Change sensor Value to double
Generally most demons we've looked at take the int64 values and
immediately translates the value to a double. As the individual demons
are doing the floating point computation anyway, there doesn't seem to be
a large benefit to keeping separating the value on dbus, only to recompute
in each demon, causing extra overhead for consumers. In some cases this
causes an extra d-bus call, for instance when reading a threshold a
consumer to know the value needs to also access the value field to get
the scale, which is arguably more expensive than having the property be a
double in the first place.
Change-Id: If47c19f77286e3764d167a6843ba91e2ca1769b8
Signed-off-by: James Feist <james.feist@linux.intel.com>
diff --git a/xyz/openbmc_project/Sensor/Value.interface.yaml b/xyz/openbmc_project/Sensor/Value.interface.yaml
index 53710e0..91f356a 100644
--- a/xyz/openbmc_project/Sensor/Value.interface.yaml
+++ b/xyz/openbmc_project/Sensor/Value.interface.yaml
@@ -19,15 +19,15 @@
properties:
- name: Value
- type: int64
+ type: double
description: >
The sensor reading.
- name: MaxValue
- type: int64
+ type: double
description: >
The Maximum supported sensor reading.
- name: MinValue
- type: int64
+ type: double
description: >
The Minimum supported sensor reading.
- name: Unit
@@ -42,11 +42,6 @@
For objects in the power namespace, Unit must be "Watts".
For objects in the energy namespace, Unit must be "Joules".
For objects in the utilization namespace, Unit must be "Percent"
- - name: Scale
- type: int64
- description: >
- The reading scaling factor N, where the actual reading is Value * 10^N.
- This is immutable once set for a sensor.
enumerations:
- name: Unit