Re: [RFC][PATCH] HACK: usb: dwc2: Workaround case where GOTGCTL state is wrong

From: John Stultz
Date: Thu Dec 08 2016 - 17:43:04 EST


On Tue, Dec 6, 2016 at 7:52 PM, John Youn <John.Youn@xxxxxxxxxxxx> wrote:
> On 12/6/2016 5:48 PM, John Stultz wrote:
>> Hey John,
>> Just wanted to send this by you, as it seems something is
>> slightly off with the GOTGCTL state when removing a otg adapter
>> cable. The following seems to work around the issue I'm seeing.
>>
>>
>> When removing a USB-A to USB-otg adapter cable, we get a change
>> status irq, and then in dwc2_conn_id_status_change, we
>> erroniously see the GOTGCTL_CONID_B flag set. This causes us to
>
> This is the correct behavior for an OTG controller. When you unplug a
> cable or plug in the B end of a cable, the ID pin floats, indicating
> it is a B-Device.
>
> When you plug in an A-cable, which is what your adapter is, it will
> ground the pin, meaning A-device.

Hrm... So normally, when I plug in the gadget cable into the OTG port,
I see the change_status irq comes in and the function sees:

dwc2 f72c0000.usb: gotgctl=4010000
dwc2 f72c0000.usb: gotgctl.b.conidsts=1
dwc2 f72c0000.usb: Do port resume before switching to device mode
dwc2 f72c0000.usb: dwc2_hsotg_enqueue_setup: failed queue (-11)
dwc2 f72c0000.usb: new device is high-speed
dwc2 f72c0000.usb: new device is high-speed
dwc2 f72c0000.usb: new device is high-speed
dwc2 f72c0000.usb: new address 37
configfs-gadget gadget: high-speed config #1: b

Then when I unplug the cable:

dwc2 f72c0000.usb: gotgctl=2200000
dwc2 f72c0000.usb: gotgctl.b.conidsts=0
usb 1-1: reset high-speed USB device number 13 using dwc2



When I plug in the OTG to USB-A adapter cable w/ a mouse plugged in
(note I see no change interrupt):

usb 1-1: USB disconnect, device number 13
usb 1-1: new low-speed USB device number 14 using dwc2
input: Logitech USB Optical Mouse as
/devices/platform/soc/f72c0000.usb/usb1/1-1/1-1:1.0/0003:046D:C058.0003/input/input3
hid-generic 0003:046D:C058.0003: input,hidraw0: USB HID v1.11 Mouse
[Logitech USB Optical Mouse] on usb-f72c0000.usb-1/input0


Then unplugging the OTG to USB-A adapter cable w/ mouse:

dwc2 f72c0000.usb: gotgctl=4010000
dwc2 f72c0000.usb: gotgctl.b.conidsts=1
dwc2 f72c0000.usb: Do port resume before switching to device mode
dwc2 f72c0000.usb: Waiting for Peripheral Mode, Mode=Host

<here we get stuck in the loop until it finishes, unless using the
patch from this thread>

usb 1-1: USB disconnect, device number 14
dwc2 f72c0000.usb: gotgctl=2200000
dwc2 f72c0000.usb: gotgctl.b.conidsts=0
usb 1-1: new high-speed USB device number 15 using dwc2
hub 1-1:1.0: USB hub found
hub 1-1:1.0: 3 ports detected


So I only get the change irq when:
* I plug in a micro-usb-B cable for gadget mode
* I remove the micro-usb-B cable being used for gadget mode
* I remove a OTG to USB-A adapter

One slight quirk, is that I don't always see the change irq when
removing the OTG to USB, as if I plug in a highspeed mass-storage
device, instead of the low-speed mouse, I don't see the change
interrupt and the device shows up and disappears the same as when I
plug into the normal USB-A host ports on the board.


>> get stuck in the "while (!dwc2_is_device_mode(hsotg))" loop,
>> spitting out "Waiting for Peripheral Mode, Mode=Host" warnings
>> until it fails out many seconds later.
>
> This is weird. Once the ID pin goes to B, the core should become a
> peripheral and this should be reflected in the status registers.
>
>>
>> This patch works around the issue by re-reading the GOTGCTL
>> state to check if the GOTGCTL_CONID_B is still set and if not
>> restarting the change status logic.
>
> This also seems weird. The connector id status shouldn't go back to A,
> assuming you've left the cable unplugged.

So I suspect this has something to do with the way the USB-A host
ports on the board are wired up. As removing the usb-b plug seems to
switch the device back into A mode.

One quirk with this board is that the USB-A ports on the board do not
function if anything is in the OTG/B plug (which is frustrating to use
at times).

Guodong or Chen Yu understand the hardware details a bit better, and
might be able to explain more if you need more information.

> Is the controller supposed to work in both peripheral and host modes?

I certainly hope so!

>> I suspect this isn't the best solution, but it seems to work
>> well for me.
>>
>
> The workaround seems fine, but still, this indicates that something
> wrong is going on somwhere.

Yea. While I'm ok with some of these workarounds, I'd be happier to
figure out what the proper thing should be.


> You can add my ack:
>
> Acked-by: John Youn <johnyoun@xxxxxxxxxxxx>

Many thanks!
-john