Unable to erase image slot using mcuboot serial command.

[CreativeIndustryAg]

Hello

Platform: NXP mimxrt1060
I'm trying to implement custom dfu, and it turned out i'm not able to write/erase anything to flash.
I always get the same error (below), and what's more i'm event can't erase slot using mcuboot shell command:

uart:~$ mcuboot erase 1
[00:00:14.038,085] flash_flexspi_nor: flash_flexspi_nor_write_enable: Enabling write
[00:00:14.046,813] os: ***** USAGE FAULT *****
[00:00:14.052,398] os: Attempt to execute undefined instruction
[00:00:14.059,722] os: r0/a1: 0x00000001 r1/a2: 0x80032174 r2/a3: 0x9e8f438e
[00:00:14.068,389] os: r3/a4: 0x800024b8 r12/ip: 0x00000001 r14/lr: 0x00001251
[00:00:14.077,087] os: xpsr: 0x81000000
[00:00:14.082,275] os: Faulting instruction address (r15/pc): 0x000005c6
[00:00:14.090,209] os: >>> ZEPHYR FATAL ERROR 0: CPU exception on CPU 0
[00:00:14.098,022] os: Current thread: 0x800133c0 (shell_uart)
[00:00:14.105,072] os: Halting system

When i use mcumgr via udp everything works.

Could you help me?

[DerekSnell]

Hi @CreativeIndustryAg ,

You mention using a custom DFU. Are there some simple steps to replicate this issue on the mimxrt1060_evk board?

The usage fault is triggered by an undefined instruction. That means the opcode fetched from memory is invalid. And it reports the faulting instruction address is at 0x5C6, which is in ITCM RAM. Since code must be copied to ITCM, is there valid executable code located there? Have you already tried using a debugger and stepping through the code to see what leads up to this fault?

Best regards

[CreativeIndustryAg]

@DerekSnell
Thanks for answer. :)
It turnout the problem disappear when i disable continuous reading ADC (adc_context).
DFU and ADC uses different tasks, and there is no data exchange between them.

OZone debug call stack:

Level	Function	Stack Frame	Source	PC	Return Address	Stack Used
0	BusFault Exception	32 @ 0x80004840		0xCCCCCCCC	[0x80004858]: 0x00000CFC	14080
0	flash_flexspi_nor_read_status	24 @ 0x80004860	flash_mcux_flexspi_nor.c:180:2	0x00000CF8	[0x80004874]: 0x00000D36	1376
0	flash_flexspi_nor_wait_bus_busy	16 @ 0x80004878	flash_mcux_flexspi_nor.c:299:3	0x00000D32	[0x80004884]: 0x00000DC2	1352
0	flash_flexspi_nor_write	48 @ 0x80004888	flash_mcux_flexspi_nor.c:372:3	0x00000DBE	[0x800048B4]: 0x6001BCEE	1336
0	flash_sync	40 @ 0x800048B8	/home/agr/zephyrproject/zephyr/include/zephyr/drivers/flash.h:205:2	0x6001BCEC	[0x800048DC]: 0x60041B84	1288
0	stream_flash_buffered_write	40 @ 0x800048E0	stream_flash.c:199:3	0x60041B80	[0x80004904]: 0x6003FA2E	1248
0	flash_img_buffered_write	16 @ 0x80004908	flash_img.c:44:2	0x6003FA2A	[0x80004914]: 0x60013964	1208
0	dfu_upload_image	32 @ 0x80004918	dfu_img_utils.c:207:3	0x60013960	[0x80004934]: 0x6003D638	1192
0	parse_image_file	8 @ 0x80004938	dfu_https_ft.cpp:236:3	0x6003D634	[0x8000493C]: 0x60013436	1160
0	download_file	1088 @ 0x80004940	dfu_https_ft.cpp:215:5	0x60013434	[0x80004D7C]: 0x60013608	1152
0	download_new_image	24 @ 0x80004D80	dfu_https_ft.cpp:307:3	0x60013604	[0x80004D94]: 0x60013712	64
0	https_ft_dfu_task_entry_point	32 @ 0x80004D98	dfu_https_ft.cpp:347:9	0x6001370E	[0x80004DB4]: 0x6003DA14	40
0	z_thread_entry	8 @ 0x80004DB8	thread_entry.c:36:2	0x6003DA12	[0x80004DBC]: 0xAAAAAAAA	8
0	Top of stack - return address out of range: 0xAAAAAAAA

I still can't find where is the problem is and i can't reproduce it on samples.

[DerekSnell]

Hi @CreativeIndustryAg ,
I see the bus fault occurs in the FlexSPI driver when writing to flash. If you are eXecuting-in-Place (XiP) from that flash while you are also writing to it, the FlexSPI driver should jump to code in ITCM, so it is not XIPing while erasing/writing the flash. I wonder if your ADC reading or interrupt is trying to jump and XIP from flash during the flash write. The app will not be able to execute from flash during those times, and that includes ISRs. So that is where I would focus to troubleshoot this.

Let us know what you find