Re: 2.3: Designing a better serial subsystem.

Alan Cox (alan@lxorguk.ukuu.org.uk)
Thu, 5 Nov 1998 15:13:02 +0000 (GMT)


> Every now and then the issue comes up: a serial driver does a lot of
> stuff that could've been in generic code. The serial subsystem needs
> a rework.

The m68K guys have had a modular serial driver for a long time

> The best thing would be if a device driver for a serial card would
> only export a few simple functions. On the other hand we'd like to be
> able to handle high performance cards efficiently.

Don't make the same mistake they did

> Also the system should be able to request "I don't need to know
> anything until there is an xx character in the data stream". Thats a
> feature some cards support, and should be usable by the PPP line
> discipline (there is a fixed end-of-packet character in PPP right?).

That sort of smart stuff is needed

> I'm not good at thinking up good names. Feel free to suggest
> others. I'm sure I forgot a few things that the system might want from
> a serial driver.

Instead of trying to jam a layer of code into the serial stuff, take
the routines that are duplicates and bugging you and make a libserial.c
that anyone can call.

Then the driver keeps control. It can simply opt to call

generic_serial_open(my_card_hangup, my_card_wait);

Alan

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/