Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1552245 > unrolled thread

Re: [PATCH] usb: dwc3: gadget: Avoid race between dwc3 interrupt handler and irq thread handler

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [PATCH] usb: dwc3: gadget: Avoid race between dwc3 interrupt  handler and irq thread handler John Youn <John.Youn@synopsys.com> - 2017-01-05 20:10 +0100
    Re: [PATCH] usb: dwc3: gadget: Avoid race between dwc3 interrupt  handler and irq thread handler Baolin Wang <baolin.wang@linaro.org> - 2017-01-06 03:50 +0100

#1552245 — Re: [PATCH] usb: dwc3: gadget: Avoid race between dwc3 interrupt handler and irq thread handler

On 12/28/2016 5:29 PM, John Youn wrote:
> 
> 
>> Janusz Dziedzic <janusz.dziedzic@gmail.com> writes:
>>>>>> On some platfroms(like x86 platform), when one core is running the
>> USB gadget
>>>>>> irq thread handler by dwc3_thread_interrupt(), meanwhile another
>> core also can
>>>>>> respond other interrupts from dwc3 controller and modify the event
>> buffer by
>>>>>> dwc3_interrupt() function, that will cause getting the wrong event
>> count in
>>>>>> irq thread handler to make the USB function abnormal.
>>>>>>
>>>>>> We should add spin_lock/unlock() in dwc3_check_event_buf() to avoid
>> this race.
>>>>>
>>>>> Why not spin_lock_irq ones? This lock seems to be used in both
>>>>> normal and interrupt threads. Or, I missed anything?
>>>>
>>>> this is top half handler. Interrupts are already disabled.
>>>>
>>> BTW,
>>> We don't use spin_lock in top half handler.
>>> Maybe we should/can switch all spin_lock_irqsave() to simple
>>> spin_lock() in the thread/callbacks?
>>
>> in theory, yes we've masked all interrupts from this controller for the
>> duration of the thread handler. However this breaks networking
>> gadgets. I can only guess network stack has a hard requirement to run
>> with IRQs disabled.
>>
> 
> Hi,
> 
> Is this version 3.00a of the core?
> 
> That version has a STAR where the interrupts cannot be masked. That
> results in similar symptoms to what you're seeing here.
> Regards,
> John
> 

Didn't see any response to this. Just want to make sure this
possibility is addressed as there is a workaround for it on mainline.

See the following commits for details:

d9fa4c63f766 ("usb: dwc3: core: add a event buffer cache")
ebbb2d59398f ("usb: dwc3: gadget: use evt->cache for processing events")
65aca3205046 ("usb: dwc3: gadget: clear events in top-half handler")
cf40b86b6ef6 ("usb: dwc3: Implement interrupt moderation")
28632b44d129 ("usb: dwc3: Workaround for irq mask issue")

Regards,
John

[toc] | [next] | [standalone]

#1552485

Hi John,

On 6 January 2017 at 03:08, John Youn <John.Youn@synopsys.com> wrote:
> On 12/28/2016 5:29 PM, John Youn wrote:
>>
>>
>>> Janusz Dziedzic <janusz.dziedzic@gmail.com> writes:
>>>>>>> On some platfroms(like x86 platform), when one core is running the
>>> USB gadget
>>>>>>> irq thread handler by dwc3_thread_interrupt(), meanwhile another
>>> core also can
>>>>>>> respond other interrupts from dwc3 controller and modify the event
>>> buffer by
>>>>>>> dwc3_interrupt() function, that will cause getting the wrong event
>>> count in
>>>>>>> irq thread handler to make the USB function abnormal.
>>>>>>>
>>>>>>> We should add spin_lock/unlock() in dwc3_check_event_buf() to avoid
>>> this race.
>>>>>>
>>>>>> Why not spin_lock_irq ones? This lock seems to be used in both
>>>>>> normal and interrupt threads. Or, I missed anything?
>>>>>
>>>>> this is top half handler. Interrupts are already disabled.
>>>>>
>>>> BTW,
>>>> We don't use spin_lock in top half handler.
>>>> Maybe we should/can switch all spin_lock_irqsave() to simple
>>>> spin_lock() in the thread/callbacks?
>>>
>>> in theory, yes we've masked all interrupts from this controller for the
>>> duration of the thread handler. However this breaks networking
>>> gadgets. I can only guess network stack has a hard requirement to run
>>> with IRQs disabled.
>>>
>>
>> Hi,
>>
>> Is this version 3.00a of the core?
>>
>> That version has a STAR where the interrupts cannot be masked. That
>> results in similar symptoms to what you're seeing here.

Sorry for late reply. The version is 2.80a.

>> Regards,
>> John
>>
>
> Didn't see any response to this. Just want to make sure this
> possibility is addressed as there is a workaround for it on mainline.

Thanks for reminding.

>
> See the following commits for details:
>
> d9fa4c63f766 ("usb: dwc3: core: add a event buffer cache")
> ebbb2d59398f ("usb: dwc3: gadget: use evt->cache for processing events")
> 65aca3205046 ("usb: dwc3: gadget: clear events in top-half handler")
> cf40b86b6ef6 ("usb: dwc3: Implement interrupt moderation")
> 28632b44d129 ("usb: dwc3: Workaround for irq mask issue")
>

-- 
Baolin.wang
Best Regards

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web