utils.cpp

/*******************************************************************************
* Copyright 2018-2024 Intel Corporation
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
*     http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*******************************************************************************/

#ifdef _WIN32
#include <malloc.h>
#include <windows.h>
#endif

#if defined __unix__ || defined __APPLE__ || defined __FreeBSD__ \
        || defined __Fuchsia__
#include <unistd.h>
#endif

#ifdef __unix__
#include <sys/stat.h>
#include <sys/types.h>
#endif

#include <algorithm>
#include <climits>
#include <cstdio>
#include <cstdlib>
#include <cstring>
#include <mutex>
#include <string>

#include "oneapi/dnnl/dnnl.h"

#include "memory_debug.hpp"
#include "utils.hpp"

#if DNNL_CPU_RUNTIME != DNNL_RUNTIME_NONE
#include "cpu/platform.hpp"
#endif

namespace dnnl {
namespace impl {

int getenv(const char *name, char *buffer, int buffer_size) {
    if (name == nullptr || buffer_size < 0
            || (buffer == nullptr && buffer_size > 0))
        return INT_MIN;

    int result = 0;
    int term_zero_idx = 0;
    size_t value_length = 0;

#ifdef _WIN32
    value_length = GetEnvironmentVariable(name, buffer, buffer_size);
#else
    const char *value = ::getenv(name);
    value_length = value == nullptr ? 0 : strlen(value);
#endif

    if (value_length > INT_MAX)
        result = INT_MIN;
    else {
        int int_value_length = (int)value_length;
        if (int_value_length >= buffer_size) {
            result = -int_value_length;
        } else {
            term_zero_idx = int_value_length;
            result = int_value_length;
#ifndef _WIN32
            if (value) strncpy(buffer, value, buffer_size - 1);
#endif
        }
    }

    if (buffer != nullptr) buffer[term_zero_idx] = '\0';
    return result;
}

int getenv_int(const char *name, int default_value) {
    int value = default_value;
    // # of digits in the longest 32-bit signed int + sign + terminating null
    const int len = 12;
    char value_str[len];
    if (getenv(name, value_str, len) > 0) value = atoi(value_str);
    return value;
}

int getenv_int_user(const char *name, int default_value) {
    int value = default_value;
    // # of digits in the longest 32-bit signed int + sign + terminating null
    const int len = 12;
    char value_str[len];
    for (const auto &prefix : {"ONEDNN_", "DNNL_"}) {
        std::string name_str = std::string(prefix) + std::string(name);
        if (getenv(name_str.c_str(), value_str, len) > 0) {
            value = atoi(value_str);
            break;
        }
    }
    return value;
}

std::string getenv_string_user(const char *name) {
    // Random number to fit possible string input.
    std::string value;
    const int len = 128;
    char value_str[len];
    for (const auto &prefix : {"ONEDNN_", "DNNL_"}) {
        std::string name_str = std::string(prefix) + std::string(name);
        if (getenv(name_str.c_str(), value_str, len) > 0) {
            value = value_str;
            break;
        }
    }
    std::transform(value.begin(), value.end(), value.begin(), ::tolower);
    return value;
}

FILE *fopen(const char *filename, const char *mode) {
#ifdef _WIN32
    FILE *fp = NULL;
    return ::fopen_s(&fp, filename, mode) ? NULL : fp;
#else
    return ::fopen(filename, mode);
#endif
}

int getpagesize() {
#ifdef _WIN32
    SYSTEM_INFO info;
    GetSystemInfo(&info);
    return info.dwPageSize;
#else
    return ::getpagesize();
#endif
}

void *malloc(size_t size, int alignment) {
    void *ptr;
    if (memory_debug::is_mem_debug())
        return memory_debug::malloc(size, alignment);

#ifdef _WIN32
    ptr = _aligned_malloc(size, alignment);
    int rc = ptr ? 0 : -1;
#else
    int rc = ::posix_memalign(&ptr, alignment, size);
#endif

    return (rc == 0) ? ptr : nullptr;
}

void free(void *p) {

    if (memory_debug::is_mem_debug()) return memory_debug::free(p);

#ifdef _WIN32
    _aligned_free(p);
#else
    ::free(p);
#endif
}

// Atomic operations
int32_t fetch_and_add(int32_t *dst, int32_t val) {
#ifdef _WIN32
    return InterlockedExchangeAdd(reinterpret_cast<long *>(dst), val);
#else
    return __sync_fetch_and_add(dst, val);
#endif
}

static setting_t<bool> jit_dump {false};
bool get_jit_dump() {
    if (!jit_dump.initialized()) {
        static bool val = getenv_int_user("JIT_DUMP", jit_dump.get());
        jit_dump.set(val);
    }
    return jit_dump.get();
}

#if defined(DNNL_AARCH64) && (DNNL_AARCH64 == 1)
static setting_t<unsigned> jit_profiling_flags {DNNL_JIT_PROFILE_LINUX_PERFMAP};
#else
static setting_t<unsigned> jit_profiling_flags {DNNL_JIT_PROFILE_VTUNE};
#endif
unsigned get_jit_profiling_flags() {
    MAYBE_UNUSED(jit_profiling_flags);
    unsigned flag = 0;
#if DNNL_CPU_RUNTIME != DNNL_RUNTIME_NONE
    if (!jit_profiling_flags.initialized()) {
        static unsigned val
                = getenv_int_user("JIT_PROFILE", jit_profiling_flags.get());
        jit_profiling_flags.set(val);
    }
    flag = jit_profiling_flags.get();
#endif
    return flag;
}

static setting_t<std::string> jit_profiling_jitdumpdir;
dnnl_status_t init_jit_profiling_jitdumpdir(
        const char *jitdumpdir, bool overwrite) {
#ifdef __linux__
    static std::mutex m;
    std::lock_guard<std::mutex> g(m);

    if (jit_profiling_jitdumpdir.initialized() && !overwrite)
        return status::success;

    if (!jitdumpdir) {
        char buf[PATH_MAX];
        if (getenv("JITDUMPDIR", buf, sizeof(buf)) > 0)
            jit_profiling_jitdumpdir.set(buf);
        else if (getenv("HOME", buf, sizeof(buf)) > 0)
            jit_profiling_jitdumpdir.set(buf);
        else
            jit_profiling_jitdumpdir.set(".");
    } else
        jit_profiling_jitdumpdir.set(jitdumpdir);

    return status::success;
#else
    UNUSED(jit_profiling_jitdumpdir);
    return status::unimplemented;
#endif
}

std::string get_jit_profiling_jitdumpdir() {
    std::string jitdumpdir;
#if DNNL_CPU_RUNTIME != DNNL_RUNTIME_NONE
    if (!jit_profiling_jitdumpdir.initialized()) {
        auto status = init_jit_profiling_jitdumpdir(nullptr, false);
        if (status != status::success) return std::string();
    }
    jitdumpdir = jit_profiling_jitdumpdir.get();
#endif
    return jitdumpdir;
}

bool is_destroying_cache_safe() {
#if defined(_WIN32) \
        && (defined(DNNL_WITH_SYCL) || DNNL_GPU_RUNTIME == DNNL_RUNTIME_OCL)
    // The ntdll.dll library is located in system32, therefore setting
    // additional environment is not required.
    HMODULE handle = LoadLibraryExA(
            "ntdll.dll", nullptr, LOAD_LIBRARY_SEARCH_SYSTEM32);
    if (!handle) { return false; }

    // RtlDllShutdownInProgress returns TRUE if the whole process terminates
    // and FALSE if DLL is being unloaded dynamically or if it’s called from
    // an executable.
    auto f = reinterpret_cast<BOOLEAN (*)(void)>(
            GetProcAddress(handle, "RtlDllShutdownInProgress"));
    if (!f) {
        auto ret = FreeLibrary(handle);
        assert(ret);
        MAYBE_UNUSED(ret);
        return false;
    }

    bool is_process_termination_in_progress = f();

    auto ret = FreeLibrary(handle);
    assert(ret);
    MAYBE_UNUSED(ret);

    if (is_process_termination_in_progress) {
        return false;
    } else {
        // Three scenarios possible:
        //    1. oneDNN is being dynamically unloaded
        //    2. Another dynamic library that contains statically linked
        //       oneDNN is dynamically unloaded
        //    3. oneDNN is statically linked in an executable which is done
        //       and now the process terminates In all these scenarios
        //       content of the cache can be safely destroyed.
        //
        // Note: the first 2 scenarios can still cause some issues in the
        // case of OpenCL runtime. There doesn't seem to be a way to distinguish
        // the 2 scenarios from the 3rd one therefore it's always
        // considered unsafe to clean up the cache.
#if DNNL_GPU_RUNTIME == DNNL_RUNTIME_OCL
        return false;
#endif
        return true;
    }
#else
    // Always destroy the content of the cache for non-Windows OSes, and
    // non-sycl and non-ocl runtimes because there is no a problem with
    // library unloading order in such cases.
    return true;
#endif
}

} // namespace impl
} // namespace dnnl

dnnl_status_t dnnl_set_jit_dump(int enabled) {
    using namespace dnnl::impl;
    jit_dump.set(enabled);
    return status::success;
}

dnnl_status_t dnnl_set_jit_profiling_flags(unsigned flags) {
    using namespace dnnl::impl;
#if DNNL_CPU_RUNTIME != DNNL_RUNTIME_NONE
    unsigned mask = DNNL_JIT_PROFILE_VTUNE;
#ifdef __linux__
    mask |= DNNL_JIT_PROFILE_LINUX_PERF;
    mask |= DNNL_JIT_PROFILE_LINUX_JITDUMP_USE_TSC;
#endif
    if (flags & ~mask) return status::invalid_arguments;
    jit_profiling_flags.set(flags);
    return status::success;
#else
    return status::unimplemented;
#endif
}

dnnl_status_t dnnl_set_jit_profiling_jitdumpdir(const char *dir) {
    auto status = dnnl::impl::status::unimplemented;
#if DNNL_CPU_RUNTIME != DNNL_RUNTIME_NONE
    status = dnnl::impl::init_jit_profiling_jitdumpdir(dir, true);
#endif
    return status;
}

dnnl_status_t dnnl_set_max_cpu_isa(dnnl_cpu_isa_t isa) {
    auto status = dnnl::impl::status::runtime_error;
#if DNNL_CPU_RUNTIME != DNNL_RUNTIME_NONE
    status = dnnl::impl::cpu::platform::set_max_cpu_isa(isa);
#endif
    return status;
}

dnnl_cpu_isa_t dnnl_get_effective_cpu_isa() {
    auto isa = dnnl_cpu_isa_default;
#if DNNL_CPU_RUNTIME != DNNL_RUNTIME_NONE
    isa = dnnl::impl::cpu::platform::get_effective_cpu_isa();
#endif
    return isa;
}

dnnl_status_t dnnl_set_cpu_isa_hints(dnnl_cpu_isa_hints_t isa_hints) {
    auto status = dnnl::impl::status::runtime_error;
#if DNNL_CPU_RUNTIME != DNNL_RUNTIME_NONE
    status = dnnl::impl::cpu::platform::set_cpu_isa_hints(isa_hints);
#endif
    return status;
}

dnnl_cpu_isa_hints_t dnnl_get_cpu_isa_hints() {
    auto isa_hint = dnnl_cpu_isa_no_hints;
#if DNNL_CPU_RUNTIME != DNNL_RUNTIME_NONE
    isa_hint = dnnl::impl::cpu::platform::get_cpu_isa_hints();
#endif
    return isa_hint;
}

#if DNNL_CPU_THREADING_RUNTIME == DNNL_RUNTIME_THREADPOOL
#include "oneapi/dnnl/dnnl_threadpool_iface.hpp"
namespace dnnl {
namespace impl {
namespace threadpool_utils {

namespace {
thread_local dnnl::threadpool_interop::threadpool_iface *active_threadpool
        = nullptr;
}

void DNNL_API activate_threadpool(
        dnnl::threadpool_interop::threadpool_iface *tp) {
    // The assert here is put to prevent from activating a test threadpool while
    // the library one was left activating(not deactivated)
    assert(IMPLICATION(active_threadpool, active_threadpool == tp));
    if (!active_threadpool) active_threadpool = tp;
}

void DNNL_API deactivate_threadpool() {
    active_threadpool = nullptr;
}

dnnl::threadpool_interop::threadpool_iface *get_active_threadpool() {
    return active_threadpool;
}

int &get_threadlocal_max_concurrency() {
    thread_local int max_concurrency
            = (int)cpu::platform::get_max_threads_to_use();
    assert(max_concurrency > 0);
    return max_concurrency;
}

int DNNL_API get_max_concurrency() {
    return get_threadlocal_max_concurrency();
}

} // namespace threadpool_utils
} // namespace impl
} // namespace dnnl
#endif