We're using the libata stack nowadays. Haven't heard anything from bugreporter/
contributer so closing this issue. Please feel free to reopen this issue if you 
think it's still valid.

thx && regards
-mika-

Sadly I still haven't had the time and opportunity to investigate on this 
issue. Any news from anyone of you?

thanks && regards,
-mika-

Ok, thanks for the answer, Mark. I will try to reproduce it on my own here so I 
can bug upstream about it, I'll tag it with "need-eg" in the meanwhile.

regards,
-mika-

Short answer is no, sorry, but thank you for asking.

After the 2.6.20 fiascos we cannot afford more debugging for Linus who needs to
fix his development philosophy.  The last report I can give is 2.6.20.15 in grml
1.0-1 with CONFIG_USB_SUSPEND turned off.

Upper-case -Y doesn't work:  it hangs the PC when issued.

Lower-case -y still behaves as before,
http://bts.grml.org/grml/msg439

Truecrypt mount/dismount on USB causes IDE drives to re-spin needlessly, too,
independent of shutdown.

Linux doesn't get that there's no reason to spin up without IDE file activity. 
So the kernel has missing and/or corrupted logic.

Therefore I recommend closing the grml bug and sending it upstream to kernel
devs.  Link back here for reference.

Mark, did you have any chance to test this issue with a recent kernel version 
and the behaviour is still reproduceable?

thx && regards,
-mika-

Lower-case -y stresses the drives.  It spins them back up just to spin them down
again.  (I tested it, you should be able as well.)  That is the worst stress for
a drive.  At shutdown drives are re-awakened (spun up) and immediately shut down
(spin down).  You hear the procedure even without a screen.

Please report kernel flaws upstream.  This may be one.  Reports have more
credibility from you than me.

Since IDE drives are already dismounted and spun down, there is no reason for
system shutdown to re-activate them.  It is unnecessary stress on the drives and
time-consuming:  a real bug.

Right now capital -Y is safer than lowercase -y because -Y does not re-activate
drives.  The manual workaround is that when "interrupt lost" message scrolls
along, do a manual power-off.  Not ideal, but better for the drives than -y.

The issue seems to be a kernel problem. Asking in #grml regarding the problem 
amne suggested to test 'hdparm -y /dev/....' because this usually works much 
better than the '-Y'. Mark, does this work for you and solve your problem?

Booting grml from USB, we often turn IDE drives off:

# hdparm -Y /dev/hda /dev/hdb /dev/hdc

This works fine but causes "shutdown -h 0" to fail with "lost interrupt"
message.  Hard manual power off is necessary to shutdown.  Note that restart
still works ("shutdown -r 0").

There are no file systems mounted on the IDE drives when we turn them off,
that's not the issue here.