Suspend/resume regression with latest ACPI code

[Kevin Oberman]

> Date: Thu, 29 Apr 2004 20:16:18 +0200 (CEST)
> From: Lukas Ertl <l.ertl at univie.ac.at>
> Sender: owner-freebsd-acpi at freebsd.org
> 
> Hi,
> 
> I'm seeing a suspend/resume regression with the latest ACPI code.  Last
> known good kernel was from April 12th, so that was before the import of
> ACPI CA 20040402.
> 
> The box in question is a ThinkPad T40, and with the old kernel, I could
> perfectly suspend/resume my laptop.  Now, only the first suspend/resume
> cycle works; if I try afterwards, it immediately wakes up again after
> suspending.  No errors or otherwise suspiscous output.
> 
> I'm not sure if this could be related to the PCI changes from Warner.

This is a me-too (tm-AOL). In my case it's an IBM T30. Looks like the
problem may be common to multiple IBM systems. I've seen some other
reports of this on IBMs, but I'm not sure if I've seen reports on other
platforms.

I'm seriously suspicious that this is really a PCI issue in the power
stuff. I suspect that something is not being properly reset on resume
and that the next attempt to suspend fails because something is not in
the right state. It gets all the way to the end of the suspend, though,
before it decides that it really can't.

I have a couple of other problems that I think are really PCI issues
(sound and the loss of IRQ11 after a while) and I think that this should
also go on the list.
-- 
R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman at es.net			Phone: +1 510 486-8634