Date: Thu, 24 Feb 2011 17:48:07 +0100
From: Jean Delvare <khali@linux-fr.org>
To: Dirk Eibach <eibach@gdsys.de>
Cc: linux-kernel@vger.kernel.org, guenter.roeck@ericsson.com,
        lm-sensors@lm-sensors.org, rdunlap@xenotime.net,
        linux-doc@vger.kernel.org
Subject: Re: [PATCH v3] hwmon: Add support for Texas Instruments ADS1015
Message-ID: <20110224174807.41c7597f@endymion.delvare>
In-Reply-To: <1298024158-21076-1-git-send-email-eibach@gdsys.de>
References: <20110217134203.0556a7ab@endymion.delvare>
 <1298024158-21076-1-git-send-email-eibach@gdsys.de>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 17239
Lines: 566

Hi Dirk,

Sorry for the late reply.

On Fri, 18 Feb 2011 11:15:58 +0100, Dirk Eibach wrote:
> Signed-off-by: Dirk Eibach <eibach@gdsys.de>
> ---
> Changes since v1:
> - fixed/extended Documentation
> - removed unused register definitions
> - hardcoded PGA fullscale table size
> - made sure patch applies against v2.6.38-rc4
> - reordered functions to avoid forward declaration
> - results from i2c_smbus_read_word_data() are handled correctly
> - moved locking into ads1015_read_value()
> - removed unnecessray clearing of bit
> - proper error handling in ads1015_read_value()
> - use DIV_ROUND_CLOSEST for scaling result
> - removed detect()
> 
> Changes since v2:
> - removed *all* leftovers from detect()
> - fixed return with mutex held
> - made sysfs representation configurable
>   (hope this will be the reference implementation for generations to come ;)

Thanks for your continued work on this driver. The changes this time
are important enough to warrant a full review. Here we go:

>  Documentation/hwmon/ads1015 |   72 +++++++++++
>  drivers/hwmon/Kconfig       |   10 ++
>  drivers/hwmon/Makefile      |    1 +
>  drivers/hwmon/ads1015.c     |  295 +++++++++++++++++++++++++++++++++++++++++++
>  include/linux/i2c/ads1015.h |   30 +++++
>  5 files changed, 408 insertions(+), 0 deletions(-)
>  create mode 100644 Documentation/hwmon/ads1015
>  create mode 100644 drivers/hwmon/ads1015.c
>  create mode 100644 include/linux/i2c/ads1015.h
> 
> diff --git a/Documentation/hwmon/ads1015 b/Documentation/hwmon/ads1015
> new file mode 100644
> index 0000000..2494e99
> --- /dev/null
> +++ b/Documentation/hwmon/ads1015
> @@ -0,0 +1,72 @@
> +Kernel driver ads1015
> +=====================
> +
> +Supported chips:
> +  * Texas Instruments ADS1015
> +    Prefix: 'ads1015'
> +    Datasheet: Publicly available at the Texas Instruments website :
> +               http://focus.ti.com/lit/ds/symlink/ads1015.pdf
> +
> +Authors:
> +        Dirk Eibach, Guntermann & Drunck GmbH <eibach@gdsys.de>
> +
> +Description
> +-----------
> +
> +This driver implements support for the Texas Instruments ADS1015.
> +
> +This device is a 12-bit A-D converter with 4 inputs.
> +
> +The inputs can be used single ended or in certain differential combinations.
> +
> +The inputs are mapped to 8 sysfs input files in0_input - in7_input.
> +The mapping can be configured using platform data or devicetree.
> +
> +Data sources for configuration:
> +0: Voltage over AIN0 and AIN1.
> +1: Voltage over AIN0 and AIN3.
> +2: Voltage over AIN1 and AIN3.
> +3: Voltage over AIN2 and AIN3.
> +4: Voltage over AIN0 and GND.
> +5: Voltage over AIN1 and GND.
> +6: Voltage over AIN2 and GND.
> +7: Voltage over AIN3 and GND.
> +Any other value: disable
> +
> +By default in0_input is mapped to source 0, in1_input to source 1 and so on.

I see that you went for dynamic naming of sysfs files. I would have
used a different strategy which would make the code much more simple.
You can keep static sysfs file names, and instantiate them
conditionally. Maybe you were not aware of this, but it is perfectly
fine for an hwmon device to number its inputs non-linearly, and as a
matter of fact many hwmon driver do this.

For example, a setup where each input is used single-ended would result
in a hwmon device with attributes in4_input, in5_input, in6_input and
in7_input.

> +
> +Platform Data
> +-------------
> +
> +In linux/i2c/ads1015.h platform data is defined as:
> +
> +struct ads1015_platform_data {
> +	int exported_channels[8];
> +};
> +
> +exported_channels contains the data sources for the 8 sysfs input files.
> +
> +Example:
> +struct ads1015_platform_data data = {
> +	4, 2, -1, -1, -1, -1, -1, -1 };
> +
> +In this case only in0_input and in1_input would be created.
> +in0_input would give the voltage over AIN0 and GND.
> +in0_input would give the voltage over AIN1 and AIN3.

With my proposal, the platform data could be a single bitfield, where
each bit says enable or disable the corresponding sysfs attribute. For
example:

struct ads1015_platform_data {
	unsigned int channels;
};

Example:
struct ads1015_platform_data data = {
	.channels = (1 << 4) | (1 << 2)
}

The only drawback of my proposal is that you can't create the
attributes by group, you have to create them individually. Not a
problem in practice though. I suggest that you give it a try and see
what you prefer.

> +
> +Devicetree
> +----------
> +
> +The ads1015 node may have an "exported-channels" property with 8 integer
> +values. The 8 values are the data sources for the 8 sysfs input files.
> +
> +Example:
> +ads1015@49 {
> +	compatible = "ti,ads1015";
> +	reg = <0x49>;
> +	exported-channels = < 4 2 0xff 0xff 0xff 0xff 0xff 0xff >;
> +};
> +
> +In this case only in0_input and in1_input would be created.
> +in0_input would give the voltage over AIN0 and GND.
> +in0_input would give the voltage over AIN1 and AIN3.

You meant "in1_input" the second time.

Do you have an actual need for this? I think devicetree attributes have
to be discussed and documented appropriately? I admit I am not too
familiar with this.

> diff --git a/drivers/hwmon/Kconfig b/drivers/hwmon/Kconfig
> index 773e484..7e247f7 100644
> --- a/drivers/hwmon/Kconfig
> +++ b/drivers/hwmon/Kconfig
> @@ -871,6 +871,16 @@ config SENSORS_SMSC47B397
>  	  This driver can also be built as a module.  If so, the module
>  	  will be called smsc47b397.
>  
> +config SENSORS_ADS1015
> +	tristate "Texas Instruments ADS1015"
> +	depends on I2C
> +	help
> +	  If you say yes here you get support for Texas Instruments ADS1015
> +	  12-bit 4-input ADC device.
> +
> +	  This driver can also be built as a module.  If so, the module
> +	  will be called ads1015.
> +
>  config SENSORS_ADS7828
>  	tristate "Texas Instruments ADS7828"
>  	depends on I2C
> diff --git a/drivers/hwmon/Makefile b/drivers/hwmon/Makefile
> index dde02d9..aae4036 100644
> --- a/drivers/hwmon/Makefile
> +++ b/drivers/hwmon/Makefile
> @@ -29,6 +29,7 @@ obj-$(CONFIG_SENSORS_ADM1026)	+= adm1026.o
>  obj-$(CONFIG_SENSORS_ADM1029)	+= adm1029.o
>  obj-$(CONFIG_SENSORS_ADM1031)	+= adm1031.o
>  obj-$(CONFIG_SENSORS_ADM9240)	+= adm9240.o
> +obj-$(CONFIG_SENSORS_ADS1015)	+= ads1015.o
>  obj-$(CONFIG_SENSORS_ADS7828)	+= ads7828.o
>  obj-$(CONFIG_SENSORS_ADS7871)	+= ads7871.o
>  obj-$(CONFIG_SENSORS_ADT7411)	+= adt7411.o
> diff --git a/drivers/hwmon/ads1015.c b/drivers/hwmon/ads1015.c
> new file mode 100644
> index 0000000..cf7aff4
> --- /dev/null
> +++ b/drivers/hwmon/ads1015.c
> @@ -0,0 +1,295 @@
> +/*
> + * ads1015.c - lm_sensors driver for ads1015 12-bit 4-input ADC
> + * (C) Copyright 2010
> + * Dirk Eibach, Guntermann & Drunck GmbH <eibach@gdsys.de>
> + *
> + * Based on the ads7828 driver by Steve Hardy.
> + *
> + * Datasheet available at: http://focus.ti.com/lit/ds/symlink/ads1015.pdf
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License as published by
> + * the Free Software Foundation; either version 2 of the License, or
> + * (at your option) any later version.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program; if not, write to the Free Software
> + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
> + */
> +
> +#include <linux/module.h>
> +#include <linux/init.h>
> +#include <linux/slab.h>
> +#include <linux/jiffies.h>
> +#include <linux/i2c.h>
> +#include <linux/hwmon.h>
> +#include <linux/hwmon-sysfs.h>
> +#include <linux/err.h>
> +#include <linux/mutex.h>

You have to include <linux/of.h> for device tree support.

> +#include <linux/i2c/ads1015.h>
> +
> +/* ADS1015 registers */
> +enum {
> +	ADS1015_CONVERSION = 0,
> +	ADS1015_CONFIG = 1,
> +};
> +
> +/* PGA fullscale voltages in mV */
> +static const unsigned int fullscale_table[8] = {
> +	6144, 4096, 2048, 1024, 512, 256, 256, 256 };
> +
> +/* Default set of exported channels */
> +#define ADS1015_CONFIG_CHANNELS 8
> +static const int default_channels[ADS1015_CONFIG_CHANNELS] = {
> +	0, 1, 2, 3, 4, 5, 6, 7 };
> +
> +/* strings for sysfs */
> +static const char *input_names[8] = {
> +	"in0_input",
> +	"in1_input",
> +	"in2_input",
> +	"in3_input",
> +	"in4_input",
> +	"in5_input",
> +	"in6_input",
> +	"in7_input"
> +};
> +
> +struct ads1015_data {
> +	struct device *hwmon_dev;
> +	struct mutex update_lock; /* mutex protect updates */
> +	struct sensor_device_attribute attr[ADS1015_CONFIG_CHANNELS];
> +	struct attribute *attr_table[ADS1015_CONFIG_CHANNELS + 1];
> +	struct attribute_group attr_group;
> +};
> +
> +static s32 ads1015_read_reg(struct i2c_client *client, unsigned int reg)
> +{
> +	s32 data = i2c_smbus_read_word_data(client, reg);
> +
> +	return (data < 0) ? data : swab16(data);
> +}
> +
> +static s32 ads1015_write_reg(struct i2c_client *client, unsigned int reg,
> +			     u16 val)
> +{
> +	return i2c_smbus_write_word_data(client, reg, swab16(val));
> +}
> +
> +static int ads1015_read_value(struct i2c_client *client, unsigned int channel,
> +			      int *value)
> +{
> +	u16 config;
> +	s16 conversion;
> +	unsigned int pga;
> +	int fullscale;
> +	unsigned int k;
> +	struct ads1015_data *data = i2c_get_clientdata(client);
> +	int res;
> +
> +	mutex_lock(&data->update_lock);
> +
> +	/* get fullscale voltage */
> +	res = ads1015_read_reg(client, ADS1015_CONFIG);
> +	if (res < 0)
> +		goto err_unlock;
> +	config = res;
> +	pga = (config >> 9) & 0x0007;
> +	fullscale = fullscale_table[pga];

You only read the fullscale value. Now that we have platform data
attached to the device, don't you think it would make sense to let the
user set it, possibly even define different values for each "virtual
channel"? I can imagine that different scaling factors make sense for
single-ended vs. differential measurements, or even for different
single-ended inputs.

This is just a question, BTW, this feature can be added later if needed.

> +
> +	/* set channel and start single conversion */
> +	config &= ~(0x0007 << 12);
> +	config |= (1 << 15) | (1 << 8) | (channel & 0x0007) << 12;
> +
> +	/* wait until conversion finished */
> +	res = ads1015_write_reg(client, ADS1015_CONFIG, config);
> +	if (res < 0)
> +		goto err_unlock;
> +	for (k = 0; k < 5; ++k) {
> +		schedule_timeout(msecs_to_jiffies(1));
> +		res = ads1015_read_reg(client, ADS1015_CONFIG);
> +		if (res < 0)
> +			goto err_unlock;
> +		config = res;
> +		if (config & (1 << 15))
> +			break;
> +	}
> +	if (k == 5) {
> +		res = -EIO;
> +		goto err_unlock;
> +	}
> +
> +	res = ads1015_read_reg(client, ADS1015_CONVERSION);
> +	if (res < 0)
> +		goto err_unlock;
> +	conversion = res;
> +
> +	mutex_unlock(&data->update_lock);
> +
> +	*value = DIV_ROUND_CLOSEST(conversion * fullscale, 0x7ff0);
> +
> +	return 0;
> +
> +err_unlock:
> +	mutex_unlock(&data->update_lock);
> +	return res;
> +}
> +
> +/* sysfs callback function */
> +static ssize_t show_in(struct device *dev, struct device_attribute *da,
> +	char *buf)
> +{
> +	struct sensor_device_attribute *attr = to_sensor_dev_attr(da);
> +	struct i2c_client *client = to_i2c_client(dev);
> +	int in;
> +	int res;
> +
> +	res = ads1015_read_value(client, attr->index, &in);
> +
> +	return (res < 0) ? res : sprintf(buf, "%d\n", in);
> +}
> +
> +/*
> + * Driver interface
> + */
> +
> +static int ads1015_remove(struct i2c_client *client)
> +{
> +	struct ads1015_data *data = i2c_get_clientdata(client);
> +	hwmon_device_unregister(data->hwmon_dev);
> +	sysfs_remove_group(&client->dev.kobj, &data->attr_group);
> +	kfree(data);
> +	return 0;
> +}
> +
> +static void ads1015_get_exported_channels(struct i2c_client *client,
> +					  int *exported_channels)
> +{
> +	struct ads1015_platform_data *pdata = dev_get_platdata(&client->dev);
> +#ifdef CONFIG_OF
> +	struct device_node *np = client->dev.of_node;
> +	const int *of_channels;
> +	int of_channels_size;
> +#endif
> +
> +	/* prefer platform data */
> +	if (pdata) {
> +		memcpy(exported_channels, pdata->exported_channels,
> +		       sizeof(default_channels));
> +		return;
> +	}
> +
> +#ifdef CONFIG_OF
> +	/* fallback on OF */
> +	of_channels = of_get_property(np, "exported-channels",
> +				      &of_channels_size);
> +	if (of_channels && (of_channels_size == sizeof(default_channels))) {
> +		memcpy(exported_channels, of_channels,
> +		       sizeof(default_channels));
> +		return;
> +	}
> +#endif
> +
> +	/* fallback on default configuration */
> +	memcpy(exported_channels, default_channels, sizeof(default_channels));
> +}

Why don't you just return a pointer to the data? You only need the data
during probe and you make no changes to it, so I see no need to copy
the data.

> +
> +/* create sysfs attribute according to channel setup */
> +static struct attribute *ads1015_export_channel(struct i2c_client *client,
> +						unsigned int input, int channel)
> +{
> +	struct ads1015_data *data = i2c_get_clientdata(client);
> +	struct sensor_device_attribute attr =
> +		SENSOR_ATTR(input, S_IRUGO, show_in, NULL, channel);
> +
> +	attr_name(attr.dev_attr) = input_names[input];
> +
> +	memcpy(&data->attr[input], &attr, sizeof(attr));
> +
> +	return &data->attr[input].dev_attr.attr;
> +}
> +
> +static int ads1015_probe(struct i2c_client *client,
> +			 const struct i2c_device_id *id)
> +{
> +	struct ads1015_data *data;
> +	int err;
> +	int exported_channels[ADS1015_CONFIG_CHANNELS];
> +	unsigned int k;
> +	unsigned int act_attr = 0;

"act"?

> +
> +	data = kzalloc(sizeof(struct ads1015_data), GFP_KERNEL);
> +	if (!data) {
> +		err = -ENOMEM;
> +		goto exit;
> +	}
> +
> +	i2c_set_clientdata(client, data);
> +	mutex_init(&data->update_lock);
> +
> +	/* Register sysfs hooks */
> +	data->attr_group.attrs = data->attr_table;
> +	ads1015_get_exported_channels(client, exported_channels);
> +	for (k = 0; k < ADS1015_CONFIG_CHANNELS; ++k) {
> +		int channel = exported_channels[k];
> +		if ((channel < 0) || (channel > 7))

You don't need all these parentheses.

> +			continue;

Is there any benefit in continuing here, as opposed to breaking?
Breaking would let you use k below to index data->attr_table, instead
of tracking act_attr separately.

> +		data->attr_table[act_attr++] =
> +			ads1015_export_channel(client, k, channel);
> +	}
> +	err = sysfs_create_group(&client->dev.kobj, &data->attr_group);
> +	if (err)
> +		goto exit_free;
> +
> +	data->hwmon_dev = hwmon_device_register(&client->dev);
> +	if (IS_ERR(data->hwmon_dev)) {
> +		err = PTR_ERR(data->hwmon_dev);
> +		goto exit_remove;
> +	}
> +
> +	return 0;
> +
> +exit_remove:
> +	sysfs_remove_group(&client->dev.kobj, &data->attr_group);
> +exit_free:
> +	kfree(data);
> +exit:
> +	return err;
> +}
> +
> +static const struct i2c_device_id ads1015_id[] = {
> +	{ "ads1015", 0 },
> +	{ }
> +};
> +MODULE_DEVICE_TABLE(i2c, ads1015_id);
> +
> +static struct i2c_driver ads1015_driver = {
> +	.driver = {
> +		.name = "ads1015",
> +	},
> +	.probe = ads1015_probe,
> +	.remove = ads1015_remove,
> +	.id_table = ads1015_id,
> +};
> +
> +static int __init sensors_ads1015_init(void)
> +{
> +	return i2c_add_driver(&ads1015_driver);
> +}
> +
> +static void __exit sensors_ads1015_exit(void)
> +{
> +	i2c_del_driver(&ads1015_driver);
> +}
> +
> +MODULE_AUTHOR("Dirk Eibach <eibach@gdsys.de>");
> +MODULE_DESCRIPTION("ADS1015 driver");
> +MODULE_LICENSE("GPL");
> +
> +module_init(sensors_ads1015_init);
> +module_exit(sensors_ads1015_exit);
> diff --git a/include/linux/i2c/ads1015.h b/include/linux/i2c/ads1015.h
> new file mode 100644
> index 0000000..152bf5f
> --- /dev/null
> +++ b/include/linux/i2c/ads1015.h
> @@ -0,0 +1,30 @@
> +/*
> + * Platform Data for ADS1015 12-bit 4-input ADC
> + * (C) Copyright 2010
> + * Dirk Eibach, Guntermann & Drunck GmbH <eibach@gdsys.de>
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License as published by
> + * the Free Software Foundation; either version 2 of the License, or
> + * (at your option) any later version.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program; if not, write to the Free Software
> + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
> + */
> +
> +#ifndef LINUX_ADS1015_H
> +#define LINUX_ADS1015_H
> +
> +#include <linux/types.h>

You don't use anything from that header file.

> +
> +struct ads1015_platform_data {
> +	int exported_channels[8];
> +};
> +
> +#endif /* LINUX_ADS1015_H */


-- 
Jean Delvare
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/