On Thu, Dec 20, 2007 at 02:32:26PM +0100, Matthias Drochner wrote:
> For me, the appended patch helps. Don't know what this
> "dying" stuff is good for. In our situation it is
> clearly wrong.

I think it is supposed to make the detach case work, but it is real mess
and I think quite wrong in many parts. Can you just make sure that the
device doesn't access the hardware when it is suspended or disabled and
that detach is removing the interrupt handler early enough so that this
can go away?

Joerg