Negative control: EFR32MG13 RTCC as a zero-answering stub
Same firmware (firmware-efr32mg13-smoke) and smoke script as the passing run,
but with rtcc typed stub instead of efm32s1_rtcc. Base, size and IRQ (31)
are unchanged. A stub is what the window would be without a model: it
answers every read with zero.
The Renode type name efr32_rtcc was tried first. The engine refuses it at
load time ("unknown peripheral type 'efr32_rtcc' ... Refusing to answer it with
a zero-filled stub"), so a stub has to be declared explicitly.
The smoke FAILS 3/5. Every check up to wtim0 passes. The RTCC check never
completes: CNT stays at zero, so OF and CC1 never set, and each bounded wait
spins until the 200000-step budget is spent. Neither the rtcc ... ok line nor
EFR32MG13 PASS is printed, and the summary word at 0x2000FC00 is never
written (expected 0xEF3107FF). See run_smoke.log and
simulation-output/uart.log.