blob: 83bd667fac30e2f2356813300b73a9cd47dfaca2 [file] [log] [blame]
Misha Brukmanc501f552004-03-01 17:47:27 +00001<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Misha Brukman76307852003-11-08 01:05:38 +00003<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencercb84e432004-08-26 20:44:00 +00006 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <meta name="author" content="Chris Lattner">
8 <meta name="description"
9 content="LLVM Assembly Language Reference Manual.">
Misha Brukman76307852003-11-08 01:05:38 +000010 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattner757528b0b2004-05-23 21:06:01 +000012
Misha Brukman76307852003-11-08 01:05:38 +000013<body>
Chris Lattner757528b0b2004-05-23 21:06:01 +000014
Chris Lattner48b383b02003-11-25 01:02:51 +000015<div class="doc_title"> LLVM Language Reference Manual </div>
Chris Lattner2f7c9632001-06-06 20:29:01 +000016<ol>
Misha Brukman76307852003-11-08 01:05:38 +000017 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +000020 <li><a href="#highlevel">High Level Structure</a>
21 <ol>
22 <li><a href="#modulestructure">Module Structure</a></li>
Chris Lattnerd79749a2004-12-09 16:36:40 +000023 <li><a href="#linkage">Linkage Types</a></li>
Chris Lattner0132aff2005-05-06 22:57:40 +000024 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +000025 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner91c15c42006-01-23 23:23:47 +000026 <li><a href="#functionstructure">Functions</a></li>
Dan Gohmanef9462f2008-10-14 16:51:45 +000027 <li><a href="#aliasstructure">Aliases</a></li>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +000028 <li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel9eb525d2008-09-26 23:51:19 +000029 <li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen71183b62007-12-10 03:18:06 +000030 <li><a href="#gc">Garbage Collector Names</a></li>
Chris Lattner91c15c42006-01-23 23:23:47 +000031 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Reid Spencer50c723a2007-02-19 23:54:10 +000032 <li><a href="#datalayout">Data Layout</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +000033 </ol>
34 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +000035 <li><a href="#typesystem">Type System</a>
36 <ol>
Chris Lattner7824d182008-01-04 04:32:38 +000037 <li><a href="#t_classifications">Type Classifications</a></li>
Robert Bocchino820bc75b2006-02-17 21:18:08 +000038 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner48b383b02003-11-25 01:02:51 +000039 <ol>
Chris Lattner7824d182008-01-04 04:32:38 +000040 <li><a href="#t_floating">Floating Point Types</a></li>
41 <li><a href="#t_void">Void Type</a></li>
42 <li><a href="#t_label">Label Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +000043 </ol>
44 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +000045 <li><a href="#t_derived">Derived Types</a>
46 <ol>
Chris Lattner9a2e3cb2007-12-18 06:18:21 +000047 <li><a href="#t_integer">Integer Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +000048 <li><a href="#t_array">Array Type</a></li>
Misha Brukman76307852003-11-08 01:05:38 +000049 <li><a href="#t_function">Function Type</a></li>
50 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +000051 <li><a href="#t_struct">Structure Type</a></li>
Andrew Lenharth8df88e22006-12-08 17:13:00 +000052 <li><a href="#t_pstruct">Packed Structure Type</a></li>
Reid Spencer404a3252007-02-15 03:07:05 +000053 <li><a href="#t_vector">Vector Type</a></li>
Chris Lattner37b6b092005-04-25 17:34:15 +000054 <li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +000055 </ol>
56 </li>
57 </ol>
58 </li>
Chris Lattner6af02f32004-12-09 16:11:40 +000059 <li><a href="#constants">Constants</a>
Chris Lattner74d3f822004-12-09 17:30:23 +000060 <ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +000061 <li><a href="#simpleconstants">Simple Constants</a></li>
62 <li><a href="#aggregateconstants">Aggregate Constants</a></li>
63 <li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
64 <li><a href="#undefvalues">Undefined Values</a></li>
65 <li><a href="#constantexprs">Constant Expressions</a></li>
Chris Lattner74d3f822004-12-09 17:30:23 +000066 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +000067 </li>
Chris Lattner98f013c2006-01-25 23:47:57 +000068 <li><a href="#othervalues">Other Values</a>
69 <ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +000070 <li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Chris Lattner98f013c2006-01-25 23:47:57 +000071 </ol>
72 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +000073 <li><a href="#instref">Instruction Reference</a>
74 <ol>
75 <li><a href="#terminators">Terminator Instructions</a>
76 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +000077 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
78 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +000079 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
80 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +000081 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner08b7d5b2004-10-16 18:04:13 +000082 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +000083 </ol>
84 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +000085 <li><a href="#binaryops">Binary Operations</a>
86 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +000087 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
88 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
89 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Reid Spencer7e80b0b2006-10-26 06:15:43 +000090 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
91 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
92 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer7eb55b32006-11-02 01:53:59 +000093 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
94 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
95 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +000096 </ol>
97 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +000098 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
99 <ol>
Reid Spencer2ab01932007-02-02 13:57:07 +0000100 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
101 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
102 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000103 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000104 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000105 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000106 </ol>
107 </li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000108 <li><a href="#vectorops">Vector Operations</a>
109 <ol>
110 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
111 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
112 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000113 </ol>
114 </li>
Dan Gohmanb9d66602008-05-12 23:51:09 +0000115 <li><a href="#aggregateops">Aggregate Operations</a>
116 <ol>
117 <li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
118 <li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
119 </ol>
120 </li>
Chris Lattner6ab66722006-08-15 00:45:58 +0000121 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000122 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000123 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
124 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
125 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000126 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
127 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
128 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000129 </ol>
130 </li>
Reid Spencer97c5fa42006-11-08 01:18:52 +0000131 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000132 <ol>
133 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
134 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
135 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
136 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
137 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencer51b07252006-11-09 23:03:26 +0000138 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
139 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
140 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
141 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000142 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
143 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5b950642006-11-11 23:08:07 +0000144 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000145 </ol>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000146 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000147 <li><a href="#otherops">Other Operations</a>
148 <ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000149 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
150 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Nate Begemand2195702008-05-12 19:01:56 +0000151 <li><a href="#i_vicmp">'<tt>vicmp</tt>' Instruction</a></li>
152 <li><a href="#i_vfcmp">'<tt>vfcmp</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000153 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000154 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000155 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattner33337472006-01-13 23:26:01 +0000156 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000157 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000158 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000159 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000160 </li>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000161 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000162 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000163 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
164 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000165 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
166 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
167 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000168 </ol>
169 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000170 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
171 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000172 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
173 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
174 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000175 </ol>
176 </li>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000177 <li><a href="#int_codegen">Code Generator Intrinsics</a>
178 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000179 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
180 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
181 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
182 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
183 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
184 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
185 <li><a href="#int_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswellaa1c3c12004-04-09 16:43:20 +0000186 </ol>
187 </li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000188 <li><a href="#int_libc">Standard C Library Intrinsics</a>
189 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000190 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
191 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
192 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
193 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
194 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohmanb6324c12007-10-15 20:30:11 +0000195 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
196 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
197 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000198 </ol>
199 </li>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000200 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000201 <ol>
Reid Spencer96a5f022007-04-04 02:42:35 +0000202 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattnerb748c672006-01-16 22:34:14 +0000203 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
204 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
205 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Reid Spencer5bf54c82007-04-11 23:23:49 +0000206 <li><a href="#int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic </a></li>
207 <li><a href="#int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic </a></li>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000208 </ol>
209 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000210 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskey2211f492007-03-14 19:31:19 +0000211 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands86e01192007-09-11 14:10:23 +0000212 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands644f9172007-07-27 12:58:54 +0000213 <ol>
214 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands644f9172007-07-27 12:58:54 +0000215 </ol>
216 </li>
Bill Wendlingf85850f2008-11-18 22:10:53 +0000217 <li><a href="#int_atomics">Atomic intrinsics</a>
218 <ol>
219 <li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
220 <li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
221 <li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
222 <li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
223 <li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
224 <li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
225 <li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
226 <li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
227 <li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
228 <li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
229 <li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
230 <li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
231 <li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
232 </ol>
233 </li>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000234 <li><a href="#int_general">General intrinsics</a>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000235 <ol>
Reid Spencer5b2cb0f2007-07-20 19:59:11 +0000236 <li><a href="#int_var_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000237 '<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000238 <li><a href="#int_annotation">
Bill Wendling14313312008-11-19 05:56:17 +0000239 '<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +0000240 <li><a href="#int_trap">
Bill Wendling14313312008-11-19 05:56:17 +0000241 '<tt>llvm.trap</tt>' Intrinsic</a></li>
242 <li><a href="#int_stackprotector">
243 '<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Tanya Lattner293c0372007-09-21 22:59:12 +0000244 </ol>
Tanya Lattnercb1b9602007-06-15 20:50:54 +0000245 </li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000246 </ol>
247 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000248</ol>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000249
250<div class="doc_author">
251 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
252 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman76307852003-11-08 01:05:38 +0000253</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000254
Chris Lattner2f7c9632001-06-06 20:29:01 +0000255<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000256<div class="doc_section"> <a name="abstract">Abstract </a></div>
257<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000258
Misha Brukman76307852003-11-08 01:05:38 +0000259<div class="doc_text">
Chris Lattner48b383b02003-11-25 01:02:51 +0000260<p>This document is a reference manual for the LLVM assembly language.
Bill Wendling6e03f9a2008-08-05 22:29:16 +0000261LLVM is a Static Single Assignment (SSA) based representation that provides
Chris Lattner67c37d12008-08-05 18:29:16 +0000262type safety, low-level operations, flexibility, and the capability of
263representing 'all' high-level languages cleanly. It is the common code
Chris Lattner48b383b02003-11-25 01:02:51 +0000264representation used throughout all phases of the LLVM compilation
265strategy.</p>
Misha Brukman76307852003-11-08 01:05:38 +0000266</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000267
Chris Lattner2f7c9632001-06-06 20:29:01 +0000268<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000269<div class="doc_section"> <a name="introduction">Introduction</a> </div>
270<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000271
Misha Brukman76307852003-11-08 01:05:38 +0000272<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000273
Chris Lattner48b383b02003-11-25 01:02:51 +0000274<p>The LLVM code representation is designed to be used in three
Gabor Greifa54634a2007-07-06 22:07:22 +0000275different forms: as an in-memory compiler IR, as an on-disk bitcode
Chris Lattner48b383b02003-11-25 01:02:51 +0000276representation (suitable for fast loading by a Just-In-Time compiler),
277and as a human readable assembly language representation. This allows
278LLVM to provide a powerful intermediate representation for efficient
279compiler transformations and analysis, while providing a natural means
280to debug and visualize the transformations. The three different forms
281of LLVM are all equivalent. This document describes the human readable
282representation and notation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000283
John Criswell4a3327e2005-05-13 22:25:59 +0000284<p>The LLVM representation aims to be light-weight and low-level
Chris Lattner48b383b02003-11-25 01:02:51 +0000285while being expressive, typed, and extensible at the same time. It
286aims to be a "universal IR" of sorts, by being at a low enough level
287that high-level ideas may be cleanly mapped to it (similar to how
288microprocessors are "universal IR's", allowing many source languages to
289be mapped to them). By providing type information, LLVM can be used as
290the target of optimizations: for example, through pointer analysis, it
291can be proven that a C automatic variable is never accessed outside of
292the current function... allowing it to be promoted to a simple SSA
293value instead of a memory location.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000294
Misha Brukman76307852003-11-08 01:05:38 +0000295</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000296
Chris Lattner2f7c9632001-06-06 20:29:01 +0000297<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000298<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000299
Misha Brukman76307852003-11-08 01:05:38 +0000300<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000301
Chris Lattner48b383b02003-11-25 01:02:51 +0000302<p>It is important to note that this document describes 'well formed'
303LLVM assembly language. There is a difference between what the parser
304accepts and what is considered 'well formed'. For example, the
305following instruction is syntactically okay, but not well formed:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000306
Bill Wendling3716c5d2007-05-29 09:04:49 +0000307<div class="doc_code">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000308<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000309%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattner757528b0b2004-05-23 21:06:01 +0000310</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000311</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000312
Chris Lattner48b383b02003-11-25 01:02:51 +0000313<p>...because the definition of <tt>%x</tt> does not dominate all of
314its uses. The LLVM infrastructure provides a verification pass that may
315be used to verify that an LLVM module is well formed. This pass is
John Criswell4a3327e2005-05-13 22:25:59 +0000316automatically run by the parser after parsing input assembly and by
Gabor Greifa54634a2007-07-06 22:07:22 +0000317the optimizer before it outputs bitcode. The violations pointed out
Chris Lattner48b383b02003-11-25 01:02:51 +0000318by the verifier pass indicate bugs in transformation passes or input to
319the parser.</p>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000320</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000321
Chris Lattner87a3dbe2007-10-03 17:34:29 +0000322<!-- Describe the typesetting conventions here. -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000323
Chris Lattner2f7c9632001-06-06 20:29:01 +0000324<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000325<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000326<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000327
Misha Brukman76307852003-11-08 01:05:38 +0000328<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000329
Reid Spencerb23b65f2007-08-07 14:34:28 +0000330 <p>LLVM identifiers come in two basic types: global and local. Global
331 identifiers (functions, global variables) begin with the @ character. Local
332 identifiers (register names, types) begin with the % character. Additionally,
Dan Gohmanef9462f2008-10-14 16:51:45 +0000333 there are three different formats for identifiers, for different purposes:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000334
Chris Lattner2f7c9632001-06-06 20:29:01 +0000335<ol>
Reid Spencerb23b65f2007-08-07 14:34:28 +0000336 <li>Named values are represented as a string of characters with their prefix.
337 For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
338 regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
Chris Lattnerd79749a2004-12-09 16:36:40 +0000339 Identifiers which require other characters in their names can be surrounded
Daniel Dunbar0f8155a2008-10-14 23:51:43 +0000340 with quotes. Special characters may be escaped using "\xx" where xx is the
341 ASCII code for the character in hexadecimal. In this way, any character can
342 be used in a name value, even quotes themselves.
Chris Lattnerd79749a2004-12-09 16:36:40 +0000343
Reid Spencerb23b65f2007-08-07 14:34:28 +0000344 <li>Unnamed values are represented as an unsigned numeric value with their
345 prefix. For example, %12, @2, %44.</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000346
Reid Spencer8f08d802004-12-09 18:02:53 +0000347 <li>Constants, which are described in a <a href="#constants">section about
348 constants</a>, below.</li>
Misha Brukman76307852003-11-08 01:05:38 +0000349</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000350
Reid Spencerb23b65f2007-08-07 14:34:28 +0000351<p>LLVM requires that values start with a prefix for two reasons: Compilers
Chris Lattnerd79749a2004-12-09 16:36:40 +0000352don't need to worry about name clashes with reserved words, and the set of
353reserved words may be expanded in the future without penalty. Additionally,
354unnamed identifiers allow a compiler to quickly come up with a temporary
355variable without having to avoid symbol table conflicts.</p>
356
Chris Lattner48b383b02003-11-25 01:02:51 +0000357<p>Reserved words in LLVM are very similar to reserved words in other
Reid Spencer5b950642006-11-11 23:08:07 +0000358languages. There are keywords for different opcodes
359('<tt><a href="#i_add">add</a></tt>',
360 '<tt><a href="#i_bitcast">bitcast</a></tt>',
361 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000362href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
Chris Lattnerd79749a2004-12-09 16:36:40 +0000363and others. These reserved words cannot conflict with variable names, because
Reid Spencerb23b65f2007-08-07 14:34:28 +0000364none of them start with a prefix character ('%' or '@').</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000365
366<p>Here is an example of LLVM code to multiply the integer variable
367'<tt>%X</tt>' by 8:</p>
368
Misha Brukman76307852003-11-08 01:05:38 +0000369<p>The easy way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000370
Bill Wendling3716c5d2007-05-29 09:04:49 +0000371<div class="doc_code">
Chris Lattnerd79749a2004-12-09 16:36:40 +0000372<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000373%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnerd79749a2004-12-09 16:36:40 +0000374</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000375</div>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000376
Misha Brukman76307852003-11-08 01:05:38 +0000377<p>After strength reduction:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000378
Bill Wendling3716c5d2007-05-29 09:04:49 +0000379<div class="doc_code">
Chris Lattnerd79749a2004-12-09 16:36:40 +0000380<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000381%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnerd79749a2004-12-09 16:36:40 +0000382</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000383</div>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000384
Misha Brukman76307852003-11-08 01:05:38 +0000385<p>And the hard way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000386
Bill Wendling3716c5d2007-05-29 09:04:49 +0000387<div class="doc_code">
Chris Lattnerd79749a2004-12-09 16:36:40 +0000388<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000389<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
390<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
391%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnerd79749a2004-12-09 16:36:40 +0000392</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000393</div>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000394
Chris Lattner48b383b02003-11-25 01:02:51 +0000395<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
396important lexical features of LLVM:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000397
Chris Lattner2f7c9632001-06-06 20:29:01 +0000398<ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000399
400 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
401 line.</li>
402
403 <li>Unnamed temporaries are created when the result of a computation is not
404 assigned to a named value.</li>
405
Misha Brukman76307852003-11-08 01:05:38 +0000406 <li>Unnamed temporaries are numbered sequentially</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000407
Misha Brukman76307852003-11-08 01:05:38 +0000408</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000409
John Criswell02fdc6f2005-05-12 16:52:32 +0000410<p>...and it also shows a convention that we follow in this document. When
Chris Lattnerd79749a2004-12-09 16:36:40 +0000411demonstrating instructions, we will follow an instruction with a comment that
412defines the type and name of value produced. Comments are shown in italic
413text.</p>
414
Misha Brukman76307852003-11-08 01:05:38 +0000415</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000416
417<!-- *********************************************************************** -->
418<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
419<!-- *********************************************************************** -->
420
421<!-- ======================================================================= -->
422<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
423</div>
424
425<div class="doc_text">
426
427<p>LLVM programs are composed of "Module"s, each of which is a
428translation unit of the input programs. Each module consists of
429functions, global variables, and symbol table entries. Modules may be
430combined together with the LLVM linker, which merges function (and
431global variable) definitions, resolves forward declarations, and merges
432symbol table entries. Here is an example of the "hello world" module:</p>
433
Bill Wendling3716c5d2007-05-29 09:04:49 +0000434<div class="doc_code">
Chris Lattner6af02f32004-12-09 16:11:40 +0000435<pre><i>; Declare the string constant as a global constant...</i>
Chris Lattner2b0bf4f2007-06-12 17:00:26 +0000436<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
437 href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
Chris Lattner6af02f32004-12-09 16:11:40 +0000438
439<i>; External declaration of the puts function</i>
Chris Lattner2b0bf4f2007-06-12 17:00:26 +0000440<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
Chris Lattner6af02f32004-12-09 16:11:40 +0000441
442<i>; Definition of main function</i>
Chris Lattner2b0bf4f2007-06-12 17:00:26 +0000443define i32 @main() { <i>; i32()* </i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000444 <i>; Convert [13x i8 ]* to i8 *...</i>
Chris Lattner6af02f32004-12-09 16:11:40 +0000445 %cast210 = <a
Chris Lattner2150cde2007-06-12 17:01:15 +0000446 href="#i_getelementptr">getelementptr</a> [13 x i8 ]* @.LC0, i64 0, i64 0 <i>; i8 *</i>
Chris Lattner6af02f32004-12-09 16:11:40 +0000447
448 <i>; Call puts function to write out the string to stdout...</i>
449 <a
Chris Lattner2b0bf4f2007-06-12 17:00:26 +0000450 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
Chris Lattner6af02f32004-12-09 16:11:40 +0000451 <a
Bill Wendling3716c5d2007-05-29 09:04:49 +0000452 href="#i_ret">ret</a> i32 0<br>}<br>
453</pre>
454</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000455
456<p>This example is made up of a <a href="#globalvars">global variable</a>
457named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
458function, and a <a href="#functionstructure">function definition</a>
459for "<tt>main</tt>".</p>
460
Chris Lattnerd79749a2004-12-09 16:36:40 +0000461<p>In general, a module is made up of a list of global values,
462where both functions and global variables are global values. Global values are
463represented by a pointer to a memory location (in this case, a pointer to an
464array of char, and a pointer to a function), and have one of the following <a
465href="#linkage">linkage types</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000466
Chris Lattnerd79749a2004-12-09 16:36:40 +0000467</div>
468
469<!-- ======================================================================= -->
470<div class="doc_subsection">
471 <a name="linkage">Linkage Types</a>
472</div>
473
474<div class="doc_text">
475
476<p>
477All Global Variables and Functions have one of the following types of linkage:
478</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000479
480<dl>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000481
Dale Johannesen4188aad2008-05-23 23:13:41 +0000482 <dt><tt><b><a name="linkage_internal">internal</a></b></tt>: </dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000483
484 <dd>Global values with internal linkage are only directly accessible by
485 objects in the current module. In particular, linking code into a module with
486 an internal global value may cause the internal to be renamed as necessary to
487 avoid collisions. Because the symbol is internal to the module, all
488 references can be updated. This corresponds to the notion of the
Chris Lattnere20b4702007-01-14 06:51:48 +0000489 '<tt>static</tt>' keyword in C.
Chris Lattner6af02f32004-12-09 16:11:40 +0000490 </dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000491
Chris Lattner6af02f32004-12-09 16:11:40 +0000492 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000493
Chris Lattnere20b4702007-01-14 06:51:48 +0000494 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
495 the same name when linkage occurs. This is typically used to implement
496 inline functions, templates, or other code which must be generated in each
497 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
498 allowed to be discarded.
Chris Lattner6af02f32004-12-09 16:11:40 +0000499 </dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000500
Dale Johannesen4188aad2008-05-23 23:13:41 +0000501 <dt><tt><b><a name="linkage_common">common</a></b></tt>: </dt>
502
503 <dd>"<tt>common</tt>" linkage is exactly the same as <tt>linkonce</tt>
504 linkage, except that unreferenced <tt>common</tt> globals may not be
505 discarded. This is used for globals that may be emitted in multiple
506 translation units, but that are not guaranteed to be emitted into every
507 translation unit that uses them. One example of this is tentative
508 definitions in C, such as "<tt>int X;</tt>" at global scope.
509 </dd>
510
Chris Lattner6af02f32004-12-09 16:11:40 +0000511 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000512
Dale Johannesen4188aad2008-05-23 23:13:41 +0000513 <dd>"<tt>weak</tt>" linkage is the same as <tt>common</tt> linkage, except
514 that some targets may choose to emit different assembly sequences for them
515 for target-dependent reasons. This is used for globals that are declared
516 "weak" in C source code.
Chris Lattner6af02f32004-12-09 16:11:40 +0000517 </dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000518
Chris Lattner6af02f32004-12-09 16:11:40 +0000519 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000520
521 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
522 pointer to array type. When two global variables with appending linkage are
523 linked together, the two global arrays are appended together. This is the
524 LLVM, typesafe, equivalent of having the system linker append together
525 "sections" with identical names when .o files are linked.
Chris Lattner6af02f32004-12-09 16:11:40 +0000526 </dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000527
Anton Korobeynikova0554d92007-01-12 19:20:47 +0000528 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
Chris Lattner67c37d12008-08-05 18:29:16 +0000529 <dd>The semantics of this linkage follow the ELF object file model: the
530 symbol is weak until linked, if not linked, the symbol becomes null instead
531 of being an undefined reference.
Anton Korobeynikova0554d92007-01-12 19:20:47 +0000532 </dd>
Anton Korobeynikova0554d92007-01-12 19:20:47 +0000533
Chris Lattner6af02f32004-12-09 16:11:40 +0000534 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000535
536 <dd>If none of the above identifiers are used, the global is externally
537 visible, meaning that it participates in linkage and can be used to resolve
538 external symbol references.
Chris Lattner6af02f32004-12-09 16:11:40 +0000539 </dd>
Reid Spencer7972c472007-04-11 23:49:50 +0000540</dl>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000541
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000542 <p>
543 The next two types of linkage are targeted for Microsoft Windows platform
544 only. They are designed to support importing (exporting) symbols from (to)
Chris Lattner67c37d12008-08-05 18:29:16 +0000545 DLLs (Dynamic Link Libraries).
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000546 </p>
547
Anton Korobeynikova0554d92007-01-12 19:20:47 +0000548 <dl>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000549 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
550
551 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
552 or variable via a global pointer to a pointer that is set up by the DLL
553 exporting the symbol. On Microsoft Windows targets, the pointer name is
554 formed by combining <code>_imp__</code> and the function or variable name.
555 </dd>
556
557 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
558
559 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
560 pointer to a pointer in a DLL, so that it can be referenced with the
561 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
562 name is formed by combining <code>_imp__</code> and the function or variable
563 name.
564 </dd>
565
Chris Lattner6af02f32004-12-09 16:11:40 +0000566</dl>
567
Dan Gohman8ef44982008-11-24 17:18:39 +0000568<p>For example, since the "<tt>.LC0</tt>"
Chris Lattner6af02f32004-12-09 16:11:40 +0000569variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
570variable and was linked with this one, one of the two would be renamed,
571preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
572external (i.e., lacking any linkage declarations), they are accessible
Reid Spencer92c671e2007-01-05 00:59:10 +0000573outside of the current module.</p>
574<p>It is illegal for a function <i>declaration</i>
575to have any linkage type other than "externally visible", <tt>dllimport</tt>,
Anton Korobeynikova0554d92007-01-12 19:20:47 +0000576or <tt>extern_weak</tt>.</p>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000577<p>Aliases can have only <tt>external</tt>, <tt>internal</tt> and <tt>weak</tt>
Dan Gohmanef9462f2008-10-14 16:51:45 +0000578linkages.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000579</div>
580
581<!-- ======================================================================= -->
582<div class="doc_subsection">
Chris Lattner0132aff2005-05-06 22:57:40 +0000583 <a name="callingconv">Calling Conventions</a>
584</div>
585
586<div class="doc_text">
587
588<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
589and <a href="#i_invoke">invokes</a> can all have an optional calling convention
590specified for the call. The calling convention of any pair of dynamic
591caller/callee must match, or the behavior of the program is undefined. The
592following calling conventions are supported by LLVM, and more may be added in
593the future:</p>
594
595<dl>
596 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
597
598 <dd>This calling convention (the default if no other calling convention is
599 specified) matches the target C calling conventions. This calling convention
John Criswell02fdc6f2005-05-12 16:52:32 +0000600 supports varargs function calls and tolerates some mismatch in the declared
Reid Spencer72ba4992006-12-31 21:30:18 +0000601 prototype and implemented declaration of the function (as does normal C).
Chris Lattner0132aff2005-05-06 22:57:40 +0000602 </dd>
603
604 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
605
606 <dd>This calling convention attempts to make calls as fast as possible
607 (e.g. by passing things in registers). This calling convention allows the
608 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner67c37d12008-08-05 18:29:16 +0000609 without having to conform to an externally specified ABI (Application Binary
610 Interface). Implementations of this convention should allow arbitrary
Arnold Schwaighofer2c6b8882008-05-14 09:17:12 +0000611 <a href="CodeGenerator.html#tailcallopt">tail call optimization</a> to be
612 supported. This calling convention does not support varargs and requires the
613 prototype of all callees to exactly match the prototype of the function
614 definition.
Chris Lattner0132aff2005-05-06 22:57:40 +0000615 </dd>
616
617 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
618
619 <dd>This calling convention attempts to make code in the caller as efficient
620 as possible under the assumption that the call is not commonly executed. As
621 such, these calls often preserve all registers so that the call does not break
622 any live ranges in the caller side. This calling convention does not support
623 varargs and requires the prototype of all callees to exactly match the
624 prototype of the function definition.
625 </dd>
626
Chris Lattner573f64e2005-05-07 01:46:40 +0000627 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000628
629 <dd>Any calling convention may be specified by number, allowing
630 target-specific calling conventions to be used. Target specific calling
631 conventions start at 64.
632 </dd>
Chris Lattner573f64e2005-05-07 01:46:40 +0000633</dl>
Chris Lattner0132aff2005-05-06 22:57:40 +0000634
635<p>More calling conventions can be added/defined on an as-needed basis, to
636support pascal conventions or any other well-known target-independent
637convention.</p>
638
639</div>
640
641<!-- ======================================================================= -->
642<div class="doc_subsection">
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000643 <a name="visibility">Visibility Styles</a>
644</div>
645
646<div class="doc_text">
647
648<p>
649All Global Variables and Functions have one of the following visibility styles:
650</p>
651
652<dl>
653 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
654
Chris Lattner67c37d12008-08-05 18:29:16 +0000655 <dd>On targets that use the ELF object file format, default visibility means
656 that the declaration is visible to other
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000657 modules and, in shared libraries, means that the declared entity may be
658 overridden. On Darwin, default visibility means that the declaration is
659 visible to other modules. Default visibility corresponds to "external
660 linkage" in the language.
661 </dd>
662
663 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
664
665 <dd>Two declarations of an object with hidden visibility refer to the same
666 object if they are in the same shared object. Usually, hidden visibility
667 indicates that the symbol will not be placed into the dynamic symbol table,
668 so no other module (executable or shared library) can reference it
669 directly.
670 </dd>
671
Anton Korobeynikov39f3cff2007-04-29 18:35:00 +0000672 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
673
674 <dd>On ELF, protected visibility indicates that the symbol will be placed in
675 the dynamic symbol table, but that references within the defining module will
676 bind to the local symbol. That is, the symbol cannot be overridden by another
677 module.
678 </dd>
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000679</dl>
680
681</div>
682
683<!-- ======================================================================= -->
684<div class="doc_subsection">
Chris Lattner6af02f32004-12-09 16:11:40 +0000685 <a name="globalvars">Global Variables</a>
686</div>
687
688<div class="doc_text">
689
Chris Lattner5d5aede2005-02-12 19:30:21 +0000690<p>Global variables define regions of memory allocated at compilation time
Chris Lattner662c8722005-11-12 00:45:07 +0000691instead of run-time. Global variables may optionally be initialized, may have
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000692an explicit section to be placed in, and may have an optional explicit alignment
693specified. A variable may be defined as "thread_local", which means that it
694will not be shared by threads (each thread will have a separated copy of the
695variable). A variable may be defined as a global "constant," which indicates
696that the contents of the variable will <b>never</b> be modified (enabling better
Chris Lattner5d5aede2005-02-12 19:30:21 +0000697optimization, allowing the global data to be placed in the read-only section of
698an executable, etc). Note that variables that need runtime initialization
John Criswell4c0cf7f2005-10-24 16:17:18 +0000699cannot be marked "constant" as there is a store to the variable.</p>
Chris Lattner5d5aede2005-02-12 19:30:21 +0000700
701<p>
702LLVM explicitly allows <em>declarations</em> of global variables to be marked
703constant, even if the final definition of the global is not. This capability
704can be used to enable slightly better optimization of the program, but requires
705the language definition to guarantee that optimizations based on the
706'constantness' are valid for the translation units that do not include the
707definition.
708</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000709
710<p>As SSA values, global variables define pointer values that are in
711scope (i.e. they dominate) all basic blocks in the program. Global
712variables always define a pointer to their "content" type because they
713describe a region of memory, and all memory objects in LLVM are
714accessed through pointers.</p>
715
Christopher Lamb308121c2007-12-11 09:31:00 +0000716<p>A global variable may be declared to reside in a target-specifc numbered
717address space. For targets that support them, address spaces may affect how
718optimizations are performed and/or what target instructions are used to access
Christopher Lamb25f50762007-12-12 08:44:39 +0000719the variable. The default address space is zero. The address space qualifier
720must precede any other attributes.</p>
Christopher Lamb308121c2007-12-11 09:31:00 +0000721
Chris Lattner662c8722005-11-12 00:45:07 +0000722<p>LLVM allows an explicit section to be specified for globals. If the target
723supports it, it will emit globals to the section specified.</p>
724
Chris Lattner54611b42005-11-06 08:02:57 +0000725<p>An explicit alignment may be specified for a global. If not present, or if
726the alignment is set to zero, the alignment of the global is set by the target
727to whatever it feels convenient. If an explicit alignment is specified, the
728global is forced to have at least that much alignment. All alignments must be
729a power of 2.</p>
730
Christopher Lamb308121c2007-12-11 09:31:00 +0000731<p>For example, the following defines a global in a numbered address space with
732an initializer, section, and alignment:</p>
Chris Lattner5760c502007-01-14 00:27:09 +0000733
Bill Wendling3716c5d2007-05-29 09:04:49 +0000734<div class="doc_code">
Chris Lattner5760c502007-01-14 00:27:09 +0000735<pre>
Christopher Lamb308121c2007-12-11 09:31:00 +0000736@G = constant float 1.0 addrspace(5), section "foo", align 4
Chris Lattner5760c502007-01-14 00:27:09 +0000737</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000738</div>
Chris Lattner5760c502007-01-14 00:27:09 +0000739
Chris Lattner6af02f32004-12-09 16:11:40 +0000740</div>
741
742
743<!-- ======================================================================= -->
744<div class="doc_subsection">
745 <a name="functionstructure">Functions</a>
746</div>
747
748<div class="doc_text">
749
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000750<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
751an optional <a href="#linkage">linkage type</a>, an optional
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000752<a href="#visibility">visibility style</a>, an optional
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000753<a href="#callingconv">calling convention</a>, a return type, an optional
754<a href="#paramattrs">parameter attribute</a> for the return type, a function
755name, a (possibly empty) argument list (each with optional
Devang Patel7e9b05e2008-10-06 18:50:38 +0000756<a href="#paramattrs">parameter attributes</a>), optional
757<a href="#fnattrs">function attributes</a>, an optional section,
758an optional alignment, an optional <a href="#gc">garbage collector name</a>,
Chris Lattnercbc4d2a2008-10-04 18:10:21 +0000759an opening curly brace, a list of basic blocks, and a closing curly brace.
Anton Korobeynikovc7f9f3d2007-01-23 12:35:46 +0000760
761LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
762optional <a href="#linkage">linkage type</a>, an optional
763<a href="#visibility">visibility style</a>, an optional
764<a href="#callingconv">calling convention</a>, a return type, an optional
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000765<a href="#paramattrs">parameter attribute</a> for the return type, a function
Gordon Henriksen71183b62007-12-10 03:18:06 +0000766name, a possibly empty list of arguments, an optional alignment, and an optional
Gordon Henriksendc5cafb2007-12-10 03:30:21 +0000767<a href="#gc">garbage collector name</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000768
Chris Lattner67c37d12008-08-05 18:29:16 +0000769<p>A function definition contains a list of basic blocks, forming the CFG
770(Control Flow Graph) for
Chris Lattner6af02f32004-12-09 16:11:40 +0000771the function. Each basic block may optionally start with a label (giving the
772basic block a symbol table entry), contains a list of instructions, and ends
773with a <a href="#terminators">terminator</a> instruction (such as a branch or
774function return).</p>
775
Chris Lattnera59fb102007-06-08 16:52:14 +0000776<p>The first basic block in a function is special in two ways: it is immediately
Chris Lattner6af02f32004-12-09 16:11:40 +0000777executed on entrance to the function, and it is not allowed to have predecessor
778basic blocks (i.e. there can not be any branches to the entry block of a
779function). Because the block can have no predecessors, it also cannot have any
780<a href="#i_phi">PHI nodes</a>.</p>
781
Chris Lattner662c8722005-11-12 00:45:07 +0000782<p>LLVM allows an explicit section to be specified for functions. If the target
783supports it, it will emit functions to the section specified.</p>
784
Chris Lattner54611b42005-11-06 08:02:57 +0000785<p>An explicit alignment may be specified for a function. If not present, or if
786the alignment is set to zero, the alignment of the function is set by the target
787to whatever it feels convenient. If an explicit alignment is specified, the
788function is forced to have at least that much alignment. All alignments must be
789a power of 2.</p>
790
Devang Patel02256232008-10-07 17:48:33 +0000791 <h5>Syntax:</h5>
792
793<div class="doc_code">
Chris Lattner0ae02092008-10-13 16:55:18 +0000794<tt>
795define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
796 [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
797 &lt;ResultType&gt; @&lt;FunctionName&gt; ([argument list])
798 [<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
799 [<a href="#gc">gc</a>] { ... }
800</tt>
Devang Patel02256232008-10-07 17:48:33 +0000801</div>
802
Chris Lattner6af02f32004-12-09 16:11:40 +0000803</div>
804
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000805
806<!-- ======================================================================= -->
807<div class="doc_subsection">
808 <a name="aliasstructure">Aliases</a>
809</div>
810<div class="doc_text">
811 <p>Aliases act as "second name" for the aliasee value (which can be either
Anton Korobeynikov25b2e822008-03-22 08:36:14 +0000812 function, global variable, another alias or bitcast of global value). Aliases
813 may have an optional <a href="#linkage">linkage type</a>, and an
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000814 optional <a href="#visibility">visibility style</a>.</p>
815
816 <h5>Syntax:</h5>
817
Bill Wendling3716c5d2007-05-29 09:04:49 +0000818<div class="doc_code">
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000819<pre>
Duncan Sands7e99a942008-09-12 20:48:21 +0000820@&lt;Name&gt; = alias [Linkage] [Visibility] &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendling2d8b9a82007-05-29 09:42:13 +0000821</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +0000822</div>
Anton Korobeynikova97b6942007-04-25 14:27:10 +0000823
824</div>
825
826
827
Chris Lattner91c15c42006-01-23 23:23:47 +0000828<!-- ======================================================================= -->
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000829<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
830<div class="doc_text">
831 <p>The return type and each parameter of a function type may have a set of
832 <i>parameter attributes</i> associated with them. Parameter attributes are
833 used to communicate additional information about the result or parameters of
Duncan Sandsad0ea2d2007-11-27 13:23:08 +0000834 a function. Parameter attributes are considered to be part of the function,
835 not of the function type, so functions with different parameter attributes
836 can have the same function type.</p>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000837
Reid Spencercf7ebf52007-01-15 18:27:39 +0000838 <p>Parameter attributes are simple keywords that follow the type specified. If
839 multiple parameter attributes are needed, they are space separated. For
Bill Wendling3716c5d2007-05-29 09:04:49 +0000840 example:</p>
841
842<div class="doc_code">
843<pre>
Devang Patel9eb525d2008-09-26 23:51:19 +0000844declare i32 @printf(i8* noalias , ...)
Chris Lattnerd2597d72008-10-04 18:33:34 +0000845declare i32 @atoi(i8 zeroext)
846declare signext i8 @returns_signed_char()
Bill Wendling3716c5d2007-05-29 09:04:49 +0000847</pre>
848</div>
849
Duncan Sandsad0ea2d2007-11-27 13:23:08 +0000850 <p>Note that any attributes for the function result (<tt>nounwind</tt>,
851 <tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000852
Anton Korobeynikova0554d92007-01-12 19:20:47 +0000853 <p>Currently, only the following parameter attributes are defined:</p>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000854 <dl>
Reid Spencer314e1cb2007-07-19 23:13:04 +0000855 <dt><tt>zeroext</tt></dt>
Chris Lattnerd2597d72008-10-04 18:33:34 +0000856 <dd>This indicates to the code generator that the parameter or return value
857 should be zero-extended to a 32-bit value by the caller (for a parameter)
858 or the callee (for a return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000859
Reid Spencer314e1cb2007-07-19 23:13:04 +0000860 <dt><tt>signext</tt></dt>
Chris Lattnerd2597d72008-10-04 18:33:34 +0000861 <dd>This indicates to the code generator that the parameter or return value
862 should be sign-extended to a 32-bit value by the caller (for a parameter)
863 or the callee (for a return value).</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000864
Anton Korobeynikove8166852007-01-28 14:30:45 +0000865 <dt><tt>inreg</tt></dt>
Dale Johannesenc50ada22008-09-25 20:47:45 +0000866 <dd>This indicates that this parameter or return value should be treated
867 in a special target-dependent fashion during while emitting code for a
868 function call or return (usually, by putting it in a register as opposed
Chris Lattnerd2597d72008-10-04 18:33:34 +0000869 to memory, though some targets use it to distinguish between two different
870 kinds of registers). Use of this attribute is target-specific.</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000871
Duncan Sands2a1d8ba2008-10-06 08:14:18 +0000872 <dt><tt><a name="byval">byval</a></tt></dt>
Chris Lattner352ab9b2008-01-15 04:34:22 +0000873 <dd>This indicates that the pointer parameter should really be passed by
874 value to the function. The attribute implies that a hidden copy of the
875 pointee is made between the caller and the callee, so the callee is unable
Chris Lattner1ca5c642008-08-05 18:21:08 +0000876 to modify the value in the callee. This attribute is only valid on LLVM
Chris Lattner352ab9b2008-01-15 04:34:22 +0000877 pointer arguments. It is generally used to pass structs and arrays by
Duncan Sands2a1d8ba2008-10-06 08:14:18 +0000878 value, but is also valid on pointers to scalars. The copy is considered to
879 belong to the caller not the callee (for example,
880 <tt><a href="#readonly">readonly</a></tt> functions should not write to
Devang Patel7e9b05e2008-10-06 18:50:38 +0000881 <tt>byval</tt> parameters). This is not a valid attribute for return
882 values. </dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000883
Anton Korobeynikove8166852007-01-28 14:30:45 +0000884 <dt><tt>sret</tt></dt>
Duncan Sandsfa4b6732008-02-18 04:19:38 +0000885 <dd>This indicates that the pointer parameter specifies the address of a
886 structure that is the return value of the function in the source program.
Chris Lattnerd2597d72008-10-04 18:33:34 +0000887 This pointer must be guaranteed by the caller to be valid: loads and stores
888 to the structure may be assumed by the callee to not to trap. This may only
Devang Patel7e9b05e2008-10-06 18:50:38 +0000889 be applied to the first parameter. This is not a valid attribute for
890 return values. </dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000891
Zhou Sheng2444a9a2007-06-05 05:28:26 +0000892 <dt><tt>noalias</tt></dt>
Nick Lewyckyf5ffcbc2008-11-24 03:41:24 +0000893 <dd>This indicates that the pointer does not alias any global or any other
894 parameter. The caller is responsible for ensuring that this is the
Nick Lewyckyd59572c2008-11-24 05:00:44 +0000895 case. On a function return value, <tt>noalias</tt> additionally indicates
896 that the pointer does not alias any other pointers visible to the
Torok Edwin9bc1a652008-11-24 08:02:24 +0000897 caller. Note that this applies only to pointers that can be used to actually
898 load/store a value: NULL, unique pointers from malloc(0), and freed pointers
899 are considered to not alias anything.</dd>
Chris Lattner5cee13f2008-01-11 06:20:47 +0000900
Duncan Sands27e91592007-07-27 19:57:41 +0000901 <dt><tt>nest</tt></dt>
Duncan Sands825bde42008-07-08 09:27:25 +0000902 <dd>This indicates that the pointer parameter can be excised using the
Devang Patel7e9b05e2008-10-06 18:50:38 +0000903 <a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
904 attribute for return values.</dd>
Anton Korobeynikova0554d92007-01-12 19:20:47 +0000905 </dl>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000906
Reid Spencerb5ebf3d2006-12-31 07:07:53 +0000907</div>
908
909<!-- ======================================================================= -->
Chris Lattner91c15c42006-01-23 23:23:47 +0000910<div class="doc_subsection">
Gordon Henriksen71183b62007-12-10 03:18:06 +0000911 <a name="gc">Garbage Collector Names</a>
912</div>
913
914<div class="doc_text">
915<p>Each function may specify a garbage collector name, which is simply a
916string.</p>
917
918<div class="doc_code"><pre
919>define void @f() gc "name" { ...</pre></div>
920
921<p>The compiler declares the supported values of <i>name</i>. Specifying a
922collector which will cause the compiler to alter its output in order to support
923the named garbage collection algorithm.</p>
924</div>
925
926<!-- ======================================================================= -->
927<div class="doc_subsection">
Devang Patel9eb525d2008-09-26 23:51:19 +0000928 <a name="fnattrs">Function Attributes</a>
Devang Patelcaacdba2008-09-04 23:05:13 +0000929</div>
930
931<div class="doc_text">
Devang Patel9eb525d2008-09-26 23:51:19 +0000932
933<p>Function attributes are set to communicate additional information about
934 a function. Function attributes are considered to be part of the function,
935 not of the function type, so functions with different parameter attributes
936 can have the same function type.</p>
937
938 <p>Function attributes are simple keywords that follow the type specified. If
939 multiple attributes are needed, they are space separated. For
940 example:</p>
Devang Patelcaacdba2008-09-04 23:05:13 +0000941
942<div class="doc_code">
Bill Wendlingb175fa42008-09-07 10:26:33 +0000943<pre>
Devang Patel9eb525d2008-09-26 23:51:19 +0000944define void @f() noinline { ... }
945define void @f() alwaysinline { ... }
946define void @f() alwaysinline optsize { ... }
947define void @f() optsize
Bill Wendlingb175fa42008-09-07 10:26:33 +0000948</pre>
Devang Patelcaacdba2008-09-04 23:05:13 +0000949</div>
950
Bill Wendlingb175fa42008-09-07 10:26:33 +0000951<dl>
Devang Patel9eb525d2008-09-26 23:51:19 +0000952<dt><tt>alwaysinline</tt></dt>
Chris Lattnerfbf60a42008-10-04 18:23:17 +0000953<dd>This attribute indicates that the inliner should attempt to inline this
954function into callers whenever possible, ignoring any active inlining size
955threshold for this caller.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +0000956
Devang Patel9eb525d2008-09-26 23:51:19 +0000957<dt><tt>noinline</tt></dt>
Chris Lattnerfbf60a42008-10-04 18:23:17 +0000958<dd>This attribute indicates that the inliner should never inline this function
Chris Lattner0625c282008-10-05 17:14:59 +0000959in any situation. This attribute may not be used together with the
Chris Lattnerfbf60a42008-10-04 18:23:17 +0000960<tt>alwaysinline</tt> attribute.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +0000961
Devang Patel9eb525d2008-09-26 23:51:19 +0000962<dt><tt>optsize</tt></dt>
Devang Patele9743902008-09-29 18:34:44 +0000963<dd>This attribute suggests that optimization passes and code generator passes
Chris Lattnerfbf60a42008-10-04 18:23:17 +0000964make choices that keep the code size of this function low, and otherwise do
965optimizations specifically to reduce code size.</dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +0000966
Devang Patel9eb525d2008-09-26 23:51:19 +0000967<dt><tt>noreturn</tt></dt>
Chris Lattnerfbf60a42008-10-04 18:23:17 +0000968<dd>This function attribute indicates that the function never returns normally.
969This produces undefined behavior at runtime if the function ever does
970dynamically return.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +0000971
972<dt><tt>nounwind</tt></dt>
Chris Lattnerfbf60a42008-10-04 18:23:17 +0000973<dd>This function attribute indicates that the function never returns with an
974unwind or exceptional control flow. If the function does unwind, its runtime
975behavior is undefined.</dd>
976
977<dt><tt>readnone</tt></dt>
Duncan Sands2a1d8ba2008-10-06 08:14:18 +0000978<dd>This attribute indicates that the function computes its result (or the
979exception it throws) based strictly on its arguments, without dereferencing any
980pointer arguments or otherwise accessing any mutable state (e.g. memory, control
981registers, etc) visible to caller functions. It does not write through any
982pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments) and
983never changes any state visible to callers.</dd>
Devang Patel9eb525d2008-09-26 23:51:19 +0000984
Duncan Sands2a1d8ba2008-10-06 08:14:18 +0000985<dt><tt><a name="readonly">readonly</a></tt></dt>
986<dd>This attribute indicates that the function does not write through any
987pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments)
988or otherwise modify any state (e.g. memory, control registers, etc) visible to
989caller functions. It may dereference pointer arguments and read state that may
990be set in the caller. A readonly function always returns the same value (or
991throws the same exception) when called with the same set of arguments and global
992state.</dd>
Bill Wendlinga8130172008-11-13 01:02:51 +0000993
994<dt><tt><a name="ssp">ssp</a></tt></dt>
Bill Wendling6e41add2008-11-26 19:19:05 +0000995<dd>This attribute indicates that the function should emit a stack smashing
Bill Wendlinga8130172008-11-13 01:02:51 +0000996protector. It is in the form of a "canary"&mdash;a random value placed on the
997stack before the local variables that's checked upon return from the function to
998see if it has been overwritten. A heuristic is used to determine if a function
Bill Wendling6e41add2008-11-26 19:19:05 +0000999needs stack protectors or not.
Bill Wendlinga8130172008-11-13 01:02:51 +00001000
Bill Wendling0f5541e2008-11-26 19:07:40 +00001001<p>If a function that has an <tt>ssp</tt> attribute is inlined into a function
1002that doesn't have an <tt>ssp</tt> attribute, then the resulting function will
1003have an <tt>ssp</tt> attribute.</p></dd>
1004
1005<dt><tt>sspreq</tt></dt>
Bill Wendling6e41add2008-11-26 19:19:05 +00001006<dd>This attribute indicates that the function should <em>always</em> emit a
Bill Wendlinga8130172008-11-13 01:02:51 +00001007stack smashing protector. This overrides the <tt><a href="#ssp">ssp</a></tt>
Bill Wendling6e41add2008-11-26 19:19:05 +00001008function attribute.
Bill Wendling0f5541e2008-11-26 19:07:40 +00001009
1010<p>If a function that has an <tt>sspreq</tt> attribute is inlined into a
1011function that doesn't have an <tt>sspreq</tt> attribute or which has
1012an <tt>ssp</tt> attribute, then the resulting function will have
1013an <tt>sspreq</tt> attribute.</p></dd>
Bill Wendlingb175fa42008-09-07 10:26:33 +00001014</dl>
1015
Devang Patelcaacdba2008-09-04 23:05:13 +00001016</div>
1017
1018<!-- ======================================================================= -->
1019<div class="doc_subsection">
Chris Lattner93564892006-04-08 04:40:53 +00001020 <a name="moduleasm">Module-Level Inline Assembly</a>
Chris Lattner91c15c42006-01-23 23:23:47 +00001021</div>
1022
1023<div class="doc_text">
1024<p>
1025Modules may contain "module-level inline asm" blocks, which corresponds to the
1026GCC "file scope inline asm" blocks. These blocks are internally concatenated by
1027LLVM and treated as a single unit, but may be separated in the .ll file if
1028desired. The syntax is very simple:
1029</p>
1030
Bill Wendling3716c5d2007-05-29 09:04:49 +00001031<div class="doc_code">
1032<pre>
1033module asm "inline asm code goes here"
1034module asm "more can go here"
1035</pre>
1036</div>
Chris Lattner91c15c42006-01-23 23:23:47 +00001037
1038<p>The strings can contain any character by escaping non-printable characters.
1039 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
1040 for the number.
1041</p>
1042
1043<p>
1044 The inline asm code is simply printed to the machine code .s file when
1045 assembly code is generated.
1046</p>
1047</div>
Chris Lattner6af02f32004-12-09 16:11:40 +00001048
Reid Spencer50c723a2007-02-19 23:54:10 +00001049<!-- ======================================================================= -->
1050<div class="doc_subsection">
1051 <a name="datalayout">Data Layout</a>
1052</div>
1053
1054<div class="doc_text">
1055<p>A module may specify a target specific data layout string that specifies how
Reid Spencer7972c472007-04-11 23:49:50 +00001056data is to be laid out in memory. The syntax for the data layout is simply:</p>
1057<pre> target datalayout = "<i>layout specification</i>"</pre>
1058<p>The <i>layout specification</i> consists of a list of specifications
1059separated by the minus sign character ('-'). Each specification starts with a
1060letter and may include other information after the letter to define some
1061aspect of the data layout. The specifications accepted are as follows: </p>
Reid Spencer50c723a2007-02-19 23:54:10 +00001062<dl>
1063 <dt><tt>E</tt></dt>
1064 <dd>Specifies that the target lays out data in big-endian form. That is, the
1065 bits with the most significance have the lowest address location.</dd>
1066 <dt><tt>e</tt></dt>
Chris Lattner67c37d12008-08-05 18:29:16 +00001067 <dd>Specifies that the target lays out data in little-endian form. That is,
Reid Spencer50c723a2007-02-19 23:54:10 +00001068 the bits with the least significance have the lowest address location.</dd>
1069 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1070 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
1071 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
1072 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
1073 too.</dd>
1074 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1075 <dd>This specifies the alignment for an integer type of a given bit
1076 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
1077 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1078 <dd>This specifies the alignment for a vector type of a given bit
1079 <i>size</i>.</dd>
1080 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1081 <dd>This specifies the alignment for a floating point type of a given bit
1082 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
1083 (double).</dd>
1084 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
1085 <dd>This specifies the alignment for an aggregate type of a given bit
1086 <i>size</i>.</dd>
1087</dl>
1088<p>When constructing the data layout for a given target, LLVM starts with a
1089default set of specifications which are then (possibly) overriden by the
1090specifications in the <tt>datalayout</tt> keyword. The default specifications
1091are given in this list:</p>
1092<ul>
1093 <li><tt>E</tt> - big endian</li>
1094 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
1095 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
1096 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
1097 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
1098 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner67c37d12008-08-05 18:29:16 +00001099 <li><tt>i64:32:64</tt> - i64 has ABI alignment of 32-bits but preferred
Reid Spencer50c723a2007-02-19 23:54:10 +00001100 alignment of 64-bits</li>
1101 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
1102 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
1103 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
1104 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
1105 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
1106</ul>
Chris Lattner1ca5c642008-08-05 18:21:08 +00001107<p>When LLVM is determining the alignment for a given type, it uses the
Dan Gohmanef9462f2008-10-14 16:51:45 +00001108following rules:</p>
Reid Spencer50c723a2007-02-19 23:54:10 +00001109<ol>
1110 <li>If the type sought is an exact match for one of the specifications, that
1111 specification is used.</li>
1112 <li>If no match is found, and the type sought is an integer type, then the
1113 smallest integer type that is larger than the bitwidth of the sought type is
1114 used. If none of the specifications are larger than the bitwidth then the the
1115 largest integer type is used. For example, given the default specifications
1116 above, the i7 type will use the alignment of i8 (next largest) while both
1117 i65 and i256 will use the alignment of i64 (largest specified).</li>
1118 <li>If no match is found, and the type sought is a vector type, then the
1119 largest vector type that is smaller than the sought vector type will be used
Dan Gohmanef9462f2008-10-14 16:51:45 +00001120 as a fall back. This happens because &lt;128 x double&gt; can be implemented
1121 in terms of 64 &lt;2 x double&gt;, for example.</li>
Reid Spencer50c723a2007-02-19 23:54:10 +00001122</ol>
1123</div>
Chris Lattner6af02f32004-12-09 16:11:40 +00001124
Chris Lattner2f7c9632001-06-06 20:29:01 +00001125<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001126<div class="doc_section"> <a name="typesystem">Type System</a> </div>
1127<!-- *********************************************************************** -->
Chris Lattner6af02f32004-12-09 16:11:40 +00001128
Misha Brukman76307852003-11-08 01:05:38 +00001129<div class="doc_text">
Chris Lattner6af02f32004-12-09 16:11:40 +00001130
Misha Brukman76307852003-11-08 01:05:38 +00001131<p>The LLVM type system is one of the most important features of the
Chris Lattner48b383b02003-11-25 01:02:51 +00001132intermediate representation. Being typed enables a number of
Chris Lattner67c37d12008-08-05 18:29:16 +00001133optimizations to be performed on the intermediate representation directly,
1134without having to do
Chris Lattner48b383b02003-11-25 01:02:51 +00001135extra analyses on the side before the transformation. A strong type
1136system makes it easier to read the generated code and enables novel
1137analyses and transformations that are not feasible to perform on normal
1138three address code representations.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +00001139
1140</div>
1141
Chris Lattner2f7c9632001-06-06 20:29:01 +00001142<!-- ======================================================================= -->
Chris Lattner7824d182008-01-04 04:32:38 +00001143<div class="doc_subsection"> <a name="t_classifications">Type
Chris Lattner48b383b02003-11-25 01:02:51 +00001144Classifications</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +00001145<div class="doc_text">
Chris Lattner7824d182008-01-04 04:32:38 +00001146<p>The types fall into a few useful
Chris Lattner48b383b02003-11-25 01:02:51 +00001147classifications:</p>
Misha Brukmanc501f552004-03-01 17:47:27 +00001148
1149<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +00001150 <tbody>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001151 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner48b383b02003-11-25 01:02:51 +00001152 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +00001153 <td><a href="#t_integer">integer</a></td>
Reid Spencer138249b2007-05-16 18:44:01 +00001154 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +00001155 </tr>
1156 <tr>
Chris Lattner7824d182008-01-04 04:32:38 +00001157 <td><a href="#t_floating">floating point</a></td>
1158 <td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +00001159 </tr>
1160 <tr>
1161 <td><a name="t_firstclass">first class</a></td>
Chris Lattner7824d182008-01-04 04:32:38 +00001162 <td><a href="#t_integer">integer</a>,
1163 <a href="#t_floating">floating point</a>,
1164 <a href="#t_pointer">pointer</a>,
Dan Gohman08783a882008-06-18 18:42:13 +00001165 <a href="#t_vector">vector</a>,
Dan Gohmanb9d66602008-05-12 23:51:09 +00001166 <a href="#t_struct">structure</a>,
1167 <a href="#t_array">array</a>,
Dan Gohmanda52d212008-05-23 22:50:26 +00001168 <a href="#t_label">label</a>.
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00001169 </td>
Chris Lattner48b383b02003-11-25 01:02:51 +00001170 </tr>
Chris Lattner7824d182008-01-04 04:32:38 +00001171 <tr>
1172 <td><a href="#t_primitive">primitive</a></td>
1173 <td><a href="#t_label">label</a>,
1174 <a href="#t_void">void</a>,
Chris Lattner7824d182008-01-04 04:32:38 +00001175 <a href="#t_floating">floating point</a>.</td>
1176 </tr>
1177 <tr>
1178 <td><a href="#t_derived">derived</a></td>
1179 <td><a href="#t_integer">integer</a>,
1180 <a href="#t_array">array</a>,
1181 <a href="#t_function">function</a>,
1182 <a href="#t_pointer">pointer</a>,
1183 <a href="#t_struct">structure</a>,
1184 <a href="#t_pstruct">packed structure</a>,
1185 <a href="#t_vector">vector</a>,
1186 <a href="#t_opaque">opaque</a>.
Dan Gohman93bf60d2008-10-14 16:32:04 +00001187 </td>
Chris Lattner7824d182008-01-04 04:32:38 +00001188 </tr>
Chris Lattner48b383b02003-11-25 01:02:51 +00001189 </tbody>
Misha Brukman76307852003-11-08 01:05:38 +00001190</table>
Misha Brukmanc501f552004-03-01 17:47:27 +00001191
Chris Lattner48b383b02003-11-25 01:02:51 +00001192<p>The <a href="#t_firstclass">first class</a> types are perhaps the
1193most important. Values of these types are the only ones which can be
1194produced by instructions, passed as arguments, or used as operands to
Dan Gohman34d1c0d2008-05-23 21:53:15 +00001195instructions.</p>
Misha Brukman76307852003-11-08 01:05:38 +00001196</div>
Chris Lattner74d3f822004-12-09 17:30:23 +00001197
Chris Lattner2f7c9632001-06-06 20:29:01 +00001198<!-- ======================================================================= -->
Chris Lattner7824d182008-01-04 04:32:38 +00001199<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner43542b32008-01-04 04:34:14 +00001200
Chris Lattner7824d182008-01-04 04:32:38 +00001201<div class="doc_text">
1202<p>The primitive types are the fundamental building blocks of the LLVM
1203system.</p>
1204
Chris Lattner43542b32008-01-04 04:34:14 +00001205</div>
1206
Chris Lattner7824d182008-01-04 04:32:38 +00001207<!-- _______________________________________________________________________ -->
1208<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
1209
1210<div class="doc_text">
1211 <table>
1212 <tbody>
1213 <tr><th>Type</th><th>Description</th></tr>
1214 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
1215 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
1216 <tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
1217 <tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
1218 <tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
1219 </tbody>
1220 </table>
1221</div>
1222
1223<!-- _______________________________________________________________________ -->
1224<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
1225
1226<div class="doc_text">
1227<h5>Overview:</h5>
1228<p>The void type does not represent any value and has no size.</p>
1229
1230<h5>Syntax:</h5>
1231
1232<pre>
1233 void
1234</pre>
1235</div>
1236
1237<!-- _______________________________________________________________________ -->
1238<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
1239
1240<div class="doc_text">
1241<h5>Overview:</h5>
1242<p>The label type represents code labels.</p>
1243
1244<h5>Syntax:</h5>
1245
1246<pre>
1247 label
1248</pre>
1249</div>
1250
1251
1252<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001253<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +00001254
Misha Brukman76307852003-11-08 01:05:38 +00001255<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +00001256
Chris Lattner48b383b02003-11-25 01:02:51 +00001257<p>The real power in LLVM comes from the derived types in the system.
1258This is what allows a programmer to represent arrays, functions,
1259pointers, and other useful types. Note that these derived types may be
1260recursive: For example, it is possible to have a two dimensional array.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00001261
Misha Brukman76307852003-11-08 01:05:38 +00001262</div>
Chris Lattner74d3f822004-12-09 17:30:23 +00001263
Chris Lattner2f7c9632001-06-06 20:29:01 +00001264<!-- _______________________________________________________________________ -->
Reid Spencer138249b2007-05-16 18:44:01 +00001265<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1266
1267<div class="doc_text">
1268
1269<h5>Overview:</h5>
1270<p>The integer type is a very simple derived type that simply specifies an
1271arbitrary bit width for the integer type desired. Any bit width from 1 bit to
12722^23-1 (about 8 million) can be specified.</p>
1273
1274<h5>Syntax:</h5>
1275
1276<pre>
1277 iN
1278</pre>
1279
1280<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1281value.</p>
1282
1283<h5>Examples:</h5>
1284<table class="layout">
Chris Lattner9a2e3cb2007-12-18 06:18:21 +00001285 <tbody>
1286 <tr>
1287 <td><tt>i1</tt></td>
1288 <td>a single-bit integer.</td>
1289 </tr><tr>
1290 <td><tt>i32</tt></td>
1291 <td>a 32-bit integer.</td>
1292 </tr><tr>
1293 <td><tt>i1942652</tt></td>
1294 <td>a really big integer of over 1 million bits.</td>
Reid Spencer138249b2007-05-16 18:44:01 +00001295 </tr>
Chris Lattner9a2e3cb2007-12-18 06:18:21 +00001296 </tbody>
Reid Spencer138249b2007-05-16 18:44:01 +00001297</table>
Bill Wendling3716c5d2007-05-29 09:04:49 +00001298</div>
Reid Spencer138249b2007-05-16 18:44:01 +00001299
1300<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001301<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +00001302
Misha Brukman76307852003-11-08 01:05:38 +00001303<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +00001304
Chris Lattner2f7c9632001-06-06 20:29:01 +00001305<h5>Overview:</h5>
Chris Lattner74d3f822004-12-09 17:30:23 +00001306
Misha Brukman76307852003-11-08 01:05:38 +00001307<p>The array type is a very simple derived type that arranges elements
Chris Lattner48b383b02003-11-25 01:02:51 +00001308sequentially in memory. The array type requires a size (number of
1309elements) and an underlying data type.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00001310
Chris Lattner590645f2002-04-14 06:13:44 +00001311<h5>Syntax:</h5>
Chris Lattner74d3f822004-12-09 17:30:23 +00001312
1313<pre>
1314 [&lt;# elements&gt; x &lt;elementtype&gt;]
1315</pre>
1316
John Criswell02fdc6f2005-05-12 16:52:32 +00001317<p>The number of elements is a constant integer value; elementtype may
Chris Lattner48b383b02003-11-25 01:02:51 +00001318be any type with a size.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00001319
Chris Lattner590645f2002-04-14 06:13:44 +00001320<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001321<table class="layout">
1322 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +00001323 <td class="left"><tt>[40 x i32]</tt></td>
1324 <td class="left">Array of 40 32-bit integer values.</td>
1325 </tr>
1326 <tr class="layout">
1327 <td class="left"><tt>[41 x i32]</tt></td>
1328 <td class="left">Array of 41 32-bit integer values.</td>
1329 </tr>
1330 <tr class="layout">
1331 <td class="left"><tt>[4 x i8]</tt></td>
1332 <td class="left">Array of 4 8-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001333 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +00001334</table>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001335<p>Here are some examples of multidimensional arrays:</p>
1336<table class="layout">
1337 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +00001338 <td class="left"><tt>[3 x [4 x i32]]</tt></td>
1339 <td class="left">3x4 array of 32-bit integer values.</td>
1340 </tr>
1341 <tr class="layout">
1342 <td class="left"><tt>[12 x [10 x float]]</tt></td>
1343 <td class="left">12x10 array of single precision floating point values.</td>
1344 </tr>
1345 <tr class="layout">
1346 <td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
1347 <td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001348 </tr>
1349</table>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +00001350
John Criswell4c0cf7f2005-10-24 16:17:18 +00001351<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1352length array. Normally, accesses past the end of an array are undefined in
Chris Lattnerc0ad71e2005-06-24 17:22:57 +00001353LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1354As a special case, however, zero length arrays are recognized to be variable
1355length. This allows implementation of 'pascal style arrays' with the LLVM
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00001356type "{ i32, [0 x float]}", for example.</p>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +00001357
Misha Brukman76307852003-11-08 01:05:38 +00001358</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001359
Chris Lattner2f7c9632001-06-06 20:29:01 +00001360<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001361<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +00001362<div class="doc_text">
Chris Lattnerda508ac2008-04-23 04:59:35 +00001363
Chris Lattner2f7c9632001-06-06 20:29:01 +00001364<h5>Overview:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +00001365
Chris Lattner48b383b02003-11-25 01:02:51 +00001366<p>The function type can be thought of as a function signature. It
Devang Patele3dfc1c2008-03-24 05:35:41 +00001367consists of a return type and a list of formal parameter types. The
Chris Lattnerda508ac2008-04-23 04:59:35 +00001368return type of a function type is a scalar type, a void type, or a struct type.
Devang Patel9c1f8b12008-03-24 20:52:42 +00001369If the return type is a struct type then all struct elements must be of first
Chris Lattnerda508ac2008-04-23 04:59:35 +00001370class types, and the struct must have at least one element.</p>
Devang Pateld6cff512008-03-10 20:49:15 +00001371
Chris Lattner2f7c9632001-06-06 20:29:01 +00001372<h5>Syntax:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +00001373
1374<pre>
1375 &lt;returntype list&gt; (&lt;parameter list&gt;)
1376</pre>
1377
John Criswell4c0cf7f2005-10-24 16:17:18 +00001378<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Misha Brukman20f9a622004-08-12 20:16:08 +00001379specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
Chris Lattner5ed60612003-09-03 00:41:47 +00001380which indicates that the function takes a variable number of arguments.
1381Variable argument functions can access their arguments with the <a
Devang Pateld6cff512008-03-10 20:49:15 +00001382 href="#int_varargs">variable argument handling intrinsic</a> functions.
1383'<tt>&lt;returntype list&gt;</tt>' is a comma-separated list of
1384<a href="#t_firstclass">first class</a> type specifiers.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +00001385
Chris Lattner2f7c9632001-06-06 20:29:01 +00001386<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001387<table class="layout">
1388 <tr class="layout">
Reid Spencer58c08712006-12-31 07:18:34 +00001389 <td class="left"><tt>i32 (i32)</tt></td>
1390 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001391 </td>
Reid Spencer58c08712006-12-31 07:18:34 +00001392 </tr><tr class="layout">
Reid Spencer314e1cb2007-07-19 23:13:04 +00001393 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencer655dcc62006-12-31 07:20:23 +00001394 </tt></td>
Reid Spencer58c08712006-12-31 07:18:34 +00001395 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1396 an <tt>i16</tt> that should be sign extended and a
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00001397 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
Reid Spencer58c08712006-12-31 07:18:34 +00001398 <tt>float</tt>.
1399 </td>
1400 </tr><tr class="layout">
1401 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1402 <td class="left">A vararg function that takes at least one
Reid Spencer3e628eb92007-01-04 16:43:23 +00001403 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
Reid Spencer58c08712006-12-31 07:18:34 +00001404 which returns an integer. This is the signature for <tt>printf</tt> in
1405 LLVM.
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001406 </td>
Devang Patele3dfc1c2008-03-24 05:35:41 +00001407 </tr><tr class="layout">
1408 <td class="left"><tt>{i32, i32} (i32)</tt></td>
Misha Brukmanc9813bd2008-11-27 06:41:20 +00001409 <td class="left">A function taking an <tt>i32</tt>, returning two
1410 <tt>i32</tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
Devang Patele3dfc1c2008-03-24 05:35:41 +00001411 </td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001412 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +00001413</table>
Misha Brukmanc501f552004-03-01 17:47:27 +00001414
Misha Brukman76307852003-11-08 01:05:38 +00001415</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +00001416<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001417<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +00001418<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +00001419<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +00001420<p>The structure type is used to represent a collection of data members
1421together in memory. The packing of the field types is defined to match
1422the ABI of the underlying processor. The elements of a structure may
1423be any type that has a size.</p>
1424<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1425and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1426field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1427instruction.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +00001428<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +00001429<pre> { &lt;type list&gt; }<br></pre>
Chris Lattner2f7c9632001-06-06 20:29:01 +00001430<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001431<table class="layout">
1432 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +00001433 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1434 <td class="left">A triple of three <tt>i32</tt> values</td>
1435 </tr><tr class="layout">
1436 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1437 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1438 second element is a <a href="#t_pointer">pointer</a> to a
1439 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1440 an <tt>i32</tt>.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001441 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +00001442</table>
Misha Brukman76307852003-11-08 01:05:38 +00001443</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001444
Chris Lattner2f7c9632001-06-06 20:29:01 +00001445<!-- _______________________________________________________________________ -->
Andrew Lenharth8df88e22006-12-08 17:13:00 +00001446<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1447</div>
1448<div class="doc_text">
1449<h5>Overview:</h5>
1450<p>The packed structure type is used to represent a collection of data members
1451together in memory. There is no padding between fields. Further, the alignment
1452of a packed structure is 1 byte. The elements of a packed structure may
1453be any type that has a size.</p>
1454<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1455and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1456field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1457instruction.</p>
1458<h5>Syntax:</h5>
1459<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1460<h5>Examples:</h5>
1461<table class="layout">
1462 <tr class="layout">
Jeff Cohen5819f182007-04-22 01:17:39 +00001463 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1464 <td class="left">A triple of three <tt>i32</tt> values</td>
1465 </tr><tr class="layout">
Bill Wendlingb175fa42008-09-07 10:26:33 +00001466 <td class="left">
1467<tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)*&nbsp;}&nbsp;&gt;</tt></td>
Jeff Cohen5819f182007-04-22 01:17:39 +00001468 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1469 second element is a <a href="#t_pointer">pointer</a> to a
1470 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1471 an <tt>i32</tt>.</td>
Andrew Lenharth8df88e22006-12-08 17:13:00 +00001472 </tr>
1473</table>
1474</div>
1475
1476<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001477<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +00001478<div class="doc_text">
Chris Lattner590645f2002-04-14 06:13:44 +00001479<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +00001480<p>As in many languages, the pointer type represents a pointer or
Christopher Lamb308121c2007-12-11 09:31:00 +00001481reference to another object, which must live in memory. Pointer types may have
1482an optional address space attribute defining the target-specific numbered
1483address space where the pointed-to object resides. The default address space is
1484zero.</p>
Chris Lattner590645f2002-04-14 06:13:44 +00001485<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +00001486<pre> &lt;type&gt; *<br></pre>
Chris Lattner590645f2002-04-14 06:13:44 +00001487<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001488<table class="layout">
1489 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +00001490 <td class="left"><tt>[4x i32]*</tt></td>
1491 <td class="left">A <a href="#t_pointer">pointer</a> to <a
1492 href="#t_array">array</a> of four <tt>i32</tt> values.</td>
1493 </tr>
1494 <tr class="layout">
1495 <td class="left"><tt>i32 (i32 *) *</tt></td>
1496 <td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00001497 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner747359f2007-12-19 05:04:11 +00001498 <tt>i32</tt>.</td>
1499 </tr>
1500 <tr class="layout">
1501 <td class="left"><tt>i32 addrspace(5)*</tt></td>
1502 <td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
1503 that resides in address space #5.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001504 </tr>
Misha Brukman76307852003-11-08 01:05:38 +00001505</table>
Misha Brukman76307852003-11-08 01:05:38 +00001506</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001507
Chris Lattnerc8cb6952004-08-12 19:12:28 +00001508<!-- _______________________________________________________________________ -->
Reid Spencer404a3252007-02-15 03:07:05 +00001509<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +00001510<div class="doc_text">
Chris Lattner37b6b092005-04-25 17:34:15 +00001511
Chris Lattnerc8cb6952004-08-12 19:12:28 +00001512<h5>Overview:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +00001513
Reid Spencer404a3252007-02-15 03:07:05 +00001514<p>A vector type is a simple derived type that represents a vector
1515of elements. Vector types are used when multiple primitive data
Chris Lattnerc8cb6952004-08-12 19:12:28 +00001516are operated in parallel using a single instruction (SIMD).
Reid Spencer404a3252007-02-15 03:07:05 +00001517A vector type requires a size (number of
Chris Lattner330ce692005-11-10 01:44:22 +00001518elements) and an underlying primitive data type. Vectors must have a power
Reid Spencer404a3252007-02-15 03:07:05 +00001519of two length (1, 2, 4, 8, 16 ...). Vector types are
Chris Lattnerc8cb6952004-08-12 19:12:28 +00001520considered <a href="#t_firstclass">first class</a>.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +00001521
Chris Lattnerc8cb6952004-08-12 19:12:28 +00001522<h5>Syntax:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +00001523
1524<pre>
1525 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1526</pre>
1527
John Criswell4a3327e2005-05-13 22:25:59 +00001528<p>The number of elements is a constant integer value; elementtype may
Chris Lattnerc0f423a2007-01-15 01:54:13 +00001529be any integer or floating point type.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +00001530
Chris Lattnerc8cb6952004-08-12 19:12:28 +00001531<h5>Examples:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +00001532
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001533<table class="layout">
1534 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +00001535 <td class="left"><tt>&lt;4 x i32&gt;</tt></td>
1536 <td class="left">Vector of 4 32-bit integer values.</td>
1537 </tr>
1538 <tr class="layout">
1539 <td class="left"><tt>&lt;8 x float&gt;</tt></td>
1540 <td class="left">Vector of 8 32-bit floating-point values.</td>
1541 </tr>
1542 <tr class="layout">
1543 <td class="left"><tt>&lt;2 x i64&gt;</tt></td>
1544 <td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +00001545 </tr>
1546</table>
Misha Brukman76307852003-11-08 01:05:38 +00001547</div>
1548
Chris Lattner37b6b092005-04-25 17:34:15 +00001549<!-- _______________________________________________________________________ -->
1550<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1551<div class="doc_text">
1552
1553<h5>Overview:</h5>
1554
1555<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksena699c4d2007-10-14 00:34:53 +00001556corresponds (for example) to the C notion of a forward declared structure type.
Chris Lattner37b6b092005-04-25 17:34:15 +00001557In LLVM, opaque types can eventually be resolved to any type (not just a
1558structure type).</p>
1559
1560<h5>Syntax:</h5>
1561
1562<pre>
1563 opaque
1564</pre>
1565
1566<h5>Examples:</h5>
1567
1568<table class="layout">
1569 <tr class="layout">
Chris Lattner747359f2007-12-19 05:04:11 +00001570 <td class="left"><tt>opaque</tt></td>
1571 <td class="left">An opaque type.</td>
Chris Lattner37b6b092005-04-25 17:34:15 +00001572 </tr>
1573</table>
1574</div>
1575
1576
Chris Lattner74d3f822004-12-09 17:30:23 +00001577<!-- *********************************************************************** -->
1578<div class="doc_section"> <a name="constants">Constants</a> </div>
1579<!-- *********************************************************************** -->
1580
1581<div class="doc_text">
1582
1583<p>LLVM has several different basic types of constants. This section describes
1584them all and their syntax.</p>
1585
1586</div>
1587
1588<!-- ======================================================================= -->
Reid Spencer8f08d802004-12-09 18:02:53 +00001589<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattner74d3f822004-12-09 17:30:23 +00001590
1591<div class="doc_text">
1592
1593<dl>
1594 <dt><b>Boolean constants</b></dt>
1595
1596 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Reid Spencer36a15422007-01-12 03:35:51 +00001597 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
Chris Lattner74d3f822004-12-09 17:30:23 +00001598 </dd>
1599
1600 <dt><b>Integer constants</b></dt>
1601
Reid Spencer8f08d802004-12-09 18:02:53 +00001602 <dd>Standard integers (such as '4') are constants of the <a
Reid Spencer3e628eb92007-01-04 16:43:23 +00001603 href="#t_integer">integer</a> type. Negative numbers may be used with
Chris Lattner74d3f822004-12-09 17:30:23 +00001604 integer types.
1605 </dd>
1606
1607 <dt><b>Floating point constants</b></dt>
1608
1609 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1610 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattner1429e6f2008-04-01 18:45:27 +00001611 notation (see below). The assembler requires the exact decimal value of
1612 a floating-point constant. For example, the assembler accepts 1.25 but
1613 rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
1614 constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00001615
1616 <dt><b>Null pointer constants</b></dt>
1617
John Criswelldfe6a862004-12-10 15:51:16 +00001618 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Chris Lattner74d3f822004-12-09 17:30:23 +00001619 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1620
1621</dl>
1622
John Criswelldfe6a862004-12-10 15:51:16 +00001623<p>The one non-intuitive notation for constants is the optional hexadecimal form
Chris Lattner74d3f822004-12-09 17:30:23 +00001624of floating point constants. For example, the form '<tt>double
16250x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
16264.5e+15</tt>'. The only time hexadecimal floating point constants are required
Reid Spencer8f08d802004-12-09 18:02:53 +00001627(and the only time that they are generated by the disassembler) is when a
1628floating point constant must be emitted but it cannot be represented as a
1629decimal floating point number. For example, NaN's, infinities, and other
1630special values are represented in their IEEE hexadecimal format so that
1631assembly and disassembly do not cause any bits to change in the constants.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00001632
1633</div>
1634
1635<!-- ======================================================================= -->
1636<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1637</div>
1638
1639<div class="doc_text">
Chris Lattner455fc8c2005-03-07 22:13:59 +00001640<p>Aggregate constants arise from aggregation of simple constants
1641and smaller aggregate constants.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00001642
1643<dl>
1644 <dt><b>Structure constants</b></dt>
1645
1646 <dd>Structure constants are represented with notation similar to structure
1647 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattnerbea11172007-12-25 20:34:52 +00001648 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
1649 where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>". Structure constants
Chris Lattner455fc8c2005-03-07 22:13:59 +00001650 must have <a href="#t_struct">structure type</a>, and the number and
Chris Lattner74d3f822004-12-09 17:30:23 +00001651 types of elements must match those specified by the type.
1652 </dd>
1653
1654 <dt><b>Array constants</b></dt>
1655
1656 <dd>Array constants are represented with notation similar to array type
1657 definitions (a comma separated list of elements, surrounded by square brackets
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00001658 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
Chris Lattner74d3f822004-12-09 17:30:23 +00001659 constants must have <a href="#t_array">array type</a>, and the number and
1660 types of elements must match those specified by the type.
1661 </dd>
1662
Reid Spencer404a3252007-02-15 03:07:05 +00001663 <dt><b>Vector constants</b></dt>
Chris Lattner74d3f822004-12-09 17:30:23 +00001664
Reid Spencer404a3252007-02-15 03:07:05 +00001665 <dd>Vector constants are represented with notation similar to vector type
Chris Lattner74d3f822004-12-09 17:30:23 +00001666 definitions (a comma separated list of elements, surrounded by
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00001667 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
Jeff Cohen5819f182007-04-22 01:17:39 +00001668 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
Reid Spencer404a3252007-02-15 03:07:05 +00001669 href="#t_vector">vector type</a>, and the number and types of elements must
Chris Lattner74d3f822004-12-09 17:30:23 +00001670 match those specified by the type.
1671 </dd>
1672
1673 <dt><b>Zero initialization</b></dt>
1674
1675 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1676 value to zero of <em>any</em> type, including scalar and aggregate types.
1677 This is often used to avoid having to print large zero initializers (e.g. for
John Criswell4c0cf7f2005-10-24 16:17:18 +00001678 large arrays) and is always exactly equivalent to using explicit zero
Chris Lattner74d3f822004-12-09 17:30:23 +00001679 initializers.
1680 </dd>
1681</dl>
1682
1683</div>
1684
1685<!-- ======================================================================= -->
1686<div class="doc_subsection">
1687 <a name="globalconstants">Global Variable and Function Addresses</a>
1688</div>
1689
1690<div class="doc_text">
1691
1692<p>The addresses of <a href="#globalvars">global variables</a> and <a
1693href="#functionstructure">functions</a> are always implicitly valid (link-time)
John Criswelldfe6a862004-12-10 15:51:16 +00001694constants. These constants are explicitly referenced when the <a
1695href="#identifiers">identifier for the global</a> is used and always have <a
Chris Lattner74d3f822004-12-09 17:30:23 +00001696href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1697file:</p>
1698
Bill Wendling3716c5d2007-05-29 09:04:49 +00001699<div class="doc_code">
Chris Lattner74d3f822004-12-09 17:30:23 +00001700<pre>
Chris Lattner00538a12007-06-06 18:28:13 +00001701@X = global i32 17
1702@Y = global i32 42
1703@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattner74d3f822004-12-09 17:30:23 +00001704</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +00001705</div>
Chris Lattner74d3f822004-12-09 17:30:23 +00001706
1707</div>
1708
1709<!-- ======================================================================= -->
Reid Spencer641f5c92004-12-09 18:13:12 +00001710<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattner74d3f822004-12-09 17:30:23 +00001711<div class="doc_text">
Reid Spencer641f5c92004-12-09 18:13:12 +00001712 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
John Criswell4a3327e2005-05-13 22:25:59 +00001713 no specific value. Undefined values may be of any type and be used anywhere
Reid Spencer641f5c92004-12-09 18:13:12 +00001714 a constant is permitted.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00001715
Reid Spencer641f5c92004-12-09 18:13:12 +00001716 <p>Undefined values indicate to the compiler that the program is well defined
1717 no matter what value is used, giving the compiler more freedom to optimize.
1718 </p>
Chris Lattner74d3f822004-12-09 17:30:23 +00001719</div>
1720
1721<!-- ======================================================================= -->
1722<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1723</div>
1724
1725<div class="doc_text">
1726
1727<p>Constant expressions are used to allow expressions involving other constants
1728to be used as constants. Constant expressions may be of any <a
John Criswell4a3327e2005-05-13 22:25:59 +00001729href="#t_firstclass">first class</a> type and may involve any LLVM operation
Chris Lattner74d3f822004-12-09 17:30:23 +00001730that does not have side effects (e.g. load and call are not supported). The
1731following is the syntax for constant expressions:</p>
1732
1733<dl>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00001734 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1735 <dd>Truncate a constant to another type. The bit size of CST must be larger
Chris Lattnerc0f423a2007-01-15 01:54:13 +00001736 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00001737
Reid Spencer59b6b7d2006-11-08 01:11:31 +00001738 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1739 <dd>Zero extend a constant to another type. The bit size of CST must be
Chris Lattnerc0f423a2007-01-15 01:54:13 +00001740 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00001741
1742 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1743 <dd>Sign extend a constant to another type. The bit size of CST must be
Chris Lattnerc0f423a2007-01-15 01:54:13 +00001744 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00001745
1746 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1747 <dd>Truncate a floating point constant to another floating point type. The
1748 size of CST must be larger than the size of TYPE. Both types must be
1749 floating point.</dd>
1750
1751 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1752 <dd>Floating point extend a constant to another type. The size of CST must be
1753 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1754
Reid Spencer753163d2007-07-31 14:40:14 +00001755 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00001756 <dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begemand4d45c22007-11-17 03:58:34 +00001757 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1758 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1759 of the same number of elements. If the value won't fit in the integer type,
1760 the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00001761
Reid Spencer51b07252006-11-09 23:03:26 +00001762 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00001763 <dd>Convert a floating point constant to the corresponding signed integer
Nate Begemand4d45c22007-11-17 03:58:34 +00001764 constant. TYPE must be a scalar or vector integer type. CST must be of scalar
1765 or vector floating point type. Both CST and TYPE must be scalars, or vectors
1766 of the same number of elements. If the value won't fit in the integer type,
1767 the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00001768
Reid Spencer51b07252006-11-09 23:03:26 +00001769 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00001770 <dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begemand4d45c22007-11-17 03:58:34 +00001771 constant. TYPE must be a scalar or vector floating point type. CST must be of
1772 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1773 of the same number of elements. If the value won't fit in the floating point
1774 type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00001775
Reid Spencer51b07252006-11-09 23:03:26 +00001776 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00001777 <dd>Convert a signed integer constant to the corresponding floating point
Nate Begemand4d45c22007-11-17 03:58:34 +00001778 constant. TYPE must be a scalar or vector floating point type. CST must be of
1779 scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
1780 of the same number of elements. If the value won't fit in the floating point
1781 type, the results are undefined.</dd>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00001782
Reid Spencer5b950642006-11-11 23:08:07 +00001783 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1784 <dd>Convert a pointer typed constant to the corresponding integer constant
1785 TYPE must be an integer type. CST must be of pointer type. The CST value is
1786 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1787
1788 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1789 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1790 pointer type. CST must be of integer type. The CST value is zero extended,
1791 truncated, or unchanged to make it fit in a pointer size. This one is
1792 <i>really</i> dangerous!</dd>
1793
1794 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00001795 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1796 identical (same number of bits). The conversion is done as if the CST value
1797 was stored to memory and read back as TYPE. In other words, no bits change
Reid Spencer5b950642006-11-11 23:08:07 +00001798 with this operator, just the type. This can be used for conversion of
Reid Spencer404a3252007-02-15 03:07:05 +00001799 vector types to any other type, as long as they have the same bit width. For
Dan Gohmanc05dca92008-09-08 16:45:59 +00001800 pointers it is only valid to cast to another pointer type. It is not valid
1801 to bitcast to or from an aggregate type.
Reid Spencer59b6b7d2006-11-08 01:11:31 +00001802 </dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00001803
1804 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1805
1806 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1807 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1808 instruction, the index list may have zero or more indexes, which are required
1809 to make sense for the type of "CSTPTR".</dd>
1810
Robert Bocchino7e97a6d2006-01-10 19:31:34 +00001811 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1812
1813 <dd>Perform the <a href="#i_select">select operation</a> on
Reid Spencer9965ee72006-12-04 19:23:19 +00001814 constants.</dd>
1815
1816 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1817 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1818
1819 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1820 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +00001821
Nate Begemand2195702008-05-12 19:01:56 +00001822 <dt><b><tt>vicmp COND ( VAL1, VAL2 )</tt></b></dt>
1823 <dd>Performs the <a href="#i_vicmp">vicmp operation</a> on constants.</dd>
1824
1825 <dt><b><tt>vfcmp COND ( VAL1, VAL2 )</tt></b></dt>
1826 <dd>Performs the <a href="#i_vfcmp">vfcmp operation</a> on constants.</dd>
1827
Robert Bocchino7e97a6d2006-01-10 19:31:34 +00001828 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1829
1830 <dd>Perform the <a href="#i_extractelement">extractelement
Dan Gohmanef9462f2008-10-14 16:51:45 +00001831 operation</a> on constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +00001832
Robert Bocchinof72fdfe2006-01-15 20:48:27 +00001833 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1834
1835 <dd>Perform the <a href="#i_insertelement">insertelement
Reid Spencer9965ee72006-12-04 19:23:19 +00001836 operation</a> on constants.</dd>
Robert Bocchinof72fdfe2006-01-15 20:48:27 +00001837
Chris Lattner016a0e52006-04-08 00:13:41 +00001838
1839 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1840
1841 <dd>Perform the <a href="#i_shufflevector">shufflevector
Reid Spencer9965ee72006-12-04 19:23:19 +00001842 operation</a> on constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +00001843
Chris Lattner74d3f822004-12-09 17:30:23 +00001844 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1845
Reid Spencer641f5c92004-12-09 18:13:12 +00001846 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1847 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
Chris Lattner74d3f822004-12-09 17:30:23 +00001848 binary</a> operations. The constraints on operands are the same as those for
1849 the corresponding instruction (e.g. no bitwise operations on floating point
John Criswell02fdc6f2005-05-12 16:52:32 +00001850 values are allowed).</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +00001851</dl>
Chris Lattner74d3f822004-12-09 17:30:23 +00001852</div>
Chris Lattnerb1652612004-03-08 16:49:10 +00001853
Chris Lattner2f7c9632001-06-06 20:29:01 +00001854<!-- *********************************************************************** -->
Chris Lattner98f013c2006-01-25 23:47:57 +00001855<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1856<!-- *********************************************************************** -->
1857
1858<!-- ======================================================================= -->
1859<div class="doc_subsection">
1860<a name="inlineasm">Inline Assembler Expressions</a>
1861</div>
1862
1863<div class="doc_text">
1864
1865<p>
1866LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1867Module-Level Inline Assembly</a>) through the use of a special value. This
1868value represents the inline assembler as a string (containing the instructions
1869to emit), a list of operand constraints (stored as a string), and a flag that
1870indicates whether or not the inline asm expression has side effects. An example
1871inline assembler expression is:
1872</p>
1873
Bill Wendling3716c5d2007-05-29 09:04:49 +00001874<div class="doc_code">
Chris Lattner98f013c2006-01-25 23:47:57 +00001875<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +00001876i32 (i32) asm "bswap $0", "=r,r"
Chris Lattner98f013c2006-01-25 23:47:57 +00001877</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +00001878</div>
Chris Lattner98f013c2006-01-25 23:47:57 +00001879
1880<p>
1881Inline assembler expressions may <b>only</b> be used as the callee operand of
1882a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1883</p>
1884
Bill Wendling3716c5d2007-05-29 09:04:49 +00001885<div class="doc_code">
Chris Lattner98f013c2006-01-25 23:47:57 +00001886<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +00001887%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
Chris Lattner98f013c2006-01-25 23:47:57 +00001888</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +00001889</div>
Chris Lattner98f013c2006-01-25 23:47:57 +00001890
1891<p>
1892Inline asms with side effects not visible in the constraint list must be marked
1893as having side effects. This is done through the use of the
1894'<tt>sideeffect</tt>' keyword, like so:
1895</p>
1896
Bill Wendling3716c5d2007-05-29 09:04:49 +00001897<div class="doc_code">
Chris Lattner98f013c2006-01-25 23:47:57 +00001898<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +00001899call void asm sideeffect "eieio", ""()
Chris Lattner98f013c2006-01-25 23:47:57 +00001900</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +00001901</div>
Chris Lattner98f013c2006-01-25 23:47:57 +00001902
1903<p>TODO: The format of the asm and constraints string still need to be
1904documented here. Constraints on what can be done (e.g. duplication, moving, etc
Chris Lattnerd5528262008-10-04 18:36:02 +00001905need to be documented). This is probably best done by reference to another
1906document that covers inline asm from a holistic perspective.
Chris Lattner98f013c2006-01-25 23:47:57 +00001907</p>
1908
1909</div>
1910
1911<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001912<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1913<!-- *********************************************************************** -->
Chris Lattner74d3f822004-12-09 17:30:23 +00001914
Misha Brukman76307852003-11-08 01:05:38 +00001915<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +00001916
Chris Lattner48b383b02003-11-25 01:02:51 +00001917<p>The LLVM instruction set consists of several different
1918classifications of instructions: <a href="#terminators">terminator
John Criswell4a3327e2005-05-13 22:25:59 +00001919instructions</a>, <a href="#binaryops">binary instructions</a>,
1920<a href="#bitwiseops">bitwise binary instructions</a>, <a
Chris Lattner48b383b02003-11-25 01:02:51 +00001921 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1922instructions</a>.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00001923
Misha Brukman76307852003-11-08 01:05:38 +00001924</div>
Chris Lattner74d3f822004-12-09 17:30:23 +00001925
Chris Lattner2f7c9632001-06-06 20:29:01 +00001926<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001927<div class="doc_subsection"> <a name="terminators">Terminator
1928Instructions</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +00001929
Misha Brukman76307852003-11-08 01:05:38 +00001930<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +00001931
Chris Lattner48b383b02003-11-25 01:02:51 +00001932<p>As mentioned <a href="#functionstructure">previously</a>, every
1933basic block in a program ends with a "Terminator" instruction, which
1934indicates which block should be executed after the current block is
1935finished. These terminator instructions typically yield a '<tt>void</tt>'
1936value: they produce control flow, not values (the one exception being
1937the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
John Criswelldfe6a862004-12-10 15:51:16 +00001938<p>There are six different terminator instructions: the '<a
Chris Lattner48b383b02003-11-25 01:02:51 +00001939 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1940instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
Chris Lattner08b7d5b2004-10-16 18:04:13 +00001941the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1942 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1943 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +00001944
Misha Brukman76307852003-11-08 01:05:38 +00001945</div>
Chris Lattner74d3f822004-12-09 17:30:23 +00001946
Chris Lattner2f7c9632001-06-06 20:29:01 +00001947<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001948<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1949Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +00001950<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +00001951<h5>Syntax:</h5>
Dan Gohmancc3132e2008-10-04 19:00:07 +00001952<pre>
1953 ret &lt;type&gt; &lt;value&gt; <i>; Return a value from a non-void function</i>
Chris Lattner590645f2002-04-14 06:13:44 +00001954 ret void <i>; Return from void function</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00001955</pre>
Chris Lattnerda508ac2008-04-23 04:59:35 +00001956
Chris Lattner2f7c9632001-06-06 20:29:01 +00001957<h5>Overview:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +00001958
Dan Gohmancc3132e2008-10-04 19:00:07 +00001959<p>The '<tt>ret</tt>' instruction is used to return control flow (and
1960optionally a value) from a function back to the caller.</p>
John Criswell417228d2004-04-09 16:48:45 +00001961<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Dan Gohmancc3132e2008-10-04 19:00:07 +00001962returns a value and then causes control flow, and one that just causes
Chris Lattner48b383b02003-11-25 01:02:51 +00001963control flow to occur.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +00001964
Chris Lattner2f7c9632001-06-06 20:29:01 +00001965<h5>Arguments:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +00001966
Dan Gohmancc3132e2008-10-04 19:00:07 +00001967<p>The '<tt>ret</tt>' instruction optionally accepts a single argument,
1968the return value. The type of the return value must be a
1969'<a href="#t_firstclass">first class</a>' type.</p>
1970
1971<p>A function is not <a href="#wellformed">well formed</a> if
1972it it has a non-void return type and contains a '<tt>ret</tt>'
1973instruction with no return value or a return value with a type that
1974does not match its type, or if it has a void return type and contains
1975a '<tt>ret</tt>' instruction with a return value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +00001976
Chris Lattner2f7c9632001-06-06 20:29:01 +00001977<h5>Semantics:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +00001978
Chris Lattner48b383b02003-11-25 01:02:51 +00001979<p>When the '<tt>ret</tt>' instruction is executed, control flow
1980returns back to the calling function's context. If the caller is a "<a
John Criswell40db33f2004-06-25 15:16:57 +00001981 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
Chris Lattner48b383b02003-11-25 01:02:51 +00001982the instruction after the call. If the caller was an "<a
1983 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
John Criswell02fdc6f2005-05-12 16:52:32 +00001984at the beginning of the "normal" destination block. If the instruction
Chris Lattner48b383b02003-11-25 01:02:51 +00001985returns a value, that value shall set the call or invoke instruction's
Dan Gohmanef9462f2008-10-14 16:51:45 +00001986return value.</p>
Chris Lattnerda508ac2008-04-23 04:59:35 +00001987
Chris Lattner2f7c9632001-06-06 20:29:01 +00001988<h5>Example:</h5>
Chris Lattnerda508ac2008-04-23 04:59:35 +00001989
1990<pre>
1991 ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner590645f2002-04-14 06:13:44 +00001992 ret void <i>; Return from a void function</i>
Dan Gohmancc3132e2008-10-04 19:00:07 +00001993 ret { i32, i8 } { i32 4, i8 2 } <i>; Return an aggregate of values 4 and 2</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00001994</pre>
Misha Brukman76307852003-11-08 01:05:38 +00001995</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +00001996<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00001997<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +00001998<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +00001999<h5>Syntax:</h5>
Reid Spencer36a15422007-01-12 03:35:51 +00002000<pre> br i1 &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002001</pre>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002002<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +00002003<p>The '<tt>br</tt>' instruction is used to cause control flow to
2004transfer to a different basic block in the current function. There are
2005two forms of this instruction, corresponding to a conditional branch
2006and an unconditional branch.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002007<h5>Arguments:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +00002008<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
Reid Spencer36a15422007-01-12 03:35:51 +00002009single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
Reid Spencer50c723a2007-02-19 23:54:10 +00002010unconditional form of the '<tt>br</tt>' instruction takes a single
2011'<tt>label</tt>' value as a target.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002012<h5>Semantics:</h5>
Reid Spencer36a15422007-01-12 03:35:51 +00002013<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Chris Lattner48b383b02003-11-25 01:02:51 +00002014argument is evaluated. If the value is <tt>true</tt>, control flows
2015to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
2016control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002017<h5>Example:</h5>
Reid Spencer36a15422007-01-12 03:35:51 +00002018<pre>Test:<br> %cond = <a href="#i_icmp">icmp</a> eq, i32 %a, %b<br> br i1 %cond, label %IfEqual, label %IfUnequal<br>IfEqual:<br> <a
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002019 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
Misha Brukman76307852003-11-08 01:05:38 +00002020</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002021<!-- _______________________________________________________________________ -->
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002022<div class="doc_subsubsection">
2023 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
2024</div>
2025
Misha Brukman76307852003-11-08 01:05:38 +00002026<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +00002027<h5>Syntax:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002028
2029<pre>
2030 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
2031</pre>
2032
Chris Lattner2f7c9632001-06-06 20:29:01 +00002033<h5>Overview:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002034
2035<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
2036several different places. It is a generalization of the '<tt>br</tt>'
Misha Brukman76307852003-11-08 01:05:38 +00002037instruction, allowing a branch to occur to one of many possible
2038destinations.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002039
2040
Chris Lattner2f7c9632001-06-06 20:29:01 +00002041<h5>Arguments:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002042
2043<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
2044comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
2045an array of pairs of comparison value constants and '<tt>label</tt>'s. The
2046table is not allowed to contain duplicate constant entries.</p>
2047
Chris Lattner2f7c9632001-06-06 20:29:01 +00002048<h5>Semantics:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002049
Chris Lattner48b383b02003-11-25 01:02:51 +00002050<p>The <tt>switch</tt> instruction specifies a table of values and
2051destinations. When the '<tt>switch</tt>' instruction is executed, this
John Criswellbcbb18c2004-06-25 16:05:06 +00002052table is searched for the given value. If the value is found, control flow is
2053transfered to the corresponding destination; otherwise, control flow is
2054transfered to the default destination.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002055
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002056<h5>Implementation:</h5>
2057
2058<p>Depending on properties of the target machine and the particular
2059<tt>switch</tt> instruction, this instruction may be code generated in different
John Criswellbcbb18c2004-06-25 16:05:06 +00002060ways. For example, it could be generated as a series of chained conditional
2061branches or with a lookup table.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002062
2063<h5>Example:</h5>
2064
2065<pre>
2066 <i>; Emulate a conditional br instruction</i>
Reid Spencer36a15422007-01-12 03:35:51 +00002067 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002068 switch i32 %Val, label %truedest [i32 0, label %falsedest ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002069
2070 <i>; Emulate an unconditional br instruction</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002071 switch i32 0, label %dest [ ]
Chris Lattnercf96c6c2004-02-24 04:54:45 +00002072
2073 <i>; Implement a jump table:</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002074 switch i32 %val, label %otherwise [ i32 0, label %onzero
2075 i32 1, label %onone
2076 i32 2, label %ontwo ]
Chris Lattner2f7c9632001-06-06 20:29:01 +00002077</pre>
Misha Brukman76307852003-11-08 01:05:38 +00002078</div>
Chris Lattner0132aff2005-05-06 22:57:40 +00002079
Chris Lattner2f7c9632001-06-06 20:29:01 +00002080<!-- _______________________________________________________________________ -->
Chris Lattner0132aff2005-05-06 22:57:40 +00002081<div class="doc_subsubsection">
2082 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
2083</div>
2084
Misha Brukman76307852003-11-08 01:05:38 +00002085<div class="doc_text">
Chris Lattner0132aff2005-05-06 22:57:40 +00002086
Chris Lattner2f7c9632001-06-06 20:29:01 +00002087<h5>Syntax:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +00002088
2089<pre>
Devang Patel02256232008-10-07 17:48:33 +00002090 &lt;result&gt; = invoke [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>] &lt;ptr to function ty&gt; &lt;function ptr val&gt;(&lt;function args&gt;) [<a href="#fnattrs">fn attrs</a>]
Chris Lattner6b7a0082006-05-14 18:23:06 +00002091 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattner0132aff2005-05-06 22:57:40 +00002092</pre>
2093
Chris Lattnera8292f32002-05-06 22:08:29 +00002094<h5>Overview:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +00002095
2096<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
2097function, with the possibility of control flow transfer to either the
John Criswell02fdc6f2005-05-12 16:52:32 +00002098'<tt>normal</tt>' label or the
2099'<tt>exception</tt>' label. If the callee function returns with the
Chris Lattner0132aff2005-05-06 22:57:40 +00002100"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
2101"normal" label. If the callee (or any indirect callees) returns with the "<a
John Criswell02fdc6f2005-05-12 16:52:32 +00002102href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Dan Gohmanef9462f2008-10-14 16:51:45 +00002103continued at the dynamically nearest "exception" label.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +00002104
Chris Lattner2f7c9632001-06-06 20:29:01 +00002105<h5>Arguments:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +00002106
Misha Brukman76307852003-11-08 01:05:38 +00002107<p>This instruction requires several arguments:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +00002108
Chris Lattner2f7c9632001-06-06 20:29:01 +00002109<ol>
Chris Lattner0132aff2005-05-06 22:57:40 +00002110 <li>
Duncan Sands16f122e2007-03-30 12:22:09 +00002111 The optional "cconv" marker indicates which <a href="#callingconv">calling
Chris Lattner0132aff2005-05-06 22:57:40 +00002112 convention</a> the call should use. If none is specified, the call defaults
2113 to using C calling conventions.
2114 </li>
Devang Patel7e9b05e2008-10-06 18:50:38 +00002115
2116 <li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
2117 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
2118 and '<tt>inreg</tt>' attributes are valid here.</li>
2119
Chris Lattner0132aff2005-05-06 22:57:40 +00002120 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
2121 function value being invoked. In most cases, this is a direct function
2122 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
2123 an arbitrary pointer to function value.
2124 </li>
2125
2126 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
2127 function to be invoked. </li>
2128
2129 <li>'<tt>function args</tt>': argument list whose types match the function
2130 signature argument types. If the function signature indicates the function
2131 accepts a variable number of arguments, the extra arguments can be
2132 specified. </li>
2133
2134 <li>'<tt>normal label</tt>': the label reached when the called function
2135 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
2136
2137 <li>'<tt>exception label</tt>': the label reached when a callee returns with
2138 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
2139
Devang Patel02256232008-10-07 17:48:33 +00002140 <li>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patel7e9b05e2008-10-06 18:50:38 +00002141 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
2142 '<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002143</ol>
Chris Lattner0132aff2005-05-06 22:57:40 +00002144
Chris Lattner2f7c9632001-06-06 20:29:01 +00002145<h5>Semantics:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +00002146
Misha Brukman76307852003-11-08 01:05:38 +00002147<p>This instruction is designed to operate as a standard '<tt><a
Chris Lattner0132aff2005-05-06 22:57:40 +00002148href="#i_call">call</a></tt>' instruction in most regards. The primary
2149difference is that it establishes an association with a label, which is used by
2150the runtime library to unwind the stack.</p>
2151
2152<p>This instruction is used in languages with destructors to ensure that proper
2153cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
2154exception. Additionally, this is important for implementation of
2155'<tt>catch</tt>' clauses in high-level languages that support them.</p>
2156
Chris Lattner2f7c9632001-06-06 20:29:01 +00002157<h5>Example:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +00002158<pre>
Nick Lewycky084ab472008-03-16 07:18:12 +00002159 %retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +00002160 unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewycky084ab472008-03-16 07:18:12 +00002161 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen5819f182007-04-22 01:17:39 +00002162 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002163</pre>
Misha Brukman76307852003-11-08 01:05:38 +00002164</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00002165
2166
Chris Lattner5ed60612003-09-03 00:41:47 +00002167<!-- _______________________________________________________________________ -->
Chris Lattner08b7d5b2004-10-16 18:04:13 +00002168
Chris Lattner48b383b02003-11-25 01:02:51 +00002169<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
2170Instruction</a> </div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00002171
Misha Brukman76307852003-11-08 01:05:38 +00002172<div class="doc_text">
Chris Lattner08b7d5b2004-10-16 18:04:13 +00002173
Chris Lattner5ed60612003-09-03 00:41:47 +00002174<h5>Syntax:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00002175<pre>
2176 unwind
2177</pre>
2178
Chris Lattner5ed60612003-09-03 00:41:47 +00002179<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00002180
2181<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
2182at the first callee in the dynamic call stack which used an <a
2183href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
2184primarily used to implement exception handling.</p>
2185
Chris Lattner5ed60612003-09-03 00:41:47 +00002186<h5>Semantics:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00002187
Chris Lattnerfe8519c2008-04-19 21:01:16 +00002188<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Chris Lattner08b7d5b2004-10-16 18:04:13 +00002189immediately halt. The dynamic call stack is then searched for the first <a
2190href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
2191execution continues at the "exceptional" destination block specified by the
2192<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
2193dynamic call chain, undefined behavior results.</p>
Misha Brukman76307852003-11-08 01:05:38 +00002194</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +00002195
2196<!-- _______________________________________________________________________ -->
2197
2198<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
2199Instruction</a> </div>
2200
2201<div class="doc_text">
2202
2203<h5>Syntax:</h5>
2204<pre>
2205 unreachable
2206</pre>
2207
2208<h5>Overview:</h5>
2209
2210<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
2211instruction is used to inform the optimizer that a particular portion of the
2212code is not reachable. This can be used to indicate that the code after a
2213no-return function cannot be reached, and other facts.</p>
2214
2215<h5>Semantics:</h5>
2216
2217<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
2218</div>
2219
2220
2221
Chris Lattner2f7c9632001-06-06 20:29:01 +00002222<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +00002223<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +00002224<div class="doc_text">
Chris Lattner48b383b02003-11-25 01:02:51 +00002225<p>Binary operators are used to do most of the computation in a
Chris Lattner81f92972008-04-01 18:47:32 +00002226program. They require two operands of the same type, execute an operation on them, and
John Criswelldfe6a862004-12-10 15:51:16 +00002227produce a single value. The operands might represent
Reid Spencer404a3252007-02-15 03:07:05 +00002228multiple data, as is the case with the <a href="#t_vector">vector</a> data type.
Chris Lattner81f92972008-04-01 18:47:32 +00002229The result value has the same type as its operands.</p>
Misha Brukman76307852003-11-08 01:05:38 +00002230<p>There are several different binary operators:</p>
Misha Brukman76307852003-11-08 01:05:38 +00002231</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002232<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002233<div class="doc_subsubsection">
2234 <a name="i_add">'<tt>add</tt>' Instruction</a>
2235</div>
2236
Misha Brukman76307852003-11-08 01:05:38 +00002237<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002238
Chris Lattner2f7c9632001-06-06 20:29:01 +00002239<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002240
2241<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002242 &lt;result&gt; = add &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002243</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002244
Chris Lattner2f7c9632001-06-06 20:29:01 +00002245<h5>Overview:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002246
Misha Brukman76307852003-11-08 01:05:38 +00002247<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002248
Chris Lattner2f7c9632001-06-06 20:29:01 +00002249<h5>Arguments:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002250
2251<p>The two arguments to the '<tt>add</tt>' instruction must be <a
2252 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>, or
2253 <a href="#t_vector">vector</a> values. Both arguments must have identical
2254 types.</p>
2255
Chris Lattner2f7c9632001-06-06 20:29:01 +00002256<h5>Semantics:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002257
Misha Brukman76307852003-11-08 01:05:38 +00002258<p>The value produced is the integer or floating point sum of the two
2259operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002260
Chris Lattner2f2427e2008-01-28 00:36:27 +00002261<p>If an integer sum has unsigned overflow, the result returned is the
2262mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2263the result.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002264
Chris Lattner2f2427e2008-01-28 00:36:27 +00002265<p>Because LLVM integers use a two's complement representation, this
2266instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002267
Chris Lattner2f7c9632001-06-06 20:29:01 +00002268<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002269
2270<pre>
2271 &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002272</pre>
Misha Brukman76307852003-11-08 01:05:38 +00002273</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002274<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002275<div class="doc_subsubsection">
2276 <a name="i_sub">'<tt>sub</tt>' Instruction</a>
2277</div>
2278
Misha Brukman76307852003-11-08 01:05:38 +00002279<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002280
Chris Lattner2f7c9632001-06-06 20:29:01 +00002281<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002282
2283<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002284 &lt;result&gt; = sub &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002285</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002286
Chris Lattner2f7c9632001-06-06 20:29:01 +00002287<h5>Overview:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002288
Misha Brukman76307852003-11-08 01:05:38 +00002289<p>The '<tt>sub</tt>' instruction returns the difference of its two
2290operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002291
2292<p>Note that the '<tt>sub</tt>' instruction is used to represent the
2293'<tt>neg</tt>' instruction present in most other intermediate
2294representations.</p>
2295
Chris Lattner2f7c9632001-06-06 20:29:01 +00002296<h5>Arguments:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002297
2298<p>The two arguments to the '<tt>sub</tt>' instruction must be <a
2299 href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2300 or <a href="#t_vector">vector</a> values. Both arguments must have identical
2301 types.</p>
2302
Chris Lattner2f7c9632001-06-06 20:29:01 +00002303<h5>Semantics:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002304
Chris Lattner48b383b02003-11-25 01:02:51 +00002305<p>The value produced is the integer or floating point difference of
2306the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002307
Chris Lattner2f2427e2008-01-28 00:36:27 +00002308<p>If an integer difference has unsigned overflow, the result returned is the
2309mathematical result modulo 2<sup>n</sup>, where n is the bit width of
2310the result.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002311
Chris Lattner2f2427e2008-01-28 00:36:27 +00002312<p>Because LLVM integers use a two's complement representation, this
2313instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002314
Chris Lattner2f7c9632001-06-06 20:29:01 +00002315<h5>Example:</h5>
Bill Wendling2d8b9a82007-05-29 09:42:13 +00002316<pre>
2317 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002318 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002319</pre>
Misha Brukman76307852003-11-08 01:05:38 +00002320</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002321
Chris Lattner2f7c9632001-06-06 20:29:01 +00002322<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002323<div class="doc_subsubsection">
2324 <a name="i_mul">'<tt>mul</tt>' Instruction</a>
2325</div>
2326
Misha Brukman76307852003-11-08 01:05:38 +00002327<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002328
Chris Lattner2f7c9632001-06-06 20:29:01 +00002329<h5>Syntax:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002330<pre> &lt;result&gt; = mul &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002331</pre>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002332<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +00002333<p>The '<tt>mul</tt>' instruction returns the product of its two
2334operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002335
Chris Lattner2f7c9632001-06-06 20:29:01 +00002336<h5>Arguments:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002337
2338<p>The two arguments to the '<tt>mul</tt>' instruction must be <a
2339href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
2340or <a href="#t_vector">vector</a> values. Both arguments must have identical
2341types.</p>
2342
Chris Lattner2f7c9632001-06-06 20:29:01 +00002343<h5>Semantics:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002344
Chris Lattner48b383b02003-11-25 01:02:51 +00002345<p>The value produced is the integer or floating point product of the
Misha Brukman76307852003-11-08 01:05:38 +00002346two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002347
Chris Lattner2f2427e2008-01-28 00:36:27 +00002348<p>If the result of an integer multiplication has unsigned overflow,
2349the result returned is the mathematical result modulo
23502<sup>n</sup>, where n is the bit width of the result.</p>
2351<p>Because LLVM integers use a two's complement representation, and the
2352result is the same width as the operands, this instruction returns the
2353correct result for both signed and unsigned integers. If a full product
2354(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
2355should be sign-extended or zero-extended as appropriate to the
2356width of the full product.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002357<h5>Example:</h5>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002358<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002359</pre>
Misha Brukman76307852003-11-08 01:05:38 +00002360</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002361
Chris Lattner2f7c9632001-06-06 20:29:01 +00002362<!-- _______________________________________________________________________ -->
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002363<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2364</a></div>
2365<div class="doc_text">
2366<h5>Syntax:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002367<pre> &lt;result&gt; = udiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002368</pre>
2369<h5>Overview:</h5>
2370<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2371operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002372
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002373<h5>Arguments:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002374
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002375<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002376<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2377values. Both arguments must have identical types.</p>
2378
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002379<h5>Semantics:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002380
Chris Lattner2f2427e2008-01-28 00:36:27 +00002381<p>The value produced is the unsigned integer quotient of the two operands.</p>
2382<p>Note that unsigned integer division and signed integer division are distinct
2383operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
2384<p>Division by zero leads to undefined behavior.</p>
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002385<h5>Example:</h5>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002386<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002387</pre>
2388</div>
2389<!-- _______________________________________________________________________ -->
2390<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2391</a> </div>
2392<div class="doc_text">
2393<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002394<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002395 &lt;result&gt; = sdiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002396</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002397
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002398<h5>Overview:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002399
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002400<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2401operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002402
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002403<h5>Arguments:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002404
2405<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2406<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2407values. Both arguments must have identical types.</p>
2408
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002409<h5>Semantics:</h5>
Chris Lattner1429e6f2008-04-01 18:45:27 +00002410<p>The value produced is the signed integer quotient of the two operands rounded towards zero.</p>
Chris Lattner2f2427e2008-01-28 00:36:27 +00002411<p>Note that signed integer division and unsigned integer division are distinct
2412operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
2413<p>Division by zero leads to undefined behavior. Overflow also leads to
2414undefined behavior; this is a rare case, but can occur, for example,
2415by doing a 32-bit division of -2147483648 by -1.</p>
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002416<h5>Example:</h5>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002417<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002418</pre>
2419</div>
2420<!-- _______________________________________________________________________ -->
2421<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
Chris Lattner48b383b02003-11-25 01:02:51 +00002422Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +00002423<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +00002424<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002425<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002426 &lt;result&gt; = fdiv &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner48b383b02003-11-25 01:02:51 +00002427</pre>
2428<h5>Overview:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002429
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002430<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
Chris Lattner48b383b02003-11-25 01:02:51 +00002431operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002432
Chris Lattner48b383b02003-11-25 01:02:51 +00002433<h5>Arguments:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002434
Jeff Cohen5819f182007-04-22 01:17:39 +00002435<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002436<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2437of floating point values. Both arguments must have identical types.</p>
2438
Chris Lattner48b383b02003-11-25 01:02:51 +00002439<h5>Semantics:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002440
Reid Spencer7e80b0b2006-10-26 06:15:43 +00002441<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002442
Chris Lattner48b383b02003-11-25 01:02:51 +00002443<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002444
2445<pre>
2446 &lt;result&gt; = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
Chris Lattner48b383b02003-11-25 01:02:51 +00002447</pre>
2448</div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002449
Chris Lattner48b383b02003-11-25 01:02:51 +00002450<!-- _______________________________________________________________________ -->
Reid Spencer7eb55b32006-11-02 01:53:59 +00002451<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2452</div>
2453<div class="doc_text">
2454<h5>Syntax:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002455<pre> &lt;result&gt; = urem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +00002456</pre>
2457<h5>Overview:</h5>
2458<p>The '<tt>urem</tt>' instruction returns the remainder from the
2459unsigned division of its two arguments.</p>
2460<h5>Arguments:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002461<p>The two arguments to the '<tt>urem</tt>' instruction must be
2462<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2463values. Both arguments must have identical types.</p>
Reid Spencer7eb55b32006-11-02 01:53:59 +00002464<h5>Semantics:</h5>
2465<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Chris Lattner1429e6f2008-04-01 18:45:27 +00002466This instruction always performs an unsigned division to get the remainder.</p>
Chris Lattner2f2427e2008-01-28 00:36:27 +00002467<p>Note that unsigned integer remainder and signed integer remainder are
2468distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
2469<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Reid Spencer7eb55b32006-11-02 01:53:59 +00002470<h5>Example:</h5>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002471<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +00002472</pre>
2473
2474</div>
2475<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002476<div class="doc_subsubsection">
2477 <a name="i_srem">'<tt>srem</tt>' Instruction</a>
2478</div>
2479
Chris Lattner48b383b02003-11-25 01:02:51 +00002480<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002481
Chris Lattner48b383b02003-11-25 01:02:51 +00002482<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002483
2484<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002485 &lt;result&gt; = srem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner48b383b02003-11-25 01:02:51 +00002486</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002487
Chris Lattner48b383b02003-11-25 01:02:51 +00002488<h5>Overview:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002489
Reid Spencer7eb55b32006-11-02 01:53:59 +00002490<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman08143e32007-11-05 23:35:22 +00002491signed division of its two operands. This instruction can also take
2492<a href="#t_vector">vector</a> versions of the values in which case
2493the elements must be integers.</p>
Chris Lattnerb8f816e2008-01-04 04:33:49 +00002494
Chris Lattner48b383b02003-11-25 01:02:51 +00002495<h5>Arguments:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002496
Reid Spencer7eb55b32006-11-02 01:53:59 +00002497<p>The two arguments to the '<tt>srem</tt>' instruction must be
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002498<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2499values. Both arguments must have identical types.</p>
2500
Chris Lattner48b383b02003-11-25 01:02:51 +00002501<h5>Semantics:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002502
Reid Spencer7eb55b32006-11-02 01:53:59 +00002503<p>This instruction returns the <i>remainder</i> of a division (where the result
Gabor Greif0f75ad02008-08-07 21:46:00 +00002504has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
2505operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
Reid Spencer806ad6a2007-03-24 22:23:39 +00002506a value. For more information about the difference, see <a
Chris Lattner48b383b02003-11-25 01:02:51 +00002507 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
Reid Spencer806ad6a2007-03-24 22:23:39 +00002508Math Forum</a>. For a table of how this is implemented in various languages,
Reid Spencerdb3b93b2007-03-24 22:40:44 +00002509please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
Reid Spencer806ad6a2007-03-24 22:23:39 +00002510Wikipedia: modulo operation</a>.</p>
Chris Lattner2f2427e2008-01-28 00:36:27 +00002511<p>Note that signed integer remainder and unsigned integer remainder are
2512distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
2513<p>Taking the remainder of a division by zero leads to undefined behavior.
2514Overflow also leads to undefined behavior; this is a rare case, but can occur,
2515for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
2516(The remainder doesn't actually overflow, but this rule lets srem be
2517implemented using instructions that return both the result of the division
2518and the remainder.)</p>
Chris Lattner48b383b02003-11-25 01:02:51 +00002519<h5>Example:</h5>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002520<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +00002521</pre>
2522
2523</div>
2524<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002525<div class="doc_subsubsection">
2526 <a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
2527
Reid Spencer7eb55b32006-11-02 01:53:59 +00002528<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002529
Reid Spencer7eb55b32006-11-02 01:53:59 +00002530<h5>Syntax:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002531<pre> &lt;result&gt; = frem &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer7eb55b32006-11-02 01:53:59 +00002532</pre>
2533<h5>Overview:</h5>
2534<p>The '<tt>frem</tt>' instruction returns the remainder from the
2535division of its two operands.</p>
2536<h5>Arguments:</h5>
2537<p>The two arguments to the '<tt>frem</tt>' instruction must be
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002538<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
2539of floating point values. Both arguments must have identical types.</p>
2540
Reid Spencer7eb55b32006-11-02 01:53:59 +00002541<h5>Semantics:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002542
Chris Lattner1429e6f2008-04-01 18:45:27 +00002543<p>This instruction returns the <i>remainder</i> of a division.
2544The remainder has the same sign as the dividend.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002545
Reid Spencer7eb55b32006-11-02 01:53:59 +00002546<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002547
2548<pre>
2549 &lt;result&gt; = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
Chris Lattner48b383b02003-11-25 01:02:51 +00002550</pre>
Misha Brukman76307852003-11-08 01:05:38 +00002551</div>
Robert Bocchino820bc75b2006-02-17 21:18:08 +00002552
Reid Spencer2ab01932007-02-02 13:57:07 +00002553<!-- ======================================================================= -->
2554<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2555Operations</a> </div>
2556<div class="doc_text">
2557<p>Bitwise binary operators are used to do various forms of
2558bit-twiddling in a program. They are generally very efficient
2559instructions and can commonly be strength reduced from other
Chris Lattner1429e6f2008-04-01 18:45:27 +00002560instructions. They require two operands of the same type, execute an operation on them,
2561and produce a single value. The resulting value is the same type as its operands.</p>
Reid Spencer2ab01932007-02-02 13:57:07 +00002562</div>
2563
Reid Spencer04e259b2007-01-31 21:39:12 +00002564<!-- _______________________________________________________________________ -->
2565<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2566Instruction</a> </div>
2567<div class="doc_text">
2568<h5>Syntax:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002569<pre> &lt;result&gt; = shl &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +00002570</pre>
Chris Lattnerf0e50112007-10-03 21:01:14 +00002571
Reid Spencer04e259b2007-01-31 21:39:12 +00002572<h5>Overview:</h5>
Chris Lattnerf0e50112007-10-03 21:01:14 +00002573
Reid Spencer04e259b2007-01-31 21:39:12 +00002574<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2575the left a specified number of bits.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +00002576
Reid Spencer04e259b2007-01-31 21:39:12 +00002577<h5>Arguments:</h5>
Chris Lattnerf0e50112007-10-03 21:01:14 +00002578
Reid Spencer04e259b2007-01-31 21:39:12 +00002579<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
Nate Begemanfecbc8c2008-07-29 15:49:41 +00002580 href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greif0f75ad02008-08-07 21:46:00 +00002581type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +00002582
Reid Spencer04e259b2007-01-31 21:39:12 +00002583<h5>Semantics:</h5>
Chris Lattnerf0e50112007-10-03 21:01:14 +00002584
Gabor Greif0f75ad02008-08-07 21:46:00 +00002585<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod 2<sup>n</sup>,
2586where n is the width of the result. If <tt>op2</tt> is (statically or dynamically) negative or
Mon P Wang68d4eee2008-12-10 08:55:09 +00002587equal to or larger than the number of bits in <tt>op1</tt>, the result is undefined.
2588If the arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
2589corresponding shift amount in <tt>op2</tt>.</p>
Chris Lattnerf0e50112007-10-03 21:01:14 +00002590
Reid Spencer04e259b2007-01-31 21:39:12 +00002591<h5>Example:</h5><pre>
2592 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2593 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2594 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +00002595 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +00002596 &lt;result&gt; = shl &lt;2 x i32&gt; &lt; i32 1, i32 1&gt;, &lt; i32 1, i32 2&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 2, i32 4&gt;</i>
Reid Spencer04e259b2007-01-31 21:39:12 +00002597</pre>
2598</div>
2599<!-- _______________________________________________________________________ -->
2600<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2601Instruction</a> </div>
2602<div class="doc_text">
2603<h5>Syntax:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002604<pre> &lt;result&gt; = lshr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +00002605</pre>
2606
2607<h5>Overview:</h5>
2608<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
Jeff Cohen5819f182007-04-22 01:17:39 +00002609operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +00002610
2611<h5>Arguments:</h5>
2612<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Nate Begemanfecbc8c2008-07-29 15:49:41 +00002613<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greif0f75ad02008-08-07 21:46:00 +00002614type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +00002615
2616<h5>Semantics:</h5>
Chris Lattnerf0e50112007-10-03 21:01:14 +00002617
Reid Spencer04e259b2007-01-31 21:39:12 +00002618<p>This instruction always performs a logical shift right operation. The most
2619significant bits of the result will be filled with zero bits after the
Gabor Greif0f75ad02008-08-07 21:46:00 +00002620shift. If <tt>op2</tt> is (statically or dynamically) equal to or larger than
Mon P Wang68d4eee2008-12-10 08:55:09 +00002621the number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
2622vectors, each vector element of <tt>op1</tt> is shifted by the corresponding shift
2623amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +00002624
2625<h5>Example:</h5>
2626<pre>
2627 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2628 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2629 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2630 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattnerf0e50112007-10-03 21:01:14 +00002631 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +00002632 &lt;result&gt; = lshr &lt;2 x i32&gt; &lt; i32 -2, i32 4&gt;, &lt; i32 1, i32 2&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0x7FFFFFFF, i32 1&gt;</i>
Reid Spencer04e259b2007-01-31 21:39:12 +00002633</pre>
2634</div>
2635
Reid Spencer2ab01932007-02-02 13:57:07 +00002636<!-- _______________________________________________________________________ -->
Reid Spencer04e259b2007-01-31 21:39:12 +00002637<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2638Instruction</a> </div>
2639<div class="doc_text">
2640
2641<h5>Syntax:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002642<pre> &lt;result&gt; = ashr &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Reid Spencer04e259b2007-01-31 21:39:12 +00002643</pre>
2644
2645<h5>Overview:</h5>
2646<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
Jeff Cohen5819f182007-04-22 01:17:39 +00002647operand shifted to the right a specified number of bits with sign extension.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +00002648
2649<h5>Arguments:</h5>
2650<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Nate Begemanfecbc8c2008-07-29 15:49:41 +00002651<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greif0f75ad02008-08-07 21:46:00 +00002652type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +00002653
2654<h5>Semantics:</h5>
2655<p>This instruction always performs an arithmetic shift right operation,
2656The most significant bits of the result will be filled with the sign bit
Gabor Greif0f75ad02008-08-07 21:46:00 +00002657of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
Mon P Wang68d4eee2008-12-10 08:55:09 +00002658larger than the number of bits in <tt>op1</tt>, the result is undefined. If the
2659arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
2660corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer04e259b2007-01-31 21:39:12 +00002661
2662<h5>Example:</h5>
2663<pre>
2664 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2665 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2666 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2667 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattnerf0e50112007-10-03 21:01:14 +00002668 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang4dd832d2008-12-09 05:46:39 +00002669 &lt;result&gt; = ashr &lt;2 x i32&gt; &lt; i32 -2, i32 4&gt;, &lt; i32 1, i32 3&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 -1, i32 0&gt;</i>
Reid Spencer04e259b2007-01-31 21:39:12 +00002670</pre>
2671</div>
2672
Chris Lattner2f7c9632001-06-06 20:29:01 +00002673<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00002674<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2675Instruction</a> </div>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002676
Misha Brukman76307852003-11-08 01:05:38 +00002677<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002678
Chris Lattner2f7c9632001-06-06 20:29:01 +00002679<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002680
2681<pre>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002682 &lt;result&gt; = and &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002683</pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002684
Chris Lattner2f7c9632001-06-06 20:29:01 +00002685<h5>Overview:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002686
Chris Lattner48b383b02003-11-25 01:02:51 +00002687<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2688its two operands.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002689
Chris Lattner2f7c9632001-06-06 20:29:01 +00002690<h5>Arguments:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002691
2692<p>The two arguments to the '<tt>and</tt>' instruction must be
2693<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2694values. Both arguments must have identical types.</p>
2695
Chris Lattner2f7c9632001-06-06 20:29:01 +00002696<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00002697<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Chris Lattner48b383b02003-11-25 01:02:51 +00002698<p> </p>
Bill Wendling5703c6e2008-09-07 10:29:20 +00002699<div>
Misha Brukman76307852003-11-08 01:05:38 +00002700<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +00002701 <tbody>
2702 <tr>
2703 <td>In0</td>
2704 <td>In1</td>
2705 <td>Out</td>
2706 </tr>
2707 <tr>
2708 <td>0</td>
2709 <td>0</td>
2710 <td>0</td>
2711 </tr>
2712 <tr>
2713 <td>0</td>
2714 <td>1</td>
2715 <td>0</td>
2716 </tr>
2717 <tr>
2718 <td>1</td>
2719 <td>0</td>
2720 <td>0</td>
2721 </tr>
2722 <tr>
2723 <td>1</td>
2724 <td>1</td>
2725 <td>1</td>
2726 </tr>
2727 </tbody>
2728</table>
Misha Brukmanc501f552004-03-01 17:47:27 +00002729</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002730<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002731<pre>
2732 &lt;result&gt; = and i32 4, %var <i>; yields {i32}:result = 4 &amp; %var</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002733 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2734 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002735</pre>
Misha Brukman76307852003-11-08 01:05:38 +00002736</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002737<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00002738<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +00002739<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +00002740<h5>Syntax:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002741<pre> &lt;result&gt; = or &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002742</pre>
Chris Lattner48b383b02003-11-25 01:02:51 +00002743<h5>Overview:</h5>
2744<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2745or of its two operands.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002746<h5>Arguments:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002747
2748<p>The two arguments to the '<tt>or</tt>' instruction must be
2749<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2750values. Both arguments must have identical types.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002751<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00002752<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Chris Lattner48b383b02003-11-25 01:02:51 +00002753<p> </p>
Bill Wendling5703c6e2008-09-07 10:29:20 +00002754<div>
Chris Lattner48b383b02003-11-25 01:02:51 +00002755<table border="1" cellspacing="0" cellpadding="4">
2756 <tbody>
2757 <tr>
2758 <td>In0</td>
2759 <td>In1</td>
2760 <td>Out</td>
2761 </tr>
2762 <tr>
2763 <td>0</td>
2764 <td>0</td>
2765 <td>0</td>
2766 </tr>
2767 <tr>
2768 <td>0</td>
2769 <td>1</td>
2770 <td>1</td>
2771 </tr>
2772 <tr>
2773 <td>1</td>
2774 <td>0</td>
2775 <td>1</td>
2776 </tr>
2777 <tr>
2778 <td>1</td>
2779 <td>1</td>
2780 <td>1</td>
2781 </tr>
2782 </tbody>
2783</table>
Misha Brukmanc501f552004-03-01 17:47:27 +00002784</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002785<h5>Example:</h5>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002786<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
2787 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
2788 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002789</pre>
Misha Brukman76307852003-11-08 01:05:38 +00002790</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002791<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00002792<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2793Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +00002794<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +00002795<h5>Syntax:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00002796<pre> &lt;result&gt; = xor &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002797</pre>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002798<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +00002799<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2800or of its two operands. The <tt>xor</tt> is used to implement the
2801"one's complement" operation, which is the "~" operator in C.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002802<h5>Arguments:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002803<p>The two arguments to the '<tt>xor</tt>' instruction must be
2804<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
2805values. Both arguments must have identical types.</p>
2806
Chris Lattner2f7c9632001-06-06 20:29:01 +00002807<h5>Semantics:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00002808
Misha Brukman76307852003-11-08 01:05:38 +00002809<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Chris Lattner48b383b02003-11-25 01:02:51 +00002810<p> </p>
Bill Wendling5703c6e2008-09-07 10:29:20 +00002811<div>
Chris Lattner48b383b02003-11-25 01:02:51 +00002812<table border="1" cellspacing="0" cellpadding="4">
2813 <tbody>
2814 <tr>
2815 <td>In0</td>
2816 <td>In1</td>
2817 <td>Out</td>
2818 </tr>
2819 <tr>
2820 <td>0</td>
2821 <td>0</td>
2822 <td>0</td>
2823 </tr>
2824 <tr>
2825 <td>0</td>
2826 <td>1</td>
2827 <td>1</td>
2828 </tr>
2829 <tr>
2830 <td>1</td>
2831 <td>0</td>
2832 <td>1</td>
2833 </tr>
2834 <tr>
2835 <td>1</td>
2836 <td>1</td>
2837 <td>0</td>
2838 </tr>
2839 </tbody>
2840</table>
Misha Brukmanc501f552004-03-01 17:47:27 +00002841</div>
Chris Lattner48b383b02003-11-25 01:02:51 +00002842<p> </p>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002843<h5>Example:</h5>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002844<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
2845 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
2846 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
2847 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00002848</pre>
Misha Brukman76307852003-11-08 01:05:38 +00002849</div>
Chris Lattner54611b42005-11-06 08:02:57 +00002850
Chris Lattner2f7c9632001-06-06 20:29:01 +00002851<!-- ======================================================================= -->
Chris Lattner54611b42005-11-06 08:02:57 +00002852<div class="doc_subsection">
Chris Lattnerce83bff2006-04-08 23:07:04 +00002853 <a name="vectorops">Vector Operations</a>
2854</div>
2855
2856<div class="doc_text">
2857
2858<p>LLVM supports several instructions to represent vector operations in a
Jeff Cohen5819f182007-04-22 01:17:39 +00002859target-independent manner. These instructions cover the element-access and
Chris Lattnerce83bff2006-04-08 23:07:04 +00002860vector-specific operations needed to process vectors effectively. While LLVM
2861does directly support these vector operations, many sophisticated algorithms
2862will want to use target-specific intrinsics to take full advantage of a specific
2863target.</p>
2864
2865</div>
2866
2867<!-- _______________________________________________________________________ -->
2868<div class="doc_subsubsection">
2869 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2870</div>
2871
2872<div class="doc_text">
2873
2874<h5>Syntax:</h5>
2875
2876<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002877 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, i32 &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +00002878</pre>
2879
2880<h5>Overview:</h5>
2881
2882<p>
2883The '<tt>extractelement</tt>' instruction extracts a single scalar
Reid Spencer404a3252007-02-15 03:07:05 +00002884element from a vector at a specified index.
Chris Lattnerce83bff2006-04-08 23:07:04 +00002885</p>
2886
2887
2888<h5>Arguments:</h5>
2889
2890<p>
2891The first operand of an '<tt>extractelement</tt>' instruction is a
Reid Spencer404a3252007-02-15 03:07:05 +00002892value of <a href="#t_vector">vector</a> type. The second operand is
Chris Lattnerce83bff2006-04-08 23:07:04 +00002893an index indicating the position from which to extract the element.
2894The index may be a variable.</p>
2895
2896<h5>Semantics:</h5>
2897
2898<p>
2899The result is a scalar of the same type as the element type of
2900<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2901<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2902results are undefined.
2903</p>
2904
2905<h5>Example:</h5>
2906
2907<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002908 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +00002909</pre>
2910</div>
2911
2912
2913<!-- _______________________________________________________________________ -->
2914<div class="doc_subsubsection">
2915 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2916</div>
2917
2918<div class="doc_text">
2919
2920<h5>Syntax:</h5>
2921
2922<pre>
Dan Gohman43ba0672008-05-12 23:38:42 +00002923 &lt;result&gt; = insertelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt;, i32 &lt;idx&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +00002924</pre>
2925
2926<h5>Overview:</h5>
2927
2928<p>
2929The '<tt>insertelement</tt>' instruction inserts a scalar
Reid Spencer404a3252007-02-15 03:07:05 +00002930element into a vector at a specified index.
Chris Lattnerce83bff2006-04-08 23:07:04 +00002931</p>
2932
2933
2934<h5>Arguments:</h5>
2935
2936<p>
2937The first operand of an '<tt>insertelement</tt>' instruction is a
Reid Spencer404a3252007-02-15 03:07:05 +00002938value of <a href="#t_vector">vector</a> type. The second operand is a
Chris Lattnerce83bff2006-04-08 23:07:04 +00002939scalar value whose type must equal the element type of the first
2940operand. The third operand is an index indicating the position at
2941which to insert the value. The index may be a variable.</p>
2942
2943<h5>Semantics:</h5>
2944
2945<p>
Reid Spencer404a3252007-02-15 03:07:05 +00002946The result is a vector of the same type as <tt>val</tt>. Its
Chris Lattnerce83bff2006-04-08 23:07:04 +00002947element values are those of <tt>val</tt> except at position
2948<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2949exceeds the length of <tt>val</tt>, the results are undefined.
2950</p>
2951
2952<h5>Example:</h5>
2953
2954<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00002955 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +00002956</pre>
2957</div>
2958
2959<!-- _______________________________________________________________________ -->
2960<div class="doc_subsubsection">
2961 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2962</div>
2963
2964<div class="doc_text">
2965
2966<h5>Syntax:</h5>
2967
2968<pre>
Mon P Wang25f01062008-11-10 04:46:22 +00002969 &lt;result&gt; = shufflevector &lt;n x &lt;ty&gt;&gt; &lt;v1&gt;, &lt;n x &lt;ty&gt;&gt; &lt;v2&gt;, &lt;m x i32&gt; &lt;mask&gt; <i>; yields &lt;m x &lt;ty&gt;&gt;</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +00002970</pre>
2971
2972<h5>Overview:</h5>
2973
2974<p>
2975The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
Mon P Wang25f01062008-11-10 04:46:22 +00002976from two input vectors, returning a vector with the same element type as
2977the input and length that is the same as the shuffle mask.
Chris Lattnerce83bff2006-04-08 23:07:04 +00002978</p>
2979
2980<h5>Arguments:</h5>
2981
2982<p>
Mon P Wang25f01062008-11-10 04:46:22 +00002983The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2984with types that match each other. The third argument is a shuffle mask whose
2985element type is always 'i32'. The result of the instruction is a vector whose
2986length is the same as the shuffle mask and whose element type is the same as
2987the element type of the first two operands.
Chris Lattnerce83bff2006-04-08 23:07:04 +00002988</p>
2989
2990<p>
2991The shuffle mask operand is required to be a constant vector with either
2992constant integer or undef values.
2993</p>
2994
2995<h5>Semantics:</h5>
2996
2997<p>
2998The elements of the two input vectors are numbered from left to right across
2999both of the vectors. The shuffle mask operand specifies, for each element of
Mon P Wang25f01062008-11-10 04:46:22 +00003000the result vector, which element of the two input vectors the result element
Chris Lattnerce83bff2006-04-08 23:07:04 +00003001gets. The element selector may be undef (meaning "don't care") and the second
3002operand may be undef if performing a shuffle from only one vector.
3003</p>
3004
3005<h5>Example:</h5>
3006
3007<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003008 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Jeff Cohen5819f182007-04-22 01:17:39 +00003009 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003010 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
3011 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i> - Identity shuffle.
Mon P Wang25f01062008-11-10 04:46:22 +00003012 %result = shufflevector &lt;8 x i32&gt; %v1, &lt;8 x i32&gt; undef,
3013 &lt;4 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3&gt; <i>; yields &lt;4 x i32&gt;</i>
3014 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
3015 &lt;8 x i32&gt; &lt;i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 &gt; <i>; yields &lt;8 x i32&gt;</i>
Chris Lattnerce83bff2006-04-08 23:07:04 +00003016</pre>
3017</div>
3018
Tanya Lattnerb138bbe2006-04-14 19:24:33 +00003019
Chris Lattnerce83bff2006-04-08 23:07:04 +00003020<!-- ======================================================================= -->
3021<div class="doc_subsection">
Dan Gohmanb9d66602008-05-12 23:51:09 +00003022 <a name="aggregateops">Aggregate Operations</a>
3023</div>
3024
3025<div class="doc_text">
3026
3027<p>LLVM supports several instructions for working with aggregate values.
3028</p>
3029
3030</div>
3031
3032<!-- _______________________________________________________________________ -->
3033<div class="doc_subsubsection">
3034 <a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
3035</div>
3036
3037<div class="doc_text">
3038
3039<h5>Syntax:</h5>
3040
3041<pre>
3042 &lt;result&gt; = extractvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;idx&gt;{, &lt;idx&gt;}*
3043</pre>
3044
3045<h5>Overview:</h5>
3046
3047<p>
Dan Gohman35a835c2008-05-13 18:16:06 +00003048The '<tt>extractvalue</tt>' instruction extracts the value of a struct field
3049or array element from an aggregate value.
Dan Gohmanb9d66602008-05-12 23:51:09 +00003050</p>
3051
3052
3053<h5>Arguments:</h5>
3054
3055<p>
3056The first operand of an '<tt>extractvalue</tt>' instruction is a
3057value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a>
Dan Gohman35a835c2008-05-13 18:16:06 +00003058type. The operands are constant indices to specify which value to extract
Dan Gohman1ecaf452008-05-31 00:58:22 +00003059in a similar manner as indices in a
Dan Gohmanb9d66602008-05-12 23:51:09 +00003060'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3061</p>
3062
3063<h5>Semantics:</h5>
3064
3065<p>
3066The result is the value at the position in the aggregate specified by
3067the index operands.
3068</p>
3069
3070<h5>Example:</h5>
3071
3072<pre>
Dan Gohman1ecaf452008-05-31 00:58:22 +00003073 %result = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +00003074</pre>
3075</div>
3076
3077
3078<!-- _______________________________________________________________________ -->
3079<div class="doc_subsubsection">
3080 <a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
3081</div>
3082
3083<div class="doc_text">
3084
3085<h5>Syntax:</h5>
3086
3087<pre>
Dan Gohman1ecaf452008-05-31 00:58:22 +00003088 &lt;result&gt; = insertvalue &lt;aggregate type&gt; &lt;val&gt;, &lt;ty&gt; &lt;val&gt;, &lt;idx&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +00003089</pre>
3090
3091<h5>Overview:</h5>
3092
3093<p>
3094The '<tt>insertvalue</tt>' instruction inserts a value
Dan Gohman35a835c2008-05-13 18:16:06 +00003095into a struct field or array element in an aggregate.
Dan Gohmanb9d66602008-05-12 23:51:09 +00003096</p>
3097
3098
3099<h5>Arguments:</h5>
3100
3101<p>
3102The first operand of an '<tt>insertvalue</tt>' instruction is a
3103value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a> type.
3104The second operand is a first-class value to insert.
Dan Gohman34d1c0d2008-05-23 21:53:15 +00003105The following operands are constant indices
Dan Gohman1ecaf452008-05-31 00:58:22 +00003106indicating the position at which to insert the value in a similar manner as
Dan Gohman35a835c2008-05-13 18:16:06 +00003107indices in a
Dan Gohmanb9d66602008-05-12 23:51:09 +00003108'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
3109The value to insert must have the same type as the value identified
Dan Gohman35a835c2008-05-13 18:16:06 +00003110by the indices.
Dan Gohmanef9462f2008-10-14 16:51:45 +00003111</p>
Dan Gohmanb9d66602008-05-12 23:51:09 +00003112
3113<h5>Semantics:</h5>
3114
3115<p>
3116The result is an aggregate of the same type as <tt>val</tt>. Its
3117value is that of <tt>val</tt> except that the value at the position
Dan Gohman35a835c2008-05-13 18:16:06 +00003118specified by the indices is that of <tt>elt</tt>.
Dan Gohmanb9d66602008-05-12 23:51:09 +00003119</p>
3120
3121<h5>Example:</h5>
3122
3123<pre>
Dan Gohman88ce1a52008-06-23 15:26:37 +00003124 %result = insertvalue {i32, float} %agg, i32 1, 0 <i>; yields {i32, float}</i>
Dan Gohmanb9d66602008-05-12 23:51:09 +00003125</pre>
3126</div>
3127
3128
3129<!-- ======================================================================= -->
3130<div class="doc_subsection">
Chris Lattner6ab66722006-08-15 00:45:58 +00003131 <a name="memoryops">Memory Access and Addressing Operations</a>
Chris Lattner54611b42005-11-06 08:02:57 +00003132</div>
3133
Misha Brukman76307852003-11-08 01:05:38 +00003134<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +00003135
Chris Lattner48b383b02003-11-25 01:02:51 +00003136<p>A key design point of an SSA-based representation is how it
3137represents memory. In LLVM, no memory locations are in SSA form, which
3138makes things very simple. This section describes how to read, write,
John Criswelldfe6a862004-12-10 15:51:16 +00003139allocate, and free memory in LLVM.</p>
Chris Lattner54611b42005-11-06 08:02:57 +00003140
Misha Brukman76307852003-11-08 01:05:38 +00003141</div>
Chris Lattner54611b42005-11-06 08:02:57 +00003142
Chris Lattner2f7c9632001-06-06 20:29:01 +00003143<!-- _______________________________________________________________________ -->
Chris Lattner54611b42005-11-06 08:02:57 +00003144<div class="doc_subsubsection">
3145 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
3146</div>
3147
Misha Brukman76307852003-11-08 01:05:38 +00003148<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +00003149
Chris Lattner2f7c9632001-06-06 20:29:01 +00003150<h5>Syntax:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00003151
3152<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003153 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00003154</pre>
Chris Lattner54611b42005-11-06 08:02:57 +00003155
Chris Lattner2f7c9632001-06-06 20:29:01 +00003156<h5>Overview:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00003157
Chris Lattner48b383b02003-11-25 01:02:51 +00003158<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lamb55c6d4f2007-12-17 01:00:21 +00003159heap and returns a pointer to it. The object is always allocated in the generic
3160address space (address space zero).</p>
Chris Lattner54611b42005-11-06 08:02:57 +00003161
Chris Lattner2f7c9632001-06-06 20:29:01 +00003162<h5>Arguments:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00003163
3164<p>The '<tt>malloc</tt>' instruction allocates
3165<tt>sizeof(&lt;type&gt;)*NumElements</tt>
John Criswella92e5862004-02-24 16:13:56 +00003166bytes of memory from the operating system and returns a pointer of the
Chris Lattner54611b42005-11-06 08:02:57 +00003167appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greifdd1fc982008-02-09 22:24:34 +00003168number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner1f17cce2008-04-02 00:38:26 +00003169If a constant alignment is specified, the value result of the allocation is guaranteed to
Gabor Greifdd1fc982008-02-09 22:24:34 +00003170be aligned to at least that boundary. If not specified, or if zero, the target can
3171choose to align the allocation on any convenient boundary.</p>
Chris Lattner54611b42005-11-06 08:02:57 +00003172
Misha Brukman76307852003-11-08 01:05:38 +00003173<p>'<tt>type</tt>' must be a sized type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +00003174
Chris Lattner2f7c9632001-06-06 20:29:01 +00003175<h5>Semantics:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00003176
Chris Lattner48b383b02003-11-25 01:02:51 +00003177<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
Nick Lewyckyf5ffcbc2008-11-24 03:41:24 +00003178a pointer is returned. The result of a zero byte allocation is undefined. The
Chris Lattnerfe8519c2008-04-19 21:01:16 +00003179result is null if there is insufficient memory available.</p>
Misha Brukman76307852003-11-08 01:05:38 +00003180
Chris Lattner54611b42005-11-06 08:02:57 +00003181<h5>Example:</h5>
3182
3183<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003184 %array = malloc [4 x i8 ] <i>; yields {[%4 x i8]*}:array</i>
Chris Lattner54611b42005-11-06 08:02:57 +00003185
Bill Wendling2d8b9a82007-05-29 09:42:13 +00003186 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
3187 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
3188 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
3189 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
3190 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00003191</pre>
Misha Brukman76307852003-11-08 01:05:38 +00003192</div>
Chris Lattner54611b42005-11-06 08:02:57 +00003193
Chris Lattner2f7c9632001-06-06 20:29:01 +00003194<!-- _______________________________________________________________________ -->
Chris Lattner54611b42005-11-06 08:02:57 +00003195<div class="doc_subsubsection">
3196 <a name="i_free">'<tt>free</tt>' Instruction</a>
3197</div>
3198
Misha Brukman76307852003-11-08 01:05:38 +00003199<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +00003200
Chris Lattner2f7c9632001-06-06 20:29:01 +00003201<h5>Syntax:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00003202
3203<pre>
3204 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00003205</pre>
Chris Lattner54611b42005-11-06 08:02:57 +00003206
Chris Lattner2f7c9632001-06-06 20:29:01 +00003207<h5>Overview:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00003208
Chris Lattner48b383b02003-11-25 01:02:51 +00003209<p>The '<tt>free</tt>' instruction returns memory back to the unused
John Criswell4a3327e2005-05-13 22:25:59 +00003210memory heap to be reallocated in the future.</p>
Chris Lattner54611b42005-11-06 08:02:57 +00003211
Chris Lattner2f7c9632001-06-06 20:29:01 +00003212<h5>Arguments:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00003213
Chris Lattner48b383b02003-11-25 01:02:51 +00003214<p>'<tt>value</tt>' shall be a pointer value that points to a value
3215that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
3216instruction.</p>
Chris Lattner54611b42005-11-06 08:02:57 +00003217
Chris Lattner2f7c9632001-06-06 20:29:01 +00003218<h5>Semantics:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00003219
John Criswelldfe6a862004-12-10 15:51:16 +00003220<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner0f103e12008-04-19 22:41:32 +00003221after this instruction executes. If the pointer is null, the operation
3222is a noop.</p>
Chris Lattner54611b42005-11-06 08:02:57 +00003223
Chris Lattner2f7c9632001-06-06 20:29:01 +00003224<h5>Example:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00003225
3226<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003227 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
3228 free [4 x i8]* %array
Chris Lattner2f7c9632001-06-06 20:29:01 +00003229</pre>
Misha Brukman76307852003-11-08 01:05:38 +00003230</div>
Chris Lattner54611b42005-11-06 08:02:57 +00003231
Chris Lattner2f7c9632001-06-06 20:29:01 +00003232<!-- _______________________________________________________________________ -->
Chris Lattner54611b42005-11-06 08:02:57 +00003233<div class="doc_subsubsection">
3234 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
3235</div>
3236
Misha Brukman76307852003-11-08 01:05:38 +00003237<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +00003238
Chris Lattner2f7c9632001-06-06 20:29:01 +00003239<h5>Syntax:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00003240
3241<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003242 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00003243</pre>
Chris Lattner54611b42005-11-06 08:02:57 +00003244
Chris Lattner2f7c9632001-06-06 20:29:01 +00003245<h5>Overview:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00003246
Jeff Cohen5819f182007-04-22 01:17:39 +00003247<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
3248currently executing function, to be automatically released when this function
Christopher Lamb55c6d4f2007-12-17 01:00:21 +00003249returns to its caller. The object is always allocated in the generic address
3250space (address space zero).</p>
Chris Lattner54611b42005-11-06 08:02:57 +00003251
Chris Lattner2f7c9632001-06-06 20:29:01 +00003252<h5>Arguments:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00003253
John Criswelldfe6a862004-12-10 15:51:16 +00003254<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
Chris Lattner48b383b02003-11-25 01:02:51 +00003255bytes of memory on the runtime stack, returning a pointer of the
Gabor Greifdd1fc982008-02-09 22:24:34 +00003256appropriate type to the program. If "NumElements" is specified, it is the
3257number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner1f17cce2008-04-02 00:38:26 +00003258If a constant alignment is specified, the value result of the allocation is guaranteed
Gabor Greifdd1fc982008-02-09 22:24:34 +00003259to be aligned to at least that boundary. If not specified, or if zero, the target
3260can choose to align the allocation on any convenient boundary.</p>
Chris Lattner54611b42005-11-06 08:02:57 +00003261
Misha Brukman76307852003-11-08 01:05:38 +00003262<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +00003263
Chris Lattner2f7c9632001-06-06 20:29:01 +00003264<h5>Semantics:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00003265
Chris Lattnerfe8519c2008-04-19 21:01:16 +00003266<p>Memory is allocated; a pointer is returned. The operation is undefiend if
3267there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
Chris Lattner48b383b02003-11-25 01:02:51 +00003268memory is automatically released when the function returns. The '<tt>alloca</tt>'
3269instruction is commonly used to represent automatic variables that must
3270have an address available. When the function returns (either with the <tt><a
John Criswellc932bef2005-05-12 16:55:34 +00003271 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
Chris Lattner1f17cce2008-04-02 00:38:26 +00003272instructions), the memory is reclaimed. Allocating zero bytes
3273is legal, but the result is undefined.</p>
Chris Lattner54611b42005-11-06 08:02:57 +00003274
Chris Lattner2f7c9632001-06-06 20:29:01 +00003275<h5>Example:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +00003276
3277<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003278 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
Bill Wendling3716c5d2007-05-29 09:04:49 +00003279 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
3280 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003281 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +00003282</pre>
Misha Brukman76307852003-11-08 01:05:38 +00003283</div>
Chris Lattner54611b42005-11-06 08:02:57 +00003284
Chris Lattner2f7c9632001-06-06 20:29:01 +00003285<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00003286<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
3287Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +00003288<div class="doc_text">
Chris Lattner095735d2002-05-06 03:03:22 +00003289<h5>Syntax:</h5>
Christopher Lambbff50202007-04-21 08:16:25 +00003290<pre> &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;]<br> &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;]<br></pre>
Chris Lattner095735d2002-05-06 03:03:22 +00003291<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00003292<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Chris Lattner095735d2002-05-06 03:03:22 +00003293<h5>Arguments:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +00003294<p>The argument to the '<tt>load</tt>' instruction specifies the memory
John Criswell4c0cf7f2005-10-24 16:17:18 +00003295address from which to load. The pointer must point to a <a
Chris Lattner10ee9652004-06-03 22:57:15 +00003296 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
John Criswell4c0cf7f2005-10-24 16:17:18 +00003297marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
Chris Lattner48b383b02003-11-25 01:02:51 +00003298the number or order of execution of this <tt>load</tt> with other
3299volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
3300instructions. </p>
Chris Lattner2a1993f2008-01-06 21:04:43 +00003301<p>
Chris Lattner1f17cce2008-04-02 00:38:26 +00003302The optional constant "align" argument specifies the alignment of the operation
Chris Lattner2a1993f2008-01-06 21:04:43 +00003303(that is, the alignment of the memory address). A value of 0 or an
3304omitted "align" argument means that the operation has the preferential
3305alignment for the target. It is the responsibility of the code emitter
3306to ensure that the alignment information is correct. Overestimating
3307the alignment results in an undefined behavior. Underestimating the
3308alignment may produce less efficient code. An alignment of 1 is always
3309safe.
3310</p>
Chris Lattner095735d2002-05-06 03:03:22 +00003311<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00003312<p>The location of memory pointed to is loaded.</p>
Chris Lattner095735d2002-05-06 03:03:22 +00003313<h5>Examples:</h5>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003314<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Chris Lattner48b383b02003-11-25 01:02:51 +00003315 <a
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003316 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
3317 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner095735d2002-05-06 03:03:22 +00003318</pre>
Misha Brukman76307852003-11-08 01:05:38 +00003319</div>
Chris Lattner095735d2002-05-06 03:03:22 +00003320<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +00003321<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
3322Instruction</a> </div>
Reid Spencera89fb182006-11-09 21:18:01 +00003323<div class="doc_text">
Chris Lattner095735d2002-05-06 03:03:22 +00003324<h5>Syntax:</h5>
Christopher Lambbff50202007-04-21 08:16:25 +00003325<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
3326 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
Chris Lattner095735d2002-05-06 03:03:22 +00003327</pre>
Chris Lattner095735d2002-05-06 03:03:22 +00003328<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +00003329<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Chris Lattner095735d2002-05-06 03:03:22 +00003330<h5>Arguments:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +00003331<p>There are two arguments to the '<tt>store</tt>' instruction: a value
Jeff Cohen5819f182007-04-22 01:17:39 +00003332to store and an address at which to store it. The type of the '<tt>&lt;pointer&gt;</tt>'
Chris Lattner1f17cce2008-04-02 00:38:26 +00003333operand must be a pointer to the <a href="#t_firstclass">first class</a> type
3334of the '<tt>&lt;value&gt;</tt>'
John Criswell4a3327e2005-05-13 22:25:59 +00003335operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
Chris Lattner48b383b02003-11-25 01:02:51 +00003336optimizer is not allowed to modify the number or order of execution of
3337this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
3338 href="#i_store">store</a></tt> instructions.</p>
Chris Lattner2a1993f2008-01-06 21:04:43 +00003339<p>
Chris Lattner1f17cce2008-04-02 00:38:26 +00003340The optional constant "align" argument specifies the alignment of the operation
Chris Lattner2a1993f2008-01-06 21:04:43 +00003341(that is, the alignment of the memory address). A value of 0 or an
3342omitted "align" argument means that the operation has the preferential
3343alignment for the target. It is the responsibility of the code emitter
3344to ensure that the alignment information is correct. Overestimating
3345the alignment results in an undefined behavior. Underestimating the
3346alignment may produce less efficient code. An alignment of 1 is always
3347safe.
3348</p>
Chris Lattner48b383b02003-11-25 01:02:51 +00003349<h5>Semantics:</h5>
3350<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
3351at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
Chris Lattner095735d2002-05-06 03:03:22 +00003352<h5>Example:</h5>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003353<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8830ffe2007-10-22 05:10:05 +00003354 store i32 3, i32* %ptr <i>; yields {void}</i>
3355 %val = <a href="#i_load">load</a> i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner095735d2002-05-06 03:03:22 +00003356</pre>
Reid Spencer443460a2006-11-09 21:15:49 +00003357</div>
3358
Chris Lattner095735d2002-05-06 03:03:22 +00003359<!-- _______________________________________________________________________ -->
Chris Lattner33fd7022004-04-05 01:30:49 +00003360<div class="doc_subsubsection">
3361 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
3362</div>
3363
Misha Brukman76307852003-11-08 01:05:38 +00003364<div class="doc_text">
Chris Lattner590645f2002-04-14 06:13:44 +00003365<h5>Syntax:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +00003366<pre>
Matthijs Kooijman0e268272008-10-13 13:44:15 +00003367 &lt;result&gt; = getelementptr &lt;pty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
Chris Lattner33fd7022004-04-05 01:30:49 +00003368</pre>
3369
Chris Lattner590645f2002-04-14 06:13:44 +00003370<h5>Overview:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +00003371
3372<p>
3373The '<tt>getelementptr</tt>' instruction is used to get the address of a
Matthijs Kooijman0e268272008-10-13 13:44:15 +00003374subelement of an aggregate data structure. It performs address calculation only
3375and does not access memory.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +00003376
Chris Lattner590645f2002-04-14 06:13:44 +00003377<h5>Arguments:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +00003378
Matthijs Kooijman0e268272008-10-13 13:44:15 +00003379<p>The first argument is always a pointer, and forms the basis of the
3380calculation. The remaining arguments are indices, that indicate which of the
3381elements of the aggregate object are indexed. The interpretation of each index
3382is dependent on the type being indexed into. The first index always indexes the
3383pointer value given as the first argument, the second index indexes a value of
3384the type pointed to (not necessarily the value directly pointed to, since the
3385first index can be non-zero), etc. The first type indexed into must be a pointer
3386value, subsequent types can be arrays, vectors and structs. Note that subsequent
3387types being indexed into can never be pointers, since that would require loading
3388the pointer before continuing calculation.</p>
3389
3390<p>The type of each index argument depends on the type it is indexing into.
3391When indexing into a (packed) structure, only <tt>i32</tt> integer
3392<b>constants</b> are allowed. When indexing into an array, pointer or vector,
3393only integers of 32 or 64 bits are allowed (also non-constants). 32-bit values
3394will be sign extended to 64-bits if required.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +00003395
Chris Lattner48b383b02003-11-25 01:02:51 +00003396<p>For example, let's consider a C code fragment and how it gets
3397compiled to LLVM:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +00003398
Bill Wendling3716c5d2007-05-29 09:04:49 +00003399<div class="doc_code">
Chris Lattner33fd7022004-04-05 01:30:49 +00003400<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +00003401struct RT {
3402 char A;
Chris Lattnera446f1b2007-05-29 15:43:56 +00003403 int B[10][20];
Bill Wendling3716c5d2007-05-29 09:04:49 +00003404 char C;
3405};
3406struct ST {
Chris Lattnera446f1b2007-05-29 15:43:56 +00003407 int X;
Bill Wendling3716c5d2007-05-29 09:04:49 +00003408 double Y;
3409 struct RT Z;
3410};
Chris Lattner33fd7022004-04-05 01:30:49 +00003411
Chris Lattnera446f1b2007-05-29 15:43:56 +00003412int *foo(struct ST *s) {
Bill Wendling3716c5d2007-05-29 09:04:49 +00003413 return &amp;s[1].Z.B[5][13];
3414}
Chris Lattner33fd7022004-04-05 01:30:49 +00003415</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +00003416</div>
Chris Lattner33fd7022004-04-05 01:30:49 +00003417
Misha Brukman76307852003-11-08 01:05:38 +00003418<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +00003419
Bill Wendling3716c5d2007-05-29 09:04:49 +00003420<div class="doc_code">
Chris Lattner33fd7022004-04-05 01:30:49 +00003421<pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +00003422%RT = type { i8 , [10 x [20 x i32]], i8 }
3423%ST = type { i32, double, %RT }
Chris Lattner33fd7022004-04-05 01:30:49 +00003424
Bill Wendling3716c5d2007-05-29 09:04:49 +00003425define i32* %foo(%ST* %s) {
3426entry:
3427 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
3428 ret i32* %reg
3429}
Chris Lattner33fd7022004-04-05 01:30:49 +00003430</pre>
Bill Wendling3716c5d2007-05-29 09:04:49 +00003431</div>
Chris Lattner33fd7022004-04-05 01:30:49 +00003432
Chris Lattner590645f2002-04-14 06:13:44 +00003433<h5>Semantics:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +00003434
Misha Brukman76307852003-11-08 01:05:38 +00003435<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003436type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
Chris Lattner33fd7022004-04-05 01:30:49 +00003437}</tt>' type, a structure. The second index indexes into the third element of
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003438the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
3439i8 }</tt>' type, another structure. The third index indexes into the second
3440element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
Chris Lattner33fd7022004-04-05 01:30:49 +00003441array. The two dimensions of the array are subscripted into, yielding an
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003442'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
3443to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +00003444
Chris Lattner48b383b02003-11-25 01:02:51 +00003445<p>Note that it is perfectly legal to index partially through a
3446structure, returning a pointer to an inner element. Because of this,
3447the LLVM code for the given testcase is equivalent to:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +00003448
3449<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003450 define i32* %foo(%ST* %s) {
3451 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen5819f182007-04-22 01:17:39 +00003452 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
3453 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003454 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
3455 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
3456 ret i32* %t5
Chris Lattner33fd7022004-04-05 01:30:49 +00003457 }
Chris Lattnera8292f32002-05-06 22:08:29 +00003458</pre>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +00003459
3460<p>Note that it is undefined to access an array out of bounds: array and
3461pointer indexes must always be within the defined bounds of the array type.
Chris Lattner851b7712008-04-24 05:59:56 +00003462The one exception for this rule is zero length arrays. These arrays are
Chris Lattnerc0ad71e2005-06-24 17:22:57 +00003463defined to be accessible as variable length arrays, which requires access
3464beyond the zero'th element.</p>
3465
Chris Lattner6ab66722006-08-15 00:45:58 +00003466<p>The getelementptr instruction is often confusing. For some more insight
3467into how it works, see <a href="GetElementPtr.html">the getelementptr
3468FAQ</a>.</p>
3469
Chris Lattner590645f2002-04-14 06:13:44 +00003470<h5>Example:</h5>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +00003471
Chris Lattner33fd7022004-04-05 01:30:49 +00003472<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003473 <i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman0e268272008-10-13 13:44:15 +00003474 %aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
3475 <i>; yields i8*:vptr</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +00003476 %vptr = getelementptr {i32, &lt;2 x i8&gt;}* %svptr, i64 0, i32 1, i32 1
Matthijs Kooijman0e268272008-10-13 13:44:15 +00003477 <i>; yields i8*:eptr</i>
3478 %eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Chris Lattner33fd7022004-04-05 01:30:49 +00003479</pre>
Chris Lattner33fd7022004-04-05 01:30:49 +00003480</div>
Reid Spencer443460a2006-11-09 21:15:49 +00003481
Chris Lattner2f7c9632001-06-06 20:29:01 +00003482<!-- ======================================================================= -->
Reid Spencer97c5fa42006-11-08 01:18:52 +00003483<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
Misha Brukman76307852003-11-08 01:05:38 +00003484</div>
Misha Brukman76307852003-11-08 01:05:38 +00003485<div class="doc_text">
Reid Spencer97c5fa42006-11-08 01:18:52 +00003486<p>The instructions in this category are the conversion instructions (casting)
3487which all take a single operand and a type. They perform various bit conversions
3488on the operand.</p>
Misha Brukman76307852003-11-08 01:05:38 +00003489</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00003490
Chris Lattnera8292f32002-05-06 22:08:29 +00003491<!-- _______________________________________________________________________ -->
Chris Lattnerb53c28d2004-03-12 05:50:16 +00003492<div class="doc_subsubsection">
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003493 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
3494</div>
3495<div class="doc_text">
3496
3497<h5>Syntax:</h5>
3498<pre>
3499 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3500</pre>
3501
3502<h5>Overview:</h5>
3503<p>
3504The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
3505</p>
3506
3507<h5>Arguments:</h5>
3508<p>
3509The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
3510be an <a href="#t_integer">integer</a> type, and a type that specifies the size
Chris Lattnerc0f423a2007-01-15 01:54:13 +00003511and type of the result, which must be an <a href="#t_integer">integer</a>
Reid Spencer51b07252006-11-09 23:03:26 +00003512type. The bit size of <tt>value</tt> must be larger than the bit size of
3513<tt>ty2</tt>. Equal sized types are not allowed.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003514
3515<h5>Semantics:</h5>
3516<p>
3517The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
Reid Spencer51b07252006-11-09 23:03:26 +00003518and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
3519larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
3520It will always truncate bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003521
3522<h5>Example:</h5>
3523<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003524 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
Reid Spencer36a15422007-01-12 03:35:51 +00003525 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
3526 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003527</pre>
3528</div>
3529
3530<!-- _______________________________________________________________________ -->
3531<div class="doc_subsubsection">
3532 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
3533</div>
3534<div class="doc_text">
3535
3536<h5>Syntax:</h5>
3537<pre>
3538 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3539</pre>
3540
3541<h5>Overview:</h5>
3542<p>The '<tt>zext</tt>' instruction zero extends its operand to type
3543<tt>ty2</tt>.</p>
3544
3545
3546<h5>Arguments:</h5>
3547<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
Chris Lattnerc0f423a2007-01-15 01:54:13 +00003548<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3549also be of <a href="#t_integer">integer</a> type. The bit size of the
Reid Spencer51b07252006-11-09 23:03:26 +00003550<tt>value</tt> must be smaller than the bit size of the destination type,
3551<tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003552
3553<h5>Semantics:</h5>
3554<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
Chris Lattnerc87f3df2007-05-24 19:13:27 +00003555bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003556
Reid Spencer07c9c682007-01-12 15:46:11 +00003557<p>When zero extending from i1, the result will always be either 0 or 1.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003558
3559<h5>Example:</h5>
3560<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003561 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencer36a15422007-01-12 03:35:51 +00003562 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003563</pre>
3564</div>
3565
3566<!-- _______________________________________________________________________ -->
3567<div class="doc_subsubsection">
3568 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3569</div>
3570<div class="doc_text">
3571
3572<h5>Syntax:</h5>
3573<pre>
3574 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3575</pre>
3576
3577<h5>Overview:</h5>
3578<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3579
3580<h5>Arguments:</h5>
3581<p>
3582The '<tt>sext</tt>' instruction takes a value to cast, which must be of
Chris Lattnerc0f423a2007-01-15 01:54:13 +00003583<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3584also be of <a href="#t_integer">integer</a> type. The bit size of the
Reid Spencer51b07252006-11-09 23:03:26 +00003585<tt>value</tt> must be smaller than the bit size of the destination type,
3586<tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003587
3588<h5>Semantics:</h5>
3589<p>
3590The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3591bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
Chris Lattnerc87f3df2007-05-24 19:13:27 +00003592the type <tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003593
Reid Spencer36a15422007-01-12 03:35:51 +00003594<p>When sign extending from i1, the extension always results in -1 or 0.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003595
3596<h5>Example:</h5>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003597<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003598 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencer36a15422007-01-12 03:35:51 +00003599 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003600</pre>
3601</div>
3602
3603<!-- _______________________________________________________________________ -->
3604<div class="doc_subsubsection">
Reid Spencer2e2740d2006-11-09 21:48:10 +00003605 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3606</div>
3607
3608<div class="doc_text">
3609
3610<h5>Syntax:</h5>
3611
3612<pre>
3613 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3614</pre>
3615
3616<h5>Overview:</h5>
3617<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3618<tt>ty2</tt>.</p>
3619
3620
3621<h5>Arguments:</h5>
3622<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
3623 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
3624cast it to. The size of <tt>value</tt> must be larger than the size of
3625<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
3626<i>no-op cast</i>.</p>
3627
3628<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +00003629<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
3630<a href="#t_floating">floating point</a> type to a smaller
3631<a href="#t_floating">floating point</a> type. If the value cannot fit within
3632the destination type, <tt>ty2</tt>, then the results are undefined.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +00003633
3634<h5>Example:</h5>
3635<pre>
3636 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
3637 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
3638</pre>
3639</div>
3640
3641<!-- _______________________________________________________________________ -->
3642<div class="doc_subsubsection">
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003643 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
3644</div>
3645<div class="doc_text">
3646
3647<h5>Syntax:</h5>
3648<pre>
3649 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3650</pre>
3651
3652<h5>Overview:</h5>
3653<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
3654floating point value.</p>
3655
3656<h5>Arguments:</h5>
3657<p>The '<tt>fpext</tt>' instruction takes a
3658<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
Reid Spencer51b07252006-11-09 23:03:26 +00003659and a <a href="#t_floating">floating point</a> type to cast it to. The source
3660type must be smaller than the destination type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003661
3662<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +00003663<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Duncan Sands16f122e2007-03-30 12:22:09 +00003664<a href="#t_floating">floating point</a> type to a larger
3665<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
Reid Spencer51b07252006-11-09 23:03:26 +00003666used to make a <i>no-op cast</i> because it always changes bits. Use
Reid Spencer5b950642006-11-11 23:08:07 +00003667<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003668
3669<h5>Example:</h5>
3670<pre>
3671 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
3672 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
3673</pre>
3674</div>
3675
3676<!-- _______________________________________________________________________ -->
3677<div class="doc_subsubsection">
Reid Spencer2eadb532007-01-21 00:29:26 +00003678 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003679</div>
3680<div class="doc_text">
3681
3682<h5>Syntax:</h5>
3683<pre>
Reid Spencer753163d2007-07-31 14:40:14 +00003684 &lt;result&gt; = fptoui &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003685</pre>
3686
3687<h5>Overview:</h5>
Reid Spencer753163d2007-07-31 14:40:14 +00003688<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003689unsigned integer equivalent of type <tt>ty2</tt>.
3690</p>
3691
3692<h5>Arguments:</h5>
Reid Spencer753163d2007-07-31 14:40:14 +00003693<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
Nate Begemand4d45c22007-11-17 03:58:34 +00003694scalar or vector <a href="#t_floating">floating point</a> value, and a type
3695to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3696type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3697vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003698
3699<h5>Semantics:</h5>
Reid Spencer753163d2007-07-31 14:40:14 +00003700<p> The '<tt>fptoui</tt>' instruction converts its
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003701<a href="#t_floating">floating point</a> operand into the nearest (rounding
3702towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
3703the results are undefined.</p>
3704
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003705<h5>Example:</h5>
3706<pre>
Reid Spencer753163d2007-07-31 14:40:14 +00003707 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +00003708 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencer753163d2007-07-31 14:40:14 +00003709 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003710</pre>
3711</div>
3712
3713<!-- _______________________________________________________________________ -->
3714<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +00003715 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003716</div>
3717<div class="doc_text">
3718
3719<h5>Syntax:</h5>
3720<pre>
Reid Spencer51b07252006-11-09 23:03:26 +00003721 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003722</pre>
3723
3724<h5>Overview:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +00003725<p>The '<tt>fptosi</tt>' instruction converts
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003726<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
Chris Lattnerb53c28d2004-03-12 05:50:16 +00003727</p>
3728
Chris Lattnera8292f32002-05-06 22:08:29 +00003729<h5>Arguments:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +00003730<p> The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
Nate Begemand4d45c22007-11-17 03:58:34 +00003731scalar or vector <a href="#t_floating">floating point</a> value, and a type
3732to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
3733type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
3734vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00003735
Chris Lattnera8292f32002-05-06 22:08:29 +00003736<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +00003737<p>The '<tt>fptosi</tt>' instruction converts its
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003738<a href="#t_floating">floating point</a> operand into the nearest (rounding
3739towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3740the results are undefined.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00003741
Chris Lattner70de6632001-07-09 00:26:23 +00003742<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00003743<pre>
Reid Spencer36a15422007-01-12 03:35:51 +00003744 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner5b95a172007-09-22 03:17:52 +00003745 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003746 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003747</pre>
3748</div>
3749
3750<!-- _______________________________________________________________________ -->
3751<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +00003752 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003753</div>
3754<div class="doc_text">
3755
3756<h5>Syntax:</h5>
3757<pre>
Reid Spencer51b07252006-11-09 23:03:26 +00003758 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003759</pre>
3760
3761<h5>Overview:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +00003762<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003763integer and converts that value to the <tt>ty2</tt> type.</p>
3764
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003765<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +00003766<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
3767scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3768to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3769type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3770floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003771
3772<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +00003773<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003774integer quantity and converts it to the corresponding floating point value. If
Jeff Cohenbeccb742007-04-22 14:56:37 +00003775the value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003776
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003777<h5>Example:</h5>
3778<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003779 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +00003780 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003781</pre>
3782</div>
3783
3784<!-- _______________________________________________________________________ -->
3785<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +00003786 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003787</div>
3788<div class="doc_text">
3789
3790<h5>Syntax:</h5>
3791<pre>
Reid Spencer51b07252006-11-09 23:03:26 +00003792 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003793</pre>
3794
3795<h5>Overview:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +00003796<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003797integer and converts that value to the <tt>ty2</tt> type.</p>
3798
3799<h5>Arguments:</h5>
Nate Begemand4d45c22007-11-17 03:58:34 +00003800<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
3801scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
3802to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
3803type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
3804floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003805
3806<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +00003807<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003808integer quantity and converts it to the corresponding floating point value. If
Jeff Cohenbeccb742007-04-22 14:56:37 +00003809the value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003810
3811<h5>Example:</h5>
3812<pre>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003813 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +00003814 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003815</pre>
3816</div>
3817
3818<!-- _______________________________________________________________________ -->
3819<div class="doc_subsubsection">
Reid Spencerb7344ff2006-11-11 21:00:47 +00003820 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
3821</div>
3822<div class="doc_text">
3823
3824<h5>Syntax:</h5>
3825<pre>
3826 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3827</pre>
3828
3829<h5>Overview:</h5>
3830<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
3831the integer type <tt>ty2</tt>.</p>
3832
3833<h5>Arguments:</h5>
3834<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
Duncan Sands16f122e2007-03-30 12:22:09 +00003835must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
Dan Gohmanef9462f2008-10-14 16:51:45 +00003836<tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +00003837
3838<h5>Semantics:</h5>
3839<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
3840<tt>ty2</tt> by interpreting the pointer value as an integer and either
3841truncating or zero extending that value to the size of the integer type. If
3842<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
3843<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
Jeff Cohen222a8a42007-04-29 01:07:00 +00003844are the same size, then nothing is done (<i>no-op cast</i>) other than a type
3845change.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +00003846
3847<h5>Example:</h5>
3848<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +00003849 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
3850 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
Reid Spencerb7344ff2006-11-11 21:00:47 +00003851</pre>
3852</div>
3853
3854<!-- _______________________________________________________________________ -->
3855<div class="doc_subsubsection">
3856 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
3857</div>
3858<div class="doc_text">
3859
3860<h5>Syntax:</h5>
3861<pre>
3862 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3863</pre>
3864
3865<h5>Overview:</h5>
3866<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
3867a pointer type, <tt>ty2</tt>.</p>
3868
3869<h5>Arguments:</h5>
Duncan Sands16f122e2007-03-30 12:22:09 +00003870<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Reid Spencerb7344ff2006-11-11 21:00:47 +00003871value to cast, and a type to cast it to, which must be a
Dan Gohmanef9462f2008-10-14 16:51:45 +00003872<a href="#t_pointer">pointer</a> type.</p>
Reid Spencerb7344ff2006-11-11 21:00:47 +00003873
3874<h5>Semantics:</h5>
3875<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
3876<tt>ty2</tt> by applying either a zero extension or a truncation depending on
3877the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
3878size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
3879the size of a pointer then a zero extension is done. If they are the same size,
3880nothing is done (<i>no-op cast</i>).</p>
3881
3882<h5>Example:</h5>
3883<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +00003884 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
3885 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
3886 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencerb7344ff2006-11-11 21:00:47 +00003887</pre>
3888</div>
3889
3890<!-- _______________________________________________________________________ -->
3891<div class="doc_subsubsection">
Reid Spencer5b950642006-11-11 23:08:07 +00003892 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003893</div>
3894<div class="doc_text">
3895
3896<h5>Syntax:</h5>
3897<pre>
Reid Spencer5b950642006-11-11 23:08:07 +00003898 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003899</pre>
3900
3901<h5>Overview:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003902
Reid Spencer5b950642006-11-11 23:08:07 +00003903<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003904<tt>ty2</tt> without changing any bits.</p>
3905
3906<h5>Arguments:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003907
Reid Spencer5b950642006-11-11 23:08:07 +00003908<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Dan Gohmanc05dca92008-09-08 16:45:59 +00003909a non-aggregate first class value, and a type to cast it to, which must also be
3910a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes of
3911<tt>value</tt>
Reid Spencere3db84c2007-01-09 20:08:58 +00003912and the destination type, <tt>ty2</tt>, must be identical. If the source
Chris Lattnerb54c30f2008-05-20 20:48:21 +00003913type is a pointer, the destination type must also be a pointer. This
3914instruction supports bitwise conversion of vectors to integers and to vectors
3915of other types (as long as they have the same size).</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003916
3917<h5>Semantics:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +00003918<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Reid Spencerb7344ff2006-11-11 21:00:47 +00003919<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
3920this conversion. The conversion is done as if the <tt>value</tt> had been
3921stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
3922converted to other pointer types with this instruction. To convert pointers to
3923other types, use the <a href="#i_inttoptr">inttoptr</a> or
3924<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +00003925
3926<h5>Example:</h5>
3927<pre>
Jeff Cohen222a8a42007-04-29 01:07:00 +00003928 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00003929 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Dan Gohmanef9462f2008-10-14 16:51:45 +00003930 %Z = bitcast &lt;2 x int&gt; %V to i64; <i>; yields i64: %V</i>
Chris Lattner70de6632001-07-09 00:26:23 +00003931</pre>
Misha Brukman76307852003-11-08 01:05:38 +00003932</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00003933
Reid Spencer97c5fa42006-11-08 01:18:52 +00003934<!-- ======================================================================= -->
3935<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
3936<div class="doc_text">
3937<p>The instructions in this category are the "miscellaneous"
3938instructions, which defy better classification.</p>
3939</div>
Reid Spencerc828a0e2006-11-18 21:50:54 +00003940
3941<!-- _______________________________________________________________________ -->
3942<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
3943</div>
3944<div class="doc_text">
3945<h5>Syntax:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +00003946<pre> &lt;result&gt; = icmp &lt;cond&gt; &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {i1} or {&lt;N x i1&gt;}:result</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +00003947</pre>
3948<h5>Overview:</h5>
Dan Gohmanc579d972008-09-09 01:02:47 +00003949<p>The '<tt>icmp</tt>' instruction returns a boolean value or
3950a vector of boolean values based on comparison
3951of its two integer, integer vector, or pointer operands.</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +00003952<h5>Arguments:</h5>
3953<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Jeff Cohen222a8a42007-04-29 01:07:00 +00003954the condition code indicating the kind of comparison to perform. It is not
3955a value, just a keyword. The possible condition code are:
Dan Gohmanef9462f2008-10-14 16:51:45 +00003956</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +00003957<ol>
3958 <li><tt>eq</tt>: equal</li>
3959 <li><tt>ne</tt>: not equal </li>
3960 <li><tt>ugt</tt>: unsigned greater than</li>
3961 <li><tt>uge</tt>: unsigned greater or equal</li>
3962 <li><tt>ult</tt>: unsigned less than</li>
3963 <li><tt>ule</tt>: unsigned less or equal</li>
3964 <li><tt>sgt</tt>: signed greater than</li>
3965 <li><tt>sge</tt>: signed greater or equal</li>
3966 <li><tt>slt</tt>: signed less than</li>
3967 <li><tt>sle</tt>: signed less or equal</li>
3968</ol>
Chris Lattnerc0f423a2007-01-15 01:54:13 +00003969<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Dan Gohmanc579d972008-09-09 01:02:47 +00003970<a href="#t_pointer">pointer</a>
3971or integer <a href="#t_vector">vector</a> typed.
3972They must also be identical types.</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +00003973<h5>Semantics:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00003974<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to
Reid Spencerc828a0e2006-11-18 21:50:54 +00003975the condition code given as <tt>cond</tt>. The comparison performed always
Dan Gohmanc579d972008-09-09 01:02:47 +00003976yields either an <a href="#t_primitive"><tt>i1</tt></a> or vector of <tt>i1</tt> result, as follows:
Dan Gohmanef9462f2008-10-14 16:51:45 +00003977</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +00003978<ol>
3979 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
3980 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3981 </li>
3982 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Dan Gohmanef9462f2008-10-14 16:51:45 +00003983 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00003984 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Gabor Greif0f75ad02008-08-07 21:46:00 +00003985 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00003986 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
Gabor Greif0f75ad02008-08-07 21:46:00 +00003987 <tt>true</tt> if <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00003988 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
Gabor Greif0f75ad02008-08-07 21:46:00 +00003989 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00003990 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
Gabor Greif0f75ad02008-08-07 21:46:00 +00003991 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00003992 <li><tt>sgt</tt>: interprets the operands as signed values and yields
Gabor Greif0f75ad02008-08-07 21:46:00 +00003993 <tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00003994 <li><tt>sge</tt>: interprets the operands as signed values and yields
Gabor Greif0f75ad02008-08-07 21:46:00 +00003995 <tt>true</tt> if <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00003996 <li><tt>slt</tt>: interprets the operands as signed values and yields
Gabor Greif0f75ad02008-08-07 21:46:00 +00003997 <tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00003998 <li><tt>sle</tt>: interprets the operands as signed values and yields
Gabor Greif0f75ad02008-08-07 21:46:00 +00003999 <tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00004000</ol>
4001<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Jeff Cohen222a8a42007-04-29 01:07:00 +00004002values are compared as if they were integers.</p>
Dan Gohmanc579d972008-09-09 01:02:47 +00004003<p>If the operands are integer vectors, then they are compared
4004element by element. The result is an <tt>i1</tt> vector with
4005the same number of elements as the values being compared.
4006Otherwise, the result is an <tt>i1</tt>.
4007</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +00004008
4009<h5>Example:</h5>
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00004010<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
4011 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
4012 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
4013 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
4014 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
4015 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +00004016</pre>
4017</div>
4018
4019<!-- _______________________________________________________________________ -->
4020<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
4021</div>
4022<div class="doc_text">
4023<h5>Syntax:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +00004024<pre> &lt;result&gt; = fcmp &lt;cond&gt; &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {i1} or {&lt;N x i1&gt;}:result</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +00004025</pre>
4026<h5>Overview:</h5>
Dan Gohmanc579d972008-09-09 01:02:47 +00004027<p>The '<tt>fcmp</tt>' instruction returns a boolean value
4028or vector of boolean values based on comparison
Dan Gohmanef9462f2008-10-14 16:51:45 +00004029of its operands.</p>
Dan Gohmanc579d972008-09-09 01:02:47 +00004030<p>
4031If the operands are floating point scalars, then the result
4032type is a boolean (<a href="#t_primitive"><tt>i1</tt></a>).
4033</p>
4034<p>If the operands are floating point vectors, then the result type
4035is a vector of boolean with the same number of elements as the
4036operands being compared.</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +00004037<h5>Arguments:</h5>
4038<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Jeff Cohen222a8a42007-04-29 01:07:00 +00004039the condition code indicating the kind of comparison to perform. It is not
Dan Gohmanef9462f2008-10-14 16:51:45 +00004040a value, just a keyword. The possible condition code are:</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +00004041<ol>
Reid Spencerf69acf32006-11-19 03:00:14 +00004042 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00004043 <li><tt>oeq</tt>: ordered and equal</li>
4044 <li><tt>ogt</tt>: ordered and greater than </li>
4045 <li><tt>oge</tt>: ordered and greater than or equal</li>
4046 <li><tt>olt</tt>: ordered and less than </li>
4047 <li><tt>ole</tt>: ordered and less than or equal</li>
4048 <li><tt>one</tt>: ordered and not equal</li>
4049 <li><tt>ord</tt>: ordered (no nans)</li>
4050 <li><tt>ueq</tt>: unordered or equal</li>
4051 <li><tt>ugt</tt>: unordered or greater than </li>
4052 <li><tt>uge</tt>: unordered or greater than or equal</li>
4053 <li><tt>ult</tt>: unordered or less than </li>
4054 <li><tt>ule</tt>: unordered or less than or equal</li>
4055 <li><tt>une</tt>: unordered or not equal</li>
4056 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00004057 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00004058</ol>
Jeff Cohen222a8a42007-04-29 01:07:00 +00004059<p><i>Ordered</i> means that neither operand is a QNAN while
Reid Spencer02e0d1d2006-12-06 07:08:07 +00004060<i>unordered</i> means that either operand may be a QNAN.</p>
Dan Gohmanc579d972008-09-09 01:02:47 +00004061<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be
4062either a <a href="#t_floating">floating point</a> type
4063or a <a href="#t_vector">vector</a> of floating point type.
4064They must have identical types.</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +00004065<h5>Semantics:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00004066<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Dan Gohmanc579d972008-09-09 01:02:47 +00004067according to the condition code given as <tt>cond</tt>.
4068If the operands are vectors, then the vectors are compared
4069element by element.
4070Each comparison performed
Dan Gohmanef9462f2008-10-14 16:51:45 +00004071always yields an <a href="#t_primitive">i1</a> result, as follows:</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +00004072<ol>
4073 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00004074 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greif0f75ad02008-08-07 21:46:00 +00004075 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00004076 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greif0f75ad02008-08-07 21:46:00 +00004077 <tt>op1</tt> is greather than <tt>op2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00004078 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greif0f75ad02008-08-07 21:46:00 +00004079 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00004080 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greif0f75ad02008-08-07 21:46:00 +00004081 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00004082 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greif0f75ad02008-08-07 21:46:00 +00004083 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00004084 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greif0f75ad02008-08-07 21:46:00 +00004085 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00004086 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
4087 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greif0f75ad02008-08-07 21:46:00 +00004088 <tt>op1</tt> is equal to <tt>op2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00004089 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greif0f75ad02008-08-07 21:46:00 +00004090 <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00004091 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greif0f75ad02008-08-07 21:46:00 +00004092 <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00004093 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greif0f75ad02008-08-07 21:46:00 +00004094 <tt>op1</tt> is less than <tt>op2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00004095 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greif0f75ad02008-08-07 21:46:00 +00004096 <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00004097 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greif0f75ad02008-08-07 21:46:00 +00004098 <tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +00004099 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +00004100 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
4101</ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +00004102
4103<h5>Example:</h5>
4104<pre> &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
Dan Gohmanc579d972008-09-09 01:02:47 +00004105 &lt;result&gt; = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
4106 &lt;result&gt; = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
4107 &lt;result&gt; = fcmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
Reid Spencerc828a0e2006-11-18 21:50:54 +00004108</pre>
4109</div>
4110
Reid Spencer97c5fa42006-11-08 01:18:52 +00004111<!-- _______________________________________________________________________ -->
Nate Begemand2195702008-05-12 19:01:56 +00004112<div class="doc_subsubsection">
4113 <a name="i_vicmp">'<tt>vicmp</tt>' Instruction</a>
4114</div>
4115<div class="doc_text">
4116<h5>Syntax:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00004117<pre> &lt;result&gt; = vicmp &lt;cond&gt; &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt; <i>; yields {ty}:result</i>
Nate Begemand2195702008-05-12 19:01:56 +00004118</pre>
4119<h5>Overview:</h5>
4120<p>The '<tt>vicmp</tt>' instruction returns an integer vector value based on
4121element-wise comparison of its two integer vector operands.</p>
4122<h5>Arguments:</h5>
4123<p>The '<tt>vicmp</tt>' instruction takes three operands. The first operand is
4124the condition code indicating the kind of comparison to perform. It is not
Dan Gohmanef9462f2008-10-14 16:51:45 +00004125a value, just a keyword. The possible condition code are:</p>
Nate Begemand2195702008-05-12 19:01:56 +00004126<ol>
4127 <li><tt>eq</tt>: equal</li>
4128 <li><tt>ne</tt>: not equal </li>
4129 <li><tt>ugt</tt>: unsigned greater than</li>
4130 <li><tt>uge</tt>: unsigned greater or equal</li>
4131 <li><tt>ult</tt>: unsigned less than</li>
4132 <li><tt>ule</tt>: unsigned less or equal</li>
4133 <li><tt>sgt</tt>: signed greater than</li>
4134 <li><tt>sge</tt>: signed greater or equal</li>
4135 <li><tt>slt</tt>: signed less than</li>
4136 <li><tt>sle</tt>: signed less or equal</li>
4137</ol>
Dan Gohmanc579d972008-09-09 01:02:47 +00004138<p>The remaining two arguments must be <a href="#t_vector">vector</a> or
Nate Begemand2195702008-05-12 19:01:56 +00004139<a href="#t_integer">integer</a> typed. They must also be identical types.</p>
4140<h5>Semantics:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00004141<p>The '<tt>vicmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begemand2195702008-05-12 19:01:56 +00004142according to the condition code given as <tt>cond</tt>. The comparison yields a
4143<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, of
4144identical type as the values being compared. The most significant bit in each
4145element is 1 if the element-wise comparison evaluates to true, and is 0
4146otherwise. All other bits of the result are undefined. The condition codes
4147are evaluated identically to the <a href="#i_icmp">'<tt>icmp</tt>'
Dan Gohmanef9462f2008-10-14 16:51:45 +00004148instruction</a>.</p>
Nate Begemand2195702008-05-12 19:01:56 +00004149
4150<h5>Example:</h5>
4151<pre>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00004152 &lt;result&gt; = vicmp eq &lt;2 x i32&gt; &lt; i32 4, i32 0&gt;, &lt; i32 5, i32 0&gt; <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0, i32 -1 &gt;</i>
4153 &lt;result&gt; = vicmp ult &lt;2 x i8 &gt; &lt; i8 1, i8 2&gt;, &lt; i8 2, i8 2 &gt; <i>; yields: result=&lt;2 x i8&gt; &lt; i8 -1, i8 0 &gt;</i>
Nate Begemand2195702008-05-12 19:01:56 +00004154</pre>
4155</div>
4156
4157<!-- _______________________________________________________________________ -->
4158<div class="doc_subsubsection">
4159 <a name="i_vfcmp">'<tt>vfcmp</tt>' Instruction</a>
4160</div>
4161<div class="doc_text">
4162<h5>Syntax:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00004163<pre> &lt;result&gt; = vfcmp &lt;cond&gt; &lt;ty&gt; &lt;op1&gt;, &lt;op2&gt;</pre>
Nate Begemand2195702008-05-12 19:01:56 +00004164<h5>Overview:</h5>
4165<p>The '<tt>vfcmp</tt>' instruction returns an integer vector value based on
4166element-wise comparison of its two floating point vector operands. The output
4167elements have the same width as the input elements.</p>
4168<h5>Arguments:</h5>
4169<p>The '<tt>vfcmp</tt>' instruction takes three operands. The first operand is
4170the condition code indicating the kind of comparison to perform. It is not
Dan Gohmanef9462f2008-10-14 16:51:45 +00004171a value, just a keyword. The possible condition code are:</p>
Nate Begemand2195702008-05-12 19:01:56 +00004172<ol>
4173 <li><tt>false</tt>: no comparison, always returns false</li>
4174 <li><tt>oeq</tt>: ordered and equal</li>
4175 <li><tt>ogt</tt>: ordered and greater than </li>
4176 <li><tt>oge</tt>: ordered and greater than or equal</li>
4177 <li><tt>olt</tt>: ordered and less than </li>
4178 <li><tt>ole</tt>: ordered and less than or equal</li>
4179 <li><tt>one</tt>: ordered and not equal</li>
4180 <li><tt>ord</tt>: ordered (no nans)</li>
4181 <li><tt>ueq</tt>: unordered or equal</li>
4182 <li><tt>ugt</tt>: unordered or greater than </li>
4183 <li><tt>uge</tt>: unordered or greater than or equal</li>
4184 <li><tt>ult</tt>: unordered or less than </li>
4185 <li><tt>ule</tt>: unordered or less than or equal</li>
4186 <li><tt>une</tt>: unordered or not equal</li>
4187 <li><tt>uno</tt>: unordered (either nans)</li>
4188 <li><tt>true</tt>: no comparison, always returns true</li>
4189</ol>
4190<p>The remaining two arguments must be <a href="#t_vector">vector</a> of
4191<a href="#t_floating">floating point</a> typed. They must also be identical
4192types.</p>
4193<h5>Semantics:</h5>
Gabor Greif0f75ad02008-08-07 21:46:00 +00004194<p>The '<tt>vfcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begemand2195702008-05-12 19:01:56 +00004195according to the condition code given as <tt>cond</tt>. The comparison yields a
4196<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, with
4197an identical number of elements as the values being compared, and each element
4198having identical with to the width of the floating point elements. The most
4199significant bit in each element is 1 if the element-wise comparison evaluates to
4200true, and is 0 otherwise. All other bits of the result are undefined. The
4201condition codes are evaluated identically to the
Dan Gohmanef9462f2008-10-14 16:51:45 +00004202<a href="#i_fcmp">'<tt>fcmp</tt>' instruction</a>.</p>
Nate Begemand2195702008-05-12 19:01:56 +00004203
4204<h5>Example:</h5>
4205<pre>
Chris Lattner0ae02092008-10-13 16:55:18 +00004206 <i>; yields: result=&lt;2 x i32&gt; &lt; i32 0, i32 -1 &gt;</i>
4207 &lt;result&gt; = vfcmp oeq &lt;2 x float&gt; &lt; float 4, float 0 &gt;, &lt; float 5, float 0 &gt;
4208
4209 <i>; yields: result=&lt;2 x i64&gt; &lt; i64 -1, i64 0 &gt;</i>
4210 &lt;result&gt; = vfcmp ult &lt;2 x double&gt; &lt; double 1, double 2 &gt;, &lt; double 2, double 2&gt;
Nate Begemand2195702008-05-12 19:01:56 +00004211</pre>
4212</div>
4213
4214<!-- _______________________________________________________________________ -->
Chris Lattnerb54c30f2008-05-20 20:48:21 +00004215<div class="doc_subsubsection">
4216 <a name="i_phi">'<tt>phi</tt>' Instruction</a>
4217</div>
4218
Reid Spencer97c5fa42006-11-08 01:18:52 +00004219<div class="doc_text">
Chris Lattnerb54c30f2008-05-20 20:48:21 +00004220
Reid Spencer97c5fa42006-11-08 01:18:52 +00004221<h5>Syntax:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00004222
Reid Spencer97c5fa42006-11-08 01:18:52 +00004223<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
4224<h5>Overview:</h5>
4225<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
4226the SSA graph representing the function.</p>
4227<h5>Arguments:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00004228
Jeff Cohen222a8a42007-04-29 01:07:00 +00004229<p>The type of the incoming values is specified with the first type
Reid Spencer97c5fa42006-11-08 01:18:52 +00004230field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
4231as arguments, with one pair for each predecessor basic block of the
4232current block. Only values of <a href="#t_firstclass">first class</a>
4233type may be used as the value arguments to the PHI node. Only labels
4234may be used as the label arguments.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00004235
Reid Spencer97c5fa42006-11-08 01:18:52 +00004236<p>There must be no non-phi instructions between the start of a basic
4237block and the PHI instructions: i.e. PHI instructions must be first in
4238a basic block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00004239
Reid Spencer97c5fa42006-11-08 01:18:52 +00004240<h5>Semantics:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00004241
Jeff Cohen222a8a42007-04-29 01:07:00 +00004242<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
4243specified by the pair corresponding to the predecessor basic block that executed
4244just prior to the current block.</p>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00004245
Reid Spencer97c5fa42006-11-08 01:18:52 +00004246<h5>Example:</h5>
Chris Lattnerb54c30f2008-05-20 20:48:21 +00004247<pre>
4248Loop: ; Infinite loop that counts from 0 on up...
4249 %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
4250 %nextindvar = add i32 %indvar, 1
4251 br label %Loop
4252</pre>
Reid Spencer97c5fa42006-11-08 01:18:52 +00004253</div>
4254
Chris Lattnerb53c28d2004-03-12 05:50:16 +00004255<!-- _______________________________________________________________________ -->
4256<div class="doc_subsubsection">
4257 <a name="i_select">'<tt>select</tt>' Instruction</a>
4258</div>
4259
4260<div class="doc_text">
4261
4262<h5>Syntax:</h5>
4263
4264<pre>
Dan Gohmanc579d972008-09-09 01:02:47 +00004265 &lt;result&gt; = select <i>selty</i> &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
4266
Dan Gohmanef9462f2008-10-14 16:51:45 +00004267 <i>selty</i> is either i1 or {&lt;N x i1&gt;}
Chris Lattnerb53c28d2004-03-12 05:50:16 +00004268</pre>
4269
4270<h5>Overview:</h5>
4271
4272<p>
4273The '<tt>select</tt>' instruction is used to choose one value based on a
4274condition, without branching.
4275</p>
4276
4277
4278<h5>Arguments:</h5>
4279
4280<p>
Dan Gohmanc579d972008-09-09 01:02:47 +00004281The '<tt>select</tt>' instruction requires an 'i1' value or
4282a vector of 'i1' values indicating the
Chris Lattnerb54c30f2008-05-20 20:48:21 +00004283condition, and two values of the same <a href="#t_firstclass">first class</a>
Dan Gohmanc579d972008-09-09 01:02:47 +00004284type. If the val1/val2 are vectors and
4285the condition is a scalar, then entire vectors are selected, not
Chris Lattnerb54c30f2008-05-20 20:48:21 +00004286individual elements.
Chris Lattnerb53c28d2004-03-12 05:50:16 +00004287</p>
4288
4289<h5>Semantics:</h5>
4290
4291<p>
Dan Gohmanc579d972008-09-09 01:02:47 +00004292If the condition is an i1 and it evaluates to 1, the instruction returns the first
John Criswell88190562005-05-16 16:17:45 +00004293value argument; otherwise, it returns the second value argument.
Chris Lattnerb53c28d2004-03-12 05:50:16 +00004294</p>
Dan Gohmanc579d972008-09-09 01:02:47 +00004295<p>
4296If the condition is a vector of i1, then the value arguments must
4297be vectors of the same size, and the selection is done element
4298by element.
4299</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00004300
4301<h5>Example:</h5>
4302
4303<pre>
Reid Spencer36a15422007-01-12 03:35:51 +00004304 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnerb53c28d2004-03-12 05:50:16 +00004305</pre>
4306</div>
4307
Robert Bocchinof72fdfe2006-01-15 20:48:27 +00004308
4309<!-- _______________________________________________________________________ -->
4310<div class="doc_subsubsection">
Chris Lattnere23c1392005-05-06 05:47:36 +00004311 <a name="i_call">'<tt>call</tt>' Instruction</a>
4312</div>
4313
Misha Brukman76307852003-11-08 01:05:38 +00004314<div class="doc_text">
Chris Lattnere23c1392005-05-06 05:47:36 +00004315
Chris Lattner2f7c9632001-06-06 20:29:01 +00004316<h5>Syntax:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +00004317<pre>
Devang Patel02256232008-10-07 17:48:33 +00004318 &lt;result&gt; = [tail] call [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>] &lt;ty&gt; [&lt;fnty&gt;*] &lt;fnptrval&gt;(&lt;function args&gt;) [<a href="#fnattrs">fn attrs</a>]
Chris Lattnere23c1392005-05-06 05:47:36 +00004319</pre>
4320
Chris Lattner2f7c9632001-06-06 20:29:01 +00004321<h5>Overview:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +00004322
Misha Brukman76307852003-11-08 01:05:38 +00004323<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +00004324
Chris Lattner2f7c9632001-06-06 20:29:01 +00004325<h5>Arguments:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +00004326
Misha Brukman76307852003-11-08 01:05:38 +00004327<p>This instruction requires several arguments:</p>
Chris Lattnere23c1392005-05-06 05:47:36 +00004328
Chris Lattnera8292f32002-05-06 22:08:29 +00004329<ol>
Chris Lattner48b383b02003-11-25 01:02:51 +00004330 <li>
Chris Lattner0132aff2005-05-06 22:57:40 +00004331 <p>The optional "tail" marker indicates whether the callee function accesses
4332 any allocas or varargs in the caller. If the "tail" marker is present, the
Chris Lattnere23c1392005-05-06 05:47:36 +00004333 function call is eligible for tail call optimization. Note that calls may
4334 be marked "tail" even if they do not occur before a <a
Dan Gohmanef9462f2008-10-14 16:51:45 +00004335 href="#i_ret"><tt>ret</tt></a> instruction.</p>
Chris Lattner48b383b02003-11-25 01:02:51 +00004336 </li>
4337 <li>
Duncan Sands16f122e2007-03-30 12:22:09 +00004338 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
Chris Lattner0132aff2005-05-06 22:57:40 +00004339 convention</a> the call should use. If none is specified, the call defaults
Dan Gohmanef9462f2008-10-14 16:51:45 +00004340 to using C calling conventions.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +00004341 </li>
Devang Patel7e9b05e2008-10-06 18:50:38 +00004342
4343 <li>
4344 <p>The optional <a href="#paramattrs">Parameter Attributes</a> list for
4345 return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
4346 and '<tt>inreg</tt>' attributes are valid here.</p>
4347 </li>
4348
Chris Lattner0132aff2005-05-06 22:57:40 +00004349 <li>
Nick Lewyckya9b13d52007-09-08 13:57:50 +00004350 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
4351 the type of the return value. Functions that return no value are marked
4352 <tt><a href="#t_void">void</a></tt>.</p>
4353 </li>
4354 <li>
4355 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
4356 value being invoked. The argument types must match the types implied by
4357 this signature. This type can be omitted if the function is not varargs
4358 and if the function type does not return a pointer to a function.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +00004359 </li>
4360 <li>
4361 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
4362 be invoked. In most cases, this is a direct function invocation, but
4363 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
John Criswell88190562005-05-16 16:17:45 +00004364 to function value.</p>
Chris Lattner48b383b02003-11-25 01:02:51 +00004365 </li>
4366 <li>
4367 <p>'<tt>function args</tt>': argument list whose types match the
Reid Spencerd845d162005-05-01 22:22:57 +00004368 function signature argument types. All arguments must be of
4369 <a href="#t_firstclass">first class</a> type. If the function signature
4370 indicates the function accepts a variable number of arguments, the extra
4371 arguments can be specified.</p>
Chris Lattner48b383b02003-11-25 01:02:51 +00004372 </li>
Devang Patel7e9b05e2008-10-06 18:50:38 +00004373 <li>
Devang Patel02256232008-10-07 17:48:33 +00004374 <p>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patel7e9b05e2008-10-06 18:50:38 +00004375 '<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
4376 '<tt>readnone</tt>' attributes are valid here.</p>
4377 </li>
Chris Lattnera8292f32002-05-06 22:08:29 +00004378</ol>
Chris Lattnere23c1392005-05-06 05:47:36 +00004379
Chris Lattner2f7c9632001-06-06 20:29:01 +00004380<h5>Semantics:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +00004381
Chris Lattner48b383b02003-11-25 01:02:51 +00004382<p>The '<tt>call</tt>' instruction is used to cause control flow to
4383transfer to a specified function, with its incoming arguments bound to
4384the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
4385instruction in the called function, control flow continues with the
4386instruction after the function call, and the return value of the
Dan Gohmanef9462f2008-10-14 16:51:45 +00004387function is bound to the result argument.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +00004388
Chris Lattner2f7c9632001-06-06 20:29:01 +00004389<h5>Example:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +00004390
4391<pre>
Nick Lewyckya9b13d52007-09-08 13:57:50 +00004392 %retval = call i32 @test(i32 %argc)
Chris Lattnerfb7c88d2008-03-21 17:24:17 +00004393 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
4394 %X = tail call i32 @foo() <i>; yields i32</i>
4395 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
4396 call void %foo(i8 97 signext)
Devang Pateld6cff512008-03-10 20:49:15 +00004397
4398 %struct.A = type { i32, i8 }
Devang Patel7e9b05e2008-10-06 18:50:38 +00004399 %r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohmancc3132e2008-10-04 19:00:07 +00004400 %gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
4401 %gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner6cbe8e92008-10-08 06:26:11 +00004402 %Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijmaneefa7df2008-10-07 10:03:45 +00004403 %ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattnere23c1392005-05-06 05:47:36 +00004404</pre>
4405
Misha Brukman76307852003-11-08 01:05:38 +00004406</div>
Chris Lattner6a4a0492004-09-27 21:51:25 +00004407
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004408<!-- _______________________________________________________________________ -->
Chris Lattner6a4a0492004-09-27 21:51:25 +00004409<div class="doc_subsubsection">
Chris Lattner33337472006-01-13 23:26:01 +00004410 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
Chris Lattner6a4a0492004-09-27 21:51:25 +00004411</div>
4412
Misha Brukman76307852003-11-08 01:05:38 +00004413<div class="doc_text">
Chris Lattner6a4a0492004-09-27 21:51:25 +00004414
Chris Lattner26ca62e2003-10-18 05:51:36 +00004415<h5>Syntax:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +00004416
4417<pre>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00004418 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattner6a4a0492004-09-27 21:51:25 +00004419</pre>
4420
Chris Lattner26ca62e2003-10-18 05:51:36 +00004421<h5>Overview:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +00004422
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00004423<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Chris Lattner6a4a0492004-09-27 21:51:25 +00004424the "variable argument" area of a function call. It is used to implement the
4425<tt>va_arg</tt> macro in C.</p>
4426
Chris Lattner26ca62e2003-10-18 05:51:36 +00004427<h5>Arguments:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +00004428
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00004429<p>This instruction takes a <tt>va_list*</tt> value and the type of
4430the argument. It returns a value of the specified argument type and
Jeff Cohen222a8a42007-04-29 01:07:00 +00004431increments the <tt>va_list</tt> to point to the next argument. The
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00004432actual type of <tt>va_list</tt> is target specific.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +00004433
Chris Lattner26ca62e2003-10-18 05:51:36 +00004434<h5>Semantics:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +00004435
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00004436<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
4437type from the specified <tt>va_list</tt> and causes the
4438<tt>va_list</tt> to point to the next argument. For more information,
4439see the variable argument handling <a href="#int_varargs">Intrinsic
4440Functions</a>.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +00004441
4442<p>It is legal for this instruction to be called in a function which does not
4443take a variable number of arguments, for example, the <tt>vfprintf</tt>
Misha Brukman76307852003-11-08 01:05:38 +00004444function.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +00004445
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00004446<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
John Criswell88190562005-05-16 16:17:45 +00004447href="#intrinsics">intrinsic function</a> because it takes a type as an
Chris Lattner6a4a0492004-09-27 21:51:25 +00004448argument.</p>
4449
Chris Lattner26ca62e2003-10-18 05:51:36 +00004450<h5>Example:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +00004451
4452<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
4453
Misha Brukman76307852003-11-08 01:05:38 +00004454</div>
Chris Lattner941515c2004-01-06 05:31:32 +00004455
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004456<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +00004457<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
4458<!-- *********************************************************************** -->
Chris Lattner941515c2004-01-06 05:31:32 +00004459
Misha Brukman76307852003-11-08 01:05:38 +00004460<div class="doc_text">
Chris Lattnerfee11462004-02-12 17:01:32 +00004461
4462<p>LLVM supports the notion of an "intrinsic function". These functions have
Reid Spencer4eefaab2007-04-01 08:04:23 +00004463well known names and semantics and are required to follow certain restrictions.
4464Overall, these intrinsics represent an extension mechanism for the LLVM
Jeff Cohen222a8a42007-04-29 01:07:00 +00004465language that does not require changing all of the transformations in LLVM when
Gabor Greifa54634a2007-07-06 22:07:22 +00004466adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
Chris Lattnerfee11462004-02-12 17:01:32 +00004467
John Criswell88190562005-05-16 16:17:45 +00004468<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Jeff Cohen222a8a42007-04-29 01:07:00 +00004469prefix is reserved in LLVM for intrinsic names; thus, function names may not
4470begin with this prefix. Intrinsic functions must always be external functions:
4471you cannot define the body of intrinsic functions. Intrinsic functions may
4472only be used in call or invoke instructions: it is illegal to take the address
4473of an intrinsic function. Additionally, because intrinsic functions are part
4474of the LLVM language, it is required if any are added that they be documented
4475here.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +00004476
Chandler Carruth7132e002007-08-04 01:51:18 +00004477<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
4478a family of functions that perform the same operation but on different data
4479types. Because LLVM can represent over 8 million different integer types,
4480overloading is used commonly to allow an intrinsic function to operate on any
4481integer type. One or more of the argument types or the result type can be
4482overloaded to accept any integer type. Argument types may also be defined as
4483exactly matching a previous argument's type or the result type. This allows an
4484intrinsic function which accepts multiple arguments, but needs all of them to
4485be of the same type, to only be overloaded with respect to a single argument or
4486the result.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +00004487
Chandler Carruth7132e002007-08-04 01:51:18 +00004488<p>Overloaded intrinsics will have the names of its overloaded argument types
4489encoded into its function name, each preceded by a period. Only those types
4490which are overloaded result in a name suffix. Arguments whose type is matched
4491against another type do not. For example, the <tt>llvm.ctpop</tt> function can
4492take an integer of any width and returns an integer of exactly the same integer
4493width. This leads to a family of functions such as
4494<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
4495Only one type, the return type, is overloaded, and only one type suffix is
4496required. Because the argument's type is matched against the return type, it
4497does not require its own name suffix.</p>
Reid Spencer4eefaab2007-04-01 08:04:23 +00004498
4499<p>To learn how to add an intrinsic function, please see the
4500<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
Chris Lattnerfee11462004-02-12 17:01:32 +00004501</p>
4502
Misha Brukman76307852003-11-08 01:05:38 +00004503</div>
Chris Lattner941515c2004-01-06 05:31:32 +00004504
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004505<!-- ======================================================================= -->
Chris Lattner941515c2004-01-06 05:31:32 +00004506<div class="doc_subsection">
4507 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
4508</div>
4509
Misha Brukman76307852003-11-08 01:05:38 +00004510<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +00004511
Misha Brukman76307852003-11-08 01:05:38 +00004512<p>Variable argument support is defined in LLVM with the <a
Chris Lattner33337472006-01-13 23:26:01 +00004513 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
Chris Lattner48b383b02003-11-25 01:02:51 +00004514intrinsic functions. These functions are related to the similarly
4515named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004516
Chris Lattner48b383b02003-11-25 01:02:51 +00004517<p>All of these functions operate on arguments that use a
4518target-specific value type "<tt>va_list</tt>". The LLVM assembly
4519language reference manual does not define what this type is, so all
Jeff Cohen222a8a42007-04-29 01:07:00 +00004520transformations should be prepared to handle these functions regardless of
4521the type used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004522
Chris Lattner30b868d2006-05-15 17:26:46 +00004523<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
Chris Lattner48b383b02003-11-25 01:02:51 +00004524instruction and the variable argument handling intrinsic functions are
4525used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004526
Bill Wendling3716c5d2007-05-29 09:04:49 +00004527<div class="doc_code">
Chris Lattnerfee11462004-02-12 17:01:32 +00004528<pre>
Anton Korobeynikov640bbe02007-03-21 23:58:04 +00004529define i32 @test(i32 %X, ...) {
Chris Lattnerfee11462004-02-12 17:01:32 +00004530 ; Initialize variable argument processing
Jeff Cohen222a8a42007-04-29 01:07:00 +00004531 %ap = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +00004532 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov640bbe02007-03-21 23:58:04 +00004533 call void @llvm.va_start(i8* %ap2)
Chris Lattnerfee11462004-02-12 17:01:32 +00004534
4535 ; Read a single integer argument
Jeff Cohen222a8a42007-04-29 01:07:00 +00004536 %tmp = va_arg i8** %ap, i32
Chris Lattnerfee11462004-02-12 17:01:32 +00004537
4538 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohen222a8a42007-04-29 01:07:00 +00004539 %aq = alloca i8*
Chris Lattnerdb0790c2007-01-08 07:55:15 +00004540 %aq2 = bitcast i8** %aq to i8*
Jeff Cohen222a8a42007-04-29 01:07:00 +00004541 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov640bbe02007-03-21 23:58:04 +00004542 call void @llvm.va_end(i8* %aq2)
Chris Lattnerfee11462004-02-12 17:01:32 +00004543
4544 ; Stop processing of arguments.
Anton Korobeynikov640bbe02007-03-21 23:58:04 +00004545 call void @llvm.va_end(i8* %ap2)
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00004546 ret i32 %tmp
Chris Lattnerfee11462004-02-12 17:01:32 +00004547}
Anton Korobeynikov640bbe02007-03-21 23:58:04 +00004548
4549declare void @llvm.va_start(i8*)
4550declare void @llvm.va_copy(i8*, i8*)
4551declare void @llvm.va_end(i8*)
Chris Lattnerfee11462004-02-12 17:01:32 +00004552</pre>
Misha Brukman76307852003-11-08 01:05:38 +00004553</div>
Chris Lattner941515c2004-01-06 05:31:32 +00004554
Bill Wendling3716c5d2007-05-29 09:04:49 +00004555</div>
4556
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004557<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +00004558<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00004559 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
Chris Lattner941515c2004-01-06 05:31:32 +00004560</div>
4561
4562
Misha Brukman76307852003-11-08 01:05:38 +00004563<div class="doc_text">
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004564<h5>Syntax:</h5>
Chris Lattnerdb0790c2007-01-08 07:55:15 +00004565<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004566<h5>Overview:</h5>
Dan Gohmanef9462f2008-10-14 16:51:45 +00004567<p>The '<tt>llvm.va_start</tt>' intrinsic initializes
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00004568<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
4569href="#i_va_arg">va_arg</a></tt>.</p>
4570
4571<h5>Arguments:</h5>
4572
Dan Gohmanef9462f2008-10-14 16:51:45 +00004573<p>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00004574
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004575<h5>Semantics:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00004576
Dan Gohmanef9462f2008-10-14 16:51:45 +00004577<p>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00004578macro available in C. In a target-dependent way, it initializes the
Jeff Cohen222a8a42007-04-29 01:07:00 +00004579<tt>va_list</tt> element to which the argument points, so that the next call to
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00004580<tt>va_arg</tt> will produce the first variable argument passed to the function.
4581Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
Jeff Cohen222a8a42007-04-29 01:07:00 +00004582last argument of the function as the compiler can figure that out.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00004583
Misha Brukman76307852003-11-08 01:05:38 +00004584</div>
Chris Lattner941515c2004-01-06 05:31:32 +00004585
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004586<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +00004587<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00004588 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
Chris Lattner941515c2004-01-06 05:31:32 +00004589</div>
4590
Misha Brukman76307852003-11-08 01:05:38 +00004591<div class="doc_text">
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004592<h5>Syntax:</h5>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00004593<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004594<h5>Overview:</h5>
Chris Lattnerdb0790c2007-01-08 07:55:15 +00004595
Jeff Cohen222a8a42007-04-29 01:07:00 +00004596<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Reid Spencer96a5f022007-04-04 02:42:35 +00004597which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
Chris Lattner48b383b02003-11-25 01:02:51 +00004598or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +00004599
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004600<h5>Arguments:</h5>
Chris Lattnerdb0790c2007-01-08 07:55:15 +00004601
Jeff Cohen222a8a42007-04-29 01:07:00 +00004602<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +00004603
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004604<h5>Semantics:</h5>
Chris Lattnerdb0790c2007-01-08 07:55:15 +00004605
Misha Brukman76307852003-11-08 01:05:38 +00004606<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Jeff Cohen222a8a42007-04-29 01:07:00 +00004607macro available in C. In a target-dependent way, it destroys the
4608<tt>va_list</tt> element to which the argument points. Calls to <a
4609href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
4610<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
4611<tt>llvm.va_end</tt>.</p>
Chris Lattnerdb0790c2007-01-08 07:55:15 +00004612
Misha Brukman76307852003-11-08 01:05:38 +00004613</div>
Chris Lattner941515c2004-01-06 05:31:32 +00004614
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004615<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +00004616<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00004617 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
Chris Lattner941515c2004-01-06 05:31:32 +00004618</div>
4619
Misha Brukman76307852003-11-08 01:05:38 +00004620<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +00004621
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004622<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004623
4624<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00004625 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattner757528b0b2004-05-23 21:06:01 +00004626</pre>
4627
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004628<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004629
Jeff Cohen222a8a42007-04-29 01:07:00 +00004630<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
4631from the source argument list to the destination argument list.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004632
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004633<h5>Arguments:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004634
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00004635<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Andrew Lenharth5305ea52005-06-22 20:38:11 +00004636The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +00004637
Chris Lattner757528b0b2004-05-23 21:06:01 +00004638
Chris Lattnerbd64b4e2003-05-08 04:57:36 +00004639<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004640
Jeff Cohen222a8a42007-04-29 01:07:00 +00004641<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
4642macro available in C. In a target-dependent way, it copies the source
4643<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
4644intrinsic is necessary because the <tt><a href="#int_va_start">
4645llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
4646example, memory allocation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004647
Misha Brukman76307852003-11-08 01:05:38 +00004648</div>
Chris Lattner941515c2004-01-06 05:31:32 +00004649
Chris Lattnerfee11462004-02-12 17:01:32 +00004650<!-- ======================================================================= -->
4651<div class="doc_subsection">
Chris Lattner757528b0b2004-05-23 21:06:01 +00004652 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
4653</div>
4654
4655<div class="doc_text">
4656
4657<p>
4658LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner67c37d12008-08-05 18:29:16 +00004659Collection</a> (GC) requires the implementation and generation of these
4660intrinsics.
Reid Spencer96a5f022007-04-04 02:42:35 +00004661These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
Chris Lattner757528b0b2004-05-23 21:06:01 +00004662stack</a>, as well as garbage collector implementations that require <a
Reid Spencer96a5f022007-04-04 02:42:35 +00004663href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
Chris Lattner757528b0b2004-05-23 21:06:01 +00004664Front-ends for type-safe garbage collected languages should generate these
4665intrinsics to make use of the LLVM garbage collectors. For more details, see <a
4666href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
4667</p>
Christopher Lamb55c6d4f2007-12-17 01:00:21 +00004668
4669<p>The garbage collection intrinsics only operate on objects in the generic
4670 address space (address space zero).</p>
4671
Chris Lattner757528b0b2004-05-23 21:06:01 +00004672</div>
4673
4674<!-- _______________________________________________________________________ -->
4675<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00004676 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004677</div>
4678
4679<div class="doc_text">
4680
4681<h5>Syntax:</h5>
4682
4683<pre>
Chris Lattner12477732007-09-21 17:30:40 +00004684 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattner757528b0b2004-05-23 21:06:01 +00004685</pre>
4686
4687<h5>Overview:</h5>
4688
John Criswelldfe6a862004-12-10 15:51:16 +00004689<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Chris Lattner757528b0b2004-05-23 21:06:01 +00004690the code generator, and allows some metadata to be associated with it.</p>
4691
4692<h5>Arguments:</h5>
4693
4694<p>The first argument specifies the address of a stack object that contains the
4695root pointer. The second pointer (which must be either a constant or a global
4696value address) contains the meta-data to be associated with the root.</p>
4697
4698<h5>Semantics:</h5>
4699
Chris Lattner851b7712008-04-24 05:59:56 +00004700<p>At runtime, a call to this intrinsic stores a null pointer into the "ptrloc"
Chris Lattner757528b0b2004-05-23 21:06:01 +00004701location. At compile-time, the code generator generates information to allow
Gordon Henriksenfb56bde2007-12-25 02:31:26 +00004702the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
4703intrinsic may only be used in a function which <a href="#gc">specifies a GC
4704algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004705
4706</div>
4707
4708
4709<!-- _______________________________________________________________________ -->
4710<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00004711 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004712</div>
4713
4714<div class="doc_text">
4715
4716<h5>Syntax:</h5>
4717
4718<pre>
Chris Lattner12477732007-09-21 17:30:40 +00004719 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattner757528b0b2004-05-23 21:06:01 +00004720</pre>
4721
4722<h5>Overview:</h5>
4723
4724<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
4725locations, allowing garbage collector implementations that require read
4726barriers.</p>
4727
4728<h5>Arguments:</h5>
4729
Chris Lattnerf9228072006-03-14 20:02:51 +00004730<p>The second argument is the address to read from, which should be an address
4731allocated from the garbage collector. The first object is a pointer to the
4732start of the referenced object, if needed by the language runtime (otherwise
4733null).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004734
4735<h5>Semantics:</h5>
4736
4737<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
4738instruction, but may be replaced with substantially more complex code by the
Gordon Henriksenfb56bde2007-12-25 02:31:26 +00004739garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
4740may only be used in a function which <a href="#gc">specifies a GC
4741algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004742
4743</div>
4744
4745
4746<!-- _______________________________________________________________________ -->
4747<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00004748 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004749</div>
4750
4751<div class="doc_text">
4752
4753<h5>Syntax:</h5>
4754
4755<pre>
Chris Lattner12477732007-09-21 17:30:40 +00004756 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattner757528b0b2004-05-23 21:06:01 +00004757</pre>
4758
4759<h5>Overview:</h5>
4760
4761<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
4762locations, allowing garbage collector implementations that require write
4763barriers (such as generational or reference counting collectors).</p>
4764
4765<h5>Arguments:</h5>
4766
Chris Lattnerf9228072006-03-14 20:02:51 +00004767<p>The first argument is the reference to store, the second is the start of the
4768object to store it to, and the third is the address of the field of Obj to
4769store to. If the runtime does not require a pointer to the object, Obj may be
4770null.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004771
4772<h5>Semantics:</h5>
4773
4774<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
4775instruction, but may be replaced with substantially more complex code by the
Gordon Henriksenfb56bde2007-12-25 02:31:26 +00004776garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
4777may only be used in a function which <a href="#gc">specifies a GC
4778algorithm</a>.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +00004779
4780</div>
4781
4782
4783
4784<!-- ======================================================================= -->
4785<div class="doc_subsection">
Chris Lattner3649c3a2004-02-14 04:08:35 +00004786 <a name="int_codegen">Code Generator Intrinsics</a>
4787</div>
4788
4789<div class="doc_text">
4790<p>
4791These intrinsics are provided by LLVM to expose special features that may only
4792be implemented with code generator support.
4793</p>
4794
4795</div>
4796
4797<!-- _______________________________________________________________________ -->
4798<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00004799 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
Chris Lattner3649c3a2004-02-14 04:08:35 +00004800</div>
4801
4802<div class="doc_text">
4803
4804<h5>Syntax:</h5>
4805<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00004806 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +00004807</pre>
4808
4809<h5>Overview:</h5>
4810
4811<p>
Chris Lattnerc1fb4262006-10-15 20:05:59 +00004812The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
4813target-specific value indicating the return address of the current function
4814or one of its callers.
Chris Lattner3649c3a2004-02-14 04:08:35 +00004815</p>
4816
4817<h5>Arguments:</h5>
4818
4819<p>
4820The argument to this intrinsic indicates which function to return the address
4821for. Zero indicates the calling function, one indicates its caller, etc. The
4822argument is <b>required</b> to be a constant integer value.
4823</p>
4824
4825<h5>Semantics:</h5>
4826
4827<p>
4828The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
4829the return address of the specified call frame, or zero if it cannot be
4830identified. The value returned by this intrinsic is likely to be incorrect or 0
4831for arguments other than zero, so it should only be used for debugging purposes.
4832</p>
4833
4834<p>
4835Note that calling this intrinsic does not prevent function inlining or other
Chris Lattner2e6eb5f2005-03-07 20:30:51 +00004836aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner3649c3a2004-02-14 04:08:35 +00004837source-language caller.
4838</p>
4839</div>
4840
4841
4842<!-- _______________________________________________________________________ -->
4843<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00004844 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
Chris Lattner3649c3a2004-02-14 04:08:35 +00004845</div>
4846
4847<div class="doc_text">
4848
4849<h5>Syntax:</h5>
4850<pre>
Chris Lattner12477732007-09-21 17:30:40 +00004851 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +00004852</pre>
4853
4854<h5>Overview:</h5>
4855
4856<p>
Chris Lattnerc1fb4262006-10-15 20:05:59 +00004857The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
4858target-specific frame pointer value for the specified stack frame.
Chris Lattner3649c3a2004-02-14 04:08:35 +00004859</p>
4860
4861<h5>Arguments:</h5>
4862
4863<p>
4864The argument to this intrinsic indicates which function to return the frame
4865pointer for. Zero indicates the calling function, one indicates its caller,
4866etc. The argument is <b>required</b> to be a constant integer value.
4867</p>
4868
4869<h5>Semantics:</h5>
4870
4871<p>
4872The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
4873the frame address of the specified call frame, or zero if it cannot be
4874identified. The value returned by this intrinsic is likely to be incorrect or 0
4875for arguments other than zero, so it should only be used for debugging purposes.
4876</p>
4877
4878<p>
4879Note that calling this intrinsic does not prevent function inlining or other
Chris Lattner2e6eb5f2005-03-07 20:30:51 +00004880aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner3649c3a2004-02-14 04:08:35 +00004881source-language caller.
4882</p>
4883</div>
4884
Chris Lattnerc8a2c222005-02-28 19:24:19 +00004885<!-- _______________________________________________________________________ -->
4886<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00004887 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
Chris Lattner2f0f0012006-01-13 02:03:13 +00004888</div>
4889
4890<div class="doc_text">
4891
4892<h5>Syntax:</h5>
4893<pre>
Chris Lattner12477732007-09-21 17:30:40 +00004894 declare i8 *@llvm.stacksave()
Chris Lattner2f0f0012006-01-13 02:03:13 +00004895</pre>
4896
4897<h5>Overview:</h5>
4898
4899<p>
4900The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
Reid Spencer96a5f022007-04-04 02:42:35 +00004901the function stack, for use with <a href="#int_stackrestore">
Chris Lattner2f0f0012006-01-13 02:03:13 +00004902<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
4903features like scoped automatic variable sized arrays in C99.
4904</p>
4905
4906<h5>Semantics:</h5>
4907
4908<p>
4909This intrinsic returns a opaque pointer value that can be passed to <a
Reid Spencer96a5f022007-04-04 02:42:35 +00004910href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
Chris Lattner2f0f0012006-01-13 02:03:13 +00004911<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
4912<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
4913state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
4914practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
4915that were allocated after the <tt>llvm.stacksave</tt> was executed.
4916</p>
4917
4918</div>
4919
4920<!-- _______________________________________________________________________ -->
4921<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00004922 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
Chris Lattner2f0f0012006-01-13 02:03:13 +00004923</div>
4924
4925<div class="doc_text">
4926
4927<h5>Syntax:</h5>
4928<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00004929 declare void @llvm.stackrestore(i8 * %ptr)
Chris Lattner2f0f0012006-01-13 02:03:13 +00004930</pre>
4931
4932<h5>Overview:</h5>
4933
4934<p>
4935The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
4936the function stack to the state it was in when the corresponding <a
Reid Spencer96a5f022007-04-04 02:42:35 +00004937href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
Chris Lattner2f0f0012006-01-13 02:03:13 +00004938useful for implementing language features like scoped automatic variable sized
4939arrays in C99.
4940</p>
4941
4942<h5>Semantics:</h5>
4943
4944<p>
Reid Spencer96a5f022007-04-04 02:42:35 +00004945See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
Chris Lattner2f0f0012006-01-13 02:03:13 +00004946</p>
4947
4948</div>
4949
4950
4951<!-- _______________________________________________________________________ -->
4952<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00004953 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
Chris Lattnerc8a2c222005-02-28 19:24:19 +00004954</div>
4955
4956<div class="doc_text">
4957
4958<h5>Syntax:</h5>
4959<pre>
Chris Lattner12477732007-09-21 17:30:40 +00004960 declare void @llvm.prefetch(i8* &lt;address&gt;, i32 &lt;rw&gt;, i32 &lt;locality&gt;)
Chris Lattnerc8a2c222005-02-28 19:24:19 +00004961</pre>
4962
4963<h5>Overview:</h5>
4964
4965
4966<p>
4967The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
John Criswell88190562005-05-16 16:17:45 +00004968a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
4969no
4970effect on the behavior of the program but can change its performance
Chris Lattnerff851072005-02-28 19:47:14 +00004971characteristics.
Chris Lattnerc8a2c222005-02-28 19:24:19 +00004972</p>
4973
4974<h5>Arguments:</h5>
4975
4976<p>
4977<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
4978determining if the fetch should be for a read (0) or write (1), and
4979<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Chris Lattnerd3e641c2005-03-07 20:31:38 +00004980locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
Chris Lattnerc8a2c222005-02-28 19:24:19 +00004981<tt>locality</tt> arguments must be constant integers.
4982</p>
4983
4984<h5>Semantics:</h5>
4985
4986<p>
4987This intrinsic does not modify the behavior of the program. In particular,
4988prefetches cannot trap and do not produce a value. On targets that support this
4989intrinsic, the prefetch can provide hints to the processor cache for better
4990performance.
4991</p>
4992
4993</div>
4994
Andrew Lenharthb4427912005-03-28 20:05:49 +00004995<!-- _______________________________________________________________________ -->
4996<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00004997 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
Andrew Lenharthb4427912005-03-28 20:05:49 +00004998</div>
4999
5000<div class="doc_text">
5001
5002<h5>Syntax:</h5>
5003<pre>
Chris Lattner12477732007-09-21 17:30:40 +00005004 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharthb4427912005-03-28 20:05:49 +00005005</pre>
5006
5007<h5>Overview:</h5>
5008
5009
5010<p>
John Criswell88190562005-05-16 16:17:45 +00005011The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
Chris Lattner67c37d12008-08-05 18:29:16 +00005012(PC) in a region of
5013code to simulators and other tools. The method is target specific, but it is
5014expected that the marker will use exported symbols to transmit the PC of the
5015marker.
5016The marker makes no guarantees that it will remain with any specific instruction
5017after optimizations. It is possible that the presence of a marker will inhibit
Chris Lattnerb40261e2006-03-24 07:16:10 +00005018optimizations. The intended use is to be inserted after optimizations to allow
John Criswell88190562005-05-16 16:17:45 +00005019correlations of simulation runs.
Andrew Lenharthb4427912005-03-28 20:05:49 +00005020</p>
5021
5022<h5>Arguments:</h5>
5023
5024<p>
5025<tt>id</tt> is a numerical id identifying the marker.
5026</p>
5027
5028<h5>Semantics:</h5>
5029
5030<p>
5031This intrinsic does not modify the behavior of the program. Backends that do not
5032support this intrinisic may ignore it.
5033</p>
5034
5035</div>
5036
Andrew Lenharth01aa5632005-11-11 16:47:30 +00005037<!-- _______________________________________________________________________ -->
5038<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005039 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
Andrew Lenharth01aa5632005-11-11 16:47:30 +00005040</div>
5041
5042<div class="doc_text">
5043
5044<h5>Syntax:</h5>
5045<pre>
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00005046 declare i64 @llvm.readcyclecounter( )
Andrew Lenharth01aa5632005-11-11 16:47:30 +00005047</pre>
5048
5049<h5>Overview:</h5>
5050
5051
5052<p>
5053The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
5054counter register (or similar low latency, high accuracy clocks) on those targets
5055that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
5056As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
5057should only be used for small timings.
5058</p>
5059
5060<h5>Semantics:</h5>
5061
5062<p>
5063When directly supported, reading the cycle counter should not modify any memory.
5064Implementations are allowed to either return a application specific value or a
5065system wide value. On backends without support, this is lowered to a constant 0.
5066</p>
5067
5068</div>
5069
Chris Lattner3649c3a2004-02-14 04:08:35 +00005070<!-- ======================================================================= -->
5071<div class="doc_subsection">
Chris Lattnerfee11462004-02-12 17:01:32 +00005072 <a name="int_libc">Standard C Library Intrinsics</a>
5073</div>
5074
5075<div class="doc_text">
5076<p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00005077LLVM provides intrinsics for a few important standard C library functions.
5078These intrinsics allow source-language front-ends to pass information about the
5079alignment of the pointer arguments to the code generator, providing opportunity
5080for more efficient code generation.
Chris Lattnerfee11462004-02-12 17:01:32 +00005081</p>
5082
5083</div>
5084
5085<!-- _______________________________________________________________________ -->
5086<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005087 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
Chris Lattnerfee11462004-02-12 17:01:32 +00005088</div>
5089
5090<div class="doc_text">
5091
5092<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +00005093<p>This is an overloaded intrinsic. You can use llvm.memcpy on any integer bit
5094width. Not all targets support all bit widths however.</p>
Chris Lattnerfee11462004-02-12 17:01:32 +00005095<pre>
Chris Lattnerdd708342008-11-21 16:42:48 +00005096 declare void @llvm.memcpy.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5097 i8 &lt;len&gt;, i32 &lt;align&gt;)
5098 declare void @llvm.memcpy.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5099 i16 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00005100 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00005101 i32 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00005102 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00005103 i64 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattnerfee11462004-02-12 17:01:32 +00005104</pre>
5105
5106<h5>Overview:</h5>
5107
5108<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +00005109The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattnerfee11462004-02-12 17:01:32 +00005110location to the destination location.
5111</p>
5112
5113<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +00005114Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
5115intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattnerfee11462004-02-12 17:01:32 +00005116</p>
5117
5118<h5>Arguments:</h5>
5119
5120<p>
5121The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner0c8b2592006-03-03 00:07:20 +00005122the source. The third argument is an integer argument
Chris Lattnerfee11462004-02-12 17:01:32 +00005123specifying the number of bytes to copy, and the fourth argument is the alignment
5124of the source and destination locations.
5125</p>
5126
Chris Lattner4c67c482004-02-12 21:18:15 +00005127<p>
5128If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattner5316e5d2006-03-04 00:02:10 +00005129the caller guarantees that both the source and destination pointers are aligned
5130to that boundary.
Chris Lattner4c67c482004-02-12 21:18:15 +00005131</p>
5132
Chris Lattnerfee11462004-02-12 17:01:32 +00005133<h5>Semantics:</h5>
5134
5135<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +00005136The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattnerfee11462004-02-12 17:01:32 +00005137location to the destination location, which are not allowed to overlap. It
5138copies "len" bytes of memory over. If the argument is known to be aligned to
5139some boundary, this can be specified as the fourth argument, otherwise it should
5140be set to 0 or 1.
5141</p>
5142</div>
5143
5144
Chris Lattnerf30152e2004-02-12 18:10:10 +00005145<!-- _______________________________________________________________________ -->
5146<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005147 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
Chris Lattnerf30152e2004-02-12 18:10:10 +00005148</div>
5149
5150<div class="doc_text">
5151
5152<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +00005153<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
5154width. Not all targets support all bit widths however.</p>
Chris Lattnerf30152e2004-02-12 18:10:10 +00005155<pre>
Chris Lattnerdd708342008-11-21 16:42:48 +00005156 declare void @llvm.memmove.i8(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5157 i8 &lt;len&gt;, i32 &lt;align&gt;)
5158 declare void @llvm.memmove.i16(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
5159 i16 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00005160 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00005161 i32 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00005162 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00005163 i64 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattnerf30152e2004-02-12 18:10:10 +00005164</pre>
5165
5166<h5>Overview:</h5>
5167
5168<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +00005169The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
5170location to the destination location. It is similar to the
Chris Lattnerec564022008-01-06 19:51:52 +00005171'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Chris Lattnerf30152e2004-02-12 18:10:10 +00005172</p>
5173
5174<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +00005175Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
5176intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattnerf30152e2004-02-12 18:10:10 +00005177</p>
5178
5179<h5>Arguments:</h5>
5180
5181<p>
5182The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner0c8b2592006-03-03 00:07:20 +00005183the source. The third argument is an integer argument
Chris Lattnerf30152e2004-02-12 18:10:10 +00005184specifying the number of bytes to copy, and the fourth argument is the alignment
5185of the source and destination locations.
5186</p>
5187
Chris Lattner4c67c482004-02-12 21:18:15 +00005188<p>
5189If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattner5316e5d2006-03-04 00:02:10 +00005190the caller guarantees that the source and destination pointers are aligned to
5191that boundary.
Chris Lattner4c67c482004-02-12 21:18:15 +00005192</p>
5193
Chris Lattnerf30152e2004-02-12 18:10:10 +00005194<h5>Semantics:</h5>
5195
5196<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +00005197The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
Chris Lattnerf30152e2004-02-12 18:10:10 +00005198location to the destination location, which may overlap. It
5199copies "len" bytes of memory over. If the argument is known to be aligned to
5200some boundary, this can be specified as the fourth argument, otherwise it should
5201be set to 0 or 1.
5202</p>
5203</div>
5204
Chris Lattner941515c2004-01-06 05:31:32 +00005205
Chris Lattner3649c3a2004-02-14 04:08:35 +00005206<!-- _______________________________________________________________________ -->
5207<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005208 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
Chris Lattner3649c3a2004-02-14 04:08:35 +00005209</div>
5210
5211<div class="doc_text">
5212
5213<h5>Syntax:</h5>
Chris Lattnerdd708342008-11-21 16:42:48 +00005214<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
5215width. Not all targets support all bit widths however.</p>
Chris Lattner3649c3a2004-02-14 04:08:35 +00005216<pre>
Chris Lattnerdd708342008-11-21 16:42:48 +00005217 declare void @llvm.memset.i8(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5218 i8 &lt;len&gt;, i32 &lt;align&gt;)
5219 declare void @llvm.memset.i16(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
5220 i16 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00005221 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00005222 i32 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00005223 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00005224 i64 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +00005225</pre>
5226
5227<h5>Overview:</h5>
5228
5229<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +00005230The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
Chris Lattner3649c3a2004-02-14 04:08:35 +00005231byte value.
5232</p>
5233
5234<p>
5235Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
5236does not return a value, and takes an extra alignment argument.
5237</p>
5238
5239<h5>Arguments:</h5>
5240
5241<p>
5242The first argument is a pointer to the destination to fill, the second is the
Chris Lattner0c8b2592006-03-03 00:07:20 +00005243byte value to fill it with, the third argument is an integer
Chris Lattner3649c3a2004-02-14 04:08:35 +00005244argument specifying the number of bytes to fill, and the fourth argument is the
5245known alignment of destination location.
5246</p>
5247
5248<p>
5249If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattner5316e5d2006-03-04 00:02:10 +00005250the caller guarantees that the destination pointer is aligned to that boundary.
Chris Lattner3649c3a2004-02-14 04:08:35 +00005251</p>
5252
5253<h5>Semantics:</h5>
5254
5255<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +00005256The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
5257the
Chris Lattner3649c3a2004-02-14 04:08:35 +00005258destination location. If the argument is known to be aligned to some boundary,
5259this can be specified as the fourth argument, otherwise it should be set to 0 or
52601.
5261</p>
5262</div>
5263
5264
Chris Lattner3b4f4372004-06-11 02:28:03 +00005265<!-- _______________________________________________________________________ -->
5266<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005267 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
Chris Lattner8a8f2e52005-07-21 01:29:16 +00005268</div>
5269
5270<div class="doc_text">
5271
5272<h5>Syntax:</h5>
Dale Johannesendd89d272007-10-02 17:47:38 +00005273<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
Dan Gohmanb6324c12007-10-15 20:30:11 +00005274floating point or vector of floating point type. Not all targets support all
Dan Gohmanef9462f2008-10-14 16:51:45 +00005275types however.</p>
Chris Lattner8a8f2e52005-07-21 01:29:16 +00005276<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +00005277 declare float @llvm.sqrt.f32(float %Val)
5278 declare double @llvm.sqrt.f64(double %Val)
5279 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
5280 declare fp128 @llvm.sqrt.f128(fp128 %Val)
5281 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattner8a8f2e52005-07-21 01:29:16 +00005282</pre>
5283
5284<h5>Overview:</h5>
5285
5286<p>
Reid Spencerb4f9a6f2006-01-16 21:12:35 +00005287The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohmanb6324c12007-10-15 20:30:11 +00005288returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Chris Lattner8a8f2e52005-07-21 01:29:16 +00005289<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattner00d7cb92008-01-29 07:00:44 +00005290negative numbers other than -0.0 (which allows for better optimization, because
5291there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
5292defined to return -0.0 like IEEE sqrt.
Chris Lattner8a8f2e52005-07-21 01:29:16 +00005293</p>
5294
5295<h5>Arguments:</h5>
5296
5297<p>
5298The argument and return value are floating point numbers of the same type.
5299</p>
5300
5301<h5>Semantics:</h5>
5302
5303<p>
Dan Gohman33988db2007-07-16 14:37:41 +00005304This function returns the sqrt of the specified operand if it is a nonnegative
Chris Lattner8a8f2e52005-07-21 01:29:16 +00005305floating point number.
5306</p>
5307</div>
5308
Chris Lattner33b73f92006-09-08 06:34:02 +00005309<!-- _______________________________________________________________________ -->
5310<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005311 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
Chris Lattner33b73f92006-09-08 06:34:02 +00005312</div>
5313
5314<div class="doc_text">
5315
5316<h5>Syntax:</h5>
Dale Johannesendd89d272007-10-02 17:47:38 +00005317<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
Dan Gohmanb6324c12007-10-15 20:30:11 +00005318floating point or vector of floating point type. Not all targets support all
Dan Gohmanef9462f2008-10-14 16:51:45 +00005319types however.</p>
Chris Lattner33b73f92006-09-08 06:34:02 +00005320<pre>
Dale Johannesendd89d272007-10-02 17:47:38 +00005321 declare float @llvm.powi.f32(float %Val, i32 %power)
5322 declare double @llvm.powi.f64(double %Val, i32 %power)
5323 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
5324 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
5325 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattner33b73f92006-09-08 06:34:02 +00005326</pre>
5327
5328<h5>Overview:</h5>
5329
5330<p>
5331The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
5332specified (positive or negative) power. The order of evaluation of
Dan Gohmanb6324c12007-10-15 20:30:11 +00005333multiplications is not defined. When a vector of floating point type is
5334used, the second argument remains a scalar integer value.
Chris Lattner33b73f92006-09-08 06:34:02 +00005335</p>
5336
5337<h5>Arguments:</h5>
5338
5339<p>
5340The second argument is an integer power, and the first is a value to raise to
5341that power.
5342</p>
5343
5344<h5>Semantics:</h5>
5345
5346<p>
5347This function returns the first value raised to the second power with an
5348unspecified sequence of rounding operations.</p>
5349</div>
5350
Dan Gohmanb6324c12007-10-15 20:30:11 +00005351<!-- _______________________________________________________________________ -->
5352<div class="doc_subsubsection">
5353 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
5354</div>
5355
5356<div class="doc_text">
5357
5358<h5>Syntax:</h5>
5359<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
5360floating point or vector of floating point type. Not all targets support all
Dan Gohmanef9462f2008-10-14 16:51:45 +00005361types however.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00005362<pre>
5363 declare float @llvm.sin.f32(float %Val)
5364 declare double @llvm.sin.f64(double %Val)
5365 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
5366 declare fp128 @llvm.sin.f128(fp128 %Val)
5367 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
5368</pre>
5369
5370<h5>Overview:</h5>
5371
5372<p>
5373The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
5374</p>
5375
5376<h5>Arguments:</h5>
5377
5378<p>
5379The argument and return value are floating point numbers of the same type.
5380</p>
5381
5382<h5>Semantics:</h5>
5383
5384<p>
5385This function returns the sine of the specified operand, returning the
5386same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmand0806a02007-10-17 18:05:13 +00005387conditions in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00005388</div>
5389
5390<!-- _______________________________________________________________________ -->
5391<div class="doc_subsubsection">
5392 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
5393</div>
5394
5395<div class="doc_text">
5396
5397<h5>Syntax:</h5>
5398<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
5399floating point or vector of floating point type. Not all targets support all
Dan Gohmanef9462f2008-10-14 16:51:45 +00005400types however.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00005401<pre>
5402 declare float @llvm.cos.f32(float %Val)
5403 declare double @llvm.cos.f64(double %Val)
5404 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
5405 declare fp128 @llvm.cos.f128(fp128 %Val)
5406 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
5407</pre>
5408
5409<h5>Overview:</h5>
5410
5411<p>
5412The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
5413</p>
5414
5415<h5>Arguments:</h5>
5416
5417<p>
5418The argument and return value are floating point numbers of the same type.
5419</p>
5420
5421<h5>Semantics:</h5>
5422
5423<p>
5424This function returns the cosine of the specified operand, returning the
5425same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohmand0806a02007-10-17 18:05:13 +00005426conditions in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00005427</div>
5428
5429<!-- _______________________________________________________________________ -->
5430<div class="doc_subsubsection">
5431 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
5432</div>
5433
5434<div class="doc_text">
5435
5436<h5>Syntax:</h5>
5437<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
5438floating point or vector of floating point type. Not all targets support all
Dan Gohmanef9462f2008-10-14 16:51:45 +00005439types however.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00005440<pre>
5441 declare float @llvm.pow.f32(float %Val, float %Power)
5442 declare double @llvm.pow.f64(double %Val, double %Power)
5443 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
5444 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
5445 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
5446</pre>
5447
5448<h5>Overview:</h5>
5449
5450<p>
5451The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
5452specified (positive or negative) power.
5453</p>
5454
5455<h5>Arguments:</h5>
5456
5457<p>
5458The second argument is a floating point power, and the first is a value to
5459raise to that power.
5460</p>
5461
5462<h5>Semantics:</h5>
5463
5464<p>
5465This function returns the first value raised to the second power,
5466returning the
5467same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmand0806a02007-10-17 18:05:13 +00005468conditions in the same way.</p>
Dan Gohmanb6324c12007-10-15 20:30:11 +00005469</div>
5470
Chris Lattner33b73f92006-09-08 06:34:02 +00005471
Andrew Lenharth1d463522005-05-03 18:01:48 +00005472<!-- ======================================================================= -->
5473<div class="doc_subsection">
Nate Begeman0f223bb2006-01-13 23:26:38 +00005474 <a name="int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +00005475</div>
5476
5477<div class="doc_text">
5478<p>
Nate Begeman0f223bb2006-01-13 23:26:38 +00005479LLVM provides intrinsics for a few important bit manipulation operations.
Andrew Lenharth1d463522005-05-03 18:01:48 +00005480These allow efficient code generation for some algorithms.
5481</p>
5482
5483</div>
5484
5485<!-- _______________________________________________________________________ -->
5486<div class="doc_subsubsection">
Reid Spencer96a5f022007-04-04 02:42:35 +00005487 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
Nate Begeman0f223bb2006-01-13 23:26:38 +00005488</div>
5489
5490<div class="doc_text">
5491
5492<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +00005493<p>This is an overloaded intrinsic function. You can use bswap on any integer
Dan Gohmanef9462f2008-10-14 16:51:45 +00005494type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
Nate Begeman0f223bb2006-01-13 23:26:38 +00005495<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +00005496 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
5497 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
5498 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman0f223bb2006-01-13 23:26:38 +00005499</pre>
5500
5501<h5>Overview:</h5>
5502
5503<p>
Reid Spencerf361c4f2007-04-02 02:25:19 +00005504The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
Reid Spencer4eefaab2007-04-01 08:04:23 +00005505values with an even number of bytes (positive multiple of 16 bits). These are
5506useful for performing operations on data that is not in the target's native
5507byte order.
Nate Begeman0f223bb2006-01-13 23:26:38 +00005508</p>
5509
5510<h5>Semantics:</h5>
5511
5512<p>
Chandler Carruth7132e002007-08-04 01:51:18 +00005513The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00005514and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
5515intrinsic returns an i32 value that has the four bytes of the input i32
5516swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the returned
Chandler Carruth7132e002007-08-04 01:51:18 +00005517i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
5518<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Reid Spencer4eefaab2007-04-01 08:04:23 +00005519additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
Nate Begeman0f223bb2006-01-13 23:26:38 +00005520</p>
5521
5522</div>
5523
5524<!-- _______________________________________________________________________ -->
5525<div class="doc_subsubsection">
Reid Spencerb4f9a6f2006-01-16 21:12:35 +00005526 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +00005527</div>
5528
5529<div class="doc_text">
5530
5531<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +00005532<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Dan Gohmanef9462f2008-10-14 16:51:45 +00005533width. Not all targets support all bit widths however.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +00005534<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +00005535 declare i8 @llvm.ctpop.i8 (i8 &lt;src&gt;)
5536 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00005537 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +00005538 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
5539 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +00005540</pre>
5541
5542<h5>Overview:</h5>
5543
5544<p>
Chris Lattner069b5bd2006-01-16 22:38:59 +00005545The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
5546value.
Andrew Lenharth1d463522005-05-03 18:01:48 +00005547</p>
5548
5549<h5>Arguments:</h5>
5550
5551<p>
Chris Lattner573f64e2005-05-07 01:46:40 +00005552The only argument is the value to be counted. The argument may be of any
Reid Spencer3e628eb92007-01-04 16:43:23 +00005553integer type. The return type must match the argument type.
Andrew Lenharth1d463522005-05-03 18:01:48 +00005554</p>
5555
5556<h5>Semantics:</h5>
5557
5558<p>
5559The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
5560</p>
5561</div>
5562
5563<!-- _______________________________________________________________________ -->
5564<div class="doc_subsubsection">
Chris Lattnerb748c672006-01-16 22:34:14 +00005565 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +00005566</div>
5567
5568<div class="doc_text">
5569
5570<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +00005571<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
Dan Gohmanef9462f2008-10-14 16:51:45 +00005572integer bit width. Not all targets support all bit widths however.</p>
Andrew Lenharth1d463522005-05-03 18:01:48 +00005573<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +00005574 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
5575 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00005576 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +00005577 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
5578 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +00005579</pre>
5580
5581<h5>Overview:</h5>
5582
5583<p>
Reid Spencerb4f9a6f2006-01-16 21:12:35 +00005584The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
5585leading zeros in a variable.
Andrew Lenharth1d463522005-05-03 18:01:48 +00005586</p>
5587
5588<h5>Arguments:</h5>
5589
5590<p>
Chris Lattner573f64e2005-05-07 01:46:40 +00005591The only argument is the value to be counted. The argument may be of any
Reid Spencer3e628eb92007-01-04 16:43:23 +00005592integer type. The return type must match the argument type.
Andrew Lenharth1d463522005-05-03 18:01:48 +00005593</p>
5594
5595<h5>Semantics:</h5>
5596
5597<p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +00005598The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
5599in a variable. If the src == 0 then the result is the size in bits of the type
Reid Spencerb5ebf3d2006-12-31 07:07:53 +00005600of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
Andrew Lenharth1d463522005-05-03 18:01:48 +00005601</p>
5602</div>
Chris Lattner3b4f4372004-06-11 02:28:03 +00005603
5604
Chris Lattnerefa20fa2005-05-15 19:39:26 +00005605
5606<!-- _______________________________________________________________________ -->
5607<div class="doc_subsubsection">
Chris Lattnerb748c672006-01-16 22:34:14 +00005608 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
Chris Lattnerefa20fa2005-05-15 19:39:26 +00005609</div>
5610
5611<div class="doc_text">
5612
5613<h5>Syntax:</h5>
Reid Spencer4eefaab2007-04-01 08:04:23 +00005614<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
Dan Gohmanef9462f2008-10-14 16:51:45 +00005615integer bit width. Not all targets support all bit widths however.</p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +00005616<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +00005617 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
5618 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovbe9c93c2007-03-22 00:02:17 +00005619 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth7132e002007-08-04 01:51:18 +00005620 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
5621 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Chris Lattnerefa20fa2005-05-15 19:39:26 +00005622</pre>
5623
5624<h5>Overview:</h5>
5625
5626<p>
Reid Spencerb4f9a6f2006-01-16 21:12:35 +00005627The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
5628trailing zeros.
Chris Lattnerefa20fa2005-05-15 19:39:26 +00005629</p>
5630
5631<h5>Arguments:</h5>
5632
5633<p>
5634The only argument is the value to be counted. The argument may be of any
Reid Spencer3e628eb92007-01-04 16:43:23 +00005635integer type. The return type must match the argument type.
Chris Lattnerefa20fa2005-05-15 19:39:26 +00005636</p>
5637
5638<h5>Semantics:</h5>
5639
5640<p>
5641The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
5642in a variable. If the src == 0 then the result is the size in bits of the type
5643of src. For example, <tt>llvm.cttz(2) = 1</tt>.
5644</p>
5645</div>
5646
Reid Spencer8a5799f2007-04-01 08:27:01 +00005647<!-- _______________________________________________________________________ -->
5648<div class="doc_subsubsection">
Reid Spencerea2945e2007-04-10 02:51:31 +00005649 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
Reid Spencer8bc7d952007-04-01 19:00:37 +00005650</div>
5651
5652<div class="doc_text">
5653
5654<h5>Syntax:</h5>
Reid Spencerea2945e2007-04-10 02:51:31 +00005655<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
Dan Gohmanef9462f2008-10-14 16:51:45 +00005656on any integer bit width.</p>
Reid Spencer8bc7d952007-04-01 19:00:37 +00005657<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +00005658 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
5659 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Reid Spencer8bc7d952007-04-01 19:00:37 +00005660</pre>
5661
5662<h5>Overview:</h5>
Reid Spencerea2945e2007-04-10 02:51:31 +00005663<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
Reid Spencer8bc7d952007-04-01 19:00:37 +00005664range of bits from an integer value and returns them in the same bit width as
5665the original value.</p>
5666
5667<h5>Arguments:</h5>
5668<p>The first argument, <tt>%val</tt> and the result may be integer types of
5669any bit width but they must have the same bit width. The second and third
Reid Spencer96a5f022007-04-04 02:42:35 +00005670arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
Reid Spencer8bc7d952007-04-01 19:00:37 +00005671
5672<h5>Semantics:</h5>
Reid Spencerea2945e2007-04-10 02:51:31 +00005673<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
Reid Spencer96a5f022007-04-04 02:42:35 +00005674of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
5675<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
5676operates in forward mode.</p>
5677<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
5678right by <tt>%loBit</tt> bits and then ANDing it with a mask with
Reid Spencer8bc7d952007-04-01 19:00:37 +00005679only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
5680<ol>
5681 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
5682 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
5683 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
5684 to determine the number of bits to retain.</li>
5685 <li>A mask of the retained bits is created by shifting a -1 value.</li>
Dan Gohmanef9462f2008-10-14 16:51:45 +00005686 <li>The mask is ANDed with <tt>%val</tt> to produce the result.</li>
Reid Spencer8bc7d952007-04-01 19:00:37 +00005687</ol>
Reid Spencer70845c02007-05-14 16:14:57 +00005688<p>In reverse mode, a similar computation is made except that the bits are
5689returned in the reverse order. So, for example, if <tt>X</tt> has the value
5690<tt>i16 0x0ACF (101011001111)</tt> and we apply
5691<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
5692<tt>i16 0x0026 (000000100110)</tt>.</p>
Reid Spencer8bc7d952007-04-01 19:00:37 +00005693</div>
5694
Reid Spencer5bf54c82007-04-11 23:23:49 +00005695<div class="doc_subsubsection">
5696 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
5697</div>
5698
5699<div class="doc_text">
5700
5701<h5>Syntax:</h5>
5702<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
Dan Gohmanef9462f2008-10-14 16:51:45 +00005703on any integer bit width.</p>
Reid Spencer5bf54c82007-04-11 23:23:49 +00005704<pre>
Chandler Carruth7132e002007-08-04 01:51:18 +00005705 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
5706 declare i29 @llvm.part.set.i29.i9 (i29 %val, i9 %repl, i32 %lo, i32 %hi)
Reid Spencer5bf54c82007-04-11 23:23:49 +00005707</pre>
5708
5709<h5>Overview:</h5>
5710<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
5711of bits in an integer value with another integer value. It returns the integer
5712with the replaced bits.</p>
5713
5714<h5>Arguments:</h5>
5715<p>The first argument, <tt>%val</tt> and the result may be integer types of
5716any bit width but they must have the same bit width. <tt>%val</tt> is the value
5717whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
5718integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
5719type since they specify only a bit index.</p>
5720
5721<h5>Semantics:</h5>
5722<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
5723of operation: forwards and reverse. If <tt>%lo</tt> is greater than
5724<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
5725operates in forward mode.</p>
5726<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
5727truncating it down to the size of the replacement area or zero extending it
5728up to that size.</p>
5729<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
5730are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
5731in <tt>%repl</tt> replaces the <tt>%lo</tt>th bit in <tt>%val</tt> and etc. up
Dan Gohmanef9462f2008-10-14 16:51:45 +00005732to the <tt>%hi</tt>th bit.</p>
Reid Spencer146281c2007-05-14 16:50:20 +00005733<p>In reverse mode, a similar computation is made except that the bits are
5734reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
Dan Gohmanef9462f2008-10-14 16:51:45 +00005735<tt>%hi</tt> bit in <tt>%val</tt> and etc. down to the <tt>%lo</tt>th bit.</p>
Reid Spencer5bf54c82007-04-11 23:23:49 +00005736<h5>Examples:</h5>
5737<pre>
Reid Spencerc70afc32007-04-12 01:03:03 +00005738 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
Reid Spencer146281c2007-05-14 16:50:20 +00005739 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
5740 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
5741 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
Reid Spencerc70afc32007-04-12 01:03:03 +00005742 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
Reid Spencer7972c472007-04-11 23:49:50 +00005743</pre>
Reid Spencer5bf54c82007-04-11 23:23:49 +00005744</div>
5745
Chris Lattner941515c2004-01-06 05:31:32 +00005746<!-- ======================================================================= -->
5747<div class="doc_subsection">
5748 <a name="int_debugger">Debugger Intrinsics</a>
5749</div>
5750
5751<div class="doc_text">
5752<p>
5753The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
5754are described in the <a
5755href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
5756Debugging</a> document.
5757</p>
5758</div>
5759
5760
Jim Laskey2211f492007-03-14 19:31:19 +00005761<!-- ======================================================================= -->
5762<div class="doc_subsection">
5763 <a name="int_eh">Exception Handling Intrinsics</a>
5764</div>
5765
5766<div class="doc_text">
5767<p> The LLVM exception handling intrinsics (which all start with
5768<tt>llvm.eh.</tt> prefix), are described in the <a
5769href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
5770Handling</a> document. </p>
5771</div>
5772
Tanya Lattnercb1b9602007-06-15 20:50:54 +00005773<!-- ======================================================================= -->
5774<div class="doc_subsection">
Duncan Sands86e01192007-09-11 14:10:23 +00005775 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands644f9172007-07-27 12:58:54 +00005776</div>
5777
5778<div class="doc_text">
5779<p>
Duncan Sands86e01192007-09-11 14:10:23 +00005780 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands644f9172007-07-27 12:58:54 +00005781 the <tt>nest</tt> attribute, from a function. The result is a callable
5782 function pointer lacking the nest parameter - the caller does not need
5783 to provide a value for it. Instead, the value to use is stored in
5784 advance in a "trampoline", a block of memory usually allocated
5785 on the stack, which also contains code to splice the nest value into the
5786 argument list. This is used to implement the GCC nested function address
5787 extension.
5788</p>
5789<p>
5790 For example, if the function is
5791 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
Bill Wendling252570f2007-09-22 09:23:55 +00005792 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as follows:</p>
Duncan Sands644f9172007-07-27 12:58:54 +00005793<pre>
Duncan Sands86e01192007-09-11 14:10:23 +00005794 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
5795 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
5796 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
5797 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands644f9172007-07-27 12:58:54 +00005798</pre>
Bill Wendling252570f2007-09-22 09:23:55 +00005799 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
5800 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands644f9172007-07-27 12:58:54 +00005801</div>
5802
5803<!-- _______________________________________________________________________ -->
5804<div class="doc_subsubsection">
5805 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
5806</div>
5807<div class="doc_text">
5808<h5>Syntax:</h5>
5809<pre>
Duncan Sands86e01192007-09-11 14:10:23 +00005810declare i8* @llvm.init.trampoline(i8* &lt;tramp&gt;, i8* &lt;func&gt;, i8* &lt;nval&gt;)
Duncan Sands644f9172007-07-27 12:58:54 +00005811</pre>
5812<h5>Overview:</h5>
5813<p>
Duncan Sands86e01192007-09-11 14:10:23 +00005814 This fills the memory pointed to by <tt>tramp</tt> with code
5815 and returns a function pointer suitable for executing it.
Duncan Sands644f9172007-07-27 12:58:54 +00005816</p>
5817<h5>Arguments:</h5>
5818<p>
5819 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
5820 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
5821 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sandsf2bcd372007-08-22 23:39:54 +00005822 intrinsic. Note that the size and the alignment are target-specific - LLVM
5823 currently provides no portable way of determining them, so a front-end that
5824 generates this intrinsic needs to have some target-specific knowledge.
5825 The <tt>func</tt> argument must hold a function bitcast to an <tt>i8*</tt>.
Duncan Sands644f9172007-07-27 12:58:54 +00005826</p>
5827<h5>Semantics:</h5>
5828<p>
5829 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands86e01192007-09-11 14:10:23 +00005830 dependent code, turning it into a function. A pointer to this function is
5831 returned, but needs to be bitcast to an
Duncan Sands644f9172007-07-27 12:58:54 +00005832 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sands86e01192007-09-11 14:10:23 +00005833 before being called. The new function's signature is the same as that of
5834 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
5835 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
5836 of pointer type. Calling the new function is equivalent to calling
5837 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
5838 missing <tt>nest</tt> argument. If, after calling
5839 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
5840 modified, then the effect of any later call to the returned function pointer is
5841 undefined.
Duncan Sands644f9172007-07-27 12:58:54 +00005842</p>
5843</div>
5844
5845<!-- ======================================================================= -->
5846<div class="doc_subsection">
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00005847 <a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
5848</div>
5849
5850<div class="doc_text">
5851<p>
5852 These intrinsic functions expand the "universal IR" of LLVM to represent
5853 hardware constructs for atomic operations and memory synchronization. This
5854 provides an interface to the hardware, not an interface to the programmer. It
Chris Lattner67c37d12008-08-05 18:29:16 +00005855 is aimed at a low enough level to allow any programming models or APIs
5856 (Application Programming Interfaces) which
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00005857 need atomic behaviors to map cleanly onto it. It is also modeled primarily on
5858 hardware behavior. Just as hardware provides a "universal IR" for source
5859 languages, it also provides a starting point for developing a "universal"
5860 atomic operation and synchronization IR.
5861</p>
5862<p>
5863 These do <em>not</em> form an API such as high-level threading libraries,
5864 software transaction memory systems, atomic primitives, and intrinsic
5865 functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
5866 application libraries. The hardware interface provided by LLVM should allow
5867 a clean implementation of all of these APIs and parallel programming models.
5868 No one model or paradigm should be selected above others unless the hardware
5869 itself ubiquitously does so.
5870
5871</p>
5872</div>
5873
5874<!-- _______________________________________________________________________ -->
5875<div class="doc_subsubsection">
5876 <a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
5877</div>
5878<div class="doc_text">
5879<h5>Syntax:</h5>
5880<pre>
5881declare void @llvm.memory.barrier( i1 &lt;ll&gt;, i1 &lt;ls&gt;, i1 &lt;sl&gt;, i1 &lt;ss&gt;,
5882i1 &lt;device&gt; )
5883
5884</pre>
5885<h5>Overview:</h5>
5886<p>
5887 The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
5888 specific pairs of memory access types.
5889</p>
5890<h5>Arguments:</h5>
5891<p>
5892 The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
5893 The first four arguments enables a specific barrier as listed below. The fith
5894 argument specifies that the barrier applies to io or device or uncached memory.
5895
5896</p>
5897 <ul>
5898 <li><tt>ll</tt>: load-load barrier</li>
5899 <li><tt>ls</tt>: load-store barrier</li>
5900 <li><tt>sl</tt>: store-load barrier</li>
5901 <li><tt>ss</tt>: store-store barrier</li>
Dan Gohmanef9462f2008-10-14 16:51:45 +00005902 <li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00005903 </ul>
5904<h5>Semantics:</h5>
5905<p>
5906 This intrinsic causes the system to enforce some ordering constraints upon
5907 the loads and stores of the program. This barrier does not indicate
5908 <em>when</em> any events will occur, it only enforces an <em>order</em> in
5909 which they occur. For any of the specified pairs of load and store operations
5910 (f.ex. load-load, or store-load), all of the first operations preceding the
5911 barrier will complete before any of the second operations succeeding the
5912 barrier begin. Specifically the semantics for each pairing is as follows:
5913</p>
5914 <ul>
5915 <li><tt>ll</tt>: All loads before the barrier must complete before any load
5916 after the barrier begins.</li>
5917
5918 <li><tt>ls</tt>: All loads before the barrier must complete before any
5919 store after the barrier begins.</li>
5920 <li><tt>ss</tt>: All stores before the barrier must complete before any
5921 store after the barrier begins.</li>
5922 <li><tt>sl</tt>: All stores before the barrier must complete before any
5923 load after the barrier begins.</li>
5924 </ul>
5925<p>
5926 These semantics are applied with a logical "and" behavior when more than one
5927 is enabled in a single memory barrier intrinsic.
5928</p>
5929<p>
5930 Backends may implement stronger barriers than those requested when they do not
5931 support as fine grained a barrier as requested. Some architectures do not
5932 need all types of barriers and on such architectures, these become noops.
5933</p>
5934<h5>Example:</h5>
5935<pre>
5936%ptr = malloc i32
5937 store i32 4, %ptr
5938
5939%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
5940 call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
5941 <i>; guarantee the above finishes</i>
5942 store i32 8, %ptr <i>; before this begins</i>
5943</pre>
5944</div>
5945
Andrew Lenharth95528942008-02-21 06:45:13 +00005946<!-- _______________________________________________________________________ -->
5947<div class="doc_subsubsection">
Mon P Wang6a490372008-06-25 08:15:39 +00005948 <a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharth95528942008-02-21 06:45:13 +00005949</div>
5950<div class="doc_text">
5951<h5>Syntax:</h5>
5952<p>
Mon P Wang2c839d42008-07-30 04:36:53 +00005953 This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
5954 any integer bit width and for different address spaces. Not all targets
5955 support all bit widths however.</p>
Andrew Lenharth95528942008-02-21 06:45:13 +00005956
5957<pre>
Mon P Wang2c839d42008-07-30 04:36:53 +00005958declare i8 @llvm.atomic.cmp.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;cmp&gt;, i8 &lt;val&gt; )
5959declare i16 @llvm.atomic.cmp.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;cmp&gt;, i16 &lt;val&gt; )
5960declare i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;cmp&gt;, i32 &lt;val&gt; )
5961declare i64 @llvm.atomic.cmp.swap.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;cmp&gt;, i64 &lt;val&gt; )
Andrew Lenharth95528942008-02-21 06:45:13 +00005962
5963</pre>
5964<h5>Overview:</h5>
5965<p>
5966 This loads a value in memory and compares it to a given value. If they are
5967 equal, it stores a new value into the memory.
5968</p>
5969<h5>Arguments:</h5>
5970<p>
Mon P Wang6a490372008-06-25 08:15:39 +00005971 The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result as
Andrew Lenharth95528942008-02-21 06:45:13 +00005972 well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
5973 same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
5974 this integer type. While any bit width integer may be used, targets may only
5975 lower representations they support in hardware.
5976
5977</p>
5978<h5>Semantics:</h5>
5979<p>
5980 This entire intrinsic must be executed atomically. It first loads the value
5981 in memory pointed to by <tt>ptr</tt> and compares it with the value
5982 <tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
5983 loaded value is yielded in all cases. This provides the equivalent of an
5984 atomic compare-and-swap operation within the SSA framework.
5985</p>
5986<h5>Examples:</h5>
5987
5988<pre>
5989%ptr = malloc i32
5990 store i32 4, %ptr
5991
5992%val1 = add i32 4, 4
Mon P Wang2c839d42008-07-30 04:36:53 +00005993%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 4, %val1 )
Andrew Lenharth95528942008-02-21 06:45:13 +00005994 <i>; yields {i32}:result1 = 4</i>
5995%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
5996%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
5997
5998%val2 = add i32 1, 1
Mon P Wang2c839d42008-07-30 04:36:53 +00005999%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 5, %val2 )
Andrew Lenharth95528942008-02-21 06:45:13 +00006000 <i>; yields {i32}:result2 = 8</i>
6001%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
6002
6003%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
6004</pre>
6005</div>
6006
6007<!-- _______________________________________________________________________ -->
6008<div class="doc_subsubsection">
6009 <a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
6010</div>
6011<div class="doc_text">
6012<h5>Syntax:</h5>
6013
6014<p>
6015 This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
6016 integer bit width. Not all targets support all bit widths however.</p>
6017<pre>
Mon P Wang2c839d42008-07-30 04:36:53 +00006018declare i8 @llvm.atomic.swap.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;val&gt; )
6019declare i16 @llvm.atomic.swap.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;val&gt; )
6020declare i32 @llvm.atomic.swap.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;val&gt; )
6021declare i64 @llvm.atomic.swap.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;val&gt; )
Andrew Lenharth95528942008-02-21 06:45:13 +00006022
6023</pre>
6024<h5>Overview:</h5>
6025<p>
6026 This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
6027 the value from memory. It then stores the value in <tt>val</tt> in the memory
6028 at <tt>ptr</tt>.
6029</p>
6030<h5>Arguments:</h5>
6031
6032<p>
Mon P Wang6a490372008-06-25 08:15:39 +00006033 The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both the
Andrew Lenharth95528942008-02-21 06:45:13 +00006034 <tt>val</tt> argument and the result must be integers of the same bit width.
6035 The first argument, <tt>ptr</tt>, must be a pointer to a value of this
6036 integer type. The targets may only lower integer representations they
6037 support.
6038</p>
6039<h5>Semantics:</h5>
6040<p>
6041 This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
6042 stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
6043 equivalent of an atomic swap operation within the SSA framework.
6044
6045</p>
6046<h5>Examples:</h5>
6047<pre>
6048%ptr = malloc i32
6049 store i32 4, %ptr
6050
6051%val1 = add i32 4, 4
Mon P Wang2c839d42008-07-30 04:36:53 +00006052%result1 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val1 )
Andrew Lenharth95528942008-02-21 06:45:13 +00006053 <i>; yields {i32}:result1 = 4</i>
6054%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
6055%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
6056
6057%val2 = add i32 1, 1
Mon P Wang2c839d42008-07-30 04:36:53 +00006058%result2 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val2 )
Andrew Lenharth95528942008-02-21 06:45:13 +00006059 <i>; yields {i32}:result2 = 8</i>
6060
6061%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
6062%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
6063</pre>
6064</div>
6065
6066<!-- _______________________________________________________________________ -->
6067<div class="doc_subsubsection">
Mon P Wang6a490372008-06-25 08:15:39 +00006068 <a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharth95528942008-02-21 06:45:13 +00006069
6070</div>
6071<div class="doc_text">
6072<h5>Syntax:</h5>
6073<p>
Mon P Wang6a490372008-06-25 08:15:39 +00006074 This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on any
Andrew Lenharth95528942008-02-21 06:45:13 +00006075 integer bit width. Not all targets support all bit widths however.</p>
6076<pre>
Mon P Wang2c839d42008-07-30 04:36:53 +00006077declare i8 @llvm.atomic.load.add.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6078declare i16 @llvm.atomic.load.add.i16..p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6079declare i32 @llvm.atomic.load.add.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6080declare i64 @llvm.atomic.load.add.i64..p0i64( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Andrew Lenharth95528942008-02-21 06:45:13 +00006081
6082</pre>
6083<h5>Overview:</h5>
6084<p>
6085 This intrinsic adds <tt>delta</tt> to the value stored in memory at
6086 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6087</p>
6088<h5>Arguments:</h5>
6089<p>
6090
6091 The intrinsic takes two arguments, the first a pointer to an integer value
6092 and the second an integer value. The result is also an integer value. These
6093 integer types can have any bit width, but they must all have the same bit
6094 width. The targets may only lower integer representations they support.
6095</p>
6096<h5>Semantics:</h5>
6097<p>
6098 This intrinsic does a series of operations atomically. It first loads the
6099 value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
6100 to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6101</p>
6102
6103<h5>Examples:</h5>
6104<pre>
6105%ptr = malloc i32
6106 store i32 4, %ptr
Mon P Wang2c839d42008-07-30 04:36:53 +00006107%result1 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 4 )
Andrew Lenharth95528942008-02-21 06:45:13 +00006108 <i>; yields {i32}:result1 = 4</i>
Mon P Wang2c839d42008-07-30 04:36:53 +00006109%result2 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 2 )
Andrew Lenharth95528942008-02-21 06:45:13 +00006110 <i>; yields {i32}:result2 = 8</i>
Mon P Wang2c839d42008-07-30 04:36:53 +00006111%result3 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 5 )
Andrew Lenharth95528942008-02-21 06:45:13 +00006112 <i>; yields {i32}:result3 = 10</i>
Mon P Wang6a490372008-06-25 08:15:39 +00006113%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharth95528942008-02-21 06:45:13 +00006114</pre>
6115</div>
6116
Mon P Wang6a490372008-06-25 08:15:39 +00006117<!-- _______________________________________________________________________ -->
6118<div class="doc_subsubsection">
6119 <a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
6120
6121</div>
6122<div class="doc_text">
6123<h5>Syntax:</h5>
6124<p>
6125 This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
Mon P Wang2c839d42008-07-30 04:36:53 +00006126 any integer bit width and for different address spaces. Not all targets
6127 support all bit widths however.</p>
Mon P Wang6a490372008-06-25 08:15:39 +00006128<pre>
Mon P Wang2c839d42008-07-30 04:36:53 +00006129declare i8 @llvm.atomic.load.sub.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6130declare i16 @llvm.atomic.load.sub.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6131declare i32 @llvm.atomic.load.sub.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6132declare i64 @llvm.atomic.load.sub.i64.p0i32( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6a490372008-06-25 08:15:39 +00006133
6134</pre>
6135<h5>Overview:</h5>
6136<p>
6137 This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
6138 <tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
6139</p>
6140<h5>Arguments:</h5>
6141<p>
6142
6143 The intrinsic takes two arguments, the first a pointer to an integer value
6144 and the second an integer value. The result is also an integer value. These
6145 integer types can have any bit width, but they must all have the same bit
6146 width. The targets may only lower integer representations they support.
6147</p>
6148<h5>Semantics:</h5>
6149<p>
6150 This intrinsic does a series of operations atomically. It first loads the
6151 value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
6152 result to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
6153</p>
6154
6155<h5>Examples:</h5>
6156<pre>
6157%ptr = malloc i32
6158 store i32 8, %ptr
Mon P Wang2c839d42008-07-30 04:36:53 +00006159%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 4 )
Mon P Wang6a490372008-06-25 08:15:39 +00006160 <i>; yields {i32}:result1 = 8</i>
Mon P Wang2c839d42008-07-30 04:36:53 +00006161%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 2 )
Mon P Wang6a490372008-06-25 08:15:39 +00006162 <i>; yields {i32}:result2 = 4</i>
Mon P Wang2c839d42008-07-30 04:36:53 +00006163%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 5 )
Mon P Wang6a490372008-06-25 08:15:39 +00006164 <i>; yields {i32}:result3 = 2</i>
6165%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
6166</pre>
6167</div>
6168
6169<!-- _______________________________________________________________________ -->
6170<div class="doc_subsubsection">
6171 <a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
6172 <a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
6173 <a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
6174 <a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
6175
6176</div>
6177<div class="doc_text">
6178<h5>Syntax:</h5>
6179<p>
6180 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_and</tt>,
6181 <tt>llvm.atomic.load_nand</tt>, <tt>llvm.atomic.load_or</tt>, and
Mon P Wang2c839d42008-07-30 04:36:53 +00006182 <tt>llvm.atomic.load_xor</tt> on any integer bit width and for different
6183 address spaces. Not all targets support all bit widths however.</p>
Mon P Wang6a490372008-06-25 08:15:39 +00006184<pre>
Mon P Wang2c839d42008-07-30 04:36:53 +00006185declare i8 @llvm.atomic.load.and.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6186declare i16 @llvm.atomic.load.and.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6187declare i32 @llvm.atomic.load.and.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6188declare i64 @llvm.atomic.load.and.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6a490372008-06-25 08:15:39 +00006189
6190</pre>
6191
6192<pre>
Mon P Wang2c839d42008-07-30 04:36:53 +00006193declare i8 @llvm.atomic.load.or.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6194declare i16 @llvm.atomic.load.or.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6195declare i32 @llvm.atomic.load.or.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6196declare i64 @llvm.atomic.load.or.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6a490372008-06-25 08:15:39 +00006197
6198</pre>
6199
6200<pre>
Mon P Wang2c839d42008-07-30 04:36:53 +00006201declare i8 @llvm.atomic.load.nand.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6202declare i16 @llvm.atomic.load.nand.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6203declare i32 @llvm.atomic.load.nand.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6204declare i64 @llvm.atomic.load.nand.i64.p0i32( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6a490372008-06-25 08:15:39 +00006205
6206</pre>
6207
6208<pre>
Mon P Wang2c839d42008-07-30 04:36:53 +00006209declare i8 @llvm.atomic.load.xor.i8.p0i32( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6210declare i16 @llvm.atomic.load.xor.i16.p0i32( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6211declare i32 @llvm.atomic.load.xor.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6212declare i64 @llvm.atomic.load.xor.i64.p0i32( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6a490372008-06-25 08:15:39 +00006213
6214</pre>
6215<h5>Overview:</h5>
6216<p>
6217 These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
6218 the value stored in memory at <tt>ptr</tt>. It yields the original value
6219 at <tt>ptr</tt>.
6220</p>
6221<h5>Arguments:</h5>
6222<p>
6223
6224 These intrinsics take two arguments, the first a pointer to an integer value
6225 and the second an integer value. The result is also an integer value. These
6226 integer types can have any bit width, but they must all have the same bit
6227 width. The targets may only lower integer representations they support.
6228</p>
6229<h5>Semantics:</h5>
6230<p>
6231 These intrinsics does a series of operations atomically. They first load the
6232 value stored at <tt>ptr</tt>. They then do the bitwise operation
6233 <tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the original
6234 value stored at <tt>ptr</tt>.
6235</p>
6236
6237<h5>Examples:</h5>
6238<pre>
6239%ptr = malloc i32
6240 store i32 0x0F0F, %ptr
Mon P Wang2c839d42008-07-30 04:36:53 +00006241%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6a490372008-06-25 08:15:39 +00006242 <i>; yields {i32}:result0 = 0x0F0F</i>
Mon P Wang2c839d42008-07-30 04:36:53 +00006243%result1 = call i32 @llvm.atomic.load.and.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang6a490372008-06-25 08:15:39 +00006244 <i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Mon P Wang2c839d42008-07-30 04:36:53 +00006245%result2 = call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6a490372008-06-25 08:15:39 +00006246 <i>; yields {i32}:result2 = 0xF0</i>
Mon P Wang2c839d42008-07-30 04:36:53 +00006247%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang6a490372008-06-25 08:15:39 +00006248 <i>; yields {i32}:result3 = FF</i>
6249%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
6250</pre>
6251</div>
6252
6253
6254<!-- _______________________________________________________________________ -->
6255<div class="doc_subsubsection">
6256 <a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
6257 <a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
6258 <a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
6259 <a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
6260
6261</div>
6262<div class="doc_text">
6263<h5>Syntax:</h5>
6264<p>
6265 These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
6266 <tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
Mon P Wang2c839d42008-07-30 04:36:53 +00006267 <tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
6268 address spaces. Not all targets
Mon P Wang6a490372008-06-25 08:15:39 +00006269 support all bit widths however.</p>
6270<pre>
Mon P Wang2c839d42008-07-30 04:36:53 +00006271declare i8 @llvm.atomic.load.max.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6272declare i16 @llvm.atomic.load.max.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6273declare i32 @llvm.atomic.load.max.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6274declare i64 @llvm.atomic.load.max.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6a490372008-06-25 08:15:39 +00006275
6276</pre>
6277
6278<pre>
Mon P Wang2c839d42008-07-30 04:36:53 +00006279declare i8 @llvm.atomic.load.min.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6280declare i16 @llvm.atomic.load.min.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6281declare i32 @llvm.atomic.load.min.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6282declare i64 @llvm.atomic.load.min.i64..p0i64( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6a490372008-06-25 08:15:39 +00006283
6284</pre>
6285
6286<pre>
Mon P Wang2c839d42008-07-30 04:36:53 +00006287declare i8 @llvm.atomic.load.umax.i8.p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6288declare i16 @llvm.atomic.load.umax.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6289declare i32 @llvm.atomic.load.umax.i32.p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6290declare i64 @llvm.atomic.load.umax.i64.p0i64( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6a490372008-06-25 08:15:39 +00006291
6292</pre>
6293
6294<pre>
Mon P Wang2c839d42008-07-30 04:36:53 +00006295declare i8 @llvm.atomic.load.umin.i8..p0i8( i8* &lt;ptr&gt;, i8 &lt;delta&gt; )
6296declare i16 @llvm.atomic.load.umin.i16.p0i16( i16* &lt;ptr&gt;, i16 &lt;delta&gt; )
6297declare i32 @llvm.atomic.load.umin.i32..p0i32( i32* &lt;ptr&gt;, i32 &lt;delta&gt; )
6298declare i64 @llvm.atomic.load.umin.i64..p0i64( i64* &lt;ptr&gt;, i64 &lt;delta&gt; )
Mon P Wang6a490372008-06-25 08:15:39 +00006299
6300</pre>
6301<h5>Overview:</h5>
6302<p>
6303 These intrinsics takes the signed or unsigned minimum or maximum of
6304 <tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
6305 original value at <tt>ptr</tt>.
6306</p>
6307<h5>Arguments:</h5>
6308<p>
6309
6310 These intrinsics take two arguments, the first a pointer to an integer value
6311 and the second an integer value. The result is also an integer value. These
6312 integer types can have any bit width, but they must all have the same bit
6313 width. The targets may only lower integer representations they support.
6314</p>
6315<h5>Semantics:</h5>
6316<p>
6317 These intrinsics does a series of operations atomically. They first load the
6318 value stored at <tt>ptr</tt>. They then do the signed or unsigned min or max
6319 <tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They yield
6320 the original value stored at <tt>ptr</tt>.
6321</p>
6322
6323<h5>Examples:</h5>
6324<pre>
6325%ptr = malloc i32
6326 store i32 7, %ptr
Mon P Wang2c839d42008-07-30 04:36:53 +00006327%result0 = call i32 @llvm.atomic.load.min.i32.p0i32( i32* %ptr, i32 -2 )
Mon P Wang6a490372008-06-25 08:15:39 +00006328 <i>; yields {i32}:result0 = 7</i>
Mon P Wang2c839d42008-07-30 04:36:53 +00006329%result1 = call i32 @llvm.atomic.load.max.i32.p0i32( i32* %ptr, i32 8 )
Mon P Wang6a490372008-06-25 08:15:39 +00006330 <i>; yields {i32}:result1 = -2</i>
Mon P Wang2c839d42008-07-30 04:36:53 +00006331%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32( i32* %ptr, i32 10 )
Mon P Wang6a490372008-06-25 08:15:39 +00006332 <i>; yields {i32}:result2 = 8</i>
Mon P Wang2c839d42008-07-30 04:36:53 +00006333%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32( i32* %ptr, i32 30 )
Mon P Wang6a490372008-06-25 08:15:39 +00006334 <i>; yields {i32}:result3 = 8</i>
6335%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
6336</pre>
6337</div>
Andrew Lenharth9b254ee2008-02-16 01:24:58 +00006338
6339<!-- ======================================================================= -->
6340<div class="doc_subsection">
Tanya Lattnercb1b9602007-06-15 20:50:54 +00006341 <a name="int_general">General Intrinsics</a>
6342</div>
6343
6344<div class="doc_text">
6345<p> This class of intrinsics is designed to be generic and has
6346no specific purpose. </p>
6347</div>
6348
6349<!-- _______________________________________________________________________ -->
6350<div class="doc_subsubsection">
6351 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
6352</div>
6353
6354<div class="doc_text">
6355
6356<h5>Syntax:</h5>
6357<pre>
Tanya Lattnerbed1d4d2007-06-18 23:42:37 +00006358 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
Tanya Lattnercb1b9602007-06-15 20:50:54 +00006359</pre>
6360
6361<h5>Overview:</h5>
6362
6363<p>
6364The '<tt>llvm.var.annotation</tt>' intrinsic
6365</p>
6366
6367<h5>Arguments:</h5>
6368
6369<p>
Tanya Lattnerbed1d4d2007-06-18 23:42:37 +00006370The first argument is a pointer to a value, the second is a pointer to a
6371global string, the third is a pointer to a global string which is the source
6372file name, and the last argument is the line number.
Tanya Lattnercb1b9602007-06-15 20:50:54 +00006373</p>
6374
6375<h5>Semantics:</h5>
6376
6377<p>
Anton Korobeynikov06cbb652008-01-15 22:31:34 +00006378This intrinsic allows annotation of local variables with arbitrary strings.
Tanya Lattnercb1b9602007-06-15 20:50:54 +00006379This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikov06cbb652008-01-15 22:31:34 +00006380annotations. These have no other defined use, they are ignored by code
6381generation and optimization.
6382</p>
Tanya Lattnercb1b9602007-06-15 20:50:54 +00006383</div>
6384
Tanya Lattner293c0372007-09-21 22:59:12 +00006385<!-- _______________________________________________________________________ -->
6386<div class="doc_subsubsection">
Tanya Lattner0186a652007-09-21 23:57:59 +00006387 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattner293c0372007-09-21 22:59:12 +00006388</div>
6389
6390<div class="doc_text">
6391
6392<h5>Syntax:</h5>
Tanya Lattner23dbd572007-09-21 23:56:27 +00006393<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
6394any integer bit width.
6395</p>
Tanya Lattner293c0372007-09-21 22:59:12 +00006396<pre>
Tanya Lattnercf3e26f2007-09-22 00:03:01 +00006397 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6398 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6399 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6400 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
6401 declare i256 @llvm.annotation.i256(i256 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
Tanya Lattner293c0372007-09-21 22:59:12 +00006402</pre>
6403
6404<h5>Overview:</h5>
Tanya Lattner23dbd572007-09-21 23:56:27 +00006405
6406<p>
6407The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattner293c0372007-09-21 22:59:12 +00006408</p>
6409
6410<h5>Arguments:</h5>
6411
6412<p>
6413The first argument is an integer value (result of some expression),
6414the second is a pointer to a global string, the third is a pointer to a global
6415string which is the source file name, and the last argument is the line number.
Tanya Lattner23dbd572007-09-21 23:56:27 +00006416It returns the value of the first argument.
Tanya Lattner293c0372007-09-21 22:59:12 +00006417</p>
6418
6419<h5>Semantics:</h5>
6420
6421<p>
6422This intrinsic allows annotations to be put on arbitrary expressions
6423with arbitrary strings. This can be useful for special purpose optimizations
6424that want to look for these annotations. These have no other defined use, they
6425are ignored by code generation and optimization.
Dan Gohmanef9462f2008-10-14 16:51:45 +00006426</p>
Tanya Lattner293c0372007-09-21 22:59:12 +00006427</div>
Jim Laskey2211f492007-03-14 19:31:19 +00006428
Anton Korobeynikov06cbb652008-01-15 22:31:34 +00006429<!-- _______________________________________________________________________ -->
6430<div class="doc_subsubsection">
6431 <a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
6432</div>
6433
6434<div class="doc_text">
6435
6436<h5>Syntax:</h5>
6437<pre>
6438 declare void @llvm.trap()
6439</pre>
6440
6441<h5>Overview:</h5>
6442
6443<p>
6444The '<tt>llvm.trap</tt>' intrinsic
6445</p>
6446
6447<h5>Arguments:</h5>
6448
6449<p>
6450None
6451</p>
6452
6453<h5>Semantics:</h5>
6454
6455<p>
6456This intrinsics is lowered to the target dependent trap instruction. If the
6457target does not have a trap instruction, this intrinsic will be lowered to the
6458call of the abort() function.
6459</p>
6460</div>
6461
Bill Wendling14313312008-11-19 05:56:17 +00006462<!-- _______________________________________________________________________ -->
6463<div class="doc_subsubsection">
Misha Brukman50de2b22008-11-22 23:55:29 +00006464 <a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
Bill Wendling14313312008-11-19 05:56:17 +00006465</div>
6466<div class="doc_text">
6467<h5>Syntax:</h5>
6468<pre>
6469declare void @llvm.stackprotector( i8* &lt;guard&gt;, i8** &lt;slot&gt; )
6470
6471</pre>
6472<h5>Overview:</h5>
6473<p>
6474 The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and stores
6475 it onto the stack at <tt>slot</tt>. The stack slot is adjusted to ensure that
6476 it is placed on the stack before local variables.
6477</p>
6478<h5>Arguments:</h5>
6479<p>
6480 The <tt>llvm.stackprotector</tt> intrinsic requires two pointer arguments. The
6481 first argument is the value loaded from the stack guard
6482 <tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt> that
6483 has enough space to hold the value of the guard.
6484</p>
6485<h5>Semantics:</h5>
6486<p>
6487 This intrinsic causes the prologue/epilogue inserter to force the position of
6488 the <tt>AllocaInst</tt> stack slot to be before local variables on the
6489 stack. This is to ensure that if a local variable on the stack is overwritten,
6490 it will destroy the value of the guard. When the function exits, the guard on
6491 the stack is checked against the original guard. If they're different, then
6492 the program aborts by calling the <tt>__stack_chk_fail()</tt> function.
6493</p>
6494</div>
6495
Chris Lattner2f7c9632001-06-06 20:29:01 +00006496<!-- *********************************************************************** -->
Chris Lattner2f7c9632001-06-06 20:29:01 +00006497<hr>
Misha Brukmanc501f552004-03-01 17:47:27 +00006498<address>
6499 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +00006500 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukmanc501f552004-03-01 17:47:27 +00006501 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +00006502 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Misha Brukmanc501f552004-03-01 17:47:27 +00006503
6504 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencerca058542006-03-14 05:39:39 +00006505 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukmanc501f552004-03-01 17:47:27 +00006506 Last modified: $Date$
6507</address>
Chris Lattnerb8f816e2008-01-04 04:33:49 +00006508
Misha Brukman76307852003-11-08 01:05:38 +00006509</body>
6510</html>