Subject: Help us fix i386 ACPI suspend/resume support
To: None <current-users@netbsd.org, port-i386@netbsd.org>
From: Perry E. Metzger <perry@piermont.com>
List: current-users
Date: 06/20/2006 12:04:22
As some of you may have noticed, we've been making rapid progress on
the kernel portions of ACPI suspend/resume support.

However, many devices still lack power hooks to save the device state
before sleep and restore it after.

If you're able to, please submit tested patches for any devices that
you have that still lack power hooks.

An explanation of how to test out the ACPI suspend code is here:

http://mail-index.netbsd.org/current-users/2006/06/17/0014.html

Most devices that call powerhook_establish(9) are good examples of
existing code that properly restores device state.

Code from FreeBSD and other open source OSes may also be a good source
of information for you on how particular devices can have their state
properly saved and restored.

When your machine is working "perfectly", "lspci -xxx" (lspci is in
pkgsrc/sysutils/pciutils) should produce identical output before and
after an S3 suspend. We don't actually have any machines working quite
perfectly yet, so don't be concerned if you can't fix *all* the
differences between before and after on your machine. Submitting
patches that make one or two of your devices work correctly after
suspend is a big help -- you need not get the whole machine perfect to
be of substantial help to the effort.

Please note that the current sysctl based hack for entering suspend is
going to go away when we come up with a proper userland interface --
it is mostly there so people can test suspend under -current.

And remember, the NetBSD project is *you*. We depend on code
contributions from our users to make improvements happen. Many hands
make work faster -- if you can help fixing even just one device, it
will make the goal of NetBSD working well with ACPI that much closer
to reality.

Perry
-- 
Perry E. Metzger		perry@piermont.com