def main(config_path, base_dir, target_path, modes, verbose, use_cache=True): global config # Load config with open(config_path) as f: config = yaml.load(f.read(), Loader=yaml.SafeLoader) options.initialize(config, config_path, base_dir, target_path) options.set("modes", modes) if verbose: options.set("verbose", True) with options.get_target_path().open("rb") as f2: rom_bytes = f2.read() if "sha1" in config: sha1 = hashlib.sha1(rom_bytes).hexdigest() e_sha1 = config["sha1"] if e_sha1 != sha1: log.error(f"sha1 mismatch: expected {e_sha1}, was {sha1}") # Create main output dir options.get_base_path().mkdir(parents=True, exist_ok=True) processed_segments: List[Segment] = [] seg_sizes: Dict[str, int] = {} seg_split: Dict[str, int] = {} seg_cached: Dict[str, int] = {} # Load cache if use_cache: try: with options.get_cache_path().open("rb") as f3: cache = pickle.load(f3) if verbose: log.write(f"Loaded cache ({len(cache.keys())} items)") except Exception: cache = {} else: cache = {} # invalidate entire cache if options change if use_cache and cache.get("__options__") != config.get("options"): if verbose: log.write("Options changed, invalidating cache") cache = { "__options__": config.get("options"), } # Initialize segments all_segments = initialize_segments(config["segments"]) # Load and process symbols if options.mode_active("code"): log.write("Loading and processing symbols") symbols.initialize(all_segments) # Resolve raster/palette siblings if options.mode_active("img"): palettes.initialize(all_segments) # Scan log.write("Starting scan") for segment in all_segments: typ = segment.type if segment.type == "bin" and segment.is_name_default(): typ = "unk" if typ not in seg_sizes: seg_sizes[typ] = 0 seg_split[typ] = 0 seg_cached[typ] = 0 seg_sizes[typ] += 0 if segment.size is None else segment.size if segment.should_scan(): # Check cache but don't write anything if use_cache: if segment.cache() == cache.get(segment.unique_id()): continue if segment.needs_symbols: segment_symbols, other_symbols = get_segment_symbols( segment, all_segments) segment.given_seg_symbols = segment_symbols segment.given_ext_symbols = other_symbols segment.did_run = True segment.scan(rom_bytes) processed_segments.append(segment) seg_split[typ] += 1 log.dot(status=segment.status()) # Split log.write("Starting split") for segment in all_segments: if use_cache: cached = segment.cache() if cached == cache.get(segment.unique_id()): # Cache hit seg_cached[typ] += 1 continue else: # Cache miss; split cache[segment.unique_id()] = cached if segment.should_split(): segment.split(rom_bytes) log.dot(status=segment.status()) if options.mode_active("ld"): global linker_writer linker_writer = LinkerWriter() for segment in all_segments: linker_writer.add(segment) linker_writer.save_linker_script() linker_writer.save_symbol_header() # Write undefined_funcs_auto.txt to_write = [ s for s in symbols.all_symbols if s.referenced and not s.defined and not s.dead and s.type == "func" ] if len(to_write) > 0: with open(options.get_undefined_funcs_auto_path(), "w", newline="\n") as f: for symbol in to_write: f.write(f"{symbol.name} = 0x{symbol.vram_start:X};\n") # write undefined_syms_auto.txt to_write = [ s for s in symbols.all_symbols if s.referenced and not s.defined and not s.dead and not s.type == "func" ] if len(to_write) > 0: with open(options.get_undefined_syms_auto_path(), "w", newline="\n") as f: for symbol in to_write: f.write(f"{symbol.name} = 0x{symbol.vram_start:X};\n") # print warnings during split for segment in all_segments: if len(segment.warnings) > 0: log.write( f"{Style.DIM}0x{segment.rom_start:06X}{Style.RESET_ALL} {segment.type} {Style.BRIGHT}{segment.name}{Style.RESET_ALL}:" ) for warn in segment.warnings: log.write("warning: " + warn, status="warn") log.write("") # empty line # Statistics do_statistics(seg_sizes, rom_bytes, seg_split, seg_cached) # Save cache if cache != {} and use_cache: if verbose: log.write("Writing cache") with open(options.get_cache_path(), "wb") as f4: pickle.dump(cache, f4)
def should_scan(self) -> bool: return options.mode_active( "code") and self.rom_start != "auto" and self.rom_end != "auto"
def should_split(self) -> bool: return self.extract and options.mode_active("img")
def should_split(self): return self.extract and (super().should_split() or options.mode_active("img"))
def should_split(self): return self.extract and options.mode_active("code")
def should_run(self): return self.extract and options.mode_active(self.type)
def should_scan(self) -> bool: return options.mode_active("code")
def main(config_path, base_dir, target_path, modes, verbose, use_cache=True): global config log.write(f"splat {VERSION} (powered by spimdisasm {spimdisasm.__version__})") # Load config config = {} for entry in config_path: with open(entry) as f: additional_config = yaml.load(f.read(), Loader=yaml.SafeLoader) config = merge_configs(config, additional_config) options.initialize(config, config_path, base_dir, target_path) options.set("modes", modes) if verbose: options.set("verbose", True) with options.get_target_path().open("rb") as f2: rom_bytes = f2.read() if "sha1" in config: sha1 = hashlib.sha1(rom_bytes).hexdigest() e_sha1 = config["sha1"].lower() if e_sha1 != sha1: log.error(f"sha1 mismatch: expected {e_sha1}, was {sha1}") # Create main output dir options.get_base_path().mkdir(parents=True, exist_ok=True) processed_segments: List[Segment] = [] seg_sizes: Dict[str, int] = {} seg_split: Dict[str, int] = {} seg_cached: Dict[str, int] = {} # Load cache if use_cache: try: with options.get_cache_path().open("rb") as f3: cache = pickle.load(f3) if verbose: log.write(f"Loaded cache ({len(cache.keys())} items)") except Exception: cache = {} else: cache = {} # invalidate entire cache if options change if use_cache and cache.get("__options__") != config.get("options"): if verbose: log.write("Options changed, invalidating cache") cache = { "__options__": config.get("options"), } configure_disassembler() # Initialize segments all_segments = initialize_segments(config["segments"]) # Load and process symbols symbols.initialize(all_segments) # Assign symbols to segments assign_symbols_to_segments() if options.mode_active("code"): symbols.initialize_spim_context(all_segments) # Resolve raster/palette siblings if options.mode_active("img"): palettes.initialize(all_segments) # Scan scan_bar = tqdm.tqdm(all_segments, total=len(all_segments)) for segment in scan_bar: assert isinstance(segment, Segment) scan_bar.set_description(f"Scanning {brief_seg_name(segment, 20)}") typ = segment.type if segment.type == "bin" and segment.is_name_default(): typ = "unk" if typ not in seg_sizes: seg_sizes[typ] = 0 seg_split[typ] = 0 seg_cached[typ] = 0 seg_sizes[typ] += 0 if segment.size is None else segment.size if segment.should_scan(): # Check cache but don't write anything if use_cache: if segment.cache() == cache.get(segment.unique_id()): continue segment.did_run = True segment.scan(rom_bytes) processed_segments.append(segment) seg_split[typ] += 1 # Split for segment in tqdm.tqdm( all_segments, total=len(all_segments), desc=f"Splitting {brief_seg_name(segment, 20)}", ): if use_cache: cached = segment.cache() if cached == cache.get(segment.unique_id()): # Cache hit seg_cached[typ] += 1 continue else: # Cache miss; split cache[segment.unique_id()] = cached if segment.should_split(): segment.split(rom_bytes) if options.mode_active("ld"): global linker_writer linker_writer = LinkerWriter() for segment in tqdm.tqdm( all_segments, total=len(all_segments), desc=f"Writing linker script {brief_seg_name(segment, 20)}", ): linker_writer.add(segment) linker_writer.save_linker_script() linker_writer.save_symbol_header() # write elf_sections.txt - this only lists the generated sections in the elf, not subsections # that the elf combines into one section if options.get_create_elf_section_list_auto(): section_list = "" for segment in all_segments: section_list += "." + to_cname(segment.name) + "\n" with open(options.get_elf_section_list_path(), "w", newline="\n") as f: f.write(section_list) # Write undefined_funcs_auto.txt if options.get_create_undefined_funcs_auto(): to_write = [ s for s in symbols.all_symbols if s.referenced and not s.defined and not s.dead and s.type == "func" ] to_write.sort(key=lambda x: x.vram_start) with open(options.get_undefined_funcs_auto_path(), "w", newline="\n") as f: for symbol in to_write: f.write(f"{symbol.name} = 0x{symbol.vram_start:X};\n") # write undefined_syms_auto.txt if options.get_create_undefined_syms_auto(): to_write = [ s for s in symbols.all_symbols if s.referenced and not s.defined and not s.dead and s.type not in {"func", "label", "jtbl_label"} ] to_write.sort(key=lambda x: x.vram_start) with open(options.get_undefined_syms_auto_path(), "w", newline="\n") as f: for symbol in to_write: f.write(f"{symbol.name} = 0x{symbol.vram_start:X};\n") # print warnings during split for segment in all_segments: if len(segment.warnings) > 0: log.write( f"{Style.DIM}0x{segment.rom_start:06X}{Style.RESET_ALL} {segment.type} {Style.BRIGHT}{segment.name}{Style.RESET_ALL}:" ) for warn in segment.warnings: log.write("warning: " + warn, status="warn") log.write("") # empty line # Statistics do_statistics(seg_sizes, rom_bytes, seg_split, seg_cached) # Save cache if cache != {} and use_cache: if verbose: log.write("Writing cache") with open(options.get_cache_path(), "wb") as f4: pickle.dump(cache, f4)
def should_run(self): return options.mode_active(self.type)
def should_run(self): return super().should_run() or options.mode_active("img")
def main(config_path, out_dir, target_path, modes, verbose, ignore_cache=False): # Load config with open(config_path) as f: config = yaml.safe_load(f.read()) options.initialize(config) options.set("modes", modes) options.set("verbose", verbose) if not out_dir: out_dir = options.get("out_dir") if not out_dir: print( "Error: Output dir not specified as a command line arg or via the config yaml (out_dir)" ) sys.exit(2) else: out_dir = os.path.join(Path(config_path).parent, out_dir) if not target_path: target_path = options.get("target_path") if not target_path: print( "Error: Target binary path not specified as a command line arg or via the config yaml (target_path)" ) sys.exit(2) else: target_path = os.path.join(out_dir, target_path) with open(target_path, "rb") as f: rom_bytes = f.read() # Create main output dir Path(out_dir).mkdir(parents=True, exist_ok=True) symbol_addrs_path = get_symbol_addrs_path(out_dir) undefined_syms_path = get_undefined_syms_path(out_dir) all_symbols = gather_symbols(symbol_addrs_path, undefined_syms_path) symbol_ranges = [s for s in all_symbols if s.size > 4] platform = get_platform() processed_segments = [] ld_sections = [] seg_sizes = {} seg_split = {} seg_cached = {} # Load cache cache_path = get_cache_path(out_dir) try: with open(cache_path, "rb") as f: cache = pickle.load(f) except Exception: cache = {} # Initialize segments all_segments = initialize_segments(config_path, config["segments"]) for segment in all_segments: if platform == "n64" and type( segment) == N64SegCode: # remove special-case sometime segment_symbols, other_symbols = get_segment_symbols( segment, all_symbols, all_segments) segment.seg_symbols = segment_symbols segment.ext_symbols = other_symbols segment.all_symbols = all_symbols segment.symbol_ranges = symbol_ranges typ = segment.type if segment.type == "bin" and segment.is_name_default(): typ = "unk" if typ not in seg_sizes: seg_sizes[typ] = 0 seg_split[typ] = 0 seg_cached[typ] = 0 seg_sizes[typ] += segment.size if len(segment.errors) == 0: if segment.should_run(): # Check cache cached = segment.cache() if not ignore_cache and cached == cache.get( segment.unique_id()): # Cache hit seg_cached[typ] += 1 else: # Cache miss; split cache[segment.unique_id()] = cached segment.did_run = True segment.split(rom_bytes, out_dir) if len(segment.errors) == 0: processed_segments.append(segment) seg_split[typ] += 1 log.dot(status=segment.status()) ld_sections.append(segment.get_ld_section()) for segment in processed_segments: segment.postsplit(processed_segments) log.dot(status=segment.status()) # Write ldscript if options.mode_active("ld") and not options.get("skip_ld"): if verbose: log.write(f"saving {config['basename']}.ld") write_ldscript(config['basename'], out_dir, ld_sections) undefined_syms_to_write = [ s for s in all_symbols if s.referenced and not s.defined and not s.type == "func" ] undefined_funcs_to_write = [ s for s in all_symbols if s.referenced and not s.defined and s.type == "func" ] # Write undefined_funcs_auto.txt undefined_funcs_auto_path = get_undefined_funcs_auto_path(out_dir) to_write = undefined_funcs_to_write if len(to_write) > 0: with open(undefined_funcs_auto_path, "w", newline="\n") as f: for symbol in to_write: f.write(f"{symbol.name} = 0x{symbol.vram_start:X};\n") # write undefined_syms_auto.txt undefined_syms_auto_path = get_undefined_syms_auto_path(out_dir) to_write = undefined_syms_to_write if len(to_write) > 0: with open(undefined_syms_auto_path, "w", newline="\n") as f: for symbol in to_write: f.write(f"{symbol.name} = 0x{symbol.vram_start:X};\n") # print warnings during split/postsplit for segment in all_segments: if len(segment.warnings) > 0: log.write( f"{Style.DIM}0x{segment.rom_start:06X}{Style.RESET_ALL} {segment.type} {Style.BRIGHT}{segment.name}{Style.RESET_ALL}:" ) for warn in segment.warnings: log.write("warning: " + warn, status="warn") log.write("") # empty line # Statistics unk_size = seg_sizes.get("unk", 0) rest_size = 0 total_size = len(rom_bytes) for typ in seg_sizes: if typ != "unk": rest_size += seg_sizes[typ] assert (unk_size + rest_size == total_size) known_ratio = rest_size / total_size unk_ratio = unk_size / total_size log.write( f"Split {fmt_size(rest_size)} ({known_ratio:.2%}) in defined segments") for typ in seg_sizes: if typ != "unk": tmp_size = seg_sizes[typ] tmp_ratio = tmp_size / total_size log.write( f"{typ:>20}: {fmt_size(tmp_size):>8} ({tmp_ratio:.2%}) {Fore.GREEN}{seg_split[typ]} split{Style.RESET_ALL}, {Style.DIM}{seg_cached[typ]} cached" ) log.write( f"{'unknown':>20}: {fmt_size(unk_size):>8} ({unk_ratio:.2%}) from unknown bin files" ) # Save cache if cache != {}: if verbose: print("Writing cache") with open(cache_path, "wb") as f: pickle.dump(cache, f) return 0 # no error
def should_run(self): return super().should_run() or ( options.mode_active("img") or options.mode_active("ci4") or options.mode_active("ci8") or options.mode_active("i4") or options.mode_active("i8") or options.mode_active("ia4") or options.mode_active("ia8") or options.mode_active("ia16"))