archive: set WithModTimeUpperBound when WithSourceDateEpoch is set

WithModTimeUpperBound sets the upper bound value of the ModTime property
of the tar entry structs.

WithSourceDateEpoch now implies WithModTimeUpperBound too,
in addition to WithWhiteoutTime.

For moby/buildkit issue 3296

Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>

archive/tar.go

@@ -97,7 +97,9 @@ func WriteDiff(ctx context.Context, w io.Writer, a, b string, opts ...WriteDiffO
 func writeDiffNaive(ctx context.Context, w io.Writer, a, b string, o WriteDiffOptions) error {
 	var opts []ChangeWriterOpt
 	if o.SourceDateEpoch != nil {
-		opts = append(opts, WithWhiteoutTime(*o.SourceDateEpoch))
+		opts = append(opts,
+			WithModTimeUpperBound(*o.SourceDateEpoch),
+			WithWhiteoutTime(*o.SourceDateEpoch))
 	}
 	cw := NewChangeWriter(w, b, opts...)
 	err := fs.Changes(ctx, a, b, cw.HandleChange)
@@ -495,6 +497,7 @@ func mkparent(ctx context.Context, path, root string, parents []string) error {
 type ChangeWriter struct {
 	tw                *tar.Writer
 	source            string
+	modTimeUpperBound *time.Time
 	whiteoutT         time.Time
 	inodeSrc          map[uint64]string
 	inodeRefs         map[uint64][]string
@@ -504,6 +507,13 @@ type ChangeWriter struct {
 // ChangeWriterOpt can be specified in NewChangeWriter.
 type ChangeWriterOpt func(cw *ChangeWriter)
 
+// WithModTimeUpperBound sets the mod time upper bound.
+func WithModTimeUpperBound(tm time.Time) ChangeWriterOpt {
+	return func(cw *ChangeWriter) {
+		cw.modTimeUpperBound = &tm
+	}
+}
+
 // WithWhiteoutTime sets the whiteout timestamp.
 func WithWhiteoutTime(tm time.Time) ChangeWriterOpt {
 	return func(cw *ChangeWriter) {
@@ -579,6 +589,9 @@ func (cw *ChangeWriter) HandleChange(k fs.ChangeKind, p string, f os.FileInfo, e
 
 		// truncate timestamp for compatibility. without PAX stdlib rounds timestamps instead
 		hdr.Format = tar.FormatPAX
+		if cw.modTimeUpperBound != nil && hdr.ModTime.After(*cw.modTimeUpperBound) {
+			hdr.ModTime = *cw.modTimeUpperBound
+		}
 		hdr.ModTime = hdr.ModTime.Truncate(time.Second)
 		hdr.AccessTime = time.Time{}
 		hdr.ChangeTime = time.Time{}

archive/tar_opts.go

@@ -91,7 +91,10 @@ type WriteDiffOptions struct {
 
 	writeDiffFunc func(context.Context, io.Writer, string, string, WriteDiffOptions) error
 
-	// SourceDateEpoch specifies the timestamp used for whiteouts to provide control for reproducibility.
+	// SourceDateEpoch specifies the following timestamps to provide control for reproducibility.
+	//   - The upper bound timestamp of the diff contents
+	//   - The timestamp of the whiteouts
+	//
 	// See also https://reproducible-builds.org/docs/source-date-epoch/ .
 	SourceDateEpoch *time.Time
 }

archive/tar_test.go

@@ -36,6 +36,7 @@ import (
 	"github.com/containerd/containerd/pkg/testutil"
 	"github.com/containerd/continuity/fs"
 	"github.com/containerd/continuity/fs/fstest"
+	"github.com/opencontainers/go-digest"
 	"github.com/stretchr/testify/require"
 	exec "golang.org/x/sys/execabs"
 )
@@ -1157,20 +1158,36 @@ func TestDiffTar(t *testing.T) {
 	}
 }
 
-func TestWhiteoutSourceDateEpoch(t *testing.T) {
+func TestSourceDateEpoch(t *testing.T) {
 	sourceDateEpoch, err := time.Parse(time.RFC3339, "2022-01-23T12:34:56Z")
 	require.NoError(t, err)
+	past, err := time.Parse(time.RFC3339, "2022-01-01T00:00:00Z")
+	require.NoError(t, err)
+	require.True(t, past.Before(sourceDateEpoch))
+	veryRecent := time.Now()
+	require.True(t, veryRecent.After(sourceDateEpoch))
+
 	opts := []WriteDiffOpt{WithSourceDateEpoch(&sourceDateEpoch)}
 	validators := []tarEntryValidator{
 		composeValidators(whiteoutEntry("f1"), requireModTime(sourceDateEpoch)),
+		composeValidators(fileEntry("f2", []byte("content2"), 0644), requireModTime(past)),
+		composeValidators(fileEntry("f3", []byte("content3"), 0644), requireModTime(sourceDateEpoch)),
 	}
 	a := fstest.Apply(
 		fstest.CreateFile("/f1", []byte("content"), 0644),
 	)
 	b := fstest.Apply(
+		// Remove f1; the timestamp of the tar entry will be sourceDateEpoch
 		fstest.RemoveAll("/f1"),
+		// Create f2 with the past timestamp; the timestamp of the tar entry will be past (< sourceDateEpoch)
+		fstest.CreateFile("/f2", []byte("content2"), 0644),
+		fstest.Chtimes("/f2", past, past),
+		// Create f3 with the veryRecent timestamp; the timestamp of the tar entry will be sourceDateEpoch
+		fstest.CreateFile("/f3", []byte("content3"), 0644),
+		fstest.Chtimes("/f3", veryRecent, veryRecent),
 	)
 	makeDiffTarTest(validators, a, b, opts...)(t)
+	makeDiffTarReproTest(a, b, opts...)(t)
 }
 
 type tarEntryValidator func(*tar.Header, []byte) error
@@ -1303,6 +1320,54 @@ func makeDiffTarTest(validators []tarEntryValidator, a, b fstest.Applier, opts .
 	}
 }
 
+func makeDiffTar(t *testing.T, a, b fstest.Applier, opts ...WriteDiffOpt) (digest.Digest, []byte) {
+	ad := t.TempDir()
+	if err := a.Apply(ad); err != nil {
+		t.Fatalf("failed to apply a: %v", err)
+	}
+
+	bd := t.TempDir()
+	if err := fs.CopyDir(bd, ad); err != nil {
+		t.Fatalf("failed to copy dir: %v", err)
+	}
+	if err := b.Apply(bd); err != nil {
+		t.Fatalf("failed to apply b: %v", err)
+	}
+
+	rc := Diff(context.Background(), ad, bd, opts...)
+	defer rc.Close()
+	var buf bytes.Buffer
+	r := io.TeeReader(rc, &buf)
+	dgst, err := digest.FromReader(r)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if err = rc.Close(); err != nil {
+		t.Fatal(err)
+	}
+	return dgst, buf.Bytes()
+}
+
+func makeDiffTarReproTest(a, b fstest.Applier, opts ...WriteDiffOpt) func(*testing.T) {
+	return func(t *testing.T) {
+		const (
+			count = 5
+			delay = 100 * time.Millisecond
+		)
+		var lastDigest digest.Digest
+		for i := 0; i < count; i++ {
+			dgst, _ := makeDiffTar(t, a, b, opts...)
+			t.Logf("#%d: digest %s", i, dgst)
+			if lastDigest == "" {
+				lastDigest = dgst
+			} else if dgst != lastDigest {
+				t.Fatalf("expected digest %s, got %s", lastDigest, dgst)
+			}
+			time.Sleep(delay)
+		}
+	}
+}
+
 type diffApplier struct{}
 
 func (d diffApplier) TestContext(ctx context.Context) (context.Context, func(), error) {