Hi!
I heard Patrick Mochel has already developed the code for a new
device/power management syste. So I'm wondering where this code is. I like
to intergrate it into my CVS tree. Thanks.
On Fri, 16 Nov 2001, James Simmons wrote:
> I heard Patrick Mochel has already developed the code for a new
> device/power management syste. So I'm wondering where this code is. I like
> to intergrate it into my CVS tree. Thanks.
There were some bits at ftp://ftp.kernel.org/pub/people/mochel last time
I looked.
regards,
Dave.
--
| Dave Jones. http://www.codemonkey.org.uk
| SuSE Labs
On Fri, 16 Nov 2001, Dave Jones wrote:
> On Fri, 16 Nov 2001, James Simmons wrote:
>
> > I heard Patrick Mochel has already developed the code for a new
> > device/power management syste. So I'm wondering where this code is. I like
> > to intergrate it into my CVS tree. Thanks.
>
> There were some bits at ftp://ftp.kernel.org/pub/people/mochel last time
> I looked.
That's where it is. The most recent drop is the -1115 patches.
I have some other patches converting drivers to use the new driver model,
and I will push those out soon..
However, that code does not have all the pieces to power management in it.
It simply provides a unified device tree based on locality; the framework
for doing things like system suspend.
The power management transitions live in the most recent ACPI code, which
you can get from Intel:
http://developer.intel.com/technology/IAPC/acpi/index.htm
and is specific to ACPI-enabled platforms.
-pat
On Fri, 16 Nov 2001, Patrick Mochel wrote:
> > There were some bits at ftp://ftp.kernel.org/pub/people/mochel last time
> > I looked.
> That's where it is. The most recent drop is the -1115 patches.
I meant /pub/linux/kernel/people/ of course, but I'm sure James was
smart enough to figure that out 8)
> The power management transitions live in the most recent ACPI code, which
> you can get from Intel:
Something I'm curious on wrt to this new work. Would it make sense for
these callbacks to get called before/after APM suspend as well as ACPI ?
(I'm thinking of older pre-ACPI compliant boxes).
Saving state of devices etc seems a logical thing to do.
regards,
Dave.
--
| Dave Jones. http://www.codemonkey.org.uk
| SuSE Labs
> Something I'm curious on wrt to this new work. Would it make sense for
> these callbacks to get called before/after APM suspend as well as ACPI ?
> (I'm thinking of older pre-ACPI compliant boxes).
>
> Saving state of devices etc seems a logical thing to do.
Yes, it's entirely possible, and seems like a good thing to do.
The original motivation behind it was to replace struct pm_dev and the
callbacks for it. I haven't looked it into replacing them in the APM code
or the power management code for other architectures, but it shouldn't be
that painful. Hopefully.
-pat
Hi!
> > The power management transitions live in the most recent ACPI code, which
> > you can get from Intel:
>
> Something I'm curious on wrt to this new work. Would it make sense for
> these callbacks to get called before/after APM suspend as well as ACPI ?
> (I'm thinking of older pre-ACPI compliant boxes).
With apm, bios/hw should do state saving itself. Doing it from os only
makes sense in order to work around bios bugs. It probably should not be
done as default.
Pavel
--
Philips Velo 1: 1"x4"x8", 300gram, 60, 12MB, 40bogomips, linux, mutt,
details at http://atrey.karlin.mff.cuni.cz/~pavel/velo/index.html.