Re: [PATCH v2] usb: dwc3: gadget: Wait for end transfer complete before free irq

[Baolin Wang]

Hi,

On 13 October 2016 at 15:51, Felipe Balbi <balbi@xxxxxxxxxx> wrote:
>
> Hi,
>
> Baolin Wang <baolin.wang@xxxxxxxxxx> writes:
>> Hi,
>>
>> On 13 October 2016 at 15:02, Felipe Balbi <balbi@xxxxxxxxxx> wrote:
>>>
>>> Hi,
>>>
>>> Baolin Wang <baolin.wang@xxxxxxxxxx> writes:
>>>> @@ -1742,6 +1791,17 @@ static int dwc3_gadget_stop(struct usb_gadget *g)
>>>>       dwc->gadget_driver      = NULL;
>>>>       spin_unlock_irqrestore(&dwc->lock, flags);
>>>>
>>>> +     /*
>>>> +      * Since the xHCI will share the same interrupt with gadget, thus when
>>>> +      * free the gadget irq, it will not shutdown this gadget irq line. Then
>>>> +      * the gadget driver can not handle the end transfer command complete
>>>> +      * event after free the gadget irq, which will hang the system to crash.
>>>> +      *
>>>> +      * So we should wait for the end transfer command complete event before
>>>> +      * free it to avoid this situation.
>>>> +      */
>>>> +     dwc3_wait_command_complete(dwc);
>>>
>>> this doesn't make sense. We have already masked all interrupts before
>>> getting here. We have also, already, disabled all endpoints.
>>
>> No, you just mask the interrupts described in DEVTEN register, and you
>> did not disable the endpoint command complete event.
>
> True that
>
>>> I'm thinking this is a bug in configfs interface of Gadget API, not
>>> dwc3. The only reason for this to happen would be if we get to
>>> ->udc_stop() with endpoints still enabled.
>>>
>>> Can you check if that's the case? i.e. can you check if any endpoints
>>> are still enabled when we get here?
>>
>> No, it is not any endpoints are still enabled. Like I said in commit
>> message, we will start end transfer command when disable the endpoint,
>> if the end transfer command complete event comes after we have freed
>> the gadget irq, it will trigger the interrupt line all the time to
>> crash the system.
>
> I see what the problem is. Databook tells us we *must* set CMDIOC when
> issuing EndTransfer command and we should always wait for Command
> Complete IRQ. However, we took a shortcut and just delayed for 100us
> after issuing End Transfer.

Yes, but 100us delay is not enough in some scenarios, like changing
function with configfs frequently, we will met problems.

>
> It seems as if *that* should be fixed. We should start actually waiting
> for Command Complete IRQ.
>
> --
> balbi



-- 
Baolin.wang
Best Regards