This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

CCSV4.2.3 problem with c6472 EVM



We have hit into a very strange issue with the CCSv4.2.3

 Here is the scenario.

What we are doing: verifying data transfer from a signal path FPGA to the DSP DDR memory.

 Setup: Stratix IV FPGA EVM connected to 6472 Rev4 EVM over SRIO (single port).

CCSv4 version 4.2.3

We load DSP code using CCSv4 and load FPGA using Altera FPGA tool.

Gel file used: evmc6272.gel provided as part of CCSv4 installation

 What we have observed: the memory content (as seen using CCS memory window), after running the test is not consistent. Occasionally (but rare) we see the right data in the memory window. 

It is not cache related since cache is not enabled.

However, if we press ALT-S in CCSv4, the right data shows up in memory window.

We then did another experiment. Keeping the HW setup as is, we used a different PC for CCSv4. This PC has CCSv4 version 4.2.1. We don’t have the above problem with this setup. It works consistently in this setup.

The second setup uses same DSP project, same gel file and same FPGA build as the first one.

 The behavior is repeatable.

Uisng 4.2.1 isntead of 4.2.3 is not a good option for us.

  • I am also experiencing some memory window issue with CCS version 4.2.3. The content of the memory window often doesn't respond after numerous hit on the refresh button  and is left with ????. Also, if I were to scroll up or down on the memory window to view the different memory address , the new memory content doesn't get updated.

    I am using EVM6472 with CCS: 4.2.3.00004

     

    Thanks,

     

    Dominic

  • Dominic, 

    A colleague reported that CCSv4.2.3 displays question marks across the window whenever a single invalid address is shown in the memory view, therefore it will require investigation. I will get back on this.

    Regards,

    Rafael

  • Can we install 4.2.1 alongside 4.2.3?

    Cheers

  • Eddie said:
    Can we install 4.2.1 alongside 4.2.3?

    Sure. Just use separate install directories and separate workspace folders.

  • SNBANIK ,

    SNBANIK said:
     What we have observed: the memory content (as seen using CCS memory window), after running the test is not consistent. Occasionally (but rare) we see the right data in the memory window. 
    It is not cache related since cache is not enabled.
    However, if we press ALT-S in CCSv4, the right data shows up in memory window.

    Is the target still running when you see incorrect data in the memory window?

    ki

  • I am not sure what you exactly meant by "Is the target still running ..."

    We stop the processor to view the memory but otherwise the EVM is connected. 

  • SNBANIK said:

    We stop the processor to view the memory but otherwise the EVM is connected. 

    hmmm. then all should work fine. Did you try refreshing the display (right-click in memory view and select 'refresh' in the context menu)? Could you also check that memory location with the disassembly view and watch view to see if it shows the correct values at that address? That way we can see if the issue is specific to the memory view only.

    ki

  • Yes we tried refreshing the display too. No difference.

    Watch View - this is a large buffer (image). Can watch view handle something like 512 x 600 array?

  • SNBANIK said:
    Watch View - this is a large buffer (image). Can watch view handle something like 512 x 600 array?

    Maybe just enter a few addresses explicitly and see if the value at the addresses looks correct

    How does the disassembly view look?

  • I am not sure about the dissembly view. I haven't used it to view data memory location. May be you can help me by putting in a few lines as to how to use that to view data locations.

     The HW setup that shows this problem is shared with other groups and currently being used for verifying other aspects. I would let you know as soon as I get to it.

    However this sa,e HW setup with the CCS running on another PC (CCS4.2.1) shows no problem.

    I am not sure what extra information you (or I)  can get with these experiments. 

  • SNBANIK said:
    I am not sure about the dissembly view. I haven't used it to view data memory location. May be you can help me by putting in a few lines as to how to use that to view data locations.

    You should just be able to enter an address in the location field and it will display the contents of memory at that location (in addition to try disassembling the value).