diff --git a/go.mod b/go.mod
index c60d9d50fc..a69ceab9b5 100644
--- a/go.mod
+++ b/go.mod
@@ -17,7 +17,7 @@ require (
 	github.com/containers/image/v5 v5.23.1-0.20221101011818-2f770d6d5a0c
 	github.com/containers/ocicrypt v1.1.6
 	github.com/containers/psgo v1.8.0
-	github.com/containers/storage v1.43.1-0.20221104122514-74e37433a2a0
+	github.com/containers/storage v1.44.1-0.20221110192950-67e9778710f8
 	github.com/coreos/go-systemd/v22 v22.5.0
 	github.com/coreos/stream-metadata-go v0.0.0-20210225230131-70edb9eb47b3
 	github.com/cyphar/filepath-securejoin v0.2.3
diff --git a/go.sum b/go.sum
index 1ee8a39b12..99c94744cd 100644
--- a/go.sum
+++ b/go.sum
@@ -281,8 +281,8 @@ github.com/containers/psgo v1.8.0 h1:2loGekmGAxM9ir5OsXWEfGwFxorMPYnc6gEDsGFQvhY
 github.com/containers/psgo v1.8.0/go.mod h1:T8ZxnX3Ur4RvnhxFJ7t8xJ1F48RhiZB4rSrOaR/qGHc=
 github.com/containers/storage v1.37.0/go.mod h1:kqeJeS0b7DO2ZT1nVWs0XufrmPFbgV3c+Q/45RlH6r4=
 github.com/containers/storage v1.43.0/go.mod h1:uZ147thiIFGdVTjMmIw19knttQnUCl3y9zjreHrg11s=
-github.com/containers/storage v1.43.1-0.20221104122514-74e37433a2a0 h1:j/+RfR57O5MEr0Irhju/5phfwoU7s3pR0Q5T1dEPeVw=
-github.com/containers/storage v1.43.1-0.20221104122514-74e37433a2a0/go.mod h1:HSfx7vUXwKPatPMqhgMw3mI3c3ijIJPZV5O0sj/mVxI=
+github.com/containers/storage v1.44.1-0.20221110192950-67e9778710f8 h1:MrQjgoKVQpD/16sfYe9C3T3y2gLvfBPADMFQ7Oq93zo=
+github.com/containers/storage v1.44.1-0.20221110192950-67e9778710f8/go.mod h1:HSfx7vUXwKPatPMqhgMw3mI3c3ijIJPZV5O0sj/mVxI=
 github.com/coreos/bbolt v1.3.2/go.mod h1:iRUV2dpdMOn7Bo10OQBFzIJO9kkE559Wcmn+qkEiiKk=
 github.com/coreos/etcd v3.3.10+incompatible/go.mod h1:uF7uidLiAD3TWHmW31ZFd/JWoc32PjwdhPthX9715RE=
 github.com/coreos/go-iptables v0.4.5/go.mod h1:/mVI274lEDI2ns62jHCDnCyBF9Iwsmekav8Dbxlm1MU=
diff --git a/vendor/github.com/containers/storage/VERSION b/vendor/github.com/containers/storage/VERSION
index d5811245b9..5b76dbea5e 100644
--- a/vendor/github.com/containers/storage/VERSION
+++ b/vendor/github.com/containers/storage/VERSION
@@ -1 +1 @@
-1.43.1-dev
+1.44.1-dev
diff --git a/vendor/github.com/containers/storage/containers.go b/vendor/github.com/containers/storage/containers.go
index d3223cd45d..a123b716bd 100644
--- a/vendor/github.com/containers/storage/containers.go
+++ b/vendor/github.com/containers/storage/containers.go
@@ -191,7 +191,7 @@ func (r *containerStore) startWritingWithReload(canReload bool) error {
 	}()
 
 	if canReload {
-		if err := r.reloadIfChanged(true); err != nil {
+		if _, err := r.reloadIfChanged(true); err != nil {
 			return err
 		}
 	}
@@ -215,18 +215,41 @@ func (r *containerStore) stopWriting() {
 // If this succeeds, the caller MUST call stopReading().
 func (r *containerStore) startReading() error {
 	r.lockfile.RLock()
-	succeeded := false
+	unlockFn := r.lockfile.Unlock // A function to call to clean up, or nil
 	defer func() {
-		if !succeeded {
-			r.lockfile.Unlock()
+		if unlockFn != nil {
+			unlockFn()
 		}
 	}()
 
-	if err := r.reloadIfChanged(false); err != nil {
-		return err
+	if tryLockedForWriting, err := r.reloadIfChanged(false); err != nil {
+		if !tryLockedForWriting {
+			return err
+		}
+		unlockFn()
+		unlockFn = nil
+
+		r.lockfile.Lock()
+		unlockFn = r.lockfile.Unlock
+		if _, err := r.load(true); err != nil {
+			return err
+		}
+		unlockFn()
+		unlockFn = nil
+
+		r.lockfile.RLock()
+		unlockFn = r.lockfile.Unlock
+		// We need to check for a reload reload once more because the on-disk state could have been modified
+		// after we released the lock.
+		// If that, _again_, finds inconsistent state, just give up.
+		// We could, plausibly, retry a few times, but that inconsistent state (duplicate container names)
+		// shouldn’t be saved (by correct implementations) in the first place.
+		if _, err := r.reloadIfChanged(false); err != nil {
+			return fmt.Errorf("(even after successfully cleaning up once:) %w", err)
+		}
 	}
 
-	succeeded = true
+	unlockFn = nil
 	return nil
 }
 
@@ -239,15 +262,23 @@ func (r *containerStore) stopReading() {
 //
 // The caller must hold r.lockfile for reading _or_ writing; lockedForWriting is true
 // if it is held for writing.
-func (r *containerStore) reloadIfChanged(lockedForWriting bool) error {
+//
+// If !lockedForWriting and this function fails, the return value indicates whether
+// load() with lockedForWriting could succeed. In that case the caller MUST
+// call load(), not reloadIfChanged() (because the “if changed” state will not
+// be detected again).
+func (r *containerStore) reloadIfChanged(lockedForWriting bool) (bool, error) {
 	r.loadMut.Lock()
 	defer r.loadMut.Unlock()
 
 	modified, err := r.lockfile.Modified()
-	if err == nil && modified {
+	if err != nil {
+		return false, err
+	}
+	if modified {
 		return r.load(lockedForWriting)
 	}
-	return err
+	return false, nil
 }
 
 func (r *containerStore) Containers() ([]Container, error) {
@@ -274,24 +305,27 @@ func (r *containerStore) datapath(id, key string) string {
 //
 // The caller must hold r.lockfile for reading _or_ writing; lockedForWriting is true
 // if it is held for writing.
-func (r *containerStore) load(lockedForWriting bool) error {
-	needSave := false
+//
+// If !lockedForWriting and this function fails, the return value indicates whether
+// retrying with lockedForWriting could succeed.
+func (r *containerStore) load(lockedForWriting bool) (bool, error) {
 	rpath := r.containerspath()
 	data, err := os.ReadFile(rpath)
 	if err != nil && !os.IsNotExist(err) {
-		return err
+		return false, err
 	}
 
 	containers := []*Container{}
 	if len(data) != 0 {
 		if err := json.Unmarshal(data, &containers); err != nil {
-			return fmt.Errorf("loading %q: %w", rpath, err)
+			return false, fmt.Errorf("loading %q: %w", rpath, err)
 		}
 	}
 	idlist := make([]string, 0, len(containers))
 	layers := make(map[string]*Container)
 	ids := make(map[string]*Container)
 	names := make(map[string]*Container)
+	var errorToResolveBySaving error // == nil
 	for n, container := range containers {
 		idlist = append(idlist, container.ID)
 		ids[container.ID] = containers[n]
@@ -299,7 +333,7 @@ func (r *containerStore) load(lockedForWriting bool) error {
 		for _, name := range container.Names {
 			if conflict, ok := names[name]; ok {
 				r.removeName(conflict, name)
-				needSave = true
+				errorToResolveBySaving = errors.New("container store is inconsistent and the current caller does not hold a write lock")
 			}
 			names[name] = containers[n]
 		}
@@ -310,14 +344,13 @@ func (r *containerStore) load(lockedForWriting bool) error {
 	r.byid = ids
 	r.bylayer = layers
 	r.byname = names
-	if needSave {
+	if errorToResolveBySaving != nil {
 		if !lockedForWriting {
-			// Eventually, the callers should be modified to retry with a write lock, instead.
-			return errors.New("container store is inconsistent and the current caller does not hold a write lock")
+			return true, errorToResolveBySaving
 		}
-		return r.Save()
+		return false, r.Save()
 	}
-	return nil
+	return false, nil
 }
 
 // Save saves the contents of the store to disk.  It should be called with
@@ -358,7 +391,7 @@ func newContainerStore(dir string) (rwContainerStore, error) {
 		return nil, err
 	}
 	defer cstore.stopWriting()
-	if err := cstore.load(true); err != nil {
+	if _, err := cstore.load(true); err != nil {
 		return nil, err
 	}
 	return &cstore, nil
diff --git a/vendor/github.com/containers/storage/drivers/quota/projectquota.go b/vendor/github.com/containers/storage/drivers/quota/projectquota.go
index ed4c7eaa5d..d831d84822 100644
--- a/vendor/github.com/containers/storage/drivers/quota/projectquota.go
+++ b/vendor/github.com/containers/storage/drivers/quota/projectquota.go
@@ -51,6 +51,7 @@ struct fsxattr {
 */
 import "C"
 import (
+	"errors"
 	"fmt"
 	"math"
 	"os"
@@ -78,6 +79,7 @@ type Control struct {
 	backingFsBlockDev string
 	nextProjectID     uint32
 	quotas            map[string]uint32
+	basePath          string
 }
 
 // Attempt to generate a unigue projectid.  Multiple directories
@@ -158,20 +160,22 @@ func NewControl(basePath string) (*Control, error) {
 		Size:   0,
 		Inodes: 0,
 	}
-	if err := setProjectQuota(backingFsBlockDev, minProjectID, quota); err != nil {
-		return nil, err
-	}
 
 	q := Control{
 		backingFsBlockDev: backingFsBlockDev,
 		nextProjectID:     minProjectID + 1,
 		quotas:            make(map[string]uint32),
+		basePath:          basePath,
+	}
+
+	if err := q.setProjectQuota(minProjectID, quota); err != nil {
+		return nil, err
 	}
 
 	//
 	// get first project id to be used for next container
 	//
-	err = q.findNextProjectID(basePath)
+	err = q.findNextProjectID()
 	if err != nil {
 		return nil, err
 	}
@@ -204,11 +208,11 @@ func (q *Control) SetQuota(targetPath string, quota Quota) error {
 	// set the quota limit for the container's project id
 	//
 	logrus.Debugf("SetQuota path=%s, size=%d, inodes=%d, projectID=%d", targetPath, quota.Size, quota.Inodes, projectID)
-	return setProjectQuota(q.backingFsBlockDev, projectID, quota)
+	return q.setProjectQuota(projectID, quota)
 }
 
 // setProjectQuota - set the quota for project id on xfs block device
-func setProjectQuota(backingFsBlockDev string, projectID uint32, quota Quota) error {
+func (q *Control) setProjectQuota(projectID uint32, quota Quota) error {
 	var d C.fs_disk_quota_t
 	d.d_version = C.FS_DQUOT_VERSION
 	d.d_id = C.__u32(projectID)
@@ -225,15 +229,35 @@ func setProjectQuota(backingFsBlockDev string, projectID uint32, quota Quota) er
 		d.d_ino_softlimit = d.d_ino_hardlimit
 	}
 
-	var cs = C.CString(backingFsBlockDev)
+	var cs = C.CString(q.backingFsBlockDev)
 	defer C.free(unsafe.Pointer(cs))
 
-	_, _, errno := unix.Syscall6(unix.SYS_QUOTACTL, C.Q_XSETPQLIM,
-		uintptr(unsafe.Pointer(cs)), uintptr(d.d_id),
-		uintptr(unsafe.Pointer(&d)), 0, 0)
-	if errno != 0 {
+	runQuotactl := func() syscall.Errno {
+		_, _, errno := unix.Syscall6(unix.SYS_QUOTACTL, C.Q_XSETPQLIM,
+			uintptr(unsafe.Pointer(cs)), uintptr(d.d_id),
+			uintptr(unsafe.Pointer(&d)), 0, 0)
+		return errno
+	}
+
+	errno := runQuotactl()
+
+	// If the backingFsBlockDev does not exist any more then try to recreate it.
+	if errors.Is(errno, unix.ENOENT) {
+		if _, err := makeBackingFsDev(q.basePath); err != nil {
+			return fmt.Errorf(
+				"failed to recreate missing backingFsBlockDev %s for projid %d: %w",
+				q.backingFsBlockDev, projectID, err,
+			)
+		}
+
+		if errno := runQuotactl(); errno != 0 {
+			return fmt.Errorf("failed to set quota limit for projid %d on %s after backingFsBlockDev recreation: %w",
+				projectID, q.backingFsBlockDev, errno)
+		}
+
+	} else if errno != 0 {
 		return fmt.Errorf("failed to set quota limit for projid %d on %s: %w",
-			projectID, backingFsBlockDev, errno)
+			projectID, q.backingFsBlockDev, errno)
 	}
 
 	return nil
@@ -332,16 +356,16 @@ func setProjectID(targetPath string, projectID uint32) error {
 
 // findNextProjectID - find the next project id to be used for containers
 // by scanning driver home directory to find used project ids
-func (q *Control) findNextProjectID(home string) error {
-	files, err := os.ReadDir(home)
+func (q *Control) findNextProjectID() error {
+	files, err := os.ReadDir(q.basePath)
 	if err != nil {
-		return fmt.Errorf("read directory failed : %s", home)
+		return fmt.Errorf("read directory failed : %s", q.basePath)
 	}
 	for _, file := range files {
 		if !file.IsDir() {
 			continue
 		}
-		path := filepath.Join(home, file.Name())
+		path := filepath.Join(q.basePath, file.Name())
 		projid, err := getProjectID(path)
 		if err != nil {
 			return err
diff --git a/vendor/github.com/containers/storage/images.go b/vendor/github.com/containers/storage/images.go
index 81aef7c9f2..260f5c389b 100644
--- a/vendor/github.com/containers/storage/images.go
+++ b/vendor/github.com/containers/storage/images.go
@@ -208,7 +208,7 @@ func (r *imageStore) startWritingWithReload(canReload bool) error {
 	}()
 
 	if canReload {
-		if err := r.reloadIfChanged(true); err != nil {
+		if _, err := r.reloadIfChanged(true); err != nil {
 			return err
 		}
 	}
@@ -235,20 +235,43 @@ func (r *imageStore) stopWriting() {
 // should use startReading() instead.
 func (r *imageStore) startReadingWithReload(canReload bool) error {
 	r.lockfile.RLock()
-	succeeded := false
+	unlockFn := r.lockfile.Unlock // A function to call to clean up, or nil
 	defer func() {
-		if !succeeded {
-			r.lockfile.Unlock()
+		if unlockFn != nil {
+			unlockFn()
 		}
 	}()
 
 	if canReload {
-		if err := r.reloadIfChanged(false); err != nil {
-			return err
+		if tryLockedForWriting, err := r.reloadIfChanged(false); err != nil {
+			if !tryLockedForWriting {
+				return err
+			}
+			unlockFn()
+			unlockFn = nil
+
+			r.lockfile.Lock()
+			unlockFn = r.lockfile.Unlock
+			if _, err := r.load(true); err != nil {
+				return err
+			}
+			unlockFn()
+			unlockFn = nil
+
+			r.lockfile.RLock()
+			unlockFn = r.lockfile.Unlock
+			// We need to check for a reload reload once more because the on-disk state could have been modified
+			// after we released the lock.
+			// If that, _again_, finds inconsistent state, just give up.
+			// We could, plausibly, retry a few times, but that inconsistent state (duplicate image names)
+			// shouldn’t be saved (by correct implementations) in the first place.
+			if _, err := r.reloadIfChanged(false); err != nil {
+				return fmt.Errorf("(even after successfully cleaning up once:) %w", err)
+			}
 		}
 	}
 
-	succeeded = true
+	unlockFn = nil
 	return nil
 }
 
@@ -267,15 +290,23 @@ func (r *imageStore) stopReading() {
 //
 // The caller must hold r.lockfile for reading _or_ writing; lockedForWriting is true
 // if it is held for writing.
-func (r *imageStore) reloadIfChanged(lockedForWriting bool) error {
+//
+// If !lockedForWriting and this function fails, the return value indicates whether
+// retrying with lockedForWriting could succeed. In that case the caller MUST
+// call load(), not reloadIfChanged() (because the “if changed” state will not
+// be detected again).
+func (r *imageStore) reloadIfChanged(lockedForWriting bool) (bool, error) {
 	r.loadMut.Lock()
 	defer r.loadMut.Unlock()
 
 	modified, err := r.lockfile.Modified()
-	if err == nil && modified {
+	if err != nil {
+		return false, err
+	}
+	if modified {
 		return r.load(lockedForWriting)
 	}
-	return err
+	return false, nil
 }
 
 func (r *imageStore) Images() ([]Image, error) {
@@ -342,36 +373,39 @@ func (i *Image) recomputeDigests() error {
 //
 // The caller must hold r.lockfile for reading _or_ writing; lockedForWriting is true
 // if it is held for writing.
-func (r *imageStore) load(lockedForWriting bool) error {
-	shouldSave := false
+//
+// If !lockedForWriting and this function fails, the return value indicates whether
+// retrying with lockedForWriting could succeed.
+func (r *imageStore) load(lockedForWriting bool) (bool, error) {
 	rpath := r.imagespath()
 	data, err := os.ReadFile(rpath)
 	if err != nil && !os.IsNotExist(err) {
-		return err
+		return false, err
 	}
 
 	images := []*Image{}
 	if len(data) != 0 {
 		if err := json.Unmarshal(data, &images); err != nil {
-			return fmt.Errorf("loading %q: %w", rpath, err)
+			return false, fmt.Errorf("loading %q: %w", rpath, err)
 		}
 	}
 	idlist := make([]string, 0, len(images))
 	ids := make(map[string]*Image)
 	names := make(map[string]*Image)
 	digests := make(map[digest.Digest][]*Image)
+	var errorToResolveBySaving error // == nil
 	for n, image := range images {
 		ids[image.ID] = images[n]
 		idlist = append(idlist, image.ID)
 		for _, name := range image.Names {
 			if conflict, ok := names[name]; ok {
 				r.removeName(conflict, name)
-				shouldSave = true
+				errorToResolveBySaving = ErrDuplicateImageNames
 			}
 		}
 		// Compute the digest list.
 		if err := image.recomputeDigests(); err != nil {
-			return fmt.Errorf("computing digests for image with ID %q (%v): %w", image.ID, image.Names, err)
+			return false, fmt.Errorf("computing digests for image with ID %q (%v): %w", image.ID, image.Names, err)
 		}
 		for _, name := range image.Names {
 			names[name] = image
@@ -383,19 +417,23 @@ func (r *imageStore) load(lockedForWriting bool) error {
 		image.ReadOnly = !r.lockfile.IsReadWrite()
 	}
 
-	if shouldSave && (!r.lockfile.IsReadWrite() || !lockedForWriting) {
-		// Eventually, the callers should be modified to retry with a write lock if IsReadWrite && !lockedForWriting, instead.
-		return ErrDuplicateImageNames
+	if errorToResolveBySaving != nil {
+		if !r.lockfile.IsReadWrite() {
+			return false, errorToResolveBySaving
+		}
+		if !lockedForWriting {
+			return true, errorToResolveBySaving
+		}
 	}
 	r.images = images
 	r.idindex = truncindex.NewTruncIndex(idlist) // Invalid values in idlist are ignored: they are not a reason to refuse processing the whole store.
 	r.byid = ids
 	r.byname = names
 	r.bydigest = digests
-	if shouldSave {
-		return r.Save()
+	if errorToResolveBySaving != nil {
+		return false, r.Save()
 	}
-	return nil
+	return false, nil
 }
 
 // Save saves the contents of the store to disk.  It should be called with
@@ -439,7 +477,7 @@ func newImageStore(dir string) (rwImageStore, error) {
 		return nil, err
 	}
 	defer istore.stopWriting()
-	if err := istore.load(true); err != nil {
+	if _, err := istore.load(true); err != nil {
 		return nil, err
 	}
 	return &istore, nil
@@ -462,7 +500,7 @@ func newROImageStore(dir string) (roImageStore, error) {
 		return nil, err
 	}
 	defer istore.stopReading()
-	if err := istore.load(false); err != nil {
+	if _, err := istore.load(false); err != nil {
 		return nil, err
 	}
 	return &istore, nil
diff --git a/vendor/github.com/containers/storage/layers.go b/vendor/github.com/containers/storage/layers.go
index 8053373f70..a0b9da4e93 100644
--- a/vendor/github.com/containers/storage/layers.go
+++ b/vendor/github.com/containers/storage/layers.go
@@ -36,6 +36,10 @@ import (
 const (
 	tarSplitSuffix = ".tar-split.gz"
 	incompleteFlag = "incomplete"
+	// maxLayerStoreCleanupIterations is the number of times we try to clean up inconsistent layer store state
+	// in readers (which, for implementation reasons, gives other writers the opportunity to create more inconsistent state)
+	// until we just give up.
+	maxLayerStoreCleanupIterations = 3
 )
 
 // A Layer is a record of a copy-on-write layer that's stored by the lower
@@ -331,7 +335,7 @@ func (r *layerStore) startWritingWithReload(canReload bool) error {
 	}()
 
 	if canReload {
-		if err := r.reloadIfChanged(true); err != nil {
+		if _, err := r.reloadIfChanged(true); err != nil {
 			return err
 		}
 	}
@@ -358,20 +362,46 @@ func (r *layerStore) stopWriting() {
 // should use startReading() instead.
 func (r *layerStore) startReadingWithReload(canReload bool) error {
 	r.lockfile.RLock()
-	succeeded := false
+	unlockFn := r.lockfile.Unlock // A function to call to clean up, or nil
 	defer func() {
-		if !succeeded {
-			r.lockfile.Unlock()
+		if unlockFn != nil {
+			unlockFn()
 		}
 	}()
 
 	if canReload {
-		if err := r.reloadIfChanged(false); err != nil {
-			return err
+		cleanupsDone := 0
+		for {
+			tryLockedForWriting, err := r.reloadIfChanged(false)
+			if err == nil {
+				break
+			}
+			if !tryLockedForWriting {
+				return err
+			}
+			if cleanupsDone >= maxLayerStoreCleanupIterations {
+				return fmt.Errorf("(even after %d cleanup attempts:) %w", cleanupsDone, err)
+			}
+			unlockFn()
+			unlockFn = nil
+
+			r.lockfile.Lock()
+			unlockFn = r.lockfile.Unlock
+			if _, err := r.load(true); err != nil {
+				return err
+			}
+			unlockFn()
+			unlockFn = nil
+
+			r.lockfile.RLock()
+			unlockFn = r.lockfile.Unlock
+			// We need to check for a reload reload again because the on-disk state could have been modified
+			// after we released the lock.
+			cleanupsDone++
 		}
 	}
 
-	succeeded = true
+	unlockFn = nil
 	return nil
 }
 
@@ -424,15 +454,23 @@ func (r *layerStore) Modified() (bool, error) {
 //
 // The caller must hold r.lockfile for reading _or_ writing; lockedForWriting is true
 // if it is held for writing.
-func (r *layerStore) reloadIfChanged(lockedForWriting bool) error {
+//
+// If !lockedForWriting and this function fails, the return value indicates whether
+// retrying with lockedForWriting could succeed. In that case the caller MUST
+// call load(), not reloadIfChanged() (because the “if changed” state will not
+// be detected again).
+func (r *layerStore) reloadIfChanged(lockedForWriting bool) (bool, error) {
 	r.loadMut.Lock()
 	defer r.loadMut.Unlock()
 
 	modified, err := r.Modified()
-	if err == nil && modified {
+	if err != nil {
+		return false, err
+	}
+	if modified {
 		return r.load(lockedForWriting)
 	}
-	return err
+	return false, nil
 }
 
 func (r *layerStore) Layers() ([]Layer, error) {
@@ -455,26 +493,28 @@ func (r *layerStore) layerspath() string {
 //
 // The caller must hold r.lockfile for reading _or_ writing; lockedForWriting is true
 // if it is held for writing.
-func (r *layerStore) load(lockedForWriting bool) error {
-	shouldSave := false
+//
+// If !lockedForWriting and this function fails, the return value indicates whether
+// retrying with lockedForWriting could succeed.
+func (r *layerStore) load(lockedForWriting bool) (bool, error) {
 	rpath := r.layerspath()
 	info, err := os.Stat(rpath)
 	if err != nil {
 		if !os.IsNotExist(err) {
-			return err
+			return false, err
 		}
 	} else {
 		r.layerspathModified = info.ModTime()
 	}
 	data, err := os.ReadFile(rpath)
 	if err != nil && !os.IsNotExist(err) {
-		return err
+		return false, err
 	}
 
 	layers := []*Layer{}
 	if len(data) != 0 {
 		if err := json.Unmarshal(data, &layers); err != nil {
-			return fmt.Errorf("loading %q: %w", rpath, err)
+			return false, fmt.Errorf("loading %q: %w", rpath, err)
 		}
 	}
 	idlist := make([]string, 0, len(layers))
@@ -482,6 +522,7 @@ func (r *layerStore) load(lockedForWriting bool) error {
 	names := make(map[string]*Layer)
 	compressedsums := make(map[digest.Digest][]string)
 	uncompressedsums := make(map[digest.Digest][]string)
+	var errorToResolveBySaving error // == nil; if there are multiple errors, this is one of them.
 	if r.lockfile.IsReadWrite() {
 		selinux.ClearLabels()
 	}
@@ -491,7 +532,7 @@ func (r *layerStore) load(lockedForWriting bool) error {
 		for _, name := range layer.Names {
 			if conflict, ok := names[name]; ok {
 				r.removeName(conflict, name)
-				shouldSave = true
+				errorToResolveBySaving = ErrDuplicateLayerNames
 			}
 			names[name] = layers[n]
 		}
@@ -505,11 +546,22 @@ func (r *layerStore) load(lockedForWriting bool) error {
 			selinux.ReserveLabel(layer.MountLabel)
 		}
 		layer.ReadOnly = !r.lockfile.IsReadWrite()
+		// The r.lockfile.IsReadWrite() condition maintains past practice:
+		// Incomplete layers in a read-only store are not treated as a reason to refuse to use other layers from that store
+		// (OTOH creating child layers on top would probably lead to problems?).
+		// We do remove incomplete layers in read-write stores so that we don’t build on top of them.
+		if layerHasIncompleteFlag(layer) && r.lockfile.IsReadWrite() {
+			errorToResolveBySaving = errors.New("an incomplete layer exists and can't be cleaned up")
+		}
 	}
 
-	if shouldSave && (!r.lockfile.IsReadWrite() || !lockedForWriting) {
-		// Eventually, the callers should be modified to retry with a write lock if IsReadWrite && !lockedForWriting, instead.
-		return ErrDuplicateLayerNames
+	if errorToResolveBySaving != nil {
+		if !r.lockfile.IsReadWrite() {
+			return false, errorToResolveBySaving
+		}
+		if !lockedForWriting {
+			return true, errorToResolveBySaving
+		}
 	}
 	r.layers = layers
 	r.idindex = truncindex.NewTruncIndex(idlist) // Invalid values in idlist are ignored: they are not a reason to refuse processing the whole store.
@@ -523,42 +575,42 @@ func (r *layerStore) load(lockedForWriting bool) error {
 		r.mountsLockfile.RLock()
 		defer r.mountsLockfile.Unlock()
 		if err := r.loadMounts(); err != nil {
-			return err
+			return false, err
 		}
+	}
 
-		// Last step: as we’re writable, try to remove anything that a previous
+	if errorToResolveBySaving != nil {
+		if !r.lockfile.IsReadWrite() {
+			return false, fmt.Errorf("internal error: layerStore.load has shouldSave but !r.lockfile.IsReadWrite")
+		}
+		// Last step: try to remove anything that a previous
 		// user of this storage area marked for deletion but didn't manage to
 		// actually delete.
 		var incompleteDeletionErrors error // = nil
-		if lockedForWriting {
-			for _, layer := range r.layers {
-				if layer.Flags == nil {
-					layer.Flags = make(map[string]interface{})
-				}
-				if layerHasIncompleteFlag(layer) {
-					logrus.Warnf("Found incomplete layer %#v, deleting it", layer.ID)
-					err = r.deleteInternal(layer.ID)
-					if err != nil {
-						// Don't return the error immediately, because deleteInternal does not saveLayers();
-						// Even if deleting one incomplete layer fails, call saveLayers() so that other possible successfully
-						// deleted incomplete layers have their metadata correctly removed.
-						incompleteDeletionErrors = multierror.Append(incompleteDeletionErrors,
-							fmt.Errorf("deleting layer %#v: %w", layer.ID, err))
-					}
-					shouldSave = true
+		for _, layer := range r.layers {
+			if layer.Flags == nil {
+				layer.Flags = make(map[string]interface{})
+			}
+			if layerHasIncompleteFlag(layer) {
+				logrus.Warnf("Found incomplete layer %#v, deleting it", layer.ID)
+				err = r.deleteInternal(layer.ID)
+				if err != nil {
+					// Don't return the error immediately, because deleteInternal does not saveLayers();
+					// Even if deleting one incomplete layer fails, call saveLayers() so that other possible successfully
+					// deleted incomplete layers have their metadata correctly removed.
+					incompleteDeletionErrors = multierror.Append(incompleteDeletionErrors,
+						fmt.Errorf("deleting layer %#v: %w", layer.ID, err))
 				}
 			}
 		}
-		if shouldSave {
-			if err := r.saveLayers(); err != nil {
-				return err
-			}
+		if err := r.saveLayers(); err != nil {
+			return false, err
 		}
 		if incompleteDeletionErrors != nil {
-			return incompleteDeletionErrors
+			return false, incompleteDeletionErrors
 		}
 	}
-	return nil
+	return false, nil
 }
 
 func (r *layerStore) loadMounts() error {
@@ -689,7 +741,7 @@ func (s *store) newLayerStore(rundir string, layerdir string, driver drivers.Dri
 		return nil, err
 	}
 	defer rlstore.stopWriting()
-	if err := rlstore.load(true); err != nil {
+	if _, err := rlstore.load(true); err != nil {
 		return nil, err
 	}
 	return &rlstore, nil
@@ -714,7 +766,7 @@ func newROLayerStore(rundir string, layerdir string, driver drivers.Driver) (roL
 		return nil, err
 	}
 	defer rlstore.stopReading()
-	if err := rlstore.load(false); err != nil {
+	if _, err := rlstore.load(false); err != nil {
 		return nil, err
 	}
 	return &rlstore, nil
diff --git a/vendor/github.com/containers/storage/pkg/parsers/kernel/uname_freebsd.go b/vendor/github.com/containers/storage/pkg/parsers/kernel/uname_freebsd.go
new file mode 100644
index 0000000000..e913fad001
--- /dev/null
+++ b/vendor/github.com/containers/storage/pkg/parsers/kernel/uname_freebsd.go
@@ -0,0 +1,17 @@
+package kernel
+
+import "golang.org/x/sys/unix"
+
+// Utsname represents the system name structure.
+// It is passthrough for unix.Utsname in order to make it portable with
+// other platforms where it is not available.
+type Utsname unix.Utsname
+
+func uname() (*unix.Utsname, error) {
+	uts := &unix.Utsname{}
+
+	if err := unix.Uname(uts); err != nil {
+		return nil, err
+	}
+	return uts, nil
+}
diff --git a/vendor/github.com/containers/storage/pkg/parsers/kernel/uname_unsupported.go b/vendor/github.com/containers/storage/pkg/parsers/kernel/uname_unsupported.go
index 052c6874a3..12671db513 100644
--- a/vendor/github.com/containers/storage/pkg/parsers/kernel/uname_unsupported.go
+++ b/vendor/github.com/containers/storage/pkg/parsers/kernel/uname_unsupported.go
@@ -1,13 +1,14 @@
-//go:build freebsd || openbsd
-// +build freebsd openbsd
+//go:build openbsd
+// +build openbsd
 
 package kernel
 
 import (
-	"errors"
+	"fmt"
+	"runtime"
 )
 
 // A stub called by kernel_unix.go .
 func uname() (*Utsname, error) {
-	return nil, errors.New("Kernel version detection is available only on linux")
+	return nil, fmt.Errorf("Kernel version detection is not available on %s", runtime.GOOS)
 }
diff --git a/vendor/github.com/containers/storage/pkg/parsers/kernel/uname_unsupported_type.go b/vendor/github.com/containers/storage/pkg/parsers/kernel/uname_unsupported_type.go
index b7e0f0c237..f515500c92 100644
--- a/vendor/github.com/containers/storage/pkg/parsers/kernel/uname_unsupported_type.go
+++ b/vendor/github.com/containers/storage/pkg/parsers/kernel/uname_unsupported_type.go
@@ -1,5 +1,5 @@
-//go:build !linux && !solaris
-// +build !linux,!solaris
+//go:build !linux && !solaris && !freebsd
+// +build !linux,!solaris,!freebsd
 
 package kernel
 
diff --git a/vendor/modules.txt b/vendor/modules.txt
index 12bd890ba7..b6854c62e7 100644
--- a/vendor/modules.txt
+++ b/vendor/modules.txt
@@ -264,7 +264,7 @@ github.com/containers/psgo/internal/dev
 github.com/containers/psgo/internal/host
 github.com/containers/psgo/internal/proc
 github.com/containers/psgo/internal/process
-# github.com/containers/storage v1.43.1-0.20221104122514-74e37433a2a0
+# github.com/containers/storage v1.44.1-0.20221110192950-67e9778710f8
 ## explicit; go 1.17
 github.com/containers/storage
 github.com/containers/storage/drivers