CB_Lab4Layer

This is the experimental CMSIS-Build Laboratory that uses project layers to compose multiple project examples from a set of different software packs.

CMSIS-Build can generate multiple reference examples for many different evaluation kits that demonstrate the functionality of complex software stacks. The range of examples is not limited, it can start from a simple RTOS based Blinky to more complex examples such as Cloud IoT connectors, or Machine Learning examples.

Directory Structure

Directory	Content Description
./build	Input: Scripts that generate reference examples from layers
./layer	Output/Input: Layers extracted from master projects that are used to generate reference examples
./project	Input: Master projects that are used to extract the layers
./docs	Documentation that explains how to create the demo examples from scratch
./demo	Demo examples (published on http://www.keil.com/iot)

Note: this is work in progress and there are several gaps that we will close in the near future:

Output projects are currently manually configured for MDK (uVision). In future, the Board pack should define the debug configuration of example projects. Board may have on-board debug adapters and therefore the debug settings depends on the target board that is used.

The various projects are based on software packs that public available.

Layers for IoT Cloud Examples

The following diagram shows the various layers that are used to compose the IoT Cloud examples.

Layer	Implements
Board	Base setup of a evaluation board; it includes CMSIS-Drivers and typically uses the device SDK
RTOS	RTOS kernel used in the example application
Socket	IP (internet) communication; frequently a TCP/IP stack
Module	WiFi Shield used in the example; a separate layer to make it easy to change Arduino shields
App	Application example; for IoT Cloud examples it contains typically a Cloud SDK

The examples are composed in a way that the device configuration is preserved and can be adopted to the requirements of a complex user application. It allows move this examples to bespoke hardware and extended it further with user code.

Working with the Examples

The following steps describe how to generate working examples from the content of the .\layer directory.

1. Start bash shell: and set the .\build directory as base directory (start bash there or use the cd command).

      /c/Work/CB_Lab4Layer/build (master)
      $

2. Setup CMSIS-Build: execute the CMSIS-Build setup script.

      $ source /c/CMSIS-Build/cbuild/etc/setup
      Set CMSIS_PACK_ROOT=/C/ARM/PACK
      Set CMSIS_BUILD_ROOT=/c/CMSIS-Build/cbuild/bin
      Set CMSIS_COMPILER_ROOT=/c/CMSIS-Build/cbuild/etc
      Added /c/CMSIS-Build/cbuild/bin into the PATH

3. Compose a single project from layers: For this step it is necessary to understand which layers can be combined. Valid combinations are in the ProjectList.txt (# is a comment line that lists future combinations). The example below composes the project AWS_MQTT_Demo for 32F746GDISCOVERY board using ESP8266 module and WiFi socket with RTX as RTOS

      $ ./gen_proj.sh App=AWS_MQTT_Demo Board=32F746GDISCOVERY Module=ESP8266 Socket=WiFi RTOS=RTX --layer=../layer
      Generate project from layers
      Output:  AWS_MQTT_Demo/32F746GDISCOVERY_ESP8266_WiFi_RTX
      Project: AWS_MQTT_Demo.cprj
      (cbuildgen): Build Process Manager 0.9.0 (C) 2020 ARM
      M650: Command completed successfully.

4. Compose multiple projects out of layers: The file ProjectList.txt provides the verified combinations. To compose all project combinations execute the following script:

      $ ./gen_proj_list.sh ProjectList.txt --layer=../layer

5. Build a composed project: Invoke CMSIS-Build using the .cprj file directly from the build directory (or from the target sub-directory)

      $ cbuild.sh AWS_MQTT_Demo/32F746GDISCOVERY_ESP8266_WiFi_RTX/AWS_MQTT_Demo.cprj
      (cbuild.sh): Build Invocation 0.9.0 (C) 2019-2020 ARM
      ... 
      cbuild.sh finished successfully!

Layers can be extracted from the master projects by executing the following script:

      $ ./extract_layers.sh