Echelon FTXL User Manual

FTXL User’s Guide

078-0363-01A

Echelon, LONWORKS, LONMARK, NodeBuilder, LonTalk, Neuron, 3120, 3150, LNS, Echelon logo are trademarks of Echelon Corporation registered in the United States and other countries. 3190, FTXL, OpenLDV, Pyxos, and LonScanner are trademarks of Echelon Corporation.

Other brand and product names are trademarks or registered trademarks of their respective holders.

Neuron Chips and other OEM Products were not designed for use in equipment or systems, which involve danger to human health or safety, or a risk of property damage and Echelon assumes no responsibility or liability for use of the Neuron Chips in such applications.

Parts manufactured by vendors other than Echelon and referenced in this document have been described for illustrative purposes only, and may not have been tested by Echelon. It is the responsibility of the customer to determine the suitability of these parts for each application.

ECHELON MAKES AND YOU RECEIVE NO WARRANTIES OR CONDITIONS, EXPRESS, IMPLIED, STATUTORY OR IN ANY COMMUNICATION WITH YOU, AND ECHELON SPECIFICALLY DISCLAIMS ANY IMPLIED WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

.LON, ShortStack, LonMaker, and the

No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of Echelon Corporation.

Echelon Corporation

www.echelon.com

Welcome

Echelon’s FTXL™ products enable any product that contains an Altera® Nios® II processor to quickly and inexpensively become a networked smart device. An FTXL device includes a complete ANSI/CEA 709.1-B (EN14908.1) implementation that runs on the Nios II embedded processor. Thus, the FTXL 3190™ Free Topology Smart Transceiver Chip provides a simple way to add

ONWORKS

L because it has a simple host application programming interface (API), a pre-built link-layer driver, a simple hardware interface, and comprehensive tool support.

networking to smart devices. The FTXL Transceiver is easy to use

This document describes how to develop an application for a L using Echelon’s FTXL Transceiver. It describes the architecture of an FTXL device and how to develop the software for an FTXL device. Development of a FTXL device includes creating a model file, running the LonTalk Developer utility, and using the FTXL API functions to program your FTXL application for the Nios II processor.

See the FTXL device, the development boards for which the FTXL Developer’s Kit provides reference designs, and FPGA design requirements for an FTXL device.

FTXL Hardware Guide

Audience

This document assumes that the reader has a good understanding of the

ONWORKS platform and programming for the Altera Nios II processor.

4 Transport End-to-end

communication reliability

Acknowledged and unacknowledged message delivery, common ordering, duplicate detection

3 Network Destination addressing Unicast and multicast addressing,

routers

2 Data Link Media access and framing Framing, data encoding, CRC error

checking, predictive carrier sense multiple access (CSMA), collision avoidance, priority, collision detection

1 Physical Electrical interconnect Media-specific interfaces and modulation

schemes

Echelon’s implementation of the ANSI/CEA-709.1 Control Network Protocol is called the

LonTalk protocol

. Echelon has implementations of the LonTalk

protocol in several product offerings, including the Neuron firmware (which is

included in a ShortStack

.LON® servers, and the FTXL LonTalk protocol stack. This document refers to

Micro Server), LNS® Server, LNS remote client,

2 Introduction to FTXL

the ANSI/CEA-709.1 (EN14908-1) Control Network Protocol as the “LonTalk protocol”, although other interoperable implementations exist.

A LONWORKS Device with a Single Processor Chip

A basic LONWORKS device consists of four primary components:

1. An application processor that implements the application layer, or both the application and presentation layers, of the LonTalk protocol

2. A protocol engine that implements layers 2 through 5 (or 2 through 7) of the LonTalk protocol

3. A network transceiver that provides the physical interface for the L

ONWORKS network communications media, and implements the physical

layer of the LonTalk protocol

4. Circuitry to implement the device I/O

These components can be combined in a physical device. For example, Echelon’s Smart Transceiver product can be used as a single-chip solution that combines all four components in a single chip. When used in this way, the Smart Transceiver runs the device’s application, implements the LonTalk protocol, and interfaces with the physical communications media through a transformer.

4 shows the seven-layer LonTalk protocol on a single Neuron Chip or Smart

page Transceiver.

A L

ONWORKS device that uses a single processor chip is called a

device, which means that the Neuron-based processor (the Smart Transceiver) runs both the application and the LonTalk protocol.

Figure 1 on

Neuron-hosted

FTXL User’s Guide 3

Figure 1. A Single-Chip L

For a Neuron-hosted device that uses a Neuron Chip or Smart Transceiver, the physical layer (layer 1) is handled by the Neuron Chip or Smart Transceiver. The middle layers (layers 2 through 6) are handled by the Neuron firmware. The application layer (layer 7) is handled by your Neuron C application program. You create the application program using the Neuron C programming language in either the NodeBuilder

Development Tool or the Mini EVK Evaluation Kit.

ONWORKS Device

A LONWORKS Device with Two Processor Chips

Some LONWORKS devices run applications that require more memory or processing capabilities than a single Neuron Chip or Smart Transceiver can provide. Other L an existing processor and application. For these applications, the device uses two processor chips working together:

• An Echelon Smart Transceiver

• A microprocessor, microcontroller, or embedded processor in a field-

programmable gate array (FPGA) device, typically called the

processor

A L

ONWORKS device that uses two processor chips is called a

which means that the device includes a Smart Transceiver plus a host processor.

4 Introduction to FTXL

ONWORKS devices are implemented by adding a transceiver to

host

host-based

device,

Compared to the single-chip device, the Smart Transceiver implements only a subset of the LonTalk protocol layers. The host processor implements the remaining layers and runs the device’s application program. The Smart Transceiver and the host processor communicate with each other through a linklayer interface.

For a single-chip, Neuron-hosted, device you write the application program in Neuron C. For a host-based device, you write the application program in ANSI C, C++, or other high-level language, using a common application framework and application programming interface (API). This API is called the addition, for a host-based device, you select a suitable host processor and use the host processor’s application development environment, rather than the NodeBuilder Development Tool or the Mini EVK application, to develop the application.

LonTalk API

. In

Echelon provides the following solutions for creating host-based L devices:

• The LonTalk Platform for ShortStack Micro Servers

• The LonTalk Platform for FTXL Transceivers

ONWORKS

LonTalk Platform for ShortStack Micro Servers

The LonTalk Platform for ShortStack Micro Servers is a set of development tools, APIs, and firmware for developing host-based L LonTalk Compact API and a ShortStack Micro Server.

A ShortStack Micro Server is a Smart Transceiver with firmware, the

firmware

protocol, as shown in application layer (layer 7) and part of the presentation layer (layer 6).

The ShortStack firmware allows you to use almost any host processor for your device’s application and I/O. The Smart Transceiver implements layers 2 to 5 (and part of layer 6) of the LonTalk protocol and provides the physical interface for the L

A simple serial communications interface provides communications between the ShortStack Micro Server and the host processor. Because a ShortStack Micro Server can work with any host processor, you must provide the serial driver implementation, although Echelon does provide the serial driver API and an example driver for some host processors. Currently, example drivers are available for an Atmel processor.

For ShortStack device development, you use the C programming language use the Echelon LonTalk Interface Developer utility to create the application framework. Your application uses an ANSI C API, the Echelon LonTalk Compact API, to manage communications with the ShortStack Micro Server and devices on the L

, that implements layers 2 to 5 (and part of layer 6) of the LonTalk

Figure 2 on page 6. The host processor implements the

ONWORKS communications channel.

ARM7 microprocessor and an Altera Nios II embedded

ONWORKS network.

ONWORKS devices that use the

ShortStack

. You

For ShortStack device development, you could alternatively use any programming language supported by the host processor if you port the LonTalk Compact API and the application framework generated by the LonTalk Interface Developer utility to that language.

FTXL User’s Guide 5

Using a ShortStack Micro Server makes it easy to add LONWORKS networks to any existing smart device.

ShortStack Device

Application in any suitable language

Link layer

Transceiver and

wiring

FT 3120, PL 3120,

FT 3150, or PL 3150

Smart Transceiver

SCI or SPI serial I/O link layer and driver software

Host Processor

Figure 2. The ShortStack Solution for a Host-Based L

ONWORKS Device

LonTalk Platform for FTXL Transceivers

The LonTalk Platform for FTXL Transceivers is a set of development tools, APIs, firmware, and chips for developing host-based L LonTalk API and an FTXL Transceiver.

An FTXL Transceiver is an FT 3190 Smart Transceiver with firmware that implements the data link layer (layer 2) of the LonTalk protocol, as shown in Figure 3 on page 8. The host processor implements the remaining layers (layers 3 to 7). Included with the FTXL development tools is the FTXL LonTalk protocol stack, which implements layers 3 to 6 of the LonTalk protocol and runs on the host processor. Your application implements the application layer (layer 7).

6 Introduction to FTXL

ONWORKS devices that use the

For an FTXL device, you use an Altera Nios II processor as the host processor for your device’s application and I/O. The Nios II processor runs on an Altera Cyclone II or Cyclone III FPGA device. The FTXL LonTalk protocol stack implements layers 3 to 6 of the LonTalk protocol, and the FTXL Transceiver implements layers 1 and 2, including the physical interface for the L

ONWORKS

communications channel.

The FTXL LonTalk protocol stack includes a communications interface driver for the parallel link layer that manages communications between the FTXL LonTalk protocol stack within the Nios II host processor and the FTXL Transceiver. You need to include the physical implementation of the parallel link layer in your FTXL device design. However, you do not need to provide the software implementation of the parallel interface driver because it is included with the FTXL LonTalk protocol stack, nor can you modify the Echelon-provided implementation.

For FTXL device development, you use a C or C++ compiler that supports the Nios II processor. As with ShortStack development, you use the Echelon LonTalk Interface Developer utility to create the application framework. Your application uses an ANSI C API, the Echelon LonTalk API, to manage communications with the FTXL LonTalk protocol stack, FTXL Transceiver, and devices on the L

ONWORKS network.

Using an FTXL Transceiver, it is easy to add L performance FPGA-based smart device.

ONWORKS networking to a high-

FTXL User’s Guide 7

FTXL Device

Application in C

Link layer

Transceiver and

wiring

FTXL 3190 Free

Topology Smart

Transceiver

Figure 3. An FTXL Device

11-pin parallel I/O link layer and driver software

Nios II

Host Processor

Comparing Neuron-Hosted, ShortStack, and FTXL Devices

Table 3 on page 9 compares some of the key characteristics of the Neuron-hosted and host-based solutions for L

ONWORKS devices.

8 Introduction to FTXL

Table 3. Comparing Neuron-Hosted and Host-Based Solutions for L

NeuronHosted

Characteristic

Maximum number of

Solution ShortStack Solution FTXL Solution

62 254

[1]

4096

network variables

Maximum number of

62 127

[2]

8192

aliases

Maximum number of

15 15 4096

addresses

Maximum number of

0 0 4096 dynamic network variables

Maximum number of

16 16 200 receive transaction records

Maximum number of

2 2 2500 transmit transaction records

ONWORKS Devices

Support for the

No No Yes LonTalk Extended Command Set

File access methods supported

FTP

DMF

[4]

FTP

[4]

, DMF FTP

Link-layer type N/A 4- or 5-line SCI

or 6- or 7-line SPI

Typical host API runtime footprint

N/A 5-6 KB code with 1 KB RAM

(includes serial driver, but does not include optional API or ISI API)

Host processor type N/A Any 8-, 16-, 32-, or 64-bit

microprocessor or microcontroller

Application

Neuron C Any (typically ANSI C) ANSI C or C++ for the development language

[3]

[4]

, DMF

11-line parallel I/O

[5]

[6]

540 KB (includes LonTalk protocol stack, but does not include the application or operating system)

Altera Nios II embedded processor

Nios II processor

FTXL User’s Guide 9

Notes:

1. ShortStack Micro Servers running on FT 3150 or PL 3150 Smart Transceivers support up to 254 network variables. ShortStack Micro Servers running on FT 3120 Smart Transceivers support up to 240 network variables, and ShortStack Micro Servers running on PL 3120 Smart Transceivers support up to 62 network variables. A custom Micro Server can support up to 254 network variables, depending on available resources.

2. ShortStack Micro Servers running on FT 3150 or PL 3150 Smart Transceivers support up to 127 aliases. ShortStack Micro Servers running on FT 3120 Smart Transceivers support up to 120 aliases. ShortStack Micro Servers running on PL 3120 Smart Transceivers support up to 62 aliases. A custom Micro Server can support up to 127 aliases, depending on available resources.

3. See the more information about the extended command set (ECS) network management commands. This document is available from the IHS Standards Store:

global.ihs.com/doc_detail.cfm?item_s_key=00391891&item_key_date=971131&rid= CEA.

4. An implementation of the L with the product.

5. For more information about the direct memory files (DMF) feature, see

Memory Files

6. The FTXL parallel I/O link-layer driver is included with the FTXL LonTalk protocol stack.

LonTalk Control Network Protocol Specification

ONWORKS file transfer protocol (FTP) is not provided

on page 96.

The FTXL solution provides the best performance and highest network capacity, but is limited using to an Altera Nios II host processor and the TP/FT-10 channel. The ShortStack solution provides support for any host processor (with available examples for both an Atmel ARM7 host processor and an Altera Nios II host processor), and supports both the TP/FT-10 and PL-20 channels. The ShortStack solution supports fewer network variables and aliases that the FTXL solution, but more network variables and aliases than the Neuron-hosted solution.

Because the ShortStack and FTXL solutions are both built on the LonTalk platform, they share a very similar API (the FTXL LonTalk API and the ShortStack LonTalk Compact API). Thus, migrating applications from one solution to the other is fairly easy. In addition, you can create applications that share a common code base for devices that use both solutions.

, EIA/CEA 709.1-B-2002, for

Using Direct

Requirements and Restrictions for FTXL

The FTXL Developer’s Kit supports only the FTXL 3190 Free Topology Smart Transceiver. It does not support other transceiver types.

The FTXL LonTalk protocol stack requires that the FTXL application use an embedded operating system. The FTXL Developer’s Kit includes an example application that uses the Micrium μC/OS-II operating system, but you can use any embedded operating system that meets your application’s requirements. And although the μC/OS-II operating system is a real-time operating system, the FTXL LonTalk protocol stack does not require the operating system to be a realtime operating system.

10 Introduction to FTXL

The FTXL LonTalk protocol stack and API require about 540 KB of program memory on the Nios II host processor, not including the application program or the operating system. In addition, you must provide sufficient additional nonvolatile memory for device configuration data and any non-volatile data that you include in your application.

You can implement configuration properties as configuration network variables or in configuration files. To access configuration files, you can implement the

ONWORKS file transfer protocol (FTP) or use the direct memory files (DMF)

L feature. See when to use FTP or the DMF feature.

Using Direct Memory Files

on page 96 for more information about

Development Tools for FTXL

To develop an application for a device that uses an FTXL Transceiver, you need a development system for the Nios II processor. In addition, you need the FTXL Developer’s Kit, which includes:

• The FTXL LonTalk API

• The LonTalk

your FTXL device and generating the application framework

• Example FTXL applications

• A reference design for a Nios II processor and associated hardware for

the FPGA device

You also need a network management tool to install and test your FTXL device. You can use the LonMaker Integration Tool, or any other tool that can install and monitor L information on the LonMaker tool.

You do not need the NodeBuilder Development Tool to use the FTXL Developer's Kit; however, the NodeBuilder Code Wizard that is included with the NodeBuilder tool, version 3 or later, can help you develop your Neuron C model file. The model file is used to define the device’s interoperable interface.

ONWORKS devices. See the

Interface Developer utility for defining the interface for

FTXL Architecture

An FTXL device consists of the following components:

• The FTXL 3190 Free Topology Smart Transceiver running the FTXL

firmware

• A Nios II embedded processor running the following software:

• An FTXL host application that uses the FTXL LonTalk API

LonMaker User's Guide

for more

• The FTXL LonTalk protocol stack

• The FTXL hardware abstraction layer (HAL)

• The FTXL non-volatile data (NVD) driver

• The FTXL operating system abstraction layer (OSAL)

• An embedded operating system

• The Altera SOPC Builder hardware abstraction layer (HAL)

FTXL User’s Guide 11

Figure 4 shows the basic architecture of an FTXL device.

Figure 4. FTXL Architecture

The FTXL Developer's Kit includes the FTXL LonTalk API and a precompiled library that implements the FTXL LonTalk protocol stack. The kit also includes source code for additional operating system and hardware APIs that you compile and link with your application. The FTXL LonTalk API defines the functions that your application calls to communicate with other devices on a L network. The API code provides ANSI C interfaces for the host application.

12 Introduction to FTXL

ONWORKS

The FTXL LonTalk API consists of the following types of functions:

• Functions to initialize the FTXL device after each reset.

• A function that the application must call periodically. This function

processes messages pending in any of the data queues.

• Various functions to initiate typical operations, such as the propagation

of network variable updates.

• Event handler functions to notify the application of events, such as the

arrival of network variable data or an error in the propagation of an application message.

• Functions to interface with the operating system.

The FTXL Developer’s Kit

The FTXL Developer’s Kit consists of two components: a hardware component and a software component. See the the hardware component of the FTXL Developer’s Kit.

The software component contains the software required to develop FTXL applications that use an FTXL Transceiver:

1. The FTXL LonTalk protocol stack library and FTXL LonTalk API

2. ANSI C source code for event handler functions.

FTXL Hardware Guide

for information about

3. Portable ANSI C source code for the reference implementations of the APIs for the operating system and hardware.

4. The LonTalk Interface Developer utility that you use to generate device interface data, device interface files, and a skeletal application framework.

5. Example applications that run on the reference design hardware.

The software component of the FTXL Developer’s Kit is available as a free download from the Echelon Web site: must acquire a licence from Echelon to use the FTXL Developer’s Kit.

www.echelon.com/downloads. You also

Overview of the FTXL Development Process

Figure 5 on page 14 shows a high-level overview of the development process for an FTXL application. The basic process includes the following steps:

1. Use the Altera Quartus II software and SOPC Builder tool, with input from FTXL hardware components and your FPGA design, to generate compiled hardware description files.

2. Use the LonTalk Interface Developer utility, with input from a model file that you create, to generate application framework files and interface files.

3. Use the Altera Nios II EDS IDE to create the FTXL application, with input from:

• The application framework files generated by the LonTalk Interface

Developer utility

FTXL User’s Guide 13

Source

FPGA

Hardware

Description

Files

FTXL Hardware

Components

• The FTXL hardware abstraction layer (HAL) files, which you might

need to modify

• The FTXL operating system abstraction layer (OSAL) files, which you

might need to modify

• The FTXL non-volatile data (NVD) driver files, which you might need

modify

• The FTXL LonTalk protocol stack

Because an FTXL device is comprised of both hardware and software components, different people can be involved in the various steps, and these steps can occur in parallel or sequentially. The figure does not imply a required order of steps.

Quartus II

Software and

SOPC Builder

Compiled

FPGA

Hardware

Description

Files

Nios II EDS

IDE

FTXL Application

Source

Model File

(*.nc)

LonTalk

Interface

Developer

Generated Application Framework

Files

FTXL HAL FTXL OSAL

FTXL NVD

Driver

FTXL LonTalk Protocol Stack

Interface Files

(*.xif and *.xfb)

Figure 5. Overview of the FTXL Development Process

For more information about hardware development for an FTXL device, see the

FTXL Hardware Guide

This manual describes the software development process for creating an FTXL device, which includes the general tasks listed in

Table 4.

Table 4. Tasks for Developing Software for an FTXL Device

Task Additional Considerations Reference

Install the FTXL Developer’s Kit and become familiar with it

Chapter

Started with FTXL

, on page 17

Getting

14 Introduction to FTXL

Task Additional Considerations Reference

Select an FPGA device and load it with Nios II processor and related hardware

Integrate the FTXL application with your device hardware

Test and verify your hardware design

Select and define the functional profiles and resource types for your device using tools such as the NodeBuilder Resource Editor and the SNVT and SCPT Master List

The FTXL application runs on a Nios II embedded processor, which is implemented on an FPGA device. You must meet the FTXL hardware and software requirements to ensure that the FTXL device has sufficient RAM and non-volatile memory.

You integrate the FTXL Transceiver with the device hardware. You can reuse many parts of a hardware design for different applications to create different FTXL devices.

You must ensure that the host processor and the FTXL Transceiver can communicate using the parallel interface. The FTXL Developer’s Kit includes a BringUp application to help test and verify the communications interface.

You must select profiles and types for use in the device’s interoperable interface for each application that you plan to implement. This selection can include the definition of user-defined types for network variables, configuration properties or functional profiles. A large set of standard definitions is also available and is sufficient for many applications.

FTXL

The

Hardware Guide

FTXL

The

Hardware Guide

Chapter 6,

Working with the Nios II Development Environment

101

page

FTXL

The

Hardware Guide

Chapter

, on

Creating a Model

, on page 23

File

Structure the layout and interoperable interface of your FTXL device by creating a model file

FTXL User’s Guide 15

You must define the interoperable interface for your device in a model file, using the Neuron C (Version 2.1) language, for every application that you implement. You can write this code by hand, derive it from an existing Neuron C or ShortStack application, or use the NodeBuilder Code Wizard included with the NodeBuilder Development Tool to create the required code using a graphical user interface.

Chapter

Creating a Model File

Appendix

Neuron C Syntax for the Model File

on page

, on page 23

125

Task Additional Considerations Reference

Use the LonTalk Interface Developer utility to generate device interface data, device interface files, and a skeleton application framework

Complete the FTXL LonTalk API event handler functions and callback handler functions to process application-specific

ONWORKS events

Modify the FTXL Operating System Abstraction Layer (OSAL) files for your application’s operating system

You must execute this utility, a simple click-through wizard, whenever the model file changes or other preferences change. The utility generates the interface files (including the XIF file) and source code that you can compile and link with your application. This source code includes data that is required for initialization and for complete implementations of some aspects of your device.

You must complete the event handler functions and callback handler functions for every application that you implement, because they provide input from network events to your application, and because they are part of your networked device’s control algorithm.

If you use the Micrium μC/OS-II operating system, you can use the OSAL files that are included with the FTXL Developer’s Kit.

Chapter

Using the LonTalk Interface Developer Utility

on page

Chapter

Developing an FTXL Application

147

FTXL

, on

on page

Appendix

LonTalk API

page

The FTXL Operating System Abstraction Layer

on page

157

Modify the non-volatile data (NVD) driver files

Modify your application to interface with a

ONWORKS network by

L using the FTXL LonTalk API function calls

Test, install, and integrate your FTXL device using a

ONWORKS network tool

L such as the LonMaker Integration Tool

Depending on the type of non-volatile memory that your device uses, you can use one of the non-volatile data drivers provided with the FTXL Developer’s Kit, make minor modifications to one of these drivers, or implement your own driver.

You must make these function calls for every application that you implement. These calls include, for example, calls to the LonPropagateNv() function that propagates an updated network variable value to the network. Together with the completion of the event and callback handler functions, this task forms the core of your networked device’s control algorithm.

The

Providing Persistent Storage for Non-Volatile

on page 77

Data

Chapter

Developing an FTXL Application

147

FTXL

, on

on page

Appendix

LonTalk API

page

LonMaker

User's Guide

16 Introduction to FTXL

Getting Started with FTXL

This chapter describes the FTXL Developer’s Kit and how to install it.

FTXL User’s Guide 17

FTXL Developer’s Kit Overview

The FTXL Developer’s Kit is a development toolkit that contains the hardware designs, software designs, and documentation needed for developing applications that use an FTXL Transceiver. The kit includes the following components:

• Hardware and software design files for the FPGA design, including

Quartus II files, SOPC Builder files, and Nios IDE files

• Hardware component files for the FPGA development board

• The FTXL LonTalk protocol stack and FTXL LonTalk API, delivered as a

C object library

• Software source files for the FTXL LonTalk API

• A set of example programs that demonstrate how to use the FTXL

LonTalk API to communicate with a L

• The LonTalk Interface Developer utility, which defines parameters for

your FTXL host application program and generates required device interface data for your device

ONWORKS network

• Documentation, including this

, and HTML documentation for the FTXL API

Guide

The FTXL Developer’s Kit also refers to three hardware development boards that are available from devboards GmbH, can also obtain these boards from EBV Elektronik GmbH, FTXL Developer’s Kit uses these boards for its examples and reference designs. These boards are:

• The

Contact your Altera representative for information about acquiring a Nios II development license.

See the development boards and the reference designs for the FTXL Developer’s Kit.

The software for the FTXL Developer’s Kit is available as a free download from

www.echelon.com/ftxl.

DBC2C20 Altera Cyclone II Development Board

FPGA device and peripheral I/O

FTXL Adapter Board

between the DBC2C20 development board and the FTXL Transceiver Board

FTXL Transceiver Board

Transceiver Chip and a L

FTXL Hardware Guide

for more information about the hardware

FTXL User’s Guide

www.devboards.de. European customers

, which primarily provides voltage regulation

, which includes the FTXL 3190 Smart

ONWORKS network connector

, the

FTXL Hardware

www.ebv.com. The

, which provides the

Installing the FTXL Developer’s Kit

The FTXL Developer’s Kit requires the following software:

• Altera Quartus II software, Version 7.2 or later

• Altera Nios II EDS integrated development environment (IDE), Version

7.2 or later

18 Getting Started with FTXL

• Driver software for the Altera USB-Blaster download cable

• FPGA configuration data and software for the DBC2C20 development

board (included with the FTXL Developer’ Kit)

For more information about the Altera software products, see see Chapter

Working with the Nios II Development Environment

Web site for the Nios II processor,

www.altera.com/products/ip/processors/nios2/ni2-index.html.

The following sections describe the hardware and software requirements, and how to install the FTXL Developer’s Kit.

Hardware Requirements

For the FTXL Developer’s Kit plus the Altera design software, your computer system must meet the following minimum requirements:

• Intel

• 256 MB RAM

• 5 GB available hard disk space (includes the space required for the Altera

• CD-ROM drive

• 1 available Universal Serial Bus (USB) port

The recommended specifications for your computer system include:

• Intel Pentium 4 2.0 GHz processor

• 1 GB RAM

Pentium® III 866 MHz processor

tools)

, on page 101, and the Altera

• 5 GB available hard disk space

• CD-ROM or DVD-ROM drive

• 2 available USB ports

In addition, you must have the following hardware for L

• L

ONWORKS compatible network interface, such as a U10 USB Network

Interface or an

• A L

ONWORKS TP/FT-10 network cable, with network terminator

.LON 100 Internet Server

Software Requirements

For the FTXL Developer’s Kit plus the Altera design software, your computer system must meet one of the following minimum requirements:

• Microsoft

• Microsoft Windows Vista

The following software is optional, depending on your requirements:

• Adobe Reader 7.0.8 or later

Windows® XP, plus Service Pack 2 or later

ONWORKS connectivity:

FTXL User’s Guide 19

DBC2C20 Software

Although the DBC2C20 Altera Cyclone II Development Board includes a set of software for general FPGA development, you do not need to install any of the DBC2C20 software to work with the FTXL Developer’s Kit. All of the necessary FPGA components and other software for the DBC2C20 development board are installed with the FTXL Developer’s Kit.

Installing the FTXL Developer’s Kit

To install the FTXL Developer’s Kit, perform the following steps:

1. Download the FTXL Developer’s Kit from Although the download is free, you must agree to the licence terms for the FTXL Developer’s Kit when you download it.

2. Double click the FtxlDevKit100.exe file that you downloaded. The Echelon FTXL Developer’s Kit main installer window opens.

3. Follow the installation dialogs to install the FTXL Developer’s Kit onto your computer.

After you install the kit, you can integrate it into your Nios II application development environment, as described in Chapter

Development Environment

In addition to the FTXL Developer’s Kit, the installation program also installs:

• L

ONMARK

• L

ONMARK Standard Program ID Calculator

• NodeBuilder Resource Editor

FTXL API Files

The FTXL LonTalk protocol stack and FTXL LonTalk API are provided as a C object library. In addition, the FTXL Developer’s Kit includes a set of portable ANSI C files that accompany the API, which are listed in contained in the [ you installed FTXL, usually C:\LonWorks\FTXL). In addition, there is a backup of these files in a ZIP file in the [

, on page 101.

Resource Files

FTXL

]\Core directory (where [

FTXL

www.echelon.com/ftxl.

Working with the Nios II

Table 5. These files are

FTXL

] is the directory in which

]\SourceArchive directory.

The LonTalk Interface Developer utility automatically copies these files into the project folder, but does not overwrite existing files with the same names.

Table 5. FTXL LonTalk API Files

File Name Description

FtxlApi.h Function definitions for the FTXL LonTalk API

FtxlHal.h

FtxlHal.c

20 Getting Started with FTXL

Functions for the FTXL hardware abstraction layer (HAL)

File Name Description

FtxlHandlers.c

FtxlNvdFlashDirect.c

FtxlNvdFlashFs.c

FtxlNvdUserDefined.c

FtxlOsal.h

FtxlOsal.c

FtxlTypes.h C type definitions that are used by the FTXL LonTalk API

libFtxl100.a C library for the FTXL LonTalk protocol stack and FTXL LonTalk

LonPlatform.h Definitions for adjusting your compiler and development

Function definitions for the FTXL event handler functions and callback handler functions

Functions for managing non-volatile data

Functions for the FTXL operating system abstraction layer (OSAL)

API

environment to the requirements of the FTXL LonTalk API

LonTalk Interface Developer

The LonTalk Interface Developer utility generates the device interface data and device interface files required to implement the device interface for your FTXL device. It also creates a skeleton application framework that you can modify and link with your application. This framework contains most of the code that is needed for initialization and other required processing.

The executable for the LonTalk Interface Developer utility is named LID.exe, and is installed in the LonTalk Interface Developer directory (usually, C:\LonWorks\InterfaceDeveloper).

The LonTalk Interface Developer utility also includes a command-line interface that allows make-file and script-driven use of the utility. For more information about the command-line interface, see Appendix

Command Line Usage

For more information about the LonTalk Interface Developer utility, see Chapter 4,

Using the LonTalk Interface Developer Utility

, on page 111.

Example FTXL Applications

The FTXL Developer’s Kit includes two example applications that run on the reference designs for the development boards that are available from devboards.de GmbH. The first example is a simple FTXL application that simulates a voltage amplifier, whereas the second example demonstrates the use of dynamic network variables and changeable-type network variables.

See Appendix about these examples.

Example FTXL Applications

LonTalk Interface Developer

, on page 55.

, on page 195, for more information

FTXL User’s Guide 21

22 Getting Started with FTXL

Creating a Model File

You use a model file to define your device’s interoperable interface, including its network inputs and outputs. The LonTalk Interface Developer utility converts the information in the model file into device interface data and a device interface file for your application. This chapter describes how to create a model file using the Neuron C programming language.

Syntax for the Neuron C statements in the model file is

described in Appendix

, on page 125.

File

Neuron C Syntax for the Model

FTXL User’s Guide 23

Model File Overview

The interoperable application interface of a LONWORKS device consists of its functional blocks, network variables, configuration properties, and their relationships. The receiving data using interoperable data types. The the device’s means of providing externally exposed configuration data, again using interoperable data types. The configuration data items can be read (and typically also written) by a network tool. The device interface is organized into

functional blocks

and configuration properties that are used to perform one task. These network variables and configuration properties are called the

The model file describes the functional blocks, network variables, configuration properties, and their relationships, that make up the interoperable interface for an FTXL device, using the Neuron C programming language. Neuron C is based on ANSI C, and is designed for creating a device’s interoperable interface and implementing its algorithms to run on Neuron Chips and Smart Transceivers. However, you do not need to be proficient in Neuron C to create a model file for an FTXL application because the model file does not include executable code. All tools required to process model files are included with FTXL; you do not need to license another Neuron C development tool to work with an FTXL model file. The model file uses Neuron C Version 2.1 declaration syntax.

The LonTalk Interface Developer utility uses the model file to generate device interface data and device interface files. You can use any of the following methods to create a model file:

network variables

, each of which groups together a collection of network variables

are the device’s means of sending and

configuration properties

functional block members

are

• Manually create a model file A model file is a text file that you can create with any text or programming editor, including Windows Notepad. Model files have the .nc file extension. This chapter describes the types of Neuron C statements you can include in a model file. Appendix syntax for the Neuron C statements.

• Reuse existing Neuron C code You can reuse an existing Neuron C application that was originally written for a Neuron Chip or a Smart Transceiver as a model file. The LonTalk Interface Developer utility uses only the device interface declarations from a Neuron C application program, and ignores all other code. You might have to delete some code from an existing Neuron C application program, or exclude this code using conditional compilation, as described later in this chapter.

• Automatically generate a model file You can use the NodeBuilder Code Wizard, included with Release 3 or later of the NodeBuilder Development Tool, to automatically generate a model file. Using the NodeBuilder Code Wizard, you can define your device interface by dragging functional profiles and type definitions from a graphical view of your resource catalog to a graphical view of your device interface, and refine them using a convenient graphical user interface. When you complete the device interface definition, click the Generate Code and Exit button to automatically generate your model file. Use the main file produced by the NodeBuilder Code Wizard as your model file. NodeBuilder software is not included with the FTXL

C describes the

24 Creating a Model File

Developer’s Kit, and must be licensed separately. See the

User’s Guide

for details about using the NodeBuilder Code Wizard.

NodeBuilder

See Appendix detailed Neuron C syntax for each type of statement that can be included in the model file.

Neuron C Syntax for the Model File

Defining the Device Interface

You use a model file to define the device interface for your device. The device interface for a L

• Functional blocks

• Network variables

• Configuration properties

functional block

properties, which are used together to perform one task. These network variables and configuration properties are called the

Functional blocks are defined by to describe common units of functional behavior. Each functional profile defines mandatory and optional network variables and configuration properties. Each functional block implements an instance of a functional profile. A functional block must implement all of the mandatory network variables and configuration properties defined by the functional profile, and can also implement any of the optional network variables and configuration properties defined by the functional profile. In addition, a functional block can implement network variables and configuration properties that are not defined by the functional profile – these are

implementation-specific

called

ONWORKS device consists of its:

is a collection of network variables and configuration

functional profiles

network variables and configuration properties.

, on page 125, for the

functional block members

. A functional profile is used

The primary inputs and outputs to a functional block are provided by network variables. A get from other devices on a network (an make available to other devices on a network (an Network variables are used for operational data such as temperatures, pressures, switch states, or actuator positions.

configuration property

A device (its network variables and functional blocks). Configuration properties are used for configuration data such as set points, alarm thresholds, or calibration factors. Configuration properties can be set by a network management tool (such as the LonMaker Integration tool or a customized plug-in created for the device), and allow a network integrator to customize a device’s behavior.

These interface components, and the resource files used to define them, are described in the following sections.

network variable

is a data item that specifies the configurations for a

is a data item that a device application expects to

input network variable

) or expects to

output network variable

Defining the Interface for an FTXL Application

Within the model file, you define a simple input network variable with the following syntax:

network input

type name

;

FTXL User’s Guide 25

Example: The following declaration defines an input network variable of type “SNVT_type” with the name “nviAmpere”.

network input SNVT_amp nviAmpere;

You define a simple output network variable using the same syntax, but with the output modifier:

network output

Example: The following declaration defines an output network variable of type “SNVT_type” with the name “nvoAmpere”.

network output SNVT_amp nvoAmpere;

By convention, input network variable names have an network variables have an

See

Network Variable Syntax

declaration syntax.

The LonTalk Interface Developer utility reads the network variable declarations in the model file to generate device-specific code. For the example of the

nviAmpere

a standard ANSI C type definition for the implements two global C-language variables:

typedef ncsLong SNVT_amp; … volatile SNVT_amp nviAmpere; SNVT_amp nvoAmpere;

The ncsLong data type defines the host equivalent of a Neuron C signed long variable. This type is defined in the LonPlatform.h file.

Your FTXL application can simply read the retrieve the most recently received value from that input network variable. Likewise, your application can write the result of a calculation to the global C variable, and call the appropriate FTXL LonTalk API function to propagate the network variable to the L

type name

and

nvoAmpere

;

nvi

prefix and output

nvo

prefix.

on page 132 for the full network variable

pair of network variables above, the utility generates

SNVT_amp

nviAmpere

network variable type and

global C variable to

nvoAmpere

ONWORKS network.

Choosing the Data Type

Many functional profiles define the exact type of each member network variable.

SNVT_amp

The different network variable type within a functional profile that requires this network variable type renders the implementation of the profile not valid.

Other profiles are generic profiles that allow various network variable types to implement a member. The

Defining a Functional Block

the functional profile. This profile defines the

SNVT_xxx

Implementing a generic profile allows you to choose the standard network variable type from a range of allowed types when you create the model file.

For added flexibility, if the specific functional profile allows it, your application can implement changeable-type network variables. A

variable

26 Creating a Model File

is network variable that is initially declared with a distinct default type

type used in the previous section is such a type. Using a

SFPTopenLoopSensor

on page 27) is an example for such a generic

nvoValue

, which means “any standard network variable type.”

functional block (described in

member to be of type

changeable-type network

(for example, different type (for example,

Using changeable-type network variables allows you to design a generic device (such as a generic proportional-integral-derivative (PID) controller) that supports a wide range of numeric network variable types for set-point, control, and process-value network variables.

SNVT_volt

), but can be changed during device installation to a

SNVT_volt_mil

Defining a Changeable-Type Network Variable

See information about implementing changeable-type network variables for FTXL applications.

You can also define your own nonstandard data types. The NodeBuilder Resource Editor utility, which is included with the FTXL Development Kit, allows you to define your own, nonstandard data types for network variables or configuration properties, and allows definition of your own, nonstandard functional profiles. These nonstandard types are called user-defined types and user-defined profiles.

Defining a Functional Block

The first step for defining a device interface is to select the functional profile, or profiles, that you want your device to implement. You can use the NodeBuilder Resource Editor to look through the standard functional profiles, as described in

Defining a Resource File

each of the standard functional profiles at

For example, if your device is a simple sensor or actuator, you can use one of the following standard profiles:

• Open-loop sensor (SFPTopenLoopSensor)

• Closed-loop sensor (SFPTclosedLoopSensor)

• Open-loop actuator (SFPTopenLoopActuator)

on page 40. You can find detailed documentation for

on page 30 for more

types.lonmark.org2.

• Closed-loop actuator (SFPTclosedLoopActuator).

If your device is more complex, look through the other functional profiles to see if any suitable standard profiles have been defined. If you cannot find an existing profile that meets your needs, you can define a user functional profile, as described in

Example: The following example shows a simple functional block declaration.

network output SNVT_amp nvoAmpere;

fblock SFPTopenLoopSensor { nvoAmpere implements nvoValue; } fbAmpMeter;

This functional block:

• Is named

• Implements the standard profile SFPTopenLoopSensor

Defining a Resource File

fbAmpMeter

you include the external_name keyword to define a more human-readable name)

(network management tools use this name unless

on page 40.

Use the Windows Internet Explorer browser to view this site.

FTXL User’s Guide 27

• Includes a single network variable, named implements the nvoValue network variable member of the standard profile

nvoAmpere

, which

Declaring a Functional Block

A functional block declaration, by itself, does not cause the LonTalk Interface Developer utility to generate any executable code, although it does create data that implements various aspects of the functional block. Principally, the functional block creates associations among network variables and configuration properties. The LonTalk Interface Developer utility uses these associations to create the self-documentation (SD) and self-identification (SI) data in the device and in its associated device interface file (.xif or .xfb extension).

The functional block information in the device interface file, or the SD and SI data, communicates the presence and names of the functional blocks contained in the device to a network management tool.

Network-variable or configuration members of a functional block also have selfdocumentation data, which is also automatically generated by the LonTalk Interface Developer utility. This self-documentation data provides details about the particular network variable or configuration property, including whether the network variable or configuration property is a member of a functional block.

Functional blocks can be implemented as single blocks or as arrays of functional blocks. In a functional block array, each member of the array implements the same functional profile, but has different network variables and typically has different configuration properties that implement its network variable and configuration property members.

Example: The following example shows a simple array of 10 functional blocks.

network output SNVT_amp nvoAmpere[10];

fblock SFPTopenLoopSensor { nvoAmpere[0] implements nvoValue; } fbAmpMeter[10];

This functional block array:

• Contains ten functional blocks, implementing the SFPTopenLoopSensor profile.

• Distributes the ten nvoAmpere network variables among the ten functional blocks, starting with the first network variable (at network variable array index zero). Each member of the network variable array applies to a different network variable member of the functional block array.

Defining a Network Variable

Every network variable has a type, called a the units, scaling, and structure of the data contained within the network variable. To connect a network variable to another network variable, both must have the same type. This type matching prevents common installation errors from occurring, such as connecting a pressure output to a temperature input.

fbAmpMeter[0]

network variable type

fbAmpMeter[9]

, that defines

, each

28 Creating a Model File

Type translators are also available to convert network variables of one type to another type. Some type translators can perform sophisticated transformations between dissimilar network variable types. Type translators are special functional blocks that require additional resources, for example, a dedicated typetranslating device in your network.

You can minimize the need for type translators by using standard network variable types (SNVTs) for commonly used types, and by using changeable-type network variables, where appropriate. You can also define your own user network variable types (UNVTs).

You can use the NodeBuilder Resource Editor to look through the standard network variable types, as described in you can browse the standard profiles online at

You can connect network variables on different devices that are of identical type, but opposite direction, to allow the devices to share information. For example, an application on a lighting device could have an input network variable of the switch type, while an application on a dimmer-switch device could have an output network variable of the same type. You can use a network tool, such as the LonMaker Integration Tool, to connect these two devices, allowing the switch to control the lighting device, as shown in

Defining a Resource File

types.lonmark.org.

Figure 6.

on page 40, or

Figure 6. Simple Switch Controlling a Single Light

A single network variable can be connected to multiple network variables of the same type but opposite direction. The example in switch being used to control three lights.

Figure 7 shows the same

Figure 7. Simple Switch Controlling Three Lights

FTXL User’s Guide 29

The FTXL application in a device does not need to know anything about where input network variables come from or where output network variables go. After the FTXL application updates a value for an output network variable, it uses a simple API function call to have the FTXL LonTalk protocol stack propagate it.

Through a process called installation, the FTXL stack is configured to know the logical address of the other devices (or groups of devices) in the network that expect a specific network variable, and the FTXL stack assembles and sends the appropriate packets to these devices. Similarly, when the FTXL stack receives an updated value for an input network variable required by its application program, it reads the data from the network and passes the data to the application program.

The binding process creates logical connections between an output network variable in one device and an input network variable in another device or group of devices. You can think of these connections as “virtual wires.” For example, the dimmer-switch device in the dimmer-switch-light example above could be replaced with an occupancy sensor, without requiring any changes to the lighting device.

Network variable processing is transparent, and typical networked applications do not need to know whether a local network variable is bound (“connected”) to one or more network variables on the same device, to one or more other devices, or not bound at all. For those applications that do require such knowledge, API functions (such as LonQueryNvConfig(), LonQueryAliasConfig(), LonNvIsBound(), and LonMtIsBound()) are supplied to query the related information.

binding

that takes place during network design and

Defining a Changeable-Type Network Variable

changeable-type network variable

installation-time changes to its type and its size.

You can use a changeable-type network variable to implement a generic functional block that works with different types of inputs and outputs. Typically, an integrator uses a network management tool plug-in that you create to change network variable types.

For example, you can create a general-purpose device that can be used with a variety of sensors or actuators, and then create a functional block that allows the integrator to select the network variable type depending on the physical sensor or actuator that is attached to the device during installation.

Restrictions:

• Each changeable-type network variable must be declared with an initial type in the model file. This initial type defines the default type and the maximum size of the network variable.

• A changeable-type network variable must be a member of a functional block.

• Only network variables that are not bound can change their type. To change the type of a bound network variable, you must first unbind (disconnect) the network variable.

is a network variable that supports

30 Creating a Model File

• Only a network management tool, such as the LonMaker Integration tool, can change the type of a changeable-type network variable. The FTXL device does not initiate type changes.

To create a changeable-type network variable for an FTXL application, perform the following tasks:

1. Declare the network variable with the changeable_type keyword. You must declare an initial type for the network variable, and the size of the initial type must be equal to the largest network variable size that your application supports. The initial type must be one of the interoperable standard or user network variable types.

2. Select Has changeable interface in the L

ONMARK Standard Program ID

Calculator (included with the LonTalk Interface Developer utility) to set the changeable-interface bit in the program ID when you create the device template.

3. Declare a SCPTnvType configuration property that applies to the changeable-type network variable. This configuration property is used by network management tools to notify your application of changes to the network variable type.

4. You can optionally also declare a SCPTmaxNVLength configuration property that applies to the changeable-type network variable. This configuration property informs network management tools of the maximum type length supported by the changeable-type network variable. This value is a constant, so declare this configuration property with the const modifier.

5. Implement code in your FTXL application to process changes to the SCPTnvType value. This code can accept or reject a type change. Ensure that your application can process all possible types that the changeabletype network variable might use at runtime.

6. Implement code to provide information about the current length of the network variable.

The LonMaker browser provides integrators with a user interface to change network variable types. However, you might want to provide a custom interface for integrators to change network variable types on your device. For example, the custom interface could restrict the available types to those types supported by your application, thus preventing configuration errors.

The LonMaker Integration tool, Turbo Edition, supports changeable-type network variables. However, if you use LonMaker 3.0 or earlier to manage an FTXL device with changeable-type network variables, you must explicitly set the CP value in the LonMaker browser (or in a device plug-in) to inform the FTXL device of the type changes in addition to using the “Change Network Variable Type” facility that is provided with LonMaker 3.0 or earlier to change the type of a network variable in the LNS database.

Handling Changes to Changeable-Type Network Variables

See

on page 88 for information about how your application should handle changes to changeabletype network variables.

FTXL User’s Guide 31

Defining a Configuration Property

Like network variables, configuration properties have types, called

property types

they contain. Unlike network variable types, configuration property types also specify the meaning of the data. For example, standard network variable types represent temperature values, whereas configuration property types represent specific types of temperature settings, such as the air temperature weighting used during daytime control, or the weighting of an air temperature sensor when calculating an air temperature alarm.

, that determine the units, scaling, and structure of the data that

configuration

Declaring a Configuration Property

You declare a configuration property in a model file. Similar to network variable types, there are standard and user-defined configuration property types. You can use the NodeBuilder Resource Editor to look through the standard configuration property types, as described in browse the standard profiles online at your own configuration property type, if needed.

You can implement a configuration property using either of the following techniques:

• A configuration property network variable

• A configuration file

configuration network variable

network variable or CPNV) uses a network variable to implement the configuration property. In this case, a L configuration property, just like any other network variable. A CPNV can also provide your application with detailed notification of updates to the configuration property. However, a CPNV is limited to a maximum of 31 bytes, and an FTXL application is limited to a maximum of 4096 network variables, including CPNVs. Use the network … config_prop syntax described in

Configuration Network Variable

property as a configuration network variable. By convention, CPNV names start

nci

with an

prefix, and configuration properties in files start with a cp prefix.

Defining a Resource File

types.lonmark.org. You can also define

(also known as a configuration property

ONWORKS device can modify the

on page 143 to implement a configuration

on page 40, or you can

Declaring a

configuration file

two blocks of data called value files, rather than as separate externally exposed data items. A value file consists of configuration property records of varying length concatenated together. Each value file must fit as contiguous bytes into the memory space in the device. When there are two value files, one contains writeable configuration properties, and the second contains read-only data. To allow a network management tool to access the data items in the value file, you specify a provided template file, which is an array of text characters that describes the elements in the value files. When you use the Direct Memory Files feature, the total size of the directory, template file, and value files cannot exceed 65 535 bytes (64 KB -1). When you use FTP, individual files cannot exceed 2 147 483 647 bytes (2 GB -1, or 2

Other devices cannot connect to or poll a configuration property implemented in a configuration file. To modify a configuration property implemented in a configuration file, a network management tool must modify the configuration file, for which your application must provide an appropriate access method.

32 Creating a Model File

implements the configuration properties for a device as one or

-1 bytes).

You must implement configuration properties within a configuration file if any of the following apply to your application:

• The total number of network variables (including configuration network

variables and dynamic network variables) exceeds the total number of available network variables (a maximum of 4096 for an FTXL device, but potentially fewer than 4096 depending on the resources available).

• The size of a single configuration property exceeds the maximum size of a

configuration network variable (31 bytes).

• Your device cannot use a configuration network variable (CPNV). For

example, for a device that uses a configuration property array that applies to several network variables or functional blocks with one instance of the configuration property array each, the configuration property array must be shared among all network variables or functional blocks to which it applies. In this case, the device must implement the configuration properties within a configuration file.

In addition, you might decide whether to implement configuration properties within a configuration file for performance reasons. Using the direct memory files (DMF) feature can be faster than using configuration network variables (CPNVs) if you have more than a few configuration properties because multiple configuration properties can be updated during a single write to memory (especially during device commissioning). However, FTP can be faster than DMF if there are many configuration properties to be updated.

Use the cp_family syntax described in

The Configuration Property Type

on page

140 to implement a configuration property as a part of a configuration file.

When implementing configuration property files, the LonTalk Interface Developer utility combines all configuration properties declared using the cp_family keyword, and creates the value files and a number of related data structures.

However, you must provide one of two supported mechanisms to access these files:

• An implementation of the L

ONWORKS file transfer protocol

• Support for the direct memory files feature

The LonTalk Interface Developer utility provides most of the required code to support direct memory files. However, if you use FTP, you must also implement

ONWORKS file transfer protocol within your application program. You would

the L typically implement the L

ONWORKS file transfer protocol only if the total amount

of related data exceeds (or is likely to exceed) the size of the direct memory file window.

See the information about the L

Files

File Transfer engineering bulletin at www.echelon.com for more

ONWORKS file transfer protocol; see

Using Direct Memory

on page 96 for more information about the direct memory files feature.

To indicate which file access method the application should use, you must declare the appropriate network variables in your model file:

• For direct memory files, declare an output network variable of type

SNVT_address. If your device implements the SFPTnodeObject functional profile, you use this network variable to implement the profile’s nvoFileDirectory member. If your device does not implement the

FTXL User’s Guide 33

SFPTnodeObject functional profile, simply add this network variable to the model file. You do not need to initialize this network variable (any initial value is ignored – the LonTalk Interface Developer utility calculates the correct value).

• For FTP, declare at least two mandatory network variables, an input

network variable of type SNVT_file_req, and an output network variable of type SNVT_file_status. You also need to define a message tag for the transfer of the data. In addition, you need an input network variable of type SNVT_file_pos to support random access to the various files. You must also implement the L application program.

The LONWORKS file transfer protocol and the direct memory files feature are mutually exclusive; your device cannot implement both.

ONWORKS file transfer protocol within your

Responding to Configuration Property Value Changes

Events are not automatically generated when a configuration property implemented in a configuration file is updated, but you can declare your configuration property so that a modification to its value causes the related functional block to be disabled and re-enabled, or causes the device to be taken offline and brought back online after the modification, or causes the entire device to reset. These state changes help to synchronize your application with new configuration property values.

Your application could monitor changes to the configuration file, and thus detect changes to a particular configuration property. Such monitoring would be implemented in the FTP server or direct memory files driver.

However, many applications do not need to know that a configuration property value has changed. For example, an application that uses a configuration property to parameterize an algorithm that uses some event as a trigger (such as a network variable update or a change to an input signal) would not typically need to know of the change to the configuration property value, but simply consider the most recent value.

Defining a Configuration Property Array

You can define a configuration property as:

• A single configuration property

• An array of configuration properties

• A configuration property array

A single configuration property either applies to one or more network variables or functional blocks within the model file for the device, or the configuration property applies to the entire device.

When you define an array of configuration properties, each element of the array can apply to one or more network variables or functional blocks within the model file.

34 Creating a Model File

When you define a configuration property array, the entire array (but not each element) applies to one or more network variables or functional blocks within the model file. That is, a configuration property array is atomic, and thus applies in its entirety to a particular item.

Assuming that the device has sufficient resources, it is always possible to define arrays of configuration properties. However, configuration property arrays are subject to the functional profile definition. For each member configuration property, the profile describes whether it can, cannot, or must be implemented as a configuration property array. The profile also describes minimum and maximum dimensions for the array. If you do not implement the configuration property array as the profile requires, the profile’s implementation becomes incorrect.

Example:

This example defines a four-channel analog-to-digital converter (ADC), with the following properties:

• Four channels (implemented as an array of functional blocks)

• One gain setting per channel (implemented as an array of configuration

properties)

• A single offset setting for the ADC (implemented as a shared

configuration property)

• A linearization setting for all channels (implemented as a configuration

property array)

#include <s32.h> #define CHANNELS 4

network output SNVT_volt nvoAnalogValue[CHANNELS];

network input cp SCPTgain nciGain[CHANNELS]; network input cp SCPToffset nciOffset; network input cp SCPTsetpoint nciLinearization[5];

fblock SFPTopenLoopSensor { // the actual network variable that implements the // mandatory 'nvoValue' member of this profile: nvoAnalogValue[0] implements nvoValue; } fbAdc[CHANNELS] external_name("Analog Input") fb_properties { // one gain factor per channel: nciGain[0], // one offset, common to all channels: static nciOffset, // one linearization array for all channels: static nciLinearization = { {0, 0}, {2, 0}, {4, 0}, {6, 0}, {8, 0} }; };

This example implements a single output network variable, of type SNVT_volt, per channel to represent the most recent ADC reading. This network variable has a fixed type, defined at compile-time, but could be defined as a changeabletype network variable if needed for the application.

FTXL User’s Guide 35

There is one gain setting per channel, implemented as an array of configuration network variables (CPNVs), of type SCPTgain, where the elements of the array are distributed among the four functional blocks contained in the functional block array. Because the SCPTgain configuration property has a default gain factor of

1.0, no explicit initialization is required for this configuration property network variable.

There is a single offset setting, implemented as a configuration network variable (CPNV), of type SCPToffset. This CPNV applies to all channels, and is shared among the elements of the functional block array. The SCPToffset configuration property has a default value of zero.

The SCPToffset configuration property is a type-inheriting configuration property. The true data type of a type-inheriting property is the type of the network variable to which the property applies. For an SFPTopenLoopSensor standard functional profile, the SCPToffset configuration property applies to the functional block, and thus implicitly applies to the profile's primary member network variable. In this example, the effective data type of this property is SNVT_volt (inherited from nvoAnalogValue).

The example also includes a five-point linearization factor, implemented as a configuration property array of type SCPTsetpoint. The SCPTsetpoint configuration property is also a type-inheriting configuration property, and its effective data type is also SNVT_volt in this example.

Because the SCPTsetpoint linearization factor is a configuration property array, it applies to the entire array of functional blocks, unlike the array of SCPTgain configuration property network variables, whose elements are distributed among the elements of the functional block array. In this example, the linearization configuration property array is implemented with configuration property network variables, and must be shared among the elements of the functional block array.

To implement the linearization array of configuration properties such that each of the four functional blocks has its own linearization data array, you must implement this configuration property array in files, and declare the configuration property with the cp_family modifier.

Table 6 shows the relationships between the members of the functional-block array. As the table shows, each channel has a unique gain value, but all channels share the offset value and linearization factor.

Table 6. Functional-Block Members for the Four-Channel ADC

Channel Gain Offset Linearization

fbAdc[0] nciGain[0]

fbAdc[1] nciGain[1]

nciOffset nciLinearization[0..4]

fbAdc[2] nciGain[2]

fbAdc[3] nciGain[3]

36 Creating a Model File

Sharing a Configuration Property

The typical instantiation of a configuration property is unique to a single device, functional block, or network variable. For example, a configuration property family whose name appears in the property list of five separate network variables has five instantiations, and each instance is specific to a single network variable. Similarly, a network variable array of five elements that includes the same configuration property family name in its property list instantiates five members of the configuration property family, and each one applies to one of the network variable array elements.

Rather than creating extra configuration property instances, you can specify that functional blocks or network variables share a configuration property by including the static or global keywords in the configuration property declaration.

The global keyword causes a configuration property member to be shared among all the functional blocks or network variables whose property list contains that configuration property family name. The functional blocks or network variables in the configuration property family can have only one such global member. Thus, if you specify a global member for both the functional blocks and the network variables in a configuration property family, the global member shared by the functional blocks is a the network variables.

The static keyword causes a configuration property family member to be shared among all elements of the array it is associated with (either network variable array or functional block array). However, the sharing of the static member does not extend to other network variables or functional blocks outside of the array.

different

member than the global member shared by

Example 1:

// CP for throttle (default 1 minute) SCPTmaxSndT cp_family cpMaxSendT = { 0, 0, 1, 0, 0 };

// NVs with shared throttle: network output SNVT_lev_percent nvoValue1 nv_properties { global cpMaxSendT }; network output SNVT_lev_percent nvoValue2 nv_properties { global cpMaxSendT // the same as the one above }; network output SNVT_lev_percent nvoValueArray[10] nv_properties { static cpMaxSendT // shared among the array // elements only };

In addition to sharing members of a configuration property family, you can use the static or global keywords for a configuration network variable (CPNV) to specify sharing. However, a shared configuration property network variable cannot appear in two or more property lists without the global keyword because there is only one instance of the network variable (configuration property families can have multiple instances).

A configuration property that applies to a device cannot be shared because there is only one device per application.

FTXL User’s Guide 37

Example 2:

The following model file defines a three-phase ammeter, implemented with an array of three SFPTopenLoopSensor functional blocks. The hardware for this device contains a separate sensor for each phase, but a common analog-to-digital converter for all three phases. Each phase has individual gain factors, but shares one property to specify the sample rate for all three phases.

#define NUM_PHASES 3

SCPTgain cp_family cpGain; SCPTupdateRate cp_family cpUpdateRate;

network output SNVT_amp nvoAmpere[NUM_PHASES];

fblock SFPTopenLoopSensor { nvoAmpere[0] implements nvoValue; } fbAmpereMeter[NUM_PHASES] external_name("AmpereMeter") fb_properties { cpGain, static cpUpdateRate };

Inheriting a Configuration Property Type

You can define a configuration property type that does not include a complete type definition, but instead references the type definition of the network variable to which it applies. A configuration property type that references another type is called a family member for a type-inheriting configuration property appears in a property list, the instantiation of the configuration property family member uses the type of the network variable. Likewise, a configuration property network variable can be type-inheriting; however, for configuration network variable arrays and arrays of configuration network variables (CPNVs), each element of the array must inherit the same type.

type-inheriting configuration property

. When the configuration property

Type-inheriting configuration properties that are listed in an nv_properties clause inherit the type from the network variable to which they apply. Typeinheriting configuration properties that are listed in an fb_property clause inherit their type from the functional profile’s principal network variable member, an attribute that is assigned to exactly one network variable member.

Recommendation: Because the type of a type-inheriting configuration property is not known until instantiation, you should specify the configuration property initializer option in the property list rather than in the declaration. Likewise, you should specify the

range-mod

Restrictions:

• Type-inheriting configuration network variables that are also shared can

• A type-inheriting configuration property cannot be used as a device

A typical example of a type-inheriting configuration property is the SCPTdefOutput configuration property type. Several functional profiles list the

38 Creating a Model File

strings can apply to different instantiations of the property.

only be shared among network variables of identical type.

property, because the device has no type from which to inherit.

range-mod

string in the property list because different

SCPTdefOutput configuration property as an optional configuration property, and use it to define the default value for the sensor's principal network variable. The functional profile itself, however, might not define the type for the principal network variable.

The following example implements a SFPTopenLoopSensor functional block with an optional SCPTdefOutput configuration property. The configuration property inherits the type from the network variable it applies to, SNVT_amp in this case.

Example 1:

SCPTdefOutput cp_family cpDefaultOutput;

network output SNVT_amp nvoAmpere nv_properties { cpDefaultOutput = 123 };

fblock SFPTopenLoopSensor { nvoAmpere implements nvoValue; } fbAmpereMeter;

The initial value (123) must be provided in the instantiation of the configuration property, because the type for cpDefaultOutput is not known until it is instantiated.

You can also combine type-inheriting configuration properties with network variables that are of changeable type. The type of such a network variable can be changed dynamically by a network integrator when the device is installed in a network.

Example 2:

SCPTdefOutput cp_family cpDefaultOutput; SCPTnvType cp_family cpNvType;

network output changeable_type SNVT_amp nvoValue nv_properties { cpDefaultOutput = 123, cpNvType };

fblock SFPTopenLoopSensor { nvoValue implements nvoValue; } fbGenericMeter;

The nvoValue principal network variable, although it is of changeable type, must still implement a default type (SNVT_amp in the example). The SCPTdefOutput type-inheriting configuration property inherits the type information from this initial type. Therefore, the initializer for cpDefaultOutput must be specific to this instantiation. Furthermore, the initializer must be valid for this initial type.

If the network integrator decides to change this type at runtime, for example, to SNVT_volt, then it is in the responsibility of the network management tool to apply the formatting rules that apply to the new type when reading or writing this configuration property. However, your application has the responsibility to propagate the new type to this network variable’s type-inheriting configuration properties (if any).

FTXL User’s Guide 39

Declaring a Message Tag

You can declare a message tag in a model file. A point for application messages. Application messages are used for the

ONWORKS file transfer protocol, and are also used to implement proprietary

L interfaces to L

Application

Message tag declarations do not generate code, but result in a simple enumeration, whose members are used to identify individual tags. There are two basic forms of message tags: bindable and nonbindable.

Example:

msg_tag myBindableMT; msg_tag bind_info(nonbind) myNotBindableMT;

Similar to network variables, you can connect bindable message tags together, thus allowing applications to communicate with each other through the message tags (rather than having to know specific device addressing details). Each bindable message tag requires one address-table space for its exclusive use.

Sending application messages through bindable message tags is also known as sending application messages with implicit addressing.

Nonbindable message tags enable (and require) the use of explicit addresses, which the sending application must provide. However, these addresses do not require address-table space.

ONWORKS devices as described in Chapter 5,

, on page 73.

message tag

Developing an FTXL

is a connection

Defining a Resource File

Functional profiles, network variable types, and configuration property types are defined in recognized by all interoperable network management tools, such as the LonMaker Integration Tool. This standard format enables device manufacturers to create definitions for user functional profiles, user network variable types (UNVTs), and user configuration property types (UCPTs) that can be used during installation by a network integrator using any interoperable network management tool.

A set of standard functional profiles, standard network variable types (SNVTs), and standard configuration property types (SCPTs) is defined by a standard resource file set distributed by L functional profile defined in a resource file is also called a

template

Resource files are grouped into specified range of program IDs. A complete resource file set consists of a type file (.TYP extension), a functional profile definitions file (.FPT extension), a format file (.FMT extension), and one or more language files (.ENG, .ENU, or other extensions).

Each set defines functional profiles, network variable types, and configuration properties for a particular type of device. The program ID range is determined by

program ID template

a scope value specifies which fields of the program ID template are used to match

resource files

. LONWORKS resource files use a standard format that is

in the file, and a

ONMARK International (www.lonmark.org). A

functional profile

resource file sets

scope

, where each set applies to a

value for the resource file set. The

40 Creating a Model File

the program ID template to the program ID of a device. That is, the range of device types to which a resource file applies is the scope of the resource file.

The program ID template has an identical structure to the program ID of a device, except that the applicable fields might be restricted by the scope. The scope value is a kind of filter that indicates the relevant parts of the program ID. For example, the scope can specify that the resource file applies to an individual device type, or to all device types.

You can specify a resource file for any of the following scopes:

0 – Standard

Applies to all devices.

1 – Device Class

Applies to all devices with the specified device class.

2 – Device Class and Subclass

Applies to all devices with the specified device class and subclass.

3 – Manufacturer

Applies to all devices from the specified manufacturer.

4 – Manufacturer and Device Class

Applies to all devices from the specified manufacturer with the specified device class.

5 – Manufacturer, Device Class, and Device Subclass

Applies to all devices from the specified manufacturer with the specified device class and device subclass.

6 – Manufacturer, Device Class, Device Subclass, and Device Model

Applies to all devices of the specified type from the specified manufacturer.

For scopes 1 through 6, the program ID template included in the resource file set specifies the components. Network management tools match this template against the program ID for a device when searching for an appropriate resource file.

For a device to be able to use a resource file set, the program ID of the device must match the program ID template of the resource file set to the degree specified by the scope. Thus, each L

ONWORKS manufacturer can create resource

files that are unique to their devices.

Example: Consider a resource file set with a program ID template of 81:23:45:01:02:05:04:00, with manufacturer and device class scope (scope 4). Any device with the manufacturer ID fields of the program ID set to 1:23:45 and the device class ID fields set to 01:02 would be able to use types defined in this resource file set. However, resources on devices of the same class, but from a different manufacturer, could not access this resource file set.

A resource file set can also use information in any resource file set that has a numerically lower scope, as long as the relevant fields of their program ID templates match. For example, a scope 4 resource file set can use resources in a scope 3 resource file set, assuming that the manufacturer ID components of the resource file sets’ program ID templates match.

Scopes 0 through 2 are reserved for standard resource definitions published by Echelon and distributed by L

ONMARK International. Scope 0 applies to all

FTXL User’s Guide 41

devices, and scopes 1 and 2 are reserved for future use. Because scope 0 applies to all devices, there is a single scope 0 resource file set called the

resource file set

The FTXL Developer's Kit includes the scope 0 standard resource file set that defines the standard functional profiles (SFPTs), SNVTs, and SCPTs (updates are also available from L also includes the NodeBuilder Resource Editor that you can use to view the standard resource file set, or use to create your own user functional profiles (UFPTs), UNVTs, and UCPTs.

You can define your own functional profiles, types, and formats in scope 3 through 6 resource files.

Most LNS tools, including the LonMaker tool assume a default scope of 3 for all user resources. LNS automatically sets the scope to the highest (most specific) applicable scope level. However, if you use LNS 3.0 or earlier with scope 4, 5, or 6 resource files, you must explicitly set the scope in LNS so that LNS uses the appropriate scope. See the developing a plug-in to set the scope, or see the help) for information about modifying a device shape to set the scope.

ONMARK International at www.lonmark.org). The kit

NodeBuilder User’s Guide

for information about

LonMaker User's Guide

standard

(or online

Implementation-Specific Scope Rules

When you add implementation-specific network variables or configuration properties to a standard or user functional profile, you must ensure that the scope of the resource definition for the additional item is numerically less than or equal to the scope of the functional profile, and that the member number is set appropriately. For example:

• If you add an implementation-specific network variable or configuration

property to a standard functional block (SFPT, scope 0), it must be defined by a standard type (SNVT, or SCPT).

• If you implement a functional block that is based on a manufacturer

scope resource file (scope 3), you can add an implementation-specific network variable or configuration property that is defined in the same scope 3 resource file, and you can also add an implementation-specific network variable or configuration property that is defined by a SNVT or SCPT (scope 0).

You can add implementation-specific members to standard functional profiles using inheritance by performing the following tasks:

1. Use the NodeBuilder Resource Editor to create a user functional profile with the same functional profile key as the standard functional profile.

2. Set Inherit members from scope 0 in the functional profile definition. This setting makes all members of the standard functional profile part of your user functional profile.

3. Declare a functional block based on the new user functional profile.

Add implementation-specific members to the functional block.

42 Creating a Model File

Writing Acceptable Neuron C Code

When processing the model file, the LonTalk Interface Developer utility distinguishes between three categories of Neuron C statements:

• Acceptable

• Ignored – ignored statements produce a warning

• Unacceptable – unacceptable statements produce an error

Appendix and ignored compiler directives for model files. All other compiler directives are not accepted by the LonTalk Interface Developer utility and cause an error if included in a model file. A statement can be unacceptable because it controls features that are meaningless in an FTXL device, or because it refers to attributes that are determined by the FTXL protocol stack or by other means.

The LonTalk Interface Developer utility ignores all executable code and I/O object declarations. These constructs cause the LonTalk Interface Developer utility to issue a warning message. The LonTalk Interface Developer utility predefines the _FTXL and _MODEL_FILE macros, so that you can use #ifdef or #ifndef compiler directives to control conditional compilation of source code that is used for standard Neuron C compilation and as an FTXL model file.

All constructs not specifically mentioned as unacceptable or ignored are acceptable.

Model File Compiler Directives

Anonymous Top-Level Types

Anonymous top-level types are not acceptable. The following Neuron C construct is not acceptable:

network output struct {int a; int b;} nvoZorro;

This statement is not acceptable because the type of the nvoZorro network variable does not have a name. The LonTalk Interface Developer utility issues an error when it detects such a construct.

, on page 119, lists the acceptable

Using a named type solves the problem, for example:

typedef struct { int a; int b; } Zorro; network output Zorro nvoZorro;

The use of anonymous sub-types is permitted. For example, the LonTalk Interface Developer utility allows the following type definition:

typedef struct { int a; int b; struct { long x; long y; long z; } c; } Zorro; network output Zorro nvoZorro;

FTXL User’s Guide 43

Legacy Neuron C Constructs

You must use the Neuron C Version 2.1 syntax described in this manual. You cannot use legacy Neuron C constructs for defining L interfaces. That is, you cannot use the config modifier for network variables, and you cannot use Neuron C legacy syntax for declaring functional blocks or configuration properties. The legacy syntax used an sd_string() modifier containing a string that starts with a ‘&’ or ‘@’ character.

ONMARK-compliant

Using Authentication for Network Variables

Authentication is a special acknowledged service between one source device and one or more (up to 63) destination devices. Authentication is used by the destination devices to verify the identity of the source device. This type of service is useful, for example, if a device containing an electronic lock receives a message to open the lock. By using authentication, the electronic lock device can verify that the “open” message comes from the owner, not from someone attempting to break into the system.

Authentication doubles the number of messages per transaction. An acknowledged message normally requires two messages: an update and an acknowledgment. An authenticated message requires four messages, as shown

Figure 8 on page 46. These extra messages can affect system response time

in and capacity.

A device can use authentication with acknowledged updates or network variable polls. However, a device cannot use authentication with unacknowledged or repeated updates.

For a program to use authenticated network variables or send authenticated messages, you must perform the following steps:

1. Declare the network variable as authenticated, or allow the network management tool to specify that the network variable is to be authenticated.

2. Specify the authentication key to be used for this device using a network management tool, and enable authentication. You can use the LonMaker Integration Tool to install a key during network integration, or your application can use the LonQueryDomainConfig() and LonUpdateDomainConfig() API functions to install a key locally.

Specifying the Authentication Key

All devices that read or write a given authenticated network variable connection must have the same authentication key. This 48-bit authentication key is used in a special way for authentication, as described in

45. If a device belongs to more than one domain, you must specify a

page separate key for each domain.

The key itself is transmitted to the device only during the initial configuration. All subsequent changes to the key do not involve sending it over the network. The network management tool can modify a device’s key over the network, in a secure fashion, with a network management message.

How Authentication Works

44 Creating a Model File

Alternatively, your application can use a combination of the LonQueryDomainConfig() and LonUpdateDomainConfig() API calls to specify the authentication keys during application start-up.

If you set the authentication key during device manufacturing, you must perform the following tasks to ensure that the key is not exposed to the network during device installation:

1. Specify that the device should use network-management authentication (set the configuration data in the LonConfigData data structure, which is defined in the FtxlTypes.h file).

2. Set the device’s state to configured. An unconfigured device does not enforce authentication.

3. Recommended: Set the device’s domain to an invalid domain value to avoid address conflicts during device installation.

If you do not set the authentication key during device manufacturing, the device installer can specify authentication for the device using the network management tool, but must specify an authentication key because the device has only a default key.

How Authentication Works

Figure 8 on page 46 illustrates the process of authentication:

1. Device A uses the acknowledged service to send an update to a network variable that is configured with the authentication attribute on Device B. If Device A does not receive the challenge (described in step retry of the initial update.

2. Device B generates a 64-bit random number and returns a challenge packet that includes the 64-bit random number to Device A. Device B then uses an encryption algorithm (part of the FTXL LonTalk protocol stack) to compute a transformation on that random number using its 48bit authentication key and the message data. The transformation is stored in Device B.

3. Device A then also uses the same encryption algorithm to compute a transformation on the random number (returned to it by Device B) using its 48-bit authentication key and the message data. Device A then sends this computed transformation to Device B.

4. Device B compares its computed transformation with the number that it receives from Device A. If the two numbers match, the identity of the sender is verified, and Device B can perform the requested action and send its acknowledgment to Device A. If the two numbers do not match, Device B does not perform the requested action, and an error is logged in the error table.

If the acknowledgment is lost and Device A tries to send the same message again, Device B remembers that the authentication was successfully completed and acknowledges it again.

2), it sends a

FTXL User’s Guide 45

ACKD Message or

Request

Device A

(Writer)

Challenge

Reply to challenge

ACK or Response

Device B

(reader)

Figure 8. Authentication Process

If Device A attempts to update an output network variable that is connected to multiple readers, each receiver device generates a different 64-bit random number and sends it in a challenge packet to Device A. Device A must then transform each of these numbers and send a reply to each receiver device.

The principal strength of authentication is that it cannot be defeated by simple record and playback of commands that implement the desired functions (for example, unlocking the lock). Authentication does not require that the specific messages and commands be secret, because they are sent unencrypted over the network, and anyone who is determined can read those messages.

It is good practice to connect a device directly to a network management tool when initially installing its authentication key. This direct connection prevents the key from being sent over the network, where it might be detected by an intruder. After a device has its authentication key, a network management tool can modify the key, over the network, by sending an increment to be added to the existing key.

You can update the device’s address without having to update the key, and you can perform authentication even if the devices’ domains do not match. Thus, an FTXL device can set its key during device manufacturing, and you can then use a network management tool to update the key securely over the network.

Managing Memory

The LonTalk Interface Developer Neuron C compiler generates four tables that affect memory usage. The FTXL LonTalk protocol stack and network management tools use these tables to define the network configuration for a device. The LonTalk Interface Developer utility allocates space for the following tables:

• The address table

• The alias table

• The domain table

• The network variable configuration table

See the for more information about these tables. This document is available from the IHS Standards Store:

46 Creating a Model File

LonTalk Control Network Protocol Specification

, EIA/CEA 709.1-B-2002,

http://global.ihs.com/doc_detail.cfm?item_s_key=00391891&item_key_date=9711 31&rid=CEA.

See Appendix 181, for information about how to calculate the memory requirements for you FTXL application.

Address Table

The address table contains the list of network addresses to which the device sends network variable updates or polls, or sends implicitly-addressed application messages. You can configure the address table through network management messages from a network management tool.

By default, the LonTalk Interface Developer utility calculates the size of the address table. The utility calculates the required number of address table entries based on parameters defined in the device’s interface, such as the number of static polling input network variables, static non-constant output network variables, bindable message tags, the number of aliases, and the number of dynamic network variables. The utility always allocates at least 15 address table entries. Within the LonTalk Interface Developer utility, you can override the automatic calculation of the table size and specify any number of entries, from 0 to 4096.

Recommendation: Whenever possible, use the LonTalk Interface Developer utility-generated size for the address table.

The maximum number of address table entries that a device could require is determined by the expected maximum number of different destination entries that the device requires for connections (network variables and bindable message tags).

Determining Memory Usage for FTXL Applications

, on page

The size of the address table affects the amount of RAM and non-volatile memory required for the device. When the LonTalk Interface Developer utility calculates the size of the address table, it attempts to balance the need to limit the amount of resources required (small address table) and the need for comprehensive coverage (large address table). Although you generally do not need to, you can override the automatically calculated value with one that reflects the use of the device.

Alias Table

An alias is an abstraction for a network variable that is managed by network management tools and the FTXL LonTalk protocol stack. Network management tools use aliases to create connections that cannot be created solely with the address and network variable tables. Aliases provide network integrators with more flexibility for how devices are installed into networks.

The alias table has no default size, and can contain up to 8192 entries. The LonTalk Interface Developer utility calculates the size of the alias table. The utility calculates the required number of alias table entries based on parameters defined in the device’s interface, such as the number of static network variables and the number of supported dynamic network variables. The utility always allocates at least 5 alias table entries, unless the device does not support any network variables. Within the LonTalk Interface Developer utility, you can

FTXL User’s Guide 47

override the automatic calculation of the table size and specify any number of entries, from 0 to 8192.

Recommendation: Whenever possible, use the LonTalk Interface Developer utility-generated size for the alias table.

The maximum number of aliases that a device could require depends on its involvement in network variable connections and the characteristics of these connections. The size of the alias table also affects the performance of the device, because the alias table must be searched whenever network variable updates arrive. When the LonTalk Interface Developer utility calculates the size of the alias table, it attempts to balance the need for performance (small alias table) and the need for comprehensive coverage (large alias table). Although you generally do not need to, you can override the automatically calculated value with one that reflects the use of the device.

Domain Table

The number of domain table entries is dependent on the network in which the device is installed; it is not dependent on the application.

The LonTalk Interface Developer utility always allocates 2 domain table entries. From the command-line interface for the LonTalk Interface Developer utility, you can override the number of entries. However, L all interoperable L the size of the domain table to one entry will prevent certification.

ONMARK International requires

ONWORKS devices to have two domain table entries. Reducing

Recommendation: Whenever possible, use the LonTalk Interface Developer utility-generated number of domain table entries.

Network Variable Configuration Table

This table contains one entry for each network variable that is declared in the model file. Each element of a network variable array counts separately.

The maximum size of the network variable configuration table is 4096 entries. You cannot change the size of this table, except by adding or deleting static network variables or by increasing or decreasing the number of dynamic network variables.

Example Model files

This section describes a few example model files, with increasing levels of complexity.

See

Network Variable and Configuration Property Declarations

information about mapping types and items declared in the model file to those shown in the LonTalk Interface Developer utility-generated application framework.

Simple Network Variable Declarations

on page 68 for

This example declares one input network variable and one output network variable. Both network variables are declared with the SNVT_count type. The names of the network variables (nviCount and nvoCount) are arbitrary.

48 Creating a Model File

However, it is a common practice to use the “nvi” prefix for input network variables and the "nvo" prefix for output network variables.

network input SNVT_count nviCount; network output SNVT_count nvoCount;

The LonTalk Interface Developer utility compiles this model file into an application framework that contains, among other things, two global C variables in the FtxlDev.c file:

volatile SNVT_count nviCount; SNVT_count nvoCount;

When an update occurs for the input network variable (nviCount), the FTXL LonTalk protocol stack stores the updated value in the global variable. The application can use this variable like any other C variable. When the application needs to update the output value, it updates the nvoCount variable, so that the FTXL LonTalk protocol stack can read the updated value and send it to the network.

For more information about how the LonTalk Interface Developer utilitygenerated framework represents network variables, see

Important: This example is not interoperable because it does not use functional blocks to define the purpose of these network variables. However, this type of declaration can define a functioning device for an initial test application.

Using Types

on page 64.

Network Variables Using Standard Types

A more complete example includes the use of more complex standard network variable types and declarations. This example provides the model for a simple electricity meter, where all input data is retrieved from the network through the nviAmpere, nviVolt, and nviCosPhi input network variables. The result is posted to the nvoWattage output network variable. A second nvoUsage output network variable is polled and uses non-volatile storage to count the meter's total lifetime.

network input SNVT_amp nviAmpere; network input SNVT_volt nviVolt; network input SNVT_angle nviCosPhi; network output SNVT_power nvoWattage; network output polled eeprom SNVT_elapsed_tm nvoUsage;

The LonTalk Interface Developer utility generates type definitions in the LonNvTypes.h file for all of the above network variables. However, it does not generate type definitions in the LonCpTypes.h file because there are no configuration properties.

In addition to the type definitions and other data, the LonTalk Interface Developer utility generates the following global C variables for this model file:

volatile SNVT_amp nviAmpere; volatile SNVT_volt nviVolt; volatile SNVT_angle nviCosPhi; SNVT_power nvoWattage; SNVT_elapsed_tm nvoUsage;

The declaration of the nvoUsage output network variable uses the network variable modifiers polled and eeprom. The LonTalk Interface Developer utility stores these attributes in the network-variable table (nvTable[]) in the FtxlDev.c

FTXL User’s Guide 49

file. The API uses this table to access the network variables when the application runs. In addition, the application can query the data in this table at runtime.

Functional Blocks without Configuration Properties

The following model file describes a similar meter application as in the previous example, but implements it using functional blocks to provide an interoperable interface:

• A node object based on the the entire device

• An array of three meters, each based on the same user-defined

UFPTenergyMeter

Configuration properties are not used in this example.

// node object: network input SNVT_obj_request nviNodeRequest; network output polled SNVT_obj_status nvoNodeStatus;

fblock SFPTnodeObject { nviNodeRequest implements nviRequest; nvoNodeStatus implements nvoStatus; } NodeObject external_name("NodeObject");

// UFPTenergyMeter // Implements the meter from the previous example. network input SNVT_amp nviAmpere[3]; network input SNVT_volt nviVoltage[3]; network input SNVT_angle nviCosPhi[3]; network output SNVT_power nvoWattage[3]; network output polled eeprom SNVT_elapsed_tm nvoUsage[3];

fblock UFPTenergyMeter { nvoWattage[0] implements nvoWattage; nviAmpere[0] implements nviAmpere; nviVoltage[0] implements nviVoltage; nviCosPhi[0] implements nviCosPhi; nvoUsage[0] implements nvoUsage; } Meter[3] external_name("Meter");

profile, implementing three identical meters.

SFPTnodeObject

functional profile to manage

Because functional blocks only provide logical grouping of network variables and configuration properties, and meaning to those groups, but do not themselves contain executable code, the functional blocks appear only in the selfdocumentation data generated by the LonTalk Interface Developer utility, but not in any generated executable code.

50 Creating a Model File

Functional Blocks with Configuration Network Variables

The following example takes the above example and adds a few configuration properties implemented as configuration network variables. A cp modifier in the network variable declaration makes the network variable a configuration network variable (CPNV). The nv_properties and fb_properties modifiers apply the configuration properties to specific network variables or the functional block.

// Configuration properties for the node object network input cp SCPTlocation nciLocation;

// network variables for the node object network input SNVT_obj_request nviNodeRequest; network output polled SNVT_obj_status nvoNodeStatus;

fblock SFPTnodeObject { nviNodeRequest implements nviRequest; nvoNodeStatus implements nvoStatus; } NodeObject external_name("NodeObject") fb_properties { nciLocation };

// config properties for the Meter network input cp SCPTminSendTime nciMinSendTime[3]; network input cp SCPTmaxSendTime nciMaxSendTime[3]; network input cp UCPTcoupling nciCoupling;

// network variables for the meter network input SNVT_amp nviAmpere[3]; network input SNVT_volt nviVoltage[3]; network input SNVT_angle nviCosPhi[3]; network output SNVT_power nvoWattage[3] nv_properties { nciMinSendTime[0], nciMaxSendTime[0] };

network output polled eeprom SNVT_elapsed_tm nvoUsage;

This example implements two arrays of configuration network variables,

nciMinSendTime

applies to one element of the

nciMinSendTime[0]

array of network variables in turn implements the

FTXL User’s Guide 51

and

nciMaxSendTime

and

nciMaxSentTime[0]

. Each element of these two arrays

nvoWattage

array, starting with

. Each element of the

nvoWattage

member of one

element of the

nvoWattage[0]

Meter

array of functional blocks, again starting with

The user-defined among all three meters, configuring the meters as three single-phase meters or as one three-phase meter in this example. There is only a single configuration property, and it applies to every element of the array of three

UFPTenergyMeter

The LonTalk Interface Developer utility creates a network variable table for the configuration network variables and the persistent

UCPTcoupling

functional blocks.

configuration property

nciCoupling

nvoUsage

is shared

nciCoupling

network variable.

Functional Blocks with Configuration Properties Implemented in a Configuration File

This example implements a device similar to the one in the previous example, with these differences:

1. All configuration properties are implemented within a configuration file instead of as a configuration network variable

2. A

SNVT_address

these files through the direct memory files feature

3. An

// config properties for the node object: SCPTlocation cp_family cpLocation;

// network variables for the node object network input SNVT_obj_request nviNodeRequest; network output polled SNVT_obj_status nvoNodeStatus; const network output polled SNVT_address nvoFileDirectory;

// node object fblock SFPTnodeObject { nviNodeRequest implements nviRequest; nvoNodeStatus implements nvoStatus; nvoFileDirectory implements nvoFileDirectory; } NodeObject external_name("NodeObject") fb_properties { cpLocation };

// config properties for the Meter SCPTminSendTime cp_family cpMinSendTime; SCPTmaxSendTime cp_family cpMaxSendTime; UCPTcoupling cp_family cpCoupling;

SFPTnodeObject

address network variable

type network variable is declared to enable access to

node object has been added to support the SNVT

52 Creating a Model File

network output polled eeprom SNVT_elapsed_tm nvoUsage[3];

The addition of the

SNVT_address

typed network variable

nvoFileDirectory

is important for enabling the direct memory files feature for access to the configuration property files. The LonTalk Interface Developer initializes this network variable’s value correctly, and creates all required structures and code for direct memory file access; see

Using Direct Memory Files

on page 96 for more

information.

Alternatively, you can use the LONWORKS File Transfer Protocol (FTP) to access the file directory and the files in the directory. In this case, you need to implement the network variables and message tags as needed for the implementation of a L application code in your host to implement the protocol. See the

ONWORKS FTP server in the model file, and provide

File Transfer

engineering bulletin at www.echelon.com for more information about the

ONWORKS file transfer protocol.

FTXL User’s Guide 53

Using the LonTalk Interface

Developer Utility

You use the model file, described in Chapter 3, and the LonTalk Interface Developer utility to define the network inputs and outputs for your device, and to create your application’s skeleton framework source code. You use this skeleton application framework as the basis for your FTXL application development.

The utility also generates device interface files that are used by a network management tool when designing a network that uses your device.

This chapter describes how to use the LonTalk Interface Developer utility and its options, and describes the files that it generates and how to use them.

FTXL User’s Guide 55

Running the LonTalk Interface Developer

You use the LonTalk Interface Developer utility to create the application framework files that are required for your FTXL application. The LonTalk Interface Developer utility also generates the device interface files (*.xif and *.xfb) that can be used by network management tools to design a network that uses your device.

To create the device interface data and device interface files for your device, perform the following tasks:

1. Create a model file as described in Chapter

23.

page

2. Start the LonTalk Interface Developer utility: from the Windows Start menu, select Programs → Echelon FTXL Developer’s Kit → LonTalk Interface Developer.

3. In the LonTalk Interface Developer utility, specify the program ID, the model file for the device, and other preferences for the utility. The utility uses this information to generate a number of files that your application uses. See

4. Add the FtxlDev.h ANSI C header file to your FTXL application with an include statement:

#include "FtxlDev.h"

Using the LonTalk Interface Developer Files

Creating a Model File

on page 61.

, on

The LonTalk Interface Developer utility creates the application framework files and copies other necessary files (such as the FTXL LonTalk API files and FTXL LonTalk protocol stack library, libFtxl100.a) to your project directory.

In general, you should limit changes to the LonTalk Interface Developer utilitygenerated files. Any changes that you make will be overwritten the next time you run the utility. However, the LonTalk Interface Developer utility does not overwrite or modify the FTXL LonTalk API files.

After you have created the LonTalk Interface Developer utility-generated files, you need to modify and add code to your application, using the FTXL LonTalk API, to implement desired L

See Chapter about how to use the FXTL LonTalk API calls to implement L

Developing an FTXL Application

ONWORKS functionality into your FXTL application.

Specifying the Project File

From the Welcome to LonTalk Interface Developer page of the utility, you can enter the name and location of a new or existing FTXL project file (.lidprj extension). The LonTalk Interface Developer utility uses this project file to maintain your preferences for this project. The base name of the project file is also used as the base name for the device interface files that the utility generates.

Recommendation: Include a project version number in the name of the project to facilitate version control and project management for your LonTalk Interface Developer projects.

, on page 73, for information

ONWORKS tasks.

The utility creates all of its output files in the same directory as the project file. Your application’s model file does not need to be in this directory; from the

56 Using the LonTalk Interface Developer Utility

utility’s Model File Selection page, you can specify the name and location of the model file.

The location of the LonTalk Interface Developer project file can be the same as your application’s project folder, but you can also generate and maintain the LonTalk Interface Developer’s project in a separate folder, and manually link the latest generated framework with your application by copying or referencing the correct location.

Specifying the FTXL Transceiver Configuration

From the FTXL Transceiver Configuration page of the utility, you can specify the clock speed for the FTXL Transceiver.

Specifying Service Pin Held Events

When you press the local service pin on the device, the FTXL LonTalk protocol stack sends a service-pin message on the L service-pin event to the application. However, when you press service pin on the device, whether the FTXL LonTalk protocol stack sends a service-pin-held event depends on how you configure it.

From the System Preferences page of the utility, you can configure the following behavior for how the FTXL LonTalk API handles a service-pin-held event:

ONWORKS network and signals a

and hold

the local

• Whether the FTXL LonTalk API sends a notification of a service-pin-held event to the application.

• How long the local service pin must be pressed and held before the FTXL LonTalk API notifies the application of the event. You can specify from 1 to 30 seconds, or accept a default of 10 seconds.

Receiving service-pin-held events is optional, and how the application processes the events depends on the requirements of the application. For example, many devices support an emergency recovery feature that is triggered by pressing and holding the service pin for a prolonged amount of time (typically 10 or 20s). Then the device moves to the unconfigured state (that is, calls LonGoUnconfigured()) or uses another method to return to a factory state.

Configuring the FTXL LonTalk Protocol Stack

From the Stack Configuration page of the utility, you can specify override values for the following system definitions:

• The size of the address table (the number of addresses)

• The size of the alias table (the number of aliases)

• The number of receive transaction records

• The number of transmit transaction records

• The maximum lifetime of a transmit transaction

If you do not specify an override value, the LonTalk Interface Developer utility generates appropriate values based on other preferences that you specify for the project.

FTXL User’s Guide 57

Recommendation: Allow the LonTalk Interface Developer utility to calculate appropriate values for the stack configuration.

See

Managing Memory

on page 46 for more information about these values.

Configuring the Buffers

From the Buffer Configuration page of the utility, you can specify the number for each of the following application buffer types:

• Input buffers

• Non-priority output buffers

• Priority output buffers

You can also specify the number of link-layer buffers.

In addition, you can specify both the size and number for the transceiver buffers:

• Input buffers

• Non-priority output buffers

• Priority output buffers

Recommendation: Allow the LonTalk Interface Developer utility to calculate appropriate values for the buffer configuration.

Configuring the Application

From the Application Configuration page of the utility, you can specify the following parameters for the application:

• The number of dynamic network variables

• The average amount of memory to reserve for self-documentation data for

dynamic network variables

By default, the number of supported dynamic network variables is zero, but you can specify up to 4096. During compilation, the utility verifies that the sum of static and dynamic network variables does not exceed a total of 4096 for the device.

The average amount of memory to reserve for dynamic network variable selfdocumation strings is used, along with the number of dynamic network variables, to calculate the maximum amount of non-volatile data that might be required for the FTXL device. The actual size of a particular dynamic variable’s selfdocumentation string can exceed the specified average, as long as the actual average size is less than or equal to the specified average size.

The default size for the dynamic network variable self-documentation data is 16 bytes, but you can specify up to 128 bytes.

Configuring Support for Non-Volatile Data

From the Non-Volatile Data Support page of the utility, you can specify the following parameters for the application:

• The non-volatile data driver model

58 Using the LonTalk Interface Developer Utility

• The non-volatile data flush guard timeout value

• The name for the top-level root segment for the non-volatile data

The non-volatile data driver model can be one of the following types, depending on your application’s requirements:

• Flash file system (such as the Micrium μC/FS embedded file system)

• Flash direct memory (with no file system) if you do not have, or do not

want to use, a flash file system for your non-volatile data

• User defined if you have another non-volatile data support model that

your application uses

You can only select one driver model for the specified application.

The non-volatile data flush timeout value determines how long the FTXL LonTalk protocol stack waits to receive additional updates before writing them to the non-volatile data.

The non-volatile root name is used to configure the non-volatile data support driver. If you use the flash file system, the non-volatile root name is used as a file system directory name in which to create non-volatile data files. If you use the direct flash model, the name represents a Nios II flash device name. If you use unstructured flash memory, leave the Root field blank.

Within the Nios development environment, the system.h file defines the root name. For the examples that are included with the FTXL Developer’s Kit, the root name is /dev/cfi_flash, which is the root directory for the flash file system.

The FTXL source files that handle non-volatile data (FtxlNvdFlashDirect.c, FtxlNvdFlashFs.c, and FtxlNvdUserDefined.c) use conditional compilation based on the selected model to include the appropriate code. If you select a user-defined model, the related callback handler functions are not defined and cause a linker error if they are not implemented.

Specifying the Device Program ID

From the Program ID Selection page of the utility, you specify the device program ID or use the L device program ID. The program ID is a 16-digit hexadecimal number that uniquely identifies the device interface for your device.

The program ID can be formatted as a standard or non-standard program ID. When formatted as a standard program ID, the 16 hexadecimal digits are organized into six fields that identify the manufacturer, classification, usage, channel type, and model number of the device. The L ID Calculator simplifies the selection of the appropriate values for these fields by allowing you to select from lists contained in a program ID definition file distributed by L

ONMARK International. A current version of this list is included

with the FTXL Developer’s Kit.

Within the device’s program ID, you must include your manufacturer ID. If your company is a member of L Manufacturer ID assigned by L within the Standard Program ID Calculator utility, or online at

www.lonmark.org/mid.

ONMARK Standard Program ID Calculator to specify the

ONMARK International, you have a permanent

ONMARK International. You can find those listed

ONMARK Standard Program

FTXL User’s Guide 59

If your company is not a member of the LONMARK International, you can obtain a temporary manufacturer ID from

ONMARK International to obtain a temporary manufacturer ID.

For prototypes and example applications, you can use the F:FF:FF manufacturer ID, but you should not release a device that uses this non-unique identifier into production.

If you want to specify a program ID that does not follow the standard program ID format, you must use the command-line interface for the LonTalk Interface Developer utility. L devices to use a standard-format program ID. Using a non-standard format for the program ID will prevent the use of functional blocks and configuration properties, and will prevent certification.

ONMARK International requires all interoperable LONWORKS

Specifying the Model File

From the Model File Selection page of the utility, you specify the model file for the device. You can also click Edit to open the model file in whatever editor is associated with the .nc file type, for example, Notepad or the NodeBuilder Development Tool.

The model file is a simple source file written using a subset of the Neuron C Version 2.1 programming language. The model file contains declarations of network variables, configuration properties, functional blocks, and their relationships.

www.lonmark.org/mid. You do not have to join

The LonTalk Interface Developer utility uses the information in the model file, combined with other user preferences, to generate the application framework files and the interface files. You must compile and link the application framework files with the host application.

See Chapter model file.

Creating a Model File

, on page 23 for more information about the

Specifying Neuron C Compiler Preferences

From the Neuron C Compiler Preferences page of the utility, you can specify macros for the Neuron C compiler preprocessor and extend the include search path for the compiler.

For the preprocessor macros (#define statements), you can only specify macros that do not require values. These macros are optional. Use separate lines to specify multiple macros.

The _FTXL macro is always predefined by the LonTalk Interface Developer utility, and does not need to be specified explicitly. You can use this macro to control conditional compilation for FTXL applications. In addition, the utility predefines the _MODEL_FILE macro for model file definitions and the _LID3 macro for LonTalk Interface Developer utility macros.

For the search path, you can specify additional directories in which the compiler should search for user-defined include files (files specified within quotation marks, for example, #include "my_header.h").

Specifying additional directories is optional. Use separate lines to specify multiple directories.

60 Using the LonTalk Interface Developer Utility

The LonTalk Interface Developer project directory is automatically included in the compiler search path, and does not need to be specified explicitly. Similarly, the Neuron C Compiler system directories (for header files specified with angled brackets, for example, #include <string.h>) are also automatically included in the compiler search path.

Specifying Code Generator Preferences

From the Interface Developer Code Generator Preferences page of the utility, you can specify preferences for the LonTalk Interface Developer compiler, such as whether to generate verbose source-code comments.

Compiling and Generating the Files

From the Summary and Confirmation page of the utility, you can view all of the information that you specified for the project. When you click Next, the LonTalk Interface Developer utility compiles the model file and generates a number of C source files and header files, as described in

Developer Files

The Build Progress and Summary page shows the results of compilation and generation of the FTXL project files.

Any warning or error messages have the following format:

on page 61.

Using the LonTalk Interface

Error-type: Model_file_name Line_number(Column_number): Message

Example: A model file named “tester.nc” includes the following single network variable declaration:

network input SNVT_volt nviVolt

Note the missing semi-colon at the end of the line. When you use this file to build a project from the LonTalk Interface Developer utility, the compiler issues the following message:

Error: TESTER.NC 1( 32): Unexpected END-OF-FILE in source file [NCC#21]

The message type is error, the line number is 1, the column number is 32 (which corresponds to the position of the error, in this case, the missing semi-colon), and the compiler message number is NCC#21. To fix this error, add a semi-colon to the end of the line.

See the

NodeBuilder Errors Guide

for information about the compiler messages.

Using the LonTalk Interface Developer Files

The LonTalk Interface Developer utility takes all of the information that you provide and automatically generates the following files that are needed for your FTXL application:

• LonNvTypes.h

• LonCpTypes.h

• FtxlDev.h

• FtxlDev.c

FTXL User’s Guide 61

•

project

•

project

These files form the FTXL application framework, which defines the FTXL device initialization data and self-identification data for use in initialization phase, including communication parameters and everything you need to begin device development. The framework includes ANSI C type definitions for network variable and configuration property types used with the application, and implements them as global application variables.

To include these files in your application, include the FtxlDev.h file in your FTXL application using an ANSI C #include statement, and add the FtxlDev.c file to your project so that it can be compiled and linked.

The following sections describe the copied and generated files.

Copied Files

The LonTalk Interface Developer utility copies the following files into your project directory if no file with the same name already exists:

• FtxlApi.h

• FtxlHal.c

• FtxlHal.h

.xif

.xfb

• FtxlHandlers.c

• FtxlNvdFlashDirect.c

• FtxlNvdFlashFs.c

• FtxlNvdUserDefined.c

• FtxlOsal.c

• FtxlOsal.h

• FtxlTypes.h

• libFtxl100.a

• LonPlatform.h

Existing files with the same name, even if they are not write-protected, are not overwritten by the utility.

Other than FtxlDev.h, you do not normally have to explicitly include any of the header files with your application source, because the FtxlDev.h file already includes the required files.

You must ensure that the libFtxl100.a library and the various C files are available to your project so that they can be compiled and linked with your application.

LonNvTypes.h and LonCpTypes.h

The LonNvTypes.h file defines network variable types, and includes type definitions for standard or user network variable types (SNVTs or UNVTs). See

Using Types

62 Using the LonTalk Interface Developer Utility

on page 64 for more information on the generated types.

The LonCpTypes.h file defines configuration property types, and includes standard or user configuration property types (SCPTs or UCPTs) for configuration properties implemented within configuration files.

Either of these files might be empty if your application does not use either network variables or configuration properties.

FtxlDev.h

The FtxlDev.h file is the main header file that the LonTalk Interface Developer utility produces. This file provides the definitions that are required for your application code to interface with the application framework and the FTXL LonTalk API, including C extern references to public functions, variables, and constants generated by the LonTalk Interface Developer utility.

You should include this file with all source files that your application uses, but you do not normally have to edit this file. Any manual changes to this file are not retained when you rerun the LonTalk Interface Developer utility. The file contains comments that describe how you can override some of the preferences and assumptions made by the utility.

FtxlDev.c

The FtxlDev.c file is the main source file that the LonTalk Interface Developer utility produces. This file includes the FtxlDev.h file header file, declares the network variables, configuration properties, and configuration files (where applicable).

It defines the device’s LonInit() function. It also defines variables and constants, including the network variable table, the device’s initialization data block, and a number of utility functions.

You must compile and link this file with your application, but you do not normally have to edit this file. Any manual changes to this file are not retained when you rerun the LonTalk Interface Developer utility, but the file contains comments that describe how you can override some of the preferences and assumptions made by the utility.

project.xif and project.xfb

The LonTalk Interface Developer utility generates the device interface file for your project in two formats:

project

•

project

For both files, Welcome to LonTalk Interface Developer window of the LonTalk Interface Developer utility. Thus, these files have the same name as the FTXL project file (.lidprj extension).

These files comply with the L

Important: If your device is defined with a non-standard program ID, the device interface file cannot contain interoperable LONMARK constructs.

.xif (a text file)

.xfb (a binary file)

project

is the name of the FTXL project that you specified in the

ONMARK device interface revision 4.401 format.

FTXL User’s Guide 63

Using Types

The LonTalk Interface Developer utility produces type definitions for the network variables and configuration properties in your model file. For maximum portability, all types defined by the utility are based on a small set of host-side equivalents to the built-in Neuron C types. For example, the LonPlatform.h file contains a type definition for a Neuron C signed integer equivalent type called ncsInt. This type must be the equivalent of a Neuron C signed integer, a signed 8-bit scalar. For most target platforms, the ncsInt type is defined as signed char type.

A network variable declared by a Neuron C built-in type does not require a hostside type definition in the LonNvTypes.h file, but is instead declared with its respective host-side Neuron C equivalent type as declared in LonPlatform.h.

Important: Network variables that use ordinary C types, such as int or long, are not interoperable. For interoperability, network variables must use types defined within the device resource files. These network variable types include standard network variable types (SNVTs) and user-defined network variable types (UNVTs). You can use the Resource Editor tool to define your own UNVT.

Example:

A model file contains the following declarations:

network input int nviInteger; network output SNVT_count nvoCount; network output SNVT_switch nvoSwitch;

• The nviInteger declaration uses a built-in Neuron-C type, so the LonTalk Interface Developer utility uses the ncsInt type defined in LonPlatform.h.

• The nvoCount declaration uses a type that is not a built-in Neuron C type. The utility produces the following type definition:

typedef ncuLong SNVT_count;

The ncuLong type represents the host-side equivalent of a Neuron C unsigned long, a 16-bit unsigned scalar. It is defined in LonPlatform.h, and typically maps to the LonWord type. LonWord is a platformindependent definition of a 16-bit scalar in big-endian notation:

typedef struct { LonByte msb; LonByte lsb; } LonWord;

To use this platform-independent type for numeric operations, you can use the optional LON_GET_UNSIGNED_WORD or LON_SET_UNSIGNED_WORD macros. Similar macros are provided for signed words (16 bit), and for signed and unsigned 32-bit scalars (DOUBLE).

Important: If a network variable or configuration property is defined with an initializer in your device’s model file, and if you change the default definition of multibyte scalars (such as the ncuLong type), you must modify the initializer generated by the LonTalk Interface Developer utility if the type is a multibyte scalar type.

64 Using the LonTalk Interface Developer Utility

• The nvoSwitch declaration is based on a structure. The LonTalk Interface Developer utility redefines this structure using built-in Neuron C equivalent types:

typedef LON_STRUCT_BEGIN(SNVT_switch){ ncuInt value; ncsInt state; } LON_STRUCT_END(SNVT_switch);

Type definitions for structures assume a padding of 0 (zero) bytes and a packing of 1 byte. The LON_STRUCT_BEGIN and LON_STRUCT_END macros enforce platform-specific byte packing and padding. These macros are defined in the LonPlatform.h file, which allows you to adjust them for your compiler.

Bit Field Members

For portability, none of the types that the LonTalk Interface Developer utility generates use bit fields. Instead, the utility defines bit fields with their enclosing bytes, and provides macros to extract or manipulate the bit field information.

By using macros to work directly with the bytes of the bit field, your code is portable to both big-endian and little-endian platforms (that is, platforms that represent the most-significant bit in the left-most position and platforms that represent the most-significant bit in the right-most position). The macros also reduce the need for anonymous bit fields to achieve the correct alignment and padding.

Example: The following macros and structure define a simple bit field of two flags, a 1-bit flag alpha and a 4-bit flag beta:

typedef LON_STRUCT_BEGIN(Example) { LonByte flags_1; // contains alpha, beta } LON_STRUCT_END(Example);

#define LON_ALPHA_MASK 0x80 #define LON_ALPHA_SHIFT 7 #define LON_ALPHA_FIELD flags_1 #define LON_BETA_MASK 0x70 #define LON_BETA_SHIFT 4 #define LON_BETA_FIELD flags_1

When your program refers to the flags_1 structure member, it can use the bit mask macros (LON_ALPHA_MASK and LON_BETA_MASK), along with the bit shift values (LON_ALPHA_SHIFT and LON_BETA_SHIFT), to retrieve the two flag values. These macros are defined in the LonNvTypes.h file. The LON_STRUCT_* macros enforce platform-specific byte packing.

To read the alpha flag, use the following example assignment:

Example var; alpha_flag = (var.LON_ALPHA_FIELD & var.LON_ALPHA_MASK) >> var.LON_ALPHA_SHIFT;

You can also use the LON_GET_ATTRIBUTE() and LON_SET_ATTRIBUTE()

var

macros to access flag values. For example, for a variable named these macros to get or set the attributes:

alpha_flag = LON_GET_ATTRIBUTE(var, LON_ALPHA); …

, you can use

FTXL User’s Guide 65

LON_SET_ATTRIBUTE(var, LON_ALPHA, alpha_flag);

These macros are defined in the FtxlTypes.h file.

Enumerations

The LonTalk Interface Developer utility does not produce enumerations. FTXL requires an enumeration to be of size byte. The ANSI C standard requires that an enumeration be an int, which is larger than one byte for many platforms.

An FTXL enumeration uses the LON_ENUM_BEGIN and LON_ENUM_END macros. For many compilers, these macros can be defined to generate native enumerations:

#define LON_ENUM_BEGIN(name) enum #define LON_ENUM_END(name) name

Some compilers support a colon notation to define the enumeration’s underlying type:

#define LON_ENUM_BEGIN(name) enum : signed char #define LON_ENUM_END(name)

When your program refers to an enumerated type in a structure or union, it should not use the enumeration’s name, but should use the LON_ENUM_* macros.

For those compilers that support byte-sized enumerations, it can be defined as:

#define LON_ENUM(name) name

For other compilers, it can be defined as:

#define LON_ENUM(name) signed char

Example: LON_ENUM_* macros, and the equivalent ANSI C enumeration.

FTXL Enumeration Equivalent ANSI C Enumeration

LON_ENUM_BEGIN(Color) {

} LON_ENUM_END(Color);

typedef struct {

} Example;

Table 7 shows an example enumeration using the FTXL

Table 7. Enumerations in FTXL

red, green, blue

… LON_ENUM(Color) color;

Floating Point Variables

Floating point variables receive special processing, because the Neuron C compiler does not have built-in support for floating point types. Instead, it offers an implementation for floating point arithmetic using a set of floating-point support functions operating on a float_type type. The LonTalk Interface Developer utility represents this type as a float_type structure, just like any other structured type.

enum {

red, green, blue

} Color;

typedef struct {

… Color color;

} Example;

66 Using the LonTalk Interface Developer Utility

This floating-point format can represent numbers with the following characteristics:

1038

•

10*1± approximate maximum value

approximate relative resolution

10*1−±

The float_type structure declaration represents a floating-point number in IEEE 754 single-precision format. This format has one sign bit, eight exponent bits, and 23 mantissa bits; the data is stored in big-endian order. The float_type type is identical to the type used to represent floating-point network variables.

For example, the LonTalk Interface Developer utility generates the following definitions for the floating point type SNVT_volt_f:

/* * Type: SNVT_volt_f */ typedef LON_STRUCT_BEGIN(SNVT_volt_f) { LonByte Flags_1; /* Use bit field macros, defined below */ LonByte Flags_2; /* Use bit field macros, defined below */ ncuLong LS_mantissa; } LON_STRUCT_END(SNVT_volt_f);

/* * Macros to access the sign bit field contained in * Flags_1 */ #define LON_SIGN_MASK 0x80 #define LON_SIGN_SHIFT 7 #define LON_SIGN_FIELD Flags_1

/* * Macros to access the MS_exponent bit field contained in * Flags_1 */ #define LON_MSEXPONENT_MASK 0x7F #define LON_MSEXPONENT_SHIFT 0 #define LON_MSEXPONENT_FIELD Flags_1

/* * Macros to access the LS_exponent bit field contained in * Flags_2 */ #define LON_LSEXPONENT_MASK 0x80 #define LON_LSEXPONENT_SHIFT 7 #define LON_LSEXPONENT_FIELD Flags_2

/* * Macros to access the MS_mantissa bit field contained in * Flags_2 */ #define LON_MSMANTISSA_MASK 0x7F #define LON_MSMANTISSA_SHIFT 0

FTXL User’s Guide 67

#define LON_MSMANTISSA_FIELD Flags_2

See the IEEE Standard for Binary Floating-Point Arithmetic (ANSI/IEEE Std 754-1985) documentation for more information.

Network Variable and Configuration Property Declarations

The LonTalk Interface Developer utility generates network variables and configuration properties using the built-in types defined in LonPlatform.h along with the types defined in LonNvTypes.h and LonCpTypes.h. Both network variables and configuration properties are declared in the FtxlDev.c file, where input network variables (including configuration network variables) appear as volatile variables of the relevant type, and configuration properties that are not implemented with network variables appear as members of configuration files.

Example:

A model file contains the following Neuron C declarations:

SCPTlocation cp_family cpLocation;

network input SNVT_obj_request nviNodeRequest; network output polled SNVT_obj_status nvoNodeStatus; const network output polled SNVT_address nvoFileDir;

fblock SFPTnodeObject { nviNodeRequest implements nviRequest; nvoNodeStatus implements nvoStatus; nvoFileDir implements nvoFileDirectory; } NodeObject external_name("NodeObject") fb_properties { cpLocation };

The LonTalk Interface Developer utility generates the following variables in the FtxlDev.c file for the nviNodeRequest, nvoNodeStatus, and nvoFileDir network variables:

volatile SNVT_obj_request nviNodeRequest; SNVT_obj_status nvoNodeStatus; SNVT_address nvoFileDir = { LON_DMF_WINDOW_START/256u, LON_DMF_WINDOW_START%256u };

The FTXL LonTalk API, upon receipt of an incoming network variable update, automatically moves data into the corresponding input network variable and signals this event by calling an event handler function, which allows your application to respond to the arrival of new network variable data. Your application then reads the input variable to obtain the latest value.

To send an update to the nvoNodeStatus output network variable, your application writes the new value to the nvoNodeStatus variable, and then calls the LonPropagateNv() function to propagate the new value onto the network.

Developing an FTXL Application

See development of a FTXL application using the LonTalk Interface Developer utility-generated code.

on page 73 for information about the

68 Using the LonTalk Interface Developer Utility

The utility generates a configuration file in FtxlDev.c for the cpLocation configuration property:

/* * * Writable configuration parameter value file */ volatile LonWriteableValueFile lonWriteableValueFile = { {{'\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0', '\x0'}} };

/* * CP template file */ const char lonTemplateFile[] = \ "1.1;" \ "1,0,0\x80,17,31;";

#ifndef LON_FILEDIR_USER_DEFINED /* * Variable: File Directory */

const LonFileDirectory lonFileDirectory = { LON_FILE_DIRECTORY_VERSION, LON_FILE_COUNT, { LON_REGISTER_FILE("template", sizeof(lonTemplateFile), LonTemplateFileType, LON_DMF_WINDOW_START+sizeof(lonFileDirectory)), LON_REGISTER_FILE("rwValues", sizeof(lonWriteableValueFile), LonValueFileType, LON_DMF_WINDOW_START+sizeof(lonFileDirectory) +sizeof(lonTemplateFile)), LON_REGISTER_FILE("roValues", 0, LonValueFileType,

} }; #endif /* LON_FILEDIR_USER_DEFINED */

The LonWriteableValueFile data structure is defined in the FtxlDev.h header file:

typedef LON_STRUCT_BEGIN(LonWriteableValueFile) { SCPTlocation cpLocation_1; /* sd_string("1,0,0\x80,17,31;") */ } LON_STRUCT_END(LonWriteableValueFile);

extern volatile LonWriteableValueFile lonWriteableValueFile;

FTXL User’s Guide 69

Similarly, a LonReadOnlyValueFile type is defined and used to declare a lonReadOnlyValueFile variable if the model file declares read-only configuration properties.

The LonTalk Interface Developer utility generates resource definitions for configuration properties and network variables defined with the eeprom keyword. Your application must provide sufficient persistent storage for these resources. You can use any type of non-volatile memory, or any other media for persistent data storage. The template file and the read-only value file would normally be declared as const, and can be linked into a code segment, which might relate to non-modifiable memory such as PROM or EPROM (these files must not be changed at runtime). However, writable, non-volatile storage must be implemented for the writable configuration property value file.

The details of such persistent storage are subject to the host platform requirements and capabilities; persistent storage options include: flash memory, EEPROM memory, non-volatile RAM, or storage in a file or database on a hard drive.

You can specify initializers for network variables or configuration properties in the model file. Alternatively, you can specify initializers for configuration properties in the resource file that defines the configuration property type or functional profile. For network variables without explicit initialization, the rules imposed by your host development environment apply. These values might have random content, or might automatically be preset to a well-defined value.

Constant Configuration Properties

In general, a configuration property can be modifiable, either from within the FTXL application or from a network management tool. However, the LonTalk Interface Developer utility declares constant configuration property files as constants (using the C const keyword), so that they are allocated in nonmodifiable memory.

A special class of configuration properties is the property. A device-specific configuration property is considered variable to the application (that is, your application can change it), but constant to the external interface. These properties might, for example, be used to store calibration data that is gathered during the device’s auto-tuning procedure.

However, a paradox arises because the network manager expects this configuration property within the read-only value file, but the read-only value file must be writable from the local application. This paradox is known as the writeable read-only value file.

FTXL presents the following solution to resolve this paradox:

• Before the inclusion of the FtxlDev.h header file into the FtxlDev.c file, you can define the LON_READONLY_FILE_IS_WRITEABLE macro to a value of 1 (one). If you do not define this macro, or define it to equate to zero, the read-only value file is constant. This is the default state. The LON_READONLY_FILE_IS_WRITEABLE macro is used within the FtxlDev.h header to define the read-only file’s storage type with the LON_READONLY_FILE_STORAGE_TYPE macro, which in turn is used in declaration and specification of the lonReadOnlyValueFile variable.

device-specific

configuration

70 Using the LonTalk Interface Developer Utility

• Defining the LON_READONLY_FILE_IS_WRITEABLE macro to 1 causes the read-only value file to be writeable by the local application. Because it is now allocated in volatile memory, your driver for nonvolatile data must also be able to read and write the read-only value file.

For the network management tool, however, the read-only file remains nonwriteable. If your application uses the direct memory files feature to access the files, the LonTalk Interface Developer utility generates code that declares this direct memory files window segment as non-modifiable. If your application uses

ONWORKS FTP to access the files, your implementation of the LONWORKS file

L transfer protocol and server must prevent write operations to the read-only value file under all circumstances.

The Network Variable Table

The network variable table lists all the network variables that are defined by your application. It contains a pointer to each network variable and the initial (or declared) length of each network variable, in bytes. It also contains an attribute byte that contains flags which define the characteristics of each network variable.

The network variable table acts as a bridge between your application and the FTXL LonTalk API. The LonTalk Interface Developer utility generates the network variable table, along with the LonInit() function that reads the table and register the network variables with the FTXL LonTalk API.

An FTXL application typically accesses a network variable value through the C global variable that implements the network variable. However, the FTXL LonTalk API also provides a function that returns the pointer to a network variable’s value as a function of its index:

void* const LonGetNvValue(unsigned index);

You can use this function for any network variable, including static network variables, dynamic network variables, and configuration property network variables. The LonGetNvValue() function returns NULL for an invalid index, or returns a pointer to the value.

For dynamic network variables, you must use the LonGetNvValue() function because there is no global C variable or network variable table entry for a dynamic network variable.

Network Variable Attributes

The network variable table (nvTable[])in the FtxlDev.c file includes a bitmask for each network variable to define the network variable’s attributes, including, for example, whether the network variable is:

• An output network variable

• Persistent

• Polled

• Synchronous

• Of changeable type

FTXL User’s Guide 71

The FtxlTypes.h file defines the bitmasks for these attributes. For example, LON_NV_IS_OUTPUT is the mask for an output network variable, LON_NV_POLLED is the mask for a polled network variable, and so on.

The FTXL LonTalk API does not propagate a polled output network variable's value to the network when your application calls the LonPropagateNv() function. For input network variables, the polled attribute changes the behavior of the network management tool’s binder, which determines how a network variable connection is managed.

Developing an FTXL Application

See propagation of network variable updates.

The Message Tag Table

Although the FTXL LonTalk protocol stack does not use the message tag table, the LonTalk Interface Developer utility declares the message tag table in FtxlDev.c if you declare one or more message tags in the model file.

The message tag table lists all the message tags that are defined by your application. It contains a flag for each message tag which indicates that the message tag is not associated with an address table entry and therefore can only be used for sending explicitly addressed application messages. This flag is set for all message tags declared with the bind_info(nonbind) modifier in the model file.

Communicating with Other Devices Using Application Messages

See for more information about using message tags.

on page 73 for more information about

on page 93

72 Using the LonTalk Interface Developer Utility

Developing an FTXL Application

This chapter describes how to develop an FTXL application. It also describes the various tasks performed by the application.

FTXL User’s Guide 73

Overview of an FTXL Application

This chapter describes how to use the FTXL LonTalk API and the application framework produced by the LonTalk Interface Developer utility to perform the following tasks:

• Use the FTXL LonTalk API and FTXL LonTalk protocol stack

• Integrate the application with an operating system

• Provide persistent storage for non-volatile data

• Initialize the FTXL device

• Periodically call the FTXL event pump

• Send information to other devices using network variables

• Receive information from other devices using network variables

• Handle network variable poll requests from other devices

• Handle updates to changeable-type network variables

• Handle dynamic network variables

• Communicate with other devices using application messages

• Handle management tasks and events

• Handle local network management commands

• Handle reset events

• Query the error log

• Use the direct memory files feature

• Shut down the FTXL device

Most FTXL applications need to perform only the tasks that relate to persistent storage, initialization, calling the event pump, and sending and receiving network variables.

This chapter shows you the basic control flow for each of the above tasks. It also provides a simple code example to illustrate some of the basic tasks.

Using the FTXL LonTalk API

Within the seven-layer OSI Model protocol, the FTXL LonTalk API forms the majority of the Presentation layer, and provides the interface between the FTXL LonTalk protocol stack in the Session layer and the host application in the Application layer, as shown in

Figure 9 on page 75.

74 Developing an FTXL Application

Host Application

FtxlApi.h FtxlDev.h

Application Framework

FTXL LonTalk API

FTXL LonTalk Protocol Stack

FTXL

Operating System Abstraction Layer

Hardware Abstraction Layer

Transceiver

ANSI/CEA-709.1 control network

Figure 9. The FTXL LonTalk API within the OSI Model

The [

FTXL

]\Core\libFtxl100.a file contains the FTXL LonTalk protocol stack, the FTXL LonTalk API, and the parallel interface driver, which together allow your FTXL application to handle network events, propagate network variables, respond to network variable poll requests, and so on.

An FTXL application must include the FtxlDev.h file to be able to use the FTXL LonTalk API. This file is generated by the LonTalk Interface Developer utility, and is located in your application project directory. The FtxlDev.h file includes

FTXL

the [

]\Core\FtxlApi.h file, which contains definitions for accessing the

FTXL LonTalk API.

The [

FTXL

]\Core\FtxlHandlers.c source file contains stubs for the event handler functions and callback handler functions that the FTXL LonTalk API calls. You must add code to these stubs to respond to specific events. For example, the LonNvUpdateOccurred() event handler function could inform the application of the arrival of new data for a set-point value, and the related code could recalculate the device’s response, assign output values to peripheral I/O devices, update the appropriate network variables, and propagate the changes to the network.

The following recommendations can help you manage your FTXL application project:

• Keep edits to LonTalk Interface Developer utility-generated files to a

minimum, that is, do not edit the LonNvTypes.h, LonCpTypes.h, FtxlDev.h or FtxlDev.c files unless necessary

• Add #include “FtxlDev.h” to your application source files to provide access

to network variable types and instantiations and the FTXL LonTalk API

• Keep changes to the FtxlHandlers.c file to a minimum

o Add calls to your own functions in files that you create and

maintain

FTXL User’s Guide 75

o Future versions or fixes to the FTXL product might affect these

API files

Callbacks and Events

The FTXL LonTalk API uses two types of notifications for occurrences within the system: callbacks and events.

The FTXL LonTalk API uses a the application immediately. A callback can occur in one of the FTXL LonTalk protocol stack contexts (tasks or threads).

When you implement a callback handler function to process a callback, you must ensure that the function completes its work as quickly as possible. Generally, a callback handler function must not call FTXL LonTalk API functions or perform time-intensive operations.

The FTXL LonTalk API uses an application. The protocol stack does not wait for the processing of the event to complete before continuing.

The FTXL LonTalk protocol stack holds events in an internal queue for processing. Thus, the application program must periodically call the LonEventPump() function to process the event queue. This function also calls the related event handler functions.

Because event processing in the event handler functions is not tied to the context of the protocol stack, an event handler function can call FTXL LonTalk API functions or perform time-intensive operations. An event handler function runs within the same context (task or thread) as its caller (the LonEventPump() function).

See Appendix handler functions and event handler functions.

FTXL LonTalk API

callback

event

when the API needs a return value from

to deliver a one-way notification to the

, on page 147, for a list of the callback

Integrating the Application with an Operating System

The FTXL LonTalk protocol stack requires an FTXL application to use an operating system. The FTXL Developer’s Kit includes example applications that use the Micrium μC/OS-II operating system, but you can use any embedded operating system that meets your application’s requirements. Although the μC/OS-II operating system is a real-time operating system, the FTXL LonTalk protocol stack does not require the operating system to be a real-time operating system.

To allow the FTXL LonTalk protocol stack to use any operating system, the FTXL LonTalk protocol stack library is linked with the FTXL Operating System Abstraction Layer (OSAL) files, FtxlOsal.h and FtxlOsal.c. The FTXL OSAL files provide macros and C functions for general operating system functions, such as creating semaphores and waiting for events. The FTXL OSAL functions also include error handling and basic debug tracing for the operating system functions.

Your FTXL application can call the FTXL OSAL functions when it needs to call operating system functions, or it can call the operating system functions directly. By calling FTXL OSAL functions, your FTXL application can be more easily ported to another operating system, if needed.

76 Developing an FTXL Application

The FTXL OSAL function prototypes are generic, and do not depend on the operating system’s syntax. For example, to create a binary semaphore, your application can call the OsalCreateBinarySemaphore() function, which in turn calls the operating system’s function to create the semaphore. The FTXL OSAL function assigns a pointer to the created semaphore and returns a status variable that indicates whether the function was successful.

The FTXL Developer’s Kit includes source code for FTXL OSAL files that use the syntax of the Micrium μC/OS-II operating system. To use a different operating system, you must modify the OSAL files to implement the API for that operating system.

For more information about the FTXL OSAL functions, see

System Abstraction Layer

operating system, see

on page 156. For information about configuring the

Configuring the Operating System

Configuring the Micrium μC/OS-II Operating System

The FTXL Operating

on page 160 and see

on page 165.

Providing Persistent Storage for Non-Volatile Data

The FTXL LonTalk protocol stack provides an API for managing non-volatile data (NVD). Because non-volatile data is stored and managed by the host processor rather than the FTXL Transciever, the FTXL application must implement the API’s functions so that both the FTXL LonTalk protocol stack and the application can read and write NVD to non-volatile memory (typically, flash memory). Two example implementations, one using a flash file system, and one using raw flash access (through the HAL flash access routines) are provided in the FtxlNvdFlashDirect.c and FtxlNvdFlashFs.c files.

The implementations of the NVD-management functions are contained in one of the following files (all of which are copied to the project directory by the LonTalk Interface Developer utility):

• FtxlNvdFlashDirect.c for direct-access flash memory management

• FtxlNvdFlashFs.c for file-system flash memory management

• FtxlNvdUserDefined.c for your own flash memory management

Typically, if you select either the direct flash model or the flash file system model, you need only specify the appropriate value for the non-volatile root in the LonTalk Interface Developer Utility. This section describes how the FTXL API uses the non-volatile memory driver, in case you need to implement your own user-defined non-volatile data driver or modify one of the provided drivers.

Non-volatile data is stored in segments. Two of the segments are used to store data maintained by the FTXL LonTalk protocol stack, and the third segment is used to store data maintained by the application. Examples of data maintained by the FTXL LonTalk protocol stack include network variable configuration and address tables. Examples of data maintained by the application include configuration network variable values and persistent memory files (used for configuration property value files and user files). Each data segment is identified by an enumeration of type LonNvdSegmentType, defined in the FtxlTypes.h file.

The FTXL LonTalk protocol stack reads non-volatile data (loads it into RAM) only during device initialization. Included with the data is a header that the FTXL LonTalk protocol stack uses for validation. Within this header is an application identifier, generated by the LonTalk Interface Developer utility, that allows the FTXL LonTalk protocol stack to ensure that the data belongs to the

FTXL User’s Guide 77

current application. The header also includes a checksum to ensure that the data is free of errors. If any of these validations fails, the FTXL LonTalk protocol stack deletes all non-volatile data in the segment and sets the device to the unconfigured state.

When data that must be stored persistently is updated in RAM, the FTXL LonTalk protocol stack does not immediately update the corresponding persistent memory. Instead, the FTXL LonTalk protocol stack defers writing the data to persistent memory so that it can continue to respond to network management commands in a timely fashion. The reasons for deferred writing of persistent memory include:

• Flash sectors sizes tend to be large and can take a long time to write.

• Each network management update generally affects only a small amount

of data, and typically, a single logical operation consists of many messages (commissioning of the device generally being the most common and most extensive).

• The FTXL LonTalk protocol stack supports large configurations.

If the FTXL LonTalk protocol stack has not received any updates to a particular segment for a short (configurable) time (for example, 1 second), it uses the application callback handler functions to write the data to persistent memory. If the FTXL LonTalk protocol stack is shut down by calling the LonExit() function, the FTXL LonTalk protocol stack completes the write process before returning from the function. However, a sudden power outage or an unexpected CPU reset can prevent an orderly shutdown. The FTXL LonTalk protocol stack maintains a set of flags (one for each segment) that survive an unorderly shutdown so that the FTXL LonTalk protocol stack can detect the unorderly shutdown at the next restart.

The FTXL LonTalk protocol stack checks the flag, by calling the LonNvdIsInTransaction() callback handler function, during device startup before it reads the non-volatile data. If the flag is set, integrity of the non-volatile data has been compromised. Even if the configuration is internally consistent, the FTXL device has likely lost updates from a network manager that it has already acknowleged. If the FTXL device reverted to the last known configuration, this inconsistency would likely be undetected and could result in errors that are difficult to isolate. Instead, the FTXL LonTalk protocol stack deletes the configuration data, logs a configuration checksum error, and goes unconfigured. You can restore the configuration by recommissiong the device from network management tool.

If you use either of the standard non-volatile drivers, you can enable tracing by setting the global variable nvdTraceEnabled to a non-zero value. If create your own custom non-volatile data driver, be sure to add some tracing capability to it.

Restoring Non-Volatile Data

During device startup, the FTXL LonTalk protocol stack reads the non-volatile data for each segment and initializes the corresponding data structures stored in RAM by performing the following steps:

1. Calling the LonNvdIsInTransaction() callback handler function. The application returns whether an NVD transaction for this segment was in progress when the FTXL LonTalk protocol stack was stopped. Typically,

78 Developing an FTXL Application

this function returns FALSE, but if the device was reset while a transaction was in progress, this function returns TRUE and the nonvolatile data segment is considered corrupt, so the restore fails.

2. Calling the LonNvdOpenForRead() callback handler function to open the segment that corresponds to the specified type.

3. Calling the LonNvdRead() callback handler function to read the header of the NVD image. This function verifies the header and, if it is valid, uses the size information in the header to allocate the appropriate buffers.

4. Calling the LonNvdRead() callback handler function again (perhaps many times) to read the entire configuration and de-serialize the image.

5. Deserializing the image and updating the FTXL LonTalk protocol stack’s control structures.

6. Calling the LonNvdClose() callback handler function to close the file.

If, at any time during this process any error occurs, the FTXL LonTalk protocol stack sets the device to the unconfigured state, generates a configuration checksum error, and calls the LonNvdDelete() callback handler function.

The FTXL LonTalk protocol stack handles the deserialization of the data for the LonNvdSegNetworkImage and LonNvdSegNodeDefinition segments, but not for the application-defined LonNvdSegApplicationData segment. Instead, the FTXL LonTalk protocol stack calls the LonNvdDeserializeSegment() callback handler function during step segment. The LonNvdDeserializeSegment() callback handler function is generated by the LonTalk Interface Developer utility.

5 above when it processes the LonNvdSegApplicationData

Writing Non-Volatile Data

When the FTXL LonTalk protocol stack processes a network management message that affects any of its configuration data, the FTXL LonTalk protocol stack checks whether there is an NVD transaction for the affected segment. If not, FTXL LonTalk protocol stack starts a timer and calls the LonNvdEnterTransaction() callback handler function for the segment. If there is already a transaction pending, the FTXL LonTalk protocol stack simply resets the timer.

When the timer expires, the FTXL LonTalk protocol stack writes the data to persistent memory by performing the following steps:

1. Determining the size of the serialized image.

2. Allocating a buffer large enough to hold the serialized image.

3. Serializing the data.

4. Calling the LonNvdOpenForWrite() callback handler function to open the segment with write access. If the segment does not already exist, this function must create it. If the segment exists, but is the wrong size, the application might need to delete it before writing to it.

5. Calling the LonNvdWrite() callback hanlder function one or more times to write the image.

6. Calling the LonNvdClose() callback handler function to close the file.

FTXL User’s Guide 79

7. Calling the LonNvdExitTransaction() callback handler function to clear the transaction.

8. Freeing the buffer that contains the serialized image.

The FTXL LonTalk protocol stack determines the size of the serialized image and handles the serialization of the data for the LonNvdSegNetworkImage and LonNvdSegNodeDefinition segments, but not for the application-defined LonNvdSegApplicationData segment. Intstead, theFTXL LonTalk protocol stack calls the LonNvdGetApplicationSegmentSize() callback handler function in step above, and the LonNvdSerializeSegment() callback handler function during step 3 above when it processes the LonNvdSegApplicationData segment. Both of these callback handler functions are generated by the LonTalk Interface Developer utility.

The FTXL LonTalk protocol stack uses a low-priority operating system task or thread (typically lower than the application task) to write NVD to persistent memory. By using a low-priority task or thread, writing NVD should not block the running of the application or the FTXL LonTalk protocol stack. In addition, FTXL LonTalk protocol stack ensures that these NVD-management functions are never called by more than one task or thread at a time.

The application can update configuration network variables (CPNVs) and user files directly, without the FTXL LonTalk protocol stack’s knowledge. The application must inform the FTXL LonTalk protocol stack when this occurs so that the FTXL LonTalk protocol stack can manage the write transaction. Thus, the application should call the LonNvdAppSegHasBeenUpdated() function to initiate an NVD transaction for the application segment.

Tasks Performed by an FTXL Application

The main() function of an FTXL application typically performs only the following actions:

1. Creates one or more operating system contexts (tasks or threads)

2. Starts the operating system (if it is not already started)

Within one of the newly created tasks, the application life cycle includes two phases:

• Initialization

• Normal processing

The initialization phase of an FTXL application includes a call to the LonInit() API function to initialize the FTXL LonTalk protocol stack and the FTXL Transceiver. The initialization phase defines basic parameters for L network communication, such as the communication parameters for the physical transceiver in use, and defines the application’s external interface: its network variables, configuration properties, and self-documentation data. Successful completion of the initialization phase causes the FTXL Transceiver to leave Quiet mode, after which it can send and receive messages over the network. During the initialization phase, the application also creates at least one operating system event (or other protected shared resource).

During normal processing, which is often implemented within an infinite loop, the application waits for an operating system event whenever it is not busy. When the event occurs, the application calls the LonEventPump() API function to

ONWORKS

80 Developing an FTXL Application

process FTXL events. This function then calls event handler functions (such as LonNvUpdateOccurred() or LonNvUpdateCompleted()).

The following sections describe the tasks that an FTXL application performs during its life cycle.

Initializing the FTXL Device

Before your application initializes the FTXL LonTalk protocol stack, it must initialize the C runtime environment and the operating system.

Your application must call the LonInit() function once during device startup. The implementation of this function is generated by the LonTalk Interface Developer utility, and is included in the FtxlDev.c file. This function initializes the FTXL LonTalk API, the FTXL LonTalk protocol stack, and the FTXL Transceiver. The main application thread must call this function before it calls any other FTXL LonTalk API functions.

The LonInit() registers the FTXL device interface data with the FTXL LonTalk protocol stack. This data defines the network parameters and device interface. If your application needs to change the network parameters or change the device interface, it can call the LonExit() function to shut down the FTXL LonTalk protocol stack, and then call the LonInit() function to restart the protocol stack with the updated interface.

Add a call the LonInit() function to the beginning of the application’s main thread. If this function is successful, your application can begin normal operations, including calling the event pump, as described in

the Event Pump

Periodically Calling

Example:

void myMainThread(void) { LonApiError sts; sts = LonInit(); if (sts == LonApiNoError) { // begin normal operations } }

Periodically Calling the Event Pump

As described in periodically call the LonEventPump() function to check if there are any

ONWORKS events to process. This function calls specific API functions based on

L the type of event, then calls event handler functions to notify the application layer of these network events. You can call this function from the idle loop within the main application thread or from any point in your application that is processed periodically. However, you must call this function from the same application context (task or thread) that called the the LonInit() function.

The FTXL LonTalk API calls the LonEventReady() callback handler function whenever an event has been posted. This function is typically called from an FTXL LonTalk protocol stack task or thread, and you must not call the LonEventPump() function directly from the callback. However, your application could define an operating system event which is signaled by the LonEventReady() callback handler function. From within your application’s

Callbacks and Events

on page 76, your FTXL application must

FTXL User’s Guide 81

main thread, the application should implement an infinite loop that waits on this operating system event. Whenever the event is signaled, the application should call the LonEventPump() API function to process FTXL events.

You can signal this same operating system event to schedule your main application thread to perform other functions as well. For example, you could signal the operating system event from within an interrupt handler to signal the main application task to process application I/O. Calling the LonEventPump() function when there are no FTXL events is acceptable.

The host application should be prepared to process the maximum rate of

ONWORKS traffic delivered to the device. Although events are enqueued within

L the FTXL LonTalk protocol stack, your application should call the LonEventPump() function frequently to process events. Use the following formula to determine the minimum call rate for the LonEventPump() function:

rate

where for this device, and your application (that is, buffers that hold incoming data until your application is ready to process it). The formula subtracts one from the number of available buffers to allow new data to arrive while other data is being processed. However, the formula also assumes that your application has more than one input buffer; having only one input buffer is generally not recommended.

If the application expects periods of inactivity, it can simply wait for the FTXL LonTalk protocol stack to post an event. If the application expects periods where it is busy for several milliseconds at a time, it should call the LonEventPump() function during the busy time to ensure that events are processed. Use the formula above to determine a baseline for how often to call the LonEventPump() function.

Recommendation: In the absence of measured data for the network, assume 90 packets per second arriving for the device. This packet rate meets the TP/FT-10 channel’s throughput figures, assuming that most traffic uses acknowledged or request/response service. Use of other service types will increase the required packet rate, but not every packet on the network is necessarily addressed to this device.

Using the formula, devices that implement two input buffers and are attached to a TP/FT-10 network that expect high throughput should call the LonEventPump() function approximately once every 10 ms.

MaxPacketRate

InputBufferCount

ateMaxPacketR

rCountInputBuffe

is the maximum number of packets per second arriving

1−

is the number of input buffers defined for

When an event occurs, the LonEventPump() function calls the appropriate event function for your host application to handle the event. Your event handler functions must be designed for this minimum call rate, and should defer timeconsuming operations (such as lengthy flash writes) whenever possible, or manage them in separate contexts (tasks or threads).

See Appendix handler and callback handler functions.

82 Developing an FTXL Application

FTXL LonTalk API

, on page 147, for a list of the available event

Example:

while (1) { // process application-specific data ... if (OsalWaitForEvent(readyHandle, OSAL_WAIT_FOREVER) == OSALSTS_SUCCESS) LonEventPump(); }

...

void LonEventReady(void) { OsalSetEvent(readyHandle); }

In the example, the readyHandle variable is the handle to an OSAL event; this handle is defined using the OsalCreateEvent() function during the application’s initialization phase, and is signaled by the LonEventReady() callback handler function whenever an event is ready to be processed.

Sending a Network Variable Update

Your FTXL device typically communicates with other LONWORKS devices by sending and receiving network variables. Each static network variable is represented by a global variable declared by the LonTalk Interface Developer utility in the FtxlDev.c file, with extern declarations provided in the FtxlDev.h file. To send an update for a static output network variable, first write the new value to the network variable declared in FtxlDev.c, and then call the LonPropagateNv() function to send the network variable update. The LonPropagateNv() function uses the index of the network variable, which is defined in the LonNvIndex enumeration in FtxlDev.h. The index names use the following format:

Name

LonNvIndex

Example: A network variable that is named nviRequest has the index name LonNvIndexNviRequest.

For dynamic network variables, the application must call the LonGetNvValue() function to retrieve the address of the value of a dynamic network variable.

The LonPropagateNv() function forwards the update to the FTXL LonTalk protocol stack, which in turn transmits the update to the network. This function returns an error status that indicates whether the update was delivered to the FTXL LonTalk protocol stack, but does not indicate successful completion of the update itself.

The FTXL device must be configured and online to be able to propagate a network variable value. If the LonPropagateNv() function is called when the FTXL device is not configured or not online, the function returns LonApiOffline.

After the update is complete, the FTXL LonTalk protocol stack informs the LonEventReady() callback handler function in the FTXL application. The application then calls the LonEventPump() function, which in turn calls your LonNvUpdateCompleted() callback handler function, to notify your application of the success or failure of the update. You can use this function for any

FTXL User’s Guide 83

application-specific processing of update completion. Figure 10 shows the control flow for processing a network variable update.

FTXL Application

LonPropagateNv()

LonEventPump()

LonNvUpdateCompleted()

FTXL LonTalk Protocol

Stack and API

Send Network Variable

Update to Network

Add “Update Complete”

Event to Queue

LonEventReady()

Figure 10. Control Flow for Sending a Network Variable Update to the Network

In the case of an unacknowledged or repeated service type, the FTXL LonTalk protocol stack considers the update complete when it has finished sending the update to the network. In the case of an acknowledged service type, the FTXL LonTalk protocol stack considers the update complete when it receives

acknowledgements from all receiving devices, or when the retry timer expires

times (where

To process an update failure, edit the LonNvUpdateCompleted() callback handler function in the FtxlHandlers.c file. This function is passed the network variable index (the same one that you passed to the LonPropagateNv() function), and is also passed a success flag. The function is initially empty, but you can edit it to add your application-specific processing. The function initially appears as:

void LonNvUpdateCompleted(const unsigned index, const LonBool success) { /* TBD */ }

84 Developing an FTXL Application

is the retry count for the network variable + 1).

Do not handle an update failure with a repeated propagation; the FTXL LonTalk protocol stack automatically retries a number of times based on the network variable’s retry count. A completion failure generally indicates a problem that should be signaled to the user interface (if any), flagged by an error or alarm output network variable (if any), or by signaled as a comm_failure error through the nvoStatus network variable of the Node Object functional block (if there is one).

Example: The following model file defines the device interface for a simple power converter. This converter accepts current and voltage inputs on its nviAmpere and nviVolt input network variables. It computes the power and sends the value on its nvoWatt output network variable:

network input SNVT_amp nviAmpere; network input SNVT_volt nviVolt; network output SNVT_power nvoWatt;

fblock UFPTpowerConverter { nvoWatt implements nvoPower; nviAmpere implements nviCurrent; nviVolt implements nviVoltage; } powerConverter;

The following code fragment, implemented in your application’s code, uses the data most recently received by either of the two input network variables, computes the product, and stores the result in the nvoWatt output network variable. It then calls the LonPropagateNv() function to send the computed value.

#include "FtxlDev.h"

void myController(void) { nvoWatt = nviAmpere * nviVolt; if (LonPropagateNv(LonNvIndexNvoWatt)!= LonApiNoError) { // handle propagation error here // such as lack of buffers or validation ... } }

Receiving a Network Variable Update from the Network

When the FTXL LonTalk protocol stack receives a network variable update from the network, it enqueues the event and signals the arrival of the event by calling the LonEventReady() callback handler function. When the application calls the LonEventPump() function, the FTXL LonTalk protocol stack writes the update to your network variable (by using the variable’s address stored in the network variable table), and then calls the LonNvUpdateOccurred() event handler function to inform your application that the update occurred. The application can read the current value of any input network variable by reading the value of the variable declared in the FtxlDev.c file.

FTXL User’s Guide 85

If a network variable update is received while the FTXL device is offline, the value of the network variable is updated, but the LonNvUpdateOccurred() event handler function is not called.

To process notification of a network variable update, modify the LonNvUpdateOccurred() event handler function (in the FtxlHandlers.c file) to call the appropriate functions in your application. The API calls this function with the index of the updated network variable. for receiving a network variable update.

Figure 11 shows the control flow

Receive Network

Variable Update from

Network

FTXL LonTalk Protocol

Stack and API

Update

Add “Update Received”

Event to Queue

LonNvUpdateOccurred()

FTXL Application

nviCount

LonEventPump()

Figure 11. Control Flow for Receiving a Network Variable Update

Configuration network variables are used much in the same way as input network variables, with the exception that the values must be kept in persistent storage, and the application does not always respond to changes immediately. Example 1, below, shows the processing flow for regular network variable updates, and example 2 shows the same flow but with the addition of a configuration network variable.

Example 1:

This example uses the same power converter model file from the example in the previous section, demonstrated how to read the network variable inputs asynchronously by reading the latest values from the network variables declared in the FtxlDev.c file.

86 Developing an FTXL Application

Sending a Network Variable Update

, on page 83. That example

This example extends the previous example and shows how your application can be notified of an update to either network variable. To receive notification of a network variable update, modify the LonNvUpdateOccurred() callback function.

In FtxlHandlers.c:

extern void myController(void);

void LonNvUpdateCompleted(unsigned index, const LonBool success) {

switch (index) {

case LonNvIndexNviAmpere: /* fall through */ case LonNvIndexNviVolt: myController(); break;

default: /* handle other NV updates (if any) */ } }

In your application source file:

#include "FtxlDev.h"

void myController(void) { // nvoWatt = nviAmpere * nviVolt; LON_SET_UNSIGNED_WORD(nvoWatt, LON_GET_UNSIGNED_WORD(nviAmpere) * LON_GET_UNSIGNED_WORD(nviVolt)); if (LonPropagateNv(LonNvIndexNvoWatt) != LonApiNoError) { // handle propagation error here ... } }

This modification calls the myController() function defined in the example in the previous section, network variable types are defined as type LonWord, this example uses the LON_GET_UNSIGNED_WORD macros to get the nviAmpere and nviVolt network variable values, and LON_SET_UNSIGNED_WORD to set the value for the nvoWatt network variable.

Example 2:

This example adds a configuration network variable to Example 1. A SCPTgain configuration property is added to the device interface in the model file:

network input SNVT_amp nviAmpere; network input SNVT_volt nviVolt; network output SNVT_power nvoWatt;

network input cp SCPTgain nciGain;

fblock UFPTpowerConverter { nvoWatt implements nvoPower; nviAmpere implements nviCurrent; nviVolt implements nviVoltage; } powerConverter fb_properties {

FTXL User’s Guide 87

Sending a Network Variable Update

, on page 83. Because

nciGain };

You can enhance the myController() function to implement the new gain factor:

void myController(void) { // nvoWatt = nviAmpere * nviVolt * nciGain.multiplier; LON_SET_UNSIGNED_WORD(nvoWatt, LON_GET_UNSIGNED_WORD(nviAmpere) * LON_GET_UNSIGNED_WORD(nviVolt) * LON_GET_UNSIGNED_WORD(nciGain.multiplier)); // nvoWatt /= nciGain.divider;

LON_SET_UNSIGNED_WORD(nvoWatt, LON_GET_UNSIGNED_WORD(nvoWatt) / LON_GET_UNSIGNED_WORD(nciGain. divider));

if (LonPropagateNv(LonNvIndexNvoWatt) != LonApiNoError) { // handle propagation error here ... } }

Configuration network variables must be persistent, that is, their values must withstand a power outage.

Handling a Network Variable Poll Request from the Network

Devices on the network can request the current value of a network variable on your device by polling or fetching the network variable. The FTXL LonTalk protocol stack responds to poll of fetch requests by sending the current value of the requested network variable.

Handling Changes to Changeable-Type Network Variables

When a network management tool plug-in or the LonMaker browser changes the type of a changeable-type network variable, it informs your application of the change by describing the new type in the SCPTnvType configuration property that is associated with the network variable.

When your application detects a change to the SCPTnvType value:

• It determines if the change is valid.

• If the change is valid, it processes the change.

• If the change is not valid, it reports an error.

Valid type changes are those that the application can support. For example, an implementation of a generic PID controller might accept any numerical floatingpoint typed network variables (such as SNVT_temp_f, SNVT_rpm_f, or SNVT_volt_f), but can reject other types of network variables. Or a data logger device might support all types that are less than 16 bytes in size, and so on.

88 Developing an FTXL Application

Echelon FTXL User Manual

Specifications and Main Features

Frequently Asked Questions

User Manual

Welcome

Audience

Related Documentation

Related Altera Product Documentation

Related devboards.de Product Documentation

Introduction to FTXL

Overview

A LONWORKS Device with a Single Processor Chip

A LONWORKS Device with Two Processor Chips

Comparing Neuron-Hosted, ShortStack, and FTXL Devices

Requirements and Restrictions for FTXL

Development Tools for FTXL

FTXL Architecture

The FTXL Developer’s Kit

Overview of the FTXL Development Process

Getting Started with FTXL

FTXL Developer’s Kit Overview

Installing the FTXL Developer’s Kit

Hardware Requirements

Software Requirements

DBC2C20 Software

Installing the FTXL Developer’s Kit

FTXL API Files

LonTalk Interface Developer

Example FTXL Applications

Creating a Model File

Model File Overview

Defining the Device Interface

Defining the Interface for an FTXL Application

Defining a Functional Block

Defining a Network Variable

Defining a Configuration Property

Declaring a Message Tag

Defining a Resource File

Writing Acceptable Neuron C Code

Anonymous Top-Level Types

Legacy Neuron C Constructs

Using Authentication for Network Variables

Specifying the Authentication Key

How Authentication Works

Managing Memory

Address Table

Alias Table

Domain Table

Network Variable Configuration Table

Example Model files

Simple Network Variable Declarations

Network Variables Using Standard Types

Functional Blocks without Configuration Properties

Functional Blocks with Configuration Network Variables

Functional Blocks with Configuration Properties Implemented in a Configuration File

Running the LonTalk Interface Developer

Specifying the Project File

Specifying the FTXL Transceiver Configuration

Specifying Service Pin Held Events

Configuring the FTXL LonTalk Protocol Stack

Configuring the Buffers

Configuring the Application

Configuring Support for Non-Volatile Data

Specifying the Device Program ID

Specifying the Model File

Specifying Neuron C Compiler Preferences

Specifying Code Generator Preferences

Compiling and Generating the Files

Using the LonTalk Interface Developer Files

Copied Files

LonNvTypes.h and LonCpTypes.h

FtxlDev.h

FtxlDev.c

project.xif and project.xfb

Using Types

Bit Field Members

Enumerations

Floating Point Variables

Network Variable and Configuration Property Declarations

Constant Configuration Properties