[Merged by Bors] - Use dedicated PoST service for Proof generation

CHANGELOG.md

@@ -14,11 +14,14 @@ See [RELEASE](./RELEASE.md) for workflow instructions.
 
 * [#5118](https://github.com/spacemeshos/go-spacemesh/pull/5118) reduce number of tortoise results returned after recovery.
 
-this is hotfix for a bug introduced in v1.2.0. in rare conditions node may loop with the following warning:
+  this is hotfix for a bug introduced in v1.2.0. in rare conditions node may loop with the following warning:
 
-> 2023-10-02T15:28:14.002+0200 WARN fd68b.sync mesh failed to process layer from sync {"node_id": "fd68b9397572556c2f329f3e5af2faf23aef85dbbbb7e38447fae2f4ef38899f", "module": "sync", "sessionId": "29422935-68d6-47d1-87a8-02293aa181f3", "layer_id": 23104, "errmsg": "requested layer 8063 is before evicted 13102", "name": "sync"}
+  > 2023-10-02T15:28:14.002+0200 WARN fd68b.sync mesh failed to process layer from sync {"node_id": "fd68b9397572556c2f329f3e5af2faf23aef85dbbbb7e38447fae2f4ef38899f", "module": "sync", "sessionId": "29422935-68d6-47d1-87a8-02293aa181f3", "layer_id": 23104, "errmsg": "requested layer 8063 is before evicted 13102", "name": "sync"}
 
 * [#5091](https://github.com/spacemeshos/go-spacemesh/pull/5091) First stage of separating PoST from the node into its own service.
+* [#5061](https://github.com/spacemeshos/go-spacemesh/pull/5061) Proof generation is now done via a dedicated service instead of the node.
+
+  Operating a node doesn't require any changes at the moment. The service will be automatically started by the node if needed and will be stopped when the node is stopped.
 
 * [#5138](https://github.com/spacemeshos/go-spacemesh/pull/5138) Bump poet to v0.9.7
 

activation/activation.go

@@ -12,7 +12,6 @@
 	"sync"
 	"time"
 
-	"github.com/spacemeshos/post/proving"
 	"github.com/spacemeshos/post/shared"
 	"go.uber.org/atomic"
 	"golang.org/x/sync/errgroup"
@@ -88,6 +87,9 @@
 	initialPost       *types.Post
 	validator         nipostValidator
 
+	postMux    sync.Mutex
+	postClient PostClient
+
 	// smeshingMutex protects `StartSmeshing` and `StopSmeshing` from concurrent access
 	smeshingMutex sync.Mutex
 
@@ -183,6 +185,41 @@
 	return b
 }
 
+func (b *Builder) Connected(client PostClient) {
+	b.postMux.Lock()
+	defer b.postMux.Unlock()
+
+	if b.postClient != nil {
+		b.log.With().Error("post service already connected")
+		return
+	}
+
+	b.postClient = client
+}
+
+func (b *Builder) Disconnected(client PostClient) {
+	b.postMux.Lock()
+	defer b.postMux.Unlock()
+
+	if b.postClient != client {
+		b.log.With().Debug("post service not connected")
+		return
+	}
+
+	b.postClient = nil
+}
+
+func (b *Builder) proof(ctx context.Context, challenge []byte) (*types.Post, *types.PostMetadata, error) {
+	b.postMux.Lock()
+	defer b.postMux.Unlock()
+
+	if b.postClient == nil {
+		return nil, nil, errors.New("post service not connected")
+	}
+
+	return b.postClient.Proof(ctx, challenge)
+}
+
 // Smeshing returns true iff atx builder is smeshing.
 func (b *Builder) Smeshing() bool {
 	return b.started.Load()
@@ -333,7 +370,7 @@
 	startTime := time.Now()
 	var err error
 	events.EmitPostStart(shared.ZeroChallenge)
-	post, metadata, err := b.postSetupProvider.GenerateProof(ctx, shared.ZeroChallenge, proving.WithPowCreator(b.nodeID.Bytes()))
+	post, metadata, err := b.proof(ctx, shared.ZeroChallenge)
 	if err != nil {
 		events.EmitPostFailure()
 		return fmt.Errorf("post execution: %w", err)

activation/activation_test.go

@@ -96,12 +96,13 @@ type testAtxBuilder struct {
 	coinbase    types.Address
 	goldenATXID types.ATXID
 
-	mpub       *mocks.MockPublisher
-	mnipost    *MocknipostBuilder
-	mpost      *MockpostSetupProvider
-	mclock     *MocklayerClock
-	msync      *Mocksyncer
-	mValidator *MocknipostValidator
+	mpub        *mocks.MockPublisher
+	mnipost     *MocknipostBuilder
+	mpost       *MockpostSetupProvider
+	mpostClient *MockPostClient
+	mclock      *MocklayerClock
+	msync       *Mocksyncer
+	mValidator  *MocknipostValidator
 }
 
 func newTestBuilder(tb testing.TB, opts ...BuilderOption) *testAtxBuilder {
@@ -118,6 +119,7 @@ func newTestBuilder(tb testing.TB, opts ...BuilderOption) *testAtxBuilder {
 		mpub:        mocks.NewMockPublisher(ctrl),
 		mnipost:     NewMocknipostBuilder(ctrl),
 		mpost:       NewMockpostSetupProvider(ctrl),
+		mpostClient: NewMockPostClient(ctrl),
 		mclock:      NewMocklayerClock(ctrl),
 		msync:       NewMocksyncer(ctrl),
 		mValidator:  NewMocknipostValidator(ctrl),
@@ -143,6 +145,7 @@ func newTestBuilder(tb testing.TB, opts ...BuilderOption) *testAtxBuilder {
 		Nonce:   0,
 		Indices: make([]byte, 10),
 	}
+	b.Connected(tab.mpostClient)
 	tab.Builder = b
 	dir := tb.TempDir()
 	tab.mnipost.EXPECT().DataDir().Return(dir).AnyTimes()
@@ -250,7 +253,7 @@ func TestBuilder_StartSmeshingCoinbase(t *testing.T) {
 	tab.mpost.EXPECT().StartSession(gomock.Any()).AnyTimes()
 	tab.mpost.EXPECT().LastOpts().Return(&PostSetupOpts{}).AnyTimes()
 	tab.mpost.EXPECT().CommitmentAtx().Return(tab.goldenATXID, nil).AnyTimes()
-	tab.mpost.EXPECT().GenerateProof(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes().Return(&types.Post{}, &types.PostMetadata{}, nil)
+	tab.mpostClient.EXPECT().Proof(gomock.Any(), gomock.Any()).AnyTimes().Return(&types.Post{}, &types.PostMetadata{}, nil)
 	tab.mValidator.EXPECT().Post(gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes().Return(nil)
 	tab.mclock.EXPECT().AwaitLayer(gomock.Any()).Return(make(chan struct{})).AnyTimes()
 	require.NoError(t, tab.StartSmeshing(coinbase, postSetupOpts))
@@ -271,7 +274,7 @@ func TestBuilder_RestartSmeshing(t *testing.T) {
 		tab.mpost.EXPECT().CommitmentAtx().Return(types.EmptyATXID, nil).AnyTimes()
 		tab.mpost.EXPECT().LastOpts().Return(&PostSetupOpts{}).AnyTimes()
 		tab.mpost.EXPECT().StartSession(gomock.Any()).AnyTimes()
-		tab.mpost.EXPECT().GenerateProof(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes().Return(&types.Post{}, &types.PostMetadata{
+		tab.mpostClient.EXPECT().Proof(gomock.Any(), gomock.Any()).AnyTimes().Return(&types.Post{}, &types.PostMetadata{
 			Challenge: shared.ZeroChallenge,
 		}, nil)
 		tab.mpost.EXPECT().Reset().AnyTimes()
@@ -382,7 +385,7 @@ func TestBuilder_StartSmeshing_PanicsOnErrInStartSession(t *testing.T) {
 	tab.log = l
 
 	// Stub these methods in case they get called
-	tab.mpost.EXPECT().GenerateProof(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes().Return(&types.Post{}, &types.PostMetadata{}, nil)
+	tab.mpostClient.EXPECT().Proof(gomock.Any(), gomock.Any()).AnyTimes().Return(&types.Post{}, &types.PostMetadata{}, nil)
 	tab.mclock.EXPECT().AwaitLayer(gomock.Any()).AnyTimes()
 
 	// Set expectations
@@ -407,7 +410,7 @@ func TestBuilder_StartSmeshing_SessionNotStartedOnFailPrepare(t *testing.T) {
 	tab.log = l
 
 	// Stub these methods in case they get called
-	tab.mpost.EXPECT().GenerateProof(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes().Return(&types.Post{}, &types.PostMetadata{}, nil)
+	tab.mpostClient.EXPECT().Proof(gomock.Any(), gomock.Any()).AnyTimes().Return(&types.Post{}, &types.PostMetadata{}, nil)
 	tab.mclock.EXPECT().AwaitLayer(gomock.Any()).AnyTimes()
 
 	// Set PrepareInitializer to fail
@@ -430,7 +433,7 @@ func TestBuilder_StopSmeshing_OnPoSTError(t *testing.T) {
 	tab.mpost.EXPECT().StartSession(gomock.Any()).Return(nil).AnyTimes()
 	tab.mpost.EXPECT().CommitmentAtx().Return(types.EmptyATXID, nil).AnyTimes()
 	tab.mpost.EXPECT().LastOpts().Return(&PostSetupOpts{}).AnyTimes()
-	tab.mpost.EXPECT().GenerateProof(gomock.Any(), gomock.Any(), gomock.Any()).Return(&types.Post{}, &types.PostMetadata{}, nil).AnyTimes()
+	tab.mpostClient.EXPECT().Proof(gomock.Any(), gomock.Any()).AnyTimes().Return(&types.Post{}, &types.PostMetadata{}, nil)
 	tab.mValidator.EXPECT().Post(gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes().Return(nil)
 	ch := make(chan struct{})
 	close(ch)
@@ -1089,7 +1092,7 @@ func TestBuilder_RetryPublishActivationTx(t *testing.T) {
 
 func TestBuilder_InitialProofGeneratedOnce(t *testing.T) {
 	tab := newTestBuilder(t, WithPoetConfig(PoetConfig{PhaseShift: layerDuration * 4}))
-	tab.mpost.EXPECT().GenerateProof(gomock.Any(), shared.ZeroChallenge, gomock.Any()).Return(&types.Post{}, &types.PostMetadata{}, nil)
+	tab.mpostClient.EXPECT().Proof(gomock.Any(), shared.ZeroChallenge).Return(&types.Post{}, &types.PostMetadata{}, nil)
 	tab.mpost.EXPECT().LastOpts().Return(&PostSetupOpts{})
 	tab.mpost.EXPECT().CommitmentAtx().Return(tab.goldenATXID, nil)
 	tab.mValidator.EXPECT().Post(gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes().Return(nil)
@@ -1121,7 +1124,7 @@ func TestBuilder_InitialPostIsPersisted(t *testing.T) {
 	tab.mpost.EXPECT().Config().AnyTimes().Return(PostConfig{})
 	tab.mpost.EXPECT().LastOpts().Return(&PostSetupOpts{}).AnyTimes()
 	tab.mpost.EXPECT().CommitmentAtx().Return(tab.goldenATXID, nil).Times(3)
-	tab.mpost.EXPECT().GenerateProof(gomock.Any(), shared.ZeroChallenge, gomock.Any()).Return(&types.Post{}, &types.PostMetadata{
+	tab.mpostClient.EXPECT().Proof(gomock.Any(), shared.ZeroChallenge).Return(&types.Post{}, &types.PostMetadata{
 		Challenge: shared.ZeroChallenge,
 	}, nil)
 	tab.mValidator.EXPECT().Post(gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes().Return(nil)
@@ -1132,7 +1135,7 @@ func TestBuilder_InitialPostIsPersisted(t *testing.T) {
 
 	// Remove the persisted post file and try again
 	require.NoError(t, os.Remove(filepath.Join(tab.nipostBuilder.DataDir(), postFilename)))
-	tab.mpost.EXPECT().GenerateProof(gomock.Any(), shared.ZeroChallenge, gomock.Any()).Return(&types.Post{}, &types.PostMetadata{}, nil)
+	tab.mpostClient.EXPECT().Proof(gomock.Any(), shared.ZeroChallenge).Return(&types.Post{}, &types.PostMetadata{}, nil)
 	require.NoError(t, tab.generateInitialPost(context.Background()))
 }