3 #ifndef __CDS_INTRUSIVE_SPLIT_LIST_RCU_H
4 #define __CDS_INTRUSIVE_SPLIT_LIST_RCU_H
6 #include <cds/intrusive/details/split_list_base.h>
7 #include <cds/details/binary_functor_wrapper.h>
9 namespace cds { namespace intrusive {
11 /// Split-ordered list RCU specialization
12 /** @ingroup cds_intrusive_map
13 \anchor cds_intrusive_SplitListSet_rcu
15 Hash table implementation based on split-ordered list algorithm discovered by Ori Shalev and Nir Shavit, see
16 - [2003] Ori Shalev, Nir Shavit "Split-Ordered Lists - Lock-free Resizable Hash Tables"
17 - [2008] Nir Shavit "The Art of Multiprocessor Programming"
19 The split-ordered list is a lock-free implementation of an extensible unbounded hash table. It uses original
20 recursive split-ordering algorithm discovered by Ori Shalev and Nir Shavit that allows to split buckets
21 without moving an item on resizing, see \ref cds_SplitList_algo_desc "short algo description".
25 Template parameters are:
26 - \p RCU - one of \ref cds_urcu_gc "RCU type"
27 - \p OrderedList - ordered list implementation used as bucket for hash set, for example, MichaelList, LazyList.
28 The intrusive ordered list implementation specifies the type \p T stored in the hash-set,
29 the comparing functor for the type \p T and other features specific for the ordered list.
30 - \p Traits - set traits, default isd \p split_list::traits.
31 Instead of defining \p Traits struct you may use option-based syntax with \p split_list::make_traits metafunction.
33 @note About reqired features of hash functor see \ref cds_SplitList_hash_functor "SplitList general description".
36 Before including <tt><cds/intrusive/split_list_rcu.h></tt> you should include appropriate RCU header file,
37 see \ref cds_urcu_gc "RCU type" for list of existing RCU class and corresponding header files.
38 For example, for \ref cds_urcu_general_buffered_gc "general-purpose buffered RCU" and
39 MichaelList-based split-list you should include:
41 #include <cds/urcu/general_buffered.h>
42 #include <cds/intrusive/michael_list_rcu.h>
43 #include <cds/intrusive/split_list_rcu.h>
45 // Declare Michael's list for type Foo and default traits:
46 typedef cds::intrusive::MichaelList< cds::urcu::gc< cds::urcu::general_buffered<> >, Foo > rcu_michael_list;
48 // Declare split-list based on rcu_michael_list
49 typedef cds::intrusive::SplitListSet< cds::urcu::gc< cds::urcu::general_buffered<> >, rcu_michael_list > rcu_split_list;
56 # ifdef CDS_DOXYGEN_INVOKED
57 class Traits = split_list::traits
62 class SplitListSet< cds::urcu::gc< RCU >, OrderedList, Traits >
65 typedef cds::urcu::gc< RCU > gc; ///< RCU garbage collector
66 typedef Traits traits; ///< Traits template parameters
68 /// Hash functor for \ref value_type and all its derivatives that you use
69 typedef typename cds::opt::v::hash_selector< typename traits::hash >::type hash;
73 typedef split_list::details::rebind_list_traits<OrderedList, traits> wrapped_ordered_list;
77 # ifdef CDS_DOXYGEN_INVOKED
78 typedef OrderedList ordered_list; ///< type of ordered list used as base for split-list
80 typedef typename wrapped_ordered_list::result ordered_list;
82 typedef typename ordered_list::value_type value_type; ///< type of value stored in the split-list
83 typedef typename ordered_list::key_comparator key_comparator; ///< key compare functor
84 typedef typename ordered_list::disposer disposer; ///< Node disposer functor
85 typedef typename ordered_list::rcu_lock rcu_lock; ///< RCU scoped lock
86 typedef typename ordered_list::exempt_ptr exempt_ptr; ///< pointer to extracted node
87 /// Group of \p extract_xxx functions require external locking if underlying ordered list requires that
88 static CDS_CONSTEXPR const bool c_bExtractLockExternal = ordered_list::c_bExtractLockExternal;
90 typedef typename traits::item_counter item_counter; ///< Item counter type
91 typedef typename traits::back_off back_off; ///< back-off strategy for spinning
92 typedef typename traits::memory_model memory_model; ///< Memory ordering. See cds::opt::memory_model option
93 typedef typename traits::stat stat; ///< Internal statistics
96 typedef typename ordered_list::node_type list_node_type; ///< Node type as declared in ordered list
97 typedef split_list::node<list_node_type> node_type; ///< split-list node type
98 typedef node_type dummy_node_type; ///< dummy node type
100 /// Split-list node traits
102 This traits is intended for converting between underlying ordered list node type \ref list_node_type
103 and split-list node type \ref node_type
105 typedef split_list::node_traits<typename ordered_list::node_traits> node_traits;
108 /// Bucket table implementation
109 typedef typename split_list::details::bucket_table_selector<
110 traits::dynamic_bucket_table
113 , opt::allocator< typename traits::allocator >
114 , opt::memory_model< memory_model >
115 >::type bucket_table;
121 /// Ordered list wrapper to access protected members of OrderedList
122 class ordered_list_wrapper: public ordered_list
124 typedef ordered_list base_class;
125 typedef typename base_class::auxiliary_head bucket_head_type;
128 bool insert_at( dummy_node_type * pHead, value_type& val )
130 assert( pHead != nullptr );
131 bucket_head_type h(pHead);
132 return base_class::insert_at( h, val );
135 template <typename Func>
136 bool insert_at( dummy_node_type * pHead, value_type& val, Func f )
138 assert( pHead != nullptr );
139 bucket_head_type h(pHead);
140 return base_class::insert_at( h, val, f );
143 template <typename Func>
144 std::pair<bool, bool> ensure_at( dummy_node_type * pHead, value_type& val, Func func )
146 assert( pHead != nullptr );
147 bucket_head_type h(pHead);
148 return base_class::ensure_at( h, val, func );
151 bool unlink_at( dummy_node_type * pHead, value_type& val )
153 assert( pHead != nullptr );
154 bucket_head_type h(pHead);
155 return base_class::unlink_at( h, val );
158 template <typename Q, typename Compare, typename Func>
159 bool erase_at( dummy_node_type * pHead, split_list::details::search_value_type<Q> const& val, Compare cmp, Func f )
161 assert( pHead != nullptr );
162 bucket_head_type h(pHead);
163 return base_class::erase_at( h, val, cmp, f );
166 template <typename Q, typename Compare>
167 bool erase_at( dummy_node_type * pHead, split_list::details::search_value_type<Q> const& val, Compare cmp )
169 assert( pHead != nullptr );
170 bucket_head_type h(pHead);
171 return base_class::erase_at( h, val, cmp );
174 template <typename Q, typename Compare>
175 value_type * extract_at( dummy_node_type * pHead, split_list::details::search_value_type<Q>& val, Compare cmp )
177 assert( pHead != nullptr );
178 bucket_head_type h(pHead);
179 return base_class::extract_at( h, val, cmp );
182 template <typename Q, typename Compare, typename Func>
183 bool find_at( dummy_node_type * pHead, split_list::details::search_value_type<Q>& val, Compare cmp, Func f ) const
185 assert( pHead != nullptr );
186 bucket_head_type h(pHead);
187 return base_class::find_at( h, val, cmp, f );
190 template <typename Q, typename Compare>
191 bool find_at( dummy_node_type * pHead, split_list::details::search_value_type<Q> const & val, Compare cmp ) const
193 assert( pHead != nullptr );
194 bucket_head_type h(pHead);
195 return base_class::find_at( h, val, cmp );
198 template <typename Q, typename Compare>
199 value_type * get_at( dummy_node_type * pHead, split_list::details::search_value_type<Q>& val, Compare cmp ) const
201 assert( pHead != nullptr );
202 bucket_head_type h(pHead);
203 return base_class::get_at( h, val, cmp );
206 bool insert_aux_node( dummy_node_type * pNode )
208 return base_class::insert_aux_node( pNode );
210 bool insert_aux_node( dummy_node_type * pHead, dummy_node_type * pNode )
212 bucket_head_type h(pHead);
213 return base_class::insert_aux_node( h, pNode );
217 template <typename Less>
218 struct less_wrapper: public cds::opt::details::make_comparator_from_less<Less>
220 typedef cds::opt::details::make_comparator_from_less<Less> base_wrapper;
222 template <typename Q1, typename Q2>
223 int operator()( split_list::details::search_value_type<Q1> const& v1, Q2 const& v2 ) const
225 return base_wrapper::operator()( v1.val, v2 );
228 template <typename Q1, typename Q2>
229 int operator()( Q1 const& v1, split_list::details::search_value_type<Q2> const& v2 ) const
231 return base_wrapper::operator()( v1, v2.val );
237 ordered_list_wrapper m_List; ///< Ordered list containing split-list items
238 bucket_table m_Buckets; ///< bucket table
239 atomics::atomic<size_t> m_nBucketCountLog2; ///< log2( current bucket count )
240 item_counter m_ItemCounter; ///< Item counter
241 hash m_HashFunctor; ///< Hash functor
242 stat m_Stat; ///< Internal stattistics accumulator
246 typedef cds::details::Allocator< dummy_node_type, typename traits::allocator > dummy_node_allocator;
248 dummy_node_type * alloc_dummy_node( size_t nHash )
250 m_Stat.onHeadNodeAllocated();
251 return dummy_node_allocator().New( nHash );
253 void free_dummy_node( dummy_node_type * p )
255 dummy_node_allocator().Delete( p );
256 m_Stat.onHeadNodeFreed();
259 /// Calculates hash value of \p key
260 template <typename Q>
261 size_t hash_value( Q const& key ) const
263 return m_HashFunctor( key );
266 size_t bucket_no( size_t nHash ) const
268 return nHash & ( (1 << m_nBucketCountLog2.load(atomics::memory_order_relaxed)) - 1 );
271 static size_t parent_bucket( size_t nBucket )
273 assert( nBucket > 0 );
274 return nBucket & ~( 1 << bitop::MSBnz( nBucket ) );
277 dummy_node_type * init_bucket( size_t nBucket )
279 assert( nBucket > 0 );
280 size_t nParent = parent_bucket( nBucket );
282 dummy_node_type * pParentBucket = m_Buckets.bucket( nParent );
283 if ( pParentBucket == nullptr ) {
284 pParentBucket = init_bucket( nParent );
285 m_Stat.onRecursiveInitBucket();
288 assert( pParentBucket != nullptr );
290 // Allocate a dummy node for new bucket
292 dummy_node_type * pBucket = alloc_dummy_node( split_list::dummy_hash( nBucket ) );
293 if ( m_List.insert_aux_node( pParentBucket, pBucket ) ) {
294 m_Buckets.bucket( nBucket, pBucket );
295 m_Stat.onNewBucket();
298 free_dummy_node( pBucket );
301 // Another thread set the bucket. Wait while it done
303 // In this point, we must wait while nBucket is empty.
304 // The compiler can decide that waiting loop can be "optimized" (stripped)
305 // To prevent this situation, we use waiting on volatile bucket_head_ptr pointer.
307 m_Stat.onBucketInitContenton();
310 dummy_node_type volatile * p = m_Buckets.bucket( nBucket );
312 return const_cast<dummy_node_type *>( p );
314 m_Stat.onBusyWaitBucketInit();
318 dummy_node_type * get_bucket( size_t nHash )
320 size_t nBucket = bucket_no( nHash );
322 dummy_node_type * pHead = m_Buckets.bucket( nBucket );
323 if ( pHead == nullptr )
324 pHead = init_bucket( nBucket );
326 assert( pHead->is_dummy() );
333 // GC and OrderedList::gc must be the same
334 static_assert( std::is_same<gc, typename ordered_list::gc>::value, "GC and OrderedList::gc must be the same");
336 // atomicity::empty_item_counter is not allowed as a item counter
337 static_assert( !std::is_same<item_counter, cds::atomicity::empty_item_counter>::value,
338 "cds::atomicity::empty_item_counter is not allowed as a item counter");
340 // Initialize bucket 0
341 dummy_node_type * pNode = alloc_dummy_node( 0 /*split_list::dummy_hash(0)*/ );
343 // insert_aux_node cannot return false for empty list
344 CDS_VERIFY( m_List.insert_aux_node( pNode ));
346 m_Buckets.bucket( 0, pNode );
349 void inc_item_count()
351 size_t sz = m_nBucketCountLog2.load(atomics::memory_order_relaxed);
352 if ( ( ++m_ItemCounter >> sz ) > m_Buckets.load_factor() && ((size_t)(1 << sz )) < m_Buckets.capacity() )
354 m_nBucketCountLog2.compare_exchange_strong( sz, sz + 1, atomics::memory_order_seq_cst, atomics::memory_order_relaxed );
358 template <typename Q, typename Compare, typename Func>
359 bool find_( Q& val, Compare cmp, Func f )
361 size_t nHash = hash_value( val );
362 split_list::details::search_value_type<Q> sv( val, split_list::regular_hash( nHash ));
363 dummy_node_type * pHead = get_bucket( nHash );
364 assert( pHead != nullptr );
366 return m_Stat.onFind( m_List.find_at( pHead, sv, cmp,
367 [&f](value_type& item, split_list::details::search_value_type<Q>& val){ f(item, val.val ); }));
370 template <typename Q, typename Compare>
371 bool find_value( Q const& val, Compare cmp )
373 size_t nHash = hash_value( val );
374 split_list::details::search_value_type<Q const> sv( val, split_list::regular_hash( nHash ));
375 dummy_node_type * pHead = get_bucket( nHash );
376 assert( pHead != nullptr );
378 return m_Stat.onFind( m_List.find_at( pHead, sv, cmp ));
381 template <typename Q, typename Compare>
382 value_type * get_( Q const& val, Compare cmp )
384 size_t nHash = hash_value( val );
385 split_list::details::search_value_type<Q const> sv( val, split_list::regular_hash( nHash ));
386 dummy_node_type * pHead = get_bucket( nHash );
387 assert( pHead != nullptr );
389 value_type * p = m_List.get_at( pHead, sv, cmp );
390 m_Stat.onFind( p != nullptr );
394 template <typename Q, typename Compare>
395 value_type * extract_( Q const& val, Compare cmp )
397 size_t nHash = hash_value( val );
398 split_list::details::search_value_type<Q const> sv( val, split_list::regular_hash( nHash ));
399 dummy_node_type * pHead = get_bucket( nHash );
400 assert( pHead != nullptr );
402 value_type * pNode = m_List.extract_at( pHead, sv, cmp );
405 m_Stat.onExtractSuccess();
408 m_Stat.onExtractFailed();
412 template <typename Q, typename Less>
413 value_type * extract_with_( Q const& val, Less pred )
415 return extract_( val, typename wrapped_ordered_list::template make_compare_from_less<Less>());
418 template <typename Q, typename Compare>
419 bool erase_( const Q& val, Compare cmp )
421 size_t nHash = hash_value( val );
422 split_list::details::search_value_type<Q const> sv( val, split_list::regular_hash( nHash ));
423 dummy_node_type * pHead = get_bucket( nHash );
424 assert( pHead != nullptr );
426 if ( m_List.erase_at( pHead, sv, cmp ) ) {
428 m_Stat.onEraseSuccess();
431 m_Stat.onEraseFailed();
435 template <typename Q, typename Compare, typename Func>
436 bool erase_( Q const& val, Compare cmp, Func f )
438 size_t nHash = hash_value( val );
439 split_list::details::search_value_type<Q const> sv( val, split_list::regular_hash( nHash ));
440 dummy_node_type * pHead = get_bucket( nHash );
441 assert( pHead != nullptr );
443 if ( m_List.erase_at( pHead, sv, cmp, f )) {
445 m_Stat.onEraseSuccess();
448 m_Stat.onEraseFailed();
455 /// Initialize split-ordered list of default capacity
457 The default capacity is defined in bucket table constructor.
458 See split_list::expandable_bucket_table, split_list::static_ducket_table
459 which selects by split_list::dynamic_bucket_table option.
462 : m_nBucketCountLog2(1)
467 /// Initialize split-ordered list
469 size_t nItemCount ///< estimate average of item count
470 , size_t nLoadFactor = 1 ///< load factor - average item count per bucket. Small integer up to 8, default is 1.
472 : m_Buckets( nItemCount, nLoadFactor )
473 , m_nBucketCountLog2(1)
481 The function inserts \p val in the set if it does not contain
482 an item with key equal to \p val.
484 The function makes RCU lock internally.
486 Returns \p true if \p val is placed into the set, \p false otherwise.
488 bool insert( value_type& val )
490 size_t nHash = hash_value( val );
491 dummy_node_type * pHead = get_bucket( nHash );
492 assert( pHead != nullptr );
494 node_traits::to_node_ptr( val )->m_nHash = split_list::regular_hash( nHash );
496 if ( m_List.insert_at( pHead, val )) {
498 m_Stat.onInsertSuccess();
501 m_Stat.onInsertFailed();
507 This function is intended for derived non-intrusive containers.
509 The function allows to split creating of new item into two part:
510 - create item with key only
511 - insert new item into the set
512 - if inserting is success, calls \p f functor to initialize value-field of \p val.
514 The functor signature is:
516 void func( value_type& val );
518 where \p val is the item inserted.
520 The function makes RCU lock internally.
522 @warning For \ref cds_intrusive_MichaelList_rcu "MichaelList" as the bucket see \ref cds_intrusive_item_creating "insert item troubleshooting".
523 \ref cds_intrusive_LazyList_rcu "LazyList" provides exclusive access to inserted item and does not require any node-level
526 template <typename Func>
527 bool insert( value_type& val, Func f )
529 size_t nHash = hash_value( val );
530 dummy_node_type * pHead = get_bucket( nHash );
531 assert( pHead != nullptr );
533 node_traits::to_node_ptr( val )->m_nHash = split_list::regular_hash( nHash );
535 if ( m_List.insert_at( pHead, val, f )) {
537 m_Stat.onInsertSuccess();
540 m_Stat.onInsertFailed();
544 /// Ensures that the \p val exists in the set
546 The operation performs inserting or changing data with lock-free manner.
548 If the item \p val is not found in the set, then \p val is inserted into the set.
549 Otherwise, the functor \p func is called with item found.
550 The functor signature is:
552 void func( bool bNew, value_type& item, value_type& val );
555 - \p bNew - \p true if the item has been inserted, \p false otherwise
556 - \p item - item of the set
557 - \p val - argument \p val passed into the \p ensure function
558 If new item has been inserted (i.e. \p bNew is \p true) then \p item and \p val arguments
559 refers to the same thing.
561 The function makes RCU lock internally.
563 Returns std::pair<bool, bool> where \p first is \p true if operation is successfull,
564 \p second is \p true if new item has been added or \p false if the item with \p key
565 already is in the set.
567 @warning For \ref cds_intrusive_MichaelList_rcu "MichaelList" as the bucket see \ref cds_intrusive_item_creating "insert item troubleshooting".
568 \ref cds_intrusive_LazyList_rcu "LazyList" provides exclusive access to inserted item and does not require any node-level
571 template <typename Func>
572 std::pair<bool, bool> ensure( value_type& val, Func func )
574 size_t nHash = hash_value( val );
575 dummy_node_type * pHead = get_bucket( nHash );
576 assert( pHead != nullptr );
578 node_traits::to_node_ptr( val )->m_nHash = split_list::regular_hash( nHash );
580 std::pair<bool, bool> bRet = m_List.ensure_at( pHead, val, func );
581 if ( bRet.first && bRet.second ) {
583 m_Stat.onEnsureNew();
586 m_Stat.onEnsureExist();
590 /// Unlinks the item \p val from the set
592 The function searches the item \p val in the set and unlinks it from the set
593 if it is found and is equal to \p val.
595 Difference between \ref erase and \p unlink functions: \p erase finds <i>a key</i>
596 and deletes the item found. \p unlink finds an item by key and deletes it
597 only if \p val is an item of that set, i.e. the pointer to item found
598 is equal to <tt> &val </tt>.
600 RCU \p synchronize method can be called, therefore, RCU should not be locked.
602 The function returns \p true if success and \p false otherwise.
604 bool unlink( value_type& val )
606 size_t nHash = hash_value( val );
607 dummy_node_type * pHead = get_bucket( nHash );
608 assert( pHead != nullptr );
610 if ( m_List.unlink_at( pHead, val ) ) {
612 m_Stat.onEraseSuccess();
615 m_Stat.onEraseFailed();
619 /// Deletes the item from the set
620 /** \anchor cds_intrusive_SplitListSet_rcu_erase
621 The function searches an item with key equal to \p key in the set,
622 unlinks it from the set, and returns \p true.
623 If the item with key equal to \p key is not found the function return \p false.
625 Difference between \ref erase and \p unlink functions: \p erase finds <i>a key</i>
626 and deletes the item found. \p unlink finds an item by key and deletes it
627 only if \p key is an item of that set, i.e. the pointer to item found
628 is equal to <tt> &key </tt>.
630 RCU \p synchronize method can be called, therefore, RCU should not be locked.
632 Note the hash functor should accept a parameter of type \p Q that can be not the same as \p value_type.
634 template <typename Q>
635 bool erase( Q const& key )
637 return erase_( key, key_comparator() );
640 /// Deletes the item from the set using \p pred for searching
642 The function is an analog of \ref cds_intrusive_SplitListSet_rcu_erase "erase(Q const&)"
643 but \p cmp is used for key compare.
644 \p Less has the interface like \p std::less.
645 \p pred must imply the same element order as the comparator used for building the set.
647 template <typename Q, typename Less>
648 bool erase_with( Q const& key, Less pred )
650 return erase_( key, typename wrapped_ordered_list::template make_compare_from_less<Less>() );
653 /// Deletes the item from the set
654 /** \anchor cds_intrusive_SplitListSet_rcu_erase_func
655 The function searches an item with key equal to \p key in the set,
656 call \p f functor with item found, unlinks it from the set, and returns \p true.
657 The \ref disposer specified by \p OrderedList class template parameter is called
658 by garbage collector \p GC asynchronously.
660 The \p Func interface is
663 void operator()( value_type const& item );
667 If the item with key equal to \p key is not found the function return \p false.
669 RCU \p synchronize method can be called, therefore, RCU should not be locked.
671 Note the hash functor should accept a parameter of type \p Q that can be not the same as \p value_type.
673 template <typename Q, typename Func>
674 bool erase( Q const& key, Func f )
676 return erase_( key, key_comparator(), f );
679 /// Deletes the item from the set using \p pred for searching
681 The function is an analog of \ref cds_intrusive_SplitListSet_rcu_erase_func "erase(Q const&, Func)"
682 but \p cmp is used for key compare.
683 \p Less has the interface like \p std::less.
684 \p pred must imply the same element order as the comparator used for building the set.
686 template <typename Q, typename Less, typename Func>
687 bool erase_with( Q const& key, Less pred, Func f )
689 return erase_( key, typename wrapped_ordered_list::template make_compare_from_less<Less>(), f );
692 /// Extracts an item from the set
693 /** \anchor cds_intrusive_SplitListSet_rcu_extract
694 The function searches an item with key equal to \p key in the set,
695 unlinks it, and returns \ref cds::urcu::exempt_ptr "exempt_ptr" pointer to the item found.
696 If the item with the key equal to \p key is not found the function returns an empty \p exempt_ptr.
698 @note The function does NOT call RCU read-side lock or synchronization,
699 and does NOT dispose the item found. It just excludes the item from the set
700 and returns a pointer to item found.
701 You should lock RCU before calling of the function, and you should synchronize RCU
702 outside the RCU lock before reusing returned pointer.
705 typedef cds::urcu::gc< general_buffered<> > rcu;
706 typedef cds::intrusive::MichaelList< rcu, Foo > rcu_michael_list;
707 typedef cds::intrusive::SplitListSet< rcu, rcu_michael_list, foo_traits > rcu_splitlist_set;
709 rcu_splitlist_set theSet;
712 rcu_splitlist_set::exempt_ptr p;
714 // first, we should lock RCU
715 rcu_splitlist_set::rcu_lock lock;
717 // Now, you can apply extract function
718 // Note that you must not delete the item found inside the RCU lock
719 p = theList.extract( 10 );
721 // do something with p
726 // We may safely release p here
727 // release() passes the pointer to RCU reclamation cycle:
728 // it invokes RCU retire_ptr function with the disposer you provided for rcu_michael_list.
732 template <typename Q>
733 exempt_ptr extract( Q const& key )
735 return exempt_ptr(extract_( key, key_comparator() ));
738 /// Extracts an item from the set using \p pred for searching
740 The function is an analog of \ref cds_intrusive_SplitListSet_rcu_extract "extract(exempt_ptr&, Q const&)"
741 but \p pred is used for key compare.
742 \p Less functor has the interface like \p std::less.
743 \p pred must imply the same element order as the comparator used for building the set.
745 template <typename Q, typename Less>
746 exempt_ptr extract_with( Q const& key, Less pred )
748 return exempt_ptr( extract_with_( key, pred ));
751 /// Finds the key \p key
752 /** \anchor cds_intrusive_SplitListSet_rcu_find_func
753 The function searches the item with key equal to \p key and calls the functor \p f for item found.
754 The interface of \p Func functor is:
757 void operator()( value_type& item, Q& key );
760 where \p item is the item found, \p key is the <tt>find</tt> function argument.
762 The functor can change non-key fields of \p item. Note that the functor is only guarantee
763 that \p item cannot be disposed during functor is executing.
764 The functor does not serialize simultaneous access to the set \p item. If such access is
765 possible you must provide your own synchronization schema on item level to exclude unsafe item modifications.
767 The \p key argument is non-const since it can be used as \p f functor destination i.e., the functor
768 can modify both arguments.
770 Note the hash functor specified for class \p Traits template parameter
771 should accept a parameter of type \p Q that can be not the same as \p value_type.
773 The function applies RCU lock internally.
775 The function returns \p true if \p key is found, \p false otherwise.
777 template <typename Q, typename Func>
778 bool find( Q& key, Func f )
780 return find_( key, key_comparator(), f );
783 /// Finds the key \p key with \p pred predicate for comparing
785 The function is an analog of \ref cds_intrusive_SplitListSet_rcu_find_func "find(Q&, Func)"
786 but \p cmp is used for key compare.
787 \p Less has the interface like \p std::less.
788 \p cmp must imply the same element order as the comparator used for building the set.
790 template <typename Q, typename Less, typename Func>
791 bool find_with( Q& key, Less pred, Func f )
793 return find_( key, typename wrapped_ordered_list::template make_compare_from_less<Less>(), f );
797 /// Finds the key \p key
798 /** \anchor cds_intrusive_SplitListSet_rcu_find_val
799 The function searches the item with key equal to \p key
800 and returns \p true if \p key found or \p false otherwise.
802 template <typename Q>
803 bool find( Q const& key )
805 return find_value( key, key_comparator() );
808 /// Finds the key \p key with \p pred predicate for comparing
810 The function is an analog of \ref cds_intrusive_SplitListSet_rcu_find_val "find(Q const&)"
811 but \p cmp is used for key compare.
812 \p Less has the interface like \p std::less.
813 \p pred must imply the same element order as the comparator used for building the set.
815 template <typename Q, typename Less>
816 bool find_with( Q const& key, Less pred )
818 return find_value( key, typename wrapped_ordered_list::template make_compare_from_less<Less>() );
821 /// Finds the key \p key and return the item found
822 /** \anchor cds_intrusive_SplitListSet_rcu_get
823 The function searches the item with key equal to \p key and returns the pointer to item found.
824 If \p key is not found it returns \p nullptr.
826 Note the compare functor should accept a parameter of type \p Q that can be not the same as \p value_type.
828 RCU should be locked before call of this function.
829 Returned item is valid only while RCU is locked:
831 cds::intrusive::SplitListSet< your_template_parameters > theSet;
835 hash_set::rcu_lock lock;
837 foo * pVal = theSet.get( 5 );
842 // Unlock RCU by rcu_lock destructor
843 // pVal can be retired by disposer at any time after RCU has been unlocked
847 template <typename Q>
848 value_type * get( Q const& key )
850 return get_( key, key_comparator() );
853 /// Finds the key \p key and return the item found
855 The function is an analog of \ref cds_intrusive_SplitListSet_rcu_get "get(Q const&)"
856 but \p pred is used for comparing the keys.
858 \p Less functor has the semantics like \p std::less but should take arguments of type \ref value_type and \p Q
860 \p pred must imply the same element order as the comparator used for building the set.
862 template <typename Q, typename Less>
863 value_type * get_with( Q const& key, Less pred )
865 return get_( key, typename wrapped_ordered_list::template make_compare_from_less<Less>());
869 /// Returns item count in the set
872 return m_ItemCounter;
875 /// Checks if the set is empty
877 Emptiness is checked by item counting: if item count is zero then the set is empty.
878 Thus, the correct item counting feature is an important part of split-list set implementation.
885 /// Clears the set (not atomic)
888 iterator it = begin();
889 while ( it != end() ) {
897 /// Returns internal statistics
898 stat const& statistics() const
905 template <bool IsConst>
907 :public split_list::details::iterator_type<node_traits, ordered_list, IsConst>
909 typedef split_list::details::iterator_type<node_traits, ordered_list, IsConst> iterator_base_class;
910 typedef typename iterator_base_class::list_iterator list_iterator;
913 : iterator_base_class()
916 iterator_type( iterator_type const& src )
917 : iterator_base_class( src )
920 // This ctor should be protected...
921 iterator_type( list_iterator itCur, list_iterator itEnd )
922 : iterator_base_class( itCur, itEnd )
929 The forward iterator for a split-list has some features:
930 - it has no post-increment operator
931 - it depends on iterator of underlying \p OrderedList
932 - The iterator cannot be moved across thread boundary since it may contain GC's guard that is thread-private GC data.
933 - Iterator ensures thread-safety even if you delete the item that iterator points to. However, in case of concurrent
934 deleting operations it is no guarantee that you iterate all item in the split-list.
936 Therefore, the use of iterators in concurrent environment is not good idea. Use the iterator on the concurrent container
937 for debug purpose only.
939 typedef iterator_type<false> iterator;
940 /// Const forward iterator
942 For iterator's features and requirements see \ref iterator
944 typedef iterator_type<true> const_iterator;
946 /// Returns a forward iterator addressing the first element in a split-list
948 For empty list \code begin() == end() \endcode
952 return iterator( m_List.begin(), m_List.end() );
955 /// Returns an iterator that addresses the location succeeding the last element in a split-list
957 Do not use the value returned by <tt>end</tt> function to access any item.
959 The returned value can be used only to control reaching the end of the split-list.
960 For empty list \code begin() == end() \endcode
964 return iterator( m_List.end(), m_List.end() );
967 /// Returns a forward const iterator addressing the first element in a split-list
968 const_iterator begin() const
972 /// Returns a forward const iterator addressing the first element in a split-list
973 const_iterator cbegin() const
975 return const_iterator( m_List.cbegin(), m_List.cend() );
978 /// Returns an const iterator that addresses the location succeeding the last element in a split-list
979 const_iterator end() const
983 /// Returns an const iterator that addresses the location succeeding the last element in a split-list
984 const_iterator cend() const
986 return const_iterator( m_List.cend(), m_List.cend() );
991 }} // namespace cds::intrusive
993 #endif // #ifndef __CDS_INTRUSIVE_SPLIT_LIST_RCU_H