On Thu, Jun 24, 2010 at 5:43 PM, Kevin Hilman
[off-list ref] wrote:

Implement the new runtime PM framework as a thin layer on top of the
omap_device API. ?Since we don't have an OMAP-specific bus, override
the runtime PM hooks for the platform_bus for the OMAP specific
implementation.

[...]

+int platform_pm_runtime_suspend(struct device *dev)
+{
+ ? ? ? struct platform_device *pdev = to_platform_device(dev);
+ ? ? ? struct omap_device *odev = to_omap_device(pdev);
+ ? ? ? int r, ret = 0;
+
+ ? ? ? dev_dbg(dev, "%s\n", __func__);
+
+ ? ? ? if (dev->driver->pm && dev->driver->pm->runtime_suspend)
+ ? ? ? ? ? ? ? ret = dev->driver->pm->runtime_suspend(dev);
+ ? ? ? if (!ret && omap_device_is_valid(odev)) {
+ ? ? ? ? ? ? ? r = omap_device_idle(pdev);
+ ? ? ? ? ? ? ? WARN_ON(r);

For the record, I should note here that this is *really* dangerous.
When handed a random platform_device pointer, it is not safe to use
to_omap_device() and dereference it with omap_device_is_valid().
There are no guarantees that the dereference is actually valid,
particularly so when the platform_device has been dynamically
allocated.

It is only okay to use to_omap_device() when the code is already
absolutely sure that the platform_device is in fact contained by a
struct omap_device.  There needs to be a different method to test if
it is contained by an omap_device before doing the dereference.