Drivers: RTC: NXP: introduce Counter-based generic driver #95087
Conversation
zephyrbot
added
area: Counter
platform: NXP
zephyrbot
requested review from
axelnxp,
bjarki-andreasen,
butok,
dbaluta,
decsny,
DerekSnell,
dleach02,
EmilioCBen,
George-Stefan,
JarmouniA,
kartben,
manuargue,
MaochenWang1,
mmahadevan108 and
nashif
Holt-Sun
force-pushed
the
gaps-generic-rtc-using-counter-api
branch
from
2565ef1 to
838fb61
Compare
|
The twister build failure is caused by: cba0742#diff-db8ba35bdefd7627bfa4c2a0d7d13c9d1f20787ddf7181333981f6c01a32cc34 |
@bjarki-andreasen Do you mean the "set_calibration", "get_calibration"? There are no such corresponding APIs in counter driver model. All other APIs are using counter APIs. Or I miss understand anything? If so, please be more specific. |
Yes, fsl anything should not be in a generic counter to rtc wrapper. Keyword here is generic. If the counter API is missing a feature, like calibration it would seem, that should be added to the counter API first, then the generic wrapper can wrap it. |
Holt-Sun
force-pushed
the
gaps-generic-rtc-using-counter-api
branch
from
838fb61 to
b06d820
Compare
Holt-Sun
force-pushed
the
gaps-generic-rtc-using-counter-api
branch
from
90f26ec to
79a8187
Compare
|
@sylvioalves Your input would be useful here I think since ESP32 can also use this new driver to provide Zephyr RTC API, following the rejection of #92253 |
Holt-Sun
force-pushed
the
gaps-generic-rtc-using-counter-api
branch
from
79a8187 to
cb69f1d
Compare
|
Need to fix the compliance check here |
Holt-Sun
force-pushed
the
gaps-generic-rtc-using-counter-api
branch
from
cb69f1d to
57cc2e5
Compare
JarmouniA
changed the title
Add new zephyr,counter-rtc.yaml for virtual RTC device wrapping counter APIs Signed-off-by: Holt Sun <[email protected]>
Fix some build warning. Signed-off-by: Holt Sun <[email protected]>
Holt-Sun
force-pushed
the
gaps-generic-rtc-using-counter-api
branch
from
57cc2e5 to
a381a40
Compare
nandojve
changed the title
Counter RTC driver is written using the counter API as a target instead of a hardware. It can adapt to all RTC device without broken-down time register but with counter. Signed-off-by: Holt Sun <[email protected]>
Add counter_rtc in nxp device and board. Signed-off-by: Holt Sun <[email protected]>
Add confs for nxp boards. Signed-off-by: Holt Sun <[email protected]>
Add rtc test confs for nxp boards. Signed-off-by: Holt Sun <[email protected]>
Holt-Sun
force-pushed
the
gaps-generic-rtc-using-counter-api
branch
from
a381a40 to
9929a45
Compare
|
| return -EINVAL; | ||
| } | ||
|
|
||
| const struct counter_rtc_config *config = dev->config; |
There was a problem hiding this comment.
For the sake of code organization, would you mind moving all declarations/variables (specially all those struct) to the beginning of the function?
| static int counter_rtc_alarm_get_supported_fields(const struct device *dev, uint16_t id, | ||
| uint16_t *mask) |
There was a problem hiding this comment.
Why don't you map the id to an alarm channel? For example NXP's SNVS RTC (Real Time Counter) has 2 channels, HP (High Power) and LP (Low Power)
Create generic NXP RTC driver using Zephyr RTC driver model. The RTC driver invokes counter APIs. It can works with existing RTC counter drivers counter_mcux_rtc.c and counter_mcux_lpc_rtc.c.