Improved HP/DHP doc

[libcds.git] / cds / gc / hp_smr.h

/*
    This file is a part of libcds - Concurrent Data Structures library

    (C) Copyright Maxim Khizhinsky (libcds.dev@gmail.com) 2006-2017

    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_GC_HP_SMR_H
#define CDSLIB_GC_HP_SMR_H

#include <exception>
#include <cds/gc/details/hp_common.h>
#include <cds/details/throw_exception.h>
#include <cds/details/static_functor.h>
#include <cds/details/marked_ptr.h>
#include <cds/user_setup/cache_line.h>

/**
    @page cds_garbage_collectors_comparison Hazard Pointer SMR implementations
    @ingroup cds_garbage_collector

    <table>
        <tr>
            <th>Feature</th>
            <th>%cds::gc::HP</th>
            <th>%cds::gc::DHP</th>
        </tr>
        <tr>
            <td>Max number of guarded (hazard) pointers per thread</td>
            <td>limited (specified at construction time)</td>
            <td>unlimited (dynamically allocated when needed)</td>
        </tr>
        <tr>
            <td>Max number of retired pointers<sup>1</sup></td>
            <td>bounded, specified at construction time</td>
            <td>bounded, adaptive, depends on current thread count and number of hazard pointer for each thread</td>
        </tr>
        <tr>
            <td>Thread count</td>
            <td>bounded, upper bound is specified at construction time</td>
            <td>unbounded</td>
        </tr>
    </table>

    <sup>1</sup>Unbounded count of retired pointers means a possibility of memory exhaustion.
*/

namespace cds {
    /// @defgroup cds_garbage_collector Garbage collectors


    /// Different safe memory reclamation schemas (garbage collectors)
    /** @ingroup cds_garbage_collector

        This namespace specifies different safe memory reclamation (SMR) algorithms.
        See \ref cds_garbage_collector "Garbage collectors"
    */
    namespace gc {
    } // namespace gc

} // namespace cds


namespace cds { namespace gc {
    /// Hazard pointer implementation details
    namespace hp {
        using namespace cds::gc::hp::common;

        /// Exception "Not enough Hazard Pointer"
        class not_enought_hazard_ptr: public std::length_error
        {
        //@cond
        public:
            not_enought_hazard_ptr()
                : std::length_error( "Not enough Hazard Pointer" )
            {}
        //@endcond
        };

        /// Exception "Hazard Pointer SMR is not initialized"
        class not_initialized: public std::runtime_error
        {
        //@cond
        public:
            not_initialized()
                : std::runtime_error( "Global Hazard Pointer SMR object is not initialized" )
            {}
        //@endcond
        };

        //@cond
        /// Per-thread hazard pointer storage
        class thread_hp_storage {
        public:
            thread_hp_storage( guard* arr, size_t nSize ) CDS_NOEXCEPT
                : free_head_( arr )
                , array_( arr )
                , capacity_( nSize )
#       ifdef CDS_ENABLE_HPSTAT
                , alloc_guard_count_(0)
                , free_guard_count_(0)
#       endif
            {
                // Initialize guards
                new( arr ) guard[nSize];

                for ( guard* pEnd = arr + nSize - 1; arr < pEnd; ++arr )
                    arr->next_ = arr + 1;
                arr->next_ = nullptr;
            }

            thread_hp_storage() = delete;
            thread_hp_storage( thread_hp_storage const& ) = delete;
            thread_hp_storage( thread_hp_storage&& ) = delete;

            size_t capacity() const CDS_NOEXCEPT
            {
                return capacity_;
            }

            bool full() const CDS_NOEXCEPT
            {
                return free_head_ == nullptr;
            }

            guard* alloc()
            {
#       ifdef CDS_DISABLE_SMR_EXCEPTION
                assert( !full());
#       else
                if ( full() )
                    CDS_THROW_EXCEPTION( not_enought_hazard_ptr());
#       endif
                guard* g = free_head_;
                free_head_ = g->next_;
                CDS_HPSTAT( ++alloc_guard_count_ );
                return g;
            }

            void free( guard* g ) CDS_NOEXCEPT
            {
                assert( g >= array_ && g < array_ + capacity() );

                if ( g ) {
                    g->clear();
                    g->next_ = free_head_;
                    free_head_ = g;
                    CDS_HPSTAT( ++free_guard_count_ );
                }
            }

            template< size_t Capacity>
            size_t alloc( guard_array<Capacity>& arr )
            {
                size_t i;
                guard* g = free_head_;
                for ( i = 0; i < Capacity && g; ++i ) {
                    arr.reset( i, g );
                    g = g->next_;
                }

#       ifdef CDS_DISABLE_SMR_EXCEPTION
                assert( i == Capacity );
#       else
                if ( i != Capacity )
                    CDS_THROW_EXCEPTION( not_enought_hazard_ptr());
#       endif
                free_head_ = g;
                CDS_HPSTAT( alloc_guard_count_ += Capacity );
                return i;
            }

            template <size_t Capacity>
            void free( guard_array<Capacity>& arr ) CDS_NOEXCEPT
            {
                guard* gList = free_head_;
                for ( size_t i = 0; i < Capacity; ++i ) {
                    guard* g = arr[i];
                    if ( g ) {
                        g->clear();
                        g->next_ = gList;
                        gList = g;
                        CDS_HPSTAT( ++free_guard_count_ );
                    }
                }
                free_head_ = gList;
            }

            // cppcheck-suppress functionConst
            void clear()
            {
                for ( guard* cur = array_, *last = array_ + capacity(); cur < last; ++cur )
                    cur->clear();
            }

            guard& operator[]( size_t idx )
            {
                assert( idx < capacity() );

                return array_[idx];
            }

            static size_t calc_array_size( size_t capacity )
            {
                return sizeof( guard ) * capacity;
            }

        private:
            guard*          free_head_; ///< Head of free guard list
            guard* const    array_;     ///< HP array
            size_t const    capacity_;  ///< HP array capacity
#       ifdef CDS_ENABLE_HPSTAT
        public:
            size_t          alloc_guard_count_;
            size_t          free_guard_count_;
#       endif
        };
        //@endcond

        //@cond
        /// Per-thread retired array
        class retired_array
        {
        public:
            retired_array( retired_ptr* arr, size_t capacity ) CDS_NOEXCEPT
                : current_( arr )
                , last_( arr + capacity )
                , retired_( arr )
#       ifdef CDS_ENABLE_HPSTAT
                , retire_call_count_(0)
#       endif
            {}

            retired_array() = delete;
            retired_array( retired_array const& ) = delete;
            retired_array( retired_array&& ) = delete;

            size_t capacity() const CDS_NOEXCEPT
            {
                return last_ - retired_;
            }

            size_t size() const CDS_NOEXCEPT
            {
                return current_ - retired_;
            }

            bool push( retired_ptr&& p ) CDS_NOEXCEPT
            {
                *current_ = p;
                CDS_HPSTAT( ++retire_call_count_ );
                return ++current_ < last_;
            }

            retired_ptr* first() const CDS_NOEXCEPT
            {
                return retired_;
            }

            retired_ptr* last() const CDS_NOEXCEPT
            {
                return current_;
            }

            void reset( size_t nSize ) CDS_NOEXCEPT
            {
                current_ = first() + nSize;
            }

            bool full() const CDS_NOEXCEPT
            {
                return current_ == last_;
            }

            static size_t calc_array_size( size_t capacity )
            {
                return sizeof( retired_ptr ) * capacity;
            }

        private:
            retired_ptr*            current_;
            retired_ptr* const      last_;
            retired_ptr* const      retired_;
#       ifdef CDS_ENABLE_HPSTAT
        public:
            size_t  retire_call_count_;
#       endif
        };
        //@endcond

        /// Internal statistics
        struct stat {
            size_t  guard_allocated;    ///< Count of allocated HP guards
            size_t  guard_freed;        ///< Count of freed HP guards
            size_t  retired_count;      ///< Count of retired pointers
            size_t  free_count;         ///< Count of free pointers
            size_t  scan_count;         ///< Count of \p scan() call
            size_t  help_scan_count;    ///< Count of \p help_scan() call

            size_t  thread_rec_count;   ///< Count of thread records

            /// Default ctor
            stat()
            {
                clear();
            }

            /// Clears all counters
            void clear()
            {
                guard_allocated =
                    guard_freed =
                    retired_count =
                    free_count =
                    scan_count =
                    help_scan_count =
                    thread_rec_count = 0;
            }
        };

        //@cond
        /// Per-thread data
        struct thread_data {
            thread_hp_storage   hazards_;   ///< Hazard pointers private to the thread
            retired_array       retired_;   ///< Retired data private to the thread

            stat                stat_;      ///< Internal statistics for the thread

            char pad1_[cds::c_nCacheLineSize];
            atomics::atomic<unsigned int> sync_; ///< dummy var to introduce synchronizes-with relationship between threads
            char pad2_[cds::c_nCacheLineSize];

            // CppCheck warn: pad1_ and pad2_ is uninitialized in ctor
            // cppcheck-suppress uninitMemberVar
            thread_data( guard* guards, size_t guard_count, retired_ptr* retired_arr, size_t retired_capacity )
                : hazards_( guards, guard_count )
                , retired_( retired_arr, retired_capacity )
                , sync_(0)
            {}

            thread_data() = delete;
            thread_data( thread_data const& ) = delete;
            thread_data( thread_data&& ) = delete;

            void sync()
            {
                sync_.fetch_add( 1, atomics::memory_order_acq_rel );
            }
        };
        //@endcond

        /// \p smr::scan() strategy
        enum scan_type {
            classic,    ///< classic scan as described in Michael's works (see smr::classic_scan() )
            inplace     ///< inplace scan without allocation (see smr::inplace_scan() )
        };

        //@cond
        /// Hazard Pointer SMR (Safe Memory Reclamation)
        class smr
        {
            struct thread_record;

        public:
            /// Returns the instance of Hazard Pointer \ref smr
            static smr& instance()
            {
#       ifdef CDS_DISABLE_SMR_EXCEPTION
                assert( instance_ != nullptr );
#       else
                if ( !instance_ )
                    CDS_THROW_EXCEPTION( not_initialized());
#       endif
                return *instance_;
            }

            /// Creates Hazard Pointer SMR singleton
            /**
                Hazard Pointer SMR is a singleton. If HP instance is not initialized then the function creates the instance.
                Otherwise it does nothing.

                The Michael's HP reclamation schema depends of three parameters:
                - \p nHazardPtrCount - HP pointer count per thread. Usually it is small number (2-4) depending from
                    the data structure algorithms. By default, if \p nHazardPtrCount = 0,
                    the function uses maximum of HP count for CDS library
                - \p nMaxThreadCount - max count of thread with using HP GC in your application. Default is 100.
                - \p nMaxRetiredPtrCount - capacity of array of retired pointers for each thread. Must be greater than
                    <tt> nHazardPtrCount * nMaxThreadCount </tt>
                    Default is <tt>2 * nHazardPtrCount * nMaxThreadCount</tt>
            */
            static CDS_EXPORT_API void construct(
                size_t nHazardPtrCount = 0,     ///< Hazard pointer count per thread
                size_t nMaxThreadCount = 0,     ///< Max count of simultaneous working thread in your application
                size_t nMaxRetiredPtrCount = 0, ///< Capacity of the array of retired objects for the thread
                scan_type nScanType = inplace   ///< Scan type (see \ref scan_type enum)
            );

            // for back-copatibility
            static void Construct(
                size_t nHazardPtrCount = 0,     ///< Hazard pointer count per thread
                size_t nMaxThreadCount = 0,     ///< Max count of simultaneous working thread in your application
                size_t nMaxRetiredPtrCount = 0, ///< Capacity of the array of retired objects for the thread
                scan_type nScanType = inplace   ///< Scan type (see \ref scan_type enum)
            )
            {
                construct( nHazardPtrCount, nMaxThreadCount, nMaxRetiredPtrCount, nScanType );
            }

            /// Destroys global instance of \ref smr
            /**
                The parameter \p bDetachAll should be used carefully: if its value is \p true,
                then the object destroyed automatically detaches all attached threads. This feature
                can be useful when you have no control over the thread termination, for example,
                when \p libcds is injected into existing external thread.
            */
            static CDS_EXPORT_API void destruct(
                bool bDetachAll = false     ///< Detach all threads
            );

            // for back-compatibility
            static void Destruct(
                bool bDetachAll = false     ///< Detach all threads
            )
            {
                destruct( bDetachAll );
            }

            /// Checks if global SMR object is constructed and may be used
            static bool isUsed() CDS_NOEXCEPT
            {
                return instance_ != nullptr;
            }

            /// Set memory management functions
            /**
                @note This function may be called <b>BEFORE</b> creating an instance
                of Hazard Pointer SMR

                SMR object allocates some memory for thread-specific data and for
                creating SMR object.
                By default, a standard \p new and \p delete operators are used for this.
            */
            static CDS_EXPORT_API void set_memory_allocator(
                void* ( *alloc_func )( size_t size ),
                void (*free_func )( void * p )
            );

            /// Returns max Hazard Pointer count per thread
            size_t get_hazard_ptr_count() const CDS_NOEXCEPT
            {
                return hazard_ptr_count_;
            }

            /// Returns max thread count
            size_t get_max_thread_count() const CDS_NOEXCEPT
            {
                return max_thread_count_;
            }

            /// Returns max size of retired objects array
            size_t get_max_retired_ptr_count() const CDS_NOEXCEPT
            {
                return max_retired_ptr_count_;
            }

            /// Get current scan strategy
            scan_type get_scan_type() const
            {
                return scan_type_;
            }

            /// Checks that required hazard pointer count \p nRequiredCount is less or equal then max hazard pointer count
            /**
                If <tt> nRequiredCount > get_hazard_ptr_count()</tt> then the exception \p not_enought_hazard_ptr is thrown
            */
            static void check_hazard_ptr_count( size_t nRequiredCount )
            {
                if ( instance().get_hazard_ptr_count() < nRequiredCount ) {
#       ifdef CDS_DISABLE_SMR_EXCEPTION
                    assert( false );    // not enough hazard ptr
#       else
                    CDS_THROW_EXCEPTION( not_enought_hazard_ptr() );
#       endif
                }
            }

            /// Returns thread-local data for the current thread
            static CDS_EXPORT_API thread_data* tls();

            static CDS_EXPORT_API void attach_thread();
            static CDS_EXPORT_API void detach_thread();

            /// Get internal statistics
            void statistics( stat& st );

        public: // for internal use only
            /// The main garbage collecting function
            /**
                This function is called internally when upper bound of thread's list of reclaimed pointers
                is reached.

                There are the following scan algorithm:
                - \ref hzp_gc_classic_scan "classic_scan" allocates memory for internal use
                - \ref hzp_gc_inplace_scan "inplace_scan" does not allocate any memory

                Use \p set_scan_type() member function to setup appropriate scan algorithm.
            */
            void scan( thread_data* pRec )
            {
                ( this->*scan_func_ )( pRec );
            }

            /// Helper scan routine
            /**
                The function guarantees that every node that is eligible for reuse is eventually freed, barring
                thread failures. To do so, after executing \p scan(), a thread executes a \p %help_scan(),
                where it checks every HP record. If an HP record is inactive, the thread moves all "lost" reclaimed pointers
                to thread's list of reclaimed pointers.

                The function is called internally by \p scan().
            */
            CDS_EXPORT_API void help_scan( thread_data* pThis );

        private:
            CDS_EXPORT_API smr(
                size_t nHazardPtrCount,     ///< Hazard pointer count per thread
                size_t nMaxThreadCount,     ///< Max count of simultaneous working thread in your application
                size_t nMaxRetiredPtrCount, ///< Capacity of the array of retired objects for the thread
                scan_type nScanType         ///< Scan type (see \ref scan_type enum)
            );

            CDS_EXPORT_API ~smr();

            CDS_EXPORT_API void detach_all_thread();

            /// Classic scan algorithm
            /** @anchor hzp_gc_classic_scan
                Classical scan algorithm as described in Michael's paper.

                A scan includes four stages. The first stage involves scanning the array HP for non-null values.
                Whenever a non-null value is encountered, it is inserted in a local list of currently protected pointer.
                Only stage 1 accesses shared variables. The following stages operate only on private variables.

                The second stage of a scan involves sorting local list of protected pointers to allow
                binary search in the third stage.

                The third stage of a scan involves checking each reclaimed node
                against the pointers in local list of protected pointers. If the binary search yields
                no match, the node is freed. Otherwise, it cannot be deleted now and must kept in thread's list
                of reclaimed pointers.

                The forth stage prepares new thread's private list of reclaimed pointers
                that could not be freed during the current scan, where they remain until the next scan.

                This algorithm allocates memory for internal HP array.

                This function is called internally by ThreadGC object when upper bound of thread's list of reclaimed pointers
                is reached.
            */
            CDS_EXPORT_API void classic_scan( thread_data* pRec );

            /// In-place scan algorithm
            /** @anchor hzp_gc_inplace_scan
                Unlike the \p classic_scan() algorithm, \p %inplace_scan() does not allocate any memory.
                All operations are performed in-place.
            */
            CDS_EXPORT_API void inplace_scan( thread_data* pRec );

        private:
            CDS_EXPORT_API thread_record* create_thread_data();
            static CDS_EXPORT_API void destroy_thread_data( thread_record* pRec );

            /// Allocates Hazard Pointer SMR thread private data
            CDS_EXPORT_API thread_record* alloc_thread_data();

            /// Free HP SMR thread-private data
            CDS_EXPORT_API void free_thread_data( thread_record* pRec );

        private:
            static CDS_EXPORT_API smr* instance_;

            atomics::atomic< thread_record*>    thread_list_;   ///< Head of thread list

            size_t const    hazard_ptr_count_;      ///< max count of thread's hazard pointer
            size_t const    max_thread_count_;      ///< max count of thread
            size_t const    max_retired_ptr_count_; ///< max count of retired ptr per thread
            scan_type const scan_type_;             ///< scan type (see \ref scan_type enum)
            void ( smr::*scan_func_ )( thread_data* pRec );
        };
        //@endcond

        //@cond
        // for backward compatibility
        typedef smr GarbageCollector;
        //@endcond

    } // namespace hp

    /// Hazard Pointer SMR (Safe Memory Reclamation)
    /**  @ingroup cds_garbage_collector

        Implementation of classic Hazard Pointer SMR

        Sources:
            - [2002] Maged M.Michael "Safe memory reclamation for dynamic lock-freeobjects using atomic reads and writes"
            - [2003] Maged M.Michael "Hazard Pointers: Safe memory reclamation for lock-free objects"
            - [2004] Andrei Alexandrescy, Maged Michael "Lock-free Data Structures with Hazard Pointers"

        Hazard Pointer SMR is a singleton. The main user-level part of Hazard Pointer schema is
        \p %cds::gc::HP class and its nested classes. Before use any HP-related class you must initialize \p %HP
        by contructing \p %cds::gc::HP object in beginning of your \p main().
        See \ref cds_how_to_use "How to use" section for details how to apply SMR schema.
    */
    class HP
    {
    public:
        /// Native guarded pointer type
        typedef hp::hazard_ptr guarded_pointer;

        /// Atomic reference
        template <typename T> using atomic_ref = atomics::atomic<T *>;

        /// Atomic marked pointer
        template <typename MarkedPtr> using atomic_marked_ptr = atomics::atomic<MarkedPtr>;

        /// Atomic type
        template <typename T> using atomic_type = atomics::atomic<T>;

        /// Exception "Not enough Hazard Pointer"
        typedef hp::not_enought_hazard_ptr not_enought_hazard_ptr_exception;

        /// Internal statistics
        typedef hp::stat stat;

        /// Hazard Pointer guard
        /**
            A guard is a hazard pointer.
            Additionally, the \p %Guard class manages allocation and deallocation of the hazard pointer.

            \p %Guard object is movable but not copyable.

            The guard object can be in two states:
            - unlinked - the guard is not linked with any internal hazard pointer.
              In this state no operation except \p link() and move assignment is supported.
            - linked (default) - the guard allocates an internal hazard pointer and completely operable.

            Due to performance reason the implementation does not check state of the guard in runtime.

            @warning Move assignment transfers the guard in unlinked state, use with care.
        */
        class Guard
        {
        public:
            /// Default ctor allocates a guard (hazard pointer) from thread-private storage
            /**
                @warning Can throw \p too_many_hazard_ptr_exception if internal hazard pointer objects are exhausted.
            */
            Guard()
                : guard_( hp::smr::tls()->hazards_.alloc() )
            {}

            /// Initilalizes an unlinked guard i.e. the guard contains no hazard pointer. Used for move semantics support
            explicit Guard( std::nullptr_t ) CDS_NOEXCEPT
                : guard_( nullptr )
            {}

            /// Move ctor - \p src guard becomes unlinked (transfer internal guard ownership)
            Guard( Guard&& src ) CDS_NOEXCEPT
                : guard_( src.guard_ )
            {
                src.guard_ = nullptr;
            }

            /// Move assignment: the internal guards are swapped between \p src and \p this
            /**
                @warning \p src will become in unlinked state if \p this was unlinked on entry.
            */
            Guard& operator=( Guard&& src ) CDS_NOEXCEPT
            {
                std::swap( guard_, src.guard_ );
                return *this;
            }

            /// Copy ctor is prohibited - the guard is not copyable
            Guard( Guard const& ) = delete;

            /// Copy assignment is prohibited
            Guard& operator=( Guard const& ) = delete;

            /// Frees the internal hazard pointer if the guard is in linked state
            ~Guard()
            {
                unlink();
            }

            /// Checks if the guard object linked with any internal hazard pointer
            bool is_linked() const
            {
                return guard_ != nullptr;
            }

            /// Links the guard with internal hazard pointer if the guard is in unlinked state
            /**
                @warning Can throw \p not_enought_hazard_ptr_exception if internal hazard pointer array is exhausted.
            */
            void link()
            {
                if ( !guard_ )
                    guard_ = hp::smr::tls()->hazards_.alloc();
            }

            /// Unlinks the guard from internal hazard pointer; the guard becomes in unlinked state
            void unlink()
            {
                if ( guard_ ) {
                    hp::smr::tls()->hazards_.free( guard_ );
                    guard_ = nullptr;
                }
            }

            /// Protects a pointer of type \p atomic<T*>
            /**
                Return the value of \p toGuard

                The function tries to load \p toGuard and to store it
                to the HP slot repeatedly until the guard's value equals \p toGuard

                @warning The guad object should be in linked state, otherwise the result is undefined
            */
            template <typename T>
            T protect( atomics::atomic<T> const& toGuard )
            {
                assert( guard_ != nullptr );

                T pCur = toGuard.load(atomics::memory_order_acquire);
                T pRet;
                do {
                    pRet = assign( pCur );
                    pCur = toGuard.load(atomics::memory_order_acquire);
                } while ( pRet != pCur );
                return pCur;
            }

            /// Protects a converted pointer of type \p atomic<T*>
            /**
                Return the value of \p toGuard

                The function tries to load \p toGuard and to store result of \p f functor
                to the HP slot repeatedly until the guard's value equals \p toGuard.

                The function is useful for intrusive containers when \p toGuard is a node pointer
                that should be converted to a pointer to the value before protecting.
                The parameter \p f of type Func is a functor that makes this conversion:
                \code
                    struct functor {
                        value_type * operator()( T * p );
                    };
                \endcode
                Actually, the result of <tt> f( toGuard.load()) </tt> is assigned to the hazard pointer.

                @warning The guad object should be in linked state, otherwise the result is undefined
            */
            template <typename T, class Func>
            T protect( atomics::atomic<T> const& toGuard, Func f )
            {
                assert( guard_ != nullptr );

                T pCur = toGuard.load(atomics::memory_order_acquire);
                T pRet;
                do {
                    pRet = pCur;
                    assign( f( pCur ));
                    pCur = toGuard.load(atomics::memory_order_acquire);
                } while ( pRet != pCur );
                return pCur;
            }

            /// Store \p p to the guard
            /**
                The function equals to a simple assignment the value \p p to guard, no loop is performed.
                Can be used for a pointer that cannot be changed concurrently or if the pointer is already
                guarded by another guard.

                @warning The guad object should be in linked state, otherwise the result is undefined
            */
            template <typename T>
            T * assign( T* p )
            {
                assert( guard_ != nullptr );

                guard_->set( p );
                hp::smr::tls()->sync();
                return p;
            }

            //@cond
            std::nullptr_t assign( std::nullptr_t )
            {
                assert( guard_ != nullptr );

                guard_->clear();
                return nullptr;
            }
            //@endcond

            /// Copy a value guarded from \p src guard to \p this guard (valid only in linked state)
            void copy( Guard const& src )
            {
                assign( src.get_native());
            }

            /// Store marked pointer \p p to the guard
            /**
                The function equals to a simple assignment of <tt>p.ptr()</tt>, no loop is performed.
                Can be used for a marked pointer that cannot be changed concurrently or if the marked pointer
                is already guarded by another guard.

                @warning The guard object should be in linked state, otherwise the result is undefined
            */
            template <typename T, int BITMASK>
            T * assign( cds::details::marked_ptr<T, BITMASK> p )
            {
                return assign( p.ptr());
            }

            /// Clear value of the guard (valid only in linked state)
            void clear()
            {
                assign( nullptr );
            }

            /// Get the value currently protected (valid only in linked state)
            template <typename T>
            T * get() const
            {
                assert( guard_ != nullptr );
                return guard_->get_as<T>();
            }

            /// Get native hazard pointer stored (valid only in linked state)
            guarded_pointer get_native() const
            {
                assert( guard_ != nullptr );
                return guard_->get();
            }

            //@cond
            hp::guard* release()
            {
                hp::guard* g = guard_;
                guard_ = nullptr;
                return g;
            }

            hp::guard*& guard_ref()
            {
                return guard_;
            }
            //@endcond

        private:
            //@cond
            hp::guard* guard_;
            //@endcond
        };

        /// Array of Hazard Pointer guards
        /**
            The class is intended for allocating an array of hazard pointer guards.
            Template parameter \p Count defines the size of the array.
        */
        template <size_t Count>
        class GuardArray
        {
        public:
            /// Rebind array for other size \p Count2
            template <size_t Count2>
            struct rebind {
                typedef GuardArray<Count2>  other;   ///< rebinding result
            };

            /// Array capacity
            static CDS_CONSTEXPR const size_t c_nCapacity = Count;

        public:
            /// Default ctor allocates \p Count hazard pointers
            GuardArray()
            {
                hp::smr::tls()->hazards_.alloc( guards_ );
            }

            /// Move ctor is prohibited
            GuardArray( GuardArray&& ) = delete;

            /// Move assignment is prohibited
            GuardArray& operator=( GuardArray&& ) = delete;

            /// Copy ctor is prohibited
            GuardArray( GuardArray const& ) = delete;

            /// Copy assignment is prohibited
            GuardArray& operator=( GuardArray const& ) = delete;

            /// Frees allocated hazard pointers
            ~GuardArray()
            {
                hp::smr::tls()->hazards_.free( guards_ );
            }

            /// Protects a pointer of type \p atomic<T*>
            /**
                Return the value of \p toGuard

                The function tries to load \p toGuard and to store it
                to the slot \p nIndex repeatedly until the guard's value equals \p toGuard
            */
            template <typename T>
            T protect( size_t nIndex, atomics::atomic<T> const& toGuard )
            {
                assert( nIndex < capacity());

                T pRet;
                do {
                    pRet = assign( nIndex, toGuard.load(atomics::memory_order_acquire));
                } while ( pRet != toGuard.load(atomics::memory_order_acquire));

                return pRet;
            }

            /// Protects a pointer of type \p atomic<T*>
            /**
                Return the value of \p toGuard

                The function tries to load \p toGuard and to store it
                to the slot \p nIndex repeatedly until the guard's value equals \p toGuard

                The function is useful for intrusive containers when \p toGuard is a node pointer
                that should be converted to a pointer to the value type before guarding.
                The parameter \p f of type Func is a functor that makes this conversion:
                \code
                    struct functor {
                        value_type * operator()( T * p );
                    };
                \endcode
                Really, the result of <tt> f( toGuard.load()) </tt> is assigned to the hazard pointer.
            */
            template <typename T, class Func>
            T protect( size_t nIndex, atomics::atomic<T> const& toGuard, Func f )
            {
                assert( nIndex < capacity());

                T pRet;
                do {
                    assign( nIndex, f( pRet = toGuard.load(atomics::memory_order_acquire)));
                } while ( pRet != toGuard.load(atomics::memory_order_acquire));

                return pRet;
            }

            /// Store \p to the slot \p nIndex
            /**
                The function equals to a simple assignment, no loop is performed.
            */
            template <typename T>
            T * assign( size_t nIndex, T * p )
            {
                assert( nIndex < capacity() );

                guards_.set( nIndex, p );
                hp::smr::tls()->sync();
                return p;
            }

            /// Store marked pointer \p p to the guard
            /**
                The function equals to a simple assignment of <tt>p.ptr()</tt>, no loop is performed.
                Can be used for a marked pointer that cannot be changed concurrently.
            */
            template <typename T, int BITMASK>
            T * assign( size_t nIndex, cds::details::marked_ptr<T, BITMASK> p )
            {
                return assign( nIndex, p.ptr());
            }

            /// Copy guarded value from \p src guard to slot at index \p nIndex
            void copy( size_t nIndex, Guard const& src )
            {
                assign( nIndex, src.get_native());
            }

            /// Copy guarded value from slot \p nSrcIndex to the slot \p nDestIndex
            void copy( size_t nDestIndex, size_t nSrcIndex )
            {
                assign( nDestIndex, get_native( nSrcIndex ));
            }

            /// Clear value of the slot \p nIndex
            void clear( size_t nIndex )
            {
                guards_.clear( nIndex );
            }

            /// Get current value of slot \p nIndex
            template <typename T>
            T * get( size_t nIndex ) const
            {
                assert( nIndex < capacity() );
                return guards_[nIndex]->template get_as<T>();
            }

            /// Get native hazard pointer stored
            guarded_pointer get_native( size_t nIndex ) const
            {
                assert( nIndex < capacity());
                return guards_[nIndex]->get();
            }

            //@cond
            hp::guard* release( size_t nIndex ) CDS_NOEXCEPT
            {
                return guards_.release( nIndex );
            }
            //@endcond

            /// Capacity of the guard array
            static CDS_CONSTEXPR size_t capacity()
            {
                return c_nCapacity;
            }

        private:
            //@cond
            hp::guard_array<c_nCapacity> guards_;
            //@endcond
        };

        /// Guarded pointer
        /**
            A guarded pointer is a pair of a pointer and GC's guard.
            Usually, it is used for returning a pointer to an element of a lock-free container.
            The guard prevents the pointer to be early disposed (freed) by SMR.
            After destructing \p %guarded_ptr object the pointer can be disposed (freed) automatically at any time.

            Template arguments:
            - \p GuardedType - a type which the guard stores
            - \p ValueType - a value type
            - \p Cast - a functor for converting <tt>GuardedType*</tt> to <tt>ValueType*</tt>. Default is \p void (no casting).

            For intrusive containers, \p GuardedType is the same as \p ValueType and no casting is needed.
            In such case the \p %guarded_ptr is:
            @code
            typedef cds::gc::HP::guarded_ptr< foo > intrusive_guarded_ptr;
            @endcode

            For standard (non-intrusive) containers \p GuardedType is not the same as \p ValueType and casting is needed.
            For example:
            @code
            struct foo {
                int const   key;
                std::string value;
            };

            struct value_accessor {
                std::string* operator()( foo* pFoo ) const
                {
                    return &(pFoo->value);
                }
            };

            // Guarded ptr
            typedef cds::gc::HP::guarded_ptr< Foo, std::string, value_accessor > nonintrusive_guarded_ptr;
            @endcode

            You don't need use this class directly.
            All set/map container classes from \p libcds declare the typedef for \p %guarded_ptr with appropriate casting functor.
        */
        template <typename GuardedType, typename ValueType=GuardedType, typename Cast=void >
        class guarded_ptr
        {
            //@cond
            struct trivial_cast {
                ValueType * operator()( GuardedType * p ) const
                {
                    return p;
                }
            };

            template <typename GT, typename VT, typename C> friend class guarded_ptr;
            //@endcond

        public:
            typedef GuardedType guarded_type; ///< Guarded type
            typedef ValueType   value_type;   ///< Value type

            /// Functor for casting \p guarded_type to \p value_type
            typedef typename std::conditional< std::is_same<Cast, void>::value, trivial_cast, Cast >::type value_cast;

        public:
            /// Creates empty guarded pointer
            guarded_ptr() CDS_NOEXCEPT
                : guard_(nullptr)
            {}

            //@cond
            explicit guarded_ptr( hp::guard* g ) CDS_NOEXCEPT
                : guard_( g )
            {}

            /// Initializes guarded pointer with \p p
            explicit guarded_ptr( guarded_type* p ) CDS_NOEXCEPT
                : guard_( nullptr )
            {
                reset(p);
            }
            explicit guarded_ptr( std::nullptr_t ) CDS_NOEXCEPT
                : guard_( nullptr )
            {}
            //@endcond

            /// Move ctor
            guarded_ptr( guarded_ptr&& gp ) CDS_NOEXCEPT
                : guard_( gp.guard_ )
            {
                gp.guard_ = nullptr;
            }

            /// Move ctor
            template <typename GT, typename VT, typename C>
            guarded_ptr( guarded_ptr<GT, VT, C>&& gp ) CDS_NOEXCEPT
                : guard_( gp.guard_ )
            {
                gp.guard_ = nullptr;
            }

            /// Ctor from \p Guard
            explicit guarded_ptr( Guard&& g ) CDS_NOEXCEPT
                : guard_( g.release())
            {}

            /// The guarded pointer is not copy-constructible
            guarded_ptr( guarded_ptr const& gp ) = delete;

            /// Clears the guarded pointer
            /**
                \ref release() is called if guarded pointer is not \ref empty()
            */
            ~guarded_ptr() CDS_NOEXCEPT
            {
                release();
            }

            /// Move-assignment operator
            guarded_ptr& operator=( guarded_ptr&& gp ) CDS_NOEXCEPT
            {
                std::swap( guard_, gp.guard_ );
                return *this;
            }

            /// Move-assignment from \p Guard
            guarded_ptr& operator=( Guard&& g ) CDS_NOEXCEPT
            {
                std::swap( guard_, g.guard_ref());
                return *this;
            }

            /// The guarded pointer is not copy-assignable
            guarded_ptr& operator=(guarded_ptr const& gp) = delete;

            /// Returns a pointer to guarded value
            value_type * operator ->() const CDS_NOEXCEPT
            {
                assert( !empty());
                return value_cast()( guard_->get_as<guarded_type>());
            }

            /// Returns a reference to guarded value
            value_type& operator *() CDS_NOEXCEPT
            {
                assert( !empty());
                return *value_cast()( guard_->get_as<guarded_type>());
            }

            /// Returns const reference to guarded value
            value_type const& operator *() const CDS_NOEXCEPT
            {
                assert( !empty());
                return *value_cast()( guard_->get_as<guarded_type>());
            }

            /// Checks if the guarded pointer is \p nullptr
            bool empty() const CDS_NOEXCEPT
            {
                return !guard_ || guard_->get( atomics::memory_order_relaxed ) == nullptr;
            }

            /// \p bool operator returns <tt>!empty()</tt>
            explicit operator bool() const CDS_NOEXCEPT
            {
                return !empty();
            }

            /// Clears guarded pointer
            /**
                If the guarded pointer has been released, the pointer can be disposed (freed) at any time.
                Dereferncing the guarded pointer after \p release() is dangerous.
            */
            void release() CDS_NOEXCEPT
            {
                free_guard();
            }

            //@cond
            // For internal use only!!!
            void reset(guarded_type * p) CDS_NOEXCEPT
            {
                alloc_guard();
                assert( guard_ );
                guard_->set(p);
            }
            //@endcond

        private:
            //@cond
            void alloc_guard()
            {
                if ( !guard_ )
                    guard_ = hp::smr::tls()->hazards_.alloc();
            }

            void free_guard()
            {
                if ( guard_ ) {
                    hp::smr::tls()->hazards_.free( guard_ );
                    guard_ = nullptr;
                }
            }
            //@endcond

        private:
            //@cond
            hp::guard* guard_;
            //@endcond
        };

    public:
        /// \p scan() type
        enum class scan_type {
            classic = hp::classic,    ///< classic scan as described in Michael's papers
            inplace = hp::inplace     ///< inplace scan without allocation
        };

        /// Initializes %HP singleton
        /**
            The constructor initializes Hazard Pointer SMR singleton with passed parameters.
            If the instance does not yet exist then the function creates the instance.
            Otherwise it does nothing.

            The Michael's %HP reclamation schema depends of three parameters:
            - \p nHazardPtrCount - hazard pointer count per thread. Usually it is small number (up to 10) depending from
                the data structure algorithms. If \p nHazardPtrCount = 0, the defaul value 8 is used
            - \p nMaxThreadCount - max count of thread with using Hazard Pointer GC in your application. Default is 100.
            - \p nMaxRetiredPtrCount - capacity of array of retired pointers for each thread. Must be greater than
                <tt> nHazardPtrCount * nMaxThreadCount </tt>. Default is <tt>2 * nHazardPtrCount * nMaxThreadCount </tt>.
        */
        HP(
            size_t nHazardPtrCount = 0,     ///< Hazard pointer count per thread
            size_t nMaxThreadCount = 0,     ///< Max count of simultaneous working thread in your application
            size_t nMaxRetiredPtrCount = 0, ///< Capacity of the array of retired objects for the thread
            scan_type nScanType = scan_type::inplace   ///< Scan type (see \p scan_type enum)
        )
        {
            hp::smr::construct(
                nHazardPtrCount,
                nMaxThreadCount,
                nMaxRetiredPtrCount,
                static_cast<hp::scan_type>(nScanType)
            );
        }

        /// Terminates GC singleton
        /**
            The destructor destroys %HP global object. After calling of this function you may \b NOT
            use CDS data structures based on \p %cds::gc::HP.
            Usually, %HP object is destroyed at the end of your \p main().
        */
        ~HP()
        {
            hp::smr::destruct( true );
        }

        /// Checks that required hazard pointer count \p nCountNeeded is less or equal then max hazard pointer count
        /**
            If <tt> nRequiredCount > get_hazard_ptr_count()</tt> then the exception \p not_enought_hazard_ptr is thrown
        */
        static void check_available_guards( size_t nCountNeeded )
        {
            hp::smr::check_hazard_ptr_count( nCountNeeded );
        }

        /// Set memory management functions
        /**
            @note This function may be called <b>BEFORE</b> creating an instance
            of Hazard Pointer SMR

            SMR object allocates some memory for thread-specific data and for
            creating SMR object.
            By default, a standard \p new and \p delete operators are used for this.
        */
        static void set_memory_allocator(
            void* ( *alloc_func )( size_t size ),   ///< \p malloc() function
            void( *free_func )( void * p )          ///< \p free() function
        )
        {
            hp::smr::set_memory_allocator( alloc_func, free_func );
        }

        /// Returns max Hazard Pointer count
        static size_t max_hazard_count()
        {
            return hp::smr::instance().get_hazard_ptr_count();
        }

        /// Returns max count of thread
        static size_t max_thread_count()
        {
            return hp::smr::instance().get_max_thread_count();
        }

        /// Returns capacity of retired pointer array
        static size_t retired_array_capacity()
        {
            return hp::smr::instance().get_max_retired_ptr_count();
        }

        /// Retire pointer \p p with function \p func
        /**
            The function places pointer \p p to array of pointers ready for removing.
            (so called retired pointer array). The pointer can be safely removed when no hazard pointer points to it.
            \p func is a disposer: when \p p can be safely removed, \p func is called.
        */
        template <typename T>
        static void retire( T * p, void( *func )( T * ))
        {
            hp::thread_data* rec = hp::smr::tls();
            if ( !rec->retired_.push( hp::retired_ptr( p, func )))
                hp::smr::instance().scan( rec );
        }

        /// Retire pointer \p p with functor of type \p Disposer
        /**
            The function places pointer \p p to array of pointers ready for removing.
            (so called retired pointer array). The pointer can be safely removed when no hazard pointer points to it.

            Deleting the pointer is an invocation of some object of type \p Disposer; the interface of \p Disposer is:
            \code
            template <typename T>
            struct disposer {
                void operator()( T * p )    ;   // disposing operator
            };
            \endcode
            Since the functor call can happen at any time after \p retire() call, additional restrictions are imposed to \p Disposer type:
            - it should be stateless functor
            - it should be default-constructible
            - the result of functor call with argument \p p should not depend on where the functor will be called.

            \par Examples:
            Operator \p delete functor:
            \code
            template <typename T>
            struct disposer {
                void operator ()( T * p ) {
                    delete p;
                }
            };

            // How to call HP::retire method
            int * p = new int;

            // ... use p in lock-free manner

            cds::gc::HP::retire<disposer>( p ) ;   // place p to retired pointer array of HP GC
            \endcode

            Functor based on \p std::allocator :
            \code
            template <typename Alloc = std::allocator<int> >
            struct disposer {
                template <typename T>
                void operator()( T * p ) {
                    typedef typename Alloc::templare rebind<T>::other alloc_t;
                    alloc_t a;
                    a.destroy( p );
                    a.deallocate( p, 1 );
                }
            };
            \endcode
        */
        template <class Disposer, typename T>
        static void retire( T * p )
        {
            if ( !hp::smr::tls()->retired_.push( hp::retired_ptr( p, cds::details::static_functor<Disposer, T>::call )))
                scan();
        }

        /// Get current scan strategy
        static scan_type getScanType()
        {
            return static_cast<scan_type>( hp::smr::instance().get_scan_type());
        }

        /// Checks if Hazard Pointer GC is constructed and may be used
        static bool isUsed()
        {
            return hp::smr::isUsed();
        }

        /// Forces SMR call for current thread
        /**
            Usually, this function should not be called directly.
        */
        static void scan()
        {
            hp::smr::instance().scan( hp::smr::tls());
        }

        /// Synonym for \p scan()
        static void force_dispose()
        {
            scan();
        }

        /// Returns internal statistics
        /**
            The function clears \p st before gathering statistics.

            @note Internal statistics is available only if you compile
            \p libcds and your program with \p -DCDS_ENABLE_HPSTAT key.
        */
        static void statistics( stat& st )
        {
            hp::smr::instance().statistics( st );
        }

        /// Returns post-mortem statistics
        /**
            Post-mortem statistics is gathered in the \p %HP object destructor
            and can be accessible after destructing the global \p %HP object.

            @note Internal statistics is available only if you compile
            \p libcds and your program with \p -DCDS_ENABLE_HPSTAT key.

            Usage:
            \code
            int main()
            {
                cds::Initialize();
                {
                    // Initialize HP SMR
                    cds::gc::HP hp;

                    // deal with HP-based data structured
                    // ...
                }

                // HP object destroyed
                // Get total post-mortem statistics
                cds::gc::HP::stat const& st = cds::gc::HP::postmortem_statistics();

                printf( "HP statistics:\n"
                    "\tthread count           = %llu\n"
                    "\tguard allocated        = %llu\n"
                    "\tguard freed            = %llu\n"
                    "\tretired data count     = %llu\n"
                    "\tfree data count        = %llu\n"
                    "\tscan() call count      = %llu\n"
                    "\thelp_scan() call count = %llu\n",
                    st.thread_rec_count,
                    st.guard_allocated, st.guard_freed,
                    st.retired_count, st.free_count,
                    st.scan_count, st.help_scan_count
                );

                cds::Terminate();
            }
            \endcode
        */
        static stat const& postmortem_statistics();
    };

}} // namespace cds::gc

#endif // #ifndef CDSLIB_GC_HP_SMR_H