Hi,
(Here's a newbie answer so take that for what it's worth).
Have only done limited JTAG debugging on the BB-X15 with CCS 7.x - when bringing up the SPL on a custom board with the same Sitara SOC as on the BB-X15.
I THINK I recall that only a hand full of the 20 pins (like 7 maybe?) actually carry data and or clock for JTAG debugging on the 20 pin connector So the subset of JTAG connections on the Pocket Beagle may be identical as on the 20 pin connector??
. ***But please look this up and correct me if I'm in error!!!!!***
There's a TI board port series on how to setup CCS to debug the Linux SPL on the BeagleBone Black. It's old, but TI mentioned on E2E that they're working on updating this video series.
When debugging the SPL, breakpoints couldn't be set in CCS. This is a known problem, and I don't know if TI has a solution yet. Alternatively in the SPL, we were able to click "run until" in the disassembly window to determine the flow of execution. Just that it was in assembly, yuck.
Haven't tried debugging anything but the SPL using JTAG. I think there have been other discussion threads on the merits of using JTAG vs. GDB and other techniques for debugging LInux. I've heard that it can be useful for debugging drivers, but I THINK part of the issue maybe kernel context switching??