dwc3: Power management for cable disconnect

dwc3: Power management for cable disconnect

[Wesley Cheng]

Hi linux-usb members!

I was looking at seeing how the DWC3 drivers are handling the USB cable
disconnect routine (VBUS low situation) when there is no UDC binded.  Is
the below behavior expected?

1.  Enumerate device w/ host.
2.  Gadget is unbinded
	- echo "" > /sys/kernel/config/usb_gadget/g1/UDC
3.  Disconnect the USB cable (VBUS low)
4.  No dwc3_gadget_disconnect_interrupt() seen (since controller is
   halted from step#1)
5.  Runtime PM autosuspend fails due to "dwc->connected" being true
(cleared in dwc3_gadget_disconnect_interrupt())
6.  Gadget binded
	- echo udc_name > /sys/kernel/config/usb_gadget/g1/UDC
7.  No runtime suspend until cable is plugged in and out

This would be problematic for devices which have to reliably enter
suspend.  I just wanted to see if this is a scenario which has been
considered?  Seems odd that we would still keep the controller out of
low power if there was no cable connected.

Thanks

Regards,
Wesley Cheng

-- 
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project

Re: dwc3: Power management for cable disconnect

[Thinh Nguyen]

Hi,

Wesley Cheng wrote:
> Hi linux-usb members!
>
> I was looking at seeing how the DWC3 drivers are handling the USB cable
> disconnect routine (VBUS low situation) when there is no UDC binded.  Is
> the below behavior expected?
>
> 1.  Enumerate device w/ host.
> 2.  Gadget is unbinded
> 	- echo "" > /sys/kernel/config/usb_gadget/g1/UDC
> 3.  Disconnect the USB cable (VBUS low)
> 4.  No dwc3_gadget_disconnect_interrupt() seen (since controller is
>    halted from step#1)

I think you mean after step#2?

> 5.  Runtime PM autosuspend fails due to "dwc->connected" being true
> (cleared in dwc3_gadget_disconnect_interrupt())

This is probably a mishap in the dwc3 driver. When the function driver
is unregistered, the pullup() callback will halt the controller. This is
a soft-disconnect, and dwc->connected should be false.

> 6.  Gadget binded
> 	- echo udc_name > /sys/kernel/config/usb_gadget/g1/UDC

After this step, the device will still remain disconnected.

> 7.  No runtime suspend until cable is plugged in and out

... until the reset and connection done interrupt telling the driver
that the device is now connected.

>
> This would be problematic for devices which have to reliably enter
> suspend.  I just wanted to see if this is a scenario which has been
> considered?  Seems odd that we would still keep the controller out of
> low power if there was no cable connected.
>

This should be a simple fix. Would you create a patch for this?

Thanks,
Thinh

Re: dwc3: Power management for cable disconnect

[Wesley Cheng]

On 10/16/2020 3:42 PM, Thinh Nguyen wrote:
> Hi,
> 
> Wesley Cheng wrote:
>> Hi linux-usb members!
>>
>> I was looking at seeing how the DWC3 drivers are handling the USB cable
>> disconnect routine (VBUS low situation) when there is no UDC binded.  Is
>> the below behavior expected?
>>
>> 1.  Enumerate device w/ host.
>> 2.  Gadget is unbinded
>> 	- echo "" > /sys/kernel/config/usb_gadget/g1/UDC
>> 3.  Disconnect the USB cable (VBUS low)
>> 4.  No dwc3_gadget_disconnect_interrupt() seen (since controller is
>>    halted from step#1)
> 
> I think you mean after step#2?
> 
>> 5.  Runtime PM autosuspend fails due to "dwc->connected" being true
>> (cleared in dwc3_gadget_disconnect_interrupt())
> 
> This is probably a mishap in the dwc3 driver. When the function driver
> is unregistered, the pullup() callback will halt the controller. This is
> a soft-disconnect, and dwc->connected should be false.
> 
>> 6.  Gadget binded
>> 	- echo udc_name > /sys/kernel/config/usb_gadget/g1/UDC
> 
> After this step, the device will still remain disconnected.
> 
>> 7.  No runtime suspend until cable is plugged in and out
> 
> ... until the reset and connection done interrupt telling the driver
> that the device is now connected.
> 
>>
>> This would be problematic for devices which have to reliably enter
>> suspend.  I just wanted to see if this is a scenario which has been
>> considered?  Seems odd that we would still keep the controller out of
>> low power if there was no cable connected.
>>
> 
> This should be a simple fix. Would you create a patch for this?
> 

Hi Thinh,

Sure, will submit a patch for this.  Will upload it soon!

Thanks

Regards,
Wesley Cheng

-- 
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project