rocksdb/port/win/port_win.cc

304 lines
7.6 KiB
C++
Raw Normal View History

// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
// This source code is licensed under both the GPLv2 (found in the
// COPYING file in the root directory) and Apache 2.0 License
// (found in the LICENSE.Apache file in the root directory).
//
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#if defined(OS_WIN)
#include "port/win/port_win.h"
Built-in support for generating unique IDs, bug fix (#8708) Summary: Env::GenerateUniqueId() works fine on Windows and on POSIX where /proc/sys/kernel/random/uuid exists. Our other implementation is flawed and easily produces collision in a new multi-threaded test. As we rely more heavily on DB session ID uniqueness, this becomes a serious issue. This change combines several individually suitable entropy sources for reliable generation of random unique IDs, with goal of uniqueness and portability, not cryptographic strength nor maximum speed. Specifically: * Moves code for getting UUIDs from the OS to port::GenerateRfcUuid rather than in Env implementation details. Callers are now told whether the operation fails or succeeds. * Adds an internal API GenerateRawUniqueId for generating high-quality 128-bit unique identifiers, by combining entropy from three "tracks": * Lots of info from default Env like time, process id, and hostname. * std::random_device * port::GenerateRfcUuid (when working) * Built-in implementations of Env::GenerateUniqueId() will now always produce an RFC 4122 UUID string, either from platform-specific API or by converting the output of GenerateRawUniqueId. DB session IDs now use GenerateRawUniqueId while DB IDs (not as critical) try to use port::GenerateRfcUuid but fall back on GenerateRawUniqueId with conversion to an RFC 4122 UUID. GenerateRawUniqueId is declared and defined under env/ rather than util/ or even port/ because of the Env dependency. Likely follow-up: enhance GenerateRawUniqueId to be faster after the first call and to guarantee uniqueness within the lifetime of a single process (imparting the same property onto DB session IDs). Pull Request resolved: https://github.com/facebook/rocksdb/pull/8708 Test Plan: A new mini-stress test in env_test checks the various public and internal APIs for uniqueness, including each track of GenerateRawUniqueId individually. We can't hope to verify anywhere close to 128 bits of entropy, but it can at least detect flaws as bad as the old code. Serial execution of the new tests takes about 350 ms on my machine. Reviewed By: zhichao-cao, mrambacher Differential Revision: D30563780 Pulled By: pdillinger fbshipit-source-id: de4c9ff4b2f581cf784fcedb5f39f16e5185c364
2021-08-30 22:19:39 +00:00
#include <assert.h>
#include <io.h>
Built-in support for generating unique IDs, bug fix (#8708) Summary: Env::GenerateUniqueId() works fine on Windows and on POSIX where /proc/sys/kernel/random/uuid exists. Our other implementation is flawed and easily produces collision in a new multi-threaded test. As we rely more heavily on DB session ID uniqueness, this becomes a serious issue. This change combines several individually suitable entropy sources for reliable generation of random unique IDs, with goal of uniqueness and portability, not cryptographic strength nor maximum speed. Specifically: * Moves code for getting UUIDs from the OS to port::GenerateRfcUuid rather than in Env implementation details. Callers are now told whether the operation fails or succeeds. * Adds an internal API GenerateRawUniqueId for generating high-quality 128-bit unique identifiers, by combining entropy from three "tracks": * Lots of info from default Env like time, process id, and hostname. * std::random_device * port::GenerateRfcUuid (when working) * Built-in implementations of Env::GenerateUniqueId() will now always produce an RFC 4122 UUID string, either from platform-specific API or by converting the output of GenerateRawUniqueId. DB session IDs now use GenerateRawUniqueId while DB IDs (not as critical) try to use port::GenerateRfcUuid but fall back on GenerateRawUniqueId with conversion to an RFC 4122 UUID. GenerateRawUniqueId is declared and defined under env/ rather than util/ or even port/ because of the Env dependency. Likely follow-up: enhance GenerateRawUniqueId to be faster after the first call and to guarantee uniqueness within the lifetime of a single process (imparting the same property onto DB session IDs). Pull Request resolved: https://github.com/facebook/rocksdb/pull/8708 Test Plan: A new mini-stress test in env_test checks the various public and internal APIs for uniqueness, including each track of GenerateRawUniqueId individually. We can't hope to verify anywhere close to 128 bits of entropy, but it can at least detect flaws as bad as the old code. Serial execution of the new tests takes about 350 ms on my machine. Reviewed By: zhichao-cao, mrambacher Differential Revision: D30563780 Pulled By: pdillinger fbshipit-source-id: de4c9ff4b2f581cf784fcedb5f39f16e5185c364
2021-08-30 22:19:39 +00:00
#include <rpc.h>
#include <stdio.h>
#include <string.h>
#include <chrono>
Built-in support for generating unique IDs, bug fix (#8708) Summary: Env::GenerateUniqueId() works fine on Windows and on POSIX where /proc/sys/kernel/random/uuid exists. Our other implementation is flawed and easily produces collision in a new multi-threaded test. As we rely more heavily on DB session ID uniqueness, this becomes a serious issue. This change combines several individually suitable entropy sources for reliable generation of random unique IDs, with goal of uniqueness and portability, not cryptographic strength nor maximum speed. Specifically: * Moves code for getting UUIDs from the OS to port::GenerateRfcUuid rather than in Env implementation details. Callers are now told whether the operation fails or succeeds. * Adds an internal API GenerateRawUniqueId for generating high-quality 128-bit unique identifiers, by combining entropy from three "tracks": * Lots of info from default Env like time, process id, and hostname. * std::random_device * port::GenerateRfcUuid (when working) * Built-in implementations of Env::GenerateUniqueId() will now always produce an RFC 4122 UUID string, either from platform-specific API or by converting the output of GenerateRawUniqueId. DB session IDs now use GenerateRawUniqueId while DB IDs (not as critical) try to use port::GenerateRfcUuid but fall back on GenerateRawUniqueId with conversion to an RFC 4122 UUID. GenerateRawUniqueId is declared and defined under env/ rather than util/ or even port/ because of the Env dependency. Likely follow-up: enhance GenerateRawUniqueId to be faster after the first call and to guarantee uniqueness within the lifetime of a single process (imparting the same property onto DB session IDs). Pull Request resolved: https://github.com/facebook/rocksdb/pull/8708 Test Plan: A new mini-stress test in env_test checks the various public and internal APIs for uniqueness, including each track of GenerateRawUniqueId individually. We can't hope to verify anywhere close to 128 bits of entropy, but it can at least detect flaws as bad as the old code. Serial execution of the new tests takes about 350 ms on my machine. Reviewed By: zhichao-cao, mrambacher Differential Revision: D30563780 Pulled By: pdillinger fbshipit-source-id: de4c9ff4b2f581cf784fcedb5f39f16e5185c364
2021-08-30 22:19:39 +00:00
#include <cstdlib>
#include <exception>
#include <memory>
#include "port/port_dirent.h"
#include "port/sys_time.h"
#ifdef ROCKSDB_WINDOWS_UTF8_FILENAMES
// utf8 <-> utf16
#include <codecvt>
#include <locale>
#include <string>
#endif
#include "logging/logging.h"
namespace ROCKSDB_NAMESPACE {
extern const bool kDefaultToAdaptiveMutex = false;
namespace port {
#ifdef ROCKSDB_WINDOWS_UTF8_FILENAMES
std::string utf16_to_utf8(const std::wstring& utf16) {
std::wstring_convert<std::codecvt_utf8_utf16<wchar_t>, wchar_t> convert;
return convert.to_bytes(utf16);
}
std::wstring utf8_to_utf16(const std::string& utf8) {
std::wstring_convert<std::codecvt_utf8_utf16<wchar_t>> converter;
return converter.from_bytes(utf8);
}
#endif
void GetTimeOfDay(TimeVal* tv, struct timezone* /* tz */) {
std::chrono::microseconds usNow(
std::chrono::duration_cast<std::chrono::microseconds>(
std::chrono::system_clock::now().time_since_epoch()));
std::chrono::seconds secNow(
std::chrono::duration_cast<std::chrono::seconds>(usNow));
tv->tv_sec = static_cast<long>(secNow.count());
tv->tv_usec = static_cast<long>(
usNow.count() -
std::chrono::duration_cast<std::chrono::microseconds>(secNow).count());
}
Mutex::~Mutex() {}
CondVar::~CondVar() {}
void CondVar::Wait() {
2015-10-12 22:41:20 +00:00
// Caller must ensure that mutex is held prior to calling this method
std::unique_lock<std::mutex> lk(mu_->getLock(), std::adopt_lock);
#ifndef NDEBUG
mu_->locked_ = false;
#endif
2015-10-12 22:41:20 +00:00
cv_.wait(lk);
#ifndef NDEBUG
mu_->locked_ = true;
#endif
2015-10-12 22:41:20 +00:00
// Release ownership of the lock as we don't want it to be unlocked when
// it goes out of scope (as we adopted the lock and didn't lock it ourselves)
lk.release();
}
bool CondVar::TimedWait(uint64_t abs_time_us) {
// MSVC++ library implements wait_until in terms of wait_for so
// we need to convert absolute wait into relative wait.
std::chrono::microseconds usAbsTime(abs_time_us);
std::chrono::microseconds usNow(
std::chrono::duration_cast<std::chrono::microseconds>(
std::chrono::system_clock::now().time_since_epoch()));
std::chrono::microseconds relTimeUs = (usAbsTime > usNow)
? (usAbsTime - usNow)
: std::chrono::microseconds::zero();
2015-10-12 22:41:20 +00:00
// Caller must ensure that mutex is held prior to calling this method
std::unique_lock<std::mutex> lk(mu_->getLock(), std::adopt_lock);
Fix MSVC-related build issues (#7439) Summary: This PR addresses some build and functional issues on MSVC targets, as a step towards an eventual goal of having RocksDB build successfully for Windows on ARM64. Addressed issues include: - BitsSetToOne and CountTrailingZeroBits do not compile on non-x64 MSVC targets. A fallback implementation of BitsSetToOne when Intel intrinsics are not available is added, based on the C++20 `<bit>` popcount implementation in Microsoft's STL. - The implementation of FloorLog2 for MSVC targets (including x64) gives incorrect results. The unit test easily detects this, but CircleCI is currently configured to only run a specific set of tests for Windows CMake builds, so this seems to have been unnoticed. - AsmVolatilePause does not use YieldProcessor on Windows ARM64 targets, even though it is available. - When CondVar::TimedWait calls Microsoft STL's condition_variable::wait_for, it can potentially trigger a bug (just recently fixed in the upcoming VS 16.8's STL) that deadlocks various tests that wait for a timer to execute, since `Timer::Run` doesn't get a chance to execute before being blocked by the test function acquiring the mutex. - In c_test, `GetTempDir` assumes a POSIX-style temp path. - `NormalizePath` did not eliminate consecutive POSIX-style path separators on Windows, resulting in test failures in e.g., wal_manager_test. - Various other test failures. In a followup PR I hope to modify CircleCI's config.yml to invoke all RocksDB unit tests in Windows CMake builds with CTest, instead of the current use of `run_ci_db_test.ps1` which requires individual tests to be specified and is missing many of the existing tests. Notes from peterd: FloorLog2 is not yet used in production code (it's for something in progress). I also added a few more inexpensive platform-dependent tests to Windows CircleCI runs. And included facebook/folly#1461 as requested Pull Request resolved: https://github.com/facebook/rocksdb/pull/7439 Reviewed By: jay-zhuang Differential Revision: D24021563 Pulled By: pdillinger fbshipit-source-id: 0ec2027c0d6a494d8a0fe38d9667fc2f7e29f7e7
2020-10-01 16:21:30 +00:00
// Work around https://github.com/microsoft/STL/issues/369
#if defined(_MSC_VER) && \
(!defined(_MSVC_STL_UPDATE) || _MSVC_STL_UPDATE < 202008L)
if (relTimeUs == std::chrono::microseconds::zero()) {
Fix MSVC-related build issues (#7439) Summary: This PR addresses some build and functional issues on MSVC targets, as a step towards an eventual goal of having RocksDB build successfully for Windows on ARM64. Addressed issues include: - BitsSetToOne and CountTrailingZeroBits do not compile on non-x64 MSVC targets. A fallback implementation of BitsSetToOne when Intel intrinsics are not available is added, based on the C++20 `<bit>` popcount implementation in Microsoft's STL. - The implementation of FloorLog2 for MSVC targets (including x64) gives incorrect results. The unit test easily detects this, but CircleCI is currently configured to only run a specific set of tests for Windows CMake builds, so this seems to have been unnoticed. - AsmVolatilePause does not use YieldProcessor on Windows ARM64 targets, even though it is available. - When CondVar::TimedWait calls Microsoft STL's condition_variable::wait_for, it can potentially trigger a bug (just recently fixed in the upcoming VS 16.8's STL) that deadlocks various tests that wait for a timer to execute, since `Timer::Run` doesn't get a chance to execute before being blocked by the test function acquiring the mutex. - In c_test, `GetTempDir` assumes a POSIX-style temp path. - `NormalizePath` did not eliminate consecutive POSIX-style path separators on Windows, resulting in test failures in e.g., wal_manager_test. - Various other test failures. In a followup PR I hope to modify CircleCI's config.yml to invoke all RocksDB unit tests in Windows CMake builds with CTest, instead of the current use of `run_ci_db_test.ps1` which requires individual tests to be specified and is missing many of the existing tests. Notes from peterd: FloorLog2 is not yet used in production code (it's for something in progress). I also added a few more inexpensive platform-dependent tests to Windows CircleCI runs. And included facebook/folly#1461 as requested Pull Request resolved: https://github.com/facebook/rocksdb/pull/7439 Reviewed By: jay-zhuang Differential Revision: D24021563 Pulled By: pdillinger fbshipit-source-id: 0ec2027c0d6a494d8a0fe38d9667fc2f7e29f7e7
2020-10-01 16:21:30 +00:00
lk.unlock();
lk.lock();
}
#endif
2015-10-12 22:41:20 +00:00
#ifndef NDEBUG
mu_->locked_ = false;
#endif
std::cv_status cvStatus = cv_.wait_for(lk, relTimeUs);
#ifndef NDEBUG
mu_->locked_ = true;
#endif
2015-10-12 22:41:20 +00:00
// Release ownership of the lock as we don't want it to be unlocked when
// it goes out of scope (as we adopted the lock and didn't lock it ourselves)
lk.release();
if (cvStatus == std::cv_status::timeout) {
return true;
}
return false;
}
void CondVar::Signal() { cv_.notify_one(); }
void CondVar::SignalAll() { cv_.notify_all(); }
support for concurrent adds to memtable Summary: This diff adds support for concurrent adds to the skiplist memtable implementations. Memory allocation is made thread-safe by the addition of a spinlock, with small per-core buffers to avoid contention. Concurrent memtable writes are made via an additional method and don't impose a performance overhead on the non-concurrent case, so parallelism can be selected on a per-batch basis. Write thread synchronization is an increasing bottleneck for higher levels of concurrency, so this diff adds --enable_write_thread_adaptive_yield (default off). This feature causes threads joining a write batch group to spin for a short time (default 100 usec) using sched_yield, rather than going to sleep on a mutex. If the timing of the yield calls indicates that another thread has actually run during the yield then spinning is avoided. This option improves performance for concurrent situations even without parallel adds, although it has the potential to increase CPU usage (and the heuristic adaptation is not yet mature). Parallel writes are not currently compatible with inplace updates, update callbacks, or delete filtering. Enable it with --allow_concurrent_memtable_write (and --enable_write_thread_adaptive_yield). Parallel memtable writes are performance neutral when there is no actual parallelism, and in my experiments (SSD server-class Linux and varying contention and key sizes for fillrandom) they are always a performance win when there is more than one thread. Statistics are updated earlier in the write path, dropping the number of DB mutex acquisitions from 2 to 1 for almost all cases. This diff was motivated and inspired by Yahoo's cLSM work. It is more conservative than cLSM: RocksDB's write batch group leader role is preserved (along with all of the existing flush and write throttling logic) and concurrent writers are blocked until all memtable insertions have completed and the sequence number has been advanced, to preserve linearizability. My test config is "db_bench -benchmarks=fillrandom -threads=$T -batch_size=1 -memtablerep=skip_list -value_size=100 --num=1000000/$T -level0_slowdown_writes_trigger=9999 -level0_stop_writes_trigger=9999 -disable_auto_compactions --max_write_buffer_number=8 -max_background_flushes=8 --disable_wal --write_buffer_size=160000000 --block_size=16384 --allow_concurrent_memtable_write" on a two-socket Xeon E5-2660 @ 2.2Ghz with lots of memory and an SSD hard drive. With 1 thread I get ~440Kops/sec. Peak performance for 1 socket (numactl -N1) is slightly more than 1Mops/sec, at 16 threads. Peak performance across both sockets happens at 30 threads, and is ~900Kops/sec, although with fewer threads there is less performance loss when the system has background work. Test Plan: 1. concurrent stress tests for InlineSkipList and DynamicBloom 2. make clean; make check 3. make clean; DISABLE_JEMALLOC=1 make valgrind_check; valgrind db_bench 4. make clean; COMPILE_WITH_TSAN=1 make all check; db_bench 5. make clean; COMPILE_WITH_ASAN=1 make all check; db_bench 6. make clean; OPT=-DROCKSDB_LITE make check 7. verify no perf regressions when disabled Reviewers: igor, sdong Reviewed By: sdong Subscribers: MarkCallaghan, IslamAbdelRahman, anthony, yhchiang, rven, sdong, guyg8, kradhakrishnan, dhruba Differential Revision: https://reviews.facebook.net/D50589
2015-08-14 23:59:07 +00:00
int PhysicalCoreID() { return GetCurrentProcessorNumber(); }
void InitOnce(OnceType* once, void (*initializer)()) {
2015-11-20 23:31:47 +00:00
std::call_once(once->flag_, initializer);
}
// Private structure, exposed only by pointer
struct DIR {
HANDLE handle_;
bool firstread_;
RX_WIN32_FIND_DATA data_;
dirent entry_;
DIR() : handle_(INVALID_HANDLE_VALUE), firstread_(true) {}
DIR(const DIR&) = delete;
DIR& operator=(const DIR&) = delete;
~DIR() {
if (INVALID_HANDLE_VALUE != handle_) {
::FindClose(handle_);
}
}
};
DIR* opendir(const char* name) {
if (!name || *name == 0) {
errno = ENOENT;
return nullptr;
}
std::string pattern(name);
pattern.append("\\").append("*");
std::unique_ptr<DIR> dir(new DIR);
dir->handle_ =
RX_FindFirstFileEx(RX_FN(pattern).c_str(),
FindExInfoBasic, // Do not want alternative name
&dir->data_, FindExSearchNameMatch,
NULL, // lpSearchFilter
0);
if (dir->handle_ == INVALID_HANDLE_VALUE) {
return nullptr;
}
RX_FILESTRING x(dir->data_.cFileName, RX_FNLEN(dir->data_.cFileName));
strcpy_s(dir->entry_.d_name, sizeof(dir->entry_.d_name), FN_TO_RX(x).c_str());
return dir.release();
}
struct dirent* readdir(DIR* dirp) {
if (!dirp || dirp->handle_ == INVALID_HANDLE_VALUE) {
errno = EBADF;
return nullptr;
}
if (dirp->firstread_) {
dirp->firstread_ = false;
return &dirp->entry_;
}
auto ret = RX_FindNextFile(dirp->handle_, &dirp->data_);
if (ret == 0) {
return nullptr;
}
RX_FILESTRING x(dirp->data_.cFileName, RX_FNLEN(dirp->data_.cFileName));
strcpy_s(dirp->entry_.d_name, sizeof(dirp->entry_.d_name),
FN_TO_RX(x).c_str());
return &dirp->entry_;
}
int closedir(DIR* dirp) {
delete dirp;
return 0;
}
int truncate(const char* path, int64_t length) {
if (path == nullptr) {
errno = EFAULT;
return -1;
}
return ROCKSDB_NAMESPACE::port::Truncate(path, length);
}
int Truncate(std::string path, int64_t len) {
if (len < 0) {
errno = EINVAL;
return -1;
}
HANDLE hFile =
RX_CreateFile(RX_FN(path).c_str(), GENERIC_READ | GENERIC_WRITE,
FILE_SHARE_READ | FILE_SHARE_WRITE | FILE_SHARE_DELETE,
NULL, // Security attrs
OPEN_EXISTING, // Truncate existing file only
FILE_ATTRIBUTE_NORMAL, NULL);
if (INVALID_HANDLE_VALUE == hFile) {
auto lastError = GetLastError();
if (lastError == ERROR_FILE_NOT_FOUND) {
errno = ENOENT;
} else if (lastError == ERROR_ACCESS_DENIED) {
errno = EACCES;
} else {
errno = EIO;
}
return -1;
}
int result = 0;
FILE_END_OF_FILE_INFO end_of_file;
end_of_file.EndOfFile.QuadPart = len;
if (!SetFileInformationByHandle(hFile, FileEndOfFileInfo, &end_of_file,
sizeof(FILE_END_OF_FILE_INFO))) {
errno = EIO;
result = -1;
}
CloseHandle(hFile);
return result;
}
void Crash(const std::string& srcfile, int srcline) {
fprintf(stdout, "Crashing at %s:%d\n", srcfile.c_str(), srcline);
fflush(stdout);
abort();
}
int GetMaxOpenFiles() { return -1; }
// Assume 4KB page size
const size_t kPageSize = 4U * 1024U;
void SetCpuPriority(ThreadId id, CpuPriority priority) {
// Not supported
(void)id;
(void)priority;
}
int64_t GetProcessID() { return GetCurrentProcessId(); }
Built-in support for generating unique IDs, bug fix (#8708) Summary: Env::GenerateUniqueId() works fine on Windows and on POSIX where /proc/sys/kernel/random/uuid exists. Our other implementation is flawed and easily produces collision in a new multi-threaded test. As we rely more heavily on DB session ID uniqueness, this becomes a serious issue. This change combines several individually suitable entropy sources for reliable generation of random unique IDs, with goal of uniqueness and portability, not cryptographic strength nor maximum speed. Specifically: * Moves code for getting UUIDs from the OS to port::GenerateRfcUuid rather than in Env implementation details. Callers are now told whether the operation fails or succeeds. * Adds an internal API GenerateRawUniqueId for generating high-quality 128-bit unique identifiers, by combining entropy from three "tracks": * Lots of info from default Env like time, process id, and hostname. * std::random_device * port::GenerateRfcUuid (when working) * Built-in implementations of Env::GenerateUniqueId() will now always produce an RFC 4122 UUID string, either from platform-specific API or by converting the output of GenerateRawUniqueId. DB session IDs now use GenerateRawUniqueId while DB IDs (not as critical) try to use port::GenerateRfcUuid but fall back on GenerateRawUniqueId with conversion to an RFC 4122 UUID. GenerateRawUniqueId is declared and defined under env/ rather than util/ or even port/ because of the Env dependency. Likely follow-up: enhance GenerateRawUniqueId to be faster after the first call and to guarantee uniqueness within the lifetime of a single process (imparting the same property onto DB session IDs). Pull Request resolved: https://github.com/facebook/rocksdb/pull/8708 Test Plan: A new mini-stress test in env_test checks the various public and internal APIs for uniqueness, including each track of GenerateRawUniqueId individually. We can't hope to verify anywhere close to 128 bits of entropy, but it can at least detect flaws as bad as the old code. Serial execution of the new tests takes about 350 ms on my machine. Reviewed By: zhichao-cao, mrambacher Differential Revision: D30563780 Pulled By: pdillinger fbshipit-source-id: de4c9ff4b2f581cf784fcedb5f39f16e5185c364
2021-08-30 22:19:39 +00:00
bool GenerateRfcUuid(std::string* output) {
UUID uuid;
UuidCreateSequential(&uuid);
RPC_CSTR rpc_str;
auto status = UuidToStringA(&uuid, &rpc_str);
if (status != RPC_S_OK) {
return false;
}
// rpc_str is nul-terminated
*output = reinterpret_cast<char*>(rpc_str);
status = RpcStringFreeA(&rpc_str);
assert(status == RPC_S_OK);
return true;
}
} // namespace port
} // namespace ROCKSDB_NAMESPACE
#endif