linux-kernel - Re: [PATCH] mux: core: Fix double get

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:   Thu, 31 Aug 2017 23:57:25 +0200
From:   Peter Rosin <peda@...ntia.se>
To:     Hans de Goede <hdegoede@...hat.com>
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] mux: core: Fix double get_device()

On 2017-08-31 17:22, Hans de Goede wrote:
> class_find_device already does a get_device on the returned device.
> So the device returned by of_find_mux_chip_by_node is already referenced
> and we should not reference it again (and unref it on error).

Oops, yes that's right.

> Also rename of_find_mux_chip_by_node to of_get_mux_chip_by_node to
> make it clear that it returns a reference.

By that logic, you sign up for a lot of churn, these are but a few
examples:

s/of_find_device_by_node/of_get_device_by_node/
s/of_find_regulator_by_node/of_get_regulator_by_node/
s/of_find_spi_controller_by_node/of_get_spi_controller_by_node/
s/of_find_i2c_device_by_node/of_get_i2c_device_by_node/
s/of_find_i2c_adapter_by_node/of_get_i2c_adapter_by_node/

But there's already an of_get_i2c_adapter_by_node (that has slightly
different semantics) -> get ready for pushback.

So, I don't like the rename and would be much happier with a comment
above the function that the returned ref is supposed to be put away
when done.

Cheers,
Peter

> Signed-off-by: Hans de Goede <hdegoede@...hat.com>
> ---
>  drivers/mux/core.c | 7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/mux/core.c b/drivers/mux/core.c
> index 7aa656ac8fb5..8864cc745506 100644
> --- a/drivers/mux/core.c
> +++ b/drivers/mux/core.c
> @@ -433,7 +433,7 @@ static int of_dev_node_match(struct device *dev, const void *data)
>  	return dev->of_node == data;
>  }
>  
> -static struct mux_chip *of_find_mux_chip_by_node(struct device_node *np)
> +static struct mux_chip *of_get_mux_chip_by_node(struct device_node *np)
>  {
>  	struct device *dev;
>  
> @@ -471,7 +471,7 @@ static struct mux_control *of_mux_control_get(struct device *dev,
>  		return ERR_PTR(ret);
>  	}
>  
> -	mux_chip = of_find_mux_chip_by_node(args.np);
> +	mux_chip = of_get_mux_chip_by_node(args.np);
>  	of_node_put(args.np);
>  	if (!mux_chip)
>  		return ERR_PTR(-EPROBE_DEFER);
> @@ -480,6 +480,7 @@ static struct mux_control *of_mux_control_get(struct device *dev,
>  	    (!args.args_count && (mux_chip->controllers > 1))) {
>  		dev_err(dev, "%pOF: wrong #mux-control-cells for %pOF\n",
>  			np, args.np);
> +		put_device(&mux_chip->dev);
>  		return ERR_PTR(-EINVAL);
>  	}
>  
> @@ -490,10 +491,10 @@ static struct mux_control *of_mux_control_get(struct device *dev,
>  	if (controller >= mux_chip->controllers) {
>  		dev_err(dev, "%pOF: bad mux controller %u specified in %pOF\n",
>  			np, controller, args.np);
> +		put_device(&mux_chip->dev);
>  		return ERR_PTR(-EINVAL);
>  	}
>  
> -	get_device(&mux_chip->dev);
>  	return &mux_chip->mux[controller];
>  }
>  
>