Hi, On 9/11/23 14:56, Andy Shevchenko wrote: > On Mon, Sep 11, 2023 at 02:49:44PM +0200, Bartosz Golaszewski wrote: >> On Sun, Sep 10, 2023 at 1:26 PM Hans de Goede <hdegoede@xxxxxxxxxx> wrote: >>> On 9/10/23 10:24, Andy Shevchenko wrote: >>>> On Sat, Sep 9, 2023 at 5:18 PM Hans de Goede <hdegoede@xxxxxxxxxx> wrote: > > ... > >>>>> + gpiod = devm_gpiod_get(&x86_android_tablet_device->dev, con_id, dflags); > > ^^^ > >>>>> - *desc = gpiod; >>>>> + if (desc) >>>>> + *desc = gpiod; >>>> >>>> Are we expecting that callers may not want the GPIO descriptor out of >>>> this function? >>>> Sounds a bit weird if so. >>> >>> Yes specifically the Charge-enable and OTG (Vboost enable) pins on the >>> bq25892 charger on the Lenovo Yoga Tab 3 just need to be set to a fixed >>> value, so we request the pins with GPIOD_OUT_LOW / _HIGH and then >>> leave them at that value. >> >> You mean you leak the descriptor? It would warrant either a comment or >> storing the descriptor somewhere and cleaning it up if the device can >> be unbound (I guess it can since the driver can be built as module). > > Note devm_*() above. Right, the GPIOs will be free-ed on unbound through the devm framework. Regards, Hans
