def execute_hardware(plan):
dma = []
hw_switch.reset()
ret_dma_base = int(
PL.ip_dict["SEG_{0}_Reg".format(metadata.DMA_names[0])][0], 16)
ret_dma_mmio = MMIO(ret_dma_base, 256)
ret_dma = DMA(ret_dma_base, 1)
ret_dma.create_buf(8388607)
prepare_execution(plan, dma, metadata.DMA[0][0][0])
hw_switch.commit()
## Timer Start
start_time = time.process_time()
ret_dma.transfer(8388607, 1)
for d in dma:
d.transfer()
for d in dma:
d.wait()
## Timer End
end_time = time.process_time()
print("Elapsed Test Time: ", end_time - start_time)
ret_dma.wait()
bytes_read = ret_dma_mmio.read(0x58)
ffi = pynq.drivers.dma.ffi
buf = ffi.buffer(ret_dma.buf, bytes_read)
view = np.frombuffer(buf, plan.dtype, -1).copy()
return view
class DMAWrapper:
def __init__(self,index):
#print('Send DMA: create index {0} name {1}'.format(index,metadata.DMA_names[index]))
base_addr=int(PL.ip_dict["SEG_{0}_Reg".format(metadata.DMA_names[index])][0],16)
#print('Send DMA: base_address {0:x}'.format(base_addr))
self.dma=DMA(base_addr,0)
self.ports=metadata.DMA[index]
def set_data(self,data,dtype):
self.length=len(data) * dtype.itemsize
#print('Send DMA: sending {0} bytes'.format(self.length))
self.dma.create_buf(self.length)
ffi=pynq.drivers.dma.ffi
buf=ffi.buffer(self.dma.buf,self.length)
view=np.frombuffer(buf,dtype,-1)
np.copyto(view,data,casting='same_kind')
def transfer(self):
#print('Send DMA: transfer started')
self.dma.transfer(self.length,0)
def wait(self):
self.dma.wait()
ffi = FFI()
# ### Constant definition:
# W and halfWidth are the dimension and the half-dimension of ROI image, CH is the number of channel of imageand TH is the value applied within the threshold filter.
W = 50
halfWidth = 0.5 * W
CH = 3
TH = 17
# ### DMA inizialization:
# Two differents DMAs are instantiated at the same address, one for reading and one for writing
dmaOut = DMA(0x40400000, 1)
dmaIn = DMA(0x40400000, 0)
# Two buffers are created containing data to be tranfered to PL (dmaIn) and output results (dmaOut)
dmaIn.create_buf(W * W * CH + 1)
dmaOut.create_buf(W * W)
# Pointers to buffers are taken in order to allow to write and read data. Another buffer is allocated (at the same position of the DMA's out one) to allow to read data using numpy.frombuffer function.
# First value of input buffer is set with threshold value.
pointIn = ffi.cast("uint8_t *", dmaIn.get_buf())
pointOut = ffi.cast("uint8_t *", dmaOut.get_buf())
c_buffer = ffi.buffer(pointOut, W * W)
pointIn[0] = TH
# ### Setting parameters for video analysis :
# Path to source folder.
source_folder = "./Assets/"
# Parameters for Cascadeclassifier function.
sc_fct = 1.3
hdmi_out = HDMI('out', video_mode=VMODE_1920x1080)
hdmi_in = HDMI('in')
hdmi_in.start()
hdmi_out.start()
print("HDMI IN Physical Address: ")
print(hex(hdmi_in.frame_phyaddr()))
print("HDMI OUT Physical Address: ")
print(hex(hdmi_out.frame_phyaddr()))
print()
# Memory allocation for data
print ("### Memory allocation data ###")
# 1. Input Conv1
dma0 = DMA(0x40400000,0)
inputConv1 = dma0.create_buf(0x17C4,0) # 0 means data is not cacheable
print("Physical InputConv1 Address")
print(dma0._bufPtr)
# 2. Output Conv1
dma1 = DMA(0x40400000,0)
outputConv2 = dma1.create_buf(0x19500,0)
print("Physical OutputConv2 Address")
print(dma1._bufPtr)
# 3. Weights Conv1
dma2 = DMA(0x40400000,0)
weightsConv1 = dma2.create_buf(0x500,0)
print("Physical Weights Conv1 Address")
print(dma2._bufPtr)
class Trace_Buffer:
"""Class for the trace buffer, leveraging the sigrok libraries.
This trace buffer class gets the traces from DMA and processes it using
the sigrok commands.
For PMODA and PMODB, pin numbers 0-7 correspond to the pins on the Pmod
interface. Although PMODA and PMODB are sharing the same trace buffer,
only one Pmod can be traced at a specific time.
For ARDUINO, pin numbers 0-5 correspond to A0-A5;
pin numbers 6-7 correspond to D0-D1;
pin numbers 8-19 correspond to D2-D13;
pin numbers 20-21 correspond to SDA and SCL.
Attributes
----------
if_id : int
The interface ID (PMODA, PMODB, ARDUINO).
pins : list
Array of pin numbers, 0-7 for PMODA or PMODB and 0-21 for ARDUINO.
protocol : str
The protocol the sigrok decoder are using, for example, I2C.
trace_csv: str
Absolute path of the `*.csv` trace that can be opened by text editor.
trace_sr: str
Absolute path of the `*.sr` trace file that can be unzipped.
trace_pd : str
Absolute path of the `*.pd` decoded file by sigrok decoder.
probes : list
The list of probes used for the trace, e.g., ['SCL','SDA'] for I2C.
dma : DMA
The DMA object associated with the trace buffer.
ctrl : MMIO
The MMIO class used to control the DMA.
rate: int
The sample rate of the traces, at most 100M samples per second.
samples : ndarray
The np array storing the 64-bit samples.
ffi: cffi.api.FFI
The FFI API to the underlying C structure
"""
def __init__(self,
if_id,
pins,
protocol,
probes=None,
trace=None,
rate=500000):
"""Return a new trace buffer object.
Users have to specify the location of the traces, even if no trace
has been imported from DMA yet. This method will construct the trace
from the DMA data.
The maximum sample rate is 100MHz.
For PMODA and PMODB, pin numbers 0-7 correspond to the pins on the
Pmod interface. Although PMODA and PMODB are sharing the same trace
buffer, only one Pmod can be traced at a specific time.
For ARDUINO, pin numbers 0-5 correspond to A0-A5;
pin numbers 6-7 correspond to D0-D1;
pin numbers 8-19 correspond to D2-D13;
pin numbers 20-21 correspond to SDA and SCL.
When using the trace buffer, only one out of the above 4 groups can be
traced in the current implementation.
The list `probes` depends on the protocol. For instance, the I2C
protocol requires a list of ['SCL','SDA'].
Parameters
----------
if_id : int
The interface ID (PMODA, PMODB, ARDUINO).
pins : list
List of pin numbers, 0-7 for PMODA or PMODB and 0-21 for ARDUINO.
protocol : str
The protocol the sigrok decoder are using, for example, I2C.
trace: str
The relative/absolute path of the trace file in `csv`/`sr` format.
rate : int
The rate of the samples, at most 100M.
"""
if os.geteuid() != 0:
raise EnvironmentError('Root permissions required.')
if not isinstance(protocol, str):
raise TypeError("Protocol name has to be a string.")
if not isinstance(rate, int):
raise TypeError("Sample rate has to be an integer.")
if not 1 <= rate <= 100000000:
raise ValueError("Sample rate out of range.")
if if_id in [PMODA, PMODB]:
dma_base, _, _ = PL.ip_dict["SEG_axi_dma_0_Reg"]
ctrl_base, ctrl_range, _ = PL.ip_dict["SEG_trace_cntrl_0_Reg2"]
elif if_id in [ARDUINO]:
dma_base, _, _ = PL.ip_dict["SEG_axi_dma_0_Reg1"]
ctrl_base, ctrl_range, _ = PL.ip_dict["SEG_trace_cntrl_0_Reg"]
else:
raise ValueError("No such IOP for instrumentation.")
self.if_id = if_id
if not pins:
raise ValueError("No pins specified to trace.")
elif if_id in [PMODA, PMODB]:
for p in pins:
if not p in range(8):
raise ValueError("Available pin numbers are 0-7.")
self.pins = np.array([7 - p for p in pins])
else:
for p in pins:
if not p in range(22):
raise ValueError("Available pin numbers are 0-21.")
self.pins = np.array([21 - p for p in pins])
if not probes:
self.probes = ['Pin {}'.format(i) for i in pins]
elif not isinstance(probes, list):
raise ValueError("Probes have to be a list.")
else:
self.probes = probes
self.dma = DMA(dma_base, direction=1)
self.ctrl = MMIO(ctrl_base, ctrl_range)
self.rate = rate
self.protocol = protocol
self.ffi = cffi.FFI()
self.samples = None
self.trace_pd = ''
if trace:
if not isinstance(trace, str):
raise TypeError("Trace path has to be a string.")
if not os.path.isfile(trace):
trace_abs = os.getcwd() + '/' + trace
else:
trace_abs = trace
if not os.path.isfile(trace_abs):
raise ValueError("Specified trace file does not exist.")
_, format = os.path.splitext(trace_abs)
if format == '.csv':
self.trace_csv = trace_abs
self.trace_sr = ''
elif format == '.sr':
self.trace_sr = trace_abs
self.trace_csv = ''
else:
raise ValueError("Only supporting csv or sr files.")
def __del__(self):
"""Destructor for trace buffer object.
Returns
-------
None
"""
del self.dma
def start(self, timeout=10):
"""Start the DMA to capture the traces.
If length is not specified, the maximum number of samples will
be captured.
Parameters
----------
timeout : int
The time in number of milliseconds to wait for DMA to be idle.
Return
------
None
"""
# Create buffer
self.dma.create_buf(MAX_NUM_SAMPLES * 8)
self.dma.transfer(MAX_NUM_SAMPLES * 8, direction=1)
# Wait for DMA to be idle
timer = timeout
while (self.ctrl.read(0x00) & 0x04) == 0:
sleep(0.001)
timer -= 1
if timer == 0:
raise RuntimeError("Timeout when waiting DMA to be idle.")
# Configuration
self.ctrl.write(TRACE_LENGTH_OFFSET, MAX_NUM_SAMPLES)
self.ctrl.write(TRACE_SAMPLE_RATE_OFFSET,
int(MAX_SAMPLE_RATE / self.rate))
self.ctrl.write(TRACE_CMP_LSW_OFFSET, 0x00000)
self.ctrl.write(TRACE_CMP_MSW_OFFSET, 0x00000)
# Start the DMA
self.ctrl.write(TRACE_CTRL_OFFSET, 0x01)
self.ctrl.write(TRACE_CTRL_OFFSET, 0x00)
def stop(self):
"""Stop the DMA after capture is done.
Note
----
There is an internal timeout mechanism in the DMA class.
Return
------
None
"""
# Wait for the DMA
self.dma.wait()
# Get samples from DMA
self.samples = np.frombuffer(self.ffi.buffer(self.dma.get_buf(64),
MAX_NUM_SAMPLES * 8),
dtype=np.uint64)
def show(self):
"""Show information about the specified protocol.
This method will print out useful information about the protocol.
Return
------
None
"""
if os.system("sigrok-cli --protocol-decoders " + \
self.protocol+" --show"):
raise RuntimeError('Sigrok-cli show failed.')
def csv2sr(self):
"""Translate the `*.csv` file to `*.sr` file.
The translated `*.sr` files can be directly used in PulseView to show
the waveform.
Note
----
This method also modifies the input `*.csv` file (the comment header,
usually 3 lines, will be removed).
Return
------
None
"""
name, _ = os.path.splitext(self.trace_csv)
self.trace_sr = name + ".sr"
temp = name + ".temp"
if os.system("rm -rf " + self.trace_sr):
raise RuntimeError('Trace sr file cannot be deleted.')
in_file = open(self.trace_csv, 'r')
out_file = open(temp, 'w')
# Copy only the contents; ignore comments
for i, line in enumerate(in_file):
if not line.startswith(';'):
out_file.write(line)
in_file.close()
out_file.close()
os.remove(self.trace_csv)
os.rename(temp, self.trace_csv)
command = "sigrok-cli -i " + self.trace_csv + \
" -I csv -o " + self.trace_sr
if os.system(command):
raise RuntimeError('Sigrok-cli csv to sr failed.')
def sr2csv(self):
"""Translate the `*.sr` file to `*.csv` file.
The translated `*.csv` files can be used for interactive plotting.
`*.csv` file is human readable, and can be opened using text editor.
Note
----
This method also removes the redundant header that is generated by
sigrok.
Return
------
None
"""
name, _ = os.path.splitext(self.trace_sr)
self.trace_csv = name + ".csv"
temp = name + ".temp"
if os.system("rm -rf " + self.trace_csv):
raise RuntimeError('Trace csv file cannot be deleted.')
command = "sigrok-cli -i " + self.trace_sr + \
" -O csv > " + temp
if os.system(command):
raise RuntimeError('Sigrok-cli sr to csv failed.')
in_file = open(temp, 'r')
out_file = open(self.trace_csv, 'w')
# Copy only the contents; ignore comments
for i, line in enumerate(in_file):
if not line.startswith(';'):
out_file.write(line)
in_file.close()
out_file.close()
os.remove(temp)
def decode(self, decoded_file, options=''):
"""Decode and record the trace based on the protocol specified.
The `decoded_file` contains the name of the output file.
The `option` specifies additional options to be passed to sigrok-cli.
For example, users can use option=':wordsize=9:cpol=1:cpha=0' to add
these options for the SPI decoder.
The decoder will also ignore the pin collected but not required for
decoding.
Note
----
The output file will have `*.pd` extension.
Note
----
The decoded file will be put into the specified path, or in the
working directory in case the path does not exist.
Parameters
----------
decoded_file : str
Name of the output file, which can be opened in text editor.
options : str
Additional options to be passed to sigrok-cli.
Return
------
None
"""
self.set_metadata()
if not isinstance(decoded_file, str):
raise TypeError("File name has to be a string.")
if not self.probes:
raise ValueError("Cannot decode without metadata.")
if os.path.isdir(os.path.dirname(decoded_file)):
decoded_abs = decoded_file
else:
decoded_abs = os.getcwd() + '/' + decoded_file
name, _ = os.path.splitext(self.trace_sr)
temp_file = name + '.temp'
if os.system('rm -rf ' + temp_file):
raise RuntimeError("Cannot remove temporary file.")
self.trace_pd = ''
if os.system('rm -rf ' + decoded_abs):
raise RuntimeError("Cannot remove old decoded file.")
pd_annotation = ''
for i in self.probes:
if not i == 'NC':
# Ignore pins not connected to device
pd_annotation += (':' + i.lower() + '=' + i)
command = "sigrok-cli -i " + self.trace_sr + " -P " + \
self.protocol + options + pd_annotation + (' > ' + temp_file)
if os.system(command):
raise RuntimeError('Sigrok-cli decode failed.')
f_decoded = open(decoded_abs, 'w')
f_temp = open(temp_file, 'r')
j = 0
for line in f_temp:
m = re.search('([0-9]+)-([0-9]+) (.*)', line)
if m:
while j < int(m.group(1)):
f_decoded.write('x\n')
j += 1
f_decoded.write(m.group(3) + '\n')
j += 1
while j < int(m.group(2)):
f_decoded.write('.\n')
j += 1
f_temp.close()
f_decoded.close()
self.trace_pd = decoded_abs
if os.system('rm -rf ' + temp_file):
raise RuntimeError("Cannot remove temporary file.")
if os.path.getsize(self.trace_pd) == 0:
raise RuntimeError("No transactions and decoded file is empty.")
def set_metadata(self):
"""Set metadata for the trace.
A `*.sr` file directly generated from `*.csv` will not have any
metadata. This method helps to set the sample rate, probe names, etc.
Return
------
None
"""
# Convert csv file to sr file, if necessary
if self.trace_sr == '':
self.csv2sr()
name, _ = os.path.splitext(self.trace_sr)
if os.system("rm -rf " + name):
raise RuntimeError('Directory cannot be deleted.')
if os.system("mkdir " + name):
raise RuntimeError('Directory cannot be created.')
if os.system("unzip -q " + self.trace_sr + " -d " + name):
raise RuntimeError('Unzip sr file failed.')
metadata = open(name + '/metadata', 'r')
temp = open(name + '/temp', 'w')
pat = "rate=0 Hz"
subst = "rate=" + str(self.rate) + " Hz"
j = 0
for i, line in enumerate(metadata):
if line.startswith("probe"):
# Set the probe names
temp.write("probe" + str(j + 1) + "=" + self.probes[j] + '\n')
j += 1
else:
# Set the sample rate
temp.write(line.replace(pat, subst))
metadata.close()
temp.close()
if os.system("rm -rf " + name + '/metadata'):
raise RuntimeError('Cannot remove metadata folder.')
if os.system("mv " + name + '/temp ' + name + '/metadata'):
raise RuntimeError('Cannot rename metadata folder.')
if os.system("cd "+ name +"; zip -rq " + \
self.trace_sr + " * ; cd .."):
raise RuntimeError('Zip sr file failed.')
if os.system("rm -rf " + name):
raise RuntimeError('Cannot remove temporary folder.')
def parse(self, parsed, start_pos, stop_pos):
"""Parse the input data and generate a `*.csv` file.
This method can be used along with the DMA. The input data is assumed
to be 64-bit. The generated `*.csv` file can be then used as the trace
file.
Note
----
PMODA and PMODB are sharing the same trace buffer with different sets
of pins, while ARDUINO has its own trace buffer.
Note
----
The parsed file will be put into the specified path, or in the
working directory in case the path does not exist.
Parameters
----------
parsed : str
Name of the parsed output file which can be opened in text editor.
start_pos : int
Starting sample number, no less than 1.
stop_pos : int
Stopping sample number, no more than the maximum number of samples.
Return
------
None
"""
if not isinstance(parsed, str):
raise TypeError("File name has to be an string.")
if not isinstance(start_pos, int):
raise TypeError("Start position has to be an integer.")
if not isinstance(stop_pos, int):
raise TypeError("Stop position has to be an integer.")
if not 1 <= start_pos <= stop_pos <= MAX_NUM_SAMPLES:
raise ValueError("Start or stop position out of range.")
if os.path.isdir(os.path.dirname(parsed)):
parsed_abs = parsed
else:
parsed_abs = os.getcwd() + '/' + parsed
if os.system('rm -rf ' + parsed_abs):
raise RuntimeError("Cannot remove old parsed file.")
with open(parsed_abs, 'w') as f:
for i in range(start_pos, stop_pos):
if self.if_id == PMODA:
sample = np.array(
list(np.binary_repr(self.samples[i], width=64))[32:])
io_direction = sample[8:16]
io_input = sample[16:24]
io_output = sample[24:]
io_direction = io_direction[self.pins]
io_input = io_input[self.pins]
io_output = io_output[self.pins]
elif self.if_id == PMODB:
sample = np.array(
list(np.binary_repr(self.samples[i], width=64))[:32])
io_direction = sample[8:16]
io_input = sample[16:24]
io_output = sample[24:]
io_direction = io_direction[self.pins]
io_input = io_input[self.pins]
io_output = io_output[self.pins]
else:
sample = np.array(
list(np.binary_repr(self.samples[i], width=64)))
io_direction = sample[:22]
io_input = sample[22:44]
io_output = np.append(sample[44:], ['0', '0'])
io_direction = io_direction[self.pins]
io_input = io_input[self.pins]
io_output = io_output[self.pins]
condition = [io_direction == '0', io_direction == '1']
list_val = np.select(condition, [io_output, io_input])
f.write(','.join(list_val) + '\n')
self.trace_csv = parsed_abs
self.trace_sr = ''
def display(self):
"""Draw digital waveforms in ipython notebook.
It utilises the wavedrom java script library, documentation for which
can be found here: https://code.google.com/p/wavedrom/.
Note
----
Only use this method in Jupyter notebook.
Note
----
WaveDrom.js and WaveDromSkin.js are required under the subdirectory js.
Example of the data format to draw waveform:
>>> data = {'signal': [
{'name': 'clk', 'wave': 'p.....|...'},
{'name': 'dat', 'wave': 'x.345x|=.x', 'data': ['D','A','T','A']},
{'name': 'req', 'wave': '0.1..0|1.0'},
{},
{'name': 'ack', 'wave': '1.....|01.'}
]}
Returns
-------
None
"""
# Copy the javascript to the notebook location
if not (os.path.isfile('./js/WaveDrom.js')
and os.path.isfile('./js/WaveDromSkin.js')):
if os.system("cp -rf " + \
os.path.dirname(os.path.realpath(__file__)) + \
'/js' + ' ./'):
raise RuntimeError('Cannnot copy wavedrom javascripts.')
# Convert sr file to csv file, if necessary
if self.trace_csv == '':
self.sr2csv()
# Read csv trace file
data_file = open(self.trace_csv, 'r')
# Construct the sample numbers and headers
head = dict()
head['text'] = ['tspan', {'class':'info h4'},
'Protocol decoder: ' + self.protocol + \
'; Sample rate: ' + str(self.rate) + ' samples/s']
head['tock'] = ''
# Setting up the json data
data = dict()
if self.trace_pd:
pd_file = open(self.trace_pd, 'r')
data['signal'] = [{
'name': '',
'wave': '',
'data': list()
} for _ in range(len(self.probes) + 1)]
i = 0
for data_line, pd_line in zip_longest(data_file, pd_file):
# Adding time line
if i % 10 == 0:
head['tock'] += (str(i) + ' ' * 10)
# Reading both raw data and decoded files
csv_data = list(data_line.rstrip().split(','))
if pd_line is not None:
pd_data = pd_line.rstrip()
else:
pd_data = 'x'
# Adding decoded data
if str(pd_data) in ['x', '.']:
data['signal'][0]['wave'] += str(pd_data)
else:
data['signal'][0]['wave'] += '4'
data['signal'][0]['data'].append(str(pd_data))
# Adding raw data
if i == 0:
ref = csv_data
for index, signal_name in enumerate(self.probes):
data['signal'][index + 1]['name'] = signal_name
data['signal'][index + 1]['wave'] += str(
csv_data[index])
else:
for index in range(len(self.probes)):
if csv_data[index] == ref[index]:
data['signal'][index + 1]['wave'] += '.'
else:
ref[index] = csv_data[index]
data['signal'][index + 1]['wave'] += str(
csv_data[index])
i += 1
# Removing NC signal and close file
for idx, val in enumerate(data['signal']):
if val['name'] == 'NC':
del data['signal'][idx]
pd_file.close()
else:
data['signal'] = [{
'name': '',
'wave': '',
'data': list()
} for _ in range(len(self.probes))]
for i, data_line in enumerate(data_file):
if i % 10 == 0:
head['tock'] += (str(i) + ' ' * 10)
csv_data = list(data_line.rstrip().split(','))
if i == 0:
ref = csv_data
for index, signal_name in enumerate(self.probes):
data['signal'][index]['name'] = signal_name
data['signal'][index]['wave'] += str(csv_data[index])
else:
for index in range(len(self.probes)):
if csv_data[index] == ref[index]:
data['signal'][index]['wave'] += '.'
else:
ref[index] = csv_data[index]
data['signal'][index]['wave'] += str(
csv_data[index])
data['head'] = head
# Close data file
data_file.close()
htmldata = '<script type="WaveDrom">' + json.dumps(data) + '</script>'
IPython.core.display.display_html(IPython.core.display.HTML(htmldata))
jsdata = 'WaveDrom.ProcessAll();'
IPython.core.display.display_javascript(
IPython.core.display.Javascript(
data=jsdata,
lib=['files/js/WaveDrom.js', 'files/js/WaveDromSkin.js']))
class Trace_Buffer:
"""Class for the trace buffer, leveraging the sigrok libraries.
This trace buffer class gets the traces from DMA and processes it using
the sigrok commands.
Note
----
The `sigrok-cli` library has to be installed before using this class.
Attributes
----------
protocol : str
The protocol the sigrok decoder are using.
trace_csv: str
The absolute path of the trace file `*.csv`.
trace_sr: str
The absolute path of the trace file `*.sr`, translated from `*.csv`.
trace_pd : str
The absolute path of the decoded file by sigrok.
probes : list
The list of probes used for the trace.
dma : DMA
The DMA object associated with the trace buffer.
ctrl : MMIO
The MMIO class used to control the DMA.
samplerate: int
The samplerate of the traces.
data : cffi.FFI.CData
The pointer to the starting address of the trace data.
"""
def __init__(self, if_id, protocol, trace=None, data=None,
samplerate=500000):
"""Return a new trace buffer object.
Users have to specify the location of the traces, even if no trace
has been imported from DMA yet. This method will construct the trace
from the DMA data.
The maximum sample rate is 100MHz.
Note
----
The probes selected by `mask` does not include any tristate probe.
Parameters
----------
if_id : int
The interface ID (PMODA, PMODB, ARDUINO).
protocol : str
The protocol the sigrok decoder are using.
trace: str
The relative/absolute path of the trace file.
data : cffi.FFI.CData
The pointer to the starting address of the data.
samplerate : int
The rate of the samples.
"""
if os.geteuid() != 0:
raise EnvironmentError('Root permissions required.')
if not isinstance(protocol, str):
raise TypeError("Protocol name has to be a string.")
if data != None:
if not isinstance(data, cffi.FFI.CData):
raise TypeError("Data pointer has wrong type.")
if not isinstance(samplerate, int):
raise TypeError("Sample rate has to be an integer.")
if not 1 <= samplerate <= 100000000:
raise ValueError("Sample rate out of range.")
if if_id in [PMODA, PMODB]:
dma_base = int(PL.ip_dict["SEG_axi_dma_0_Reg"][0],16)
ctrl_base = int(PL.ip_dict["SEG_trace_cntrl_0_Reg2"][0],16)
ctrl_range = int(PL.ip_dict["SEG_trace_cntrl_0_Reg2"][1],16)
elif if_id in [ARDUINO]:
dma_base = int(PL.ip_dict["SEG_axi_dma_0_Reg1"][0],16)
ctrl_base = int(PL.ip_dict["SEG_trace_cntrl_0_Reg"][0],16)
ctrl_range = int(PL.ip_dict["SEG_trace_cntrl_0_Reg"][1],16)
else:
raise ValueError("No such IOP for instrumentation.")
self.dma = DMA(dma_base, direction=1)
self.ctrl = MMIO(ctrl_base, ctrl_range)
self.samplerate = samplerate
self.protocol = protocol
self.data = data
self.probes = []
self.trace_pd = ''
if trace != None:
if not isinstance(trace, str):
raise TypeError("Trace path has to be a string.")
if not os.path.isfile(trace):
trace_abs = os.getcwd() + '/' + trace
else:
trace_abs = trace
if not os.path.isfile(trace_abs):
raise ValueError("Specified trace file does not exist.")
_, format = os.path.splitext(trace_abs)
if format == '.csv':
self.trace_csv = trace_abs
self.trace_sr = ''
elif format == '.sr':
self.trace_sr = trace_abs
self.trace_csv = ''
else:
raise ValueError("Only supporting csv or sr files.")
def __del__(self):
"""Destructor for trace buffer object.
Parameters
----------
None
Returns
-------
None
"""
del(self.dma)
def start(self, timeout=10):
"""Start the DMA to capture the traces.
Parameters
----------
timeout : int
The time in number of milliseconds to wait for DMA to be idle.
Return
------
None
"""
# Create buffer
self.dma.create_buf(MAX_NUM_SAMPLES*8)
self.dma.transfer(MAX_NUM_SAMPLES*8, direction=1)
# Wait for DMA to be idle
timer = timeout
while (self.ctrl.read(0x00) & 0x04)==0:
sleep(0.001)
timer -= 1
if (timer==0):
raise RuntimeError("Timeout when waiting DMA to be idle.")
# Configuration
self.ctrl.write(TRACE_LENGTH_OFFSET, MAX_NUM_SAMPLES)
self.ctrl.write(TRACE_SAMPLE_RATE_OFFSET, \
int(MAX_SAMPLE_RATE / self.samplerate))
self.ctrl.write(TRACE_CMP_LSW_OFFSET, 0x00000)
self.ctrl.write(TRACE_CMP_MSW_OFFSET, 0x00000)
# Start the DMA
self.ctrl.write(TRACE_CTRL_OFFSET,0x01)
self.ctrl.write(TRACE_CTRL_OFFSET,0x00)
def stop(self):
"""Stop the DMA after capture is done.
Note
----
There is an internal timeout mechanism in the DMA class.
Parameters
----------
None
Return
------
None
"""
# Wait for the DMA
self.dma.wait()
# Get 64-bit samples from DMA
self.data = self.dma.get_buf(64)
def show(self):
"""Show information about the specified protocol.
Parameters
----------
None
Return
------
None
"""
if os.system("sigrok-cli --protocol-decoders " + \
self.protocol+" --show"):
raise RuntimeError('Sigrok-cli show failed.')
def csv2sr(self):
"""Translate the `*.csv` file to `*.sr` file.
The translated `*.sr` files can be directly used in PulseView to show
the waveform.
Note
----
This method also modifies the input `*.csv` file (the comment header,
usually 3 lines, will be removed).
Parameters
----------
None
Return
------
None
"""
name, _ = os.path.splitext(self.trace_csv)
self.trace_sr = name + ".sr"
temp = name + ".temp"
if os.system("rm -rf " + self.trace_sr):
raise RuntimeError('Trace sr file cannot be deleted.')
in_file = open(self.trace_csv, 'r')
out_file = open(temp, 'w')
# Copy only the contents; ignore comments
for i, line in enumerate(in_file):
if not line.startswith(';'):
out_file.write(line)
in_file.close()
out_file.close()
os.remove(self.trace_csv)
os.rename(temp, self.trace_csv)
command = "sigrok-cli -i " + self.trace_csv + \
" -I csv -o " + self.trace_sr
if os.system(command):
raise RuntimeError('Sigrok-cli csv to sr failed.')
def sr2csv(self):
"""Translate the `*.sr` file to `*.csv` file.
The translated `*.csv` files can be used for interactive plotting.
It is human readable.
Note
----
This method also removes the redundant header that is generated by
sigrok.
Parameters
----------
None
Return
------
None
"""
name, _ = os.path.splitext(self.trace_sr)
self.trace_csv = name + ".csv"
temp = name + ".temp"
if os.system("rm -rf " + self.trace_csv):
raise RuntimeError('Trace csv file cannot be deleted.')
command = "sigrok-cli -i " + self.trace_sr + \
" -O csv > " + temp
if os.system(command):
raise RuntimeError('Sigrok-cli sr to csv failed.')
in_file = open(temp, 'r')
out_file = open(self.trace_csv, 'w')
# Copy only the contents; ignore comments
for i, line in enumerate(in_file):
if not line.startswith(';'):
out_file.write(line)
in_file.close()
out_file.close()
os.remove(temp)
def decode(self, decoded_file, options=''):
"""Decode and record the trace based on the protocol specified.
The `decoded_file` contains the name of the output file.
The `option` specifies additional options to be passed to sigrok-cli.
For example, users can use option=':wordsize=9:cpol=1:cpha=0' to add
these options for the SPI decoder.
The decoder will also ignore the pin collected but not required for
decoding.
Note
----
The output file will have `*.pd` extension.
Note
----
The decoded file will be put into the specified path, or in the
working directory in case the path does not exist.
Parameters
----------
decoded_file : str
The name of the file recording the outputs.
options : str
Additional options to be passed to sigrok-cli.
Return
------
None
"""
if not isinstance(decoded_file, str):
raise TypeError("File name has to be a string.")
if self.probes == []:
raise ValueError("Cannot decode without metadata.")
if os.path.isdir(os.path.dirname(decoded_file)):
decoded_abs = decoded_file
else:
decoded_abs = os.getcwd() + '/' + decoded_file
name, _ = os.path.splitext(self.trace_sr)
temp_file = name + '.temp'
if os.system('rm -rf ' + temp_file):
raise RuntimeError("Cannot remove temporary file.")
self.trace_pd = ''
if os.system('rm -rf ' + decoded_abs):
raise RuntimeError("Cannot remove old decoded file.")
pd_annotation = ''
for i in self.probes:
if not i=='NC':
# Ignore pins not connected to device
pd_annotation += (':'+i.lower()+'='+i)
command = "sigrok-cli -i " + self.trace_sr + " -P " + \
self.protocol + options + pd_annotation + (' > ' + temp_file)
if os.system(command):
raise RuntimeError('Sigrok-cli decode failed.')
f_decoded = open(decoded_abs, 'w')
f_temp = open(temp_file, 'r')
j = 0
for line in f_temp:
m = re.search('([0-9]+)-([0-9]+) (.*)', line)
if m:
while (j < int(m.group(1))):
f_decoded.write('\n')
j += 1
while (j <= int(m.group(2))):
f_decoded.write(m.group(3) + '\n')
j += 1
f_temp.close()
f_decoded.close()
self.trace_pd = decoded_abs
if os.system('rm -rf ' + temp_file):
raise RuntimeError("Cannot remove temporary file.")
if os.path.getsize(self.trace_pd)==0:
raise RuntimeError("No transactions and decoded file is empty.")
def set_metadata(self, probes):
"""Set metadata for the trace.
A `*.sr` file directly generated from `*.csv` will not have any
metadata. This method helps to set the sample rate, probe names, etc.
The list `probes` depends on the protocol. For instance, the I2C
protocol requires a list of ['SDA','SCL'].
Parameters
----------
probes : list
A list of probe names.
Return
------
None
"""
if not isinstance(probes, list):
raise TypeError("Probes have to be in a list.")
# Convert csv file to sr file, if necessary
if self.trace_sr == '':
self.csv2sr()
self.probes = probes
name, _ = os.path.splitext(self.trace_sr)
if os.system("rm -rf " + name):
raise RuntimeError('Directory cannot be deleted.')
if os.system("mkdir " + name):
raise RuntimeError('Directory cannot be created.')
if os.system("unzip -q "+ self.trace_sr + " -d " + name):
raise RuntimeError('Unzip sr file failed.')
metadata = open(name + '/metadata', 'r')
temp = open(name + '/temp', 'w')
pat = "samplerate=0 Hz"
subst = "samplerate=" + str(self.samplerate) +" Hz"
j = 0
for i, line in enumerate(metadata):
if line.startswith("probe"):
# Set the probe names
temp.write("probe"+str(j+1)+"="+probes[j]+'\n')
j += 1
else:
# Set the sample rate
temp.write(line.replace(pat, subst))
metadata.close()
temp.close()
if os.system("rm -rf "+ name + '/metadata'):
raise RuntimeError('Cannot remove metadata folder.')
if os.system("mv " + name + '/temp ' + name + '/metadata'):
raise RuntimeError('Cannot rename metadata folder.')
if os.system("cd "+ name +"; zip -rq " + \
self.trace_sr + " * ; cd .."):
raise RuntimeError('Zip sr file failed.')
if os.system("rm -rf " + name):
raise RuntimeError('Cannnot remove temporary folder.')
def parse(self, parsed, start=0, stop=MAX_NUM_SAMPLES, mask=MASK_ALL,
tri_sel=[], tri_0=[], tri_1=[]):
"""Parse the input data and generate a `*.csv` file.
This method can be used along with the DMA. The input data is assumed
to be 64-bit. The generated `*.csv` file can be then used as the trace
file.
To extract certain bits from the 64-bit data, use the parameter
`mask`.
Note
----
The probe pins selected by `mask` does not include any tristate probe.
To specify a set of tristate probe pins, e.g., users can set
tri_sel = [0x0000000000000004],
tri_0 = [0x0000000000000010], and
tri_1 = [0x0000000000000100].
In this example, the 3rd probe from the LSB is the selection probe;
the 5th probe is selected if selection probe is 0, otherwise the 9th
probe is selected. There can be multiple sets of tristate probe pins.
Note
----
The parsed file will be put into the specified path, or in the
working directory in case the path does not exist.
Parameters
----------
parsed : str
The file name of the parsed output.
start : int
The first 64-bit sample of the trace.
stop : int
The last 64-bit sample of the trace.
mask : int
A 64-bit mask to be applied to the 64-bit samples.
tri_sel : list
The list of tristate selection probe pins.
tri_0 : list
The list of probe pins selected when the selection probe is 0.
tri_1 : list
The list probe pins selected when the selection probe is 1.
Return
------
None
"""
if not isinstance(parsed, str):
raise TypeError("File name has to be an string.")
if not isinstance(start, int):
raise TypeError("Sample number has to be an integer.")
if not isinstance(stop, int):
raise TypeError("Sample number has to be an integer.")
if not 1 <= (stop-start) <= MAX_NUM_SAMPLES:
raise ValueError("Data length has to be in [1,{}]."\
.format(MAX_NUM_SAMPLES))
if not isinstance(mask, int):
raise TypeError("Data mask has to be an integer.")
if not 0<=mask<=MASK_ALL:
raise ValueError("Data mask out of range.")
if not isinstance(tri_sel, list):
raise TypeError("Selection probe pins have to be in a list.")
if not isinstance(tri_0, list) or not isinstance(tri_1, list):
raise TypeError("Data probe pins have to be in a list.")
if not len(tri_sel)==len(tri_0)==len(tri_1):
raise ValueError("Inconsistent length for tristate lists.")
for element in tri_sel:
if not isinstance(element, int) or not 0<element<=MASK_ALL:
raise TypeError("Selection probe has to be an integer.")
if not (element & element-1)==0:
raise ValueError("Selection probe can only have 1-bit set.")
if not (element & mask)==0:
raise ValueError("Selection probe has be excluded from mask.")
for element in tri_0:
if not isinstance(element, int) or not 0<element<=MASK_ALL:
raise TypeError("Data probe has to be an integer.")
if not (element & element-1)==0:
raise ValueError("Data probe can only have 1-bit set.")
if not (element & mask)==0:
raise ValueError("Data probe has be excluded from mask.")
for element in tri_1:
if not isinstance(element, int) or not 0<element<=MASK_ALL:
raise TypeError("Data probe has to be an integer.")
if not (element & element-1)==0:
raise ValueError("Data probe can only have 1-bit set.")
if not (element & mask)==0:
raise ValueError("Data probe has be excluded from mask.")
if os.path.isdir(os.path.dirname(parsed)):
parsed_abs = parsed
else:
parsed_abs = os.getcwd() + '/' + parsed
if os.system('rm -rf ' + parsed_abs):
raise RuntimeError("Cannot remove old parsed file.")
with open(parsed_abs, 'w') as f:
for i in range(start, stop):
raw_val = self.data[i] & MASK_ALL
list_val = []
for j in range(63,-1,-1):
if (mask & 1<<j)>>j:
list_val.append(str((raw_val & 1<<j)>>j))
else:
for selection in tri_sel:
idx = tri_sel.index(selection)
if (selection & 1<<j)>>j:
if ((raw_val & 1<<j)>>j)==0:
log = tri_0[idx].bit_length()-1
list_val.append(
str((raw_val & 1<<log)>>log))
else:
log = tri_1[idx].bit_length()-1
list_val.append(
str((raw_val & 1<<log)>>log))
temp = ','.join(list_val)
f.write(temp + '\n')
self.trace_csv = parsed_abs
self.trace_sr = ''
def display(self, start_pos, stop_pos):
"""Draw digital waveforms in ipython notebook.
It utilises the wavedrom java script library, documentation for which
can be found here: https://code.google.com/p/wavedrom/.
Note
----
Only use this method in Jupyter notebook.
Note
----
WaveDrom.js and WaveDromSkin.js are required under the subdirectory js.
Example of the data format to draw waveform:
>>> data = {'signal': [
{'name': 'clk', 'wave': 'p.....|...'},
{'name': 'dat', 'wave': 'x.345x|=.x', 'data': ['D','A','T','A']},
{'name': 'req', 'wave': '0.1..0|1.0'},
{},
{'name': 'ack', 'wave': '1.....|01.'}
]}
Parameters
----------
start_pos : int
The starting sample number (relative to the trace).
stop_pos : int
The stopping sample number (relative to the trace).
Returns
-------
None
"""
if self.probes == []:
raise ValueError("Cannot display without metadata.")
if not isinstance(start_pos, int):
raise TypeError("Start position has to be an integer.")
if not 1 <= start_pos <= MAX_NUM_SAMPLES:
raise ValueError("Start position out of range.")
if not isinstance(stop_pos, int):
raise TypeError("Stop position has to be an integer.")
if not 1 <= stop_pos <= MAX_NUM_SAMPLES:
raise ValueError("Stop position out of range.")
# Copy the javascript to the notebook location
if os.system("cp -rf " + \
os.path.dirname(os.path.realpath(__file__)) + \
'/js' + ' ./'):
raise RuntimeError('Cannnot copy wavedrom javascripts.')
# Convert sr file to csv file, if necessary
if self.trace_csv == '':
self.sr2csv()
# Read csv trace file
with open(self.trace_csv, 'r') as data_file:
csv_data = list(csv.reader(data_file))
# Read decoded file
with open(self.trace_pd, 'r') as pd_file:
pd_data = list(csv.reader(pd_file))
# Construct the decoded transactions
data = {}
data['signal']=[]
if self.trace_pd != '':
temp_val = {'name': '', 'wave': '', 'data': []}
for i in range(start_pos, stop_pos):
if i==start_pos:
ref = pd_data[i]
if not ref:
temp_val['wave'] += 'x'
else:
temp_val['wave'] += '4'
temp_val['data'].append(''.join(pd_data[i]))
else:
if pd_data[i] == ref:
temp_val['wave'] += '.'
else:
ref = pd_data[i]
if not ref:
temp_val['wave'] += 'x'
else:
temp_val['wave'] += '4'
temp_val['data'].append(''.join(pd_data[i]))
data['signal'].append(temp_val)
# Construct the jason format data
for signal_name in self.probes:
index = self.probes.index(signal_name)
temp_val = {'name': signal_name, 'wave': ''}
for i in range(start_pos, stop_pos):
if i==start_pos:
ref = csv_data[i][index]
temp_val['wave'] += str(csv_data[i][index])
else:
if csv_data[i][index] == ref:
temp_val['wave'] += '.'
else:
ref = csv_data[i][index]
temp_val['wave'] += str(csv_data[i][index])
data['signal'].append(temp_val)
# Construct the sample numbers and headers
head = {}
head['text'] = ['tspan', {'class':'info h4'}, \
'Protocol decoder: ' + self.protocol + \
'; Sample rate: ' + str(self.samplerate) + ' samples/s']
head['tock'] = ''
for i in range(start_pos, stop_pos):
if i%2:
head['tock'] += ' '
else:
head['tock'] += (str(i)+' ')
data['head'] = head
htmldata = '<script type="WaveDrom">' + json.dumps(data) + '</script>'
IPython.core.display.display_html(IPython.core.display.HTML(htmldata))
jsdata = 'WaveDrom.ProcessAll();'
IPython.core.display.display_javascript(
IPython.core.display.Javascript(
data=jsdata, \
lib=['files/js/WaveDrom.js', 'files/js/WaveDromSkin.js']))