linux-kernel - Re: [PATCH 7/7] clk: st: clkgen-pll: Add clk_unregister for odf_clk in clkgen_c32_pll

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <aWpe7MWiJlduga23@redhat.com>
Date: Fri, 16 Jan 2026 10:53:16 -0500
From: Brian Masney <bmasney@...hat.com>
To: Haoxiang Li <lihaoxiang@...c.iscas.ac.cn>
Cc: mturquette@...libre.com, sboyd@...nel.org, linux-clk@...r.kernel.org,
	linux-kernel@...r.kernel.org, stable@...r.kernel.org
Subject: Re: [PATCH 7/7] clk: st: clkgen-pll: Add clk_unregister for odf_clk
 in clkgen_c32_pll_setup()

Hi Haoxiang,

On Fri, Jan 16, 2026 at 07:38:47PM +0800, Haoxiang Li wrote:
> In clkgen_c32_pll_setup(), clkgen_odf_register() allocated
> clk_gate and clk_divider memory and registered a clk. Add
> clk_unregister() and kfree() to release the memory if
> error occurs. Initialize odf to zero for safe.
> 
> Fixes: b9b8e614b580 ("clk: st: Support for PLLs inside ClockGenA(s)")
> Cc: stable@...r.kernel.org
> Signed-off-by: Haoxiang Li <lihaoxiang@...c.iscas.ac.cn>
> ---
>  drivers/clk/st/clkgen-pll.c | 16 +++++++++++++---
>  1 file changed, 13 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/clk/st/clkgen-pll.c b/drivers/clk/st/clkgen-pll.c
> index 89f0454fa72e..3fc0af4b77c6 100644
> --- a/drivers/clk/st/clkgen-pll.c
> +++ b/drivers/clk/st/clkgen-pll.c
> @@ -761,10 +761,12 @@ static void __init clkgen_c32_pll_setup(struct device_node *np,
>  	struct clk *pll_clk;
>  	const char *parent_name, *pll_name;
>  	void __iomem *pll_base;
> -	int num_odfs, odf;
> +	int num_odfs, odf = 0;
>  	struct clk_onecell_data *clk_data;
>  	unsigned long pll_flags = 0;
>  	struct clkgen_pll *pll;
> +	struct clk_gate *gate;
> +	struct clk_divider *div;
>  
>  	parent_name = of_clk_get_parent_name(np, 0);
>  	if (!parent_name)
> @@ -808,7 +810,7 @@ static void __init clkgen_c32_pll_setup(struct device_node *np,
>  			if (of_property_read_string_index(np,
>  							  "clock-output-names",
>  							  odf, &clk_name))
> -				return;
> +				goto err_odf_unregister;
>  
>  			of_clk_detect_critical(np, odf, &odf_flags);
>  		}
> @@ -816,8 +818,8 @@ static void __init clkgen_c32_pll_setup(struct device_node *np,
>  		odf_clk = clkgen_odf_register(pll_name, pll_base, datac->data,
>  				odf_flags, odf, &clkgena_c32_odf_lock,
>  				clk_name);
> -			goto err;
>  		if (IS_ERR(odf_clk))
> +			goto err_odf_unregister;
>  
>  		clk_data->clks[odf] = odf_clk;
>  	}
> @@ -825,6 +827,14 @@ static void __init clkgen_c32_pll_setup(struct device_node *np,
>  	of_clk_add_provider(np, of_clk_src_onecell_get, clk_data);
>  	return;
>  
> +err_odf_unregister:
> +	while (--odf >= 0) {

I think the prefix -- is not appropriate here. If clkgen_odf_register()
fails for the first odf (ie odf=0), then when we jump to
err_odf_unregister, odf will still be set to 0, --odf will set it to -1,
the while loop will not run, and won't free anything.

What do you think about using the postfix operator instead?

	while (odf-- >= 0)

Brian