Re: [PATCH] clk: bcm2835: Fix memory leak in bcm2835_register_pll

[Alexander Stein]

Hello,

On Thursday, October 10, 2019, 3:30:58 AM CEST Navid Emamdoost wrote:
> In the implementation of bcm2835_register_pll(), the allocated memory
> for pll should be released if devm_clk_hw_register() fails.
> 
> Fixes: b19f009d4510 ("clk: bcm2835: Migrate to clk_hw based registration and OF APIs")
> Signed-off-by: Navid Emamdoost <navid.emamdoost@xxxxxxxxx>
> ---
>  drivers/clk/bcm/clk-bcm2835.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/clk/bcm/clk-bcm2835.c b/drivers/clk/bcm/clk-bcm2835.c
> index 802e488fd3c3..99549642110a 100644
> --- a/drivers/clk/bcm/clk-bcm2835.c
> +++ b/drivers/clk/bcm/clk-bcm2835.c
> @@ -1320,8 +1320,10 @@ static struct clk_hw *bcm2835_register_pll(struct bcm2835_cprman *cprman,
>  	pll->hw.init = &init;
>  
>  	ret = devm_clk_hw_register(cprman->dev, &pll->hw);
> -	if (ret)
> +	if (ret) {
> +		kfree(pll);
>  		return NULL;
> +	}
>  	return &pll->hw;
>  }

Eh, is pll freed at all, even in successful case? I failed to find a corresponding kfree().
The pointer itself is lost once the function returns.
The solution would rather be to use devm_kzalloc instead of kzalloc, like the other clocks
in e.g. bcm2835_register_pll()

Best regards,
Alexander