netdev - Re: [PATCH v2] net: mdio: fix unbalanced fwnode reference count in mdio_device

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] [day] [month] [year] [list]

Message-ID: <833ea8e7-f15b-ff66-84bf-2bcd77710fd8@kylinos.cn>
Date: Fri, 22 Nov 2024 15:59:08 +0800
From: luoxuanqiang <luoxuanqiang@...inos.cn>
To: Zeng Heng <zengheng4@...wei.com>, hkallweit1@...il.com,
 edumazet@...gle.com, pabeni@...hat.com, kuba@...nel.org,
 davem@...emloft.net, andrew@...n.ch, f.fainelli@...il.com,
 linux@...linux.org.uk
Cc: liwei391@...wei.com, netdev@...r.kernel.org
Subject: Re: [PATCH v2] net: mdio: fix unbalanced fwnode reference count in
 mdio_device_release()

Hi Heng,

在 2022/12/3 15:34, Zeng Heng 写道:
> There is warning report about of_node refcount leak
> while probing mdio device:
>
> OF: ERROR: memory leak, expected refcount 1 instead of 2,
> of_node_get()/of_node_put() unbalanced - destroy cset entry:
> attach overlay node /spi/soc@...dio@...700c0/ethernet@4
>
> In of_mdiobus_register_device(), we increase fwnode refcount
> by fwnode_handle_get() before associating the of_node with
> mdio device, but it has never been decreased in normal path.
> Since that, in mdio_device_release(), it needs to call
> fwnode_handle_put() in addition instead of calling kfree()
> directly.
>
> After above, just calling mdio_device_free() in the error handle
> path of of_mdiobus_register_device() is enough to keep the
> refcount balanced.
>
> Fixes: a9049e0c513c ("mdio: Add support for mdio drivers.")
> Signed-off-by: Zeng Heng <zengheng4@...wei.com>
> Reviewed-by: Yang Yingliang <yangyingliang@...wei.com>
> ---
>   changes in v2:
>    - Add operation about setting device node as NULL-pointer.
>      There is no practical changes.
>    - Add reviewed-by tag.
> ---
>   drivers/net/mdio/of_mdio.c    | 3 ++-
>   drivers/net/phy/mdio_device.c | 2 ++
>   2 files changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/mdio/of_mdio.c b/drivers/net/mdio/of_mdio.c
> index 796e9c7857d0..510822d6d0d9 100644
> --- a/drivers/net/mdio/of_mdio.c
> +++ b/drivers/net/mdio/of_mdio.c
> @@ -68,8 +68,9 @@ static int of_mdiobus_register_device(struct mii_bus *mdio,
>   	/* All data is now stored in the mdiodev struct; register it. */
>   	rc = mdio_device_register(mdiodev);
>   	if (rc) {
> +		device_set_node(&mdiodev->dev, NULL);
> +		fwnode_handle_put(fwnode);
According to my understanding, the process flow of mdio_device_free() is
as follows:
mdio_device_free()
     -> put_device()
         ->kobject_put()
             ->kobject_release()
                 ->device_release()
                     ->mdio_device_release()
                     //Here, it will be called once fwnode_handle_put();

Why is it necessary to add fwnode_handle_put(fwnode) again here? In the
body of your email, you described that mdio_device_free() is sufficient
to keep refcount balanced, and it was not added in the v1.
Could you please explain the reason for this?
I am looking forward to your response :)

Thanks
Xuanqiang
>   		mdio_device_free(mdiodev);
> -		of_node_put(child);
>   		return rc;
>   	}
>   
> diff --git a/drivers/net/phy/mdio_device.c b/drivers/net/phy/mdio_device.c
> index 250742ffdfd9..044828d081d2 100644
> --- a/drivers/net/phy/mdio_device.c
> +++ b/drivers/net/phy/mdio_device.c
> @@ -21,6 +21,7 @@
>   #include <linux/slab.h>
>   #include <linux/string.h>
>   #include <linux/unistd.h>
> +#include <linux/property.h>
>   
>   void mdio_device_free(struct mdio_device *mdiodev)
>   {
> @@ -30,6 +31,7 @@ EXPORT_SYMBOL(mdio_device_free);
>   
>   static void mdio_device_release(struct device *dev)
>   {
> +	fwnode_handle_put(dev->fwnode);
>   	kfree(to_mdio_device(dev));
>   }
>