Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / cc68939ba63e840b317789e10c1fb4645d95c328 / . / docs / LangRef.html
blob: 308dd207e5d51158a2f2ac5d77047068d19918d1 [file] [log] [blame]
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencer3921c742004-08-26 20:44:00 +0000[diff] [blame]6 <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 Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]10 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]12
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]13<body>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]14
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]15<div class="doc_title"> LLVM Language Reference Manual </div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]16<ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]17 <li><a href="#abstract">Abstract</a></li>
18 <li><a href="#introduction">Introduction</a></li>
19 <li><a href="#identifiers">Identifiers</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]20 <li><a href="#highlevel">High Level Structure</a>
21 <ol>
22 <li><a href="#modulestructure">Module Structure</a></li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]23 <li><a href="#linkage">Linkage Types</a></li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]24 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]25 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]26 <li><a href="#functionstructure">Functions</a></li>
Anton Korobeynikovc6c98af2007-04-29 18:02:48 +0000[diff] [blame]27 <li><a href="#aliasstructure">Aliases</a>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]28 <li><a href="#paramattrs">Parameter Attributes</a></li>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]29 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]30 <li><a href="#datalayout">Data Layout</a></li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]31 </ol>
32 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]33 <li><a href="#typesystem">Type System</a>
34 <ol>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]35 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]36 <ol>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]37 <li><a href="#t_classifications">Type Classifications</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]38 </ol>
39 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]40 <li><a href="#t_derived">Derived Types</a>
41 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]42 <li><a href="#t_array">Array Type</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]43 <li><a href="#t_function">Function Type</a></li>
44 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]45 <li><a href="#t_struct">Structure Type</a></li>
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]46 <li><a href="#t_pstruct">Packed Structure Type</a></li>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]47 <li><a href="#t_vector">Vector Type</a></li>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]48 <li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]49 </ol>
50 </li>
51 </ol>
52 </li>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]53 <li><a href="#constants">Constants</a>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]54 <ol>
55 <li><a href="#simpleconstants">Simple Constants</a>
56 <li><a href="#aggregateconstants">Aggregate Constants</a>
57 <li><a href="#globalconstants">Global Variable and Function Addresses</a>
58 <li><a href="#undefvalues">Undefined Values</a>
59 <li><a href="#constantexprs">Constant Expressions</a>
60 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]61 </li>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]62 <li><a href="#othervalues">Other Values</a>
63 <ol>
64 <li><a href="#inlineasm">Inline Assembler Expressions</a>
65 </ol>
66 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]67 <li><a href="#instref">Instruction Reference</a>
68 <ol>
69 <li><a href="#terminators">Terminator Instructions</a>
70 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]71 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
72 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]73 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
74 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]75 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]76 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]77 </ol>
78 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]79 <li><a href="#binaryops">Binary Operations</a>
80 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]81 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
82 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
83 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]84 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
85 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
86 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]87 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
88 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
89 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]90 </ol>
91 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]92 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
93 <ol>
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]94 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
95 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
96 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]97 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]98 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]99 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]100 </ol>
101 </li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]102 <li><a href="#vectorops">Vector Operations</a>
103 <ol>
104 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
105 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
106 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]107 </ol>
108 </li>
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]109 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]110 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]111 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
112 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
113 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]114 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
115 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
116 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]117 </ol>
118 </li>
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]119 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]120 <ol>
121 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
122 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
123 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
124 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
125 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]126 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
127 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
128 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
129 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]130 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
131 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]132 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]133 </ol>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]134 <li><a href="#otherops">Other Operations</a>
135 <ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]136 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
137 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]138 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]139 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]140 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]141 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]142 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]143 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]144 </ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]145 </li>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]146 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]147 <ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]148 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
149 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]150 <li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
151 <li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
152 <li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]153 </ol>
154 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]155 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
156 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]157 <li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
158 <li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
159 <li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]160 </ol>
161 </li>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]162 <li><a href="#int_codegen">Code Generator Intrinsics</a>
163 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]164 <li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
165 <li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
166 <li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
167 <li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
168 <li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
169 <li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
170 <li><a href="#int_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell7123e272004-04-09 16:43:20 +0000[diff] [blame]171 </ol>
172 </li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]173 <li><a href="#int_libc">Standard C Library Intrinsics</a>
174 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]175 <li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
176 <li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
177 <li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
178 <li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
179 <li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]180 </ol>
181 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]182 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]183 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]184 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]185 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
186 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
187 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]188 <li><a href="#int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic </a></li>
189 <li><a href="#int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]190 </ol>
191 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]192 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]193 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]194 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]195 <ol>
196 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]197 </ol>
198 </li>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]199 <li><a href="#int_general">General intrinsics</a>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]200 <ol>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]201 <li><a href="#int_var_annotation">
Tanya Lattner91d0b882007-09-22 00:01:26 +0000[diff] [blame]202 <tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]203 </ol>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]204 <ol>
205 <li><a href="#int_annotation">
Tanya Lattner91d0b882007-09-22 00:01:26 +0000[diff] [blame]206 <tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]207 </ol>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]208 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]209 </ol>
210 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]211</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]212
213<div class="doc_author">
214 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
215 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]216</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]217
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]218<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]219<div class="doc_section"> <a name="abstract">Abstract </a></div>
220<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]221
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]222<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]223<p>This document is a reference manual for the LLVM assembly language.
224LLVM is an SSA based representation that provides type safety,
225low-level operations, flexibility, and the capability of representing
226'all' high-level languages cleanly. It is the common code
227representation used throughout all phases of the LLVM compilation
228strategy.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]229</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]230
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]231<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]232<div class="doc_section"> <a name="introduction">Introduction</a> </div>
233<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]234
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]235<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]236
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]237<p>The LLVM code representation is designed to be used in three
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]238different forms: as an in-memory compiler IR, as an on-disk bitcode
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]239representation (suitable for fast loading by a Just-In-Time compiler),
240and as a human readable assembly language representation. This allows
241LLVM to provide a powerful intermediate representation for efficient
242compiler transformations and analysis, while providing a natural means
243to debug and visualize the transformations. The three different forms
244of LLVM are all equivalent. This document describes the human readable
245representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]246
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]247<p>The LLVM representation aims to be light-weight and low-level
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]248while being expressive, typed, and extensible at the same time. It
249aims to be a "universal IR" of sorts, by being at a low enough level
250that high-level ideas may be cleanly mapped to it (similar to how
251microprocessors are "universal IR's", allowing many source languages to
252be mapped to them). By providing type information, LLVM can be used as
253the target of optimizations: for example, through pointer analysis, it
254can be proven that a C automatic variable is never accessed outside of
255the current function... allowing it to be promoted to a simple SSA
256value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]257
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]258</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]259
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]260<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]261<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]262
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]263<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]264
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]265<p>It is important to note that this document describes 'well formed'
266LLVM assembly language. There is a difference between what the parser
267accepts and what is considered 'well formed'. For example, the
268following instruction is syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]269
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]270<div class="doc_code">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]271<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]272%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]273</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]274</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]275
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]276<p>...because the definition of <tt>%x</tt> does not dominate all of
277its uses. The LLVM infrastructure provides a verification pass that may
278be used to verify that an LLVM module is well formed. This pass is
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]279automatically run by the parser after parsing input assembly and by
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]280the optimizer before it outputs bitcode. The violations pointed out
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]281by the verifier pass indicate bugs in transformation passes or input to
282the parser.</p>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]283</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]284
Chris Lattnercc689392007-10-03 17:34:29 +0000[diff] [blame^]285<!-- Describe the typesetting conventions here. -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]286
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]287<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]288<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]289<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]290
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]291<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]292
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]293 <p>LLVM identifiers come in two basic types: global and local. Global
294 identifiers (functions, global variables) begin with the @ character. Local
295 identifiers (register names, types) begin with the % character. Additionally,
296 there are three different formats for identifiers, for different purposes:
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]297
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]298<ol>
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]299 <li>Named values are represented as a string of characters with their prefix.
300 For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
301 regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]302 Identifiers which require other characters in their names can be surrounded
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]303 with quotes. In this way, anything except a <tt>&quot;</tt> character can
304 be used in a named value.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]305
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]306 <li>Unnamed values are represented as an unsigned numeric value with their
307 prefix. For example, %12, @2, %44.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]308
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]309 <li>Constants, which are described in a <a href="#constants">section about
310 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]311</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]312
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]313<p>LLVM requires that values start with a prefix for two reasons: Compilers
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]314don't need to worry about name clashes with reserved words, and the set of
315reserved words may be expanded in the future without penalty. Additionally,
316unnamed identifiers allow a compiler to quickly come up with a temporary
317variable without having to avoid symbol table conflicts.</p>
318
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]319<p>Reserved words in LLVM are very similar to reserved words in other
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]320languages. There are keywords for different opcodes
321('<tt><a href="#i_add">add</a></tt>',
322 '<tt><a href="#i_bitcast">bitcast</a></tt>',
323 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]324href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]325and others. These reserved words cannot conflict with variable names, because
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]326none of them start with a prefix character ('%' or '@').</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]327
328<p>Here is an example of LLVM code to multiply the integer variable
329'<tt>%X</tt>' by 8:</p>
330
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]331<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]332
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]333<div class="doc_code">
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]334<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]335%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]336</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]337</div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]338
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]339<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]340
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]341<div class="doc_code">
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]342<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]343%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]344</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]345</div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]346
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]347<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]348
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]349<div class="doc_code">
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]350<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]351<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
352<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
353%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]354</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]355</div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]356
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]357<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
358important lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]359
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]360<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]361
362 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
363 line.</li>
364
365 <li>Unnamed temporaries are created when the result of a computation is not
366 assigned to a named value.</li>
367
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]368 <li>Unnamed temporaries are numbered sequentially</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]369
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]370</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]371
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]372<p>...and it also shows a convention that we follow in this document. When
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]373demonstrating instructions, we will follow an instruction with a comment that
374defines the type and name of value produced. Comments are shown in italic
375text.</p>
376
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]377</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]378
379<!-- *********************************************************************** -->
380<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
381<!-- *********************************************************************** -->
382
383<!-- ======================================================================= -->
384<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
385</div>
386
387<div class="doc_text">
388
389<p>LLVM programs are composed of "Module"s, each of which is a
390translation unit of the input programs. Each module consists of
391functions, global variables, and symbol table entries. Modules may be
392combined together with the LLVM linker, which merges function (and
393global variable) definitions, resolves forward declarations, and merges
394symbol table entries. Here is an example of the "hello world" module:</p>
395
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]396<div class="doc_code">
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]397<pre><i>; Declare the string constant as a global constant...</i>
Chris Lattnera89e5f12007-06-12 17:00:26 +0000[diff] [blame]398<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
399 href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]400
401<i>; External declaration of the puts function</i>
Chris Lattnera89e5f12007-06-12 17:00:26 +0000[diff] [blame]402<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]403
404<i>; Definition of main function</i>
Chris Lattnera89e5f12007-06-12 17:00:26 +0000[diff] [blame]405define i32 @main() { <i>; i32()* </i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]406 <i>; Convert [13x i8 ]* to i8 *...</i>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]407 %cast210 = <a
Chris Lattner6c0955b2007-06-12 17:01:15 +0000[diff] [blame]408 href="#i_getelementptr">getelementptr</a> [13 x i8 ]* @.LC0, i64 0, i64 0 <i>; i8 *</i>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]409
410 <i>; Call puts function to write out the string to stdout...</i>
411 <a
Chris Lattnera89e5f12007-06-12 17:00:26 +0000[diff] [blame]412 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]413 <a
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]414 href="#i_ret">ret</a> i32 0<br>}<br>
415</pre>
416</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]417
418<p>This example is made up of a <a href="#globalvars">global variable</a>
419named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
420function, and a <a href="#functionstructure">function definition</a>
421for "<tt>main</tt>".</p>
422
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]423<p>In general, a module is made up of a list of global values,
424where both functions and global variables are global values. Global values are
425represented by a pointer to a memory location (in this case, a pointer to an
426array of char, and a pointer to a function), and have one of the following <a
427href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]428
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]429</div>
430
431<!-- ======================================================================= -->
432<div class="doc_subsection">
433 <a name="linkage">Linkage Types</a>
434</div>
435
436<div class="doc_text">
437
438<p>
439All Global Variables and Functions have one of the following types of linkage:
440</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]441
442<dl>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]443
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]444 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]445
446 <dd>Global values with internal linkage are only directly accessible by
447 objects in the current module. In particular, linking code into a module with
448 an internal global value may cause the internal to be renamed as necessary to
449 avoid collisions. Because the symbol is internal to the module, all
450 references can be updated. This corresponds to the notion of the
Chris Lattner4887bd82007-01-14 06:51:48 +0000[diff] [blame]451 '<tt>static</tt>' keyword in C.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]452 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]453
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]454 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]455
Chris Lattner4887bd82007-01-14 06:51:48 +0000[diff] [blame]456 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
457 the same name when linkage occurs. This is typically used to implement
458 inline functions, templates, or other code which must be generated in each
459 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
460 allowed to be discarded.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]461 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]462
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]463 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]464
465 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
466 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
Chris Lattner4887bd82007-01-14 06:51:48 +0000[diff] [blame]467 used for globals that may be emitted in multiple translation units, but that
468 are not guaranteed to be emitted into every translation unit that uses them.
469 One example of this are common globals in C, such as "<tt>int X;</tt>" at
470 global scope.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]471 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]472
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]473 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]474
475 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
476 pointer to array type. When two global variables with appending linkage are
477 linked together, the two global arrays are appended together. This is the
478 LLVM, typesafe, equivalent of having the system linker append together
479 "sections" with identical names when .o files are linked.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]480 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]481
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]482 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
483 <dd>The semantics of this linkage follow the ELF model: the symbol is weak
484 until linked, if not linked, the symbol becomes null instead of being an
485 undefined reference.
486 </dd>
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]487
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]488 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]489
490 <dd>If none of the above identifiers are used, the global is externally
491 visible, meaning that it participates in linkage and can be used to resolve
492 external symbol references.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]493 </dd>
Reid Spencerc8910842007-04-11 23:49:50 +0000[diff] [blame]494</dl>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]495
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]496 <p>
497 The next two types of linkage are targeted for Microsoft Windows platform
498 only. They are designed to support importing (exporting) symbols from (to)
499 DLLs.
500 </p>
501
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]502 <dl>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]503 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
504
505 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
506 or variable via a global pointer to a pointer that is set up by the DLL
507 exporting the symbol. On Microsoft Windows targets, the pointer name is
508 formed by combining <code>_imp__</code> and the function or variable name.
509 </dd>
510
511 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
512
513 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
514 pointer to a pointer in a DLL, so that it can be referenced with the
515 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
516 name is formed by combining <code>_imp__</code> and the function or variable
517 name.
518 </dd>
519
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]520</dl>
521
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]522<p><a name="linkage_external"></a>For example, since the "<tt>.LC0</tt>"
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]523variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
524variable and was linked with this one, one of the two would be renamed,
525preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
526external (i.e., lacking any linkage declarations), they are accessible
Reid Spencerac8d2762007-01-05 00:59:10 +0000[diff] [blame]527outside of the current module.</p>
528<p>It is illegal for a function <i>declaration</i>
529to have any linkage type other than "externally visible", <tt>dllimport</tt>,
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]530or <tt>extern_weak</tt>.</p>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]531<p>Aliases can have only <tt>external</tt>, <tt>internal</tt> and <tt>weak</tt>
532linkages.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]533</div>
534
535<!-- ======================================================================= -->
536<div class="doc_subsection">
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]537 <a name="callingconv">Calling Conventions</a>
538</div>
539
540<div class="doc_text">
541
542<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
543and <a href="#i_invoke">invokes</a> can all have an optional calling convention
544specified for the call. The calling convention of any pair of dynamic
545caller/callee must match, or the behavior of the program is undefined. The
546following calling conventions are supported by LLVM, and more may be added in
547the future:</p>
548
549<dl>
550 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
551
552 <dd>This calling convention (the default if no other calling convention is
553 specified) matches the target C calling conventions. This calling convention
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]554 supports varargs function calls and tolerates some mismatch in the declared
Reid Spencerc28d2bc2006-12-31 21:30:18 +0000[diff] [blame]555 prototype and implemented declaration of the function (as does normal C).
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]556 </dd>
557
558 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
559
560 <dd>This calling convention attempts to make calls as fast as possible
561 (e.g. by passing things in registers). This calling convention allows the
562 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner8cdc5bc2005-05-06 23:08:23 +0000[diff] [blame]563 without having to conform to an externally specified ABI. Implementations of
564 this convention should allow arbitrary tail call optimization to be supported.
565 This calling convention does not support varargs and requires the prototype of
566 all callees to exactly match the prototype of the function definition.
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]567 </dd>
568
569 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
570
571 <dd>This calling convention attempts to make code in the caller as efficient
572 as possible under the assumption that the call is not commonly executed. As
573 such, these calls often preserve all registers so that the call does not break
574 any live ranges in the caller side. This calling convention does not support
575 varargs and requires the prototype of all callees to exactly match the
576 prototype of the function definition.
577 </dd>
578
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]579 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]580
581 <dd>Any calling convention may be specified by number, allowing
582 target-specific calling conventions to be used. Target specific calling
583 conventions start at 64.
584 </dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]585</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]586
587<p>More calling conventions can be added/defined on an as-needed basis, to
588support pascal conventions or any other well-known target-independent
589convention.</p>
590
591</div>
592
593<!-- ======================================================================= -->
594<div class="doc_subsection">
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]595 <a name="visibility">Visibility Styles</a>
596</div>
597
598<div class="doc_text">
599
600<p>
601All Global Variables and Functions have one of the following visibility styles:
602</p>
603
604<dl>
605 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
606
607 <dd>On ELF, default visibility means that the declaration is visible to other
608 modules and, in shared libraries, means that the declared entity may be
609 overridden. On Darwin, default visibility means that the declaration is
610 visible to other modules. Default visibility corresponds to "external
611 linkage" in the language.
612 </dd>
613
614 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
615
616 <dd>Two declarations of an object with hidden visibility refer to the same
617 object if they are in the same shared object. Usually, hidden visibility
618 indicates that the symbol will not be placed into the dynamic symbol table,
619 so no other module (executable or shared library) can reference it
620 directly.
621 </dd>
622
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000[diff] [blame]623 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
624
625 <dd>On ELF, protected visibility indicates that the symbol will be placed in
626 the dynamic symbol table, but that references within the defining module will
627 bind to the local symbol. That is, the symbol cannot be overridden by another
628 module.
629 </dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]630</dl>
631
632</div>
633
634<!-- ======================================================================= -->
635<div class="doc_subsection">
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]636 <a name="globalvars">Global Variables</a>
637</div>
638
639<div class="doc_text">
640
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]641<p>Global variables define regions of memory allocated at compilation time
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]642instead of run-time. Global variables may optionally be initialized, may have
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]643an explicit section to be placed in, and may have an optional explicit alignment
644specified. A variable may be defined as "thread_local", which means that it
645will not be shared by threads (each thread will have a separated copy of the
646variable). A variable may be defined as a global "constant," which indicates
647that the contents of the variable will <b>never</b> be modified (enabling better
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]648optimization, allowing the global data to be placed in the read-only section of
649an executable, etc). Note that variables that need runtime initialization
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]650cannot be marked "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]651
652<p>
653LLVM explicitly allows <em>declarations</em> of global variables to be marked
654constant, even if the final definition of the global is not. This capability
655can be used to enable slightly better optimization of the program, but requires
656the language definition to guarantee that optimizations based on the
657'constantness' are valid for the translation units that do not include the
658definition.
659</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]660
661<p>As SSA values, global variables define pointer values that are in
662scope (i.e. they dominate) all basic blocks in the program. Global
663variables always define a pointer to their "content" type because they
664describe a region of memory, and all memory objects in LLVM are
665accessed through pointers.</p>
666
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]667<p>LLVM allows an explicit section to be specified for globals. If the target
668supports it, it will emit globals to the section specified.</p>
669
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]670<p>An explicit alignment may be specified for a global. If not present, or if
671the alignment is set to zero, the alignment of the global is set by the target
672to whatever it feels convenient. If an explicit alignment is specified, the
673global is forced to have at least that much alignment. All alignments must be
674a power of 2.</p>
675
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]676<p>For example, the following defines a global with an initializer, section,
677 and alignment:</p>
678
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]679<div class="doc_code">
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]680<pre>
Chris Lattner3e63a9d2007-07-13 20:01:46 +0000[diff] [blame]681@G = constant float 1.0, section "foo", align 4
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]682</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]683</div>
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]684
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]685</div>
686
687
688<!-- ======================================================================= -->
689<div class="doc_subsection">
690 <a name="functionstructure">Functions</a>
691</div>
692
693<div class="doc_text">
694
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]695<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
696an optional <a href="#linkage">linkage type</a>, an optional
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]697<a href="#visibility">visibility style</a>, an optional
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]698<a href="#callingconv">calling convention</a>, a return type, an optional
699<a href="#paramattrs">parameter attribute</a> for the return type, a function
700name, a (possibly empty) argument list (each with optional
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]701<a href="#paramattrs">parameter attributes</a>), an optional section, an
702optional alignment, an opening curly brace, a list of basic blocks, and a
703closing curly brace.
704
705LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
706optional <a href="#linkage">linkage type</a>, an optional
707<a href="#visibility">visibility style</a>, an optional
708<a href="#callingconv">calling convention</a>, a return type, an optional
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]709<a href="#paramattrs">parameter attribute</a> for the return type, a function
710name, a possibly empty list of arguments, and an optional alignment.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]711
712<p>A function definition contains a list of basic blocks, forming the CFG for
713the function. Each basic block may optionally start with a label (giving the
714basic block a symbol table entry), contains a list of instructions, and ends
715with a <a href="#terminators">terminator</a> instruction (such as a branch or
716function return).</p>
717
Chris Lattner4a3c9012007-06-08 16:52:14 +0000[diff] [blame]718<p>The first basic block in a function is special in two ways: it is immediately
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]719executed on entrance to the function, and it is not allowed to have predecessor
720basic blocks (i.e. there can not be any branches to the entry block of a
721function). Because the block can have no predecessors, it also cannot have any
722<a href="#i_phi">PHI nodes</a>.</p>
723
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]724<p>LLVM allows an explicit section to be specified for functions. If the target
725supports it, it will emit functions to the section specified.</p>
726
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]727<p>An explicit alignment may be specified for a function. If not present, or if
728the alignment is set to zero, the alignment of the function is set by the target
729to whatever it feels convenient. If an explicit alignment is specified, the
730function is forced to have at least that much alignment. All alignments must be
731a power of 2.</p>
732
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]733</div>
734
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]735
736<!-- ======================================================================= -->
737<div class="doc_subsection">
738 <a name="aliasstructure">Aliases</a>
739</div>
740<div class="doc_text">
741 <p>Aliases act as "second name" for the aliasee value (which can be either
Anton Korobeynikova80e1182007-04-28 13:45:00 +0000[diff] [blame]742 function or global variable or bitcast of global value). Aliases may have an
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]743 optional <a href="#linkage">linkage type</a>, and an
744 optional <a href="#visibility">visibility style</a>.</p>
745
746 <h5>Syntax:</h5>
747
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]748<div class="doc_code">
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]749<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]750@&lt;Name&gt; = [Linkage] [Visibility] alias &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]751</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]752</div>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]753
754</div>
755
756
757
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]758<!-- ======================================================================= -->
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]759<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
760<div class="doc_text">
761 <p>The return type and each parameter of a function type may have a set of
762 <i>parameter attributes</i> associated with them. Parameter attributes are
763 used to communicate additional information about the result or parameters of
764 a function. Parameter attributes are considered to be part of the function
765 type so two functions types that differ only by the parameter attributes
766 are different function types.</p>
767
Reid Spencer950e9f82007-01-15 18:27:39 +0000[diff] [blame]768 <p>Parameter attributes are simple keywords that follow the type specified. If
769 multiple parameter attributes are needed, they are space separated. For
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]770 example:</p>
771
772<div class="doc_code">
773<pre>
Reid Spencer9445e9a2007-07-19 23:13:04 +0000[diff] [blame]774%someFunc = i16 (i8 signext %someParam) zeroext
775%someFunc = i16 (i8 zeroext %someParam) zeroext
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]776</pre>
777</div>
778
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]779 <p>Note that the two function types above are unique because the parameter has
Reid Spencer9445e9a2007-07-19 23:13:04 +0000[diff] [blame]780 a different attribute (<tt>signext</tt> in the first one, <tt>zeroext</tt> in
781 the second). Also note that the attribute for the function result
782 (<tt>zeroext</tt>) comes immediately after the argument list.</p>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]783
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]784 <p>Currently, only the following parameter attributes are defined:</p>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]785 <dl>
Reid Spencer9445e9a2007-07-19 23:13:04 +0000[diff] [blame]786 <dt><tt>zeroext</tt></dt>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]787 <dd>This indicates that the parameter should be zero extended just before
788 a call to this function.</dd>
Reid Spencer9445e9a2007-07-19 23:13:04 +0000[diff] [blame]789 <dt><tt>signext</tt></dt>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]790 <dd>This indicates that the parameter should be sign extended just before
791 a call to this function.</dd>
Anton Korobeynikov34d6dea2007-01-28 14:30:45 +0000[diff] [blame]792 <dt><tt>inreg</tt></dt>
793 <dd>This indicates that the parameter should be placed in register (if
Anton Korobeynikov66a8c8c2007-01-28 15:27:21 +0000[diff] [blame]794 possible) during assembling function call. Support for this attribute is
795 target-specific</dd>
Anton Korobeynikov34d6dea2007-01-28 14:30:45 +0000[diff] [blame]796 <dt><tt>sret</tt></dt>
Anton Korobeynikov66a8c8c2007-01-28 15:27:21 +0000[diff] [blame]797 <dd>This indicates that the parameter specifies the address of a structure
Reid Spencer67606122007-03-22 02:02:11 +0000[diff] [blame]798 that is the return value of the function in the source program.</dd>
Zhou Shengfebca342007-06-05 05:28:26 +0000[diff] [blame]799 <dt><tt>noalias</tt></dt>
800 <dd>This indicates that the parameter not alias any other object or any
801 other "noalias" objects during the function call.
Reid Spencer2dc52012007-03-22 02:18:56 +0000[diff] [blame]802 <dt><tt>noreturn</tt></dt>
803 <dd>This function attribute indicates that the function never returns. This
804 indicates to LLVM that every call to this function should be treated as if
805 an <tt>unreachable</tt> instruction immediately followed the call.</dd>
Reid Spencer67606122007-03-22 02:02:11 +0000[diff] [blame]806 <dt><tt>nounwind</tt></dt>
807 <dd>This function attribute indicates that the function type does not use
808 the unwind instruction and does not allow stack unwinding to propagate
809 through it.</dd>
Duncan Sands50f19f52007-07-27 19:57:41 +0000[diff] [blame]810 <dt><tt>nest</tt></dt>
811 <dd>This indicates that the parameter can be excised using the
812 <a href="#int_trampoline">trampoline intrinsics</a>.</dd>
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]813 </dl>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]814
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]815</div>
816
817<!-- ======================================================================= -->
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]818<div class="doc_subsection">
Chris Lattner1eeeb0c2006-04-08 04:40:53 +0000[diff] [blame]819 <a name="moduleasm">Module-Level Inline Assembly</a>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]820</div>
821
822<div class="doc_text">
823<p>
824Modules may contain "module-level inline asm" blocks, which corresponds to the
825GCC "file scope inline asm" blocks. These blocks are internally concatenated by
826LLVM and treated as a single unit, but may be separated in the .ll file if
827desired. The syntax is very simple:
828</p>
829
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]830<div class="doc_code">
831<pre>
832module asm "inline asm code goes here"
833module asm "more can go here"
834</pre>
835</div>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]836
837<p>The strings can contain any character by escaping non-printable characters.
838 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
839 for the number.
840</p>
841
842<p>
843 The inline asm code is simply printed to the machine code .s file when
844 assembly code is generated.
845</p>
846</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]847
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]848<!-- ======================================================================= -->
849<div class="doc_subsection">
850 <a name="datalayout">Data Layout</a>
851</div>
852
853<div class="doc_text">
854<p>A module may specify a target specific data layout string that specifies how
Reid Spencerc8910842007-04-11 23:49:50 +0000[diff] [blame]855data is to be laid out in memory. The syntax for the data layout is simply:</p>
856<pre> target datalayout = "<i>layout specification</i>"</pre>
857<p>The <i>layout specification</i> consists of a list of specifications
858separated by the minus sign character ('-'). Each specification starts with a
859letter and may include other information after the letter to define some
860aspect of the data layout. The specifications accepted are as follows: </p>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]861<dl>
862 <dt><tt>E</tt></dt>
863 <dd>Specifies that the target lays out data in big-endian form. That is, the
864 bits with the most significance have the lowest address location.</dd>
865 <dt><tt>e</tt></dt>
866 <dd>Specifies that hte target lays out data in little-endian form. That is,
867 the bits with the least significance have the lowest address location.</dd>
868 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
869 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
870 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
871 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
872 too.</dd>
873 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
874 <dd>This specifies the alignment for an integer type of a given bit
875 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
876 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
877 <dd>This specifies the alignment for a vector type of a given bit
878 <i>size</i>.</dd>
879 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
880 <dd>This specifies the alignment for a floating point type of a given bit
881 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
882 (double).</dd>
883 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
884 <dd>This specifies the alignment for an aggregate type of a given bit
885 <i>size</i>.</dd>
886</dl>
887<p>When constructing the data layout for a given target, LLVM starts with a
888default set of specifications which are then (possibly) overriden by the
889specifications in the <tt>datalayout</tt> keyword. The default specifications
890are given in this list:</p>
891<ul>
892 <li><tt>E</tt> - big endian</li>
893 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
894 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
895 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
896 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
897 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
898 <li><tt>i64:32:64</tt> - i64 has abi alignment of 32-bits but preferred
899 alignment of 64-bits</li>
900 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
901 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
902 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
903 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
904 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
905</ul>
906<p>When llvm is determining the alignment for a given type, it uses the
907following rules:
908<ol>
909 <li>If the type sought is an exact match for one of the specifications, that
910 specification is used.</li>
911 <li>If no match is found, and the type sought is an integer type, then the
912 smallest integer type that is larger than the bitwidth of the sought type is
913 used. If none of the specifications are larger than the bitwidth then the the
914 largest integer type is used. For example, given the default specifications
915 above, the i7 type will use the alignment of i8 (next largest) while both
916 i65 and i256 will use the alignment of i64 (largest specified).</li>
917 <li>If no match is found, and the type sought is a vector type, then the
918 largest vector type that is smaller than the sought vector type will be used
919 as a fall back. This happens because <128 x double> can be implemented in
920 terms of 64 <2 x double>, for example.</li>
921</ol>
922</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]923
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]924<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]925<div class="doc_section"> <a name="typesystem">Type System</a> </div>
926<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]927
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]928<div class="doc_text">
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]929
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]930<p>The LLVM type system is one of the most important features of the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]931intermediate representation. Being typed enables a number of
932optimizations to be performed on the IR directly, without having to do
933extra analyses on the side before the transformation. A strong type
934system makes it easier to read the generated code and enables novel
935analyses and transformations that are not feasible to perform on normal
936three address code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]937
938</div>
939
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]940<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]941<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]942<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +0000[diff] [blame]943<p>The primitive types are the fundamental building blocks of the LLVM
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000[diff] [blame]944system. The current set of primitive types is as follows:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]945
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]946<table class="layout">
947 <tr class="layout">
948 <td class="left">
949 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]950 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]951 <tr><th>Type</th><th>Description</th></tr>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]952 <tr><td><tt><a name="t_void">void</a></tt></td><td>No value</td></tr>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]953 <tr><td><tt>label</tt></td><td>Branch destination</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]954 </tbody>
955 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]956 </td>
957 <td class="right">
958 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]959 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]960 <tr><th>Type</th><th>Description</th></tr>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]961 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]962 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]963 </tbody>
964 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]965 </td>
966 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]967</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]968</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]969
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]970<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]971<div class="doc_subsubsection"> <a name="t_classifications">Type
972Classifications</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]973<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]974<p>These different primitive types fall into a few useful
975classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]976
977<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]978 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]979 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]980 <tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]981 <td><a name="t_integer">integer</a></td>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]982 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]983 </tr>
984 <tr>
985 <td><a name="t_floating">floating point</a></td>
986 <td><tt>float, double</tt></td>
987 </tr>
988 <tr>
989 <td><a name="t_firstclass">first class</a></td>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]990 <td><tt>i1, ..., float, double, <br/>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]991 <a href="#t_pointer">pointer</a>,<a href="#t_vector">vector</a></tt>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]992 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]993 </tr>
994 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]995</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]996
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]997<p>The <a href="#t_firstclass">first class</a> types are perhaps the
998most important. Values of these types are the only ones which can be
999produced by instructions, passed as arguments, or used as operands to
1000instructions. This means that all structures and arrays must be
1001manipulated either by pointer or by component.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1002</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1003
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1004<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1005<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1006
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1007<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1008
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1009<p>The real power in LLVM comes from the derived types in the system.
1010This is what allows a programmer to represent arrays, functions,
1011pointers, and other useful types. Note that these derived types may be
1012recursive: For example, it is possible to have a two dimensional array.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1013
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1014</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1015
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1016<!-- _______________________________________________________________________ -->
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1017<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1018
1019<div class="doc_text">
1020
1021<h5>Overview:</h5>
1022<p>The integer type is a very simple derived type that simply specifies an
1023arbitrary bit width for the integer type desired. Any bit width from 1 bit to
10242^23-1 (about 8 million) can be specified.</p>
1025
1026<h5>Syntax:</h5>
1027
1028<pre>
1029 iN
1030</pre>
1031
1032<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1033value.</p>
1034
1035<h5>Examples:</h5>
1036<table class="layout">
1037 <tr class="layout">
1038 <td class="left">
1039 <tt>i1</tt><br/>
1040 <tt>i4</tt><br/>
1041 <tt>i8</tt><br/>
1042 <tt>i16</tt><br/>
1043 <tt>i32</tt><br/>
1044 <tt>i42</tt><br/>
1045 <tt>i64</tt><br/>
1046 <tt>i1942652</tt><br/>
1047 </td>
1048 <td class="left">
1049 A boolean integer of 1 bit<br/>
1050 A nibble sized integer of 4 bits.<br/>
1051 A byte sized integer of 8 bits.<br/>
1052 A half word sized integer of 16 bits.<br/>
1053 A word sized integer of 32 bits.<br/>
1054 An integer whose bit width is the answer. <br/>
1055 A double word sized integer of 64 bits.<br/>
1056 A really big integer of over 1 million bits.<br/>
1057 </td>
1058 </tr>
1059</table>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1060</div>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1061
1062<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1063<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1064
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1065<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1066
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1067<h5>Overview:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1068
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1069<p>The array type is a very simple derived type that arranges elements
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1070sequentially in memory. The array type requires a size (number of
1071elements) and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1072
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1073<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1074
1075<pre>
1076 [&lt;# elements&gt; x &lt;elementtype&gt;]
1077</pre>
1078
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]1079<p>The number of elements is a constant integer value; elementtype may
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1080be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1081
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1082<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1083<table class="layout">
1084 <tr class="layout">
1085 <td class="left">
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1086 <tt>[40 x i32 ]</tt><br/>
1087 <tt>[41 x i32 ]</tt><br/>
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1088 <tt>[40 x i8]</tt><br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1089 </td>
1090 <td class="left">
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1091 Array of 40 32-bit integer values.<br/>
1092 Array of 41 32-bit integer values.<br/>
1093 Array of 40 8-bit integer values.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1094 </td>
1095 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1096</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1097<p>Here are some examples of multidimensional arrays:</p>
1098<table class="layout">
1099 <tr class="layout">
1100 <td class="left">
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1101 <tt>[3 x [4 x i32]]</tt><br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1102 <tt>[12 x [10 x float]]</tt><br/>
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1103 <tt>[2 x [3 x [4 x i16]]]</tt><br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1104 </td>
1105 <td class="left">
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1106 3x4 array of 32-bit integer values.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1107 12x10 array of single precision floating point values.<br/>
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1108 2x3x4 array of 16-bit integer values.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1109 </td>
1110 </tr>
1111</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1112
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]1113<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1114length array. Normally, accesses past the end of an array are undefined in
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1115LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1116As a special case, however, zero length arrays are recognized to be variable
1117length. This allows implementation of 'pascal style arrays' with the LLVM
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1118type "{ i32, [0 x float]}", for example.</p>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1119
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1120</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1121
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1122<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1123<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1124<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1125<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1126<p>The function type can be thought of as a function signature. It
1127consists of a return type and a list of formal parameter types.
John Criswell009900b2003-11-25 21:45:46 +0000[diff] [blame]1128Function types are usually used to build virtual function tables
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1129(which are structures of pointers to functions), for indirect function
1130calls, and when defining a function.</p>
John Criswell009900b2003-11-25 21:45:46 +0000[diff] [blame]1131<p>
1132The return type of a function type cannot be an aggregate type.
1133</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1134<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1135<pre> &lt;returntype&gt; (&lt;parameter list&gt;)<br></pre>
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]1136<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Misha Brukmanc24b7582004-08-12 20:16:08 +0000[diff] [blame]1137specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1138which indicates that the function takes a variable number of arguments.
1139Variable argument functions can access their arguments with the <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1140 href="#int_varargs">variable argument handling intrinsic</a> functions.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1141<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1142<table class="layout">
1143 <tr class="layout">
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1144 <td class="left"><tt>i32 (i32)</tt></td>
1145 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1146 </td>
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1147 </tr><tr class="layout">
Reid Spencer9445e9a2007-07-19 23:13:04 +0000[diff] [blame]1148 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencerf17a0b72006-12-31 07:20:23 +0000[diff] [blame]1149 </tt></td>
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1150 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1151 an <tt>i16</tt> that should be sign extended and a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1152 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1153 <tt>float</tt>.
1154 </td>
1155 </tr><tr class="layout">
1156 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1157 <td class="left">A vararg function that takes at least one
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1158 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1159 which returns an integer. This is the signature for <tt>printf</tt> in
1160 LLVM.
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1161 </td>
1162 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1163</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1164
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1165</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1166<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1167<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1168<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1169<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1170<p>The structure type is used to represent a collection of data members
1171together in memory. The packing of the field types is defined to match
1172the ABI of the underlying processor. The elements of a structure may
1173be any type that has a size.</p>
1174<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1175and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1176field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1177instruction.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1178<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1179<pre> { &lt;type list&gt; }<br></pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1180<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1181<table class="layout">
1182 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]1183 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1184 <td class="left">A triple of three <tt>i32</tt> values</td>
1185 </tr><tr class="layout">
1186 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1187 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1188 second element is a <a href="#t_pointer">pointer</a> to a
1189 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1190 an <tt>i32</tt>.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1191 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1192</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1193</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1194
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1195<!-- _______________________________________________________________________ -->
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]1196<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1197</div>
1198<div class="doc_text">
1199<h5>Overview:</h5>
1200<p>The packed structure type is used to represent a collection of data members
1201together in memory. There is no padding between fields. Further, the alignment
1202of a packed structure is 1 byte. The elements of a packed structure may
1203be any type that has a size.</p>
1204<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1205and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1206field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1207instruction.</p>
1208<h5>Syntax:</h5>
1209<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1210<h5>Examples:</h5>
1211<table class="layout">
1212 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]1213 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1214 <td class="left">A triple of three <tt>i32</tt> values</td>
1215 </tr><tr class="layout">
1216 <td class="left"><tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}&nbsp;&gt;</tt></td>
1217 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1218 second element is a <a href="#t_pointer">pointer</a> to a
1219 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1220 an <tt>i32</tt>.</td>
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]1221 </tr>
1222</table>
1223</div>
1224
1225<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1226<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1227<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1228<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1229<p>As in many languages, the pointer type represents a pointer or
1230reference to another object, which must live in memory.</p>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1231<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1232<pre> &lt;type&gt; *<br></pre>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1233<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1234<table class="layout">
1235 <tr class="layout">
1236 <td class="left">
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1237 <tt>[4x i32]*</tt><br/>
1238 <tt>i32 (i32 *) *</tt><br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1239 </td>
1240 <td class="left">
1241 A <a href="#t_pointer">pointer</a> to <a href="#t_array">array</a> of
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1242 four <tt>i32</tt> values<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1243 A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1244 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
1245 <tt>i32</tt>.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1246 </td>
1247 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1248</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1249</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1250
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1251<!-- _______________________________________________________________________ -->
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1252<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1253<div class="doc_text">
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1254
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1255<h5>Overview:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1256
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1257<p>A vector type is a simple derived type that represents a vector
1258of elements. Vector types are used when multiple primitive data
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1259are operated in parallel using a single instruction (SIMD).
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1260A vector type requires a size (number of
Chris Lattnerb8d172f2005-11-10 01:44:22 +0000[diff] [blame]1261elements) and an underlying primitive data type. Vectors must have a power
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1262of two length (1, 2, 4, 8, 16 ...). Vector types are
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1263considered <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1264
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1265<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1266
1267<pre>
1268 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1269</pre>
1270
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1271<p>The number of elements is a constant integer value; elementtype may
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]1272be any integer or floating point type.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1273
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1274<h5>Examples:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1275
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1276<table class="layout">
1277 <tr class="layout">
1278 <td class="left">
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1279 <tt>&lt;4 x i32&gt;</tt><br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1280 <tt>&lt;8 x float&gt;</tt><br/>
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1281 <tt>&lt;2 x i64&gt;</tt><br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1282 </td>
1283 <td class="left">
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1284 Vector of 4 32-bit integer values.<br/>
1285 Vector of 8 floating-point values.<br/>
1286 Vector of 2 64-bit integer values.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1287 </td>
1288 </tr>
1289</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1290</div>
1291
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1292<!-- _______________________________________________________________________ -->
1293<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1294<div class="doc_text">
1295
1296<h5>Overview:</h5>
1297
1298<p>Opaque types are used to represent unknown types in the system. This
1299corresponds (for example) to the C notion of a foward declared structure type.
1300In LLVM, opaque types can eventually be resolved to any type (not just a
1301structure type).</p>
1302
1303<h5>Syntax:</h5>
1304
1305<pre>
1306 opaque
1307</pre>
1308
1309<h5>Examples:</h5>
1310
1311<table class="layout">
1312 <tr class="layout">
1313 <td class="left">
1314 <tt>opaque</tt>
1315 </td>
1316 <td class="left">
1317 An opaque type.<br/>
1318 </td>
1319 </tr>
1320</table>
1321</div>
1322
1323
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1324<!-- *********************************************************************** -->
1325<div class="doc_section"> <a name="constants">Constants</a> </div>
1326<!-- *********************************************************************** -->
1327
1328<div class="doc_text">
1329
1330<p>LLVM has several different basic types of constants. This section describes
1331them all and their syntax.</p>
1332
1333</div>
1334
1335<!-- ======================================================================= -->
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]1336<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1337
1338<div class="doc_text">
1339
1340<dl>
1341 <dt><b>Boolean constants</b></dt>
1342
1343 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]1344 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1345 </dd>
1346
1347 <dt><b>Integer constants</b></dt>
1348
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]1349 <dd>Standard integers (such as '4') are constants of the <a
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1350 href="#t_integer">integer</a> type. Negative numbers may be used with
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1351 integer types.
1352 </dd>
1353
1354 <dt><b>Floating point constants</b></dt>
1355
1356 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1357 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1358 notation (see below). Floating point constants must have a <a
1359 href="#t_floating">floating point</a> type. </dd>
1360
1361 <dt><b>Null pointer constants</b></dt>
1362
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1363 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1364 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1365
1366</dl>
1367
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1368<p>The one non-intuitive notation for constants is the optional hexadecimal form
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1369of floating point constants. For example, the form '<tt>double
13700x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
13714.5e+15</tt>'. The only time hexadecimal floating point constants are required
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]1372(and the only time that they are generated by the disassembler) is when a
1373floating point constant must be emitted but it cannot be represented as a
1374decimal floating point number. For example, NaN's, infinities, and other
1375special values are represented in their IEEE hexadecimal format so that
1376assembly and disassembly do not cause any bits to change in the constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1377
1378</div>
1379
1380<!-- ======================================================================= -->
1381<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1382</div>
1383
1384<div class="doc_text">
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000[diff] [blame]1385<p>Aggregate constants arise from aggregation of simple constants
1386and smaller aggregate constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1387
1388<dl>
1389 <dt><b>Structure constants</b></dt>
1390
1391 <dd>Structure constants are represented with notation similar to structure
1392 type definitions (a comma separated list of elements, surrounded by braces
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1393 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* %G }</tt>",
Chris Lattner3e63a9d2007-07-13 20:01:46 +0000[diff] [blame]1394 where "<tt>%G</tt>" is declared as "<tt>@G = external global i32</tt>". Structure constants
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000[diff] [blame]1395 must have <a href="#t_struct">structure type</a>, and the number and
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1396 types of elements must match those specified by the type.
1397 </dd>
1398
1399 <dt><b>Array constants</b></dt>
1400
1401 <dd>Array constants are represented with notation similar to array type
1402 definitions (a comma separated list of elements, surrounded by square brackets
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1403 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1404 constants must have <a href="#t_array">array type</a>, and the number and
1405 types of elements must match those specified by the type.
1406 </dd>
1407
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1408 <dt><b>Vector constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1409
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1410 <dd>Vector constants are represented with notation similar to vector type
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1411 definitions (a comma separated list of elements, surrounded by
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1412 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; i32 42,
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]1413 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1414 href="#t_vector">vector type</a>, and the number and types of elements must
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1415 match those specified by the type.
1416 </dd>
1417
1418 <dt><b>Zero initialization</b></dt>
1419
1420 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1421 value to zero of <em>any</em> type, including scalar and aggregate types.
1422 This is often used to avoid having to print large zero initializers (e.g. for
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]1423 large arrays) and is always exactly equivalent to using explicit zero
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1424 initializers.
1425 </dd>
1426</dl>
1427
1428</div>
1429
1430<!-- ======================================================================= -->
1431<div class="doc_subsection">
1432 <a name="globalconstants">Global Variable and Function Addresses</a>
1433</div>
1434
1435<div class="doc_text">
1436
1437<p>The addresses of <a href="#globalvars">global variables</a> and <a
1438href="#functionstructure">functions</a> are always implicitly valid (link-time)
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1439constants. These constants are explicitly referenced when the <a
1440href="#identifiers">identifier for the global</a> is used and always have <a
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1441href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1442file:</p>
1443
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1444<div class="doc_code">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1445<pre>
Chris Lattnera18a4242007-06-06 18:28:13 +0000[diff] [blame]1446@X = global i32 17
1447@Y = global i32 42
1448@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1449</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1450</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1451
1452</div>
1453
1454<!-- ======================================================================= -->
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1455<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1456<div class="doc_text">
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1457 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1458 no specific value. Undefined values may be of any type and be used anywhere
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1459 a constant is permitted.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1460
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1461 <p>Undefined values indicate to the compiler that the program is well defined
1462 no matter what value is used, giving the compiler more freedom to optimize.
1463 </p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1464</div>
1465
1466<!-- ======================================================================= -->
1467<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1468</div>
1469
1470<div class="doc_text">
1471
1472<p>Constant expressions are used to allow expressions involving other constants
1473to be used as constants. Constant expressions may be of any <a
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1474href="#t_firstclass">first class</a> type and may involve any LLVM operation
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1475that does not have side effects (e.g. load and call are not supported). The
1476following is the syntax for constant expressions:</p>
1477
1478<dl>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1479 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1480 <dd>Truncate a constant to another type. The bit size of CST must be larger
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]1481 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1482
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1483 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1484 <dd>Zero extend a constant to another type. The bit size of CST must be
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]1485 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1486
1487 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1488 <dd>Sign extend a constant to another type. The bit size of CST must be
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]1489 smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1490
1491 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1492 <dd>Truncate a floating point constant to another floating point type. The
1493 size of CST must be larger than the size of TYPE. Both types must be
1494 floating point.</dd>
1495
1496 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1497 <dd>Floating point extend a constant to another type. The size of CST must be
1498 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1499
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]1500 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1501 <dd>Convert a floating point constant to the corresponding unsigned integer
1502 constant. TYPE must be an integer type. CST must be floating point. If the
1503 value won't fit in the integer type, the results are undefined.</dd>
1504
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]1505 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1506 <dd>Convert a floating point constant to the corresponding signed integer
1507 constant. TYPE must be an integer type. CST must be floating point. If the
1508 value won't fit in the integer type, the results are undefined.</dd>
1509
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]1510 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1511 <dd>Convert an unsigned integer constant to the corresponding floating point
1512 constant. TYPE must be floating point. CST must be of integer type. If the
Jeff Cohencb757312007-04-22 14:56:37 +0000[diff] [blame]1513 value won't fit in the floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1514
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]1515 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1516 <dd>Convert a signed integer constant to the corresponding floating point
1517 constant. TYPE must be floating point. CST must be of integer type. If the
Jeff Cohencb757312007-04-22 14:56:37 +0000[diff] [blame]1518 value won't fit in the floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1519
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]1520 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1521 <dd>Convert a pointer typed constant to the corresponding integer constant
1522 TYPE must be an integer type. CST must be of pointer type. The CST value is
1523 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1524
1525 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1526 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1527 pointer type. CST must be of integer type. The CST value is zero extended,
1528 truncated, or unchanged to make it fit in a pointer size. This one is
1529 <i>really</i> dangerous!</dd>
1530
1531 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1532 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1533 identical (same number of bits). The conversion is done as if the CST value
1534 was stored to memory and read back as TYPE. In other words, no bits change
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]1535 with this operator, just the type. This can be used for conversion of
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1536 vector types to any other type, as long as they have the same bit width. For
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]1537 pointers it is only valid to cast to another pointer type.
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1538 </dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1539
1540 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1541
1542 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1543 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1544 instruction, the index list may have zero or more indexes, which are required
1545 to make sense for the type of "CSTPTR".</dd>
1546
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]1547 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1548
1549 <dd>Perform the <a href="#i_select">select operation</a> on
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]1550 constants.</dd>
1551
1552 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1553 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1554
1555 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1556 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]1557
1558 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1559
1560 <dd>Perform the <a href="#i_extractelement">extractelement
1561 operation</a> on constants.
1562
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]1563 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1564
1565 <dd>Perform the <a href="#i_insertelement">insertelement
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]1566 operation</a> on constants.</dd>
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]1567
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]1568
1569 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1570
1571 <dd>Perform the <a href="#i_shufflevector">shufflevector
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]1572 operation</a> on constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]1573
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1574 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1575
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1576 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1577 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1578 binary</a> operations. The constraints on operands are the same as those for
1579 the corresponding instruction (e.g. no bitwise operations on floating point
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]1580 values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1581</dl>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1582</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +0000[diff] [blame]1583
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1584<!-- *********************************************************************** -->
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1585<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1586<!-- *********************************************************************** -->
1587
1588<!-- ======================================================================= -->
1589<div class="doc_subsection">
1590<a name="inlineasm">Inline Assembler Expressions</a>
1591</div>
1592
1593<div class="doc_text">
1594
1595<p>
1596LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1597Module-Level Inline Assembly</a>) through the use of a special value. This
1598value represents the inline assembler as a string (containing the instructions
1599to emit), a list of operand constraints (stored as a string), and a flag that
1600indicates whether or not the inline asm expression has side effects. An example
1601inline assembler expression is:
1602</p>
1603
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1604<div class="doc_code">
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1605<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1606i32 (i32) asm "bswap $0", "=r,r"
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1607</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1608</div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1609
1610<p>
1611Inline assembler expressions may <b>only</b> be used as the callee operand of
1612a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1613</p>
1614
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1615<div class="doc_code">
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1616<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1617%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1618</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1619</div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1620
1621<p>
1622Inline asms with side effects not visible in the constraint list must be marked
1623as having side effects. This is done through the use of the
1624'<tt>sideeffect</tt>' keyword, like so:
1625</p>
1626
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1627<div class="doc_code">
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1628<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1629call void asm sideeffect "eieio", ""()
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1630</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1631</div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1632
1633<p>TODO: The format of the asm and constraints string still need to be
1634documented here. Constraints on what can be done (e.g. duplication, moving, etc
1635need to be documented).
1636</p>
1637
1638</div>
1639
1640<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1641<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1642<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1643
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1644<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1645
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1646<p>The LLVM instruction set consists of several different
1647classifications of instructions: <a href="#terminators">terminator
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1648instructions</a>, <a href="#binaryops">binary instructions</a>,
1649<a href="#bitwiseops">bitwise binary instructions</a>, <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1650 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1651instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1652
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1653</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1654
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1655<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1656<div class="doc_subsection"> <a name="terminators">Terminator
1657Instructions</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1658
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1659<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1660
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1661<p>As mentioned <a href="#functionstructure">previously</a>, every
1662basic block in a program ends with a "Terminator" instruction, which
1663indicates which block should be executed after the current block is
1664finished. These terminator instructions typically yield a '<tt>void</tt>'
1665value: they produce control flow, not values (the one exception being
1666the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1667<p>There are six different terminator instructions: the '<a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1668 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1669instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1670the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1671 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1672 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1673
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1674</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1675
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1676<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1677<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1678Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1679<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1680<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1681<pre> ret &lt;type&gt; &lt;value&gt; <i>; Return a value from a non-void function</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1682 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1683</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1684<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1685<p>The '<tt>ret</tt>' instruction is used to return control flow (and a
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1686value) from a function back to the caller.</p>
John Criswell4457dc92004-04-09 16:48:45 +0000[diff] [blame]1687<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1688returns a value and then causes control flow, and one that just causes
1689control flow to occur.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1690<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1691<p>The '<tt>ret</tt>' instruction may return any '<a
1692 href="#t_firstclass">first class</a>' type. Notice that a function is
1693not <a href="#wellformed">well formed</a> if there exists a '<tt>ret</tt>'
1694instruction inside of the function that returns a value that does not
1695match the return type of the function.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1696<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1697<p>When the '<tt>ret</tt>' instruction is executed, control flow
1698returns back to the calling function's context. If the caller is a "<a
John Criswellfa081872004-06-25 15:16:57 +0000[diff] [blame]1699 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1700the instruction after the call. If the caller was an "<a
1701 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]1702at the beginning of the "normal" destination block. If the instruction
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1703returns a value, that value shall set the call or invoke instruction's
1704return value.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1705<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1706<pre> ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1707 ret void <i>; Return from a void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1708</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1709</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1710<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1711<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1712<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1713<h5>Syntax:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]1714<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 Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1715</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1716<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1717<p>The '<tt>br</tt>' instruction is used to cause control flow to
1718transfer to a different basic block in the current function. There are
1719two forms of this instruction, corresponding to a conditional branch
1720and an unconditional branch.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1721<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1722<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]1723single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1724unconditional form of the '<tt>br</tt>' instruction takes a single
1725'<tt>label</tt>' value as a target.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1726<h5>Semantics:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]1727<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1728argument is evaluated. If the value is <tt>true</tt>, control flows
1729to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1730control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1731<h5>Example:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]1732<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 Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1733 href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1734</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1735<!-- _______________________________________________________________________ -->
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1736<div class="doc_subsubsection">
1737 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1738</div>
1739
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1740<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1741<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1742
1743<pre>
1744 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
1745</pre>
1746
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1747<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1748
1749<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1750several different places. It is a generalization of the '<tt>br</tt>'
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1751instruction, allowing a branch to occur to one of many possible
1752destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1753
1754
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1755<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1756
1757<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1758comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1759an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1760table is not allowed to contain duplicate constant entries.</p>
1761
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1762<h5>Semantics:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1763
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1764<p>The <tt>switch</tt> instruction specifies a table of values and
1765destinations. When the '<tt>switch</tt>' instruction is executed, this
John Criswell84114752004-06-25 16:05:06 +0000[diff] [blame]1766table is searched for the given value. If the value is found, control flow is
1767transfered to the corresponding destination; otherwise, control flow is
1768transfered to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1769
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1770<h5>Implementation:</h5>
1771
1772<p>Depending on properties of the target machine and the particular
1773<tt>switch</tt> instruction, this instruction may be code generated in different
John Criswell84114752004-06-25 16:05:06 +0000[diff] [blame]1774ways. For example, it could be generated as a series of chained conditional
1775branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1776
1777<h5>Example:</h5>
1778
1779<pre>
1780 <i>; Emulate a conditional br instruction</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]1781 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1782 switch i32 %Val, label %truedest [i32 0, label %falsedest ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1783
1784 <i>; Emulate an unconditional br instruction</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1785 switch i32 0, label %dest [ ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1786
1787 <i>; Implement a jump table:</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1788 switch i32 %val, label %otherwise [ i32 0, label %onzero
1789 i32 1, label %onone
1790 i32 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1791</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1792</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1793
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1794<!-- _______________________________________________________________________ -->
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1795<div class="doc_subsubsection">
1796 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1797</div>
1798
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1799<div class="doc_text">
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1800
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1801<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1802
1803<pre>
1804 &lt;result&gt; = invoke [<a href="#callingconv">cconv</a>] &lt;ptr to function ty&gt; %&lt;function ptr val&gt;(&lt;function args&gt;)
Chris Lattner76b8a332006-05-14 18:23:06 +0000[diff] [blame]1805 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1806</pre>
1807
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]1808<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1809
1810<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1811function, with the possibility of control flow transfer to either the
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]1812'<tt>normal</tt>' label or the
1813'<tt>exception</tt>' label. If the callee function returns with the
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1814"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1815"normal" label. If the callee (or any indirect callees) returns with the "<a
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]1816href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
1817continued at the dynamically nearest "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1818
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1819<h5>Arguments:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1820
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1821<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1822
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1823<ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1824 <li>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]1825 The optional "cconv" marker indicates which <a href="#callingconv">calling
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1826 convention</a> the call should use. If none is specified, the call defaults
1827 to using C calling conventions.
1828 </li>
1829 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1830 function value being invoked. In most cases, this is a direct function
1831 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1832 an arbitrary pointer to function value.
1833 </li>
1834
1835 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1836 function to be invoked. </li>
1837
1838 <li>'<tt>function args</tt>': argument list whose types match the function
1839 signature argument types. If the function signature indicates the function
1840 accepts a variable number of arguments, the extra arguments can be
1841 specified. </li>
1842
1843 <li>'<tt>normal label</tt>': the label reached when the called function
1844 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1845
1846 <li>'<tt>exception label</tt>': the label reached when a callee returns with
1847 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
1848
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1849</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1850
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1851<h5>Semantics:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1852
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1853<p>This instruction is designed to operate as a standard '<tt><a
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1854href="#i_call">call</a></tt>' instruction in most regards. The primary
1855difference is that it establishes an association with a label, which is used by
1856the runtime library to unwind the stack.</p>
1857
1858<p>This instruction is used in languages with destructors to ensure that proper
1859cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
1860exception. Additionally, this is important for implementation of
1861'<tt>catch</tt>' clauses in high-level languages that support them.</p>
1862
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1863<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1864<pre>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]1865 %retval = invoke i32 %Test(i32 15) to label %Continue
1866 unwind label %TestCleanup <i>; {i32}:retval set</i>
1867 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Test(i32 15) to label %Continue
1868 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1869</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1870</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1871
1872
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1873<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1874
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1875<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
1876Instruction</a> </div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1877
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1878<div class="doc_text">
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1879
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1880<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1881<pre>
1882 unwind
1883</pre>
1884
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1885<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1886
1887<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
1888at the first callee in the dynamic call stack which used an <a
1889href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
1890primarily used to implement exception handling.</p>
1891
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1892<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1893
1894<p>The '<tt>unwind</tt>' intrinsic causes execution of the current function to
1895immediately halt. The dynamic call stack is then searched for the first <a
1896href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
1897execution continues at the "exceptional" destination block specified by the
1898<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
1899dynamic call chain, undefined behavior results.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1900</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1901
1902<!-- _______________________________________________________________________ -->
1903
1904<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
1905Instruction</a> </div>
1906
1907<div class="doc_text">
1908
1909<h5>Syntax:</h5>
1910<pre>
1911 unreachable
1912</pre>
1913
1914<h5>Overview:</h5>
1915
1916<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
1917instruction is used to inform the optimizer that a particular portion of the
1918code is not reachable. This can be used to indicate that the code after a
1919no-return function cannot be reached, and other facts.</p>
1920
1921<h5>Semantics:</h5>
1922
1923<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
1924</div>
1925
1926
1927
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1928<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1929<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1930<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1931<p>Binary operators are used to do most of the computation in a
1932program. They require two operands, execute an operation on them, and
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1933produce a single value. The operands might represent
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1934multiple data, as is the case with the <a href="#t_vector">vector</a> data type.
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1935The result value of a binary operator is not
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1936necessarily the same type as its operands.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1937<p>There are several different binary operators:</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1938</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1939<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1940<div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
1941Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1942<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1943<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1944<pre> &lt;result&gt; = add &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1945</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1946<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1947<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1948<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1949<p>The two arguments to the '<tt>add</tt>' instruction must be either <a
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1950 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1951 This instruction can also take <a href="#t_vector">vector</a> versions of the values.
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1952Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1953<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1954<p>The value produced is the integer or floating point sum of the two
1955operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1956<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1957<pre> &lt;result&gt; = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1958</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1959</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1960<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1961<div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
1962Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1963<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1964<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1965<pre> &lt;result&gt; = sub &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1966</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1967<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1968<p>The '<tt>sub</tt>' instruction returns the difference of its two
1969operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1970<p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
1971instruction present in most other intermediate representations.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1972<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1973<p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1974 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1975values.
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1976This instruction can also take <a href="#t_vector">vector</a> versions of the values.
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1977Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1978<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1979<p>The value produced is the integer or floating point difference of
1980the two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1981<h5>Example:</h5>
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]1982<pre>
1983 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1984 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1985</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1986</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1987<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1988<div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
1989Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1990<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1991<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1992<pre> &lt;result&gt; = mul &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1993</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1994<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1995<p>The '<tt>mul</tt>' instruction returns the product of its two
1996operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1997<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1998<p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1999 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2000values.
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2001This instruction can also take <a href="#t_vector">vector</a> versions of the values.
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2002Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2003<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2004<p>The value produced is the integer or floating point product of the
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2005two operands.</p>
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]2006<p>Because the operands are the same width, the result of an integer
2007multiplication is the same whether the operands should be deemed unsigned or
2008signed.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2009<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2010<pre> &lt;result&gt; = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2011</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2012</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2013<!-- _______________________________________________________________________ -->
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2014<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2015</a></div>
2016<div class="doc_text">
2017<h5>Syntax:</h5>
2018<pre> &lt;result&gt; = udiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2019</pre>
2020<h5>Overview:</h5>
2021<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2022operands.</p>
2023<h5>Arguments:</h5>
2024<p>The two arguments to the '<tt>udiv</tt>' instruction must be
2025<a href="#t_integer">integer</a> values. Both arguments must have identical
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2026types. This instruction can also take <a href="#t_vector">vector</a> versions
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2027of the values in which case the elements must be integers.</p>
2028<h5>Semantics:</h5>
2029<p>The value produced is the unsigned integer quotient of the two operands. This
2030instruction always performs an unsigned division operation, regardless of
2031whether the arguments are unsigned or not.</p>
2032<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2033<pre> &lt;result&gt; = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2034</pre>
2035</div>
2036<!-- _______________________________________________________________________ -->
2037<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2038</a> </div>
2039<div class="doc_text">
2040<h5>Syntax:</h5>
2041<pre> &lt;result&gt; = sdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2042</pre>
2043<h5>Overview:</h5>
2044<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2045operands.</p>
2046<h5>Arguments:</h5>
2047<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2048<a href="#t_integer">integer</a> values. Both arguments must have identical
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2049types. This instruction can also take <a href="#t_vector">vector</a> versions
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2050of the values in which case the elements must be integers.</p>
2051<h5>Semantics:</h5>
2052<p>The value produced is the signed integer quotient of the two operands. This
2053instruction always performs a signed division operation, regardless of whether
2054the arguments are signed or not.</p>
2055<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2056<pre> &lt;result&gt; = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2057</pre>
2058</div>
2059<!-- _______________________________________________________________________ -->
2060<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2061Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2062<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2063<h5>Syntax:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2064<pre> &lt;result&gt; = fdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2065</pre>
2066<h5>Overview:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2067<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2068operands.</p>
2069<h5>Arguments:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2070<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2071<a href="#t_floating">floating point</a> values. Both arguments must have
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2072identical types. This instruction can also take <a href="#t_vector">vector</a>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2073versions of floating point values.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2074<h5>Semantics:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2075<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2076<h5>Example:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2077<pre> &lt;result&gt; = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2078</pre>
2079</div>
2080<!-- _______________________________________________________________________ -->
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]2081<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2082</div>
2083<div class="doc_text">
2084<h5>Syntax:</h5>
2085<pre> &lt;result&gt; = urem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2086</pre>
2087<h5>Overview:</h5>
2088<p>The '<tt>urem</tt>' instruction returns the remainder from the
2089unsigned division of its two arguments.</p>
2090<h5>Arguments:</h5>
2091<p>The two arguments to the '<tt>urem</tt>' instruction must be
2092<a href="#t_integer">integer</a> values. Both arguments must have identical
2093types.</p>
2094<h5>Semantics:</h5>
2095<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
2096This instruction always performs an unsigned division to get the remainder,
2097regardless of whether the arguments are unsigned or not.</p>
2098<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2099<pre> &lt;result&gt; = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]2100</pre>
2101
2102</div>
2103<!-- _______________________________________________________________________ -->
2104<div class="doc_subsubsection"> <a name="i_srem">'<tt>srem</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2105Instruction</a> </div>
2106<div class="doc_text">
2107<h5>Syntax:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]2108<pre> &lt;result&gt; = srem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2109</pre>
2110<h5>Overview:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]2111<p>The '<tt>srem</tt>' instruction returns the remainder from the
2112signed division of its two operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2113<h5>Arguments:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]2114<p>The two arguments to the '<tt>srem</tt>' instruction must be
2115<a href="#t_integer">integer</a> values. Both arguments must have identical
2116types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2117<h5>Semantics:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]2118<p>This instruction returns the <i>remainder</i> of a division (where the result
Reid Spencerc9fdfc82007-03-24 22:23:39 +0000[diff] [blame]2119has the same sign as the dividend, <tt>var1</tt>), not the <i>modulo</i>
2120operator (where the result has the same sign as the divisor, <tt>var2</tt>) of
2121a value. For more information about the difference, see <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2122 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
Reid Spencerc9fdfc82007-03-24 22:23:39 +0000[diff] [blame]2123Math Forum</a>. For a table of how this is implemented in various languages,
Reid Spencer64f5c6c2007-03-24 22:40:44 +0000[diff] [blame]2124please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
Reid Spencerc9fdfc82007-03-24 22:23:39 +0000[diff] [blame]2125Wikipedia: modulo operation</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2126<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2127<pre> &lt;result&gt; = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]2128</pre>
2129
2130</div>
2131<!-- _______________________________________________________________________ -->
2132<div class="doc_subsubsection"> <a name="i_frem">'<tt>frem</tt>'
2133Instruction</a> </div>
2134<div class="doc_text">
2135<h5>Syntax:</h5>
2136<pre> &lt;result&gt; = frem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2137</pre>
2138<h5>Overview:</h5>
2139<p>The '<tt>frem</tt>' instruction returns the remainder from the
2140division of its two operands.</p>
2141<h5>Arguments:</h5>
2142<p>The two arguments to the '<tt>frem</tt>' instruction must be
2143<a href="#t_floating">floating point</a> values. Both arguments must have
2144identical types.</p>
2145<h5>Semantics:</h5>
2146<p>This instruction returns the <i>remainder</i> of a division.</p>
2147<h5>Example:</h5>
2148<pre> &lt;result&gt; = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2149</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2150</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]2151
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]2152<!-- ======================================================================= -->
2153<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2154Operations</a> </div>
2155<div class="doc_text">
2156<p>Bitwise binary operators are used to do various forms of
2157bit-twiddling in a program. They are generally very efficient
2158instructions and can commonly be strength reduced from other
2159instructions. They require two operands, execute an operation on them,
2160and produce a single value. The resulting value of the bitwise binary
2161operators is always the same type as its first operand.</p>
2162</div>
2163
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2164<!-- _______________________________________________________________________ -->
2165<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2166Instruction</a> </div>
2167<div class="doc_text">
2168<h5>Syntax:</h5>
2169<pre> &lt;result&gt; = shl &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2170</pre>
2171<h5>Overview:</h5>
2172<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2173the left a specified number of bits.</p>
2174<h5>Arguments:</h5>
2175<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
2176 href="#t_integer">integer</a> type.</p>
2177<h5>Semantics:</h5>
2178<p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup>.</p>
2179<h5>Example:</h5><pre>
2180 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2181 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2182 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
2183</pre>
2184</div>
2185<!-- _______________________________________________________________________ -->
2186<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2187Instruction</a> </div>
2188<div class="doc_text">
2189<h5>Syntax:</h5>
2190<pre> &lt;result&gt; = lshr &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2191</pre>
2192
2193<h5>Overview:</h5>
2194<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2195operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2196
2197<h5>Arguments:</h5>
2198<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
2199<a href="#t_integer">integer</a> type.</p>
2200
2201<h5>Semantics:</h5>
2202<p>This instruction always performs a logical shift right operation. The most
2203significant bits of the result will be filled with zero bits after the
2204shift.</p>
2205
2206<h5>Example:</h5>
2207<pre>
2208 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2209 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2210 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2211 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
2212</pre>
2213</div>
2214
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]2215<!-- _______________________________________________________________________ -->
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2216<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2217Instruction</a> </div>
2218<div class="doc_text">
2219
2220<h5>Syntax:</h5>
2221<pre> &lt;result&gt; = ashr &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2222</pre>
2223
2224<h5>Overview:</h5>
2225<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2226operand shifted to the right a specified number of bits with sign extension.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2227
2228<h5>Arguments:</h5>
2229<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
2230<a href="#t_integer">integer</a> type.</p>
2231
2232<h5>Semantics:</h5>
2233<p>This instruction always performs an arithmetic shift right operation,
2234The most significant bits of the result will be filled with the sign bit
2235of <tt>var1</tt>.</p>
2236
2237<h5>Example:</h5>
2238<pre>
2239 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2240 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2241 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2242 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
2243</pre>
2244</div>
2245
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2246<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2247<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2248Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2249<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2250<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2251<pre> &lt;result&gt; = and &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2252</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2253<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2254<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2255its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2256<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2257<p>The two arguments to the '<tt>and</tt>' instruction must be <a
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]2258 href="#t_integer">integer</a> values. Both arguments must have
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2259identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2260<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2261<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2262<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2263<div style="align: center">
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2264<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2265 <tbody>
2266 <tr>
2267 <td>In0</td>
2268 <td>In1</td>
2269 <td>Out</td>
2270 </tr>
2271 <tr>
2272 <td>0</td>
2273 <td>0</td>
2274 <td>0</td>
2275 </tr>
2276 <tr>
2277 <td>0</td>
2278 <td>1</td>
2279 <td>0</td>
2280 </tr>
2281 <tr>
2282 <td>1</td>
2283 <td>0</td>
2284 <td>0</td>
2285 </tr>
2286 <tr>
2287 <td>1</td>
2288 <td>1</td>
2289 <td>1</td>
2290 </tr>
2291 </tbody>
2292</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2293</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2294<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2295<pre> &lt;result&gt; = and i32 4, %var <i>; yields {i32}:result = 4 &amp; %var</i>
2296 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2297 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2298</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2299</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2300<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2301<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2302<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2303<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2304<pre> &lt;result&gt; = or &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2305</pre>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2306<h5>Overview:</h5>
2307<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2308or of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2309<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2310<p>The two arguments to the '<tt>or</tt>' instruction must be <a
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]2311 href="#t_integer">integer</a> values. Both arguments must have
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2312identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2313<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2314<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2315<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2316<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2317<table border="1" cellspacing="0" cellpadding="4">
2318 <tbody>
2319 <tr>
2320 <td>In0</td>
2321 <td>In1</td>
2322 <td>Out</td>
2323 </tr>
2324 <tr>
2325 <td>0</td>
2326 <td>0</td>
2327 <td>0</td>
2328 </tr>
2329 <tr>
2330 <td>0</td>
2331 <td>1</td>
2332 <td>1</td>
2333 </tr>
2334 <tr>
2335 <td>1</td>
2336 <td>0</td>
2337 <td>1</td>
2338 </tr>
2339 <tr>
2340 <td>1</td>
2341 <td>1</td>
2342 <td>1</td>
2343 </tr>
2344 </tbody>
2345</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2346</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2347<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2348<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
2349 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
2350 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2351</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2352</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2353<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2354<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2355Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2356<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2357<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2358<pre> &lt;result&gt; = xor &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2359</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2360<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2361<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2362or of its two operands. The <tt>xor</tt> is used to implement the
2363"one's complement" operation, which is the "~" operator in C.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2364<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2365<p>The two arguments to the '<tt>xor</tt>' instruction must be <a
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]2366 href="#t_integer">integer</a> values. Both arguments must have
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2367identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2368<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2369<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2370<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2371<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2372<table border="1" cellspacing="0" cellpadding="4">
2373 <tbody>
2374 <tr>
2375 <td>In0</td>
2376 <td>In1</td>
2377 <td>Out</td>
2378 </tr>
2379 <tr>
2380 <td>0</td>
2381 <td>0</td>
2382 <td>0</td>
2383 </tr>
2384 <tr>
2385 <td>0</td>
2386 <td>1</td>
2387 <td>1</td>
2388 </tr>
2389 <tr>
2390 <td>1</td>
2391 <td>0</td>
2392 <td>1</td>
2393 </tr>
2394 <tr>
2395 <td>1</td>
2396 <td>1</td>
2397 <td>0</td>
2398 </tr>
2399 </tbody>
2400</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2401</div>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2402<p> </p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2403<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2404<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
2405 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
2406 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
2407 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2408</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2409</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2410
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2411<!-- ======================================================================= -->
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2412<div class="doc_subsection">
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2413 <a name="vectorops">Vector Operations</a>
2414</div>
2415
2416<div class="doc_text">
2417
2418<p>LLVM supports several instructions to represent vector operations in a
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2419target-independent manner. These instructions cover the element-access and
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2420vector-specific operations needed to process vectors effectively. While LLVM
2421does directly support these vector operations, many sophisticated algorithms
2422will want to use target-specific intrinsics to take full advantage of a specific
2423target.</p>
2424
2425</div>
2426
2427<!-- _______________________________________________________________________ -->
2428<div class="doc_subsubsection">
2429 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2430</div>
2431
2432<div class="doc_text">
2433
2434<h5>Syntax:</h5>
2435
2436<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2437 &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 Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2438</pre>
2439
2440<h5>Overview:</h5>
2441
2442<p>
2443The '<tt>extractelement</tt>' instruction extracts a single scalar
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2444element from a vector at a specified index.
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2445</p>
2446
2447
2448<h5>Arguments:</h5>
2449
2450<p>
2451The first operand of an '<tt>extractelement</tt>' instruction is a
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2452value of <a href="#t_vector">vector</a> type. The second operand is
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2453an index indicating the position from which to extract the element.
2454The index may be a variable.</p>
2455
2456<h5>Semantics:</h5>
2457
2458<p>
2459The result is a scalar of the same type as the element type of
2460<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2461<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2462results are undefined.
2463</p>
2464
2465<h5>Example:</h5>
2466
2467<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2468 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2469</pre>
2470</div>
2471
2472
2473<!-- _______________________________________________________________________ -->
2474<div class="doc_subsubsection">
2475 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2476</div>
2477
2478<div class="doc_text">
2479
2480<h5>Syntax:</h5>
2481
2482<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2483 &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 Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2484</pre>
2485
2486<h5>Overview:</h5>
2487
2488<p>
2489The '<tt>insertelement</tt>' instruction inserts a scalar
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2490element into a vector at a specified index.
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2491</p>
2492
2493
2494<h5>Arguments:</h5>
2495
2496<p>
2497The first operand of an '<tt>insertelement</tt>' instruction is a
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2498value of <a href="#t_vector">vector</a> type. The second operand is a
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2499scalar value whose type must equal the element type of the first
2500operand. The third operand is an index indicating the position at
2501which to insert the value. The index may be a variable.</p>
2502
2503<h5>Semantics:</h5>
2504
2505<p>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2506The result is a vector of the same type as <tt>val</tt>. Its
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2507element values are those of <tt>val</tt> except at position
2508<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2509exceeds the length of <tt>val</tt>, the results are undefined.
2510</p>
2511
2512<h5>Example:</h5>
2513
2514<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2515 %result = insertelement &lt;4 x i32&gt; %vec, i32 1, i32 0 <i>; yields &lt;4 x i32&gt;</i>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2516</pre>
2517</div>
2518
2519<!-- _______________________________________________________________________ -->
2520<div class="doc_subsubsection">
2521 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2522</div>
2523
2524<div class="doc_text">
2525
2526<h5>Syntax:</h5>
2527
2528<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2529 &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;n x i32&gt; &lt;mask&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2530</pre>
2531
2532<h5>Overview:</h5>
2533
2534<p>
2535The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
2536from two input vectors, returning a vector of the same type.
2537</p>
2538
2539<h5>Arguments:</h5>
2540
2541<p>
2542The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2543with types that match each other and types that match the result of the
2544instruction. The third argument is a shuffle mask, which has the same number
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2545of elements as the other vector type, but whose element type is always 'i32'.
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2546</p>
2547
2548<p>
2549The shuffle mask operand is required to be a constant vector with either
2550constant integer or undef values.
2551</p>
2552
2553<h5>Semantics:</h5>
2554
2555<p>
2556The elements of the two input vectors are numbered from left to right across
2557both of the vectors. The shuffle mask operand specifies, for each element of
2558the result vector, which element of the two input registers the result element
2559gets. The element selector may be undef (meaning "don't care") and the second
2560operand may be undef if performing a shuffle from only one vector.
2561</p>
2562
2563<h5>Example:</h5>
2564
2565<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2566 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2567 &lt;4 x i32&gt; &lt;i32 0, i32 4, i32 1, i32 5&gt; <i>; yields &lt;4 x i32&gt;</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2568 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
2569 &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.
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2570</pre>
2571</div>
2572
Tanya Lattner09474292006-04-14 19:24:33 +0000[diff] [blame]2573
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2574<!-- ======================================================================= -->
2575<div class="doc_subsection">
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]2576 <a name="memoryops">Memory Access and Addressing Operations</a>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2577</div>
2578
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2579<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2580
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2581<p>A key design point of an SSA-based representation is how it
2582represents memory. In LLVM, no memory locations are in SSA form, which
2583makes things very simple. This section describes how to read, write,
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2584allocate, and free memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2585
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2586</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2587
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2588<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2589<div class="doc_subsubsection">
2590 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
2591</div>
2592
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2593<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2594
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2595<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2596
2597<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2598 &lt;result&gt; = malloc &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2599</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2600
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2601<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2602
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2603<p>The '<tt>malloc</tt>' instruction allocates memory from the system
2604heap and returns a pointer to it.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2605
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2606<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2607
2608<p>The '<tt>malloc</tt>' instruction allocates
2609<tt>sizeof(&lt;type&gt;)*NumElements</tt>
John Criswell6e4ca612004-02-24 16:13:56 +0000[diff] [blame]2610bytes of memory from the operating system and returns a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2611appropriate type to the program. If "NumElements" is specified, it is the
2612number of elements allocated. If an alignment is specified, the value result
2613of the allocation is guaranteed to be aligned to at least that boundary. If
2614not specified, or if zero, the target can choose to align the allocation on any
2615convenient boundary.</p>
2616
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2617<p>'<tt>type</tt>' must be a sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2618
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2619<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2620
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2621<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
2622a pointer is returned.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2623
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2624<h5>Example:</h5>
2625
2626<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2627 %array = malloc [4 x i8 ] <i>; yields {[%4 x i8]*}:array</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2628
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]2629 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
2630 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
2631 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
2632 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
2633 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2634</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2635</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2636
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2637<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2638<div class="doc_subsubsection">
2639 <a name="i_free">'<tt>free</tt>' Instruction</a>
2640</div>
2641
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2642<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2643
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2644<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2645
2646<pre>
2647 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2648</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2649
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2650<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2651
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2652<p>The '<tt>free</tt>' instruction returns memory back to the unused
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]2653memory heap to be reallocated in the future.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2654
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2655<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2656
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2657<p>'<tt>value</tt>' shall be a pointer value that points to a value
2658that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
2659instruction.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2660
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2661<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2662
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2663<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2664after this instruction executes.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2665
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2666<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2667
2668<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2669 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
2670 free [4 x i8]* %array
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2671</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2672</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2673
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2674<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2675<div class="doc_subsubsection">
2676 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
2677</div>
2678
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2679<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2680
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2681<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2682
2683<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2684 &lt;result&gt; = alloca &lt;type&gt;[, i32 &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2685</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2686
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2687<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2688
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2689<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
2690currently executing function, to be automatically released when this function
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2691returns to its caller.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2692
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2693<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2694
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2695<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2696bytes of memory on the runtime stack, returning a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2697appropriate type to the program. If "NumElements" is specified, it is the
2698number of elements allocated. If an alignment is specified, the value result
2699of the allocation is guaranteed to be aligned to at least that boundary. If
2700not specified, or if zero, the target can choose to align the allocation on any
2701convenient boundary.</p>
2702
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2703<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2704
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2705<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2706
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]2707<p>Memory is allocated; a pointer is returned. '<tt>alloca</tt>'d
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2708memory is automatically released when the function returns. The '<tt>alloca</tt>'
2709instruction is commonly used to represent automatic variables that must
2710have an address available. When the function returns (either with the <tt><a
John Criswelldae2e932005-05-12 16:55:34 +0000[diff] [blame]2711 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2712instructions), the memory is reclaimed.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2713
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2714<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2715
2716<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2717 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2718 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
2719 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2720 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2721</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2722</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2723
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2724<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2725<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
2726Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2727<div class="doc_text">
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2728<h5>Syntax:</h5>
Christopher Lamb2330e4d2007-04-21 08:16:25 +0000[diff] [blame]2729<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 Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2730<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2731<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2732<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2733<p>The argument to the '<tt>load</tt>' instruction specifies the memory
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]2734address from which to load. The pointer must point to a <a
Chris Lattnere53e5082004-06-03 22:57:15 +0000[diff] [blame]2735 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]2736marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2737the number or order of execution of this <tt>load</tt> with other
2738volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
2739instructions. </p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2740<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2741<p>The location of memory pointed to is loaded.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2742<h5>Examples:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2743<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2744 <a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2745 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
2746 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2747</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2748</div>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2749<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2750<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
2751Instruction</a> </div>
Reid Spencer035ab572006-11-09 21:18:01 +0000[diff] [blame]2752<div class="doc_text">
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2753<h5>Syntax:</h5>
Christopher Lamb2330e4d2007-04-21 08:16:25 +0000[diff] [blame]2754<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
2755 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2756</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2757<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2758<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2759<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2760<p>There are two arguments to the '<tt>store</tt>' instruction: a value
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2761to store and an address at which to store it. The type of the '<tt>&lt;pointer&gt;</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2762operand must be a pointer to the type of the '<tt>&lt;value&gt;</tt>'
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]2763operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2764optimizer is not allowed to modify the number or order of execution of
2765this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
2766 href="#i_store">store</a></tt> instructions.</p>
2767<h5>Semantics:</h5>
2768<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
2769at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2770<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2771<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2772 <a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2773 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
2774 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2775</pre>
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]2776</div>
2777
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2778<!-- _______________________________________________________________________ -->
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2779<div class="doc_subsubsection">
2780 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
2781</div>
2782
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2783<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2784<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2785<pre>
2786 &lt;result&gt; = getelementptr &lt;ty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
2787</pre>
2788
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2789<h5>Overview:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2790
2791<p>
2792The '<tt>getelementptr</tt>' instruction is used to get the address of a
2793subelement of an aggregate data structure.</p>
2794
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2795<h5>Arguments:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2796
Reid Spencer85f5b5b2006-12-04 21:29:24 +0000[diff] [blame]2797<p>This instruction takes a list of integer operands that indicate what
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2798elements of the aggregate object to index to. The actual types of the arguments
2799provided depend on the type of the first pointer argument. The
2800'<tt>getelementptr</tt>' instruction is used to index down through the type
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]2801levels of a structure or to a specific index in an array. When indexing into a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2802structure, only <tt>i32</tt> integer constants are allowed. When indexing
Reid Spencer85f5b5b2006-12-04 21:29:24 +0000[diff] [blame]2803into an array or pointer, only integers of 32 or 64 bits are allowed, and will
2804be sign extended to 64-bit values.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2805
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2806<p>For example, let's consider a C code fragment and how it gets
2807compiled to LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2808
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2809<div class="doc_code">
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2810<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2811struct RT {
2812 char A;
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]2813 int B[10][20];
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2814 char C;
2815};
2816struct ST {
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]2817 int X;
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2818 double Y;
2819 struct RT Z;
2820};
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2821
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]2822int *foo(struct ST *s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2823 return &amp;s[1].Z.B[5][13];
2824}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2825</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2826</div>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2827
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2828<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2829
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2830<div class="doc_code">
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2831<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2832%RT = type { i8 , [10 x [20 x i32]], i8 }
2833%ST = type { i32, double, %RT }
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2834
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2835define i32* %foo(%ST* %s) {
2836entry:
2837 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
2838 ret i32* %reg
2839}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2840</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2841</div>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2842
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2843<h5>Semantics:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2844
2845<p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]2846on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
Reid Spencer85f5b5b2006-12-04 21:29:24 +0000[diff] [blame]2847and <a href="#t_array">array</a> types can use a 32-bit or 64-bit
Reid Spencer42ddd842006-12-03 16:53:48 +0000[diff] [blame]2848<a href="#t_integer">integer</a> type but the value will always be sign extended
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2849to 64-bits. <a href="#t_struct">Structure</a> types require <tt>i32</tt>
Reid Spencer42ddd842006-12-03 16:53:48 +0000[diff] [blame]2850<b>constants</b>.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2851
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2852<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2853type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2854}</tt>' type, a structure. The second index indexes into the third element of
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2855the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
2856i8 }</tt>' type, another structure. The third index indexes into the second
2857element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2858array. The two dimensions of the array are subscripted into, yielding an
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2859'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
2860to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2861
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2862<p>Note that it is perfectly legal to index partially through a
2863structure, returning a pointer to an inner element. Because of this,
2864the LLVM code for the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2865
2866<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2867 define i32* %foo(%ST* %s) {
2868 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2869 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
2870 %t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2871 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
2872 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
2873 ret i32* %t5
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2874 }
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2875</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]2876
2877<p>Note that it is undefined to access an array out of bounds: array and
2878pointer indexes must always be within the defined bounds of the array type.
2879The one exception for this rules is zero length arrays. These arrays are
2880defined to be accessible as variable length arrays, which requires access
2881beyond the zero'th element.</p>
2882
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]2883<p>The getelementptr instruction is often confusing. For some more insight
2884into how it works, see <a href="GetElementPtr.html">the getelementptr
2885FAQ</a>.</p>
2886
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2887<h5>Example:</h5>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]2888
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2889<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2890 <i>; yields [12 x i8]*:aptr</i>
2891 %aptr = getelementptr {i32, [12 x i8]}* %sptr, i64 0, i32 1
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2892</pre>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2893</div>
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]2894
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2895<!-- ======================================================================= -->
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]2896<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2897</div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2898<div class="doc_text">
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]2899<p>The instructions in this category are the conversion instructions (casting)
2900which all take a single operand and a type. They perform various bit conversions
2901on the operand.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2902</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2903
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2904<!-- _______________________________________________________________________ -->
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2905<div class="doc_subsubsection">
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2906 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
2907</div>
2908<div class="doc_text">
2909
2910<h5>Syntax:</h5>
2911<pre>
2912 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2913</pre>
2914
2915<h5>Overview:</h5>
2916<p>
2917The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
2918</p>
2919
2920<h5>Arguments:</h5>
2921<p>
2922The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
2923be an <a href="#t_integer">integer</a> type, and a type that specifies the size
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]2924and type of the result, which must be an <a href="#t_integer">integer</a>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]2925type. The bit size of <tt>value</tt> must be larger than the bit size of
2926<tt>ty2</tt>. Equal sized types are not allowed.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2927
2928<h5>Semantics:</h5>
2929<p>
2930The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]2931and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
2932larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
2933It will always truncate bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2934
2935<h5>Example:</h5>
2936<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2937 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]2938 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
2939 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2940</pre>
2941</div>
2942
2943<!-- _______________________________________________________________________ -->
2944<div class="doc_subsubsection">
2945 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
2946</div>
2947<div class="doc_text">
2948
2949<h5>Syntax:</h5>
2950<pre>
2951 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2952</pre>
2953
2954<h5>Overview:</h5>
2955<p>The '<tt>zext</tt>' instruction zero extends its operand to type
2956<tt>ty2</tt>.</p>
2957
2958
2959<h5>Arguments:</h5>
2960<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]2961<a href="#t_integer">integer</a> type, and a type to cast it to, which must
2962also be of <a href="#t_integer">integer</a> type. The bit size of the
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]2963<tt>value</tt> must be smaller than the bit size of the destination type,
2964<tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2965
2966<h5>Semantics:</h5>
2967<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
Chris Lattnerd1d25172007-05-24 19:13:27 +0000[diff] [blame]2968bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2969
Reid Spencerb5929522007-01-12 15:46:11 +0000[diff] [blame]2970<p>When zero extending from i1, the result will always be either 0 or 1.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2971
2972<h5>Example:</h5>
2973<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2974 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]2975 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2976</pre>
2977</div>
2978
2979<!-- _______________________________________________________________________ -->
2980<div class="doc_subsubsection">
2981 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
2982</div>
2983<div class="doc_text">
2984
2985<h5>Syntax:</h5>
2986<pre>
2987 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2988</pre>
2989
2990<h5>Overview:</h5>
2991<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
2992
2993<h5>Arguments:</h5>
2994<p>
2995The '<tt>sext</tt>' instruction takes a value to cast, which must be of
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]2996<a href="#t_integer">integer</a> type, and a type to cast it to, which must
2997also be of <a href="#t_integer">integer</a> type. The bit size of the
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]2998<tt>value</tt> must be smaller than the bit size of the destination type,
2999<tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3000
3001<h5>Semantics:</h5>
3002<p>
3003The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3004bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
Chris Lattnerd1d25172007-05-24 19:13:27 +0000[diff] [blame]3005the type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3006
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3007<p>When sign extending from i1, the extension always results in -1 or 0.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3008
3009<h5>Example:</h5>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3010<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3011 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3012 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3013</pre>
3014</div>
3015
3016<!-- _______________________________________________________________________ -->
3017<div class="doc_subsubsection">
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]3018 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3019</div>
3020
3021<div class="doc_text">
3022
3023<h5>Syntax:</h5>
3024
3025<pre>
3026 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3027</pre>
3028
3029<h5>Overview:</h5>
3030<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3031<tt>ty2</tt>.</p>
3032
3033
3034<h5>Arguments:</h5>
3035<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
3036 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
3037cast it to. The size of <tt>value</tt> must be larger than the size of
3038<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
3039<i>no-op cast</i>.</p>
3040
3041<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3042<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
3043<a href="#t_floating">floating point</a> type to a smaller
3044<a href="#t_floating">floating point</a> type. If the value cannot fit within
3045the destination type, <tt>ty2</tt>, then the results are undefined.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]3046
3047<h5>Example:</h5>
3048<pre>
3049 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
3050 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
3051</pre>
3052</div>
3053
3054<!-- _______________________________________________________________________ -->
3055<div class="doc_subsubsection">
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3056 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
3057</div>
3058<div class="doc_text">
3059
3060<h5>Syntax:</h5>
3061<pre>
3062 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3063</pre>
3064
3065<h5>Overview:</h5>
3066<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
3067floating point value.</p>
3068
3069<h5>Arguments:</h5>
3070<p>The '<tt>fpext</tt>' instruction takes a
3071<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3072and a <a href="#t_floating">floating point</a> type to cast it to. The source
3073type must be smaller than the destination type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3074
3075<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3076<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3077<a href="#t_floating">floating point</a> type to a larger
3078<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3079used to make a <i>no-op cast</i> because it always changes bits. Use
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]3080<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3081
3082<h5>Example:</h5>
3083<pre>
3084 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
3085 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
3086</pre>
3087</div>
3088
3089<!-- _______________________________________________________________________ -->
3090<div class="doc_subsubsection">
Reid Spencer24d6da52007-01-21 00:29:26 +0000[diff] [blame]3091 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3092</div>
3093<div class="doc_text">
3094
3095<h5>Syntax:</h5>
3096<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]3097 &lt;result&gt; = fptoui &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3098</pre>
3099
3100<h5>Overview:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]3101<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3102unsigned integer equivalent of type <tt>ty2</tt>.
3103</p>
3104
3105<h5>Arguments:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]3106<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3107<a href="#t_floating">floating point</a> value, and a type to cast it to, which
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]3108must be an <a href="#t_integer">integer</a> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3109
3110<h5>Semantics:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]3111<p> The '<tt>fptoui</tt>' instruction converts its
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3112<a href="#t_floating">floating point</a> operand into the nearest (rounding
3113towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
3114the results are undefined.</p>
3115
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3116<h5>Example:</h5>
3117<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]3118 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]3119 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]3120 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3121</pre>
3122</div>
3123
3124<!-- _______________________________________________________________________ -->
3125<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3126 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3127</div>
3128<div class="doc_text">
3129
3130<h5>Syntax:</h5>
3131<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3132 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3133</pre>
3134
3135<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3136<p>The '<tt>fptosi</tt>' instruction converts
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3137<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3138</p>
3139
3140
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3141<h5>Arguments:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3142<p> The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3143<a href="#t_floating">floating point</a> value, and a type to cast it to, which
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]3144must also be an <a href="#t_integer">integer</a> type.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3145
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3146<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3147<p>The '<tt>fptosi</tt>' instruction converts its
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3148<a href="#t_floating">floating point</a> operand into the nearest (rounding
3149towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3150the results are undefined.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3151
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]3152<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3153<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3154 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]3155 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3156 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3157</pre>
3158</div>
3159
3160<!-- _______________________________________________________________________ -->
3161<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3162 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3163</div>
3164<div class="doc_text">
3165
3166<h5>Syntax:</h5>
3167<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3168 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3169</pre>
3170
3171<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3172<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3173integer and converts that value to the <tt>ty2</tt> type.</p>
3174
3175
3176<h5>Arguments:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3177<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be an
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]3178<a href="#t_integer">integer</a> value, and a type to cast it to, which must
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3179be a <a href="#t_floating">floating point</a> type.</p>
3180
3181<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3182<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3183integer quantity and converts it to the corresponding floating point value. If
Jeff Cohencb757312007-04-22 14:56:37 +0000[diff] [blame]3184the value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3185
3186
3187<h5>Example:</h5>
3188<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3189 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3190 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3191</pre>
3192</div>
3193
3194<!-- _______________________________________________________________________ -->
3195<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3196 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3197</div>
3198<div class="doc_text">
3199
3200<h5>Syntax:</h5>
3201<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3202 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3203</pre>
3204
3205<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3206<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3207integer and converts that value to the <tt>ty2</tt> type.</p>
3208
3209<h5>Arguments:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3210<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be an
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]3211<a href="#t_integer">integer</a> value, and a type to cast it to, which must be
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3212a <a href="#t_floating">floating point</a> type.</p>
3213
3214<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3215<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3216integer quantity and converts it to the corresponding floating point value. If
Jeff Cohencb757312007-04-22 14:56:37 +0000[diff] [blame]3217the value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3218
3219<h5>Example:</h5>
3220<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3221 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3222 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3223</pre>
3224</div>
3225
3226<!-- _______________________________________________________________________ -->
3227<div class="doc_subsubsection">
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]3228 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
3229</div>
3230<div class="doc_text">
3231
3232<h5>Syntax:</h5>
3233<pre>
3234 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3235</pre>
3236
3237<h5>Overview:</h5>
3238<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
3239the integer type <tt>ty2</tt>.</p>
3240
3241<h5>Arguments:</h5>
3242<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3243must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]3244<tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.
3245
3246<h5>Semantics:</h5>
3247<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
3248<tt>ty2</tt> by interpreting the pointer value as an integer and either
3249truncating or zero extending that value to the size of the integer type. If
3250<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
3251<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3252are the same size, then nothing is done (<i>no-op cast</i>) other than a type
3253change.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]3254
3255<h5>Example:</h5>
3256<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3257 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
3258 %Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]3259</pre>
3260</div>
3261
3262<!-- _______________________________________________________________________ -->
3263<div class="doc_subsubsection">
3264 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
3265</div>
3266<div class="doc_text">
3267
3268<h5>Syntax:</h5>
3269<pre>
3270 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3271</pre>
3272
3273<h5>Overview:</h5>
3274<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
3275a pointer type, <tt>ty2</tt>.</p>
3276
3277<h5>Arguments:</h5>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3278<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]3279value to cast, and a type to cast it to, which must be a
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]3280<a href="#t_pointer">pointer</a> type.
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]3281
3282<h5>Semantics:</h5>
3283<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
3284<tt>ty2</tt> by applying either a zero extension or a truncation depending on
3285the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
3286size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
3287the size of a pointer then a zero extension is done. If they are the same size,
3288nothing is done (<i>no-op cast</i>).</p>
3289
3290<h5>Example:</h5>
3291<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3292 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
3293 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
3294 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]3295</pre>
3296</div>
3297
3298<!-- _______________________________________________________________________ -->
3299<div class="doc_subsubsection">
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]3300 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3301</div>
3302<div class="doc_text">
3303
3304<h5>Syntax:</h5>
3305<pre>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]3306 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3307</pre>
3308
3309<h5>Overview:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]3310<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3311<tt>ty2</tt> without changing any bits.</p>
3312
3313<h5>Arguments:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]3314<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3315a first class value, and a type to cast it to, which must also be a <a
3316 href="#t_firstclass">first class</a> type. The bit sizes of <tt>value</tt>
Reid Spencer19b569f2007-01-09 20:08:58 +0000[diff] [blame]3317and the destination type, <tt>ty2</tt>, must be identical. If the source
3318type is a pointer, the destination type must also be a pointer.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3319
3320<h5>Semantics:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]3321<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]3322<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
3323this conversion. The conversion is done as if the <tt>value</tt> had been
3324stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
3325converted to other pointer types with this instruction. To convert pointers to
3326other types, use the <a href="#i_inttoptr">inttoptr</a> or
3327<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3328
3329<h5>Example:</h5>
3330<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3331 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3332 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
3333 %Z = bitcast <2xint> %V to i64; <i>; yields i64: %V</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]3334</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3335</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3336
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]3337<!-- ======================================================================= -->
3338<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
3339<div class="doc_text">
3340<p>The instructions in this category are the "miscellaneous"
3341instructions, which defy better classification.</p>
3342</div>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3343
3344<!-- _______________________________________________________________________ -->
3345<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
3346</div>
3347<div class="doc_text">
3348<h5>Syntax:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3349<pre> &lt;result&gt; = icmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {i1}:result</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3350</pre>
3351<h5>Overview:</h5>
3352<p>The '<tt>icmp</tt>' instruction returns a boolean value based on comparison
3353of its two integer operands.</p>
3354<h5>Arguments:</h5>
3355<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3356the condition code indicating the kind of comparison to perform. It is not
3357a value, just a keyword. The possible condition code are:
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3358<ol>
3359 <li><tt>eq</tt>: equal</li>
3360 <li><tt>ne</tt>: not equal </li>
3361 <li><tt>ugt</tt>: unsigned greater than</li>
3362 <li><tt>uge</tt>: unsigned greater or equal</li>
3363 <li><tt>ult</tt>: unsigned less than</li>
3364 <li><tt>ule</tt>: unsigned less or equal</li>
3365 <li><tt>sgt</tt>: signed greater than</li>
3366 <li><tt>sge</tt>: signed greater or equal</li>
3367 <li><tt>slt</tt>: signed less than</li>
3368 <li><tt>sle</tt>: signed less or equal</li>
3369</ol>
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]3370<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Reid Spencer350f8aa2007-01-04 05:19:58 +0000[diff] [blame]3371<a href="#t_pointer">pointer</a> typed. They must also be identical types.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3372<h5>Semantics:</h5>
3373<p>The '<tt>icmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3374the condition code given as <tt>cond</tt>. The comparison performed always
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3375yields a <a href="#t_primitive">i1</a> result, as follows:
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3376<ol>
3377 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
3378 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3379 </li>
3380 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
3381 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3382 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
3383 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3384 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
3385 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3386 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
3387 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3388 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
3389 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3390 <li><tt>sgt</tt>: interprets the operands as signed values and yields
3391 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3392 <li><tt>sge</tt>: interprets the operands as signed values and yields
3393 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3394 <li><tt>slt</tt>: interprets the operands as signed values and yields
3395 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3396 <li><tt>sle</tt>: interprets the operands as signed values and yields
3397 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3398</ol>
3399<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3400values are compared as if they were integers.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3401
3402<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3403<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
3404 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
3405 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
3406 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
3407 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
3408 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3409</pre>
3410</div>
3411
3412<!-- _______________________________________________________________________ -->
3413<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
3414</div>
3415<div class="doc_text">
3416<h5>Syntax:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3417<pre> &lt;result&gt; = fcmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {i1}:result</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3418</pre>
3419<h5>Overview:</h5>
3420<p>The '<tt>fcmp</tt>' instruction returns a boolean value based on comparison
3421of its floating point operands.</p>
3422<h5>Arguments:</h5>
3423<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3424the condition code indicating the kind of comparison to perform. It is not
3425a value, just a keyword. The possible condition code are:
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3426<ol>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3427 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3428 <li><tt>oeq</tt>: ordered and equal</li>
3429 <li><tt>ogt</tt>: ordered and greater than </li>
3430 <li><tt>oge</tt>: ordered and greater than or equal</li>
3431 <li><tt>olt</tt>: ordered and less than </li>
3432 <li><tt>ole</tt>: ordered and less than or equal</li>
3433 <li><tt>one</tt>: ordered and not equal</li>
3434 <li><tt>ord</tt>: ordered (no nans)</li>
3435 <li><tt>ueq</tt>: unordered or equal</li>
3436 <li><tt>ugt</tt>: unordered or greater than </li>
3437 <li><tt>uge</tt>: unordered or greater than or equal</li>
3438 <li><tt>ult</tt>: unordered or less than </li>
3439 <li><tt>ule</tt>: unordered or less than or equal</li>
3440 <li><tt>une</tt>: unordered or not equal</li>
3441 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3442 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3443</ol>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3444<p><i>Ordered</i> means that neither operand is a QNAN while
Reid Spencer93a49852006-12-06 07:08:07 +0000[diff] [blame]3445<i>unordered</i> means that either operand may be a QNAN.</p>
Reid Spencer350f8aa2007-01-04 05:19:58 +0000[diff] [blame]3446<p>The <tt>val1</tt> and <tt>val2</tt> arguments must be
3447<a href="#t_floating">floating point</a> typed. They must have identical
3448types.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3449<h5>Semantics:</h5>
3450<p>The '<tt>fcmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3451the condition code given as <tt>cond</tt>. The comparison performed always
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3452yields a <a href="#t_primitive">i1</a> result, as follows:
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3453<ol>
3454 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3455 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3456 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3457 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3458 <tt>var1</tt> is greather than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3459 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3460 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3461 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3462 <tt>var1</tt> is less than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3463 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3464 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3465 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3466 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3467 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
3468 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3469 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3470 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3471 <tt>var1</tt> is greater than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3472 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3473 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3474 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3475 <tt>var1</tt> is less than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3476 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3477 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3478 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3479 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3480 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3481 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
3482</ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3483
3484<h5>Example:</h5>
3485<pre> &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
3486 &lt;result&gt; = icmp one float 4.0, 5.0 <i>; yields: result=true</i>
3487 &lt;result&gt; = icmp olt float 4.0, 5.0 <i>; yields: result=true</i>
3488 &lt;result&gt; = icmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
3489</pre>
3490</div>
3491
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]3492<!-- _______________________________________________________________________ -->
3493<div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
3494Instruction</a> </div>
3495<div class="doc_text">
3496<h5>Syntax:</h5>
3497<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
3498<h5>Overview:</h5>
3499<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
3500the SSA graph representing the function.</p>
3501<h5>Arguments:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3502<p>The type of the incoming values is specified with the first type
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]3503field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
3504as arguments, with one pair for each predecessor basic block of the
3505current block. Only values of <a href="#t_firstclass">first class</a>
3506type may be used as the value arguments to the PHI node. Only labels
3507may be used as the label arguments.</p>
3508<p>There must be no non-phi instructions between the start of a basic
3509block and the PHI instructions: i.e. PHI instructions must be first in
3510a basic block.</p>
3511<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3512<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
3513specified by the pair corresponding to the predecessor basic block that executed
3514just prior to the current block.</p>
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]3515<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3516<pre>Loop: ; Infinite loop that counts from 0 on up...<br> %indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]<br> %nextindvar = add i32 %indvar, 1<br> br label %Loop<br></pre>
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]3517</div>
3518
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3519<!-- _______________________________________________________________________ -->
3520<div class="doc_subsubsection">
3521 <a name="i_select">'<tt>select</tt>' Instruction</a>
3522</div>
3523
3524<div class="doc_text">
3525
3526<h5>Syntax:</h5>
3527
3528<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3529 &lt;result&gt; = select i1 &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3530</pre>
3531
3532<h5>Overview:</h5>
3533
3534<p>
3535The '<tt>select</tt>' instruction is used to choose one value based on a
3536condition, without branching.
3537</p>
3538
3539
3540<h5>Arguments:</h5>
3541
3542<p>
3543The '<tt>select</tt>' instruction requires a boolean value indicating the condition, and two values of the same <a href="#t_firstclass">first class</a> type.
3544</p>
3545
3546<h5>Semantics:</h5>
3547
3548<p>
3549If the boolean condition evaluates to true, the instruction returns the first
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]3550value argument; otherwise, it returns the second value argument.
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3551</p>
3552
3553<h5>Example:</h5>
3554
3555<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3556 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3557</pre>
3558</div>
3559
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]3560
3561<!-- _______________________________________________________________________ -->
3562<div class="doc_subsubsection">
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3563 <a name="i_call">'<tt>call</tt>' Instruction</a>
3564</div>
3565
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3566<div class="doc_text">
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3567
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3568<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3569<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +0000[diff] [blame]3570 &lt;result&gt; = [tail] call [<a href="#callingconv">cconv</a>] &lt;ty&gt; [&lt;fnty&gt;*] &lt;fnptrval&gt;(&lt;param list&gt;)
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3571</pre>
3572
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3573<h5>Overview:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3574
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3575<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3576
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3577<h5>Arguments:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3578
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3579<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3580
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3581<ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3582 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3583 <p>The optional "tail" marker indicates whether the callee function accesses
3584 any allocas or varargs in the caller. If the "tail" marker is present, the
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3585 function call is eligible for tail call optimization. Note that calls may
3586 be marked "tail" even if they do not occur before a <a
3587 href="#i_ret"><tt>ret</tt></a> instruction.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3588 </li>
3589 <li>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3590 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3591 convention</a> the call should use. If none is specified, the call defaults
3592 to using C calling conventions.
3593 </li>
3594 <li>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +0000[diff] [blame]3595 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
3596 the type of the return value. Functions that return no value are marked
3597 <tt><a href="#t_void">void</a></tt>.</p>
3598 </li>
3599 <li>
3600 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
3601 value being invoked. The argument types must match the types implied by
3602 this signature. This type can be omitted if the function is not varargs
3603 and if the function type does not return a pointer to a function.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3604 </li>
3605 <li>
3606 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
3607 be invoked. In most cases, this is a direct function invocation, but
3608 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]3609 to function value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3610 </li>
3611 <li>
3612 <p>'<tt>function args</tt>': argument list whose types match the
Reid Spencera7e302a2005-05-01 22:22:57 +0000[diff] [blame]3613 function signature argument types. All arguments must be of
3614 <a href="#t_firstclass">first class</a> type. If the function signature
3615 indicates the function accepts a variable number of arguments, the extra
3616 arguments can be specified.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3617 </li>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3618</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3619
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3620<h5>Semantics:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3621
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3622<p>The '<tt>call</tt>' instruction is used to cause control flow to
3623transfer to a specified function, with its incoming arguments bound to
3624the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
3625instruction in the called function, control flow continues with the
3626instruction after the function call, and the return value of the
3627function is bound to the result argument. This is a simpler case of
3628the <a href="#i_invoke">invoke</a> instruction.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3629
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3630<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3631
3632<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +0000[diff] [blame]3633 %retval = call i32 @test(i32 %argc)
3634 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42);
3635 %X = tail call i32 @foo()
3636 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo()
3637 %Z = call void %foo(i8 97 signext)
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3638</pre>
3639
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3640</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3641
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3642<!-- _______________________________________________________________________ -->
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3643<div class="doc_subsubsection">
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]3644 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3645</div>
3646
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3647<div class="doc_text">
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3648
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]3649<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3650
3651<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3652 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3653</pre>
3654
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]3655<h5>Overview:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3656
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3657<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3658the "variable argument" area of a function call. It is used to implement the
3659<tt>va_arg</tt> macro in C.</p>
3660
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]3661<h5>Arguments:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3662
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3663<p>This instruction takes a <tt>va_list*</tt> value and the type of
3664the argument. It returns a value of the specified argument type and
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3665increments the <tt>va_list</tt> to point to the next argument. The
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3666actual type of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3667
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]3668<h5>Semantics:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3669
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3670<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
3671type from the specified <tt>va_list</tt> and causes the
3672<tt>va_list</tt> to point to the next argument. For more information,
3673see the variable argument handling <a href="#int_varargs">Intrinsic
3674Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3675
3676<p>It is legal for this instruction to be called in a function which does not
3677take a variable number of arguments, for example, the <tt>vfprintf</tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3678function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3679
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3680<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]3681href="#intrinsics">intrinsic function</a> because it takes a type as an
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3682argument.</p>
3683
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]3684<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3685
3686<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
3687
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3688</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3689
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3690<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3691<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
3692<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3693
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3694<div class="doc_text">
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3695
3696<p>LLVM supports the notion of an "intrinsic function". These functions have
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]3697well known names and semantics and are required to follow certain restrictions.
3698Overall, these intrinsics represent an extension mechanism for the LLVM
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3699language that does not require changing all of the transformations in LLVM when
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]3700adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3701
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]3702<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3703prefix is reserved in LLVM for intrinsic names; thus, function names may not
3704begin with this prefix. Intrinsic functions must always be external functions:
3705you cannot define the body of intrinsic functions. Intrinsic functions may
3706only be used in call or invoke instructions: it is illegal to take the address
3707of an intrinsic function. Additionally, because intrinsic functions are part
3708of the LLVM language, it is required if any are added that they be documented
3709here.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3710
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]3711<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
3712a family of functions that perform the same operation but on different data
3713types. Because LLVM can represent over 8 million different integer types,
3714overloading is used commonly to allow an intrinsic function to operate on any
3715integer type. One or more of the argument types or the result type can be
3716overloaded to accept any integer type. Argument types may also be defined as
3717exactly matching a previous argument's type or the result type. This allows an
3718intrinsic function which accepts multiple arguments, but needs all of them to
3719be of the same type, to only be overloaded with respect to a single argument or
3720the result.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3721
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]3722<p>Overloaded intrinsics will have the names of its overloaded argument types
3723encoded into its function name, each preceded by a period. Only those types
3724which are overloaded result in a name suffix. Arguments whose type is matched
3725against another type do not. For example, the <tt>llvm.ctpop</tt> function can
3726take an integer of any width and returns an integer of exactly the same integer
3727width. This leads to a family of functions such as
3728<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
3729Only one type, the return type, is overloaded, and only one type suffix is
3730required. Because the argument's type is matched against the return type, it
3731does not require its own name suffix.</p>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]3732
3733<p>To learn how to add an intrinsic function, please see the
3734<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3735</p>
3736
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3737</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3738
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3739<!-- ======================================================================= -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3740<div class="doc_subsection">
3741 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
3742</div>
3743
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3744<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3745
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3746<p>Variable argument support is defined in LLVM with the <a
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]3747 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3748intrinsic functions. These functions are related to the similarly
3749named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3750
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3751<p>All of these functions operate on arguments that use a
3752target-specific value type "<tt>va_list</tt>". The LLVM assembly
3753language reference manual does not define what this type is, so all
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3754transformations should be prepared to handle these functions regardless of
3755the type used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3756
Chris Lattner374ab302006-05-15 17:26:46 +0000[diff] [blame]3757<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3758instruction and the variable argument handling intrinsic functions are
3759used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3760
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]3761<div class="doc_code">
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3762<pre>
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]3763define i32 @test(i32 %X, ...) {
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3764 ; Initialize variable argument processing
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3765 %ap = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3766 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]3767 call void @llvm.va_start(i8* %ap2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3768
3769 ; Read a single integer argument
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3770 %tmp = va_arg i8** %ap, i32
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3771
3772 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3773 %aq = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3774 %aq2 = bitcast i8** %aq to i8*
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3775 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]3776 call void @llvm.va_end(i8* %aq2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3777
3778 ; Stop processing of arguments.
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]3779 call void @llvm.va_end(i8* %ap2)
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3780 ret i32 %tmp
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3781}
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]3782
3783declare void @llvm.va_start(i8*)
3784declare void @llvm.va_copy(i8*, i8*)
3785declare void @llvm.va_end(i8*)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3786</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3787</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3788
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]3789</div>
3790
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3791<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3792<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3793 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3794</div>
3795
3796
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3797<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3798<h5>Syntax:</h5>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3799<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3800<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3801<P>The '<tt>llvm.va_start</tt>' intrinsic initializes
3802<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
3803href="#i_va_arg">va_arg</a></tt>.</p>
3804
3805<h5>Arguments:</h5>
3806
3807<P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
3808
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3809<h5>Semantics:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3810
3811<P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
3812macro available in C. In a target-dependent way, it initializes the
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3813<tt>va_list</tt> element to which the argument points, so that the next call to
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3814<tt>va_arg</tt> will produce the first variable argument passed to the function.
3815Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3816last argument of the function as the compiler can figure that out.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3817
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3818</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3819
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3820<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3821<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3822 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3823</div>
3824
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3825<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3826<h5>Syntax:</h5>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]3827<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3828<h5>Overview:</h5>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3829
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3830<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3831which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3832or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3833
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3834<h5>Arguments:</h5>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3835
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3836<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3837
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3838<h5>Semantics:</h5>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3839
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3840<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3841macro available in C. In a target-dependent way, it destroys the
3842<tt>va_list</tt> element to which the argument points. Calls to <a
3843href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
3844<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
3845<tt>llvm.va_end</tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3846
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3847</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3848
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3849<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3850<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3851 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3852</div>
3853
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3854<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3855
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3856<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3857
3858<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]3859 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3860</pre>
3861
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3862<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3863
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3864<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
3865from the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3866
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3867<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3868
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3869<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Andrew Lenharthd0a4c622005-06-22 20:38:11 +0000[diff] [blame]3870The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3871
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3872
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3873<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3874
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3875<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
3876macro available in C. In a target-dependent way, it copies the source
3877<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
3878intrinsic is necessary because the <tt><a href="#int_va_start">
3879llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
3880example, memory allocation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3881
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3882</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3883
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3884<!-- ======================================================================= -->
3885<div class="doc_subsection">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3886 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
3887</div>
3888
3889<div class="doc_text">
3890
3891<p>
3892LLVM support for <a href="GarbageCollection.html">Accurate Garbage
3893Collection</a> requires the implementation and generation of these intrinsics.
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3894These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3895stack</a>, as well as garbage collector implementations that require <a
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3896href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3897Front-ends for type-safe garbage collected languages should generate these
3898intrinsics to make use of the LLVM garbage collectors. For more details, see <a
3899href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
3900</p>
3901</div>
3902
3903<!-- _______________________________________________________________________ -->
3904<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3905 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3906</div>
3907
3908<div class="doc_text">
3909
3910<h5>Syntax:</h5>
3911
3912<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]3913 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3914</pre>
3915
3916<h5>Overview:</h5>
3917
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]3918<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3919the code generator, and allows some metadata to be associated with it.</p>
3920
3921<h5>Arguments:</h5>
3922
3923<p>The first argument specifies the address of a stack object that contains the
3924root pointer. The second pointer (which must be either a constant or a global
3925value address) contains the meta-data to be associated with the root.</p>
3926
3927<h5>Semantics:</h5>
3928
3929<p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
3930location. At compile-time, the code generator generates information to allow
3931the runtime to find the pointer at GC safe points.
3932</p>
3933
3934</div>
3935
3936
3937<!-- _______________________________________________________________________ -->
3938<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3939 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3940</div>
3941
3942<div class="doc_text">
3943
3944<h5>Syntax:</h5>
3945
3946<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]3947 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3948</pre>
3949
3950<h5>Overview:</h5>
3951
3952<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
3953locations, allowing garbage collector implementations that require read
3954barriers.</p>
3955
3956<h5>Arguments:</h5>
3957
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]3958<p>The second argument is the address to read from, which should be an address
3959allocated from the garbage collector. The first object is a pointer to the
3960start of the referenced object, if needed by the language runtime (otherwise
3961null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3962
3963<h5>Semantics:</h5>
3964
3965<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
3966instruction, but may be replaced with substantially more complex code by the
3967garbage collector runtime, as needed.</p>
3968
3969</div>
3970
3971
3972<!-- _______________________________________________________________________ -->
3973<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3974 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3975</div>
3976
3977<div class="doc_text">
3978
3979<h5>Syntax:</h5>
3980
3981<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]3982 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3983</pre>
3984
3985<h5>Overview:</h5>
3986
3987<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
3988locations, allowing garbage collector implementations that require write
3989barriers (such as generational or reference counting collectors).</p>
3990
3991<h5>Arguments:</h5>
3992
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]3993<p>The first argument is the reference to store, the second is the start of the
3994object to store it to, and the third is the address of the field of Obj to
3995store to. If the runtime does not require a pointer to the object, Obj may be
3996null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3997
3998<h5>Semantics:</h5>
3999
4000<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
4001instruction, but may be replaced with substantially more complex code by the
4002garbage collector runtime, as needed.</p>
4003
4004</div>
4005
4006
4007
4008<!-- ======================================================================= -->
4009<div class="doc_subsection">
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4010 <a name="int_codegen">Code Generator Intrinsics</a>
4011</div>
4012
4013<div class="doc_text">
4014<p>
4015These intrinsics are provided by LLVM to expose special features that may only
4016be implemented with code generator support.
4017</p>
4018
4019</div>
4020
4021<!-- _______________________________________________________________________ -->
4022<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4023 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4024</div>
4025
4026<div class="doc_text">
4027
4028<h5>Syntax:</h5>
4029<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4030 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4031</pre>
4032
4033<h5>Overview:</h5>
4034
4035<p>
Chris Lattner32b5d712006-10-15 20:05:59 +0000[diff] [blame]4036The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
4037target-specific value indicating the return address of the current function
4038or one of its callers.
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4039</p>
4040
4041<h5>Arguments:</h5>
4042
4043<p>
4044The argument to this intrinsic indicates which function to return the address
4045for. Zero indicates the calling function, one indicates its caller, etc. The
4046argument is <b>required</b> to be a constant integer value.
4047</p>
4048
4049<h5>Semantics:</h5>
4050
4051<p>
4052The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
4053the return address of the specified call frame, or zero if it cannot be
4054identified. The value returned by this intrinsic is likely to be incorrect or 0
4055for arguments other than zero, so it should only be used for debugging purposes.
4056</p>
4057
4058<p>
4059Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +0000[diff] [blame]4060aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4061source-language caller.
4062</p>
4063</div>
4064
4065
4066<!-- _______________________________________________________________________ -->
4067<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4068 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4069</div>
4070
4071<div class="doc_text">
4072
4073<h5>Syntax:</h5>
4074<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]4075 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4076</pre>
4077
4078<h5>Overview:</h5>
4079
4080<p>
Chris Lattner32b5d712006-10-15 20:05:59 +0000[diff] [blame]4081The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
4082target-specific frame pointer value for the specified stack frame.
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4083</p>
4084
4085<h5>Arguments:</h5>
4086
4087<p>
4088The argument to this intrinsic indicates which function to return the frame
4089pointer for. Zero indicates the calling function, one indicates its caller,
4090etc. The argument is <b>required</b> to be a constant integer value.
4091</p>
4092
4093<h5>Semantics:</h5>
4094
4095<p>
4096The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
4097the frame address of the specified call frame, or zero if it cannot be
4098identified. The value returned by this intrinsic is likely to be incorrect or 0
4099for arguments other than zero, so it should only be used for debugging purposes.
4100</p>
4101
4102<p>
4103Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +0000[diff] [blame]4104aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4105source-language caller.
4106</p>
4107</div>
4108
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]4109<!-- _______________________________________________________________________ -->
4110<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4111 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4112</div>
4113
4114<div class="doc_text">
4115
4116<h5>Syntax:</h5>
4117<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]4118 declare i8 *@llvm.stacksave()
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4119</pre>
4120
4121<h5>Overview:</h5>
4122
4123<p>
4124The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4125the function stack, for use with <a href="#int_stackrestore">
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4126<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
4127features like scoped automatic variable sized arrays in C99.
4128</p>
4129
4130<h5>Semantics:</h5>
4131
4132<p>
4133This intrinsic returns a opaque pointer value that can be passed to <a
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4134href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4135<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
4136<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
4137state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
4138practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
4139that were allocated after the <tt>llvm.stacksave</tt> was executed.
4140</p>
4141
4142</div>
4143
4144<!-- _______________________________________________________________________ -->
4145<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4146 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4147</div>
4148
4149<div class="doc_text">
4150
4151<h5>Syntax:</h5>
4152<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4153 declare void @llvm.stackrestore(i8 * %ptr)
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4154</pre>
4155
4156<h5>Overview:</h5>
4157
4158<p>
4159The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
4160the function stack to the state it was in when the corresponding <a
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4161href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4162useful for implementing language features like scoped automatic variable sized
4163arrays in C99.
4164</p>
4165
4166<h5>Semantics:</h5>
4167
4168<p>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4169See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4170</p>
4171
4172</div>
4173
4174
4175<!-- _______________________________________________________________________ -->
4176<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4177 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]4178</div>
4179
4180<div class="doc_text">
4181
4182<h5>Syntax:</h5>
4183<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]4184 declare void @llvm.prefetch(i8* &lt;address&gt;, i32 &lt;rw&gt;, i32 &lt;locality&gt;)
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]4185</pre>
4186
4187<h5>Overview:</h5>
4188
4189
4190<p>
4191The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]4192a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
4193no
4194effect on the behavior of the program but can change its performance
Chris Lattner2a615362005-02-28 19:47:14 +0000[diff] [blame]4195characteristics.
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]4196</p>
4197
4198<h5>Arguments:</h5>
4199
4200<p>
4201<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
4202determining if the fetch should be for a read (0) or write (1), and
4203<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Chris Lattneraeffb4a2005-03-07 20:31:38 +0000[diff] [blame]4204locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]4205<tt>locality</tt> arguments must be constant integers.
4206</p>
4207
4208<h5>Semantics:</h5>
4209
4210<p>
4211This intrinsic does not modify the behavior of the program. In particular,
4212prefetches cannot trap and do not produce a value. On targets that support this
4213intrinsic, the prefetch can provide hints to the processor cache for better
4214performance.
4215</p>
4216
4217</div>
4218
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]4219<!-- _______________________________________________________________________ -->
4220<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4221 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]4222</div>
4223
4224<div class="doc_text">
4225
4226<h5>Syntax:</h5>
4227<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]4228 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]4229</pre>
4230
4231<h5>Overview:</h5>
4232
4233
4234<p>
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]4235The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
4236(PC) in a region of
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]4237code to simulators and other tools. The method is target specific, but it is
4238expected that the marker will use exported symbols to transmit the PC of the marker.
Jeff Cohen25d4f7e2005-11-11 02:15:27 +0000[diff] [blame]4239The marker makes no guarantees that it will remain with any specific instruction
Chris Lattnerd07c3f42005-11-15 06:07:55 +0000[diff] [blame]4240after optimizations. It is possible that the presence of a marker will inhibit
Chris Lattnerb3e7afd2006-03-24 07:16:10 +0000[diff] [blame]4241optimizations. The intended use is to be inserted after optimizations to allow
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]4242correlations of simulation runs.
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]4243</p>
4244
4245<h5>Arguments:</h5>
4246
4247<p>
4248<tt>id</tt> is a numerical id identifying the marker.
4249</p>
4250
4251<h5>Semantics:</h5>
4252
4253<p>
4254This intrinsic does not modify the behavior of the program. Backends that do not
4255support this intrinisic may ignore it.
4256</p>
4257
4258</div>
4259
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]4260<!-- _______________________________________________________________________ -->
4261<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4262 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]4263</div>
4264
4265<div class="doc_text">
4266
4267<h5>Syntax:</h5>
4268<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4269 declare i64 @llvm.readcyclecounter( )
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]4270</pre>
4271
4272<h5>Overview:</h5>
4273
4274
4275<p>
4276The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
4277counter register (or similar low latency, high accuracy clocks) on those targets
4278that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
4279As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
4280should only be used for small timings.
4281</p>
4282
4283<h5>Semantics:</h5>
4284
4285<p>
4286When directly supported, reading the cycle counter should not modify any memory.
4287Implementations are allowed to either return a application specific value or a
4288system wide value. On backends without support, this is lowered to a constant 0.
4289</p>
4290
4291</div>
4292
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4293<!-- ======================================================================= -->
4294<div class="doc_subsection">
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4295 <a name="int_libc">Standard C Library Intrinsics</a>
4296</div>
4297
4298<div class="doc_text">
4299<p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4300LLVM provides intrinsics for a few important standard C library functions.
4301These intrinsics allow source-language front-ends to pass information about the
4302alignment of the pointer arguments to the code generator, providing opportunity
4303for more efficient code generation.
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4304</p>
4305
4306</div>
4307
4308<!-- _______________________________________________________________________ -->
4309<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4310 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4311</div>
4312
4313<div class="doc_text">
4314
4315<h5>Syntax:</h5>
4316<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4317 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4318 i32 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4319 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4320 i64 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4321</pre>
4322
4323<h5>Overview:</h5>
4324
4325<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4326The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4327location to the destination location.
4328</p>
4329
4330<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4331Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
4332intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4333</p>
4334
4335<h5>Arguments:</h5>
4336
4337<p>
4338The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4339the source. The third argument is an integer argument
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4340specifying the number of bytes to copy, and the fourth argument is the alignment
4341of the source and destination locations.
4342</p>
4343
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]4344<p>
4345If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +0000[diff] [blame]4346the caller guarantees that both the source and destination pointers are aligned
4347to that boundary.
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]4348</p>
4349
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4350<h5>Semantics:</h5>
4351
4352<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4353The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4354location to the destination location, which are not allowed to overlap. It
4355copies "len" bytes of memory over. If the argument is known to be aligned to
4356some boundary, this can be specified as the fourth argument, otherwise it should
4357be set to 0 or 1.
4358</p>
4359</div>
4360
4361
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4362<!-- _______________________________________________________________________ -->
4363<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4364 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4365</div>
4366
4367<div class="doc_text">
4368
4369<h5>Syntax:</h5>
4370<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4371 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4372 i32 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4373 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4374 i64 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4375</pre>
4376
4377<h5>Overview:</h5>
4378
4379<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4380The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
4381location to the destination location. It is similar to the
4382'<tt>llvm.memcmp</tt>' intrinsic but allows the two memory locations to overlap.
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4383</p>
4384
4385<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4386Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
4387intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4388</p>
4389
4390<h5>Arguments:</h5>
4391
4392<p>
4393The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4394the source. The third argument is an integer argument
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4395specifying the number of bytes to copy, and the fourth argument is the alignment
4396of the source and destination locations.
4397</p>
4398
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]4399<p>
4400If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +0000[diff] [blame]4401the caller guarantees that the source and destination pointers are aligned to
4402that boundary.
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]4403</p>
4404
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4405<h5>Semantics:</h5>
4406
4407<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4408The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4409location to the destination location, which may overlap. It
4410copies "len" bytes of memory over. If the argument is known to be aligned to
4411some boundary, this can be specified as the fourth argument, otherwise it should
4412be set to 0 or 1.
4413</p>
4414</div>
4415
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]4416
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4417<!-- _______________________________________________________________________ -->
4418<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4419 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4420</div>
4421
4422<div class="doc_text">
4423
4424<h5>Syntax:</h5>
4425<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4426 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4427 i32 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4428 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4429 i64 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4430</pre>
4431
4432<h5>Overview:</h5>
4433
4434<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4435The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4436byte value.
4437</p>
4438
4439<p>
4440Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
4441does not return a value, and takes an extra alignment argument.
4442</p>
4443
4444<h5>Arguments:</h5>
4445
4446<p>
4447The first argument is a pointer to the destination to fill, the second is the
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4448byte value to fill it with, the third argument is an integer
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4449argument specifying the number of bytes to fill, and the fourth argument is the
4450known alignment of destination location.
4451</p>
4452
4453<p>
4454If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattnerf0afc2c2006-03-04 00:02:10 +0000[diff] [blame]4455the caller guarantees that the destination pointer is aligned to that boundary.
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4456</p>
4457
4458<h5>Semantics:</h5>
4459
4460<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4461The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
4462the
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4463destination location. If the argument is known to be aligned to some boundary,
4464this can be specified as the fourth argument, otherwise it should be set to 0 or
44651.
4466</p>
4467</div>
4468
4469
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]4470<!-- _______________________________________________________________________ -->
4471<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4472 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]4473</div>
4474
4475<div class="doc_text">
4476
4477<h5>Syntax:</h5>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]4478<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
4479floating point type. Not all targets support all types however.
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]4480<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]4481 declare float @llvm.sqrt.f32(float %Val)
4482 declare double @llvm.sqrt.f64(double %Val)
4483 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
4484 declare fp128 @llvm.sqrt.f128(fp128 %Val)
4485 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]4486</pre>
4487
4488<h5>Overview:</h5>
4489
4490<p>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]4491The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]4492returning the same value as the libm '<tt>sqrt</tt>' function would. Unlike
4493<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
4494negative numbers (which allows for better optimization).
4495</p>
4496
4497<h5>Arguments:</h5>
4498
4499<p>
4500The argument and return value are floating point numbers of the same type.
4501</p>
4502
4503<h5>Semantics:</h5>
4504
4505<p>
Dan Gohmand6257fe2007-07-16 14:37:41 +0000[diff] [blame]4506This function returns the sqrt of the specified operand if it is a nonnegative
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]4507floating point number.
4508</p>
4509</div>
4510
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]4511<!-- _______________________________________________________________________ -->
4512<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4513 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]4514</div>
4515
4516<div class="doc_text">
4517
4518<h5>Syntax:</h5>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]4519<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
4520floating point type. Not all targets support all types however.
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]4521<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]4522 declare float @llvm.powi.f32(float %Val, i32 %power)
4523 declare double @llvm.powi.f64(double %Val, i32 %power)
4524 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
4525 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
4526 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]4527</pre>
4528
4529<h5>Overview:</h5>
4530
4531<p>
4532The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
4533specified (positive or negative) power. The order of evaluation of
4534multiplications is not defined.
4535</p>
4536
4537<h5>Arguments:</h5>
4538
4539<p>
4540The second argument is an integer power, and the first is a value to raise to
4541that power.
4542</p>
4543
4544<h5>Semantics:</h5>
4545
4546<p>
4547This function returns the first value raised to the second power with an
4548unspecified sequence of rounding operations.</p>
4549</div>
4550
4551
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4552<!-- ======================================================================= -->
4553<div class="doc_subsection">
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]4554 <a name="int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4555</div>
4556
4557<div class="doc_text">
4558<p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]4559LLVM provides intrinsics for a few important bit manipulation operations.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4560These allow efficient code generation for some algorithms.
4561</p>
4562
4563</div>
4564
4565<!-- _______________________________________________________________________ -->
4566<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4567 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]4568</div>
4569
4570<div class="doc_text">
4571
4572<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]4573<p>This is an overloaded intrinsic function. You can use bswap on any integer
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4574type that is an even number of bytes (i.e. BitWidth % 16 == 0).
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]4575<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4576 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
4577 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
4578 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]4579</pre>
4580
4581<h5>Overview:</h5>
4582
4583<p>
Reid Spencer338ea092007-04-02 02:25:19 +0000[diff] [blame]4584The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]4585values with an even number of bytes (positive multiple of 16 bits). These are
4586useful for performing operations on data that is not in the target's native
4587byte order.
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]4588</p>
4589
4590<h5>Semantics:</h5>
4591
4592<p>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4593The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4594and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
4595intrinsic returns an i32 value that has the four bytes of the input i32
4596swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the returned
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4597i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
4598<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]4599additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]4600</p>
4601
4602</div>
4603
4604<!-- _______________________________________________________________________ -->
4605<div class="doc_subsubsection">
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]4606 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4607</div>
4608
4609<div class="doc_text">
4610
4611<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]4612<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
4613width. Not all targets support all bit widths however.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4614<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4615 declare i8 @llvm.ctpop.i8 (i8 &lt;src&gt;)
4616 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4617 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4618 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
4619 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4620</pre>
4621
4622<h5>Overview:</h5>
4623
4624<p>
Chris Lattnerec6cb612006-01-16 22:38:59 +0000[diff] [blame]4625The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
4626value.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4627</p>
4628
4629<h5>Arguments:</h5>
4630
4631<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]4632The only argument is the value to be counted. The argument may be of any
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]4633integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4634</p>
4635
4636<h5>Semantics:</h5>
4637
4638<p>
4639The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
4640</p>
4641</div>
4642
4643<!-- _______________________________________________________________________ -->
4644<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]4645 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4646</div>
4647
4648<div class="doc_text">
4649
4650<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]4651<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
4652integer bit width. Not all targets support all bit widths however.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4653<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4654 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
4655 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4656 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4657 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
4658 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4659</pre>
4660
4661<h5>Overview:</h5>
4662
4663<p>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]4664The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
4665leading zeros in a variable.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4666</p>
4667
4668<h5>Arguments:</h5>
4669
4670<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]4671The only argument is the value to be counted. The argument may be of any
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]4672integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4673</p>
4674
4675<h5>Semantics:</h5>
4676
4677<p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]4678The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
4679in a variable. If the src == 0 then the result is the size in bits of the type
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4680of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4681</p>
4682</div>
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]4683
4684
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]4685
4686<!-- _______________________________________________________________________ -->
4687<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]4688 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]4689</div>
4690
4691<div class="doc_text">
4692
4693<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]4694<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
4695integer bit width. Not all targets support all bit widths however.
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]4696<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4697 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
4698 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4699 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4700 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
4701 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]4702</pre>
4703
4704<h5>Overview:</h5>
4705
4706<p>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]4707The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
4708trailing zeros.
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]4709</p>
4710
4711<h5>Arguments:</h5>
4712
4713<p>
4714The only argument is the value to be counted. The argument may be of any
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]4715integer type. The return type must match the argument type.
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]4716</p>
4717
4718<h5>Semantics:</h5>
4719
4720<p>
4721The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
4722in a variable. If the src == 0 then the result is the size in bits of the type
4723of src. For example, <tt>llvm.cttz(2) = 1</tt>.
4724</p>
4725</div>
4726
Reid Spencer497d93e2007-04-01 08:27:01 +0000[diff] [blame]4727<!-- _______________________________________________________________________ -->
4728<div class="doc_subsubsection">
Reid Spencerbeacf662007-04-10 02:51:31 +0000[diff] [blame]4729 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
Reid Spencera13ba7d2007-04-01 19:00:37 +0000[diff] [blame]4730</div>
4731
4732<div class="doc_text">
4733
4734<h5>Syntax:</h5>
Reid Spencerbeacf662007-04-10 02:51:31 +0000[diff] [blame]4735<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
Reid Spencera13ba7d2007-04-01 19:00:37 +0000[diff] [blame]4736on any integer bit width.
4737<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4738 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
4739 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Reid Spencera13ba7d2007-04-01 19:00:37 +0000[diff] [blame]4740</pre>
4741
4742<h5>Overview:</h5>
Reid Spencerbeacf662007-04-10 02:51:31 +0000[diff] [blame]4743<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
Reid Spencera13ba7d2007-04-01 19:00:37 +0000[diff] [blame]4744range of bits from an integer value and returns them in the same bit width as
4745the original value.</p>
4746
4747<h5>Arguments:</h5>
4748<p>The first argument, <tt>%val</tt> and the result may be integer types of
4749any bit width but they must have the same bit width. The second and third
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4750arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
Reid Spencera13ba7d2007-04-01 19:00:37 +0000[diff] [blame]4751
4752<h5>Semantics:</h5>
Reid Spencerbeacf662007-04-10 02:51:31 +0000[diff] [blame]4753<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4754of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
4755<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
4756operates in forward mode.</p>
4757<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
4758right by <tt>%loBit</tt> bits and then ANDing it with a mask with
Reid Spencera13ba7d2007-04-01 19:00:37 +0000[diff] [blame]4759only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
4760<ol>
4761 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
4762 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
4763 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
4764 to determine the number of bits to retain.</li>
4765 <li>A mask of the retained bits is created by shifting a -1 value.</li>
4766 <li>The mask is ANDed with <tt>%val</tt> to produce the result.
4767</ol>
Reid Spencerd6a85b52007-05-14 16:14:57 +0000[diff] [blame]4768<p>In reverse mode, a similar computation is made except that the bits are
4769returned in the reverse order. So, for example, if <tt>X</tt> has the value
4770<tt>i16 0x0ACF (101011001111)</tt> and we apply
4771<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
4772<tt>i16 0x0026 (000000100110)</tt>.</p>
Reid Spencera13ba7d2007-04-01 19:00:37 +0000[diff] [blame]4773</div>
4774
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]4775<div class="doc_subsubsection">
4776 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
4777</div>
4778
4779<div class="doc_text">
4780
4781<h5>Syntax:</h5>
4782<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
4783on any integer bit width.
4784<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4785 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
4786 declare i29 @llvm.part.set.i29.i9 (i29 %val, i9 %repl, i32 %lo, i32 %hi)
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]4787</pre>
4788
4789<h5>Overview:</h5>
4790<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
4791of bits in an integer value with another integer value. It returns the integer
4792with the replaced bits.</p>
4793
4794<h5>Arguments:</h5>
4795<p>The first argument, <tt>%val</tt> and the result may be integer types of
4796any bit width but they must have the same bit width. <tt>%val</tt> is the value
4797whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
4798integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
4799type since they specify only a bit index.</p>
4800
4801<h5>Semantics:</h5>
4802<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
4803of operation: forwards and reverse. If <tt>%lo</tt> is greater than
4804<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
4805operates in forward mode.</p>
4806<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
4807truncating it down to the size of the replacement area or zero extending it
4808up to that size.</p>
4809<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
4810are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
4811in <tt>%repl</tt> replaces the <tt>%lo</tt>th bit in <tt>%val</tt> and etc. up
4812to the <tt>%hi</tt>th bit.
Reid Spencerc6749c42007-05-14 16:50:20 +0000[diff] [blame]4813<p>In reverse mode, a similar computation is made except that the bits are
4814reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
4815<tt>%hi</tt> bit in <tt>%val</tt> and etc. down to the <tt>%lo</tt>th bit.
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]4816<h5>Examples:</h5>
4817<pre>
Reid Spencerf0dbf642007-04-12 01:03:03 +0000[diff] [blame]4818 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
Reid Spencerc6749c42007-05-14 16:50:20 +0000[diff] [blame]4819 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
4820 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
4821 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
Reid Spencerf0dbf642007-04-12 01:03:03 +0000[diff] [blame]4822 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
Reid Spencerc8910842007-04-11 23:49:50 +0000[diff] [blame]4823</pre>
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]4824</div>
4825
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]4826<!-- ======================================================================= -->
4827<div class="doc_subsection">
4828 <a name="int_debugger">Debugger Intrinsics</a>
4829</div>
4830
4831<div class="doc_text">
4832<p>
4833The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
4834are described in the <a
4835href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
4836Debugging</a> document.
4837</p>
4838</div>
4839
4840
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]4841<!-- ======================================================================= -->
4842<div class="doc_subsection">
4843 <a name="int_eh">Exception Handling Intrinsics</a>
4844</div>
4845
4846<div class="doc_text">
4847<p> The LLVM exception handling intrinsics (which all start with
4848<tt>llvm.eh.</tt> prefix), are described in the <a
4849href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
4850Handling</a> document. </p>
4851</div>
4852
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]4853<!-- ======================================================================= -->
4854<div class="doc_subsection">
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]4855 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]4856</div>
4857
4858<div class="doc_text">
4859<p>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]4860 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]4861 the <tt>nest</tt> attribute, from a function. The result is a callable
4862 function pointer lacking the nest parameter - the caller does not need
4863 to provide a value for it. Instead, the value to use is stored in
4864 advance in a "trampoline", a block of memory usually allocated
4865 on the stack, which also contains code to splice the nest value into the
4866 argument list. This is used to implement the GCC nested function address
4867 extension.
4868</p>
4869<p>
4870 For example, if the function is
4871 <tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
Bill Wendling03295ca2007-09-22 09:23:55 +0000[diff] [blame]4872 pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as follows:</p>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]4873<pre>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]4874 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
4875 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
4876 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
4877 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]4878</pre>
Bill Wendling03295ca2007-09-22 09:23:55 +0000[diff] [blame]4879 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
4880 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]4881</div>
4882
4883<!-- _______________________________________________________________________ -->
4884<div class="doc_subsubsection">
4885 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
4886</div>
4887<div class="doc_text">
4888<h5>Syntax:</h5>
4889<pre>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]4890declare i8* @llvm.init.trampoline(i8* &lt;tramp&gt;, i8* &lt;func&gt;, i8* &lt;nval&gt;)
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]4891</pre>
4892<h5>Overview:</h5>
4893<p>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]4894 This fills the memory pointed to by <tt>tramp</tt> with code
4895 and returns a function pointer suitable for executing it.
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]4896</p>
4897<h5>Arguments:</h5>
4898<p>
4899 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
4900 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
4901 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sandsc00c2ba2007-08-22 23:39:54 +0000[diff] [blame]4902 intrinsic. Note that the size and the alignment are target-specific - LLVM
4903 currently provides no portable way of determining them, so a front-end that
4904 generates this intrinsic needs to have some target-specific knowledge.
4905 The <tt>func</tt> argument must hold a function bitcast to an <tt>i8*</tt>.
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]4906</p>
4907<h5>Semantics:</h5>
4908<p>
4909 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]4910 dependent code, turning it into a function. A pointer to this function is
4911 returned, but needs to be bitcast to an
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]4912 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]4913 before being called. The new function's signature is the same as that of
4914 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
4915 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
4916 of pointer type. Calling the new function is equivalent to calling
4917 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
4918 missing <tt>nest</tt> argument. If, after calling
4919 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
4920 modified, then the effect of any later call to the returned function pointer is
4921 undefined.
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]4922</p>
4923</div>
4924
4925<!-- ======================================================================= -->
4926<div class="doc_subsection">
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]4927 <a name="int_general">General Intrinsics</a>
4928</div>
4929
4930<div class="doc_text">
4931<p> This class of intrinsics is designed to be generic and has
4932no specific purpose. </p>
4933</div>
4934
4935<!-- _______________________________________________________________________ -->
4936<div class="doc_subsubsection">
4937 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
4938</div>
4939
4940<div class="doc_text">
4941
4942<h5>Syntax:</h5>
4943<pre>
Tanya Lattnerd2e84422007-06-18 23:42:37 +0000[diff] [blame]4944 declare void @llvm.var.annotation(i8* &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]4945</pre>
4946
4947<h5>Overview:</h5>
4948
4949<p>
4950The '<tt>llvm.var.annotation</tt>' intrinsic
4951</p>
4952
4953<h5>Arguments:</h5>
4954
4955<p>
Tanya Lattnerd2e84422007-06-18 23:42:37 +0000[diff] [blame]4956The first argument is a pointer to a value, the second is a pointer to a
4957global string, the third is a pointer to a global string which is the source
4958file name, and the last argument is the line number.
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]4959</p>
4960
4961<h5>Semantics:</h5>
4962
4963<p>
4964This intrinsic allows annotation of local variables with arbitrary strings.
4965This can be useful for special purpose optimizations that want to look for these
4966 annotations. These have no other defined use, they are ignored by code
4967 generation and optimization.
4968</div>
4969
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]4970<!-- _______________________________________________________________________ -->
4971<div class="doc_subsubsection">
Tanya Lattnere1a8da02007-09-21 23:57:59 +0000[diff] [blame]4972 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]4973</div>
4974
4975<div class="doc_text">
4976
4977<h5>Syntax:</h5>
Tanya Lattner39cfba62007-09-21 23:56:27 +0000[diff] [blame]4978<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
4979any integer bit width.
4980</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]4981<pre>
Tanya Lattnerd3989a82007-09-22 00:03:01 +0000[diff] [blame]4982 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
4983 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
4984 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
4985 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
4986 declare i256 @llvm.annotation.i256(i256 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]4987</pre>
4988
4989<h5>Overview:</h5>
Tanya Lattner39cfba62007-09-21 23:56:27 +0000[diff] [blame]4990
4991<p>
4992The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]4993</p>
4994
4995<h5>Arguments:</h5>
4996
4997<p>
4998The first argument is an integer value (result of some expression),
4999the second is a pointer to a global string, the third is a pointer to a global
5000string which is the source file name, and the last argument is the line number.
Tanya Lattner39cfba62007-09-21 23:56:27 +0000[diff] [blame]5001It returns the value of the first argument.
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]5002</p>
5003
5004<h5>Semantics:</h5>
5005
5006<p>
5007This intrinsic allows annotations to be put on arbitrary expressions
5008with arbitrary strings. This can be useful for special purpose optimizations
5009that want to look for these annotations. These have no other defined use, they
5010are ignored by code generation and optimization.
5011</div>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]5012
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5013<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5014<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]5015<address>
5016 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
5017 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
5018 <a href="http://validator.w3.org/check/referer"><img
5019 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
5020
5021 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]5022 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]5023 Last modified: $Date$
5024</address>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5025</body>
5026</html>