Merge pull request #2329 from nspcc-dev/dont-request-queued-blocks

Don't request queued blocks again
This commit is contained in:
Roman Khimov 2022-01-18 18:12:51 +03:00 committed by GitHub
commit 01e9af6a30
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
3 changed files with 36 additions and 15 deletions

View file

@ -17,8 +17,9 @@ type Blockqueuer interface {
type blockQueue struct {
log *zap.Logger
queueLock sync.Mutex
queueLock sync.RWMutex
queue []*block.Block
lastQ uint32
checkBlocks chan struct{}
chain Blockqueuer
relayF func(*block.Block)
@ -32,6 +33,10 @@ const (
blockCacheSize = 2000
)
func indexToPosition(i uint32) int {
return int(i) % blockCacheSize
}
func newBlockQueue(capacity int, bc Blockqueuer, log *zap.Logger, relayer func(*block.Block)) *blockQueue {
if log == nil {
return nil
@ -56,12 +61,12 @@ func (bq *blockQueue) run() {
}
for {
h := bq.chain.BlockHeight()
pos := int(h+1) % blockCacheSize
pos := indexToPosition(h + 1)
bq.queueLock.Lock()
b := bq.queue[pos]
// The chain moved forward using blocks from other sources (consensus).
for i := lastHeight; i < h; i++ {
old := int(i+1) % blockCacheSize
old := indexToPosition(i + 1)
if bq.queue[old] != nil && bq.queue[old].Index == i {
bq.len--
bq.queue[old] = nil
@ -106,11 +111,15 @@ func (bq *blockQueue) putBlock(block *block.Block) error {
bq.queueLock.Unlock()
return nil
}
pos := block.Index % blockCacheSize
pos := indexToPosition(block.Index)
// If we already have it, keep the old block, throw away new one.
if bq.queue[pos] == nil || bq.queue[pos].Index < block.Index {
bq.len++
bq.queue[pos] = block
for pos < blockCacheSize && bq.queue[pos] != nil && bq.lastQ+1 == bq.queue[pos].Index {
bq.lastQ = bq.queue[pos].Index
pos++
}
}
l := bq.len
bq.queueLock.Unlock()
@ -125,6 +134,12 @@ func (bq *blockQueue) putBlock(block *block.Block) error {
return nil
}
func (bq *blockQueue) lastQueued() uint32 {
bq.queueLock.RLock()
defer bq.queueLock.RUnlock()
return bq.lastQ
}
func (bq *blockQueue) discard() {
if bq.discarded.CAS(false, true) {
close(bq.checkBlocks)

View file

@ -22,6 +22,7 @@ func TestBlockQueue(t *testing.T) {
for i := 3; i < 5; i++ {
assert.NoError(t, bq.putBlock(blocks[i]))
}
assert.Equal(t, uint32(0), bq.lastQueued())
// nothing should be put into the blockchain
assert.Equal(t, uint32(0), chain.BlockHeight())
assert.Equal(t, 2, bq.length())
@ -30,6 +31,7 @@ func TestBlockQueue(t *testing.T) {
assert.NoError(t, bq.putBlock(blocks[i]))
}
// but they're still not put into the blockchain, because bq isn't running
assert.Equal(t, uint32(4), bq.lastQueued())
assert.Equal(t, uint32(0), chain.BlockHeight())
assert.Equal(t, 4, bq.length())
// block with too big index is dropped
@ -37,17 +39,15 @@ func TestBlockQueue(t *testing.T) {
assert.Equal(t, 4, bq.length())
go bq.run()
// run() is asynchronous, so we need some kind of timeout anyway and this is the simplest one
for i := 0; i < 5; i++ {
if chain.BlockHeight() != 4 {
time.Sleep(time.Second)
}
}
assert.Eventually(t, func() bool { return chain.BlockHeight() == 4 }, 4*time.Second, 100*time.Millisecond)
assert.Equal(t, uint32(4), bq.lastQueued())
assert.Equal(t, 0, bq.length())
assert.Equal(t, uint32(4), chain.BlockHeight())
// put some old blocks
for i := 1; i < 5; i++ {
assert.NoError(t, bq.putBlock(blocks[i]))
}
assert.Equal(t, uint32(4), bq.lastQueued())
assert.Equal(t, 0, bq.length())
assert.Equal(t, uint32(4), chain.BlockHeight())
// unexpected blocks with run() active
@ -64,11 +64,8 @@ func TestBlockQueue(t *testing.T) {
assert.NoError(t, bq.putBlock(blocks[6]))
assert.NoError(t, bq.putBlock(blocks[5]))
// run() is asynchronous, so we need some kind of timeout anyway and this is the simplest one
for i := 0; i < 5; i++ {
if chain.BlockHeight() != 8 {
time.Sleep(time.Second)
}
}
assert.Eventually(t, func() bool { return chain.BlockHeight() == 8 }, 4*time.Second, 100*time.Millisecond)
assert.Equal(t, uint32(8), bq.lastQueued())
assert.Equal(t, 1, bq.length())
assert.Equal(t, uint32(8), chain.BlockHeight())
bq.discard()

View file

@ -1070,7 +1070,16 @@ func (s *Server) handleGetAddrCmd(p Peer) error {
// 2. Send requests for chunk in increasing order.
// 3. After all requests were sent, request random height.
func (s *Server) requestBlocks(bq Blockqueuer, p Peer) error {
pl := getRequestBlocksPayload(p, bq.BlockHeight(), &s.lastRequestedBlock)
h := bq.BlockHeight()
pl := getRequestBlocksPayload(p, h, &s.lastRequestedBlock)
lq := s.bQueue.lastQueued()
if lq > pl.IndexStart {
c := int16(h + blockCacheSize - lq)
if c < payload.MaxHashesCount {
pl.Count = c
}
pl.IndexStart = lq + 1
}
return p.EnqueueP2PMessage(NewMessage(CMDGetBlockByIndex, pl))
}