zephyr/boards/lairdconnect/bl654_dvk/board.cmake

7 lines
261 B
CMake
Raw Normal View History

# SPDX-License-Identifier: Apache-2.0
runners: nrfjprog: boilerplate and recover rework Rework the runner to improve various issues. Every board.cmake file for an nRF SoC target is repeating boilerplate needed for the nrfjprog runner's --nrf-family argument. The information we need to decide the --nrf-family is already available in Kconfig, so just get it from there instead. Keep the --nrf-family argument around for compatibility, though. This cuts boilerplate burden for board maintainers. We also need to revisit how this runner handles recovery to fix it in nRF53 and keep things consistent everywhere else. To cleanly handle additional readback protection features in nRF53, add a --recover option that does an 'nrfjprog --recover' before flashing. Keep the behavior consistent across SoCs by supporting it on those too. Because this is expected to be a bit tricky for users to understand, check if a --recover is needed if the 'nrfjprog --program' fails because of protection, and tell the user how to fix it. Finally, instead of performing a separate 'nrfjprog --eraseall', just give --chiperase to 'nrfjprog --program' process's arguments instead of --sectorerase. This is cleaner, resulting in fewer subprocesses and avoiding an extra chip reset. Having a separate 'west flash --recover' option doubles the number of test cases if we want to keep exhaustively enumerating them. That doesn't feel worthwhile, so update the test cases by picking a representative subset of the possibilities. Each test now has enough state that it's worth wrapping it up in a named tuple for readability. Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
2020-11-24 05:19:23 +08:00
board_runner_args(nrfjprog "--softreset")
board_runner_args(jlink "--device=nRF52840_xxAA" "--speed=4000")
include(${ZEPHYR_BASE}/boards/common/nrfjprog.board.cmake)
include(${ZEPHYR_BASE}/boards/common/jlink.board.cmake)