> I ran into a possibly related problem trying to get my StyleWriter
> driver (which currently works with the old StyleWriter 1) to work with
> the Color StyleWriter 2400.  Basically, talking to the printer over the
> serial port from a userland program causes the machine to hang when the
> printer should start sending data back.  Unplugging the serial cable
> causes things to start happening again for a few seconds, but the
> machine hangs again a short time later.  Rebooting without unplugging
> the cable also causes the machine to crash again sometime during the
> boot sequence.
> 
> (BTW, this thing has some wierd energy-saving behavior.  It turns
> itself off after being idle for a certain amount of time, and turns
> itself on again when it starts getting serial data.  It might make
> sense for such a beast to only start up an external clock signal when a
> print job starts, I suppose...)
> 
> If the StyleWriter were pulling the same trick as these laser printers,
> could it cause the symptoms I described?

Yes. I'd also suspect it to be more-likely on a Color printer as it has
to transfer much more data to the printer.

> Unfortunately, I don't have access to anything like an oscilliscope
> that I could use to look into this.  If I can compile -current kernels,
> is there anything I can do to find out if this might be my problem?
> (i.e. do you have test code you could send me that would tell us
> anything useful?)

If you can compile test kernels, try removing the ZSWR15_DCD_IE from
dev/ser.c (around line 262, I think). This change is basically what I'd
do to "support" these printers (plus set the code up to use the external
input). If it stops hanging, we know what's wrong.

Note: After doing this, you won't be able to notice carrier loss on
the modem. But it's just a test. :-)

Take care,

Bill