speedtch speedtch.c,1.44,1.45

[Duncan Sands]

Hi Roman,

> Well I just thought there might have been some known correspondence
> between the info the modem reported about itself and the way the driver
> should treat it.  The comments in the driver also suggest that the
> usability of the interrupt endpoint for status updates depends on the
> revision, so I thought bcdDevice can be used here too.  But I agree it's
> always hard to know for sure without specs.

for the interrupt endpoint you are right.  On the other hand, you can
always just try the int endpoint and fall back to polling if it fails
(and I think that it is wiser in general to probe for a feature, rather
than relying on revisions etc).  To be honest, I'm tempted to just poll
always, but dwmw2 would leap on me if I did :)

> FWIW I at first made cxacru treat the modem as an older revision based
> on presence of the user-supplied extra firmware file.  Now that I've got
> statistically significant confirmation that the distinction can be made
> by the product ID, I've removed this configuration option, and now
> associate proper values with the corresponding usb_device_id via
> .driver_info.

Well, if no-one ever complains that is fine.  The reason for adding the
speedtch altsetting parameter is that someone really does need it.

Ciao,

Duncan.