[PATCH 1/2] usb: gadget: udc: atmel: check rc of devm_gpiod_get_optional()

[PATCH 1/2] usb: gadget: udc: atmel: check rc of devm_gpiod_get_optional()

[Michael Walle]

devm_gpiod_get_optional() might still return an error code, esp.
EPROBE_DEFER. Return any errors.

Signed-off-by: Michael Walle <michael@walle.cc>
---
 drivers/usb/gadget/udc/atmel_usba_udc.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/usb/gadget/udc/atmel_usba_udc.c b/drivers/usb/gadget/udc/atmel_usba_udc.c
index ae2bfbac603e..48355e0cee76 100644
--- a/drivers/usb/gadget/udc/atmel_usba_udc.c
+++ b/drivers/usb/gadget/udc/atmel_usba_udc.c
@@ -2165,6 +2165,8 @@ static struct usba_ep * atmel_udc_of_init(struct platform_device *pdev,
 
 	udc->vbus_pin = devm_gpiod_get_optional(&pdev->dev, "atmel,vbus",
 						GPIOD_IN);
+	if (IS_ERR(udc->vbus_pin))
+		return ERR_CAST(udc->vbus_pin);
 
 	if (fifo_mode == 0) {
 		udc->num_ep = udc_config->num_ep;
-- 
2.30.2

[PATCH 2/2] usb: gadget: udc: atmel: convert to platform driver

[Michael Walle]

The driver won't probe on a LAN9668 because the pinctrl driver isn't
ready yet. Probe deferral is not supported because the init section
is already discarded. With fw_devlink enabled, the probe won't even
be called. Convert the driver to a proper platform driver.

Signed-off-by: Michael Walle <michael@walle.cc>
---
 drivers/usb/gadget/udc/atmel_usba_udc.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/usb/gadget/udc/atmel_usba_udc.c b/drivers/usb/gadget/udc/atmel_usba_udc.c
index 48355e0cee76..53ca38c4b3ec 100644
--- a/drivers/usb/gadget/udc/atmel_usba_udc.c
+++ b/drivers/usb/gadget/udc/atmel_usba_udc.c
@@ -2060,7 +2060,7 @@ static const struct usba_udc_errata at91sam9g45_errata = {
 	.pulse_bias = at91sam9g45_pulse_bias,
 };
 
-static const struct usba_ep_config ep_config_sam9[] __initconst = {
+static const struct usba_ep_config ep_config_sam9[] = {
 	{ .nr_banks = 1 },				/* ep 0 */
 	{ .nr_banks = 2, .can_dma = 1, .can_isoc = 1 },	/* ep 1 */
 	{ .nr_banks = 2, .can_dma = 1, .can_isoc = 1 },	/* ep 2 */
@@ -2070,7 +2070,7 @@ static const struct usba_ep_config ep_config_sam9[] __initconst = {
 	{ .nr_banks = 3, .can_dma = 1, .can_isoc = 1 },	/* ep 6 */
 };
 
-static const struct usba_ep_config ep_config_sama5[] __initconst = {
+static const struct usba_ep_config ep_config_sama5[] = {
 	{ .nr_banks = 1 },				/* ep 0 */
 	{ .nr_banks = 3, .can_dma = 1, .can_isoc = 1 },	/* ep 1 */
 	{ .nr_banks = 3, .can_dma = 1, .can_isoc = 1 },	/* ep 2 */
@@ -2449,6 +2449,7 @@ static int usba_udc_resume(struct device *dev)
 static SIMPLE_DEV_PM_OPS(usba_udc_pm_ops, usba_udc_suspend, usba_udc_resume);
 
 static struct platform_driver udc_driver = {
+	.probe		= usba_udc_probe,
 	.remove		= usba_udc_remove,
 	.driver		= {
 		.name		= "atmel_usba_udc",
@@ -2456,8 +2457,7 @@ static struct platform_driver udc_driver = {
 		.of_match_table	= atmel_udc_dt_ids,
 	},
 };
-
-module_platform_driver_probe(udc_driver, usba_udc_probe);
+module_platform_driver(udc_driver);
 
 MODULE_DESCRIPTION("Atmel USBA UDC driver");
 MODULE_AUTHOR("Haavard Skinnemoen (Atmel)");
-- 
2.30.2

Re: [PATCH 1/2] usb: gadget: udc: atmel: check rc of devm_gpiod_get_optional()

[Saravana Kannan]

On Tue, Jul 5, 2022 at 6:19 AM Michael Walle <michael@walle.cc> wrote:
>
> devm_gpiod_get_optional() might still return an error code, esp.
> EPROBE_DEFER. Return any errors.
>
> Signed-off-by: Michael Walle <michael@walle.cc>
> ---
>  drivers/usb/gadget/udc/atmel_usba_udc.c | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/drivers/usb/gadget/udc/atmel_usba_udc.c b/drivers/usb/gadget/udc/atmel_usba_udc.c
> index ae2bfbac603e..48355e0cee76 100644
> --- a/drivers/usb/gadget/udc/atmel_usba_udc.c
> +++ b/drivers/usb/gadget/udc/atmel_usba_udc.c
> @@ -2165,6 +2165,8 @@ static struct usba_ep * atmel_udc_of_init(struct platform_device *pdev,
>
>         udc->vbus_pin = devm_gpiod_get_optional(&pdev->dev, "atmel,vbus",
>                                                 GPIOD_IN);
> +       if (IS_ERR(udc->vbus_pin))
> +               return ERR_CAST(udc->vbus_pin);

I'm confused. Is it really an optional resource if you treat a failure to get it
as a reason to fail a probe?

-Saravana

>
>         if (fifo_mode == 0) {
>                 udc->num_ep = udc_config->num_ep;
> --
> 2.30.2
>

Re: [PATCH 1/2] usb: gadget: udc: atmel: check rc of devm_gpiod_get_optional()

[Michael Walle]

Am 2022-07-05 20:53, schrieb Saravana Kannan:
> On Tue, Jul 5, 2022 at 6:19 AM Michael Walle <michael@walle.cc> wrote:
>> 
>> devm_gpiod_get_optional() might still return an error code, esp.
>> EPROBE_DEFER. Return any errors.
>> 
>> Signed-off-by: Michael Walle <michael@walle.cc>
>> ---
>>  drivers/usb/gadget/udc/atmel_usba_udc.c | 2 ++
>>  1 file changed, 2 insertions(+)
>> 
>> diff --git a/drivers/usb/gadget/udc/atmel_usba_udc.c 
>> b/drivers/usb/gadget/udc/atmel_usba_udc.c
>> index ae2bfbac603e..48355e0cee76 100644
>> --- a/drivers/usb/gadget/udc/atmel_usba_udc.c
>> +++ b/drivers/usb/gadget/udc/atmel_usba_udc.c
>> @@ -2165,6 +2165,8 @@ static struct usba_ep * atmel_udc_of_init(struct 
>> platform_device *pdev,
>> 
>>         udc->vbus_pin = devm_gpiod_get_optional(&pdev->dev, 
>> "atmel,vbus",
>>                                                 GPIOD_IN);
>> +       if (IS_ERR(udc->vbus_pin))
>> +               return ERR_CAST(udc->vbus_pin);
> 
> I'm confused. Is it really an optional resource if you treat a failure 
> to get it
> as a reason to fail a probe?

If the gpio isn't found NULL is returned.

-michael

Re: [PATCH 1/2] usb: gadget: udc: atmel: check rc of devm_gpiod_get_optional()

[Saravana Kannan]

On Tue, Jul 5, 2022 at 1:56 PM Michael Walle <michael@walle.cc> wrote:
>
> Am 2022-07-05 20:53, schrieb Saravana Kannan:
> > On Tue, Jul 5, 2022 at 6:19 AM Michael Walle <michael@walle.cc> wrote:
> >>
> >> devm_gpiod_get_optional() might still return an error code, esp.
> >> EPROBE_DEFER. Return any errors.
> >>
> >> Signed-off-by: Michael Walle <michael@walle.cc>
> >> ---
> >>  drivers/usb/gadget/udc/atmel_usba_udc.c | 2 ++
> >>  1 file changed, 2 insertions(+)
> >>
> >> diff --git a/drivers/usb/gadget/udc/atmel_usba_udc.c
> >> b/drivers/usb/gadget/udc/atmel_usba_udc.c
> >> index ae2bfbac603e..48355e0cee76 100644
> >> --- a/drivers/usb/gadget/udc/atmel_usba_udc.c
> >> +++ b/drivers/usb/gadget/udc/atmel_usba_udc.c
> >> @@ -2165,6 +2165,8 @@ static struct usba_ep * atmel_udc_of_init(struct
> >> platform_device *pdev,
> >>
> >>         udc->vbus_pin = devm_gpiod_get_optional(&pdev->dev,
> >> "atmel,vbus",
> >>                                                 GPIOD_IN);
> >> +       if (IS_ERR(udc->vbus_pin))
> >> +               return ERR_CAST(udc->vbus_pin);
> >
> > I'm confused. Is it really an optional resource if you treat a failure
> > to get it
> > as a reason to fail a probe?
>
> If the gpio isn't found NULL is returned.

Ah, ok.

-Saravana