2 * Copyright 2015 Facebook, Inc.
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
8 * http://www.apache.org/licenses/LICENSE-2.0
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
18 * Two Read-Write spin lock implementations.
20 * Ref: http://locklessinc.com/articles/locks
22 * Both locks here are faster than pthread_rwlock and have very low
23 * overhead (usually 20-30ns). They don't use any system mutexes and
24 * are very compact (4/8 bytes), so are suitable for per-instance
25 * based locking, particularly when contention is not expected.
27 * In most cases, RWSpinLock is a reasonable choice. It has minimal
28 * overhead, and comparable contention performance when the number of
29 * competing threads is less than or equal to the number of logical
30 * CPUs. Even as the number of threads gets larger, RWSpinLock can
31 * still be very competitive in READ, although it is slower on WRITE,
32 * and also inherently unfair to writers.
34 * RWTicketSpinLock shows more balanced READ/WRITE performance. If
35 * your application really needs a lot more threads, and a
36 * higher-priority writer, prefer one of the RWTicketSpinLock locks.
40 * RWTicketSpinLock locks can only be used with GCC on x86/x86-64
43 * RWTicketSpinLock<32> only allows up to 2^8 - 1 concurrent
44 * readers and writers.
46 * RWTicketSpinLock<64> only allows up to 2^16 - 1 concurrent
47 * readers and writers.
49 * RWTicketSpinLock<..., true> (kFavorWriter = true, that is, strict
50 * writer priority) is NOT reentrant, even for lock_shared().
52 * The lock will not grant any new shared (read) accesses while a thread
53 * attempting to acquire the lock in write mode is blocked. (That is,
54 * if the lock is held in shared mode by N threads, and a thread attempts
55 * to acquire it in write mode, no one else can acquire it in shared mode
56 * until these N threads release the lock and then the blocked thread
57 * acquires and releases the exclusive lock.) This also applies for
58 * attempts to reacquire the lock in shared mode by threads that already
59 * hold it in shared mode, making the lock non-reentrant.
61 * RWSpinLock handles 2^30 - 1 concurrent readers.
63 * @author Xin Liu <xliux@fb.com>
66 #ifndef FOLLY_RWSPINLOCK_H_
67 #define FOLLY_RWSPINLOCK_H_
70 ========================================================================
71 Benchmark on (Intel(R) Xeon(R) CPU L5630 @ 2.13GHz) 8 cores(16 HTs)
72 ========================================================================
74 ------------------------------------------------------------------------------
75 1. Single thread benchmark (read/write lock + unlock overhead)
76 Benchmark Iters Total t t/iter iter/sec
77 -------------------------------------------------------------------------------
78 * BM_RWSpinLockRead 100000 1.786 ms 17.86 ns 53.4M
79 +30.5% BM_RWSpinLockWrite 100000 2.331 ms 23.31 ns 40.91M
80 +85.7% BM_RWTicketSpinLock32Read 100000 3.317 ms 33.17 ns 28.75M
81 +96.0% BM_RWTicketSpinLock32Write 100000 3.5 ms 35 ns 27.25M
82 +85.6% BM_RWTicketSpinLock64Read 100000 3.315 ms 33.15 ns 28.77M
83 +96.0% BM_RWTicketSpinLock64Write 100000 3.5 ms 35 ns 27.25M
84 +85.7% BM_RWTicketSpinLock32FavorWriterRead 100000 3.317 ms 33.17 ns 28.75M
85 +29.7% BM_RWTicketSpinLock32FavorWriterWrite 100000 2.316 ms 23.16 ns 41.18M
86 +85.3% BM_RWTicketSpinLock64FavorWriterRead 100000 3.309 ms 33.09 ns 28.82M
87 +30.2% BM_RWTicketSpinLock64FavorWriterWrite 100000 2.325 ms 23.25 ns 41.02M
88 + 175% BM_PThreadRWMutexRead 100000 4.917 ms 49.17 ns 19.4M
89 + 166% BM_PThreadRWMutexWrite 100000 4.757 ms 47.57 ns 20.05M
91 ------------------------------------------------------------------------------
92 2. Contention Benchmark 90% read 10% write
93 Benchmark hits average min max sigma
94 ------------------------------------------------------------------------------
95 ---------- 8 threads ------------
96 RWSpinLock Write 142666 220ns 78ns 40.8us 269ns
97 RWSpinLock Read 1282297 222ns 80ns 37.7us 248ns
98 RWTicketSpinLock Write 85692 209ns 71ns 17.9us 252ns
99 RWTicketSpinLock Read 769571 215ns 78ns 33.4us 251ns
100 pthread_rwlock_t Write 84248 2.48us 99ns 269us 8.19us
101 pthread_rwlock_t Read 761646 933ns 101ns 374us 3.25us
103 ---------- 16 threads ------------
104 RWSpinLock Write 124236 237ns 78ns 261us 801ns
105 RWSpinLock Read 1115807 236ns 78ns 2.27ms 2.17us
106 RWTicketSpinLock Write 81781 231ns 71ns 31.4us 351ns
107 RWTicketSpinLock Read 734518 238ns 78ns 73.6us 379ns
108 pthread_rwlock_t Write 83363 7.12us 99ns 785us 28.1us
109 pthread_rwlock_t Read 754978 2.18us 101ns 1.02ms 14.3us
111 ---------- 50 threads ------------
112 RWSpinLock Write 131142 1.37us 82ns 7.53ms 68.2us
113 RWSpinLock Read 1181240 262ns 78ns 6.62ms 12.7us
114 RWTicketSpinLock Write 83045 397ns 73ns 7.01ms 31.5us
115 RWTicketSpinLock Read 744133 386ns 78ns 11ms 31.4us
116 pthread_rwlock_t Write 80849 112us 103ns 4.52ms 263us
117 pthread_rwlock_t Read 728698 24us 101ns 7.28ms 194us
121 #include <folly/Portability.h>
123 #if defined(__GNUC__) && \
124 (defined(__i386) || FOLLY_X64 || \
126 #define RW_SPINLOCK_USE_X86_INTRINSIC_
127 #include <x86intrin.h>
129 #undef RW_SPINLOCK_USE_X86_INTRINSIC_
132 // iOS doesn't define _mm_cvtsi64_si128 and friends
133 #if defined(__SSE2__) && !TARGET_OS_IPHONE
134 #define RW_SPINLOCK_USE_SSE_INSTRUCTIONS_
136 #undef RW_SPINLOCK_USE_SSE_INSTRUCTIONS_
142 #include <boost/noncopyable.hpp>
145 #include <glog/logging.h>
147 #include <folly/Likely.h>
152 * A simple, small (4-bytes), but unfair rwlock. Use it when you want
153 * a nice writer and don't expect a lot of write/read contention, or
154 * when you need small rwlocks since you are creating a large number
157 * Note that the unfairness here is extreme: if the lock is
158 * continually accessed for read, writers will never get a chance. If
159 * the lock can be that highly contended this class is probably not an
160 * ideal choice anyway.
162 * It currently implements most of the Lockable, SharedLockable and
163 * UpgradeLockable concepts except the TimedLockable related locking/unlocking
166 class RWSpinLock : boost::noncopyable {
167 enum : int32_t { READER = 4, UPGRADED = 2, WRITER = 1 };
169 RWSpinLock() : bits_(0) {}
174 while (!LIKELY(try_lock())) {
175 if (++count > 1000) sched_yield();
179 // Writer is responsible for clearing up both the UPGRADED and WRITER bits.
181 static_assert(READER > WRITER + UPGRADED, "wrong bits!");
182 bits_.fetch_and(~(WRITER | UPGRADED), std::memory_order_release);
185 // SharedLockable Concept
188 while (!LIKELY(try_lock_shared())) {
189 if (++count > 1000) sched_yield();
193 void unlock_shared() {
194 bits_.fetch_add(-READER, std::memory_order_release);
197 // Downgrade the lock from writer status to reader status.
198 void unlock_and_lock_shared() {
199 bits_.fetch_add(READER, std::memory_order_acquire);
203 // UpgradeLockable Concept
204 void lock_upgrade() {
206 while (!try_lock_upgrade()) {
207 if (++count > 1000) sched_yield();
211 void unlock_upgrade() {
212 bits_.fetch_add(-UPGRADED, std::memory_order_acq_rel);
215 // unlock upgrade and try to acquire write lock
216 void unlock_upgrade_and_lock() {
218 while (!try_unlock_upgrade_and_lock()) {
219 if (++count > 1000) sched_yield();
223 // unlock upgrade and read lock atomically
224 void unlock_upgrade_and_lock_shared() {
225 bits_.fetch_add(READER - UPGRADED, std::memory_order_acq_rel);
228 // write unlock and upgrade lock atomically
229 void unlock_and_lock_upgrade() {
230 // need to do it in two steps here -- as the UPGRADED bit might be OR-ed at
231 // the same time when other threads are trying do try_lock_upgrade().
232 bits_.fetch_or(UPGRADED, std::memory_order_acquire);
233 bits_.fetch_add(-WRITER, std::memory_order_release);
237 // Attempt to acquire writer permission. Return false if we didn't get it.
240 return bits_.compare_exchange_strong(expect, WRITER,
241 std::memory_order_acq_rel);
244 // Try to get reader permission on the lock. This can fail if we
245 // find out someone is a writer or upgrader.
246 // Setting the UPGRADED bit would allow a writer-to-be to indicate
247 // its intention to write and block any new readers while waiting
248 // for existing readers to finish and release their read locks. This
249 // helps avoid starving writers (promoted from upgraders).
250 bool try_lock_shared() {
251 // fetch_add is considerably (100%) faster than compare_exchange,
252 // so here we are optimizing for the common (lock success) case.
253 int32_t value = bits_.fetch_add(READER, std::memory_order_acquire);
254 if (UNLIKELY(value & (WRITER|UPGRADED))) {
255 bits_.fetch_add(-READER, std::memory_order_release);
261 // try to unlock upgrade and write lock atomically
262 bool try_unlock_upgrade_and_lock() {
263 int32_t expect = UPGRADED;
264 return bits_.compare_exchange_strong(expect, WRITER,
265 std::memory_order_acq_rel);
268 // try to acquire an upgradable lock.
269 bool try_lock_upgrade() {
270 int32_t value = bits_.fetch_or(UPGRADED, std::memory_order_acquire);
272 // Note: when failed, we cannot flip the UPGRADED bit back,
273 // as in this case there is either another upgrade lock or a write lock.
274 // If it's a write lock, the bit will get cleared up when that lock's done
276 return ((value & (UPGRADED | WRITER)) == 0);
279 // mainly for debugging purposes.
280 int32_t bits() const { return bits_.load(std::memory_order_acquire); }
283 class UpgradedHolder;
288 explicit ReadHolder(RWSpinLock* lock = nullptr) : lock_(lock) {
289 if (lock_) lock_->lock_shared();
292 explicit ReadHolder(RWSpinLock& lock) : lock_(&lock) {
293 lock_->lock_shared();
296 ReadHolder(ReadHolder&& other) noexcept : lock_(other.lock_) {
297 other.lock_ = nullptr;
301 explicit ReadHolder(UpgradedHolder&& upgraded) : lock_(upgraded.lock_) {
302 upgraded.lock_ = nullptr;
303 if (lock_) lock_->unlock_upgrade_and_lock_shared();
306 explicit ReadHolder(WriteHolder&& writer) : lock_(writer.lock_) {
307 writer.lock_ = nullptr;
308 if (lock_) lock_->unlock_and_lock_shared();
311 ReadHolder& operator=(ReadHolder&& other) {
313 swap(lock_, other.lock_);
317 ReadHolder(const ReadHolder& other) = delete;
318 ReadHolder& operator=(const ReadHolder& other) = delete;
320 ~ReadHolder() { if (lock_) lock_->unlock_shared(); }
322 void reset(RWSpinLock* lock = nullptr) {
323 if (lock == lock_) return;
324 if (lock_) lock_->unlock_shared();
326 if (lock_) lock_->lock_shared();
329 void swap(ReadHolder* other) {
330 std::swap(lock_, other->lock_);
334 friend class UpgradedHolder;
335 friend class WriteHolder;
339 class UpgradedHolder {
341 explicit UpgradedHolder(RWSpinLock* lock = nullptr) : lock_(lock) {
342 if (lock_) lock_->lock_upgrade();
345 explicit UpgradedHolder(RWSpinLock& lock) : lock_(&lock) {
346 lock_->lock_upgrade();
349 explicit UpgradedHolder(WriteHolder&& writer) {
350 lock_ = writer.lock_;
351 writer.lock_ = nullptr;
352 if (lock_) lock_->unlock_and_lock_upgrade();
355 UpgradedHolder(UpgradedHolder&& other) noexcept : lock_(other.lock_) {
356 other.lock_ = nullptr;
359 UpgradedHolder& operator =(UpgradedHolder&& other) {
361 swap(lock_, other.lock_);
365 UpgradedHolder(const UpgradedHolder& other) = delete;
366 UpgradedHolder& operator =(const UpgradedHolder& other) = delete;
368 ~UpgradedHolder() { if (lock_) lock_->unlock_upgrade(); }
370 void reset(RWSpinLock* lock = nullptr) {
371 if (lock == lock_) return;
372 if (lock_) lock_->unlock_upgrade();
374 if (lock_) lock_->lock_upgrade();
377 void swap(UpgradedHolder* other) {
379 swap(lock_, other->lock_);
383 friend class WriteHolder;
384 friend class ReadHolder;
390 explicit WriteHolder(RWSpinLock* lock = nullptr) : lock_(lock) {
391 if (lock_) lock_->lock();
394 explicit WriteHolder(RWSpinLock& lock) : lock_(&lock) {
398 // promoted from an upgrade lock holder
399 explicit WriteHolder(UpgradedHolder&& upgraded) {
400 lock_ = upgraded.lock_;
401 upgraded.lock_ = nullptr;
402 if (lock_) lock_->unlock_upgrade_and_lock();
405 WriteHolder(WriteHolder&& other) noexcept : lock_(other.lock_) {
406 other.lock_ = nullptr;
409 WriteHolder& operator =(WriteHolder&& other) {
411 swap(lock_, other.lock_);
415 WriteHolder(const WriteHolder& other) = delete;
416 WriteHolder& operator =(const WriteHolder& other) = delete;
418 ~WriteHolder () { if (lock_) lock_->unlock(); }
420 void reset(RWSpinLock* lock = nullptr) {
421 if (lock == lock_) return;
422 if (lock_) lock_->unlock();
424 if (lock_) lock_->lock();
427 void swap(WriteHolder* other) {
429 swap(lock_, other->lock_);
433 friend class ReadHolder;
434 friend class UpgradedHolder;
438 // Synchronized<> adaptors
439 friend void acquireRead(RWSpinLock& l) { return l.lock_shared(); }
440 friend void acquireReadWrite(RWSpinLock& l) { return l.lock(); }
441 friend void releaseRead(RWSpinLock& l) { return l.unlock_shared(); }
442 friend void releaseReadWrite(RWSpinLock& l) { return l.unlock(); }
445 std::atomic<int32_t> bits_;
449 #ifdef RW_SPINLOCK_USE_X86_INTRINSIC_
450 // A more balanced Read-Write spin lock implemented based on GCC intrinsics.
453 template <size_t kBitWidth> struct RWTicketIntTrait {
454 static_assert(kBitWidth == 32 || kBitWidth == 64,
455 "bit width has to be either 32 or 64 ");
459 struct RWTicketIntTrait<64> {
460 typedef uint64_t FullInt;
461 typedef uint32_t HalfInt;
462 typedef uint16_t QuarterInt;
464 #ifdef RW_SPINLOCK_USE_SSE_INSTRUCTIONS_
465 static __m128i make128(const uint16_t v[4]) {
466 return _mm_set_epi16(0, 0, 0, 0, v[3], v[2], v[1], v[0]);
468 static inline __m128i fromInteger(uint64_t from) {
469 return _mm_cvtsi64_si128(from);
471 static inline uint64_t toInteger(__m128i in) {
472 return _mm_cvtsi128_si64(in);
474 static inline uint64_t addParallel(__m128i in, __m128i kDelta) {
475 return toInteger(_mm_add_epi16(in, kDelta));
481 struct RWTicketIntTrait<32> {
482 typedef uint32_t FullInt;
483 typedef uint16_t HalfInt;
484 typedef uint8_t QuarterInt;
486 #ifdef RW_SPINLOCK_USE_SSE_INSTRUCTIONS_
487 static __m128i make128(const uint8_t v[4]) {
488 return _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0,
489 0, 0, 0, 0, v[3], v[2], v[1], v[0]);
491 static inline __m128i fromInteger(uint32_t from) {
492 return _mm_cvtsi32_si128(from);
494 static inline uint32_t toInteger(__m128i in) {
495 return _mm_cvtsi128_si32(in);
497 static inline uint32_t addParallel(__m128i in, __m128i kDelta) {
498 return toInteger(_mm_add_epi8(in, kDelta));
505 template<size_t kBitWidth, bool kFavorWriter=false>
506 class RWTicketSpinLockT : boost::noncopyable {
507 typedef detail::RWTicketIntTrait<kBitWidth> IntTraitType;
508 typedef typename detail::RWTicketIntTrait<kBitWidth>::FullInt FullInt;
509 typedef typename detail::RWTicketIntTrait<kBitWidth>::HalfInt HalfInt;
510 typedef typename detail::RWTicketIntTrait<kBitWidth>::QuarterInt
516 __extension__ struct {
523 private: // Some x64-specific utilities for atomic access to ticket.
524 template<class T> static T load_acquire(T* addr) {
525 T t = *addr; // acquire barrier
526 asm volatile("" : : : "memory");
531 static void store_release(T* addr, T v) {
532 asm volatile("" : : : "memory");
533 *addr = v; // release barrier
538 RWTicketSpinLockT() {
539 store_release(&ticket.whole, FullInt(0));
544 writeLockAggressive();
551 * Both try_lock and try_lock_shared diverge in our implementation from the
552 * lock algorithm described in the link above.
554 * In the read case, it is undesirable that the readers could wait
555 * for another reader (before increasing ticket.read in the other
556 * implementation). Our approach gives up on
557 * first-come-first-serve, but our benchmarks showed improve
558 * performance for both readers and writers under heavily contended
559 * cases, particularly when the number of threads exceeds the number
562 * We have writeLockAggressive() using the original implementation
563 * for a writer, which gives some advantage to the writer over the
564 * readers---for that path it is guaranteed that the writer will
565 * acquire the lock after all the existing readers exit.
569 FullInt old = t.whole = load_acquire(&ticket.whole);
570 if (t.users != t.write) return false;
572 return __sync_bool_compare_and_swap(&ticket.whole, old, t.whole);
576 * Call this if you want to prioritize writer to avoid starvation.
577 * Unlike writeLockNice, immediately acquires the write lock when
578 * the existing readers (arriving before the writer) finish their
581 void writeLockAggressive() {
582 // sched_yield() is needed here to avoid a pathology if the number
583 // of threads attempting concurrent writes is >= the number of real
584 // cores allocated to this process. This is less likely than the
585 // corresponding situation in lock_shared(), but we still want to
588 QuarterInt val = __sync_fetch_and_add(&ticket.users, 1);
589 while (val != load_acquire(&ticket.write)) {
590 asm_volatile_pause();
591 if (UNLIKELY(++count > 1000)) sched_yield();
595 // Call this when the writer should be nicer to the readers.
596 void writeLockNice() {
597 // Here it doesn't cpu-relax the writer.
599 // This is because usually we have many more readers than the
600 // writers, so the writer has less chance to get the lock when
601 // there are a lot of competing readers. The aggressive spinning
602 // can help to avoid starving writers.
604 // We don't worry about sched_yield() here because the caller
605 // has already explicitly abandoned fairness.
606 while (!try_lock()) {}
609 // Atomically unlock the write-lock from writer and acquire the read-lock.
610 void unlock_and_lock_shared() {
611 QuarterInt val = __sync_fetch_and_add(&ticket.read, 1);
614 // Release writer permission on the lock.
617 t.whole = load_acquire(&ticket.whole);
618 FullInt old = t.whole;
620 #ifdef RW_SPINLOCK_USE_SSE_INSTRUCTIONS_
621 // SSE2 can reduce the lock and unlock overhead by 10%
622 static const QuarterInt kDeltaBuf[4] = { 1, 1, 0, 0 }; // write/read/user
623 static const __m128i kDelta = IntTraitType::make128(kDeltaBuf);
624 __m128i m = IntTraitType::fromInteger(old);
625 t.whole = IntTraitType::addParallel(m, kDelta);
630 store_release(&ticket.readWrite, t.readWrite);
634 // sched_yield() is important here because we can't grab the
635 // shared lock if there is a pending writeLockAggressive, so we
636 // need to let threads that already have a shared lock complete
638 while (!LIKELY(try_lock_shared())) {
639 asm_volatile_pause();
640 if (UNLIKELY((++count & 1023) == 0)) sched_yield();
644 bool try_lock_shared() {
646 old.whole = t.whole = load_acquire(&ticket.whole);
647 old.users = old.read;
648 #ifdef RW_SPINLOCK_USE_SSE_INSTRUCTIONS_
649 // SSE2 may reduce the total lock and unlock overhead by 10%
650 static const QuarterInt kDeltaBuf[4] = { 0, 1, 1, 0 }; // write/read/user
651 static const __m128i kDelta = IntTraitType::make128(kDeltaBuf);
652 __m128i m = IntTraitType::fromInteger(old.whole);
653 t.whole = IntTraitType::addParallel(m, kDelta);
658 return __sync_bool_compare_and_swap(&ticket.whole, old.whole, t.whole);
661 void unlock_shared() {
662 QuarterInt val = __sync_fetch_and_add(&ticket.write, 1);
667 typedef RWTicketSpinLockT<kBitWidth, kFavorWriter> RWSpinLock;
668 class ReadHolder : boost::noncopyable {
670 explicit ReadHolder(RWSpinLock *lock = nullptr) :
672 if (lock_) lock_->lock_shared();
675 explicit ReadHolder(RWSpinLock &lock) : lock_ (&lock) {
676 if (lock_) lock_->lock_shared();
679 // atomically unlock the write-lock from writer and acquire the read-lock
680 explicit ReadHolder(WriteHolder *writer) : lock_(nullptr) {
681 std::swap(this->lock_, writer->lock_);
683 lock_->unlock_and_lock_shared();
688 if (lock_) lock_->unlock_shared();
691 void reset(RWSpinLock *lock = nullptr) {
692 if (lock_) lock_->unlock_shared();
694 if (lock_) lock_->lock_shared();
697 void swap(ReadHolder *other) {
698 std::swap(this->lock_, other->lock_);
705 class WriteHolder : boost::noncopyable {
707 explicit WriteHolder(RWSpinLock *lock = nullptr) : lock_(lock) {
708 if (lock_) lock_->lock();
710 explicit WriteHolder(RWSpinLock &lock) : lock_ (&lock) {
711 if (lock_) lock_->lock();
715 if (lock_) lock_->unlock();
718 void reset(RWSpinLock *lock = nullptr) {
719 if (lock == lock_) return;
720 if (lock_) lock_->unlock();
722 if (lock_) lock_->lock();
725 void swap(WriteHolder *other) {
726 std::swap(this->lock_, other->lock_);
730 friend class ReadHolder;
734 // Synchronized<> adaptors.
735 friend void acquireRead(RWTicketSpinLockT& mutex) {
738 friend void acquireReadWrite(RWTicketSpinLockT& mutex) {
741 friend bool acquireReadWrite(RWTicketSpinLockT& mutex,
742 unsigned int milliseconds) {
746 friend void releaseRead(RWTicketSpinLockT& mutex) {
747 mutex.unlock_shared();
749 friend void releaseReadWrite(RWTicketSpinLockT& mutex) {
754 typedef RWTicketSpinLockT<32> RWTicketSpinLock32;
755 typedef RWTicketSpinLockT<64> RWTicketSpinLock64;
757 #endif // RW_SPINLOCK_USE_X86_INTRINSIC_
761 #ifdef RW_SPINLOCK_USE_X86_INTRINSIC_
762 #undef RW_SPINLOCK_USE_X86_INTRINSIC_
765 #endif // FOLLY_RWSPINLOCK_H_