Re: [PATCH v2] serial: qcom_geni: Fix BT failure regression on RB2 platform

[Mayank Rana <mayank.rana@oss.qualcomm.com>]

On 1/7/2026 8:26 AM, Praveen Talari wrote:
> Commit 10904d725f6e ("serial: qcom-geni: Enable PM runtime for serial
> driver") caused BT init to fail during bootup on the RB2 platform,
> preventing proper BT initialization. However, BT works correctly after
> bootup completes.
> 
> The issue occurs when runtime PM is enabled and uart_add_one_port() is
> called before wakeup IRQ setup. The uart_add_one_port() call activates
> the device through runtime PM, which configures GPIOs to the "qup_x"
> pinmux function during runtime resume. When wakeup IRQ registration
> happens afterward using dev_pm_set_dedicated_wake_irq(), these GPIOs
> are reset back to the "gpio" pinmux function, which impacts the RX GPIO
> and leads to Bluetooth failures.
> 
> Fix this by ensuring wakeup IRQ setup is completed before calling
> uart_add_one_port() to prevent the pinmux function conflict.
> 
> Tested-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> Reported-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> Closes: https://lore.kernel.org/all/20251110101043.2108414-4-praveen.talari@oss.qualcomm.com/
> Fixes: 10904d725f6e ("serial: qcom-geni: Enable PM runtime for serial driver")
> Signed-off-by: Praveen Talari <praveen.talari@oss.qualcomm.com>
> ---
> v1 -> v2
> - Updated commit text.
> - Added Tested-by in commit text
> ---
>   drivers/tty/serial/qcom_geni_serial.c | 12 ++++++------
>   1 file changed, 6 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/tty/serial/qcom_geni_serial.c b/drivers/tty/serial/qcom_geni_serial.c
> index 6ce6528f5c10..46a9c71630d5 100644
> --- a/drivers/tty/serial/qcom_geni_serial.c
> +++ b/drivers/tty/serial/qcom_geni_serial.c
> @@ -1888,12 +1888,6 @@ static int qcom_geni_serial_probe(struct platform_device *pdev)
>   	if (ret)
>   		goto error;
>   
> -	devm_pm_runtime_enable(port->se.dev);
> -
> -	ret = uart_add_one_port(drv, uport);
> -	if (ret)
> -		goto error;
> -
>   	if (port->wakeup_irq > 0) {
>   		device_init_wakeup(&pdev->dev, true);
>   		ret = dev_pm_set_dedicated_wake_irq(&pdev->dev,

do you need to handle error handling part here as calling 
uart_remove_one_port() shall result into NULL access related crash while 
accessing port->port_dev as port_dev is not allocated 
(serial_base_port_add() add is not being called due to moving 
uart_add_one_port() later)

--- a/drivers/tty/serial/qcom_geni_serial.c
+++ b/drivers/tty/serial/qcom_geni_serial.c
@@ -1901,7 +1901,6 @@ static int qcom_geni_serial_probe(struct 
platform_device *pdev)
                 if (ret) {
                         device_init_wakeup(&pdev->dev, false);
                         ida_free(&port_ida, uport->line);
-                       uart_remove_one_port(drv, uport);
                         goto error;
                 }
         }

> @@ -1906,6 +1900,12 @@ static int qcom_geni_serial_probe(struct platform_device *pdev)
>   		}
>   	}
>   
> +	devm_pm_runtime_enable(port->se.dev);
> +
> +	ret = uart_add_one_port(drv, uport);
> +	if (ret)
> +		goto error;
> +
>   	return 0;
>   
>   error:
> 
> base-commit: 6cd6c12031130a349a098dbeb19d8c3070d2dfbe
Regards,
Mayank