> Hmm, you mean that linesw[tp->t_line].l_rint() is inside an spltty()
> protected section in the new driver, but not in the old one?

Yes, that's exactly what I mean.  8-)

It's not entirely clear that the old behaviour is a 'right' (by the
definition of how the tty l_rint routine should work) but if it's not
then it's impossible to do a split-level serial driver on the Alpha at
all, unless all serial drivers are split-level drivers...  8-)

> I'm not
> sure that his protection is even necessary in the new one (at first glance),
> so that might easily be fixed.

I'm not sure that that protection is necessary, either...

It's worth noting that i don't particularly trust the way the input
fifo is managed.  Looks like there's the potential for weird
(undetected) wraprounds...




chris