Merge f0f6d1666c into d5a8f8a5ae

KawPow: fixed data race when building programs

CHANGELOG.md

@@ -1,3 +1,63 @@
+# v6.19.3
+- [#3245](https://github.com/xmrig/xmrig/issues/3245) Improved algorithm negotiation for donation rounds by sending extra information about current mining job.
+- [#3254](https://github.com/xmrig/xmrig/pull/3254) Tweaked auto-tuning for Intel CPUs.
+- [#3271](https://github.com/xmrig/xmrig/pull/3271) RandomX: optimized program generation.
+- [#3273](https://github.com/xmrig/xmrig/pull/3273) RandomX: fixed undefined behavior.
+- [#3275](https://github.com/xmrig/xmrig/pull/3275) RandomX: fixed `jccErratum` list.
+- [#3280](https://github.com/xmrig/xmrig/pull/3280) Updated example scripts.
+
+# v6.19.2
+- [#3230](https://github.com/xmrig/xmrig/pull/3230) Fixed parsing of `TX_EXTRA_MERGE_MINING_TAG`.
+- [#3232](https://github.com/xmrig/xmrig/pull/3232) Added new `X-Hash-Difficulty` HTTP header.
+- [#3240](https://github.com/xmrig/xmrig/pull/3240) Improved .cmd files when run by shortcuts on another drive.
+- [#3241](https://github.com/xmrig/xmrig/pull/3241) Added view tag calculation (fixes Wownero solo mining issue).
+
+# v6.19.1
+- Resolved deprecated methods warnings with OpenSSL 3.0.
+- [#3213](https://github.com/xmrig/xmrig/pull/3213) Fixed build with 32-bit clang 15.
+- [#3218](https://github.com/xmrig/xmrig/pull/3218) Fixed: `--randomx-wrmsr=-1` worked only on Intel.
+- [#3228](https://github.com/xmrig/xmrig/pull/3228) Fixed build with gcc 13.
+
+# v6.19.0
+- [#3144](https://github.com/xmrig/xmrig/pull/3144) Update to latest `sse2neon.h`.
+- [#3161](https://github.com/xmrig/xmrig/pull/3161) MSVC build: enabled parallel compilation.
+- [#3163](https://github.com/xmrig/xmrig/pull/3163) Improved Zen 3 MSR mod.
+- [#3176](https://github.com/xmrig/xmrig/pull/3176) Update cmake required version to 3.1.
+- [#3182](https://github.com/xmrig/xmrig/pull/3182) DragonflyBSD compilation fixes.
+- [#3196](https://github.com/xmrig/xmrig/pull/3196) Show IP address for failed connections.
+- [#3185](https://github.com/xmrig/xmrig/issues/3185) Fixed macOS DMI reader.
+- [#3198](https://github.com/xmrig/xmrig/pull/3198) Fixed broken RandomX light mode mining.
+- [#3202](https://github.com/xmrig/xmrig/pull/3202) Solo mining: added job timeout (default is 15 seconds).
+
+# v6.18.1
+- [#3129](https://github.com/xmrig/xmrig/pull/3129) Fix: protectRX flushed CPU cache only on MacOS/iOS.
+- [#3126](https://github.com/xmrig/xmrig/pull/3126) Don't reset when pool sends the same job blob.
+- [#3120](https://github.com/xmrig/xmrig/pull/3120) RandomX: optimized `CFROUND` elimination.
+- [#3109](https://github.com/xmrig/xmrig/pull/3109) RandomX: added Blake2 AVX2 version.
+- [#3082](https://github.com/xmrig/xmrig/pull/3082) Fixed GCC 12 warnings.
+- [#3075](https://github.com/xmrig/xmrig/pull/3075) Recognize `armv7ve` as valid ARMv7 target.
+- [#3132](https://github.com/xmrig/xmrig/pull/3132) RandomX: added MSR mod for Zen 4.
+- [#3134](https://github.com/xmrig/xmrig/pull/3134) Added Zen4 to `randomx_boost.sh`.
+
+# v6.18.0
+- [#3067](https://github.com/xmrig/xmrig/pull/3067) Monero v15 network upgrade support and more house keeping.
+  - Removed deprecated AstroBWTv1 and v2.
+  - Fixed debug GhostRider build.
+  - Monero v15 network upgrade support.
+  - Fixed ZMQ debug log.
+  - Improved daemon ZMQ mining stability.
+- [#3054](https://github.com/xmrig/xmrig/pull/3054) Fixes for 32-bit ARM.
+- [#3042](https://github.com/xmrig/xmrig/pull/3042) Fixed being unable to resume from `pause-on-battery`.
+- [#3031](https://github.com/xmrig/xmrig/pull/3031) Fixed `--cpu-priority` not working sometimes.
+- [#3020](https://github.com/xmrig/xmrig/pull/3020) Removed old AstroBWT algorithm.
+
+# v6.17.0
+- [#2954](https://github.com/xmrig/xmrig/pull/2954) **Dero HE fork support (`astrobwt/v2` algorithm).**
+  - [#2961](https://github.com/xmrig/xmrig/pull/2961) Dero HE (`astrobwt/v2`) CUDA config generator.
+  - [#2969](https://github.com/xmrig/xmrig/pull/2969) Dero HE (`astrobwt/v2`) OpenCL support.
+- Fixed displayed DMI memory information for empty slots.
+- [#2932](https://github.com/xmrig/xmrig/pull/2932) Fixed GhostRider with hwloc disabled.
+
 # v6.16.4
 - [#2904](https://github.com/xmrig/xmrig/pull/2904) Fixed unaligned memory accesses.
 - [#2908](https://github.com/xmrig/xmrig/pull/2908) Added MSVC/2022 to `version.h`.

CMakeLists.txt

@@ -1,4 +1,4 @@
-cmake_minimum_required(VERSION 2.8.12)
+cmake_minimum_required(VERSION 3.1)
 project(xmrig)
 
 option(WITH_HWLOC           "Enable hwloc support" ON)
@@ -8,7 +8,6 @@ option(WITH_CN_PICO         "Enable CryptoNight-Pico algorithm" ON)
 option(WITH_CN_FEMTO        "Enable CryptoNight-UPX2 algorithm" ON)
 option(WITH_RANDOMX         "Enable RandomX algorithms family" ON)
 option(WITH_ARGON2          "Enable Argon2 algorithms family" ON)
-option(WITH_ASTROBWT        "Enable AstroBWT algorithms family" ON)
 option(WITH_KAWPOW          "Enable KawPow algorithms family" ON)
 option(WITH_GHOSTRIDER      "Enable GhostRider algorithm" ON)
 option(WITH_HTTP            "Enable HTTP protocol support (client/server)" ON)
@@ -28,13 +27,16 @@ option(WITH_STRICT_CACHE    "Enable strict checks for OpenCL cache" ON)
 option(WITH_INTERLEAVE_DEBUG_LOG "Enable debug log for threads interleave" OFF)
 option(WITH_PROFILING       "Enable profiling for developers" OFF)
 option(WITH_SSE4_1          "Enable SSE 4.1 for Blake2" ON)
+option(WITH_AVX2            "Enable AVX2 for Blake2" ON)
 option(WITH_VAES            "Enable VAES instructions for Cryptonight" ON)
 option(WITH_BENCHMARK       "Enable builtin RandomX benchmark and stress test" ON)
 option(WITH_SECURE_JIT      "Enable secure access to JIT memory" OFF)
 option(WITH_DMI             "Enable DMI/SMBIOS reader" ON)
+option(WITH_BUNDLED_FMT     "Force use of bundled fmt library" ON)
 
 option(BUILD_STATIC         "Build static binary" OFF)
-option(ARM_TARGET           "Force use specific ARM target 8 or 7" 0)
+option(ARM_V8               "Force ARMv8 (64 bit) architecture, use with caution if automatic detection fails, but you sure it may work" OFF)
+option(ARM_V7               "Force ARMv7 (32 bit) architecture, use with caution if automatic detection fails, but you sure it may work" OFF)
 option(HWLOC_DEBUG          "Enable hwloc debug helpers and log" OFF)
 
 
@@ -197,9 +199,9 @@ add_definitions(-D__STDC_FORMAT_MACROS -DUNICODE -D_FILE_OFFSET_BITS=64)
 find_package(UV REQUIRED)
 
 include(cmake/flags.cmake)
+include(cmake/fmt.cmake)
 include(cmake/randomx.cmake)
 include(cmake/argon2.cmake)
-include(cmake/astrobwt.cmake)
 include(cmake/kawpow.cmake)
 include(cmake/ghostrider.cmake)
 include(cmake/OpenSSL.cmake)
@@ -229,7 +231,7 @@ include(src/hw/api/api.cmake)
 include(src/hw/dmi/dmi.cmake)
 
 include_directories(src)
-include_directories(src/3rdparty)
+include_directories(src/3rdparty/CL)
 include_directories(${UV_INCLUDE_DIR})
 
 if (WITH_DEBUG_LOG)
@@ -237,7 +239,7 @@ if (WITH_DEBUG_LOG)
 endif()
 
 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} ${GHOSTRIDER_LIBRARY})
+target_link_libraries(${CMAKE_PROJECT_NAME} ${XMRIG_ASM_LIBRARY} ${OPENSSL_LIBRARIES} ${UV_LIBRARIES} ${EXTRA_LIBS} ${CPUID_LIB} ${ARGON2_LIBRARY} ${ETHASH_LIBRARY} ${GHOSTRIDER_LIBRARY} ${FMT_LIBRARY})
 
 if (WIN32)
     add_custom_command(TARGET ${CMAKE_PROJECT_NAME} POST_BUILD COMMAND ${CMAKE_COMMAND} -E copy_if_different "${CMAKE_SOURCE_DIR}/bin/WinRing0/WinRing0x64.sys" $<TARGET_FILE_DIR:${CMAKE_PROJECT_NAME}>)

README.md

@@ -7,10 +7,10 @@
 [![GitHub stars](https://img.shields.io/github/stars/xmrig/xmrig.svg)](https://github.com/xmrig/xmrig/stargazers)
 [![GitHub forks](https://img.shields.io/github/forks/xmrig/xmrig.svg)](https://github.com/xmrig/xmrig/network)
 
-XMRig is a high performance, open source, cross platform RandomX, KawPow, CryptoNight, AstroBWT and [GhostRider](https://github.com/xmrig/xmrig/tree/master/src/crypto/ghostrider#readme) unified CPU/GPU miner and [RandomX benchmark](https://xmrig.com/benchmark). Official binaries are available for Windows, Linux, macOS and FreeBSD.
+XMRig is a high performance, open source, cross platform RandomX, KawPow, CryptoNight and [GhostRider](https://github.com/xmrig/xmrig/tree/master/src/crypto/ghostrider#readme) unified CPU/GPU miner and [RandomX benchmark](https://xmrig.com/benchmark). Official binaries are available for Windows, Linux, macOS and FreeBSD.
 
 ## Mining backends
-- **CPU** (x64/ARMv8)
+- **CPU** (x86/x64/ARMv7/ARMv8)
 - **OpenCL** for AMD GPUs.
 - **CUDA** for NVIDIA GPUs via external [CUDA plugin](https://github.com/xmrig/xmrig-cuda).
 

cmake/astrobwt.cmake

@@ -1,51 +0,0 @@
-if (WITH_ASTROBWT)
-    add_definitions(/DXMRIG_ALGO_ASTROBWT)
-
-    list(APPEND HEADERS_CRYPTO
-        src/crypto/astrobwt/AstroBWT.h
-    )
-
-    list(APPEND SOURCES_CRYPTO
-        src/crypto/astrobwt/AstroBWT.cpp
-    )
-
-    if (XMRIG_ARM)
-        list(APPEND HEADERS_CRYPTO
-            src/crypto/astrobwt/salsa20_ref/ecrypt-config.h
-            src/crypto/astrobwt/salsa20_ref/ecrypt-machine.h
-            src/crypto/astrobwt/salsa20_ref/ecrypt-portable.h
-            src/crypto/astrobwt/salsa20_ref/ecrypt-sync.h
-        )
-
-        list(APPEND SOURCES_CRYPTO
-            src/crypto/astrobwt/salsa20_ref/salsa20.c
-        )
-    else()
-        if (CMAKE_SIZEOF_VOID_P EQUAL 8)
-            add_definitions(/DASTROBWT_AVX2)
-            list(APPEND SOURCES_CRYPTO src/crypto/astrobwt/xmm6int/salsa20_xmm6int-avx2.c)
-
-            if (CMAKE_C_COMPILER_ID MATCHES GNU OR CMAKE_C_COMPILER_ID MATCHES Clang)
-                set_source_files_properties(src/crypto/astrobwt/xmm6int/salsa20_xmm6int-avx2.c PROPERTIES COMPILE_FLAGS -mavx2)
-            endif()
-
-            if (CMAKE_C_COMPILER_ID MATCHES MSVC)
-                enable_language(ASM_MASM)
-                list(APPEND SOURCES_CRYPTO src/crypto/astrobwt/sha3_256_avx2.asm)
-            else()
-                enable_language(ASM)
-                list(APPEND SOURCES_CRYPTO src/crypto/astrobwt/sha3_256_avx2.S)
-            endif()
-        endif()
-
-        list(APPEND HEADERS_CRYPTO
-            src/crypto/astrobwt/Salsa20.hpp
-        )
-
-        list(APPEND SOURCES_CRYPTO
-            src/crypto/astrobwt/Salsa20.cpp
-        )
-    endif()
-else()
-    remove_definitions(/DXMRIG_ALGO_ASTROBWT)
-endif()

cmake/cpu.cmake

@@ -25,13 +25,20 @@ if (XMRIG_64_BIT AND CMAKE_SYSTEM_PROCESSOR MATCHES "^(x86_64|AMD64)$")
     add_definitions(-DRAPIDJSON_SSE2)
 else()
     set(WITH_SSE4_1 OFF)
+    set(WITH_AVX2 OFF)
     set(WITH_VAES OFF)
 endif()
 
+if (ARM_V8)
+    set(ARM_TARGET 8)
+elseif (ARM_V7)
+    set(ARM_TARGET 7)
+endif()
+
 if (NOT ARM_TARGET)
     if (CMAKE_SYSTEM_PROCESSOR MATCHES "^(aarch64|arm64|armv8-a)$")
         set(ARM_TARGET 8)
-    elseif (CMAKE_SYSTEM_PROCESSOR MATCHES "^(armv7|armv7f|armv7s|armv7k|armv7-a|armv7l)$")
+    elseif (CMAKE_SYSTEM_PROCESSOR MATCHES "^(armv7|armv7f|armv7s|armv7k|armv7-a|armv7l|armv7ve)$")
         set(ARM_TARGET 7)
     endif()
 endif()
@@ -57,3 +64,7 @@ endif()
 if (WITH_SSE4_1)
     add_definitions(-DXMRIG_FEATURE_SSE4_1)
 endif()
+
+if (WITH_AVX2)
+    add_definitions(-DXMRIG_FEATURE_AVX2)
+endif()

cmake/flags.cmake

@@ -10,7 +10,7 @@ if ("${CMAKE_BUILD_TYPE}" STREQUAL "")
 endif()
 
 if (CMAKE_BUILD_TYPE STREQUAL "Release")
-    add_definitions(/DNDEBUG)
+    add_definitions(-DNDEBUG)
 endif()
 
 include(CheckSymbolExists)
@@ -32,7 +32,7 @@ if (CMAKE_CXX_COMPILER_ID MATCHES GNU)
         set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -maes")
         set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -maes")
 
-        add_definitions(/DHAVE_ROTR)
+        add_definitions(-DHAVE_ROTR)
     endif()
 
     if (WIN32)
@@ -49,28 +49,16 @@ if (CMAKE_CXX_COMPILER_ID MATCHES GNU)
         set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -static")
     endif()
 
-    add_definitions(/D_GNU_SOURCE)
-
-    if (${CMAKE_VERSION} VERSION_LESS "3.1.0")
-        set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -std=c99")
-        set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11")
-    endif()
-
-    #set(CMAKE_C_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -gdwarf-2")
-
-    add_definitions(/DHAVE_BUILTIN_CLEAR_CACHE)
+    add_definitions(-D_GNU_SOURCE -DHAVE_BUILTIN_CLEAR_CACHE)
 
 elseif (CMAKE_CXX_COMPILER_ID MATCHES MSVC)
-    set(CMAKE_C_FLAGS_RELEASE "/MT /O2 /Oi /DNDEBUG /GL")
-    set(CMAKE_CXX_FLAGS_RELEASE "/MT /O2 /Oi /DNDEBUG /GL")
+    set(CMAKE_C_FLAGS_RELEASE "/MP /MT /O2 /Oi /DNDEBUG /GL")
+    set(CMAKE_CXX_FLAGS_RELEASE "/MP /MT /O2 /Oi /DNDEBUG /GL")
 
-    set(CMAKE_C_FLAGS_RELWITHDEBINFO "/Ob1 /Zi /DRELWITHDEBINFO")
-    set(CMAKE_CXX_FLAGS_RELWITHDEBINFO "/Ob1 /Zi /DRELWITHDEBINFO")
+    set(CMAKE_C_FLAGS_RELWITHDEBINFO "/MP /Ob1 /Zi /DRELWITHDEBINFO")
+    set(CMAKE_CXX_FLAGS_RELWITHDEBINFO "/MP /Ob1 /Zi /DRELWITHDEBINFO")
 
-    add_definitions(/D_CRT_SECURE_NO_WARNINGS)
-    add_definitions(/D_CRT_NONSTDC_NO_WARNINGS)
-    add_definitions(/DNOMINMAX)
-    add_definitions(/DHAVE_ROTR)
+    add_definitions(-D_CRT_SECURE_NO_WARNINGS -D_CRT_NONSTDC_NO_WARNINGS -DNOMINMAX -DHAVE_ROTR)
 
 elseif (CMAKE_CXX_COMPILER_ID MATCHES Clang)
 
@@ -92,7 +80,7 @@ elseif (CMAKE_CXX_COMPILER_ID MATCHES Clang)
 
         check_symbol_exists("_rotr" "x86intrin.h" HAVE_ROTR)
         if (HAVE_ROTR)
-            add_definitions(/DHAVE_ROTR)
+            add_definitions(-DHAVE_ROTR)
         endif()
     endif()
 
@@ -105,6 +93,6 @@ endif()
 if (NOT WIN32)
     check_symbol_exists("__builtin___clear_cache" "stdlib.h" HAVE_BUILTIN_CLEAR_CACHE)
     if (HAVE_BUILTIN_CLEAR_CACHE)
-        add_definitions(/DHAVE_BUILTIN_CLEAR_CACHE)
+        add_definitions(-DHAVE_BUILTIN_CLEAR_CACHE)
     endif()
 endif()

cmake/fmt.cmake

@@ -0,0 +1,13 @@
+# SPDX-FileCopyrightText: © 2023 Jean-Pierre De Jesus DIAZ <me@jeandudey.tech>
+# SPDX-License-Identifier: GPL-3.0-or-later
+
+
+if(WITH_BUNDLED_FMT)
+    add_library(fmt INTERFACE)
+    target_sources(fmt INTERFACE ${CMAKE_CURRENT_SOURCE_DIR}/src/3rdparty/fmt/format.cc)
+    target_include_directories(fmt INTERFACE ${CMAKE_CURRENT_SOURCE_DIR}/src/3rdparty/fmt)
+else()
+    set(FMT_LIBRARY fmt)
+    find_package(fmt REQUIRED)
+    set(FMT_LIBRARY fmt::fmt)
+endif()

cmake/os.cmake

@@ -15,7 +15,7 @@ else()
         set(XMRIG_OS_ANDROID ON)
     elseif(CMAKE_SYSTEM_NAME MATCHES "Linux")
         set(XMRIG_OS_LINUX ON)
-    elseif(CMAKE_SYSTEM_NAME STREQUAL FreeBSD)
+    elseif(CMAKE_SYSTEM_NAME STREQUAL FreeBSD OR CMAKE_SYSTEM_NAME STREQUAL DragonFly)
         set(XMRIG_OS_FREEBSD ON)
     endif()
 endif()

cmake/randomx.cmake

@@ -76,7 +76,15 @@ if (WITH_RANDOMX)
         list(APPEND SOURCES_CRYPTO src/crypto/randomx/blake2/blake2b_sse41.c)
 
         if (CMAKE_C_COMPILER_ID MATCHES GNU OR CMAKE_C_COMPILER_ID MATCHES Clang)
-            set_source_files_properties(src/crypto/randomx/blake2/blake2b_sse41.c PROPERTIES COMPILE_FLAGS -msse4.1)
+            set_source_files_properties(src/crypto/randomx/blake2/blake2b_sse41.c PROPERTIES COMPILE_FLAGS "-Ofast -msse4.1")
+        endif()
+    endif()
+
+    if (WITH_AVX2)
+        list(APPEND SOURCES_CRYPTO src/crypto/randomx/blake2/avx2/blake2b_avx2.c)
+
+        if (CMAKE_C_COMPILER_ID MATCHES GNU OR CMAKE_C_COMPILER_ID MATCHES Clang)
+            set_source_files_properties(src/crypto/randomx/blake2/avx2/blake2b_avx2.c PROPERTIES COMPILE_FLAGS "-Ofast -mavx2")
         endif()
     endif()
 

scripts/benchmark_10M.cmd

@@ -1,4 +1,4 @@
 @echo off
-cd %~dp0
+cd /d "%~dp0"
 xmrig.exe --bench=10M --submit
 pause

scripts/benchmark_1M.cmd

@@ -1,4 +1,4 @@
 @echo off
-cd %~dp0
+cd /d "%~dp0"
 xmrig.exe --bench=1M --submit
 pause

scripts/build.hwloc.sh

@@ -1,6 +1,10 @@
 #!/bin/bash -e
 
-HWLOC_VERSION="2.5.0"
+HWLOC_VERSION_MAJOR="2"
+HWLOC_VERSION_MINOR="9"
+HWLOC_VERSION_PATCH="0"
+
+HWLOC_VERSION="${HWLOC_VERSION_MAJOR}.${HWLOC_VERSION_MINOR}.${HWLOC_VERSION_PATCH}"
 
 mkdir -p deps
 mkdir -p deps/include
@@ -8,7 +12,7 @@ mkdir -p deps/lib
 
 mkdir -p build && cd build
 
-wget https://download.open-mpi.org/release/hwloc/v2.5/hwloc-${HWLOC_VERSION}.tar.gz -O hwloc-${HWLOC_VERSION}.tar.gz
+wget https://download.open-mpi.org/release/hwloc/v${HWLOC_VERSION_MAJOR}.${HWLOC_VERSION_MINOR}/hwloc-${HWLOC_VERSION}.tar.gz -O hwloc-${HWLOC_VERSION}.tar.gz
 tar -xzf hwloc-${HWLOC_VERSION}.tar.gz
 
 cd hwloc-${HWLOC_VERSION}
@@ -16,4 +20,4 @@ cd hwloc-${HWLOC_VERSION}
 make -j$(nproc || sysctl -n hw.ncpu || sysctl -n hw.logicalcpu)
 cp -fr include ../../deps
 cp hwloc/.libs/libhwloc.a ../../deps/lib
-cd ..
+cd ..

scripts/build.libressl.sh

@@ -1,6 +1,6 @@
 #!/bin/bash -e
 
-LIBRESSL_VERSION="3.0.2"
+LIBRESSL_VERSION="3.5.2"
 
 mkdir -p deps
 mkdir -p deps/include
@@ -17,4 +17,4 @@ make -j$(nproc || sysctl -n hw.ncpu || sysctl -n hw.logicalcpu)
 cp -fr include ../../deps
 cp crypto/.libs/libcrypto.a ../../deps/lib
 cp ssl/.libs/libssl.a ../../deps/lib
-cd ..
+cd ..

scripts/build.openssl.sh

@@ -1,6 +1,6 @@
 #!/bin/bash -e
 
-OPENSSL_VERSION="1.1.1l"
+OPENSSL_VERSION="1.1.1s"
 
 mkdir -p deps
 mkdir -p deps/include

scripts/build.openssl3.sh

@@ -0,0 +1,20 @@
+#!/bin/bash -e
+
+OPENSSL_VERSION="3.0.7"
+
+mkdir -p deps
+mkdir -p deps/include
+mkdir -p deps/lib
+
+mkdir -p build && cd build
+
+wget https://www.openssl.org/source/openssl-${OPENSSL_VERSION}.tar.gz -O openssl-${OPENSSL_VERSION}.tar.gz
+tar -xzf openssl-${OPENSSL_VERSION}.tar.gz
+
+cd openssl-${OPENSSL_VERSION}
+./config -no-shared -no-asm -no-zlib -no-comp -no-dgram -no-filenames -no-cms
+make -j$(nproc || sysctl -n hw.ncpu || sysctl -n hw.logicalcpu)
+cp -fr include ../../deps
+cp libcrypto.a ../../deps/lib
+cp libssl.a ../../deps/lib
+cd ..

scripts/build.uv.sh

@@ -1,6 +1,6 @@
 #!/bin/bash -e
 
-UV_VERSION="1.42.0"
+UV_VERSION="1.44.2"
 
 mkdir -p deps
 mkdir -p deps/include
@@ -17,4 +17,4 @@ sh autogen.sh
 make -j$(nproc || sysctl -n hw.ncpu || sysctl -n hw.logicalcpu)
 cp -fr include ../../deps
 cp .libs/libuv.a ../../deps/lib
-cd ..
+cd ..

scripts/generate_cl.js

@@ -67,15 +67,6 @@ function rx()
 }
 
 
-function astrobwt()
-{
-    const astrobwt = opencl_minify(addIncludes('astrobwt.cl', [ 'BWT.cl', 'salsa20.cl', 'sha3.cl' ]));
-
-    // fs.writeFileSync('astrobwt_gen.cl', astrobwt);
-    fs.writeFileSync('astrobwt_cl.h', text2h(astrobwt, 'xmrig', 'astrobwt_cl'));
-}
-
-
 function kawpow()
 {
     const kawpow = opencl_minify(addIncludes('kawpow.cl', [ 'defs.h' ]));
@@ -97,11 +88,6 @@ process.chdir(path.resolve('src/backend/opencl/cl/rx'));
 
 rx();
 
-process.chdir(cwd);
-process.chdir(path.resolve('src/backend/opencl/cl/astrobwt'));
-
-astrobwt();
-
 process.chdir(cwd);
 process.chdir(path.resolve('src/backend/opencl/cl/kawpow'));
 

scripts/pool_mine_example.cmd

@@ -15,6 +15,6 @@
 :: Choose pools outside of top 5 to help Monero network be more decentralized!
 :: Smaller pools also often have smaller fees/payout limits.
 
-cd %~dp0
-xmrig.exe -o pool.hashvault.pro:3333 -u 48edfHu7V9Z84YzzMa6fUueoELZ9ZRXq9VetWzYGzKt52XU5xvqgzYnDK9URnRoJMk1j8nLwEVsaSWJ4fhdUyZijBGUicoD -p x
+cd /d "%~dp0"
+xmrig.exe -o xmrpool.eu:3333 -u 48edfHu7V9Z84YzzMa6fUueoELZ9ZRXq9VetWzYGzKt52XU5xvqgzYnDK9URnRoJMk1j8nLwEVsaSWJ4fhdUyZijBGUicoD -p x
 pause

scripts/randomx_boost.sh

@@ -10,14 +10,24 @@ fi
 
 if grep -E 'AMD Ryzen|AMD EPYC' /proc/cpuinfo > /dev/null;
 	then
-	if grep "cpu family[[:space:]]:[[:space:]]25" /proc/cpuinfo > /dev/null;
+	if grep "cpu family[[:space:]]\{1,\}:[[:space:]]25" /proc/cpuinfo > /dev/null;
 		then
-			echo "Detected Zen3 CPU"
-			wrmsr -a 0xc0011020 0x4480000000000
-			wrmsr -a 0xc0011021 0x1c000200000040
-			wrmsr -a 0xc0011022 0xc000000401500000
-			wrmsr -a 0xc001102b 0x2000cc14
-			echo "MSR register values for Zen3 applied"
+			if grep "model[[:space:]]\{1,\}:[[:space:]]97" /proc/cpuinfo > /dev/null;
+				then
+					echo "Detected Zen4 CPU"
+					wrmsr -a 0xc0011020 0x4400000000000
+					wrmsr -a 0xc0011021 0x4000000000040
+					wrmsr -a 0xc0011022 0x8680000401570000
+					wrmsr -a 0xc001102b 0x2040cc10
+					echo "MSR register values for Zen4 applied"
+				else
+					echo "Detected Zen3 CPU"
+					wrmsr -a 0xc0011020 0x4480000000000
+					wrmsr -a 0xc0011021 0x1c000200000040
+					wrmsr -a 0xc0011022 0xc000000401570000
+					wrmsr -a 0xc001102b 0x2000cc10
+					echo "MSR register values for Zen3 applied"
+				fi
 		else
 			echo "Detected Zen1/Zen2 CPU"
 			wrmsr -a 0xc0011020 0

scripts/rtm_ghostrider_example.cmd

@@ -15,7 +15,7 @@
 :: Choose pools outside of top 5 to help Raptoreum network be more decentralized!
 :: Smaller pools also often have smaller fees/payout limits.
 
-cd %~dp0
+cd /d "%~dp0"
 :: Use this command line to connect to non-SSL port
 xmrig.exe -a gr -o raptoreumemporium.com:3008 -u WALLET_ADDRESS -p x
 :: Or use this command line to connect to an SSL port

scripts/solo_mine_example.cmd

@@ -11,6 +11,6 @@
 :: Mining solo is the best way to help Monero network be more decentralized!
 :: But you will only get a payout when you find a block which can take more than a year for a single low-end PC.
 
-cd %~dp0
-xmrig.exe -o node.xmr.to:18081 -a rx/0 -u 48edfHu7V9Z84YzzMa6fUueoELZ9ZRXq9VetWzYGzKt52XU5xvqgzYnDK9URnRoJMk1j8nLwEVsaSWJ4fhdUyZijBGUicoD --daemon
+cd /d "%~dp0"
+xmrig.exe -o YOUR_NODE_IP:18081 -a rx/0 -u 48edfHu7V9Z84YzzMa6fUueoELZ9ZRXq9VetWzYGzKt52XU5xvqgzYnDK9URnRoJMk1j8nLwEVsaSWJ4fhdUyZijBGUicoD --daemon
 pause

src/3rdparty/argon2/CMakeLists.txt

@@ -1,4 +1,4 @@
-cmake_minimum_required(VERSION 2.8.12)
+cmake_minimum_required(VERSION 3.1)
 
 project(argon2 C)
 set(CMAKE_C_STANDARD 99)

src/3rdparty/cl.h

@@ -29,7 +29,7 @@
 #if defined(__APPLE__)
 #   include <OpenCL/cl.h>
 #else
-#   include "3rdparty/CL/cl.h"
+#   include <CL/cl.h>
 #endif
 
 

src/3rdparty/fmt/fmt/format.cc

@@ -5,7 +5,7 @@
 //
 // For the license information refer to format.h.
 
-#include "3rdparty/fmt/format-inl.h"
+#include "fmt/format-inl.h"
 
 FMT_BEGIN_NAMESPACE
 namespace detail {

src/3rdparty/hwloc/CMakeLists.txt

@@ -1,4 +1,4 @@
-cmake_minimum_required (VERSION  2.8.12)
+cmake_minimum_required(VERSION 3.1)
 project (hwloc C)
 
 include_directories(include)

src/3rdparty/hwloc/NEWS

@@ -1,5 +1,5 @@
 Copyright © 2009 CNRS
-Copyright © 2009-2021 Inria.  All rights reserved.
+Copyright © 2009-2022 Inria.  All rights reserved.
 Copyright © 2009-2013 Université Bordeaux
 Copyright © 2009-2011 Cisco Systems, Inc.  All rights reserved.
 Copyright © 2020 Hewlett Packard Enterprise.  All rights reserved.
@@ -17,6 +17,132 @@ bug fixes (and other actions) for each version of hwloc since version
 0.9.
 
 
+Version 2.9.0
+-------------
+* Backends
+  + Expose the memory size of CXL memory devices (Type 3) on Linux.
+  + The LevelZero backend now reports the "XeLinkBandwidth" distance
+    matrix between L0 devices (and subdevices) when available.
+  + Add support for CUDA compute capability up to 9.0.
+* Tools
+  + lstopo now switches to console mode when its output is redirected.
+    Graphical window mode may be forced back with --of window.
+  + hwloc-calc now accepts "numa" in -H, and I/O subtypes such as "gpu"
+    in -I and -N.
+
+
+Version 2.8.0
+-------------
+* API
+  + Add HWLOC_TOPOLOGY_FLAG_NO_DISTANCES, _NO_MEMATTRS and _NO_CPUKINDS
+    to reduce the overhead when unneeded.
+  + Add separate Read/Write Bandwidth/Latency memory attributes and
+    implement them on Linux.
+* Backends
+  + NUMA nodes may now have a subtype such as DRAM, HBM, SPM, or NVM
+    on heterogeneous memory platforms on Linux.
+    - Add DAXType and DAXParent attributes on Linux to tell where a
+      DAX device or its corresponding NUMA node come from (SPM for
+      Specific-Purpose or NVM for Non-Volatile Memory).
+  + Detect heterogeneous caches in hybrid CPUs on MacOS X,
+    thanks to Paul Bone for the help.
+  + Max frequencies are not ignored in Linux cpukinds anymore (they were
+    ignored in hwloc 2.7.0), but they may be slightly adjusted to avoid
+    reporting hybrid CPUs because Intel Turbo Boost Max 3.0.
+    - See the documentation of environment variable HWLOC_CPUKINDS_MAXFREQ.
+  + Hardwire the PCI locality of HPE Cray EX235a nodes.
+* Tools
+  + lstopo and other tools may now load Linux and x86 cpuid topology files
+    from a tarball.
+  + lstopo may now replace the P# and L# index prefixes with custom strings
+    thanks to --os-index-prefix and --logical-index-prefix options.
+* Misc
+  + Add --disable-readme to avoid regenerating the top-level hwloc README
+    file from the documentation.
+
+
+Version 2.7.1
+-------------
+* Workaround crashes when virtual machines report incoherent x86 CPUID
+  information about numbers of cores and threads.
+  Thanks to Peter Bense for the report.
+* Use setenv() instead of putenv() when trying to force enable oneAPI L0
+  support, to avoid issues with applications that touch the environment,
+  thanks to Josh Hursey for the patch.
+* Add some warnings at the end of configure when GPU libraries are
+  missing on the system or their path is missing in the environment.
+
+
+Version 2.7.0
+-------------
+* Backends
+  + Add support for NUMA nodes and caches with more than 64 PUs across
+    multiple processor groups on Windows 11 and Windows Server 2022.
+  + Group objects are not created for Windows processor groups anymore,
+    except if HWLOC_WINDOWS_PROCESSOR_GROUP_OBJS=1 in the environment.
+  + Expose "Cluster" group objects on Linux kernel 5.16+ for CPUs
+    that share some internal cache or bus. This can be equivalent
+    to the L2 Cache level on some platforms (e.g. x86) or a specific
+    level between L2 and L3 on others (e.g. ARM Kungpeng 920).
+    Thanks to Jonathan Cameron for the help.
+    - HWLOC_DONT_MERGE_CLUSTER_GROUPS=1 may be set in the environment
+      to prevent these groups from being merged with identical caches, etc.
+  + Improve the oneAPI LevelZero backend:
+    - Expose subdevices such as "ze0.1" inside root OS devices ("ze0")
+      when the hardware contains multiple subdevices.
+    - Add many new attributes to describe device type, and the
+      numbers of slices, subslices, execution units and threads.
+    - Expose the memory information as LevelZeroHBM/DDR/MemorySize infos.
+  + Ignore the max frequencies of cores in Linux cpukinds when the
+    base frequencies are available (to avoid exposing hybrid CPUs
+    when Intel Turbo Boost Max 3.0 gives slightly different max
+    frequencies to CPU cores).
+    - May be reverted by setting HWLOC_CPUKINDS_MAXFREQ=1 in the environment.
+* Tools
+  + Add --grey and --palette options to switch lstopo to greyscale or
+    white-background-only graphics, or to tune individual colors.
+* Build
+  + Windows CMake builds now support non-MSVC compilers, detect several
+    features at build time, can build/run tests, etc.
+    Thanks to Michael Hirsch and Alexander Neumann .
+
+
+Version 2.6.0
+-------------
+* Backends
+  + Expose two cpukinds for energy-efficient cores (icestorm) and
+    high-performance cores (firestorm) on Apple M1 on Mac OS X.
+  + Use sysfs CPU "capacity" to rank hybrid cores by efficiency
+    on Linux when available (mostly on recent ARM platforms for now).
+  + Improve HWLOC_MEMBIND_BIND (without the STRICT flag) on Linux kernel
+    >= 5.15: If more than one node is given, the kernel may now use all
+    of them instead of only the first one before falling back to others.
+  + Expose cache os_index when available on Linux, it may be needed
+    when using resctrl to configure cache partitioning, memory bandwidth
+    monitoring, etc.
+  + Add a "XGMIHops" distances matrix in the RSMI backend for AMD GPU
+    interconnected through XGMI links.
+  + Expose AMD GPU memory information (VRAM and GTT) in the RSMI backend.
+  + Add OS devices such as "bxi0" for Atos/Bull BXI HCAs on Linux.
+* Tools
+  + lstopo has a better placement algorithm with respect to I/O
+    objects, see --children-order in the manpage for details.
+  + hwloc-annotate may now change object subtypes and cache or memory
+    sizes.
+* Build
+  + Allow to specify the ROCm installation for building the RSMI backend:
+    - Use a custom installation path if specified with --with-rocm=<dir>.
+    - Use /opt/rocm-<version> if specified with --with-rocm-version=<version>
+      or the ROCM_VERSION environment variable.
+    - Try /opt/rocm if it exists.
+    - See "How do I enable ROCm SMI and select which version to use?"
+      in the FAQ for details.
+  + Add a CMakeLists for Windows under contrib/windows-cmake/ .
+* Documentation
+  + Add FAQ entry "How do I create a custom heterogeneous and
+     asymmetric topology?"
+
+
 Version 2.5.0
 -------------
 * API

src/3rdparty/hwloc/README

@@ -78,7 +78,7 @@ debug and report issues.
 Questions may be sent to the users or developers mailing lists (https://
 www.open-mpi.org/community/lists/hwloc.php).
 
-There is also a #hwloc IRC channel on Freenode (irc.freenode.net).
+There is also a #hwloc IRC channel on Libera Chat (irc.libera.chat).
 
 
 

src/3rdparty/hwloc/VERSION

@@ -8,7 +8,7 @@
 # Please update HWLOC_VERSION* in contrib/windows/hwloc_config.h too.
 
 major=2
-minor=5
+minor=9
 release=0
 
 # greek is used for alpha or beta release tags.  If it is non-empty,
@@ -22,7 +22,7 @@ greek=
 
 # The date when this release was created
 
-date="Jun 14, 2021"
+date="Dec 14, 2022"
 
 # If snapshot=1, then use the value from snapshot_version as the
 # entire hwloc version (i.e., ignore major, minor, release, and
@@ -41,7 +41,7 @@ snapshot_version=${major}.${minor}.${release}${greek}-git
 # 2. Version numbers are described in the Libtool current:revision:age
 # format.
 
-libhwloc_so_version=20:0:5
+libhwloc_so_version=21:1:6
 libnetloc_so_version=0:0:0
 
 # Please also update the <TargetName> lines in contrib/windows/libhwloc.vcxproj

src/3rdparty/hwloc/include/hwloc.h

@@ -1,6 +1,6 @@
 /*
  * Copyright © 2009 CNRS
- * Copyright © 2009-2021 Inria.  All rights reserved.
+ * Copyright © 2009-2022 Inria.  All rights reserved.
  * Copyright © 2009-2012 Université Bordeaux
  * Copyright © 2009-2020 Cisco Systems, Inc.  All rights reserved.
  * See COPYING in top-level directory.
@@ -29,7 +29,7 @@
  * THAT IS IN THE PDF/HTML THAT IS ***NOT*** IN hwloc.h!
  *
  * There are entire paragraph-length descriptions, discussions, and
- * pretty prictures to explain subtle corner cases, provide concrete
+ * pretty pictures to explain subtle corner cases, provide concrete
  * examples, etc.
  *
  * Please, go read the documentation.  :-)
@@ -93,7 +93,7 @@ extern "C" {
  * Two stable releases of the same series usually have the same ::HWLOC_API_VERSION
  * even if their HWLOC_VERSION are different.
  */
-#define HWLOC_API_VERSION 0x00020500
+#define HWLOC_API_VERSION 0x00020800
 
 /** \brief Indicate at runtime which hwloc API version was used at build time.
  *
@@ -346,7 +346,8 @@ typedef enum hwloc_obj_osdev_type_e {
 				  * For instance the "eth0" interface on Linux. */
   HWLOC_OBJ_OSDEV_OPENFABRICS,	/**< \brief Operating system openfabrics device.
 				  * For instance the "mlx4_0" InfiniBand HCA,
-				  * or "hfi1_0" Omni-Path interface on Linux. */
+				  * "hfi1_0" Omni-Path interface,
+				  * or "bxi0" Atos/Bull BXI HCA on Linux. */
   HWLOC_OBJ_OSDEV_DMA,		/**< \brief Operating system dma engine device.
 				  * For instance the "dma0chan0" DMA channel on Linux. */
   HWLOC_OBJ_OSDEV_COPROC	/**< \brief Operating system co-processor device.
@@ -516,7 +517,7 @@ struct hwloc_obj {
                                           * objects).
                                           *
                                           * If the ::HWLOC_TOPOLOGY_FLAG_INCLUDE_DISALLOWED configuration flag is set,
-                                          * some of these CPUs may not be allowed for binding,
+                                          * some of these CPUs may be online but not allowed for binding,
                                           * see hwloc_topology_get_allowed_cpuset().
                                           *
 					  * \note All objects have non-NULL CPU and node sets except Misc and I/O objects.
@@ -548,7 +549,7 @@ struct hwloc_obj {
                                           * nodes more precisely.
                                           *
                                           * If the ::HWLOC_TOPOLOGY_FLAG_INCLUDE_DISALLOWED configuration flag is set,
-                                          * some of these nodes may not be allowed for allocation,
+                                          * some of these nodes may be online but not allowed for allocation,
                                           * see hwloc_topology_get_allowed_nodeset().
                                           *
                                           * If there are no NUMA nodes in the machine, all the memory is close to this
@@ -641,7 +642,7 @@ union hwloc_obj_attr_u {
     unsigned char revision;
     float linkspeed; /* in GB/s */
   } pcidev;
-  /** \brief Bridge specific Object Attribues */
+  /** \brief Bridge specific Object Attributes */
   struct hwloc_bridge_attr_s {
     union {
       struct hwloc_pcidev_attr_s pci;
@@ -970,7 +971,7 @@ HWLOC_DECLSPEC const char * hwloc_obj_type_string (hwloc_obj_type_t type) __hwlo
  *
  * If \p size is 0, \p string may safely be \c NULL.
  *
- * \return the number of character that were actually written if not truncating,
+ * \return the number of characters that were actually written if not truncating,
  * or that would have been written (not including the ending \\0).
  */
 HWLOC_DECLSPEC int hwloc_obj_type_snprintf(char * __hwloc_restrict string, size_t size,
@@ -985,7 +986,7 @@ HWLOC_DECLSPEC int hwloc_obj_type_snprintf(char * __hwloc_restrict string, size_
  *
  * If \p size is 0, \p string may safely be \c NULL.
  *
- * \return the number of character that were actually written if not truncating,
+ * \return the number of characters that were actually written if not truncating,
  * or that would have been written (not including the ending \\0).
  */
 HWLOC_DECLSPEC int hwloc_obj_attr_snprintf(char * __hwloc_restrict string, size_t size,
@@ -1088,7 +1089,7 @@ HWLOC_DECLSPEC int hwloc_obj_add_info(hwloc_obj_t obj, const char *name, const c
  *
  * Some operating systems only support binding threads or processes to a single PU.
  * Others allow binding to larger sets such as entire Cores or Packages or
- * even random sets of invididual PUs. In such operating system, the scheduler
+ * even random sets of individual PUs. In such operating system, the scheduler
  * is free to run the task on one of these PU, then migrate it to another PU, etc.
  * It is often useful to call hwloc_bitmap_singlify() on the target CPU set before
  * passing it to the binding function to avoid these expensive migrations.
@@ -1166,7 +1167,7 @@ typedef enum {
    * CPUs are idle, operating systems may execute the thread/process
    * on those other CPUs instead of the designated CPUs, to let them
    * progress anyway.  Strict binding means that the thread/process
-   * will _never_ execute on other cpus than the designated CPUs, even
+   * will _never_ execute on other CPUs than the designated CPUs, even
    * when those are busy with other tasks and other CPUs are idle.
    *
    * \note Depending on the operating system, strict binding may not
@@ -1203,7 +1204,7 @@ typedef enum {
   HWLOC_CPUBIND_NOMEMBIND = (1<<3)
 } hwloc_cpubind_flags_t;
 
-/** \brief Bind current process or thread on cpus given in physical bitmap \p set.
+/** \brief Bind current process or thread on CPUs given in physical bitmap \p set.
  *
  * \return -1 with errno set to ENOSYS if the action is not supported
  * \return -1 with errno set to EXDEV if the binding cannot be enforced
@@ -1212,12 +1213,13 @@ HWLOC_DECLSPEC int hwloc_set_cpubind(hwloc_topology_t topology, hwloc_const_cpus
 
 /** \brief Get current process or thread binding.
  *
- * Writes into \p set the physical cpuset which the process or thread (according to \e
- * flags) was last bound to.
+ * The CPU-set \p set (previously allocated by the caller)
+ * is filled with the list of PUs which the process or
+ * thread (according to \e flags) was last bound to.
  */
 HWLOC_DECLSPEC int hwloc_get_cpubind(hwloc_topology_t topology, hwloc_cpuset_t set, int flags);
 
-/** \brief Bind a process \p pid on cpus given in physical bitmap \p set.
+/** \brief Bind a process \p pid on CPUs given in physical bitmap \p set.
  *
  * \note \p hwloc_pid_t is \p pid_t on Unix platforms,
  * and \p HANDLE on native Windows platforms.
@@ -1231,6 +1233,10 @@ HWLOC_DECLSPEC int hwloc_get_cpubind(hwloc_topology_t topology, hwloc_cpuset_t s
 HWLOC_DECLSPEC int hwloc_set_proc_cpubind(hwloc_topology_t topology, hwloc_pid_t pid, hwloc_const_cpuset_t set, int flags);
 
 /** \brief Get the current physical binding of process \p pid.
+ *
+ * The CPU-set \p set (previously allocated by the caller)
+ * is filled with the list of PUs which the process
+ * was last bound to.
  *
  * \note \p hwloc_pid_t is \p pid_t on Unix platforms,
  * and \p HANDLE on native Windows platforms.
@@ -1244,7 +1250,7 @@ HWLOC_DECLSPEC int hwloc_set_proc_cpubind(hwloc_topology_t topology, hwloc_pid_t
 HWLOC_DECLSPEC int hwloc_get_proc_cpubind(hwloc_topology_t topology, hwloc_pid_t pid, hwloc_cpuset_t set, int flags);
 
 #ifdef hwloc_thread_t
-/** \brief Bind a thread \p thread on cpus given in physical bitmap \p set.
+/** \brief Bind a thread \p thread on CPUs given in physical bitmap \p set.
  *
  * \note \p hwloc_thread_t is \p pthread_t on Unix platforms,
  * and \p HANDLE on native Windows platforms.
@@ -1256,6 +1262,10 @@ HWLOC_DECLSPEC int hwloc_set_thread_cpubind(hwloc_topology_t topology, hwloc_thr
 
 #ifdef hwloc_thread_t
 /** \brief Get the current physical binding of thread \p tid.
+ *
+ * The CPU-set \p set (previously allocated by the caller)
+ * is filled with the list of PUs which the thread
+ * was last bound to.
  *
  * \note \p hwloc_thread_t is \p pthread_t on Unix platforms,
  * and \p HANDLE on native Windows platforms.
@@ -1266,6 +1276,10 @@ HWLOC_DECLSPEC int hwloc_get_thread_cpubind(hwloc_topology_t topology, hwloc_thr
 #endif
 
 /** \brief Get the last physical CPU where the current process or thread ran.
+ *
+ * The CPU-set \p set (previously allocated by the caller)
+ * is filled with the list of PUs which the process or
+ * thread (according to \e flags) last ran on.
  *
  * The operating system may move some tasks from one processor
  * to another at any time according to their binding,
@@ -1281,6 +1295,10 @@ HWLOC_DECLSPEC int hwloc_get_thread_cpubind(hwloc_topology_t topology, hwloc_thr
 HWLOC_DECLSPEC int hwloc_get_last_cpu_location(hwloc_topology_t topology, hwloc_cpuset_t set, int flags);
 
 /** \brief Get the last physical CPU where a process ran.
+ *
+ * The CPU-set \p set (previously allocated by the caller)
+ * is filled with the list of PUs which the process
+ * last ran on.
  *
  * The operating system may move some tasks from one processor
  * to another at any time according to their binding,
@@ -1511,6 +1529,9 @@ HWLOC_DECLSPEC int hwloc_set_membind(hwloc_topology_t topology, hwloc_const_bitm
 /** \brief Query the default memory binding policy and physical locality of the
  * current process or thread.
  *
+ * The bitmap \p set (previously allocated by the caller)
+ * is filled with the process or thread memory binding.
+ *
  * This function has two output parameters: \p set and \p policy.
  * The values returned in these parameters depend on both the \p flags
  * passed in and the current memory binding policies and nodesets in
@@ -1571,6 +1592,9 @@ HWLOC_DECLSPEC int hwloc_set_proc_membind(hwloc_topology_t topology, hwloc_pid_t
 /** \brief Query the default memory binding policy and physical locality of the
  * specified process.
  *
+ * The bitmap \p set (previously allocated by the caller)
+ * is filled with the process memory binding.
+ *
  * This function has two output parameters: \p set and \p policy.
  * The values returned in these parameters depend on both the \p flags
  * passed in and the current memory binding policies and nodesets in
@@ -1624,6 +1648,9 @@ HWLOC_DECLSPEC int hwloc_set_area_membind(hwloc_topology_t topology, const void
 /** \brief Query the CPUs near the physical NUMA node(s) and binding policy of
  * the memory identified by (\p addr, \p len ).
  *
+ * The bitmap \p set (previously allocated by the caller)
+ * is filled with the memory area binding.
+ *
  * This function has two output parameters: \p set and \p policy.
  * The values returned in these parameters depend on both the \p flags
  * passed in and the memory binding policies and nodesets of the pages
@@ -1652,7 +1679,8 @@ HWLOC_DECLSPEC int hwloc_get_area_membind(hwloc_topology_t topology, const void
 
 /** \brief Get the NUMA nodes where memory identified by (\p addr, \p len ) is physically allocated.
  *
- * Fills \p set according to the NUMA nodes where the memory area pages
+ * The bitmap \p set (previously allocated by the caller)
+ * is filled according to the NUMA nodes where the memory area pages
  * are physically allocated. If no page is actually allocated yet,
  * \p set may be empty.
  *
@@ -1698,9 +1726,12 @@ HWLOC_DECLSPEC void *hwloc_alloc_membind(hwloc_topology_t topology, size_t len,
 
 /** \brief Allocate some memory on NUMA memory nodes specified by \p set
  *
- * This is similar to hwloc_alloc_membind_nodeset() except that it is allowed to change
- * the current memory binding policy, thus providing more binding support, at
- * the expense of changing the current state.
+ * First, try to allocate properly with hwloc_alloc_membind().
+ * On failure, the current process or thread memory binding policy
+ * is changed with hwloc_set_membind() before allocating memory.
+ * Thus this function works in more cases, at the expense of changing
+ * the current state (possibly affecting future allocations that
+ * would not specify any policy).
  *
  * If ::HWLOC_MEMBIND_BYNODESET is specified, set is considered a nodeset.
  * Otherwise it's a cpuset.
@@ -1883,8 +1914,9 @@ HWLOC_DECLSPEC int hwloc_topology_set_components(hwloc_topology_t __hwloc_restri
 enum hwloc_topology_flags_e {
  /** \brief Detect the whole system, ignore reservations, include disallowed objects.
    *
-   * Gather all resources, even if some were disabled by the administrator.
+   * Gather all online resources, even if some were disabled by the administrator.
    * For instance, ignore Linux Cgroup/Cpusets and gather all processors and memory nodes.
+   * However offline PUs and NUMA nodes are still ignored.
    *
    * When this flag is not set, PUs and NUMA nodes that are disallowed are not added to the topology.
    * Parent objects (package, core, cache, etc.) are added only if some of their children are allowed.
@@ -2028,17 +2060,38 @@ enum hwloc_topology_flags_e {
    * not change to due thread binding changes on Windows
    * (see ::HWLOC_TOPOLOGY_FLAG_RESTRICT_TO_CPUBINDING).
    */
-  HWLOC_TOPOLOGY_FLAG_DONT_CHANGE_BINDING = (1UL<<6)
+  HWLOC_TOPOLOGY_FLAG_DONT_CHANGE_BINDING = (1UL<<6),
+
+  /** \brief Ignore distances.
+   *
+   * Ignore distance information from the operating systems (and from XML)
+   * and hence do not use distances for grouping.
+   */
+  HWLOC_TOPOLOGY_FLAG_NO_DISTANCES = (1UL<<7),
+
+  /** \brief Ignore memory attributes.
+   *
+   * Ignore memory attribues from the operating systems (and from XML).
+   */
+  HWLOC_TOPOLOGY_FLAG_NO_MEMATTRS = (1UL<<8),
+
+  /** \brief Ignore CPU Kinds.
+   *
+   * Ignore CPU kind information from the operating systems (and from XML).
+   */
+  HWLOC_TOPOLOGY_FLAG_NO_CPUKINDS = (1UL<<9)
 };
 
 /** \brief Set OR'ed flags to non-yet-loaded topology.
  *
  * Set a OR'ed set of ::hwloc_topology_flags_e onto a topology that was not yet loaded.
  *
- * If this function is called multiple times, the last invokation will erase
+ * If this function is called multiple times, the last invocation will erase
  * and replace the set of flags that was previously set.
  *
- * The flags set in a topology may be retrieved with hwloc_topology_get_flags()
+ * By default, no flags are set (\c 0).
+ *
+ * The flags set in a topology may be retrieved with hwloc_topology_get_flags().
  */
 HWLOC_DECLSPEC int hwloc_topology_set_flags (hwloc_topology_t topology, unsigned long flags);
 
@@ -2046,6 +2099,9 @@ HWLOC_DECLSPEC int hwloc_topology_set_flags (hwloc_topology_t topology, unsigned
  *
  * Get the OR'ed set of ::hwloc_topology_flags_e of a topology.
  *
+ * If hwloc_topology_set_flags() was not called earlier,
+ * no flags are set (\c 0 is returned).
+ *
  * \return the flags previously set with hwloc_topology_set_flags().
  */
 HWLOC_DECLSPEC unsigned long hwloc_topology_get_flags (hwloc_topology_t topology);

src/3rdparty/hwloc/include/hwloc/autogen/config.h

@@ -1,6 +1,6 @@
 /*
  * Copyright © 2009 CNRS
- * Copyright © 2009-2020 Inria.  All rights reserved.
+ * Copyright © 2009-2022 Inria.  All rights reserved.
  * Copyright © 2009-2012 Université Bordeaux
  * Copyright © 2009-2011 Cisco Systems, Inc.  All rights reserved.
  * See COPYING in top-level directory.
@@ -11,9 +11,9 @@
 #ifndef HWLOC_CONFIG_H
 #define HWLOC_CONFIG_H
 
-#define HWLOC_VERSION "2.5.0"
+#define HWLOC_VERSION "2.9.0"
 #define HWLOC_VERSION_MAJOR 2
-#define HWLOC_VERSION_MINOR 5
+#define HWLOC_VERSION_MINOR 9
 #define HWLOC_VERSION_RELEASE 0
 #define HWLOC_VERSION_GREEK ""
 

src/3rdparty/hwloc/include/hwloc/bitmap.h

@@ -1,6 +1,6 @@
 /*
  * Copyright © 2009 CNRS
- * Copyright © 2009-2020 Inria.  All rights reserved.
+ * Copyright © 2009-2022 Inria.  All rights reserved.
  * Copyright © 2009-2012 Université Bordeaux
  * Copyright © 2009-2011 Cisco Systems, Inc.  All rights reserved.
  * See COPYING in top-level directory.
@@ -112,7 +112,7 @@ HWLOC_DECLSPEC int hwloc_bitmap_copy(hwloc_bitmap_t dst, hwloc_const_bitmap_t sr
  *
  * If \p buflen is 0, \p buf may safely be \c NULL.
  *
- * \return the number of character that were actually written if not truncating,
+ * \return the number of characters that were actually written if not truncating,
  * or that would have been written (not including the ending \\0).
  */
 HWLOC_DECLSPEC int hwloc_bitmap_snprintf(char * __hwloc_restrict buf, size_t buflen, hwloc_const_bitmap_t bitmap);
@@ -137,7 +137,7 @@ HWLOC_DECLSPEC int hwloc_bitmap_sscanf(hwloc_bitmap_t bitmap, const char * __hwl
  *
  * If \p buflen is 0, \p buf may safely be \c NULL.
  *
- * \return the number of character that were actually written if not truncating,
+ * \return the number of characters that were actually written if not truncating,
  * or that would have been written (not including the ending \\0).
  */
 HWLOC_DECLSPEC int hwloc_bitmap_list_snprintf(char * __hwloc_restrict buf, size_t buflen, hwloc_const_bitmap_t bitmap);
@@ -161,7 +161,7 @@ HWLOC_DECLSPEC int hwloc_bitmap_list_sscanf(hwloc_bitmap_t bitmap, const char *
  *
  * If \p buflen is 0, \p buf may safely be \c NULL.
  *
- * \return the number of character that were actually written if not truncating,
+ * \return the number of characters that were actually written if not truncating,
  * or that would have been written (not including the ending \\0).
  */
 HWLOC_DECLSPEC int hwloc_bitmap_taskset_snprintf(char * __hwloc_restrict buf, size_t buflen, hwloc_const_bitmap_t bitmap);
@@ -357,11 +357,11 @@ HWLOC_DECLSPEC int hwloc_bitmap_last_unset(hwloc_const_bitmap_t bitmap) __hwloc_
  * The loop must start with hwloc_bitmap_foreach_begin() and end
  * with hwloc_bitmap_foreach_end() followed by a terminating ';'.
  *
- * \p index is the loop variable; it should be an unsigned int.  The
- * first iteration will set \p index to the lowest index in the bitmap.
+ * \p id is the loop variable; it should be an unsigned int.  The
+ * first iteration will set \p id to the lowest index in the bitmap.
  * Successive iterations will iterate through, in order, all remaining
  * indexes set in the bitmap.  To be specific: each iteration will return a
- * value for \p index such that hwloc_bitmap_isset(bitmap, index) is true.
+ * value for \p id such that hwloc_bitmap_isset(bitmap, id) is true.
  *
  * The assert prevents the loop from being infinite if the bitmap is infinitely set.
  *

src/3rdparty/hwloc/include/hwloc/cpukinds.h

@@ -1,5 +1,5 @@
 /*
- * Copyright © 2020 Inria.  All rights reserved.
+ * Copyright © 2020-2021 Inria.  All rights reserved.
  * See COPYING in top-level directory.
  */
 
@@ -42,18 +42,23 @@ extern "C" {
  * (for instance the "CoreType" and "FrequencyMaxMHz",
  *  see \ref topoattrs_cpukinds).
  *
- * A higher efficiency value means intrinsic greater performance
+ * A higher efficiency value means greater intrinsic performance
  * (and possibly less performance/power efficiency).
- * Kinds with lower efficiency are ranked first:
+ * Kinds with lower efficiency values are ranked first:
  * Passing 0 as \p kind_index to hwloc_cpukinds_get_info() will
- * return information about the less efficient CPU kind.
+ * return information about the CPU kind with lower performance
+ * but higher energy-efficiency.
+ * Higher \p kind_index values would rather return information
+ * about power-hungry high-performance cores.
  *
- * When available, efficiency values are gathered from the operating
- * system (when \p cpukind_efficiency is set in the
- * struct hwloc_topology_discovery_support array, only on Windows 10 for now).
- * Otherwise hwloc tries to compute efficiencies
- * by comparing CPU kinds using frequencies (on ARM),
- * or core types and frequencies (on other architectures).
+ * When available, efficiency values are gathered from the operating system.
+ * If so, \p cpukind_efficiency is set in the struct hwloc_topology_discovery_support array.
+ * This is currently available on Windows 10, Mac OS X (Darwin),
+ * and on some Linux platforms where core "capacity" is exposed in sysfs.
+ *
+ * If the operating system does not expose core efficiencies natively,
+ * hwloc tries to compute efficiencies by comparing CPU kinds using
+ * frequencies (on ARM), or core types and frequencies (on other architectures).
  * The environment variable HWLOC_CPUKINDS_RANKING may be used
  * to change this heuristics, see \ref envvar.
  *

src/3rdparty/hwloc/include/hwloc/deprecated.h

@@ -1,6 +1,6 @@
 /*
  * Copyright © 2009 CNRS
- * Copyright © 2009-2021 Inria.  All rights reserved.
+ * Copyright © 2009-2022 Inria.  All rights reserved.
  * Copyright © 2009-2012 Université Bordeaux
  * Copyright © 2009-2010 Cisco Systems, Inc.  All rights reserved.
  * See COPYING in top-level directory.
@@ -55,7 +55,7 @@ hwloc_topology_insert_misc_object_by_parent(hwloc_topology_t topology, hwloc_obj
  *
  * If \p size is 0, \p string may safely be \c NULL.
  *
- * \return the number of character that were actually written if not truncating,
+ * \return the number of characters that were actually written if not truncating,
  * or that would have been written (not including the ending \\0).
  */
 static __hwloc_inline int

src/3rdparty/hwloc/include/hwloc/distances.h

@@ -1,5 +1,5 @@
 /*
- * Copyright © 2010-2021 Inria.  All rights reserved.
+ * Copyright © 2010-2022 Inria.  All rights reserved.
  * See COPYING in top-level directory.
  */
 
@@ -35,7 +35,8 @@ extern "C" {
  * from a core in another node.
  * The corresponding kind is ::HWLOC_DISTANCES_KIND_FROM_OS | ::HWLOC_DISTANCES_KIND_FROM_USER.
  * The name of this distances structure is "NUMALatency".
- * Others distance structures include and "XGMIBandwidth" and "NVLinkBandwidth".
+ * Others distance structures include and "XGMIBandwidth", "XGMIHops",
+ * "XeLinkBandwidth" and "NVLinkBandwidth".
  *
  * The matrix may also contain bandwidths between random sets of objects,
  * possibly provided by the user, as specified in the \p kind attribute.
@@ -159,7 +160,8 @@ hwloc_distances_get_by_type(hwloc_topology_t topology, hwloc_obj_type_t type,
  * Usually only one distances structure may match a given name.
  *
  * The name of the most common structure is "NUMALatency".
- * Others include "XGMIBandwidth" and "NVLinkBandwidth".
+ * Others include "XGMIBandwidth", "XGMIHops", "XeLinkBandwidth",
+ * and "NVLinkBandwidth".
  */
 HWLOC_DECLSPEC int
 hwloc_distances_get_by_name(hwloc_topology_t topology, const char *name,

src/3rdparty/hwloc/include/hwloc/helper.h

@@ -1,6 +1,6 @@
 /*
  * Copyright © 2009 CNRS
- * Copyright © 2009-2021 Inria.  All rights reserved.
+ * Copyright © 2009-2022 Inria.  All rights reserved.
  * Copyright © 2009-2012 Université Bordeaux
  * Copyright © 2009-2010 Cisco Systems, Inc.  All rights reserved.
  * See COPYING in top-level directory.
@@ -886,9 +886,6 @@ enum hwloc_distrib_flags_e {
  * \p flags should be 0 or a OR'ed set of ::hwloc_distrib_flags_e.
  *
  * \note This function requires the \p roots objects to have a CPU set.
- *
- * \note This function replaces the now deprecated hwloc_distribute()
- * and hwloc_distributev() functions.
  */
 static __hwloc_inline int
 hwloc_distrib(hwloc_topology_t topology,

src/3rdparty/hwloc/include/hwloc/intel-mic.h

@@ -1,136 +0,0 @@
-/*
- * Copyright © 2013-2016 Inria.  All rights reserved.
- * See COPYING in top-level directory.
- */
-
-/** \file
- * \brief Macros to help interaction between hwloc and Intel Xeon Phi (MIC).
- *
- * Applications that use both hwloc and Intel Xeon Phi (MIC) may want to
- * include this file so as to get topology information for MIC devices.
- */
-
-#ifndef HWLOC_INTEL_MIC_H
-#define HWLOC_INTEL_MIC_H
-
-#include "hwloc.h"
-#include "hwloc/autogen/config.h"
-#include "hwloc/helper.h"
-
-#ifdef HWLOC_LINUX_SYS
-#include "hwloc/linux.h"
-
-#include <dirent.h>
-#include <string.h>
-#endif
-
-#include <stdio.h>
-#include <stdlib.h>
-
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-
-/** \defgroup hwlocality_intel_mic Interoperability with Intel Xeon Phi (MIC)
- *
- * This interface offers ways to retrieve topology information about
- * Intel Xeon Phi (MIC) devices.
- *
- * @{
- */
-
-/** \brief Get the CPU set of logical processors that are physically
- * close to MIC device whose index is \p idx.
- *
- * Return the CPU set describing the locality of the MIC device whose index is \p idx.
- *
- * Topology \p topology and device index \p idx must match the local machine.
- * I/O devices detection is not needed in the topology.
- *
- * The function only returns the locality of the device.
- * If more information about the device is needed, OS objects should
- * be used instead, see hwloc_intel_mic_get_device_osdev_by_index().
- *
- * This function is currently only implemented in a meaningful way for
- * Linux; other systems will simply get a full cpuset.
- */
-static __hwloc_inline int
-hwloc_intel_mic_get_device_cpuset(hwloc_topology_t topology __hwloc_attribute_unused,
-				  int idx __hwloc_attribute_unused,
-				  hwloc_cpuset_t set)
-{
-#ifdef HWLOC_LINUX_SYS
-	/* If we're on Linux, use the sysfs mechanism to get the local cpus */
-#define HWLOC_INTEL_MIC_DEVICE_SYSFS_PATH_MAX 128
-	char path[HWLOC_INTEL_MIC_DEVICE_SYSFS_PATH_MAX];
-	DIR *sysdir = NULL;
-	struct dirent *dirent;
-	unsigned pcibus, pcidev, pcifunc;
-
-	if (!hwloc_topology_is_thissystem(topology)) {
-		errno = EINVAL;
-		return -1;
-	}
-
-	sprintf(path, "/sys/class/mic/mic%d", idx);
-	sysdir = opendir(path);
-	if (!sysdir)
-		return -1;
-
-	while ((dirent = readdir(sysdir)) != NULL) {
-		if (sscanf(dirent->d_name, "pci_%02x:%02x.%02x", &pcibus, &pcidev, &pcifunc) == 3) {
-			sprintf(path, "/sys/class/mic/mic%d/pci_%02x:%02x.%02x/local_cpus", idx, pcibus, pcidev, pcifunc);
-			if (hwloc_linux_read_path_as_cpumask(path, set) < 0
-			    || hwloc_bitmap_iszero(set))
-				hwloc_bitmap_copy(set, hwloc_topology_get_complete_cpuset(topology));
-			break;
-		}
-	}
-
-	closedir(sysdir);
-#else
-	/* Non-Linux systems simply get a full cpuset */
-	hwloc_bitmap_copy(set, hwloc_topology_get_complete_cpuset(topology));
-#endif
-	return 0;
-}
-
-/** \brief Get the hwloc OS device object corresponding to the
- * MIC device for the given index.
- *
- * Return the OS device object describing the MIC device whose index is \p idx.
- * Return NULL if there is none.
- *
- * The topology \p topology does not necessarily have to match the current
- * machine. For instance the topology may be an XML import of a remote host.
- * I/O devices detection must be enabled in the topology.
- *
- * \note The corresponding PCI device object can be obtained by looking
- * at the OS device parent object.
- */
-static __hwloc_inline hwloc_obj_t
-hwloc_intel_mic_get_device_osdev_by_index(hwloc_topology_t topology,
-					  unsigned idx)
-{
-	hwloc_obj_t osdev = NULL;
-	while ((osdev = hwloc_get_next_osdev(topology, osdev)) != NULL) {
-		if (HWLOC_OBJ_OSDEV_COPROC == osdev->attr->osdev.type
-                    && osdev->name
-		    && !strncmp("mic", osdev->name, 3)
-		    && atoi(osdev->name + 3) == (int) idx)
-                        return osdev;
-        }
-        return NULL;
-}
-
-/** @} */
-
-
-#ifdef __cplusplus
-} /* extern "C" */
-#endif
-
-
-#endif /* HWLOC_INTEL_MIC_H */

src/3rdparty/hwloc/include/hwloc/linux.h

@@ -1,6 +1,6 @@
 /*
  * Copyright © 2009 CNRS
- * Copyright © 2009-2016 Inria.  All rights reserved.
+ * Copyright © 2009-2021 Inria.  All rights reserved.
  * Copyright © 2009-2011 Université Bordeaux
  * See COPYING in top-level directory.
  */
@@ -44,6 +44,10 @@ extern "C" {
 HWLOC_DECLSPEC int hwloc_linux_set_tid_cpubind(hwloc_topology_t topology, pid_t tid, hwloc_const_cpuset_t set);
 
 /** \brief Get the current binding of thread \p tid
+ *
+ * The CPU-set \p set (previously allocated by the caller)
+ * is filled with the list of PUs which the thread
+ * was last bound to.
  *
  * The behavior is exactly the same as the Linux sched_getaffinity system call,
  * but uses a hwloc cpuset.
@@ -54,6 +58,9 @@ HWLOC_DECLSPEC int hwloc_linux_set_tid_cpubind(hwloc_topology_t topology, pid_t
 HWLOC_DECLSPEC int hwloc_linux_get_tid_cpubind(hwloc_topology_t topology, pid_t tid, hwloc_cpuset_t set);
 
 /** \brief Get the last physical CPU where thread \p tid ran.
+ *
+ * The CPU-set \p set (previously allocated by the caller)
+ * is filled with the PU which the thread last ran on.
  *
  * \note This is equivalent to calling hwloc_get_proc_last_cpu_location() with
  * ::HWLOC_CPUBIND_THREAD as flags.

src/3rdparty/hwloc/include/hwloc/memattrs.h

@@ -1,5 +1,5 @@
 /*
- * Copyright © 2019-2020 Inria.  All rights reserved.
+ * Copyright © 2019-2022 Inria.  All rights reserved.
  * See COPYING in top-level directory.
  */
 
@@ -54,6 +54,8 @@ extern "C" {
  * Attribute values for these nodes, if any, may then be obtained with
  * hwloc_memattr_get_value() and manually compared with the desired criteria.
  *
+ * \sa An example is available in doc/examples/memory-attributes.c in the source tree.
+ *
  * \note The API also supports specific objects as initiator,
  * but it is currently not used internally by hwloc.
  * Users may for instance use it to provide custom performance
@@ -65,19 +67,19 @@ extern "C" {
 
 /** \brief Memory node attributes. */
 enum hwloc_memattr_id_e {
-  /** \brief "Capacity".
-   * The capacity is returned in bytes
-   * (local_memory attribute in objects).
+  /** \brief
+   * The \"Capacity\" is returned in bytes (local_memory attribute in objects).
    *
    * Best capacity nodes are nodes with <b>higher capacity</b>.
    *
    * No initiator is involved when looking at this attribute.
    * The corresponding attribute flags are ::HWLOC_MEMATTR_FLAG_HIGHER_FIRST.
+   * \hideinitializer
    */
   HWLOC_MEMATTR_ID_CAPACITY = 0,
 
-  /** \brief "Locality".
-   * The locality is returned as the number of PUs in that locality
+  /** \brief
+   * The \"Locality\" is returned as the number of PUs in that locality
    * (e.g. the weight of its cpuset).
    *
    * Best locality nodes are nodes with <b>smaller locality</b>
@@ -87,26 +89,87 @@ enum hwloc_memattr_id_e {
    *
    * No initiator is involved when looking at this attribute.
    * The corresponding attribute flags are ::HWLOC_MEMATTR_FLAG_HIGHER_FIRST.
+   * \hideinitializer
    */
   HWLOC_MEMATTR_ID_LOCALITY = 1,
 
-  /** \brief "Bandwidth".
-   * The bandwidth is returned in MiB/s, as seen from the given initiator location.
+  /** \brief
+   * The \"Bandwidth\" is returned in MiB/s, as seen from the given initiator location.
+   *
    * Best bandwidth nodes are nodes with <b>higher bandwidth</b>.
+   *
    * The corresponding attribute flags are ::HWLOC_MEMATTR_FLAG_HIGHER_FIRST
    * and ::HWLOC_MEMATTR_FLAG_NEED_INITIATOR.
+   *
+   * This is the average bandwidth for read and write accesses. If the platform
+   * provides individual read and write bandwidths but no explicit average value,
+   * hwloc computes and returns the average.
+   * \hideinitializer
    */
   HWLOC_MEMATTR_ID_BANDWIDTH = 2,
 
-  /** \brief "Latency".
-   * The latency is returned as nanoseconds, as seen from the given initiator location.
+  /** \brief
+   * The \"ReadBandwidth\" is returned in MiB/s, as seen from the given initiator location.
+   *
+   * Best bandwidth nodes are nodes with <b>higher bandwidth</b>.
+   *
+   * The corresponding attribute flags are ::HWLOC_MEMATTR_FLAG_HIGHER_FIRST
+   * and ::HWLOC_MEMATTR_FLAG_NEED_INITIATOR.
+   * \hideinitializer
+   */
+  HWLOC_MEMATTR_ID_READ_BANDWIDTH = 4,
+
+  /** \brief
+   * The \"WriteBandwidth\" is returned in MiB/s, as seen from the given initiator location.
+   *
+   * Best bandwidth nodes are nodes with <b>higher bandwidth</b>.
+   *
+   * The corresponding attribute flags are ::HWLOC_MEMATTR_FLAG_HIGHER_FIRST
+   * and ::HWLOC_MEMATTR_FLAG_NEED_INITIATOR.
+   * \hideinitializer
+   */
+  HWLOC_MEMATTR_ID_WRITE_BANDWIDTH = 5,
+
+  /** \brief
+   * The \"Latency\" is returned as nanoseconds, as seen from the given initiator location.
+   *
    * Best latency nodes are nodes with <b>smaller latency</b>.
+   *
    * The corresponding attribute flags are ::HWLOC_MEMATTR_FLAG_LOWER_FIRST
    * and ::HWLOC_MEMATTR_FLAG_NEED_INITIATOR.
+   *
+   * This is the average latency for read and write accesses. If the platform
+   * provides individual read and write latencies but no explicit average value,
+   * hwloc computes and returns the average.
+   * \hideinitializer
    */
-  HWLOC_MEMATTR_ID_LATENCY = 3
+  HWLOC_MEMATTR_ID_LATENCY = 3,
 
-  /* TODO read vs write, persistence? */
+  /** \brief
+   * The \"ReadLatency\" is returned as nanoseconds, as seen from the given initiator location.
+   *
+   * Best latency nodes are nodes with <b>smaller latency</b>.
+   *
+   * The corresponding attribute flags are ::HWLOC_MEMATTR_FLAG_LOWER_FIRST
+   * and ::HWLOC_MEMATTR_FLAG_NEED_INITIATOR.
+   * \hideinitializer
+   */
+  HWLOC_MEMATTR_ID_READ_LATENCY = 6,
+
+  /** \brief
+   * The \"WriteLatency\" is returned as nanoseconds, as seen from the given initiator location.
+   *
+   * Best latency nodes are nodes with <b>smaller latency</b>.
+   *
+   * The corresponding attribute flags are ::HWLOC_MEMATTR_FLAG_LOWER_FIRST
+   * and ::HWLOC_MEMATTR_FLAG_NEED_INITIATOR.
+   * \hideinitializer
+   */
+  HWLOC_MEMATTR_ID_WRITE_LATENCY = 7,
+
+  /* TODO persistence? */
+
+  HWLOC_MEMATTR_ID_MAX /**< \private Sentinel value */
 };
 
 /** \brief A memory attribute identifier.
@@ -354,7 +417,7 @@ hwloc_memattr_register(hwloc_topology_t topology,
  * \p flags must be \c 0 for now.
  *
  * \note The initiator \p initiator should be of type ::HWLOC_LOCATION_TYPE_CPUSET
- * when refering to accesses performed by CPU cores.
+ * when referring to accesses performed by CPU cores.
  * ::HWLOC_LOCATION_TYPE_OBJECT is currently unused internally by hwloc,
  * but users may for instance use it to provide custom information about
  * host memory accesses performed by GPUs.
@@ -398,7 +461,7 @@ hwloc_memattr_set_value(hwloc_topology_t topology,
  * values.
  *
  * \note The initiator \p initiator should be of type ::HWLOC_LOCATION_TYPE_CPUSET
- * when refering to accesses performed by CPU cores.
+ * when referring to accesses performed by CPU cores.
  * ::HWLOC_LOCATION_TYPE_OBJECT is currently unused internally by hwloc,
  * but users may for instance use it to provide custom information about
  * host memory accesses performed by GPUs.
@@ -408,7 +471,7 @@ hwloc_memattr_get_targets(hwloc_topology_t topology,
                           hwloc_memattr_id_t attribute,
                           struct hwloc_location *initiator,
                           unsigned long flags,
-                          unsigned *nrp, hwloc_obj_t *targets, hwloc_uint64_t *values);
+                          unsigned *nr, hwloc_obj_t *targets, hwloc_uint64_t *values);
 
 /** \brief Return the initiators that have values for a given attribute for a specific target NUMA node.
  *

src/3rdparty/hwloc/include/hwloc/plugins.h

@@ -1,5 +1,5 @@
 /*
- * Copyright © 2013-2021 Inria.  All rights reserved.
+ * Copyright © 2013-2022 Inria.  All rights reserved.
  * Copyright © 2016 Cisco Systems, Inc.  All rights reserved.
  * See COPYING in top-level directory.
  */
@@ -338,9 +338,15 @@ struct hwloc_component {
  * This function return 1 by default (show critical only),
  * 0 in lstopo (show all),
  * or anything set in HWLOC_HIDE_ERRORS in the environment.
+ *
+ * Use macros HWLOC_SHOW_CRITICAL_ERRORS() and HWLOC_SHOW_ALL_ERRORS()
+ * for clarity.
  */
 HWLOC_DECLSPEC int hwloc_hide_errors(void);
 
+#define HWLOC_SHOW_CRITICAL_ERRORS() (hwloc_hide_errors() < 2)
+#define HWLOC_SHOW_ALL_ERRORS() (hwloc_hide_errors() == 0)
+
 /** \brief Add an object to the topology.
  *
  * Insert new object \p obj in the topology starting under existing object \p root
@@ -497,9 +503,11 @@ hwloc_filter_check_pcidev_subtype_important(unsigned classid)
   return (baseclass == 0x03 /* PCI_BASE_CLASS_DISPLAY */
 	  || baseclass == 0x02 /* PCI_BASE_CLASS_NETWORK */
 	  || baseclass == 0x01 /* PCI_BASE_CLASS_STORAGE */
+	  || baseclass == 0x00 /* Unclassified, for Atos/Bull BXI */
 	  || baseclass == 0x0b /* PCI_BASE_CLASS_PROCESSOR */
 	  || classid == 0x0c04 /* PCI_CLASS_SERIAL_FIBER */
 	  || classid == 0x0c06 /* PCI_CLASS_SERIAL_INFINIBAND */
+          || classid == 0x0502 /* PCI_CLASS_MEMORY_CXL */
           || baseclass == 0x06 /* PCI_BASE_CLASS_BRIDGE with non-PCI downstream. the core will drop the useless ones later */
 	  || baseclass == 0x12 /* Processing Accelerators */);
 }

src/3rdparty/hwloc/include/hwloc/rename.h

@@ -1,6 +1,6 @@
 /*
  * Copyright © 2009-2011 Cisco Systems, Inc.  All rights reserved.
- * Copyright © 2010-2021 Inria.  All rights reserved.
+ * Copyright © 2010-2022 Inria.  All rights reserved.
  * See COPYING in top-level directory.
  */
 
@@ -123,6 +123,9 @@ extern "C" {
 #define HWLOC_TOPOLOGY_FLAG_RESTRICT_TO_CPUBINDING HWLOC_NAME_CAPS(TOPOLOGY_FLAG_RESTRICT_TO_CPUBINDING)
 #define HWLOC_TOPOLOGY_FLAG_RESTRICT_TO_MEMBINDING HWLOC_NAME_CAPS(TOPOLOGY_FLAG_RESTRICT_TO_MEMBINDING)
 #define HWLOC_TOPOLOGY_FLAG_DONT_CHANGE_BINDING HWLOC_NAME_CAPS(TOPOLOGY_FLAG_DONT_CHANGE_BINDING)
+#define HWLOC_TOPOLOGY_FLAG_NO_DISTANCES HWLOC_NAME_CAPS(TOPOLOGY_FLAG_NO_DISTANCES)
+#define HWLOC_TOPOLOGY_FLAG_NO_MEMATTRS HWLOC_NAME_CAPS(TOPOLOGY_FLAG_NO_MEMATTRS)
+#define HWLOC_TOPOLOGY_FLAG_NO_CPUKINDS HWLOC_NAME_CAPS(TOPOLOGY_FLAG_NO_CPUKINDS)
 
 #define hwloc_topology_set_pid HWLOC_NAME(topology_set_pid)
 #define hwloc_topology_set_synthetic HWLOC_NAME(topology_set_synthetic)
@@ -381,6 +384,11 @@ extern "C" {
 #define HWLOC_MEMATTR_ID_LOCALITY HWLOC_NAME_CAPS(MEMATTR_ID_LOCALITY)
 #define HWLOC_MEMATTR_ID_BANDWIDTH HWLOC_NAME_CAPS(MEMATTR_ID_BANDWIDTH)
 #define HWLOC_MEMATTR_ID_LATENCY HWLOC_NAME_CAPS(MEMATTR_ID_LATENCY)
+#define HWLOC_MEMATTR_ID_READ_BANDWIDTH HWLOC_NAME_CAPS(MEMATTR_ID_READ_BANDWIDTH)
+#define HWLOC_MEMATTR_ID_WRITE_BANDWIDTH HWLOC_NAME_CAPS(MEMATTR_ID_WRITE_BANDWIDTH)
+#define HWLOC_MEMATTR_ID_READ_LATENCY HWLOC_NAME_CAPS(MEMATTR_ID_READ_LATENCY)
+#define HWLOC_MEMATTR_ID_WRITE_LATENCY HWLOC_NAME_CAPS(MEMATTR_ID_WRITE_LATENCY)
+#define HWLOC_MEMATTR_ID_MAX HWLOC_NAME_CAPS(MEMATTR_ID_MAX)
 
 #define hwloc_memattr_id_t HWLOC_NAME(memattr_id_t)
 #define hwloc_memattr_get_by_name HWLOC_NAME(memattr_get_by_name)
@@ -862,6 +870,7 @@ extern "C" {
 #define hwloc_internal_memattrs_destroy HWLOC_NAME(internal_memattrs_destroy)
 #define hwloc_internal_memattrs_need_refresh HWLOC_NAME(internal_memattrs_need_refresh)
 #define hwloc_internal_memattrs_refresh HWLOC_NAME(internal_memattrs_refresh)
+#define hwloc_internal_memattrs_guess_memory_tiers HWLOC_NAME(internal_memattrs_guess_memory_tiers)
 
 #define hwloc_internal_cpukind_s HWLOC_NAME(internal_cpukind_s)
 #define hwloc_internal_cpukinds_init HWLOC_NAME(internal_cpukinds_init)

src/3rdparty/hwloc/include/private/autogen/config.h

@@ -1,6 +1,6 @@
 /*
  * Copyright © 2009, 2011, 2012 CNRS.  All rights reserved.
- * Copyright © 2009-2020 Inria.  All rights reserved.
+ * Copyright © 2009-2021 Inria.  All rights reserved.
  * Copyright © 2009, 2011, 2012, 2015 Université Bordeaux.  All rights reserved.
  * Copyright © 2009-2020 Cisco Systems, Inc.  All rights reserved.
  * $COPYRIGHT$
@@ -290,10 +290,6 @@
 /* Define to '1' if sysctlbyname is present and usable */
 /* #undef HAVE_SYSCTLBYNAME */
 
-/* Define to 1 if the system has the type
-   `SYSTEM_LOGICAL_PROCESSOR_INFORMATION'. */
-#define HAVE_SYSTEM_LOGICAL_PROCESSOR_INFORMATION 1
-
 /* Define to 1 if the system has the type
    `SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX'. */
 #define HAVE_SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX 1

src/3rdparty/hwloc/include/private/misc.h

@@ -504,7 +504,7 @@ hwloc__obj_type_is_icache(hwloc_obj_type_t type)
   }                                    \
 } while(0)
 #else /* HAVE_USELOCALE */
-#if __HWLOC_HAVE_ATTRIBUTE_UNUSED
+#if HWLOC_HAVE_ATTRIBUTE_UNUSED
 #define hwloc_localeswitch_declare int __dummy_nolocale __hwloc_attribute_unused
 #define hwloc_localeswitch_init()
 #else

src/3rdparty/hwloc/include/private/private.h

@@ -1,6 +1,6 @@
 /*
  * Copyright © 2009      CNRS
- * Copyright © 2009-2021 Inria.  All rights reserved.
+ * Copyright © 2009-2022 Inria.  All rights reserved.
  * Copyright © 2009-2012, 2020 Université Bordeaux
  * Copyright © 2009-2011 Cisco Systems, Inc.  All rights reserved.
  *
@@ -259,6 +259,7 @@ struct hwloc_topology {
     unsigned bus_first, bus_last;
     hwloc_bitmap_t cpuset;
   } * pci_forced_locality;
+  hwloc_uint64_t pci_locality_quirks;
 
   /* component blacklisting */
   unsigned nr_blacklisted_components;
@@ -419,6 +420,7 @@ extern void hwloc_internal_memattrs_need_refresh(hwloc_topology_t topology);
 extern void hwloc_internal_memattrs_refresh(hwloc_topology_t topology);
 extern int hwloc_internal_memattrs_dup(hwloc_topology_t new, hwloc_topology_t old);
 extern int hwloc_internal_memattr_set_value(hwloc_topology_t topology, hwloc_memattr_id_t id, hwloc_obj_type_t target_type, hwloc_uint64_t target_gp_index, unsigned target_os_index, struct hwloc_internal_location_s *initiator, hwloc_uint64_t value);
+extern int hwloc_internal_memattrs_guess_memory_tiers(hwloc_topology_t topology);
 
 extern void hwloc_internal_cpukinds_init(hwloc_topology_t topology);
 extern int hwloc_internal_cpukinds_rank(hwloc_topology_t topology);
@@ -480,6 +482,7 @@ extern char * hwloc_progname(struct hwloc_topology *topology);
 #define HWLOC_GROUP_KIND_AIX_SDL_UNKNOWN		210	/* subkind is SDL level */
 #define HWLOC_GROUP_KIND_WINDOWS_PROCESSOR_GROUP	220	/* no subkind */
 #define HWLOC_GROUP_KIND_WINDOWS_RELATIONSHIP_UNKNOWN	221	/* no subkind */
+#define HWLOC_GROUP_KIND_LINUX_CLUSTER                  222     /* no subkind */
 /* distance groups */
 #define HWLOC_GROUP_KIND_DISTANCE			900	/* subkind is round of adding these groups during distance based grouping */
 /* finally, hwloc-specific groups required to insert something else, should disappear as soon as possible */

src/3rdparty/hwloc/include/private/windows.h

@@ -0,0 +1,30 @@
+/*
+ * Copyright © 2009 Université Bordeaux
+ * Copyright © 2020-2022 Inria.  All rights reserved.
+ *
+ * See COPYING in top-level directory.
+ */
+
+#ifndef HWLOC_PRIVATE_WINDOWS_H
+#define HWLOC_PRIVATE_WINDOWS_H
+
+#ifndef _ANONYMOUS_UNION
+#ifdef __GNUC__
+#define _ANONYMOUS_UNION __extension__
+#else
+#define _ANONYMOUS_UNION
+#endif /* __GNUC__ */
+#endif /* _ANONYMOUS_UNION */
+
+#ifndef _ANONYMOUS_STRUCT
+#ifdef __GNUC__
+#define _ANONYMOUS_STRUCT __extension__
+#else
+#define _ANONYMOUS_STRUCT
+#endif /* __GNUC__ */
+#endif /* _ANONYMOUS_STRUCT */
+
+#define DUMMYUNIONNAME
+#define DUMMYSTRUCTNAME
+
+#endif /* HWLOC_PRIVATE_WINDOWS_H */

src/3rdparty/hwloc/src/components.c

@@ -1,5 +1,5 @@
 /*
- * Copyright © 2009-2021 Inria.  All rights reserved.
+ * Copyright © 2009-2022 Inria.  All rights reserved.
  * Copyright © 2012 Université Bordeaux
  * See COPYING in top-level directory.
  */
@@ -386,7 +386,7 @@ hwloc_disc_component_register(struct hwloc_disc_component *component,
 				   |HWLOC_DISC_PHASE_MISC
 				   |HWLOC_DISC_PHASE_ANNOTATE
 				   |HWLOC_DISC_PHASE_TWEAK))) {
-    if (hwloc_hide_errors() < 2)
+    if (HWLOC_SHOW_CRITICAL_ERRORS())
       fprintf(stderr, "hwloc: Cannot register discovery component `%s' with invalid phases 0x%x\n",
               component->name, component->phases);
     return -1;
@@ -476,7 +476,7 @@ hwloc_components_init(void)
   /* hwloc_static_components is created by configure in static-components.h */
   for(i=0; NULL != hwloc_static_components[i]; i++) {
     if (hwloc_static_components[i]->flags) {
-      if (hwloc_hide_errors() < 2)
+      if (HWLOC_SHOW_CRITICAL_ERRORS())
         fprintf(stderr, "hwloc: Ignoring static component with invalid flags %lx\n",
                 hwloc_static_components[i]->flags);
       continue;
@@ -505,7 +505,7 @@ hwloc_components_init(void)
 #ifdef HWLOC_HAVE_PLUGINS
   for(desc = hwloc_plugins; NULL != desc; desc = desc->next) {
     if (desc->component->flags) {
-      if (hwloc_hide_errors() < 2)
+      if (HWLOC_SHOW_CRITICAL_ERRORS())
         fprintf(stderr, "hwloc: Ignoring plugin `%s' component with invalid flags %lx\n",
                 desc->name, desc->component->flags);
       continue;
@@ -738,7 +738,7 @@ hwloc_disc_component_try_enable(struct hwloc_topology *topology,
   backend = comp->instantiate(topology, comp, topology->backend_excluded_phases | blacklisted_phases,
 			      NULL, NULL, NULL);
   if (!backend) {
-    if (hwloc_components_verbose || (envvar_forced && hwloc_hide_errors() < 2))
+    if (hwloc_components_verbose || (envvar_forced && HWLOC_SHOW_CRITICAL_ERRORS()))
       fprintf(stderr, "hwloc: Failed to instantiate discovery component `%s'\n", comp->name);
     return -1;
   }
@@ -835,7 +835,7 @@ hwloc_disc_components_enable_others(struct hwloc_topology *topology)
 	  if (comp->phases & ~blacklisted_phases)
 	    hwloc_disc_component_try_enable(topology, comp, 1 /* envvar forced */, blacklisted_phases);
 	} else {
-          if (hwloc_hide_errors() < 2)
+          if (HWLOC_SHOW_CRITICAL_ERRORS())
             fprintf(stderr, "hwloc: Cannot find discovery component `%s'\n", name);
 	}
 
@@ -967,7 +967,7 @@ hwloc_backend_enable(struct hwloc_backend *backend)
 
   /* check backend flags */
   if (backend->flags) {
-    if (hwloc_hide_errors() < 2)
+    if (HWLOC_SHOW_CRITICAL_ERRORS())
       fprintf(stderr, "hwloc: Cannot enable discovery component `%s' phases 0x%x with unknown flags %lx\n",
               backend->component->name, backend->component->phases, backend->flags);
     return -1;

src/3rdparty/hwloc/src/cpukinds.c

@@ -1,5 +1,5 @@
 /*
- * Copyright © 2020-2021 Inria.  All rights reserved.
+ * Copyright © 2020-2022 Inria.  All rights reserved.
  * See COPYING in top-level directory.
  */
 
@@ -42,6 +42,9 @@ hwloc_internal_cpukinds_dup(hwloc_topology_t new, hwloc_topology_t old)
   struct hwloc_internal_cpukind_s *kinds;
   unsigned i;
 
+  if (!old->nr_cpukinds)
+    return 0;
+
   kinds = hwloc_tma_malloc(tma, old->nr_cpukinds * sizeof(*kinds));
   if (!kinds)
     return -1;
@@ -445,7 +448,9 @@ static int hwloc__cpukinds_compare_ranking_values(const void *_a, const void *_b
 {
   const struct hwloc_internal_cpukind_s *a = _a;
   const struct hwloc_internal_cpukind_s *b = _b;
-  return a->ranking_value - b->ranking_value;
+  uint64_t arv = a->ranking_value;
+  uint64_t brv = b->ranking_value;
+  return arv < brv ? -1 : arv > brv ? 1 : 0;
 }
 
 /* this function requires ranking values to be unique */
@@ -499,7 +504,7 @@ hwloc_internal_cpukinds_rank(struct hwloc_topology *topology)
       heuristics = HWLOC_CPUKINDS_RANKING_FORCED_EFFICIENCY;
     else if (!strcmp(env, "no_forced_efficiency"))
       heuristics = HWLOC_CPUKINDS_RANKING_NO_FORCED_EFFICIENCY;
-    else if (hwloc_hide_errors() < 2)
+    else if (HWLOC_SHOW_CRITICAL_ERRORS())
       fprintf(stderr, "hwloc: Failed to recognize HWLOC_CPUKINDS_RANKING value %s\n", env);
   }
 

src/3rdparty/hwloc/src/diff.c

@@ -1,5 +1,5 @@
 /*
- * Copyright © 2013-2020 Inria.  All rights reserved.
+ * Copyright © 2013-2022 Inria.  All rights reserved.
  * See COPYING in top-level directory.
  */
 
@@ -218,7 +218,7 @@ hwloc_diff_trees(hwloc_topology_t topo1, hwloc_obj_t obj1,
 		struct hwloc_info_s *info1 = &obj1->infos[i], *info2 = &obj2->infos[i];
 		if (strcmp(info1->name, info2->name))
 			goto out_too_complex;
-		if (strcmp(obj1->infos[i].value, obj2->infos[i].value)) {
+		if (strcmp(info1->value, info2->value)) {
 			err = hwloc_append_diff_obj_attr_string(obj1,
 								HWLOC_TOPOLOGY_DIFF_OBJ_ATTR_INFO,
 								info1->name,

src/3rdparty/hwloc/src/distances.c

@@ -1,5 +1,5 @@
 /*
- * Copyright © 2010-2021 Inria.  All rights reserved.
+ * Copyright © 2010-2022 Inria.  All rights reserved.
  * Copyright © 2011-2012 Université Bordeaux
  * Copyright © 2011 Cisco Systems, Inc.  All rights reserved.
  * See COPYING in top-level directory.
@@ -860,7 +860,7 @@ struct hwloc_distances_container_s {
   struct hwloc_distances_s distances;
 };
 
-#define HWLOC_DISTANCES_CONTAINER_OFFSET ((char*)&((struct hwloc_distances_container_s*)NULL)->distances - (char*)NULL)
+#define HWLOC_DISTANCES_CONTAINER_OFFSET ((uintptr_t)(&((struct hwloc_distances_container_s*)NULL)->distances) - (uintptr_t)NULL)
 #define HWLOC_DISTANCES_CONTAINER(_d) (struct hwloc_distances_container_s *) ( ((char*)_d) - HWLOC_DISTANCES_CONTAINER_OFFSET )
 
 static struct hwloc_internal_distances_s *

src/3rdparty/hwloc/src/memattrs.c

@@ -1,11 +1,12 @@
 /*
- * Copyright © 2020 Inria.  All rights reserved.
+ * Copyright © 2020-2022 Inria.  All rights reserved.
  * See COPYING in top-level directory.
  */
 
 #include "private/autogen/config.h"
 #include "hwloc.h"
 #include "private/private.h"
+#include "private/debug.h"
 
 
 /*****************************
@@ -49,36 +50,51 @@ hwloc__setup_memattr(struct hwloc_internal_memattr_s *imattr,
 void
 hwloc_internal_memattrs_prepare(struct hwloc_topology *topology)
 {
-#define NR_DEFAULT_MEMATTRS 4
-  topology->memattrs = malloc(NR_DEFAULT_MEMATTRS * sizeof(*topology->memattrs));
+  topology->memattrs = malloc(HWLOC_MEMATTR_ID_MAX * sizeof(*topology->memattrs));
   if (!topology->memattrs)
     return;
 
-  assert(HWLOC_MEMATTR_ID_CAPACITY < NR_DEFAULT_MEMATTRS);
   hwloc__setup_memattr(&topology->memattrs[HWLOC_MEMATTR_ID_CAPACITY],
                        (char *) "Capacity",
                        HWLOC_MEMATTR_FLAG_HIGHER_FIRST,
                        HWLOC_IMATTR_FLAG_STATIC_NAME|HWLOC_IMATTR_FLAG_CONVENIENCE);
 
-  assert(HWLOC_MEMATTR_ID_LOCALITY < NR_DEFAULT_MEMATTRS);
   hwloc__setup_memattr(&topology->memattrs[HWLOC_MEMATTR_ID_LOCALITY],
                        (char *) "Locality",
                        HWLOC_MEMATTR_FLAG_LOWER_FIRST,
                        HWLOC_IMATTR_FLAG_STATIC_NAME|HWLOC_IMATTR_FLAG_CONVENIENCE);
 
-  assert(HWLOC_MEMATTR_ID_BANDWIDTH < NR_DEFAULT_MEMATTRS);
   hwloc__setup_memattr(&topology->memattrs[HWLOC_MEMATTR_ID_BANDWIDTH],
                        (char *) "Bandwidth",
                        HWLOC_MEMATTR_FLAG_HIGHER_FIRST|HWLOC_MEMATTR_FLAG_NEED_INITIATOR,
                        HWLOC_IMATTR_FLAG_STATIC_NAME);
 
-  assert(HWLOC_MEMATTR_ID_LATENCY < NR_DEFAULT_MEMATTRS);
+  hwloc__setup_memattr(&topology->memattrs[HWLOC_MEMATTR_ID_READ_BANDWIDTH],
+                       (char *) "ReadBandwidth",
+                       HWLOC_MEMATTR_FLAG_HIGHER_FIRST|HWLOC_MEMATTR_FLAG_NEED_INITIATOR,
+                       HWLOC_IMATTR_FLAG_STATIC_NAME);
+
+  hwloc__setup_memattr(&topology->memattrs[HWLOC_MEMATTR_ID_WRITE_BANDWIDTH],
+                       (char *) "WriteBandwidth",
+                       HWLOC_MEMATTR_FLAG_HIGHER_FIRST|HWLOC_MEMATTR_FLAG_NEED_INITIATOR,
+                       HWLOC_IMATTR_FLAG_STATIC_NAME);
+
   hwloc__setup_memattr(&topology->memattrs[HWLOC_MEMATTR_ID_LATENCY],
                        (char *) "Latency",
                        HWLOC_MEMATTR_FLAG_LOWER_FIRST|HWLOC_MEMATTR_FLAG_NEED_INITIATOR,
                        HWLOC_IMATTR_FLAG_STATIC_NAME);
 
-  topology->nr_memattrs = NR_DEFAULT_MEMATTRS;
+  hwloc__setup_memattr(&topology->memattrs[HWLOC_MEMATTR_ID_READ_LATENCY],
+                       (char *) "ReadLatency",
+                       HWLOC_MEMATTR_FLAG_LOWER_FIRST|HWLOC_MEMATTR_FLAG_NEED_INITIATOR,
+                       HWLOC_IMATTR_FLAG_STATIC_NAME);
+
+  hwloc__setup_memattr(&topology->memattrs[HWLOC_MEMATTR_ID_WRITE_LATENCY],
+                       (char *) "WriteLatency",
+                       HWLOC_MEMATTR_FLAG_LOWER_FIRST|HWLOC_MEMATTR_FLAG_NEED_INITIATOR,
+                       HWLOC_IMATTR_FLAG_STATIC_NAME);
+
+  topology->nr_memattrs = HWLOC_MEMATTR_ID_MAX;
 }
 
 static void
@@ -127,6 +143,8 @@ hwloc_internal_memattrs_dup(struct hwloc_topology *new, struct hwloc_topology *o
   struct hwloc_internal_memattr_s *imattrs;
   hwloc_memattr_id_t id;
 
+  /* old->nr_memattrs is always > 0 thanks to default memattrs */
+
   imattrs = hwloc_tma_malloc(tma, old->nr_memattrs * sizeof(*imattrs));
   if (!imattrs)
     return -1;
@@ -1195,3 +1213,214 @@ hwloc_get_local_numanode_objs(hwloc_topology_t topology,
   *nrp = i;
   return 0;
 }
+
+
+/**************************************
+ * Using memattrs to identify HBM/DRAM
+ */
+
+struct hwloc_memory_tier_s {
+  hwloc_obj_t node;
+  uint64_t local_bw;
+  enum hwloc_memory_tier_type_e {
+    /* warning the order is important for guess_memory_tiers() after qsort() */
+    HWLOC_MEMORY_TIER_UNKNOWN,
+    HWLOC_MEMORY_TIER_DRAM,
+    HWLOC_MEMORY_TIER_HBM,
+    HWLOC_MEMORY_TIER_SPM, /* Specific-Purpose Memory is usually HBM, we'll use BW to confirm */
+    HWLOC_MEMORY_TIER_NVM,
+    HWLOC_MEMORY_TIER_GPU,
+  } type;
+};
+
+static int compare_tiers(const void *_a, const void *_b)
+{
+  const struct hwloc_memory_tier_s *a = _a, *b = _b;
+  /* sort by type of tier first */
+  if (a->type != b->type)
+    return a->type - b->type;
+  /* then by bandwidth */
+  if (a->local_bw > b->local_bw)
+    return -1;
+  else if (a->local_bw < b->local_bw)
+    return 1;
+  return 0;
+}
+
+int
+hwloc_internal_memattrs_guess_memory_tiers(hwloc_topology_t topology)
+{
+  struct hwloc_internal_memattr_s *imattr;
+  struct hwloc_memory_tier_s *tiers;
+  unsigned i, j, n;
+  const char *env;
+  int spm_is_hbm = -1; /* -1 will guess from BW, 0 no, 1 forced */
+  int mark_dram = 1;
+  unsigned first_spm, first_nvm;
+  hwloc_uint64_t max_unknown_bw, min_spm_bw;
+
+  env = getenv("HWLOC_MEMTIERS_GUESS");
+  if (env) {
+    if (!strcmp(env, "none")) {
+      return 0;
+    } else if (!strcmp(env, "default")) {
+      /* nothing */
+    } else if (!strcmp(env, "spm_is_hbm")) {
+      hwloc_debug("Assuming SPM-tier is HBM, ignore bandwidth\n");
+      spm_is_hbm = 1;
+    } else if (HWLOC_SHOW_CRITICAL_ERRORS()) {
+      fprintf(stderr, "hwloc: Failed to recognize HWLOC_MEMTIERS_GUESS value %s\n", env);
+    }
+  }
+
+  imattr = &topology->memattrs[HWLOC_MEMATTR_ID_BANDWIDTH];
+
+  if (!(imattr->iflags & HWLOC_IMATTR_FLAG_CACHE_VALID))
+    hwloc__imattr_refresh(topology, imattr);
+
+  n = hwloc_get_nbobjs_by_depth(topology, HWLOC_TYPE_DEPTH_NUMANODE);
+  assert(n);
+
+  tiers = malloc(n * sizeof(*tiers));
+  if (!tiers)
+    return -1;
+
+  for(i=0; i<n; i++) {
+    hwloc_obj_t node;
+    const char *daxtype;
+    struct hwloc_internal_location_s iloc;
+    struct hwloc_internal_memattr_target_s *imtg = NULL;
+    struct hwloc_internal_memattr_initiator_s *imi;
+
+    node = hwloc_get_obj_by_depth(topology, HWLOC_TYPE_DEPTH_NUMANODE, i);
+    assert(node);
+    tiers[i].node = node;
+
+    /* defaults */
+    tiers[i].type = HWLOC_MEMORY_TIER_UNKNOWN;
+    tiers[i].local_bw = 0; /* unknown */
+
+    daxtype = hwloc_obj_get_info_by_name(node, "DAXType");
+    /* mark NVM, SPM and GPU nodes */
+    if (daxtype && !strcmp(daxtype, "NVM"))
+      tiers[i].type = HWLOC_MEMORY_TIER_NVM;
+    if (daxtype && !strcmp(daxtype, "SPM"))
+      tiers[i].type = HWLOC_MEMORY_TIER_SPM;
+    if (node->subtype && !strcmp(node->subtype, "GPUMemory"))
+      tiers[i].type = HWLOC_MEMORY_TIER_GPU;
+
+    if (spm_is_hbm == -1) {
+      for(j=0; j<imattr->nr_targets; j++)
+        if (imattr->targets[j].obj == node) {
+          imtg = &imattr->targets[j];
+          break;
+        }
+      if (imtg && !hwloc_bitmap_iszero(node->cpuset)) {
+        iloc.type = HWLOC_LOCATION_TYPE_CPUSET;
+        iloc.location.cpuset = node->cpuset;
+        imi = hwloc__memattr_target_get_initiator(imtg, &iloc, 0);
+        if (imi)
+          tiers[i].local_bw = imi->value;
+      }
+    }
+  }
+
+  /* sort tiers */
+  qsort(tiers, n, sizeof(*tiers), compare_tiers);
+  hwloc_debug("Sorting memory tiers...\n");
+  for(i=0; i<n; i++)
+    hwloc_debug("  tier %u = node L#%u P#%u with tier type %d and local BW #%llu\n",
+                i,
+                tiers[i].node->logical_index, tiers[i].node->os_index,
+                tiers[i].type, (unsigned long long) tiers[i].local_bw);
+
+  /* now we have UNKNOWN tiers (sorted by BW), then SPM tiers (sorted by BW), then NVM, then GPU */
+
+  /* iterate over UNKNOWN tiers, and find their BW */
+  for(i=0; i<n; i++) {
+    if (tiers[i].type > HWLOC_MEMORY_TIER_UNKNOWN)
+      break;
+  }
+  first_spm = i;
+  /* get max BW from first */
+  if (first_spm > 0)
+    max_unknown_bw = tiers[0].local_bw;
+  else
+    max_unknown_bw = 0;
+
+  /* there are no DRAM or HBM tiers yet */
+
+  /* iterate over SPM tiers, and find their BW */
+  for(i=first_spm; i<n; i++) {
+    if (tiers[i].type > HWLOC_MEMORY_TIER_SPM)
+      break;
+  }
+  first_nvm = i;
+  /* get min BW from last */
+  if (first_nvm > first_spm)
+    min_spm_bw = tiers[first_nvm-1].local_bw;
+  else
+    min_spm_bw = 0;
+
+  /* FIXME: if there's more than 10% between some sets of nodes inside a tier, split it? */
+  /* FIXME: if there are cpuset-intersecting nodes in same tier, abort? */
+
+  if (spm_is_hbm == -1) {
+    /* if we have BW for all SPM and UNKNOWN
+     * and all SPM BW are 2x superior to all UNKNOWN BW
+     */
+    hwloc_debug("UNKNOWN-memory-tier max bandwidth %llu\n", (unsigned long long) max_unknown_bw);
+    hwloc_debug("SPM-memory-tier min bandwidth %llu\n", (unsigned long long) min_spm_bw);
+    if (max_unknown_bw > 0 && min_spm_bw > 0 && max_unknown_bw*2 < min_spm_bw) {
+      hwloc_debug("assuming SPM means HBM and !SPM means DRAM since bandwidths are very different\n");
+      spm_is_hbm = 1;
+    } else {
+      hwloc_debug("cannot assume SPM means HBM\n");
+      spm_is_hbm = 0;
+    }
+  }
+
+  if (spm_is_hbm) {
+    for(i=0; i<first_spm; i++)
+      tiers[i].type = HWLOC_MEMORY_TIER_DRAM;
+    for(i=first_spm; i<first_nvm; i++)
+      tiers[i].type = HWLOC_MEMORY_TIER_HBM;
+  }
+
+  if (first_spm == n)
+    mark_dram = 0;
+
+    /* now apply subtypes */
+  for(i=0; i<n; i++) {
+    const char *type = NULL;
+    if (tiers[i].node->subtype) /* don't overwrite the existing subtype */
+      continue;
+    switch (tiers[i].type) {
+    case HWLOC_MEMORY_TIER_DRAM:
+      if (mark_dram)
+        type = "DRAM";
+      break;
+    case HWLOC_MEMORY_TIER_HBM:
+      type = "HBM";
+      break;
+    case HWLOC_MEMORY_TIER_SPM:
+      type = "SPM";
+      break;
+    case HWLOC_MEMORY_TIER_NVM:
+      type = "NVM";
+      break;
+    default:
+      /* GPU memory is already marked with subtype="GPUMemory",
+       * UNKNOWN doesn't deserve any subtype
+       */
+      break;
+    }
+    if (type) {
+      hwloc_debug("Marking node L#%u P#%u as %s\n", tiers[i].node->logical_index, tiers[i].node->os_index, type);
+      tiers[i].node->subtype = strdup(type);
+    }
+  }
+
+  free(tiers);
+  return 0;
+}

src/3rdparty/hwloc/src/pci-common.c

@@ -1,5 +1,5 @@
 /*
- * Copyright © 2009-2021 Inria.  All rights reserved.
+ * Copyright © 2009-2022 Inria.  All rights reserved.
  * See COPYING in top-level directory.
  */
 
@@ -119,6 +119,13 @@ hwloc_pci_discovery_init(struct hwloc_topology *topology)
   topology->pci_forced_locality = NULL;
 
   topology->first_pci_locality = topology->last_pci_locality = NULL;
+
+#define HWLOC_PCI_LOCALITY_QUIRK_CRAY_EX235A (1ULL<<0)
+#define HWLOC_PCI_LOCALITY_QUIRK_FAKE (1ULL<<62)
+  topology->pci_locality_quirks = (uint64_t) -1;
+  /* -1 is unknown, 0 is disabled, >0 is bitmask of enabled quirks.
+   * bit 63 should remain unused so that -1 is unaccessible as a bitmask.
+   */
 }
 
 void
@@ -146,7 +153,7 @@ hwloc_pci_discovery_prepare(struct hwloc_topology *topology)
 	  }
 	  free(buffer);
 	} else {
-          if (hwloc_hide_errors() < 2)
+          if (HWLOC_SHOW_CRITICAL_ERRORS())
             fprintf(stderr, "hwloc/pci: Ignoring HWLOC_PCI_LOCALITY file `%s' too large (%lu bytes)\n",
                     env, (unsigned long) st.st_size);
 	}
@@ -333,7 +340,7 @@ hwloc_pci_add_object(struct hwloc_obj *parent, struct hwloc_obj **parent_io_firs
     }
     case HWLOC_PCI_BUSID_EQUAL: {
       static int reported = 0;
-      if (!reported && hwloc_hide_errors() < 2) {
+      if (!reported && HWLOC_SHOW_CRITICAL_ERRORS()) {
         fprintf(stderr, "*********************************************************\n");
         fprintf(stderr, "* hwloc %s received invalid PCI information.\n", HWLOC_VERSION);
         fprintf(stderr, "*\n");
@@ -442,13 +449,90 @@ hwloc_pcidisc_add_hostbridges(struct hwloc_topology *topology,
   return new;
 }
 
-static struct hwloc_obj *
-hwloc_pci_fixup_busid_parent(struct hwloc_topology *topology __hwloc_attribute_unused,
-			     struct hwloc_pcidev_attr_s *busid __hwloc_attribute_unused,
-			     struct hwloc_obj *parent __hwloc_attribute_unused)
+/* return 1 if a quirk was applied */
+static int
+hwloc__pci_find_busid_parent_quirk(struct hwloc_topology *topology,
+                                   struct hwloc_pcidev_attr_s *busid,
+                                   hwloc_cpuset_t cpuset)
 {
-  /* no quirk for now */
-  return parent;
+  if (topology->pci_locality_quirks == (uint64_t)-1 /* unknown */) {
+    const char *dmi_board_name, *env;
+
+    /* first invokation, detect which quirks are needed */
+    topology->pci_locality_quirks = 0; /* no quirk yet */
+
+    dmi_board_name = hwloc_obj_get_info_by_name(hwloc_get_root_obj(topology), "DMIBoardName");
+    if (dmi_board_name && !strcmp(dmi_board_name, "HPE CRAY EX235A")) {
+      hwloc_debug("enabling for PCI locality quirk for HPE Cray EX235A\n");
+      topology->pci_locality_quirks |= HWLOC_PCI_LOCALITY_QUIRK_CRAY_EX235A;
+    }
+
+    env = getenv("HWLOC_PCI_LOCALITY_QUIRK_FAKE");
+    if (env && atoi(env)) {
+      hwloc_debug("enabling for PCI locality fake quirk (attaching everything to last PU)\n");
+      topology->pci_locality_quirks |= HWLOC_PCI_LOCALITY_QUIRK_FAKE;
+    }
+  }
+
+  if (topology->pci_locality_quirks & HWLOC_PCI_LOCALITY_QUIRK_FAKE) {
+    unsigned last = hwloc_bitmap_last(hwloc_topology_get_topology_cpuset(topology));
+    hwloc_bitmap_set(cpuset, last);
+    return 1;
+  }
+
+  if (topology->pci_locality_quirks & HWLOC_PCI_LOCALITY_QUIRK_CRAY_EX235A) {
+    /* AMD Trento has xGMI ports connected to individual CCDs (8 cores + L3)
+     * instead of NUMA nodes (pairs of CCDs within Trento) as is usual in AMD EPYC CPUs.
+     * This is not described by the ACPI tables, hence we need to manually hardwire
+     * the xGMI locality for the (currently single) server that currently uses that CPU.
+     * It's not clear if ACPI tables can/will ever be fixed (would require one initiator
+     * proximity domain per CCD), or if Linux can/will work around the issue.
+     */
+    if (busid->domain == 0) {
+      if (busid->bus >= 0xd0 && busid->bus <= 0xd1) {
+        hwloc_bitmap_set_range(cpuset, 0, 7);
+        hwloc_bitmap_set_range(cpuset, 64, 71);
+        return 1;
+      }
+      if (busid->bus >= 0xd4 && busid->bus <= 0xd6) {
+        hwloc_bitmap_set_range(cpuset, 8, 15);
+        hwloc_bitmap_set_range(cpuset, 72, 79);
+        return 1;
+      }
+      if (busid->bus >= 0xc8 && busid->bus <= 0xc9) {
+        hwloc_bitmap_set_range(cpuset, 16, 23);
+        hwloc_bitmap_set_range(cpuset, 80, 87);
+        return 1;
+      }
+      if (busid->bus >= 0xcc && busid->bus <= 0xce) {
+        hwloc_bitmap_set_range(cpuset, 24, 31);
+        hwloc_bitmap_set_range(cpuset, 88, 95);
+        return 1;
+      }
+      if (busid->bus >= 0xd8 && busid->bus <= 0xd9) {
+        hwloc_bitmap_set_range(cpuset, 32, 39);
+        hwloc_bitmap_set_range(cpuset, 96, 103);
+        return 1;
+      }
+      if (busid->bus >= 0xdc && busid->bus <= 0xde) {
+        hwloc_bitmap_set_range(cpuset, 40, 47);
+        hwloc_bitmap_set_range(cpuset, 104, 111);
+        return 1;
+      }
+      if (busid->bus >= 0xc0 && busid->bus <= 0xc1) {
+        hwloc_bitmap_set_range(cpuset, 48, 55);
+        hwloc_bitmap_set_range(cpuset, 112, 119);
+        return 1;
+      }
+      if (busid->bus >= 0xc4 && busid->bus <= 0xc6) {
+        hwloc_bitmap_set_range(cpuset, 56, 63);
+        hwloc_bitmap_set_range(cpuset, 120, 127);
+        return 1;
+      }
+    }
+  }
+
+  return 0;
 }
 
 static struct hwloc_obj *
@@ -457,7 +541,7 @@ hwloc__pci_find_busid_parent(struct hwloc_topology *topology, struct hwloc_pcide
   hwloc_bitmap_t cpuset = hwloc_bitmap_alloc();
   hwloc_obj_t parent;
   int forced = 0;
-  int noquirks = 0;
+  int noquirks = 0, got_quirked = 0;
   unsigned i;
   int err;
 
@@ -490,7 +574,7 @@ hwloc__pci_find_busid_parent(struct hwloc_topology *topology, struct hwloc_pcide
     if (env) {
       static int reported = 0;
       if (!topology->pci_has_forced_locality && !reported) {
-        if (!hwloc_hide_errors())
+        if (HWLOC_SHOW_ALL_ERRORS())
           fprintf(stderr, "hwloc/pci: Environment variable %s is deprecated, please use HWLOC_PCI_LOCALITY instead.\n", env);
 	reported = 1;
       }
@@ -505,7 +589,13 @@ hwloc__pci_find_busid_parent(struct hwloc_topology *topology, struct hwloc_pcide
     }
   }
 
-  if (!forced) {
+  if (!forced && !noquirks && topology->pci_locality_quirks /* either quirks are unknown yet, or some are enabled */) {
+    err = hwloc__pci_find_busid_parent_quirk(topology, busid, cpuset);
+    if (err > 0)
+      got_quirked = 1;
+  }
+
+  if (!forced && !got_quirked) {
     /* get the cpuset by asking the backend that provides the relevant hook, if any. */
     struct hwloc_backend *backend = topology->get_pci_busid_cpuset_backend;
     if (backend)
@@ -520,11 +610,7 @@ hwloc__pci_find_busid_parent(struct hwloc_topology *topology, struct hwloc_pcide
   hwloc_debug_bitmap("  will attach PCI bus to cpuset %s\n", cpuset);
 
   parent = hwloc_find_insert_io_parent_by_complete_cpuset(topology, cpuset);
-  if (parent) {
-    if (!noquirks)
-      /* We found a valid parent. Check that the OS didn't report invalid locality */
-      parent = hwloc_pci_fixup_busid_parent(topology, busid, parent);
-  } else {
+  if (!parent) {
     /* Fallback to root */
     parent = hwloc_get_root_obj(topology);
   }
@@ -805,18 +891,28 @@ hwloc_pcidisc_find_linkspeed(const unsigned char *config,
   memcpy(&linksta, &config[offset + HWLOC_PCI_EXP_LNKSTA], 4);
   speed = linksta & HWLOC_PCI_EXP_LNKSTA_SPEED; /* PCIe generation */
   width = (linksta & HWLOC_PCI_EXP_LNKSTA_WIDTH) >> 4; /* how many lanes */
-  /* PCIe Gen1 = 2.5GT/s signal-rate per lane with 8/10 encoding    = 0.25GB/s data-rate per lane
-   * PCIe Gen2 = 5  GT/s signal-rate per lane with 8/10 encoding    = 0.5 GB/s data-rate per lane
-   * PCIe Gen3 = 8  GT/s signal-rate per lane with 128/130 encoding = 1   GB/s data-rate per lane
-   * PCIe Gen4 = 16 GT/s signal-rate per lane with 128/130 encoding = 2   GB/s data-rate per lane
-   * PCIe Gen5 = 32 GT/s signal-rate per lane with 128/130 encoding = 4   GB/s data-rate per lane
+  /*
+   * These are single-direction bandwidths only.
+   *
+   * Gen1 used NRZ with 8/10 encoding.
+   * PCIe Gen1 = 2.5GT/s signal-rate per lane x 8/10    =  0.25GB/s data-rate per lane
+   * PCIe Gen2 = 5  GT/s signal-rate per lane x 8/10    =  0.5 GB/s data-rate per lane
+   * Gen3 switched to NRZ with 128/130 encoding.
+   * PCIe Gen3 = 8  GT/s signal-rate per lane x 128/130 =  1   GB/s data-rate per lane
+   * PCIe Gen4 = 16 GT/s signal-rate per lane x 128/130 =  2   GB/s data-rate per lane
+   * PCIe Gen5 = 32 GT/s signal-rate per lane x 128/130 =  4   GB/s data-rate per lane
+   * Gen6 switched to PAM with with 242/256 FLIT (242B payload protected by 8B CRC + 6B FEC).
+   * PCIe Gen6 = 64 GT/s signal-rate per lane x 242/256 =  8   GB/s data-rate per lane
+   * PCIe Gen7 = 128GT/s signal-rate per lane x 242/256 = 16   GB/s data-rate per lane
    */
 
   /* lanespeed in Gbit/s */
   if (speed <= 2)
     lanespeed = 2.5f * speed * 0.8f;
+  else if (speed <= 5)
+    lanespeed = 8.0f * (1<<(speed-3)) * 128/130;
   else
-    lanespeed = 8.0f * (1<<(speed-3)) * 128/130; /* assume Gen6 will be 64 GT/s and so on */
+    lanespeed = 8.0f * (1<<(speed-3)) * 242/256; /* assume Gen8 will be 256 GT/s and so on */
 
   /* linkspeed in GB/s */
   *linkspeed = lanespeed * width / 8;
@@ -943,6 +1039,7 @@ hwloc_pci_class_string(unsigned short class_id)
       switch (class_id) {
 	case 0x0500: return "RAM";
 	case 0x0501: return "Flash";
+        case 0x0502: return "CXLMem";
       }
       return "Memory";
     case 0x06:

src/3rdparty/hwloc/src/topology-synthetic.c

@@ -1,6 +1,6 @@
 /*
  * Copyright © 2009 CNRS
- * Copyright © 2009-2020 Inria.  All rights reserved.
+ * Copyright © 2009-2022 Inria.  All rights reserved.
  * Copyright © 2009-2010 Université Bordeaux
  * Copyright © 2009-2011 Cisco Systems, Inc.  All rights reserved.
  * See COPYING in top-level directory.
@@ -323,17 +323,29 @@ hwloc_synthetic_parse_memory_attr(const char *attr, const char **endp)
   hwloc_uint64_t size;
   size = strtoull(attr, (char **) &endptr, 0);
   if (!hwloc_strncasecmp(endptr, "TB", 2)) {
+    size *= 1000ULL*1000ULL*1000ULL*1000ULL;
+    endptr += 2;
+  } else if (!hwloc_strncasecmp(endptr, "TiB", 3)) {
     size <<= 40;
-    endptr += 2;
+    endptr += 3;
   } else if (!hwloc_strncasecmp(endptr, "GB", 2)) {
+    size *= 1000ULL*1000ULL*1000ULL;
+    endptr += 2;
+  } else if (!hwloc_strncasecmp(endptr, "GiB", 3)) {
     size <<= 30;
-    endptr += 2;
+    endptr += 3;
   } else if (!hwloc_strncasecmp(endptr, "MB", 2)) {
+    size *= 1000ULL*1000ULL;
+    endptr += 2;
+  } else if (!hwloc_strncasecmp(endptr, "MiB", 3)) {
     size <<= 20;
-    endptr += 2;
+    endptr += 3;
   } else if (!hwloc_strncasecmp(endptr, "kB", 2)) {
-    size <<= 10;
+    size *= 1000ULL;
     endptr += 2;
+  } else if (!hwloc_strncasecmp(endptr, "kiB", 3)) {
+    size <<= 10;
+    endptr += 3;
   }
   *endp = endptr;
   return size;
@@ -802,15 +814,15 @@ hwloc_backend_synthetic_init(struct hwloc_synthetic_backend_data_s *data,
     } else if (hwloc__obj_type_is_cache(type)) {
       if (!curlevel->attr.memorysize) {
 	if (1 == curlevel->attr.depth)
-	  /* 32Kb in L1 */
+	  /* 32KiB in L1 */
 	  curlevel->attr.memorysize = 32*1024;
 	else
-	  /* *4 at each level, starting from 1MB for L2, unified */
+	  /* *4 at each level, starting from 1MiB for L2, unified */
 	  curlevel->attr.memorysize = 256ULL*1024 << (2*curlevel->attr.depth);
       }
 
     } else if (type == HWLOC_OBJ_NUMANODE && !curlevel->attr.memorysize) {
-      /* 1GB in memory nodes. */
+      /* 1GiB in memory nodes. */
       curlevel->attr.memorysize = 1024*1024*1024;
     }
 

src/3rdparty/hwloc/src/topology-windows.c

@@ -1,6 +1,6 @@
 /*
  * Copyright © 2009 CNRS
- * Copyright © 2009-2021 Inria.  All rights reserved.
+ * Copyright © 2009-2022 Inria.  All rights reserved.
  * Copyright © 2009-2012, 2020 Université Bordeaux
  * Copyright © 2011 Cisco Systems, Inc.  All rights reserved.
  * See COPYING in top-level directory.
@@ -13,6 +13,7 @@
 #include "hwloc.h"
 #include "hwloc/windows.h"
 #include "private/private.h"
+#include "private/windows.h" /* must be before windows.h */
 #include "private/debug.h"
 
 #include <windows.h>
@@ -65,26 +66,6 @@ typedef enum _LOGICAL_PROCESSOR_RELATIONSHIP {
 #  endif /* HAVE_RELATIONPROCESSORPACKAGE */
 #endif /* HAVE_LOGICAL_PROCESSOR_RELATIONSHIP */
 
-#ifndef HAVE_SYSTEM_LOGICAL_PROCESSOR_INFORMATION
-typedef struct _SYSTEM_LOGICAL_PROCESSOR_INFORMATION {
-  ULONG_PTR ProcessorMask;
-  LOGICAL_PROCESSOR_RELATIONSHIP Relationship;
-  _ANONYMOUS_UNION
-  union {
-    struct {
-      BYTE flags;
-    } ProcessorCore;
-    struct {
-      DWORD NodeNumber;
-    } NumaNode;
-    CACHE_DESCRIPTOR Cache;
-    ULONGLONG Reserved[2];
-  } DUMMYUNIONNAME;
-} SYSTEM_LOGICAL_PROCESSOR_INFORMATION, *PSYSTEM_LOGICAL_PROCESSOR_INFORMATION;
-#endif
-
-/* Extended interface, for group support */
-
 #ifndef HAVE_GROUP_AFFINITY
 typedef struct _GROUP_AFFINITY {
   KAFFINITY Mask;
@@ -93,35 +74,40 @@ typedef struct _GROUP_AFFINITY {
 } GROUP_AFFINITY, *PGROUP_AFFINITY;
 #endif
 
-#ifndef HAVE_PROCESSOR_RELATIONSHIP
+/* always use our own structure because the EfficiencyClass field didn't exist before Win10 */
 typedef struct HWLOC_PROCESSOR_RELATIONSHIP {
   BYTE Flags;
-  BYTE EfficiencyClass; /* for RelationProcessorCore, higher means greater performance but less efficiency, only available in Win10+ */
+  BYTE EfficiencyClass; /* for RelationProcessorCore, higher means greater performance but less efficiency */
   BYTE Reserved[20];
   WORD GroupCount;
   GROUP_AFFINITY GroupMask[ANYSIZE_ARRAY];
-} PROCESSOR_RELATIONSHIP, *PPROCESSOR_RELATIONSHIP;
-#endif
+} HWLOC_PROCESSOR_RELATIONSHIP;
 
-#ifndef HAVE_NUMA_NODE_RELATIONSHIP
-typedef struct _NUMA_NODE_RELATIONSHIP {
+/* always use our own structure because the GroupCount and GroupMasks fields didn't exist in some Win10 */
+typedef struct HWLOC_NUMA_NODE_RELATIONSHIP {
   DWORD NodeNumber;
-  BYTE Reserved[20];
-  GROUP_AFFINITY GroupMask;
-} NUMA_NODE_RELATIONSHIP, *PNUMA_NODE_RELATIONSHIP;
-#endif
+  BYTE Reserved[18];
+  WORD GroupCount;
+  _ANONYMOUS_UNION
+  union {
+    GROUP_AFFINITY GroupMask;
+    GROUP_AFFINITY GroupMasks[ANYSIZE_ARRAY];
+  } DUMMYUNIONNAME;
+} HWLOC_NUMA_NODE_RELATIONSHIP;
 
-#ifndef HAVE_CACHE_RELATIONSHIP
-typedef struct _CACHE_RELATIONSHIP {
+typedef struct HWLOC_CACHE_RELATIONSHIP {
   BYTE Level;
   BYTE Associativity;
   WORD LineSize;
   DWORD CacheSize;
   PROCESSOR_CACHE_TYPE Type;
-  BYTE Reserved[20];
-  GROUP_AFFINITY GroupMask;
-} CACHE_RELATIONSHIP, *PCACHE_RELATIONSHIP;
-#endif
+  BYTE Reserved[18];
+  WORD GroupCount;
+  union {
+    GROUP_AFFINITY GroupMask;
+    GROUP_AFFINITY GroupMasks[ANYSIZE_ARRAY];
+  } DUMMYUNIONNAME;
+} HWLOC_CACHE_RELATIONSHIP;
 
 #ifndef HAVE_PROCESSOR_GROUP_INFO
 typedef struct _PROCESSOR_GROUP_INFO {
@@ -141,20 +127,19 @@ typedef struct _GROUP_RELATIONSHIP {
 } GROUP_RELATIONSHIP, *PGROUP_RELATIONSHIP;
 #endif
 
-#ifndef HAVE_SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX
-typedef struct _SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX {
+/* always use our own structure because we need our own HWLOC_PROCESSOR/CACHE/NUMA_NODE_RELATIONSHIP */
+typedef struct HWLOC_SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX {
   LOGICAL_PROCESSOR_RELATIONSHIP Relationship;
   DWORD Size;
   _ANONYMOUS_UNION
   union {
-    PROCESSOR_RELATIONSHIP Processor;
-    NUMA_NODE_RELATIONSHIP NumaNode;
-    CACHE_RELATIONSHIP Cache;
+    HWLOC_PROCESSOR_RELATIONSHIP Processor;
+    HWLOC_NUMA_NODE_RELATIONSHIP NumaNode;
+    HWLOC_CACHE_RELATIONSHIP Cache;
     GROUP_RELATIONSHIP Group;
     /* Odd: no member to tell the cpu mask of the package... */
   } DUMMYUNIONNAME;
-} SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX, *PSYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX;
-#endif
+} HWLOC_SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX;
 
 #ifndef HAVE_PSAPI_WORKING_SET_EX_BLOCK
 typedef union _PSAPI_WORKING_SET_EX_BLOCK {
@@ -200,10 +185,7 @@ static PFN_GETCURRENTPROCESSORNUMBER GetCurrentProcessorNumberProc;
 typedef VOID (WINAPI *PFN_GETCURRENTPROCESSORNUMBEREX)(PPROCESSOR_NUMBER);
 static PFN_GETCURRENTPROCESSORNUMBEREX GetCurrentProcessorNumberExProc;
 
-typedef BOOL (WINAPI *PFN_GETLOGICALPROCESSORINFORMATION)(PSYSTEM_LOGICAL_PROCESSOR_INFORMATION Buffer, PDWORD ReturnLength);
-static PFN_GETLOGICALPROCESSORINFORMATION GetLogicalProcessorInformationProc;
-
-typedef BOOL (WINAPI *PFN_GETLOGICALPROCESSORINFORMATIONEX)(LOGICAL_PROCESSOR_RELATIONSHIP relationship, PSYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX Buffer, PDWORD ReturnLength);
+typedef BOOL (WINAPI *PFN_GETLOGICALPROCESSORINFORMATIONEX)(LOGICAL_PROCESSOR_RELATIONSHIP relationship, HWLOC_SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX *Buffer, PDWORD ReturnLength);
 static PFN_GETLOGICALPROCESSORINFORMATIONEX GetLogicalProcessorInformationExProc;
 
 typedef BOOL (WINAPI *PFN_SETTHREADGROUPAFFINITY)(HANDLE hThread, const GROUP_AFFINITY *GroupAffinity, PGROUP_AFFINITY PreviousGroupAffinity);
@@ -244,8 +226,6 @@ static void hwloc_win_get_function_ptrs(void)
 	(PFN_GETACTIVEPROCESSORGROUPCOUNT) GetProcAddress(kernel32, "GetActiveProcessorGroupCount");
       GetActiveProcessorCountProc =
 	(PFN_GETACTIVEPROCESSORCOUNT) GetProcAddress(kernel32, "GetActiveProcessorCount");
-      GetLogicalProcessorInformationProc =
-	(PFN_GETLOGICALPROCESSORINFORMATION) GetProcAddress(kernel32, "GetLogicalProcessorInformation");
       GetCurrentProcessorNumberProc =
 	(PFN_GETCURRENTPROCESSORNUMBER) GetProcAddress(kernel32, "GetCurrentProcessorNumber");
       GetCurrentProcessorNumberExProc =
@@ -370,13 +350,13 @@ static hwloc_cpuset_t * processor_group_cpusets = NULL;
 static void
 hwloc_win_get_processor_groups(void)
 {
-  PSYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX procInfoTotal, tmpprocInfoTotal, procInfo;
+  HWLOC_SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX *procInfoTotal, *tmpprocInfoTotal, *procInfo;
   DWORD length;
   unsigned i;
 
   hwloc_debug("querying windows processor groups\n");
 
-  if (!GetActiveProcessorGroupCountProc || !GetLogicalProcessorInformationExProc)
+  if (!GetLogicalProcessorInformationExProc)
     goto error;
 
   nr_processor_groups = GetActiveProcessorGroupCountProc();
@@ -386,7 +366,7 @@ hwloc_win_get_processor_groups(void)
   hwloc_debug("found %lu windows processor groups\n", nr_processor_groups);
 
   if (nr_processor_groups > 1 && SIZEOF_VOID_P == 4) {
-    if (!hwloc_hide_errors())
+    if (HWLOC_SHOW_ALL_ERRORS())
       fprintf(stderr, "hwloc: multiple processor groups found on 32bits Windows, topology may be invalid/incomplete.\n");
   }
 
@@ -415,6 +395,8 @@ hwloc_win_get_processor_groups(void)
 
     assert(procInfo->Relationship == RelationGroup);
 
+    hwloc_debug("Found %u active windows processor groups\n",
+                (unsigned) procInfo->Group.ActiveGroupCount);
     for (id = 0; id < procInfo->Group.ActiveGroupCount; id++) {
       KAFFINITY mask;
       hwloc_bitmap_t set;
@@ -424,8 +406,8 @@ hwloc_win_get_processor_groups(void)
         goto error_with_cpusets;
 
       mask = procInfo->Group.GroupInfo[id].ActiveProcessorMask;
-      hwloc_debug("group %u %d cpus mask %lx\n", id,
-                  procInfo->Group.GroupInfo[id].ActiveProcessorCount, mask);
+      hwloc_debug("group %u with %u cpus mask 0x%llx\n", id,
+                  (unsigned) procInfo->Group.GroupInfo[id].ActiveProcessorCount, (unsigned long long) mask);
       /* KAFFINITY is ULONG_PTR */
       hwloc_bitmap_set_ith_ULONG_PTR(set, id, mask);
       /* FIXME: what if running 32bits on a 64bits windows with 64-processor groups?
@@ -1008,6 +990,8 @@ hwloc_look_windows(struct hwloc_backend *backend, struct hwloc_disc_status *dsta
   unsigned hostname_size = sizeof(hostname);
   int has_efficiencyclass = 0;
   struct hwloc_win_efficiency_classes eclasses;
+  char *env = getenv("HWLOC_WINDOWS_PROCESSOR_GROUP_OBJS");
+  int keep_pgroup_objs = (env && atoi(env));
 
   assert(dstatus->phase == HWLOC_DISC_PHASE_CPU);
 
@@ -1038,137 +1022,8 @@ hwloc_look_windows(struct hwloc_backend *backend, struct hwloc_disc_status *dsta
 
   GetSystemInfo(&SystemInfo);
 
-  if (!GetLogicalProcessorInformationExProc && GetLogicalProcessorInformationProc) {
-      PSYSTEM_LOGICAL_PROCESSOR_INFORMATION procInfo, tmpprocInfo;
-      unsigned id;
-      unsigned i;
-      struct hwloc_obj *obj;
-      hwloc_obj_type_t type;
-
-      length = 0;
-      procInfo = NULL;
-
-      while (1) {
-	if (GetLogicalProcessorInformationProc(procInfo, &length))
-	  break;
-	if (GetLastError() != ERROR_INSUFFICIENT_BUFFER)
-	  return -1;
-	tmpprocInfo = realloc(procInfo, length);
-	if (!tmpprocInfo) {
-	  free(procInfo);
-	  goto out;
-	}
-	procInfo = tmpprocInfo;
-      }
-
-      assert(!length || procInfo);
-
-      for (i = 0; i < length / sizeof(*procInfo); i++) {
-
-        /* Ignore unknown caches */
-	if (procInfo->Relationship == RelationCache
-		&& procInfo->Cache.Type != CacheUnified
-		&& procInfo->Cache.Type != CacheData
-		&& procInfo->Cache.Type != CacheInstruction)
-	  continue;
-
-	id = HWLOC_UNKNOWN_INDEX;
-	switch (procInfo[i].Relationship) {
-	  case RelationNumaNode:
-	    type = HWLOC_OBJ_NUMANODE;
-	    id = procInfo[i].NumaNode.NodeNumber;
-	    gotnuma++;
-	    if (id > max_numanode_index)
-	      max_numanode_index = id;
-	    break;
-	  case RelationProcessorPackage:
-	    type = HWLOC_OBJ_PACKAGE;
-	    break;
-	  case RelationCache:
-	    type = (procInfo[i].Cache.Type == CacheInstruction ? HWLOC_OBJ_L1ICACHE : HWLOC_OBJ_L1CACHE) + procInfo[i].Cache.Level - 1;
-	    break;
-	  case RelationProcessorCore:
-	    type = HWLOC_OBJ_CORE;
-	    break;
-	  case RelationGroup:
-	  default:
-	    type = HWLOC_OBJ_GROUP;
-	    break;
-	}
-
-	if (!hwloc_filter_check_keep_object_type(topology, type))
-	  continue;
-
-	obj = hwloc_alloc_setup_object(topology, type, id);
-        obj->cpuset = hwloc_bitmap_alloc();
-	hwloc_debug("%s#%u mask %llx\n", hwloc_obj_type_string(type), id, (unsigned long long) procInfo[i].ProcessorMask);
-	/* ProcessorMask is a ULONG_PTR */
-	hwloc_bitmap_set_ith_ULONG_PTR(obj->cpuset, 0, procInfo[i].ProcessorMask);
-	hwloc_debug_2args_bitmap("%s#%u bitmap %s\n", hwloc_obj_type_string(type), id, obj->cpuset);
-
-	switch (type) {
-	  case HWLOC_OBJ_NUMANODE:
-	    {
-	      ULONGLONG avail;
-	      obj->nodeset = hwloc_bitmap_alloc();
-	      hwloc_bitmap_set(obj->nodeset, id);
-	      if ((GetNumaAvailableMemoryNodeExProc && GetNumaAvailableMemoryNodeExProc(id, &avail))
-		  || (GetNumaAvailableMemoryNodeProc && GetNumaAvailableMemoryNodeProc(id, &avail))) {
-		obj->attr->numanode.local_memory = avail;
-		gotnumamemory++;
-	      }
-	      obj->attr->numanode.page_types_len = 2;
-	      obj->attr->numanode.page_types = malloc(2 * sizeof(*obj->attr->numanode.page_types));
-	      memset(obj->attr->numanode.page_types, 0, 2 * sizeof(*obj->attr->numanode.page_types));
-	      obj->attr->numanode.page_types_len = 1;
-	      obj->attr->numanode.page_types[0].size = SystemInfo.dwPageSize;
-#if HAVE_DECL__SC_LARGE_PAGESIZE
-	      obj->attr->numanode.page_types_len++;
-	      obj->attr->numanode.page_types[1].size = sysconf(_SC_LARGE_PAGESIZE);
-#endif
-	      break;
-	    }
-	  case HWLOC_OBJ_L1CACHE:
-	  case HWLOC_OBJ_L2CACHE:
-	  case HWLOC_OBJ_L3CACHE:
-	  case HWLOC_OBJ_L4CACHE:
-	  case HWLOC_OBJ_L5CACHE:
-	  case HWLOC_OBJ_L1ICACHE:
-	  case HWLOC_OBJ_L2ICACHE:
-	  case HWLOC_OBJ_L3ICACHE:
-	    obj->attr->cache.size = procInfo[i].Cache.Size;
-	    obj->attr->cache.associativity = procInfo[i].Cache.Associativity == CACHE_FULLY_ASSOCIATIVE ? -1 : procInfo[i].Cache.Associativity ;
-	    obj->attr->cache.linesize = procInfo[i].Cache.LineSize;
-	    obj->attr->cache.depth = procInfo[i].Cache.Level;
-	    switch (procInfo->Cache.Type) {
-	      case CacheUnified:
-		obj->attr->cache.type = HWLOC_OBJ_CACHE_UNIFIED;
-		break;
-	      case CacheData:
-		obj->attr->cache.type = HWLOC_OBJ_CACHE_DATA;
-		break;
-	      case CacheInstruction:
-		obj->attr->cache.type = HWLOC_OBJ_CACHE_INSTRUCTION;
-		break;
-	      default:
-		hwloc_free_unlinked_object(obj);
-		continue;
-	    }
-	    break;
-	  case HWLOC_OBJ_GROUP:
-	    obj->attr->group.kind = procInfo[i].Relationship == RelationGroup ? HWLOC_GROUP_KIND_WINDOWS_PROCESSOR_GROUP : HWLOC_GROUP_KIND_WINDOWS_RELATIONSHIP_UNKNOWN;
-	    break;
-	  default:
-	    break;
-	}
-	hwloc__insert_object_by_cpuset(topology, NULL, obj, "windows:GetLogicalProcessorInformation");
-      }
-
-      free(procInfo);
-  }
-
   if (GetLogicalProcessorInformationExProc) {
-      PSYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX procInfoTotal, tmpprocInfoTotal, procInfo;
+      HWLOC_SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX *procInfoTotal, *tmpprocInfoTotal, *procInfo;
       unsigned id;
       struct hwloc_obj *obj;
       hwloc_obj_type_t type;
@@ -1207,8 +1062,16 @@ hwloc_look_windows(struct hwloc_backend *backend, struct hwloc_disc_status *dsta
 	switch (procInfo->Relationship) {
 	  case RelationNumaNode:
 	    type = HWLOC_OBJ_NUMANODE;
-            num = 1;
-            GroupMask = &procInfo->NumaNode.GroupMask;
+            /* Starting with Windows 11 and Server 2022, the GroupCount field is valid and >=1
+             * and we may read GroupMasks[]. Older releases have GroupCount==0 and we must read GroupMask.
+             */
+            if (procInfo->NumaNode.GroupCount) {
+              num = procInfo->NumaNode.GroupCount;
+              GroupMask = procInfo->NumaNode.GroupMasks;
+            } else {
+              num = 1;
+              GroupMask = &procInfo->NumaNode.GroupMask;
+            }
 	    id = procInfo->NumaNode.NodeNumber;
 	    gotnuma++;
 	    if (id > max_numanode_index)
@@ -1221,18 +1084,20 @@ hwloc_look_windows(struct hwloc_backend *backend, struct hwloc_disc_status *dsta
 	    break;
 	  case RelationCache:
 	    type = (procInfo->Cache.Type == CacheInstruction ? HWLOC_OBJ_L1ICACHE : HWLOC_OBJ_L1CACHE) + procInfo->Cache.Level - 1;
-            num = 1;
-            GroupMask = &procInfo->Cache.GroupMask;
+            /* GroupCount added approximately with NumaNode.GroupCount above */
+            if (procInfo->Cache.GroupCount) {
+              num = procInfo->Cache.GroupCount;
+              GroupMask = procInfo->Cache.GroupMasks;
+            } else {
+              num = 1;
+              GroupMask = &procInfo->Cache.GroupMask;
+            }
 	    break;
 	  case RelationProcessorCore:
 	    type = HWLOC_OBJ_CORE;
             num = procInfo->Processor.GroupCount;
             GroupMask = procInfo->Processor.GroupMask;
-            if (has_efficiencyclass)
-              /* the EfficiencyClass field didn't exist before Windows10 and recent MSVC headers,
-               * so just access it manually instead of trying to detect it.
-               */
-              efficiency_class = * ((&procInfo->Processor.Flags) + 1);
+            efficiency_class = procInfo->Processor.EfficiencyClass;
 	    break;
 	  case RelationGroup:
 	    /* So strange an interface... */
@@ -1257,11 +1122,12 @@ hwloc_look_windows(struct hwloc_backend *backend, struct hwloc_disc_status *dsta
 		groups_pu_set = hwloc_bitmap_alloc();
 	      hwloc_bitmap_or(groups_pu_set, groups_pu_set, set);
 
-	      if (hwloc_filter_check_keep_object_type(topology, HWLOC_OBJ_GROUP)) {
+              /* Ignore processor groups unless requested and filtered-in */
+              if (keep_pgroup_objs && hwloc_filter_check_keep_object_type(topology, HWLOC_OBJ_GROUP)) {
 		obj = hwloc_alloc_setup_object(topology, HWLOC_OBJ_GROUP, id);
 		obj->cpuset = set;
 		obj->attr->group.kind = HWLOC_GROUP_KIND_WINDOWS_PROCESSOR_GROUP;
-		hwloc__insert_object_by_cpuset(topology, NULL, obj, "windows:GetLogicalProcessorInformation:ProcessorGroup");
+		hwloc__insert_object_by_cpuset(topology, NULL, obj, "windows:GetLogicalProcessorInformationEx:ProcessorGroup");
 	      } else
 		hwloc_bitmap_free(set);
 	    }

src/3rdparty/hwloc/src/topology-x86.c

@@ -1,5 +1,5 @@
 /*
- * Copyright © 2010-2021 Inria.  All rights reserved.
+ * Copyright © 2010-2022 Inria.  All rights reserved.
  * Copyright © 2010-2013 Université Bordeaux
  * Copyright © 2010-2011 Cisco Systems, Inc.  All rights reserved.
  * See COPYING in top-level directory.
@@ -500,7 +500,8 @@ static void read_amd_cores_topoext(struct procinfo *infos, unsigned long flags,
       nodes_per_proc = ((ecx >> 8) & 7) + 1;
     }
     if ((infos->cpufamilynumber == 0x15 && nodes_per_proc > 2)
-	|| ((infos->cpufamilynumber == 0x17 || infos->cpufamilynumber == 0x18) && nodes_per_proc > 4)) {
+	|| ((infos->cpufamilynumber == 0x17 || infos->cpufamilynumber == 0x18) && nodes_per_proc > 4)
+        || (infos->cpufamilynumber == 0x19 && nodes_per_proc > 1)) {
       hwloc_debug("warning: undefined nodes_per_proc value %u, assuming it means %u\n", nodes_per_proc, nodes_per_proc);
     }
   }
@@ -613,10 +614,13 @@ static void look_proc(struct hwloc_backend *backend, struct procinfo *infos, uns
   eax = 0x01;
   cpuid_or_from_dump(&eax, &ebx, &ecx, &edx, src_cpuiddump);
   infos->apicid = ebx >> 24;
-  if (edx & (1 << 28))
+  if (edx & (1 << 28)) {
     legacy_max_log_proc = 1 << hwloc_flsl(((ebx >> 16) & 0xff) - 1);
-  else
+  } else {
+    hwloc_debug("HTT bit not set in CPUID 0x01.edx, assuming legacy_max_log_proc = 1\n");
     legacy_max_log_proc = 1;
+  }
+
   hwloc_debug("APIC ID 0x%02x legacy_max_log_proc %u\n", infos->apicid, legacy_max_log_proc);
   infos->ids[PKG] = infos->apicid / legacy_max_log_proc;
   legacy_log_proc_id = infos->apicid % legacy_max_log_proc;
@@ -679,12 +683,23 @@ static void look_proc(struct hwloc_backend *backend, struct procinfo *infos, uns
       unsigned max_nbcores;
       unsigned max_nbthreads;
       unsigned threadid __hwloc_attribute_unused;
+      hwloc_debug("Trying to get core/thread IDs from 0x04...\n");
       max_nbcores = ((eax >> 26) & 0x3f) + 1;
-      max_nbthreads = legacy_max_log_proc / max_nbcores;
-      hwloc_debug("thus %u threads\n", max_nbthreads);
-      threadid = legacy_log_proc_id % max_nbthreads;
-      infos->ids[CORE] = legacy_log_proc_id / max_nbthreads;
-      hwloc_debug("this is thread %u of core %u\n", threadid, infos->ids[CORE]);
+      hwloc_debug("found %u cores max\n", max_nbcores);
+      /* some VMs (e.g. issue#525) don't report valid information, check things before dividing by 0. */
+      if (!max_nbcores) {
+        hwloc_debug("cannot detect core/thread IDs from 0x04 without a valid max of cores\n");
+      } else {
+        max_nbthreads = legacy_max_log_proc / max_nbcores;
+        hwloc_debug("found %u threads max\n", max_nbthreads);
+        if (!max_nbthreads) {
+          hwloc_debug("cannot detect core/thread IDs from 0x04 without a valid max of threads\n");
+        } else {
+          threadid = legacy_log_proc_id % max_nbthreads;
+          infos->ids[CORE] = legacy_log_proc_id / max_nbthreads;
+          hwloc_debug("this is thread %u of core %u\n", threadid, infos->ids[CORE]);
+        }
+      }
     }
   }
 
@@ -775,13 +790,19 @@ static void look_proc(struct hwloc_backend *backend, struct procinfo *infos, uns
 
     } else if (cpuid_type == amd) {
       /* AMD quirks */
-      if (infos->cpufamilynumber == 0x17
-	  && cache->level == 3 && cache->nbthreads_sharing == 6) {
-	/* AMD family 0x17 always shares L3 between 8 APIC ids,
-	 * even when only 6 APIC ids are enabled and reported in nbthreads_sharing
-	 * (on 24-core CPUs).
+      if (infos->cpufamilynumber >= 0x17 && cache->level == 3) {
+	/* AMD family 0x19 always shares L3 between 16 APIC ids (8 HT cores).
+         * while Family 0x17 shares between 8 APIC ids (4 HT cores).
+         * But many models have less APIC ids enabled and reported in nbthreads_sharing.
+         * It means we must round-up nbthreads_sharing to the nearest power of 2
+         * before computing cacheid.
 	 */
-	cache->cacheid = infos->apicid / 8;
+        unsigned nbapics_sharing = cache->nbthreads_sharing;
+        if (nbapics_sharing & (nbapics_sharing-1))
+          /* not a power of two, round-up */
+          nbapics_sharing = 1U<<(1+hwloc_ffsl(nbapics_sharing));
+
+	cache->cacheid = infos->apicid / nbapics_sharing;
 
       } else if (infos->cpufamilynumber== 0x10 && infos->cpumodelnumber == 0x9
 	  && cache->level == 3
@@ -807,7 +828,7 @@ static void look_proc(struct hwloc_backend *backend, struct procinfo *infos, uns
       } else if (infos->cpufamilynumber == 0x15
 		 && (infos->cpumodelnumber == 0x1 /* Bulldozer */ || infos->cpumodelnumber == 0x2 /* Piledriver */)
 		 && cache->level == 3 && cache->nbthreads_sharing == 6) {
-	/* AMD Bulldozer and Piledriver 12-core processors have same APIC ids as Magny-Cours below,
+	/* AMD Bulldozer and Piledriver 12-core processors have same APIC ids as Magny-Cours above,
 	 * but we can't merge the checks because the original nbthreads_sharing must be exactly 6 here.
 	 */
 	cache->cacheid = (infos->apicid % legacy_max_log_proc) / cache->nbthreads_sharing /* cacheid within the package */
@@ -1231,6 +1252,18 @@ static void summarize(struct hwloc_backend *backend, struct procinfo *infos, uns
 	    }
 	  }
 	  cache = hwloc_alloc_setup_object(topology, otype, HWLOC_UNKNOWN_INDEX);
+          /* We don't specify the os_index of caches because we want to be
+           * 100% sure they are identical to what the Linux kernel reports
+           * (so that things like resctrl work).
+           * However, vendor/model-specific quirks in the x86 code above
+           * make this difficult.
+           *
+           * Caveat: if the x86 backend is used on Linux to avoid kernel bugs,
+           * IDs won't be available to resctrl users. But resctrl heavily
+           * relies on the kernel x86 discovery being non-buggy anyway.
+           *
+           * TODO: make this optional? or only disable it on Linux?
+           */
 	  cache->attr->cache.depth = level;
 	  cache->attr->cache.size = infos[i].cache[l].size;
 	  cache->attr->cache.linesize = infos[i].cache[l].linesize;
@@ -1316,7 +1349,7 @@ look_procs(struct hwloc_backend *backend, struct procinfo *infos, unsigned long
   if (data->apicid_unique) {
     summarize(backend, infos, flags);
 
-    if (has_hybrid(features)) {
+    if (has_hybrid(features) && !(topology->flags & HWLOC_TOPOLOGY_FLAG_NO_CPUKINDS)) {
       /* use hybrid info for cpukinds */
       hwloc_bitmap_t atomset = hwloc_bitmap_alloc();
       hwloc_bitmap_t coreset = hwloc_bitmap_alloc();
@@ -1688,7 +1721,7 @@ hwloc_x86_check_cpuiddump_input(const char *src_cpuiddump_path, hwloc_bitmap_t s
   char line [32];
 
   dir = opendir(src_cpuiddump_path);
-  if (!dir)
+  if (!dir) 
     return -1;
 
   path = malloc(strlen(src_cpuiddump_path) + strlen("/hwloc-cpuid-info") + 1);

src/3rdparty/hwloc/src/topology-xml.c

@@ -1,6 +1,6 @@
 /*
  * Copyright © 2009 CNRS
- * Copyright © 2009-2021 Inria.  All rights reserved.
+ * Copyright © 2009-2022 Inria.  All rights reserved.
  * Copyright © 2009-2011, 2020 Université Bordeaux
  * Copyright © 2009-2018 Cisco Systems, Inc.  All rights reserved.
  * See COPYING in top-level directory.
@@ -123,6 +123,17 @@ hwloc__xml_import_object_attr(struct hwloc_topology *topology,
       fprintf(stderr, "%s: unexpected zero gp_index, topology may be invalid\n", state->global->msgprefix);
     if (obj->gp_index >= topology->next_gp_index)
       topology->next_gp_index = obj->gp_index + 1;
+  } else if (!strcmp(name, "id")) { /* forward compat */
+    if (!strncmp(value, "obj", 3)) {
+      obj->gp_index = strtoull(value+3, NULL, 10);
+      if (!obj->gp_index && hwloc__xml_verbose())
+        fprintf(stderr, "%s: unexpected zero id, topology may be invalid\n", state->global->msgprefix);
+      if (obj->gp_index >= topology->next_gp_index)
+        topology->next_gp_index = obj->gp_index + 1;
+    } else {
+      if (hwloc__xml_verbose())
+        fprintf(stderr, "%s: unexpected id `%s' not-starting with `obj', ignoring\n", state->global->msgprefix, value);
+    }
   } else if (!strcmp(name, "cpuset")) {
     if (!obj->cpuset)
       obj->cpuset = hwloc_bitmap_alloc();
@@ -243,7 +254,7 @@ hwloc__xml_import_object_attr(struct hwloc_topology *topology,
   else if (!strcmp(name, "dont_merge")) {
     unsigned long lvalue = strtoul(value, NULL, 10);
     if (obj->type == HWLOC_OBJ_GROUP)
-      obj->attr->group.dont_merge = lvalue;
+      obj->attr->group.dont_merge = (unsigned char) lvalue;
     else if (hwloc__xml_verbose())
       fprintf(stderr, "%s: ignoring dont_merge attribute for non-group object type\n",
 	      state->global->msgprefix);
@@ -263,7 +274,7 @@ hwloc__xml_import_object_attr(struct hwloc_topology *topology,
 #ifndef HWLOC_HAVE_32BITS_PCI_DOMAIN
       } else if (domain > 0xffff) {
 	static int warned = 0;
-	if (!warned && hwloc_hide_errors() < 2)
+	if (!warned && HWLOC_SHOW_ALL_ERRORS())
 	  fprintf(stderr, "hwloc/xml: Ignoring PCI device with non-16bit domain.\nPass --enable-32bits-pci-domain to configure to support such devices\n(warning: it would break the library ABI, don't enable unless really needed).\n");
 	warned = 1;
 	*ignore = 1;
@@ -363,7 +374,7 @@ hwloc__xml_import_object_attr(struct hwloc_topology *topology,
 #ifndef HWLOC_HAVE_32BITS_PCI_DOMAIN
       } else if (domain > 0xffff) {
 	static int warned = 0;
-	if (!warned && hwloc_hide_errors() < 2)
+	if (!warned && HWLOC_SHOW_ALL_ERRORS())
 	  fprintf(stderr, "hwloc/xml: Ignoring bridge to PCI with non-16bit domain.\nPass --enable-32bits-pci-domain to configure to support such devices\n(warning: it would break the library ABI, don't enable unless really needed).\n");
 	warned = 1;
 	*ignore = 1;
@@ -1235,7 +1246,7 @@ hwloc__xml_import_object(hwloc_topology_t topology,
 	/* next should be before cur */
 	if (!childrengotignored) {
 	  static int reported = 0;
-	  if (!reported && hwloc_hide_errors() < 2) {
+	  if (!reported && HWLOC_SHOW_CRITICAL_ERRORS()) {
 	    hwloc__xml_import_report_outoforder(topology, next, cur);
 	    reported = 1;
 	  }
@@ -1568,6 +1579,9 @@ hwloc__xml_v2import_distances(hwloc_topology_t topology,
     }
   }
 
+  if (topology->flags & HWLOC_TOPOLOGY_FLAG_NO_DISTANCES)
+    goto out_ignore;
+
   hwloc_internal_distances_add_by_index(topology, name, unique_type, different_types, nbobjs, indexes, u64values, kind, 0 /* assume grouping was applied when this matrix was discovered before exporting to XML */);
 
   /* prevent freeing below */
@@ -1722,7 +1736,8 @@ hwloc__xml_import_memattr(hwloc_topology_t topology,
     }
   }
 
-  if (name && flags != (unsigned long) -1) {
+  if (name && flags != (unsigned long) -1
+      && !(topology->flags & HWLOC_TOPOLOGY_FLAG_NO_MEMATTRS)) {
     hwloc_memattr_id_t _id;
 
     ret = hwloc_memattr_get_by_name(topology, name, &_id);
@@ -1833,7 +1848,13 @@ hwloc__xml_import_cpukind(hwloc_topology_t topology,
     goto error;
   }
 
-  hwloc_internal_cpukinds_register(topology, cpuset, forced_efficiency, infos, nr_infos, HWLOC_CPUKINDS_REGISTER_FLAG_OVERWRITE_FORCED_EFFICIENCY);
+  if (topology->flags & HWLOC_TOPOLOGY_FLAG_NO_CPUKINDS) {
+    hwloc__free_infos(infos, nr_infos);
+    hwloc_bitmap_free(cpuset);
+  } else {
+    hwloc_internal_cpukinds_register(topology, cpuset, forced_efficiency, infos, nr_infos, HWLOC_CPUKINDS_REGISTER_FLAG_OVERWRITE_FORCED_EFFICIENCY);
+    hwloc__free_infos(infos, nr_infos);
+  }
 
   return state->global->close_tag(state);
 
@@ -2168,7 +2189,8 @@ done:
        * but it would require to have those objects in the original XML order (like the first_numanode cousin-list).
        * because the topology order can be different if some parents are ignored during load.
        */
-      if (nbobjs == data->nbnumanodes) {
+      if (nbobjs == data->nbnumanodes
+          && !(topology->flags & HWLOC_TOPOLOGY_FLAG_NO_DISTANCES)) {
 	hwloc_obj_t *objs = malloc(nbobjs*sizeof(hwloc_obj_t));
 	uint64_t *values = malloc(nbobjs*nbobjs*sizeof(*values));
         assert(data->nbnumanodes > 0); /* v1dist->nbobjs is >0 after import */
@@ -2650,7 +2672,7 @@ hwloc__xml_export_object_contents (hwloc__xml_export_state_t state, hwloc_topolo
 
       logical_to_v2array = malloc(nbobjs * sizeof(*logical_to_v2array));
       if (!logical_to_v2array) {
-        if (!hwloc_hide_errors())
+        if (HWLOC_SHOW_ALL_ERRORS())
           fprintf(stderr, "hwloc/xml/export/v1: failed to allocated logical_to_v2array\n");
 	continue;
       }
@@ -2825,6 +2847,7 @@ hwloc__xml_v1export_object_with_memory(hwloc__xml_export_state_t parentstate, hw
     /* child has sibling, we must add a Group around those memory children */
     hwloc_obj_t group = parentstate->global->v1_memory_group;
     parentstate->new_child(parentstate, &gstate, "object");
+    group->parent = obj->parent;
     group->cpuset = obj->cpuset;
     group->complete_cpuset = obj->complete_cpuset;
     group->nodeset = obj->nodeset;
@@ -3123,9 +3146,11 @@ hwloc__xml_export_memattrs(hwloc__xml_export_state_t state, hwloc_topology_t top
       continue;
 
     imattr = &topology->memattrs[id];
-    if ((id == HWLOC_MEMATTR_ID_LATENCY || id == HWLOC_MEMATTR_ID_BANDWIDTH)
-        && !imattr->nr_targets)
-      /* no need to export target-less attributes for initial attributes, no release support attributes without those definitions */
+    if (id < HWLOC_MEMATTR_ID_MAX && !imattr->nr_targets)
+      /* no need to export standard attributes without any target,
+       * their definition is now standardized,
+       * the old hwloc importing this XML may recreate these attributes just like it would for a non-imported topology.
+       */
       continue;
 
     state->new_child(state, &mstate, "memattr");

src/3rdparty/hwloc/src/topology.c

@@ -1,8 +1,9 @@
 /*
  * Copyright © 2009 CNRS
- * Copyright © 2009-2021 Inria.  All rights reserved.
+ * Copyright © 2009-2022 Inria.  All rights reserved.
  * Copyright © 2009-2012, 2020 Université Bordeaux
  * Copyright © 2009-2011 Cisco Systems, Inc.  All rights reserved.
+ * Copyright © 2022 IBM Corporation.  All rights reserved.
  * See COPYING in top-level directory.
  */
 
@@ -52,29 +53,41 @@
 #include <windows.h>
 #endif
 
+
+#ifdef HWLOC_HAVE_LEVELZERO
 /*
  * Define ZES_ENABLE_SYSMAN=1 early so that the LevelZero backend gets Sysman enabled.
+ *
+ * Only if the levelzero was enabled in this build so that we don't enable sysman
+ * for external levelzero users when hwloc doesn't need it. If somebody ever loads
+ * an external levelzero plugin in a hwloc library built without levelzero (unlikely),
+ * he may have to manually set ZES_ENABLE_SYSMAN=1.
+ *
  * Use the constructor if supported and/or the Windows DllMain callback.
  * Do it in the main hwloc library instead of the levelzero component because
  * the latter could be loaded later as a plugin.
  *
  * L0 seems to be using getenv() to check this variable on Windows
  * (at least in the Intel Compute-Runtime of March 2021),
- * so use putenv() to set the variable.
+ * but setenv() doesn't seem to exist on Windows, hence use putenv() to set the variable.
  *
  * For the record, Get/SetEnvironmentVariable() is not exactly the same as getenv/putenv():
  * - getenv() doesn't see what was set with SetEnvironmentVariable()
  * - GetEnvironmentVariable() doesn't see putenv() in cygwin (while it does in MSVC and MinGW).
  * Hence, if L0 ever switches from getenv() to GetEnvironmentVariable(),
  * it will break in cygwin, we'll have to use both putenv() and SetEnvironmentVariable().
- * Hopefully L0 will be provide a way to enable Sysman without env vars before it happens.
+ * Hopefully L0 will provide a way to enable Sysman without env vars before it happens.
  */
-#ifdef HWLOC_HAVE_ATTRIBUTE_CONSTRUCTOR
+#if HWLOC_HAVE_ATTRIBUTE_CONSTRUCTOR
 static void hwloc_constructor(void) __attribute__((constructor));
 static void hwloc_constructor(void)
 {
   if (!getenv("ZES_ENABLE_SYSMAN"))
-    putenv((char *) "ZES_ENABLE_SYSMAN=1");
+#ifdef HWLOC_WIN_SYS
+    putenv("ZES_ENABLE_SYSMAN=1");
+#else
+    setenv("ZES_ENABLE_SYSMAN", "1", 1);
+#endif
 }
 #endif
 #ifdef HWLOC_WIN_SYS
@@ -82,11 +95,14 @@ BOOL WINAPI DllMain(HINSTANCE hinstDLL, DWORD fdwReason, LPVOID lpReserved)
 {
   if (fdwReason == DLL_PROCESS_ATTACH) {
     if (!getenv("ZES_ENABLE_SYSMAN"))
+      /* Windows does not have a setenv, so use putenv. */
       putenv((char *) "ZES_ENABLE_SYSMAN=1");
   }
   return TRUE;
 }
 #endif
+#endif /* HWLOC_HAVE_LEVELZERO */
+
 
 unsigned hwloc_get_api_version(void)
 {
@@ -98,14 +114,25 @@ int hwloc_topology_abi_check(hwloc_topology_t topology)
   return topology->topology_abi != HWLOC_TOPOLOGY_ABI ? -1 : 0;
 }
 
+/* callers should rather use wrappers HWLOC_SHOW_ALL_ERRORS() and HWLOC_SHOW_CRITICAL_ERRORS() for clarity */
 int hwloc_hide_errors(void)
 {
   static int hide = 1; /* only show critical errors by default. lstopo will show others */
   static int checked = 0;
   if (!checked) {
     const char *envvar = getenv("HWLOC_HIDE_ERRORS");
-    if (envvar)
+    if (envvar) {
       hide = atoi(envvar);
+#ifdef HWLOC_DEBUG
+    } else {
+      /* if debug is enabled and HWLOC_DEBUG_VERBOSE isn't forced to 0,
+       * show all errors jus like we show all debug messages.
+       */
+      envvar = getenv("HWLOC_DEBUG_VERBOSE");
+      if (!envvar || atoi(envvar))
+        hide = 0;
+#endif
+    }
     checked = 1;
   }
   return hide;
@@ -142,7 +169,7 @@ static void report_insert_error(hwloc_obj_t new, hwloc_obj_t old, const char *ms
 {
   static int reported = 0;
 
-  if (reason && !reported && hwloc_hide_errors() < 2) {
+  if (reason && !reported && HWLOC_SHOW_CRITICAL_ERRORS()) {
     char newstr[512];
     char oldstr[512];
     report_insert_error_format_obj(newstr, sizeof(newstr), new);
@@ -1901,6 +1928,9 @@ hwloc_topology_alloc_group_object(struct hwloc_topology *topology)
 static void hwloc_propagate_symmetric_subtree(hwloc_topology_t topology, hwloc_obj_t root);
 static void propagate_total_memory(hwloc_obj_t obj);
 static void hwloc_set_group_depth(hwloc_topology_t topology);
+static void hwloc_connect_children(hwloc_obj_t parent);
+static int hwloc_connect_levels(hwloc_topology_t topology);
+static int hwloc_connect_special_levels(hwloc_topology_t topology);
 
 hwloc_obj_t
 hwloc_topology_insert_group_object(struct hwloc_topology *topology, hwloc_obj_t obj)
@@ -2474,13 +2504,26 @@ hwloc_compare_levels_structure(hwloc_topology_t topology, unsigned i)
   return 0;
 }
 
-/* return > 0 if any level was removed, which means reconnect is needed */
-static void
+/* return > 0 if any level was removed.
+ * performs its own reconnect internally if needed
+ */
+static int
 hwloc_filter_levels_keep_structure(hwloc_topology_t topology)
 {
   unsigned i, j;
   int res = 0;
 
+  if (topology->modified) {
+    /* WARNING: hwloc_topology_reconnect() is duplicated partially here
+     * and at the end of this function:
+     * - we need normal levels before merging.
+     * - and we'll need to update special levels after merging.
+     */
+    hwloc_connect_children(topology->levels[0][0]);
+    if (hwloc_connect_levels(topology) < 0)
+      return -1;
+  }
+
   /* start from the bottom since we'll remove intermediate levels */
   for(i=topology->nb_levels-1; i>0; i--) {
     int replacechild = 0, replaceparent = 0;
@@ -2646,6 +2689,22 @@ hwloc_filter_levels_keep_structure(hwloc_topology_t topology)
 	topology->type_depth[type] = HWLOC_TYPE_DEPTH_MULTIPLE;
     }
   }
+
+
+  if (res > 0 || topology-> modified) {
+    /* WARNING: hwloc_topology_reconnect() is duplicated partially here
+     * and at the beginning of this function.
+     * If we merged some levels, some child+parent special children lisst
+     * may have been merged, hence specials level might need reordering,
+     * So reconnect special levels only here at the end
+     * (it's not needed at the beginning of this function).
+     */
+    if (hwloc_connect_special_levels(topology) < 0)
+      return -1;
+    topology->modified = 0;
+  }
+
+  return 0;
 }
 
 static void
@@ -2963,9 +3022,9 @@ hwloc_list_special_objects(hwloc_topology_t topology, hwloc_obj_t obj)
   }
 }
 
-/* Build I/O levels */
+/* Build Memory, I/O and Misc levels */
 static int
-hwloc_connect_io_misc_levels(hwloc_topology_t topology)
+hwloc_connect_special_levels(hwloc_topology_t topology)
 {
   unsigned i;
 
@@ -3130,7 +3189,7 @@ hwloc_connect_levels(hwloc_topology_t topology)
       tmpnbobjs = realloc(topology->level_nbobjects,
 			  2 * topology->nb_levels_allocated * sizeof(*topology->level_nbobjects));
       if (!tmplevels || !tmpnbobjs) {
-        if (hwloc_hide_errors() < 2)
+        if (HWLOC_SHOW_CRITICAL_ERRORS())
           fprintf(stderr, "hwloc: failed to realloc level arrays to %u\n", topology->nb_levels_allocated * 2);
 
 	/* if one realloc succeeded, make sure the caller will free the new buffer */
@@ -3176,6 +3235,10 @@ hwloc_connect_levels(hwloc_topology_t topology)
 int
 hwloc_topology_reconnect(struct hwloc_topology *topology, unsigned long flags)
 {
+  /* WARNING: when updating this function, the replicated code must
+   * also be updated inside hwloc_filter_levels_keep_structure()
+   */
+
   if (flags) {
     errno = EINVAL;
     return -1;
@@ -3188,7 +3251,7 @@ hwloc_topology_reconnect(struct hwloc_topology *topology, unsigned long flags)
   if (hwloc_connect_levels(topology) < 0)
     return -1;
 
-  if (hwloc_connect_io_misc_levels(topology) < 0)
+  if (hwloc_connect_special_levels(topology) < 0)
     return -1;
 
   topology->modified = 0;
@@ -3484,6 +3547,8 @@ hwloc_discover(struct hwloc_topology *topology,
   /*
    * Additional discovery
    */
+  hwloc_pci_discovery_prepare(topology);
+
   if (topology->backend_phases & HWLOC_DISC_PHASE_PCI) {
     dstatus->phase = HWLOC_DISC_PHASE_PCI;
     hwloc_discover_by_phase(topology, dstatus, "PCI");
@@ -3501,6 +3566,8 @@ hwloc_discover(struct hwloc_topology *topology,
     hwloc_discover_by_phase(topology, dstatus, "ANNOTATE");
   }
 
+  hwloc_pci_discovery_exit(topology); /* pci needed up to annotate */
+
   if (getenv("HWLOC_DEBUG_SORT_CHILDREN"))
     hwloc_debug_sort_children(topology->levels[0][0]);
 
@@ -3513,31 +3580,28 @@ hwloc_discover(struct hwloc_topology *topology,
   hwloc_debug("%s", "\nRemoving empty objects\n");
   remove_empty(topology, &topology->levels[0][0]);
   if (!topology->levels[0][0]) {
-    if (hwloc_hide_errors() < 2)
+    if (HWLOC_SHOW_CRITICAL_ERRORS())
       fprintf(stderr, "hwloc: Topology became empty, aborting!\n");
     return -1;
   }
   if (hwloc_bitmap_iszero(topology->levels[0][0]->cpuset)) {
-    if (hwloc_hide_errors() < 2)
+    if (HWLOC_SHOW_CRITICAL_ERRORS())
       fprintf(stderr, "hwloc: Topology does not contain any PU, aborting!\n");
     return -1;
   }
   if (hwloc_bitmap_iszero(topology->levels[0][0]->nodeset)) {
-    if (hwloc_hide_errors() < 2)
+    if (HWLOC_SHOW_CRITICAL_ERRORS())
       fprintf(stderr, "hwloc: Topology does not contain any NUMA node, aborting!\n");
     return -1;
   }
   hwloc_debug_print_objects(0, topology->levels[0][0]);
 
-  /* Reconnect things after all these changes.
-   * Often needed because of Groups inserted for I/Os.
-   * And required for KEEP_STRUCTURE below.
-   */
-  if (hwloc_topology_reconnect(topology, 0) < 0)
-    return -1;
-
   hwloc_debug("%s", "\nRemoving levels with HWLOC_TYPE_FILTER_KEEP_STRUCTURE\n");
-  hwloc_filter_levels_keep_structure(topology);
+  if (hwloc_filter_levels_keep_structure(topology) < 0)
+    return -1;
+  /* takes care of reconnecting children/levels internally,
+   * because it needs normal levels.
+   * and it's often needed below because of Groups inserted for I/Os anyway */
   hwloc_debug_print_objects(0, topology->levels[0][0]);
 
   /* accumulate children memory in total_memory fields (only once parent is set) */
@@ -3762,7 +3826,16 @@ hwloc_topology_set_flags (struct hwloc_topology *topology, unsigned long flags)
     return -1;
   }
 
-  if (flags & ~(HWLOC_TOPOLOGY_FLAG_INCLUDE_DISALLOWED|HWLOC_TOPOLOGY_FLAG_IS_THISSYSTEM|HWLOC_TOPOLOGY_FLAG_THISSYSTEM_ALLOWED_RESOURCES|HWLOC_TOPOLOGY_FLAG_IMPORT_SUPPORT|HWLOC_TOPOLOGY_FLAG_RESTRICT_TO_CPUBINDING|HWLOC_TOPOLOGY_FLAG_RESTRICT_TO_MEMBINDING|HWLOC_TOPOLOGY_FLAG_DONT_CHANGE_BINDING)) {
+  if (flags & ~(HWLOC_TOPOLOGY_FLAG_INCLUDE_DISALLOWED
+                |HWLOC_TOPOLOGY_FLAG_IS_THISSYSTEM
+                |HWLOC_TOPOLOGY_FLAG_THISSYSTEM_ALLOWED_RESOURCES
+                |HWLOC_TOPOLOGY_FLAG_IMPORT_SUPPORT
+                |HWLOC_TOPOLOGY_FLAG_RESTRICT_TO_CPUBINDING
+                |HWLOC_TOPOLOGY_FLAG_RESTRICT_TO_MEMBINDING
+                |HWLOC_TOPOLOGY_FLAG_DONT_CHANGE_BINDING
+                |HWLOC_TOPOLOGY_FLAG_NO_DISTANCES
+                |HWLOC_TOPOLOGY_FLAG_NO_MEMATTRS
+                |HWLOC_TOPOLOGY_FLAG_NO_CPUKINDS)) {
     errno = EINVAL;
     return -1;
   }
@@ -4027,15 +4100,11 @@ hwloc_topology_load (struct hwloc_topology *topology)
    */
   hwloc_set_binding_hooks(topology);
 
-  hwloc_pci_discovery_prepare(topology);
-
   /* actual topology discovery */
   err = hwloc_discover(topology, &dstatus);
   if (err < 0)
     goto out;
 
-  hwloc_pci_discovery_exit(topology);
-
 #ifndef HWLOC_DEBUG
   if (getenv("HWLOC_DEBUG_CHECK"))
 #endif
@@ -4057,6 +4126,7 @@ hwloc_topology_load (struct hwloc_topology *topology)
   /* Same for memattrs */
   hwloc_internal_memattrs_need_refresh(topology);
   hwloc_internal_memattrs_refresh(topology);
+  hwloc_internal_memattrs_guess_memory_tiers(topology);
 
   topology->is_loaded = 1;
 
@@ -4360,14 +4430,13 @@ hwloc_topology_restrict(struct hwloc_topology *topology, hwloc_const_bitmap_t se
   hwloc_bitmap_free(droppedcpuset);
   hwloc_bitmap_free(droppednodeset);
 
-  if (hwloc_topology_reconnect(topology, 0) < 0)
+  if (hwloc_filter_levels_keep_structure(topology) < 0) /* takes care of reconnecting internally */
     goto out;
 
   /* some objects may have disappeared, we need to update distances objs arrays */
   hwloc_internal_distances_invalidate_cached_objs(topology);
   hwloc_internal_memattrs_need_refresh(topology);
 
-  hwloc_filter_levels_keep_structure(topology);
   hwloc_propagate_symmetric_subtree(topology, topology->levels[0][0]);
   propagate_total_memory(topology->levels[0][0]);
   hwloc_internal_cpukinds_restrict(topology);

src/3rdparty/libethash/CMakeLists.txt

@@ -1,4 +1,4 @@
-cmake_minimum_required (VERSION 2.8.12)
+cmake_minimum_required(VERSION 3.1)
 project (ethash C)
 
 set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} -Os")

src/Summary.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 2017-2019 XMR-Stak    <https://github.com/fireice-uk>, <https://github.com/psychocrypt>
- * Copyright 2018-2021 SChernykh   <https://github.com/SChernykh>
- * Copyright 2016-2021 XMRig       <support@xmrig.com>
+ * Copyright (c) 2018-2022 SChernykh   <https://github.com/SChernykh>
+ * Copyright (c) 2016-2022 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
@@ -22,7 +16,6 @@
  *   along with this program. If not, see <http://www.gnu.org/licenses/>.
  */
 
-
 #include <cinttypes>
 #include <cstdio>
 #include <uv.h>
@@ -158,7 +151,7 @@ static void print_memory(const Config *config)
                        "", memory.id().data(), memory.size() / oneGiB, memory.type(), memory.speed() / 1000000ULL, memory.product().data());
         }
         else if (printEmpty) {
-            Log::print(WHITE_BOLD("   %-13s") "%s: " BLACK_BOLD("<empty>"), "", memory.slot().data());
+            Log::print(WHITE_BOLD("   %-13s") "%s: " BLACK_BOLD("<empty>"), "", memory.id().data());
         }
     }
 

src/Summary.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 2018-2021 SChernykh   <https://github.com/SChernykh>
- * Copyright 2016-2021 XMRig       <https://github.com/xmrig>, <support@xmrig.com>
+ * Copyright (c) 2018-2022 SChernykh   <https://github.com/SChernykh>
+ * Copyright (c) 2016-2022 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

src/backend/common/WorkerJob.h

@@ -109,7 +109,7 @@ private:
     }
 
 
-    alignas(16) uint8_t m_blobs[2][Job::kMaxBlobSize * N]{};
+    alignas(8) uint8_t m_blobs[2][Job::kMaxBlobSize * N]{};
     Job m_jobs[2];
     uint32_t m_rounds[2] = { 0, 0 };
     uint64_t m_nonce_mask[2] = { 0, 0 };

src/backend/cpu/Cpu.h

@@ -1,6 +1,6 @@
 /* XMRig
- * Copyright (c) 2018-2020 SChernykh   <https://github.com/SChernykh>
- * Copyright (c) 2016-2020 XMRig       <https://github.com/xmrig>, <support@xmrig.com>
+ * Copyright (c) 2018-2023 SChernykh   <https://github.com/SChernykh>
+ * Copyright (c) 2016-2023 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
@@ -37,7 +37,7 @@ public:
 };
 
 
-} /* namespace xmrig */
+} // namespace xmrig
 
 
-#endif /* XMRIG_CPU_H */
+#endif // XMRIG_CPU_H

src/backend/cpu/CpuBackend.cpp

@@ -436,10 +436,6 @@ rapidjson::Value xmrig::CpuBackend::toJSON(rapidjson::Document &doc) const
     out.AddMember("argon2-impl", argon2::Impl::name().toJSON(), allocator);
 #   endif
 
-#   ifdef XMRIG_ALGO_ASTROBWT
-    out.AddMember("astrobwt-max-size", cpu.astrobwtMaxSize(), allocator);
-#   endif
-
     out.AddMember("hugepages", d_ptr->hugePages(2, doc), allocator);
     out.AddMember("memory",    static_cast<uint64_t>(d_ptr->algo.isValid() ? (d_ptr->ways() * d_ptr->algo.l3()) : 0), allocator);
 

src/backend/cpu/CpuConfig.cpp

@@ -45,11 +45,6 @@ const char *CpuConfig::kAsm                 = "asm";
 const char *CpuConfig::kArgon2Impl          = "argon2-impl";
 #endif
 
-#ifdef XMRIG_ALGO_ASTROBWT
-const char *CpuConfig::kAstroBWTMaxSize     = "astrobwt-max-size";
-const char *CpuConfig::kAstroBWTAVX2        = "astrobwt-avx2";
-#endif
-
 
 extern template class Threads<CpuThreads>;
 
@@ -89,11 +84,6 @@ rapidjson::Value xmrig::CpuConfig::toJSON(rapidjson::Document &doc) const
     obj.AddMember(StringRef(kArgon2Impl), m_argon2Impl.toJSON(), allocator);
 #   endif
 
-#   ifdef XMRIG_ALGO_ASTROBWT
-    obj.AddMember(StringRef(kAstroBWTMaxSize),  m_astrobwtMaxSize, allocator);
-    obj.AddMember(StringRef(kAstroBWTAVX2),     m_astrobwtAVX2, allocator);
-#   endif
-
     m_threads.toJSON(obj, doc);
 
     return obj;
@@ -158,24 +148,6 @@ void xmrig::CpuConfig::read(const rapidjson::Value &value)
         m_argon2Impl = Json::getString(value, kArgon2Impl);
 #       endif
 
-#       ifdef XMRIG_ALGO_ASTROBWT
-        const auto& astroBWTMaxSize = Json::getValue(value, kAstroBWTMaxSize);
-        if (astroBWTMaxSize.IsNull() || !astroBWTMaxSize.IsInt()) {
-            m_shouldSave = true;
-        }
-        else {
-            m_astrobwtMaxSize = std::min(std::max(astroBWTMaxSize.GetInt(), 400), 1200);
-        }
-
-        const auto& astroBWTAVX2 = Json::getValue(value, kAstroBWTAVX2);
-        if (astroBWTAVX2.IsNull() || !astroBWTAVX2.IsBool()) {
-            m_shouldSave = true;
-        }
-        else {
-            m_astrobwtAVX2 = astroBWTAVX2.GetBool();
-        }
-#       endif
-
         m_threads.read(value);
 
         generate();
@@ -206,7 +178,6 @@ void xmrig::CpuConfig::generate()
     count += xmrig::generate<Algorithm::CN_FEMTO>(m_threads, m_limit);
     count += xmrig::generate<Algorithm::RANDOM_X>(m_threads, m_limit);
     count += xmrig::generate<Algorithm::ARGON2>(m_threads, m_limit);
-    count += xmrig::generate<Algorithm::ASTROBWT>(m_threads, m_limit);
     count += xmrig::generate<Algorithm::GHOSTRIDER>(m_threads, m_limit);
 
     m_shouldSave |= count > 0;

src/backend/cpu/CpuConfig.h

@@ -56,11 +56,6 @@ public:
     static const char *kArgon2Impl;
 #   endif
 
-#   ifdef XMRIG_ALGO_ASTROBWT
-    static const char *kAstroBWTMaxSize;
-    static const char *kAstroBWTAVX2;
-#   endif
-
     CpuConfig() = default;
 
     bool isHwAES() const;
@@ -69,7 +64,6 @@ public:
     std::vector<CpuLaunchData> get(const Miner *miner, const Algorithm &algorithm) const;
     void read(const rapidjson::Value &value);
 
-    inline bool astrobwtAVX2() const                    { return m_astrobwtAVX2; }
     inline bool isEnabled() const                       { return m_enabled; }
     inline bool isHugePages() const                     { return m_hugePageSize > 0; }
     inline bool isHugePagesJit() const                  { return m_hugePagesJit; }
@@ -78,7 +72,6 @@ public:
     inline const Assembly &assembly() const             { return m_assembly; }
     inline const String &argon2Impl() const             { return m_argon2Impl; }
     inline const Threads<CpuThreads> &threads() const   { return m_threads; }
-    inline int astrobwtMaxSize() const                  { return m_astrobwtMaxSize; }
     inline int priority() const                         { return m_priority; }
     inline size_t hugePageSize() const                  { return m_hugePageSize * 1024U; }
     inline uint32_t limit() const                       { return m_limit; }
@@ -96,12 +89,10 @@ private:
 
     AesMode m_aes           = AES_AUTO;
     Assembly m_assembly;
-    bool m_astrobwtAVX2     = false;
     bool m_enabled          = true;
     bool m_hugePagesJit     = false;
     bool m_shouldSave       = false;
     bool m_yield            = true;
-    int m_astrobwtMaxSize   = 550;
     int m_memoryPool        = 0;
     int m_priority          = -1;
     size_t m_hugePageSize   = kDefaultHugePageSizeKb;

src/backend/cpu/CpuConfig_gen.h

@@ -153,14 +153,6 @@ size_t inline generate<Algorithm::ARGON2>(Threads<CpuThreads> &threads, uint32_t
 #endif
 
 
-#ifdef XMRIG_ALGO_ASTROBWT
-template<>
-size_t inline generate<Algorithm::ASTROBWT>(Threads<CpuThreads>& threads, uint32_t limit)
-{
-    return generate(Algorithm::kASTROBWT, threads, Algorithm::ASTROBWT_DERO, limit);
-}
-#endif
-
 #ifdef XMRIG_ALGO_GHOSTRIDER
 template<>
 size_t inline generate<Algorithm::GHOSTRIDER>(Threads<CpuThreads>& threads, uint32_t limit)

src/backend/cpu/CpuLaunchData.cpp

@@ -35,11 +35,9 @@
 xmrig::CpuLaunchData::CpuLaunchData(const Miner *miner, const Algorithm &algorithm, const CpuConfig &config, const CpuThread &thread, size_t threads, const std::vector<int64_t>& affinities) :
     algorithm(algorithm),
     assembly(config.assembly()),
-    astrobwtAVX2(config.astrobwtAVX2()),
     hugePages(config.isHugePages()),
     hwAES(config.isHwAES()),
     yield(config.isYield()),
-    astrobwtMaxSize(config.astrobwtMaxSize()),
     priority(config.priority()),
     affinity(thread.affinity()),
     miner(miner),

src/backend/cpu/CpuLaunchData.h

@@ -58,11 +58,9 @@ public:
 
     const Algorithm algorithm;
     const Assembly assembly;
-    const bool astrobwtAVX2;
     const bool hugePages;
     const bool hwAES;
     const bool yield;
-    const int astrobwtMaxSize;
     const int priority;
     const int64_t affinity;
     const Miner *miner;

src/backend/cpu/CpuWorker.cpp

@@ -33,6 +33,7 @@
 #include "crypto/common/Nonce.h"
 #include "crypto/common/VirtualMemory.h"
 #include "crypto/rx/Rx.h"
+#include "crypto/rx/RxCache.h"
 #include "crypto/rx/RxDataset.h"
 #include "crypto/rx/RxVm.h"
 #include "crypto/ghostrider/ghostrider.h"
@@ -44,11 +45,6 @@
 #endif
 
 
-#ifdef XMRIG_ALGO_ASTROBWT
-#   include "crypto/astrobwt/AstroBWT.h"
-#endif
-
-
 #ifdef XMRIG_FEATURE_BENCHMARK
 #   include "backend/common/benchmark/BenchState.h"
 #endif
@@ -73,19 +69,20 @@ xmrig::CpuWorker<N>::CpuWorker(size_t id, const CpuLaunchData &data) :
     Worker(id, data.affinity, data.priority),
     m_algorithm(data.algorithm),
     m_assembly(data.assembly),
-    m_astrobwtAVX2(data.astrobwtAVX2),
     m_hwAES(data.hwAES),
     m_yield(data.yield),
     m_av(data.av()),
-    m_astrobwtMaxSize(data.astrobwtMaxSize * 1000),
     m_miner(data.miner),
     m_threads(data.threads),
     m_ctx()
 {
 #   ifdef XMRIG_ALGO_CN_HEAVY
     // cn-heavy optimization for Zen3 CPUs
-    const bool is_vermeer = (Cpu::info()->arch() == ICpuInfo::ARCH_ZEN3) && (Cpu::info()->model() == 0x21);
-    if ((N == 1) && (m_av == CnHash::AV_SINGLE) && (m_algorithm.family() == Algorithm::CN_HEAVY) && (m_assembly != Assembly::NONE) && is_vermeer) {
+    const auto arch = Cpu::info()->arch();
+    const uint32_t model = Cpu::info()->model();
+    const bool is_vermeer = (arch == ICpuInfo::ARCH_ZEN3) && (model == 0x21);
+    const bool is_raphael = (arch == ICpuInfo::ARCH_ZEN4) && (model == 0x61);
+    if ((N == 1) && (m_av == CnHash::AV_SINGLE) && (m_algorithm.family() == Algorithm::CN_HEAVY) && (m_assembly != Assembly::NONE) && (is_vermeer || is_raphael)) {
         std::lock_guard<std::mutex> lock(cn_heavyZen3MemoryMutex);
         if (!cn_heavyZen3Memory) {
             // Round up number of threads to the multiple of 8
@@ -149,6 +146,11 @@ void xmrig::CpuWorker<N>::allocateRandomX_VM()
         uint8_t* scratchpad = m_memory->isHugePages() ? m_memory->scratchpad() : dataset->tryAllocateScrathpad();
         m_vm = RxVm::create(dataset, scratchpad ? scratchpad : m_memory->scratchpad(), !m_hwAES, m_assembly, node());
     }
+    else if (!dataset->get() && (m_job.currentJob().seed() != m_seed)) {
+        // Update RandomX light VM with the new seed
+        randomx_vm_set_cache(m_vm, dataset->cache()->get());
+    }
+    m_seed = m_job.currentJob().seed();
 }
 #endif
 
@@ -223,12 +225,6 @@ bool xmrig::CpuWorker<N>::selfTest()
     }
 #   endif
 
-#   ifdef XMRIG_ALGO_ASTROBWT
-    if (m_algorithm.family() == Algorithm::ASTROBWT) {
-        return verify(Algorithm::ASTROBWT_DERO, astrobwt_dero_test_out);
-    }
-#   endif
-
     return false;
 }
 
@@ -318,14 +314,6 @@ void xmrig::CpuWorker<N>::start()
             {
                 switch (job.algorithm().family()) {
 
-#               ifdef XMRIG_ALGO_ASTROBWT
-                case Algorithm::ASTROBWT:
-                    if (!astrobwt::astrobwt_dero(m_job.blob(), job.size(), m_ctx[0]->memory, m_hash, m_astrobwtMaxSize, m_astrobwtAVX2)) {
-                        valid = false;
-                    }
-                    break;
-#               endif
-
 #               ifdef XMRIG_ALGO_GHOSTRIDER
                 case Algorithm::GHOSTRIDER:
                     if (N == 8) {

src/backend/cpu/CpuWorker.h

@@ -83,14 +83,12 @@ private:
     void allocateCnCtx();
     void consumeJob();
 
-    alignas(16) uint8_t m_hash[N * 32]{ 0 };
+    alignas(8) uint8_t m_hash[N * 32]{ 0 };
     const Algorithm m_algorithm;
     const Assembly m_assembly;
-    const bool m_astrobwtAVX2;
     const bool m_hwAES;
     const bool m_yield;
     const CnHash::AlgoVariant m_av;
-    const int m_astrobwtMaxSize;
     const Miner *m_miner;
     const size_t m_threads;
     cryptonight_ctx *m_ctx[N];
@@ -99,6 +97,7 @@ private:
 
 #   ifdef XMRIG_ALGO_RANDOMX
     randomx_vm *m_vm        = nullptr;
+    Buffer m_seed;
 #   endif
 
 #   ifdef XMRIG_ALGO_GHOSTRIDER

src/backend/cpu/interfaces/ICpuInfo.h

@@ -1,6 +1,6 @@
 /* XMRig
- * Copyright (c) 2018-2021 SChernykh   <https://github.com/SChernykh>
- * Copyright (c) 2016-2021 XMRig       <support@xmrig.com>
+ * Copyright (c) 2018-2023 SChernykh   <https://github.com/SChernykh>
+ * Copyright (c) 2016-2023 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
@@ -26,6 +26,12 @@
 #include "crypto/common/Assembly.h"
 
 
+#ifdef XMRIG_FEATURE_HWLOC
+using hwloc_const_bitmap_t  = const struct hwloc_bitmap_s *;
+using hwloc_topology_t      = struct hwloc_topology *;
+#endif
+
+
 namespace xmrig {
 
 
@@ -45,19 +51,21 @@ public:
         ARCH_ZEN,
         ARCH_ZEN_PLUS,
         ARCH_ZEN2,
-        ARCH_ZEN3
+        ARCH_ZEN3,
+        ARCH_ZEN4
     };
 
     enum MsrMod : uint32_t {
         MSR_MOD_NONE,
         MSR_MOD_RYZEN_17H,
         MSR_MOD_RYZEN_19H,
+        MSR_MOD_RYZEN_19H_ZEN4,
         MSR_MOD_INTEL,
         MSR_MOD_CUSTOM,
         MSR_MOD_MAX
     };
 
-#   define MSR_NAMES_LIST "none", "ryzen_17h", "ryzen_19h", "intel", "custom"
+#   define MSR_NAMES_LIST "none", "ryzen_17h", "ryzen_19h", "ryzen_19h_zen4", "intel", "custom"
 
     enum Flag : uint32_t {
         FLAG_AES,
@@ -114,10 +122,16 @@ public:
     virtual size_t threads() const                                                  = 0;
     virtual Vendor vendor() const                                                   = 0;
     virtual uint32_t model() const                                                  = 0;
+
+#   ifdef XMRIG_FEATURE_HWLOC
+    virtual bool membind(hwloc_const_bitmap_t nodeset)                              = 0;
+    virtual const std::vector<uint32_t> &nodeset() const                            = 0;
+    virtual hwloc_topology_t topology() const                                       = 0;
+#   endif
 };
 
 
-} /* namespace xmrig */
+} // namespace xmrig
 
 
 #endif // XMRIG_CPUINFO_H