Initial commit

14 files changed, 1483 insertions, 0 deletions

diff --git a/tools/decode_logic_analzyer.py b/tools/decode_logic_analzyer.py
new file mode 100644
index 0000000..be1d4da
--- /dev/null
+++ b/tools/decode_logic_analzyer.py
@@ -0,0 +1,88 @@
+#!/usr/bin/env python
+
+import sys
+import subprocess
+import binascii
+from statistics import mean
+from dataclasses import dataclass, fields
+import struct
+from pprint import pprint
+
+from cobs import cobs
+
+time = [*sys.argv, '1s'][1]
+proc = subprocess.run(f'sigrok-cli --driver dreamsourcelab-dslogic --config samplerate=10M --channels 0,1 --protocol-decoders uart:baudrate=250000:rx=1 --protocol-decoder-annotations uart=rx-data,uart=tx-data --time {time}'.split(), check=True, capture_output=True, text=True)
+data = [line.partition(':')[2] for line in proc.stdout.splitlines()]
+data = bytes([int(x, 16) for x in data if x])
+
+class Serialized:
+    @classmethod
+    def deserialize(kls, data):
+        fields = struct.unpack(kls._struct_format(), data)
+        mapped = [cast(val) for cast, val in zip(kls._struct_casts(), fields)]
+        return kls(*mapped)
+
+    @classmethod
+    def _struct_format(kls):
+        return kls._parse_fields()[0]
+
+    @classmethod
+    def _struct_casts(kls):
+        return kls._parse_fields()[1]
+
+    @classmethod
+    def _parse_fields(kls):
+        fmt = '<'
+        casts = []
+        for field in fields(kls):
+            if isinstance(field.type, tuple):
+                struct_type, cast = field.type
+            else:
+                struct_type, cast = field.type, int
+            fmt += struct_type
+            casts.append(cast)
+        return fmt, casts
+
+@dataclass
+class Header(Serialized):
+    crc: 'I'
+    src: 'B'
+    dst: 'B'
+    pid: 'B'
+    packet_type: 'B'
+
+@dataclass
+class ADCPacket(Serialized):
+    timestamp: 'Q'
+    sampling_interval: 'I'
+    total_samples: 'I'
+    sample_count: 'I'
+    samples: ('96s', bytes)
+
+    def __post_init__(self):
+        data = self.samples
+        foo = lambda x: x if x < 0x800000 else x-0x1000000
+        self.samples = [[
+            foo(struct.unpack('<I', data[3*(2*sample + channel):][:3] + b'\0')[0])
+            for sample in range(16)
+            ] for channel in range(2)]
+
+
+norm_a, norm_b = 0, 0
+for packet in data.split(b'\0'):
+    try:
+        packet = cobs.decode(packet)
+        hdr = Header.deserialize(packet[:8])
+        if hdr.packet_type == 2:
+            packet = ADCPacket.deserialize(packet[8:])
+            diff_a = max([abs(x - norm_a) for x in packet.samples[0][:packet.sample_count]])
+            diff_b = max([abs(x - norm_b) for x in packet.samples[1][:packet.sample_count]])
+            if diff_a > 10000 or diff_b > 10000:
+                pprint(packet)
+                norm_a = mean(packet.samples[0][:packet.sample_count])
+                norm_b = mean(packet.samples[1][:packet.sample_count])
+            elif any(x != 0 for x in packet.samples[0][packet.sample_count:] + packet.samples[1][packet.sample_count:]):
+                pprint('nonzero', packet)
+    except (cobs.DecodeError, struct.error):
+        print('Decoding error')
+
diff --git a/tools/extract_pinmap.py b/tools/extract_pinmap.py
new file mode 100644
index 0000000..d2c52e8
--- /dev/null
+++ b/tools/extract_pinmap.py
@@ -0,0 +1,11 @@
+#!/usr/bin/env python
+
+import click
+
+@click.command()
+@click.option('sch_file')
+def cli():
+
+
+if __name__ == '__main__':
+    cli()
diff --git a/tools/gen_isr_header.py b/tools/gen_isr_header.py
new file mode 100644
index 0000000..6382b1f
--- /dev/null
+++ b/tools/gen_isr_header.py
@@ -0,0 +1,66 @@
+#!/usr/bin/env python3
+
+import subprocess
+import os
+import re
+import datetime
+
+def cpp_preprocess(input_path, cpp='cpp'):
+    return subprocess.check_output([cpp, '-P', input_path]).decode()
+
+def gen_isr_header(f, cpp='cpp'):
+    stripped_code = cpp_preprocess(args.input, args.use_cpp)
+
+    armed = False
+    for line in stripped_code.splitlines():
+        line = line.strip()
+
+        if armed:
+            if not line.startswith('.word'):
+                break
+
+            word, value = line.split()
+            assert word == '.word'
+            if value == '0':
+                yield None
+            else:
+                yield value
+
+        else:
+            if line.startswith('g_pfnVectors:'):
+                armed = True
+
+    else:
+        raise ValueError('Cannot find interrupt vector definition!')
+
+if __name__ == '__main__':
+    import argparse
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--use-cpp', type=str, default=os.getenv('CPP', 'cpp'), help='cpp (C preprocessor) executable to use')
+    parser.add_argument('-g', '--generate-include-guards', action='store_true', help='Whether to generate include guards')
+    parser.add_argument('input', help='Input stm32****_startup.s file')
+    args = parser.parse_args()
+
+    print('/* AUTOGENERATED FILE! DO NOT MODIFY! */')
+    print('/* Generated {datetime.datetime.utcnow()} from {args.input} */')
+    if args.generate_include_guards:
+        include_guard_id = '__ISR_HEADER_' + re.sub('[^A-Za-z0-9]', '_', args.input.split('/')[-1]) + '__'
+        print(f'#ifndef {include_guard_id}')
+        print(f'#define {include_guard_id}')
+
+    print()
+    for i, handler_name in enumerate(gen_isr_header(args.input, args.use_cpp)):
+        if handler_name is None:
+            print(f'/* IRQ {i} is undefined for this part. */')
+        else:
+            print(f'void {handler_name}(void); {" " * (30-len(handler_name))} /* {i:> 3} */')
+    print()
+
+    print(f'#define NUM_IRQs {i+1}')
+    print('extern uint32_t g_pfnVectors[NUM_IRQs];')
+    print('#define isr_vector g_pfnVectors')
+    print()
+
+    if args.generate_include_guards:
+        print(f'#endif /* {include_guard_id} */')
+    
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..452471d
--- /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(r'^(\.[0-9a-zA-Z-_.]+)$', line):
+                line_cont = line
+                sec_name = None
+                continue
+
+            if line_cont:
+                line = line_cont + ' ' + line
+                line_cont = None
+
+            # -ffunction-sections/-fdata-sections section
+            if m := re.match(r'^(\.[0-9a-zA-Z-_.]+)\.([0-9a-zA-Z-_.]+)\s+(0x[0-9a-f]+)\s+(0x[0-9a-f]+)\s+(\S+)$', 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(r'^(\.[0-9a-zA-Z-_]+)\s+(0x[0-9a-f]+)\s+(0x[0-9a-f]+)\s+(\S+)$', line):
+                sec, _loc, size, obj = m.groups()
+                size = int(size, 16)
+                obj = path.abspath(obj)
+
+                if sec in total_sections:
+                    objs[obj] += size
+
+                sec_name = sec
+                last_loc, last_sym = None, None
+                continue
+            
+            # symbol def
+            if m := re.match(r'^(0x[0-9a-f]+)\s+(\S+)$', 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(r'^([^()]+)(\((.+)\))?$', 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(fr'\.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/musl_include_shims/bits/alltypes.h b/tools/musl_include_shims/bits/alltypes.h
new file mode 100644
index 0000000..581ca85
--- /dev/null
+++ b/tools/musl_include_shims/bits/alltypes.h
@@ -0,0 +1,23 @@
+
+/* shim file for musl */
+
+#ifndef __MUSL_SHIM_BITS_ALLTYPES_H__
+#define __MUSL_SHIM_BITS_ALLTYPES_H__
+
+#define _REDIR_TIME64 1
+#define _Addr int
+#define _Int64 long long
+#define _Reg int
+
+#define __BYTE_ORDER 1234
+
+#define __LONG_MAX 0x7fffffffL
+
+#ifndef __cplusplus
+typedef unsigned wchar_t;
+#endif
+
+typedef float float_t;
+typedef double double_t;
+
+#endif /* __MUSL_SHIM_BITS_ALLTYPES_H__ */
diff --git a/tools/musl_include_shims/endian.h b/tools/musl_include_shims/endian.h
new file mode 100644
index 0000000..172c432
--- /dev/null
+++ b/tools/musl_include_shims/endian.h
@@ -0,0 +1,80 @@
+#ifndef _ENDIAN_H
+#define _ENDIAN_H
+
+#include <features.h>
+
+#define __NEED_uint16_t
+#define __NEED_uint32_t
+#define __NEED_uint64_t
+
+#include <bits/alltypes.h>
+
+#define __PDP_ENDIAN 3412
+
+#define BIG_ENDIAN __BIG_ENDIAN
+#define LITTLE_ENDIAN __LITTLE_ENDIAN
+#define PDP_ENDIAN __PDP_ENDIAN
+#define BYTE_ORDER __BYTE_ORDER
+
+static __inline uint16_t __bswap16(uint16_t __x)
+{
+	return __x<<8 | __x>>8;
+}
+
+static __inline uint32_t __bswap32(uint32_t __x)
+{
+	return __x>>24 | __x>>8&0xff00 | __x<<8&0xff0000 | __x<<24;
+}
+
+static __inline uint64_t __bswap64(uint64_t __x)
+{
+	return __bswap32(__x)+0ULL<<32 | __bswap32(__x>>32);
+}
+
+#if __BYTE_ORDER == __LITTLE_ENDIAN
+#define htobe16(x) __bswap16(x)
+#define be16toh(x) __bswap16(x)
+#define htobe32(x) __bswap32(x)
+#define be32toh(x) __bswap32(x)
+#define htobe64(x) __bswap64(x)
+#define be64toh(x) __bswap64(x)
+#define htole16(x) (uint16_t)(x)
+#define le16toh(x) (uint16_t)(x)
+#define htole32(x) (uint32_t)(x)
+#define le32toh(x) (uint32_t)(x)
+#define htole64(x) (uint64_t)(x)
+#define le64toh(x) (uint64_t)(x)
+#else
+#define htobe16(x) (uint16_t)(x)
+#define be16toh(x) (uint16_t)(x)
+#define htobe32(x) (uint32_t)(x)
+#define be32toh(x) (uint32_t)(x)
+#define htobe64(x) (uint64_t)(x)
+#define be64toh(x) (uint64_t)(x)
+#define htole16(x) __bswap16(x)
+#define le16toh(x) __bswap16(x)
+#define htole32(x) __bswap32(x)
+#define le32toh(x) __bswap32(x)
+#define htole64(x) __bswap64(x)
+#define le64toh(x) __bswap64(x)
+#endif
+
+#if defined(_GNU_SOURCE) || defined(_BSD_SOURCE)
+#if __BYTE_ORDER == __LITTLE_ENDIAN
+#define betoh16(x) __bswap16(x)
+#define betoh32(x) __bswap32(x)
+#define betoh64(x) __bswap64(x)
+#define letoh16(x) (uint16_t)(x)
+#define letoh32(x) (uint32_t)(x)
+#define letoh64(x) (uint64_t)(x)
+#else
+#define betoh16(x) (uint16_t)(x)
+#define betoh32(x) (uint32_t)(x)
+#define betoh64(x) (uint64_t)(x)
+#define letoh16(x) __bswap16(x)
+#define letoh32(x) __bswap32(x)
+#define letoh64(x) __bswap64(x)
+#endif
+#endif
+
+#endif
diff --git a/tools/musl_include_shims/features.h b/tools/musl_include_shims/features.h
new file mode 100644
index 0000000..85cfb72
--- /dev/null
+++ b/tools/musl_include_shims/features.h
@@ -0,0 +1,40 @@
+#ifndef _FEATURES_H
+#define _FEATURES_H
+
+#if defined(_ALL_SOURCE) && !defined(_GNU_SOURCE)
+#define _GNU_SOURCE 1
+#endif
+
+#if defined(_DEFAULT_SOURCE) && !defined(_BSD_SOURCE)
+#define _BSD_SOURCE 1
+#endif
+
+#if !defined(_POSIX_SOURCE) && !defined(_POSIX_C_SOURCE) \
+ && !defined(_XOPEN_SOURCE) && !defined(_GNU_SOURCE) \
+ && !defined(_BSD_SOURCE) && !defined(__STRICT_ANSI__)
+#define _BSD_SOURCE 1
+#define _XOPEN_SOURCE 700
+#endif
+
+#if __STDC_VERSION__ >= 199901L
+#define __restrict restrict
+#elif !defined(__GNUC__)
+#define __restrict
+#endif
+
+#if __STDC_VERSION__ >= 199901L || defined(__cplusplus)
+#define __inline inline
+#elif !defined(__GNUC__)
+#define __inline
+#endif
+
+#if __STDC_VERSION__ >= 201112L
+#elif defined(__GNUC__)
+#define _Noreturn __attribute__((__noreturn__))
+#else
+#define _Noreturn
+#endif
+
+#define __REDIR(x,y) __typeof__(x) x __asm__(#y)
+
+#endif
diff --git a/tools/musl_include_shims/fp_arch.h b/tools/musl_include_shims/fp_arch.h
new file mode 100644
index 0000000..f5bab6d
--- /dev/null
+++ b/tools/musl_include_shims/fp_arch.h
@@ -0,0 +1,6 @@
+#ifndef __MUSL_SHIM_FP_ARCH_H__
+#define __MUSL_SHIM_FP_ARCH_H__
+
+#define hidden
+
+#endif /* __MUSL_SHIM_FP_ARCH_H__ */
diff --git a/tools/musl_include_shims/libm.h b/tools/musl_include_shims/libm.h
new file mode 100644
index 0000000..d48135d
--- /dev/null
+++ b/tools/musl_include_shims/libm.h
@@ -0,0 +1,270 @@
+#ifndef _LIBM_H
+#define _LIBM_H
+
+#include <stdint.h>
+#include <float.h>
+#include <math.h>
+#include <endian.h>
+#include "fp_arch.h"
+
+#if LDBL_MANT_DIG == 53 && LDBL_MAX_EXP == 1024
+#elif LDBL_MANT_DIG == 64 && LDBL_MAX_EXP == 16384 && __BYTE_ORDER == __LITTLE_ENDIAN
+union ldshape {
+	long double f;
+	struct {
+		uint64_t m;
+		uint16_t se;
+	} i;
+};
+#elif LDBL_MANT_DIG == 64 && LDBL_MAX_EXP == 16384 && __BYTE_ORDER == __BIG_ENDIAN
+/* This is the m68k variant of 80-bit long double, and this definition only works
+ * on archs where the alignment requirement of uint64_t is <= 4. */
+union ldshape {
+	long double f;
+	struct {
+		uint16_t se;
+		uint16_t pad;
+		uint64_t m;
+	} i;
+};
+#elif LDBL_MANT_DIG == 113 && LDBL_MAX_EXP == 16384 && __BYTE_ORDER == __LITTLE_ENDIAN
+union ldshape {
+	long double f;
+	struct {
+		uint64_t lo;
+		uint32_t mid;
+		uint16_t top;
+		uint16_t se;
+	} i;
+	struct {
+		uint64_t lo;
+		uint64_t hi;
+	} i2;
+};
+#elif LDBL_MANT_DIG == 113 && LDBL_MAX_EXP == 16384 && __BYTE_ORDER == __BIG_ENDIAN
+union ldshape {
+	long double f;
+	struct {
+		uint16_t se;
+		uint16_t top;
+		uint32_t mid;
+		uint64_t lo;
+	} i;
+	struct {
+		uint64_t hi;
+		uint64_t lo;
+	} i2;
+};
+#else
+#error Unsupported long double representation
+#endif
+
+/* Support non-nearest rounding mode.  */
+#define WANT_ROUNDING 1
+/* Support signaling NaNs.  */
+#define WANT_SNAN 0
+
+#if WANT_SNAN
+#error SNaN is unsupported
+#else
+#define issignalingf_inline(x) 0
+#define issignaling_inline(x) 0
+#endif
+
+#ifndef TOINT_INTRINSICS
+#define TOINT_INTRINSICS 0
+#endif
+
+#if TOINT_INTRINSICS
+/* Round x to nearest int in all rounding modes, ties have to be rounded
+   consistently with converttoint so the results match.  If the result
+   would be outside of [-2^31, 2^31-1] then the semantics is unspecified.  */
+static double_t roundtoint(double_t);
+
+/* Convert x to nearest int in all rounding modes, ties have to be rounded
+   consistently with roundtoint.  If the result is not representible in an
+   int32_t then the semantics is unspecified.  */
+static int32_t converttoint(double_t);
+#endif
+
+/* Helps static branch prediction so hot path can be better optimized.  */
+#ifdef __GNUC__
+#define predict_true(x) __builtin_expect(!!(x), 1)
+#define predict_false(x) __builtin_expect(x, 0)
+#else
+#define predict_true(x) (x)
+#define predict_false(x) (x)
+#endif
+
+/* Evaluate an expression as the specified type. With standard excess
+   precision handling a type cast or assignment is enough (with
+   -ffloat-store an assignment is required, in old compilers argument
+   passing and return statement may not drop excess precision).  */
+
+static inline float eval_as_float(float x)
+{
+	float y = x;
+	return y;
+}
+
+static inline double eval_as_double(double x)
+{
+	double y = x;
+	return y;
+}
+
+/* fp_barrier returns its input, but limits code transformations
+   as if it had a side-effect (e.g. observable io) and returned
+   an arbitrary value.  */
+
+#ifndef fp_barrierf
+#define fp_barrierf fp_barrierf
+static inline float fp_barrierf(float x)
+{
+	volatile float y = x;
+	return y;
+}
+#endif
+
+#ifndef fp_barrier
+#define fp_barrier fp_barrier
+static inline double fp_barrier(double x)
+{
+	volatile double y = x;
+	return y;
+}
+#endif
+
+#ifndef fp_barrierl
+#define fp_barrierl fp_barrierl
+static inline long double fp_barrierl(long double x)
+{
+	volatile long double y = x;
+	return y;
+}
+#endif
+
+/* fp_force_eval ensures that the input value is computed when that's
+   otherwise unused.  To prevent the constant folding of the input
+   expression, an additional fp_barrier may be needed or a compilation
+   mode that does so (e.g. -frounding-math in gcc). Then it can be
+   used to evaluate an expression for its fenv side-effects only.   */
+
+#ifndef fp_force_evalf
+#define fp_force_evalf fp_force_evalf
+static inline void fp_force_evalf(float x)
+{
+	volatile float y;
+	y = x;
+}
+#endif
+
+#ifndef fp_force_eval
+#define fp_force_eval fp_force_eval
+static inline void fp_force_eval(double x)
+{
+	volatile double y;
+	y = x;
+}
+#endif
+
+#ifndef fp_force_evall
+#define fp_force_evall fp_force_evall
+static inline void fp_force_evall(long double x)
+{
+	volatile long double y;
+	y = x;
+}
+#endif
+
+#define FORCE_EVAL(x) do {                        \
+	if (sizeof(x) == sizeof(float)) {         \
+		fp_force_evalf(x);                \
+	} else if (sizeof(x) == sizeof(double)) { \
+		fp_force_eval(x);                 \
+	} else {                                  \
+		fp_force_evall(x);                \
+	}                                         \
+} while(0)
+
+#define asuint(f) ((union{float _f; uint32_t _i;}){f})._i
+#define asfloat(i) ((union{uint32_t _i; float _f;}){i})._f
+#define asuint64(f) ((union{double _f; uint64_t _i;}){f})._i
+#define asdouble(i) ((union{uint64_t _i; double _f;}){i})._f
+
+#define EXTRACT_WORDS(hi,lo,d)                    \
+do {                                              \
+  uint64_t __u = asuint64(d);                     \
+  (hi) = __u >> 32;                               \
+  (lo) = (uint32_t)__u;                           \
+} while (0)
+
+#define GET_HIGH_WORD(hi,d)                       \
+do {                                              \
+  (hi) = asuint64(d) >> 32;                       \
+} while (0)
+
+#define GET_LOW_WORD(lo,d)                        \
+do {                                              \
+  (lo) = (uint32_t)asuint64(d);                   \
+} while (0)
+
+#define INSERT_WORDS(d,hi,lo)                     \
+do {                                              \
+  (d) = asdouble(((uint64_t)(hi)<<32) | (uint32_t)(lo)); \
+} while (0)
+
+#define SET_HIGH_WORD(d,hi)                       \
+  INSERT_WORDS(d, hi, (uint32_t)asuint64(d))
+
+#define SET_LOW_WORD(d,lo)                        \
+  INSERT_WORDS(d, asuint64(d)>>32, lo)
+
+#define GET_FLOAT_WORD(w,d)                       \
+do {                                              \
+  (w) = asuint(d);                                \
+} while (0)
+
+#define SET_FLOAT_WORD(d,w)                       \
+do {                                              \
+  (d) = asfloat(w);                               \
+} while (0)
+
+hidden int    __rem_pio2_large(double*,double*,int,int,int);
+
+hidden int    __rem_pio2(double,double*);
+hidden double __sin(double,double,int);
+hidden double __cos(double,double);
+hidden double __tan(double,double,int);
+hidden double __expo2(double);
+
+hidden int    __rem_pio2f(float,double*);
+hidden float  __sindf(double);
+hidden float  __cosdf(double);
+hidden float  __tandf(double,int);
+hidden float  __expo2f(float);
+
+hidden int __rem_pio2l(long double, long double *);
+hidden long double __sinl(long double, long double, int);
+hidden long double __cosl(long double, long double);
+hidden long double __tanl(long double, long double, int);
+
+hidden long double __polevll(long double, const long double *, int);
+hidden long double __p1evll(long double, const long double *, int);
+
+hidden double __lgamma_r(double, int *);
+hidden float __lgammaf_r(float, int *);
+
+/* error handling functions */
+hidden float __math_xflowf(uint32_t, float);
+hidden float __math_uflowf(uint32_t);
+hidden float __math_oflowf(uint32_t);
+hidden float __math_divzerof(uint32_t);
+hidden float __math_invalidf(float);
+hidden double __math_xflow(uint32_t, double);
+hidden double __math_uflow(uint32_t);
+hidden double __math_oflow(uint32_t);
+hidden double __math_divzero(uint32_t);
+hidden double __math_invalid(double);
+
+#endif
diff --git a/tools/usb_test.py b/tools/usb_test.py
new file mode 100644
index 0000000..0f61593
--- /dev/null
+++ b/tools/usb_test.py
@@ -0,0 +1,188 @@
+#!/usr/bin/env python
+
+import time
+from pprint import pprint
+from enum import Enum
+from functools import cache
+from dataclasses import dataclass, fields, astuple
+import struct
+import binascii
+
+import numpy as np
+import click
+import serial
+from cobs import cobs
+
+class CobsSerial:
+    def __init__(self, port, timeout):
+        self.ser = serial.Serial(port, timeout=timeout)
+        self.ser.flushOutput()
+        self.ser.flushInput()
+        self.ser.write(bytes([0])) # synchronize
+        self.ser.flushOutput()
+
+    def write_packet(self, data):
+        self.ser.write(cobs.encode(data))
+        self.ser.write(bytes([0]))
+        self.ser.flushOutput()
+
+    def read_packet(self):
+        data = b''
+        while (b := self.ser.read(1)):
+            if b[0] == 0:
+                break
+            data += b
+
+        if data:
+            return parse_packet(cobs.decode(data))
+        else:
+            return None
+
+    def command(self, command, args=b''):
+        self.write_packet(bytes([command.value]) + args)
+        return self.read_packet()
+
+
+class SerializableEnum(Enum):
+    def __int__(self):
+        return self.value
+
+class PacketType(SerializableEnum):
+    USBP_GET_STATUS = 0
+    USBP_GET_MEASUREMENTS = 1
+    USBP_SET_MOTOR = 2
+
+class ErrorCode(Enum):
+    ERR_SUCCESS = 0
+    ERR_TIMEOUT = 1
+    ERR_PHYSICAL_LAYER = 2
+    ERR_FRAMING = 3
+    ERR_PROTOCOL = 4
+    ERR_DMA = 5
+    ERR_BUSY = 6
+    ERR_BUFFER_OVERFLOW = 7
+    ERR_RX_OVERRUN = 8
+    ERR_TX_OVERRUN = 9
+
+class BoardConfig(Enum):
+    BCFG_UNCONFIGURED = 0
+    BCFG_DISPLAY = 1
+    BCFG_MOTOR = 2
+    BCFG_MEAS = 3
+
+class Serialized:
+    @classmethod
+    def deserialize(kls, data):
+        fields = struct.unpack(kls._struct_format(), data)
+        mapped = [cast(val) for cast, val in zip(kls._struct_casts(), fields)]
+        return kls(*mapped)
+
+    def serialize(self):
+        mapped = [uncast(val) for uncast, val in zip(self._struct_uncasts(), astuple(self))]
+        return struct.pack(self._struct_format(), *mapped)
+
+    @classmethod
+    @cache
+    def _struct_format(kls):
+        return kls._parse_fields()[0]
+
+    @classmethod
+    @cache
+    def _struct_casts(kls):
+        return kls._parse_fields()[1]
+
+    @classmethod
+    @cache
+    def _struct_uncasts(kls):
+        return kls._parse_fields()[2]
+
+    @classmethod
+    def _parse_fields(kls):
+        fmt = '<'
+        casts = []
+        uncasts = []
+        for field in fields(kls):
+            if isinstance(field.type, tuple):
+                struct_type, cast, uncast, *_ = *field.type, int
+            else:
+                struct_type, cast, uncast = field.type, int, int
+            fmt += struct_type
+            casts.append(cast)
+            uncasts.append(uncast)
+        return fmt, casts, uncasts
+
+def timestamp(value):
+    return float(value) / 1e6
+
+@dataclass
+class StatusPacket(Serialized):
+    packet_type: ('B', PacketType)
+    sys_time_us: ('Q', timestamp)
+    has_lcd: ('B', bool)
+    has_adc: ('B', bool)
+    board_config: ('B', BoardConfig)
+    bus_addr: 'B'
+    last_uart_error: ('B', ErrorCode)
+    last_uart_error_timestamp: ('Q', timestamp)
+    last_uart_rx: ('Q', timestamp)
+    last_uart_tx: ('Q', timestamp)
+    last_bus_error: ('B', ErrorCode)
+    last_bus_error_timestamp: ('Q', timestamp)
+
+@dataclass
+class MotorPacket(Serialized):
+    packet_type: ('B', PacketType)
+    speed_rpm: 'i'
+
+def parse_packet(data):
+    packet_type = PacketType(data[0])
+    if packet_type == PacketType.USBP_GET_STATUS:
+        return StatusPacket.deserialize(data)
+    if packet_type == PacketType.USBP_GET_MEASUREMENTS:
+        return MeasurementPacket.deserialize(data)
+    else:
+        raise ValueError(f'Unsupported packet type {packet_type}')
+
+@dataclass
+class MeasurementPacket(Serialized):
+    packet_type: ('B', PacketType)
+    num_channels: 'B'
+    _num_samples_a: 'I'
+    _num_samples_b: 'I'
+    _measurements_raw: ('240s', bytes)
+
+    @property
+    def measurements(self):
+        return np.frombuffer(self._measurements_raw, np.dtype(np.int32).newbyteorder('<')).reshape([2, 2, -1])
+
+    @property
+    def num_samples(self):
+        return [self._num_samples_a, self._num_samples_b]
+
+@click.group()
+def cli():
+    pass
+
+@cli.command()
+@click.argument('port')
+@click.option('--timeout', type=float, default=1)
+def probe(port, timeout):
+    ser = CobsSerial(port, timeout)
+    pprint(ser.command(PacketType.USBP_GET_STATUS))
+    while True:
+        time.sleep(0.01)
+        packet = ser.command(PacketType.USBP_GET_MEASUREMENTS)
+        for i in range(packet.num_samples[1]):
+            print(packet.measurements[1,1,i], packet.num_samples[1])
+
+@cli.command()
+@click.argument('port')
+@click.argument('speed_rpm', type=int, default=0)
+@click.option('--timeout', type=float, default=1)
+def motor(port, speed_rpm, timeout):
+    ser = CobsSerial(port, timeout)
+    packet = MotorPacket(PacketType.USBP_SET_MOTOR, speed_rpm)
+    ser.write_packet(packet.serialize())
+
+if __name__ == '__main__':
+    cli()