> I have had the same problem with all recent versions of NetBSD and
> the PC I have as the console.  When the console reboots, some where
> in its hardware reset sequence, the serial ports send a BREAK.

Are you sure of that?  I've seen some "do special things on BREAK"
devices that more accurately can be described as "do special things on
any framing error".  (I don't *think* SPARC consoles are among them,
but don't recall checking to be sure.)

> A break has nothing to do with the DTR line.  A break is formed by
> holding the TX line inactive for a second or two.

Active, actually, I think, and the duration depends on the baudrate of
the break - I think a normal time for a break is on the order of 500ms,
which can be less insufficient at low baudrates but is plenty at high
baudrates.  (As for active/inactive, I'm assuming that by inactive you
mean the line state during normal inter-character time.)

> Anyway, getting back to the point, I would like to add an option in
> the kernel which will ignore the console break.  I looked thru the
> source code a long time ago, and didn't see any good way of doing
> this except by commenting out the call to the debugger.

Wouldn't it be enough to just ignore the break bit in the chip's status
register?

Of course, you then lose the ability to *deliberately* drop into
ddb/monitor.

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B