Write the application as normal. There must be a main() routine in the application.

Link the application against the Boot Monitor platform library file platform.a. The file platform.a is in one of the target build subdirectories (install_dir\software\firmware\Platform\Builds\target_build).

Choose the Builds subdirectory that matches your application. For example, Release_ARM_Little_Endian for ARM code. If the subdirectory does not exist, see Rebuilding the Boot Monitor or platform library for details on rebuilding platform.a.

If required, select the link time platform library options.

Platform selection uses special symbols in your application:

The platform options are listed in Table E.2.

Scatter loading is supported for applications using the platform library, however the scatter file must follow Example E.1. The execution regions INIT and SDRAM must be present, the execution regions ITCM and DTCM are optional, if they are present, the relevant Tightly Coupled Memory (TCM) is enabled and the base address must be set to the address supplied in the scatter file. The sys_vectors.o object must be located at address zero. Additional execution regions, such as one for the SSRAM, can be added. An example scatter loading application can be found in the examples directory.

LR_ROOT 0x8000
{
    INIT +0 FIXED
    {
        sys_boot.o (!!!_platform_area_boot, +FIRST)
        *(+RO)
    }
     SDRAM +0
    {
        *(+RW,+ZI)
    } 
    ITCM 0x0
    {
        sys_vectors.o(_platform_area_vectors, +FIRST)
        ‘application code’.o(+RO)
    }
    DTCM 0x08000000
    {
        ‘application code’.o(+RW,+ZI)
	}
}

To run the image from RAM, load the image with a debugger and execute as normal. The image uses the procedure described in Redirecting character output to hardware devices to redirect standard I/O either to the debugger or to be handled by the application itself.

See Loading and running an application from NOR flash for more information on running from flash.