Subject: Re: filthy, evil bochs patches
To: Julian Assange <proff@iq.org>
From: Matthias Drochner <drochner@zel459.zel.kfa-juelich.de>
List: port-i386
Date: 10/27/1999 15:03:45
proff@iq.org said:
> Thus far I've managed to emulate a 1.4.1 install without problem.
The bochs which is now in pkgsrc could run an installation
floppy too, with some minor problems remaining:
-The isapnp probe stepped on VGA registers, leading to panic.
With the console driver changed to wscons this doesn't happen
because the wscons drivers participate in the address space
accounting. I'd consider the bochs panic a helpful development
aid here and wouldn't hack around it.
-If no harddisk is specified, one with 0 sectors is seen.
Needs a thorough look at the wd probe...
-The BIOS call which does the microsecond delay is not supported,
this breaks the countdown of the bootloader.
-Some timer problem - wskbd's software key repeat is too fast.
To your patches:
harddrv: My patch in pkgsrc is similar, we should try to understand
the underlying problem however.
keyboard: The offending command is the beginning of a "set scanset"
sequence. This is currently not supported by bochs. It happens
to work if it is simply ignored because our console drivers
use scanset 1 only. My patch doesn't ACK the command to
discourage the driver from sending the second byte.
vga: It's a hack, I wouldn't do this. Or perhaps print a warning
and continue.
best regards
Matthias