Re: [PATCH] [RFC]power: battery: Generic battery driver using IIO

[anish kumar]

Thanks for your time.
On Fri, 2012-09-07 at 08:52 +0100, Jonathan Cameron wrote:
> On 02/09/12 16:39, anish kumar wrote:
> > From: anish kumar <anish198519851985@xxxxxxxxx>
> >
> > This patch is to use IIO to write a generic batttery driver.
> battery
> > There are some inherent assumptions here:
> > 1.User is having both main battery and backup battery.
> Seems rather like that could be easily enough relaxed or configured via
> platform data?
Yes
> > 2.Both batteries use same channel to get the information.
> Do you mean same actual channel (physical pin) or same ADC
> (with appropriate mux etc)?
As lars pointed out it would be better if we have multiple channels
per battery.The channel what I am referring here is the name of 
different channels which will be exported by adc driver to get
voltage, current and power.
> >
> Mostly fine. There are a few corners I didn't understand as I don't
> have time to dive into the battery framework in the near future. Will 
> leave it up to those more familiar with that side of things!
> 
> My only real issue is that it would be cleaner to do the individual
> channels by name rather than a get all as you 'know' here how many
> channels are expected.
agreed.
> 
> Also don't release the IIO channels until you are actually finished
> using them as holding them should prevent the ADC module from going away
> underneath you (which is nasty :)
agreed.
> 
> Sorry it took so long to get back to you on this.
That is the biggest advantage of working in open-source.
We can take our own time.
> 
> So what happened in 1985?
Was hurrying to get an email id :)
> 
> Jonathan
> > Signed-off-by: anish kumar <anish198519851985@xxxxxxxxx>
> > ---
> >   drivers/power/Kconfig                 |    8 +
> >   drivers/power/Makefile                |    1 +
> >   drivers/power/generic-battery.c       |  374 +++++++++++++++++++++++++++++++++
> >   include/linux/power/generic-battery.h |   26 +++
> >   4 files changed, 409 insertions(+), 0 deletions(-)
> >   create mode 100644 drivers/power/generic-battery.c
> >   create mode 100644 include/linux/power/generic-battery.h
> >
> > diff --git a/drivers/power/Kconfig b/drivers/power/Kconfig
> > index fcc1bb0..546e86b 100644
> > --- a/drivers/power/Kconfig
> > +++ b/drivers/power/Kconfig
> > @@ -309,6 +309,14 @@ config AB8500_BATTERY_THERM_ON_BATCTRL
> >   	help
> >   	  Say Y to enable battery temperature measurements using
> >   	  thermistor connected on BATCTRL ADC.
> > +
> > +config GENERIC_BATTERY
> > +	tristate "Generic battery support using IIO"
> > +	depends on IIO
> > +	help
> > +	  Say Y here to enable support for the generic battery driver
> > +	  which uses IIO framework to read adc for it's main and backup
> ADC
> > +	  battery.
> >   endif # POWER_SUPPLY
> >
> >   source "drivers/power/avs/Kconfig"
> > diff --git a/drivers/power/Makefile b/drivers/power/Makefile
> > index ee58afb..8284f9c 100644
> > --- a/drivers/power/Makefile
> > +++ b/drivers/power/Makefile
> > @@ -45,3 +45,4 @@ obj-$(CONFIG_CHARGER_MAX8997)	+= max8997_charger.o
> >   obj-$(CONFIG_CHARGER_MAX8998)	+= max8998_charger.o
> >   obj-$(CONFIG_POWER_AVS)		+= avs/
> >   obj-$(CONFIG_CHARGER_SMB347)	+= smb347-charger.o
> > +obj-$(CONFIG_GENERIC_BATTERY)	+= generic-battery.o
> > diff --git a/drivers/power/generic-battery.c b/drivers/power/generic-battery.c
> > new file mode 100644
> > index 0000000..33170b7
> > --- /dev/null
> > +++ b/drivers/power/generic-battery.c
> > @@ -0,0 +1,374 @@
> > +/*
> > + * Generice battery driver code using IIO
> Generic
> > + * Copyright (C) 2012, Anish Kumar <anish198519851985@xxxxxxxxx>
> > + * based on jz4740-battery.c
> > + * based on s3c_adc_battery.c
> > + *
> > + * This file is subject to the terms and conditions of the GNU General Public
> > + * License.  See the file COPYING in the main directory of this archive for
> > + * more details.
> > + *
> bonus blank line to remove
> > + */
> > +
> > +#include <linux/interrupt.h>
> > +#include <linux/platform_device.h>
> > +#include <linux/power_supply.h>
> > +#include <linux/gpio.h>
> > +#include <linux/err.h>
> > +#include <linux/timer.h>
> > +#include <linux/jiffies.h>
> > +#include <linux/errno.h>
> > +#include <linux/init.h>
> > +#include <linux/module.h>
> > +#include <linux/slab.h>
> > +#include <linux/iio/consumer.h>
> > +#include <linux/iio/types.h>
> > +
> > +#include <linux/power/generic-battery.h>
> > +
> > +#define BAT_POLL_INTERVAL		10000 /* ms */
> > +#define JITTER_DELAY			500 /* ms */
> Elements to make configurable?
Yes probably make it platform data?
> > +
> > +enum BAT {
> > +	MAIN_BAT = 0,
> > +	BACKUP_BAT,
> > +	BAT_MAX,
> > +};
> > +
> Document this please. It's not immediately obvious what
> channel_index is.
This will be removed as we will be using only one device.
> 
> Why not just have a direct pointer to the correct channel?
> > +struct generic_adc_bat {
> > +	struct power_supply		psy;
> > +	struct iio_channel		*channels;
> > +	int				channel_index;
> > +};
> > +
> > +struct generic_bat {
> Spacing is a bit random in here
> > +	struct generic_adc_bat bat[BAT_MAX];
> > +	struct generic_platform_data	pdata;
> > +	int				volt_value;
> > +	int				cur_value;
> > +	unsigned int			timestamp;
> > +	int				level;
> > +	int				status;
> > +	int				cable_plugged:1;
> > +};
> > +
> > +static struct generic_bat generic_bat = {
> > +	.bat[MAIN_BAT] = {
> > +		.psy.name = "main-bat",
> > +	},
> > +	.bat[BACKUP_BAT] = {
> > +		.psy.name = "backup-bat",
> > +	},
> > +};
> > +
> > +static struct delayed_work bat_work;
> > +
> > +static void generic_adc_bat_ext_power_changed(struct power_supply *psy)
> > +{
> > +	schedule_delayed_work(&bat_work,
> > +			msecs_to_jiffies(JITTER_DELAY));
> > +}
> > +
> > +static enum power_supply_property generic_adc_main_bat_props[] = {
> > +	POWER_SUPPLY_PROP_STATUS,
> > +	POWER_SUPPLY_PROP_CHARGE_FULL_DESIGN,
> > +	POWER_SUPPLY_PROP_CHARGE_EMPTY_DESIGN,
> > +	POWER_SUPPLY_PROP_CHARGE_NOW,
> > +	POWER_SUPPLY_PROP_VOLTAGE_NOW,
> > +	POWER_SUPPLY_PROP_CURRENT_NOW,
> > +	POWER_SUPPLY_PROP_TECHNOLOGY,
> > +	POWER_SUPPLY_PROP_VOLTAGE_MIN_DESIGN,
> > +	POWER_SUPPLY_PROP_VOLTAGE_MAX_DESIGN,
> > +	POWER_SUPPLY_PROP_MODEL_NAME,
> > +};
> > +
> > +static int charge_finished(struct generic_bat *bat)
> > +{
> > +	return bat->pdata.gpio_inverted ?
> > +		!gpio_get_value(bat->pdata.gpio_charge_finished) :
> > +		gpio_get_value(bat->pdata.gpio_charge_finished);
> > +}
> > +
> > +static int generic_adc_bat_get_property(struct power_supply *psy,
> > +		enum power_supply_property psp,
> > +		union power_supply_propval *val)
> > +{
> > +	struct generic_adc_bat *adc_bat;
> > +	struct generic_bat *bat;
> > +	int ret, scaleint, scalepart, iio_val;
> > +	long result = 0;
> > +
> > +	if (!strcmp(psy->name, "main-battery")) {
> > +		adc_bat = container_of(psy,
> > +					struct generic_adc_bat, psy);
> This first statement is I think shared so move it out of the if / else
This whole logic will change as we will be using only one device and
if has user has multiple batteries then he can register multiple
devices.
> 
> > +		bat = container_of(adc_bat,
> > +				struct generic_bat, bat[MAIN_BAT]);
> > +	} else if (!strcmp(psy->name, "backup-battery")) {
> > +		adc_bat = container_of(psy, struct generic_adc_bat, psy);
> > +		bat = container_of(adc_bat,
> > +				struct generic_bat, bat[BACKUP_BAT]);
> > +	} else {
> > +		/* Ideally this should never happen */
> > +		WARN_ON(1);
> > +		return -EINVAL;
> > +	}
> > +
> > +	if (!bat) {
> 	> +		dev_err(psy->dev, "no battery infos ?!\n");
> > +		return -EINVAL;
> > +	
> > +	ret = iio_read_channel_raw(&adc_bat->channels[adc_bat->channel_index],
> > +			&iio_val);
> > +	ret = iio_read_channel_scale(&adc_bat->channels[adc_bat->channel_index],
> > +			&scaleint, &scalepart);
> > +	if (ret < 0)
> > +		return ret;
> > +
> A quick comment here on what the battery framework is expecting vs IIO
> supplying would be good.  It may be that this particular output is one 
> that we should lift over into the core rather than end up with multiple
> drivers doing the same thing.
I think for this we need to get all the channels supported by a
particular adc device and once we get the information regarding the
channel(current/voltage/power) supported we can find out "What
information is exported about this channel which may include scale or
raw adc value".This can be used to call the corresponding API's.
> 
> > +	switch (ret) {
> > +	case IIO_VAL_INT:
> > +		result = iio_val * scaleint;
> > +		break;
> > +	case IIO_VAL_INT_PLUS_MICRO:
> > +		result = (s64)iio_val * (s64)scaleint +
> > +			div_s64((s64)iio_val * (s64)scalepart, 1000000LL);
> > +		break;
> > +	case IIO_VAL_INT_PLUS_NANO:
> > +		result = (s64)iio_val * (s64)scaleint +
> > +			div_s64((s64)iio_val * (s64)scalepart, 1000000000LL);
> > +		break;
> > +	}
> > +
> > +	switch (psp) {
> > +	case POWER_SUPPLY_PROP_STATUS:
> > +		if (bat->pdata.gpio_charge_finished < 0)
> > +			val->intval = bat->level == 100000 ?
> > +				POWER_SUPPLY_STATUS_FULL : bat->status;
> > +		else
> > +			val->intval = bat->status;
> > +		return 0;
> > +	case POWER_SUPPLY_PROP_CHARGE_EMPTY_DESIGN:
> > +		val->intval = 0;
> > +		return 0;
> > +	case POWER_SUPPLY_PROP_CHARGE_NOW:
> > +		val->intval = bat->pdata.cal_charge(result);
> > +		return 0;
> > +	case POWER_SUPPLY_PROP_VOLTAGE_NOW:
> > +		val->intval = result;
> > +		return 0;
> > +	case POWER_SUPPLY_PROP_CURRENT_NOW:
> > +		val->intval = result;
> > +		return 0;
> why return from these and break from the later ones (to return much the 
> same I think...)?
my mistake.
> > +	case POWER_SUPPLY_PROP_TECHNOLOGY:
> > +		val->intval = bat->pdata.battery_info.technology;
> > +		break;
> > +	case POWER_SUPPLY_PROP_VOLTAGE_MIN_DESIGN:
> > +		val->intval = bat->pdata.battery_info.voltage_min_design;
> > +		break;
> > +	case POWER_SUPPLY_PROP_VOLTAGE_MAX_DESIGN:
> > +		val->intval = bat->pdata.battery_info.voltage_max_design;
> > +		break;
> > +	case POWER_SUPPLY_PROP_CHARGE_FULL_DESIGN:
> > +		val->intval = bat->pdata.battery_info.charge_full_design;
> > +		break;
> > +	case POWER_SUPPLY_PROP_MODEL_NAME:
> > +		val->strval = bat->pdata.battery_info.name;
> > +		break;
> > +	default:
> > +		return -EINVAL;
> > +	}
> > +	return ret;
> > +}
> > +
> > +static void generic_adc_bat_work(struct work_struct *work)
> > +{
> > +	struct generic_bat *bat = &generic_bat;
> > +	int is_charged;
> > +	int is_plugged;
> > +	static int was_plugged;
> That's nasty. Prevents even the theoretical posibility of multiple 
> copies of the driver. That should be in the device specific data.
yes should be part of device specific data.
> > +
> > +	/* backup battery doesn't have current monitoring capability anyway */
> Always or in this particular configuration?
I was told by Anton Vorontsov.Anyway now this driver will be per device
so shouldn't matter.
> 
> > +	is_plugged = power_supply_am_i_supplied(&bat->bat[MAIN_BAT].psy);
> > +	bat->cable_plugged = is_plugged;
> > +	if (is_plugged != was_plugged) {
> > +		was_plugged = is_plugged;
> > +		if (is_plugged)
> > +			bat->status = POWER_SUPPLY_STATUS_CHARGING;
> > +		else
> > +			bat->status = POWER_SUPPLY_STATUS_DISCHARGING;
> > +	} else {
> > +		if ((bat->pdata.gpio_charge_finished >= 0) && is_plugged) {
> > +			is_charged = charge_finished(&generic_bat);
> > +			if (is_charged)
> > +				bat->status = POWER_SUPPLY_STATUS_FULL;
> > +			else
> > +				bat->status = POWER_SUPPLY_STATUS_CHARGING;
> > +		}
> > +	}
> > +
> > +	power_supply_changed(&bat->bat[MAIN_BAT].psy);
> > +}
> > +
> > +static irqreturn_t generic_adc_bat_charged(int irq, void *dev_id)
> > +{
> > +	schedule_delayed_work(&bat_work,
> > +			msecs_to_jiffies(JITTER_DELAY));
> > +	return IRQ_HANDLED;
> > +}
> > +
> > +static int __devinit generic_adc_bat_probe(struct platform_device *pdev)
> > +{
> > +	struct generic_platform_data *pdata = pdev->dev.platform_data;
> > +	struct iio_channel *channels;
> > +	int num_channels = 0;
> > +	int ret, i, j, k = 0;
> > +	enum iio_chan_type type;
> > +
> > +	for (i = 0; i < ARRAY_SIZE(generic_bat.bat); i++) {
> > +		generic_bat.bat[i].psy.type = POWER_SUPPLY_TYPE_BATTERY;
> > +		generic_bat.bat[i].psy.properties =
> > +					generic_adc_main_bat_props;
> > +		generic_bat.bat[i].psy.num_properties =
> > +					ARRAY_SIZE(generic_adc_main_bat_props);
> > +		generic_bat.bat[i].psy.get_property =
> > +					generic_adc_bat_get_property;
> > +		generic_bat.bat[i].psy.external_power_changed =
> > +					generic_adc_bat_ext_power_changed;
> Could you do this with a static const array?  Looks like there is 
> nothing dynamic in here and that would make for cleaner code to read. 
> Given other elements of psy are clearly dynamic (dev pointer) you will 
> have to memcpy the static version in which is tedious. Still easier
> to read though in my view.
> > +	}
right.
> > +
> > +	generic_bat.volt_value = -1;
> > +	generic_bat.cur_value = -1;
> > +	generic_bat.cable_plugged = 0;
> > +	generic_bat.status = POWER_SUPPLY_STATUS_DISCHARGING;
> > +
> > +	memcpy(&generic_bat.pdata, pdata, sizeof(struct generic_platform_data));
> > +
> Is there anything dynamic in the pdata?  If not it would be cleaner just 
> to use a pointer to it rather than make a copy (I can't immediately spot 
> anything but them I'm not familiar with the battery
> side of things.
We can just assign it to generic_bat so that it can be used in the
relevant functions.Otherwise we need to use dev_set_drvdata and
dev_get_drvdata. 
> > +	for (i = 0; i < ARRAY_SIZE(generic_bat.bat); i++) {
> > +		ret = power_supply_register(&pdev->dev,
> > +				&generic_bat.bat[i].psy);
> > +		if (ret)
> > +			goto err_reg_main;
> > +	}
> > +
> > +	channels = iio_channel_get_all(dev_name(&pdev->dev));
> I would give these explicit names and get the two individual channels by 
> name. I think it will give you cleaner code.
Yes, now it will be based on pdata->voltage_channel,
pdata->current_channel and so on.We will use this to get the channel.
> > +	if (IS_ERR(channels)) {
> > +		ret = PTR_ERR(channels);
> > +		goto err_reg_backup;
> > +	}
> > +
> > +	while (channels[num_channels].indio_dev)
> > +		num_channels++;
> > +
> > +	for (i = 0; i < ARRAY_SIZE(generic_bat.bat); i++)
> > +		generic_bat.bat[i].channels = kzalloc(sizeof(struct iio_channel)
> > +							* BAT_MAX, GFP_KERNEL);
> > +
> > +	/* assuming main battery and backup battery is using the same channel */
> > +	for (i = 0; i < num_channels; i++) {
> > +		ret = iio_get_channel_type(&channels[i], &type);
> Using named channels you should 'know' you have the correct type.  You 
> can of course check if you don't trust your platform data though!
right.
> > +		if (ret < 0)
> > +			goto err_gpio;
> > +
> > +		if (type == IIO_VOLTAGE || type == IIO_CURRENT) {
> > +			for (j = 0; j < BAT_MAX; j++) {
> > +				generic_bat.bat[j].channel_index = k;
> > +				generic_bat.bat[j].channels[k] = channels[i];
> > +			}
> > +			k++;
> > +		}
> > +		continue;
> > +	}
> > +
> > +	INIT_DELAYED_WORK(&bat_work, generic_adc_bat_work);
> > +
> > +	if (pdata->gpio_charge_finished >= 0) {
> > +		ret = gpio_request(pdata->gpio_charge_finished, "charged");
> > +		if (ret)
> > +			goto err_gpio;
> > +
> > +		ret = request_irq(gpio_to_irq(pdata->gpio_charge_finished),
> > +				generic_adc_bat_charged,
> > +				IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING,
> > +				"battery charged", NULL);
> > +		if (ret)
> > +			goto err_gpio;
> > +	}
> > +
> > +	dev_info(&pdev->dev, "successfully loaded\n");
> > +
> > +	/* Schedule timer to check current status */
> > +	schedule_delayed_work(&bat_work,
> > +			msecs_to_jiffies(JITTER_DELAY));
> > +
> > +	iio_channel_release_all(channels);
> Not if you want to prevent the adc driver from being removed from under 
> you.  They should only be released in the driver removal.
right.
> > +
> > +	return 0;
> > +
> > +err_gpio:
> > +	iio_channel_release_all(channels);
> > +err_reg_backup:
> > +	for (i = 0; i < ARRAY_SIZE(generic_bat.bat); i++)
> > +		power_supply_unregister(&generic_bat.bat[i].psy);
> > +err_reg_main:
> > +	return ret;
> > +}
> > +
> > +static int generic_adc_bat_remove(struct platform_device *pdev)
> > +{
> > +	int i;
> > +	struct generic_platform_data *pdata = pdev->dev.platform_data;
> > +
> > +	for (i = 0; i < ARRAY_SIZE(generic_bat.bat); i++)
> > +		power_supply_unregister(&generic_bat.bat[i].psy);
> 
> You should release the IIO channels here when they are no longer needed.
right.
> > +
> > +	if (pdata->gpio_charge_finished >= 0) {
> I forget, is a gpio index of 0 definitely invalid?
> > +		free_irq(gpio_to_irq(pdata->gpio_charge_finished), NULL);
> > +		gpio_free(pdata->gpio_charge_finished);
> > +	}
> > +
> > +	cancel_delayed_work(&bat_work);
> > +
> > +	return 0;
> > +}
> > +
> > +#ifdef CONFIG_PM
> > +static int generic_adc_bat_suspend(struct platform_device *pdev,
> > +		pm_message_t state)
> > +{
> > +	struct generic_platform_data *pdata = pdev->dev.platform_data;
> > +	struct generic_bat *bat = container_of(pdata,
> > +					struct generic_bat, pdata);
> > +
> > +	cancel_delayed_work_sync(&bat_work);
> > +	bat->status = POWER_SUPPLY_STATUS_UNKNOWN;
> > +
> > +	return 0;
> > +}
> > +
> > +static int generic_adc_bat_resume(struct platform_device *pdev)
> > +{
> > +	/* Schedule timer to check current status */
> > +	schedule_delayed_work(&bat_work,
> > +			msecs_to_jiffies(JITTER_DELAY));
> > +
> > +	return 0;
> > +}
> > +#else
> > +#define generic_adc_bat_suspend NULL
> > +#define generic_adc_bat_resume NULL
> > +#endif
> > +
> > +static struct platform_driver generic_adc_bat_driver = {
> > +	.driver		= {
> > +		.name	= "generic-adc-battery",
> > +	},
> > +	.probe		= generic_adc_bat_probe,
> > +	.remove		= generic_adc_bat_remove,
> > +	.suspend	= generic_adc_bat_suspend,
> > +	.resume		= generic_adc_bat_resume,
> > +};
> > +
> > +module_platform_driver(generic_adc_bat_driver);
> > +
> > +MODULE_AUTHOR("anish kumar <anish198519851985@xxxxxxxxx>");
> > +MODULE_DESCRIPTION("generic battery driver using IIO");
> > +MODULE_LICENSE("GPL");
> > diff --git a/include/linux/power/generic-battery.h b/include/linux/power/generic-battery.h
> > new file mode 100644
> > index 0000000..7a43aa0
> > --- /dev/null
> > +++ b/include/linux/power/generic-battery.h
> > @@ -0,0 +1,26 @@
> > +/*
> > + * This program is free software; you can redistribute it and/or modify
> > + * it under the terms of the GNU General Public License version 2 as
> > + * published by the Free Software Foundation.
> > + */
> > +
> > +#ifndef GENERIC_BATTERY_H
> > +#define GENERIC_BATTERY_H
> > +
> > +#include <linux/power_supply.h>
> > +
> > +/**
> > + * struct generic_platform_data - platform data for generic battery
> A little inconsistent on capitalization.  (might as well do it now or
> one of those people who delight in trivial patches will 'tidy' it
> up for you :)
I don't mind as it is a good start for people to get familiar with the 
linux patch process.
However I will address this valuable comments.
> > + * @battery_info: Information about the battery
> > + * @cal_charge: platform callback to calculate charge
> > + * @gpio_charge_finished: GPIO number used for interrupts
> > + * @gpio_inverted: Is the gpio inverted?
> > + */
> > +struct generic_platform_data {
> > +	struct power_supply_info battery_info;
> > +	int	(*cal_charge)(long);
> > +	int	gpio_charge_finished;
> > +	int	gpio_inverted;
> > +};
> > +
> > +#endif /* GENERIC_BATTERY_H */
> >
> 


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