1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382
|
// Copyright (c) 2006-2018 Maxim Khizhinsky
//
// Distributed under the Boost Software License, Version 1.0. (See accompanying
// file LICENSE or copy at http://www.boost.org/LICENSE_1_0.txt)
#ifndef CDSLIB_SYNC_SPINLOCK_H
#define CDSLIB_SYNC_SPINLOCK_H
#include <cds/algo/atomic.h>
#include <cds/os/thread.h>
#include <cds/algo/backoff_strategy.h>
namespace cds {
/// Synchronization primitives
namespace sync {
/// Spin lock
/**
Simple and light-weight spin-lock critical section
It is useful to gain access to small (short-timed) code
Algorithm:
TATAS (test-and-test-and-lock)
[1984] L. Rudolph, Z. Segall. Dynamic Decentralized Cache Schemes for MIMD Parallel Processors.
No serialization performed - any of waiting threads may owns the spin-lock.
This spin-lock is NOT recursive: the thread owned the lock cannot call \p lock() method without deadlock.
The method \p unlock() can call any thread
DEBUG version: The spinlock stores owner thead id. Assertion is raised when:
- double lock attempt encountered by same thread (deadlock)
- unlock by another thread
If spin-lock is locked the \p Backoff algorithm is called. Predefined \p backoff::LockDefault class yields current
thread and repeats lock attempts later
Template parameters:
- \p Backoff - backoff strategy. Used when spin lock is locked
*/
template <typename Backoff >
class spin_lock
{
public:
typedef Backoff backoff_strategy; ///< back-off strategy type
private:
atomics::atomic<bool> m_spin; ///< Spin
# ifdef CDS_DEBUG
typename OS::ThreadId m_dbgOwnerId; ///< Owner thread id (only for debug mode)
# endif
public:
/// Construct free (unlocked) spin-lock
spin_lock() noexcept
# ifdef CDS_DEBUG
: m_dbgOwnerId( OS::c_NullThreadId )
# endif
{
m_spin.store( false, atomics::memory_order_release );
}
/// Construct spin-lock in specified state
/**
In debug mode: if \p bLocked = true then spin-lock is made owned by current thread
*/
explicit spin_lock( bool bLocked ) noexcept
# ifdef CDS_DEBUG
: m_dbgOwnerId( bLocked ? cds::OS::get_current_thread_id() : cds::OS::c_NullThreadId )
# endif
{
m_spin.store( bLocked, atomics::memory_order_release );
}
/// Dummy copy constructor
/**
The ctor initializes the spin to free (unlocked) state like the default ctor.
*/
spin_lock(const spin_lock<Backoff>& ) noexcept
: m_spin( false )
# ifdef CDS_DEBUG
, m_dbgOwnerId( cds::OS::c_NullThreadId )
# endif
{
CDS_TSAN_ANNOTATE_MUTEX_CREATE( this );
}
/// Destructor. On debug time it checks whether spin-lock is free
~spin_lock()
{
assert( !m_spin.load( atomics::memory_order_relaxed ));
CDS_TSAN_ANNOTATE_MUTEX_DESTROY( this );
}
/// Check if the spin is locked
bool is_locked() const noexcept
{
return m_spin.load( atomics::memory_order_relaxed );
}
/// Try to lock the object
/**
Returns \p true if locking is succeeded
otherwise (if the spin is already locked) returns \p false
Debug version: deadlock can be detected
*/
bool try_lock() noexcept
{
# ifdef CDS_THREAD_SANITIZER_ENABLED
bool bCurrent = m_spin.exchange( true, atomics::memory_order_acq_rel );
if ( !bCurrent )
CDS_TSAN_ANNOTATE_MUTEX_ACQUIRED( this );
# else
bool bCurrent = m_spin.exchange( true, atomics::memory_order_acquire );
# endif
CDS_DEBUG_ONLY(
if ( !bCurrent ) {
m_dbgOwnerId = OS::get_current_thread_id();
}
)
return !bCurrent;
}
/// Try to lock the object, repeat \p nTryCount times if failed
/**
Returns \p true if locking is succeeded
otherwise (if the spin is already locked) returns \p false
*/
bool try_lock( unsigned int nTryCount ) noexcept( noexcept( backoff_strategy()()))
{
backoff_strategy backoff;
while ( nTryCount-- ) {
if ( try_lock())
return true;
backoff();
}
return false;
}
/// Lock the spin-lock. Waits infinitely while spin-lock is locked. Debug version: deadlock may be detected
void lock() noexcept(noexcept( backoff_strategy()()))
{
backoff_strategy backoff;
// Deadlock detected
CDS_TSAN_ANNOTATE_IGNORE_READS_BEGIN;
assert( m_dbgOwnerId != OS::get_current_thread_id());
CDS_TSAN_ANNOTATE_IGNORE_READS_END;
// TATAS algorithm
while ( !try_lock()) {
while ( m_spin.load( atomics::memory_order_acquire ))
backoff();
}
assert( m_dbgOwnerId == OS::get_current_thread_id());
}
/// Unlock the spin-lock. Debug version: deadlock may be detected
void unlock() noexcept
{
assert( m_spin.load( atomics::memory_order_relaxed ));
assert( m_dbgOwnerId == OS::get_current_thread_id());
CDS_DEBUG_ONLY( m_dbgOwnerId = OS::c_NullThreadId; )
CDS_TSAN_ANNOTATE_MUTEX_RELEASED( this );
m_spin.store( false, atomics::memory_order_release );
}
};
/// Spin-lock implementation default for the current platform
typedef spin_lock<backoff::LockDefault > spin;
/// Recursive spin lock.
/**
Allows recursive calls: the owner thread may recursive enter to critical section guarded by the spin-lock.
Template parameters:
- \p Integral one of integral atomic type: <tt>unsigned int</tt>, \p int, and others
- \p Backoff backoff strategy. Used when spin lock is locked
*/
template <typename Integral, class Backoff>
class reentrant_spin_lock
{
typedef OS::ThreadId thread_id; ///< The type of thread id
public:
typedef Integral integral_type; ///< The integral type
typedef Backoff backoff_strategy; ///< The backoff type
private:
//@cond
atomics::atomic<integral_type> m_spin; ///< spin-lock atomic
atomics::atomic<thread_id> m_OwnerId; ///< Owner thread id. If spin-lock is not locked it usually equals to \p OS::c_NullThreadId
//@endcond
private:
//@cond
void take( thread_id tid ) noexcept
{
m_OwnerId.store( tid, atomics::memory_order_relaxed );
}
void free() noexcept
{
m_OwnerId.store( OS::c_NullThreadId, atomics::memory_order_relaxed );
}
bool is_taken( thread_id tid ) const noexcept
{
return m_OwnerId.load( atomics::memory_order_relaxed ) == tid;
}
bool try_taken_lock( thread_id tid ) noexcept
{
if ( is_taken( tid )) {
m_spin.fetch_add( 1, atomics::memory_order_relaxed );
return true;
}
return false;
}
bool try_acquire() noexcept
{
integral_type nCurrent = 0;
bool bRet = m_spin.compare_exchange_weak( nCurrent, 1, atomics::memory_order_acquire, atomics::memory_order_acquire );
# ifdef CDS_THREAD_SANITIZER_ENABLED
if ( bRet )
CDS_TSAN_ANNOTATE_MUTEX_ACQUIRED( this );
# endif
return bRet;
}
bool try_acquire( unsigned int nTryCount ) noexcept( noexcept( backoff_strategy()()))
{
backoff_strategy bkoff;
while ( nTryCount-- ) {
if ( try_acquire())
return true;
bkoff();
}
return false;
}
void acquire() noexcept( noexcept( backoff_strategy()()))
{
// TATAS algorithm
backoff_strategy bkoff;
while ( !try_acquire()) {
while ( m_spin.load( atomics::memory_order_acquire ))
bkoff();
}
}
//@endcond
public:
/// Default constructor initializes spin to free (unlocked) state
reentrant_spin_lock() noexcept
: m_spin(0)
, m_OwnerId( OS::c_NullThreadId )
{
CDS_TSAN_ANNOTATE_MUTEX_CREATE( this );
}
/// Dummy copy constructor
/**
In theory, spin-lock cannot be copied. However, it is not practical.
Therefore, we provide dummy copy constructor that do no copy in fact. The ctor
initializes the spin to free (unlocked) state like default ctor.
*/
reentrant_spin_lock( const reentrant_spin_lock<Integral, Backoff>& ) noexcept
: m_spin(0)
, m_OwnerId( OS::c_NullThreadId )
{
CDS_TSAN_ANNOTATE_MUTEX_CREATE( this );
}
/// Construct object in specified state
explicit reentrant_spin_lock( bool bLocked )
: m_spin(0)
, m_OwnerId( OS::c_NullThreadId )
{
CDS_TSAN_ANNOTATE_MUTEX_CREATE( this );
if ( bLocked )
lock();
}
/// Dtor. Spin-lock must be unlocked
~reentrant_spin_lock()
{
assert( m_spin.load( atomics::memory_order_acquire ) == 0 );
assert( m_OwnerId.load( atomics::memory_order_relaxed ) == OS::c_NullThreadId );
CDS_TSAN_ANNOTATE_MUTEX_DESTROY( this );
}
/// Checks if the spin is locked
/**
The spin is locked if lock count > 0 and the current thread is not an owner of the lock.
Otherwise (i.e. lock count == 0 or the curren thread owns the spin) the spin is unlocked.
*/
bool is_locked() const noexcept
{
return !( m_spin.load( atomics::memory_order_relaxed ) == 0 || is_taken( cds::OS::get_current_thread_id()));
}
/// Try to lock the spin-lock
bool try_lock() noexcept( noexcept( std::declval<reentrant_spin_lock>().try_acquire()))
{
thread_id tid = OS::get_current_thread_id();
if ( try_taken_lock( tid ))
return true;
if ( try_acquire()) {
take( tid );
return true;
}
return false;
}
/// Try to lock up to \p nTryCount attempts
bool try_lock( unsigned int nTryCount ) noexcept( noexcept( std::declval<reentrant_spin_lock>().try_acquire( nTryCount )))
{
thread_id tid = OS::get_current_thread_id();
if ( try_taken_lock( tid ))
return true;
if ( try_acquire( nTryCount )) {
take( tid );
return true;
}
return false;
}
/// Lock the object waits if it is busy
void lock() noexcept( noexcept( std::declval<reentrant_spin_lock>().acquire()))
{
thread_id tid = OS::get_current_thread_id();
if ( !try_taken_lock( tid )) {
acquire();
take( tid );
}
}
/// Unlock the spin-lock
void unlock() noexcept
{
assert( is_taken( OS::get_current_thread_id()));
integral_type n = m_spin.load( atomics::memory_order_relaxed );
if ( n > 1 )
m_spin.store( n - 1, atomics::memory_order_relaxed );
else {
free();
CDS_TSAN_ANNOTATE_MUTEX_RELEASED( this );
m_spin.store( 0, atomics::memory_order_release );
}
}
/// Change the owner of locked spin-lock. May be called by thread that owns spin-lock
void change_owner( OS::ThreadId newOwnerId ) noexcept
{
assert( is_taken( OS::get_current_thread_id()));
assert( newOwnerId != OS::c_NullThreadId );
m_OwnerId.store( newOwnerId, atomics::memory_order_relaxed );
}
};
/// Recursive 32bit spin-lock
typedef reentrant_spin_lock<uint32_t, backoff::LockDefault> reentrant_spin32;
/// Default recursive spin-lock
typedef reentrant_spin32 reentrant_spin;
/// Recursive 64bit spin-lock
typedef reentrant_spin_lock<uint64_t, backoff::LockDefault> reentrant_spin64;
} // namespace sync
} // namespace cds
#endif // #ifndef CDSLIB_SYNC_SPINLOCK_H
|