Merge pull request from GHSA-55r9-5mx9-qq7r

when a client pushes an image zot's inline dedupe
will try to find the blob path corresponding with the blob digest
that it's currently pushed and if it's found in the cache
then zot will make a symbolic link to that cache entry and report
to the client that the blob already exists on the location.

Before this patch authorization was not applied on this process meaning
that a user could copy blobs without having permissions on the source repo.

Added a rule which says that the client should have read permissions on the source repo
before deduping, otherwise just Stat() the blob and return the corresponding status code.

Signed-off-by: Petu Eusebiu <peusebiu@cisco.com>
Co-authored-by: Petu Eusebiu <peusebiu@cisco.com>

pkg/api/controller_test.go

@@ -5142,6 +5142,122 @@ func TestGetUsername(t *testing.T) {
 	})
 }
 
+func TestAuthorizationMountBlob(t *testing.T) {
+	Convey("Make a new controller", t, func() {
+		port := test.GetFreePort()
+		baseURL := test.GetBaseURL(port)
+
+		conf := config.New()
+		conf.HTTP.Port = port
+		// have two users: one for  user Policy, and another for default policy
+		username1, _ := test.GenerateRandomString()
+		password1, _ := test.GenerateRandomString()
+		username2, _ := test.GenerateRandomString()
+		password2, _ := test.GenerateRandomString()
+		username1 = strings.ToLower(username1)
+		username2 = strings.ToLower(username2)
+
+		content := test.GetCredString(username1, password1) + test.GetCredString(username2, password2)
+		htpasswdPath := test.MakeHtpasswdFileFromString(content)
+		defer os.Remove(htpasswdPath)
+
+		conf.HTTP.Auth = &config.AuthConfig{
+			HTPasswd: config.AuthHTPasswd{
+				Path: htpasswdPath,
+			},
+		}
+
+		user1Repo := fmt.Sprintf("%s/**", username1)
+		user2Repo := fmt.Sprintf("%s/**", username2)
+
+		// config with all policy types, to test that the correct one is applied in each case
+		conf.HTTP.AccessControl = &config.AccessControlConfig{
+			Repositories: config.Repositories{
+				user1Repo: config.PolicyGroup{
+					Policies: []config.Policy{
+						{
+							Users: []string{username1},
+							Actions: []string{
+								constants.ReadPermission,
+								constants.CreatePermission,
+							},
+						},
+					},
+				},
+				user2Repo: config.PolicyGroup{
+					Policies: []config.Policy{
+						{
+							Users: []string{username2},
+							Actions: []string{
+								constants.ReadPermission,
+								constants.CreatePermission,
+							},
+						},
+					},
+				},
+			},
+		}
+
+		dir := t.TempDir()
+
+		ctlr := api.NewController(conf)
+		ctlr.Config.Storage.RootDirectory = dir
+
+		cm := test.NewControllerManager(ctlr)
+		cm.StartAndWait(port)
+		defer cm.StopServer()
+
+		userClient1 := resty.New()
+		userClient1.SetBasicAuth(username1, password1)
+
+		userClient2 := resty.New()
+		userClient2.SetBasicAuth(username2, password2)
+
+		img := CreateImageWith().RandomLayers(1, 2).DefaultConfig().Build()
+
+		repoName1 := username1 + "/" + "myrepo"
+		tag := "1.0"
+
+		// upload image with user1 on repoName1
+		err := UploadImageWithBasicAuth(img, baseURL, repoName1, tag, username1, password1)
+		So(err, ShouldBeNil)
+
+		repoName2 := username2 + "/" + "myrepo"
+
+		blobDigest := img.Manifest.Layers[0].Digest
+
+		/* a HEAD request by user2 on blob digest (found in user1Repo) should return 404
+		because user2 doesn't have permissions to read user1Repo */
+		resp, err := userClient2.R().Head(baseURL + fmt.Sprintf("/v2/%s/blobs/%s", repoName2, blobDigest))
+		So(err, ShouldBeNil)
+		So(resp.StatusCode(), ShouldEqual, http.StatusNotFound)
+
+		params := make(map[string]string)
+		params["mount"] = blobDigest.String()
+
+		// trying to mount a blob which can be found in cache, but user doesn't have permission
+		// should return 202 instead of 201
+		resp, err = userClient2.R().SetQueryParams(params).Post(baseURL + "/v2/" + repoName2 + "/blobs/uploads/")
+		So(err, ShouldBeNil)
+		So(resp.StatusCode(), ShouldEqual, http.StatusAccepted)
+
+		/* a HEAD request by user1 on blob digest (found in user1Repo) should return 200
+		because user1 has permission to read user1Repo */
+		resp, err = userClient1.R().Head(baseURL + fmt.Sprintf("/v2/%s/blobs/%s", username1+"/"+"mysecondrepo", blobDigest))
+		So(err, ShouldBeNil)
+		So(resp.StatusCode(), ShouldEqual, http.StatusOK)
+
+		// user2 can upload without dedupe
+		err = UploadImageWithBasicAuth(img, baseURL, repoName2, tag, username2, password2)
+		So(err, ShouldBeNil)
+
+		// trying to mount a blob which can be found in cache and user has permission should return 201 instead of 202
+		resp, err = userClient2.R().SetQueryParams(params).Post(baseURL + "/v2/" + repoName2 + "/blobs/uploads/")
+		So(err, ShouldBeNil)
+		So(resp.StatusCode(), ShouldEqual, http.StatusCreated)
+	})
+}
+
 func TestAuthorizationWithOnlyAnonymousPolicy(t *testing.T) {
 	Convey("Make a new controller", t, func() {
 		const TestRepo = "my-repos/repo"

pkg/api/routes.go

@@ -873,6 +873,37 @@ func (rh *RouteHandler) DeleteManifest(response http.ResponseWriter, request *ht
 	response.WriteHeader(http.StatusAccepted)
 }
 
+// canMount checks if a user has read permission on cached blobs with this specific digest.
+// returns true if the user have permission to copy blob from cache.
+func canMount(userAc *reqCtx.UserAccessControl, imgStore storageTypes.ImageStore, digest godigest.Digest,
+) (bool, error) {
+	canMount := true
+
+	// authz enabled
+	if userAc != nil {
+		canMount = false
+
+		repos, err := imgStore.GetAllDedupeReposCandidates(digest)
+		if err != nil {
+			// first write
+			return false, err
+		}
+
+		if len(repos) == 0 {
+			canMount = false
+		}
+
+		// check if user can read any repo which contain this blob
+		for _, repo := range repos {
+			if userAc.Can(constants.ReadPermission, repo) {
+				canMount = true
+			}
+		}
+	}
+
+	return canMount, nil
+}
+
 // CheckBlob godoc
 // @Summary Check image blob/layer
 // @Description Check an image's blob/layer given a digest
@@ -905,7 +936,31 @@ func (rh *RouteHandler) CheckBlob(response http.ResponseWriter, request *http.Re
 
 	digest := godigest.Digest(digestStr)
 
-	ok, blen, err := imgStore.CheckBlob(name, digest)
+	userAc, err := reqCtx.UserAcFromContext(request.Context())
+	if err != nil {
+		response.WriteHeader(http.StatusInternalServerError)
+
+		return
+	}
+
+	userCanMount, err := canMount(userAc, imgStore, digest)
+	if err != nil {
+		rh.c.Log.Error().Err(err).Msg("unexpected error")
+	}
+
+	var blen int64
+
+	if userCanMount {
+		ok, blen, err = imgStore.CheckBlob(name, digest)
+	} else {
+		var lockLatency time.Time
+
+		imgStore.RLock(&lockLatency)
+		defer imgStore.RUnlock(&lockLatency)
+
+		ok, blen, _, err = imgStore.StatBlob(name, digest)
+	}
+
 	if err != nil {
 		details := zerr.GetDetails(err)
 		if errors.Is(err, zerr.ErrBadBlobDigest) { //nolint:gocritic // errorslint conflicts with gocritic:IfElseChain
@@ -1191,11 +1246,27 @@ func (rh *RouteHandler) CreateBlobUpload(response http.ResponseWriter, request *
 		}
 
 		mountDigest := godigest.Digest(mountDigests[0])
+
+		userAc, err := reqCtx.UserAcFromContext(request.Context())
+		if err != nil {
+			response.WriteHeader(http.StatusInternalServerError)
+
+			return
+		}
+
+		userCanMount, err := canMount(userAc, imgStore, mountDigest)
+		if err != nil {
+			rh.c.Log.Error().Err(err).Msg("unexpected error")
+		}
+
 		// zot does not support cross mounting directly and do a workaround creating using hard link.
 		// check blob looks for actual path (name+mountDigests[0]) first then look for cache and
 		// if found in cache, will do hard link and if fails we will start new upload.
-		_, _, err := imgStore.CheckBlob(name, mountDigest)
-		if err != nil {
+		if userCanMount {
+			_, _, err = imgStore.CheckBlob(name, mountDigest)
+		}
+
+		if err != nil || !userCanMount {
 			upload, err := imgStore.NewBlobUpload(name)
 			if err != nil {
 				details := zerr.GetDetails(err)

pkg/storage/cache/boltdb.go

@@ -174,6 +174,57 @@ func (d *BoltDBDriver) PutBlob(digest godigest.Digest, path string) error {
 	return nil
 }
 
+func (d *BoltDBDriver) GetAllBlobs(digest godigest.Digest) ([]string, error) {
+	var blobPath strings.Builder
+
+	blobPaths := []string{}
+
+	if err := d.db.View(func(tx *bbolt.Tx) error {
+		root := tx.Bucket([]byte(constants.BlobsCache))
+		if root == nil {
+			// this is a serious failure
+			err := zerr.ErrCacheRootBucket
+			d.log.Error().Err(err).Msg("failed to access root bucket")
+
+			return err
+		}
+
+		bucket := root.Bucket([]byte(digest.String()))
+		if bucket != nil {
+			origin := bucket.Bucket([]byte(constants.OriginalBucket))
+			blobPath.Write(d.getOne(origin))
+			originBlob := blobPath.String()
+
+			blobPaths = append(blobPaths, originBlob)
+
+			deduped := bucket.Bucket([]byte(constants.DuplicatesBucket))
+			if deduped != nil {
+				cursor := deduped.Cursor()
+
+				for k, _ := cursor.First(); k != nil; k, _ = cursor.Next() {
+					var blobPath strings.Builder
+
+					blobPath.Write(k)
+
+					duplicateBlob := blobPath.String()
+
+					if duplicateBlob != originBlob {
+						blobPaths = append(blobPaths, duplicateBlob)
+					}
+				}
+
+				return nil
+			}
+		}
+
+		return zerr.ErrCacheMiss
+	}); err != nil {
+		return nil, err
+	}
+
+	return blobPaths, nil
+}
+
 func (d *BoltDBDriver) GetBlob(digest godigest.Digest) (string, error) {
 	var blobPath strings.Builder
 

pkg/storage/cache/boltdb_test.go

@@ -122,4 +122,47 @@ func TestBoltDBCache(t *testing.T) {
 		So(err, ShouldNotBeNil)
 		So(val, ShouldBeEmpty)
 	})
+
+	Convey("Test cache.GetAllBlos()", t, func() {
+		dir := t.TempDir()
+
+		log := log.NewLogger("debug", "")
+		So(log, ShouldNotBeNil)
+
+		_, err := storage.Create("boltdb", "failTypeAssertion", log)
+		So(err, ShouldNotBeNil)
+
+		cacheDriver, _ := storage.Create("boltdb", cache.BoltDBDriverParameters{dir, "cache_test", false}, log)
+		So(cacheDriver, ShouldNotBeNil)
+
+		err = cacheDriver.PutBlob("digest", "first")
+		So(err, ShouldBeNil)
+
+		err = cacheDriver.PutBlob("digest", "second")
+		So(err, ShouldBeNil)
+
+		err = cacheDriver.PutBlob("digest", "third")
+		So(err, ShouldBeNil)
+
+		blobs, err := cacheDriver.GetAllBlobs("digest")
+		So(err, ShouldBeNil)
+
+		So(blobs, ShouldResemble, []string{"first", "second", "third"})
+
+		err = cacheDriver.DeleteBlob("digest", "first")
+		So(err, ShouldBeNil)
+
+		blobs, err = cacheDriver.GetAllBlobs("digest")
+		So(err, ShouldBeNil)
+
+		So(blobs, ShouldResemble, []string{"second", "third"})
+
+		err = cacheDriver.DeleteBlob("digest", "third")
+		So(err, ShouldBeNil)
+
+		blobs, err = cacheDriver.GetAllBlobs("digest")
+		So(err, ShouldBeNil)
+
+		So(blobs, ShouldResemble, []string{"second"})
+	})
 }

pkg/storage/cache/cacheinterface.go

@@ -11,6 +11,8 @@ type Cache interface {
 	// Retrieves the blob matching provided digest.
 	GetBlob(digest godigest.Digest) (string, error)
 
+	GetAllBlobs(digest godigest.Digest) ([]string, error)
+
 	// Uploads blob to cachedb.
 	PutBlob(digest godigest.Digest, path string) error
 

pkg/storage/cache/dynamodb.go

@@ -141,6 +141,42 @@ func (d *DynamoDBDriver) GetBlob(digest godigest.Digest) (string, error) {
 	return out.OriginalBlobPath, nil
 }
 
+func (d *DynamoDBDriver) GetAllBlobs(digest godigest.Digest) ([]string, error) {
+	blobPaths := []string{}
+
+	resp, err := d.client.GetItem(context.TODO(), &dynamodb.GetItemInput{
+		TableName: aws.String(d.tableName),
+		Key: map[string]types.AttributeValue{
+			"Digest": &types.AttributeValueMemberS{Value: digest.String()},
+		},
+	})
+	if err != nil {
+		d.log.Error().Err(err).Str("tableName", d.tableName).Msg("failed to get blob")
+
+		return nil, err
+	}
+
+	out := Blob{}
+
+	if resp.Item == nil {
+		d.log.Debug().Err(zerr.ErrCacheMiss).Str("digest", string(digest)).Msg("failed to find blob in cache")
+
+		return nil, zerr.ErrCacheMiss
+	}
+
+	_ = attributevalue.UnmarshalMap(resp.Item, &out)
+
+	blobPaths = append(blobPaths, out.OriginalBlobPath)
+
+	for _, item := range out.DuplicateBlobPath {
+		if item != out.OriginalBlobPath {
+			blobPaths = append(blobPaths, item)
+		}
+	}
+
+	return blobPaths, nil
+}
+
 func (d *DynamoDBDriver) PutBlob(digest godigest.Digest, path string) error {
 	if path == "" {
 		d.log.Error().Err(zerr.ErrEmptyValue).Str("digest", digest.String()).

pkg/storage/cache/dynamodb_test.go

@@ -144,6 +144,48 @@ func TestDynamoDB(t *testing.T) {
 		So(err, ShouldNotBeNil)
 		So(val, ShouldBeEmpty)
 	})
+
+	Convey("Test dynamoDB", t, func(c C) {
+		log := log.NewLogger("debug", "")
+
+		cacheDriver, err := storage.Create("dynamodb", cache.DynamoDBDriverParameters{
+			Endpoint:  os.Getenv("DYNAMODBMOCK_ENDPOINT"),
+			TableName: "BlobTable",
+			Region:    "us-east-2",
+		}, log)
+		So(cacheDriver, ShouldNotBeNil)
+		So(err, ShouldBeNil)
+
+		err = cacheDriver.PutBlob("digest", "first")
+		So(err, ShouldBeNil)
+
+		err = cacheDriver.PutBlob("digest", "second")
+		So(err, ShouldBeNil)
+
+		err = cacheDriver.PutBlob("digest", "third")
+		So(err, ShouldBeNil)
+
+		blobs, err := cacheDriver.GetAllBlobs("digest")
+		So(err, ShouldBeNil)
+
+		So(blobs, ShouldResemble, []string{"first", "second", "third"})
+
+		err = cacheDriver.DeleteBlob("digest", "first")
+		So(err, ShouldBeNil)
+
+		blobs, err = cacheDriver.GetAllBlobs("digest")
+		So(err, ShouldBeNil)
+
+		So(blobs, ShouldResemble, []string{"second", "third"})
+
+		err = cacheDriver.DeleteBlob("digest", "third")
+		So(err, ShouldBeNil)
+
+		blobs, err = cacheDriver.GetAllBlobs("digest")
+		So(err, ShouldBeNil)
+
+		So(blobs, ShouldResemble, []string{"second"})
+	})
 }
 
 func TestDynamoDBError(t *testing.T) {

pkg/storage/imagestore/imagestore.go

@@ -1114,6 +1114,37 @@ func (is *ImageStore) BlobPath(repo string, digest godigest.Digest) string {
 	return path.Join(is.rootDir, repo, "blobs", digest.Algorithm().String(), digest.Encoded())
 }
 
+func (is *ImageStore) GetAllDedupeReposCandidates(digest godigest.Digest) ([]string, error) {
+	var lockLatency time.Time
+
+	if err := digest.Validate(); err != nil {
+		return nil, err
+	}
+
+	is.RLock(&lockLatency)
+	defer is.RUnlock(&lockLatency)
+
+	blobsPaths, err := is.cache.GetAllBlobs(digest)
+	if err != nil {
+		return nil, err
+	}
+
+	repos := []string{}
+
+	for _, blobPath := range blobsPaths {
+		// these can be both full paths or relative paths depending on the cache options
+		if !is.cache.UsesRelativePaths() && path.IsAbs(blobPath) {
+			blobPath, _ = filepath.Rel(is.rootDir, blobPath)
+		}
+
+		blobsDirIndex := strings.LastIndex(blobPath, "/blobs/")
+
+		repos = append(repos, blobPath[:blobsDirIndex])
+	}
+
+	return repos, nil
+}
+
 /*
 	CheckBlob verifies a blob and returns true if the blob is correct
 

pkg/storage/storage_test.go

@@ -10,6 +10,7 @@ import (
 	"io"
 	"os"
 	"path"
+	"slices"
 	"strings"
 	"sync"
 	"testing"
@@ -132,6 +133,75 @@ func TestStorageNew(t *testing.T) {
 	})
 }
 
+func TestGetAllDedupeReposCandidates(t *testing.T) {
+	for _, testcase := range testCases {
+		testcase := testcase
+		t.Run(testcase.testCaseName, func(t *testing.T) {
+			var imgStore storageTypes.ImageStore
+			if testcase.storageType == storageConstants.S3StorageDriverName {
+				tskip.SkipS3(t)
+
+				uuid, err := guuid.NewV4()
+				if err != nil {
+					panic(err)
+				}
+
+				testDir := path.Join("/oci-repo-test", uuid.String())
+				tdir := t.TempDir()
+
+				var store driver.StorageDriver
+				store, imgStore, _ = createObjectsStore(testDir, tdir)
+				defer cleanupStorage(store, testDir)
+			} else {
+				dir := t.TempDir()
+
+				log := zlog.Logger{Logger: zerolog.New(os.Stdout)}
+				metrics := monitoring.NewMetricsServer(false, log)
+				cacheDriver, _ := storage.Create("boltdb", cache.BoltDBDriverParameters{
+					RootDir:     dir,
+					Name:        "cache",
+					UseRelPaths: true,
+				}, log)
+
+				driver := local.New(true)
+
+				imgStore = imagestore.NewImageStore(dir, dir, true, true, log, metrics, nil, driver, cacheDriver)
+			}
+
+			Convey("Push repos with deduped blobs", t, func(c C) {
+				repoNames := []string{
+					"first",
+					"second",
+					"repo/a",
+					"repo/a/b/c/d/e/f",
+					"repo/repo-b/blobs",
+					"foo/bar/baz",
+					"blobs/foo/bar/blobs",
+					"blobs",
+					"blobs/foo",
+				}
+
+				storeController := storage.StoreController{DefaultStore: imgStore}
+
+				image := CreateRandomImage()
+
+				for _, repoName := range repoNames {
+					err := WriteImageToFileSystem(image, repoName, tag, storeController)
+					So(err, ShouldBeNil)
+				}
+
+				randomBlobDigest := image.Manifest.Layers[0].Digest
+
+				repos, err := imgStore.GetAllDedupeReposCandidates(randomBlobDigest)
+				So(err, ShouldBeNil)
+				slices.Sort(repoNames)
+				slices.Sort(repos)
+				So(repoNames, ShouldResemble, repos)
+			})
+		})
+	}
+}
+
 func TestStorageAPIs(t *testing.T) {
 	for _, testcase := range testCases {
 		testcase := testcase

pkg/storage/types/types.go

@@ -63,6 +63,7 @@ type ImageStore interface { //nolint:interfacebloat
 	GetAllBlobs(repo string) ([]string, error)
 	PopulateStorageMetrics(interval time.Duration, sch *scheduler.Scheduler)
 	VerifyBlobDigestValue(repo string, digest godigest.Digest) error
+	GetAllDedupeReposCandidates(digest godigest.Digest) ([]string, error)
 }
 
 type Driver interface { //nolint:interfacebloat

pkg/test/mocks/cache_mock.go

@@ -6,6 +6,8 @@ type CacheMock struct {
 	// Returns the human-readable "name" of the driver.
 	NameFn func() string
 
+	GetAllBlobsFn func(digest godigest.Digest) ([]string, error)
+
 	// Retrieves the blob matching provided digest.
 	GetBlobFn func(digest godigest.Digest) (string, error)
 
@@ -68,3 +70,11 @@ func (cacheMock CacheMock) DeleteBlob(digest godigest.Digest, path string) error
 
 	return nil
 }
+
+func (cacheMock CacheMock) GetAllBlobs(digest godigest.Digest) ([]string, error) {
+	if cacheMock.GetAllBlobsFn != nil {
+		return cacheMock.GetAllBlobsFn(digest)
+	}
+
+	return []string{}, nil
+}

pkg/test/mocks/image_store_mock.go

@@ -57,6 +57,7 @@ type MockedImageStore struct {
 	PopulateStorageMetricsFn      func(interval time.Duration, sch *scheduler.Scheduler)
 	StatIndexFn                   func(repo string) (bool, int64, time.Time, error)
 	VerifyBlobDigestValueFn       func(repo string, digest godigest.Digest) error
+	GetAllDedupeReposCandidatesFn func(digest godigest.Digest) ([]string, error)
 }
 
 func (is MockedImageStore) StatIndex(repo string) (bool, int64, time.Time, error) {
@@ -419,3 +420,11 @@ func (is MockedImageStore) VerifyBlobDigestValue(repo string, digest godigest.Di
 
 	return nil
 }
+
+func (is MockedImageStore) GetAllDedupeReposCandidates(digest godigest.Digest) ([]string, error) {
+	if is.GetAllBlobsFn != nil {
+		return is.GetAllDedupeReposCandidatesFn(digest)
+	}
+
+	return []string{}, nil
+}