On Mon, Jan 28, 2013 at 5:42 PM, Felipe Balbi <balbi@xxxxxx> wrote:
> Hi,
>
> On Mon, Jan 28, 2013 at 05:28:30PM +0530, Vivek Gautam wrote:
>> >> +static int dwc3_exynos_runtime_resume(struct device *dev)
>> >> +{
>> >> + struct dwc3_exynos *exynos = dev_get_drvdata(dev);
>> >> + struct platform_device *pdev_dwc = exynos->dwc3;
>> >> + struct dwc3 *dwc = NULL;
>> >> +
>> >> + dwc = platform_get_drvdata(pdev_dwc);
>> >> +
>> >> + clk_enable(exynos->clk);
>> >> +
>> >> + if (!dwc)
>> >> + return 0;
>> >> +
>> >> + pm_runtime_get_sync(dwc->usb3_phy->dev);
>> >
>> > dude, this is wrong :-)
>> >
>> > look at this:
>> >
>> > pm_runtime_get() -> dwc3_exynos_runtime_resume() ->
>> > pm_runtime_get_sync() -> dwc3_exynos_runtime_resume() -> ...
>> >
>> > only your clock enalbe should do
>> >
>>
>> We want to wake up "dwc->usb3_phy" so tried to call pm_runtime_get_sync()
>> with "dwc->usb3_phy->dev".
>> Missing something ? :-(
>
> oh, my bad. That's the PHY... But we can't really do that for samsung
> only. It needs to be done generically for the entire dwc3 core driver,
> and for that we need to introduce usb_phy_autopm_get(),
> usb_phy_autopm_get_sync() and friends.
>
aah!! Ok. I definitely missed that part. :-(
> Then, from dwc_probe() we call:
>
> phy = usb_get_phy();
> usb_phy_autopm_enable(phy);
> usb_phy_autopm_get_sync(phy);
>
> or something similar ;-) Bottom line, you shouldn't fiddle with phy->dev
> directly.
>
Ok, the core should actually be handling the 'phy' not the glue layers.
Right ?
Will try putting these helper functions in place and come up with a
solution. :-)
--
Thanks & Regards
Vivek
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
