[
https://issues.apache.org/jira/browse/MYNEWT-744?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vipul Rahane updated MYNEWT-744:
--------------------------------
Description:
- Change data structure to SLIST instead of TAILQ so that it can get
initialized statically. This is needed so that sysinit could be used for
initializing the sensors.
- Remove dummy functions in the bsp and add the sensor device initialization
functions instead.
- Adding syscfg based configuration for onboard Vs offboard sensors
- Adding sensor creator package for creating external sensor devices if they
are not on the board itself. This is just for convenience sake. The right way
to do this would be from the bsp. As a result, removed sensor device creation
from nrf52dk and nrf52840pdk bsps.
- Add sensor interface struct so that it can get used by the individual drivers
so that we can talk to multiple sensors using the same driver using same or
different interfaces.
- For TSL2561, apart from the above changes, various fixes were made:
1. Making a few changes to this driver along with sensor_itf changes
2. Gain, integration time and enabled flag was a global instead of reading
every time gain is set or integration time is set. Changing it so that we read
the timing register before setting either gain or integration time.
3. Also changing a few functions in the driver to handle errors more
gracefully. Changing shell to work with new functions.
4. Making changes so that it goes in synch with the drivers.
5. Removing delays from the read function. the poller should be set
according to the other config values being set so that the sensor is read at
the correct interval.
- Add a mask field to the sensor structure so that differentiating between
supported Vs configured sensors is possible
- Change the sensor_mgr_match_bytype() function to check w.r.t mask and make it
a public API.
- Adding config variable for mask for different drivers in their respective
structures.
- Adding an API to set the sensor types based on the mask.
- Moving config_sensor() out of the app into the creator package and the bsp.
Objective is to show how a developer would initialize and configure the sensor
in the bsp which is normally the practice.
- Changing sensor_shell command "list" to show only configured types.
- sensor_shell cmd "type" shows supported types set by the driver.
- In the sensor_oic source file changing code to check type only while
allocating resources.
- Changing device name in the sensors_test app to be "sensy" instead of "pi"
- For the LSM303DLHC driver the address field in the "sensor_itf" structure is
not enough since it talks to the sensor using two addresses: one for
accelerometer and one for magnetometer. The best way to deal with this is to
make the addresses config variables in the sensor driver structure. For any
such sensor devices which have multiple addresses, a similar approach should be
followed.
- Have changes the drivers to work with all the fixes mentioned above.
was:
1. Changa data structure to SLIST instead of TAILQ so that it can get
initialized statically.
2. Remove dummy functions in the bsp and add the sensor device initialization
functions instead.
3. sensor_dev_create() should be moved out of hal_bsp.c.
> SensorAPI: improvements
> -----------------------
>
> Key: MYNEWT-744
> URL: https://issues.apache.org/jira/browse/MYNEWT-744
> Project: Mynewt
> Issue Type: Improvement
> Security Level: Public(Viewable by anyone)
> Reporter: Vipul Rahane
> Assignee: Vipul Rahane
> Fix For: v1_1_0_rel
>
>
> - Change data structure to SLIST instead of TAILQ so that it can get
> initialized statically. This is needed so that sysinit could be used for
> initializing the sensors.
> - Remove dummy functions in the bsp and add the sensor device initialization
> functions instead.
> - Adding syscfg based configuration for onboard Vs offboard sensors
> - Adding sensor creator package for creating external sensor devices if they
> are not on the board itself. This is just for convenience sake. The right way
> to do this would be from the bsp. As a result, removed sensor device creation
> from nrf52dk and nrf52840pdk bsps.
> - Add sensor interface struct so that it can get used by the individual
> drivers so that we can talk to multiple sensors using the same driver using
> same or different interfaces.
> - For TSL2561, apart from the above changes, various fixes were made:
> 1. Making a few changes to this driver along with sensor_itf changes
> 2. Gain, integration time and enabled flag was a global instead of
> reading every time gain is set or integration time is set. Changing it so
> that we read the timing register before setting either gain or integration
> time.
> 3. Also changing a few functions in the driver to handle errors more
> gracefully. Changing shell to work with new functions.
> 4. Making changes so that it goes in synch with the drivers.
> 5. Removing delays from the read function. the poller should be set
> according to the other config values being set so that the sensor is read at
> the correct interval.
> - Add a mask field to the sensor structure so that differentiating between
> supported Vs configured sensors is possible
> - Change the sensor_mgr_match_bytype() function to check w.r.t mask and make
> it a public API.
> - Adding config variable for mask for different drivers in their respective
> structures.
> - Adding an API to set the sensor types based on the mask.
> - Moving config_sensor() out of the app into the creator package and the bsp.
> Objective is to show how a developer would initialize and configure the
> sensor in the bsp which is normally the practice.
> - Changing sensor_shell command "list" to show only configured types.
> - sensor_shell cmd "type" shows supported types set by the driver.
> - In the sensor_oic source file changing code to check type only while
> allocating resources.
> - Changing device name in the sensors_test app to be "sensy" instead of "pi"
> - For the LSM303DLHC driver the address field in the "sensor_itf" structure
> is not enough since it talks to the sensor using two addresses: one for
> accelerometer and one for magnetometer. The best way to deal with this is to
> make the addresses config variables in the sensor driver structure. For any
> such sensor devices which have multiple addresses, a similar approach should
> be followed.
> - Have changes the drivers to work with all the fixes mentioned above.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
