TRACECLK and CPU clock

RE: TRACECLK and CPU clock

Guilherme de Paula — Wed, 06 Nov 2019 21:01:10 GMT

But this link from Segger says it is fine. Maybe there is a way to make it work.

https://wiki.segger.com/Tracing_on_Nordic_Semiconductor_nRF52832

RE: TRACECLK and CPU clock

Guilherme de Paula — Wed, 06 Nov 2019 20:40:11 GMT

Wow. This is a big problem for us. The nRF52840 has this limitation too. Our problem is even more complicated, as the J-Trace Pro that we bought are only supported by IAR in "Sampling" mode and not in "Streaming" mode. We are considering switching to SES (Segger Embedded Studio).

RE: TRACECLK and CPU clock

Kenneth — Tue, 20 Aug 2019 07:24:38 GMT

Yes, I expect this will be addressed, but I can't comment on when and how etc.

Best regards,
Kenneth

RE: TRACECLK and CPU clock

eblot — Thu, 15 Aug 2019 15:40:55 GMT

Ok.

Is it something that will be fixed in the next iteration of the nRF52, because the ETM may be used, among other needs, to debug performance issues *without* altering the normal behaviour of the CPU, and it looks like the implementation of the ETM port by Nordic defeats this very purpose. It seems quite paradoxal to me.

Thanks

RE: TRACECLK and CPU clock

Kenneth — Thu, 15 Aug 2019 07:19:37 GMT

[quote user="eblot0"]Am I correct?[/quote]

Yes, though I expect if you disable cache that should help, so just call this before ble_stack_init()::

NRF_NVMC->ICACHECNF = 0x00;

RE: TRACECLK and CPU clock

eblot — Wed, 14 Aug 2019 15:36:42 GMT

Ok, so that means there is no way to get a full valid trace output out of nRF52 devices, it all depends on parameters that cannot be tweaked or even guessed. Am I correct?

RE: TRACECLK and CPU clock

Kenneth — Wed, 14 Aug 2019 11:06:05 GMT

Afaik:

1. No, not possible.

2. The speed of the processor will heavily depend on the cache hit/miss, and the flash accessing. In certain cases, where code is being executed from RAM, there can be potentials for trace overflow.