On 08/14/13 18:08, Angelo Compagnucci wrote: > This revision solves various problems of the previous one. > Most important is locking of the chip configuration during updates > and the use of devm_* API. > For each channel there are now two sysfs attributes, so you can read > the raw value or the scaled one. > > To Peter: Chip versions differs in address assignment (MCP3422 has a fixed address) > and channel number (MCP3423 has 2 channels, MCP3424 has 4) but the chip manages > internally the lack of channels (it remaps non exixtent channels to the existent ones) > so I opted to keep the driver clean and implemented all the four channels. > > > > Signed-off-by: Angelo Compagnucci <angelo.compagnucci@xxxxxxxxx> pga is not in the abi. Unless I'm mistaken that's just a scale setting? I note you already have scale elements of the info_mask so you will need to combine the two (already done for the read?) and provide the relevant write callback. Various other bits and piece inline. Often they are not saying what you are doing is wrong, merely that there are accepted ways of implementing some things and using them saves me having to think and that's always good ;) As I stated in the other branch of this thread, please just have 3 different iio_chan_spec arrays to cover the 3 chip variants. Thanks, Jonathan > --- > drivers/iio/adc/Kconfig | 10 ++ > drivers/iio/adc/Makefile | 1 + > drivers/iio/adc/mcp3422.c | 420 +++++++++++++++++++++++++++++++++++++++++++++ > 3 files changed, 431 insertions(+) > create mode 100644 drivers/iio/adc/mcp3422.c > > diff --git a/drivers/iio/adc/Kconfig b/drivers/iio/adc/Kconfig > index f725b45..e9879e6 100644 > --- a/drivers/iio/adc/Kconfig > +++ b/drivers/iio/adc/Kconfig > @@ -145,6 +145,16 @@ config MCP320X > This driver can also be built as a module. If so, the module will be > called mcp320x. > > +config MCP3422 > + tristate "Microchip Technology MCP3422/3/4 driver" > + depends on I2C > + help > + Say yes here to build support for Microchip Technology's MCP3422, > + MCP3423 or MCP3424 analog to digital converters. > + > + This driver can also be built as a module. If so, the module will be > + called mcp3422. > + > config NAU7802 > tristate "Nuvoton NAU7802 ADC driver" > depends on I2C > diff --git a/drivers/iio/adc/Makefile b/drivers/iio/adc/Makefile > index 2a4324e..9772e3b 100644 > --- a/drivers/iio/adc/Makefile > +++ b/drivers/iio/adc/Makefile > @@ -16,6 +16,7 @@ obj-$(CONFIG_EXYNOS_ADC) += exynos_adc.o > obj-$(CONFIG_LP8788_ADC) += lp8788_adc.o > obj-$(CONFIG_MAX1363) += max1363.o > obj-$(CONFIG_MCP320X) += mcp320x.o > +obj-$(CONFIG_MCP3422) += mcp3422.o > obj-$(CONFIG_NAU7802) += nau7802.o > obj-$(CONFIG_TI_ADC081C) += ti-adc081c.o > obj-$(CONFIG_TI_AM335X_ADC) += ti_am335x_adc.o > diff --git a/drivers/iio/adc/mcp3422.c b/drivers/iio/adc/mcp3422.c > new file mode 100644 > index 0000000..c6058da > --- /dev/null > +++ b/drivers/iio/adc/mcp3422.c > @@ -0,0 +1,420 @@ > +/* > + * mcp3422.c - driver for the Microchip mcp3422/3/4 chip family > + * > + * Copyright (C) 2013, Angelo Compagnucci > + * Author: Angelo Compagnucci <angelo.compagnucci@xxxxxxxxx> > + * > + * Datasheet: http://ww1.microchip.com/downloads/en/devicedoc/22088b.pdf > + * > + * This driver exports the value of analog input voltage to sysfs, the > + * voltage unit is nV. > + * > + * 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. > + */ > + > +#include <linux/err.h> > +#include <linux/i2c.h> > +#include <linux/module.h> > +#include <linux/delay.h> > +#include <linux/sysfs.h> > + > +#include <linux/iio/iio.h> > +#include <linux/iio/sysfs.h> > + > +/* Masks */ > +#define MCP3422_CHANNEL_MASK 0x60 > +#define MCP3422_PGA_MASK 0x03 > +#define MCP3422_SRATE_MASK 0x0C > +#define MCP3422_SRATE_240 0x0 > +#define MCP3422_SRATE_60 0x1 > +#define MCP3422_SRATE_15 0x2 > +#define MCP3422_SRATE_3 0x3 > +#define MCP3422_PGA_1 0 > +#define MCP3422_PGA_2 1 > +#define MCP3422_PGA_4 2 > +#define MCP3422_PGA_8 3 > +#define MCP3422_CONT_SAMPLING 0x10 > + > +#define MCP3422_CHANNEL(config) ((((config) & MCP3422_CHANNEL_MASK) >> 5) + 1) > +#define MCP3422_PGA(config) ((config) & MCP3422_PGA_MASK) > +#define MCP3422_SAMPLE_RATE(config) (((config) & MCP3422_SRATE_MASK) >> 2) > + > +#define MCP3422_CHANNEL_VALUE(value) (((value - 1) << 5) & MCP3422_CHANNEL_MASK) > +#define MCP3422_PGA_VALUE(value) ((value) & MCP3422_PGA_MASK) > +#define MCP3422_SAMPLE_RATE_VALUE(value) ((value << 2) & MCP3422_SRATE_MASK) > + > +#define VREF 2048 > + I'd define this just before use. Makes for easier reviewing as I we don't have to jump around in the code. Only true for structure filling macros like this though. Leave allthe other defines here at the top. > +#define MCP3422_CHAN(index) \ > + { \ > + .type = IIO_VOLTAGE, \ > + .indexed = 1, \ > + .channel = index, \ > + .info_mask_separate = BIT(IIO_CHAN_INFO_RAW) \ > + | BIT(IIO_CHAN_INFO_SCALE), \ > + } > + > +/* LSB is in nV to eliminate floating point */ > +static const u32 rates_to_lsb[] = {1000000, 250000, 62500, 15625}; > + > +/* > + * Client data (each client gets its own) > + */ > +struct mcp3422 { > + struct i2c_client *i2c; > + u8 config; > + struct mutex lock; > +}; > + > +static int mcp3422_update_config(struct mcp3422 *adc, u8 newconfig) > +{ > + u8 buf; > + int ret; > + > + mutex_lock(&adc->lock); > + Why the copy of newconfig? I may be half asleep but can't see the disadvantage in just using it directly? > + buf = newconfig; > + ret = i2c_master_send(adc->i2c, &buf, 1); > + if (ret > 0) > + adc->config = buf; > + > + mutex_unlock(&adc->lock); > + > + return ret; > +} > + > +static int mcp3422_read(struct mcp3422 *adc, int *value, u8 *config) > +{ > + int ret = 0; > + u8 sample_rate = MCP3422_SAMPLE_RATE(adc->config); > + u8 buf[4] = {'\0', '\0', '\0', '\0'}; Why not just set it to 0 rather than explicity the NULL character? > + > + *value = 0; > + > + if (sample_rate == MCP3422_SRATE_3) { > + ret = i2c_master_recv(adc->i2c, buf, 4); > + *value += (s8)buf[0]; At this point *value is always equal to 0? Is there any reason to set that then add this value rather than just setting *value = (s8)buf[0]; ? > + *value <<= 16; > + *value |= buf[1] << 8; > + *value |= buf[2]; Do the above as u32 temp; temp = (buf[0] << 16) | buf[1] << 8 | buf[0]; *value = sign_extend32(temp, 24); Not necessarily better or cleaner than your version, just the 'standard' way which makes it easier for maintainers to deal with it. > + *config = buf[3]; > + if (*value == 0xFFFFFFFF) value is signed, so use it. if (*value == -1) > + *value = 0; > + } else { > + ret = i2c_master_recv(adc->i2c, buf, 3); Here a standard endian conversion would be preferred followed by sign extension. > + *value += (s8)buf[0]; > + *value <<= 8; > + *value |= buf[1]; > + *config = buf[2]; > + } > + > + return ret; > +} > + > +static int mcp3422_read_channel(struct mcp3422 *adc, > + struct iio_chan_spec const *channel, int *value) > +{ > + int ret, mtime = 0; > + u8 config = 0; > + u8 req_channel = channel->channel; > + > + if (req_channel != MCP3422_CHANNEL(adc->config)) { > + config = adc->config; > + config &= ~MCP3422_CHANNEL_MASK; > + config |= MCP3422_CHANNEL_VALUE(req_channel); > + mcp3422_update_config(adc, config); > + switch (MCP3422_SAMPLE_RATE(config)) { > + case MCP3422_SRATE_240: > + mtime = 1000 / 240; > + break; > + case MCP3422_SRATE_60: > + mtime = 1000 / 60; > + break; > + case MCP3422_SRATE_15: > + mtime = 1000 / 15; > + break; > + case MCP3422_SRATE_3: > + mtime = 1000 / 3; > + break; > + } > + msleep(mtime); > + } > + > + ret = mcp3422_read(adc, value, &config); > + return ret; > +} > + > +static int mcp3422_read_raw(struct iio_dev *iio, > + struct iio_chan_spec const *channel, int *val1, > + int *val2, long mask) > +{ > + struct mcp3422 *adc = iio_priv(iio); > + int err, temp = 0; > + > + u8 sample_rate = MCP3422_SAMPLE_RATE(adc->config); > + u8 pga = MCP3422_PGA(adc->config); > + > + err = mcp3422_read_channel(adc, channel, &temp); > + if (err < 0) > + return err; > + > + switch (mask) { > + case IIO_CHAN_INFO_RAW: > + *val1 = temp; > + > + return IIO_VAL_INT; > + > + case IIO_CHAN_INFO_SCALE: > + > + temp = (temp * rates_to_lsb[sample_rate]) > + / (pga ? 1 << pga : 1); Ah, so you have alrady applied the pga to the read version of scale. Just have the equivalent for write. I can see it will be a little fiddly given the dependence on the sample_rate, but such is life ;) Much less fiddly than forcing all userspace users to cope with a new bit of ABI. Actually this is just complicated enough that you might want to just precompute the values and have them as const data in the driver. You have 4 pga values and 4 sampling frequencies so hardly enormous. Up to you which way you do it. > + > + *val1 = temp / 1000000000; > + *val2 = temp % 1000000000; > + > + return IIO_VAL_INT_PLUS_NANO; > + > + default: > + break; > + } > + > + return -EINVAL; > +} > + > +static ssize_t mcp3422_set_sample_rate(struct device *dev, > + struct device_attribute *attr, > + const char *buf, size_t count) > +{ > + struct i2c_client *client = to_i2c_client(dev); > + struct iio_dev *iio = i2c_get_clientdata(client); > + struct mcp3422 *adc = iio_priv(iio); > + u8 value = 0; > + u8 config = adc->config; > + > + if (kstrtou8(buf, 10, &value) < 0) > + return -EINVAL; > + > + switch (value) { > + case 240: > + value = MCP3422_SRATE_240; > + break; > + case 60: > + value = MCP3422_SRATE_60; > + break; > + case 15: > + value = MCP3422_SRATE_15; > + break; > + case 3: > + value = MCP3422_SRATE_3; > + break; > + default: > + return -EINVAL; > + } > + > + config &= ~MCP3422_SRATE_MASK; > + config |= MCP3422_SAMPLE_RATE_VALUE(value); > + > + mcp3422_update_config(adc, config); > + > + return count; > +} > + > +static ssize_t mcp3422_get_sample_rate(struct device *dev, > + struct device_attribute *attr, > + char *buf) > +{ > + struct i2c_client *client = to_i2c_client(dev); > + struct iio_dev *iio = i2c_get_clientdata(client); > + struct mcp3422 *adc = iio_priv(iio); > + u8 value = 0; > + > + switch (MCP3422_SAMPLE_RATE(adc->config)) { > + case 0: > + value = 240; > + break; > + case 1: > + value = 60; > + break; > + case 2: > + value = 15; > + break; > + case 3: > + value = 3; > + break; > + } > + > + return sprintf(buf, "%d\n", value); > +} > + > +static ssize_t mcp3422_set_pga(struct device *dev, > + struct device_attribute *attr, > + const char *buf, size_t count) > +{ > + struct i2c_client *client = to_i2c_client(dev); > + struct iio_dev *iio = i2c_get_clientdata(client); > + struct mcp3422 *adc = iio_priv(iio); > + u8 value = 0; > + u8 config = adc->config; > + > + if (kstrtou8(buf, 10, &value) < 0) > + return -EINVAL; > + Could take advantage of the simple mapping to these constants and have value = 1 << (value - 1); > + switch (value) { > + case 1: > + value = MCP3422_PGA_1; > + break; > + case 2: > + value = MCP3422_PGA_2; > + break; > + case 4: > + value = MCP3422_PGA_4; > + break; > + case 8: > + value = MCP3422_PGA_8; > + break; > + default: > + return -EINVAL; > + } > + > + config &= ~MCP3422_PGA_MASK; > + config |= MCP3422_PGA_VALUE(value); > + > + mcp3422_update_config(adc, config); > + > + return count; > +} > + > +static ssize_t mcp3422_get_pga(struct device *dev, > + struct device_attribute *attr, > + char *buf) > +{ > + struct i2c_client *client = to_i2c_client(dev); > + struct iio_dev *iio = i2c_get_clientdata(client); > + struct mcp3422 *adc = iio_priv(iio); > + > + return sprintf(buf, "%d\n", 1 << MCP3422_PGA(adc->config)); > +} > + > +static IIO_DEV_ATTR_SAMP_FREQ(S_IRUGO | S_IWUSR, > + mcp3422_get_sample_rate, > + mcp3422_set_sample_rate); > +static IIO_DEVICE_ATTR(pga, > + S_IRUGO | S_IWUSR, > + mcp3422_get_pga, > + mcp3422_set_pga, 0); > +static IIO_CONST_ATTR(sampling_frequency_available, "240 60 15 3"); > +static IIO_CONST_ATTR(pga_available, "1 2 4 8"); > + > +static struct attribute *mcp3422_attributes[] = { > + &iio_dev_attr_sampling_frequency.dev_attr.attr, > + &iio_const_attr_sampling_frequency_available.dev_attr.attr, > + &iio_dev_attr_pga.dev_attr.attr, > + &iio_const_attr_pga_available.dev_attr.attr, Basic principal of driver writing. If it isn't in existing ABI documents have a very strong arguement why. In this case the PGA is just applying a predivide to the input voltage and hence acts as a 'scale' to be applied to the output. > + NULL, > +}; > + > +static const struct attribute_group mcp3422_attribute_group = { > + .attrs = mcp3422_attributes, > +}; > + > +static const struct iio_chan_spec mcp3422_channels[] = { > + MCP3422_CHAN(1), > + MCP3422_CHAN(2), > + MCP3422_CHAN(3), > + MCP3422_CHAN(4), > +}; > + > +static const struct iio_info mcp3422_info = { > + .read_raw = mcp3422_read_raw, > + .attrs = &mcp3422_attribute_group, > + .driver_module = THIS_MODULE, > +}; > + > +static int mcp3422_probe(struct i2c_client *client, > + const struct i2c_device_id *id) > +{ > + struct iio_dev *indio_dev; > + struct mcp3422 *adc; > + int err; > + u8 config; > + > + if (!i2c_check_functionality(client->adapter, I2C_FUNC_I2C)) > + return -ENODEV; > + > + indio_dev = devm_iio_device_alloc(&client->dev, sizeof(*adc)); > + if (!indio_dev) > + return -ENOMEM; > + > + adc = iio_priv(indio_dev); > + adc->i2c = client; > + > + mutex_init(&adc->lock); > + > + indio_dev->dev.parent = &client->dev; > + indio_dev->name = dev_name(&client->dev); > + indio_dev->modes = INDIO_DIRECT_MODE; > + indio_dev->info = &mcp3422_info; > + > + indio_dev->channels = mcp3422_channels; > + indio_dev->num_channels = ARRAY_SIZE(mcp3422_channels); > + > + err = iio_device_register(indio_dev); > + if (err < 0) > + return err; > + > + i2c_set_clientdata(client, indio_dev); > + > + /* meaningful default configuration */ > + config = (MCP3422_CONT_SAMPLING > + | MCP3422_CHANNEL_VALUE(1) > + | MCP3422_PGA_VALUE(MCP3422_PGA_1) > + | MCP3422_SAMPLE_RATE_VALUE(MCP3422_SRATE_240)); > + mcp3422_update_config(adc, config); > + > + return 0; > +} > + > +static int mcp3422_remove(struct i2c_client *client) > +{ > + struct iio_dev *iio = i2c_get_clientdata(client); > + > + iio_device_unregister(iio); > + iio_device_free(iio); > + > + return 0; > +} > + > +static const struct i2c_device_id mcp3422_id[] = { > + { "mcp3422", 2 }, > + { "mcp3423", 3 }, > + { "mcp3424", 4 }, > + { } > +}; > +MODULE_DEVICE_TABLE(i2c, mcp3422_id); > + > +#ifdef CONFIG_OF > +static const struct of_device_id mcp3422_of_match[] = { > + { .compatible = "mcp3422" }, > + { } > +}; > +MODULE_DEVICE_TABLE(of, mcp3422_of_match); > +#endif > + > +static struct i2c_driver mcp3422_driver = { > + .driver = { > + .name = "mcp3422", > + .owner = THIS_MODULE, > + .of_match_table = of_match_ptr(mcp3422_of_match), > + }, > + .probe = mcp3422_probe, > + .remove = mcp3422_remove, > + .id_table = mcp3422_id, > +}; > +module_i2c_driver(mcp3422_driver); > + > +MODULE_AUTHOR("Angelo Compagnucci <angelo.compagnucci@xxxxxxxxx>"); > +MODULE_DESCRIPTION("Microchip mcp3422/3/4 driver"); > +MODULE_LICENSE("GPL v2"); > -- To unsubscribe from this list: send the line "unsubscribe linux-iio" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
