Andrey Smirnov wrote:
> On Wed, Feb 2, 2022 at 6:54 PM Thinh Nguyen <Thinh.Nguyen@xxxxxxxxxxxx> wrote:
>>
>> Andrey Smirnov wrote:
>>> It is necessary that:
>>>
>>> ROLE_SWITCH && device_property_read_bool(dwc->dev, "usb-role-switch")
>>>
>>> is true in order for dwc3_get_dr_mode() to _not_ force us from OTG to
>>> PERIPHERAL mode here:
>>>
>>> if (mode == USB_DR_MODE_OTG &&
>>> (!IS_ENABLED(CONFIG_USB_ROLE_SWITCH) ||
>>> !device_property_read_bool(dwc->dev, "usb-role-switch")) &&
>>> !DWC3_VER_IS_PRIOR(DWC3, 330A))
>>> mode = USB_DR_MODE_PERIPHERAL;
>>>
>>> and dwc3_drd_init() to be called later in dwc3_core_init_mode(). So,
>>> to avoid always ignoring extcon device returned by dwc3_get_extcon()
>>> change dwc3_drd_init() to check and use it first, before checking if
>>> dwc3_setup_role_switch() should be called.
>>>
>>> Cc: Felipe Balbi <balbi@xxxxxxxxxx>
>>> Cc: Thinh Nguyen <thinhn@xxxxxxxxxxxx>
>>> Cc: linux-usb@xxxxxxxxxxxxxxx
>>> Cc: linux-kernel@xxxxxxxxxxxxxxx
>>> Signed-off-by: Andrey Smirnov <andrew.smirnov@xxxxxxxxx>
>>> ---
>>>
>>> Hopefully I didn't miss something important making this patch
>>> unnecessary. Don't know if this is a good solution or not, part of me
>>> thinks than maybe changing the aforementioned code in
>>> dwc3_get_dr_mode() to account for extcon wopuld be
>>> simpler/better. Happy to rework this.
>>
>> The driver either use extcon or usb-role-switch. It doesn't make sense to
>> enable usb-role-switch for extcon just so the driver doesn't default
>> the dr_mode to peripheral.
>>
>> Perhaps, this is what you're looking for? (code is not tested)
>>
>
> Yeah that kind of what I meant by "maybe changing the aforementioned code in
> dwc3_get_dr_mode() to account for extcon wopuld be
> simpler/better." OK let me rework the patch to do that.
>
>> diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c
>> index bffd719b8b52..a52331ea7a0d 100644
>> --- a/drivers/usb/dwc3/core.c
>> +++ b/drivers/usb/dwc3/core.c
>> @@ -86,7 +86,7 @@ static int dwc3_get_dr_mode(struct dwc3 *dwc)
>> * mode. If the controller supports DRD but the dr_mode is not
>> * specified or set to OTG, then set the mode to peripheral.
>> */
>> - if (mode == USB_DR_MODE_OTG &&
>> + if (mode == USB_DR_MODE_OTG && !dwc->edev &&
>> (!IS_ENABLED(CONFIG_USB_ROLE_SWITCH) ||
>> !device_property_read_bool(dwc->dev, "usb-role-switch")) &&
>> !DWC3_VER_IS_PRIOR(DWC3, 330A))
>> @@ -1715,6 +1715,13 @@ static int dwc3_probe(struct platform_device *pdev)
>> goto err2;
>> }
>>
>> + dwc->edev = dwc3_get_extcon(dwc);
>> + if (IS_ERR(dwc->edev)) {
>> + ret = PTR_ERR(dwc->edev);
>> + dev_err(dwc->dev, "failed to get extcon %d\n", ret);
>> + goto err3;
>> + }
>> +
>> ret = dwc3_get_dr_mode(dwc);
>> if (ret)
>> goto err3;
>> diff --git a/drivers/usb/dwc3/drd.c b/drivers/usb/dwc3/drd.c
>> index 63089c7fb530..d02fcfdc74db 100644
>> --- a/drivers/usb/dwc3/drd.c
>> +++ b/drivers/usb/dwc3/drd.c
>> @@ -585,16 +585,7 @@ int dwc3_drd_init(struct dwc3 *dwc)
>> {
>> int ret, irq;
>>
>> - dwc->edev = dwc3_get_extcon(dwc);
>> - if (IS_ERR(dwc->edev))
>> - return PTR_ERR(dwc->edev);
>> -
>> - if (ROLE_SWITCH &&
>> - device_property_read_bool(dwc->dev, "usb-role-switch")) {
>> - ret = dwc3_setup_role_switch(dwc);
>> - if (ret < 0)
>> - return ret;
>> - } else if (dwc->edev) {
>
> We probably don't even need to move this around. If "usb-role-switch"
> isn't going to be set then that "if" could go first and not interfere.
>
Yeah, this is unnecessary.
>> + if (dwc->edev) {
>> dwc->edev_nb.notifier_call = dwc3_drd_notifier;
>> ret = extcon_register_notifier(dwc->edev, EXTCON_USB_HOST,
>> &dwc->edev_nb);
>> @@ -604,6 +595,11 @@ int dwc3_drd_init(struct dwc3 *dwc)
>> }
>>
>> dwc3_drd_update(dwc);
>> + } else if (ROLE_SWITCH &&
>> + device_property_read_bool(dwc->dev, "usb-role-switch")) {
>> + ret = dwc3_setup_role_switch(dwc);
>> + if (ret < 0)
>> + return ret;
>> } else {
>> dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_OTG);
>>
>>
Thanks,
Thinh
