TockOS Support Tool


Keywords
pip, python-3, tock
License
MIT
Install
pip install tockloader==1.0.0

Documentation

TockLoader

Tool for programming Tock onto hardware boards.

Install

pip3 install tockloader --user

If you want tab completions:

register-python-argcomplete tockloader >> ~/.bashrc

Usage

This tool installs a binary called tockloader, which supports several commands.

Primary Commands

These are the main commands for managing apps on a board.

tockloader install

Load Tock applications on to the board. Use --no-replace to install multiple copies of the same app.

tockloader update

Update an application that is already flashed to the board with a new binary.

tockloader uninstall [application name(s)]

Remove an application from flash by its name.

Board Inspection Commands

These query the board for its current state.

tockloader list

Print information about the apps currently loaded onto the board.

tockloader info

Show all properties of the board.

Utility Commands

These provide other helpful features.

tockloader listen

Listen to UART printf() data from a board. Use the option --rtt to use Segger's RTT listener instead of using a serial port.

Other Commands

These provide more internal functionality.

tockloader flash

Load binaries onto hardware platforms that are running a compatible bootloader. This is used by the TockOS Make system when kernel binaries are programmed to the board with make program.

tockloader inspect-tab

Show details about a compiled TAB file.

tockloader enable-app [application name(s)]

Enable an app so that the kernel will run it at boot.

tockloader disable-app [application name(s)]

Disable an app so that the kernel will not start it at boot.

tockloader sticky-app [application name(s)]

Mark an app as sticky so that the --force flag is required to uninstall it.

tockloader unsticky-app [application name(s)]

Remove the sticky flag from an app.

tockloader list-attributes

Show all of the attributes that are stored on the board.

tockloader set-attribute [attribute key] [attribute value]

Set a particular attribute key to the specified value. This will overwrite an existing attribute if the key matches.

tockloader remove-attribute [attribute key]

Remove a particular attribute from the board.

tockloader dump-flash-page [page number]

Show the contents of a page of flash.

tockloader read [address] [# bytes]

Read arbitrary flash memory from the board.

tockloader write [address] [# bytes] [value]

Write arbitrary flash memory on the board with a specific value.

tockloader list-known-boards

Print which boards tockloader has default settings for built-in.

tockloader set-start-address [address]

Set the jump address the bootloader uses for the location of the kernel.

tockloader tbf tlv add|modify|delete [TLVNAME]

Interact with TLV structures within a TBF.

tockloader tbf credential add|delete [credential type]

Add and remove credentials in the TBF footer.

tockloader tickv get|append|invalidate|dump|cleanup|reset [key] [value]

Interact with a TicKV key-value database.

Specifying the Board

For tockloader to know how to interface with a particular hardware board, it tries several options:

  1. Read the parameters from the bootloader. Tockloader assumes it can open a serial connection to a tock-bootloader on the board.

  2. Use JLinkExe and OpenOCD to discover known boards.

  3. Use the --board command line flag and a list of known boards.

  4. Use individual command line flags that specify how to interact with the board.

If command line flags are passed they take priority over any automatically discovered options.

Tockloader has hardcoded parameters for a variety of boards. You can list these with:

tockloader list-known-boards

To use a known board, if it is not automatically discovered, you can:

tockloader [command] --board [board]

If your board is not a known board, you can specify the required parameters via command line options. Note, you also need to provide a name for the board.

tockloader [command] --board [board] --arch [arch] --page-size [page_size]
  • board: The name of the board. This helps prevent incompatible applications from being flashed on the wrong board.
  • arch: The architecture of the board. Likely cortex-m0 or cortex-m4.
  • page_size: The size in bytes of the smallest erasable unit in flash.

Specifying the Communication Channel

Tockloader defaults to using a serial connection to an on-chip bootloader to program and interact with a board. If you need to use a different communication mechanism, you can specify what Tockloader should use with command line arguments. Note, Tockloader's board autodiscovery process also selects different communication channels based on which board it finds.

To use a JTAG interface using JLinkExe, specify --jlink. JLinkExe requires knowing the device type of the MCU on the board.

tockloader [command] --board [board] --arch [arch] --page-size [page_size] \
                     --jlink --jlink-cmd [jlink_cmd] --jlink-device [device] \
                     --jlink-speed [speed] --jlink-if [if]
  • jlink_cmd: The JLink executable to invoke. Defaults to JLinkExe on Mac/Linux, and JLink on Windows.
  • device: The JLinkExe device identifier.
  • speed: The speed value to pass to JLink. Defaults to 1200.
  • if: The interface to pass to JLink.

Tockloader can also do JTAG using OpenOCD. OpenOCD needs to know which config file to use.

tockloader [command] --board [board] --arch [arch] --page-size [page_size] \
                     --openocd --openocd-board [openocd_board] \
                     --openocd-cmd [openocd_cmd] \
                     --openocd-options [openocd_options] \
                     --openocd-commands [openocd_commands]
  • openocd_board: The .cfg file in the board folder in OpenOCD to use.

  • openocd_cmd: The OpenOCD executable to invoke. Defaults to openocd.

  • openocd_options: A list of Tock-specific flags used to customize how Tockloader calls OpenOCD based on experience with various boards and their quirks. Options include:

    • noreset: Removes the command reset init; from OpenOCD commands.
    • nocmdprefix: Removes the commands init; reset init; halt; from OpenOCD commands.
    • workareazero: Adds the command set WORKAREASIZE 0; to OpenOCD commands.
    • resume: Adds the commands soft_reset_halt; resume; to OpenOCD commands.
  • openocd_commands: This sets a custom OpenOCD command string to allow Tockloader to program arbitrary chips with OpenOCD before support for the board is officially include in Tockloader. The following main operations can be customized:

    • program: Operation used to write a binary to the chip.
    • read: Operation used to read arbitrary flash memory on the chip.
    • erase: Operation that erases arbitrary ranges of flash memory on the chip.

    The custom values are specified as key=value pairs, for example, --openocd_commands 'program=write_image; halt;' 'erase=flash fillb {address:#x} 0xff 512;'. Operation strings can include wildcards which will get set with the correct value by Tockloader:

    • {{binary}}: The binary file path.
    • {address:#x}: The specified address for the binary to be programmed at.
    • {length}: The number of bytes. Only valid for the read operation.

Finally, Tockloader can treat a local file as though it were the flash contents of a board. The file can then be loaded separately onto a board.

tockloader [command] --flash-file [filepath]
  • filepath: The file to use as the flash contents. Will be created if it doesn't exist.

Example Usage

Install an app, make sure it's up to date, and make sure it's the only app on the board:

tockloader install --make --erase

Get all info from the board that can be used to help debugging:

tockloader info

Print additionally debugging information. This can be helpful when using JTAG.

tockloader install --debug

Get printf() data from a board:

tockloader listen

Additional Flags

There are additional flags that might be useful for customizing tockloader's operation based on the requirements of a particular hardware platform.

  • --app-address: Manually specify the address at the beginning of where apps are stored. This can be in hex or decimal.
  • --bundle-apps: This forces tockloader to write all apps as a concatenated bundle using only a single flash command. This will require that anytime any app changes in any way (e.g. its header changes or the app is updated or a new app is installed) all apps are re-written.

Credentials and Integrity Support

Tockloader supports working with credentials stored in the TBF footer. Tockloader will attempt to verify that stored credentials are valid for the given TBF. For credentials that require keys to verify, Tockloader can check the credential using:

$ tockloader inspect-tab --verify-credentials [list of key files]
example:
$ tockloader inspect-tab --verify-credentials tockkey.public.der

Tockloader can also add credentials. To add a hash:

$ tockloader tbf credential add sha256

To add an RSA signature:

$ tockloader tbf credential add rsa2048 --private-key tockkey2048.private.der --public-key tockkey2048.public.der

To remove credentials:

$ tockloader tbf credential delete sha256

Features

  • Supported communication protocols
    • Serial over USB
    • Segger JLinkExe JTAG support
    • OpenOCD JTAG support
  • JLink RTT listener
  • JSON output using --output-format json for certain commands.

Complete Install Instructions

Tockloader is a Python script that is installed as an executable. To use Tockloader, you need python3, a couple dependencies, and the Tockloader package.

  • Ubuntu

    sudo apt install python3-pip
    pip3 install -U pip --user     # update pip
    pip3 install tockloader --user
    
  • MacOS

    brew install python3
    pip3 install tockloader
    
  • Windows

Internal Notes

Test Locally

To test the code locally without installing as a package, from the top-level directory:

python3 -m tockloader.main <COMMANDS>

Upload to PyPI

python3 setup.py sdist bdist_wheel
twine upload dist/*

Build Docs

pip3 install mkdocs
cd docs
./generate_docs.py
cd ..
mkdocs serve --dev-addr=0.0.0.0:8001