Replay: the ZG222 GPIO-less negative control, after the ARMv6-M fault fix
Same firmware (firmware-efm32zg222-smoke) and smoke assertions as
2026-11-08-negative-control-no-gpio-timer, against a temporary copy of the
current configs/chips/onboarding/efm32zg222.yaml with only USART1 kept
(efm32zg222-negctl.yaml here; flash now at 0x0). The firmware still faults on
its first GPIO access (0x40006050) and the smoke still FAILS 2/5.
What changed is the report. Before, on this Cortex-M0+:
FAULT HardFault escalated from a precise BusFault: data access at 0x4000_6050 (BFAR valid), ...
HFSR=0x40000000 CFSR=0x00008200 [FORCED PRECISERR] BFAR=0x40006050
After:
FAULT HardFault: data access at 0x4000_6050, at PC 0x0000_01CC in `core::ptr::read_volatile` (mod.rs:2091), called from `firmware_efm32zg222_smoke::__cortex_m_rt_main`.
ARMv6-M (no HFSR/CFSR); engine-recorded cause [PRECISERR] address=0x40006050
ARMv6-M has one fault exception and no CFSR/HFSR/MMFAR/BFAR (ARMv6-M ARM B3.2.2), so nothing "escalates" and there is no BFAR to name. The engine still records the cause; the report now says that is what it is. Firmware on an M0/M0+ now also reads zero from those reserved offsets, and SHCSR keeps only SVCALLPENDED.
Run with the debug CLI from examples/renode-onboarding
(labwired test --script efm32zg222-negctl-smoke.yaml, system file pointing at
the chip copy above), 2026-11-08.