RT1052 debugger fil

[JRIIS]

Hi

Prevas continue the R&D for the Volvo CE project and use support debugged the IMXRT1050-EVKB rev. A1 that is modified to use QSPI flash.

Prevas have used following guidelines and somehow got it to function, but still miss to download the file – right now they use: '/opt/SEGGER/JLink/Devices/NXP/iMXRT102x/NXP_iMXRT102x_QSPI.elf', Prevas would preferred to use the right version 'iMXRT105x_QSPI.elf' in stead

Is it passable to generate that? Or is perfectly fine to continue use “NXP_iMXRT102x_QSPI.elf”?

Prevas have also used following references and guides:

https://docs.zephyrproject.org/latest/boards/arm/mimxrt1050_evk/doc/index.html?highlight=rt1050#opti...
https://www.nxp.com/docs/en/application-note/AN12108.pdf o

https://docs.zephyrproject.org/latest/boards/arm/mimxrt1050_evk/doc/index.html?highlight=rt1050#opti...

https://www.nxp.com/docs/en/nxp/application-notes/AN12183.pdf

Thanks,

/Jakob

[kerryzhou]

Hi @JRIIS ,

    In the previous time, I mean the old JLINK driver, eg, JLink_V644i, it contains the NXP_iMXRT105x_QSPI.elf, but I find the new Segger JLINK driver already remove the NXP_iMXRT105x_QSPI.elf from folder:

So, if you want to use the JLINK with QSPI firmware for the RT1050, I suggest you use our AE designed RT-UFL, that is very easy to use.

More details, please refer to Jay Heng's blog:

https://github.com/JayHeng/RT-UFL/blob/master/README-en.md

Download it:

https://github.com/JayHeng/RT-UFL/archive/refs/tags/v1.0.zip

Then unzip it, copy the file in

   \RT-UFL-1.0\algo\SEGGER\JLink_Vxxx 

 to:

C:\Program Files\SEGGER\JLink

 Then, when you select the device, select:

  MIMXRT1050_UFL_L0

Then it will helps you use the QSPI flash.

_L0 : QSPI Flash and Octal Flash（Page  size 256 Bytes，Sector size 4KB）

_L1/2 :  Hyper Flash（Page size512 Bytes，Sector size 4KB/64KB）.

Wish it helps you!

Best Regards,

Kerry

[JRIIS]

With your proposal it stopped working - so Prevas started form beginning again, but it is not very stable.

Prevas got several error message, but there terminal have filtered most of the error out so this is this is the only one left:

J-Link>loadfile /home/jepod/repos/pressalit-clean-me/build/zephyr/zephyr.bin 0x60000000

Downloading file [/home/jepod/repos/pressalit-clean-me/build/zephyr/zephyr.bin]...

****** Error: Failed to prepare RAMCode using RAM

Error while determining flash info (Bank @ 0x60000000)

Unspecified error -1

There where likewise issues with erase sectors.

Prevas have also used a board witch is not modified and L1/L2 is selected, but that also stopped working.

Prevas is now one step back using the hack version where they have adjusted in JLinkDevices.xml and use "MIMXRT1052xxxxA" when they program QSPI:

  <Device>

    <ChipInfo Vendor="NXP" Name="MIMXRT1052xxxxA" WorkRAMAddr="0x20000000" WorkRAMSize="0x00080000" Core="JLINK_CORE_CORTEX_M7" Aliases="MCIMXRT1052; MIMXRT1052xxx5A; MIMXRT1052CVL5A; MIMXRT1052xxx6A; MIMXRT1052DVL6A" />

    <FlashBankInfo Name="HyperFlash" BaseAddr="0x60000000" MaxSize="0x01000000" Loader="Devices/NXP/iMXRT102x/NXP_iMXRT102x_QSPI.elf" LoaderType="FLASH_ALGO_TYPE_OPEN" />

  </Device>

Do you have any suggestions to fix this issue with your proposal?

[kerryzhou]

Hi @JRIIS ,

   Do you test any other app code? Eg, test the SDK code, led_blinky, modify it to QSPI flash, then test it, whether it works, just to check whether it is related to the zephyr.bin.

  About the flashloader, do you try my mentioned RT-UFL , that is very good to use with JLINK.

Wish it helps you!

Best Regards,

Kerry

[JRIIS]

Prevas is working continued with the "MIMXRT1052xxxxA" hack R&D implemented.

Would still appreciate to understand why RT-UFL not work/are unstable.

It is still a challenge to use NXP_iMXRT102x_QSPI.elf witch Prevas don’t know the consequence off.

Any chance that you can bring Prevas closer to a solution?

Thanks,

/Jakob

[kerryzhou]

Hi @JRIIS ,

   How do you test the RT-UFL? Do you modify the JLinkDevices.xml, add the RT-UFL?

   In the previous time, I also test  it on my MIMXRT1050-EVKB board which is modified to the QSPI, that totally works OK.

   This is the related blog from Jay heng, the author of RT-UFL:

https://www.cnblogs.com/henjay724/p/14942574.html

https://www.cnblogs.com/henjay724/p/15430619.html

https://www.cnblogs.com/henjay724/p/15448766.html

https://www.cnblogs.com/henjay724/p/15465655.html

If you still have issues about it, you also can send email to Jay directly, as you are our NXP colleague.

Best Regards,

Kerry