9.4.1 Step-by-step guide to enable semihosting in your application

1. For MPLAB XC32 projects:
   1. Type -lrdimon-musl into the Project Properties, XC32 (Global Options) > Additional options text box.
   2. If it has not already been specified, enter a heap size in Project Properties, XC32 (Global Options) > xc32-ld > Heap size (bytes). For more information, see Allocating Heap Memory and MPLAB® XC32 C/C++ Compiler User's Guide for PIC32C/SAM MCUs.

      

   For Arm GCC projects:

   1. Type -lrdimon into the Project Properties, Arm GCC (Global Options) > arm-ld > Additional options text box.
   2. If you get a link error for the symbol end, add the following definition to your application linker script. It should have the same value as the existing _end symbol:

      end = DEFINED(end) ? end : . ;
      _end = .;

2. Call Initialise_monitor_handles in your code before you make any semihosting calls. For example:

   extern void initialise_monitor_handles(void);
   
   int main (void){
      initialise_monitor_handles();
      
      printf(“hello world!”);
    }
   

3. Add autoload.py script to your project by right clicking the project and selecting Add > Add autoload.py. Open the autoload.py file (located in project > Script Files) and copy the following code to the autoload.py file:

   def on_halted(pc): 
       bp = memPgm.Read16(pc) 
       if bp != 48811: #0xbeab 
           return False 
       semihost() 
       deb.SetNamedReg("PC", pc+2) #move pc over swbp 
       return True 
   
   def semihost(): 
       r = deb.GetNamedReg("R0") 
       if r == 12: 
           r = deb.SetNamedReg("R0",1) #file length 1 for :tt 
           return  
       
       if r == 5:         
           block = deb.GetNamedReg("R1") 
           buf = memPgm.Read32(block+4) 
           len = memPgm.Read16(block+8) 
           message = bytearray(len) 
           memPgm.Read(buf ,message,0,len) 
           msg.printToTab(message.decode(), "Printf output") 
    
       r = deb.SetNamedReg("R0",0)  
       return 

   Note: autoload.py enables scripting with Jython inside the IDE. It is a powerful way to automate and get control (e.g., for debugging, memory access or advanced breakpoints.)
4. Reload scripts by right clicking on the project and selecting Add > Reload Scripting Commands.
5. (Optional) For improved throughput use memory buffer by adding the following before the first printf call:

   // Optional memory buffer for printf for enhanced performance
   charbuff[1024];
   setvbuf(stdout,buff,_IOFBF,1024);

You have now configured your project to direct printf data to the Output window > Printf output tab. Add printf freely in your application code and #include <stdio.h>.

#include <stdio.h> 
#include <xc.h> 

extern void initialise_monitor_handles(void); 

void main(void) { 
    initialise_monitor_handles(); 

    char buff[1024]; 
    setvbuf(stdout,buff,_IOFBF,1024); 

    volatile int i = 0; 
    while(1)
    { 
        i= i + 1; 
        printf("Count %i\n", i);     
    } 
}