Data Match breakpoints - do they work?

[chriscowdery]

Hi All,

 Data Match breakpoints are broken for iMXRT1021 on IAR Embedded Workbench. I can set a breakpoint, but it is never triggered.

IAR are aware of the problem, and may investigate...

BUT - has anybody tried them using a different toolchain? I would like definite confirmation that there is no 'errata' with the chip itself? And that the problem is _definitely_ with IAR.

Thanks,

Chris.

[Chozyigit]

Hello Chris,

I know it is too late discuss it for you but I would like share for people who has the same problem. 

I believe problem belongs to IAR since we use Tiva Series TI MCU, but data match works in very very few occasions. On the other hand you should use code breakpoint then edit it, instead of data breakpoint. The video below may help, but I can ensure the methods in this video do not work everytime. 

https://www.youtube.com/watch?v=FykMIrKIFlA

[kerryzhou]

Hi Chris Cowdery,

    I think this function need to support by the toolchain instead of the processor.

    The toolchain can read the related address, when the address data is changed, then put a breakpoint.

    Do you check with IAR side, why that IDE have the problems, or any workaround about it? 

    You can contact with IAR support, the IAR expert should more familiar with it.

Have a great day,
Kerry

-------------------------------------------------------------------------------
Note:
- If this post answers your question, please click the "Mark Correct" button. Thank you!

- We are following threads for 7 weeks after the last post, later replies are ignored
Please open a new thread and refer to the closed one, if you have a related question at a later point in time.
-------------------------------------------------------------------------------

[chriscowdery]

Hi Kerry,

 I am checking with IAR.

But can you confirm that the iMXRT1021 does support full Data Match Write breakpoints?

Thanks,

Chris.

[kerryzhou]

Hi Chris Cowdery,

     RT can support the breakpoint,  but the full Data Match write breakpoints, it should be realized by the IDE, not just the MCU. RT is the ARM cortex-M7 core, it can support the Cortex M7 core breakpoint type.

      You can check with IAR side, whether this feature need the specific support in the ARM core side?

Have a great day,
Kerry

-------------------------------------------------------------------------------
Note:
- If this post answers your question, please click the "Mark Correct" button. Thank you!

- We are following threads for 7 weeks after the last post, later replies are ignored
Please open a new thread and refer to the closed one, if you have a related question at a later point in time.
-------------------------------------------------------------------------------

[chriscowdery]

OK - many thanks Kerry. I will discuss with IAR.

Chris.

[kerryzhou]

Hi Chris Cowdery 

  Please also tell your IAR version, the debugger which you are using?

  Do you try the official RT1020 SDK IAR project? When you set the breakpoint, whether that works or not?

  You also can share some screenshots about your IDE breakpoints problems.

Have a great day,
Kerry

-------------------------------------------------------------------------------
Note:
- If this post answers your question, please click the "Mark Correct" button. Thank you!

- We are following threads for 7 weeks after the last post, later replies are ignored
Please open a new thread and refer to the closed one, if you have a related question at a later point in time.
-------------------------------------------------------------------------------

[chriscowdery]

Hi Kerry,

 I am using IAR EWARM 8.50.1, but the problem was there on older versions.

The breakpoint triggers _every time_ memory location 0x00000000 is written to (so both lines in the test code).

But it should _never_ be triggered with the configuration above.

I am using a J-Link.

IAR know it's a problem, but I want to know if anybody has tested this feature on a different IDE.

Thanks,

Chris.

[kerryzhou]

Hi Chris,

  Thanks for your updated information.

   So the IAR breakpoints issue, you still need to check it with IAR side.

  About the other IDE, I checked the MCUXPresso user manual MCUXpresso_IDE_User_Guide.pdf under:

C:\nxp\MCUXpressoIDE_11.1.0_3209

, it has this description:

If you have interest, you can try the MCUXpresso IDE by refering the user manual.

Have a great day,
Kerry

-------------------------------------------------------------------------------
Note:
- If this post answers your question, please click the "Mark Correct" button. Thank you!

- We are following threads for 7 weeks after the last post, later replies are ignored
Please open a new thread and refer to the closed one, if you have a related question at a later point in time.
-------------------------------------------------------------------------------

[chriscowdery]

Hi Kerry,

  That MCUXpresso example is not the same.

It is a breakpoint on data write to 0x10007d00.

I require data write of a value to 0x10007d00. So a data match breakpoint.

The example given for MCUXpresso does work on IAR. It is the data match that I am asking about.

Has anybody tested data match on MCUXpresso?

Thanks,

Chris.

[kerryzhou]

Hi Chris Cowdery,

  It seems the MCUXPresso IDE doesn't have the data match function.

  Can you also check with MDK side(www.keil.com), whether they have this function or not.

  This function is seldomly used.

Have a great day,
Kerry

-------------------------------------------------------------------------------
Note:
- If this post answers your question, please click the "Mark Correct" button. Thank you!

- We are following threads for 7 weeks after the last post, later replies are ignored
Please open a new thread and refer to the closed one, if you have a related question at a later point in time.
-------------------------------------------------------------------------------

[chriscowdery]

Hi Kerry,

 I don't have the Keil tools, so I can't check those. Maybe someone at NXP has them?

Are you able to verify that it is supported by the processor? I.e. it should work if the toolchain was OK.

Thanks,

Chris.