Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 40119ceeecafdd6d47bd7bb7f520262858b931df / . / docs / GetElementPtr.html
blob: 3bf38a366d02c95928effd78bc79ba8567be2082 [file] [log] [blame]
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<html>
4<head>
5 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
6 <title>The Often Misunderstood GEP Instruction</title>
7 <link rel="stylesheet" href="llvm.css" type="text/css">
Reid Spencerf19ccf82006-08-10 21:01:14 +0000[diff] [blame]8 <style type="text/css">
9 TABLE { text-align: left; border: 1px solid black; border-collapse: collapse; margin: 0 0 0 0; }
10 </style>
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]11</head>
12<body>
13
14<div class="doc_title">
15 The Often Misunderstood GEP Instruction
16</div>
17
18<ol>
19 <li><a href="#intro">Introduction</a></li>
Dan Gohmanb02c08c2010-02-25 18:16:03 +0000[diff] [blame]20 <li><a href="#addresses">Address Computation</a>
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]21 <ol>
22 <li><a href="#extra_index">Why is the extra 0 index required?</a></li>
23 <li><a href="#deref">What is dereferenced by GEP?</a></li>
24 <li><a href="#firstptr">Why can you index through the first pointer but not
25 subsequent ones?</a></li>
26 <li><a href="#lead0">Why don't GEP x,0,0,1 and GEP x,1 alias? </a></li>
27 <li><a href="#trail0">Why do GEP x,1,0,0 and GEP x,1 alias? </a></li>
Dan Gohmanb02c08c2010-02-25 18:16:03 +0000[diff] [blame]28 <li><a href="#vectors">Can GEP index into vector elements?</a>
29 <li><a href="#unions">Can GEP index into unions?</a>
30 <li><a href="#addrspace">What effect do address spaces have on GEPs?</a>
31 <li><a href="#int">How is GEP different from ptrtoint, arithmetic, and inttoptr?</a></li>
32 <li><a href="#be">I'm writing a backend for a target which needs custom lowering for GEP. How do I do this?</a>
33 <li><a href="#vla">How does VLA addressing work with GEPs?</a>
34 </ol></li>
35 <li><a href="#rules">Rules</a>
36 <ol>
37 <li><a href="#bounds">What happens if an array index is out of bounds?</a>
38 <li><a href="#negative">Can array indices be negative?</a>
39 <li><a href="#compare">Can I compare two values computed with GEPs?</a>
40 <li><a href="#types">Can I do GEP with a different pointer type than the type of the underlying object?</a>
41 <li><a href="#null">Can I cast an object's address to integer and add it to null?</a>
42 <li><a href="#ptrdiff">Can I compute the distance between two objects, and add that value to one address to compute the other address?</a>
43 <li><a href="#tbaa">Can I do type-based alias analysis on LLVM IR?</a>
44 <li><a href="#overflow">What happens if a GEP computation overflows?</a>
45 <li><a href="#check">How can I tell if my front-end is following the rules?</a>
46 </ol></li>
47 <li><a href="#rationale">Rationale</a>
48 <ol>
49 <li><a href="#goals">Why is GEP designed this way?</a></li>
50 <li><a href="#i32">Why do struct member indices always use i32?</a></li>
51 <li><a href="#uglygep">What's an uglygep?</a>
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]52 </ol></li>
53 <li><a href="#summary">Summary</a></li>
54</ol>
55
56<div class="doc_author">
57 <p>Written by: <a href="mailto:rspencer@reidspencer.com">Reid Spencer</a>.</p>
58</div>
59
60
61<!-- *********************************************************************** -->
62<div class="doc_section"><a name="intro"><b>Introduction</b></a></div>
63<!-- *********************************************************************** -->
Dan Gohmanb02c08c2010-02-25 18:16:03 +0000[diff] [blame]64
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]65<div class="doc_text">
66 <p>This document seeks to dispel the mystery and confusion surrounding LLVM's
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]67 <a href="LangRef.html#i_getelementptr">GetElementPtr</a> (GEP) instruction.
68 Questions about the wily GEP instruction are
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]69 probably the most frequently occurring questions once a developer gets down to
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]70 coding with LLVM. Here we lay out the sources of confusion and show that the
71 GEP instruction is really quite simple.
72 </p>
73</div>
74
75<!-- *********************************************************************** -->
Dan Gohmanb02c08c2010-02-25 18:16:03 +0000[diff] [blame]76<div class="doc_section"><a name="addresses"><b>Address Computation</b></a></div>
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]77<!-- *********************************************************************** -->
78<div class="doc_text">
79 <p>When people are first confronted with the GEP instruction, they tend to
80 relate it to known concepts from other programming paradigms, most notably C
Dan Gohmanb02c08c2010-02-25 18:16:03 +0000[diff] [blame]81 array indexing and field selection. GEP closely resembles C array indexing
82 and field selection, however it's is a little different and this leads to
83 the following questions.</p>
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]84</div>
85
86<!-- *********************************************************************** -->
87<div class="doc_subsection">
Reid Spencer919d3712006-08-15 03:32:10 +0000[diff] [blame]88 <a name="firstptr"><b>What is the first index of the GEP instruction?</b></a>
89</div>
90<div class="doc_text">
Reid Spencer80a4d052006-08-15 03:43:31 +0000[diff] [blame]91 <p>Quick answer: The index stepping through the first operand.</p>
92 <p>The confusion with the first index usually arises from thinking about
93 the GetElementPtr instruction as if it was a C index operator. They aren't the
94 same. For example, when we write, in "C":</p>
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]95
96<div class="doc_code">
97<pre>
98AType *Foo;
99...
100X = &amp;Foo-&gt;F;
101</pre>
102</div>
103
Reid Spencereda573d2006-08-15 04:00:29 +0000[diff] [blame]104 <p>it is natural to think that there is only one index, the selection of the
105 field <tt>F</tt>. However, in this example, <tt>Foo</tt> is a pointer. That
Dan Gohmanb02c08c2010-02-25 18:16:03 +0000[diff] [blame]106 pointer must be indexed explicitly in LLVM. C, on the other hand, indices
Jim Laskeye3c312f2006-08-15 08:14:19 +0000[diff] [blame]107 through it transparently. To arrive at the same address location as the C
Reid Spencereda573d2006-08-15 04:00:29 +0000[diff] [blame]108 code, you would provide the GEP instruction with two index operands. The
109 first operand indexes through the pointer; the second operand indexes the
110 field <tt>F</tt> of the structure, just as if you wrote:</p>
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]111
112<div class="doc_code">
113<pre>
114X = &amp;Foo[0].F;
115</pre>
116</div>
117
Reid Spencer80a4d052006-08-15 03:43:31 +0000[diff] [blame]118 <p>Sometimes this question gets rephrased as:</p>
Chris Lattner4a5dfee2006-08-17 03:26:50 +0000[diff] [blame]119 <blockquote><p><i>Why is it okay to index through the first pointer, but
120 subsequent pointers won't be dereferenced?</i></p></blockquote>
Reid Spencer919d3712006-08-15 03:32:10 +0000[diff] [blame]121 <p>The answer is simply because memory does not have to be accessed to
122 perform the computation. The first operand to the GEP instruction must be a
123 value of a pointer type. The value of the pointer is provided directly to
Reid Spencer1c6f87d2006-08-15 03:57:05 +0000[diff] [blame]124 the GEP instruction as an operand without any need for accessing memory. It
125 must, therefore be indexed and requires an index operand. Consider this
126 example:</p>
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]127
128<div class="doc_code">
129<pre>
130struct munger_struct {
131 int f1;
132 int f2;
133};
134void munge(struct munger_struct *P) {
135 P[0].f1 = P[1].f1 + P[2].f2;
136}
137...
138munger_struct Array[3];
139...
140munge(Array);
141</pre>
142</div>
143
Reid Spencer919d3712006-08-15 03:32:10 +0000[diff] [blame]144 <p>In this "C" example, the front end compiler (llvm-gcc) will generate three
145 GEP instructions for the three indices through "P" in the assignment
146 statement. The function argument <tt>P</tt> will be the first operand of each
Reid Spencer10146472006-08-16 05:53:32 +0000[diff] [blame]147 of these GEP instructions. The second operand indexes through that pointer.
148 The third operand will be the field offset into the
149 <tt>struct munger_struct</tt> type, for either the <tt>f1</tt> or
Reid Spencer919d3712006-08-15 03:32:10 +0000[diff] [blame]150 <tt>f2</tt> field. So, in LLVM assembly the <tt>munge</tt> function looks
151 like:</p>
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]152
153<div class="doc_code">
154<pre>
155void %munge(%struct.munger_struct* %P) {
156entry:
157 %tmp = getelementptr %struct.munger_struct* %P, i32 1, i32 0
158 %tmp = load i32* %tmp
159 %tmp6 = getelementptr %struct.munger_struct* %P, i32 2, i32 1
160 %tmp7 = load i32* %tmp6
161 %tmp8 = add i32 %tmp7, %tmp
162 %tmp9 = getelementptr %struct.munger_struct* %P, i32 0, i32 0
163 store i32 %tmp8, i32* %tmp9
164 ret void
165}
166</pre>
167</div>
168
Reid Spencer919d3712006-08-15 03:32:10 +0000[diff] [blame]169 <p>In each case the first operand is the pointer through which the GEP
170 instruction starts. The same is true whether the first operand is an
171 argument, allocated memory, or a global variable. </p>
172 <p>To make this clear, let's consider a more obtuse example:</p>
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]173
174<div class="doc_code">
175<pre>
Dan Gohmanb02c08c2010-02-25 18:16:03 +0000[diff] [blame]176%MyVar = uninitialized global i32
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]177...
178%idx1 = getelementptr i32* %MyVar, i64 0
179%idx2 = getelementptr i32* %MyVar, i64 1
180%idx3 = getelementptr i32* %MyVar, i64 2
181</pre>
182</div>
183
Reid Spencer919d3712006-08-15 03:32:10 +0000[diff] [blame]184 <p>These GEP instructions are simply making address computations from the
185 base address of <tt>MyVar</tt>. They compute, as follows (using C syntax):
186 </p>
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]187
188<div class="doc_code">
189<pre>
190idx1 = (char*) &amp;MyVar + 0
191idx2 = (char*) &amp;MyVar + 4
192idx3 = (char*) &amp;MyVar + 8
193</pre>
194</div>
195
Reid Spencerb913a512007-02-09 17:56:02 +0000[diff] [blame]196 <p>Since the type <tt>i32</tt> is known to be four bytes long, the indices
Reid Spencer919d3712006-08-15 03:32:10 +0000[diff] [blame]197 0, 1 and 2 translate into memory offsets of 0, 4, and 8, respectively. No
198 memory is accessed to make these computations because the address of
199 <tt>%MyVar</tt> is passed directly to the GEP instructions.</p>
200 <p>The obtuse part of this example is in the cases of <tt>%idx2</tt> and
201 <tt>%idx3</tt>. They result in the computation of addresses that point to
202 memory past the end of the <tt>%MyVar</tt> global, which is only one
Reid Spencerb913a512007-02-09 17:56:02 +0000[diff] [blame]203 <tt>i32</tt> long, not three <tt>i32</tt>s long. While this is legal in LLVM,
Reid Spencer919d3712006-08-15 03:32:10 +0000[diff] [blame]204 it is inadvisable because any load or store with the pointer that results
205 from these GEP instructions would produce undefined results.</p>
206</div>
207
208<!-- *********************************************************************** -->
209<div class="doc_subsection">
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]210 <a name="extra_index"><b>Why is the extra 0 index required?</b></a>
211</div>
212<!-- *********************************************************************** -->
213<div class="doc_text">
214 <p>Quick answer: there are no superfluous indices.</p>
215 <p>This question arises most often when the GEP instruction is applied to a
216 global variable which is always a pointer type. For example, consider
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]217 this:</p>
218
219<div class="doc_code">
220<pre>
221%MyStruct = uninitialized global { float*, i32 }
222...
223%idx = getelementptr { float*, i32 }* %MyStruct, i64 0, i32 1
224</pre>
225</div>
226
Reid Spencerb913a512007-02-09 17:56:02 +0000[diff] [blame]227 <p>The GEP above yields an <tt>i32*</tt> by indexing the <tt>i32</tt> typed
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]228 field of the structure <tt>%MyStruct</tt>. When people first look at it, they
Reid Spencerb913a512007-02-09 17:56:02 +0000[diff] [blame]229 wonder why the <tt>i64 0</tt> index is needed. However, a closer inspection
Reid Spencer919d3712006-08-15 03:32:10 +0000[diff] [blame]230 of how globals and GEPs work reveals the need. Becoming aware of the following
Dan Gohmanb02c08c2010-02-25 18:16:03 +0000[diff] [blame]231 facts will dispel the confusion:</p>
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]232 <ol>
Reid Spencerb913a512007-02-09 17:56:02 +0000[diff] [blame]233 <li>The type of <tt>%MyStruct</tt> is <i>not</i> <tt>{ float*, i32 }</tt>
234 but rather <tt>{ float*, i32 }*</tt>. That is, <tt>%MyStruct</tt> is a
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]235 pointer to a structure containing a pointer to a <tt>float</tt> and an
Reid Spencerb913a512007-02-09 17:56:02 +0000[diff] [blame]236 <tt>i32</tt>.</li>
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]237 <li>Point #1 is evidenced by noticing the type of the first operand of
238 the GEP instruction (<tt>%MyStruct</tt>) which is
Reid Spencerb913a512007-02-09 17:56:02 +0000[diff] [blame]239 <tt>{ float*, i32 }*</tt>.</li>
240 <li>The first index, <tt>i64 0</tt> is required to step over the global
Reid Spencer919d3712006-08-15 03:32:10 +0000[diff] [blame]241 variable <tt>%MyStruct</tt>. Since the first argument to the GEP
242 instruction must always be a value of pointer type, the first index
243 steps through that pointer. A value of 0 means 0 elements offset from that
244 pointer.</li>
Reid Spencerb913a512007-02-09 17:56:02 +0000[diff] [blame]245 <li>The second index, <tt>i32 1</tt> selects the second field of the
246 structure (the <tt>i32</tt>). </li>
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]247 </ol>
248</div>
249
250<!-- *********************************************************************** -->
251<div class="doc_subsection">
252 <a name="deref"><b>What is dereferenced by GEP?</b></a>
253</div>
254<div class="doc_text">
255 <p>Quick answer: nothing.</p>
256 <p>The GetElementPtr instruction dereferences nothing. That is, it doesn't
Reid Spencer919d3712006-08-15 03:32:10 +0000[diff] [blame]257 access memory in any way. That's what the Load and Store instructions are for.
258 GEP is only involved in the computation of addresses. For example, consider
259 this:</p>
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]260
261<div class="doc_code">
262<pre>
263%MyVar = uninitialized global { [40 x i32 ]* }
264...
265%idx = getelementptr { [40 x i32]* }* %MyVar, i64 0, i32 0, i64 0, i64 17
266</pre>
267</div>
268
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]269 <p>In this example, we have a global variable, <tt>%MyVar</tt> that is a
270 pointer to a structure containing a pointer to an array of 40 ints. The
Reid Spencer80a4d052006-08-15 03:43:31 +0000[diff] [blame]271 GEP instruction seems to be accessing the 18th integer of the structure's
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]272 array of ints. However, this is actually an illegal GEP instruction. It
273 won't compile. The reason is that the pointer in the structure <i>must</i>
274 be dereferenced in order to index into the array of 40 ints. Since the
275 GEP instruction never accesses memory, it is illegal.</p>
276 <p>In order to access the 18th integer in the array, you would need to do the
277 following:</p>
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]278
279<div class="doc_code">
280<pre>
281%idx = getelementptr { [40 x i32]* }* %, i64 0, i32 0
282%arr = load [40 x i32]** %idx
283%idx = getelementptr [40 x i32]* %arr, i64 0, i64 17
284</pre>
285</div>
286
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]287 <p>In this case, we have to load the pointer in the structure with a load
288 instruction before we can index into the array. If the example was changed
289 to:</p>
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]290
291<div class="doc_code">
292<pre>
293%MyVar = uninitialized global { [40 x i32 ] }
294...
295%idx = getelementptr { [40 x i32] }*, i64 0, i32 0, i64 17
296</pre>
297</div>
298
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]299 <p>then everything works fine. In this case, the structure does not contain a
Reid Spencer80a4d052006-08-15 03:43:31 +0000[diff] [blame]300 pointer and the GEP instruction can index through the global variable,
Reid Spencerb913a512007-02-09 17:56:02 +0000[diff] [blame]301 into the first field of the structure and access the 18th <tt>i32</tt> in the
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]302 array there.</p>
303</div>
304
305<!-- *********************************************************************** -->
306<div class="doc_subsection">
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]307 <a name="lead0"><b>Why don't GEP x,0,0,1 and GEP x,1 alias?</b></a>
308</div>
309<div class="doc_text">
310 <p>Quick Answer: They compute different address locations.</p>
311 <p>If you look at the first indices in these GEP
312 instructions you find that they are different (0 and 1), therefore the address
313 computation diverges with that index. Consider this example:</p>
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]314
315<div class="doc_code">
316<pre>
317%MyVar = global { [10 x i32 ] }
Dan Gohmanb02c08c2010-02-25 18:16:03 +0000[diff] [blame]318%idx1 = getelementptr { [10 x i32 ] }* %MyVar, i64 0, i32 0, i64 1
319%idx2 = getelementptr { [10 x i32 ] }* %MyVar, i64 1
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]320</pre>
321</div>
322
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]323 <p>In this example, <tt>idx1</tt> computes the address of the second integer
Misha Brukmanfc13d1c2009-08-18 19:18:40 +0000[diff] [blame]324 in the array that is in the structure in <tt>%MyVar</tt>, that is
325 <tt>MyVar+4</tt>. The type of <tt>idx1</tt> is <tt>i32*</tt>. However,
326 <tt>idx2</tt> computes the address of <i>the next</i> structure after
327 <tt>%MyVar</tt>. The type of <tt>idx2</tt> is <tt>{ [10 x i32] }*</tt> and its
328 value is equivalent to <tt>MyVar + 40</tt> because it indexes past the ten
329 4-byte integers in <tt>MyVar</tt>. Obviously, in such a situation, the
330 pointers don't alias.</p>
331
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]332</div>
333
334<!-- *********************************************************************** -->
335<div class="doc_subsection">
Jim Laskey19af0e12006-08-15 12:15:08 +0000[diff] [blame]336 <a name="trail0"><b>Why do GEP x,1,0,0 and GEP x,1 alias?</b></a>
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]337</div>
338<div class="doc_text">
339 <p>Quick Answer: They compute the same address location.</p>
340 <p>These two GEP instructions will compute the same address because indexing
341 through the 0th element does not change the address. However, it does change
342 the type. Consider this example:</p>
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]343
344<div class="doc_code">
345<pre>
346%MyVar = global { [10 x i32 ] }
Dan Gohmanb02c08c2010-02-25 18:16:03 +0000[diff] [blame]347%idx1 = getelementptr { [10 x i32 ] }* %MyVar, i64 1, i32 0, i64 0
348%idx2 = getelementptr { [10 x i32 ] }* %MyVar, i64 1
Bill Wendling32759082008-01-04 12:04:32 +0000[diff] [blame]349</pre>
350</div>
351
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]352 <p>In this example, the value of <tt>%idx1</tt> is <tt>%MyVar+40</tt> and
Reid Spencerb913a512007-02-09 17:56:02 +0000[diff] [blame]353 its type is <tt>i32*</tt>. The value of <tt>%idx2</tt> is also
354 <tt>MyVar+40</tt> but its type is <tt>{ [10 x i32] }*</tt>.</p>
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]355</div>
356
357<!-- *********************************************************************** -->
Dan Gohmanb02c08c2010-02-25 18:16:03 +0000[diff] [blame]358
359<div class="doc_subsection">
360 <a name="vectors"><b>Can GEP index into vector elements?</b></a>
361</div>
362<div class="doc_text">
363 <p>This hasn't always been forcefully disallowed, though it's not recommended.
364 It leads to awkward special cases in the optimizers, and fundamental
365 inconsistency in the IR. In the future, it will probably be outright
366 disallowed.</p>
367
368</div>
369
370<!-- *********************************************************************** -->
371
372<div class="doc_subsection">
373 <a name="unions"><b>Can GEP index into unions?</b></a>
374</div>
375<div class="doc_text">
376 <p>Unknown.</p>
377
378</div>
379
380<!-- *********************************************************************** -->
381
382<div class="doc_subsection">
383 <a name="addrspace"><b>What effect do address spaces have on GEPs?</b></a>
384</div>
385<div class="doc_text">
386 <p>None, except that the address space qualifier on the first operand pointer
387 type always matches the address space qualifier on the result type.</p>
388
389</div>
390
391<!-- *********************************************************************** -->
392
393<div class="doc_subsection">
394 <a name="int"><b>How is GEP different from ptrtoint, arithmetic,
395 and inttoptr?</b></a>
396</div>
397<div class="doc_text">
398 <p>It's very similar; there are only subtle differences.</p>
399
400 <p>With ptrtoint, you have to pick an integer type. One approach is to pick i64;
401 this is safe on everything LLVM supports (LLVM internally assumes pointers
402 are never wider than 64 bits in many places), and the optimizer will actually
403 narrow the i64 arithmetic down to the actual pointer size on targets which
404 don't support 64-bit arithmetic in most cases. However, there are some cases
405 where it doesn't do this. With GEP you can avoid this problem.
406
407 <p>Also, GEP carries additional pointer aliasing rules. It's invalid to take a
408 GEP from one object, address into a different separately allocated
409 object, and dereference it. IR producers (front-ends) must follow this rule,
410 and consumers (optimizers, specifically alias analysis) benefit from being
411 able to rely on it. See the <a href="#rules">Rules</a> section for more
412 information.</p>
413
414 <p>And, GEP is more concise in common cases.</p>
415
416 <p>However, for the underlying integer computation implied, there
417 is no difference.</p>
418
419</div>
420
421<!-- *********************************************************************** -->
422
423<div class="doc_subsection">
424 <a name="be"><b>I'm writing a backend for a target which needs custom
425 lowering for GEP. How do I do this?</b></a>
426</div>
427<div class="doc_text">
428 <p>You don't. The integer computation implied by a GEP is target-independent.
429 Typically what you'll need to do is make your backend pattern-match
430 expressions trees involving ADD, MUL, etc., which are what GEP is lowered
431 into. This has the advantage of letting your code work correctly in more
432 cases.</p>
433
434 <p>GEP does use target-dependent parameters for the size and layout of data
435 types, which targets can customize.</p>
436
437 <p>If you require support for addressing units which are not 8 bits, you'll
438 need to fix a lot of code in the backend, with GEP lowering being only a
439 small piece of the overall picture.</p>
440
441</div>
442
443<!-- *********************************************************************** -->
444
445<div class="doc_subsection">
446 <a name="vla"><b>How does VLA addressing work with GEPs?</b></a>
447</div>
448<div class="doc_text">
449 <p>GEPs don't natively support VLAs. LLVM's type system is entirely static,
450 and GEP address computations are guided by an LLVM type.</p>
451
452 <p>VLA indices can be implemented as linearized indices. For example, an
453 expression like X[a][b][c], must be effectively lowered into a form
454 like X[a*m+b*n+c], so that it appears to the GEP as a single-dimensional
455 array reference.</p>
456
457 <p>This means if you want to write an analysis which understands array
458 indices and you want to support VLAs, your code will have to be
459 prepared to reverse-engineer the linearization. One way to solve this
460 problem is to use the ScalarEvolution library, which always presents
461 VLA and non-VLA indexing in the same manner.</p>
462</div>
463
464<!-- *********************************************************************** -->
465<div class="doc_section"><a name="rules"><b>Rules</b></a></div>
466<!-- *********************************************************************** -->
467
468<!-- *********************************************************************** -->
469
470<div class="doc_subsection">
471 <a name="bounds"><b>What happens if an array index is out of bounds?</b></a>
472</div>
473<div class="doc_text">
474 <p>There are two senses in which an array index can be out of bounds.</p>
475
476 <p>First, there's the array type which comes from the (static) type of
477 the first operand to the GEP. Indices greater than the number of elements
478 in the corresponding static array type are valid. There is no problem with
479 out of bounds indices in this sense. Indexing into an array only depends
480 on the size of the array element, not the number of elements.</p>
481
482 <p>A common example of how this is used is arrays where the size is not known.
483 It's common to use array types with zero length to represent these. The
484 fact that the static type says there are zero elements is irrelevant; it's
485 perfectly valid to compute arbitrary element indices, as the computation
486 only depends on the size of the array element, not the number of
487 elements. Note that zero-sized arrays are not a special case here.</p>
488
489 <p>This sense is unconnected with <tt>inbounds</tt> keyword. The
490 <tt>inbounds</tt> keyword is designed to describe low-level pointer
491 arithmetic overflow conditions, rather than high-level array
492 indexing rules.
493
494 <p>Analysis passes which wish to understand array indexing should not
495 assume that the static array type bounds are respected.</p>
496
497 <p>The second sense of being out of bounds is computing an address that's
498 beyond the actual underlying allocated object.</p>
499
500 <p>With the <tt>inbounds</tt> keyword, the result value of the GEP is
501 undefined if the address is outside the actual underlying allocated
502 object and not the address one-past-the-end.</p>
503
504 <p>Without the <tt>inbounds</tt> keyword, there are no restrictions
505 on computing out-of-bounds addresses. Obviously, performing a load or
506 a store requires an address of allocated and sufficiently aligned
507 memory. But the GEP itself is only concerned with computing addresses.</p>
508
509</div>
510
511<!-- *********************************************************************** -->
512<div class="doc_subsection">
513 <a name="negative"><b>Can array indices be negative?</b></a>
514</div>
515<div class="doc_text">
516 <p>Yes. This is basically a special case of array indices being out
517 of bounds.</p>
518
519</div>
520
521<!-- *********************************************************************** -->
522<div class="doc_subsection">
523 <a name="compare"><b>Can I compare two values computed with GEPs?</b></a>
524</div>
525<div class="doc_text">
526 <p>Yes. If both addresses are within the same allocated object, or
527 one-past-the-end, you'll get the comparison result you expect. If either
528 is outside of it, integer arithmetic wrapping may occur, so the
529 comparison may not be meaningful.</p>
530
531</div>
532
533<!-- *********************************************************************** -->
534<div class="doc_subsection">
535 <a name="types"><b>Can I do GEP with a different pointer type than the type of
536 the underlying object?</b></a>
537</div>
538<div class="doc_text">
539 <p>Yes. There are no restrictions on bitcasting a pointer value to an arbitrary
540 pointer type. The types in a GEP serve only to define the parameters for the
541 underlying integer computation. They need not correspond with the actual
542 type of the underlying object.</p>
543
544 <p>Furthermore, loads and stores don't have to use the same types as the type
545 of the underlying object. Types in this context serve only to specify
546 memory size and alignment. Beyond that there are merely a hint to the
547 optimizer indicating how the value will likely be used.</p>
548
549</div>
550
551<!-- *********************************************************************** -->
552<div class="doc_subsection">
553 <a name="null"><b>Can I cast an object's address to integer and add it
554 to null?</b></a>
555</div>
556<div class="doc_text">
557 <p>You can compute an address that way, but if you use GEP to do the add,
558 you can't use that pointer to actually access the object, unless the
559 object is managed outside of LLVM.</p>
560
561 <p>The underlying integer computation is sufficiently defined; null has a
562 defined value -- zero -- and you can add whatever value you want to it.</p>
563
564 <p>However, it's invalid to access (load from or store to) an LLVM-aware
565 object with such a pointer. This includes GlobalVariables, Allocas, and
566 objects pointed to by noalias pointers.</p>
567
568 <p>If you really need this functionality, you can do the arithmetic with
569 explicit integer instructions, and use inttoptr to convert the result to
570 an address. Most of GEP's special aliasing rules do not apply to pointers
571 computed from ptrtoint, arithmetic, and inttoptr sequences.</p>
572
573</div>
574
575<!-- *********************************************************************** -->
576<div class="doc_subsection">
577 <a name="ptrdiff"><b>Can I compute the distance between two objects, and add
578 that value to one address to compute the other address?</b></a>
579</div>
580<div class="doc_text">
581 <p>As with arithmetic on null, You can use GEP to compute an address that
582 way, but you can't use that pointer to actually access the object if you
583 do, unless the object is managed outside of LLVM.</p>
584
585 <p>Also as above, ptrtoint and inttoptr provide an alternative way to do this
586 which do not have this restriction.</p>
587
588</div>
589
590<!-- *********************************************************************** -->
591<div class="doc_subsection">
592 <a name="tbaa"><b>Can I do type-based alias analysis on LLVM IR?</b></a>
593</div>
594<div class="doc_text">
595 <p>You can't do type-based alias analysis using LLVM's built-in type system,
596 because LLVM has no restrictions on mixing types in addressing, loads or
597 stores.</p>
598
599 <p>It would be possible to add special annotations to the IR, probably using
600 metadata, to describe a different type system (such as the C type system),
601 and do type-based aliasing on top of that. This is a much bigger
602 undertaking though.</p>
603
604</div>
605
606<!-- *********************************************************************** -->
607
608<div class="doc_subsection">
609 <a name="overflow"><b>What happens if a GEP computation overflows?</b></a>
610</div>
611<div class="doc_text">
612 <p>If the GEP has the <tt>inbounds</tt> keyword, the result value is
613 undefined.</p>
614
615 <p>Otherwise, the result value is the result from evaluating the implied
616 two's complement integer computation. However, since there's no
617 guarantee of where an object will be allocated in the address space,
618 such values have limited meaning.</p>
619
620</div>
621
622<!-- *********************************************************************** -->
623
624<div class="doc_subsection">
625 <a name="check"><b>How can I tell if my front-end is following the
626 rules?</b></a>
627</div>
628<div class="doc_text">
629 <p>There is currently no checker for the getelementptr rules. Currently,
630 the only way to do this is to manually check each place in your front-end
631 where GetElementPtr operators are created.</p>
632
633 <p>It's not possible to write a checker which could find all rule
634 violations statically. It would be possible to write a checker which
635 works by instrumenting the code with dynamic checks though. Alternatively,
636 it would be possible to write a static checker which catches a subset of
637 possible problems. However, no such checker exists today.</p>
638
639</div>
640
641<!-- *********************************************************************** -->
642<div class="doc_section"><a name="rationale"><b>Rationale</b></a></div>
643<!-- *********************************************************************** -->
644
645<!-- *********************************************************************** -->
646
647<div class="doc_subsection">
648 <a name="goals"><b>Why is GEP designed this way?</b></a>
649</div>
650<div class="doc_text">
651 <p>The design of GEP has the following goals, in rough unofficial
652 order of priority:</p>
653 <ul>
654 <li>Support C, C-like languages, and languages which can be
655 conceptually lowered into C (this covers a lot).</li>
656 <li>Support optimizations such as those that are common in
Dan Gohmanff70fe42010-07-06 15:26:33 +0000[diff] [blame]657 C compilers. In particular, GEP is a cornerstone of LLVM's
658 <a href="LangRef.html#pointeraliasing">pointer aliasing model</a>.</li>
Dan Gohmanb02c08c2010-02-25 18:16:03 +0000[diff] [blame]659 <li>Provide a consistent method for computing addresses so that
660 address computations don't need to be a part of load and
661 store instructions in the IR.</li>
662 <li>Support non-C-like languages, to the extent that it doesn't
663 interfere with other goals.</li>
664 <li>Minimize target-specific information in the IR.</li>
665 </ul>
666</div>
667
668<!-- *********************************************************************** -->
669<div class="doc_subsection">
670 <a name="i32"><b>Why do struct member indices always use i32?</b></a>
671</div>
672<div class="doc_text">
673 <p>The specific type i32 is probably just a historical artifact, however it's
674 wide enough for all practical purposes, so there's been no need to change it.
675 It doesn't necessarily imply i32 address arithmetic; it's just an identifier
676 which identifies a field in a struct. Requiring that all struct indices be
677 the same reduces the range of possibilities for cases where two GEPs are
678 effectively the same but have distinct operand types.</p>
679
680</div>
681
682<!-- *********************************************************************** -->
683
684<div class="doc_subsection">
685 <a name="uglygep"><b>What's an uglygep?</b></a>
686</div>
687<div class="doc_text">
688 <p>Some LLVM optimizers operate on GEPs by internally lowering them into
689 more primitive integer expressions, which allows them to be combined
690 with other integer expressions and/or split into multiple separate
691 integer expressions. If they've made non-trivial changes, translating
692 back into LLVM IR can involve reverse-engineering the structure of
693 the addressing in order to fit it into the static type of the original
694 first operand. It isn't always possibly to fully reconstruct this
695 structure; sometimes the underlying addressing doesn't correspond with
696 the static type at all. In such cases the optimizer instead will emit
697 a GEP with the base pointer casted to a simple address-unit pointer,
698 using the name "uglygep". This isn't pretty, but it's just as
699 valid, and it's sufficient to preserve the pointer aliasing guarantees
700 that GEP provides.</p>
701
702</div>
703
704<!-- *********************************************************************** -->
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]705<div class="doc_section"><a name="summary"><b>Summary</b></a></div>
706<!-- *********************************************************************** -->
707
708<div class="doc_text">
709 <p>In summary, here's some things to always remember about the GetElementPtr
710 instruction:</p>
711 <ol>
712 <li>The GEP instruction never accesses memory, it only provides pointer
713 computations.</li>
714 <li>The first operand to the GEP instruction is always a pointer and it must
715 be indexed.</li>
716 <li>There are no superfluous indices for the GEP instruction.</li>
717 <li>Trailing zero indices are superfluous for pointer aliasing, but not for
718 the types of the pointers.</li>
719 <li>Leading zero indices are not superfluous for pointer aliasing nor the
720 types of the pointers.</li>
721 </ol>
722</div>
723
724<!-- *********************************************************************** -->
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]725
726<hr>
727<address>
728 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]729 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]730 <a href="http://validator.w3.org/check/referer"><img
Misha Brukmanf00ddb02008-12-11 18:23:24 +0000[diff] [blame]731 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Reid Spencere00906f2006-08-10 20:15:58 +0000[diff] [blame]732 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br/>
733 Last modified: $Date$
734</address>
735</body>
736</html>