Improved push and pull with upload manager and download manager

This commit adds a transfer manager which deduplicates and schedules
transfers, and also an upload manager and download manager that build on
top of the transfer manager to provide high-level interfaces for uploads
and downloads. The push and pull code is modified to use these building
blocks.

Some benefits of the changes:

- Simplification of push/pull code
- Pushes can upload layers concurrently
- Failed downloads and uploads are retried after backoff delays
- Cancellation is supported, but individual transfers will only be
  cancelled if all pushes or pulls using them are cancelled.
- The distribution code is decoupled from Docker Engine packages and API
  conventions (i.e. streamformatter), which will make it easier to split
  out.

This commit also includes unit tests for the new distribution/xfer
package. The tests cover 87.8% of the statements in the package.

Signed-off-by: Aaron Lehmann <aaron.lehmann@docker.com>

integration-cli/docker_cli_pull_test.go

@@ -140,7 +140,7 @@ func (s *DockerHubPullSuite) TestPullAllTagsFromCentralRegistry(c *check.C) {
 }
 
 // TestPullClientDisconnect kills the client during a pull operation and verifies that the operation
-// still succesfully completes on the daemon side.
+// gets cancelled.
 //
 // Ref: docker/docker#15589
 func (s *DockerHubPullSuite) TestPullClientDisconnect(c *check.C) {
@@ -161,14 +161,8 @@ func (s *DockerHubPullSuite) TestPullClientDisconnect(c *check.C) {
 	err = pullCmd.Process.Kill()
 	c.Assert(err, checker.IsNil)
 
-	maxAttempts := 20
-	for i := 0; ; i++ {
-		if _, err := s.CmdWithError("inspect", repoName); err == nil {
-			break
-		}
-		if i >= maxAttempts {
-			c.Fatal("timeout reached: image was not pulled after client disconnected")
-		}
-		time.Sleep(500 * time.Millisecond)
+	time.Sleep(2 * time.Second)
+	if _, err := s.CmdWithError("inspect", repoName); err == nil {
+		c.Fatal("image was pulled after client disconnected")
 	}
 }