15:51 <maribu[m]> Sorry for nagging again. I'm still fighting with the TI CC2650 with the integrated ICEPick JTAG Router. I cleaned up the icepick.cfg to follow the sequence given in the datasheet to the latter. However, the icepick_c_tapenable procedure is for some reason called twice. It seems that the event tap-enable is for some reason triggered twice. Is this expected behavior?

15:54 <maribu[m]> Also: I fail to find a good documentation regarding the events that are generated by OpenOCD. E.g. as someone new to OpenOCD it is not obvious to me what the actual high level flow is.

19:12 <borneoa_> maribu[m]: hi, there was one patch in Gerrit that is probably related with your problem https://review.openocd.org/5658/ it was also taken by zephyr port in an old branch as https://github.com/zephyrproject-rtos/openocd/commit/d091d39201c4 but disappears from latest version

19:24 <maribu[m]> I am currently working around this by using the ICEPick warm reset feature and not asserting srst at all. In the first tap-enable I put the device in the wait-in-reset state, in the second I am releasing it. I think with that OpenOCD is reliably attaching. However, flashing still fails. This may actually be a different issue.

19:25 <maribu[m]> At least I seem to have no issue with connecting via GDB

21:27 <maribu[m]> Note: flash verify_image works reliably every time (after having written the image with JLinkExe). Writing the image with OpenOCD fails reliably - except for the very first time when the MCU was still in factory state. It usually take a second or two before I see any output. Then I mostly get Error: Invalid ACK (0) in DAP response.