Re: ACPI S1 state

[Joerg Sonnenberger <joerg%britannica.bec.de@localhost>]

On Wed, Jun 11, 2008 at 03:45:04PM -0500, David Young wrote:
> On Wed, Jun 11, 2008 at 02:13:13PM +0200, Joerg Sonnenberger wrote:
> > On Wed, Jun 11, 2008 at 03:43:36AM -0500, David Young wrote:
> > > This patch helps my ancient HP Pavilion N3270 enter/leave S1 state.
> > > (The N3270 does not have an S3 state.)  I can put the system to sleep
> > > with 'sysctl -w machdep.sleep_state=1', and I can wake it by tapping
> > > the power button.
> > 
> > The first part assumes that interrupts where not already disabled. I
> > don't think we should make that assumption. So if we want to go this
> > way, disable interrupt should actually return the current state.
> 
> My x86 assembly is rusty, but let me see what I can do.

Just grep for the disable function in the existing code, no assmebly
involved.

> 
> > > If I do not put the devices to sleep, then the system hangs instead
> > > of waking up.  It seems to have something to do with interrupts from
> > > PCI devices.
> > 
> > Can you try what happens if you mask all ioapic/pic pins before entering
> > S1?
> 
> I will try that if it is easy, just for curiosity's sake. :-) Is there
> any single routine that I can call to do that?

ioapic or pic?

Joerg