bootloader project in nCS

[Ubaid_M]

Hi,

1. Is bootloader file and firmware file, different for projects build on nCS..?
    Or a single executable build file consists both bootloader & firmware..?

2. Can you suggest a sample program to build a bootloader project..?

Thanks,

[Edvin]

Hello,

In NCS, the bootloader is enabled through configurations, mostly. 

I suggest that you check out the documentation of FOTA upgrades:

https://developer.nordicsemi.com/nRF_Connect_SDK/doc/1.9.1/nrf/ug_nrf52.html#fota-upgrades

And there is also a written guide on how to add DFU to your project, that I can recommend:

https://devzone.nordicsemi.com/guides/nrf-connect-sdk-guides/b/software/posts/ncs-dfu

If you struggle with the implementation, I suggest that you follow the guide first, and add DFU to the smp_svr sample from NCS, so that you know that everything around (uploading via nRF Connect for iOS/Android/mcumgr) is working as it should, before you start looking into potential issues in your custom application.

BR,

Edvin

[Ubaid_M]

Hello Edvin,

Edvin said:

And there is also a written guide on how to add DFU to your project,

Towards DFU, i see that we make the following configuration to the project:

# Ensure an MCUboot-compatible binary is generated.
CONFIG_BOOTLOADER_MCUBOOT=y

So,
1. By enabling MCUBOOT what happens..?
2. Where is the bootloader image file in the device memory..?
3. Is the boot loader image file different than firmware image file or bootloader code is integrated into the firmware image file..?

Thanks,

[Ubaid_M]

Hello Edvin,

Thank you for the info,

Further,

1. How does how does the image boot if I have not enabled CONFIG_BOOTLOADER_MCUBOOT=y..?

2. Some preexisting bootloader does the job or if system does not find anything at the address,

Edvin said:

bootloader is located at address 0x0

It will straight away jump to the firmware file address 

Edvin said:

app is located at 0xc200

..??

3. Can you please suggest if this "BOOTLOADER_MCUBOOT" checks for firmware file image code signing before booting the image..?

4. If it does, how does it use the key for code signing check on the firmware file..?

5. How can I update this key..?

Thanks,

[Edvin]

Hello,

Ubaid_M said:

1. How does how does the image boot if I have not enabled CONFIG_BOOTLOADER_MCUBOOT=y..?

Then there is no bootloader. The application will be located at 0x00.

Ubaid_M said:

Some preexisting bootloader does the job or if system does not find anything at the address,

no. See 1.

Ubaid_M said:

3. Can you please suggest if this "BOOTLOADER_MCUBOOT" checks for firmware file image code signing before booting the image..?

Yes it does.

Ubaid_M said:

4. If it does, how does it use the key for code signing check on the firmware file..?

It uses a default key, which you should see when you compile. Some message saying that this is a default key which should not be used for production, and that you should change it.

Ubaid_M said:

5. How can I update this key..?

I believe it says in the same warning.

Best regards,

Edvin

[Ubaid_M]

Thank you Edvin,

Edvin said:

Then there is no bootloader. The application will be located at 0x00.

It is clearer now.

Now if I want to generate a private key and add it into the MCUBOOT, what process do I follow..?

Please suggest.

Thanks

[Edvin]

Please see this link on how to generate new key pairs.

https://developer.nordicsemi.com/nRF_Connect_SDK/doc/latest/tfm/docs/technical_references/design_docs/tfm_secure_boot.html#second-stage-bootloader-in-tf-m

There is also a Kconfig definition

CONFIG_BOOT_SIGNATURE_KEY_FILE (described as "config BOOT_SIGNATURE_KEY_FILE" in NCS\nrf\modules\mcuboot\Kconfig) where you can set the absolute path of the key file that you have generated.

Best regards,

Edvin

[Ubaid_M]

Hello Edvin,

Edvin said:

CONFIG_BOOT_SIGNATURE_KEY_FILE

Thank you for the suggestion, it helps.

Further, I understand the usual way mcuboot works with SLOT-0, SLOT-1 & SCRATCH memory locations towards swapping of images for firmware update.

I want to customize a few things here,

1. I want to map SLOT-1 & SCRATCH memory locations to external serial flash locations.

2. I want to provide mcuboot SPI interface APIs to interact with external flash.

So, for the above 2 requirements, kindly suggest how and where I can tweak & build the mcuboot source code.

If any sample is available wherein mcuboot source code is modified, it would be great.

Thanks,

Ubaid

[Edvin]

Ubaid_M said:

1. I want to map SLOT-1 & SCRATCH memory locations to external serial flash locations.

2. I want to provide mcuboot SPI interface APIs to interact with external flash.

What chip are you using? I believe it is possible to offload this to an external QSPI flash chip by using CONFIG_NORDIC_QSPI_NOR=y. Without knowing the details, perhaps you can look into what this does, and try to implement something similar using SPI. 

The MCUBOOT bootloader project is located in ncs <vX.Y.Z>\bootloader\mcuboot.

BR,
Edvin

[Ubaid_M]

Hello Edvin,

Edvin said:

The MCUBOOT bootloader project is located in ncs <vX.Y.Z>\bootloader\mcuboot.

1. So i can tweak in the and make changes to mcuboot source code from here right..?

Another query i had was with copying of mcuboot,

Edvin said:

So in this case you can see that the bootloader is located at address 0x0

2. WHO does this copying of mcuboot onto device memory location 0x00..?

Thanking you,

Ubaid

[Ubaid_M]

Hello Edvin,

Edvin said:

The MCUBOOT bootloader project is located in ncs <vX.Y.Z>\bootloader\mcuboot.

Where can i find the mcuboot source code..?

Edvin said:

What chip are you using?

I am using nrf52833dk_nrf52833 board with an external flash IS25LQ040B.
The interface is SPI single line & not QSPI.

That is, i need to change mcuboot source code to provide, SPI access to it, so that  mcuboot can copy new firmware file from external flash to internal flash.

Thanks,

[Edvin]

Hello,

Please see the v1.9.1\bootloader\mcuboot\boot\zephyr (that is the project folder, containing the main.c for the bootloader. You can see that if you follow the guide from one of the links, it should print something like 

I: Starting bootloader

I: Primary image: magic=bad, swap_type=0x0, copy_done=0x2, image_ok=0x2

I: Secondary image: magic=unset, swap_type=0x1, copy_done=0x3, image_ok=0x3

I: Boot source: none

I: Swap type: none

I: Bootloader chainload address offset: 0xc000

I: Jumping to the first image slot
*** Booting Zephyr OS build v2.7.99-ncs1  ***

The first and last line is coming from that main.c file.

BR,
Edvin

[Ubaid_M]

Thank you Edvin,

Edvin said:

the bootloader is located at address 0x0

Can you please suggest who copies bootloader code from hex file onto memory location 0x000..?

And also what is the size overhead added by enabling MCU_BOOT on the firmware file..?

Can i compile mcboot separately..?

Thanks,

Ubaid