diff --git a/CHANGELOG.md b/CHANGELOG.md index 8ef9573d..c5849269 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,3 +1,12 @@ +# v2.8.0 +- **[#753](https://github.com/xmrig/xmrig/issues/753) Added new algorithm [CryptoNight variant 2](https://github.com/xmrig/xmrig/issues/753) for Monero fork, thanks [@SChernykh](https://github.com/SChernykh).** + - Added global and per thread option `"asm"` and and command line equivalent. +- **[#758](https://github.com/xmrig/xmrig/issues/758) Added SSL/TLS support for secure connections to pools.** + - Added per pool options `"tls"` and `"tls-fingerprint"` and command line equivalents. +- [#767](https://github.com/xmrig/xmrig/issues/767) Added config autosave feature, same with GPU miners. +- [#245](https://github.com/xmrig/xmrig-proxy/issues/245) Fixed API ID collision when run multiple miners on same machine. +- [#757](https://github.com/xmrig/xmrig/issues/757) Fixed send buffer overflow. + # v2.6.4 - [#700](https://github.com/xmrig/xmrig/issues/700) `cryptonight-lite/ipbc` replaced to `cryptonight-heavy/tube` for **Bittube (TUBE)**. - Added `cryptonight/rto` (cryptonight variant 1 with IPBC/TUBE mod) variant for **Arto (RTO)** coin. diff --git a/CMakeLists.txt b/CMakeLists.txt index e263808e..1becac5c 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -1,11 +1,14 @@ cmake_minimum_required(VERSION 2.8) project(xmrig) -option(WITH_LIBCPUID "Use Libcpuid" ON) -option(WITH_AEON "CryptoNight-Lite support" ON) -option(WITH_SUMO "CryptoNight-Heavy support" ON) -option(WITH_HTTPD "HTTP REST API" ON) -option(BUILD_STATIC "Build static binary" OFF) +option(WITH_LIBCPUID "Use Libcpuid" ON) +option(WITH_AEON "CryptoNight-Lite support" ON) +option(WITH_SUMO "CryptoNight-Heavy support" ON) +option(WITH_HTTPD "HTTP REST API" ON) +option(WITH_DEBUG_LOG "Enable debug log output" OFF) +option(WITH_TLS "Enable OpenSSL support" ON) +option(WITH_ASM "Enable ASM PoW implementations" ON) +option(BUILD_STATIC "Build static binary" OFF) include (CheckIncludeFile) include (cmake/cpu.cmake) @@ -18,6 +21,7 @@ set(HEADERS src/common/config/ConfigLoader.h src/common/config/ConfigWatcher.h src/common/Console.h + src/common/cpu/Cpu.h src/common/crypto/Algorithm.h src/common/crypto/keccak.h src/common/interfaces/IClientListener.h @@ -25,10 +29,12 @@ set(HEADERS src/common/interfaces/IConfigCreator.h src/common/interfaces/IConsoleListener.h src/common/interfaces/IControllerListener.h + src/common/interfaces/ICpuInfo.h src/common/interfaces/ILogBackend.h src/common/interfaces/IStrategy.h src/common/interfaces/IStrategyListener.h src/common/interfaces/IWatcherListener.h + src/common/log/BasicLog.h src/common/log/ConsoleLog.h src/common/log/FileLog.h src/common/log/Log.h @@ -46,7 +52,6 @@ set(HEADERS src/common/xmrig.h src/core/ConfigLoader_platform.h src/core/Controller.h - src/Cpu.h src/interfaces/IJobResultListener.h src/interfaces/IThread.h src/interfaces/IWorker.h @@ -94,6 +99,7 @@ set(SOURCES src/common/Console.cpp src/common/crypto/Algorithm.cpp src/common/crypto/keccak.cpp + src/common/log/BasicLog.cpp src/common/log/ConsoleLog.cpp src/common/log/FileLog.cpp src/common/log/Log.cpp @@ -131,7 +137,6 @@ if (WIN32) res/app.rc src/App_win.cpp src/common/Platform_win.cpp - src/Cpu_win.cpp src/Mem_win.cpp ) @@ -141,18 +146,16 @@ elseif (APPLE) set(SOURCES_OS src/App_unix.cpp src/common/Platform_mac.cpp - src/Cpu_mac.cpp src/Mem_unix.cpp ) else() set(SOURCES_OS src/App_unix.cpp src/common/Platform_unix.cpp - src/Cpu_unix.cpp src/Mem_unix.cpp ) - set(EXTRA_LIBS pthread rt) + set(EXTRA_LIBS pthread rt dl) endif() if (CMAKE_SYSTEM_NAME STREQUAL FreeBSD) @@ -168,7 +171,6 @@ endif() add_definitions(/D__STDC_FORMAT_MACROS) add_definitions(/DUNICODE) -#add_definitions(/DAPP_DEBUG) set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} "${CMAKE_SOURCE_DIR}/cmake") @@ -181,17 +183,21 @@ if (WITH_LIBCPUID) include_directories(src/3rdparty/libcpuid) set(CPUID_LIB cpuid) - set(SOURCES_CPUID src/Cpu.cpp) + set(SOURCES_CPUID src/core/cpu/AdvancedCpuInfo.h src/core/cpu/AdvancedCpuInfo.cpp src/core/cpu/Cpu.cpp) else() add_definitions(/DXMRIG_NO_LIBCPUID) + set(SOURCES_CPUID src/common/cpu/BasicCpuInfo.h src/common/cpu/Cpu.cpp) if (XMRIG_ARM) - set(SOURCES_CPUID src/Cpu_arm.cpp) + set(SOURCES_CPUID ${SOURCES_CPUID} src/common/cpu/BasicCpuInfo_arm.cpp) else() - set(SOURCES_CPUID src/Cpu_stub.cpp) + set(SOURCES_CPUID ${SOURCES_CPUID} src/common/cpu/BasicCpuInfo.cpp) endif() endif() +include(cmake/OpenSSL.cmake) +include(cmake/asm.cmake) + CHECK_INCLUDE_FILE (syslog.h HAVE_SYSLOG_H) if (HAVE_SYSLOG_H) add_definitions(/DHAVE_SYSLOG_H) @@ -231,6 +237,7 @@ if (WITH_HTTPD) message(FATAL_ERROR "microhttpd NOT found: use `-DWITH_HTTPD=OFF` to build without http deamon support") endif() else() + set(HTTPD_SOURCES "") set(MHD_LIBRARY "") add_definitions(/DXMRIG_NO_HTTPD) add_definitions(/DXMRIG_NO_API) @@ -244,5 +251,9 @@ if (BUILD_STATIC) set(CMAKE_EXE_LINKER_FLAGS " -static") endif() -add_executable(${PROJECT_NAME} ${HEADERS} ${SOURCES} ${SOURCES_OS} ${SOURCES_CPUID} ${HEADERS_CRYPTO} ${SOURCES_CRYPTO} ${SOURCES_SYSLOG} ${HTTPD_SOURCES}) -target_link_libraries(${PROJECT_NAME} ${UV_LIBRARIES} ${MHD_LIBRARY} ${EXTRA_LIBS} ${CPUID_LIB}) +if (WITH_DEBUG_LOG) + add_definitions(/DAPP_DEBUG) +endif() + +add_executable(${PROJECT_NAME} ${HEADERS} ${SOURCES} ${SOURCES_OS} ${SOURCES_CPUID} ${HEADERS_CRYPTO} ${SOURCES_CRYPTO} ${SOURCES_SYSLOG} ${HTTPD_SOURCES} ${TLS_SOURCES} ${XMRIG_ASM_SOURCES}) +target_link_libraries(${PROJECT_NAME} ${XMRIG_ASM_LIBRARY} ${OPENSSL_LIBRARIES} ${UV_LIBRARIES} ${MHD_LIBRARY} ${EXTRA_LIBS} ${CPUID_LIB}) diff --git a/README.md b/README.md index 6054b9f8..77b8f394 100644 --- a/README.md +++ b/README.md @@ -1,6 +1,6 @@ # XMRig -:warning: **If you mine Monero, Aeon, Sumokoin, Turtlecoin, Stellite, GRAFT, Haven Protocol, IPBC, [PLEASE READ](https://github.com/xmrig/xmrig/issues/482)!** :warning: +:warning: **[Monero will change PoW algorithm on October 18](https://github.com/xmrig/xmrig/issues/753), all miners and proxy should be updated to v2.8+** :warning: [![Github All Releases](https://img.shields.io/github/downloads/xmrig/xmrig/total.svg)](https://github.com/xmrig/xmrig/releases) [![GitHub release](https://img.shields.io/github/release/xmrig/xmrig/all.svg)](https://github.com/xmrig/xmrig/releases) @@ -87,11 +87,22 @@ Use [config.xmrig.com](https://config.xmrig.com/xmrig) to generate, edit or shar Also you can use configuration via config file, default **config.json**. You can load multiple config files and combine it with command line options. ## Algorithm variations -Since version 0.8.0. -* `--av=1` For CPUs with hardware AES. -* `--av=2` Lower power mode (double hash) of `1`. -* `--av=3` Software AES implementation. -* `--av=4` Lower power mode (double hash) of `3`. + +- `av` option used for automatic and simple threads mode (when you specify only threads count). +- For [advanced threads mode](https://github.com/xmrig/xmrig/issues/563) each thread configured individually and `av` option not used. + +| av | Hashes per round | Hardware AES | +|----|------------------|--------------| +| 1 | 1 (Single) | yes | +| 2 | 2 (Double) | yes | +| 3 | 1 (Single) | no | +| 4 | 2 (Double) | no | +| 5 | 3 (Triple) | yes | +| 6 | 4 (Quard) | yes | +| 7 | 5 (Penta) | yes | +| 8 | 3 (Triple) | no | +| 9 | 4 (Quard) | no | +| 10 | 5 (Penta) | no | ## Common Issues ### HUGE PAGES unavailable @@ -100,8 +111,7 @@ Since version 0.8.0. ## Other information * No HTTP support, only stratum protocol support. -* No TLS support. -* Default donation 5% (5 minutes in 100 minutes) can be reduced to 1% via command line option `--donate-level`. +* Default donation 5% (5 minutes in 100 minutes) can be reduced to 1% via option `donate-level`. ### CPU mining performance diff --git a/cmake/OpenSSL.cmake b/cmake/OpenSSL.cmake new file mode 100644 index 00000000..ed287e7e --- /dev/null +++ b/cmake/OpenSSL.cmake @@ -0,0 +1,23 @@ +if (WITH_TLS) + set(OPENSSL_ROOT_DIR ${XMRIG_DEPS}) + + if (WIN32) + set(OPENSSL_USE_STATIC_LIBS TRUE) + set(OPENSSL_MSVC_STATIC_RT TRUE) + + set(EXTRA_LIBS ${EXTRA_LIBS} Crypt32) + endif() + + find_package(OpenSSL) + + if (OPENSSL_FOUND) + set(TLS_SOURCES src/common/net/Tls.h src/common/net/Tls.cpp) + include_directories(${OPENSSL_INCLUDE_DIR}) + else() + message(FATAL_ERROR "OpenSSL NOT found: use `-DWITH_TLS=OFF` to build without TLS support") + endif() +else() + set(TLS_SOURCES "") + set(OPENSSL_LIBRARIES "") + add_definitions(/DXMRIG_NO_TLS) +endif() diff --git a/cmake/asm.cmake b/cmake/asm.cmake new file mode 100644 index 00000000..cb50f0d9 --- /dev/null +++ b/cmake/asm.cmake @@ -0,0 +1,27 @@ +if (WITH_ASM AND NOT XMRIG_ARM AND CMAKE_SIZEOF_VOID_P EQUAL 8) + set(XMRIG_ASM_LIBRARY "xmrig-asm") + + if (CMAKE_C_COMPILER_ID MATCHES MSVC) + enable_language(ASM_MASM) + set(XMRIG_ASM_FILE "src/crypto/asm/cnv2_main_loop.asm") + set_property(SOURCE ${XMRIG_ASM_FILE} PROPERTY ASM_MASM) + else() + enable_language(ASM) + + if (WIN32 AND CMAKE_C_COMPILER_ID MATCHES GNU) + set(XMRIG_ASM_FILE "src/crypto/asm/cnv2_main_loop_win.S") + else() + set(XMRIG_ASM_FILE "src/crypto/asm/cnv2_main_loop.S") + endif() + + set_property(SOURCE ${XMRIG_ASM_FILE} PROPERTY C) + endif() + + add_library(${XMRIG_ASM_LIBRARY} STATIC ${XMRIG_ASM_FILE}) + set(XMRIG_ASM_SOURCES src/crypto/Asm.h src/crypto/Asm.cpp) + set_property(TARGET ${XMRIG_ASM_LIBRARY} PROPERTY LINKER_LANGUAGE C) +else() + set(XMRIG_ASM_SOURCES "") + set(XMRIG_ASM_LIBRARY "") + add_definitions(/DXMRIG_NO_ASM) +endif() diff --git a/cmake/flags.cmake b/cmake/flags.cmake index 498f2165..8bc14804 100644 --- a/cmake/flags.cmake +++ b/cmake/flags.cmake @@ -15,7 +15,7 @@ if (CMAKE_CXX_COMPILER_ID MATCHES GNU) set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -Wall -Wno-strict-aliasing") set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} -Ofast") - set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall -fno-exceptions -fno-rtti") + set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall -fno-exceptions -fno-rtti -Wno-class-memaccess") set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -Ofast -s") if (XMRIG_ARMv8) diff --git a/src/3rdparty/libcpuid/asm-bits.c b/src/3rdparty/libcpuid/asm-bits.c index b8e32284..bfabd404 100644 --- a/src/3rdparty/libcpuid/asm-bits.c +++ b/src/3rdparty/libcpuid/asm-bits.c @@ -1,825 +1,836 @@ -/* - * Copyright 2008 Veselin Georgiev, - * anrieffNOSPAM @ mgail_DOT.com (convert to gmail) - * - * Redistribution and use in source and binary forms, with or without - * modification, are permitted provided that the following conditions - * are met: - * - * 1. Redistributions of source code must retain the above copyright - * notice, this list of conditions and the following disclaimer. - * 2. 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 AUTHOR ``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 AUTHOR 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. - */ - -#include "libcpuid.h" -#include "asm-bits.h" - -int cpuid_exists_by_eflags(void) -{ -#if defined(PLATFORM_X64) - return 1; /* CPUID is always present on the x86_64 */ -#elif defined(PLATFORM_X86) -# if defined(COMPILER_GCC) - int result; - __asm __volatile( - " pushfl\n" - " pop %%eax\n" - " mov %%eax, %%ecx\n" - " xor $0x200000, %%eax\n" - " push %%eax\n" - " popfl\n" - " pushfl\n" - " pop %%eax\n" - " xor %%ecx, %%eax\n" - " mov %%eax, %0\n" - " push %%ecx\n" - " popfl\n" - : "=m"(result) - : :"eax", "ecx", "memory"); - return (result != 0); -# elif defined(COMPILER_MICROSOFT) - int result; - __asm { - pushfd - pop eax - mov ecx, eax - xor eax, 0x200000 - push eax - popfd - pushfd - pop eax - xor eax, ecx - mov result, eax - push ecx - popfd - }; - return (result != 0); -# else - return 0; -# endif /* COMPILER_MICROSOFT */ -#else - return 0; -#endif /* PLATFORM_X86 */ -} - -#ifdef INLINE_ASM_SUPPORTED -/* - * with MSVC/AMD64, the exec_cpuid() and cpu_rdtsc() functions - * are implemented in separate .asm files. Otherwise, use inline assembly - */ -void exec_cpuid(uint32_t *regs) -{ -#ifdef COMPILER_GCC -# ifdef PLATFORM_X64 - __asm __volatile( - " mov %0, %%rdi\n" - - " push %%rbx\n" - " push %%rcx\n" - " push %%rdx\n" - - " mov (%%rdi), %%eax\n" - " mov 4(%%rdi), %%ebx\n" - " mov 8(%%rdi), %%ecx\n" - " mov 12(%%rdi), %%edx\n" - - " cpuid\n" - - " movl %%eax, (%%rdi)\n" - " movl %%ebx, 4(%%rdi)\n" - " movl %%ecx, 8(%%rdi)\n" - " movl %%edx, 12(%%rdi)\n" - " pop %%rdx\n" - " pop %%rcx\n" - " pop %%rbx\n" - : - :"m"(regs) - :"memory", "eax", "rdi" - ); -# else - __asm __volatile( - " mov %0, %%edi\n" - - " push %%ebx\n" - " push %%ecx\n" - " push %%edx\n" - - " mov (%%edi), %%eax\n" - " mov 4(%%edi), %%ebx\n" - " mov 8(%%edi), %%ecx\n" - " mov 12(%%edi), %%edx\n" - - " cpuid\n" - - " mov %%eax, (%%edi)\n" - " mov %%ebx, 4(%%edi)\n" - " mov %%ecx, 8(%%edi)\n" - " mov %%edx, 12(%%edi)\n" - " pop %%edx\n" - " pop %%ecx\n" - " pop %%ebx\n" - : - :"m"(regs) - :"memory", "eax", "edi" - ); -# endif /* COMPILER_GCC */ -#else -# ifdef COMPILER_MICROSOFT - __asm { - push ebx - push ecx - push edx - push edi - mov edi, regs - - mov eax, [edi] - mov ebx, [edi+4] - mov ecx, [edi+8] - mov edx, [edi+12] - - cpuid - - mov [edi], eax - mov [edi+4], ebx - mov [edi+8], ecx - mov [edi+12], edx - - pop edi - pop edx - pop ecx - pop ebx - } -# else -# error "Unsupported compiler" -# endif /* COMPILER_MICROSOFT */ -#endif -} -#endif /* INLINE_ASSEMBLY_SUPPORTED */ - -#ifdef INLINE_ASM_SUPPORTED -void cpu_rdtsc(uint64_t* result) -{ - uint32_t low_part, hi_part; -#ifdef COMPILER_GCC - __asm __volatile ( - " rdtsc\n" - " mov %%eax, %0\n" - " mov %%edx, %1\n" - :"=m"(low_part), "=m"(hi_part)::"memory", "eax", "edx" - ); -#else -# ifdef COMPILER_MICROSOFT - __asm { - rdtsc - mov low_part, eax - mov hi_part, edx - }; -# else -# error "Unsupported compiler" -# endif /* COMPILER_MICROSOFT */ -#endif /* COMPILER_GCC */ - *result = (uint64_t)low_part + (((uint64_t) hi_part) << 32); -} -#endif /* INLINE_ASM_SUPPORTED */ - -#ifdef INLINE_ASM_SUPPORTED -void busy_sse_loop(int cycles) -{ -#ifdef COMPILER_GCC -#ifndef __APPLE__ -# define XALIGN ".balign 16\n" -#else -# define XALIGN ".align 4\n" -#endif - __asm __volatile ( - " xorps %%xmm0, %%xmm0\n" - " xorps %%xmm1, %%xmm1\n" - " xorps %%xmm2, %%xmm2\n" - " xorps %%xmm3, %%xmm3\n" - " xorps %%xmm4, %%xmm4\n" - " xorps %%xmm5, %%xmm5\n" - " xorps %%xmm6, %%xmm6\n" - " xorps %%xmm7, %%xmm7\n" - XALIGN - /* ".bsLoop:\n" */ - "1:\n" - // 0: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - // 1: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - // 2: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - // 3: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - // 4: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - // 5: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - // 6: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - // 7: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - // 8: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - // 9: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //10: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //11: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //12: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //13: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //14: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //15: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //16: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //17: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //18: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //19: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //20: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //21: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //22: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //23: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //24: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //25: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //26: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //27: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //28: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //29: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //30: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - //31: - " addps %%xmm1, %%xmm0\n" - " addps %%xmm2, %%xmm1\n" - " addps %%xmm3, %%xmm2\n" - " addps %%xmm4, %%xmm3\n" - " addps %%xmm5, %%xmm4\n" - " addps %%xmm6, %%xmm5\n" - " addps %%xmm7, %%xmm6\n" - " addps %%xmm0, %%xmm7\n" - - " dec %%eax\n" - /* "jnz .bsLoop\n" */ - " jnz 1b\n" - ::"a"(cycles) - ); -#else -# ifdef COMPILER_MICROSOFT - __asm { - mov eax, cycles - xorps xmm0, xmm0 - xorps xmm1, xmm1 - xorps xmm2, xmm2 - xorps xmm3, xmm3 - xorps xmm4, xmm4 - xorps xmm5, xmm5 - xorps xmm6, xmm6 - xorps xmm7, xmm7 - //-- - align 16 -bsLoop: - // 0: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 1: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 2: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 3: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 4: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 5: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 6: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 7: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 8: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 9: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 10: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 11: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 12: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 13: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 14: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 15: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 16: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 17: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 18: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 19: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 20: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 21: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 22: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 23: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 24: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 25: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 26: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 27: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 28: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 29: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 30: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - // 31: - addps xmm0, xmm1 - addps xmm1, xmm2 - addps xmm2, xmm3 - addps xmm3, xmm4 - addps xmm4, xmm5 - addps xmm5, xmm6 - addps xmm6, xmm7 - addps xmm7, xmm0 - //---------------------- - dec eax - jnz bsLoop - } -# else -# error "Unsupported compiler" -# endif /* COMPILER_MICROSOFT */ -#endif /* COMPILER_GCC */ -} -#endif /* INLINE_ASSEMBLY_SUPPORTED */ +/* + * Copyright 2008 Veselin Georgiev, + * anrieffNOSPAM @ mgail_DOT.com (convert to gmail) + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. 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 AUTHOR ``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 AUTHOR 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. + */ + +#include "libcpuid.h" +#include "asm-bits.h" + +int cpuid_exists_by_eflags(void) +{ +#if defined(PLATFORM_X64) + return 1; /* CPUID is always present on the x86_64 */ +#elif defined(PLATFORM_X86) +# if defined(COMPILER_GCC) || defined(COMPILER_CLANG) + int result; + __asm __volatile( + " pushfl\n" + " pop %%eax\n" + " mov %%eax, %%ecx\n" + " xor $0x200000, %%eax\n" + " push %%eax\n" + " popfl\n" + " pushfl\n" + " pop %%eax\n" + " xor %%ecx, %%eax\n" + " mov %%eax, %0\n" + " push %%ecx\n" + " popfl\n" + : "=m"(result) + : :"eax", "ecx", "memory"); + return (result != 0); +# elif defined(COMPILER_MICROSOFT) + int result; + __asm { + pushfd + pop eax + mov ecx, eax + xor eax, 0x200000 + push eax + popfd + pushfd + pop eax + xor eax, ecx + mov result, eax + push ecx + popfd + }; + return (result != 0); +# else + return 0; +# endif /* COMPILER_MICROSOFT */ +#elif defined(PLATFORM_ARM) + return 0; +#else + return 0; +#endif /* PLATFORM_X86 */ +} + +#ifdef INLINE_ASM_SUPPORTED +/* + * with MSVC/AMD64, the exec_cpuid() and cpu_rdtsc() functions + * are implemented in separate .asm files. Otherwise, use inline assembly + */ +void exec_cpuid(uint32_t *regs) +{ +# if defined(COMPILER_GCC) || defined(COMPILER_CLANG) +# ifdef PLATFORM_X64 + __asm __volatile( + " mov %0, %%rdi\n" + + " push %%rbx\n" + " push %%rcx\n" + " push %%rdx\n" + + " mov (%%rdi), %%eax\n" + " mov 4(%%rdi), %%ebx\n" + " mov 8(%%rdi), %%ecx\n" + " mov 12(%%rdi), %%edx\n" + + " cpuid\n" + + " movl %%eax, (%%rdi)\n" + " movl %%ebx, 4(%%rdi)\n" + " movl %%ecx, 8(%%rdi)\n" + " movl %%edx, 12(%%rdi)\n" + " pop %%rdx\n" + " pop %%rcx\n" + " pop %%rbx\n" + : + :"m"(regs) + :"memory", "eax", "rdi" + ); +# elif defined(PLATFORM_X86) + __asm __volatile( + " mov %0, %%edi\n" + + " push %%ebx\n" + " push %%ecx\n" + " push %%edx\n" + + " mov (%%edi), %%eax\n" + " mov 4(%%edi), %%ebx\n" + " mov 8(%%edi), %%ecx\n" + " mov 12(%%edi), %%edx\n" + + " cpuid\n" + + " mov %%eax, (%%edi)\n" + " mov %%ebx, 4(%%edi)\n" + " mov %%ecx, 8(%%edi)\n" + " mov %%edx, 12(%%edi)\n" + " pop %%edx\n" + " pop %%ecx\n" + " pop %%ebx\n" + : + :"m"(regs) + :"memory", "eax", "edi" + ); +# elif defined(PLATFORM_ARM) +# endif /* COMPILER_GCC */ +#else +# ifdef COMPILER_MICROSOFT + __asm { + push ebx + push ecx + push edx + push edi + mov edi, regs + + mov eax, [edi] + mov ebx, [edi+4] + mov ecx, [edi+8] + mov edx, [edi+12] + + cpuid + + mov [edi], eax + mov [edi+4], ebx + mov [edi+8], ecx + mov [edi+12], edx + + pop edi + pop edx + pop ecx + pop ebx + } +# else +# error "Unsupported compiler" +# endif /* COMPILER_MICROSOFT */ +#endif +} +#endif /* INLINE_ASSEMBLY_SUPPORTED */ + +#ifdef INLINE_ASM_SUPPORTED +void cpu_rdtsc(uint64_t* result) +{ + uint32_t low_part, hi_part; +#if defined(COMPILER_GCC) || defined(COMPILER_CLANG) +#ifdef PLATFORM_ARM + low_part = 0; + hi_part = 0; +#else + __asm __volatile ( + " rdtsc\n" + " mov %%eax, %0\n" + " mov %%edx, %1\n" + :"=m"(low_part), "=m"(hi_part)::"memory", "eax", "edx" + ); +#endif +#else +# ifdef COMPILER_MICROSOFT + __asm { + rdtsc + mov low_part, eax + mov hi_part, edx + }; +# else +# error "Unsupported compiler" +# endif /* COMPILER_MICROSOFT */ +#endif /* COMPILER_GCC */ + *result = (uint64_t)low_part + (((uint64_t) hi_part) << 32); +} +#endif /* INLINE_ASM_SUPPORTED */ + +#ifdef INLINE_ASM_SUPPORTED +void busy_sse_loop(int cycles) +{ +# if defined(COMPILER_GCC) || defined(COMPILER_CLANG) +#ifndef __APPLE__ +# define XALIGN ".balign 16\n" +#else +# define XALIGN ".align 4\n" +#endif +#ifdef PLATFORM_ARM +#else + __asm __volatile ( + " xorps %%xmm0, %%xmm0\n" + " xorps %%xmm1, %%xmm1\n" + " xorps %%xmm2, %%xmm2\n" + " xorps %%xmm3, %%xmm3\n" + " xorps %%xmm4, %%xmm4\n" + " xorps %%xmm5, %%xmm5\n" + " xorps %%xmm6, %%xmm6\n" + " xorps %%xmm7, %%xmm7\n" + XALIGN + /* ".bsLoop:\n" */ + "1:\n" + // 0: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + // 1: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + // 2: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + // 3: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + // 4: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + // 5: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + // 6: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + // 7: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + // 8: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + // 9: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //10: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //11: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //12: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //13: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //14: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //15: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //16: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //17: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //18: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //19: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //20: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //21: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //22: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //23: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //24: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //25: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //26: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //27: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //28: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //29: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //30: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + //31: + " addps %%xmm1, %%xmm0\n" + " addps %%xmm2, %%xmm1\n" + " addps %%xmm3, %%xmm2\n" + " addps %%xmm4, %%xmm3\n" + " addps %%xmm5, %%xmm4\n" + " addps %%xmm6, %%xmm5\n" + " addps %%xmm7, %%xmm6\n" + " addps %%xmm0, %%xmm7\n" + + " dec %%eax\n" + /* "jnz .bsLoop\n" */ + " jnz 1b\n" + ::"a"(cycles) + ); +#endif +#else +# ifdef COMPILER_MICROSOFT + __asm { + mov eax, cycles + xorps xmm0, xmm0 + xorps xmm1, xmm1 + xorps xmm2, xmm2 + xorps xmm3, xmm3 + xorps xmm4, xmm4 + xorps xmm5, xmm5 + xorps xmm6, xmm6 + xorps xmm7, xmm7 + //-- + align 16 +bsLoop: + // 0: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 1: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 2: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 3: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 4: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 5: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 6: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 7: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 8: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 9: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 10: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 11: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 12: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 13: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 14: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 15: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 16: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 17: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 18: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 19: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 20: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 21: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 22: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 23: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 24: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 25: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 26: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 27: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 28: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 29: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 30: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + // 31: + addps xmm0, xmm1 + addps xmm1, xmm2 + addps xmm2, xmm3 + addps xmm3, xmm4 + addps xmm4, xmm5 + addps xmm5, xmm6 + addps xmm6, xmm7 + addps xmm7, xmm0 + //---------------------- + dec eax + jnz bsLoop + } +# else +# error "Unsupported compiler" +# endif /* COMPILER_MICROSOFT */ +#endif /* COMPILER_GCC */ +} +#endif /* INLINE_ASSEMBLY_SUPPORTED */ \ No newline at end of file diff --git a/src/3rdparty/libcpuid/asm-bits.h b/src/3rdparty/libcpuid/asm-bits.h index 3a03e11c..9049e2fe 100644 --- a/src/3rdparty/libcpuid/asm-bits.h +++ b/src/3rdparty/libcpuid/asm-bits.h @@ -1,53 +1,71 @@ -/* - * Copyright 2008 Veselin Georgiev, - * anrieffNOSPAM @ mgail_DOT.com (convert to gmail) - * - * Redistribution and use in source and binary forms, with or without - * modification, are permitted provided that the following conditions - * are met: - * - * 1. Redistributions of source code must retain the above copyright - * notice, this list of conditions and the following disclaimer. - * 2. 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 AUTHOR ``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 AUTHOR 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. - */ -#ifndef __ASM_BITS_H__ -#define __ASM_BITS_H__ -#include "libcpuid.h" - -/* Determine Compiler: */ -#if defined(_MSC_VER) -# define COMPILER_MICROSOFT -#elif defined(__GNUC__) -# define COMPILER_GCC -#endif - -/* Determine Platform */ -#if defined(__x86_64__) || defined(_M_AMD64) -# define PLATFORM_X64 -#elif defined(__i386__) || defined(_M_IX86) -# define PLATFORM_X86 -#endif - -/* Under Windows/AMD64 with MSVC, inline assembly isn't supported */ -#if (defined(COMPILER_GCC) && defined(PLATFORM_X64)) || defined(PLATFORM_X86) -# define INLINE_ASM_SUPPORTED -#endif - -int cpuid_exists_by_eflags(void); -void exec_cpuid(uint32_t *regs); -void busy_sse_loop(int cycles); - -#endif /* __ASM_BITS_H__ */ +/* + * Copyright 2008 Veselin Georgiev, + * anrieffNOSPAM @ mgail_DOT.com (convert to gmail) + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. 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 AUTHOR ``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 AUTHOR 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. + */ +#ifndef __ASM_BITS_H__ +#define __ASM_BITS_H__ +#include "libcpuid.h" + +/* Determine Compiler: */ +#if defined(_MSC_VER) +#if !defined(COMPILER_MICROSOFT) +# define COMPILER_MICROSOFT +#endif +#elif defined(__GNUC__) +#if !defined(COMPILER_GCC) +# define COMPILER_GCC +#endif +#elif defined(__clang__) +#if !defined(COMPILER_CLANG) +# define COMPILER_CLANG +#endif +#endif + +/* Determine Platform */ +#if defined(__x86_64__) || defined(_M_AMD64) +#if !defined(PLATFORM_X64) +# define PLATFORM_X64 +#endif +#elif defined(__i386__) || defined(_M_IX86) +#if !defined(PLATFORM_X86) +# define PLATFORM_X86 +#endif +#elif defined(__ARMEL__) +#if !defined(PLATFORM_ARM) +# define PLATFORM_ARM +#endif +#endif + +/* Under Windows/AMD64 with MSVC, inline assembly isn't supported */ +#if (((defined(COMPILER_GCC) || defined(COMPILER_CLANG))) && \ + (defined(PLATFORM_X64) || defined(PLATFORM_X86) || defined(PLATFORM_ARM))) || \ + (defined(COMPILER_MICROSOFT) && defined(PLATFORM_X86)) +# define INLINE_ASM_SUPPORTED +#endif + +int cpuid_exists_by_eflags(void); +void exec_cpuid(uint32_t *regs); +void busy_sse_loop(int cycles); + +#endif /* __ASM_BITS_H__ */ diff --git a/src/3rdparty/libcpuid/cpuid_main.c b/src/3rdparty/libcpuid/cpuid_main.c index f22c7dd6..61cb638d 100644 --- a/src/3rdparty/libcpuid/cpuid_main.c +++ b/src/3rdparty/libcpuid/cpuid_main.c @@ -221,42 +221,42 @@ static void load_features_common(struct cpu_raw_data_t* raw, struct cpu_id_t* da static cpu_vendor_t cpuid_vendor_identify(const uint32_t *raw_vendor, char *vendor_str) { - int i; - cpu_vendor_t vendor = VENDOR_UNKNOWN; - const struct { cpu_vendor_t vendor; char match[16]; } - matchtable[NUM_CPU_VENDORS] = { - /* source: http://www.sandpile.org/ia32/cpuid.htm */ - { VENDOR_INTEL , "GenuineIntel" }, - { VENDOR_AMD , "AuthenticAMD" }, - { VENDOR_CYRIX , "CyrixInstead" }, - { VENDOR_NEXGEN , "NexGenDriven" }, - { VENDOR_TRANSMETA , "GenuineTMx86" }, - { VENDOR_UMC , "UMC UMC UMC " }, - { VENDOR_CENTAUR , "CentaurHauls" }, - { VENDOR_RISE , "RiseRiseRise" }, - { VENDOR_SIS , "SiS SiS SiS " }, - { VENDOR_NSC , "Geode by NSC" }, - }; + int i; + cpu_vendor_t vendor = VENDOR_UNKNOWN; + const struct { cpu_vendor_t vendor; char match[16]; } + matchtable[NUM_CPU_VENDORS] = { + /* source: http://www.sandpile.org/ia32/cpuid.htm */ + { VENDOR_INTEL , "GenuineIntel" }, + { VENDOR_AMD , "AuthenticAMD" }, + { VENDOR_CYRIX , "CyrixInstead" }, + { VENDOR_NEXGEN , "NexGenDriven" }, + { VENDOR_TRANSMETA , "GenuineTMx86" }, + { VENDOR_UMC , "UMC UMC UMC " }, + { VENDOR_CENTAUR , "CentaurHauls" }, + { VENDOR_RISE , "RiseRiseRise" }, + { VENDOR_SIS , "SiS SiS SiS " }, + { VENDOR_NSC , "Geode by NSC" }, + }; - memcpy(vendor_str + 0, &raw_vendor[1], 4); - memcpy(vendor_str + 4, &raw_vendor[3], 4); - memcpy(vendor_str + 8, &raw_vendor[2], 4); - vendor_str[12] = 0; + memcpy(vendor_str + 0, &raw_vendor[1], 4); + memcpy(vendor_str + 4, &raw_vendor[3], 4); + memcpy(vendor_str + 8, &raw_vendor[2], 4); + vendor_str[12] = 0; - /* Determine vendor: */ - for (i = 0; i < NUM_CPU_VENDORS; i++) - if (!strcmp(vendor_str, matchtable[i].match)) { - vendor = matchtable[i].vendor; - break; - } - return vendor; + /* Determine vendor: */ + for (i = 0; i < NUM_CPU_VENDORS; i++) + if (!strcmp(vendor_str, matchtable[i].match)) { + vendor = matchtable[i].vendor; + break; + } + return vendor; } static int cpuid_basic_identify(struct cpu_raw_data_t* raw, struct cpu_id_t* data) { int i, j, basic, xmodel, xfamily, ext; char brandstr[64] = {0}; - data->vendor = cpuid_vendor_identify(raw->basic_cpuid[0], data->vendor_str); + data->vendor = cpuid_vendor_identify(raw->basic_cpuid[0], data->vendor_str); if (data->vendor == VENDOR_UNKNOWN) return set_error(ERR_CPU_UNKN); diff --git a/src/3rdparty/libcpuid/libcpuid.h b/src/3rdparty/libcpuid/libcpuid.h index c44990c3..847e5a4a 100644 --- a/src/3rdparty/libcpuid/libcpuid.h +++ b/src/3rdparty/libcpuid/libcpuid.h @@ -60,7 +60,7 @@ */ /** @mainpage A simple libcpuid introduction - * + * * LibCPUID provides CPU identification and access to the CPUID and RDTSC * instructions on the x86. *

@@ -82,6 +82,7 @@ */ /** @defgroup libcpuid LibCPUID + * @brief LibCPUID provides CPU identification @{ */ /* Include some integer type specifications: */ @@ -535,23 +536,23 @@ typedef enum { * @brief Describes common library error codes */ typedef enum { - ERR_OK = 0, /*!< "No error" */ - ERR_NO_CPUID = -1, /*!< "CPUID instruction is not supported" */ - ERR_NO_RDTSC = -2, /*!< "RDTSC instruction is not supported" */ - ERR_NO_MEM = -3, /*!< "Memory allocation failed" */ - ERR_OPEN = -4, /*!< "File open operation failed" */ - ERR_BADFMT = -5, /*!< "Bad file format" */ - ERR_NOT_IMP = -6, /*!< "Not implemented" */ - ERR_CPU_UNKN = -7, /*!< "Unsupported processor" */ - ERR_NO_RDMSR = -8, /*!< "RDMSR instruction is not supported" */ - ERR_NO_DRIVER= -9, /*!< "RDMSR driver error (generic)" */ - ERR_NO_PERMS = -10, /*!< "No permissions to install RDMSR driver" */ - ERR_EXTRACT = -11, /*!< "Cannot extract RDMSR driver (read only media?)" */ - ERR_HANDLE = -12, /*!< "Bad handle" */ - ERR_INVMSR = -13, /*!< "Invalid MSR" */ - ERR_INVCNB = -14, /*!< "Invalid core number" */ - ERR_HANDLE_R = -15, /*!< "Error on handle read" */ - ERR_INVRANGE = -16, /*!< "Invalid given range" */ + ERR_OK = 0, /*!< No error */ + ERR_NO_CPUID = -1, /*!< CPUID instruction is not supported */ + ERR_NO_RDTSC = -2, /*!< RDTSC instruction is not supported */ + ERR_NO_MEM = -3, /*!< Memory allocation failed */ + ERR_OPEN = -4, /*!< File open operation failed */ + ERR_BADFMT = -5, /*!< Bad file format */ + ERR_NOT_IMP = -6, /*!< Not implemented */ + ERR_CPU_UNKN = -7, /*!< Unsupported processor */ + ERR_NO_RDMSR = -8, /*!< RDMSR instruction is not supported */ + ERR_NO_DRIVER= -9, /*!< RDMSR driver error (generic) */ + ERR_NO_PERMS = -10, /*!< No permissions to install RDMSR driver */ + ERR_EXTRACT = -11, /*!< Cannot extract RDMSR driver (read only media?) */ + ERR_HANDLE = -12, /*!< Bad handle */ + ERR_INVMSR = -13, /*!< Invalid MSR */ + ERR_INVCNB = -14, /*!< Invalid core number */ + ERR_HANDLE_R = -15, /*!< Error on handle read */ + ERR_INVRANGE = -16, /*!< Invalid given range */ } cpu_error_t; /** @@ -668,7 +669,7 @@ struct cpu_epc_t cpuid_get_epc(int index, const struct cpu_raw_data_t* raw); const char* cpuid_lib_version(void); #ifdef __cplusplus -}; /* extern "C" */ +} /* extern "C" */ #endif diff --git a/src/3rdparty/libcpuid/libcpuid_internal.h b/src/3rdparty/libcpuid/libcpuid_internal.h index 038aa209..64804616 100644 --- a/src/3rdparty/libcpuid/libcpuid_internal.h +++ b/src/3rdparty/libcpuid/libcpuid_internal.h @@ -75,8 +75,9 @@ enum _intel_bits_t { _3 = LBIT( 14 ), _5 = LBIT( 15 ), _7 = LBIT( 16 ), - XEON_ = LBIT( 17 ), - ATOM_ = LBIT( 18 ), + _9 = LBIT( 17 ), + XEON_ = LBIT( 18 ), + ATOM_ = LBIT( 19 ), }; typedef enum _intel_bits_t intel_bits_t; diff --git a/src/3rdparty/libcpuid/libcpuid_types.h b/src/3rdparty/libcpuid/libcpuid_types.h index 9e897275..0e667aa6 100644 --- a/src/3rdparty/libcpuid/libcpuid_types.h +++ b/src/3rdparty/libcpuid/libcpuid_types.h @@ -32,6 +32,32 @@ #ifndef __LIBCPUID_TYPES_H__ #define __LIBCPUID_TYPES_H__ -#include +#if !defined(_MSC_VER) || _MSC_VER >= 1600 +# include +#else +/* we have to provide our own: */ +# if !defined(__int32_t_defined) +typedef int int32_t; +# endif + +# if !defined(__uint32_t_defined) +typedef unsigned uint32_t; +# endif + +typedef signed char int8_t; +typedef unsigned char uint8_t; +typedef signed short int16_t; +typedef unsigned short uint16_t; +#if (defined _MSC_VER) && (_MSC_VER <= 1300) + /* MSVC 6.0: no long longs ... */ + typedef signed __int64 int64_t; + typedef unsigned __int64 uint64_t; +#else + /* all other sane compilers: */ + typedef signed long long int64_t; + typedef unsigned long long uint64_t; +#endif + +#endif #endif /* __LIBCPUID_TYPES_H__ */ diff --git a/src/3rdparty/libcpuid/recog_amd.c b/src/3rdparty/libcpuid/recog_amd.c index 352d733b..4726f633 100644 --- a/src/3rdparty/libcpuid/recog_amd.c +++ b/src/3rdparty/libcpuid/recog_amd.c @@ -49,6 +49,10 @@ enum _amd_model_codes_t { _1400, _1500, _1600, + _1900, + _2400, + _2500, + _2700, }; static void load_amd_features(struct cpu_raw_data_t* raw, struct cpu_id_t* data) diff --git a/src/3rdparty/libcpuid/recog_intel.c b/src/3rdparty/libcpuid/recog_intel.c index 5467c19f..397d750e 100644 --- a/src/3rdparty/libcpuid/recog_intel.c +++ b/src/3rdparty/libcpuid/recog_intel.c @@ -376,7 +376,7 @@ static intel_code_and_bits_t get_brand_code_and_bits(struct cpu_id_t* data) bits |= bit_matchtable[i].bit; } - if ((i = match_pattern(bs, "Core(TM) [im][357]")) != 0) { + if ((i = match_pattern(bs, "Core(TM) [im][3579]")) != 0) { bits |= CORE_; i--; switch (bs[i + 9]) { @@ -387,6 +387,7 @@ static intel_code_and_bits_t get_brand_code_and_bits(struct cpu_id_t* data) case '3': bits |= _3; break; case '5': bits |= _5; break; case '7': bits |= _7; break; + case '9': bits |= _9; break; } } for (i = 0; i < COUNT_OF(matchtable); i++) diff --git a/src/App.cpp b/src/App.cpp index adcc5752..134e4ef5 100644 --- a/src/App.cpp +++ b/src/App.cpp @@ -29,11 +29,11 @@ #include "api/Api.h" #include "App.h" #include "common/Console.h" +#include "common/cpu/Cpu.h" #include "common/log/Log.h" #include "common/Platform.h" #include "core/Config.h" #include "core/Controller.h" -#include "Cpu.h" #include "crypto/CryptoNight.h" #include "Mem.h" #include "net/Network.h" diff --git a/src/Summary.cpp b/src/Summary.cpp index fe538fda..3c1d06a7 100644 --- a/src/Summary.cpp +++ b/src/Summary.cpp @@ -27,32 +27,31 @@ #include +#include "common/cpu/Cpu.h" #include "common/log/Log.h" #include "common/net/Pool.h" #include "core/Config.h" #include "core/Controller.h" -#include "Cpu.h" +#include "crypto/Asm.h" #include "Mem.h" #include "Summary.h" #include "version.h" -static void print_versions(xmrig::Config *config) +#ifndef XMRIG_NO_ASM +static const char *coloredAsmNames[] = { + "\x1B[1;31mnone\x1B[0m", + "auto", + "\x1B[1;32mintel\x1B[0m", + "\x1B[1;32mryzen\x1B[0m" +}; + + +inline static const char *asmName(xmrig::Assembly assembly, bool colors) { - char buf[16] = { 0 }; - -# if defined(__clang__) - snprintf(buf, 16, " clang/%d.%d.%d", __clang_major__, __clang_minor__, __clang_patchlevel__); -# elif defined(__GNUC__) - snprintf(buf, 16, " gcc/%d.%d.%d", __GNUC__, __GNUC_MINOR__, __GNUC_PATCHLEVEL__); -# elif defined(_MSC_VER) - snprintf(buf, 16, " MSVC/%d", MSVC_VERSION); -# endif - - Log::i()->text(config->isColors() ? GREEN_BOLD(" * ") WHITE_BOLD("%-13s") CYAN_BOLD("%s/%s") WHITE_BOLD(" libuv/%s%s") - : " * %-13s%s/%s libuv/%s%s", - "VERSIONS", APP_NAME, APP_VERSION, uv_version_string(), buf); + return colors ? coloredAsmNames[assembly] : xmrig::Asm::toString(assembly); } +#endif static void print_memory(xmrig::Config *config) { @@ -70,21 +69,23 @@ static void print_memory(xmrig::Config *config) { static void print_cpu(xmrig::Config *config) { + using namespace xmrig; + if (config->isColors()) { - Log::i()->text(GREEN_BOLD(" * ") WHITE_BOLD("%-13s") "%s (%d) %sx64 %sAES-NI", + Log::i()->text(GREEN_BOLD(" * ") WHITE_BOLD("%-13s%s (%d)") " %sx64 %sAES", "CPU", - Cpu::brand(), - Cpu::sockets(), - Cpu::isX64() ? "\x1B[1;32m" : "\x1B[1;31m-", - Cpu::hasAES() ? "\x1B[1;32m" : "\x1B[1;31m-"); + Cpu::info()->brand(), + Cpu::info()->sockets(), + Cpu::info()->isX64() ? "\x1B[1;32m" : "\x1B[1;31m-", + Cpu::info()->hasAES() ? "\x1B[1;32m" : "\x1B[1;31m-"); # ifndef XMRIG_NO_LIBCPUID - Log::i()->text(GREEN_BOLD(" * ") WHITE_BOLD("%-13s") "%.1f MB/%.1f MB", "CPU L2/L3", Cpu::l2() / 1024.0, Cpu::l3() / 1024.0); + Log::i()->text(GREEN_BOLD(" * ") WHITE_BOLD("%-13s%.1f MB/%.1f MB"), "CPU L2/L3", Cpu::info()->L2() / 1024.0, Cpu::info()->L3() / 1024.0); # endif } else { - Log::i()->text(" * %-13s%s (%d) %sx64 %sAES-NI", "CPU", Cpu::brand(), Cpu::sockets(), Cpu::isX64() ? "" : "-", Cpu::hasAES() ? "" : "-"); + Log::i()->text(" * %-13s%s (%d) %sx64 %sAES", "CPU", Cpu::info()->brand(), Cpu::info()->sockets(), Cpu::info()->isX64() ? "" : "-", Cpu::info()->hasAES() ? "" : "-"); # ifndef XMRIG_NO_LIBCPUID - Log::i()->text(" * %-13s%.1f MB/%.1f MB", "CPU L2/L3", Cpu::l2() / 1024.0, Cpu::l3() / 1024.0); + Log::i()->text(" * %-13s%.1f MB/%.1f MB", "CPU L2/L3", Cpu::info()->L2() / 1024.0, Cpu::info()->L3() / 1024.0); # endif } } @@ -117,45 +118,21 @@ static void print_threads(xmrig::Config *config) config->isColors() && config->donateLevel() == 0 ? "\x1B[1;31m" : "", config->donateLevel()); } -} +# ifndef XMRIG_NO_ASM + if (config->assembly() == xmrig::ASM_AUTO) { + const xmrig::Assembly assembly = xmrig::Cpu::info()->assembly(); -static void print_pools(xmrig::Config *config) -{ - const std::vector &pools = config->pools(); - - for (size_t i = 0; i < pools.size(); ++i) { - Log::i()->text(config->isColors() ? GREEN_BOLD(" * ") WHITE_BOLD("POOL #%-7zu") CYAN_BOLD("%s") " variant " WHITE_BOLD("%s") - : " * POOL #%-7d%s variant %s", - i + 1, - pools[i].url(), - pools[i].algorithm().variantName() - ); + Log::i()->text(config->isColors() ? GREEN_BOLD(" * ") WHITE_BOLD("%-13sauto:%s") + : " * %-13sauto:%s", "ASSEMBLY", asmName(assembly, config->isColors())); } - -# ifdef APP_DEBUG - for (const Pool &pool : pools) { - pool.print(); + else { + Log::i()->text(config->isColors() ? GREEN_BOLD(" * ") WHITE_BOLD("%-13s%s") : " * %-13s%s", "ASSEMBLY", asmName(config->assembly(), config->isColors())); } # endif } -#ifndef XMRIG_NO_API -static void print_api(xmrig::Config *config) -{ - const int port = config->apiPort(); - if (port == 0) { - return; - } - - Log::i()->text(config->isColors() ? GREEN_BOLD(" * ") WHITE_BOLD("%-13s") CYAN("%s:") CYAN_BOLD("%d") - : " * %-13s%s:%d", - "API BIND", config->isApiIPv6() ? "[::]" : "0.0.0.0", port); -} -#endif - - static void print_commands(xmrig::Config *config) { if (config->isColors()) { @@ -171,15 +148,12 @@ static void print_commands(xmrig::Config *config) void Summary::print(xmrig::Controller *controller) { - print_versions(controller->config()); + controller->config()->printVersions(); print_memory(controller->config()); print_cpu(controller->config()); print_threads(controller->config()); - print_pools(controller->config()); - -# ifndef XMRIG_NO_API - print_api(controller->config()); -# endif + controller->config()->printPools(); + controller->config()->printAPI(); print_commands(controller->config()); } diff --git a/src/api/ApiRouter.cpp b/src/api/ApiRouter.cpp index 07e425f1..dd7accf6 100644 --- a/src/api/ApiRouter.cpp +++ b/src/api/ApiRouter.cpp @@ -35,14 +35,13 @@ #include "api/ApiRouter.h" #include "common/api/HttpReply.h" #include "common/api/HttpRequest.h" +#include "common/cpu/Cpu.h" #include "common/crypto/keccak.h" #include "common/net/Job.h" #include "common/Platform.h" #include "core/Config.h" #include "core/Controller.h" -#include "Cpu.h" #include "interfaces/IThread.h" -#include "Mem.h" #include "rapidjson/document.h" #include "rapidjson/prettywriter.h" #include "rapidjson/stringbuffer.h" @@ -67,7 +66,7 @@ ApiRouter::ApiRouter(xmrig::Controller *controller) : memset(m_workerId, 0, sizeof(m_workerId)); setWorkerId(controller->config()->apiWorkerId()); - genId(); + genId(controller->config()->apiId()); } @@ -145,10 +144,15 @@ void ApiRouter::finalize(xmrig::HttpReply &reply, rapidjson::Document &doc) cons } -void ApiRouter::genId() +void ApiRouter::genId(const char *id) { memset(m_id, 0, sizeof(m_id)); + if (id && strlen(id) > 0) { + strncpy(m_id, id, sizeof(m_id) - 1); + return; + } + uv_interface_address_t *interfaces; int count = 0; @@ -160,11 +164,13 @@ void ApiRouter::genId() if (!interfaces[i].is_internal && interfaces[i].address.address4.sin_family == AF_INET) { uint8_t hash[200]; const size_t addrSize = sizeof(interfaces[i].phys_addr); - const size_t inSize = strlen(APP_KIND) + addrSize; + const size_t inSize = strlen(APP_KIND) + addrSize + sizeof(uint16_t); + const uint16_t port = static_cast(m_controller->config()->apiPort()); uint8_t *input = new uint8_t[inSize](); - memcpy(input, interfaces[i].phys_addr, addrSize); - memcpy(input + addrSize, APP_KIND, strlen(APP_KIND)); + memcpy(input, &port, sizeof(uint16_t)); + memcpy(input + sizeof(uint16_t), interfaces[i].phys_addr, addrSize); + memcpy(input + sizeof(uint16_t) + addrSize, APP_KIND, strlen(APP_KIND)); xmrig::keccak(input, inSize, hash); Job::toHex(hash, 8, m_id); @@ -232,13 +238,14 @@ void ApiRouter::getIdentify(rapidjson::Document &doc) const void ApiRouter::getMiner(rapidjson::Document &doc) const { + using namespace xmrig; auto &allocator = doc.GetAllocator(); rapidjson::Value cpu(rapidjson::kObjectType); - cpu.AddMember("brand", rapidjson::StringRef(Cpu::brand()), allocator); - cpu.AddMember("aes", Cpu::hasAES(), allocator); - cpu.AddMember("x64", Cpu::isX64(), allocator); - cpu.AddMember("sockets", Cpu::sockets(), allocator); + cpu.AddMember("brand", rapidjson::StringRef(Cpu::info()->brand()), allocator); + cpu.AddMember("aes", Cpu::info()->hasAES(), allocator); + cpu.AddMember("x64", Cpu::info()->isX64(), allocator); + cpu.AddMember("sockets", Cpu::info()->sockets(), allocator); doc.AddMember("version", APP_VERSION, allocator); doc.AddMember("kind", APP_KIND, allocator); diff --git a/src/api/ApiRouter.h b/src/api/ApiRouter.h index 9e32cdae..b781d5a2 100644 --- a/src/api/ApiRouter.h +++ b/src/api/ApiRouter.h @@ -56,7 +56,7 @@ protected: private: void finalize(xmrig::HttpReply &reply, rapidjson::Document &doc) const; - void genId(); + void genId(const char *id); void getConnection(rapidjson::Document &doc) const; void getHashrate(rapidjson::Document &doc) const; void getIdentify(rapidjson::Document &doc) const; @@ -66,7 +66,7 @@ private: void setWorkerId(const char *id); void updateWorkerId(const char *id, const char *previousId); - char m_id[17]; + char m_id[32]; char m_workerId[128]; NetworkState m_network; xmrig::Controller *m_controller; diff --git a/src/common/Platform.cpp b/src/common/Platform.cpp index 52b55987..a95f78e7 100644 --- a/src/common/Platform.cpp +++ b/src/common/Platform.cpp @@ -26,6 +26,12 @@ #include +#ifndef XMRIG_NO_TLS +# include +# include +#endif + + #include "Platform.h" @@ -61,3 +67,23 @@ const char *Platform::defaultConfigName() *m_defaultConfigName = '\0'; return nullptr; } + + +void Platform::init(const char *userAgent) +{ +# ifndef XMRIG_NO_TLS + SSL_library_init(); + SSL_load_error_strings(); + ERR_load_BIO_strings(); + ERR_load_crypto_strings(); + SSL_load_error_strings(); + OpenSSL_add_all_digests(); +# endif + + if (userAgent) { + m_userAgent = userAgent; + } + else { + m_userAgent = createUserAgent(); + } +} diff --git a/src/common/Platform.h b/src/common/Platform.h index 8704604a..5dfb9ff7 100644 --- a/src/common/Platform.h +++ b/src/common/Platform.h @@ -21,8 +21,8 @@ * along with this program. If not, see . */ -#ifndef __PLATFORM_H__ -#define __PLATFORM_H__ +#ifndef XMRIG_PLATFORM_H +#define XMRIG_PLATFORM_H #include @@ -43,9 +43,11 @@ public: static inline const char *userAgent() { return m_userAgent.data(); } private: + static char *createUserAgent(); + static char m_defaultConfigName[520]; static xmrig::c_str m_userAgent; }; -#endif /* __PLATFORM_H__ */ +#endif /* XMRIG_PLATFORM_H */ diff --git a/src/common/Platform_mac.cpp b/src/common/Platform_mac.cpp index b8181cc4..d0c533b0 100644 --- a/src/common/Platform_mac.cpp +++ b/src/common/Platform_mac.cpp @@ -38,7 +38,7 @@ #endif -static inline char *createUserAgent() +char *Platform::createUserAgent() { const size_t max = 160; @@ -65,17 +65,6 @@ bool Platform::setThreadAffinity(uint64_t cpu_id) } -void Platform::init(const char *userAgent) -{ - if (userAgent) { - m_userAgent = userAgent; - } - else { - m_userAgent = createUserAgent(); - } -} - - void Platform::setProcessPriority(int priority) { diff --git a/src/common/Platform_unix.cpp b/src/common/Platform_unix.cpp index 97b32ee8..058920ec 100644 --- a/src/common/Platform_unix.cpp +++ b/src/common/Platform_unix.cpp @@ -52,7 +52,7 @@ typedef cpuset_t cpu_set_t; #endif -static inline char *createUserAgent() +char *Platform::createUserAgent() { const size_t max = 160; @@ -92,23 +92,11 @@ bool Platform::setThreadAffinity(uint64_t cpu_id) } -void Platform::init(const char *userAgent) -{ - if (userAgent) { - m_userAgent = userAgent; - } - else { - m_userAgent = createUserAgent(); - } -} - - void Platform::setProcessPriority(int priority) { } - void Platform::setThreadPriority(int priority) { if (priority == -1) { diff --git a/src/common/Platform_win.cpp b/src/common/Platform_win.cpp index 47f41867..32b850d1 100644 --- a/src/common/Platform_win.cpp +++ b/src/common/Platform_win.cpp @@ -55,7 +55,7 @@ static inline OSVERSIONINFOEX winOsVersion() } -static inline char *createUserAgent() +char *Platform::createUserAgent() { const auto osver = winOsVersion(); const size_t max = 160; @@ -94,17 +94,6 @@ bool Platform::setThreadAffinity(uint64_t cpu_id) } -void Platform::init(const char *userAgent) -{ - if (userAgent) { - m_userAgent = userAgent; - } - else { - m_userAgent = createUserAgent(); - } -} - - void Platform::setProcessPriority(int priority) { if (priority == -1) { diff --git a/src/common/api/HttpRequest.cpp b/src/common/api/HttpRequest.cpp index 01245dfc..6898a385 100644 --- a/src/common/api/HttpRequest.cpp +++ b/src/common/api/HttpRequest.cpp @@ -147,7 +147,7 @@ int xmrig::HttpRequest::end(int status, MHD_Response *rsp) MHD_add_response_header(rsp, "Content-Type", "application/json"); MHD_add_response_header(rsp, "Access-Control-Allow-Origin", "*"); MHD_add_response_header(rsp, "Access-Control-Allow-Methods", "GET, PUT"); - MHD_add_response_header(rsp, "Access-Control-Allow-Headers", "Authorization"); + MHD_add_response_header(rsp, "Access-Control-Allow-Headers", "Authorization, Content-Type"); const int ret = MHD_queue_response(m_connection, status, rsp); MHD_destroy_response(rsp); diff --git a/src/common/config/CommonConfig.cpp b/src/common/config/CommonConfig.cpp index 7e43b39d..beb2d0c9 100644 --- a/src/common/config/CommonConfig.cpp +++ b/src/common/config/CommonConfig.cpp @@ -29,12 +29,37 @@ #include +#ifndef XMRIG_NO_HTTPD +# include +#endif + + +#ifndef XMRIG_NO_TLS +# include +#endif + + +#ifdef XMRIG_AMD_PROJECT +# if defined(__APPLE__) +# include +# else +# include "3rdparty/CL/cl.h" +# endif +#endif + + +#ifdef XMRIG_NVIDIA_PROJECT +# include "nvidia/cryptonight.h" +#endif + + #include "common/config/CommonConfig.h" #include "common/log/Log.h" #include "donate.h" #include "rapidjson/document.h" #include "rapidjson/filewritestream.h" #include "rapidjson/prettywriter.h" +#include "version.h" xmrig::CommonConfig::CommonConfig() : @@ -42,6 +67,7 @@ xmrig::CommonConfig::CommonConfig() : m_adjusted(false), m_apiIPv6(false), m_apiRestricted(true), + m_autoSave(true), m_background(false), m_colors(true), m_dryRun(false), @@ -69,8 +95,102 @@ xmrig::CommonConfig::CommonConfig() : } -xmrig::CommonConfig::~CommonConfig() +void xmrig::CommonConfig::printAPI() { +# ifndef XMRIG_NO_API + if (apiPort() == 0) { + return; + } + + Log::i()->text(isColors() ? GREEN_BOLD(" * ") WHITE_BOLD("%-13s") CYAN("%s:") CYAN_BOLD("%d") + : " * %-13s%s:%d", + "API BIND", isApiIPv6() ? "[::]" : "0.0.0.0", apiPort()); +# endif +} + + +void xmrig::CommonConfig::printPools() +{ + for (size_t i = 0; i < m_activePools.size(); ++i) { + if (!isColors()) { + Log::i()->text(" * POOL #%-7zu%s variant=%s, TLS=%d", + i + 1, + m_activePools[i].url(), + m_activePools[i].algorithm().variantName(), + static_cast(m_activePools[i].isTLS()) + ); + } + else { + Log::i()->text(GREEN_BOLD(" * ") WHITE_BOLD("POOL #%-7zu") "\x1B[1;%dm%s\x1B[0m variant " WHITE_BOLD("%s"), + i + 1, + m_activePools[i].isTLS() ? 32 : 36, + m_activePools[i].url(), + m_activePools[i].algorithm().variantName() + ); + } + } + +# ifdef APP_DEBUG + LOG_NOTICE("POOLS --------------------------------------------------------------------"); + for (const Pool &pool : m_activePools) { + pool.print(); + } + LOG_NOTICE("--------------------------------------------------------------------------"); +# endif +} + + +void xmrig::CommonConfig::printVersions() +{ + char buf[256] = { 0 }; + +# if defined(__clang__) + snprintf(buf, sizeof buf, "clang/%d.%d.%d", __clang_major__, __clang_minor__, __clang_patchlevel__); +# elif defined(__GNUC__) + snprintf(buf, sizeof buf, "gcc/%d.%d.%d", __GNUC__, __GNUC_MINOR__, __GNUC_PATCHLEVEL__); +# elif defined(_MSC_VER) + snprintf(buf, sizeof buf, "MSVC/%d", MSVC_VERSION); +# endif + + Log::i()->text(isColors() ? GREEN_BOLD(" * ") WHITE_BOLD("%-13s") CYAN_BOLD("%s/%s") WHITE_BOLD(" %s") + : " * %-13s%s/%s %s", + "ABOUT", APP_NAME, APP_VERSION, buf); + +# if defined(XMRIG_AMD_PROJECT) +# if CL_VERSION_2_0 + const char *ocl = "2.0"; +# elif CL_VERSION_1_2 + const char *ocl = "1.2"; +# elif CL_VERSION_1_1 + const char *ocl = "1.1"; +# elif CL_VERSION_1_0 + const char *ocl = "1.0"; +# else + const char *ocl = "0.0"; +# endif + int length = snprintf(buf, sizeof buf, "OpenCL/%s ", ocl); +# elif defined(XMRIG_NVIDIA_PROJECT) + const int cudaVersion = cuda_get_runtime_version(); + int length = snprintf(buf, sizeof buf, "CUDA/%d.%d ", cudaVersion / 1000, cudaVersion % 100); +# else + memset(buf, 0, 16); + int length = 0; +# endif + +# if !defined(XMRIG_NO_TLS) && defined(OPENSSL_VERSION_TEXT) + { + constexpr const char *v = OPENSSL_VERSION_TEXT + 8; + length += snprintf(buf + length, (sizeof buf) - length, "OpenSSL/%.*s ", static_cast(strchr(v, ' ') - v), v); + } +# endif + +# ifndef XMRIG_NO_HTTPD + length += snprintf(buf + length, (sizeof buf) - length, "microhttpd/%s ", MHD_get_version()); +# endif + + Log::i()->text(isColors() ? GREEN_BOLD(" * ") WHITE_BOLD("%-13slibuv/%s %s") + : " * %-13slibuv/%s %s", + "LIBS", uv_version_string(), buf); } @@ -98,7 +218,7 @@ bool xmrig::CommonConfig::save() rapidjson::PrettyWriter writer(os); doc.Accept(writer); - fclose(fp); + fflush(fp); uv_fs_close(uv_default_loop(), &req, fd, nullptr); uv_fs_req_cleanup(&req); @@ -126,6 +246,12 @@ bool xmrig::CommonConfig::finalize() pool.adjust(m_algorithm); if (pool.isValid() && pool.algorithm().isValid()) { +# ifdef XMRIG_NO_TLS + if (pool.isTLS()) { + continue; + } +# endif + m_activePools.push_back(std::move(pool)); } } @@ -157,6 +283,10 @@ bool xmrig::CommonConfig::parseBoolean(int key, bool enable) m_pools.back().setKeepAlive(enable ? Pool::kKeepAliveTimeout : 0); break; + case TlsKey: /* --tls */ + m_pools.back().setTLS(enable); + break; + # ifndef XMRIG_PROXY_PROJECT case NicehashKey: /* --nicehash */ m_pools.back().setNicehash(enable); @@ -179,10 +309,14 @@ bool xmrig::CommonConfig::parseBoolean(int key, bool enable) m_apiRestricted = enable; break; - case IConfig::DryRunKey: /* --dry-run */ + case DryRunKey: /* --dry-run */ m_dryRun = enable; break; + case AutoSaveKey: + m_autoSave = enable; + break; + default: break; } @@ -235,6 +369,10 @@ bool xmrig::CommonConfig::parseString(int key, const char *arg) m_pools.back().setRigId(arg); break; + case FingerprintKey: /* --tls-fingerprint */ + m_pools.back().setFingerprint(arg); + break; + case VariantKey: /* --variant */ m_pools.back().algorithm().parseVariant(arg); break; @@ -251,6 +389,10 @@ bool xmrig::CommonConfig::parseString(int key, const char *arg) m_apiWorkerId = arg; break; + case ApiIdKey: /* --api-id */ + m_apiId = arg; + break; + case UserAgentKey: /* --user-agent */ m_userAgent = arg; break; @@ -265,6 +407,7 @@ bool xmrig::CommonConfig::parseString(int key, const char *arg) case SyslogKey: /* --syslog */ case KeepAliveKey: /* --keepalive */ case NicehashKey: /* --nicehash */ + case TlsKey: /* --tls */ case ApiIPv6Key: /* --api-ipv6 */ case DryRunKey: /* --dry-run */ return parseBoolean(key, true); diff --git a/src/common/config/CommonConfig.h b/src/common/config/CommonConfig.h index ffebb6b2..422a6bb2 100644 --- a/src/common/config/CommonConfig.h +++ b/src/common/config/CommonConfig.h @@ -21,8 +21,8 @@ * along with this program. If not, see . */ -#ifndef __COMMONCONFIG_H__ -#define __COMMONCONFIG_H__ +#ifndef XMRIG_COMMONCONFIG_H +#define XMRIG_COMMONCONFIG_H #include @@ -41,14 +41,15 @@ class CommonConfig : public IConfig { public: CommonConfig(); - ~CommonConfig(); inline bool isApiIPv6() const { return m_apiIPv6; } inline bool isApiRestricted() const { return m_apiRestricted; } + inline bool isAutoSave() const { return m_autoSave; } inline bool isBackground() const { return m_background; } inline bool isColors() const { return m_colors; } inline bool isDryRun() const { return m_dryRun; } inline bool isSyslog() const { return m_syslog; } + inline const char *apiId() const { return m_apiId.data(); } inline const char *apiToken() const { return m_apiToken.data(); } inline const char *apiWorkerId() const { return m_apiWorkerId.data(); } inline const char *logFile() const { return m_logFile.data(); } @@ -67,6 +68,10 @@ public: bool save() override; + void printAPI(); + void printPools(); + void printVersions(); + protected: enum State { NoneState, @@ -84,6 +89,7 @@ protected: bool m_adjusted; bool m_apiIPv6; bool m_apiRestricted; + bool m_autoSave; bool m_background; bool m_colors; bool m_dryRun; @@ -97,6 +103,7 @@ protected: State m_state; std::vector m_activePools; std::vector m_pools; + xmrig::c_str m_apiId; xmrig::c_str m_apiToken; xmrig::c_str m_apiWorkerId; xmrig::c_str m_fileName; diff --git a/src/common/config/ConfigLoader.cpp b/src/common/config/ConfigLoader.cpp index cc5d9a49..484c2f8f 100644 --- a/src/common/config/ConfigLoader.cpp +++ b/src/common/config/ConfigLoader.cpp @@ -32,6 +32,11 @@ #endif +#ifndef XMRIG_NO_TLS +# include +#endif + + #include "common/config/ConfigLoader.h" #include "common/config/ConfigWatcher.h" #include "common/interfaces/IConfig.h" @@ -313,6 +318,13 @@ void xmrig::ConfigLoader::showVersion() printf("\nlibuv/%s\n", uv_version_string()); # ifndef XMRIG_NO_HTTPD - printf("libmicrohttpd/%s\n", MHD_get_version()); + printf("microhttpd/%s\n", MHD_get_version()); +# endif + +# if !defined(XMRIG_NO_TLS) && defined(OPENSSL_VERSION_TEXT) + { + constexpr const char *v = OPENSSL_VERSION_TEXT + 8; + printf("OpenSSL/%.*s\n", static_cast(strchr(v, ' ') - v), v); + } # endif } diff --git a/src/Cpu_stub.cpp b/src/common/cpu/BasicCpuInfo.cpp similarity index 77% rename from src/Cpu_stub.cpp rename to src/common/cpu/BasicCpuInfo.cpp index 8b27ad65..66af53cc 100644 --- a/src/Cpu_stub.cpp +++ b/src/common/cpu/BasicCpuInfo.cpp @@ -21,6 +21,9 @@ * along with this program. If not, see . */ +#include +#include + #ifdef _MSC_VER # include @@ -32,14 +35,8 @@ # define bit_AES (1 << 25) #endif -#ifndef bit_BMI2 -# define bit_BMI2 (1 << 8) -#endif -#include - - -#include "Cpu.h" +#include "common/cpu/BasicCpuInfo.h" #define VENDOR_ID (0) @@ -96,43 +93,18 @@ static inline bool has_aes_ni() } -static inline bool has_bmi2() { - int cpu_info[4] = { 0 }; - cpuid(EXTENDED_FEATURES, cpu_info); - - return (cpu_info[EBX_Reg] & bit_BMI2) != 0; -} - - -char Cpu::m_brand[64] = { 0 }; -int Cpu::m_flags = 0; -int Cpu::m_l2_cache = 0; -int Cpu::m_l3_cache = 0; -int Cpu::m_sockets = 1; -int Cpu::m_totalCores = 0; -size_t Cpu::m_totalThreads = 0; - - -size_t Cpu::optimalThreadsCount(size_t size, int maxCpuUsage) -{ - const size_t count = m_totalThreads / 2; - return count < 1 ? 1 : count; -} - - -void Cpu::initCommon() +xmrig::BasicCpuInfo::BasicCpuInfo() : + m_aes(has_aes_ni()), + m_brand(), + m_threads(std::thread::hardware_concurrency()) { cpu_brand_string(m_brand); - -# if defined(__x86_64__) || defined(_M_AMD64) - m_flags |= X86_64; -# endif - - if (has_aes_ni()) { - m_flags |= AES; - } - - if (has_bmi2()) { - m_flags |= BMI2; - } +} + + +size_t xmrig::BasicCpuInfo::optimalThreadsCount(size_t memSize, int maxCpuUsage) const +{ + const size_t count = threads() / 2; + + return count < 1 ? 1 : count; } diff --git a/src/common/cpu/BasicCpuInfo.h b/src/common/cpu/BasicCpuInfo.h new file mode 100644 index 00000000..f9d710d6 --- /dev/null +++ b/src/common/cpu/BasicCpuInfo.h @@ -0,0 +1,70 @@ +/* XMRig + * Copyright 2010 Jeff Garzik + * Copyright 2012-2014 pooler + * Copyright 2014 Lucas Jones + * Copyright 2014-2016 Wolf9466 + * Copyright 2016 Jay D Dee + * Copyright 2017-2018 XMR-Stak , + * Copyright 2016-2018 XMRig , + * + * + * 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 3 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, see . + */ + +#ifndef XMRIG_BASICCPUINFO_H +#define XMRIG_BASICCPUINFO_H + + +#include "common/interfaces/ICpuInfo.h" + + +namespace xmrig { + + +class BasicCpuInfo : public ICpuInfo +{ +public: + BasicCpuInfo(); + +protected: + size_t optimalThreadsCount(size_t memSize, int maxCpuUsage) const override; + + inline Assembly assembly() const override { return ASM_NONE; } + inline bool hasAES() const override { return m_aes; } + inline bool isSupported() const override { return true; } + inline const char *brand() const override { return m_brand; } + inline int32_t cores() const override { return -1; } + inline int32_t L2() const override { return -1; } + inline int32_t L3() const override { return -1; } + inline int32_t nodes() const override { return -1; } + inline int32_t sockets() const override { return 1; } + inline int32_t threads() const override { return m_threads; } + +# if defined(__x86_64__) || defined(_M_AMD64) || defined (__arm64__) || defined (__aarch64__) + inline bool isX64() const override { return true; } +# else + inline bool isX64() const override { return false; } +# endif + +private: + bool m_aes; + char m_brand[64]; + int32_t m_threads; +}; + + +} /* namespace xmrig */ + + +#endif /* XMRIG_BASICCPUINFO_H */ diff --git a/src/Cpu_arm.cpp b/src/common/cpu/BasicCpuInfo_arm.cpp similarity index 73% rename from src/Cpu_arm.cpp rename to src/common/cpu/BasicCpuInfo_arm.cpp index 3e259d0d..c1c127db 100644 --- a/src/Cpu_arm.cpp +++ b/src/common/cpu/BasicCpuInfo_arm.cpp @@ -21,37 +21,27 @@ * along with this program. If not, see . */ - #include +#include -#include "Cpu.h" +#include "common/cpu/BasicCpuInfo.h" -char Cpu::m_brand[64] = { 0 }; -int Cpu::m_flags = 0; -int Cpu::m_l2_cache = 0; -int Cpu::m_l3_cache = 0; -int Cpu::m_sockets = 1; -int Cpu::m_totalCores = 0; -size_t Cpu::m_totalThreads = 0; - - -size_t Cpu::optimalThreadsCount(size_t size, int maxCpuUsage) -{ - return m_totalThreads; -} - - -void Cpu::initCommon() +xmrig::BasicCpuInfo::BasicCpuInfo() : + m_aes(false), + m_brand(), + m_threads(std::thread::hardware_concurrency()) { memcpy(m_brand, "Unknown", 7); -# if defined (__arm64__) || defined (__aarch64__) - m_flags |= X86_64; -# endif - # if __ARM_FEATURE_CRYPTO - m_flags |= AES; + m_aes = true; # endif } + + +size_t xmrig::BasicCpuInfo::optimalThreadsCount(size_t memSize, int maxCpuUsage) const +{ + return threads(); +} diff --git a/src/Cpu_unix.cpp b/src/common/cpu/Cpu.cpp similarity index 71% rename from src/Cpu_unix.cpp rename to src/common/cpu/Cpu.cpp index b895c897..b1bb28ac 100644 --- a/src/Cpu_unix.cpp +++ b/src/common/cpu/Cpu.cpp @@ -22,33 +22,36 @@ */ -#ifdef __FreeBSD__ -# include -# include -# include -# include -#endif +#include -#include -#include -#include -#include +#include "common/cpu/BasicCpuInfo.h" +#include "common/cpu/Cpu.h" -#include "Cpu.h" +static xmrig::ICpuInfo *cpuInfo = nullptr; -#ifdef __FreeBSD__ -typedef cpuset_t cpu_set_t; -#endif - - -void Cpu::init() +xmrig::ICpuInfo *xmrig::Cpu::info() { -# ifdef XMRIG_NO_LIBCPUID - m_totalThreads = sysconf(_SC_NPROCESSORS_CONF); -# endif + assert(cpuInfo != nullptr); - initCommon(); + return cpuInfo; +} + + +void xmrig::Cpu::init() +{ + assert(cpuInfo == nullptr); + + cpuInfo = new BasicCpuInfo(); +} + + +void xmrig::Cpu::release() +{ + assert(cpuInfo != nullptr); + + delete cpuInfo; + cpuInfo = nullptr; } diff --git a/src/Cpu_mac.cpp b/src/common/cpu/Cpu.h similarity index 75% rename from src/Cpu_mac.cpp rename to src/common/cpu/Cpu.h index 085148bc..1d5a9fb1 100644 --- a/src/Cpu_mac.cpp +++ b/src/common/cpu/Cpu.h @@ -4,7 +4,7 @@ * Copyright 2014 Lucas Jones * Copyright 2014-2016 Wolf9466 * Copyright 2016 Jay D Dee - * Copyright 2016 Jay D Dee + * Copyright 2017-2018 XMR-Stak , * Copyright 2016-2018 XMRig , * * This program is free software: you can redistribute it and/or modify @@ -21,20 +21,26 @@ * along with this program. If not, see . */ - -#include -#include -#include +#ifndef XMRIG_CPU_H +#define XMRIG_CPU_H -#include "Cpu.h" +#include "common/interfaces/ICpuInfo.h" -void Cpu::init() +namespace xmrig { + + +class Cpu { -# ifdef XMRIG_NO_LIBCPUID - m_totalThreads = sysconf(_SC_NPROCESSORS_CONF); -# endif +public: + static ICpuInfo *info(); + static void init(); + static void release(); +}; - initCommon(); -} + +} /* namespace xmrig */ + + +#endif /* XMRIG_CPU_H */ diff --git a/src/common/crypto/Algorithm.cpp b/src/common/crypto/Algorithm.cpp index 31035fb1..a3cf48b2 100644 --- a/src/common/crypto/Algorithm.cpp +++ b/src/common/crypto/Algorithm.cpp @@ -6,6 +6,7 @@ * Copyright 2016 Jay D Dee * Copyright 2017-2018 XMR-Stak , * Copyright 2018 Lee Clagett + * Copyright 2018 SChernykh * Copyright 2016-2018 XMRig , * * This program is free software: you can redistribute it and/or modify @@ -60,6 +61,7 @@ static AlgoData const algorithms[] = { { "cryptonight/msr", "cn/msr", xmrig::CRYPTONIGHT, xmrig::VARIANT_MSR }, { "cryptonight/xao", "cn/xao", xmrig::CRYPTONIGHT, xmrig::VARIANT_XAO }, { "cryptonight/rto", "cn/rto", xmrig::CRYPTONIGHT, xmrig::VARIANT_RTO }, + { "cryptonight/2", "cn/2", xmrig::CRYPTONIGHT, xmrig::VARIANT_2 }, # ifndef XMRIG_NO_AEON { "cryptonight-lite", "cn-lite", xmrig::CRYPTONIGHT_LITE, xmrig::VARIANT_AUTO }, @@ -81,6 +83,8 @@ static AlgoData const algorithms[] = { static AlgoData const xmrStakAlgorithms[] = { { "cryptonight-monerov7", nullptr, xmrig::CRYPTONIGHT, xmrig::VARIANT_1 }, { "cryptonight_v7", nullptr, xmrig::CRYPTONIGHT, xmrig::VARIANT_1 }, + { "cryptonight-monerov8", nullptr, xmrig::CRYPTONIGHT, xmrig::VARIANT_2 }, + { "cryptonight_v8", nullptr, xmrig::CRYPTONIGHT, xmrig::VARIANT_2 }, { "cryptonight_v7_stellite", nullptr, xmrig::CRYPTONIGHT, xmrig::VARIANT_XTL }, { "cryptonight_lite", nullptr, xmrig::CRYPTONIGHT_LITE, xmrig::VARIANT_0 }, { "cryptonight-aeonv7", nullptr, xmrig::CRYPTONIGHT_LITE, xmrig::VARIANT_1 }, @@ -103,7 +107,8 @@ static const char *variants[] = { "msr", "xhv", "xao", - "rto" + "rto", + "2", }; @@ -172,11 +177,21 @@ void xmrig::Algorithm::parseVariant(const char *variant) void xmrig::Algorithm::parseVariant(int variant) { - if (variant >= VARIANT_AUTO && variant < VARIANT_MAX) { - m_variant = static_cast(variant); - } - else { - assert(false); + assert(variant >= -1 && variant <= 2); + + switch (variant) { + case -1: + case 0: + case 1: + m_variant = static_cast(variant); + break; + + case 2: + m_variant = VARIANT_2; + break; + + default: + break; } } diff --git a/src/common/crypto/Algorithm.h b/src/common/crypto/Algorithm.h index bcf029d8..731fa793 100644 --- a/src/common/crypto/Algorithm.h +++ b/src/common/crypto/Algorithm.h @@ -6,6 +6,7 @@ * Copyright 2016 Jay D Dee * Copyright 2017-2018 XMR-Stak , * Copyright 2018 Lee Clagett + * Copyright 2018 SChernykh * Copyright 2016-2018 XMRig , * * This program is free software: you can redistribute it and/or modify @@ -22,8 +23,8 @@ * along with this program. If not, see . */ -#ifndef __ALGORITHM_H__ -#define __ALGORITHM_H__ +#ifndef XMRIG_ALGORITHM_H +#define XMRIG_ALGORITHM_H #include diff --git a/src/common/crypto/keccak.h b/src/common/crypto/keccak.h index 0413ec2d..6121044a 100644 --- a/src/common/crypto/keccak.h +++ b/src/common/crypto/keccak.h @@ -23,8 +23,8 @@ */ -#ifndef KECCAK_H_ -#define KECCAK_H_ +#ifndef XMRIG_KECCAK_H +#define XMRIG_KECCAK_H #include #include @@ -41,9 +41,15 @@ inline void keccak(const uint8_t *in, size_t inlen, uint8_t *md) keccak(in, static_cast(inlen), md, 200); } + +inline void keccak(const char *in, size_t inlen, uint8_t *md) +{ + keccak(reinterpret_cast(in), static_cast(inlen), md, 200); +} + // update the state void keccakf(uint64_t st[25], int norounds); } /* namespace xmrig */ -#endif /* KECCAK_H_ */ +#endif /* XMRIG_KECCAK_H */ diff --git a/src/common/interfaces/IConfig.h b/src/common/interfaces/IConfig.h index 62c7ba94..69f2ffab 100644 --- a/src/common/interfaces/IConfig.h +++ b/src/common/interfaces/IConfig.h @@ -20,8 +20,8 @@ * along with this program. If not, see . */ -#ifndef __ICONFIG_H__ -#define __ICONFIG_H__ +#ifndef XMRIG_ICONFIG_H +#define XMRIG_ICONFIG_H #include "common/crypto/Algorithm.h" @@ -42,6 +42,7 @@ public: ApiPort = 4000, ApiRestrictedKey = 4004, ApiWorkerIdKey = 4002, + ApiIdKey = 4005, BackgroundKey = 'B', ColorKey = 1002, ConfigKey = 'c', @@ -62,6 +63,9 @@ public: VerboseKey = 1100, VersionKey = 'V', WatchKey = 1105, + TlsKey = 1013, + FingerprintKey = 1014, + AutoSaveKey = 1016, // xmrig common CPUPriorityKey = 1021, @@ -77,12 +81,20 @@ public: SafeKey = 1005, ThreadsKey = 't', HardwareAESKey = 1011, + AssemblyKey = 1015, // xmrig amd - OclPlatform = 1400, - OclAffinity = 1401, - OclDevices = 1402, - OclLaunch = 1403, + OclPlatformKey = 1400, + OclAffinityKey = 1401, + OclDevicesKey = 1402, + OclLaunchKey = 1403, + OclCacheKey = 1404, + OclPrintKey = 1405, + OclLoaderKey = 1406, + OclSridedIndexKey = 1407, + OclMemChunkKey = 1408, + OclUnrollKey = 1409, + OclCompModeKey = 1410, // xmrig-proxy AccessLogFileKey = 'A', @@ -94,6 +106,16 @@ public: PoolCoinKey = 'C', ReuseTimeoutKey = 1106, WorkersKey = 1103, + WorkersAdvKey = 1107, + + // xmrig nvidia + CudaMaxThreadsKey = 1200, + CudaBFactorKey = 1201, + CudaBSleepKey = 1202, + CudaDevicesKey = 1203, + CudaLaunchKey = 1204, + CudaAffinityKey = 1205, + CudaMaxUsageKey = 1206, }; virtual ~IConfig() {} @@ -115,4 +137,4 @@ public: } /* namespace xmrig */ -#endif // __ICONFIG_H__ +#endif // XMRIG_ICONFIG_H diff --git a/src/common/interfaces/IControllerListener.h b/src/common/interfaces/IControllerListener.h index d6077138..35249bcd 100644 --- a/src/common/interfaces/IControllerListener.h +++ b/src/common/interfaces/IControllerListener.h @@ -21,8 +21,8 @@ * along with this program. If not, see . */ -#ifndef __ICONTROLLERLISTENER_H__ -#define __ICONTROLLERLISTENER_H__ +#ifndef XMRIG_ICONTROLLERLISTENER_H +#define XMRIG_ICONTROLLERLISTENER_H namespace xmrig { @@ -43,4 +43,4 @@ public: } /* namespace xmrig */ -#endif // __ICONTROLLERLISTENER_H__ +#endif // XMRIG_ICONTROLLERLISTENER_H diff --git a/src/common/interfaces/ICpuInfo.h b/src/common/interfaces/ICpuInfo.h new file mode 100644 index 00000000..267616d0 --- /dev/null +++ b/src/common/interfaces/ICpuInfo.h @@ -0,0 +1,60 @@ +/* XMRig + * Copyright 2010 Jeff Garzik + * Copyright 2012-2014 pooler + * Copyright 2014 Lucas Jones + * Copyright 2014-2016 Wolf9466 + * Copyright 2016 Jay D Dee + * Copyright 2016-2018 XMRig + * + * 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 3 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, see . + */ + +#ifndef XMRIG_CPUINFO_H +#define XMRIG_CPUINFO_H + + +#include +#include + + +#include "common/xmrig.h" + + +namespace xmrig { + + +class ICpuInfo +{ +public: + virtual ~ICpuInfo() {} + + virtual bool hasAES() const = 0; + virtual bool isSupported() const = 0; + virtual bool isX64() const = 0; + virtual const char *brand() const = 0; + virtual int32_t cores() const = 0; + virtual int32_t L2() const = 0; + virtual int32_t L3() const = 0; + virtual int32_t nodes() const = 0; + virtual int32_t sockets() const = 0; + virtual int32_t threads() const = 0; + virtual size_t optimalThreadsCount(size_t memSize, int maxCpuUsage) const = 0; + virtual xmrig::Assembly assembly() const = 0; +}; + + +} /* namespace xmrig */ + + +#endif // XMRIG_CPUINFO_H diff --git a/src/Cpu.h b/src/common/log/BasicLog.cpp similarity index 50% rename from src/Cpu.h rename to src/common/log/BasicLog.cpp index a125bae8..cb4defcd 100644 --- a/src/Cpu.h +++ b/src/common/log/BasicLog.cpp @@ -21,46 +21,69 @@ * along with this program. If not, see . */ -#ifndef __CPU_H__ -#define __CPU_H__ + +#include +#include +#include +#include +#include + +#ifdef WIN32 +# include +# include +#endif -#include +#include "common/log/BasicLog.h" +#include "common/log/Log.h" -class Cpu +BasicLog::BasicLog() { -public: - enum Flags { - X86_64 = 1, - AES = 2, - BMI2 = 4 - }; - - static size_t optimalThreadsCount(size_t size, int maxCpuUsage); - static void init(); - - static inline bool hasAES() { return (m_flags & AES) != 0; } - static inline bool isX64() { return (m_flags & X86_64) != 0; } - static inline const char *brand() { return m_brand; } - static inline int cores() { return m_totalCores; } - static inline int l2() { return m_l2_cache; } - static inline int l3() { return m_l3_cache; } - static inline int sockets() { return m_sockets; } - static inline int threads() { return m_totalThreads; } - -private: - static void initCommon(); - - static bool m_l2_exclusive; - static char m_brand[64]; - static int m_flags; - static int m_l2_cache; - static int m_l3_cache; - static int m_sockets; - static int m_totalCores; - static size_t m_totalThreads; -}; +} -#endif /* __CPU_H__ */ +void BasicLog::message(Level level, const char* fmt, va_list args) +{ + time_t now = time(nullptr); + tm stime; + +# ifdef _WIN32 + localtime_s(&stime, &now); +# else + localtime_r(&now, &stime); +# endif + + snprintf(m_fmt, sizeof(m_fmt) - 1, "[%d-%02d-%02d %02d:%02d:%02d]%s %s%s", + stime.tm_year + 1900, + stime.tm_mon + 1, + stime.tm_mday, + stime.tm_hour, + stime.tm_min, + stime.tm_sec, + Log::colorByLevel(level, false), + fmt, + Log::endl(false) + ); + + print(args); +} + + +void BasicLog::text(const char* fmt, va_list args) +{ + snprintf(m_fmt, sizeof(m_fmt) - 1, "%s%s", fmt, Log::endl(false)); + + print(args); +} + + +void BasicLog::print(va_list args) +{ + if (vsnprintf(m_buf, sizeof(m_buf) - 1, m_fmt, args) <= 0) { + return; + } + + fputs(m_buf, stdout); + fflush(stdout); +} diff --git a/src/common/log/BasicLog.h b/src/common/log/BasicLog.h new file mode 100644 index 00000000..523538e9 --- /dev/null +++ b/src/common/log/BasicLog.h @@ -0,0 +1,55 @@ +/* XMRig + * Copyright 2010 Jeff Garzik + * Copyright 2012-2014 pooler + * Copyright 2014 Lucas Jones + * Copyright 2014-2016 Wolf9466 + * Copyright 2016 Jay D Dee + * Copyright 2017-2018 XMR-Stak , + * Copyright 2016-2018 XMRig , + * + * 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 3 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, see . + */ + +#ifndef __BASICLOG_H__ +#define __BASICLOG_H__ + + +#include + + +#include "common/interfaces/ILogBackend.h" + + +namespace xmrig { + class Controller; +} + + +class BasicLog : public ILogBackend +{ +public: + BasicLog(); + + void message(Level level, const char *fmt, va_list args) override; + void text(const char *fmt, va_list args) override; + +private: + bool isWritable() const; + void print(va_list args); + + char m_buf[kBufferSize]; + char m_fmt[256]; +}; + +#endif /* __BASICLOG_H__ */ diff --git a/src/common/log/Log.cpp b/src/common/log/Log.cpp index ccf38008..2af90209 100644 --- a/src/common/log/Log.cpp +++ b/src/common/log/Log.cpp @@ -30,6 +30,7 @@ #include "common/interfaces/ILogBackend.h" +#include "common/log/BasicLog.h" #include "common/log/Log.h" @@ -109,6 +110,14 @@ const char *Log::endl(bool isColors) } +void Log::defaultInit() +{ + m_self = new Log(); + + add(new BasicLog()); +} + + Log::~Log() { for (auto backend : m_backends) { diff --git a/src/common/log/Log.h b/src/common/log/Log.h index bfa30717..2774ae0c 100644 --- a/src/common/log/Log.h +++ b/src/common/log/Log.h @@ -36,7 +36,7 @@ class Log { public: - static inline Log* i() { assert(m_self != nullptr); return m_self; } + static inline Log* i() { if (!m_self) { defaultInit(); } return m_self; } static inline void add(ILogBackend *backend) { i()->m_backends.push_back(backend); } static inline void init() { if (!m_self) { new Log(); } } static inline void release() { assert(m_self != nullptr); delete m_self; } @@ -46,6 +46,7 @@ public: static const char *colorByLevel(ILogBackend::Level level, bool isColors = true); static const char *endl(bool isColors = true); + static void defaultInit(); private: inline Log() { @@ -68,6 +69,8 @@ private: #define RED(x) "\x1B[0;31m" x "\x1B[0m" #define GREEN_BOLD(x) "\x1B[1;32m" x "\x1B[0m" #define GREEN(x) "\x1B[0;32m" x "\x1B[0m" +#define YELLOW(x) "\x1B[0;33m" x "\x1B[0m" +#define YELLOW_BOLD(x) "\x1B[1;33m" x "\x1B[0m" #define MAGENTA_BOLD(x) "\x1B[1;35m" x "\x1B[0m" #define MAGENTA(x) "\x1B[0;35m" x "\x1B[0m" #define CYAN_BOLD(x) "\x1B[1;36m" x "\x1B[0m" diff --git a/src/common/net/Client.cpp b/src/common/net/Client.cpp index f4553d97..6a79749d 100644 --- a/src/common/net/Client.cpp +++ b/src/common/net/Client.cpp @@ -29,6 +29,13 @@ #include +#ifndef XMRIG_NO_TLS +# include +# include +# include "common/net/Tls.h" +#endif + + #include "common/interfaces/IClientListener.h" #include "common/log/Log.h" #include "common/net/Client.h" @@ -48,6 +55,17 @@ int64_t Client::m_sequence = 1; xmrig::Storage Client::m_storage; +#ifdef APP_DEBUG +static const char *states[] = { + "unconnected", + "host-lookup", + "connecting", + "connected", + "closing" +}; +#endif + + Client::Client(int id, const char *agent, IClientListener *listener) : m_ipv6(false), m_nicehash(false), @@ -61,6 +79,7 @@ Client::Client(int id, const char *agent, IClientListener *listener) : m_failures(0), m_recvBufPos(0), m_state(UnconnectedState), + m_tls(nullptr), m_expire(0), m_jobs(0), m_keepAlive(0), @@ -92,6 +111,12 @@ Client::~Client() void Client::connect() { +# ifndef XMRIG_NO_TLS + if (m_pool.isTLS()) { + m_tls = new Tls(this); + } +# endif + resolve(m_pool.host()); } @@ -122,6 +147,7 @@ void Client::deleteLater() } + void Client::setPool(const Pool &pool) { if (!pool.isValid()) { @@ -160,6 +186,30 @@ bool Client::disconnect() } +const char *Client::tlsFingerprint() const +{ +# ifndef XMRIG_NO_TLS + if (isTLS() && m_pool.fingerprint() == nullptr) { + return m_tls->fingerprint(); + } +# endif + + return nullptr; +} + + +const char *Client::tlsVersion() const +{ +# ifndef XMRIG_NO_TLS + if (isTLS()) { + return m_tls->version(); + } +# endif + + return nullptr; +} + + int64_t Client::submit(const JobResult &result) { using namespace rapidjson; @@ -245,6 +295,16 @@ bool Client::isCriticalError(const char *message) } +bool Client::isTLS() const +{ +# ifndef XMRIG_NO_TLS + return m_pool.isTLS() && m_tls; +# else + return false; +# endif +} + + bool Client::parseJob(const rapidjson::Value ¶ms, int *code) { if (!params.IsObject()) { @@ -270,17 +330,17 @@ bool Client::parseJob(const rapidjson::Value ¶ms, int *code) } if (params.HasMember("algo")) { - job.algorithm().parseAlgorithm(params["algo"].GetString()); + job.setAlgorithm(params["algo"].GetString()); } if (params.HasMember("variant")) { const rapidjson::Value &variant = params["variant"]; if (variant.IsInt()) { - job.algorithm().parseVariant(variant.GetInt()); + job.setVariant(variant.GetInt()); } else if (variant.IsString()){ - job.algorithm().parseVariant(variant.GetString()); + job.setVariant(variant.GetString()); } } @@ -330,6 +390,39 @@ bool Client::parseLogin(const rapidjson::Value &result, int *code) } +bool Client::send(BIO *bio) +{ +# ifndef XMRIG_NO_TLS + uv_buf_t buf; + buf.len = BIO_get_mem_data(bio, &buf.base); + + if (buf.len == 0) { + return true; + } + + LOG_DEBUG("[%s] TLS send (%d bytes)", m_pool.url(), static_cast(buf.len)); + + bool result = false; + if (state() == ConnectedState && uv_is_writable(m_stream)) { + result = uv_try_write(m_stream, &buf, 1) > 0; + + if (!result) { + close(); + } + } + else { + LOG_DEBUG_ERR("[%s] send failed, invalid state: %d", m_pool.url(), m_state); + } + + (void) BIO_reset(bio); + + return result; +# else + return false; +# endif +} + + bool Client::verifyAlgorithm(const xmrig::Algorithm &algorithm) const { # ifdef XMRIG_PROXY_PROJECT @@ -389,7 +482,9 @@ int64_t Client::send(const rapidjson::Document &doc) doc.Accept(writer); const size_t size = buffer.GetSize(); - if (size > (sizeof(m_buf) - 2)) { + if (size > (sizeof(m_sendBuf) - 2)) { + LOG_ERR("[%s] send failed: \"send buffer overflow: %zu > %zu\"", m_pool.url(), size, (sizeof(m_sendBuf) - 2)); + close(); return -1; } @@ -404,16 +499,27 @@ int64_t Client::send(const rapidjson::Document &doc) int64_t Client::send(size_t size) { LOG_DEBUG("[%s] send (%d bytes): \"%s\"", m_pool.url(), size, m_sendBuf); - if (state() != ConnectedState || !uv_is_writable(m_stream)) { - LOG_DEBUG_ERR("[%s] send failed, invalid state: %d", m_pool.url(), m_state); - return -1; + +# ifndef XMRIG_NO_TLS + if (isTLS()) { + if (!m_tls->send(m_sendBuf, size)) { + return -1; + } } + else +# endif + { + if (state() != ConnectedState || !uv_is_writable(m_stream)) { + LOG_DEBUG_ERR("[%s] send failed, invalid state: %d", m_pool.url(), m_state); + return -1; + } - uv_buf_t buf = uv_buf_init(m_sendBuf, (unsigned int) size); + uv_buf_t buf = uv_buf_init(m_sendBuf, (unsigned int) size); - if (uv_try_write(m_stream, &buf, 1) < 0) { - close(); - return -1; + if (uv_try_write(m_stream, &buf, 1) < 0) { + close(); + return -1; + } } m_expire = uv_now(uv_default_loop()) + kResponseTimeout; @@ -463,6 +569,22 @@ void Client::connect(sockaddr *addr) } +void Client::handshake() +{ +# ifndef XMRIG_NO_TLS + if (isTLS()) { + m_expire = uv_now(uv_default_loop()) + kResponseTimeout; + + m_tls->handshake(); + } + else +# endif + { + login(); + } +} + + void Client::login() { using namespace rapidjson; @@ -511,6 +633,13 @@ void Client::onClose() m_socket = nullptr; setState(UnconnectedState); +# ifndef XMRIG_NO_TLS + if (m_tls) { + delete m_tls; + m_tls = nullptr; + } +# endif + reconnect(); } @@ -665,6 +794,35 @@ void Client::ping() } +void Client::read() +{ + char* end; + char* start = m_recvBuf.base; + size_t remaining = m_recvBufPos; + + while ((end = static_cast(memchr(start, '\n', remaining))) != nullptr) { + end++; + size_t len = end - start; + parse(start, len); + + remaining -= len; + start = end; + } + + if (remaining == 0) { + m_recvBufPos = 0; + return; + } + + if (start == m_recvBuf.base) { + return; + } + + memcpy(m_recvBuf.base, start, remaining); + m_recvBufPos = remaining; +} + + void Client::reconnect() { if (!m_listener) { @@ -689,7 +847,7 @@ void Client::reconnect() void Client::setState(SocketState state) { - LOG_DEBUG("[%s] state: %d", m_pool.url(), state); + LOG_DEBUG("[%s] state: \"%s\"", m_pool.url(), states[state]); if (m_state == state) { return; @@ -757,7 +915,7 @@ void Client::onConnect(uv_connect_t *req, int status) uv_read_start(client->m_stream, Client::onAllocBuffer, Client::onRead); delete req; - client->login(); + client->handshake(); } @@ -789,30 +947,18 @@ void Client::onRead(uv_stream_t *stream, ssize_t nread, const uv_buf_t *buf) client->m_recvBufPos += nread; - char* end; - char* start = client->m_recvBuf.base; - size_t remaining = client->m_recvBufPos; +# ifndef XMRIG_NO_TLS + if (client->isTLS()) { + LOG_DEBUG("[%s] TLS received (%d bytes)", client->m_pool.url(), static_cast(nread)); - while ((end = static_cast(memchr(start, '\n', remaining))) != nullptr) { - end++; - size_t len = end - start; - client->parse(start, len); - - remaining -= len; - start = end; - } - - if (remaining == 0) { + client->m_tls->read(client->m_recvBuf.base, client->m_recvBufPos); client->m_recvBufPos = 0; - return; } - - if (start == client->m_recvBuf.base) { - return; + else +# endif + { + client->read(); } - - memcpy(client->m_recvBuf.base, start, remaining); - client->m_recvBufPos = remaining; } diff --git a/src/common/net/Client.h b/src/common/net/Client.h index 4be8badb..d6418338 100644 --- a/src/common/net/Client.h +++ b/src/common/net/Client.h @@ -21,8 +21,8 @@ * along with this program. If not, see . */ -#ifndef __CLIENT_H__ -#define __CLIENT_H__ +#ifndef XMRIG_CLIENT_H +#define XMRIG_CLIENT_H #include @@ -43,6 +43,9 @@ class IClientListener; class JobResult; +typedef struct bio_st BIO; + + class Client { public: @@ -54,12 +57,20 @@ public: ClosingState }; - constexpr static int kResponseTimeout = 20 * 1000; + constexpr static int kResponseTimeout = 20 * 1000; + +# ifndef XMRIG_NO_TLS + constexpr static int kInputBufferSize = 1024 * 16; +# else + constexpr static int kInputBufferSize = 1024 * 2; +# endif Client(int id, const char *agent, IClientListener *listener); ~Client(); bool disconnect(); + const char *tlsFingerprint() const; + const char *tlsVersion() const; int64_t submit(const JobResult &result); void connect(); void connect(const Pool &pool); @@ -80,6 +91,9 @@ public: inline void setRetryPause(int ms) { m_retryPause = ms; } private: + class Tls; + + enum Extensions { NicehashExt = 1, AlgoExt = 2 @@ -87,14 +101,17 @@ private: bool close(); bool isCriticalError(const char *message); + bool isTLS() const; bool parseJob(const rapidjson::Value ¶ms, int *code); bool parseLogin(const rapidjson::Value &result, int *code); + bool send(BIO *bio); bool verifyAlgorithm(const xmrig::Algorithm &algorithm) const; int resolve(const char *host); int64_t send(const rapidjson::Document &doc); int64_t send(size_t size); void connect(const std::vector &ipv4, const std::vector &ipv6); void connect(sockaddr *addr); + void handshake(); void login(); void onClose(); void parse(char *line, size_t len); @@ -102,6 +119,7 @@ private: void parseNotification(const char *method, const rapidjson::Value ¶ms, const rapidjson::Value &error); void parseResponse(int64_t id, const rapidjson::Value &result, const rapidjson::Value &error); void ping(); + void read(); void reconnect(); void setState(SocketState state); void startTimeout(); @@ -120,9 +138,9 @@ private: bool m_ipv6; bool m_nicehash; bool m_quiet; - char m_buf[2048]; + char m_buf[kInputBufferSize]; char m_ip[46]; - char m_sendBuf[768]; + char m_sendBuf[2048]; const char *m_agent; IClientListener *m_listener; int m_extensions; @@ -135,6 +153,7 @@ private: size_t m_recvBufPos; SocketState m_state; std::map m_results; + Tls *m_tls; uint64_t m_expire; uint64_t m_jobs; uint64_t m_keepAlive; @@ -150,4 +169,4 @@ private: }; -#endif /* __CLIENT_H__ */ +#endif /* XMRIG_CLIENT_H */ diff --git a/src/common/net/Job.cpp b/src/common/net/Job.cpp index 80b521ea..2bfb39f0 100644 --- a/src/common/net/Job.cpp +++ b/src/common/net/Job.cpp @@ -6,6 +6,7 @@ * Copyright 2016 Jay D Dee * Copyright 2017-2018 XMR-Stak , * Copyright 2018 Lee Clagett + * Copyright 2018 SChernykh * Copyright 2016-2018 XMRig , * * This program is free software: you can redistribute it and/or modify @@ -58,6 +59,7 @@ static inline char hf_bin2hex(unsigned char c) Job::Job() : + m_autoVariant(false), m_nicehash(false), m_poolId(-2), m_threadId(-1), @@ -69,7 +71,8 @@ Job::Job() : } -Job::Job(int poolId, bool nicehash, xmrig::Algorithm algorithm, const xmrig::Id &clientId) : +Job::Job(int poolId, bool nicehash, const xmrig::Algorithm &algorithm, const xmrig::Id &clientId) : + m_autoVariant(algorithm.variant() == xmrig::VARIANT_AUTO), m_nicehash(nicehash), m_poolId(poolId), m_threadId(-1), @@ -112,6 +115,10 @@ bool Job::setBlob(const char *blob) m_nicehash = true; } + if (m_autoVariant) { + m_algorithm.setVariant(variant()); + } + # ifdef XMRIG_PROXY_PROJECT memset(m_rawBlob, 0, sizeof(m_rawBlob)); memcpy(m_rawBlob, blob, m_size * 2); @@ -163,25 +170,13 @@ bool Job::setTarget(const char *target) } -xmrig::Variant Job::variant() const +void Job::setAlgorithm(const char *algo) { - if (m_algorithm.variant() == xmrig::VARIANT_XTL && m_blob[0] < 4) { - return xmrig::VARIANT_1; - } - - if (m_algorithm.variant() == xmrig::VARIANT_MSR && m_blob[0] < 7) { - return xmrig::VARIANT_1; - } - - if (m_algorithm.variant() == xmrig::VARIANT_XHV && m_blob[0] < 3) { - return xmrig::VARIANT_0; - } + m_algorithm.parseAlgorithm(algo); if (m_algorithm.variant() == xmrig::VARIANT_AUTO) { - return m_algorithm.algo() == xmrig::CRYPTONIGHT_HEAVY ? xmrig::VARIANT_0 : xmrig::VARIANT_1; + m_algorithm.setVariant(variant()); } - - return m_algorithm.variant(); } @@ -229,3 +224,25 @@ bool Job::operator!=(const Job &other) const { return m_id != other.m_id || memcmp(m_blob, other.m_blob, sizeof(m_blob)) != 0; } + + +xmrig::Variant Job::variant() const +{ + using namespace xmrig; + + switch (m_algorithm.algo()) { + case CRYPTONIGHT: + return (m_blob[0] >= 8) ? VARIANT_2 : VARIANT_1; + + case CRYPTONIGHT_LITE: + return VARIANT_1; + + case CRYPTONIGHT_HEAVY: + return VARIANT_0; + + default: + break; + } + + return m_algorithm.variant(); +} diff --git a/src/common/net/Job.h b/src/common/net/Job.h index 049eb7d4..b561b9c1 100644 --- a/src/common/net/Job.h +++ b/src/common/net/Job.h @@ -6,6 +6,7 @@ * Copyright 2016 Jay D Dee * Copyright 2017-2018 XMR-Stak , * Copyright 2018 Lee Clagett + * Copyright 2018 SChernykh * Copyright 2016-2018 XMRig , * * This program is free software: you can redistribute it and/or modify @@ -22,8 +23,8 @@ * along with this program. If not, see . */ -#ifndef __JOB_H__ -#define __JOB_H__ +#ifndef XMRIG_JOB_H +#define XMRIG_JOB_H #include @@ -38,12 +39,12 @@ class Job { public: Job(); - Job(int poolId, bool nicehash, xmrig::Algorithm algorithm, const xmrig::Id &clientId); + Job(int poolId, bool nicehash, const xmrig::Algorithm &algorithm, const xmrig::Id &clientId); ~Job(); bool setBlob(const char *blob); bool setTarget(const char *target); - xmrig::Variant variant() const; + void setAlgorithm(const char *algo); inline bool isNicehash() const { return m_nicehash; } inline bool isValid() const { return m_size > 0 && m_diff > 0; } @@ -63,7 +64,8 @@ public: inline void setClientId(const xmrig::Id &id) { m_clientId = id; } inline void setPoolId(int poolId) { m_poolId = poolId; } inline void setThreadId(int threadId) { m_threadId = threadId; } - inline xmrig::Algorithm &algorithm() { return m_algorithm; } + inline void setVariant(const char *variant) { m_algorithm.parseVariant(variant); } + inline void setVariant(int variant) { m_algorithm.parseVariant(variant); } # ifdef XMRIG_PROXY_PROJECT inline char *rawBlob() { return m_rawBlob; } @@ -83,6 +85,9 @@ public: bool operator!=(const Job &other) const; private: + xmrig::Variant variant() const; + + bool m_autoVariant; bool m_nicehash; int m_poolId; int m_threadId; @@ -100,4 +105,4 @@ private: # endif }; -#endif /* __JOB_H__ */ +#endif /* XMRIG_JOB_H */ diff --git a/src/common/net/Pool.cpp b/src/common/net/Pool.cpp index 053f2507..141e5115 100644 --- a/src/common/net/Pool.cpp +++ b/src/common/net/Pool.cpp @@ -5,6 +5,7 @@ * Copyright 2014-2016 Wolf9466 * Copyright 2016 Jay D Dee * Copyright 2017-2018 XMR-Stak , + * Copyright 2018 SChernykh * Copyright 2016-2018 XMRig , * * This program is free software: you can redistribute it and/or modify @@ -45,6 +46,7 @@ Pool::Pool() : m_nicehash(false), + m_tls(false), m_keepAlive(0), m_port(kDefaultPort) { @@ -64,6 +66,7 @@ Pool::Pool() : */ Pool::Pool(const char *url) : m_nicehash(false), + m_tls(false), m_keepAlive(0), m_port(kDefaultPort) { @@ -71,8 +74,9 @@ Pool::Pool(const char *url) : } -Pool::Pool(const char *host, uint16_t port, const char *user, const char *password, int keepAlive, bool nicehash) : +Pool::Pool(const char *host, uint16_t port, const char *user, const char *password, int keepAlive, bool nicehash, bool tls) : m_nicehash(nicehash), + m_tls(tls), m_keepAlive(keepAlive), m_port(port), m_host(host), @@ -113,15 +117,17 @@ bool Pool::isCompatible(const xmrig::Algorithm &algorithm) const bool Pool::isEqual(const Pool &other) const { - return (m_nicehash == other.m_nicehash - && m_keepAlive == other.m_keepAlive - && m_port == other.m_port - && m_algorithm == other.m_algorithm - && m_host == other.m_host - && m_password == other.m_password - && m_rigId == other.m_rigId - && m_url == other.m_url - && m_user == other.m_user); + return (m_nicehash == other.m_nicehash + && m_tls == other.m_tls + && m_keepAlive == other.m_keepAlive + && m_port == other.m_port + && m_algorithm == other.m_algorithm + && m_fingerprint == other.m_fingerprint + && m_host == other.m_host + && m_password == other.m_password + && m_rigId == other.m_rigId + && m_url == other.m_url + && m_user == other.m_user); } @@ -133,7 +139,13 @@ bool Pool::parse(const char *url) const char *base = url; if (p) { - if (strncasecmp(url, "stratum+tcp://", 14)) { + if (strncasecmp(url, "stratum+tcp://", 14) == 0) { + m_tls = false; + } + else if (strncasecmp(url, "stratum+ssl://", 14) == 0) { + m_tls = true; + } + else { return false; } @@ -211,6 +223,7 @@ rapidjson::Value Pool::toJSON(rapidjson::Document &doc) const case xmrig::VARIANT_AUTO: case xmrig::VARIANT_0: case xmrig::VARIANT_1: + case xmrig::VARIANT_2: obj.AddMember("variant", m_algorithm.variant(), allocator); break; @@ -219,6 +232,9 @@ rapidjson::Value Pool::toJSON(rapidjson::Document &doc) const break; } + obj.AddMember("tls", isTLS(), allocator); + obj.AddMember("tls-fingerprint", fingerprint() ? Value(StringRef(fingerprint())).Move() : Value(kNullType).Move(), allocator); + return obj; } @@ -359,7 +375,7 @@ void Pool::adjustVariant(const xmrig::Variant variantHint) if (m_algorithm.algo() == CRYPTONIGHT_HEAVY) { m_algorithm.setVariant(VARIANT_0); } - else { + else if (m_algorithm.algo() == CRYPTONIGHT_LITE) { m_algorithm.setVariant(VARIANT_1); } # endif @@ -377,6 +393,7 @@ void Pool::rebuild() m_algorithms.push_back(m_algorithm); # ifndef XMRIG_PROXY_PROJECT + addVariant(xmrig::VARIANT_2); addVariant(xmrig::VARIANT_1); addVariant(xmrig::VARIANT_0); addVariant(xmrig::VARIANT_XTL); diff --git a/src/common/net/Pool.h b/src/common/net/Pool.h index 57a30d1e..123cc131 100644 --- a/src/common/net/Pool.h +++ b/src/common/net/Pool.h @@ -5,6 +5,7 @@ * Copyright 2014-2016 Wolf9466 * Copyright 2016 Jay D Dee * Copyright 2017-2018 XMR-Stak , + * Copyright 2018 SChernykh * Copyright 2016-2018 XMRig , * * This program is free software: you can redistribute it and/or modify @@ -21,8 +22,8 @@ * along with this program. If not, see . */ -#ifndef __POOL_H__ -#define __POOL_H__ +#ifndef XMRIG_POOL_H +#define XMRIG_POOL_H #include @@ -44,30 +45,35 @@ public: Pool(); Pool(const char *url); Pool(const char *host, - uint16_t port, - const char *user = nullptr, - const char *password = nullptr, - int keepAlive = 0, - bool nicehash = false + uint16_t port, + const char *user = nullptr, + const char *password = nullptr, + int keepAlive = 0, + bool nicehash = false, + bool tls = false ); - inline bool isNicehash() const { return m_nicehash; } - inline bool isValid() const { return !m_host.isNull() && m_port > 0; } - inline const char *host() const { return m_host.data(); } - inline const char *password() const { return !m_password.isNull() ? m_password.data() : kDefaultPassword; } - inline const char *rigId() const { return m_rigId.data(); } - inline const char *url() const { return m_url.data(); } - inline const char *user() const { return !m_user.isNull() ? m_user.data() : kDefaultUser; } - inline const xmrig::Algorithm &algorithm() const { return m_algorithm; } - inline const xmrig::Algorithms &algorithms() const { return m_algorithms; } - inline int keepAlive() const { return m_keepAlive; } - inline uint16_t port() const { return m_port; } - inline void setKeepAlive(int keepAlive) { m_keepAlive = keepAlive >= 0 ? keepAlive : 0; } - inline void setNicehash(bool nicehash) { m_nicehash = nicehash; } - inline void setPassword(const char *password) { m_password = password; } - inline void setRigId(const char *rigId) { m_rigId = rigId; } - inline void setUser(const char *user) { m_user = user; } - inline xmrig::Algorithm &algorithm() { return m_algorithm; } + inline bool isNicehash() const { return m_nicehash; } + inline bool isTLS() const { return m_tls; } + inline bool isValid() const { return !m_host.isNull() && m_port > 0; } + inline const char *fingerprint() const { return m_fingerprint.data(); } + inline const char *host() const { return m_host.data(); } + inline const char *password() const { return !m_password.isNull() ? m_password.data() : kDefaultPassword; } + inline const char *rigId() const { return m_rigId.data(); } + inline const char *url() const { return m_url.data(); } + inline const char *user() const { return !m_user.isNull() ? m_user.data() : kDefaultUser; } + inline const xmrig::Algorithm &algorithm() const { return m_algorithm; } + inline const xmrig::Algorithms &algorithms() const { return m_algorithms; } + inline int keepAlive() const { return m_keepAlive; } + inline uint16_t port() const { return m_port; } + inline void setFingerprint(const char *fingerprint) { m_fingerprint = fingerprint; } + inline void setKeepAlive(int keepAlive) { m_keepAlive = keepAlive >= 0 ? keepAlive : 0; } + inline void setNicehash(bool nicehash) { m_nicehash = nicehash; } + inline void setPassword(const char *password) { m_password = password; } + inline void setRigId(const char *rigId) { m_rigId = rigId; } + inline void setTLS(bool tls) { m_tls = tls; } + inline void setUser(const char *user) { m_user = user; } + inline xmrig::Algorithm &algorithm() { return m_algorithm; } inline bool operator!=(const Pool &other) const { return !isEqual(other); } inline bool operator==(const Pool &other) const { return isEqual(other); } @@ -91,10 +97,12 @@ private: void rebuild(); bool m_nicehash; + bool m_tls; int m_keepAlive; uint16_t m_port; xmrig::Algorithm m_algorithm; xmrig::Algorithms m_algorithms; + xmrig::c_str m_fingerprint; xmrig::c_str m_host; xmrig::c_str m_password; xmrig::c_str m_rigId; @@ -105,4 +113,4 @@ private: typedef std::vector Pools; -#endif /* __POOL_H__ */ +#endif /* XMRIG_POOL_H */ diff --git a/src/common/net/Tls.cpp b/src/common/net/Tls.cpp new file mode 100644 index 00000000..182d86ff --- /dev/null +++ b/src/common/net/Tls.cpp @@ -0,0 +1,190 @@ +/* XMRig + * Copyright 2010 Jeff Garzik + * Copyright 2012-2014 pooler + * Copyright 2014 Lucas Jones + * Copyright 2014-2016 Wolf9466 + * Copyright 2016 Jay D Dee + * Copyright 2017-2018 XMR-Stak , + * Copyright 2018 Lee Clagett + * Copyright 2018 SChernykh + * Copyright 2016-2018 XMRig , + * + * 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 3 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, see . + */ + + +#include + + +#include "common/net/Client.h" +#include "common/net/Tls.h" +#include "common/log/Log.h" + + +#ifdef _MSC_VER +# define strncasecmp(x,y,z) _strnicmp(x,y,z) +#endif + + +Client::Tls::Tls(Client *client) : + m_ready(false), + m_buf(), + m_fingerprint(), + m_client(client), + m_ssl(nullptr) +{ + m_ctx = SSL_CTX_new(SSLv23_method()); + assert(m_ctx != nullptr); + + if (!m_ctx) { + return; + } + + m_writeBio = BIO_new(BIO_s_mem()); + m_readBio = BIO_new(BIO_s_mem()); + SSL_CTX_set_options(m_ctx, SSL_OP_NO_SSLv2 | SSL_OP_NO_SSLv3); +} + + +Client::Tls::~Tls() +{ + if (m_ctx) { + SSL_CTX_free(m_ctx); + } + + if (m_ssl) { + SSL_free(m_ssl); + } +} + + +bool Client::Tls::handshake() +{ + m_ssl = SSL_new(m_ctx); + assert(m_ssl != nullptr); + + if (!m_ssl) { + return false; + } + + SSL_set_connect_state(m_ssl); + SSL_set_bio(m_ssl, m_readBio, m_writeBio); + SSL_do_handshake(m_ssl); + + return send(); +} + + +bool Client::Tls::send(const char *data, size_t size) +{ + SSL_write(m_ssl, data, size); + + return send(); +} + + +const char *Client::Tls::fingerprint() const +{ + return m_ready ? m_fingerprint : nullptr; +} + + +const char *Client::Tls::version() const +{ + return m_ready ? SSL_get_version(m_ssl) : nullptr; +} + + +void Client::Tls::read(const char *data, size_t size) +{ + BIO_write(m_readBio, data, size); + + if (!SSL_is_init_finished(m_ssl)) { + const int rc = SSL_connect(m_ssl); + + if (rc < 0 && SSL_get_error(m_ssl, rc) == SSL_ERROR_WANT_READ) { + send(); + } else if (rc == 1) { + X509 *cert = SSL_get_peer_certificate(m_ssl); + if (!verify(cert)) { + X509_free(cert); + m_client->close(); + + return; + } + + X509_free(cert); + m_ready = true; + m_client->login(); + } + + return; + } + + int bytes_read = 0; + while ((bytes_read = SSL_read(m_ssl, m_buf, sizeof(m_buf))) > 0) { + m_client->parse(m_buf, bytes_read); + } +} + + +bool Client::Tls::send() +{ + return m_client->send(m_writeBio); +} + + +bool Client::Tls::verify(X509 *cert) +{ + if (cert == nullptr) { + LOG_ERR("[%s] Failed to get server certificate", m_client->m_pool.url()); + + return false; + } + + if (!verifyFingerprint(cert)) { + LOG_ERR("[%s] Failed to verify server certificate fingerprint", m_client->m_pool.url()); + + const char *fingerprint = m_client->m_pool.fingerprint(); + if (strlen(m_fingerprint) == 64 && fingerprint != nullptr) { + LOG_ERR("\"%s\" was given", m_fingerprint); + LOG_ERR("\"%s\" was configured", fingerprint); + } + + return false; + } + + return true; +} + + +bool Client::Tls::verifyFingerprint(X509 *cert) +{ + const EVP_MD *digest = EVP_get_digestbyname("sha256"); + if (digest == nullptr) { + return false; + } + + unsigned char md[EVP_MAX_MD_SIZE]; + unsigned int dlen; + + if (X509_digest(cert, digest, md, &dlen) != 1) { + return false; + } + + Job::toHex(md, 32, m_fingerprint); + const char *fingerprint = m_client->m_pool.fingerprint(); + + return fingerprint == nullptr || strncasecmp(m_fingerprint, fingerprint, 64) == 0; +} diff --git a/src/common/net/Tls.h b/src/common/net/Tls.h new file mode 100644 index 00000000..6e38f32f --- /dev/null +++ b/src/common/net/Tls.h @@ -0,0 +1,62 @@ +/* XMRig + * Copyright 2010 Jeff Garzik + * Copyright 2012-2014 pooler + * Copyright 2014 Lucas Jones + * Copyright 2014-2016 Wolf9466 + * Copyright 2016 Jay D Dee + * Copyright 2017-2018 XMR-Stak , + * Copyright 2016-2018 XMRig , + * + * 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 3 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, see . + */ + +#ifndef XMRIG_TLS_H +#define XMRIG_TLS_H + + +#include + + +#include "common/net/Client.h" + + +class Client::Tls +{ +public: + Tls(Client *client); + ~Tls(); + + bool handshake(); + bool send(const char *data, size_t size); + const char *fingerprint() const; + const char *version() const; + void read(const char *data, size_t size); + +private: + bool send(); + bool verify(X509 *cert); + bool verifyFingerprint(X509 *cert); + + BIO *m_readBio; + BIO *m_writeBio; + bool m_ready; + char m_buf[1024 * 2]; + char m_fingerprint[32 * 2 + 8]; + Client *m_client; + SSL *m_ssl; + SSL_CTX *m_ctx; +}; + + +#endif /* XMRIG_TLS_H */ diff --git a/src/common/utils/timestamp.h b/src/common/utils/timestamp.h new file mode 100644 index 00000000..6b6a8ab2 --- /dev/null +++ b/src/common/utils/timestamp.h @@ -0,0 +1,47 @@ +/* XMRig + * Copyright 2010 Jeff Garzik + * Copyright 2012-2014 pooler + * Copyright 2014 Lucas Jones + * Copyright 2014-2016 Wolf9466 + * Copyright 2016 Jay D Dee + * Copyright 2017-2018 XMR-Stak , + * Copyright 2016-2018 XMRig , + * + * 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 3 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, see . + */ + +#ifndef XMRIG_TIMESTAMP_H +#define XMRIG_TIMESTAMP_H + + +#include + + +namespace xmrig { + + +static inline int64_t currentMSecsSinceEpoch() +{ + using namespace std::chrono; + if (high_resolution_clock::is_steady) { + return time_point_cast(high_resolution_clock::now()).time_since_epoch().count(); + } + + return time_point_cast(steady_clock::now()).time_since_epoch().count(); +} + + +} /* namespace xmrig */ + +#endif /* XMRIG_TIMESTAMP_H */ diff --git a/src/common/xmrig.h b/src/common/xmrig.h index 0ff945b9..52650f0d 100644 --- a/src/common/xmrig.h +++ b/src/common/xmrig.h @@ -21,8 +21,8 @@ * along with this program. If not, see . */ -#ifndef __XMRIG_H__ -#define __XMRIG_H__ +#ifndef XMRIG_XMRIG_H +#define XMRIG_XMRIG_H namespace xmrig @@ -33,7 +33,7 @@ enum Algo { INVALID_ALGO = -1, CRYPTONIGHT, /* CryptoNight (Monero) */ CRYPTONIGHT_LITE, /* CryptoNight-Lite (AEON) */ - CRYPTONIGHT_HEAVY /* CryptoNight-Heavy (SUMO) */ + CRYPTONIGHT_HEAVY /* CryptoNight-Heavy (RYO) */ }; @@ -65,8 +65,9 @@ enum Variant { VARIANT_XTL = 3, // Modified CryptoNight variant 1 (Stellite only) VARIANT_MSR = 4, // Modified CryptoNight variant 1 (Masari only) VARIANT_XHV = 5, // Modified CryptoNight-Heavy (Haven Protocol only) - VARIANT_XAO = 6, // Modified CryptoNight variant 1 (Alloy only) + VARIANT_XAO = 6, // Modified CryptoNight variant 0 (Alloy only) VARIANT_RTO = 7, // Modified CryptoNight variant 1 (Arto only) + VARIANT_2 = 8, // CryptoNight variant 2 VARIANT_MAX }; @@ -84,7 +85,25 @@ enum AesMode { }; +enum OclVendor { + OCL_VENDOR_UNKNOWN = -2, + OCL_VENDOR_MANUAL = -1, + OCL_VENDOR_AMD = 0, + OCL_VENDOR_NVIDIA = 1, + OCL_VENDOR_INTEL = 2 +}; + + +enum Assembly { + ASM_NONE, + ASM_AUTO, + ASM_INTEL, + ASM_RYZEN, + ASM_MAX +}; + + } /* namespace xmrig */ -#endif /* __XMRIG_H__ */ +#endif /* XMRIG_XMRIG_H */ diff --git a/src/config.json b/src/config.json index b2dad4c9..ae85b1f7 100644 --- a/src/config.json +++ b/src/config.json @@ -3,10 +3,13 @@ "api": { "port": 0, "access-token": null, + "id": null, "worker-id": null, "ipv6": false, "restricted": true }, + "asm": true, + "autosave": true, "av": 0, "background": false, "colors": true, @@ -19,13 +22,15 @@ "max-cpu-usage": 75, "pools": [ { - "url": "proxy.fee.xmrig.com:9999", - "user": "YOUR_WALLET", + "url": "donate.v2.xmrig.com:3333", + "user": "YOUR_WALLET_ADDRESS", "pass": "x", "rig-id": null, "nicehash": false, "keepalive": false, - "variant": 1 + "variant": -1, + "tls": false, + "tls-fingerprint": null } ], "print-time": 60, diff --git a/src/core/Config.cpp b/src/core/Config.cpp index fa6afdb4..c8891c77 100644 --- a/src/core/Config.cpp +++ b/src/core/Config.cpp @@ -27,9 +27,10 @@ #include "common/config/ConfigLoader.h" +#include "common/cpu/Cpu.h" #include "core/Config.h" #include "core/ConfigCreator.h" -#include "Cpu.h" +#include "crypto/Asm.h" #include "crypto/CryptoNight_constants.h" #include "rapidjson/document.h" #include "rapidjson/filewritestream.h" @@ -43,19 +44,16 @@ static char affinity_tmp[20] = { 0 }; xmrig::Config::Config() : xmrig::CommonConfig(), m_aesMode(AES_AUTO), m_algoVariant(AV_AUTO), + m_assembly(ASM_AUTO), m_hugePages(true), m_safe(false), + m_shouldSave(false), m_maxCpuUsage(75), m_priority(-1) { } -xmrig::Config::~Config() -{ -} - - bool xmrig::Config::reload(const char *json) { return xmrig::ConfigLoader::reload(this, json); @@ -75,11 +73,17 @@ void xmrig::Config::getJSON(rapidjson::Document &doc) const Value api(kObjectType); api.AddMember("port", apiPort(), allocator); api.AddMember("access-token", apiToken() ? Value(StringRef(apiToken())).Move() : Value(kNullType).Move(), allocator); + api.AddMember("id", apiId() ? Value(StringRef(apiId())).Move() : Value(kNullType).Move(), allocator); api.AddMember("worker-id", apiWorkerId() ? Value(StringRef(apiWorkerId())).Move() : Value(kNullType).Move(), allocator); api.AddMember("ipv6", isApiIPv6(), allocator); api.AddMember("restricted", isApiRestricted(), allocator); doc.AddMember("api", api, allocator); +# ifndef XMRIG_NO_ASM + doc.AddMember("asm", Asm::toJSON(m_assembly), allocator); +# endif + + doc.AddMember("autosave", isAutoSave(), allocator); doc.AddMember("av", algoVariant(), allocator); doc.AddMember("background", isBackground(), allocator); doc.AddMember("colors", isColors(), allocator); @@ -111,7 +115,7 @@ void xmrig::Config::getJSON(rapidjson::Document &doc) const doc.AddMember("retry-pause", retryPause(), allocator); doc.AddMember("safe", m_safe, allocator); - if (threadsMode() == Advanced) { + if (threadsMode() != Simple) { Value threads(kArrayType); for (const IThread *thread : m_threads.list) { @@ -121,7 +125,7 @@ void xmrig::Config::getJSON(rapidjson::Document &doc) const doc.AddMember("threads", threads, allocator); } else { - doc.AddMember("threads", threadsMode() == Automatic ? Value(kNullType) : Value(threadsCount()), allocator); + doc.AddMember("threads", threadsCount(), allocator); } doc.AddMember("user-agent", userAgent() ? Value(StringRef(userAgent())).Move() : Value(kNullType).Move(), allocator); @@ -152,7 +156,7 @@ bool xmrig::Config::finalize() if (!m_threads.cpu.empty()) { m_threads.mode = Advanced; - const bool softAES = (m_aesMode == AES_AUTO ? (Cpu::hasAES() ? AES_HW : AES_SOFT) : m_aesMode) == AES_SOFT; + const bool softAES = (m_aesMode == AES_AUTO ? (Cpu::info()->hasAES() ? AES_HW : AES_SOFT) : m_aesMode) == AES_SOFT; for (size_t i = 0; i < m_threads.cpu.size(); ++i) { m_threads.list.push_back(CpuThread::createFromData(i, m_algorithm.algo(), m_threads.cpu[i], m_priority, softAES)); @@ -161,25 +165,26 @@ bool xmrig::Config::finalize() return true; } - const AlgoVariant av = getAlgoVariant(); + const AlgoVariant av = getAlgoVariant(); m_threads.mode = m_threads.count ? Simple : Automatic; const size_t size = CpuThread::multiway(av) * cn_select_memory(m_algorithm.algo()) / 1024; if (!m_threads.count) { - m_threads.count = Cpu::optimalThreadsCount(size, m_maxCpuUsage); + m_threads.count = Cpu::info()->optimalThreadsCount(size, m_maxCpuUsage); } else if (m_safe) { - const size_t count = Cpu::optimalThreadsCount(size, m_maxCpuUsage); + const size_t count = Cpu::info()->optimalThreadsCount(size, m_maxCpuUsage); if (m_threads.count > count) { m_threads.count = count; } } for (size_t i = 0; i < m_threads.count; ++i) { - m_threads.list.push_back(CpuThread::createFromAV(i, m_algorithm.algo(), av, m_threads.mask, m_priority)); + m_threads.list.push_back(CpuThread::createFromAV(i, m_algorithm.algo(), av, m_threads.mask, m_priority, m_assembly)); } + m_shouldSave = m_threads.mode == Automatic; return true; } @@ -203,6 +208,12 @@ bool xmrig::Config::parseBoolean(int key, bool enable) m_aesMode = enable ? AES_HW : AES_SOFT; break; +# ifndef XMRIG_NO_ASM + case AssemblyKey: + m_assembly = Asm::parse(enable); + break; +# endif + default: break; } @@ -231,7 +242,7 @@ bool xmrig::Config::parseString(int key, const char *arg) case ThreadsKey: /* --threads */ if (strncmp(arg, "all", 3) == 0) { - m_threads.count = Cpu::threads(); + m_threads.count = Cpu::info()->threads(); return true; } @@ -243,6 +254,12 @@ bool xmrig::Config::parseString(int key, const char *arg) return parseUint64(key, p ? strtoull(p, nullptr, 16) : strtoull(arg, nullptr, 10)); } +# ifndef XMRIG_NO_ASM + case AssemblyKey: /* --asm */ + m_assembly = Asm::parse(arg); + break; +# endif + default: break; } @@ -338,10 +355,10 @@ xmrig::AlgoVariant xmrig::Config::getAlgoVariant() const # endif if (m_algoVariant <= AV_AUTO || m_algoVariant >= AV_MAX) { - return Cpu::hasAES() ? AV_SINGLE : AV_SINGLE_SOFT; + return Cpu::info()->hasAES() ? AV_SINGLE : AV_SINGLE_SOFT; } - if (m_safe && !Cpu::hasAES() && m_algoVariant <= AV_DOUBLE) { + if (m_safe && !Cpu::info()->hasAES() && m_algoVariant <= AV_DOUBLE) { return static_cast(m_algoVariant + 2); } @@ -353,10 +370,10 @@ xmrig::AlgoVariant xmrig::Config::getAlgoVariant() const xmrig::AlgoVariant xmrig::Config::getAlgoVariantLite() const { if (m_algoVariant <= AV_AUTO || m_algoVariant >= AV_MAX) { - return Cpu::hasAES() ? AV_DOUBLE : AV_DOUBLE_SOFT; + return Cpu::info()->hasAES() ? AV_DOUBLE : AV_DOUBLE_SOFT; } - if (m_safe && !Cpu::hasAES() && m_algoVariant <= AV_DOUBLE) { + if (m_safe && !Cpu::info()->hasAES() && m_algoVariant <= AV_DOUBLE) { return static_cast(m_algoVariant + 2); } diff --git a/src/core/Config.h b/src/core/Config.h index f0f1404f..eb33ee14 100644 --- a/src/core/Config.h +++ b/src/core/Config.h @@ -21,8 +21,8 @@ * along with this program. If not, see . */ -#ifndef __CONFIG_H__ -#define __CONFIG_H__ +#ifndef XMRIG_CONFIG_H +#define XMRIG_CONFIG_H #include @@ -69,7 +69,6 @@ public: Config(); - ~Config(); bool reload(const char *json); @@ -77,7 +76,9 @@ public: inline AesMode aesMode() const { return m_aesMode; } inline AlgoVariant algoVariant() const { return m_algoVariant; } + inline Assembly assembly() const { return m_assembly; } inline bool isHugePages() const { return m_hugePages; } + inline bool isShouldSave() const { return m_shouldSave && isAutoSave(); } inline const std::vector &threads() const { return m_threads.list; } inline int priority() const { return m_priority; } inline int threadsCount() const { return m_threads.list.size(); } @@ -116,8 +117,10 @@ private: AesMode m_aesMode; AlgoVariant m_algoVariant; + Assembly m_assembly; bool m_hugePages; bool m_safe; + bool m_shouldSave; int m_maxCpuUsage; int m_priority; Threads m_threads; @@ -126,4 +129,4 @@ private: } /* namespace xmrig */ -#endif /* __CONFIG_H__ */ +#endif /* XMRIG_CONFIG_H */ diff --git a/src/core/ConfigLoader_platform.h b/src/core/ConfigLoader_platform.h index bc6657d1..a932b235 100644 --- a/src/core/ConfigLoader_platform.h +++ b/src/core/ConfigLoader_platform.h @@ -22,8 +22,8 @@ * along with this program. If not, see . */ -#ifndef __CONFIGLOADER_PLATFORM_H__ -#define __CONFIGLOADER_PLATFORM_H__ +#ifndef XMRIG_CONFIGLOADER_PLATFORM_H +#define XMRIG_CONFIGLOADER_PLATFORM_H #ifdef _MSC_VER @@ -86,6 +86,7 @@ Options:\n\ --api-port=N port for the miner API\n\ --api-access-token=T access token for API\n\ --api-worker-id=ID custom worker-id for API\n\ + --api-id=ID custom instance ID for API\n\ --api-ipv6 enable IPv6 support for API\n\ --api-no-restricted enable full remote access (only if API token set)\n\ -h, --help display this help and exit\n\ @@ -101,6 +102,7 @@ static struct option const options[] = { { "api-access-token", 1, nullptr, xmrig::IConfig::ApiAccessTokenKey }, { "api-port", 1, nullptr, xmrig::IConfig::ApiPort }, { "api-worker-id", 1, nullptr, xmrig::IConfig::ApiWorkerIdKey }, + { "api-id", 1, nullptr, xmrig::IConfig::ApiIdKey }, { "api-ipv6", 0, nullptr, xmrig::IConfig::ApiIPv6Key }, { "api-no-restricted", 0, nullptr, xmrig::IConfig::ApiRestrictedKey }, { "av", 1, nullptr, xmrig::IConfig::AVKey }, @@ -130,8 +132,11 @@ static struct option const options[] = { { "user-agent", 1, nullptr, xmrig::IConfig::UserAgentKey }, { "userpass", 1, nullptr, xmrig::IConfig::UserpassKey }, { "rig-id", 1, nullptr, xmrig::IConfig::RigIdKey }, + { "tls", 0, nullptr, xmrig::IConfig::TlsKey }, + { "tls-fingerprint", 1, nullptr, xmrig::IConfig::FingerprintKey }, { "version", 0, nullptr, xmrig::IConfig::VersionKey }, - { 0, 0, 0, 0 } + { "asm", 1, nullptr, xmrig::IConfig::AssemblyKey }, + { nullptr, 0, nullptr, 0 } }; @@ -155,20 +160,24 @@ static struct option const config_options[] = { { "threads", 1, nullptr, xmrig::IConfig::ThreadsKey }, { "user-agent", 1, nullptr, xmrig::IConfig::UserAgentKey }, { "hw-aes", 0, nullptr, xmrig::IConfig::HardwareAESKey }, - { 0, 0, 0, 0 } + { "asm", 1, nullptr, xmrig::IConfig::AssemblyKey }, + { "autosave", 0, nullptr, xmrig::IConfig::AutoSaveKey }, + { nullptr, 0, nullptr, 0 } }; static struct option const pool_options[] = { - { "url", 1, nullptr, xmrig::IConfig::UrlKey }, - { "pass", 1, nullptr, xmrig::IConfig::PasswordKey }, - { "user", 1, nullptr, xmrig::IConfig::UserKey }, - { "userpass", 1, nullptr, xmrig::IConfig::UserpassKey }, - { "nicehash", 0, nullptr, xmrig::IConfig::NicehashKey }, - { "keepalive", 2, nullptr, xmrig::IConfig::KeepAliveKey }, - { "variant", 1, nullptr, xmrig::IConfig::VariantKey }, - { "rig-id", 1, nullptr, xmrig::IConfig::RigIdKey }, - { 0, 0, 0, 0 } + { "url", 1, nullptr, xmrig::IConfig::UrlKey }, + { "pass", 1, nullptr, xmrig::IConfig::PasswordKey }, + { "user", 1, nullptr, xmrig::IConfig::UserKey }, + { "userpass", 1, nullptr, xmrig::IConfig::UserpassKey }, + { "nicehash", 0, nullptr, xmrig::IConfig::NicehashKey }, + { "keepalive", 2, nullptr, xmrig::IConfig::KeepAliveKey }, + { "variant", 1, nullptr, xmrig::IConfig::VariantKey }, + { "rig-id", 1, nullptr, xmrig::IConfig::RigIdKey }, + { "tls", 0, nullptr, xmrig::IConfig::TlsKey }, + { "tls-fingerprint", 1, nullptr, xmrig::IConfig::FingerprintKey }, + { nullptr, 0, nullptr, 0 } }; @@ -178,10 +187,11 @@ static struct option const api_options[] = { { "worker-id", 1, nullptr, xmrig::IConfig::ApiWorkerIdKey }, { "ipv6", 0, nullptr, xmrig::IConfig::ApiIPv6Key }, { "restricted", 0, nullptr, xmrig::IConfig::ApiRestrictedKey }, - { 0, 0, 0, 0 } + { "id", 1, nullptr, xmrig::IConfig::ApiIdKey }, + { nullptr, 0, nullptr, 0 } }; } /* namespace xmrig */ -#endif /* __CONFIGLOADER_PLATFORM_H__ */ +#endif /* XMRIG_CONFIGLOADER_PLATFORM_H */ diff --git a/src/core/Controller.cpp b/src/core/Controller.cpp index ce73f037..792ac939 100644 --- a/src/core/Controller.cpp +++ b/src/core/Controller.cpp @@ -26,6 +26,7 @@ #include "common/config/ConfigLoader.h" +#include "common/cpu/Cpu.h" #include "common/interfaces/IControllerListener.h" #include "common/log/ConsoleLog.h" #include "common/log/FileLog.h" @@ -33,7 +34,6 @@ #include "common/Platform.h" #include "core/Config.h" #include "core/Controller.h" -#include "Cpu.h" #include "net/Network.h" diff --git a/src/Cpu.cpp b/src/core/cpu/AdvancedCpuInfo.cpp similarity index 57% rename from src/Cpu.cpp rename to src/core/cpu/AdvancedCpuInfo.cpp index eebe585d..1f86a420 100644 --- a/src/Cpu.cpp +++ b/src/core/cpu/AdvancedCpuInfo.cpp @@ -21,65 +21,25 @@ * along with this program. If not, see . */ - #include #include #include +#include -#include "Cpu.h" +#include "core/cpu/AdvancedCpuInfo.h" -bool Cpu::m_l2_exclusive = false; -char Cpu::m_brand[64] = { 0 }; -int Cpu::m_flags = 0; -int Cpu::m_l2_cache = 0; -int Cpu::m_l3_cache = 0; -int Cpu::m_sockets = 1; -int Cpu::m_totalCores = 0; -size_t Cpu::m_totalThreads = 0; - - -size_t Cpu::optimalThreadsCount(size_t size, int maxCpuUsage) -{ - if (m_totalThreads == 1) { - return 1; - } - - size_t cache = 0; - if (m_l3_cache) { - cache = m_l2_exclusive ? (m_l2_cache + m_l3_cache) : m_l3_cache; - } - else { - cache = m_l2_cache; - } - - size_t count = 0; - - if (cache) { - count = cache / size; - - if (cache % size >= size / 2) { - count++; - } - } - else { - count = m_totalThreads / 2; - } - - if (count > m_totalThreads) { - count = m_totalThreads; - } - - if (((float) count / m_totalThreads * 100) > maxCpuUsage) { - count = (int) ceil((float) m_totalThreads * (maxCpuUsage / 100.0)); - } - - return count < 1 ? 1 : count; -} - - -void Cpu::initCommon() +xmrig::AdvancedCpuInfo::AdvancedCpuInfo() : + m_assembly(ASM_NONE), + m_aes(false), + m_L2_exclusive(false), + m_brand(), + m_cores(0), + m_L2(0), + m_L3(0), + m_sockets(1), + m_threads(std::thread::hardware_concurrency()) { struct cpu_raw_data_t raw = { 0 }; struct cpu_id_t data = { 0 }; @@ -87,42 +47,78 @@ void Cpu::initCommon() cpuid_get_raw_data(&raw); cpu_identify(&raw, &data); - strncpy(m_brand, data.brand_str, sizeof(m_brand) - 1); - - m_totalThreads = data.total_logical_cpus; - m_sockets = m_totalThreads / data.num_logical_cpus; + strncpy(m_brand, data.brand_str, sizeof(m_brand)); + m_sockets = threads() / data.num_logical_cpus; if (m_sockets == 0) { m_sockets = 1; } - m_totalCores = data.num_cores * m_sockets; - m_l3_cache = data.l3_cache > 0 ? data.l3_cache * m_sockets : 0; + m_cores = data.num_cores * m_sockets; + m_L3 = data.l3_cache > 0 ? data.l3_cache * m_sockets : 0; // Workaround for AMD CPUs https://github.com/anrieff/libcpuid/issues/97 if (data.vendor == VENDOR_AMD && data.ext_family >= 0x15 && data.ext_family < 0x17) { - m_l2_cache = data.l2_cache * (m_totalCores / 2) * m_sockets; - m_l2_exclusive = true; + m_L2 = data.l2_cache * (cores() / 2) * m_sockets; + m_L2_exclusive = true; } // Workaround for Intel Pentium Dual-Core, Core Duo, Core 2 Duo, Core 2 Quad and their Xeon homologue // These processors have L2 cache shared by 2 cores. else if (data.vendor == VENDOR_INTEL && data.ext_family == 0x06 && (data.ext_model == 0x0E || data.ext_model == 0x0F || data.ext_model == 0x17)) { - int l2_count_per_socket = m_totalCores > 1 ? m_totalCores / 2 : 1; - m_l2_cache = data.l2_cache > 0 ? data.l2_cache * l2_count_per_socket * m_sockets : 0; + int l2_count_per_socket = cores() > 1 ? cores() / 2 : 1; + m_L2 = data.l2_cache > 0 ? data.l2_cache * l2_count_per_socket * m_sockets : 0; } else{ - m_l2_cache = data.l2_cache > 0 ? data.l2_cache * m_totalCores * m_sockets : 0; + m_L2 = data.l2_cache > 0 ? data.l2_cache * cores() * m_sockets : 0; } -# if defined(__x86_64__) || defined(_M_AMD64) - m_flags |= X86_64; -# endif - if (data.flags[CPU_FEATURE_AES]) { - m_flags |= AES; - } + m_aes = true; - if (data.flags[CPU_FEATURE_BMI2]) { - m_flags |= BMI2; + if (data.vendor == VENDOR_AMD && data.ext_family >= 23) { + m_assembly = ASM_RYZEN; + } + else if (data.vendor == VENDOR_INTEL && data.ext_model >= 42) { + m_assembly = ASM_INTEL; + } } } + + +size_t xmrig::AdvancedCpuInfo::optimalThreadsCount(size_t memSize, int maxCpuUsage) const +{ + if (threads() == 1) { + return 1; + } + + size_t cache = 0; + if (m_L3) { + cache = m_L2_exclusive ? (m_L2 + m_L3) : m_L3; + } + else { + cache = m_L2; + } + + size_t count = 0; + + if (cache) { + count = cache / memSize; + + if (cache % memSize >= memSize / 2) { + count++; + } + } + else { + count = threads() / 2; + } + + if (count > (size_t) threads()) { + count = threads(); + } + + if (((float) count / threads() * 100) > maxCpuUsage) { + count = (int) ceil((float) threads() * (maxCpuUsage / 100.0)); + } + + return count < 1 ? 1 : count; +} diff --git a/src/core/cpu/AdvancedCpuInfo.h b/src/core/cpu/AdvancedCpuInfo.h new file mode 100644 index 00000000..5e8967ad --- /dev/null +++ b/src/core/cpu/AdvancedCpuInfo.h @@ -0,0 +1,75 @@ +/* XMRig + * Copyright 2010 Jeff Garzik + * Copyright 2012-2014 pooler + * Copyright 2014 Lucas Jones + * Copyright 2014-2016 Wolf9466 + * Copyright 2016 Jay D Dee + * Copyright 2017-2018 XMR-Stak , + * Copyright 2016-2018 XMRig , + * + * 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 3 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, see . + */ + +#ifndef XMRIG_ADVANCEDCPUINFO_H +#define XMRIG_ADVANCEDCPUINFO_H + + +#include "common/interfaces/ICpuInfo.h" + + +namespace xmrig { + + +class AdvancedCpuInfo : public ICpuInfo +{ +public: + AdvancedCpuInfo(); + +protected: + size_t optimalThreadsCount(size_t memSize, int maxCpuUsage) const override; + + inline Assembly assembly() const override { return m_assembly; } + inline bool hasAES() const override { return m_aes; } + inline bool isSupported() const override { return true; } + inline const char *brand() const override { return m_brand; } + inline int32_t cores() const override { return m_cores; } + inline int32_t L2() const override { return m_L2; } + inline int32_t L3() const override { return m_L3; } + inline int32_t nodes() const override { return -1; } + inline int32_t sockets() const override { return m_sockets; } + inline int32_t threads() const override { return m_threads; } + +# if defined(__x86_64__) || defined(_M_AMD64) + inline bool isX64() const override { return true; } +# else + inline bool isX64() const override { return false; } +# endif + +private: + Assembly m_assembly; + bool m_aes; + bool m_L2_exclusive; + char m_brand[64]; + int32_t m_cores; + int32_t m_L2; + int32_t m_L3; + int32_t m_sockets; + int32_t m_threads; +}; + + +} /* namespace xmrig */ + + +#endif /* XMRIG_ADVANCEDCPUINFO_H */ diff --git a/src/Cpu_win.cpp b/src/core/cpu/Cpu.cpp similarity index 69% rename from src/Cpu_win.cpp rename to src/core/cpu/Cpu.cpp index 7258f726..773255d2 100644 --- a/src/Cpu_win.cpp +++ b/src/core/cpu/Cpu.cpp @@ -22,20 +22,40 @@ */ -#include +#include -#include "Cpu.h" +#include "common/cpu/Cpu.h" -void Cpu::init() +#ifndef XMRIG_NO_LIBCPUID +# include "core/cpu/AdvancedCpuInfo.h" +#endif + + +static xmrig::ICpuInfo *cpuInfo = nullptr; + + +xmrig::ICpuInfo *xmrig::Cpu::info() { -# ifdef XMRIG_NO_LIBCPUID - SYSTEM_INFO sysinfo; - GetSystemInfo(&sysinfo); + assert(cpuInfo != nullptr); - m_totalThreads = sysinfo.dwNumberOfProcessors; -# endif - - initCommon(); + return cpuInfo; +} + + +void xmrig::Cpu::init() +{ + assert(cpuInfo == nullptr); + + cpuInfo = new AdvancedCpuInfo(); +} + + +void xmrig::Cpu::release() +{ + assert(cpuInfo != nullptr); + + delete cpuInfo; + cpuInfo = nullptr; } diff --git a/src/crypto/Asm.cpp b/src/crypto/Asm.cpp new file mode 100644 index 00000000..48c1beb8 --- /dev/null +++ b/src/crypto/Asm.cpp @@ -0,0 +1,100 @@ +/* XMRig + * Copyright 2010 Jeff Garzik + * Copyright 2012-2014 pooler + * Copyright 2014 Lucas Jones + * Copyright 2014-2016 Wolf9466 + * Copyright 2016 Jay D Dee + * Copyright 2017-2018 XMR-Stak , + * Copyright 2016-2018 XMRig , + * + * 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 3 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, see . + */ + + +#include +#include + + +#ifdef _MSC_VER +# define strncasecmp _strnicmp +# define strcasecmp _stricmp +#endif + + +#include "crypto/Asm.h" +#include "rapidjson/document.h" + + +static const char *asmNames[] = { + "none", + "auto", + "intel", + "ryzen" +}; + + +xmrig::Assembly xmrig::Asm::parse(const char *assembly, Assembly defaultValue) +{ + constexpr size_t const size = sizeof(asmNames) / sizeof((asmNames)[0]); + assert(assembly != nullptr); + assert(ASM_MAX == size); + + if (assembly == nullptr) { + return defaultValue; + } + + for (size_t i = 0; i < size; i++) { + if (strcasecmp(assembly, asmNames[i]) == 0) { + return static_cast(i); + } + } + + return defaultValue; +} + + +xmrig::Assembly xmrig::Asm::parse(const rapidjson::Value &value, Assembly defaultValue) +{ + if (value.IsBool()) { + return parse(value.GetBool()); + } + + if (value.IsString()) { + return parse(value.GetString(), defaultValue); + } + + return defaultValue; +} + + +const char *xmrig::Asm::toString(Assembly assembly) +{ + return asmNames[assembly]; +} + + +rapidjson::Value xmrig::Asm::toJSON(Assembly assembly) +{ + using namespace rapidjson; + + if (assembly == ASM_NONE) { + return Value(false); + } + + if (assembly == ASM_AUTO) { + return Value(true); + } + + return Value(StringRef(toString(assembly))); +} diff --git a/src/crypto/Asm.h b/src/crypto/Asm.h new file mode 100644 index 00000000..3b755fd6 --- /dev/null +++ b/src/crypto/Asm.h @@ -0,0 +1,50 @@ +/* XMRig + * Copyright 2010 Jeff Garzik + * Copyright 2012-2014 pooler + * Copyright 2014 Lucas Jones + * Copyright 2014-2016 Wolf9466 + * Copyright 2016 Jay D Dee + * Copyright 2017-2018 XMR-Stak , + * Copyright 2016-2018 XMRig , + * + * 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 3 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, see . + */ + +#ifndef XMRIG_ASM_H +#define XMRIG_ASM_H + + +#include "common/xmrig.h" +#include "rapidjson/fwd.h" + + +namespace xmrig { + + +class Asm +{ +public: + static Assembly parse(const char *assembly, Assembly defaultValue = ASM_AUTO); + static Assembly parse(const rapidjson::Value &value, Assembly defaultValue = ASM_AUTO); + static const char *toString(Assembly assembly); + static rapidjson::Value toJSON(Assembly assembly); + + inline static Assembly parse(bool enable) { return enable ? ASM_AUTO : ASM_NONE; } +}; + + +} /* namespace xmrig */ + + +#endif /* XMRIG_ASM_H */ diff --git a/src/crypto/CryptoNight.h b/src/crypto/CryptoNight.h index e8e86dc4..680f1740 100644 --- a/src/crypto/CryptoNight.h +++ b/src/crypto/CryptoNight.h @@ -22,8 +22,8 @@ * along with this program. If not, see . */ -#ifndef __CRYPTONIGHT_H__ -#define __CRYPTONIGHT_H__ +#ifndef XMRIG_CRYPTONIGHT_H +#define XMRIG_CRYPTONIGHT_H #include @@ -31,9 +31,9 @@ struct cryptonight_ctx { - alignas(16) uint8_t state[200]; - alignas(16) uint8_t* memory; + alignas(16) uint8_t state[224]; + alignas(16) uint8_t *memory; }; -#endif /* __CRYPTONIGHT_H__ */ +#endif /* XMRIG_CRYPTONIGHT_H */ diff --git a/src/crypto/CryptoNight_arm.h b/src/crypto/CryptoNight_arm.h index efb5759e..4fcebc3e 100644 --- a/src/crypto/CryptoNight_arm.h +++ b/src/crypto/CryptoNight_arm.h @@ -7,6 +7,7 @@ * Copyright 2016 Imran Yusuff * Copyright 2017-2018 XMR-Stak , * Copyright 2018 Lee Clagett + * Copyright 2018 SChernykh * Copyright 2016-2018 XMRig , * * This program is free software: you can redistribute it and/or modify @@ -23,8 +24,8 @@ * along with this program. If not, see . */ -#ifndef __CRYPTONIGHT_ARM_H__ -#define __CRYPTONIGHT_ARM_H__ +#ifndef XMRIG_CRYPTONIGHT_ARM_H +#define XMRIG_CRYPTONIGHT_ARM_H #include "common/crypto/keccak.h" @@ -95,9 +96,6 @@ static inline __attribute__((always_inline)) uint64_t _mm_cvtsi128_si64(__m128i } -#define EXTRACT64(X) _mm_cvtsi128_si64(X) - - #if defined (__arm64__) || defined (__aarch64__) static inline uint64_t __umul128(uint64_t a, uint64_t b, uint64_t* hi) { @@ -404,19 +402,27 @@ static inline __m128i aes_round_tweak_div(const __m128i &in, const __m128i &key) } -template -static inline void cryptonight_monero_tweak(uint64_t* mem_out, __m128i tmp) +template +static inline void cryptonight_monero_tweak(const uint8_t* l, uint64_t idx, __m128i ax0, __m128i bx0, __m128i bx1, __m128i cx) { - mem_out[0] = EXTRACT64(tmp); + uint64_t* mem_out = (uint64_t*)&l[idx]; - uint64_t vh = vgetq_lane_u64(tmp, 1); + if (VARIANT == xmrig::VARIANT_2) { + VARIANT2_SHUFFLE(l, idx, ax0, bx0, bx1); + _mm_store_si128((__m128i *)mem_out, _mm_xor_si128(bx0, cx)); + } else { + __m128i tmp = _mm_xor_si128(bx0, cx); + mem_out[0] = _mm_cvtsi128_si64(tmp); - uint8_t x = vh >> 24; - static const uint16_t table = 0x7531; - const uint8_t index = (((x >> SHIFT) & 6) | (x & 1)) << 1; - vh ^= ((table >> index) & 0x3) << 28; + uint64_t vh = vgetq_lane_u64(tmp, 1); - mem_out[1] = vh; + uint8_t x = vh >> 24; + static const uint16_t table = 0x7531; + const uint8_t index = (((x >> (VARIANT == xmrig::VARIANT_XTL ? 4 : 3)) & 6) | (x & 1)) << 1; + vh ^= ((table >> index) & 0x3) << 28; + + mem_out[1] = vh; + } } @@ -426,27 +432,29 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si constexpr size_t MASK = xmrig::cn_select_mask(); constexpr size_t ITERATIONS = xmrig::cn_select_iter(); constexpr size_t MEM = xmrig::cn_select_memory(); - constexpr bool IS_MONERO = xmrig::cn_is_monero(); + constexpr bool IS_V1 = xmrig::cn_base_variant() == xmrig::VARIANT_1; - if (IS_MONERO && size < 43) { + if (IS_V1 && size < 43) { memset(output, 0, 32); return; } xmrig::keccak(input, size, ctx[0]->state); - VARIANT1_INIT(0); - cn_explode_scratchpad((__m128i*) ctx[0]->state, (__m128i*) ctx[0]->memory); const uint8_t* l0 = ctx[0]->memory; uint64_t* h0 = reinterpret_cast(ctx[0]->state); + VARIANT1_INIT(0); + VARIANT2_INIT(0); + uint64_t al0 = h0[0] ^ h0[4]; uint64_t ah0 = h0[1] ^ h0[5]; __m128i bx0 = _mm_set_epi64x(h0[3] ^ h0[7], h0[2] ^ h0[6]); + __m128i bx1 = _mm_set_epi64x(h0[9] ^ h0[11], h0[8] ^ h0[10]); - uint64_t idx0 = h0[0] ^ h0[4]; + uint64_t idx0 = al0; for (size_t i = 0; i < ITERATIONS; i++) { __m128i cx; @@ -454,44 +462,47 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si cx = _mm_load_si128((__m128i *) &l0[idx0 & MASK]); } + const __m128i ax0 = _mm_set_epi64x(ah0, al0); if (VARIANT == xmrig::VARIANT_TUBE) { - cx = aes_round_tweak_div(cx, _mm_set_epi64x(ah0, al0)); + cx = aes_round_tweak_div(cx, ax0); } else if (SOFT_AES) { - cx = soft_aesenc((uint32_t*)&l0[idx0 & MASK], _mm_set_epi64x(ah0, al0)); + cx = soft_aesenc((uint32_t*)&l0[idx0 & MASK], ax0); } else { - cx = _mm_aesenc_si128(cx, _mm_set_epi64x(ah0, al0)); + cx = _mm_aesenc_si128(cx, ax0); } - if (IS_MONERO) { - cryptonight_monero_tweak((uint64_t*)&l0[idx0 & MASK], _mm_xor_si128(bx0, cx)); + if (IS_V1 || VARIANT == xmrig::VARIANT_2) { + cryptonight_monero_tweak(l0, idx0 & MASK, ax0, bx0, bx1, cx); } else { _mm_store_si128((__m128i *)&l0[idx0 & MASK], _mm_xor_si128(bx0, cx)); } - idx0 = EXTRACT64(cx); - bx0 = cx; + idx0 = _mm_cvtsi128_si64(cx); uint64_t hi, lo, cl, ch; cl = ((uint64_t*) &l0[idx0 & MASK])[0]; ch = ((uint64_t*) &l0[idx0 & MASK])[1]; - lo = __umul128(idx0, cl, &hi); + if (VARIANT == xmrig::VARIANT_2) { + VARIANT2_INTEGER_MATH(0, cl, cx); + lo = __umul128(idx0, cl, &hi); + VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx0, bx1, hi, lo); + } + else { + lo = __umul128(idx0, cl, &hi); + } al0 += hi; ah0 += lo; ((uint64_t*)&l0[idx0 & MASK])[0] = al0; - if (IS_MONERO) { - if (VARIANT == xmrig::VARIANT_TUBE || VARIANT == xmrig::VARIANT_RTO) { - ((uint64_t*)&l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0 ^ al0; - } - else { - ((uint64_t*)&l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0; - } - } - else { + if (IS_V1 && (VARIANT == xmrig::VARIANT_TUBE || VARIANT == xmrig::VARIANT_RTO)) { + ((uint64_t*)&l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0 ^ al0; + } else if (IS_V1) { + ((uint64_t*)&l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0; + } else { ((uint64_t*)&l0[idx0 & MASK])[1] = ah0; } @@ -514,6 +525,10 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si idx0 = d ^ q; } } + if (VARIANT == xmrig::VARIANT_2) { + bx1 = bx0; + } + bx0 = cx; } cn_implode_scratchpad((__m128i*) ctx[0]->memory, (__m128i*) ctx[0]->state); @@ -529,9 +544,9 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si constexpr size_t MASK = xmrig::cn_select_mask(); constexpr size_t ITERATIONS = xmrig::cn_select_iter(); constexpr size_t MEM = xmrig::cn_select_memory(); - constexpr bool IS_MONERO = xmrig::cn_is_monero(); + constexpr bool IS_V1 = xmrig::cn_base_variant() == xmrig::VARIANT_1; - if (IS_MONERO && size < 43) { + if (IS_V1 && size < 43) { memset(output, 0, 64); return; } @@ -539,14 +554,16 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si xmrig::keccak(input, size, ctx[0]->state); xmrig::keccak(input + size, size, ctx[1]->state); - VARIANT1_INIT(0); - VARIANT1_INIT(1); - const uint8_t* l0 = ctx[0]->memory; const uint8_t* l1 = ctx[1]->memory; uint64_t* h0 = reinterpret_cast(ctx[0]->state); uint64_t* h1 = reinterpret_cast(ctx[1]->state); + VARIANT1_INIT(0); + VARIANT1_INIT(1); + VARIANT2_INIT(0); + VARIANT2_INIT(1); + cn_explode_scratchpad((__m128i*) h0, (__m128i*) l0); cn_explode_scratchpad((__m128i*) h1, (__m128i*) l1); @@ -555,11 +572,13 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si uint64_t ah0 = h0[1] ^ h0[5]; uint64_t ah1 = h1[1] ^ h1[5]; - __m128i bx0 = _mm_set_epi64x(h0[3] ^ h0[7], h0[2] ^ h0[6]); - __m128i bx1 = _mm_set_epi64x(h1[3] ^ h1[7], h1[2] ^ h1[6]); + __m128i bx00 = _mm_set_epi64x(h0[3] ^ h0[7], h0[2] ^ h0[6]); + __m128i bx01 = _mm_set_epi64x(h0[9] ^ h0[11], h0[8] ^ h0[10]); + __m128i bx10 = _mm_set_epi64x(h1[3] ^ h1[7], h1[2] ^ h1[6]); + __m128i bx11 = _mm_set_epi64x(h1[9] ^ h1[11], h1[8] ^ h1[10]); - uint64_t idx0 = h0[0] ^ h0[4]; - uint64_t idx1 = h1[0] ^ h1[4]; + uint64_t idx0 = al0; + uint64_t idx1 = al1; for (size_t i = 0; i < ITERATIONS; i++) { __m128i cx0, cx1; @@ -568,52 +587,53 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si cx1 = _mm_load_si128((__m128i *) &l1[idx1 & MASK]); } + const __m128i ax0 = _mm_set_epi64x(ah0, al0); + const __m128i ax1 = _mm_set_epi64x(ah1, al1); if (VARIANT == xmrig::VARIANT_TUBE) { - cx0 = aes_round_tweak_div(cx0, _mm_set_epi64x(ah0, al0)); - cx1 = aes_round_tweak_div(cx1, _mm_set_epi64x(ah1, al1)); + cx0 = aes_round_tweak_div(cx0, ax0); + cx1 = aes_round_tweak_div(cx1, ax1); } else if (SOFT_AES) { - cx0 = soft_aesenc((uint32_t*)&l0[idx0 & MASK], _mm_set_epi64x(ah0, al0)); - cx1 = soft_aesenc((uint32_t*)&l1[idx1 & MASK], _mm_set_epi64x(ah1, al1)); + cx0 = soft_aesenc((uint32_t*)&l0[idx0 & MASK], ax0); + cx1 = soft_aesenc((uint32_t*)&l1[idx1 & MASK], ax1); } else { - cx0 = _mm_aesenc_si128(cx0, _mm_set_epi64x(ah0, al0)); - cx1 = _mm_aesenc_si128(cx1, _mm_set_epi64x(ah1, al1)); + cx0 = _mm_aesenc_si128(cx0, ax0); + cx1 = _mm_aesenc_si128(cx1, ax1); } - if (IS_MONERO) { - cryptonight_monero_tweak((uint64_t*)&l0[idx0 & MASK], _mm_xor_si128(bx0, cx0)); - cryptonight_monero_tweak((uint64_t*)&l1[idx1 & MASK], _mm_xor_si128(bx1, cx1)); + if (IS_V1 || (VARIANT == xmrig::VARIANT_2)) { + cryptonight_monero_tweak(l0, idx0 & MASK, ax0, bx00, bx01, cx0); + cryptonight_monero_tweak(l1, idx1 & MASK, ax1, bx10, bx11, cx1); } else { - _mm_store_si128((__m128i *) &l0[idx0 & MASK], _mm_xor_si128(bx0, cx0)); - _mm_store_si128((__m128i *) &l1[idx1 & MASK], _mm_xor_si128(bx1, cx1)); - }; + _mm_store_si128((__m128i *) &l0[idx0 & MASK], _mm_xor_si128(bx00, cx0)); + _mm_store_si128((__m128i *) &l1[idx1 & MASK], _mm_xor_si128(bx10, cx1)); + } - idx0 = EXTRACT64(cx0); - idx1 = EXTRACT64(cx1); - - bx0 = cx0; - bx1 = cx1; + idx0 = _mm_cvtsi128_si64(cx0); + idx1 = _mm_cvtsi128_si64(cx1); uint64_t hi, lo, cl, ch; cl = ((uint64_t*) &l0[idx0 & MASK])[0]; ch = ((uint64_t*) &l0[idx0 & MASK])[1]; - lo = __umul128(idx0, cl, &hi); + if (VARIANT == xmrig::VARIANT_2) { + VARIANT2_INTEGER_MATH(0, cl, cx0); + lo = __umul128(idx0, cl, &hi); + VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx00, bx01, hi, lo); + } else { + lo = __umul128(idx0, cl, &hi); + } al0 += hi; ah0 += lo; ((uint64_t*)&l0[idx0 & MASK])[0] = al0; - if (IS_MONERO) { - if (VARIANT == xmrig::VARIANT_TUBE || VARIANT == xmrig::VARIANT_RTO) { - ((uint64_t*)&l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0 ^ al0; - } - else { - ((uint64_t*)&l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0; - } - } - else { + if (IS_V1 && (VARIANT == xmrig::VARIANT_TUBE || VARIANT == xmrig::VARIANT_RTO)) { + ((uint64_t*)&l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0 ^ al0; + } else if (IS_V1) { + ((uint64_t*)&l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0; + } else { ((uint64_t*)&l0[idx0 & MASK])[1] = ah0; } @@ -639,22 +659,24 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si cl = ((uint64_t*) &l1[idx1 & MASK])[0]; ch = ((uint64_t*) &l1[idx1 & MASK])[1]; - lo = __umul128(idx1, cl, &hi); + if (VARIANT == xmrig::VARIANT_2) { + VARIANT2_INTEGER_MATH(1, cl, cx1); + lo = __umul128(idx1, cl, &hi); + VARIANT2_SHUFFLE2(l1, idx1 & MASK, ax1, bx10, bx11, hi, lo); + } else { + lo = __umul128(idx1, cl, &hi); + } al1 += hi; ah1 += lo; ((uint64_t*)&l1[idx1 & MASK])[0] = al1; - if (IS_MONERO) { - if (VARIANT == xmrig::VARIANT_TUBE || VARIANT == xmrig::VARIANT_RTO) { - ((uint64_t*)&l1[idx1 & MASK])[1] = ah1 ^ tweak1_2_1 ^ al1; - } - else { - ((uint64_t*)&l1[idx1 & MASK])[1] = ah1 ^ tweak1_2_1; - } - } - else { + if (IS_V1 && (VARIANT == xmrig::VARIANT_TUBE || VARIANT == xmrig::VARIANT_RTO)) { + ((uint64_t*)&l1[idx1 & MASK])[1] = ah1 ^ tweak1_2_1 ^ al1; + } else if (IS_V1) { + ((uint64_t*)&l1[idx1 & MASK])[1] = ah1 ^ tweak1_2_1; + } else { ((uint64_t*)&l1[idx1 & MASK])[1] = ah1; } @@ -677,6 +699,12 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si idx1 = d ^ q; } } + if (VARIANT == xmrig::VARIANT_2) { + bx01 = bx00; + bx11 = bx10; + } + bx00 = cx0; + bx10 = cx1; } cn_implode_scratchpad((__m128i*) l0, (__m128i*) h0); diff --git a/src/crypto/CryptoNight_constants.h b/src/crypto/CryptoNight_constants.h index 08a755d4..f13891a7 100644 --- a/src/crypto/CryptoNight_constants.h +++ b/src/crypto/CryptoNight_constants.h @@ -107,6 +107,7 @@ inline uint32_t cn_select_mask(Algo algorithm) template inline constexpr uint32_t cn_select_iter() { return 0; } template<> inline constexpr uint32_t cn_select_iter() { return CRYPTONIGHT_ITER; } template<> inline constexpr uint32_t cn_select_iter() { return CRYPTONIGHT_ITER; } +template<> inline constexpr uint32_t cn_select_iter() { return CRYPTONIGHT_ITER; } template<> inline constexpr uint32_t cn_select_iter() { return CRYPTONIGHT_ITER; } template<> inline constexpr uint32_t cn_select_iter() { return CRYPTONIGHT_MSR_ITER; } template<> inline constexpr uint32_t cn_select_iter() { return CRYPTONIGHT_XAO_ITER; } @@ -150,29 +151,16 @@ inline uint32_t cn_select_iter(Algo algorithm, Variant variant) } -template inline constexpr bool cn_is_monero() { return false; } -template<> inline constexpr bool cn_is_monero() { return false; } -template<> inline constexpr bool cn_is_monero() { return true; } -template<> inline constexpr bool cn_is_monero() { return true; } -template<> inline constexpr bool cn_is_monero() { return true; } -template<> inline constexpr bool cn_is_monero() { return true; } -template<> inline constexpr bool cn_is_monero() { return false; } -template<> inline constexpr bool cn_is_monero() { return false; } -template<> inline constexpr bool cn_is_monero() { return true; } - - -inline bool cn_is_monero(Variant variant) -{ - switch (variant) { - case VARIANT_0: - case VARIANT_XHV: - case VARIANT_RTO: - return false; - - default: - return true; - } -} +template inline constexpr Variant cn_base_variant() { return VARIANT_0; } +template<> inline constexpr Variant cn_base_variant() { return VARIANT_0; } +template<> inline constexpr Variant cn_base_variant() { return VARIANT_1; } +template<> inline constexpr Variant cn_base_variant() { return VARIANT_1; } +template<> inline constexpr Variant cn_base_variant() { return VARIANT_1; } +template<> inline constexpr Variant cn_base_variant() { return VARIANT_1; } +template<> inline constexpr Variant cn_base_variant() { return VARIANT_0; } +template<> inline constexpr Variant cn_base_variant() { return VARIANT_0; } +template<> inline constexpr Variant cn_base_variant() { return VARIANT_1; } +template<> inline constexpr Variant cn_base_variant() { return VARIANT_2; } } /* namespace xmrig */ diff --git a/src/crypto/CryptoNight_monero.h b/src/crypto/CryptoNight_monero.h index a758fdbc..52229026 100644 --- a/src/crypto/CryptoNight_monero.h +++ b/src/crypto/CryptoNight_monero.h @@ -6,6 +6,7 @@ * Copyright 2016 Jay D Dee * Copyright 2017-2018 XMR-Stak , * Copyright 2018 Lee Clagett + * Copyright 2018 SChernykh * Copyright 2016-2018 XMRig , * * This program is free software: you can redistribute it and/or modify @@ -22,29 +23,31 @@ * along with this program. If not, see . */ -#ifndef __CRYPTONIGHT_MONERO_H__ -#define __CRYPTONIGHT_MONERO_H__ +#ifndef XMRIG_CRYPTONIGHT_MONERO_H +#define XMRIG_CRYPTONIGHT_MONERO_H +#include +#include // VARIANT ALTERATIONS #ifndef XMRIG_ARM # define VARIANT1_INIT(part) \ uint64_t tweak1_2_##part = 0; \ - if (IS_MONERO) { \ + if (IS_V1) { \ tweak1_2_##part = (*reinterpret_cast(input + 35 + part * size) ^ \ *(reinterpret_cast(ctx[part]->state) + 24)); \ } #else # define VARIANT1_INIT(part) \ uint64_t tweak1_2_##part = 0; \ - if (IS_MONERO) { \ + if (IS_V1) { \ memcpy(&tweak1_2_##part, input + 35 + part * size, sizeof tweak1_2_##part); \ tweak1_2_##part ^= *(reinterpret_cast(ctx[part]->state) + 24); \ } #endif #define VARIANT1_1(p) \ - if (IS_MONERO) { \ + if (IS_V1) { \ const uint8_t tmp = reinterpret_cast(p)[11]; \ static const uint32_t table = 0x75310; \ const uint8_t index = (((tmp >> 3) & 6) | (tmp & 1)) << 1; \ @@ -52,9 +55,96 @@ } #define VARIANT1_2(p, part) \ - if (IS_MONERO) { \ + if (IS_V1) { \ (p) ^= tweak1_2_##part; \ } -#endif /* __CRYPTONIGHT_MONERO_H__ */ +#ifndef XMRIG_ARM +# define VARIANT2_INIT(part) \ + __m128i division_result_xmm_##part = _mm_cvtsi64_si128(h##part[12]); \ + __m128i sqrt_result_xmm_##part = _mm_cvtsi64_si128(h##part[13]); + +#ifdef _MSC_VER +# define VARIANT2_SET_ROUNDING_MODE() if (VARIANT == xmrig::VARIANT_2) { _control87(RC_DOWN, MCW_RC); } +#else +# define VARIANT2_SET_ROUNDING_MODE() if (VARIANT == xmrig::VARIANT_2) { fesetround(FE_DOWNWARD); } +#endif + +# define VARIANT2_INTEGER_MATH(part, cl, cx) \ + do { \ + const uint64_t sqrt_result = static_cast(_mm_cvtsi128_si64(sqrt_result_xmm_##part)); \ + const uint64_t cx_0 = _mm_cvtsi128_si64(cx); \ + cl ^= static_cast(_mm_cvtsi128_si64(division_result_xmm_##part)) ^ (sqrt_result << 32); \ + const uint32_t d = static_cast(cx_0 + (sqrt_result << 1)) | 0x80000001UL; \ + const uint64_t cx_1 = _mm_cvtsi128_si64(_mm_srli_si128(cx, 8)); \ + const uint64_t division_result = static_cast(cx_1 / d) + ((cx_1 % d) << 32); \ + division_result_xmm_##part = _mm_cvtsi64_si128(static_cast(division_result)); \ + sqrt_result_xmm_##part = int_sqrt_v2(cx_0 + division_result); \ + } while (0) + +# define VARIANT2_SHUFFLE(base_ptr, offset, _a, _b, _b1) \ + do { \ + const __m128i chunk1 = _mm_load_si128((__m128i *)((base_ptr) + ((offset) ^ 0x10))); \ + const __m128i chunk2 = _mm_load_si128((__m128i *)((base_ptr) + ((offset) ^ 0x20))); \ + const __m128i chunk3 = _mm_load_si128((__m128i *)((base_ptr) + ((offset) ^ 0x30))); \ + _mm_store_si128((__m128i *)((base_ptr) + ((offset) ^ 0x10)), _mm_add_epi64(chunk3, _b1)); \ + _mm_store_si128((__m128i *)((base_ptr) + ((offset) ^ 0x20)), _mm_add_epi64(chunk1, _b)); \ + _mm_store_si128((__m128i *)((base_ptr) + ((offset) ^ 0x30)), _mm_add_epi64(chunk2, _a)); \ + } while (0) + +# define VARIANT2_SHUFFLE2(base_ptr, offset, _a, _b, _b1, hi, lo) \ + do { \ + const __m128i chunk1 = _mm_xor_si128(_mm_load_si128((__m128i *)((base_ptr) + ((offset) ^ 0x10))), _mm_set_epi64x(lo, hi)); \ + const __m128i chunk2 = _mm_load_si128((__m128i *)((base_ptr) + ((offset) ^ 0x20))); \ + hi ^= ((uint64_t*)((base_ptr) + ((offset) ^ 0x20)))[0]; \ + lo ^= ((uint64_t*)((base_ptr) + ((offset) ^ 0x20)))[1]; \ + const __m128i chunk3 = _mm_load_si128((__m128i *)((base_ptr) + ((offset) ^ 0x30))); \ + _mm_store_si128((__m128i *)((base_ptr) + ((offset) ^ 0x10)), _mm_add_epi64(chunk3, _b1)); \ + _mm_store_si128((__m128i *)((base_ptr) + ((offset) ^ 0x20)), _mm_add_epi64(chunk1, _b)); \ + _mm_store_si128((__m128i *)((base_ptr) + ((offset) ^ 0x30)), _mm_add_epi64(chunk2, _a)); \ + } while (0) + +#else +# define VARIANT2_INIT(part) \ + uint64_t division_result_##part = h##part[12]; \ + uint64_t sqrt_result_##part = h##part[13]; + +# define VARIANT2_INTEGER_MATH(part, cl, cx) \ + do { \ + const uint64_t cx_0 = _mm_cvtsi128_si64(cx); \ + cl ^= division_result_##part ^ (sqrt_result_##part << 32); \ + const uint32_t d = static_cast(cx_0 + (sqrt_result_##part << 1)) | 0x80000001UL; \ + const uint64_t cx_1 = _mm_cvtsi128_si64(_mm_srli_si128(cx, 8)); \ + division_result_##part = static_cast(cx_1 / d) + ((cx_1 % d) << 32); \ + const uint64_t sqrt_input = cx_0 + division_result_##part; \ + sqrt_result_##part = sqrt(sqrt_input + 18446744073709551616.0) * 2.0 - 8589934592.0; \ + const uint64_t s = sqrt_result_##part >> 1; \ + const uint64_t b = sqrt_result_##part & 1; \ + const uint64_t r2 = (uint64_t)(s) * (s + b) + (sqrt_result_##part << 32); \ + sqrt_result_##part += ((r2 + b > sqrt_input) ? -1 : 0) + ((r2 + (1ULL << 32) < sqrt_input - s) ? 1 : 0); \ + } while (0) + +# define VARIANT2_SHUFFLE(base_ptr, offset, _a, _b, _b1) \ + do { \ + const uint64x2_t chunk1 = vld1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x10))); \ + const uint64x2_t chunk2 = vld1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x20))); \ + const uint64x2_t chunk3 = vld1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x30))); \ + vst1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x10)), vaddq_u64(chunk3, vreinterpretq_u64_u8(_b1))); \ + vst1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x20)), vaddq_u64(chunk1, vreinterpretq_u64_u8(_b))); \ + vst1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x30)), vaddq_u64(chunk2, vreinterpretq_u64_u8(_a))); \ + } while (0) + +# define VARIANT2_SHUFFLE2(base_ptr, offset, _a, _b, _b1, hi, lo) \ + do { \ + const uint64x2_t chunk1 = veorq_u64(vld1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x10))), vcombine_u64(vcreate_u64(hi), vcreate_u64(lo))); \ + const uint64x2_t chunk2 = vld1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x20))); \ + hi ^= ((uint64_t*)((base_ptr) + ((offset) ^ 0x20)))[0]; \ + lo ^= ((uint64_t*)((base_ptr) + ((offset) ^ 0x20)))[1]; \ + const uint64x2_t chunk3 = vld1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x30))); \ + vst1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x10)), vaddq_u64(chunk3, vreinterpretq_u64_u8(_b1))); \ + vst1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x20)), vaddq_u64(chunk1, vreinterpretq_u64_u8(_b))); \ + vst1q_u64((uint64_t*)((base_ptr) + ((offset) ^ 0x30)), vaddq_u64(chunk2, vreinterpretq_u64_u8(_a))); \ + } while (0) +#endif +#endif /* XMRIG_CRYPTONIGHT_MONERO_H */ diff --git a/src/crypto/CryptoNight_test.h b/src/crypto/CryptoNight_test.h index 16296efb..95e12197 100644 --- a/src/crypto/CryptoNight_test.h +++ b/src/crypto/CryptoNight_test.h @@ -69,7 +69,7 @@ const static uint8_t test_output_v0[160] = { }; -// Monero v7 +// Cryptonight variant 1 (Monero v7) const static uint8_t test_output_v1[160] = { 0xF2, 0x2D, 0x3D, 0x62, 0x03, 0xD2, 0xA0, 0x8B, 0x41, 0xD9, 0x02, 0x72, 0x78, 0xD8, 0xBC, 0xC9, 0x83, 0xAC, 0xAD, 0xA9, 0xB6, 0x8E, 0x52, 0xE3, 0xC6, 0x89, 0x69, 0x2A, 0x50, 0xE9, 0x21, 0xD9, @@ -84,6 +84,21 @@ const static uint8_t test_output_v1[160] = { }; +// Cryptonight variant 2 (Monero v8) +const static uint8_t test_output_v2[160] = { + 0x97, 0x37, 0x82, 0x82, 0xCF, 0x10, 0xE7, 0xAD, 0x03, 0x3F, 0x7B, 0x80, 0x74, 0xC4, 0x0E, 0x14, + 0xD0, 0x6E, 0x7F, 0x60, 0x9D, 0xDD, 0xDA, 0x78, 0x76, 0x80, 0xB5, 0x8C, 0x05, 0xF4, 0x3D, 0x21, + 0x87, 0x1F, 0xCD, 0x68, 0x23, 0xF6, 0xA8, 0x79, 0xBB, 0x3F, 0x33, 0x95, 0x1C, 0x8E, 0x8E, 0x89, + 0x1D, 0x40, 0x43, 0x88, 0x0B, 0x02, 0xDF, 0xA1, 0xBB, 0x3B, 0xE4, 0x98, 0xB5, 0x0E, 0x75, 0x78, + 0xE6, 0x0D, 0x24, 0x0F, 0x65, 0x85, 0x60, 0x3A, 0x4A, 0xE5, 0x5F, 0x54, 0x9B, 0xC8, 0x79, 0x93, + 0xEB, 0x3D, 0x98, 0x2C, 0xFE, 0x9B, 0xFB, 0x15, 0xB6, 0x88, 0x21, 0x94, 0xB0, 0x05, 0x86, 0x5C, + 0x59, 0x8B, 0x93, 0x7A, 0xDA, 0xD2, 0xA2, 0x14, 0xED, 0xB7, 0xC4, 0x5D, 0xA1, 0xEF, 0x26, 0xF3, + 0xC7, 0x73, 0x29, 0x4D, 0xF1, 0xC8, 0x2C, 0xE0, 0xD0, 0xE9, 0xED, 0x0C, 0x70, 0x75, 0x05, 0x3E, + 0x5B, 0xF6, 0xA0, 0x6E, 0xEA, 0xDE, 0x87, 0x0B, 0x06, 0x29, 0x03, 0xBF, 0xB4, 0x85, 0x9D, 0x04, + 0x75, 0x1A, 0xCD, 0x1E, 0xD6, 0xAA, 0x1B, 0x05, 0x24, 0x6A, 0x2C, 0x80, 0x69, 0x68, 0xDC, 0x97 +}; + + // Stellite (XTL) const static uint8_t test_output_xtl[160] = { 0x8F, 0xE5, 0xF0, 0x5F, 0x02, 0x2A, 0x61, 0x7D, 0xE5, 0x3F, 0x79, 0x36, 0x4B, 0x25, 0xCB, 0xC3, diff --git a/src/crypto/CryptoNight_x86.h b/src/crypto/CryptoNight_x86.h index 199c190b..8dcdd414 100644 --- a/src/crypto/CryptoNight_x86.h +++ b/src/crypto/CryptoNight_x86.h @@ -6,6 +6,7 @@ * Copyright 2016 Jay D Dee * Copyright 2017-2018 XMR-Stak , * Copyright 2018 Lee Clagett + * Copyright 2018 SChernykh * Copyright 2016-2018 XMRig , * * This program is free software: you can redistribute it and/or modify @@ -22,8 +23,8 @@ * along with this program. If not, see . */ -#ifndef __CRYPTONIGHT_X86_H__ -#define __CRYPTONIGHT_X86_H__ +#ifndef XMRIG_CRYPTONIGHT_X86_H +#define XMRIG_CRYPTONIGHT_X86_H #ifdef __GNUC__ @@ -73,10 +74,7 @@ static inline void do_skein_hash(const uint8_t *input, size_t len, uint8_t *outp void (* const extra_hashes[4])(const uint8_t *, size_t, uint8_t *) = {do_blake_hash, do_groestl_hash, do_jh_hash, do_skein_hash}; - #if defined(__x86_64__) || defined(_M_AMD64) -# define EXTRACT64(X) _mm_cvtsi128_si64(X) - # ifdef __GNUC__ static inline uint64_t __umul128(uint64_t a, uint64_t b, uint64_t* hi) { @@ -88,13 +86,14 @@ static inline uint64_t __umul128(uint64_t a, uint64_t b, uint64_t* hi) #define __umul128 _umul128 # endif #elif defined(__i386__) || defined(_M_IX86) -# define HI32(X) \ - _mm_srli_si128((X), 4) +static inline int64_t _mm_cvtsi128_si64(__m128i a) +{ + return ((uint64_t)(uint32_t)_mm_cvtsi128_si32(a) | ((uint64_t)(uint32_t)_mm_cvtsi128_si32(_mm_srli_si128(a, 4)) << 32)); +} - -# define EXTRACT64(X) \ - ((uint64_t)(uint32_t)_mm_cvtsi128_si32(X) | \ - ((uint64_t)(uint32_t)_mm_cvtsi128_si32(HI32(X)) << 32)) +static inline __m128i _mm_cvtsi64_si128(int64_t a) { + return _mm_set_epi64x(0, a); +} static inline uint64_t __umul128(uint64_t multiplier, uint64_t multiplicand, uint64_t *product_hi) { // multiplier = ab = a * 2^32 + b @@ -408,20 +407,46 @@ static inline __m128i aes_round_tweak_div(const __m128i &in, const __m128i &key) } -template -static inline void cryptonight_monero_tweak(uint64_t* mem_out, __m128i tmp) +static inline __m128i int_sqrt_v2(const uint64_t n0) { - mem_out[0] = EXTRACT64(tmp); + __m128d x = _mm_castsi128_pd(_mm_add_epi64(_mm_cvtsi64_si128(n0 >> 12), _mm_set_epi64x(0, 1023ULL << 52))); + x = _mm_sqrt_sd(_mm_setzero_pd(), x); + uint64_t r = static_cast(_mm_cvtsi128_si64(_mm_castpd_si128(x))); - tmp = _mm_castps_si128(_mm_movehl_ps(_mm_castsi128_ps(tmp), _mm_castsi128_ps(tmp))); - uint64_t vh = EXTRACT64(tmp); + const uint64_t s = r >> 20; + r >>= 19; - uint8_t x = vh >> 24; - static const uint16_t table = 0x7531; - const uint8_t index = (((x >> SHIFT) & 6) | (x & 1)) << 1; - vh ^= ((table >> index) & 0x3) << 28; + uint64_t x2 = (s - (1022ULL << 32)) * (r - s - (1022ULL << 32) + 1); +# if (defined(_MSC_VER) || __GNUC__ > 7 || (__GNUC__ == 7 && __GNUC_MINOR__ > 1)) && (defined(__x86_64__) || defined(_M_AMD64)) + _addcarry_u64(_subborrow_u64(0, x2, n0, (unsigned long long int*)&x2), r, 0, (unsigned long long int*)&r); +# else + if (x2 < n0) ++r; +# endif - mem_out[1] = vh; + return _mm_cvtsi64_si128(r); +} + + +template +static inline void cryptonight_monero_tweak(uint64_t* mem_out, const uint8_t* l, uint64_t idx, __m128i ax0, __m128i bx0, __m128i bx1, __m128i cx) +{ + if (VARIANT == xmrig::VARIANT_2) { + VARIANT2_SHUFFLE(l, idx, ax0, bx0, bx1); + _mm_store_si128((__m128i *)mem_out, _mm_xor_si128(bx0, cx)); + } else { + __m128i tmp = _mm_xor_si128(bx0, cx); + mem_out[0] = _mm_cvtsi128_si64(tmp); + + tmp = _mm_castps_si128(_mm_movehl_ps(_mm_castsi128_ps(tmp), _mm_castsi128_ps(tmp))); + uint64_t vh = _mm_cvtsi128_si64(tmp); + + uint8_t x = static_cast(vh >> 24); + static const uint16_t table = 0x7531; + const uint8_t index = (((x >> (VARIANT == xmrig::VARIANT_XTL ? 4 : 3)) & 6) | (x & 1)) << 1; + vh ^= ((table >> index) & 0x3) << 28; + + mem_out[1] = vh; + } } @@ -431,25 +456,28 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si constexpr size_t MASK = xmrig::cn_select_mask(); constexpr size_t ITERATIONS = xmrig::cn_select_iter(); constexpr size_t MEM = xmrig::cn_select_memory(); - constexpr bool IS_MONERO = xmrig::cn_is_monero(); + constexpr bool IS_V1 = xmrig::cn_base_variant() == xmrig::VARIANT_1; - if (IS_MONERO && size < 43) { + if (IS_V1 && size < 43) { memset(output, 0, 32); return; } xmrig::keccak(input, size, ctx[0]->state); - VARIANT1_INIT(0) - cn_explode_scratchpad((__m128i*) ctx[0]->state, (__m128i*) ctx[0]->memory); const uint8_t* l0 = ctx[0]->memory; uint64_t* h0 = reinterpret_cast(ctx[0]->state); + VARIANT1_INIT(0); + VARIANT2_INIT(0); + VARIANT2_SET_ROUNDING_MODE(); + uint64_t al0 = h0[0] ^ h0[4]; uint64_t ah0 = h0[1] ^ h0[5]; __m128i bx0 = _mm_set_epi64x(h0[3] ^ h0[7], h0[2] ^ h0[6]); + __m128i bx1 = _mm_set_epi64x(h0[9] ^ h0[11], h0[8] ^ h0[10]); uint64_t idx0 = al0; @@ -459,44 +487,47 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si cx = _mm_load_si128((__m128i *) &l0[idx0 & MASK]); } + const __m128i ax0 = _mm_set_epi64x(ah0, al0); if (VARIANT == xmrig::VARIANT_TUBE) { - cx = aes_round_tweak_div(cx, _mm_set_epi64x(ah0, al0)); + cx = aes_round_tweak_div(cx, ax0); } else if (SOFT_AES) { - cx = soft_aesenc((uint32_t*)&l0[idx0 & MASK], _mm_set_epi64x(ah0, al0)); + cx = soft_aesenc((uint32_t*)&l0[idx0 & MASK], ax0); } else { - cx = _mm_aesenc_si128(cx, _mm_set_epi64x(ah0, al0)); + cx = _mm_aesenc_si128(cx, ax0); } - if (IS_MONERO) { - cryptonight_monero_tweak((uint64_t*)&l0[idx0 & MASK], _mm_xor_si128(bx0, cx)); + if (IS_V1 || VARIANT == xmrig::VARIANT_2) { + cryptonight_monero_tweak((uint64_t*)&l0[idx0 & MASK], l0, idx0 & MASK, ax0, bx0, bx1, cx); } else { _mm_store_si128((__m128i *)&l0[idx0 & MASK], _mm_xor_si128(bx0, cx)); } - idx0 = EXTRACT64(cx); - bx0 = cx; + idx0 = _mm_cvtsi128_si64(cx); uint64_t hi, lo, cl, ch; cl = ((uint64_t*) &l0[idx0 & MASK])[0]; ch = ((uint64_t*) &l0[idx0 & MASK])[1]; - lo = __umul128(idx0, cl, &hi); + if (VARIANT == xmrig::VARIANT_2) { + VARIANT2_INTEGER_MATH(0, cl, cx); + lo = __umul128(idx0, cl, &hi); + VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx0, bx1, hi, lo); + } + else { + lo = __umul128(idx0, cl, &hi); + } al0 += hi; ah0 += lo; ((uint64_t*)&l0[idx0 & MASK])[0] = al0; - if (IS_MONERO) { - if (VARIANT == xmrig::VARIANT_TUBE || VARIANT == xmrig::VARIANT_RTO) { - ((uint64_t*)&l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0 ^ al0; - } - else { - ((uint64_t*)&l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0; - } - } - else { + if (IS_V1 && (VARIANT == xmrig::VARIANT_TUBE || VARIANT == xmrig::VARIANT_RTO)) { + ((uint64_t*)&l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0 ^ al0; + } else if (IS_V1) { + ((uint64_t*)&l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0; + } else { ((uint64_t*)&l0[idx0 & MASK])[1] = ah0; } @@ -517,6 +548,10 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si idx0 = d ^ q; } + if (VARIANT == xmrig::VARIANT_2) { + bx1 = bx0; + } + bx0 = cx; } cn_implode_scratchpad((__m128i*) ctx[0]->memory, (__m128i*) ctx[0]->state); @@ -526,15 +561,67 @@ inline void cryptonight_single_hash(const uint8_t *__restrict__ input, size_t si } +#ifndef XMRIG_NO_ASM +extern "C" void cnv2_mainloop_ivybridge_asm(cryptonight_ctx *ctx); +extern "C" void cnv2_mainloop_ryzen_asm(cryptonight_ctx *ctx); +extern "C" void cnv2_double_mainloop_sandybridge_asm(cryptonight_ctx* ctx0, cryptonight_ctx* ctx1); + + +template +inline void cryptonight_single_hash_asm(const uint8_t *__restrict__ input, size_t size, uint8_t *__restrict__ output, cryptonight_ctx **__restrict__ ctx) +{ + constexpr size_t MEM = xmrig::cn_select_memory(); + + xmrig::keccak(input, size, ctx[0]->state); + cn_explode_scratchpad(reinterpret_cast<__m128i*>(ctx[0]->state), reinterpret_cast<__m128i*>(ctx[0]->memory)); + + if (ASM == xmrig::ASM_INTEL) { + cnv2_mainloop_ivybridge_asm(ctx[0]); + } + else { + cnv2_mainloop_ryzen_asm(ctx[0]); + } + + cn_implode_scratchpad(reinterpret_cast<__m128i*>(ctx[0]->memory), reinterpret_cast<__m128i*>(ctx[0]->state)); + xmrig::keccakf(reinterpret_cast(ctx[0]->state), 24); + extra_hashes[ctx[0]->state[0] & 3](ctx[0]->state, 200, output); +} + + +template +inline void cryptonight_double_hash_asm(const uint8_t *__restrict__ input, size_t size, uint8_t *__restrict__ output, cryptonight_ctx **__restrict__ ctx) +{ + constexpr size_t MEM = xmrig::cn_select_memory(); + + xmrig::keccak(input, size, ctx[0]->state); + xmrig::keccak(input + size, size, ctx[1]->state); + + cn_explode_scratchpad(reinterpret_cast<__m128i*>(ctx[0]->state), reinterpret_cast<__m128i*>(ctx[0]->memory)); + cn_explode_scratchpad(reinterpret_cast<__m128i*>(ctx[1]->state), reinterpret_cast<__m128i*>(ctx[1]->memory)); + + cnv2_double_mainloop_sandybridge_asm(ctx[0], ctx[1]); + + cn_implode_scratchpad(reinterpret_cast<__m128i*>(ctx[0]->memory), reinterpret_cast<__m128i*>(ctx[0]->state)); + cn_implode_scratchpad(reinterpret_cast<__m128i*>(ctx[1]->memory), reinterpret_cast<__m128i*>(ctx[1]->state)); + + xmrig::keccakf(reinterpret_cast(ctx[0]->state), 24); + xmrig::keccakf(reinterpret_cast(ctx[1]->state), 24); + + extra_hashes[ctx[0]->state[0] & 3](ctx[0]->state, 200, output); + extra_hashes[ctx[1]->state[0] & 3](ctx[1]->state, 200, output + 32); +} +#endif + + template inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t size, uint8_t *__restrict__ output, cryptonight_ctx **__restrict__ ctx) { constexpr size_t MASK = xmrig::cn_select_mask(); constexpr size_t ITERATIONS = xmrig::cn_select_iter(); constexpr size_t MEM = xmrig::cn_select_memory(); - constexpr bool IS_MONERO = xmrig::cn_is_monero(); + constexpr bool IS_V1 = xmrig::cn_base_variant() == xmrig::VARIANT_1; - if (IS_MONERO && size < 43) { + if (IS_V1 && size < 43) { memset(output, 0, 64); return; } @@ -542,14 +629,17 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si xmrig::keccak(input, size, ctx[0]->state); xmrig::keccak(input + size, size, ctx[1]->state); - VARIANT1_INIT(0); - VARIANT1_INIT(1); - const uint8_t* l0 = ctx[0]->memory; const uint8_t* l1 = ctx[1]->memory; uint64_t* h0 = reinterpret_cast(ctx[0]->state); uint64_t* h1 = reinterpret_cast(ctx[1]->state); + VARIANT1_INIT(0); + VARIANT1_INIT(1); + VARIANT2_INIT(0); + VARIANT2_INIT(1); + VARIANT2_SET_ROUNDING_MODE(); + cn_explode_scratchpad((__m128i*) h0, (__m128i*) l0); cn_explode_scratchpad((__m128i*) h1, (__m128i*) l1); @@ -558,8 +648,10 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si uint64_t ah0 = h0[1] ^ h0[5]; uint64_t ah1 = h1[1] ^ h1[5]; - __m128i bx0 = _mm_set_epi64x(h0[3] ^ h0[7], h0[2] ^ h0[6]); - __m128i bx1 = _mm_set_epi64x(h1[3] ^ h1[7], h1[2] ^ h1[6]); + __m128i bx00 = _mm_set_epi64x(h0[3] ^ h0[7], h0[2] ^ h0[6]); + __m128i bx01 = _mm_set_epi64x(h0[9] ^ h0[11], h0[8] ^ h0[10]); + __m128i bx10 = _mm_set_epi64x(h1[3] ^ h1[7], h1[2] ^ h1[6]); + __m128i bx11 = _mm_set_epi64x(h1[9] ^ h1[11], h1[8] ^ h1[10]); uint64_t idx0 = al0; uint64_t idx1 = al1; @@ -571,53 +663,54 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si cx1 = _mm_load_si128((__m128i *) &l1[idx1 & MASK]); } + const __m128i ax0 = _mm_set_epi64x(ah0, al0); + const __m128i ax1 = _mm_set_epi64x(ah1, al1); if (VARIANT == xmrig::VARIANT_TUBE) { - cx0 = aes_round_tweak_div(cx0, _mm_set_epi64x(ah0, al0)); - cx1 = aes_round_tweak_div(cx1, _mm_set_epi64x(ah1, al1)); + cx0 = aes_round_tweak_div(cx0, ax0); + cx1 = aes_round_tweak_div(cx1, ax1); } else if (SOFT_AES) { - cx0 = soft_aesenc((uint32_t*)&l0[idx0 & MASK], _mm_set_epi64x(ah0, al0)); - cx1 = soft_aesenc((uint32_t*)&l1[idx1 & MASK], _mm_set_epi64x(ah1, al1)); + cx0 = soft_aesenc((uint32_t*)&l0[idx0 & MASK], ax0); + cx1 = soft_aesenc((uint32_t*)&l1[idx1 & MASK], ax1); } else { - cx0 = _mm_aesenc_si128(cx0, _mm_set_epi64x(ah0, al0)); - cx1 = _mm_aesenc_si128(cx1, _mm_set_epi64x(ah1, al1)); + cx0 = _mm_aesenc_si128(cx0, ax0); + cx1 = _mm_aesenc_si128(cx1, ax1); } - if (IS_MONERO) { - cryptonight_monero_tweak((uint64_t*)&l0[idx0 & MASK], _mm_xor_si128(bx0, cx0)); - cryptonight_monero_tweak((uint64_t*)&l1[idx1 & MASK], _mm_xor_si128(bx1, cx1)); + if (IS_V1 || (VARIANT == xmrig::VARIANT_2)) { + cryptonight_monero_tweak((uint64_t*)&l0[idx0 & MASK], l0, idx0 & MASK, ax0, bx00, bx01, cx0); + cryptonight_monero_tweak((uint64_t*)&l1[idx1 & MASK], l1, idx1 & MASK, ax1, bx10, bx11, cx1); } else { - _mm_store_si128((__m128i *) &l0[idx0 & MASK], _mm_xor_si128(bx0, cx0)); - _mm_store_si128((__m128i *) &l1[idx1 & MASK], _mm_xor_si128(bx1, cx1)); + _mm_store_si128((__m128i *) &l0[idx0 & MASK], _mm_xor_si128(bx00, cx0)); + _mm_store_si128((__m128i *) &l1[idx1 & MASK], _mm_xor_si128(bx10, cx1)); } - idx0 = EXTRACT64(cx0); - idx1 = EXTRACT64(cx1); - - bx0 = cx0; - bx1 = cx1; + idx0 = _mm_cvtsi128_si64(cx0); + idx1 = _mm_cvtsi128_si64(cx1); uint64_t hi, lo, cl, ch; cl = ((uint64_t*) &l0[idx0 & MASK])[0]; ch = ((uint64_t*) &l0[idx0 & MASK])[1]; - lo = __umul128(idx0, cl, &hi); + if (VARIANT == xmrig::VARIANT_2) { + VARIANT2_INTEGER_MATH(0, cl, cx0); + lo = __umul128(idx0, cl, &hi); + VARIANT2_SHUFFLE2(l0, idx0 & MASK, ax0, bx00, bx01, hi, lo); + } else { + lo = __umul128(idx0, cl, &hi); + } al0 += hi; ah0 += lo; ((uint64_t*)&l0[idx0 & MASK])[0] = al0; - if (IS_MONERO) { - if (VARIANT == xmrig::VARIANT_TUBE || VARIANT == xmrig::VARIANT_RTO) { - ((uint64_t*)&l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0 ^ al0; - } - else { - ((uint64_t*)&l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0; - } - } - else { - ((uint64_t*)&l0[idx0 & MASK])[1] = ah0; + if (IS_V1 && (VARIANT == xmrig::VARIANT_TUBE || VARIANT == xmrig::VARIANT_RTO)) { + ((uint64_t*) &l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0 ^ al0; + } else if (IS_V1) { + ((uint64_t*) &l0[idx0 & MASK])[1] = ah0 ^ tweak1_2_0; + } else { + ((uint64_t*) &l0[idx0 & MASK])[1] = ah0; } al0 ^= cl; @@ -640,22 +733,24 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si cl = ((uint64_t*) &l1[idx1 & MASK])[0]; ch = ((uint64_t*) &l1[idx1 & MASK])[1]; - lo = __umul128(idx1, cl, &hi); + if (VARIANT == xmrig::VARIANT_2) { + VARIANT2_INTEGER_MATH(1, cl, cx1); + lo = __umul128(idx1, cl, &hi); + VARIANT2_SHUFFLE2(l1, idx1 & MASK, ax1, bx10, bx11, hi, lo); + } else { + lo = __umul128(idx1, cl, &hi); + } al1 += hi; ah1 += lo; ((uint64_t*)&l1[idx1 & MASK])[0] = al1; - if (IS_MONERO) { - if (VARIANT == xmrig::VARIANT_TUBE || VARIANT == xmrig::VARIANT_RTO) { - ((uint64_t*)&l1[idx1 & MASK])[1] = ah1 ^ tweak1_2_1 ^ al1; - } - else { - ((uint64_t*)&l1[idx1 & MASK])[1] = ah1 ^ tweak1_2_1; - } - } - else { + if (IS_V1 && (VARIANT == xmrig::VARIANT_TUBE || VARIANT == xmrig::VARIANT_RTO)) { + ((uint64_t*)&l1[idx1 & MASK])[1] = ah1 ^ tweak1_2_1 ^ al1; + } else if (IS_V1) { + ((uint64_t*)&l1[idx1 & MASK])[1] = ah1 ^ tweak1_2_1; + } else { ((uint64_t*)&l1[idx1 & MASK])[1] = ah1; } @@ -676,6 +771,13 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si idx1 = d ^ q; } + + if (VARIANT == xmrig::VARIANT_2) { + bx01 = bx00; + bx11 = bx10; + } + bx00 = cx0; + bx10 = cx1; } cn_implode_scratchpad((__m128i*) l0, (__m128i*) h0); @@ -689,12 +791,12 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si } -#define CN_STEP1(a, b, c, l, ptr, idx) \ +#define CN_STEP1(a, b0, b1, c, l, ptr, idx) \ ptr = reinterpret_cast<__m128i*>(&l[idx & MASK]); \ c = _mm_load_si128(ptr); -#define CN_STEP2(a, b, c, l, ptr, idx) \ +#define CN_STEP2(a, b0, b1, c, l, ptr, idx) \ if (VARIANT == xmrig::VARIANT_TUBE) { \ c = aes_round_tweak_div(c, a); \ } \ @@ -704,26 +806,31 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si c = _mm_aesenc_si128(c, a); \ } \ \ - b = _mm_xor_si128(b, c); \ - \ - if (IS_MONERO) { \ - cryptonight_monero_tweak(reinterpret_cast(ptr), b); \ + if (IS_V1 || (VARIANT == xmrig::VARIANT_2)) { \ + cryptonight_monero_tweak((uint64_t*)ptr, l, idx & MASK, a, b0, b1, c); \ } else { \ - _mm_store_si128(ptr, b); \ + _mm_store_si128(ptr, _mm_xor_si128(b0, c)); \ } -#define CN_STEP3(a, b, c, l, ptr, idx) \ - idx = EXTRACT64(c); \ +#define CN_STEP3(part, a, b0, b1, c, l, ptr, idx) \ + idx = _mm_cvtsi128_si64(c); \ ptr = reinterpret_cast<__m128i*>(&l[idx & MASK]); \ - b = _mm_load_si128(ptr); + uint64_t cl##part = ((uint64_t*)ptr)[0]; \ + uint64_t ch##part = ((uint64_t*)ptr)[1]; -#define CN_STEP4(a, b, c, l, mc, ptr, idx) \ - lo = __umul128(idx, EXTRACT64(b), &hi); \ +#define CN_STEP4(part, a, b0, b1, c, l, mc, ptr, idx) \ + if (VARIANT == xmrig::VARIANT_2) { \ + VARIANT2_INTEGER_MATH(part, cl##part, c); \ + lo = __umul128(idx, cl##part, &hi); \ + VARIANT2_SHUFFLE2(l, idx & MASK, a, b0, b1, hi, lo); \ + } else { \ + lo = __umul128(idx, cl##part, &hi); \ + } \ a = _mm_add_epi64(a, _mm_set_epi64x(lo, hi)); \ \ - if (IS_MONERO) { \ + if (IS_V1) { \ _mm_store_si128(ptr, _mm_xor_si128(a, mc)); \ \ if (VARIANT == xmrig::VARIANT_TUBE || \ @@ -734,8 +841,8 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si _mm_store_si128(ptr, a); \ } \ \ - a = _mm_xor_si128(a, b); \ - idx = EXTRACT64(a); \ + a = _mm_xor_si128(a, _mm_set_epi64x(ch##part, cl##part)); \ + idx = _mm_cvtsi128_si64(a); \ \ if (ALGO == xmrig::CRYPTONIGHT_HEAVY) { \ int64_t n = ((int64_t*)&l[idx & MASK])[0]; \ @@ -747,15 +854,29 @@ inline void cryptonight_double_hash(const uint8_t *__restrict__ input, size_t si } \ \ idx = d ^ q; \ - } + } \ + if (VARIANT == xmrig::VARIANT_2) { \ + b1 = b0; \ + } \ + b0 = c; #define CONST_INIT(ctx, n) \ __m128i mc##n; \ - if (IS_MONERO) { \ + __m128i division_result_xmm_##n; \ + __m128i sqrt_result_xmm_##n; \ + if (IS_V1) { \ mc##n = _mm_set_epi64x(*reinterpret_cast(input + n * size + 35) ^ \ *(reinterpret_cast((ctx)->state) + 24), 0); \ - } + } \ + if (VARIANT == xmrig::VARIANT_2) { \ + division_result_xmm_##n = _mm_cvtsi64_si128(h##n[12]); \ + sqrt_result_xmm_##n = _mm_cvtsi64_si128(h##n[13]); \ + } \ + __m128i ax##n = _mm_set_epi64x(h##n[1] ^ h##n[5], h##n[0] ^ h##n[4]); \ + __m128i bx##n##0 = _mm_set_epi64x(h##n[3] ^ h##n[7], h##n[2] ^ h##n[6]); \ + __m128i bx##n##1 = _mm_set_epi64x(h##n[9] ^ h##n[11], h##n[8] ^ h##n[10]); \ + __m128i cx##n = _mm_setzero_si128(); template @@ -764,9 +885,9 @@ inline void cryptonight_triple_hash(const uint8_t *__restrict__ input, size_t si constexpr size_t MASK = xmrig::cn_select_mask(); constexpr size_t ITERATIONS = xmrig::cn_select_iter(); constexpr size_t MEM = xmrig::cn_select_memory(); - constexpr bool IS_MONERO = xmrig::cn_is_monero(); + constexpr bool IS_V1 = xmrig::cn_base_variant() == xmrig::VARIANT_1; - if (IS_MONERO && size < 43) { + if (IS_V1 && size < 43) { memset(output, 0, 32 * 3); return; } @@ -776,10 +897,6 @@ inline void cryptonight_triple_hash(const uint8_t *__restrict__ input, size_t si cn_explode_scratchpad(reinterpret_cast<__m128i*>(ctx[i]->state), reinterpret_cast<__m128i*>(ctx[i]->memory)); } - CONST_INIT(ctx[0], 0); - CONST_INIT(ctx[1], 1); - CONST_INIT(ctx[2], 2); - uint8_t* l0 = ctx[0]->memory; uint8_t* l1 = ctx[1]->memory; uint8_t* l2 = ctx[2]->memory; @@ -787,58 +904,35 @@ inline void cryptonight_triple_hash(const uint8_t *__restrict__ input, size_t si uint64_t* h1 = reinterpret_cast(ctx[1]->state); uint64_t* h2 = reinterpret_cast(ctx[2]->state); - __m128i ax0 = _mm_set_epi64x(h0[1] ^ h0[5], h0[0] ^ h0[4]); - __m128i bx0 = _mm_set_epi64x(h0[3] ^ h0[7], h0[2] ^ h0[6]); - __m128i ax1 = _mm_set_epi64x(h1[1] ^ h1[5], h1[0] ^ h1[4]); - __m128i bx1 = _mm_set_epi64x(h1[3] ^ h1[7], h1[2] ^ h1[6]); - __m128i ax2 = _mm_set_epi64x(h2[1] ^ h2[5], h2[0] ^ h2[4]); - __m128i bx2 = _mm_set_epi64x(h2[3] ^ h2[7], h2[2] ^ h2[6]); - __m128i cx0 = _mm_set_epi64x(0, 0); - __m128i cx1 = _mm_set_epi64x(0, 0); - __m128i cx2 = _mm_set_epi64x(0, 0); + CONST_INIT(ctx[0], 0); + CONST_INIT(ctx[1], 1); + CONST_INIT(ctx[2], 2); + VARIANT2_SET_ROUNDING_MODE(); uint64_t idx0, idx1, idx2; - idx0 = EXTRACT64(ax0); - idx1 = EXTRACT64(ax1); - idx2 = EXTRACT64(ax2); + idx0 = _mm_cvtsi128_si64(ax0); + idx1 = _mm_cvtsi128_si64(ax1); + idx2 = _mm_cvtsi128_si64(ax2); - for (size_t i = 0; i < ITERATIONS / 2; i++) { + for (size_t i = 0; i < ITERATIONS; i++) { uint64_t hi, lo; __m128i *ptr0, *ptr1, *ptr2; - // EVEN ROUND - CN_STEP1(ax0, bx0, cx0, l0, ptr0, idx0); - CN_STEP1(ax1, bx1, cx1, l1, ptr1, idx1); - CN_STEP1(ax2, bx2, cx2, l2, ptr2, idx2); + CN_STEP1(ax0, bx00, bx01, cx0, l0, ptr0, idx0); + CN_STEP1(ax1, bx10, bx11, cx1, l1, ptr1, idx1); + CN_STEP1(ax2, bx20, bx21, cx2, l2, ptr2, idx2); - CN_STEP2(ax0, bx0, cx0, l0, ptr0, idx0); - CN_STEP2(ax1, bx1, cx1, l1, ptr1, idx1); - CN_STEP2(ax2, bx2, cx2, l2, ptr2, idx2); + CN_STEP2(ax0, bx00, bx01, cx0, l0, ptr0, idx0); + CN_STEP2(ax1, bx10, bx11, cx1, l1, ptr1, idx1); + CN_STEP2(ax2, bx20, bx21, cx2, l2, ptr2, idx2); - CN_STEP3(ax0, bx0, cx0, l0, ptr0, idx0); - CN_STEP3(ax1, bx1, cx1, l1, ptr1, idx1); - CN_STEP3(ax2, bx2, cx2, l2, ptr2, idx2); + CN_STEP3(0, ax0, bx00, bx01, cx0, l0, ptr0, idx0); + CN_STEP3(1, ax1, bx10, bx11, cx1, l1, ptr1, idx1); + CN_STEP3(2, ax2, bx20, bx21, cx2, l2, ptr2, idx2); - CN_STEP4(ax0, bx0, cx0, l0, mc0, ptr0, idx0); - CN_STEP4(ax1, bx1, cx1, l1, mc1, ptr1, idx1); - CN_STEP4(ax2, bx2, cx2, l2, mc2, ptr2, idx2); - - // ODD ROUND - CN_STEP1(ax0, cx0, bx0, l0, ptr0, idx0); - CN_STEP1(ax1, cx1, bx1, l1, ptr1, idx1); - CN_STEP1(ax2, cx2, bx2, l2, ptr2, idx2); - - CN_STEP2(ax0, cx0, bx0, l0, ptr0, idx0); - CN_STEP2(ax1, cx1, bx1, l1, ptr1, idx1); - CN_STEP2(ax2, cx2, bx2, l2, ptr2, idx2); - - CN_STEP3(ax0, cx0, bx0, l0, ptr0, idx0); - CN_STEP3(ax1, cx1, bx1, l1, ptr1, idx1); - CN_STEP3(ax2, cx2, bx2, l2, ptr2, idx2); - - CN_STEP4(ax0, cx0, bx0, l0, mc0, ptr0, idx0); - CN_STEP4(ax1, cx1, bx1, l1, mc1, ptr1, idx1); - CN_STEP4(ax2, cx2, bx2, l2, mc2, ptr2, idx2); + CN_STEP4(0, ax0, bx00, bx01, cx0, l0, mc0, ptr0, idx0); + CN_STEP4(1, ax1, bx10, bx11, cx1, l1, mc1, ptr1, idx1); + CN_STEP4(2, ax2, bx20, bx21, cx2, l2, mc2, ptr2, idx2); } for (size_t i = 0; i < 3; i++) { @@ -855,9 +949,9 @@ inline void cryptonight_quad_hash(const uint8_t *__restrict__ input, size_t size constexpr size_t MASK = xmrig::cn_select_mask(); constexpr size_t ITERATIONS = xmrig::cn_select_iter(); constexpr size_t MEM = xmrig::cn_select_memory(); - constexpr bool IS_MONERO = xmrig::cn_is_monero(); + constexpr bool IS_V1 = xmrig::cn_base_variant() == xmrig::VARIANT_1;; - if (IS_MONERO && size < 43) { + if (IS_V1 && size < 43) { memset(output, 0, 32 * 4); return; } @@ -867,11 +961,6 @@ inline void cryptonight_quad_hash(const uint8_t *__restrict__ input, size_t size cn_explode_scratchpad(reinterpret_cast<__m128i*>(ctx[i]->state), reinterpret_cast<__m128i*>(ctx[i]->memory)); } - CONST_INIT(ctx[0], 0); - CONST_INIT(ctx[1], 1); - CONST_INIT(ctx[2], 2); - CONST_INIT(ctx[3], 3); - uint8_t* l0 = ctx[0]->memory; uint8_t* l1 = ctx[1]->memory; uint8_t* l2 = ctx[2]->memory; @@ -881,71 +970,42 @@ inline void cryptonight_quad_hash(const uint8_t *__restrict__ input, size_t size uint64_t* h2 = reinterpret_cast(ctx[2]->state); uint64_t* h3 = reinterpret_cast(ctx[3]->state); - __m128i ax0 = _mm_set_epi64x(h0[1] ^ h0[5], h0[0] ^ h0[4]); - __m128i bx0 = _mm_set_epi64x(h0[3] ^ h0[7], h0[2] ^ h0[6]); - __m128i ax1 = _mm_set_epi64x(h1[1] ^ h1[5], h1[0] ^ h1[4]); - __m128i bx1 = _mm_set_epi64x(h1[3] ^ h1[7], h1[2] ^ h1[6]); - __m128i ax2 = _mm_set_epi64x(h2[1] ^ h2[5], h2[0] ^ h2[4]); - __m128i bx2 = _mm_set_epi64x(h2[3] ^ h2[7], h2[2] ^ h2[6]); - __m128i ax3 = _mm_set_epi64x(h3[1] ^ h3[5], h3[0] ^ h3[4]); - __m128i bx3 = _mm_set_epi64x(h3[3] ^ h3[7], h3[2] ^ h3[6]); - __m128i cx0 = _mm_set_epi64x(0, 0); - __m128i cx1 = _mm_set_epi64x(0, 0); - __m128i cx2 = _mm_set_epi64x(0, 0); - __m128i cx3 = _mm_set_epi64x(0, 0); + CONST_INIT(ctx[0], 0); + CONST_INIT(ctx[1], 1); + CONST_INIT(ctx[2], 2); + CONST_INIT(ctx[3], 3); + VARIANT2_SET_ROUNDING_MODE(); uint64_t idx0, idx1, idx2, idx3; - idx0 = EXTRACT64(ax0); - idx1 = EXTRACT64(ax1); - idx2 = EXTRACT64(ax2); - idx3 = EXTRACT64(ax3); + idx0 = _mm_cvtsi128_si64(ax0); + idx1 = _mm_cvtsi128_si64(ax1); + idx2 = _mm_cvtsi128_si64(ax2); + idx3 = _mm_cvtsi128_si64(ax3); - for (size_t i = 0; i < ITERATIONS / 2; i++) + for (size_t i = 0; i < ITERATIONS; i++) { uint64_t hi, lo; __m128i *ptr0, *ptr1, *ptr2, *ptr3; - // EVEN ROUND - CN_STEP1(ax0, bx0, cx0, l0, ptr0, idx0); - CN_STEP1(ax1, bx1, cx1, l1, ptr1, idx1); - CN_STEP1(ax2, bx2, cx2, l2, ptr2, idx2); - CN_STEP1(ax3, bx3, cx3, l3, ptr3, idx3); + CN_STEP1(ax0, bx00, bx01, cx0, l0, ptr0, idx0); + CN_STEP1(ax1, bx10, bx11, cx1, l1, ptr1, idx1); + CN_STEP1(ax2, bx20, bx21, cx2, l2, ptr2, idx2); + CN_STEP1(ax3, bx30, bx31, cx3, l3, ptr3, idx3); - CN_STEP2(ax0, bx0, cx0, l0, ptr0, idx0); - CN_STEP2(ax1, bx1, cx1, l1, ptr1, idx1); - CN_STEP2(ax2, bx2, cx2, l2, ptr2, idx2); - CN_STEP2(ax3, bx3, cx3, l3, ptr3, idx3); + CN_STEP2(ax0, bx00, bx01, cx0, l0, ptr0, idx0); + CN_STEP2(ax1, bx10, bx11, cx1, l1, ptr1, idx1); + CN_STEP2(ax2, bx20, bx21, cx2, l2, ptr2, idx2); + CN_STEP2(ax3, bx30, bx31, cx3, l3, ptr3, idx3); - CN_STEP3(ax0, bx0, cx0, l0, ptr0, idx0); - CN_STEP3(ax1, bx1, cx1, l1, ptr1, idx1); - CN_STEP3(ax2, bx2, cx2, l2, ptr2, idx2); - CN_STEP3(ax3, bx3, cx3, l3, ptr3, idx3); + CN_STEP3(0, ax0, bx00, bx01, cx0, l0, ptr0, idx0); + CN_STEP3(1, ax1, bx10, bx11, cx1, l1, ptr1, idx1); + CN_STEP3(2, ax2, bx20, bx21, cx2, l2, ptr2, idx2); + CN_STEP3(3, ax3, bx30, bx31, cx3, l3, ptr3, idx3); - CN_STEP4(ax0, bx0, cx0, l0, mc0, ptr0, idx0); - CN_STEP4(ax1, bx1, cx1, l1, mc1, ptr1, idx1); - CN_STEP4(ax2, bx2, cx2, l2, mc2, ptr2, idx2); - CN_STEP4(ax3, bx3, cx3, l3, mc3, ptr3, idx3); - - // ODD ROUND - CN_STEP1(ax0, cx0, bx0, l0, ptr0, idx0); - CN_STEP1(ax1, cx1, bx1, l1, ptr1, idx1); - CN_STEP1(ax2, cx2, bx2, l2, ptr2, idx2); - CN_STEP1(ax3, cx3, bx3, l3, ptr3, idx3); - - CN_STEP2(ax0, cx0, bx0, l0, ptr0, idx0); - CN_STEP2(ax1, cx1, bx1, l1, ptr1, idx1); - CN_STEP2(ax2, cx2, bx2, l2, ptr2, idx2); - CN_STEP2(ax3, cx3, bx3, l3, ptr3, idx3); - - CN_STEP3(ax0, cx0, bx0, l0, ptr0, idx0); - CN_STEP3(ax1, cx1, bx1, l1, ptr1, idx1); - CN_STEP3(ax2, cx2, bx2, l2, ptr2, idx2); - CN_STEP3(ax3, cx3, bx3, l3, ptr3, idx3); - - CN_STEP4(ax0, cx0, bx0, l0, mc0, ptr0, idx0); - CN_STEP4(ax1, cx1, bx1, l1, mc1, ptr1, idx1); - CN_STEP4(ax2, cx2, bx2, l2, mc2, ptr2, idx2); - CN_STEP4(ax3, cx3, bx3, l3, mc3, ptr3, idx3); + CN_STEP4(0, ax0, bx00, bx01, cx0, l0, mc0, ptr0, idx0); + CN_STEP4(1, ax1, bx10, bx11, cx1, l1, mc1, ptr1, idx1); + CN_STEP4(2, ax2, bx20, bx21, cx2, l2, mc2, ptr2, idx2); + CN_STEP4(3, ax3, bx30, bx31, cx3, l3, mc3, ptr3, idx3); } for (size_t i = 0; i < 4; i++) { @@ -962,9 +1022,9 @@ inline void cryptonight_penta_hash(const uint8_t *__restrict__ input, size_t siz constexpr size_t MASK = xmrig::cn_select_mask(); constexpr size_t ITERATIONS = xmrig::cn_select_iter(); constexpr size_t MEM = xmrig::cn_select_memory(); - constexpr bool IS_MONERO = xmrig::cn_is_monero(); + constexpr bool IS_V1 = xmrig::cn_base_variant() == xmrig::VARIANT_1; - if (IS_MONERO && size < 43) { + if (IS_V1 && size < 43) { memset(output, 0, 32 * 5); return; } @@ -974,12 +1034,6 @@ inline void cryptonight_penta_hash(const uint8_t *__restrict__ input, size_t siz cn_explode_scratchpad(reinterpret_cast<__m128i*>(ctx[i]->state), reinterpret_cast<__m128i*>(ctx[i]->memory)); } - CONST_INIT(ctx[0], 0); - CONST_INIT(ctx[1], 1); - CONST_INIT(ctx[2], 2); - CONST_INIT(ctx[3], 3); - CONST_INIT(ctx[4], 4); - uint8_t* l0 = ctx[0]->memory; uint8_t* l1 = ctx[1]->memory; uint8_t* l2 = ctx[2]->memory; @@ -991,83 +1045,48 @@ inline void cryptonight_penta_hash(const uint8_t *__restrict__ input, size_t siz uint64_t* h3 = reinterpret_cast(ctx[3]->state); uint64_t* h4 = reinterpret_cast(ctx[4]->state); - __m128i ax0 = _mm_set_epi64x(h0[1] ^ h0[5], h0[0] ^ h0[4]); - __m128i bx0 = _mm_set_epi64x(h0[3] ^ h0[7], h0[2] ^ h0[6]); - __m128i ax1 = _mm_set_epi64x(h1[1] ^ h1[5], h1[0] ^ h1[4]); - __m128i bx1 = _mm_set_epi64x(h1[3] ^ h1[7], h1[2] ^ h1[6]); - __m128i ax2 = _mm_set_epi64x(h2[1] ^ h2[5], h2[0] ^ h2[4]); - __m128i bx2 = _mm_set_epi64x(h2[3] ^ h2[7], h2[2] ^ h2[6]); - __m128i ax3 = _mm_set_epi64x(h3[1] ^ h3[5], h3[0] ^ h3[4]); - __m128i bx3 = _mm_set_epi64x(h3[3] ^ h3[7], h3[2] ^ h3[6]); - __m128i ax4 = _mm_set_epi64x(h4[1] ^ h4[5], h4[0] ^ h4[4]); - __m128i bx4 = _mm_set_epi64x(h4[3] ^ h4[7], h4[2] ^ h4[6]); - __m128i cx0 = _mm_set_epi64x(0, 0); - __m128i cx1 = _mm_set_epi64x(0, 0); - __m128i cx2 = _mm_set_epi64x(0, 0); - __m128i cx3 = _mm_set_epi64x(0, 0); - __m128i cx4 = _mm_set_epi64x(0, 0); + CONST_INIT(ctx[0], 0); + CONST_INIT(ctx[1], 1); + CONST_INIT(ctx[2], 2); + CONST_INIT(ctx[3], 3); + CONST_INIT(ctx[4], 4); + VARIANT2_SET_ROUNDING_MODE(); uint64_t idx0, idx1, idx2, idx3, idx4; - idx0 = EXTRACT64(ax0); - idx1 = EXTRACT64(ax1); - idx2 = EXTRACT64(ax2); - idx3 = EXTRACT64(ax3); - idx4 = EXTRACT64(ax4); + idx0 = _mm_cvtsi128_si64(ax0); + idx1 = _mm_cvtsi128_si64(ax1); + idx2 = _mm_cvtsi128_si64(ax2); + idx3 = _mm_cvtsi128_si64(ax3); + idx4 = _mm_cvtsi128_si64(ax4); - for (size_t i = 0; i < ITERATIONS / 2; i++) + for (size_t i = 0; i < ITERATIONS; i++) { uint64_t hi, lo; __m128i *ptr0, *ptr1, *ptr2, *ptr3, *ptr4; - // EVEN ROUND - CN_STEP1(ax0, bx0, cx0, l0, ptr0, idx0); - CN_STEP1(ax1, bx1, cx1, l1, ptr1, idx1); - CN_STEP1(ax2, bx2, cx2, l2, ptr2, idx2); - CN_STEP1(ax3, bx3, cx3, l3, ptr3, idx3); - CN_STEP1(ax4, bx4, cx4, l4, ptr4, idx4); + CN_STEP1(ax0, bx00, bx01, cx0, l0, ptr0, idx0); + CN_STEP1(ax1, bx10, bx11, cx1, l1, ptr1, idx1); + CN_STEP1(ax2, bx20, bx21, cx2, l2, ptr2, idx2); + CN_STEP1(ax3, bx30, bx31, cx3, l3, ptr3, idx3); + CN_STEP1(ax4, bx40, bx41, cx4, l4, ptr4, idx4); - CN_STEP2(ax0, bx0, cx0, l0, ptr0, idx0); - CN_STEP2(ax1, bx1, cx1, l1, ptr1, idx1); - CN_STEP2(ax2, bx2, cx2, l2, ptr2, idx2); - CN_STEP2(ax3, bx3, cx3, l3, ptr3, idx3); - CN_STEP2(ax4, bx4, cx4, l4, ptr4, idx4); + CN_STEP2(ax0, bx00, bx01, cx0, l0, ptr0, idx0); + CN_STEP2(ax1, bx10, bx11, cx1, l1, ptr1, idx1); + CN_STEP2(ax2, bx20, bx21, cx2, l2, ptr2, idx2); + CN_STEP2(ax3, bx30, bx31, cx3, l3, ptr3, idx3); + CN_STEP2(ax4, bx40, bx41, cx4, l4, ptr4, idx4); - CN_STEP3(ax0, bx0, cx0, l0, ptr0, idx0); - CN_STEP3(ax1, bx1, cx1, l1, ptr1, idx1); - CN_STEP3(ax2, bx2, cx2, l2, ptr2, idx2); - CN_STEP3(ax3, bx3, cx3, l3, ptr3, idx3); - CN_STEP3(ax4, bx4, cx4, l4, ptr4, idx4); + CN_STEP3(0, ax0, bx00, bx01, cx0, l0, ptr0, idx0); + CN_STEP3(1, ax1, bx10, bx11, cx1, l1, ptr1, idx1); + CN_STEP3(2, ax2, bx20, bx21, cx2, l2, ptr2, idx2); + CN_STEP3(3, ax3, bx30, bx31, cx3, l3, ptr3, idx3); + CN_STEP3(4, ax4, bx40, bx41, cx4, l4, ptr4, idx4); - CN_STEP4(ax0, bx0, cx0, l0, mc0, ptr0, idx0); - CN_STEP4(ax1, bx1, cx1, l1, mc1, ptr1, idx1); - CN_STEP4(ax2, bx2, cx2, l2, mc2, ptr2, idx2); - CN_STEP4(ax3, bx3, cx3, l3, mc3, ptr3, idx3); - CN_STEP4(ax4, bx4, cx4, l4, mc4, ptr4, idx4); - - // ODD ROUND - CN_STEP1(ax0, cx0, bx0, l0, ptr0, idx0); - CN_STEP1(ax1, cx1, bx1, l1, ptr1, idx1); - CN_STEP1(ax2, cx2, bx2, l2, ptr2, idx2); - CN_STEP1(ax3, cx3, bx3, l3, ptr3, idx3); - CN_STEP1(ax4, cx4, bx4, l4, ptr4, idx4); - - CN_STEP2(ax0, cx0, bx0, l0, ptr0, idx0); - CN_STEP2(ax1, cx1, bx1, l1, ptr1, idx1); - CN_STEP2(ax2, cx2, bx2, l2, ptr2, idx2); - CN_STEP2(ax3, cx3, bx3, l3, ptr3, idx3); - CN_STEP2(ax4, cx4, bx4, l4, ptr4, idx4); - - CN_STEP3(ax0, cx0, bx0, l0, ptr0, idx0); - CN_STEP3(ax1, cx1, bx1, l1, ptr1, idx1); - CN_STEP3(ax2, cx2, bx2, l2, ptr2, idx2); - CN_STEP3(ax3, cx3, bx3, l3, ptr3, idx3); - CN_STEP3(ax4, cx4, bx4, l4, ptr4, idx4); - - CN_STEP4(ax0, cx0, bx0, l0, mc0, ptr0, idx0); - CN_STEP4(ax1, cx1, bx1, l1, mc1, ptr1, idx1); - CN_STEP4(ax2, cx2, bx2, l2, mc2, ptr2, idx2); - CN_STEP4(ax3, cx3, bx3, l3, mc3, ptr3, idx3); - CN_STEP4(ax4, cx4, bx4, l4, mc4, ptr4, idx4); + CN_STEP4(0, ax0, bx00, bx01, cx0, l0, mc0, ptr0, idx0); + CN_STEP4(1, ax1, bx10, bx11, cx1, l1, mc1, ptr1, idx1); + CN_STEP4(2, ax2, bx20, bx21, cx2, l2, mc2, ptr2, idx2); + CN_STEP4(3, ax3, bx30, bx31, cx3, l3, mc3, ptr3, idx3); + CN_STEP4(4, ax4, bx40, bx41, cx4, l4, mc4, ptr4, idx4); } for (size_t i = 0; i < 5; i++) { @@ -1077,4 +1096,4 @@ inline void cryptonight_penta_hash(const uint8_t *__restrict__ input, size_t siz } } -#endif /* __CRYPTONIGHT_X86_H__ */ +#endif /* XMRIG_CRYPTONIGHT_X86_H */ diff --git a/src/crypto/asm/cnv2_double_main_loop_sandybridge.inc b/src/crypto/asm/cnv2_double_main_loop_sandybridge.inc new file mode 100644 index 00000000..e8251bc7 --- /dev/null +++ b/src/crypto/asm/cnv2_double_main_loop_sandybridge.inc @@ -0,0 +1,410 @@ + mov rax, rsp + push rbx + push rbp + push rsi + push rdi + push r12 + push r13 + push r14 + push r15 + sub rsp, 184 + + stmxcsr DWORD PTR [rsp+272] + mov DWORD PTR [rsp+276], 24448 + ldmxcsr DWORD PTR [rsp+276] + + mov r13, QWORD PTR [rcx+224] + mov r9, rdx + mov r10, QWORD PTR [rcx+32] + mov r8, rcx + xor r10, QWORD PTR [rcx] + mov r14d, 524288 + mov r11, QWORD PTR [rcx+40] + xor r11, QWORD PTR [rcx+8] + mov rsi, QWORD PTR [rdx+224] + mov rdx, QWORD PTR [rcx+56] + xor rdx, QWORD PTR [rcx+24] + mov rdi, QWORD PTR [r9+32] + xor rdi, QWORD PTR [r9] + mov rbp, QWORD PTR [r9+40] + xor rbp, QWORD PTR [r9+8] + movq xmm0, rdx + movaps XMMWORD PTR [rax-88], xmm6 + movaps XMMWORD PTR [rax-104], xmm7 + movaps XMMWORD PTR [rax-120], xmm8 + movaps XMMWORD PTR [rsp+112], xmm9 + movaps XMMWORD PTR [rsp+96], xmm10 + movaps XMMWORD PTR [rsp+80], xmm11 + movaps XMMWORD PTR [rsp+64], xmm12 + movaps XMMWORD PTR [rsp+48], xmm13 + movaps XMMWORD PTR [rsp+32], xmm14 + movaps XMMWORD PTR [rsp+16], xmm15 + mov rdx, r10 + movq xmm4, QWORD PTR [r8+96] + and edx, 2097136 + mov rax, QWORD PTR [rcx+48] + xorps xmm13, xmm13 + xor rax, QWORD PTR [rcx+16] + mov rcx, QWORD PTR [rcx+88] + xor rcx, QWORD PTR [r8+72] + movq xmm5, QWORD PTR [r8+104] + movq xmm7, rax + + mov eax, 1 + shl rax, 52 + movq xmm14, rax + punpcklqdq xmm14, xmm14 + + mov eax, 1023 + shl rax, 52 + movq xmm12, rax + punpcklqdq xmm12, xmm12 + + mov rax, QWORD PTR [r8+80] + xor rax, QWORD PTR [r8+64] + punpcklqdq xmm7, xmm0 + movq xmm0, rcx + mov rcx, QWORD PTR [r9+56] + xor rcx, QWORD PTR [r9+24] + movq xmm3, rax + mov rax, QWORD PTR [r9+48] + xor rax, QWORD PTR [r9+16] + punpcklqdq xmm3, xmm0 + movq xmm0, rcx + mov QWORD PTR [rsp], r13 + mov rcx, QWORD PTR [r9+88] + xor rcx, QWORD PTR [r9+72] + movq xmm6, rax + mov rax, QWORD PTR [r9+80] + xor rax, QWORD PTR [r9+64] + punpcklqdq xmm6, xmm0 + movq xmm0, rcx + mov QWORD PTR [rsp+256], r10 + mov rcx, rdi + mov QWORD PTR [rsp+264], r11 + movq xmm8, rax + and ecx, 2097136 + punpcklqdq xmm8, xmm0 + movq xmm0, QWORD PTR [r9+96] + punpcklqdq xmm4, xmm0 + movq xmm0, QWORD PTR [r9+104] + lea r8, QWORD PTR [rcx+rsi] + movdqu xmm11, XMMWORD PTR [r8] + punpcklqdq xmm5, xmm0 + lea r9, QWORD PTR [rdx+r13] + movdqu xmm15, XMMWORD PTR [r9] + + ALIGN 16 +main_loop_double_sandybridge: + movdqu xmm9, xmm15 + mov eax, edx + mov ebx, edx + xor eax, 16 + xor ebx, 32 + xor edx, 48 + + movq xmm0, r11 + movq xmm2, r10 + punpcklqdq xmm2, xmm0 + aesenc xmm9, xmm2 + + movdqu xmm0, XMMWORD PTR [rax+r13] + movdqu xmm1, XMMWORD PTR [rbx+r13] + paddq xmm0, xmm7 + paddq xmm1, xmm2 + movdqu XMMWORD PTR [rbx+r13], xmm0 + movdqu xmm0, XMMWORD PTR [rdx+r13] + movdqu XMMWORD PTR [rdx+r13], xmm1 + paddq xmm0, xmm3 + movdqu XMMWORD PTR [rax+r13], xmm0 + + movq r11, xmm9 + mov edx, r11d + and edx, 2097136 + movdqa xmm0, xmm9 + pxor xmm0, xmm7 + movdqu XMMWORD PTR [r9], xmm0 + + lea rbx, QWORD PTR [rdx+r13] + mov r10, QWORD PTR [rdx+r13] + + movdqu xmm10, xmm11 + movq xmm0, rbp + movq xmm11, rdi + punpcklqdq xmm11, xmm0 + aesenc xmm10, xmm11 + + mov eax, ecx + mov r12d, ecx + xor eax, 16 + xor r12d, 32 + xor ecx, 48 + + movdqu xmm0, XMMWORD PTR [rax+rsi] + paddq xmm0, xmm6 + movdqu xmm1, XMMWORD PTR [r12+rsi] + movdqu XMMWORD PTR [r12+rsi], xmm0 + paddq xmm1, xmm11 + movdqu xmm0, XMMWORD PTR [rcx+rsi] + movdqu XMMWORD PTR [rcx+rsi], xmm1 + paddq xmm0, xmm8 + movdqu XMMWORD PTR [rax+rsi], xmm0 + + movq rcx, xmm10 + and ecx, 2097136 + + movdqa xmm0, xmm10 + pxor xmm0, xmm6 + movdqu XMMWORD PTR [r8], xmm0 + mov r12, QWORD PTR [rcx+rsi] + + mov r9, QWORD PTR [rbx+8] + + xor edx, 16 + mov r8d, edx + mov r15d, edx + + movq rdx, xmm5 + shl rdx, 32 + movq rax, xmm4 + xor rdx, rax + xor r10, rdx + mov rax, r10 + mul r11 + mov r11d, r8d + xor r11d, 48 + movq xmm0, rdx + xor rdx, [r11+r13] + movq xmm1, rax + xor rax, [r11+r13+8] + punpcklqdq xmm0, xmm1 + + pxor xmm0, XMMWORD PTR [r8+r13] + xor r8d, 32 + movdqu xmm1, XMMWORD PTR [r11+r13] + paddq xmm0, xmm7 + paddq xmm1, xmm2 + movdqu XMMWORD PTR [r11+r13], xmm0 + movdqu xmm0, XMMWORD PTR [r8+r13] + movdqu XMMWORD PTR [r8+r13], xmm1 + paddq xmm0, xmm3 + movdqu XMMWORD PTR [r15+r13], xmm0 + + mov r11, QWORD PTR [rsp+256] + add r11, rdx + mov rdx, QWORD PTR [rsp+264] + add rdx, rax + mov QWORD PTR [rbx], r11 + xor r11, r10 + mov QWORD PTR [rbx+8], rdx + xor rdx, r9 + mov QWORD PTR [rsp+256], r11 + and r11d, 2097136 + mov QWORD PTR [rsp+264], rdx + mov QWORD PTR [rsp+8], r11 + lea r15, QWORD PTR [r11+r13] + movdqu xmm15, XMMWORD PTR [r11+r13] + lea r13, QWORD PTR [rsi+rcx] + movdqa xmm0, xmm5 + psrldq xmm0, 8 + movaps xmm2, xmm13 + movq r10, xmm0 + psllq xmm5, 1 + shl r10, 32 + movdqa xmm0, xmm9 + psrldq xmm0, 8 + movdqa xmm1, xmm10 + movq r11, xmm0 + psrldq xmm1, 8 + movq r8, xmm1 + psrldq xmm4, 8 + movaps xmm0, xmm13 + movq rax, xmm4 + xor r10, rax + movaps xmm1, xmm13 + xor r10, r12 + lea rax, QWORD PTR [r11+1] + shr rax, 1 + movdqa xmm3, xmm9 + punpcklqdq xmm3, xmm10 + paddq xmm5, xmm3 + movq rdx, xmm5 + psrldq xmm5, 8 + cvtsi2sd xmm2, rax + or edx, -2147483647 + lea rax, QWORD PTR [r8+1] + shr rax, 1 + movq r9, xmm5 + cvtsi2sd xmm0, rax + or r9d, -2147483647 + cvtsi2sd xmm1, rdx + unpcklpd xmm2, xmm0 + movaps xmm0, xmm13 + cvtsi2sd xmm0, r9 + unpcklpd xmm1, xmm0 + divpd xmm2, xmm1 + paddq xmm2, xmm14 + cvttsd2si rax, xmm2 + psrldq xmm2, 8 + mov rbx, rax + imul rax, rdx + sub r11, rax + js div_fix_1_sandybridge +div_fix_1_ret_sandybridge: + + cvttsd2si rdx, xmm2 + mov rax, rdx + imul rax, r9 + movd xmm2, r11d + movd xmm4, ebx + sub r8, rax + js div_fix_2_sandybridge +div_fix_2_ret_sandybridge: + + movd xmm1, r8d + movd xmm0, edx + punpckldq xmm2, xmm1 + punpckldq xmm4, xmm0 + punpckldq xmm4, xmm2 + paddq xmm3, xmm4 + movdqa xmm0, xmm3 + psrlq xmm0, 12 + paddq xmm0, xmm12 + sqrtpd xmm1, xmm0 + movq r9, xmm1 + movdqa xmm5, xmm1 + psrlq xmm5, 19 + test r9, 524287 + je sqrt_fix_1_sandybridge +sqrt_fix_1_ret_sandybridge: + + movq r9, xmm10 + psrldq xmm1, 8 + movq r8, xmm1 + test r8, 524287 + je sqrt_fix_2_sandybridge +sqrt_fix_2_ret_sandybridge: + + mov r12d, ecx + mov r8d, ecx + xor r12d, 16 + xor r8d, 32 + xor ecx, 48 + mov rax, r10 + mul r9 + movq xmm0, rax + movq xmm3, rdx + punpcklqdq xmm3, xmm0 + + movdqu xmm0, XMMWORD PTR [r12+rsi] + pxor xmm0, xmm3 + movdqu xmm1, XMMWORD PTR [r8+rsi] + xor rdx, [r8+rsi] + xor rax, [r8+rsi+8] + movdqu xmm3, XMMWORD PTR [rcx+rsi] + paddq xmm0, xmm6 + paddq xmm1, xmm11 + paddq xmm3, xmm8 + movdqu XMMWORD PTR [r8+rsi], xmm0 + movdqu XMMWORD PTR [rcx+rsi], xmm1 + movdqu XMMWORD PTR [r12+rsi], xmm3 + + add rdi, rdx + mov QWORD PTR [r13], rdi + xor rdi, r10 + mov ecx, edi + and ecx, 2097136 + lea r8, QWORD PTR [rcx+rsi] + + mov rdx, QWORD PTR [r13+8] + add rbp, rax + mov QWORD PTR [r13+8], rbp + movdqu xmm11, XMMWORD PTR [rcx+rsi] + xor rbp, rdx + mov r13, QWORD PTR [rsp] + movdqa xmm3, xmm7 + mov rdx, QWORD PTR [rsp+8] + movdqa xmm8, xmm6 + mov r10, QWORD PTR [rsp+256] + movdqa xmm7, xmm9 + mov r11, QWORD PTR [rsp+264] + movdqa xmm6, xmm10 + mov r9, r15 + dec r14d + jne main_loop_double_sandybridge + + ldmxcsr DWORD PTR [rsp+272] + movaps xmm13, XMMWORD PTR [rsp+48] + lea r11, QWORD PTR [rsp+184] + movaps xmm6, XMMWORD PTR [r11-24] + movaps xmm7, XMMWORD PTR [r11-40] + movaps xmm8, XMMWORD PTR [r11-56] + movaps xmm9, XMMWORD PTR [r11-72] + movaps xmm10, XMMWORD PTR [r11-88] + movaps xmm11, XMMWORD PTR [r11-104] + movaps xmm12, XMMWORD PTR [r11-120] + movaps xmm14, XMMWORD PTR [rsp+32] + movaps xmm15, XMMWORD PTR [rsp+16] + mov rsp, r11 + pop r15 + pop r14 + pop r13 + pop r12 + pop rdi + pop rsi + pop rbp + pop rbx + jmp cnv2_double_mainloop_asm_sandybridge_endp + +div_fix_1_sandybridge: + dec rbx + add r11, rdx + jmp div_fix_1_ret_sandybridge + +div_fix_2_sandybridge: + dec rdx + add r8, r9 + jmp div_fix_2_ret_sandybridge + +sqrt_fix_1_sandybridge: + movq r8, xmm3 + movdqa xmm0, xmm5 + psrldq xmm0, 8 + dec r9 + mov r11d, -1022 + shl r11, 32 + mov rax, r9 + shr r9, 19 + shr rax, 20 + mov rdx, r9 + sub rdx, rax + lea rdx, [rdx+r11+1] + add rax, r11 + imul rdx, rax + sub rdx, r8 + adc r9, 0 + movq xmm5, r9 + punpcklqdq xmm5, xmm0 + jmp sqrt_fix_1_ret_sandybridge + +sqrt_fix_2_sandybridge: + psrldq xmm3, 8 + movq r11, xmm3 + dec r8 + mov ebx, -1022 + shl rbx, 32 + mov rax, r8 + shr r8, 19 + shr rax, 20 + mov rdx, r8 + sub rdx, rax + lea rdx, [rdx+rbx+1] + add rax, rbx + imul rdx, rax + sub rdx, r11 + adc r8, 0 + movq xmm0, r8 + punpcklqdq xmm5, xmm0 + jmp sqrt_fix_2_ret_sandybridge + +cnv2_double_mainloop_asm_sandybridge_endp: diff --git a/src/crypto/asm/cnv2_main_loop.S b/src/crypto/asm/cnv2_main_loop.S new file mode 100644 index 00000000..4dbcbbda --- /dev/null +++ b/src/crypto/asm/cnv2_main_loop.S @@ -0,0 +1,37 @@ +#define ALIGN .align +.intel_syntax noprefix +#ifdef __APPLE__ +# define FN_PREFIX(fn) _ ## fn +.text +#else +# define FN_PREFIX(fn) fn +.section .text +#endif +.global FN_PREFIX(cnv2_mainloop_ivybridge_asm) +.global FN_PREFIX(cnv2_mainloop_ryzen_asm) +.global FN_PREFIX(cnv2_double_mainloop_sandybridge_asm) + +ALIGN 16 +FN_PREFIX(cnv2_mainloop_ivybridge_asm): + sub rsp, 48 + mov rcx, rdi + #include "cnv2_main_loop_ivybridge.inc" + add rsp, 48 + ret 0 + +ALIGN 16 +FN_PREFIX(cnv2_mainloop_ryzen_asm): + sub rsp, 48 + mov rcx, rdi + #include "cnv2_main_loop_ryzen.inc" + add rsp, 48 + ret 0 + +ALIGN 16 +FN_PREFIX(cnv2_double_mainloop_sandybridge_asm): + sub rsp, 48 + mov rcx, rdi + mov rdx, rsi + #include "cnv2_double_main_loop_sandybridge.inc" + add rsp, 48 + ret 0 diff --git a/src/crypto/asm/cnv2_main_loop.asm b/src/crypto/asm/cnv2_main_loop.asm new file mode 100644 index 00000000..d9522267 --- /dev/null +++ b/src/crypto/asm/cnv2_main_loop.asm @@ -0,0 +1,25 @@ +_TEXT_CNV2_MAINLOOP SEGMENT PAGE READ EXECUTE +PUBLIC cnv2_mainloop_ivybridge_asm +PUBLIC cnv2_mainloop_ryzen_asm +PUBLIC cnv2_double_mainloop_sandybridge_asm + +ALIGN 64 +cnv2_mainloop_ivybridge_asm PROC + INCLUDE cnv2_main_loop_ivybridge.inc + ret 0 +cnv2_mainloop_ivybridge_asm ENDP + +ALIGN 64 +cnv2_mainloop_ryzen_asm PROC + INCLUDE cnv2_main_loop_ryzen.inc + ret 0 +cnv2_mainloop_ryzen_asm ENDP + +ALIGN 64 +cnv2_double_mainloop_sandybridge_asm PROC + INCLUDE cnv2_double_main_loop_sandybridge.inc + ret 0 +cnv2_double_mainloop_sandybridge_asm ENDP + +_TEXT_CNV2_MAINLOOP ENDS +END diff --git a/src/crypto/asm/cnv2_main_loop_ivybridge.inc b/src/crypto/asm/cnv2_main_loop_ivybridge.inc new file mode 100644 index 00000000..8c2c2d3b --- /dev/null +++ b/src/crypto/asm/cnv2_main_loop_ivybridge.inc @@ -0,0 +1,186 @@ + mov QWORD PTR [rsp+24], rbx + push rbp + push rsi + push rdi + push r12 + push r13 + push r14 + push r15 + sub rsp, 80 + + stmxcsr DWORD PTR [rsp] + mov DWORD PTR [rsp+4], 24448 + ldmxcsr DWORD PTR [rsp+4] + + mov rax, QWORD PTR [rcx+48] + mov r9, rcx + xor rax, QWORD PTR [rcx+16] + mov esi, 524288 + mov r8, QWORD PTR [rcx+32] + mov r13d, -2147483647 + xor r8, QWORD PTR [rcx] + mov r11, QWORD PTR [rcx+40] + mov r10, r8 + mov rdx, QWORD PTR [rcx+56] + movq xmm4, rax + xor rdx, QWORD PTR [rcx+24] + xor r11, QWORD PTR [rcx+8] + mov rbx, QWORD PTR [rcx+224] + mov rax, QWORD PTR [r9+80] + xor rax, QWORD PTR [r9+64] + movq xmm0, rdx + mov rcx, QWORD PTR [rcx+88] + xor rcx, QWORD PTR [r9+72] + movq xmm3, QWORD PTR [r9+104] + movaps XMMWORD PTR [rsp+64], xmm6 + movaps XMMWORD PTR [rsp+48], xmm7 + movaps XMMWORD PTR [rsp+32], xmm8 + and r10d, 2097136 + movq xmm5, rax + + xor eax, eax + mov QWORD PTR [rsp+16], rax + + mov ax, 1023 + shl rax, 52 + movq xmm8, rax + mov r15, QWORD PTR [r9+96] + punpcklqdq xmm4, xmm0 + movq xmm0, rcx + punpcklqdq xmm5, xmm0 + movdqu xmm6, XMMWORD PTR [r10+rbx] + + ALIGN 16 +main_loop_ivybridge: + lea rdx, QWORD PTR [r10+rbx] + mov ecx, r10d + mov eax, r10d + mov rdi, r15 + xor ecx, 16 + xor eax, 32 + xor r10d, 48 + movq xmm0, r11 + movq xmm7, r8 + punpcklqdq xmm7, xmm0 + aesenc xmm6, xmm7 + movq rbp, xmm6 + mov r9, rbp + and r9d, 2097136 + movdqu xmm2, XMMWORD PTR [rcx+rbx] + movdqu xmm1, XMMWORD PTR [rax+rbx] + movdqu xmm0, XMMWORD PTR [r10+rbx] + paddq xmm1, xmm7 + paddq xmm0, xmm5 + paddq xmm2, xmm4 + movdqu XMMWORD PTR [rcx+rbx], xmm0 + movdqu XMMWORD PTR [rax+rbx], xmm2 + movdqu XMMWORD PTR [r10+rbx], xmm1 + mov r10, r9 + xor r10d, 32 + movq rcx, xmm3 + mov rax, rcx + shl rax, 32 + xor rdi, rax + movdqa xmm0, xmm6 + pxor xmm0, xmm4 + movdqu XMMWORD PTR [rdx], xmm0 + xor rdi, QWORD PTR [r9+rbx] + lea r14, QWORD PTR [r9+rbx] + mov r12, QWORD PTR [r14+8] + xor edx, edx + lea r9d, DWORD PTR [ecx+ecx] + add r9d, ebp + movdqa xmm0, xmm6 + psrldq xmm0, 8 + or r9d, r13d + movq rax, xmm0 + div r9 + xorps xmm3, xmm3 + mov eax, eax + shl rdx, 32 + add rdx, rax + lea r9, QWORD PTR [rdx+rbp] + mov r15, rdx + mov rax, r9 + shr rax, 12 + movq xmm0, rax + paddq xmm0, xmm8 + sqrtsd xmm3, xmm0 + psubq xmm3, XMMWORD PTR [rsp+16] + movq rdx, xmm3 + test edx, 524287 + je sqrt_fixup_ivybridge + psrlq xmm3, 19 +sqrt_fixup_ivybridge_ret: + + mov ecx, r10d + mov rax, rdi + mul rbp + movq xmm2, rdx + xor rdx, [rcx+rbx] + add r8, rdx + mov QWORD PTR [r14], r8 + xor r8, rdi + mov edi, r8d + and edi, 2097136 + movq xmm0, rax + xor rax, [rcx+rbx+8] + add r11, rax + mov QWORD PTR [r14+8], r11 + punpcklqdq xmm2, xmm0 + + mov r9d, r10d + xor r9d, 48 + xor r10d, 16 + pxor xmm2, XMMWORD PTR [r9+rbx] + movdqu xmm0, XMMWORD PTR [r10+rbx] + paddq xmm0, xmm5 + movdqu xmm1, XMMWORD PTR [rcx+rbx] + paddq xmm2, xmm4 + paddq xmm1, xmm7 + movdqa xmm5, xmm4 + movdqu XMMWORD PTR [r9+rbx], xmm0 + movdqa xmm4, xmm6 + movdqu XMMWORD PTR [rcx+rbx], xmm2 + movdqu XMMWORD PTR [r10+rbx], xmm1 + movdqu xmm6, [rdi+rbx] + mov r10d, edi + xor r11, r12 + dec rsi + jne main_loop_ivybridge + + ldmxcsr DWORD PTR [rsp] + mov rbx, QWORD PTR [rsp+160] + movaps xmm6, XMMWORD PTR [rsp+64] + movaps xmm7, XMMWORD PTR [rsp+48] + movaps xmm8, XMMWORD PTR [rsp+32] + add rsp, 80 + pop r15 + pop r14 + pop r13 + pop r12 + pop rdi + pop rsi + pop rbp + jmp cnv2_main_loop_ivybridge_endp + +sqrt_fixup_ivybridge: + dec rdx + mov r13d, -1022 + shl r13, 32 + mov rax, rdx + shr rdx, 19 + shr rax, 20 + mov rcx, rdx + sub rcx, rax + add rax, r13 + not r13 + sub rcx, r13 + mov r13d, -2147483647 + imul rcx, rax + sub rcx, r9 + adc rdx, 0 + movq xmm3, rdx + jmp sqrt_fixup_ivybridge_ret + +cnv2_main_loop_ivybridge_endp: diff --git a/src/crypto/asm/cnv2_main_loop_ryzen.inc b/src/crypto/asm/cnv2_main_loop_ryzen.inc new file mode 100644 index 00000000..d386aa2d --- /dev/null +++ b/src/crypto/asm/cnv2_main_loop_ryzen.inc @@ -0,0 +1,179 @@ + mov QWORD PTR [rsp+16], rbx + mov QWORD PTR [rsp+24], rbp + mov QWORD PTR [rsp+32], rsi + push rdi + push r12 + push r13 + push r14 + push r15 + sub rsp, 64 + + stmxcsr DWORD PTR [rsp] + mov DWORD PTR [rsp+4], 24448 + ldmxcsr DWORD PTR [rsp+4] + + mov rax, QWORD PTR [rcx+48] + mov r9, rcx + xor rax, QWORD PTR [rcx+16] + mov ebp, 524288 + mov r8, QWORD PTR [rcx+32] + xor r8, QWORD PTR [rcx] + mov r11, QWORD PTR [rcx+40] + mov r10, r8 + mov rdx, QWORD PTR [rcx+56] + movq xmm3, rax + xor rdx, QWORD PTR [rcx+24] + xor r11, QWORD PTR [rcx+8] + mov rbx, QWORD PTR [rcx+224] + mov rax, QWORD PTR [r9+80] + xor rax, QWORD PTR [r9+64] + movq xmm0, rdx + mov rcx, QWORD PTR [rcx+88] + xor rcx, QWORD PTR [r9+72] + mov rdi, QWORD PTR [r9+104] + and r10d, 2097136 + movaps XMMWORD PTR [rsp+48], xmm6 + movq xmm4, rax + movaps XMMWORD PTR [rsp+32], xmm7 + movaps XMMWORD PTR [rsp+16], xmm8 + xorps xmm8, xmm8 + mov ax, 1023 + shl rax, 52 + movq xmm7, rax + mov r15, QWORD PTR [r9+96] + punpcklqdq xmm3, xmm0 + movq xmm0, rcx + punpcklqdq xmm4, xmm0 + + ALIGN 16 +main_loop_ryzen: + movdqa xmm5, XMMWORD PTR [r10+rbx] + movq xmm0, r11 + movq xmm6, r8 + punpcklqdq xmm6, xmm0 + lea rdx, QWORD PTR [r10+rbx] + lea r9, QWORD PTR [rdi+rdi] + shl rdi, 32 + + mov ecx, r10d + mov eax, r10d + xor ecx, 16 + xor eax, 32 + xor r10d, 48 + aesenc xmm5, xmm6 + movdqa xmm2, XMMWORD PTR [rcx+rbx] + movdqa xmm1, XMMWORD PTR [rax+rbx] + movdqa xmm0, XMMWORD PTR [r10+rbx] + paddq xmm2, xmm3 + paddq xmm1, xmm6 + paddq xmm0, xmm4 + movdqa XMMWORD PTR [rcx+rbx], xmm0 + movdqa XMMWORD PTR [rax+rbx], xmm2 + movdqa XMMWORD PTR [r10+rbx], xmm1 + + movaps xmm1, xmm8 + mov rsi, r15 + xor rsi, rdi + movq r14, xmm5 + movdqa xmm0, xmm5 + pxor xmm0, xmm3 + mov r10, r14 + and r10d, 2097136 + movdqa XMMWORD PTR [rdx], xmm0 + xor rsi, QWORD PTR [r10+rbx] + lea r12, QWORD PTR [r10+rbx] + mov r13, QWORD PTR [r10+rbx+8] + + add r9d, r14d + or r9d, -2147483647 + xor edx, edx + movdqa xmm0, xmm5 + psrldq xmm0, 8 + movq rax, xmm0 + + div r9 + movq xmm0, rax + movq xmm1, rdx + punpckldq xmm0, xmm1 + movq r15, xmm0 + paddq xmm0, xmm5 + movdqa xmm2, xmm0 + psrlq xmm0, 12 + paddq xmm0, xmm7 + sqrtsd xmm1, xmm0 + movq rdi, xmm1 + test rdi, 524287 + je sqrt_fixup_ryzen + shr rdi, 19 + +sqrt_fixup_ryzen_ret: + mov rax, rsi + mul r14 + movq xmm1, rax + movq xmm0, rdx + punpcklqdq xmm0, xmm1 + + mov r9d, r10d + mov ecx, r10d + xor r9d, 16 + xor ecx, 32 + xor r10d, 48 + movdqa xmm1, XMMWORD PTR [rcx+rbx] + xor rdx, [rcx+rbx] + xor rax, [rcx+rbx+8] + movdqa xmm2, XMMWORD PTR [r9+rbx] + pxor xmm2, xmm0 + paddq xmm4, XMMWORD PTR [r10+rbx] + paddq xmm2, xmm3 + paddq xmm1, xmm6 + movdqa XMMWORD PTR [r9+rbx], xmm4 + movdqa XMMWORD PTR [rcx+rbx], xmm2 + movdqa XMMWORD PTR [r10+rbx], xmm1 + + movdqa xmm4, xmm3 + add r8, rdx + add r11, rax + mov QWORD PTR [r12], r8 + xor r8, rsi + mov QWORD PTR [r12+8], r11 + mov r10, r8 + xor r11, r13 + and r10d, 2097136 + movdqa xmm3, xmm5 + dec ebp + jne main_loop_ryzen + + ldmxcsr DWORD PTR [rsp] + movaps xmm6, XMMWORD PTR [rsp+48] + lea r11, QWORD PTR [rsp+64] + mov rbx, QWORD PTR [r11+56] + mov rbp, QWORD PTR [r11+64] + mov rsi, QWORD PTR [r11+72] + movaps xmm8, XMMWORD PTR [r11-48] + movaps xmm7, XMMWORD PTR [rsp+32] + mov rsp, r11 + pop r15 + pop r14 + pop r13 + pop r12 + pop rdi + jmp cnv2_main_loop_ryzen_endp + +sqrt_fixup_ryzen: + movq r9, xmm2 + dec rdi + mov edx, -1022 + shl rdx, 32 + mov rax, rdi + shr rdi, 19 + shr rax, 20 + mov rcx, rdi + sub rcx, rax + lea rcx, [rcx+rdx+1] + add rax, rdx + imul rcx, rax + sub rcx, r9 + adc rdi, 0 + jmp sqrt_fixup_ryzen_ret + +cnv2_main_loop_ryzen_endp: diff --git a/src/crypto/asm/cnv2_main_loop_win.S b/src/crypto/asm/cnv2_main_loop_win.S new file mode 100644 index 00000000..f06e4fa4 --- /dev/null +++ b/src/crypto/asm/cnv2_main_loop_win.S @@ -0,0 +1,21 @@ +#define ALIGN .align +.intel_syntax noprefix +.section .text +.global cnv2_mainloop_ivybridge_asm +.global cnv2_mainloop_ryzen_asm +.global cnv2_double_mainloop_sandybridge_asm + +ALIGN 16 +cnv2_mainloop_ivybridge_asm: + #include "cnv2_main_loop_ivybridge.inc" + ret 0 + +ALIGN 16 +cnv2_mainloop_ryzen_asm: + #include "cnv2_main_loop_ryzen.inc" + ret 0 + +ALIGN 16 +cnv2_double_mainloop_sandybridge_asm: + #include "cnv2_double_main_loop_sandybridge.inc" + ret 0 diff --git a/src/interfaces/IThread.h b/src/interfaces/IThread.h index 2e9e3c39..3a8708e6 100644 --- a/src/interfaces/IThread.h +++ b/src/interfaces/IThread.h @@ -20,8 +20,8 @@ * along with this program. If not, see . */ -#ifndef __ITHREAD_H__ -#define __ITHREAD_H__ +#ifndef XMRIG_ITHREAD_H +#define XMRIG_ITHREAD_H #include @@ -64,10 +64,14 @@ public: # ifndef XMRIG_NO_API virtual rapidjson::Value toAPI(rapidjson::Document &doc) const = 0; # endif + +# ifdef APP_DEBUG + virtual void print() const = 0; +# endif }; } /* namespace xmrig */ -#endif // __ITHREAD_H__ +#endif // XMRIG_ITHREAD_H diff --git a/src/net/Network.cpp b/src/net/Network.cpp index 7293a0ac..703e0ccf 100644 --- a/src/net/Network.cpp +++ b/src/net/Network.cpp @@ -101,7 +101,15 @@ void Network::onActive(IStrategy *strategy, Client *client) m_state.setPool(client->host(), client->port(), client->ip()); - LOG_INFO(isColors() ? "\x1B[01;37muse pool \x1B[01;36m%s:%d \x1B[01;30m%s" : "use pool %s:%d %s", client->host(), client->port(), client->ip()); + const char *tlsVersion = client->tlsVersion(); + LOG_INFO(isColors() ? WHITE_BOLD("use pool ") CYAN_BOLD("%s:%d ") GREEN_BOLD("%s") " \x1B[1;30m%s " + : "use pool %s:%d %s %s", + client->host(), client->port(), tlsVersion ? tlsVersion : "", client->ip()); + + const char *fingerprint = client->tlsFingerprint(); + if (fingerprint != nullptr) { + LOG_INFO("\x1B[1;30mfingerprint (SHA-256): \"%s\"", fingerprint); + } } diff --git a/src/net/strategies/DonateStrategy.cpp b/src/net/strategies/DonateStrategy.cpp index 395c53bb..6fc90842 100644 --- a/src/net/strategies/DonateStrategy.cpp +++ b/src/net/strategies/DonateStrategy.cpp @@ -33,10 +33,6 @@ #include "net/strategies/DonateStrategy.h" -const static char *kDonatePool1 = "miner.fee.xmrig.com"; -const static char *kDonatePool2 = "emergency.fee.xmrig.com"; - - static inline float randomf(float min, float max) { return (max - min) * ((((float) rand()) / (float) RAND_MAX)) + min; } @@ -55,17 +51,11 @@ DonateStrategy::DonateStrategy(int level, const char *user, xmrig::Algo algo, IS xmrig::keccak(reinterpret_cast(user), strlen(user), hash); Job::toHex(hash, 32, userId); - if (algo == xmrig::CRYPTONIGHT) { - m_pools.push_back(Pool(kDonatePool1, 6666, userId, nullptr, false, true)); - m_pools.push_back(Pool(kDonatePool1, 80, userId, nullptr, false, true)); - m_pools.push_back(Pool(kDonatePool2, 5555, "48edfHu7V9Z84YzzMa6fUueoELZ9ZRXq9VetWzYGzKt52XU5xvqgzYnDK9URnRoJMk1j8nLwEVsaSWJ4fhdUyZijBGUicoD", "emergency", false, false)); - } - else if (algo == xmrig::CRYPTONIGHT_HEAVY) { - m_pools.push_back(Pool(kDonatePool1, 8888, userId, nullptr, false, true)); - } - else { - m_pools.push_back(Pool(kDonatePool1, 5555, userId, nullptr, false, true)); - } +# ifndef XMRIG_NO_TLS + m_pools.push_back(Pool("donate.ssl.xmrig.com", 443, userId, nullptr, false, true, true)); +# endif + + m_pools.push_back(Pool("donate.v2.xmrig.com", 3333, userId, nullptr, false, true)); for (Pool &pool : m_pools) { pool.adjust(xmrig::Algorithm(algo, xmrig::VARIANT_AUTO)); diff --git a/src/net/strategies/DonateStrategy.h b/src/net/strategies/DonateStrategy.h index 95ff6608..e75e41a4 100644 --- a/src/net/strategies/DonateStrategy.h +++ b/src/net/strategies/DonateStrategy.h @@ -21,8 +21,8 @@ * along with this program. If not, see . */ -#ifndef __DONATESTRATEGY_H__ -#define __DONATESTRATEGY_H__ +#ifndef XMRIG_DONATESTRATEGY_H +#define XMRIG_DONATESTRATEGY_H #include @@ -76,4 +76,4 @@ private: uv_timer_t m_timer; }; -#endif /* __DONATESTRATEGY_H__ */ +#endif /* XMRIG_DONATESTRATEGY_H */ diff --git a/src/version.h b/src/version.h index 76a3c77f..2a1720ea 100644 --- a/src/version.h +++ b/src/version.h @@ -21,21 +21,21 @@ * along with this program. If not, see . */ -#ifndef __VERSION_H__ -#define __VERSION_H__ +#ifndef XMRIG_VERSION_H +#define XMRIG_VERSION_H #define APP_ID "xmrig" #define APP_NAME "XMRig" #define APP_DESC "XMRig CPU miner" -#define APP_VERSION "2.6.4" +#define APP_VERSION "2.8.0-dev" #define APP_DOMAIN "xmrig.com" #define APP_SITE "www.xmrig.com" #define APP_COPYRIGHT "Copyright (C) 2016-2018 xmrig.com" #define APP_KIND "cpu" #define APP_VER_MAJOR 2 -#define APP_VER_MINOR 6 -#define APP_VER_PATCH 4 +#define APP_VER_MINOR 8 +#define APP_VER_PATCH 0 #ifdef _MSC_VER # if (_MSC_VER >= 1910) @@ -53,4 +53,4 @@ # endif #endif -#endif /* __VERSION_H__ */ +#endif /* XMRIG_VERSION_H */ diff --git a/src/workers/CpuThread.cpp b/src/workers/CpuThread.cpp index 7cef4f3a..4b528148 100644 --- a/src/workers/CpuThread.cpp +++ b/src/workers/CpuThread.cpp @@ -24,7 +24,10 @@ #include +#include "common/cpu/Cpu.h" +#include "common/log/Log.h" #include "common/net/Pool.h" +#include "crypto/Asm.h" #include "rapidjson/document.h" #include "workers/CpuThread.h" @@ -36,9 +39,10 @@ #endif -xmrig::CpuThread::CpuThread(size_t index, Algo algorithm, AlgoVariant av, Multiway multiway, int64_t affinity, int priority, bool softAES, bool prefetch) : +xmrig::CpuThread::CpuThread(size_t index, Algo algorithm, AlgoVariant av, Multiway multiway, int64_t affinity, int priority, bool softAES, bool prefetch, Assembly assembly) : m_algorithm(algorithm), m_av(av), + m_assembly(assembly), m_prefetch(prefetch), m_softAES(softAES), m_priority(priority), @@ -49,22 +53,23 @@ xmrig::CpuThread::CpuThread(size_t index, Algo algorithm, AlgoVariant av, Multiw } -xmrig::CpuThread::~CpuThread() -{ -} - - bool xmrig::CpuThread::isSoftAES(AlgoVariant av) { return av == AV_SINGLE_SOFT || av == AV_DOUBLE_SOFT || av > AV_PENTA; } -xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant av, Variant variant) +xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant av, Variant variant, Assembly assembly) { assert(variant >= VARIANT_0 && variant < VARIANT_MAX); - static const cn_hash_fun func_table[VARIANT_MAX * 10 * 3] = { +# ifndef XMRIG_NO_ASM + constexpr const size_t count = VARIANT_MAX * 10 * 3 + 3; +# else + constexpr const size_t count = VARIANT_MAX * 10 * 3; +# endif + + static const cn_hash_fun func_table[count] = { cryptonight_single_hash, cryptonight_double_hash, cryptonight_single_hash, @@ -135,6 +140,17 @@ xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant a cryptonight_quad_hash, cryptonight_penta_hash, + cryptonight_single_hash, + cryptonight_double_hash, + cryptonight_single_hash, + cryptonight_double_hash, + cryptonight_triple_hash, + cryptonight_quad_hash, + cryptonight_penta_hash, + cryptonight_triple_hash, + cryptonight_quad_hash, + cryptonight_penta_hash, + # ifndef XMRIG_NO_AEON cryptonight_single_hash, cryptonight_double_hash, @@ -164,6 +180,7 @@ xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant a nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_XHV nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_XAO nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_RTO + nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_2 # else nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, @@ -173,6 +190,7 @@ xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant a nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, + nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, # endif # ifndef XMRIG_NO_SUMO @@ -216,6 +234,7 @@ xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant a nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_XAO nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_RTO + nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, // VARIANT_2 # else nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, @@ -225,12 +244,17 @@ xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant a nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, + nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, +# endif +# ifndef XMRIG_NO_ASM + cryptonight_single_hash_asm, + cryptonight_single_hash_asm, + cryptonight_double_hash_asm # endif }; - const size_t index = VARIANT_MAX * 10 * algorithm + 10 * variant + av - 1; - # ifndef NDEBUG + const size_t index = fnIndex(algorithm, av, variant, assembly); cn_hash_fun func = func_table[index]; assert(index < sizeof(func_table) / sizeof(func_table[0])); @@ -238,12 +262,12 @@ xmrig::CpuThread::cn_hash_fun xmrig::CpuThread::fn(Algo algorithm, AlgoVariant a return func; # else - return func_table[index]; + return func_table[fnIndex(algorithm, av, variant, assembly)]; # endif } -xmrig::CpuThread *xmrig::CpuThread::createFromAV(size_t index, Algo algorithm, AlgoVariant av, int64_t affinity, int priority) +xmrig::CpuThread *xmrig::CpuThread::createFromAV(size_t index, Algo algorithm, AlgoVariant av, int64_t affinity, int priority, Assembly assembly) { assert(av > AV_AUTO && av < AV_MAX); @@ -266,7 +290,7 @@ xmrig::CpuThread *xmrig::CpuThread::createFromAV(size_t index, Algo algorithm, A } } - return new CpuThread(index, algorithm, av, multiway(av), cpuId, priority, isSoftAES(av), false); + return new CpuThread(index, algorithm, av, multiway(av), cpuId, priority, isSoftAES(av), false, assembly); } @@ -284,7 +308,7 @@ xmrig::CpuThread *xmrig::CpuThread::createFromData(size_t index, Algo algorithm, assert(av > AV_AUTO && av < AV_MAX); - return new CpuThread(index, algorithm, static_cast(av), multiway, data.affinity, priority, softAES, false); + return new CpuThread(index, algorithm, static_cast(av), multiway, data.affinity, priority, softAES, false, data.assembly); } @@ -306,11 +330,14 @@ xmrig::CpuThread::Data xmrig::CpuThread::parse(const rapidjson::Value &object) } const auto &affinity = object["affine_to_cpu"]; - if (affinity.IsUint64()) { data.affinity = affinity.GetInt64(); } +# ifndef XMRIG_NO_ASM + data.assembly = Asm::parse(object["asm"]); +# endif + return data; } @@ -346,6 +373,21 @@ xmrig::IThread::Multiway xmrig::CpuThread::multiway(AlgoVariant av) } +#ifdef APP_DEBUG +void xmrig::CpuThread::print() const +{ + LOG_DEBUG(GREEN_BOLD("CPU thread: ") " index " WHITE_BOLD("%zu") ", multiway " WHITE_BOLD("%d") ", av " WHITE_BOLD("%d") ",", + index(), static_cast(multiway()), static_cast(m_av)); + +# ifndef XMRIG_NO_ASM + LOG_DEBUG(" assembly: %s, affine_to_cpu: %" PRId64, Asm::toString(m_assembly), affinity()); +# else + LOG_DEBUG(" affine_to_cpu: %" PRId64, affinity()); +# endif +} +#endif + + #ifndef XMRIG_NO_API rapidjson::Value xmrig::CpuThread::toAPI(rapidjson::Document &doc) const { @@ -376,5 +418,39 @@ rapidjson::Value xmrig::CpuThread::toConfig(rapidjson::Document &doc) const obj.AddMember("low_power_mode", multiway(), allocator); obj.AddMember("affine_to_cpu", affinity() == -1L ? Value(kFalseType) : Value(affinity()), allocator); +# ifndef XMRIG_NO_ASM + obj.AddMember("asm", Asm::toJSON(m_assembly), allocator); +# endif + return obj; } + + +size_t xmrig::CpuThread::fnIndex(Algo algorithm, AlgoVariant av, Variant variant, Assembly assembly) +{ + const size_t index = VARIANT_MAX * 10 * algorithm + 10 * variant + av - 1; + +# ifndef XMRIG_NO_ASM + if (assembly == ASM_AUTO) { + assembly = Cpu::info()->assembly(); + } + + if (assembly == ASM_NONE) { + return index; + } + + constexpr const size_t offset = VARIANT_MAX * 10 * 3; + + if (algorithm == CRYPTONIGHT && variant == VARIANT_2) { + if (av == AV_SINGLE) { + return offset + assembly - 2; + } + + if (av == AV_DOUBLE) { + return offset + 2; + } + } +# endif + + return index; +} diff --git a/src/workers/CpuThread.h b/src/workers/CpuThread.h index 0e364764..29ab9696 100644 --- a/src/workers/CpuThread.h +++ b/src/workers/CpuThread.h @@ -21,8 +21,8 @@ * along with this program. If not, see . */ -#ifndef __CPUTHREAD_H__ -#define __CPUTHREAD_H__ +#ifndef XMRIG_CPUTHREAD_H +#define XMRIG_CPUTHREAD_H #include "common/xmrig.h" @@ -40,7 +40,7 @@ class CpuThread : public IThread public: struct Data { - inline Data() : valid(false), affinity(-1L), multiway(SingleWay) {} + inline Data() : assembly(ASM_AUTO), valid(false), affinity(-1L), multiway(SingleWay) {} inline void setMultiway(int value) { @@ -50,27 +50,27 @@ public: } } + Assembly assembly; bool valid; int64_t affinity; Multiway multiway; }; - CpuThread(size_t index, Algo algorithm, AlgoVariant av, Multiway multiway, int64_t affinity, int priority, bool softAES, bool prefetch); - ~CpuThread(); + CpuThread(size_t index, Algo algorithm, AlgoVariant av, Multiway multiway, int64_t affinity, int priority, bool softAES, bool prefetch, Assembly assembly); typedef void (*cn_hash_fun)(const uint8_t *input, size_t size, uint8_t *output, cryptonight_ctx **ctx); static bool isSoftAES(AlgoVariant av); - static cn_hash_fun fn(Algo algorithm, AlgoVariant av, Variant variant); - static CpuThread *createFromAV(size_t index, Algo algorithm, AlgoVariant av, int64_t affinity, int priority); + static cn_hash_fun fn(Algo algorithm, AlgoVariant av, Variant variant, Assembly assembly); + static CpuThread *createFromAV(size_t index, Algo algorithm, AlgoVariant av, int64_t affinity, int priority, Assembly assembly); static CpuThread *createFromData(size_t index, Algo algorithm, const CpuThread::Data &data, int priority, bool softAES); static Data parse(const rapidjson::Value &object); static Multiway multiway(AlgoVariant av); inline bool isPrefetch() const { return m_prefetch; } inline bool isSoftAES() const { return m_softAES; } - inline cn_hash_fun fn(Variant variant) const { return fn(m_algorithm, m_av, variant); } + inline cn_hash_fun fn(Variant variant) const { return fn(m_algorithm, m_av, variant, m_assembly); } inline Algo algorithm() const override { return m_algorithm; } inline int priority() const override { return m_priority; } @@ -80,6 +80,10 @@ public: inline Type type() const override { return CPU; } protected: +# ifdef APP_DEBUG + void print() const override; +# endif + # ifndef XMRIG_NO_API rapidjson::Value toAPI(rapidjson::Document &doc) const override; # endif @@ -87,8 +91,11 @@ protected: rapidjson::Value toConfig(rapidjson::Document &doc) const override; private: + static size_t fnIndex(Algo algorithm, AlgoVariant av, Variant variant, Assembly assembly); + const Algo m_algorithm; const AlgoVariant m_av; + const Assembly m_assembly; const bool m_prefetch; const bool m_softAES; const int m_priority; @@ -101,4 +108,4 @@ private: } /* namespace xmrig */ -#endif /* __CPUTHREAD_H__ */ +#endif /* XMRIG_CPUTHREAD_H */ diff --git a/src/workers/MultiWorker.cpp b/src/workers/MultiWorker.cpp index 5d43875c..a6dbc73a 100644 --- a/src/workers/MultiWorker.cpp +++ b/src/workers/MultiWorker.cpp @@ -6,6 +6,7 @@ * Copyright 2016 Jay D Dee * Copyright 2017-2018 XMR-Stak , * Copyright 2018 Lee Clagett + * Copyright 2018 SChernykh * Copyright 2016-2018 XMRig , * * This program is free software: you can redistribute it and/or modify @@ -55,6 +56,7 @@ bool MultiWorker::selfTest() if (m_thread->algorithm() == CRYPTONIGHT) { return verify(VARIANT_0, test_output_v0) && verify(VARIANT_1, test_output_v1) && + verify(VARIANT_2, test_output_v2) && verify(VARIANT_XTL, test_output_xtl) && verify(VARIANT_MSR, test_output_msr) && verify(VARIANT_XAO, test_output_xao) && @@ -102,7 +104,7 @@ void MultiWorker::start() storeStats(); } - m_thread->fn(m_state.job.variant())(m_state.blob, m_state.job.size(), m_hash, m_ctx); + m_thread->fn(m_state.job.algorithm().variant())(m_state.blob, m_state.job.size(), m_hash, m_ctx); for (size_t i = 0; i < N; ++i) { if (*reinterpret_cast(m_hash + (i * 32) + 24) < m_state.job.target()) { diff --git a/src/workers/MultiWorker.h b/src/workers/MultiWorker.h index d4a6910e..c08e4fbe 100644 --- a/src/workers/MultiWorker.h +++ b/src/workers/MultiWorker.h @@ -6,6 +6,7 @@ * Copyright 2016 Jay D Dee * Copyright 2017-2018 XMR-Stak , * Copyright 2018 Lee Clagett + * Copyright 2018 SChernykh * Copyright 2016-2018 XMRig , * * This program is free software: you can redistribute it and/or modify @@ -22,8 +23,8 @@ * along with this program. If not, see . */ -#ifndef __MULTIWORKER_H__ -#define __MULTIWORKER_H__ +#ifndef XMRIG_MULTIWORKER_H +#define XMRIG_MULTIWORKER_H #include "common/net/Job.h" @@ -71,4 +72,4 @@ private: }; -#endif /* __MULTIWORKER_H__ */ +#endif /* XMRIG_MULTIWORKER_H */ diff --git a/src/workers/Worker.cpp b/src/workers/Worker.cpp index 567b3e08..c569908c 100644 --- a/src/workers/Worker.cpp +++ b/src/workers/Worker.cpp @@ -24,8 +24,8 @@ #include +#include "common/cpu/Cpu.h" #include "common/Platform.h" -#include "Cpu.h" #include "workers/CpuThread.h" #include "workers/Handle.h" #include "workers/Worker.h" @@ -41,7 +41,7 @@ Worker::Worker(Handle *handle) : m_sequence(0), m_thread(static_cast(handle->config())) { - if (Cpu::threads() > 1 && m_thread->affinity() != -1L) { + if (xmrig::Cpu::info()->threads() > 1 && m_thread->affinity() != -1L) { Platform::setThreadAffinity(m_thread->affinity()); } diff --git a/src/workers/Worker.h b/src/workers/Worker.h index aad9e3c5..73e25033 100644 --- a/src/workers/Worker.h +++ b/src/workers/Worker.h @@ -21,8 +21,8 @@ * along with this program. If not, see . */ -#ifndef __WORKER_H__ -#define __WORKER_H__ +#ifndef XMRIG_WORKER_H +#define XMRIG_WORKER_H #include @@ -33,7 +33,6 @@ #include "Mem.h" -struct cryptonight_ctx; class Handle; @@ -67,4 +66,4 @@ protected: }; -#endif /* __WORKER_H__ */ +#endif /* XMRIG_WORKER_H */ diff --git a/src/workers/Workers.cpp b/src/workers/Workers.cpp index 0e75e736..a5109e9b 100644 --- a/src/workers/Workers.cpp +++ b/src/workers/Workers.cpp @@ -160,6 +160,14 @@ void Workers::setJob(const Job &job, bool donate) void Workers::start(xmrig::Controller *controller) { +# ifdef APP_DEBUG + LOG_NOTICE("THREADS ------------------------------------------------------------------"); + for (const xmrig::IThread *thread : controller->config()->threads()) { + thread->print(); + } + LOG_NOTICE("--------------------------------------------------------------------------"); +# endif + m_controller = controller; const std::vector &threads = controller->config()->threads(); @@ -192,6 +200,10 @@ void Workers::start(xmrig::Controller *controller) m_workers.push_back(handle); handle->start(Workers::onReady); } + + if (controller->config()->isShouldSave()) { + controller->config()->save(); + } }