peerplays_migrated/libraries/chain/db_witness_schedule.cpp
2019-11-08 15:42:09 +03:00

261 lines
9.7 KiB
C++

/*
* Copyright (c) 2015 Cryptonomex, Inc., and contributors.
*
* The MIT License
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#include <graphene/chain/database.hpp>
#include <graphene/chain/global_property_object.hpp>
#include <graphene/chain/witness_object.hpp>
#include <graphene/chain/witness_schedule_object.hpp>
namespace graphene { namespace chain {
using boost::container::flat_set;
witness_id_type database::get_scheduled_witness( uint32_t slot_num )const
{
witness_id_type wid;
const global_property_object& gpo = get_global_properties();
if (gpo.parameters.witness_schedule_algorithm == GRAPHENE_WITNESS_SHUFFLED_ALGORITHM)
{
const dynamic_global_property_object& dpo = get_dynamic_global_properties();
const witness_schedule_object& wso = get_witness_schedule_object();;
uint64_t current_aslot = dpo.current_aslot + slot_num;
return wso.current_shuffled_witnesses[ current_aslot % wso.current_shuffled_witnesses.size() ];
}
if (gpo.parameters.witness_schedule_algorithm == GRAPHENE_WITNESS_SCHEDULED_ALGORITHM &&
slot_num != 0 )
{
const witness_schedule_object& wso = get_witness_schedule_object();;
// ask the near scheduler who goes in the given slot
bool slot_is_near = wso.scheduler.get_slot(slot_num-1, wid);
if(! slot_is_near)
{
// if the near scheduler doesn't know, we have to extend it to
// a far scheduler.
// n.b. instantiating it is slow, but block gaps long enough to
// need it are likely pretty rare.
witness_scheduler_rng far_rng(wso.rng_seed.begin(), GRAPHENE_FAR_SCHEDULE_CTR_IV);
far_future_witness_scheduler far_scheduler =
far_future_witness_scheduler(wso.scheduler, far_rng);
if(!far_scheduler.get_slot(slot_num-1, wid))
{
// no scheduled witness -- somebody set up us the bomb
// n.b. this code path is impossible, the present
// implementation of far_future_witness_scheduler
// returns true unconditionally
assert( false );
}
}
}
return wid;
}
fc::time_point_sec database::get_slot_time(uint32_t slot_num)const
{
if( slot_num == 0 )
return fc::time_point_sec();
auto interval = block_interval();
const dynamic_global_property_object& dpo = get_dynamic_global_properties();
if( head_block_num() == 0 )
{
// n.b. first block is at genesis_time plus one block interval
fc::time_point_sec genesis_time = dpo.time;
return genesis_time + slot_num * interval;
}
int64_t head_block_abs_slot = head_block_time().sec_since_epoch() / interval;
fc::time_point_sec head_slot_time(head_block_abs_slot * interval);
const global_property_object& gpo = get_global_properties();
if( dpo.dynamic_flags & dynamic_global_property_object::maintenance_flag )
slot_num += gpo.parameters.maintenance_skip_slots;
// "slot 0" is head_slot_time
// "slot 1" is head_slot_time,
// plus maint interval if head block is a maint block
// plus block interval if head block is not a maint block
return head_slot_time + (slot_num * interval);
}
uint32_t database::get_slot_at_time(fc::time_point_sec when)const
{
fc::time_point_sec first_slot_time = get_slot_time( 1 );
//@ROL std::cout << "@get_slot_at_time " << when.to_iso_string() << " " << first_slot_time.to_iso_string() << "\n";
if( when < first_slot_time )
return 0;
return (when - first_slot_time).to_seconds() / block_interval() + 1;
}
void database::update_witness_schedule()
{
const witness_schedule_object& wso = get_witness_schedule_object();
const global_property_object& gpo = get_global_properties();
if( head_block_num() % gpo.active_witnesses.size() == 0 )
{
modify( wso, [&]( witness_schedule_object& _wso )
{
_wso.current_shuffled_witnesses.clear();
_wso.current_shuffled_witnesses.reserve( gpo.active_witnesses.size() );
for( const witness_id_type& w : gpo.active_witnesses )
_wso.current_shuffled_witnesses.push_back( w );
auto now_hi = uint64_t(head_block_time().sec_since_epoch()) << 32;
for( uint32_t i = 0; i < _wso.current_shuffled_witnesses.size(); ++i )
{
/// High performance random generator
/// http://xorshift.di.unimi.it/
uint64_t k = now_hi + uint64_t(i)*2685821657736338717ULL;
k ^= (k >> 12);
k ^= (k << 25);
k ^= (k >> 27);
k *= 2685821657736338717ULL;
uint32_t jmax = _wso.current_shuffled_witnesses.size() - i;
uint32_t j = i + k%jmax;
std::swap( _wso.current_shuffled_witnesses[i],
_wso.current_shuffled_witnesses[j] );
}
});
}
}
vector<witness_id_type> database::get_near_witness_schedule()const
{
const witness_schedule_object& wso = get_witness_schedule_object();
vector<witness_id_type> result;
result.reserve(wso.scheduler.size());
uint32_t slot_num = 0;
witness_id_type wid;
while( wso.scheduler.get_slot(slot_num++, wid) )
result.emplace_back(wid);
return result;
}
void database::update_witness_schedule(const signed_block& next_block)
{
auto start = fc::time_point::now();
const global_property_object& gpo = get_global_properties();
const witness_schedule_object& wso = get_witness_schedule_object();
uint32_t schedule_needs_filled = gpo.active_witnesses.size();
uint32_t schedule_slot = get_slot_at_time(next_block.timestamp);
// We shouldn't be able to generate _pending_block with timestamp
// in the past, and incoming blocks from the network with timestamp
// in the past shouldn't be able to make it this far without
// triggering FC_ASSERT elsewhere
assert( schedule_slot > 0 );
witness_id_type first_witness;
bool slot_is_near = wso.scheduler.get_slot( schedule_slot-1, first_witness );
witness_id_type wit;
const dynamic_global_property_object& dpo = get_dynamic_global_properties();
assert( dpo.random.data_size() == witness_scheduler_rng::seed_length );
assert( witness_scheduler_rng::seed_length == wso.rng_seed.size() );
modify(wso, [&](witness_schedule_object& _wso)
{
_wso.slots_since_genesis += schedule_slot;
witness_scheduler_rng rng(wso.rng_seed.data, _wso.slots_since_genesis);
_wso.scheduler._min_token_count = std::max(int(gpo.active_witnesses.size()) / 2, 1);
if( slot_is_near )
{
uint32_t drain = schedule_slot;
while( drain > 0 )
{
if( _wso.scheduler.size() == 0 )
break;
_wso.scheduler.consume_schedule();
--drain;
}
}
else
{
_wso.scheduler.reset_schedule( first_witness );
}
while( !_wso.scheduler.get_slot(schedule_needs_filled, wit) )
{
if( _wso.scheduler.produce_schedule(rng) & emit_turn )
memcpy(_wso.rng_seed.begin(), dpo.random.data(), dpo.random.data_size());
}
_wso.last_scheduling_block = next_block.block_num();
_wso.recent_slots_filled = (
(_wso.recent_slots_filled << 1)
+ 1) << (schedule_slot - 1);
});
auto end = fc::time_point::now();
static uint64_t total_time = 0;
static uint64_t calls = 0;
total_time += (end - start).count();
if( ++calls % 1000 == 0 )
idump( ( double(total_time/1000000.0)/calls) );
}
uint32_t database::update_witness_missed_blocks( const signed_block& b )
{
uint32_t missed_blocks = get_slot_at_time( b.timestamp );
FC_ASSERT( missed_blocks != 0, "Trying to push double-produced block onto current block?!" );
missed_blocks--;
const auto& witnesses = witness_schedule_id_type()(*this).current_shuffled_witnesses;
if( missed_blocks < witnesses.size() )
for( uint32_t i = 0; i < missed_blocks; ++i ) {
const auto& witness_missed = get_scheduled_witness( i+1 )(*this);
modify( witness_missed, []( witness_object& w ) {
w.total_missed++;
});
}
return missed_blocks;
}
uint32_t database::witness_participation_rate()const
{
const global_property_object& gpo = get_global_properties();
if (gpo.parameters.witness_schedule_algorithm == GRAPHENE_WITNESS_SHUFFLED_ALGORITHM)
{
const dynamic_global_property_object& dpo = get_dynamic_global_properties();
return uint64_t(GRAPHENE_100_PERCENT) * dpo.recent_slots_filled.popcount() / 128;
}
if (gpo.parameters.witness_schedule_algorithm == GRAPHENE_WITNESS_SCHEDULED_ALGORITHM)
{
const witness_schedule_object& wso = get_witness_schedule_object();
return uint64_t(GRAPHENE_100_PERCENT) * wso.recent_slots_filled.popcount() / 128;
}
return 0;
}
} }