docs: fault-injection evidence for recovery ladder

This commit is contained in:
Lars Nolden
2026-09-01 11:21:04 +02:00
parent e131687264
commit 70762362a4
2 changed files with 45 additions and 0 deletions
+15
View File
@@ -109,6 +109,21 @@
rtl_sdr probes and disables direct-sampling mode); it is not a wedge rtl_sdr probes and disables direct-sampling mode); it is not a wedge
indicator. The real symptom is railed, gain-independent noise. indicator. The real symptom is railed, gain-independent noise.
## CONFIRMED (failure recovery)
- The recovery ladder works and escalates correctly, verified by fault
injection rather than assumption:
`inject standby` -> heartbeat sees chip_mode 2 -> level 1 restores RX;
`inject sleep` (configuration destroyed) -> heartbeat sees chip_mode 0 and
`config_intact:false` -> level 1 reports "failed" -> level 2 (full radio
reinit) restores RX, confirmed by chip_mode 5 + config intact + plausible
RSSI on the following heartbeats.
- Checking chip mode alone is NOT sufficient to detect a broken receiver: a
radio that lost its configuration can sit in RX looking healthy while
deaf to C1. The firmware therefore reads the sync-word registers back
every heartbeat (`config_intact`), and recovery levels only report success
when both the mode and the configuration verify.
## NEEDS TESTING ## NEEDS TESTING
- Narrower RX bandwidth (156 kHz) vs current 234.3 kHz on weak meters. - Narrower RX bandwidth (156 kHz) vs current 234.3 kHz on weak meters.
+30
View File
@@ -72,3 +72,33 @@ Every run records `SDR sensitivity: full | degraded` from
`hil/sdr_health.py`. Treat cross-run comparisons as invalid unless both `hil/sdr_health.py`. Treat cross-run comparisons as invalid unless both
runs report the same state -- this dongle loses ~17 dB within minutes of runs report the same state -- this dongle loses ~17 dB within minutes of
streaming (docs/HARDWARE.md). streaming (docs/HARDWARE.md).
## Long reliability soak (DUT only, no SDR)
```
nix develop --command ./hil/soak.sh 9000 # seconds
```
Measures firmware-side reliability over hours: repeat success rate, timing
stability, radio recoveries, config-integrity failures and MCU resets, and
reports progress against the milestone targets. Deliberately does not use
the SDR, which cannot stream for hours without faulting.
## Fault injection (verifying the recovery ladder)
```
./result-py/bin/python3 hil/dut_cmd.py <port> --listen 26 "inject standby"
./result-py/bin/python3 hil/dut_cmd.py <port> --listen 26 "inject sleep"
```
`standby` leaves the radio out of RX with its configuration intact;
`sleep` additionally destroys the configuration. The heartbeat probe must
detect both and the recovery ladder must restore RX. Verified behaviour:
```
inject standby -> heartbeat chip_mode=2, config_intact=true
-> radio_recovery level 1 -> rx_restored
inject sleep -> heartbeat chip_mode=0, config_intact=false
-> radio_recovery level 1 -> failed (correctly refuses)
-> radio_recovery level 2 -> rx_restored
```