On 12/12/2024 19:04:03+0900, Joe Hattori wrote:
> Current code leaves the device's wakeup enabled in the error path of
> .probe(), which results in a memory leak. Call device_init_wakeup() as
> the last step in the .probe() to avoid this leak.
Do you have more info on where the memory is allocated?
Coudln't we have a devm_ version of device_init_wakeup instead?
>
> This bug was found by an experimental static analysis tool that I am
> developing.
>
> Fixes: 32a4a4ebf768 ("rtc: bd70528: Initial support for ROHM bd70528 RTC")
> Signed-off-by: Joe Hattori <joe@xxxxxxxxxxxxxxxxxxxxx>
> ---
> Changes in V2:
> - Move the device_init_wakeup() to the last step of the .probe() to make
> the cleanup simpler.
> ---
> drivers/rtc/rtc-bd70528.c | 10 ++++++----
> 1 file changed, 6 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/rtc/rtc-bd70528.c b/drivers/rtc/rtc-bd70528.c
> index 954ac4ef53e8..d5cc4993f918 100644
> --- a/drivers/rtc/rtc-bd70528.c
> +++ b/drivers/rtc/rtc-bd70528.c
> @@ -312,9 +312,6 @@ static int bd70528_probe(struct platform_device *pdev)
> }
> }
>
> - device_set_wakeup_capable(&pdev->dev, true);
> - device_wakeup_enable(&pdev->dev);
> -
> rtc = devm_rtc_allocate_device(&pdev->dev);
> if (IS_ERR(rtc)) {
> dev_err(&pdev->dev, "RTC device creation failed\n");
> @@ -331,7 +328,12 @@ static int bd70528_probe(struct platform_device *pdev)
> if (ret)
> return ret;
>
> - return devm_rtc_register_device(rtc);
> + ret = devm_rtc_register_device(rtc);
> + if (ret)
> + return ret;
> +
> + device_init_wakeup(&pdev->dev, true);
> + return 0;
> }
>
> static const struct platform_device_id bd718x7_rtc_id[] = {
> --
> 2.34.1
>
--
Alexandre Belloni, co-owner and COO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
