First commit

.gitignore

@@ -0,0 +1,2 @@
+*.pyc
+elmo/*

README.md

@@ -0,0 +1,11 @@
+# ELMO Online
+
+ELMO Online is a Python library which proposes an encapsulation of the binary project ELMO.
+
+[MOW17] **Towards Practical Tools for Side
+Channel Aware Software Engineering : ’Grey Box’ Modelling for Instruction Leakages**
+by _David McCann, Elisabeth Oswald et Carolyn Whitnall_.
+https://www.usenix.org/conference/
+usenixsecurity17/technical-sessions/presentation/mccann.
+
+**ELMO GitHub**: https://github.com/sca-research/ELMO

engine.py

@@ -0,0 +1,192 @@
+import numpy as np
+import os
+
+def hweight(n):
+    c = 0
+    while n>0:
+        c += (n & 1)
+        n >>= 1
+    return c
+
+def hdistance(x,y):
+    return hweight(x^y)
+
+def binary_writing(n, nb_bits=32, with_hamming=False):
+    n = np.array(n)
+    w, h = np.zeros((nb_bits, len(n))), np.zeros((len(n)))
+
+    for ind in range(nb_bits):
+        w[ind] = (n & 1)
+        h += w[ind]
+
+        n >>= 1
+        ind += 1
+    
+    return (w, h) if with_hamming else w
+    
+PREVIOUS = 0
+CURRENT = 1
+SUBSEQUENT = 2
+
+class ELMOEngine:
+    def __init__(self):
+        self.load_coefficients()
+        self.reset_points()
+    
+    def _extract_data(self, nb):
+        coeffs = self.coefficients[self.pos:self.pos+nb]
+        self.pos += nb
+        return coeffs
+
+    def load_coefficients(self):
+        filename = os.path.dirname(os.path.abspath(__file__))+'/elmo/coeffs.txt'
+        self.coefficients = None
+        with open(filename, 'r') as _file:
+            self.coefficients = np.array([list(map(float, line.split())) for line in _file.readlines()[:2153]])
+        
+        if self.coefficients is None:
+            raise IOError('Problem to read the coefficients.')
+        
+        self.pos = 0
+        
+        self.constant = np.squeeze(self._extract_data(1))
+        
+        self.PrvInstr = self._extract_data(4)
+        self.SubInstr = self._extract_data(4)
+        
+        self.Operand1 = self._extract_data(32)
+        self.Operand2 = self._extract_data(32)
+        self.BitFlip1 = self._extract_data(32)
+        self.BitFlip2 = self._extract_data(32)
+        
+        self.HWOp1PrvInstr = self._extract_data(4)
+        self.HWOp2PrvInstr = self._extract_data(4)
+        self.HDOp1PrvInstr = self._extract_data(4)
+        self.HDOp2PrvInstr = self._extract_data(4)
+        self.HWOp1SubInstr = self._extract_data(4)
+        self.HWOp2SubInstr = self._extract_data(4)
+        self.HDOp1SubInstr = self._extract_data(4)
+        self.HDOp2SubInstr = self._extract_data(4)
+        
+        self.Operand1_bitinteractions = self._extract_data(496)
+        self.Operand2_bitinteractions = self._extract_data(496)
+        self.BitFlip1_bitinteractions = self._extract_data(496)
+        self.BitFlip2_bitinteractions = self._extract_data(496)
+    
+    def reset_points(self):
+        self.points = []
+        self.power = None
+        
+    def add_point(self, triplet, previous_ops, current_ops):
+        self.points.append((triplet, previous_ops, current_ops))
+    
+    def _dot(self, a, b):
+        return np.sum(a * b, axis=0)
+        
+    def calculate_point(self, triplet, previous_ops, current_ops, debug=False):
+        nb_points = triplet.shape[1]
+        instructiontype = triplet[CURRENT]
+        instructiontype = instructiontype % 5 # Type 5 = Instruction was not profiled
+                
+        # Previous
+        previous_instruction_typedec = triplet[PREVIOUS]
+        previous_instruction_type = np.zeros((5, nb_points))
+        for i in range(nb_points):
+            if previous_instruction_typedec[i] < 5:
+                previous_instruction_type[previous_instruction_typedec[i],i] = 1
+        
+        # Current
+        (current_op1_binary, hw_op1) = binary_writing(current_ops[0], with_hamming=True)
+        (current_op2_binary, hw_op2) = binary_writing(current_ops[1], with_hamming=True)
+
+        (current_op1_bitflip, hd_op1) = binary_writing(previous_ops[0] ^ current_ops[0], with_hamming=True)
+        (current_op2_bitflip, hd_op2) = binary_writing(previous_ops[1] ^ current_ops[1], with_hamming=True)
+        
+        current_instruction_typedec = instructiontype
+        current_instruction_type = np.zeros((5, nb_points))
+        for i in range(nb_points):
+            if triplet[CURRENT,i] < 5:
+                current_instruction_type[current_instruction_typedec[i],i] = 1
+        
+        # Subsequent
+        subsequent_instruction_typedec = triplet[SUBSEQUENT]
+        subsequent_instruction_type = np.zeros((5, nb_points))
+        for i in range(nb_points):
+            if subsequent_instruction_typedec[i] < 5:
+                subsequent_instruction_type[subsequent_instruction_typedec[i],i] = 1
+
+        # Component variables
+        PrvInstr_data = self._dot( previous_instruction_type[1:], self.PrvInstr[:,instructiontype] )
+        SubInstr_data = self._dot( subsequent_instruction_type[1:], self.SubInstr[:,instructiontype] )
+        
+        Operand1_data = self._dot( current_op1_binary, self.Operand1[:,instructiontype] )
+        Operand2_data = self._dot( current_op2_binary, self.Operand2[:,instructiontype] )
+        BitFlip1_data = self._dot( current_op1_bitflip, self.BitFlip1[:,instructiontype] )
+        BitFlip2_data = self._dot( current_op2_bitflip, self.BitFlip2[:,instructiontype] )
+        
+        HWOp1PrvInstr_data = hw_op1 * self._dot(previous_instruction_type[1:], self.HWOp1PrvInstr[:,instructiontype])
+        HWOp2PrvInstr_data = hw_op2 * self._dot(previous_instruction_type[1:], self.HWOp2PrvInstr[:,instructiontype])
+        HDOp1PrvInstr_data = hd_op1 * self._dot(previous_instruction_type[1:], self.HDOp1PrvInstr[:,instructiontype])
+        HDOp2PrvInstr_data = hd_op2 * self._dot(previous_instruction_type[1:], self.HDOp2PrvInstr[:,instructiontype])
+        HWOp1SubInstr_data = hw_op1 * self._dot(subsequent_instruction_type[1:], self.HWOp1SubInstr[:,instructiontype])
+        HWOp2SubInstr_data = hw_op2 * self._dot(subsequent_instruction_type[1:], self.HWOp2SubInstr[:,instructiontype])
+        HDOp1SubInstr_data = hd_op1 * self._dot(subsequent_instruction_type[1:], self.HDOp1SubInstr[:,instructiontype])
+        HDOp2SubInstr_data = hd_op2 * self._dot(subsequent_instruction_type[1:], self.HDOp2SubInstr[:,instructiontype])
+        
+        Operand1_bitinteractions_data = np.zeros((nb_points))
+        Operand2_bitinteractions_data = np.zeros((nb_points))
+        BitFlip1_bitinteractions_data = np.zeros((nb_points))
+        BitFlip2_bitinteractions_data = np.zeros((nb_points))
+        
+        count = 0
+        for i in range(32):
+            for j in range(i+1,32):
+                Operand1_bitinteractions_data += self.Operand1_bitinteractions[count,instructiontype] * current_op1_binary[i] * current_op1_binary[j]
+                Operand2_bitinteractions_data += self.Operand2_bitinteractions[count,instructiontype] * current_op2_binary[i] * current_op2_binary[j]
+
+                BitFlip1_bitinteractions_data += self.BitFlip1_bitinteractions[count,instructiontype] * current_op1_bitflip[i] * current_op1_bitflip[j]
+                BitFlip2_bitinteractions_data += self.BitFlip2_bitinteractions[count,instructiontype] * current_op2_bitflip[i] * current_op2_bitflip[j]
+
+                count += 1
+                
+        power = self.constant[instructiontype] \
+                    + PrvInstr_data + SubInstr_data \
+                    + Operand1_data + Operand2_data \
+                    + BitFlip1_data + BitFlip2_data \
+                    + HWOp1PrvInstr_data + HWOp2PrvInstr_data \
+                    + HDOp1PrvInstr_data + HDOp2PrvInstr_data \
+                    + HWOp1SubInstr_data + HWOp2SubInstr_data \
+                    + HDOp1SubInstr_data + HDOp2SubInstr_data \
+                    + Operand1_bitinteractions_data + Operand2_bitinteractions_data \
+                    + BitFlip1_bitinteractions_data + BitFlip2_bitinteractions_data
+        
+        for i in range(nb_points):
+            if triplet[CURRENT,i] == 5:
+                power[i] = self.constant[triplet[CURRENT,i]]
+                
+        if debug:
+            print([self.constant[instructiontype], \
+                       PrvInstr_data, SubInstr_data, \
+                       Operand1_data, Operand2_data, \
+                       BitFlip1_data, BitFlip2_data, \
+                       HWOp1PrvInstr_data, HWOp2PrvInstr_data, \
+                       HDOp1PrvInstr_data, HDOp2PrvInstr_data, \
+                       HWOp1SubInstr_data, HWOp2SubInstr_data, \
+                       HDOp1SubInstr_data, HDOp2SubInstr_data, \
+                       Operand1_bitinteractions_data, Operand2_bitinteractions_data, \
+                       BitFlip1_bitinteractions_data, BitFlip2_bitinteractions_data])
+        return power
+    
+    def run(self):
+        nb_points = len(self.points)
+        triplet = np.array([p[0] for p in self.points]).T # shape = (3, nb_points)
+        previous_ops = np.array([p[1] for p in self.points]).T # shape = (2, nb_points)
+        current_ops = np.array([p[2] for p in self.points]).T # shape = (2, nb_points)
+        
+        self.power = self.calculate_point(triplet, previous_ops, current_ops)
+    
+    def oneshot_point(self, triplet, previous_ops, current_ops):
+        self.reset_points()
+        self.add_point(triplet, previous_ops, current_ops)
+        self.run()
+        return self.power

executorthread.py

@@ -0,0 +1,66 @@
+from servicethread import OneShotServiceThread
+import subprocess
+
+import shutil
+from project_reader import ProjectReader
+
+global_variables = {}
+
+class ExecutorThread(OneShotServiceThread):
+    def __init__(self, ip, port, clientsocket, **kwargs):
+        super().__init__(ip, port, clientsocket)
+        self.projects = kwargs['projects'] if 'projects' in kwargs else None
+
+    def execute(self):
+        projects = self.projects
+        if project is None:
+            reader = ProjectReader()
+            projects = {sc.get_project_label(): sc for sc in reader.get_project_classes()}
+            print('Warning: need to research the projects.')
+        else:
+            print('Already have projects')
+        
+        data = self.protocol.get_data()
+        self.protocol.please_assert(data)
+        self.protocol.please_assert('project' in data)
+        self.protocol.please_assert(data['project'] in projects)
+        self.protocol.send_ack()
+
+        # Get the project
+        project = projects[data['project']]
+        
+        # Make the compilation
+        if False:
+            print('Compiling binary...')
+            
+            # Adapt the project source
+            with open('binaries/Frodo/frodo-base.c', 'r') as _src_file:
+                content = _src_file.read()
+                with open('binaries/Frodo/frodo.c', 'w') as _dst_file:
+                    _dst_file.write(content.replace('%NEED_TO_FILL%', str(n)))
+            
+            # Compile the project
+            make_directory = 'projects/{}/{}'.format(project.get_project_directory(), project.get_make_directory())
+            process = subprocess.Popen('make', shell=True, cwd=make_directory, executable='/bin/bash', stdout=subprocess.PIPE, stderr=subprocess.PIPE)
+            output, error = process.communicate()
+            if error and ('error' in error.decode('latin-1')):
+                print("Error to compile")
+                print(error)
+                raise Exception()
+            
+            # Save last compilation data
+            global_variables[project_name] = {
+                'last_n': n,
+            }
+        
+        # Generate the trace by launching ELMO
+        command = './elmo ../projects/{}/{}'.format(project.get_project_directory(), project.get_binary())
+        process = subprocess.Popen(command, shell=True, cwd='elmo_online/elmo/', executable='/bin/bash', stdout=subprocess.PIPE)
+        output, error = process.communicate()
+
+        # Send results
+        self.protocol.send_data({
+            'output': output.decode('latin-1') if output else None,
+            'error': error.decode('latin-1') if error else None,
+        })
+        self.protocol.close()

launch.py

@@ -0,0 +1,6 @@
+from .project_reader import Project
+
+reader = ProjectReader()
+projects = reader.get_projects()
+for key, project in projects.items():
+    globals()[key] = project

listeningthread.py

@@ -0,0 +1,39 @@
+from servicethread import PermanentServiceThread
+import socket
+
+class ListeningThread(PermanentServiceThread):
+    def __init__(self, host, port, threadclass, **kwargs):
+        super().__init__()
+        self.hostname = host
+        self.port = port
+        self.threadclass = threadclass
+        self.kwargs = kwargs
+    
+    def execute(self):
+        self.tcpsock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+        self.tcpsock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
+        self.tcpsock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEPORT, 1)
+        # self.tcpsock.setsockopt(socket.SOL_SOCKET, socket.SO_ATTACH_REUSEPORT_CBPF, 1)
+        self.tcpsock.bind((self.hostname, self.port))
+        self.tcpsock.listen(5)
+        print('[port][%s] Listening' % self.port)
+        
+        while self.is_running():
+            try:
+                (clientsocket, (ip, port)) = self.tcpsock.accept()
+                print('[port][{}] Accepted: {} <=> {}'.format(
+                    self.port,
+                    clientsocket.getsockname(),
+                    clientsocket.getpeername(),
+                ))
+                newthread = self.threadclass(ip, port, clientsocket, **self.kwargs)
+                newthread.start()
+            except socket.timeout:
+                pass
+
+    def stop(self):
+        super().stop()
+        clientsocker = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+        clientsocker.connect( (self.hostname, self.port) )
+        self.tcpsock.close()
+        print('[port][%s] Stop listening' % self.port)

project_base.py

@@ -0,0 +1,214 @@
+import os, re
+import socket
+from protocol import SocketTool
+import numpy as np
+
+def to_hex(v, nb_bits=16):
+    try:
+        v_hex = v.hex()
+    except AttributeError:
+        v_hex = hex(v)[2:]
+    return '0'*(nb_bits//4-len(v_hex)) + v_hex
+
+def split_octet(hexstr):
+    return [hexstr[i:i+2] for i in range(0, len(hexstr), 2)]
+
+def to_signed_hex(v, nb_bits=16):
+    return split_octet(to_hex(v & (2**nb_bits-1), nb_bits=nb_bits))
+
+def write(_input, uintXX, nb_bits=16):
+    uintXX = to_signed_hex(uintXX, nb_bits=nb_bits)
+    for i in range(nb_bits//8):
+        _input.write(uintXX[i]+'\n')
+    
+def write_list(_input, uint16_list):
+    for uint16 in uint16_list:
+        write(_input, uint16)
+
+def launch_simulation(quiet=False, **kwargs):
+    try:
+        s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+        s.connect(('localhost', 5000))
+        SocketTool.send_data(s, kwargs)
+        if not SocketTool.get_ack(s):
+            raise RuntimeError("The request has been refused !")
+        else:
+            data = SocketTool.get_data(s)
+            if data['error'] and not quiet:
+                raise Exception("The simulation return an error")
+            return data['output'], data['error']
+        s.close()
+    except IOError:
+        raise RuntimeError("The connection refused. Has the ELMO server been switch on ?")
+
+
+class SimulationProject:
+    _nb_bits_for_nb_challenges = 16
+    _project_directory = None
+    
+    ### Define the project
+    @classmethod
+    def get_project_directory(cl):
+        if cl._project_directory:
+            return cl._project_directory
+        else:
+            raise NotImplementedError()
+    
+    @classmethod
+    def set_project_directory(cl, project_directory):
+        cl._project_directory = project_directory
+
+    @classmethod
+    def get_project_label(cl):
+        return cl.get_project_directory()
+
+    @classmethod
+    def get_make_directory(cl):
+        return ''
+        
+    @classmethod
+    def get_binary(cl):
+        raise NotImplementedError()
+        
+    @classmethod
+    def get_parameters_names(cl):
+        return set()
+        
+    @classmethod
+    def adapt_project(cl, parameters):
+        return
+        
+    def get_challenge_format(self):
+        raise NotImplementedError()
+
+
+    ### Tools to realize the simulation of the project
+    def __init__(self, challenges=None):
+        self.elmo_folder = os.path.dirname(os.path.abspath(__file__))+'/elmo'
+        self.challenges = challenges
+        self.is_executed = False
+    
+    def set_challenges(self, challenges):
+        self.challenges = challenges
+
+    def get_writable_input_file(self):
+        return open('{}/input.txt'.format(self.elmo_folder), 'w')
+
+    def set_input_for_each_challenge(self, input, challenge):
+        format = self.get_challenge_format()
+
+        def aux(sizes, data):
+            if len(sizes) == 0:
+                write(input, data)
+            else:
+                assert len(data) == sizes[0], 'Incorrect format for challenge. Get {} instead of {}'.format(len(data), sizes[0])
+                for i in range(sizes[0]):
+                    aux(sizes[1:], data[i])
+
+        for num_part in range(len(format)):                
+            aux(format[num_part], challenge[num_part])
+
+    def set_input(self, input):
+        assert len(self.challenges) < 2**16, 'The number of challenges must be strictly lower than 65536. Currently, there are {} challenges.'.format(len(self.challenges))
+        write(input, len(self.challenges), nb_bits=self._nb_bits_for_nb_challenges)
+        for challenge in self.challenges:
+            self.set_input_for_each_challenge(input, challenge)
+            
+    def run(self):
+        with open('{}/input.txt'.format(self.elmo_folder), 'w') as _input:
+            self.set_input(_input)
+        launch_simulation(project=self.get_project_label(), quiet=False)
+        self.is_executed = True
+        
+    def get_asmtrace_filename(self):
+        return '{}/output/asmoutput/asmtrace00001.txt'.format(self.elmo_folder)
+        
+    def get_indexes_of(self, condition):
+        with open(self.get_asmtrace_filename(), 'r') as _file:
+            asmtrace = _file.readlines()
+            return [i for i, instr in enumerate(asmtrace) if condition(instr)]
+    
+    def get_number_of_traces(self):
+        return len(self.challenges)
+    
+    def get_results(self, only_filenames=False, reorganise=None, indexes=None):
+        assert self.is_executed
+        nb_traces = self.get_number_of_traces()
+
+        trace_filenames = []
+        for filename in os.listdir('{}/output/traces/'.format(self.elmo_folder)):
+            if re.search(r'^trace\d+\.trc$', filename):
+                trace_filenames.append('{}/output/traces/{}'.format(self.elmo_folder, filename))
+                if len(trace_filenames) >= nb_traces:
+                    break
+        
+        assert len(trace_filenames) == nb_traces
+        results = trace_filenames
+        
+        if not only_filenames:
+            for i in range(len(results)):
+                with open(results[i], 'r') as _file:
+                    if indexes is not None:
+                        results[i] = list(map(float,  _file.readlines()[indexes]))
+                    else:
+                        results[i] = list(map(float,  _file.readlines()))
+
+        if reorganise is not None:
+            results = reorganise(results)
+
+        return results
+    
+    def get_traces(self, reorganise=None, indexes=None):
+        results = self.get_results(only_filenames=False, reorganise=reorganise,indexes=indexes)
+
+        nb_traces = self.get_number_of_traces()
+        trace_length = len(results[0])
+
+        traces = np.zeros((nb_traces, trace_length))
+        for i in range(nb_traces):
+            traces[i,:] = results[i]
+            
+        if reorganise is not None:
+            traces = reorganise(traces)
+            
+        return traces
+
+    def get_printed_data(self):
+        with open('{}/output/printdata.txt'.format(self.elmo_folder), 'r') as _file:
+            data = list(map(lambda x: int(x, 16), _file.readlines()))
+        
+        nb_traces = self.get_number_of_traces()
+        nb_data_per_trace = len(data) // nb_traces
+        
+        return [data[nb_data_per_trace*i:nb_data_per_trace*(i+1)] for i in range(nb_traces)]        
+        
+    def analyse_operands(self, num_line, num_trace=1):
+        num_str = str(num_trace)
+        num_str = '0'*(5-len(num_str)) + num_str
+        
+        operands_filename = self.elmo_folder + '/output/operands/operands{}.txt'.format(num_str)
+        trace_filename = self.elmo_folder + '/output/traces/trace0000{}.trc'.format(num_trace)
+        
+        is_multiple = (type(num_line) is list)
+        if not is_multiple:
+            num_line = [num_line]
+        output = [{}]*len(num_line)
+        
+        with open(operands_filename, 'r') as _file:
+            lines = _file.readlines()
+            for i, num in enumerate(num_line):
+                line = lines[num].split()
+                data = list(map(int, line[0:7]))
+                output[i]['previous'] = data[0:2]
+                output[i]['current'] = data[2:4]
+                output[i]['triplet'] = data[4:7]
+                output[i]['other'] = list(map(float, line[7:]))
+        
+        with open(trace_filename, 'r') as _file:
+            lines = _file.readlines()
+            for i, num in enumerate(num_line):
+                line = lines[num]
+                output[i]['power'] = float(line)
+            
+        return output if is_multiple else output[0]
+

project_reader.py

@@ -0,0 +1,42 @@
+from project_base import SimulationProject
+import os, re
+import inspect
+
+PROJECTS_REPOSITORY = 'projects'
+
+class ProjectReader:
+    def __init__(self):
+        pass
+    
+    def get_projects(self):
+        projects = {}
+        
+        for root, repositories, files in os.walk(PROJECTS_REPOSITORY):
+            for filename in files:
+                if re.fullmatch(r'.*project.*\.py', filename):
+                    # Encapsulation the project
+                    complete_filename = root+'/'+filename
+                    globals = {
+                        #'__builtins__': {'__build_class__': __build_class__},
+                        'SimulationProject': SimulationProject,
+                    }
+                    locals = {}
+                    
+                    # Read the project code
+                    with open(complete_filename, 'r') as _file:
+                        project = '\n'.join(_file.readlines())
+                        exec(project, globals, locals)
+                    
+                    # Extract the simulations
+                    for key, obj in locals.item():
+                        if inspect.isclass(obj) and issubclass(obj, SimulationProject):
+                            if key in projects:
+                                print('Warning! Multiplie simulation with the same name. Simulation ignored: {} in {}'.format(key, complete_filename[len(PROJECTS_REPOSITORY)+1:]))
+                            else:
+                                obj.set_project_directory(root[len(PROJECTS_REPOSITORY)+1:])
+                                projects[key] = obj
+        
+        return projects
+        
+    def get_project_classes(self):
+        return self.get_projects().values()

projects/.gitignore

@@ -0,0 +1,3 @@
+*
+!.gitignore
+!elmoasmfunctions*

projects/elmoasmfunctions.s

@@ -0,0 +1,257 @@
+##
+## University of Bristol – Open Access Software Licence
+## Copyright (c) 2016, The University of Bristol, a chartered
+## corporation having Royal Charter number RC000648 and a charity
+## (number X1121) and its place of administration being at Senate
+## House, Tyndall Avenue, Bristol, BS8 1TH, United Kingdom.
+## All rights reserved
+##
+## Redistribution and use in source and binary forms, with or without
+## modification, are permitted provided that the following conditions
+## are met:
+##
+## 1. Redistributions of source code must retain the above copyright
+## notice, this list of conditions and the following disclaimer.
+##
+## 2. Redistributions in binary form must reproduce the above
+## copyright notice, this list of conditions and the following
+## disclaimer in the documentation and/or other materials provided
+## with the distribution.
+##
+## THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+## "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+## LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
+## FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
+## COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT,
+## INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+## (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+## SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+## HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+## STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+## ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+## OF THE POSSIBILITY OF SUCH DAMAGE.
+##
+## Any use of the software for scientific publications or commercial
+## purposes should be reported to the University of Bristol
+## (OSI-notifications@bristol.ac.uk and quote reference 2668). This is
+## for impact and usage monitoring purposes only.
+##
+## Enquiries about further applications and development opportunities
+## are welcome. Please contact elisabeth.oswald@bristol.ac.uk
+##
+
+.syntax unified
+.text
+.thumb
+
+.func starttrigger
+.global starttrigger
+
+starttrigger:
+    push {r0-r7}
+    movs r4, #0xE0
+    lsls r4, #24
+    movs r5, #0x04
+    eors r4, r5
+    movs r5, #1
+    str r5, [r4, #0]
+    pop {r0-r7}
+bx lr
+
+.endfunc
+
+.func endtrigger
+.global endtrigger
+
+endtrigger:
+    push {r0-r7}
+    movs r4, #0xE0
+    lsls r4, #24
+    movs r5, #0x04
+    eors r4, r5
+    movs r5, #0
+    str r5, [r4, #0]
+    pop {r0-r7}
+bx lr
+
+.endfunc
+
+.func readbyte
+.global readbyte
+
+readbyte:
+    push {r0-r7}
+    movs r4, #0xE1
+    lsls r4, #24
+    ldr r1, [r4, #0]
+    strb r1, [r0, #0]
+    pop {r0-r7}
+
+bx lr
+
+.endfunc
+
+.func LoadN
+.global LoadN
+
+LoadN:
+    push {r0-r7}
+    movs r4, #0xE1
+    lsls r4, #24
+    movs r5, #0x10
+    eors r4, r5
+    ldr r1, [r4, #0]
+    str r1, [r0, #0]
+    pop {r0-r7}
+
+bx lr
+
+.endfunc
+
+.func randbyte
+.global randbyte
+
+randbyte:
+    push {r0-r7}
+    movs r4, #0xE1
+    lsls r4, #24
+    movs r5, #0x04
+    eors r4, r5
+    ldr r1, [r4, #0]
+    strb r1, [r0, #0]
+    pop {r0-r7}
+
+bx lr
+
+.endfunc
+
+.func getstart
+.global getstart
+
+getstart:
+    push {r0-r7}
+    movs r4, #0xE1
+    lsls r4, #24
+    movs r5, #0x08
+    eors r4, r5
+    ldr r1, [r4, #0]
+    str r1, [r0, #0]
+    pop {r0-r7}
+
+bx lr
+
+.endfunc
+
+
+.func getruncount
+.global getruncount
+
+getruncount:
+    push {r0-r7}
+    movs r4, #0xE1
+    lsls r4, #24
+    movs r5, #0x0C
+    eors r4, r5
+    ldr r1, [r4, #0]
+    str r1, [r0, #0]
+    pop {r0-r7}
+
+bx lr
+
+.endfunc
+
+
+
+.func printbyte
+.global printbyte
+
+printbyte:
+    push {r0-r7}
+    movs r4, #0xE0
+    lsls r4, #24
+    ldrb r5, [r0]
+    str r5, [r4]
+    pop {r0-r7}
+bx lr
+
+.endfunc
+
+.func endprogram
+.global endprogram
+
+endprogram:
+    push {r0-r7}
+    movs r4, #0xF0
+    lsls r4, #24
+    movs r5, #0
+
+    str r5, [r4]
+    pop {r0-r7}
+bx lr
+
+.endfunc
+
+.func initialisemaskflow
+.global initialisemaskflow
+
+# Takes address of key as input (r0)
+initialisemaskflow:
+    push {r0-r7}
+    movs r4, #0xE0
+    lsls r4, #24
+    movs r5, #0x40
+    eors r4, r5
+    str r0, [r4]
+    pop {r0-r7}
+bx lr
+
+.endfunc
+
+.func resetmaskflow
+.global resetmaskflow
+
+resetmaskflow:
+    push {r0-r7}
+    movs r4, #0xE0
+    lsls r4, #24
+    movs r5, #0x42
+    eors r4, r5
+    movs r5, #0
+    str r5, [r4]
+    pop {r0-r7}
+bx lr
+
+.endfunc
+
+.func setmaskflowstart
+.global setmaskflowstart
+
+# Takes r0 as start number
+
+setmaskflowstart:
+    push {r0-r7}
+    movs r4, #0xE0
+    lsls r4, #24
+    movs r5, #0x44
+    eors r4, r5
+    str r0, [r4]
+    pop {r0-r7}
+bx lr
+
+.endfunc
+
+.func resetdatafile
+.global resetdatafile
+
+resetdatafile:
+    push {r0-r7}
+    movs r4, #0xE0
+    lsls r4, #24
+    movs r5, #0x46
+    eors r4, r5
+    movs r5, #0
+    str r5, [r4]
+    pop {r0-r7}
+bx lr
+
+.endfunc

projects/elmoasmfunctionsdef.h

@@ -0,0 +1,55 @@
+/*
+ * University of Bristol – Open Access Software Licence
+ * Copyright (c) 2016, The University of Bristol, a chartered
+ * corporation having Royal Charter number RC000648 and a charity
+ * (number X1121) and its place of administration being at Senate
+ * House, Tyndall Avenue, Bristol, BS8 1TH, United Kingdom.
+ * All rights reserved
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ *
+ * 2. Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials provided
+ * with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
+ * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
+ * COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT,
+ * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+ * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * Any use of the software for scientific publications or commercial
+ * purposes should be reported to the University of Bristol
+ * (OSI-notifications@bristol.ac.uk and quote reference 2668). This is
+ * for impact and usage monitoring purposes only.
+ *
+ * Enquiries about further applications and development opportunities
+ * are welcome. Please contact elisabeth.oswald@bristol.ac.uk
+ */
+
+extern void starttrigger(void);
+extern void endtrigger(void);
+extern void randbyte(unsigned char * pointer);
+extern void LoadN(void* addr);
+extern void readbyte(unsigned char * pointer);
+extern void printbyte(unsigned char * pointer);
+extern void endprogram(void);
+extern void getstart(unsigned int * pointer);
+extern void getruncount(unsigned int * pointer);
+extern void initialisemaskflow(unsigned char * pointer);
+extern void resetmaskflow(void);
+extern void setmaskflowstart(unsigned int start);
+extern void resetdatafile(void);

protocol.py

@@ -0,0 +1,134 @@
+import os
+import json
+
+
+class ClosureException(Exception):
+   pass
+
+
+# https://stackoverflow.com/questions/20007319/how-to-do-a-large-text-file-transfer-in-python
+
+class SocketTool:
+    @classmethod
+    def convert_to_bytes(cl, no):
+        result = bytearray()
+        result.append(no & 255)
+        for i in range(3):
+            no = no >> 8
+            result.append(no & 255)
+        return result
+
+    @classmethod
+    def bytes_to_number(cl, b):
+        import sys
+        b = b if (sys.version_info > (3, 0)) else map(ord, b)
+        
+        res = 0
+        for i in range(4):
+            res += b[i] << (i*8)
+        return res        
+        
+    @classmethod
+    def send_file(cl, s, filename, mode='rb'):
+        length = os.path.getsize(filename)
+        s.send(cl.convert_to_bytes(length)) # has to be 4 bytes
+        with open(filename, mode) as infile:
+            d = infile.read(1024*64) # We send by pack of 64 ko
+            while d:
+                s.send(d)
+                d = infile.read(1024*64)
+
+    @classmethod
+    def get_file(cl, s):
+        size = s.recv(4) # assuming that the size won't be bigger then 1GB
+        size = cl.bytes_to_number(size)
+        current_size = 0
+        buffer = b''
+        while current_size < size:
+            data = s.recv(1024*64)
+            if not data:
+                break
+            if len(data) + current_size > size:
+                data = data[:size-current_size] # trim additional data
+            buffer += data
+            # you can stream here to disk
+            current_size += len(data)
+        # you have entire file in memory
+        return buffer
+        
+    @classmethod
+    def send_data(cl, s, data):
+        data = json.dumps(data)
+        data = data.encode('utf-8')
+        s.send(cl.convert_to_bytes(len(data))) # has to be 4 bytes
+        s.send(data)
+        
+    @classmethod
+    def get_data(cl, s):
+        import sys
+        exception_class = json.decoder.JSONDecodeError if (sys.version_info > (3, 0)) else ValueError
+
+        try:
+            size = s.recv(4)
+            if not size:
+                return None
+            size = cl.bytes_to_number(size)
+            data = s.recv(size)
+            data = data.decode('utf-8')
+            data = json.loads(data)
+            return data
+        except exception_class:
+            return None
+
+    @classmethod
+    def send_ack(cl, s, positive=True):
+        data = 'OK' if positive else 'NO'
+        s.send(data.encode('ascii'))
+
+    @classmethod
+    def get_ack(cl, s):
+        data = s.recv(2).decode('ascii')
+        return (data == 'OK')
+
+
+class Protocol:
+    def __init__(self, clientsocket):
+        self.clientsocket = clientsocket
+    
+    def send_file(self, filename, mode='rb'):
+        SocketTool.send_file(self.clientsocket, filename, mode)
+    
+    def get_file(self):
+        return SocketTool.get_file(self.clientsocket)
+        
+    def send_data(self, data):
+        SocketTool.send_data(self.clientsocket, data)
+    
+    def get_data(self):
+        return SocketTool.get_data(self.clientsocket)
+        
+    def send_ack(self, positive=True):
+        SocketTool.send_ack(self.clientsocket, positive)
+
+    def send_nack(self, positive=True):
+        SocketTool.send_ack(self.clientsocket, not positive)
+
+    def get_ack(self):
+        return SocketTool.get_ack(self.clientsocket)
+        
+    def please_assert(self, condition):
+        if not condition:
+            self.send_nack()
+            self.clientsocket.close()
+            raise ClosureException
+    
+    def get_object_or_nack(self, model, **kwargs):
+        try:
+            return model.objects.get(**kwargs)
+        except model.DoesNotExist:
+            self.send_nack()
+            self.clientsocket.close()
+            raise ClosureException
+    
+    def close(self):
+        self.clientsocket.close()

run

@@ -0,0 +1,24 @@
+elmo_repository="elmo"
+elmo_source="https://github.com/sca-research/ELMO.git"
+
+echo "====== ELMO Online ======"
+if [ -d ${elmo_repository} ]; then
+    echo " - ELMO tool already installed."
+else
+    # Download the tool
+    git clone --depth=1 --branch=master ${elmo_source} ${elmo_repository}
+    rm -rf ./${elmo_repository}/.git
+    
+    # Compile the tool
+    cd ./${elmo_repository}
+    make
+    cd ..
+fi
+
+#current_directory=${PWD##*/}
+#echo " - Current directory: ${current_directory}"
+echo 
+
+#cd ..
+#python3 -m ${current_directory}.run_server
+python3 run_server.py

run_server.py

@@ -0,0 +1,36 @@
+from listeningthread import ListeningThread
+from executorthread import ExecutorThread
+
+def do_main_program(projects):
+    global thread, stop
+    thread = ListeningThread('localhost', 5000, ExecutorThread, projects=projects)
+    thread.start()
+
+def program_cleanup(signum, frame):
+    global thread, stop
+    thread.stop()
+    stop = True
+
+thread = None
+stop = False
+
+# Information
+from project_reader import ProjectReader
+reader = ProjectReader()
+projects = {sc.get_project_label(): sc for sc in reader.get_project_classes()}
+print('Available projects: %s' % list(projects.keys()))
+print('')
+
+# Execute
+print("Executing...")
+do_main_program(projects)
+print("Done ! And now, listening...")
+
+import signal
+signal.signal(signal.SIGINT, program_cleanup)
+signal.signal(signal.SIGTERM, program_cleanup)
+
+# Wait
+import time
+while not stop:
+    time.sleep(1)

servicethread.py

@@ -0,0 +1,43 @@
+import threading
+from protocol import Protocol, ClosureException
+
+class ServiceThread(threading.Thread):
+    def run(self):
+        self.execute()
+
+    def execute(self):
+        # Method where the service runs
+        pass
+
+
+class OneShotServiceThread(ServiceThread):
+    def __init__(self, ip, port, clientsocket):
+        threading.Thread.__init__(self)
+        self.ip = ip
+        self.port = port
+        self.clientsocket = clientsocket
+        self.protocol = Protocol(clientsocket)
+
+    def run(self):
+        try:
+            self.execute()
+        except ClosureException:
+            return
+            
+    def execute(self):
+        # Method where the service runs
+        pass
+
+
+class PermanentServiceThread(ServiceThread):
+    def __init__(self):
+        threading.Thread.__init__(self)
+        self._is_running = True
+
+    def is_running(self):
+        return self._is_running
+
+    def stop(self):
+        self._is_running = False
+
+