• State Resolved
  • Locked Locked
  • Replies 1 reply
  • Subscribers 99 subscribers
  • Views 354 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

TDA4VM: C7x: write-back/invalidate system

Tsurumoto
Genius 15995 points

Part Number: TDA4VM

Hi

The registers in the write-back/invalidate system need to wait for the completion of execution until the bit field is cleared after the start operation.

However, your CSL implementation does not wait for execution to complete.

■Questions

In C7, is it unnecessary to wait for an operation after write-back/invalidate?

　→For example, is it OK to perform another write-back-invalidate operation before the completion of the first write-back-invalidate operation?

If so, why is this not done in CSL?

 

Thanks and Best regards,

HaTa

  • 0
    TI__Intellectual 2570 points

    There are two questions here:

    1. For C7x writeback-n-invalidate to be executed cleanly, I would recommend to wait for the completion by checking the completion bit. 

    2. For CSL function, my understanding is there might be two CSL's provided: one does writeback-n-invalidate as you observed, one does the check of completion. Therefore, users are given the flexibility to use them with respect to their applications. For example, user can:

    #1. Choose to use them back-to-back. 

    #2. Or, writeback-n-invalidate first, followed by some other functions, then execute the completion check if he/she needs to do so. 

    #3. Or, check if there is any previous writeback is completed or not by executing completion check first, then execute the writeback-n-invalidate. This sounds odd but usually is used to be a safer coding convention if multiple modules from different programmers are integrated together and you might "trust" the other programmers did the check. 

    So by splitting the two functions into two separate CSL's, users are given the maximum flexibility for their software work. 

    Thanks,

    Kai