In message <20000118223838.J21588@fundy.ca>David Maxwell writes
>On Tue, Jan 18, 2000 at 05:45:46PM -0800, Jonathan Stone wrote:

>
>Sorry, perhaps I'm being thrown by the choice of mailing-list. What OS
>Version are you setting up here? 

A late-december-ish -current. It never crossed my mind to send
something so horribly x86-specific to current-users; thatd be
net.abuse to every other port. In retrospect I can see how it'd be
confusing.


>Got it now I think - you're running -current. What serial console
>'policy' did you build your bootblocks with? ala boot_console(8)

Whimper. I loked at all the man pages execpt boot(4).  None of them
mentioned boot_console, nor did GENERIC. I UTLSed and built
biosboot_com0 using the makefile in arch/i386/stand/biosboot_com0.

Which (it seems) has the "hit a key" test backwards for any sane use
of a serial console. I want serial consoles because I want to bput the
boxes in a wiring closet, to get logs of the console if they crash or
wedge too hard to log tan error, etc., etc.  For that setup, I ant to
use a serial console _unless_ someone types on the keyboard during
reboot, in which case I'd (maybe) like the bootblocks to switch to the
graphics console/pckbd.

Seems like the test for the provided "biosboot_com0" bootblocks is
exactly inverted: normal "internal" console, unless someone hits a key
on the serial console. I cant think of _any_ sane use for that.  Whoat
on earth is that _for_? Debugging graphic-console I/O?  Shouldn't the
default be tailored to real serial consoles-- i.e., switch the sense
of the test, for other than CONSDEV_AUTO?

And forcing XON/XOFF with DIRECT_CONSOLE is just ... broken, given the
existence of BIOS serial consoles which want RTS/CTS hardware handshaking.


Sigh.  Miles is right. What we really need is a way to ask the BIOs
what its console deviceis, and I have no clue how to do that.

Meanwhile, the console(4) manpage is woefully out of date.  Shouldn't
it .Xr to boot_console() and warn that its for the old, pre-1.4,
bootblocks?