feat: Add PROFILE command — heap, stack watermarks, CPU runtime stats

- PROFILE command returns heap usage (free/min/dram/iram),
  per-task stack high watermark, and per-task CPU % (when
  CONFIG_FREERTOS_GENERATE_RUN_TIME_STATS is enabled)
- Enable FreeRTOS runtime stats in sdkconfig.defaults
- Enlarge cmd reply buffer to 1400 bytes for multi-line output
- Add esp_heap_caps.h include

get-started/csi_recv_router/main/app_main.c

@@ -31,6 +31,7 @@
 #include "esp_now.h"
 #include "esp_timer.h"
 #include "esp_task_wdt.h"
+#include "esp_heap_caps.h"
 #include "esp_ota_ops.h"
 #include "esp_https_ota.h"
 #include "esp_http_client.h"
@@ -848,6 +849,55 @@ static int cmd_handle(const char *cmd, char *reply, size_t reply_size)
         return strlen(reply);
     }
 
+    /* PROFILE */
+    if (strcmp(cmd, "PROFILE") == 0) {
+        int pos = 0;
+        /* Heap info */
+        size_t free_heap = esp_get_free_heap_size();
+        size_t min_heap = esp_get_minimum_free_heap_size();
+        size_t free_dram = heap_caps_get_free_size(MALLOC_CAP_8BIT);
+        size_t total_dram = heap_caps_get_total_size(MALLOC_CAP_8BIT);
+        size_t free_iram = heap_caps_get_free_size(MALLOC_CAP_IRAM_8BIT);
+        pos += snprintf(reply + pos, reply_size - pos,
+                        "OK PROFILE\nHEAP free=%u min=%u dram=%u/%u iram=%u\n",
+                        (unsigned)free_heap, (unsigned)min_heap,
+                        (unsigned)free_dram, (unsigned)total_dram, (unsigned)free_iram);
+
+        /* Per-task stack watermarks */
+        pos += snprintf(reply + pos, reply_size - pos, "TASKS\n");
+        const char *task_names[] = {"led_task", "cmd_task", "adaptive", "ble_host", "main", NULL};
+        for (int i = 0; task_names[i] != NULL && pos < (int)reply_size - 60; i++) {
+            TaskHandle_t th = xTaskGetHandle(task_names[i]);
+            if (th) {
+                UBaseType_t hwm = uxTaskGetStackHighWaterMark(th);
+                pos += snprintf(reply + pos, reply_size - pos,
+                                "  %-12s stack_free=%u\n", task_names[i], (unsigned)(hwm * sizeof(StackType_t)));
+            }
+        }
+
+#if defined(CONFIG_FREERTOS_GENERATE_RUN_TIME_STATS) && CONFIG_FREERTOS_GENERATE_RUN_TIME_STATS
+        /* CPU runtime stats */
+        UBaseType_t n = uxTaskGetNumberOfTasks();
+        TaskStatus_t *tasks = malloc(n * sizeof(TaskStatus_t));
+        if (tasks) {
+            uint32_t total_time;
+            n = uxTaskGetSystemState(tasks, n, &total_time);
+            if (total_time > 0) {
+                pos += snprintf(reply + pos, reply_size - pos, "CPU\n");
+                for (UBaseType_t i = 0; i < n && pos < (int)reply_size - 60; i++) {
+                    uint32_t pct = (tasks[i].ulRunTimeCounter * 100) / total_time;
+                    if (pct > 0 || tasks[i].ulRunTimeCounter > 0) {
+                        pos += snprintf(reply + pos, reply_size - pos,
+                                        "  %-12s %3lu%%\n", tasks[i].pcTaskName, (unsigned long)pct);
+                    }
+                }
+            }
+            free(tasks);
+        }
+#endif
+        return pos;
+    }
+
     /* OTA <url> */
     if (strncmp(cmd, "OTA ", 4) == 0) {
         const char *url = cmd + 4;
@@ -899,7 +949,7 @@ static void cmd_task(void *arg)
     ESP_LOGI(TAG, "Command listener on UDP port %d", CONFIG_CSI_CMD_PORT);
 
     char rx_buf[128];
-    char reply_buf[256];
+    char reply_buf[1400];
     struct sockaddr_in src_addr;
     socklen_t src_len;
 

get-started/csi_recv_router/sdkconfig.defaults

@@ -35,6 +35,8 @@ CONFIG_COMPILER_OPTIMIZATION_PERF=y
 # FreeRTOS
 #
 CONFIG_FREERTOS_HZ=1000
+CONFIG_FREERTOS_GENERATE_RUN_TIME_STATS=y
+CONFIG_FREERTOS_USE_STATS_FORMATTING_FUNCTIONS=y
 
 #
 # ESP32-specific