The three ways of temperature sensing in Linux
Hi all,
[Not sure if this is TECH or CORE, as it touches more than one subsystem]
Not sure how late it is to post this, but I am trying anyway.
During LPC 2015 last week it came out the topic of the interaction between the thermal subsystem with HWMON. On top of it, we also got a proposal [1] of having thermal exposing its registered devices as IIO.
I know the three subsystem coexist today for some years. They also have their own design goals. However, the fact that we have three different subsystems for doing very similar task may be confusing, for the driver writer, and for user space too.
Another aspect of it is the possible code duplication. For example, it would interesting to get board temperature sensors, typically registered as hwmon devices, to be available in the thermal subsystem while constructing thermal zones out of them. Some user process would expect the opposite though, to read the thermal zone sensors as hwmon devices, or as IIO devices.
Therefore, it would be interesting to have a better convergence on how these subsystems talk to each other, and how this is exposed to user space.
The involved maintainers are:
HWMON:
M: Jean Delvare <[email protected]> M: Guenter Roeck <[email protected]>
Thermal:
M: Zhang Rui <[email protected]> M: Eduardo Valentin <[email protected]>
IIO:
M: Jonathan Cameron <[email protected]>
BR,
Eduardo Valentin
[1] - http://comments.gmane.org/gmane.linux.kernel.iio/19627
Quoted 16 lines · Eduardo Valentin
Hi all,
[Not sure if this is TECH or CORE, as it touches more than one subsystem]
Not sure how late it is to post this, but I am trying anyway.
During LPC 2015 last week it came out the topic of the interaction between the thermal subsystem with HWMON. On top of it, we also got a proposal [1] of having thermal exposing its registered devices as IIO.
I know the three subsystem coexist today for some years. They also have their own design goals. However, the fact that we have three different subsystems for doing very similar task may be confusing, for the driver writer, and for user space too.
Quoted 6 lines
Another aspect of it is the possible code duplication. For example, it would interesting to get board temperature sensors, typically registered as hwmon devices, to be available in the thermal subsystem while constructing thermal zones out of them. Some user process would expect the opposite though, to read the thermal zone sensors as hwmon devices, or as IIO devices.
I would rather expect using only one subsystem for exporting this information to user-space. If the susbsytem lacks some features then it could be expanded.
Some power supply drivers (fuel gauges or chargers) also provide temperature data. They use the thermal for that. It would be weird to export them through IIO...
Anyway from that point of view the topic is interesting to me.
Best regards,
Krzysztof
Quoted 30 lines
Therefore, it would be interesting to have a better convergence on how these subsystems talk to each other, and how this is exposed to user space.
The involved maintainers are:
HWMON:M: Jean Delvare <[email protected]> M: Guenter Roeck <[email protected]>Thermal:
M: Zhang Rui <[email protected]> M: Eduardo Valentin <[email protected]>IIO:
M: Jonathan Cameron <[email protected]>BR,
Eduardo Valentin
[1] - http://comments.gmane.org/gmane.linux.kernel.iio/19627
_______________________________________________
Ksummit-discuss mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/ksummit-discuss
Quoted 24 lines · Krzysztof Kozlowski
Quoted 16 lines · Eduardo Valentin
Hi all,
[Not sure if this is TECH or CORE, as it touches more than one subsystem]
Not sure how late it is to post this, but I am trying anyway.
During LPC 2015 last week it came out the topic of the interaction between the thermal subsystem with HWMON. On top of it, we also got a proposal [1] of having thermal exposing its registered devices as IIO.
I know the three subsystem coexist today for some years. They also have their own design goals. However, the fact that we have three different subsystems for doing very similar task may be confusing, for the driver writer, and for user space too.
Quoted 6 lines
Another aspect of it is the possible code duplication. For example, it would interesting to get board temperature sensors, typically registered as hwmon devices, to be available in the thermal subsystem while constructing thermal zones out of them. Some user process would expect the opposite though, to read the thermal zone sensors as hwmon devices, or as IIO devices.
Quoted 8 lines
I would rather expect using only one subsystem for exporting this information to user-space. If the susbsytem lacks some features then it could be expanded.
Some power supply drivers (fuel gauges or chargers) also provide temperature data. They use the thermal for that. It would be weird to export them through IIO...
I agree, the current situation is less than optimal.
There are bridges from thermal and iio to hwmon, but they are not perfect. Resulting hwmon device names are odd, and iio doesn't support limits. The thermal-hwmon bridge only supports one trip point, and it doesn't pass alarms.
Both the iio-hwmon and the thermal-hwmon bridge break the hwmon model of reporting sensor data per chip, which makes it difficult to convert drivers from hwmon to thermal (if that is suggested) without breaking the userspace ABI.
At the same time, there is no bridge from the hwmon core to thermal. Currently hwmon drivers supporting thermal register with both hwmon and thermal, which is really undesirable. Implementing a hwmon-thermal bridge in the hwmon core would require implementing a new hwmon registration API. That is easier said then done, and would require more time than at least I have available.
This, I think, is the core of the problem. Somehow would have to go in and create that new hwmon registration API, if that is a solution acceptable to the thermal maintainers. At the same time, the thermal-hwmon bridge should be improved to export provide available information to hwmon.
An alternative would be for thermal to 'take over', convert all hwmon drivers to thermal drivers, and retain a compatibility layer to provide the data as hwmon device data to user space without breaking the existing hwmon ABI. Again, someone would have to do that work.
Guenter
Hey Guenter!
Quoted 46 lines · Guenter Roeck
Quoted 24 lines · Krzysztof Kozlowski
Quoted 16 lines · Eduardo Valentin
Hi all,
[Not sure if this is TECH or CORE, as it touches more than one subsystem]
Not sure how late it is to post this, but I am trying anyway.
During LPC 2015 last week it came out the topic of the interaction between the thermal subsystem with HWMON. On top of it, we also got a proposal [1] of having thermal exposing its registered devices as IIO.
I know the three subsystem coexist today for some years. They also have their own design goals. However, the fact that we have three different subsystems for doing very similar task may be confusing, for the driver writer, and for user space too.
Quoted 6 lines
Another aspect of it is the possible code duplication. For example, it would interesting to get board temperature sensors, typically registered as hwmon devices, to be available in the thermal subsystem while constructing thermal zones out of them. Some user process would expect the opposite though, to read the thermal zone sensors as hwmon devices, or as IIO devices.
Quoted 8 lines
I would rather expect using only one subsystem for exporting this information to user-space. If the susbsytem lacks some features then it could be expanded.
Some power supply drivers (fuel gauges or chargers) also provide temperature data. They use the thermal for that. It would be weird to export them through IIO...
I agree, the current situation is less than optimal.
There are bridges from thermal and iio to hwmon, but they are not perfect. Resulting hwmon device names are odd, and iio doesn't support limits. The thermal-hwmon bridge only supports one trip point, and it doesn't pass alarms.
Both the iio-hwmon and the thermal-hwmon bridge break the hwmon model of reporting sensor data per chip, which makes it difficult to convert drivers from hwmon to thermal (if that is suggested) without breaking the userspace ABI.
Quoted 6 lines
At the same time, there is no bridge from the hwmon core to thermal. Currently hwmon drivers supporting thermal register with both hwmon and thermal, which is really undesirable. Implementing a hwmon-thermal bridge in the hwmon core would require implementing a new hwmon registration API. That is easier said then done, and would require more time than at least I have available.
Yes, we do have some bridges in place, and more coming (as thermal-iio RFC). I just would like to have a more coordinated or a better overall agreegation of these bridges. Specially to avoid breaking compatibility, as you mentioned in the existing thermal-hwmon, and iio-hwmon cases. This is the original motivation of this proposal, define a clear path for this overlap.
Quoted 6 lines
This, I think, is the core of the problem. Somehow would have to go in and create that new hwmon registration API, if that is a solution acceptable to the thermal maintainers. At the same time, the thermal-hwmon bridge should be improved to export provide available information to hwmon.
I don't think this is much of a concern from thermal subsystem, mainly because most of the work would be first at the hwmon side.
On the other hand, I definitely agree that we need to sit down and architecture what are the needed shared information between the subsystems. And on top of that, how do we present them to userspace? To avoid, for instance, having a half baked hwmon device showing up in sysfs, if that is coming from thermal.
Quoted 5 lines
An alternative would be for thermal to 'take over', convert all hwmon drivers to thermal drivers, and retain a compatibility layer to provide the data as hwmon device data to user space without breaking the existing hwmon ABI. Again, someone would have to do that work.
This is more in the direction of having a single subsystem to perform the task of monitoring temperature. The amount of work will be similar in either cases above, I would say. There is no simple way out, we would need to touch all drivers.
Of course, this is a view in the hwmon <-> thermal case, but we need to consider the iio interactions too. This is why I though of starting this topic :-)
I think a first step would be to clarify the data that would be needed to be shared between the subsystems. What are the involved callbacks, and monitoring modes? Once that is clear enough, we can properly design and write better bridges across the three subsystems. A latter step would be to discuss the convergence, if needed at all, to a single way of monitoring temperature (within kernel and userspace).
Quoted 2 lines
Guenter
BR,
Eduardo Valentin
Quoted 24 lines · Krzysztof Kozlowski
Quoted 16 lines · Eduardo Valentin
Hi all,
[Not sure if this is TECH or CORE, as it touches more than one subsystem]
Not sure how late it is to post this, but I am trying anyway.
During LPC 2015 last week it came out the topic of the interaction between the thermal subsystem with HWMON. On top of it, we also got a proposal [1] of having thermal exposing its registered devices as IIO.
I know the three subsystem coexist today for some years. They also have their own design goals. However, the fact that we have three different subsystems for doing very similar task may be confusing, for the driver writer, and for user space too.
Quoted 6 lines
Another aspect of it is the possible code duplication. For example, it would interesting to get board temperature sensors, typically registered as hwmon devices, to be available in the thermal subsystem while constructing thermal zones out of them. Some user process would expect the opposite though, to read the thermal zone sensors as hwmon devices, or as IIO devices.
Quoted 3 lines
I would rather expect using only one subsystem for exporting this information to user-space. If the susbsytem lacks some features then it could be expanded.
Yes. That would be the best scenario. I agree here. But in practice, we have three different subsystem doing similar things around temperature monitoring. Then, the topic is really to discuss how to either converge to having one (which at this point, I doubt it would be that straight forward, specially considering the fact that we must avoid breaking userspace) or at least having the subsystems having a better interaction between each other.
Quoted 4 lines
Some power supply drivers (fuel gauges or chargers) also provide temperature data. They use the thermal for that. It would be weird to export them through IIO...
Agreed here too.
Quoted 2 lines
Anyway from that point of view the topic is interesting to me.
Cool!
BR,
Eduardo Valentin
Quoted 35 lines
Best regards,
KrzysztofQuoted 30 lines
Therefore, it would be interesting to have a better convergence on how these subsystems talk to each other, and how this is exposed to user space.
The involved maintainers are:
HWMON:M: Jean Delvare <[email protected]> M: Guenter Roeck <[email protected]>Thermal:
M: Zhang Rui <[email protected]> M: Eduardo Valentin <[email protected]>IIO:
M: Jonathan Cameron <[email protected]>BR,
Eduardo Valentin
[1] - http://comments.gmane.org/gmane.linux.kernel.iio/19627
_______________________________________________
Ksummit-discuss mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/ksummit-discuss
Am Montag, 24. August 2015, 23:12:10 schrieb Eduardo Valentin:
Quoted 23 lines
Hi all,
[Not sure if this is TECH or CORE, as it touches more than one subsystem]
Not sure how late it is to post this, but I am trying anyway.
During LPC 2015 last week it came out the topic of the interaction between the thermal subsystem with HWMON. On top of it, we also got a proposal [1] of having thermal exposing its registered devices as IIO.
I know the three subsystem coexist today for some years. They also have their own design goals. However, the fact that we have three different subsystems for doing very similar task may be confusing, for the driver writer, and for user space too.
Another aspect of it is the possible code duplication. For example, it would interesting to get board temperature sensors, typically registered as hwmon devices, to be available in the thermal subsystem while constructing thermal zones out of them. Some user process would expect the opposite though, to read the thermal zone sensors as hwmon devices, or as IIO devices.
Quoted 3 lines
Therefore, it would be interesting to have a better convergence on how these subsystems talk to each other, and how this is exposed to user space.
On Rockchip we currently have two scenarios:
rk3066 simply has a "dumb" (ts-)adc just emitting voltages and is actually the same ip as the saradc and thus an iio adc driver. It will need something like the iio-thermal driver that was posted by someone quite some time ago to actually get temperatures and act on them.
The rk3288 on the other hand has a much more intelligent tsadc - including temperature-specific actions, which also already has a thermal driver in the kernel.
So far I always thought hwmon was some sort of grandfather to the thermal subsystem without any specific api, but I've also not needed one at all yet.
Heiko
Quoted 4 lines · Heiko Stuebner
So far I always thought hwmon was some sort of grandfather to the thermal subsystem without any specific api, but I've also not needed one at all yet.
Yes, I have heard that before.
Its scope is quite different, though. One does thermal management, one monitors the hardware. Monitoring the hardware is a bit more than just monitoring its temperature. It also means to monitor and report voltages, current, power, energy, humidity, and fan speeds, as well as associated alarms. Thermal management takes an active role, hardware monitoring by its nature doesn't.
On top of that, many devices report much more than temperatures. System managers may report and sometimes control all of the above. SuperIO chips monitor voltages, temperatures, and fan speeds, and often implement fan control. PMBus devices primarily control voltages, but may report literally everything except for humidity (including fan speeds). And so on.
Maybe just have a look into drivers/hwmon/pmbus/ltc2978.c, and let me know where you would like to have this driver (and all the other PMBus drivers) moved to.
Thanks,
Guenter
Quoted 6 lines · Guenter Roeck
Quoted 4 lines · Heiko Stuebner
So far I always thought hwmon was some sort of grandfather to the thermal subsystem without any specific api, but I've also not needed one at all yet.
Yes, I have heard that before.
And that is a misconception, I would say, as both have different design goals. Keep in mind that my original proposal is to discuss what touches the monitoring part.
Quoted 6 lines
Its scope is quite different, though. One does thermal management, one monitors the hardware. Monitoring the hardware is a bit more than just monitoring its temperature. It also means to monitor and report voltages, current, power, energy, humidity, and fan speeds, as well as associated alarms. Thermal management takes an active role, hardware monitoring by its nature doesn't.
I agree here. Also, keep in mind that the scope of thermal subsystem is not only monitor temperature, but also cover what needs to be done when the system state changes and needs action, typically on top of temperature thresholds/trip points.
Quoted 6 lines
On top of that, many devices report much more than temperatures. System managers may report and sometimes control all of the above. SuperIO chips monitor voltages, temperatures, and fan speeds, and often implement fan control. PMBus devices primarily control voltages, but may report literally everything except for humidity (including fan speeds). And so on.
There are discussions to have power / voltage / current (and humidity too) inputs to the thermal subsystem. And that's probably where the issue gets worse. Another take we could have is to centralize the hw monitoring part of the problem in, well, hwmon, and make thermal a user of it. Then we would move the thermal drivers, at least the sensing part of them, to hwmon. But then again, we still have the IIO case :-).
BR,
Eduardo Valentin
Quoted 14 lines
Maybe just have a look into
drivers/hwmon/pmbus/ltc2978.c, and let me know where you would like to have this driver (and all the other PMBus drivers) moved to.Thanks,
Guenter
_______________________________________________
Ksummit-discuss mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/ksummit-discuss
Quoted 12 lines · Eduardo Valentin
Quoted 6 lines · Guenter Roeck
Quoted 4 lines · Heiko Stuebner
So far I always thought hwmon was some sort of grandfather to the thermal subsystem without any specific api, but I've also not needed one at all yet.
Yes, I have heard that before.
And that is a misconception, I would say, as both have different design goals. Keep in mind that my original proposal is to discuss what touches the monitoring part.
Is that still true for thermal ? I have the impression that there is a desire to move the monitoring part into thermal.
Quoted 20 lines
Quoted 6 lines
Its scope is quite different, though. One does thermal management, one monitors the hardware. Monitoring the hardware is a bit more than just monitoring its temperature. It also means to monitor and report voltages, current, power, energy, humidity, and fan speeds, as well as associated alarms. Thermal management takes an active role, hardware monitoring by its nature doesn't.
I agree here. Also, keep in mind that the scope of thermal subsystem is not only monitor temperature, but also cover what needs to be done when the system state changes and needs action, typically on top of temperature thresholds/trip points.
Quoted 6 lines
On top of that, many devices report much more than temperatures. System managers may report and sometimes control all of the above. SuperIO chips monitor voltages, temperatures, and fan speeds, and often implement fan control. PMBus devices primarily control voltages, but may report literally everything except for humidity (including fan speeds). And so on.
Quoted 7 lines
There are discussions to have power / voltage / current (and humidity too) inputs to the thermal subsystem. And that's probably where the issue gets worse. Another take we could have is to centralize the hw monitoring part of the problem in, well, hwmon, and make thermal a user of it. Then we would move the thermal drivers, at least the sensing part of them, to hwmon. But then again, we still have the IIO case :-).
My suggestion would have been to separate subsystems into logical entities. Let thermal handle thermal policies and decisions, let hwmon read (and write) low speed sensor values, let iio handle high speed io, and interconnect it all with sensible APIs.
However, it seems to me that the direction in the thermal community is more along the line of handling both monitoring and the decision making process in thermal (and maybe some of it in iio). Or, in other words, to rewrite all hwmon drivers, give them a new API, and move them into, say, drivers/thermal/chips/.
I personally think that this might not be such a good idea, as it will distract the thermal subsystem from its original focus. But the focus may be shifting, or maybe I am just plain wrong.
Given that, I would not object if that is where things are going. All I am asking for is to keep supporting the existing ABI.
Thanks,
Guenter
Quoted 43 lines · Eduardo Valentin
Quoted 3 lines · Guenter Roeck
Quoted 2 lines · Heiko Stuebner
So far I always thought hwmon was some sort of grandfather to the
thermal
Quoted 1 line
Quoted 1 line
subsystem without any specific api, but I've also not needed one at
all yet.
Quoted 2 lines
Quoted 1 line
Yes, I have heard that before.
And that is a misconception, I would say, as both have different design goals. Keep in mind that my original proposal is to discuss what touches
the monitoring part.Quoted 2 lines
Its scope is quite different, though. One does thermal management,
one monitors
Quoted 1 line
the hardware. Monitoring the hardware is a bit more than just
monitoring its
Quoted 1 line
temperature. It also means to monitor and report voltages, current,
power,
Quoted 1 line
energy, humidity, and fan speeds, as well as associated alarms.
Thermal
Quoted 1 line
management takes an active role, hardware monitoring by its nature
doesn't.
I agree here. Also, keep in mind that the scope of thermal subsystem is not only monitor temperature, but also cover what needs to be done when the system state changes and needs action, typically on top of temperature thresholds/trip points.
Quoted 3 lines
On top of that, many devices report much more than temperatures. System managers may report and sometimes control all of the above.
SuperIO
Quoted 1 line
chips monitor voltages, temperatures, and fan speeds, and often
implement fan
Quoted 1 line
control. PMBus devices primarily control voltages, but may report
literally
Quoted 1 line
everything except for humidity (including fan speeds). And so on.
Quoted 9 lines
There are discussions to have power / voltage / current (and humidity too) inputs to the thermal subsystem. And that's probably where the issue
gets worse. Another take we could have is to centralize the hw monitoring
part of the problem in, well, hwmon, and make thermal a user of it. Then
we would move the thermal drivers, at least the sensing part of them, to hwmon. But then again, we still have the IIO case :-).
At least for temperature monitoring with temp specific sensors, right now IIO
doesn't really have any overlap.
We do have temp sensors as part of IMUs and humidity sensors plus a few unusual temp sensors (remote IR sensors and high speed thermo couple devices)
The iio hwmon bridge was meant for case where a fast ADC happens to be used for a slow hwmon job.
As I understand it, the proposed IIO interface for thermal is due to a need for a low overhead fat pipe to userspace governors.
Current bit IIO doesn't do that are relevant:
1) event access within the kernel (trip points?) Userspace interface been present from the start.
2) multiple trip points of the same type.
(Not a big addition but not relevant for
our core use cases and mostly only present on hwmon chips)
Note ultimate intent for a long time has been to fully separate IIO backend and
frontend (userspace interface). No one has had time to work on this for a while though. So right now you can sit what you like on IIO but the IIO interface will be there as well, if unused.
Speaking of which, don't forget the other overlapping subsystems.
E.g. input and battery/power.
We have a non mainline input bridge and
a few users for battery monitoring in mainline.
At the end of the day lots of things (fast and slow) can be wired up to ridiculous rate ADCs. Not unusual on SOCs for a
1Msp ADC to be used to read the battery voltage.
Plus there will always be someone out there using a £1000+ IMUs in their gaming rig
Jonathan
Quoted 28 lines
BR,
Eduardo Valentin
Quoted 2 lines
Maybe just have a look into
drivers/hwmon/pmbus/ltc2978.c, and let meknow
Quoted 1 line
where you would like to have this driver (and all the other PMBus
drivers)
Quoted 11 lines
moved to.
Thanks,
Guenter
_______________________________________________
Ksummit-discuss mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/ksummit-discuss------------------------------------------------------------------------
_______________________________________________
Ksummit-discuss mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/ksummit-discuss
Signature · 1 line
Sent from my Android device with K-9 Mail. Please excuse my brevity.