Quite a few of us are shackled with a useless WinModem through no
fault of our own. I bought a laptop; it has a WinModem built in, I
have to carry a PCMCIA card when I go on the road. The CPU loading is
not a problem for me, the laptop has a PII/300 in it that hardly ever
breaks a sweat. I expect that I am not the only one in this boat.
I personally have never written a driver for anything but Solaris but
the requirements seem pretty modest. It is just that the code paths in
the interrupt handlers might be a bit long. My CPU would probably be
happy to chew on something for a while, even kernel compiles don't
seem to take very long anymore.
Richard,
I would reccomend getting a copy of:
Linux Device Drivers by Alessandro Rubini, O'Reilly and Asscociates
ISBN 1-56592-292-1
http://www.amazon.com/exec/obidos/ASIN/1565922921/ref=sim_books/002-7777075-8168254
and
Linux Kernel Internals 2nd edition by M. Beck et al. Addison-Wesley
ISBN 201331438
http://www.amazon.com/exec/obidos/ASIN/0201331438/qid=923689511/sr=1-1/002-7777075-8168254
> On Thu, 8 Apr 1999, Richard Reynolds wrote:
> > someone saying that this modem was all software driven (duh I knew that) and
> > that the drivers were somekind of seceret (did not know that can't even see
> > how, sence i found them?) and that they were not available. Well as of now i
> > have the technology (drivers source code) to make the drivers required to
> > run my modem. I can recreate the drivers for windozs make new ones for DOS
> > (also not "supported" by the vendor) but i have no ability to program in
> > Linux and don't even know where to start. The drivers would need to do 3
> > major things
> I've heard that the timing was very important to these kinds of modem
> trying to support these brain-dead things would mess up scheduling as it
> needs constant real-time dsp in order to work.
> IMHO well-behaved hardware should use buffering, dma, etc to be time
> constraint friendly to the OS. Real modems are serial ports buffered with
> 16550A's or better. Or maybe USB modems as I've been rumored as to their
> existence.
> > 1. initialize the device maybee at a very early stage the ports are turned
> > off until initialize is complete and successful.
> > 2. process every 32bit word before being sent to the modem.
> > 3. process every 32bit word after being recieved by the modem.
> Play dsp cause the manufactorer was to cheap to be good?
> Chew up 20% or more of your cpu? Constantly allow it to preempt the things
> running on your system?
> Sounds real ugly to me.
> Of course I'll be quiet now as I'm a little extreme in the fact that I
> would pay extra for a modem without the evil fax capabilities.
> I really hate fax;->
>
> Please read the FAQ at http://www.tux.org/lkml/
-
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/