TCP settings

backend/cmd/config.go

@@ -15,7 +15,7 @@ type Network struct {
 }
 
 type Transport struct {
-	PropagateFault bool
+	PropagateFault bool `toml:"propagate_fault"`
 }
 
 type TFTP struct {
@@ -27,9 +27,18 @@ type TFTP struct {
 }
 
 type Blcu struct {
-	IP              string
-	DownloadOrderId uint16
-	UploadOrderId   uint16
+	IP              string `toml:"ip"`
+	DownloadOrderId uint16 `toml:"download_order_id"`
+	UploadOrderId   uint16 `toml:"upload_order_id"`
+}
+
+type TCP struct {
+	BackoffMinMs       int     `toml:"backoff_min_ms"`        // Minimum backoff duration in milliseconds
+	BackoffMaxMs       int     `toml:"backoff_max_ms"`        // Maximum backoff duration in milliseconds
+	BackoffMultiplier  float64 `toml:"backoff_multiplier"`    // Exponential backoff multiplier
+	MaxRetries         int     `toml:"max_retries"`           // Maximum number of retries before cycling (0 or negative = infinite)
+	ConnectionTimeout  int     `toml:"connection_timeout_ms"` // Connection timeout in milliseconds
+	KeepAlive          int     `toml:"keep_alive_ms"`         // Keep-alive interval in milliseconds
 }
 
 type Config struct {
@@ -39,5 +48,6 @@ type Config struct {
 	Network   Network
 	Transport Transport
 	TFTP      TFTP
+	TCP       TCP
 	Blcu      Blcu
 }

backend/cmd/config.toml

@@ -1,16 +1,56 @@
 # Hyperloop UPV Backend Configuration
 # Configuration file for the H10 Control Station backend server
 
+# <-- CHECKLIST -->
+# 1. Check that all the boards you want to use are declared in the [vehicle] section
+# 2. Set the branch you want to use for the ADJ configuration
+# 3. Toggle the Fault Propagation to your needs (treu/false)
+# 4. Check the TCP configuration and make sure to use the needed Keep Alive settings
+
+
 # Vehicle Configuration
 [vehicle]
 boards = ["HVSCU", "PCU", "BLCU"]
 
+# ADJ (Architecture Description JSON) Configuration
+[adj]
+branch = "software" # Leave blank when using ADJ as a submodule (like this: "")
+test = true         # Enable test mode
+
+# Network Configuration
+[network]
+manual = false # Manual network device selection
+
+# Transport Configuration
+[transport]
+propagate_fault = true
+
+# TCP Configuration
+# These settings control how the backend reconnects to boards when connections are lost
+[tcp]
+backoff_min_ms = 100         # Minimum backoff duration in milliseconds
+backoff_max_ms = 5000        # Maximum backoff duration in milliseconds
+backoff_multiplier = 1.5     # Exponential backoff multiplier (e.g., 1.5 means each retry waits 1.5x longer)
+max_retries = 0              # Maximum retries before cycling (0 = infinite retries, recommended for persistent reconnection)
+connection_timeout_ms = 1000 # Connection timeout in milliseconds
+keep_alive_ms = 1000         # Keep-alive interval in milliseconds
+
 # BLCU (Boot Loader Control Unit) Configuration
 [blcu]
 ip = "127.0.0.1"      # TFTP server IP address
 download_order_id = 1 # Packet ID for download orders (0 = use default)
 upload_order_id = 2   # Packet ID for upload orders (0 = use default)
 
+# TFTP Configuration
+[tftp]
+block_size = 131072    # TFTP block size in bytes (128kB)
+retries = 3            # Maximum number of retries before aborting transfer
+timeout_ms = 5000      # Timeout between retries in milliseconds
+backoff_factor = 2     # Backoff multiplier for retry delays
+enable_progress = true # Enable progress callbacks during transfers
+
+# <-- DO NOT TOUCH BELOW THIS LINE -->
+
 # Server Configuration
 [server.ethernet-view]
 address = "127.0.0.1:4040"
@@ -33,23 +73,3 @@ order_data = "/orderStructures"
 programable_boards = "/uploadableBoards"
 connections = "/backend"
 files = "/"
-
-# ADJ (Architecture Description JSON) Configuration
-[adj]
-branch = "software" # Leave blank when using ADJ as a submodule (like this: "")
-test = true         # Enable test mode
-
-# Network Configuration
-[network]
-manual = true # Manual network device selection
-
-# Transport Configuration
-[transport]
-propagate_fault = true
-
-[tftp]
-block_size = 131072   # TFTP block size in bytes (128kB)
-retries = 3          # Maximum number of retries before aborting transfer
-timeout_ms = 5000    # Timeout between retries in milliseconds
-backoff_factor = 2   # Backoff multiplier for retry delays
-enable_progress = true # Enable progress callbacks during transfers

backend/cmd/main.go

@@ -287,7 +287,42 @@ func main() {
 		if err != nil {
 			panic("Failed to resolve local backend TCP client address")
 		}
-		go transp.HandleClient(tcp.NewClientConfig(backendTcpClientAddr), fmt.Sprintf("%s:%d", adj.Info.Addresses[board.Name], adj.Info.Ports[TcpServer]))
+		// Create TCP client config with custom parameters from config
+		clientConfig := tcp.NewClientConfig(backendTcpClientAddr)
+
+		// Apply custom timeout if specified
+		if config.TCP.ConnectionTimeout > 0 {
+			clientConfig.Timeout = time.Duration(config.TCP.ConnectionTimeout) * time.Millisecond
+		}
+
+		// Apply custom keep-alive if specified
+		if config.TCP.KeepAlive > 0 {
+			clientConfig.KeepAlive = time.Duration(config.TCP.KeepAlive) * time.Millisecond
+		}
+
+		// Apply custom backoff parameters
+		if config.TCP.BackoffMinMs > 0 || config.TCP.BackoffMaxMs > 0 || config.TCP.BackoffMultiplier > 0 {
+			minBackoff := 100 * time.Millisecond // default
+			maxBackoff := 5 * time.Second // default
+			multiplier := 1.5 // default
+
+			if config.TCP.BackoffMinMs > 0 {
+				minBackoff = time.Duration(config.TCP.BackoffMinMs) * time.Millisecond
+			}
+			if config.TCP.BackoffMaxMs > 0 {
+				maxBackoff = time.Duration(config.TCP.BackoffMaxMs) * time.Millisecond
+			}
+			if config.TCP.BackoffMultiplier > 0 {
+				multiplier = config.TCP.BackoffMultiplier
+			}
+
+			clientConfig.ConnectionBackoffFunction = tcp.NewExponentialBackoff(minBackoff, multiplier, maxBackoff)
+		}
+
+		// Apply max retries (0 or negative means infinite)
+		clientConfig.MaxConnectionRetries = config.TCP.MaxRetries
+
+		go transp.HandleClient(clientConfig, fmt.Sprintf("%s:%d", adj.Info.Addresses[board.Name], adj.Info.Ports[TcpServer]))
 		i++
 	}
 

backend/go.mod

@@ -13,6 +13,7 @@ require (
 	github.com/pin/tftp/v3 v3.0.0
 	github.com/pkg/browser v0.0.0-20240102092130-5ac0b6a4141c
 	github.com/rs/zerolog v1.29.0
+	github.com/stretchr/testify v1.9.0
 	golang.org/x/exp v0.0.0-20230522175609-2e198f4a06a1
 )
 
@@ -22,6 +23,7 @@ require (
 	github.com/ProtonMail/go-crypto v1.0.0 // indirect
 	github.com/cloudflare/circl v1.3.7 // indirect
 	github.com/cyphar/filepath-securejoin v0.2.4 // indirect
+	github.com/davecgh/go-spew v1.1.1 // indirect
 	github.com/emirpasic/gods v1.18.1 // indirect
 	github.com/go-git/gcfg v1.5.1-0.20230307220236-3a3c6141e376 // indirect
 	github.com/go-git/go-billy/v5 v5.5.0 // indirect
@@ -32,6 +34,7 @@ require (
 	github.com/mattn/go-isatty v0.0.17 // indirect
 	github.com/pjbgf/sha1cd v0.3.0 // indirect
 	github.com/pkg/errors v0.9.1 // indirect
+	github.com/pmezard/go-difflib v1.0.0 // indirect
 	github.com/sergi/go-diff v1.3.2-0.20230802210424-5b0b94c5c0d3 // indirect
 	github.com/skeema/knownhosts v1.2.2 // indirect
 	github.com/xanzy/ssh-agent v0.3.3 // indirect
@@ -40,6 +43,7 @@ require (
 	golang.org/x/sys v0.31.0 // indirect
 	golang.org/x/tools v0.13.0 // indirect
 	gopkg.in/warnings.v0 v0.1.2 // indirect
+	gopkg.in/yaml.v3 v3.0.1 // indirect
 )
 
 require (

backend/pkg/transport/network/tcp/client.go

@@ -39,29 +39,37 @@ func (client *Client) Dial() (net.Conn, error) {
 	var err error
 	var conn net.Conn
 	client.logger.Info().Msg("dialing")
-	// The max connection retries will not work because the for loop never completes, it always returns and the function is called again by transport.
+	// Reset retry counter when starting a new dial attempt
+	client.currentRetries = 0
+
 	for client.config.MaxConnectionRetries <= 0 || client.currentRetries < client.config.MaxConnectionRetries {
-		client.currentRetries++
 		conn, err = client.config.DialContext(client.config.Context, "tcp", client.address)
 
-		backoffDuration := client.config.ConnectionBackoffFunction(client.currentRetries)
-		client.logger.Error().Stack().Err(err).Dur("backoff", backoffDuration).Int("retries", client.currentRetries+1).Msg("retrying")
-		time.Sleep(backoffDuration)
-
 		if err == nil {
 			client.logger.Info().Msg("connected")
 			client.currentRetries = 0
 			return conn, nil
 		}
+
+		// Check if context was cancelled
 		if client.config.Context.Err() != nil {
 			client.logger.Error().Stack().Err(client.config.Context.Err()).Msg("canceled")
 			return nil, client.config.Context.Err()
 		}
 
+		// Check if we should retry this error
 		if netErr, ok := err.(net.Error); !client.config.TryReconnect || (!errors.Is(err, syscall.ECONNREFUSED) && (!ok || !netErr.Timeout())) {
-			client.logger.Error().Stack().Err(err).Msg("failed")
+			client.logger.Error().Stack().Err(err).Msg("failed with non-retryable error")
 			return nil, err
 		}
+
+		// Increment retry counter and calculate backoff
+		client.currentRetries++
+		backoffDuration := client.config.ConnectionBackoffFunction(client.currentRetries)
+		client.logger.Error().Stack().Err(err).Dur("backoff", backoffDuration).Int("retry", client.currentRetries).Msg("retrying after backoff")
+
+		// Sleep for backoff duration
+		time.Sleep(backoffDuration)
 	}
 
 	client.logger.Debug().Int("max", client.config.MaxConnectionRetries).Msg("max connection retries exceeded")