Re: [RFC] PCI: Runtime power management

From: Jesse Barnes <jbarnes@virtuousgeek.org>
To: Matthew Garrett <mjg59@srcf.ucam.org>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>,
	Alan Stern <stern@rowland.harvard.edu>, Greg KH <gregkh@suse.de>,
	LKML <linux-kernel@vger.kernel.org>,
	Linux-pm mailing list <linux-pm@lists.linux-foundation.org>,
	linux-pci@vger.kernel.org, linux-usb@vger.kernel.org
Subject: Re: [RFC] PCI: Runtime power management
Date: Fri, 14 Aug 2009 10:37:19 -0700	[thread overview]
Message-ID: <20090814103719.79e99d91@jbarnes-g45> (raw)
In-Reply-To: <20090813002925.GA2532@srcf.ucam.org>

On Thu, 13 Aug 2009 01:29:25 +0100
Matthew Garrett <mjg59@srcf.ucam.org> wrote:

> I got a fixed BIOS from Dell and have been able to get this working
> now. It seems entirely happy with USB, but I'd like some sanity
> checks on whether I'm doing this correctly. There's certainly a
> couple of quirks related to setting the ACPI GPE type that would need
> a little bit of work in the ACPI layer, and it breaks ACPI-mediated
> PCI hotplug though that's easy enough to fix by just calling into the
> hotplug code from the core notifier.
> 
> This patch builds on top of Rafael's work on systemwide runtime power
> management. It supports suspending and resuming PCI devices at
> runtime, enabling platform wakeup events that allow the devices to
> automatically resume when appropriate. It currently requires platform
> support, but PCIe setups could be supported natively once native PCIe
> PME code has been added to the kernel.

PCI bits look pretty good to me, though Rafael should take a look too.
Card readers and firewire could benefit from similar treatment, maybe
that would get us to .5W territory on some machines.

-- 
Jesse Barnes, Intel Open Source Technology Center