linux-kernel - Re: [PATCH 3/3] staging:iio:ad2s1210: Add write

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] [thread-next>] [day] [month] [year] [list]

Message-ID: <20180313090249.cdurikolfwc5qyol@mwanda>
Date:   Tue, 13 Mar 2018 12:02:49 +0300
From:   Dan Carpenter <dan.carpenter@...cle.com>
To:     Rodrigo Siqueira <rodrigosiqueiramelo@...il.com>
Cc:     Lars-Peter Clausen <lars@...afoo.de>,
        Michael Hennerich <Michael.Hennerich@...log.com>,
        Jonathan Cameron <jic23@...nel.org>,
        Hartmut Knaack <knaack.h@....de>,
        Peter Meerwald-Stadler <pmeerw@...erw.net>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Graff Yang <graff.yang@...il.com>, linux-iio@...r.kernel.org,
        devel@...verdev.osuosl.org, daniel.baluta@....com,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 3/3] staging:iio:ad2s1210: Add write_raw to handle
 frequency

On Mon, Mar 12, 2018 at 03:21:52PM -0300, Rodrigo Siqueira wrote:
> The write interface of AD2S1210 utilizes IIO_DEVICE_ATTR, which violate
> the official IIO ABI. This patch, add the write_raw function responsible
> for handling the fclkin and fexcit channel; also it removes the use of
> IIO_DEVICE_ATTR for fclkin and fexcit.
> 
> Signed-off-by: Rodrigo Siqueira <rodrigosiqueiramelo@...il.com>
> ---
>  drivers/staging/iio/resolver/ad2s1210.c | 117 +++++++++++++++-----------------
>  1 file changed, 55 insertions(+), 62 deletions(-)
> 
> diff --git a/drivers/staging/iio/resolver/ad2s1210.c b/drivers/staging/iio/resolver/ad2s1210.c
> index 27a42ed10fcd..ea6ade4e563c 100644
> --- a/drivers/staging/iio/resolver/ad2s1210.c
> +++ b/drivers/staging/iio/resolver/ad2s1210.c
> @@ -60,6 +60,8 @@
>  
>  #define AD2S1210_DEF_EXCIT	10000
>  
> +#define ERROR_MESSAGE "ad2s1210: %s out of range\n"
> +

Don't make this a define.  That's horrible.  It's a pointless layer of
abstraction.


>  enum ad2s1210_mode {
>  	MOD_POS = 0,
>  	MOD_VEL,
> @@ -210,64 +212,6 @@ static inline int ad2s1210_soft_reset(struct ad2s1210_state *st)
>  	return ad2s1210_config_write(st, 0x0);
>  }
>  
> -static ssize_t ad2s1210_store_fclkin(struct device *dev,
> -				     struct device_attribute *attr,
> -				     const char *buf,
> -				     size_t len)
> -{
> -	struct ad2s1210_state *st = iio_priv(dev_to_iio_dev(dev));
> -	unsigned int fclkin;
> -	int ret;
> -
> -	ret = kstrtouint(buf, 10, &fclkin);
> -	if (ret)
> -		return ret;
> -	if (fclkin < AD2S1210_MIN_CLKIN || fclkin > AD2S1210_MAX_CLKIN) {
> -		dev_err(dev, "ad2s1210: fclkin out of range\n");
> -		return -EINVAL;
> -	}
> -
> -	mutex_lock(&st->lock);
> -	st->fclkin = fclkin;
> -
> -	ret = ad2s1210_update_frequency_control_word(st);
> -	if (ret < 0)
> -		goto error_ret;
> -	ret = ad2s1210_soft_reset(st);
> -error_ret:
> -	mutex_unlock(&st->lock);
> -
> -	return ret < 0 ? ret : len;
> -}
> -
> -static ssize_t ad2s1210_store_fexcit(struct device *dev,
> -				     struct device_attribute *attr,
> -				     const char *buf, size_t len)
> -{
> -	struct ad2s1210_state *st = iio_priv(dev_to_iio_dev(dev));
> -	unsigned int fexcit;
> -	int ret;
> -
> -	ret = kstrtouint(buf, 10, &fexcit);
> -	if (ret < 0)
> -		return ret;
> -	if (fexcit < AD2S1210_MIN_EXCIT || fexcit > AD2S1210_MAX_EXCIT) {
> -		dev_err(dev,
> -			"ad2s1210: excitation frequency out of range\n");
> -		return -EINVAL;
> -	}
> -	mutex_lock(&st->lock);
> -	st->fexcit = fexcit;
> -	ret = ad2s1210_update_frequency_control_word(st);
> -	if (ret < 0)
> -		goto error_ret;
> -	ret = ad2s1210_soft_reset(st);
> -error_ret:
> -	mutex_unlock(&st->lock);
> -
> -	return ret < 0 ? ret : len;
> -}
> -
>  static ssize_t ad2s1210_show_control(struct device *dev,
>  				     struct device_attribute *attr,
>  				     char *buf)
> @@ -545,8 +489,58 @@ static int ad2s1210_read_raw(struct iio_dev *indio_dev,
>  	return ret;
>  }
>  
> -static IIO_DEVICE_ATTR(fclkin, 0644, NULL, ad2s1210_store_fclkin, 0);
> -static IIO_DEVICE_ATTR(fexcit, 0644, NULL, ad2s1210_store_fexcit, 0);
> +static int ad2s1210_write_raw(struct iio_dev *indio_dev,
> +			      struct iio_chan_spec const *chan, int val,
> +			      int val2, long mask)
> +{
> +	struct ad2s1210_state *st = iio_priv(indio_dev);
> +	unsigned int clk = val;
> +	int ret;
> +
> +	switch (mask) {
> +	case IIO_CHAN_INFO_FREQUENCY:
> +		switch (chan->channel) {

This is a switch statement with only one case.  Make it an if statement
and chop out some indenting.

	if (mask != IIO_CHAN_INFO_FREQUENCY)
		return -EINVAL;

> +		case FCLKIN:
> +			if (clk < AD2S1210_MIN_CLKIN ||
> +			    clk > AD2S1210_MAX_CLKIN) {
> +				dev_err(&indio_dev->dev, ERROR_MESSAGE,

Ah...  I see why you did the ERROR_MESSAGE define, to get around the 80
character limit.  Don't do that.  Just go over 80 characters if you need
to.


> +					"fclkin");
> +				ret = -EINVAL;
> +				goto error_ret;

Direct returns are better.  Less chance of bugs statistically.

regards,
dan carpenter