peakrdl-python 2.3.0

Generate Python Register Access Layer (RAL) from SystemRDL

Introduction

PeakRDL Python is used to generate a python based Register Access Layer (RAL) from SystemRDL.

Documentation

See the peakrdl-python Documentation for more details

Getting Started

Installation

1. Install a recent version of Python 3
2. Install peakrdl-python

   python3 -m pip install peakrdl-python
   

3. (Optional) Install peakrdl, this is needed if you want to use peakrdl python from the command line

   python3 -m pip install peakrdl
   

Demo

This demonstration relies on downloading the systemRDL example from Accelera, available here: accelera-generic_example.rdl. This demonstration also assumes that peakrdl has been installed.

1. Build the Register Access Layer (RAL) from the systemRDL code

   peakrdl python accelera-generic_example.rdl -o .
   

   This will create a python package called some_register_map containing the python RAL
2. In addition to the RAL, peakrdl-python also generates a simulator that can be used to exercise the RAL without connecting to real hardware. Enter the following code into a file:

   """
   A demonstration of using peakrdl-python using the accelera generic example
   """
   # import the top level RAL class
   from some_register_map.reg_model import RegModel
   # import the simulator class
   from some_register_map.sim import Simulator
   
   from some_register_map.lib import NormalCallbackSet
   
   if __name__ == '__main__':
       # create an instance of the RAL with the callbacks directed at the hardware simulator
       hw_sim = Simulator(0)
       ral = RegModel(callbacks=NormalCallbackSet(read_callback=hw_sim.read,
                                                  write_callback=hw_sim.write))
   
       # read chip ID
       chip_id_part_number = ral.chip_id_reg.part_num.read()
       chip_id_revision_number = ral.chip_id_reg.part_num.read()
       print(f'Chip ID:{chip_id_part_number}.{chip_id_revision_number}')
   

   save it as some_register_map_demo.py
3. Run the example

   python3 -m some_register_map_demo
   

   This will generate the following output on the console:

   Chip ID:0.0
   

Usage

To make use of the RAL with real hardware or a different simulation, the callbacks will need to be connected to the appropriate access function in order to perform an address space reads and writes

Upgrading from previous versions (some important changes)

Upgrading from pre 0.9.0

In order to address a major limitation of peakrdl-python that prevented it from implementing the full systemRDL specification, a breaking API change was needed for handling blocks:

• registers (in register array)
• memory entries in a memory

Users are encouraged to upgrade in order to avoid this limitation. However, there is a legacy mode to support users with existing designs, see: Legacy Block Callback and Block Access in the documentation

Upgrading from pre 1.2.0

Version 1.2 introduced a new way to define the enumerations for the field encoding. This allows metadata from the systemRDL to propagate through to the generated code. This may break advanced usage of the python enumerations. User are encouraged to use the new feature, however, if there are problems with the old enumeration types (based on IntEnum) can be used, see Legacy Enumeration Types in the documentation

Upgrading from pre 2.0.0

Version 2.0.0 introduced a significant change to the process for building the register model python code. This change was intended to reduce the size of the generated code by only generating python classes for systemRDL components that required unique classes. The previous versions were more conservative and tended to generate a lot of duplicate classes.

Version 2.1.0 has improved this to ensure field encoding enumerations were correctly deduplicated.

The implementation requires a hash to be generated of each node in order to determine whether it is unique or not. This hash was incorperated within the class names which resulted in the code changing each time it was regenerated, version 2.1.0 introduces a option to either:

• The builtin python hash function, this is fast but is a salted hash so changes hashes export to export
• Use the SHA256 hash from the python hashlib standard library, this may slow down the export of large register models but will be consistent, therefore is useful if the resultant code is being checked into a version control system (such as GIT) and the differences are being reviewed