I am using a EZ-Kit with an ADSP-21469.
I have this problem that happens sometime and then disappear and then come again.
When I start my program, nothing happens, the DSP seems stuck.
If I halt the target, I always find the PC stuck on the same instruction, just after a call to twidfftf.
Then if I continue the execution, the programs executes normally.
If I don't halt the program with the debugger, it remains stuck. I can see with statistical profiling that I am 100% in _twidfftf.
If I just change some unrelated instructions and compile again, the problem disappear.
So I suspect a problem of data or code alignment but can put my finger on it.
Could the DSP be in a trapped state without the debugger being notified ?
Here is the disassembly of the code where the DSP is stuck:
 cjump _twidfftf(db);
[12524B] dm(0xfffffffe,i6)=m13 <= stuck on this line
Thanks for any tip or advice to investigate this problem.