Media limitations of test devices

Primary media and primary bootloaders

Methods like deploy to TFTP rely on LAVA interrupting the primary bootloader on the DUT and changing the boot process to use the files provided by the test writer. Many test devices require this bootloader to be installed onto re-writeable media which can be modified from the running system, for example an SD card.

The critical element for LAVA is the first point where the boot can be interrupted or modified. Devices which lack some kind of BMC rely on this bootloader to be able to automatically recover from a broken deployment. This bootloader can be considered as the primary bootloader and the medium where this is installed can be considered as the primary medium which must be protected from deployments which would replace its entire contents. For example, a panda board has an SD card and USB host support, the primary bootloader (U-Boot) must be on the SD card, so V2 uses the SD card as primary media. It is therefore not supportable for a panda board to deploy to the SD card in LAVA.


Primary and secondary media relate to devices where the primary bootloader is installed on writeable media and where that bootloader needs to operate before new software can be deployed. Fastboot devices often differ and allow all media on the device to be modified in test jobs as long as jumpers or dip switches are set to force the device to boot into fastboot mode. Modifying the fastboot support on the device may involve changing those jumpers and/or dip switches, i.e. an admin task whilst the device is offline.

It is important to consider the constraints of primary media as a limitation of the hardware for automation. The risks of allowing test writers to easily brick devices outweigh the usefulness of the primary media for files other than the primary bootloader. LAVA has tried to use partitions on the primary media in the past and this has proven to be unreliable.

Devices which only provide primary media can still support deployment methods like TFTP to use ramdisk and NFS test jobs.


Many devices allow test writers to access the primary media from within a test shell even if the test job has deployed into a ramdisk, NFS or secondary media. It remains the responsibility of the test writer to not modify the primary media from within a test shell just because you can. Test writers who do so may have submission privileges revoked by the admins.

Secondary media

If the device supports any media other than primary media, all these media are collectively termed secondary media. Any deployment to the secondary media can write a complete image including partition tables and complete filesystems without affecting the ability of LAVA to recover from a broken test job.

For example, a beaglebone-black can have a USB stick as secondary media. A cubietruck can have a single SATA drive. A mustang device can have several SATA drives attached. Usually, USB is the least useful secondary media as it is so slow to write and the test job will be trying to write a full size filesystem image of many gigabytes.

Bootloader limitations

The bootloader may be unable to write to permanent media. Usually this is a good thing, bootloaders which can write to media can be problematic, slow and/or greedy of resources on the worker e.g. fastboot, necessitating the use of LXC. U-Boot and Grub are each restricted to reading files from media and not writing new files during the boot process. To deploy a test image to the secondary media, the device needs to first boot into a full OS environment with threading, multiple cores, network and full kernel support. A standard test shell can then write to the media in various ways. Depending on the device configuration, test writers may still need to respect the location of the primary bootloader to avoid bricking the device.

LAVA can use secondary media in a number of different ways, depending on the use case.

Occassional debugging

Combined with a hacking session, the boot commands can be overridden to test support on selected devices. The primary bootloader will still be used, for example to read files from the filesystem. Optionally, the kernel can be loaded over TFTP to use the filesystem on the secondary media as a form of persistence.


Unless the device is restricted to particular submitters and any health check has been disabled, the next test job could replace, corrupt or update the root filesystem on the secondary media. The full process would have to be reset.

Installer testing

An operating system installer is the traditional form of secondary media deployments. The installer boots into a ramdisk to allow complete access to the device, including all media. The installer can be used to do the deployment to the secondary media in LAVA, although support will be required to automate the questions and prompts normally raised during the process.

This method has the advantage that the final system is a fresh, clean install and the disadvantage that the whole system has to be recreated each test job, as well as the overhead of starting and running the installer.


  • Installer may try to write new UEFI or new UBoot - the automation of the installer will need to prevent overwriting the primary bootloader. (If your test job bricks the device, the admin could revoke or suspend your submission rights.)
  • Not advised for most UBoot devices - many installer programs need special support to install onto UBoot devices and it could be hard to both update the kernel in the installed system and prevent modification of the primary bootloader.
  • Chainload installed Grub from the primary Grub bootloader - Writing a second bootloader to the secondary media can work, as long as the second bootloader can be chainloaded from the primary bootloader by issuing commands to the primary bootloader.
  • Wait for the installer to run
    • Large SATA drives can take long time to partition.
    • Downloads from mirrors may take time to install

Secondary media deployment of images

Secondary media deployments are a way of automating the deployment of a filesystem image directly to the secondary media. The image will need to contain the partition(s) and filesystem(s) for the test system.

Unlike the installer support, secondary media deployments can work with UBoot devices although many ARMv7 devices are limited by slow media like USB drives instead of SATA.


  • Make sure all tools are installed - The test job will download and apply the image after completing a test shell, ensure wget is installed.
  • New image may include new UEFI or new UBoot - The image to be deployed will need to avoid overwriting the primary bootloader. (If your test job bricks the device, the admin could revoke or suspend your submission rights.)
  • Production images can be a risk - LAVA still needs to interrupt the primary bootloader and add files to the deployed image to be able to run test shell definitions. Production images often include security settings which will disable this access, causing your tests to fail.
  • Single write operation - LAVA downloads the image and then simply writes the data to the media before rebooting. The image must be fully configured to work in this way, including raising usable network interfaces directly upon boot.

Principles and Requirements

Secondary deployments are done by the device under test, using actions defined by LAVA and tools provided by the initial deployment. Test writers need to ensure that the initial deployment has enough support to complete the second deployment.

Images on remote servers are downloaded to the dispatcher (and decompressed where relevant) so that the device does not need to do the decompression or need lots of storage in the initial deployment.

By keeping the downloaded image intact, it becomes possible to put the LAVA extensions alongside the image instead of inside.

To make this work, several requirements must be met:

  • The initial deployment must provide or support installation of all tools necessary to complete the second deployment - it is a TestError if there is insufficient space or the deployment cannot complete this step.
  • The initial deployment does not need enough space for the decompressed image, however, the initial deployment is responsible for writing the decompressed image to the secondary media from stdin, so the amount of memory taken up by the initial deployment can have an impact on the speed or success of the write.
  • The operation of the second deployment is an action which precedes the second boot. There is no provision for getting data back from this test shell into the boot arguments for the next boot. Any data which is genuinely persistent needs to be specified in advance.
  • LAVA manages the path to which the second deployment is written, based on the media supported by the device and the ID of that media. Where a device supports multiple options for secondary media, the job specifies which media is to be used.
  • LAVA will need to support instructions in the job definition which determine whether a failed test shell should allow or skip the boot action following.
  • LAVA will declare available media using the kernel interface as the label. A SATA drive which can only be attached to devices of a particular device type using USB is still a USB device as it is constrained by the USB interface being present in the test image kernel. A SATA drive attached to a SATA connector on the board is a SATA device in LAVA (irrespective of how the board actually delivers the SATA interface on that connector).
  • If a device has multiple media of the same type, it is up to the test writer to determine how to ensure that the correct image is booted. The blkid of a partition within an image is a permanent UUID within that image and needs to be determined in advance if this is to be used in arguments to the bootloader as the root filesystem.
  • The manufacturer ID and serial number of the hardware to be used for the secondary deployment must be set in the device configuration. This makes it possible for test images to use such support as is available (e.g. udev) to boot the correct device.
  • The job definition needs to specify which hardware to use for the second deployment - if this label is based on a device node, it is a TestError if the use of this label does not result in a successful boot.
  • The job definition also needs to specify the path to the kernel, dtb and the partition containing the rootfs within the deployed image.
  • The job definition needs to include the bootloader commands, although defaults can be provided in some cases.

Test Writer steps

  • always ensure you have set a usable root password in the image / test media or set the root user to not have a password.
    • If a password is set for the root user, the password must be declared in the test job submission.
  • always ensure you have set the bootable flag on the boot partition when building the image.
  • always ensure you have installed a kernel into the image
    • note down the paths to the kernel and initramfs etc. These will need to be specified in the test job submission.
  • always ensure you have the UUID of the new filesystem containing the root filesystem. This will need to be specified in the test job submission.
  • ensure that if a bootloader is present in the image to be deployed that this bootloader can be chainloaded by the primary bootloader already on the device.


Deploy commands

This is an example block - the actual data values here are known not to work as the deploy step is for a panda but the boot step in the next example comes from a working cubietruck job.

This example uses a device configuration where UUID-required is True.

For simplicity, this example also omits the initial deployment and boot, at the start of this block, the device is already running a kernel with a ramdisk or rootfs which provides enough support to complete this second deployment.

# secondary media - use the first deploy to get to a system which can deploy the next
# in testing, assumed to already be deployed
- deploy:
      minutes: 10
    to: usb
    os: debian
    # not a real job, just used for unit tests
    compression: gz
    device: SanDisk_Ultra # needs to be exposed in the device-specific UI
    download: /usr/bin/wget
  1. Ensure that the deploy action has sufficient time to download the decompressed image and write that image directly to the media using STDOUT. In the example, the deploy timeout has been set to ten minutes - in a test on the panda, the actual time required to write the specified image to a USB device was around 6 minutes.
  2. Note the deployment strategy - to: usb. This is a direct mapping to the kernel interface used to deploy and boot this image. The bootloader must also support reading files over this interface.
  3. The compression method used by the specified image is explicitly set.
  4. The image is downloaded and decompressed by the dispatcher, then made available to the device to retrieve and write to the specified media.
  5. The device is specified as a label so that the correct UUID can be constructed from the device configuration data.
  6. The download tool is specified as a full path which must exist inside the currently deployed system. This tool will be used to retrieve the decompressed image from the dispatcher and pass STDOUT to the writer tool, dd by default. If the download tool is the default /usr/bin/wget, LAVA will add the following options: --no-check-certificate --no-proxy --connect-timeout=30 -S --progress=dot:giga -O - If different download tools are required for particular images, these can be specified, however, if those tools require options, the test writer can either ensure that a script exists in the image which wraps those options or file a bug to have the alternative tool options supported.

The default writer tool is dd but it is possible to specify an alternative one. In particular, bmaptool is usually a much better choice for USB or SD card devices. It will typically flash the image faster and extend the lifetime of the storage media. It needs a .bmap file which contains a block map alongside the actual image file. For this reason, two files need to be downloaded and stored in the same directory on the dispatcher. The example below illustrates how to do this:

# secondary media deployment using bmaptool
- deploy:
      minutes: 10
    to: usb
    os: debian
    # not a real job, just used for illustrative purposes
    compression: gz
    uniquify: false
    device: SanDisk_Ultra # needs to be exposed in the device-specific UI
      tool: /usr/bin/bmaptool
      options: copy {DOWNLOAD_URL} {DEVICE}
      prompt: 'bmaptool: info'
      prompts: ['copying time: [0-9ms\.\ ]+, copying speed [0-9\.]+ MiB\/sec']
  1. The images list needs to contain one image entry and can have others as well such as bmap in this case. They will all be downloaded separately to the dispatcher and made available to the board via HTTP. The URL of the image file is available in the job definition as {DOWNLOAD_URL}. The URL of the other images will need to be determined by other means. Say, the image file could be a manifest with the list of the actual binary images (not the case with this bmaptool example).
  2. Each item in the images list is normally downloaded into a separate sub-directory such as image or bmap in this example. As the bmaptool expects both files to be in the same path, the uniquify: false option is used so all the files are downloaded directly at the root of the job’s storage-deploy-* directory. Please note that if several image files have the same name, they will overwrite each other when uniquify is set to false. For this reason, if not specified in the job it will be set to true by default.
  3. To use an alternative writer tool, the writer parameters are used. The absolute path to the tool must be provided with tool as well as the options required to call it. The prompt is used to detect that the flashing has started.
  4. The writer tool will normally also be responsible for downloading the image file, hence the {DOWNLOAD_URL} option passed to it in the example. It is also possible to provide both download and writer parameters, in which case the standard output of the downloader tool will be piped into the standard input of the writer tool.
  5. The tool prompts parameter is to detect when the writer tool has completed the flashing operation. When LAVA has matched a prompt with the tool output, it will then proceed with the secondary boot action. The prompts parameters defaults are to match the output of dd, so they should be defined appropriately when using an alternative writer tool.

The kernel inside the initial deployment MUST support UUID when deployed on a device where UUID is required, as it is this kernel which needs to make /dev/disk/by-id/$path exist for dd to use. Remember not to quote the UUID:

root_uuid: UUID=159d17cc-697c-4125-95a0-a3775e1deabe

Boot commands

- boot:
    method: u-boot
    commands: usb
      shutdown-message: "reboot: Restarting system"
    # these files are part of the image already deployed and are known to the test writer
    kernel: /boot/vmlinuz-3.16.0-4-armmp-lpae
    ramdisk: /boot/initrd.img-3.16.0-4-armmp-lpae.u-boot
    dtb: /boot/dtb-3.16.0-4-armmp-lpae'
    root_uuid: UUID=159d17cc-697c-4125-95a0-a3775e1deabe  # comes from the supplied image.
    boot_part: 1  # the partition on the media from which the bootloader can read the kernel, ramdisk & dtb
      - 'linaro-test'
      - 'root@debian:~#'

The kernel and (if specified) the ramdisk and dtb paths are the paths used by the bootloader to load the files in order to boot the image deployed onto the secondary media. These are not necessarily the same as the paths to the same files as they would appear inside the image after booting, depending on whether any boot partition is mounted at a particular mountpoint.

The root_uuid is the full option for the root= command to the kernel, including the UUID= prefix.

The boot_part is the number of the partition from which the bootloader can read the files to boot the image. This will be combined with the device configuration interface name and device_id to create the command to the bootloader, e.g.:

"setenv loadfdt 'load usb 0:1 ${fdt_addr_r} /boot/dtb-3.16.0-4-armmp-lpae''",

The dispatcher does NOT analyze the incoming image - internal UUIDs inside an image do not change as the refactored dispatcher does not break up or reorganise the partitions. Therefore, the UUIDs of partitions inside the image MUST be declared by the job submissions.