On Tue, 18 Jan 2000, Jonathan Stone wrote:

> Or is the biosboot_com0 I built getting horribly confused by the BIOS
> redirecting "normal" (VGA/kbd) console to the serial port?

Given how easy I got the serial console working on an old 486, and how
much grief you are experiencing, this is my first suspect.  It's darned
sweet of Intel to finally admit the existence of serial consoles, but
since their firmware architecture has no way of passing console device
information to the bootblocks it sounds like this ill-conceived
``feature'' does more harm than good.  Exactly how is a modern
rasterconsole OS supposed to do serial output under the Intel Kludge?  And
does the kludge work at all after the OS switches into protected mode and
tells the BIOS to kiss off?  Or are we talking about some INT13h serial
console for DOS and maybe Netware?

You might try using CONSDEV_COM0 instead of CONSDEV_COM0KBD or
CONSDEV_AUTO, and don't use COMCONS_KEYPRESS.  These options are set
according to BASE in biosboot/Makefile.  This will make sure your
bootblocks are dead-set on using serial console, and are not trying to
autodetect.

However, depending on how your funky chipset works, I don't see any
guarantee that the Intel Kludge will be compatible with NetBSD's sane
serial console architecture.  Are you even certain the serial hardware is
available to the operating system, when the Kludge is turned on?  Or is it
masked out by the chipset/BIOS conspiracy?

Maybe my condemnation is a bit premature, but it sure looks to me like yet
another case of an advertised ``feature'' turning out to be completely
useless due to poor implementation.  I'm really getting sick of these
charletans (sp?).

-- 
Miles Nordin / v:+1 720 841-8308 fax:+1 530 579-8680
555 Bryant Street PMB 182 / Palo Alto, CA 94301-1700 / US