In some email I received from anthony baxter, sie wrote:
> 
> 
> David Miller wrote:
> > It is a Zilog 85C30, it cannot go faster than 38400 at least on sun4c
> > because of the crystal clock frequency being fed into it.
> 
> matthew green wrote:
> >    I know that at least two 8530 uarts can speak with each other at baudrate
> >    generator value 0 (from a target hardware which even has the same quartz).
> > the z8530 is certainly capable of doing faster tha 38400 ...
> >    So the problem is "you *can* run faster, but you won't get 115k" ?
> > but not on a sparc.  the clock is the problem.
> 
> FWIW, we run Sparc 2's with SunOS 4.1.3 at 76.8K all the time - they handle
> the load fine. Actually, now that I think about it, there's an ELC doing it,
> too.

Umm, do you get sustained load at 76.8k ?  I've found that at 19.2k, there
is occasional lossage, but maybe that's just a modem-ss2 thingy.

Oh, I think I also tried PC-SS2 doing PPP and above 9600(?), the SS2 just
didn't want to know about it.

FIFO is too small, perhaps ?

darren