Microbit v2 specific features #763
Conversation
-Added ROM setting to enable/disable feature via .cfg files. -Implemented hook at flash_decoder layer once the first 12 vectors are obtained (flash decoder min size)
-hook for custom reset button handling -hook for 30ms event tick -hook for power down event -usb state udpate
-Add KL27Z NXP drivers -Power monitor (detect power source) -Board Rev Id detection -i2c support -low power support -button and LED handling
-build filesystem hook for error blob app and DATA.BIN file creation -Added microbit custom BL project for custom URL in HELP_FAQ.HTM -I2C comms and flash interface -SWD clock speed improvements
-Make power LED remain active in VLPS mode. Changed FlexIO source clock to one that is active in VLPS. -Removed NRF Power mode property -Added Power LED sleep state property -Changed size of properties 0x01-03 to 2B
…le. Add documentation on the feautre.
…, .txt, .csv, .htm, .wav).
-Add property for Kl27 user event (wakeup or long button press) -Blink LED when attempting to enter deep sleep but are PC connected -Hold interrupt line to send I2C message on reset long press -Clear I2C responses after first read in the comms interface
-Change wake up when the reset button is pressed instead of released -Add hook in bootloader project to bypass reset button check to jump directly to the interface -Reset the nRF whenever we wakeup from the reset button independently of the press duration
-Keep red LED ON after a long press. -When the reset button is pressed, the KL27 will reset the `Power LED Sleep state` value to the default ON -Red LED in deep sleep mode now depends on the `Power LED Sleep state` it can be either HIGH or LOW but no PWM.
…ew event so the nrf can see the new edges.
…this command doesn't assert the interrupt line.
…SENSOR_INT caused by WAKE_ON_EDGE
…e.g attaching usb device to VM)
… of waiting to the main loop poll to update the status.
… getting PWR_SOURCE_NONE.
-Add support for reading flash config data -Change filesize to bytes instead of KB -Fix address ranges in erase storage cmd -Add more validations to different commands
- Set backdoor key to 'MICROBIT'. - Enable FPROT for first 32 kB of flash. - Fix FPROTx numbering according to KL27 RM, where FPROT3 is at address 0x408. - Enable backdoor key in FSEC. - Added asterisks next to selected options in comments for FOPT and FSEC.
|
Thanks for the PR! I probably won't be able to get to it until this weekend at the earliest. |
# Conflicts: # records/tools/version.yaml Resolved using DAPLink version 256 from master
update formatting made microbit specific changes to be board configurable via yaml files
|
|
||
| if (sector_offset != 0) { | ||
| if (sector_offset > 1) { |
There was a problem hiding this comment.
This should probably depend on the value of BOARD_EXTRA_BUFFER
There was a problem hiding this comment.
Updated calculation.
Consider BOARD_EXTRA_BUFFER in sector offset calculation Follow ADIv6 for powering down the debug interface Add comment to .yaml file on DELAY_FAST_CYCLES Don't erase target image if incompatible image is enabled and triggered.
gerargz
force-pushed
the
microbit_features
branch
from
90b6221 to
44a2522
Compare
| // TODO: put the interface chip in sleep mode | ||
| } |
There was a problem hiding this comment.
Just noticed that the while(1); from main() is missing here.
It's probably ok if the main loop keeps running, and might actually be better: RTX will block on the osThreadFlagsWait() call and enter the idle thread which will WFI.
But has this been tested on a build without the micro:bit override of this hook? (Is there even a way to enter powerdown for "normal" builds?)
There was a problem hiding this comment.
If I remember correctly I initially had the while(1); in the weak function, but this had strange behavior where the code wasn't executing correctly in the "strong" custom function. I think it had to do with how the compiler optimizations were working. Something about the optimizations being done per file, without knowing that there was a "strong" function definition elsewhere. That and the combination of having an endless loop in the weak function, made the optimizations assume wrong things.
I don't see any build that uses the main_powerdown_event() to trigger this event, so the while(1) was probably a placeholder and there's no impact by removing it, I think.
There was a problem hiding this comment.
Ok, cool. Nothing to worry about then. Thanks for the quick answer!
There was a problem hiding this comment.
Approved. Really fantastic work @gerargz!
The only outstanding question is the new one about the missing while(1); from the powerdown hook. Ideally I'd like to get an answer before this is actually merged, but it's not a big deal either way.
|
All questions addressed, ready for merge! 🚀 |
Draft PR with microbit v2 specific features. This contains the commit ID for the DAPLink build used by the factory release of the microbit v2 boards.
This PR is still a draft as it contains modifications to the DAPLink core that may not be upstream friendly.