Re: [PATCH] spmi: hisi-spmi-controller: fix potential memory leak in spmi_controller_probe()

[Stephen Boyd]

Quoting Ma Ke (2023-09-21 19:52:16)
> spmi_controller_alloc() allocates a memory space for ctrl. When some
> errors occur, ctrl should be handled by spmi_controller_put() and set
> spmi_controller->controller = NULL because spmi_controller->controller
> has a dangling pointer to the freed memory. When the failure happens,
> the function returns without calling spmi_controller_put() and setting
> spmi_controller->controller = NULL, which will lead to a memory leak.
> 
> When the failure happens, we can fix it by calling spmi_controller_put()
> and setting spmi_controller->controller = NULL in all of the places
> where we call spmi_controller_put().
> 
> Signed-off-by: Ma Ke <make_ruc2021@xxxxxxx>
> ---
>  drivers/spmi/hisi-spmi-controller.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/spmi/hisi-spmi-controller.c b/drivers/spmi/hisi-spmi-controller.c
> index 9cbd473487cb..5b3cfa7f5056 100644
> --- a/drivers/spmi/hisi-spmi-controller.c
> +++ b/drivers/spmi/hisi-spmi-controller.c
> @@ -321,6 +321,7 @@ static int spmi_controller_probe(struct platform_device *pdev)
>  
>  err_put_controller:
>         spmi_controller_put(ctrl);
> +       spmi_controller->controller = NULL;

The controller will most likely be freed after the put call one line
above so this will oops. No thanks.