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.

TMS320C6678: C6678 CorePac execution during XDS560V2 access to CorePac memory.

Part Number: TMS320C6678
Other Parts Discussed in Thread: TMS320C6202

TMS320C6678

I have ordered Spectrum Digital XDS560V2 PRO TRACE emulators for our TMS320C6678 DSP design debugging/verification effort.
Our DSP design includes an external watchdog timer which drives a Power On Reset of the DSP when the watchdog timer is not reset before hitting its terminal count.
While I wait for our target HW and the emulator, I am writing test procedures, and I am trying to determine if we should plan on disabling the watchdog timer during debug sessions when 'poking' a value into C6678 CorePac data memory becomes necessary.

Two Questions:
(1) Can an XDS560V2 PRO TRACE emulator access (read or write) CorePac L2 cache memory space (being used as a CorePac Data Memory segment) without first halting the CorePac?

(2) If yes to (1), does a runtime XDS560V2 PRO TRACE emulator access to CorePac L2 cache memory space (being used as a CorePac Data Memory segment) temporarily suspend CorePac - C66x DSP execution while the access is taking place?

I remember when performing similar accesses 14 years ago with the PCI based XDS560 emulator to debug a TMS320C6202 DSP design that runtime emulator write accesses to data memory (by updating the contents of a memory window, for example) with the XDS560 emulator produced unpredictable results (sometimes the emulator would lock up) if we didn't first halt the C6202 DSP using the XDS560 emulator before using the emulator to overwrite a value in DSP internal memory, and then allow the C6202 DSP code execution to resume again.  I want to determine whether or not the XDS560V2 behaves similarly.

Thank you,
Jim  Sanchez

  • Hi,

    I've forwarded this to the design team. Their feedback will be posted here.

    Best Regards,
    Yordan
  • James

    As I look at   processors.wiki.ti.com/.../Category:Emulation    it tells me that the JTAG access is not intrusive and real-time, so I think that the answer to your question is yes.

    But there are few issues. First is the speed of the JTAG and if it can keep up with the processor.  I believe that the priority of JTAG when accessing memory is lower than the core priority (at least the default) so you need to be aware of it

    The second issue is accessing physical memory L2 that is configured as cache.  I am not sure (you have to confirm) if the architecture lets external host accessing cache memory directly.  Please check on it (for example, define large L2 cache and then try to write to the physical address from CCS and see what happens.

    Does it make sense to you?

    regards

    Ran