Browse Source

Removed libcpuid support.

pull/1966/head
XMRig 4 years ago
parent
commit
ab45794b7c
No known key found for this signature in database GPG Key ID: 446A53638BE94409
  1. 3
      CMakeLists.txt
  2. 1
      cmake/cpu.cmake
  3. 38
      src/3rdparty/libcpuid/CMakeLists.txt
  4. 39
      src/3rdparty/libcpuid/amd_code_t.h
  5. 836
      src/3rdparty/libcpuid/asm-bits.c
  6. 71
      src/3rdparty/libcpuid/asm-bits.h
  7. 389
      src/3rdparty/libcpuid/cpuid_main.c
  8. 58
      src/3rdparty/libcpuid/intel_code_t.h
  9. 678
      src/3rdparty/libcpuid/libcpuid.h
  10. 47
      src/3rdparty/libcpuid/libcpuid_constants.h
  11. 107
      src/3rdparty/libcpuid/libcpuid_internal.h
  12. 63
      src/3rdparty/libcpuid/libcpuid_types.h
  13. 93
      src/3rdparty/libcpuid/libcpuid_util.c
  14. 78
      src/3rdparty/libcpuid/libcpuid_util.h
  15. 359
      src/3rdparty/libcpuid/masm-x64.asm
  16. 173
      src/3rdparty/libcpuid/recog_amd.c
  17. 31
      src/3rdparty/libcpuid/recog_amd.h
  18. 543
      src/3rdparty/libcpuid/recog_intel.c
  19. 31
      src/3rdparty/libcpuid/recog_intel.h
  20. 2
      src/Summary.cpp
  21. 14
      src/backend/cpu/Cpu.cpp
  22. 10
      src/backend/cpu/Cpu.h
  23. 39
      src/backend/cpu/cpu.cmake
  24. 130
      src/backend/cpu/platform/AdvancedCpuInfo.cpp
  25. 63
      src/backend/cpu/platform/AdvancedCpuInfo.h
  26. 3
      src/base/api/Api.cpp

3
CMakeLists.txt

@ -1,7 +1,6 @@
cmake_minimum_required(VERSION 2.8)
project(xmrig)
option(WITH_LIBCPUID "Enable libcpuid support" ON)
option(WITH_HWLOC "Enable hwloc support" ON)
option(WITH_CN_LITE "Enable CryptoNight-Lite algorithms family" ON)
option(WITH_CN_HEAVY "Enable CryptoNight-Heavy algorithms family" ON)
@ -205,7 +204,7 @@ if (WITH_DEBUG_LOG)
add_definitions(/DAPP_DEBUG)
endif()
add_executable(${CMAKE_PROJECT_NAME} ${HEADERS} ${SOURCES} ${SOURCES_OS} ${SOURCES_CPUID} ${HEADERS_CRYPTO} ${SOURCES_CRYPTO} ${SOURCES_SYSLOG} ${TLS_SOURCES} ${XMRIG_ASM_SOURCES})
add_executable(${CMAKE_PROJECT_NAME} ${HEADERS} ${SOURCES} ${SOURCES_OS} ${HEADERS_CRYPTO} ${SOURCES_CRYPTO} ${SOURCES_SYSLOG} ${TLS_SOURCES} ${XMRIG_ASM_SOURCES})
target_link_libraries(${CMAKE_PROJECT_NAME} ${XMRIG_ASM_LIBRARY} ${OPENSSL_LIBRARIES} ${UV_LIBRARIES} ${EXTRA_LIBS} ${CPUID_LIB} ${ARGON2_LIBRARY} ${ETHASH_LIBRARY})
if (WIN32)

1
cmake/cpu.cmake

@ -18,7 +18,6 @@ endif()
if (ARM_TARGET AND ARM_TARGET GREATER 6)
set(XMRIG_ARM ON)
set(WITH_LIBCPUID OFF)
add_definitions(/DXMRIG_ARM)
message(STATUS "Use ARM_TARGET=${ARM_TARGET} (${CMAKE_SYSTEM_PROCESSOR})")

38
src/3rdparty/libcpuid/CMakeLists.txt

@ -1,38 +0,0 @@
cmake_minimum_required (VERSION 2.8)
project (cpuid C)
add_definitions(/DVERSION="0.4.0")
set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} -Os")
set(HEADERS
libcpuid.h
libcpuid_types.h
libcpuid_constants.h
libcpuid_internal.h
amd_code_t.h
intel_code_t.h
recog_amd.h
recog_intel.h
asm-bits.h
libcpuid_util.h
)
set(SOURCES
cpuid_main.c
asm-bits.c
recog_amd.c
recog_intel.c
libcpuid_util.c
)
if (CMAKE_CL_64)
enable_language(ASM_MASM)
set(SOURCES_ASM masm-x64.asm)
endif()
add_library(cpuid STATIC
${HEADERS}
${SOURCES}
${SOURCES_ASM}
)

39
src/3rdparty/libcpuid/amd_code_t.h

@ -1,39 +0,0 @@
/*
* Copyright 2016 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.
*/
/*
* This file contains a list of internal codes we use in detection. It is
* of no external use and isn't a complete list of AMD products.
*/
CODE2(OPTERON_800, 1000),
CODE(PHENOM),
CODE(PHENOM2),
CODE(FUSION_C),
CODE(FUSION_E),
CODE(FUSION_EA),
CODE(FUSION_Z),
CODE(FUSION_A),

836
src/3rdparty/libcpuid/asm-bits.c

@ -1,836 +0,0 @@
/*
* 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 */

71
src/3rdparty/libcpuid/asm-bits.h

@ -1,71 +0,0 @@
/*
* 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__ */

389
src/3rdparty/libcpuid/cpuid_main.c

@ -1,389 +0,0 @@
/*
* 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 "libcpuid_internal.h"
#include "recog_intel.h"
#include "recog_amd.h"
#include "asm-bits.h"
#include "libcpuid_util.h"
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
/* Implementation: */
static int _libcpiud_errno = ERR_OK;
int set_error(cpu_error_t err)
{
_libcpiud_errno = (int) err;
return (int) err;
}
static void cpu_id_t_constructor(struct cpu_id_t* id)
{
memset(id, 0, sizeof(struct cpu_id_t));
id->l1_data_cache = id->l1_instruction_cache = id->l2_cache = id->l3_cache = id->l4_cache = -1;
id->l1_assoc = id->l2_assoc = id->l3_assoc = id->l4_assoc = -1;
id->l1_cacheline = id->l2_cacheline = id->l3_cacheline = id->l4_cacheline = -1;
id->sse_size = -1;
}
/* get_total_cpus() system specific code: uses OS routines to determine total number of CPUs */
#ifdef __APPLE__
#include <unistd.h>
#include <mach/clock_types.h>
#include <mach/clock.h>
#include <mach/mach.h>
static int get_total_cpus(void)
{
kern_return_t kr;
host_basic_info_data_t basic_info;
host_info_t info = (host_info_t)&basic_info;
host_flavor_t flavor = HOST_BASIC_INFO;
mach_msg_type_number_t count = HOST_BASIC_INFO_COUNT;
kr = host_info(mach_host_self(), flavor, info, &count);
if (kr != KERN_SUCCESS) return 1;
return basic_info.avail_cpus;
}
#define GET_TOTAL_CPUS_DEFINED
#endif
#ifdef _WIN32
#include <windows.h>
static int get_total_cpus(void)
{
SYSTEM_INFO system_info;
GetSystemInfo(&system_info);
return system_info.dwNumberOfProcessors;
}
#define GET_TOTAL_CPUS_DEFINED
#endif
#if defined linux || defined __linux__ || defined __sun
#include <sys/sysinfo.h>
#include <unistd.h>
static int get_total_cpus(void)
{
return sysconf(_SC_NPROCESSORS_ONLN);
}
#define GET_TOTAL_CPUS_DEFINED
#endif
#if defined __FreeBSD__ || defined __OpenBSD__ || defined __NetBSD__ || defined __bsdi__ || defined __QNX__
#include <sys/types.h>
#include <sys/sysctl.h>
static int get_total_cpus(void)
{
int mib[2] = { CTL_HW, HW_NCPU };
int ncpus;
size_t len = sizeof(ncpus);
if (sysctl(mib, 2, &ncpus, &len, (void *) 0, 0) != 0) return 1;
return ncpus;
}
#define GET_TOTAL_CPUS_DEFINED
#endif
#ifndef GET_TOTAL_CPUS_DEFINED
static int get_total_cpus(void)
{
static int warning_printed = 0;
if (!warning_printed) {
warning_printed = 1;
warnf("Your system is not supported by libcpuid -- don't know how to detect the\n");
warnf("total number of CPUs on your system. It will be reported as 1.\n");
printf("Please use cpu_id_t.logical_cpus field instead.\n");
}
return 1;
}
#endif /* GET_TOTAL_CPUS_DEFINED */
static void load_features_common(struct cpu_raw_data_t* raw, struct cpu_id_t* data)
{
const struct feature_map_t matchtable_edx1[] = {
{ 0, CPU_FEATURE_FPU },
{ 1, CPU_FEATURE_VME },
{ 2, CPU_FEATURE_DE },
{ 3, CPU_FEATURE_PSE },
{ 4, CPU_FEATURE_TSC },
{ 5, CPU_FEATURE_MSR },
{ 6, CPU_FEATURE_PAE },
{ 7, CPU_FEATURE_MCE },
{ 8, CPU_FEATURE_CX8 },
{ 9, CPU_FEATURE_APIC },
{ 11, CPU_FEATURE_SEP },
{ 12, CPU_FEATURE_MTRR },
{ 13, CPU_FEATURE_PGE },
{ 14, CPU_FEATURE_MCA },
{ 15, CPU_FEATURE_CMOV },
{ 16, CPU_FEATURE_PAT },
{ 17, CPU_FEATURE_PSE36 },
{ 19, CPU_FEATURE_CLFLUSH },
{ 23, CPU_FEATURE_MMX },
{ 24, CPU_FEATURE_FXSR },
{ 25, CPU_FEATURE_SSE },
{ 26, CPU_FEATURE_SSE2 },
{ 28, CPU_FEATURE_HT },
};
const struct feature_map_t matchtable_ecx1[] = {
{ 0, CPU_FEATURE_PNI },
{ 1, CPU_FEATURE_PCLMUL },
{ 3, CPU_FEATURE_MONITOR },
{ 9, CPU_FEATURE_SSSE3 },
{ 12, CPU_FEATURE_FMA3 },
{ 13, CPU_FEATURE_CX16 },
{ 19, CPU_FEATURE_SSE4_1 },
{ 20, CPU_FEATURE_SSE4_2 },
{ 22, CPU_FEATURE_MOVBE },
{ 23, CPU_FEATURE_POPCNT },
{ 25, CPU_FEATURE_AES },
{ 26, CPU_FEATURE_XSAVE },
{ 27, CPU_FEATURE_OSXSAVE },
{ 28, CPU_FEATURE_AVX },
{ 29, CPU_FEATURE_F16C },
{ 30, CPU_FEATURE_RDRAND },
};
const struct feature_map_t matchtable_ebx7[] = {
{ 3, CPU_FEATURE_BMI1 },
{ 5, CPU_FEATURE_AVX2 },
{ 8, CPU_FEATURE_BMI2 },
};
const struct feature_map_t matchtable_edx81[] = {
{ 11, CPU_FEATURE_SYSCALL },
{ 27, CPU_FEATURE_RDTSCP },
{ 29, CPU_FEATURE_LM },
};
const struct feature_map_t matchtable_ecx81[] = {
{ 0, CPU_FEATURE_LAHF_LM },
};
const struct feature_map_t matchtable_edx87[] = {
{ 8, CPU_FEATURE_CONSTANT_TSC },
};
if (raw->basic_cpuid[0][0] >= 1) {
match_features(matchtable_edx1, COUNT_OF(matchtable_edx1), raw->basic_cpuid[1][3], data);
match_features(matchtable_ecx1, COUNT_OF(matchtable_ecx1), raw->basic_cpuid[1][2], data);
}
if (raw->basic_cpuid[0][0] >= 7) {
match_features(matchtable_ebx7, COUNT_OF(matchtable_ebx7), raw->basic_cpuid[7][1], data);
}
if (raw->ext_cpuid[0][0] >= 0x80000001) {
match_features(matchtable_edx81, COUNT_OF(matchtable_edx81), raw->ext_cpuid[1][3], data);
match_features(matchtable_ecx81, COUNT_OF(matchtable_ecx81), raw->ext_cpuid[1][2], data);
}
if (raw->ext_cpuid[0][0] >= 0x80000007) {
match_features(matchtable_edx87, COUNT_OF(matchtable_edx87), raw->ext_cpuid[7][3], data);
}
if (data->flags[CPU_FEATURE_SSE]) {
/* apply guesswork to check if the SSE unit width is 128 bit */
switch (data->vendor) {
case VENDOR_AMD:
data->sse_size = (data->ext_family >= 16 && data->ext_family != 17) ? 128 : 64;
break;
case VENDOR_INTEL:
data->sse_size = (data->family == 6 && data->ext_model >= 15) ? 128 : 64;
break;
default:
break;
}
/* leave the CPU_FEATURE_128BIT_SSE_AUTH 0; the advanced per-vendor detection routines
* will set it accordingly if they detect the needed bit */
}
}
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" },
};
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;
}
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);
if (data->vendor == VENDOR_UNKNOWN)
return set_error(ERR_CPU_UNKN);
basic = raw->basic_cpuid[0][0];
if (basic >= 1) {
data->family = (raw->basic_cpuid[1][0] >> 8) & 0xf;
data->model = (raw->basic_cpuid[1][0] >> 4) & 0xf;
data->stepping = raw->basic_cpuid[1][0] & 0xf;
xmodel = (raw->basic_cpuid[1][0] >> 16) & 0xf;
xfamily = (raw->basic_cpuid[1][0] >> 20) & 0xff;
if (data->vendor == VENDOR_AMD && data->family < 0xf)
data->ext_family = data->family;
else
data->ext_family = data->family + xfamily;
data->ext_model = data->model + (xmodel << 4);
}
ext = raw->ext_cpuid[0][0] - 0x8000000;
/* obtain the brand string, if present: */
if (ext >= 4) {
for (i = 0; i < 3; i++)
for (j = 0; j < 4; j++)
memcpy(brandstr + i * 16 + j * 4,
&raw->ext_cpuid[2 + i][j], 4);
brandstr[48] = 0;
i = 0;
while (brandstr[i] == ' ') i++;
strncpy(data->brand_str, brandstr + i, sizeof(data->brand_str));
data->brand_str[48] = 0;
}
load_features_common(raw, data);
data->total_logical_cpus = get_total_cpus();
return set_error(ERR_OK);
}
/* Interface: */
int cpuid_get_total_cpus(void)
{
return get_total_cpus();
}
int cpuid_present(void)
{
return cpuid_exists_by_eflags();
}
void cpu_exec_cpuid(uint32_t eax, uint32_t* regs)
{
regs[0] = eax;
regs[1] = regs[2] = regs[3] = 0;
exec_cpuid(regs);
}
void cpu_exec_cpuid_ext(uint32_t* regs)
{
exec_cpuid(regs);
}
int cpuid_get_raw_data(struct cpu_raw_data_t* data)
{
unsigned i;
if (!cpuid_present())
return set_error(ERR_NO_CPUID);
for (i = 0; i < 32; i++)
cpu_exec_cpuid(i, data->basic_cpuid[i]);
for (i = 0; i < 32; i++)
cpu_exec_cpuid(0x80000000 + i, data->ext_cpuid[i]);
for (i = 0; i < MAX_INTELFN4_LEVEL; i++) {
memset(data->intel_fn4[i], 0, sizeof(data->intel_fn4[i]));
data->intel_fn4[i][0] = 4;
data->intel_fn4[i][2] = i;
cpu_exec_cpuid_ext(data->intel_fn4[i]);
}
for (i = 0; i < MAX_INTELFN11_LEVEL; i++) {
memset(data->intel_fn11[i], 0, sizeof(data->intel_fn11[i]));
data->intel_fn11[i][0] = 11;
data->intel_fn11[i][2] = i;
cpu_exec_cpuid_ext(data->intel_fn11[i]);
}
for (i = 0; i < MAX_INTELFN12H_LEVEL; i++) {
memset(data->intel_fn12h[i], 0, sizeof(data->intel_fn12h[i]));
data->intel_fn12h[i][0] = 0x12;
data->intel_fn12h[i][2] = i;
cpu_exec_cpuid_ext(data->intel_fn12h[i]);
}
for (i = 0; i < MAX_INTELFN14H_LEVEL; i++) {
memset(data->intel_fn14h[i], 0, sizeof(data->intel_fn14h[i]));
data->intel_fn14h[i][0] = 0x14;
data->intel_fn14h[i][2] = i;
cpu_exec_cpuid_ext(data->intel_fn14h[i]);
}
return set_error(ERR_OK);
}
int cpu_ident_internal(struct cpu_raw_data_t* raw, struct cpu_id_t* data, struct internal_id_info_t* internal)
{
int r;
struct cpu_raw_data_t myraw;
if (!raw) {
if ((r = cpuid_get_raw_data(&myraw)) < 0)
return set_error(r);
raw = &myraw;
}
cpu_id_t_constructor(data);
if ((r = cpuid_basic_identify(raw, data)) < 0)
return set_error(r);
switch (data->vendor) {
case VENDOR_INTEL:
r = cpuid_identify_intel(raw, data, internal);
break;
case VENDOR_AMD:
r = cpuid_identify_amd(raw, data, internal);
break;
default:
break;
}
return set_error(r);
}
int cpu_identify(struct cpu_raw_data_t* raw, struct cpu_id_t* data)
{
struct internal_id_info_t throwaway;
return cpu_ident_internal(raw, data, &throwaway);
}
const char* cpuid_lib_version(void)
{
return VERSION;
}

58
src/3rdparty/libcpuid/intel_code_t.h

@ -1,58 +0,0 @@
/*
* Copyright 2016 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.
*/
/*
* This file contains a list of internal codes we use in detection. It is
* of no external use and isn't a complete list of intel products.
*/
CODE2(PENTIUM, 2000),
CODE(IRWIN),
CODE(POTOMAC),
CODE(GAINESTOWN),
CODE(WESTMERE),
CODE(PENTIUM_M),
CODE(NOT_CELERON),
CODE(CORE_SOLO),
CODE(MOBILE_CORE_SOLO),
CODE(CORE_DUO),
CODE(MOBILE_CORE_DUO),
CODE(WOLFDALE),
CODE(MEROM),
CODE(PENRYN),
CODE(QUAD_CORE),
CODE(DUAL_CORE_HT),
CODE(QUAD_CORE_HT),
CODE(MORE_THAN_QUADCORE),
CODE(PENTIUM_D),
CODE(SILVERTHORNE),
CODE(DIAMONDVILLE),
CODE(PINEVIEW),
CODE(CEDARVIEW),

678
src/3rdparty/libcpuid/libcpuid.h

@ -1,678 +0,0 @@
/*
* 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 __LIBCPUID_H__
#define __LIBCPUID_H__
/**
* \file libcpuid.h
* \author Veselin Georgiev
* \date Oct 2008
* \version 0.4.0
*
* Version history:
*
* * 0.1.0 (2008-10-15): initial adaptation from wxfractgui sources
* * 0.1.1 (2009-07-06): Added intel_fn11 fields to cpu_raw_data_t to handle
* new processor topology enumeration required on Core i7
* * 0.1.2 (2009-09-26): Added support for MSR reading through self-extracting
* kernel driver on Win32.
* * 0.1.3 (2010-04-20): Added support for greater more accurate CPU clock
* measurements with cpu_clock_by_ic()
* * 0.2.0 (2011-10-11): Support for AMD Bulldozer CPUs, 128-bit SSE unit size
* checking. A backwards-incompatible change, since the
* sizeof cpu_id_t is now different.
* * 0.2.1 (2012-05-26): Support for Ivy Bridge, and detecting the presence of
* the RdRand instruction.
* * 0.2.2 (2015-11-04): Support for newer processors up to Haswell and Vishera.
* Fix clock detection in cpu_clock_by_ic() for Bulldozer.
* More entries supported in cpu_msrinfo().
* *BSD and Solaris support (unofficial).
* * 0.3.0 (2016-07-09): Support for Skylake; MSR ops in FreeBSD; INFO_VOLTAGE
* for AMD CPUs. Level 4 cache support for Crystalwell
* (a backwards-incompatible change since the sizeof
* cpu_raw_data_t is now different).
* * 0.4.0 (2016-09-30): Better detection of AMD clock multiplier with msrinfo.
* Support for Intel SGX detection
* (a backwards-incompatible change since the sizeof
* cpu_raw_data_t and cpu_id_t is now different).
*/
/** @mainpage A simple libcpuid introduction
*
* LibCPUID provides CPU identification and access to the CPUID and RDTSC
* instructions on the x86.
* <p>
* To execute CPUID, use \ref cpu_exec_cpuid <br>
* To execute RDTSC, use \ref cpu_rdtsc <br>
* To fetch the CPUID info needed for CPU identification, use
* \ref cpuid_get_raw_data <br>
* To make sense of that data (decode, extract features), use \ref cpu_identify <br>
* To detect the CPU speed, use either \ref cpu_clock, \ref cpu_clock_by_os,
* \ref cpu_tsc_mark + \ref cpu_tsc_unmark + \ref cpu_clock_by_mark,
* \ref cpu_clock_measure or \ref cpu_clock_by_ic.
* Read carefully for pros/cons of each method. <br>
*
* To read MSRs, use \ref cpu_msr_driver_open to get a handle, and then
* \ref cpu_rdmsr for querying abilities. Some MSR decoding is available on recent
* CPUs, and can be queried through \ref cpu_msrinfo; the various types of queries
* are described in \ref cpu_msrinfo_request_t.
* </p>
*/
/** @defgroup libcpuid LibCPUID
* @brief LibCPUID provides CPU identification
@{ */
/* Include some integer type specifications: */
#include "libcpuid_types.h"
/* Some limits and other constants */
#include "libcpuid_constants.h"
#ifdef __cplusplus
extern "C" {
#endif
/**
* @brief CPU vendor, as guessed from the Vendor String.
*/
typedef enum {
VENDOR_INTEL = 0, /*!< Intel CPU */
VENDOR_AMD, /*!< AMD CPU */
VENDOR_CYRIX, /*!< Cyrix CPU */
VENDOR_NEXGEN, /*!< NexGen CPU */
VENDOR_TRANSMETA, /*!< Transmeta CPU */
VENDOR_UMC, /*!< x86 CPU by UMC */
VENDOR_CENTAUR, /*!< x86 CPU by IDT */
VENDOR_RISE, /*!< x86 CPU by Rise Technology */
VENDOR_SIS, /*!< x86 CPU by SiS */
VENDOR_NSC, /*!< x86 CPU by National Semiconductor */
NUM_CPU_VENDORS, /*!< Valid CPU vendor ids: 0..NUM_CPU_VENDORS - 1 */
VENDOR_UNKNOWN = -1,
} cpu_vendor_t;
#define NUM_CPU_VENDORS NUM_CPU_VENDORS
/**
* @brief Contains just the raw CPUID data.
*
* This contains only the most basic CPU data, required to do identification
* and feature recognition. Every processor should be identifiable using this
* data only.
*/
struct cpu_raw_data_t {
/** contains results of CPUID for eax = 0, 1, ...*/
uint32_t basic_cpuid[MAX_CPUID_LEVEL][4];
/** contains results of CPUID for eax = 0x80000000, 0x80000001, ...*/
uint32_t ext_cpuid[MAX_EXT_CPUID_LEVEL][4];
/** when the CPU is intel and it supports deterministic cache
information: this contains the results of CPUID for eax = 4
and ecx = 0, 1, ... */
uint32_t intel_fn4[MAX_INTELFN4_LEVEL][4];
/** when the CPU is intel and it supports leaf 0Bh (Extended Topology
enumeration leaf), this stores the result of CPUID with
eax = 11 and ecx = 0, 1, 2... */
uint32_t intel_fn11[MAX_INTELFN11_LEVEL][4];
/** when the CPU is intel and supports leaf 12h (SGX enumeration leaf),
* this stores the result of CPUID with eax = 0x12 and
* ecx = 0, 1, 2... */
uint32_t intel_fn12h[MAX_INTELFN12H_LEVEL][4];
/** when the CPU is intel and supports leaf 14h (Intel Processor Trace
* capabilities leaf).
* this stores the result of CPUID with eax = 0x12 and
* ecx = 0, 1, 2... */
uint32_t intel_fn14h[MAX_INTELFN14H_LEVEL][4];
};
/**
* @brief This contains information about SGX features of the processor
* Example usage:
* @code
* ...
* struct cpu_raw_data_t raw;
* struct cpu_id_t id;
*
* if (cpuid_get_raw_data(&raw) == 0 && cpu_identify(&raw, &id) == 0 && id.sgx.present) {
* printf("SGX is present.\n");
* printf("SGX1 instructions: %s.\n", id.sgx.flags[INTEL_SGX1] ? "present" : "absent");
* printf("SGX2 instructions: %s.\n", id.sgx.flags[INTEL_SGX2] ? "present" : "absent");
* printf("Max 32-bit enclave size: 2^%d bytes.\n", id.sgx.max_enclave_32bit);
* printf("Max 64-bit enclave size: 2^%d bytes.\n", id.sgx.max_enclave_64bit);
* for (int i = 0; i < id.sgx.num_epc_sections; i++) {
* struct cpu_epc_t epc = cpuid_get_epc(i, NULL);
* printf("EPC section #%d: address = %x, size = %d bytes.\n", epc.address, epc.size);
* }
* } else {
* printf("SGX is not present.\n");
* }
* @endcode
*/
struct cpu_sgx_t {
/** Whether SGX is present (boolean) */
uint32_t present;
/** Max enclave size in 32-bit mode. This is a power-of-two value:
* if it is "31", then the max enclave size is 2^31 bytes (2 GiB).
*/
uint8_t max_enclave_32bit;
/** Max enclave size in 64-bit mode. This is a power-of-two value:
* if it is "36", then the max enclave size is 2^36 bytes (64 GiB).
*/
uint8_t max_enclave_64bit;
/**
* contains SGX feature flags. See the \ref cpu_sgx_feature_t
* "INTEL_SGX*" macros below.
*/
uint8_t flags[SGX_FLAGS_MAX];
/** number of Enclave Page Cache (EPC) sections. Info for each
* section is available through the \ref cpuid_get_epc() function
*/
int num_epc_sections;
/** bit vector of the supported extended features that can be written
* to the MISC region of the SSA (Save State Area)
*/
uint32_t misc_select;
/** a bit vector of the attributes that can be set to SECS.ATTRIBUTES
* via ECREATE. Corresponds to bits 0-63 (incl.) of SECS.ATTRIBUTES.
*/
uint64_t secs_attributes;
/** a bit vector of the bits that can be set in the XSAVE feature
* request mask; Corresponds to bits 64-127 of SECS.ATTRIBUTES.
*/
uint64_t secs_xfrm;
};
/**
* @brief This contains the recognized CPU features/info
*/
struct cpu_id_t {
/** contains the CPU vendor string, e.g. "GenuineIntel" */
char vendor_str[VENDOR_STR_MAX];
/** contains the brand string, e.g. "Intel(R) Xeon(TM) CPU 2.40GHz" */
char brand_str[BRAND_STR_MAX];
/** contains the recognized CPU vendor */
cpu_vendor_t vendor;
/**
* contain CPU flags. Used to test for features. See
* the \ref cpu_feature_t "CPU_FEATURE_*" macros below.
* @see Features
*/
uint8_t flags[CPU_FLAGS_MAX];
/** CPU family */
int32_t family;
/** CPU model */
int32_t model;
/** CPU stepping */
int32_t stepping;
/** CPU extended family */
int32_t ext_family;
/** CPU extended model */
int32_t ext_model;
/** Number of CPU cores on the current processor */
int32_t num_cores;
/**
* Number of logical processors on the current processor.
* Could be more than the number of physical cores,
* e.g. when the processor has HyperThreading.
*/
int32_t num_logical_cpus;
/**
* The total number of logical processors.
* The same value is availabe through \ref cpuid_get_total_cpus.
*
* This is num_logical_cpus * {total physical processors in the system}
* (but only on a real system, under a VM this number may be lower).
*
* If you're writing a multithreaded program and you want to run it on
* all CPUs, this is the number of threads you need.
*
* @note in a VM, this will exactly match the number of CPUs set in
* the VM's configuration.
*
*/
int32_t total_logical_cpus;
/**
* L1 data cache size in KB. Could be zero, if the CPU lacks cache.
* If the size cannot be determined, it will be -1.
*/
int32_t l1_data_cache;
/**
* L1 instruction cache size in KB. Could be zero, if the CPU lacks
* cache. If the size cannot be determined, it will be -1.
* @note On some Intel CPUs, whose instruction cache is in fact
* a trace cache, the size will be expressed in K uOps.
*/
int32_t l1_instruction_cache;
/**
* L2 cache size in KB. Could be zero, if the CPU lacks L2 cache.
* If the size of the cache could not be determined, it will be -1
*/
int32_t l2_cache;
/** L3 cache size in KB. Zero on most systems */
int32_t l3_cache;
/** L4 cache size in KB. Zero on most systems */
int32_t l4_cache;
/** Cache associativity for the L1 data cache. -1 if undetermined */
int32_t l1_assoc;
/** Cache associativity for the L2 cache. -1 if undetermined */
int32_t l2_assoc;
/** Cache associativity for the L3 cache. -1 if undetermined */
int32_t l3_assoc;
/** Cache associativity for the L4 cache. -1 if undetermined */
int32_t l4_assoc;
/** Cache-line size for L1 data cache. -1 if undetermined */
int32_t l1_cacheline;
/** Cache-line size for L2 cache. -1 if undetermined */
int32_t l2_cacheline;
/** Cache-line size for L3 cache. -1 if undetermined */
int32_t l3_cacheline;
/** Cache-line size for L4 cache. -1 if undetermined */
int32_t l4_cacheline;
/**
* The brief and human-friendly CPU codename, which was recognized.<br>
* Examples:
* @code
* +--------+--------+-------+-------+-------+---------------------------------------+-----------------------+
* | Vendor | Family | Model | Step. | Cache | Brand String | cpu_id_t.cpu_codename |
* +--------+--------+-------+-------+-------+---------------------------------------+-----------------------+
* | AMD | 6 | 8 | 0 | 256 | (not available - will be ignored) | "K6-2" |
* | Intel | 15 | 2 | 5 | 512 | "Intel(R) Xeon(TM) CPU 2.40GHz" | "Xeon (Prestonia)" |
* | Intel | 6 | 15 | 11 | 4096 | "Intel(R) Core(TM)2 Duo CPU E6550..." | "Conroe (Core 2 Duo)" |
* | AMD | 15 | 35 | 2 | 1024 | "Dual Core AMD Opteron(tm) Proces..." | "Opteron (Dual Core)" |
* +--------+--------+-------+-------+-------+---------------------------------------+-----------------------+
* @endcode
*/
char cpu_codename[64];
/** SSE execution unit size (64 or 128; -1 if N/A) */
int32_t sse_size;
/**
* contain miscellaneous detection information. Used to test about specifics of
* certain detected features. See \ref cpu_hint_t "CPU_HINT_*" macros below.
* @see Hints
*/
uint8_t detection_hints[CPU_HINTS_MAX];
/** contains information about SGX features if the processor, if present */
struct cpu_sgx_t sgx;
};
/**
* @brief CPU feature identifiers
*
* Usage:
* @code
* ...
* struct cpu_raw_data_t raw;
* struct cpu_id_t id;
* if (cpuid_get_raw_data(&raw) == 0 && cpu_identify(&raw, &id) == 0) {
* if (id.flags[CPU_FEATURE_SSE2]) {
* // The CPU has SSE2...
* ...
* } else {
* // no SSE2
* }
* } else {
* // processor cannot be determined.
* }
* @endcode
*/
typedef enum {
CPU_FEATURE_FPU = 0, /*!< Floating point unit */
CPU_FEATURE_VME, /*!< Virtual mode extension */
CPU_FEATURE_DE, /*!< Debugging extension */
CPU_FEATURE_PSE, /*!< Page size extension */
CPU_FEATURE_TSC, /*!< Time-stamp counter */
CPU_FEATURE_MSR, /*!< Model-specific regsisters, RDMSR/WRMSR supported */
CPU_FEATURE_PAE, /*!< Physical address extension */
CPU_FEATURE_MCE, /*!< Machine check exception */
CPU_FEATURE_CX8, /*!< CMPXCHG8B instruction supported */
CPU_FEATURE_APIC, /*!< APIC support */
CPU_FEATURE_MTRR, /*!< Memory type range registers */
CPU_FEATURE_SEP, /*!< SYSENTER / SYSEXIT instructions supported */
CPU_FEATURE_PGE, /*!< Page global enable */
CPU_FEATURE_MCA, /*!< Machine check architecture */
CPU_FEATURE_CMOV, /*!< CMOVxx instructions supported */
CPU_FEATURE_PAT, /*!< Page attribute table */
CPU_FEATURE_PSE36, /*!< 36-bit page address extension */
CPU_FEATURE_PN, /*!< Processor serial # implemented (Intel P3 only) */
CPU_FEATURE_CLFLUSH, /*!< CLFLUSH instruction supported */
CPU_FEATURE_DTS, /*!< Debug store supported */
CPU_FEATURE_ACPI, /*!< ACPI support (power states) */
CPU_FEATURE_MMX, /*!< MMX instruction set supported */
CPU_FEATURE_FXSR, /*!< FXSAVE / FXRSTOR supported */
CPU_FEATURE_SSE, /*!< Streaming-SIMD Extensions (SSE) supported */
CPU_FEATURE_SSE2, /*!< SSE2 instructions supported */
CPU_FEATURE_SS, /*!< Self-snoop */
CPU_FEATURE_HT, /*!< Hyper-threading supported (but might be disabled) */
CPU_FEATURE_TM, /*!< Thermal monitor */
CPU_FEATURE_IA64, /*!< IA64 supported (Itanium only) */
CPU_FEATURE_PBE, /*!< Pending-break enable */
CPU_FEATURE_PNI, /*!< PNI (SSE3) instructions supported */
CPU_FEATURE_PCLMUL, /*!< PCLMULQDQ instruction supported */
CPU_FEATURE_DTS64, /*!< 64-bit Debug store supported */
CPU_FEATURE_MONITOR, /*!< MONITOR / MWAIT supported */
CPU_FEATURE_DS_CPL, /*!< CPL Qualified Debug Store */
CPU_FEATURE_VMX, /*!< Virtualization technology supported */
CPU_FEATURE_SMX, /*!< Safer mode exceptions */
CPU_FEATURE_EST, /*!< Enhanced SpeedStep */
CPU_FEATURE_TM2, /*!< Thermal monitor 2 */
CPU_FEATURE_SSSE3, /*!< SSSE3 instructionss supported (this is different from SSE3!) */
CPU_FEATURE_CID, /*!< Context ID supported */
CPU_FEATURE_CX16, /*!< CMPXCHG16B instruction supported */
CPU_FEATURE_XTPR, /*!< Send Task Priority Messages disable */
CPU_FEATURE_PDCM, /*!< Performance capabilities MSR supported */
CPU_FEATURE_DCA, /*!< Direct cache access supported */
CPU_FEATURE_SSE4_1, /*!< SSE 4.1 instructions supported */
CPU_FEATURE_SSE4_2, /*!< SSE 4.2 instructions supported */
CPU_FEATURE_SYSCALL, /*!< SYSCALL / SYSRET instructions supported */
CPU_FEATURE_XD, /*!< Execute disable bit supported */
CPU_FEATURE_MOVBE, /*!< MOVBE instruction supported */
CPU_FEATURE_POPCNT, /*!< POPCNT instruction supported */
CPU_FEATURE_AES, /*!< AES* instructions supported */
CPU_FEATURE_XSAVE, /*!< XSAVE/XRSTOR/etc instructions supported */
CPU_FEATURE_OSXSAVE, /*!< non-privileged copy of OSXSAVE supported */
CPU_FEATURE_AVX, /*!< Advanced vector extensions supported */
CPU_FEATURE_MMXEXT, /*!< AMD MMX-extended instructions supported */
CPU_FEATURE_3DNOW, /*!< AMD 3DNow! instructions supported */
CPU_FEATURE_3DNOWEXT, /*!< AMD 3DNow! extended instructions supported */
CPU_FEATURE_NX, /*!< No-execute bit supported */
CPU_FEATURE_FXSR_OPT, /*!< FFXSR: FXSAVE and FXRSTOR optimizations */
CPU_FEATURE_RDTSCP, /*!< RDTSCP instruction supported (AMD-only) */
CPU_FEATURE_LM, /*!< Long mode (x86_64/EM64T) supported */
CPU_FEATURE_LAHF_LM, /*!< LAHF/SAHF supported in 64-bit mode */
CPU_FEATURE_CMP_LEGACY, /*!< core multi-processing legacy mode */
CPU_FEATURE_SVM, /*!< AMD Secure virtual machine */
CPU_FEATURE_ABM, /*!< LZCNT instruction support */
CPU_FEATURE_MISALIGNSSE,/*!< Misaligned SSE supported */
CPU_FEATURE_SSE4A, /*!< SSE 4a from AMD */
CPU_FEATURE_3DNOWPREFETCH, /*!< PREFETCH/PREFETCHW support */
CPU_FEATURE_OSVW, /*!< OS Visible Workaround (AMD) */
CPU_FEATURE_IBS, /*!< Instruction-based sampling */
CPU_FEATURE_SSE5, /*!< SSE 5 instructions supported (deprecated, will never be 1) */
CPU_FEATURE_SKINIT, /*!< SKINIT / STGI supported */
CPU_FEATURE_WDT, /*!< Watchdog timer support */
CPU_FEATURE_TS, /*!< Temperature sensor */
CPU_FEATURE_FID, /*!< Frequency ID control */
CPU_FEATURE_VID, /*!< Voltage ID control */
CPU_FEATURE_TTP, /*!< THERMTRIP */
CPU_FEATURE_TM_AMD, /*!< AMD-specified hardware thermal control */
CPU_FEATURE_STC, /*!< Software thermal control */
CPU_FEATURE_100MHZSTEPS,/*!< 100 MHz multiplier control */
CPU_FEATURE_HWPSTATE, /*!< Hardware P-state control */
CPU_FEATURE_CONSTANT_TSC, /*!< TSC ticks at constant rate */
CPU_FEATURE_XOP, /*!< The XOP instruction set (same as the old CPU_FEATURE_SSE5) */
CPU_FEATURE_FMA3, /*!< The FMA3 instruction set */
CPU_FEATURE_FMA4, /*!< The FMA4 instruction set */
CPU_FEATURE_TBM, /*!< Trailing bit manipulation instruction support */
CPU_FEATURE_F16C, /*!< 16-bit FP convert instruction support */
CPU_FEATURE_RDRAND, /*!< RdRand instruction */
CPU_FEATURE_X2APIC, /*!< x2APIC, APIC_BASE.EXTD, MSRs 0000_0800h...0000_0BFFh 64-bit ICR (+030h but not +031h), no DFR (+00Eh), SELF_IPI (+040h) also see standard level 0000_000Bh */
CPU_FEATURE_CPB, /*!< Core performance boost */
CPU_FEATURE_APERFMPERF, /*!< MPERF/APERF MSRs support */
CPU_FEATURE_PFI, /*!< Processor Feedback Interface support */
CPU_FEATURE_PA, /*!< Processor accumulator */
CPU_FEATURE_AVX2, /*!< AVX2 instructions */
CPU_FEATURE_BMI1, /*!< BMI1 instructions */
CPU_FEATURE_BMI2, /*!< BMI2 instructions */
CPU_FEATURE_HLE, /*!< Hardware Lock Elision prefixes */
CPU_FEATURE_RTM, /*!< Restricted Transactional Memory instructions */
CPU_FEATURE_AVX512F, /*!< AVX-512 Foundation */
CPU_FEATURE_AVX512DQ, /*!< AVX-512 Double/Quad granular insns */
CPU_FEATURE_AVX512PF, /*!< AVX-512 Prefetch */
CPU_FEATURE_AVX512ER, /*!< AVX-512 Exponential/Reciprocal */
CPU_FEATURE_AVX512CD, /*!< AVX-512 Conflict detection */
CPU_FEATURE_SHA_NI, /*!< SHA-1/SHA-256 instructions */
CPU_FEATURE_AVX512BW, /*!< AVX-512 Byte/Word granular insns */
CPU_FEATURE_AVX512VL, /*!< AVX-512 128/256 vector length extensions */
CPU_FEATURE_SGX, /*!< SGX extensions. Non-autoritative, check cpu_id_t::sgx::present to verify presence */
CPU_FEATURE_RDSEED, /*!< RDSEED instruction */
CPU_FEATURE_ADX, /*!< ADX extensions (arbitrary precision) */
/* termination: */
NUM_CPU_FEATURES,
} cpu_feature_t;
/**
* @brief CPU detection hints identifiers
*
* Usage: similar to the flags usage
*/
typedef enum {
CPU_HINT_SSE_SIZE_AUTH = 0, /*!< SSE unit size is authoritative (not only a Family/Model guesswork, but based on an actual CPUID bit) */
/* termination */
NUM_CPU_HINTS,
} cpu_hint_t;
/**
* @brief SGX features flags
* \see cpu_sgx_t
*
* Usage:
* @code
* ...
* struct cpu_raw_data_t raw;
* struct cpu_id_t id;
* if (cpuid_get_raw_data(&raw) == 0 && cpu_identify(&raw, &id) == 0 && id.sgx.present) {
* if (id.sgx.flags[INTEL_SGX1])
* // The CPU has SGX1 instructions support...
* ...
* } else {
* // no SGX
* }
* } else {
* // processor cannot be determined.
* }
* @endcode
*/
typedef enum {
INTEL_SGX1, /*!< SGX1 instructions support */
INTEL_SGX2, /*!< SGX2 instructions support */
/* termination: */
NUM_SGX_FEATURES,
} cpu_sgx_feature_t;
/**
* @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 */
} cpu_error_t;
/**
* @brief Internal structure, used in cpu_tsc_mark, cpu_tsc_unmark and
* cpu_clock_by_mark
*/
struct cpu_mark_t {
uint64_t tsc; /*!< Time-stamp from RDTSC */
uint64_t sys_clock; /*!< In microsecond resolution */
};
/**
* @brief Returns the total number of logical CPU threads (even if CPUID is not present).
*
* Under VM, this number (and total_logical_cpus, since they are fetched with the same code)
* may be nonsensical, i.e. might not equal NumPhysicalCPUs*NumCoresPerCPU*HyperThreading.
* This is because no matter how many logical threads the host machine has, you may limit them
* in the VM to any number you like. **This** is the number returned by cpuid_get_total_cpus().
*
* @returns Number of logical CPU threads available. Equals the \ref cpu_id_t::total_logical_cpus.
*/
int cpuid_get_total_cpus(void);
/**
* @brief Checks if the CPUID instruction is supported
* @retval 1 if CPUID is present
* @retval 0 the CPU doesn't have CPUID.
*/
int cpuid_present(void);
/**
* @brief Executes the CPUID instruction
* @param eax - the value of the EAX register when executing CPUID
* @param regs - the results will be stored here. regs[0] = EAX, regs[1] = EBX, ...
* @note CPUID will be executed with EAX set to the given value and EBX, ECX,
* EDX set to zero.
*/
void cpu_exec_cpuid(uint32_t eax, uint32_t* regs);
/**
* @brief Executes the CPUID instruction with the given input registers
* @note This is just a bit more generic version of cpu_exec_cpuid - it allows
* you to control all the registers.
* @param regs - Input/output. Prior to executing CPUID, EAX, EBX, ECX and
* EDX will be set to regs[0], regs[1], regs[2] and regs[3].
* After CPUID, this array will contain the results.
*/
void cpu_exec_cpuid_ext(uint32_t* regs);
/**
* @brief Obtains the raw CPUID data from the current CPU
* @param data - a pointer to cpu_raw_data_t structure
* @returns zero if successful, and some negative number on error.
* The error message can be obtained by calling \ref cpuid_error.
* @see cpu_error_t
*/
int cpuid_get_raw_data(struct cpu_raw_data_t* data);
/**
* @brief Identifies the CPU
* @param raw - Input - a pointer to the raw CPUID data, which is obtained
* either by cpuid_get_raw_data or cpuid_deserialize_raw_data.
* Can also be NULL, in which case the functions calls
* cpuid_get_raw_data itself.
* @param data - Output - the decoded CPU features/info is written here.
* @note The function will not fail, even if some of the information
* cannot be obtained. Even when the CPU is new and thus unknown to
* libcpuid, some generic info, such as "AMD K9 family CPU" will be
* written to data.cpu_codename, and most other things, such as the
* CPU flags, cache sizes, etc. should be detected correctly anyway.
* However, the function CAN fail, if the CPU is completely alien to
* libcpuid.
* @note While cpu_identify() and cpuid_get_raw_data() are fast for most
* purposes, running them several thousand times per second can hamper
* performance significantly. Specifically, avoid writing "cpu feature
* checker" wrapping function, which calls cpu_identify and returns the
* value of some flag, if that function is going to be called frequently.
* @returns zero if successful, and some negative number on error.
* The error message can be obtained by calling \ref cpuid_error.
* @see cpu_error_t
*/
int cpu_identify(struct cpu_raw_data_t* raw, struct cpu_id_t* data);
/**
* @brief The return value of cpuid_get_epc().
* @details
* Describes an EPC (Enclave Page Cache) layout (physical address and size).
* A CPU may have one or more EPC areas, and information about each is
* fetched via \ref cpuid_get_epc.
*/
struct cpu_epc_t {
uint64_t start_addr;
uint64_t length;
};
/**
* @brief Fetches information about an EPC (Enclave Page Cache) area.
* @param index - zero-based index, valid range [0..cpu_id_t.egx.num_epc_sections)
* @param raw - a pointer to fetched raw CPUID data. Needed only for testing,
* you can safely pass NULL here (if you pass a real structure,
* it will be used for fetching the leaf 12h data if index < 2;
* otherwise the real CPUID instruction will be used).
* @returns the requested data. If the CPU doesn't support SGX, or if
* index >= cpu_id_t.egx.num_epc_sections, both fields of the returned
* structure will be zeros.
*/
struct cpu_epc_t cpuid_get_epc(int index, const struct cpu_raw_data_t* raw);
/**
* @brief Returns the libcpuid version
*
* @returns the string representation of the libcpuid version, like "0.1.1"
*/
const char* cpuid_lib_version(void);
#ifdef __cplusplus
} /* extern "C" */
#endif
/** @} */
#endif /* __LIBCPUID_H__ */

47
src/3rdparty/libcpuid/libcpuid_constants.h

@ -1,47 +0,0 @@
/*
* 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.
*/
/**
* @File libcpuid_constants.h
* @Author Veselin Georgiev
* @Brief Some limits and constants for libcpuid
*/
#ifndef __LIBCPUID_CONSTANTS_H__
#define __LIBCPUID_CONSTANTS_H__
#define VENDOR_STR_MAX 16
#define BRAND_STR_MAX 64
#define CPU_FLAGS_MAX 128
#define MAX_CPUID_LEVEL 32
#define MAX_EXT_CPUID_LEVEL 32
#define MAX_INTELFN4_LEVEL 8
#define MAX_INTELFN11_LEVEL 4
#define MAX_INTELFN12H_LEVEL 4
#define MAX_INTELFN14H_LEVEL 4
#define CPU_HINTS_MAX 16
#define SGX_FLAGS_MAX 14
#endif /* __LIBCPUID_CONSTANTS_H__ */

107
src/3rdparty/libcpuid/libcpuid_internal.h

@ -1,107 +0,0 @@
/*
* Copyright 2016 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 __LIBCPUID_INTERNAL_H__
#define __LIBCPUID_INTERNAL_H__
/*
* This file contains internal undocumented declarations and function prototypes
* for the workings of the internal library infrastructure.
*/
enum _common_codes_t {
NA = 0,
NC, /* No code */
};
#define CODE(x) x
#define CODE2(x, y) x = y
enum _amd_code_t {
#include "amd_code_t.h"
};
typedef enum _amd_code_t amd_code_t;
enum _intel_code_t {
#include "intel_code_t.h"
};
typedef enum _intel_code_t intel_code_t;
#undef CODE
#undef CODE2
struct internal_id_info_t {
union {
amd_code_t amd;
intel_code_t intel;
} code;
uint64_t bits;
int score; // detection (matchtable) score
};
#define LBIT(x) (((long long) 1) << x)
enum _common_bits_t {
_M_ = LBIT( 0 ),
MOBILE_ = LBIT( 1 ),
_MP_ = LBIT( 2 ),
};
// additional detection bits for Intel CPUs:
enum _intel_bits_t {
PENTIUM_ = LBIT( 10 ),
CELERON_ = LBIT( 11 ),
CORE_ = LBIT( 12 ),
_I_ = LBIT( 13 ),
_3 = LBIT( 14 ),
_5 = LBIT( 15 ),
_7 = LBIT( 16 ),
_9 = LBIT( 17 ),
XEON_ = LBIT( 18 ),
ATOM_ = LBIT( 19 ),
};
typedef enum _intel_bits_t intel_bits_t;
enum _amd_bits_t {
ATHLON_ = LBIT( 10 ),
_XP_ = LBIT( 11 ),
DURON_ = LBIT( 12 ),
SEMPRON_ = LBIT( 13 ),
OPTERON_ = LBIT( 14 ),
TURION_ = LBIT( 15 ),
_LV_ = LBIT( 16 ),
_64_ = LBIT( 17 ),
_X2 = LBIT( 18 ),
_X3 = LBIT( 19 ),
_X4 = LBIT( 20 ),
_X6 = LBIT( 21 ),
_FX = LBIT( 22 ),
_APU_ = LBIT( 23 ),
};
typedef enum _amd_bits_t amd_bits_t;
int cpu_ident_internal(struct cpu_raw_data_t* raw, struct cpu_id_t* data,
struct internal_id_info_t* internal);
#endif /* __LIBCPUID_INTERNAL_H__ */

63
src/3rdparty/libcpuid/libcpuid_types.h

@ -1,63 +0,0 @@
/*
* 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.
*/
/**
* @File libcpuid_types.h
* @Author Veselin Georgiev
* @Brief Type specifications for libcpuid.
*/
#ifndef __LIBCPUID_TYPES_H__
#define __LIBCPUID_TYPES_H__
#if !defined(_MSC_VER) || _MSC_VER >= 1600
# include <stdint.h>
#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__ */

93
src/3rdparty/libcpuid/libcpuid_util.c

@ -1,93 +0,0 @@
/*
* 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 <stdio.h>
#include <stdlib.h>
#include <stdarg.h>
#include <string.h>
#include <ctype.h>
#include "libcpuid.h"
#include "libcpuid_util.h"
void match_features(const struct feature_map_t* matchtable, int count, uint32_t reg, struct cpu_id_t* data)
{
int i;
for (i = 0; i < count; i++)
if (reg & (1u << matchtable[i].bit))
data->flags[matchtable[i].feature] = 1;
}
static int xmatch_entry(char c, const char* p)
{
int i, j;
if (c == 0) return -1;
if (c == p[0]) return 1;
if (p[0] == '.') return 1;
if (p[0] == '#' && isdigit(c)) return 1;
if (p[0] == '[') {
j = 1;
while (p[j] && p[j] != ']') j++;
if (!p[j]) return -1;
for (i = 1; i < j; i++)
if (p[i] == c) return j + 1;
}
return -1;
}
int match_pattern(const char* s, const char* p)
{
int i, j, dj, k, n, m;
n = (int) strlen(s);
m = (int) strlen(p);
for (i = 0; i < n; i++) {
if (xmatch_entry(s[i], p) != -1) {
j = 0;
k = 0;
while (j < m && ((dj = xmatch_entry(s[i + k], p + j)) != -1)) {
k++;
j += dj;
}
if (j == m) return i + 1;
}
}
return 0;
}
struct cpu_id_t* get_cached_cpuid(void)
{
static int initialized = 0;
static struct cpu_id_t id;
if (initialized) return &id;
if (cpu_identify(NULL, &id))
memset(&id, 0, sizeof(id));
initialized = 1;
return &id;
}
int match_all(uint64_t bits, uint64_t mask)
{
return (bits & mask) == mask;
}

78
src/3rdparty/libcpuid/libcpuid_util.h

@ -1,78 +0,0 @@
/*
* 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 __LIBCPUID_UTIL_H__
#define __LIBCPUID_UTIL_H__
#define COUNT_OF(array) (sizeof(array) / sizeof(array[0]))
struct feature_map_t {
unsigned bit;
cpu_feature_t feature;
};
void match_features(const struct feature_map_t* matchtable, int count,
uint32_t reg, struct cpu_id_t* data);
struct match_entry_t {
int family, model, stepping, ext_family, ext_model;
int ncores, l2cache, l3cache, brand_code;
uint64_t model_bits;
int model_code;
char name[32];
};
// returns the match score:
int match_cpu_codename(const struct match_entry_t* matchtable, int count,
struct cpu_id_t* data, int brand_code, uint64_t bits,
int model_code);
/*
* Seek for a pattern in `haystack'.
* Pattern may be an fixed string, or contain the special metacharacters
* '.' - match any single character
* '#' - match any digit
* '[<chars>] - match any of the given chars (regex-like ranges are not
* supported)
* Return val: 0 if the pattern is not found. Nonzero if it is found (actually,
* x + 1 where x is the index where the match is found).
*/
int match_pattern(const char* haystack, const char* pattern);
/*
* Gets an initialized cpu_id_t. It is cached, so that internal libcpuid
* machinery doesn't need to issue cpu_identify more than once.
*/
struct cpu_id_t* get_cached_cpuid(void);
/* returns true if all bits of mask are present in `bits'. */
int match_all(uint64_t bits, uint64_t mask);
/*
* Sets the current errno
*/
int set_error(cpu_error_t err);
#endif /* __LIBCPUID_UTIL_H__ */

359
src/3rdparty/libcpuid/masm-x64.asm

@ -1,359 +0,0 @@
.code
; procedure exec_cpuid
; Signature: void exec_cpiud(uint32_t *regs)
exec_cpuid Proc
push rbx
push rcx
push rdx
push rdi
mov rdi, rcx
mov eax, [rdi]
mov ebx, [rdi+4]
mov ecx, [rdi+8]
mov edx, [rdi+12]
cpuid
mov [rdi], eax
mov [rdi+4], ebx
mov [rdi+8], ecx
mov [rdi+12], edx
pop rdi
pop rdx
pop rcx
pop rbx
ret
exec_cpuid endp
; procedure cpu_rdtsc
; Signature: void cpu_rdtsc(uint64_t *result)
cpu_rdtsc Proc
push rdx
rdtsc
mov [rcx], eax
mov [rcx+4], edx
pop rdx
ret
cpu_rdtsc endp
; procedure busy_sse_loop
; Signature: void busy_sse_loop(int cycles)
busy_sse_loop Proc
; save xmm6 & xmm7 into the shadow area, as Visual C++ 2008
; expects that we don't touch them:
movups [rsp + 8], xmm6
movups [rsp + 24], xmm7
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 ecx
jnz bsLoop
; restore xmm6 & xmm7:
movups xmm6, [rsp + 8]
movups xmm7, [rsp + 24]
ret
busy_sse_loop endp
END

173
src/3rdparty/libcpuid/recog_amd.c

@ -1,173 +0,0 @@
/*
* 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 <stdio.h>
#include <string.h>
#include <ctype.h>
#include "libcpuid.h"
#include "libcpuid_util.h"
#include "libcpuid_internal.h"
#include "recog_amd.h"
const struct amd_code_str { amd_code_t code; char *str; } amd_code_str[] = {
#define CODE(x) { x, #x }
#define CODE2(x, y) CODE(x)
#include "amd_code_t.h"
#undef CODE
};
struct amd_code_and_bits_t {
int code;
uint64_t bits;
};
enum _amd_model_codes_t {
// Only for Ryzen CPUs:
_1400,
_1500,
_1600,
_1900,
_2400,
_2500,
_2700,
};
static void load_amd_features(struct cpu_raw_data_t* raw, struct cpu_id_t* data)
{
const struct feature_map_t matchtable_edx81[] = {
{ 20, CPU_FEATURE_NX },
{ 22, CPU_FEATURE_MMXEXT },
{ 25, CPU_FEATURE_FXSR_OPT },
{ 30, CPU_FEATURE_3DNOWEXT },
{ 31, CPU_FEATURE_3DNOW },
};
const struct feature_map_t matchtable_ecx81[] = {
{ 1, CPU_FEATURE_CMP_LEGACY },
{ 2, CPU_FEATURE_SVM },
{ 5, CPU_FEATURE_ABM },
{ 6, CPU_FEATURE_SSE4A },
{ 7, CPU_FEATURE_MISALIGNSSE },
{ 8, CPU_FEATURE_3DNOWPREFETCH },
{ 9, CPU_FEATURE_OSVW },
{ 10, CPU_FEATURE_IBS },
{ 11, CPU_FEATURE_XOP },
{ 12, CPU_FEATURE_SKINIT },
{ 13, CPU_FEATURE_WDT },
{ 16, CPU_FEATURE_FMA4 },
{ 21, CPU_FEATURE_TBM },
};
const struct feature_map_t matchtable_edx87[] = {
{ 0, CPU_FEATURE_TS },
{ 1, CPU_FEATURE_FID },
{ 2, CPU_FEATURE_VID },
{ 3, CPU_FEATURE_TTP },
{ 4, CPU_FEATURE_TM_AMD },
{ 5, CPU_FEATURE_STC },
{ 6, CPU_FEATURE_100MHZSTEPS },
{ 7, CPU_FEATURE_HWPSTATE },
/* id 8 is handled in common */
{ 9, CPU_FEATURE_CPB },
{ 10, CPU_FEATURE_APERFMPERF },
{ 11, CPU_FEATURE_PFI },
{ 12, CPU_FEATURE_PA },
};
if (raw->ext_cpuid[0][0] >= 0x80000001) {
match_features(matchtable_edx81, COUNT_OF(matchtable_edx81), raw->ext_cpuid[1][3], data);
match_features(matchtable_ecx81, COUNT_OF(matchtable_ecx81), raw->ext_cpuid[1][2], data);
}
if (raw->ext_cpuid[0][0] >= 0x80000007)
match_features(matchtable_edx87, COUNT_OF(matchtable_edx87), raw->ext_cpuid[7][3], data);
if (raw->ext_cpuid[0][0] >= 0x8000001a) {
/* We have the extended info about SSE unit size */
data->detection_hints[CPU_HINT_SSE_SIZE_AUTH] = 1;
data->sse_size = (raw->ext_cpuid[0x1a][0] & 1) ? 128 : 64;
}
}
static void decode_amd_cache_info(struct cpu_raw_data_t* raw, struct cpu_id_t* data)
{
int l3_result;
const int assoc_table[16] = {
0, 1, 2, 0, 4, 0, 8, 0, 16, 0, 32, 48, 64, 96, 128, 255
};
unsigned n = raw->ext_cpuid[0][0];
if (n >= 0x80000005) {
data->l1_data_cache = (raw->ext_cpuid[5][2] >> 24) & 0xff;
data->l1_assoc = (raw->ext_cpuid[5][2] >> 16) & 0xff;
data->l1_cacheline = (raw->ext_cpuid[5][2]) & 0xff;
data->l1_instruction_cache = (raw->ext_cpuid[5][3] >> 24) & 0xff;
}
if (n >= 0x80000006) {
data->l2_cache = (raw->ext_cpuid[6][2] >> 16) & 0xffff;
data->l2_assoc = assoc_table[(raw->ext_cpuid[6][2] >> 12) & 0xf];
data->l2_cacheline = (raw->ext_cpuid[6][2]) & 0xff;
l3_result = (raw->ext_cpuid[6][3] >> 18);
if (l3_result > 0) {
l3_result = 512 * l3_result; /* AMD spec says it's a range,
but we take the lower bound */
data->l3_cache = l3_result;
data->l3_assoc = assoc_table[(raw->ext_cpuid[6][3] >> 12) & 0xf];
data->l3_cacheline = (raw->ext_cpuid[6][3]) & 0xff;
} else {
data->l3_cache = 0;
}
}
}
static void decode_amd_number_of_cores(struct cpu_raw_data_t* raw, struct cpu_id_t* data)
{
int logical_cpus = -1, num_cores = -1;
if (raw->basic_cpuid[0][0] >= 1) {
logical_cpus = (raw->basic_cpuid[1][1] >> 16) & 0xff;
if (raw->ext_cpuid[0][0] >= 8) {
num_cores = 1 + (raw->ext_cpuid[8][2] & 0xff);
}
}
if (data->flags[CPU_FEATURE_HT]) {
if (num_cores > 1) {
if (data->ext_family >= 23)
num_cores /= 2; // e.g., Ryzen 7 reports 16 "real" cores, but they are really just 8.
data->num_cores = num_cores;
data->num_logical_cpus = logical_cpus;
} else {
data->num_cores = 1;
data->num_logical_cpus = (logical_cpus >= 2 ? logical_cpus : 2);
}
} else {
data->num_cores = data->num_logical_cpus = 1;
}
}
int cpuid_identify_amd(struct cpu_raw_data_t* raw, struct cpu_id_t* data, struct internal_id_info_t* internal)
{
load_amd_features(raw, data);
decode_amd_cache_info(raw, data);
decode_amd_number_of_cores(raw, data);
return 0;
}

31
src/3rdparty/libcpuid/recog_amd.h

@ -1,31 +0,0 @@
/*
* 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 __RECOG_AMD_H__
#define __RECOG_AMD_H__
int cpuid_identify_amd(struct cpu_raw_data_t* raw, struct cpu_id_t* data, struct internal_id_info_t* internal);
#endif /* __RECOG_AMD_H__ */

543
src/3rdparty/libcpuid/recog_intel.c

@ -1,543 +0,0 @@
/*
* 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 <string.h>
#include <ctype.h>
#include "libcpuid.h"
#include "libcpuid_util.h"
#include "libcpuid_internal.h"
#include "recog_intel.h"
const struct intel_bcode_str { intel_code_t code; char *str; } intel_bcode_str[] = {
#define CODE(x) { x, #x }
#define CODE2(x, y) CODE(x)
#include "intel_code_t.h"
#undef CODE
};
typedef struct {
int code;
uint64_t bits;
} intel_code_and_bits_t;
enum _intel_model_t {
UNKNOWN = -1,
_3000 = 100,
_3100,
_3200,
X3200,
_3300,
X3300,
_5100,
_5200,
_5300,
_5400,
_2xxx, /* Core i[357] 2xxx */
_3xxx, /* Core i[357] 3xxx */
};
typedef enum _intel_model_t intel_model_t;
static void load_intel_features(struct cpu_raw_data_t* raw, struct cpu_id_t* data)
{
const struct feature_map_t matchtable_edx1[] = {
{ 18, CPU_FEATURE_PN },
{ 21, CPU_FEATURE_DTS },
{ 22, CPU_FEATURE_ACPI },
{ 27, CPU_FEATURE_SS },
{ 29, CPU_FEATURE_TM },
{ 30, CPU_FEATURE_IA64 },
{ 31, CPU_FEATURE_PBE },
};
const struct feature_map_t matchtable_ecx1[] = {
{ 2, CPU_FEATURE_DTS64 },
{ 4, CPU_FEATURE_DS_CPL },
{ 5, CPU_FEATURE_VMX },
{ 6, CPU_FEATURE_SMX },
{ 7, CPU_FEATURE_EST },
{ 8, CPU_FEATURE_TM2 },
{ 10, CPU_FEATURE_CID },
{ 14, CPU_FEATURE_XTPR },
{ 15, CPU_FEATURE_PDCM },
{ 18, CPU_FEATURE_DCA },
{ 21, CPU_FEATURE_X2APIC },
};
const struct feature_map_t matchtable_edx81[] = {
{ 20, CPU_FEATURE_XD },
};
const struct feature_map_t matchtable_ebx7[] = {
{ 2, CPU_FEATURE_SGX },
{ 4, CPU_FEATURE_HLE },
{ 11, CPU_FEATURE_RTM },
{ 16, CPU_FEATURE_AVX512F },
{ 17, CPU_FEATURE_AVX512DQ },
{ 18, CPU_FEATURE_RDSEED },
{ 19, CPU_FEATURE_ADX },
{ 26, CPU_FEATURE_AVX512PF },
{ 27, CPU_FEATURE_AVX512ER },
{ 28, CPU_FEATURE_AVX512CD },
{ 29, CPU_FEATURE_SHA_NI },
{ 30, CPU_FEATURE_AVX512BW },
{ 31, CPU_FEATURE_AVX512VL },
};
if (raw->basic_cpuid[0][0] >= 1) {
match_features(matchtable_edx1, COUNT_OF(matchtable_edx1), raw->basic_cpuid[1][3], data);
match_features(matchtable_ecx1, COUNT_OF(matchtable_ecx1), raw->basic_cpuid[1][2], data);
}
if (raw->ext_cpuid[0][0] >= 1) {
match_features(matchtable_edx81, COUNT_OF(matchtable_edx81), raw->ext_cpuid[1][3], data);
}
// detect TSX/AVX512:
if (raw->basic_cpuid[0][0] >= 7) {
match_features(matchtable_ebx7, COUNT_OF(matchtable_ebx7), raw->basic_cpuid[7][1], data);
}
}
enum _cache_type_t {
L1I,
L1D,
L2,
L3,
L4
};
typedef enum _cache_type_t cache_type_t;
static void check_case(uint8_t on, cache_type_t cache, int size, int assoc, int linesize, struct cpu_id_t* data)
{
if (!on) return;
switch (cache) {
case L1I:
data->l1_instruction_cache = size;
break;
case L1D:
data->l1_data_cache = size;
data->l1_assoc = assoc;
data->l1_cacheline = linesize;
break;
case L2:
data->l2_cache = size;
data->l2_assoc = assoc;
data->l2_cacheline = linesize;
break;
case L3:
data->l3_cache = size;
data->l3_assoc = assoc;
data->l3_cacheline = linesize;
break;
case L4:
data->l4_cache = size;
data->l4_assoc = assoc;
data->l4_cacheline = linesize;
break;
default:
break;
}
}
static void decode_intel_oldstyle_cache_info(struct cpu_raw_data_t* raw, struct cpu_id_t* data)
{
uint8_t f[256] = {0};
int reg, off;
uint32_t x;
for (reg = 0; reg < 4; reg++) {
x = raw->basic_cpuid[2][reg];
if (x & 0x80000000) continue;
for (off = 0; off < 4; off++) {
f[x & 0xff] = 1;
x >>= 8;
}
}
check_case(f[0x06], L1I, 8, 4, 32, data);
check_case(f[0x08], L1I, 16, 4, 32, data);
check_case(f[0x0A], L1D, 8, 2, 32, data);
check_case(f[0x0C], L1D, 16, 4, 32, data);
check_case(f[0x22], L3, 512, 4, 64, data);
check_case(f[0x23], L3, 1024, 8, 64, data);
check_case(f[0x25], L3, 2048, 8, 64, data);
check_case(f[0x29], L3, 4096, 8, 64, data);
check_case(f[0x2C], L1D, 32, 8, 64, data);
check_case(f[0x30], L1I, 32, 8, 64, data);
check_case(f[0x39], L2, 128, 4, 64, data);
check_case(f[0x3A], L2, 192, 6, 64, data);
check_case(f[0x3B], L2, 128, 2, 64, data);
check_case(f[0x3C], L2, 256, 4, 64, data);
check_case(f[0x3D], L2, 384, 6, 64, data);
check_case(f[0x3E], L2, 512, 4, 64, data);
check_case(f[0x41], L2, 128, 4, 32, data);
check_case(f[0x42], L2, 256, 4, 32, data);
check_case(f[0x43], L2, 512, 4, 32, data);
check_case(f[0x44], L2, 1024, 4, 32, data);
check_case(f[0x45], L2, 2048, 4, 32, data);
check_case(f[0x46], L3, 4096, 4, 64, data);
check_case(f[0x47], L3, 8192, 8, 64, data);
check_case(f[0x4A], L3, 6144, 12, 64, data);
check_case(f[0x4B], L3, 8192, 16, 64, data);
check_case(f[0x4C], L3, 12288, 12, 64, data);
check_case(f[0x4D], L3, 16384, 16, 64, data);
check_case(f[0x4E], L2, 6144, 24, 64, data);
check_case(f[0x60], L1D, 16, 8, 64, data);
check_case(f[0x66], L1D, 8, 4, 64, data);
check_case(f[0x67], L1D, 16, 4, 64, data);
check_case(f[0x68], L1D, 32, 4, 64, data);
/* The following four entries are trace cache. Intel does not
* specify a cache-line size, so we use -1 instead
*/
check_case(f[0x70], L1I, 12, 8, -1, data);
check_case(f[0x71], L1I, 16, 8, -1, data);
check_case(f[0x72], L1I, 32, 8, -1, data);
check_case(f[0x73], L1I, 64, 8, -1, data);
check_case(f[0x78], L2, 1024, 4, 64, data);
check_case(f[0x79], L2, 128, 8, 64, data);
check_case(f[0x7A], L2, 256, 8, 64, data);
check_case(f[0x7B], L2, 512, 8, 64, data);
check_case(f[0x7C], L2, 1024, 8, 64, data);
check_case(f[0x7D], L2, 2048, 8, 64, data);
check_case(f[0x7F], L2, 512, 2, 64, data);
check_case(f[0x82], L2, 256, 8, 32, data);
check_case(f[0x83], L2, 512, 8, 32, data);
check_case(f[0x84], L2, 1024, 8, 32, data);
check_case(f[0x85], L2, 2048, 8, 32, data);
check_case(f[0x86], L2, 512, 4, 64, data);
check_case(f[0x87], L2, 1024, 8, 64, data);
if (f[0x49]) {
/* This flag is overloaded with two meanings. On Xeon MP
* (family 0xf, model 0x6) this means L3 cache. On all other
* CPUs (notably Conroe et al), this is L2 cache. In both cases
* it means 4MB, 16-way associative, 64-byte line size.
*/
if (data->family == 0xf && data->model == 0x6) {
data->l3_cache = 4096;
data->l3_assoc = 16;
data->l3_cacheline = 64;
} else {
data->l2_cache = 4096;
data->l2_assoc = 16;
data->l2_cacheline = 64;
}
}
if (f[0x40]) {
/* Again, a special flag. It means:
* 1) If no L2 is specified, then CPU is w/o L2 (0 KB)
* 2) If L2 is specified by other flags, then, CPU is w/o L3.
*/
if (data->l2_cache == -1) {
data->l2_cache = 0;
} else {
data->l3_cache = 0;
}
}
}
static void decode_intel_deterministic_cache_info(struct cpu_raw_data_t* raw,
struct cpu_id_t* data)
{
int ecx;
int ways, partitions, linesize, sets, size, level, typenumber;
cache_type_t type;
for (ecx = 0; ecx < MAX_INTELFN4_LEVEL; ecx++) {
typenumber = raw->intel_fn4[ecx][0] & 0x1f;
if (typenumber == 0) break;
level = (raw->intel_fn4[ecx][0] >> 5) & 0x7;
if (level == 1 && typenumber == 1)
type = L1D;
else if (level == 1 && typenumber == 2)
type = L1I;
else if (level == 2 && typenumber == 3)
type = L2;
else if (level == 3 && typenumber == 3)
type = L3;
else if (level == 4 && typenumber == 3)
type = L4;
else {
continue;
}
ways = ((raw->intel_fn4[ecx][1] >> 22) & 0x3ff) + 1;
partitions = ((raw->intel_fn4[ecx][1] >> 12) & 0x3ff) + 1;
linesize = (raw->intel_fn4[ecx][1] & 0xfff) + 1;
sets = raw->intel_fn4[ecx][2] + 1;
size = ways * partitions * linesize * sets / 1024;
check_case(1, type, size, ways, linesize, data);
}
}
static int decode_intel_extended_topology(struct cpu_raw_data_t* raw,
struct cpu_id_t* data)
{
int i, level_type, num_smt = -1, num_core = -1;
for (i = 0; i < MAX_INTELFN11_LEVEL; i++) {
level_type = (raw->intel_fn11[i][2] & 0xff00) >> 8;
switch (level_type) {
case 0x01:
num_smt = raw->intel_fn11[i][1] & 0xffff;
break;
case 0x02:
num_core = raw->intel_fn11[i][1] & 0xffff;
break;
default:
break;
}
}
if (num_smt == -1 || num_core == -1) return 0;
data->num_logical_cpus = num_core;
data->num_cores = num_core / num_smt;
// make sure num_cores is at least 1. In VMs, the CPUID instruction
// is rigged and may give nonsensical results, but we should at least
// avoid outputs like data->num_cores == 0.
if (data->num_cores <= 0) data->num_cores = 1;
return 1;
}
static void decode_intel_number_of_cores(struct cpu_raw_data_t* raw,
struct cpu_id_t* data)
{
int logical_cpus = -1, num_cores = -1;
if (raw->basic_cpuid[0][0] >= 11) {
if (decode_intel_extended_topology(raw, data)) return;
}
if (raw->basic_cpuid[0][0] >= 1) {
logical_cpus = (raw->basic_cpuid[1][1] >> 16) & 0xff;
if (raw->basic_cpuid[0][0] >= 4) {
num_cores = 1 + ((raw->basic_cpuid[4][0] >> 26) & 0x3f);
}
}
if (data->flags[CPU_FEATURE_HT]) {
if (num_cores > 1) {
data->num_cores = num_cores;
data->num_logical_cpus = logical_cpus;
} else {
data->num_cores = 1;
data->num_logical_cpus = (logical_cpus >= 1 ? logical_cpus : 1);
if (data->num_logical_cpus == 1)
data->flags[CPU_FEATURE_HT] = 0;
}
} else {
data->num_cores = data->num_logical_cpus = 1;
}
}
static intel_code_and_bits_t get_brand_code_and_bits(struct cpu_id_t* data)
{
intel_code_t code = (intel_code_t) NC;
intel_code_and_bits_t result;
uint64_t bits = 0;
int i = 0;
const char* bs = data->brand_str;
const char* s;
const struct { intel_code_t c; const char *search; } matchtable[] = {
{ PENTIUM_M, "Pentium(R) M" },
{ CORE_SOLO, "Pentium(R) Dual CPU" },
{ CORE_SOLO, "Pentium(R) Dual-Core" },
{ PENTIUM_D, "Pentium(R) D" },
{ CORE_SOLO, "Genuine Intel(R) CPU" },
{ CORE_SOLO, "Intel(R) Core(TM)" },
{ DIAMONDVILLE, "CPU [N ][23]## " },
{ SILVERTHORNE, "CPU Z" },
{ PINEVIEW, "CPU [ND][45]## " },
{ CEDARVIEW, "CPU [ND]#### " },
};
const struct { uint64_t bit; const char* search; } bit_matchtable[] = {
{ XEON_, "Xeon" },
{ _MP_, " MP" },
{ ATOM_, "Atom(TM) CPU" },
{ MOBILE_, "Mobile" },
{ CELERON_, "Celeron" },
{ PENTIUM_, "Pentium" },
};
for (i = 0; i < COUNT_OF(bit_matchtable); i++) {
if (match_pattern(bs, bit_matchtable[i].search))
bits |= bit_matchtable[i].bit;
}
if ((i = match_pattern(bs, "Core(TM) [im][3579]")) != 0) {
bits |= CORE_;
i--;
switch (bs[i + 9]) {
case 'i': bits |= _I_; break;
case 'm': bits |= _M_; break;
}
switch (bs[i + 10]) {
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++)
if (match_pattern(bs, matchtable[i].search)) {
code = matchtable[i].c;
break;
}
if (bits & XEON_) {
if (match_pattern(bs, "W35##") || match_pattern(bs, "[ELXW]75##"))
bits |= _7;
else if (match_pattern(bs, "[ELXW]55##"))
code = GAINESTOWN;
else if (match_pattern(bs, "[ELXW]56##"))
code = WESTMERE;
else if (data->l3_cache > 0 && data->family == 16)
/* restrict by family, since later Xeons also have L3 ... */
code = IRWIN;
}
if (match_all(bits, XEON_ + _MP_) && data->l3_cache > 0)
code = POTOMAC;
if (code == CORE_SOLO) {
s = strstr(bs, "CPU");
if (s) {
s += 3;
while (*s == ' ') s++;
if (*s == 'T')
bits |= MOBILE_;
}
}
if (code == CORE_SOLO) {
switch (data->num_cores) {
case 1: break;
case 2:
{
code = CORE_DUO;
if (data->num_logical_cpus > 2)
code = DUAL_CORE_HT;
break;
}
case 4:
{
code = QUAD_CORE;
if (data->num_logical_cpus > 4)
code = QUAD_CORE_HT;
break;
}
default:
code = MORE_THAN_QUADCORE; break;
}
}
if (code == CORE_DUO && (bits & MOBILE_) && data->model != 14) {
if (data->ext_model < 23) {
code = MEROM;
} else {
code = PENRYN;
}
}
if (data->ext_model == 23 &&
(code == CORE_DUO || code == PENTIUM_D || (bits & CELERON_))) {
code = WOLFDALE;
}
result.code = code;
result.bits = bits;
return result;
}
static void decode_intel_sgx_features(const struct cpu_raw_data_t* raw, struct cpu_id_t* data)
{
struct cpu_epc_t epc;
int i;
if (raw->basic_cpuid[0][0] < 0x12) return; // no 12h leaf
if (raw->basic_cpuid[0x12][0] == 0) return; // no sub-leafs available, probably it's disabled by BIOS
// decode sub-leaf 0:
if (raw->basic_cpuid[0x12][0] & 1) data->sgx.flags[INTEL_SGX1] = 1;
if (raw->basic_cpuid[0x12][0] & 2) data->sgx.flags[INTEL_SGX2] = 1;
if (data->sgx.flags[INTEL_SGX1] || data->sgx.flags[INTEL_SGX2])
data->sgx.present = 1;
data->sgx.misc_select = raw->basic_cpuid[0x12][1];
data->sgx.max_enclave_32bit = (raw->basic_cpuid[0x12][3] ) & 0xff;
data->sgx.max_enclave_64bit = (raw->basic_cpuid[0x12][3] >> 8) & 0xff;
// decode sub-leaf 1:
data->sgx.secs_attributes = raw->intel_fn12h[1][0] | (((uint64_t) raw->intel_fn12h[1][1]) << 32);
data->sgx.secs_xfrm = raw->intel_fn12h[1][2] | (((uint64_t) raw->intel_fn12h[1][3]) << 32);
// decode higher-order subleafs, whenever present:
data->sgx.num_epc_sections = -1;
for (i = 0; i < 1000000; i++) {
epc = cpuid_get_epc(i, raw);
if (epc.length == 0) {
data->sgx.num_epc_sections = i;
break;
}
}
if (data->sgx.num_epc_sections == -1) {
data->sgx.num_epc_sections = 1000000;
}
}
struct cpu_epc_t cpuid_get_epc(int index, const struct cpu_raw_data_t* raw)
{
uint32_t regs[4];
struct cpu_epc_t retval = {0, 0};
if (raw && index < MAX_INTELFN12H_LEVEL - 2) {
// this was queried already, use the data:
memcpy(regs, raw->intel_fn12h[2 + index], sizeof(regs));
} else {
// query this ourselves:
regs[0] = 0x12;
regs[2] = 2 + index;
regs[1] = regs[3] = 0;
cpu_exec_cpuid_ext(regs);
}
// decode values:
if ((regs[0] & 0xf) == 0x1) {
retval.start_addr |= (regs[0] & 0xfffff000); // bits [12, 32) -> bits [12, 32)
retval.start_addr |= ((uint64_t) (regs[1] & 0x000fffff)) << 32; // bits [0, 20) -> bits [32, 52)
retval.length |= (regs[2] & 0xfffff000); // bits [12, 32) -> bits [12, 32)
retval.length |= ((uint64_t) (regs[3] & 0x000fffff)) << 32; // bits [0, 20) -> bits [32, 52)
}
return retval;
}
int cpuid_identify_intel(struct cpu_raw_data_t* raw, struct cpu_id_t* data, struct internal_id_info_t* internal)
{
intel_code_and_bits_t brand;
load_intel_features(raw, data);
if (raw->basic_cpuid[0][0] >= 4) {
/* Deterministic way is preferred, being more generic */
decode_intel_deterministic_cache_info(raw, data);
} else if (raw->basic_cpuid[0][0] >= 2) {
decode_intel_oldstyle_cache_info(raw, data);
}
decode_intel_number_of_cores(raw, data);
brand = get_brand_code_and_bits(data);
internal->code.intel = brand.code;
internal->bits = brand.bits;
if (data->flags[CPU_FEATURE_SGX]) {
// if SGX is indicated by the CPU, verify its presence:
decode_intel_sgx_features(raw, data);
}
return 0;
}

31
src/3rdparty/libcpuid/recog_intel.h

@ -1,31 +0,0 @@
/*
* 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 __RECOG_INTEL_H__
#define __RECOG_INTEL_H__
int cpuid_identify_intel(struct cpu_raw_data_t* raw, struct cpu_id_t* data, struct internal_id_info_t* internal);
#endif /*__RECOG_INTEL_H__*/

2
src/Summary.cpp

@ -95,7 +95,7 @@ static void print_cpu(Config *)
info->isX64() ? GREEN_BOLD_S : RED_BOLD_S "-",
info->hasAES() ? GREEN_BOLD_S : RED_BOLD_S "-"
);
# if defined(XMRIG_FEATURE_LIBCPUID) || defined (XMRIG_FEATURE_HWLOC)
# if defined(XMRIG_FEATURE_HWLOC)
Log::print(WHITE_BOLD(" %-13s") BLACK_BOLD("L2:") WHITE_BOLD("%.1f MB") BLACK_BOLD(" L3:") WHITE_BOLD("%.1f MB")
CYAN_BOLD(" %zu") "C" BLACK_BOLD("/") CYAN_BOLD("%zu") "T"
# ifdef XMRIG_FEATURE_HWLOC

14
src/backend/cpu/Cpu.cpp

@ -1,12 +1,6 @@
/* XMRig
* Copyright 2010 Jeff Garzik <jgarzik@pobox.com>
* Copyright 2012-2014 pooler <pooler@litecoinpool.org>
* Copyright 2014 Lucas Jones <https://github.com/lucasjones>
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2017-2018 XMR-Stak <https://github.com/fireice-uk>, <https://github.com/psychocrypt>
* Copyright 2018-2020 SChernykh <https://github.com/SChernykh>
* Copyright 2016-2020 XMRig <https://github.com/xmrig>, <support@xmrig.com>
* Copyright (c) 2018-2020 SChernykh <https://github.com/SChernykh>
* Copyright (c) 2016-2020 XMRig <https://github.com/xmrig>, <support@xmrig.com>
*
* 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
@ -32,8 +26,6 @@
#if defined(XMRIG_FEATURE_HWLOC)
# include "backend/cpu/platform/HwlocCpuInfo.h"
#elif defined(XMRIG_FEATURE_LIBCPUID)
# include "backend/cpu/platform/AdvancedCpuInfo.h"
#else
# include "backend/cpu/platform/BasicCpuInfo.h"
#endif
@ -47,8 +39,6 @@ xmrig::ICpuInfo *xmrig::Cpu::info()
if (cpuInfo == nullptr) {
# if defined(XMRIG_FEATURE_HWLOC)
cpuInfo = new HwlocCpuInfo();
# elif defined(XMRIG_FEATURE_LIBCPUID)
cpuInfo = new AdvancedCpuInfo();
# else
cpuInfo = new BasicCpuInfo();
# endif

10
src/backend/cpu/Cpu.h

@ -1,12 +1,6 @@
/* XMRig
* Copyright 2010 Jeff Garzik <jgarzik@pobox.com>
* Copyright 2012-2014 pooler <pooler@litecoinpool.org>
* Copyright 2014 Lucas Jones <https://github.com/lucasjones>
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2017-2018 XMR-Stak <https://github.com/fireice-uk>, <https://github.com/psychocrypt>
* Copyright 2018-2020 SChernykh <https://github.com/SChernykh>
* Copyright 2016-2020 XMRig <https://github.com/xmrig>, <support@xmrig.com>
* Copyright (c) 2018-2020 SChernykh <https://github.com/SChernykh>
* Copyright (c) 2016-2020 XMRig <https://github.com/xmrig>, <support@xmrig.com>
*
* 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

39
src/backend/cpu/cpu.cmake

@ -1,13 +1,14 @@
set(HEADERS_BACKEND_CPU
src/backend/cpu/Cpu.h
src/backend/cpu/CpuBackend.h
src/backend/cpu/CpuConfig.h
src/backend/cpu/CpuConfig_gen.h
src/backend/cpu/CpuConfig.h
src/backend/cpu/CpuLaunchData.cpp
src/backend/cpu/CpuThread.h
src/backend/cpu/CpuThreads.h
src/backend/cpu/CpuWorker.h
src/backend/cpu/interfaces/ICpuInfo.h
src/backend/cpu/platform/BasicCpuInfo.h
)
set(SOURCES_BACKEND_CPU
@ -20,7 +21,6 @@ set(SOURCES_BACKEND_CPU
src/backend/cpu/CpuWorker.cpp
)
if (WITH_HWLOC)
if (CMAKE_CXX_COMPILER_ID MATCHES MSVC)
add_subdirectory(src/3rdparty/hwloc)
@ -32,51 +32,26 @@ if (WITH_HWLOC)
set(CPUID_LIB ${HWLOC_LIBRARY})
endif()
set(WITH_LIBCPUID OFF)
remove_definitions(/DXMRIG_FEATURE_LIBCPUID)
add_definitions(/DXMRIG_FEATURE_HWLOC)
if (HWLOC_DEBUG)
add_definitions(/DXMRIG_HWLOC_DEBUG)
endif()
set(SOURCES_CPUID
src/backend/cpu/platform/BasicCpuInfo.h
src/backend/cpu/platform/HwlocCpuInfo.cpp
src/backend/cpu/platform/HwlocCpuInfo.h
)
elseif (WITH_LIBCPUID)
message(WARNING, "libcpuid support is deprecated and will be removed in future versions.")
set(WITH_HWLOC OFF)
add_subdirectory(src/3rdparty/libcpuid)
include_directories(src/3rdparty/libcpuid)
add_definitions(/DXMRIG_FEATURE_LIBCPUID)
remove_definitions(/DXMRIG_FEATURE_HWLOC)
set(CPUID_LIB cpuid)
set(SOURCES_CPUID
src/backend/cpu/platform/AdvancedCpuInfo.cpp
src/backend/cpu/platform/AdvancedCpuInfo.h
)
list(APPEND HEADERS_BACKEND_CPU src/backend/cpu/platform/HwlocCpuInfo.h)
list(APPEND SOURCES_BACKEND_CPU src/backend/cpu/platform/HwlocCpuInfo.cpp)
else()
remove_definitions(/DXMRIG_FEATURE_LIBCPUID)
remove_definitions(/DXMRIG_FEATURE_HWLOC)
set(CPUID_LIB "")
set(SOURCES_CPUID
src/backend/cpu/platform/BasicCpuInfo.h
)
endif()
if (XMRIG_ARM)
list(APPEND SOURCES_CPUID src/backend/cpu/platform/BasicCpuInfo_arm.cpp)
list(APPEND SOURCES_BACKEND_CPU src/backend/cpu/platform/BasicCpuInfo_arm.cpp)
if (XMRIG_OS_UNIX)
list(APPEND SOURCES_CPUID src/backend/cpu/platform/lscpu_arm.cpp)
list(APPEND SOURCES_BACKEND_CPU src/backend/cpu/platform/lscpu_arm.cpp)
endif()
else()
list(APPEND SOURCES_CPUID src/backend/cpu/platform/BasicCpuInfo.cpp)
list(APPEND SOURCES_BACKEND_CPU src/backend/cpu/platform/BasicCpuInfo.cpp)
endif()

130
src/backend/cpu/platform/AdvancedCpuInfo.cpp

@ -1,130 +0,0 @@
/* XMRig
* Copyright 2010 Jeff Garzik <jgarzik@pobox.com>
* Copyright 2012-2014 pooler <pooler@litecoinpool.org>
* Copyright 2014 Lucas Jones <https://github.com/lucasjones>
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2017-2019 XMR-Stak <https://github.com/fireice-uk>, <https://github.com/psychocrypt>
* Copyright 2018-2020 SChernykh <https://github.com/SChernykh>
* Copyright 2016-2020 XMRig <support@xmrig.com>
*
* 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 <http://www.gnu.org/licenses/>.
*/
#include "backend/cpu/platform/AdvancedCpuInfo.h"
#include "3rdparty/libcpuid/libcpuid.h"
#include <algorithm>
#include <cassert>
#include <cmath>
#include <cstdio>
#include <cstring>
xmrig::AdvancedCpuInfo::AdvancedCpuInfo()
{
struct cpu_raw_data_t raw = {};
struct cpu_id_t data = {};
cpuid_get_raw_data(&raw);
cpu_identify(&raw, &data);
snprintf(m_backend, sizeof m_backend, "libcpuid/%s", cpuid_lib_version());
m_threads = static_cast<size_t>(data.total_logical_cpus);
m_packages = std::max<size_t>(m_threads / static_cast<size_t>(data.num_logical_cpus), 1);
m_cores = static_cast<size_t>(data.num_cores) * m_packages;
m_L3 = data.l3_cache > 0 ? static_cast<size_t>(data.l3_cache) * m_packages : 0;
const auto l2 = static_cast<size_t>(data.l2_cache);
// Workaround for AMD CPUs https://github.com/anrieff/libcpuid/issues/97
if (m_vendor == VENDOR_AMD && data.ext_family >= 0x15 && data.ext_family < 0x17) {
m_L2 = l2 * (cores() / 2) * m_packages;
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 (m_vendor == VENDOR_INTEL && data.ext_family == 0x06 && (data.ext_model == 0x0E || data.ext_model == 0x0F || data.ext_model == 0x17)) {
size_t l2_count_per_socket = cores() > 1 ? cores() / 2 : 1;
m_L2 = data.l2_cache > 0 ? l2 * l2_count_per_socket * m_packages : 0;
}
else{
m_L2 = data.l2_cache > 0 ? l2 * cores() * m_packages : 0;
}
m_L2 *= 1024;
m_L3 *= 1024;
}
xmrig::CpuThreads xmrig::AdvancedCpuInfo::threads(const Algorithm &algorithm, uint32_t limit) const
{
if (threads() == 1) {
return 1;
}
size_t cache = 0;
size_t count = 0;
# ifdef XMRIG_ALGO_ASTROBWT
if (algorithm == Algorithm::ASTROBWT_DERO) {
CpuThreads t;
count = threads();
for (size_t i = 0; i < count; ++i) {
t.add(i, 0);
}
return t;
}
# endif
if (m_L3) {
cache = m_L2_exclusive ? (m_L2 + m_L3) : m_L3;
}
else {
cache = m_L2;
}
if (cache) {
const size_t memory = algorithm.l3();
assert(memory > 0);
count = cache / memory;
if (cache % memory >= memory / 2) {
count++;
}
}
else {
count = threads() / 2;
}
uint32_t intensity = algorithm.maxIntensity() == 1 ? 0 : 1;
# ifdef XMRIG_ALGO_CN_PICO
if (algorithm == Algorithm::CN_PICO_0 && (count / cores()) >= 2) {
intensity = 2;
}
# endif
if (limit > 0 && limit < 100) {
count = std::min(count, static_cast<size_t>(round(threads() * (limit / 100.0))));
}
else {
count = std::min(count, threads());
}
return CpuThreads(std::max<size_t>(count, 1), intensity);
}

63
src/backend/cpu/platform/AdvancedCpuInfo.h

@ -1,63 +0,0 @@
/* XMRig
* Copyright 2010 Jeff Garzik <jgarzik@pobox.com>
* Copyright 2012-2014 pooler <pooler@litecoinpool.org>
* Copyright 2014 Lucas Jones <https://github.com/lucasjones>
* Copyright 2014-2016 Wolf9466 <https://github.com/OhGodAPet>
* Copyright 2016 Jay D Dee <jayddee246@gmail.com>
* Copyright 2017-2019 XMR-Stak <https://github.com/fireice-uk>, <https://github.com/psychocrypt>
* Copyright 2018-2020 SChernykh <https://github.com/SChernykh>
* Copyright 2016-2020 XMRig <support@xmrig.com>
*
* 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 <http://www.gnu.org/licenses/>.
*/
#ifndef XMRIG_ADVANCEDCPUINFO_H
#define XMRIG_ADVANCEDCPUINFO_H
#include "backend/cpu/platform/BasicCpuInfo.h"
namespace xmrig {
class AdvancedCpuInfo : public BasicCpuInfo
{
public:
AdvancedCpuInfo();
protected:
CpuThreads threads(const Algorithm &algorithm, uint32_t limit) const override;
inline const char *backend() const override { return m_backend; }
inline size_t cores() const override { return m_cores; }
inline size_t L2() const override { return m_L2; }
inline size_t L3() const override { return m_L3; }
inline size_t packages() const override { return m_packages; }
inline size_t threads() const override { return m_threads; }
private:
bool m_L2_exclusive = false;
char m_backend[32]{};
size_t m_cores = 0;
size_t m_L2 = 0;
size_t m_L3 = 0;
size_t m_packages = 1;
};
} /* namespace xmrig */
#endif /* XMRIG_ADVANCEDCPUINFO_H */

3
src/base/api/Api.cpp

@ -169,9 +169,6 @@ void xmrig::Api::exec(IApiRequest &request)
# ifdef XMRIG_FEATURE_HTTP
features.PushBack("http", allocator);
# endif
# ifdef XMRIG_FEATURE_LIBCPUID
features.PushBack("cpuid", allocator);
# endif
# ifdef XMRIG_FEATURE_HWLOC
features.PushBack("hwloc", allocator);
# endif

Loading…
Cancel
Save