Skip to contentSkip to conversation

Creating a driver for MAX31827 temperature switch

DISCUSSION Daniel Matyas 4 participants 4 messages Apr 2, 2023 – Apr 3, 2023

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

Dear Kernel community,

I am developing an IIO driver for a temperature switch, which communicates through I2C at Analog Devices Inc.

When implementing the event handling for the comparator mode of the device, I faced a problem: I don't know how to differentiate the underTemp event from the overTemp event. To understand better, I suggest you check out the device's data sheet for Address map (page 12), Configuration Register Definition (page 13) and OT/UT Status Bits and ALARM Pin Behavior (page 15) - https://www.analog.com/media/en/technical-documentation/data-sheets/MAX31827-MAX31829.pdf

I had the idea to make 2 channels with the exact same attributes, but with different indexes, so that I can store the overTemp related events on channel 0 and underTemp related events on channel 1. Even so, I don't really feel like this is the right solution. Can anyone give me some advice on this?

Also, I was suggested that I convert my driver from IIO to HWMON. Do you think that this is needed?

Yours faithfully,
Daniel Matyas

#1 lore raw

Am 02.04.23 um 23:31 schrieb Daniel Matyas:

Quoted 4 lines

Dear Kernel community,

I am developing an IIO driver for a temperature switch, which communicates through I2C at Analog Devices Inc.

Quoted 8 lines

When implementing the event handling for the comparator mode of the device, I faced a problem: I don't know how to differentiate the underTemp event from the overTemp event. To understand better, I suggest you check out the device's data sheet for Address map (page 12), Configuration Register Definition (page 13) and OT/UT Status Bits and ALARM Pin Behavior (page 15) - https://www.analog.com/media/en/technical-documentation/data-sheets/MAX31827-MAX31829.pdf

I had the idea to make 2 channels with the exact same attributes, but with different indexes, so that I can store the overTemp related events on channel 0 and underTemp related events on channel 1. Even so, I don't really feel like this is the right solution. Can anyone give me some advice on this?

Also, I was suggested that I convert my driver from IIO to HWMON. Do you think that this is needed?

Yours faithfully,
Daniel Matyas

Hi,

regarding the question whether the driver should be a hwmon driver: i depends on the usage of the temperature sensor. If its a low sample rate temperature sensor used to mainly monitor the temperature of the computer system itself, then the driver should be a hwmon driver. The datasheet suggests that the sensor device is intended for monitoring computer systems, so a hwmon driver would make sense (see https://www.kernel.org/doc/html/latest/driver-api/iio/intro.html).

Armin Wolf

maintainer
#2 lore raw
In reply to Daniel Matyas
Quoted 10 lines · Daniel Matyas

Dear Kernel community,

I am developing an IIO driver for a temperature switch, which communicates through I2C at Analog Devices Inc.

When implementing the event handling for the comparator mode of the device, I faced a problem: I don't know how to differentiate the underTemp event from the overTemp event. To understand better, I suggest you check out the device's data sheet for Address map (page 12), Configuration Register Definition (page 13) and OT/UT Status Bits and ALARM Pin Behavior (page 15) - https://www.analog.com/media/en/technical-documentation/data-sheets/MAX31827-MAX31829.pdf

I had the idea to make 2 channels with the exact same attributes, but with different indexes, so that I can store the overTemp related events on channel 0 and underTemp related events on channel 1. Even so, I don't really feel like this is the right solution. Can anyone give me some advice on this?

Also, I was suggested that I convert my driver from IIO to HWMON. Do you think that this is needed?

This is a typical hardware monitoring chip, and the driver should be a hardware monitoring driver.

Guenter

Quoted 2 lines

Yours faithfully,
Daniel Matyas

maintainer
#3 lore raw
In reply to Daniel Matyas
Quoted 7 lines · On Mon, 3 Apr 2023 00:31:21 +0300 Daniel Matyas <[email protected]>…

Dear Kernel community,

I am developing an IIO driver for a temperature switch, which communicates through I2C at Analog Devices Inc.

When implementing the event handling for the comparator mode of the device, I faced a problem: I don't know how to differentiate the underTemp event from the overTemp event. To understand better, I suggest you check out the device's data sheet for Address map (page 12), Configuration Register Definition (page 13) and OT/UT Status Bits and ALARM Pin Behavior (page 15) - https://www.analog.com/media/en/technical-documentation/data-sheets/MAX31827-MAX31829.pdf

I had the idea to make 2 channels with the exact same attributes, but with different indexes, so that I can store the overTemp related events on channel 0 and underTemp related events on channel 1. Even so, I don't really feel like this is the right solution. Can anyone give me some advice on this?

If it's just a question of over and under temp, I'd imagine these map to two different directions of IIO event. You can have one channel with RISING and FALLING events on it.

e.g.
https://elixir.bootlin.com/linux/v6.2.9/source/drivers/iio/adc/max1363.c#L451

The place we run into IIO limitations is if you have two events in the same direction (warn and critical for example) When that happens it is almost always a device that fits better in hwmon anyway so we've never figured out how to enable this sort of thing. Doesn't seem to apply here though.

Quoted 2 lines

Also, I was suggested that I convert my driver from IIO to HWMON. Do you think that this is needed?

No obvious reason why this benefits from being in IIO so I'm agreeing with the others who have replied that this looks more suited to hwmon.

Jonathan

Quoted 3 lines

Yours faithfully,
Daniel Matyas