Replace sync/atomic with uber-go/atomic (#2935)

* Replace sync/atomic with uber-go/atomic

Replaces the usage of sync/atomic with uber-go/atomic across the whole
codebase.

This makes all access in the atomic variables atomic everywhere.

Related with: #2928

Signed-off-by: Javier Palomo <javier.palomo.almena@gmail.com>

* scripts: Verify that we are not using restricted packages

It checks that we are not directly importing 'sync/atomic'.

Signed-off-by: Javier Palomo <javier.palomo.almena@gmail.com>

* Remove unnecesary addition to CHANGELOG

Signed-off-by: Javier Palomo <javier.palomo.almena@gmail.com>

* Use faillint to ensure that sync/atomic is not being used

Signed-off-by: Javier Palomo <javier.palomo.almena@gmail.com>

* Faillint: Add a replacement suggestion for 'sync/atomic'

Signed-off-by: Javier Palomo <javier.palomo.almena@gmail.com>

* Kick off CI

Signed-off-by: Giedrius Statkevičius <giedriuswork@gmail.com>

Co-authored-by: Giedrius Statkevičius <giedriuswork@gmail.com>
Signed-off-by: soniasingla <soniasingla.1812@gmail.com>

Makefile

@@ -266,7 +266,8 @@ github.com/prometheus/prometheus/pkg/testutils=github.com/thanos-io/thanos/pkg/t
 github.com/prometheus/client_golang/prometheus.{DefaultGatherer,DefBuckets,NewUntypedFunc,UntypedFunc},\
 github.com/prometheus/client_golang/prometheus.{NewCounter,NewCounterVec,NewCounterVec,NewGauge,NewGaugeVec,NewGaugeFunc,\
 NewHistorgram,NewHistogramVec,NewSummary,NewSummaryVec}=github.com/prometheus/client_golang/prometheus/promauto.{NewCounter,\
-NewCounterVec,NewCounterVec,NewGauge,NewGaugeVec,NewGaugeFunc,NewHistorgram,NewHistogramVec,NewSummary,NewSummaryVec}" ./...
+NewCounterVec,NewCounterVec,NewGauge,NewGaugeVec,NewGaugeFunc,NewHistorgram,NewHistogramVec,NewSummary,NewSummaryVec},\
+sync/atomic=go.uber.org/atomic" ./...
 	@$(FAILLINT) -paths "fmt.{Print,Println,Sprint}" -ignore-tests ./...
 	@echo ">> linting all of the Go files GOGC=${GOGC}"
 	@$(GOLANGCI_LINT) run

go.mod

@@ -54,6 +54,7 @@ require (
 	github.com/uber/jaeger-lib v2.2.0+incompatible
 	go.elastic.co/apm v1.5.0
 	go.elastic.co/apm/module/apmot v1.5.0
+	go.uber.org/atomic v1.6.0
 	go.uber.org/automaxprocs v1.2.0
 	golang.org/x/crypto v0.0.0-20200622213623-75b288015ac9
 	golang.org/x/oauth2 v0.0.0-20200107190931-bf48bf16ab8d

pkg/alert/alert.go

@@ -14,7 +14,6 @@ import (
 	"net/url"
 	"path"
 	"sync"
-	"sync/atomic"
 	"time"
 
 	"github.com/go-kit/kit/log"
@@ -25,6 +24,7 @@ import (
 	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/client_golang/prometheus/promauto"
 	"github.com/prometheus/prometheus/pkg/labels"
+	"go.uber.org/atomic"
 
 	"github.com/thanos-io/thanos/pkg/runutil"
 	"github.com/thanos-io/thanos/pkg/tracing"
@@ -370,7 +370,7 @@ func (s *Sender) Send(ctx context.Context, alerts []*Alert) {
 
 	var (
 		wg         sync.WaitGroup
-		numSuccess uint64
+		numSuccess atomic.Uint64
 	)
 	for _, am := range s.alertmanagers {
 		for _, u := range am.dispatcher.Endpoints() {
@@ -396,14 +396,14 @@ func (s *Sender) Send(ctx context.Context, alerts []*Alert) {
 					s.latency.WithLabelValues(u.Host).Observe(time.Since(start).Seconds())
 					s.sent.WithLabelValues(u.Host).Add(float64(len(alerts)))
 
-					atomic.AddUint64(&numSuccess, 1)
+					numSuccess.Inc()
 				})
 			}(am, *u)
 		}
 	}
 	wg.Wait()
 
-	if numSuccess > 0 {
+	if numSuccess.Load() > 0 {
 		return
 	}
 

pkg/prober/http.go

@@ -6,18 +6,18 @@ package prober
 import (
 	"io"
 	"net/http"
-	"sync/atomic"
 
 	"github.com/go-kit/kit/log"
 	"github.com/go-kit/kit/log/level"
+	"go.uber.org/atomic"
 )
 
 type check func() bool
 
 // HTTPProbe represents health and readiness status of given component, and provides HTTP integration.
 type HTTPProbe struct {
-	ready   uint32
-	healthy uint32
+	ready   atomic.Uint32
+	healthy atomic.Uint32
 }
 
 // NewHTTP returns HTTPProbe representing readiness and healthiness of given component.
@@ -49,34 +49,33 @@ func (p *HTTPProbe) handler(logger log.Logger, c check) http.HandlerFunc {
 
 // isReady returns true if component is ready.
 func (p *HTTPProbe) isReady() bool {
-	ready := atomic.LoadUint32(&p.ready)
+	ready := p.ready.Load()
 	return ready > 0
 }
 
 // isHealthy returns true if component is healthy.
 func (p *HTTPProbe) isHealthy() bool {
-	healthy := atomic.LoadUint32(&p.healthy)
+	healthy := p.healthy.Load()
 	return healthy > 0
 }
 
 // Ready sets components status to ready.
 func (p *HTTPProbe) Ready() {
-	atomic.SwapUint32(&p.ready, 1)
+	p.ready.Swap(1)
 }
 
 // NotReady sets components status to not ready with given error as a cause.
 func (p *HTTPProbe) NotReady(err error) {
-	atomic.SwapUint32(&p.ready, 0)
+	p.ready.Swap(0)
 
 }
 
 // Healthy sets components status to healthy.
 func (p *HTTPProbe) Healthy() {
-	atomic.SwapUint32(&p.healthy, 1)
-
+	p.healthy.Swap(1)
 }
 
 // NotHealthy sets components status to not healthy with given error as a cause.
 func (p *HTTPProbe) NotHealthy(err error) {
-	atomic.SwapUint32(&p.healthy, 0)
+	p.healthy.Swap(0)
 }

pkg/reloader/reloader_test.go

@@ -15,12 +15,12 @@ import (
 	"path/filepath"
 	"strings"
 	"sync"
-	"sync/atomic"
 	"testing"
 	"time"
 
 	"github.com/fortytw2/leaktest"
 	"github.com/thanos-io/thanos/pkg/testutil"
+	"go.uber.org/atomic"
 )
 
 func TestReloader_ConfigApply(t *testing.T) {

pkg/store/bucket_test.go

@@ -19,7 +19,6 @@ import (
 	"sort"
 	"strconv"
 	"sync"
-	"sync/atomic"
 	"testing"
 	"time"
 
@@ -52,6 +51,7 @@ import (
 	storetestutil "github.com/thanos-io/thanos/pkg/store/storepb/testutil"
 	"github.com/thanos-io/thanos/pkg/testutil"
 	"github.com/thanos-io/thanos/pkg/testutil/e2eutil"
+	"go.uber.org/atomic"
 )
 
 var emptyRelabelConfig = make([]*relabel.Config, 0)
@@ -1281,10 +1281,10 @@ func benchBucketSeries(t testutil.TB, samplesPerSeries, totalSeries int, request
 
 	if !t.IsBenchmark() {
 		// Make sure the pool is correctly used. This is expected for 200k numbers.
-		testutil.Equals(t, numOfBlocks, int(chunkPool.(*mockedPool).gets))
+		testutil.Equals(t, numOfBlocks, int(chunkPool.(*mockedPool).gets.Load()))
 		// TODO(bwplotka): This is wrong negative for large number of samples (1mln). Investigate.
-		testutil.Equals(t, 0, int(chunkPool.(*mockedPool).balance))
-		chunkPool.(*mockedPool).gets = 0
+		testutil.Equals(t, 0, int(chunkPool.(*mockedPool).balance.Load()))
+		chunkPool.(*mockedPool).gets.Store(0)
 
 		for _, b := range blocks {
 			// NOTE(bwplotka): It is 4 x 1.0 for 100mln samples. Kind of make sense: long series.
@@ -1306,22 +1306,22 @@ func (m fakePool) Put(_ *[]byte) {}
 
 type mockedPool struct {
 	parent  pool.BytesPool
-	balance uint64
-	gets    uint64
+	balance atomic.Uint64
+	gets    atomic.Uint64
 }
 
 func (m *mockedPool) Get(sz int) (*[]byte, error) {
 	b, err := m.parent.Get(sz)
 	if err != nil {
 		return nil, err
 	}
-	atomic.AddUint64(&m.balance, uint64(cap(*b)))
-	atomic.AddUint64(&m.gets, uint64(1))
+	m.balance.Add(uint64(cap(*b)))
+	m.gets.Add(uint64(1))
 	return b, nil
 }
 
 func (m *mockedPool) Put(b *[]byte) {
-	atomic.AddUint64(&m.balance, ^uint64(cap(*b)-1))
+	m.balance.Sub(uint64(cap(*b)))
 	m.parent.Put(b)
 }
 

pkg/store/limiter.go

@@ -5,10 +5,10 @@ package store
 
 import (
 	"sync"
-	"sync/atomic"
 
 	"github.com/pkg/errors"
 	"github.com/prometheus/client_golang/prometheus"
+	"go.uber.org/atomic"
 )
 
 type ChunksLimiter interface {
@@ -25,7 +25,7 @@ type ChunksLimiterFactory func(failedCounter prometheus.Counter) ChunksLimiter
 // Limiter is a simple mechanism for checking if something has passed a certain threshold.
 type Limiter struct {
 	limit    uint64
-	reserved uint64
+	reserved atomic.Uint64
 
 	// Counter metric which we will increase if limit is exceeded.
 	failedCounter prometheus.Counter
@@ -42,7 +42,7 @@ func (l *Limiter) Reserve(num uint64) error {
 	if l.limit == 0 {
 		return nil
 	}
-	if reserved := atomic.AddUint64(&l.reserved, num); reserved > l.limit {
+	if reserved := l.reserved.Add(num); reserved > l.limit {
 		// We need to protect from the counter being incremented twice due to concurrency
 		// while calling Reserve().
 		l.failedOnce.Do(l.failedCounter.Inc)