Matter Linux Fabric Bridge Example

An example application to implement the Aggregator device type with Fabric Synchronization condition met and demonstrates the end-to-end Fabric Synchronization feature using dynamic endpoints.

Fabric Synchronization feature will facilitate the commissioning of end devices from one fabric to another without requiring user intervention for every end device. It defines mechanisms that can be used by multiple ecosystems/controllers to communicate with one another to simplify the experience for users.

This doc is tested on Ubuntu 22.04 LTS (aarch64)

Theory of Operation

Dynamic Endpoints

The Bridge Example makes use of Dynamic Endpoints. Current SDK support is limited for dynamic endpoints, since endpoints are typically defined (along with the clusters and attributes they contain) in a .zap file which then generates code and static structures to define the endpoints.

To support endpoints that are not statically defined, the ZCL attribute storage mechanisms will hold additional endpoint information for NUM_DYNAMIC_ENDPOINTS additional endpoints. These additional endpoint structures must be defined by the application and can change at runtime.

To facilitate the creation of these endpoint structures, several macros are defined:

DECLARE_DYNAMIC_ATTRIBUTE_LIST_BEGIN(attrListName) DECLARE_DYNAMIC_ATTRIBUTE(attId, attType, attSizeBytes, attrMask) DECLARE_DYNAMIC_ATTRIBUTE_LIST_END(clusterRevision)

  • These three macros are used to declare a list of attributes for use within a cluster. The declaration must begin with the DECLARE_DYNAMIC_ATTRIBUTE_LIST_BEGIN macro which will define the name of the allocated attribute structure. Each attribute is then added by the DECLARE_DYNAMIC_ATTRIBUTE macro. Finally, DECLARE_DYNAMIC_ATTRIBUTE_LIST_END macro should be used to close the definition.

  • All attributes defined with these macros will be configured as ATTRIBUTE_MASK_EXTERNAL_STORAGE in the ZCL database and therefore will rely on the application to maintain storage for the attribute. Consequently, reads or writes to these attributes must be handled within the application by the emberAfExternalAttributeWriteCallback and emberAfExternalAttributeReadCallback functions. See the bridge application's main.cpp for an example of this implementation.

DECLARE_DYNAMIC_CLUSTER_LIST_BEGIN(clusterListName) DECLARE_DYNAMIC_CLUSTER(clusterId, clusterAttrs, role, incomingCommands, outgoingCommands) DECLARE_DYNAMIC_CLUSTER_LIST_END

  • These three macros are used to declare a list of clusters for use within a endpoint. The declaration must begin with the DECLARE_DYNAMIC_CLUSTER_LIST_BEGIN macro which will define the name of the allocated cluster structure. Each cluster is then added by the DECLARE_DYNAMIC_CLUSTER macro referencing attribute list previously defined by the DECLARE_DYNAMIC_ATTRIBUTE... macros and the lists of incoming/outgoing commands terminated by kInvalidCommandId (or nullptr if there aren't any commands in the list). Finally, DECLARE_DYNAMIC_CLUSTER_LIST_END macro should be used to close the definition.

DECLARE_DYNAMIC_ENDPOINT(endpointName, clusterList)

  • This macro is used to declare an endpoint and its associated cluster list, which must be previously defined by the DECLARE_DYNAMIC_CLUSTER... macros.

Building

  • Install tool chain

    sudo apt-get install git gcc g++ python pkg-config libssl-dev libdbus-1-dev libglib2.0-dev ninja-build python3-venv python3-dev unzip
    
  • Build the example application:

    cd ~/connectedhomeip/examples/fabric-bridge-app/linux
    git submodule update --init
    source third_party/connectedhomeip/scripts/activate.sh
    gn gen out/debug
    ninja -C out/debug
    
  • To delete generated executable, libraries and object files use:

    cd ~/connectedhomeip/examples/fabric-bridge-app/linux
    rm -rf out/
    

Running the Complete Example on Ubuntu

  • Building

    Follow Building section of this document.

  • Run Linux Fabric Bridge Example App

    cd ~/connectedhomeip/examples/fabric-bridge-app/linux
    sudo out/debug/fabric-bridge-app
    
  • Test the device using ChipDeviceController on your laptop / workstation etc.