There is no guarantee that a fix delay is actually enough, because with
the BT_EN gpio, we only control activation of bluetooth on the chip. But
we don't know the status of the whole chip.
Also hciattach has anyway a timeout of 5 seconds, so 0.1 seconds does
probably not make a big difference in this regard.
So in order to make it the safest possible, we can only rely a failure
handling and retrying. The number 5 for retries has been defined since
repeated restart of the service with only 3 retries leads to failure
pretty quickly. At 4 there is no failure anymore. So with 5 we are sure
the risk of failure is really low.
BugzID: 70195