archive: preserve hardlinks in Tar and Untar

* integration test for preserving hardlinks

Signed-off-by: Vincent Batts <vbatts@redhat.com>
Signed-off-by: Vincent Batts <vbatts@hashbangbash.com>

archive/archive.go

@@ -153,7 +153,15 @@ func (compression *Compression) Extension() string {
 	return ""
 }
 
-func addTarFile(path, name string, tw *tar.Writer, twBuf *bufio.Writer) error {
+type tarAppender struct {
+	TarWriter *tar.Writer
+	Buffer    *bufio.Writer
+
+	// for hardlink mapping
+	SeenFiles map[uint64]string
+}
+
+func (ta *tarAppender) addTarFile(path, name string) error {
 	fi, err := os.Lstat(path)
 	if err != nil {
 		return err
@@ -188,13 +196,28 @@ func addTarFile(path, name string, tw *tar.Writer, twBuf *bufio.Writer) error {
 
 	}
 
+	// if it's a regular file and has more than 1 link,
+	// it's hardlinked, so set the type flag accordingly
+	if fi.Mode().IsRegular() && stat.Nlink > 1 {
+		// a link should have a name that it links too
+		// and that linked name should be first in the tar archive
+		ino := uint64(stat.Ino)
+		if oldpath, ok := ta.SeenFiles[ino]; ok {
+			hdr.Typeflag = tar.TypeLink
+			hdr.Linkname = oldpath
+			hdr.Size = 0 // This Must be here for the writer math to add up!
+		} else {
+			ta.SeenFiles[ino] = name
+		}
+	}
+
 	capability, _ := system.Lgetxattr(path, "security.capability")
 	if capability != nil {
 		hdr.Xattrs = make(map[string]string)
 		hdr.Xattrs["security.capability"] = string(capability)
 	}
 
-	if err := tw.WriteHeader(hdr); err != nil {
+	if err := ta.TarWriter.WriteHeader(hdr); err != nil {
 		return err
 	}
 
@@ -204,17 +227,17 @@ func addTarFile(path, name string, tw *tar.Writer, twBuf *bufio.Writer) error {
 			return err
 		}
 
-		twBuf.Reset(tw)
+		ta.Buffer.Reset(ta.TarWriter)
-		_, err = io.Copy(twBuf, file)
+		_, err = io.Copy(ta.Buffer, file)
 		file.Close()
 		if err != nil {
 			return err
 		}
-		err = twBuf.Flush()
+		err = ta.Buffer.Flush()
 		if err != nil {
 			return err
 		}
-		twBuf.Reset(nil)
+		ta.Buffer.Reset(nil)
 	}
 
 	return nil
@@ -345,9 +368,15 @@ func TarWithOptions(srcPath string, options *TarOptions) (io.ReadCloser, error)
 		return nil, err
 	}
 
-	tw := tar.NewWriter(compressWriter)
-
 	go func() {
+		ta := &tarAppender{
+			TarWriter: tar.NewWriter(compressWriter),
+			Buffer:    pools.BufioWriter32KPool.Get(nil),
+			SeenFiles: make(map[uint64]string),
+		}
+		// this buffer is needed for the duration of this piped stream
+		defer pools.BufioWriter32KPool.Put(ta.Buffer)
+
 		// In general we log errors here but ignore them because
 		// during e.g. a diff operation the container can continue
 		// mutating the filesystem and we can see transient errors
@@ -357,9 +386,6 @@ func TarWithOptions(srcPath string, options *TarOptions) (io.ReadCloser, error)
 			options.Includes = []string{"."}
 		}
 
-		twBuf := pools.BufioWriter32KPool.Get(nil)
-		defer pools.BufioWriter32KPool.Put(twBuf)
-
 		var renamedRelFilePath string // For when tar.Options.Name is set
 		for _, include := range options.Includes {
 			filepath.Walk(filepath.Join(srcPath, include), func(filePath string, f os.FileInfo, err error) error {
@@ -395,7 +421,7 @@ func TarWithOptions(srcPath string, options *TarOptions) (io.ReadCloser, error)
 					relFilePath = strings.Replace(relFilePath, renamedRelFilePath, options.Name, 1)
 				}
 
-				if err := addTarFile(filePath, relFilePath, tw, twBuf); err != nil {
+				if err := ta.addTarFile(filePath, relFilePath); err != nil {
 					log.Debugf("Can't add file %s to tar: %s", srcPath, err)
 				}
 				return nil
@@ -403,7 +429,7 @@ func TarWithOptions(srcPath string, options *TarOptions) (io.ReadCloser, error)
 		}
 
 		// Make sure to check the error on Close.
-		if err := tw.Close(); err != nil {
+		if err := ta.TarWriter.Close(); err != nil {
 			log.Debugf("Can't close tar writer: %s", err)
 		}
 		if err := compressWriter.Close(); err != nil {

archive/archive_test.go

@@ -249,6 +249,64 @@ func TestUntarUstarGnuConflict(t *testing.T) {
 	}
 }
 
+func TestTarWithHardLink(t *testing.T) {
+	origin, err := ioutil.TempDir("", "docker-test-tar-hardlink")
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer os.RemoveAll(origin)
+	if err := ioutil.WriteFile(path.Join(origin, "1"), []byte("hello world"), 0700); err != nil {
+		t.Fatal(err)
+	}
+	if err := os.Link(path.Join(origin, "1"), path.Join(origin, "2")); err != nil {
+		t.Fatal(err)
+	}
+
+	var i1, i2 uint64
+	if i1, err = getNlink(path.Join(origin, "1")); err != nil {
+		t.Fatal(err)
+	}
+	// sanity check that we can hardlink
+	if i1 != 2 {
+		t.Skipf("skipping since hardlinks don't work here; expected 2 links, got %d", i1)
+	}
+
+	dest, err := ioutil.TempDir("", "docker-test-tar-hardlink-dest")
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer os.RemoveAll(dest)
+
+	// we'll do this in two steps to separate failure
+	fh, err := Tar(origin, Uncompressed)
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	// ensure we can read the whole thing with no error, before writing back out
+	buf, err := ioutil.ReadAll(fh)
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	bRdr := bytes.NewReader(buf)
+	err = Untar(bRdr, dest, &TarOptions{Compression: Uncompressed})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	if i1, err = getInode(path.Join(dest, "1")); err != nil {
+		t.Fatal(err)
+	}
+	if i2, err = getInode(path.Join(dest, "2")); err != nil {
+		t.Fatal(err)
+	}
+
+	if i1 != i2 {
+		t.Errorf("expected matching inodes, but got %d and %d", i1, i2)
+	}
+}
+
 func getNlink(path string) (uint64, error) {
 	stat, err := os.Stat(path)
 	if err != nil {

archive/changes.go

@@ -368,11 +368,15 @@ func minor(device uint64) uint64 {
 // ExportChanges produces an Archive from the provided changes, relative to dir.
 func ExportChanges(dir string, changes []Change) (Archive, error) {
 	reader, writer := io.Pipe()
-	tw := tar.NewWriter(writer)
-
 	go func() {
-		twBuf := pools.BufioWriter32KPool.Get(nil)
+		ta := &tarAppender{
-		defer pools.BufioWriter32KPool.Put(twBuf)
+			TarWriter: tar.NewWriter(writer),
+			Buffer:    pools.BufioWriter32KPool.Get(nil),
+			SeenFiles: make(map[uint64]string),
+		}
+		// this buffer is needed for the duration of this piped stream
+		defer pools.BufioWriter32KPool.Put(ta.Buffer)
+
 		// In general we log errors here but ignore them because
 		// during e.g. a diff operation the container can continue
 		// mutating the filesystem and we can see transient errors
@@ -390,19 +394,19 @@ func ExportChanges(dir string, changes []Change) (Archive, error) {
 					AccessTime: timestamp,
 					ChangeTime: timestamp,
 				}
-				if err := tw.WriteHeader(hdr); err != nil {
+				if err := ta.TarWriter.WriteHeader(hdr); err != nil {
 					log.Debugf("Can't write whiteout header: %s", err)
 				}
 			} else {
 				path := filepath.Join(dir, change.Path)
-				if err := addTarFile(path, change.Path[1:], tw, twBuf); err != nil {
+				if err := ta.addTarFile(path, change.Path[1:]); err != nil {
 					log.Debugf("Can't add file %s to tar: %s", path, err)
 				}
 			}
 		}
 
 		// Make sure to check the error on Close.
-		if err := tw.Close(); err != nil {
+		if err := ta.TarWriter.Close(); err != nil {
 			log.Debugf("Can't close layer: %s", err)
 		}
 		writer.Close()