Hi,
I have been trying to get Busblaster v2 connect to a Pandaboard for some time. I am using OpenOCD as the debugger and have downloaded the bitstream for jtagkey layout v1.3 to the CPLD. The steps I am following are similar to http://dangerousprototypes.com/docs/Bus ... nOCD_guide (http://dangerousprototypes.com/docs/Bus_Blaster_OpenOCD_guide). The source code of OpenOCD came from prdownload.berlios.de (Version : openocd-0.5.0.zip).
I am hitting the 'Warn : Invalid ACK 0x6 in JTAG-DP transaction' condition consistently and the OpenOCD keeps looping there. The logs are like (I tried latest version from git source to see if its a bug from OpenOCD, but its the same),
Open On-Chip Debugger 0.6.0-dev-00180-g444f202 (2011-11-02-23:22)
<snip>
Info : JTAG tap: omap4430.jrc tap/device found: 0x3b95c02f (mfg:
0x017, part: 0xb95c, ver: 0x3)
Info : JTAG tap: omap4430.dap enabled
Polling target failed, GDB will be halted. Polling again in 100ms
Polling target failed, GDB will be halted. Polling again in 300ms
Polling target failed, GDB will be halted. Polling again in 700ms
Info : JTAG tap: omap4430.m30_dap enabled
Info : JTAG tap: omap4430.m31_dap enabled
Warn : Invalid ACK 0x6 in JTAG-DP transaction
Polling target failed, GDB will be halted. Polling again in 1500ms
Polling target failed, GDB will be halted. Polling again in 3100ms
Polling target failed, GDB will be halted. Polling again in 6300ms
Polling target failed, GDB will be halted. Polling again in 6300ms
^C
This condition looks quite common based on Google searches but little information is found on the reasons to hit this. As I followed through the code a bit, I noticed that the debugger fails to get a ACK when it tries to read from the DP CTRL/STAT register. It needs to read this register to check whether the reset of power to system and debug domain are acknowledged or not. And from this point, it keeps looping doing the read operation. I have posted this query on OpenOCD forum some time back but haven't found anything significant to make a progress.
BTW, I have tried with SD card removed from the slot as I read elsewhere saying that it might interfere with JTAG debugging in some cases. I have also tried with Linux (Fedora 14) as well as cygwin over Windows Vista but it doesn't seem to make any difference.
Any advise or any means to swim away from this is greatly appreciated.
Regards,
Ram
have you tried to mail in the OpenOCD mailing list ? They might know better the internals of Cortex-A9 code.
Hi,
Thank you for your reply. I have tried posting this question on OpenOCD development mailing lists and Sparkfun forum. I didn't supposedly receive much response. I had a feeling that I am missing something too obvious but it looks like overall procedure I am following is okay.
http://forum.sparkfun.com/viewtopic.php?f=18&t=30585 (http://forum.sparkfun.com/viewtopic.php?f=18&t=30585)
https://lists.berlios.de/pipermail/open ... 21520.html (https://lists.berlios.de/pipermail/openocd-development/2011-November/021520.html)
I am pretty sure many users would be using this combination. Can some one check whether they are able to get the OpenOCD up with the SD card removed from the board. I am suspecting whether the board requires x-loader for setting some registers to enable debugging. I am not sure.
Thanks and Regards,
Ram