gold-plugin.cpp revision c170f5f31cd697818bf470b57b72c11a1da78010 
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/ToolOutputFile.h"
21#include "llvm/Support/Errno.h"
22#include "llvm/Support/Path.h"
23#include "llvm/Support/Program.h"
24
25#include <cerrno>
26#include <cstdlib>
27#include <cstring>
28#include <fstream>
29#include <list>
30#include <vector>
31
32// Support Windows/MinGW crazyness.
33#ifdef _WIN32
34# include <io.h>
35# define lseek _lseek
36# define read _read
37#endif
38
39using namespace llvm;
40
41namespace {
42  ld_plugin_status discard_message(int level, const char *format, ...) {
43    // Die loudly. Recent versions of Gold pass ld_plugin_message as the first
44    // callback in the transfer vector. This should never be called.
45    abort();
46  }
47
48  ld_plugin_add_symbols add_symbols = NULL;
49  ld_plugin_get_symbols get_symbols = NULL;
50  ld_plugin_add_input_file add_input_file = NULL;
51  ld_plugin_add_input_library add_input_library = NULL;
52  ld_plugin_set_extra_library_path set_extra_library_path = NULL;
53  ld_plugin_get_view get_view = NULL;
54  ld_plugin_message message = discard_message;
55
56  int api_version = 0;
57  int gold_version = 0;
58
59  struct claimed_file {
60    void *handle;
61    std::vector<ld_plugin_symbol> syms;
62  };
63
64  lto_codegen_model output_type = LTO_CODEGEN_PIC_MODEL_STATIC;
65  std::string output_name = "";
66  std::list<claimed_file> Modules;
67  std::vector<sys::Path> Cleanup;
68  lto_code_gen_t code_gen = NULL;
69}
70
71namespace options {
72  enum generate_bc { BC_NO, BC_ALSO, BC_ONLY };
73  static bool generate_api_file = false;
74  static generate_bc generate_bc_file = BC_NO;
75  static std::string bc_path;
76  static std::string obj_path;
77  static std::string extra_library_path;
78  static std::string triple;
79  static std::string mcpu;
80  // Additional options to pass into the code generator.
81  // Note: This array will contain all plugin options which are not claimed
82  // as plugin exclusive to pass to the code generator.
83  // For example, "generate-api-file" and "as"options are for the plugin
84  // use only and will not be passed.
85  static std::vector<std::string> extra;
86
87  static void process_plugin_option(const char* opt_)
88  {
89    if (opt_ == NULL)
90      return;
91    llvm::StringRef opt = opt_;
92
93    if (opt == "generate-api-file") {
94      generate_api_file = true;
95    } else if (opt.startswith("mcpu=")) {
96      mcpu = opt.substr(strlen("mcpu="));
97    } else if (opt.startswith("extra-library-path=")) {
98      extra_library_path = opt.substr(strlen("extra_library_path="));
99    } else if (opt.startswith("mtriple=")) {
100      triple = opt.substr(strlen("mtriple="));
101    } else if (opt.startswith("obj-path=")) {
102      obj_path = opt.substr(strlen("obj-path="));
103    } else if (opt == "emit-llvm") {
104      generate_bc_file = BC_ONLY;
105    } else if (opt == "also-emit-llvm") {
106      generate_bc_file = BC_ALSO;
107    } else if (opt.startswith("also-emit-llvm=")) {
108      llvm::StringRef path = opt.substr(strlen("also-emit-llvm="));
109      generate_bc_file = BC_ALSO;
110      if (!bc_path.empty()) {
111        (*message)(LDPL_WARNING, "Path to the output IL file specified twice. "
112                   "Discarding %s", opt_);
113      } else {
114        bc_path = path;
115      }
116    } else {
117      // Save this option to pass to the code generator.
118      extra.push_back(opt);
119    }
120  }
121}
122
123static ld_plugin_status claim_file_hook(const ld_plugin_input_file *file,
124                                        int *claimed);
125static ld_plugin_status all_symbols_read_hook(void);
126static ld_plugin_status cleanup_hook(void);
127
128extern "C" ld_plugin_status onload(ld_plugin_tv *tv);
129ld_plugin_status onload(ld_plugin_tv *tv) {
130  // We're given a pointer to the first transfer vector. We read through them
131  // until we find one where tv_tag == LDPT_NULL. The REGISTER_* tagged values
132  // contain pointers to functions that we need to call to register our own
133  // hooks. The others are addresses of functions we can use to call into gold
134  // for services.
135
136  bool registeredClaimFile = false;
137
138  for (; tv->tv_tag != LDPT_NULL; ++tv) {
139    switch (tv->tv_tag) {
140      case LDPT_API_VERSION:
141        api_version = tv->tv_u.tv_val;
142        break;
143      case LDPT_GOLD_VERSION:  // major * 100 + minor
144        gold_version = tv->tv_u.tv_val;
145        break;
146      case LDPT_OUTPUT_NAME:
147        output_name = tv->tv_u.tv_string;
148        break;
149      case LDPT_LINKER_OUTPUT:
150        switch (tv->tv_u.tv_val) {
151          case LDPO_REL:  // .o
152          case LDPO_DYN:  // .so
153            output_type = LTO_CODEGEN_PIC_MODEL_DYNAMIC;
154            break;
155          case LDPO_EXEC:  // .exe
156            output_type = LTO_CODEGEN_PIC_MODEL_STATIC;
157            break;
158          default:
159            (*message)(LDPL_ERROR, "Unknown output file type %d",
160                       tv->tv_u.tv_val);
161            return LDPS_ERR;
162        }
163        // TODO: add an option to disable PIC.
164        //output_type = LTO_CODEGEN_PIC_MODEL_DYNAMIC_NO_PIC;
165        break;
166      case LDPT_OPTION:
167        options::process_plugin_option(tv->tv_u.tv_string);
168        break;
169      case LDPT_REGISTER_CLAIM_FILE_HOOK: {
170        ld_plugin_register_claim_file callback;
171        callback = tv->tv_u.tv_register_claim_file;
172
173        if ((*callback)(claim_file_hook) != LDPS_OK)
174          return LDPS_ERR;
175
176        registeredClaimFile = true;
177      } break;
178      case LDPT_REGISTER_ALL_SYMBOLS_READ_HOOK: {
179        ld_plugin_register_all_symbols_read callback;
180        callback = tv->tv_u.tv_register_all_symbols_read;
181
182        if ((*callback)(all_symbols_read_hook) != LDPS_OK)
183          return LDPS_ERR;
184
185        code_gen = lto_codegen_create();
186      } break;
187      case LDPT_REGISTER_CLEANUP_HOOK: {
188        ld_plugin_register_cleanup callback;
189        callback = tv->tv_u.tv_register_cleanup;
190
191        if ((*callback)(cleanup_hook) != LDPS_OK)
192          return LDPS_ERR;
193      } break;
194      case LDPT_ADD_SYMBOLS:
195        add_symbols = tv->tv_u.tv_add_symbols;
196        break;
197      case LDPT_GET_SYMBOLS:
198        get_symbols = tv->tv_u.tv_get_symbols;
199        break;
200      case LDPT_ADD_INPUT_FILE:
201        add_input_file = tv->tv_u.tv_add_input_file;
202        break;
203      case LDPT_ADD_INPUT_LIBRARY:
204        add_input_library = tv->tv_u.tv_add_input_file;
205        break;
206      case LDPT_SET_EXTRA_LIBRARY_PATH:
207        set_extra_library_path = tv->tv_u.tv_set_extra_library_path;
208        break;
209      case LDPT_GET_VIEW:
210        get_view = tv->tv_u.tv_get_view;
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  lto_module_t M;
238
239  if (get_view) {
240    const void *view;
241    if (get_view(file->handle, &view) != LDPS_OK) {
242      (*message)(LDPL_ERROR, "Failed to get a view of %s", file->name);
243      return LDPS_ERR;
244    }
245    M = lto_module_create_from_memory(view, file->filesize);
246  } else if (file->offset) {
247    // Gold has found what might be IR part-way inside of a file, such as
248    // an .a archive.
249    M = lto_module_create_from_fd_at_offset(file->fd, file->name, -1,
250                                            file->filesize, file->offset);
251  } else {
252    M = lto_module_create_from_fd(file->fd, file->name, file->filesize);
253  }
254  if (!M) {
255    if (const char* msg = lto_get_error_message()) {
256      (*message)(LDPL_ERROR, "Failed to create LTO module: %s", msg);
257      return LDPS_ERR;
258    }
259    return LDPS_OK;
260  }
261
262  *claimed = 1;
263  Modules.resize(Modules.size() + 1);
264  claimed_file &cf = Modules.back();
265
266  if (!options::triple.empty())
267    lto_module_set_target_triple(M, options::triple.c_str());
268
269  cf.handle = file->handle;
270  unsigned sym_count = lto_module_get_num_symbols(M);
271  cf.syms.reserve(sym_count);
272
273  for (unsigned i = 0; i != sym_count; ++i) {
274    lto_symbol_attributes attrs = lto_module_get_symbol_attribute(M, i);
275    if ((attrs & LTO_SYMBOL_SCOPE_MASK) == LTO_SYMBOL_SCOPE_INTERNAL)
276      continue;
277
278    cf.syms.push_back(ld_plugin_symbol());
279    ld_plugin_symbol &sym = cf.syms.back();
280    sym.name = const_cast<char *>(lto_module_get_symbol_name(M, i));
281    sym.name = strdup(sym.name);
282    sym.version = NULL;
283
284    int scope = attrs & LTO_SYMBOL_SCOPE_MASK;
285    switch (scope) {
286      case LTO_SYMBOL_SCOPE_HIDDEN:
287        sym.visibility = LDPV_HIDDEN;
288        break;
289      case LTO_SYMBOL_SCOPE_PROTECTED:
290        sym.visibility = LDPV_PROTECTED;
291        break;
292      case 0: // extern
293      case LTO_SYMBOL_SCOPE_DEFAULT:
294        sym.visibility = LDPV_DEFAULT;
295        break;
296      default:
297        (*message)(LDPL_ERROR, "Unknown scope attribute: %d", scope);
298        return LDPS_ERR;
299    }
300
301    int definition = attrs & LTO_SYMBOL_DEFINITION_MASK;
302    sym.comdat_key = NULL;
303    switch (definition) {
304      case LTO_SYMBOL_DEFINITION_REGULAR:
305        sym.def = LDPK_DEF;
306        break;
307      case LTO_SYMBOL_DEFINITION_UNDEFINED:
308        sym.def = LDPK_UNDEF;
309        break;
310      case LTO_SYMBOL_DEFINITION_TENTATIVE:
311        sym.def = LDPK_COMMON;
312        break;
313      case LTO_SYMBOL_DEFINITION_WEAK:
314        sym.comdat_key = sym.name;
315        sym.def = LDPK_WEAKDEF;
316        break;
317      case LTO_SYMBOL_DEFINITION_WEAKUNDEF:
318        sym.def = LDPK_WEAKUNDEF;
319        break;
320      default:
321        (*message)(LDPL_ERROR, "Unknown definition attribute: %d", definition);
322        return LDPS_ERR;
323    }
324
325    sym.size = 0;
326
327    sym.resolution = LDPR_UNKNOWN;
328  }
329
330  cf.syms.reserve(cf.syms.size());
331
332  if (!cf.syms.empty()) {
333    if ((*add_symbols)(cf.handle, cf.syms.size(), &cf.syms[0]) != LDPS_OK) {
334      (*message)(LDPL_ERROR, "Unable to add symbols!");
335      return LDPS_ERR;
336    }
337  }
338
339  if (code_gen)
340    lto_codegen_add_module(code_gen, M);
341
342  lto_module_dispose(M);
343
344  return LDPS_OK;
345}
346
347/// all_symbols_read_hook - gold informs us that all symbols have been read.
348/// At this point, we use get_symbols to see if any of our definitions have
349/// been overridden by a native object file. Then, perform optimization and
350/// codegen.
351static ld_plugin_status all_symbols_read_hook(void) {
352  std::ofstream api_file;
353  assert(code_gen);
354
355  if (options::generate_api_file) {
356    api_file.open("apifile.txt", std::ofstream::out | std::ofstream::trunc);
357    if (!api_file.is_open()) {
358      (*message)(LDPL_FATAL, "Unable to open apifile.txt for writing.");
359      abort();
360    }
361  }
362
363  // If we don't preserve any symbols, libLTO will assume that all symbols are
364  // needed. Keep all symbols unless we're producing a final executable.
365  bool anySymbolsPreserved = false;
366  for (std::list<claimed_file>::iterator I = Modules.begin(),
367         E = Modules.end(); I != E; ++I) {
368    if (I->syms.empty())
369      continue;
370    (*get_symbols)(I->handle, I->syms.size(), &I->syms[0]);
371    for (unsigned i = 0, e = I->syms.size(); i != e; i++) {
372      if (I->syms[i].resolution == LDPR_PREVAILING_DEF) {
373        lto_codegen_add_must_preserve_symbol(code_gen, I->syms[i].name);
374        anySymbolsPreserved = true;
375
376        if (options::generate_api_file)
377          api_file << I->syms[i].name << "\n";
378      }
379    }
380  }
381
382  if (options::generate_api_file)
383    api_file.close();
384
385  if (!anySymbolsPreserved) {
386    // All of the IL is unnecessary!
387    lto_codegen_dispose(code_gen);
388    return LDPS_OK;
389  }
390
391  lto_codegen_set_pic_model(code_gen, output_type);
392  lto_codegen_set_debug_model(code_gen, LTO_DEBUG_MODEL_DWARF);
393  if (!options::mcpu.empty())
394    lto_codegen_set_cpu(code_gen, options::mcpu.c_str());
395
396  // Pass through extra options to the code generator.
397  if (!options::extra.empty()) {
398    for (std::vector<std::string>::iterator it = options::extra.begin();
399         it != options::extra.end(); ++it) {
400      lto_codegen_debug_options(code_gen, (*it).c_str());
401    }
402  }
403
404  if (options::generate_bc_file != options::BC_NO) {
405    std::string path;
406    if (options::generate_bc_file == options::BC_ONLY)
407      path = output_name;
408    else if (!options::bc_path.empty())
409      path = options::bc_path;
410    else
411      path = output_name + ".bc";
412    bool err = lto_codegen_write_merged_modules(code_gen, path.c_str());
413    if (err)
414      (*message)(LDPL_FATAL, "Failed to write the output file.");
415    if (options::generate_bc_file == options::BC_ONLY)
416      exit(0);
417  }
418  const char *objPath;
419  if (lto_codegen_compile_to_file(code_gen, &objPath)) {
420    (*message)(LDPL_ERROR, "Could not produce a combined object file\n");
421  }
422
423  lto_codegen_dispose(code_gen);
424  for (std::list<claimed_file>::iterator I = Modules.begin(),
425         E = Modules.end(); I != E; ++I) {
426    for (unsigned i = 0; i != I->syms.size(); ++i) {
427      ld_plugin_symbol &sym = I->syms[i];
428      free(sym.name);
429    }
430  }
431
432  if ((*add_input_file)(objPath) != LDPS_OK) {
433    (*message)(LDPL_ERROR, "Unable to add .o file to the link.");
434    (*message)(LDPL_ERROR, "File left behind in: %s", objPath);
435    return LDPS_ERR;
436  }
437
438  if (!options::extra_library_path.empty() &&
439      set_extra_library_path(options::extra_library_path.c_str()) != LDPS_OK) {
440    (*message)(LDPL_ERROR, "Unable to set the extra library path.");
441    return LDPS_ERR;
442  }
443
444  if (options::obj_path.empty())
445    Cleanup.push_back(sys::Path(objPath));
446
447  return LDPS_OK;
448}
449
450static ld_plugin_status cleanup_hook(void) {
451  std::string ErrMsg;
452
453  for (int i = 0, e = Cleanup.size(); i != e; ++i)
454    if (Cleanup[i].eraseFromDisk(false, &ErrMsg))
455      (*message)(LDPL_ERROR, "Failed to delete '%s': %s", Cleanup[i].c_str(),
456                 ErrMsg.c_str());
457
458  return LDPS_OK;
459}
460