Codewarrior 10.1 Downloading process is getting corrupted

[timias]

I am using Codewarrior 10.1

I have MQX 3.7 Installed and I am trying to debug a program on a board with a MCF51JE256 processor.

Immediately after I created a new project using the MQX 3.7 Wizard, and moving my known good app source files over. I connect to my target using  a PEMicro USB Multilink Pod. This worked for a few sessions, until now. I terminated my debug session, recompiled and attempted to download the new image, but this time the debugger prompted me with a new message. " The target MCU is not responding. Please turn MCU power off (Under 0.1V), turn MCU power on, and then click OK...."

I did and then it told me "Device is secure. Erase to unsecure?" I Clicked Yes and the image seems to download, but instead of going to main it goes to __boot (in boot.c) and Resume now only single steps through code.

Things I have already done:

I have reinstalled CW 10.1, Reinstalled MQX 3.7, rebooted the computer, recycled target power, unplugged the pod,  replaced the pod, swapped out hardware, and none of that had any effect.

I was successfully using the same setup prior to installing MQX 3.7 (but I am unsure if this is the cause), because I also changed the BSP I was using from twrmcf51MM to the twrmcf51JE, as it is now available and is technically more correct.

The only way to stop the error is to completely scrap my project, and start over (VERY FRUSTRATING). This is the second time this has happened today.

What information can I provide to expedite the troubleshooting of this problem?

[J2MEJediMaster]

It is time to get an engineer formally involved here. To do that, file a service request. Click here to go to the service request page. Pick the category of Software Product Support and topic CodeWarrior. Be prepared to submit a project that has the condition described.

---Tom

[timias · Accepted solution]

I thought I would provide an update. I opened a support case with Freescale, and we seemed to intially identify a discrepency with my MQX installation. There was a patch for CW 10.1 I didn't have installed. For a while that seemed to fix the issue, and then about a month later, it reocurred. This time though I tried a few new things.

I had a bare project created from troubleshooting a completely seperate issue in my workspace. While my existing project failed to download, giving me the "error device is secure... " the bare project downloaded successfully everytime. Comparing between the two projects on the debug configuration, yielded two seperate paths to an identical memory config file.

Being that a MCF51MM256 is nearly identical to a MCF51JE256 ( my processor) at least from MQX's standpoint I switched to the MM. Low and behold it suddenly started downloading.

I suspect that some temporary files are being corrupted and changing between them , resets it to a good state.

The settings are located at:

debug configuration->remote system[edit]->system  (both target init file and memory config files)

[timias]

Root cause - The Coldfire MCF51JE processor has two onboard flash controllers with reserved memory locations which control flash lock. The first one is around address 0x420 which is great because you can link your code around it. The second one is is right smack in the middle of the flash, so sometimes your code (if it gets long enough)  will jack up that register and screw up the debugger.  The default linker files didn't address this, and there is no good way to modify them to, other than manually placing each file around the second controller's control register.

Bottom line we are changing our processor as soon as we can, but in the mean time keep the size down.