Repository navigation
zephyr-cp: stm32wba65i_dk1 overlay has the same missing "ranges" defect as the RP2040 boards #7
Description
Found while fixing #6, deliberately not patched because it is outside that scope and unverified.
ports/zephyr-cp/boards/stm32wba65i_dk1.overlay does /delete-node/ partitions; and rebuilds the node without ranges;, while zephyr/boards/st/stm32wba65i_dk1/stm32wba65i_dk1.dts supplies one. Its partitions use zephyr,mapped-partition, and flash sits at 0x08000000, so by the same mechanism as #6 the code partition should resolve to an untranslated offset and the image be mis-linked.
Not confirmed. I have no STM32WBA hardware and did not build it. The right next step is to build the board and compare rom_start VMA/LMA before and after adding ranges;, exactly as was done for the three RP2040 boards.
For completeness, overlays that share the /delete-node/ partitions; pattern but are harmless: the nRF52840, nRF5340 and native_sim overlays. Their flash base is 0x0, so the untranslated offset coincidentally equals the absolute address.
Activity
tyeth commentedon Sep 8, 2026
Confirmed and fixed upstream in the same commit as #6.
adafruit/circuitpythonffd62e1c59addsranges;toports/zephyr-cp/boards/st/stm32wba65i_dk1/board.overlay(line 5 onmain121489fe70). Its commit message independently confirms the mechanism this issue predicted: withoutranges;the partitions resolved to0x0,0x10000,0x108000and0x1e0000instead of0x08000000upwards, and because flash_map computes a mapped partition's offset asDT_REG_ADDR(partition) - DT_REG_ADDR(memory), everyflash_areaoffset on the board underflowed.🤖 Generated with Claude Code