Re: [PATCH v3] clk: starfive: jh7110: fix memory leak in jh7110_reset_controller_register() error path

[Brian Masney <bmasney@redhat.com>]

Hi Guangshuo,

I missed that you sent a new version. My same comment from the v2 still
applies. See below for details.

On Mon, Apr 13, 2026 at 10:36:43PM +0800, Guangshuo Li wrote:
> jh7110_reset_controller_register() allocates a jh71x0_reset_adev with
> kzalloc() and sets jh7110_reset_adev_release() as the release callback
> for its embedded auxiliary_device before calling auxiliary_device_init().
> 
> If auxiliary_device_init() fails, the function returns immediately
> without freeing the allocated rdev. The release callback is not
> available for this path, because it is only reached after a successful
> auxiliary_device_init(), for example when auxiliary_device_add() fails
> and auxiliary_device_uninit() is called.
> 
> The issue was identified by a static analysis tool I developed and
> confirmed by manual review. Free rdev explicitly when
> auxiliary_device_init() returns an error.
> 
> Fixes: edab7204afe5 ("clk: starfive: Add StarFive JH7110 system clock driver")
> Cc: stable@vger.kernel.org
> Signed-off-by: Guangshuo Li <lgs201920130244@gmail.com>
> ---
> v3:
>   - clarify the changelog to describe the exact failure path
>   - note that the issue was identified by a static analysis tool
>     developed by me and confirmed by manual review
>   - apologize for sending the initial public posting as v2 by mistake
> 
> v2:
>   - initial public posting; v1 was mistakenly skipped
> 
>  drivers/clk/starfive/clk-starfive-jh7110-sys.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/clk/starfive/clk-starfive-jh7110-sys.c b/drivers/clk/starfive/clk-starfive-jh7110-sys.c
> index 52833d4241c5..55cd0ccbdb84 100644
> --- a/drivers/clk/starfive/clk-starfive-jh7110-sys.c
> +++ b/drivers/clk/starfive/clk-starfive-jh7110-sys.c
> @@ -360,8 +360,10 @@ int jh7110_reset_controller_register(struct jh71x0_clk_priv *priv,
>  	adev->id = adev_id;
>  
>  	ret = auxiliary_device_init(adev);
> -	if (ret)
> +	if (ret) {
> +		kfree(rdev);
>  		return ret;
> +	}
>  
>  	ret = auxiliary_device_add(adev);
>  	if (ret) {

There's actually another leak in the error path for
auxiliary_device_add(). I think this code should be
converted to devm_kzalloc().

There is no devm_kzalloc_obj() yet, however according to [1] that should
be coming soon.

[1] https://lore.kernel.org/lkml/20260330154108.GA3389518@killaraus.ideasonboard.com/

Brian