The main tool in flashcontainer is called Pargen. Pargen is generating data files for parameter values or other data structures that can to be stored in flash memory. It allows to change the data without recompilation by reading definitions from an XML file.

Flashcontainter also includes tailored front ends for Pargen to simplify special use cases. See the TC3xx tool description chapter for generating Infineon Aurix TC3xx boot mode headers structures.

• Read input from a schema validated XML file
• Generate C-Source stubs for embedding into the application source amd testing environments
• Generate Intel Hex files for flashing with a programmer
• Generate GNU linker include file for mapping the parameter to absolute addresses

The package is available on Pypi and can be installed using

pip install flashcontainer

A python 3.8 or higher version is required.

The parameter generator tool can then by called on cmdline using

$ pargen -h
usage: pargen [-h] [--ihex] [--csrc] [--gld] [--a2l] [--pyhexdump] [--destdir directory] [--filename basename] [--static] [--modify name=value] [--version] [-v] file

pargen 0.4.0: A tool for generating flashable parameter container.

positional arguments:
  file                  name of the XML parameter definition file

optional arguments:
  -h, --help            show this help message and exit
  --ihex                generate intelhex file
  --csrc                generate c/c++ header and source files
  --gld                 generate GNU linker include file for parameter symbol generation
  --a2l                 generate A2L parameter description file
  --pyhexdump           generate pyHexDump print configuration file
  --destdir directory, -o directory
                        specify output directory for generated files
  --filename basename, -f basename
                        set basename for generated files
  --static, -s          create static comment output without dynamic elements like date and time
  --modify name=value, -m name=value
                        modify parameter value using name=value notation
  --version             show program's version number and exit
  -v, --verbose         increase logging level

Copyright (c) 2022-2023 Haju Schulz <[email protected]>. Visit https://github.com/nhjschulz/flashcontainer for full documentation and examples.

The flashcontainer Developing page on Github explains how to use unreleased development builds or how to setup a development environment.

The flash container configuration for Pargen is a XML definition file with the format explained below.

The examples folder shows how to configure Pargen for various use cases. Most of it is likely self explanatory. Read the examples.md files inside the examples folder to learn more about them. To understand Pargen's XML capabilities in depth, read on.

The XML follows an XSD-schema defined in pargen_1.1.xsd. It is highly recommended to use an XML editor with schema validation support to avoid or detect validations already while editing. Visual Studio Code is a perfect choice, given the "XML Language Support" extension from Red Hat is installed. This extensions brings validation and "IntelliSense" to editing XML files.

The file defines the following data element hierarchy. The "..." lines mean that the preceding element may appear multiple times:

  <pd:struct>
      <pd:fields>
          <pd:field> or <pd:arrayfield> or <pd:crcfield> or <pd:padding>
          ...
      </pd:fields>
  </pd:struct>
  ...
  <pd:container>
      <pd:blocks>
          <pd:block>
              <pd:parameter> or <pd:crc>
              ...
          </pd:block>
          ...
      </pd:blocks>
      ...
  </pd:container>

The XML file uses XSD schema validation and a namespace. This requires the following (static) XML element to be used as the root XML element at the beginning of the file:

?xml version="1.0" encoding="utf-8"?>
<pd:pargen xmlns:pd="http://nhjschulz.github.io/1.1/pargen"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://nhjschulz.github.io/1.1/pargen http://nhjschulz.github.io/xsd/pargen_1.1.xsd" >

The top level configuration element is the container. A container maps its parameter blocks to absolute addresses by the at attribute. Address attributes inside block elements are offsets to the at attribute value.

<pd:container name="UCBRange" at="0xAF400000">
  <pd:blocks>
  ....
  </pd:blocks>
</pd:container>

A block element defines a contiguous memory area inside a container. Blocks contain an optional header and parameters at unique offsets inside the block memory range. Gaps between parameters are filled with the byte value specified using the 'fill' attribute. A container may have 1 to many block children inside a blocks element.

Pargen blocks may contain a header at the beginning of the block memory area. This optional header contains block identification, version and length information. The header supports parameter validation to verify correctness and compatibility with the using application during runtime. The safety example in the examples folder shows how to use the header in combination with a CRC for this purpose.

The header data is a 16 byte long data structure with the following layout:

struct sruct_pargen_header_type
{
    uint16_t id;        /* from id header attribute */
    uint16_t major;     /* from major header attribute */
    uint16_t minor;     /* from minor header attribute */
    uint16_t dataver;   /* from version header attribute */
    uint32_t reserved;  /* reserved = 0x00000000 */
    uint32_t length;    /* from length block attribute */
};

Application can freely decide how to use the id, major, minor and dataver header fields. The proposed usage is as follows:

• id - A unique id to identify the blocks purpose. (Example: 1:CAN Bus settings, 2 :Motor parameters, ...)
• Major.Minor - A version pair defining the layout of the block
• dataver - The version or build number of the parameter set

The header data is internally handled as a parameter. Space for any further parameter starts at offset 16 (0x10) if the header is used.

<pd:block offset="0x0000" name="UCB_BMHD0" length="0x1F4" fill="0x00" endianness="LE">
    <pd:header id="0x0A" major="1" minor="0" version="1"></pd:header>
    <pd:comment>Aurix Bootmode Headers</pd:comment>
    <pd:data>
      ....
    </pd:data>
</pd:block>

A parameter element defines a single parameter inside a block. Blocks may have one to many parameter elements.

<pd:param offset="0x004" name="STAD" type="uint32">
  <pd:comment>Application entry point address</pd:comment>
  <pd:value>0x80028000</pd:value>

The value element of a parameter holds the parameter value inside its text element using a JSON style syntax. The following subset of JSON definitions are supported:

For complex parameters (that have the type of a defined struct), JSON input is parsed as key: value pairs to populate the structs fields. Input is required for every Field or Array Field in the struct, Padding and CRCs need no input in the value element. This means the set of keys in the input JSON has to match the set of Field names and Array Field names of the struct, or pargen will detect the input as invalid.

The crc element defines an integer parameter. The difference to a normal integer parameter is the automatic value calculation using a crc algorithm. Instead of a parameter value child element, memory and config elements are used to define crc calculation parameters.

The memory element defines the memory range used to calculate the crc and the access method to this memory range if byte swapping is needed. The bytes at the range boundaries are included into the crc calculation.

The config element defines the crc calculation parameters to enable arbitrary crc methods. The values for common used crc methods can be taken from this crc catalog page. The default values select the IEEE802.3 crc calculation also known as CRC-32. Note that the bit size of the crc is not part of these parameters, but derived from the type attribute of the crc element.

<pd:crc offset="0x008" name="CRCBMHD" type="uint32">
  <pd:memory from="0x0000" to="0x0007" access="32" swap="true"/>
  <pd:config polynomial="0x04C11DB7" init="0xFFFFFFFF" rev_in="true" rev_out="true" final_xor="true" ></pd:config>
</pd:crc>

A struct is a way to group multiple simple types into a combined data structure. By referencing a structs name in a block parameter, the parameters type will be interpreted as the struct. A struct features multiple attributes that define how the struct and the data of the encompassed fields.

The fill value will be used as padding, with the additional options of "parent". Choosing parent will use the fill value of a parameters encompassing block.

A field is the basic building block of a struct and serves as a placeholder for a value of a basic type in instances of the struct. Unlike parameter elements in a block, a field does not hold offset or alignment information, instead alignment is handled by padding tags. The basic type of the field defines its size and content interpretation.

The CRC Field Element functions similarly to a CRC Element, except it does not feature offset and alignment attributes.

Array Fields are similar to fields, but feature a size attribute that determines the resulting arrays size

The padding element is used to introduce gaps / offsets between a structs fields. It features a size attribute that determines the padding size.

The tc3xx tool is used to generate Pargen definition XML files for boot mode header structures. A Boot mode header structure is a data block in flash memory, which is defined by Infineon TC3xx processors. It gets processed by startup software to select which image to boot into.

General usage of the tc3xx command:

tc3xx  [-h] [--version] {tool name} ...

Tc3xx supports the abmhd tool to a generate an alternate boot mode header structure for firmware hex files. The tc3xx abmhd tool is controlled by options and produces a Pargen definition file. The definitions file is written to stdout id no output file option is given. This allows to "pipe" the definitions directly into Pargen.

tc3xx abmhd usage:

usage: tc3xx abmhd [-h] [--stad STADABM] [--from CHKSTART] [--to CHKEND] [--abmhdid ABMHDID] [--output filename] ADDRESS HEXFILE    

Generate TC3XX alternate boot mode header Pargen definition file.   

Example:
    tc3xx abmhd --stad 0x80028000 --from 0x8002000 --to 0x8004000  0x80000100 fw.hex | pargen --ihex -f abmhd -   

positional arguments:
  ADDRESS               flash address of alternate boot mode header.
  HEXFILE               name of hexfile with user data content    

options:
  -h, --help            show this help message and exit
  --stad STADABM, -s STADABM
                        user code start address (default: lowest address in hexfile)
  --from CHKSTART, -f CHKSTART
                        begin of range to be checked (default: lowest address in hexfile)
  --to CHKEND, -t CHKEND
                        end of range to be checked (default: highest address in hexfile + 1)
  --abmhdid ABMHDID, -i ABMHDID
                        alternate Boot Mode Header Identifier value (default=0xFA7CB359)
  --output filename, -o filename
                        file name of generated Pargen xml file (default: <stdout>)

If you have further ideas or you found some bugs, great! Create an issue or if you are able and willing to fix it by yourself, clone the repository and create a pull request.

The project uses the following non standard python libraries:

The whole source code is published under the BSD 3-Clause License. Consider also the different licenses of used third party libraries too!

We welcome contribution, but unless you explicitly state otherwise: Any contribution intentionally submitted for inclusion in the work by you, shall be licensed as above, without any additional terms or conditions.