So I was watching pf interact with epat last night and I finally got it
through my head that instead of doing something like initialize the chip
once during pi->proto->init_proto(pi); and then just manage the connection
status with the connect() & disconnect() calls all the protocol drivers
appear to be fully initializing their respective chips on every
connect()/disconnect(). Is there any good reason to be doing it this
way? I'd rather just put the initialization stuff in something like
proto_init() (or whatever it was called) and call it once when the IDE
portion is loaded. For instance, with the epat protocol, there's lots of
stuff going on in connect() that doesn't *really* have any business being
called repeatedly over time.
So ... if I go ahead and change this so the initialization is done once
and then the connection is just managed will this break anything that I'm
not thinking of?
Josh
(incidentally, this should also make it easier to make the epat chip daisy
chain work)
-- To unsubscribe, send mail to: linux-parport-request@torque.net --
-- with the single word "unsubscribe" in the body of the message. --
This archive was generated by hypermail 2b29 : Tue Sep 12 2000 - 12:31:55 EDT