On Wednesday, November 28, 2012 03:25:59 PM Bjorn Helgaas wrote:

On Tue, Nov 20, 2012 at 1:08 AM, Huang Ying [off-list ref] wrote:

quoted

For unbound PCI devices, what we need is:

 - Always in D0 state, because some devices does not work again after
   being put into D3 by the PCI bus.

 - In SUSPENDED state if allowed, so that the parent devices can still
   be put into low power state.

To satisfy these requirement, the runtime PM for the unbound PCI
devices are disabled and set to SUSPENDED state.  One issue of this
solution is that the PCI devices will be put into SUSPENDED state even
if the SUSPENDED state is forbidden via the sysfs interface
(.../power/control) of the device.  This is not an issue for most
devices, because most PCI devices are not used at all if unbounded.
But there are exceptions.  For example, unbound VGA card can be used
for display, but suspend its parents make it stop working.

To fix the issue, we keep the runtime PM enabled when the PCI devices
are unbound.  But the runtime PM callbacks will do nothing if the PCI
devices are unbound.  This way, we can put the PCI devices into
SUSPENDED state without put the PCI devices into D3 state.

Does this fix a reported problem?  Is there a bug report URL?  What
problem would a user notice?

There is a BZ: https://bugzilla.kernel.org/show_bug.cgi?id=48201

Unfortunately, the reporter hasn't confirmed that the bug is fixed,
although we're quite confident that it will be.

This doesn't look critical enough to try to put in v3.7 (correct me if
I'm wrong).  It's getting pretty late for the v3.8 merge window, but
it looks like it qualifies as a bug fix that we would merge even after
the merge window, so I'll put it in my -next branch headed for v3.8.
Then it can be backported to v3.6 and v3.7.

I think that's OK.

Thanks,
Rafael


-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.