Qt for VxWorks

Contact The Qt Company for more information: https://www.qt.io/contact-us/

Qt for VxWorks Licensing

Qt for VxWorks is available under the Qt for Device Creation Professional (DCP) license. For more information, see Qt Pricing.

Supported Architectures and VxWorks Releases

Qt 6.8 is verified on VxWorks 24.03.

Supported architectures:

• ARM-v7—Tested on BD-SL-i.MX6. BSP used: fsl_imx6
• x86_64—Tested on Intel NUC6i3SYH. BSP used: itl_generic

If you are interested in later Qt releases, please contact Qt professional services.

Requirements for VxWorks

Qt Widgets Applications

• POSIX support
• C++17 support

Qt Quick 2 Applications

All features which are required for Qt Widgets applications, and in addition the following:

• GPU device (GPUDEV) for OpenGL ES 2.0

Supported Modules

Most essential Qt modules and some add-on modules are supported.

Supported Essential modules

Supported Add-Ons

Platform Notes

With the release of Qt 5.0, Qt no longer contains its own window system implementation: QWS is no longer a supported platform. For single-process use cases, the Qt Platform Abstraction is a superior solution.

There is one plugin that is usable on VxWorks: EGLFS. The availability of this plugins depends on the configuration of Qt.

Configuring for a Specific Device

Prepare your environment by installing VxWorks SDK and obtaining WindRiver license which is needed by installer and for building VxWorks images. Search for appropriate installer in {https://gallery.windriver.com/portal/products}.

Building Qt for a given device requires a Qt 6 installation for compilation host, a toolchain and a sysroot. Additionally, some devices require vendor specific adaptation code for EGL and OpenGL 2.0 support.

VxWorks Image Requirements

Qt for VxWorks requires certain VxWorks image bundles and components to be embedded in base software to make Qt compile and work properly. The list is by no means complete as it depends on hardware, software and system requirements of your project. However, the following table contains those that are mandatory for certain functionalities to work (including touchscreen support). Note also that these might change with VxWorks versions.

Example VSB configuration for the fsl_imx6 and itl_generic BSPs

You need to add the following features to VSB for Qt to work properly. For more details please check WindRiver VxWorks Documentation.

The following config variables are added implicitly and you need to remove them:

You need to add the following variables to the config:

Example VIP configuration for fsl_imx6 and itl_generic BSPs

For more details please check WindRiver VxWorks Documentation.

The following components are added implicitly and you need to remove them:

VIP Parameters

You need to add the following parameters for Qt to run properly on VxWorks:

Building Qt 6 for VxWorks

Building Qt 6 for Host

When cross-building Qt 6 for VxWorks, it's best practice to use host tools that use the same source code as cross-build does. This requires building Qt 6 for host first, but only with limited subset of submodules:

• qtbase
• qtdeclarative
• qtquick3d
• qtshadertools

Make sure to have all necessary prerequisites for building Qt 6 for host. Check details in Building Qt Sources.

Running the Configure Script

The commands to configure, build, and install Qt 6 for host are the following:

 ./configure \
     -cmake-generator "Ninja" \
     -extprefix <path/to/Qt6/host/installation/dir> \
     -submodules qtbase,qtdeclarative,qtquick3d,qtshadertools \
     -nomake tests \
     -nomake examples \
     -- \
     -B <host/build/directory>
 cd <host/build/directory>
 ninja
 ninja install

These commands install Qt 6 for the host in <path_to_Qt 6_host_installation_dir>.

Building Qt 6 for the Target

Setting Up the Build Environment

Before you run configure and build Qt 6, you must open VxWorks Development Shell in command prompt.

• Linux:

   cd <VxWorks installation directory>
   ./wrenv.sh -p vxworks
   export WIND_CC_SYSROOT=<path to VxWorks VSB directory>
  

• Windows:

   cd <VxWorks installation directory>
   wrenv -p vxworks
   export WIND_CC_SYSROOT=<path to VxWorks VSB directory>
  

Preparing the Custom CMake Toolchain File

To include graphics libraries from the BSP, you need to first create a CMake toolchain file. An example CMake toolchain file for BD-SL-i.MX6 device:

 # Copyright (C) 2024 The Qt Company Ltd.
 # SPDX-License-Identifier: BSD-3-Clause

 if(QT_VXWORKS_TOOLCHAIN_FILE)
     set(_original_toolchain_file "${QT_VXWORKS_TOOLCHAIN_FILE}")
 elseif(DEFINED ENV{QT_VXWORKS_TOOLCHAIN_FILE})
     set(_original_toolchain_file "$ENV{QT_VXWORKS_TOOLCHAIN_FILE}")
 else()
     message(FATAL_ERROR "QT_VXWORKS_TOOLCHAIN_FILE is not set.")
 endif()

 if(NOT EXISTS "${_original_toolchain_file}")
     message(FATAL_ERORR "${_original_toolchain_file} doesn't exists.")
 endif()

 include("${_original_toolchain_file}")
 unset(_original_toolchain_file)

 list(APPEND CMAKE_TRY_COMPILE_PLATFORM_VARIABLES QT_VXWORKS_TOOLCHAIN_FILE)

 set(_common_lib_path "${CMAKE_SYSROOT}/usr/lib/common")
 set(EGL_INCLUDE_DIR ${CMAKE_SYSROOT}/usr/h/public CACHE PATH "Path to EGL include directory" FORCE)
 set(EGL_LIBRARY ${_common_lib_path}/libgfxFslVivEGL.so CACHE PATH "Path to EGL lib" FORCE)
 set(GLESv2_INCLUDE_DIR ${CMAKE_SYSROOT}/usr/h/public CACHE PATH "Path to GLES include directory" FORCE)
 set(GLESv2_LIBRARY ${_common_lib_path}/libgfxFslVivGLESv2.so CACHE PATH "Path to GLES lib" FORCE)

 set(VxWorksPlatformGraphics_DEFINES "-D_FSLVIV")
 set(VxWorksPlatformGraphics_LIBRARIES_PACK
     "${_common_lib_path}/libgfxFslVivGAL.so"
     "${_common_lib_path}/libgfxFslVivGLSLC.so"
     "${_common_lib_path}/libgfxFslVivVDK.so"
     "${_common_lib_path}/libxml.so"
 )

 set(VxWorksPlatformGraphics_REQUIRED_LIBRARIES
     ${VxWorksPlatformGraphics_LIBRARIES_PACK}
     ${EGL_LIBRARY}
     ${GLESv2_LIBRARY}
 )
 unset(_common_lib_path)

An example for Intel NUC6i3SYH:

 # Copyright (C) 2024 The Qt Company Ltd.
 # SPDX-License-Identifier: BSD-3-Clause

 if(QT_VXWORKS_TOOLCHAIN_FILE)
     set(_original_toolchain_file "${QT_VXWORKS_TOOLCHAIN_FILE}")
 elseif(DEFINED ENV{QT_VXWORKS_TOOLCHAIN_FILE})
     set(_original_toolchain_file "$ENV{QT_VXWORKS_TOOLCHAIN_FILE}")
 else()
     message(FATAL_ERROR "QT_VXWORKS_TOOLCHAIN_FILE is not set.")
 endif()

 if(NOT EXISTS "${_original_toolchain_file}")
     message(FATAL_ERORR "${_original_toolchain_file} doesn't exists.")
 endif()

 include("${_original_toolchain_file}")
 unset(_original_toolchain_file)

 list(APPEND CMAKE_TRY_COMPILE_PLATFORM_VARIABLES QT_VXWORKS_TOOLCHAIN_FILE)

 set(_common_lib_path "${CMAKE_SYSROOT}/usr/lib/common")
 set(EGL_INCLUDE_DIR ${CMAKE_SYSROOT}/usr/h/public CACHE PATH "Path to EGL include directory" FORCE)
 set(EGL_LIBRARY ${_common_lib_path}/libgfxMesaEGL.so CACHE PATH "Path to EGL lib" FORCE)
 set(GLESv2_INCLUDE_DIR ${CMAKE_SYSROOT}/usr/h/public CACHE PATH "Path to GLES include directory" FORCE)
 set(GLESv2_LIBRARY ${_common_lib_path}/libgfxMesaGL.so CACHE PATH "Path to GLES lib" FORCE)

 set(VxWorksPlatformGraphics_DEFINES "-D__vxworks -D__VXWORKS")
 set(VxWorksPlatformGraphics_LIBRARIES_PACK
     "${_common_lib_path}/libgfxMesaSw.so"
     "${_common_lib_path}/libgfxMesa.so"
     "${_common_lib_path}/libgfxMesaDriI965.so"
     "${_common_lib_path}/libgfxLibDRM.so"
     "${_common_lib_path}/libxml.so"
 )

 set(VxWorksPlatformGraphics_REQUIRED_LIBRARIES
     ${VxWorksPlatformGraphics_LIBRARIES_PACK}
     ${EGL_LIBRARY}
     ${GLESv2_LIBRARY}
 )
 unset(_common_lib_path)

Running Qt 6 Configure for the Target

Below is an example build configuration for the BD-SL-i.MX6. For most VxWorks boards the configure command looks very similar. By default, Qt 6 is configured to use shared libraries. To build Qt 6 statically, add -static option to configure.

Make sure that {WIND_CC_SYSROOT} environment variable is set to VxWorks VSB root directory.

 ./configure \
     -cmake-generator "Ninja" \
     -icu \
     -eglfs \
     -qpa eglfs \
     -submodules "qtbase,qtdeclarative,qtmultimedia,qtquick3d,qtgraphs,qtimageformats,qtsvg,qtshadertools,qtvirtualkeyboard" \
     -- \
     -B <target/build/directory> \
     -DQT_VXWORKS_TOOLCHAIN_FILE=<path/to/BDP/toolchain.cmake> \
     -DFEATURE_vulkan=OFF \
     -DQT_QMAKE_TARGET_MKSPEC=vxworks-clang \
     -DQT_HOST_PATH=<path/to/Qt6/host/installation/dir> \
     -DCMAKE_STAGING_PREFIX:PATH=<path/to/Qt6/target/installation/dir> \
     -DCMAKE_TOOLCHAIN_FILE=<path/to/cmake/for/graphics/file/created/above> \

In case of building for DKM rather than RTP, use -static option and change the CMAKE_TOOLCHAIN_FILE value to "$WIND_CC_SYSROOT/mk/dkm.toolchain.cmake"

It is recommended to build Qt 6 using a shadow build. See Qt Configure Options for more information.

After configuration, build and install Qt 6 for VxWorks with these commands:

 cd <host build directory>
 ninja
 ninja install

Platform Plugins for VxWorks Devices

Qt for VxWorks supports EGLFS platform plugin for a native window substitution. Read more about its configuration in EGLFS.

Building Qt Applications

To build a Qt application, you must set the QT_VXWORKS_TOOLCHAIN_FILE as an environment variable or as a define.

An example to build the clocks Qt demo:

 export WIND_CC_SYSROOT=<path/to/VxWorks/VSB/directory>
 export QT_VXWORKS_TOOLCHAIN_FILE=<path/to/the/BSP/toolchain.cmake>
 <path/to/Qt6/target/installation/dir>/bin/qt-cmake -G Ninja -S ~/Qt/Examples/Qt-6.8.0/demos/clocks -B <path/to/build/directory>
 cd <path/to/build/directory>
 ninja

Running Qt Applications

The following example shows how to start an application when Qt 6 is built using shared libraries. With a statically built Qt 6, you don't need to use the LD_LIBRARY_PATH environment variable for the Qt libraries, but it needs to point to the location of VxWorks shared libraries (for example libc and OpenGL ES 2.0). It is not needed for Qt 6 static libraries.

 putenv "LD_LIBRARY_PATH=/sd0:1/lib"
 cd "/sd0:1"
 rtpSp("<Qt 6_app>", 200, 0x100000, 0, 0x01000000)

Limitations

Video Memory

Systems with a fixed amount of dedicated video memory may need extra care before running Qt application based on Qt Quick or classes like QOpenGLWidget. The default setting might be insufficient for such applications, especially when they are displayed on a high resolution (for example, full HD) screen. In this case they might start failing in unexpected ways. It is therefore recommended to ensure that there is at least 128 MB of GPU memory available. For systems that do not have a fixed amount of memory reserved for the GPU this is not an issue.