2005-04-11 20:20:23

by Shawn Starr

[permalink] [raw]
Subject: Re: [ACPI] [2.6.12-rc2][suspend] Suspending Thinkpad: drive bay light in S3 mode stays on

Sure, I suppose you can, but most suspend tools just
echo stuff to /sys (or still /proc/acpi/sleep) which
makes it harder to script it. Besides, when a laptop
goes into suspend to RAM there should be no extra
power on except a Moon or some other icon.

That said, the ACPI thinkpad extras was designed to do
all of this so why shouldn't the driver do S3 suspend
if it hooks into it already?

Shawn.

--- Matthew Garrett <[email protected]> wrote:
> On Mon, 2005-04-11 at 16:03 -0400, Shawn Starr
> wrote:
> > I notice in Linux and in XP the drive bay light
> > remains on while the laptop is in suspend-to-RAM.
> I
> > know the ACPI thinkpad extras added to the kernel
> > recently can turn this off. I wonder if we can/or
> need
> > to write hooks to turn the light off so to
> conserve
> > power when we're in S3
>
> Just disable it in your suspend script. There's no
> reason to push that
> sort of policy into the kernel.
>
> --
> Matthew Garrett | [email protected]
>
>


2005-04-11 20:35:37

by Matthew Garrett

[permalink] [raw]
Subject: Re: [ACPI] [2.6.12-rc2][suspend] Suspending Thinkpad: drive bay light in S3 mode stays on

On Mon, 2005-04-11 at 16:17 -0400, Shawn Starr wrote:
> Sure, I suppose you can, but most suspend tools just
> echo stuff to /sys (or still /proc/acpi/sleep) which
> makes it harder to script it. Besides, when a laptop
> goes into suspend to RAM there should be no extra
> power on except a Moon or some other icon.

Most suspend tools are depressingly stupid. That's not a good reason to
push functionality into the kernel. The vast majority of hardware won't
work with that approach at the moment.

> That said, the ACPI thinkpad extras was designed to do
> all of this so why shouldn't the driver do S3 suspend
> if it hooks into it already?

Because, well, strictly it wasn't. The LED control functionality in the
IBM-acpi code exists because it exposes methods that are used by the
BIOS in normal usage. It gives some degree of extra flexibility -
there's no point in removing that for the sake of having one fewer line
of shell in a suspend script. I might want the LEDs to be in different
states depending on what triggered the suspend.

--
Matthew Garrett | [email protected]