Issue about debugging TEE environment on i.MX8MQ using JLink Plus

[_Bradley_]

Hi. I'm new to NXP board and JLink.

I have the board "imx8mqevk", and I'm trying to debug TEE using JLink Plus and Segger Ozone.

The Feature of Jlink Plus I have is like :

SEGGER J-Link Commander V7.94g (Compiled Jan 24 2024 15:48:18)
DLL version V7.94g, compiled Jan 24 2024 15:48:07

Connecting to J-Link via USB...O.K.
Firmware: J-Link V10 compiled Jan 30 2023 11:28:07
Hardware version: V10.10
J-Link uptime (since boot): N/A (Not supported by this model)
S/N: XXXXXXXXX
License(s): RDI, FlashBP, FlashDL, JFlash, GDB
VTref=3.393V

As I know, the board has 4 cortex-A53 cores. 

But, whenever I start the debugging session on Ozone and try debugging the board, it always gets into Kernel Panic due to SMP Preemption or cpu_idle something.

Linux says like :

[ 30.989659] Internal error: Oops - Undefined instruction: 0000000002000000 [#1] PREEMPT SMP
[ 30.998407] Modules linked in:
[ 31.001605] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.2.0-rc3-dirty #22
[ 31.008692] Hardware name: NXP i.MX8MQ EVK (DT)
[ 31.013422] pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[ 31.020690] pc : arch_cpu_idle+0x18/0x2c
[ 31.024797] lr : arch_cpu_idle+0x14/0x2c
[ 31.028898] sp : ffffffc00a1a3d70
[ 31.032358] x29: ffffffc00a1a3d70 x28: 0000000041fa80a0 x27: 00000000fdf63be0
[ 31.039814] x26: ffffffc00a1b4880 x25: 0000000000000000 x24: 0000000000000000
[ 31.047271] x23: ffffffc00a1a9c1c x22: ffffffc009b52b78 x21: ffffffc00a1a9a68
[ 31.054731] x20: ffffffc00a1a9c10 x19: 0000000000000000 x18: 0000000000000000
[ 31.062188] x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000000
[ 31.069643] x14: 00000000000002fe x13: 0000000000000001 x12: 0000000000000000
[ 31.077097] x11: 0000000000000001 x10: 0000000000000a40 x9 : ffffffc00a1a3cd0
[ 31.084552] x8 : ffffffc00a1b5320 x7 : 0000000000000000 x6 : 0000000010ed03ee
[ 31.092008] x5 : 4000000000000000 x4 : ffffffc0b3e2a000 x3 : 0000000000000000
[ 31.099463] x2 : 0000000000002c39 x1 : ffffff80bd97ac50 x0 : ffffffc008fd7c4c
[ 31.106923] Call trace:
[ 31.109479] arch_cpu_idle+0x18/0x2c
[ 31.113219] default_idle_call+0x30/0x6c
[ 31.117318] do_idle+0x244/0x2c0
[ 31.120696] cpu_startup_entry+0x28/0x30
[ 31.124794] kernel_init+0x0/0x12c
[ 31.128349] arch_post_acpi_subsys_init+0x0/0x18
[ 31.133174] start_kernel+0x658/0x698
[ 31.136998] __primary_switched+0xbc/0xc4
[ 31.141197] Code: d2801c00 910003fd 97fffff7 d50343ff (a8c17bfd)
[ 31.147561] ---[ end trace 0000000000000000 ]---
[ 31.152393] Kernel panic - not syncing: Attempted to kill the idle task!
[ 31.159389] SMP: stopping secondary CPUs
[ 31.163488] Kernel Offset: disabled
[ 31.167129] CPU features: 0x00000,01000108,0000421b
[ 31.172222] Memory Limit: none
[ 31.175416] ---[ end Kernel panic - not syncing: Attempted to kill the idle task! ]---

The Segger Ozone console says like:

File.Open ("/Users/xxxxxxx/Desktop/tee/xxxxxxxxx/tee.elf");

File.Open: completed in 182 ms

Program segments:

  Address         Size         Code      RO Data      RW Data      ZI Data    Flg

---------    ---------    ---------    ---------    ---------    ---------    ---

 FE000000    1 003 856      467 872      176 208      248 408      111 368    RWE

 FE200000            0            0            0            0            0    RW

 FE200000            0            0            0            0            0    RW

 00000000            0            0            0            0            0    RW

---------    ---------    ---------    ---------    ---------    ---------    ---

   Total:    1 003 856      467 872      176 208      248 408      111 368

---------    ---------    ---------    ---------    ---------    ---------    ---

For further information on ELF file data sections, execute command Elf.PrintSectionInfo(0).

Cannot break at symbol "main": symbol address could not be determined.

Invalid startup completion point "main": symbol address could not be determined.

Debug.ReadIntoInstCache: updated instruction information within 1 code ranges (0xFE000000-0xFE0723A0)

Debug.SetConnectMode (CM_ATTACH);

Debug.Start();

Device "CORTEX-A53" selected.

TotalIRLen = 4, IRPrint = 0x01

JTAG chain detection found 1 devices:

 #0 Id: 0x5BA00477, IRLen: 04, CoreSight JTAG-DP

DPv0 detected

Scanning AP map to find all available APs

AP[5]: Stopped AP scan as end of AP map has been reached

AP[0]: AHB-AP (IDR: 0x64770001)

AP[1]: APB-AP (IDR: 0x44770002)

AP[2]: JTAG-AP (IDR: 0x24760010)

AP[3]: JTAG-AP (IDR: 0x001C0030)

AP[4]: AHB-AP (IDR: 0x24770011)

Iterating through AP map to find APB-AP to use

AP[0]: Skipped. Not an APB-AP

AP[1]: APB-AP found

No AP map specified by user. Setting auto-detected AP map.

Scanning ROMTbl @ 0x80000000

[0]Comp[0] @ 0x80400000: ROM Table

Scanning ROMTbl @ 0x80400000

[1]Comp[0] @ 0x80410000: Cortex-A53

[1]Comp[1] @ 0x80420000: CSS600-CTI

Core found. Stopped ROM table scan: https://wiki.segger.com/ROMTableScan

Cortex-A53 @ 0x80410000 (detected)

CoreCTI @ 0x80420000 (detected)

Debug architecture: ARMv8.0

6 code breakpoints, 4 data breakpoints

Processor features:

  EL0 support: AArch64 + AArch32

  EL1 support: AArch64 + AArch32

  EL2 support: AArch64 + AArch32

  EL3 support: AArch64 + AArch32

  FPU support: Single + Double + Conversion

ARMv8-A/R: The connected J-Link (S/N 600101144) uses an old firmware module V0 with known problems / limitations.

Add. info (CPU temp. halted)

Current exception level: EL2

Exception level AArch usage:

  EL0: AArch32

  EL1: AArch32

  EL2: AArch64

  EL3: AArch64

Non-secure status: Non-secure

Cache info:

  Inner cache boundary: L1

  LoU Uniprocessor: 2

  LoC: 0

  LoU Inner Shareable: 2

I-Cache L1: 491 KB, 629 Sets, 32 Bytes/Line, 25-Way

VMSAv8-64: Supports 48-bit VAs

Connected to target device.

Startup complete (PC=0x00000001)

Memory access: CPU temp. halted: https://wiki.segger.com/Memory_accesses#Legacy_stop_mode

CPU could not be halted

CPU could not be halted

I think there are lots of problems like memory access, CPU synchronization, Cache size that the debugging tool recognizing and etc...

As I think, the most suspicious thing on the log is "ARMv8-A/R: The connected J-Link (S/N 600101144) uses an old firmware module V0 with known problems / limitations."

So I've checked the firmware version of my JLink Plus, but the JLinkConfig says it has up-to-date FW version. 

I think there isn't problem with building the image. because my coworker's environment can debug the board with same environment. the only thing that is different is JLink. As I said above, I'm using JLink Plus, but my coworker uses JLink(the basic one I think). 

it has been about 3 months since I get into this issue ...

Is there anyone who has experienced like me? 

Best Regards, 
Bradley. 

[Chavira]

Hi @_Bradley_,

Thank you for contacting NXP Support!

Your J‑Link Plus is using an older ARMv8‑A firmware module (V0), which is why Ozone displays the warning:

“ARMv8‑A/R: The connected J‑Link uses an old firmware module V0 with known problems/limitations.”  

This older module is not fully compatible with Cortex‑A53 debugging, and it can lead to kernel panics when attaching, while your coworker’s newer J‑Link works without issues.
Recommended Solution

Update to the latest J‑Link Software & Documentation Pack to ensure both the host tools and the probe firmware are fully up to date. 

If the same warning still appears after updating, the limitation is due to the hardware revision of your J‑Link Plus, not the software.

In that case, you will need a newer J‑Link model (e.g., V11, V12, or V13), as older hardware cannot fully support modern ARMv8‑A debug features.  

Best regards,
Chavira