10:17 <pigmoral> paulk: hi, I'm trying to use jtag on A733. OpenOCD seems to recognize the chip, but I can't halt the CPU or get some useful infomation from "dap info". I'm new to bringing up OpenOCD on a new platform, so I tried some old cfgs from your repo. And here's some OpenOCD output which might be useful:

10:17 <pigmoral> https://gist.github.com/pigmoral/87dd50b0b72a5732dcca35443550b32f

10:18 <pigmoral> Do you have any suggestions for what might be wrong or what I should check next? Thanks!

10:18 <gamiee> pigmoral : before issuing dap info, I would rather halt the core. How did you tried to halt it?

10:24 <pigmoral> Yes, I have tried "-c init -c halt", but got "Command 'halt' failed with error code -302" from openocd.

10:25 <gamiee> pigmoral: there are two options: connect to OpenOCD's server for commands, or connect to OpenOCD with GDB and issue halt with it.

10:25 <gamiee> don't do it with -c for right now

10:36 <pigmoral> got it, but I still get "error code -302" in the OpenOCD debug output, whether I use "halt" from telnet or "monitor halt" in gdb.

10:36 <pigmoral> The cfg I used: https://github.com/paulkocialkowski/openocd-config-sunxi/blob/main/sun50iw10.cfg

11:05 <gamiee> pigmoral: only error code -302?

11:05 <gamiee> I never saw that openocd returning just that

11:05 <gamiee> can you show some error log or something ,because i really didnt saw such thing

12:23 <paulk> pigmoral: it's probably using adiv6, not v5 like sun50iw10

12:24 <paulk> then you need to use ARM_IDCODE_SOC_600_4BIT

12:28 <paulk> mhh or maybe I'm reading the long wrong and it's actually ARM_IDCODE_SOC_400 with adiv5

12:29 <paulk> maybe try "dap info 1"

12:45 <pigmoral> gamiee: the error code -302 appeared in the OpenOCD output with "--debug" option, twice: first when gdb connects and again when I issue "monitor halt" in gdb. Here is the detail log: https://gist.github.com/pigmoral/e3cb7d2e1fcbc74c886ca8253ee9fada

12:45 <pigmoral> I also checked OpenOCD source code, -302 happens when it times out waiting for the CPU to enter the halted state after issuing halt.

12:48 <gamiee> pigmoral: oooh, hmm, never seen this message yet, but I worked mostly with RISC-V

12:50 <pigmoral> paulk: yap, using ARM_IDCODE_SOC_600_4BIT gives me a jtag tap error, bug ARM_IDCODE_SOC_400 works fine

12:52 <pigmoral> I also tried "dap info 1", you can check it at the end of https://gist.github.com/pigmoral/87dd50b0b72a5732dcca35443550b32f

12:55 <paulk> mhh it's probably not finding the blocks on the debug apb

13:01 <pigmoral> also, I noticed something different between a527 and a733

13:01 <pigmoral> on a527, jtag auto tap only finds 1 device: 0x2ba06477 (mfg: 0x23b (ARM Ltd), part: 0xba06, ver: 0x2)

13:02 <pigmoral> but 2 on a733: 0x5ba00477 (mfg: 0x23b (ARM Ltd), part: 0xba00, ver: 0x5) and 0x2ba06477 (mfg: 0x23b (ARM Ltd), part: 0xba06, ver: 0x2)

13:02 <pigmoral> I'm not sure if this requires any extra handling?

13:08 <paulk> mhh interesting

13:19 <pigmoral> wow, It works when I add an extra tap for the second device. And I got dap info from the second tap successfully

13:20 <pigmoral> FYI, dap info for a733 and corresponding openocd cfg: https://gist.github.com/pigmoral/d24dbd800c74e6fa5f1addaa9536bd30

13:34 <juanesf91> gamiee: https://forum.armbian.com/topic/46922-orangepi-zero-2w-tv-out/#comment-205935

13:58 <juanesf91> gamiee: https://github.com/armbian/build/pull/8086