tools/gold/gold-plugin.cpp

   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 "plugin-api.h"
  16
  17 #include "llvm-c/lto.h"
  18
  19 #include "llvm/Support/raw_ostream.h"
  20 #include "llvm/System/Path.h"
  21
  22 #include <cerrno>
  23 #include <cstdlib>
  24 #include <cstring>
  25 #include <fstream>
  26 #include <list>
  27 #include <vector>
  28
  29 using namespace llvm;
  30
  31 namespace {
  32   ld_plugin_status discard_message(int level, const char *format, ...) {
  33     // Die loudly. Recent versions of Gold pass ld_plugin_message as the first
  34     // callback in the transfer vector. This should never be called.
  35     abort();
  36   }
  37
  38   ld_plugin_add_symbols add_symbols = NULL;
  39   ld_plugin_get_symbols get_symbols = NULL;
  40   ld_plugin_add_input_file add_input_file = NULL;
  41   ld_plugin_message message = discard_message;
  42
  43   int api_version = 0;
  44   int gold_version = 0;
  45
  46   bool generate_api_file = false;
  47   const char *gcc_path = NULL;
  48
  49   struct claimed_file {
  50     lto_module_t M;
  51     void *handle;
  52     std::vector<ld_plugin_symbol> syms;
  53   };
  54
  55   lto_codegen_model output_type = LTO_CODEGEN_PIC_MODEL_STATIC;
  56   std::list<claimed_file> Modules;
  57   std::vector<sys::Path> Cleanup;
  58 }
  59
  60 ld_plugin_status claim_file_hook(const ld_plugin_input_file *file,
  61                                  int *claimed);
  62 ld_plugin_status all_symbols_read_hook(void);
  63 ld_plugin_status cleanup_hook(void);
  64
  65 extern "C" ld_plugin_status onload(ld_plugin_tv *tv);
  66 ld_plugin_status onload(ld_plugin_tv *tv) {
  67   // We're given a pointer to the first transfer vector. We read through them
  68   // until we find one where tv_tag == LDPT_NULL. The REGISTER_* tagged values
  69   // contain pointers to functions that we need to call to register our own
  70   // hooks. The others are addresses of functions we can use to call into gold
  71   // for services.
  72
  73   bool registeredClaimFile = false;
  74   bool registeredAllSymbolsRead = false;
  75   bool registeredCleanup = false;
  76
  77   for (; tv->tv_tag != LDPT_NULL; ++tv) {
  78     switch (tv->tv_tag) {
  79       case LDPT_API_VERSION:
  80         api_version = tv->tv_u.tv_val;
  81         break;
  82       case LDPT_GOLD_VERSION:  // major * 100 + minor
  83         gold_version = tv->tv_u.tv_val;
  84         break;
  85       case LDPT_LINKER_OUTPUT:
  86         switch (tv->tv_u.tv_val) {
  87           case LDPO_REL:  // .o
  88           case LDPO_DYN:  // .so
  89             output_type = LTO_CODEGEN_PIC_MODEL_DYNAMIC;
  90             break;
  91           case LDPO_EXEC:  // .exe
  92             output_type = LTO_CODEGEN_PIC_MODEL_STATIC;
  93             break;
  94           default:
  95             (*message)(LDPL_ERROR, "Unknown output file type %d",
  96                        tv->tv_u.tv_val);
  97             return LDPS_ERR;
  98         }
  99         // TODO: add an option to disable PIC.
 100         //output_type = LTO_CODEGEN_PIC_MODEL_DYNAMIC_NO_PIC;
 101         break;
 102       case LDPT_OPTION:
 103         if (strcmp("generate-api-file", tv->tv_u.tv_string) == 0) {
 104           generate_api_file = true;
 105         } else if (strncmp("gcc=", tv->tv_u.tv_string, 4) == 0) {
 106           if (gcc_path) {
 107             (*message)(LDPL_WARNING, "Path to gcc specified twice. "
 108                        "Discarding %s", tv->tv_u.tv_string);
 109           } else {
 110             gcc_path = strdup(tv->tv_u.tv_string + 4);
 111           }
 112         } else {
 113           (*message)(LDPL_WARNING, "Ignoring flag %s", tv->tv_u.tv_string);
 114         }
 115         break;
 116       case LDPT_REGISTER_CLAIM_FILE_HOOK: {
 117         ld_plugin_register_claim_file callback;
 118         callback = tv->tv_u.tv_register_claim_file;
 119
 120         if ((*callback)(claim_file_hook) != LDPS_OK)
 121           return LDPS_ERR;
 122
 123         registeredClaimFile = true;
 124       } break;
 125       case LDPT_REGISTER_ALL_SYMBOLS_READ_HOOK: {
 126         ld_plugin_register_all_symbols_read callback;
 127         callback = tv->tv_u.tv_register_all_symbols_read;
 128
 129         if ((*callback)(all_symbols_read_hook) != LDPS_OK)
 130           return LDPS_ERR;
 131
 132         registeredAllSymbolsRead = true;
 133       } break;
 134       case LDPT_REGISTER_CLEANUP_HOOK: {
 135         ld_plugin_register_cleanup callback;
 136         callback = tv->tv_u.tv_register_cleanup;
 137
 138         if ((*callback)(cleanup_hook) != LDPS_OK)
 139           return LDPS_ERR;
 140
 141         registeredCleanup = true;
 142       } break;
 143       case LDPT_ADD_SYMBOLS:
 144         add_symbols = tv->tv_u.tv_add_symbols;
 145         break;
 146       case LDPT_GET_SYMBOLS:
 147         get_symbols = tv->tv_u.tv_get_symbols;
 148         break;
 149       case LDPT_ADD_INPUT_FILE:
 150         add_input_file = tv->tv_u.tv_add_input_file;
 151         break;
 152       case LDPT_MESSAGE:
 153         message = tv->tv_u.tv_message;
 154         break;
 155       default:
 156         break;
 157     }
 158   }
 159
 160   if (!registeredClaimFile) {
 161     (*message)(LDPL_ERROR, "register_claim_file not passed to LLVMgold.");
 162     return LDPS_ERR;
 163   }
 164   if (!add_symbols) {
 165     (*message)(LDPL_ERROR, "add_symbols not passed to LLVMgold.");
 166     return LDPS_ERR;
 167   }
 168
 169   return LDPS_OK;
 170 }
 171
 172 /// claim_file_hook - called by gold to see whether this file is one that
 173 /// our plugin can handle. We'll try to open it and register all the symbols
 174 /// with add_symbol if possible.
 175 ld_plugin_status claim_file_hook(const ld_plugin_input_file *file,
 176                                  int *claimed) {
 177   void *buf = NULL;
 178   if (file->offset) {
 179     // Gold has found what might be IR part-way inside of a file, such as
 180     // an .a archive.
 181     if (lseek(file->fd, file->offset, SEEK_SET) == -1) {
 182       (*message)(LDPL_ERROR,
 183                  "Failed to seek to archive member of %s at offset %d: %s\n",
 184                  file->name,
 185                  file->offset, strerror(errno));
 186       return LDPS_ERR;
 187     }
 188     buf = malloc(file->filesize);
 189     if (!buf) {
 190       (*message)(LDPL_ERROR,
 191                  "Failed to allocate buffer for archive member of size: %d\n",
 192                  file->filesize);
 193       return LDPS_ERR;
 194     }
 195     if (read(file->fd, buf, file->filesize) != file->filesize) {
 196       (*message)(LDPL_ERROR,
 197                  "Failed to read archive member of %s at offset %d: %s\n",
 198                  file->name,
 199                  file->offset,
 200                  strerror(errno));
 201       free(buf);
 202       return LDPS_ERR;
 203     }
 204     if (!lto_module_is_object_file_in_memory(buf, file->filesize)) {
 205       free(buf);
 206       return LDPS_OK;
 207     }
 208   } else if (!lto_module_is_object_file(file->name))
 209     return LDPS_OK;
 210
 211   *claimed = 1;
 212   Modules.resize(Modules.size() + 1);
 213   claimed_file &cf = Modules.back();
 214
 215   cf.M = buf ? lto_module_create_from_memory(buf, file->filesize) :
 216                lto_module_create(file->name);
 217   free(buf);
 218   if (!cf.M) {
 219     (*message)(LDPL_ERROR, "Failed to create LLVM module: %s",
 220                lto_get_error_message());
 221     return LDPS_ERR;
 222   }
 223   cf.handle = file->handle;
 224   unsigned sym_count = lto_module_get_num_symbols(cf.M);
 225   cf.syms.reserve(sym_count);
 226
 227   for (unsigned i = 0; i != sym_count; ++i) {
 228     lto_symbol_attributes attrs = lto_module_get_symbol_attribute(cf.M, i);
 229     if ((attrs & LTO_SYMBOL_SCOPE_MASK) == LTO_SYMBOL_SCOPE_INTERNAL)
 230       continue;
 231
 232     cf.syms.push_back(ld_plugin_symbol());
 233     ld_plugin_symbol &sym = cf.syms.back();
 234     sym.name = const_cast<char *>(lto_module_get_symbol_name(cf.M, i));
 235     sym.version = NULL;
 236
 237     int scope = attrs & LTO_SYMBOL_SCOPE_MASK;
 238     switch (scope) {
 239       case LTO_SYMBOL_SCOPE_HIDDEN:
 240         sym.visibility = LDPV_HIDDEN;
 241         break;
 242       case LTO_SYMBOL_SCOPE_PROTECTED:
 243         sym.visibility = LDPV_PROTECTED;
 244         break;
 245       case 0: // extern
 246       case LTO_SYMBOL_SCOPE_DEFAULT:
 247         sym.visibility = LDPV_DEFAULT;
 248         break;
 249       default:
 250         (*message)(LDPL_ERROR, "Unknown scope attribute: %d", scope);
 251         return LDPS_ERR;
 252     }
 253
 254     int definition = attrs & LTO_SYMBOL_DEFINITION_MASK;
 255     switch (definition) {
 256       case LTO_SYMBOL_DEFINITION_REGULAR:
 257         sym.def = LDPK_DEF;
 258         break;
 259       case LTO_SYMBOL_DEFINITION_UNDEFINED:
 260         sym.def = LDPK_UNDEF;
 261         break;
 262       case LTO_SYMBOL_DEFINITION_TENTATIVE:
 263         sym.def = LDPK_COMMON;
 264         break;
 265       case LTO_SYMBOL_DEFINITION_WEAK:
 266         sym.def = LDPK_WEAKDEF;
 267         break;
 268       case LTO_SYMBOL_DEFINITION_WEAKUNDEF:
 269         sym.def = LDPK_WEAKUNDEF;
 270         break;
 271       default:
 272         (*message)(LDPL_ERROR, "Unknown definition attribute: %d", definition);
 273         return LDPS_ERR;
 274     }
 275
 276     // LLVM never emits COMDAT.
 277     sym.size = 0;
 278     sym.comdat_key = NULL;
 279
 280     sym.resolution = LDPR_UNKNOWN;
 281   }
 282
 283   cf.syms.reserve(cf.syms.size());
 284
 285   if (!cf.syms.empty()) {
 286     if ((*add_symbols)(cf.handle, cf.syms.size(), &cf.syms[0]) != LDPS_OK) {
 287       (*message)(LDPL_ERROR, "Unable to add symbols!");
 288       return LDPS_ERR;
 289     }
 290   }
 291
 292   return LDPS_OK;
 293 }
 294
 295 /// all_symbols_read_hook - gold informs us that all symbols have been read.
 296 /// At this point, we use get_symbols to see if any of our definitions have
 297 /// been overridden by a native object file. Then, perform optimization and
 298 /// codegen.
 299 ld_plugin_status all_symbols_read_hook(void) {
 300   lto_code_gen_t cg = lto_codegen_create();
 301
 302   for (std::list<claimed_file>::iterator I = Modules.begin(),
 303        E = Modules.end(); I != E; ++I)
 304     lto_codegen_add_module(cg, I->M);
 305
 306   std::ofstream api_file;
 307   if (generate_api_file) {
 308     api_file.open("apifile.txt", std::ofstream::out | std::ofstream::trunc);
 309     if (!api_file.is_open()) {
 310       (*message)(LDPL_FATAL, "Unable to open apifile.txt for writing.");
 311       abort();
 312     }
 313   }
 314
 315   // If we don't preserve any symbols, libLTO will assume that all symbols are
 316   // needed. Keep all symbols unless we're producing a final executable.
 317   if (output_type == LTO_CODEGEN_PIC_MODEL_STATIC) {
 318     bool anySymbolsPreserved = false;
 319     for (std::list<claimed_file>::iterator I = Modules.begin(),
 320          E = Modules.end(); I != E; ++I) {
 321       (*get_symbols)(I->handle, I->syms.size(), &I->syms[0]);
 322       for (unsigned i = 0, e = I->syms.size(); i != e; i++) {
 323         if (I->syms[i].resolution == LDPR_PREVAILING_DEF ||
 324             (I->syms[i].def == LDPK_COMMON &&
 325              I->syms[i].resolution == LDPR_RESOLVED_IR)) {
 326           lto_codegen_add_must_preserve_symbol(cg, I->syms[i].name);
 327           anySymbolsPreserved = true;
 328
 329           if (generate_api_file)
 330             api_file << I->syms[i].name << "\n";
 331         }
 332       }
 333     }
 334
 335     if (generate_api_file)
 336       api_file.close();
 337
 338     if (!anySymbolsPreserved) {
 339       // This entire file is unnecessary!
 340       lto_codegen_dispose(cg);
 341       return LDPS_OK;
 342     }
 343   }
 344
 345   lto_codegen_set_pic_model(cg, output_type);
 346   lto_codegen_set_debug_model(cg, LTO_DEBUG_MODEL_DWARF);
 347   if (gcc_path)
 348     lto_codegen_set_gcc_path(cg, gcc_path);
 349
 350   size_t bufsize = 0;
 351   const char *buffer = static_cast<const char *>(lto_codegen_compile(cg,
 352                                                                      &bufsize));
 353
 354   std::string ErrMsg;
 355
 356   sys::Path uniqueObjPath("/tmp/llvmgold.o");
 357   if (uniqueObjPath.createTemporaryFileOnDisk(true, &ErrMsg)) {
 358     (*message)(LDPL_ERROR, "%s", ErrMsg.c_str());
 359     return LDPS_ERR;
 360   }
 361   raw_fd_ostream *objFile = new raw_fd_ostream(uniqueObjPath.c_str(), true,
 362                                                ErrMsg);
 363   if (!ErrMsg.empty()) {
 364     delete objFile;
 365     (*message)(LDPL_ERROR, "%s", ErrMsg.c_str());
 366     return LDPS_ERR;
 367   }
 368
 369   objFile->write(buffer, bufsize);
 370   objFile->close();
 371
 372   lto_codegen_dispose(cg);
 373
 374   if ((*add_input_file)(const_cast<char*>(uniqueObjPath.c_str())) != LDPS_OK) {
 375     (*message)(LDPL_ERROR, "Unable to add .o file to the link.");
 376     (*message)(LDPL_ERROR, "File left behind in: %s", uniqueObjPath.c_str());
 377     return LDPS_ERR;
 378   }
 379
 380   Cleanup.push_back(uniqueObjPath);
 381
 382   return LDPS_OK;
 383 }
 384
 385 ld_plugin_status cleanup_hook(void) {
 386   std::string ErrMsg;
 387
 388   for (int i = 0, e = Cleanup.size(); i != e; ++i)
 389     if (Cleanup[i].eraseFromDisk(false, &ErrMsg))
 390       (*message)(LDPL_ERROR, "Failed to delete '%s': %s", Cleanup[i].c_str(),
 391                  ErrMsg.c_str());
 392
 393   return LDPS_OK;
 394 }