Zygo Blaxell wrote:
> That 25% of CPU time is probably non-interruptable too, in order to
> meet the strict realtime requirements of most of the line protocols.
Unlikely. From reading the v8 and v34 spec, the RT requirement
is in the order of tens of ms, so certainly doable without
disabling interrupts (except for the pathological cases, such
as IDE with ints masked).
> I would suspect that there's very little information to know about the
> WinModem itself. It's probably just a DSP (or even just a simple pair of
If you'd care to look at the scarce information available, then
you'd know that for example the Rockwell "Winmodem" chipset
(called WinRPI) still has the data pump in hardware, it just
delegates the MNP/v42 compression/packetization protocol to the
host.
> The _hard_ part is writing signal processing code for speeds above
> 2400 bps without at least an oscilloscope and access to a library full
I've done modem stuff with 9600 bps on a 486 machine, so what's
the problem?
Tom
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to [EMAIL PROTECTED]
