Good explanation. So the SetWriteMask() is only an optimization for non-interleaved planar bitmaps? It now sounds like SetWriteMask() should be removed in all cases this "optimization" is used for AmigaOS 3+.
It's not that terrible, really. The blitter works exactly like it always did regardless of whether you restart it several times over for each individual plane, or let it rip for a single consecutive chunk of memory which comprises all planes. Noticeable gains are only possible on an AGA machine (higher bandwidth). On an ECS machine the effects will be less pronounced. In Super72 mode at four colours, ECS already has so much trouble at hand that it cannot always detect if the blitter is still running :-(
Where the write mask makes a big difference, however, is with RTG screens. Instead of jush pushing chunky pixels around, CyberGraphX/Picasso96/etc. now have to treat the screen as if it were a real planar bitmap, so that only the relevant data is affected when text is rendered or scrolled. This requires significant extra effort.
Ironically, the "optimization" of restricting rendering operations to a single plane now cause exactly the opposite of the intended goal.