Re: Alpha XL300 (aka XLT) problems

[Robb Romans <rr44607@swt.edu>]

On Thu, 28 Jan 1999 Jay.Estabrook@digital.com wrote:
>
> XLT and ALCOR are not the same, and it is just that little difference that
> makes all the difference... ;-}
> 
> Running an ALCOR kernel on a true XLT will cause additional bits
> in the interrupt summary register to be treated as real interrupts, causing
> im most circumstances what looks like a hang but is really just taking
> continuous bogus interrupts.
> 
> The last 2.0.35 kernel alpha sources/patches I released may have attempted
> to determine dynamically which of XLT or ALCOR was really present. I received
> reports that this detection was not accurate, so, I backed that out of my
> patches (which I'm not sure I've ever released).
> 
> As for 2.2.0 and the pre's, I'm not sure what is in there any more WRT XLT
> and ALCOR, especially when attempting "generic". I'll have a look and see
> what can be done about bugfixing for 2.2.x.
> 
> Is anyone willing to try some patches, see if we *can* get the dynamic 
> detection stuff to work?

Sure, fire away.

I have an XLT300, running fine with 2.2.0pre9. It identifies itself as a
"Bret" on bootup. I initially had problems with no visible screen after
"jumping to linux.." with anything newer than 2.1.131 but this was solved
by upgrading from egcs 1.03a to 1.1.1. The machine did not really hang, as
I could login from a serial terminal.

I'm using the milo from RH5.2, though an earlier milo, and one from debian
made no difference on the invisible console.

	Robb

-- 
Robb Romans
Power by Linux.Alpha
Southwest Texas State University
http://taz.cs.swt.edu/~rr44607