Dag-Erling Smorgrav wrote:
> Robert Nordier <firstname.lastname@example.org> writes:
> > The -P is completely handled by boot2. The effect is to set the
> > bootblock -D (dual) and -h (serial) options, if no keyboard is found.
> Unfortunately, the keyboard probe returns false positives on far too
> many motherboards for this to be of any use.
The new bootblocks don't do a keyboard "probe" as such. The present
test is whether an "enhanced" keyboard is installed, as determined
by the BIOS (any standard i386 keyboard qualifies as enhanced,
except -- in some cases -- when running under i386 emulation).
Whether this works is really a function of the BIOS rather than
the hardware. It will inevitably require the keyboard is (un)plugged
before POST, and some BIOSes may require that the presence or
absence of the keyboard is explicitly noted in BIOS setup.
The approach certainly does work in a substantial number of cases:
and, where it does work, it is highly likely to work reliably (since,
if any code understands the given hardware, it is likely to be
However, if there is a concensus that the approach used in the old
(biosboot) bootblocks, which attempt to interrogate the keyboard
controller directly, gives better results in practice, I don't have
any problem with changing to that way of doing things.
A primary constraint, though, is that only ~100 bytes are available
in boot2 for keyboard detection logic, by whatever means.
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message