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/Errno.h"
21 #include "llvm/System/Path.h"
22 #include "llvm/System/Program.h"
23 
24 #include <cerrno>
25 #include <cstdlib>
26 #include <cstring>
27 #include <fstream>
28 #include <list>
29 #include <vector>
30 
31 using namespace llvm;
32 
33 namespace {
34   ld_plugin_status discard_message(int level, const char *format, ...) {
35     // Die loudly. Recent versions of Gold pass ld_plugin_message as the first
36     // callback in the transfer vector. This should never be called.
37     abort();
38   }
39 
40   ld_plugin_add_symbols add_symbols = NULL;
41   ld_plugin_get_symbols get_symbols = NULL;
42   ld_plugin_add_input_file add_input_file = NULL;
43   ld_plugin_message message = discard_message;
44 
45   int api_version = 0;
46   int gold_version = 0;
47 
48   bool generate_api_file = false;
49   const char *as_path = NULL;
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::list<claimed_file> Modules;
59   std::vector<sys::Path> Cleanup;
60 }
61 
62 ld_plugin_status claim_file_hook(const ld_plugin_input_file *file,
63                                  int *claimed);
64 ld_plugin_status all_symbols_read_hook(void);
65 ld_plugin_status cleanup_hook(void);
66 
67 extern "C" ld_plugin_status onload(ld_plugin_tv *tv);
68 ld_plugin_status onload(ld_plugin_tv *tv) {
69   // We're given a pointer to the first transfer vector. We read through them
70   // until we find one where tv_tag == LDPT_NULL. The REGISTER_* tagged values
71   // contain pointers to functions that we need to call to register our own
72   // hooks. The others are addresses of functions we can use to call into gold
73   // for services.
74 
75   bool registeredClaimFile = false;
76   bool registeredAllSymbolsRead = false;
77   bool registeredCleanup = false;
78 
79   for (; tv->tv_tag != LDPT_NULL; ++tv) {
80     switch (tv->tv_tag) {
81       case LDPT_API_VERSION:
82         api_version = tv->tv_u.tv_val;
83         break;
84       case LDPT_GOLD_VERSION:  // major * 100 + minor
85         gold_version = tv->tv_u.tv_val;
86         break;
87       case LDPT_LINKER_OUTPUT:
88         switch (tv->tv_u.tv_val) {
89           case LDPO_REL:  // .o
90           case LDPO_DYN:  // .so
91             output_type = LTO_CODEGEN_PIC_MODEL_DYNAMIC;
92             break;
93           case LDPO_EXEC:  // .exe
94             output_type = LTO_CODEGEN_PIC_MODEL_STATIC;
95             break;
96           default:
97             (*message)(LDPL_ERROR, "Unknown output file type %d",
98                        tv->tv_u.tv_val);
99             return LDPS_ERR;
100         }
101         // TODO: add an option to disable PIC.
102         //output_type = LTO_CODEGEN_PIC_MODEL_DYNAMIC_NO_PIC;
103         break;
104       case LDPT_OPTION:
105         if (strcmp("generate-api-file", tv->tv_u.tv_string) == 0) {
106           generate_api_file = true;
107         } else if (strncmp("as=", tv->tv_u.tv_string, 3) == 0) {
108           if (as_path) {
109             (*message)(LDPL_WARNING, "Path to as specified twice. "
110                        "Discarding %s", tv->tv_u.tv_string);
111           } else {
112             as_path = strdup(tv->tv_u.tv_string + 3);
113           }
114         } else {
115           (*message)(LDPL_WARNING, "Ignoring flag %s", tv->tv_u.tv_string);
116         }
117         break;
118       case LDPT_REGISTER_CLAIM_FILE_HOOK: {
119         ld_plugin_register_claim_file callback;
120         callback = tv->tv_u.tv_register_claim_file;
121 
122         if ((*callback)(claim_file_hook) != LDPS_OK)
123           return LDPS_ERR;
124 
125         registeredClaimFile = true;
126       } break;
127       case LDPT_REGISTER_ALL_SYMBOLS_READ_HOOK: {
128         ld_plugin_register_all_symbols_read callback;
129         callback = tv->tv_u.tv_register_all_symbols_read;
130 
131         if ((*callback)(all_symbols_read_hook) != LDPS_OK)
132           return LDPS_ERR;
133 
134         registeredAllSymbolsRead = true;
135       } break;
136       case LDPT_REGISTER_CLEANUP_HOOK: {
137         ld_plugin_register_cleanup callback;
138         callback = tv->tv_u.tv_register_cleanup;
139 
140         if ((*callback)(cleanup_hook) != LDPS_OK)
141           return LDPS_ERR;
142 
143         registeredCleanup = true;
144       } break;
145       case LDPT_ADD_SYMBOLS:
146         add_symbols = tv->tv_u.tv_add_symbols;
147         break;
148       case LDPT_GET_SYMBOLS:
149         get_symbols = tv->tv_u.tv_get_symbols;
150         break;
151       case LDPT_ADD_INPUT_FILE:
152         add_input_file = tv->tv_u.tv_add_input_file;
153         break;
154       case LDPT_MESSAGE:
155         message = tv->tv_u.tv_message;
156         break;
157       default:
158         break;
159     }
160   }
161 
162   if (!registeredClaimFile) {
163     (*message)(LDPL_ERROR, "register_claim_file not passed to LLVMgold.");
164     return LDPS_ERR;
165   }
166   if (!add_symbols) {
167     (*message)(LDPL_ERROR, "add_symbols not passed to LLVMgold.");
168     return LDPS_ERR;
169   }
170 
171   return LDPS_OK;
172 }
173 
174 /// claim_file_hook - called by gold to see whether this file is one that
175 /// our plugin can handle. We'll try to open it and register all the symbols
176 /// with add_symbol if possible.
177 ld_plugin_status claim_file_hook(const ld_plugin_input_file *file,
178                                  int *claimed) {
179   void *buf = NULL;
180   if (file->offset) {
181     // Gold has found what might be IR part-way inside of a file, such as
182     // an .a archive.
183     if (lseek(file->fd, file->offset, SEEK_SET) == -1) {
184       (*message)(LDPL_ERROR,
185                  "Failed to seek to archive member of %s at offset %d: %s\n",
186                  file->name,
187                  file->offset, sys::StrError(errno).c_str());
188       return LDPS_ERR;
189     }
190     buf = malloc(file->filesize);
191     if (!buf) {
192       (*message)(LDPL_ERROR,
193                  "Failed to allocate buffer for archive member of size: %d\n",
194                  file->filesize);
195       return LDPS_ERR;
196     }
197     if (read(file->fd, buf, file->filesize) != file->filesize) {
198       (*message)(LDPL_ERROR,
199                  "Failed to read archive member of %s at offset %d: %s\n",
200                  file->name,
201                  file->offset,
202                  sys::StrError(errno).c_str());
203       free(buf);
204       return LDPS_ERR;
205     }
206     if (!lto_module_is_object_file_in_memory(buf, file->filesize)) {
207       free(buf);
208       return LDPS_OK;
209     }
210   } else if (!lto_module_is_object_file(file->name))
211     return LDPS_OK;
212 
213   *claimed = 1;
214   Modules.resize(Modules.size() + 1);
215   claimed_file &cf = Modules.back();
216 
217   cf.M = buf ? lto_module_create_from_memory(buf, file->filesize) :
218                lto_module_create(file->name);
219   free(buf);
220   if (!cf.M) {
221     (*message)(LDPL_ERROR, "Failed to create LLVM module: %s",
222                lto_get_error_message());
223     return LDPS_ERR;
224   }
225   cf.handle = file->handle;
226   unsigned sym_count = lto_module_get_num_symbols(cf.M);
227   cf.syms.reserve(sym_count);
228 
229   for (unsigned i = 0; i != sym_count; ++i) {
230     lto_symbol_attributes attrs = lto_module_get_symbol_attribute(cf.M, i);
231     if ((attrs & LTO_SYMBOL_SCOPE_MASK) == LTO_SYMBOL_SCOPE_INTERNAL)
232       continue;
233 
234     cf.syms.push_back(ld_plugin_symbol());
235     ld_plugin_symbol &sym = cf.syms.back();
236     sym.name = const_cast<char *>(lto_module_get_symbol_name(cf.M, i));
237     sym.version = NULL;
238 
239     int scope = attrs & LTO_SYMBOL_SCOPE_MASK;
240     switch (scope) {
241       case LTO_SYMBOL_SCOPE_HIDDEN:
242         sym.visibility = LDPV_HIDDEN;
243         break;
244       case LTO_SYMBOL_SCOPE_PROTECTED:
245         sym.visibility = LDPV_PROTECTED;
246         break;
247       case 0: // extern
248       case LTO_SYMBOL_SCOPE_DEFAULT:
249         sym.visibility = LDPV_DEFAULT;
250         break;
251       default:
252         (*message)(LDPL_ERROR, "Unknown scope attribute: %d", scope);
253         return LDPS_ERR;
254     }
255 
256     int definition = attrs & LTO_SYMBOL_DEFINITION_MASK;
257     switch (definition) {
258       case LTO_SYMBOL_DEFINITION_REGULAR:
259         sym.def = LDPK_DEF;
260         break;
261       case LTO_SYMBOL_DEFINITION_UNDEFINED:
262         sym.def = LDPK_UNDEF;
263         break;
264       case LTO_SYMBOL_DEFINITION_TENTATIVE:
265         sym.def = LDPK_COMMON;
266         break;
267       case LTO_SYMBOL_DEFINITION_WEAK:
268         sym.def = LDPK_WEAKDEF;
269         break;
270       case LTO_SYMBOL_DEFINITION_WEAKUNDEF:
271         sym.def = LDPK_WEAKUNDEF;
272         break;
273       default:
274         (*message)(LDPL_ERROR, "Unknown definition attribute: %d", definition);
275         return LDPS_ERR;
276     }
277 
278     // LLVM never emits COMDAT.
279     sym.size = 0;
280     sym.comdat_key = NULL;
281 
282     sym.resolution = LDPR_UNKNOWN;
283   }
284 
285   cf.syms.reserve(cf.syms.size());
286 
287   if (!cf.syms.empty()) {
288     if ((*add_symbols)(cf.handle, cf.syms.size(), &cf.syms[0]) != LDPS_OK) {
289       (*message)(LDPL_ERROR, "Unable to add symbols!");
290       return LDPS_ERR;
291     }
292   }
293 
294   return LDPS_OK;
295 }
296 
297 /// all_symbols_read_hook - gold informs us that all symbols have been read.
298 /// At this point, we use get_symbols to see if any of our definitions have
299 /// been overridden by a native object file. Then, perform optimization and
300 /// codegen.
301 ld_plugin_status all_symbols_read_hook(void) {
302   lto_code_gen_t cg = lto_codegen_create();
303 
304   for (std::list<claimed_file>::iterator I = Modules.begin(),
305        E = Modules.end(); I != E; ++I)
306     lto_codegen_add_module(cg, I->M);
307 
308   std::ofstream api_file;
309   if (generate_api_file) {
310     api_file.open("apifile.txt", std::ofstream::out | std::ofstream::trunc);
311     if (!api_file.is_open()) {
312       (*message)(LDPL_FATAL, "Unable to open apifile.txt for writing.");
313       abort();
314     }
315   }
316 
317   // If we don't preserve any symbols, libLTO will assume that all symbols are
318   // needed. Keep all symbols unless we're producing a final executable.
319   if (output_type == LTO_CODEGEN_PIC_MODEL_STATIC) {
320     bool anySymbolsPreserved = false;
321     for (std::list<claimed_file>::iterator I = Modules.begin(),
322          E = Modules.end(); I != E; ++I) {
323       (*get_symbols)(I->handle, I->syms.size(), &I->syms[0]);
324       for (unsigned i = 0, e = I->syms.size(); i != e; i++) {
325         if (I->syms[i].resolution == LDPR_PREVAILING_DEF ||
326             (I->syms[i].def == LDPK_COMMON &&
327              I->syms[i].resolution == LDPR_RESOLVED_IR)) {
328           lto_codegen_add_must_preserve_symbol(cg, I->syms[i].name);
329           anySymbolsPreserved = true;
330 
331           if (generate_api_file)
332             api_file << I->syms[i].name << "\n";
333         }
334       }
335     }
336 
337     if (generate_api_file)
338       api_file.close();
339 
340     if (!anySymbolsPreserved) {
341       // This entire file is unnecessary!
342       lto_codegen_dispose(cg);
343       return LDPS_OK;
344     }
345   }
346 
347   lto_codegen_set_pic_model(cg, output_type);
348   lto_codegen_set_debug_model(cg, LTO_DEBUG_MODEL_DWARF);
349   if (as_path) {
350     sys::Path p = sys::Program::FindProgramByName(as_path);
351     lto_codegen_set_assembler_path(cg, p.c_str());
352   }
353 
354   size_t bufsize = 0;
355   const char *buffer = static_cast<const char *>(lto_codegen_compile(cg,
356                                                                      &bufsize));
357 
358   std::string ErrMsg;
359 
360   sys::Path uniqueObjPath("/tmp/llvmgold.o");
361   if (uniqueObjPath.createTemporaryFileOnDisk(true, &ErrMsg)) {
362     (*message)(LDPL_ERROR, "%s", ErrMsg.c_str());
363     return LDPS_ERR;
364   }
365   raw_fd_ostream *objFile = new raw_fd_ostream(uniqueObjPath.c_str(),
366                                                /*Binary=*/true,
367                                                /*Force=*/true,
368                                                ErrMsg);
369   if (!ErrMsg.empty()) {
370     delete objFile;
371     (*message)(LDPL_ERROR, "%s", ErrMsg.c_str());
372     return LDPS_ERR;
373   }
374 
375   objFile->write(buffer, bufsize);
376   objFile->close();
377 
378   lto_codegen_dispose(cg);
379 
380   if ((*add_input_file)(const_cast<char*>(uniqueObjPath.c_str())) != LDPS_OK) {
381     (*message)(LDPL_ERROR, "Unable to add .o file to the link.");
382     (*message)(LDPL_ERROR, "File left behind in: %s", uniqueObjPath.c_str());
383     return LDPS_ERR;
384   }
385 
386   Cleanup.push_back(uniqueObjPath);
387 
388   return LDPS_OK;
389 }
390 
391 ld_plugin_status cleanup_hook(void) {
392   std::string ErrMsg;
393 
394   for (int i = 0, e = Cleanup.size(); i != e; ++i)
395     if (Cleanup[i].eraseFromDisk(false, &ErrMsg))
396       (*message)(LDPL_ERROR, "Failed to delete '%s': %s", Cleanup[i].c_str(),
397                  ErrMsg.c_str());
398 
399   return LDPS_OK;
400 }
401