Getting Started Guides

First Things First

Prerequisite

Trusted Firmware M provides a reference implementation of the Platform Security Architecture (PSA) specifications, aligning with PSA Certified guidelines. It is assumed that the reader is familiar with the specifications that can be found here.

The current TF-M implementation on Armv8-M leverages TrustZone for Armv8-M so a good understanding of the v8-M architecture is also necessary. A good place to get started with Armv8-M is developer.arm.com.

Build and run instructions

Trusted Firmware M source code is available on git.trustedfirmware.org.

To build & run TF-M:

To port TF-M to a another system or OS, follow the OS Integration Guide

Contributing Guidelines contains guidance on how to contribute to this project.

Set up build environments

TF-M officially supports a limited set of build environments and setups. In this context, official support means that the environments listed below are actively used by team members and active developers, hence users should be able to recreate the same configurations by following the instructions described below. In case of problems, the TF-M team provides support only for these environments, but building in other environments can still be possible.

The following environments are supported:

  1. version supported:

    Ubuntu 18.04 x64+

  2. install dependencies:

sudo apt-get install -y git curl wget build-essential libssl-dev python3 \
python3-pip cmake make

  1. verify cmake version:

cmake --version

Note

Please download CMake version 3.21 or later from https://cmake.org/download/.

  1. add CMake path into environment:

export PATH=<CMake path>/bin:$PATH

Install python dependencies

Clone the TF-M source code, and then install the TF-M’s additional Python dependencies.

  1. get the TF-M source code:

git clone https://git.trustedfirmware.org/TF-M/trusted-firmware-m.git

  1. TF-M’s tools/requirements.txt file declares additional Python dependencies. Install them with pip3:

pip3 install --upgrade pip
cd trusted-firmware-m
pip3 install -r tools/requirements.txt

Install a toolchain

To compile TF-M code, at least one of the supported compiler toolchains have to be available in the build environment. The currently supported compiler versions are:

  • Arm Compiler v6.13 ~ v6.14, v6.18+

    • Download the standalone packages from here.

    • Add Arm Compiler into environment:

      export PATH=<ARM_CLANG_PATH>/bin:$PATH
export ARM_PRODUCT_PATH=<ARM_CLANG_PATH>/sw/mappings

    • Configure proper tool variant and license.

    Note

    Arm compiler v6.15 ~ v6.17 may cause MemManage fault. This defect has been fixed since Arm compiler v6.18. See [SDCOMP-59788] in Armclang v6.18 release note for details.

  • GNU Arm compiler v7.3.1+

    • Download the GNU Arm compiler from here.

    • Add GNU Arm into environment:

      export PATH=<GNU_ARM_PATH>/bin:$PATH

    Note

    GNU Arm compiler version 10-2020-q4-major has an issue in CMSE support. The bug is reported in here. Select other GNU Arm compiler versions instead.

  • IAR Arm compiler v8.42.x, v8.50.x

    • Download IAR build tools from here.

    • Add IAR Arm compiler into environment:

      export PATH=<IAR_COMPILER_PATH>/bin:$PATH

Build AN521 regression sample

Here, we take building TF-M for AN521 platform with regression tests using GCC as an example:

Get the TF-M tests source code:

git clone https://git.trustedfirmware.org/TF-M/tf-m-tests.git

Build SPE and NSPE.

cd </tf-m-tests/tests_reg>
cmake -S spe -B build_spe -DTFM_PLATFORM=arm/mps2/an521 -DCONFIG_TFM_SOURCE_PATH=<TF-M source dir absolute path> \
      -DCMAKE_BUILD_TYPE=Debug -DTFM_TOOLCHAIN_FILE=<TF-M source dir absolute path>/toolchain_GNUARM.cmake \
      -DTEST_S=ON -DTEST_NS=ON \
cmake --build build_spe -- install

cmake -S . -B build_test -DCONFIG_SPE_PATH=<tf-m-tests absolute path>/tests_reg/build_spe/api_ns \
      -DCMAKE_BUILD_TYPE=Debug -DTFM_TOOLCHAIN_FILE=<tf-m-tests absolute path>/tests_reg/build_spe/api_ns/cmake/toolchain_ns_GNUARM.cmake
cmake --build build_test

Run AN521 regression sample

Run the sample code on SSE-200 Fast-Model, using FVP_MPS2_AEMv8M provided by Arm Development Studio.

Note

Arm Development Studio is not essential to develop TF-M, you can skip this section if don’t want to try on Arm develop boards.

  1. install Arm Development Studio to get the fast-model.

    Download Arm Development Studio from here.

  2. Add bl2.axf and tfm_s_ns_signed.bin to symbol files in Debug Configuration menu.

<DS_PATH>/sw/models/bin/FVP_MPS2_AEMv8M  \
--parameter fvp_mps2.platform_type=2 \
--parameter cpu0.baseline=0 \
--parameter cpu0.INITVTOR_S=0x10000000 \
--parameter cpu0.semihosting-enable=0 \
--parameter fvp_mps2.DISABLE_GATING=0 \
--parameter fvp_mps2.telnetterminal0.start_telnet=1 \
--parameter fvp_mps2.telnetterminal1.start_telnet=0 \
--parameter fvp_mps2.telnetterminal2.start_telnet=0 \
--parameter fvp_mps2.telnetterminal0.quiet=0 \
--parameter fvp_mps2.telnetterminal1.quiet=1 \
--parameter fvp_mps2.telnetterminal2.quiet=1 \
--application cpu0=<build_spe>/api_ns/bin/bl2.axf \
--data cpu0=<build_test>/tfm_s_ns_signed.bin@0x10080000

Note

The log is output to telnet by default. It can be also redirected to stdout by adding the following parameter.

--parameter fvp_mps2.UART0.out_file=/dev/stdout

To automatically terminate the fast-model when it finishes running, you can add the following parameters:

--parameter fvp_mps2.UART0.shutdown_on_eot=1

After completing the procedure you should see the following messages on the DAPLink UART (baud 115200 8n1):

...
#### Execute test suites for the Secure area ####
Running Test Suite PSA protected storage S interface tests (TFM_S_PS_TEST_1XXX)...
> Executing 'TFM_S_PS_TEST_1001'
Description: 'Set interface'
TEST: TFM_S_PS_TEST_1001 - PASSED!
> Executing 'TFM_S_PS_TEST_1002'
Description: 'Set interface with create flags'
TEST: TFM_S_PS_TEST_1002 - PASSED!
> Executing 'TFM_S_PS_TEST_1003'
Description: 'Set interface with NULL data pointer'
TEST: TFM_S_PS_TEST_1003 - PASSED!
> Executing 'TFM_S_PS_TEST_1005'
Description: 'Set interface with write once UID'
TEST: TFM_S_PS_TEST_1005 - PASSED!
....

Tool & Dependency overview

To build the TF-M firmware the following tools are needed:

  • C compiler of supported toolchains

  • CMake version 3.21 or later

  • Git

  • gmake, aka GNU Make

  • Python v3.x

  • a set of python modules listed in tools/requirements.txt

Dependency chain

@startuml skinparam state { BackgroundColor #92AEE0 FontColor black FontSize 16 AttributeFontColor black AttributeFontSize 16 } state fw as "Firmware" : TF-M binary state c_comp as "C Compiler" : C99 state python as "Python" : v3.x fw --> c_comp fw --> CMake CMake --> gmake CMake --> Ninja fw --> cryptography fw --> pyasn1 fw --> yaml fw --> jinja2 fw --> cbor2 fw --> click fw --> imgtool c_comp --> GCC c_comp --> CLANG c_comp --> IAR cryptography --> python pyasn1 --> python yaml --> python jinja2 --> python cbor2 --> python click --> python imgtool --> python kconfiglib --> python @enduml

Next steps

Here are some next steps for exploring TF-M:

Copyright (c) 2017-2024, Arm Limited. All rights reserved. Copyright (c) 2022 Cypress Semiconductor Corporation (an Infineon company) or an affiliate of Cypress Semiconductor Corporation. All rights reserved.