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" // plugin-api.h requires HAVE_STDINT_H
16 #include "llvm-c/lto.h"
17 #include "llvm/ADT/OwningPtr.h"
18 #include "llvm/ADT/StringSet.h"
19 #include "llvm/Support/Errno.h"
20 #include "llvm/Support/FileSystem.h"
21 #include "llvm/Support/MemoryBuffer.h"
22 #include "llvm/Support/Path.h"
23 #include "llvm/Support/Program.h"
24 #include "llvm/Support/ToolOutputFile.h"
25 #include "llvm/Support/system_error.h"
26 #include <cerrno>
27 #include <cstdlib>
28 #include <cstring>
29 #include <fstream>
30 #include <list>
31 #include <plugin-api.h>
32 #include <vector>
33 
34 // Support Windows/MinGW crazyness.
35 #ifdef _WIN32
36 # include <io.h>
37 # define lseek _lseek
38 # define read _read
39 #endif
40 
41 #ifndef LDPO_PIE
42 // FIXME: remove this declaration when we stop maintaining Ubuntu Quantal and
43 // Precise and Debian Wheezy (binutils 2.23 is required)
44 # define LDPO_PIE 3
45 #endif
46 
47 using namespace llvm;
48 
49 namespace {
50   ld_plugin_status discard_message(int level, const char *format, ...) {
51     // Die loudly. Recent versions of Gold pass ld_plugin_message as the first
52     // callback in the transfer vector. This should never be called.
53     abort();
54   }
55 
56   ld_plugin_add_symbols add_symbols = NULL;
57   ld_plugin_get_symbols get_symbols = NULL;
58   ld_plugin_add_input_file add_input_file = NULL;
59   ld_plugin_add_input_library add_input_library = NULL;
60   ld_plugin_set_extra_library_path set_extra_library_path = NULL;
61   ld_plugin_get_view get_view = NULL;
62   ld_plugin_message message = discard_message;
63 
64   int api_version = 0;
65   int gold_version = 0;
66 
67   struct claimed_file {
68     void *handle;
69     std::vector<ld_plugin_symbol> syms;
70   };
71 
72   lto_codegen_model output_type = LTO_CODEGEN_PIC_MODEL_STATIC;
73   std::string output_name = "";
74   std::list<claimed_file> Modules;
75   std::vector<std::string> Cleanup;
76   lto_code_gen_t code_gen = NULL;
77   StringSet<> CannotBeHidden;
78 }
79 
80 namespace options {
81   enum generate_bc { BC_NO, BC_ALSO, BC_ONLY };
82   static bool generate_api_file = false;
83   static generate_bc generate_bc_file = BC_NO;
84   static std::string bc_path;
85   static std::string obj_path;
86   static std::string extra_library_path;
87   static std::string triple;
88   static std::string mcpu;
89   // Additional options to pass into the code generator.
90   // Note: This array will contain all plugin options which are not claimed
91   // as plugin exclusive to pass to the code generator.
92   // For example, "generate-api-file" and "as"options are for the plugin
93   // use only and will not be passed.
94   static std::vector<std::string> extra;
95 
96   static void process_plugin_option(const char* opt_)
97   {
98     if (opt_ == NULL)
99       return;
100     llvm::StringRef opt = opt_;
101 
102     if (opt == "generate-api-file") {
103       generate_api_file = true;
104     } else if (opt.startswith("mcpu=")) {
105       mcpu = opt.substr(strlen("mcpu="));
106     } else if (opt.startswith("extra-library-path=")) {
107       extra_library_path = opt.substr(strlen("extra_library_path="));
108     } else if (opt.startswith("mtriple=")) {
109       triple = opt.substr(strlen("mtriple="));
110     } else if (opt.startswith("obj-path=")) {
111       obj_path = opt.substr(strlen("obj-path="));
112     } else if (opt == "emit-llvm") {
113       generate_bc_file = BC_ONLY;
114     } else if (opt == "also-emit-llvm") {
115       generate_bc_file = BC_ALSO;
116     } else if (opt.startswith("also-emit-llvm=")) {
117       llvm::StringRef path = opt.substr(strlen("also-emit-llvm="));
118       generate_bc_file = BC_ALSO;
119       if (!bc_path.empty()) {
120         (*message)(LDPL_WARNING, "Path to the output IL file specified twice. "
121                    "Discarding %s", opt_);
122       } else {
123         bc_path = path;
124       }
125     } else {
126       // Save this option to pass to the code generator.
127       extra.push_back(opt);
128     }
129   }
130 }
131 
132 static ld_plugin_status claim_file_hook(const ld_plugin_input_file *file,
133                                         int *claimed);
134 static ld_plugin_status all_symbols_read_hook(void);
135 static ld_plugin_status cleanup_hook(void);
136 
137 extern "C" ld_plugin_status onload(ld_plugin_tv *tv);
138 ld_plugin_status onload(ld_plugin_tv *tv) {
139   // We're given a pointer to the first transfer vector. We read through them
140   // until we find one where tv_tag == LDPT_NULL. The REGISTER_* tagged values
141   // contain pointers to functions that we need to call to register our own
142   // hooks. The others are addresses of functions we can use to call into gold
143   // for services.
144 
145   bool registeredClaimFile = false;
146 
147   for (; tv->tv_tag != LDPT_NULL; ++tv) {
148     switch (tv->tv_tag) {
149       case LDPT_API_VERSION:
150         api_version = tv->tv_u.tv_val;
151         break;
152       case LDPT_GOLD_VERSION:  // major * 100 + minor
153         gold_version = tv->tv_u.tv_val;
154         break;
155       case LDPT_OUTPUT_NAME:
156         output_name = tv->tv_u.tv_string;
157         break;
158       case LDPT_LINKER_OUTPUT:
159         switch (tv->tv_u.tv_val) {
160           case LDPO_REL:  // .o
161           case LDPO_DYN:  // .so
162           case LDPO_PIE:  // position independent executable
163             output_type = LTO_CODEGEN_PIC_MODEL_DYNAMIC;
164             break;
165           case LDPO_EXEC:  // .exe
166             output_type = LTO_CODEGEN_PIC_MODEL_STATIC;
167             break;
168           default:
169             (*message)(LDPL_ERROR, "Unknown output file type %d",
170                        tv->tv_u.tv_val);
171             return LDPS_ERR;
172         }
173         // TODO: add an option to disable PIC.
174         //output_type = LTO_CODEGEN_PIC_MODEL_DYNAMIC_NO_PIC;
175         break;
176       case LDPT_OPTION:
177         options::process_plugin_option(tv->tv_u.tv_string);
178         break;
179       case LDPT_REGISTER_CLAIM_FILE_HOOK: {
180         ld_plugin_register_claim_file callback;
181         callback = tv->tv_u.tv_register_claim_file;
182 
183         if ((*callback)(claim_file_hook) != LDPS_OK)
184           return LDPS_ERR;
185 
186         registeredClaimFile = true;
187       } break;
188       case LDPT_REGISTER_ALL_SYMBOLS_READ_HOOK: {
189         ld_plugin_register_all_symbols_read callback;
190         callback = tv->tv_u.tv_register_all_symbols_read;
191 
192         if ((*callback)(all_symbols_read_hook) != LDPS_OK)
193           return LDPS_ERR;
194 
195         code_gen = lto_codegen_create();
196       } break;
197       case LDPT_REGISTER_CLEANUP_HOOK: {
198         ld_plugin_register_cleanup callback;
199         callback = tv->tv_u.tv_register_cleanup;
200 
201         if ((*callback)(cleanup_hook) != LDPS_OK)
202           return LDPS_ERR;
203       } break;
204       case LDPT_ADD_SYMBOLS:
205         add_symbols = tv->tv_u.tv_add_symbols;
206         break;
207       case LDPT_GET_SYMBOLS_V2:
208         get_symbols = tv->tv_u.tv_get_symbols;
209         break;
210       case LDPT_ADD_INPUT_FILE:
211         add_input_file = tv->tv_u.tv_add_input_file;
212         break;
213       case LDPT_ADD_INPUT_LIBRARY:
214         add_input_library = tv->tv_u.tv_add_input_file;
215         break;
216       case LDPT_SET_EXTRA_LIBRARY_PATH:
217         set_extra_library_path = tv->tv_u.tv_set_extra_library_path;
218         break;
219       case LDPT_GET_VIEW:
220         get_view = tv->tv_u.tv_get_view;
221         break;
222       case LDPT_MESSAGE:
223         message = tv->tv_u.tv_message;
224         break;
225       default:
226         break;
227     }
228   }
229 
230   if (!registeredClaimFile) {
231     (*message)(LDPL_ERROR, "register_claim_file not passed to LLVMgold.");
232     return LDPS_ERR;
233   }
234   if (!add_symbols) {
235     (*message)(LDPL_ERROR, "add_symbols not passed to LLVMgold.");
236     return LDPS_ERR;
237   }
238 
239   return LDPS_OK;
240 }
241 
242 /// claim_file_hook - called by gold to see whether this file is one that
243 /// our plugin can handle. We'll try to open it and register all the symbols
244 /// with add_symbol if possible.
245 static ld_plugin_status claim_file_hook(const ld_plugin_input_file *file,
246                                         int *claimed) {
247   lto_module_t M;
248   const void *view;
249   OwningPtr<MemoryBuffer> buffer;
250   if (get_view) {
251     if (get_view(file->handle, &view) != LDPS_OK) {
252       (*message)(LDPL_ERROR, "Failed to get a view of %s", file->name);
253       return LDPS_ERR;
254     }
255   } else {
256     int64_t offset = 0;
257     // Gold has found what might be IR part-way inside of a file, such as
258     // an .a archive.
259     if (file->offset) {
260       offset = file->offset;
261     }
262     if (error_code ec = MemoryBuffer::getOpenFileSlice(
263             file->fd, file->name, buffer, file->filesize, offset)) {
264       (*message)(LDPL_ERROR, ec.message().c_str());
265       return LDPS_ERR;
266     }
267     view = buffer->getBufferStart();
268   }
269 
270   if (!lto_module_is_object_file_in_memory(view, file->filesize))
271     return LDPS_OK;
272 
273   M = lto_module_create_from_memory(view, file->filesize);
274   if (!M) {
275     if (const char* msg = lto_get_error_message()) {
276       (*message)(LDPL_ERROR,
277                  "LLVM gold plugin has failed to create LTO module: %s",
278                  msg);
279       return LDPS_ERR;
280     }
281     return LDPS_OK;
282   }
283 
284   *claimed = 1;
285   Modules.resize(Modules.size() + 1);
286   claimed_file &cf = Modules.back();
287 
288   if (!options::triple.empty())
289     lto_module_set_target_triple(M, options::triple.c_str());
290 
291   cf.handle = file->handle;
292   unsigned sym_count = lto_module_get_num_symbols(M);
293   cf.syms.reserve(sym_count);
294 
295   for (unsigned i = 0; i != sym_count; ++i) {
296     lto_symbol_attributes attrs = lto_module_get_symbol_attribute(M, i);
297     if ((attrs & LTO_SYMBOL_SCOPE_MASK) == LTO_SYMBOL_SCOPE_INTERNAL)
298       continue;
299 
300     cf.syms.push_back(ld_plugin_symbol());
301     ld_plugin_symbol &sym = cf.syms.back();
302     sym.name = const_cast<char *>(lto_module_get_symbol_name(M, i));
303     sym.name = strdup(sym.name);
304     sym.version = NULL;
305 
306     int scope = attrs & LTO_SYMBOL_SCOPE_MASK;
307     bool CanBeHidden = scope == LTO_SYMBOL_SCOPE_DEFAULT_CAN_BE_HIDDEN;
308     if (!CanBeHidden)
309       CannotBeHidden.insert(sym.name);
310     switch (scope) {
311       case LTO_SYMBOL_SCOPE_HIDDEN:
312         sym.visibility = LDPV_HIDDEN;
313         break;
314       case LTO_SYMBOL_SCOPE_PROTECTED:
315         sym.visibility = LDPV_PROTECTED;
316         break;
317       case 0: // extern
318       case LTO_SYMBOL_SCOPE_DEFAULT:
319       case LTO_SYMBOL_SCOPE_DEFAULT_CAN_BE_HIDDEN:
320         sym.visibility = LDPV_DEFAULT;
321         break;
322       default:
323         (*message)(LDPL_ERROR, "Unknown scope attribute: %d", scope);
324         return LDPS_ERR;
325     }
326 
327     int definition = attrs & LTO_SYMBOL_DEFINITION_MASK;
328     sym.comdat_key = NULL;
329     switch (definition) {
330       case LTO_SYMBOL_DEFINITION_REGULAR:
331         sym.def = LDPK_DEF;
332         break;
333       case LTO_SYMBOL_DEFINITION_UNDEFINED:
334         sym.def = LDPK_UNDEF;
335         break;
336       case LTO_SYMBOL_DEFINITION_TENTATIVE:
337         sym.def = LDPK_COMMON;
338         break;
339       case LTO_SYMBOL_DEFINITION_WEAK:
340         sym.comdat_key = sym.name;
341         sym.def = LDPK_WEAKDEF;
342         break;
343       case LTO_SYMBOL_DEFINITION_WEAKUNDEF:
344         sym.def = LDPK_WEAKUNDEF;
345         break;
346       default:
347         (*message)(LDPL_ERROR, "Unknown definition attribute: %d", definition);
348         return LDPS_ERR;
349     }
350 
351     sym.size = 0;
352 
353     sym.resolution = LDPR_UNKNOWN;
354   }
355 
356   cf.syms.reserve(cf.syms.size());
357 
358   if (!cf.syms.empty()) {
359     if ((*add_symbols)(cf.handle, cf.syms.size(), &cf.syms[0]) != LDPS_OK) {
360       (*message)(LDPL_ERROR, "Unable to add symbols!");
361       return LDPS_ERR;
362     }
363   }
364 
365   if (code_gen) {
366     if (lto_codegen_add_module(code_gen, M)) {
367       (*message)(LDPL_ERROR, "Error linking module: %s",
368                  lto_get_error_message());
369       return LDPS_ERR;
370     }
371   }
372 
373   lto_module_dispose(M);
374 
375   return LDPS_OK;
376 }
377 
378 static bool mustPreserve(const claimed_file &F, int i) {
379   if (F.syms[i].resolution == LDPR_PREVAILING_DEF)
380     return true;
381   if (F.syms[i].resolution == LDPR_PREVAILING_DEF_IRONLY_EXP)
382     return CannotBeHidden.count(F.syms[i].name);
383   return false;
384 }
385 
386 /// all_symbols_read_hook - gold informs us that all symbols have been read.
387 /// At this point, we use get_symbols to see if any of our definitions have
388 /// been overridden by a native object file. Then, perform optimization and
389 /// codegen.
390 static ld_plugin_status all_symbols_read_hook(void) {
391   std::ofstream api_file;
392   assert(code_gen);
393 
394   if (options::generate_api_file) {
395     api_file.open("apifile.txt", std::ofstream::out | std::ofstream::trunc);
396     if (!api_file.is_open()) {
397       (*message)(LDPL_FATAL, "Unable to open apifile.txt for writing.");
398       abort();
399     }
400   }
401 
402   for (std::list<claimed_file>::iterator I = Modules.begin(),
403          E = Modules.end(); I != E; ++I) {
404     if (I->syms.empty())
405       continue;
406     (*get_symbols)(I->handle, I->syms.size(), &I->syms[0]);
407     for (unsigned i = 0, e = I->syms.size(); i != e; i++) {
408       if (mustPreserve(*I, i)) {
409         lto_codegen_add_must_preserve_symbol(code_gen, I->syms[i].name);
410 
411         if (options::generate_api_file)
412           api_file << I->syms[i].name << "\n";
413       }
414     }
415   }
416 
417   if (options::generate_api_file)
418     api_file.close();
419 
420   lto_codegen_set_pic_model(code_gen, output_type);
421   lto_codegen_set_debug_model(code_gen, LTO_DEBUG_MODEL_DWARF);
422   if (!options::mcpu.empty())
423     lto_codegen_set_cpu(code_gen, options::mcpu.c_str());
424 
425   // Pass through extra options to the code generator.
426   if (!options::extra.empty()) {
427     for (std::vector<std::string>::iterator it = options::extra.begin();
428          it != options::extra.end(); ++it) {
429       lto_codegen_debug_options(code_gen, (*it).c_str());
430     }
431   }
432 
433   if (options::generate_bc_file != options::BC_NO) {
434     std::string path;
435     if (options::generate_bc_file == options::BC_ONLY)
436       path = output_name;
437     else if (!options::bc_path.empty())
438       path = options::bc_path;
439     else
440       path = output_name + ".bc";
441     bool err = lto_codegen_write_merged_modules(code_gen, path.c_str());
442     if (err)
443       (*message)(LDPL_FATAL, "Failed to write the output file.");
444     if (options::generate_bc_file == options::BC_ONLY) {
445       lto_codegen_dispose(code_gen);
446       exit(0);
447     }
448   }
449 
450   std::string ObjPath;
451   {
452     const char *Temp;
453     if (lto_codegen_compile_to_file(code_gen, &Temp)) {
454       (*message)(LDPL_ERROR, "Could not produce a combined object file\n");
455     }
456     ObjPath = Temp;
457   }
458 
459   lto_codegen_dispose(code_gen);
460   for (std::list<claimed_file>::iterator I = Modules.begin(),
461          E = Modules.end(); I != E; ++I) {
462     for (unsigned i = 0; i != I->syms.size(); ++i) {
463       ld_plugin_symbol &sym = I->syms[i];
464       free(sym.name);
465     }
466   }
467 
468   if ((*add_input_file)(ObjPath.c_str()) != LDPS_OK) {
469     (*message)(LDPL_ERROR, "Unable to add .o file to the link.");
470     (*message)(LDPL_ERROR, "File left behind in: %s", ObjPath.c_str());
471     return LDPS_ERR;
472   }
473 
474   if (!options::extra_library_path.empty() &&
475       set_extra_library_path(options::extra_library_path.c_str()) != LDPS_OK) {
476     (*message)(LDPL_ERROR, "Unable to set the extra library path.");
477     return LDPS_ERR;
478   }
479 
480   if (options::obj_path.empty())
481     Cleanup.push_back(ObjPath);
482 
483   return LDPS_OK;
484 }
485 
486 static ld_plugin_status cleanup_hook(void) {
487   for (int i = 0, e = Cleanup.size(); i != e; ++i) {
488     error_code EC = sys::fs::remove(Cleanup[i]);
489     if (EC)
490       (*message)(LDPL_ERROR, "Failed to delete '%s': %s", Cleanup[i].c_str(),
491                  EC.message().c_str());
492   }
493 
494   return LDPS_OK;
495 }
496