summaryrefslogtreecommitdiff
path: root/tools
diff options
context:
space:
mode:
Diffstat (limited to 'tools')
-rw-r--r--tools/__pycache__/dsss_demod_test_waveform_gen.cpython-38.pycbin0 -> 3107 bytes
-rw-r--r--tools/__pycache__/presig_gen.cpython-38.pycbin0 -> 5250 bytes
-rw-r--r--tools/butter_filter_gen.py93
-rw-r--r--tools/crypto_test.c46
-rw-r--r--tools/crypto_test_runner.py46
-rw-r--r--tools/cwt_wavelet_header_gen.py29
-rw-r--r--tools/dsss_demod_test.c109
-rw-r--r--tools/dsss_demod_test_runner.py241
-rw-r--r--tools/dsss_demod_test_waveform_gen.py86
-rw-r--r--tools/e2e_test.c111
-rw-r--r--tools/fft_window_header_gen.py59
-rw-r--r--tools/fmeas_export_ocxo_2day.binbin0 -> 4908532 bytes
-rw-r--r--tools/freq_meas_test.c106
-rw-r--r--tools/freq_meas_test_runner.py39
-rw-r--r--tools/gold_code_header_gen.py70
-rw-r--r--tools/grid_freq_psd_spl_108pt.json1
-rw-r--r--tools/hum_generator.py111
-rw-r--r--tools/ldparser.py126
-rw-r--r--tools/linkmem.py276
-rw-r--r--tools/linksize.py62
-rw-r--r--tools/linktracer.py118
-rw-r--r--tools/mapparse.py129
-rw-r--r--tools/presig_gen.py141
-rw-r--r--tools/reed_solomon.py91
24 files changed, 2090 insertions, 0 deletions
diff --git a/tools/__pycache__/dsss_demod_test_waveform_gen.cpython-38.pyc b/tools/__pycache__/dsss_demod_test_waveform_gen.cpython-38.pyc
new file mode 100644
index 0000000..1512990
--- /dev/null
+++ b/tools/__pycache__/dsss_demod_test_waveform_gen.cpython-38.pyc
Binary files differ
diff --git a/tools/__pycache__/presig_gen.cpython-38.pyc b/tools/__pycache__/presig_gen.cpython-38.pyc
new file mode 100644
index 0000000..b28b1a9
--- /dev/null
+++ b/tools/__pycache__/presig_gen.cpython-38.pyc
Binary files differ
diff --git a/tools/butter_filter_gen.py b/tools/butter_filter_gen.py
new file mode 100644
index 0000000..0bb81bc
--- /dev/null
+++ b/tools/butter_filter_gen.py
@@ -0,0 +1,93 @@
+#!/usr/bin/env python3
+
+import math
+import sys
+import contextlib
+
+import scipy.signal as sig
+import numpy as np
+
+
+@contextlib.contextmanager
+def wrap(left='{', right='}', file=None, end=''):
+ print(left, file=file, end=end)
+ yield
+ print(right, file=file, end=end)
+
+@contextlib.contextmanager
+def print_include_guards(macro_name):
+ print(f'#ifndef {macro_name}')
+ print(f'#define {macro_name}')
+ print()
+ yield
+ print()
+ print(f'#endif /* {macro_name} */')
+
+macro_float = lambda f: f'{f}'.replace('.', 'F').replace('-', 'N').replace('+', 'P')
+
+ordinal = lambda n: "%d%s" % (n,"tsnrhtdd"[(n//10%10!=1)*(n%10<4)*n%10::4])
+
+SI_TABLE = {-18: 'a', -15: 'f', -12: 'p', -9: 'n', -6: 'µ', -3: 'm', 0: '', 3: 'k', 6: 'M', 9: 'G', 12: 'T', 15: 'P', 18: 'E'}
+def siprefix(x, space=' ', unit=''):
+ l = math.log10(x)//3*3
+ if l in SI_TABLE:
+ return f'{x/10**l}{space}{SI_TABLE[l]}{unit}'
+ return f'{x}{space}{unit}'
+
+if __name__ == '__main__':
+ import argparse
+ parser = argparse.ArgumentParser()
+ parser.add_argument('-m', '--macro-name', default='butter_filter', help='Prefix for output macro names')
+ parser.add_argument('fc', type=float, help='Corner frequency [Hz]')
+ parser.add_argument('fs', type=float, help='Sampling rate [Hz]')
+ parser.add_argument('n', type=int, nargs='?', default=6, help='Filter order')
+ args = parser.parse_args()
+
+ sos = sig.butter(args.n, args.fc, fs=args.fs, output='sos')
+
+ print('/* THIS IS A GENERATED FILE. DO NOT EDIT! */')
+ print()
+ with print_include_guards(f'__BUTTER_FILTER_GENERATED_{args.n}_{macro_float(args.fc)}_{macro_float(args.fs)}__'):
+
+ print(f'/* {ordinal(args.n)} order Butterworth IIR filter coefficients')
+ print(f' *')
+ print(f' * corner frequency f_c = {siprefix(args.fc)}Hz')
+ print(f' * sampling rate f_s = {siprefix(args.fs)}Hz')
+ print(f' */')
+ print()
+ print(f'#define {args.macro_name.upper()}_ORDER {args.n}')
+ print(f'#define {args.macro_name.upper()}_CLEN {(args.n+1)//2}')
+
+ # scipy.signal.butter by default returns extremely small bs for the first biquad and large ones for subsequent
+ # sections. Balance magnitudes to reduce possible rounding errors.
+ first_biquad_bs = sos[0][:3]
+ approx_mag = round(math.log10(np.mean(first_biquad_bs)))
+ mags = [approx_mag // len(sos)] * len(sos)
+ mags[0] += approx_mag - sum(mags)
+ sos[0][:3] /= 10**approx_mag
+ sos = np.array([ sec * np.array([10**mag, 10**mag, 10**mag, 1, 1, 1]) for mag, sec in zip(mags, sos) ])
+
+ ones = np.ones([100000])
+ _, steady_state = sig.sosfilt(sos, ones, zi=np.zeros([(args.n+1)//2, 2]))
+
+ print(f'#define {args.macro_name.upper()}_COEFF ', end='')
+ for sec in sos:
+ bs, ases = sec[:3], sec[4:6]
+
+ with wrap():
+ print('.b=', end='')
+ with wrap():
+ print(', '.join(f'{v}' for v in bs), end='')
+ print(', .a=', end='')
+ with wrap():
+ print(', '.join(f'{v}' for v in ases), end='')
+ print(', ', end='')
+ print()
+
+ print(f'#define {args.macro_name.upper()}_STEADY_STATE ', end='')
+ for sec in steady_state:
+ with wrap():
+ print(', '.join(f'{v}' for v in sec), end='')
+ print(', ', end='')
+ print()
+
diff --git a/tools/crypto_test.c b/tools/crypto_test.c
new file mode 100644
index 0000000..410fac2
--- /dev/null
+++ b/tools/crypto_test.c
@@ -0,0 +1,46 @@
+
+#include <stdint.h>
+#include <math.h>
+#include <unistd.h>
+#include <stdio.h>
+#include <string.h>
+#include <errno.h>
+#include <stdlib.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <sys/fcntl.h>
+
+#include "crypto.h"
+
+void oob_trigger_activated(enum trigger_domain domain, int serial) {
+ printf("oob_trigger_activated(%d, %d)\n", domain, serial);
+ fflush(stdout);
+}
+
+void print_usage() {
+ fprintf(stderr, "Usage: crypto_test [auth_key_hex]\n");
+}
+
+int main(int argc, char **argv) {
+ if (argc != 2) {
+ fprintf(stderr, "Error: Invalid arguments.\n");
+ print_usage();
+ return 1;
+ }
+
+ uint8_t auth_key[16];
+
+ for (size_t i=0; argv[1][i+0] != '\0' && argv[1][i+1] != '\0' && i/2<sizeof(auth_key); i+= 2) {
+ char buf[3] = { argv[1][i+0], argv[1][i+1], 0};
+ char *endptr;
+ auth_key[i/2] = strtoul(buf, &endptr, 16);
+ if (!endptr || *endptr != '\0') {
+ fprintf(stderr, "Invalid authkey\n");
+ return 1;
+ }
+ }
+
+ printf("rc=%d\n", oob_message_received(auth_key));
+
+ return 0;
+}
diff --git a/tools/crypto_test_runner.py b/tools/crypto_test_runner.py
new file mode 100644
index 0000000..34c8b59
--- /dev/null
+++ b/tools/crypto_test_runner.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+import subprocess
+from os import path
+import binascii
+import re
+
+import presig_gen
+
+def do_test(domain, value, height, root_key, binary, expect_fail=False):
+ auth = presig_gen.gen_at_height(domain, value, height, root_key)
+ auth = binascii.hexlify(auth).decode()
+
+ output = subprocess.check_output([binary, auth])
+ *lines, rc_line = output.decode().splitlines()
+ rc = int(re.match('^rc=(\d+)$', rc_line).group(1))
+ assert expect_fail == (rc == 0)
+
+def run_tests(root_key, max_height, binary):
+ for domain, value in {
+ 'all': 'all',
+ 'vendor': presig_gen.TEST_VENDOR,
+ 'series': presig_gen.TEST_SERIES,
+ 'country': presig_gen.TEST_COUNTRY,
+ 'region': presig_gen.TEST_REGION,
+ }.items():
+ for height in range(max_height):
+ do_test(domain, value, height, root_key, binary)
+ do_test(domain, 'fail', height, root_key, binary, expect_fail=True)
+ do_test('fail', 'fail', height, root_key, binary, expect_fail=True)
+ do_test('', '', height, root_key, binary, expect_fail=True)
+ do_test(domain, value, max_height, root_key, binary, expect_fail=True)
+ do_test(domain, value, max_height+1, root_key, binary, expect_fail=True)
+
+if __name__ == '__main__':
+ import argparse
+ parser = argparse.ArgumentParser()
+ parser.add_argument('keyfile', help='Root key file')
+ parser.add_argument('max_height', type=int, default=8, nargs='?', help='Height of generated prekeys')
+ default_binary = path.abspath(path.join(path.dirname(__file__), '../build/tools/crypto_test'))
+ parser.add_argument('binary', default=default_binary, nargs='?', help='crypto_test binary to use')
+ args = parser.parse_args()
+
+ with open(args.keyfile, 'r') as f:
+ root_key = binascii.unhexlify(f.read().strip())
+
+ run_tests(root_key, args.max_height, args.binary)
diff --git a/tools/cwt_wavelet_header_gen.py b/tools/cwt_wavelet_header_gen.py
new file mode 100644
index 0000000..8be785b
--- /dev/null
+++ b/tools/cwt_wavelet_header_gen.py
@@ -0,0 +1,29 @@
+#!/usr/bin/env python3
+
+import textwrap
+
+import scipy.signal as sig
+import numpy as np
+
+if __name__ == '__main__':
+ import argparse
+ parser = argparse.ArgumentParser()
+ parser.add_argument('n', type=int, help='Window size')
+ parser.add_argument('w', type=float, help='Wavelet width')
+ parser.add_argument('-v', '--variable', default='cwt_ricker_table', help='Name for alias variable pointing to generated wavelet LUT')
+ args = parser.parse_args()
+
+ print(f'/* CWT Ricker wavelet LUT for {args.n} sample window of width {args.w}. */')
+ varname = f'cwt_ricker_{args.n}_window_{str(args.w).replace(".", "F")}'
+ print(f'const float {varname}[{args.n}] = {{')
+
+ win = sig.ricker(args.n, args.w)
+ par = ' '.join(f'{f:>015.12e}f,' for f in win)
+ print(textwrap.fill(par,
+ initial_indent=' '*4, subsequent_indent=' '*4,
+ width=120,
+ replace_whitespace=False, drop_whitespace=False))
+ print('};')
+ print()
+ print(f'const float * const {args.variable} __attribute__((weak)) = {varname};')
+
diff --git a/tools/dsss_demod_test.c b/tools/dsss_demod_test.c
new file mode 100644
index 0000000..f7df111
--- /dev/null
+++ b/tools/dsss_demod_test.c
@@ -0,0 +1,109 @@
+
+#include <stdint.h>
+#include <math.h>
+#include <unistd.h>
+#include <stdio.h>
+#include <string.h>
+#include <errno.h>
+#include <stdlib.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <sys/fcntl.h>
+
+#include "dsss_demod.h"
+
+void handle_dsss_received(symbol_t data[static TRANSMISSION_SYMBOLS]) {
+ printf("data sequence received: [ ");
+ for (size_t i=0; i<TRANSMISSION_SYMBOLS; i++) {
+ //printf("%+3d", ((data[i]&1) ? 1 : -1) * (data[i]>>1));
+ printf("%2d", data[i]);
+ if (i+1 < TRANSMISSION_SYMBOLS)
+ printf(", ");
+ }
+ printf(" ]\n");
+}
+
+void print_usage() {
+ fprintf(stderr, "Usage: dsss_demod_test [test_data.bin] [optional recording channel number]\n");
+}
+
+int main(int argc, char **argv) {
+ if (argc != 2 && argc != 3) {
+ fprintf(stderr, "Error: Invalid arguments.\n");
+ print_usage();
+ return 1;
+ }
+
+ int fd = open(argv[1], O_RDONLY);
+ struct stat st;
+ if (fstat(fd, &st)) {
+ fprintf(stderr, "Error querying test data file size: %s\n", strerror(errno));
+ return 2;
+ }
+
+ if (st.st_size < 0 || st.st_size > 10000000) {
+ fprintf(stderr, "Error reading test data: too much test data (size=%zd)\n", st.st_size);
+ return 2;
+ }
+
+ if (st.st_size % sizeof(float) != 0) {
+ fprintf(stderr, "Error reading test data: file size is not divisible by %zd (size=%zd)\n", sizeof(float), st.st_size);
+ return 2;
+ }
+
+ char *buf = malloc(st.st_size);
+ if (!buf) {
+ fprintf(stderr, "Error allocating memory");
+ return 2;
+ }
+
+ int record_channel = -1;
+ if (argc == 3) {
+ char *endptr;
+ record_channel = strtoul(argv[2], &endptr, 10);
+ if (!endptr || *endptr != '\0') {
+ fprintf(stderr, "Invalid channel number \"%s\"\n", argv[2]);
+ return 1;
+ }
+ }
+
+ if (record_channel != -1)
+ fprintf(stderr, "Reading %zd samples test data...", st.st_size/sizeof(float));
+ ssize_t nread = 0;
+ while (nread < st.st_size) {
+ ssize_t rc = read(fd, buf + nread, st.st_size - nread);
+
+ if (rc == -EINTR || rc == -EAGAIN)
+ continue;
+
+ if (rc < 0) {
+ fprintf(stderr, "\nError reading test data: %s\n", strerror(errno));
+ return 2;
+ }
+
+ if (rc == 0) {
+ fprintf(stderr, "\nError reading test data: Unexpected end of file\n");
+ return 2;
+ }
+
+ nread += rc;
+ }
+ if (record_channel != -1)
+ fprintf(stderr, " done.\n");
+
+ const size_t n_samples = st.st_size / sizeof(float);
+ float *buf_f = (float *)buf;
+
+ if (record_channel != -1)
+ fprintf(stderr, "Starting simulation.\n");
+
+ struct dsss_demod_state demod;
+ dsss_demod_init(&demod);
+ for (size_t i=0; i<n_samples; i++) {
+ //fprintf(stderr, "Iteration %zd/%zd\n", i, n_samples);
+ dsss_demod_step(&demod, buf_f[i], i);
+ }
+
+ free(buf);
+ return 0;
+}
diff --git a/tools/dsss_demod_test_runner.py b/tools/dsss_demod_test_runner.py
new file mode 100644
index 0000000..d3c3cfc
--- /dev/null
+++ b/tools/dsss_demod_test_runner.py
@@ -0,0 +1,241 @@
+#!/usr/bin/env python3
+
+import os
+import sys
+from os import path
+import subprocess
+import json
+from collections import namedtuple, defaultdict
+from tqdm import tqdm
+import uuid
+import multiprocessing
+import sqlite3
+import time
+from urllib.parse import urlparse
+import tempfile
+import itertools
+
+import numpy as np
+np.set_printoptions(linewidth=240)
+
+from dsss_demod_test_waveform_gen import load_noise_gen, modulate as dsss_modulate
+
+
+def build_test_binary(nbits, thf, decimation, symbols, cachedir):
+ build_id = str(uuid.uuid4())
+ builddir = path.join(cachedir, build_id)
+ os.mkdir(builddir)
+
+ cwd = path.join(path.dirname(__file__), '..')
+
+ env = os.environ.copy()
+ env['BUILDDIR'] = path.abspath(builddir)
+ env['DSSS_GOLD_CODE_NBITS'] = str(nbits)
+ env['DSSS_DECIMATION'] = str(decimation)
+ env['DSSS_THRESHOLD_FACTOR'] = str(thf)
+ env['DSSS_WAVELET_WIDTH'] = str(0.73 * decimation)
+ env['DSSS_WAVELET_LUT_SIZE'] = str(10 * decimation)
+ env['TRANSMISSION_SYMBOLS'] = str(symbols)
+
+ with open(path.join(builddir, 'make_stdout.txt'), 'w') as stdout,\
+ open(path.join(builddir, 'make_stderr.txt'), 'w') as stderr:
+ subprocess.run(['make', 'clean', os.path.abspath(path.join(builddir, 'tools/dsss_demod_test'))],
+ env=env, cwd=cwd, check=True, stdout=stdout, stderr=stderr)
+
+ return build_id
+
+def sequence_matcher(test_data, decoded, max_shift=3):
+ match_result = []
+ for shift in range(-max_shift, max_shift):
+ failures = -shift if shift < 0 else 0 # we're skipping the first $shift symbols
+ a = test_data if shift > 0 else test_data[-shift:]
+ b = decoded if shift < 0 else decoded[shift:]
+ for i, (ref, found) in enumerate(itertools.zip_longest(a, b)):
+ if ref is None: # end of signal
+ break
+ if ref != found:
+ failures += 1
+ match_result.append(failures)
+ failures = min(match_result)
+ return failures/len(test_data)
+
+ResultParams = namedtuple('ResultParams', ['nbits', 'thf', 'decimation', 'symbols', 'seed', 'amplitude', 'background'])
+
+def run_test(seed, amplitude_spec, background, nbits, decimation, symbols, thfs, lookup_binary, cachedir):
+ noise_gen, noise_params = load_noise_gen(background)
+
+ test_data = np.random.RandomState(seed=seed).randint(0, 2 * (2**nbits), symbols)
+
+ signal = np.repeat(dsss_modulate(test_data, nbits) * 2.0 - 1, decimation)
+ # We're re-using the seed here. This is not a problem.
+ noise = noise_gen(seed, len(signal), *noise_params)
+ amplitudes = amplitude_spec[0] * 10 ** np.linspace(0, amplitude_spec[1], amplitude_spec[2])
+ # DEBUG
+ my_pid = multiprocessing.current_process().pid
+ wql = len(amplitudes) * len(thfs)
+ print(f'[{my_pid}] starting, got workqueue of length {wql}')
+ i = 0
+ # Map lsb to sign to match test program
+ # test_data = (test_data>>1) * (2*(test_data&1) - 1)
+ # END DEBUG
+
+ output = []
+ for amp in amplitudes:
+ with tempfile.NamedTemporaryFile(dir=cachedir) as f:
+ waveform = signal*amp + noise
+ f.write(waveform.astype('float32').tobytes())
+ f.flush()
+ # DEBUG
+ fcopy = f'/tmp/test-{path.basename(f.name)}'
+ import shutil
+ shutil.copy(f.name, fcopy)
+ # END DEBUG
+
+ for thf in thfs:
+ rpars = ResultParams(nbits, thf, decimation, symbols, seed, amp, background)
+ cmdline = [lookup_binary(nbits, thf, decimation, symbols), f.name]
+ # DEBUG
+ starttime = time.time()
+ # END DEBUG
+ try:
+ proc = subprocess.run(cmdline, stdout=subprocess.PIPE, encoding='utf-8', check=True, timeout=300)
+
+ lines = proc.stdout.splitlines()
+ matched = [ l.partition('[')[2].partition(']')[0]
+ for l in lines if l.strip().startswith('data sequence received:') ]
+ matched = [ [ int(elem) for elem in l.split(',') ] for l in matched ]
+
+ ser = min(sequence_matcher(test_data, match) for match in matched) if matched else None
+ output.append((rpars, ser))
+ # DEBUG
+ #print(f'[{my_pid}] ran {i}/{wql}: time={time.time() - starttime}\n {ser=}\n {rpars}\n {" ".join(cmdline)}\n {fcopy}', flush=True)
+ i += 1
+ # END DEBUG
+
+ except subprocess.TimeoutExpired:
+ output.append((rpars, None))
+ # DEBUG
+ print(f'[{my_pid}] ran {i}/{wql}: Timeout!\n {rpars}\n {" ".join(cmdline)}\n {fcopy}', flush=True)
+ i += 1
+ # END DEBUG
+ print(f'[{my_pid}] finished.')
+ return output
+
+def parallel_generator(db, table, columns, builder, param_list, desc, context={}, params_mapper=lambda *args: args,
+ disable_cache=False):
+ with multiprocessing.Pool(multiprocessing.cpu_count()) as pool:
+ with db as conn:
+ jobs = []
+ for params in param_list:
+ found_res = conn.execute(
+ f'SELECT result FROM {table} WHERE ({",".join(columns)}) = ({",".join("?"*len(columns))})',
+ params_mapper(*params)).fetchone()
+
+ if found_res and not disable_cache:
+ yield params, json.loads(*found_res)
+
+ else:
+ jobs.append((params, pool.apply_async(builder, params, context)))
+
+ pool.close()
+ print('Using', len(param_list) - len(jobs), 'cached jobs', flush=True)
+ with tqdm(total=len(jobs), desc=desc) as tq:
+ for i, (params, res) in enumerate(jobs):
+ # DEBUG
+ print('Got result', i, params, res)
+ # END DEBUG
+ tq.update(1)
+ result = res.get()
+ with db as conn:
+ conn.execute(f'INSERT INTO {table} VALUES ({"?,"*len(params)}?,?)',
+ (*params_mapper(*params), json.dumps(result), timestamp()))
+ yield params, result
+ pool.join()
+
+if __name__ == '__main__':
+ import argparse
+ parser = argparse.ArgumentParser()
+ parser.add_argument('-d', '--dump', help='Write results to JSON file')
+ parser.add_argument('-c', '--cachedir', default='dsss_test_cache', help='Directory to store build output and data in')
+ parser.add_argument('-n', '--no-cache', action='store_true', help='Disable result cache')
+ parser.add_argument('-b', '--batches', type=int, default=1, help='Number of batches to split the computation into')
+ parser.add_argument('-i', '--index', type=int, default=0, help='Batch index to compute')
+ parser.add_argument('-p', '--prepare', action='store_true', help='Prepare mode: compile runners, then exit.')
+ args = parser.parse_args()
+
+ DecoderParams = namedtuple('DecoderParams', ['nbits', 'thf', 'decimation', 'symbols'])
+# dec_paramses = [ DecoderParams(nbits=nbits, thf=thf, decimation=decimation, symbols=20)
+# for nbits in [5, 6]
+# for thf in [4.5, 4.0, 5.0]
+# for decimation in [10, 5, 22] ]
+ dec_paramses = [ DecoderParams(nbits=nbits, thf=thf, decimation=decimation, symbols=100)
+ for nbits in [5, 6]
+ for thf in [3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 8.5, 9.0, 9.5, 10.0]
+ for decimation in [1, 2, 3, 4, 5, 6, 9, 10, 11, 12, 16, 22, 30, 40, 50] ]
+# dec_paramses = [ DecoderParams(nbits=nbits, thf=thf, decimation=decimation, symbols=100)
+# for nbits in [5, 6, 7, 8]
+# for thf in [1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 8.5, 9.0, 9.5, 10.0]
+# for decimation in [1, 2, 3, 4, 5, 6, 9, 10, 11, 12, 16, 22, 30, 40, 50] ]
+
+ build_cache_dir = path.join(args.cachedir, 'builds')
+ data_cache_dir = path.join(args.cachedir, 'data')
+ os.makedirs(build_cache_dir, exist_ok=True)
+ os.makedirs(data_cache_dir, exist_ok=True)
+
+ build_db = sqlite3.connect(path.join(args.cachedir, 'build_db.sqlite3'))
+ build_db.execute('CREATE TABLE IF NOT EXISTS builds (nbits, thf, decimation, symbols, result, timestamp)')
+ timestamp = lambda: int(time.time()*1000)
+
+ builds = dict(parallel_generator(build_db, table='builds', columns=['nbits', 'thf', 'decimation', 'symbols'],
+ builder=build_test_binary, param_list=dec_paramses, desc='Building decoders',
+ context=dict(cachedir=build_cache_dir)))
+ print('Done building decoders.')
+ if args.prepare:
+ sys.exit(0)
+
+ GeneratorParams = namedtuple('GeneratorParams', ['seed', 'amplitude_spec', 'background'])
+ gen_params = [ GeneratorParams(rep, (5e-3, 1, 5), background)
+ #GeneratorParams(rep, (0.05e-3, 3.5, 50), background)
+ for rep in range(50)
+ for background in ['meas://fmeas_export_ocxo_2day.bin', 'synth://grid_freq_psd_spl_108pt.json'] ]
+# gen_params = [ GeneratorParams(rep, (5e-3, 1, 5), background)
+# for rep in range(1)
+# for background in ['meas://fmeas_export_ocxo_2day.bin'] ]
+
+ data_db = sqlite3.connect(path.join(args.cachedir, 'data_db.sqlite3'))
+ data_db.execute('CREATE TABLE IF NOT EXISTS waveforms'
+ '(seed, amplitude_spec, background, nbits, decimation, symbols, thresholds, result, timestamp)')
+
+ 'SELECT FROM waveforms GROUP BY (amplitude_spec, background, nbits, decimation, symbols, thresholds, result)'
+
+ dec_param_groups = defaultdict(lambda: [])
+ for nbits, thf, decimation, symbols in dec_paramses:
+ dec_param_groups[(nbits, decimation, symbols)].append(thf)
+ waveform_params = [ (*gp, *dp, thfs) for gp in gen_params for dp, thfs in dec_param_groups.items() ]
+ print(f'Generated {len(waveform_params)} parameter sets')
+
+ # Separate out our batch
+ waveform_params = waveform_params[args.index::args.batches]
+
+ def lookup_binary(*params):
+ return path.join(build_cache_dir, builds[tuple(params)], 'tools/dsss_demod_test')
+
+ def params_mapper(seed, amplitude_spec, background, nbits, decimation, symbols, thresholds):
+ amplitude_spec = ','.join(str(x) for x in amplitude_spec)
+ thresholds = ','.join(str(x) for x in thresholds)
+ return seed, amplitude_spec, background, nbits, decimation, symbols, thresholds
+
+ results = []
+ for _params, chunk in parallel_generator(data_db, 'waveforms',
+ ['seed', 'amplitude_spec', 'background', 'nbits', 'decimation', 'symbols', 'thresholds'],
+ params_mapper=params_mapper,
+ builder=run_test,
+ param_list=waveform_params, desc='Simulating demodulation',
+ context=dict(cachedir=data_cache_dir, lookup_binary=lookup_binary),
+ disable_cache=args.no_cache):
+ results += chunk
+
+ if args.dump:
+ with open(args.dump, 'w') as f:
+ json.dump(results, f)
+
diff --git a/tools/dsss_demod_test_waveform_gen.py b/tools/dsss_demod_test_waveform_gen.py
new file mode 100644
index 0000000..414c553
--- /dev/null
+++ b/tools/dsss_demod_test_waveform_gen.py
@@ -0,0 +1,86 @@
+
+from os import path
+import json
+import functools
+
+import numpy as np
+import numbers
+import math
+from scipy import signal as sig
+import scipy.fftpack
+
+sampling_rate = 10 # sp/s
+
+# From https://github.com/mubeta06/python/blob/master/signal_processing/sp/gold.py
+preferred_pairs = {5:[[2],[1,2,3]], 6:[[5],[1,4,5]], 7:[[4],[4,5,6]],
+ 8:[[1,2,3,6,7],[1,2,7]], 9:[[5],[3,5,6]],
+ 10:[[2,5,9],[3,4,6,8,9]], 11:[[9],[3,6,9]]}
+
+def gen_gold(seq1, seq2):
+ gold = [seq1, seq2]
+ for shift in range(len(seq1)):
+ gold.append(seq1 ^ np.roll(seq2, -shift))
+ return gold
+
+def gold(n):
+ n = int(n)
+ if not n in preferred_pairs:
+ raise KeyError('preferred pairs for %s bits unknown' % str(n))
+ t0, t1 = preferred_pairs[n]
+ (seq0, _st0), (seq1, _st1) = sig.max_len_seq(n, taps=t0), sig.max_len_seq(n, taps=t1)
+ return gen_gold(seq0, seq1)
+
+def modulate(data, nbits=5):
+ # 0, 1 -> -1, 1
+ mask = np.array(gold(nbits))*2 - 1
+
+ sel = mask[data>>1]
+ data_lsb_centered = ((data&1)*2 - 1)
+
+ signal = (np.multiply(sel, np.tile(data_lsb_centered, (2**nbits-1, 1)).T).flatten() + 1) // 2
+ return np.hstack([ np.zeros(len(mask)), signal, np.zeros(len(mask)) ])
+
+def load_noise_meas_params(capture_file):
+ with open(capture_file, 'rb') as f:
+ meas_data = np.copy(np.frombuffer(f.read(), dtype='float32'))
+ meas_data -= np.mean(meas_data)
+ return (meas_data,)
+
+def mains_noise_measured(seed, n, meas_data):
+ last_valid = len(meas_data) - n
+ st = np.random.RandomState(seed)
+ start = st.randint(last_valid)
+ return meas_data[start:start+n] + 50.00
+
+def load_noise_synth_params(specfile):
+ with open(specfile) as f:
+ d = json.load(f)
+ return {'spl_x': np.linspace(*d['x_spec']),
+ 'spl_N': d['x_spec'][2],
+ 'psd_spl': (d['t'], d['c'], d['k']) }
+
+def mains_noise_synthetic(seed, n, psd_spl, spl_N, spl_x):
+ st = np.random.RandomState(seed)
+ noise = st.normal(size=spl_N) * 2
+ spec = scipy.fftpack.fft(noise) **2
+
+ spec *= np.exp(scipy.interpolate.splev(spl_x, psd_spl))
+
+ spec **= 1/2
+
+ renoise = scipy.fftpack.ifft(spec)
+ return renoise[10000:][:n] + 50.00
+
+@functools.lru_cache()
+def load_noise_gen(url):
+ schema, refpath = url.split('://')
+ if not path.isabs(refpath):
+ refpath = path.abspath(path.join(path.dirname(__file__), refpath))
+
+ if schema == 'meas':
+ return mains_noise_measured, load_noise_meas_params(refpath)
+ elif schema == 'synth':
+ return mains_noise_synthetic, load_noise_synth_params(refpath)
+ else:
+ raise ValueError('Invalid schema', schema)
+
diff --git a/tools/e2e_test.c b/tools/e2e_test.c
new file mode 100644
index 0000000..935f70d
--- /dev/null
+++ b/tools/e2e_test.c
@@ -0,0 +1,111 @@
+
+#include <stdint.h>
+#include <math.h>
+#include <unistd.h>
+#include <stdio.h>
+#include <string.h>
+#include <errno.h>
+#include <stdlib.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <sys/fcntl.h>
+
+#include "freq_meas.h"
+#include "dsss_demod.h"
+
+typedef uint16_t adc_data_t;
+
+void handle_dsss_received(uint8_t data[static TRANSMISSION_SYMBOLS]) {
+ printf("data sequence received: [ ");
+ for (size_t i=0; i<TRANSMISSION_SYMBOLS; i++) {
+ printf("%+3d", ((data[i]&1) ? 1 : -1) * (data[i]>>1));
+ if (i+1 < TRANSMISSION_SYMBOLS)
+ printf(", ");
+ }
+ printf(" ]\n");
+}
+
+void print_usage(void);
+void print_usage() {
+ fprintf(stderr, "Usage: e2e_test [emulated_adc_data.bin]\n");
+}
+
+int main(int argc, char **argv) {
+ if (argc != 2) {
+ fprintf(stderr, "Error: Invalid arguments.\n");
+ print_usage();
+ return 1;
+ }
+
+ int fd = open(argv[1], O_RDONLY);
+ struct stat st;
+ if (fstat(fd, &st)) {
+ fprintf(stderr, "Error querying test data file size: %s\n", strerror(errno));
+ return 2;
+ }
+
+ if (st.st_size < 0 || st.st_size > 100000000) {
+ fprintf(stderr, "Error reading test data: too much test data (size=%zd)\n", st.st_size);
+ return 2;
+ }
+
+ if (st.st_size % sizeof(adc_data_t) != 0) {
+ fprintf(stderr, "Error reading test data: file size is not divisible by %zd (size=%zd)\n", sizeof(adc_data_t), st.st_size);
+ return 2;
+ }
+
+ char *buf = malloc(st.st_size);
+ if (!buf) {
+ fprintf(stderr, "Error allocating memory");
+ return 2;
+ }
+
+ const size_t n_samples = st.st_size / sizeof(adc_data_t);
+ fprintf(stderr, "Reading %zd samples test data...", n_samples);
+ ssize_t nread = 0;
+ while (nread < st.st_size) {
+ ssize_t rc = read(fd, buf + nread, st.st_size - nread);
+
+ if (rc == -EINTR || rc == -EAGAIN)
+ continue;
+
+ if (rc < 0) {
+ fprintf(stderr, "\nError reading test data: %s\n", strerror(errno));
+ return 2;
+ }
+
+ if (rc == 0) {
+ fprintf(stderr, "\nError reading test data: Unexpected end of file\n");
+ return 2;
+ }
+
+ nread += rc;
+ }
+ fprintf(stderr, " done. Read %zd bytes.\n", nread);
+
+ adc_data_t *buf_d = (adc_data_t *)buf;
+
+ struct dsss_demod_state demod;
+ dsss_demod_init(&demod);
+
+ fprintf(stderr, "Starting simulation.\n");
+ size_t iterations = (n_samples-FMEAS_FFT_LEN)/(FMEAS_FFT_LEN/2);
+ for (size_t i=0; i<iterations; i++) {
+
+ /*
+ fprintf(stderr, "Iteration %zd/%zd\n", i, iterations);
+ */
+ float res = NAN;
+ int rc = adc_buf_measure_freq(buf_d + i*(FMEAS_FFT_LEN/2), &res);
+ if (rc)
+ printf("ERROR: Simulation error in iteration %zd at position %zd: %d\n", i, i*(FMEAS_FFT_LEN/2), rc);
+
+ dsss_demod_step(&demod, res, i);
+ /*
+ printf("%09zd %12f\n", i, res);
+ */
+ }
+
+ free(buf);
+ return 0;
+}
diff --git a/tools/fft_window_header_gen.py b/tools/fft_window_header_gen.py
new file mode 100644
index 0000000..7df2ee3
--- /dev/null
+++ b/tools/fft_window_header_gen.py
@@ -0,0 +1,59 @@
+#!/usr/bin/env python3
+
+import textwrap
+
+import scipy.signal as sig
+import numpy as np
+
+WINDOW_TYPES = [
+ 'boxcar',
+ 'triang',
+ 'blackman',
+ 'hamming',
+ 'hann',
+ 'bartlett',
+ 'flattop',
+ 'parzen',
+ 'bohman',
+ 'blackmanharris',
+ 'nuttall',
+ 'barthann',
+ 'kaiser',
+ 'gaussian',
+ 'general_gaussian',
+ 'slepian',
+ 'dpss',
+ 'chebwin',
+ 'exponential',
+ 'tukey',
+ ]
+
+if __name__ == '__main__':
+ import argparse
+ parser = argparse.ArgumentParser()
+ parser.add_argument('window', choices=WINDOW_TYPES, help='Type of window function to use')
+ parser.add_argument('n', type=int, help='Width of window in samples')
+ parser.add_argument('window_args', nargs='*', type=float,
+ help='''Window argument(s) if required. See https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.get_window.html#scipy.signal.get_window for details.''')
+ parser.add_argument('-v', '--variable', default='fft_window_table', help='Name for alias variable pointing to generated window')
+ args = parser.parse_args()
+
+ print(f'/* FTT window table for {args.n} sample {args.window} window.')
+ if args.window_args:
+ print(f' * Window arguments were: ({" ,".join(str(arg) for arg in args.window_args)})')
+ print(f' */')
+ winargs = ''.join(f'_{arg:.4g}'.replace('.', 'F') for arg in args.window_args)
+ varname = f'fft_{args.n}_window_{args.window}{winargs}'
+ print(f'const float {varname}[{args.n}] = {{')
+
+ win = sig.get_window(args.window if not args.window_args else (args.window, *args.window_args),
+ Nx=args.n, fftbins=True)
+ par = ' '.join(f'{f:>013.8g},' for f in win)
+ print(textwrap.fill(par,
+ initial_indent=' '*4, subsequent_indent=' '*4,
+ width=120,
+ replace_whitespace=False, drop_whitespace=False))
+ print('};')
+ print()
+ print(f'const float * const {args.variable} __attribute__((weak)) = {varname};')
+
diff --git a/tools/fmeas_export_ocxo_2day.bin b/tools/fmeas_export_ocxo_2day.bin
new file mode 100644
index 0000000..c0cd8a8
--- /dev/null
+++ b/tools/fmeas_export_ocxo_2day.bin
Binary files differ
diff --git a/tools/freq_meas_test.c b/tools/freq_meas_test.c
new file mode 100644
index 0000000..e2900ad
--- /dev/null
+++ b/tools/freq_meas_test.c
@@ -0,0 +1,106 @@
+
+#include <stdint.h>
+#include <math.h>
+#include <unistd.h>
+#include <stdio.h>
+#include <string.h>
+#include <errno.h>
+#include <stdlib.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <sys/fcntl.h>
+
+#include "freq_meas.h"
+
+void print_usage(void);
+
+void print_usage() {
+ fprintf(stderr, "Usage: freq_meas_test [test_data.bin]\n");
+}
+
+int main(int argc, char **argv) {
+ if (argc != 2) {
+ fprintf(stderr, "Error: Invalid arguments.\n");
+ print_usage();
+ return 1;
+ }
+
+ int fd = open(argv[1], O_RDONLY);
+ struct stat st;
+ if (fstat(fd, &st)) {
+ fprintf(stderr, "Error querying test data file size: %s\n", strerror(errno));
+ return 2;
+ }
+
+ if (st.st_size < 0 || st.st_size > 1000000) {
+ fprintf(stderr, "Error reading test data: too much test data (size=%zd)\n", st.st_size);
+ return 2;
+ }
+
+ if (st.st_size % sizeof(float) != 0) {
+ fprintf(stderr, "Error reading test data: file size is not divisible by %zd (size=%zd)\n", sizeof(float), st.st_size);
+ return 2;
+ }
+
+ char *buf = malloc(st.st_size);
+ if (!buf) {
+ fprintf(stderr, "Error allocating memory");
+ return 2;
+ }
+
+ fprintf(stderr, "Reading %zd samples test data...", st.st_size/sizeof(float));
+ ssize_t nread = 0;
+ while (nread < st.st_size) {
+ ssize_t rc = read(fd, buf + nread, st.st_size - nread);
+
+ if (rc == -EINTR || rc == -EAGAIN)
+ continue;
+
+ if (rc < 0) {
+ fprintf(stderr, "\nError reading test data: %s\n", strerror(errno));
+ return 2;
+ }
+
+ if (rc == 0) {
+ fprintf(stderr, "\nError reading test data: Unexpected end of file\n");
+ return 2;
+ }
+
+ nread += rc;
+ }
+ fprintf(stderr, " done.\n");
+
+ const size_t n_samples = st.st_size / sizeof(float);
+ float *buf_f = (float *)buf;
+
+ int16_t *sim_adc_buf = calloc(sizeof(int16_t), n_samples);
+ if (!sim_adc_buf) {
+ fprintf(stderr, "Error allocating memory\n");
+ return 2;
+ }
+
+ fprintf(stderr, "Converting and truncating test data...");
+ for (size_t i=0; i<n_samples; i++)
+ /* Note on scaling: We can't simply scale by 0x8000 (1/2 full range) here. Our test data is nominally 1Vp-p but
+ * certain tests such as the interharmonics one can have some samples exceeding that range. */
+ sim_adc_buf[i] = buf_f[i] * (0x4000-1);
+ fprintf(stderr, " done.\n");
+
+ fprintf(stderr, "Starting simulation.\n");
+
+ size_t iterations = (n_samples-FMEAS_FFT_LEN)/(FMEAS_FFT_LEN/2);
+ for (size_t i=0; i<iterations; i++) {
+
+ fprintf(stderr, "Iteration %zd/%zd\n", i, iterations);
+ float res = NAN;
+ int rc = adc_buf_measure_freq(sim_adc_buf + i*(FMEAS_FFT_LEN/2), &res);
+ if (rc)
+ printf("ERROR: Simulation error in iteration %zd at position %zd: %d\n", i, i*(FMEAS_FFT_LEN/2), rc);
+
+ printf("%09zd %12f\n", i, res);
+ }
+
+ free(buf);
+ free(sim_adc_buf);
+ return 0;
+}
diff --git a/tools/freq_meas_test_runner.py b/tools/freq_meas_test_runner.py
new file mode 100644
index 0000000..779922a
--- /dev/null
+++ b/tools/freq_meas_test_runner.py
@@ -0,0 +1,39 @@
+#!/usr/bin/env python3
+
+import os
+from os import path
+import subprocess
+import json
+
+import numpy as np
+np.set_printoptions(linewidth=240)
+
+
+if __name__ == '__main__':
+ import argparse
+ parser = argparse.ArgumentParser()
+ parser.add_argument(metavar='test_data_directory', dest='dir', help='Directory with test data .bin files')
+ default_binary = path.abspath(path.join(path.dirname(__file__), '../build/tools/freq_meas_test'))
+ parser.add_argument(metavar='test_binary', dest='binary', nargs='?', default=default_binary)
+ parser.add_argument('-d', '--dump', help='Write raw measurements to JSON file')
+ args = parser.parse_args()
+
+ bin_files = [ path.join(args.dir, d) for d in os.listdir(args.dir) if d.lower().endswith('.bin') ]
+
+ savedata = {}
+ for p in bin_files:
+ output = subprocess.check_output([args.binary, p], stderr=subprocess.DEVNULL)
+ measurements = np.array([ float(value) for _offset, value in [ line.split() for line in output.splitlines() ] ])
+ savedata[p] = list(measurements)
+
+ # Cut off first and last sample for mean and RMS calculations as these show boundary effects.
+ measurements = measurements[1:-1]
+ mean = np.mean(measurements)
+ rms = np.sqrt(np.mean(np.square(measurements - mean)))
+
+ print(f'{path.basename(p):<60}: mean={mean:<8.4f}Hz rms={rms*1000:.3f}mHz')
+
+ if args.dump:
+ with open(args.dump, 'w') as f:
+ json.dump(savedata, f)
+
diff --git a/tools/gold_code_header_gen.py b/tools/gold_code_header_gen.py
new file mode 100644
index 0000000..fa98fce
--- /dev/null
+++ b/tools/gold_code_header_gen.py
@@ -0,0 +1,70 @@
+#!/usr/bin/env python3
+
+import sys
+import math
+import textwrap
+import contextlib
+
+import numpy as np
+import scipy.signal as sig
+
+# From https://github.com/mubeta06/python/blob/master/signal_processing/sp/gold.py
+preferred_pairs = {5:[[2],[1,2,3]], 6:[[5],[1,4,5]], 7:[[4],[4,5,6]],
+ 8:[[1,2,3,6,7],[1,2,7]], 9:[[5],[3,5,6]],
+ 10:[[2,5,9],[3,4,6,8,9]], 11:[[9],[3,6,9]]}
+
+def gen_gold(seq1, seq2):
+ gold = [seq1, seq2]
+ for shift in range(len(seq1)):
+ gold.append(seq1 ^ np.roll(seq2, -shift))
+ return gold
+
+def gold(n):
+ n = int(n)
+ if not n in preferred_pairs:
+ raise KeyError('preferred pairs for %s bits unknown' % str(n))
+ t0, t1 = preferred_pairs[n]
+ (seq0, _st0), (seq1, _st1) = sig.max_len_seq(n, taps=t0), sig.max_len_seq(n, taps=t1)
+ return gen_gold(seq0, seq1)
+
+@contextlib.contextmanager
+def print_include_guards(macro_name):
+ print(f'#ifndef {macro_name}')
+ print(f'#define {macro_name}')
+ yield
+ print(f'#endif /* {macro_name} */')
+
+if __name__ == '__main__':
+ import argparse
+ parser = argparse.ArgumentParser(add_help=False)
+ parser.add_argument('n', type=int, choices=preferred_pairs, help='bit width of shift register. Generate 2**n + 1 sequences of length 2**n - 1.')
+ parser.add_argument('-v', '--variable', default='gold_code_table', help='Name for weak alias of generated table')
+ parser.add_argument('-h', '--header', action='store_true', help='Generate header file')
+ parser.add_argument('-c', '--source', action='store_true', help='Generate table source file')
+ args = parser.parse_args()
+
+ if not args.header != args.source:
+ print('Exactly one of --header and --source must be given.', file=sys.stderr)
+ sys.exit(1)
+
+ nbytes = math.ceil((2**args.n-1)/8)
+
+ if args.source:
+ print('/* THIS IS A GENERATED FILE. DO NOT EDIT! */')
+ print('#include <unistd.h>')
+ print('#include <stdint.h>')
+ print()
+ print(f'/* {args.n} bit gold sequences: {2**args.n+1} sequences of length {2**args.n-1} bit.')
+ print(f' *')
+ print(f' * Each code is packed left-aligned into {nbytes} bytes in big-endian byte order.')
+ print(f' */')
+ print(f'const uint8_t {args.variable}[{2**args.n+1}][{nbytes}] = {{')
+ for i, code in enumerate(gold(args.n)):
+ par = '{' + ' '.join(f'0x{d:02x},' for d in np.packbits(code)) + f'}}, /* {i: 3d} "{"".join(str(x) for x in code)}" */'
+ print(textwrap.fill(par, initial_indent=' '*4, subsequent_indent=' '*4, width=120))
+ print('};')
+ print()
+ else:
+ print('/* THIS IS A GENERATED FILE. DO NOT EDIT! */')
+ with print_include_guards(f'__GOLD_CODE_GENERATED_HEADER_{args.n}__'):
+ print(f'extern const uint8_t {args.variable}[{2**args.n+1}][{nbytes}];')
diff --git a/tools/grid_freq_psd_spl_108pt.json b/tools/grid_freq_psd_spl_108pt.json
new file mode 100644
index 0000000..5a0ff41
--- /dev/null
+++ b/tools/grid_freq_psd_spl_108pt.json
@@ -0,0 +1 @@
+{"x_spec": [3.2595692805152726e-05, 5.0, 613575], "t": [3.2595692805152726e-05, 3.2595692805152726e-05, 3.2595692805152726e-05, 3.2595692805152726e-05, 0.0001423024947075771, 0.00015800362803968106, 0.00017543716661470822, 0.00019479425764873777, 0.0002162871388378975, 0.00024015146540428407, 0.00026664889389955537, 0.00029606995109590574, 0.00032873721941990017, 0.0003650088738553592, 0.0004052826090950758, 0.00045000000000000004, 0.000499651343175437, 0.0005547810327489297, 0.0006159935292916862, 0.0006839599873288199, 0.0007594256141046668, 0.0008432178402871724, 0.0009362553921977272, 0.0010395583650374223, 0.0011542594075560205, 0.001281616140796111, 0.0014230249470757708, 0.001580036280396809, 0.0017543716661470824, 0.0019479425764873776, 0.002162871388378975, 0.0024015146540428403, 0.002666488938995554, 0.002960699510959057, 0.0032873721941990056, 0.0036500887385535925, 0.004052826090950754, 0.0045000000000000005, 0.00499651343175437, 0.005547810327489296, 0.006159935292916869, 0.0068395998732882, 0.007594256141046669, 0.008432178402871724, 0.009362553921977271, 0.010395583650374221, 0.011542594075560205, 0.012816161407961109, 0.014230249470757707, 0.01580036280396809, 0.017543716661470823, 0.01947942576487376, 0.02162871388378975, 0.024015146540428405, 0.026664889389955565, 0.02960699510959057, 0.03287372194199005, 0.036500887385535925, 0.04052826090950754, 0.045, 0.0499651343175437, 0.05547810327489296, 0.06159935292916863, 0.06839599873288206, 0.07594256141046668, 0.08432178402871732, 0.09362553921977272, 0.10395583650374222, 0.11542594075560206, 0.12816161407961107, 0.14230249470757705, 0.15800362803968088, 0.1754371666147082, 0.1947942576487376, 0.21628713883789774, 0.24015146540428406, 0.26664889389955565, 0.2960699510959057, 0.32873721941990053, 0.36500887385535924, 0.40528260909507535, 0.45, 0.499651343175437, 0.5547810327489296, 0.6159935292916868, 0.6839599873288206, 0.7594256141046669, 0.8432178402871732, 0.9362553921977271, 1.0395583650374223, 1.1542594075560206, 1.2816161407961109, 1.4230249470757708, 1.5800362803968104, 1.7543716661470823, 1.9479425764873777, 2.162871388378975, 2.4015146540428405, 2.6664889389955535, 2.960699510959057, 3.287372194199002, 3.6500887385535927, 4.052826090950758, 4.5, 5.0, 5.0, 5.0, 5.0], "c": [0.7720161468716866, -0.5547528253056444, 0.30706059086000753, 0.19422577014134906, -1.1954636661840032, 0.9215976941641111, -0.6668136393976918, -1.341269161156733, -0.16311330594842666, -1.7639636752234251, -1.238385544822954, -0.32649555618555554, -0.03086589610280171, -2.358195657381619, -0.5759152419849985, 0.1892225800004134, -1.8122889670546236, -0.8109120798216202, -0.5500991736738969, -4.680192969256771, -2.8007700704649876, 0.16866469558571784, -1.1040811840849307, -3.0243574268705546, -4.018139927365795, -4.100581028618109, -0.556354762846191, -7.414377514669229, 1.36396325920194, -6.002559557058508, -2.2113451390305365, -4.578944771104116, -4.372644849632638, -3.945339124673235, -4.778747958903158, -2.370174137632325, -5.7372466088109295, -4.707506574819875, -4.834404729330929, -5.005244244061701, -5.82644896783577, -4.717966026411524, -6.146374820241562, -4.972788381244952, -5.854957092953355, -5.702174935205885, -6.222035857079607, -6.2128389666872, -6.212821706753751, -6.253599689326325, -6.681685577659057, -6.372364384360678, -6.771223202540934, -6.856809137231159, -6.986412256164045, -7.190466178818742, -7.577896455149433, -7.515731696006047, -7.598155006351761, -7.824526916149126, -8.141496591776512, -8.36794927682997, -8.80307396767114, -8.828816533544659, -9.357524260470413, -9.658130054343863, -10.005768472049466, -10.499801262514108, -11.028689820560558, -11.413688641742898, -11.906162042727946, -12.232342460719975, -12.438432746733596, -13.088338100203112, -12.308710772618745, -11.685074853925329, -11.397838681243094, -12.265219694936695, -13.600359694898529, -14.031425961884718, -12.236885080485473, -13.527508426900974, -13.698402018452601, -13.397911198962568, -14.144410560196603, -13.905769594095293, -14.410874830544122, -14.531727635304264, -14.59275291853806, -14.35404826562502, -14.58670053318149, -14.432515268864977, -14.363428024828353, -14.429222027493264, -14.73947634127499, -14.717315405960353, -14.678539669792505, -14.825278423641382, -14.80936417940876, -14.943375264882789, -14.680885181815674, -14.54841244844906, -14.634365225950589, -14.609444790868906, 0.0, 0.0, 0.0, 0.0], "k": 3} \ No newline at end of file
diff --git a/tools/hum_generator.py b/tools/hum_generator.py
new file mode 100644
index 0000000..fbcabac
--- /dev/null
+++ b/tools/hum_generator.py
@@ -0,0 +1,111 @@
+#!/usr/bin/env python
+# coding: utf-8
+
+import binascii
+import struct
+
+import numpy as np
+import pydub
+
+from dsss_demod_test_waveform_gen import load_noise_gen, modulate as dsss_modulate
+
+np.set_printoptions(linewidth=240)
+
+def generate_noisy_signal(
+ test_data=32,
+ test_nbits=5,
+ test_decimation=10,
+ test_signal_amplitude=20e-3,
+ noise_level=10e-3,
+ noise_spec='synth://grid_freq_psd_spl_108pt.json',
+ seed=0):
+
+ #test_data = np.random.RandomState(seed=0).randint(0, 2 * (2**test_nbits), test_duration)
+ #test_data = np.array([0, 1, 2, 3] * 50)
+ if isinstance(test_data, int):
+ test_data = np.array(range(test_data))
+
+
+ signal = np.repeat(dsss_modulate(test_data, test_nbits) * 2.0 - 1, test_decimation)
+
+ noise_gen, noise_params = load_noise_gen(noise_spec)
+ noise = noise_gen(seed, len(signal), **noise_params)
+ return np.absolute(noise + signal*test_signal_amplitude)
+
+def write_raw_frequencies_bin(outfile, **kwargs):
+ with open(outfile, 'wb') as f:
+ for x in generate_noisy_signal(**kwargs):
+ f.write(struct.pack('f', x))
+
+def synthesize_sine(freqs, freqs_sampling_rate=10.0, output_sampling_rate=44100):
+ duration = len(freqs) / freqs_sampling_rate # seconds
+ afreq_out = np.interp(np.linspace(0, duration, int(duration*output_sampling_rate)), np.linspace(0, duration, len(freqs)), freqs)
+ return np.sin(np.cumsum(2*np.pi * afreq_out / output_sampling_rate))
+
+def write_flac(filename, signal, sampling_rate=44100):
+ signal -= np.min(signal)
+ signal /= np.max(signal)
+ signal -= 0.5
+ signal *= 2**16 - 1
+ le_bytes = signal.astype(np.int16).tobytes()
+ seg = pydub.AudioSegment(data=le_bytes, sample_width=2, frame_rate=sampling_rate, channels=1)
+ seg.export(filename, format='flac')
+
+def write_synthetic_hum_flac(filename, output_sampling_rate=44100, freqs_sampling_rate=10.0, **kwargs):
+ signal = generate_noisy_signal(**kwargs)
+ print(signal)
+ write_flac(filename, synthesize_sine(signal, freqs_sampling_rate, output_sampling_rate),
+ sampling_rate=output_sampling_rate)
+
+def emulate_adc_signal(adc_bits=12, adc_offset=0.4, adc_amplitude=0.25, freq_sampling_rate=10.0, output_sampling_rate=1000, **kwargs):
+ signal = synthesize_sine(generate_noisy_signal(), freq_sampling_rate, output_sampling_rate)
+ signal = signal*adc_amplitude + adc_offset
+ smin, smax = np.min(signal), np.max(signal)
+ if smin < 0.0 or smax > 1.0:
+ raise UserWarning('Amplitude or offset too large: Signal out of bounds with min/max [{smin}, {smax}] of ADC range')
+ signal *= 2**adc_bits -1
+ return signal
+
+def save_adc_signal(fn, signal, dtype=np.uint16):
+ with open(fn, 'wb') as f:
+ f.write(signal.astype(dtype).tobytes())
+
+def write_emulated_adc_signal_bin(filename, **kwargs):
+ save_adc_signal(filename, emulate_adc_signal(**kwargs))
+
+def hum_cmd(args):
+ write_synthetic_hum_flac(args.out_flac,
+ output_sampling_rate=args.audio_sampling_rate,
+ freqs_sampling_rate=args.frequency_sampling_rate,
+ test_data = np.array(list(binascii.unhexlify(args.data))),
+ test_nbits = args.symbol_bits,
+ test_decimation = args.decimation,
+ test_signal_amplitude = args.signal_level/1e3,
+ noise_level = args.noise_level/1e3,
+ noise_spec=args.noise_spec,
+ seed = args.random_seed)
+
+
+if __name__ == '__main__':
+ import argparse
+ parser = argparse.ArgumentParser()
+ cmd_parser = parser.add_subparsers(required=True)
+ hum_parser = cmd_parser.add_parser('hum', help='Generated artificial modulated mains hum')
+ # output parameters
+ hum_parser.add_argument('-a', '--audio-sampling-rate', type=int, default=44100)
+
+ # modulation parameters
+ hum_parser.add_argument('-f', '--frequency-sampling-rate', type=float, default=10.0*100/128)
+ hum_parser.add_argument('-b', '--symbol-bits', type=int, default=5, help='bits per symbol (excluding sign bit)')
+ hum_parser.add_argument('-n', '--noise-level', type=float, default=1.0, help='Scale synthetic noise level')
+ hum_parser.add_argument('-s', '--signal-level', type=float, default=20.0, help='Synthetic noise level in mHz')
+ hum_parser.add_argument('-d', '--decimation', type=int, default=10, help='DSSS modulation decimation in frequency measurement cycles')
+ hum_parser.add_argument('-r', '--random-seed', type=int, default=0)
+ hum_parser.add_argument('--noise-spec', type=str, default='synth://grid_freq_psd_spl_108pt.json')
+ hum_parser.add_argument('out_flac', metavar='out.flac', help='FLAC output file')
+ hum_parser.add_argument('data', help='modulation data hex string')
+ hum_parser.set_defaults(func=hum_cmd)
+
+ args = parser.parse_args()
+ args.func(args)
+
diff --git a/tools/ldparser.py b/tools/ldparser.py
new file mode 100644
index 0000000..c620fe2
--- /dev/null
+++ b/tools/ldparser.py
@@ -0,0 +1,126 @@
+
+import sys
+
+import pyparsing as pp
+from pyparsing import pyparsing_common as ppc
+
+LPAREN, RPAREN, LBRACE, RBRACE, LBROK, RBROK, COLON, SEMICOLON, EQUALS, COMMA = map(pp.Suppress, '(){}<>:;=,')
+
+parse_suffix_int = lambda lit: int(lit[:-1]) * (10**(3*(1 + 'kmgtpe'.find(lit[-1].lower()))))
+si_suffix = pp.oneOf('k m g t p e', caseless=True)
+
+numeric_literal = pp.Regex('0x[0-9a-fA-F]+').setName('hex int').setParseAction(pp.tokenMap(int, 16)) \
+ | (pp.Regex('[0-9]+[kKmMgGtTpPeE]')).setName('size int').setParseAction(pp.tokenMap(parse_suffix_int)) \
+ | pp.Word(pp.nums).setName('int').setParseAction(pp.tokenMap(int))
+access_def = pp.Regex('[rR]?[wW]?[xX]?').setName('access literal').setParseAction(pp.tokenMap(str.lower))
+
+origin_expr = pp.Suppress(pp.CaselessKeyword('ORIGIN')) + EQUALS + numeric_literal
+length_expr = pp.Suppress(pp.CaselessKeyword('LENGTH')) + EQUALS + numeric_literal
+mem_expr = pp.Group(ppc.identifier + LPAREN + access_def + RPAREN + COLON + origin_expr + COMMA + length_expr)
+mem_contents = pp.ZeroOrMore(mem_expr)
+
+mem_toplevel = pp.CaselessKeyword("MEMORY") + pp.Group(LBRACE + pp.Optional(mem_contents, []) + RBRACE)
+
+glob = pp.Word(pp.alphanums + '._*')
+match_expr = pp.Forward()
+assignment = pp.Forward()
+funccall = pp.Group(pp.Word(pp.alphas + '_') + LPAREN + (assignment | numeric_literal | match_expr | glob | ppc.identifier) + RPAREN + pp.Optional(SEMICOLON))
+value = numeric_literal | funccall | ppc.identifier | '.'
+formula = (value + pp.oneOf('+ = * / %') + value) | value
+# suppress stray semicolons
+assignment << (SEMICOLON | pp.Group((ppc.identifier | '.') + EQUALS + (formula | value) + pp.Optional(SEMICOLON)))
+match_expr << (glob + LPAREN + pp.OneOrMore(funccall | glob) + RPAREN)
+
+section_contents = pp.ZeroOrMore(assignment | funccall | match_expr);
+
+section_name = pp.Regex('\.[a-zA-Z0-9_.]+')
+section_def = pp.Group(section_name + pp.Optional(numeric_literal) + COLON + LBRACE + pp.Group(section_contents) +
+ RBRACE + pp.Optional(RBROK + ppc.identifier + pp.Optional('AT' + RBROK + ppc.identifier)))
+sec_contents = pp.ZeroOrMore(section_def | assignment)
+
+sections_toplevel = pp.Group(pp.CaselessKeyword("SECTIONS").suppress() + LBRACE + sec_contents + RBRACE)
+
+toplevel_elements = mem_toplevel | funccall | sections_toplevel | assignment
+ldscript = pp.Group(pp.ZeroOrMore(toplevel_elements))
+ldscript.ignore(pp.cppStyleComment)
+
+if __name__ == '__main__':
+ import argparse
+ parser = argparse.ArgumentParser()
+ parser.add_argument('linker_script', type=argparse.FileType('r'))
+ args = parser.parse_args()
+
+ #print(mem_expr.parseString('FLASH (rx) : ORIGIN = 0x0800000, LENGTH = 512K', parseAll=True))
+ # print(ldscript.parseString('''
+ # /* Entry Point */
+ # ENTRY(Reset_Handler)
+ #
+ # /* Highest address of the user mode stack */
+ # _estack = 0x20020000; /* end of RAM */
+ # /* Generate a link error if heap and stack don't fit into RAM */
+ # _Min_Heap_Size = 0x200;; /* required amount of heap */
+ # _Min_Stack_Size = 0x400;; /* required amount of stack */
+ # ''', parseAll=True))
+
+ print(ldscript.parseFile(args.linker_script, parseAll=True))
+ #print(funccall.parseString('KEEP(*(.isr_vector))'))
+ #print(section_contents.parseString('''
+ # . = ALIGN(4);
+ # KEEP(*(.isr_vector)) /* Startup code */
+ # . = ALIGN(4);
+ # ''', parseAll=True))
+
+ #print(section_def.parseString('''
+ # .text :
+ # {
+ # . = ALIGN(4);
+ # *(.text) /* .text sections (code) */
+ # *(.text*) /* .text* sections (code) */
+ # *(.glue_7) /* glue arm to thumb code */
+ # *(.glue_7t) /* glue thumb to arm code */
+ # *(.eh_frame)
+ #
+ # KEEP (*(.init))
+ # KEEP (*(.fini))
+ #
+ # . = ALIGN(4);
+ # _etext = .; /* define a global symbols at end of code */
+ # } >FLASH
+ # ''', parseAll=True))
+
+ #print(section_def.parseString('.ARM.extab : { *(.ARM.extab* .gnu.linkonce.armextab.*) } >FLASH', parseAll=True))
+
+ #print(assignment.parseString('__preinit_array_start = .', parseAll=True))
+ #print(assignment.parseString('a = 23', parseAll=True))
+ #print(funccall.parseString('foo (a=23)', parseAll=True))
+ #print(funccall.parseString('PROVIDE_HIDDEN (__preinit_array_start = .);', parseAll=True))
+ #print(section_def.parseString('''
+ # .preinit_array :
+ # {
+ # PROVIDE_HIDDEN (__preinit_array_start = .);
+ # KEEP (*(.preinit_array*))
+ # PROVIDE_HIDDEN (__preinit_array_end = .);
+ # } >FLASH''', parseAll=True))
+ #print(match_expr.parseString('*(SORT(.init_array.*))', parseAll=True))
+ #print(funccall.parseString('KEEP (*(SORT(.init_array.*)))', parseAll=True))
+ #print(section_def.parseString('''
+ # .init_array :
+ # {
+ # PROVIDE_HIDDEN (__init_array_start = .);
+ # KEEP (*(SORT(.init_array.*)))
+ # KEEP (*(.init_array*))
+ # PROVIDE_HIDDEN (__init_array_end = .);
+ # } >FLASH
+ # ''', parseAll=True))
+
+ #print(match_expr.parseString('*(.ARM.extab* .gnu.linkonce.armextab.*)', parseAll=True))
+ #print(formula.parseString('. + _Min_Heap_Size', parseAll=True))
+ #print(assignment.parseString('. = . + _Min_Heap_Size;', parseAll=True))
+ #print(sections_toplevel.parseString('''
+ # SECTIONS
+ # {
+ # .ARMattributes : { }
+ # }
+ # ''', parseAll=True))
+ #sys.exit(0)
+
diff --git a/tools/linkmem.py b/tools/linkmem.py
new file mode 100644
index 0000000..934a571
--- /dev/null
+++ b/tools/linkmem.py
@@ -0,0 +1,276 @@
+
+import tempfile
+import os
+from os import path
+import sys
+import re
+import subprocess
+from contextlib import contextmanager
+from collections import defaultdict
+import colorsys
+
+import cxxfilt
+from elftools.elf.elffile import ELFFile
+from elftools.elf.enums import ENUM_ST_SHNDX
+from elftools.elf.descriptions import describe_symbol_type, describe_sh_type
+import libarchive
+import matplotlib.cm
+
+@contextmanager
+def chdir(newdir):
+ old_cwd = os.getcwd()
+ try:
+ os.chdir(newdir)
+ yield
+ finally:
+ os.chdir(old_cwd)
+
+def keep_last(it, first=None):
+ last = first
+ for elem in it:
+ yield last, elem
+ last = elem
+
+def delim(start, end, it, first_only=True):
+ found = False
+ for elem in it:
+ if end(elem):
+ if first_only:
+ return
+ found = False
+ elif start(elem):
+ found = True
+ elif found:
+ yield elem
+
+def delim_prefix(start, end, it):
+ yield from delim(lambda l: l.startswith(start), lambda l: end is not None and l.startswith(end), it)
+
+def trace_source_files(linker, cmdline, trace_sections=[], total_sections=['.text', '.data', '.rodata']):
+ with tempfile.TemporaryDirectory() as tempdir:
+ out_path = path.join(tempdir, 'output.elf')
+ output = subprocess.check_output([linker, '-o', out_path, f'-Wl,--print-map', *cmdline])
+ lines = [ line.strip() for line in output.decode().splitlines() ]
+ # FIXME also find isr vector table references
+
+ defs = {}
+ objs = defaultdict(lambda: 0)
+ aliases = {}
+ sec_name = None
+ last_loc = None
+ last_sym = None
+ line_cont = None
+ for last_line, line in keep_last(delim_prefix('Linker script and memory map', 'OUTPUT', lines), first=''):
+ if not line or line.startswith('LOAD '):
+ sec_name = None
+ continue
+
+ # first part of continuation line
+ if m := re.match('^(\.[0-9a-zA-Z-_.]+)$', line):
+ line_cont = line
+ sec_name = None
+ continue
+
+ if line_cont:
+ line = line_cont + ' ' + line
+ line_cont = None
+
+ # -ffunction-sections/-fdata-sections section
+ if m := re.match('^(\.[0-9a-zA-Z-_.]+)\.([0-9a-zA-Z-_.]+)\s+(0x[0-9a-f]+)\s+(0x[0-9a-f]+)\s+(\S+)$', line):
+ sec, sym, loc, size, obj = m.groups()
+ *_, sym = sym.rpartition('.')
+ sym = cxxfilt.demangle(sym)
+ size = int(size, 16)
+ obj = path.abspath(obj)
+
+ if sec not in total_sections:
+ size = 0
+
+ objs[obj] += size
+ defs[sym] = (sec, size, obj)
+
+ sec_name, last_loc, last_sym = sec, loc, sym
+ continue
+
+ # regular (no -ffunction-sections/-fdata-sections) section
+ if m := re.match('^(\.[0-9a-zA-Z-_]+)\s+(0x[0-9a-f]+)\s+(0x[0-9a-f]+)\s+(\S+)$', line):
+ sec, _loc, size, obj = m.groups()
+ size = int(size, 16)
+ obj = path.abspath(obj)
+
+ if sec in total_sections:
+ objs[obj] += size
+
+ sec_name = sec
+ last_loc, last_sym = None, None
+ continue
+
+ # symbol def
+ if m := re.match('^(0x[0-9a-f]+)\s+(\S+)$', line):
+ loc, sym = m.groups()
+ sym = cxxfilt.demangle(sym)
+ loc = int(loc, 16)
+ if sym in defs:
+ continue
+
+ if loc == last_loc:
+ assert last_sym is not None
+ aliases[sym] = last_sym
+ else:
+ assert sec_name
+ defs[sym] = (sec_name, None, obj)
+ last_loc, last_sym = loc, sym
+
+ continue
+
+ refs = defaultdict(lambda: set())
+ for sym, (sec, size, obj) in defs.items():
+ fn, _, member = re.match('^([^()]+)(\((.+)\))?$', obj).groups()
+ fn = path.abspath(fn)
+
+ if member:
+ subprocess.check_call(['ar', 'x', '--output', tempdir, fn, member])
+ fn = path.join(tempdir, member)
+
+ with open(fn, 'rb') as f:
+ elf = ELFFile(f)
+
+ symtab = elf.get_section_by_name('.symtab')
+
+ symtab_demangled = { cxxfilt.demangle(nsym.name).replace(' ', ''): i
+ for i, nsym in enumerate(symtab.iter_symbols()) }
+
+ s = set()
+ sec_map = { sec.name: i for i, sec in enumerate(elf.iter_sections()) }
+ matches = [ i for name, i in sec_map.items() if re.match(f'\.rel\..*\.{sym}', name) ]
+ if matches:
+ sec = elf.get_section(matches[0])
+ for reloc in sec.iter_relocations():
+ refsym = symtab.get_symbol(reloc['r_info_sym'])
+ name = refsym.name if refsym.name else elf.get_section(refsym['st_shndx']).name.split('.')[-1]
+ s.add(name)
+ refs[sym] = s
+
+ for tsec in trace_sections:
+ matches = [ i for name, i in sec_map.items() if name == f'.rel{tsec}' ]
+ s = set()
+ if matches:
+ sec = elf.get_section(matches[0])
+ for reloc in sec.iter_relocations():
+ refsym = symtab.get_symbol(reloc['r_info_sym'])
+ s.add(refsym.name)
+ refs[tsec.replace('.', '_')] |= s
+
+ return objs, aliases, defs, refs
+
+@contextmanager
+def wrap(leader='', print=print, left='{', right='}'):
+ print(leader, left)
+ yield lambda *args, **kwargs: print(' ', *args, **kwargs)
+ print(right)
+
+def mangle(name):
+ return re.sub('[^a-zA-Z0-9_]', '_', name)
+
+hexcolor = lambda r, g, b, *_a: f'#{int(r*255):02x}{int(g*255):02x}{int(b*255):02x}'
+def vhex(val):
+ r,g,b,_a = matplotlib.cm.viridis(1.0-val)
+ fc = hexcolor(r, g, b)
+ h,s,v = colorsys.rgb_to_hsv(r,g,b)
+ cc = '#000000' if v > 0.8 else '#ffffff'
+ return fc, cc
+
+if __name__ == '__main__':
+ import argparse
+ parser = argparse.ArgumentParser()
+ parser.add_argument('--trace-sections', type=str, action='append', default=[])
+ parser.add_argument('--trim-stubs', type=str, action='append', default=[])
+ parser.add_argument('--highlight-subdirs', type=str, default=None)
+ parser.add_argument('linker_binary')
+ parser.add_argument('linker_args', nargs=argparse.REMAINDER)
+ args = parser.parse_args()
+
+ trace_sections = args.trace_sections
+ trace_sections_mangled = { sec.replace('.', '_') for sec in trace_sections }
+ objs, aliases, syms, refs = trace_source_files(args.linker_binary, args.linker_args, trace_sections)
+
+ clusters = defaultdict(lambda: [])
+ for sym, (sec, size, obj) in syms.items():
+ clusters[obj].append((sym, sec, size))
+
+ max_ssize = max(size or 0 for _sec, size, _obj in syms.values())
+ max_osize = max(objs.values())
+
+ subdir_prefix = path.abspath(args.highlight_subdirs) + '/' if args.highlight_subdirs else '### NO HIGHLIGHT ###'
+ first_comp = lambda le_path: path.dirname(le_path).partition(os.sep)[0]
+ subdir_colors = sorted({ first_comp(obj[len(subdir_prefix):]) for obj in objs if obj.startswith(subdir_prefix) })
+ subdir_colors = { path: hexcolor(*matplotlib.cm.Pastel1(i/len(subdir_colors))) for i, path in enumerate(subdir_colors) }
+
+ subdir_sizes = defaultdict(lambda: 0)
+ for obj, size in objs.items():
+ if not isinstance(size, int):
+ continue
+ if obj.startswith(subdir_prefix):
+ subdir_sizes[first_comp(obj[len(subdir_prefix):])] += size
+ else:
+ subdir_sizes['<others>'] += size
+
+ print('Subdir sizes:', file=sys.stderr)
+ for subdir, size in sorted(subdir_sizes.items(), key=lambda x: x[1]):
+ print(f'{subdir:>20}: {size:>6,d} B', file=sys.stderr)
+
+ def lookup_highlight(path):
+ if args.highlight_subdirs:
+ if obj.startswith(subdir_prefix):
+ highlight_head = first_comp(path[len(subdir_prefix):])
+ return subdir_colors[highlight_head], highlight_head
+ else:
+ return '#e0e0e0', None
+ else:
+ return '#ddf7f4', None
+
+ with wrap('digraph G', print) as lvl1print:
+ print('size="23.4,16.5!";')
+ print('graph [fontsize=40];')
+ print('node [fontsize=40];')
+ #print('ratio="fill";')
+
+ print('rankdir=LR;')
+ print('ranksep=5;')
+ print('nodesep=0.2;')
+ print()
+
+ for i, (obj, obj_syms) in enumerate(clusters.items()):
+ with wrap(f'subgraph cluster_{i}', lvl1print) as lvl2print:
+ print('style = "filled";')
+ highlight_color, highlight_head = lookup_highlight(obj)
+ print(f'bgcolor = "{highlight_color}";')
+ print('pencolor = none;')
+ fc, cc = vhex(objs[obj]/max_osize)
+ highlight_subdir_part = f'<font face="carlito" color="{cc}" point-size="40">{highlight_head} / </font>' if highlight_head else ''
+ lvl2print(f'label = <<table border="0"><tr><td border="0" cellpadding="5" bgcolor="{fc}">'
+ f'{highlight_subdir_part}'
+ f'<font face="carlito" color="{cc}"><b>{path.basename(obj)} ({objs[obj]}B)</b></font>'
+ f'</td></tr></table>>;')
+ lvl2print()
+ for sym, sec, size in obj_syms:
+ has_size = isinstance(size, int) and size > 0
+ size_s = f' ({size}B)' if has_size else ''
+ fc, cc = vhex(size/max_ssize) if has_size else ('#ffffff', '#000000')
+ shape = 'box' if sec == '.text' else 'oval'
+ lvl2print(f'{mangle(sym)}[label = "{sym}{size_s}", style="rounded,filled", shape="{shape}", fillcolor="{fc}", fontname="carlito", fontcolor="{cc}" color=none];')
+ lvl1print()
+
+ edges = set()
+ for start, ends in refs.items():
+ for end in ends:
+ end = aliases.get(end, end)
+ if (start in syms or start in trace_sections_mangled) and end in syms:
+ edges.add((start, end))
+
+ for start, end in edges:
+ lvl1print(f'{mangle(start)} -> {mangle(end)} [style="bold", color="#333333"];')
+
+ for sec in trace_sections:
+ lvl1print(f'{sec.replace(".", "_")} [label = "section {sec}", shape="box", style="filled,bold"];')
+
diff --git a/tools/linksize.py b/tools/linksize.py
new file mode 100644
index 0000000..c41a951
--- /dev/null
+++ b/tools/linksize.py
@@ -0,0 +1,62 @@
+#!/usr/bin/env python3
+
+def parse_linker_script(data):
+ pass
+
+def link(groups):
+ defined_symbols = {}
+ undefined_symbols = set()
+ for group, files in groups:
+ while True:
+ found_something = False
+
+ for fn in files:
+ symbols = load_symbols(fn)
+ for symbol in symbols:
+ if symbol in defined_symbols:
+
+ if not group or not found_something:
+ break
+
+
+if __name__ == '__main__':
+
+ import argparse
+ parser = argparse.ArgumentParser()
+ parser.add_argument('-T', '--script', type=str, help='Linker script to use')
+ parser.add_argument('-o', '--output', type=str, help='Output file to produce')
+ args, rest = parser.parse_known_intermixed_args()
+ print(rest)
+
+ addprefix = lambda *xs: [ prefix + opt for opt in xs for prefix in ('', '-Wl,') ]
+ START_GROUP = addprefix('-(', '--start-group')
+ END_GROUP = addprefix('-)', '--end-group')
+ GROUP_OPTS = [*START_GROUP, *END_GROUP]
+ input_files = [ arg for arg in rest if not arg.startswith('-') or arg in GROUP_OPTS ]
+
+ def input_file_iter(input_files):
+ group = False
+ files = []
+ for arg in input_files:
+ if arg in START_GROUP:
+ assert not group
+
+ if files:
+ yield False, files # nested -Wl,--start-group
+ group, files = True, []
+
+ elif arg in END_GROUP:
+ assert group # missing -Wl,--start-group
+ if files:
+ yield True, files
+ group, files = False, []
+
+ else:
+ files.append(arg)
+
+ assert not group # missing -Wl,--end-group
+ if files:
+ yield False, files
+
+
+
diff --git a/tools/linktracer.py b/tools/linktracer.py
new file mode 100644
index 0000000..0c53a60
--- /dev/null
+++ b/tools/linktracer.py
@@ -0,0 +1,118 @@
+#!/usr/bin/env python3
+
+import re
+import subprocess
+import tempfile
+import pprint
+
+ARCHIVE_RE = r'([^(]*)(\([^)]*\))?'
+
+def trace_source_files(linker, cmdline):
+ with tempfile.NamedTemporaryFile() as mapfile:
+ output = subprocess.check_output([linker, f'-Wl,--Map={mapfile.name}', *cmdline])
+
+ # intentionally use generator here
+ idx = 0
+ lines = [ line.rstrip() for line in mapfile.read().decode().splitlines() if line.strip() ]
+
+ for idx, line in enumerate(lines[idx:], start=idx):
+ #print('Dropping', line)
+ if line == 'Linker script and memory map':
+ break
+
+ idx += 1
+ objects = []
+ symbols = {}
+ sections = {}
+ current_object = None
+ last_offset = None
+ last_symbol = None
+ cont_sec = None
+ cont_ind = None
+ current_section = None
+ for idx, line in enumerate(lines[idx:], start=idx):
+ print(f'Processing >{line}')
+ if line.startswith('LOAD'):
+ _load, obj = line.split()
+ objects.append(obj)
+ continue
+
+ if line.startswith('OUTPUT'):
+ break
+
+ m = re.match(r'^( ?)([^ ]+)? +(0x[0-9a-z]+) +(0x[0-9a-z]+)?(.*)?$', line)
+ if m is None:
+ m = re.match(r'^( ?)([^ ]+)?$', line)
+ if m:
+ cont_ind, cont_sec = m.groups()
+ else:
+ cont_ind, cont_sec = None, None
+ last_offset, last_symbol = None, None
+ continue
+ indent, sec, offx, size, sym_or_src = m.groups()
+ if sec is None:
+ sec = cont_sec
+ ind = cont_ind
+ cont_sec = None
+ cont_ind = None
+ print(f'vals: indent={indent} sec={sec} offx={offx} size={size} sym_or_src={sym_or_src}')
+ if not re.match('^[a-zA-Z_0-9<>():*]+$', sym_or_src):
+ continue
+
+ if indent == '':
+ print(f'Section: {sec} 0x{size:x}')
+ current_section = sec
+ sections[sec] = size
+ last_offset = None
+ last_symbol = None
+ continue
+
+ if offx is not None:
+ offx = int(offx, 16)
+ if size is not None:
+ size = int(size, 16)
+
+ if size is not None and sym_or_src is not None:
+ # archive/object line
+ archive, _member = re.match(ARCHIVE_RE, sym_or_src).groups()
+ current_object = archive
+ last_offset = offx
+ else:
+ if sym_or_src is not None:
+ assert size is None
+ if last_offset is not None:
+ last_size = offx - last_offset
+ symbols[last_symbol] = (last_size, current_section)
+ print(f'Symbol: {last_symbol} 0x{last_size:x} @{current_section}')
+ last_offset = offx
+ last_symbol = sym_or_src
+
+ idx += 1
+
+ for idx, line in enumerate(lines[idx:], start=idx):
+ if line == 'Cross Reference Table':
+ break
+
+ idx += 1
+
+ # map which symbol was pulled from which object in the end
+ used_defs = {}
+ for line in lines:
+ *left, right = line.split()
+
+ archive, _member = re.match(ARCHIVE_RE, right).groups()
+ if left:
+ used_defs[''.join(left)] = archive
+
+ #pprint.pprint(symbols)
+
+
+if __name__ == '__main__':
+ import argparse
+ parser = argparse.ArgumentParser()
+ parser.add_argument('linker_binary')
+ parser.add_argument('linker_args', nargs=argparse.REMAINDER)
+ args = parser.parse_args()
+
+ source_files = trace_source_files(args.linker_binary, args.linker_args)
+
diff --git a/tools/mapparse.py b/tools/mapparse.py
new file mode 100644
index 0000000..c1f460a
--- /dev/null
+++ b/tools/mapparse.py
@@ -0,0 +1,129 @@
+
+import re
+from collections import defaultdict, namedtuple
+
+Section = namedtuple('Section', ['name', 'offset', 'objects'])
+ObjectEntry = namedtuple('ObjectEntry', ['filename', 'object', 'offset', 'size'])
+FileEntry = namedtuple('FileEntry', ['section', 'object', 'offset', 'length'])
+
+class Memory:
+ def __init__(self, name, origin, length, attrs=''):
+ self.name, self.origin, self.length, self.attrs = name, origin, length, attrs
+ self.sections = {}
+ self.files = defaultdict(lambda: [])
+ self.totals = defaultdict(lambda: 0)
+
+ def add_toplevel(self, name, offx, length):
+ self.sections[name] = Section(offx, length, [])
+
+ def add_obj(self, name, offx, length, fn, obj):
+ base_section, sep, subsec = name[1:].partition('.')
+ base_section = '.'+base_section
+ if base_section in self.sections:
+ sec = secname, secoffx, secobjs = self.sections[base_section]
+ secobjs.append(ObjectEntry(fn, obj, offx, length))
+ else:
+ sec = None
+ self.files[fn].append(FileEntry(sec, obj, offx, length))
+ self.totals[fn] += length
+
+class MapFile:
+ def __init__(self, s):
+ self._lines = s.splitlines()
+ self.memcfg = {}
+ self.defaultmem = Memory('default', 0, 0xffffffffffffffff)
+ self._parse()
+
+ def __getitem__(self, offx_or_name):
+ ''' Lookup a memory area by name or address '''
+ if offx_or_name in self.memcfg:
+ return self.memcfg[offx_or_name]
+
+ elif isinstance(offx_or_name, int):
+ for mem in self.memcfg.values():
+ if mem.origin <= offx_or_name < mem.origin+mem.length:
+ return mem
+ else:
+ return self.defaultmem
+
+ raise ValueError('Invalid argument type for indexing')
+
+ def _skip(self, regex):
+ matcher = re.compile(regex)
+ for l in self:
+ if matcher.match(l):
+ break
+
+ def __iter__(self):
+ while self._lines:
+ yield self._lines.pop(0)
+
+ def _parse(self):
+ self._skip('^Memory Configuration')
+
+ # Parse memory segmentation info
+ self._skip('^Name')
+ for l in self:
+ if not l:
+ break
+ name, origin, length, *attrs = l.split()
+ if not name.startswith('*'):
+ self.memcfg[name] = Memory(name, int(origin, 16), int(length, 16), attrs[0] if attrs else '')
+
+ # Parse section information
+ toplevel_m = re.compile('^(\.[a-zA-Z0-9_.]+)\s+(0x[0-9a-fA-F]+)\s+(0x[0-9a-fA-F]+)')
+ secondlevel_m = re.compile('^ (\.[a-zA-Z0-9_.]+)\s+(0x[0-9a-fA-F]+)\s+(0x[0-9a-fA-F]+)\s+(.*)$')
+ secondlevel_linebreak_m = re.compile('^ (\.[a-zA-Z0-9_.]+)\n')
+ filelike = re.compile('^(/?[^()]*\.[a-zA-Z0-9-_]+)(\(.*\))?')
+ linebreak_section = None
+ for l in self:
+ # Toplevel section
+ match = toplevel_m.match(l)
+ if match:
+ name, offx, length = match.groups()
+ offx, length = int(offx, 16), int(length, 16)
+ self[offx].add_toplevel(name, offx, length)
+
+ match = secondlevel_linebreak_m.match(l)
+ if match:
+ linebreak_section, = match.groups()
+ continue
+
+ if linebreak_section:
+ l = ' {} {}'.format(linebreak_section, l)
+ linebreak_section = None
+
+ # Second-level section
+ match = secondlevel_m.match(l)
+ if match:
+ name, offx, length, misc = match.groups()
+ match = filelike.match(misc)
+ if match:
+ fn, obj = match.groups()
+ obj = obj.strip('()') if obj else None
+ offx, length = int(offx, 16), int(length, 16)
+ self[offx].add_obj(name, offx, length, fn, obj)
+
+
+if __name__ == '__main__':
+ import argparse
+ parser = argparse.ArgumentParser(description='Parser GCC map file')
+ parser.add_argument('mapfile', type=argparse.FileType('r'), help='The GCC .map file to parse')
+ parser.add_argument('-m', '--memory', type=str, help='The memory segments to print, comma-separated')
+ args = parser.parse_args()
+ mf = MapFile(args.mapfile.read())
+ args.mapfile.close()
+
+ mems = args.memory.split(',') if args.memory else mf.memcfg.keys()
+
+ for name in mems:
+ mem = mf.memcfg[name]
+ print('Symbols by file for memory', name)
+ for tot, fn in reversed(sorted( (tot, fn) for fn, tot in mem.totals.items() )):
+ print(' {:>8} {}'.format(tot, fn))
+ for length, offx, sec, obj in reversed(sorted(( (length, offx, sec, obj) for sec, obj, offx, length in
+ mem.files[fn] ), key=lambda e: e[0] )):
+ name = sec.name if sec else None
+ print(' {:>8} {:>#08x} {}'.format(length, offx, obj))
+ #print('{:>16} 0x{:016x} 0x{:016x} ({:>24}) {}'.format(name, origin, length, length, attrs))
+
diff --git a/tools/presig_gen.py b/tools/presig_gen.py
new file mode 100644
index 0000000..c5dafe7
--- /dev/null
+++ b/tools/presig_gen.py
@@ -0,0 +1,141 @@
+#!/usr/bin/env python3
+
+import os
+import sys
+import textwrap
+import uuid
+import hmac
+import binascii
+import time
+from datetime import datetime
+
+LINKING_KEY_SIZE = 15
+PRESIG_VERSION = '000.001'
+DOMAINS = ['all', 'country', 'region', 'vendor', 'series']
+
+def format_hex(data, indent=4, wrap=True):
+ indent = ' '*indent
+ par = ', '.join(f'0x{b:02x}' for b in data)
+ par = textwrap.fill(par, width=120,
+ initial_indent=indent, subsequent_indent=indent,
+ replace_whitespace=False, drop_whitespace=False)
+ if wrap:
+ return f'{{\n{par}\n}}'
+ return par
+
+def domain_string(domain, value):
+ return f'smart reset domain string v{PRESIG_VERSION}: domain:{domain}={value}'
+
+def keygen_cmd(args):
+ if os.path.exists(args.keyfile) and not args.force:
+ print("Error: keyfile already exists. We won't overwrite it. Instead please remove it manually.",
+ file=sys.stderr)
+ return 1
+
+ root_key = os.urandom(LINKING_KEY_SIZE)
+
+ with open(args.keyfile, 'wb') as f:
+ f.write(binascii.hexlify(root_key))
+ f.write(b'\n')
+ return 0
+
+def gen_at_height(domain, value, height, key):
+ # nanananananana BLOCKCHAIN!
+
+ ds = domain_string(domain, value).encode('utf-8')
+
+ for height in range(height+1):
+ key = hmac.digest(key, ds, 'sha512')[:LINKING_KEY_SIZE]
+
+ return key
+
+def auth_cmd(args):
+ with open(args.keyfile, 'r') as f:
+ root_key = binascii.unhexlify(f.read().strip())
+
+ vals = [ (domain, getattr(args, domain)) for domain in DOMAINS if getattr(args, domain) is not None ]
+ if not vals:
+ vals = [('all', 'all')]
+ for domain, value in vals:
+ auth = gen_at_height(domain, value, args.height, root_key)
+ print(f'{domain}="{value}" @{args.height}: {binascii.hexlify(auth).decode()}')
+
+
+def prekey_cmd(args):
+ with open(args.keyfile, 'r') as f:
+ root_key = binascii.unhexlify(f.read().strip())
+
+ print('#include <stdint.h>')
+ print('#include <assert.h>')
+ print()
+ print('#include "crypto.h"')
+ print()
+
+ bundle_id = uuid.uuid4().bytes
+ print(f'/* bundle id {binascii.hexlify(bundle_id).decode()} */')
+ print(f'uint8_t presig_bundle_id[16] = {format_hex(bundle_id)};')
+ print()
+ print(f'/* generated on {datetime.now()} */')
+ print(f'uint64_t bundle_timestamp = {int(time.time())};')
+ print()
+ print(f'int presig_height = {args.max_height};')
+ print()
+
+ print('const char *presig_domain_strings[_TRIGGER_DOMAIN_COUNT] = {')
+ for domain in DOMAINS:
+ ds = domain_string(domain, getattr(args, domain))
+ assert '"' not in ds
+ print(f' [TRIGGER_DOMAIN_{domain.upper()}] = "{ds}",')
+ print('};')
+ print()
+
+ print('uint8_t presig_keys[_TRIGGER_DOMAIN_COUNT][PRESIG_MSG_LEN] = {')
+ for domain in DOMAINS:
+ key = gen_at_height(domain, getattr(args, domain), args.max_height, root_key)
+ print(f' [TRIGGER_DOMAIN_{domain.upper()}] = {{{format_hex(key, indent=0, wrap=False)}}},')
+ print('};')
+
+ print()
+ print('static inline void __hack_asserts_only(void) {')
+ print(f' static_assert(_TRIGGER_DOMAIN_COUNT == {len(DOMAINS)});')
+ print(f' static_assert(PRESIG_MSG_LEN == {LINKING_KEY_SIZE});')
+ print('}')
+ print()
+
+
+
+TEST_VENDOR = 'Darthenschmidt Cyberei und Verschleierungstechnik GmbH'
+TEST_SERIES = 'Frobnicator v0.23.7'
+TEST_REGION = 'Neuland'
+TEST_COUNTRY = 'Germany'
+
+if __name__ == '__main__':
+ import argparse
+ parser = argparse.ArgumentParser()
+ parser.add_argument('keyfile', help='Key file to use')
+
+ subparsers = parser.add_subparsers(title='subcommands')
+ keygen_parser = subparsers.add_parser('keygen', help='Generate a new key')
+ keygen_parser.add_argument('-f', '--force', action='store_true', help='Force overwriting existing keyfile')
+ keygen_parser.set_defaults(func=keygen_cmd)
+
+ auth_parser = subparsers.add_parser('auth', help='Generate one-time authentication string')
+ auth_parser.add_argument('height', type=int, help='Authentication string height, counting from 0 (root key)')
+ auth_parser.set_defaults(func=auth_cmd)
+ auth_parser.add_argument('-a', '--all', action='store_const', const='all', help='Vendor name for vendor domain')
+ auth_parser.add_argument('-v', '--vendor', type=str, nargs='?', const=TEST_VENDOR, help='Vendor name for vendor domain')
+ auth_parser.add_argument('-s', '--series', type=str, nargs='?', const=TEST_SERIES, help='Series identifier for series domain')
+ auth_parser.add_argument('-r', '--region', type=str, nargs='?', const=TEST_REGION, help='Region name for region domain')
+ auth_parser.add_argument('-c', '--country', type=str, nargs='?', const=TEST_COUNTRY, help='Country name for country domain')
+
+ prekey_parser = subparsers.add_parser('prekey', help='Generate prekey data .C source code file')
+ prekey_parser.add_argument('-m', '--max-height', type=int, default=8, help='Height of generated prekey')
+ prekey_parser.add_argument('-v', '--vendor', type=str, default=TEST_VENDOR, help='Vendor name for vendor domain')
+ prekey_parser.add_argument('-s', '--series', type=str, default=TEST_SERIES, help='Series identifier for series domain')
+ prekey_parser.add_argument('-r', '--region', type=str, default=TEST_REGION, help='Region name for region domain')
+ prekey_parser.add_argument('-c', '--country', type=str, default=TEST_COUNTRY, help='Country name for country domain')
+ prekey_parser.set_defaults(func=prekey_cmd, all='all')
+
+ args = parser.parse_args()
+ sys.exit(args.func(args))
+
diff --git a/tools/reed_solomon.py b/tools/reed_solomon.py
new file mode 100644
index 0000000..c4ca6e4
--- /dev/null
+++ b/tools/reed_solomon.py
@@ -0,0 +1,91 @@
+import os, sys
+import ctypes as C
+import argparse
+import binascii
+import numpy as np
+import timeit
+import statistics
+
+lib = C.CDLL('rslib.so')
+
+lib.rslib_encode.argtypes = [C.c_int, C.c_size_t, C.POINTER(C.c_char), C.POINTER(C.c_char)]
+lib.rslib_decode.argtypes = [C.c_int, C.c_size_t, C.POINTER(C.c_char)]
+lib.rslib_gexp.argtypes = [C.c_int, C.c_int]
+lib.rslib_gexp.restype = C.c_int
+lib.rslib_decode.restype = C.c_int
+lib.rslib_npar.restype = C.c_size_t
+
+def npar():
+ return lib.rslib_npar()
+
+def encode(data: bytes, nbits=8):
+ out = C.create_string_buffer(len(data) + lib.rslib_npar())
+ lib.rslib_encode(nbits, len(data), data, out)
+ return out.raw
+
+def decode(data: bytes, nbits=8):
+ inout = C.create_string_buffer(data)
+ lib.rslib_decode(nbits, len(data), inout)
+ return inout.raw[:-lib.rslib_npar() - 1]
+
+def cmdline_func_test(args, print=lambda *args, **kwargs: None, benchmark=False):
+ st = np.random.RandomState(seed=args.seed)
+
+ lfsr = [lib.rslib_gexp(i, args.bits) for i in range(2**args.bits - 1)]
+ print('LFSR', len(set(lfsr)), lfsr)
+ assert all(0 < x < 2**args.bits for x in lfsr)
+ assert len(set(lfsr)) == 2**args.bits - 1
+
+ print('Seed', args.seed)
+ for i in range(args.repeat):
+ print(f'Run {i}')
+ test_data = bytes(st.randint(2**args.bits, size=args.message_length, dtype=np.uint8))
+ print(' Raw:', binascii.hexlify(test_data).decode())
+ encoded = encode(test_data, nbits=args.bits)
+ print(' Encoded:', binascii.hexlify(encoded).decode())
+
+ indices = st.permutation(len(encoded))
+ encoded = list(encoded)
+ for pos in indices[:args.errors]:
+ encoded[pos] = st.randint(2**args.bits)
+ encoded = bytes(encoded)
+ print(' Modified:', ''.join(f'\033[91m{b:02x}\033[0m' if pos in indices[:args.errors] else f'{b:02x}' for pos, b in enumerate(encoded)))
+
+ if benchmark:
+ rpt = 10000
+ delta = timeit.timeit('decode(encoded, nbits=args.bits)',
+ globals={'args': args, 'decode': decode, 'encoded': encoded},
+ number=rpt)/rpt
+ print(f'Decoding runtime: {delta*1e6:.3f}μs')
+ decoded = decode(encoded, nbits=args.bits)
+ print(' Decoded:', binascii.hexlify(decoded).decode())
+ print(' Delta:', binascii.hexlify(
+ bytes(x^y for x, y in zip(test_data, decoded))
+ ).decode().replace('0', '.'))
+ assert test_data == decoded
+
+def cmdline_func_encode(args, **kwargs):
+ data = np.frombuffer(binascii.unhexlify(args.hex_str), dtype=np.uint8)
+ # Map 8 bit input to 6 bit symbol string
+ data = np.packbits(np.pad(np.unpackbits(data).reshape((-1, 6)), ((0,0),(2, 0))).flatten())
+ encoded = encode(data.tobytes(), nbits=args.bits)
+ print('symbol array:', ', '.join(f'0x{x:02x}' for x in encoded))
+ print('hex string:', binascii.hexlify(encoded).decode())
+
+if __name__ == '__main__':
+ parser = argparse.ArgumentParser()
+ cmd_parser = parser.add_subparsers(required=True)
+ test_parser = cmd_parser.add_parser('test', help='Test reed-solomon implementation')
+ test_parser.add_argument('-m', '--message-length', type=int, default=6, help='Test message (plaintext) length in bytes')
+ test_parser.add_argument('-e', '--errors', type=int, default=2, help='Number of byte errors to insert into simulation')
+ test_parser.add_argument('-r', '--repeat', type=int, default=1000, help='Repeat experiment -r times')
+ test_parser.add_argument('-b', '--bits', type=int, default=8, help='Symbol bit size')
+ test_parser.add_argument('-s', '--seed', type=int, default=0, help='Random seed')
+ test_parser.set_defaults(func=cmdline_func_test)
+ enc_parser = cmd_parser.add_parser('encode', help='RS-Encode given hex string')
+ enc_parser.set_defaults(func=cmdline_func_encode)
+ enc_parser.add_argument('-b', '--bits', type=int, default=8, help='Symbol bit size')
+ enc_parser.add_argument('hex_str', type=str, help='Input data as hex string')
+ args = parser.parse_args()
+ args.func(args)
+