|
Hi All! We are working in a multiprocessor environment with 7 ADSP 2191s connected in a sequential chain as described in the application notes. The system is passing the scan test after detecting the 7 Jtag devices. Earlier we couldn't load the emulator session. The VDSP++ 3.0 environment got hung when we tried changing the session. But we overcame the problem by holding the board in reset, when it is getting loaded. The environment is loaded with the following errors. Jtag scan failed. Check target power and connection Failed to set 16-bit IO memory. But then we cud load and execute the program into the DSPs. In this session we are facing problems in coming out of an infinite loop. Once we get into an infinite loop the VDSP++ lost its control over the device. And we couldn't restart the program without restarting the computer. Expecting help from you, to overcome these problems. Have a nice day, regards, Unmai. |
|
|
APEX ICE - Multiprocessor
Started by ●October 31, 2003
Reply by ●October 31, 20032003-10-31
|
On 31 Oct 2003, Unmai R.R wrote: > In this session we are facing problems in coming out of an infinite >loop. Once we get into an infinite loop the VDSP++ lost its control over >the device. And we couldn't restart the program without restarting the >computer. Do you know which processor is entering the loop first? I think it will be hard to regain control via the jtag chain unless you halt all the processors. Can you set individual break points in each processor? Maybe you can have each processor that can enter the loop send a signal to the others which will cause them to all hit a break point, then you can use the jtag chain to force the one that's still running to pay attention. It's not a simple problem. The jtag chain has to pass the data thru all the processors, so you need to know which one has failed. Or else you do need to halt them all externally (maybe just checking a button switch and halt if pressed will work). Patience, persistence, truth, Dr. mike |
Reply by ●October 31, 20032003-10-31
|
>Do you know which processor is entering the loop first? We have checked by loading the program into only one processor. And now, by properly selecting the session from the menu, instead of selecting from the multiprocessor window and following the sequence of power-up and power-down, we could able to gain control over the device. But still we are getting the initial error message when starting up the emulator session. Thanks a lot! Regards, Unmai. |
