create-update-payload: add option to use format for an MBL OS version

[jonathanhaigh-arm]

create-update-payload now supports creating payloads using update payload formats from previous versions of MBL OS, but the user must know the integer payload format version in order to specify it.

The user shouldn't have to know about payload format versions, so add another option that accepts MBL OS versions (e.g. "mbl-os-0.9") and sets the payload format version to the correct value for that version of MBL OS.

For IOTMBL-2613: Robust update: provide mechanism to update across payload format versions

Jenkins builds

• http://jenkins.mbed-linux.arm.com/view/jh/job/jh-test/197/ (succeeded).

Lava results

• http://e120856-lin/results.php?image_version=custom&image_number=197&custom_version=jh-test&lava_query=&lava_owner=&failures=false&submitter=jonhai01&sort=by_device. There were multiple failures.

Expected failures:

• Helloworld update failures (expected until we update the tests).
• systemd test failures due to bluetooth service failure on rpi3 (currently expected).
• lxc container build/boot failures (currently expected).

Unexpected failures:

• Bootloader1 test failure due to timeout: https://lava.mbedcloudtesting.com/scheduler/job/82143
• imx_usb failed on imx7d-pico-oulab-05: https://lava.mbedcloudtesting.com/scheduler/job/82105
• block device not found on imx7d-pico-oulab-07: https://lava.mbedcloudtesting.com/scheduler/job/82111
• Some sort of provisioning failure on imx7d-pico-oulab-06: https://lava.mbedcloudtesting.com/scheduler/job/82106.
• Wi-Fi access failure on imx8mmevk-oulab-04: https://lava.mbedcloudtesting.com/scheduler/job/82075. The same board failed on its last warrior-dev Wi-Fi access test (https://lava.mbedcloudtesting.com/scheduler/job/81744) so it's not due to changes in this PR.
• Wi-Fi access failure on imx6ul-pico-oulab-03: https://lava.mbedcloudtesting.com/scheduler/job/82103. The same board failed on its last warrior-dev Wi-Fi access test (https://lava.mbedcloudtesting.com/results/82185) so it's not due to changes in this PR.

None of these unexpected failures seem to be related to the changes in this PR...