store/tikv: refine code, backoffer is-a go context (#5276)

store/tikv/backoff.go

@@ -40,10 +40,10 @@ const (
 // NewBackoffFn creates a backoff func which implements exponential backoff with
 // optional jitters.
 // See http://www.awsarchitectureblog.com/2015/03/backoff.html
-func NewBackoffFn(base, cap, jitter int) func() int {
+func NewBackoffFn(base, cap, jitter int) func(goCtx goctx.Context) int {
 	attempts := 0
 	lastSleep := base
-	return func() int {
+	return func(goCtx goctx.Context) int {
 		var sleep int
 		switch jitter {
 		case NoJitter:
@@ -57,7 +57,11 @@ func NewBackoffFn(base, cap, jitter int) func() int {
 		case DecorrJitter:
 			sleep = int(math.Min(float64(cap), float64(base+rand.Intn(lastSleep*3-base))))
 		}
-		time.Sleep(time.Duration(sleep) * time.Millisecond)
+
+		select {
+		case <-time.After(time.Duration(sleep) * time.Millisecond):
+		case <-goCtx.Done():
+		}
 
 		attempts++
 		lastSleep = sleep
@@ -81,7 +85,7 @@ const (
 	boServerBusy
 )
 
-func (t backoffType) createFn() func() int {
+func (t backoffType) createFn() func(goctx.Context) int {
 	switch t {
 	case boTiKVRPC:
 		return NewBackoffFn(100, 2000, EqualJitter)
@@ -154,19 +158,20 @@ var commitMaxBackoff = 20000
 
 // Backoffer is a utility for retrying queries.
 type Backoffer struct {
-	fn         map[backoffType]func() int
+	goctx.Context
+
+	fn         map[backoffType]func(goctx.Context) int
 	maxSleep   int
 	totalSleep int
 	errors     []error
-	ctx        goctx.Context
 	types      []backoffType
 }
 
 // NewBackoffer creates a Backoffer with maximum sleep time(in ms).
 func NewBackoffer(maxSleep int, ctx goctx.Context) *Backoffer {
 	return &Backoffer{
+		Context:  ctx,
 		maxSleep: maxSleep,
-		ctx:      ctx,
 	}
 }
 
@@ -174,7 +179,7 @@ func NewBackoffer(maxSleep int, ctx goctx.Context) *Backoffer {
 // It returns a retryable error if total sleep time exceeds maxSleep.
 func (b *Backoffer) Backoff(typ backoffType, err error) error {
 	select {
-	case <-b.ctx.Done():
+	case <-b.Context.Done():
 		return errors.Trace(err)
 	default:
 	}
@@ -182,7 +187,7 @@ func (b *Backoffer) Backoff(typ backoffType, err error) error {
 	backoffCounter.WithLabelValues(typ.String()).Inc()
 	// Lazy initialize.
 	if b.fn == nil {
-		b.fn = make(map[backoffType]func() int)
+		b.fn = make(map[backoffType]func(goctx.Context) int)
 	}
 	f, ok := b.fn[typ]
 	if !ok {
@@ -190,7 +195,7 @@ func (b *Backoffer) Backoff(typ backoffType, err error) error {
 		b.fn[typ] = f
 	}
 
-	b.totalSleep += f()
+	b.totalSleep += f(b)
 	b.types = append(b.types, typ)
 
 	log.Debugf("%v, retry later(totalSleep %dms, maxSleep %dms)", err, b.totalSleep, b.maxSleep)
@@ -221,21 +226,21 @@ func (b *Backoffer) String() string {
 // current Backoffer's context.
 func (b *Backoffer) Clone() *Backoffer {
 	return &Backoffer{
+		Context:    b.Context,
 		maxSleep:   b.maxSleep,
 		totalSleep: b.totalSleep,
 		errors:     b.errors,
-		ctx:        b.ctx,
 	}
 }
 
 // Fork creates a new Backoffer which keeps current Backoffer's sleep time and errors, and holds
 // a child context of current Backoffer's context.
 func (b *Backoffer) Fork() (*Backoffer, goctx.CancelFunc) {
-	ctx, cancel := goctx.WithCancel(b.ctx)
+	ctx, cancel := goctx.WithCancel(b.Context)
 	return &Backoffer{
+		Context:    ctx,
 		maxSleep:   b.maxSleep,
 		totalSleep: b.totalSleep,
 		errors:     b.errors,
-		ctx:        ctx,
 	}, cancel
 }

store/tikv/kv.go

@@ -371,7 +371,7 @@ func (s *tikvStore) CurrentVersion() (kv.Version, error) {
 
 func (s *tikvStore) getTimestampWithRetry(bo *Backoffer) (uint64, error) {
 	for {
-		startTS, err := s.oracle.GetTimestamp(bo.ctx)
+		startTS, err := s.oracle.GetTimestamp(bo)
 		if err == nil {
 			return startTS, nil
 		}

store/tikv/region_cache.go

@@ -333,7 +333,7 @@ func (c *RegionCache) loadRegion(bo *Backoffer, key []byte) (*Region, error) {
 			}
 		}
 
-		meta, leader, err := c.pdClient.GetRegion(bo.ctx, key)
+		meta, leader, err := c.pdClient.GetRegion(bo, key)
 		if err != nil {
 			backoffErr = errors.Errorf("loadRegion from PD failed, key: %q, err: %v", key, err)
 			continue
@@ -367,7 +367,7 @@ func (c *RegionCache) loadRegionByID(bo *Backoffer, regionID uint64) (*Region, e
 			}
 		}
 
-		meta, leader, err := c.pdClient.GetRegionByID(bo.ctx, regionID)
+		meta, leader, err := c.pdClient.GetRegionByID(bo, regionID)
 		if err != nil {
 			backoffErr = errors.Errorf("loadRegion from PD failed, regionID: %v, err: %v", regionID, err)
 			continue
@@ -427,7 +427,7 @@ func (c *RegionCache) ClearStoreByID(id uint64) {
 
 func (c *RegionCache) loadStoreAddr(bo *Backoffer, id uint64) (string, error) {
 	for {
-		store, err := c.pdClient.GetStore(bo.ctx, id)
+		store, err := c.pdClient.GetStore(bo, id)
 		if err != nil {
 			if errors.Cause(err) == goctx.Canceled {
 				return "", errors.Trace(err)

store/tikv/region_request.go

@@ -104,7 +104,7 @@ func (s *RegionRequestSender) sendReqToRegion(bo *Backoffer, ctx *RPCContext, re
 	if e := tikvrpc.SetContext(req, ctx.Meta, ctx.Peer); e != nil {
 		return nil, false, errors.Trace(e)
 	}
-	context, cancel := goctx.WithTimeout(bo.ctx, timeout)
+	context, cancel := goctx.WithTimeout(bo, timeout)
 	defer cancel()
 	resp, err = s.client.SendReq(context, ctx.Addr, req)
 	if err != nil {
@@ -124,7 +124,7 @@ func (s *RegionRequestSender) onSendFail(bo *Backoffer, ctx *RPCContext, err err
 	}
 	if grpc.Code(errors.Cause(err)) == codes.Canceled {
 		select {
-		case <-bo.ctx.Done():
+		case <-bo.Done():
 			return errors.Trace(err)
 		default:
 			// If we don't cancel, but the error code is Canceled, it must be from grpc remote.