libata and legacy ide pcmcia failure

[Robert de Rooy]

Mark Lord wrote:
> I still don't see much evidence that interrupts are actually 
> functioning here.
> It would be good to see /proc/interrupts before/after libata tries to 
> talk to it.
>
> Let's assume for the moment that interrupts are b0rken.
> The legacy IDE driver can talk to such devices completely without 
> interrupts,
> if the IDE polling patch (below) is applied.
>
> So, Robert:  could you try again with the old IDE driver,
> except apply this patch to it first?  This will give valuable info.
I tried the patch, but a get a kernel panic on boot with it. I can try 
to get a screenshot of it tomorrow.