Re: ACPI problems with HEAD and IBM ThinkPad T23

----- Original Message ----- 
From: "Kevin Oberman" <oberman_at_es.net>
To: "Matthew Emmerton" <matt_at_gsicomp.on.ca>
Cc: <freebsd-current_at_freebsd.org>
Sent: Wednesday, June 04, 2003 10:21 AM
Subject: Re: ACPI problems with HEAD and IBM ThinkPad T23


> > From: "Matthew Emmerton" <matt_at_gsicomp.on.ca>
> > Date: Wed, 4 Jun 2003 10:00:53 -0400
> > Sender: owner-freebsd-current_at_freebsd.org
> >
> > This is a multi-part message in MIME format.
> >
> > ------=_NextPart_000_0069_01C32A80.2F346810
> > Content-Type: text/plain;
> > charset="iso-8859-1"
> > Content-Transfer-Encoding: 7bit
> >
> > I cvsup'd from 5.0-R to HEAD and enabled ACPI to see if anything had
changed
> > in the past 6 months.
> > Indeed, things have, and a few workarounds I had implemented before were
no
> > longer needed.
> >
> > However, I still got a host of errors from ACPI upon boot, and some
hacks to
> > device.hints were still neccessary to keep my laptop functional through
a
> > boot.
> >
> > Attached are my device.hints file and a copy of the dmesg.
> >
> > In particular, the things that bother me are:
> > 1) All of the "Method execution failed" errors
> > 2) The fact that I have to have a device.hints entry in order to see
console
> > message upon boot.  (Without this, my system boots up silently and after
> > about 30 seconds, displays the familiar login prompt.)
>
> On my T30, I get similar messages at startup. Some patches currently
> under discussion on the ACPI list may (or may not) eliminate them. The
> messages that the patches (if committed) eliminate are bogus and yours
> are of the type to be in this class.

Ahh.  Good to hear.

> Take a look at the output of 'sysctl hw.acpi'. The T30 does not support
> S1, so you need to re-define anything using S1 to use something else or
> nothing at all.

Same here.

> Does S3 work properly? When I suspend my T30, I lose my USB ports until
> I reboot and suspend does not turn off the backlight on my LCD. (This
> provides a wonderful demonstration of literal bit rot.)

I actually make it a habit to never use power-save modes on my T23 when I'm
using FreeBSD - I got bit badly once.  When I did try it, I lose my
backlight, my ethernet (since I was using my Cisco Aironet 350 pccard at the
time) and IIRC the system locked up shortly thereafter.

> The requirements for hints to get the console to work at boot are
> surprising. I have never seen this on my T30.

As pointed out in another note, this is my fault - I trashed the default
device.hints and made a "minimal" version.  In my case, it requires entries
for sc0 and fd0.

> How recent is your BIOS? The latest release was on March 6, 2003. I
> found significant improvement in ACPI after upgrading my firmware!

I upgraded to the latest BIOS before sending this report - the dmesg output
was the same between my older BIOS (December, IIRC) and the latest BIOS.

--
Matt Emmerton