linux-kernel - Re: [PATCH v2] gpio: vt8500: memory cleanup missing

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]
Date:	Mon, 14 Jan 2013 14:14:22 +0000
From:	Grant Likely <grant.likely@...retlab.ca>
To:	Tony Prisk <linux@...sktech.co.nz>,
	Linux Walleij <linus.walleij@...aro.org>
Cc:	vt8500-wm8505-linux-kernel@...glegroups.com,
	linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
	linux@....linux.org.uk, Tony Prisk <linux@...sktech.co.nz>
Subject: Re: [PATCH v2] gpio: vt8500: memory cleanup missing

On Fri, 11 Jan 2013 08:09:46 +1300, Tony Prisk <linux@...sktech.co.nz> wrote:
> This driver is missing a .remove callback, and the fail path on
> probe is incomplete.
> 
> If an error occurs in vt8500_add_chips, gpio_base is not unmapped.
> The driver is also ignoring the return value from this function so
> if a chip fails to register it completes as successful.
> 
> Replaced pr_err with dev_err in vt8500_add_chips since the device is
> available.
> 
> There is also no .remove callback defined. To allow removing the
> registered chips, I have moved *vtchip to be a static global.
> 
> Signed-off-by: Tony Prisk <linux@...sktech.co.nz>
> ---
> v2:
> Remove global variable and use platform_set_drvdata instead.
> 
>  drivers/gpio/gpio-vt8500.c |   51 +++++++++++++++++++++++++++++++++++---------
>  1 file changed, 41 insertions(+), 10 deletions(-)
> 
> diff --git a/drivers/gpio/gpio-vt8500.c b/drivers/gpio/gpio-vt8500.c
> index b53320a..87e59b5 100644
> --- a/drivers/gpio/gpio-vt8500.c
> +++ b/drivers/gpio/gpio-vt8500.c
> @@ -233,10 +233,12 @@ static int vt8500_add_chips(struct platform_device *pdev, void __iomem *base,
>  			sizeof(struct vt8500_gpio_chip) * data->num_banks,
>  			GFP_KERNEL);
>  	if (!vtchip) {
> -		pr_err("%s: failed to allocate chip memory\n", __func__);
> +		dev_err(&pdev->dev, "failed to allocate chip memory\n");
>  		return -ENOMEM;
>  	}
>  
> +	platform_set_drvdata(pdev, vtchip);
> +
>  	for (i = 0; i < data->num_banks; i++) {
>  		vtchip[i].base = base;
>  		vtchip[i].regs = &data->banks[i];
> @@ -261,6 +263,7 @@ static int vt8500_add_chips(struct platform_device *pdev, void __iomem *base,
>  
>  		gpiochip_add(chip);
>  	}
> +
>  	return 0;
>  }
>  

Watch out; irrelevant whitespace change here. From a maintainer point of
voiew, I'm less confident about a patch when I see unrelated whitespace
changes because it suggests that there are things in the patch that you
don't intend. It helps me out a lot if this stuff gets scrubbed before I
see it.

> @@ -273,36 +276,64 @@ static struct of_device_id vt8500_gpio_dt_ids[] = {
>  
>  static int vt8500_gpio_probe(struct platform_device *pdev)
>  {
> +	int ret;
>  	void __iomem *gpio_base;
> -	struct device_node *np;
> +	struct device_node *np = pdev->dev.of_node;
>  	const struct of_device_id *of_id =
>  				of_match_device(vt8500_gpio_dt_ids, &pdev->dev);
>  
> -	if (!of_id) {
> -		dev_err(&pdev->dev, "Failed to find gpio controller\n");
> +	if (!np) {
> +		dev_err(&pdev->dev, "GPIO node missing in devicetree\n");
>  		return -ENODEV;
>  	}
>  
> -	np = pdev->dev.of_node;
> -	if (!np) {
> -		dev_err(&pdev->dev, "Missing GPIO description in devicetree\n");
> -		return -EFAULT;
> +	if (!of_id) {
> +		dev_err(&pdev->dev, "No matching driver data\n");
> +		return -ENODEV;
>  	}

Why is this flipped around? I don't see any functional reason for this
change.

In actual fact, since the driver needs both it only needs to test for
the of_id. If there is no node, then of_id will never be set.

>  
>  	gpio_base = of_iomap(np, 0);
>  	if (!gpio_base) {
>  		dev_err(&pdev->dev, "Unable to map GPIO registers\n");
> -		of_node_put(np);
>  		return -ENOMEM;
>  	}
>  
> -	vt8500_add_chips(pdev, gpio_base, of_id->data);
> +	ret = vt8500_add_chips(pdev, gpio_base, of_id->data);
> +	if (ret) {
> +		iounmap(gpio_base);
> +		return ret;
> +	}
> +
> +	return 0;
> +}
> +
> +static int vt8500_gpio_remove(struct platform_device *pdev)
> +{
> +	int i;
> +	int ret;
> +	const struct vt8500_gpio_data *data;
> +	struct vt8500_gpio_chip *vtchip = platform_get_drvdata(pdev);
> +	void __iomem *gpio_base = vtchip[0].base;
> +	const struct of_device_id *of_id =
> +				of_match_device(vt8500_gpio_dt_ids, &pdev->dev);
> +
> +	data = of_id->data;
> +
> +	for (i = 0; i < data->num_banks; i++) {

It would make for simpler code all around if num_banks was cached in the
vt8500_gpio_chip structure during the .probe() routine. It looks wrong
to be calling of_match_device() in the remove hook.

Otherwise this iteration looks much better.

g.

> +		ret = gpiochip_remove(&vtchip[i].chip);
> +		if (ret)
> +			dev_warn(&pdev->dev, "gpiochip_remove returned %d\n",
> +				 ret);
> +	}
> +
> +	iounmap(gpio_base);
>  
>  	return 0;
>  }
>  
>  static struct platform_driver vt8500_gpio_driver = {
>  	.probe		= vt8500_gpio_probe,
> +	.remove		= vt8500_gpio_remove,
>  	.driver		= {
>  		.name	= "vt8500-gpio",
>  		.owner	= THIS_MODULE,
> -- 
> 1.7.9.5
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@...r.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

-- 
Grant Likely, B.Sc, P.Eng.
Secret Lab Technologies, Ltd.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/