gold-plugin.cpp revision f29140106f74d15ba357aa0a7f109adc939c3104
1//===-- gold-plugin.cpp - Plugin to gold for Link Time Optimization ------===// 2// 3// The LLVM Compiler Infrastructure 4// 5// This file is distributed under the University of Illinois Open Source 6// License. See LICENSE.TXT for details. 7// 8//===----------------------------------------------------------------------===// 9// 10// This is a gold plugin for LLVM. It provides an LLVM implementation of the 11// interface described in http://gcc.gnu.org/wiki/whopr/driver . 12// 13//===----------------------------------------------------------------------===// 14 15#include "llvm/Config/config.h" 16#include "plugin-api.h" 17 18#include "llvm-c/lto.h" 19 20#include "llvm/Support/raw_ostream.h" 21#include "llvm/System/Errno.h" 22#include "llvm/System/Path.h" 23#include "llvm/System/Program.h" 24 25#include <cerrno> 26#include <cstdlib> 27#include <cstring> 28#include <fstream> 29#include <list> 30#include <vector> 31 32using namespace llvm; 33 34namespace { 35 ld_plugin_status discard_message(int level, const char *format, ...) { 36 // Die loudly. Recent versions of Gold pass ld_plugin_message as the first 37 // callback in the transfer vector. This should never be called. 38 abort(); 39 } 40 41 ld_plugin_add_symbols add_symbols = NULL; 42 ld_plugin_get_symbols get_symbols = NULL; 43 ld_plugin_add_input_file add_input_file = NULL; 44 ld_plugin_add_input_library add_input_library = NULL; 45 ld_plugin_set_extra_library_path set_extra_library_path = NULL; 46 ld_plugin_message message = discard_message; 47 48 int api_version = 0; 49 int gold_version = 0; 50 51 struct claimed_file { 52 lto_module_t M; 53 void *handle; 54 std::vector<ld_plugin_symbol> syms; 55 }; 56 57 lto_codegen_model output_type = LTO_CODEGEN_PIC_MODEL_STATIC; 58 std::string output_name = ""; 59 std::list<claimed_file> Modules; 60 std::vector<sys::Path> Cleanup; 61} 62 63namespace options { 64 enum generate_bc { BC_NO, BC_ALSO, BC_ONLY }; 65 static bool generate_api_file = false; 66 static generate_bc generate_bc_file = BC_NO; 67 static std::string bc_path; 68 static std::string as_path; 69 static std::vector<std::string> as_args; 70 static std::vector<std::string> pass_through; 71 static std::string extra_library_path; 72 static std::string triple; 73 static std::string mcpu; 74 // Additional options to pass into the code generator. 75 // Note: This array will contain all plugin options which are not claimed 76 // as plugin exclusive to pass to the code generator. 77 // For example, "generate-api-file" and "as"options are for the plugin 78 // use only and will not be passed. 79 static std::vector<std::string> extra; 80 81 static void process_plugin_option(const char* opt_) 82 { 83 if (opt_ == NULL) 84 return; 85 llvm::StringRef opt = opt_; 86 87 if (opt == "generate-api-file") { 88 generate_api_file = true; 89 } else if (opt.startswith("mcpu=")) { 90 mcpu = opt.substr(strlen("mcpu=")); 91 } else if (opt.startswith("as=")) { 92 if (!as_path.empty()) { 93 (*message)(LDPL_WARNING, "Path to as specified twice. " 94 "Discarding %s", opt_); 95 } else { 96 as_path = opt.substr(strlen("as=")); 97 } 98 } else if (opt.startswith("as-arg=")) { 99 llvm::StringRef item = opt.substr(strlen("as-arg=")); 100 as_args.push_back(item.str()); 101 } else if (opt.startswith("extra-library-path=")) { 102 extra_library_path = opt.substr(strlen("extra_library_path=")); 103 } else if (opt.startswith("pass-through=")) { 104 llvm::StringRef item = opt.substr(strlen("pass-through=")); 105 pass_through.push_back(item.str()); 106 } else if (opt.startswith("mtriple=")) { 107 triple = opt.substr(strlen("mtriple=")); 108 } else if (opt == "emit-llvm") { 109 generate_bc_file = BC_ONLY; 110 } else if (opt == "also-emit-llvm") { 111 generate_bc_file = BC_ALSO; 112 } else if (opt.startswith("also-emit-llvm=")) { 113 llvm::StringRef path = opt.substr(strlen("also-emit-llvm=")); 114 generate_bc_file = BC_ALSO; 115 if (!bc_path.empty()) { 116 (*message)(LDPL_WARNING, "Path to the output IL file specified twice. " 117 "Discarding %s", opt_); 118 } else { 119 bc_path = path; 120 } 121 } else { 122 // Save this option to pass to the code generator. 123 extra.push_back(opt); 124 } 125 } 126} 127 128static ld_plugin_status claim_file_hook(const ld_plugin_input_file *file, 129 int *claimed); 130static ld_plugin_status all_symbols_read_hook(void); 131static ld_plugin_status cleanup_hook(void); 132 133extern "C" ld_plugin_status onload(ld_plugin_tv *tv); 134ld_plugin_status onload(ld_plugin_tv *tv) { 135 // We're given a pointer to the first transfer vector. We read through them 136 // until we find one where tv_tag == LDPT_NULL. The REGISTER_* tagged values 137 // contain pointers to functions that we need to call to register our own 138 // hooks. The others are addresses of functions we can use to call into gold 139 // for services. 140 141 bool registeredClaimFile = false; 142 143 for (; tv->tv_tag != LDPT_NULL; ++tv) { 144 switch (tv->tv_tag) { 145 case LDPT_API_VERSION: 146 api_version = tv->tv_u.tv_val; 147 break; 148 case LDPT_GOLD_VERSION: // major * 100 + minor 149 gold_version = tv->tv_u.tv_val; 150 break; 151 case LDPT_OUTPUT_NAME: 152 output_name = tv->tv_u.tv_string; 153 break; 154 case LDPT_LINKER_OUTPUT: 155 switch (tv->tv_u.tv_val) { 156 case LDPO_REL: // .o 157 case LDPO_DYN: // .so 158 output_type = LTO_CODEGEN_PIC_MODEL_DYNAMIC; 159 break; 160 case LDPO_EXEC: // .exe 161 output_type = LTO_CODEGEN_PIC_MODEL_STATIC; 162 break; 163 default: 164 (*message)(LDPL_ERROR, "Unknown output file type %d", 165 tv->tv_u.tv_val); 166 return LDPS_ERR; 167 } 168 // TODO: add an option to disable PIC. 169 //output_type = LTO_CODEGEN_PIC_MODEL_DYNAMIC_NO_PIC; 170 break; 171 case LDPT_OPTION: 172 options::process_plugin_option(tv->tv_u.tv_string); 173 break; 174 case LDPT_REGISTER_CLAIM_FILE_HOOK: { 175 ld_plugin_register_claim_file callback; 176 callback = tv->tv_u.tv_register_claim_file; 177 178 if ((*callback)(claim_file_hook) != LDPS_OK) 179 return LDPS_ERR; 180 181 registeredClaimFile = true; 182 } break; 183 case LDPT_REGISTER_ALL_SYMBOLS_READ_HOOK: { 184 ld_plugin_register_all_symbols_read callback; 185 callback = tv->tv_u.tv_register_all_symbols_read; 186 187 if ((*callback)(all_symbols_read_hook) != LDPS_OK) 188 return LDPS_ERR; 189 } break; 190 case LDPT_REGISTER_CLEANUP_HOOK: { 191 ld_plugin_register_cleanup callback; 192 callback = tv->tv_u.tv_register_cleanup; 193 194 if ((*callback)(cleanup_hook) != LDPS_OK) 195 return LDPS_ERR; 196 } break; 197 case LDPT_ADD_SYMBOLS: 198 add_symbols = tv->tv_u.tv_add_symbols; 199 break; 200 case LDPT_GET_SYMBOLS: 201 get_symbols = tv->tv_u.tv_get_symbols; 202 break; 203 case LDPT_ADD_INPUT_FILE: 204 add_input_file = tv->tv_u.tv_add_input_file; 205 break; 206 case LDPT_ADD_INPUT_LIBRARY: 207 add_input_library = tv->tv_u.tv_add_input_file; 208 break; 209 case LDPT_SET_EXTRA_LIBRARY_PATH: 210 set_extra_library_path = tv->tv_u.tv_set_extra_library_path; 211 break; 212 case LDPT_MESSAGE: 213 message = tv->tv_u.tv_message; 214 break; 215 default: 216 break; 217 } 218 } 219 220 if (!registeredClaimFile) { 221 (*message)(LDPL_ERROR, "register_claim_file not passed to LLVMgold."); 222 return LDPS_ERR; 223 } 224 if (!add_symbols) { 225 (*message)(LDPL_ERROR, "add_symbols not passed to LLVMgold."); 226 return LDPS_ERR; 227 } 228 229 return LDPS_OK; 230} 231 232/// claim_file_hook - called by gold to see whether this file is one that 233/// our plugin can handle. We'll try to open it and register all the symbols 234/// with add_symbol if possible. 235static ld_plugin_status claim_file_hook(const ld_plugin_input_file *file, 236 int *claimed) { 237 void *buf = NULL; 238 if (file->offset) { 239 // Gold has found what might be IR part-way inside of a file, such as 240 // an .a archive. 241 if (lseek(file->fd, file->offset, SEEK_SET) == -1) { 242 (*message)(LDPL_ERROR, 243 "Failed to seek to archive member of %s at offset %d: %s\n", 244 file->name, 245 file->offset, sys::StrError(errno).c_str()); 246 return LDPS_ERR; 247 } 248 buf = malloc(file->filesize); 249 if (!buf) { 250 (*message)(LDPL_ERROR, 251 "Failed to allocate buffer for archive member of size: %d\n", 252 file->filesize); 253 return LDPS_ERR; 254 } 255 if (read(file->fd, buf, file->filesize) != file->filesize) { 256 (*message)(LDPL_ERROR, 257 "Failed to read archive member of %s at offset %d: %s\n", 258 file->name, 259 file->offset, 260 sys::StrError(errno).c_str()); 261 free(buf); 262 return LDPS_ERR; 263 } 264 if (!lto_module_is_object_file_in_memory(buf, file->filesize)) { 265 free(buf); 266 return LDPS_OK; 267 } 268 } else if (!lto_module_is_object_file(file->name)) 269 return LDPS_OK; 270 271 *claimed = 1; 272 Modules.resize(Modules.size() + 1); 273 claimed_file &cf = Modules.back(); 274 275 cf.M = buf ? lto_module_create_from_memory(buf, file->filesize) : 276 lto_module_create(file->name); 277 free(buf); 278 if (!cf.M) { 279 (*message)(LDPL_ERROR, "Failed to create LLVM module: %s", 280 lto_get_error_message()); 281 return LDPS_ERR; 282 } 283 284 if (!options::triple.empty()) 285 lto_module_set_target_triple(cf.M, options::triple.c_str()); 286 287 cf.handle = file->handle; 288 unsigned sym_count = lto_module_get_num_symbols(cf.M); 289 cf.syms.reserve(sym_count); 290 291 for (unsigned i = 0; i != sym_count; ++i) { 292 lto_symbol_attributes attrs = lto_module_get_symbol_attribute(cf.M, i); 293 if ((attrs & LTO_SYMBOL_SCOPE_MASK) == LTO_SYMBOL_SCOPE_INTERNAL) 294 continue; 295 296 cf.syms.push_back(ld_plugin_symbol()); 297 ld_plugin_symbol &sym = cf.syms.back(); 298 sym.name = const_cast<char *>(lto_module_get_symbol_name(cf.M, i)); 299 sym.version = NULL; 300 301 int scope = attrs & LTO_SYMBOL_SCOPE_MASK; 302 switch (scope) { 303 case LTO_SYMBOL_SCOPE_HIDDEN: 304 sym.visibility = LDPV_HIDDEN; 305 break; 306 case LTO_SYMBOL_SCOPE_PROTECTED: 307 sym.visibility = LDPV_PROTECTED; 308 break; 309 case 0: // extern 310 case LTO_SYMBOL_SCOPE_DEFAULT: 311 sym.visibility = LDPV_DEFAULT; 312 break; 313 default: 314 (*message)(LDPL_ERROR, "Unknown scope attribute: %d", scope); 315 return LDPS_ERR; 316 } 317 318 int definition = attrs & LTO_SYMBOL_DEFINITION_MASK; 319 switch (definition) { 320 case LTO_SYMBOL_DEFINITION_REGULAR: 321 sym.def = LDPK_DEF; 322 break; 323 case LTO_SYMBOL_DEFINITION_UNDEFINED: 324 sym.def = LDPK_UNDEF; 325 break; 326 case LTO_SYMBOL_DEFINITION_TENTATIVE: 327 sym.def = LDPK_COMMON; 328 break; 329 case LTO_SYMBOL_DEFINITION_WEAK: 330 sym.def = LDPK_WEAKDEF; 331 break; 332 case LTO_SYMBOL_DEFINITION_WEAKUNDEF: 333 sym.def = LDPK_WEAKUNDEF; 334 break; 335 default: 336 (*message)(LDPL_ERROR, "Unknown definition attribute: %d", definition); 337 return LDPS_ERR; 338 } 339 340 // LLVM never emits COMDAT. 341 sym.size = 0; 342 sym.comdat_key = NULL; 343 344 sym.resolution = LDPR_UNKNOWN; 345 } 346 347 cf.syms.reserve(cf.syms.size()); 348 349 if (!cf.syms.empty()) { 350 if ((*add_symbols)(cf.handle, cf.syms.size(), &cf.syms[0]) != LDPS_OK) { 351 (*message)(LDPL_ERROR, "Unable to add symbols!"); 352 return LDPS_ERR; 353 } 354 } 355 356 return LDPS_OK; 357} 358 359/// all_symbols_read_hook - gold informs us that all symbols have been read. 360/// At this point, we use get_symbols to see if any of our definitions have 361/// been overridden by a native object file. Then, perform optimization and 362/// codegen. 363static ld_plugin_status all_symbols_read_hook(void) { 364 lto_code_gen_t cg = lto_codegen_create(); 365 366 for (std::list<claimed_file>::iterator I = Modules.begin(), 367 E = Modules.end(); I != E; ++I) 368 lto_codegen_add_module(cg, I->M); 369 370 std::ofstream api_file; 371 if (options::generate_api_file) { 372 api_file.open("apifile.txt", std::ofstream::out | std::ofstream::trunc); 373 if (!api_file.is_open()) { 374 (*message)(LDPL_FATAL, "Unable to open apifile.txt for writing."); 375 abort(); 376 } 377 } 378 379 // If we don't preserve any symbols, libLTO will assume that all symbols are 380 // needed. Keep all symbols unless we're producing a final executable. 381 bool anySymbolsPreserved = false; 382 for (std::list<claimed_file>::iterator I = Modules.begin(), 383 E = Modules.end(); I != E; ++I) { 384 (*get_symbols)(I->handle, I->syms.size(), &I->syms[0]); 385 for (unsigned i = 0, e = I->syms.size(); i != e; i++) { 386 if (I->syms[i].resolution == LDPR_PREVAILING_DEF) { 387 lto_codegen_add_must_preserve_symbol(cg, I->syms[i].name); 388 anySymbolsPreserved = true; 389 390 if (options::generate_api_file) 391 api_file << I->syms[i].name << "\n"; 392 } 393 } 394 } 395 396 if (options::generate_api_file) 397 api_file.close(); 398 399 if (!anySymbolsPreserved) { 400 // All of the IL is unnecessary! 401 lto_codegen_dispose(cg); 402 return LDPS_OK; 403 } 404 405 lto_codegen_set_pic_model(cg, output_type); 406 lto_codegen_set_debug_model(cg, LTO_DEBUG_MODEL_DWARF); 407 if (!options::as_path.empty()) { 408 sys::Path p = sys::Program::FindProgramByName(options::as_path); 409 lto_codegen_set_assembler_path(cg, p.c_str()); 410 } 411 if (!options::as_args.empty()) { 412 std::vector<const char *> as_args_p; 413 for (std::vector<std::string>::iterator I = options::as_args.begin(), 414 E = options::as_args.end(); I != E; ++I) { 415 as_args_p.push_back(I->c_str()); 416 } 417 lto_codegen_set_assembler_args(cg, &as_args_p[0], as_args_p.size()); 418 } 419 if (!options::mcpu.empty()) 420 lto_codegen_set_cpu(cg, options::mcpu.c_str()); 421 422 // Pass through extra options to the code generator. 423 if (!options::extra.empty()) { 424 for (std::vector<std::string>::iterator it = options::extra.begin(); 425 it != options::extra.end(); ++it) { 426 lto_codegen_debug_options(cg, (*it).c_str()); 427 } 428 } 429 430 431 if (options::generate_bc_file != options::BC_NO) { 432 std::string path; 433 if (options::generate_bc_file == options::BC_ONLY) 434 path = output_name; 435 else if (!options::bc_path.empty()) 436 path = options::bc_path; 437 else 438 path = output_name + ".bc"; 439 bool err = lto_codegen_write_merged_modules(cg, path.c_str()); 440 if (err) 441 (*message)(LDPL_FATAL, "Failed to write the output file."); 442 if (options::generate_bc_file == options::BC_ONLY) 443 exit(0); 444 } 445 size_t bufsize = 0; 446 const char *buffer = static_cast<const char *>(lto_codegen_compile(cg, 447 &bufsize)); 448 449 std::string ErrMsg; 450 451 sys::Path uniqueObjPath("/tmp/llvmgold.o"); 452 if (uniqueObjPath.createTemporaryFileOnDisk(true, &ErrMsg)) { 453 (*message)(LDPL_ERROR, "%s", ErrMsg.c_str()); 454 return LDPS_ERR; 455 } 456 tool_output_file objFile(uniqueObjPath.c_str(), ErrMsg, 457 raw_fd_ostream::F_Binary); 458 if (!ErrMsg.empty()) { 459 (*message)(LDPL_ERROR, "%s", ErrMsg.c_str()); 460 return LDPS_ERR; 461 } 462 463 objFile.write(buffer, bufsize); 464 objFile.close(); 465 if (objFile.has_error()) { 466 (*message)(LDPL_ERROR, "Error writing output file '%s'", 467 uniqueObjPath.c_str()); 468 objFile.clear_error(); 469 return LDPS_ERR; 470 } 471 objFile.keep(); 472 473 lto_codegen_dispose(cg); 474 475 if ((*add_input_file)(uniqueObjPath.c_str()) != LDPS_OK) { 476 (*message)(LDPL_ERROR, "Unable to add .o file to the link."); 477 (*message)(LDPL_ERROR, "File left behind in: %s", uniqueObjPath.c_str()); 478 return LDPS_ERR; 479 } 480 481 if (!options::extra_library_path.empty() && 482 set_extra_library_path(options::extra_library_path.c_str()) != LDPS_OK) { 483 (*message)(LDPL_ERROR, "Unable to set the extra library path."); 484 return LDPS_ERR; 485 } 486 487 for (std::vector<std::string>::iterator i = options::pass_through.begin(), 488 e = options::pass_through.end(); 489 i != e; ++i) { 490 std::string &item = *i; 491 const char *item_p = item.c_str(); 492 if (llvm::StringRef(item).startswith("-l")) { 493 if (add_input_library(item_p + 2) != LDPS_OK) { 494 (*message)(LDPL_ERROR, "Unable to add library to the link."); 495 return LDPS_ERR; 496 } 497 } else { 498 if (add_input_file(item_p) != LDPS_OK) { 499 (*message)(LDPL_ERROR, "Unable to add .o file to the link."); 500 return LDPS_ERR; 501 } 502 } 503 } 504 505 Cleanup.push_back(uniqueObjPath); 506 507 return LDPS_OK; 508} 509 510static ld_plugin_status cleanup_hook(void) { 511 std::string ErrMsg; 512 513 for (int i = 0, e = Cleanup.size(); i != e; ++i) 514 if (Cleanup[i].eraseFromDisk(false, &ErrMsg)) 515 (*message)(LDPL_ERROR, "Failed to delete '%s': %s", Cleanup[i].c_str(), 516 ErrMsg.c_str()); 517 518 return LDPS_OK; 519} 520