>> [...] within one part in 128 error [...]

>> If you ask for 115200, boards with 20MHz clock (like mine) give you
>> 113636.36363636...  is 1.3573+% low
> Isn't that > 1 part in 128?

Duh.  Sorry about that.  Let me rephrase it:

The closest boards with 20MHz clock can come to 115200 is
113636.36.363636..., which is 1.3573+% low.

This means either asking for a more precise rate or
relaxing/eliminating the driver's error tolerance checks.

(Of course, this is all with my current driver, which is not in-tree,
though the comments about what the hardware can achieve are
driver-independent.)

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B