hwmon: (pmbus/ltc2978) Add regulator ops
v4 cover hwmon: (pmbus/ltc2978) Add regulator ops
Hi!
The LTC2978 driver supports a wide range of power regulator chips, but it has limited functionality for using it in a dynamic regulator framework. Since standard functions for setting and getting voltage are missing as pmbus core operations this patchset adds it.
These patches have been tested on an ARTPEC-8 developer board with a group of LTC2977 power regulators.
Kind regards
Mårten Lindahl
Changes in v2:
- Add pmbus core _pmbus_write_byte_data to check for driver specific callback
- Change pmbus_update_byte_data to use _pmbus_read/write_byte_data
- Change pmbus_regulator_is_enabled to use _pmbus_read_byte_data
- Export pmbus core functions enable/disable/is_enabled
Changes in v3:
- Split patch "hwmon: (pmbus/ltc2978) Use driver specific ops if they exist" into two patches: (1) pmbus core, (2) ltc2978.
- Move ltc2978_regulator_get/set_voltage functions to pmbus core.
Changes in v4:
- Split (and rename) patch "hwmon: (pmbus) Use driver specific ops if they exist" into two patches where the first handle _pmbus_write_byte_data, and the second handle _pmbus_read_byte_data.
- Use voltage conversion functions in pmbus_regulator_get/set_voltage.
Mårten Lindahl (4):
hwmon: (pmbus) Introduce and use write_byte_data callback
hwmon: (pmbus) Use _pmbus_read_byte_data with callback
hwmon: (pmbus/ltc2978) Add chip specific write_byte_data
hwmon: (pmbus) Add get_voltage/set_voltage ops
drivers/hwmon/pmbus/ltc2978.c | 12 +++ drivers/hwmon/pmbus/pmbus.h | 2 + drivers/hwmon/pmbus/pmbus_core.c | 133 +++++++++++++++++++++++++------ 3 files changed, 121 insertions(+), 26 deletions(-)
Signature · 1 line
2.30.2
PATCH v4 1/4 hwmon: (pmbus) Introduce and use write_byte_data callback
Some of the pmbus core functions uses pmbus_write_byte_data, which does not support driver callbacks for chip specific write operations. This could potentially influence some specific regulator chips that for example need a time delay before each data access.
Lets add support for driver callback with _pmbus_write_byte_data.
--- drivers/hwmon/pmbus/pmbus.h | 2 ++ drivers/hwmon/pmbus/pmbus_core.c | 24 +++++++++++++++++++++--- 2 files changed, 23 insertions(+), 3 deletions(-)
drivers/hwmon/pmbus/pmbus.h +2 -0
@@ -438,6 +438,8 @@ struct pmbus_driver_info {int (*read_byte_data)(struct i2c_client *client,int page,int reg);int (*read_word_data)(struct i2c_client *client,int page,int phase,int reg); +int (*write_byte_data)(struct i2c_client *client,int page,int reg, +u8 byte);int (*write_word_data)(struct i2c_client *client,int page,int reg,u16 word);int (*write_byte)(struct i2c_client *client,int page,u8 value);
drivers/hwmon/pmbus/pmbus_core.c +21 -3
@@ -276,6 +276,24 @@ static int _pmbus_write_word_data(struct i2c_client *client, int page, int reg,return pmbus_write_word_data (client, page, reg, word); } +/* +* _pmbus_write_byte_data() is similar to pmbus_write_byte_data(), but checks if +* a device specific mapping function exists and calls it if necessary. +*/ +static int _pmbus_write_byte_data (struct i2c_client *client,int page,int reg,u8 value) +{ +struct pmbus_data *data =i2c_get_clientdata (client); +const struct pmbus_driver_info *info = data->info; +int status; + +if (info->write_byte_data) { + status = info->write_byte_data (client, page, reg, value); +if (status != -ENODATA ) +return status; + } +return pmbus_write_byte_data (client, page, reg, value); +} +int pmbus_update_fan (struct i2c_client *client,int page,int id,u8 config,u8 mask,u16 command) {
@@ -290,7 +308,7 @@ int pmbus_update_fan(struct i2c_client *client, int page, int id, to = (from & ~mask) | (config & mask);if (to != from) { - rv =pmbus_write_byte_data (client, page, + rv =_pmbus_write_byte_data (client, page, pmbus_fan_config_registers[id], to);if (rv <0 )return rv;
@@ -397,7 +415,7 @@ int pmbus_update_byte_data(struct i2c_client *client, int page, u8 reg, tmp = (rv & ~mask) | (value & mask);if (tmp != rv) - rv =pmbus_write_byte_data (client, page, reg, tmp); + rv =_pmbus_write_byte_data (client, page, reg, tmp);return rv; }
@@ -912,7 +930,7 @@ static int pmbus_get_boolean(struct i2c_client *client, struct pmbus_boolean *b, regval = status & mask;if (regval) { - ret =pmbus_write_byte_data (client, page, reg, regval); + ret =_pmbus_write_byte_data (client, page, reg, regval);if (ret) goto unlock; }
Signature · 1 line
2.30.2
Re: [PATCH v4 1/4] hwmon: (pmbus) Introduce and use write_byte_data callback
Quoted 8 lines · Mårten Lindahl
Some of the pmbus core functions uses
pmbus_write_byte_data, which does not support driver callbacks for chip specific write operations. This could potentially influence some specific regulator chips that for example need a time delay before each data access.Lets add support for driver callback with
_pmbus_write_byte_data.Signed-off-by Mårten Lindahl <[email protected]>
Applied.
Thanks,
Guenter
Quoted 74 lines
--- drivers/hwmon/pmbus/pmbus.h | 2 ++ drivers/hwmon/pmbus/pmbus_core.c | 24 +++++++++++++++++++++--- 2 files changed, 23 insertions(+), 3 deletions(-)drivers/hwmon/pmbus/pmbus.h +2
-0@@ -438,6 +438,8 @@ struct pmbus_driver_info {int (*read_byte_data)(struct i2c_client *client,int page,int reg);int (*read_word_data)(struct i2c_client *client,int page,int phase,int reg); +int (*write_byte_data)(struct i2c_client *client,int page,int reg, +u8 byte);int (*write_word_data)(struct i2c_client *client,int page,int reg,u16 word);int (*write_byte)(struct i2c_client *client,int page,u8 value);drivers/hwmon/pmbus/pmbus_core.c +21
-3@@ -276,6 +276,24 @@ static int _pmbus_write_word_data(struct i2c_client *client, int page, int reg,return pmbus_write_word_data (client, page, reg, word); } +/* +* _pmbus_write_byte_data() is similar to pmbus_write_byte_data(), but checks if +* a device specific mapping function exists and calls it if necessary. +*/ +static int _pmbus_write_byte_data (struct i2c_client *client,int page,int reg,u8 value) +{ +struct pmbus_data *data =i2c_get_clientdata (client); +const struct pmbus_driver_info *info = data->info; +int status; + +if (info->write_byte_data) { + status = info->write_byte_data (client, page, reg, value); +if (status != -ENODATA ) +return status; + } +return pmbus_write_byte_data (client, page, reg, value); +} +int pmbus_update_fan (struct i2c_client *client,int page,int id,u8 config,u8 mask,u16 command) {@@ -290,7 +308,7 @@ int pmbus_update_fan(struct i2c_client *client, int page, int id, to = (from & ~mask) | (config & mask);if (to != from) { - rv =pmbus_write_byte_data (client, page, + rv =_pmbus_write_byte_data (client, page, pmbus_fan_config_registers[id], to);if (rv <0 )return rv;@@ -397,7 +415,7 @@ int pmbus_update_byte_data(struct i2c_client *client, int page, u8 reg, tmp = (rv & ~mask) | (value & mask);if (tmp != rv) - rv =pmbus_write_byte_data (client, page, reg, tmp); + rv =_pmbus_write_byte_data (client, page, reg, tmp);return rv; }@@ -912,7 +930,7 @@ static int pmbus_get_boolean(struct i2c_client *client, struct pmbus_boolean *b, regval = status & mask;if (regval) { - ret =pmbus_write_byte_data (client, page, reg, regval); + ret =_pmbus_write_byte_data (client, page, reg, regval);if (ret) goto unlock; }
PATCH v4 2/4 hwmon: (pmbus) Use _pmbus_read_byte_data with callback
In reply to Mårten LindahlSome of the pmbus core functions uses pmbus_read_byte_data, which does not support driver callbacks for chip specific write operations. This could potentially influence some specific regulator chips that for example need a time delay before each data access.
Lets use _pmbus_read_byte_data with callback check.
--- drivers/hwmon/pmbus/pmbus_core.c | 46 ++++++++++++++++---------------- 1 file changed, 23 insertions(+), 23 deletions(-)
drivers/hwmon/pmbus/pmbus_core.c +23 -23
@@ -294,6 +294,24 @@ static int _pmbus_write_byte_data(struct i2c_client *client, int page, int reg,return pmbus_write_byte_data (client, page, reg, value); } +/* +* _pmbus_read_byte_data() is similar to pmbus_read_byte_data(), but checks if +* a device specific mapping function exists and calls it if necessary. +*/ +static int _pmbus_read_byte_data (struct i2c_client *client,int page,int reg) +{ +struct pmbus_data *data =i2c_get_clientdata (client); +const struct pmbus_driver_info *info = data->info; +int status; + +if (info->read_byte_data) { + status = info->read_byte_data (client, page, reg); +if (status != -ENODATA ) +return status; + } +return pmbus_read_byte_data (client, page, reg); +} +int pmbus_update_fan (struct i2c_client *client,int page,int id,u8 config,u8 mask,u16 command) {
@@ -301,7 +319,7 @@ int pmbus_update_fan(struct i2c_client *client, int page, int id,int rv;u8 to; - from =pmbus_read_byte_data (client, page, + from =_pmbus_read_byte_data (client, page, pmbus_fan_config_registers[id]);if (from <0 )return from;
@@ -408,7 +426,7 @@ int pmbus_update_byte_data(struct i2c_client *client, int page, u8 reg,unsigned int tmp;int rv; - rv =pmbus_read_byte_data (client, page, reg); + rv =_pmbus_read_byte_data (client, page, reg);if (rv <0 )return rv;
@@ -421,24 +439,6 @@ int pmbus_update_byte_data(struct i2c_client *client, int page, u8 reg, }EXPORT_SYMBOL_NS_GPL (pmbus_update_byte_data,PMBUS ); -/* -* _pmbus_read_byte_data() is similar to pmbus_read_byte_data(), but checks if -* a device specific mapping function exists and calls it if necessary. -*/ -static int _pmbus_read_byte_data (struct i2c_client *client,int page,int reg) -{ -struct pmbus_data *data =i2c_get_clientdata (client); -const struct pmbus_driver_info *info = data->info; -int status; - -if (info->read_byte_data) { - status = info->read_byte_data (client, page, reg); -if (status != -ENODATA ) -return status; - } -return pmbus_read_byte_data (client, page, reg); -} -static struct pmbus_sensor *pmbus_find_sensor (struct pmbus_data *data,int page,int reg) {
@@ -473,7 +473,7 @@ static int pmbus_get_fan_rate(struct i2c_client *client, int page, int id,return s->data; } - config =pmbus_read_byte_data (client, page, + config =_pmbus_read_byte_data (client, page, pmbus_fan_config_registers[id]);if (config <0 )return config;
@@ -2414,7 +2414,7 @@ static int pmbus_regulator_is_enabled(struct regulator_dev *rdev)int ret;mutex_lock (&data->update_lock); - ret =pmbus_read_byte_data (client, page,PMBUS_OPERATION ); + ret =_pmbus_read_byte_data (client, page,PMBUS_OPERATION );mutex_unlock (&data->update_lock);if (ret <0 )
@@ -2513,7 +2513,7 @@ static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsignedif (!(func & cat->func))continue ; - status =pmbus_read_byte_data (client, page, cat->reg); + status =_pmbus_read_byte_data (client, page, cat->reg);if (status <0 ) {mutex_unlock (&data->update_lock);return status;
Signature · 1 line
2.30.2
Re: [PATCH v4 2/4] hwmon: (pmbus) Use _pmbus_read_byte_data with callback
Quoted 8 lines · Mårten Lindahl
Some of the pmbus core functions uses
pmbus_read_byte_data, which does not support driver callbacks for chip specific write operations. This could potentially influence some specific regulator chips that for example need a time delay before each data access.Lets use
_pmbus_read_byte_datawith callback check.Signed-off-by Mårten Lindahl <[email protected]>
Applied.
Thanks,
Guenter
Quoted 103 lines
--- drivers/hwmon/pmbus/pmbus_core.c | 46 ++++++++++++++++---------------- 1 file changed, 23 insertions(+), 23 deletions(-)drivers/hwmon/pmbus/pmbus_core.c +23
-23@@ -294,6 +294,24 @@ static int _pmbus_write_byte_data(struct i2c_client *client, int page, int reg,return pmbus_write_byte_data (client, page, reg, value); } +/* +* _pmbus_read_byte_data() is similar to pmbus_read_byte_data(), but checks if +* a device specific mapping function exists and calls it if necessary. +*/ +static int _pmbus_read_byte_data (struct i2c_client *client,int page,int reg) +{ +struct pmbus_data *data =i2c_get_clientdata (client); +const struct pmbus_driver_info *info = data->info; +int status; + +if (info->read_byte_data) { + status = info->read_byte_data (client, page, reg); +if (status != -ENODATA ) +return status; + } +return pmbus_read_byte_data (client, page, reg); +} +int pmbus_update_fan (struct i2c_client *client,int page,int id,u8 config,u8 mask,u16 command) {@@ -301,7 +319,7 @@ int pmbus_update_fan(struct i2c_client *client, int page, int id,int rv;u8 to; - from =pmbus_read_byte_data (client, page, + from =_pmbus_read_byte_data (client, page, pmbus_fan_config_registers[id]);if (from <0 )return from;@@ -408,7 +426,7 @@ int pmbus_update_byte_data(struct i2c_client *client, int page, u8 reg,unsigned int tmp;int rv; - rv =pmbus_read_byte_data (client, page, reg); + rv =_pmbus_read_byte_data (client, page, reg);if (rv <0 )return rv;@@ -421,24 +439,6 @@ int pmbus_update_byte_data(struct i2c_client *client, int page, u8 reg, }EXPORT_SYMBOL_NS_GPL (pmbus_update_byte_data,PMBUS ); -/* -* _pmbus_read_byte_data() is similar to pmbus_read_byte_data(), but checks if -* a device specific mapping function exists and calls it if necessary. -*/ -static int _pmbus_read_byte_data (struct i2c_client *client,int page,int reg) -{ -struct pmbus_data *data =i2c_get_clientdata (client); -const struct pmbus_driver_info *info = data->info; -int status; - -if (info->read_byte_data) { - status = info->read_byte_data (client, page, reg); -if (status != -ENODATA ) -return status; - } -return pmbus_read_byte_data (client, page, reg); -} -static struct pmbus_sensor *pmbus_find_sensor (struct pmbus_data *data,int page,int reg) {@@ -473,7 +473,7 @@ static int pmbus_get_fan_rate(struct i2c_client *client, int page, int id,return s->data; } - config =pmbus_read_byte_data (client, page, + config =_pmbus_read_byte_data (client, page, pmbus_fan_config_registers[id]);if (config <0 )return config;@@ -2414,7 +2414,7 @@ static int pmbus_regulator_is_enabled(struct regulator_dev *rdev)int ret;mutex_lock (&data->update_lock); - ret =pmbus_read_byte_data (client, page,PMBUS_OPERATION ); + ret =_pmbus_read_byte_data (client, page,PMBUS_OPERATION );mutex_unlock (&data->update_lock);if (ret <0 )@@ -2513,7 +2513,7 @@ static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsignedif (!(func & cat->func))continue ; - status =pmbus_read_byte_data (client, page, cat->reg); + status =_pmbus_read_byte_data (client, page, cat->reg);if (status <0 ) {mutex_unlock (&data->update_lock);return status;
PATCH v4 3/4 hwmon: (pmbus/ltc2978) Add chip specific write_byte_data
In reply to Mårten LindahlSeveral of the manuals for devices supported by this driver describes the need for a minimum wait time before the chip is ready to receive next command.
This wait time is already implemented in the driver as a ltc_wait_ready function with a driver defined wait time of 100 ms, and is considered for specific devices before reading/writing data on the pmbus.
Since this driver uses the default pmbus_regulator_ops for the enable/ disable/is_enabled functions we should add a driver specific callback for write_byte_data to prevent bypassing the wait time recommendations for the following devices: ltc3880/ltc3882/ltc3883/ltc3884/ltc3886/ ltc3887/ltc3889/ltm4664/ltm4675/ltm4676/ltm4677/ltm4678/ltm4680/ltm4686/ ltm4700/ltc7880.
--- drivers/hwmon/pmbus/ltc2978.c | 12 ++++++++++++ 1 file changed, 12 insertions(+)
drivers/hwmon/pmbus/ltc2978.c +12 -0
@@ -196,6 +196,17 @@ static int ltc_read_byte_data(struct i2c_client *client, int page, int reg)return pmbus_read_byte_data (client, page, reg); } +static int ltc_write_byte_data (struct i2c_client *client,int page,int reg,u8 value) +{ +int ret; + + ret =ltc_wait_ready (client); +if (ret <0 ) +return ret; + +return pmbus_write_byte_data (client, page, reg, value); +} +static int ltc_write_byte (struct i2c_client *client,int page,u8 byte) {int ret;
@@ -681,6 +692,7 @@ static int ltc2978_probe(struct i2c_client *client) info = &data->info; info->write_word_data = ltc2978_write_word_data; info->write_byte = ltc_write_byte; + info->write_byte_data = ltc_write_byte_data; info->read_word_data = ltc_read_word_data; info->read_byte_data = ltc_read_byte_data;
Signature · 1 line
2.30.2
Re: [PATCH v4 3/4] hwmon: (pmbus/ltc2978) Add chip specific write_byte_data
Quoted 8 lines · Mårten Lindahl
Several of the manuals for devices supported by this driver describes the need for a minimum wait time before the chip is ready to receive next command.
This wait time is already implemented in the driver as a
ltc_wait_readyfunction with a driver defined wait time of 100 ms, and is considered for specific devices before reading/writing data on the pmbus.
Quoted 8 lines
Since this driver uses the default
pmbus_regulator_opsfor the enable/ disable/is_enabled functions we should add a driver specific callback forwrite_byte_datato prevent bypassing the wait time recommendations for the following devices: ltc3880/ltc3882/ltc3883/ltc3884/ltc3886/ ltc3887/ltc3889/ltm4664/ltm4675/ltm4676/ltm4677/ltm4678/ltm4680/ltm4686/ ltm4700/ltc7880.Signed-off-by Mårten Lindahl <[email protected]>
Applied.
Thanks,
Guenter
Quoted 34 lines
--- drivers/hwmon/pmbus/ltc2978.c | 12 ++++++++++++ 1 file changed, 12 insertions(+)drivers/hwmon/pmbus/ltc2978.c +12
-0@@ -196,6 +196,17 @@ static int ltc_read_byte_data(struct i2c_client *client, int page, int reg)return pmbus_read_byte_data (client, page, reg); } +static int ltc_write_byte_data (struct i2c_client *client,int page,int reg,u8 value) +{ +int ret; + + ret =ltc_wait_ready (client); +if (ret <0 ) +return ret; + +return pmbus_write_byte_data (client, page, reg, value); +} +static int ltc_write_byte (struct i2c_client *client,int page,u8 byte) {int ret;@@ -681,6 +692,7 @@ static int ltc2978_probe(struct i2c_client *client) info = &data->info; info->write_word_data = ltc2978_write_word_data; info->write_byte = ltc_write_byte; + info->write_byte_data = ltc_write_byte_data; info->read_word_data = ltc_read_word_data; info->read_byte_data = ltc_read_byte_data;
PATCH v4 4/4 hwmon: (pmbus) Add get_voltage/set_voltage ops
In reply to Mårten LindahlThe pmbus core does not have operations for getting or setting voltage. Add functions get/set voltage for the dynamic regulator framework.
--- drivers/hwmon/pmbus/pmbus_core.c | 63 ++++++++++++++++++++++++++++++++ 1 file changed, 63 insertions(+)
drivers/hwmon/pmbus/pmbus_core.c +63 -0
@@ -1531,6 +1531,11 @@ static const struct pmbus_sensor_attr voltage_attributes[] = { .gbit = PB_STATUS_VOUT_OV, .limit = vout_limit_attrs, .nlimit = ARRAY_SIZE(vout_limit_attrs), + }, { + .reg = PMBUS_VOUT_COMMAND, + .class = PSC_VOLTAGE_OUT, + .paged = true, + .func = PMBUS_HAVE_VOUT, } };
@@ -2563,11 +2568,69 @@ static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsignedreturn 0 ; } +static int pmbus_regulator_get_voltage (struct regulator_dev *rdev) +{ +struct device *dev =rdev_get_dev (rdev); +struct i2c_client *client =to_i2c_client (dev->parent); +struct pmbus_data *data =i2c_get_clientdata (client); +struct pmbus_sensor *sensor; +u8 page =rdev_get_id (rdev); +int ret; + + sensor =pmbus_find_sensor (data, page,PMBUS_READ_VOUT ); +if (IS_ERR (sensor)) +return -ENODATA ; + +mutex_lock (&data->update_lock); +pmbus_update_sensor_data (client, sensor); +if (sensor->data <0 ) + ret = sensor->data; +else + ret = (int )pmbus_reg2data (data, sensor) *1000 ;/* unit is uV */ +mutex_unlock (&data->update_lock); + +return ret; +} + +static int pmbus_regulator_set_voltage (struct regulator_dev *rdev,int min_uV, +int max_uV,unsigned int *selector) +{ +struct device *dev =rdev_get_dev (rdev); +struct i2c_client *client =to_i2c_client (dev->parent); +struct pmbus_data *data =i2c_get_clientdata (client); +struct pmbus_sensor *sensor; +u8 page =rdev_get_id (rdev); +s64 tmp =DIV_ROUND_CLOSEST_ULL (min_uV,1000 );/* convert to mV */ +u16 val; +int ret; + *selector =0 ; + + sensor =pmbus_find_sensor (data, page,PMBUS_VOUT_COMMAND ); +if (IS_ERR (sensor)) +return -ENODATA ; + + ret =_pmbus_read_word_data (client, page,0xff ,PMBUS_VOUT_MARGIN_LOW ); +if (ret <0 ) +return ret; + + val =pmbus_data2reg (data, sensor, tmp); + +/* Do not fall shorter than low margin */ +if (ret > val) + val = ret; + + ret =_pmbus_write_word_data (client, page,PMBUS_VOUT_COMMAND , val); + +return ret; +} +const struct regulator_ops pmbus_regulator_ops = { .enable = pmbus_regulator_enable, .disable = pmbus_regulator_disable, .is_enabled = pmbus_regulator_is_enabled, .get_error_flags = pmbus_regulator_get_error_flags, + .get_voltage = pmbus_regulator_get_voltage, + .set_voltage = pmbus_regulator_set_voltage, };EXPORT_SYMBOL_NS_GPL (pmbus_regulator_ops,PMBUS );
Signature · 1 line
2.30.2
Guenter Roeck
maintainerRe: [PATCH v4 4/4] hwmon: (pmbus) Add get_voltage/set_voltage ops
Quoted 13 lines · Mårten Lindahl
The pmbus core does not have operations for getting or setting voltage. Add functions get/set voltage for the dynamic regulator framework.
Signed-off-by Mårten Lindahl <[email protected]>--- drivers/hwmon/pmbus/pmbus_core.c | 63 ++++++++++++++++++++++++++++++++ 1 file changed, 63 insertions(+)drivers/hwmon/pmbus/pmbus_core.c +0
-0
drivers/hwmon/pmbus/pmbus_core.c static const struct pmbus_sensor_attr voltage_attributes[] = { · line 1539
Quoted 9 lines
drivers/hwmon/pmbus/pmbus_core.c +5
-0@@ -1531,6 +1531,11 @@ static const struct pmbus_sensor_attr voltage_attributes[] = { .gbit = PB_STATUS_VOUT_OV, .limit = vout_limit_attrs, .nlimit = ARRAY_SIZE(vout_limit_attrs), + }, { + .reg = PMBUS_VOUT_COMMAND, + .class = PSC_VOLTAGE_OUT, + .paged = true, + .func = PMBUS_HAVE_VOUT, }
Ok, you lost me here. This adds an inX_input attribute. Why ? This is completely unrelated to the intended scope of this patch. It also doesn't report a measured voltage, but a configuration value. If anything, it would have to be a separate patch, and you'd have to argue hard why it makes sense to report it as measured voltage.
Quoted 17 lines
drivers/hwmon/pmbus/pmbus_core.c +11
-0};@@ -2563,11 +2568,69 @@ static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsignedreturn 0 ; } +static int pmbus_regulator_get_voltage (struct regulator_dev *rdev) +{ +struct device *dev =rdev_get_dev (rdev); +struct i2c_client *client =to_i2c_client (dev->parent); +struct pmbus_data *data =i2c_get_clientdata (client); +struct pmbus_sensor *sensor; +u8 page =rdev_get_id (rdev); +int ret; + + sensor =pmbus_find_sensor (data, page,PMBUS_READ_VOUT ); +if (IS_ERR (sensor))
drivers/hwmon/pmbus/pmbus_core.c static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsigned · line 2591
Quoted 10 lines
drivers/hwmon/pmbus/pmbus_core.c +10
-0+return -ENODATA ; + +mutex_lock (&data->update_lock); +pmbus_update_sensor_data (client, sensor); +if (sensor->data <0 ) + ret = sensor->data; +else + ret = (int )pmbus_reg2data (data, sensor) *1000 ;/* unit is uV */ +mutex_unlock (&data->update_lock); +
Same question. Why ?
Quoted 14 lines
drivers/hwmon/pmbus/pmbus_core.c +14
-0+return ret; +} + +static int pmbus_regulator_set_voltage (struct regulator_dev *rdev,int min_uV, +int max_uV,unsigned int *selector) +{ +struct device *dev =rdev_get_dev (rdev); +struct i2c_client *client =to_i2c_client (dev->parent); +struct pmbus_data *data =i2c_get_clientdata (client); +struct pmbus_sensor *sensor; +u8 page =rdev_get_id (rdev); +s64 tmp =DIV_ROUND_CLOSEST_ULL (min_uV,1000 );/* convert to mV */ +u16 val; +int ret;
drivers/hwmon/pmbus/pmbus_core.c static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsigned · line 2615
Quoted 10 lines
drivers/hwmon/pmbus/pmbus_core.c +10
-0+ *selector =0 ; + + sensor =pmbus_find_sensor (data, page,PMBUS_VOUT_COMMAND ); +if (IS_ERR (sensor)) +return -ENODATA ; + + ret =_pmbus_read_word_data (client, page,0xff ,PMBUS_VOUT_MARGIN_LOW ); +if (ret <0 ) +return ret; +
That actually makes me wonder: What about VOUT_MARGIN_HIGH ? Also, there are optional MFR_VOUT_MIN and MFR_VOUT_MAX registers. Would it possibly make sense to determine the valid range once during probe and then compare against it ?
Thanks,
Guenter
Quoted 21 lines
drivers/hwmon/pmbus/pmbus_core.c +13
-0+ val =pmbus_data2reg (data, sensor, tmp); + +/* Do not fall shorter than low margin */ +if (ret > val) + val = ret; + + ret =_pmbus_write_word_data (client, page,PMBUS_VOUT_COMMAND , val); + +return ret; +} +const struct regulator_ops pmbus_regulator_ops = { .enable = pmbus_regulator_enable, .disable = pmbus_regulator_disable, .is_enabled = pmbus_regulator_is_enabled, .get_error_flags = pmbus_regulator_get_error_flags, + .get_voltage = pmbus_regulator_get_voltage, + .set_voltage = pmbus_regulator_set_voltage, };EXPORT_SYMBOL_NS_GPL (pmbus_regulator_ops,PMBUS );
Re: [PATCH v4 4/4] hwmon: (pmbus) Add get_voltage/set_voltage ops
Quoted 21 lines · Guenter Roeck
Quoted 20 lines · Mårten Lindahl
The pmbus core does not have operations for getting or setting voltage. Add functions get/set voltage for the dynamic regulator framework.
Signed-off-by Mårten Lindahl <[email protected]>--- drivers/hwmon/pmbus/pmbus_core.c | 63 ++++++++++++++++++++++++++++++++ 1 file changed, 63 insertions(+)drivers/hwmon/pmbus/pmbus_core.c +4
-0@@ -1531,6 +1531,11 @@ static const struct pmbus_sensor_attr voltage_attributes[] = { .gbit = PB_STATUS_VOUT_OV, .limit = vout_limit_attrs, .nlimit = ARRAY_SIZE(vout_limit_attrs), + }, { + .reg = PMBUS_VOUT_COMMAND, + .class = PSC_VOLTAGE_OUT, + .paged = true,
Quoted 8 lines
drivers/hwmon/pmbus/pmbus_core.c static const struct pmbus_sensor_attr voltage_attributes[] = { · line 1539
Quoted 2 lines
drivers/hwmon/pmbus/pmbus_core.c +1
-0+ .func = PMBUS_HAVE_VOUT, }Ok, you lost me here. This adds an
inX_inputattribute. Why ? This is completely unrelated to the intended scope of this patch. It also doesn't report a measured voltage, but a configuration value. If anything, it would have to be a separate patch, and you'd have to argue hard why it makes sense to report it as measured voltage.
I see. The reason for adding this is as simple as I now understand it is wrong. Please remember, my first version of the set/get_voltage functions where hardcoded with L16 input/output. Then in order to use the already existing convertion functions pmbus_data2reg and pmbus_reg2data I added this only for the need of a sensor object, as those functions are tailored for a sensor object.
So now I have to ask you for advice. Should I use the existing convertion functions, or do you suggest new variants of them? If reusing them, I guess I have two options:
1: Modify them to take class, page, and data outside of a sensor object as input. 2: Use them as they are, but create a local 'dummy' sensor object with class, page,
and data to use when calling the convertion functions.
I hope I made it more clear for you now how I was thinking. There is absolutely no intention of having sensor inX_input attributes for reading the setpoint values. This was just an unwanted sideeffect, and I will glady remove it again.
Quoted 22 lines
Quoted 22 lines
drivers/hwmon/pmbus/pmbus_core.c +15
-0};@@ -2563,11 +2568,69 @@ static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsignedreturn 0 ; } +static int pmbus_regulator_get_voltage (struct regulator_dev *rdev) +{ +struct device *dev =rdev_get_dev (rdev); +struct i2c_client *client =to_i2c_client (dev->parent); +struct pmbus_data *data =i2c_get_clientdata (client); +struct pmbus_sensor *sensor; +u8 page =rdev_get_id (rdev); +int ret; + + sensor =pmbus_find_sensor (data, page,PMBUS_READ_VOUT ); +if (IS_ERR (sensor)) +return -ENODATA ; + +mutex_lock (&data->update_lock); +pmbus_update_sensor_data (client, sensor);
Quoted 8 lines
drivers/hwmon/pmbus/pmbus_core.c static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsigned · line 2591
Quoted 6 lines
drivers/hwmon/pmbus/pmbus_core.c +6
-0+if (sensor->data <0 ) + ret = sensor->data; +else + ret = (int )pmbus_reg2data (data, sensor) *1000 ;/* unit is uV */ +mutex_unlock (&data->update_lock); +Same question. Why ?
Same reason as above. Only to get the sensor object for pmbus_reg2data.
Quoted 18 lines
Quoted 18 lines
drivers/hwmon/pmbus/pmbus_core.c +17
-0+return ret; +} + +static int pmbus_regulator_set_voltage (struct regulator_dev *rdev,int min_uV, +int max_uV,unsigned int *selector) +{ +struct device *dev =rdev_get_dev (rdev); +struct i2c_client *client =to_i2c_client (dev->parent); +struct pmbus_data *data =i2c_get_clientdata (client); +struct pmbus_sensor *sensor; +u8 page =rdev_get_id (rdev); +s64 tmp =DIV_ROUND_CLOSEST_ULL (min_uV,1000 );/* convert to mV */ +u16 val; +int ret; + *selector =0 ; + + sensor =pmbus_find_sensor (data, page,PMBUS_VOUT_COMMAND );
Quoted 8 lines
drivers/hwmon/pmbus/pmbus_core.c static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsigned · line 2615
Quoted 7 lines
drivers/hwmon/pmbus/pmbus_core.c +7
-0+if (IS_ERR (sensor)) +return -ENODATA ; + + ret =_pmbus_read_word_data (client, page,0xff ,PMBUS_VOUT_MARGIN_LOW ); +if (ret <0 ) +return ret; +That actually makes me wonder: What about
VOUT_MARGIN_HIGH?
Ok, I will add a check for VOUT_MARGIN_HIGH also.
Quoted 3 lines
Also, there are optional
MFR_VOUT_MINandMFR_VOUT_MAXregisters. Would it possibly make sense to determine the valid range once during probe and then compare against it ?
Maybe this could be a good thing, so that we don't need to read both margins every time. But I guess that would need a new kind of page list with margins added to the pmbus_driver_info struct?
I would prefer to make that change in a separate patch if it's ok with you?
Kind regards
Mårten
Quoted 26 lines
Thanks,
GuenterQuoted 21 lines
drivers/hwmon/pmbus/pmbus_core.c +13
-0+ val =pmbus_data2reg (data, sensor, tmp); + +/* Do not fall shorter than low margin */ +if (ret > val) + val = ret; + + ret =_pmbus_write_word_data (client, page,PMBUS_VOUT_COMMAND , val); + +return ret; +} +const struct regulator_ops pmbus_regulator_ops = { .enable = pmbus_regulator_enable, .disable = pmbus_regulator_disable, .is_enabled = pmbus_regulator_is_enabled, .get_error_flags = pmbus_regulator_get_error_flags, + .get_voltage = pmbus_regulator_get_voltage, + .set_voltage = pmbus_regulator_set_voltage, };EXPORT_SYMBOL_NS_GPL (pmbus_regulator_ops,PMBUS );
Guenter Roeck
maintainerRe: [PATCH v4 4/4] hwmon: (pmbus) Add get_voltage/set_voltage ops
Quoted 37 lines · Marten Lindahl
Quoted 21 lines · Guenter Roeck
Quoted 20 lines · Mårten Lindahl
The pmbus core does not have operations for getting or setting voltage. Add functions get/set voltage for the dynamic regulator framework.
Signed-off-by Mårten Lindahl <[email protected]>--- drivers/hwmon/pmbus/pmbus_core.c | 63 ++++++++++++++++++++++++++++++++ 1 file changed, 63 insertions(+)drivers/hwmon/pmbus/pmbus_core.c +4
-0@@ -1531,6 +1531,11 @@ static const struct pmbus_sensor_attr voltage_attributes[] = { .gbit = PB_STATUS_VOUT_OV, .limit = vout_limit_attrs, .nlimit = ARRAY_SIZE(vout_limit_attrs), + }, { + .reg = PMBUS_VOUT_COMMAND, + .class = PSC_VOLTAGE_OUT, + .paged = true,Quoted 8 lines
drivers/hwmon/pmbus/pmbus_core.c static const struct pmbus_sensor_attr voltage_attributes[] = { · line 1539
Quoted 2 lines
drivers/hwmon/pmbus/pmbus_core.c +1
-0+ .func = PMBUS_HAVE_VOUT, }Ok, you lost me here. This adds an
inX_inputattribute. Why ? This is completely unrelated to the intended scope of this patch. It also doesn't report a measured voltage, but a configuration value. If anything, it would have to be a separate patch, and you'd have to argue hard why it makes sense to report it as measured voltage.I see. The reason for adding this is as simple as I now understand it is wrong. Please remember, my first version of the set/get_voltage functions where hardcoded with L16 input/output. Then in order to use the already existing convertion functions
pmbus_data2regandpmbus_reg2dataI added this only for the need of a sensor object, as those functions are tailored for a sensor object.
Quoted 7 lines
So now I have to ask you for advice. Should I use the existing convertion functions, or do you suggest new variants of them? If reusing them, I guess I have two options:
1: Modify them to take class, page, and data outside of a sensor object as input. 2: Use them as they are, but create a local 'dummy' sensor object with class, page,
and data to use when calling the convertion functions.
I think 2) is the easier and less complex solution for now.
Quoted 4 lines
I hope I made it more clear for you now how I was thinking. There is absolutely no intention of having sensor
inX_inputattributes for reading the setpoint values. This was just an unwanted sideeffect, and I will glady remove it again.
No problem. Thanks for the explanation.
Quoted 66 lines
Quoted 22 lines
Quoted 22 lines
drivers/hwmon/pmbus/pmbus_core.c +15
-0};@@ -2563,11 +2568,69 @@ static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsignedreturn 0 ; } +static int pmbus_regulator_get_voltage (struct regulator_dev *rdev) +{ +struct device *dev =rdev_get_dev (rdev); +struct i2c_client *client =to_i2c_client (dev->parent); +struct pmbus_data *data =i2c_get_clientdata (client); +struct pmbus_sensor *sensor; +u8 page =rdev_get_id (rdev); +int ret; + + sensor =pmbus_find_sensor (data, page,PMBUS_READ_VOUT ); +if (IS_ERR (sensor)) +return -ENODATA ; + +mutex_lock (&data->update_lock); +pmbus_update_sensor_data (client, sensor);Quoted 8 lines
drivers/hwmon/pmbus/pmbus_core.c static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsigned · line 2591
Quoted 6 lines
drivers/hwmon/pmbus/pmbus_core.c +6
-0+if (sensor->data <0 ) + ret = sensor->data; +else + ret = (int )pmbus_reg2data (data, sensor) *1000 ;/* unit is uV */ +mutex_unlock (&data->update_lock); +Same question. Why ?
Same reason as above. Only to get the sensor object for
pmbus_reg2data.Quoted 18 lines
Quoted 18 lines
drivers/hwmon/pmbus/pmbus_core.c +17
-0+return ret; +} + +static int pmbus_regulator_set_voltage (struct regulator_dev *rdev,int min_uV, +int max_uV,unsigned int *selector) +{ +struct device *dev =rdev_get_dev (rdev); +struct i2c_client *client =to_i2c_client (dev->parent); +struct pmbus_data *data =i2c_get_clientdata (client); +struct pmbus_sensor *sensor; +u8 page =rdev_get_id (rdev); +s64 tmp =DIV_ROUND_CLOSEST_ULL (min_uV,1000 );/* convert to mV */ +u16 val; +int ret; + *selector =0 ; + + sensor =pmbus_find_sensor (data, page,PMBUS_VOUT_COMMAND );Quoted 8 lines
drivers/hwmon/pmbus/pmbus_core.c static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsigned · line 2615
Quoted 7 lines
drivers/hwmon/pmbus/pmbus_core.c +7
-0+if (IS_ERR (sensor)) +return -ENODATA ; + + ret =_pmbus_read_word_data (client, page,0xff ,PMBUS_VOUT_MARGIN_LOW ); +if (ret <0 ) +return ret; +That actually makes me wonder: What about
VOUT_MARGIN_HIGH?Ok, I will add a check for
VOUT_MARGIN_HIGHalso.Quoted 3 lines
Also, there are optional
MFR_VOUT_MINandMFR_VOUT_MAXregisters. Would it possibly make sense to determine the valid range once during probe and then compare against it ?
Quoted 6 lines
Maybe this could be a good thing, so that we don't need to read both margins every time. But I guess that would need a new kind of page list with margins added to the
pmbus_driver_infostruct?
I would prefer to make that change in a separate patch if it's ok with you?
I think you need to check for four values right now:
- Try to read MFR_VOUT_MIN. If that does not work, read VOUT_MARGIN_LOW.
- Same for high values.
Ultimately, yes, I think we should add a list of limits. I think it would make more sense though to add the limits to a new regulator specific data structure. Maybe we should create a separate data structure for regulators. Right now we pass struct pmbus_data. Maybe we need struct pmbus_regulator_data which would contain a pointer to struct pmbus_data as well as additional information needed for regulators.
Thanks,
Guenter
Re: [PATCH v4 4/4] hwmon: (pmbus) Add get_voltage/set_voltage ops
Quoted 46 lines · Guenter Roeck
Quoted 37 lines · Marten Lindahl
Quoted 21 lines · Guenter Roeck
Quoted 20 lines · Mårten Lindahl
The pmbus core does not have operations for getting or setting voltage. Add functions get/set voltage for the dynamic regulator framework.
Signed-off-by Mårten Lindahl <[email protected]>--- drivers/hwmon/pmbus/pmbus_core.c | 63 ++++++++++++++++++++++++++++++++ 1 file changed, 63 insertions(+)drivers/hwmon/pmbus/pmbus_core.c +4
-0@@ -1531,6 +1531,11 @@ static const struct pmbus_sensor_attr voltage_attributes[] = { .gbit = PB_STATUS_VOUT_OV, .limit = vout_limit_attrs, .nlimit = ARRAY_SIZE(vout_limit_attrs), + }, { + .reg = PMBUS_VOUT_COMMAND, + .class = PSC_VOLTAGE_OUT, + .paged = true,Quoted 8 lines
drivers/hwmon/pmbus/pmbus_core.c static const struct pmbus_sensor_attr voltage_attributes[] = { · line 1539
Quoted 2 lines
drivers/hwmon/pmbus/pmbus_core.c +1
-0+ .func = PMBUS_HAVE_VOUT, }Ok, you lost me here. This adds an
inX_inputattribute. Why ? This is completely unrelated to the intended scope of this patch. It also doesn't report a measured voltage, but a configuration value. If anything, it would have to be a separate patch, and you'd have to argue hard why it makes sense to report it as measured voltage.I see. The reason for adding this is as simple as I now understand it is wrong. Please remember, my first version of the set/get_voltage functions where hardcoded with L16 input/output. Then in order to use the already existing convertion functions
pmbus_data2regandpmbus_reg2dataI added this only for the need of a sensor object, as those functions are tailored for a sensor object.Quoted 7 lines
So now I have to ask you for advice. Should I use the existing convertion functions, or do you suggest new variants of them? If reusing them, I guess I have two options:
1: Modify them to take class, page, and data outside of a sensor object as input. 2: Use them as they are, but create a local 'dummy' sensor object with class, page,
and data to use when calling the convertion functions.
Quoted 1 line
I think 2) is the easier and less complex solution for now.
Hi Guenter!
This seems to work fine.
Quoted 81 lines
Quoted 5 lines
I hope I made it more clear for you now how I was thinking. There is absolutely no intention of having sensor
inX_inputattributes for reading the setpoint values. This was just an unwanted sideeffect, and I will glady remove it again.No problem. Thanks for the explanation.
Quoted 66 lines
Quoted 22 lines
Quoted 22 lines
drivers/hwmon/pmbus/pmbus_core.c +15
-0};@@ -2563,11 +2568,69 @@ static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsignedreturn 0 ; } +static int pmbus_regulator_get_voltage (struct regulator_dev *rdev) +{ +struct device *dev =rdev_get_dev (rdev); +struct i2c_client *client =to_i2c_client (dev->parent); +struct pmbus_data *data =i2c_get_clientdata (client); +struct pmbus_sensor *sensor; +u8 page =rdev_get_id (rdev); +int ret; + + sensor =pmbus_find_sensor (data, page,PMBUS_READ_VOUT ); +if (IS_ERR (sensor)) +return -ENODATA ; + +mutex_lock (&data->update_lock); +pmbus_update_sensor_data (client, sensor);Quoted 8 lines
drivers/hwmon/pmbus/pmbus_core.c static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsigned · line 2591
Quoted 6 lines
drivers/hwmon/pmbus/pmbus_core.c +6
-0+if (sensor->data <0 ) + ret = sensor->data; +else + ret = (int )pmbus_reg2data (data, sensor) *1000 ;/* unit is uV */ +mutex_unlock (&data->update_lock); +Same question. Why ?
Same reason as above. Only to get the sensor object for
pmbus_reg2data.Quoted 18 lines
Quoted 18 lines
drivers/hwmon/pmbus/pmbus_core.c +17
-0+return ret; +} + +static int pmbus_regulator_set_voltage (struct regulator_dev *rdev,int min_uV, +int max_uV,unsigned int *selector) +{ +struct device *dev =rdev_get_dev (rdev); +struct i2c_client *client =to_i2c_client (dev->parent); +struct pmbus_data *data =i2c_get_clientdata (client); +struct pmbus_sensor *sensor; +u8 page =rdev_get_id (rdev); +s64 tmp =DIV_ROUND_CLOSEST_ULL (min_uV,1000 );/* convert to mV */ +u16 val; +int ret; + *selector =0 ; + + sensor =pmbus_find_sensor (data, page,PMBUS_VOUT_COMMAND );Quoted 8 lines
drivers/hwmon/pmbus/pmbus_core.c static int pmbus_regulator_get_error_flags(struct regulator_dev *rdev, unsigned · line 2615
Quoted 7 lines
drivers/hwmon/pmbus/pmbus_core.c +7
-0+if (IS_ERR (sensor)) +return -ENODATA ; + + ret =_pmbus_read_word_data (client, page,0xff ,PMBUS_VOUT_MARGIN_LOW ); +if (ret <0 ) +return ret; +That actually makes me wonder: What about
VOUT_MARGIN_HIGH?Ok, I will add a check for
VOUT_MARGIN_HIGHalso.Quoted 3 lines
Also, there are optional
MFR_VOUT_MINandMFR_VOUT_MAXregisters. Would it possibly make sense to determine the valid range once during probe and then compare against it ?Quoted 6 lines
Maybe this could be a good thing, so that we don't need to read both margins every time. But I guess that would need a new kind of page list with margins added to the
pmbus_driver_infostruct?
I would prefer to make that change in a separate patch if it's ok with you?
Quoted 4 lines
I think you need to check for four values right now:
- Try to read
MFR_VOUT_MIN. If that does not work, readVOUT_MARGIN_LOW.
- Same for high values.
Ok, I think I first need to check the MFR_VOUT_MIN/MFR_VOUT_MAX registers to make sure they are supported. On LTC2977 they are not supported, so I need to write PMBUS_CLEAR_FAULTS after trying to access them.
I'll send a new patch for this.
Kind regards
Mårten
Quoted 11 lines
Ultimately, yes, I think we should add a list of limits. I think it would make more sense though to add the limits to a new regulator specific data structure. Maybe we should create a separate data structure for regulators. Right now we pass struct
pmbus_data. Maybe we need structpmbus_regulator_datawhich would contain a pointer to structpmbus_dataas well as additional information needed for regulators.Thanks,
Guenter