[2.0.0-dev] add xxhash

.gitmodules

@@ -34,3 +34,6 @@
 [submodule "libraries/libfc/libraries/boringssl/bssl"]
 	path = libraries/libfc/libraries/boringssl/bssl
 	url = https://boringssl.googlesource.com/boringssl
+[submodule "libraries/libfc/libraries/xxhash/xxHash"]
+	path = libraries/libfc/libraries/xxhash/xxHash
+	url = https://github.com/Cyan4973/xxHash

CMakeLists.txt

@@ -245,6 +245,7 @@ configure_file(libraries/prometheus/prometheus-cpp/LICENSE                 licen
 configure_file(programs/cleos/LICENSE.CLI11                                licenses/spring/LICENSE.CLI11     COPYONLY)
 configure_file(libraries/libfc/libraries/bls12-381/LICENSE                 licenses/spring/LICENSE.bls12-381 COPYONLY)
 configure_file(libraries/libfc/libraries/boringssl/bssl/LICENSE            licenses/spring/LICENSE.boringssl COPYONLY)
+configure_file(libraries/libfc/libraries/xxhash/xxHash/LICENSE             licenses/spring/LICENSE.xxHash    COPYONLY)
 
 install(DIRECTORY "${CMAKE_CURRENT_BINARY_DIR}/licenses/spring" DESTINATION "${CMAKE_INSTALL_FULL_DATAROOTDIR}/licenses/" COMPONENT base)
 

CMakeModules/EosioTester.cmake.in

@@ -37,6 +37,7 @@ endif ( APPLE )
 
 add_compile_definitions(BOOST_TEST_DEFAULTS_TO_CORE_DUMP)
 add_compile_definitions(BOOST_UNORDERED_DISABLE_NEON)
+add_compile_definitions(XXH_INLINE_ALL=1)
 
 set( Boost_USE_MULTITHREADED      ON )
 set( Boost_USE_STATIC_LIBS ON CACHE STRING "ON or OFF" )
@@ -118,6 +119,7 @@ target_include_directories(EosioChain INTERFACE
                               @CMAKE_INSTALL_FULL_INCLUDEDIR@
                               @CMAKE_INSTALL_FULL_INCLUDEDIR@/wasm-jit
                               @CMAKE_INSTALL_FULL_INCLUDEDIR@/springboringssl
+                              @CMAKE_INSTALL_FULL_INCLUDEDIR@/springxxhash
                               @CMAKE_INSTALL_FULL_INCLUDEDIR@/softfloat )
 
 #adds -lrt and -latomic. Ubuntu eosio.contracts build breaks without this

CMakeModules/EosioTesterBuild.cmake.in

@@ -34,6 +34,7 @@ endif ( APPLE )
 
 add_compile_definitions(BOOST_TEST_DEFAULTS_TO_CORE_DUMP)
 add_compile_definitions(BOOST_UNORDERED_DISABLE_NEON)
+add_compile_definitions(XXH_INLINE_ALL=1)
 
 set( Boost_USE_MULTITHREADED      ON )
 set( Boost_USE_STATIC_LIBS ON CACHE STRING "ON or OFF" )
@@ -116,6 +117,7 @@ target_include_directories(EosioChain INTERFACE
                               @CMAKE_SOURCE_DIR@/libraries/libfc/include
                               @CMAKE_SOURCE_DIR@/libraries/libfc/libraries/boringssl/bssl/include
                               @CMAKE_SOURCE_DIR@/libraries/libfc/libraries/bls12-381/include
+                              @CMAKE_SOURCE_DIR@/libraries/libfc/libraries/xxhash
                               @CMAKE_SOURCE_DIR@/libraries/softfloat/source/include
                               @CMAKE_SOURCE_DIR@/libraries/appbase/include
                               @CMAKE_SOURCE_DIR@/libraries/chaindb/include

benchmark/hash.cpp

@@ -4,6 +4,7 @@
 #include <fc/crypto/sha256.hpp>
 #include <fc/crypto/sha512.hpp>
 #include <fc/crypto/ripemd160.hpp>
+#include <fc/crypto/xxh3.hpp>
 #include <fc/utility.hpp>
 
 #include <benchmark.hpp>
@@ -13,7 +14,11 @@ using namespace fc;
 namespace eosio::benchmark {
 
 void hash_benchmarking() {
-   std::string small_message = "abcdefghijklmnopqrstuvwxyz0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ01";
+   const std::string really_small_message = "abcdefghijklmnop";
+   const std::string small_message = "abcdefghijklmnopqrstuvwxyz0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ01";
+
+   std::span<const char, 16> really_small_message_span(really_small_message.data(), 16);
+   std::span<const char, 64> small_message_span(small_message.data(), 64);
 
    // build a large message
    constexpr auto large_msg_size = 4096;
@@ -84,6 +89,31 @@ void hash_benchmarking() {
    };
    benchmarking("keccak256 (" + std::to_string(large_message.length()) + " bytes)", keccak_large_msg);
 
+   auto xxh3_really_small_message = [&]() {
+      fc::xxh3::hash(really_small_message);
+   };
+   benchmarking("xxh3 (" + std::to_string(really_small_message.length()) + " bytes)", xxh3_really_small_message);
+
+   auto xxh3_small_msg = [&]() {
+      fc::xxh3::hash(small_message);
+   };
+   benchmarking("xxh3 (" + std::to_string(small_message.length()) + " bytes)", xxh3_small_msg);
+
+   auto xxh3_large_msg = [&]() {
+      fc::xxh3::hash(large_message);
+   };
+   benchmarking("xxh3 (" + std::to_string(large_message.length()) + " bytes)", xxh3_large_msg);
+
+   auto xxh3_really_small_fixed_message = [&]() {
+      fc::xxh3::hash_raw(really_small_message_span);
+   };
+   benchmarking("xxh3 fixed (" + std::to_string(really_small_message_span.size()) + " bytes)", xxh3_really_small_fixed_message);
+
+   auto xxh3_small_fixed_msg = [&]() {
+      fc::xxh3::hash_raw(small_message_span);
+   };
+   benchmarking("xxh3 fixed (" + std::to_string(small_message_span.size()) + " bytes)", xxh3_small_fixed_msg);
+
 }
 
 } // benchmark

libraries/libfc/CMakeLists.txt

@@ -2,6 +2,7 @@ add_subdirectory( secp256k1 )
 add_subdirectory( libraries/bn256/src )
 add_subdirectory( libraries/bls12-381 )
 add_subdirectory( libraries/boringssl )
+add_subdirectory( libraries/xxhash )
 
 set(CMAKE_THREAD_PREFER_PTHREAD TRUE)
 set(THREADS_PREFER_PTHREAD_FLAG TRUE)
@@ -59,6 +60,7 @@ set( fc_sources
      src/crypto/modular_arithmetic.cpp
      src/crypto/blake2.cpp
      src/crypto/k1_recover.cpp
+     src/crypto/xxh3.cpp
      src/network/url.cpp
      src/network/http/http_client.cpp
      src/compress/zlib.cpp
@@ -105,7 +107,7 @@ endif()
 
 target_link_libraries( fc PUBLIC Boost::date_time Boost::chrono Boost::iostreams Boost::interprocess Boost::multi_index Boost::dll
                                  Boost::multiprecision Boost::beast Boost::asio Boost::thread Boost::crc Threads::Threads
-                                 boringssl ZLIB::ZLIB ${PLATFORM_SPECIFIC_LIBS} ${CMAKE_DL_LIBS} secp256k1 bls12-381
+                                 boringssl ZLIB::ZLIB ${PLATFORM_SPECIFIC_LIBS} ${CMAKE_DL_LIBS} secp256k1 bls12-381 xxhash
                                  ${security_framework} ${corefoundation_framework})
 
 add_subdirectory( test )

libraries/libfc/include/fc/crypto/xxh3.hpp

@@ -0,0 +1,62 @@
+#pragma once
+
+#include <fc/fwd.hpp>
+#include <fc/crypto/packhash.hpp>
+#include <fc/io/raw_fwd.hpp>
+
+#include <xxHash/xxhash.h>
+
+namespace fc {
+
+namespace detail {
+struct alignas(64) alignmeto64 {};
+
+template <typename T>
+concept ContiguousCharSource = requires(const T& obj) {
+   { obj.data() } -> std::convertible_to<const char*>;
+   { obj.size() } -> std::unsigned_integral;
+};
+}
+
+struct xxh3 : public add_packhash_to_hash<xxh3> {
+   xxh3() = default;
+   explicit xxh3(const uint64_t h) : _hash(h) {}
+
+   ///no data() since would that be confusing LE/BE? but that breaks common interface across hash types
+
+   static xxh3 hash(const char* d, uint32_t dlen);
+   static xxh3 hash(const std::string&);
+
+   template<typename T>
+   static xxh3 hash( const T& t ) {
+      return packhash(t);
+   }
+
+   static xxh3 hash_raw(const detail::ContiguousCharSource auto& r) {
+      return xxh3{XXH3_64bits(r.data(), r.size())};
+   }
+
+   class encoder {
+      public:
+         encoder();
+
+         void write( const char* d, uint32_t dlen );
+         void put( char c ) { write( &c, 1 ); }
+         void reset();
+         xxh3 result();
+
+      private:
+         struct impl;
+         //XXH3_state_t documents requirement of 64 byte alignment.
+         fc::fwd<impl,576,detail::alignmeto64> my;
+   };
+
+   friend auto operator<=>(const xxh3&, const xxh3&) = default;
+
+   uint64_t _hash = 0;
+};
+
+}
+
+FC_REFLECT(xxh3, (_hash))
+

libraries/libfc/libraries/xxhash/CMakeLists.txt

@@ -0,0 +1,5 @@
+add_library(xxhash INTERFACE)
+target_include_directories(xxhash INTERFACE .)
+target_compile_definitions(xxhash INTERFACE -DXXH_INLINE_ALL=1)
+
+install(FILES xxHash/xxhash.h DESTINATION "${CMAKE_INSTALL_FULL_INCLUDEDIR}/springxxhash/xxHash" COMPONENT dev EXCLUDE_FROM_ALL)

libraries/libfc/src/crypto/xxh3.cpp

@@ -0,0 +1,34 @@
+#include <fc/crypto/xxh3.hpp>
+#include <fc/fwd_impl.hpp>
+
+namespace fc {
+
+struct xxh3::encoder::impl {
+   XXH3_state_t ctx;
+};
+
+xxh3::encoder::encoder() {
+   reset();
+}
+void xxh3::encoder::reset() {
+   XXH3_64bits_reset(&my->ctx);
+}
+void xxh3::encoder::write(const char* d, uint32_t dlen) {
+   XXH3_64bits_update(&my->ctx, d, dlen);
+}
+xxh3 xxh3::encoder::result() {
+   xxh3 h;
+   h._hash = XXH3_64bits_digest(&my->ctx);
+   return h;
+}
+
+xxh3 xxh3::hash(const char* d, uint32_t dlen) {
+  encoder e;
+  e.write(d,dlen);
+  return e.result();
+}
+xxh3 xxh3::hash(const std::string& s) {
+  return hash(s.data(), s.size());
+}
+
+}

libraries/libfc/test/crypto/test_hash_functions.cpp

@@ -2,8 +2,11 @@
 
 #include <fc/crypto/hex.hpp>
 #include <fc/crypto/sha3.hpp>
+#include <fc/crypto/xxh3.hpp>
 #include <fc/utility.hpp>
 
+#include <span>
+
 using namespace fc;
 
 BOOST_AUTO_TEST_SUITE(hash_functions)
@@ -66,4 +69,34 @@ BOOST_AUTO_TEST_CASE(keccak256) try {
 
 } FC_LOG_AND_RETHROW();
 
+#define XSUM_U8 uint8_t
+#define XSUM_U32 uint32_t
+#define XSUM_U64 uint64_t
+typedef struct {
+   uint64_t low64;
+   uint64_t high64;
+} XXH128_hash_t;
+#include <xxHash/tests/sanity_test_vectors.h>
+
+BOOST_AUTO_TEST_CASE(xhh3) try {
+   //reconstruct test vector input from xxhash library's tests
+   char sanity_buffer[4096+64+1];
+
+   uint64_t byte_gen = 2654435761U;
+   for(unsigned i = 0; i < sizeof(sanity_buffer); ++i) {
+      sanity_buffer[i] = (char)(byte_gen>>56);
+      byte_gen *= 11400714785074694797ULL;
+   }
+
+   for(const XSUM_testdata64_t& t : XSUM_XXH3_testdata) {
+      //non-zero seed not exposed by fc's xxh3 interface
+      if(t.seed != 0)
+         continue;
+
+      BOOST_CHECK_EQUAL(t.Nresult, xxh3::hash(sanity_buffer, t.len)._hash);
+      BOOST_CHECK_EQUAL(t.Nresult, xxh3::hash_raw(std::span(sanity_buffer, t.len))._hash);
+   }
+
+} FC_LOG_AND_RETHROW();
+
 BOOST_AUTO_TEST_SUITE_END()