syz-cluster: collect information about base crashes

Track base crashes for (commit hash, config, arch) tuples.

9 files changed, 272 insertions, 10 deletions

diff --git a/syz-cluster/pkg/api/client.go b/syz-cluster/pkg/api/client.go
index 1b571c628..bd9e4466e 100644
--- a/syz-cluster/pkg/api/client.go
+++ b/syz-cluster/pkg/api/client.go
@@ -114,6 +114,24 @@ func (client Client) UploadSession(ctx context.Context, req *NewSession) (*Uploa
 	return postJSON[NewSession, UploadSessionResp](ctx, client.baseURL+"/sessions/upload", req)
 }
 
+type BaseFindingInfo struct {
+	BuildID string `json:"buildID"`
+	Title   string `json:"title"`
+}
+
+func (client Client) UploadBaseFinding(ctx context.Context, req *BaseFindingInfo) error {
+	_, err := postJSON[BaseFindingInfo, any](ctx, client.baseURL+"/base_findings/upload", req)
+	return err
+}
+
+type BaseFindingStatus struct {
+	Observed bool `json:"observed"`
+}
+
+func (client Client) BaseFindingStatus(ctx context.Context, req *BaseFindingInfo) (*BaseFindingStatus, error) {
+	return postJSON[BaseFindingInfo, BaseFindingStatus](ctx, client.baseURL+"/base_findings/status", req)
+}
+
 const requestTimeout = time.Minute
 
 func finishRequest[Resp any](httpReq *http.Request) (*Resp, error) {
diff --git a/syz-cluster/pkg/controller/api.go b/syz-cluster/pkg/controller/api.go
index de8545319..4a230a35d 100644
--- a/syz-cluster/pkg/controller/api.go
+++ b/syz-cluster/pkg/controller/api.go
@@ -16,20 +16,22 @@ import (
 )
 
 type APIServer struct {
-	seriesService  *service.SeriesService
-	sessionService *service.SessionService
-	buildService   *service.BuildService
-	testService    *service.SessionTestService
-	findingService *service.FindingService
+	seriesService      *service.SeriesService
+	sessionService     *service.SessionService
+	buildService       *service.BuildService
+	testService        *service.SessionTestService
+	findingService     *service.FindingService
+	baseFindingService *service.BaseFindingService
 }
 
 func NewAPIServer(env *app.AppEnvironment) *APIServer {
 	return &APIServer{
-		seriesService:  service.NewSeriesService(env),
-		sessionService: service.NewSessionService(env),
-		buildService:   service.NewBuildService(env),
-		testService:    service.NewSessionTestService(env),
-		findingService: service.NewFindingService(env),
+		seriesService:      service.NewSeriesService(env),
+		sessionService:     service.NewSessionService(env),
+		buildService:       service.NewBuildService(env),
+		testService:        service.NewSessionTestService(env),
+		findingService:     service.NewFindingService(env),
+		baseFindingService: service.NewBaseFindingService(env),
 	}
 }
 
@@ -46,6 +48,8 @@ func (c APIServer) Mux() *http.ServeMux {
 	mux.HandleFunc("/tests/upload_artifacts", c.uploadTestArtifact)
 	mux.HandleFunc("/tests/upload", c.uploadTest)
 	mux.HandleFunc("/trees", c.getTrees)
+	mux.HandleFunc("/base_findings/upload", c.uploadBaseFinding)
+	mux.HandleFunc("/base_findings/status", c.baseFindingStatus)
 	return mux
 }
 
@@ -206,3 +210,32 @@ func (c APIServer) getTrees(w http.ResponseWriter, r *http.Request) {
 		FuzzConfigs: api.FuzzConfigs,
 	})
 }
+
+func (c APIServer) uploadBaseFinding(w http.ResponseWriter, r *http.Request) {
+	req := api.ParseJSON[api.BaseFindingInfo](w, r)
+	if req == nil {
+		return
+	}
+	err := c.baseFindingService.Upload(r.Context(), req)
+	if errors.Is(err, service.ErrBuildNotFound) {
+		http.Error(w, fmt.Sprint(err), http.StatusNotFound)
+		return
+	} else if err != nil {
+		http.Error(w, fmt.Sprint(err), http.StatusInternalServerError)
+		return
+	}
+	api.ReplyJSON[interface{}](w, nil)
+}
+
+func (c APIServer) baseFindingStatus(w http.ResponseWriter, r *http.Request) {
+	req := api.ParseJSON[api.BaseFindingInfo](w, r)
+	if req == nil {
+		return
+	}
+	resp, err := c.baseFindingService.Status(r.Context(), req)
+	if err != nil {
+		http.Error(w, fmt.Sprint(err), http.StatusInternalServerError)
+		return
+	}
+	api.ReplyJSON[*api.BaseFindingStatus](w, resp)
+}
diff --git a/syz-cluster/pkg/controller/api_test.go b/syz-cluster/pkg/controller/api_test.go
index e64f3bc78..69dd02826 100644
--- a/syz-cluster/pkg/controller/api_test.go
+++ b/syz-cluster/pkg/controller/api_test.go
@@ -139,6 +139,46 @@ func TestAPIUploadTestArtifacts(t *testing.T) {
 	assert.NoError(t, err)
 }
 
+func TestAPIBaseFindings(t *testing.T) {
+	env, ctx := app.TestEnvironment(t)
+	client := TestServer(t, env)
+	buildResp := UploadTestBuild(t, ctx, client, testBuild)
+
+	err := client.UploadBaseFinding(ctx, &api.BaseFindingInfo{
+		BuildID: buildResp.ID,
+		Title:   "title 1",
+	})
+	assert.NoError(t, err)
+
+	// Let's upload a different build for the same revision.
+	buildResp2 := UploadTestBuild(t, ctx, client, testBuild)
+	assert.NotEqual(t, buildResp.ID, buildResp2.ID)
+
+	resp, err := client.BaseFindingStatus(ctx, &api.BaseFindingInfo{
+		BuildID: buildResp2.ID,
+		Title:   "title 1",
+	})
+	assert.NoError(t, err)
+	assert.True(t, resp.Observed)
+
+	t.Run("unseen title", func(t *testing.T) {
+		resp, err := client.BaseFindingStatus(ctx, &api.BaseFindingInfo{
+			BuildID: buildResp2.ID,
+			Title:   "title 2",
+		})
+		assert.NoError(t, err)
+		assert.False(t, resp.Observed)
+	})
+
+	t.Run("invalid build id", func(t *testing.T) {
+		_, err := client.BaseFindingStatus(ctx, &api.BaseFindingInfo{
+			BuildID: "unknown id",
+			Title:   "title 1",
+		})
+		assert.Error(t, err)
+	})
+}
+
 var testSeries = &api.Series{
 	ExtID:       "ext-id",
 	AuthorEmail: "some@email.com",
diff --git a/syz-cluster/pkg/db/base_finding_repo.go b/syz-cluster/pkg/db/base_finding_repo.go
new file mode 100644
index 000000000..e963cb691
--- /dev/null
+++ b/syz-cluster/pkg/db/base_finding_repo.go
@@ -0,0 +1,49 @@
+// Copyright 2025 syzkaller project authors. All rights reserved.
+// Use of this source code is governed by Apache 2 LICENSE that can be found in the LICENSE file.
+
+package db
+
+import (
+	"context"
+
+	"cloud.google.com/go/spanner"
+)
+
+type BaseFindingRepository struct {
+	client *spanner.Client
+}
+
+func NewBaseFindingRepository(client *spanner.Client) *BaseFindingRepository {
+	return &BaseFindingRepository{
+		client: client,
+	}
+}
+
+func (repo *BaseFindingRepository) Save(ctx context.Context, info *BaseFinding) error {
+	_, err := repo.client.ReadWriteTransaction(ctx,
+		func(ctx context.Context, txn *spanner.ReadWriteTransaction) error {
+			m, err := spanner.InsertOrUpdateStruct("BaseFindings", info)
+			if err != nil {
+				return err
+			}
+			return txn.BufferWrite([]*spanner.Mutation{m})
+		})
+	return err
+}
+
+func (repo *BaseFindingRepository) Exists(ctx context.Context, info *BaseFinding) (bool, error) {
+	entity, err := readEntity[BaseFinding](ctx, repo.client.Single(), spanner.Statement{
+		SQL: `SELECT * FROM BaseFindings WHERE
+CommitHash = @commit AND
+Config = @config AND
+Arch = @arch AND
+Title = @title`,
+		Params: map[string]interface{}{
+			"commit": info.CommitHash,
+			"config": info.Config,
+			"arch":   info.Arch,
+			"title":  info.Title,
+		},
+	})
+	return entity != nil, err
+}
diff --git a/syz-cluster/pkg/db/base_finding_repo_test.go b/syz-cluster/pkg/db/base_finding_repo_test.go
new file mode 100644
index 000000000..de7eb07b6
--- /dev/null
+++ b/syz-cluster/pkg/db/base_finding_repo_test.go
@@ -0,0 +1,40 @@
+// Copyright 2025 syzkaller project authors. All rights reserved.
+// Use of this source code is governed by Apache 2 LICENSE that can be found in the LICENSE file.
+
+package db
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestBaseFindingRepository(t *testing.T) {
+	client, ctx := NewTransientDB(t)
+	repo := NewBaseFindingRepository(client)
+
+	// It works fine on unknown titles.
+	exists, err := repo.Exists(ctx, &BaseFinding{
+		CommitHash: "abcd",
+		Config:     "cfg",
+		Arch:       "x86",
+	})
+	require.NoError(t, err)
+	assert.False(t, exists)
+
+	// Add some new title.
+	finding := &BaseFinding{
+		CommitHash: "hash",
+		Config:     "config",
+		Arch:       "arch",
+		Title:      "title",
+	}
+	err = repo.Save(ctx, finding)
+	require.NoError(t, err)
+
+	// Verify it exists.
+	exists, err = repo.Exists(ctx, finding)
+	require.NoError(t, err)
+	assert.True(t, exists)
+}
diff --git a/syz-cluster/pkg/db/entities.go b/syz-cluster/pkg/db/entities.go
index 53e52b4ea..13ba2fb85 100644
--- a/syz-cluster/pkg/db/entities.go
+++ b/syz-cluster/pkg/db/entities.go
@@ -161,3 +161,12 @@ type ReportReply struct {
 	ReportID  string    `spanner:"ReportID"`
 	Time      time.Time `spanner:"Time"`
 }
+
+// BaseFinding collects all crashes observed on the base kernel tree.
+// It will be used to avoid unnecessary bug reproduction attempts.
+type BaseFinding struct {
+	CommitHash string `spanner:"CommitHash"`
+	Config     string `spanner:"Config"`
+	Arch       string `spanner:"Arch"`
+	Title      string `spanner:"Title"`
+}
diff --git a/syz-cluster/pkg/db/migrations/6_base_findings.down.sql b/syz-cluster/pkg/db/migrations/6_base_findings.down.sql
new file mode 100644
index 000000000..c30ac67a0
--- /dev/null
+++ b/syz-cluster/pkg/db/migrations/6_base_findings.down.sql
@@ -0,0 +1 @@
+DROP TABLE BaseFindings;
diff --git a/syz-cluster/pkg/db/migrations/6_base_findings.up.sql b/syz-cluster/pkg/db/migrations/6_base_findings.up.sql
new file mode 100644
index 000000000..8ec8f245a
--- /dev/null
+++ b/syz-cluster/pkg/db/migrations/6_base_findings.up.sql
@@ -0,0 +1,6 @@
+CREATE TABLE BaseFindings (
+    CommitHash STRING(64) NOT NULL,
+    Config STRING(256) NOT NULL,
+    Arch STRING(64) NOT NULL,
+    Title STRING(512) NOT NULL,
+) PRIMARY KEY (CommitHash, Config, Arch, Title);
diff --git a/syz-cluster/pkg/service/base_finding.go b/syz-cluster/pkg/service/base_finding.go
new file mode 100644
index 000000000..e7b59c18c
--- /dev/null
+++ b/syz-cluster/pkg/service/base_finding.go
@@ -0,0 +1,66 @@
+// Copyright 2025 syzkaller project authors. All rights reserved.
+// Use of this source code is governed by Apache 2 LICENSE that can be found in the LICENSE file.
+
+package service
+
+import (
+	"context"
+	"errors"
+	"fmt"
+
+	"github.com/google/syzkaller/syz-cluster/pkg/api"
+	"github.com/google/syzkaller/syz-cluster/pkg/app"
+	"github.com/google/syzkaller/syz-cluster/pkg/db"
+)
+
+type BaseFindingService struct {
+	baseFindingRepo *db.BaseFindingRepository
+	buildRepo       *db.BuildRepository
+}
+
+func NewBaseFindingService(env *app.AppEnvironment) *BaseFindingService {
+	return &BaseFindingService{
+		baseFindingRepo: db.NewBaseFindingRepository(env.Spanner),
+		buildRepo:       db.NewBuildRepository(env.Spanner),
+	}
+}
+
+var ErrBuildNotFound = errors.New("build not found")
+
+func (s *BaseFindingService) Upload(ctx context.Context, info *api.BaseFindingInfo) error {
+	finding, err := s.makeBaseFinding(ctx, info)
+	if err != nil {
+		return err
+	}
+	return s.baseFindingRepo.Save(ctx, finding)
+}
+
+func (s *BaseFindingService) Status(ctx context.Context, info *api.BaseFindingInfo) (
+	*api.BaseFindingStatus, error) {
+	finding, err := s.makeBaseFinding(ctx, info)
+	if err != nil {
+		return nil, err
+	}
+	exists, err := s.baseFindingRepo.Exists(ctx, finding)
+	if err != nil {
+		return nil, err
+	}
+	return &api.BaseFindingStatus{
+		Observed: exists,
+	}, nil
+}
+
+func (s *BaseFindingService) makeBaseFinding(ctx context.Context, info *api.BaseFindingInfo) (*db.BaseFinding, error) {
+	build, err := s.buildRepo.GetByID(ctx, info.BuildID)
+	if err != nil {
+		return nil, fmt.Errorf("failed to query build: %w", err)
+	} else if build == nil {
+		return nil, ErrBuildNotFound
+	}
+	return &db.BaseFinding{
+		CommitHash: build.CommitHash,
+		Config:     build.ConfigName,
+		Arch:       build.Arch,
+		Title:      info.Title,
+	}, nil
+}