Re: boot-floppies 2.2.11 for Alpha building now, testers needed
Adam C Powell IV <firstname.lastname@example.org> writes:
> My mistake. I was trying to install on a serial console machine, it boots fine on an LX164 with a
> video card.
Yeah, that's a known issue. The problem is that the kernel does not
detect whether it has been booted from a serial console or not, and so
it will send output to the screen unless you give it a "console=ttyS0"
(or whichever serial port) argument on the kernel command line.
The goal is eventually to do this guessing either in the bootloader or
the kernel. This is tricky to do in aboot because the CTB (console
terminal block) is seriously underspecified. (I speculate that this
is because many (most?) implementations of SRM take input on the
serial port *and* the keyboard if both are present, though they only
sends output to one at a time). It should be pretty easy to do in
MILO since MILO is actually a console rather than a bootloader :)
> The 2.0.36 slink kernel showed startup message on the serial port, can that option be turned on for
> potato too? I realize it may require a different root image, as did slink...
Hm, that's interesting. So 2.0 would automatically detect a serial
console? Or was there a separate boot disk for it? (I seem to
remember that there was no "console on serial port" config option in
> --- /mirrors/debian/dists/potato/main/disks-alpha/2.2.11-2000-04-21/README.txt Fri Apr 21 14:22:02 2000
> +++ README.txt Tue Apr 25 10:11:03 2000
I'll apply this ASAP.
David Huggins-Daines, Senior Linux Consultant, Linuxcare, Inc.
Linuxcare. Support for the revolution.