In D47073#1073123, @markj wrote:

Seems like it'd be easy enough to plumb the device_t in here and use device_printf()...

Except that some (all?) of these routines may be used as part of early console where no device_t exists (though maybe we don't flush there, it's a generic interface).
And that would be a lot of replumbing in every single client uart driver.
It might be worth doing in the fullness of time, but it's a much better project than I want to bite off here to make this more correct (since the lines will be proximately close to the uart probe messages typically.
I'll add it to the list of backlog items for uart...