>So I'm wondering how many of the people having fifo overruns while
>ppp'ing are actually getting them on the port in use (remembering that
>the modem is zs0a)?

Well, I for one often get fifo overruns (the hardware kind) on zs0b (the
printer port) even when I'm not using the printer, even when the printer is
turned off!  I've found that disconnecting the printer cable from the Mac
takes care of most of the overruns.

I still regularly get other strange things happening.  I use the modem
(zs0a) port for my ppp link to the world, and often I find that I have to
turn the modem off and on a couple or three times after starting pppd
before it will start sending the init and dial strings to the modem.  Then,
things will be going along nicely, and I'll try to reply to an E-mail I'm
reading, and the dratted serial port will stop interrupting altogether.
Once this happens, pppd is irretrievably hung, and not even a kill -ABORT
will get rid of it - it hangs in an "exit pending" state, and only a reboot
will recover it.

This exact behaviour has existed since long before Bill's ser.c driver, so
I suspect that there's something else somewhere that's screwing things up.

One more piece of info that may or may not be relevant:  It has been
several month's of -current kernels since the ite console code has been
able to ring my bell.  It used to at least include the bell sound during a
shutdown sequence, but not any more.  And recently, with 1.1Alpha and 1.1,
I've actually had a couple of hangs during shutdown.  I've caught a couple
of them and pressed the debugger switch, and gdb's trace always includes a
reference to an asc (Apple Sound Chip) routine.

------------------------------------------------------------------
| Paul Goyette            | PGP Public key available on request  |
| Paul@pgoyette.bdt.com   | Fingerprint: 0E 40 D2 FC 2A 13 74 A0 |
|                         |              E4 69 D5 BE 65 E4 56 C6 |
------------------------------------------------------------------