Unidentified Emulex card from 1991

Har. The same chip is used for ttyb. And even ttya can have its parameters changed. Not nearly as friendly as it could be. Some of it is spazzes which, as you say, just complicate init code (such as the mapping between the 0xc0 bits of write register 3 and character size). But some of it is real problems: no FIFO to speak of, a view of hardware flow control which means either wiring signals up to the `wrong' pins and compensating in software or disagreeing with almost everybody else about what signals drive flow control, a weird set of BRG pre-divisors which impose an unpleasantly low cap on the baudrate if you want reliable communication, software setting of hardware things like "we're using a directly-attached crystal to drive the BRG". Perhaps those are covered by your meaning of "awesome".... And then there are the interface things. In addition to the stuff mentioned above, I see reset bits in the same register as some interrupt control, most of the interrupt bits in the same register as the write-enable bit for another register, putting some of the channel B bits in channel A's registers...and possibly others I'm just not competent to notice. But a right mess of a design. I can't imagine any of what I cite above would be particularly difficult to fix or would require impairing its capabilities. Somehow, it's the "kids today" I'd expect to be dismissing as ignorable extra code bloat compelled by hardware insanities and people like us complaining about it. /~\ The ASCII Mouse \ / Ribbon Campaign X Against HTML mouse at rodents-montreal.org / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B