blob: afd7fd6f93e9aa321392366fe0047eb5628192ef [file] [log] [blame]
Nick Lewycky3e62b2d2009-02-03 07:13:24 +00001//===-- 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 <cstdlib>
23#include <cstring>
24#include <list>
25#include <vector>
Torok Edwin3e5a0d82009-02-04 17:39:30 +000026#include <cerrno>
Nick Lewycky3e62b2d2009-02-03 07:13:24 +000027
28using namespace llvm;
29
30namespace {
31 ld_plugin_status discard_message(int level, const char *format, ...) {
32 // Die loudly. Recent versions of Gold pass ld_plugin_message as the first
33 // callback in the transfer vector. This should never be called.
34 abort();
35 }
36
37 ld_plugin_add_symbols add_symbols = NULL;
38 ld_plugin_get_symbols get_symbols = NULL;
39 ld_plugin_add_input_file add_input_file = NULL;
40 ld_plugin_message message = discard_message;
41
42 int api_version = 0;
43 int gold_version = 0;
44
45 struct claimed_file {
46 lto_module_t M;
47 void *handle;
Torok Edwin3e5a0d82009-02-04 17:39:30 +000048 void *buf;
Nick Lewycky3e62b2d2009-02-03 07:13:24 +000049 std::vector<ld_plugin_symbol> syms;
50 };
51
52 lto_codegen_model output_type = LTO_CODEGEN_PIC_MODEL_STATIC;
53 std::list<claimed_file> Modules;
54 std::vector<sys::Path> Cleanup;
55}
56
57ld_plugin_status claim_file_hook(const ld_plugin_input_file *file,
58 int *claimed);
59ld_plugin_status all_symbols_read_hook(void);
60ld_plugin_status cleanup_hook(void);
61
62extern "C" ld_plugin_status onload(ld_plugin_tv *tv);
63ld_plugin_status onload(ld_plugin_tv *tv) {
64 // We're given a pointer to the first transfer vector. We read through them
65 // until we find one where tv_tag == LDPT_NULL. The REGISTER_* tagged values
66 // contain pointers to functions that we need to call to register our own
67 // hooks. The others are addresses of functions we can use to call into gold
68 // for services.
69
70 bool registeredClaimFile = false;
71 bool registeredAllSymbolsRead = false;
72 bool registeredCleanup = false;
73
74 for (; tv->tv_tag != LDPT_NULL; ++tv) {
75 switch (tv->tv_tag) {
76 case LDPT_API_VERSION:
77 api_version = tv->tv_u.tv_val;
78 break;
79 case LDPT_GOLD_VERSION: // major * 100 + minor
80 gold_version = tv->tv_u.tv_val;
81 break;
82 case LDPT_LINKER_OUTPUT:
83 switch (tv->tv_u.tv_val) {
84 case LDPO_REL: // .o
85 case LDPO_DYN: // .so
86 output_type = LTO_CODEGEN_PIC_MODEL_DYNAMIC;
87 break;
88 case LDPO_EXEC: // .exe
89 output_type = LTO_CODEGEN_PIC_MODEL_STATIC;
90 break;
91 default:
92 (*message)(LDPL_ERROR, "Unknown output file type %d",
93 tv->tv_u.tv_val);
94 return LDPS_ERR;
95 }
96 // TODO: add an option to disable PIC.
97 //output_type = LTO_CODEGEN_PIC_MODEL_DYNAMIC_NO_PIC;
98 break;
99 case LDPT_OPTION:
100 (*message)(LDPL_WARNING, "Ignoring flag %s", tv->tv_u.tv_string);
101 break;
102 case LDPT_REGISTER_CLAIM_FILE_HOOK: {
103 ld_plugin_register_claim_file callback;
104 callback = tv->tv_u.tv_register_claim_file;
105
106 if ((*callback)(claim_file_hook) != LDPS_OK)
107 return LDPS_ERR;
108
109 registeredClaimFile = true;
110 } break;
111 case LDPT_REGISTER_ALL_SYMBOLS_READ_HOOK: {
112 ld_plugin_register_all_symbols_read callback;
113 callback = tv->tv_u.tv_register_all_symbols_read;
114
115 if ((*callback)(all_symbols_read_hook) != LDPS_OK)
116 return LDPS_ERR;
117
118 registeredAllSymbolsRead = true;
119 } break;
120 case LDPT_REGISTER_CLEANUP_HOOK: {
121 ld_plugin_register_cleanup callback;
122 callback = tv->tv_u.tv_register_cleanup;
123
124 if ((*callback)(cleanup_hook) != LDPS_OK)
125 return LDPS_ERR;
126
127 registeredCleanup = true;
128 } break;
129 case LDPT_ADD_SYMBOLS:
130 add_symbols = tv->tv_u.tv_add_symbols;
131 break;
132 case LDPT_GET_SYMBOLS:
133 get_symbols = tv->tv_u.tv_get_symbols;
134 break;
135 case LDPT_ADD_INPUT_FILE:
136 add_input_file = tv->tv_u.tv_add_input_file;
137 break;
138 case LDPT_MESSAGE:
139 message = tv->tv_u.tv_message;
140 break;
141 default:
142 break;
143 }
144 }
145
146 if (!registeredClaimFile || !registeredAllSymbolsRead || !registeredCleanup ||
147 !add_symbols || !get_symbols || !add_input_file) {
148 (*message)(LDPL_ERROR, "Not all hooks registered for LLVMgold.");
149 return LDPS_ERR;
150 }
151
152 return LDPS_OK;
153}
154
155/// claim_file_hook - called by gold to see whether this file is one that
156/// our plugin can handle. We'll try to open it and register all the symbols
157/// with add_symbol if possible.
158ld_plugin_status claim_file_hook(const ld_plugin_input_file *file,
159 int *claimed) {
Torok Edwin3e5a0d82009-02-04 17:39:30 +0000160 void *buf = NULL;
161 printf("%s,%d,%d\n",file->name, file->offset, file->filesize);
Nick Lewycky3e62b2d2009-02-03 07:13:24 +0000162 // If set, this means gold found IR in an ELF section. LLVM doesn't wrap its
Torok Edwin3e5a0d82009-02-04 17:39:30 +0000163 // IR in ELF, so we know it's not us. But it can also be an .a file containing
164 // LLVM IR.
165 if (file->offset) {
166 if (lseek(file->fd, file->offset, SEEK_SET) == -1) {
167 (*message)(LDPL_ERROR,
168 "Failed to seek to archive member of %s at offset %d: %s\n",
169 file->name,
170 file->offset, strerror(errno));
171 return LDPS_ERR;
172 }
173 buf = malloc(file->filesize);
174 if (!buf) {
175 (*message)(LDPL_ERROR,
176 "Failed to allocate buffer for archive member of size: %d\n",
177 file->filesize);
178 return LDPS_ERR;
179 }
180 if (read(file->fd, buf, file->filesize) != file->filesize) {
181 (*message)(LDPL_ERROR,
182 "Failed to read archive member of %s at offset %d: %s\n",
183 file->name,
184 file->offset,
185 strerror(errno));
186 free(buf);
187 return LDPS_ERR;
188 }
189 if (!lto_module_is_object_file_in_memory(buf, file->filesize)) {
190 free(buf);
191 return LDPS_OK;
192 }
193 } else if (!lto_module_is_object_file(file->name))
Nick Lewycky3e62b2d2009-02-03 07:13:24 +0000194 return LDPS_OK;
195
196 *claimed = 1;
197 Modules.resize(Modules.size() + 1);
198 claimed_file &cf = Modules.back();
199
Torok Edwin3e5a0d82009-02-04 17:39:30 +0000200 cf.M = buf ? lto_module_create_from_memory(buf, file->filesize) :
201 lto_module_create(file->name);
202 cf.buf = buf;
Nick Lewycky3e62b2d2009-02-03 07:13:24 +0000203 if (!cf.M) {
204 (*message)(LDPL_ERROR, "Failed to create LLVM module: %s",
205 lto_get_error_message());
206 return LDPS_ERR;
207 }
208 cf.handle = file->handle;
209 unsigned sym_count = lto_module_get_num_symbols(cf.M);
210 cf.syms.reserve(sym_count);
211
212 for (unsigned i = 0; i != sym_count; ++i) {
213 lto_symbol_attributes attrs = lto_module_get_symbol_attribute(cf.M, i);
214 if ((attrs & LTO_SYMBOL_SCOPE_MASK) == LTO_SYMBOL_SCOPE_INTERNAL)
215 continue;
216
217 cf.syms.push_back(ld_plugin_symbol());
218 ld_plugin_symbol &sym = cf.syms.back();
219 sym.name = const_cast<char *>(lto_module_get_symbol_name(cf.M, i));
220 sym.version = NULL;
221
222 int scope = attrs & LTO_SYMBOL_SCOPE_MASK;
223 switch (scope) {
224 case LTO_SYMBOL_SCOPE_HIDDEN:
225 sym.visibility = LDPV_HIDDEN;
226 break;
227 case LTO_SYMBOL_SCOPE_PROTECTED:
228 sym.visibility = LDPV_PROTECTED;
229 break;
230 case 0: // extern
231 case LTO_SYMBOL_SCOPE_DEFAULT:
232 sym.visibility = LDPV_DEFAULT;
233 break;
234 default:
235 (*message)(LDPL_ERROR, "Unknown scope attribute: %d", scope);
Torok Edwin3e5a0d82009-02-04 17:39:30 +0000236 free(buf);
Nick Lewycky3e62b2d2009-02-03 07:13:24 +0000237 return LDPS_ERR;
238 }
239
240 int definition = attrs & LTO_SYMBOL_DEFINITION_MASK;
241 switch (definition) {
242 case LTO_SYMBOL_DEFINITION_REGULAR:
243 sym.def = LDPK_DEF;
244 break;
245 case LTO_SYMBOL_DEFINITION_UNDEFINED:
246 sym.def = LDPK_UNDEF;
247 break;
248 case LTO_SYMBOL_DEFINITION_TENTATIVE:
249 sym.def = LDPK_COMMON;
250 break;
251 case LTO_SYMBOL_DEFINITION_WEAK:
252 sym.def = LDPK_WEAKDEF;
253 break;
254 default:
255 (*message)(LDPL_ERROR, "Unknown definition attribute: %d", definition);
Torok Edwin3e5a0d82009-02-04 17:39:30 +0000256 free(buf);
Nick Lewycky3e62b2d2009-02-03 07:13:24 +0000257 return LDPS_ERR;
258 }
259
260 // LLVM never emits COMDAT.
261 sym.size = 0;
262 sym.comdat_key = NULL;
263
264 sym.resolution = LDPR_UNKNOWN;
265 }
266
267 cf.syms.reserve(cf.syms.size());
268
269 if (!cf.syms.empty()) {
270 if ((*add_symbols)(cf.handle, cf.syms.size(), &cf.syms[0]) != LDPS_OK) {
271 (*message)(LDPL_ERROR, "Unable to add symbols!");
Torok Edwin3e5a0d82009-02-04 17:39:30 +0000272 free(buf);
Nick Lewycky3e62b2d2009-02-03 07:13:24 +0000273 return LDPS_ERR;
274 }
275 }
276
277 return LDPS_OK;
278}
279
280/// all_symbols_read_hook - gold informs us that all symbols have been read.
281/// At this point, we use get_symbols to see if any of our definitions have
282/// been overridden by a native object file. Then, perform optimization and
283/// codegen.
284ld_plugin_status all_symbols_read_hook(void) {
285 lto_code_gen_t cg = lto_codegen_create();
286
287 for (std::list<claimed_file>::iterator I = Modules.begin(),
288 E = Modules.end(); I != E; ++I)
289 lto_codegen_add_module(cg, I->M);
290
291 // If we don't preserve any symbols, libLTO will assume that all symbols are
292 // needed. Keep all symbols unless we're producing a final executable.
293 if (output_type == LTO_CODEGEN_PIC_MODEL_STATIC) {
294 bool anySymbolsPreserved = false;
295 for (std::list<claimed_file>::iterator I = Modules.begin(),
296 E = Modules.end(); I != E; ++I) {
297 (*get_symbols)(I->handle, I->syms.size(), &I->syms[0]);
298 for (unsigned i = 0, e = I->syms.size(); i != e; i++) {
299 (*message)(LDPL_WARNING, "def: %d visibility: %d resolution %d",
300 I->syms[i].def, I->syms[i].visibility, I->syms[i].resolution);
301 if (I->syms[i].resolution == LDPR_PREVAILING_DEF) {
302 lto_codegen_add_must_preserve_symbol(cg, I->syms[i].name);
303 anySymbolsPreserved = true;
304 }
305 }
306 }
307
308 if (!anySymbolsPreserved) {
309 // This entire file is unnecessary!
310 lto_codegen_dispose(cg);
311 return LDPS_OK;
312 }
313 }
314
315 lto_codegen_set_pic_model(cg, output_type);
316 lto_codegen_set_debug_model(cg, LTO_DEBUG_MODEL_DWARF);
317
318 size_t bufsize = 0;
319 const char *buffer = static_cast<const char *>(lto_codegen_compile(cg,
320 &bufsize));
321
322 std::string ErrMsg;
323
324 sys::Path uniqueObjPath("/tmp/llvmgold.o");
325 if (uniqueObjPath.createTemporaryFileOnDisk(true, &ErrMsg)) {
326 (*message)(LDPL_ERROR, "%s", ErrMsg.c_str());
327 return LDPS_ERR;
328 }
329 raw_fd_ostream *objFile = new raw_fd_ostream(uniqueObjPath.c_str(), true,
330 ErrMsg);
331 if (!ErrMsg.empty()) {
332 delete objFile;
333 (*message)(LDPL_ERROR, "%s", ErrMsg.c_str());
334 return LDPS_ERR;
335 }
336
337 objFile->write(buffer, bufsize);
338 objFile->close();
339
340 lto_codegen_dispose(cg);
Torok Edwin3e5a0d82009-02-04 17:39:30 +0000341 for (std::list<claimed_file>::iterator I = Modules.begin(),
342 E = Modules.end(); I != E; ++I) {
343 free(I->buf);
344 }
Nick Lewycky3e62b2d2009-02-03 07:13:24 +0000345
346 if ((*add_input_file)(const_cast<char*>(uniqueObjPath.c_str())) != LDPS_OK) {
347 (*message)(LDPL_ERROR, "Unable to add .o file to the link.");
348 (*message)(LDPL_ERROR, "File left behind in: %s", uniqueObjPath.c_str());
349 return LDPS_ERR;
350 }
351
352 Cleanup.push_back(uniqueObjPath);
353
354 return LDPS_OK;
355}
356
357ld_plugin_status cleanup_hook(void) {
358 std::string ErrMsg;
359
360 for (int i = 0, e = Cleanup.size(); i != e; ++i)
361 if (Cleanup[i].eraseFromDisk(false, &ErrMsg))
362 (*message)(LDPL_ERROR, "Failed to delete '%s': %s", Cleanup[i].c_str(),
363 ErrMsg.c_str());
364
365 return LDPS_OK;
366}