Posts by awneil

So what I was doing was opening Tools->J-Link Settings:

seeing that it already had the correct Serial number, I just clicked 'Save to Project' - and that doesn't add the current Serial number to the Project.
in the Console, it says:

Tools.JLinkSettings();
J-Link settings were written to the project file.

But if I do choose the '...' for 'Host Interface', select the required Serial number, and then do 'Save to Project' - it is, indeed, saved.
in the Console, it says:

Tools.JLinkSettings();
J-Link settings were written to the project 
file.Project.SetHostIF ("USB", "777578294");

Quote from SEGGER - Nino

Where is the difference

Ozone always creates the .user file anyhow - so there is no user effort to create it.

It would just take the setting from the standard dialogue, so no need to transcribe the serial number and mess about in some other tool to create a "script" of some sort - which would, itself, have to be "dragged about".

Also, I think (but haven't tried) that the .user file would be portable between Windows, Linux, etc - which a "script" wouldn't be ?

Quote from SEGGER - Nino

Works flawlessly for me for years.
Could you provide a short clip on how you try to do this?

could it be that I am using on-board ST-Links which have been converted to J-Link - is there some restriction related to the free conversion?

Quote from SEGGER - Nino

The prompt should only appear if more that 1 J-Link are detected to be connected to the host PC.

Yes, that is the case

Quote from SEGGER - Nino

Is this a common occurrence with you and your colleagues

Yes: it's a radio system, with a transmitter and a receiver.

Quote from SEGGER - Nino

Simply slelect the one you want

I do that

Quote from SEGGER - Nino

and it will be saved in your Ozone project

but it doesn't do that.

Quote from SEGGER - Nino

You can call Ozone with -usb <S/N> and Ozone will open the last project with your J-Link selected

It's a workaround.

I still think it would be nicer to have it in the per-user setup file

OK - thanks.

Quote from awneil

Presumably, the J-Link Serial Number should appear as the 2nd argument to Project.SetHostIF() ?

Nope - if I do that, it still prompts me to select the J-Link each time.

Actually, the J-Link serial number doesn't seem to be getting saved to the project at all:

I press 'Save to Project', and the following appears in the console:

Tools.JLinkSettings();
J-Link settings were written to the project file.

but the serial number is not added to the .jdebug project file, and there's nothing in the .jdebug.user either.

There is this in the .jdebug project file:

void OnProjectLoad (void) {
  //
  // Dialog-generated settings
  //
  Project.SetDevice ("STM32L072CZ");
  Project.SetHostIF ("USB", "");
  Project.SetTargetIF ("SWD");
  Project.SetTIFSpeed ("4 MHz");

Presumably, the J-Link Serial Number should appear as the 2nd argument to Project.SetHostIF() ?

Aha: Ozone itself has Target.WriteU32()

So this seems to work:

/*********************************************************************
*
*       AfterTargetReset
*
* Function description
*   Event handler routine. Optional.
*   The default implementation initializes SP and PC to reset values.
**
**********************************************************************
*/
void AfterTargetReset (void) {
  Target.WriteU32( 0x40015808, 0x1800 ); // Set DBG_IWDG_STOP and DBG_WWDG_STOP bits in DBG_APB1_FZ
  _SetupTarget();
}

Working on an STM32L072 system where the watchdog is enabled.

If execution breaks or is halted manually, the WD will restart the system.

The STM32L072 has a DBG_APB1_FZ register to "freeze" the WD (among other things) when the core is halted.

This should be accessible to the J-Link - so how do I do it.

I see the JLINK_MEM_WriteU32() command in the J-Link script files wiki:

https://wiki.segger.com/J-Link_script_…_WriteU32.28.29

and there's Exec.Command() in the Ozone manual.

So is it just a case of putting Exec.Command( "JLINK_MEM_WriteU32( 0x4001 5808, 0x1800)" ) at an appropriate place in the Ozone project?

And where would be an "appropriate place" ?

Currently, a specific J-Link can be defined in the .jdebug file (by its serial number).

This isn't particularly useful for a project shared between team members - as different people are likely to have their "own" J-Link(s).

So is it possible to specify a J-Link (by serial number) in the .jdebug.user file?

If so, how?

If not, can this be considered a feature request?

Quote from SEGGER - Alex

More a question for a Linux forum.

I guess - but you'll be familiar with the behaviour?

Quote from SEGGER - Alex

Usually, the 1st J-Link that enumerates gets ttyACM0, the next …1, and so on

and that happens anew each time? Unlike Windows, which remembers that it's seen this J-Link before - so gives it the same COM number that it had before.

Thanks.

On Windows (on a given PC), the COMx: number will always be the same for the same J-Link.

Thus, when (not if) I have multiple J-Links, I can tell which one I'm connecting to from its COM port number.

Does the same apply to Ubuntu, or does it just use ttyACM0 for whichever J-Link (or other ACM device) is connected first, ttyACM1 for whichever is connected second, etc ... ?

I am using Ubuntu 20.04.3 LTS (in a VirtualBox VM on Win10-Pro).

Ozone (x64) v3.24g on Ubuntu 20.04.3 LTS (in a VirtualBox VM on Win10-Pro).

If I do 'Save Project as' to a folder where I don't have write rights, the save fails (of course) but Ozone gives no indication of the failure; in fact, the Console suggests that it succeeded:

The first attempt (ringed red) failed;
The second attempt (ringed green) succeeded.

In both cases, there was no error message or warning, and the console output actually suggests that both succeeded.

Quote from vtran5

if I reconnect the chip again

How long was it disconnected?

Was there a reset in between?

Most likely, whatever's running in the chip has overwritten your value?

Quote from st0ne

i have a problem with a custom board

Is this a board that has previously worked - or a brand new design that has not yet been proven ?

Does your code disable the debug port and/or the pins it uses?

Does your code put the processor to sleep?

More tips for debugging "cannot connect to target" problems:

https://wiki.segger.com/J-Link_cannot_connect_to_the_CPU

sounds like some files had got corrupted - maybe line endings, or some other unprintable characters ... ?

By "commas" you mean "quotes" ?

Remember that the preprocessor just does a simple text substitution

So, for this to work, your #define must expand to something that would be valid as a printf argument.

So, if you want MODULE_NAME to expand to a string, you will have to define it as such; eg,

#define MODULE_NAME "MODULE1"

Then you can put MODULE_NAME anywhere that a string literal would be legal.

Note that this is standard 'C' stuff - nothing specific to Segger or SES

Yes, target debug connection is SWD, and VTREF is connected - as it's required by the SWD.

The baud rate looked correct on a scope and, as mentioned, it works fine with an FTDI USB-to-1.8V-UART cable.

I had studied the links provided.

is there any restriction on using the VCOM and SWD simultaneously ?

Target is nRF52 - Cortex-M4

Quote from yu_bl

the file cc1.exe can be found

But the error message says that the problem is starting the process - not finding the file.

Can you run that file from the command line?

Have you ever been able to build anything successfully?

Should the J-Link VCOM (Virtual COM port) work at 1.8V ?

I just want to receive stuff from a 1.8V target UART at 115200 baud.

It works OK with an FTDI USB-to-UART cable (1.8V version), but a terminal connected to the J-Link VCOM just shows junk - as if the baud rate were wrong.