Re: Some tests with my Macintosh Centris 650
On Sat, 24 Jan 2015, John Paul Adrian Glaubitz wrote:
> On 01/24/2015 12:01 AM, Finn Thain wrote:
> > That's a very early crash. It could be caused by a NuBus slot
> > interrupt being asserted, since these are non-maskable IRQs on some
> > models. This can happen with SONIC ethernet, prior to the Linux
> > macsonic driver being loaded.
I should add that it is possible to disable all slot interrupts
altogether; but impossible to disable them individually. Your
Centris/Quadra 650 is one of the models that suffers from this limitation.
> >
> > Try booting MacOS with no extensions loaded (hold down shift key when
> > you hear the boot chime, and release it when you see the words
> > "Extensions disabled").
>
> Ok, I will give that a try next week. The Mac is located in my office
> and I am not going to work over the weekend :).
>
> I would actually buy a serial cable to help with the debugging. Is there
> any particular cable that you can recommend?
I've soldered up mini-DIN8 to DB9 cables in the past but it is fiddly and
time consuming. (The Mac serial ports are RS-422 but they become RS-232
compatible by grounding the appropriate pin.) In my experience it is
easier to find an old serial printer (aka cross-over) cable: these are
male to male, mini-DIN8 to mini-DIN8, DTR to DTR.
A lot of low-end Macs were bundled with low-end Apple or HP inkjets, with
such a serial cable. An equivalent cable was used with Newtons and
QuickTake digital cameras. They are common, and can also be used to create
a two-node localtalk network between any two macs.
The Keyspan adapter, part no. USA-28X is the one I use; it's well
supported on every operating system of interest to me. You can find all
this gear for sale on ebay.
But this bug you are chasing looks to be earlier than any serial port
output; if your kernel has CONFIG_EARLY_PRINTK enabled, then the screen
will be painted black before anything is sent to the serial port. Any
character sent to the serial port will be simulataneously painted on the
screen in white.
It looks to me like the bug you are chasing is an early unhandled
interrupt, and may be a bootloader bug. If you've already tried unpacking
the gzipped kernel, and tried disabling AppleTalk (see Chooser in the
Apple menu), then lastly I'd try disabling all extensions (restart with
shift key pressed).
--
Reply to: