/usr/share/doc/llvm-3.5-doc/html/GetElementPtr.html is in llvm-3.5-doc 1:3.5.2-3ubuntu1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 | <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<title>The Often Misunderstood GEP Instruction — LLVM 3.5 documentation</title>
<link rel="stylesheet" href="_static/llvm-theme.css" type="text/css" />
<link rel="stylesheet" href="_static/pygments.css" type="text/css" />
<script type="text/javascript">
var DOCUMENTATION_OPTIONS = {
URL_ROOT: './',
VERSION: '3.5',
COLLAPSE_INDEX: false,
FILE_SUFFIX: '.html',
HAS_SOURCE: true
};
</script>
<script type="text/javascript" src="_static/jquery.js"></script>
<script type="text/javascript" src="_static/underscore.js"></script>
<script type="text/javascript" src="_static/doctools.js"></script>
<link rel="top" title="LLVM 3.5 documentation" href="index.html" />
<link rel="next" title="MCJIT Design and Implementation" href="MCJITDesignAndImplementation.html" />
<link rel="prev" title="YAML I/O" href="YamlIO.html" />
<style type="text/css">
table.right { float: right; margin-left: 20px; }
table.right td { border: 1px solid #ccc; }
</style>
</head>
<body role="document">
<div class="logo">
<a href="index.html">
<img src="_static/logo.png"
alt="LLVM Logo" width="250" height="88"/></a>
</div>
<div class="related" role="navigation" aria-label="related navigation">
<h3>Navigation</h3>
<ul>
<li class="right" style="margin-right: 10px">
<a href="genindex.html" title="General Index"
accesskey="I">index</a></li>
<li class="right" >
<a href="MCJITDesignAndImplementation.html" title="MCJIT Design and Implementation"
accesskey="N">next</a> |</li>
<li class="right" >
<a href="YamlIO.html" title="YAML I/O"
accesskey="P">previous</a> |</li>
<li><a href="http://llvm.org/">LLVM Home</a> | </li>
<li><a href="index.html">Documentation</a>»</li>
</ul>
</div>
<div class="document">
<div class="documentwrapper">
<div class="body" role="main">
<div class="section" id="the-often-misunderstood-gep-instruction">
<h1>The Often Misunderstood GEP Instruction<a class="headerlink" href="#the-often-misunderstood-gep-instruction" title="Permalink to this headline">¶</a></h1>
<div class="contents local topic" id="contents">
<ul class="simple">
<li><a class="reference internal" href="#introduction" id="id2">Introduction</a></li>
<li><a class="reference internal" href="#address-computation" id="id3">Address Computation</a><ul>
<li><a class="reference internal" href="#what-is-the-first-index-of-the-gep-instruction" id="id4">What is the first index of the GEP instruction?</a></li>
<li><a class="reference internal" href="#why-is-the-extra-0-index-required" id="id5">Why is the extra 0 index required?</a></li>
<li><a class="reference internal" href="#what-is-dereferenced-by-gep" id="id6">What is dereferenced by GEP?</a></li>
<li><a class="reference internal" href="#why-don-t-gep-x-0-0-1-and-gep-x-1-alias" id="id7">Why don’t GEP x,0,0,1 and GEP x,1 alias?</a></li>
<li><a class="reference internal" href="#why-do-gep-x-1-0-0-and-gep-x-1-alias" id="id8">Why do GEP x,1,0,0 and GEP x,1 alias?</a></li>
<li><a class="reference internal" href="#can-gep-index-into-vector-elements" id="id9">Can GEP index into vector elements?</a></li>
<li><a class="reference internal" href="#what-effect-do-address-spaces-have-on-geps" id="id10">What effect do address spaces have on GEPs?</a></li>
<li><a class="reference internal" href="#how-is-gep-different-from-ptrtoint-arithmetic-and-inttoptr" id="id11">How is GEP different from <code class="docutils literal"><span class="pre">ptrtoint</span></code>, arithmetic, and <code class="docutils literal"><span class="pre">inttoptr</span></code>?</a></li>
<li><a class="reference internal" href="#i-m-writing-a-backend-for-a-target-which-needs-custom-lowering-for-gep-how-do-i-do-this" id="id12">I’m writing a backend for a target which needs custom lowering for GEP. How do I do this?</a></li>
<li><a class="reference internal" href="#how-does-vla-addressing-work-with-geps" id="id13">How does VLA addressing work with GEPs?</a></li>
</ul>
</li>
<li><a class="reference internal" href="#rules" id="id14">Rules</a><ul>
<li><a class="reference internal" href="#what-happens-if-an-array-index-is-out-of-bounds" id="id15">What happens if an array index is out of bounds?</a></li>
<li><a class="reference internal" href="#can-array-indices-be-negative" id="id16">Can array indices be negative?</a></li>
<li><a class="reference internal" href="#can-i-compare-two-values-computed-with-geps" id="id17">Can I compare two values computed with GEPs?</a></li>
<li><a class="reference internal" href="#can-i-do-gep-with-a-different-pointer-type-than-the-type-of-the-underlying-object" id="id18">Can I do GEP with a different pointer type than the type of the underlying object?</a></li>
<li><a class="reference internal" href="#can-i-cast-an-object-s-address-to-integer-and-add-it-to-null" id="id19">Can I cast an object’s address to integer and add it to null?</a></li>
<li><a class="reference internal" href="#can-i-compute-the-distance-between-two-objects-and-add-that-value-to-one-address-to-compute-the-other-address" id="id20">Can I compute the distance between two objects, and add that value to one address to compute the other address?</a></li>
<li><a class="reference internal" href="#can-i-do-type-based-alias-analysis-on-llvm-ir" id="id21">Can I do type-based alias analysis on LLVM IR?</a></li>
<li><a class="reference internal" href="#what-happens-if-a-gep-computation-overflows" id="id22">What happens if a GEP computation overflows?</a></li>
<li><a class="reference internal" href="#how-can-i-tell-if-my-front-end-is-following-the-rules" id="id23">How can I tell if my front-end is following the rules?</a></li>
</ul>
</li>
<li><a class="reference internal" href="#rationale" id="id24">Rationale</a><ul>
<li><a class="reference internal" href="#why-is-gep-designed-this-way" id="id25">Why is GEP designed this way?</a></li>
<li><a class="reference internal" href="#why-do-struct-member-indices-always-use-i32" id="id26">Why do struct member indices always use <code class="docutils literal"><span class="pre">i32</span></code>?</a></li>
<li><a class="reference internal" href="#what-s-an-uglygep" id="id27">What’s an uglygep?</a></li>
</ul>
</li>
<li><a class="reference internal" href="#summary" id="id28">Summary</a></li>
</ul>
</div>
<div class="section" id="introduction">
<h2><a class="toc-backref" href="#id2">Introduction</a><a class="headerlink" href="#introduction" title="Permalink to this headline">¶</a></h2>
<p>This document seeks to dispel the mystery and confusion surrounding LLVM’s
<a class="reference external" href="LangRef.html#i_getelementptr">GetElementPtr</a> (GEP) instruction. Questions
about the wily GEP instruction are probably the most frequently occurring
questions once a developer gets down to coding with LLVM. Here we lay out the
sources of confusion and show that the GEP instruction is really quite simple.</p>
</div>
<div class="section" id="address-computation">
<h2><a class="toc-backref" href="#id3">Address Computation</a><a class="headerlink" href="#address-computation" title="Permalink to this headline">¶</a></h2>
<p>When people are first confronted with the GEP instruction, they tend to relate
it to known concepts from other programming paradigms, most notably C array
indexing and field selection. GEP closely resembles C array indexing and field
selection, however it is a little different and this leads to the following
questions.</p>
<div class="section" id="what-is-the-first-index-of-the-gep-instruction">
<h3><a class="toc-backref" href="#id4">What is the first index of the GEP instruction?</a><a class="headerlink" href="#what-is-the-first-index-of-the-gep-instruction" title="Permalink to this headline">¶</a></h3>
<p>Quick answer: The index stepping through the first operand.</p>
<p>The confusion with the first index usually arises from thinking about the
GetElementPtr instruction as if it was a C index operator. They aren’t the
same. For example, when we write, in “C”:</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="n">AType</span> <span class="o">*</span><span class="n">Foo</span><span class="p">;</span>
<span class="p">...</span>
<span class="n">X</span> <span class="o">=</span> <span class="o">&</span><span class="n">Foo</span><span class="o">-></span><span class="n">F</span><span class="p">;</span>
</pre></div>
</div>
<p>it is natural to think that there is only one index, the selection of the field
<code class="docutils literal"><span class="pre">F</span></code>. However, in this example, <code class="docutils literal"><span class="pre">Foo</span></code> is a pointer. That pointer
must be indexed explicitly in LLVM. C, on the other hand, indices through it
transparently. To arrive at the same address location as the C code, you would
provide the GEP instruction with two index operands. The first operand indexes
through the pointer; the second operand indexes the field <code class="docutils literal"><span class="pre">F</span></code> of the
structure, just as if you wrote:</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="n">X</span> <span class="o">=</span> <span class="o">&</span><span class="n">Foo</span><span class="p">[</span><span class="mi">0</span><span class="p">].</span><span class="n">F</span><span class="p">;</span>
</pre></div>
</div>
<p>Sometimes this question gets rephrased as:</p>
<blockquote id="gep-index-through-first-pointer">
<div><em>Why is it okay to index through the first pointer, but subsequent pointers
won’t be dereferenced?</em></div></blockquote>
<p>The answer is simply because memory does not have to be accessed to perform the
computation. The first operand to the GEP instruction must be a value of a
pointer type. The value of the pointer is provided directly to the GEP
instruction as an operand without any need for accessing memory. It must,
therefore be indexed and requires an index operand. Consider this example:</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="k">struct</span> <span class="n">munger_struct</span> <span class="p">{</span>
<span class="kt">int</span> <span class="n">f1</span><span class="p">;</span>
<span class="kt">int</span> <span class="n">f2</span><span class="p">;</span>
<span class="p">};</span>
<span class="kt">void</span> <span class="nf">munge</span><span class="p">(</span><span class="k">struct</span> <span class="n">munger_struct</span> <span class="o">*</span><span class="n">P</span><span class="p">)</span> <span class="p">{</span>
<span class="n">P</span><span class="p">[</span><span class="mi">0</span><span class="p">].</span><span class="n">f1</span> <span class="o">=</span> <span class="n">P</span><span class="p">[</span><span class="mi">1</span><span class="p">].</span><span class="n">f1</span> <span class="o">+</span> <span class="n">P</span><span class="p">[</span><span class="mi">2</span><span class="p">].</span><span class="n">f2</span><span class="p">;</span>
<span class="p">}</span>
<span class="p">...</span>
<span class="n">munger_struct</span> <span class="n">Array</span><span class="p">[</span><span class="mi">3</span><span class="p">];</span>
<span class="p">...</span>
<span class="n">munge</span><span class="p">(</span><span class="n">Array</span><span class="p">);</span>
</pre></div>
</div>
<p>In this “C” example, the front end compiler (Clang) will generate three GEP
instructions for the three indices through “P” in the assignment statement. The
function argument <code class="docutils literal"><span class="pre">P</span></code> will be the first operand of each of these GEP
instructions. The second operand indexes through that pointer. The third
operand will be the field offset into the <code class="docutils literal"><span class="pre">struct</span> <span class="pre">munger_struct</span></code> type, for
either the <code class="docutils literal"><span class="pre">f1</span></code> or <code class="docutils literal"><span class="pre">f2</span></code> field. So, in LLVM assembly the <code class="docutils literal"><span class="pre">munge</span></code> function
looks like:</p>
<div class="highlight-llvm"><div class="highlight"><pre><span class="kt">void</span> <span class="nv">%munge</span><span class="p">(</span><span class="nv">%struct.munger_struct</span><span class="p">*</span> <span class="nv">%P</span><span class="p">)</span> <span class="p">{</span>
<span class="nl">entry:</span>
<span class="nv">%tmp</span> <span class="p">=</span> <span class="k">getelementptr</span> <span class="nv">%struct.munger_struct</span><span class="p">*</span> <span class="nv">%P</span><span class="p">,</span> <span class="k">i32</span> <span class="m">1</span><span class="p">,</span> <span class="k">i32</span> <span class="m">0</span>
<span class="nv">%tmp</span> <span class="p">=</span> <span class="k">load</span> <span class="k">i32</span><span class="p">*</span> <span class="nv">%tmp</span>
<span class="nv">%tmp6</span> <span class="p">=</span> <span class="k">getelementptr</span> <span class="nv">%struct.munger_struct</span><span class="p">*</span> <span class="nv">%P</span><span class="p">,</span> <span class="k">i32</span> <span class="m">2</span><span class="p">,</span> <span class="k">i32</span> <span class="m">1</span>
<span class="nv">%tmp7</span> <span class="p">=</span> <span class="k">load</span> <span class="k">i32</span><span class="p">*</span> <span class="nv">%tmp6</span>
<span class="nv">%tmp8</span> <span class="p">=</span> <span class="k">add</span> <span class="k">i32</span> <span class="nv">%tmp7</span><span class="p">,</span> <span class="nv">%tmp</span>
<span class="nv">%tmp9</span> <span class="p">=</span> <span class="k">getelementptr</span> <span class="nv">%struct.munger_struct</span><span class="p">*</span> <span class="nv">%P</span><span class="p">,</span> <span class="k">i32</span> <span class="m">0</span><span class="p">,</span> <span class="k">i32</span> <span class="m">0</span>
<span class="k">store</span> <span class="k">i32</span> <span class="nv">%tmp8</span><span class="p">,</span> <span class="k">i32</span><span class="p">*</span> <span class="nv">%tmp9</span>
<span class="k">ret</span> <span class="kt">void</span>
<span class="p">}</span>
</pre></div>
</div>
<p>In each case the first operand is the pointer through which the GEP instruction
starts. The same is true whether the first operand is an argument, allocated
memory, or a global variable.</p>
<p>To make this clear, let’s consider a more obtuse example:</p>
<div class="highlight-llvm"><div class="highlight"><pre>%MyVar = uninitialized global i32
...
%idx1 = getelementptr i32* %MyVar, i64 0
%idx2 = getelementptr i32* %MyVar, i64 1
%idx3 = getelementptr i32* %MyVar, i64 2
</pre></div>
</div>
<p>These GEP instructions are simply making address computations from the base
address of <code class="docutils literal"><span class="pre">MyVar</span></code>. They compute, as follows (using C syntax):</p>
<div class="highlight-c++"><div class="highlight"><pre><span class="n">idx1</span> <span class="o">=</span> <span class="p">(</span><span class="kt">char</span><span class="o">*</span><span class="p">)</span> <span class="o">&</span><span class="n">MyVar</span> <span class="o">+</span> <span class="mi">0</span>
<span class="n">idx2</span> <span class="o">=</span> <span class="p">(</span><span class="kt">char</span><span class="o">*</span><span class="p">)</span> <span class="o">&</span><span class="n">MyVar</span> <span class="o">+</span> <span class="mi">4</span>
<span class="n">idx3</span> <span class="o">=</span> <span class="p">(</span><span class="kt">char</span><span class="o">*</span><span class="p">)</span> <span class="o">&</span><span class="n">MyVar</span> <span class="o">+</span> <span class="mi">8</span>
</pre></div>
</div>
<p>Since the type <code class="docutils literal"><span class="pre">i32</span></code> is known to be four bytes long, the indices 0, 1 and 2
translate into memory offsets of 0, 4, and 8, respectively. No memory is
accessed to make these computations because the address of <code class="docutils literal"><span class="pre">%MyVar</span></code> is passed
directly to the GEP instructions.</p>
<p>The obtuse part of this example is in the cases of <code class="docutils literal"><span class="pre">%idx2</span></code> and <code class="docutils literal"><span class="pre">%idx3</span></code>. They
result in the computation of addresses that point to memory past the end of the
<code class="docutils literal"><span class="pre">%MyVar</span></code> global, which is only one <code class="docutils literal"><span class="pre">i32</span></code> long, not three <code class="docutils literal"><span class="pre">i32</span></code>s long.
While this is legal in LLVM, it is inadvisable because any load or store with
the pointer that results from these GEP instructions would produce undefined
results.</p>
</div>
<div class="section" id="why-is-the-extra-0-index-required">
<h3><a class="toc-backref" href="#id5">Why is the extra 0 index required?</a><a class="headerlink" href="#why-is-the-extra-0-index-required" title="Permalink to this headline">¶</a></h3>
<p>Quick answer: there are no superfluous indices.</p>
<p>This question arises most often when the GEP instruction is applied to a global
variable which is always a pointer type. For example, consider this:</p>
<div class="highlight-llvm"><div class="highlight"><pre>%MyStruct = uninitialized global { float*, i32 }
...
%idx = getelementptr { float*, i32 }* %MyStruct, i64 0, i32 1
</pre></div>
</div>
<p>The GEP above yields an <code class="docutils literal"><span class="pre">i32*</span></code> by indexing the <code class="docutils literal"><span class="pre">i32</span></code> typed field of the
structure <code class="docutils literal"><span class="pre">%MyStruct</span></code>. When people first look at it, they wonder why the <code class="docutils literal"><span class="pre">i64</span>
<span class="pre">0</span></code> index is needed. However, a closer inspection of how globals and GEPs work
reveals the need. Becoming aware of the following facts will dispel the
confusion:</p>
<ol class="arabic simple">
<li>The type of <code class="docutils literal"><span class="pre">%MyStruct</span></code> is <em>not</em> <code class="docutils literal"><span class="pre">{</span> <span class="pre">float*,</span> <span class="pre">i32</span> <span class="pre">}</span></code> but rather <code class="docutils literal"><span class="pre">{</span> <span class="pre">float*,</span>
<span class="pre">i32</span> <span class="pre">}*</span></code>. That is, <code class="docutils literal"><span class="pre">%MyStruct</span></code> is a pointer to a structure containing a
pointer to a <code class="docutils literal"><span class="pre">float</span></code> and an <code class="docutils literal"><span class="pre">i32</span></code>.</li>
<li>Point #1 is evidenced by noticing the type of the first operand of the GEP
instruction (<code class="docutils literal"><span class="pre">%MyStruct</span></code>) which is <code class="docutils literal"><span class="pre">{</span> <span class="pre">float*,</span> <span class="pre">i32</span> <span class="pre">}*</span></code>.</li>
<li>The first index, <code class="docutils literal"><span class="pre">i64</span> <span class="pre">0</span></code> is required to step over the global variable
<code class="docutils literal"><span class="pre">%MyStruct</span></code>. Since the first argument to the GEP instruction must always
be a value of pointer type, the first index steps through that pointer. A
value of 0 means 0 elements offset from that pointer.</li>
<li>The second index, <code class="docutils literal"><span class="pre">i32</span> <span class="pre">1</span></code> selects the second field of the structure (the
<code class="docutils literal"><span class="pre">i32</span></code>).</li>
</ol>
</div>
<div class="section" id="what-is-dereferenced-by-gep">
<h3><a class="toc-backref" href="#id6">What is dereferenced by GEP?</a><a class="headerlink" href="#what-is-dereferenced-by-gep" title="Permalink to this headline">¶</a></h3>
<p>Quick answer: nothing.</p>
<p>The GetElementPtr instruction dereferences nothing. That is, it doesn’t access
memory in any way. That’s what the Load and Store instructions are for. GEP is
only involved in the computation of addresses. For example, consider this:</p>
<div class="highlight-llvm"><div class="highlight"><pre>%MyVar = uninitialized global { [40 x i32 ]* }
...
%idx = getelementptr { [40 x i32]* }* %MyVar, i64 0, i32 0, i64 0, i64 17
</pre></div>
</div>
<p>In this example, we have a global variable, <code class="docutils literal"><span class="pre">%MyVar</span></code> that is a pointer to a
structure containing a pointer to an array of 40 ints. The GEP instruction seems
to be accessing the 18th integer of the structure’s array of ints. However, this
is actually an illegal GEP instruction. It won’t compile. The reason is that the
pointer in the structure <em>must</em> be dereferenced in order to index into the
array of 40 ints. Since the GEP instruction never accesses memory, it is
illegal.</p>
<p>In order to access the 18th integer in the array, you would need to do the
following:</p>
<div class="highlight-llvm"><div class="highlight"><pre>%idx = getelementptr { [40 x i32]* }* %, i64 0, i32 0
%arr = load [40 x i32]** %idx
%idx = getelementptr [40 x i32]* %arr, i64 0, i64 17
</pre></div>
</div>
<p>In this case, we have to load the pointer in the structure with a load
instruction before we can index into the array. If the example was changed to:</p>
<div class="highlight-llvm"><div class="highlight"><pre>%MyVar = uninitialized global { [40 x i32 ] }
...
%idx = getelementptr { [40 x i32] }*, i64 0, i32 0, i64 17
</pre></div>
</div>
<p>then everything works fine. In this case, the structure does not contain a
pointer and the GEP instruction can index through the global variable, into the
first field of the structure and access the 18th <code class="docutils literal"><span class="pre">i32</span></code> in the array there.</p>
</div>
<div class="section" id="why-don-t-gep-x-0-0-1-and-gep-x-1-alias">
<h3><a class="toc-backref" href="#id7">Why don’t GEP x,0,0,1 and GEP x,1 alias?</a><a class="headerlink" href="#why-don-t-gep-x-0-0-1-and-gep-x-1-alias" title="Permalink to this headline">¶</a></h3>
<p>Quick Answer: They compute different address locations.</p>
<p>If you look at the first indices in these GEP instructions you find that they
are different (0 and 1), therefore the address computation diverges with that
index. Consider this example:</p>
<div class="highlight-llvm"><div class="highlight"><pre><span class="nv">%MyVar</span> <span class="p">=</span> <span class="k">global</span> <span class="p">{</span> <span class="p">[</span><span class="m">10</span> <span class="k">x</span> <span class="k">i32</span> <span class="p">]</span> <span class="p">}</span>
<span class="nv">%idx1</span> <span class="p">=</span> <span class="k">getelementptr</span> <span class="p">{</span> <span class="p">[</span><span class="m">10</span> <span class="k">x</span> <span class="k">i32</span> <span class="p">]</span> <span class="p">}*</span> <span class="nv">%MyVar</span><span class="p">,</span> <span class="k">i64</span> <span class="m">0</span><span class="p">,</span> <span class="k">i32</span> <span class="m">0</span><span class="p">,</span> <span class="k">i64</span> <span class="m">1</span>
<span class="nv">%idx2</span> <span class="p">=</span> <span class="k">getelementptr</span> <span class="p">{</span> <span class="p">[</span><span class="m">10</span> <span class="k">x</span> <span class="k">i32</span> <span class="p">]</span> <span class="p">}*</span> <span class="nv">%MyVar</span><span class="p">,</span> <span class="k">i64</span> <span class="m">1</span>
</pre></div>
</div>
<p>In this example, <code class="docutils literal"><span class="pre">idx1</span></code> computes the address of the second integer in the
array that is in the structure in <code class="docutils literal"><span class="pre">%MyVar</span></code>, that is <code class="docutils literal"><span class="pre">MyVar+4</span></code>. The type of
<code class="docutils literal"><span class="pre">idx1</span></code> is <code class="docutils literal"><span class="pre">i32*</span></code>. However, <code class="docutils literal"><span class="pre">idx2</span></code> computes the address of <em>the next</em>
structure after <code class="docutils literal"><span class="pre">%MyVar</span></code>. The type of <code class="docutils literal"><span class="pre">idx2</span></code> is <code class="docutils literal"><span class="pre">{</span> <span class="pre">[10</span> <span class="pre">x</span> <span class="pre">i32]</span> <span class="pre">}*</span></code> and its
value is equivalent to <code class="docutils literal"><span class="pre">MyVar</span> <span class="pre">+</span> <span class="pre">40</span></code> because it indexes past the ten 4-byte
integers in <code class="docutils literal"><span class="pre">MyVar</span></code>. Obviously, in such a situation, the pointers don’t
alias.</p>
</div>
<div class="section" id="why-do-gep-x-1-0-0-and-gep-x-1-alias">
<h3><a class="toc-backref" href="#id8">Why do GEP x,1,0,0 and GEP x,1 alias?</a><a class="headerlink" href="#why-do-gep-x-1-0-0-and-gep-x-1-alias" title="Permalink to this headline">¶</a></h3>
<p>Quick Answer: They compute the same address location.</p>
<p>These two GEP instructions will compute the same address because indexing
through the 0th element does not change the address. However, it does change the
type. Consider this example:</p>
<div class="highlight-llvm"><div class="highlight"><pre><span class="nv">%MyVar</span> <span class="p">=</span> <span class="k">global</span> <span class="p">{</span> <span class="p">[</span><span class="m">10</span> <span class="k">x</span> <span class="k">i32</span> <span class="p">]</span> <span class="p">}</span>
<span class="nv">%idx1</span> <span class="p">=</span> <span class="k">getelementptr</span> <span class="p">{</span> <span class="p">[</span><span class="m">10</span> <span class="k">x</span> <span class="k">i32</span> <span class="p">]</span> <span class="p">}*</span> <span class="nv">%MyVar</span><span class="p">,</span> <span class="k">i64</span> <span class="m">1</span><span class="p">,</span> <span class="k">i32</span> <span class="m">0</span><span class="p">,</span> <span class="k">i64</span> <span class="m">0</span>
<span class="nv">%idx2</span> <span class="p">=</span> <span class="k">getelementptr</span> <span class="p">{</span> <span class="p">[</span><span class="m">10</span> <span class="k">x</span> <span class="k">i32</span> <span class="p">]</span> <span class="p">}*</span> <span class="nv">%MyVar</span><span class="p">,</span> <span class="k">i64</span> <span class="m">1</span>
</pre></div>
</div>
<p>In this example, the value of <code class="docutils literal"><span class="pre">%idx1</span></code> is <code class="docutils literal"><span class="pre">%MyVar+40</span></code> and its type is
<code class="docutils literal"><span class="pre">i32*</span></code>. The value of <code class="docutils literal"><span class="pre">%idx2</span></code> is also <code class="docutils literal"><span class="pre">MyVar+40</span></code> but its type is <code class="docutils literal"><span class="pre">{</span> <span class="pre">[10</span> <span class="pre">x</span>
<span class="pre">i32]</span> <span class="pre">}*</span></code>.</p>
</div>
<div class="section" id="can-gep-index-into-vector-elements">
<h3><a class="toc-backref" href="#id9">Can GEP index into vector elements?</a><a class="headerlink" href="#can-gep-index-into-vector-elements" title="Permalink to this headline">¶</a></h3>
<p>This hasn’t always been forcefully disallowed, though it’s not recommended. It
leads to awkward special cases in the optimizers, and fundamental inconsistency
in the IR. In the future, it will probably be outright disallowed.</p>
</div>
<div class="section" id="what-effect-do-address-spaces-have-on-geps">
<h3><a class="toc-backref" href="#id10">What effect do address spaces have on GEPs?</a><a class="headerlink" href="#what-effect-do-address-spaces-have-on-geps" title="Permalink to this headline">¶</a></h3>
<p>None, except that the address space qualifier on the first operand pointer type
always matches the address space qualifier on the result type.</p>
</div>
<div class="section" id="how-is-gep-different-from-ptrtoint-arithmetic-and-inttoptr">
<h3><a class="toc-backref" href="#id11">How is GEP different from <code class="docutils literal"><span class="pre">ptrtoint</span></code>, arithmetic, and <code class="docutils literal"><span class="pre">inttoptr</span></code>?</a><a class="headerlink" href="#how-is-gep-different-from-ptrtoint-arithmetic-and-inttoptr" title="Permalink to this headline">¶</a></h3>
<p>It’s very similar; there are only subtle differences.</p>
<p>With ptrtoint, you have to pick an integer type. One approach is to pick i64;
this is safe on everything LLVM supports (LLVM internally assumes pointers are
never wider than 64 bits in many places), and the optimizer will actually narrow
the i64 arithmetic down to the actual pointer size on targets which don’t
support 64-bit arithmetic in most cases. However, there are some cases where it
doesn’t do this. With GEP you can avoid this problem.</p>
<p>Also, GEP carries additional pointer aliasing rules. It’s invalid to take a GEP
from one object, address into a different separately allocated object, and
dereference it. IR producers (front-ends) must follow this rule, and consumers
(optimizers, specifically alias analysis) benefit from being able to rely on
it. See the <a class="reference internal" href="#rules">Rules</a> section for more information.</p>
<p>And, GEP is more concise in common cases.</p>
<p>However, for the underlying integer computation implied, there is no
difference.</p>
</div>
<div class="section" id="i-m-writing-a-backend-for-a-target-which-needs-custom-lowering-for-gep-how-do-i-do-this">
<h3><a class="toc-backref" href="#id12">I’m writing a backend for a target which needs custom lowering for GEP. How do I do this?</a><a class="headerlink" href="#i-m-writing-a-backend-for-a-target-which-needs-custom-lowering-for-gep-how-do-i-do-this" title="Permalink to this headline">¶</a></h3>
<p>You don’t. The integer computation implied by a GEP is target-independent.
Typically what you’ll need to do is make your backend pattern-match expressions
trees involving ADD, MUL, etc., which are what GEP is lowered into. This has the
advantage of letting your code work correctly in more cases.</p>
<p>GEP does use target-dependent parameters for the size and layout of data types,
which targets can customize.</p>
<p>If you require support for addressing units which are not 8 bits, you’ll need to
fix a lot of code in the backend, with GEP lowering being only a small piece of
the overall picture.</p>
</div>
<div class="section" id="how-does-vla-addressing-work-with-geps">
<h3><a class="toc-backref" href="#id13">How does VLA addressing work with GEPs?</a><a class="headerlink" href="#how-does-vla-addressing-work-with-geps" title="Permalink to this headline">¶</a></h3>
<p>GEPs don’t natively support VLAs. LLVM’s type system is entirely static, and GEP
address computations are guided by an LLVM type.</p>
<p>VLA indices can be implemented as linearized indices. For example, an expression
like <code class="docutils literal"><span class="pre">X[a][b][c]</span></code>, must be effectively lowered into a form like
<code class="docutils literal"><span class="pre">X[a*m+b*n+c]</span></code>, so that it appears to the GEP as a single-dimensional array
reference.</p>
<p>This means if you want to write an analysis which understands array indices and
you want to support VLAs, your code will have to be prepared to reverse-engineer
the linearization. One way to solve this problem is to use the ScalarEvolution
library, which always presents VLA and non-VLA indexing in the same manner.</p>
</div>
</div>
<div class="section" id="rules">
<span id="id1"></span><h2><a class="toc-backref" href="#id14">Rules</a><a class="headerlink" href="#rules" title="Permalink to this headline">¶</a></h2>
<div class="section" id="what-happens-if-an-array-index-is-out-of-bounds">
<h3><a class="toc-backref" href="#id15">What happens if an array index is out of bounds?</a><a class="headerlink" href="#what-happens-if-an-array-index-is-out-of-bounds" title="Permalink to this headline">¶</a></h3>
<p>There are two senses in which an array index can be out of bounds.</p>
<p>First, there’s the array type which comes from the (static) type of the first
operand to the GEP. Indices greater than the number of elements in the
corresponding static array type are valid. There is no problem with out of
bounds indices in this sense. Indexing into an array only depends on the size of
the array element, not the number of elements.</p>
<p>A common example of how this is used is arrays where the size is not known.
It’s common to use array types with zero length to represent these. The fact
that the static type says there are zero elements is irrelevant; it’s perfectly
valid to compute arbitrary element indices, as the computation only depends on
the size of the array element, not the number of elements. Note that zero-sized
arrays are not a special case here.</p>
<p>This sense is unconnected with <code class="docutils literal"><span class="pre">inbounds</span></code> keyword. The <code class="docutils literal"><span class="pre">inbounds</span></code> keyword is
designed to describe low-level pointer arithmetic overflow conditions, rather
than high-level array indexing rules.</p>
<p>Analysis passes which wish to understand array indexing should not assume that
the static array type bounds are respected.</p>
<p>The second sense of being out of bounds is computing an address that’s beyond
the actual underlying allocated object.</p>
<p>With the <code class="docutils literal"><span class="pre">inbounds</span></code> keyword, the result value of the GEP is undefined if the
address is outside the actual underlying allocated object and not the address
one-past-the-end.</p>
<p>Without the <code class="docutils literal"><span class="pre">inbounds</span></code> keyword, there are no restrictions on computing
out-of-bounds addresses. Obviously, performing a load or a store requires an
address of allocated and sufficiently aligned memory. But the GEP itself is only
concerned with computing addresses.</p>
</div>
<div class="section" id="can-array-indices-be-negative">
<h3><a class="toc-backref" href="#id16">Can array indices be negative?</a><a class="headerlink" href="#can-array-indices-be-negative" title="Permalink to this headline">¶</a></h3>
<p>Yes. This is basically a special case of array indices being out of bounds.</p>
</div>
<div class="section" id="can-i-compare-two-values-computed-with-geps">
<h3><a class="toc-backref" href="#id17">Can I compare two values computed with GEPs?</a><a class="headerlink" href="#can-i-compare-two-values-computed-with-geps" title="Permalink to this headline">¶</a></h3>
<p>Yes. If both addresses are within the same allocated object, or
one-past-the-end, you’ll get the comparison result you expect. If either is
outside of it, integer arithmetic wrapping may occur, so the comparison may not
be meaningful.</p>
</div>
<div class="section" id="can-i-do-gep-with-a-different-pointer-type-than-the-type-of-the-underlying-object">
<h3><a class="toc-backref" href="#id18">Can I do GEP with a different pointer type than the type of the underlying object?</a><a class="headerlink" href="#can-i-do-gep-with-a-different-pointer-type-than-the-type-of-the-underlying-object" title="Permalink to this headline">¶</a></h3>
<p>Yes. There are no restrictions on bitcasting a pointer value to an arbitrary
pointer type. The types in a GEP serve only to define the parameters for the
underlying integer computation. They need not correspond with the actual type of
the underlying object.</p>
<p>Furthermore, loads and stores don’t have to use the same types as the type of
the underlying object. Types in this context serve only to specify memory size
and alignment. Beyond that there are merely a hint to the optimizer indicating
how the value will likely be used.</p>
</div>
<div class="section" id="can-i-cast-an-object-s-address-to-integer-and-add-it-to-null">
<h3><a class="toc-backref" href="#id19">Can I cast an object’s address to integer and add it to null?</a><a class="headerlink" href="#can-i-cast-an-object-s-address-to-integer-and-add-it-to-null" title="Permalink to this headline">¶</a></h3>
<p>You can compute an address that way, but if you use GEP to do the add, you can’t
use that pointer to actually access the object, unless the object is managed
outside of LLVM.</p>
<p>The underlying integer computation is sufficiently defined; null has a defined
value — zero — and you can add whatever value you want to it.</p>
<p>However, it’s invalid to access (load from or store to) an LLVM-aware object
with such a pointer. This includes <code class="docutils literal"><span class="pre">GlobalVariables</span></code>, <code class="docutils literal"><span class="pre">Allocas</span></code>, and objects
pointed to by noalias pointers.</p>
<p>If you really need this functionality, you can do the arithmetic with explicit
integer instructions, and use inttoptr to convert the result to an address. Most
of GEP’s special aliasing rules do not apply to pointers computed from ptrtoint,
arithmetic, and inttoptr sequences.</p>
</div>
<div class="section" id="can-i-compute-the-distance-between-two-objects-and-add-that-value-to-one-address-to-compute-the-other-address">
<h3><a class="toc-backref" href="#id20">Can I compute the distance between two objects, and add that value to one address to compute the other address?</a><a class="headerlink" href="#can-i-compute-the-distance-between-two-objects-and-add-that-value-to-one-address-to-compute-the-other-address" title="Permalink to this headline">¶</a></h3>
<p>As with arithmetic on null, you can use GEP to compute an address that way, but
you can’t use that pointer to actually access the object if you do, unless the
object is managed outside of LLVM.</p>
<p>Also as above, ptrtoint and inttoptr provide an alternative way to do this which
do not have this restriction.</p>
</div>
<div class="section" id="can-i-do-type-based-alias-analysis-on-llvm-ir">
<h3><a class="toc-backref" href="#id21">Can I do type-based alias analysis on LLVM IR?</a><a class="headerlink" href="#can-i-do-type-based-alias-analysis-on-llvm-ir" title="Permalink to this headline">¶</a></h3>
<p>You can’t do type-based alias analysis using LLVM’s built-in type system,
because LLVM has no restrictions on mixing types in addressing, loads or stores.</p>
<p>LLVM’s type-based alias analysis pass uses metadata to describe a different type
system (such as the C type system), and performs type-based aliasing on top of
that. Further details are in the <a class="reference external" href="LangRef.html#tbaa">language reference</a>.</p>
</div>
<div class="section" id="what-happens-if-a-gep-computation-overflows">
<h3><a class="toc-backref" href="#id22">What happens if a GEP computation overflows?</a><a class="headerlink" href="#what-happens-if-a-gep-computation-overflows" title="Permalink to this headline">¶</a></h3>
<p>If the GEP lacks the <code class="docutils literal"><span class="pre">inbounds</span></code> keyword, the value is the result from
evaluating the implied two’s complement integer computation. However, since
there’s no guarantee of where an object will be allocated in the address space,
such values have limited meaning.</p>
<p>If the GEP has the <code class="docutils literal"><span class="pre">inbounds</span></code> keyword, the result value is undefined (a “trap
value”) if the GEP overflows (i.e. wraps around the end of the address space).</p>
<p>As such, there are some ramifications of this for inbounds GEPs: scales implied
by array/vector/pointer indices are always known to be “nsw” since they are
signed values that are scaled by the element size. These values are also
allowed to be negative (e.g. “<code class="docutils literal"><span class="pre">gep</span> <span class="pre">i32</span> <span class="pre">*%P,</span> <span class="pre">i32</span> <span class="pre">-1</span></code>”) but the pointer itself
is logically treated as an unsigned value. This means that GEPs have an
asymmetric relation between the pointer base (which is treated as unsigned) and
the offset applied to it (which is treated as signed). The result of the
additions within the offset calculation cannot have signed overflow, but when
applied to the base pointer, there can be signed overflow.</p>
</div>
<div class="section" id="how-can-i-tell-if-my-front-end-is-following-the-rules">
<h3><a class="toc-backref" href="#id23">How can I tell if my front-end is following the rules?</a><a class="headerlink" href="#how-can-i-tell-if-my-front-end-is-following-the-rules" title="Permalink to this headline">¶</a></h3>
<p>There is currently no checker for the getelementptr rules. Currently, the only
way to do this is to manually check each place in your front-end where
GetElementPtr operators are created.</p>
<p>It’s not possible to write a checker which could find all rule violations
statically. It would be possible to write a checker which works by instrumenting
the code with dynamic checks though. Alternatively, it would be possible to
write a static checker which catches a subset of possible problems. However, no
such checker exists today.</p>
</div>
</div>
<div class="section" id="rationale">
<h2><a class="toc-backref" href="#id24">Rationale</a><a class="headerlink" href="#rationale" title="Permalink to this headline">¶</a></h2>
<div class="section" id="why-is-gep-designed-this-way">
<h3><a class="toc-backref" href="#id25">Why is GEP designed this way?</a><a class="headerlink" href="#why-is-gep-designed-this-way" title="Permalink to this headline">¶</a></h3>
<p>The design of GEP has the following goals, in rough unofficial order of
priority:</p>
<ul class="simple">
<li>Support C, C-like languages, and languages which can be conceptually lowered
into C (this covers a lot).</li>
<li>Support optimizations such as those that are common in C compilers. In
particular, GEP is a cornerstone of LLVM’s <a class="reference external" href="LangRef.html#pointeraliasing">pointer aliasing
model</a>.</li>
<li>Provide a consistent method for computing addresses so that address
computations don’t need to be a part of load and store instructions in the IR.</li>
<li>Support non-C-like languages, to the extent that it doesn’t interfere with
other goals.</li>
<li>Minimize target-specific information in the IR.</li>
</ul>
</div>
<div class="section" id="why-do-struct-member-indices-always-use-i32">
<h3><a class="toc-backref" href="#id26">Why do struct member indices always use <code class="docutils literal"><span class="pre">i32</span></code>?</a><a class="headerlink" href="#why-do-struct-member-indices-always-use-i32" title="Permalink to this headline">¶</a></h3>
<p>The specific type i32 is probably just a historical artifact, however it’s wide
enough for all practical purposes, so there’s been no need to change it. It
doesn’t necessarily imply i32 address arithmetic; it’s just an identifier which
identifies a field in a struct. Requiring that all struct indices be the same
reduces the range of possibilities for cases where two GEPs are effectively the
same but have distinct operand types.</p>
</div>
<div class="section" id="what-s-an-uglygep">
<h3><a class="toc-backref" href="#id27">What’s an uglygep?</a><a class="headerlink" href="#what-s-an-uglygep" title="Permalink to this headline">¶</a></h3>
<p>Some LLVM optimizers operate on GEPs by internally lowering them into more
primitive integer expressions, which allows them to be combined with other
integer expressions and/or split into multiple separate integer expressions. If
they’ve made non-trivial changes, translating back into LLVM IR can involve
reverse-engineering the structure of the addressing in order to fit it into the
static type of the original first operand. It isn’t always possibly to fully
reconstruct this structure; sometimes the underlying addressing doesn’t
correspond with the static type at all. In such cases the optimizer instead will
emit a GEP with the base pointer casted to a simple address-unit pointer, using
the name “uglygep”. This isn’t pretty, but it’s just as valid, and it’s
sufficient to preserve the pointer aliasing guarantees that GEP provides.</p>
</div>
</div>
<div class="section" id="summary">
<h2><a class="toc-backref" href="#id28">Summary</a><a class="headerlink" href="#summary" title="Permalink to this headline">¶</a></h2>
<p>In summary, here’s some things to always remember about the GetElementPtr
instruction:</p>
<ol class="arabic simple">
<li>The GEP instruction never accesses memory, it only provides pointer
computations.</li>
<li>The first operand to the GEP instruction is always a pointer and it must be
indexed.</li>
<li>There are no superfluous indices for the GEP instruction.</li>
<li>Trailing zero indices are superfluous for pointer aliasing, but not for the
types of the pointers.</li>
<li>Leading zero indices are not superfluous for pointer aliasing nor the types
of the pointers.</li>
</ol>
</div>
</div>
</div>
</div>
<div class="clearer"></div>
</div>
<div class="related" role="navigation" aria-label="related navigation">
<h3>Navigation</h3>
<ul>
<li class="right" style="margin-right: 10px">
<a href="genindex.html" title="General Index"
>index</a></li>
<li class="right" >
<a href="MCJITDesignAndImplementation.html" title="MCJIT Design and Implementation"
>next</a> |</li>
<li class="right" >
<a href="YamlIO.html" title="YAML I/O"
>previous</a> |</li>
<li><a href="http://llvm.org/">LLVM Home</a> | </li>
<li><a href="index.html">Documentation</a>»</li>
</ul>
</div>
<div class="footer" role="contentinfo">
© Copyright 2003-2014, LLVM Project.
Last updated on 2016-04-18.
Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.3.6.
</div>
</body>
</html>
|