As I mentioned previously, I've prepared and submitted a patch which
allows the specification of two separated speed rates for the serial
line when talking to the sick lms: one for initial connection attempt,
and another for the speed to be set during initialization for data transfer.
I have tested the patch and it removes all the problems we were having
with our lasers. I'm a bit suspicious because I assume a lot of people
has sick lasers and nobody mentioned any issues with them.