How do I avoid halting the CPU when reading/writing RAM from debugger?

RE: How do I avoid halting the CPU when reading/writing RAM from debugger?

Susheel Nuguru — Wed, 20 Apr 2016 09:53:26 GMT

I have looked into the log file and it seems that something is wrong here. I have forwarded this to the developer responsible for this tool. I will update you when I have some information from him.

RE: How do I avoid halting the CPU when reading/writing RAM from debugger?

Claus Andersen — Wed, 20 Apr 2016 09:18:02 GMT

The linked archive contains: source for a test program binaries ready to run log showing the behavior (line 138 should say false instead of true)

www.dropbox.com/.../nrfjprog test.7z

PS should I be able to attach a file directly to the message? I can only insert links.

RE: How do I avoid halting the CPU when reading/writing RAM from debugger?

Susheel Nuguru — Wed, 20 Apr 2016 08:38:15 GMT

Can you please attach your script here. I will try it on my desk.

RE: How do I avoid halting the CPU when reading/writing RAM from debugger?

Claus Andersen — Wed, 20 Apr 2016 08:07:40 GMT

I have a loop where I read from RAM, test if the CPU is halted and in case it is call go. Occasionally this loop succeeds a few times but then go stops having any effect.

RE: How do I avoid halting the CPU when reading/writing RAM from debugger?

Claus Andersen — Wed, 20 Apr 2016 08:01:45 GMT

I'm using version 8.4.0 according to the release notes. The files are taken from the install directory at C:\Program Files (x86)\Nordic Semiconductor\nrf5x\bin.

RE: How do I avoid halting the CPU when reading/writing RAM from debugger?

Susheel Nuguru — Wed, 20 Apr 2016 07:10:54 GMT

Hi ca, which version of nrfjprog dll are you using?

RE: How do I avoid halting the CPU when reading/writing RAM from debugger?

Claus Andersen — Wed, 20 Apr 2016 03:06:25 GMT

Yes I have tried the things you suggest but they do not start the CPU again. Calling for instance the NRFJPROG_go function I get a SUCCESS error code back but the CPU does not start running again and the function NRFJPROG_is_halted tells me that the CPU is in fact still halted. I get the same result when using the nrfjprog.exe.

The CPU does not NEED to be halted to read the memory. This can bee seen when using the JLink Commander as I said. So this must be a decision made by the developers of the nrfjprogdll. Would it be possible to have an extra version of NRFJPROG_read that did not halt the CPU?

RE: How do I avoid halting the CPU when reading/writing RAM from debugger?

Susheel Nuguru — Tue, 19 Apr 2016 11:35:46 GMT

update: I have updated news from the developer responsible for nrfjprog tools. They do halt the CPU and keeps it halted after the mem read/write operation is complete. The user have to then call "GO" for dll or --run from the command line tool and then it should work.

RE: How do I avoid halting the CPU when reading/writing RAM from debugger?

Susheel Nuguru — Tue, 19 Apr 2016 11:23:14 GMT

Hi CA,

The nrfjprog might halt the CPU when it is attempting to read memory addresses.

Did you try below?

nrfjprog --run