Tim Parker wrote:
>
> Hi all,
>
> [EMAIL PROTECTED] wrote:
> >
> > Simen Thoresen wrote:
> > >
> > > Hi!
> > >
> [snip background]
> > >
> > > The symptom. When no devices are attached to either 7890, the box boots
> > > nicely from a Slackware 4.0.0 install floppy. No disks are available tho.
> > >
> > > When /any/ disk is attached to either aic7890, the box crashes totally
> > > while initializing the kernel (right after the PIIX4 not supported
> > > message). The boot disk uses kernel v 2.2.5, compiled with aic78xx support
> > > (non smp).
> > >
> > > The crash is total - no aiee. no panic. Just a black screen. I need to do a
> > > hard poweroff on the powersupply to turn the power off.
> > >
> > > When no devices are attached, both 7890 are detected and initialized
> > > without comments.
> [snippity snip]
> > > Any comments?
> >
> > This is an FAQ by now. Use the option aic7xxx=no_probe when booting the
> > machine if the aic7xxx driver is compiled directly into the kernel or if
> > trying to do an install then make sure the aic7xxx driver is loaded with the
> > option aic7xxx=no_probe specified on the module insertion command line, aka
> > "insmod aic7xxx aic7xxx=no_probe".
> >
>
> ..I wish it was. By happy accident I was searching for (and not finding
> yet..)
> reasons why the install I was trying on a Tyan 1696DLUA kept screwing up
> when
> the boot check got the drive(s) on the aic7895 [status read errors
> mainly, and
> then hanging completely], when this mail came through. Thanks for the
> reply
> Doug - i'll try it this evening.
The option above shouldn't affect your motherboard. It is specific to the
Dell PowerEdge 6300 motherboards (and other machines that use the SC450NX
chipset, which is the Quad Xeon chipset). All it does is disable the probes
for the EISA/VLB aic7xxx controllers since the process of writing to/reading
from the card signature I/O locations causes the SC450NX chipset some
problems. You system is more likely to need a BIOS update from Tyan. Check
their web site. I think your motherboard needs *at least* version 1.17 of the
motherboard BIOS in order to work properly.
> Everything else looks OK from a h/w
> side,
> 2 x Celeron 366/128 on PC100 slockets (boots up fine) - although one
> thing
> spooked : when the kernel reports the bogomips, I get around 2.75. This
> was at
> the end of the (long) evening so I put it down as something to do
> tonight.
> I'm presuming this is just a spurious cache enable problem and will
> troll
> through the BIOS and docs again later - unless i'm going into complete
> no-brain mode and anyone has any comments.
Hmmm...did you modify the Celeron processors? I assume you did becuase they
don't tend to do SMP at all without modifications on the chip. In any case,
the MTRR code in the kernel should solve your bogomips discrepancy. That's
another BIOS bug (the BIOS isn't initializing the second processor
completely).
--
Doug Ledford <[EMAIL PROTECTED]>
Opinions expressed are my own, but
they should be everybody's.
-
Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/mentre/smp-faq/
To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
