[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240510174937.0a710104@donnerap.manchester.arm.com>
Date: Fri, 10 May 2024 17:49:37 +0100
From: Andre Przywara <andre.przywara@....com>
To: Javier Carrasco <javier.carrasco.cruz@...il.com>
Cc: Yangtao Li <tiny.windzz@...il.com>, "Rafael J. Wysocki"
<rafael@...nel.org>, Viresh Kumar <viresh.kumar@...aro.org>, Chen-Yu Tsai
<wens@...e.org>, Jernej Skrabec <jernej.skrabec@...il.com>, Samuel Holland
<samuel@...lland.org>, Jonathan Cameron <Jonathan.Cameron@...wei.com>,
linux-pm@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
linux-sunxi@...ts.linux.dev, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/2] cpufreq: sun50i: fix memory leak in
dt_has_supported_hw()
On Fri, 03 May 2024 19:52:32 +0200
Javier Carrasco <javier.carrasco.cruz@...il.com> wrote:
Hi Javier,
> The for_each_child_of_node() loop does not decrement the child node
> refcount before the break instruction, even though the node is no
> longer required.
Ah, thanks for spotting this, there is indeed a leak. Sorry for the
blunder!
> This can be avoided with the new for_each_child_of_node_scoped() macro
> that removes the need for any of_node_put().
Wow, that's the typical convoluted Linux macro, but it looks correct to me.
It would call the put even if the loop ends naturally, but there is a NULL
test in there, so that's fine.
> Fixes: fa5aec9561cf ("cpufreq: sun50i: Add support for opp_supported_hw")
> Signed-off-by: Javier Carrasco <javier.carrasco.cruz@...il.com>
Reviewed-by: Andre Przywara <andre.przywara@....com>
Thanks!
Andre
> ---
> drivers/cpufreq/sun50i-cpufreq-nvmem.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/cpufreq/sun50i-cpufreq-nvmem.c b/drivers/cpufreq/sun50i-cpufreq-nvmem.c
> index 0b882765cd66..ef83e4bf2639 100644
> --- a/drivers/cpufreq/sun50i-cpufreq-nvmem.c
> +++ b/drivers/cpufreq/sun50i-cpufreq-nvmem.c
> @@ -131,7 +131,7 @@ static const struct of_device_id cpu_opp_match_list[] = {
> static bool dt_has_supported_hw(void)
> {
> bool has_opp_supported_hw = false;
> - struct device_node *np, *opp;
> + struct device_node *np;
> struct device *cpu_dev;
>
> cpu_dev = get_cpu_device(0);
> @@ -142,7 +142,7 @@ static bool dt_has_supported_hw(void)
> if (!np)
> return false;
>
> - for_each_child_of_node(np, opp) {
> + for_each_child_of_node_scoped(np, opp) {
> if (of_find_property(opp, "opp-supported-hw", NULL)) {
> has_opp_supported_hw = true;
> break;
>
Powered by blists - more mailing lists