Re: [PATCH] usb: dwc3: Support the dwc3 host suspend/resume

From: Manu Gautam
Date: Fri Jul 21 2017 - 05:13:51 EST


Hi,

On 7/21/2017 2:31 PM, Baolin Wang wrote:
> On 21 July 2017 at 16:45, Manu Gautam <mgautam@xxxxxxxxxxxxxx> wrote:
>> Hi,
>>
>>
>> On 7/21/2017 12:28 PM, Baolin Wang wrote:
>>> For some mobile devices with strict power management, we also want to
>>> suspend the host when the slave was detached for power saving. Thus
>>> adding the host suspend/resume functions to support this requirement.
>>
>> USB/PM core already takes care of suspending root-HUB/XHCI when
>> no device connected.
> Correct, but what I mean is we can power off the dwc3 controller when
> there are no device connected.
>
>>> We will issue the pm_suspend_ignore_children() for the dwc3 device,
>>> since we will resume the child device (xHCI device) in runtime resume
>>> callback (dwc3_host_resume()) of dwc3 device, now the dwc3 device's
>>> runtime state is not RPM_ACTIVE, which will block to resume its
>>> child device (xHCI device). Add ignore children flag can avoid this
>>> situation.
>> This defeats the basic purpose of runtime PM. Without ignore_children
>> once USB bus is suspended, dwc3 gets suspended and then dwc3 glue device.
>> Only requirement I see from the patch is to resume XHCI/root-hub on
>> dwc3 resume. I am sure there must be other way to deal with that e.g.
>> doing same from glue driver by using a wq or use pm_request_resume()
> Ah, I will try pm_request_resume().
>
>>> Signed-off-by: Baolin Wang <baolin.wang@xxxxxxxxxx>
>>> ---
>>> drivers/usb/dwc3/core.c | 26 +++++++++++++++++++++++++-
>>> drivers/usb/dwc3/core.h | 7 +++++++
>>> drivers/usb/dwc3/host.c | 15 +++++++++++++++
>>> 3 files changed, 47 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c
>>> index 326b302..2be7ddc 100644
>>> --- a/drivers/usb/dwc3/core.c
>>> +++ b/drivers/usb/dwc3/core.c
>>> @@ -1193,6 +1193,7 @@ static int dwc3_probe(struct platform_device *pdev)
>>> pm_runtime_use_autosuspend(dev);
>>> pm_runtime_set_autosuspend_delay(dev, DWC3_DEFAULT_AUTOSUSPEND_DELAY);
>>> pm_runtime_enable(dev);
>>> + pm_suspend_ignore_children(dev, true);
>>> ret = pm_runtime_get_sync(dev);
>>> if (ret < 0)
>>> goto err1;
>>> @@ -1292,15 +1293,27 @@ static int dwc3_remove(struct platform_device *pdev)
>>> static int dwc3_suspend_common(struct dwc3 *dwc)
>> What is the trigger for runtime suspend now that you have ignore_children set.
>>> {
>>> unsigned long flags;
>>> + int ret;
>>>
>>> switch (dwc->dr_mode) {
>>> case USB_DR_MODE_PERIPHERAL:
>>> + spin_lock_irqsave(&dwc->lock, flags);
>>> + dwc3_gadget_suspend(dwc);
>>> + spin_unlock_irqrestore(&dwc->lock, flags);
>>> + break;
>>> case USB_DR_MODE_OTG:
>>> + ret = dwc3_host_suspend(dwc);
>> With DRD/OTG, if current mode is device and dwc3->xhci won't be valid.
> If current mode is device, then xHCI device is always in suspend state.
dwc->xhci is allocated in dwc3_host_init. And dwc->xhci device is unregistred
from dwc3_host_exit that is called from __dwc3_set_mode.
>
>> You can refer to the patch that I pushed to address this.
>>
>>> + if (ret)
>>> + return ret;
>>> +
>>> spin_lock_irqsave(&dwc->lock, flags);
>>> dwc3_gadget_suspend(dwc);
>>> spin_unlock_irqrestore(&dwc->lock, flags);
>>> break;
>>> case USB_DR_MODE_HOST:
>>> + ret = dwc3_host_suspend(dwc);
>>> + if (ret)
>>> + return ret;
>>> default:
>>> /* do nothing */
>>> break;
>>> @@ -1322,12 +1335,23 @@ static int dwc3_resume_common(struct dwc3 *dwc)
>>>
>>> switch (dwc->dr_mode) {
>>> case USB_DR_MODE_PERIPHERAL:
>>> + spin_lock_irqsave(&dwc->lock, flags);
>>> + dwc3_gadget_resume(dwc);
>>> + spin_unlock_irqrestore(&dwc->lock, flags);
>>> + break;
>>> case USB_DR_MODE_OTG:
>>> + ret = dwc3_host_resume(dwc);
>>> + if (ret)
>>> + return ret;
>>> +
>>> spin_lock_irqsave(&dwc->lock, flags);
>>> dwc3_gadget_resume(dwc);
>>> spin_unlock_irqrestore(&dwc->lock, flags);
>>> - /* FALLTHROUGH */
>>> + break;
>>> case USB_DR_MODE_HOST:
>>> + ret = dwc3_host_resume(dwc);
>>> + if (ret)
>>> + return ret;
>>> default:
>>> /* do nothing */
>>> break;
>>> diff --git a/drivers/usb/dwc3/core.h b/drivers/usb/dwc3/core.h
>>> index ea910ac..9b5a4eb 100644
>>> --- a/drivers/usb/dwc3/core.h
>>> +++ b/drivers/usb/dwc3/core.h
>>> @@ -1193,11 +1193,18 @@ static inline bool dwc3_is_usb31(struct dwc3 *dwc)
>>> #if IS_ENABLED(CONFIG_USB_DWC3_HOST) || IS_ENABLED(CONFIG_USB_DWC3_DUAL_ROLE)
>>> int dwc3_host_init(struct dwc3 *dwc);
>>> void dwc3_host_exit(struct dwc3 *dwc);
>>> +int dwc3_host_suspend(struct dwc3 *dwc);
>>> +int dwc3_host_resume(struct dwc3 *dwc);
>>> #else
>>> static inline int dwc3_host_init(struct dwc3 *dwc)
>>> { return 0; }
>>> static inline void dwc3_host_exit(struct dwc3 *dwc)
>>> { }
>>> +
>>> +static inline int dwc3_host_suspend(struct dwc3 *dwc)
>>> +{ return 0; }
>>> +static inline int dwc3_host_resume(struct dwc3 *dwc)
>>> +{ return 0; }
>>> #endif
>>>
>>> #if IS_ENABLED(CONFIG_USB_DWC3_GADGET) || IS_ENABLED(CONFIG_USB_DWC3_DUAL_ROLE)
>>> diff --git a/drivers/usb/dwc3/host.c b/drivers/usb/dwc3/host.c
>>> index 76f0b0d..3fa4414 100644
>>> --- a/drivers/usb/dwc3/host.c
>>> +++ b/drivers/usb/dwc3/host.c
>>> @@ -16,6 +16,7 @@
>>> */
>>>
>>> #include <linux/platform_device.h>
>>> +#include <linux/pm_runtime.h>
>>>
>>> #include "core.h"
>>>
>>> @@ -151,3 +152,17 @@ void dwc3_host_exit(struct dwc3 *dwc)
>>> dev_name(dwc->dev));
>>> platform_device_unregister(dwc->xhci);
>>> }
>>> +
>>> +int dwc3_host_suspend(struct dwc3 *dwc)
>>> +{
>>> + struct device *xhci = &dwc->xhci->dev;
>>> +
>>> + return pm_runtime_suspend(xhci);
>> If ignore_children is not marked then xhci would indeed be suspended.
>>
>>> +}
>>> +
>>> +int dwc3_host_resume(struct dwc3 *dwc)
>>> +{
>>> + struct device *xhci = &dwc->xhci->dev;
>>> +
>>> + return pm_runtime_resume(xhci);
>> Other than what I pushed in my patch -
>> ("usb: dwc3: Don't reinitialize core during host bus-suspend/resume")
>> Just performing pm_request_resume or handling same in dwc3 glue
>> driver should be sufficient.
> Yes.
>
>> Also, what is trigger for runtime_resume for your platform?
> In our platform glue layer, we have extcon events to notify there are
> devices connected, then gule layer will resume dwc3 device.
>

IMO you can just perform resume of &dwc->xhci->dev instead of resuming dwc3.
Since, dwc3 is parent of xhci that will trigger resume of dwc3 as well.