diff --git a/hack/vendor.sh b/hack/vendor.sh
index 2ee530a736..b0a89db68d 100755
--- a/hack/vendor.sh
+++ b/hack/vendor.sh
@@ -53,9 +53,9 @@ clone hg code.google.com/p/go.net 84a4013f96e0
 
 clone hg code.google.com/p/gosqlite 74691fb6f837
 
-# get Go tip's archive/tar, for xattr support
-# TODO after Go 1.3 drops, bump our minimum supported version and drop this vendored dep
-clone hg code.google.com/p/go 3458ba248590
+# get Go tip's archive/tar, for xattr support and improved performance
+# TODO after Go 1.4 drops, bump our minimum supported version and drop this vendored dep
+clone hg code.google.com/p/go 17404efd6b02
 mv src/code.google.com/p/go/src/pkg/archive/tar tmp-tar
 rm -rf src/code.google.com/p/go
 mkdir -p src/code.google.com/p/go/src/pkg/archive
diff --git a/vendor/src/code.google.com/p/go/src/pkg/archive/tar/reader.go b/vendor/src/code.google.com/p/go/src/pkg/archive/tar/reader.go
index 920a9b08f9..a27559d0f0 100644
--- a/vendor/src/code.google.com/p/go/src/pkg/archive/tar/reader.go
+++ b/vendor/src/code.google.com/p/go/src/pkg/archive/tar/reader.go
@@ -29,10 +29,11 @@ const maxNanoSecondIntSize = 9
 // The Next method advances to the next file in the archive (including the first),
 // and then it can be treated as an io.Reader to access the file's data.
 type Reader struct {
-	r    io.Reader
-	err  error
-	pad  int64          // amount of padding (ignored) after current file entry
-	curr numBytesReader // reader for current file entry
+	r       io.Reader
+	err     error
+	pad     int64           // amount of padding (ignored) after current file entry
+	curr    numBytesReader  // reader for current file entry
+	hdrBuff [blockSize]byte // buffer to use in readHeader
 }
 
 // A numBytesReader is an io.Reader with a numBytes method, returning the number
@@ -426,7 +427,9 @@ func (tr *Reader) verifyChecksum(header []byte) bool {
 }
 
 func (tr *Reader) readHeader() *Header {
-	header := make([]byte, blockSize)
+	header := tr.hdrBuff[:]
+	copy(header, zeroBlock)
+
 	if _, tr.err = io.ReadFull(tr.r, header); tr.err != nil {
 		return nil
 	}
diff --git a/vendor/src/code.google.com/p/go/src/pkg/archive/tar/writer.go b/vendor/src/code.google.com/p/go/src/pkg/archive/tar/writer.go
index 6eff6f6f84..d107dbbb51 100644
--- a/vendor/src/code.google.com/p/go/src/pkg/archive/tar/writer.go
+++ b/vendor/src/code.google.com/p/go/src/pkg/archive/tar/writer.go
@@ -37,8 +37,9 @@ type Writer struct {
 	nb         int64 // number of unwritten bytes for current file entry
 	pad        int64 // amount of padding to write after current file entry
 	closed     bool
-	usedBinary bool // whether the binary numeric field extension was used
-	preferPax  bool // use pax header instead of binary numeric header
+	usedBinary bool            // whether the binary numeric field extension was used
+	preferPax  bool            // use pax header instead of binary numeric header
+	hdrBuff    [blockSize]byte // buffer to use in writeHeader
 }
 
 // NewWriter creates a new Writer writing to w.
@@ -160,7 +161,8 @@ func (tw *Writer) writeHeader(hdr *Header, allowPax bool) error {
 	// subsecond time resolution, but for now let's just capture
 	// too long fields or non ascii characters
 
-	header := make([]byte, blockSize)
+	header := tw.hdrBuff[:]
+	copy(header, zeroBlock)
 	s := slicer(header)
 
 	// keep a reference to the filename to allow to overwrite it later if we detect that we can use ustar longnames instead of pax