Building the host

Building with a hypervisor buildfile variant

If the BSP you're using to build your hypervisor system provides a hypervisor buildfile variant, follow these steps to build the host:

1. Unzip the archive that contains the BSP.
2. Locate the hypervisor buildfile variant within your BSP. This buildfile is in the same directory as the generic buildfile but its name contains -hypervisor; for example, /images/mek/imx8qm-cpu-mek-hypervisor.build.
3. If you want to use the shmem-host shared memory demonstration application, you need to build it. For more information, go to Adding the shmem-host demonstration application.
4. Use the following command to make the BSP's prebuilt binaries available to the BSP's build system:

   make prebuilt

5. Navigate to the /images directory (e.g., cd images/), and then use the following command to build the host IFS:

   make hyp

   Note:

   If you run simply make, the build utility uses the default buildfile that comes with the BSP and, thus, won't build a hypervisor host image. The command make hyp won't work if a QNX hypervisor product is not installed.
6. Use the following command to build the host disk image:

   make disk_image

Building without a hypervisor buildfile variant

If the BSP you're using does not provide a hypervisor buildfile variant, follow these steps to build the host:

1. Unzip the archive that contains the BSP.
2. If you want to include the shmem-host shared memory demonstration application, you need to build it using source code you obtain from a BSP provided with a QNX hypervisor product. For more information, go to Adding the shmem-host demonstration application.
3. Use the following command to make the BSP's prebuilt binaries available to the BSP's build system:

   make prebuilt

4. Navigate to the /images directory (e.g., cd images/).
5. Modify the buildfile to include the files you need for your hypervisor system at the appropriate target locations. These include:
   • The QNX virtual machine binary, qvm, and, optionally for x86, the qvm-check utility. For example:

     /sbin/qvm = qvm
     /bin/qvm-check = qvm-check

   • The configuration files (*.qvmconf) for each VM that you will run in your hypervisor system:

     infotainment_ubuntu.qvmconf

   • The shared object file (vdev-*.so) for every vdev you want to make available for your guests. For example:

     vdev-8259.so
     vdev-hpet.so
     vdev-ioapic.so
     ...
     vdev-virtio-console.so
     vdev-virtio-net.so

   • The smmuman service, architecture, and board support libraries, and its configuration files. For example, for a QNX hypervisor system on a supported x86 board:

     /bin/smmuman = smmuman
     /bin/smmu-vtd.so = smmu-vtd.so
     /etc/smmuman/vtd.smmu = ./smmuman-config/vtd.smmu

6. Modify the buildfile to include -Q enable in the startup driver arguments, inside the virtual attribute that specifies arguments to the startup driver and kernel.
7. Run make to build the host IFS using the host buildfile.
8. If it's applicable in your environment, build a bootable disk image:
   • Use the make disk_image command, if your BSP supports it.
   • For other BSPs, you can create a system disk image using the diskimage utility (go to the diskimage entry in the QNX OS Utilities Reference).
   • For an example of host disk image creation, in a QNX-supplied BSP that supports the QNX hypervisor product that you're using, examine the Makefile located at images/board_variant/Makefile. The entry for the disk_image target illustrates which tools and scripts are called, and which configuration and buildfiles are used (some of which are found in the images/ folder).
   • For more information, go to Hypervisor disk images and Transferring the disk image.

Adding the shmem-host demonstration application

The shmem-host application expects to communicate with a peer via the shared memory (shmem) vdev. In most cases, this peer is either shmem-guest or shmem-linux, running inside of a guest equipped with the appropriate vdev.