seamless crashing

TODO.md

@@ -9,5 +9,7 @@
 		- needs to have a CLI table output
 		- needs to have a GUI output (matplotlib)
 - [ ] Refactor; make sure everything adheres to naming conventions
-- [ ] Bluetooth pairing
 - [ ] Renaming wifi (should reboot the wifi/web comms to take effect)
+- [ ] Make sure external RTC crystal is actually in use
+- [ ] Warn if time is de-synced from client by more than 5 minutes
+- [ ] Bluetooth pairing

main/control_fsm.h

@@ -42,7 +42,8 @@ typedef enum {
 	STATE_CALIBRATE_DRIVE_DELAY,
 	STATE_CALIBRATE_DRIVE_MOVE
 } fsm_state_t;
-#define LOG_TYPE_BAT 100
+#define LOG_TYPE_BAT   100
+#define LOG_TYPE_CRASH 101
 
 typedef enum {
 	RELAY_SENSORS = 0,

main/main.c

@@ -1,4 +1,5 @@
 #include "esp_task_wdt.h"
+#include "esp_system.h"
 #include "i2c.h"
 #include "log_test.h"
 #include "storage.h"
@@ -109,6 +110,7 @@ void app_main(void) {esp_task_wdt_add(NULL);
 	//run_all_log_tests();
     
     if (rtc_xtal_init() != ESP_OK) ESP_LOGE(TAG, "RTC FAILED");
+    rtc_restore_time();  // Restore time from RTC backup if we crashed
 
     // Say hello; turn on the lights
     esp_sleep_wakeup_cause_t cause = rtc_wakeup_cause();
@@ -171,6 +173,21 @@ void app_main(void) {esp_task_wdt_add(NULL);
 	if (adc_init()      != ESP_OK) ESP_LOGE(TAG, "ADC FAILED");
     if (storage_init()  != ESP_OK) ESP_LOGE(TAG, "STORAGE FAILED");
     if (log_init()      != ESP_OK) ESP_LOGE(TAG, "LOG FAILED");
+
+    // Write a crash log entry if we rebooted unexpectedly
+    esp_reset_reason_t reset_reason = esp_reset_reason();
+    if (reset_reason == ESP_RST_PANIC    ||
+        reset_reason == ESP_RST_INT_WDT  ||
+        reset_reason == ESP_RST_TASK_WDT ||
+        reset_reason == ESP_RST_WDT) {
+        ESP_LOGW(TAG, "Crash detected! Reset reason: %d", reset_reason);
+        uint8_t crash_entry[9] = {};
+        uint64_t ts = rtc_get_ms();
+        memcpy(&crash_entry[0], &ts, 8);
+        crash_entry[8] = (uint8_t)reset_reason;
+        log_write(crash_entry, sizeof(crash_entry), LOG_TYPE_CRASH);
+    }
+
     if (solar_run_fsm() != ESP_OK) ESP_LOGE(TAG, "SOLAR FAILED");
     // TODO: Do a 12V check and enter deep sleep if there's a problem
 
@@ -317,6 +334,7 @@ void app_main(void) {esp_task_wdt_add(NULL);
         }
 
         solar_run_fsm();
+        rtc_save_time();   // Keep RTC backup fresh so crashes don't lose time
 
         rtc_check_shutdown_timer();
         esp_task_wdt_reset();

main/rtc.c

@@ -34,9 +34,10 @@
 
 uint64_t last_activity_tick = 0;
 
-// RTC_DATA_ATTR keeps this var in RTC memory; persists across sleeps (but not across boots)
+// RTC_DATA_ATTR keeps these in RTC memory; persists across deep sleep AND software resets (panics, WDT)
 RTC_DATA_ATTR int64_t next_alarm_time_s = -1;
 RTC_DATA_ATTR bool rtc_set = false;
+RTC_DATA_ATTR int64_t rtc_backup_s = 0;  // Crash-safe time snapshot
 bool rtc_is_set() {
 	return rtc_set;
 }
@@ -94,10 +95,25 @@ void rtc_set_s(int64_t tv_sec)
 {
 	rtc_set = true;
     settimeofday(&(struct timeval){.tv_sec = tv_sec, .tv_usec=0}, NULL);
+    rtc_backup_s = tv_sec;
     solar_reset_fsm();
     rtc_schedule_next_alarm();
 }
 
+void rtc_save_time(void)
+{
+    if (rtc_set)
+        rtc_backup_s = rtc_get_s();
+}
+
+void rtc_restore_time(void)
+{
+    if (rtc_set && rtc_backup_s > 0) {
+        settimeofday(&(struct timeval){.tv_sec = rtc_backup_s, .tv_usec = 0}, NULL);
+        ESP_LOGI("RTC", "Time restored from crash backup: %lld", (long long)rtc_backup_s);
+    }
+}
+
 int64_t rtc_get_ms(void)
 {
     struct timeval tv;

main/rtc.h

@@ -43,6 +43,9 @@ int64_t rtc_get_s (void);
 int64_t rtc_get_ms(void);
 void rtc_set_s(int64_t);
 
+void rtc_save_time(void);    // Snapshot current time to RTC memory (call periodically)
+void rtc_restore_time(void); // Restore time from RTC backup after a crash reboot
+
 void    rtc_schedule_next_alarm(void);
 int64_t rtc_get_next_alarm_s();