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 #include "llvm/System/Program.h"
22 
23 #include <cerrno>
24 #include <cstdlib>
25 #include <cstring>
26 #include <fstream>
27 #include <list>
28 #include <vector>
29 
30 using namespace llvm;
31 
32 namespace {
33   ld_plugin_status discard_message(int level, const char *format, ...) {
34     // Die loudly. Recent versions of Gold pass ld_plugin_message as the first
35     // callback in the transfer vector. This should never be called.
36     abort();
37   }
38 
39   ld_plugin_add_symbols add_symbols = NULL;
40   ld_plugin_get_symbols get_symbols = NULL;
41   ld_plugin_add_input_file add_input_file = NULL;
42   ld_plugin_message message = discard_message;
43 
44   int api_version = 0;
45   int gold_version = 0;
46 
47   bool generate_api_file = false;
48   const char *as_path = NULL;
49 
50   struct claimed_file {
51     lto_module_t M;
52     void *handle;
53     std::vector<ld_plugin_symbol> syms;
54   };
55 
56   lto_codegen_model output_type = LTO_CODEGEN_PIC_MODEL_STATIC;
57   std::list<claimed_file> Modules;
58   std::vector<sys::Path> Cleanup;
59 }
60 
61 ld_plugin_status claim_file_hook(const ld_plugin_input_file *file,
62                                  int *claimed);
63 ld_plugin_status all_symbols_read_hook(void);
64 ld_plugin_status cleanup_hook(void);
65 
66 extern "C" ld_plugin_status onload(ld_plugin_tv *tv);
67 ld_plugin_status onload(ld_plugin_tv *tv) {
68   // We're given a pointer to the first transfer vector. We read through them
69   // until we find one where tv_tag == LDPT_NULL. The REGISTER_* tagged values
70   // contain pointers to functions that we need to call to register our own
71   // hooks. The others are addresses of functions we can use to call into gold
72   // for services.
73 
74   bool registeredClaimFile = false;
75   bool registeredAllSymbolsRead = false;
76   bool registeredCleanup = false;
77 
78   for (; tv->tv_tag != LDPT_NULL; ++tv) {
79     switch (tv->tv_tag) {
80       case LDPT_API_VERSION:
81         api_version = tv->tv_u.tv_val;
82         break;
83       case LDPT_GOLD_VERSION:  // major * 100 + minor
84         gold_version = tv->tv_u.tv_val;
85         break;
86       case LDPT_LINKER_OUTPUT:
87         switch (tv->tv_u.tv_val) {
88           case LDPO_REL:  // .o
89           case LDPO_DYN:  // .so
90             output_type = LTO_CODEGEN_PIC_MODEL_DYNAMIC;
91             break;
92           case LDPO_EXEC:  // .exe
93             output_type = LTO_CODEGEN_PIC_MODEL_STATIC;
94             break;
95           default:
96             (*message)(LDPL_ERROR, "Unknown output file type %d",
97                        tv->tv_u.tv_val);
98             return LDPS_ERR;
99         }
100         // TODO: add an option to disable PIC.
101         //output_type = LTO_CODEGEN_PIC_MODEL_DYNAMIC_NO_PIC;
102         break;
103       case LDPT_OPTION:
104         if (strcmp("generate-api-file", tv->tv_u.tv_string) == 0) {
105           generate_api_file = true;
106         } else if (strncmp("as=", tv->tv_u.tv_string, 3) == 0) {
107           if (as_path) {
108             (*message)(LDPL_WARNING, "Path to as specified twice. "
109                        "Discarding %s", tv->tv_u.tv_string);
110           } else {
111             as_path = strdup(tv->tv_u.tv_string + 3);
112           }
113         } else {
114           (*message)(LDPL_WARNING, "Ignoring flag %s", tv->tv_u.tv_string);
115         }
116         break;
117       case LDPT_REGISTER_CLAIM_FILE_HOOK: {
118         ld_plugin_register_claim_file callback;
119         callback = tv->tv_u.tv_register_claim_file;
120 
121         if ((*callback)(claim_file_hook) != LDPS_OK)
122           return LDPS_ERR;
123 
124         registeredClaimFile = true;
125       } break;
126       case LDPT_REGISTER_ALL_SYMBOLS_READ_HOOK: {
127         ld_plugin_register_all_symbols_read callback;
128         callback = tv->tv_u.tv_register_all_symbols_read;
129 
130         if ((*callback)(all_symbols_read_hook) != LDPS_OK)
131           return LDPS_ERR;
132 
133         registeredAllSymbolsRead = true;
134       } break;
135       case LDPT_REGISTER_CLEANUP_HOOK: {
136         ld_plugin_register_cleanup callback;
137         callback = tv->tv_u.tv_register_cleanup;
138 
139         if ((*callback)(cleanup_hook) != LDPS_OK)
140           return LDPS_ERR;
141 
142         registeredCleanup = true;
143       } break;
144       case LDPT_ADD_SYMBOLS:
145         add_symbols = tv->tv_u.tv_add_symbols;
146         break;
147       case LDPT_GET_SYMBOLS:
148         get_symbols = tv->tv_u.tv_get_symbols;
149         break;
150       case LDPT_ADD_INPUT_FILE:
151         add_input_file = tv->tv_u.tv_add_input_file;
152         break;
153       case LDPT_MESSAGE:
154         message = tv->tv_u.tv_message;
155         break;
156       default:
157         break;
158     }
159   }
160 
161   if (!registeredClaimFile) {
162     (*message)(LDPL_ERROR, "register_claim_file not passed to LLVMgold.");
163     return LDPS_ERR;
164   }
165   if (!add_symbols) {
166     (*message)(LDPL_ERROR, "add_symbols not passed to LLVMgold.");
167     return LDPS_ERR;
168   }
169 
170   return LDPS_OK;
171 }
172 
173 /// claim_file_hook - called by gold to see whether this file is one that
174 /// our plugin can handle. We'll try to open it and register all the symbols
175 /// with add_symbol if possible.
176 ld_plugin_status claim_file_hook(const ld_plugin_input_file *file,
177                                  int *claimed) {
178   void *buf = NULL;
179   if (file->offset) {
180     // Gold has found what might be IR part-way inside of a file, such as
181     // an .a archive.
182     if (lseek(file->fd, file->offset, SEEK_SET) == -1) {
183       (*message)(LDPL_ERROR,
184                  "Failed to seek to archive member of %s at offset %d: %s\n",
185                  file->name,
186                  file->offset, strerror(errno));
187       return LDPS_ERR;
188     }
189     buf = malloc(file->filesize);
190     if (!buf) {
191       (*message)(LDPL_ERROR,
192                  "Failed to allocate buffer for archive member of size: %d\n",
193                  file->filesize);
194       return LDPS_ERR;
195     }
196     if (read(file->fd, buf, file->filesize) != file->filesize) {
197       (*message)(LDPL_ERROR,
198                  "Failed to read archive member of %s at offset %d: %s\n",
199                  file->name,
200                  file->offset,
201                  strerror(errno));
202       free(buf);
203       return LDPS_ERR;
204     }
205     if (!lto_module_is_object_file_in_memory(buf, file->filesize)) {
206       free(buf);
207       return LDPS_OK;
208     }
209   } else if (!lto_module_is_object_file(file->name))
210     return LDPS_OK;
211 
212   *claimed = 1;
213   Modules.resize(Modules.size() + 1);
214   claimed_file &cf = Modules.back();
215 
216   cf.M = buf ? lto_module_create_from_memory(buf, file->filesize) :
217                lto_module_create(file->name);
218   free(buf);
219   if (!cf.M) {
220     (*message)(LDPL_ERROR, "Failed to create LLVM module: %s",
221                lto_get_error_message());
222     return LDPS_ERR;
223   }
224   cf.handle = file->handle;
225   unsigned sym_count = lto_module_get_num_symbols(cf.M);
226   cf.syms.reserve(sym_count);
227 
228   for (unsigned i = 0; i != sym_count; ++i) {
229     lto_symbol_attributes attrs = lto_module_get_symbol_attribute(cf.M, i);
230     if ((attrs & LTO_SYMBOL_SCOPE_MASK) == LTO_SYMBOL_SCOPE_INTERNAL)
231       continue;
232 
233     cf.syms.push_back(ld_plugin_symbol());
234     ld_plugin_symbol &sym = cf.syms.back();
235     sym.name = const_cast<char *>(lto_module_get_symbol_name(cf.M, i));
236     sym.version = NULL;
237 
238     int scope = attrs & LTO_SYMBOL_SCOPE_MASK;
239     switch (scope) {
240       case LTO_SYMBOL_SCOPE_HIDDEN:
241         sym.visibility = LDPV_HIDDEN;
242         break;
243       case LTO_SYMBOL_SCOPE_PROTECTED:
244         sym.visibility = LDPV_PROTECTED;
245         break;
246       case 0: // extern
247       case LTO_SYMBOL_SCOPE_DEFAULT:
248         sym.visibility = LDPV_DEFAULT;
249         break;
250       default:
251         (*message)(LDPL_ERROR, "Unknown scope attribute: %d", scope);
252         return LDPS_ERR;
253     }
254 
255     int definition = attrs & LTO_SYMBOL_DEFINITION_MASK;
256     switch (definition) {
257       case LTO_SYMBOL_DEFINITION_REGULAR:
258         sym.def = LDPK_DEF;
259         break;
260       case LTO_SYMBOL_DEFINITION_UNDEFINED:
261         sym.def = LDPK_UNDEF;
262         break;
263       case LTO_SYMBOL_DEFINITION_TENTATIVE:
264         sym.def = LDPK_COMMON;
265         break;
266       case LTO_SYMBOL_DEFINITION_WEAK:
267         sym.def = LDPK_WEAKDEF;
268         break;
269       case LTO_SYMBOL_DEFINITION_WEAKUNDEF:
270         sym.def = LDPK_WEAKUNDEF;
271         break;
272       default:
273         (*message)(LDPL_ERROR, "Unknown definition attribute: %d", definition);
274         return LDPS_ERR;
275     }
276 
277     // LLVM never emits COMDAT.
278     sym.size = 0;
279     sym.comdat_key = NULL;
280 
281     sym.resolution = LDPR_UNKNOWN;
282   }
283 
284   cf.syms.reserve(cf.syms.size());
285 
286   if (!cf.syms.empty()) {
287     if ((*add_symbols)(cf.handle, cf.syms.size(), &cf.syms[0]) != LDPS_OK) {
288       (*message)(LDPL_ERROR, "Unable to add symbols!");
289       return LDPS_ERR;
290     }
291   }
292 
293   return LDPS_OK;
294 }
295 
296 /// all_symbols_read_hook - gold informs us that all symbols have been read.
297 /// At this point, we use get_symbols to see if any of our definitions have
298 /// been overridden by a native object file. Then, perform optimization and
299 /// codegen.
300 ld_plugin_status all_symbols_read_hook(void) {
301   lto_code_gen_t cg = lto_codegen_create();
302 
303   for (std::list<claimed_file>::iterator I = Modules.begin(),
304        E = Modules.end(); I != E; ++I)
305     lto_codegen_add_module(cg, I->M);
306 
307   std::ofstream api_file;
308   if (generate_api_file) {
309     api_file.open("apifile.txt", std::ofstream::out | std::ofstream::trunc);
310     if (!api_file.is_open()) {
311       (*message)(LDPL_FATAL, "Unable to open apifile.txt for writing.");
312       abort();
313     }
314   }
315 
316   // If we don't preserve any symbols, libLTO will assume that all symbols are
317   // needed. Keep all symbols unless we're producing a final executable.
318   if (output_type == LTO_CODEGEN_PIC_MODEL_STATIC) {
319     bool anySymbolsPreserved = false;
320     for (std::list<claimed_file>::iterator I = Modules.begin(),
321          E = Modules.end(); I != E; ++I) {
322       (*get_symbols)(I->handle, I->syms.size(), &I->syms[0]);
323       for (unsigned i = 0, e = I->syms.size(); i != e; i++) {
324         if (I->syms[i].resolution == LDPR_PREVAILING_DEF ||
325             (I->syms[i].def == LDPK_COMMON &&
326              I->syms[i].resolution == LDPR_RESOLVED_IR)) {
327           lto_codegen_add_must_preserve_symbol(cg, I->syms[i].name);
328           anySymbolsPreserved = true;
329 
330           if (generate_api_file)
331             api_file << I->syms[i].name << "\n";
332         }
333       }
334     }
335 
336     if (generate_api_file)
337       api_file.close();
338 
339     if (!anySymbolsPreserved) {
340       // This entire file is unnecessary!
341       lto_codegen_dispose(cg);
342       return LDPS_OK;
343     }
344   }
345 
346   lto_codegen_set_pic_model(cg, output_type);
347   lto_codegen_set_debug_model(cg, LTO_DEBUG_MODEL_DWARF);
348   if (as_path) {
349     sys::Path p = sys::Program::FindProgramByName(as_path);
350     lto_codegen_set_assembler_path(cg, p.c_str());
351   }
352 
353   size_t bufsize = 0;
354   const char *buffer = static_cast<const char *>(lto_codegen_compile(cg,
355                                                                      &bufsize));
356 
357   std::string ErrMsg;
358 
359   sys::Path uniqueObjPath("/tmp/llvmgold.o");
360   if (uniqueObjPath.createTemporaryFileOnDisk(true, &ErrMsg)) {
361     (*message)(LDPL_ERROR, "%s", ErrMsg.c_str());
362     return LDPS_ERR;
363   }
364   raw_fd_ostream *objFile = new raw_fd_ostream(uniqueObjPath.c_str(), true,
365                                                ErrMsg);
366   if (!ErrMsg.empty()) {
367     delete objFile;
368     (*message)(LDPL_ERROR, "%s", ErrMsg.c_str());
369     return LDPS_ERR;
370   }
371 
372   objFile->write(buffer, bufsize);
373   objFile->close();
374 
375   lto_codegen_dispose(cg);
376 
377   if ((*add_input_file)(const_cast<char*>(uniqueObjPath.c_str())) != LDPS_OK) {
378     (*message)(LDPL_ERROR, "Unable to add .o file to the link.");
379     (*message)(LDPL_ERROR, "File left behind in: %s", uniqueObjPath.c_str());
380     return LDPS_ERR;
381   }
382 
383   Cleanup.push_back(uniqueObjPath);
384 
385   return LDPS_OK;
386 }
387 
388 ld_plugin_status cleanup_hook(void) {
389   std::string ErrMsg;
390 
391   for (int i = 0, e = Cleanup.size(); i != e; ++i)
392     if (Cleanup[i].eraseFromDisk(false, &ErrMsg))
393       (*message)(LDPL_ERROR, "Failed to delete '%s': %s", Cleanup[i].c_str(),
394                  ErrMsg.c_str());
395 
396   return LDPS_OK;
397 }
398