NOTE

First full run of the extended smoke (2026-11-09) with the final firmware, before the engine had an STM32F0 SYSCFG model:

  • exti pads ... pf1=0 pc13=0 FAIL: a syscfg entry with no profile is built as a read-zero stub, so EXTICR1..4 did not exist. Both buttons toggled their pads (presses=1), but no edge reached EXTI lines 1/13 or NVIC 5/7. Fixed by Stm32F0Syscfg (profile: stm32f0, A..F = 0..5).

Everything else already passed: summary 0x072FDFFF (bit 13 missing). Earlier probe runs (not kept) found two firmware mistakes, not engine bugs: AHBENR enabled IOPA..IOPE only (the GPIOF clock gate correctly ignored PF writes), and the second TMP102 read expected 0x1900 although configs/devices/tmp102.yaml documents +0.5 °C self-drift per read.