nrfjprog.dll - nrfjprogdll_go not starting SoC

[Martin]

Hi

I used nrfjprog.dll in our internal software. implementation is in C#.

Everything works just fine - except for the last call to NRFJPROGDLL_go - it seems like it starts the proc for a couple mseccs, then it returns to halted state.

If I call it from command line (nrfjprog -f nrf51 --run) everything is fine and the device starts to work...

Is there any more detailed information on how the commands --run and --go are implemtend in nrfjprog?
I already had a look pynrfjprog but this does not give any insight on this issue.

I'm usijng the latest version of nrf command line tools.

Any hints?

[Simen August Tinderholt]

Do you perform any actions after _go() ? Some functions in the nRFJProg DLL halt the processor to ensure data integrity.

If possible, could you post the nRFJProg log, or a pseudocode of your script?

[Martin]

Hi

pseudo code below:

void flashingDevice(){
    
    NRFJProg.NRFJPROG_open_dll ();
    
    NRFJProg.NRFJPROG_is_dll_open( ref open );
    
    if(open){
        NRFJProg.NRFJPROG_is_connected_to_emu( ref connected );
        
        if(!connected){
            NRFJProg.NRFJPROG_connect_to_emu_without_snr( 8000 );
            
            NRFJProg.NRFJPROG_read_connected_emu_snr( ref serial );
            
            NRFJProg.NRFJPROG_connect_to_device();
            
            // GET SOME DEVICE INFOS
            NRFJProg.NRFJPROG_is_connected_to_device( ref deviceConnected );
            if ( deviceConnected ) {
                NRFJProg.NRFJPROG_halt();
                Thread.Sleep( 200 );
                
                NRFJProg.NRFJPROG_is_halted( ref halted );
                
                NRFJProg.NRFJPROG_erase_all();
                
                foreach(cell in sdHex){
                    NRFJProg.NRFJPROG_write(cell, ...);
                }
                
                foreach(cell in fwHex){
                    NRFJProg.NRFJPROG_write(cell, ...);
                }
                
                // SET DATA TO UICR
                NRFJProg.NRFJPROG_write_u32();

                if(protect)
                NRFJProg.NRFJPROG_readback_protect(NRFJProg.readback_protection_status_t.ALL);
                
                // Write some more data to RAM
                NRFJProg.NRFJPROG_write_u32();
                Thread.Sleep( 200 );
                
                // JUST A TRY
                //NRFJProg.NRFJPROG_sys_reset();
                //Thread.Sleep( 400 );
                
                NRFJProg.NRFJPROG_go();
                Thread.Sleep( 400 );
                
                NRFJProg.NRFJPROG_disconnect_from_device();
                
                // becoming desperate
                NRFJProg.NRFJPROG_go();
            }
        }
        
        NRFJProg.NRFJPROG_disconnect_from_emu();
    }
}

[Simen August Tinderholt]

Ah, you're protecting, then doing go?

You should perform a debug reset instead, this will reset your device and start the program. Since you've enabled readback protection, the go probably fails, as you no longer have access to the core.

[Martin]

currently I'm not setting protection, this was also one of my first thoughts.

should I do a debug_reset instead of sys_reset anyways?

[Simen August Tinderholt]

If you're not protected you should be able to do a sysreset, but a debug reset is mostly equivalent, and should always work.

After programming you should at least reset the device to make sure your program counter and stack pointers are read correctly.

[Martin]

okay, I will make the changes and report back.

BTW: any chance you can make a flow diagram available how you are handling the commandline args in nrfjprog?