rtel wrote on Friday, August 22, 2014:
If you don’t use an interrupt wrapper then you have the standard interrupt entry/exit code generated by the compiler. That makes the code simple to write but also means potentially nested interrupts use the stack of whichever task was running when the interrupt occurred. That can consume a lot of RAM as each task must allocate enough RAM for an entire [potentially nested 6 deep] interrupt stack frame.
If you use a wrapper then you use the FreeRTOS interrupt entry code rather than the compiler’s interrupt entry code. The FreeRTOS interrupt entry code handles interrupt nesting and switches to a dedicated interrupt/system stack on interrupt entry - meaning you can allocate a smaller stack to each task and save a lot of RAM.
Please see the documentation page for this port as you will find this information there: