mybeacon-daemon/cmd/beacon-daemon/network_manager.go

// Network Manager - manages all network interfaces with priority-based logic
package main

import (
	"fmt"
	"log"
	"os"
	"os/exec"
	"strings"
	"syscall"
	"time"
)

const (
	eth0Interface   = "eth0"
	wlan0Interface  = "wlan0"
	apFallbackDelay = 120 * time.Second // 120 sec without connection
)

// NetworkState represents current network state
type NetworkState int

const (
	StateNoNetwork NetworkState = iota
	StateEth0Online
	StateWLAN0ClientOnline
	StateWLAN0APFallback
)

// NetworkManager manages all network interfaces with priority-based logic
// Priority 1: eth0 (carrier detect)
// Priority 2: wlan0 client (if configured)
// Fallback: wlan0 AP (120 sec without connection AND without eth0 carrier)
type NetworkManager struct {
	cfg      *Config
	debug    bool
	stopChan chan struct{}

	// Scanners reference (needed to stop WiFi scanner when using wlan0)
	scanners *ScannerManager

	// State tracking
	eth0Carrier     bool
	eth0HasIP       bool
	eth0DhcpRunning bool
	wlan0HasIP      bool
	currentState    NetworkState
	lastOnlineTime  time.Time
	apRunning       bool
	eth0DhcpPid     int
}

// NewNetworkManager creates a new network manager
func NewNetworkManager(cfg *Config, scanners *ScannerManager) *NetworkManager {
	return &NetworkManager{
		cfg:            cfg,
		debug:          cfg.Debug,
		stopChan:       make(chan struct{}),
		scanners:       scanners,
		currentState:   StateNoNetwork,
		lastOnlineTime: time.Now(), // Start timer immediately
	}
}

// UpdateConfig updates network manager configuration
func (nm *NetworkManager) UpdateConfig(cfg *Config) {
	nm.cfg = cfg
}

// HasIP returns true if any interface has an IP address
func (nm *NetworkManager) HasIP() bool {
	return nm.eth0HasIP || nm.wlan0HasIP
}

// HasCarrier returns true if eth0 has carrier
func (nm *NetworkManager) HasCarrier() bool {
	return nm.eth0Carrier
}

// IsOnline returns true if device is online (eth0 or wlan0 client)
func (nm *NetworkManager) IsOnline() bool {
	return nm.currentState == StateEth0Online || nm.currentState == StateWLAN0ClientOnline
}

// Start begins network management
func (nm *NetworkManager) Start() {
	log.Println("[netmgr] Starting network manager...")
	log.Println("[netmgr] Priority: eth0 → wlan0 client → wlan0 AP fallback (120s)")

	// Start main monitoring loop
	go nm.monitorLoop()
}

// Stop stops the network manager
func (nm *NetworkManager) Stop() {
	log.Println("[netmgr] Stopping network manager...")
	close(nm.stopChan)

	// Clean up WiFi interfaces only (leave eth0 running for SSH access)
	nm.stopWLAN0Client()
	nm.stopAP()

	// Note: eth0 is NOT stopped to maintain SSH connectivity
	// Only stop eth0 DHCP process, but keep the IP address
	if nm.eth0DhcpPid > 0 {
		syscall.Kill(nm.eth0DhcpPid, syscall.SIGTERM)
	}
	exec.Command("killall", "-q", "udhcpc").Run()
}

// monitorLoop is the main network management loop
// Polls every second and applies priority-based logic
func (nm *NetworkManager) monitorLoop() {
	ticker := time.NewTicker(1 * time.Second)
	defer ticker.Stop()

	for {
		select {
		case <-nm.stopChan:
			return
		case <-ticker.C:
			nm.tick()
		}
	}
}

// tick performs one iteration of network state check and management
func (nm *NetworkManager) tick() {
	// Update state
	nm.eth0Carrier = nm.checkCarrier(eth0Interface)
	nm.eth0HasIP = nm.hasIP(eth0Interface)
	nm.wlan0HasIP = nm.hasIP(wlan0Interface)

	online := nm.eth0HasIP || nm.wlan0HasIP
	if online {
		nm.lastOnlineTime = time.Now()
	}

	// Priority 1: eth0 (independent, always configure if carrier UP)
	if nm.eth0Carrier {
		if !nm.eth0HasIP && !nm.eth0DhcpRunning {
			// Carrier UP but no IP and DHCP not running - configure network
			if nm.cfg.Network.Eth0.Static {
				log.Println("[netmgr] eth0 carrier UP - configuring static IP")
				nm.configureEth0Static()
			} else {
				log.Println("[netmgr] eth0 carrier UP - starting DHCP")
				nm.startEth0DHCP()
			}
		}
	} else {
		// No eth0 carrier - stop eth0 DHCP/IP
		if nm.eth0HasIP || nm.eth0DhcpRunning {
			log.Println("[netmgr] eth0 carrier DOWN - stopping network")
			nm.stopEth0()
		}
	}

	// Priority 2: wlan0 client (INDEPENDENT of eth0, always connect if configured)
	if nm.cfg.WiFi.ClientEnabled && nm.cfg.WiFi.SSID != "" {
		if !nm.wlan0HasIP {
			// Should connect
			log.Printf("[netmgr] Connecting wlan0 client to %s...", nm.cfg.WiFi.SSID)
			nm.stopAP() // Stop AP before connecting client
			if nm.scanners.IsWiFiRunning() {
				log.Println("[netmgr] Stopping WiFi scanner before connecting client")
				nm.scanners.StopWiFi()
			}
			if err := nm.connectWLAN0Client(); err != nil {
				log.Printf("[netmgr] wlan0 client connection failed: %v", err)
				// Will fallback to AP after 120s
			}
		}

		if nm.wlan0HasIP {
			if nm.currentState != StateWLAN0ClientOnline {
				log.Println("[netmgr] wlan0 client online")
				nm.currentState = StateWLAN0ClientOnline
			}
			nm.stopAP() // Ensure AP is stopped
			// Stop WiFi scanner (can't run with client)
			if nm.scanners.IsWiFiRunning() {
				log.Println("[netmgr] Stopping WiFi scanner (wlan0 client active)")
				nm.scanners.StopWiFi()
			}
		}
		return // wlan0 client configured - don't start scanner or AP
	} else {
		// wlan0 client not configured - stop it if running
		if nm.wlan0HasIP {
			log.Println("[netmgr] wlan0 client disabled - disconnecting")
			nm.stopWLAN0Client()
		}
	}

	// WiFi Scanner: ONLY if wlan0 is free (no client, no AP) AND monitor enabled
	if nm.cfg.WiFi.MonitorEnabled && !nm.apRunning {
		if !nm.scanners.IsWiFiRunning() {
			log.Println("[netmgr] Starting WiFi scanner (wlan0 free)")
			nm.scanners.StartWiFi(nm.cfg.ZMQAddrWiFi, nm.cfg.WiFiIface)
		}
	} else {
		if nm.scanners.IsWiFiRunning() {
			log.Println("[netmgr] Stopping WiFi scanner (wlan0 busy or monitor disabled)")
			nm.scanners.StopWiFi()
		}
	}

	// Determine current state
	if nm.eth0HasIP {
		nm.currentState = StateEth0Online
	} else if nm.wlan0HasIP {
		nm.currentState = StateWLAN0ClientOnline
	} else {
		nm.currentState = StateNoNetwork
	}

	// Fallback: wlan0 AP (if offline for 120 seconds)
	timeSinceOnline := time.Since(nm.lastOnlineTime)
	if !online && timeSinceOnline >= apFallbackDelay {
		if !nm.apRunning {
			log.Printf("[netmgr] No network for %v - starting AP fallback", timeSinceOnline)
			// Stop scanner before starting AP
			if nm.scanners.IsWiFiRunning() {
				nm.scanners.StopWiFi()
			}
			nm.startAP()
		}
		// Keep trying wlan0 client even with AP running (if configured)
		if nm.cfg.WiFi.ClientEnabled && nm.cfg.WiFi.SSID != "" {
			// Try to reconnect every 30 seconds
			if int(timeSinceOnline.Seconds())%30 == 0 {
				log.Println("[netmgr] AP running - retrying wlan0 client...")
				if err := nm.connectWLAN0Client(); err != nil {
					if nm.debug {
						log.Printf("[netmgr] wlan0 client retry failed: %v", err)
					}
				} else if nm.hasIP(wlan0Interface) {
					log.Println("[netmgr] wlan0 client connected - stopping AP")
					nm.stopAP()
					nm.currentState = StateWLAN0ClientOnline
				}
			}
		}
	} else {
		// Online - stop AP if running
		if nm.apRunning {
			log.Println("[netmgr] Network online - stopping AP fallback")
			nm.stopAP()
		}
	}
}

// =======================================================================================
// eth0 management
// =======================================================================================

func (nm *NetworkManager) startEth0DHCP() {
	// Stop any existing DHCP first
	nm.stopEth0DHCP()

	// Bring interface up (don't flush IP - udhcpc will handle it)
	exec.Command("ip", "link", "set", eth0Interface, "up").Run()

	// Start udhcpc
	pidFile := fmt.Sprintf("/var/run/udhcpc.%s.pid", eth0Interface)
	cmd := exec.Command("udhcpc", "-i", eth0Interface, "-b", "-p", pidFile)
	if err := cmd.Start(); err != nil {
		log.Printf("[netmgr] Failed to start DHCP on eth0: %v", err)
		return
	}

	nm.eth0DhcpRunning = true
	time.Sleep(500 * time.Millisecond)

	// Read PID
	pidData, err := os.ReadFile(pidFile)
	if err == nil {
		fmt.Sscanf(string(pidData), "%d", &nm.eth0DhcpPid)
		if nm.debug {
			log.Printf("[netmgr] eth0 DHCP started (PID: %d)", nm.eth0DhcpPid)
		}
	} else {
		log.Printf("[netmgr] Warning: Could not read udhcpc PID file: %v", err)
	}
}

func (nm *NetworkManager) configureEth0Static() {
	nm.stopEth0DHCP()

	// Bring interface up
	exec.Command("ip", "link", "set", eth0Interface, "up").Run()

	// Configure IP
	if nm.cfg.Network.Eth0.Address != "" {
		cmd := exec.Command("ip", "addr", "add", nm.cfg.Network.Eth0.Address, "dev", eth0Interface)
		if err := cmd.Run(); err != nil && !strings.Contains(err.Error(), "exists") {
			log.Printf("[netmgr] Failed to configure IP %s on eth0: %v", nm.cfg.Network.Eth0.Address, err)
			return
		}
		log.Printf("[netmgr] Configured static IP %s on eth0", nm.cfg.Network.Eth0.Address)
	}

	// Configure gateway
	if nm.cfg.Network.Eth0.Gateway != "" {
		exec.Command("ip", "route", "del", "default").Run()
		if err := exec.Command("ip", "route", "add", "default", "via", nm.cfg.Network.Eth0.Gateway).Run(); err != nil {
			log.Printf("[netmgr] Failed to configure gateway %s: %v", nm.cfg.Network.Eth0.Gateway, err)
		} else {
			log.Printf("[netmgr] Configured gateway %s", nm.cfg.Network.Eth0.Gateway)
		}
	}

	// Configure DNS
	if nm.cfg.Network.Eth0.DNS != "" {
		dnsContent := fmt.Sprintf("nameserver %s\n", nm.cfg.Network.Eth0.DNS)
		os.WriteFile("/etc/resolv.conf", []byte(dnsContent), 0644)
	}
}

func (nm *NetworkManager) stopEth0DHCP() {
	pidFile := fmt.Sprintf("/var/run/udhcpc.%s.pid", eth0Interface)
	pidData, err := os.ReadFile(pidFile)
	if err == nil {
		var pid int
		if _, err := fmt.Sscanf(string(pidData), "%d", &pid); err == nil {
			syscall.Kill(pid, syscall.SIGTERM)
			os.Remove(pidFile)
		}
	}
	exec.Command("killall", "-q", "udhcpc").Run()
	nm.eth0DhcpPid = 0
	nm.eth0DhcpRunning = false
}

func (nm *NetworkManager) stopEth0() {
	nm.stopEth0DHCP()
	exec.Command("ip", "addr", "flush", "dev", eth0Interface).Run()
}

// =======================================================================================
// wlan0 client management
// =======================================================================================

func (nm *NetworkManager) connectWLAN0Client() error {
	// Stop WiFi scanner (can't run both)
	if nm.scanners.IsWiFiRunning() {
		log.Println("[netmgr] Stopping WiFi scanner before connecting client")
		nm.scanners.StopWiFi()
	}

	// Temporarily stop BLE scanner during WiFi connection (AIC8800 combo chip issue)
	bleWasRunning := nm.scanners.IsBLERunning()
	if bleWasRunning {
		log.Println("[netmgr] Temporarily stopping BLE scanner for WiFi connection")
		nm.scanners.StopBLE()
		defer func() {
			if bleWasRunning && nm.cfg.BLE.Enabled {
				log.Println("[netmgr] Restarting BLE scanner")
				nm.scanners.StartBLE(nm.cfg.ZMQAddrBLE)
			}
		}()
		time.Sleep(500 * time.Millisecond)
	}

	// Use wifi-connect.sh script
	scriptPath := "/opt/mybeacon/bin/wifi-connect.sh"
	cmd := exec.Command(scriptPath, nm.cfg.WiFi.SSID, nm.cfg.WiFi.PSK)
	output, err := cmd.CombinedOutput()
	if err != nil {
		return fmt.Errorf("wifi-connect.sh failed: %w (output: %s)", err, string(output))
	}

	return nil
}

func (nm *NetworkManager) stopWLAN0Client() {
	exec.Command("killall", "wpa_supplicant", "udhcpc", "dhcpcd").Run()
	exec.Command("ip", "addr", "flush", "dev", wlan0Interface).Run()
	os.RemoveAll("/var/run/wpa_supplicant/wlan0")
}

// =======================================================================================
// wlan0 AP fallback management
// =======================================================================================

func (nm *NetworkManager) startAP() {
	if nm.apRunning {
		return
	}

	log.Println("[netmgr] Starting AP fallback mode...")

	// Stop client if running
	nm.stopWLAN0Client()

	// Stop WiFi scanner
	if nm.scanners.IsWiFiRunning() {
		nm.scanners.StopWiFi()
	}

	// Create ap0 interface
	exec.Command("iw", "phy", "phy0", "interface", "add", "ap0", "type", "__ap").Run()
	time.Sleep(500 * time.Millisecond)

	// Configure AP IP
	exec.Command("ifconfig", "ap0", "192.168.4.1", "netmask", "255.255.255.0", "up").Run()

	// Start hostapd
	exec.Command("hostapd", "-B", "/etc/hostapd/hostapd.conf").Start()

	// Start dnsmasq for DHCP
	exec.Command("dnsmasq", "-C", "/etc/dnsmasq.d/ap0.conf").Start()

	nm.apRunning = true
	nm.currentState = StateWLAN0APFallback
	log.Println("[netmgr] AP fallback started: SSID=Luckfox_Setup, IP=192.168.4.1")
}

func (nm *NetworkManager) stopAP() {
	if !nm.apRunning {
		return
	}

	log.Println("[netmgr] Stopping AP fallback...")

	exec.Command("killall", "hostapd", "dnsmasq").Run()
	exec.Command("ip", "link", "del", "ap0").Run()

	nm.apRunning = false
}

// =======================================================================================
// Helper functions
// =======================================================================================

func (nm *NetworkManager) checkCarrier(iface string) bool {
	carrierPath := fmt.Sprintf("/sys/class/net/%s/carrier", iface)
	data, err := os.ReadFile(carrierPath)
	if err != nil {
		return false
	}
	return strings.TrimSpace(string(data)) == "1"
}

func (nm *NetworkManager) hasIP(iface string) bool {
	cmd := exec.Command("ip", "addr", "show", iface)
	output, err := cmd.Output()
	if err != nil {
		return false
	}
	return strings.Contains(string(output), "inet ")
}