Re: [PATCH linux-next] phy/tegra: add put_device() after of_find_device_by_node()

[Thierry Reding <thierry.reding@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 1799 bytes --]

On Thu, Aug 19, 2021 at 07:46:33PM -0700, jing yangyang wrote:
> This was found by coccicheck:
> ./drivers/phy/tegra/xusb-tegra210.c:3174: 2-8:ERROR:
> missing put_device;call of_find_device_by_node on line 3167,
> but without a corresponding object release within this function.
> 
> Reported-by: Zeal Robot <zealci@zte.com.cn>
> Signed-off-by: jing yangyang <jing.yangyang@zte.com.cn>
> ---
>  drivers/phy/tegra/xusb-tegra210.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/phy/tegra/xusb-tegra210.c b/drivers/phy/tegra/xusb-tegra210.c
> index eedfc7c..f26eb5a 100644
> --- a/drivers/phy/tegra/xusb-tegra210.c
> +++ b/drivers/phy/tegra/xusb-tegra210.c
> @@ -3170,8 +3170,10 @@ static int tegra210_utmi_port_reset(struct phy *phy)
>  		goto out;
>  	}
>  
> -	if (!platform_get_drvdata(pdev))
> +	if (!platform_get_drvdata(pdev)) {
> +		put_device(&pdev->dev);
>  		return ERR_PTR(-EPROBE_DEFER);
> +	}
>  
>  	padctl->regmap = dev_get_regmap(&pdev->dev, "usb_sleepwalk");
>  	if (!padctl->regmap)

I think we also need that put_device() after getting the regmap because
we loose the reference to the platform device after that.

On the other hand, I suppose we might have to hang on to the reference
until tegra210_xusb_padctl_remove() so that the platform device that the
regmap is provided by doesn't suddenly go away.

Mark, I couldn't find how the lifetime of regmaps is getting tracked.
dev_get_regmap() seems to only return a pointer to the regmap if it
exists, but doesn't actually increment a reference or anything. So do we
need to hold on to the struct device that we get it from until we no
longer need the reference? Or is there some other way to make sure the
regmap doesn't disappear on us?

Thierry

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]