/)
 include_directories(third_party/dynarmic/src)
 include_directories(third_party/cryptopp/)
-include_directories(third_party/result/include/)
+include_directories(third_party/cityhash/include)
+include_directories(third_party/result/include)
+include_directories(third_party/xxhash/include)
 
 add_compile_definitions(NOMINMAX)
 add_compile_definitions(SDL_MAIN_HANDLED)
@@ -48,6 +51,10 @@ add_subdirectory(third_party/cryptopp)
 # Check for x64
 if (CMAKE_SYSTEM_PROCESSOR STREQUAL "AMD64" OR CMAKE_SYSTEM_PROCESSOR STREQUAL "x86-64" OR CMAKE_SYSTEM_PROCESSOR STREQUAL "x86_64")
     set(HOST_X64 TRUE)
+    add_subdirectory(third_party/xbyak) # Add xbyak submodule for x86 JITs
+    include_directories(third_party/xbyak)
+    add_compile_definitions(PANDA3DS_DYNAPICA_SUPPORTED)
+    add_compile_definitions(PANDA3DS_X64_HOST)
 else()
     set(HOST_X64 FALSE)
 endif()
@@ -55,6 +62,7 @@ endif()
 # Check for arm64
 if (CMAKE_SYSTEM_PROCESSOR STREQUAL "aarch64" OR CMAKE_SYSTEM_PROCESSOR STREQUAL "arm64")
     set(HOST_ARM64 TRUE)
+    add_compile_definitions(PANDA3DS_ARM64_HOST)
 else()
     set(HOST_ARM64 FALSE)
 endif()
@@ -90,8 +98,10 @@ set(SERVICE_SOURCE_FILES src/core/services/service_manager.cpp src/core/services
                          src/core/services/act.cpp src/core/services/nfc.cpp src/core/services/dlp_srvr.cpp
 )
 set(PICA_SOURCE_FILES src/core/PICA/gpu.cpp src/core/PICA/regs.cpp src/core/PICA/shader_unit.cpp
-                      src/core/PICA/shader_interpreter.cpp
+                      src/core/PICA/shader_interpreter.cpp src/core/PICA/dynapica/shader_rec.cpp
+                      src/core/PICA/dynapica/shader_rec_emitter_x64.cpp src/core/PICA/pica_hash.cpp
 )
+
 set(RENDERER_GL_SOURCE_FILES src/core/renderer_gl/renderer_gl.cpp src/core/renderer_gl/textures.cpp src/core/renderer_gl/etc1.cpp)
 
 set(LOADER_SOURCE_FILES src/core/loader/elf.cpp src/core/loader/ncsd.cpp src/core/loader/ncch.cpp src/core/loader/lz77.cpp)
@@ -117,7 +127,9 @@ set(HEADER_FILES include/emulator.hpp include/helpers.hpp include/opengl.hpp inc
                  include/fs/archive_ext_save_data.hpp include/services/shared_font.hpp include/fs/archive_ncch.hpp
                  include/renderer_gl/textures.hpp include/colour.hpp include/services/y2r.hpp include/services/cam.hpp
                  include/services/ldr_ro.hpp include/ipc.hpp include/services/act.hpp include/services/nfc.hpp
-                 include/system_models.hpp include/services/dlp_srvr.hpp include/result/result.hpp
+                 include/system_models.hpp include/services/dlp_srvr.hpp include/PICA/dynapica/pica_recs.hpp
+                 include/PICA/dynapica/x64_regs.hpp include/PICA/dynapica/vertex_loader_rec.hpp include/PICA/dynapica/shader_rec.hpp
+                 include/PICA/dynapica/shader_rec_emitter_x64.hpp include/PICA/pica_hash.hpp include/result/result.hpp
                  include/result/result_common.hpp include/result/result_fs.hpp include/result/result_fnd.hpp
                  include/result/result_gsp.hpp include/result/result_kernel.hpp include/result/result_os.hpp
                  include/crypto/aes_engine.hpp include/metaprogramming.hpp
@@ -128,6 +140,9 @@ set(THIRD_PARTY_SOURCE_FILES third_party/imgui/imgui.cpp
                              third_party/imgui/imgui_tables.cpp
                              third_party/imgui/imgui_widgets.cpp
                              third_party/imgui/imgui_demo.cpp
+
+                             third_party/cityhash/cityhash.cpp
+                             third_party/xxhash/xxhash.c
 )
 
 source_group("Header Files\\Core" FILES ${HEADER_FILES})
@@ -143,9 +158,13 @@ source_group("Source Files\\Third Party" FILES ${THIRD_PARTY_SOURCE_FILES})
 
 add_executable(Alber ${SOURCE_FILES} ${FS_SOURCE_FILES} ${CRYPTO_SOURCE_FILES} ${KERNEL_SOURCE_FILES} ${LOADER_SOURCE_FILES} ${SERVICE_SOURCE_FILES}
 ${PICA_SOURCE_FILES} ${RENDERER_GL_SOURCE_FILES} ${THIRD_PARTY_SOURCE_FILES} ${HEADER_FILES})
-target_link_libraries(Alber PRIVATE dynarmic SDL2-static glad cryptopp)
 
+if(ENABLE_LTO)
+  set_target_properties(Alber PROPERTIES INTERPROCEDURAL_OPTIMIZATION TRUE)
+endif()
+
+target_link_libraries(Alber PRIVATE dynarmic SDL2-static glad cryptopp)
 
 if(GPU_DEBUG_INFO)
   target_compile_definitions(Alber PRIVATE GPU_DEBUG_INFO=1)
-endif()
+endif()

include/PICA/dynapica/pica_recs.hpp

@@ -0,0 +1,13 @@
+#pragma once
+#include "helpers.hpp"
+#include "vertex_loader_rec.hpp"
+
+// Common file for our PICA JITs (From vertex config -> CPU assembly and from PICA shader -> CPU assembly)
+
+namespace Dynapica {
+#ifdef PANDA3DS_DYNAPICA_SUPPORTED
+	static constexpr bool supported() { return true; }
+#else
+	static constexpr bool supported() { return false; }
+#endif
+}

include/PICA/dynapica/shader_rec.hpp

@@ -0,0 +1,53 @@
+#pragma once
+#include "PICA/shader.hpp"
+
+#if defined(PANDA3DS_DYNAPICA_SUPPORTED) && defined(PANDA3DS_X64_HOST)
+#define PANDA3DS_SHADER_JIT_SUPPORTED
+#include <memory>
+#include <unordered_map>
+
+#ifdef PANDA3DS_X64_HOST
+#include "shader_rec_emitter_x64.hpp"
+#endif
+#endif
+
+class ShaderJIT {
+#ifdef PANDA3DS_SHADER_JIT_SUPPORTED
+	using Hash = PICAShader::Hash;
+	using ShaderCache = std::unordered_map<Hash, std::unique_ptr<ShaderEmitter>>;
+	ShaderEmitter::PrologueCallback prologueCallback;
+	ShaderEmitter::InstructionCallback entrypointCallback;
+
+	ShaderCache cache;
+#endif
+
+public:
+#ifdef PANDA3DS_SHADER_JIT_SUPPORTED
+	// Call this before starting to process a batch of vertices
+	// This will read the PICA config (uploaded shader and shader operand descriptors) and search if we've already compiled this shader
+	// If yes, it sets it as the active shader. if not, then it compiles it, adds it to the cache, and sets it as active,
+	// The caller must make sure the entrypoint has been properly set beforehand
+	void prepare(PICAShader& shaderUnit);
+	void reset();
+	void run(PICAShader& shaderUnit) {
+		prologueCallback(shaderUnit, entrypointCallback);
+	}
+
+	static constexpr bool isAvailable() { return true; }
+#else
+	void prepare(PICAShader& shaderUnit) {
+		Helpers::panic("Vertex Loader JIT: Tried to run ShaderJIT::Prepare on platform that does not support shader jit");
+	}
+
+	void run(PICAShader& shaderUnit) {
+		Helpers::panic("Vertex Loader JIT: Tried to run ShaderJIT::Run on platform that does not support shader jit");
+	}
+
+	// Define dummy callback. This should never be called if the shader JIT is not supported
+	using Callback = void(*)(PICAShader& shaderUnit);
+	Callback activeShaderCallback = nullptr;
+
+	void reset() {}
+	static constexpr bool isAvailable() { return false; }
+#endif
+};

include/PICA/dynapica/shader_rec_emitter_x64.hpp

@@ -0,0 +1,139 @@
+#pragma once
+
+// Only do anything if we're on an x64 target with JIT support enabled
+#if defined(PANDA3DS_DYNAPICA_SUPPORTED) && defined(PANDA3DS_X64_HOST)
+#include "helpers.hpp"
+#include "logger.hpp"
+#include "PICA/shader.hpp"
+#include "xbyak/xbyak.h"
+#include "xbyak/xbyak_util.h"
+#include "x64_regs.hpp"
+
+#include <vector>
+
+class ShaderEmitter : public Xbyak::CodeGenerator {
+	static constexpr size_t executableMemorySize = PICAShader::maxInstructionCount * 96; // How much executable memory to alloc for each shader
+	// Allocate some extra space as padding for security purposes in the extremely unlikely occasion we manage to overflow the above size
+	static constexpr size_t allocSize = executableMemorySize + 0x1000;
+
+	// If the swizzle field is this value then the swizzle pattern is .xyzw so we don't need a shuffle
+	static constexpr uint noSwizzle = 0x1B;
+
+	using f24 = Floats::f24;
+	using vec4f = OpenGL::Vector<f24, 4>;
+
+	// An array of labels (incl pointers) to each compiled (to x64) PICA instruction
+	std::array<Xbyak::Label, PICAShader::maxInstructionCount> instructionLabels;
+	// A vector of PCs that can potentially return based on the state of the PICA callstack.
+	// Filled before compiling a shader by scanning the code for call instructions
+	std::vector<u32> returnPCs;
+
+	// Vector value of (-0.0, -0.0, -0.0, -0.0) for negating vectors via pxor
+	Label negateVector;
+	// Vector value of (1.0, 1.0, 1.0, 1.0) for SLT(i)/SGE(i)
+	Label onesVector;
+
+	u32 recompilerPC = 0; // PC the recompiler is currently recompiling @
+	u32 loopLevel = 0;    // The current loop nesting level (0 = not in a loop)
+
+	bool haveSSE4_1 = false;  // Shows if the CPU supports SSE4.1
+	bool haveAVX = false;     // Shows if the CPU supports AVX (NOT AVX2, NOT AVX512. Regular AVX)
+	bool haveFMA3 = false;    // Shows if the CPU supports FMA3
+
+	// Compile all instructions from [current recompiler PC, end)
+	void compileUntil(const PICAShader& shaderUnit, u32 endPC);
+	// Compile instruction "instr"
+	void compileInstruction(const PICAShader& shaderUnit);
+
+	bool isCall(u32 instruction) {
+		const u32 opcode = instruction >> 26;
+		return (opcode == ShaderOpcodes::CALL) || (opcode == ShaderOpcodes::CALLC) || (opcode == ShaderOpcodes::CALLU);
+	}
+	// Scan the shader code for call instructions to fill up the returnPCs vector before starting compilation
+	void scanForCalls(const PICAShader& shaderUnit);
+
+	// Load register with number "srcReg" indexed by index "idx" into the xmm register "reg"
+	template <int sourceIndex>
+	void loadRegister(Xmm dest, const PICAShader& shader, u32 src, u32 idx, u32 operandDescriptor);
+	void storeRegister(Xmm source, const PICAShader& shader, u32 dest, u32 operandDescriptor);
+
+	const vec4f& getSourceRef(const PICAShader& shader, u32 src);
+	const vec4f& getDestRef(const PICAShader& shader, u32 dest);
+
+	// Check the value of the cmp register for instructions like ifc and callc
+	// Result is returned in the zero flag. If the comparison is true then zero == 1, else zero == 0
+	void checkCmpRegister(const PICAShader& shader, u32 instruction);
+
+	// Check the value of the bool uniform for instructions like ifu and callu
+	// Result is returned in the zero flag. If the comparison is true then zero == 0, else zero == 1 (Opposite of checkCmpRegister)
+	void checkBoolUniform(const PICAShader& shader, u32 instruction);
+
+	// Prints a log. This is not meant to be used outside of debugging so it is very slow with our internal ABI.
+	void emitPrintLog(const PICAShader& shaderUnit);
+	static void printLog(const PICAShader& shaderUnit);
+
+	// Instruction recompilation functions
+	void recADD(const PICAShader& shader, u32 instruction);
+	void recCALL(const PICAShader& shader, u32 instruction);
+	void recCALLC(const PICAShader& shader, u32 instruction);
+	void recCALLU(const PICAShader& shader, u32 instruction);
+	void recCMP(const PICAShader& shader, u32 instruction);
+	void recDP3(const PICAShader& shader, u32 instruction);
+	void recDP4(const PICAShader& shader, u32 instruction);
+	void recEMIT(const PICAShader& shader, u32 instruction);
+	void recEND(const PICAShader& shader, u32 instruction);
+	void recFLR(const PICAShader& shader, u32 instruction);
+	void recIFC(const PICAShader& shader, u32 instruction);
+	void recIFU(const PICAShader& shader, u32 instruction);
+	void recJMPC(const PICAShader& shader, u32 instruction);
+	void recJMPU(const PICAShader& shader, u32 instruction);
+	void recLOOP(const PICAShader& shader, u32 instruction);
+	void recMAD(const PICAShader& shader, u32 instruction);
+	void recMAX(const PICAShader& shader, u32 instruction);
+	void recMIN(const PICAShader& shader, u32 instruction);
+	void recMOVA(const PICAShader& shader, u32 instruction);
+	void recMOV(const PICAShader& shader, u32 instruction);
+	void recMUL(const PICAShader& shader, u32 instruction);
+	void recRCP(const PICAShader& shader, u32 instruction);
+	void recRSQ(const PICAShader& shader, u32 instruction);
+	void recSETEMIT(const PICAShader& shader, u32 instruction);
+	void recSGE(const PICAShader& shader, u32 instruction);
+	void recSGEI(const PICAShader& shader, u32 instruction);
+	void recSLT(const PICAShader& shader, u32 instruction);
+
+	MAKE_LOG_FUNCTION(log, shaderJITLogger)
+
+public:
+	using InstructionCallback = const void(*)(PICAShader& shaderUnit); // Callback type used for instructions
+	// Callback type used for the JIT prologue. This is what the caller will call
+	using PrologueCallback = const void(*)(PICAShader& shaderUnit, InstructionCallback cb);
+	PrologueCallback prologueCb = nullptr;
+
+	// Initialize our emitter with "allocSize" bytes of RWX memory
+	ShaderEmitter() : Xbyak::CodeGenerator(allocSize) {
+		const auto cpu = Xbyak::util::Cpu();
+
+		haveSSE4_1 = cpu.has(Xbyak::util::Cpu::tSSE41);
+		haveAVX = cpu.has(Xbyak::util::Cpu::tAVX);
+		haveFMA3 = cpu.has(Xbyak::util::Cpu::tFMA);
+
+		if (!cpu.has(Xbyak::util::Cpu::tSSE3)) {
+			Helpers::panic("This CPU does not support SSE3. Please use the shader interpreter instead");
+		}
+	}
+	
+	void compile(const PICAShader& shaderUnit);
+
+	// PC must be a valid entrypoint here. It doesn't have that much overhead in this case, so we use std::array<>::at() to assert it does
+	InstructionCallback getInstructionCallback(u32 pc) {
+		// Cast away the constness because casting to a function pointer is hard otherwise. Legal as long as we don't write to *ptr
+		uint8_t* ptr = const_cast<uint8_t*>(instructionLabels.at(pc).getAddress());
+		return reinterpret_cast<InstructionCallback>(ptr);
+	}
+
+	PrologueCallback getPrologueCallback() {
+		return prologueCb;
+	}
+};
+
+#endif // x64 recompiler check

include/PICA/dynapica/vertex_loader_rec.hpp

@@ -0,0 +1,28 @@
+#pragma once
+#include "helpers.hpp"
+#include "renderer_gl/renderer_gl.hpp"
+#include "x64_regs.hpp"
+
+// Recompiler that takes the current vertex attribute configuration, ie the format of vertices (VAO in OpenGL) and emits optimized
+// code in our CPU's native architecture for loading vertices. Unimplemented at the moment, we've only got a skeleton of it here for later
+
+class VertexLoaderJIT {
+	using PICARegs = const std::array<u32, 0x300>&;
+	using Callback = void(*)(Vertex* output, size_t count);  // A function pointer to JIT-emitted code
+	Callback compileConfig(PICARegs regs);
+
+public:
+#if defined(PANDA3DS_DYNAPICA_SUPPORTED) && defined(PANDA3DS_X64_HOST)
+	#define PANDA3DS_VERTEX_LOADER_JIT_SUPPORTED
+
+	void loadVertices(Vertex* output, size_t count, PICARegs regs);
+	static constexpr bool isAvailable() { return true; }
+
+#else
+	void loadVertices(Vertex* output, size_t count, PICARegs regs) {
+		Helpers::panic("Vertex Loader JIT: Tried to load vertices with JIT on platform that does not support vertex loader jit");
+	}
+
+	static constexpr bool isAvailable() { return false; }
+#endif
+};

include/PICA/dynapica/x64_regs.hpp

@@ -0,0 +1,41 @@
+#pragma once
+
+#ifdef PANDA3DS_X64_HOST
+#include "xbyak/xbyak.h"
+using namespace Xbyak;
+using namespace Xbyak::util;
+
+#if defined(WIN32) || defined(_WIN32) || defined(__WIN32) && !defined(__CYGWIN__)
+#define PANDA3DS_MS_ABI
+constexpr Reg32 arg1 = ecx;  // register where first arg is stored
+constexpr Reg32 arg2 = edx;  // register where second arg is stored
+constexpr Reg32 arg3 = r8d;  // register where third arg is stored
+constexpr Reg32 arg4 = r9d;  // register where fourth arg is stored
+
+// Similar for floating point and vector arguemnts.
+constexpr Xmm arg1f = xmm0;
+constexpr Xmm arg2f = xmm1;
+constexpr Xmm arg3f = xmm2;
+constexpr Xmm arg4f = xmm3;
+
+constexpr bool isWindows() { return true; }
+
+#else  // System V calling convention
+#define PANDA3DS_SYSV_ABI
+constexpr Reg32 arg1 = edi;
+constexpr Reg32 arg2 = esi;
+constexpr Reg32 arg3 = edx;
+constexpr Reg32 arg4 = ecx;
+
+constexpr Xmm arg1f = xmm0;
+constexpr Xmm arg2f = xmm1;
+constexpr Xmm arg3f = xmm2;
+constexpr Xmm arg4f = xmm3;
+constexpr Xmm arg5f = xmm4;
+constexpr Xmm arg6f = xmm5;
+constexpr Xmm arg7f = xmm6;
+constexpr Xmm arg8f = xmm7;
+
+constexpr bool isWindows() { return false; }
+#endif
+#endif  // PANDA3DS_X64_HOST

include/PICA/gpu.hpp

@@ -6,6 +6,7 @@
 #include "PICA/float_types.hpp"
 #include "PICA/regs.hpp"
 #include "PICA/shader_unit.hpp"
+#include "PICA/dynapica/shader_rec.hpp"
 #include "renderer_gl/renderer_gl.hpp"
 
 class GPU {
@@ -15,6 +16,8 @@ class GPU {
 
 	Memory& mem;
 	ShaderUnit shaderUnit;
+	ShaderJIT shaderJIT; // Doesn't do anything if JIT is disabled or not supported
+
 	u8* vram = nullptr;
 	MAKE_LOG_FUNCTION(log, gpuLogger)
 
@@ -28,7 +31,7 @@ class GPU {
 	uint immediateModeVertIndex;
 	uint immediateModeAttrIndex; // Index of the immediate mode attribute we're uploading
 
-	template <bool indexed>
+	template <bool indexed, bool useShaderJIT>
 	void drawArrays();
 
 	// Silly method of avoiding linking problems. TODO: Change to something less silly

include/PICA/pica_hash.hpp

@@ -0,0 +1,20 @@
+#pragma once
+#include <cstdint>
+#include <cstddef>
+
+// Defines to pick which hash algorithm to use for the PICA (For hashing shaders, etc)
+// Please only define one of them
+// Available algorithms:
+// xxh3: 64-bit non-cryptographic hash using SIMD, default.
+// Google CityHash64: 64-bit non-cryptographic hash, generated using regular 64-bit arithmetic
+
+//#define PANDA3DS_PICA_CITYHASH
+#define PANDA3DS_PICA_XXHASH3
+
+namespace PICAHash {
+	#if defined(PANDA3DS_PICA_CITYHASH) || defined(PANDA3DS_PICA_XXHASH3)
+	using HashType = std::uint64_t;
+	#endif
+
+	HashType computeHash(const char* buf, std::size_t len);
+} // namespace PICAHash 

include/PICA/regs.hpp

@@ -13,9 +13,17 @@ namespace PICA {
 			DepthScale = 0x4D,
 			DepthOffset = 0x4E,
 			ShaderOutputCount = 0x4F,
+			ShaderOutmap0 = 0x50,
 
 			DepthmapEnable = 0x6D,
+
+			// Texture registers
 			TexUnitCfg = 0x80,
+			Tex0BorderColor = 0x81,
+			Tex1BorderColor = 0x91,
+			Tex2BorderColor = 0x99,
+			TexEnvUpdateBuffer = 0xE0,
+			TexEnvBufferColor = 0xFD,
 
 			// clang-format off
 			#define defineTexEnv(index, offset)   \

include/PICA/shader.hpp

@@ -5,6 +5,7 @@
 #include "helpers.hpp"
 #include "opengl.hpp"
 #include "PICA/float_types.hpp"
+#include "PICA/pica_hash.hpp"
 
 enum class ShaderType {
 	Vertex, Geometry
@@ -15,6 +16,11 @@ namespace ShaderOpcodes {
 		ADD = 0x00,
 		DP3 = 0x01,
 		DP4 = 0x02,
+		DPH = 0x03,
+		DST = 0x04,
+		EX2 = 0x05,
+		LG2 = 0x06,
+		LIT = 0x07,
 		MUL = 0x08,
 		SLT = 0x0A,
 		FLR = 0x0B,
@@ -24,6 +30,8 @@ namespace ShaderOpcodes {
 		RSQ = 0x0F,
 		MOVA = 0x12,
 		MOV = 0x13,
+		DPHI = 0x18,
+		DSTI = 0x19,
 		SGEI = 0x1A,
 		SLTI = 0x1B,
 		NOP = 0x21,
@@ -34,6 +42,8 @@ namespace ShaderOpcodes {
 		IFU = 0x27,
 		IFC = 0x28,
 		LOOP = 0x29,
+		EMIT = 0x2A,
+		SETEMIT = 0x2B,
 		JMPC = 0x2C,
 		JMPU = 0x2D,
 		CMP1 = 0x2E, // Both of these instructions are CMP
@@ -42,6 +52,7 @@ namespace ShaderOpcodes {
 	};
 }
 
+// Note: All PICA f24 vec4 registers must have the alignas(16) specifier to make them easier to access in SSE/NEON code in the JIT
 class PICAShader {
 	using f24 = Floats::f24;
 	using vec4f = OpenGL::Vector<f24, 4>;
@@ -71,8 +82,22 @@ class PICAShader {
 	bool f32UniformTransfer = false; // Are we transferring an f32 uniform or an f24 uniform?
 
 	std::array<u32, 4> floatUniformBuffer; // Buffer for temporarily caching float uniform data
+
+public:
+	// These are placed close to the temp registers and co because it helps the JIT generate better code
+	u32 entrypoint = 0; // Initial shader PC
+	u32 boolUniform;
+	std::array<OpenGL::Vector<u8, 4>, 4> intUniforms;
+	alignas(16) std::array<vec4f, 96> floatUniforms;
+
+	alignas(16) std::array<vec4f, 16> fixedAttributes; // Fixed vertex attributes
+	alignas(16) std::array<vec4f, 16> inputs; // Attributes passed to the shader
+	alignas(16) std::array<vec4f, 16> outputs;
+	alignas(16) vec4f dummy = vec4f({ f24::zero(), f24::zero(), f24::zero(), f24::zero() }); // Dummy register used by the JIT
+
+protected:
 	std::array<u32, 128> operandDescriptors;
-	std::array<vec4f, 16> tempRegisters; // General purpose registers the shader can use for temp values
+	alignas(16) std::array<vec4f, 16> tempRegisters; // General purpose registers the shader can use for temp values
 	OpenGL::Vector<s32, 2> addrRegister; // Address register
 	bool cmpRegister[2]; // Comparison registers where the result of CMP is stored in
 	u32 loopCounter;
@@ -85,13 +110,28 @@ class PICAShader {
 	std::array<Loop, 4> loopInfo;
 	std::array<ConditionalInfo, 8> conditionalInfo;
 	std::array<CallInfo, 4> callInfo;
-
 	ShaderType type;
 
+	// We use a hashmap for matching 3DS shaders to their equivalent compiled code in our shader cache in the shader JIT
+	// We choose our hash type to be a 64-bit integer by default, as the collision chance is very tiny and generating it is decently optimal
+	// Ideally we want to be able to support multiple different types of hash depending on compilation settings, but let's get this working first
+	using Hash = PICAHash::HashType;
+
+	Hash lastCodeHash = 0; // Last hash computed for the shader code (Used for the JIT caching mechanism)
+	Hash lastOpdescHash = 0;  // Last hash computed for the operand descriptors (Also used for the JIT)
+
+	bool codeHashDirty = false;
+	bool opdescHashDirty = false;
+
+	// Add these as friend classes for the JIT so it has access to all important state
+	friend class ShaderJIT;
+	friend class ShaderEmitter;
+
 	vec4f getSource(u32 source);
 	vec4f& getDest(u32 dest);
 
-	// Shader opcodes
+private:
+	// Interpreter functions for the various shader functions
 	void add(u32 instruction);
 	void call(u32 instruction);
 	void callc(u32 instruction);
@@ -99,11 +139,14 @@ class PICAShader {
 	void cmp(u32 instruction);
 	void dp3(u32 instruction);
 	void dp4(u32 instruction);
+	void dphi(u32 instruction);
+	void ex2(u32 instruction);
 	void flr(u32 instruction);
 	void ifc(u32 instruction);
 	void ifu(u32 instruction);
 	void jmpc(u32 instruction);
 	void jmpu(u32 instruction);
+	void lg2(u32 instruction);
 	void loop(u32 instruction);
 	void mad(u32 instruction);
 	void madi(u32 instruction);
@@ -127,15 +170,16 @@ class PICAShader {
 		u32 compSwizzle;
 		bool negate;
 
+		using namespace Helpers;
 		if constexpr (sourceIndex == 1) { // SRC1
-			negate = (Helpers::getBit<4>(opDescriptor)) != 0;
-			compSwizzle = Helpers::getBits<5, 8>(opDescriptor);
+			negate = (getBit<4>(opDescriptor)) != 0;
+			compSwizzle = getBits<5, 8>(opDescriptor);
 		} else if constexpr (sourceIndex == 2) { // SRC2
-			negate = (Helpers::getBit<13>(opDescriptor)) != 0;
-			compSwizzle = Helpers::getBits<14, 8>(opDescriptor);
+			negate = (getBit<13>(opDescriptor)) != 0;
+			compSwizzle = getBits<14, 8>(opDescriptor);
 		} else if constexpr (sourceIndex == 3) { // SRC3
-			negate = (Helpers::getBit<22>(opDescriptor)) != 0;
-			compSwizzle = Helpers::getBits<23, 8>(opDescriptor);
+			negate = (getBit<22>(opDescriptor)) != 0;
+			compSwizzle = getBits<23, 8>(opDescriptor);
 		}
 
 		// Iterate through every component of the swizzled vector in reverse order
@@ -169,17 +213,9 @@ class PICAShader {
 	bool isCondTrue(u32 instruction);
 
 public:
-	std::array<u32, 4096> loadedShader; // Currently loaded & active shader
-	std::array<u32, 4096> bufferedShader; // Shader to be transferred when the SH_CODETRANSFER_END reg gets written to
-
-	u32 entrypoint = 0; // Initial shader PC
-	u32 boolUniform;
-	std::array<OpenGL::Vector<u8, 4>, 4> intUniforms;
-	std::array<vec4f, 96> floatUniforms;
-
-	std::array<vec4f, 16> fixedAttributes; // Fixed vertex attributes
-	std::array<vec4f, 16> inputs; // Attributes passed to the shader
-	std::array<vec4f, 16> outputs;
+	static constexpr size_t maxInstructionCount = 4096;
+	std::array<u32, maxInstructionCount> loadedShader; // Currently loaded & active shader
+	std::array<u32, maxInstructionCount> bufferedShader; // Shader to be transferred when the SH_CODETRANSFER_END reg gets written to
 
 	PICAShader(ShaderType type) : type(type) {}
 
@@ -200,11 +236,15 @@ public:
 		if (bufferIndex >= 4095) Helpers::panic("o no, shader upload overflew");
 		bufferedShader[bufferIndex++] = word;
 		bufferIndex &= 0xfff;
+
+		codeHashDirty = true; // Signal the JIT if necessary that the program hash has potentially changed
 	}
 
 	void uploadDescriptor(u32 word) {
 		operandDescriptors[opDescriptorIndex++] = word;
 		opDescriptorIndex &= 0x7f;
+
+		opdescHashDirty = true; // Signal the JIT if necessary that the program hash has potentially changed
 	}
 
 	void setFloatUniformIndex(u32 word) {
@@ -237,13 +277,18 @@ public:
 	}
 
 	void uploadIntUniform(int index, u32 word) {
+		using namespace Helpers;
+
 		auto& u = intUniforms[index];
 		u.x() = word & 0xff;
-		u.y() = Helpers::getBits<8, 8>(word);
-		u.z() = Helpers::getBits<16, 8>(word);
-		u.w() = Helpers::getBits<24, 8>(word);
+		u.y() = getBits<8, 8>(word);
+		u.z() = getBits<16, 8>(word);
+		u.w() = getBits<24, 8>(word);
 	}
 
 	void run();
 	void reset();
+
+	Hash getCodeHash();
+	Hash getOpdescHash();
 };

include/logger.hpp

@@ -26,6 +26,7 @@ namespace Log {
     static Logger<false> threadLogger;
     static Logger<false> gpuLogger;
     static Logger<false> rendererLogger;
+    static Logger<false> shaderJITLogger;
 
     // Service loggers
     static Logger<false> acLogger;

include/memory.hpp

@@ -34,7 +34,10 @@ namespace VirtualAddrs {
 
 		NormalHeapStart = 0x08000000,
 		LinearHeapStartOld = 0x14000000, // If kernel version < 0x22C
+		LinearHeapEndOld = 0x1C000000,
+
 		LinearHeapStartNew = 0x30000000,
+		LinearHeapEndNew = 0x40000000,
 
 		// Start of TLS for first thread. Next thread's storage will be at TLSBase + 0x1000, and so on
 		TLSBase = 0xFF400000,

include/opengl.hpp

@@ -20,6 +20,7 @@
 #pragma once
 #include <array>
 #include <cassert>
+#include <cstddef>
 #include <cstdarg>
 #include <functional>
 #include <initializer_list>
@@ -627,7 +628,8 @@ namespace OpenGL {
         T& y() { return g(); }
         T& z() { return b(); }
         T& w() { return a(); }
-        T& operator[](int index) { return m_storage[index]; }
+        T& operator[](size_t index) { return m_storage[index]; }
+        const T& operator[](size_t index) const { return m_storage[index]; }
 
         T& u() { return r(); }
         T& v() { return g(); }

include/renderer_gl/renderer_gl.hpp

@@ -2,6 +2,7 @@
 #include <array>
 #include <span>
 
+#include "PICA/float_types.hpp"
 #include "helpers.hpp"
 #include "logger.hpp"
 #include "opengl.hpp"
@@ -15,7 +16,11 @@ class GPU;
 struct Vertex {
 	OpenGL::vec4 position;
 	OpenGL::vec4 colour;
-	OpenGL::vec2 UVs;
+	OpenGL::vec2 texcoord0;
+	OpenGL::vec2 texcoord1;
+	Floats::f24 texcoord0_w;
+	u32 padding; // pad so that texcoord2 is 64-bit aligned
+	OpenGL::vec2 texcoord2;
 };
 
 class Renderer {
@@ -28,6 +33,15 @@ class Renderer {
 	GLint alphaControlLoc = -1;
 	GLint texUnitConfigLoc = -1;
 
+	// TEV configuration uniform locations
+	GLint textureEnvSourceLoc = -1;
+	GLint textureEnvOperandLoc = -1;
+	GLint textureEnvCombinerLoc = -1;
+	GLint textureEnvColorLoc = -1;
+	GLint textureEnvScaleLoc = -1;
+	GLint textureEnvUpdateBufferLoc = -1;
+	GLint textureEnvBufferColorLoc = -1;
+
 	// Depth configuration uniform locations
 	GLint depthOffsetLoc = -1;
 	GLint depthScaleLoc = -1;
@@ -60,12 +74,17 @@ class Renderer {
 	static constexpr u32 regNum = 0x300;  // Number of internal PICA registers
 	const std::array<u32, regNum>& regs;
 
+	OpenGL::Texture screenTexture;
+	OpenGL::Framebuffer screenFramebuffer;
+
 	OpenGL::Framebuffer getColourFBO();
 	OpenGL::Texture getTexture(Texture& tex);
 
 	MAKE_LOG_FUNCTION(log, rendererLogger)
 	void setupBlending();
 	void bindDepthBuffer();
+	void setupTextureEnvState();
+	void bindTexturesToSlots();
 
   public:
 	Renderer(GPU& gpu, const std::array<u32, regNum>& internalRegs) : gpu(gpu), regs(internalRegs) {}

include/renderer_gl/surface_cache.hpp

@@ -13,6 +13,7 @@
 // Are concerned. We could overload the == operator, but that implies full equality
 // Including equality of the allocated OpenGL resources, which we don't want
 // - A "valid" member that tells us whether the function is still valid or not
+// - A "location" member which tells us which location in 3DS memory this surface occupies
 template <typename SurfaceType, size_t capacity>
 class SurfaceCache {
     // Vanilla std::optional can't hold actual references
@@ -40,6 +41,15 @@ public:
         return std::nullopt;
     }
 
+    OptionalRef findFromAddress(u32 address) {
+        for (auto& e : buffer) {
+            if (e.location == address && e.valid)
+                return e;
+        }
+
+        return std::nullopt;
+    }
+
     // Adds a surface object to the cache and returns it
     SurfaceType& add(const SurfaceType& surface) {
         if (size >= capacity) {

readme.md

@@ -8,7 +8,7 @@ Join our Discord server by pressing on the banner below!
 
 [![Discord Banner 2](https://discord.com/api/guilds/1118695732958994532/widget.png?style=banner2)](https://discord.gg/ZYbugsEmsw)
 
-![screenshot1](docs/img/MK7_Car_Select.png) ![screenshot2](docs/img/OoT_Title.png) ![screenshot3](docs/img/Mayro.png)
+![screenshot1](docs/img/MK7.png) ![screenshot2](docs/img/OoT_Title.png) ![screenshot3](docs/img/pokegang.png)
 
 # Compatibility
 Panda3DS is still in the early stages of development. Many games boot, many don't. Most games have at least some hilariously broken graphics, audio is not supported, performance leaves a bit to be desired mainly thanks to lack of shader acceleration, and most QoL features (including a GUI) are missing.
@@ -29,7 +29,7 @@ The 3DS emulation scene is already pretty mature, with offerings such as [Citra]
 Keep in mind, these are all long-term plans. Until then, the main focus is just improving compatibility
 
 # How to build
-Panda3DS compiles on Windows, Linux and MacOS (if you use a compiler other than AppleClang), without needing to download any system dependencies.
+Panda3DS compiles on Windows, Linux and MacOS, without needing to download any system dependencies.
 
 All you need is CMake and a generator of your choice (Make, Visual Studio, Ninja, etc). Simply clone the repo recursively and build it like your average CMake project.
 

src/core/PICA/dynapica/shader_rec.cpp

@@ -0,0 +1,32 @@
+#include "PICA/dynapica/shader_rec.hpp"
+#include <bit>
+
+#ifdef PANDA3DS_SHADER_JIT_SUPPORTED
+void ShaderJIT::reset() {
+	cache.clear();
+}
+
+void ShaderJIT::prepare(PICAShader& shaderUnit) {
+	shaderUnit.pc = shaderUnit.entrypoint;
+	// We combine the code and operand descriptor hashes into a single hash
+	// This is so that if only one of them changes, we still properly recompile the shader
+	// The combine does rotl(x, 1) ^ y for the merging instead of x ^ y because xor is commutative, hence creating possible collisions
+	// re: https://github.com/wheremyfoodat/Panda3DS/pull/15#discussion_r1229925372
+	Hash hash = std::rotl(shaderUnit.getCodeHash(), 1) ^ shaderUnit.getOpdescHash();
+	auto it = cache.find(hash);
+
+	if (it == cache.end()) { // Block has not been compiled yet
+		auto emitter = std::make_unique<ShaderEmitter>();
+		emitter->compile(shaderUnit);
+		// Get pointer to callbacks
+		entrypointCallback = emitter->getInstructionCallback(shaderUnit.entrypoint);
+		prologueCallback = emitter->getPrologueCallback();
+
+		cache.emplace_hint(it, hash, std::move(emitter));
+	} else { // Block has been compiled and found, use it
+		auto emitter = it->second.get();
+		entrypointCallback = emitter->getInstructionCallback(shaderUnit.entrypoint);
+		prologueCallback = emitter->getPrologueCallback();
+	}
+}
+#endif // PANDA3DS_SHADER_JIT_SUPPORTED

src/core/PICA/dynapica/shader_rec_emitter_x64.cpp

@@ -0,0 +1,952 @@
+#if defined(PANDA3DS_DYNAPICA_SUPPORTED) && defined(PANDA3DS_X64_HOST)
+#include "PICA/dynapica/shader_rec_emitter_x64.hpp"
+
+#include <algorithm>
+#include <bit>
+#include <cassert>
+#include <cstddef>
+#include <immintrin.h>
+#include <smmintrin.h>
+
+using namespace Xbyak;
+using namespace Xbyak::util;
+using namespace Helpers;
+
+// The shader recompiler uses quite an odd internal ABI
+// We make use of the fact that in regular conditions, we should pretty much never be calling C++ code from recompiled shader code
+// This allows us to establish an ABI that's optimized for this sort of workflow, statically allocating volatile host registers
+// To avoid pushing/popping registers, not properly maintaining stack alignment, etc
+// This generates faster recompiled code at the cost of being actively hostile against C++ interop
+// To do C++ interop, we're going to have our callCppFunc function call the C++ function, and take extreme measures to ensure we don't violate
+// The host ABI, such as pushing/popping our temporary registers (derp), force aligning the stack and setting up an entire stack frame, etc
+// This is slow, but we do not care as we should never be calling C++ code in normal, non-debugging conditions
+// The only code that might be called are helpers that are also written in assembly, for things like log2
+
+static constexpr Xmm scratch1 = xmm0;
+static constexpr Xmm scratch2 = xmm1;
+static constexpr Xmm src1_xmm = xmm2;
+static constexpr Xmm src2_xmm = xmm3;
+static constexpr Xmm src3_xmm = xmm4;
+
+#if defined(PANDA3DS_MS_ABI)
+// Register that points to PICA state. Must be volatile for the aforementioned reasons
+static constexpr Reg64 statePointer = r8;
+#elif defined(PANDA3DS_SYSV_ABI)
+static constexpr Reg64 statePointer = rdi;
+#else
+#error Unknown ABI for x86-64 shader JIT
+#endif
+
+void ShaderEmitter::compile(const PICAShader& shaderUnit) {
+	// Constants
+	align(16);
+	L(negateVector);
+	dd(0x80000000); dd(0x80000000); dd(0x80000000); dd(0x80000000); // -0.0 4 times
+	L(onesVector);
+	dd(0x3f800000); dd(0x3f800000); dd(0x3f800000); dd(0x3f800000); // 1.0 4 times
+
+	// Emit prologue first
+	align(16);
+	prologueCb = getCurr<PrologueCallback>();
+
+	// Set state pointer to the proper pointer
+	// state pointer is volatile, no need to preserve it
+	mov(statePointer, arg1.cvt64());
+
+	// Push a return guard on the stack. This happens due to the way we handle the PICA callstack, by pushing the return PC to stack
+	// By pushing ffff'ffff, we make it impossible for a return check to erroneously pass
+	push(qword, 0xffffffff);
+	// Lower rsp by 8 for the PICA return stack to be consistent
+	sub(rsp, 8);
+
+	// Tail call to shader code entrypoint
+	jmp(arg2);
+	align(16);
+	// Scan the shader code for call instructions and add them to the list of possible return PCs. We need to do this because the PICA callstack works
+	// Pretty weirdly
+	scanForCalls(shaderUnit);
+
+	// Compile every instruction in the shader
+	// This sounds horrible but the PICA instruction memory is tiny, and most of the time it's padded wtih nops that compile to nothing
+	recompilerPC = 0;
+	loopLevel = 0;
+	compileUntil(shaderUnit, PICAShader::maxInstructionCount);
+}
+
+void ShaderEmitter::scanForCalls(const PICAShader& shaderUnit) {
+	returnPCs.clear();
+
+	for (u32 i = 0; i < PICAShader::maxInstructionCount; i++) {
+		const u32 instruction = shaderUnit.loadedShader[i];
+		if (isCall(instruction)) {
+			const u32 num = instruction & 0xff;
+			const u32 dest = getBits<10, 12>(instruction);
+			const u32 returnPC = num + dest; // Add them to get the return PC
+
+			returnPCs.push_back(returnPC);
+		}
+	}
+
+	// Sort return PCs so they can be binary searched
+	std::sort(returnPCs.begin(), returnPCs.end());
+}
+
+void ShaderEmitter::compileUntil(const PICAShader& shaderUnit, u32 end) {
+	while (recompilerPC < end) {
+		compileInstruction(shaderUnit);
+	}
+}
+
+void ShaderEmitter::compileInstruction(const PICAShader& shaderUnit) {
+	// Write current location to label for this instruction
+	L(instructionLabels[recompilerPC]);
+
+	// See if PC is a possible return PC and emit the proper code if so
+	if (std::binary_search(returnPCs.begin(), returnPCs.end(), recompilerPC)) {
+		constexpr uintptr_t stackOffsetForPC = 8;
+
+		Label end;
+		// Check if return address == recompilerPC, ie if we should return
+		cmp(dword[rsp + stackOffsetForPC], recompilerPC);
+		jne(end);  // If not, continue with our lives
+		ret();     // If yes, return
+
+		L(end);
+	}
+
+	// Fetch instruction and inc PC
+	const u32 instruction = shaderUnit.loadedShader[recompilerPC++];
+	const u32 opcode = instruction >> 26;
+
+	switch (opcode) {
+		case ShaderOpcodes::ADD: recADD(shaderUnit, instruction); break;
+		case ShaderOpcodes::CALL:
+			recCALL(shaderUnit, instruction);
+			break;
+		case ShaderOpcodes::CALLC:
+			recCALLC(shaderUnit, instruction);
+			break;
+		case ShaderOpcodes::CALLU:
+			recCALLU(shaderUnit, instruction);
+			break;
+		case ShaderOpcodes::CMP1: case ShaderOpcodes::CMP2:
+			recCMP(shaderUnit, instruction);
+			break;
+		case ShaderOpcodes::DP3: recDP3(shaderUnit, instruction); break;
+		case ShaderOpcodes::DP4: recDP4(shaderUnit, instruction); break;
+		case ShaderOpcodes::END: recEND(shaderUnit, instruction); break;
+		case ShaderOpcodes::FLR: recFLR(shaderUnit, instruction); break;
+		case ShaderOpcodes::IFC: recIFC(shaderUnit, instruction); break;
+		case ShaderOpcodes::IFU: recIFU(shaderUnit, instruction); break;
+		case ShaderOpcodes::JMPC: recJMPC(shaderUnit, instruction); break;
+		case ShaderOpcodes::JMPU: recJMPU(shaderUnit, instruction); break;
+		case ShaderOpcodes::LOOP: recLOOP(shaderUnit, instruction); break;
+		case ShaderOpcodes::MOV: recMOV(shaderUnit, instruction); break;
+		case ShaderOpcodes::MOVA: recMOVA(shaderUnit, instruction); break;
+		case ShaderOpcodes::MAX: recMAX(shaderUnit, instruction); break;
+		case ShaderOpcodes::MIN: recMIN(shaderUnit, instruction); break;
+		case ShaderOpcodes::MUL: recMUL(shaderUnit, instruction); break;
+		case ShaderOpcodes::NOP: break;
+		case ShaderOpcodes::RCP: recRCP(shaderUnit, instruction); break;
+		case ShaderOpcodes::RSQ: recRSQ(shaderUnit, instruction); break;
+
+		// Unimplemented opcodes that don't seem to actually be used but exist in the binary
+		// EMIT/SETEMIT are used in geometry shaders, however are sometimes found in vertex shaders?
+		case ShaderOpcodes::EX2:
+		case ShaderOpcodes::LG2:
+		case ShaderOpcodes::EMIT:
+		case ShaderOpcodes::SETEMIT:
+			log("[ShaderJIT] Unknown PICA opcode: %02X\n", opcode);
+			emitPrintLog(shaderUnit);
+			break;
+
+		// We consider both MAD and MADI to be the same instruction and decode which one we actually have in recMAD
+		case 0x30: case 0x31: case 0x32: case 0x33: case 0x34: case 0x35: case 0x36: case 0x37:
+		case 0x38: case 0x39: case 0x3A: case 0x3B: case 0x3C: case 0x3D: case 0x3E: case 0x3F:
+			recMAD(shaderUnit, instruction);
+			break;
+
+		case ShaderOpcodes::SLT:
+		case ShaderOpcodes::SLTI:
+			recSLT(shaderUnit, instruction); break;
+
+		default:
+			Helpers::panic("Shader JIT: Unimplemented PICA opcode %X", opcode);
+	}
+}
+
+const ShaderEmitter::vec4f& ShaderEmitter::getSourceRef(const PICAShader& shader, u32 src) {
+	if (src < 0x10)
+		return shader.inputs[src];
+	else if (src < 0x20)
+		return shader.tempRegisters[src - 0x10];
+	else if (src <= 0x7f)
+		return shader.floatUniforms[src - 0x20];
+	else {
+		Helpers::warn("[Shader JIT] Unimplemented source value: %X\n", src);
+		return shader.dummy;
+	}
+}
+
+const ShaderEmitter::vec4f& ShaderEmitter::getDestRef(const PICAShader& shader, u32 dest) {
+	if (dest < 0x10) {
+		return shader.outputs[dest];
+	} else if (dest < 0x20) {
+		return shader.tempRegisters[dest - 0x10];
+	}
+	Helpers::panic("[Shader JIT] Unimplemented dest: %X", dest);
+}
+
+// See shader.hpp header for docs on how the swizzle and negate works
+template <int sourceIndex>
+void ShaderEmitter::loadRegister(Xmm dest, const PICAShader& shader, u32 src, u32 index, u32 operandDescriptor) {
+	u32 compSwizzle; // Component swizzle pattern for the register
+	bool negate;     // If true, negate all lanes of the register
+
+	if constexpr (sourceIndex == 1) { // SRC1
+		negate = (getBit<4>(operandDescriptor)) != 0;
+		compSwizzle = getBits<5, 8>(operandDescriptor);
+	}
+	else if constexpr (sourceIndex == 2) { // SRC2
+		negate = (getBit<13>(operandDescriptor)) != 0;
+		compSwizzle = getBits<14, 8>(operandDescriptor);
+	}
+	else if constexpr (sourceIndex == 3) { // SRC3
+		negate = (getBit<22>(operandDescriptor)) != 0;
+		compSwizzle = getBits<23, 8>(operandDescriptor);
+	}
+
+	// PICA has the swizzle descriptor inverted in comparison to x86. For the PICA, the descriptor is (lowest to highest bits) wzyx while it's xyzw for x86
+	u32 convertedSwizzle = ((compSwizzle >> 6) & 0b11) | (((compSwizzle >> 4) & 0b11) << 2) | (((compSwizzle >> 2) & 0b11) << 4) | ((compSwizzle & 0b11) << 6);
+
+	switch (index) {
+		case 0: [[likely]] { // Keep src as is, no need to offset it
+			const vec4f& srcRef = getSourceRef(shader, src);
+			const uintptr_t offset = uintptr_t(&srcRef) - uintptr_t(&shader); // Calculate offset of register from start of the state struct
+
+			if (compSwizzle == noSwizzle) // Avoid emitting swizzle if not necessary
+				movaps(dest, xword[statePointer + offset]);
+			else // Swizzle is not trivial so we need to emit a shuffle instruction
+				pshufd(dest, xword[statePointer + offset], convertedSwizzle);
+
+			// Negate the register if necessary
+			if (negate) {
+				pxor(dest, xword[rip + negateVector]);
+			}
+			return; // Return. Rest of the function handles indexing which is not used if index == 0
+		}
+
+		case 1: {
+			const uintptr_t addrXOffset = uintptr_t(&shader.addrRegister[0]) - uintptr_t(&shader);
+			movsxd(rax, dword[statePointer + addrXOffset]); // rax = address register x
+			break;
+		}
+
+		case 2: {
+			const uintptr_t addrYOffset = uintptr_t(&shader.addrRegister[1]) - uintptr_t(&shader);
+			movsxd(rax, dword[statePointer + addrYOffset]); // rax = address register y
+			break;
+		}
+
+		case 3: {
+			const uintptr_t loopCounterOffset = uintptr_t(&shader.loopCounter) - uintptr_t(&shader);
+			mov(eax, dword[statePointer + loopCounterOffset]); // rax = loop counter
+			break;
+		}
+		
+		default:
+			Helpers::panic("[ShaderJIT]: Unimplemented source index type %d", index);
+	}
+
+	// Swizzle and load register into dest, from [state pointer + rcx + offset] and apply the relevant swizzle
+	auto swizzleAndLoadReg = [this, &dest, &compSwizzle, &convertedSwizzle](size_t offset) {
+		if (compSwizzle == noSwizzle)  // Avoid emitting swizzle if not necessary
+			movaps(dest, xword[statePointer + rcx + offset]);
+		else  // Swizzle is not trivial so we need to emit a shuffle instruction
+			pshufd(dest, xword[statePointer + rcx + offset], convertedSwizzle);
+	};
+
+	// Here we handle what happens when using indexed addressing & we can't predict what register will be read at compile time
+	// The index of the access is assumed to be in rax
+	// Add source register (src) and index (rax) to form the final register
+	add(rax, src);
+
+	Label maybeTemp, maybeUniform, unknownReg, end;
+	const uintptr_t inputOffset = uintptr_t(&shader.inputs[0]) - uintptr_t(&shader);
+	const uintptr_t tempOffset = uintptr_t(&shader.tempRegisters[0]) - uintptr_t(&shader);
+	const uintptr_t uniformOffset = uintptr_t(&shader.floatUniforms[0]) - uintptr_t(&shader);
+
+	// If reg < 0x10, return inputRegisters[reg]
+	cmp(rax, 0x10);
+	jae(maybeTemp);
+	mov(rcx, rax);
+	shl(rcx, 4);  // rcx = rax * sizeof(vec4 of floats) = rax * 16
+	swizzleAndLoadReg(inputOffset);
+	jmp(end);
+	
+	// If (reg < 0x1F) return tempRegisters[reg - 0x10]
+	L(maybeTemp);
+	cmp(rax, 0x20);
+	jae(maybeUniform);
+	lea(rcx, qword[rax - 0x10]);
+	shl(rcx, 4);
+	swizzleAndLoadReg(tempOffset);
+	jmp(end);
+
+	// If (reg < 0x80) return floatUniforms[reg - 0x20]
+	L(maybeUniform);
+	cmp(rax, 0x80);
+	jae(unknownReg);
+	lea(rcx, qword[rax - 0x20]);
+	shl(rcx, 4);
+	swizzleAndLoadReg(uniformOffset);
+	jmp(end);
+
+	L(unknownReg);
+	pxor(dest, dest); // Set dest to 0 if we're reading from a garbage register
+
+	L(end);
+	// Negate the register if necessary
+	if (negate) {
+		pxor(dest, xword[rip + negateVector]);
+	}
+}
+
+void ShaderEmitter::storeRegister(Xmm source, const PICAShader& shader, u32 dest, u32 operandDescriptor) {
+	const vec4f& destRef = getDestRef(shader, dest);
+	const uintptr_t offset = uintptr_t(&destRef) - uintptr_t(&shader); // Calculate offset of register from start of the state struct
+
+	// Mask of which lanes to write
+	u32 writeMask = operandDescriptor & 0xf;
+	if (writeMask == 0xf) { // No lanes are masked, just movaps
+		movaps(xword[statePointer + offset], source);
+	} else if (std::popcount(writeMask) == 1) { // Only 1 register needs to be written back. This can be done with a simple shift right + movss
+		int bit = std::countr_zero(writeMask); // Get which PICA register needs to be written to (0 = w, 1 = z, etc)
+		size_t index = 3 - bit;
+		const uintptr_t lane_offset = offset + index * sizeof(float);
+
+		if (index == 0) { // Bottom lane, no need to shift
+			movss(dword[statePointer + lane_offset], source);
+		} else { // Shift right by 32 * index, then write bottom lane
+			if (haveAVX) {
+				vpsrldq(scratch1, source, index * sizeof(float));
+			} else {
+				movaps(scratch1, source);
+				psrldq(scratch1, index * sizeof(float));
+			}
+			movss(dword[statePointer + lane_offset], scratch1);
+		}
+	} else if (haveSSE4_1) {
+		// Bit reverse the write mask because that is what blendps expects
+		u32 adjustedMask = ((writeMask >> 3) & 0b1) | ((writeMask >> 1) & 0b10) | ((writeMask << 1) & 0b100) | ((writeMask << 3) & 0b1000);
+		// Don't accidentally overwrite scratch1 if that is what we're writing derp
+		Xmm temp = (source == scratch1) ? scratch2 : scratch1;
+
+		movaps(temp, xword[statePointer + offset]);     // Read current value of dest
+		blendps(temp, source, adjustedMask);            // Blend with source
+		movaps(xword[statePointer + offset], temp);     // Write back
+	} else {
+		// Blend algo referenced from Citra
+		const u8 selector = (((writeMask & 0b1000) ? 1 : 0) << 0) |
+			(((writeMask & 0b0100) ? 3 : 2) << 2) |
+			(((writeMask & 0b0010) ? 0 : 1) << 4) |
+			(((writeMask & 0b0001) ? 2 : 3) << 6);
+		
+		// Reorder instructions based on whether the source == scratch1. This is to avoid overwriting scratch1 if it's the source,
+		// While also having the memory load come first to mitigate execution hazards and give the load more time to complete before reading if possible
+		if (source != scratch1) {
+			movaps(scratch1, xword[statePointer + offset]);
+			movaps(scratch2, source);
+		} else {
+			movaps(scratch2, source);
+			movaps(scratch1, xword[statePointer + offset]);
+		}
+		
+		unpckhps(scratch2, scratch1); // Unpack X/Y components of source and destination
+		unpcklps(scratch1, source);   // Unpack Z/W components of source and destination
+		shufps(scratch1, scratch2, selector); // "merge-shuffle" dest and source using selecto
+		movaps(xword[statePointer + offset], scratch1); // Write back
+	}
+}
+
+void ShaderEmitter::checkCmpRegister(const PICAShader& shader, u32 instruction) {
+	static_assert(sizeof(bool) == 1 && sizeof(shader.cmpRegister) == 2); // The code below relies on bool being 1 byte exactly
+	const size_t cmpRegXOffset = uintptr_t(&shader.cmpRegister[0]) - uintptr_t(&shader);
+	const size_t cmpRegYOffset = cmpRegXOffset + sizeof(bool);
+
+	const u32 condition = getBits<22, 2>(instruction);
+	const uint refY = getBit<24>(instruction);
+	const uint refX = getBit<25>(instruction);
+
+	// refX in the bottom byte, refY in the top byte. This is done for condition codes 0 and 1 which check both x and y, so we can emit a single instruction that checks both
+	const u16 refX_refY_merged = refX | (refY << 8);
+
+	switch (condition) {
+		case 0: // Either cmp register matches 
+			// Z flag is 0 if at least 1 of them is set
+			test(word[statePointer + cmpRegXOffset], refX_refY_merged);
+
+			// Invert z flag
+			setz(al);
+			test(al, al);
+			break;
+		case 1: // Both cmp registers match
+			cmp(word[statePointer + cmpRegXOffset], refX_refY_merged);
+			break;
+		case 2: // At least cmp.x matches
+			cmp(byte[statePointer + cmpRegXOffset], refX);
+			break;
+		default: // At least cmp.y matches
+			cmp(byte[statePointer + cmpRegYOffset], refY);
+			break;
+	}
+}
+
+void ShaderEmitter::checkBoolUniform(const PICAShader& shader, u32 instruction) {
+	const u32 bit = getBits<22, 4>(instruction); // Bit of the bool uniform to check
+	const uintptr_t boolUniformOffset = uintptr_t(&shader.boolUniform) - uintptr_t(&shader);
+
+	test(word[statePointer + boolUniformOffset], 1 << bit);
+}
+
+void ShaderEmitter::recEND(const PICAShader& shader, u32 instruction) {
+	// Undo anything the prologue did and return
+	// Deallocate the 8 bytes taken up for the return guard + the 8 bytes of rsp padding we inserted in the prologue
+	add(rsp, 16);
+	ret();
+}
+
+void ShaderEmitter::recMOV(const PICAShader& shader, u32 instruction) {
+	const u32 operandDescriptor = shader.operandDescriptors[instruction & 0x7f];
+	const u32 src = getBits<12, 7>(instruction);
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 dest = getBits<21, 5>(instruction);
+
+	loadRegister<1>(src1_xmm, shader, src, idx, operandDescriptor); // Load source 1 into scratch1
+	storeRegister(src1_xmm, shader, dest, operandDescriptor);
+}
+
+void ShaderEmitter::recFLR(const PICAShader& shader, u32 instruction) {
+	const u32 operandDescriptor = shader.operandDescriptors[instruction & 0x7f];
+	const u32 src = getBits<12, 7>(instruction);
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 dest = getBits<21, 5>(instruction);
+
+	loadRegister<1>(src1_xmm, shader, src, idx, operandDescriptor); // Load source 1 into scratch1
+	if (haveSSE4_1) {
+		roundps(src1_xmm, src1_xmm, _MM_FROUND_FLOOR);
+	} else {
+		cvttps2dq(src1_xmm, src1_xmm); // Truncate and convert to integer
+		cvtdq2ps(src1_xmm, src1_xmm);  // Convert from integer back to float
+	}
+
+	storeRegister(src1_xmm, shader, dest, operandDescriptor);
+}
+
+void ShaderEmitter::recMOVA(const PICAShader& shader, u32 instruction) {
+	const u32 operandDescriptor = shader.operandDescriptors[instruction & 0x7f];
+	const u32 src = getBits<12, 7>(instruction);
+	const u32 idx = getBits<19, 2>(instruction);
+
+	const bool writeX = getBit<3>(operandDescriptor); // Should we write the x component of the address register?
+	const bool writeY = getBit<2>(operandDescriptor);
+
+	static_assert(sizeof(shader.addrRegister) == 2 * sizeof(s32)); // Assert that the address register is 2 s32s
+	const uintptr_t addrRegisterOffset = uintptr_t(&shader.addrRegister[0]) - uintptr_t(&shader);
+	const uintptr_t addrRegisterYOffset = addrRegisterOffset + sizeof(shader.addrRegister[0]);
+
+	// If no register is being written to then it is a nop. Probably not common but whatever
+	if (!writeX && !writeY) return;
+	loadRegister<1>(src1_xmm, shader, src, idx, operandDescriptor); // Load source 1 into scratch1
+
+	// Write both
+	if (writeX && writeY) {
+		cvttps2dq(scratch1, src1_xmm); // Convert all lanes of src1 with truncation
+		movsd(qword[statePointer + addrRegisterOffset], scratch1); // Write back bottom 2 to addr register x and ys
+	}
+	else if (writeX) {
+		cvttss2si(eax, src1_xmm); // Convert bottom lane
+		mov(dword[statePointer + addrRegisterOffset], eax); // Write it back
+	}
+	else if (writeY) {
+		psrldq(src1_xmm, sizeof(float)); // Shift y component to bottom lane
+		cvttss2si(eax, src1_xmm); // Convert bottom lane
+		mov(dword[statePointer + addrRegisterYOffset], eax); // Write it back to y component
+	}
+}
+
+void ShaderEmitter::recADD(const PICAShader& shader, u32 instruction) {
+	const u32 operandDescriptor = shader.operandDescriptors[instruction & 0x7f];
+	const u32 src1 = getBits<12, 7>(instruction);
+	const u32 src2 = getBits<7, 5>(instruction); // src2 coming first because PICA moment
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 dest = getBits<21, 5>(instruction);
+
+	loadRegister<1>(src1_xmm, shader, src1, idx, operandDescriptor);
+	loadRegister<2>(src2_xmm, shader, src2, 0, operandDescriptor);
+	addps(src1_xmm, src2_xmm); // Dot product between the 2 register
+	storeRegister(src1_xmm, shader, dest, operandDescriptor);
+}
+
+void ShaderEmitter::recDP3(const PICAShader& shader, u32 instruction) {
+	const u32 operandDescriptor = shader.operandDescriptors[instruction & 0x7f];
+	const u32 src1 = getBits<12, 7>(instruction);
+	const u32 src2 = getBits<7, 5>(instruction); // src2 coming first because PICA moment
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 dest = getBits<21, 5>(instruction);
+
+	// TODO: Safe multiplication equivalent (Multiplication is not IEEE compliant on the PICA)
+	loadRegister<1>(src1_xmm, shader, src1, idx, operandDescriptor);
+	loadRegister<2>(src2_xmm, shader, src2, 0, operandDescriptor);
+	dpps(src1_xmm, src2_xmm, 0b01111111); // 3-lane dot product between the 2 registers, store the result in all lanes of scratch1 similarly to PICA 
+	storeRegister(src1_xmm, shader, dest, operandDescriptor);
+}
+
+void ShaderEmitter::recDP4(const PICAShader& shader, u32 instruction) {
+	const u32 operandDescriptor = shader.operandDescriptors[instruction & 0x7f];
+	const u32 src1 = getBits<12, 7>(instruction);
+	const u32 src2 = getBits<7, 5>(instruction); // src2 coming first because PICA moment
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 dest = getBits<21, 5>(instruction);
+
+	// TODO: Safe multiplication equivalent (Multiplication is not IEEE compliant on the PICA)
+	loadRegister<1>(src1_xmm, shader, src1, idx, operandDescriptor);
+	loadRegister<2>(src2_xmm, shader, src2, 0, operandDescriptor);
+	dpps(src1_xmm, src2_xmm, 0b11111111); // 4-lane dot product between the 2 registers, store the result in all lanes of scratch1 similarly to PICA 
+	storeRegister(src1_xmm, shader, dest, operandDescriptor);
+}
+
+void ShaderEmitter::recMAX(const PICAShader& shader, u32 instruction) {
+	const u32 operandDescriptor = shader.operandDescriptors[instruction & 0x7f];
+	const u32 src1 = getBits<12, 7>(instruction);
+	const u32 src2 = getBits<7, 5>(instruction); // src2 coming first because PICA moment
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 dest = getBits<21, 5>(instruction);
+
+	loadRegister<1>(src1_xmm, shader, src1, idx, operandDescriptor);
+	loadRegister<2>(src2_xmm, shader, src2, 0, operandDescriptor);
+	maxps(src1_xmm, src2_xmm);
+	storeRegister(src1_xmm, shader, dest, operandDescriptor);
+}
+
+void ShaderEmitter::recMIN(const PICAShader& shader, u32 instruction) {
+	const u32 operandDescriptor = shader.operandDescriptors[instruction & 0x7f];
+	const u32 src1 = getBits<12, 7>(instruction);
+	const u32 src2 = getBits<7, 5>(instruction); // src2 coming first because PICA moment
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 dest = getBits<21, 5>(instruction);
+
+	loadRegister<1>(src1_xmm, shader, src1, idx, operandDescriptor);
+	loadRegister<2>(src2_xmm, shader, src2, 0, operandDescriptor);
+	minps(src1_xmm, src2_xmm);
+	storeRegister(src1_xmm, shader, dest, operandDescriptor);
+}
+
+void ShaderEmitter::recMUL(const PICAShader& shader, u32 instruction) {
+	const u32 operandDescriptor = shader.operandDescriptors[instruction & 0x7f];
+	const u32 src1 = getBits<12, 7>(instruction);
+	const u32 src2 = getBits<7, 5>(instruction); // src2 coming first because PICA moment
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 dest = getBits<21, 5>(instruction);
+
+	// TODO: Safe multiplication equivalent (Multiplication is not IEEE compliant on the PICA)
+	loadRegister<1>(src1_xmm, shader, src1, idx, operandDescriptor);
+	loadRegister<2>(src2_xmm, shader, src2, 0, operandDescriptor);
+	mulps(src1_xmm, src2_xmm);
+	storeRegister(src1_xmm, shader, dest, operandDescriptor);
+}
+
+void ShaderEmitter::recRCP(const PICAShader& shader, u32 instruction) {
+	const u32 operandDescriptor = shader.operandDescriptors[instruction & 0x7f];
+	const u32 src = getBits<12, 7>(instruction);
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 dest = getBits<21, 5>(instruction);
+	const u32 writeMask = operandDescriptor & 0xf;
+
+	loadRegister<1>(src1_xmm, shader, src, idx, operandDescriptor); // Load source 1 into scratch1
+	rcpss(src1_xmm, src1_xmm); // Compute rcp approximation
+
+	// If we only write back the x component to the result, we needn't perform a shuffle to do res = res.xxxx
+	// Otherwise we do
+	if (writeMask != 0x8) {// Copy bottom lane to all lanes if we're not simply writing back x
+		shufps(src1_xmm, src1_xmm, 0); // src1_xmm = src1_xmm.xxxx
+	}
+
+	storeRegister(src1_xmm, shader, dest, operandDescriptor);
+}
+
+void ShaderEmitter::recRSQ(const PICAShader& shader, u32 instruction) {
+	const u32 operandDescriptor = shader.operandDescriptors[instruction & 0x7f];
+	const u32 src = getBits<12, 7>(instruction);
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 dest = getBits<21, 5>(instruction);
+	const u32 writeMask = operandDescriptor & 0xf;
+
+	loadRegister<1>(src1_xmm, shader, src, idx, operandDescriptor); // Load source 1 into scratch1
+	rsqrtss(src1_xmm, src1_xmm); // Compute rsqrt approximation
+
+	// If we only write back the x component to the result, we needn't perform a shuffle to do res = res.xxxx
+	// Otherwise we do
+	if (writeMask != 0x8) {// Copy bottom lane to all lanes if we're not simply writing back x
+		shufps(src1_xmm, src1_xmm, 0); // src1_xmm = src1_xmm.xxxx
+	}
+
+	storeRegister(src1_xmm, shader, dest, operandDescriptor);
+}
+
+void ShaderEmitter::recMAD(const PICAShader& shader, u32 instruction) {
+	const bool isMADI = getBit<29>(instruction) == 0;
+
+	const u32 operandDescriptor = shader.operandDescriptors[instruction & 0x1f];
+	const u32 src1 = getBits<17, 5>(instruction);
+	const u32 src2 = isMADI ? getBits<12, 5>(instruction) : getBits<10, 7>(instruction);
+	const u32 src3 = isMADI ? getBits<5, 7>(instruction) : getBits<5, 5>(instruction);
+	const u32 idx = getBits<22, 2>(instruction);
+	const u32 dest = getBits<24, 5>(instruction);
+
+	loadRegister<1>(src1_xmm, shader, src1, 0, operandDescriptor);
+	loadRegister<2>(src2_xmm, shader, src2, isMADI ? 0 : idx, operandDescriptor);
+	loadRegister<3>(src3_xmm, shader, src3, isMADI ? idx : 0, operandDescriptor);
+
+	// TODO: Implement safe PICA mul
+	// If we have FMA3, optimize MAD to use FMA
+	if (haveFMA3) {
+		vfmadd213ps(src1_xmm, src2_xmm, src3_xmm);
+		storeRegister(src1_xmm, shader, dest, operandDescriptor);
+	}
+	
+	// If we don't have FMA3, do a multiplication and addition
+	else {
+		// Multiply src1 * src2
+		if (haveAVX) {
+			vmulps(scratch1, src1_xmm, src2_xmm);
+		} else {
+			movaps(scratch1, src1_xmm);
+			mulps(scratch1, src2_xmm);
+		}
+
+		// Add src3
+		addps(scratch1, src3_xmm);
+		storeRegister(scratch1, shader, dest, operandDescriptor);
+	}
+}
+
+void ShaderEmitter::recSLT(const PICAShader& shader, u32 instruction) {
+	const bool isSLTI = (instruction >> 26) == ShaderOpcodes::SLTI;
+	const u32 operandDescriptor = shader.operandDescriptors[instruction & 0x7f];
+
+	const u32 src1 = isSLTI ? getBits<14, 5>(instruction) : getBits<12, 7>(instruction);
+	const u32 src2 = isSLTI ? getBits<7, 7>(instruction) : getBits<7, 5>(instruction);
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 dest = getBits<21, 5>(instruction);
+
+	loadRegister<1>(src1_xmm, shader, src1, isSLTI ? 0 : idx, operandDescriptor);
+	loadRegister<2>(src2_xmm, shader, src2, isSLTI ? idx : 0, operandDescriptor);
+	cmpltps(src1_xmm, src2_xmm);
+	andps(src1_xmm, xword[rip + onesVector]);
+	storeRegister(src1_xmm, shader, dest, operandDescriptor);
+}
+
+void ShaderEmitter::recCMP(const PICAShader& shader, u32 instruction) {
+	const u32 operandDescriptor = shader.operandDescriptors[instruction & 0x7f];
+	const u32 src1 = getBits<12, 7>(instruction);
+	const u32 src2 = getBits<7, 5>(instruction); // src2 coming first because PICA moment
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 cmpY = getBits<21, 3>(instruction);
+	const u32 cmpX = getBits<24, 3>(instruction);
+
+	loadRegister<1>(src1_xmm, shader, src1, idx, operandDescriptor);
+	loadRegister<2>(src2_xmm, shader, src2, 0, operandDescriptor);
+
+	// Condition codes for cmpps
+	enum : u8 {
+		CMP_EQ = 0,
+		CMP_LT = 1,
+		CMP_LE = 2,
+		CMP_UNORD = 3,
+		CMP_NEQ = 4,
+		CMP_NLT = 5,
+		CMP_NLE = 6,
+		CMP_ORD = 7,
+		CMP_TRUE = 15
+	};
+
+	// Map from PICA condition codes (used as index) to x86 condition codes
+	static constexpr std::array<u8, 8> conditionCodes = { CMP_EQ, CMP_NEQ, CMP_LT, CMP_LE, CMP_LT, CMP_LE, CMP_TRUE, CMP_TRUE };
+
+	// SSE does not offer GT or GE comparisons in the cmpps instruction, so we need to flip the left and right operands in that case and use LT/LE
+	const bool invertX = (cmpX == 4 || cmpX == 5);
+	const bool invertY = (cmpY == 4 || cmpY == 5);
+	Xmm lhs_x = invertX ? src2_xmm : src1_xmm;
+	Xmm rhs_x = invertX ? src1_xmm : src2_xmm;
+	Xmm lhs_y = invertY ? src2_xmm : src1_xmm;
+	Xmm rhs_y = invertY ? src1_xmm : src2_xmm;
+
+	const u8 compareFuncX = conditionCodes[cmpX];
+	const u8 compareFuncY = conditionCodes[cmpY];
+
+	static_assert(sizeof(shader.cmpRegister[0]) == 1 && sizeof(shader.cmpRegister) == 2); // The code below relies on bool being 1 byte exactly
+	const size_t cmpRegXOffset = uintptr_t(&shader.cmpRegister[0]) - uintptr_t(&shader);
+	const size_t cmpRegYOffset = cmpRegXOffset + sizeof(bool);
+
+	// Cmp x and y are the same compare function, we can use a single cmp instruction
+	if (cmpX == cmpY) {
+		cmpps(lhs_x, rhs_x, compareFuncX);
+		movq(rax, lhs_x); // Move both comparison results to rax
+		test(eax, eax);   // Check bottom 32 bits first
+		setne(byte[statePointer + cmpRegXOffset]); // set cmp.x
+
+		shr(rax, 32);     // Check top 32 bits (shr will set the zero flag properly)
+		setne(byte[statePointer + cmpRegYOffset]); // set cmp.y
+	} else {
+		if (haveAVX) {
+			vcmpps(scratch1, lhs_x, rhs_x, compareFuncX); // Perform comparison for X component and store result in scratch1
+			vcmpps(scratch2, lhs_y, rhs_y, compareFuncY); // Perform comparison for Y component and store result in scratch2
+		} else {
+			movaps(scratch1, lhs_x); // Copy the left hand operands to temp registers
+			movaps(scratch2, lhs_y);
+
+			cmpps(scratch1, rhs_x, compareFuncX); // Perform the compares
+			cmpps(scratch2, rhs_y, compareFuncY);
+		}
+
+		movd(eax, scratch1); // Move results to eax for X and edx for Y
+		movq(rdx, scratch2);
+
+		test(eax, eax);      // Write back results with setne
+		setne(byte[statePointer + cmpRegXOffset]);
+		shr(rdx, 32);        // We want the y component for the second comparison. This shift will set zero flag to 0 if the comparison is true
+		setne(byte[statePointer + cmpRegYOffset]);
+	}
+}
+
+void ShaderEmitter::recIFC(const PICAShader& shader, u32 instruction) {
+	// z is 1 if true, else 0
+	checkCmpRegister(shader, instruction);
+	const u32 num = instruction & 0xff;
+	const u32 dest = getBits<10, 12>(instruction);
+
+	if (dest < recompilerPC) {
+		Helpers::warn("Shader JIT: IFC instruction with dest < current PC\n");
+	}
+	Label elseBlock, endIf;
+
+	// Jump to else block if z is 0
+	jnz(elseBlock, T_NEAR);
+	compileUntil(shader, dest);
+
+	if (num == 0) { // Else block is empty,
+		L(elseBlock);
+	} else { // Else block is NOT empty
+		jmp(endIf, T_NEAR); // Skip executing the else branch if the if branch was ran
+		L(elseBlock);
+		compileUntil(shader, dest + num);
+		L(endIf);
+	}
+}
+
+void ShaderEmitter::recIFU(const PICAShader& shader, u32 instruction) {
+	// z is 0 if true, else 1
+	checkBoolUniform(shader, instruction);
+	const u32 num = instruction & 0xff;
+	const u32 dest = getBits<10, 12>(instruction);
+
+	if (dest < recompilerPC) {
+		Helpers::warn("Shader JIT: IFC instruction with dest < current PC\n");
+	}
+	Label elseBlock, endIf;
+
+	// Jump to else block if z is 1
+	jz(elseBlock, T_NEAR);
+	compileUntil(shader, dest);
+
+	if (num == 0) { // Else block is empty,
+		L(elseBlock);
+	} else { // Else block is NOT empty
+		jmp(endIf, T_NEAR); // Skip executing the else branch if the if branch was ran
+		L(elseBlock);
+		compileUntil(shader, dest + num);
+		L(endIf);
+	}
+}
+
+void ShaderEmitter::recCALL(const PICAShader& shader, u32 instruction) {
+	const u32 num = instruction & 0xff;
+	const u32 dest = getBits<10, 12>(instruction);
+
+	// Push return PC as stack parameter. This is a decently fast solution and Citra does the same but we should probably switch to a proper PICA-like
+	// Callstack, because it's not great to have an infinitely expanding call stack where popping from empty stack is undefined as hell
+	push(qword, dest + num);
+	// Call subroutine, Xbyak will update the label if it hasn't been initialized yet
+	call(instructionLabels[dest]);
+
+	// Fix up stack after returning by adding 8 to rsp and "popping" the return PC
+	add(rsp, 8);
+}
+
+void ShaderEmitter::recCALLC(const PICAShader& shader, u32 instruction) {
+	Label skipCall;
+
+	// z is 1 if the call should be taken, 0 otherwise
+	checkCmpRegister(shader, instruction);
+	jnz(skipCall);
+	recCALL(shader, instruction);
+
+	L(skipCall);
+}
+
+void ShaderEmitter::recCALLU(const PICAShader& shader, u32 instruction) {
+	Label skipCall;
+
+	// z is 0 if the call should be taken, 1 otherwise
+	checkBoolUniform(shader, instruction);
+	jz(skipCall);
+	recCALL(shader, instruction);
+
+	L(skipCall);
+}
+
+void ShaderEmitter::recJMPC(const PICAShader& shader, u32 instruction) {
+	const u32 dest = getBits<10, 12>(instruction);
+
+	Label& l = instructionLabels[dest];
+	// Z is 1 if the comparison is true
+	checkCmpRegister(shader, instruction);
+	jz(l, T_NEAR);
+}
+
+void ShaderEmitter::recJMPU(const PICAShader& shader, u32 instruction) {
+	bool jumpIfFalse = instruction & 1; // If the LSB is 0 we want to compare to true, otherwise compare to false
+	const u32 dest = getBits<10, 12>(instruction);
+
+	Label& l = instructionLabels[dest];
+	// Z is 0 if the uniform is true
+	checkBoolUniform(shader, instruction);
+
+	if (jumpIfFalse) {
+		jz(l, T_NEAR);
+	} else {
+		jnz(l, T_NEAR);
+	}
+}
+
+void ShaderEmitter::recLOOP(const PICAShader& shader, u32 instruction) {
+	const u32 dest = getBits<10, 12>(instruction);
+	const u32 uniformIndex = getBits<22, 2>(instruction);
+
+	if (loopLevel > 0) {
+		log("[Shader JIT] Detected nested loop. Might be broken?\n");
+	}
+
+	if (dest < recompilerPC) {
+		Helpers::panic("[Shader JIT] Detected backwards loop\n");
+	}
+
+	loopLevel++;
+
+	// Offset of the uniform
+	const auto& uniform = shader.intUniforms[uniformIndex];
+	const uintptr_t uniformOffset = uintptr_t(&uniform[0]) - uintptr_t(&shader);
+	// Offset of the loop register
+	const uintptr_t loopRegOffset = uintptr_t(&shader.loopCounter) - uintptr_t(&shader);
+
+	movzx(eax, byte[statePointer + uniformOffset]); // eax = loop iteration count
+	movzx(ecx, byte[statePointer + uniformOffset + sizeof(u8)]); // ecx = initial loop counter value
+	movzx(edx, byte[statePointer + uniformOffset + 2 * sizeof(u8)]); // edx = loop increment
+
+	add(eax, 1); // The iteration count is actually uniform.x + 1
+	mov(dword[statePointer + loopRegOffset], ecx); // Set loop counter
+	
+	// TODO: This might break if an instruction in a loop decides to yield...
+	push(rax);  // Push loop iteration counter
+	push(rdx);  // Push loop increment
+
+	Label loopStart;
+	L(loopStart);
+	compileUntil(shader, dest + 1);
+
+	const size_t stackOffsetOfLoopIncrement = 0;
+	const size_t stackOffsetOfIterationCounter = stackOffsetOfLoopIncrement + 8;
+
+	mov(ecx, dword[rsp + stackOffsetOfLoopIncrement]);   // ecx = Loop increment
+	add(dword[statePointer + loopRegOffset], ecx);       // Increment loop counter
+	sub(dword[rsp + stackOffsetOfIterationCounter], 1);  // Subtract 1 from loop iteration counter
+
+	jnz(loopStart);  // Back to loop start if not over
+	add(rsp, 16);
+	loopLevel--;
+}
+
+void ShaderEmitter::printLog(const PICAShader& shaderUnit) {
+	printf("PC: %04X\n", shaderUnit.pc);
+
+	for (int i = 0; i < shaderUnit.tempRegisters.size(); i++) {
+		const auto& r = shaderUnit.tempRegisters[i];
+		printf("t%d: (%.2f, %.2f, %.2f, %.2f)\n", i, r[0].toFloat64(), r[1].toFloat64(), r[2].toFloat64(), r[3].toFloat64());
+	}
+
+	for (int i = 0; i < shaderUnit.outputs.size(); i++) {
+		const auto& r = shaderUnit.outputs[i];
+		printf("o%d: (%.2f, %.2f, %.2f, %.2f)\n", i, r[0].toFloat64(), r[1].toFloat64(), r[2].toFloat64(), r[3].toFloat64());
+	}
+
+	printf("addr: (%d, %d)\n", shaderUnit.addrRegister[0], shaderUnit.addrRegister[1]);
+	printf("cmp: (%d, %d)\n", shaderUnit.cmpRegister[0], shaderUnit.cmpRegister[1]);
+}
+
+// As we mentioned above, this function is uber slow because we don't expect the shader JIT to call HLL functions in real scenarios
+// Aside from debugging code. So we don't care for this function to be performant or anything of the like.  It is quick and dirty
+// And mostly meant to be used for generating logs to diff the JIT and interpreter
+// We also don't support stack arguments atm unless it becomes actually necessary
+void ShaderEmitter::emitPrintLog(const PICAShader& shaderUnit) {
+	const uintptr_t pcOffset = uintptr_t(&shaderUnit.pc) - uintptr_t(&shaderUnit);
+	// Write back PC to print it
+	mov(dword[statePointer + pcOffset], recompilerPC);
+
+	// Push all registers because our JIT assumes everything is non volatile
+	push(rbp);
+	push(rax);
+	push(rbx);
+	push(rcx);
+	push(rdx);
+	push(rsi);
+	push(rdi);
+	push(r8);
+	push(r9);
+	push(r10);
+	push(r11);
+	push(r12);
+	push(r13);
+	push(r14);
+	push(r15);
+
+	mov(rbp, rsp);
+	// Reserve a bunch of stack space for Windows shadow stack et al, then force align rsp to 16 bytes to respect the ABI
+	sub(rsp, 64);
+	and_(rsp, ~0xF);
+
+	// Call function
+	mov(arg1, statePointer);
+	mov(rax, uintptr_t(printLog));
+	call(rax);
+
+	// Undo anything we did
+	mov(rsp, rbp);
+	pop(r15);
+	pop(r14);
+	pop(r13);
+	pop(r12);
+	pop(r11);
+	pop(r10);
+	pop(r9);
+	pop(r8);
+	pop(rdi);
+	pop(rsi);
+	pop(rdx);
+	pop(rcx);
+	pop(rbx);
+	pop(rax);
+	pop(rbp);
+}
+
+#endif

src/core/PICA/gpu.cpp

@@ -1,13 +1,52 @@
 #include "PICA/gpu.hpp"
 
 #include <array>
+#include <bitset>
 #include <cstdio>
+#include <cstddef>
 
 #include "PICA/float_types.hpp"
 #include "PICA/regs.hpp"
 
 using namespace Floats;
 
+// A representation of the output vertex as it comes out of the vertex shader, with padding and all
+struct OutputVertex {
+	using vec2f = OpenGL::Vector<f24, 2>;
+	using vec3f = OpenGL::Vector<f24, 3>;
+	using vec4f = OpenGL::Vector<f24, 4>;
+
+	union {
+		struct {
+			vec4f positions;   // Vertex position
+			vec4f quaternion;  // Quaternion specifying the normal/tangent frame (for fragment lighting)
+			vec4f colour;      // Vertex color
+			vec2f texcoord0;   // Texcoords for texture unit 0 (Only U and V, W is stored separately for 3D textures!)
+			vec2f texcoord1;   // Texcoords for TU 1
+			f24 texcoord0_w;   // W component for texcoord 0 if using a 3D texture
+			u32 padding;       // Unused
+
+			vec3f view;       // View vector (for fragment lighting)
+			u32 padding2;     // Unused
+			vec2f texcoord2;  // Texcoords for TU 2
+		} s;
+
+		// The software, non-accelerated vertex loader writes here and then reads specific components from the above struct
+		f24 raw[0x20];
+	};
+	OutputVertex() {}
+};
+#define ASSERT_POS(member, pos) static_assert(offsetof(OutputVertex, s.member) == pos * sizeof(f24), "OutputVertex struct is broken!");
+
+ASSERT_POS(positions, 0)
+ASSERT_POS(quaternion, 4)
+ASSERT_POS(colour, 8)
+ASSERT_POS(texcoord0, 12)
+ASSERT_POS(texcoord1, 14)
+ASSERT_POS(texcoord0_w, 16)
+ASSERT_POS(view, 18)
+ASSERT_POS(texcoord2, 22)
+
 GPU::GPU(Memory& mem) : mem(mem), renderer(*this, regs) {
 	vram = new u8[vramSize];
 	mem.setVRAM(vram); // Give the bus a pointer to our VRAM
@@ -16,6 +55,7 @@ GPU::GPU(Memory& mem) : mem(mem), renderer(*this, regs) {
 void GPU::reset() {
 	regs.fill(0);
 	shaderUnit.reset();
+	shaderJIT.reset();
 	std::memset(vram, 0, vramSize);
 
 	totalAttribCount = 0;
@@ -37,17 +77,32 @@ void GPU::reset() {
 	renderer.reset();
 }
 
+// Call the correct version of drawArrays based on whether this is an indexed draw (first template parameter)
+// And whether we are going to use the shader JIT (second template parameter)
 void GPU::drawArrays(bool indexed) {
-	if (indexed)
-		drawArrays<true>();
-	else
-		drawArrays<false>();
+	constexpr bool shaderJITEnabled = false; // TODO: Make a configurable option
+
+	if (indexed) {
+		if constexpr (ShaderJIT::isAvailable() && shaderJITEnabled)
+			drawArrays<true, true>();
+		else
+			drawArrays<true, false>();
+	} else {
+		if constexpr (ShaderJIT::isAvailable() && shaderJITEnabled)
+			drawArrays<false, true>();
+		else
+			drawArrays<false, false>();
+	}
 }
 
 static std::array<Vertex, Renderer::vertexBufferSize> vertices;
 
-template <bool indexed>
+template <bool indexed, bool useShaderJIT>
 void GPU::drawArrays() {
+	if constexpr (useShaderJIT) {
+		shaderJIT.prepare(shaderUnit.vs);
+	}
+
 	// Base address for vertex attributes
 	// The vertex base is always on a quadword boundary because the PICA does weird alignment shit any time possible
 	const u32 vertexBase = ((regs[PICA::InternalRegs::VertexAttribLoc] >> 1) & 0xfffffff) * 16;
@@ -56,11 +111,11 @@ void GPU::drawArrays() {
 	// Configures the type of primitive and the number of vertex shader outputs
 	const u32 primConfig = regs[PICA::InternalRegs::PrimitiveConfig];
 	const PICA::PrimType primType = static_cast<PICA::PrimType>(Helpers::getBits<8, 2>(primConfig));
-	if (primType == PICA::PrimType::TriangleFan) Helpers::panic("[PICA] Tried to draw unimplemented shape %d\n", primType);
 	if (vertexCount > Renderer::vertexBufferSize) Helpers::panic("[PICA] vertexCount > vertexBufferSize");
 
 	if ((primType == PICA::PrimType::TriangleList && vertexCount % 3) ||
-		(primType == PICA::PrimType::TriangleStrip && vertexCount < 3)) {
+		(primType == PICA::PrimType::TriangleStrip && vertexCount < 3) ||
+		(primType == PICA::PrimType::TriangleFan && vertexCount < 3)) {
 		Helpers::panic("Invalid vertex count for primitive. Type: %d, vert count: %d\n", primType, vertexCount);
 	}
 
@@ -83,9 +138,19 @@ void GPU::drawArrays() {
 	// Total number of input attributes to shader. Differs between GS and VS. Currently stubbed to the VS one, as we don't have geometry shaders.
 	const u32 inputAttrCount = (regs[PICA::InternalRegs::VertexShaderInputBufferCfg] & 0xf) + 1;
 	const u64 inputAttrCfg = getVertexShaderInputConfig();
+
+	// When doing indexed rendering, we have a cache of vertices to avoid processing attributes and shaders for a single vertex many times
+	constexpr bool vertexCacheEnabled = true;
+	constexpr size_t vertexCacheSize = 64;
+
+	struct {
+		std::bitset<vertexCacheSize> validBits{0};           // Shows which tags are valid. If the corresponding bit is 1, then there's an entry
+		std::array<u32, vertexCacheSize> ids;                // IDs (ie indices of the cached vertices in the 3DS vertex buffer)
+		std::array<u32, vertexCacheSize> bufferPositions;    // Positions of the cached vertices in our own vertex buffer
+	} vertexCache;
 		
 	for (u32 i = 0; i < vertexCount; i++) {
-		u32 vertexIndex; // Index of the vertex in the VBO
+		u32 vertexIndex; // Index of the vertex in the VBO for indexed rendering
 
 		if constexpr (!indexed) {
 			vertexIndex = i + regs[PICA::InternalRegs::VertexOffsetReg];
@@ -101,6 +166,24 @@ void GPU::drawArrays() {
 			}
 		}
 
+		// Check if the vertex corresponding to the index is in cache
+		if constexpr (indexed && vertexCacheEnabled) {
+			auto& cache = vertexCache;
+			size_t tag = vertexIndex % vertexCacheSize;
+			// Cache hit
+			if (cache.validBits[tag] && cache.ids[tag] == vertexIndex) {
+				vertices[i] = vertices[cache.bufferPositions[tag]];
+				continue;
+			}
+
+			// Cache miss. Set cache entry, fetch attributes and run shaders as normal
+			else {
+				cache.validBits[tag] = true;
+				cache.ids[tag] = vertexIndex;
+				cache.bufferPositions[tag] = i;
+			}
+		}
+
 		int attrCount = 0;
 		int buffer = 0; // Vertex buffer index for non-fixed attributes
 
@@ -194,10 +277,30 @@ void GPU::drawArrays() {
 			std::memcpy(&shaderUnit.vs.inputs[mapping], &currentAttributes[j], sizeof(vec4f));
 		}
 		
-		shaderUnit.vs.run();
-		std::memcpy(&vertices[i].position, &shaderUnit.vs.outputs[0], sizeof(vec4f));
-		std::memcpy(&vertices[i].colour, &shaderUnit.vs.outputs[1], sizeof(vec4f));
-		std::memcpy(&vertices[i].UVs, &shaderUnit.vs.outputs[2], 2 * sizeof(f24));
+        if constexpr (useShaderJIT) {
+			shaderJIT.run(shaderUnit.vs);
+		} else {
+			shaderUnit.vs.run();
+		}
+
+		OutputVertex out;
+		// Map shader outputs to fixed function properties
+		const u32 totalShaderOutputs = regs[PICA::InternalRegs::ShaderOutputCount] & 7;
+		for (int i = 0; i < totalShaderOutputs; i++) {
+			const u32 config = regs[PICA::InternalRegs::ShaderOutmap0 + i];
+
+			for (int j = 0; j < 4; j++) { // pls unroll
+				const u32 mapping = (config >> (j * 8)) & 0x1F;
+				out.raw[mapping] = shaderUnit.vs.outputs[i][j];
+			}
+		}
+
+		std::memcpy(&vertices[i].position, &out.s.positions, sizeof(vec4f));
+		std::memcpy(&vertices[i].colour, &out.s.colour, sizeof(vec4f));
+		std::memcpy(&vertices[i].texcoord0, &out.s.texcoord0, 2 * sizeof(f24));
+		std::memcpy(&vertices[i].texcoord1, &out.s.texcoord1, 2 * sizeof(f24));
+		std::memcpy(&vertices[i].texcoord0_w, &out.s.texcoord0_w, sizeof(f24));
+		std::memcpy(&vertices[i].texcoord2, &out.s.texcoord2, 2 * sizeof(f24));
 
 		//printf("(x, y, z, w) = (%f, %f, %f, %f)\n", (double)vertices[i].position.x(), (double)vertices[i].position.y(), (double)vertices[i].position.z(), (double)vertices[i].position.w());
 		//printf("(r, g, b, a) = (%f, %f, %f, %f)\n", (double)vertices[i].colour.r(), (double)vertices[i].colour.g(), (double)vertices[i].colour.b(), (double)vertices[i].colour.a());
@@ -220,11 +323,11 @@ Vertex GPU::getImmediateModeVertex() {
 	shaderUnit.vs.run();
 	std::memcpy(&v.position, &shaderUnit.vs.outputs[0], sizeof(vec4f));
 	std::memcpy(&v.colour, &shaderUnit.vs.outputs[1], sizeof(vec4f));
-	std::memcpy(&v.UVs, &shaderUnit.vs.outputs[2], 2 * sizeof(f24));
+	std::memcpy(&v.texcoord0, &shaderUnit.vs.outputs[2], 2 * sizeof(f24));
 
 	printf("(x, y, z, w) = (%f, %f, %f, %f)\n", (double)v.position.x(), (double)v.position.y(), (double)v.position.z(), (double)v.position.w());
 	printf("(r, g, b, a) = (%f, %f, %f, %f)\n", (double)v.colour.r(), (double)v.colour.g(), (double)v.colour.b(), (double)v.colour.a());
-	printf("(u, v      ) = (%f, %f)\n", v.UVs.u(), v.UVs.v());
+	printf("(u, v      ) = (%f, %f)\n", v.texcoord0.u(), v.texcoord0.v());
 
 	return v;
 }

src/core/PICA/pica_hash.cpp

@@ -0,0 +1,42 @@
+#include "PICA/pica_hash.hpp"
+#include "PICA/shader.hpp"
+
+#ifdef PANDA3DS_PICA_CITYHASH
+#include "cityhash.hpp"
+#elif defined(PANDA3DS_PICA_XXHASH3)
+#include "xxhash/xxhash.h"
+#else
+#error No hashing algorithm defined for the PICA (See pica_hash.hpp for details)
+#endif
+
+PICAHash::HashType PICAHash::computeHash(const char* buf, std::size_t len) {
+#ifdef PANDA3DS_PICA_CITYHASH
+	return CityHash::CityHash64(buf, len);
+#elif defined(PANDA3DS_PICA_XXHASH3)
+	return XXH3_64bits(buf, len);
+#else
+#error No hashing algorithm defined for PICAHash::computeHash
+#endif
+}
+
+PICAShader::Hash PICAShader::getCodeHash() {
+	// Hash the code again if the code changed
+	if (codeHashDirty) {
+		codeHashDirty = false;
+		lastCodeHash = PICAHash::computeHash((const char*)&loadedShader[0], loadedShader.size() * sizeof(loadedShader[0]));
+	}
+
+	// Return the code hash
+	return lastCodeHash;
+}
+
+PICAShader::Hash PICAShader::getOpdescHash() {
+	// Hash the code again if the operand descriptors changed
+	if (opdescHashDirty) {
+		opdescHashDirty = false;
+		lastOpdescHash = PICAHash::computeHash((const char*)&operandDescriptors[0], operandDescriptors.size() * sizeof(operandDescriptors[0]));
+	}
+
+	// Return the code hash
+	return lastOpdescHash;
+}

src/core/PICA/shader_interpreter.cpp

@@ -23,12 +23,15 @@ void PICAShader::run() {
 				break;
 			case ShaderOpcodes::DP3: dp3(instruction); break;
 			case ShaderOpcodes::DP4: dp4(instruction); break;
+			case ShaderOpcodes::DPHI: dphi(instruction); break;
 			case ShaderOpcodes::END: return; // Stop running shader
+			case ShaderOpcodes::EX2: ex2(instruction); break;
 			case ShaderOpcodes::FLR: flr(instruction); break;
 			case ShaderOpcodes::IFC: ifc(instruction); break;
 			case ShaderOpcodes::IFU: ifu(instruction); break;
 			case ShaderOpcodes::JMPC: jmpc(instruction); break;
 			case ShaderOpcodes::JMPU: jmpu(instruction); break;
+			case ShaderOpcodes::LG2: lg2(instruction); break;
 			case ShaderOpcodes::LOOP: loop(instruction); break;
 			case ShaderOpcodes::MAX: max(instruction); break;
 			case ShaderOpcodes::MIN: min(instruction); break;
@@ -325,6 +328,29 @@ void PICAShader::dp4(u32 instruction) {
 	}
 }
 
+void PICAShader::dphi(u32 instruction) {
+	const u32 operandDescriptor = operandDescriptors[instruction & 0x7f];
+	const u32 src1 = getBits<14, 5>(instruction);
+	u32 src2 = getBits<7, 7>(instruction);
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 dest = getBits<21, 5>(instruction);
+
+	src2 = getIndexedSource(src2, idx);
+	vec4f srcVec1 = getSourceSwizzled<1>(src1, operandDescriptor);
+	vec4f srcVec2 = getSourceSwizzled<2>(src2, operandDescriptor);
+
+	vec4f& destVector = getDest(dest);
+	// srcVec1[3] is supposed to be replaced with 1.0 in the dot product, so we just add srcVec2[3] without multiplying it with anything
+	f24 dot = srcVec1[0] * srcVec2[0] + srcVec1[1] * srcVec2[1] + srcVec1[2] * srcVec2[2] + srcVec2[3];
+
+	u32 componentMask = operandDescriptor & 0xf;
+	for (int i = 0; i < 4; i++) {
+		if (componentMask & (1 << i)) {
+			destVector[3 - i] = dot;
+		}
+	}
+}
+
 void PICAShader::rcp(u32 instruction) {
 	const u32 operandDescriptor = operandDescriptors[instruction & 0x7f];
 	const u32 src1 = getBits<12, 7>(instruction);
@@ -365,6 +391,46 @@ void PICAShader::rsq(u32 instruction) {
 	}
 }
 
+void PICAShader::ex2(u32 instruction) {
+	const u32 operandDescriptor = operandDescriptors[instruction & 0x7f];
+	u32 src = getBits<12, 7>(instruction);
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 dest = getBits<21, 5>(instruction);
+
+	src = getIndexedSource(src, idx);
+	vec4f srcVec = getSourceSwizzled<1>(src, operandDescriptor);
+
+	vec4f& destVector = getDest(dest);
+	f24 res = f24::fromFloat32(std::exp2(srcVec[0].toFloat32()));
+
+	u32 componentMask = operandDescriptor & 0xf;
+	for (int i = 0; i < 4; i++) {
+		if (componentMask & (1 << i)) {
+			destVector[3 - i] = res;
+		}
+	}
+}
+
+void PICAShader::lg2(u32 instruction) {
+	const u32 operandDescriptor = operandDescriptors[instruction & 0x7f];
+	u32 src = getBits<12, 7>(instruction);
+	const u32 idx = getBits<19, 2>(instruction);
+	const u32 dest = getBits<21, 5>(instruction);
+
+	src = getIndexedSource(src, idx);
+	vec4f srcVec = getSourceSwizzled<1>(src, operandDescriptor);
+
+	vec4f& destVector = getDest(dest);
+	f24 res = f24::fromFloat32(std::log2(srcVec[0].toFloat32()));
+
+	u32 componentMask = operandDescriptor & 0xf;
+	for (int i = 0; i < 4; i++) {
+		if (componentMask & (1 << i)) {
+			destVector[3 - i] = res;
+		}
+	}
+}
+
 void PICAShader::mad(u32 instruction) {
 	const u32 operandDescriptor = operandDescriptors[instruction & 0x1f];
 	const u32 src1 = getBits<17, 5>(instruction);

src/core/PICA/shader_unit.cpp

@@ -1,4 +1,5 @@
 #include "PICA/shader_unit.hpp"
+#include "cityhash.hpp"
 
 void ShaderUnit::reset() {
 	vs.reset();
@@ -30,4 +31,7 @@ void PICAShader::reset() {
 	addrRegister.x() = 0;
 	addrRegister.y() = 0;
 	loopCounter = 0;
+
+	codeHashDirty = true;
+	opdescHashDirty = true;
 }

src/core/renderer_gl/renderer_gl.cpp

@@ -11,45 +11,221 @@ using namespace Helpers;
 const char* vertexShader = R"(
 	#version 410 core
 	
-	layout (location = 0) in vec4 coords;
-	layout (location = 1) in vec4 vertexColour;
-	layout (location = 2) in vec2 inUVs_texture0;
+	layout (location = 0) in vec4 a_coords;
+	layout (location = 1) in vec4 a_vertexColour;
+	layout (location = 2) in vec2 a_texcoord0;
+	layout (location = 3) in vec2 a_texcoord1;
+	layout (location = 4) in float a_texcoord0_w;
+	layout (location = 5) in vec2 a_texcoord2;
 
-	out vec4 colour;
-	out vec2 tex0_UVs;
+	out vec4 v_colour;
+	out vec3 v_texcoord0;
+	out vec2 v_texcoord1;
+	out vec2 v_texcoord2;
+	flat out vec4 v_textureEnvColor[6];
+	flat out vec4 v_textureEnvBufferColor;
+
+	// TEV uniforms
+	uniform uint u_textureEnvColor[6];
+	uniform uint u_textureEnvBufferColor;
+
+	vec4 abgr8888ToVec4(uint abgr) {
+		const float scale = 1.0 / 255.0;
+
+		return scale * vec4(
+			float(abgr & 0xffu),
+			float((abgr >> 8) & 0xffu),
+			float((abgr >> 16) & 0xffu),
+			float(abgr >> 24)
+		);
+	}
 
 	void main() {
-		gl_Position = coords;
-		colour = vertexColour;
+		gl_Position = a_coords;
+		v_colour = a_vertexColour;
 
 		// Flip y axis of UVs because OpenGL uses an inverted y for texture sampling compared to the PICA
-		tex0_UVs = vec2(inUVs_texture0.x, 1.0 - inUVs_texture0.y);
+		v_texcoord0 = vec3(a_texcoord0.x, 1.0 - a_texcoord0.y, a_texcoord0_w);
+		v_texcoord1 = vec2(a_texcoord1.x, 1.0 - a_texcoord1.y);
+		v_texcoord2 = vec2(a_texcoord2.x, 1.0 - a_texcoord2.y);
+
+		for (int i = 0; i < 6; i++) {
+			v_textureEnvColor[i] = abgr8888ToVec4(u_textureEnvColor[i]);
+		}
+
+		v_textureEnvBufferColor = abgr8888ToVec4(u_textureEnvBufferColor);
 	}
 )";
 
 const char* fragmentShader = R"(
 	#version 410 core
 	
-	in vec4 colour;
-	in vec2 tex0_UVs;
+	in vec4 v_colour;
+	in vec3 v_texcoord0;
+	in vec2 v_texcoord1;
+	in vec2 v_texcoord2;
+	flat in vec4 v_textureEnvColor[6];
+	flat in vec4 v_textureEnvBufferColor;
 
 	out vec4 fragColour;
 
 	uniform uint u_alphaControl;
 	uniform uint u_textureConfig;
 
+	// TEV uniforms
+	uniform uint u_textureEnvSource[6];
+	uniform uint u_textureEnvOperand[6];
+	uniform uint u_textureEnvCombiner[6];
+	uniform uint u_textureEnvScale[6];
+	uniform uint u_textureEnvUpdateBuffer;
+
 	// Depth control uniforms
 	uniform float u_depthScale;
 	uniform float u_depthOffset;
 	uniform bool u_depthmapEnable;
 
 	uniform sampler2D u_tex0;
+	uniform sampler2D u_tex1;
+	uniform sampler2D u_tex2;
+
+	vec4 tevSources[16];
+	vec4 tevNextPreviousBuffer;
+	bool tevUnimplementedSourceFlag = false;
+
+	// OpenGL ES 1.1 reference pages for TEVs (this is what the PICA200 implements):
+	// https://registry.khronos.org/OpenGL-Refpages/es1.1/xhtml/glTexEnv.xml
+
+	vec4 tevFetchSource(uint src_id) {
+		if (src_id >= 6u && src_id < 13u) {
+			tevUnimplementedSourceFlag = true;
+		}
+
+		return tevSources[src_id];
+	}
+
+	vec4 tevGetColorAndAlphaSource(int tev_id, int src_id) {
+		vec4 result;
+
+		vec4 colorSource = tevFetchSource((u_textureEnvSource[tev_id] >> (src_id * 4)) & 15u);
+		vec4 alphaSource = tevFetchSource((u_textureEnvSource[tev_id] >> (src_id * 4 + 16)) & 15u);
+
+		uint colorOperand = (u_textureEnvOperand[tev_id] >> (src_id * 4)) & 15u;
+		uint alphaOperand = (u_textureEnvOperand[tev_id] >> (12 + src_id * 4)) & 7u;
+
+		// TODO: figure out what the undocumented values do
+		switch (colorOperand) {
+			case  0u: result.rgb = colorSource.rgb; break;            // Source color
+			case  1u: result.rgb = 1.0 - colorSource.rgb; break;      // One minus source color
+			case  2u: result.rgb = vec3(colorSource.a); break;        // Source alpha
+			case  3u: result.rgb = vec3(1.0 - colorSource.a); break;  // One minus source alpha
+			case  4u: result.rgb = vec3(colorSource.r); break;        // Source red
+			case  5u: result.rgb = vec3(1.0 - colorSource.r); break;  // One minus source red
+			case  8u: result.rgb = vec3(colorSource.g); break;        // Source green
+			case  9u: result.rgb = vec3(1.0 - colorSource.g); break;  // One minus source green
+			case 12u: result.rgb = vec3(colorSource.b); break;        // Source blue
+			case 13u: result.rgb = vec3(1.0 - colorSource.b); break;  // One minus source blue
+			default: break;
+		}
+
+		// TODO: figure out what the undocumented values do
+		switch (alphaOperand) {
+			case 0u: result.a = alphaSource.a; break;        // Source alpha
+			case 1u: result.a = 1.0 - alphaSource.a; break;  // One minus source alpha
+			case 2u: result.a = alphaSource.r; break;        // Source red
+			case 3u: result.a = 1.0 - alphaSource.r; break;  // One minus source red
+			case 4u: result.a = alphaSource.g; break;        // Source green
+			case 5u: result.a = 1.0 - alphaSource.g; break;  // One minus source green
+			case 6u: result.a = alphaSource.b; break;        // Source blue
+			case 7u: result.a = 1.0 - alphaSource.b; break;  // One minus source blue
+			default: break;
+		}
+
+		return result;
+	}
+
+	vec4 tevCalculateCombiner(int tev_id) {
+		vec4 source0 = tevGetColorAndAlphaSource(tev_id, 0);
+		vec4 source1 = tevGetColorAndAlphaSource(tev_id, 1);
+		vec4 source2 = tevGetColorAndAlphaSource(tev_id, 2);
+
+		uint colorCombine = u_textureEnvCombiner[tev_id] & 15u;
+		uint alphaCombine = (u_textureEnvCombiner[tev_id] >> 16) & 15u;
+
+		vec4 result = vec4(1.0);
+
+		// TODO: figure out what the undocumented values do
+		switch (colorCombine) {
+			case 0u: result.rgb = source0.rgb; break;                                       // Replace
+			case 1u: result.rgb = source0.rgb * source1.rgb; break;                         // Modulate
+			case 2u: result.rgb = min(vec3(1.0), source0.rgb + source1.rgb); break;         // Add
+			case 3u: result.rgb = clamp(source0.rgb + source1.rgb - 0.5, 0.0, 1.0); break;  // Add signed
+			case 4u: result.rgb = mix(source1.rgb, source0.rgb, source2.rgb); break;        // Interpolate
+			case 5u: result.rgb = max(source0.rgb - source1.rgb, 0.0); break;               // Subtract
+			case 6u: result.rgb = vec3(4.0 * dot(source0.rgb - 0.5 , source1.rgb - 0.5)); break;  // Dot3 RGB
+			case 7u: result     = vec4(4.0 * dot(source0.rgb - 0.5 , source1.rgb - 0.5)); break;  // Dot3 RGBA
+			case 8u: result.rgb = min(source0.rgb * source1.rgb + source2.rgb, 1.0); break;       // Multiply then add
+			case 9u: result.rgb = min((source0.rgb + source1.rgb) * source2.rgb, 1.0); break;     // Add then multiply
+			default: break;
+		}
+
+		if (colorCombine != 7u) { // The color combiner also writes the alpha channel in the "Dot3 RGBA" mode.
+			// TODO: figure out what the undocumented values do
+			// TODO: test if the alpha combiner supports all the same modes as the color combiner.
+			switch (alphaCombine) {
+				case 0u: result.a = source0.a; break;                                      // Replace
+				case 1u: result.a = source0.a * source1.a; break;                          // Modulate
+				case 2u: result.a = min(1.0, source0.a + source1.a); break;                // Add
+				case 3u: result.a = clamp(source0.a + source1.a - 0.5, 0.0, 1.0); break;   // Add signed
+				case 4u: result.a = mix(source1.a, source0.a, source2.a); break;           // Interpolate
+				case 5u: result.a = max(0.0, source0.a - source1.a); break;                // Subtract
+				case 8u: result.a = min(1.0, source0.a * source1.a + source2.a); break;    // Multiply then add
+				case 9u: result.a = min(1.0, (source0.a + source1.a) * source2.a); break;  // Add then multiply
+				default: break;
+			}
+		}
+
+		result.rgb *= float(1 << (u_textureEnvScale[tev_id] & 3u));
+		result.a   *= float(1 << ((u_textureEnvScale[tev_id] >> 16) & 3u));
+
+		return result;
+	}
 
 	void main() {
-		if ((u_textureConfig & 1u) != 0) { // Render texture 0 if enabled
-			fragColour = texture(u_tex0, tex0_UVs);
-		} else {
-			fragColour = colour;
+		vec2 tex2UV = (u_textureConfig & (1u << 13)) != 0u ? v_texcoord1 : v_texcoord2;
+
+		// TODO: what do invalid sources and disabled textures read as?
+		// And what does the "previous combiner" source read initially?
+		tevSources[0] = v_colour; // Primary/vertex color
+		tevSources[1] = vec4(vec3(0.5), 1.0); // Fragment primary color
+		tevSources[2] = vec4(vec3(0.5), 1.0); // Fragment secondary color
+		if ((u_textureConfig & 1u) != 0u) tevSources[3] = texture(u_tex0, v_texcoord0.xy);
+		if ((u_textureConfig & 2u) != 0u) tevSources[4] = texture(u_tex1, v_texcoord1);
+		if ((u_textureConfig & 4u) != 0u) tevSources[5] = texture(u_tex2, tex2UV);
+		tevSources[13] = vec4(0.0); // Previous buffer
+		tevSources[15] = vec4(0.0); // Previous combiner
+
+		tevNextPreviousBuffer = v_textureEnvBufferColor;
+
+		for (int i = 0; i < 6; i++) {
+			tevSources[14] = v_textureEnvColor[i]; // Constant color
+			tevSources[15] = tevCalculateCombiner(i);
+			tevSources[13] = tevNextPreviousBuffer;
+
+			if (i < 4) {
+				if ((u_textureEnvUpdateBuffer & (0x100u << i)) != 0u) {
+					tevNextPreviousBuffer.rgb = tevSources[15].rgb;
+				}
+
+				if ((u_textureEnvUpdateBuffer & (0x1000u << i)) != 0u) {
+					tevNextPreviousBuffer.a = tevSources[15].a;
+				}
+			}
+		}
+
+		fragColour = tevSources[15];
+
+		if (tevUnimplementedSourceFlag) {
+			 // fragColour = vec4(1.0, 0.0, 1.0, 1.0);
 		}
 
 		// Get original depth value by converting from [near, far] = [0, 1] to [-1, 1]
@@ -119,10 +295,10 @@ const char* displayVertexShader = R"(
 				vec2(1.0, 0.0), // Bottom-right
 				vec2(0.0, 1.0), // Top-left
 				vec2(0.0, 0.0)  // Bottom-left
-        );
+	);
 
 		gl_Position = positions[gl_VertexID];
-        UV = texcoords[gl_VertexID];
+	UV = texcoords[gl_VertexID];
 	}
 )";
 
@@ -179,16 +355,28 @@ void Renderer::initGraphicsContext() {
 	alphaControlLoc = OpenGL::uniformLocation(triangleProgram, "u_alphaControl");
 	texUnitConfigLoc = OpenGL::uniformLocation(triangleProgram, "u_textureConfig");
 
+	textureEnvSourceLoc = OpenGL::uniformLocation(triangleProgram, "u_textureEnvSource");
+	textureEnvOperandLoc = OpenGL::uniformLocation(triangleProgram, "u_textureEnvOperand");
+	textureEnvCombinerLoc = OpenGL::uniformLocation(triangleProgram, "u_textureEnvCombiner");
+	textureEnvColorLoc = OpenGL::uniformLocation(triangleProgram, "u_textureEnvColor");
+	textureEnvScaleLoc = OpenGL::uniformLocation(triangleProgram, "u_textureEnvScale");
+	textureEnvUpdateBufferLoc = OpenGL::uniformLocation(triangleProgram, "u_textureEnvUpdateBuffer");
+	textureEnvBufferColorLoc = OpenGL::uniformLocation(triangleProgram, "u_textureEnvBufferColor");
+
 	depthScaleLoc = OpenGL::uniformLocation(triangleProgram, "u_depthScale");
 	depthOffsetLoc = OpenGL::uniformLocation(triangleProgram, "u_depthOffset");
 	depthmapEnableLoc = OpenGL::uniformLocation(triangleProgram, "u_depthmapEnable");
-	glUniform1i(OpenGL::uniformLocation(triangleProgram, "u_tex0"), 0); // Init sampler object
+
+	// Init sampler objects
+	glUniform1i(OpenGL::uniformLocation(triangleProgram, "u_tex0"), 0);
+	glUniform1i(OpenGL::uniformLocation(triangleProgram, "u_tex1"), 1);
+	glUniform1i(OpenGL::uniformLocation(triangleProgram, "u_tex2"), 2);
 
 	OpenGL::Shader vertDisplay(displayVertexShader, OpenGL::Vertex);
 	OpenGL::Shader fragDisplay(displayFragmentShader, OpenGL::Fragment);
 	displayProgram.create({ vertDisplay, fragDisplay });
-	displayProgram.use();
 
+	displayProgram.use();
 	glUniform1i(OpenGL::uniformLocation(displayProgram, "u_texture"), 0); // Init sampler object
 
 	vbo.createFixedSize(sizeof(Vertex) * vertexBufferSize, GL_STREAM_DRAW);
@@ -202,23 +390,50 @@ void Renderer::initGraphicsContext() {
 	// Colour attribute
 	vao.setAttributeFloat<float>(1, 4, sizeof(Vertex), offsetof(Vertex, colour));
 	vao.enableAttribute(1);
-	// UV attribute
-	vao.setAttributeFloat<float>(2, 2, sizeof(Vertex), offsetof(Vertex, UVs));
+	// UV 0 attribute
+	vao.setAttributeFloat<float>(2, 2, sizeof(Vertex), offsetof(Vertex, texcoord0));
 	vao.enableAttribute(2);
+	// UV 1 attribute
+	vao.setAttributeFloat<float>(3, 2, sizeof(Vertex), offsetof(Vertex, texcoord1));
+	vao.enableAttribute(3);
+	// UV 0 W-component attribute
+	vao.setAttributeFloat<float>(4, 1, sizeof(Vertex), offsetof(Vertex, texcoord0_w));
+	vao.enableAttribute(4);
+	// UV 2 attribute
+	vao.setAttributeFloat<float>(5, 2, sizeof(Vertex), offsetof(Vertex, texcoord2));
+	vao.enableAttribute(5);
 
 	dummyVBO.create();
 	dummyVAO.create();
+
+	// Create texture and framebuffer for the 3DS screen
+	const u32 screenTextureWidth = 2 * 400; // Top screen is 400 pixels wide, bottom is 320
+	const u32 screenTextureHeight = 2 * 240; // Both screens are 240 pixels tall
+
+	auto prevTexture = OpenGL::getTex2D();
+	screenTexture.create(screenTextureWidth, screenTextureHeight, GL_RGBA8);
+	screenTexture.bind();
+	screenTexture.setMinFilter(OpenGL::Linear);
+	screenTexture.setMagFilter(OpenGL::Linear);
+	glBindTexture(GL_TEXTURE_2D, prevTexture);
+
+	screenFramebuffer.createWithDrawTexture(screenTexture);
+	screenFramebuffer.bind(OpenGL::DrawAndReadFramebuffer);
+
+	if (glCheckFramebufferStatus(GL_FRAMEBUFFER) != GL_FRAMEBUFFER_COMPLETE)
+		Helpers::panic("Incomplete framebuffer");
+
+	// TODO: This should not clear the framebuffer contents. It should load them from VRAM.
+	GLint oldViewport[4];
+	glGetIntegerv(GL_VIEWPORT, oldViewport);
+	OpenGL::setViewport(screenTextureWidth, screenTextureHeight);
+	OpenGL::setClearColor(0.0, 0.0, 0.0, 1.0);
+	OpenGL::clearColor();
+	OpenGL::setViewport(oldViewport[0], oldViewport[1], oldViewport[2], oldViewport[3]);
+
 	reset();
 }
 
-void Renderer::getGraphicsContext() {
-	OpenGL::disableScissor();
-
-	vbo.bind();
-	vao.bind();
-	triangleProgram.use();
-}
-
 // Set up the OpenGL blending context to match the emulated PICA
 void Renderer::setupBlending() {
 	const bool blendingEnabled = (regs[PICA::InternalRegs::ColourOperation] & (1 << 8)) != 0;
@@ -264,13 +479,91 @@ void Renderer::setupBlending() {
 	}
 }
 
+void Renderer::setupTextureEnvState() {
+	// TODO: Only update uniforms when the TEV config changed. Use an UBO potentially.
+
+	static constexpr std::array<u32, 6> ioBases = {
+	  PICA::InternalRegs::TexEnv0Source, PICA::InternalRegs::TexEnv1Source,
+	  PICA::InternalRegs::TexEnv2Source, PICA::InternalRegs::TexEnv3Source,
+	  PICA::InternalRegs::TexEnv4Source, PICA::InternalRegs::TexEnv5Source
+	};
+
+	u32 textureEnvSourceRegs[6];
+	u32 textureEnvOperandRegs[6];
+	u32 textureEnvCombinerRegs[6];
+	u32 textureEnvColourRegs[6];
+	u32 textureEnvScaleRegs[6];
+
+	for (int i = 0; i < 6; i++) {
+		const u32 ioBase = ioBases[i];
+
+		textureEnvSourceRegs[i] = regs[ioBase];
+		textureEnvOperandRegs[i] = regs[ioBase + 1];
+		textureEnvCombinerRegs[i] = regs[ioBase + 2];
+		textureEnvColourRegs[i] = regs[ioBase + 3];
+		textureEnvScaleRegs[i] = regs[ioBase + 4];
+	}
+
+	glUniform1uiv(textureEnvSourceLoc, 6, textureEnvSourceRegs);
+	glUniform1uiv(textureEnvOperandLoc, 6, textureEnvOperandRegs);
+	glUniform1uiv(textureEnvCombinerLoc, 6, textureEnvCombinerRegs);
+	glUniform1uiv(textureEnvColorLoc, 6, textureEnvColourRegs);
+	glUniform1uiv(textureEnvScaleLoc, 6, textureEnvScaleRegs);
+	glUniform1ui(textureEnvUpdateBufferLoc, regs[PICA::InternalRegs::TexEnvUpdateBuffer]);
+	glUniform1ui(textureEnvBufferColorLoc, regs[PICA::InternalRegs::TexEnvBufferColor]);
+}
+
+void Renderer::bindTexturesToSlots() {
+	static constexpr std::array<u32, 3> ioBases = {
+	  PICA::InternalRegs::Tex0BorderColor, PICA::InternalRegs::Tex1BorderColor, PICA::InternalRegs::Tex2BorderColor
+	};
+
+	for (int i = 0; i < 3; i++) {
+		if ((regs[PICA::InternalRegs::TexUnitCfg] & (1 << i)) == 0) {
+			continue;
+		}
+
+		const size_t ioBase = ioBases[i];
+
+		const u32 dim = regs[ioBase + 1];
+		const u32 config = regs[ioBase + 2];
+		const u32 height = dim & 0x7ff;
+		const u32 width = getBits<16, 11>(dim);
+		const u32 addr = (regs[ioBase + 4] & 0x0FFFFFFF) << 3;
+		u32 format = regs[ioBase + (i == 0 ? 13 : 5)] & 0xF;
+
+		glActiveTexture(GL_TEXTURE0 + i);
+		Texture targetTex(addr, static_cast<PICA::TextureFmt>(format), width, height, config);
+		OpenGL::Texture tex = getTexture(targetTex);
+		tex.bind();
+	}
+
+	glActiveTexture(GL_TEXTURE0);
+
+	// Update the texture unit configuration uniform if it changed
+	const u32 texUnitConfig = regs[PICA::InternalRegs::TexUnitCfg];
+	if (oldTexUnitConfig != texUnitConfig) {
+		oldTexUnitConfig = texUnitConfig;
+		glUniform1ui(texUnitConfigLoc, texUnitConfig);
+	}
+}
+
 void Renderer::drawVertices(PICA::PrimType primType, std::span<const Vertex> vertices) {
 	// The fourth type is meant to be "Geometry primitive". TODO: Find out what that is
 	static constexpr std::array<OpenGL::Primitives, 4> primTypes = {
-		OpenGL::Triangle, OpenGL::TriangleStrip, OpenGL::TriangleFan, OpenGL::Triangle
+	  OpenGL::Triangle, OpenGL::TriangleStrip, OpenGL::TriangleFan, OpenGL::Triangle
 	};
 	const auto primitiveTopology = primTypes[static_cast<usize>(primType)];
 
+    // TODO: We should implement a GL state tracker that tracks settings like scissor, blending, bound program, etc
+    // This way if we attempt to eg do multiple glEnable(GL_BLEND) calls in a row, it will say "Oh blending is already enabled"
+    // And not actually perform the very expensive driver call for it
+	OpenGL::disableScissor();
+
+	vbo.bind();
+	vao.bind();
+	triangleProgram.use();
+
 	// Adjust alpha test if necessary
 	const u32 alphaControl = regs[PICA::InternalRegs::AlphaTestConfig];
 	if (alphaControl != oldAlphaControl) {
@@ -313,26 +606,8 @@ void Renderer::drawVertices(PICA::PrimType primType, std::span<const Vertex> ver
 		glUniform1i(depthmapEnableLoc, depthMapEnable);
 	}
 
-	// Hack for rendering texture 1
-	if (regs[0x80] & 1) {
-		const u32 dim = regs[0x82];
-		const u32 config = regs[0x83];
-		const u32 height = dim & 0x7ff;
-		const u32 width = getBits<16, 11>(dim);
-		const u32 addr = (regs[0x85] & 0x0FFFFFFF) << 3;
-		const u32 format = regs[0x8E] & 0xF;
-
-		Texture targetTex(addr, static_cast<PICA::TextureFmt>(format), width, height, config);
-		OpenGL::Texture tex = getTexture(targetTex);
-		tex.bind();
-	}
-
-	// Update the texture unit configuration uniform if it changed
-	const u32 texUnitConfig = regs[PICA::InternalRegs::TexUnitCfg];
-	if (oldTexUnitConfig != texUnitConfig) {
-		oldTexUnitConfig = texUnitConfig;
-		glUniform1ui(texUnitConfigLoc, texUnitConfig);
-	}
+	setupTextureEnvState();
+	bindTexturesToSlots();
 
 	// TODO: Actually use this
 	float viewportWidth = f24::fromRaw(regs[PICA::InternalRegs::ViewportWidth] & 0xffffff).toFloat32() * 2.0;
@@ -366,20 +641,11 @@ constexpr u32 bottomScreenBuffer = 0x1f05dc00;
 
 // Quick hack to display top screen for now
 void Renderer::display() {
-	OpenGL::disableBlend();
-	OpenGL::disableDepth();
 	OpenGL::disableScissor();
 
-	OpenGL::bindScreenFramebuffer();
-	colourBufferCache[0].texture.bind();
-
-	displayProgram.use();
-
-	dummyVAO.bind();
-	OpenGL::setClearColor(0.0, 0.0, 1.0, 1.0); // Clear screen colour
-	OpenGL::clearColor();
-	OpenGL::setViewport(0, 240, 400, 240); // Actually draw our 3DS screen
-	OpenGL::draw(OpenGL::TriangleStrip, 4);
+	glBindFramebuffer(GL_DRAW_FRAMEBUFFER, 0);
+	screenFramebuffer.bind(OpenGL::ReadFramebuffer);
+	glBlitFramebuffer(0, 0, 400, 480, 0, 0, 400, 480, GL_COLOR_BUFFER_BIT, GL_LINEAR);
 }
 
 void Renderer::clearBuffer(u32 startAddress, u32 endAddress, u32 value, u32 control) {
@@ -455,4 +721,28 @@ void Renderer::displayTransfer(u32 inputAddr, u32 outputAddr, u32 inputSize, u32
 
 	const u32 outputWidth = outputSize & 0xffff;
 	const u32 outputGap = outputSize >> 16;
+
+	auto framebuffer = colourBufferCache.findFromAddress(inputAddr);
+	// If there's a framebuffer at this address, use it. Otherwise go back to our old hack and display framebuffer 0
+	// Displays are hard I really don't want to try implementing them because getting a fast solution is terrible
+	OpenGL::Texture& tex = framebuffer.has_value() ? framebuffer.value().get().texture : colourBufferCache[0].texture;
+
+	tex.bind();
+	screenFramebuffer.bind(OpenGL::DrawFramebuffer);
+
+	OpenGL::disableBlend();
+	OpenGL::disableDepth();
+	OpenGL::disableScissor();
+	displayProgram.use();
+
+	// Hack: Detect whether we are writing to the top or bottom screen by checking output gap and drawing to the proper part of the output texture
+	// We consider output gap == 320 to mean bottom, and anything else to mean top
+	if (outputGap == 320) {
+		OpenGL::setViewport(40, 0, 320, 240); // Bottom screen viewport
+	} else {
+		OpenGL::setViewport(0, 240, 400, 240); // Top screen viewport
+	}
+
+	dummyVAO.bind();
+	OpenGL::draw(OpenGL::TriangleStrip, 4); // Actually draw our 3DS screen
 }

src/core/services/gsp_gpu.cpp

@@ -324,9 +324,31 @@ void GPUService::memoryFill(u32* cmd) {
 	}
 }
 
+static u32 VaddrToPaddr(u32 addr) {
+	if (addr >= VirtualAddrs::VramStart && addr < (VirtualAddrs::VramStart + VirtualAddrs::VramSize)) [[likely]] {
+		return addr - VirtualAddrs::VramStart + PhysicalAddrs::VRAM;
+	}
+	
+	else if (addr >= VirtualAddrs::LinearHeapStartOld && addr < VirtualAddrs::LinearHeapEndOld) {
+		return addr - VirtualAddrs::LinearHeapStartOld + PhysicalAddrs::FCRAM;
+	}
+
+	else if (addr >= VirtualAddrs::LinearHeapStartNew && addr < VirtualAddrs::LinearHeapEndNew) {
+		return addr - VirtualAddrs::LinearHeapStartNew + PhysicalAddrs::FCRAM;
+	}
+
+	else if (addr == 0) {
+		return 0;
+	}
+
+	Helpers::warn("[GSP::GPU VaddrToPaddr] Unknown virtual address %08X", addr);
+	// Obviously garbage address
+	return 0xF3310932;
+}
+
 void GPUService::triggerDisplayTransfer(u32* cmd) {
-	const u32 inputAddr = cmd[1];
-	const u32 outputAddr = cmd[2];
+	const u32 inputAddr = VaddrToPaddr(cmd[1]);
+	const u32 outputAddr = VaddrToPaddr(cmd[2]);
 	const u32 inputSize = cmd[3];
 	const u32 outputSize = cmd[4];
 	const u32 flags = cmd[5];

src/emulator.cpp

@@ -62,10 +62,9 @@ void Emulator::step() {}
 void Emulator::render() {}
 
 void Emulator::run() {
-	while (running) {
-		gpu.getGraphicsContext();  // Give the GPU a rendering context
-		runFrame();                // Run 1 frame of instructions
-		gpu.display();             // Display graphics
+    while (running) {
+        runFrame(); // Run 1 frame of instructions
+        gpu.display(); // Display graphics
 
 		ServiceManager& srv = kernel.getServiceManager();
 

third_party/cityhash/cityhash.cpp

@@ -0,0 +1,340 @@
+// Copyright (c) 2011 Google, Inc.
+//
+// Permission is hereby granted, free of charge, to any person obtaining a copy
+// of this software and associated documentation files (the "Software"), to deal
+// in the Software without restriction, including without limitation the rights
+// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+// copies of the Software, and to permit persons to whom the Software is
+// furnished to do so, subject to the following conditions:
+//
+// The above copyright notice and this permission notice shall be included in
+// all copies or substantial portions of the Software.
+//
+// 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 AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+// THE SOFTWARE.
+//
+// CityHash, by Geoff Pike and Jyrki Alakuijala
+//
+// This file provides CityHash64() and related functions.
+//
+// It's probably possible to create even faster hash functions by
+// writing a program that systematically explores some of the space of
+// possible hash functions, by using SIMD instructions, or by
+// compromising on hash quality.
+
+#include <algorithm>
+#include <string.h> // for memcpy and memset
+#include "include/cityhash.hpp"
+#include "swap.hpp"
+
+// #include "config.h"
+#ifdef __GNUC__
+#define HAVE_BUILTIN_EXPECT 1
+#endif
+#ifdef COMMON_BIG_ENDIAN
+#define WORDS_BIGENDIAN 1
+#endif
+
+using namespace std;
+
+typedef uint8_t uint8;
+typedef uint32_t uint32;
+typedef uint64_t uint64;
+
+namespace CityHash {
+
+static uint64 UNALIGNED_LOAD64(const char* p) {
+    uint64 result;
+    memcpy(&result, p, sizeof(result));
+    return result;
+}
+
+static uint32 UNALIGNED_LOAD32(const char* p) {
+    uint32 result;
+    memcpy(&result, p, sizeof(result));
+    return result;
+}
+
+#ifdef WORDS_BIGENDIAN
+#define uint32_in_expected_order(x) (swap32(x))
+#define uint64_in_expected_order(x) (swap64(x))
+#else
+#define uint32_in_expected_order(x) (x)
+#define uint64_in_expected_order(x) (x)
+#endif
+
+#if !defined(LIKELY)
+#if HAVE_BUILTIN_EXPECT
+#define LIKELY(x) (__builtin_expect(!!(x), 1))
+#else
+#define LIKELY(x) (x)
+#endif
+#endif
+
+static uint64 Fetch64(const char* p) {
+    return uint64_in_expected_order(UNALIGNED_LOAD64(p));
+}
+
+static uint32 Fetch32(const char* p) {
+    return uint32_in_expected_order(UNALIGNED_LOAD32(p));
+}
+
+// Some primes between 2^63 and 2^64 for various uses.
+static const uint64 k0 = 0xc3a5c85c97cb3127ULL;
+static const uint64 k1 = 0xb492b66fbe98f273ULL;
+static const uint64 k2 = 0x9ae16a3b2f90404fULL;
+
+// Bitwise right rotate.  Normally this will compile to a single
+// instruction, especially if the shift is a manifest constant.
+static uint64 Rotate(uint64 val, int shift) {
+    // Avoid shifting by 64: doing so yields an undefined result.
+    return shift == 0 ? val : ((val >> shift) | (val << (64 - shift)));
+}
+
+static uint64 ShiftMix(uint64 val) {
+    return val ^ (val >> 47);
+}
+
+static uint64 HashLen16(uint64 u, uint64 v) {
+    return Hash128to64(uint128(u, v));
+}
+
+static uint64 HashLen16(uint64 u, uint64 v, uint64 mul) {
+    // Murmur-inspired hashing.
+    uint64 a = (u ^ v) * mul;
+    a ^= (a >> 47);
+    uint64 b = (v ^ a) * mul;
+    b ^= (b >> 47);
+    b *= mul;
+    return b;
+}
+
+static uint64 HashLen0to16(const char* s, std::size_t len) {
+    if (len >= 8) {
+        uint64 mul = k2 + len * 2;
+        uint64 a = Fetch64(s) + k2;
+        uint64 b = Fetch64(s + len - 8);
+        uint64 c = Rotate(b, 37) * mul + a;
+        uint64 d = (Rotate(a, 25) + b) * mul;
+        return HashLen16(c, d, mul);
+    }
+    if (len >= 4) {
+        uint64 mul = k2 + len * 2;
+        uint64 a = Fetch32(s);
+        return HashLen16(len + (a << 3), Fetch32(s + len - 4), mul);
+    }
+    if (len > 0) {
+        uint8 a = s[0];
+        uint8 b = s[len >> 1];
+        uint8 c = s[len - 1];
+        uint32 y = static_cast<uint32>(a) + (static_cast<uint32>(b) << 8);
+        uint32 z = static_cast<uint32>(len) + (static_cast<uint32>(c) << 2);
+        return ShiftMix(y * k2 ^ z * k0) * k2;
+    }
+    return k2;
+}
+
+// This probably works well for 16-byte strings as well, but it may be overkill
+// in that case.
+static uint64 HashLen17to32(const char* s, std::size_t len) {
+    uint64 mul = k2 + len * 2;
+    uint64 a = Fetch64(s) * k1;
+    uint64 b = Fetch64(s + 8);
+    uint64 c = Fetch64(s + len - 8) * mul;
+    uint64 d = Fetch64(s + len - 16) * k2;
+    return HashLen16(Rotate(a + b, 43) + Rotate(c, 30) + d, a + Rotate(b + k2, 18) + c, mul);
+}
+
+// Return a 16-byte hash for 48 bytes.  Quick and dirty.
+// Callers do best to use "random-looking" values for a and b.
+static pair<uint64, uint64> WeakHashLen32WithSeeds(uint64 w, uint64 x, uint64 y, uint64 z, uint64 a,
+                                                   uint64 b) {
+    a += w;
+    b = Rotate(b + a + z, 21);
+    uint64 c = a;
+    a += x;
+    a += y;
+    b += Rotate(a, 44);
+    return make_pair(a + z, b + c);
+}
+
+// Return a 16-byte hash for s[0] ... s[31], a, and b.  Quick and dirty.
+static pair<uint64, uint64> WeakHashLen32WithSeeds(const char* s, uint64 a, uint64 b) {
+    return WeakHashLen32WithSeeds(Fetch64(s), Fetch64(s + 8), Fetch64(s + 16), Fetch64(s + 24), a,
+                                  b);
+}
+
+// Return an 8-byte hash for 33 to 64 bytes.
+static uint64 HashLen33to64(const char* s, std::size_t len) {
+    uint64 mul = k2 + len * 2;
+    uint64 a = Fetch64(s) * k2;
+    uint64 b = Fetch64(s + 8);
+    uint64 c = Fetch64(s + len - 24);
+    uint64 d = Fetch64(s + len - 32);
+    uint64 e = Fetch64(s + 16) * k2;
+    uint64 f = Fetch64(s + 24) * 9;
+    uint64 g = Fetch64(s + len - 8);
+    uint64 h = Fetch64(s + len - 16) * mul;
+    uint64 u = Rotate(a + g, 43) + (Rotate(b, 30) + c) * 9;
+    uint64 v = ((a + g) ^ d) + f + 1;
+    uint64 w = Common::swap64((u + v) * mul) + h;
+    uint64 x = Rotate(e + f, 42) + c;
+    uint64 y = (Common::swap64((v + w) * mul) + g) * mul;
+    uint64 z = e + f + c;
+    a = Common::swap64((x + z) * mul + y) + b;
+    b = ShiftMix((z + a) * mul + d + h) * mul;
+    return b + x;
+}
+
+uint64 CityHash64(const char* s, std::size_t len) {
+    if (len <= 32) {
+        if (len <= 16) {
+            return HashLen0to16(s, len);
+        } else {
+            return HashLen17to32(s, len);
+        }
+    } else if (len <= 64) {
+        return HashLen33to64(s, len);
+    }
+
+    // For strings over 64 bytes we hash the end first, and then as we
+    // loop we keep 56 bytes of state: v, w, x, y, and z.
+    uint64 x = Fetch64(s + len - 40);
+    uint64 y = Fetch64(s + len - 16) + Fetch64(s + len - 56);
+    uint64 z = HashLen16(Fetch64(s + len - 48) + len, Fetch64(s + len - 24));
+    pair<uint64, uint64> v = WeakHashLen32WithSeeds(s + len - 64, len, z);
+    pair<uint64, uint64> w = WeakHashLen32WithSeeds(s + len - 32, y + k1, x);
+    x = x * k1 + Fetch64(s);
+
+    // Decrease len to the nearest multiple of 64, and operate on 64-byte chunks.
+    len = (len - 1) & ~static_cast<std::size_t>(63);
+    do {
+        x = Rotate(x + y + v.first + Fetch64(s + 8), 37) * k1;
+        y = Rotate(y + v.second + Fetch64(s + 48), 42) * k1;
+        x ^= w.second;
+        y += v.first + Fetch64(s + 40);
+        z = Rotate(z + w.first, 33) * k1;
+        v = WeakHashLen32WithSeeds(s, v.second * k1, x + w.first);
+        w = WeakHashLen32WithSeeds(s + 32, z + w.second, y + Fetch64(s + 16));
+        std::swap(z, x);
+        s += 64;
+        len -= 64;
+    } while (len != 0);
+    return HashLen16(HashLen16(v.first, w.first) + ShiftMix(y) * k1 + z,
+                     HashLen16(v.second, w.second) + x);
+}
+
+uint64 CityHash64WithSeed(const char* s, std::size_t len, uint64 seed) {
+    return CityHash64WithSeeds(s, len, k2, seed);
+}
+
+uint64 CityHash64WithSeeds(const char* s, std::size_t len, uint64 seed0, uint64 seed1) {
+    return HashLen16(CityHash64(s, len) - seed0, seed1);
+}
+
+// A subroutine for CityHash128().  Returns a decent 128-bit hash for strings
+// of any length representable in signed long.  Based on City and Murmur.
+static uint128 CityMurmur(const char* s, std::size_t len, uint128 seed) {
+    uint64 a = Uint128Low64(seed);
+    uint64 b = Uint128High64(seed);
+    uint64 c = 0;
+    uint64 d = 0;
+    signed long l = static_cast<long>(len) - 16;
+    if (l <= 0) { // len <= 16
+        a = ShiftMix(a * k1) * k1;
+        c = b * k1 + HashLen0to16(s, len);
+        d = ShiftMix(a + (len >= 8 ? Fetch64(s) : c));
+    } else { // len > 16
+        c = HashLen16(Fetch64(s + len - 8) + k1, a);
+        d = HashLen16(b + len, c + Fetch64(s + len - 16));
+        a += d;
+        do {
+            a ^= ShiftMix(Fetch64(s) * k1) * k1;
+            a *= k1;
+            b ^= a;
+            c ^= ShiftMix(Fetch64(s + 8) * k1) * k1;
+            c *= k1;
+            d ^= c;
+            s += 16;
+            l -= 16;
+        } while (l > 0);
+    }
+    a = HashLen16(a, c);
+    b = HashLen16(d, b);
+    return uint128(a ^ b, HashLen16(b, a));
+}
+
+uint128 CityHash128WithSeed(const char* s, std::size_t len, uint128 seed) {
+    if (len < 128) {
+        return CityMurmur(s, len, seed);
+    }
+
+    // We expect len >= 128 to be the common case.  Keep 56 bytes of state:
+    // v, w, x, y, and z.
+    pair<uint64, uint64> v, w;
+    uint64 x = Uint128Low64(seed);
+    uint64 y = Uint128High64(seed);
+    uint64 z = len * k1;
+    v.first = Rotate(y ^ k1, 49) * k1 + Fetch64(s);
+    v.second = Rotate(v.first, 42) * k1 + Fetch64(s + 8);
+    w.first = Rotate(y + z, 35) * k1 + x;
+    w.second = Rotate(x + Fetch64(s + 88), 53) * k1;
+
+    // This is the same inner loop as CityHash64(), manually unrolled.
+    do {
+        x = Rotate(x + y + v.first + Fetch64(s + 8), 37) * k1;
+        y = Rotate(y + v.second + Fetch64(s + 48), 42) * k1;
+        x ^= w.second;
+        y += v.first + Fetch64(s + 40);
+        z = Rotate(z + w.first, 33) * k1;
+        v = WeakHashLen32WithSeeds(s, v.second * k1, x + w.first);
+        w = WeakHashLen32WithSeeds(s + 32, z + w.second, y + Fetch64(s + 16));
+        std::swap(z, x);
+        s += 64;
+        x = Rotate(x + y + v.first + Fetch64(s + 8), 37) * k1;
+        y = Rotate(y + v.second + Fetch64(s + 48), 42) * k1;
+        x ^= w.second;
+        y += v.first + Fetch64(s + 40);
+        z = Rotate(z + w.first, 33) * k1;
+        v = WeakHashLen32WithSeeds(s, v.second * k1, x + w.first);
+        w = WeakHashLen32WithSeeds(s + 32, z + w.second, y + Fetch64(s + 16));
+        std::swap(z, x);
+        s += 64;
+        len -= 128;
+    } while (LIKELY(len >= 128));
+    x += Rotate(v.first + z, 49) * k0;
+    y = y * k0 + Rotate(w.second, 37);
+    z = z * k0 + Rotate(w.first, 27);
+    w.first *= 9;
+    v.first *= k0;
+    // If 0 < len < 128, hash up to 4 chunks of 32 bytes each from the end of s.
+    for (std::size_t tail_done = 0; tail_done < len;) {
+        tail_done += 32;
+        y = Rotate(x + y, 42) * k0 + v.second;
+        w.first += Fetch64(s + len - tail_done + 16);
+        x = x * k0 + w.first;
+        z += w.second + Fetch64(s + len - tail_done);
+        w.second += v.first;
+        v = WeakHashLen32WithSeeds(s + len - tail_done, v.first + z, v.second);
+        v.first *= k0;
+    }
+    // At this point our 56 bytes of state should contain more than
+    // enough information for a strong 128-bit hash.  We use two
+    // different 56-byte-to-8-byte hashes to get a 16-byte final result.
+    x = HashLen16(x, v.first);
+    y = HashLen16(y + z, w.first);
+    return uint128(HashLen16(x + v.second, w.second) + y, HashLen16(x + w.second, y + v.second));
+}
+
+uint128 CityHash128(const char* s, std::size_t len) {
+    return len >= 16
+               ? CityHash128WithSeed(s + 16, len - 16, uint128(Fetch64(s), Fetch64(s + 8) + k0))
+               : CityHash128WithSeed(s, len, uint128(k0, k1));
+}
+
+} // namespace CityHash

third_party/cityhash/include/cityhash.hpp

@@ -0,0 +1,111 @@
+// Copyright (c) 2011 Google, Inc.
+//
+// Permission is hereby granted, free of charge, to any person obtaining a copy
+// of this software and associated documentation files (the "Software"), to deal
+// in the Software without restriction, including without limitation the rights
+// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+// copies of the Software, and to permit persons to whom the Software is
+// furnished to do so, subject to the following conditions:
+//
+// The above copyright notice and this permission notice shall be included in
+// all copies or substantial portions of the Software.
+//
+// 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 AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+// THE SOFTWARE.
+//
+// CityHash, by Geoff Pike and Jyrki Alakuijala
+//
+// http://code.google.com/p/cityhash/
+//
+// This file provides a few functions for hashing strings.  All of them are
+// high-quality functions in the sense that they pass standard tests such
+// as Austin Appleby's SMHasher.  They are also fast.
+//
+// For 64-bit x86 code, on short strings, we don't know of anything faster than
+// CityHash64 that is of comparable quality.  We believe our nearest competitor
+// is Murmur3.  For 64-bit x86 code, CityHash64 is an excellent choice for hash
+// tables and most other hashing (excluding cryptography).
+//
+// For 64-bit x86 code, on long strings, the picture is more complicated.
+// On many recent Intel CPUs, such as Nehalem, Westmere, Sandy Bridge, etc.,
+// CityHashCrc128 appears to be faster than all competitors of comparable
+// quality.  CityHash128 is also good but not quite as fast.  We believe our
+// nearest competitor is Bob Jenkins' Spooky.  We don't have great data for
+// other 64-bit CPUs, but for long strings we know that Spooky is slightly
+// faster than CityHash on some relatively recent AMD x86-64 CPUs, for example.
+// Note that CityHashCrc128 is declared in citycrc.h.
+//
+// For 32-bit x86 code, we don't know of anything faster than CityHash32 that
+// is of comparable quality.  We believe our nearest competitor is Murmur3A.
+// (On 64-bit CPUs, it is typically faster to use the other CityHash variants.)
+//
+// Functions in the CityHash family are not suitable for cryptography.
+//
+// Please see CityHash's README file for more details on our performance
+// measurements and so on.
+//
+// WARNING: This code has been only lightly tested on big-endian platforms!
+// It is known to work well on little-endian platforms that have a small penalty
+// for unaligned reads, such as current Intel and AMD moderate-to-high-end CPUs.
+// It should work on all 32-bit and 64-bit platforms that allow unaligned reads;
+// bug reports are welcome.
+//
+// By the way, for some hash functions, given strings a and b, the hash
+// of a+b is easily derived from the hashes of a and b.  This property
+// doesn't hold for any hash functions in this file.
+
+#pragma once
+
+#include <cstddef>
+#include <cstdint>
+#include <utility>
+
+namespace CityHash {
+
+using uint128 = std::pair<uint64_t, uint64_t>;
+
+[[nodiscard]] inline uint64_t Uint128Low64(const uint128& x) {
+    return x.first;
+}
+[[nodiscard]] inline uint64_t Uint128High64(const uint128& x) {
+    return x.second;
+}
+
+// Hash function for a byte array.
+[[nodiscard]] uint64_t CityHash64(const char* buf, std::size_t len);
+
+// Hash function for a byte array.  For convenience, a 64-bit seed is also
+// hashed into the result.
+[[nodiscard]] uint64_t CityHash64WithSeed(const char* buf, std::size_t len, uint64_t seed);
+
+// Hash function for a byte array.  For convenience, two seeds are also
+// hashed into the result.
+[[nodiscard]] uint64_t CityHash64WithSeeds(const char* buf, std::size_t len, uint64_t seed0,
+                                           uint64_t seed1);
+
+// Hash function for a byte array.
+[[nodiscard]] uint128 CityHash128(const char* s, std::size_t len);
+
+// Hash function for a byte array.  For convenience, a 128-bit seed is also
+// hashed into the result.
+[[nodiscard]] uint128 CityHash128WithSeed(const char* s, std::size_t len, uint128 seed);
+
+// Hash 128 input bits down to 64 bits of output.
+// This is intended to be a reasonably good hash function.
+[[nodiscard]] inline uint64_t Hash128to64(const uint128& x) {
+    // Murmur-inspired hashing.
+    const uint64_t kMul = 0x9ddfea08eb382d69ULL;
+    uint64_t a = (Uint128Low64(x) ^ Uint128High64(x)) * kMul;
+    a ^= (a >> 47);
+    uint64_t b = (Uint128High64(x) ^ a) * kMul;
+    b ^= (b >> 47);
+    b *= kMul;
+    return b;
+}
+
+} // namespace CityHash

third_party/cityhash/swap.hpp

@@ -0,0 +1,718 @@
+// Copyright (c) 2012- PPSSPP Project / Dolphin Project.
+
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU General Public License as published by
+// the Free Software Foundation, version 2.0 or later versions.
+
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU General Public License 2.0 for more details.
+
+// A copy of the GPL 2.0 should have been included with the program.
+// If not, see http://www.gnu.org/licenses/
+
+// Official git repository and contact information can be found at
+// https://github.com/hrydgard/ppsspp and http://www.ppsspp.org/.
+
+#pragma once
+
+#include <type_traits>
+
+#if defined(_MSC_VER)
+#include <cstdlib>
+#endif
+#include <cstring>
+#include "helpers.hpp"
+
+// GCC
+#ifdef __GNUC__
+
+#if __BYTE_ORDER__ && (__BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) && !defined(COMMON_LITTLE_ENDIAN)
+#define COMMON_LITTLE_ENDIAN 1
+#elif __BYTE_ORDER__ && (__BYTE_ORDER__ == __ORDER_BIG_ENDIAN__) && !defined(COMMON_BIG_ENDIAN)
+#define COMMON_BIG_ENDIAN 1
+#endif
+
+// LLVM/clang
+#elif defined(__clang__)
+
+#if __LITTLE_ENDIAN__ && !defined(COMMON_LITTLE_ENDIAN)
+#define COMMON_LITTLE_ENDIAN 1
+#elif __BIG_ENDIAN__ && !defined(COMMON_BIG_ENDIAN)
+#define COMMON_BIG_ENDIAN 1
+#endif
+
+// MSVC
+#elif defined(_MSC_VER) && !defined(COMMON_BIG_ENDIAN) && !defined(COMMON_LITTLE_ENDIAN)
+
+#define COMMON_LITTLE_ENDIAN 1
+#endif
+
+// Worst case, default to little endian.
+#if !COMMON_BIG_ENDIAN && !COMMON_LITTLE_ENDIAN
+#define COMMON_LITTLE_ENDIAN 1
+#endif
+
+namespace Common {
+
+#ifdef _MSC_VER
+[[nodiscard]] inline u16 swap16(u16 data) noexcept {
+    return _byteswap_ushort(data);
+}
+[[nodiscard]] inline u32 swap32(u32 data) noexcept {
+    return _byteswap_ulong(data);
+}
+[[nodiscard]] inline u64 swap64(u64 data) noexcept {
+    return _byteswap_uint64(data);
+}
+#elif defined(__clang__) || defined(__GNUC__)
+#if defined(__Bitrig__) || defined(__OpenBSD__)
+// redefine swap16, swap32, swap64 as inline functions
+#undef swap16
+#undef swap32
+#undef swap64
+#endif
+[[nodiscard]] inline u16 swap16(u16 data) noexcept {
+    return __builtin_bswap16(data);
+}
+[[nodiscard]] inline u32 swap32(u32 data) noexcept {
+    return __builtin_bswap32(data);
+}
+[[nodiscard]] inline u64 swap64(u64 data) noexcept {
+    return __builtin_bswap64(data);
+}
+#else
+// Generic implementation.
+[[nodiscard]] inline u16 swap16(u16 data) noexcept {
+    return (data >> 8) | (data << 8);
+}
+[[nodiscard]] inline u32 swap32(u32 data) noexcept {
+    return ((data & 0xFF000000U) >> 24) | ((data & 0x00FF0000U) >> 8) |
+           ((data & 0x0000FF00U) << 8) | ((data & 0x000000FFU) << 24);
+}
+[[nodiscard]] inline u64 swap64(u64 data) noexcept {
+    return ((data & 0xFF00000000000000ULL) >> 56) | ((data & 0x00FF000000000000ULL) >> 40) |
+           ((data & 0x0000FF0000000000ULL) >> 24) | ((data & 0x000000FF00000000ULL) >> 8) |
+           ((data & 0x00000000FF000000ULL) << 8) | ((data & 0x0000000000FF0000ULL) << 24) |
+           ((data & 0x000000000000FF00ULL) << 40) | ((data & 0x00000000000000FFULL) << 56);
+}
+#endif
+
+[[nodiscard]] inline float swapf(float f) noexcept {
+    static_assert(sizeof(u32) == sizeof(float), "float must be the same size as uint32_t.");
+
+    u32 value;
+    std::memcpy(&value, &f, sizeof(u32));
+
+    value = swap32(value);
+    std::memcpy(&f, &value, sizeof(u32));
+
+    return f;
+}
+
+[[nodiscard]] inline double swapd(double f) noexcept {
+    static_assert(sizeof(u64) == sizeof(double), "double must be the same size as uint64_t.");
+
+    u64 value;
+    std::memcpy(&value, &f, sizeof(u64));
+
+    value = swap64(value);
+    std::memcpy(&f, &value, sizeof(u64));
+
+    return f;
+}
+
+} // Namespace Common
+
+template <typename T, typename F>
+struct swap_struct_t {
+    using swapped_t = swap_struct_t;
+
+protected:
+    T value;
+
+    static T swap(T v) {
+        return F::swap(v);
+    }
+
+public:
+    T swap() const {
+        return swap(value);
+    }
+    swap_struct_t() = default;
+    swap_struct_t(const T& v) : value(swap(v)) {}
+
+    template <typename S>
+    swapped_t& operator=(const S& source) {
+        value = swap(static_cast<T>(source));
+        return *this;
+    }
+
+    operator s8() const {
+        return static_cast<s8>(swap());
+    }
+    operator u8() const {
+        return static_cast<u8>(swap());
+    }
+    operator s16() const {
+        return static_cast<s16>(swap());
+    }
+    operator u16() const {
+        return static_cast<u16>(swap());
+    }
+    operator s32() const {
+        return static_cast<s32>(swap());
+    }
+    operator u32() const {
+        return static_cast<u32>(swap());
+    }
+    operator s64() const {
+        return static_cast<s64>(swap());
+    }
+    operator u64() const {
+        return static_cast<u64>(swap());
+    }
+    operator float() const {
+        return static_cast<float>(swap());
+    }
+    operator double() const {
+        return static_cast<double>(swap());
+    }
+
+    // +v
+    swapped_t operator+() const {
+        return +swap();
+    }
+    // -v
+    swapped_t operator-() const {
+        return -swap();
+    }
+
+    // v / 5
+    swapped_t operator/(const swapped_t& i) const {
+        return swap() / i.swap();
+    }
+    template <typename S>
+    swapped_t operator/(const S& i) const {
+        return swap() / i;
+    }
+
+    // v * 5
+    swapped_t operator*(const swapped_t& i) const {
+        return swap() * i.swap();
+    }
+    template <typename S>
+    swapped_t operator*(const S& i) const {
+        return swap() * i;
+    }
+
+    // v + 5
+    swapped_t operator+(const swapped_t& i) const {
+        return swap() + i.swap();
+    }
+    template <typename S>
+    swapped_t operator+(const S& i) const {
+        return swap() + static_cast<T>(i);
+    }
+    // v - 5
+    swapped_t operator-(const swapped_t& i) const {
+        return swap() - i.swap();
+    }
+    template <typename S>
+    swapped_t operator-(const S& i) const {
+        return swap() - static_cast<T>(i);
+    }
+
+    // v += 5
+    swapped_t& operator+=(const swapped_t& i) {
+        value = swap(swap() + i.swap());
+        return *this;
+    }
+    template <typename S>
+    swapped_t& operator+=(const S& i) {
+        value = swap(swap() + static_cast<T>(i));
+        return *this;
+    }
+    // v -= 5
+    swapped_t& operator-=(const swapped_t& i) {
+        value = swap(swap() - i.swap());
+        return *this;
+    }
+    template <typename S>
+    swapped_t& operator-=(const S& i) {
+        value = swap(swap() - static_cast<T>(i));
+        return *this;
+    }
+
+    // ++v
+    swapped_t& operator++() {
+        value = swap(swap() + 1);
+        return *this;
+    }
+    // --v
+    swapped_t& operator--() {
+        value = swap(swap() - 1);
+        return *this;
+    }
+
+    // v++
+    swapped_t operator++(int) {
+        swapped_t old = *this;
+        value = swap(swap() + 1);
+        return old;
+    }
+    // v--
+    swapped_t operator--(int) {
+        swapped_t old = *this;
+        value = swap(swap() - 1);
+        return old;
+    }
+    // Comparaison
+    // v == i
+    bool operator==(const swapped_t& i) const {
+        return swap() == i.swap();
+    }
+    template <typename S>
+    bool operator==(const S& i) const {
+        return swap() == i;
+    }
+
+    // v != i
+    bool operator!=(const swapped_t& i) const {
+        return swap() != i.swap();
+    }
+    template <typename S>
+    bool operator!=(const S& i) const {
+        return swap() != i;
+    }
+
+    // v > i
+    bool operator>(const swapped_t& i) const {
+        return swap() > i.swap();
+    }
+    template <typename S>
+    bool operator>(const S& i) const {
+        return swap() > i;
+    }
+
+    // v < i
+    bool operator<(const swapped_t& i) const {
+        return swap() < i.swap();
+    }
+    template <typename S>
+    bool operator<(const S& i) const {
+        return swap() < i;
+    }
+
+    // v >= i
+    bool operator>=(const swapped_t& i) const {
+        return swap() >= i.swap();
+    }
+    template <typename S>
+    bool operator>=(const S& i) const {
+        return swap() >= i;
+    }
+
+    // v <= i
+    bool operator<=(const swapped_t& i) const {
+        return swap() <= i.swap();
+    }
+    template <typename S>
+    bool operator<=(const S& i) const {
+        return swap() <= i;
+    }
+
+    // logical
+    swapped_t operator!() const {
+        return !swap();
+    }
+
+    // bitmath
+    swapped_t operator~() const {
+        return ~swap();
+    }
+
+    swapped_t operator&(const swapped_t& b) const {
+        return swap() & b.swap();
+    }
+    template <typename S>
+    swapped_t operator&(const S& b) const {
+        return swap() & b;
+    }
+    swapped_t& operator&=(const swapped_t& b) {
+        value = swap(swap() & b.swap());
+        return *this;
+    }
+    template <typename S>
+    swapped_t& operator&=(const S b) {
+        value = swap(swap() & b);
+        return *this;
+    }
+
+    swapped_t operator|(const swapped_t& b) const {
+        return swap() | b.swap();
+    }
+    template <typename S>
+    swapped_t operator|(const S& b) const {
+        return swap() | b;
+    }
+    swapped_t& operator|=(const swapped_t& b) {
+        value = swap(swap() | b.swap());
+        return *this;
+    }
+    template <typename S>
+    swapped_t& operator|=(const S& b) {
+        value = swap(swap() | b);
+        return *this;
+    }
+
+    swapped_t operator^(const swapped_t& b) const {
+        return swap() ^ b.swap();
+    }
+    template <typename S>
+    swapped_t operator^(const S& b) const {
+        return swap() ^ b;
+    }
+    swapped_t& operator^=(const swapped_t& b) {
+        value = swap(swap() ^ b.swap());
+        return *this;
+    }
+    template <typename S>
+    swapped_t& operator^=(const S& b) {
+        value = swap(swap() ^ b);
+        return *this;
+    }
+
+    template <typename S>
+    swapped_t operator<<(const S& b) const {
+        return swap() << b;
+    }
+    template <typename S>
+    swapped_t& operator<<=(const S& b) const {
+        value = swap(swap() << b);
+        return *this;
+    }
+
+    template <typename S>
+    swapped_t operator>>(const S& b) const {
+        return swap() >> b;
+    }
+    template <typename S>
+    swapped_t& operator>>=(const S& b) const {
+        value = swap(swap() >> b);
+        return *this;
+    }
+
+    // Member
+    /** todo **/
+
+    // Arithmetics
+    template <typename S, typename T2, typename F2>
+    friend S operator+(const S& p, const swapped_t v);
+
+    template <typename S, typename T2, typename F2>
+    friend S operator-(const S& p, const swapped_t v);
+
+    template <typename S, typename T2, typename F2>
+    friend S operator/(const S& p, const swapped_t v);
+
+    template <typename S, typename T2, typename F2>
+    friend S operator*(const S& p, const swapped_t v);
+
+    template <typename S, typename T2, typename F2>
+    friend S operator%(const S& p, const swapped_t v);
+
+    // Arithmetics + assignments
+    template <typename S, typename T2, typename F2>
+    friend S operator+=(const S& p, const swapped_t v);
+
+    template <typename S, typename T2, typename F2>
+    friend S operator-=(const S& p, const swapped_t v);
+
+    // Bitmath
+    template <typename S, typename T2, typename F2>
+    friend S operator&(const S& p, const swapped_t v);
+
+    // Comparison
+    template <typename S, typename T2, typename F2>
+    friend bool operator<(const S& p, const swapped_t v);
+
+    template <typename S, typename T2, typename F2>
+    friend bool operator>(const S& p, const swapped_t v);
+
+    template <typename S, typename T2, typename F2>
+    friend bool operator<=(const S& p, const swapped_t v);
+
+    template <typename S, typename T2, typename F2>
+    friend bool operator>=(const S& p, const swapped_t v);
+
+    template <typename S, typename T2, typename F2>
+    friend bool operator!=(const S& p, const swapped_t v);
+
+    template <typename S, typename T2, typename F2>
+    friend bool operator==(const S& p, const swapped_t v);
+};
+
+// Arithmetics
+template <typename S, typename T, typename F>
+S operator+(const S& i, const swap_struct_t<T, F> v) {
+    return i + v.swap();
+}
+
+template <typename S, typename T, typename F>
+S operator-(const S& i, const swap_struct_t<T, F> v) {
+    return i - v.swap();
+}
+
+template <typename S, typename T, typename F>
+S operator/(const S& i, const swap_struct_t<T, F> v) {
+    return i / v.swap();
+}
+
+template <typename S, typename T, typename F>
+S operator*(const S& i, const swap_struct_t<T, F> v) {
+    return i * v.swap();
+}
+
+template <typename S, typename T, typename F>
+S operator%(const S& i, const swap_struct_t<T, F> v) {
+    return i % v.swap();
+}
+
+// Arithmetics + assignments
+template <typename S, typename T, typename F>
+S& operator+=(S& i, const swap_struct_t<T, F> v) {
+    i += v.swap();
+    return i;
+}
+
+template <typename S, typename T, typename F>
+S& operator-=(S& i, const swap_struct_t<T, F> v) {
+    i -= v.swap();
+    return i;
+}
+
+// Logical
+template <typename S, typename T, typename F>
+S operator&(const S& i, const swap_struct_t<T, F> v) {
+    return i & v.swap();
+}
+
+template <typename S, typename T, typename F>
+S operator&(const swap_struct_t<T, F> v, const S& i) {
+    return static_cast<S>(v.swap() & i);
+}
+
+// Comparaison
+template <typename S, typename T, typename F>
+bool operator<(const S& p, const swap_struct_t<T, F> v) {
+    return p < v.swap();
+}
+template <typename S, typename T, typename F>
+bool operator>(const S& p, const swap_struct_t<T, F> v) {
+    return p > v.swap();
+}
+template <typename S, typename T, typename F>
+bool operator<=(const S& p, const swap_struct_t<T, F> v) {
+    return p <= v.swap();
+}
+template <typename S, typename T, typename F>
+bool operator>=(const S& p, const swap_struct_t<T, F> v) {
+    return p >= v.swap();
+}
+template <typename S, typename T, typename F>
+bool operator!=(const S& p, const swap_struct_t<T, F> v) {
+    return p != v.swap();
+}
+template <typename S, typename T, typename F>
+bool operator==(const S& p, const swap_struct_t<T, F> v) {
+    return p == v.swap();
+}
+
+template <typename T>
+struct swap_64_t {
+    static T swap(T x) {
+        return static_cast<T>(Common::swap64(x));
+    }
+};
+
+template <typename T>
+struct swap_32_t {
+    static T swap(T x) {
+        return static_cast<T>(Common::swap32(x));
+    }
+};
+
+template <typename T>
+struct swap_16_t {
+    static T swap(T x) {
+        return static_cast<T>(Common::swap16(x));
+    }
+};
+
+template <typename T>
+struct swap_float_t {
+    static T swap(T x) {
+        return static_cast<T>(Common::swapf(x));
+    }
+};
+
+template <typename T>
+struct swap_double_t {
+    static T swap(T x) {
+        return static_cast<T>(Common::swapd(x));
+    }
+};
+
+template <typename T>
+struct swap_enum_t {
+    static_assert(std::is_enum_v<T>);
+    using base = std::underlying_type_t<T>;
+
+public:
+    swap_enum_t() = default;
+    swap_enum_t(const T& v) : value(swap(v)) {}
+
+    swap_enum_t& operator=(const T& v) {
+        value = swap(v);
+        return *this;
+    }
+
+    operator T() const {
+        return swap(value);
+    }
+
+    explicit operator base() const {
+        return static_cast<base>(swap(value));
+    }
+
+protected:
+    T value{};
+    // clang-format off
+    using swap_t = std::conditional_t<
+        std::is_same_v<base, u16>, swap_16_t<u16>, std::conditional_t<
+        std::is_same_v<base, s16>, swap_16_t<s16>, std::conditional_t<
+        std::is_same_v<base, u32>, swap_32_t<u32>, std::conditional_t<
+        std::is_same_v<base, s32>, swap_32_t<s32>, std::conditional_t<
+        std::is_same_v<base, u64>, swap_64_t<u64>, std::conditional_t<
+        std::is_same_v<base, s64>, swap_64_t<s64>, void>>>>>>;
+    // clang-format on
+    static T swap(T x) {
+        return static_cast<T>(swap_t::swap(static_cast<base>(x)));
+    }
+};
+
+struct SwapTag {}; // Use the different endianness from the system
+struct KeepTag {}; // Use the same endianness as the system
+
+template <typename T, typename Tag>
+struct AddEndian;
+
+// KeepTag specializations
+
+template <typename T>
+struct AddEndian<T, KeepTag> {
+    using type = T;
+};
+
+// SwapTag specializations
+
+template <>
+struct AddEndian<u8, SwapTag> {
+    using type = u8;
+};
+
+template <>
+struct AddEndian<u16, SwapTag> {
+    using type = swap_struct_t<u16, swap_16_t<u16>>;
+};
+
+template <>
+struct AddEndian<u32, SwapTag> {
+    using type = swap_struct_t<u32, swap_32_t<u32>>;
+};
+
+template <>
+struct AddEndian<u64, SwapTag> {
+    using type = swap_struct_t<u64, swap_64_t<u64>>;
+};
+
+template <>
+struct AddEndian<s8, SwapTag> {
+    using type = s8;
+};
+
+template <>
+struct AddEndian<s16, SwapTag> {
+    using type = swap_struct_t<s16, swap_16_t<s16>>;
+};
+
+template <>
+struct AddEndian<s32, SwapTag> {
+    using type = swap_struct_t<s32, swap_32_t<s32>>;
+};
+
+template <>
+struct AddEndian<s64, SwapTag> {
+    using type = swap_struct_t<s64, swap_64_t<s64>>;
+};
+
+template <>
+struct AddEndian<float, SwapTag> {
+    using type = swap_struct_t<float, swap_float_t<float>>;
+};
+
+template <>
+struct AddEndian<double, SwapTag> {
+    using type = swap_struct_t<double, swap_double_t<double>>;
+};
+
+template <typename T>
+struct AddEndian<T, SwapTag> {
+    static_assert(std::is_enum_v<T>);
+    using type = swap_enum_t<T>;
+};
+
+// Alias LETag/BETag as KeepTag/SwapTag depending on the system
+#if COMMON_LITTLE_ENDIAN
+
+using LETag = KeepTag;
+using BETag = SwapTag;
+
+#else
+
+using BETag = KeepTag;
+using LETag = SwapTag;
+
+#endif
+
+// Aliases for LE types
+using u16_le = AddEndian<u16, LETag>::type;
+using u32_le = AddEndian<u32, LETag>::type;
+using u64_le = AddEndian<u64, LETag>::type;
+
+using s16_le = AddEndian<s16, LETag>::type;
+using s32_le = AddEndian<s32, LETag>::type;
+using s64_le = AddEndian<s64, LETag>::type;
+
+template <typename T>
+using enum_le = std::enable_if_t<std::is_enum_v<T>, typename AddEndian<T, LETag>::type>;
+
+using float_le = AddEndian<float, LETag>::type;
+using double_le = AddEndian<double, LETag>::type;
+
+// Aliases for BE types
+using u16_be = AddEndian<u16, BETag>::type;
+using u32_be = AddEndian<u32, BETag>::type;
+using u64_be = AddEndian<u64, BETag>::type;
+
+using s16_be = AddEndian<s16, BETag>::type;
+using s32_be = AddEndian<s32, BETag>::type;
+using s64_be = AddEndian<s64, BETag>::type;
+
+template <typename T>
+using enum_be = std::enable_if_t<std::is_enum_v<T>, typename AddEndian<T, BETag>::type>;
+
+using float_be = AddEndian<float, BETag>::type;
+using double_be = AddEndian<double, BETag>::type;

third_party/xbyak

@@ -0,0 +1 @@
+Subproject commit ad5276fa4d86d9579ff626719cea43c809a6cab8

third_party/xxhash/.gitattributes

@@ -0,0 +1,10 @@
+# Set the default behavior
+* text eol=lf
+
+# Explicitly declare source files
+*.c text eol=lf
+*.h text eol=lf
+
+# Denote files that should not be modified.
+*.odt binary
+

third_party/xxhash/.github/dependabot.yml

@@ -0,0 +1,7 @@
+# https://docs.github.com/en/code-security/dependabot/dependabot-version-updates/configuration-options-for-the-dependabot.yml-file
+version: 2
+updates:
+  - package-ecosystem: "github-actions"
+    directory: "/"
+    schedule:
+      interval: "weekly"

third_party/xxhash/.github/workflows/ci.yml

@@ -0,0 +1,589 @@
+# Known critical issues:
+# - AVX512 related tests are incomplete.  Because default environment of
+#   GitHub Actions doesn't guarantee to support AVX512.
+#   As of May 2021, they're using Xeon E5-2673 (which doesn't support
+#   AVX512) and Xeon Platinum 8171M (which supports AVX512).
+#   See also https://github.com/actions/runner/issues/1069
+#
+#   In this CI script, it always run `make default` which compiles xxHash
+#   with AVX512 intrinsics.  But if test runner doesn't support AVX512,
+#   it doesn't run `make check` which tests runtime error/consistency.
+#   It means that this test stochastically detects a failure in AVX512
+#   code path.
+#
+# Known issues:
+# - This test script ignores exit code of cppcheck which can see under
+#   Job:Linux x64 misc tests > cppcheck in the GitHub Actions report.
+#   Because xxHash project doesn't 100% follow their recommendation.
+#   Also sometimes it reports false positives.
+#
+# - GitHub Actions doesn't support Visual Studio 2015 and 2013.
+#   https://github.com/actions/virtual-environments/issues/387
+#
+# - Setup procedure for msys2 environment is painfully slow.  It takes
+#   3..5 minutes.
+#
+# Notes:
+# - You can investigate various information at the right pane of GitHub
+#   Actions report page.
+#
+#   | Item                      | Section in the right pane             |
+#   | ------------------------- | ------------------------------------- |
+#   | OS, VM                    | Set up job                            |
+#   | git repo, commit hash     | Run actions/checkout@v3               |
+#   | gcc, tools                | Environment info                      |
+#
+# - To fail earlier, order of tests in the same job are roughly sorted by
+#   elapsed time.
+#
+# Todos:
+# - [ ] Linux: Add native ARM runner.
+# - [ ] Linux: Add native ARM64 runner.
+# - [ ] Linux: Add native PPC64LE runner.
+# - [ ] Linux: Add native S390X runner.
+# - [ ] Windows: Add VS2013.
+# - [ ] Windows: Add VS2015.
+# - [ ] Windows: Add clang for msys2.
+# - [ ] Windows: Add native or emulated ARM runner.
+# - [ ] Windows: Add native or emulated ARM64 runner.
+
+
+# Name of the workflow is also displayed as a SVG badge
+name: xxHash CI tests
+
+on: [push, pull_request]
+
+concurrency:
+  group: fast-${{ github.ref }}
+  cancel-in-progress: true
+
+permissions:
+  contents: read
+
+jobs:
+  xxhash-c-compilers:
+    name: CC=${{ matrix.cc }}, ${{ matrix.os }}
+    strategy:
+      fail-fast: false  # 'false' means Don't stop matrix workflows even if some matrix entry fails.
+      matrix:
+        include: [
+          # You can access the following values via ${{ matrix.??? }}
+          #
+          #   pkgs    : apt-get package names.  It can include multiple package names which are delimited by space.
+          #   cc      : C compiler executable.
+          #   cxx     : C++ compiler executable for `make ctocpptest`.
+          #   avx512  : Set 'true' if compiler supports avx512.  Otherwise, set 'false'.
+          #   os      : GitHub Actions YAML workflow label.  See https://github.com/actions/virtual-environments#available-environments
+
+          # cc
+          { pkgs: '',                                  cc: cc,        cxx: c++,         avx512: 'true',  os: ubuntu-latest, },
+
+          # gcc
+          { pkgs: '',                                  cc: gcc,       cxx: g++,         avx512: 'true',  os: ubuntu-latest, },
+          { pkgs: 'gcc-12  g++-12  lib32gcc-12-dev',   cc: gcc-12,    cxx: g++-12,      avx512: 'true',  os: ubuntu-22.04,  },
+          { pkgs: 'gcc-11  g++-11  lib32gcc-11-dev',   cc: gcc-11,    cxx: g++-11,      avx512: 'true',  os: ubuntu-22.04,  },
+          { pkgs: 'gcc-10  g++-10  lib32gcc-10-dev',   cc: gcc-10,    cxx: g++-10,      avx512: 'true',  os: ubuntu-22.04,  },
+          { pkgs: 'gcc-9   g++-9   lib32gcc-9-dev',    cc: gcc-9,     cxx: g++-9,       avx512: 'true',  os: ubuntu-22.04,  },
+          { pkgs: 'gcc-8   g++-8   lib32gcc-8-dev',    cc: gcc-8,     cxx: g++-8,       avx512: 'true',  os: ubuntu-20.04,  },
+          { pkgs: 'gcc-7   g++-7   lib32gcc-7-dev',    cc: gcc-7,     cxx: g++-7,       avx512: 'true',  os: ubuntu-20.04,  },
+
+          # clang
+          { pkgs: '',                                  cc: clang,     cxx: clang++,     avx512: 'true',  os: ubuntu-latest, },
+          { pkgs: 'clang-12',                          cc: clang-12,  cxx: clang++-12,  avx512: 'true',  os: ubuntu-22.04,  },
+          { pkgs: 'clang-11',                          cc: clang-11,  cxx: clang++-11,  avx512: 'true',  os: ubuntu-22.04,  },
+          { pkgs: 'clang-10',                          cc: clang-10,  cxx: clang++-10,  avx512: 'true',  os: ubuntu-20.04,  },
+          { pkgs: 'clang-9',                           cc: clang-9,   cxx: clang++-9,   avx512: 'true',  os: ubuntu-20.04,  },
+          { pkgs: 'clang-8',                           cc: clang-8,   cxx: clang++-8,   avx512: 'true',  os: ubuntu-20.04,  },
+          { pkgs: 'clang-7',                           cc: clang-7,   cxx: clang++-7,   avx512: 'true',  os: ubuntu-20.04,  },
+          { pkgs: 'clang-6.0',                         cc: clang-6.0, cxx: clang++-6.0, avx512: 'true',  os: ubuntu-20.04,  },
+        ]
+
+    runs-on: ${{ matrix.os }}
+    env:                        # Set environment variables
+      # We globally set CC and CXX to improve compatibility with .travis.yml
+      CC: ${{ matrix.cc }}
+      CXX: ${{ matrix.cxx }}
+    steps:
+    - uses: actions/checkout@c85c95e3d7251135ab7dc9ce3241c5835cc595a9 # v3.5.3
+
+    - name: apt-get install
+      run: |
+        sudo apt-get update
+        sudo apt-get install gcc-multilib
+        sudo apt-get install ${{ matrix.pkgs }}
+
+    - name: Environment info
+      run: |
+        echo && type $CC && which $CC && $CC --version
+        echo && type $CXX && which $CXX && $CXX --version
+        echo && type make && make -v
+        echo && cat /proc/cpuinfo || echo /proc/cpuinfo is not present
+
+    - name: C90 + no-long-long compliance
+      if: always()
+      run: |
+        CFLAGS="-std=c90 -pedantic -Wno-long-long -Werror" make clean xxhsum
+
+    - name: C90 + XXH_NO_LONG_LONG
+      if: always()
+      run: |
+        # strict c90, with no long long support; resulting in no XXH64_* symbol
+        make clean c90test
+
+    - name: dispatch
+      if: always()
+      run: |
+        # removing sign conversion warnings due to a bug in gcc-5's definition of some AVX512 intrinsics
+        CFLAGS="-Werror" MOREFLAGS="-Wno-sign-conversion" make clean dispatch
+
+    - name: DISPATCH=1
+      if: always()
+      run: |
+        CFLAGS="-Wall -Wextra -Werror" make DISPATCH=1 clean default
+
+    - name: XXH_SIZE_OPT == 2
+      if: always()
+      run: |
+        CFLAGS="-Os -DXXH_SIZE_OPT=2 -Wall -Wextra -Werror" make clean xxhsum
+
+    - name: noxxh3test
+      if: always()
+      run: |
+        # check library can be compiled with XXH_NO_XXH3, resulting in no XXH3_* symbol
+        make clean noxxh3test
+
+    - name: nostreamtest
+      if: always()
+      run: |
+        # check library can be compiled with XXH_NO_STREAM, resulting in no streaming symbols
+        make clean noxxh3test
+
+    - name: make avx512f
+      if: ${{ matrix.avx512 == 'true' }}
+      run: |
+        CFLAGS="-O1 -mavx512f -Werror" make clean default
+
+    - name: test-all
+      if: always()
+      run: |
+        make clean test-all
+
+
+  ubuntu-consistency:
+    name: Linux x64 check results consistency
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@c85c95e3d7251135ab7dc9ce3241c5835cc595a9 # v3.5.3
+
+    - name: Environment info
+      run: |
+        echo && gcc --version
+        echo && make -v
+        echo && cat /proc/cpuinfo || echo /proc/cpuinfo is not present
+
+    - name: Scalar code path
+      run: |
+        CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR make clean check
+
+    - name: SSE2 code path
+      run: |
+        CPPFLAGS=-DXXH_VECTOR=XXH_SSE2 make clean check
+
+    - name: AVX2 code path
+      run: |
+        CPPFLAGS="-mavx2 -DXXH_VECTOR=XXH_AVX2" make clean check
+
+    # As for AVX512, see "Known critical issues" at the top of this file
+    - name: AVX512 code path
+      run: |
+        # Run "make check" if /proc/cpuinfo has flags for avx512.
+        grep -q "^flags.*\bavx512\b" /proc/cpuinfo && CPPFLAGS="-mavx512f -DXXH_VECTOR=XXH_AVX512" make clean check || (echo This test runner does not support AVX512. && $(exit 0))
+
+    - name: reroll code path (#240)
+      run: |
+        CPPFLAGS=-DXXH_REROLL=1 make clean check
+
+    - name: tests/bench
+      run: |
+        make -C tests/bench
+
+
+  ubuntu-misc:
+    name: Linux x64 misc tests
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@c85c95e3d7251135ab7dc9ce3241c5835cc595a9 # v3.5.3
+
+    - name: apt-get install
+      run: |
+        sudo apt-get update
+        sudo apt-get install valgrind cppcheck
+
+    - name: Environment info
+      run: |
+        echo && gcc --version
+        echo && clang --version
+        echo && valgrind --version
+        echo && cppcheck --version
+        echo && make -v
+        echo && cat /proc/cpuinfo || echo /proc/cpuinfo is not present
+
+    - name: cppcheck
+      run: |
+        # This test script ignores exit code of cppcheck.  See knowin issues
+        # at the top of this file.
+        make clean cppcheck || echo There are some cppcheck reports
+
+    - name: test-mem (valgrind)
+      run: |
+        make clean test-mem
+
+    - name: usan
+      run: |
+        make clean usan
+
+    - name: Lint Unicode in root-dir, cli/, tests/, tests/bench/, tests/collisions/.
+      run: |
+        make lint-unicode
+
+    - name: test-filename-escape
+      # See also issue #695 - https://github.com/Cyan4973/xxHash/issues/695
+      run: |
+        make clean test-filename-escape
+
+  ubuntu-cmake-unofficial:
+    name: Linux x64 cmake unofficial build test
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@c85c95e3d7251135ab7dc9ce3241c5835cc595a9 # v3.5.3
+
+    - name: Environment info
+      run: |
+        echo && gcc --version
+        echo && cmake --version
+        echo && make -v
+        echo && cat /proc/cpuinfo || echo /proc/cpuinfo is not present
+
+    - name: cmake
+      run: |
+        cd cmake_unofficial
+        mkdir build
+        cd build
+        cmake ..
+        CFLAGS=-Werror make
+        mkdir test_install_dir
+        DESTDIR=test_install_dir cmake --install .
+
+    - name: cmake minimum version v2.8.12 test
+      run: |
+        mkdir -p cmake_bins
+        cd cmake_bins
+        wget https://cmake.org/files/v2.8/cmake-2.8.12.2-Linux-i386.tar.gz
+        tar xzf cmake-2.8.12.2-Linux-i386.tar.gz
+        cd ../cmake_unofficial
+        mkdir -p build
+        cd build
+        ../../cmake_bins/cmake-2.8.12.2-Linux-i386/bin/cmake --version
+        ../../cmake_bins/cmake-2.8.12.2-Linux-i386/bin/cmake ..
+        ../../cmake_bins/cmake-2.8.12.2-Linux-i386/bin/cmake --build .
+        mkdir -p test_install_dir
+        DESTDIR=test_install_dir ../../cmake_bins/cmake-2.8.12.2-Linux-i386/bin/cmake --install .
+        rm -rf *
+        ../../cmake_bins/cmake-2.8.12.2-Linux-i386/bin/cmake -DCMAKE_BUILD_TYPE=Debug ..
+        ../../cmake_bins/cmake-2.8.12.2-Linux-i386/bin/cmake --build .
+
+
+
+  # Linux, { ARM, ARM64, PPC64LE, PPC64, S390X }
+  # All tests are using QEMU and gcc cross compiler.
+
+  qemu-consistency:
+    name: QEMU ${{ matrix.name }}
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false  # 'false' means Don't stop matrix workflows even if some matrix failed.
+      matrix:
+        include: [
+          { name: 'ARM',             xcc_pkg: gcc-arm-linux-gnueabi,        xcc: arm-linux-gnueabi-gcc,        xemu_pkg: qemu-system-arm,   xemu: qemu-arm-static,     os: ubuntu-latest, },
+          { name: 'AARCH64',         xcc_pkg: gcc-aarch64-linux-gnu,        xcc: aarch64-linux-gnu-gcc,        xemu_pkg: qemu-system-arm,   xemu: qemu-aarch64-static, os: ubuntu-latest, },
+          { name: 'PPC64LE',         xcc_pkg: gcc-powerpc64le-linux-gnu,    xcc: powerpc64le-linux-gnu-gcc,    xemu_pkg: qemu-system-ppc,   xemu: qemu-ppc64le-static, os: ubuntu-latest, },
+          { name: 'PPC64',           xcc_pkg: gcc-powerpc64-linux-gnu,      xcc: powerpc64-linux-gnu-gcc,      xemu_pkg: qemu-system-ppc,   xemu: qemu-ppc64-static,   os: ubuntu-latest, },
+          { name: 'S390X',           xcc_pkg: gcc-s390x-linux-gnu,          xcc: s390x-linux-gnu-gcc,          xemu_pkg: qemu-system-s390x, xemu: qemu-s390x-static,   os: ubuntu-latest, },
+          { name: 'MIPS',            xcc_pkg: gcc-mips-linux-gnu,           xcc: mips-linux-gnu-gcc,           xemu_pkg: qemu-system-mips,  xemu: qemu-mips-static,    os: ubuntu-latest, },
+          { name: 'M68K',            xcc_pkg: gcc-m68k-linux-gnu,           xcc: m68k-linux-gnu-gcc,           xemu_pkg: qemu-system-m68k,  xemu: qemu-m68k-static,    os: ubuntu-latest, },
+
+          { name: 'ARM, gcc-10',     xcc_pkg: gcc-10-arm-linux-gnueabi,     xcc: arm-linux-gnueabi-gcc-10,     xemu_pkg: qemu-system-arm,   xemu: qemu-arm-static,     os: ubuntu-20.04, },
+          { name: 'AARCH64, gcc-10', xcc_pkg: gcc-10-aarch64-linux-gnu,     xcc: aarch64-linux-gnu-gcc-10,     xemu_pkg: qemu-system-arm,   xemu: qemu-aarch64-static, os: ubuntu-20.04, },
+          { name: 'PPC64LE, gcc-10', xcc_pkg: gcc-10-powerpc64le-linux-gnu, xcc: powerpc64le-linux-gnu-gcc-10, xemu_pkg: qemu-system-ppc,   xemu: qemu-ppc64le-static, os: ubuntu-20.04, },
+          { name: 'PPC64, gcc-10',   xcc_pkg: gcc-10-powerpc64-linux-gnu,   xcc: powerpc64-linux-gnu-gcc-10,   xemu_pkg: qemu-system-ppc,   xemu: qemu-ppc64-static,   os: ubuntu-20.04, },
+          { name: 'S390X, gcc-10',   xcc_pkg: gcc-10-s390x-linux-gnu,       xcc: s390x-linux-gnu-gcc-10,       xemu_pkg: qemu-system-s390x, xemu: qemu-s390x-static,   os: ubuntu-20.04, },
+          { name: 'MIPS, gcc-10',    xcc_pkg: gcc-10-mips-linux-gnu,        xcc: mips-linux-gnu-gcc-10,        xemu_pkg: qemu-system-mips,  xemu: qemu-mips-static,    os: ubuntu-20.04, },
+
+          { name: 'ARM, gcc-9',      xcc_pkg: gcc-9-arm-linux-gnueabi,      xcc: arm-linux-gnueabi-gcc-9,      xemu_pkg: qemu-system-arm,   xemu: qemu-arm-static,     os: ubuntu-20.04, },
+          { name: 'AARCH64, gcc-9',  xcc_pkg: gcc-9-aarch64-linux-gnu,      xcc: aarch64-linux-gnu-gcc-9,      xemu_pkg: qemu-system-arm,   xemu: qemu-aarch64-static, os: ubuntu-20.04, },
+          { name: 'PPC64LE, gcc-9',  xcc_pkg: gcc-9-powerpc64le-linux-gnu,  xcc: powerpc64le-linux-gnu-gcc-9,  xemu_pkg: qemu-system-ppc,   xemu: qemu-ppc64le-static, os: ubuntu-20.04, },
+          { name: 'PPC64, gcc-9',    xcc_pkg: gcc-9-powerpc64-linux-gnu,    xcc: powerpc64-linux-gnu-gcc-9,    xemu_pkg: qemu-system-ppc,   xemu: qemu-ppc64-static,   os: ubuntu-20.04, },
+          { name: 'S390X, gcc-9',    xcc_pkg: gcc-9-s390x-linux-gnu,        xcc: s390x-linux-gnu-gcc-9,        xemu_pkg: qemu-system-s390x, xemu: qemu-s390x-static,   os: ubuntu-20.04, },
+          { name: 'MIPS, gcc-9',     xcc_pkg: gcc-9-mips-linux-gnu,         xcc: mips-linux-gnu-gcc-9,         xemu_pkg: qemu-system-mips,  xemu: qemu-mips-static,    os: ubuntu-20.04, },
+
+          { name: 'ARM, gcc-8',      xcc_pkg: gcc-8-arm-linux-gnueabi,      xcc: arm-linux-gnueabi-gcc-8,      xemu_pkg: qemu-system-arm,   xemu: qemu-arm-static,     os: ubuntu-20.04, },
+          # aarch64-linux-gnu-gcc-8 linker has an issue for LDFLAGS="-static"
+          # { name: 'AARCH64, gcc-8',  xcc_pkg: gcc-8-aarch64-linux-gnu,      xcc: aarch64-linux-gnu-gcc-8,      xemu_pkg: qemu-system-arm,   xemu: qemu-aarch64-static, os: ubuntu-20.04, },
+          { name: 'PPC64LE, gcc-8',  xcc_pkg: gcc-8-powerpc64le-linux-gnu,  xcc: powerpc64le-linux-gnu-gcc-8,  xemu_pkg: qemu-system-ppc,   xemu: qemu-ppc64le-static, os: ubuntu-20.04, },
+          { name: 'PPC64, gcc-8',    xcc_pkg: gcc-8-powerpc64-linux-gnu,    xcc: powerpc64-linux-gnu-gcc-8,    xemu_pkg: qemu-system-ppc,   xemu: qemu-ppc64-static,   os: ubuntu-20.04, },
+          { name: 'S390X, gcc-8',    xcc_pkg: gcc-8-s390x-linux-gnu,        xcc: s390x-linux-gnu-gcc-8,        xemu_pkg: qemu-system-s390x, xemu: qemu-s390x-static,   os: ubuntu-20.04, },
+          # ubuntu-20.04 fails to retrieve gcc-8-mips-linux-gnu for some reason.
+          # { name: 'MIPS, gcc-8',   xcc_pkg: gcc-8-mips-linux-gnu,         xcc: mips-linux-gnu-gcc-8,         xemu_pkg: qemu-system-mips,  xemu: qemu-mips-static,    os: ubuntu-20.04, },
+        ]
+    env:                        # Set environment variables
+      XCC: ${{ matrix.xcc }}
+      XEMU: ${{ matrix.xemu }}
+    steps:
+    - uses: actions/checkout@c85c95e3d7251135ab7dc9ce3241c5835cc595a9 # v3.5.3
+    - name: apt update & install (1)
+      run: |
+        sudo apt-get update
+        sudo apt-get install gcc-multilib g++-multilib qemu-utils qemu-user-static
+
+    - name: Environment info (1)
+      run: |
+        echo && apt-cache search "^gcc-" | grep "linux" | sort
+
+    - name: apt update & install (2)
+      run: |
+        sudo apt-get install ${{ matrix.xcc_pkg }} ${{ matrix.xemu_pkg }}
+
+    - name: Environment info (2)
+      run: |
+        echo && which $XCC
+        echo && $XCC --version
+        echo && $XCC -v  # Show built-in specs
+        echo && which $XEMU
+        echo && $XEMU --version
+
+    - name: ARM (XXH_VECTOR=[ scalar, NEON ])
+      if: ${{ startsWith(matrix.name, 'ARM') }}
+      run: |
+        CPPFLAGS="-DXXH_VECTOR=XXH_SCALAR" LDFLAGS="-static" CC=$XCC RUN_ENV=$XEMU make clean check
+        CPPFLAGS="-DXXH_VECTOR=XXH_NEON" CFLAGS="-O3 -march=armv7-a -fPIC -mfloat-abi=softfp -mfpu=neon-vfpv4" LDFLAGS="-static" CC=$XCC RUN_ENV=$XEMU make clean check
+
+    - name: AARCH64 (XXH_VECTOR=[ scalar, NEON, SVE ])
+      if: ${{ startsWith(matrix.name, 'AARCH64') }}
+      run: |
+        CPPFLAGS="-DXXH_VECTOR=XXH_SCALAR" LDFLAGS="-static" CC=$XCC RUN_ENV=$XEMU make clean check
+        CPPFLAGS="-DXXH_VECTOR=XXH_NEON" LDFLAGS="-static" CC=$XCC RUN_ENV=$XEMU make clean check
+        CPPFLAGS="-DXXH_VECTOR=XXH_SVE" LDFLAGS="-static" CC=$XCC RUN_ENV="$XEMU -cpu max,sve128=on,sve256=off,sve512=off,sve1024=off,sve2048=off" make clean check
+        CPPFLAGS="-DXXH_VECTOR=XXH_SVE" LDFLAGS="-static" CC=$XCC RUN_ENV="$XEMU -cpu max,sve128=on,sve256=on,sve512=off,sve1024=off,sve2048=off" make clean check
+        CPPFLAGS="-DXXH_VECTOR=XXH_SVE" LDFLAGS="-static" CC=$XCC RUN_ENV="$XEMU -cpu max,sve128=on,sve256=on,sve512=on,sve1024=off,sve2048=off" make clean check
+        CPPFLAGS="-DXXH_VECTOR=XXH_SVE" LDFLAGS="-static" CC=$XCC RUN_ENV="$XEMU -cpu max,sve128=on,sve256=on,sve512=on,sve1024=on,sve2048=off" make clean check
+        CPPFLAGS="-DXXH_VECTOR=XXH_SVE" LDFLAGS="-static" CC=$XCC RUN_ENV="$XEMU -cpu max,sve128=on,sve256=on,sve512=on,sve1024=on,sve2048=on" make clean check
+
+    - name: PPC64(LE) (XXH_VECTOR=[ scalar, VSX ])
+      if: ${{ startsWith(matrix.name, 'PPC64') }}
+      run: |
+        CPPFLAGS="-DXXH_VECTOR=XXH_SCALAR" LDFLAGS="-static" CC=$XCC RUN_ENV=$XEMU make clean check
+        CPPFLAGS="-DXXH_VECTOR=XXH_VSX" CFLAGS="-O3 -maltivec -mvsx -mpower8-vector -mcpu=power8" LDFLAGS="-static" CC=$XCC RUN_ENV=$XEMU make clean check
+
+    - name: S390X (XXH_VECTOR=[ scalar, VSX ])
+      if: ${{ startsWith(matrix.name, 'S390X') }}
+      run: |
+        CPPFLAGS="-DXXH_VECTOR=XXH_SCALAR" LDFLAGS="-static" CC=$XCC RUN_ENV=$XEMU make clean check
+        CPPFLAGS=-DXXH_VECTOR=XXH_VSX CFLAGS="-O3 -march=arch11 -mzvector" LDFLAGS="-static" CC=$XCC RUN_ENV=$XEMU make clean check
+
+    - name: MIPS (XXH_VECTOR=[ scalar ])
+      if: ${{ startsWith(matrix.name, 'MIPS') }}
+      run: |
+        LDFLAGS="-static" CC=$XCC RUN_ENV=$XEMU make clean check
+
+    - name: M68K (XXH_VECTOR=[ scalar ])
+      if: ${{ startsWith(matrix.name, 'M68K') }}
+      run: |
+        LDFLAGS="-static" CC=$XCC RUN_ENV=$XEMU make clean check
+
+
+  # macOS, { 11 }
+
+  macos-general:
+    name: ${{ matrix.system.os }}
+    runs-on: ${{ matrix.system.os }}
+    strategy:
+      fail-fast: false  # 'false' means Don't stop matrix workflows even if some matrix failed.
+      matrix:
+        system: [
+          { os: macos-11    },
+          { os: macos-12    },
+        ]
+    steps:
+    - uses: actions/checkout@c85c95e3d7251135ab7dc9ce3241c5835cc595a9 # v3.5.3
+
+    - name: Environment info
+      run: |
+        echo && clang --version
+        echo && sysctl -a | grep machdep.cpu   # cpuinfo
+
+    - name: make
+      run: |
+        CFLAGS="-Werror" make clean default
+
+    - name: make test
+      run: |
+        # test scenario where "stdout" is not the console
+        make clean test MOREFLAGS='-Werror' | tee
+
+
+  # Windows, { VC++2022, VC++2019, VC++2017 } x { x64, Win32, ARM, ARM64 }
+  #
+  # - Default shell for Windows environment is PowerShell Core.
+  #   https://docs.github.com/en/actions/reference/workflow-syntax-for-github-actions#using-a-specific-shell
+  #
+  # - "windows-2022" uses Visual Studio 2022.
+  #   https://github.com/actions/virtual-environments/blob/main/images/win/Windows2022-Readme.md#visual-studio-enterprise-2022
+  #
+  # - "windows-2019" uses Visual Studio 2019.
+  #   https://github.com/actions/virtual-environments/blob/main/images/win/Windows2019-Readme.md#visual-studio-enterprise-2019
+
+  windows-visualc-general:
+    name: ${{ matrix.system.vc }}, ${{ matrix.arch }}
+    runs-on: ${{ matrix.system.os }}   # Runs-on foreach value of strategy.matrix.system.os
+    strategy:
+      fail-fast: false  # 'false' means Don't stop matrix workflows even if some matrix failed.
+      matrix:
+        system: [
+          { os: windows-2022, vc: "VC++ 2022", clangcl: 'false', }, # CMake failed to configure clang-cl with VC++2022.
+          { os: windows-2019, vc: "VC++ 2019", clangcl: 'true',  },
+        ]
+        arch: [ x64, Win32, ARM, ARM64 ]
+    steps:
+    - uses: actions/checkout@c85c95e3d7251135ab7dc9ce3241c5835cc595a9 # v3.5.3
+
+    - name: Build ${{ matrix.system.os }}, ${{ matrix.arch }}
+      run: |
+        cd cmake_unofficial
+        mkdir build
+        cd build
+        cmake .. -DCMAKE_BUILD_TYPE=Release -A ${{ matrix.arch }} -DXXHASH_C_FLAGS="/WX"
+        cmake --build . --config Release
+
+    - name: Test
+      # Run benchmark for testing only if target arch is x64 or Win32.
+      if: ${{ matrix.arch == 'x64' || matrix.arch == 'Win32' }}
+      run: |
+        .\cmake_unofficial\build\Release\xxhsum.exe -bi1
+
+    - name: Build ${{ matrix.system.os }}, clang-cl, ${{ matrix.arch }}
+      if: ${{ matrix.system.clangcl == 'true' }}
+      run: |
+        cd cmake_unofficial
+        mkdir build-clang-cl
+        cd build-clang-cl
+        cmake .. -DCMAKE_BUILD_TYPE=Release -A x64 -DCMAKE_GENERATOR_TOOLSET=ClangCL
+        cmake --build . --config Release
+
+    - name: Test (clang-cl)
+      # Run benchmark for testing only if target arch is x64 or Win32.
+      if: ${{ matrix.system.clangcl == 'true' && ( matrix.arch == 'x64' || matrix.arch == 'Win32' ) }}
+      run: |
+        .\cmake_unofficial\build-clang-cl\Release\xxhsum.exe -bi1
+
+
+  # Windows, { mingw64, mingw32 }
+  #
+  # - Shell for msys2 is sh (msys2).  defaults.run.shell is for this setting.
+  #
+  # https://github.com/msys2/MINGW-packages/blob/master/.github/workflows/main.yml
+  # https://github.com/actions/starter-workflows/issues/95
+
+  windows-msys2-general:
+    name: Windows ${{ matrix.msystem }}
+    runs-on: windows-latest
+    strategy:
+      fail-fast: false  # 'false' means Don't stop matrix workflows even if some matrix failed.
+      matrix:
+        include: [
+          { msystem: mingw64, toolchain: mingw-w64-x86_64-toolchain },
+          { msystem: mingw32, toolchain: mingw-w64-i686-toolchain },
+        ]
+    defaults:
+      run:
+        shell: msys2 {0}
+    steps:
+      - uses: actions/checkout@c85c95e3d7251135ab7dc9ce3241c5835cc595a9 # v3.5.3
+      - uses: msys2/setup-msys2@7efe20baefed56359985e327d329042cde2434ff # v2
+        with:
+          msystem: MSYS
+          install: mingw-w64-i686-make ${{ matrix.toolchain }}
+          update: true
+
+      - name: Update
+        run: |
+          pacman --noconfirm -Suuy
+          pacman --noconfirm -Suu
+
+      - name: mingw64
+        if: ${{ matrix.msystem == 'mingw64' }}
+        run: |
+          PATH=/mingw64/bin:$PATH /mingw32/bin/mingw32-make clean test MOREFLAGS=-Werror
+          PATH=/mingw64/bin:$PATH /mingw32/bin/mingw32-make -C tests/bench
+          # Abort if result of "file ./xxhsum.exe" doesn't contain 'x86-64'.
+          # Expected output is "./xxhsum.exe: PE32+ executable (console) x86-64, for MS Windows"
+          file ./xxhsum.exe | grep -q 'x86-64' || $(exit 1)
+          ./xxhsum.exe --version
+
+      - name: mingw32
+        if: ${{ matrix.msystem == 'mingw32' }}
+        run: |
+          PATH=/mingw32/bin:$PATH /mingw32/bin/mingw32-make.exe clean test MOREFLAGS=-Werror
+          PATH=/mingw32/bin:$PATH /mingw32/bin/mingw32-make.exe -C tests/bench
+          # Abort if result of "file ./xxhsum.exe" doesn't contain '80386'.
+          # Expected output is "./xxhsum.exe: PE32 executable (console) Intel 80386, for MS Windows"
+          file ./xxhsum.exe | grep -q '80386' || $(exit 1)
+          ./xxhsum.exe --version
+
+  tipi-build-linux:
+    name: tipi.build project build and dependency resolution
+    runs-on: ubuntu-latest
+    container: tipibuild/tipi-ubuntu
+
+    env:
+      HOME: /root
+
+    steps:
+      - uses: actions/checkout@c85c95e3d7251135ab7dc9ce3241c5835cc595a9 # v3.5.3
+        # FIX: we currently need a full clone
+        with:
+          fetch-depth: '0'
+      - run: |
+          mkdir -p /usr/local/share/.tipi
+          # FIX: Hack for github action
+          git config --global --add safe.directory /usr/local/share/.tipi
+          git config --global --add safe.directory /__w/xxHash/xxHash/
+
+      # checking if the xxHash project builds and passes tests
+      - name: Build as project target linux-cxx17 (run test multiInclude)
+        run: tipi . --dont-upgrade --verbose --test multiInclude -t linux-cxx17
+
+      - name: Build as project target linux-cxx20 (run test multiInclude)
+        run: tipi . --dont-upgrade --verbose --test multiInclude -t linux-cxx20
+
+      - name: Cleanup project builds
+        run: rm -r ./build
+
+      # trying if pulling the dependency with tipi works properly
+      #
+      # note: the xxhashsum sources include xxhash using a #include "../xxhash.h"
+      # this defeats the purpose of this test AND because a bug post tipi v0.0.35
+      # because of source mirroring (not) supporting relative include to locations
+      # outside of the project tree.
+      #
+      # because of this we create a copy of the ./cli and apply some sed magic
+      # to make the includes proper 'library' includes to simulate what someone
+      # consuming xxHash would do
+      #
+      # e.g. turning #include "../xxhash.h" => #include <xxhash.h>
+      - name: Build as dependency
+        run: |
+          cp -a ./cli ./cli-tipi
+          cd ./cli-tipi
+          find ./ -type f -iname "*.c" | xargs sed -i 's;"../xxhash.h";<xxhash.h>;g'
+          tipi . --dont-upgrade --verbose -t linux-cxx17
+          ./build/linux-cxx17/bin/xsum_os_specific

third_party/xxhash/.gitignore

@@ -0,0 +1,49 @@
+# objects
+*.o
+*.obj
+*.s
+
+# libraries
+libxxhash.*
+!libxxhash.pc.in
+
+# Executables
+*.exe
+xxh32sum
+xxh64sum
+xxh128sum
+xxhsum
+xxhsum32
+xxhsum_privateXXH
+xxhsum_inlinedXXH
+dispatch
+tests/generate_unicode_test
+tests/sanity_test
+tests/sanity_test_vectors_generator
+
+# local conf
+.clang_complete
+
+# Mac OS-X artefacts
+*.dSYM
+.DS_Store
+
+# Wasm / emcc / emscripten artefacts
+*.html
+*.wasm
+*.js
+
+# CMake build directories
+build*/
+
+# project managers artifacts
+.projectile
+
+# analyzer artifacts
+infer-out
+
+# test artifacts
+.test*
+tmp*
+tests/*.unicode
+tests/unicode_test*

third_party/xxhash/.tipi/deps

@@ -0,0 +1 @@
+{ }

third_party/xxhash/.travis.yml

@@ -0,0 +1,142 @@
+language: c
+
+# Dump CPU info before start
+before_install:
+  - cat /proc/cpuinfo || echo /proc/cpuinfo is not present
+
+matrix:
+  fast_finish: true
+  include:
+
+    - name: General linux x64 tests
+      arch: amd64
+      addons:
+        apt:
+          packages:
+            - g++-multilib
+            - gcc-multilib
+            - cppcheck
+      script:
+        - make -B test-all
+        - make clean
+        - CFLAGS="-Werror" MOREFLAGS="-Wno-sign-conversion" make dispatch  # removing sign conversion warnings due to a bug in gcc-5's definition of some AVX512 intrinsics
+        - make clean
+        - CFLAGS="-O1 -mavx512f -Werror" make
+        - make clean
+        - CFLAGS="-Wall -Wextra -Werror" make DISPATCH=1
+        - make clean
+        - CFLAGS="-std=c90 -pedantic -Wno-long-long -Werror" make xxhsum  # check C90 + long long compliance
+        - make c90test # strict c90, with no long long support; resulting in no XXH64_* symbol
+        - make noxxh3test # check library can be compiled with XXH_NO_XXH3, resulting in no XXH3_* symbol
+
+
+    - name: Check results consistency on x64
+      arch: amd64
+      script:
+        - CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR make check   # Scalar code path
+        - make clean
+        - CPPFLAGS=-DXXH_VECTOR=XXH_SSE2 make check   # SSE2 code path
+        - make clean
+        - CPPFLAGS="-mavx2 -DXXH_VECTOR=XXH_AVX2" make check   # AVX2 code path
+        - make clean
+        - CPPFLAGS="-mavx512f -DXXH_VECTOR=XXH_AVX512" make check   # AVX512 code path
+        - make clean
+        - CPPFLAGS=-DXXH_REROLL=1 make check   # reroll code path (#240)
+        - make -C tests/bench
+
+    - name: macOS General Test
+      os: osx
+      compiler: clang
+      script:
+        - CFLAGS="-Werror" make   # test library build
+        - make clean
+        - make test MOREFLAGS='-Werror' | tee # test scenario where `stdout` is not the console
+
+    - name: ARM compilation and consistency checks (Qemu)
+      dist: xenial
+      arch: amd64
+      addons:
+        apt:
+          packages:
+            - qemu-system-arm
+            - qemu-user-static
+            - gcc-arm-linux-gnueabi
+            - libc6-dev-armel-cross
+      script:
+        # arm (32-bit)
+        - CC=arm-linux-gnueabi-gcc CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR LDFLAGS=-static RUN_ENV=qemu-arm-static make check   # Scalar code path
+        - make clean
+        # NEON (32-bit)
+        - CC=arm-linux-gnueabi-gcc CPPFLAGS=-DXXH_VECTOR=XXH_NEON CFLAGS="-O3 -march=armv7-a -fPIC -mfloat-abi=softfp -mfpu=neon-vfpv4" LDFLAGS=-static RUN_ENV=qemu-arm-static make check   # NEON code path
+
+    - name: aarch64 compilation and consistency checks
+      dist: xenial
+      arch: arm64
+      script:
+        # aarch64
+        - CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR make check   # Scalar code path
+        # NEON (64-bit)
+        - make clean
+        - CPPFLAGS=-DXXH_VECTOR=XXH_NEON make check   # NEON code path
+        # clang
+        - make clean
+        - CC=clang CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR make check   # Scalar code path
+        # clang + NEON
+        - make clean
+        - CC=clang CPPFLAGS=-DXXH_VECTOR=XXH_NEON make check   # NEON code path
+
+    # We need Bionic here because the QEMU versions shipped in the older repos
+    # do not support POWER8 emulation, and compiling QEMU from source is a pain.
+    - name: PowerPC + PPC64 compilation and consistency checks (Qemu on Bionic)
+      dist: bionic
+      arch: amd64
+      addons:
+        apt:
+          packages:
+            - qemu-system-ppc
+            - qemu-user-static
+            - gcc-powerpc-linux-gnu
+            - gcc-powerpc64-linux-gnu
+            - libc6-dev-powerpc-cross
+            - libc6-dev-ppc64-cross
+      script:
+        - CC=powerpc-linux-gnu-gcc RUN_ENV=qemu-ppc-static LDFLAGS=-static make check   # Scalar code path
+        - make clean
+        - CC=powerpc64-linux-gnu-gcc RUN_ENV=qemu-ppc64-static CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR CFLAGS="-O3" LDFLAGS="-static -m64" make check   # Scalar code path
+        # VSX code
+        - make clean
+        - CC=powerpc64-linux-gnu-gcc RUN_ENV="qemu-ppc64-static -cpu power8" CPPFLAGS=-DXXH_VECTOR=XXH_VSX CFLAGS="-O3 -maltivec -mvsx -mcpu=power8 -mpower8-vector" LDFLAGS="-static -m64" make check   # VSX code path
+        # altivec.h redefinition issue #426
+        - make clean
+        - CC=powerpc64-linux-gnu-gcc CPPFLAGS=-DXXH_VECTOR=XXH_VSX CFLAGS="-maltivec -mvsx -mcpu=power8 -mpower8-vector" make -C tests test_ppc_redefine
+
+    - name: PPC64LE compilation and consistency checks
+      dist: xenial
+      arch: ppc64le
+      script:
+        # Scalar (universal) code path
+        - CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR LDFLAGS=-static make check
+        # VSX code path (64-bit)
+        - make clean
+        - CPPFLAGS=-DXXH_VECTOR=XXH_VSX CFLAGS="-O3 -maltivec -mvsx -mpower8-vector -mcpu=power8" LDFLAGS="-static" make check
+        # altivec.h redefinition issue #426
+        - make clean
+        - CPPFLAGS=-DXXH_VECTOR=XXH_VSX CFLAGS="-maltivec -mvsx -mcpu=power8 -mpower8-vector" make -C tests test_ppc_redefine
+
+    - name: IBM s390x compilation and consistency checks
+      dist: bionic
+      arch: s390x
+      script:
+        # Scalar (universal) code path
+        - CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR LDFLAGS=-static make check
+        # s390x code path (64-bit)
+        - make clean
+        - CPPFLAGS=-DXXH_VECTOR=XXH_VSX CFLAGS="-O3 -march=arch11 -mzvector" LDFLAGS="-static" make check
+
+    - name: cmake build test
+      script:
+        - cd cmake_unofficial
+        - mkdir build
+        - cd build
+        - cmake ..
+        - CFLAGS=-Werror make

third_party/xxhash/CHANGELOG

@@ -0,0 +1,71 @@
+v0.8.1
+- perf : much improved performance for XXH3 streaming variants, notably on gcc and msvc
+- perf : improved XXH64 speed and latency on small inputs
+- perf : small XXH32 speed and latency improvement on small inputs of random size
+- perf : minor stack usage improvement for XXH32 and XXH64
+- api  : new experimental variants XXH3_*_withSecretandSeed()
+- api  : update XXH3_generateSecret(), can no generate secret of any size (>= XXH3_SECRET_SIZE_MIN)
+- cli  : xxhsum can now generate and check XXH3 checksums, using command `-H3`
+- build: can build xxhash without XXH3, with new build macro XXH_NO_XXH3
+- build: fix xxh_x86dispatch build with MSVC, by @apankrat
+- build: XXH_INLINE_ALL can always be used safely, even after XXH_NAMESPACE or a previous XXH_INLINE_ALL
+- build: improved PPC64LE vector support, by @mpe
+- install: fix pkgconfig, by @ellert
+- install: compatibility with Haiku, by @Begasus
+- doc  : code comments made compatible with doxygen, by @easyaspi314
+- misc : XXH_ACCEPT_NULL_INPUT_POINTER is no longer necessary, all functions can accept NULL input pointers, as long as size == 0
+- misc : complete refactor of CI tests on Github Actions, offering much larger coverage, by @t-mat
+- misc : xxhsum code base split into multiple specialized units, within directory cli/, by @easyaspi314
+
+v0.8.0
+- api : stabilize XXH3
+- cli : xxhsum can parse BSD-style --check lines, by @WayneD
+- cli : `xxhsum -` accepts console input, requested by @jaki
+- cli : xxhsum accepts -- separator, by @jaki
+- cli : fix : print correct default algo for symlinked helpers, by @martinetd
+- install: improved pkgconfig script, allowing custom install locations, requested by @ellert
+
+v0.7.4
+- perf: automatic vector detection and selection at runtime (`xxh_x86dispatch.h`), initiated by @easyaspi314
+- perf: added AVX512 support, by @gzm55
+- api : new: secret generator `XXH_generateSecret()`, suggested by @koraa
+- api : fix: XXH3_state_t is movable, identified by @koraa
+- api : fix: state is correctly aligned in AVX mode (unlike `malloc()`), by @easyaspi314
+- api : fix: streaming generated wrong values in some combination of random ingestion lengths, reported by @WayneD
+- cli : fix unicode print on Windows, by @easyaspi314
+- cli : can `-c` check file generated by sfv
+- build: `make DISPATCH=1` generates `xxhsum` and `libxxhash` with runtime vector detection (x86/x64 only)
+- install: cygwin installation support
+- doc : Cryptol specification of XXH32 and XXH64, by @weaversa
+
+v0.7.3
+- perf: improved speed for large inputs (~+20%)
+- perf: improved latency for small inputs (~10%)
+- perf: s390x Vectorial code, by @easyaspi314
+- cli: improved support for Unicode filenames on Windows, thanks to @easyaspi314 and @t-mat
+- api: `xxhash.h` can now be included in any order, with and without `XXH_STATIC_LINKING_ONLY` and `XXH_INLINE_ALL`
+- build: xxHash's implementation transferred into `xxhash.h`. No more need to have `xxhash.c` in the `/include` directory for `XXH_INLINE_ALL` to work
+- install: created pkg-config file, by @bket
+- install: VCpkg installation instructions, by @LilyWangL
+- doc: Highly improved code documentation, by @easyaspi314
+- misc: New test tool in `/tests/collisions`: brute force collision tester for 64-bit hashes
+
+v0.7.2
+- Fixed collision ratio of `XXH128` for some specific input lengths, reported by @svpv
+- Improved `VSX` and `NEON` variants, by @easyaspi314
+- Improved performance of scalar code path (`XXH_VECTOR=0`), by @easyaspi314
+- `xxhsum`: can generate 128-bit hashes with the `-H2` option (note: for experimental purposes only! `XXH128` is not yet frozen)
+- `xxhsum`: option `-q` removes status notifications
+
+v0.7.1
+- Secret first: the algorithm computation can be altered by providing a "secret", which is any blob of bytes, of size >= `XXH3_SECRET_SIZE_MIN`.
+- `seed` is still available, and acts as a secret generator
+- updated `ARM NEON` variant by @easyaspi314
+- Streaming implementation is available
+- Improve compatibility and performance with Visual Studio, with help from @aras-p
+- Better integration when using `XXH_INLINE_ALL`: do not pollute host namespace, use its own macros, such as `XXH_ASSERT()`, `XXH_ALIGN`, etc.
+- 128-bit variant provides helper functions for comparison of hashes.
+- Better `clang` generation of `rotl` instruction, thanks to @easyaspi314
+- `XXH_REROLL` build macro to reduce binary size, by @easyaspi314
+- Improved `cmake` script, by @Mezozoysky
+- Full benchmark program provided in `/tests/bench`

third_party/xxhash/Doxyfile

@@ -0,0 +1,61 @@
+# Doxygen config for xxHash
+DOXYFILE_ENCODING      = UTF-8
+
+PROJECT_NAME           = "xxHash"
+PROJECT_NUMBER         = "0.8.1"
+PROJECT_BRIEF          = "Extremely fast non-cryptographic hash function"
+OUTPUT_DIRECTORY       = doxygen
+OUTPUT_LANGUAGE        = English
+
+# We already separate the internal docs.
+INTERNAL_DOCS          = YES
+# Consistency
+SORT_MEMBER_DOCS       = NO
+BRIEF_MEMBER_DESC      = YES
+REPEAT_BRIEF           = YES
+
+# Warnings
+QUIET                  = YES
+# Until we document everything
+WARN_IF_UNDOCUMENTED   = NO
+
+# TODO: Add the other files. It is just xxhash.h for now.
+FILE_PATTERNS          = xxhash.h xxh_x86dispatch.c
+# Note: xxHash's source files are technically ASCII only.
+INPUT_ENCODING         = UTF-8
+TAB_SIZE               = 4
+MARKDOWN_SUPPORT       = YES
+
+# xxHash is a C library
+OPTIMIZE_OUTPUT_FOR_C  = YES
+# So we can document the internals
+EXTRACT_STATIC         = YES
+# Document the macros
+MACRO_EXPANSION        = YES
+EXPAND_ONLY_PREDEF     = YES
+# Predefine some macros to clean up the output.
+PREDEFINED             = "XXH_DOXYGEN=" \
+                         "XXH_PUBLIC_API=" \
+                         "XXH_FORCE_INLINE=static inline" \
+                         "XXH_NO_INLINE=static" \
+                         "XXH_RESTRICT=restrict" \
+                         "XSUM_API=" \
+                         "XXH_STATIC_LINKING_ONLY" \
+                         "XXH_IMPLEMENTATION" \
+                         "XXH_PUREF=[[gnu::pure]]" \
+                         "XXH_CONSTF=[[gnu::const]]" \
+                         "XXH_MALLOCF=[[gnu::malloc]]" \
+                         "XXH_ALIGN(N)=alignas(N)" \
+                         "XXH_ALIGN_MEMBER(align,type)=alignas(align) type"
+
+# We want HTML docs
+GENERATE_HTML          = YES
+HTML_OUTPUT            = html
+HTML_FILE_EXTENSION    = .html
+# Tweak the colors a bit
+HTML_COLORSTYLE_HUE    = 220
+HTML_COLORSTYLE_GAMMA  = 100
+HTML_COLORSTYLE_SAT    = 100
+
+# We don't want LaTeX.
+GENERATE_LATEX         = NO

third_party/xxhash/LICENSE

@@ -0,0 +1,26 @@
+xxHash Library
+Copyright (c) 2012-2021 Yann Collet
+All rights reserved.
+
+BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php)
+
+Redistribution and use in source and binary forms, with or without modification,
+are permitted provided that the following conditions are met:
+
+* Redistributions of source code must retain the above copyright notice, this
+  list of conditions and the following disclaimer.
+
+* Redistributions in binary form must reproduce the above copyright notice, this
+  list of conditions and the following disclaimer in the documentation and/or
+  other materials provided with the distribution.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR
+ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
+ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

third_party/xxhash/Makefile

@@ -0,0 +1,623 @@
+# ################################################################
+# xxHash Makefile
+# Copyright (C) 2012-2021 Yann Collet
+#
+# GPL v2 License
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License along
+# with this program; if not, write to the Free Software Foundation, Inc.,
+# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+#
+# You can contact the author at:
+#   - xxHash homepage: https://www.xxhash.com
+#   - xxHash source repository: https://github.com/Cyan4973/xxHash
+# ################################################################
+# xxhsum: provides 32/64 bits hash of one or multiple files, or stdin
+# ################################################################
+Q = $(if $(filter 1,$(V) $(VERBOSE)),,@)
+
+# Version numbers
+SED ?= sed
+SED_ERE_OPT ?= -E
+LIBVER_MAJOR_SCRIPT:=`$(SED) -n '/define XXH_VERSION_MAJOR/s/.*[[:blank:]]\([0-9][0-9]*\).*/\1/p' < xxhash.h`
+LIBVER_MINOR_SCRIPT:=`$(SED) -n '/define XXH_VERSION_MINOR/s/.*[[:blank:]]\([0-9][0-9]*\).*/\1/p' < xxhash.h`
+LIBVER_PATCH_SCRIPT:=`$(SED) -n '/define XXH_VERSION_RELEASE/s/.*[[:blank:]]\([0-9][0-9]*\).*/\1/p' < xxhash.h`
+LIBVER_MAJOR := $(shell echo $(LIBVER_MAJOR_SCRIPT))
+LIBVER_MINOR := $(shell echo $(LIBVER_MINOR_SCRIPT))
+LIBVER_PATCH := $(shell echo $(LIBVER_PATCH_SCRIPT))
+LIBVER := $(LIBVER_MAJOR).$(LIBVER_MINOR).$(LIBVER_PATCH)
+
+CFLAGS ?= -O3
+DEBUGFLAGS+=-Wall -Wextra -Wconversion -Wcast-qual -Wcast-align -Wshadow \
+            -Wstrict-aliasing=1 -Wswitch-enum -Wdeclaration-after-statement \
+            -Wstrict-prototypes -Wundef -Wpointer-arith -Wformat-security \
+            -Wvla -Wformat=2 -Winit-self -Wfloat-equal -Wwrite-strings \
+            -Wredundant-decls -Wstrict-overflow=2
+CFLAGS += $(DEBUGFLAGS) $(MOREFLAGS)
+FLAGS   = $(CFLAGS) $(CPPFLAGS)
+XXHSUM_VERSION = $(LIBVER)
+UNAME := $(shell uname)
+
+# Define *.exe as extension for Windows systems
+ifneq (,$(filter Windows%,$(OS)))
+EXT =.exe
+else
+EXT =
+endif
+
+# OS X linker doesn't support -soname, and use different extension
+# see: https://developer.apple.com/library/mac/documentation/DeveloperTools/Conceptual/DynamicLibraries/100-Articles/DynamicLibraryDesignGuidelines.html
+ifeq ($(UNAME), Darwin)
+	SHARED_EXT = dylib
+	SHARED_EXT_MAJOR = $(LIBVER_MAJOR).$(SHARED_EXT)
+	SHARED_EXT_VER = $(LIBVER).$(SHARED_EXT)
+	SONAME_FLAGS = -install_name $(LIBDIR)/libxxhash.$(SHARED_EXT_MAJOR) -compatibility_version $(LIBVER_MAJOR) -current_version $(LIBVER)
+else
+	SONAME_FLAGS = -Wl,-soname=libxxhash.$(SHARED_EXT).$(LIBVER_MAJOR)
+	SHARED_EXT = so
+	SHARED_EXT_MAJOR = $(SHARED_EXT).$(LIBVER_MAJOR)
+	SHARED_EXT_VER = $(SHARED_EXT).$(LIBVER)
+endif
+
+LIBXXH = libxxhash.$(SHARED_EXT_VER)
+
+XXHSUM_SRC_DIR = cli
+XXHSUM_SPLIT_SRCS = $(XXHSUM_SRC_DIR)/xxhsum.c \
+                    $(XXHSUM_SRC_DIR)/xsum_os_specific.c \
+                    $(XXHSUM_SRC_DIR)/xsum_output.c \
+                    $(XXHSUM_SRC_DIR)/xsum_sanity_check.c \
+                    $(XXHSUM_SRC_DIR)/xsum_bench.c
+XXHSUM_SPLIT_OBJS = $(XXHSUM_SPLIT_SRCS:.c=.o)
+XXHSUM_HEADERS = $(XXHSUM_SRC_DIR)/xsum_config.h \
+                 $(XXHSUM_SRC_DIR)/xsum_arch.h \
+                 $(XXHSUM_SRC_DIR)/xsum_os_specific.h \
+                 $(XXHSUM_SRC_DIR)/xsum_output.h \
+                 $(XXHSUM_SRC_DIR)/xsum_sanity_check.h \
+                 $(XXHSUM_SRC_DIR)/xsum_bench.h
+
+## generate CLI and libraries in release mode (default for `make`)
+.PHONY: default
+default: DEBUGFLAGS=
+default: lib xxhsum_and_links
+
+.PHONY: all
+all: lib xxhsum xxhsum_inlinedXXH
+
+## xxhsum is the command line interface (CLI)
+ifeq ($(DISPATCH),1)
+xxhsum: CPPFLAGS += -DXXHSUM_DISPATCH=1
+xxhsum: xxh_x86dispatch.o
+endif
+xxhsum: xxhash.o $(XXHSUM_SPLIT_OBJS)
+	$(CC) $(FLAGS) $^ $(LDFLAGS) -o $@$(EXT)
+
+xxhsum32: CFLAGS += -m32  ## generate CLI in 32-bits mode
+xxhsum32: xxhash.c $(XXHSUM_SPLIT_SRCS) ## do not generate object (avoid mixing different ABI)
+	$(CC) $(FLAGS) $^ $(LDFLAGS) -o $@$(EXT)
+
+## dispatch only works for x86/x64 systems
+dispatch: CPPFLAGS += -DXXHSUM_DISPATCH=1
+dispatch: xxhash.o xxh_x86dispatch.o $(XXHSUM_SPLIT_SRCS)
+	$(CC) $(FLAGS) $^ $(LDFLAGS) -o $@$(EXT)
+
+xxhash.o: xxhash.c xxhash.h
+xxhsum.o: $(XXHSUM_SRC_DIR)/xxhsum.c $(XXHSUM_HEADERS) \
+    xxhash.h xxh_x86dispatch.h
+xxh_x86dispatch.o: xxh_x86dispatch.c xxh_x86dispatch.h xxhash.h
+
+.PHONY: xxhsum_and_links
+xxhsum_and_links: xxhsum xxh32sum xxh64sum xxh128sum
+
+xxh32sum xxh64sum xxh128sum: xxhsum
+	ln -sf $<$(EXT) $@$(EXT)
+
+xxhsum_inlinedXXH: CPPFLAGS += -DXXH_INLINE_ALL
+xxhsum_inlinedXXH: $(XXHSUM_SPLIT_SRCS)
+	$(CC) $(FLAGS) $< -o $@$(EXT)
+
+
+# library
+
+libxxhash.a: ARFLAGS = rcs
+libxxhash.a: xxhash.o
+	$(AR) $(ARFLAGS) $@ $^
+
+$(LIBXXH): LDFLAGS += -shared
+ifeq (,$(filter Windows%,$(OS)))
+$(LIBXXH): CFLAGS += -fPIC
+endif
+ifeq ($(DISPATCH),1)
+$(LIBXXH): xxh_x86dispatch.c
+endif
+$(LIBXXH): xxhash.c
+	$(CC) $(FLAGS) $^ $(LDFLAGS) $(SONAME_FLAGS) -o $@
+	ln -sf $@ libxxhash.$(SHARED_EXT_MAJOR)
+	ln -sf $@ libxxhash.$(SHARED_EXT)
+
+.PHONY: libxxhash
+libxxhash:  ## generate dynamic xxhash library
+libxxhash: $(LIBXXH)
+
+.PHONY: lib
+lib:  ## generate static and dynamic xxhash libraries
+lib: libxxhash.a libxxhash
+
+# helper targets
+
+AWK  = awk
+GREP = grep
+SORT = sort
+NM   = nm
+
+.PHONY: list
+list:  ## list all Makefile targets
+	$(Q)$(MAKE) -pRrq -f $(lastword $(MAKEFILE_LIST)) : 2>/dev/null | $(AWK) -v RS= -F: '/^# File/,/^# Finished Make data base/ {if ($$1 !~ "^[#.]") {print $$1}}' | $(SORT) | egrep -v -e '^[^[:alnum:]]' -e '^$@$$' | xargs
+
+.PHONY: help
+help:  ## list documented targets
+	$(Q)$(GREP) -E '^[0-9a-zA-Z_-]+:.*?## .*$$' $(MAKEFILE_LIST) | \
+	$(SORT) | \
+	$(AWK) 'BEGIN {FS = ":.*?## "}; {printf "\033[36m%-30s\033[0m %s\n", $$1, $$2}'
+
+.PHONY: clean
+clean:  ## remove all build artifacts
+	$(Q)$(RM) -r *.dSYM   # Mac OS-X specific
+	$(Q)$(RM) core *.o *.obj *.$(SHARED_EXT) *.$(SHARED_EXT).* *.a libxxhash.pc
+	$(Q)$(RM) xxhsum$(EXT) xxhsum32$(EXT) xxhsum_inlinedXXH$(EXT) dispatch$(EXT)
+	$(Q)$(RM) xxh32sum$(EXT) xxh64sum$(EXT) xxh128sum$(EXT)
+	$(Q)$(RM) $(XXHSUM_SRC_DIR)/*.o $(XXHSUM_SRC_DIR)/*.obj
+	$(MAKE) -C tests clean
+	$(MAKE) -C tests/bench clean
+	$(MAKE) -C tests/collisions clean
+	@echo cleaning completed
+
+
+# =================================================
+# tests
+# =================================================
+
+# make check can be run with cross-compiled binaries on emulated environments (qemu user mode)
+# by setting $(RUN_ENV) to the target emulation environment
+.PHONY: check
+check: xxhsum test_sanity   ## basic tests for xxhsum CLI, set RUN_ENV for emulated environments
+	# stdin
+	$(RUN_ENV) ./xxhsum$(EXT) < xxhash.c
+	# multiple files
+	$(RUN_ENV) ./xxhsum$(EXT) xxhash.*
+	# internal bench
+	$(RUN_ENV) ./xxhsum$(EXT) -bi0
+	# long bench command
+	$(RUN_ENV) ./xxhsum$(EXT) --benchmark-all -i0
+	# bench multiple variants
+	$(RUN_ENV) ./xxhsum$(EXT) -b1,2,3 -i0
+	# file bench
+	$(RUN_ENV) ./xxhsum$(EXT) -bi0 xxhash.c
+	# 32-bit
+	$(RUN_ENV) ./xxhsum$(EXT) -H0 xxhash.c
+	# 128-bit
+	$(RUN_ENV) ./xxhsum$(EXT) -H2 xxhash.c
+	# XXH3 (enforce BSD style)
+	$(RUN_ENV) ./xxhsum$(EXT) -H3 xxhash.c | grep "XXH3"
+	# request incorrect variant
+	$(RUN_ENV) ./xxhsum$(EXT) -H9 xxhash.c ; test $$? -eq 1
+	@printf "\n .......   checks completed successfully   ....... \n"
+
+.PHONY: test-unicode
+test-unicode:
+	$(MAKE) -C tests test_unicode
+
+.PHONY: test_sanity
+test_sanity:
+	$(MAKE) -C tests test_sanity
+
+.PHONY: test-mem
+VALGRIND = valgrind --leak-check=yes --error-exitcode=1
+test-mem: RUN_ENV = $(VALGRIND)
+test-mem: xxhsum check
+
+.PHONY: test32
+test32: xxhsum32
+	@echo ---- test 32-bit ----
+	./xxhsum32 -bi0 xxhash.c
+
+TEST_FILES = xxhsum$(EXT) xxhash.c xxhash.h
+.PHONY: test-xxhsum-c
+test-xxhsum-c: xxhsum
+	# xxhsum to/from pipe
+	./xxhsum $(TEST_FILES) | ./xxhsum -c -
+	./xxhsum -H0 $(TEST_FILES) | ./xxhsum -c -
+	# xxhsum -c is unable to verify checksum of file from STDIN (#470)
+	./xxhsum < README.md > .test.README.md.xxh
+	./xxhsum -c .test.README.md.xxh < README.md
+	# xxhsum -q does not display "Loading" message into stderr (#251)
+	! ./xxhsum -q $(TEST_FILES) 2>&1 | grep Loading
+	# xxhsum does not display "Loading" message into stderr either
+	! ./xxhsum $(TEST_FILES) 2>&1 | grep Loading
+	# Check that xxhsum do display filename that it failed to open.
+	LC_ALL=C ./xxhsum nonexistent 2>&1 | grep "Error: Could not open 'nonexistent'"
+	# xxhsum to/from file, shell redirection
+	./xxhsum $(TEST_FILES) > .test.xxh64
+	./xxhsum --tag $(TEST_FILES) > .test.xxh64_tag
+	./xxhsum --little-endian $(TEST_FILES) > .test.le_xxh64
+	./xxhsum --tag --little-endian $(TEST_FILES) > .test.le_xxh64_tag
+	./xxhsum -H0 $(TEST_FILES) > .test.xxh32
+	./xxhsum -H0 --tag $(TEST_FILES) > .test.xxh32_tag
+	./xxhsum -H0 --little-endian $(TEST_FILES) > .test.le_xxh32
+	./xxhsum -H0 --tag --little-endian $(TEST_FILES) > .test.le_xxh32_tag
+	./xxhsum -H2 $(TEST_FILES) > .test.xxh128
+	./xxhsum -H2 --tag $(TEST_FILES) > .test.xxh128_tag
+	./xxhsum -H2 --little-endian $(TEST_FILES) > .test.le_xxh128
+	./xxhsum -H2 --tag --little-endian $(TEST_FILES) > .test.le_xxh128_tag
+	./xxhsum -H3 $(TEST_FILES) > .test.xxh3
+	./xxhsum -H3 --tag $(TEST_FILES) > .test.xxh3_tag
+	./xxhsum -H3 --little-endian $(TEST_FILES) > .test.le_xxh3
+	./xxhsum -H3 --tag --little-endian $(TEST_FILES) > .test.le_xxh3_tag
+	./xxhsum -c .test.xxh*
+	./xxhsum -c --little-endian .test.le_xxh*
+	./xxhsum -c .test.*_tag
+	# read list of files from stdin
+	./xxhsum -c < .test.xxh32
+	./xxhsum -c < .test.xxh64
+	./xxhsum -c < .test.xxh128
+	./xxhsum -c < .test.xxh3
+	cat .test.xxh* | ./xxhsum -c -
+	# check variant with '*' marker as second separator
+	$(SED) 's/  / \*/' .test.xxh32 | ./xxhsum -c
+	# bsd-style output
+	./xxhsum --tag xxhsum* | $(GREP) XXH64
+	./xxhsum --tag -H0 xxhsum* | $(GREP) XXH32
+	./xxhsum --tag -H1 xxhsum* | $(GREP) XXH64
+	./xxhsum --tag -H2 xxhsum* | $(GREP) XXH128
+	./xxhsum --tag -H3 xxhsum* | $(GREP) XXH3
+	./xxhsum       -H3 xxhsum* | $(GREP) XXH3  # --tag is implicit for H3
+	./xxhsum --tag -H32 xxhsum* | $(GREP) XXH32
+	./xxhsum --tag -H64 xxhsum* | $(GREP) XXH64
+	./xxhsum --tag -H128 xxhsum* | $(GREP) XXH128
+	./xxhsum --tag -H0 --little-endian xxhsum* | $(GREP) XXH32_LE
+	./xxhsum --tag -H1 --little-endian xxhsum* | $(GREP) XXH64_LE
+	./xxhsum --tag -H2 --little-endian xxhsum* | $(GREP) XXH128_LE
+	./xxhsum       -H3 --little-endian xxhsum* | $(GREP) XXH3_LE
+	./xxhsum --tag -H32 --little-endian xxhsum* | $(GREP) XXH32_LE
+	./xxhsum --tag -H64 --little-endian xxhsum* | $(GREP) XXH64_LE
+	./xxhsum --tag -H128 --little-endian xxhsum* | $(GREP) XXH128_LE
+	# check bsd-style
+	./xxhsum --tag xxhsum* | ./xxhsum -c
+	./xxhsum --tag -H32 --little-endian xxhsum* | ./xxhsum -c
+	# xxhsum -c warns improperly format lines.
+	echo '12345678 ' >>.test.xxh32
+	./xxhsum -c .test.xxh32 | $(GREP) improperly
+	echo '123456789  file' >>.test.xxh64
+	./xxhsum -c .test.xxh64 | $(GREP) improperly
+	# Expects "FAILED"
+	echo "0000000000000000  LICENSE" | ./xxhsum -c -; test $$? -eq 1
+	echo "00000000  LICENSE" | ./xxhsum -c -; test $$? -eq 1
+	# Expects "FAILED open or read"
+	echo "0000000000000000  test-expects-file-not-found" | ./xxhsum -c -; test $$? -eq 1
+	echo "00000000  test-expects-file-not-found" | ./xxhsum -c -; test $$? -eq 1
+	@$(RM) .test.*
+
+.PHONY: test-filename-escape
+test-filename-escape:
+	$(MAKE) -C tests test_filename_escape
+
+.PHONY: armtest
+armtest: clean
+	@echo ---- test ARM compilation ----
+	CC=arm-linux-gnueabi-gcc MOREFLAGS="-Werror -static" $(MAKE) xxhsum
+
+.PHONY: clangtest
+clangtest: clean
+	@echo ---- test clang compilation ----
+	CC=clang MOREFLAGS="-Werror -Wconversion -Wno-sign-conversion" $(MAKE) all
+
+.PHONY: gcc-og-test
+gcc-og-test: clean
+	@echo ---- test gcc -Og compilation ----
+	CFLAGS="-Og -Wall -Wextra -Wundef -Wshadow -Wcast-align -Werror -fPIC" MOREFLAGS="-Werror" $(MAKE) all
+
+.PHONY: cxxtest
+cxxtest: clean
+	@echo ---- test C++ compilation ----
+	CC="$(CXX) -Wno-deprecated" $(MAKE) all CFLAGS="-O3 -Wall -Wextra -Wundef -Wshadow -Wcast-align -Werror -fPIC"
+
+.PHONY: c90test
+ifeq ($(NO_C90_TEST),true)
+c90test:
+	@echo no c90 compatibility test
+else
+c90test: CPPFLAGS += -DXXH_NO_LONG_LONG
+c90test: CFLAGS += -std=c90 -Werror -pedantic
+c90test: xxhash.c
+	@echo ---- test strict C90 compilation [xxh32 only] ----
+	$(RM) xxhash.o
+	$(CC) $(FLAGS) $^ -c
+	$(NM) xxhash.o | $(GREP) XXH64 ; test $$? -eq 1
+	$(RM) xxhash.o
+endif
+
+.PHONY: noxxh3test
+noxxh3test: CPPFLAGS += -DXXH_NO_XXH3
+noxxh3test: CFLAGS += -Werror -pedantic -Wno-long-long  # XXH64 requires long long support
+noxxh3test: OFILE = xxh_noxxh3.o
+noxxh3test: xxhash.c
+	@echo ---- test compilation without XXH3 ----
+	$(CC) $(FLAGS) -c $^ -o $(OFILE)
+	$(NM) $(OFILE) | $(GREP) XXH3_ ; test $$? -eq 1
+	$(RM) $(OFILE)
+
+.PHONY: nostreamtest
+nostreamtest: CPPFLAGS += -DXXH_NO_STREAM
+nostreamtest: CFLAGS += -Werror -pedantic -Wno-long-long  # XXH64 requires long long support
+nostreamtest: OFILE = xxh_nostream.o
+nostreamtest: xxhash.c
+	@echo ---- test compilation without streaming ----
+	$(CC) $(FLAGS) -c $^ -o $(OFILE)
+	$(NM) $(OFILE) | $(GREP) update ; test $$? -eq 1
+	$(RM) $(OFILE)
+
+.PHONY: nostdlibtest
+nostdlibtest: CPPFLAGS += -DXXH_NO_STDLIB
+nostdlibtest: CFLAGS += -Werror -pedantic -Wno-long-long  # XXH64 requires long long support
+nostdlibtest: OFILE = xxh_nostdlib.o
+nostdlibtest: xxhash.c
+	@echo ---- test compilation without \<stdlib.h\> ----
+	$(CC) $(FLAGS) -c $^ -o $(OFILE)
+	$(NM) $(OFILE) | $(GREP) "U _free\|U free" ; test $$? -eq 1
+	$(RM) $(OFILE)
+
+.PHONY: usan
+usan: CC=clang
+usan: CXX=clang++
+usan:  ## check CLI runtime for undefined behavior, using clang's sanitizer
+	@echo ---- check undefined behavior - sanitize ----
+	$(MAKE) clean
+	$(MAKE) test CC=$(CC) CXX=$(CXX) MOREFLAGS="-g -fsanitize=undefined -fno-sanitize-recover=all"
+
+.PHONY: staticAnalyze
+SCANBUILD ?= scan-build
+staticAnalyze: clean  ## check C source files using $(SCANBUILD) static analyzer
+	@echo ---- static analyzer - $(SCANBUILD) ----
+	CFLAGS="-g -Werror" $(SCANBUILD) --status-bugs -v $(MAKE) all
+
+CPPCHECK ?= cppcheck
+.PHONY: cppcheck
+cppcheck:  ## check C source files using $(CPPCHECK) static analyzer
+	@echo ---- static analyzer - $(CPPCHECK) ----
+	$(CPPCHECK) . --force --enable=warning,portability,performance,style --error-exitcode=1 > /dev/null
+
+.PHONY: namespaceTest
+namespaceTest:  ## ensure XXH_NAMESPACE redefines all public symbols
+	$(CC) -c xxhash.c
+	$(CC) -DXXH_NAMESPACE=TEST_ -c xxhash.c -o xxhash2.o
+	$(CC) xxhash.o xxhash2.o $(XXHSUM_SPLIT_SRCS)  -o xxhsum2  # will fail if one namespace missing (symbol collision)
+	$(RM) *.o xxhsum2  # clean
+
+MAN = $(XXHSUM_SRC_DIR)/xxhsum.1
+MD2ROFF ?= ronn
+MD2ROFF_FLAGS ?= --roff --warnings --manual="User Commands" --organization="xxhsum $(XXHSUM_VERSION)"
+$(MAN): $(XXHSUM_SRC_DIR)/xxhsum.1.md xxhash.h
+	cat $< | $(MD2ROFF) $(MD2ROFF_FLAGS) | $(SED) -n '/^\.\\\".*/!p' > $@
+
+.PHONY: man
+man: $(MAN)  ## generate man page from markdown source
+
+.PHONY: clean-man
+clean-man:
+	$(RM) xxhsum.1
+
+.PHONY: preview-man
+preview-man: man
+	man ./xxhsum.1
+
+.PHONY: test
+test: DEBUGFLAGS += -DXXH_DEBUGLEVEL=1
+test: all namespaceTest check test-xxhsum-c c90test test-tools noxxh3test nostdlibtest
+
+.PHONY: test-inline
+test-inline:
+	$(MAKE) -C tests test_multiInclude
+
+.PHONY: test-all
+test-all: CFLAGS += -Werror
+test-all: test test32 test-unicode clangtest gcc-og-test cxxtest usan test-inline listL120 trailingWhitespace test-xxh-nnn-sums
+
+.PHONY: test-tools
+test-tools:
+	CFLAGS=-Werror $(MAKE) -C tests/bench
+	CFLAGS=-Werror $(MAKE) -C tests/collisions
+
+.PHONY: test-xxh-nnn-sums
+test-xxh-nnn-sums: xxhsum_and_links
+	./xxhsum    README.md > tmp.xxhsum.out    # xxhsum outputs xxh64
+	./xxh32sum  README.md > tmp.xxh32sum.out
+	./xxh64sum  README.md > tmp.xxh64sum.out
+	./xxh128sum README.md > tmp.xxh128sum.out
+	cat tmp.xxhsum.out
+	cat tmp.xxh32sum.out
+	cat tmp.xxh64sum.out
+	cat tmp.xxh128sum.out
+	./xxhsum -c tmp.xxhsum.out
+	./xxhsum -c tmp.xxh32sum.out
+	./xxhsum -c tmp.xxh64sum.out
+	./xxhsum -c tmp.xxh128sum.out
+	./xxh32sum -c tmp.xxhsum.out            ; test $$? -eq 1  # expects "no properly formatted"
+	./xxh32sum -c tmp.xxh32sum.out
+	./xxh32sum -c tmp.xxh64sum.out          ; test $$? -eq 1  # expects "no properly formatted"
+	./xxh32sum -c tmp.xxh128sum.out         ; test $$? -eq 1  # expects "no properly formatted"
+	./xxh64sum -c tmp.xxhsum.out
+	./xxh64sum -c tmp.xxh32sum.out          ; test $$? -eq 1  # expects "no properly formatted"
+	./xxh64sum -c tmp.xxh64sum.out
+	./xxh64sum -c tmp.xxh128sum.out         ; test $$? -eq 1  # expects "no properly formatted"
+	./xxh128sum -c tmp.xxhsum.out           ; test $$? -eq 1  # expects "no properly formatted"
+	./xxh128sum -c tmp.xxh32sum.out         ; test $$? -eq 1  # expects "no properly formatted"
+	./xxh128sum -c tmp.xxh64sum.out         ; test $$? -eq 1  # expects "no properly formatted"
+	./xxh128sum -c tmp.xxh128sum.out
+
+.PHONY: listL120
+listL120:  # extract lines >= 120 characters in *.{c,h}, by Takayuki Matsuoka (note: $$, for Makefile compatibility)
+	find . -type f -name '*.c' -o -name '*.h' | while read -r filename; do awk 'length > 120 {print FILENAME "(" FNR "): " $$0}' $$filename; done
+
+.PHONY: trailingWhitespace
+trailingWhitespace:
+	! $(GREP) -E "`printf '[ \\t]$$'`" cli/*.c cli/*.h cli/*.1 *.c *.h LICENSE Makefile cmake_unofficial/CMakeLists.txt
+
+.PHONY: lint-unicode
+lint-unicode:
+	./tests/unicode_lint.sh
+
+# =========================================================
+# make install is validated only for the following targets
+# =========================================================
+ifneq (,$(filter Linux Darwin GNU/kFreeBSD GNU Haiku OpenBSD FreeBSD NetBSD DragonFly SunOS CYGWIN% , $(UNAME)))
+
+DESTDIR     ?=
+# directory variables: GNU conventions prefer lowercase
+# see https://www.gnu.org/prep/standards/html_node/Makefile-Conventions.html
+# support both lower and uppercase (BSD), use uppercase in script
+prefix      ?= /usr/local
+PREFIX      ?= $(prefix)
+exec_prefix ?= $(PREFIX)
+EXEC_PREFIX ?= $(exec_prefix)
+libdir      ?= $(EXEC_PREFIX)/lib
+LIBDIR      ?= $(libdir)
+includedir  ?= $(PREFIX)/include
+INCLUDEDIR  ?= $(includedir)
+bindir      ?= $(EXEC_PREFIX)/bin
+BINDIR      ?= $(bindir)
+datarootdir ?= $(PREFIX)/share
+mandir      ?= $(datarootdir)/man
+man1dir     ?= $(mandir)/man1
+
+ifneq (,$(filter $(UNAME),FreeBSD NetBSD DragonFly))
+PKGCONFIGDIR ?= $(PREFIX)/libdata/pkgconfig
+else
+PKGCONFIGDIR ?= $(LIBDIR)/pkgconfig
+endif
+
+ifneq (,$(filter $(UNAME),OpenBSD FreeBSD NetBSD DragonFly SunOS))
+MANDIR  ?= $(PREFIX)/man/man1
+else
+MANDIR  ?= $(man1dir)
+endif
+
+ifneq (,$(filter $(UNAME),SunOS))
+INSTALL ?= ginstall
+else
+INSTALL ?= install
+endif
+
+INSTALL_PROGRAM ?= $(INSTALL)
+INSTALL_DATA    ?= $(INSTALL) -m 644
+INSTALL_DIR     ?= $(INSTALL) -d -m 755
+
+
+# Escape special symbols by putting each character into its separate class
+EXEC_PREFIX_REGEX ?= $(shell echo "$(EXEC_PREFIX)" | $(SED) $(SED_ERE_OPT) -e "s/([^^])/[\1]/g" -e "s/\\^/\\\\^/g")
+PREFIX_REGEX ?= $(shell echo "$(PREFIX)" | $(SED) $(SED_ERE_OPT) -e "s/([^^])/[\1]/g" -e "s/\\^/\\\\^/g")
+
+PCLIBDIR ?= $(shell echo "$(LIBDIR)"     | $(SED) -n $(SED_ERE_OPT) -e "s@^$(EXEC_PREFIX_REGEX)(/|$$)@@p")
+PCINCDIR ?= $(shell echo "$(INCLUDEDIR)" | $(SED) -n $(SED_ERE_OPT) -e "s@^$(PREFIX_REGEX)(/|$$)@@p")
+PCEXECDIR?= $(if $(filter $(PREFIX),$(EXEC_PREFIX)),$$\{prefix\},$(EXEC_PREFIX))
+
+ifeq (,$(PCLIBDIR))
+# Additional prefix check is required, since the empty string is technically a
+# valid PCLIBDIR
+ifeq (,$(shell echo "$(LIBDIR)" | $(SED) -n $(SED_ERE_OPT) -e "\\@^$(EXEC_PREFIX_REGEX)(/|$$)@ p"))
+$(error configured libdir ($(LIBDIR)) is outside of exec_prefix ($(EXEC_PREFIX)), can't generate pkg-config file)
+endif
+endif
+
+ifeq (,$(PCINCDIR))
+# Additional prefix check is required, since the empty string is technically a
+# valid PCINCDIR
+ifeq (,$(shell echo "$(INCLUDEDIR)" | $(SED) -n $(SED_ERE_OPT) -e "\\@^$(PREFIX_REGEX)(/|$$)@ p"))
+$(error configured includedir ($(INCLUDEDIR)) is outside of prefix ($(PREFIX)), can't generate pkg-config file)
+endif
+endif
+
+libxxhash.pc: libxxhash.pc.in
+	@echo creating pkgconfig
+	$(Q)$(SED) $(SED_ERE_OPT) -e 's|@PREFIX@|$(PREFIX)|' \
+          -e 's|@EXECPREFIX@|$(PCEXECDIR)|' \
+          -e 's|@LIBDIR@|$(PCLIBDIR)|' \
+          -e 's|@INCLUDEDIR@|$(PCINCDIR)|' \
+          -e 's|@VERSION@|$(LIBVER)|' \
+          $< > $@
+
+
+install_libxxhash.a: libxxhash.a
+	@echo Installing libxxhash.a
+	$(Q)$(INSTALL_DIR) $(DESTDIR)$(LIBDIR)
+	$(Q)$(INSTALL_DATA) libxxhash.a $(DESTDIR)$(LIBDIR)
+
+install_libxxhash: libxxhash
+	@echo Installing libxxhash
+	$(Q)$(INSTALL_DIR) $(DESTDIR)$(LIBDIR)
+	$(Q)$(INSTALL_PROGRAM) $(LIBXXH) $(DESTDIR)$(LIBDIR)
+	$(Q)ln -sf $(LIBXXH) $(DESTDIR)$(LIBDIR)/libxxhash.$(SHARED_EXT_MAJOR)
+	$(Q)ln -sf $(LIBXXH) $(DESTDIR)$(LIBDIR)/libxxhash.$(SHARED_EXT)
+
+install_libxxhash.includes:
+	$(Q)$(INSTALL) -d -m 755 $(DESTDIR)$(INCLUDEDIR)   # includes
+	$(Q)$(INSTALL_DATA) xxhash.h $(DESTDIR)$(INCLUDEDIR)
+	$(Q)$(INSTALL_DATA) xxh3.h $(DESTDIR)$(INCLUDEDIR) # for compatibility, will be removed in v0.9.0
+ifeq ($(DISPATCH),1)
+	$(Q)$(INSTALL_DATA) xxh_x86dispatch.h $(DESTDIR)$(INCLUDEDIR)
+endif
+
+install_libxxhash.pc: libxxhash.pc
+	@echo Installing pkgconfig
+	$(Q)$(INSTALL_DIR) $(DESTDIR)$(PKGCONFIGDIR)/
+	$(Q)$(INSTALL_DATA) libxxhash.pc $(DESTDIR)$(PKGCONFIGDIR)/
+
+install_xxhsum: xxhsum
+	@echo Installing xxhsum
+	$(Q)$(INSTALL_DIR) $(DESTDIR)$(BINDIR)/
+	$(Q)$(INSTALL_PROGRAM) xxhsum $(DESTDIR)$(BINDIR)/xxhsum
+	$(Q)ln -sf xxhsum $(DESTDIR)$(BINDIR)/xxh32sum
+	$(Q)ln -sf xxhsum $(DESTDIR)$(BINDIR)/xxh64sum
+	$(Q)ln -sf xxhsum $(DESTDIR)$(BINDIR)/xxh128sum
+
+install_man:
+	@echo Installing man pages
+	$(Q)$(INSTALL_DIR) $(DESTDIR)$(MANDIR)/
+	$(Q)$(INSTALL_DATA) $(MAN) $(DESTDIR)$(MANDIR)/xxhsum.1
+	$(Q)ln -sf xxhsum.1 $(DESTDIR)$(MANDIR)/xxh32sum.1
+	$(Q)ln -sf xxhsum.1 $(DESTDIR)$(MANDIR)/xxh64sum.1
+	$(Q)ln -sf xxhsum.1 $(DESTDIR)$(MANDIR)/xxh128sum.1
+
+.PHONY: install
+install: install_libxxhash.a install_libxxhash install_libxxhash.includes install_libxxhash.pc install_xxhsum install_man ## install libraries, CLI, links and man page
+	@echo xxhash installation completed
+
+.PHONY: uninstall
+uninstall:  ## uninstall libraries, CLI, links and man page
+	$(Q)$(RM) $(DESTDIR)$(LIBDIR)/libxxhash.a
+	$(Q)$(RM) $(DESTDIR)$(LIBDIR)/libxxhash.$(SHARED_EXT)
+	$(Q)$(RM) $(DESTDIR)$(LIBDIR)/libxxhash.$(SHARED_EXT_MAJOR)
+	$(Q)$(RM) $(DESTDIR)$(LIBDIR)/$(LIBXXH)
+	$(Q)$(RM) $(DESTDIR)$(INCLUDEDIR)/xxhash.h
+	$(Q)$(RM) $(DESTDIR)$(INCLUDEDIR)/xxh3.h
+	$(Q)$(RM) $(DESTDIR)$(INCLUDEDIR)/xxh_x86dispatch.h
+	$(Q)$(RM) $(DESTDIR)$(PKGCONFIGDIR)/libxxhash.pc
+	$(Q)$(RM) $(DESTDIR)$(BINDIR)/xxh32sum
+	$(Q)$(RM) $(DESTDIR)$(BINDIR)/xxh64sum
+	$(Q)$(RM) $(DESTDIR)$(BINDIR)/xxh128sum
+	$(Q)$(RM) $(DESTDIR)$(BINDIR)/xxhsum
+	$(Q)$(RM) $(DESTDIR)$(MANDIR)/xxh32sum.1
+	$(Q)$(RM) $(DESTDIR)$(MANDIR)/xxh64sum.1
+	$(Q)$(RM) $(DESTDIR)$(MANDIR)/xxh128sum.1
+	$(Q)$(RM) $(DESTDIR)$(MANDIR)/xxhsum.1
+	@echo xxhsum successfully uninstalled
+
+endif

third_party/xxhash/README.md

@@ -0,0 +1,271 @@
+
+xxHash - Extremely fast hash algorithm
+======================================
+
+xxHash is an Extremely fast Hash algorithm, processing at RAM speed limits.
+Code is highly portable, and produces hashes identical across all platforms (little / big endian).
+The library includes the following algorithms :
+- XXH32 : generates 32-bit hashes, using 32-bit arithmetic
+- XXH64 : generates 64-bit hashes, using 64-bit arithmetic
+- XXH3 (since `v0.8.0`): generates 64 or 128-bit hashes, using vectorized arithmetic.
+  The 128-bit variant is called XXH128.
+
+All variants successfully complete the [SMHasher](https://code.google.com/p/smhasher/wiki/SMHasher) test suite
+which evaluates the quality of hash functions (collision, dispersion and randomness).
+Additional tests, which evaluate more thoroughly speed and collision properties of 64-bit hashes, [are also provided](https://github.com/Cyan4973/xxHash/tree/dev/tests).
+
+|Branch      |Status   |
+|------------|---------|
+|release     | [![Build Status](https://github.com/Cyan4973/xxHash/actions/workflows/ci.yml/badge.svg?branch=release)](https://github.com/Cyan4973/xxHash/actions?query=branch%3Arelease+) |
+|dev         | [![Build Status](https://github.com/Cyan4973/xxHash/actions/workflows/ci.yml/badge.svg?branch=dev)](https://github.com/Cyan4973/xxHash/actions?query=branch%3Adev+) |
+
+
+Benchmarks
+-------------------------
+
+The benchmarked reference system uses an Intel i7-9700K cpu, and runs Ubuntu x64 20.04.
+The [open source benchmark program] is compiled with `clang` v10.0 using `-O3` flag.
+
+| Hash Name     | Width | Bandwidth (GB/s) | Small Data Velocity | Quality | Comment |
+| ---------     | ----- | ---------------- | ----- | --- | --- |
+| __XXH3__ (SSE2) |  64 | 31.5 GB/s        | 133.1 | 10
+| __XXH128__ (SSE2) | 128 | 29.6 GB/s      | 118.1 | 10
+| _RAM sequential read_ | N/A | 28.0 GB/s  |   N/A | N/A | _for reference_
+| City64        |    64 | 22.0 GB/s        |  76.6 | 10
+| T1ha2         |    64 | 22.0 GB/s        |  99.0 |  9 | Slightly worse [collisions]
+| City128       |   128 | 21.7 GB/s        |  57.7 | 10
+| __XXH64__     |    64 | 19.4 GB/s        |  71.0 | 10
+| SpookyHash    |    64 | 19.3 GB/s        |  53.2 | 10
+| Mum           |    64 | 18.0 GB/s        |  67.0 |  9 | Slightly worse [collisions]
+| __XXH32__     |    32 |  9.7 GB/s        |  71.9 | 10
+| City32        |    32 |  9.1 GB/s        |  66.0 | 10
+| Murmur3       |    32 |  3.9 GB/s        |  56.1 | 10
+| SipHash       |    64 |  3.0 GB/s        |  43.2 | 10
+| FNV64         |    64 |  1.2 GB/s        |  62.7 |  5 | Poor avalanche properties
+| Blake2        |   256 |  1.1 GB/s        |   5.1 | 10 | Cryptographic
+| SHA1          |   160 |  0.8 GB/s        |   5.6 | 10 | Cryptographic but broken
+| MD5           |   128 |  0.6 GB/s        |   7.8 | 10 | Cryptographic but broken
+
+[open source benchmark program]: https://github.com/Cyan4973/xxHash/tree/release/tests/bench
+[collisions]: https://github.com/Cyan4973/xxHash/wiki/Collision-ratio-comparison#collision-study
+
+note 1: Small data velocity is a _rough_ evaluation of algorithm's efficiency on small data. For more detailed analysis, please refer to next paragraph.
+
+note 2: some algorithms feature _faster than RAM_ speed. In which case, they can only reach their full speed potential when input is already in CPU cache (L3 or better). Otherwise, they max out on RAM speed limit.
+
+### Small data
+
+Performance on large data is only one part of the picture.
+Hashing is also very useful in constructions like hash tables and bloom filters.
+In these use cases, it's frequent to hash a lot of small data (starting at a few bytes).
+Algorithm's performance can be very different for such scenarios, since parts of the algorithm,
+such as initialization or finalization, become fixed cost.
+The impact of branch mis-prediction also becomes much more present.
+
+XXH3 has been designed for excellent performance on both long and small inputs,
+which can be observed in the following graph:
+
+![XXH3, latency, random size](https://user-images.githubusercontent.com/750081/61976089-aedeab00-af9f-11e9-9239-e5375d6c080f.png)
+
+For a more detailed analysis, please visit the wiki :
+https://github.com/Cyan4973/xxHash/wiki/Performance-comparison#benchmarks-concentrating-on-small-data-
+
+Quality
+-------------------------
+
+Speed is not the only property that matters.
+Produced hash values must respect excellent dispersion and randomness properties,
+so that any sub-section of it can be used to maximally spread out a table or index,
+as well as reduce the amount of collisions to the minimal theoretical level, following the [birthday paradox].
+
+`xxHash` has been tested with Austin Appleby's excellent SMHasher test suite,
+and passes all tests, ensuring reasonable quality levels.
+It also passes extended tests from [newer forks of SMHasher], featuring additional scenarios and conditions.
+
+Finally, xxHash provides its own [massive collision tester](https://github.com/Cyan4973/xxHash/tree/dev/tests/collisions),
+able to generate and compare billions of hashes to test the limits of 64-bit hash algorithms.
+On this front too, xxHash features good results, in line with the [birthday paradox].
+A more detailed analysis is documented [in the wiki](https://github.com/Cyan4973/xxHash/wiki/Collision-ratio-comparison).
+
+[birthday paradox]: https://en.wikipedia.org/wiki/Birthday_problem
+[newer forks of SMHasher]: https://github.com/rurban/smhasher
+
+
+### Build modifiers
+
+The following macros can be set at compilation time to modify libxxhash's behavior. They are generally disabled by default.
+
+- `XXH_INLINE_ALL`: Make all functions `inline`, with implementations being directly included within `xxhash.h`.
+                    Inlining functions is beneficial for speed on small keys.
+                    It's _extremely effective_ when key length is expressed as _a compile time constant_,
+                    with performance improvements observed in the +200% range .
+                    See [this article](https://fastcompression.blogspot.com/2018/03/xxhash-for-small-keys-impressive-power.html) for details.
+- `XXH_PRIVATE_API`: same outcome as `XXH_INLINE_ALL`. Still available for legacy support.
+                     The name underlines that `XXH_*` symbol names will not be exported.
+- `XXH_NAMESPACE`: Prefixes all symbols with the value of `XXH_NAMESPACE`.
+                   This macro can only use compilable character set.
+                   Useful to evade symbol naming collisions,
+                   in case of multiple inclusions of xxHash's source code.
+                   Client applications still use the regular function names,
+                   as symbols are automatically translated through `xxhash.h`.
+- `XXH_FORCE_ALIGN_CHECK`: Use a faster direct read path when input is aligned.
+                           This option can result in dramatic performance improvement when input to hash is aligned on 32 or 64-bit boundaries,
+                           when running on architectures unable to load memory from unaligned addresses, or suffering a performance penalty from it.
+                           It is (slightly) detrimental on platform with good unaligned memory access performance (same instruction for both aligned and unaligned accesses).
+                           This option is automatically disabled on `x86`, `x64` and `aarch64`, and enabled on all other platforms.
+- `XXH_FORCE_MEMORY_ACCESS`: The default method `0` uses a portable `memcpy()` notation.
+                             Method `1` uses a gcc-specific `packed` attribute, which can provide better performance for some targets.
+                             Method `2` forces unaligned reads, which is not standard compliant, but might sometimes be the only way to extract better read performance.
+                             Method `3` uses a byteshift operation, which is best for old compilers which don't inline `memcpy()` or big-endian systems without a byteswap instruction.
+- `XXH_VECTOR` : manually select a vector instruction set (default: auto-selected at compilation time). Available instruction sets are `XXH_SCALAR`, `XXH_SSE2`, `XXH_AVX2`, `XXH_AVX512`, `XXH_NEON` and `XXH_VSX`. Compiler may require additional flags to ensure proper support (for example, `gcc` on linux will require `-mavx2` for `AVX2`, and `-mavx512f` for `AVX512`).
+- `XXH_NO_PREFETCH` : disable prefetching. Some platforms or situations may perform better without prefetching. XXH3 only.
+- `XXH_PREFETCH_DIST` : select prefetching distance. For close-to-metal adaptation to specific hardware platforms. XXH3 only.
+- `XXH_NO_STREAM`: Disables the streaming API, limiting it to single shot variants only.
+- `XXH_SIZE_OPT`: `0`: default, optimize for speed
+                  `1`: default for `-Os` and `-Oz`: disables some speed hacks for size optimization
+                  `2`: makes code as small as possible, performance may cry
+- `XXH_NO_INLINE_HINTS`: By default, xxHash uses `__attribute__((always_inline))` and `__forceinline` to improve performance at the cost of code size.
+                         Defining this macro to 1 will mark all internal functions as `static`, allowing the compiler to decide whether to inline a function or not.
+                         This is very useful when optimizing for smallest binary size,
+                         and is automatically defined when compiling with `-O0`, `-Os`, `-Oz`, or `-fno-inline` on GCC and Clang.
+                         This may also increase performance depending on compiler and architecture.
+- `XXH32_ENDJMP`: Switch multi-branch finalization stage of XXH32 by a single jump.
+                  This is generally undesirable for performance, especially when hashing inputs of random sizes.
+                  But depending on exact architecture and compiler, a jump might provide slightly better performance on small inputs. Disabled by default.
+- `XXH_NO_STDLIB`: Disable invocation of `<stdlib.h>` functions, notably `malloc()` and `free()`.
+                   `libxxhash`'s `XXH*_createState()` will always fail and return `NULL`.
+                   But one-shot hashing (like `XXH32()`) or streaming using statically allocated states
+                   still work as expected.
+                   This build flag is useful for embedded environments without dynamic allocation.
+- `XXH_STATIC_LINKING_ONLY`: gives access to internal state declaration, required for static allocation.
+                             Incompatible with dynamic linking, due to risks of ABI changes.
+- `XXH_NO_XXH3` : removes symbols related to `XXH3` (both 64 & 128 bits) from generated binary.
+                  Useful to reduce binary size, notably for applications which do not employ `XXH3`.
+- `XXH_NO_LONG_LONG`: removes compilation of algorithms relying on 64-bit types (`XXH3` and `XXH64`). Only `XXH32` will be compiled.
+                      Useful for targets (architectures and compilers) without 64-bit support.
+- `XXH_IMPORT`: MSVC specific: should only be defined for dynamic linking, as it prevents linkage errors.
+- `XXH_CPU_LITTLE_ENDIAN`: By default, endianness is determined by a runtime test resolved at compile time.
+                           If, for some reason, the compiler cannot simplify the runtime test, it can cost performance.
+                           It's possible to skip auto-detection and simply state that the architecture is little-endian by setting this macro to 1.
+                           Setting it to 0 states big-endian.
+- `XXH_DEBUGLEVEL` : When set to any value >= 1, enables `assert()` statements.
+                     This (slightly) slows down execution, but may help finding bugs during debugging sessions.
+
+When compiling the Command Line Interface `xxhsum` using `make`, the following environment variables can also be set :
+- `DISPATCH=1` : use `xxh_x86dispatch.c`, to automatically select between `scalar`, `sse2`, `avx2` or `avx512` instruction set at runtime, depending on local host. This option is only valid for `x86`/`x64` systems.
+- `XXH_1ST_SPEED_TARGET` : select an initial speed target, expressed in MB/s, for the first speed test in benchmark mode. Benchmark will adjust the target at subsequent iterations, but the first test is made "blindly" by targeting this speed. Currently conservatively set to 10 MB/s, to support very slow (emulated) platforms.
+
+### Building xxHash - Using vcpkg
+
+You can download and install xxHash using the [vcpkg](https://github.com/Microsoft/vcpkg) dependency manager:
+
+    git clone https://github.com/Microsoft/vcpkg.git
+    cd vcpkg
+    ./bootstrap-vcpkg.sh
+    ./vcpkg integrate install
+    ./vcpkg install xxhash
+
+The xxHash port in vcpkg is kept up to date by Microsoft team members and community contributors. If the version is out of date, please [create an issue or pull request](https://github.com/Microsoft/vcpkg) on the vcpkg repository.
+
+### Building and Using xxHash - tipi.build
+
+You can work on xxHash and depend on it in your [tipi.build](https://tipi.build) projects by adding the following entry to your `.tipi/deps`:
+
+```json
+{
+    "Cyan4973/xxHash": { "@": "v0.8.1" }
+}
+```
+
+An example of such usage can be found in the `/cli` folder of this project which, if built as root project will depend on the release `v0.8.1` of xxHash
+
+
+To contribute to xxHash itself use tipi.build on this repository (change the target name appropriately to `linux` or `macos` or `windows`):
+
+```bash
+tipi . -t <target> --test all
+```
+
+### Example
+
+The simplest example calls xxhash 64-bit variant as a one-shot function
+generating a hash value from a single buffer, and invoked from a C/C++ program:
+
+```C
+#include "xxhash.h"
+
+    (...)
+    XXH64_hash_t hash = XXH64(buffer, size, seed);
+}
+```
+
+Streaming variant is more involved, but makes it possible to provide data incrementally:
+
+```C
+#include "stdlib.h"   /* abort() */
+#include "xxhash.h"
+
+
+XXH64_hash_t calcul_hash_streaming(FileHandler fh)
+{
+    /* create a hash state */
+    XXH64_state_t* const state = XXH64_createState();
+    if (state==NULL) abort();
+
+    size_t const bufferSize = SOME_SIZE;
+    void* const buffer = malloc(bufferSize);
+    if (buffer==NULL) abort();
+
+    /* Initialize state with selected seed */
+    XXH64_hash_t const seed = 0;   /* or any other value */
+    if (XXH64_reset(state, seed) == XXH_ERROR) abort();
+
+    /* Feed the state with input data, any size, any number of times */
+    (...)
+    while ( /* some data left */ ) {
+        size_t const length = get_more_data(buffer, bufferSize, fh);
+        if (XXH64_update(state, buffer, length) == XXH_ERROR) abort();
+        (...)
+    }
+    (...)
+
+    /* Produce the final hash value */
+    XXH64_hash_t const hash = XXH64_digest(state);
+
+    /* State could be re-used; but in this example, it is simply freed  */
+    free(buffer);
+    XXH64_freeState(state);
+
+    return hash;
+}
+```
+
+
+### License
+
+The library files `xxhash.c` and `xxhash.h` are BSD licensed.
+The utility `xxhsum` is GPL licensed.
+
+
+### Other programming languages
+
+Beyond the C reference version,
+xxHash is also available from many different programming languages,
+thanks to great contributors.
+They are [listed here](http://www.xxhash.com/#other-languages).
+
+
+### Packaging status
+
+Many distributions bundle a package manager
+which allows easy xxhash installation as both a `libxxhash` library
+and `xxhsum` command line interface.
+
+[![Packaging status](https://repology.org/badge/vertical-allrepos/xxhash.svg)](https://repology.org/project/xxhash/versions)
+
+
+### Special Thanks
+
+- Takayuki Matsuoka, aka @t-mat, for creating `xxhsum -c` and great support during early xxh releases
+- Mathias Westerdahl, aka @JCash, for introducing the first version of `XXH64`
+- Devin Hussey, aka @easyaspi314, for incredible low-level optimizations on `XXH3` and `XXH128`

third_party/xxhash/SECURITY.md

@@ -0,0 +1,13 @@
+# Security Policy
+
+## Supported Versions
+
+Security updates are applied only to the latest release.
+
+## Reporting a Vulnerability
+
+If you have discovered a security vulnerability in this project, please report it privately. **Do not disclose it as a public issue.** This gives us time to work with you to fix the issue before public exposure, reducing the chance that the exploit will be used before a patch is released.
+
+Please disclose it at [security advisory](https://github.com/Cyan4973/xxHash/security/advisories/new).
+
+This project is maintained by a team of volunteers on a reasonable-effort basis. As such, please give us at least 90 days to work on a fix before public exposure.

third_party/xxhash/appveyor.yml

@@ -0,0 +1,115 @@
+#---------------------------------#
+#      general configuration      #
+#---------------------------------#
+version: 1.0.{build}
+max_jobs: 2
+
+#---------------------------------#
+#    environment configuration    #
+#---------------------------------#
+clone_depth: 2
+environment:
+  matrix:
+  - COMPILER: "visual"
+    ARCH: "x64"
+    TEST_XXHSUM: "true"
+  - COMPILER: "visual"
+    ARCH: "Win32"
+    TEST_XXHSUM: "true"
+  - COMPILER: "visual"
+    ARCH: "Win32"
+    APPVEYOR_BUILD_WORKER_IMAGE: Visual Studio 2013
+    TEST_XXHSUM: "true"
+  - COMPILER: "visual"
+    ARCH: "ARM"
+# Below tests are now disabled due to redundancy.
+# Their equivalent already runs correctly on Github Actions.
+#  - COMPILER: "visual"
+#    ARCH: "x64"
+#    APPVEYOR_BUILD_WORKER_IMAGE: Visual Studio 2017
+#    TEST_XXHSUM: "true"
+#  - COMPILER: "visual"
+#    ARCH: "ARM64"
+#    APPVEYOR_BUILD_WORKER_IMAGE: Visual Studio 2017
+#    # note: ARM64 is not available with Visual Studio 14 2015, which is default for Appveyor
+
+# The following tests were also flacky on Appveyor, for various reasons.
+#  - COMPILER: "gcc"
+#    PLATFORM: "mingw64"
+#  - COMPILER: "gcc"
+#    PLATFORM: "mingw32"
+#  - COMPILER: "gcc"
+#    PLATFORM: "clang"
+
+install:
+  - ECHO Installing %COMPILER% %PLATFORM% %ARCH%
+  - MKDIR bin
+  - if [%COMPILER%]==[gcc] SET PATH_ORIGINAL=%PATH%
+  - if [%COMPILER%]==[gcc] (
+      SET "PATH_MINGW32=c:\MinGW\bin;c:\MinGW\usr\bin" &&
+      SET "PATH_MINGW64=c:\msys64\mingw64\bin;c:\msys64\usr\bin" &&
+      COPY C:\MinGW\bin\mingw32-make.exe C:\MinGW\bin\make.exe &&
+      COPY C:\MinGW\bin\gcc.exe C:\MinGW\bin\cc.exe
+    )
+
+#---------------------------------#
+#       build configuration       #
+#---------------------------------#
+build_script:
+  - if [%PLATFORM%]==[mingw32] SET PATH=%PATH_MINGW32%;%PATH_ORIGINAL%
+  - if [%PLATFORM%]==[mingw64] SET PATH=%PATH_MINGW64%;%PATH_ORIGINAL%
+  - if [%PLATFORM%]==[clang]   SET PATH=%PATH_MINGW64%;%PATH_ORIGINAL%
+  - ECHO ***
+  - ECHO Building %COMPILER% %PLATFORM% %ARCH%
+  - ECHO ***
+
+  - if [%COMPILER%]==[gcc] (
+      if [%PLATFORM%]==[clang] (
+        clang -v
+      ) ELSE (
+        gcc -v
+      )
+    )
+  - if [%COMPILER%]==[gcc] (
+      echo ----- &&
+      make -v &&
+      echo ----- &&
+      if not [%PLATFORM%]==[clang] (
+        if [%PLATFORM%]==[mingw32] ( SET CPPFLAGS=-DPOOL_MT=0 ) &&
+        make -B clean test MOREFLAGS=-Werror
+      ) ELSE (
+        SET CXXFLAGS=--std=c++14 &&
+        make -B clean test CC=clang CXX=clang++ MOREFLAGS="--target=x86_64-w64-mingw32 -Werror -Wno-pass-failed" NO_C90_TEST=true
+      ) &&
+      make -C tests/bench
+    )
+    # note 1: strict c90 tests with clang fail, due to (erroneous) presence on `inline` keyword in some included system file
+    # note 2: multi-threading code doesn't work with mingw32, disabled through POOL_MT=0
+    # note 3: clang requires C++14 to compile sort because its own code contains c++14-only code
+
+  - if [%COMPILER%]==[visual] (
+      cd cmake_unofficial &&
+      cmake . -DCMAKE_BUILD_TYPE=Release -A %ARCH% -DXXHASH_C_FLAGS="/WX" &&
+      cmake --build . --config Release
+    )
+
+#---------------------------------#
+#       tests configuration       #
+#---------------------------------#
+test_script:
+  # note: can only run x86 and x64 binaries on Appveyor
+  # note: if %COMPILER%==gcc, xxhsum was already tested within `make test`
+  - if [%TEST_XXHSUM%]==[true] (
+      ECHO ***  &&
+      ECHO Testing %COMPILER% %PLATFORM% %ARCH%  &&
+      ECHO ***  &&
+      cd Release &&
+      xxhsum.exe -bi1 &&
+      ECHO ------- xxhsum tested -------
+    )
+
+
+#---------------------------------#
+#      artifacts configuration    #
+#---------------------------------#
+# none yet

third_party/xxhash/cli/.tipi/deps

@@ -0,0 +1,3 @@
+{
+    "Cyan4973/xxHash": { }
+}

third_party/xxhash/cli/COPYING

@@ -0,0 +1,339 @@
+                    GNU GENERAL PUBLIC LICENSE
+                       Version 2, June 1991
+
+ Copyright (C) 1989, 1991 Free Software Foundation, Inc.,
+ 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+                            Preamble
+
+  The licenses for most software are designed to take away your
+freedom to share and change it.  By contrast, the GNU General Public
+License is intended to guarantee your freedom to share and change free
+software--to make sure the software is free for all its users.  This
+General Public License applies to most of the Free Software
+Foundation's software and to any other program whose authors commit to
+using it.  (Some other Free Software Foundation software is covered by
+the GNU Lesser General Public License instead.)  You can apply it to
+your programs, too.
+
+  When we speak of free software, we are referring to freedom, not
+price.  Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+this service if you wish), that you receive source code or can get it
+if you want it, that you can change the software or use pieces of it
+in new free programs; and that you know you can do these things.
+
+  To protect your rights, we need to make restrictions that forbid
+anyone to deny you these rights or to ask you to surrender the rights.
+These restrictions translate to certain responsibilities for you if you
+distribute copies of the software, or if you modify it.
+
+  For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must give the recipients all the rights that
+you have.  You must make sure that they, too, receive or can get the
+source code.  And you must show them these terms so they know their
+rights.
+
+  We protect your rights with two steps: (1) copyright the software, and
+(2) offer you this license which gives you legal permission to copy,
+distribute and/or modify the software.
+
+  Also, for each author's protection and ours, we want to make certain
+that everyone understands that there is no warranty for this free
+software.  If the software is modified by someone else and passed on, we
+want its recipients to know that what they have is not the original, so
+that any problems introduced by others will not reflect on the original
+authors' reputations.
+
+  Finally, any free program is threatened constantly by software
+patents.  We wish to avoid the danger that redistributors of a free
+program will individually obtain patent licenses, in effect making the
+program proprietary.  To prevent this, we have made it clear that any
+patent must be licensed for everyone's free use or not licensed at all.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+
+                    GNU GENERAL PUBLIC LICENSE
+   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
+
+  0. This License applies to any program or other work which contains
+a notice placed by the copyright holder saying it may be distributed
+under the terms of this General Public License.  The "Program", below,
+refers to any such program or work, and a "work based on the Program"
+means either the Program or any derivative work under copyright law:
+that is to say, a work containing the Program or a portion of it,
+either verbatim or with modifications and/or translated into another
+language.  (Hereinafter, translation is included without limitation in
+the term "modification".)  Each licensee is addressed as "you".
+
+Activities other than copying, distribution and modification are not
+covered by this License; they are outside its scope.  The act of
+running the Program is not restricted, and the output from the Program
+is covered only if its contents constitute a work based on the
+Program (independent of having been made by running the Program).
+Whether that is true depends on what the Program does.
+
+  1. You may copy and distribute verbatim copies of the Program's
+source code as you receive it, in any medium, provided that you
+conspicuously and appropriately publish on each copy an appropriate
+copyright notice and disclaimer of warranty; keep intact all the
+notices that refer to this License and to the absence of any warranty;
+and give any other recipients of the Program a copy of this License
+along with the Program.
+
+You may charge a fee for the physical act of transferring a copy, and
+you may at your option offer warranty protection in exchange for a fee.
+
+  2. You may modify your copy or copies of the Program or any portion
+of it, thus forming a work based on the Program, and copy and
+distribute such modifications or work under the terms of Section 1
+above, provided that you also meet all of these conditions:
+
+    a) You must cause the modified files to carry prominent notices
+    stating that you changed the files and the date of any change.
+
+    b) You must cause any work that you distribute or publish, that in
+    whole or in part contains or is derived from the Program or any
+    part thereof, to be licensed as a whole at no charge to all third
+    parties under the terms of this License.
+
+    c) If the modified program normally reads commands interactively
+    when run, you must cause it, when started running for such
+    interactive use in the most ordinary way, to print or display an
+    announcement including an appropriate copyright notice and a
+    notice that there is no warranty (or else, saying that you provide
+    a warranty) and that users may redistribute the program under
+    these conditions, and telling the user how to view a copy of this
+    License.  (Exception: if the Program itself is interactive but
+    does not normally print such an announcement, your work based on
+    the Program is not required to print an announcement.)
+
+These requirements apply to the modified work as a whole.  If
+identifiable sections of that work are not derived from the Program,
+and can be reasonably considered independent and separate works in
+themselves, then this License, and its terms, do not apply to those
+sections when you distribute them as separate works.  But when you
+distribute the same sections as part of a whole which is a work based
+on the Program, the distribution of the whole must be on the terms of
+this License, whose permissions for other licensees extend to the
+entire whole, and thus to each and every part regardless of who wrote it.
+
+Thus, it is not the intent of this section to claim rights or contest
+your rights to work written entirely by you; rather, the intent is to
+exercise the right to control the distribution of derivative or
+collective works based on the Program.
+
+In addition, mere aggregation of another work not based on the Program
+with the Program (or with a work based on the Program) on a volume of
+a storage or distribution medium does not bring the other work under
+the scope of this License.
+
+  3. You may copy and distribute the Program (or a work based on it,
+under Section 2) in object code or executable form under the terms of
+Sections 1 and 2 above provided that you also do one of the following:
+
+    a) Accompany it with the complete corresponding machine-readable
+    source code, which must be distributed under the terms of Sections
+    1 and 2 above on a medium customarily used for software interchange; or,
+
+    b) Accompany it with a written offer, valid for at least three
+    years, to give any third party, for a charge no more than your
+    cost of physically performing source distribution, a complete
+    machine-readable copy of the corresponding source code, to be
+    distributed under the terms of Sections 1 and 2 above on a medium
+    customarily used for software interchange; or,
+
+    c) Accompany it with the information you received as to the offer
+    to distribute corresponding source code.  (This alternative is
+    allowed only for noncommercial distribution and only if you
+    received the program in object code or executable form with such
+    an offer, in accord with Subsection b above.)
+
+The source code for a work means the preferred form of the work for
+making modifications to it.  For an executable work, complete source
+code means all the source code for all modules it contains, plus any
+associated interface definition files, plus the scripts used to
+control compilation and installation of the executable.  However, as a
+special exception, the source code distributed need not include
+anything that is normally distributed (in either source or binary
+form) with the major components (compiler, kernel, and so on) of the
+operating system on which the executable runs, unless that component
+itself accompanies the executable.
+
+If distribution of executable or object code is made by offering
+access to copy from a designated place, then offering equivalent
+access to copy the source code from the same place counts as
+distribution of the source code, even though third parties are not
+compelled to copy the source along with the object code.
+
+  4. You may not copy, modify, sublicense, or distribute the Program
+except as expressly provided under this License.  Any attempt
+otherwise to copy, modify, sublicense or distribute the Program is
+void, and will automatically terminate your rights under this License.
+However, parties who have received copies, or rights, from you under
+this License will not have their licenses terminated so long as such
+parties remain in full compliance.
+
+  5. You are not required to accept this License, since you have not
+signed it.  However, nothing else grants you permission to modify or
+distribute the Program or its derivative works.  These actions are
+prohibited by law if you do not accept this License.  Therefore, by
+modifying or distributing the Program (or any work based on the
+Program), you indicate your acceptance of this License to do so, and
+all its terms and conditions for copying, distributing or modifying
+the Program or works based on it.
+
+  6. Each time you redistribute the Program (or any work based on the
+Program), the recipient automatically receives a license from the
+original licensor to copy, distribute or modify the Program subject to
+these terms and conditions.  You may not impose any further
+restrictions on the recipients' exercise of the rights granted herein.
+You are not responsible for enforcing compliance by third parties to
+this License.
+
+  7. If, as a consequence of a court judgment or allegation of patent
+infringement or for any other reason (not limited to patent issues),
+conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot
+distribute so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you
+may not distribute the Program at all.  For example, if a patent
+license would not permit royalty-free redistribution of the Program by
+all those who receive copies directly or indirectly through you, then
+the only way you could satisfy both it and this License would be to
+refrain entirely from distribution of the Program.
+
+If any portion of this section is held invalid or unenforceable under
+any particular circumstance, the balance of the section is intended to
+apply and the section as a whole is intended to apply in other
+circumstances.
+
+It is not the purpose of this section to induce you to infringe any
+patents or other property right claims or to contest validity of any
+such claims; this section has the sole purpose of protecting the
+integrity of the free software distribution system, which is
+implemented by public license practices.  Many people have made
+generous contributions to the wide range of software distributed
+through that system in reliance on consistent application of that
+system; it is up to the author/donor to decide if he or she is willing
+to distribute software through any other system and a licensee cannot
+impose that choice.
+
+This section is intended to make thoroughly clear what is believed to
+be a consequence of the rest of this License.
+
+  8. If the distribution and/or use of the Program is restricted in
+certain countries either by patents or by copyrighted interfaces, the
+original copyright holder who places the Program under this License
+may add an explicit geographical distribution limitation excluding
+those countries, so that distribution is permitted only in or among
+countries not thus excluded.  In such case, this License incorporates
+the limitation as if written in the body of this License.
+
+  9. The Free Software Foundation may publish revised and/or new versions
+of the General Public License from time to time.  Such new versions will
+be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+Each version is given a distinguishing version number.  If the Program
+specifies a version number of this License which applies to it and "any
+later version", you have the option of following the terms and conditions
+either of that version or of any later version published by the Free
+Software Foundation.  If the Program does not specify a version number of
+this License, you may choose any version ever published by the Free Software
+Foundation.
+
+  10. If you wish to incorporate parts of the Program into other free
+programs whose distribution conditions are different, write to the author
+to ask for permission.  For software which is copyrighted by the Free
+Software Foundation, write to the Free Software Foundation; we sometimes
+make exceptions for this.  Our decision will be guided by the two goals
+of preserving the free status of all derivatives of our free software and
+of promoting the sharing and reuse of software generally.
+
+                            NO WARRANTY
+
+  11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
+FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.  EXCEPT WHEN
+OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
+PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
+OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS
+TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
+PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
+REPAIR OR CORRECTION.
+
+  12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
+REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
+INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
+OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
+TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
+YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
+PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
+POSSIBILITY OF SUCH DAMAGES.
+
+                     END OF TERMS AND CONDITIONS
+
+            How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+convey the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    <one line to give the program's name and a brief idea of what it does.>
+    Copyright (C) <year>  <name of author>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+Also add information on how to contact you by electronic and paper mail.
+
+If the program is interactive, make it output a short notice like this
+when it starts in an interactive mode:
+
+    Gnomovision version 69, Copyright (C) year name of author
+    Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+    This is free software, and you are welcome to redistribute it
+    under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License.  Of course, the commands you use may
+be called something other than `show w' and `show c'; they could even be
+mouse-clicks or menu items--whatever suits your program.
+
+You should also get your employer (if you work as a programmer) or your
+school, if any, to sign a "copyright disclaimer" for the program, if
+necessary.  Here is a sample; alter the names:
+
+  Yoyodyne, Inc., hereby disclaims all copyright interest in the program
+  `Gnomovision' (which makes passes at compilers) written by James Hacker.
+
+  <signature of Ty Coon>, 1 April 1989
+  Ty Coon, President of Vice
+
+This General Public License does not permit incorporating your program into
+proprietary programs.  If your program is a subroutine library, you may
+consider it more useful to permit linking proprietary applications with the
+library.  If this is what you want to do, use the GNU Lesser General
+Public License instead of this License.

third_party/xxhash/cli/README.md

@@ -0,0 +1,4 @@
+This directory contains source code dedicated to the `xxhsum` command line utility,
+which is a user program of `libxxhash`.
+
+Note that, in contrast with the library `libxxhash`, the command line utility `xxhsum` ships with GPLv2 license.

third_party/xxhash/cli/xsum_arch.h

@@ -0,0 +1,159 @@
+/*
+ * xxhsum - Command line interface for xxhash algorithms
+ * Copyright (C) 2013-2021 Yann Collet
+ *
+ * GPL v2 License
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * You can contact the author at:
+ *   - xxHash homepage: https://www.xxhash.com
+ *   - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+/*
+ * Checks for predefined macros by the compiler to try and get both the arch
+ * and the compiler version.
+ */
+#ifndef XSUM_ARCH_H
+#define XSUM_ARCH_H
+
+#include "xsum_config.h"
+
+#define XSUM_LIB_VERSION XXH_VERSION_MAJOR.XXH_VERSION_MINOR.XXH_VERSION_RELEASE
+#define XSUM_QUOTE(str) #str
+#define XSUM_EXPAND_AND_QUOTE(str) XSUM_QUOTE(str)
+#define XSUM_PROGRAM_VERSION XSUM_EXPAND_AND_QUOTE(XSUM_LIB_VERSION)
+
+
+/* Show compiler versions in WELCOME_MESSAGE. XSUM_CC_VERSION_FMT will return the printf specifiers,
+ * and VERSION will contain the comma separated list of arguments to the XSUM_CC_VERSION_FMT string. */
+#if defined(__clang_version__)
+/* Clang does its own thing. */
+#  ifdef __apple_build_version__
+#    define XSUM_CC_VERSION_FMT "Apple Clang %s"
+#  else
+#    define XSUM_CC_VERSION_FMT "Clang %s"
+#  endif
+#  define XSUM_CC_VERSION  __clang_version__
+#elif defined(__VERSION__)
+/* GCC and ICC */
+#  define XSUM_CC_VERSION_FMT "%s"
+#  ifdef __INTEL_COMPILER /* icc adds its prefix */
+#    define XSUM_CC_VERSION __VERSION__
+#  else /* assume GCC */
+#    define XSUM_CC_VERSION "GCC " __VERSION__
+#  endif
+#elif defined(_MSC_FULL_VER) && defined(_MSC_BUILD)
+/*
+ * MSVC
+ *  "For example, if the version number of the Visual C++ compiler is
+ *   15.00.20706.01, the _MSC_FULL_VER macro evaluates to 150020706."
+ *
+ *   https://docs.microsoft.com/en-us/cpp/preprocessor/predefined-macros?view=vs-2017
+ */
+#  define XSUM_CC_VERSION_FMT "MSVC %02i.%02i.%05i.%02i"
+#  define XSUM_CC_VERSION  _MSC_FULL_VER / 10000000 % 100, _MSC_FULL_VER / 100000 % 100, _MSC_FULL_VER % 100000, _MSC_BUILD
+#elif defined(_MSC_VER) /* old MSVC */
+#  define XSUM_CC_VERSION_FMT "MSVC %02i.%02i"
+#  define XSUM_CC_VERSION _MSC_VER / 100, _MSC_VER % 100
+#elif defined(__TINYC__)
+/* tcc stores its version in the __TINYC__ macro. */
+#  define XSUM_CC_VERSION_FMT "tcc %i.%i.%i"
+#  define XSUM_CC_VERSION __TINYC__ / 10000 % 100, __TINYC__ / 100 % 100, __TINYC__ % 100
+#else
+#  define XSUM_CC_VERSION_FMT "%s"
+#  define XSUM_CC_VERSION "unknown compiler"
+#endif
+
+/* makes the next part easier */
+#if (defined(__x86_64__) || defined(_M_AMD64) || defined(_M_X64)) && !defined(_M_ARM64EC)
+#   define XSUM_ARCH_X64 1
+#   define XSUM_ARCH_X86 "x86_64"
+#elif defined(__i386__) || defined(_M_IX86) || defined(_M_IX86_FP)
+#   define XSUM_ARCH_X86 "i386"
+#endif
+
+/* Try to detect the architecture. */
+#if defined(XSUM_ARCH_X86)
+#  if defined(XXHSUM_DISPATCH)
+#    define XSUM_ARCH XSUM_ARCH_X86 " autoVec"
+#  elif defined(__AVX512F__)
+#    define XSUM_ARCH XSUM_ARCH_X86 " + AVX512"
+#  elif defined(__AVX2__)
+#    define XSUM_ARCH XSUM_ARCH_X86 " + AVX2"
+#  elif defined(__AVX__)
+#    define XSUM_ARCH XSUM_ARCH_X86 " + AVX"
+#  elif defined(_M_X64) || defined(_M_AMD64) || defined(__x86_64__) \
+      || defined(__SSE2__) || (defined(_M_IX86_FP) && _M_IX86_FP == 2)
+#     define XSUM_ARCH XSUM_ARCH_X86 " + SSE2"
+#  else
+#     define XSUM_ARCH XSUM_ARCH_X86
+#  endif
+#elif defined(__aarch64__) || defined(__arm64__) || defined(_M_ARM64) || defined(_M_ARM64EC)
+#  define XSUM_ARCH "aarch64 + NEON"
+#elif defined(__arm__) || defined(__thumb__) || defined(__thumb2__) || defined(_M_ARM)
+/* ARM has a lot of different features that can change xxHash significantly. */
+#  ifdef __ARM_ARCH
+#    define XSUM_ARCH_ARM_VER XSUM_EXPAND_AND_QUOTE(__ARM_ARCH)
+#  else
+#    define XSUM_ARCH_ARM_VER XSUM_EXPAND_AND_QUOTE(_M_ARM)
+#  endif
+#  if defined(_M_ARM) /* windows arm is always thumb-2 */ \
+    || defined(__thumb2__) || (defined(__thumb__) && (__thumb__ == 2 || __ARM_ARCH >= 7))
+#    define XSUM_ARCH_THUMB " Thumb-2"
+#  elif defined(__thumb__)
+#    define XSUM_ARCH_THUMB " Thumb-1"
+#  else
+#    define XSUM_ARCH_THUMB ""
+#  endif
+/* ARMv7 has unaligned by default */
+#  if defined(__ARM_FEATURE_UNALIGNED) || __ARM_ARCH >= 7 || defined(_M_ARM)
+#    define XSUM_ARCH_UNALIGNED " + unaligned"
+#  else
+#    define XSUM_ARCH_UNALIGNED ""
+#  endif
+#  if defined(__ARM_NEON) || defined(__ARM_NEON__) || defined(_M_ARM)
+#    define XSUM_ARCH_NEON " + NEON"
+#  else
+#    define XSUM_ARCH_NEON ""
+#  endif
+#  define XSUM_ARCH "ARMv" XSUM_ARCH_ARM_VER XSUM_ARCH_THUMB XSUM_ARCH_NEON XSUM_ARCH_UNALIGNED
+#elif defined(__powerpc64__) || defined(__ppc64__) || defined(__PPC64__)
+#  if defined(__GNUC__) && defined(__POWER9_VECTOR__)
+#    define XSUM_ARCH "ppc64 + POWER9 vector"
+#  elif defined(__GNUC__) && defined(__POWER8_VECTOR__)
+#    define XSUM_ARCH "ppc64 + POWER8 vector"
+#  else
+#    define XSUM_ARCH "ppc64"
+#  endif
+#elif defined(__powerpc__) || defined(__ppc__) || defined(__PPC__)
+#  define XSUM_ARCH "ppc"
+#elif defined(__AVR)
+#  define XSUM_ARCH "AVR"
+#elif defined(__mips64)
+#  define XSUM_ARCH "mips64"
+#elif defined(__mips)
+#  define XSUM_ARCH "mips"
+#elif defined(__s390x__)
+#  define XSUM_ARCH "s390x"
+#elif defined(__s390__)
+#  define XSUM_ARCH "s390"
+#else
+#  define XSUM_ARCH "unknown"
+#endif
+
+
+#endif /* XSUM_ARCH_H */

third_party/xxhash/cli/xsum_bench.c

@@ -0,0 +1,448 @@
+/*
+ * xsum_bench - Benchmark functions for xxhsum
+ * Copyright (C) 2013-2021 Yann Collet
+ *
+ * GPL v2 License
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * You can contact the author at:
+ *   - xxHash homepage: https://www.xxhash.com
+ *   - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+#include "xsum_output.h"  /* XSUM_logLevel */
+#include "xsum_bench.h"
+#include "xsum_sanity_check.h" /* XSUM_fillTestBuffer */
+#include "xsum_os_specific.h"  /* XSUM_getFileSize */
+#ifndef XXH_STATIC_LINKING_ONLY
+#  define XXH_STATIC_LINKING_ONLY
+#endif
+#include "../xxhash.h"
+#ifdef XXHSUM_DISPATCH
+#  include "../xxh_x86dispatch.h"  /* activate _dispatch() redirectors */
+#endif
+
+#include <stdlib.h>  /* malloc, free */
+#include <assert.h>
+#include <string.h>  /* strlen, memcpy */
+#include <time.h>   /* clock_t, clock, CLOCKS_PER_SEC */
+#include <errno.h>  /* errno */
+
+#define TIMELOOP_S 1
+#define TIMELOOP  (TIMELOOP_S * CLOCKS_PER_SEC)   /* target timing per iteration */
+#define TIMELOOP_MIN (TIMELOOP / 2)               /* minimum timing to validate a result */
+
+/* Each benchmark iteration attempts to match TIMELOOP (1 second).
+ * The nb of loops is adjusted at each iteration to reach that target.
+ * However, initially, there is no information, so 1st iteration blindly targets an arbitrary speed.
+ * If it's too small, it will be adjusted, and a new attempt will be made.
+ * But if it's too large, the first iteration can be very long,
+ * before being fixed at second attempt.
+ * So prefer starting with small speed targets.
+ * XXH_1ST_SPEED_TARGET is defined in MB/s */
+#ifndef XXH_1ST_SPEED_TARGET
+# define XXH_1ST_SPEED_TARGET 10
+#endif
+
+#define MAX_MEM    (2 GB - 64 MB)
+
+static clock_t XSUM_clockSpan( clock_t start )
+{
+    return clock() - start;   /* works even if overflow; Typical max span ~ 30 mn */
+}
+
+static size_t XSUM_findMaxMem(XSUM_U64 requiredMem)
+{
+    size_t const step = 64 MB;
+    void* testmem = NULL;
+
+    requiredMem = (((requiredMem >> 26) + 1) << 26);
+    requiredMem += 2*step;
+    if (requiredMem > MAX_MEM) requiredMem = MAX_MEM;
+
+    while (!testmem) {
+        if (requiredMem > step) requiredMem -= step;
+        else requiredMem >>= 1;
+        testmem = malloc ((size_t)requiredMem);
+    }
+    free (testmem);
+
+    /* keep some space available */
+    if (requiredMem > step) requiredMem -= step;
+    else requiredMem >>= 1;
+
+    return (size_t)requiredMem;
+}
+
+/*
+ * A secret buffer used for benchmarking XXH3's withSecret variants.
+ *
+ * In order for the bench to be realistic, the secret buffer would need to be
+ * pre-generated.
+ *
+ * Adding a pointer to the parameter list would be messy.
+ */
+static XSUM_U8 g_benchSecretBuf[XXH3_SECRET_SIZE_MIN];
+
+/*
+ * Wrappers for the benchmark.
+ *
+ * If you would like to add other hashes to the bench, create a wrapper and add
+ * it to the g_hashesToBench table. It will automatically be added.
+ */
+typedef XSUM_U32 (*hashFunction)(const void* buffer, size_t bufferSize, XSUM_U32 seed);
+
+static XSUM_U32 localXXH32(const void* buffer, size_t bufferSize, XSUM_U32 seed)
+{
+    return XXH32(buffer, bufferSize, seed);
+}
+static XSUM_U32 localXXH32_stream(const void* buffer, size_t bufferSize, XSUM_U32 seed)
+{
+    XXH32_state_t state;
+    (void)seed;
+    XXH32_reset(&state, seed);
+    XXH32_update(&state, buffer, bufferSize);
+    return (XSUM_U32)XXH32_digest(&state);
+}
+static XSUM_U32 localXXH64(const void* buffer, size_t bufferSize, XSUM_U32 seed)
+{
+    return (XSUM_U32)XXH64(buffer, bufferSize, seed);
+}
+static XSUM_U32 localXXH64_stream(const void* buffer, size_t bufferSize, XSUM_U32 seed)
+{
+    XXH64_state_t state;
+    (void)seed;
+    XXH64_reset(&state, seed);
+    XXH64_update(&state, buffer, bufferSize);
+    return (XSUM_U32)XXH64_digest(&state);
+}
+static XSUM_U32 localXXH3_64b(const void* buffer, size_t bufferSize, XSUM_U32 seed)
+{
+    (void)seed;
+    return (XSUM_U32)XXH3_64bits(buffer, bufferSize);
+}
+static XSUM_U32 localXXH3_64b_seeded(const void* buffer, size_t bufferSize, XSUM_U32 seed)
+{
+    return (XSUM_U32)XXH3_64bits_withSeed(buffer, bufferSize, seed);
+}
+static XSUM_U32 localXXH3_64b_secret(const void* buffer, size_t bufferSize, XSUM_U32 seed)
+{
+    (void)seed;
+    return (XSUM_U32)XXH3_64bits_withSecret(buffer, bufferSize, g_benchSecretBuf, sizeof(g_benchSecretBuf));
+}
+static XSUM_U32 localXXH3_128b(const void* buffer, size_t bufferSize, XSUM_U32 seed)
+{
+    (void)seed;
+    return (XSUM_U32)(XXH3_128bits(buffer, bufferSize).low64);
+}
+static XSUM_U32 localXXH3_128b_seeded(const void* buffer, size_t bufferSize, XSUM_U32 seed)
+{
+    return (XSUM_U32)(XXH3_128bits_withSeed(buffer, bufferSize, seed).low64);
+}
+static XSUM_U32 localXXH3_128b_secret(const void* buffer, size_t bufferSize, XSUM_U32 seed)
+{
+    (void)seed;
+    return (XSUM_U32)(XXH3_128bits_withSecret(buffer, bufferSize, g_benchSecretBuf, sizeof(g_benchSecretBuf)).low64);
+}
+static XSUM_U32 localXXH3_stream(const void* buffer, size_t bufferSize, XSUM_U32 seed)
+{
+    XXH3_state_t state;
+    (void)seed;
+    XXH3_64bits_reset(&state);
+    XXH3_64bits_update(&state, buffer, bufferSize);
+    return (XSUM_U32)XXH3_64bits_digest(&state);
+}
+static XSUM_U32 localXXH3_stream_seeded(const void* buffer, size_t bufferSize, XSUM_U32 seed)
+{
+    XXH3_state_t state;
+    XXH3_INITSTATE(&state);
+    XXH3_64bits_reset_withSeed(&state, (XXH64_hash_t)seed);
+    XXH3_64bits_update(&state, buffer, bufferSize);
+    return (XSUM_U32)XXH3_64bits_digest(&state);
+}
+static XSUM_U32 localXXH128_stream(const void* buffer, size_t bufferSize, XSUM_U32 seed)
+{
+    XXH3_state_t state;
+    (void)seed;
+    XXH3_128bits_reset(&state);
+    XXH3_128bits_update(&state, buffer, bufferSize);
+    return (XSUM_U32)(XXH3_128bits_digest(&state).low64);
+}
+static XSUM_U32 localXXH128_stream_seeded(const void* buffer, size_t bufferSize, XSUM_U32 seed)
+{
+    XXH3_state_t state;
+    XXH3_INITSTATE(&state);
+    XXH3_128bits_reset_withSeed(&state, (XXH64_hash_t)seed);
+    XXH3_128bits_update(&state, buffer, bufferSize);
+    return (XSUM_U32)(XXH3_128bits_digest(&state).low64);
+}
+
+
+typedef struct {
+    const char*  name;
+    hashFunction func;
+} hashInfo;
+
+static const hashInfo g_hashesToBench[] = {
+    { "XXH32",             &localXXH32 },
+    { "XXH64",             &localXXH64 },
+    { "XXH3_64b",          &localXXH3_64b },
+    { "XXH3_64b w/seed",   &localXXH3_64b_seeded },
+    { "XXH3_64b w/secret", &localXXH3_64b_secret },
+    { "XXH128",            &localXXH3_128b },
+    { "XXH128 w/seed",     &localXXH3_128b_seeded },
+    { "XXH128 w/secret",   &localXXH3_128b_secret },
+    { "XXH32_stream",      &localXXH32_stream },
+    { "XXH64_stream",      &localXXH64_stream },
+    { "XXH3_stream",       &localXXH3_stream },
+    { "XXH3_stream w/seed",&localXXH3_stream_seeded },
+    { "XXH128_stream",     &localXXH128_stream },
+    { "XXH128_stream w/seed",&localXXH128_stream_seeded },
+};
+#define NB_HASHFUNC (sizeof(g_hashesToBench) / sizeof(*g_hashesToBench))
+
+#define NB_TESTFUNC (1 + 2 * NB_HASHFUNC)
+int const g_nbTestFunctions = NB_TESTFUNC;
+char g_testIDs[NB_TESTFUNC] = { 0 };
+const char k_testIDs_default[NB_TESTFUNC] = { 0,
+        1 /*XXH32*/, 0,
+        1 /*XXH64*/, 0,
+        1 /*XXH3*/, 0, 0, 0, 0, 0,
+        1 /*XXH128*/ };
+
+int g_nbIterations = NBLOOPS_DEFAULT;
+#define HASHNAME_MAX 29
+static void XSUM_benchHash(hashFunction h, const char* hName, int testID,
+                           const void* buffer, size_t bufferSize)
+{
+    XSUM_U32 nbh_perIteration = (XSUM_U32)((XXH_1ST_SPEED_TARGET MB) / (bufferSize+1)) + 1;
+    int iterationNb, nbIterations = g_nbIterations + !g_nbIterations /* min 1 */;
+    double fastestH = 100000000.;
+    assert(HASHNAME_MAX > 2);
+    XSUM_logVerbose(2, "\r%80s\r", "");       /* Clean display line */
+
+    for (iterationNb = 1; iterationNb <= nbIterations; iterationNb++) {
+        XSUM_U32 r=0;
+        clock_t cStart;
+
+        XSUM_logVerbose(2, "%2i-%-*.*s : %10u ->\r",
+                        iterationNb,
+                        HASHNAME_MAX, HASHNAME_MAX, hName,
+                        (unsigned)bufferSize);
+        cStart = clock();
+        while (clock() == cStart);   /* starts clock() at its exact beginning */
+        cStart = clock();
+
+        {   XSUM_U32 u;
+            for (u=0; u<nbh_perIteration; u++)
+                r += h(buffer, bufferSize, u);
+        }
+        if (r==0) XSUM_logVerbose(3,".\r");  /* do something with r to defeat compiler "optimizing" hash away */
+
+        {   clock_t const nbTicks = XSUM_clockSpan(cStart);
+            double const ticksPerHash = ((double)nbTicks / TIMELOOP) / nbh_perIteration;
+            /*
+             * clock() is the only decent portable timer, but it isn't very
+             * precise.
+             *
+             * Sometimes, this lack of precision is enough that the benchmark
+             * finishes before there are enough ticks to get a meaningful result.
+             *
+             * For example, on a Core 2 Duo (without any sort of Turbo Boost),
+             * the imprecise timer caused peculiar results like so:
+             *
+             *    XXH3_64b                   4800.0 MB/s // conveniently even
+             *    XXH3_64b unaligned         4800.0 MB/s
+             *    XXH3_64b seeded            9600.0 MB/s // magical 2x speedup?!
+             *    XXH3_64b seeded unaligned  4800.0 MB/s
+             *
+             * If we sense a suspiciously low number of ticks, we increase the
+             * iterations until we can get something meaningful.
+             */
+            if (nbTicks < TIMELOOP_MIN) {
+                /* Not enough time spent in benchmarking, risk of rounding bias */
+                if (nbTicks == 0) { /* faster than resolution timer */
+                    nbh_perIteration *= 100;
+                } else {
+                    /*
+                     * update nbh_perIteration so that the next round lasts
+                     * approximately 1 second.
+                     */
+                    double nbh_perSecond = (1 / ticksPerHash) + 1;
+                    if (nbh_perSecond > (double)(4000U<<20)) nbh_perSecond = (double)(4000U<<20);   /* avoid overflow */
+                    nbh_perIteration = (XSUM_U32)nbh_perSecond;
+                }
+                /* g_nbIterations==0 => quick evaluation, no claim of accuracy */
+                if (g_nbIterations>0) {
+                    iterationNb--;   /* new round for a more accurate speed evaluation */
+                    continue;
+                }
+            }
+            if (ticksPerHash < fastestH) fastestH = ticksPerHash;
+            if (fastestH>0.) { /* avoid div by zero */
+                XSUM_logVerbose(2, "%2i-%-*.*s : %10u -> %8.0f it/s (%7.1f MB/s) \r",
+                            iterationNb,
+                            HASHNAME_MAX, HASHNAME_MAX, hName,
+                            (unsigned)bufferSize,
+                            (double)1 / fastestH,
+                            ((double)bufferSize / (1 MB)) / fastestH);
+        }   }
+        {   double nbh_perSecond = (1 / fastestH) + 1;
+            if (nbh_perSecond > (double)(4000U<<20)) nbh_perSecond = (double)(4000U<<20);   /* avoid overflow */
+            nbh_perIteration = (XSUM_U32)nbh_perSecond;
+        }
+    }
+    XSUM_logVerbose(1, "%2i#%-*.*s : %10u -> %8.0f it/s (%7.1f MB/s) \n",
+                    testID,
+                    HASHNAME_MAX, HASHNAME_MAX, hName,
+                    (unsigned)bufferSize,
+                    (double)1 / fastestH,
+                    ((double)bufferSize / (1 MB)) / fastestH);
+    if (XSUM_logLevel<1)
+        XSUM_logVerbose(0, "%u, ", (unsigned)((double)1 / fastestH));
+}
+
+
+/*
+ * Allocates a string containing s1 and s2 concatenated. Acts like strdup.
+ * The result must be freed.
+ */
+static char* XSUM_strcatDup(const char* s1, const char* s2)
+{
+    assert(s1 != NULL);
+    assert(s2 != NULL);
+    {   size_t len1 = strlen(s1);
+        size_t len2 = strlen(s2);
+        char* buf = (char*)malloc(len1 + len2 + 1);
+        if (buf != NULL) {
+            /* strcpy(buf, s1) */
+            memcpy(buf, s1, len1);
+            /* strcat(buf, s2) */
+            memcpy(buf + len1, s2, len2 + 1);
+        }
+        return buf;
+    }
+}
+
+
+/*!
+ * XSUM_benchMem():
+ * buffer: Must be 16-byte aligned.
+ * The real allocated size of buffer is supposed to be >= (bufferSize+3).
+ * returns: 0 on success, 1 if error (invalid mode selected)
+ */
+static void XSUM_benchMem(const void* buffer, size_t bufferSize)
+{
+    assert((((size_t)buffer) & 15) == 0);  /* ensure alignment */
+    XSUM_fillTestBuffer(g_benchSecretBuf, sizeof(g_benchSecretBuf));
+    {   int i;
+        for (i = 1; i < (int)NB_TESTFUNC; i++) {
+            int const hashFuncID = (i-1) / 2;
+            assert(g_hashesToBench[hashFuncID].name != NULL);
+            if (g_testIDs[i] == 0) continue;
+            /* aligned */
+            if ((i % 2) == 1) {
+                XSUM_benchHash(g_hashesToBench[hashFuncID].func, g_hashesToBench[hashFuncID].name, i, buffer, bufferSize);
+            }
+            /* unaligned */
+            if ((i % 2) == 0) {
+                /* Append "unaligned". */
+                char* const hashNameBuf = XSUM_strcatDup(g_hashesToBench[hashFuncID].name, " unaligned");
+                assert(hashNameBuf != NULL);
+                XSUM_benchHash(g_hashesToBench[hashFuncID].func, hashNameBuf, i, ((const char*)buffer)+3, bufferSize);
+                free(hashNameBuf);
+            }
+    }   }
+}
+
+static size_t XSUM_selectBenchedSize(const char* fileName)
+{
+    XSUM_U64 const inFileSize = XSUM_getFileSize(fileName);
+    size_t benchedSize = (size_t) XSUM_findMaxMem(inFileSize);
+    if ((XSUM_U64)benchedSize > inFileSize) benchedSize = (size_t)inFileSize;
+    if (benchedSize < inFileSize) {
+        XSUM_log("Not enough memory for '%s' full size; testing %i MB only...\n", fileName, (int)(benchedSize>>20));
+    }
+    return benchedSize;
+}
+
+
+int XSUM_benchFiles(const char* fileNamesTable[], int nbFiles)
+{
+    int fileIdx;
+    for (fileIdx=0; fileIdx<nbFiles; fileIdx++) {
+        const char* const inFileName = fileNamesTable[fileIdx];
+        assert(inFileName != NULL);
+
+        {   FILE* const inFile = XSUM_fopen( inFileName, "rb" );
+            size_t const benchedSize = XSUM_selectBenchedSize(inFileName);
+            char* const buffer = (char*)calloc(benchedSize+16+3, 1);
+            void* const alignedBuffer = (buffer+15) - (((size_t)(buffer+15)) & 0xF);  /* align on next 16 bytes */
+
+            /* Checks */
+            if (inFile==NULL){
+                XSUM_log("Error: Could not open '%s': %s.\n", inFileName, strerror(errno));
+                free(buffer);
+                exit(11);
+            }
+            if(!buffer) {
+                XSUM_log("\nError: Out of memory.\n");
+                fclose(inFile);
+                exit(12);
+            }
+
+            /* Fill input buffer */
+            {   size_t const readSize = fread(alignedBuffer, 1, benchedSize, inFile);
+                fclose(inFile);
+                if(readSize != benchedSize) {
+                    XSUM_log("\nError: Could not read '%s': %s.\n", inFileName, strerror(errno));
+                    free(buffer);
+                    exit(13);
+            }   }
+
+            /* bench */
+            XSUM_benchMem(alignedBuffer, benchedSize);
+
+            free(buffer);
+    }   }
+    return 0;
+}
+
+
+int XSUM_benchInternal(size_t keySize)
+{
+    void* const buffer = calloc(keySize+16+3, 1);
+    if (buffer == NULL) {
+        XSUM_log("\nError: Out of memory.\n");
+        exit(12);
+    }
+
+    {   const void* const alignedBuffer = ((char*)buffer+15) - (((size_t)((char*)buffer+15)) & 0xF);  /* align on next 16 bytes */
+
+        /* bench */
+        XSUM_logVerbose(1, "Sample of ");
+        if (keySize > 10 KB) {
+            XSUM_logVerbose(1, "%u KB", (unsigned)(keySize >> 10));
+        } else {
+            XSUM_logVerbose(1, "%u bytes", (unsigned)keySize);
+        }
+        XSUM_logVerbose(1, "...        \n");
+
+        XSUM_benchMem(alignedBuffer, keySize);
+        free(buffer);
+    }
+    return 0;
+}

third_party/xxhash/cli/xsum_bench.h

@@ -0,0 +1,51 @@
+/*
+ * xsum_bench - Benchmark functions for xxhsum
+ * Copyright (C) 2013-2021 Yann Collet
+ *
+ * GPL v2 License
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * You can contact the author at:
+ *   - xxHash homepage: https://www.xxhash.com
+ *   - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+#ifndef XSUM_BENCH_H
+#define XSUM_BENCH_H
+
+#include <stddef.h>  /* size_t */
+
+#define NBLOOPS_DEFAULT    3    /* Default number of benchmark iterations */
+
+extern int const g_nbTestFunctions;
+extern char g_testIDs[];  /* size : g_nbTestFunctions */
+extern const char k_testIDs_default[];
+extern int g_nbIterations;
+
+int XSUM_benchInternal(size_t keySize);
+int XSUM_benchFiles(const char* fileNamesTable[], int nbFiles);
+
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* XSUM_BENCH_H */

third_party/xxhash/cli/xsum_config.h

@@ -0,0 +1,214 @@
+/*
+ * xxhsum - Command line interface for xxhash algorithms
+ * Copyright (C) 2013-2021 Yann Collet
+ *
+ * GPL v2 License
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * You can contact the author at:
+ *   - xxHash homepage: https://www.xxhash.com
+ *   - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+/*
+ * This contains various configuration parameters and feature detection for
+ * xxhsum.
+ *
+ * Similar to config.h in Autotools, this should be the first header included.
+ */
+
+#ifndef XSUM_CONFIG_H
+#define XSUM_CONFIG_H
+
+
+/* ************************************
+ *  Compiler Options
+ **************************************/
+/*
+ * Disable Visual C's warnings when using the "insecure" CRT functions instead
+ * of the "secure" _s functions.
+ *
+ * These functions are not portable, and aren't necessary if you are using the
+ * original functions properly.
+ */
+#if defined(_MSC_VER) || defined(_WIN32)
+#  ifndef _CRT_SECURE_NO_WARNINGS
+#    define _CRT_SECURE_NO_WARNINGS
+#  endif
+#endif
+
+/* Under Linux at least, pull in the *64 commands */
+#ifndef _LARGEFILE64_SOURCE
+#  define _LARGEFILE64_SOURCE
+#endif
+#ifndef _FILE_OFFSET_BITS
+#  define _FILE_OFFSET_BITS 64
+#endif
+
+/*
+ * So we can use __attribute__((__format__))
+ */
+#ifdef __GNUC__
+#  define XSUM_ATTRIBUTE(x) __attribute__(x)
+#else
+#  define XSUM_ATTRIBUTE(x)
+#endif
+
+#if !defined(_WIN32) && (defined(__unix__) || defined(__unix) || (defined(__APPLE__) && defined(__MACH__)) /* UNIX-like OS */ \
+   || defined(__midipix__) || defined(__VMS))
+#  if (defined(__APPLE__) && defined(__MACH__)) || defined(__SVR4) || defined(_AIX) || defined(__hpux) /* POSIX.1-2001 (SUSv3) conformant */ \
+     || defined(__DragonFly__) || defined(__FreeBSD__) || defined(__NetBSD__) || defined(__OpenBSD__)  /* BSD distros */
+#    define XSUM_PLATFORM_POSIX_VERSION 200112L
+#  else
+#    if defined(__linux__) || defined(__linux)
+#      ifndef _POSIX_C_SOURCE
+#        define _POSIX_C_SOURCE 200112L  /* use feature test macro */
+#      endif
+#    endif
+#    include <unistd.h>  /* declares _POSIX_VERSION */
+#    if defined(_POSIX_VERSION)  /* POSIX compliant */
+#      define XSUM_PLATFORM_POSIX_VERSION _POSIX_VERSION
+#    else
+#      define XSUM_PLATFORM_POSIX_VERSION 0
+#    endif
+#  endif
+#endif
+#if !defined(XSUM_PLATFORM_POSIX_VERSION)
+#  define XSUM_PLATFORM_POSIX_VERSION -1
+#endif
+
+#if !defined(S_ISREG)
+#  define S_ISREG(x) (((x) & S_IFMT) == S_IFREG)
+#endif
+
+
+/* ************************************
+ * Windows helpers
+ **************************************/
+
+/*
+ * Whether to use the Windows UTF-16 APIs instead of the portable libc 8-bit
+ * ("ANSI") APIs.
+ *
+ * Windows is not UTF-8 clean by default, and the only way to access every file
+ * on the OS is to use UTF-16.
+ *
+ * Do note that xxhsum uses UTF-8 internally and only uses UTF-16 for command
+ * line arguments, console I/O, and opening files.
+ *
+ * Additionally, this guarantees all piped output is UTF-8.
+ */
+#if defined(XSUM_WIN32_USE_WCHAR) && !defined(_WIN32)
+/* We use Windows APIs, only use this on Windows. */
+#  undef XSUM_WIN32_USE_WCHAR
+#endif
+
+#ifndef XSUM_WIN32_USE_WCHAR
+#  if defined(_WIN32)
+#    include <wchar.h>
+#    if WCHAR_MAX == 0xFFFFU /* UTF-16 wchar_t */
+#       define XSUM_WIN32_USE_WCHAR 1
+#    else
+#       define XSUM_WIN32_USE_WCHAR 0
+#    endif
+#  else
+#    define XSUM_WIN32_USE_WCHAR 0
+#  endif
+#endif
+
+#if !XSUM_WIN32_USE_WCHAR
+/*
+ * It doesn't make sense to have one without the other.
+ * Due to XSUM_WIN32_USE_WCHAR being undef'd, this also handles
+ * non-WIN32 platforms.
+ */
+#  undef  XSUM_WIN32_USE_WMAIN
+#  define XSUM_WIN32_USE_WMAIN 0
+#else
+/*
+ * Whether to use wmain() or main().
+ *
+ * wmain() is preferred because we don't have to mess with internal hidden
+ * APIs.
+ *
+ * It always works on MSVC, but in MinGW, it only works on MinGW-w64 with the
+ * -municode flag.
+ *
+ * Therefore we have to use main() -- there is no better option.
+ */
+#  ifndef XSUM_WIN32_USE_WMAIN
+#    if defined(_UNICODE) || defined(UNICODE) /* MinGW -municode */ \
+        || defined(_MSC_VER) /* MSVC */
+#      define XSUM_WIN32_USE_WMAIN 1
+#    else
+#      define XSUM_WIN32_USE_WMAIN 0
+#    endif
+#  endif
+/*
+ * It is always good practice to define these to prevent accidental use of the
+ * ANSI APIs, even if the program primarily uses UTF-8.
+ */
+#  ifndef _UNICODE
+#    define _UNICODE
+#  endif
+#  ifndef UNICODE
+#    define UNICODE
+#  endif
+#endif /* XSUM_WIN32_USE_WCHAR */
+
+#ifndef XSUM_API
+#  ifdef XXH_INLINE_ALL
+#    define XSUM_API static
+#  else
+#    define XSUM_API
+#  endif
+#endif
+
+#ifndef XSUM_NO_TESTS
+#  define XSUM_NO_TESTS 0
+#endif
+
+/* ***************************
+ * Basic types
+ * ***************************/
+
+#if defined(__cplusplus) /* C++ */ \
+ || (defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L)  /* C99 */
+#  include <stdint.h>
+    typedef uint8_t  XSUM_U8;
+    typedef uint32_t XSUM_U32;
+    typedef uint64_t XSUM_U64;
+# else
+#   include <limits.h>
+    typedef unsigned char      XSUM_U8;
+#   if UINT_MAX == 0xFFFFFFFFUL
+      typedef unsigned int     XSUM_U32;
+#   else
+      typedef unsigned long    XSUM_U32;
+#   endif
+    typedef unsigned long long XSUM_U64;
+#endif /* not C++/C99 */
+
+/* ***************************
+ * Common constants
+ * ***************************/
+
+#define KB *( 1<<10)
+#define MB *( 1<<20)
+#define GB *(1U<<30)
+
+
+#endif /* XSUM_CONFIG_H */

third_party/xxhash/cli/xsum_os_specific.c

@@ -0,0 +1,483 @@
+/*
+ * xxhsum - Command line interface for xxhash algorithms
+ * Copyright (C) 2013-2021 Yann Collet
+ *
+ * GPL v2 License
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * You can contact the author at:
+ *   - xxHash homepage: https://www.xxhash.com
+ *   - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+#include "xsum_os_specific.h"  /* XSUM_API */
+#include <sys/stat.h>   /* stat() / _stat64() */
+
+/*
+ * This file contains all of the ugly boilerplate to make xxhsum work across
+ * platforms.
+ */
+#if defined(_MSC_VER) || XSUM_WIN32_USE_WCHAR
+    typedef struct __stat64 XSUM_stat_t;
+# if defined(_MSC_VER)
+    typedef int mode_t;
+# endif
+#else
+    typedef struct stat XSUM_stat_t;
+#endif
+
+#if (defined(__linux__) && (XSUM_PLATFORM_POSIX_VERSION >= 1)) \
+ || (XSUM_PLATFORM_POSIX_VERSION >= 200112L) \
+ || defined(__DJGPP__) \
+ || defined(__MSYS__) \
+ || defined(__HAIKU__)
+#  include <unistd.h>   /* isatty */
+#  define XSUM_IS_CONSOLE(stdStream) isatty(fileno(stdStream))
+#elif defined(MSDOS) || defined(OS2)
+#  include <io.h>       /* _isatty */
+#  define XSUM_IS_CONSOLE(stdStream) _isatty(_fileno(stdStream))
+#elif defined(WIN32) || defined(_WIN32)
+#  include <io.h>      /* _isatty */
+#  include <windows.h> /* DeviceIoControl, HANDLE, FSCTL_SET_SPARSE */
+#  include <stdio.h>   /* FILE */
+static __inline int XSUM_IS_CONSOLE(FILE* stdStream)
+{
+    DWORD dummy;
+    return _isatty(_fileno(stdStream)) && GetConsoleMode((HANDLE)_get_osfhandle(_fileno(stdStream)), &dummy);
+}
+#else
+#  define XSUM_IS_CONSOLE(stdStream) 0
+#endif
+
+#if defined(MSDOS) || defined(OS2) || defined(WIN32) || defined(_WIN32)
+#  include <fcntl.h>   /* _O_BINARY */
+#  include <io.h>      /* _setmode, _fileno, _get_osfhandle */
+#  if !defined(__DJGPP__)
+#    include <windows.h> /* DeviceIoControl, HANDLE, FSCTL_SET_SPARSE */
+#    include <winioctl.h> /* FSCTL_SET_SPARSE */
+#    define XSUM_SET_BINARY_MODE(file) { int const unused=_setmode(_fileno(file), _O_BINARY); (void)unused; }
+#  else
+#    define XSUM_SET_BINARY_MODE(file) setmode(fileno(file), O_BINARY)
+#  endif
+#else
+#  define XSUM_SET_BINARY_MODE(file) ((void)file)
+#endif
+
+XSUM_API int XSUM_isConsole(FILE* stream)
+{
+    return XSUM_IS_CONSOLE(stream);
+}
+
+XSUM_API void XSUM_setBinaryMode(FILE* stream)
+{
+    XSUM_SET_BINARY_MODE(stream);
+}
+
+#if !XSUM_WIN32_USE_WCHAR
+
+XSUM_API FILE* XSUM_fopen(const char* filename, const char* mode)
+{
+    return fopen(filename, mode);
+}
+XSUM_ATTRIBUTE((__format__(__printf__, 2, 0)))
+XSUM_API int XSUM_vfprintf(FILE* stream, const char* format, va_list ap)
+{
+    return vfprintf(stream, format, ap);
+}
+
+static int XSUM_stat(const char* infilename, XSUM_stat_t* statbuf)
+{
+#if defined(_MSC_VER)
+    return _stat64(infilename, statbuf);
+#else
+    return stat(infilename, statbuf);
+#endif
+}
+
+#ifndef XSUM_NO_MAIN
+int main(int argc, const char* argv[])
+{
+    return XSUM_main(argc, argv);
+}
+#endif
+
+/* Unicode helpers for Windows to make UTF-8 act as it should. */
+#else
+#  include <windows.h>
+#  include <wchar.h>
+
+/*****************************************************************************
+ *                       Unicode conversion tools
+ *****************************************************************************/
+
+/*
+ * Converts a UTF-8 string to UTF-16. Acts like strdup. The string must be freed afterwards.
+ * This version allows keeping the output length.
+ */
+static wchar_t* XSUM_widenString(const char* str, int* lenOut)
+{
+    int const len = MultiByteToWideChar(CP_UTF8, 0, str, -1, NULL, 0);
+    if (lenOut != NULL) *lenOut = len;
+    if (len == 0) return NULL;
+    {   wchar_t* buf = (wchar_t*)malloc((size_t)len * sizeof(wchar_t));
+        if (buf != NULL) {
+            if (MultiByteToWideChar(CP_UTF8, 0, str, -1, buf, len) == 0) {
+                free(buf);
+                return NULL;
+       }    }
+       return buf;
+    }
+}
+
+/*
+ * Converts a UTF-16 string to UTF-8. Acts like strdup. The string must be freed afterwards.
+ * This version allows keeping the output length.
+ */
+static char* XSUM_narrowString(const wchar_t *str, int *lenOut)
+{
+    int len = WideCharToMultiByte(CP_UTF8, 0, str, -1, NULL, 0, NULL, NULL);
+    if (lenOut != NULL) *lenOut = len;
+    if (len == 0) return NULL;
+    {   char* const buf = (char*)malloc((size_t)len * sizeof(char));
+        if (buf != NULL) {
+            if (WideCharToMultiByte(CP_UTF8, 0, str, -1, buf, len, NULL, NULL) == 0) {
+                free(buf);
+                return NULL;
+        }    }
+        return buf;
+    }
+}
+
+
+
+/*****************************************************************************
+ *                             File helpers
+ *****************************************************************************/
+/*
+ * fopen wrapper that supports UTF-8
+ *
+ * fopen will only accept ANSI filenames, which means that we can't open Unicode filenames.
+ *
+ * In order to open a Unicode filename, we need to convert filenames to UTF-16 and use _wfopen.
+ */
+XSUM_API FILE* XSUM_fopen(const char* filename, const char* mode)
+{
+    FILE* f = NULL;
+    wchar_t* const wide_filename = XSUM_widenString(filename, NULL);
+    if (wide_filename != NULL) {
+        wchar_t* const wide_mode = XSUM_widenString(mode, NULL);
+        if (wide_mode != NULL) {
+            f = _wfopen(wide_filename, wide_mode);
+            free(wide_mode);
+        }
+        free(wide_filename);
+    }
+    return f;
+}
+
+/*
+ * stat() wrapper which supports UTF-8 filenames.
+ */
+static int XSUM_stat(const char* infilename, XSUM_stat_t* statbuf)
+{
+    int r = -1;
+    wchar_t* const wide_filename = XSUM_widenString(infilename, NULL);
+    if (wide_filename != NULL) {
+        r = _wstat64(wide_filename, statbuf);
+        free(wide_filename);
+    }
+    return r;
+}
+
+/*
+ * In case it isn't available, this is what MSVC 2019 defines in stdarg.h.
+ */
+#if defined(_MSC_VER) && !defined(__clang__) && !defined(va_copy)
+#  define XSUM_va_copy(destination, source) ((destination) = (source))
+#else
+#  define XSUM_va_copy(destination, source) va_copy(destination, source)
+#endif
+
+/*
+ * vasprintf for Windows.
+ */
+XSUM_ATTRIBUTE((__format__(__printf__, 2, 0)))
+static int XSUM_vasprintf(char** strp, const char* format, va_list ap)
+{
+    int size;
+    va_list copy;
+    /*
+     * To be safe, make a va_copy.
+     *
+     * Note that Microsoft doesn't use va_copy in its sample code:
+     *   https://docs.microsoft.com/en-us/cpp/c-runtime-library/reference/vsprintf-vsprintf-l-vswprintf-vswprintf-l-vswprintf-l?view=vs-2019
+     */
+    XSUM_va_copy(copy, ap);
+    /* Calculate how many characters we need */
+    size = _vscprintf(format, ap);
+    va_end(copy);
+
+    if (size < 0) {
+        *strp = NULL;
+        return size;
+    } else {
+        int ret;
+        *strp = (char*) malloc((size_t)size + 1);
+        if (*strp == NULL) {
+            return -1;
+        }
+        /* vsprintf into the new buffer */
+        ret = vsprintf(*strp, format, ap);
+        if (ret < 0) {
+            free(*strp);
+            *strp = NULL;
+        }
+        return ret;
+    }
+}
+
+/*
+ * fprintf wrapper that supports UTF-8.
+ *
+ * fprintf doesn't properly handle Unicode on Windows.
+ *
+ * Additionally, it is codepage sensitive on console and may crash the program.
+ *
+ * Instead, we use vsnprintf, and either print with fwrite or convert to UTF-16
+ * for console output and use the codepage-independent WriteConsoleW.
+ *
+ * Credit to t-mat: https://github.com/t-mat/xxHash/commit/5691423
+ */
+XSUM_ATTRIBUTE((__format__(__printf__, 2, 0)))
+XSUM_API int XSUM_vfprintf(FILE *stream, const char *format, va_list ap)
+{
+    int result;
+    char* u8_str = NULL;
+
+    /*
+     * Generate the UTF-8 output string with vasprintf.
+     */
+    result = XSUM_vasprintf(&u8_str, format, ap);
+
+    if (result >= 0) {
+        const size_t nchar = (size_t)result + 1;
+
+        /*
+         * Check if we are outputting to a console. Don't use XSUM_isConsole
+         * directly -- we don't need to call _get_osfhandle twice.
+         */
+        int fileNb = _fileno(stream);
+        intptr_t handle_raw = _get_osfhandle(fileNb);
+        HANDLE handle = (HANDLE)handle_raw;
+        DWORD dwTemp;
+
+        if (handle_raw < 0) {
+             result = -1;
+        } else if (_isatty(fileNb) && GetConsoleMode(handle, &dwTemp)) {
+            /*
+             * Convert to UTF-16 and output with WriteConsoleW.
+             *
+             * This is codepage independent and works on Windows XP's default
+             * msvcrt.dll.
+             */
+            int len;
+            wchar_t* const u16_buf = XSUM_widenString(u8_str, &len);
+            if (u16_buf == NULL) {
+                result = -1;
+            } else {
+                if (WriteConsoleW(handle, u16_buf, (DWORD)len - 1, &dwTemp, NULL)) {
+                    result = (int)dwTemp;
+                } else {
+                    result = -1;
+                }
+                free(u16_buf);
+            }
+        } else {
+            /* fwrite the UTF-8 string if we are printing to a file */
+            result = (int)fwrite(u8_str, 1, nchar - 1, stream);
+            if (result == 0) {
+                result = -1;
+            }
+        }
+        free(u8_str);
+    }
+    return result;
+}
+
+#ifndef XSUM_NO_MAIN
+/*****************************************************************************
+ *                    Command Line argument parsing
+ *****************************************************************************/
+
+/* Converts a UTF-16 argv to UTF-8. */
+static char** XSUM_convertArgv(int argc, wchar_t* utf16_argv[])
+{
+    char** const utf8_argv = (char**)malloc((size_t)(argc + 1) * sizeof(char*));
+    if (utf8_argv != NULL) {
+        int i;
+        for (i = 0; i < argc; i++) {
+            utf8_argv[i] = XSUM_narrowString(utf16_argv[i], NULL);
+            if (utf8_argv[i] == NULL) {
+                /* Out of memory, whoops. */
+                while (i-- > 0) {
+                    free(utf8_argv[i]);
+                }
+                free(utf8_argv);
+                return NULL;
+            }
+        }
+        utf8_argv[argc] = NULL;
+    }
+    return utf8_argv;
+}
+/* Frees arguments returned by XSUM_convertArgv */
+static void XSUM_freeArgv(int argc, char** argv)
+{
+    int i;
+    if (argv == NULL) {
+        return;
+    }
+    for (i = 0; i < argc; i++) {
+        free(argv[i]);
+    }
+    free(argv);
+}
+
+static int XSUM_wmain(int argc, wchar_t* utf16_argv[])
+{
+    /* Convert the UTF-16 arguments to UTF-8. */
+    char** utf8_argv = XSUM_convertArgv(argc, utf16_argv);
+
+    if (utf8_argv == NULL) {
+        /* An unfortunate but incredibly unlikely error. */
+        fprintf(stderr, "xxhsum: error converting command line arguments!\n");
+        abort();
+    } else {
+        int ret;
+
+        /*
+         * MinGW's terminal uses full block buffering for stderr.
+         *
+         * This is nonstandard behavior and causes text to not display until
+         * the buffer fills.
+         *
+         * `setvbuf()` can easily correct this to make text display instantly.
+         */
+        setvbuf(stderr, NULL, _IONBF, 0);
+
+        /* Call our real main function */
+        ret = XSUM_main(argc, (void*)utf8_argv);
+
+        /* Cleanup */
+        XSUM_freeArgv(argc, utf8_argv);
+        return ret;
+    }
+}
+
+#if XSUM_WIN32_USE_WMAIN
+
+/*
+ * The preferred method of obtaining the real UTF-16 arguments. Always works
+ * on MSVC, sometimes works on MinGW-w64 depending on the compiler flags.
+ */
+#ifdef __cplusplus
+extern "C"
+#endif
+int __cdecl wmain(int argc, wchar_t* utf16_argv[])
+{
+    return XSUM_wmain(argc, utf16_argv);
+}
+#else /* !XSUM_WIN32_USE_WMAIN */
+
+/*
+ * Wrap `XSUM_wmain()` using `main()` and `__wgetmainargs()` on MinGW without
+ * Unicode support.
+ *
+ * `__wgetmainargs()` is used in the CRT startup to retrieve the arguments for
+ * `wmain()`, so we use it on MinGW to emulate `wmain()`.
+ *
+ * It is an internal function and not declared in any public headers, so we
+ * have to declare it manually.
+ *
+ * An alternative that doesn't mess with internal APIs is `GetCommandLineW()`
+ * with `CommandLineToArgvW()`, but the former doesn't expand wildcards and the
+ * latter requires linking to Shell32.dll and its numerous dependencies.
+ *
+ * This method keeps our dependencies to kernel32.dll and the CRT.
+ *
+ * https://docs.microsoft.com/en-us/cpp/c-runtime-library/getmainargs-wgetmainargs?view=vs-2019
+ */
+typedef struct {
+    int newmode;
+} _startupinfo;
+
+#ifdef __cplusplus
+extern "C"
+#endif
+int __cdecl __wgetmainargs(
+    int*          Argc,
+    wchar_t***    Argv,
+    wchar_t***    Env,
+    int           DoWildCard,
+    _startupinfo* StartInfo
+);
+
+int main(int ansi_argc, const char* ansi_argv[])
+{
+    int       utf16_argc;
+    wchar_t** utf16_argv;
+    wchar_t** utf16_envp;         /* Unused but required */
+    _startupinfo startinfo = {0}; /* 0 == don't change new mode */
+
+    /* Get wmain's UTF-16 arguments. Make sure we expand wildcards. */
+    if (__wgetmainargs(&utf16_argc, &utf16_argv, &utf16_envp, 1, &startinfo) < 0)
+        /* In the very unlikely case of an error, use the ANSI arguments. */
+        return XSUM_main(ansi_argc, ansi_argv);
+
+    /* Call XSUM_wmain with our UTF-16 arguments */
+    return XSUM_wmain(utf16_argc, utf16_argv);
+}
+
+#endif /* !XSUM_WIN32_USE_WMAIN */
+#endif /* !XSUM_NO_MAIN */
+#endif /* XSUM_WIN32_USE_WCHAR */
+
+
+/*
+ * Determines whether the file at filename is a directory.
+ */
+XSUM_API int XSUM_isDirectory(const char* filename)
+{
+    XSUM_stat_t statbuf;
+    int r = XSUM_stat(filename, &statbuf);
+#ifdef _MSC_VER
+    if (!r && (statbuf.st_mode & _S_IFDIR)) return 1;
+#else
+    if (!r && S_ISDIR(statbuf.st_mode)) return 1;
+#endif
+    return 0;
+}
+
+/*
+ * Returns the filesize of the file at filename.
+ */
+XSUM_API XSUM_U64 XSUM_getFileSize(const char* filename)
+{
+    XSUM_stat_t statbuf;
+    int r = XSUM_stat(filename, &statbuf);
+    if (r || !S_ISREG(statbuf.st_mode)) return 0;   /* No good... */
+    return (XSUM_U64)statbuf.st_size;
+}

third_party/xxhash/cli/xsum_os_specific.h

@@ -0,0 +1,89 @@
+/*
+ * xxhsum - Command line interface for xxhash algorithms
+ * Copyright (C) 2013-2021 Yann Collet
+ *
+ * GPL v2 License
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * You can contact the author at:
+ *   - xxHash homepage: https://www.xxhash.com
+ *   - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+#ifndef XSUM_OS_SPECIFIC_H
+#define XSUM_OS_SPECIFIC_H
+
+#include "xsum_config.h"
+#include <stdio.h>
+#include <stdarg.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/*
+ * Declared here to be implemented in user code.
+ *
+ * Functions like main(), but is passed UTF-8 arguments even on Windows.
+ */
+XSUM_API int XSUM_main(int argc, const char* argv[]);
+
+/*
+ * Returns whether stream is a console.
+ *
+ * Functionally equivalent to isatty(fileno(stream)).
+ */
+XSUM_API int XSUM_isConsole(FILE* stream);
+
+/*
+ * Sets stream to pure binary mode (a.k.a. no CRLF conversions).
+ */
+XSUM_API void XSUM_setBinaryMode(FILE* stream);
+
+/*
+ * Returns whether the file at filename is a directory.
+ */
+XSUM_API int XSUM_isDirectory(const char* filename);
+
+/*
+ * Returns the file size of the file at filename.
+ */
+XSUM_API XSUM_U64 XSUM_getFileSize(const char* filename);
+
+/*
+ * UTF-8 stdio wrappers primarily for Windows
+ */
+
+/*
+ * fopen() wrapper. Accepts UTF-8 filenames on Windows.
+ *
+ * Specifically, on Windows, the arguments will be converted to UTF-16
+ * and passed to _wfopen().
+ */
+XSUM_API FILE* XSUM_fopen(const char* filename, const char* mode);
+
+/*
+ * vfprintf() wrapper which prints UTF-8 strings to Windows consoles
+ * if applicable.
+ */
+XSUM_ATTRIBUTE((__format__(__printf__, 2, 0)))
+XSUM_API int XSUM_vfprintf(FILE* stream, const char* format, va_list ap);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* XSUM_OS_SPECIFIC_H */

third_party/xxhash/cli/xsum_output.c

@@ -0,0 +1,66 @@
+/*
+ * xxhsum - Command line interface for xxhash algorithms
+ * Copyright (C) 2013-2021 Yann Collet
+ *
+ * GPL v2 License
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * You can contact the author at:
+ *   - xxHash homepage: https://www.xxhash.com
+ *   - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+#include "xsum_output.h"
+#include "xsum_os_specific.h"  /* XSUM_API */
+
+int XSUM_logLevel = 2;
+
+XSUM_ATTRIBUTE((__format__(__printf__, 1, 2)))
+XSUM_API int XSUM_log(const char* format, ...)
+{
+    int ret;
+    va_list ap;
+    va_start(ap, format);
+    ret = XSUM_vfprintf(stderr, format, ap);
+    va_end(ap);
+    return ret;
+}
+
+
+XSUM_ATTRIBUTE((__format__(__printf__, 1, 2)))
+XSUM_API int XSUM_output(const char* format, ...)
+{
+    int ret;
+    va_list ap;
+    va_start(ap, format);
+    ret = XSUM_vfprintf(stdout, format, ap);
+    va_end(ap);
+    return ret;
+}
+
+XSUM_ATTRIBUTE((__format__(__printf__, 2, 3)))
+XSUM_API int XSUM_logVerbose(int minLevel, const char* format, ...)
+{
+    if (XSUM_logLevel >= minLevel) {
+        int ret;
+        va_list ap;
+        va_start(ap, format);
+        ret = XSUM_vfprintf(stderr, format, ap);
+        va_end(ap);
+        return ret;
+    }
+    return 0;
+}

third_party/xxhash/cli/xsum_output.h

@@ -0,0 +1,62 @@
+/*
+ * xxhsum - Command line interface for xxhash algorithms
+ * Copyright (C) 2013-2021 Yann Collet
+ *
+ * GPL v2 License
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * You can contact the author at:
+ *   - xxHash homepage: https://www.xxhash.com
+ *   - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+#ifndef XSUM_OUTPUT_H
+#define XSUM_OUTPUT_H
+
+#include "xsum_config.h"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/*
+ * How verbose the output is.
+ */
+extern int XSUM_logLevel;
+
+/*
+ * Same as fprintf(stderr, format, ...)
+ */
+XSUM_ATTRIBUTE((__format__(__printf__, 1, 2)))
+XSUM_API int XSUM_log(const char *format, ...);
+
+/*
+ * Like XSUM_log, but only outputs if XSUM_logLevel >= minLevel.
+ */
+XSUM_ATTRIBUTE((__format__(__printf__, 2, 3)))
+XSUM_API int XSUM_logVerbose(int minLevel, const char *format, ...);
+
+/*
+ * Same as printf(format, ...)
+ */
+XSUM_ATTRIBUTE((__format__(__printf__, 1, 2)))
+XSUM_API int XSUM_output(const char *format, ...);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* XSUM_OUTPUT_H */

third_party/xxhash/cli/xsum_sanity_check.c

@@ -0,0 +1,694 @@
+/*
+ * xxhsum - Command line interface for xxhash algorithms
+ * Copyright (C) 2013-2021 Yann Collet
+ *
+ * GPL v2 License
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * You can contact the author at:
+ *   - xxHash homepage: https://www.xxhash.com
+ *   - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+#include "xsum_sanity_check.h"
+#include "xsum_output.h"  /* XSUM_log */
+#ifndef XXH_STATIC_LINKING_ONLY
+#  define XXH_STATIC_LINKING_ONLY
+#endif
+#include "../xxhash.h"
+
+#include <stdlib.h>  /* exit */
+#include <assert.h>
+#include <string.h>  /* memcmp */
+
+/* use #define to make them constant, required for initialization */
+#define PRIME32 2654435761U
+#define PRIME64 11400714785074694797ULL
+
+/*
+ * Fills a test buffer with pseudorandom data.
+ *
+ * This is used in the sanity check - its values must not be changed.
+ */
+XSUM_API void XSUM_fillTestBuffer(XSUM_U8* buffer, size_t len)
+{
+    XSUM_U64 byteGen = PRIME32;
+    size_t i;
+
+    assert(buffer != NULL);
+
+    for (i=0; i<len; i++) {
+        buffer[i] = (XSUM_U8)(byteGen>>56);
+        byteGen *= PRIME64;
+    }
+}
+
+
+
+/* ************************************************
+ * Self-test:
+ * ensure results consistency across platforms
+ *********************************************** */
+#if XSUM_NO_TESTS
+XSUM_API void XSUM_sanityCheck(void)
+{
+    XSUM_log("This version of xxhsum is not verified.\n");
+}
+#else
+
+/*
+ * Test data vectors
+ */
+typedef struct {
+    XSUM_U32 len;
+    XSUM_U32 seed;
+    XSUM_U32 Nresult;
+} XSUM_testdata32_t;
+
+typedef struct {
+    XSUM_U32 len;
+    XSUM_U64 seed;
+    XSUM_U64 Nresult;
+} XSUM_testdata64_t;
+
+typedef struct {
+    XSUM_U32 len;
+    XSUM_U64 seed;
+    XXH128_hash_t Nresult;
+} XSUM_testdata128_t;
+
+#define SECRET_SAMPLE_NBBYTES 5
+typedef struct {
+    XSUM_U32 seedLen;
+    XSUM_U32 secretLen;
+    XSUM_U8 byte[SECRET_SAMPLE_NBBYTES];
+} XSUM_testdata_sample_t;
+
+/* XXH32 */
+static const XSUM_testdata32_t XSUM_XXH32_testdata[] = {
+    {   0,       0, 0x02CC5D05U },
+    {   0, PRIME32, 0x36B78AE7U },
+    {   1,       0, 0xCF65B03EU },
+    {   1, PRIME32, 0xB4545AA4U },
+    {  14,       0, 0x1208E7E2U },
+    {  14, PRIME32, 0x6AF1D1FEU },
+    { 222,       0, 0x5BD11DBDU },
+    { 222, PRIME32, 0x58803C5FU }
+};
+
+/* XXH64 */
+static const XSUM_testdata64_t XSUM_XXH64_testdata[] = {
+    {   0,       0, 0xEF46DB3751D8E999ULL },
+    {   0, PRIME32, 0xAC75FDA2929B17EFULL },
+    {   1,       0, 0xE934A84ADB052768ULL },
+    {   1, PRIME32, 0x5014607643A9B4C3ULL },
+    {   4,       0, 0x9136A0DCA57457EEULL },
+    {  14,       0, 0x8282DCC4994E35C8ULL },
+    {  14, PRIME32, 0xC3BD6BF63DEB6DF0ULL },
+    { 222,       0, 0xB641AE8CB691C174ULL },
+    { 222, PRIME32, 0x20CB8AB7AE10C14AULL }
+};
+/*
+ * XXH3:
+ * Due to being a more complex hash function with specializations for certain
+ * lengths, a more extensive test is used for XXH3.
+ */
+
+/* XXH3_64bits, seeded */
+static const XSUM_testdata64_t XSUM_XXH3_testdata[] = {
+    {    0,       0, 0x2D06800538D394C2ULL },  /* empty string */
+    {    0, PRIME64, 0xA8A6B918B2F0364AULL },
+    {    1,       0, 0xC44BDFF4074EECDBULL },  /*  1 -  3 */
+    {    1, PRIME64, 0x032BE332DD766EF8ULL },
+    {    6,       0, 0x27B56A84CD2D7325ULL },  /*  4 -  8 */
+    {    6, PRIME64, 0x84589C116AB59AB9ULL },
+    {   12,       0, 0xA713DAF0DFBB77E7ULL },  /*  9 - 16 */
+    {   12, PRIME64, 0xE7303E1B2336DE0EULL },
+    {   24,       0, 0xA3FE70BF9D3510EBULL },  /* 17 - 32 */
+    {   24, PRIME64, 0x850E80FC35BDD690ULL },
+    {   48,       0, 0x397DA259ECBA1F11ULL },  /* 33 - 64 */
+    {   48, PRIME64, 0xADC2CBAA44ACC616ULL },
+    {   80,       0, 0xBCDEFBBB2C47C90AULL },  /* 65 - 96 */
+    {   80, PRIME64, 0xC6DD0CB699532E73ULL },
+    {  195,       0, 0xCD94217EE362EC3AULL },  /* 129-240 */
+    {  195, PRIME64, 0xBA68003D370CB3D9ULL },
+
+    {  403,       0, 0xCDEB804D65C6DEA4ULL },  /* one block, last stripe is overlapping */
+    {  403, PRIME64, 0x6259F6ECFD6443FDULL },
+    {  512,       0, 0x617E49599013CB6BULL },  /* one block, finishing at stripe boundary */
+    {  512, PRIME64, 0x3CE457DE14C27708ULL },
+    { 2048,       0, 0xDD59E2C3A5F038E0ULL },  /* 2 blocks, finishing at block boundary */
+    { 2048, PRIME64, 0x66F81670669ABABCULL },
+    { 2099,       0, 0xC6B9D9B3FC9AC765ULL },  /* 2 blocks + 1 partial block, to detect off-by-one scrambling issues, like #816 */
+    { 2099, PRIME64, 0x184F316843663974ULL },
+    { 2240,       0, 0x6E73A90539CF2948ULL },  /* 3 blocks, finishing at stripe boundary */
+    { 2240, PRIME64, 0x757BA8487D1B5247ULL },
+    { 2367,       0, 0xCB37AEB9E5D361EDULL },  /* 3 blocks, last stripe is overlapping */
+    { 2367, PRIME64, 0xD2DB3415B942B42AULL }
+};
+/* XXH3_64bits, custom secret */
+static const XSUM_testdata64_t XSUM_XXH3_withSecret_testdata[] = {
+    {       0, 0, 0x3559D64878C5C66CULL },  /* empty string */
+    {       1, 0, 0x8A52451418B2DA4DULL },  /*  1 -  3 */
+    {       6, 0, 0x82C90AB0519369ADULL },  /*  4 -  8 */
+    {      12, 0, 0x14631E773B78EC57ULL },  /*  9 - 16 */
+    {      24, 0, 0xCDD5542E4A9D9FE8ULL },  /* 17 - 32 */
+    {      48, 0, 0x33ABD54D094B2534ULL },  /* 33 - 64 */
+    {      80, 0, 0xE687BA1684965297ULL },  /* 65 - 96 */
+    {     195, 0, 0xA057273F5EECFB20ULL },  /* 129-240 */
+
+    {     403, 0, 0x14546019124D43B8ULL },  /* one block, last stripe is overlapping */
+    {     512, 0, 0x7564693DD526E28DULL },  /* one block, finishing at stripe boundary */
+    {    2048, 0, 0xD32E975821D6519FULL },  /* >= 2 blodcks, at least one scrambling */
+    {    2367, 0, 0x293FA8E5173BB5E7ULL },  /* >= 2 blocks, at least one scrambling, last stripe unaligned */
+
+    { 64*10*3, 0, 0x751D2EC54BC6038BULL }   /* exactly 3 full blocks, not a multiple of 256 */
+};
+/* XXH3_128bits, seeded */
+static const XSUM_testdata128_t XSUM_XXH128_testdata[] = {
+    {    0,       0, { 0x6001C324468D497FULL, 0x99AA06D3014798D8ULL } },  /* empty string */
+    {    0, PRIME32, { 0x5444F7869C671AB0ULL, 0x92220AE55E14AB50ULL } },
+    {    1,       0, { 0xC44BDFF4074EECDBULL, 0xA6CD5E9392000F6AULL } },  /*  1 -  3 */
+    {    1, PRIME32, { 0xB53D5557E7F76F8DULL, 0x89B99554BA22467CULL } },
+    {    6,       0, { 0x3E7039BDDA43CFC6ULL, 0x082AFE0B8162D12AULL } },  /*  4 -  8 */
+    {    6, PRIME32, { 0x269D8F70BE98856EULL, 0x5A865B5389ABD2B1ULL } },
+    {   12,       0, { 0x061A192713F69AD9ULL, 0x6E3EFD8FC7802B18ULL } },  /*  9 - 16 */
+    {   12, PRIME32, { 0x9BE9F9A67F3C7DFBULL, 0xD7E09D518A3405D3ULL } },
+    {   24,       0, { 0x1E7044D28B1B901DULL, 0x0CE966E4678D3761ULL } },  /* 17 - 32 */
+    {   24, PRIME32, { 0xD7304C54EBAD40A9ULL, 0x3162026714A6A243ULL } },
+    {   48,       0, { 0xF942219AED80F67BULL, 0xA002AC4E5478227EULL } },  /* 33 - 64 */
+    {   48, PRIME32, { 0x7BA3C3E453A1934EULL, 0x163ADDE36C072295ULL } },
+    {   81,       0, { 0x5E8BAFB9F95FB803ULL, 0x4952F58181AB0042ULL } },  /* 65 - 96 */
+    {   81, PRIME32, { 0x703FBB3D7A5F755CULL, 0x2724EC7ADC750FB6ULL } },
+    {  222,       0, { 0xF1AEBD597CEC6B3AULL, 0x337E09641B948717ULL } },  /* 129-240 */
+    {  222, PRIME32, { 0xAE995BB8AF917A8DULL, 0x91820016621E97F1ULL } },
+
+    {  403,       0, { 0xCDEB804D65C6DEA4ULL, 0x1B6DE21E332DD73DULL } },  /* one block, last stripe is overlapping */
+    {  403, PRIME64, { 0x6259F6ECFD6443FDULL, 0xBED311971E0BE8F2ULL } },
+    {  512,       0, { 0x617E49599013CB6BULL, 0x18D2D110DCC9BCA1ULL } },  /* one block, finishing at stripe boundary */
+    {  512, PRIME64, { 0x3CE457DE14C27708ULL, 0x925D06B8EC5B8040ULL } },
+    { 2048,       0, { 0xDD59E2C3A5F038E0ULL, 0xF736557FD47073A5ULL } },  /* 2 blocks, finishing at block boundary */
+    { 2048, PRIME32, { 0x230D43F30206260BULL, 0x7FB03F7E7186C3EAULL } },
+    { 2240,       0, { 0x6E73A90539CF2948ULL, 0xCCB134FBFA7CE49DULL } },  /* 3 blocks, finishing at stripe boundary */
+    { 2240, PRIME32, { 0xED385111126FBA6FULL, 0x50A1FE17B338995FULL } },
+    { 2367,       0, { 0xCB37AEB9E5D361EDULL, 0xE89C0F6FF369B427ULL } },  /* 3 blocks, last stripe is overlapping */
+    { 2367, PRIME32, { 0x6F5360AE69C2F406ULL, 0xD23AAE4B76C31ECBULL } }
+};
+
+/* XXH128, custom secret */
+static const XSUM_testdata128_t XSUM_XXH128_withSecret_testdata[] = {
+    {  0, 0, { 0x005923CCEECBE8AEULL, 0x5F70F4EA232F1D38ULL } },  /* empty string */
+    {  1, 0, { 0x8A52451418B2DA4DULL, 0x3A66AF5A9819198EULL } },  /*  1 -  3 */
+    {  6, 0, { 0x0B61C8ACA7D4778FULL, 0x376BD91B6432F36DULL } },  /*  4 -  8 */
+    { 12, 0, { 0xAF82F6EBA263D7D8ULL, 0x90A3C2D839F57D0FULL } }   /*  9 - 16 */
+};
+
+#define SECRET_SIZE_MAX 9867
+static const XSUM_testdata_sample_t XSUM_XXH3_generateSecret_testdata[] = {
+    {                              0, 192, { 0xE7, 0x8C, 0x77, 0x77, 0x00 } },
+    {                              1, 240, { 0x2B, 0x3E, 0xDE, 0xC1, 0x00 } },
+    {     XXH3_SECRET_SIZE_MIN -   1, 277, { 0xE8, 0x39, 0x6C, 0xCC, 0x7B } },
+    { XXH3_SECRET_DEFAULT_SIZE + 500, SECRET_SIZE_MAX, { 0xD6, 0x1C, 0x41, 0x17, 0xB3 } }
+};
+
+static void XSUM_checkResult32(XXH32_hash_t r1, XXH32_hash_t r2)
+{
+    static int nbTests = 1;
+    if (r1!=r2) {
+        XSUM_log("\rError: 32-bit hash test %i: Internal sanity check failed!\n", nbTests);
+        XSUM_log("\rGot 0x%08X, expected 0x%08X.\n", (unsigned)r1, (unsigned)r2);
+        XSUM_log("\rNote: If you modified the hash functions, make sure to either update the values\n"
+                  "or temporarily recompile with XSUM_NO_TESTS=1.\n");
+        exit(1);
+    }
+    nbTests++;
+}
+
+static void XSUM_checkResult64(XXH64_hash_t r1, XXH64_hash_t r2)
+{
+    static int nbTests = 1;
+    if (r1!=r2) {
+        XSUM_log("\rError: 64-bit hash test %i: Internal sanity check failed!\n", nbTests);
+        XSUM_log("\rGot 0x%08X%08XULL, expected 0x%08X%08XULL.\n",
+                (unsigned)(r1>>32), (unsigned)r1, (unsigned)(r2>>32), (unsigned)r2);
+        XSUM_log("\rNote: If you modified the hash functions, make sure to either update the values\n"
+                  "or temporarily recompile with XSUM_NO_TESTS=1.\n");
+        exit(1);
+    }
+    nbTests++;
+}
+
+static void XSUM_checkResult128(XXH128_hash_t r1, XXH128_hash_t r2)
+{
+    static int nbTests = 1;
+    if ((r1.low64 != r2.low64) || (r1.high64 != r2.high64)) {
+        XSUM_log("\rError: 128-bit hash test %i: Internal sanity check failed.\n", nbTests);
+        XSUM_log("\rGot { 0x%08X%08XULL, 0x%08X%08XULL }, expected { 0x%08X%08XULL, 0x%08X%08XULL } \n",
+                (unsigned)(r1.low64>>32), (unsigned)r1.low64, (unsigned)(r1.high64>>32), (unsigned)r1.high64,
+                (unsigned)(r2.low64>>32), (unsigned)r2.low64, (unsigned)(r2.high64>>32), (unsigned)r2.high64 );
+        XSUM_log("\rNote: If you modified the hash functions, make sure to either update the values\n"
+                  "or temporarily recompile with XSUM_NO_TESTS=1.\n");
+        exit(1);
+    }
+    nbTests++;
+}
+
+
+static void XSUM_testXXH32(const void* data, const XSUM_testdata32_t* testData)
+{
+    XXH32_state_t *state = XXH32_createState();
+    size_t pos;
+
+    size_t len = testData->len;
+    XSUM_U32 seed = testData->seed;
+    XSUM_U32 Nresult = testData->Nresult;
+
+    if (len == 0) {
+        data = NULL;
+    } else {
+        assert(data != NULL);
+    }
+
+    assert(state != NULL);
+
+    XSUM_checkResult32(XXH32(data, len, seed), Nresult);
+
+    (void)XXH32_reset(state, seed);
+    (void)XXH32_update(state, data, len);
+    XSUM_checkResult32(XXH32_digest(state), Nresult);
+
+    (void)XXH32_reset(state, seed);
+    for (pos=0; pos<len; pos++)
+        (void)XXH32_update(state, ((const char*)data)+pos, 1);
+    XSUM_checkResult32(XXH32_digest(state), Nresult);
+    XXH32_freeState(state);
+}
+
+static void XSUM_testXXH64(const void* data, const XSUM_testdata64_t* testData)
+{
+    XXH64_state_t *state = XXH64_createState();
+    size_t pos;
+
+    size_t len = (size_t)testData->len;
+    XSUM_U64 seed = testData->seed;
+    XSUM_U64 Nresult = testData->Nresult;
+
+    if (len == 0) {
+        data = NULL;
+    } else {
+        assert(data != NULL);
+    }
+
+    assert(state != NULL);
+
+    XSUM_checkResult64(XXH64(data, len, seed), Nresult);
+
+    (void)XXH64_reset(state, seed);
+    (void)XXH64_update(state, data, len);
+    XSUM_checkResult64(XXH64_digest(state), Nresult);
+
+    (void)XXH64_reset(state, seed);
+    for (pos=0; pos<len; pos++)
+        (void)XXH64_update(state, ((const char*)data)+pos, 1);
+    XSUM_checkResult64(XXH64_digest(state), Nresult);
+    XXH64_freeState(state);
+}
+
+/*
+ * Used to get "random" (but actually 100% reproducible) lengths for
+ * XSUM_XXH3_randomUpdate.
+ */
+static XSUM_U32 XSUM_rand(void)
+{
+    static XSUM_U64 seed = PRIME32;
+    seed *= PRIME64;
+    return (XSUM_U32)(seed >> 40);
+}
+
+/*
+ * Technically, XXH3_64bits_update is identical to XXH3_128bits_update as of
+ * v0.8.0, but we treat them as separate.
+ */
+typedef XXH_errorcode (*XSUM_XXH3_update_t)(XXH3_state_t* state, const void* input, size_t length);
+
+/*
+ * Runs the passed XXH3_update variant on random lengths. This is to test the
+ * more complex logic of the update function, catching bugs like this one:
+ *    https://github.com/Cyan4973/xxHash/issues/378
+ */
+static void XSUM_XXH3_randomUpdate(XXH3_state_t* state, const void* data,
+                                   size_t len, XSUM_XXH3_update_t update_fn)
+{
+    size_t p = 0;
+    while (p < len) {
+        size_t const modulo = len > 2 ? len : 2;
+        size_t l = (size_t)(XSUM_rand()) % modulo;
+        if (p + l > len) l = len - p;
+        (void)update_fn(state, (const char*)data+p, l);
+        p += l;
+    }
+}
+
+static void XSUM_testXXH3(const void* data, const XSUM_testdata64_t* testData)
+{
+    size_t len = testData->len;
+    XSUM_U64 seed = testData->seed;
+    XSUM_U64 Nresult = testData->Nresult;
+    if (len == 0) {
+        data = NULL;
+    } else {
+        assert(data != NULL);
+    }
+    {   XSUM_U64 const Dresult = XXH3_64bits_withSeed(data, len, seed);
+        XSUM_checkResult64(Dresult, Nresult);
+    }
+
+    /* check that the no-seed variant produces same result as seed==0 */
+    if (seed == 0) {
+        XSUM_U64 const Dresult = XXH3_64bits(data, len);
+        XSUM_checkResult64(Dresult, Nresult);
+    }
+
+    /* check that the combination of
+     * XXH3_generateSecret_fromSeed() and XXH3_64bits_withSecretandSeed()
+     * results in exactly the same hash generation as XXH3_64bits_withSeed() */
+    {   char secretBuffer[XXH3_SECRET_DEFAULT_SIZE+1];
+        char* const secret = secretBuffer + 1;  /* intentional unalignment */
+        XXH3_generateSecret_fromSeed(secret, seed);
+        {   XSUM_U64 const Dresult = XXH3_64bits_withSecretandSeed(data, len, secret, XXH3_SECRET_DEFAULT_SIZE, seed);
+            XSUM_checkResult64(Dresult, Nresult);
+    }   }
+
+    /* streaming API test */
+    {   XXH3_state_t* const state = XXH3_createState();
+        assert(state != NULL);
+        /* single ingestion */
+        (void)XXH3_64bits_reset_withSeed(state, seed);
+        (void)XXH3_64bits_update(state, data, len);
+        XSUM_checkResult64(XXH3_64bits_digest(state), Nresult);
+
+        /* random ingestion */
+        (void)XXH3_64bits_reset_withSeed(state, seed);
+        XSUM_XXH3_randomUpdate(state, data, len, &XXH3_64bits_update);
+        XSUM_checkResult64(XXH3_64bits_digest(state), Nresult);
+
+        /* byte by byte ingestion */
+        {   size_t pos;
+            (void)XXH3_64bits_reset_withSeed(state, seed);
+            for (pos=0; pos<len; pos++)
+                (void)XXH3_64bits_update(state, ((const char*)data)+pos, 1);
+            XSUM_checkResult64(XXH3_64bits_digest(state), Nresult);
+        }
+
+        /* check that streaming with a combination of
+         * XXH3_generateSecret_fromSeed() and XXH3_64bits_reset_withSecretandSeed()
+         * results in exactly the same hash generation as XXH3_64bits_reset_withSeed() */
+        {   char secretBuffer[XXH3_SECRET_DEFAULT_SIZE+1];
+            char* const secret = secretBuffer + 1;  /* intentional unalignment */
+            XXH3_generateSecret_fromSeed(secret, seed);
+            /* single ingestion */
+            (void)XXH3_64bits_reset_withSecretandSeed(state, secret, XXH3_SECRET_DEFAULT_SIZE, seed);
+            (void)XXH3_64bits_update(state, data, len);
+            XSUM_checkResult64(XXH3_64bits_digest(state), Nresult);
+        }
+
+        XXH3_freeState(state);
+    }
+
+}
+
+
+#ifndef XXH3_MIDSIZE_MAX
+# define XXH3_MIDSIZE_MAX 240
+#endif
+
+static void XSUM_testXXH3_withSecret(const void* data, const void* secret,
+                                     size_t secretSize, const XSUM_testdata64_t* testData)
+{
+    size_t len = (size_t)testData->len;
+    XSUM_U64 Nresult = testData->Nresult;
+
+    if (len == 0) {
+        data = NULL;
+    } else {
+        assert(data != NULL);
+    }
+    {   XSUM_U64 const Dresult = XXH3_64bits_withSecret(data, len, secret, secretSize);
+        XSUM_checkResult64(Dresult, Nresult);
+    }
+
+    /* check that XXH3_64bits_withSecretandSeed()
+     * results in exactly the same return value as XXH3_64bits_withSecret() */
+    if (len > XXH3_MIDSIZE_MAX)
+    {   XSUM_U64 const Dresult = XXH3_64bits_withSecretandSeed(data, len, secret, secretSize, 0);
+        XSUM_checkResult64(Dresult, Nresult);
+    }
+
+    /* streaming API test */
+    {   XXH3_state_t *state = XXH3_createState();
+        assert(state != NULL);
+        (void)XXH3_64bits_reset_withSecret(state, secret, secretSize);
+        (void)XXH3_64bits_update(state, data, len);
+        XSUM_checkResult64(XXH3_64bits_digest(state), Nresult);
+
+        /* random ingestion */
+        (void)XXH3_64bits_reset_withSecret(state, secret, secretSize);
+        XSUM_XXH3_randomUpdate(state, data, len, &XXH3_64bits_update);
+        XSUM_checkResult64(XXH3_64bits_digest(state), Nresult);
+
+        /* byte by byte ingestion */
+        {   size_t pos;
+            (void)XXH3_64bits_reset_withSecret(state, secret, secretSize);
+            for (pos=0; pos<len; pos++)
+                (void)XXH3_64bits_update(state, ((const char*)data)+pos, 1);
+            XSUM_checkResult64(XXH3_64bits_digest(state), Nresult);
+        }
+
+        /* check that XXH3_64bits_reset_withSecretandSeed()
+         * results in exactly the same return value as XXH3_64bits_reset_withSecret() */
+         if (len > XXH3_MIDSIZE_MAX) {
+            /* single ingestion */
+            (void)XXH3_64bits_reset_withSecretandSeed(state, secret, secretSize, 0);
+            (void)XXH3_64bits_update(state, data, len);
+            XSUM_checkResult64(XXH3_64bits_digest(state), Nresult);
+        }
+
+        XXH3_freeState(state);
+    }
+}
+
+static void XSUM_testXXH128(const void* data, const XSUM_testdata128_t* testData)
+{
+    size_t len = (size_t)testData->len;
+    XSUM_U64 seed = testData->seed;
+    XXH128_hash_t const Nresult = testData->Nresult;
+    if (len == 0) {
+        data = NULL;
+    } else {
+        assert(data != NULL);
+    }
+
+    {   XXH128_hash_t const Dresult = XXH3_128bits_withSeed(data, len, seed);
+        XSUM_checkResult128(Dresult, Nresult);
+    }
+
+    /* check that XXH128() is identical to XXH3_128bits_withSeed() */
+    {   XXH128_hash_t const Dresult2 = XXH128(data, len, seed);
+        XSUM_checkResult128(Dresult2, Nresult);
+    }
+
+    /* check that the no-seed variant produces same result as seed==0 */
+    if (seed == 0) {
+        XXH128_hash_t const Dresult = XXH3_128bits(data, len);
+        XSUM_checkResult128(Dresult, Nresult);
+    }
+
+    /* check that the combination of
+     * XXH3_generateSecret_fromSeed() and XXH3_128bits_withSecretandSeed()
+     * results in exactly the same hash generation as XXH3_64bits_withSeed() */
+    {   char secretBuffer[XXH3_SECRET_DEFAULT_SIZE+1];
+        char* const secret = secretBuffer + 1;  /* intentional unalignment */
+        XXH3_generateSecret_fromSeed(secret, seed);
+        {   XXH128_hash_t const Dresult = XXH3_128bits_withSecretandSeed(data, len, secret, XXH3_SECRET_DEFAULT_SIZE, seed);
+            XSUM_checkResult128(Dresult, Nresult);
+    }   }
+
+    /* streaming API test */
+    {   XXH3_state_t *state = XXH3_createState();
+        assert(state != NULL);
+
+        /* single ingestion */
+        (void)XXH3_128bits_reset_withSeed(state, seed);
+        (void)XXH3_128bits_update(state, data, len);
+        XSUM_checkResult128(XXH3_128bits_digest(state), Nresult);
+
+        /* random ingestion */
+        (void)XXH3_128bits_reset_withSeed(state, seed);
+        XSUM_XXH3_randomUpdate(state, data, len, &XXH3_128bits_update);
+        XSUM_checkResult128(XXH3_128bits_digest(state), Nresult);
+
+        /* byte by byte ingestion */
+        {   size_t pos;
+            (void)XXH3_128bits_reset_withSeed(state, seed);
+            for (pos=0; pos<len; pos++)
+                (void)XXH3_128bits_update(state, ((const char*)data)+pos, 1);
+            XSUM_checkResult128(XXH3_128bits_digest(state), Nresult);
+        }
+
+        /* check that streaming with a combination of
+         * XXH3_generateSecret_fromSeed() and XXH3_128bits_reset_withSecretandSeed()
+         * results in exactly the same hash generation as XXH3_128bits_reset_withSeed() */
+        {   char secretBuffer[XXH3_SECRET_DEFAULT_SIZE+1];
+            char* const secret = secretBuffer + 1;  /* intentional unalignment */
+            XXH3_generateSecret_fromSeed(secret, seed);
+            /* single ingestion */
+            (void)XXH3_128bits_reset_withSecretandSeed(state, secret, XXH3_SECRET_DEFAULT_SIZE, seed);
+            (void)XXH3_128bits_update(state, data, len);
+            XSUM_checkResult128(XXH3_128bits_digest(state), Nresult);
+        }
+
+        XXH3_freeState(state);
+    }
+}
+
+static void XSUM_testXXH128_withSecret(const void* data, const void* secret, size_t secretSize, const XSUM_testdata128_t* testData)
+{
+    size_t len = testData->len;
+    XXH128_hash_t Nresult = testData->Nresult;
+    if (len == 0) {
+        data = NULL;
+    } else {
+        assert(data != NULL);
+    }
+    {   XXH128_hash_t const Dresult = XXH3_128bits_withSecret(data, len, secret, secretSize);
+        XSUM_checkResult128(Dresult, Nresult);
+    }
+
+    /* check that XXH3_128bits_withSecretandSeed()
+     * results in exactly the same return value as XXH3_128bits_withSecret() */
+    if (len > XXH3_MIDSIZE_MAX)
+    {   XXH128_hash_t const Dresult = XXH3_128bits_withSecretandSeed(data, len, secret, secretSize, 0);
+        XSUM_checkResult128(Dresult, Nresult);
+    }
+
+    /* streaming API test */
+    {   XXH3_state_t* const state = XXH3_createState();
+        assert(state != NULL);
+        (void)XXH3_128bits_reset_withSecret(state, secret, secretSize);
+        (void)XXH3_128bits_update(state, data, len);
+        XSUM_checkResult128(XXH3_128bits_digest(state), Nresult);
+
+        /* random ingestion */
+        (void)XXH3_128bits_reset_withSecret(state, secret, secretSize);
+        XSUM_XXH3_randomUpdate(state, data, len, &XXH3_128bits_update);
+        XSUM_checkResult128(XXH3_128bits_digest(state), Nresult);
+
+        /* byte by byte ingestion */
+        {   size_t pos;
+            (void)XXH3_128bits_reset_withSecret(state, secret, secretSize);
+            for (pos=0; pos<len; pos++)
+                (void)XXH3_128bits_update(state, ((const char*)data)+pos, 1);
+            XSUM_checkResult128(XXH3_128bits_digest(state), Nresult);
+        }
+
+        /* check that XXH3_128bits_reset_withSecretandSeed()
+         * results in exactly the same return value as XXH3_128bits_reset_withSecret() */
+         if (len > XXH3_MIDSIZE_MAX) {
+            /* single ingestion */
+            (void)XXH3_128bits_reset_withSecretandSeed(state, secret, secretSize, 0);
+            (void)XXH3_128bits_update(state, data, len);
+            XSUM_checkResult128(XXH3_128bits_digest(state), Nresult);
+        }
+
+        XXH3_freeState(state);
+    }
+}
+
+static void XSUM_testSecretGenerator(const void* customSeed, const XSUM_testdata_sample_t* testData)
+{
+    static int nbTests = 1;
+    const int sampleIndex[SECRET_SAMPLE_NBBYTES] = { 0, 62, 131, 191, 241 };  /* position of sampled bytes */
+    XSUM_U8 secretBuffer[SECRET_SIZE_MAX] = {0};
+    XSUM_U8 samples[SECRET_SAMPLE_NBBYTES];
+    int i;
+
+    assert(testData->secretLen <= SECRET_SIZE_MAX);
+    XXH3_generateSecret(secretBuffer, testData->secretLen, customSeed, testData->seedLen);
+    for (i=0; i<SECRET_SAMPLE_NBBYTES; i++) {
+        samples[i] = secretBuffer[sampleIndex[i]];
+    }
+    if (memcmp(samples, testData->byte, sizeof(testData->byte))) {
+        XSUM_log("\rError: Secret generation test %i: Internal sanity check failed. \n", nbTests);
+        XSUM_log("\rGot { 0x%02X, 0x%02X, 0x%02X, 0x%02X, 0x%02X }, expected { 0x%02X, 0x%02X, 0x%02X, 0x%02X, 0x%02X } \n",
+                samples[0], samples[1], samples[2], samples[3], samples[4],
+                testData->byte[0], testData->byte[1], testData->byte[2], testData->byte[3], testData->byte[4] );
+        exit(1);
+    }
+    nbTests++;
+}
+
+/*!
+ * XSUM_sanityCheck():
+ * Runs a sanity check before the benchmark.
+ *
+ * Exits on an incorrect output.
+ */
+XSUM_API void XSUM_sanityCheck(void)
+{
+    size_t i;
+#define SANITY_BUFFER_SIZE 2367
+    XSUM_U8 sanityBuffer[SANITY_BUFFER_SIZE];
+    const void* const secret = sanityBuffer + 7;
+    const size_t secretSize = XXH3_SECRET_SIZE_MIN + 11;
+    assert(sizeof(sanityBuffer) >= 7 + secretSize);
+
+    XSUM_fillTestBuffer(sanityBuffer, sizeof(sanityBuffer));
+
+    /* XXH32 */
+    for (i = 0; i < (sizeof(XSUM_XXH32_testdata)/sizeof(XSUM_XXH32_testdata[0])); i++) {
+        XSUM_testXXH32(sanityBuffer, &XSUM_XXH32_testdata[i]);
+    }
+    /* XXH64 */
+    for (i = 0; i < (sizeof(XSUM_XXH64_testdata)/sizeof(XSUM_XXH64_testdata[0])); i++) {
+        XSUM_testXXH64(sanityBuffer, &XSUM_XXH64_testdata[i]);
+    }
+    /* XXH3_64bits, seeded */
+    for (i = 0; i < (sizeof(XSUM_XXH3_testdata)/sizeof(XSUM_XXH3_testdata[0])); i++) {
+        XSUM_testXXH3(sanityBuffer, &XSUM_XXH3_testdata[i]);
+    }
+    /* XXH3_64bits, custom secret */
+    for (i = 0; i < (sizeof(XSUM_XXH3_withSecret_testdata)/sizeof(XSUM_XXH3_withSecret_testdata[0])); i++) {
+        XSUM_testXXH3_withSecret(sanityBuffer, secret, secretSize, &XSUM_XXH3_withSecret_testdata[i]);
+    }
+    /* XXH128 */
+    for (i = 0; i < (sizeof(XSUM_XXH128_testdata)/sizeof(XSUM_XXH128_testdata[0])); i++) {
+        XSUM_testXXH128(sanityBuffer, &XSUM_XXH128_testdata[i]);
+    }
+    /* XXH128 with custom Secret */
+    for (i = 0; i < (sizeof(XSUM_XXH128_withSecret_testdata)/sizeof(XSUM_XXH128_withSecret_testdata[0])); i++) {
+        XSUM_testXXH128_withSecret(sanityBuffer, secret, secretSize, &XSUM_XXH128_withSecret_testdata[i]);
+    }
+    /* secret generator */
+    for (i = 0; i < (sizeof(XSUM_XXH3_generateSecret_testdata)/sizeof(XSUM_XXH3_generateSecret_testdata[0])); i++) {
+        assert(XSUM_XXH3_generateSecret_testdata[i].seedLen <= SANITY_BUFFER_SIZE);
+        XSUM_testSecretGenerator(sanityBuffer, &XSUM_XXH3_generateSecret_testdata[i]);
+    }
+
+    XSUM_logVerbose(3, "\r%70s\r", "");       /* Clean display line */
+    XSUM_logVerbose(3, "Sanity check -- all tests ok\n");
+}
+
+#endif /* !XSUM_NO_TESTS */

third_party/xxhash/cli/xsum_sanity_check.h

@@ -0,0 +1,60 @@
+/*
+ * xxhsum - Command line interface for xxhash algorithms
+ * Copyright (C) 2013-2021 Yann Collet
+ *
+ * GPL v2 License
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * You can contact the author at:
+ *   - xxHash homepage: https://www.xxhash.com
+ *   - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+#ifndef XSUM_SANITY_CHECK_H
+#define XSUM_SANITY_CHECK_H
+
+#include "xsum_config.h"  /* XSUM_API, XSUM_U8 */
+
+#include <stddef.h>   /* size_t */
+
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/*
+ * Runs a series of self-tests.
+ *
+ * Exits if any of these tests fail, printing a message to stderr.
+ *
+ * If XSUM_NO_TESTS is defined to non-zero,
+ * this will instead print a warning if this is called (e.g. via xxhsum -b).
+ */
+XSUM_API void XSUM_sanityCheck(void);
+
+/*
+ * Fills a test buffer with pseudorandom data.
+ *
+ * This is used in the sanity check and the benchmarks.
+ * Its values must not be changed.
+ */
+XSUM_API void XSUM_fillTestBuffer(XSUM_U8* buffer, size_t len);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* XSUM_SANITY_CHECK_H */

third_party/xxhash/cli/xxhsum.1

@@ -0,0 +1,106 @@
+.TH "XXHSUM" "1" "December 2021" "xxhsum 0.8.1" "User Commands"
+.SH "NAME"
+\fBxxhsum\fR \- print or check xxHash non\-cryptographic checksums
+.SH "SYNOPSIS"
+\fBxxhsum [<OPTION>] \|\.\|\.\|\. [<FILE>] \|\.\|\.\|\.\fR \fBxxhsum \-b [<OPTION>] \|\.\|\.\|\.\fR
+.P
+\fBxxh32sum\fR is equivalent to \fBxxhsum \-H0\fR, \fBxxh64sum\fR is equivalent to \fBxxhsum \-H1\fR, \fBxxh128sum\fR is equivalent to \fBxxhsum \-H2\fR
+.SH "DESCRIPTION"
+Print or check xxHash (32, 64 or 128 bits) checksums\. When no \fIFILE\fR, read standard input, except if it\'s the console\. When \fIFILE\fR is \fB\-\fR, read standard input even if it\'s the console\.
+.P
+\fBxxhsum\fR supports a command line syntax similar but not identical to md5sum(1)\. Differences are: \fBxxhsum\fR doesn\'t have text/binary mode switch (\fB\-b\fR, \fB\-t\fR); \fBxxhsum\fR always treats files as binary file; \fBxxhsum\fR has a hash selection switch (\fB\-H\fR);
+.P
+As xxHash is a fast non\-cryptographic checksum algorithm, \fBxxhsum\fR should not be used for security related purposes\.
+.P
+\fBxxhsum \-b\fR invokes benchmark mode\. See \fIOPTIONS\fR and \fIEXAMPLES\fR for details\.
+.SH "OPTIONS"
+.TP
+\fB\-V\fR, \fB\-\-version\fR
+Displays xxhsum version and exits
+.TP
+\fB\-H\fR\fIHASHTYPE\fR
+Hash selection\. \fIHASHTYPE\fR means \fB0\fR=XXH32, \fB1\fR=XXH64, \fB2\fR=XXH128, \fB3\fR=XXH3\. Note that \fB\-H3\fR triggers \fB\-\-tag\fR, which can\'t be skipped (this is to reduce risks of confusion with \fB\-H2\fR (\fBXXH64\fR))\. Alternatively, \fIHASHTYPE\fR \fB32\fR=XXH32, \fB64\fR=XXH64, \fB128\fR=XXH128\. Default value is \fB1\fR (64bits)
+.TP
+\fB\-\-tag\fR
+Output in the BSD style\.
+.TP
+\fB\-\-little\-endian\fR
+Set output hexadecimal checksum value as little endian convention\. By default, value is displayed as big endian\.
+.TP
+\fB\-h\fR, \fB\-\-help\fR
+Displays help and exits
+.P
+\fBThe following four options are useful only when verifying checksums (\fB\-c\fR)\fR
+.TP
+\fB\-c\fR, \fB\-\-check\fR \fIFILE\fR
+Read xxHash sums from \fIFILE\fR and check them
+.TP
+\fB\-q\fR, \fB\-\-quiet\fR
+Don\'t print OK for each successfully verified file
+.TP
+\fB\-\-strict\fR
+Return an error code if any line in the file is invalid, not just if some checksums are wrong\. This policy is disabled by default, though UI will prompt an informational message if any line in the file is detected invalid\.
+.TP
+\fB\-\-status\fR
+Don\'t output anything\. Status code shows success\.
+.TP
+\fB\-w\fR, \fB\-\-warn\fR
+Emit a warning message about each improperly formatted checksum line\.
+.P
+\fBThe following options are useful only benchmark purpose\fR
+.TP
+\fB\-b\fR
+Benchmark mode\. See \fIEXAMPLES\fR for details\.
+.TP
+\fB\-b#\fR
+Specify ID of variant to be tested\. Multiple variants can be selected, separated by a \',\' comma\.
+.TP
+\fB\-B\fR\fIBLOCKSIZE\fR
+Only useful for benchmark mode (\fB\-b\fR)\. See \fIEXAMPLES\fR for details\. \fIBLOCKSIZE\fR specifies benchmark mode\'s test data block size in bytes\. Default value is 102400
+.TP
+\fB\-i\fR\fIITERATIONS\fR
+Only useful for benchmark mode (\fB\-b\fR)\. See \fIEXAMPLES\fR for details\. \fIITERATIONS\fR specifies number of iterations in benchmark\. Single iteration lasts approximately 1000 milliseconds\. Default value is 3
+.SH "EXIT STATUS"
+\fBxxhsum\fR exit \fB0\fR on success, \fB1\fR if at least one file couldn\'t be read or doesn\'t have the same checksum as the \fB\-c\fR option\.
+.SH "EXAMPLES"
+Output xxHash (64bit) checksum values of specific files to standard output
+.IP "" 4
+.nf
+$ xxhsum \-H1 foo bar baz
+.fi
+.IP "" 0
+.P
+Output xxHash (32bit and 64bit) checksum values of specific files to standard output, and redirect it to \fBxyz\.xxh32\fR and \fBqux\.xxh64\fR
+.IP "" 4
+.nf
+$ xxhsum \-H0 foo bar baz > xyz\.xxh32
+$ xxhsum \-H1 foo bar baz > qux\.xxh64
+.fi
+.IP "" 0
+.P
+Read xxHash sums from specific files and check them
+.IP "" 4
+.nf
+$ xxhsum \-c xyz\.xxh32 qux\.xxh64
+.fi
+.IP "" 0
+.P
+Benchmark xxHash algorithm\. By default, \fBxxhsum\fR benchmarks xxHash main variants on a synthetic sample of 100 KB, and print results into standard output\. The first column is the algorithm, the second column is the source data size in bytes, the third column is the number of hashes generated per second (throughput), and finally the last column translates speed in megabytes per second\.
+.IP "" 4
+.nf
+$ xxhsum \-b
+.fi
+.IP "" 0
+.P
+In the following example, the sample to hash is set to 16384 bytes, the variants to be benched are selected by their IDs, and each benchmark test is repeated 10 times, for increased accuracy\.
+.IP "" 4
+.nf
+$ xxhsum \-b1,2,3 \-i10 \-B16384
+.fi
+.IP "" 0
+.SH "BUGS"
+Report bugs at: https://github\.com/Cyan4973/xxHash/issues/
+.SH "AUTHOR"
+Yann Collet
+.SH "SEE ALSO"
+md5sum(1)

third_party/xxhash/cli/xxhsum.1.md

@@ -0,0 +1,150 @@
+xxhsum(1) -- print or check xxHash non-cryptographic checksums
+==============================================================
+
+SYNOPSIS
+--------
+
+`xxhsum [<OPTION>] ... [<FILE>] ...`
+`xxhsum -b [<OPTION>] ...`
+
+`xxh32sum` is equivalent to `xxhsum -H0`,
+`xxh64sum` is equivalent to `xxhsum -H1`,
+`xxh128sum` is equivalent to `xxhsum -H2`.
+
+
+DESCRIPTION
+-----------
+
+Print or check xxHash (32, 64 or 128 bits) checksums.
+When no <FILE>, read standard input, except if it's the console.
+When <FILE> is `-`, read standard input even if it's the console.
+
+`xxhsum` supports a command line syntax similar but not identical to md5sum(1).
+Differences are:
+`xxhsum` doesn't have text/binary mode switch (`-b`, `-t`);
+`xxhsum` always treats files as binary file;
+`xxhsum` has a hash selection switch (`-H`);
+
+As xxHash is a fast non-cryptographic checksum algorithm,
+`xxhsum` should not be used for security related purposes.
+
+`xxhsum -b` invokes benchmark mode. See [OPTIONS](#OPTIONS) and [EXAMPLES](#EXAMPLES) for details.
+
+OPTIONS
+-------
+
+* `-V`, `--version`:
+  Displays xxhsum version and exits
+
+* `-H`<HASHTYPE>:
+  Hash selection. <HASHTYPE> means `0`=XXH32, `1`=XXH64, `2`=XXH128, `3`=XXH3.
+  Note that `-H3` triggers `--tag`, which can't be skipped
+  (this is to reduce risks of confusion with `-H2` (`XXH64`)).
+  Alternatively, <HASHTYPE> `32`=XXH32, `64`=XXH64, `128`=XXH128.
+  Default value is `1` (XXH64)
+
+* `--tag`:
+  Output in the BSD style.
+
+* `--little-endian`:
+  Set output hexadecimal checksum value as little endian convention.
+  By default, value is displayed as big endian.
+
+* `-h`, `--help`:
+  Displays help and exits
+
+**The following four options are useful only when verifying checksums (`-c`)**
+
+* `-c`, `--check` <FILE>:
+  Read xxHash sums from <FILE> and check them
+
+* `-q`, `--quiet`:
+  Don't print OK for each successfully verified file
+
+* `--strict`:
+  Return an error code if any line in the file is invalid,
+  not just if some checksums are wrong.
+  This policy is disabled by default,
+  though UI will prompt an informational message
+  if any line in the file is detected invalid.
+
+* `--status`:
+  Don't output anything. Status code shows success.
+
+* `-w`, `--warn`:
+  Emit a warning message about each improperly formatted checksum line.
+
+**The following options are useful only benchmark purpose**
+
+* `-b`:
+  Benchmark mode.  See [EXAMPLES](#EXAMPLES) for details.
+
+* `-b#`:
+  Specify ID of variant to be tested.
+  Multiple variants can be selected, separated by a ',' comma.
+
+* `-B`<BLOCKSIZE>:
+  Only useful for benchmark mode (`-b`). See [EXAMPLES](#EXAMPLES) for details.
+  <BLOCKSIZE> specifies benchmark mode's test data block size in bytes.
+  Default value is 102400
+
+* `-i`<ITERATIONS>:
+  Only useful for benchmark mode (`-b`). See [EXAMPLES](#EXAMPLES) for details.
+  <ITERATIONS> specifies number of iterations in benchmark. Single iteration
+  lasts approximately 1000 milliseconds. Default value is 3
+
+EXIT STATUS
+-----------
+
+`xxhsum` exit `0` on success, `1` if at least one file couldn't be read or
+doesn't have the same checksum as the `-c` option.
+
+EXAMPLES
+--------
+
+Output xxHash (64bit) checksum values of specific files to standard output
+
+    $ xxhsum -H1 foo bar baz
+
+Output xxHash (32bit and 64bit) checksum values of specific files to standard
+output, and redirect it to `xyz.xxh32` and `qux.xxh64`
+
+    $ xxhsum -H0 foo bar baz > xyz.xxh32
+    $ xxhsum -H1 foo bar baz > qux.xxh64
+
+Read xxHash sums from specific files and check them
+
+    $ xxhsum -c xyz.xxh32 qux.xxh64
+
+Benchmark xxHash algorithm.
+By default, `xxhsum` benchmarks xxHash main variants
+on a synthetic sample of 100 KB,
+and print results into standard output.
+The first column is the algorithm,
+the second column is the source data size in bytes,
+the third column is the number of hashes generated per second (throughput),
+and finally the last column translates speed in megabytes per second.
+
+    $ xxhsum -b
+
+In the following example,
+the sample to hash is set to 16384 bytes,
+the variants to be benched are selected by their IDs,
+and each benchmark test is repeated 10 times, for increased accuracy.
+
+    $ xxhsum -b1,2,3 -i10 -B16384
+
+BUGS
+----
+
+Report bugs at: https://github.com/Cyan4973/xxHash/issues/
+
+AUTHOR
+------
+
+Yann Collet
+
+SEE ALSO
+--------
+
+md5sum(1)

third_party/xxhash/cmake_unofficial/.gitignore

@@ -0,0 +1,12 @@
+# cmake artifacts
+
+CMakeCache.txt
+CMakeFiles
+Makefile
+cmake_install.cmake
+
+
+# make compilation results
+
+*.dylib
+*.a

third_party/xxhash/cmake_unofficial/CMakeLists.txt

@@ -0,0 +1,204 @@
+# To the extent possible under law, the author(s) have dedicated all
+# copyright and related and neighboring rights to this software to
+# the public domain worldwide. This software is distributed without
+# any warranty.
+#
+# For details, see <https://creativecommons.org/publicdomain/zero/1.0/>.
+
+cmake_minimum_required (VERSION 2.8.12 FATAL_ERROR)
+
+set(XXHASH_DIR "${CMAKE_CURRENT_SOURCE_DIR}/..")
+
+file(STRINGS "${XXHASH_DIR}/xxhash.h" XXHASH_VERSION_MAJOR REGEX "^#define XXH_VERSION_MAJOR +([0-9]+) *$")
+string(REGEX REPLACE "^#define XXH_VERSION_MAJOR +([0-9]+) *$" "\\1" XXHASH_VERSION_MAJOR "${XXHASH_VERSION_MAJOR}")
+file(STRINGS "${XXHASH_DIR}/xxhash.h" XXHASH_VERSION_MINOR REGEX "^#define XXH_VERSION_MINOR +([0-9]+) *$")
+string(REGEX REPLACE "^#define XXH_VERSION_MINOR +([0-9]+) *$" "\\1" XXHASH_VERSION_MINOR "${XXHASH_VERSION_MINOR}")
+file(STRINGS "${XXHASH_DIR}/xxhash.h" XXHASH_VERSION_RELEASE REGEX "^#define XXH_VERSION_RELEASE +([0-9]+) *$")
+string(REGEX REPLACE "^#define XXH_VERSION_RELEASE +([0-9]+) *$" "\\1" XXHASH_VERSION_RELEASE "${XXHASH_VERSION_RELEASE}")
+set(XXHASH_VERSION_STRING "${XXHASH_VERSION_MAJOR}.${XXHASH_VERSION_MINOR}.${XXHASH_VERSION_RELEASE}")
+set(XXHASH_LIB_VERSION ${XXHASH_VERSION_STRING})
+set(XXHASH_LIB_SOVERSION "${XXHASH_VERSION_MAJOR}")
+mark_as_advanced(XXHASH_VERSION_MAJOR XXHASH_VERSION_MINOR XXHASH_VERSION_RELEASE XXHASH_VERSION_STRING XXHASH_LIB_VERSION XXHASH_LIB_SOVERSION)
+
+if("${CMAKE_VERSION}" VERSION_LESS "3.13")
+    #message(WARNING "CMake ${CMAKE_VERSION} has no CMP0077 policy: options will erase uncached/untyped normal vars!")
+else()
+    cmake_policy (SET CMP0077 NEW)
+endif()
+if("${CMAKE_VERSION}" VERSION_LESS "3.0")
+    project(xxHash C)
+else()
+    cmake_policy (SET CMP0048 NEW)
+    project(xxHash
+        VERSION ${XXHASH_VERSION_STRING}
+        LANGUAGES C)
+endif()
+
+if(NOT CMAKE_BUILD_TYPE AND NOT CMAKE_CONFIGURATION_TYPES)
+  set(CMAKE_BUILD_TYPE "Release" CACHE STRING "Project build type" FORCE)
+  set_property(CACHE CMAKE_BUILD_TYPE
+    PROPERTY STRINGS "Debug" "Release" "RelWithDebInfo" "MinSizeRel")
+endif()
+if(NOT CMAKE_CONFIGURATION_TYPES)
+  message(STATUS "xxHash build type: ${CMAKE_BUILD_TYPE}")
+endif()
+
+# Enable assert() statements in debug builds
+if("${CMAKE_BUILD_TYPE}" STREQUAL "Debug")
+    if("${CMAKE_VERSION}" VERSION_LESS "3.12")
+        # add_compile_definitions is not available for older cmake => do nothing
+    else()
+        add_compile_definitions(XXH_DEBUGLEVEL=1)
+    endif()
+endif()
+
+option(BUILD_SHARED_LIBS "Build shared library" ON)
+option(XXHASH_BUILD_XXHSUM "Build the xxhsum binary" ON)
+
+# If XXHASH is being bundled in another project, we don't want to
+# install anything.  However, we want to let people override this, so
+# we'll use the XXHASH_BUNDLED_MODE variable to let them do that; just
+# set it to OFF in your project before you add_subdirectory(xxhash/cmake_unofficial).
+if(NOT DEFINED XXHASH_BUNDLED_MODE)
+  if("${PROJECT_SOURCE_DIR}" STREQUAL "${CMAKE_SOURCE_DIR}")
+    set(XXHASH_BUNDLED_MODE OFF)
+  else()
+    set(XXHASH_BUNDLED_MODE ON)
+  endif()
+endif()
+set(XXHASH_BUNDLED_MODE ${XXHASH_BUNDLED_MODE} CACHE BOOL "" FORCE)
+mark_as_advanced(XXHASH_BUNDLED_MODE)
+
+# Allow people to choose whether to build shared or static libraries
+# via the BUILD_SHARED_LIBS option unless we are in bundled mode, in
+# which case we always use static libraries.
+include(CMakeDependentOption)
+CMAKE_DEPENDENT_OPTION(BUILD_SHARED_LIBS "Build shared libraries" ON "NOT XXHASH_BUNDLED_MODE" OFF)
+
+if("${CMAKE_VERSION}" VERSION_LESS "3.10")
+  # Can not enable DISPATCH mode since it fails to recognize architecture.
+else()
+  CMAKE_HOST_SYSTEM_INFORMATION(RESULT PLATFORM QUERY OS_PLATFORM)
+  message(STATUS "Architecture: ${PLATFORM}")
+endif()
+
+# libxxhash
+if((DEFINED DISPATCH) AND (DEFINED PLATFORM))
+  # Only support DISPATCH option on x86_64.
+  if("${PLATFORM}" STREQUAL "x86_64")
+    message(STATUS "Enable xxHash dispatch mode")
+    add_library(xxhash "${XXHASH_DIR}/xxh_x86dispatch.c"
+                       "${XXHASH_DIR}/xxhash.c"
+               )
+    set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -DXXHSUM_DISPATCH=1")
+  else()
+    add_library(xxhash "${XXHASH_DIR}/xxhash.c")
+  endif()
+else()
+  add_library(xxhash "${XXHASH_DIR}/xxhash.c")
+endif()
+add_library(${PROJECT_NAME}::xxhash ALIAS xxhash)
+
+target_include_directories(xxhash
+  PUBLIC
+    $<BUILD_INTERFACE:${XXHASH_DIR}>
+    $<INSTALL_INTERFACE:include/>)
+if (BUILD_SHARED_LIBS)
+  target_compile_definitions(xxhash PUBLIC XXH_EXPORT)
+endif ()
+set_target_properties(xxhash PROPERTIES
+  SOVERSION "${XXHASH_LIB_SOVERSION}"
+  VERSION "${XXHASH_VERSION_STRING}")
+
+if(XXHASH_BUILD_XXHSUM)
+  set(XXHSUM_DIR "${XXHASH_DIR}/cli")
+  # xxhsum
+  add_executable(xxhsum "${XXHSUM_DIR}/xxhsum.c"
+                        "${XXHSUM_DIR}/xsum_os_specific.c"
+                        "${XXHSUM_DIR}/xsum_output.c"
+                        "${XXHSUM_DIR}/xsum_sanity_check.c"
+                        "${XXHSUM_DIR}/xsum_bench.c"
+                )
+  add_executable(${PROJECT_NAME}::xxhsum ALIAS xxhsum)
+
+  target_link_libraries(xxhsum PRIVATE xxhash)
+  target_include_directories(xxhsum PRIVATE "${XXHASH_DIR}")
+endif(XXHASH_BUILD_XXHSUM)
+
+# Extra warning flags
+include (CheckCCompilerFlag)
+if (XXHASH_C_FLAGS)
+  set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${XXHASH_C_FLAGS}")
+endif()
+
+if(NOT XXHASH_BUNDLED_MODE)
+  include(GNUInstallDirs)
+
+  install(TARGETS xxhash
+    EXPORT xxHashTargets
+    RUNTIME DESTINATION "${CMAKE_INSTALL_BINDIR}"
+    LIBRARY DESTINATION "${CMAKE_INSTALL_LIBDIR}"
+    ARCHIVE DESTINATION "${CMAKE_INSTALL_LIBDIR}")
+  install(FILES "${XXHASH_DIR}/xxhash.h"
+    DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}")
+  install(FILES "${XXHASH_DIR}/xxh3.h"
+    DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}")
+  if(DISPATCH)
+    install(FILES "${XXHASH_DIR}/xxh_x86dispatch.h"
+      DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}")
+  endif()
+  if(XXHASH_BUILD_XXHSUM)
+    install(TARGETS xxhsum
+      EXPORT xxHashTargets
+      RUNTIME DESTINATION "${CMAKE_INSTALL_BINDIR}")
+    install(FILES "${XXHSUM_DIR}/xxhsum.1"
+      DESTINATION "${CMAKE_INSTALL_MANDIR}/man1")
+  endif(XXHASH_BUILD_XXHSUM)
+
+  include(CMakePackageConfigHelpers)
+
+  set(xxHash_VERSION_CONFIG "${PROJECT_BINARY_DIR}/xxHashConfigVersion.cmake")
+  set(xxHash_PROJECT_CONFIG "${PROJECT_BINARY_DIR}/xxHashConfig.cmake")
+  set(xxHash_TARGETS_CONFIG "${PROJECT_BINARY_DIR}/xxHashTargets.cmake")
+  set(xxHash_CONFIG_INSTALL_DIR "${CMAKE_INSTALL_LIBDIR}/cmake/xxHash/")
+  write_basic_package_version_file(${xxHash_VERSION_CONFIG}
+    VERSION ${XXHASH_VERSION_STRING}
+    COMPATIBILITY AnyNewerVersion)
+  configure_package_config_file(
+    ${PROJECT_SOURCE_DIR}/xxHashConfig.cmake.in
+    ${xxHash_PROJECT_CONFIG}
+    INSTALL_DESTINATION ${xxHash_CONFIG_INSTALL_DIR})
+  if("${CMAKE_VERSION}" VERSION_LESS "3.0")
+      set(XXHASH_EXPORT_SET xxhash)
+      if(XXHASH_BUILD_XXHSUM)
+        set(XXHASH_EXPORT_SET ${XXHASH_EXPORT_SET} xxhsum)
+      endif()
+      export(TARGETS ${XXHASH_EXPORT_SET}
+      FILE ${xxHash_TARGETS_CONFIG}
+      NAMESPACE ${PROJECT_NAME}::)
+  else()
+    export(EXPORT xxHashTargets
+      FILE ${xxHash_TARGETS_CONFIG}
+      NAMESPACE ${PROJECT_NAME}::)
+  endif()
+
+  install(FILES ${xxHash_PROJECT_CONFIG} ${xxHash_VERSION_CONFIG}
+    DESTINATION ${xxHash_CONFIG_INSTALL_DIR})
+  install(EXPORT xxHashTargets
+    DESTINATION ${xxHash_CONFIG_INSTALL_DIR}
+    NAMESPACE ${PROJECT_NAME}::)
+
+  # configure and install pkg-config
+  set(PREFIX ${CMAKE_INSTALL_PREFIX})
+  set(EXECPREFIX "\${prefix}")
+  set(INCLUDEDIR "${CMAKE_INSTALL_INCLUDEDIR}")
+  set(LIBDIR "${CMAKE_INSTALL_LIBDIR}")
+  set(VERSION "${XXHASH_VERSION_STRING}")
+  configure_file(${XXHASH_DIR}/libxxhash.pc.in ${CMAKE_BINARY_DIR}/libxxhash.pc @ONLY)
+
+  install(FILES ${CMAKE_BINARY_DIR}/libxxhash.pc
+    DESTINATION ${CMAKE_INSTALL_LIBDIR}/pkgconfig)
+
+endif(NOT XXHASH_BUNDLED_MODE)
+
+include(CPack)

third_party/xxhash/cmake_unofficial/README.md

@@ -0,0 +1,37 @@
+
+## Usage
+
+### Way 1: import targets
+Build xxHash targets:
+
+    cd </path/to/xxHash/>
+    mkdir build
+    cd build
+    cmake ../cmake_unofficial [options]
+    cmake --build .
+    cmake --build . --target install #optional
+
+Where possible options are:
+- `-DXXHASH_BUILD_ENABLE_INLINE_API=<ON|OFF>`: adds xxhash.c for the `-DXXH_INLINE_ALL` api. ON by default.
+- `-DXXHASH_BUILD_XXHSUM=<ON|OFF>`: build the command line binary. ON by default
+- `-DBUILD_SHARED_LIBS=<ON|OFF>`: build dynamic library. ON by default.
+- `-DCMAKE_INSTALL_PREFIX=<path>`: use custom install prefix path.
+- `-DDISPATCH=<ON|OFF>`: enable dispatch mode. OFF by default.
+
+Add lines into downstream CMakeLists.txt:
+
+    find_package(xxHash 0.7 CONFIG REQUIRED)
+    ...
+    target_link_libraries(MyTarget PRIVATE xxHash::xxhash)
+
+### Way 2: Add subdirectory
+Add lines into downstream CMakeLists.txt:
+
+    option(BUILD_SHARED_LIBS "Build shared libs" OFF) #optional
+    ...
+    set(XXHASH_BUILD_ENABLE_INLINE_API OFF) #optional
+    set(XXHASH_BUILD_XXHSUM OFF) #optional
+    add_subdirectory(</path/to/xxHash/cmake_unofficial/> </path/to/xxHash/build/> EXCLUDE_FROM_ALL)
+    ...
+    target_link_libraries(MyTarget PRIVATE xxHash::xxhash)
+

third_party/xxhash/cmake_unofficial/xxHashConfig.cmake.in

@@ -0,0 +1,4 @@
+@PACKAGE_INIT@
+
+include(${CMAKE_CURRENT_LIST_DIR}/xxHashTargets.cmake)
+

third_party/xxhash/doc/README.md

@@ -0,0 +1,9 @@
+xxHash Specification
+=======================
+
+This directory contains material defining the xxHash algorithm.
+It's described in [this specification document](xxhash_spec.md).
+
+The algorithm is also be illustrated by a [simple educational library](https://github.com/easyaspi314/xxhash-clean),
+written by @easyaspi314 and designed for readability
+(as opposed to the reference library which is designed for speed).

third_party/xxhash/doc/xxhash.cry

@@ -0,0 +1,206 @@
+module xxhash where
+
+/**
+ * The 32-bit variant of xxHash. The first argument is the sequence
+ * of L bytes to hash. The second argument is a seed value.
+ */
+XXH32 : {L} (fin L) => [L][8] -> [32] -> [32]
+XXH32 input seed = XXH32_avalanche acc1
+  where (stripes16 # stripes4 # stripes1) = input
+        accR = foldl XXH32_rounds (XXH32_init seed) (split stripes16 : [L/16][16][8])
+        accL = `(L % 2^^32) + if (`L:Integer) < 16
+                              then seed + PRIME32_5
+                              else XXH32_converge accR
+        acc4 = foldl XXH32_digest4 accL (split stripes4 : [(L%16)/4][4][8])
+        acc1 = foldl XXH32_digest1 acc4 (stripes1 : [L%4][8])
+
+/**
+ * The 64-bit variant of xxHash. The first argument is the sequence
+ * of L bytes to hash. The second argument is a seed value.
+ */
+XXH64 : {L} (fin L) => [L][8] -> [64] -> [64]
+XXH64 input seed = XXH64_avalanche acc1
+  where (stripes32 # stripes8 # stripes4 # stripes1) = input
+        accR = foldl XXH64_rounds (XXH64_init seed) (split stripes32 : [L/32][32][8])
+        accL = `(L % 2^^64) + if (`L:Integer) < 32
+                              then seed + PRIME64_5
+                              else XXH64_converge accR
+        acc8 = foldl XXH64_digest8 accL (split stripes8 : [(L%32)/8][8][8])
+        acc4 = foldl XXH64_digest4 acc8 (split stripes4 : [(L%8)/4][4][8])
+        acc1 = foldl XXH64_digest1 acc4 (stripes1 : [L%4][8])
+
+private
+
+  //Utility functions
+
+  /**
+   * Combines a sequence of bytes into a word using the little-endian
+   * convention.
+   */
+  toLE bytes = join (reverse bytes)
+
+  //32-bit xxHash helper functions
+
+  //32-bit prime number constants
+  PRIME32_1 = 0x9E3779B1 : [32]
+  PRIME32_2 = 0x85EBCA77 : [32]
+  PRIME32_3 = 0xC2B2AE3D : [32]
+  PRIME32_4 = 0x27D4EB2F : [32]
+  PRIME32_5 = 0x165667B1 : [32]
+
+  /**
+   * The property shows that the hexadecimal representation of the
+   * PRIME32 constants is the same as the binary representation.
+   */
+  property PRIME32s_as_bits_correct =
+    (PRIME32_1 == 0b10011110001101110111100110110001) /\
+    (PRIME32_2 == 0b10000101111010111100101001110111) /\
+    (PRIME32_3 == 0b11000010101100101010111000111101) /\
+    (PRIME32_4 == 0b00100111110101001110101100101111) /\
+    (PRIME32_5 == 0b00010110010101100110011110110001)
+
+  /**
+   * This function initializes the four internal accumulators of XXH32.
+   */
+  XXH32_init : [32] -> [4][32]
+  XXH32_init seed = [acc1, acc2, acc3, acc4]
+    where acc1 = seed + PRIME32_1 + PRIME32_2
+          acc2 = seed + PRIME32_2
+          acc3 = seed + 0
+          acc4 = seed - PRIME32_1
+
+  /**
+   * This processes a single lane of the main round function of XXH32.
+   */
+  XXH32_round : [32] -> [32] -> [32]
+  XXH32_round accN laneN = ((accN + laneN * PRIME32_2) <<< 13) * PRIME32_1
+
+  /**
+   * This is the main round function of XXH32 and processes a stripe,
+   * i.e. 4 lanes with 4 bytes each.
+   */
+  XXH32_rounds : [4][32] -> [16][8] -> [4][32]
+  XXH32_rounds accs stripe =
+    [ XXH32_round accN (toLE laneN) | accN <- accs | laneN <- split stripe ]
+
+  /**
+   * This function combines the four lane accumulators into a single
+   * 32-bit value.
+   */
+  XXH32_converge : [4][32] -> [32]
+  XXH32_converge [acc1, acc2, acc3, acc4] =
+    (acc1 <<< 1) + (acc2 <<< 7) + (acc3 <<< 12) + (acc4 <<< 18)
+
+  /**
+   * This function digests a four byte lane
+   */
+  XXH32_digest4 : [32] -> [4][8] -> [32]
+  XXH32_digest4 acc lane = ((acc + toLE lane * PRIME32_3) <<< 17) * PRIME32_4
+
+  /**
+   * This function digests a single byte lane
+   */
+  XXH32_digest1 : [32] -> [8] -> [32]
+  XXH32_digest1 acc lane = ((acc + (0 # lane) * PRIME32_5) <<< 11) * PRIME32_1
+
+  /**
+   * This function ensures that all input bits have a chance to impact
+   * any bit in the output digest, resulting in an unbiased
+   * distribution.
+   */
+  XXH32_avalanche : [32] -> [32]
+  XXH32_avalanche acc0 = acc5
+    where acc1 = acc0 ^ (acc0 >> 15)
+          acc2 = acc1 * PRIME32_2
+          acc3 = acc2 ^ (acc2 >> 13)
+          acc4 = acc3 * PRIME32_3
+          acc5 = acc4 ^ (acc4 >> 16)
+
+  //64-bit xxHash helper functions
+
+  //64-bit prime number constants
+  PRIME64_1 = 0x9E3779B185EBCA87 : [64]
+  PRIME64_2 = 0xC2B2AE3D27D4EB4F : [64]
+  PRIME64_3 = 0x165667B19E3779F9 : [64]
+  PRIME64_4 = 0x85EBCA77C2B2AE63 : [64]
+  PRIME64_5 = 0x27D4EB2F165667C5 : [64]
+
+  /**
+   * The property shows that the hexadecimal representation of the
+   * PRIME64 constants is the same as the binary representation.
+   */
+  property PRIME64s_as_bits_correct =
+    (PRIME64_1 == 0b1001111000110111011110011011000110000101111010111100101010000111) /\
+    (PRIME64_2 == 0b1100001010110010101011100011110100100111110101001110101101001111) /\
+    (PRIME64_3 == 0b0001011001010110011001111011000110011110001101110111100111111001) /\
+    (PRIME64_4 == 0b1000010111101011110010100111011111000010101100101010111001100011) /\
+    (PRIME64_5 == 0b0010011111010100111010110010111100010110010101100110011111000101)
+
+  /**
+   * This function initializes the four internal accumulators of XXH64.
+   */
+  XXH64_init : [64] -> [4][64]
+  XXH64_init seed = [acc1, acc2, acc3, acc4]
+    where acc1 = seed + PRIME64_1 + PRIME64_2
+          acc2 = seed + PRIME64_2
+          acc3 = seed + 0
+          acc4 = seed - PRIME64_1
+
+  /**
+   * This processes a single lane of the main round function of XXH64.
+   */
+  XXH64_round : [64] -> [64] -> [64]
+  XXH64_round accN laneN = ((accN + laneN * PRIME64_2) <<< 31) * PRIME64_1
+
+  /**
+   * This is the main round function of XXH64 and processes a stripe,
+   * i.e. 4 lanes with 8 bytes each.
+   */
+  XXH64_rounds : [4][64] -> [32][8] -> [4][64]
+  XXH64_rounds accs stripe =
+    [ XXH64_round accN (toLE laneN) | accN <- accs | laneN <- split stripe ]
+
+  /**
+   * This is a helper function, used to merge the four lane accumulators.
+   */
+  mergeAccumulator : [64] -> [64] -> [64]
+  mergeAccumulator acc accN = (acc ^ XXH64_round 0 accN) * PRIME64_1 + PRIME64_4
+
+  /**
+   * This function combines the four lane accumulators into a single
+   * 64-bit value.
+   */
+  XXH64_converge : [4][64] -> [64]
+  XXH64_converge [acc1, acc2, acc3, acc4] =
+    foldl mergeAccumulator ((acc1 <<< 1) + (acc2 <<< 7) + (acc3 <<< 12) + (acc4 <<< 18)) [acc1, acc2, acc3, acc4]
+
+  /**
+   * This function digests an eight byte lane
+   */
+  XXH64_digest8 : [64] -> [8][8] -> [64]
+  XXH64_digest8 acc lane = ((acc ^ XXH64_round 0 (toLE lane)) <<< 27) * PRIME64_1 + PRIME64_4
+
+  /**
+   * This function digests a four byte lane
+   */
+  XXH64_digest4 : [64] -> [4][8] -> [64]
+  XXH64_digest4 acc lane = ((acc ^ (0 # toLE lane) * PRIME64_1) <<< 23) * PRIME64_2 + PRIME64_3
+
+  /**
+   * This function digests a single byte lane
+   */
+  XXH64_digest1 : [64] -> [8] -> [64]
+  XXH64_digest1 acc lane = ((acc ^ (0 # lane) * PRIME64_5) <<< 11) * PRIME64_1
+
+  /**
+   * This function ensures that all input bits have a chance to impact
+   * any bit in the output digest, resulting in an unbiased
+   * distribution.
+   */
+  XXH64_avalanche : [64] -> [64]
+  XXH64_avalanche acc0 = acc5
+    where acc1 = acc0 ^ (acc0 >> 33)
+          acc2 = acc1 * PRIME64_2
+          acc3 = acc2 ^ (acc2 >> 29)
+          acc4 = acc3 * PRIME64_3
+          acc5 = acc4 ^ (acc4 >> 32)

third_party/xxhash/doc/xxhash_spec.md

@@ -0,0 +1,351 @@
+xxHash fast digest algorithm
+======================
+
+### Notices
+
+Copyright (c) Yann Collet
+
+Permission is granted to copy and distribute this document
+for any purpose and without charge,
+including translations into other languages
+and incorporation into compilations,
+provided that the copyright notice and this notice are preserved,
+and that any substantive changes or deletions from the original
+are clearly marked.
+Distribution of this document is unlimited.
+
+### Version
+
+0.1.1 (10/10/18)
+
+
+Table of Contents
+---------------------
+- [Introduction](#introduction)
+- [XXH32 algorithm description](#xxh32-algorithm-description)
+- [XXH64 algorithm description](#xxh64-algorithm-description)
+- [Performance considerations](#performance-considerations)
+- [Reference Implementation](#reference-implementation)
+
+
+Introduction
+----------------
+
+This document describes the xxHash digest algorithm for both 32-bit and 64-bit variants, named `XXH32` and `XXH64`. The algorithm takes an input a message of arbitrary length and an optional seed value, then produces an output of 32 or 64-bit as "fingerprint" or "digest".
+
+xxHash is primarily designed for speed. It is labeled non-cryptographic, and is not meant to avoid intentional collisions (same digest for 2 different messages), or to prevent producing a message with a predefined digest.
+
+XXH32 is designed to be fast on 32-bit machines.
+XXH64 is designed to be fast on 64-bit machines.
+Both variants produce different output.
+However, a given variant shall produce exactly the same output, irrespective of the cpu / os used. In particular, the result remains identical whatever the endianness and width of the cpu is.
+
+### Operation notations
+
+All operations are performed modulo {32,64} bits. Arithmetic overflows are expected.
+`XXH32` uses 32-bit modular operations. `XXH64` uses 64-bit modular operations.
+
+- `+`: denotes modular addition
+- `*`: denotes modular multiplication
+- `X <<< s`: denotes the value obtained by circularly shifting (rotating) `X` left by `s` bit positions.
+- `X >> s`: denotes the value obtained by shifting `X` right by s bit positions. Upper `s` bits become `0`.
+- `X xor Y`: denotes the bit-wise XOR of `X` and `Y` (same width).
+
+
+XXH32 Algorithm Description
+-------------------------------------
+
+### Overview
+
+We begin by supposing that we have a message of any length `L` as input, and that we wish to find its digest. Here `L` is an arbitrary nonnegative integer; `L` may be zero. The following steps are performed to compute the digest of the message.
+
+The algorithm collect and transform input in _stripes_ of 16 bytes. The transforms are stored inside 4 "accumulators", each one storing an unsigned 32-bit value. Each accumulator can be processed independently in parallel, speeding up processing for cpu with multiple execution units.
+
+The algorithm uses 32-bits addition, multiplication, rotate, shift and xor operations. Many operations require some 32-bits prime number constants, all defined below:
+
+```c
+  static const u32 PRIME32_1 = 0x9E3779B1U;  // 0b10011110001101110111100110110001
+  static const u32 PRIME32_2 = 0x85EBCA77U;  // 0b10000101111010111100101001110111
+  static const u32 PRIME32_3 = 0xC2B2AE3DU;  // 0b11000010101100101010111000111101
+  static const u32 PRIME32_4 = 0x27D4EB2FU;  // 0b00100111110101001110101100101111
+  static const u32 PRIME32_5 = 0x165667B1U;  // 0b00010110010101100110011110110001
+```
+
+These constants are prime numbers, and feature a good mix of bits 1 and 0, neither too regular, nor too dissymmetric. These properties help dispersion capabilities.
+
+### Step 1. Initialize internal accumulators
+
+Each accumulator gets an initial value based on optional `seed` input. Since the `seed` is optional, it can be `0`.
+
+```c
+  u32 acc1 = seed + PRIME32_1 + PRIME32_2;
+  u32 acc2 = seed + PRIME32_2;
+  u32 acc3 = seed + 0;
+  u32 acc4 = seed - PRIME32_1;
+```
+
+#### Special case: input is less than 16 bytes
+
+When the input is too small (< 16 bytes), the algorithm will not process any stripes. Consequently, it will not make use of parallel accumulators.
+
+In this case, a simplified initialization is performed, using a single accumulator:
+
+```c
+  u32 acc  = seed + PRIME32_5;
+```
+
+The algorithm then proceeds directly to step 4.
+
+### Step 2. Process stripes
+
+A stripe is a contiguous segment of 16 bytes.
+It is evenly divided into 4 _lanes_, of 4 bytes each.
+The first lane is used to update accumulator 1, the second lane is used to update accumulator 2, and so on.
+
+Each lane read its associated 32-bit value using __little-endian__ convention.
+
+For each {lane, accumulator}, the update process is called a _round_, and applies the following formula:
+
+```c
+  accN = accN + (laneN * PRIME32_2);
+  accN = accN <<< 13;
+  accN = accN * PRIME32_1;
+```
+
+This shuffles the bits so that any bit from input _lane_ impacts several bits in output _accumulator_. All operations are performed modulo 2^32.
+
+Input is consumed one full stripe at a time. Step 2 is looped as many times as necessary to consume the whole input, except for the last remaining bytes which cannot form a stripe (< 16 bytes).
+When that happens, move to step 3.
+
+### Step 3. Accumulator convergence
+
+All 4 lane accumulators from the previous steps are merged to produce a single remaining accumulator of the same width (32-bit). The associated formula is as follows:
+
+```c
+  acc = (acc1 <<< 1) + (acc2 <<< 7) + (acc3 <<< 12) + (acc4 <<< 18);
+```
+
+### Step 4. Add input length
+
+The input total length is presumed known at this stage. This step is just about adding the length to accumulator, so that it participates to final mixing.
+
+```c
+  acc = acc + (u32)inputLength;
+```
+
+Note that, if input length is so large that it requires more than 32-bits, only the lower 32-bits are added to the accumulator.
+
+### Step 5. Consume remaining input
+
+There may be up to 15 bytes remaining to consume from the input.
+The final stage will digest them according to following pseudo-code:
+
+```c
+  while (remainingLength >= 4) {
+      lane = read_32bit_little_endian(input_ptr);
+      acc = acc + lane * PRIME32_3;
+      acc = (acc <<< 17) * PRIME32_4;
+      input_ptr += 4; remainingLength -= 4;
+  }
+
+  while (remainingLength >= 1) {
+      lane = read_byte(input_ptr);
+      acc = acc + lane * PRIME32_5;
+      acc = (acc <<< 11) * PRIME32_1;
+      input_ptr += 1; remainingLength -= 1;
+  }
+```
+
+This process ensures that all input bytes are present in the final mix.
+
+### Step 6. Final mix (avalanche)
+
+The final mix ensures that all input bits have a chance to impact any bit in the output digest, resulting in an unbiased distribution. This is also called avalanche effect.
+
+```c
+  acc = acc xor (acc >> 15);
+  acc = acc * PRIME32_2;
+  acc = acc xor (acc >> 13);
+  acc = acc * PRIME32_3;
+  acc = acc xor (acc >> 16);
+```
+
+### Step 7. Output
+
+The `XXH32()` function produces an unsigned 32-bit value as output.
+
+For systems which require to store and/or display the result in binary or hexadecimal format, the canonical format is defined to reproduce the same value as the natural decimal format, hence follows __big-endian__ convention (most significant byte first).
+
+
+XXH64 Algorithm Description
+-------------------------------------
+
+### Overview
+
+`XXH64`'s algorithm structure is very similar to `XXH32` one. The major difference is that `XXH64` uses 64-bit arithmetic, speeding up memory transfer for 64-bit compliant systems, but also relying on cpu capability to efficiently perform 64-bit operations.
+
+The algorithm collects and transforms input in _stripes_ of 32 bytes. The transforms are stored inside 4 "accumulators", each one storing an unsigned 64-bit value. Each accumulator can be processed independently in parallel, speeding up processing for cpu with multiple execution units.
+
+The algorithm uses 64-bit addition, multiplication, rotate, shift and xor operations. Many operations require some 64-bit prime number constants, all defined below:
+
+```c
+  static const u64 PRIME64_1 = 0x9E3779B185EBCA87ULL;  // 0b1001111000110111011110011011000110000101111010111100101010000111
+  static const u64 PRIME64_2 = 0xC2B2AE3D27D4EB4FULL;  // 0b1100001010110010101011100011110100100111110101001110101101001111
+  static const u64 PRIME64_3 = 0x165667B19E3779F9ULL;  // 0b0001011001010110011001111011000110011110001101110111100111111001
+  static const u64 PRIME64_4 = 0x85EBCA77C2B2AE63ULL;  // 0b1000010111101011110010100111011111000010101100101010111001100011
+  static const u64 PRIME64_5 = 0x27D4EB2F165667C5ULL;  // 0b0010011111010100111010110010111100010110010101100110011111000101
+```
+
+These constants are prime numbers, and feature a good mix of bits 1 and 0, neither too regular, nor too dissymmetric. These properties help dispersion capabilities.
+
+### Step 1. Initialize internal accumulators
+
+Each accumulator gets an initial value based on optional `seed` input. Since the `seed` is optional, it can be `0`.
+
+```c
+  u64 acc1 = seed + PRIME64_1 + PRIME64_2;
+  u64 acc2 = seed + PRIME64_2;
+  u64 acc3 = seed + 0;
+  u64 acc4 = seed - PRIME64_1;
+```
+
+#### Special case: input is less than 32 bytes
+
+When the input is too small (< 32 bytes), the algorithm will not process any stripes. Consequently, it will not make use of parallel accumulators.
+
+In this case, a simplified initialization is performed, using a single accumulator:
+
+```c
+  u64 acc  = seed + PRIME64_5;
+```
+
+The algorithm then proceeds directly to step 4.
+
+### Step 2. Process stripes
+
+A stripe is a contiguous segment of 32 bytes.
+It is evenly divided into 4 _lanes_, of 8 bytes each.
+The first lane is used to update accumulator 1, the second lane is used to update accumulator 2, and so on.
+
+Each lane read its associated 64-bit value using __little-endian__ convention.
+
+For each {lane, accumulator}, the update process is called a _round_, and applies the following formula:
+
+```c
+  round(accN,laneN):
+  accN = accN + (laneN * PRIME64_2);
+  accN = accN <<< 31;
+  return accN * PRIME64_1;
+```
+
+This shuffles the bits so that any bit from input _lane_ impacts several bits in output _accumulator_. All operations are performed modulo 2^64.
+
+Input is consumed one full stripe at a time. Step 2 is looped as many times as necessary to consume the whole input, except for the last remaining bytes which cannot form a stripe (< 32 bytes).
+When that happens, move to step 3.
+
+### Step 3. Accumulator convergence
+
+All 4 lane accumulators from previous steps are merged to produce a single remaining accumulator of same width (64-bit). The associated formula is as follows.
+
+Note that accumulator convergence is more complex than 32-bit variant, and requires to define another function called _mergeAccumulator()_:
+
+```c
+  mergeAccumulator(acc,accN):
+  acc  = acc xor round(0, accN);
+  acc  = acc * PRIME64_1;
+  return acc + PRIME64_4;
+```
+
+which is then used in the convergence formula:
+
+```c
+  acc = (acc1 <<< 1) + (acc2 <<< 7) + (acc3 <<< 12) + (acc4 <<< 18);
+  acc = mergeAccumulator(acc, acc1);
+  acc = mergeAccumulator(acc, acc2);
+  acc = mergeAccumulator(acc, acc3);
+  acc = mergeAccumulator(acc, acc4);
+```
+
+### Step 4. Add input length
+
+The input total length is presumed known at this stage. This step is just about adding the length to accumulator, so that it participates to final mixing.
+
+```c
+  acc = acc + inputLength;
+```
+
+### Step 5. Consume remaining input
+
+There may be up to 31 bytes remaining to consume from the input.
+The final stage will digest them according to following pseudo-code:
+
+```c
+  while (remainingLength >= 8) {
+      lane = read_64bit_little_endian(input_ptr);
+      acc = acc xor round(0, lane);
+      acc = (acc <<< 27) * PRIME64_1;
+      acc = acc + PRIME64_4;
+      input_ptr += 8; remainingLength -= 8;
+  }
+
+  if (remainingLength >= 4) {
+      lane = read_32bit_little_endian(input_ptr);
+      acc = acc xor (lane * PRIME64_1);
+      acc = (acc <<< 23) * PRIME64_2;
+      acc = acc + PRIME64_3;
+      input_ptr += 4; remainingLength -= 4;
+  }
+
+  while (remainingLength >= 1) {
+      lane = read_byte(input_ptr);
+      acc = acc xor (lane * PRIME64_5);
+      acc = (acc <<< 11) * PRIME64_1;
+      input_ptr += 1; remainingLength -= 1;
+  }
+```
+
+This process ensures that all input bytes are present in the final mix.
+
+### Step 6. Final mix (avalanche)
+
+The final mix ensures that all input bits have a chance to impact any bit in the output digest, resulting in an unbiased distribution. This is also called avalanche effect.
+
+```c
+  acc = acc xor (acc >> 33);
+  acc = acc * PRIME64_2;
+  acc = acc xor (acc >> 29);
+  acc = acc * PRIME64_3;
+  acc = acc xor (acc >> 32);
+```
+
+### Step 7. Output
+
+The `XXH64()` function produces an unsigned 64-bit value as output.
+
+For systems which require to store and/or display the result in binary or hexadecimal format, the canonical format is defined to reproduce the same value as the natural decimal format, hence follows __big-endian__ convention (most significant byte first).
+
+Performance considerations
+----------------------------------
+
+The xxHash algorithms are simple and compact to implement. They provide a system independent "fingerprint" or digest of a message of arbitrary length.
+
+The algorithm allows input to be streamed and processed in multiple steps. In such case, an internal buffer is needed to ensure data is presented to the algorithm in full stripes.
+
+On 64-bit systems, the 64-bit variant `XXH64` is generally faster to compute, so it is a recommended variant, even when only 32-bit are needed.
+
+On 32-bit systems though, positions are reversed: `XXH64` performance is reduced, due to its usage of 64-bit arithmetic. `XXH32` becomes a faster variant.
+
+
+Reference Implementation
+----------------------------------------
+
+A reference library written in C is available at https://www.xxhash.com.
+The web page also links to multiple other implementations written in many different languages.
+It links to the [github project page](https://github.com/Cyan4973/xxHash) where an [issue board](https://github.com/Cyan4973/xxHash/issues) can be used for further public discussions on the topic.
+
+
+Version changes
+--------------------
+v0.7.3: Minor fixes
+v0.1.1: added a note on rationale for selection of constants
+v0.1.0: initial release

third_party/xxhash/include/xxhash/xxh3.h

@@ -0,0 +1,55 @@
+/*
+ * xxHash - Extremely Fast Hash algorithm
+ * Development source file for `xxh3`
+ * Copyright (C) 2019-2021 Yann Collet
+ *
+ * BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php)
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are
+ * met:
+ *
+ *    * Redistributions of source code must retain the above copyright
+ *      notice, this list of conditions and the following disclaimer.
+ *    * Redistributions in binary form must reproduce the above
+ *      copyright notice, this list of conditions and the following disclaimer
+ *      in the documentation and/or other materials provided with the
+ *      distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * You can contact the author at:
+ *   - xxHash homepage: https://www.xxhash.com
+ *   - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+/*
+ * Note: This file used to host the source code of XXH3_* variants.
+ * during the development period.
+ * The source code is now properly integrated within xxhash.h.
+ *
+ * xxh3.h is no longer useful,
+ * but it is still provided for compatibility with source code
+ * which used to include it directly.
+ *
+ * Programs are now highly discouraged to include xxh3.h.
+ * Include `xxhash.h` instead, which is the officially supported interface.
+ *
+ * In the future, xxh3.h will start to generate warnings, then errors,
+ * then it will be removed from source package and from include directory.
+ */
+
+/* Simulate the same impact as including the old xxh3.h source file */
+
+#define XXH_INLINE_ALL
+#include "xxhash/xxhash.h"

third_party/xxhash/include/xxhash/xxh_x86dispatch.h

@@ -0,0 +1,85 @@
+/*
+ * xxHash - XXH3 Dispatcher for x86-based targets
+ * Copyright (C) 2020-2021 Yann Collet
+ *
+ * BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php)
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are
+ * met:
+ *
+ *    * Redistributions of source code must retain the above copyright
+ *      notice, this list of conditions and the following disclaimer.
+ *    * Redistributions in binary form must reproduce the above
+ *      copyright notice, this list of conditions and the following disclaimer
+ *      in the documentation and/or other materials provided with the
+ *      distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * You can contact the author at:
+ *   - xxHash homepage: https://www.xxhash.com
+ *   - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+#ifndef XXH_X86DISPATCH_H_13563687684
+#define XXH_X86DISPATCH_H_13563687684
+
+#include "xxhash/xxhash.h"  /* XXH64_hash_t, XXH3_state_t */
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+XXH_PUBLIC_API XXH64_hash_t  XXH3_64bits_dispatch(XXH_NOESCAPE const void* input, size_t len);
+XXH_PUBLIC_API XXH64_hash_t  XXH3_64bits_withSeed_dispatch(XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed);
+XXH_PUBLIC_API XXH64_hash_t  XXH3_64bits_withSecret_dispatch(XXH_NOESCAPE const void* input, size_t len, XXH_NOESCAPE const void* secret, size_t secretLen);
+XXH_PUBLIC_API XXH_errorcode XXH3_64bits_update_dispatch(XXH_NOESCAPE XXH3_state_t* state, XXH_NOESCAPE const void* input, size_t len);
+
+XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_dispatch(XXH_NOESCAPE const void* input, size_t len);
+XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_withSeed_dispatch(XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed);
+XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_withSecret_dispatch(XXH_NOESCAPE const void* input, size_t len, XXH_NOESCAPE const void* secret, size_t secretLen);
+XXH_PUBLIC_API XXH_errorcode XXH3_128bits_update_dispatch(XXH_NOESCAPE XXH3_state_t* state, XXH_NOESCAPE const void* input, size_t len);
+
+#if defined (__cplusplus)
+}
+#endif
+
+
+/* automatic replacement of XXH3 functions.
+ * can be disabled by setting XXH_DISPATCH_DISABLE_REPLACE */
+#ifndef XXH_DISPATCH_DISABLE_REPLACE
+
+# undef  XXH3_64bits
+# define XXH3_64bits XXH3_64bits_dispatch
+# undef  XXH3_64bits_withSeed
+# define XXH3_64bits_withSeed XXH3_64bits_withSeed_dispatch
+# undef  XXH3_64bits_withSecret
+# define XXH3_64bits_withSecret XXH3_64bits_withSecret_dispatch
+# undef  XXH3_64bits_update
+# define XXH3_64bits_update XXH3_64bits_update_dispatch
+
+# undef  XXH128
+# define XXH128 XXH3_128bits_withSeed_dispatch
+# undef  XXH3_128bits
+# define XXH3_128bits XXH3_128bits_dispatch
+# undef  XXH3_128bits_withSeed
+# define XXH3_128bits_withSeed XXH3_128bits_withSeed_dispatch
+# undef  XXH3_128bits_withSecret
+# define XXH3_128bits_withSecret XXH3_128bits_withSecret_dispatch
+# undef  XXH3_128bits_update
+# define XXH3_128bits_update XXH3_128bits_update_dispatch
+
+#endif /* XXH_DISPATCH_DISABLE_REPLACE */
+
+#endif /* XXH_X86DISPATCH_H_13563687684 */

third_party/xxhash/libxxhash.pc.in

@@ -0,0 +1,15 @@
+#   xxHash - Extremely fast hash algorithm
+#   Copyright (C) 2012-2021, Yann Collet, Facebook
+#   BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php)
+
+prefix=@PREFIX@
+exec_prefix=@EXECPREFIX@
+includedir=${prefix}/@INCLUDEDIR@
+libdir=${exec_prefix}/@LIBDIR@
+
+Name: xxhash
+Description: extremely fast hash algorithm
+URL: http://www.xxhash.com/
+Version: @VERSION@
+Libs: -L${libdir} -lxxhash
+Cflags: -I${includedir}

third_party/xxhash/tests/Makefile

@@ -0,0 +1,127 @@
+# ################################################################
+# xxHash Makefile
+# Copyright (C) 2012-2021 Yann Collet
+#
+# GPL v2 License
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License along
+# with this program; if not, write to the Free Software Foundation, Inc.,
+# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+#
+# You can contact the author at:
+#   - xxHash homepage: https://www.xxhash.com
+#   - xxHash source repository: https://github.com/Cyan4973/xxHash
+# ################################################################
+
+CFLAGS += -Wall -Wextra -Wundef -g
+
+CP = cp
+NM = nm
+GREP = grep
+XXHSUM_DIR = ..
+XXHSUM = $(XXHSUM_DIR)/xxhsum
+
+# Define *.exe as extension for Windows systems
+ifneq (,$(filter Windows%,$(OS)))
+EXT =.exe
+else
+EXT =
+endif
+
+ifneq (,$(filter %UTF-8,$(LANG)))
+ENABLE_UNICODE ?= 1
+else
+ENABLE_UNICODE ?= 0
+endif
+
+.PHONY: default
+default: all
+
+.PHONY: all
+all: test
+
+.PHONY: test
+test: test_multiInclude test_unicode test_sanity
+
+.PHONY: test_multiInclude
+test_multiInclude:
+	@$(MAKE) clean
+	# compile without xxhash.o, ensure symbols exist within target
+	# Note: built using only default rules
+	$(MAKE) multiInclude
+	@$(MAKE) clean
+	# compile with xxhash.o, to detect duplicated symbols
+	$(MAKE) multiInclude_withxxhash
+	@$(MAKE) clean
+	# compile with XXH_NAMESPACE before XXH_INLINE_ALL
+	CPPFLAGS=-DXXH_NAMESPACE=TESTN_ $(MAKE) multiInclude
+	# no symbol prefixed TESTN_ should exist
+	! $(NM) multiInclude | $(GREP) TESTN_
+	$(MAKE) clean
+	# compile xxhash.o with XXH_NAMESPACE
+	CPPFLAGS=-DXXH_NAMESPACE=TESTN_ $(MAKE) multiInclude_withxxhash
+	# symbols prefixed TESTN_ should exist in xxhash.o (though not be invoked)
+	$(NM) multiInclude_withxxhash | $(GREP) TESTN_
+	$(MAKE) clean
+
+.PHONY: test_ppc_redefine
+test_ppc_redefine: ppc_define.c
+	@$(MAKE) clean
+	$(CC) $(CPPFLAGS) $(CFLAGS) -c $^
+
+.PHONY: $(XXHSUM)
+$(XXHSUM):
+	$(MAKE) -C $(XXHSUM_DIR) xxhsum
+	$(CP) $(XXHSUM) .
+
+# Make sure that Unicode filenames work.
+# https://github.com/Cyan4973/xxHash/issues/293
+.PHONY: test_unicode
+ifeq (0,$(ENABLE_UNICODE))
+test_unicode:
+	@echo "Skipping Unicode test, your terminal doesn't appear to support UTF-8."
+	@echo "Try with ENABLE_UNICODE=1"
+else
+test_unicode: $(XXHSUM) generate_unicode_test.c
+	# Generate a Unicode filename test dynamically
+	# to keep UTF-8 out of the source tree.
+	$(CC) $(CFLAGS) $(LDFLAGS) generate_unicode_test.c -o generate_unicode_test$(EXT)
+	./generate_unicode_test$(EXT)
+	$(SHELL) ./unicode_test.sh
+endif
+
+.PHONY: test_filename_escape
+test_filename_escape: $(XXHSUM)
+	./filename-escape.sh
+
+.PHONY: test_sanity
+test_sanity: sanity_test.c
+	$(CC) $(CFLAGS) $(LDFLAGS) sanity_test.c -o sanity_test$(EXT)
+	$(RUN_ENV) ./sanity_test$(EXT)
+
+.PHONY: sanity_test_vectors.h
+sanity_test_vectors.h: sanity_test_vectors_generator.c
+	$(CC) $(CFLAGS) $(LDFLAGS) sanity_test_vectors_generator.c -o sanity_test_vectors_generator$(EXT)
+	./sanity_test_vectors_generator$(EXT)
+
+xxhash.o: ../xxhash.c ../xxhash.h
+	$(CC) $(CFLAGS) $(CPPFLAGS) $(LDFLAGS) -c -o $@ $<
+
+multiInclude_withxxhash: multiInclude.o xxhash.o
+	$(CC) $(CFLAGS) $(CPPFLAGS) $(LDFLAGS) -o $@ $^
+
+clean:
+	@$(RM) *.o
+	@$(RM) multiInclude multiInclude_withxxhash
+	@$(RM) *.unicode generate_unicode_test$(EXT) unicode_test.* xxhsum*
+	@$(RM) sanity_test$(EXT) sanity_test_vectors_generator$(EXT)

third_party/xxhash/tests/bench/.gitignore

@@ -0,0 +1,11 @@
+# build artifacts
+
+*.o
+benchHash
+benchHash32
+benchHash_avx2
+benchHash_hw
+
+# test files
+
+test*

third_party/xxhash/tests/bench/LICENSE

@@ -0,0 +1,339 @@
+                    GNU GENERAL PUBLIC LICENSE
+                       Version 2, June 1991
+
+ Copyright (C) 1989, 1991 Free Software Foundation, Inc.,
+ 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+                            Preamble
+
+  The licenses for most software are designed to take away your
+freedom to share and change it.  By contrast, the GNU General Public
+License is intended to guarantee your freedom to share and change free
+software--to make sure the software is free for all its users.  This
+General Public License applies to most of the Free Software
+Foundation's software and to any other program whose authors commit to
+using it.  (Some other Free Software Foundation software is covered by
+the GNU Lesser General Public License instead.)  You can apply it to
+your programs, too.
+
+  When we speak of free software, we are referring to freedom, not
+price.  Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+this service if you wish), that you receive source code or can get it
+if you want it, that you can change the software or use pieces of it
+in new free programs; and that you know you can do these things.
+
+  To protect your rights, we need to make restrictions that forbid
+anyone to deny you these rights or to ask you to surrender the rights.
+These restrictions translate to certain responsibilities for you if you
+distribute copies of the software, or if you modify it.
+
+  For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must give the recipients all the rights that
+you have.  You must make sure that they, too, receive or can get the
+source code.  And you must show them these terms so they know their
+rights.
+
+  We protect your rights with two steps: (1) copyright the software, and
+(2) offer you this license which gives you legal permission to copy,
+distribute and/or modify the software.
+
+  Also, for each author's protection and ours, we want to make certain
+that everyone understands that there is no warranty for this free
+software.  If the software is modified by someone else and passed on, we
+want its recipients to know that what they have is not the original, so
+that any problems introduced by others will not reflect on the original
+authors' reputations.
+
+  Finally, any free program is threatened constantly by software
+patents.  We wish to avoid the danger that redistributors of a free
+program will individually obtain patent licenses, in effect making the
+program proprietary.  To prevent this, we have made it clear that any
+patent must be licensed for everyone's free use or not licensed at all.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+
+                    GNU GENERAL PUBLIC LICENSE
+   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
+
+  0. This License applies to any program or other work which contains
+a notice placed by the copyright holder saying it may be distributed
+under the terms of this General Public License.  The "Program", below,
+refers to any such program or work, and a "work based on the Program"
+means either the Program or any derivative work under copyright law:
+that is to say, a work containing the Program or a portion of it,
+either verbatim or with modifications and/or translated into another
+language.  (Hereinafter, translation is included without limitation in
+the term "modification".)  Each licensee is addressed as "you".
+
+Activities other than copying, distribution and modification are not
+covered by this License; they are outside its scope.  The act of
+running the Program is not restricted, and the output from the Program
+is covered only if its contents constitute a work based on the
+Program (independent of having been made by running the Program).
+Whether that is true depends on what the Program does.
+
+  1. You may copy and distribute verbatim copies of the Program's
+source code as you receive it, in any medium, provided that you
+conspicuously and appropriately publish on each copy an appropriate
+copyright notice and disclaimer of warranty; keep intact all the
+notices that refer to this License and to the absence of any warranty;
+and give any other recipients of the Program a copy of this License
+along with the Program.
+
+You may charge a fee for the physical act of transferring a copy, and
+you may at your option offer warranty protection in exchange for a fee.
+
+  2. You may modify your copy or copies of the Program or any portion
+of it, thus forming a work based on the Program, and copy and
+distribute such modifications or work under the terms of Section 1
+above, provided that you also meet all of these conditions:
+
+    a) You must cause the modified files to carry prominent notices
+    stating that you changed the files and the date of any change.
+
+    b) You must cause any work that you distribute or publish, that in
+    whole or in part contains or is derived from the Program or any
+    part thereof, to be licensed as a whole at no charge to all third
+    parties under the terms of this License.
+
+    c) If the modified program normally reads commands interactively
+    when run, you must cause it, when started running for such
+    interactive use in the most ordinary way, to print or display an
+    announcement including an appropriate copyright notice and a
+    notice that there is no warranty (or else, saying that you provide
+    a warranty) and that users may redistribute the program under
+    these conditions, and telling the user how to view a copy of this
+    License.  (Exception: if the Program itself is interactive but
+    does not normally print such an announcement, your work based on
+    the Program is not required to print an announcement.)
+
+These requirements apply to the modified work as a whole.  If
+identifiable sections of that work are not derived from the Program,
+and can be reasonably considered independent and separate works in
+themselves, then this License, and its terms, do not apply to those
+sections when you distribute them as separate works.  But when you
+distribute the same sections as part of a whole which is a work based
+on the Program, the distribution of the whole must be on the terms of
+this License, whose permissions for other licensees extend to the
+entire whole, and thus to each and every part regardless of who wrote it.
+
+Thus, it is not the intent of this section to claim rights or contest
+your rights to work written entirely by you; rather, the intent is to
+exercise the right to control the distribution of derivative or
+collective works based on the Program.
+
+In addition, mere aggregation of another work not based on the Program
+with the Program (or with a work based on the Program) on a volume of
+a storage or distribution medium does not bring the other work under
+the scope of this License.
+
+  3. You may copy and distribute the Program (or a work based on it,
+under Section 2) in object code or executable form under the terms of
+Sections 1 and 2 above provided that you also do one of the following:
+
+    a) Accompany it with the complete corresponding machine-readable
+    source code, which must be distributed under the terms of Sections
+    1 and 2 above on a medium customarily used for software interchange; or,
+
+    b) Accompany it with a written offer, valid for at least three
+    years, to give any third party, for a charge no more than your
+    cost of physically performing source distribution, a complete
+    machine-readable copy of the corresponding source code, to be
+    distributed under the terms of Sections 1 and 2 above on a medium
+    customarily used for software interchange; or,
+
+    c) Accompany it with the information you received as to the offer
+    to distribute corresponding source code.  (This alternative is
+    allowed only for noncommercial distribution and only if you
+    received the program in object code or executable form with such
+    an offer, in accord with Subsection b above.)
+
+The source code for a work means the preferred form of the work for
+making modifications to it.  For an executable work, complete source
+code means all the source code for all modules it contains, plus any
+associated interface definition files, plus the scripts used to
+control compilation and installation of the executable.  However, as a
+special exception, the source code distributed need not include
+anything that is normally distributed (in either source or binary
+form) with the major components (compiler, kernel, and so on) of the
+operating system on which the executable runs, unless that component
+itself accompanies the executable.
+
+If distribution of executable or object code is made by offering
+access to copy from a designated place, then offering equivalent
+access to copy the source code from the same place counts as
+distribution of the source code, even though third parties are not
+compelled to copy the source along with the object code.
+
+  4. You may not copy, modify, sublicense, or distribute the Program
+except as expressly provided under this License.  Any attempt
+otherwise to copy, modify, sublicense or distribute the Program is
+void, and will automatically terminate your rights under this License.
+However, parties who have received copies, or rights, from you under
+this License will not have their licenses terminated so long as such
+parties remain in full compliance.
+
+  5. You are not required to accept this License, since you have not
+signed it.  However, nothing else grants you permission to modify or
+distribute the Program or its derivative works.  These actions are
+prohibited by law if you do not accept this License.  Therefore, by
+modifying or distributing the Program (or any work based on the
+Program), you indicate your acceptance of this License to do so, and
+all its terms and conditions for copying, distributing or modifying
+the Program or works based on it.
+
+  6. Each time you redistribute the Program (or any work based on the
+Program), the recipient automatically receives a license from the
+original licensor to copy, distribute or modify the Program subject to
+these terms and conditions.  You may not impose any further
+restrictions on the recipients' exercise of the rights granted herein.
+You are not responsible for enforcing compliance by third parties to
+this License.
+
+  7. If, as a consequence of a court judgment or allegation of patent
+infringement or for any other reason (not limited to patent issues),
+conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot
+distribute so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you
+may not distribute the Program at all.  For example, if a patent
+license would not permit royalty-free redistribution of the Program by
+all those who receive copies directly or indirectly through you, then
+the only way you could satisfy both it and this License would be to
+refrain entirely from distribution of the Program.
+
+If any portion of this section is held invalid or unenforceable under
+any particular circumstance, the balance of the section is intended to
+apply and the section as a whole is intended to apply in other
+circumstances.
+
+It is not the purpose of this section to induce you to infringe any
+patents or other property right claims or to contest validity of any
+such claims; this section has the sole purpose of protecting the
+integrity of the free software distribution system, which is
+implemented by public license practices.  Many people have made
+generous contributions to the wide range of software distributed
+through that system in reliance on consistent application of that
+system; it is up to the author/donor to decide if he or she is willing
+to distribute software through any other system and a licensee cannot
+impose that choice.
+
+This section is intended to make thoroughly clear what is believed to
+be a consequence of the rest of this License.
+
+  8. If the distribution and/or use of the Program is restricted in
+certain countries either by patents or by copyrighted interfaces, the
+original copyright holder who places the Program under this License
+may add an explicit geographical distribution limitation excluding
+those countries, so that distribution is permitted only in or among
+countries not thus excluded.  In such case, this License incorporates
+the limitation as if written in the body of this License.
+
+  9. The Free Software Foundation may publish revised and/or new versions
+of the General Public License from time to time.  Such new versions will
+be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+Each version is given a distinguishing version number.  If the Program
+specifies a version number of this License which applies to it and "any
+later version", you have the option of following the terms and conditions
+either of that version or of any later version published by the Free
+Software Foundation.  If the Program does not specify a version number of
+this License, you may choose any version ever published by the Free Software
+Foundation.
+
+  10. If you wish to incorporate parts of the Program into other free
+programs whose distribution conditions are different, write to the author
+to ask for permission.  For software which is copyrighted by the Free
+Software Foundation, write to the Free Software Foundation; we sometimes
+make exceptions for this.  Our decision will be guided by the two goals
+of preserving the free status of all derivatives of our free software and
+of promoting the sharing and reuse of software generally.
+
+                            NO WARRANTY
+
+  11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
+FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.  EXCEPT WHEN
+OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
+PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
+OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS
+TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
+PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
+REPAIR OR CORRECTION.
+
+  12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
+REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
+INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
+OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
+TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
+YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
+PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
+POSSIBILITY OF SUCH DAMAGES.
+
+                     END OF TERMS AND CONDITIONS
+
+            How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+convey the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    <one line to give the program's name and a brief idea of what it does.>
+    Copyright (C) <year>  <name of author>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+Also add information on how to contact you by electronic and paper mail.
+
+If the program is interactive, make it output a short notice like this
+when it starts in an interactive mode:
+
+    Gnomovision version 69, Copyright (C) year name of author
+    Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+    This is free software, and you are welcome to redistribute it
+    under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License.  Of course, the commands you use may
+be called something other than `show w' and `show c'; they could even be
+mouse-clicks or menu items--whatever suits your program.
+
+You should also get your employer (if you work as a programmer) or your
+school, if any, to sign a "copyright disclaimer" for the program, if
+necessary.  Here is a sample; alter the names:
+
+  Yoyodyne, Inc., hereby disclaims all copyright interest in the program
+  `Gnomovision' (which makes passes at compilers) written by James Hacker.
+
+  <signature of Ty Coon>, 1 April 1989
+  Ty Coon, President of Vice
+
+This General Public License does not permit incorporating your program into
+proprietary programs.  If your program is a subroutine library, you may
+consider it more useful to permit linking proprietary applications with the
+library.  If this is what you want to do, use the GNU Lesser General
+Public License instead of this License.

third_party/xxhash/tests/bench/Makefile

@@ -0,0 +1,68 @@
+# ################################################################
+# xxHash benchHash Makefile
+# Copyright (C) 2019-2021 Yann Collet
+#
+# GPL v2 License
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License along
+# with this program; if not, write to the Free Software Foundation, Inc.,
+# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+#
+# You can contact the author at:
+#   - xxHash homepage: https://www.xxhash.com
+#   - xxHash source repository: https://github.com/Cyan4973/xxHash
+# ################################################################
+# benchHash: A generic benchmark for hash algorithms
+#            measuring throughput, latency and bandwidth
+# ################################################################
+
+
+CPPFLAGS += -I../..   # directory of xxHash source files
+CFLAGS   ?= -O3
+CFLAGS   += -Wall -Wextra -Wstrict-aliasing=1 \
+            -std=c99
+CFLAGS   += $(MOREFLAGS)   # custom way to add flags
+CXXFLAGS ?= -O3
+LDFLAGS  += $(MOREFLAGS)
+
+
+OBJ_LIST  = main.o bhDisplay.o benchHash.o benchfn.o timefn.o
+
+
+default: benchHash
+
+all: benchHash
+
+benchHash32: CFLAGS   += -m32
+benchHash32: CXXFLAGS += -m32
+
+benchHash_avx2: CFLAGS   += -mavx2
+benchHash_avx2: CXXFLAGS += -mavx2
+
+benchHash_hw: CPPFLAGS += -DHARDWARE_SUPPORT
+benchHash_hw: CFLAGS   += -mavx2 -maes
+benchHash_hw: CXXFLAGS += -mavx2 -mpclmul -std=c++14
+
+benchHash benchHash32 benchHash_avx2 benchHash_nosimd benchHash_hw: $(OBJ_LIST)
+	$(CXX) $(CPPFLAGS) $(CXXFLAGS) $^ $(LDFLAGS) -o $@
+
+
+main.o: bhDisplay.h hashes.h
+
+bhDisplay.o: bhDisplay.h benchHash.h
+
+benchHash.o: benchHash.h
+
+
+clean:
+	$(RM) *.o benchHash benchHash32 benchHash_avx2 benchHash_hw

third_party/xxhash/tests/bench/benchHash.c

@@ -0,0 +1,166 @@
+/*
+*  Hash benchmark module
+*  Part of the xxHash project
+*  Copyright (C) 2019-2021 Yann Collet
+*
+*  GPL v2 License
+*
+*  This program is free software; you can redistribute it and/or modify
+*  it under the terms of the GNU General Public License as published by
+*  the Free Software Foundation; either version 2 of the License, or
+*  (at your option) any later version.
+*
+*  This program is distributed in the hope that it will be useful,
+*  but WITHOUT ANY WARRANTY; without even the implied warranty of
+*  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+*  GNU General Public License for more details.
+*
+*  You should have received a copy of the GNU General Public License along
+*  with this program; if not, write to the Free Software Foundation, Inc.,
+*  51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+*
+*  You can contact the author at:
+*  - xxHash homepage: https://www.xxhash.com
+*  - xxHash source repository: https://github.com/Cyan4973/xxHash
+*/
+
+/* benchmark hash functions */
+
+#include <stdlib.h>   // malloc
+#include <assert.h>
+#include <string.h>
+
+#include "benchHash.h"
+
+
+static void initBuffer(void* buffer, size_t size)
+{
+    const unsigned long long k1 = 11400714785074694791ULL;   /* 0b1001111000110111011110011011000110000101111010111100101010000111 */
+    const unsigned long long k2 = 14029467366897019727ULL;   /* 0b1100001010110010101011100011110100100111110101001110101101001111 */
+    unsigned long long acc = k2;
+    unsigned char* const p = (unsigned char*)buffer;
+    for (size_t s = 0; s < size; s++) {
+        acc *= k1;
+        p[s] = (unsigned char)(acc >> 56);
+    }
+}
+
+
+#define MARGIN_FOR_LATENCY 1024
+#define START_MASK (MARGIN_FOR_LATENCY-1)
+
+typedef size_t (*sizeFunction_f)(size_t targetSize);
+
+/*
+ * bench_hash_internal():
+ * Benchmarks hashfn repeateadly over single input of size `size`
+ * return: nb of hashes per second
+ */
+static double
+bench_hash_internal(BMK_benchFn_t hashfn, void* payload,
+                    size_t nbBlocks, sizeFunction_f selectSize, size_t size,
+                    unsigned total_time_ms, unsigned iter_time_ms)
+{
+    BMK_timedFnState_shell shell;
+    BMK_timedFnState_t* const txf = BMK_initStatic_timedFnState(&shell, sizeof(shell), total_time_ms, iter_time_ms);
+    assert(txf != NULL);
+
+    size_t const srcSize = (size_t)size;
+    size_t const srcBufferSize = srcSize + MARGIN_FOR_LATENCY;
+    void* const srcBuffer = malloc(srcBufferSize);
+    assert(srcBuffer != NULL);
+    initBuffer(srcBuffer, srcBufferSize);
+    #define FAKE_DSTSIZE 32
+    size_t const dstSize = FAKE_DSTSIZE;
+    char dstBuffer_static[FAKE_DSTSIZE] = {0};
+
+    #define NB_BLOCKS_MAX 1024
+    const void* srcBuffers[NB_BLOCKS_MAX];
+    size_t srcSizes[NB_BLOCKS_MAX];
+    void* dstBuffers[NB_BLOCKS_MAX];
+    size_t dstCapacities[NB_BLOCKS_MAX];
+    assert(nbBlocks < NB_BLOCKS_MAX);
+
+    assert(size > 0);
+    for (size_t n=0; n < nbBlocks; n++) {
+        srcBuffers[n] = srcBuffer;
+        srcSizes[n] = selectSize(size);
+        dstBuffers[n] = dstBuffer_static;
+        dstCapacities[n] = dstSize;
+    }
+
+
+    BMK_benchParams_t params = {
+        .benchFn = hashfn,
+        .benchPayload = payload,
+        .initFn = NULL,
+        .initPayload = NULL,
+        .errorFn = NULL,
+        .blockCount = nbBlocks,
+        .srcBuffers = srcBuffers,
+        .srcSizes = srcSizes,
+        .dstBuffers = dstBuffers,
+        .dstCapacities = dstCapacities,
+        .blockResults = NULL
+    };
+    BMK_runOutcome_t result;
+
+    memset(&result, 0, sizeof(result));
+    while (!BMK_isCompleted_TimedFn(txf)) {
+        result = BMK_benchTimedFn(txf, params);
+        assert(BMK_isSuccessful_runOutcome(result));
+    }
+
+    BMK_runTime_t const runTime = BMK_extract_runTime(result);
+
+    free(srcBuffer);
+    assert(runTime.nanoSecPerRun != 0);
+    return (1000000000U / runTime.nanoSecPerRun) * nbBlocks;
+
+}
+
+
+static size_t rand_1_N(size_t N) { return ((size_t)rand() % N)  + 1; }
+
+static size_t identity(size_t s) { return s; }
+
+static size_t
+benchLatency(const void* src, size_t srcSize,
+                   void* dst, size_t dstCapacity,
+                   void* customPayload)
+{
+    (void)dst; (void)dstCapacity;
+    BMK_benchFn_t benchfn = (BMK_benchFn_t)customPayload;
+    static size_t hash = 0;
+
+    const void* const start = (const char*)src + (hash & START_MASK);
+
+    return hash = benchfn(start, srcSize, dst, dstCapacity, NULL);
+}
+
+
+
+#ifndef SIZE_TO_HASH_PER_ROUND
+#  define SIZE_TO_HASH_PER_ROUND 200000
+#endif
+
+#ifndef NB_HASH_ROUNDS_MAX
+#  define NB_HASH_ROUNDS_MAX 1000
+#endif
+
+double bench_hash(BMK_benchFn_t hashfn,
+                  BMK_benchMode benchMode,
+                  size_t size, BMK_sizeMode sizeMode,
+                  unsigned total_time_ms, unsigned iter_time_ms)
+{
+    sizeFunction_f const sizef = (sizeMode == BMK_fixedSize) ? identity : rand_1_N;
+    BMK_benchFn_t const benchfn = (benchMode == BMK_throughput) ? hashfn : benchLatency;
+    BMK_benchFn_t const payload = (benchMode == BMK_throughput) ? NULL : hashfn;
+
+    size_t nbBlocks = (SIZE_TO_HASH_PER_ROUND / size) + 1;
+    if (nbBlocks > NB_HASH_ROUNDS_MAX) nbBlocks = NB_HASH_ROUNDS_MAX;
+
+    return bench_hash_internal(benchfn, payload,
+                               nbBlocks, sizef, size,
+                               total_time_ms, iter_time_ms);
+}

third_party/xxhash/tests/bench/benchHash.h

@@ -0,0 +1,67 @@
+/*
+*  Hash benchmark module
+*  Part of the xxHash project
+*  Copyright (C) 2019-2021 Yann Collet
+*
+*  GPL v2 License
+*
+*  This program is free software; you can redistribute it and/or modify
+*  it under the terms of the GNU General Public License as published by
+*  the Free Software Foundation; either version 2 of the License, or
+*  (at your option) any later version.
+*
+*  This program is distributed in the hope that it will be useful,
+*  but WITHOUT ANY WARRANTY; without even the implied warranty of
+*  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+*  GNU General Public License for more details.
+*
+*  You should have received a copy of the GNU General Public License along
+*  with this program; if not, write to the Free Software Foundation, Inc.,
+*  51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+*
+*  You can contact the author at:
+*  - xxHash homepage: https://www.xxhash.com
+*  - xxHash source repository: https://github.com/Cyan4973/xxHash
+*/
+
+
+#ifndef BENCH_HASH_H_983426678
+#define BENCH_HASH_H_983426678
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+
+/* ===  Dependencies  === */
+
+#include "benchfn.h"   /* BMK_benchFn_t */
+
+
+/* ===  Declarations  === */
+
+typedef enum { BMK_throughput, BMK_latency } BMK_benchMode;
+
+typedef enum { BMK_fixedSize,   /* hash always `size` bytes */
+               BMK_randomSize,  /* hash a random nb of bytes, between 1 and `size` (inclusive) */
+} BMK_sizeMode;
+
+/*
+ * bench_hash():
+ * Returns speed expressed as nb hashes per second.
+ * total_time_ms: time spent benchmarking the hash function with given parameters
+ * iter_time_ms: time spent for one round. If multiple rounds are run,
+ *               bench_hash() will report the speed of best round.
+ */
+double bench_hash(BMK_benchFn_t hashfn,
+                  BMK_benchMode benchMode,
+                  size_t size, BMK_sizeMode sizeMode,
+                  unsigned total_time_ms, unsigned iter_time_ms);
+
+
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif /* BENCH_HASH_H_983426678 */

third_party/xxhash/tests/bench/benchfn.c

@@ -0,0 +1,253 @@
+/*
+ * Copyright (C) 2016-2021 Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+
+/* *************************************
+*  Includes
+***************************************/
+#include <stdlib.h>      /* malloc, free */
+#include <string.h>      /* memset */
+#undef NDEBUG            /* assert must not be disabled */
+#include <assert.h>      /* assert */
+
+#include "timefn.h"        /* UTIL_time_t, UTIL_getTime */
+#include "benchfn.h"
+
+
+/* *************************************
+*  Constants
+***************************************/
+#define TIMELOOP_MICROSEC     SEC_TO_MICRO      /* 1 second */
+#define TIMELOOP_NANOSEC      (1*1000000000ULL) /* 1 second */
+
+#define KB *(1 <<10)
+#define MB *(1 <<20)
+#define GB *(1U<<30)
+
+
+/* *************************************
+*  Debug errors
+***************************************/
+#if defined(DEBUG) && (DEBUG >= 1)
+#  include <stdio.h>       /* fprintf */
+#  define DISPLAY(...)       fprintf(stderr, __VA_ARGS__)
+#  define DEBUGOUTPUT(...) { if (DEBUG) DISPLAY(__VA_ARGS__); }
+#else
+#  define DEBUGOUTPUT(...)
+#endif
+
+
+/* error without displaying */
+#define RETURN_QUIET_ERROR(retValue, ...) {           \
+    DEBUGOUTPUT("%s: %i: \n", __FILE__, __LINE__);    \
+    DEBUGOUTPUT("Error : ");                          \
+    DEBUGOUTPUT(__VA_ARGS__);                         \
+    DEBUGOUTPUT(" \n");                               \
+    return retValue;                                  \
+}
+
+
+/* *************************************
+*  Benchmarking an arbitrary function
+***************************************/
+
+int BMK_isSuccessful_runOutcome(BMK_runOutcome_t outcome)
+{
+    return outcome.error_tag_never_ever_use_directly == 0;
+}
+
+/* warning : this function will stop program execution if outcome is invalid !
+ *           check outcome validity first, using BMK_isValid_runResult() */
+BMK_runTime_t BMK_extract_runTime(BMK_runOutcome_t outcome)
+{
+    assert(outcome.error_tag_never_ever_use_directly == 0);
+    return outcome.internal_never_ever_use_directly;
+}
+
+size_t BMK_extract_errorResult(BMK_runOutcome_t outcome)
+{
+    assert(outcome.error_tag_never_ever_use_directly != 0);
+    return outcome.error_result_never_ever_use_directly;
+}
+
+static BMK_runOutcome_t BMK_runOutcome_error(size_t errorResult)
+{
+    BMK_runOutcome_t b;
+    memset(&b, 0, sizeof(b));
+    b.error_tag_never_ever_use_directly = 1;
+    b.error_result_never_ever_use_directly = errorResult;
+    return b;
+}
+
+static BMK_runOutcome_t BMK_setValid_runTime(BMK_runTime_t runTime)
+{
+    BMK_runOutcome_t outcome;
+    memset(&outcome, 0, sizeof(outcome));
+    outcome.error_tag_never_ever_use_directly = 0;
+    outcome.internal_never_ever_use_directly = runTime;
+    return outcome;
+}
+
+
+/* initFn will be measured once, benchFn will be measured `nbLoops` times */
+/* initFn is optional, provide NULL if none */
+/* benchFn must return a size_t value that errorFn can interpret */
+/* takes # of blocks and list of size & stuff for each. */
+/* can report result of benchFn for each block into blockResult. */
+/* blockResult is optional, provide NULL if this information is not required */
+/* note : time per loop can be reported as zero if run time < timer resolution */
+BMK_runOutcome_t BMK_benchFunction(BMK_benchParams_t p,
+                                   unsigned nbLoops)
+{
+    /* init */
+    {   size_t i;
+        for (i = 0; i < p.blockCount; i++) {
+            memset(p.dstBuffers[i], 0xE5, p.dstCapacities[i]);  /* warm up and erase result buffer */
+    }   }
+
+    /* benchmark */
+    {   UTIL_time_t const clockStart = UTIL_getTime();
+        size_t dstSize = 0;
+        unsigned loopNb, blockNb;
+        nbLoops += !nbLoops;   /* minimum nbLoops is 1 */
+        if (p.initFn != NULL) p.initFn(p.initPayload);
+        for (loopNb = 0; loopNb < nbLoops; loopNb++) {
+            for (blockNb = 0; blockNb < p.blockCount; blockNb++) {
+                size_t const res = p.benchFn(p.srcBuffers[blockNb], p.srcSizes[blockNb],
+                                   p.dstBuffers[blockNb], p.dstCapacities[blockNb],
+                                   p.benchPayload);
+                if (loopNb == 0) {
+                    if (p.blockResults != NULL) p.blockResults[blockNb] = res;
+                    if ((p.errorFn != NULL) && (p.errorFn(res))) {
+                        RETURN_QUIET_ERROR(BMK_runOutcome_error(res),
+                            "Function benchmark failed on block %u (of size %u) with error %i",
+                            blockNb, (unsigned)p.srcSizes[blockNb], (int)res);
+                    }
+                    dstSize += res;
+            }   }
+        }  /* for (loopNb = 0; loopNb < nbLoops; loopNb++) */
+
+        {   PTime const totalTime = UTIL_clockSpanNano(clockStart);
+            BMK_runTime_t rt;
+            rt.nanoSecPerRun = (double)totalTime / nbLoops;
+            rt.sumOfReturn = dstSize;
+            return BMK_setValid_runTime(rt);
+    }   }
+}
+
+
+/* ====  Benchmarking any function, providing intermediate results  ==== */
+
+struct BMK_timedFnState_s {
+    PTime timeSpent_ns;
+    PTime timeBudget_ns;
+    PTime runBudget_ns;
+    BMK_runTime_t fastestRun;
+    unsigned nbLoops;
+    UTIL_time_t coolTime;
+};  /* typedef'd to BMK_timedFnState_t within bench.h */
+
+BMK_timedFnState_t* BMK_createTimedFnState(unsigned total_ms, unsigned run_ms)
+{
+    BMK_timedFnState_t* const r = (BMK_timedFnState_t*)malloc(sizeof(*r));
+    if (r == NULL) return NULL;   /* malloc() error */
+    BMK_resetTimedFnState(r, total_ms, run_ms);
+    return r;
+}
+
+void BMK_freeTimedFnState(BMK_timedFnState_t* state) { free(state); }
+
+BMK_timedFnState_t*
+BMK_initStatic_timedFnState(void* buffer, size_t size, unsigned total_ms, unsigned run_ms)
+{
+    typedef char check_size[ 2 * (sizeof(BMK_timedFnState_shell) >= sizeof(struct BMK_timedFnState_s)) - 1];  /* static assert : a compilation failure indicates that BMK_timedFnState_shell is not large enough */
+    typedef struct { check_size c; BMK_timedFnState_t tfs; } tfs_align;  /* force tfs to be aligned at its next best position */
+    size_t const tfs_alignment = offsetof(tfs_align, tfs); /* provides the minimal alignment restriction for BMK_timedFnState_t */
+    BMK_timedFnState_t* const r = (BMK_timedFnState_t*)buffer;
+    if (buffer == NULL) return NULL;
+    if (size < sizeof(struct BMK_timedFnState_s)) return NULL;
+    if ((size_t)buffer % tfs_alignment) return NULL;  /* buffer must be properly aligned */
+    BMK_resetTimedFnState(r, total_ms, run_ms);
+    return r;
+}
+
+void BMK_resetTimedFnState(BMK_timedFnState_t* timedFnState, unsigned total_ms, unsigned run_ms)
+{
+    if (!total_ms) total_ms = 1 ;
+    if (!run_ms) run_ms = 1;
+    if (run_ms > total_ms) run_ms = total_ms;
+    timedFnState->timeSpent_ns = 0;
+    timedFnState->timeBudget_ns = (PTime)total_ms * TIMELOOP_NANOSEC / 1000;
+    timedFnState->runBudget_ns = (PTime)run_ms * TIMELOOP_NANOSEC / 1000;
+    timedFnState->fastestRun.nanoSecPerRun = (double)TIMELOOP_NANOSEC * 2000000000;  /* hopefully large enough : must be larger than any potential measurement */
+    timedFnState->fastestRun.sumOfReturn = (size_t)(-1LL);
+    timedFnState->nbLoops = 1;
+    timedFnState->coolTime = UTIL_getTime();
+}
+
+/* Tells if nb of seconds set in timedFnState for all runs is spent.
+ * note : this function will return 1 if BMK_benchFunctionTimed() has actually errored. */
+int BMK_isCompleted_TimedFn(const BMK_timedFnState_t* timedFnState)
+{
+    return (timedFnState->timeSpent_ns >= timedFnState->timeBudget_ns);
+}
+
+
+#undef MIN
+#define MIN(a,b)   ( (a) < (b) ? (a) : (b) )
+
+#define MINUSABLETIME  (TIMELOOP_NANOSEC / 2)  /* 0.5 seconds */
+
+BMK_runOutcome_t BMK_benchTimedFn(BMK_timedFnState_t* cont,
+                                  BMK_benchParams_t p)
+{
+    PTime const runBudget_ns = cont->runBudget_ns;
+    PTime const runTimeMin_ns = runBudget_ns / 2;
+    BMK_runTime_t bestRunTime = cont->fastestRun;
+
+    for (;;) {
+        BMK_runOutcome_t const runResult = BMK_benchFunction(p, cont->nbLoops);
+
+        if (!BMK_isSuccessful_runOutcome(runResult)) { /* error : move out */
+            return runResult;
+        }
+
+        {   BMK_runTime_t const newRunTime = BMK_extract_runTime(runResult);
+            double const loopDuration_ns = newRunTime.nanoSecPerRun * cont->nbLoops;
+
+            cont->timeSpent_ns += (unsigned long long)loopDuration_ns;
+
+            /* estimate nbLoops for next run to last approximately 1 second */
+            if (loopDuration_ns > (runBudget_ns / 50)) {
+                double const fastestRun_ns = MIN(bestRunTime.nanoSecPerRun, newRunTime.nanoSecPerRun);
+                cont->nbLoops = (unsigned)(runBudget_ns / fastestRun_ns) + 1;
+            } else {
+                /* previous run was too short : blindly increase workload by x multiplier */
+                const unsigned multiplier = 10;
+                assert(cont->nbLoops < ((unsigned)-1) / multiplier);  /* avoid overflow */
+                cont->nbLoops *= multiplier;
+            }
+
+            if (loopDuration_ns < runTimeMin_ns) {
+                /* When benchmark run time is too small : don't report results.
+                 * increased risks of rounding errors */
+                continue;
+            }
+
+            if (newRunTime.nanoSecPerRun < bestRunTime.nanoSecPerRun) {
+                bestRunTime = newRunTime;
+            }
+        }
+        break;
+    }   /* while (!completed) */
+
+    return BMK_setValid_runTime(bestRunTime);
+}

third_party/xxhash/tests/bench/benchfn.h

@@ -0,0 +1,183 @@
+/*
+ * Copyright (C) 2016-2021 Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+/* benchfn :
+ * benchmark any function on a set of input
+ * providing result in nanoSecPerRun
+ * or detecting and returning an error
+ */
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+#ifndef BENCH_FN_H_23876
+#define BENCH_FN_H_23876
+
+/* ===  Dependencies  === */
+#include <stddef.h>   /* size_t */
+
+
+/* ====  Benchmark any function, iterated on a set of blocks  ==== */
+
+/* BMK_runTime_t: valid result return type */
+
+typedef struct {
+    double nanoSecPerRun;  /* time per iteration (over all blocks) */
+    size_t sumOfReturn;         /* sum of return values */
+} BMK_runTime_t;
+
+
+/* BMK_runOutcome_t:
+ * type expressing the outcome of a benchmark run by BMK_benchFunction(),
+ * which can be either valid or invalid.
+ * benchmark outcome can be invalid if errorFn is provided.
+ * BMK_runOutcome_t must be considered "opaque" : never access its members directly.
+ * Instead, use its assigned methods :
+ * BMK_isSuccessful_runOutcome, BMK_extract_runTime, BMK_extract_errorResult.
+ * The structure is only described here to allow its allocation on stack. */
+
+typedef struct {
+    BMK_runTime_t internal_never_ever_use_directly;
+    size_t error_result_never_ever_use_directly;
+    int error_tag_never_ever_use_directly;
+} BMK_runOutcome_t;
+
+
+/* prototypes for benchmarked functions */
+typedef size_t (*BMK_benchFn_t)(const void* src, size_t srcSize, void* dst, size_t dstCapacity, void* customPayload);
+typedef size_t (*BMK_initFn_t)(void* initPayload);
+typedef unsigned (*BMK_errorFn_t)(size_t);
+
+
+/* BMK_benchFunction() parameters are provided via the following structure.
+ * A structure is preferable for readability,
+ * as the number of parameters required is fairly large.
+ * No initializer is provided, because it doesn't make sense to provide some "default" :
+ * all parameters must be specified by the caller.
+ * optional parameters are labelled explicitly, and accept value NULL when not used */
+typedef struct {
+    BMK_benchFn_t benchFn;    /* the function to benchmark, over the set of blocks */
+    void* benchPayload;       /* pass custom parameters to benchFn  :
+                               * (*benchFn)(srcBuffers[i], srcSizes[i], dstBuffers[i], dstCapacities[i], benchPayload) */
+    BMK_initFn_t initFn;      /* (*initFn)(initPayload) is run once per run, at the beginning. */
+    void* initPayload;        /* Both arguments can be NULL, in which case nothing is run. */
+    BMK_errorFn_t errorFn;    /* errorFn will check each return value of benchFn over each block, to determine if it failed or not.
+                               * errorFn can be NULL, in which case no check is performed.
+                               * errorFn must return 0 when benchFn was successful, and >= 1 if it detects an error.
+                               * Execution is stopped as soon as an error is detected.
+                               * the triggering return value can be retrieved using BMK_extract_errorResult(). */
+    size_t blockCount;        /* number of blocks to operate benchFn on.
+                               * It's also the size of all array parameters :
+                               * srcBuffers, srcSizes, dstBuffers, dstCapacities, blockResults */
+    const void *const * srcBuffers; /* read-only array of buffers to be operated on by benchFn */
+    const size_t* srcSizes;   /* read-only array containing sizes of srcBuffers */
+    void *const * dstBuffers; /* array of buffers to be written into by benchFn. This array is not optional, it must be provided even if unused by benchfn. */
+    const size_t* dstCapacities; /* read-only array containing capacities of dstBuffers. This array must be present. */
+    size_t* blockResults;     /* Optional: store the return value of benchFn for each block. Use NULL if this result is not requested. */
+} BMK_benchParams_t;
+
+
+/* BMK_benchFunction() :
+ * This function benchmarks benchFn and initFn, providing a result.
+ *
+ * params : see description of BMK_benchParams_t above.
+ * nbLoops: defines number of times benchFn is run over the full set of blocks.
+ *          Minimum value is 1. A 0 is interpreted as a 1.
+ *
+ * @return: can express either an error or a successful result.
+ *          Use BMK_isSuccessful_runOutcome() to check if benchmark was successful.
+ *          If yes, extract the result with BMK_extract_runTime(),
+ *          it will contain :
+ *              .sumOfReturn : the sum of all return values of benchFn through all of blocks
+ *              .nanoSecPerRun : time per run of benchFn + (time for initFn / nbLoops)
+ *          .sumOfReturn is generally intended for functions which return a # of bytes written into dstBuffer,
+ *              in which case, this value will be the total amount of bytes written into dstBuffer.
+ *
+ * blockResults : when provided (!= NULL), and when benchmark is successful,
+ *                params.blockResults contains all return values of `benchFn` over all blocks.
+ *                when provided (!= NULL), and when benchmark failed,
+ *                params.blockResults contains return values of `benchFn` over all blocks preceding and including the failed block.
+ */
+BMK_runOutcome_t BMK_benchFunction(BMK_benchParams_t params, unsigned nbLoops);
+
+
+
+/* check first if the benchmark was successful or not */
+int BMK_isSuccessful_runOutcome(BMK_runOutcome_t outcome);
+
+/* If the benchmark was successful, extract the result.
+ * note : this function will abort() program execution if benchmark failed !
+ *        always check if benchmark was successful first !
+ */
+BMK_runTime_t BMK_extract_runTime(BMK_runOutcome_t outcome);
+
+/* when benchmark failed, it means one invocation of `benchFn` failed.
+ * The failure was detected by `errorFn`, operating on return values of `benchFn`.
+ * Returns the faulty return value.
+ * note : this function will abort() program execution if benchmark did not failed.
+ *        always check if benchmark failed first !
+ */
+size_t BMK_extract_errorResult(BMK_runOutcome_t outcome);
+
+
+
+/* ====  Benchmark any function, returning intermediate results  ==== */
+
+/* state information tracking benchmark session */
+typedef struct BMK_timedFnState_s BMK_timedFnState_t;
+
+/* BMK_benchTimedFn() :
+ * Similar to BMK_benchFunction(), most arguments being identical.
+ * Automatically determines `nbLoops` so that each result is regularly produced at interval of about run_ms.
+ * Note : minimum `nbLoops` is 1, therefore a run may last more than run_ms, and possibly even more than total_ms.
+ * Usage - initialize timedFnState, select benchmark duration (total_ms) and each measurement duration (run_ms)
+ *         call BMK_benchTimedFn() repetitively, each measurement is supposed to last about run_ms
+ *         Check if total time budget is spent or exceeded, using BMK_isCompleted_TimedFn()
+ */
+BMK_runOutcome_t BMK_benchTimedFn(BMK_timedFnState_t* timedFnState,
+                                  BMK_benchParams_t params);
+
+/* Tells if duration of all benchmark runs has exceeded total_ms
+ */
+int BMK_isCompleted_TimedFn(const BMK_timedFnState_t* timedFnState);
+
+/* BMK_createTimedFnState() and BMK_resetTimedFnState() :
+ * Create/Set BMK_timedFnState_t for next benchmark session,
+ * which shall last a minimum of total_ms milliseconds,
+ * producing intermediate results, paced at interval of (approximately) run_ms.
+ */
+BMK_timedFnState_t* BMK_createTimedFnState(unsigned total_ms, unsigned run_ms);
+void BMK_resetTimedFnState(BMK_timedFnState_t* timedFnState, unsigned total_ms, unsigned run_ms);
+void BMK_freeTimedFnState(BMK_timedFnState_t* state);
+
+
+/* BMK_timedFnState_shell and BMK_initStatic_timedFnState() :
+ * Makes it possible to statically allocate a BMK_timedFnState_t on stack.
+ * BMK_timedFnState_shell is only there to allocate space,
+ * never ever access its members.
+ * BMK_timedFnState_t() actually accepts any buffer.
+ * It will check if provided buffer is large enough and is correctly aligned,
+ * and will return NULL if conditions are not respected.
+ */
+#define BMK_TIMEDFNSTATE_SIZE 64
+typedef union {
+    char never_access_space[BMK_TIMEDFNSTATE_SIZE];
+    long long alignment_enforcer;  /* must be aligned on 8-bytes boundaries */
+} BMK_timedFnState_shell;
+BMK_timedFnState_t* BMK_initStatic_timedFnState(void* buffer, size_t size, unsigned total_ms, unsigned run_ms);
+
+
+#endif   /* BENCH_FN_H_23876 */
+
+#if defined (__cplusplus)
+}
+#endif

third_party/xxhash/tests/bench/bhDisplay.c

@@ -0,0 +1,160 @@
+/*
+*  CSV Display module for the hash benchmark program
+*  Part of the xxHash project
+*  Copyright (C) 2019-2021 Yann Collet
+*
+*  GPL v2 License
+*
+*  This program is free software; you can redistribute it and/or modify
+*  it under the terms of the GNU General Public License as published by
+*  the Free Software Foundation; either version 2 of the License, or
+*  (at your option) any later version.
+*
+*  This program is distributed in the hope that it will be useful,
+*  but WITHOUT ANY WARRANTY; without even the implied warranty of
+*  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+*  GNU General Public License for more details.
+*
+*  You should have received a copy of the GNU General Public License along
+*  with this program; if not, write to the Free Software Foundation, Inc.,
+*  51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+*
+*  You can contact the author at :
+*  - xxHash homepage : https://www.xxhash.com
+*  - xxHash source repository : https://github.com/Cyan4973/xxHash
+*/
+
+
+/* ===  Dependencies  === */
+
+#include <stdlib.h>   /* rand */
+#include <stdio.h>    /* printf */
+#include <assert.h>
+
+#include "benchHash.h"
+#include "bhDisplay.h"
+
+
+/* ===  benchmark large input  === */
+
+#define MB_UNIT           1000000
+#define BENCH_LARGE_ITER_MS   490
+#define BENCH_LARGE_TOTAL_MS 1010
+static void bench_oneHash_largeInput(Bench_Entry hashDesc, int minlog, int maxlog)
+{
+    printf("%-7s", hashDesc.name);
+    for (int sizelog=minlog; sizelog<=maxlog; sizelog++) {
+        size_t const inputSize = (size_t)1 << sizelog;
+        double const nbhps = bench_hash(hashDesc.hash, BMK_throughput,
+                                        inputSize, BMK_fixedSize,
+                                        BENCH_LARGE_TOTAL_MS, BENCH_LARGE_ITER_MS);
+        printf(",%6.0f", nbhps * inputSize / MB_UNIT); fflush(NULL);
+    }
+    printf("\n");
+}
+
+void bench_largeInput(Bench_Entry const* hashDescTable, int nbHashes, int minlog, int maxlog)
+{
+    assert(maxlog <  31);
+    assert(minlog >=  0);
+    printf("benchmarking large inputs : from %u bytes (log%i) to %u MB (log%i) \n",
+        1U << minlog, minlog,
+        (1U << maxlog) >> 20, maxlog);
+    for (int i=0; i<nbHashes; i++)
+        bench_oneHash_largeInput(hashDescTable[i], minlog, maxlog);
+}
+
+
+
+/* ===  Benchmark small inputs  === */
+
+#define BENCH_SMALL_ITER_MS   170
+#define BENCH_SMALL_TOTAL_MS  490
+static void bench_throughput_oneHash_smallInputs(Bench_Entry hashDesc, size_t sizeMin, size_t sizeMax)
+{
+    printf("%-7s", hashDesc.name);
+    for (size_t s=sizeMin; s<sizeMax+1; s++) {
+        double const nbhps = bench_hash(hashDesc.hash, BMK_throughput,
+                                        s, BMK_fixedSize,
+                                        BENCH_SMALL_TOTAL_MS, BENCH_SMALL_ITER_MS);
+        printf(",%10.0f", nbhps); fflush(NULL);
+    }
+    printf("\n");
+}
+
+void bench_throughput_smallInputs(Bench_Entry const* hashDescTable, int nbHashes, size_t sizeMin, size_t sizeMax)
+{
+    printf("Throughput small inputs of fixed size (from %zu to %zu bytes): \n",
+            sizeMin, sizeMax);
+    for (int i=0; i<nbHashes; i++)
+        bench_throughput_oneHash_smallInputs(hashDescTable[i], sizeMin, sizeMax);
+}
+
+
+
+/* ===   Latency measurements (small keys)   === */
+
+static void bench_latency_oneHash_smallInputs(Bench_Entry hashDesc, size_t size_min, size_t size_max)
+{
+    printf("%-7s", hashDesc.name);
+    for (size_t s=size_min; s<size_max+1; s++) {
+        double const nbhps = bench_hash(hashDesc.hash, BMK_latency,
+                                        s, BMK_fixedSize,
+                                        BENCH_SMALL_TOTAL_MS, BENCH_SMALL_ITER_MS);
+        printf(",%10.0f", nbhps); fflush(NULL);
+    }
+    printf("\n");
+}
+
+void bench_latency_smallInputs(Bench_Entry const* hashDescTable, int nbHashes, size_t size_min, size_t size_max)
+{
+    printf("Latency for small inputs of fixed size : \n");
+    for (int i=0; i<nbHashes; i++)
+        bench_latency_oneHash_smallInputs(hashDescTable[i], size_min, size_max);
+}
+
+
+/* ===   Random input Length   === */
+
+static void bench_randomInputLength_withOneHash(Bench_Entry hashDesc, size_t size_min, size_t size_max)
+{
+    printf("%-7s", hashDesc.name);
+    for (size_t s=size_min; s<size_max+1; s++) {
+        srand((unsigned)s);   /* ensure random sequence of length will be the same for a given s */
+        double const nbhps = bench_hash(hashDesc.hash, BMK_throughput,
+                                        s, BMK_randomSize,
+                                        BENCH_SMALL_TOTAL_MS, BENCH_SMALL_ITER_MS);
+        printf(",%10.0f", nbhps); fflush(NULL);
+    }
+    printf("\n");
+}
+
+void bench_throughput_randomInputLength(Bench_Entry const* hashDescTable, int nbHashes, size_t size_min, size_t size_max)
+{
+    printf("benchmarking random size inputs [1-N] : \n");
+    for (int i=0; i<nbHashes; i++)
+        bench_randomInputLength_withOneHash(hashDescTable[i], size_min, size_max);
+}
+
+
+/* ===   Latency with Random input Length   === */
+
+static void bench_latency_oneHash_randomInputLength(Bench_Entry hashDesc, size_t size_min, size_t size_max)
+{
+    printf("%-7s", hashDesc.name);
+    for (size_t s=size_min; s<size_max+1; s++) {
+        srand((unsigned)s);   /* ensure random sequence of length will be the same for a given s */
+        double const nbhps = bench_hash(hashDesc.hash, BMK_latency,
+                                        s, BMK_randomSize,
+                                        BENCH_SMALL_TOTAL_MS, BENCH_SMALL_ITER_MS);
+        printf(",%10.0f", nbhps); fflush(NULL);
+    }
+    printf("\n");
+}
+
+void bench_latency_randomInputLength(Bench_Entry const* hashDescTable, int nbHashes, size_t size_min, size_t size_max)
+{
+    printf("Latency for small inputs of random size [1-N] : \n");
+    for (int i=0; i<nbHashes; i++)
+        bench_latency_oneHash_randomInputLength(hashDescTable[i], size_min, size_max);
+}

third_party/xxhash/tests/bench/bhDisplay.h

@@ -0,0 +1,61 @@
+/*
+*  CSV Display module for the hash benchmark program
+*  Part of the xxHash project
+*  Copyright (C) 2019-2021 Yann Collet
+*
+*  GPL v2 License
+*
+*  This program is free software; you can redistribute it and/or modify
+*  it under the terms of the GNU General Public License as published by
+*  the Free Software Foundation; either version 2 of the License, or
+*  (at your option) any later version.
+*
+*  This program is distributed in the hope that it will be useful,
+*  but WITHOUT ANY WARRANTY; without even the implied warranty of
+*  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+*  GNU General Public License for more details.
+*
+*  You should have received a copy of the GNU General Public License along
+*  with this program; if not, write to the Free Software Foundation, Inc.,
+*  51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+*
+*  You can contact the author at:
+*  - xxHash homepage: https://www.xxhash.com
+*  - xxHash source repository: https://github.com/Cyan4973/xxHash
+*/
+
+#ifndef BH_DISPLAY_H_192088098
+#define BH_DISPLAY_H_192088098
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+
+/* ===  Dependencies  === */
+
+#include "benchfn.h"   /* BMK_benchFn_t */
+
+
+/* ===  Declarations  === */
+
+typedef struct {
+    const char* name;
+    BMK_benchFn_t hash;
+} Bench_Entry;
+
+void bench_largeInput(Bench_Entry const* hashDescTable, int nbHashes, int sizeLogMin, int sizeLogMax);
+
+void bench_throughput_smallInputs(Bench_Entry const* hashDescTable, int nbHashes, size_t sizeMin, size_t sizeMax);
+void bench_throughput_randomInputLength(Bench_Entry const* hashDescTable, int nbHashes, size_t sizeMin, size_t sizeMax);
+
+void bench_latency_smallInputs(Bench_Entry const* hashDescTable, int nbHashes, size_t sizeMin, size_t sizeMax);
+void bench_latency_randomInputLength(Bench_Entry const* hashDescTable, int nbHashes, size_t sizeMin, size_t sizeMax);
+
+
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif   /* BH_DISPLAY_H_192088098 */

third_party/xxhash/tests/bench/hashes.h

@@ -0,0 +1,118 @@
+/*
+*  List hash algorithms to benchmark
+*  Part of xxHash project
+*  Copyright (C) 2019-2021 Yann Collet
+*
+*  GPL v2 License
+*
+*  This program is free software; you can redistribute it and/or modify
+*  it under the terms of the GNU General Public License as published by
+*  the Free Software Foundation; either version 2 of the License, or
+*  (at your option) any later version.
+*
+*  This program is distributed in the hope that it will be useful,
+*  but WITHOUT ANY WARRANTY; without even the implied warranty of
+*  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+*  GNU General Public License for more details.
+*
+*  You should have received a copy of the GNU General Public License along
+*  with this program; if not, write to the Free Software Foundation, Inc.,
+*  51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+*
+*  You can contact the author at:
+*  - xxHash homepage: https://www.xxhash.com
+*  - xxHash source repository: https://github.com/Cyan4973/xxHash
+*/
+
+
+/* ===   Dependencies   === */
+
+#include <stddef.h>   /* size_t */
+
+
+/* ==================================================
+ *   Non-portable hash algorithms
+ * =============================================== */
+
+
+#ifdef HARDWARE_SUPPORT
+
+/*
+ * List any hash algorithms that depend on specific hardware support,
+ * including for example:
+ * - Hardware crc32c
+ * - Hardware AES support
+ * - Carryless Multipliers (clmul)
+ * - AVX2
+ */
+
+#endif
+
+
+
+/* ==================================================
+ * List of hashes
+ * ==================================================
+ * Each hash must be wrapped in a thin redirector conformant with the BMK_benchfn_t.
+ * BMK_benchfn_t is generic, not specifically designed for hashes.
+ * For hashes, the following parameters are expected to be useless:
+ * dst, dstCapacity, customPayload.
+ *
+ * The result of each hash is assumed to be provided as function return value.
+ * This condition is important for latency measurements.
+ */
+
+ /* ===  xxHash  === */
+#define XXH_INLINE_ALL
+#include "xxhash.h"
+
+size_t XXH32_wrapper(const void* src, size_t srcSize, void* dst, size_t dstCapacity, void* customPayload)
+{
+    (void)dst; (void)dstCapacity; (void)customPayload;
+    return (size_t) XXH32(src, srcSize, 0);
+}
+
+
+size_t XXH64_wrapper(const void* src, size_t srcSize, void* dst, size_t dstCapacity, void* customPayload)
+{
+    (void)dst; (void)dstCapacity; (void)customPayload;
+    return (size_t) XXH64(src, srcSize, 0);
+}
+
+
+size_t xxh3_wrapper(const void* src, size_t srcSize, void* dst, size_t dstCapacity, void* customPayload)
+{
+    (void)dst; (void)dstCapacity; (void)customPayload;
+    return (size_t) XXH3_64bits(src, srcSize);
+}
+
+
+size_t XXH128_wrapper(const void* src, size_t srcSize, void* dst, size_t dstCapacity, void* customPayload)
+{
+    (void)dst; (void)dstCapacity; (void)customPayload;
+    return (size_t) XXH3_128bits(src, srcSize).low64;
+}
+
+
+
+/* ==================================================
+ * Table of hashes
+ * =============================================== */
+
+#include "bhDisplay.h"   /* Bench_Entry */
+
+#ifndef HARDWARE_SUPPORT
+#  define NB_HASHES 4
+#else
+#  define NB_HASHES 4
+#endif
+
+Bench_Entry const hashCandidates[NB_HASHES] = {
+    { "xxh3"  , xxh3_wrapper },
+    { "XXH32" , XXH32_wrapper },
+    { "XXH64" , XXH64_wrapper },
+    { "XXH128", XXH128_wrapper },
+#ifdef HARDWARE_SUPPORT
+    /* list here codecs which require specific hardware support, such SSE4.1, PCLMUL, AVX2, etc. */
+#endif
+};

third_party/xxhash/tests/bench/main.c

@@ -0,0 +1,220 @@
+/*
+ * Main program to benchmark hash functions
+ * Part of the xxHash project
+ * Copyright (C) 2019-2021 Yann Collet
+ * GPL v2 License
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * You can contact the author at:
+ * - xxHash homepage: https://www.xxhash.com
+ * - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+
+/* ===  dependencies  === */
+
+#include <stdio.h>       /* printf */
+#include <limits.h>      /* INT_MAX */
+#include "bhDisplay.h"   /* bench_x */
+
+
+/* ===  defines list of hashes `hashCandidates` and NB_HASHES  *** */
+
+#include "hashes.h"
+
+
+/* ===  parse command line  === */
+
+#undef NDEBUG
+#include <assert.h>
+
+
+/*!
+ * readIntFromChar():
+ * Allows and interprets K, KB, KiB, M, MB and MiB suffix.
+ * Will also modify `*stringPtr`, advancing it to position where it stopped reading.
+ */
+static int readIntFromChar(const char** stringPtr)
+{
+    static int const max = (INT_MAX / 10) - 1;
+    int result = 0;
+    while ((**stringPtr >='0') && (**stringPtr <='9')) {
+        assert(result < max);
+        result *= 10;
+        result += (unsigned)(**stringPtr - '0');
+        (*stringPtr)++ ;
+    }
+    if ((**stringPtr=='K') || (**stringPtr=='M')) {
+        int const maxK = INT_MAX >> 10;
+        assert(result < maxK);
+        result <<= 10;
+        if (**stringPtr=='M') {
+            assert(result < maxK);
+            result <<= 10;
+        }
+        (*stringPtr)++;  /* skip `K` or `M` */
+        if (**stringPtr=='i') (*stringPtr)++;
+        if (**stringPtr=='B') (*stringPtr)++;
+    }
+    return result;
+}
+
+
+/**
+ * isCommand():
+ * Checks if string is the same as longCommand.
+ * If yes, @return 1, otherwise @return 0
+ */
+static int isCommand(const char* string, const char* longCommand)
+{
+    assert(string);
+    assert(longCommand);
+    size_t const comSize = strlen(longCommand);
+    return !strncmp(string, longCommand, comSize);
+}
+
+/*
+ * longCommandWArg():
+ * Checks if *stringPtr is the same as longCommand.
+ * If yes, @return 1 and advances *stringPtr to the position which immediately
+ * follows longCommand.
+ * @return 0 and doesn't modify *stringPtr otherwise.
+ */
+static int longCommandWArg(const char** stringPtr, const char* longCommand)
+{
+    assert(stringPtr);
+    assert(longCommand);
+    size_t const comSize = strlen(longCommand);
+    int const result = isCommand(*stringPtr, longCommand);
+    if (result) *stringPtr += comSize;
+    return result;
+}
+
+
+/* ===   default values - can be redefined at compilation time   === */
+
+#ifndef SMALL_SIZE_MIN_DEFAULT
+#  define SMALL_SIZE_MIN_DEFAULT   1
+#endif
+#ifndef SMALL_SIZE_MAX_DEFAULT
+#  define SMALL_SIZE_MAX_DEFAULT 127
+#endif
+#ifndef LARGE_SIZELOG_MIN_DEFAULT
+#  define LARGE_SIZELOG_MIN_DEFAULT   9
+#endif
+#ifndef LARGE_SIZELOG_MAX_DEFAULT
+#  define LARGE_SIZELOG_MAX_DEFAULT  27
+#endif
+
+
+static int display_hash_names(void)
+{
+    int i;
+    printf("available hashes : \n");
+    for (i=0; i<NB_HASHES; i++) {
+        printf("%s, ", hashCandidates[i].name);
+    }
+    printf("\b\b  \n");
+    return 0;
+}
+
+/*
+ * @return: hashID (necessarily between 0 and NB_HASHES) if present
+ *          -1 on error (hname not present)
+ */
+static int hashID(const char* hname)
+{
+    int id;
+    assert(hname);
+    for (id=0; id < NB_HASHES; id++) {
+        assert(hashCandidates[id].name);
+        if (strlen(hname) != strlen(hashCandidates[id].name)) continue;
+        if (isCommand(hname, hashCandidates[id].name)) return id;
+    }
+    return -1;
+}
+
+static int help(const char* exename)
+{
+    printf("Usage: %s [options]... [hash]\n", exename);
+    printf("Runs various benchmarks at various lengths for the listed hash functions\n");
+    printf("and outputs them in a CSV format.\n\n");
+    printf("Options: \n");
+    printf("  --list       Name available hash algorithms and exit \n");
+    printf("  --mins=LEN   Starting length for small size bench (default: %i) \n", SMALL_SIZE_MIN_DEFAULT);
+    printf("  --maxs=LEN   End length for small size bench (default: %i) \n", SMALL_SIZE_MAX_DEFAULT);
+    printf("  --minl=LEN   Starting log2(length) for large size bench (default: %i) \n", LARGE_SIZELOG_MIN_DEFAULT);
+    printf("  --maxl=LEN   End log2(length) for large size bench (default: %i) \n", LARGE_SIZELOG_MAX_DEFAULT);
+    printf("  [hash]       Optional, bench all available hashes if not provided \n");
+    return 0;
+}
+
+static int badusage(const char* exename)
+{
+    printf("Bad command ... \n");
+    help(exename);
+    return 1;
+}
+
+int main(int argc, const char* argv[])
+{
+    const char* const exename = argv[0];
+    int hashNb = 0;
+    int nb_h_test = NB_HASHES;
+    int largeTest_log_min = LARGE_SIZELOG_MIN_DEFAULT;
+    int largeTest_log_max = LARGE_SIZELOG_MAX_DEFAULT;
+    size_t smallTest_size_min = SMALL_SIZE_MIN_DEFAULT;
+    size_t smallTest_size_max = SMALL_SIZE_MAX_DEFAULT;
+
+    int arg_nb;
+    for (arg_nb = 1; arg_nb < argc; arg_nb++) {
+        const char** arg = argv + arg_nb;
+        if (isCommand(*arg, "-h")) { assert(argc >= 1); return help(exename); }
+        if (isCommand(*arg, "--list")) { return display_hash_names(); }
+        if (longCommandWArg(arg, "--n=")) { nb_h_test = readIntFromChar(arg); continue; }  /* hidden command */
+        if (longCommandWArg(arg, "--minl=")) { largeTest_log_min = readIntFromChar(arg); continue; }
+        if (longCommandWArg(arg, "--maxl=")) { largeTest_log_max = readIntFromChar(arg); continue; }
+        if (longCommandWArg(arg, "--mins=")) { smallTest_size_min = (size_t)readIntFromChar(arg); continue; }
+        if (longCommandWArg(arg, "--maxs=")) { smallTest_size_max = (size_t)readIntFromChar(arg); continue; }
+        /* not a command: must be a hash name */
+        hashNb = hashID(*arg);
+        if (hashNb >= 0) {
+            nb_h_test = 1;
+        } else {
+            /* not a hash name: error */
+            return badusage(exename);
+        }
+    }
+
+    /* border case (requires (mis)using hidden command `--n=#`) */
+    if (hashNb + nb_h_test > NB_HASHES) {
+        printf("wrong hash selection \n");
+        return 1;
+    }
+
+    printf(" ===  benchmarking %i hash functions  === \n", nb_h_test);
+    if (largeTest_log_max >= largeTest_log_min) {
+        bench_largeInput(hashCandidates+hashNb, nb_h_test, largeTest_log_min, largeTest_log_max);
+    }
+    if (smallTest_size_max >= smallTest_size_min) {
+        bench_throughput_smallInputs(hashCandidates+hashNb, nb_h_test, smallTest_size_min, smallTest_size_max);
+        bench_throughput_randomInputLength(hashCandidates+hashNb, nb_h_test, smallTest_size_min, smallTest_size_max);
+        bench_latency_smallInputs(hashCandidates+hashNb, nb_h_test, smallTest_size_min, smallTest_size_max);
+        bench_latency_randomInputLength(hashCandidates+hashNb, nb_h_test, smallTest_size_min, smallTest_size_max);
+    }
+
+    return 0;
+}

third_party/xxhash/tests/bench/timefn.c

@@ -0,0 +1,168 @@
+/*
+ * Copyright (C) 2019-2021 Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+/* ===  Dependencies  === */
+
+#include "timefn.h"
+
+
+/*-****************************************
+*  Time functions
+******************************************/
+
+#if defined(_WIN32)   /* Windows */
+
+#include <stdlib.h>   /* abort */
+#include <stdio.h>    /* perror */
+
+UTIL_time_t UTIL_getTime(void) { UTIL_time_t x; QueryPerformanceCounter(&x); return x; }
+
+PTime UTIL_getSpanTimeMicro(UTIL_time_t clockStart, UTIL_time_t clockEnd)
+{
+    static LARGE_INTEGER ticksPerSecond;
+    static int init = 0;
+    if (!init) {
+        if (!QueryPerformanceFrequency(&ticksPerSecond)) {
+            perror("timefn::QueryPerformanceFrequency");
+            abort();
+        }
+        init = 1;
+    }
+    return 1000000ULL*(clockEnd.QuadPart - clockStart.QuadPart)/ticksPerSecond.QuadPart;
+}
+
+PTime UTIL_getSpanTimeNano(UTIL_time_t clockStart, UTIL_time_t clockEnd)
+{
+    static LARGE_INTEGER ticksPerSecond;
+    static int init = 0;
+    if (!init) {
+        if (!QueryPerformanceFrequency(&ticksPerSecond)) {
+            perror("timefn::QueryPerformanceFrequency");
+            abort();
+        }
+        init = 1;
+    }
+    return 1000000000ULL*(clockEnd.QuadPart - clockStart.QuadPart)/ticksPerSecond.QuadPart;
+}
+
+
+
+#elif defined(__APPLE__) && defined(__MACH__)
+
+UTIL_time_t UTIL_getTime(void) { return mach_absolute_time(); }
+
+PTime UTIL_getSpanTimeMicro(UTIL_time_t clockStart, UTIL_time_t clockEnd)
+{
+    static mach_timebase_info_data_t rate;
+    static int init = 0;
+    if (!init) {
+        mach_timebase_info(&rate);
+        init = 1;
+    }
+    return (((clockEnd - clockStart) * (PTime)rate.numer) / ((PTime)rate.denom))/1000ULL;
+}
+
+PTime UTIL_getSpanTimeNano(UTIL_time_t clockStart, UTIL_time_t clockEnd)
+{
+    static mach_timebase_info_data_t rate;
+    static int init = 0;
+    if (!init) {
+        mach_timebase_info(&rate);
+        init = 1;
+    }
+    return ((clockEnd - clockStart) * (PTime)rate.numer) / ((PTime)rate.denom);
+}
+
+
+
+#elif (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) /* C11 */) \
+    && defined(TIME_UTC) /* C11 requires timespec_get, but FreeBSD 11 lacks it, while still claiming C11 compliance */
+
+#include <stdlib.h>   /* abort */
+#include <stdio.h>    /* perror */
+
+UTIL_time_t UTIL_getTime(void)
+{
+    /* time must be initialized, othersize it may fail msan test.
+     * No good reason, likely a limitation of timespec_get() for some target */
+    UTIL_time_t time = UTIL_TIME_INITIALIZER;
+    if (timespec_get(&time, TIME_UTC) != TIME_UTC) {
+        perror("timefn::timespec_get");
+        abort();
+    }
+    return time;
+}
+
+static UTIL_time_t UTIL_getSpanTime(UTIL_time_t begin, UTIL_time_t end)
+{
+    UTIL_time_t diff;
+    if (end.tv_nsec < begin.tv_nsec) {
+        diff.tv_sec = (end.tv_sec - 1) - begin.tv_sec;
+        diff.tv_nsec = (end.tv_nsec + 1000000000ULL) - begin.tv_nsec;
+    } else {
+        diff.tv_sec = end.tv_sec - begin.tv_sec;
+        diff.tv_nsec = end.tv_nsec - begin.tv_nsec;
+    }
+    return diff;
+}
+
+PTime UTIL_getSpanTimeMicro(UTIL_time_t begin, UTIL_time_t end)
+{
+    UTIL_time_t const diff = UTIL_getSpanTime(begin, end);
+    PTime micro = 0;
+    micro += 1000000ULL * diff.tv_sec;
+    micro += diff.tv_nsec / 1000ULL;
+    return micro;
+}
+
+PTime UTIL_getSpanTimeNano(UTIL_time_t begin, UTIL_time_t end)
+{
+    UTIL_time_t const diff = UTIL_getSpanTime(begin, end);
+    PTime nano = 0;
+    nano += 1000000000ULL * diff.tv_sec;
+    nano += diff.tv_nsec;
+    return nano;
+}
+
+
+
+#else   /* relies on standard C90 (note : clock_t measurements can be wrong when using multi-threading) */
+
+UTIL_time_t UTIL_getTime(void) { return clock(); }
+PTime UTIL_getSpanTimeMicro(UTIL_time_t clockStart, UTIL_time_t clockEnd) { return 1000000ULL * (clockEnd - clockStart) / CLOCKS_PER_SEC; }
+PTime UTIL_getSpanTimeNano(UTIL_time_t clockStart, UTIL_time_t clockEnd) { return 1000000000ULL * (clockEnd - clockStart) / CLOCKS_PER_SEC; }
+
+#endif
+
+
+
+/* returns time span in microseconds */
+PTime UTIL_clockSpanMicro(UTIL_time_t clockStart )
+{
+    UTIL_time_t const clockEnd = UTIL_getTime();
+    return UTIL_getSpanTimeMicro(clockStart, clockEnd);
+}
+
+/* returns time span in microseconds */
+PTime UTIL_clockSpanNano(UTIL_time_t clockStart )
+{
+    UTIL_time_t const clockEnd = UTIL_getTime();
+    return UTIL_getSpanTimeNano(clockStart, clockEnd);
+}
+
+void UTIL_waitForNextTick(void)
+{
+    UTIL_time_t const clockStart = UTIL_getTime();
+    UTIL_time_t clockEnd;
+    do {
+        clockEnd = UTIL_getTime();
+    } while (UTIL_getSpanTimeNano(clockStart, clockEnd) == 0);
+}

third_party/xxhash/tests/bench/timefn.h

@@ -0,0 +1,89 @@
+/*
+ * Copyright (c) 2016-2021 Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef TIME_FN_H_MODULE_287987
+#define TIME_FN_H_MODULE_287987
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+
+/*-****************************************
+*  Dependencies
+******************************************/
+#include <sys/types.h>    /* utime */
+#if defined(_MSC_VER)
+#  include <sys/utime.h>  /* utime */
+#else
+#  include <utime.h>      /* utime */
+#endif
+#include <time.h>         /* clock_t, clock, CLOCKS_PER_SEC */
+
+
+
+/*-****************************************
+*  Local Types
+******************************************/
+
+#if !defined (__VMS) && (defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) )
+# include <stdint.h>
+  typedef uint64_t           PTime;  /* Precise Time */
+#else
+  typedef unsigned long long PTime;  /* does not support compilers without long long support */
+#endif
+
+
+
+/*-****************************************
+*  Time functions
+******************************************/
+#if defined(_WIN32)   /* Windows */
+
+    #include <Windows.h>   /* LARGE_INTEGER */
+    typedef LARGE_INTEGER UTIL_time_t;
+    #define UTIL_TIME_INITIALIZER { { 0, 0 } }
+
+#elif defined(__APPLE__) && defined(__MACH__)
+
+    #include <mach/mach_time.h>
+    typedef PTime UTIL_time_t;
+    #define UTIL_TIME_INITIALIZER 0
+
+#elif (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) /* C11 */) \
+    && defined(TIME_UTC) /* C11 requires timespec_get, but FreeBSD 11 lacks it, while still claiming C11 compliance */
+
+    typedef struct timespec UTIL_time_t;
+    #define UTIL_TIME_INITIALIZER { 0, 0 }
+
+#else   /* relies on standard C90 (note : clock_t measurements can be wrong when using multi-threading) */
+
+    typedef clock_t UTIL_time_t;
+    #define UTIL_TIME_INITIALIZER 0
+
+#endif
+
+
+UTIL_time_t UTIL_getTime(void);
+PTime UTIL_getSpanTimeMicro(UTIL_time_t clockStart, UTIL_time_t clockEnd);
+PTime UTIL_getSpanTimeNano(UTIL_time_t clockStart, UTIL_time_t clockEnd);
+
+#define SEC_TO_MICRO ((PTime)1000000)
+PTime UTIL_clockSpanMicro(UTIL_time_t clockStart);
+PTime UTIL_clockSpanNano(UTIL_time_t clockStart);
+
+void UTIL_waitForNextTick(void);
+
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif /* TIME_FN_H_MODULE_287987 */

third_party/xxhash/tests/collisions/.gitignore

@@ -0,0 +1,2 @@
+#build artefacts
+collisionsTest

third_party/xxhash/tests/collisions/LICENSE

@@ -0,0 +1,339 @@
+                    GNU GENERAL PUBLIC LICENSE
+                       Version 2, June 1991
+
+ Copyright (C) 1989, 1991 Free Software Foundation, Inc.,
+ 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+                            Preamble
+
+  The licenses for most software are designed to take away your
+freedom to share and change it.  By contrast, the GNU General Public
+License is intended to guarantee your freedom to share and change free
+software--to make sure the software is free for all its users.  This
+General Public License applies to most of the Free Software
+Foundation's software and to any other program whose authors commit to
+using it.  (Some other Free Software Foundation software is covered by
+the GNU Lesser General Public License instead.)  You can apply it to
+your programs, too.
+
+  When we speak of free software, we are referring to freedom, not
+price.  Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+this service if you wish), that you receive source code or can get it
+if you want it, that you can change the software or use pieces of it
+in new free programs; and that you know you can do these things.
+
+  To protect your rights, we need to make restrictions that forbid
+anyone to deny you these rights or to ask you to surrender the rights.
+These restrictions translate to certain responsibilities for you if you
+distribute copies of the software, or if you modify it.
+
+  For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must give the recipients all the rights that
+you have.  You must make sure that they, too, receive or can get the
+source code.  And you must show them these terms so they know their
+rights.
+
+  We protect your rights with two steps: (1) copyright the software, and
+(2) offer you this license which gives you legal permission to copy,
+distribute and/or modify the software.
+
+  Also, for each author's protection and ours, we want to make certain
+that everyone understands that there is no warranty for this free
+software.  If the software is modified by someone else and passed on, we
+want its recipients to know that what they have is not the original, so
+that any problems introduced by others will not reflect on the original
+authors' reputations.
+
+  Finally, any free program is threatened constantly by software
+patents.  We wish to avoid the danger that redistributors of a free
+program will individually obtain patent licenses, in effect making the
+program proprietary.  To prevent this, we have made it clear that any
+patent must be licensed for everyone's free use or not licensed at all.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+
+                    GNU GENERAL PUBLIC LICENSE
+   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
+
+  0. This License applies to any program or other work which contains
+a notice placed by the copyright holder saying it may be distributed
+under the terms of this General Public License.  The "Program", below,
+refers to any such program or work, and a "work based on the Program"
+means either the Program or any derivative work under copyright law:
+that is to say, a work containing the Program or a portion of it,
+either verbatim or with modifications and/or translated into another
+language.  (Hereinafter, translation is included without limitation in
+the term "modification".)  Each licensee is addressed as "you".
+
+Activities other than copying, distribution and modification are not
+covered by this License; they are outside its scope.  The act of
+running the Program is not restricted, and the output from the Program
+is covered only if its contents constitute a work based on the
+Program (independent of having been made by running the Program).
+Whether that is true depends on what the Program does.
+
+  1. You may copy and distribute verbatim copies of the Program's
+source code as you receive it, in any medium, provided that you
+conspicuously and appropriately publish on each copy an appropriate
+copyright notice and disclaimer of warranty; keep intact all the
+notices that refer to this License and to the absence of any warranty;
+and give any other recipients of the Program a copy of this License
+along with the Program.
+
+You may charge a fee for the physical act of transferring a copy, and
+you may at your option offer warranty protection in exchange for a fee.
+
+  2. You may modify your copy or copies of the Program or any portion
+of it, thus forming a work based on the Program, and copy and
+distribute such modifications or work under the terms of Section 1
+above, provided that you also meet all of these conditions:
+
+    a) You must cause the modified files to carry prominent notices
+    stating that you changed the files and the date of any change.
+
+    b) You must cause any work that you distribute or publish, that in
+    whole or in part contains or is derived from the Program or any
+    part thereof, to be licensed as a whole at no charge to all third
+    parties under the terms of this License.
+
+    c) If the modified program normally reads commands interactively
+    when run, you must cause it, when started running for such
+    interactive use in the most ordinary way, to print or display an
+    announcement including an appropriate copyright notice and a
+    notice that there is no warranty (or else, saying that you provide
+    a warranty) and that users may redistribute the program under
+    these conditions, and telling the user how to view a copy of this
+    License.  (Exception: if the Program itself is interactive but
+    does not normally print such an announcement, your work based on
+    the Program is not required to print an announcement.)
+
+These requirements apply to the modified work as a whole.  If
+identifiable sections of that work are not derived from the Program,
+and can be reasonably considered independent and separate works in
+themselves, then this License, and its terms, do not apply to those
+sections when you distribute them as separate works.  But when you
+distribute the same sections as part of a whole which is a work based
+on the Program, the distribution of the whole must be on the terms of
+this License, whose permissions for other licensees extend to the
+entire whole, and thus to each and every part regardless of who wrote it.
+
+Thus, it is not the intent of this section to claim rights or contest
+your rights to work written entirely by you; rather, the intent is to
+exercise the right to control the distribution of derivative or
+collective works based on the Program.
+
+In addition, mere aggregation of another work not based on the Program
+with the Program (or with a work based on the Program) on a volume of
+a storage or distribution medium does not bring the other work under
+the scope of this License.
+
+  3. You may copy and distribute the Program (or a work based on it,
+under Section 2) in object code or executable form under the terms of
+Sections 1 and 2 above provided that you also do one of the following:
+
+    a) Accompany it with the complete corresponding machine-readable
+    source code, which must be distributed under the terms of Sections
+    1 and 2 above on a medium customarily used for software interchange; or,
+
+    b) Accompany it with a written offer, valid for at least three
+    years, to give any third party, for a charge no more than your
+    cost of physically performing source distribution, a complete
+    machine-readable copy of the corresponding source code, to be
+    distributed under the terms of Sections 1 and 2 above on a medium
+    customarily used for software interchange; or,
+
+    c) Accompany it with the information you received as to the offer
+    to distribute corresponding source code.  (This alternative is
+    allowed only for noncommercial distribution and only if you
+    received the program in object code or executable form with such
+    an offer, in accord with Subsection b above.)
+
+The source code for a work means the preferred form of the work for
+making modifications to it.  For an executable work, complete source
+code means all the source code for all modules it contains, plus any
+associated interface definition files, plus the scripts used to
+control compilation and installation of the executable.  However, as a
+special exception, the source code distributed need not include
+anything that is normally distributed (in either source or binary
+form) with the major components (compiler, kernel, and so on) of the
+operating system on which the executable runs, unless that component
+itself accompanies the executable.
+
+If distribution of executable or object code is made by offering
+access to copy from a designated place, then offering equivalent
+access to copy the source code from the same place counts as
+distribution of the source code, even though third parties are not
+compelled to copy the source along with the object code.
+
+  4. You may not copy, modify, sublicense, or distribute the Program
+except as expressly provided under this License.  Any attempt
+otherwise to copy, modify, sublicense or distribute the Program is
+void, and will automatically terminate your rights under this License.
+However, parties who have received copies, or rights, from you under
+this License will not have their licenses terminated so long as such
+parties remain in full compliance.
+
+  5. You are not required to accept this License, since you have not
+signed it.  However, nothing else grants you permission to modify or
+distribute the Program or its derivative works.  These actions are
+prohibited by law if you do not accept this License.  Therefore, by
+modifying or distributing the Program (or any work based on the
+Program), you indicate your acceptance of this License to do so, and
+all its terms and conditions for copying, distributing or modifying
+the Program or works based on it.
+
+  6. Each time you redistribute the Program (or any work based on the
+Program), the recipient automatically receives a license from the
+original licensor to copy, distribute or modify the Program subject to
+these terms and conditions.  You may not impose any further
+restrictions on the recipients' exercise of the rights granted herein.
+You are not responsible for enforcing compliance by third parties to
+this License.
+
+  7. If, as a consequence of a court judgment or allegation of patent
+infringement or for any other reason (not limited to patent issues),
+conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot
+distribute so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you
+may not distribute the Program at all.  For example, if a patent
+license would not permit royalty-free redistribution of the Program by
+all those who receive copies directly or indirectly through you, then
+the only way you could satisfy both it and this License would be to
+refrain entirely from distribution of the Program.
+
+If any portion of this section is held invalid or unenforceable under
+any particular circumstance, the balance of the section is intended to
+apply and the section as a whole is intended to apply in other
+circumstances.
+
+It is not the purpose of this section to induce you to infringe any
+patents or other property right claims or to contest validity of any
+such claims; this section has the sole purpose of protecting the
+integrity of the free software distribution system, which is
+implemented by public license practices.  Many people have made
+generous contributions to the wide range of software distributed
+through that system in reliance on consistent application of that
+system; it is up to the author/donor to decide if he or she is willing
+to distribute software through any other system and a licensee cannot
+impose that choice.
+
+This section is intended to make thoroughly clear what is believed to
+be a consequence of the rest of this License.
+
+  8. If the distribution and/or use of the Program is restricted in
+certain countries either by patents or by copyrighted interfaces, the
+original copyright holder who places the Program under this License
+may add an explicit geographical distribution limitation excluding
+those countries, so that distribution is permitted only in or among
+countries not thus excluded.  In such case, this License incorporates
+the limitation as if written in the body of this License.
+
+  9. The Free Software Foundation may publish revised and/or new versions
+of the General Public License from time to time.  Such new versions will
+be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+Each version is given a distinguishing version number.  If the Program
+specifies a version number of this License which applies to it and "any
+later version", you have the option of following the terms and conditions
+either of that version or of any later version published by the Free
+Software Foundation.  If the Program does not specify a version number of
+this License, you may choose any version ever published by the Free Software
+Foundation.
+
+  10. If you wish to incorporate parts of the Program into other free
+programs whose distribution conditions are different, write to the author
+to ask for permission.  For software which is copyrighted by the Free
+Software Foundation, write to the Free Software Foundation; we sometimes
+make exceptions for this.  Our decision will be guided by the two goals
+of preserving the free status of all derivatives of our free software and
+of promoting the sharing and reuse of software generally.
+
+                            NO WARRANTY
+
+  11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
+FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.  EXCEPT WHEN
+OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
+PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
+OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS
+TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
+PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
+REPAIR OR CORRECTION.
+
+  12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
+REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
+INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
+OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
+TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
+YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
+PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
+POSSIBILITY OF SUCH DAMAGES.
+
+                     END OF TERMS AND CONDITIONS
+
+            How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+convey the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    <one line to give the program's name and a brief idea of what it does.>
+    Copyright (C) <year>  <name of author>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License along
+    with this program; if not, write to the Free Software Foundation, Inc.,
+    51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+Also add information on how to contact you by electronic and paper mail.
+
+If the program is interactive, make it output a short notice like this
+when it starts in an interactive mode:
+
+    Gnomovision version 69, Copyright (C) year name of author
+    Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+    This is free software, and you are welcome to redistribute it
+    under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License.  Of course, the commands you use may
+be called something other than `show w' and `show c'; they could even be
+mouse-clicks or menu items--whatever suits your program.
+
+You should also get your employer (if you work as a programmer) or your
+school, if any, to sign a "copyright disclaimer" for the program, if
+necessary.  Here is a sample; alter the names:
+
+  Yoyodyne, Inc., hereby disclaims all copyright interest in the program
+  `Gnomovision' (which makes passes at compilers) written by James Hacker.
+
+  <signature of Ty Coon>, 1 April 1989
+  Ty Coon, President of Vice
+
+This General Public License does not permit incorporating your program into
+proprietary programs.  If your program is a subroutine library, you may
+consider it more useful to permit linking proprietary applications with the
+library.  If this is what you want to do, use the GNU Lesser General
+Public License instead of this License.

third_party/xxhash/tests/collisions/Makefile

@@ -0,0 +1,75 @@
+#  Brute force collision tester for 64-bit hashes
+#  Part of xxHash project
+#  Copyright (C) 2019-2021 Yann Collet
+#
+# GPL v2 License
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License along
+# with this program; if not, write to the Free Software Foundation, Inc.,
+# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+#
+#  You can contact the author at:
+#  - xxHash homepage: https://www.xxhash.com
+#  - xxHash source repository: https://github.com/Cyan4973/xxHash
+#
+
+SRC_DIRS = ./ ../../ allcodecs/
+VPATH = $(SRC_DIRS)
+CPPFLAGS += $(addprefix -I ,$(SRC_DIRS))
+CFLAGS   += -Wall -Wextra -Wconversion \
+            -std=c99
+CXXFLAGS += -Wall -Wextra -Wconversion \
+            -std=c++11
+LDFLAGS  += -pthread
+TESTHASHES = 110000000
+
+HASH_SRC := $(sort $(wildcard allcodecs/*.c allcodecs/*.cc))
+HASH_OBJ := $(patsubst %.c,%.o,$(HASH_SRC))
+
+
+.PHONY: default
+default: release
+
+.PHONY: all
+all: release
+
+collisionsTest: main.o pool.o threading.o sort.o $(HASH_OBJ)
+	$(CXX) $(CPPFLAGS) $(CXXFLAGS) $^ $(LDFLAGS) -o $@
+
+main.o: hashes.h xxhash.h
+
+release: CXXFLAGS += -O3
+release: CFLAGS += -O3
+release: collisionsTest
+
+debug: CXXFLAGS += -g3 -O0 -DDEBUG
+debug: CFLAGS += -g3 -O0 -DDEBUG
+debug: collisionsTest
+
+.PHONY: check
+check: test
+
+.PHONY: test
+test: debug
+	@echo ""
+	@echo "## $(TESTHASHES) hashes with original and 0 threads"
+	@time ./collisionsTest --nbh=$(TESTHASHES)
+	@echo ""
+	@echo "## $(TESTHASHES) hashes with original and 4 threads"
+	@time ./collisionsTest --nbh=$(TESTHASHES) --threadlog=2
+	@echo ""
+
+.PHONY: clean
+clean:
+	$(RM) *.o allcodecs/*.o
+	$(RM) collisionsTest

third_party/xxhash/tests/collisions/README.md

@@ -0,0 +1,122 @@
+
+__collisionsTest__ is a brute force hash analyzer
+which will measure a 64-bit hash algorithm's collision rate
+by generating billions of hashes,
+and comparing the result to an "ideal" target.
+
+The test requires a very large amount of memory.
+By default, it will generate 24 billion of 64-bit hashes,
+requiring __192 GB of RAM__ for their storage.
+The number of hashes can be modified using command `--nbh=`.
+Be aware that testing the collision ratio of 64-bit hashes
+requires a very large amount of hashes (several billion) for meaningful measurements.
+
+To reduce RAM usage, an optional filter can be requested, with `--filter`.
+It reduces the nb of candidates to analyze, hence associated RAM budget.
+Note that the filter itself requires a lot of RAM
+(32 GB by default, can be modified using `--filterlog=`,
+a too small filter will not be efficient, aim at ~2 bytes per hash),
+and reading and writing into filter cost a significant CPU budget,
+so this method is slower.
+It also doesn't allow advanced analysis of partial bitfields,
+since most hashes will be discarded and not stored.
+
+When using the filter, the RAM budget consists of the filter and a list of candidates,
+which will be a fraction of the original hash list.
+Using default settings (24 billion hashes, 32 GB filter),
+the number of potential candidates should be reduced to less than 2 billion,
+requiring ~14 GB for their storage.
+Such a result also depends on hash algorithm's efficiency.
+The number of effective candidates is likely to be lower, at ~ 1 billion,
+but storage must allocate an upper bound.
+
+For the default test, the expected "optimal" collision rate for a 64-bit hash function is ~18 collisions.
+
+#### How to build
+```
+make
+```
+
+Note: the code is a mix of C99 and C++14,
+it's not compatible with a C90-only compiler.
+
+#### Build modifier
+
+- `SLAB5`: use alternative pattern generator, friendlier for weak hash algorithms
+- `POOL_MT`: if `=0`, disable multi-threading code (enabled by default)
+
+#### How to integrate any hash in the tester
+
+The build script will compile files found in `./allcodecs`.
+Put the source code here.
+This also works if the hash is a single `*.h` file.
+
+The glue happens in `hashes.h`.
+In this file, there are 2 sections:
+- Adds the required `#include "header.h"`, and creates a wrapper
+to respect the format expected by the function pointer.
+- Adds the wrapper, along with the name and an indication of the output width,
+to the table, at the end of `hashes.h`
+
+Build with `make`. Locate your new hash with `./collisionsTest -h`,
+it should be listed.
+
+
+#### Usage
+
+```
+usage: ./collisionsTest [hashName] [opt]
+
+list of hashNames: (...)
+
+Optional parameters:
+  --nbh=NB       Select nb of hashes to generate (25769803776 by default)
+  --filter       Enable the filter. Slower, but reduces memory usage for same nb of hashes.
+  --threadlog=NB Use 2^NB threads
+  --len=NB       Select length of input (255 bytes by default)
+```
+
+#### Some advises on how to setup a collisions test
+
+Most tests are primarily driven by the amount of RAM available.
+Here's a method to decide the size of the test.
+
+Presuming that RAM budget is not plentiful, for this example 32 GB,
+the `--filter` mode is actually compulsory to measure anything meaningful.
+Let's plan 50% of memory for the filter, that's 16 GB.
+This will be good enough to filter about 10% less hashes than this size.
+Let's round down to 14 G.
+
+By requesting 14G, the expectation is that the program will automatically
+size the filter to 16 GB, and expect to store ~1G candidates,
+leaving enough room to breeze for the system.
+
+The command line becomes:
+```
+./collisionsTest --nbh=14G --filter NameOfHash
+```
+
+#### Examples:
+
+Here are a few results produced with this tester:
+
+| Algorithm | Input Len | Nb Hashes | Expected | Nb Collisions | Notes |
+| ---        | --- | ---    | ---   | --- | --- |
+| __XXH3__   | 255 | 100 Gi | 312.5 | 326 |  |
+| __XXH64__  | 255 | 100 Gi | 312.5 | 294 |  |
+| __XXH128__ low64 | 512 | 100 Gi | 312.5 | 321 |  |
+| __XXH128__ high64| 512 | 100 Gi | 312.5 | 325 |  |
+| __XXH128__ | 255 | 100 Gi |   0.0 |   0 | a 128-bit hash is expected to generate 0 collisions |
+
+Test on small inputs:
+
+| Algorithm  | Input Len | Nb Hashes | Expected | Nb Collisions | Notes |
+| ---        | --- | ---    | --- | --- | --- |
+| __XXH64__  |   8 | 100 Gi | 312.5 | __0__ | `XXH64` is bijective for `len==8` |
+| __XXH3__   |   8 | 100 Gi | 312.5 | __0__ | `XXH3` is also bijective for `len==8` |
+| __XXH3__   |  16 | 100 Gi | 312.5 | 332 |  |
+| __XXH3__   |  32 |  14 Gi |   6.1 |   3 |  |
+| __XXH128__ |  16 |  25 Gi |   0.0 |   0 | test range 9-16 |
+| __XXH128__ |  32 |  25 Gi |   0.0 |   0 | test range 17-128 |
+| __XXH128__ | 100 |  13 Gi |   0.0 |   0 | test range 17-128 |
+| __XXH128__ | 200 |  13 Gi |   0.0 |   0 | test range 129-240 |

third_party/xxhash/tests/collisions/allcodecs/README.md

@@ -0,0 +1 @@
+Put in this directory all hash algorithms to test

third_party/xxhash/tests/collisions/allcodecs/dummy.c

@@ -0,0 +1,38 @@
+/*
+ * dummy.c, a fake hash algorithm, just to test integration capabilities.
+ * Part of the xxHash project
+ * Copyright (C) 2020 Yann Collet
+ *
+ * GPL v2 License
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * You can contact the author at:
+ * - xxHash homepage: https://www.xxhash.com
+ * - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+
+#include <dummy.h>
+
+unsigned badsum32(const void* input, size_t len, unsigned seed)
+{
+    unsigned sum = seed;
+    const unsigned char* in8 = input;
+    size_t c;
+    for (c=0; c<len; c++)
+        sum += in8[c];
+    return sum;
+}

third_party/xxhash/tests/collisions/allcodecs/dummy.h

@@ -0,0 +1,45 @@
+/*
+ * dummy.c,
+ * A fake hash algorithm, just to test integration capabilities.
+ * Part of the xxHash project
+ * Copyright (C) 2020 Yann Collet
+ *
+ * GPL v2 License
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * You can contact the author at:
+ * - xxHash homepage: https://www.xxhash.com
+ * - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+#ifndef DUMMY_H_987987
+#define DUMMY_H_987987
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+
+#include <stddef.h> /* size_t */
+
+unsigned badsum32(const void* input, size_t len, unsigned seed);
+
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif  /* DUMMY_H_987987 */

third_party/xxhash/tests/collisions/hashes.h

@@ -0,0 +1,127 @@
+/*
+ * List of hashes for the brute force collision tester
+ * Part of xxHash project
+ * Copyright (C) 2019-2021 Yann Collet
+ *
+ * GPL v2 License
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * You can contact the author at:
+ * - xxHash homepage: https://www.xxhash.com
+ * - xxHash source repository: https://github.com/Cyan4973/xxHash
+ */
+
+#ifndef HASHES_H_1235465
+#define HASHES_H_1235465
+
+#include <stddef.h>      /* size_t */
+#include <stdint.h>      /* uint64_t */
+#define XXH_INLINE_ALL   /* XXH128_hash_t */
+#include "xxhash.h"
+
+
+/* return type */
+
+typedef union {
+    uint64_t       h64;
+    XXH128_hash_t h128;
+} UniHash;
+
+UniHash uniHash32(uint64_t v32)
+{   UniHash unih;
+    unih.h64 = v32;
+    return unih;
+}
+
+UniHash uniHash64(uint64_t v64)
+{   UniHash unih;
+    unih.h64 = v64;
+    return unih;
+}
+
+UniHash uniHash128(XXH128_hash_t v128)
+{   UniHash unih;
+    unih.h128 = v128;
+    return unih;
+}
+
+
+/* ===  xxHash  === */
+
+UniHash XXH3_wrapper (const void* data, size_t size)
+{
+    return uniHash64( XXH3_64bits(data, size) );
+}
+
+UniHash XXH128_wrapper (const void* data, size_t size)
+{
+    return uniHash128( XXH3_128bits(data, size) );
+}
+
+UniHash XXH128l_wrapper (const void* data, size_t size)
+{
+    return uniHash64( XXH3_128bits(data, size).low64 );
+}
+
+UniHash XXH128h_wrapper (const void* data, size_t size)
+{
+    return uniHash64( XXH3_128bits(data, size).high64 );
+}
+
+UniHash XXH64_wrapper (const void* data, size_t size)
+{
+    return uniHash64 ( XXH64(data, size, 0) );
+}
+
+UniHash XXH32_wrapper (const void* data, size_t size)
+{
+    return uniHash32( XXH32(data, size, 0) );
+}
+
+/* ===  Dummy integration example  === */
+
+#include "dummy.h"
+
+UniHash badsum32_wrapper (const void* data, size_t size)
+{
+    return uniHash32( badsum32(data, size, 0) );
+}
+
+
+
+/* ===  Table  === */
+
+typedef UniHash (*hashfn) (const void* data, size_t size);
+
+typedef struct {
+    const char* name;
+    hashfn fn;
+    int bits;
+} hashDescription;
+
+#define HASH_FN_TOTAL 7
+
+hashDescription hashfnTable[HASH_FN_TOTAL] = {
+    { "xxh3"  ,  XXH3_wrapper,     64 },
+    { "xxh64" ,  XXH64_wrapper,    64 },
+    { "xxh128",  XXH128_wrapper,  128 },
+    { "xxh128l", XXH128l_wrapper,  64 },
+    { "xxh128h", XXH128h_wrapper,  64 },
+    { "xxh32" ,  XXH32_wrapper,    32 },
+    { "badsum32",badsum32_wrapper, 32 },
+};
+
+#endif   /* HASHES_H_1235465 */