I would like to compile with CMake a bootloader that looks as below:

my-wolfboot-project/
├─ CMakeLists.txt
├─ src/
│  ├─ main.c             # Your custom entry point
│  └─ hw_init.c          # Optional hardware setup
├─ include/
│  └─ hw_init.h
└─ wolfboot/             # WolfBoot source (submodule or copy)
   ├─ ... (wolfBoot sources)

In main.c, I'll have a main function that performs some specific hardware configuration (GPIO, etc.) to then call "loader_main" and let Wolfboot does its magic.

I cannot find a place in the documentation where this case is described. I do not understant if I need to include Wolfboot sources or link against a static library.

Does anyone has some experience regarding this approach?

Thanks!

Hash generated from tools/keytools/sign differs from the sha256 calculated from the bootloader when I compile my application with -g -g3 or -Ofast -g

To be precise, it fails in the integrity check (verified using wolfboot config DEBUG=1)

int wolfBoot_start(void)
{
    struct wolfBoot_image os_image;
    int ret = 0;
    memset(&os_image, 0, sizeof(os_image));

    os_image.hdr = (uint8_t*)gImage;

    if ((ret = wolfBoot_open_image_address(&os_image, (uint8_t*)gImage)) < 0) {
        goto exit;
    }

    if ((ret = wolfBoot_verify_integrity(&os_image)) < 0) {
        goto exit; // <= Application images compiled with debug flags fail this check
    }

    if ((ret = wolfBoot_verify_authenticity(&os_image)) < 0) {
        goto exit;
    }

    wolfBoot_printf("Firmware Valid\n");

    do_boot((uint32_t*)os_image.fw_base);
    /* ... */
}

Are there any tips as to what I should look to integrate my app with the bootloader? I have exhausted all online docs I could find.

The app was created long before we needed a bootloader, so there had been no considerations taken. It is an STM32h7 project mostly bootstrapped with CubeMX.

WolfBoot version v2.5.0

My project file tree

/build.sh
/CMakeLists.txt
/bootloader/wolfBoot/
link_with_bootloader.ld
link_no_bootloader.ld
/* ... Application source files ... */

I have zip'ped relevant build files as an attachment

ARCH?=ARM
TARGET?=stm32h7
SIGN?=ECC256
HASH?=SHA256
DEBUG?=1
DEBUG_UART?=0
VTOR?=1
NO_ASM?=0
EXT_FLASH?=0
SPI_FLASH?=0
QSPI_FLASH?=0
OCTOSPI_FLASH?=0
ALLOW_DOWNGRADE?=0
NVM_FLASH_WRITEONCE?=0
WOLFBOOT_VERSION?=1
V?=0
SPMATH?=1
RAM_CODE?=0
DUALBANK_SWAP?=0
WOLFBOOT_PARTITION_SIZE?=0xC0000
WOLFBOOT_SECTOR_SIZE?=0x20000
WOLFBOOT_PARTITION_BOOT_ADDRESS?=0x8020000
WOLFBOOT_PARTITION_UPDATE_ADDRESS?=0x80E0000
WOLFBOOT_PARTITION_SWAP_ADDRESS?=0x81A0000
CFLAGS_EXTRA?=-mfloat-abi=hard -mfpu=fpv4-sp-d16
PULL_LINKER_DEFINES?=1

Notes

- I always bin-assemble the bootloader.bin and the application_v1_signed.bin into a factory.bin before flashing it to the target
- I changed the WOLFBOOT_PARTITION_SIZE from 0xD0000 to 0xC0000, because 0xD0000 is NOT a multiple of WOLFBOOT_SECTOR_SIZE=0x20000
- I made an early return in the clock setup for the bootloader. My board is custom and the configuration failed when setting up the external high-speed oscillator (HSE)

/* This implementation will setup HSI RC 64 MHz as System Clock Source */
static void clock_pll_on(int powersave)
{
    /* ... */
    pllm = 1;
    plln = 120;
    pllp = 2;
    pllq = 20;
    pllr = 2;
    d1cpre =   RCC_PRESCALER_DIV_NONE;
    hpre  =    RCC_PRESCALER_DIV_2;
    d1ppre =  (RCC_PRESCALER_DIV_2 >> 1);
    d2ppre1 = (RCC_PRESCALER_DIV_2 >> 1);
    d2ppre2 = (RCC_PRESCALER_DIV_2 >> 1);
    d3ppre =  (RCC_PRESCALER_DIV_2 >> 1);
    flash_waitstates = 4;

    flash_set_waitstates(flash_waitstates);

    /* Enable internal high-speed oscillator. */
    RCC_CR |= RCC_CR_HSION;
    DMB();
    while ((RCC_CR & RCC_CR_HSIRDY) == 0) {};

    /* Select HSI as SYSCLK source. */
    reg32 = RCC_CFGR;
    reg32 &= ~((1 << 2) |(1 << 1) | (1 << 0));
    RCC_CFGR = (reg32 | RCC_CFGR_SW_HSISYS);
    DMB();
    return; // <= Early return
    /* Enable external high-speed oscillator. */
    reg32 = RCC_CR;
    reg32 |= RCC_CR_HSEBYP;
    /* ... */
}

I have entered the following in macOS terminal:

git clone https://github.com/wolfSSL/wolfboot.git

cd wolfboot

cp ./config/examples/stm32u5.config .config && make

After this I receive the following error:

[CC-ARM] src/string.o
In file included from include/image.h:35,
                 from src/string.c:45:
include/wolfboot/wolfboot.h:167:13: fatal error: wolfssl/wolfcrypt/sha256.h: No such file or directory
  167 | #   include "wolfssl/wolfcrypt/sha256.h"
      |             ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make: *** [src/string.o] Error 1

I'm pretty lost here as I'm not familiar with Make or with WolfBoot. I hope someone can give me some advice.

What is the reason behind it ? Why not 0xA000_0000 ?

Is there a way to edit this and recompile everything? It seems that to edit the UCB STAD address I need a licence for winIDEA that I do not own and I dont know how to get it

I'm referring to this page https://www.wolfssl.com/products/wolfhsm/

Thank you

trying to evaluate the product, using wolfBoot version 2.0.2 and gcc-arm-none-eabi (gcc 13.2.1) toolchain.

using preconfigured configuration files (from config/examples) for now, just to get started.

compilation of STM32H7 and STM32L5 targets were successful.

STM32G0 compilation gives warnings about negative array indices.

STM32F4 might be missing something, it does not compile properly, gives error about overlapping sections; a F4-targeted linker script is not found in test-app (while H7 & L5 & G0 each have their own scripts).

I was sort of expecting the provided examples to work out-of-the-box. any idea what could be wrong, do I _need_ to use 'make config'?

The makefile contains a general rule for building .o files from .c files. This seems not to work.

I haven't modified any file I fetched from Github. So are there some environment settings missing on my PC?

I am trying to creating the factory.bin for stm32l452 using the .config file given in the /examples directory. I see that in the hal directory file- stm32l4.c, stm32l4xx_hal.h file is required. This file follows more dependencies which I have included. However, I get the following error when I run the make command.

make clean && make
/bin/sh: line 1: test: 0x200: integer expression expected
/bin/sh: line 1: test: 0x200: integer expression expected
/bin/sh: line 1: test: 0x200: integer expression expected
/bin/sh: line 1: test: 0x200: integer expression expected
/bin/sh: line 1: test: 0x200: integer expression expected
\t[CC-ARM] hal/stm32l4.o
\t[CC-ARM] src/string.o
\t[CC-ARM] src/image.o
\t[CC-ARM] src/libwolfboot.o
\t[CC-ARM] src/keystore.o
\t[CC-ARM] src/loader.o
\t[CC-ARM] src/boot_arm.o
make: *** No rule to make target '/home/ghost/STM32Cube/Repository/STM32Cube_FW_L4_V1.17.0/Drivers/STM32L4xx_HAL_Driver/Src/stm32l4xx_hal_flash.o', needed by 'wolfboot.elf'.  Stop.

I am not sure how to correct this error. I know that the file- stm32l4xx_hal_flash needs to be included as well but when i include it in the hal or src directory, the error remains. Please let me know what I can do fix this.

Thank you!

I've written the appropriate HAL functions for write and erase, etc.

The problem I'm having is that when I call wolfBoot_success(), it does not set the sector flags at the end of the boot partition as expected.

I can step it through and see it call wolfBoot_set_partition_state() in libwolfboot.c, then call set_partition_magic(). This works and I can see the flash is now programmed with "BOOT" at address 0x80117FC to 0x80117FF (ie. the end of the boot partition, as expected).

Then "state = get_partition_state(part);" is called, which erases the page that it just programmed. If I step through get_partition_state() I can see it ultimately calls nvm_select_fresh_sector() and it's in here that it erases the same flash page that it just programmed.

In nvm_select_fresh_sector(), this code:

    /* Erase the non-selected partition */
    addr_align = (uint32_t)(base - ((!sel) * WOLFBOOT_SECTOR_SIZE))
        & (~(NVM_CACHE_SIZE - 1));
    if (*((uint32_t*)(addr_align + WOLFBOOT_SECTOR_SIZE - sizeof(uint32_t)))
            != FLASH_WORD_ERASED) {
        hal_flash_erase(addr_align, WOLFBOOT_SECTOR_SIZE);
    }

Seems to want to erase the wrong page:
If "sel == 0", then:
addr_align = 0x8011800 - (1 * 0x400) = 0x8011400. So this will erase 0x8011400 to 0x80117FF.
If "sel == 1", then:
addr_align = 0x8011800 - (0 * 0x400) = 0x8011800. So this will erase 0x8011800 to 0x8011BFF.
This is outside the boot partition (ie. it's the start of 'update' partition).

How is this supposed to work? Have I got some configuration wrong somewhere?

Thanks,
Gareth.

# Transform it to object so it can be linked to final elf
../build/obj/image_v1_signed.o: ../build/image_v1_signed.bin
    @echo "\t[OBJCOPY]" $^ " --> " $@
    @$(OBJCOPY) -v -I binary -O elf32-littlearm -B arm \
        --rename-section  .data=.image_app \
        --set-section-alignment .data=8 \
        --set-section-flag .data=code \
        $^ $@

wolfboot.o: wolfboot.bin
    @echo "\t[OBJCOPY]" $^ " --> " $@
    @$(OBJCOPY) -I binary -O elf32-littlearm -B arm --rename-section .data=.boot3,code,alloc,contents $^ $@

As i'm not experimented with wolfboot, i'm not quiet sure that this solution is the best and i wanted to have your opinion about this. Porting Wolfboot on pico was only for a PoC during my project that's why you may find some nasty code but in case you are interested my modifications, here is my fork: https://github.com/Thisora/wolfBoot-rp2040.

I didn't plane to work anymore on this solution but if you like my suggestion i can submit a PR.

I'm looking for information on supported image file formats.

I don't see any mention of image file formats anywhere in the documentation but "Signing.md" provides only examples for binary images.

Is Intel hex format supported as well?

Thanks,
Galadrius

I'm new to this wonderful wolfBoot. I'm very keen on using this with my new project on STM32F7. But every time I tried to build it on window, some weird error occurred. To be honest i'm not so familiar with using Cygwin/Msys2 to build library on window OS.
Last time I tried build it on Ubuntu, everything seem to be good until an error appear:

Traceback (most recent call last):
  File "tools/keytools/keygen.py", line 25, in <module>
    from wolfcrypt import ciphers
  File "/usr/local/lib/python3.6/dist-packages/wolfcrypt/ciphers.py", line 23, in <module>
    from wolfcrypt._ffi import ffi as _ffi
ImportError: libwolfssl.so.24: cannot open shared object file: No such file or directory
Makefile:220: recipe for target 'ed25519.der' failed
make: *** [ed25519.der] Error 1

Ofcourse, it's my mistake. But could you, anyone, please guide me step by step how to build/ make this bootloader on window/ ubuntu OS.
Thank in advance.

I am building on windows 10 using MSYS/mingw and have the GNU MCU Eclipse build tools (2.12-20190422-1053) together with the arm-none-eabi-gcc\8.2.1-1.7.1 gcc version and am targeting the STM32F4 device.

In both cases when I run
>make config [all defaults seem appropriate for STM32F4 target]
>make

I see the error

>make: *** [Makefile:142: ed25519.der] Error 9009

Looking into this the make command is: @python3 tools/keytools/keygen.py $(KEYGEN_OPTIONS) src/ed25519_pub_key.c

The problem is in neither version of the repository (archive download or github) can I find the file src/ed25519_pub_key.c.

>python3 tools\keytools\keygen.py --ed25519 src\ed25519_pub_key.c

I think the python script is supposed to create this file but calling from the command line does not produce an error or a file.

[edit::]

So it seems I am missing the wolfcrypt python package since the import fails when I run the keygen, do I need to download and build this entirely seperately or because there does seem to be some of it availabl in the wolfboot libs directory.

I seem to be missing something fundamental here. What have I done wrong?