INSTALL

0. Building on *nix from git repository

    Run the autogen script to generate configure, then proceed to step 1.
    Prerequisites: You'll need autoconf, automake and libtool installed.

    $ ./autogen.sh

1. Building on *nix from a release

    $ ./configure
    $ make
    $ make check   # (optional, but highly recommended)
    $ sudo make install

    Note: Building with configure generates a wolfssl/options.h file that contains
    all the generated build options. This file needs to be included in your application
    before any other wolfSSL headers. Optionally your application can define
    WOLFSSL_USE_OPTIONS_H to do this automatically.

2. Building on iOS

    Use on the xcode project in IDE/iOS/wolfssl.xcodeproj
    There is a README in IDE/iOS with more information

3. Building for Apple ARM64

    When building for an Apple ARM64 platform, ensure the host CPU type is detected as "aarch64" during configure, if not, pass --host=aarch64-apple-darwin to configure.

4. Building on Windows

    Use the Visual Studio Solution wolfssl64.sln

5. Building with IAR

    Please see the README in IDE/IAR-EWARM for detailed instructions

6. Building with Keil

    Please see the Keil Projects in IDE/MDK5-ARM/Projects

7. Building with Microchip tools

    Please see the README in mplabx

8. Building with Freescale MQX

    Please see the README in mqx

9. Building with Rowley CrossWorks for ARM

    Use the CrossWorks project in IDE/ROWLEY-CROSSWORKS-ARM/wolfssl.hzp
    There is a README.md in IDE/ROWLEY-CROSSWORKS-ARM with more information

10. Building with Arduino

    Use the script IDE/ARDUINO/wolfssl-arduino.sh to reformat the wolfSSL
    library for compatibility with the Arduino IDE. There is a README.md in
    IDE/ARDUINO for detailed instructions.

11. Building for Android with Visual Studio 2017

    Please see the README in IDE/VS-ARM.
    Use the Visual Studio solution IDE/VS-ARM/wolfssl.sln.

12. Building for Yocto Project or OpenEmbedded

    Please see the README in the "meta-wolfssl" repository. This repository
    holds wolfSSL's Yocto and OpenEmbedded layer, which contains recipes
    for wolfSSL, wolfSSH, wolfMQTT, wolfTPM, wolfCrypt examples, and OSS
    project bbappend files.

    https://github.com/wolfssl/meta-wolfssl

    The wolfSSL recipe can also be found in the OpenEmbedded
    "meta-openembedded/meta-networking/recipes-connectivity" layer:

    https://github.com/openembedded/meta-openembedded

13. Porting to a new platform

    Please see section 2.4 in the manual:
    https://www.wolfssl.com/documentation/manuals/wolfssl/chapter02.html#customizing-or-porting-wolfssl

14. Building with CMake
    Note: Primary development uses automake (./configure). The support for CMake
    is still under development.

    For configuring wolfssl using CMake, we recommend downloading the CMake
    GUI (https://cmake.org/download/). This tool allows you to see all of
    wolfssl's configuration variables, set them, and view their descriptions.
    Looking at the GUI or CMakeCache.txt (generated after running cmake once) is
    the best way to find out what configuration options are available and what
    they do. You can also invoke CMake from the GUI, which is described in the
    Windows instructions below. For Unix-based systems, we describe the command
    line work flow. Regardless of your chosen workflow, cmake will generate
    a header options.h in the wolfssl directory that contains the options used
    to configure the build.

    Note: Building with configure generates a wolfssl/options.h file that contains
    all the generated build options. This file needs to be included in your application
    before any other wolfSSL headers. Optionally your application can define
    WOLFSSL_USE_OPTIONS_H to do this automatically.

    Unix-based Platforms
    ---
    1) Navigate to the wolfssl root directory containing "CMakeLists.txt".
    2) Create a directory called "build" and change into it. This is where
       CMake will store build files.
    3) Run `cmake ..` to generate the target build files (e.g. UNIX Makefiles).
       To enable or disable features, set them using -D<option>=[yes/no]. For
       example, to disable TLS 1.3 support, run cmake .. -DWOLFSSL_TLS13=no
       (autoconf equivalent: ./configure --disable-tls13) To enable DSA, run
       cmake .. -DWOLFSSL_DSA=yes (autoconf equivalent: ./configure
       --enable-dsa). Again, you can find a list of these options and their
       descriptions either using the CMake GUI or by looking at CMakeCache.txt.
    5) The build directory should now contain the generated build files. Build
       with `cmake --build .`. Under the hood, this runs the target build tool
       (by default, make). You can also invoke the target build tool directly
       (e.g. make).

       To build with debugging use: `cmake .. -DCMAKE_BUILD_TYPE=Debug`.

    In the simplest form:

    # create a root directory for wolfssl repo
    git clone https://github.com/wolfSSL/wolfssl.git
    cd wolfssl


    # From the root of the wolfSSL repo:

    mkdir -p out
    pushd out
    cmake ..
    cmake --build .

    # View the available ciphers with:
    ./examples/client/client -e
    popd


    ARIA Cipher Suite.

    The ARIA cipher needs a 3rd party source binary, typically called
    `MagicCrypto.tar.gz`.

    The MagicCrypto files can be either copied to the local `wolfssl` directory,
    or an environment variable `ARIA_DIR` can be set to point to the location.

    Simply having the environment variable or local `MagicCrypto` directory
    will not automatically enable the ARIA Ciphers.

    To enable ARIA Ciphers in wolfSSL for `CMake`:

    # From the root of the wolfSSL repo:

    # set to your path
    export ARIA_DIR=~/workspace/MagicCrypto

    mkdir -p out
    pushd out
    cmake .. -DWOLFSSL_ARIA=yes
    cmake --build .

    # View the available ciphers with:
    ./examples/client/client -e
    popd


    Windows (Visual Studio)
    ---
    1) Go to this page, download the appropriate Windows installer, and install
       to get the CMake GUI: https://cmake.org/download/ Native CMake support in
       Visual Studio 16 2019 (and possibly older versions) has proven buggy. We
       recommend using the CMake GUI in concert with Visual Studio, as described
    in these steps.
    2) Open CMake.
    3) Where is the source code: <root directory of wolfssl containing
       CMakeLists.txt>
    4) Where to build the binaries: <build directory, e.g. wolfssl/build>
    5) Hit Configure. CMake runs the code in CMakeLists.txt and builds up an
       internal representation of the project.
    6) Hit Generate. CMake generates the build files. For Windows, this will
       be Visual Studio project (.vcxproj) and solution (.sln) files.
    7) Open Visual Studio and select "Open a project or solution".
    8) Navigate to the build directory and select wolfssl.sln to load the
       project.

    Windows (command line)
    ---
    1) Open Command Prompt
    2) Run the Visual Studio batch to setup command line variables, e.g. C:\Program Files (x86)\Microsoft Visual
       Studio\2017\Community\VC\Auxiliary\Build\vcvars64.bat
    3) Follow steps in "Unix-based Platforms" above.

15. Building with liboqs for TLS 1.3 [EXPERIMENTAL]
    In order be able to use liboqs, you must have it built and installed on your
    system. We support liboqs at a specific git commit.

    NOTE: Even if you have already installed liboqs, you need to follow these
          steps to install liboqs again as we support sphincs variants that are
          disabled by default in OQS's fork of OpenSSL.

    Here are instructions for obtaining and building liboqs:

    $ mkdir ~/oqs
    $ cd ~/oqs
    $ git clone --single-branch https://github.com/open-quantum-safe/liboqs.git
    $ cd liboqs/
    $ git checkout 0.8.0
    $ mkdir build
    $ cd build
    $ cmake -DOQS_USE_OPENSSL=0 ..
    $ make all
    $ sudo make install

    And then for building wolfssl, the following is sufficient:

    $ cd wolfssl
    $ ./autogen.sh (Might not be necessary)
    $ ./configure --with-liboqs
    $ make all

    Execute the following to see the liboqs-related options for KEM groups near
    the end of the output of these commands:

    $ ./examples/server/server -?
    $ ./examples/client/client -?

    For a quick start, you can run the client and server like this:

    $ ./examples/server/server -v 4 --pqc P521_KYBER_LEVEL5
    $ ./examples/client/client -v 4 --pqc P521_KYBER_LEVEL5

    Look for the following line in the output of the server and client:

    ```
    Using Post-Quantum KEM: P521_KYBER_LEVEL5
    ```

    For authentication, you can generate a certificate chain using a patch on
    top of the Open Quantum Safe project's fork of OpenSSL. We support
    certificates and keys generated by the patched version which is maintained
    in our OSP repo.

    Instructions for obtaining and building our patched version of OQS's fork of
    OpenSSL can be found at:

    https://github.com/wolfSSL/osp/tree/master/oqs/README.md

    There are scripts for generating FALCON, Dilithium and SPHINCS+ certificate
    chains which can be found in the same directory as the `README.md` file in
    the `osp` github repo. Please find instructions on how to generate the keys
    and certificates in the `README.md` file.

    Once the certificates and keys are generated, copy them from the
    to the certs directory of wolfssl. Now you can run the server and client
    like this:

    $ examples/server/server -v 4 -l TLS_AES_256_GCM_SHA384 \
      -A certs/falcon_level5_root_cert.pem \
      -c certs/falcon_level1_entity_cert.pem \
      -k certs/falcon_level1_entity_key.pem \
      --pqc P521_KYBER_LEVEL5

    $ examples/client/client -v 4 -l TLS_AES_256_GCM_SHA384 \
      -A certs/falcon_level1_root_cert.pem \
      -c certs/falcon_level5_entity_cert.pem \
      -k certs/falcon_level5_entity_key.pem \
      --pqc P521_KYBER_LEVEL5

    Congratulations! You have just achieved a fully quantum-safe TLS 1.3
    connection!

    The following NIST Competition winning algorithms are supported:
    - CRYSTALS-KYBER (KEM)
    - Dilithium (signature scheme)
    - FALCON (signature scheme)
    - SPHINCS+ (signature scheme)

    The following NIST Competition Round 3 finalist algorithms were supported,
    but have been removed after 5.3.3
    - SABER (KEM)
    - NTRU (KEM)

    Links to more information about all of these algorithms can be found here:

    https://csrc.nist.gov/projects/post-quantum-cryptography/round-3-submissions

    NOTE: The quantum-safe algorithms provided by liboqs are unstandardized and
          experimental. It is highly advised that they NOT be used in production
          environments. All OIDs and codepoints are temporary and expected to
          change in the future. You should have no expectation of backwards
          compatibility.

16. Building with vcpkg

# Building wolfssl - Using vcpkg

 You can download and install wolfssl using the [vcpkg](https://github.com/Microsoft/vcpkg):

    git clone https://github.com/Microsoft/vcpkg.git
    cd vcpkg
    ./bootstrap-vcpkg.sh
    OR for Windows
    bootstrap-vcpkg.bat

    ./vcpkg integrate install
    ./vcpkg install wolfssl

The wolfssl port in vcpkg is kept up to date by wolfSSL.

We also have vcpkg ports for wolftpm, wolfmqtt and curl.

17. Building with hash-sigs lib for LMS/HSS support [EXPERIMENTAL]

    Using LMS/HSS requires that the hash-sigs lib has been built on
    your system. We support hash-sigs lib at this git commit:
      b0631b8891295bf2929e68761205337b7c031726
    At the time of writing this, this is the HEAD of the master
    branch of the hash-sigs project.

    Currently the hash-sigs project only builds static libraries:
      - hss_verify.a: a single-threaded verify-only static lib.
      - hss_lib.a: a single-threaded static lib.
      - hss_lib_thread.a: a multi-threaded static lib.

    The multi-threaded version will mainly have speedups for key
    generation and signing.

    The default LMS build (--enable-lms) will look for
    hss_lib.a first, and hss_lib_thread.a second, in a specified
    hash-sigs dir.

    The LMS verify-only build (--enable-lms=verify-only) will look
    for hss_verify.a only, which is a slimmer library that includes
    only the minimal functions necessary for signature verification.

    How to get and build the hash-sigs library:
      $ mkdir ~/hash_sigs
      $ cd ~/hash_sigs
      $ git clone https://github.com/cisco/hash-sigs.git src
      $ cd src
      $ git checkout b0631b8891295bf2929e68761205337b7c031726

    In sha256.h, set USE_OPENSSL to 0:
      #define USE_OPENSSL 0

    To build the single-threaded version:
      $ make hss_lib.a
      $ ls *.a
      hss_lib.a

    To build multi-threaded:
      $ make hss_lib_thread.a
      $ ls *.a
      hss_lib_thread.a

    To build verify-only:
      $ make hss_verify.a
      $ ls *.a
      hss_verify.a

    Build wolfSSL with
    $ ./configure \
        --enable-static \
        --disable-shared \
        --enable-lms \
        --with-liblms=<path to dir containing hss_lib.a or hss_lib_thread.a>
    $ make

    Run the benchmark against LMS/HSS with:
    $ ./wolfcrypt/benchmark/benchmark -lms_hss

18. Building for Debian, Ubuntu, Linux Mint, and derivatives

    To generate a .deb package, configure wolfSSL with the desired
    configuration. Then run `make deb` to generate a Debian package
    with the current configuration. To build the package inside a
    Docker container, use `make deb-docker`. In both cases the
    resulting packages are placed in the root directory of the
    project.

19. Building for RHEL, Fedora, CentOS, SUSE, and openSUSE

    To generate a .rpm package, configure wolfSSL with the desired
    configuration. Then run `make rpm` to generate a .rpm package
    with the current configuration. To build the package inside a
    Docker container, use `make rpm-docker`. In both cases the
    resulting packages are placed in the root directory of the
    project.

20. Building with xmss-reference lib for XMSS/XMSS^MT support [EXPERIMENTAL]

    Experimental support for XMSS/XMSS^MT has been achieved by integration
    with the xmss-reference implementation from RFC 8391 (XMSS: eXtended
    Merkle Signature Scheme). We support a patched version of xmss-reference
    based on this git commit:
      171ccbd26f098542a67eb5d2b128281c80bd71a6
    At the time of writing this, this is the HEAD of the master branch of
    the xmss-reference project.

    How to get the xmss-reference library:
      $ mkdir ~/xmss
      $ cd ~/xmss
      $ git clone https://github.com/XMSS/xmss-reference.git src
      $ cd src
      $ git checkout 171ccbd26f098542a67eb5d2b128281c80bd71a6
      $ git apply <path to xmss reference patch>

    The patch may be found in the wolfssl-examples repo here:
      pq/stateful_hash_sig/0001-Patch-to-support-wolfSSL-xmss-reference-integration.patch

    To build patched xmss-reference:
      $ make xmss_lib.a

    To build verify-only patched xmss-reference:
      $ make xmss_verify_lib.a

    Note that this patch changes xmss-reference to use wolfCrypt SHA256 hashing,
    by registering a SHA callback function in xmss-reference. It
    thus benefits from all the same asm speedups as wolfCrypt SHA hashing.
    Depending on architecture you may build with --enable-intelasm, or
    --enable-armasm, and see 30-40% speedups in XMSS/XMSS^MT.

    For full keygen, signing, verifying, and benchmarking support, build
    wolfSSL with:
      $ ./configure \
          --enable-xmss \
          --with-libxmss=<path to xmss src dir>
      $ make

    Run the benchmark against XMSS/XMSS^MT with:
      $ ./wolfcrypt/benchmark/benchmark -xmss_xmssmt

    For a leaner xmss verify-only build, build with
      $ ./configure \
          --enable-xmss=verify-only \
          --with-libxmss=<path to xmss src dir>
      $ make