Which part are you using?
I'm using the BCM920737 dev kit.
I'm not using the whole resistor stuff attached to the cable. I assumed that this is rather an internal J-Link adapter circuit. .... or could it have something to do with that?
Are you using the instructions provided in the WICED Smart - Debugging Support Using a J-Link Probe Application Note, or within the blog here, which outlines the exact Segger debugger we used to test this functionality: WICED Smart - Debugging Support Using a J-Link Probe Application Note
Also note that with this tool, single step debug with access to register values is not provided, so there are some limitations.
I believe simple breakpoints should still work though.
ok...., but on page 16 of the "Wiced Smart - Debugging Support Using a J-Link Probe" it says that "10. Step into, over, and out of instructions using the IDE buttons" is possible.
It seems so that stepping within the J-Link debugger terminal is also working (but just giving the assambler code as output)...
Is there a list of limitations somewhere?
When stopping within APPLICATION_INIT() at BUSY_WAIT_TILL_MANUAL_CONTINUE_IF_DEBUG_ENABLED() I can access all the variables within my "expression"-window. But as soon as it stop at some further breakpoint it just always tells me that it could not read from some address. Something is really going wrong here and debugging makes definitely no fun like this.
If I don't add any breakpoints the program works as expected. So I don't have any other fundamental problem preventing the actual code from working....
update to the topic:
I realized that the SWD and the HCI_Uart are multiplexed an so I added BLE_TRACE_DISABLE and I call ble_traceDisable() within the APPLICATION_INIT.
But still if I halt the device or I stop at a break point I'm not able to read any variable...
Does anybody else have this problem or does it work anywhere else?
When I stop the debugger I get the following out put:
Your CPU voltage is 1.7V is too low.
JLINK needs 3.3V.
I was confused about the 1.77 V anyway. But it seems that the USB is just delivering 1.7 and not more. Even with the battery I will not reach more than 3V at best. Is there a different way to power the board?
ok, I tried it with an external power supply of 3.3V (soldered to the battery pins)... but it is still not working at all. The only thing I can do is stop at break points. But reading/write of variables nor stepping nor continue to run works. This way the debugging really makes no sense.
Is this just the way it is or does it work for somebody else?
Please take a look at the pwm_tones App as an example where you can setup the HW Debugging:
* To debug, add DEBUG=1 to this application's make target in the IDE.
* Then hook up J-Link to J10 on the tag board. You will need J-Link
* GDB server, J-Link Commander and J-Link (Base version is sufficient).
* Once you download the application from the IDE, start J-Link GBD server
* and then from the debug menu, choose WICED Smart debug configuration.
* This will start GDB client which will connect to the GDB server and
* start a debug session. Make sure that when starting J-Link GDB server,
* you have have the following configurations:
* Target device = Cortex-M3, Little endian
* Target interface = SWD
* Speed = 1000 kHz
* Init regs on start = unchecked
* KNOWN ISSUE: Eclipse will not be able to halt the CPU sometimes. Open
* J-Link commander and wait for it to connect. Once it establishes a
* connection with the debug HW block, type 'halt', without the quites
* in the commander window and Eclipse will regain control of the CPU.
Hope this helps,
1 of 1 people found this helpful
I can't believe it, it works now. The second SETUP_APP_FOR_DEBUG_IF_DEBUG_ENABLED() call within the create function was missing.
Thank you so much.