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.

TCI6638K2K, CCSv6, Pro Trace Memory Transaction Logs shows wrong EDMA_CC


We capture memory transaction logs via XDS560v2 Pro Trace with the settings below:

We modified the settings for capturing DSP1 LSRAM (L2_1) EDMA WR Access only.

Our application uses EDMA_CC1_TC0 for default, and we confirmed this with the logs. But when we modify the application for using EDMA_CC4_TC0 instead of EDMA_CC1_TC0, even though the register&paramset movements confirm that we are using the EDMA_CC4_TC0, memory transaction logs show that we are using EDMA_CC2_TC0.

Moreover, there is no EDMA_CC4 in the transaction master list below.

So far I am sure that there is no paramset in EDMA_CC2 which addresses DSP1 L2SRAM as destination. We are using CCS Version: Any suggestions about that?

Thanks in advance..

  • Hi,

    Please apologize for the delay; we can confirm some of the issues found by you on the tool and are working to properly identify their source.

    I will get back to you on Monday.

  • Hi Rafael,

    It's already been Tuesday here :)

  • Hi,

    Sorry again - I am in the middle of an office move and most of my equipment is packed and stored away.

    We found out the issue is located in two files that need to be modified to add this support (and also fix the typos).These files are located under C:\ti\ccsv7\ccs_base\emulation\analysis\xmldb.
    - devices\device_kepler.xml
    - aet_config\AET_PropertyCP_Tracer_kepler.xml

    We are gathering the necessary information to modify them and will get back to you.

  • Mi mgunal--

    The Transaction Logging feature in Keystone 2 only supports a 7b Master ID (vs the SOC's 8b Master ID). Using Table 8-6 in the TCI6638K2K datasheet as a reference: most Master IDs are allocated an ID below 128, but there are a few with IDs >= 128.

    The result is that any master with a Master ID >= 128 will alias with (Master ID - 128). In the case of EDMA4 (Master IDs 169-173) they will alias with EDMA2 (Master IDs 41-45). This, I believe, explains the issue that you are seeing.

    I have filed an internal ticket against the Tools so that any aliasing that may exist is clear.

  • Thanks Jason,

    Could you suggest any quick workaround we can apply? It is very important for us to make sure which masters are being used.