On Sun, Jul 06, 2003 at 10:14:47PM +0200, Manuel Bouyer wrote:
> I think I already looked at it. The problem is that the MI code contains some
> hacks to please some brocken ATAPI devices. I suspect one of these cause
> the problem, but I don't know which one. Maybe just reading a register
> to clear a status on the drive would be enouth.

I have this problem. It started after I added a cheap cdrw to a
machine with some SiS allinone-for-amd chipset, and is caused by
that device. 

It looks to me like the last probe-time transaction never completes,
and the bus stays active.  There's nothing else on that ide channel.

The light stays on until I do something like cdrecord -checkdrive
(or, probably, anything else that would tickle the drive over the
bus).  At that time I get (from memory) an ata command timeout in
syslog, the checkdrive completes normally, and the light goes out.

The machine in question is presently in a shipping container, but
I can test it again in more detail next week when it arrives.

--
Dan.