On Tue, May 18, 2004 at 03:37:01PM -0400, Allen Briggs wrote:
> On Tue, May 18, 2004 at 07:29:33PM +0000, Matthias Scheler wrote:
> > So it was broken? Was that a 3Com 3C996B-T or another BCM 570x based card?
> 
> I think it was a 3Com 3C996B-T -- BCM 5701-based.  I have it around
> here somewhere.  Worked in x86, failed on an XScale.  It just never
> seemed to get any interrupts, if I recall correctly.  A lot of the
> initialization is done by writing to the card, so it was hard to
> see how much of that was actually working.  I think I could assign
> an address to the card, but it would soon start getting timeouts,
> fail to stop properly, and just be pretty hosed.  As I recall, I
> put some code to printf() to the console in the interrupt routine
> and never saw those.  I also seem to recall that calling the
> interrupt routine from the timeout was not fruitful, either, but
> I'm not so sure about that anymore.  It's been a while since I
> played with it.
> 

In any case I'd start checking if interrupt routing set properly,
adding printf to card's interrupt handler and MD interrupt receiving
routine it shouldn't be hard to find out. 
With Matthias' dmesg 'ivec 0' doesn't look right, I'll look why so,
but it'll take some time as I don't have card.

	Andrey