i2c_transfer() returns the number of messages successfully
transferred, not only a negative errno on failure. When called with
two messages (write command byte followed by a read of the 4-byte
response), a return value of 1 means the command write succeeded but
the read did not complete. In that case, rspbuf remains uninitialized
and must not be interpreted as valid data.
Treat any return value other than ARRAY_SIZE(msg) as an error, and
return -EIO for partial completion. Also return 0 on success instead
of the message count, since the caller only needs to distinguish
success from failure.
Signed-off-by: Sanman Pradhan <psanman@juniper.net>
Link: https://lore.kernel.org/r/20260321181052.27129-4-sanman.pradhan@hpe.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The MAX31785 driver currently uses driver-local wrappers around PMBus
core accesses to enforce a 250us inter-access delay needed to work
around occasional NACKs from the device. This duplicates the PMBus
core delay mechanism already provided by pmbus_driver_info.access_delay
and adds unnecessary complexity.
Replace the PMBus wrapper approach with access_delay for normal
PMBus-mediated accesses, while keeping the minimal local delay handling
needed for raw pre-probe SMBus operations.
For the raw i2c_transfer() long-read path, use pmbus_wait() and
pmbus_update_ts() to keep the PMBus core timing state consistent with
the raw transfer.
Also:
- allow PMBUS_FAN_CONFIG_12 physical-page accesses to fall back to the
PMBus core, while remapping only virtual pages
- use pmbus_update_fan() directly for fan configuration updates
- use the delayed raw read helper for MFR_REVISION during probe
- add a final max31785_wait() before pmbus_do_probe() to bridge the
timing gap between pre-probe accesses and PMBus core registration
- rename 'virtual' to 'vpage', 'driver_data' to 'data', and drop the
unused to_max31785_data() macro
Signed-off-by: Sanman Pradhan <psanman@juniper.net>
Link: https://lore.kernel.org/r/20260321181052.27129-3-sanman.pradhan@hpe.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Export pmbus_wait() and pmbus_update_ts() so that PMBus device
drivers which perform raw I2C transfers outside the core helpers
can keep the PMBus core delay bookkeeping in sync.
Move PMBUS_OP_WRITE and PMBUS_OP_PAGE_CHANGE from pmbus_core.c to
pmbus.h so device drivers can pass the correct operation type flags
to pmbus_update_ts().
This is needed by the max31785 driver, which performs raw
i2c_transfer() calls for its 4-byte extended fan speed reads that
cannot use the standard PMBus word read path.
Signed-off-by: Sanman Pradhan <psanman@juniper.net>
Link: https://lore.kernel.org/r/20260321181052.27129-2-sanman.pradhan@hpe.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The G and J coefficients provided by T-Head TH1520 manual (which calls
them A and C coefficients and calls H coefficient in the binding as B)
have 1/100 degree Celsius precision (the values are 42.74 and -0.16
respectively), however the binding currently only allows coefficients as
precise as 100 milli-Celsius (1/10 degree Celsius).
Change the multipleOf value of these two coefficients to 10 (in the unit
of milli-Celsius) to satisfy the need of TH1520.
Signed-off-by: Icenowy Zheng <zhengxingda@iscas.ac.cn>
Reviewed-by: Drew Fustini <fustini@kernel.org>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Link: https://lore.kernel.org/r/20260309162457.4128205-2-zhengxingda@iscas.ac.cn
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The voltage_show() function previously ignored negative error codes
returned by the underlying SPI read/write functions. Because negative
numbers have their most significant bits set in two's complement, a
failed SPI read returning -EIO (-5) would incorrectly evaluate to true
when masked with MUX_CNV_BM (0x80).
This would cause the driver to enter the polling loop even when the SPI bus
failed, eventually returning a misleading -ETIMEDOUT error to userspace
instead of the actual hardware error. Furthermore, the return values of
the initial SPI write and the final 16-bit SPI read were completely
ignored.
Add proper error checking after every SPI operation to ensure hardware
failures are immediately propagated back to userspace.
Suggested-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Tabrez Ahmed <tabreztalks@gmail.com>
Link: https://lore.kernel.org/r/20260308124714.84715-1-tabreztalks@gmail.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The local wrapper max31785_i2c_write_byte_data() declares its data
parameter as u16 but passes it directly to i2c_smbus_write_byte_data()
which takes u8. Fix the type to match the underlying API.
No functional change; all current callers pass values that fit in u8.
Signed-off-by: Sanman Pradhan <psanman@juniper.net>
Link: https://lore.kernel.org/r/20260307224517.38316-2-sanman.p211993@gmail.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The voltage_show() function returns -1 when the A/D conversion
fails to complete within the polling loop. -1 maps to -EPERM
(operation not permitted), which does not describe the actual
failure.
Replace this -1 error code with -ETIMEDOUT to better indicate
the timeout condition to userspace.
Drop the else block after return.
Note: not runtime tested due to lack of hardware.
Signed-off-by: Tabrez Ahmed <tabreztalks@gmail.com>
Link: https://lore.kernel.org/r/20260307115226.25757-1-tabreztalks@gmail.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Use sysfs_emit() instead of sprintf() in the sysfs show function
voltage_show() to comply with the preferred kernel interface for
writing to sysfs buffers, which ensures PAGE_SIZE buffer limits
are respected.
No functional change intended.
Note: Not runtime tested due to lack of hardware.
Signed-off-by: Tabrez Ahmed <tabreztalks@gmail.com>
Link: https://lore.kernel.org/r/20260307083815.12095-1-tabreztalks@gmail.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This can instead be done with
i2c_client_get_device_id(). For this driver functionality should
not change. Switch over to remove the last couple users of the
i2c_match_id() function from kernel.
Signed-off-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20260306171652.951274-12-afd@ti.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This can instead be done with
i2c_client_get_device_id(). For this driver functionality should
not change. Switch over to remove the last couple users of the
i2c_match_id() function from kernel.
Signed-off-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20260306171652.951274-11-afd@ti.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This can be done instead with
i2c_client_get_device_id() which doesn't need the i2c_device_id
passed in so we do not need to have that forward declared, allowing
us to move the i2c_device_id table down to its more natural spot
with the other module info.
Signed-off-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20260306171652.951274-10-afd@ti.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().
This helper has another benefit:
* It also checks for device match data, which means we do not have
to manually check that first.
Signed-off-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20260306171652.951274-9-afd@ti.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().
This helper has another benefit:
* It also checks for device match data, which means we do not have
to manually check that first.
Signed-off-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20260306171652.951274-8-afd@ti.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().
This helper has another benefit:
* It doesn't need the i2c_device_id passed in so we do not need
to have that forward declared, allowing us to remove that.
Signed-off-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20260306171652.951274-7-afd@ti.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().
This helper has another benefit:
* It doesn't need the i2c_device_id passed in so we do not need
to have that forward declared, allowing us to remove that.
Signed-off-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20260306171652.951274-6-afd@ti.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().
This helper has another benefit:
* It also checks for device match data. That means we do not have
to manually check that first.
Signed-off-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20260306171652.951274-5-afd@ti.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().
This helper has a couple other benefits:
* It doesn't need the i2c_device_id passed in so we do not need
to have that forward declared, allowing us to remove that.
* It also checks for device match data, which allows for OF and
ACPI based probing.
Signed-off-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20260306171652.951274-4-afd@ti.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().
This helper has another benefit:
* It also checks for device match data, which allows for OF based
probing. That means we do not have to manually check those first
and can remove that check.
As i2c_get_match_data() return NULL/0 on failure which also matches
the enum for "cffps1", switch around the enum order so cffps_unknown
is index 0 and existing behavior is preserved.
Signed-off-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20260306171652.951274-3-afd@ti.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().
This helper has another benefit:
* It doesn't need the i2c_device_id passed in so we do not need
to have that forward declared, allowing us to remove that.
Signed-off-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20260306171652.951274-2-afd@ti.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
In all cases in which a struct acpi_driver is used for binding a driver
to an ACPI device object, a corresponding platform device is created by
the ACPI core and that device is regarded as a proper representation of
underlying hardware. Accordingly, a struct platform_driver should be
used by driver code to bind to that device. There are multiple reasons
why drivers should not bind directly to ACPI device objects [1].
Overall, it is better to bind drivers to platform devices than to their
ACPI companions, so convert the asus_atk0110 ACPI driver to a platform
one.
After this change, the subordinate hwmon device will be registered under
the platform device used for driver binding and messages will be printed
relative to that device instead of its ACPI companion.
While this is not expected to alter functionality, it changes sysfs
layout and so it will be visible to user space.
Link: https://lore.kernel.org/all/2396510.ElGaqSPkdT@rafael.j.wysocki/ [1]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/3691136.iIbC2pHGDl@rafael.j.wysocki
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Instead of failing probe when the "adi,rsense-nano-ohms" firmware property
is not provided, default rsense to (NANO/MILLI), or 1 milli-Ohm. This
allows the device to probe without requiring firmware properties, which
might be useful for some high level testing.
Signed-off-by: Nuno Sá <nuno.sa@analog.com>
Link: https://lore.kernel.org/r/20260304-hwmon-ltc4282-minor-improvs-v1-2-344622924d3a@analog.com
[groeck: Clarify that the default is 1 milli-Ohm. No functional change.]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The INA219 has the lowest three bits of the bus voltage register
zero-reserved, the bus_voltage_shift ina2xx_config field was introduced
to accommodate for that.
The INA234 has four bits of the bus voltage, of the shunt voltage, and
of the current registers zero-reserved but the latter two were
implemented by choosing a 16x higher shunt_div instead of a separate
field specifying a bit shift.
This is possible because shunt voltage and current are divided by
shunt_div, hence a 16x higher shunt_div results in a 16x smaller LSB for
both the shunt voltage and the current register, perfectly accounting
for the missing bit shift.
For consistency and correctness, account for the reserved bits via
shunt_voltage_shift and current_shift configuration fields as already
done for voltage registers and use the conversion constants given in the
INA234 datasheet.
Signed-off-by: Jonas Rebmann <jre@pengutronix.de>
Link: https://lore.kernel.org/r/20260303-ina234-shift-v1-2-318c33ac4480@pengutronix.de
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The list of supported chips in the header is incomplete and contains no
other information not readily available. Remove the list and instead
hint that the chips supported by this driver have 219/226 compatible
register layout [unlike the ones supported by e.g. ina238].
Remove the unused INA226_DIE_ID define.
Signed-off-by: Jonas Rebmann <jre@pengutronix.de>
Link: https://lore.kernel.org/r/20260303-ina234-shift-v1-1-318c33ac4480@pengutronix.de
[groeck: macro -> define in commit message]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Add hardware monitoring support for the LattePanda Sigma SBC
(DFRobot, ITE IT8613E EC). The driver reads fan speed and
temperatures via direct port I/O, as the BIOS disables the
ACPI EC interface.
Signed-off-by: Mariano Abad <weimaraner@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
In all cases in which a struct acpi_driver is used for binding a driver
to an ACPI device object, a corresponding platform device is created by
the ACPI core and that device is regarded as a proper representation of
underlying hardware. Accordingly, a struct platform_driver should be
used by driver code to bind to that device. There are multiple reasons
why drivers should not bind directly to ACPI device objects [1].
Overall, it is better to bind drivers to platform devices than to their
ACPI companions, so convert the hwmon ACPI power meter driver to a
platform one.
After this change, the subordinate hwmon device will be registered
under the platform device representing the ACPI power meter, sysfs
notifications will trigger on that device, and diagnostic messages
will be printed relative to it instead of its ACPI companion.
Link: https://lore.kernel.org/all/2396510.ElGaqSPkdT@rafael.j.wysocki/ [1]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/1952740.tdWV9SEqCh@rafael.j.wysocki
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Since acpi_power_meter_notify() and acpi_power_meter_remove() are
.notify() and .remove() callback functions of an ACPI driver,
respectively, the first argument of the former and the only argument
of the latter cannot be NULL. Likewise, the acpi_power_meter_resume()
argument cannot be NULL because it is a system resume callback
function.
Moreover, since all of these functions can only run after
acpi_power_meter_add() has returned 0, the driver_data field in the
struct acpi_device object used by them cannot be NULL either.
Accordingly, drop the redundant "device" checks against NULL from
acpi_power_meter_notify() and acpi_power_meter_remove(), drop the
redundant "dev" check against NULL from acpi_power_meter_resume(),
and drop the redundant acpi_driver_data() checks against NULL from
all of these functions.
Additionally, combine the initialization of the "resource" local
variable in acpi_power_meter_notify() and acpi_power_meter_remove()
with its declaration.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/5085645.31r3eYUQgx@rafael.j.wysocki
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Extend the PMBus core vrm_version handling to support NVIDIA nvidia195mv
VID mode. This adds a new VRM/VID encoding type and the corresponding
voltage conversion logic so devices reporting nvidia195mv can have their
VOUT/VID values interpreted correctly by the hwmon PMBus core.
Signed-off-by: Ashish Yadav <ashish.yadav@infineon.com>
Link: https://lore.kernel.org/r/20260223050804.4287-3-Ashish.Yadav@infineon.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
