[mbedmicro/pyOCD] Loading the .elf file will give an warning，file could not be loaded (#624)

[PpgpOs]

Use
DAPlink v1
STM32F407ZGT6

Python 3.7
MSYS2
User@DESKTOP-U3LLAHV MINGW32 /c/Users/User/Desktop/MyProject/Stm32/Stm32f407zg/Project1
PyOCD 0.19

# pyocd gdbserver

0001226:INFO:board:Target type is cortex_m

0001256:INFO:dap:DP IDR = 0x2ba01477

0001273:INFO:dap:AP#0 IDR = 0x24770011

0001283:INFO:rom_table:AP#0 ROM table #0 @ 0xe00ff000 (designer=020 part=411)

0001293:INFO:rom_table:[0]<e000e000:SCS-M4 class=14 designer=43b part=00c>

0001299:INFO:rom_table:[1]<e0001000:DWT class=14 designer=43b part=002>

0001305:INFO:rom_table:[2]<e0002000:FPB class=14 designer=43b part=003>

0001312:INFO:rom_table:[3]<e0000000:ITM class=14 designer=43b part=001>

0001318:INFO:rom_table:[4]<e0040000:TPIU-M4 class=9 designer=43b part=9a1 devtype=11 archid=0000 devid=0:0:ca1>

0001325:INFO:rom_table:[5]<e0041000:ETM-M4 class=9 designer=43b part=925 devtype=13 archid=0000 devid=0:0:0>

0001329:INFO:cortex_m:CPU core #0 is Cortex-M4 r0p1

0001337:INFO:cortex_m:FPU present: FPv4-SP

0001346:INFO:dwt:4 hardware watchpoints

0001351:INFO:fpb:6 hardware breakpoints, 4 literal comparators

0001377:INFO:server:Semihost server started on port 4444

0001377:INFO:gdbserver:GDB server started on port 3333

The original Led flashing program stopped running，GPIO state retention
# arm-none-eabi-gdb build/Project1.elf
Then hit the Enter key twice to skip the message.
...

Type "apropos word" to search for commands related to "word"...

Reading symbols from build/Project1.elf...done.

(gdb) target remote localhost:3333

Remote debugging using localhost:3333

warning: while parsing target memory map (at line 3): Required element <memory> is missing

0x080012f6 in ?? ()

Report warning

(gdb) load

Loading section .isr_vector, size 0x188 lma 0x8000000

Loading section .text, size 0xe98 lma 0x8000188

Loading section .rodata, size 0x10 lma 0x8001020

Loading section .ARM, size 0x8 lma 0x8001030

Loading section .init_array, size 0x4 lma 0x8001038

Loading section .fini_array, size 0x4 lma 0x800103c

Loading section .data, size 0xc lma 0x8001040

Start address 0x8000f5c, load size 4172

Transfer rate: 18 KB/sec, 521 bytes/write.

(gdb) monitor reset

(gdb)

MCU not respond,Unplug the DAP cable and plug it back in,MCU run,but the original program is still running.

It was only done yesterday, it took a few hours... I can’t use it today...

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

[Chris Reed]

Hi @PpgpOs, you need to set the target type. You can see the first line in the pyocd gdbserver log where it says "Target type is cortex_m", indicating it's using the generic cortex_m target type (no flash programming or memory map). Currently, pyocd does not support auto-detection of the target with STLink probes (this will be merged in soon).

Since the STM32F103ZE target is not builtin to pyOCD, you'll need to use a CMSIS-Pack to enable support. With pyOCD 0.19.0 you can run this:

pyocd pack -u -i stm32f103ze

Then run this to list the supported targets:

pyocd list --targets

You'll see a target with the name stm32f103ze in the list. You can now pass this target name for the --target command line argument to pyocd gdbserver in order to enable flash programming and the memory map.

(I'm going to be working on much better documentation soon!)

[Chris Reed]

Just noticed you first mentioned the F407… The same process applies for enabling support for this target type.

[Chris Reed]

Fyi, the target support documentation covers much of this information and more detail in certain areas.

[PpgpOs]

@flit
I probably understand，but...

# pyocd pack --find stm32f103ze

usage: pyocd [-h] [-V] [--help-options]  ...

pyocd: error: argument : invalid choice: 'pack' (choose from 'commander', 'cmd', 'erase', 'flash', 'gdbserver', 'gdb', 'json', 'list')

The same result if you type pyocd pack -u -i stm32f103ze

[PpgpOs]

@flit Hi
pyocd gdbserver --pack C:/Users/User/Desktop/Keil.STM32F1xx_DFP.2.3.0.pack -t stm32f103ze is that
Run it successfully!
Thank you!

Why do you say that the pack command is a pyocd pack...?

[Chris Reed]

Ooops, I'm sorry! The pack command hasn't yet been released, it's only on the master branch. I'll be releasing version 0.20.0 quite soon that includes it. My apologies for the confusion!

[Savastaro Vegas]

Oh Ok. I was also confused. I thaught i had missed something.
Hi @flit, daily dose 😊: automated_test have 4/9 success on stm32f429zi (and all other boards i tested). All gdb tests, flash_test, and flash_loader_test do not pass. Have i missed something? Btw, debug with pyocd using all these boards works fine. Only tests that cause problems.

[Chris Reed]

No, you didn't miss anything! There are some known issues with STM32 devices not passing some of the tests: #579, #580, #586. Some fixes will be coming in 0.20.0, but that won't fix everything yet. But I haven't had any issues actually using it for debugging these device.

[PpgpOs]

@flit
0047452:WARNING:freertos:FreeRTOS: pxReadyTasksLists size is unexpected, maybe an unsupported version of FreeRTOS
This warning will appear when starting gdbserver when using FreeRTOS.
At this point, I can still enter debugging, but once I run osKernelStart(); (Start scheduler), I will have problems.
Will eventually stop in the vPortRaiseBASEPRI function.

portmacro.h
...
portFORCE_INLINE static void vPortRaiseBASEPRI( void )
{
uint32_t ulNewBASEPRI;

	__asm volatile //Will eventually stop here,single step debugging did not respond
	(
		"	mov %0, %1												\n"	\
		"	msr basepri, %0											\n" \
		"	isb														\n" \
		"	dsb														\n" \
		:"=r" (ulNewBASEPRI) : "i" ( configMAX_SYSCALL_INTERRUPT_PRIORITY )
	);
}