>> Of course, at really high data rates, the CPU will enter the
>> interrupt handler once and scarf up all available input without
>> leaving it.

> I can testify that on a sparc IPC running SunOS 4.1.3, full-speed
> input at 38400 causes the machine to freeze HARD until the serial
> cable is yanked out,

38400 isn't *nearly* "really high data rates" in this sense.  I mean
data rates high enough that in the time it takes to read one character
from the chip and stuff it in the ring buffer, the next arrives.

Of course, at that sort of data rate, in the time it takes to enter the
interrupt handler, you've dropped dozens of characters.  Better pad the
beginnings of those packets. :-/  Or, I suppose, fast-path that
interrupt handler, maybe even make it run in the trap window....

But it's all pretty much irrelevant, since data rates that high have to
use X1, and PCLK-driven data reception at X1 is dubious at best.

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B