stm32l4r5zitx - Loadable section "RW_IRAM1" outside of ELF segments

Hi

I'm using mbed studio and when trying to debug on a stm32l4r5zitx with a STLink.

I'm getting the following log when trying to debug:

Selected port 50000 for debugging
0000297:INFO:board:Target type is stm32l4s5vitx
0000428:INFO:coresight_target:Asserting reset prior to connect
0000581:INFO:dap:DP IDR = 0x2ba01477 (v1 rev2)
0000640:INFO:ap:AP#0 IDR = 0x24770011 (AHB-AP var1 rev2)
0000643:INFO:rom_table:AP#0 ROM table #0 @ 0xe00ff000 (designer=020 part=470)
0000645:INFO:rom_table:[0]
0000646:INFO:rom_table:[1]
0000647:INFO:rom_table:[2]
0000648:INFO:rom_table:[3]
0000649:INFO:rom_table:[4]
0000651:INFO:rom_table:[5]
0000652:INFO:cortex_m:CPU core #0 is Cortex-M4 r0p1
0000657:INFO:dwt:4 hardware watchpoints
0000660:INFO:fpb:6 hardware breakpoints, 4 literal comparators
0000668:INFO:coresight_target:Deasserting reset post connect
0000674:INFO:server:Semihost server started on port 4444
0000675:INFO:gdbserver:GDB server started on port 50000
Reading symbols from c:\Users\User\source\repos\arm-mbed\BUILD\MY_BOARD\ARMC6\arm-mbed.elf...
warning: 
**Loadable section "RW_IRAM1" outside of ELF segments**
done.
0000789:INFO:gdbserver:One client connected!
**warning: Loadable section "RW_IRAM1" outside of ELF segments
warning: Overlapping regions in memory map: ignoring**
0x080003d4 in Reset_Handler ()
0000812:INFO:gdbserver:Attempting to load Argon
0000812:INFO:gdbserver:Attempting to load FreeRTOS
0000812:INFO:gdbserver:Attempting to load Zephyr
0000813:INFO:gdbserver:Attempting to load RTX5
0000813:INFO:gdbserver:RTX5 loaded successfully
Attached to debugger on port 50000
Image loaded: c:\Users\User\source\repos\arm-mbed\BUILD\MY_BOARD\ARMC6\arm-mbed.elf
0001809:ERROR:gdbserver:Unhandled exception in handle_message: individual memory accesses must not cross memory region boundaries
Traceback (most recent call last):
  File "c:\Users\User\AppData\Local\Mbed Studio\mbed-studio-tools\python\lib\site-packages\pyocd\gdbserver\gdbserver.py", line 549, in handle_message
    reply = handler(msg[msgStart:])
  File "c:\Users\User\AppData\Local\Mbed Studio\mbed-studio-tools\python\lib\site-packages\pyocd\gdbserver\gdbserver.py", line 776, in v_command
    return self.v_cont(cmd)
  File "c:\Users\User\AppData\Local\Mbed Studio\mbed-studio-tools\python\lib\site-packages\pyocd\gdbserver\gdbserver.py", line 791, in v_cont
    if self.is_threading_enabled():
  File "c:\Users\User\AppData\Local\Mbed Studio\mbed-studio-tools\python\lib\site-packages\pyocd\gdbserver\gdbserver.py", line 1320, in is_threading_enabled
    return (self.thread_provider is not None) and self.thread_provider.is_enabled \
  File "c:\Users\User\AppData\Local\Mbed Studio\mbed-studio-tools\python\lib\site-packages\pyocd\rtos\rtx5.py", line 408, in is_enabled
    return self.get_kernel_state() != 0 and not self._target.in_thread_mode_on_main_stack()
  File "c:\Users\User\AppData\Local\Mbed Studio\mbed-studio-tools\python\lib\site-packages\pyocd\rtos\rtx5.py", line 446, in get_kernel_state
    return self._target_context.read8(self._os_rtx_info + RTX5ThreadProvider.KERNEL_STATE_OFFSET)
  File "c:\Users\User\AppData\Local\Mbed Studio\mbed-studio-tools\python\lib\site-packages\pyocd\core\memory_interface.py", line 64, in read8
    return self.read_memory(addr, 8, now)
  File "c:\Users\User\AppData\Local\Mbed Studio\mbed-studio-tools\python\lib\site-packages\pyocd\debug\cache.py", line 516, in read_memory
    return self._memcache.read_memory(addr, transfer_size, now)
  File "c:\Users\User\AppData\Local\Mbed Studio\mbed-studio-tools\python\lib\site-packages\pyocd\debug\cache.py", line 422, in read_memory
    data = self.read_memory_block8(addr, 1)[0]
  File "c:\Users\User\AppData\Local\Mbed Studio\mbed-studio-tools\python\lib\site-packages\pyocd\debug\cache.py", line 442, in read_memory_block8
    if not self._check_regions(addr, size):
  File "c:\Users\User\AppData\Local\Mbed Studio\mbed-studio-tools\python\lib\site-packages\pyocd\debug\cache.py", line 414, in _check_regions
    raise MemoryAccessError("individual memory accesses must not cross memory region boundaries")
pyocd.debug.cache.MemoryAccessError: individual memory accesses must not cross memory region boundaries
**"0001809:ERROR:gdbserver:Unhandled exception in handle_message: individual memory accesses must not cross memory region boundaries"
warning: Remote failure reply: E01**
0001809:ERROR:gdbserver:Unhandled exception in handle_message: individual memory accesses must not cross memory region boundaries

Key lines in there are:

warning: **Loadable section "RW_IRAM1" outside of ELF segments** done.         
0000789:INFO:gdbserver:One client connected! **warning: Loadable section "RW_IRAM1" outside of ELF segments warning:

In the map file, the RAM looks ok:

Execution Region RW_IRAM1 (Base: 0x200001c0, Size: 0x00002cc8, Max: 0x0009fa40, ABSOLUTE)
Execution Region ARM_LIB_STACK (Base: 0x2009fc00, Size: 0x00000400, Max: 0x00000400, ABSOLUTE)

Processor datasheets say:

The STM32L4S5xx, STM32L4S7xx and STM32L4S9xx devices feature 640 Kbytes of
embedded SRAM. This SRAM is split into three blocks:
• 192 Kbytes mapped at address 0x2000 0000 (SRAM1).
• 64 Kbytes located at address 0x1000 0000 with hardware parity check (SRAM2).
This memory is also mapped at address 0x2003 0000 offering a contiguous address
space with the SRAM1.
This block is accessed through the ICode/DCode buses for maximum performance.
These 64 Kbytes SRAM can also be retained in Standby mode.
The SRAM2 can be write-protected with 1 Kbyte granularity.
• 384 Kbytes mapped at address 0x2004 0000 - (SRAM3).

So it does sound like memory is 0x2000 0000 - 0x200A 0000 which matches the map file's use of IRAM. So why does pyOCD say the ELF file is wrong?

Thank you.

Start with the ELF-loading warning and memory-region checks in pyocd/gdbserver/gdbserver.py and pyocd/debug/cache.py, then inspect the RTX5 access path in pyocd/rtos/rtx5.py. Reproduce with the reported STM32L4 ELF and determine why RW_IRAM1 is outside the recognized segments; done means debugging proceeds without the ELF or memory-boundary errors.