CCS could not load xerf5 file

Did you try disabling any antivirus/security software running on your PC? This is a common root cause.

im not the admin so i dont have the access to disable antivirus/security software, but my collegue's PCs dont have such trouble and they run the same antivirus/security software&rules deployed by my company. So i dont think this is the root cause.

There is likely some process that is running that is preventing the debugger from accessing the file. When this problem occurs, are you able to access the file normally? For example can you move the file or rename it while CCS is closed?

i dont know where the GEL file is, but im sure i can delete the xerf5 file freely and even re-generate it by recompiling the project while CCS is opened but not in debug mode.

When this issue occurs again, can you:

Assuming you are on Windows, open Task Manager and go the Performance tab and Open Resource Monitor:

In Resource Manager, go to the CPU tab and under Associated Handles, search for the xerf5 file in question:

Provide the results of the search.

Thanks

ki

sorry that my OS is Win7 SP1 and language is simplified chinese

But you can see that when this issue occurs, no handles occupy the xerf5 file.

btw: Before this issue occured, i tried to flash bin files to my EVM via uniflash and failed because of COM occupied. Then i re-powered and re-connected my EVM, flashed bin files successfully using python. i dont know if it would help solve the issue.

Thanks

Thank you for checking.

kaisheng liu said:

sorry that my OS is Win7 SP1

I have not tried this with Windows 7. I'll give this a try tomorrow.

kaisheng liu said:

btw: Before this issue occured, i tried to flash bin files to my EVM via uniflash and failed because of COM occupied.

Do you recall if your previous failures also occurred after using UniFlash?

im sure that the previous two failures came after using uniflash, but i cant recall other failures before

I have escalated this issue to engineering. It is under further investigation.

Engineering has done some more analysis and it appears that it is due to an issue with a third party library that we rely on. This library as a dependency on a specific windows event and if it can't find it in the event log (whether it was lost of overwritten), it was cause the issue

If you are interested in the exact details:

https://stackoverflow.com/questions/38045857/boostinterprocess-exception-library-error-exception-when-creating-shared-mem

As you are aware, a reboot resolves the issue. We are looking to improve our implementation so that it removes this dependency. Unfortunately we don't have a good current workaround besides the reboot.

so will this issue be fixed in the future version of CCS or Uniflash?

and in the link you provided, it's a bug in boost lib 5 years ago, is it fixed in the latest boost 1.81+?

i understand that i can do nothing but a reboot after this issue happened, and i wonder what action may cause this issue?how can i avoid this issue before you fix it?

kaisheng liu said:

so will this issue be fixed in the future version of CCS or Uniflash?

Yes this is the plan. It would be fixed in both versions.

kaisheng liu said:

and in the link you provided, it's a bug in boost lib 5 years ago, is it fixed in the latest boost 1.81+?

It is unclear if this is resolved. In any case, our plan is to modify how the program is loaded so that this issue will not come into play at all.

kaisheng liu said:

i wonder what action may cause this issue?

Basically, the boost code is looking for a windows event 6005 which should be there but it can't find it and the exception occurs. Why the event is missing is a bit unknown. It could be that if there are a lot of events in the log, perhaps it got overwritten? This is speculation.

The first answer in the link I provided explains it better.