Make bootstrapping handle its own timeouts

Currently, an engine registers timeouts into the handler, which schedules the timeouts on behalf of the the engine.
The handler then notifies the engine when the timeout expired.

However, the only engine that uses this mechanism is the bootstrapping engine, and not the other engine types such as the snowman and state sync engines.

It therefore makes sense that the bootstrapper handle its own timeouts instead of delegating them to the handler.
By moving the timeout handling into the bootstrapper, we can make the API of the common.Engine be slimmer by removing the Timeout() method from it.

Signed-off-by: Yacov Manevich <yacov.manevich@avalabs.org>

Author: yacovm

chains/manager.go

@@ -962,7 +962,6 @@ func (m *manager) createAvalancheChain(
 		StartupTracker:                 startupTracker,
 		Sender:                         snowmanMessageSender,
 		BootstrapTracker:               sb,
-		Timer:                          h,
 		PeerTracker:                    peerTracker,
 		AncestorsMaxContainersReceived: m.BootstrapAncestorsMaxContainersReceived,
 		DB:                             blockBootstrappingDB,
@@ -1358,7 +1357,6 @@ func (m *manager) createSnowmanChain(
 		StartupTracker:                 startupTracker,
 		Sender:                         messageSender,
 		BootstrapTracker:               sb,
-		Timer:                          h,
 		PeerTracker:                    peerTracker,
 		AncestorsMaxContainersReceived: m.BootstrapAncestorsMaxContainersReceived,
 		DB:                             bootstrappingDB,

message/internal_msg_builder.go

@@ -16,7 +16,6 @@ import (
 var (
 	disconnected  = &Disconnected{}
 	gossipRequest = &GossipRequest{}
-	timeout       = &Timeout{}
 
 	_ fmt.Stringer    = (*GetStateSummaryFrontierFailed)(nil)
 	_ chainIDGetter   = (*GetStateSummaryFrontierFailed)(nil)
@@ -50,8 +49,6 @@ var (
 	_ fmt.Stringer = (*Disconnected)(nil)
 
 	_ fmt.Stringer = (*GossipRequest)(nil)
-
-	_ fmt.Stringer = (*Timeout)(nil)
 )
 
 type GetStateSummaryFrontierFailed struct {
@@ -391,18 +388,3 @@ func InternalGossipRequest(
 		expiration: mockable.MaxTime,
 	}
 }
-
-type Timeout struct{}
-
-func (Timeout) String() string {
-	return ""
-}
-
-func InternalTimeout(nodeID ids.NodeID) InboundMessage {
-	return &inboundMessage{
-		nodeID:     nodeID,
-		op:         TimeoutOp,
-		message:    timeout,
-		expiration: mockable.MaxTime,
-	}
-}

snow/engine/common/bootstrap_tracker.go

@@ -13,6 +13,4 @@ type BootstrapTracker interface {
 
 	// Bootstrapped marks the named chain as being bootstrapped
 	Bootstrapped(chainID ids.ID)
-
-	OnBootstrapCompleted() chan struct{}
 }

snow/engine/common/engine.go

@@ -415,9 +415,6 @@ type InternalHandler interface {
 	// Notify this engine of peer changes.
 	validators.Connector
 
-	// Notify this engine that a registered timeout has fired.
-	Timeout(context.Context) error
-
 	// Gossip to the network a container on the accepted frontier
 	Gossip(context.Context) error
 

snow/engine/common/timer.go

@@ -2,13 +2,3 @@
 // See the file LICENSE for licensing terms.
 
 package common
-
-import "time"
-
-// Timer describes the standard interface for specifying a timeout
-type Timer interface {
-	// RegisterTimeout specifies how much time to delay the next timeout message
-	// by. If the subnet has been bootstrapped, the timeout will fire
-	// immediately.
-	RegisterTimeout(time.Duration)
-}

snow/engine/common/traced_engine.go

@@ -329,13 +329,6 @@ func (e *tracedEngine) Disconnected(ctx context.Context, nodeID ids.NodeID) erro
 	return e.engine.Disconnected(ctx, nodeID)
 }
 
-func (e *tracedEngine) Timeout(ctx context.Context) error {
-	ctx, span := e.tracer.Start(ctx, "tracedEngine.Timeout")
-	defer span.End()
-
-	return e.engine.Timeout(ctx)
-}
-
 func (e *tracedEngine) Gossip(ctx context.Context) error {
 	ctx, span := e.tracer.Start(ctx, "tracedEngine.Gossip")
 	defer span.End()

snow/engine/enginetest/engine.go

@@ -19,7 +19,6 @@ import (
 )
 
 var (
-	errTimeout                       = errors.New("unexpectedly called Timeout")
 	errGossip                        = errors.New("unexpectedly called Gossip")
 	errNotify                        = errors.New("unexpectedly called Notify")
 	errGetStateSummaryFrontier       = errors.New("unexpectedly called GetStateSummaryFrontier")
@@ -189,19 +188,6 @@ func (e *Engine) Start(ctx context.Context, startReqID uint32) error {
 	return errStart
 }
 
-func (e *Engine) Timeout(ctx context.Context) error {
-	if e.TimeoutF != nil {
-		return e.TimeoutF(ctx)
-	}
-	if !e.CantTimeout {
-		return nil
-	}
-	if e.T != nil {
-		require.FailNow(e.T, errTimeout.Error())
-	}
-	return errTimeout
-}
-
 func (e *Engine) Gossip(ctx context.Context) error {
 	if e.GossipF != nil {
 		return e.GossipF(ctx)

snow/engine/enginetest/timer.go

@@ -8,12 +8,8 @@ import (
 	"time"
 
 	"github.com/stretchr/testify/require"
-
-	"github.com/ava-labs/avalanchego/snow/engine/common"
 )
 
-var _ common.Timer = (*Timer)(nil)
-
 // Timer is a test timer
 type Timer struct {
 	T *testing.T
@@ -23,15 +19,14 @@ type Timer struct {
 	RegisterTimeoutF func(time.Duration)
 }
 
-// Default set the default callable value to [cant]
-func (t *Timer) Default(cant bool) {
-	t.CantRegisterTimout = cant
-}
-
 func (t *Timer) RegisterTimeout(delay time.Duration) {
 	if t.RegisterTimeoutF != nil {
 		t.RegisterTimeoutF(delay)
 	} else if t.CantRegisterTimout && t.T != nil {
 		require.FailNow(t.T, "Unexpectedly called RegisterTimeout")
 	}
 }
+
+func (t *Timer) Preempt() {
+	t.RegisterTimeout(time.Duration(0))
+}

snow/engine/snowman/bootstrap/bootstrapper.go

@@ -70,7 +70,7 @@ type Bootstrapper struct {
 	Config
 	shouldHalt func() bool
 	*metrics
-
+	TimeoutRegistry TimeoutRegistry
 	// list of NoOpsHandler for messages dropped by bootstrapper
 	common.StateSummaryFrontierHandler
 	common.AcceptedStateSummaryHandler
@@ -119,7 +119,7 @@ type Bootstrapper struct {
 
 func New(config Config, onFinished func(ctx context.Context, lastReqID uint32) error) (*Bootstrapper, error) {
 	metrics, err := newMetrics(config.Ctx.Registerer)
-	return &Bootstrapper{
+	bs := &Bootstrapper{
 		shouldHalt:                  config.ShouldHalt,
 		nonVerifyingParser:          config.NonVerifyingParse,
 		Config:                      config,
@@ -139,7 +139,15 @@ func New(config Config, onFinished func(ctx context.Context, lastReqID uint32) e
 
 		executedStateTransitions: math.MaxInt,
 		onFinished:               onFinished,
-	}, err
+	}
+
+	bs.TimeoutRegistry = newTimeoutHandler(func() {
+		if err := bs.Timeout(context.TODO()); err != nil {
+			bs.Config.Ctx.Log.Warn("Encountered error during bootstrapping: %w", zap.Error(err))
+		}
+	})
+
+	return bs, err
 }
 
 func (b *Bootstrapper) Context() *snow.ConsensusContext {
@@ -696,14 +704,15 @@ func (b *Bootstrapper) tryStartExecuting(ctx context.Context) error {
 
 	// Notify the subnet that this chain is synced
 	b.Config.BootstrapTracker.Bootstrapped(b.Ctx.ChainID)
+	b.TimeoutRegistry.Preempt()
 
 	// If the subnet hasn't finished bootstrapping, this chain should remain
 	// syncing.
 	if !b.Config.BootstrapTracker.IsBootstrapped() {
 		log("waiting for the remaining chains in this subnet to finish syncing")
 		// Restart bootstrapping after [bootstrappingDelay] to keep up to date
 		// on the latest tip.
-		b.Config.Timer.RegisterTimeout(bootstrappingDelay)
+		b.TimeoutRegistry.RegisterTimeout(bootstrappingDelay)
 		b.awaitingTimeout = true
 		return nil
 	}

snow/engine/snowman/bootstrap/bootstrapper_test.go

@@ -103,7 +103,6 @@ func newConfig(t *testing.T) (Config, ids.NodeID, *enginetest.Sender, *blocktest
 		PeerTracker:                    peerTracker,
 		Sender:                         sender,
 		BootstrapTracker:               bootstrapTracker,
-		Timer:                          &enginetest.Timer{},
 		AncestorsMaxContainersReceived: 2000,
 		DB:                             memdb.New(),
 		VM:                             vm,
@@ -155,7 +154,6 @@ func TestBootstrapperStartsOnlyIfEnoughStakeIsConnected(t *testing.T) {
 		PeerTracker:                    peerTracker,
 		Sender:                         sender,
 		BootstrapTracker:               &enginetest.BootstrapTracker{},
-		Timer:                          &enginetest.Timer{},
 		AncestorsMaxContainersReceived: 2000,
 		DB:                             memdb.New(),
 		VM:                             vm,
@@ -180,6 +178,7 @@ func TestBootstrapperStartsOnlyIfEnoughStakeIsConnected(t *testing.T) {
 	}
 	bs, err := New(cfg, dummyCallback)
 	require.NoError(err)
+	bs.TimeoutRegistry = &enginetest.Timer{}
 
 	vm.CantSetState = false
 	vm.CantConnected = true
@@ -236,6 +235,7 @@ func TestBootstrapperSingleFrontier(t *testing.T) {
 		},
 	)
 	require.NoError(err)
+	bs.TimeoutRegistry = &enginetest.Timer{}
 
 	require.NoError(bs.Start(context.Background(), 0))
 
@@ -264,6 +264,7 @@ func TestBootstrapperUnknownByzantineResponse(t *testing.T) {
 		},
 	)
 	require.NoError(err)
+	bs.TimeoutRegistry = &enginetest.Timer{}
 
 	require.NoError(bs.Start(context.Background(), 0))
 
@@ -309,6 +310,7 @@ func TestBootstrapperPartialFetch(t *testing.T) {
 		},
 	)
 	require.NoError(err)
+	bs.TimeoutRegistry = &enginetest.Timer{}
 
 	require.NoError(bs.Start(context.Background(), 0))
 
@@ -359,6 +361,7 @@ func TestBootstrapperEmptyResponse(t *testing.T) {
 		},
 	)
 	require.NoError(err)
+	bs.TimeoutRegistry = &enginetest.Timer{}
 
 	require.NoError(bs.Start(context.Background(), 0))
 
@@ -407,6 +410,7 @@ func TestBootstrapperAncestors(t *testing.T) {
 		},
 	)
 	require.NoError(err)
+	bs.TimeoutRegistry = &enginetest.Timer{}
 
 	require.NoError(bs.Start(context.Background(), 0))
 
@@ -452,6 +456,7 @@ func TestBootstrapperFinalized(t *testing.T) {
 		},
 	)
 	require.NoError(err)
+	bs.TimeoutRegistry = &enginetest.Timer{}
 
 	require.NoError(bs.Start(context.Background(), 0))
 
@@ -494,6 +499,7 @@ func TestRestartBootstrapping(t *testing.T) {
 		},
 	)
 	require.NoError(err)
+	bs.TimeoutRegistry = &enginetest.Timer{}
 
 	require.NoError(bs.Start(context.Background(), 0))
 
@@ -558,6 +564,7 @@ func TestBootstrapOldBlockAfterStateSync(t *testing.T) {
 		},
 	)
 	require.NoError(err)
+	bs.TimeoutRegistry = &enginetest.Timer{}
 
 	require.NoError(bs.Start(context.Background(), 0))
 
@@ -598,6 +605,7 @@ func TestBootstrapContinueAfterHalt(t *testing.T) {
 		},
 	)
 	require.NoError(err)
+	bs.TimeoutRegistry = &enginetest.Timer{}
 
 	getBlockF := vm.GetBlockF
 	vm.GetBlockF = func(ctx context.Context, blkID ids.ID) (snowman.Block, error) {
@@ -690,7 +698,6 @@ func TestBootstrapNoParseOnNew(t *testing.T) {
 		PeerTracker:                    peerTracker,
 		Sender:                         sender,
 		BootstrapTracker:               bootstrapTracker,
-		Timer:                          &enginetest.Timer{},
 		AncestorsMaxContainersReceived: 2000,
 		DB:                             intervalDB,
 		VM:                             vm,
@@ -728,6 +735,7 @@ func TestBootstrapperReceiveStaleAncestorsMessage(t *testing.T) {
 		},
 	)
 	require.NoError(err)
+	bs.TimeoutRegistry = &enginetest.Timer{}
 
 	require.NoError(bs.Start(context.Background(), 0))
 
@@ -772,6 +780,7 @@ func TestBootstrapperRollbackOnSetState(t *testing.T) {
 			return nil
 		},
 	)
+	bs.TimeoutRegistry = &enginetest.Timer{}
 	require.NoError(err)
 
 	vm.SetStateF = func(context.Context, snow.State) error {