This example demonstrates using wolfBoot on the Infineon AURIX TC3xx family of microcontrollers. The example is based on the TC375 Lite-Kit V2, but should be easily adaptable to other TC3xx devices. This README assumes basic familiarity with the TC375 SoC, the AURIX IDE, and Lauterbach Trace32 debugger.
- WolfBoot on Infineon AURIX TC3xx
The example contains two projects: wolfBoot-tc3xx
and test-app
. The wolfBoot-tc3xx
project contains the wolfBoot bootloader, and the test-app
project contains a simple firmware application that will be loaded and executed by wolfBoot. The test-app
project is a simple blinky application that blinks LED2 on the TC375 Lite-Kit V2 once per second when running the base image, and rapidly (~3x/sec) when running the update image. The test app determines if it is a base or update image by inspecting the firmware version (obtained through the wolfBoot API). The firmware version is set in the image header by the wolfBoot keytools when signing the test app binaries. The same test app binary is used for both the base and update images, with the only difference being the firmware version set by the keytools.
- In the TC375 UCBs, BMDHx.STAD must point to the wolfBoot entrypoint
0xA00A_0000
. You can modify this in theUCB
section of the TRACE32 IDE as described in the steps later in this document. Please refer to the TRACE32 manual and the TC37xx user manual for more information on the UCBs. - Because TC3xx PFLASH ECC prevents reading from erased flash, the
EXT_FLASH
option is used to redirect flash reads to theext_flash_read()
HAL API, where the flash pages requested to be read can be blank-checked by hardware before reading. - TC3xx PFLASH is write-once (
NVM_FLASH_WRITEONCE
), however wolfBootNVM_FLASH_WRITEONCE
does not supportEXT_FLASH
. Therefore the write-once functionality is re-implemented in theHAL
layer. - This demo app is only compatible with the GCC toolchain build configurations shipped with the AURIX IDE. The TASKING compiler build configurations are not yet supported.
The TC3xx AURIX port of wolfBoot places all images in PFLASH, and uses both PFLASH0 and PFLASH1 banks. The wolfBoot executable code and the image swap sector are located in PFLASH0, with the remainder available for use. PFLASH1 is divided in half, with the first half holding the BOOT partition and the second half holding the UPDATE partition. User firmware images are directly executed in place from the BOOT partition in PFLASH1, and so must be linked to execute within this address space, with an offset of IMAGE_HEADER_SIZE
to account for the wolfBoot image header.
+==========+
| PFLASH0 |
+----------+ <-- 0x8000_0000
| Unused | 640K
+==========+ <-- 0x800A_0000
| wolfBoot | 172K
+----------+ <-- 0x8002_B000
| Unused | ~2.8M
+----------+ <-- 0x8030_0000
+==========+
| PFLASH1 |
+==========+ <-- 0x8030_0000
| BOOT | 1.5M (0x17E000)
+----------+ <-- 0x8047_E000
| UPDATE | 1.5M (0x17E000)
+----------+ <-- 0x805F_C000
| SWAP | 16K (0x4000)
+----------+ <-- 0x8060_0000
Please refer to the wolfBoot and test-app linker script templates for the exact memory configuration.
wolfBoot relies extensively on its build system in order to properly set configuration macros, linker addresses, and other target-specific items. Because wolfBoot on AURIX uses the AURIX studio IDE to build, changing things like the signature algorithm would require manually editing IDE settings, linker scripts, etc. which is error prone and tedious. The wbaurixtool.sh
script provides a single tool that automates the generation of all configurable items required for building wolfBoot and the test application on aurix given the chosen signature and hashing algorithms, including managing all configuration macros, linker scripts, as well as handling the actual key generation and image signing process. wbaurixtool.sh
can also generate wolfHSM NVM images containing the generated image signing key when used in conjunction with wolfHSM.
The general usage of wbaurixtool.sh
is as follows:
$ ./wbaurixtool.sh [global options] <subcommand> [subcommand options]
where <subcommand>
is one of the following:
keygen
: Generate a new signing key pairsign
: Sign a firmware imagetarget
: Generate thetarget.h
header file for the tc375 flash configurationmacros
: Generate thewolfBoot_macros.txt
file from thewolfBoot_macros.in
template in thewolfBoot-tc3xx
directorynvm
: Generate a wolfHSM NVM image containing the signing key, based on the nvm configuration filetools/scripts/tc3xx/wolfBoot-wolfHSM-keys.nvminit
lcf
: Generate theLcf_Gnu_Tricore_Tc.lsl
file from theLcf_Gnu_Tricore_Tc.lcf.in
template in thetest-app
directoryclean
: Remove all generated keys and images
The keygen
, sign
, macros
and lcf
commands will use ecc256
as the default signature algorithm, and sha256
as the default hashing algorithm. Each subcommand can take the --sign-algo
and --hash-algo
options to specify different algorithms.
Subcommands can be chained together, and can inherit options from previous commands. For example, ./wbaurixtool.sh keygen --sign-algo ecc256 macros lcf
will generate a new signing key pair using ECC 256, generate the wolfBoot_macros.txt
file from the wolfBoot_macros.in
template in the wolfBoot-tc3xx
directory, and generate the Lcf_Gnu_Tricore_Tc.lsl
file from the Lcf_Gnu_Tricore_Tc.lcf.in
template in the test-app
directory. If the global option --hsm
is specified, then the subsequent subcommands will apply to the wolfHSM AURIX projects instead of the standard projects (e.g. wolfBoot-tc3xx-wolfHSM
and test-app-wolfHSM
).
wbaurixtool.sh
can also be used to generate wolfHSM NVM images containing the generated image signing key via the nvm
subcommand when used in conjunction with wolfHSM. See the wolfHSM Compatibility section for more information.
For more information on the wbaurixtool.sh
script, run ./wbaurixtool.sh --help
for a full list of options and subcommands.
- A Windows 10 computer with the Infineon AURIX IDE installed
- A WSL2 distro (tested on Ubuntu 22.04) with the
build-essential
package installed (sudo apt install build-essential
) - A TC375 AURIX Lite-Kit V2
- Clone the wolfBoot repository and initialize the repository submodules (
git submodule update --init
)
- Open a WSL2 terminal and navigate to the top level
wolfBoot
directory - Compile the keytools by running
make keytools
- Use the helper script to generate a new signing key pair using ECC 256
- Navigate to
wolfBoot/tools/scripts/tc3xx
- Run
./wbaurixtool.sh keygen --sign-algo ecc256 macros lcf
. This:- Generates the signing private key
wolfBoot/priv.der
and adds the public key to the wolfBoot keystore (see keygen for more information) - Generates the
wolfBoot_macros.txt
file from thewolfBoot_macros.in
template in thewolfBoot-tc3xx
directory, which sets the appropriate wolfBoot preprocessor macros based on the hash and signature algorithms. ThewolfBoot_macros.txt
file is then passed to the compiler in the AURIX project - Generates the
Lcf_Gnu_Tricore_Tc.lsl
file from theLcf_Gnu_Tricore_Tc.lcf.in
template in thetest-app
directory, which sets appropriate values for Linker addresses (e.g. wolfBoot header size)based on the selected signature algorithm. TheLcf_Gnu_Tricore_Tc.lsl
file is then passed to the linker in the AURIX project - Note that if you already have generated keys, you can use
./wbaurixtool.sh clean
to remove them first
- Generates the signing private key
- Navigate to
$ ./wbaurixtool.sh keygen --sign-algo ecc256 macros
Generating keys with algorithm: ecc256
Keytype: ECC256
Generating key (type: ECC256)
Associated key file: priv.der
Partition ids mask: ffffffff
Key type : ECC256
Public key slot: 0
Done.
Generating macros file with sign_algo=ecc256, hash_algo=sha256
Because of repository size constraints and differing licenses, the required Infineon low level drivers ("iLLD") and auto-generated SDK configuration code that are usually included in AURIX projects are not included in this demo app. It is therefore required to locate them in your AURIX install and extract them to the location that the wolfBoot AURIX projects expect them to be at. The remainder of these instructions will use variables to reference the following three paths:
$AURIX_INSTALL
: The AURIX IDE installation location. This is usuallyC:\Infineon\AURIX-Studio-<version>
$SDK_ARCHIVE
: The zip archive of the iLLD SDK. This is usually at$AURIX_INSTALL\build_system\bundled-artefacts-repo\project-initializer\tricore-tc3xx\<version>\iLLDs\Full_Set\iLLD_<version>__TC37A.zip
$SDK_CONFIG
: The directory containing the iLLD SDK configuration for the specific chip. This is usually at$AURIX_INSTALL\build_system\bundled-artefacts-repo\project-initializer\tricore-tc3xx\<version>\ProjectTemplates\TC37A\TriCore\Configurations
Perform the following two steps to add the iLLD SDK drivers to the wolfBoot project:
- Extract the iLLD package for the TC375TP from
$SDK_ARCHIVE
into thewolfBoot/IDE/AURIX/SDK
directory. The contents of thewolfBoot/IDE/AURIX/SDK
directory should now be:
wolfBoot/IDE/AURIX/SDK
├── Infra/
├── Service/
├── iLLD/
└── placeholder.txt
- Copy the SDK configuration sources from
$SDK_CONFIG
into thewolfBoot/IDE/AURIX/Configurations
directory. The contents of thewolfBoot/IDE/AURIX/Configurations
directory should now be:
wolfBoot/IDE/AURIX/Configurations/
├── Debug
├── Ifx_Cfg.h
├── Ifx_Cfg_Ssw.c
├── Ifx_Cfg_Ssw.h
├── Ifx_Cfg_SswBmhd.c
└── placeholder.txt
- Generate the 'target.h` header file for the tc375 flash configuration
- Open a WSL terminal and navigate to
wolfBoot/tools/scripts/tc3xx
- Run
./wbaurixtool.sh target
- Open a WSL terminal and navigate to
- Open the AURIX IDE and create a new workspace directory, if you do not already have a workspace you wish to use
- Import the wolfBoot project
- Click "File" -> Open Projects From File System"
- Click "Directory" to select an import source, and choose the wolfBoot/IDE/AURIX/wolfBoot-tc3xx directory in the system file explorer
- Click "Finish" to import the project
- Build the wolfBoot Project
- Right-click the wolfBoot-tc3xx project and choose "Set active project"
- Right-click the wolfBoot-tc3xx project, and from the "Build Configurations" -> "Set Active" menu, select either the "TriCore Debug (GCC)" or "TriCore Release (GCC)" build configuration
- Click the hammer icon to build the active project. This will compile wolfBoot.
- Import the test-app project using the same procedure as in step (3), except using
wolfBoot/IDE/AURIX/test-app
as the directory - Build the test-app project using the same procedure as in step (4), except choosing the
test-app
eclipse project. Note that the build process contains a custom post-build step that converts the applicationelf
file into a.bin
file usingtricore-elf-objcopy
, which can then be signed by the wolfBoot key tools in the following step - Sign the generated test-app binary using the wolfBoot keytools
- Open a WSL terminal and navigate to
wolfBoot/tools/scripts/tc3xx
- Run
./wbaurixtool.sh sign --debug
or./wbaurixtool.sh sign
to sign either the debug or release build, respectively. This creates the signed image filestest-app_v1_signed.bin
andtest-app_v2_signed.bin
in the test-app output build directory. The v1 image is the initial image that will be loaded to theBOOT
partition, and the v2 image is the update image that will be loaded to theUPDATE
partition.
- Open a WSL terminal and navigate to
$ ./wbaurixtool.sh sign
Signing binaries with ecc256 and sha256
wolfBoot KeyTools (Compiled C version)
wolfBoot version 2020000
Update type: Firmware
Input image: ../../../IDE/AURIX/test-app/TriCore Release (GCC)/test-app.bin
Selected cipher: ECC256
Selected hash : SHA256
Public key: ../../../priv.der
Output image: ../../../IDE/AURIX/test-app/TriCore Release (GCC)/test-app_v1_signed.bin
Target partition id : 1
image header size calculated at runtime (256 bytes)
Calculating SHA256 digest...
Signing the digest...
Output image(s) successfully created.
wolfBoot KeyTools (Compiled C version)
wolfBoot version 2020000
Update type: Firmware
Input image: ../../../IDE/AURIX/test-app/TriCore Release (GCC)/test-app.bin
Selected cipher: ECC256
Selected hash : SHA256
Public key: ../../../priv.der
Output image: ../../../IDE/AURIX/test-app/TriCore Release (GCC)/test-app_v2_signed.bin
Target partition id : 2
image header size calculated at runtime (256 bytes)
Calculating SHA256 digest...
Signing the digest...
Output image(s) successfully created.
- Ensure the Lauterbach probe is connected to the debug port of the tc375 LiteKit
- Open Trace32 Power View for Tricore
- Open the SYStem menu and click "DETECT" to detect the tc375 device. Click "CONTINUE" in the pop-up window, and then choose "Set TC375xx" when the device is detected
The default Boot Mode Header (BMHD) start address on a new TC375 0xA0000000
but the wolfBoot application has a start address of 0xA00A0000
. We must therefore update the BMHD UCBs with the correct entry point such that it can boot wolfBoot out of reset.
- Select the TC37x dropdown menu and click UCBs
- Expand
BMHD0_COPY
- Click "Edit"
- Set the
STAD
to0xA00A0000
- Click "Update" to recompute the CRC
- Click "Check" to verify the new CRC
- Click "Write" to update the UCB in flash
- Perform the same operations (2-7) on the
BMHD0_ORIG
UCB
The device is now configured to boot from 0xA00A0000
out of reset.
We can now load wolfBoot and the firmware application images to the tc3xx device using Trace32 and a Lauterbach probe
- Click "File" -> "ChangeDir and Run Script" and choose the
wolfBoot/tools/scripts/tc3xx/wolfBoot-loadAll-$BUILD.cmm
script, where $BUILD should be either "debug" or "release" depending on your build type in (4) and (6).
wolfBoot and the demo applications are now loaded into flash, and core0 will be halted at the wolfBoot entry point (core0_main()
).
-
Run the application by clicking "Go" to release the core. This will run wolfBoot which will eventually boot into the application in the
BOOT
partition. You should see LED2 on the board blink once per second. -
Reset the application to trigger the firmware update. Click "System Down", "System Up", then "Go" to reset the tc3xx. If the device halts again at
core0_main
, click "Go" one more time to release the core. You should see LED2 turn on for ~5sec while wolfBoot swaps the images betweenUPDATE
andBOOT
partitions, then you should see LED2 blink rapidly (~3x/sec) indicating that the firmware update was successful and the new image has booted. Subsequent resets should continue to boot into to the new image.
To rerun the demo, simply rerun the loader script in Trace32 and repeat the above steps
wolfBoot has full support for wolfHSM on the AURIX TC3xx platform. The wolfBoot application functions as the HSM client, and all cryptographic operations required to verify application images are offloaded to the HSM. When used in tandem with wolfHSM, wolfBoot can be configured to use keys stored on the HSM for cryptographic operations, or to store keys in the default keystore and send them on-demand to the HSM for usage. The former option is the default configuration, and is recommended for most use cases, as key material will never leave the secure boundary of the HSM. The latter option is useful for development and testing, before keys have been preloaded onto the HSM.
Note that information regarding the AURIX TC3xx HSM core is restricted by NDA with Infineon. Source code for the wolfHSM TC3xx platform port is therefore not publicly available and cannot be included for distribution in wolfBoot. Instructions to build wolfBoot with wolfHSM compatibility are provided here, but the wolfHSM TC3xx port must be obtained separately from wolfSSL. To obtain the wolfHSM TC3xx port, please contact wolfSSL at [email protected].
Steps to build wolfBoot on TC3xx with wolfHSM are largely similar to the non-HSM case, with a few key differences.
- Obtain the wolfHSM release for the AURIX TC3xx from wolfSSL
- Extract the contents of the
infineon/tc3xx
directory from the wolfHSM TC3xx release you obtained from wolfSSL into the wolfBoot/IDE/AURIX/wolfHSM-infineon-tc3xx directory. The contents of this directory should now be:
IDE/AURIX/wolfHSM-infineon-tc3xx/
├── README.md
├── T32
├── placeholder.txt
├── port
├── tchsm-client
├── tchsm-server
├── wolfHSM
└── wolfssl
- Build the wolfHSM server application and load it onto the HSM core, following the instructions provided in the release you obtained from wolfSSL. You do not need to build or load the demo client application, as wolfBoot will act as the client.
- Follow all of the steps in Building and Running the wolfBoot Demo for the non-HSM enabled case, but with the following key differences:
- The wolfBoot-tc3xx-wolfHSM AURIX Studio project should be used instead of
wolfBoot-tc3xx
- Use the
wolfBoot-wolfHSM-loadAll-XXX.cmm
lauterbach scripts instead ofwolfBoot-loadAll-XXX.cmm
to load the wolfBoot and test-app images in the TRACE32 GUI - Provide the
--hsm
global option to thewbaurixtool.sh
script when invoking it, so the wolfHSM projects are used instead of the standard wolfBoot projects - If using the default build options in wolfBoot-tc3xx-wolfHSM, wolfBoot will expect the public key for image verification to be stored at a specific keyId for the wolfBoot client ID. You can use whnvmtool to generate a loadable NVM image that contains the required keys automatically via
wbaurixtool.sh
through thenvm
subcommand. This generates an NVM image containing the generated image signing key based on the wolfBoot-wolfHSM-keys.nvminit configuration file, which can then be loaded to the device via a flash programming tool. See thewhnvmtool
documentation and the documentation included in your wolfHSM AURIX release for more details. Note: if you want to use the standard wolfBoot keystore functionality in conjunction with wolfHSM for testing purposes (doesn't require pre-loading keys on the HSM) you can configure wolfBoot to send the keys to the HSM on-the-fly as ephemeral keys. To do this, ensureWOLFBOOT_USE_WOLFHSM_PUBKEY_ID
is NOT defined, and add the--localkeys
argument to then./wbaurixtool.sh keygen
command, which invokes thekeygen
tool without the default--nolocalkeys
option.
- The wolfBoot-tc3xx-wolfHSM AURIX Studio project should be used instead of
The following pseudo command sequence shows a brief overview of the commands needed to build wolfBoot on AURIX (optionally with wolfHSM). The signature and hashing algorithms used in the example are ECC 256 and SHA 256 and specified explicitly for clarity. Note that these algorithms are the default, so do not need to be explicitly specified. Optional arguments are shown in square brackets (e.g. if targeting wolfHSM, the --hsm
option must be provided as a global option to wbaurixtool.sh
).
# Navigate to wolfBoot directory
WOLFBOOT_DIR=/path/to/wolfBoot
SCRIPTS_DIR=$WOLFBOOT_DIR/tools/scripts/tc3xx
cd $WOLFBOOT_DIR
# Copy source files to appropriate location as listed in the steps above
# ...
# Start with a clean build
make clean && make keysclean && cd $WOLFBOOT_DIR/tools/keytools && make clean
cd $SCRIPTS_DIR && ./wbaurixtool.sh clean
# Delete any build artifacts in wolfBoot-tc3xx (or wolfBoot-tc3xx-wolfHSM) and test-app (or test-app-wolfHSM) AURIX Studio projects
# ...
# Make keytools (NOTE: THIS OVERRIDES TARGET.H WITH SIM VALUES)
cd $WOLFBOOT_DIR
make keytools
# Generate target.h
cd $SCRIPTS_DIR
./wbaurixtool.sh target
# Generate keys, as well as configuration macros and linker script based on the selected signature algorithm
./wbaurixtool.sh [--hsm] keygen --sign-algo ecc256 --hash-algo sha256 macros lcf
# If using wolfHSM, generate key NVM image
./wbaurixtool.sh nvm
# Load NVM image hexfile to the device
# ...
# Build wolfHSM AURIX Studio project
# ....
# Build test-app AURIX Studio project
# ....
# Sign test app
./wbaurixtool.sh [--hsm] sign --sign-algo ecc256 --hash-algo sha256 [--debug]
# Load wolfBoot + app in Lauterbach using tools/scripts/tc3xx/wolfBoot-loadAll-XXX.cmm
# ...
When running a shell script in WSL, you may see the following error:
$ ./wbaurixtool.sh:
/bin/bash^M: bad interpreter: No such file or directory
This occurs because your local git repository is configured with the default core.autocrlf true
configuration, meaning that git is checking out files with Windows-style CRLF line endings which the bash interpreter cannot handle. To fix this, you need to either configure git to not checkout windows line endings for shell scripts (GitHub docs), or you can run the dos2unix
(sudo apt install dos2unix
) utility on the script before running it.
When compiling wolfBoot to use wolfHSM with ML-DSA for verification, you must ensure to compile the keytools with a .config file specifying the ML-DSA parameters. Otherwise, the image header size will be incorrect, and attempts to boot the application will fail in unpredictable ways. Before compiling the keytools in the steps above, copy [config/examples/sim-wolfHSM-mldsa.config] to .config
. You can now run make keytools
and the keytools will be able to handle ML-DSA keys.