In the past, a utility did the determination and remapping via off-disk software. My understanding was the traditional "bad block" behavior was actually hidden

In the past, a utility did the determination and remapping via off-disk software. My understanding was the traditional "bad block" behavior was actually hidden - there are bad blocks that occur but the firmware silently remaps everything. When you see a bad block at the level you can see via e.g. smartctl program, that means the space set aside for bad block remapping is full - the drive has already been silently failing for a while now.