lkml.org 
[lkml]   [2022]   [Jul]   [20]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: [PATCH v2 3/5] usb: dwc3: gadget: Adjust IRQ management during soft disconnect/connect
Date
Hi Wesley,

On 7/20/2022, Wesley Cheng wrote:
> Hi Thinh,
>
> On 7/14/2022 10:38 AM, Thinh Nguyen wrote:
>> On 7/12/2022, Wesley Cheng wrote:
>>> Local interrupts are currently being disabled as part of aquiring the
>>> spin lock before issuing the endxfer command.  Leave interrupts
>>> enabled, so
>>> that EP0 events can continue to be processed.  Also, ensure that
>>> there are
>>> no pending interrupts before attempting to handle any soft
>>> connect/disconnect.
>>>
>>> Fixes: 861c010a2ee1 ("usb: dwc3: gadget: Refactor pullup()")
>>> Signed-off-by: Wesley Cheng <quic_wcheng@quicinc.com>
>>> ---
>>>    drivers/usb/dwc3/gadget.c | 21 ++++++++++++---------
>>>    1 file changed, 12 insertions(+), 9 deletions(-)
>>>
>>> diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
>>> index a455f8d4631d..ee85b773e3fe 100644
>>> --- a/drivers/usb/dwc3/gadget.c
>>> +++ b/drivers/usb/dwc3/gadget.c
>>> @@ -1674,6 +1674,7 @@ static int __dwc3_gadget_get_frame(struct dwc3
>>> *dwc)
>>>    static int __dwc3_stop_active_transfer(struct dwc3_ep *dep, bool
>>> force, bool interrupt)
>>>    {
>>>        struct dwc3_gadget_ep_cmd_params params;
>>> +    struct dwc3 *dwc = dep->dwc;
>>>        u32 cmd;
>>>        int ret;
>>>    @@ -1682,7 +1683,9 @@ static int
>>> __dwc3_stop_active_transfer(struct dwc3_ep *dep, bool force, bool int
>>>        cmd |= interrupt ? DWC3_DEPCMD_CMDIOC : 0;
>>>        cmd |= DWC3_DEPCMD_PARAM(dep->resource_index);
>>>        memset(&params, 0, sizeof(params));
>>> +    spin_unlock(&dwc->lock);
>>>        ret = dwc3_send_gadget_ep_cmd(dep, cmd, &params);
>>> +    spin_lock(&dwc->lock);
>>>        WARN_ON_ONCE(ret);
>>>        dep->resource_index = 0;
>>>    @@ -2029,12 +2032,11 @@ static int dwc3_gadget_ep_dequeue(struct
>>> usb_ep *ep,
>>>        struct dwc3_ep            *dep = to_dwc3_ep(ep);
>>>        struct dwc3            *dwc = dep->dwc;
>>>    -    unsigned long            flags;
>>>        int                ret = 0;
>>>           trace_dwc3_ep_dequeue(req);
>>>    -    spin_lock_irqsave(&dwc->lock, flags);
>>> +    spin_lock(&dwc->lock);
>>>           list_for_each_entry(r, &dep->cancelled_list, list) {
>>>            if (r == req)
>>> @@ -2073,7 +2075,7 @@ static int dwc3_gadget_ep_dequeue(struct
>>> usb_ep *ep,
>>>            request, ep->name);
>>>        ret = -EINVAL;
>>>    out:
>>> -    spin_unlock_irqrestore(&dwc->lock, flags);
>>> +    spin_unlock(&dwc->lock);
>>>           return ret;
>>>    }
>>> @@ -2489,9 +2491,7 @@ static int __dwc3_gadget_start(struct dwc3 *dwc);
>>>       static int dwc3_gadget_soft_disconnect(struct dwc3 *dwc)
>>>    {
>>> -    unsigned long flags;
>>> -
>>> -    spin_lock_irqsave(&dwc->lock, flags);
>>> +    spin_lock(&dwc->lock);
>>>        dwc->connected = false;
>>>           /*
>>> @@ -2506,10 +2506,10 @@ static int
>>> dwc3_gadget_soft_disconnect(struct dwc3 *dwc)
>>>               reinit_completion(&dwc->ep0_in_setup);
>>>    -        spin_unlock_irqrestore(&dwc->lock, flags);
>>> +        spin_unlock(&dwc->lock);
>>>            ret = wait_for_completion_timeout(&dwc->ep0_in_setup,
>>>                    msecs_to_jiffies(DWC3_PULL_UP_TIMEOUT));
>>> -        spin_lock_irqsave(&dwc->lock, flags);
>>> +        spin_lock(&dwc->lock);
>>>            if (ret == 0)
>>>                dev_warn(dwc->dev, "timed out waiting for SETUP
>>> phase\n");
>>>        }
>>> @@ -2523,7 +2523,7 @@ static int dwc3_gadget_soft_disconnect(struct
>>> dwc3 *dwc)
>>>         */
>>>        dwc3_stop_active_transfers(dwc);
>>>        __dwc3_gadget_stop(dwc);
>>> -    spin_unlock_irqrestore(&dwc->lock, flags);
>>> +    spin_unlock(&dwc->lock);
>>>           /*
>>>         * Note: if the GEVNTCOUNT indicates events in the event
>>> buffer, the
>>> @@ -2569,6 +2569,8 @@ static int dwc3_gadget_pullup(struct
>>> usb_gadget *g, int is_on)
>>>            return 0;
>>>        }
>>>    +    synchronize_irq(dwc->irq_gadget);
>>> +
>>>        if (!is_on) {
>>>            ret = dwc3_gadget_soft_disconnect(dwc);
>>>        } else {
>>> @@ -3729,6 +3731,7 @@ void dwc3_stop_active_transfer(struct dwc3_ep
>>> *dep, bool force,
>>>         */
>>>           __dwc3_stop_active_transfer(dep, force, interrupt);
>>> +
>>>    }
>>>       static void dwc3_clear_stall_all_ep(struct dwc3 *dwc)
>>
>> Hi Greg,
>>
>> Please don't pick up this patch yet. We're still in discussion with
>> this. I have some concern with unlocking/locking when sending End
>> Transfer command. For example, this patch may cause issues with
>> DWC3_EP_END_TRANSFER_PENDING checks.
>>
>> Hi Wesley,
>>
>> Did you try out my suggestion yet?
>>
>
> Just providing a quick update.
>
> So with your suggestion, I was able to consistently reproduce the
> controller halt issue after a day or so of testing.  However, when I
> took a further look, I believe the problem is due to the DWC3 event
> handler:
>
> static void dwc3_endpoint_interrupt(struct dwc3 *dwc,
>         const struct dwc3_event_depevt *event)
> {
> ...
>     if (!(dep->flags & DWC3_EP_ENABLED)) {
>         if (!(dep->flags & DWC3_EP_TRANSFER_STARTED))
>             return;
>
>         /* Handle only EPCMDCMPLT when EP disabled */
>         if (event->endpoint_event != DWC3_DEPEVT_EPCMDCMPLT)
>             return;
>     }
>
> The soft disconnect routine reached to the run/stop polling point, and
> I could see that DWC3_EP_DELAYED_STOP was set, and we got a
> xfercomplete event for the STATUS phase.  However, since we exit early
> in the event handler (due to __dwc3_gadget_stop() being called and
> disabling EP0), the STATUS complete is never handled, and we do not
> issue the endxfer command.
>
> I don't think I saw this issue with my change, as we allowed the
> STATUS phase handling to happen BEFORE gadget stop was called (since I
> released the lock in the stop active transfers API).
>
> However, I think even with my approach, we'd eventually run into a
> possibility of this issue, as we aren't truly handling EP0 events
> while polling for the halted status due to the above.  It was just
> reducing the chances.  The scenario of this issue is coming because
> the host took a long time to complete the STATUS phase, so we ran into
> a "timed out waiting for SETUP phase," which allowed us to call the
> run/stop routine while we were not yet in the SETUP phase.
>
> I'm currently running a change to add a EP num check to this IF
> condition:
>
>     if ((epnum > 1) && !(dep->flags & DWC3_EP_ENABLED)) {
>         if (!(dep->flags & DWC3_EP_TRANSFER_STARTED))
>             return;
>
>         /* Handle only EPCMDCMPLT when EP disabled */
>         if (event->endpoint_event != DWC3_DEPEVT_EPCMDCMPLT)
>             return;
>     }
>

Ah... good catch! Thanks for all the testings.

BR,
Thinh
\
 
 \ /
  Last update: 2022-07-21 02:27    [W:1.847 / U:0.412 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site