X-Git-Url: http://plrg.eecs.uci.edu/git/?a=blobdiff_plain;f=cds%2Fintrusive%2Fsplit_list.h;h=9a7e198e55a61d31bacc2671a1557cb0b22b074c;hb=cd515d6402be81b84e2eb0c9d4cf0a1ca9e4d95a;hp=acab1ea388f0765eb6a9d90f92a29e2d21dd6c6c;hpb=1503681e52c392bee2329cf66c910be0f8640105;p=libcds.git diff --git a/cds/intrusive/split_list.h b/cds/intrusive/split_list.h index acab1ea3..9a7e198e 100644 --- a/cds/intrusive/split_list.h +++ b/cds/intrusive/split_list.h @@ -1,8 +1,37 @@ -//$$CDS-header$$ - -#ifndef __CDS_INTRUSIVE_SPLIT_LIST_H -#define __CDS_INTRUSIVE_SPLIT_LIST_H - +/* + This file is a part of libcds - Concurrent Data Structures library + + (C) Copyright Maxim Khizhinsky (libcds.dev@gmail.com) 2006-2016 + + Source code repo: http://github.com/khizmax/libcds/ + Download: http://sourceforge.net/projects/libcds/files/ + + Redistribution and use in source and binary forms, with or without + modification, are permitted provided that the following conditions are met: + + * Redistributions of source code must retain the above copyright notice, this + list of conditions and the following disclaimer. + + * Redistributions in binary form must reproduce the above copyright notice, + this list of conditions and the following disclaimer in the documentation + and/or other materials provided with the distribution. + + THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" + AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE + DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE + FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR + SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER + CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, + OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. +*/ + +#ifndef CDSLIB_INTRUSIVE_SPLIT_LIST_H +#define CDSLIB_INTRUSIVE_SPLIT_LIST_H + +#include #include namespace cds { namespace intrusive { @@ -17,30 +46,30 @@ namespace cds { namespace intrusive { The split-ordered list is a lock-free implementation of an extensible unbounded hash table. It uses original recursive split-ordering algorithm discovered by Ori Shalev and Nir Shavit that allows to split buckets - without moving an item on resizing. + without item moving on resizing. \anchor cds_SplitList_algo_desc Short description [from [2003] Ori Shalev, Nir Shavit "Split-Ordered Lists - Lock-free Resizable Hash Tables"] The algorithm keeps all the items in one lock-free linked list, and gradually assigns the bucket pointers to - the places in the list where a sublist of “correct” items can be found. A bucket is initialized upon first - access by assigning it to a new “dummy” node (dashed contour) in the list, preceding all items that should be - in that bucket. A newly created bucket splits an older bucket’s chain, reducing the access cost to its items. The - table uses a modulo 2**i hash (there are known techniques for “pre-hashing” before a modulo 2**i hash + the places in the list where a sublist of 'correct' items can be found. A bucket is initialized upon first + access by assigning it to a new 'dummy' node (dashed contour) in the list, preceding all items that should be + in that bucket. A newly created bucket splits an older bucket's chain, reducing the access cost to its items. The + table uses a modulo 2**i hash (there are known techniques for 'pre-hashing' before a modulo 2**i hash to overcome possible binary correlations among values). The table starts at size 2 and repeatedly doubles in size. Unlike moving an item, the operation of directing a bucket pointer can be done - in a single CAS operation, and since items are not moved, they are never “lost”. + in a single CAS operation, and since items are not moved, they are never 'lost'. However, to make this approach work, one must be able to keep the items in the - list sorted in such a way that any bucket’s sublist can be “split” by directing a new + list sorted in such a way that any bucket's sublist can be 'split' by directing a new bucket pointer within it. This operation must be recursively repeatable, as every split bucket may be split again and again as the hash table grows. To achieve this goal the authors introduced recursive split-ordering, a new ordering on keys that keeps items in a given bucket adjacent in the list throughout the repeated splitting process. Magically, yet perhaps not surprisingly, recursive split-ordering is achieved by - simple binary reversal: reversing the bits of the hash key so that the new key’s + simple binary reversal: reversing the bits of the hash key so that the new key's most significant bits (MSB) are those that were originally its least significant. The split-order keys of regular nodes are exactly the bit-reverse image of the original keys after turning on their MSB. For example, items 9 and 13 are in the 1 mod @@ -49,7 +78,7 @@ namespace cds { namespace intrusive { To insert (respectively delete or search for) an item in the hash table, hash its key to the appropriate bucket using recursive split-ordering, follow the pointer to - the appropriate location in the sorted items list, and traverse the list until the key’s + the appropriate location in the sorted items list, and traverse the list until the key's proper location in the split-ordering (respectively until the key or a key indicating the item is not in the list is found). Because of the combinatorial structure induced by the split-ordering, this will require traversal of no more than an expected constant number of items. @@ -60,13 +89,13 @@ namespace cds { namespace intrusive { Implementation Template parameters are: - - \p GC - Garbage collector used. Note the \p GC must be the same as the GC used for \p OrderedList - - \p OrderedList - ordered list implementation used as bucket for hash set, for example, MichaelList, LazyList. + - \p GC - Garbage collector. Note the \p GC must be the same as the \p GC used for \p OrderedList + - \p OrderedList - ordered list implementation used as a bucket for hash set, for example, \p MichaelList, \p LazyList. The intrusive ordered list implementation specifies the type \p T stored in the hash-set, the reclamation schema \p GC used by hash-set, the comparison functor for the type \p T and other features specific for the ordered list. - - \p Traits - type traits. See split_list::type_traits for explanation. - Instead of defining \p Traits struct you may use option-based syntax with split_list::make_traits metafunction. + - \p Traits - split-list traits, default is \p split_list::traits. + Instead of defining \p Traits struct you may use option-based syntax with \p split_list::make_traits metafunction. There are several specialization of the split-list class for different \p GC: - for \ref cds_urcu_gc "RCU type" include - see @@ -80,7 +109,7 @@ namespace cds { namespace intrusive { Some member functions of split-ordered list accept the key parameter of type \p Q which differs from \p value_type. It is expected that type \p Q contains full key of \p value_type, and for equal keys of type \p Q and \p value_type the hash values of these keys must be equal too. - The hash functor Traits::hash should accept parameters of both type: + The hash functor \p Traits::hash should accept parameters of both type: \code // Our node type struct Foo { @@ -133,7 +162,6 @@ namespace cds { namespace intrusive { }; // Declare base ordered-list type for split-list - // It may be any ordered list type like MichaelList, LazyList typedef cds::intrusive::MichaelList< cds::gc::HP, Foo, typename cds::intrusive::michael_list::make_traits< // hook option @@ -180,13 +208,12 @@ namespace cds { namespace intrusive { // and so on ... \endcode - */ template < class GC, class OrderedList, # ifdef CDS_DOXYGEN_INVOKED - class Traits = split_list::type_traits + class Traits = split_list::traits # else class Traits # endif @@ -194,54 +221,54 @@ namespace cds { namespace intrusive { class SplitListSet { public: - typedef Traits options ; ///< Traits template parameters - typedef GC gc ; ///< Garbage collector + typedef GC gc; ///< Garbage collector + typedef Traits traits; ///< Set traits protected: //@cond - typedef split_list::details::rebind_list_options wrapped_ordered_list; + typedef split_list::details::rebind_list_traits wrapped_ordered_list; //@endcond public: # ifdef CDS_DOXYGEN_INVOKED - typedef OrderedList ordered_list ; ///< type of ordered list used as base for split-list + typedef OrderedList ordered_list; ///< type of ordered list used as a base for split-list # else typedef typename wrapped_ordered_list::result ordered_list; # endif - typedef typename ordered_list::value_type value_type ; ///< type of value stored in the split-list - typedef typename ordered_list::key_comparator key_comparator ; ///< key comparison functor - typedef typename ordered_list::disposer disposer ; ///< Node disposer functor + typedef typename ordered_list::value_type value_type; ///< type of value stored in the split-list + typedef typename ordered_list::key_comparator key_comparator; ///< key comparison functor + typedef typename ordered_list::disposer disposer; ///< Node disposer functor /// Hash functor for \p %value_type and all its derivatives that you use - typedef typename cds::opt::v::hash_selector< typename options::hash >::type hash; + typedef typename cds::opt::v::hash_selector< typename traits::hash >::type hash; - typedef typename options::item_counter item_counter ; ///< Item counter type - typedef typename options::back_off back_off ; ///< back-off strategy for spinning - typedef typename options::memory_model memory_model ; ///< Memory ordering. See cds::opt::memory_model option - typedef typename ordered_list::guarded_ptr guarded_ptr; ///< Guarded pointer + typedef typename traits::item_counter item_counter; ///< Item counter type + typedef typename traits::back_off back_off; ///< back-off strategy for spinning + typedef typename traits::memory_model memory_model; ///< Memory ordering. See cds::opt::memory_model option + typedef typename traits::stat stat; ///< Internal statistics, see \p spit_list::stat + typedef typename ordered_list::guarded_ptr guarded_ptr; ///< Guarded pointer protected: - typedef typename ordered_list::node_type list_node_type ; ///< Node type as declared in ordered list - typedef split_list::node node_type ; ///< split-list node type - typedef node_type dummy_node_type ; ///< dummy node type + typedef typename ordered_list::node_type list_node_type; ///< Node type as declared in ordered list + typedef split_list::node node_type; ///< split-list node type + typedef node_type dummy_node_type; ///< dummy node type /// Split-list node traits /** - This traits is intended for converting between underlying ordered list node type \ref list_node_type - and split-list node type \ref node_type + This traits is intended for converting between underlying ordered list node type \p list_node_type + and split-list node type \p node_type */ typedef split_list::node_traits node_traits; //@cond /// Bucket table implementation typedef typename split_list::details::bucket_table_selector< - options::dynamic_bucket_table + traits::dynamic_bucket_table , gc , dummy_node_type - , opt::allocator< typename options::allocator > + , opt::allocator< typename traits::allocator > , opt::memory_model< memory_model > >::type bucket_table; - //@endcond protected: @@ -269,11 +296,11 @@ namespace cds { namespace intrusive { } template - std::pair ensure_at( dummy_node_type * pHead, value_type& val, Func func ) + std::pair update_at( dummy_node_type * pHead, value_type& val, Func func, bool bAllowInsert ) { assert( pHead != nullptr ); bucket_head_type h(pHead); - return base_class::ensure_at( h, val, func ); + return base_class::update_at( h, val, func, bAllowInsert ); } bool unlink_at( dummy_node_type * pHead, value_type& val ) @@ -300,7 +327,7 @@ namespace cds { namespace intrusive { } template - bool extract_at( dummy_node_type * pHead, typename gc::Guard& guard, split_list::details::search_value_type const& val, Compare cmp ) + bool extract_at( dummy_node_type * pHead, typename guarded_ptr::native_guard& guard, split_list::details::search_value_type const& val, Compare cmp ) { assert( pHead != nullptr ); bucket_head_type h(pHead); @@ -324,7 +351,7 @@ namespace cds { namespace intrusive { } template - bool get_at( dummy_node_type * pHead, typename gc::Guard& guard, split_list::details::search_value_type const& val, Compare cmp ) + bool get_at( dummy_node_type * pHead, typename guarded_ptr::native_guard& guard, split_list::details::search_value_type const& val, Compare cmp ) { assert( pHead != nullptr ); bucket_head_type h(pHead); @@ -344,22 +371,27 @@ namespace cds { namespace intrusive { //@endcond protected: - ordered_list_wrapper m_List ; ///< Ordered list containing split-list items - bucket_table m_Buckets ; ///< bucket table - atomics::atomic m_nBucketCountLog2 ; ///< log2( current bucket count ) - item_counter m_ItemCounter ; ///< Item counter - hash m_HashFunctor ; ///< Hash functor + ordered_list_wrapper m_List; ///< Ordered list containing split-list items + bucket_table m_Buckets; ///< bucket table + atomics::atomic m_nBucketCountLog2; ///< log2( current bucket count ) + atomics::atomic m_nMaxItemCount; ///< number of items container can hold, before we have to resize + item_counter m_ItemCounter; ///< Item counter + hash m_HashFunctor; ///< Hash functor + stat m_Stat; ///< Internal statistics protected: //@cond - typedef cds::details::Allocator< dummy_node_type, typename options::allocator > dummy_node_allocator; - static dummy_node_type * alloc_dummy_node( size_t nHash ) + typedef cds::details::Allocator< dummy_node_type, typename traits::allocator > dummy_node_allocator; + + dummy_node_type * alloc_dummy_node( size_t nHash ) { + m_Stat.onHeadNodeAllocated(); return dummy_node_allocator().New( nHash ); } - static void free_dummy_node( dummy_node_type * p ) + void free_dummy_node( dummy_node_type * p ) { dummy_node_allocator().Delete( p ); + m_Stat.onHeadNodeFreed(); } /// Calculates hash value of \p key @@ -371,13 +403,13 @@ namespace cds { namespace intrusive { size_t bucket_no( size_t nHash ) const { - return nHash & ( (1 << m_nBucketCountLog2.load(atomics::memory_order_relaxed)) - 1 ); + return nHash & ( (1 << m_nBucketCountLog2.load(memory_model::memory_order_relaxed)) - 1 ); } static size_t parent_bucket( size_t nBucket ) { assert( nBucket > 0 ); - return nBucket & ~( 1 << bitop::MSBnz( nBucket ) ); + return nBucket & ~( 1 << bitop::MSBnz( nBucket )); } dummy_node_type * init_bucket( size_t nBucket ) @@ -388,15 +420,17 @@ namespace cds { namespace intrusive { dummy_node_type * pParentBucket = m_Buckets.bucket( nParent ); if ( pParentBucket == nullptr ) { pParentBucket = init_bucket( nParent ); + m_Stat.onRecursiveInitBucket(); } assert( pParentBucket != nullptr ); // Allocate a dummy node for new bucket { - dummy_node_type * pBucket = alloc_dummy_node( split_list::dummy_hash( nBucket ) ); - if ( m_List.insert_aux_node( pParentBucket, pBucket ) ) { + dummy_node_type * pBucket = alloc_dummy_node( split_list::dummy_hash( nBucket )); + if ( m_List.insert_aux_node( pParentBucket, pBucket )) { m_Buckets.bucket( nBucket, pBucket ); + m_Stat.onNewBucket(); return pBucket; } free_dummy_node( pBucket ); @@ -407,13 +441,14 @@ namespace cds { namespace intrusive { // In this point, we must wait while nBucket is empty. // The compiler can decide that waiting loop can be "optimized" (stripped) // To prevent this situation, we use waiting on volatile bucket_head_ptr pointer. - // + m_Stat.onBucketInitContenton(); back_off bkoff; while ( true ) { dummy_node_type volatile * p = m_Buckets.bucket( nBucket ); if ( p != nullptr ) return const_cast( p ); bkoff(); + m_Stat.onBusyWaitBucketInit(); } } @@ -425,7 +460,7 @@ namespace cds { namespace intrusive { if ( pHead == nullptr ) pHead = init_bucket( nBucket ); - assert( pHead->is_dummy() ); + assert( pHead->is_dummy()); return pHead; } @@ -433,10 +468,11 @@ namespace cds { namespace intrusive { void init() { // GC and OrderedList::gc must be the same - static_assert(( std::is_same::value ), "GC and OrderedList::gc must be the same"); + static_assert( std::is_same::value, "GC and OrderedList::gc must be the same"); // atomicity::empty_item_counter is not allowed as a item counter - static_assert(( !std::is_same::value ), "atomicity::empty_item_counter is not allowed as a item counter"); + static_assert( !std::is_same::value, + "cds::atomicity::empty_item_counter is not allowed as a item counter"); // Initialize bucket 0 dummy_node_type * pNode = alloc_dummy_node( 0 /*split_list::dummy_hash(0)*/ ); @@ -447,13 +483,31 @@ namespace cds { namespace intrusive { m_Buckets.bucket( 0, pNode ); } - void inc_item_count() + static size_t max_item_count( size_t nBucketCount, size_t nLoadFactor ) { - size_t sz = m_nBucketCountLog2.load(atomics::memory_order_relaxed); - if ( ( ++m_ItemCounter >> sz ) > m_Buckets.load_factor() && ((size_t)(1 << sz )) < m_Buckets.capacity() ) - { - m_nBucketCountLog2.compare_exchange_strong( sz, sz + 1, atomics::memory_order_seq_cst, atomics::memory_order_relaxed ); + return nBucketCount * nLoadFactor; + } + + void inc_item_count() + { + size_t nMaxCount = m_nMaxItemCount.load(memory_model::memory_order_relaxed); + if ( ++m_ItemCounter <= nMaxCount ) + return; + + size_t sz = m_nBucketCountLog2.load(memory_model::memory_order_relaxed); + const size_t nBucketCount = static_cast(1) << sz; + if ( nBucketCount < m_Buckets.capacity()) { + // we may grow the bucket table + const size_t nLoadFactor = m_Buckets.load_factor(); + if ( nMaxCount < max_item_count( nBucketCount, nLoadFactor )) + return; // someone already have updated m_nBucketCountLog2, so stop here + + m_nMaxItemCount.compare_exchange_strong( nMaxCount, max_item_count( nBucketCount << 1, nLoadFactor ), + memory_model::memory_order_relaxed, atomics::memory_order_relaxed ); + m_nBucketCountLog2.compare_exchange_strong( sz, sz + 1, memory_model::memory_order_relaxed, atomics::memory_order_relaxed ); } + else + m_nMaxItemCount.store( std::numeric_limits::max(), memory_model::memory_order_relaxed ); } template @@ -464,8 +518,10 @@ namespace cds { namespace intrusive { dummy_node_type * pHead = get_bucket( nHash ); assert( pHead != nullptr ); - return m_List.find_at( pHead, sv, cmp, - [&f](value_type& item, split_list::details::search_value_type& val){ f(item, val.val ); }); + return m_Stat.onFind( + m_List.find_at( pHead, sv, cmp, + [&f](value_type& item, split_list::details::search_value_type& val){ f(item, val.val ); }) + ); } template @@ -476,28 +532,28 @@ namespace cds { namespace intrusive { dummy_node_type * pHead = get_bucket( nHash ); assert( pHead != nullptr ); - return m_List.find_at( pHead, sv, cmp ); + return m_Stat.onFind( m_List.find_at( pHead, sv, cmp )); } template - bool get_( typename gc::Guard& guard, Q const& val, Compare cmp ) + bool get_( typename guarded_ptr::native_guard& guard, Q const& val, Compare cmp ) { size_t nHash = hash_value( val ); split_list::details::search_value_type sv( val, split_list::regular_hash( nHash )); dummy_node_type * pHead = get_bucket( nHash ); assert( pHead != nullptr ); - return m_List.get_at( pHead, guard, sv, cmp ); + return m_Stat.onFind( m_List.get_at( pHead, guard, sv, cmp )); } template - bool get_( typename gc::Guard& guard, Q const& key) + bool get_( typename guarded_ptr::native_guard& guard, Q const& key ) { return get_( guard, key, key_comparator()); } template - bool get_with_( typename gc::Guard& guard, Q const& key, Less ) + bool get_with_( typename guarded_ptr::native_guard& guard, Q const& key, Less ) { return get_( guard, key, typename wrapped_ordered_list::template make_compare_from_less()); } @@ -512,8 +568,10 @@ namespace cds { namespace intrusive { if ( m_List.erase_at( pHead, sv, cmp, f )) { --m_ItemCounter; + m_Stat.onEraseSuccess(); return true; } + m_Stat.onEraseFailed(); return false; } @@ -525,51 +583,55 @@ namespace cds { namespace intrusive { dummy_node_type * pHead = get_bucket( nHash ); assert( pHead != nullptr ); - if ( m_List.erase_at( pHead, sv, cmp ) ) { + if ( m_List.erase_at( pHead, sv, cmp )) { --m_ItemCounter; + m_Stat.onEraseSuccess(); return true; } + m_Stat.onEraseFailed(); return false; } template - bool extract_( typename gc::Guard& guard, Q const& val, Compare cmp ) + bool extract_( typename guarded_ptr::native_guard& guard, Q const& val, Compare cmp ) { size_t nHash = hash_value( val ); - split_list::details::search_value_type sv( val, split_list::regular_hash( nHash )); + split_list::details::search_value_type sv( val, split_list::regular_hash( nHash )); dummy_node_type * pHead = get_bucket( nHash ); assert( pHead != nullptr ); - if ( m_List.extract_at( pHead, guard, sv, cmp ) ) { + if ( m_List.extract_at( pHead, guard, sv, cmp )) { --m_ItemCounter; + m_Stat.onExtractSuccess(); return true; } + m_Stat.onExtractFailed(); return false; } template - bool extract_( typename gc::Guard& guard, Q const& key) + bool extract_( typename guarded_ptr::native_guard& guard, Q const& key ) { return extract_( guard, key, key_comparator()); } template - bool extract_with_( typename gc::Guard& guard, Q const& key, Less ) + bool extract_with_( typename guarded_ptr::native_guard& guard, Q const& key, Less ) { - return extract_( guard, key, typename wrapped_ordered_list::template make_compare_from_less() ); + return extract_( guard, key, typename wrapped_ordered_list::template make_compare_from_less()); } - //@endcond public: /// Initialize split-ordered list of default capacity /** The default capacity is defined in bucket table constructor. - See split_list::expandable_bucket_table, split_list::static_ducket_table - which selects by split_list::dynamic_bucket_table option. + See \p split_list::expandable_bucket_table, \p split_list::static_bucket_table + which selects by \p split_list::dynamic_bucket_table option. */ SplitListSet() : m_nBucketCountLog2(1) + , m_nMaxItemCount( max_item_count(2, m_Buckets.load_factor())) { init(); } @@ -581,6 +643,7 @@ namespace cds { namespace intrusive { ) : m_Buckets( nItemCount, nLoadFactor ) , m_nBucketCountLog2(1) + , m_nMaxItemCount( max_item_count(2, m_Buckets.load_factor())) { init(); } @@ -603,8 +666,10 @@ namespace cds { namespace intrusive { if ( m_List.insert_at( pHead, val )) { inc_item_count(); + m_Stat.onInsertSuccess(); return true; } + m_Stat.onInsertFailed(); return false; } @@ -621,10 +686,12 @@ namespace cds { namespace intrusive { \code void func( value_type& val ); \endcode - where \p val is the item inserted. User-defined functor \p f should guarantee that during changing - \p val no any other changes could be made on this set's item by concurrent threads. - The user-defined functor is called only if the inserting is success and may be passed by reference - using \p std::ref. + where \p val is the item inserted. + The user-defined functor is called only if the inserting is success. + + @warning For \ref cds_intrusive_MichaelList_hp "MichaelList" as the bucket see \ref cds_intrusive_item_creating "insert item troubleshooting". + \ref cds_intrusive_LazyList_hp "LazyList" provides exclusive access to inserted item and does not require any node-level + synchronization. */ template bool insert( value_type& val, Func f ) @@ -637,16 +704,19 @@ namespace cds { namespace intrusive { if ( m_List.insert_at( pHead, val, f )) { inc_item_count(); + m_Stat.onInsertSuccess(); return true; } + m_Stat.onInsertFailed(); return false; } - /// Ensures that the \p val exists in the set + /// Updates the node /** The operation performs inserting or changing data with lock-free manner. - If the item \p val is not found in the set, then \p val is inserted into the set. + If the item \p val is not found in the set, then \p val is inserted + iff \p bAllowInsert is \p true. Otherwise, the functor \p func is called with item found. The functor signature is: \code @@ -655,21 +725,22 @@ namespace cds { namespace intrusive { with arguments: - \p bNew - \p true if the item has been inserted, \p false otherwise - \p item - item of the set - - \p val - argument \p val passed into the \p ensure function + - \p val - argument \p val passed into the \p update() function If new item has been inserted (i.e. \p bNew is \p true) then \p item and \p val arguments refers to the same thing. - The functor can change non-key fields of the \p item; however, \p func must guarantee - that during changing no any other modifications could be made on this item by concurrent threads. - - You can pass \p func argument by value or by reference using \p std::ref. + The functor may change non-key fields of the \p item. Returns std::pair where \p first is \p true if operation is successfull, - \p second is \p true if new item has been added or \p false if the item with \p key - already is in the set. + \p second is \p true if new item has been added or \p false if the item with \p val + already is in the list. + + @warning For \ref cds_intrusive_MichaelList_hp "MichaelList" as the bucket see \ref cds_intrusive_item_creating "insert item troubleshooting". + \ref cds_intrusive_LazyList_hp "LazyList" provides exclusive access to inserted item and does not require any node-level + synchronization. */ template - std::pair ensure( value_type& val, Func func ) + std::pair update( value_type& val, Func func, bool bAllowInsert = true ) { size_t nHash = hash_value( val ); dummy_node_type * pHead = get_bucket( nHash ); @@ -677,11 +748,23 @@ namespace cds { namespace intrusive { node_traits::to_node_ptr( val )->m_nHash = split_list::regular_hash( nHash ); - std::pair bRet = m_List.ensure_at( pHead, val, func ); - if ( bRet.first && bRet.second ) + std::pair bRet = m_List.update_at( pHead, val, func, bAllowInsert ); + if ( bRet.first && bRet.second ) { inc_item_count(); + m_Stat.onEnsureNew(); + } + else + m_Stat.onEnsureExist(); return bRet; } + //@cond + template + CDS_DEPRECATED("ensure() is deprecated, use update()") + std::pair ensure( value_type& val, Func func ) + { + return update( val, func, true ); + } + //@endcond /// Unlinks the item \p val from the set /** @@ -701,48 +784,52 @@ namespace cds { namespace intrusive { dummy_node_type * pHead = get_bucket( nHash ); assert( pHead != nullptr ); - if ( m_List.unlink_at( pHead, val ) ) { + if ( m_List.unlink_at( pHead, val )) { --m_ItemCounter; + m_Stat.onEraseSuccess(); return true; } + m_Stat.onEraseFailed(); return false; } /// Deletes the item from the set /** \anchor cds_intrusive_SplitListSet_hp_erase - The function searches an item with key equal to \p val in the set, + The function searches an item with key equal to \p key in the set, unlinks it from the set, and returns \p true. - If the item with key equal to \p val is not found the function return \p false. + If the item with key equal to \p key is not found the function return \p false. Difference between \ref erase and \p unlink functions: \p erase finds a key and deletes the item found. \p unlink finds an item by key and deletes it - only if \p val is an item of that set, i.e. the pointer to item found - is equal to &val . + only if \p key is an item of that set, i.e. the pointer to item found + is equal to &key . Note the hash functor should accept a parameter of type \p Q that can be not the same as \p value_type. */ template - bool erase( Q const& val ) + bool erase( Q const& key ) { - return erase_( val, key_comparator() ); + return erase_( key, key_comparator()); } /// Deletes the item from the set with comparing functor \p pred /** + The function is an analog of \ref cds_intrusive_SplitListSet_hp_erase "erase(Q const&)" but \p pred predicate is used for key comparing. \p Less has the interface like \p std::less. \p pred must imply the same element order as the comparator used for building the set. */ template - bool erase_with( const Q& val, Less pred ) + bool erase_with( const Q& key, Less pred ) { - return erase_( val, typename wrapped_ordered_list::template make_compare_from_less() ); + CDS_UNUSED( pred ); + return erase_( key, typename wrapped_ordered_list::template make_compare_from_less()); } /// Deletes the item from the set /** \anchor cds_intrusive_SplitListSet_hp_erase_func - The function searches an item with key equal to \p val in the set, + The function searches an item with key equal to \p key in the set, call \p f functor with item found, unlinks it from the set, and returns \p true. The \ref disposer specified by \p OrderedList class template parameter is called by garbage collector \p GC asynchronously. @@ -753,16 +840,15 @@ namespace cds { namespace intrusive { void operator()( value_type const& item ); }; \endcode - The functor can be passed by reference with boost:ref - If the item with key equal to \p val is not found the function return \p false. + If the item with key equal to \p key is not found the function return \p false. Note the hash functor should accept a parameter of type \p Q that can be not the same as \p value_type. */ template - bool erase( Q const& val, Func f ) + bool erase( Q const& key, Func f ) { - return erase_( val, key_comparator(), f ); + return erase_( key, key_comparator(), f ); } /// Deletes the item from the set with comparing functor \p pred @@ -773,21 +859,22 @@ namespace cds { namespace intrusive { \p pred must imply the same element order as the comparator used for building the set. */ template - bool erase_with( Q const& val, Less pred, Func f ) + bool erase_with( Q const& key, Less pred, Func f ) { - return erase_( val, typename wrapped_ordered_list::template make_compare_from_less(), f ); + CDS_UNUSED( pred ); + return erase_( key, typename wrapped_ordered_list::template make_compare_from_less(), f ); } /// Extracts the item with specified \p key /** \anchor cds_intrusive_SplitListSet_hp_extract The function searches an item with key equal to \p key, - unlinks it from the set, and returns it in \p dest parameter. - If the item with key equal to \p key is not found the function returns \p false. + unlinks it from the set, and returns it as \p guarded_ptr. + If \p key is not found the function returns an empty guarded pointer. Note the compare functor should accept a parameter of type \p Q that may be not the same as \p value_type. - The \ref disposer specified in \p OrderedList class' template parameter is called automatically - by garbage collector \p GC when returned \ref guarded_ptr object will be destroyed or released. + The \p disposer specified in \p OrderedList class' template parameter is called automatically + by garbage collector \p GC when returned \p guarded_ptr object will be destroyed or released. @note Each \p guarded_ptr object uses the GC's guard that can be limited resource. Usage: @@ -796,24 +883,26 @@ namespace cds { namespace intrusive { splitlist_set theSet; // ... { - splitlist_set::guarded_ptr gp; - theSet.extract( gp, 5 ); - // Deal with gp - // ... - + splitlist_set::guarded_ptr gp( theSet.extract( 5 )); + if ( gp) { + // Deal with gp + // ... + } // Destructor of gp releases internal HP guard } \endcode */ template - bool extract( guarded_ptr& dest, Q const& key ) + guarded_ptr extract( Q const& key ) { - return extract_( dest.guard(), key ); + guarded_ptr gp; + extract_( gp.guard(), key ); + return gp; } /// Extracts the item using compare functor \p pred /** - The function is an analog of \ref cds_intrusive_SplitListSet_hp_extract "extract(guarded_ptr&, Q const&)" + The function is an analog of \ref cds_intrusive_SplitListSet_hp_extract "extract(Q const&)" but \p pred predicate is used for key comparing. \p Less functor has the semantics like \p std::less but should take arguments of type \ref value_type and \p Q @@ -821,45 +910,48 @@ namespace cds { namespace intrusive { \p pred must imply the same element order as the comparator used for building the set. */ template - bool extract_with( guarded_ptr& dest, Q const& key, Less pred ) + guarded_ptr extract_with( Q const& key, Less pred ) { - return extract_with_( dest.guard(), key, pred ); + guarded_ptr gp; + extract_with_( gp.guard(), key, pred ); + return gp; } - - /// Finds the key \p val + /// Finds the key \p key /** \anchor cds_intrusive_SplitListSet_hp_find_func - The function searches the item with key equal to \p val and calls the functor \p f for item found. + The function searches the item with key equal to \p key and calls the functor \p f for item found. The interface of \p Func functor is: \code struct functor { - void operator()( value_type& item, Q& val ); + void operator()( value_type& item, Q& key ); }; \endcode - where \p item is the item found, \p val is the find function argument. - - You can pass \p f argument by value or by reference using \p std::ref. + where \p item is the item found, \p key is the find function argument. The functor can change non-key fields of \p item. Note that the functor is only guarantee that \p item cannot be disposed during functor is executing. The functor does not serialize simultaneous access to the set \p item. If such access is possible you must provide your own synchronization schema on item level to exclude unsafe item modifications. - The \p val argument is non-const since it can be used as \p f functor destination i.e., the functor - can modify both arguments. - Note the hash functor specified for class \p Traits template parameter should accept a parameter of type \p Q that can be not the same as \p value_type. - The function returns \p true if \p val is found, \p false otherwise. + The function returns \p true if \p key is found, \p false otherwise. */ template - bool find( Q& val, Func f ) + bool find( Q& key, Func f ) + { + return find_( key, key_comparator(), f ); + } + //@cond + template + bool find( Q const& key, Func f ) { - return find_( val, key_comparator(), f ); + return find_( key, key_comparator(), f ); } + //@endcond - /// Finds the key \p val with \p pred predicate for comparing + /// Finds the key \p key with \p pred predicate for comparing /** The function is an analog of \ref cds_intrusive_SplitListSet_hp_find_func "find(Q&, Func)" but \p cmp is used for key compare. @@ -867,90 +959,72 @@ namespace cds { namespace intrusive { \p cmp must imply the same element order as the comparator used for building the set. */ template - bool find_with( Q& val, Less pred, Func f ) + bool find_with( Q& key, Less pred, Func f ) { - return find_( val, typename wrapped_ordered_list::template make_compare_from_less(), f ); + CDS_UNUSED( pred ); + return find_( key, typename wrapped_ordered_list::template make_compare_from_less(), f ); } - - /// Finds the key \p val - /** \anchor cds_intrusive_SplitListSet_hp_find_cfunc - The function searches the item with key equal to \p val and calls the functor \p f for item found. - The interface of \p Func functor is: - \code - struct functor { - void operator()( value_type& item, Q const& val ); - }; - \endcode - where \p item is the item found, \p val is the find function argument. - - You can pass \p f argument by value or by reference using \p std::ref. - - The functor can change non-key fields of \p item. Note that the functor is only guarantee - that \p item cannot be disposed during functor is executing. - The functor does not serialize simultaneous access to the set \p item. If such access is - possible you must provide your own synchronization schema on item level to exclude unsafe item modifications. - - Note the hash functor specified for class \p Traits template parameter - should accept a parameter of type \p Q that can be not the same as \p value_type. - - The function returns \p true if \p val is found, \p false otherwise. - */ - template - bool find( Q const& val, Func f ) - { - return find_( val, key_comparator(), f ); - } - - /// Finds the key \p val with \p pred predicate for comparing - /** - The function is an analog of \ref cds_intrusive_SplitListSet_hp_find_cfunc "find(Q const&, Func)" - but \p cmp is used for key compare. - \p Less has the interface like \p std::less. - \p cmp must imply the same element order as the comparator used for building the set. - */ + //@cond template - bool find_with( Q const& val, Less pred, Func f ) + bool find_with( Q const& key, Less pred, Func f ) { - return find_( val, typename wrapped_ordered_list::template make_compare_from_less(), f ); + CDS_UNUSED( pred ); + return find_( key, typename wrapped_ordered_list::template make_compare_from_less(), f ); } + //@endcond - /// Finds the key \p val - /** \anchor cds_intrusive_SplitListSet_hp_find_val - The function searches the item with key equal to \p val + /// Checks whether the set contains \p key + /** + The function searches the item with key equal to \p key and returns \p true if it is found, and \p false otherwise. Note the hash functor specified for class \p Traits template parameter should accept a parameter of type \p Q that can be not the same as \p value_type. - Otherwise, you may use \p find_with functions with explicit predicate for key comparing. + Otherwise, you may use \p contains( Q const&, Less pred ) functions with explicit predicate for key comparing. */ template - bool find( Q const& val ) + bool contains( Q const& key ) { - return find_( val, key_comparator() ); + return find_( key, key_comparator()); } + //@cond + template + CDS_DEPRECATED("deprecated, use contains()") + bool find( Q const& key ) + { + return contains( key ); + } + //@endcond - /// Finds the key \p val with \p pred predicate for comparing + /// Checks whether the set contains \p key using \p pred predicate for searching /** - The function is an analog of \ref cds_intrusive_SplitListSet_hp_find_val "find(Q const&)" - but \p cmp is used for key compare. - \p Less has the interface like \p std::less. - \p cmp must imply the same element order as the comparator used for building the set. + The function is an analog of contains( key ) but \p pred is used for key comparing. + \p Less functor has the interface like \p std::less. + \p Less must imply the same element order as the comparator used for building the set. */ template - bool find_with( Q const& val, Less pred ) + bool contains( Q const& key, Less pred ) + { + CDS_UNUSED( pred ); + return find_( key, typename wrapped_ordered_list::template make_compare_from_less()); + } + //@cond + template + CDS_DEPRECATED("deprecated, use contains()") + bool find_with( Q const& key, Less pred ) { - return find_( val, typename wrapped_ordered_list::template make_compare_from_less() ); + return contains( key, pred ); } + //@endcond - /// Finds the key \p val and return the item found + /// Finds the key \p key and return the item found /** \anchor cds_intrusive_SplitListSet_hp_get - The function searches the item with key equal to \p val - and assigns the item found to guarded pointer \p ptr. - The function returns \p true if \p val is found, and \p false otherwise. - If \p val is not found the \p ptr parameter is not changed. + The function searches the item with key equal to \p key + and returns the item found as \p guarded_ptr. + If \p key is not found the function returns an empty guarded pointer. - The \ref disposer specified in \p OrderedList class' template parameter is called - by garbage collector \p GC automatically when returned \ref guarded_ptr object + The \p disposer specified in \p OrderedList class' template parameter is called + by garbage collector \p GC automatically when returned \p guarded_ptr object will be destroyed or released. @note Each \p guarded_ptr object uses one GC's guard which can be limited resource. @@ -960,8 +1034,8 @@ namespace cds { namespace intrusive { splitlist_set theSet; // ... { - splitlist_set::guarded_ptr gp; - if ( theSet.get( gp, 5 )) { + splitlist_set::guarded_ptr gp = theSet.get( 5 ); + if ( gp ) { // Deal with gp //... } @@ -973,14 +1047,16 @@ namespace cds { namespace intrusive { should accept a parameter of type \p Q that can be not the same as \p value_type. */ template - bool get( guarded_ptr& ptr, Q const& val ) + guarded_ptr get( Q const& key ) { - return get_( ptr.guard(), val ); + guarded_ptr gp; + get_( gp.guard(), key ); + return gp; } - /// Finds the key \p val and return the item found + /// Finds the key \p key and return the item found /** - The function is an analog of \ref cds_intrusive_SplitListSet_hp_get "get( guarded_ptr& ptr, Q const&)" + The function is an analog of \ref cds_intrusive_SplitListSet_hp_get "get( Q const&)" but \p pred is used for comparing the keys. \p Less functor has the semantics like \p std::less but should take arguments of type \ref value_type and \p Q @@ -988,9 +1064,11 @@ namespace cds { namespace intrusive { \p pred must imply the same element order as the comparator used for building the set. */ template - bool get_with( guarded_ptr& ptr, Q const& val, Less pred ) + guarded_ptr get_with( Q const& key, Less pred ) { - return get_with_( ptr.guard(), val, pred ); + guarded_ptr gp; + get_with_( gp.guard(), key, pred ); + return gp; } /// Returns item count in the set @@ -1019,7 +1097,7 @@ namespace cds { namespace intrusive { void clear() { iterator it = begin(); - while ( it != end() ) { + while ( it != end()) { iterator i(it); ++i; unlink( *it ); @@ -1027,6 +1105,12 @@ namespace cds { namespace intrusive { } } + /// Returns internal statistics + stat const& statistics() const + { + return m_Stat; + } + protected: //@cond template @@ -1076,7 +1160,7 @@ namespace cds { namespace intrusive { */ iterator begin() { - return iterator( m_List.begin(), m_List.end() ); + return iterator( m_List.begin(), m_List.end()); } /// Returns an iterator that addresses the location succeeding the last element in a split-list @@ -1088,23 +1172,33 @@ namespace cds { namespace intrusive { */ iterator end() { - return iterator( m_List.end(), m_List.end() ); + return iterator( m_List.end(), m_List.end()); } /// Returns a forward const iterator addressing the first element in a split-list const_iterator begin() const { - return const_iterator( m_List.begin(), m_List.end() ); + return cbegin(); + } + /// Returns a forward const iterator addressing the first element in a split-list + const_iterator cbegin() const + { + return const_iterator( m_List.cbegin(), m_List.cend()); } /// Returns an const iterator that addresses the location succeeding the last element in a split-list const_iterator end() const { - return const_iterator( m_List.end(), m_List.end() ); + return cend(); + } + /// Returns an const iterator that addresses the location succeeding the last element in a split-list + const_iterator cend() const + { + return const_iterator( m_List.cend(), m_List.cend()); } }; }} // namespace cds::intrusive -#endif // #ifndef __CDS_INTRUSIVE_SPLIT_LIST_H +#endif // #ifndef CDSLIB_INTRUSIVE_SPLIT_LIST_H