Auto-recover a wedged radio, fix SMS send timeout (v0.9.0) - #22
Merged
Conversation
A SIM800C that the network de-registers can settle into AT+CREG? state 0 (not registered, not searching) and stay there indefinitely while still reporting a healthy signal. AT+COPS=0 answers ERROR in that state; only a radio power-cycle re-attaches the module. Observed in the wild: six days unregistered, noticed only when an SMS failed. The background monitor now re-reads registration every minute and cycles the radio (AT+CFUN=0/1) once it has been down for 10 minutes, waiting up to 90s for re-registration and re-running initialize() afterwards, since AT+CFUN resets SMS text mode and caller-ID reporting. Cycles are spaced 30 minutes apart and skipped during a call. send_sms makes one recovery attempt of its own before giving up. Also fixes the send path that turned this into a dead modem: the wait for +CMGS was 15s against a network that needs 16-25s for a multi-part UCS2 message, and on timeout the modem stayed in text-entry mode, swallowing every later AT command. The timeout is now 60s and any failure inside the transaction sends ESC to leave text-entry mode. Claude-Session: https://claude.ai/code/session_01M4WjTTexBxNnym1Gq79pLV
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The problem
A SIM800C that the network de-registers can settle into
AT+CREG?state0— not registered, not searching — and stay there indefinitely, while the signal sensor keeps reporting a healthy value.AT+COPS=0answersERRORin that state; only a radio power-cycle re-attaches the module. Seen in the wild: six days unregistered, noticed only when a scheduled SMS failed withNotRegistered.Recovering by hand then exposed a second bug: the wait for
+CMGSwas 15s, but the network needs 16–25s to accept a multi-part UCS2 (Cyrillic) message. On timeout the modem stayed in text-entry mode and swallowed every subsequent AT command, so all three retries — and the call poll with them — failed against a modem that looked completely dead.Changes
Modem.reset_radio()—AT+CFUN=0→ settle →AT+CFUN=1, then pollAT+CREG?for up to 90s. Re-runsinitialize()on success, becauseAT+CFUNresets SMS text mode and caller-ID reporting.ModemHub.async_check_registration()— registration watchdog, run every minute from the existing monitor loop; triggers recovery after 10 minutes down.ModemHub.async_recover_registration()— guarded recovery: single-flight lock, 30-minute cooldown, skipped while a call is in progress.sim800c.send_smsmakes one recovery attempt onNotRegisteredand retries the message once, instead of surfacing the error immediately.+CMGStimeout 15s → 60s, and any failure inside the transaction sendsESCso the modem leaves text-entry mode.sensor.sim800c_networknow refreshes once a minute as a side effect (previously only on the 5-minute sensor poll).Testing
99 passed(9 new),ruff checkandruff format --checkclean. New tests cover the CFUN sequence and its re-initialization, the re-registration timeout, the watchdog's grace period / cooldown / call guard, and both send-path outcomes. The manual CFUN cycle and the 16–23s+CMGStimings were verified against real hardware (SIM800 R14.18, MegaFon).https://claude.ai/code/session_01M4WjTTexBxNnym1Gq79pLV