1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" 2 "http://www.w3.org/TR/html4/strict.dtd"> 3 <html> 4 <head> 5 <title>LLVM Link Time Optimization: Design and Implementation</title> 6 <link rel="stylesheet" href="llvm.css" type="text/css"> 7 </head> 8 9 <h1> 10 LLVM Link Time Optimization: Design and Implementation 11 </h1> 12 13 <ul> 14 <li><a href="#desc">Description</a></li> 15 <li><a href="#design">Design Philosophy</a> 16 <ul> 17 <li><a href="#example1">Example of link time optimization</a></li> 18 <li><a href="#alternative_approaches">Alternative Approaches</a></li> 19 </ul></li> 20 <li><a href="#multiphase">Multi-phase communication between LLVM and linker</a> 21 <ul> 22 <li><a href="#phase1">Phase 1 : Read LLVM Bitcode Files</a></li> 23 <li><a href="#phase2">Phase 2 : Symbol Resolution</a></li> 24 <li><a href="#phase3">Phase 3 : Optimize Bitcode Files</a></li> 25 <li><a href="#phase4">Phase 4 : Symbol Resolution after optimization</a></li> 26 </ul></li> 27 <li><a href="#lto">libLTO</a> 28 <ul> 29 <li><a href="#lto_module_t">lto_module_t</a></li> 30 <li><a href="#lto_code_gen_t">lto_code_gen_t</a></li> 31 </ul> 32 </ul> 33 34 <div class="doc_author"> 35 <p>Written by Devang Patel and Nick Kledzik</p> 36 </div> 37 38 <!-- *********************************************************************** --> 39 <h2> 40 <a name="desc">Description</a> 41 </h2> 42 <!-- *********************************************************************** --> 43 44 <div> 45 <p> 46 LLVM features powerful intermodular optimizations which can be used at link 47 time. Link Time Optimization (LTO) is another name for intermodular optimization 48 when performed during the link stage. This document describes the interface 49 and design between the LTO optimizer and the linker.</p> 50 </div> 51 52 <!-- *********************************************************************** --> 53 <h2> 54 <a name="design">Design Philosophy</a> 55 </h2> 56 <!-- *********************************************************************** --> 57 58 <div> 59 <p> 60 The LLVM Link Time Optimizer provides complete transparency, while doing 61 intermodular optimization, in the compiler tool chain. Its main goal is to let 62 the developer take advantage of intermodular optimizations without making any 63 significant changes to the developer's makefiles or build system. This is 64 achieved through tight integration with the linker. In this model, the linker 65 treates LLVM bitcode files like native object files and allows mixing and 66 matching among them. The linker uses <a href="#lto">libLTO</a>, a shared 67 object, to handle LLVM bitcode files. This tight integration between 68 the linker and LLVM optimizer helps to do optimizations that are not possible 69 in other models. The linker input allows the optimizer to avoid relying on 70 conservative escape analysis. 71 </p> 72 73 <!-- ======================================================================= --> 74 <h3> 75 <a name="example1">Example of link time optimization</a> 76 </h3> 77 78 <div> 79 <p>The following example illustrates the advantages of LTO's integrated 80 approach and clean interface. This example requires a system linker which 81 supports LTO through the interface described in this document. Here, 82 llvm-gcc transparently invokes system linker. </p> 83 <ul> 84 <li> Input source file <tt>a.c</tt> is compiled into LLVM bitcode form. 85 <li> Input source file <tt>main.c</tt> is compiled into native object code. 86 </ul> 87 <pre class="doc_code"> 88 --- a.h --- 89 extern int foo1(void); 90 extern void foo2(void); 91 extern void foo4(void); 92 --- a.c --- 93 #include "a.h" 94 95 static signed int i = 0; 96 97 void foo2(void) { 98 i = -1; 99 } 100 101 static int foo3() { 102 foo4(); 103 return 10; 104 } 105 106 int foo1(void) { 107 int data = 0; 108 109 if (i < 0) { data = foo3(); } 110 111 data = data + 42; 112 return data; 113 } 114 115 --- main.c --- 116 #include <stdio.h> 117 #include "a.h" 118 119 void foo4(void) { 120 printf ("Hi\n"); 121 } 122 123 int main() { 124 return foo1(); 125 } 126 127 --- command lines --- 128 $ llvm-gcc --emit-llvm -c a.c -o a.o # <-- a.o is LLVM bitcode file 129 $ llvm-gcc -c main.c -o main.o # <-- main.o is native object file 130 $ llvm-gcc a.o main.o -o main # <-- standard link command without any modifications 131 </pre> 132 <p>In this example, the linker recognizes that <tt>foo2()</tt> is an 133 externally visible symbol defined in LLVM bitcode file. The linker completes 134 its usual symbol resolution 135 pass and finds that <tt>foo2()</tt> is not used anywhere. This information 136 is used by the LLVM optimizer and it removes <tt>foo2()</tt>. As soon as 137 <tt>foo2()</tt> is removed, the optimizer recognizes that condition 138 <tt>i < 0</tt> is always false, which means <tt>foo3()</tt> is never 139 used. Hence, the optimizer removes <tt>foo3()</tt>, also. And this in turn, 140 enables linker to remove <tt>foo4()</tt>. This example illustrates the 141 advantage of tight integration with the linker. Here, the optimizer can not 142 remove <tt>foo3()</tt> without the linker's input. 143 </p> 144 </div> 145 146 <!-- ======================================================================= --> 147 <h3> 148 <a name="alternative_approaches">Alternative Approaches</a> 149 </h3> 150 151 <div> 152 <dl> 153 <dt><b>Compiler driver invokes link time optimizer separately.</b></dt> 154 <dd>In this model the link time optimizer is not able to take advantage of 155 information collected during the linker's normal symbol resolution phase. 156 In the above example, the optimizer can not remove <tt>foo2()</tt> without 157 the linker's input because it is externally visible. This in turn prohibits 158 the optimizer from removing <tt>foo3()</tt>.</dd> 159 <dt><b>Use separate tool to collect symbol information from all object 160 files.</b></dt> 161 <dd>In this model, a new, separate, tool or library replicates the linker's 162 capability to collect information for link time optimization. Not only is 163 this code duplication difficult to justify, but it also has several other 164 disadvantages. For example, the linking semantics and the features 165 provided by the linker on various platform are not unique. This means, 166 this new tool needs to support all such features and platforms in one 167 super tool or a separate tool per platform is required. This increases 168 maintenance cost for link time optimizer significantly, which is not 169 necessary. This approach also requires staying synchronized with linker 170 developements on various platforms, which is not the main focus of the link 171 time optimizer. Finally, this approach increases end user's build time due 172 to the duplication of work done by this separate tool and the linker itself. 173 </dd> 174 </dl> 175 </div> 176 177 </div> 178 179 <!-- *********************************************************************** --> 180 <h2> 181 <a name="multiphase">Multi-phase communication between libLTO and linker</a> 182 </h2> 183 184 <div> 185 <p>The linker collects information about symbol defininitions and uses in 186 various link objects which is more accurate than any information collected 187 by other tools during typical build cycles. The linker collects this 188 information by looking at the definitions and uses of symbols in native .o 189 files and using symbol visibility information. The linker also uses 190 user-supplied information, such as a list of exported symbols. LLVM 191 optimizer collects control flow information, data flow information and knows 192 much more about program structure from the optimizer's point of view. 193 Our goal is to take advantage of tight integration between the linker and 194 the optimizer by sharing this information during various linking phases. 195 </p> 196 197 <!-- ======================================================================= --> 198 <h3> 199 <a name="phase1">Phase 1 : Read LLVM Bitcode Files</a> 200 </h3> 201 202 <div> 203 <p>The linker first reads all object files in natural order and collects 204 symbol information. This includes native object files as well as LLVM bitcode 205 files. To minimize the cost to the linker in the case that all .o files 206 are native object files, the linker only calls <tt>lto_module_create()</tt> 207 when a supplied object file is found to not be a native object file. If 208 <tt>lto_module_create()</tt> returns that the file is an LLVM bitcode file, 209 the linker 210 then iterates over the module using <tt>lto_module_get_symbol_name()</tt> and 211 <tt>lto_module_get_symbol_attribute()</tt> to get all symbols defined and 212 referenced. 213 This information is added to the linker's global symbol table. 214 </p> 215 <p>The lto* functions are all implemented in a shared object libLTO. This 216 allows the LLVM LTO code to be updated independently of the linker tool. 217 On platforms that support it, the shared object is lazily loaded. 218 </p> 219 </div> 220 221 <!-- ======================================================================= --> 222 <h3> 223 <a name="phase2">Phase 2 : Symbol Resolution</a> 224 </h3> 225 226 <div> 227 <p>In this stage, the linker resolves symbols using global symbol table. 228 It may report undefined symbol errors, read archive members, replace 229 weak symbols, etc. The linker is able to do this seamlessly even though it 230 does not know the exact content of input LLVM bitcode files. If dead code 231 stripping is enabled then the linker collects the list of live symbols. 232 </p> 233 </div> 234 235 <!-- ======================================================================= --> 236 <h3> 237 <a name="phase3">Phase 3 : Optimize Bitcode Files</a> 238 </h3> 239 <div> 240 <p>After symbol resolution, the linker tells the LTO shared object which 241 symbols are needed by native object files. In the example above, the linker 242 reports that only <tt>foo1()</tt> is used by native object files using 243 <tt>lto_codegen_add_must_preserve_symbol()</tt>. Next the linker invokes 244 the LLVM optimizer and code generators using <tt>lto_codegen_compile()</tt> 245 which returns a native object file creating by merging the LLVM bitcode files 246 and applying various optimization passes. 247 </p> 248 </div> 249 250 <!-- ======================================================================= --> 251 <h3> 252 <a name="phase4">Phase 4 : Symbol Resolution after optimization</a> 253 </h3> 254 255 <div> 256 <p>In this phase, the linker reads optimized a native object file and 257 updates the internal global symbol table to reflect any changes. The linker 258 also collects information about any changes in use of external symbols by 259 LLVM bitcode files. In the example above, the linker notes that 260 <tt>foo4()</tt> is not used any more. If dead code stripping is enabled then 261 the linker refreshes the live symbol information appropriately and performs 262 dead code stripping.</p> 263 <p>After this phase, the linker continues linking as if it never saw LLVM 264 bitcode files.</p> 265 </div> 266 267 </div> 268 269 <!-- *********************************************************************** --> 270 <h2> 271 <a name="lto">libLTO</a> 272 </h2> 273 274 <div> 275 <p><tt>libLTO</tt> is a shared object that is part of the LLVM tools, and 276 is intended for use by a linker. <tt>libLTO</tt> provides an abstract C 277 interface to use the LLVM interprocedural optimizer without exposing details 278 of LLVM's internals. The intention is to keep the interface as stable as 279 possible even when the LLVM optimizer continues to evolve. It should even 280 be possible for a completely different compilation technology to provide 281 a different libLTO that works with their object files and the standard 282 linker tool.</p> 283 284 <!-- ======================================================================= --> 285 <h3> 286 <a name="lto_module_t">lto_module_t</a> 287 </h3> 288 289 <div> 290 291 <p>A non-native object file is handled via an <tt>lto_module_t</tt>. 292 The following functions allow the linker to check if a file (on disk 293 or in a memory buffer) is a file which libLTO can process:</p> 294 295 <pre class="doc_code"> 296 lto_module_is_object_file(const char*) 297 lto_module_is_object_file_for_target(const char*, const char*) 298 lto_module_is_object_file_in_memory(const void*, size_t) 299 lto_module_is_object_file_in_memory_for_target(const void*, size_t, const char*) 300 </pre> 301 302 <p>If the object file can be processed by libLTO, the linker creates a 303 <tt>lto_module_t</tt> by using one of</p> 304 305 <pre class="doc_code"> 306 lto_module_create(const char*) 307 lto_module_create_from_memory(const void*, size_t) 308 </pre> 309 310 <p>and when done, the handle is released via</p> 311 312 <pre class="doc_code"> 313 lto_module_dispose(lto_module_t) 314 </pre> 315 316 <p>The linker can introspect the non-native object file by getting the number of 317 symbols and getting the name and attributes of each symbol via:</p> 318 319 <pre class="doc_code"> 320 lto_module_get_num_symbols(lto_module_t) 321 lto_module_get_symbol_name(lto_module_t, unsigned int) 322 lto_module_get_symbol_attribute(lto_module_t, unsigned int) 323 </pre> 324 325 <p>The attributes of a symbol include the alignment, visibility, and kind.</p> 326 </div> 327 328 <!-- ======================================================================= --> 329 <h3> 330 <a name="lto_code_gen_t">lto_code_gen_t</a> 331 </h3> 332 333 <div> 334 335 <p>Once the linker has loaded each non-native object files into an 336 <tt>lto_module_t</tt>, it can request libLTO to process them all and 337 generate a native object file. This is done in a couple of steps. 338 First, a code generator is created with:</p> 339 340 <pre class="doc_code">lto_codegen_create()</pre> 341 342 <p>Then, each non-native object file is added to the code generator with:</p> 343 344 <pre class="doc_code"> 345 lto_codegen_add_module(lto_code_gen_t, lto_module_t) 346 </pre> 347 348 <p>The linker then has the option of setting some codegen options. Whether or 349 not to generate DWARF debug info is set with:</p> 350 351 <pre class="doc_code">lto_codegen_set_debug_model(lto_code_gen_t)</pre> 352 353 <p>Which kind of position independence is set with:</p> 354 355 <pre class="doc_code">lto_codegen_set_pic_model(lto_code_gen_t) </pre> 356 357 <p>And each symbol that is referenced by a native object file or otherwise must 358 not be optimized away is set with:</p> 359 360 <pre class="doc_code"> 361 lto_codegen_add_must_preserve_symbol(lto_code_gen_t, const char*) 362 </pre> 363 364 <p>After all these settings are done, the linker requests that a native object 365 file be created from the modules with the settings using:</p> 366 367 <pre class="doc_code">lto_codegen_compile(lto_code_gen_t, size*)</pre> 368 369 <p>which returns a pointer to a buffer containing the generated native 370 object file. The linker then parses that and links it with the rest 371 of the native object files.</p> 372 373 </div> 374 375 </div> 376 377 <!-- *********************************************************************** --> 378 379 <hr> 380 <address> 381 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img 382 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a> 383 <a href="http://validator.w3.org/check/referer"><img 384 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a> 385 386 Devang Patel and Nick Kledzik<br> 387 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br> 388 Last modified: $Date$ 389 </address> 390 391 </body> 392 </html> 393 394
