As of August 2020 the site you are on (wiki.newae.com) is deprecated, and content is now at rtfm.newae.com.

Adding Modules/Parameters

From ChipWhisperer Wiki
Revision as of 19:56, 14 July 2016 by Adriel (Talk | contribs) (Adding New Modules)

Jump to: navigation, search

Adding New Modules

In the new CW plugin architecture, all modules are scanned during the tool initialization, so new functionalities can be added by just dropping its file inside the respective folder:

  • chipwhisperer/software/chipwhisperer/common/results
  • chipwhisperer/software/chipwhisperer/common/traces
  • chipwhisperer/software/chipwhisperer/capture/acq_patterns
  • chipwhisperer/software/chipwhisperer/capture/auxiliary
  • chipwhisperer/software/chipwhisperer/capture/scopes
  • chipwhisperer/software/chipwhisperer/capture/scripts
  • chipwhisperer/software/chipwhisperer/capture/targets
  • chipwhisperer/software/chipwhisperer/analyzer/attacks
  • chipwhisperer/software/chipwhisperer/analyzer/preprocessing
  • chipwhisperer/software/chipwhisperer/analyzer/scripts
  • and some of its subfolders.

These paths are checked both inside the tool's root directory (where the tool is installed), and in the Project Folder (default is ~/chipwhisperer_projects), allowing the usage of custom modules without the requirement of being system administer. The CW tools scan these directories looking for classes that inherits from the Plugin class (a marker interface actually) in each public module (that doesn't begin with "_").

IMPORTANT: Accessing "Help->List Enabled/Disabled Plugins" in the tool's menu you'll find a list with all modules it tried to load. In case of any problem, you can check in the table the error message and its details. It should be easier to visuallise if you copy and paste the text to a text editor (example: notepad).

These folders usually have a file called base.py or _base.py that contains the base class to all plugins in these directories. Ex.:

from .base import PreprocessingBase
from chipwhisperer.common.utils.pluginmanager import Plugin


class AddNoiseRandom(PreprocessingBase, Plugin):
    _name = "Add Noise: Amplitude"
    _description = "Add random noise"
     
...

Adding Parameters

Parameters are used to allow easy access and manipulation of all object's main attibutes and actions. All parameters can be accessed anywhere in the code throught the Parameter class. It means that if you want to set/get any parameter, you can do it easily adding the follow lines to your code:

from chipwhisperer.common.utils.parameter import Parameter
...
Parameter.setParameter([path,..., value])
value = Parameter.getParameter([path,...])

or, if you have access to the api:

api.setParameter([path,..., value])
value = api.getParameter([path,...])

The easiest way to add parameters to your class, is to make it Parameterized (extending this class). It is an abstract class that declares a public interface and implements two manipulation methods to create/get the parameters and find it. As a general rule, you just need to:

  • import the Parameterized class: from chipwhisperer.common.utils.parameter import Parameterized
  • make your class extend it (no construtor call is needed here since the idea is to use it as an interface to avoid the problems with multiple inheritance - i.e.: the diamont one)
  • define a _name and a _description
  • register it if it is not readily accessible through a higher parameter hierarchy: self.getParams().register()
  • call self.getParams().addChildren([...])

The getParams() method does four things:

  • create a new Parameter if it doesn't exist;
  • create a group called _name;
  • create a child description parameter with the specified _description label;
  • return a reference to the parent group parameter.

Search is performed using the findParam([fullpath]) method.

Each parameter stores the data internally or externally, using a set/get pair - usefull to retrieve dynamic data. In this case, the @setupSetParam(nameOrPath) decorator should be used in the set method in order to syncronize the GUI when the method is called directly without using the parameter class.

More information about the Parameterized and the Parameter class can be found in its docstrings.

Basic example:

from chipwhisperer.common.utils.parameter import Parameterized, setupSetParam


class ResultsSave(Parameterized):
    _name = "Save to Files"
    _description = "Save correlation output to files."

    def __init__(self):
        # self.getParams().register()
        self.getParams().addChildren([
            {'name':'Save Raw Results', 'type':'bool', 'get':self.getEnabled, 'set':self.setEnabled},  # With value saved externally
            {'name':'Symbol', 'type':'list', 'values':['o', 's', 't', 'd', '+'], 'value':'o'},         # With value saved internally
        ])

        self.findParam("Symbol").setValue('t')
        s = self.findParam("Symbol").getValue()  # s = 't'

    def getEnabled(self):
        return self._enabled

    @setupSetParam("Save Raw Results")
    def setEnabled(self, enabled):
        self._enabled = enabled