Instead of first checking the result of a global gptp command and then
checking the result of a detailed gptp command for the specific
network interface, only check the detailed command. This prevents
getting potential unexpected misalignment between the two commands.
The reason for issuing two commands was to generate less verbose output
while looping to get the expected state on the first one.

Also, ensure that the Zephyr prompt is there before starting the tests.

Signed-off-by: Ambroise Vincent <ambroise.vincent@arm.com>