dangernoodle-io / TaipanMiner / 24139865449

90%

push

github

fix: race conditions, concurrency bugs, and mDNS unresponsiveness (#73)

- replace vTaskSuspend with mining_pause in HTTP OTA upload to prevent
  SHA peripheral deadlock
- add mutex guard to mining_pause so only one caller can hold pause
- make s_ota_in_progress check-and-set atomic under spinlock
- protect s_cached_check OTA struct with s_ota_status_mux spinlock
- clear stale s_job on stratum reconnect
- increase stratum line buffer from 2048 to 4096 bytes and gracefully
  handle overflow instead of triggering reconnect loop
- drain result_queue of stale shares on stratum reconnect
- log warning when xQueueSend drops a share (result queue full)
- remove redundant xQueueReset before xQueueOverwrite
- add 10-minute job staleness watchdog to trigger reconnect
- move NVS stats save from esp_timer callback to dedicated low-priority
  task to avoid blocking timer service
- raise CONFIG_MDNS_TASK_PRIORITY from 1 to 5 so mDNS can process
  queries under mining load
- guard xEventGroupSetBits against NULL event group on WiFi reconnect
- check and log mDNS init return values

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

221 of 280 branches covered (78.93%)

Branch coverage included in aggregate %.

681 of 697 relevant lines covered (97.7%)

596617.43 hits per line