Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / f04d584cae72b1172e83c47c69ea65cc55c190ac / . / docs / LangRef.html
blob: c6cbe295465a920565d82c43e0285ba8646a9adf [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>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]180 <li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
181 <li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
182 <li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]183 </ol>
184 </li>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]185 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]186 <ol>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]187 <li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]188 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
189 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
190 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]191 <li><a href="#int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic </a></li>
192 <li><a href="#int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic </a></li>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]193 </ol>
194 </li>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]195 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]196 <li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]197 <li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]198 <ol>
199 <li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]200 </ol>
201 </li>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]202 <li><a href="#int_general">General intrinsics</a>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]203 <ol>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]204 <li><a href="#int_var_annotation">
Tanya Lattner91d0b882007-09-22 00:01:26 +0000[diff] [blame]205 <tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Reid Spencer20677642007-07-20 19:59:11 +0000[diff] [blame]206 </ol>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]207 <ol>
208 <li><a href="#int_annotation">
Tanya Lattner91d0b882007-09-22 00:01:26 +0000[diff] [blame]209 <tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]210 </ol>
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]211 </li>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]212 </ol>
213 </li>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]214</ol>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]215
216<div class="doc_author">
217 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
218 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]219</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]220
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]221<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]222<div class="doc_section"> <a name="abstract">Abstract </a></div>
223<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]224
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]225<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]226<p>This document is a reference manual for the LLVM assembly language.
227LLVM is an SSA based representation that provides type safety,
228low-level operations, flexibility, and the capability of representing
229'all' high-level languages cleanly. It is the common code
230representation used throughout all phases of the LLVM compilation
231strategy.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]232</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]233
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]234<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]235<div class="doc_section"> <a name="introduction">Introduction</a> </div>
236<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]237
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]238<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]239
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]240<p>The LLVM code representation is designed to be used in three
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]241different forms: as an in-memory compiler IR, as an on-disk bitcode
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]242representation (suitable for fast loading by a Just-In-Time compiler),
243and as a human readable assembly language representation. This allows
244LLVM to provide a powerful intermediate representation for efficient
245compiler transformations and analysis, while providing a natural means
246to debug and visualize the transformations. The three different forms
247of LLVM are all equivalent. This document describes the human readable
248representation and notation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]249
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]250<p>The LLVM representation aims to be light-weight and low-level
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]251while being expressive, typed, and extensible at the same time. It
252aims to be a "universal IR" of sorts, by being at a low enough level
253that high-level ideas may be cleanly mapped to it (similar to how
254microprocessors are "universal IR's", allowing many source languages to
255be mapped to them). By providing type information, LLVM can be used as
256the target of optimizations: for example, through pointer analysis, it
257can be proven that a C automatic variable is never accessed outside of
258the current function... allowing it to be promoted to a simple SSA
259value instead of a memory location.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]260
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]261</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]262
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]263<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]264<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]265
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]266<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]267
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]268<p>It is important to note that this document describes 'well formed'
269LLVM assembly language. There is a difference between what the parser
270accepts and what is considered 'well formed'. For example, the
271following instruction is syntactically okay, but not well formed:</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]272
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]273<div class="doc_code">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]274<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]275%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]276</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]277</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]278
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]279<p>...because the definition of <tt>%x</tt> does not dominate all of
280its uses. The LLVM infrastructure provides a verification pass that may
281be used to verify that an LLVM module is well formed. This pass is
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]282automatically run by the parser after parsing input assembly and by
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]283the optimizer before it outputs bitcode. The violations pointed out
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]284by the verifier pass indicate bugs in transformation passes or input to
285the parser.</p>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]286</div>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]287
Chris Lattnercc689392007-10-03 17:34:29 +0000[diff] [blame]288<!-- Describe the typesetting conventions here. -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]289
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]290<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]291<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]292<!-- *********************************************************************** -->
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]293
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]294<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]295
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]296 <p>LLVM identifiers come in two basic types: global and local. Global
297 identifiers (functions, global variables) begin with the @ character. Local
298 identifiers (register names, types) begin with the % character. Additionally,
299 there are three different formats for identifiers, for different purposes:
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]300
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]301<ol>
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]302 <li>Named values are represented as a string of characters with their prefix.
303 For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
304 regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]305 Identifiers which require other characters in their names can be surrounded
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]306 with quotes. In this way, anything except a <tt>&quot;</tt> character can
307 be used in a named value.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]308
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]309 <li>Unnamed values are represented as an unsigned numeric value with their
310 prefix. For example, %12, @2, %44.</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]311
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]312 <li>Constants, which are described in a <a href="#constants">section about
313 constants</a>, below.</li>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]314</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]315
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]316<p>LLVM requires that values start with a prefix for two reasons: Compilers
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]317don't need to worry about name clashes with reserved words, and the set of
318reserved words may be expanded in the future without penalty. Additionally,
319unnamed identifiers allow a compiler to quickly come up with a temporary
320variable without having to avoid symbol table conflicts.</p>
321
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]322<p>Reserved words in LLVM are very similar to reserved words in other
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]323languages. There are keywords for different opcodes
324('<tt><a href="#i_add">add</a></tt>',
325 '<tt><a href="#i_bitcast">bitcast</a></tt>',
326 '<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]327href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]328and others. These reserved words cannot conflict with variable names, because
Reid Spencer2c452282007-08-07 14:34:28 +0000[diff] [blame]329none of them start with a prefix character ('%' or '@').</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]330
331<p>Here is an example of LLVM code to multiply the integer variable
332'<tt>%X</tt>' by 8:</p>
333
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]334<p>The easy way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]335
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]336<div class="doc_code">
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]337<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]338%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]339</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]340</div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]341
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]342<p>After strength reduction:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]343
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]344<div class="doc_code">
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]345<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]346%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]347</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]348</div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]349
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]350<p>And the hard way:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]351
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]352<div class="doc_code">
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]353<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]354<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
355<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
356%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]357</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]358</div>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]359
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]360<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
361important lexical features of LLVM:</p>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]362
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]363<ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]364
365 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
366 line.</li>
367
368 <li>Unnamed temporaries are created when the result of a computation is not
369 assigned to a named value.</li>
370
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]371 <li>Unnamed temporaries are numbered sequentially</li>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]372
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]373</ol>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]374
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]375<p>...and it also shows a convention that we follow in this document. When
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]376demonstrating instructions, we will follow an instruction with a comment that
377defines the type and name of value produced. Comments are shown in italic
378text.</p>
379
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]380</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]381
382<!-- *********************************************************************** -->
383<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
384<!-- *********************************************************************** -->
385
386<!-- ======================================================================= -->
387<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
388</div>
389
390<div class="doc_text">
391
392<p>LLVM programs are composed of "Module"s, each of which is a
393translation unit of the input programs. Each module consists of
394functions, global variables, and symbol table entries. Modules may be
395combined together with the LLVM linker, which merges function (and
396global variable) definitions, resolves forward declarations, and merges
397symbol table entries. Here is an example of the "hello world" module:</p>
398
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]399<div class="doc_code">
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]400<pre><i>; Declare the string constant as a global constant...</i>
Chris Lattnera89e5f12007-06-12 17:00:26 +0000[diff] [blame]401<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
402 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]403
404<i>; External declaration of the puts function</i>
Chris Lattnera89e5f12007-06-12 17:00:26 +0000[diff] [blame]405<a href="#functionstructure">declare</a> i32 @puts(i8 *) <i>; i32(i8 *)* </i>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]406
407<i>; Definition of main function</i>
Chris Lattnera89e5f12007-06-12 17:00:26 +0000[diff] [blame]408define i32 @main() { <i>; i32()* </i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]409 <i>; Convert [13x i8 ]* to i8 *...</i>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]410 %cast210 = <a
Chris Lattner6c0955b2007-06-12 17:01:15 +0000[diff] [blame]411 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]412
413 <i>; Call puts function to write out the string to stdout...</i>
414 <a
Chris Lattnera89e5f12007-06-12 17:00:26 +0000[diff] [blame]415 href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]416 <a
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]417 href="#i_ret">ret</a> i32 0<br>}<br>
418</pre>
419</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]420
421<p>This example is made up of a <a href="#globalvars">global variable</a>
422named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
423function, and a <a href="#functionstructure">function definition</a>
424for "<tt>main</tt>".</p>
425
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]426<p>In general, a module is made up of a list of global values,
427where both functions and global variables are global values. Global values are
428represented by a pointer to a memory location (in this case, a pointer to an
429array of char, and a pointer to a function), and have one of the following <a
430href="#linkage">linkage types</a>.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]431
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]432</div>
433
434<!-- ======================================================================= -->
435<div class="doc_subsection">
436 <a name="linkage">Linkage Types</a>
437</div>
438
439<div class="doc_text">
440
441<p>
442All Global Variables and Functions have one of the following types of linkage:
443</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]444
445<dl>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]446
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]447 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]448
449 <dd>Global values with internal linkage are only directly accessible by
450 objects in the current module. In particular, linking code into a module with
451 an internal global value may cause the internal to be renamed as necessary to
452 avoid collisions. Because the symbol is internal to the module, all
453 references can be updated. This corresponds to the notion of the
Chris Lattner4887bd82007-01-14 06:51:48 +0000[diff] [blame]454 '<tt>static</tt>' keyword in C.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]455 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]456
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]457 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]458
Chris Lattner4887bd82007-01-14 06:51:48 +0000[diff] [blame]459 <dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
460 the same name when linkage occurs. This is typically used to implement
461 inline functions, templates, or other code which must be generated in each
462 translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
463 allowed to be discarded.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]464 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]465
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]466 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]467
468 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
469 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
Chris Lattner4887bd82007-01-14 06:51:48 +0000[diff] [blame]470 used for globals that may be emitted in multiple translation units, but that
471 are not guaranteed to be emitted into every translation unit that uses them.
472 One example of this are common globals in C, such as "<tt>int X;</tt>" at
473 global scope.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]474 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]475
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]476 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]477
478 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
479 pointer to array type. When two global variables with appending linkage are
480 linked together, the two global arrays are appended together. This is the
481 LLVM, typesafe, equivalent of having the system linker append together
482 "sections" with identical names when .o files are linked.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]483 </dd>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]484
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]485 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
486 <dd>The semantics of this linkage follow the ELF model: the symbol is weak
487 until linked, if not linked, the symbol becomes null instead of being an
488 undefined reference.
489 </dd>
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]490
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]491 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnere5d947b2004-12-09 16:36:40 +0000[diff] [blame]492
493 <dd>If none of the above identifiers are used, the global is externally
494 visible, meaning that it participates in linkage and can be used to resolve
495 external symbol references.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]496 </dd>
Reid Spencerc8910842007-04-11 23:49:50 +0000[diff] [blame]497</dl>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]498
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]499 <p>
500 The next two types of linkage are targeted for Microsoft Windows platform
501 only. They are designed to support importing (exporting) symbols from (to)
502 DLLs.
503 </p>
504
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]505 <dl>
Anton Korobeynikovb74ed072006-09-14 18:23:27 +0000[diff] [blame]506 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
507
508 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
509 or variable via a global pointer to a pointer that is set up by the DLL
510 exporting the symbol. On Microsoft Windows targets, the pointer name is
511 formed by combining <code>_imp__</code> and the function or variable name.
512 </dd>
513
514 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
515
516 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
517 pointer to a pointer in a DLL, so that it can be referenced with the
518 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
519 name is formed by combining <code>_imp__</code> and the function or variable
520 name.
521 </dd>
522
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]523</dl>
524
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]525<p><a name="linkage_external"></a>For example, since the "<tt>.LC0</tt>"
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]526variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
527variable and was linked with this one, one of the two would be renamed,
528preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
529external (i.e., lacking any linkage declarations), they are accessible
Reid Spencerac8d2762007-01-05 00:59:10 +0000[diff] [blame]530outside of the current module.</p>
531<p>It is illegal for a function <i>declaration</i>
532to have any linkage type other than "externally visible", <tt>dllimport</tt>,
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]533or <tt>extern_weak</tt>.</p>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]534<p>Aliases can have only <tt>external</tt>, <tt>internal</tt> and <tt>weak</tt>
535linkages.
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]536</div>
537
538<!-- ======================================================================= -->
539<div class="doc_subsection">
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]540 <a name="callingconv">Calling Conventions</a>
541</div>
542
543<div class="doc_text">
544
545<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
546and <a href="#i_invoke">invokes</a> can all have an optional calling convention
547specified for the call. The calling convention of any pair of dynamic
548caller/callee must match, or the behavior of the program is undefined. The
549following calling conventions are supported by LLVM, and more may be added in
550the future:</p>
551
552<dl>
553 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
554
555 <dd>This calling convention (the default if no other calling convention is
556 specified) matches the target C calling conventions. This calling convention
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]557 supports varargs function calls and tolerates some mismatch in the declared
Reid Spencerc28d2bc2006-12-31 21:30:18 +0000[diff] [blame]558 prototype and implemented declaration of the function (as does normal C).
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]559 </dd>
560
561 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
562
563 <dd>This calling convention attempts to make calls as fast as possible
564 (e.g. by passing things in registers). This calling convention allows the
565 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner8cdc5bc2005-05-06 23:08:23 +0000[diff] [blame]566 without having to conform to an externally specified ABI. Implementations of
567 this convention should allow arbitrary tail call optimization to be supported.
568 This calling convention does not support varargs and requires the prototype of
569 all callees to exactly match the prototype of the function definition.
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]570 </dd>
571
572 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
573
574 <dd>This calling convention attempts to make code in the caller as efficient
575 as possible under the assumption that the call is not commonly executed. As
576 such, these calls often preserve all registers so that the call does not break
577 any live ranges in the caller side. This calling convention does not support
578 varargs and requires the prototype of all callees to exactly match the
579 prototype of the function definition.
580 </dd>
581
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]582 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]583
584 <dd>Any calling convention may be specified by number, allowing
585 target-specific calling conventions to be used. Target specific calling
586 conventions start at 64.
587 </dd>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]588</dl>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]589
590<p>More calling conventions can be added/defined on an as-needed basis, to
591support pascal conventions or any other well-known target-independent
592convention.</p>
593
594</div>
595
596<!-- ======================================================================= -->
597<div class="doc_subsection">
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]598 <a name="visibility">Visibility Styles</a>
599</div>
600
601<div class="doc_text">
602
603<p>
604All Global Variables and Functions have one of the following visibility styles:
605</p>
606
607<dl>
608 <dt><b>"<tt>default</tt>" - Default style</b>:</dt>
609
610 <dd>On ELF, default visibility means that the declaration is visible to other
611 modules and, in shared libraries, means that the declared entity may be
612 overridden. On Darwin, default visibility means that the declaration is
613 visible to other modules. Default visibility corresponds to "external
614 linkage" in the language.
615 </dd>
616
617 <dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
618
619 <dd>Two declarations of an object with hidden visibility refer to the same
620 object if they are in the same shared object. Usually, hidden visibility
621 indicates that the symbol will not be placed into the dynamic symbol table,
622 so no other module (executable or shared library) can reference it
623 directly.
624 </dd>
625
Anton Korobeynikov6f9896f2007-04-29 18:35:00 +0000[diff] [blame]626 <dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
627
628 <dd>On ELF, protected visibility indicates that the symbol will be placed in
629 the dynamic symbol table, but that references within the defining module will
630 bind to the local symbol. That is, the symbol cannot be overridden by another
631 module.
632 </dd>
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]633</dl>
634
635</div>
636
637<!-- ======================================================================= -->
638<div class="doc_subsection">
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]639 <a name="globalvars">Global Variables</a>
640</div>
641
642<div class="doc_text">
643
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]644<p>Global variables define regions of memory allocated at compilation time
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]645instead of run-time. Global variables may optionally be initialized, may have
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]646an explicit section to be placed in, and may have an optional explicit alignment
647specified. A variable may be defined as "thread_local", which means that it
648will not be shared by threads (each thread will have a separated copy of the
649variable). A variable may be defined as a global "constant," which indicates
650that the contents of the variable will <b>never</b> be modified (enabling better
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]651optimization, allowing the global data to be placed in the read-only section of
652an executable, etc). Note that variables that need runtime initialization
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]653cannot be marked "constant" as there is a store to the variable.</p>
Chris Lattner3689a342005-02-12 19:30:21 +0000[diff] [blame]654
655<p>
656LLVM explicitly allows <em>declarations</em> of global variables to be marked
657constant, even if the final definition of the global is not. This capability
658can be used to enable slightly better optimization of the program, but requires
659the language definition to guarantee that optimizations based on the
660'constantness' are valid for the translation units that do not include the
661definition.
662</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]663
664<p>As SSA values, global variables define pointer values that are in
665scope (i.e. they dominate) all basic blocks in the program. Global
666variables always define a pointer to their "content" type because they
667describe a region of memory, and all memory objects in LLVM are
668accessed through pointers.</p>
669
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]670<p>LLVM allows an explicit section to be specified for globals. If the target
671supports it, it will emit globals to the section specified.</p>
672
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]673<p>An explicit alignment may be specified for a global. If not present, or if
674the alignment is set to zero, the alignment of the global is set by the target
675to whatever it feels convenient. If an explicit alignment is specified, the
676global is forced to have at least that much alignment. All alignments must be
677a power of 2.</p>
678
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]679<p>For example, the following defines a global with an initializer, section,
680 and alignment:</p>
681
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]682<div class="doc_code">
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]683<pre>
Chris Lattner3e63a9d2007-07-13 20:01:46 +0000[diff] [blame]684@G = constant float 1.0, section "foo", align 4
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]685</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]686</div>
Chris Lattner68027ea2007-01-14 00:27:09 +0000[diff] [blame]687
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]688</div>
689
690
691<!-- ======================================================================= -->
692<div class="doc_subsection">
693 <a name="functionstructure">Functions</a>
694</div>
695
696<div class="doc_text">
697
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]698<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
699an optional <a href="#linkage">linkage type</a>, an optional
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]700<a href="#visibility">visibility style</a>, an optional
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]701<a href="#callingconv">calling convention</a>, a return type, an optional
702<a href="#paramattrs">parameter attribute</a> for the return type, a function
703name, a (possibly empty) argument list (each with optional
Anton Korobeynikov8cea37b2007-01-23 12:35:46 +0000[diff] [blame]704<a href="#paramattrs">parameter attributes</a>), an optional section, an
705optional alignment, an opening curly brace, a list of basic blocks, and a
706closing curly brace.
707
708LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
709optional <a href="#linkage">linkage type</a>, an optional
710<a href="#visibility">visibility style</a>, an optional
711<a href="#callingconv">calling convention</a>, a return type, an optional
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]712<a href="#paramattrs">parameter attribute</a> for the return type, a function
713name, a possibly empty list of arguments, and an optional alignment.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]714
715<p>A function definition contains a list of basic blocks, forming the CFG for
716the function. Each basic block may optionally start with a label (giving the
717basic block a symbol table entry), contains a list of instructions, and ends
718with a <a href="#terminators">terminator</a> instruction (such as a branch or
719function return).</p>
720
Chris Lattner4a3c9012007-06-08 16:52:14 +0000[diff] [blame]721<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]722executed on entrance to the function, and it is not allowed to have predecessor
723basic blocks (i.e. there can not be any branches to the entry block of a
724function). Because the block can have no predecessors, it also cannot have any
725<a href="#i_phi">PHI nodes</a>.</p>
726
Chris Lattner88f6c462005-11-12 00:45:07 +0000[diff] [blame]727<p>LLVM allows an explicit section to be specified for functions. If the target
728supports it, it will emit functions to the section specified.</p>
729
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]730<p>An explicit alignment may be specified for a function. If not present, or if
731the alignment is set to zero, the alignment of the function is set by the target
732to whatever it feels convenient. If an explicit alignment is specified, the
733function is forced to have at least that much alignment. All alignments must be
734a power of 2.</p>
735
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]736</div>
737
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]738
739<!-- ======================================================================= -->
740<div class="doc_subsection">
741 <a name="aliasstructure">Aliases</a>
742</div>
743<div class="doc_text">
744 <p>Aliases act as "second name" for the aliasee value (which can be either
Anton Korobeynikova80e1182007-04-28 13:45:00 +0000[diff] [blame]745 function or global variable or bitcast of global value). Aliases may have an
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]746 optional <a href="#linkage">linkage type</a>, and an
747 optional <a href="#visibility">visibility style</a>.</p>
748
749 <h5>Syntax:</h5>
750
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]751<div class="doc_code">
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]752<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]753@&lt;Name&gt; = [Linkage] [Visibility] alias &lt;AliaseeTy&gt; @&lt;Aliasee&gt;
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]754</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]755</div>
Anton Korobeynikov8b0a8c82007-04-25 14:27:10 +0000[diff] [blame]756
757</div>
758
759
760
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]761<!-- ======================================================================= -->
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]762<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
763<div class="doc_text">
764 <p>The return type and each parameter of a function type may have a set of
765 <i>parameter attributes</i> associated with them. Parameter attributes are
766 used to communicate additional information about the result or parameters of
767 a function. Parameter attributes are considered to be part of the function
768 type so two functions types that differ only by the parameter attributes
769 are different function types.</p>
770
Reid Spencer950e9f82007-01-15 18:27:39 +0000[diff] [blame]771 <p>Parameter attributes are simple keywords that follow the type specified. If
772 multiple parameter attributes are needed, they are space separated. For
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]773 example:</p>
774
775<div class="doc_code">
776<pre>
Reid Spencer9445e9a2007-07-19 23:13:04 +0000[diff] [blame]777%someFunc = i16 (i8 signext %someParam) zeroext
778%someFunc = i16 (i8 zeroext %someParam) zeroext
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]779</pre>
780</div>
781
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]782 <p>Note that the two function types above are unique because the parameter has
Reid Spencer9445e9a2007-07-19 23:13:04 +0000[diff] [blame]783 a different attribute (<tt>signext</tt> in the first one, <tt>zeroext</tt> in
784 the second). Also note that the attribute for the function result
785 (<tt>zeroext</tt>) comes immediately after the argument list.</p>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]786
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]787 <p>Currently, only the following parameter attributes are defined:</p>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]788 <dl>
Reid Spencer9445e9a2007-07-19 23:13:04 +0000[diff] [blame]789 <dt><tt>zeroext</tt></dt>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]790 <dd>This indicates that the parameter should be zero extended just before
791 a call to this function.</dd>
Reid Spencer9445e9a2007-07-19 23:13:04 +0000[diff] [blame]792 <dt><tt>signext</tt></dt>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]793 <dd>This indicates that the parameter should be sign extended just before
794 a call to this function.</dd>
Anton Korobeynikov34d6dea2007-01-28 14:30:45 +0000[diff] [blame]795 <dt><tt>inreg</tt></dt>
796 <dd>This indicates that the parameter should be placed in register (if
Anton Korobeynikov66a8c8c2007-01-28 15:27:21 +0000[diff] [blame]797 possible) during assembling function call. Support for this attribute is
798 target-specific</dd>
Anton Korobeynikov34d6dea2007-01-28 14:30:45 +0000[diff] [blame]799 <dt><tt>sret</tt></dt>
Anton Korobeynikov66a8c8c2007-01-28 15:27:21 +0000[diff] [blame]800 <dd>This indicates that the parameter specifies the address of a structure
Reid Spencer67606122007-03-22 02:02:11 +0000[diff] [blame]801 that is the return value of the function in the source program.</dd>
Zhou Shengfebca342007-06-05 05:28:26 +0000[diff] [blame]802 <dt><tt>noalias</tt></dt>
803 <dd>This indicates that the parameter not alias any other object or any
804 other "noalias" objects during the function call.
Reid Spencer2dc52012007-03-22 02:18:56 +0000[diff] [blame]805 <dt><tt>noreturn</tt></dt>
806 <dd>This function attribute indicates that the function never returns. This
807 indicates to LLVM that every call to this function should be treated as if
808 an <tt>unreachable</tt> instruction immediately followed the call.</dd>
Reid Spencer67606122007-03-22 02:02:11 +0000[diff] [blame]809 <dt><tt>nounwind</tt></dt>
810 <dd>This function attribute indicates that the function type does not use
811 the unwind instruction and does not allow stack unwinding to propagate
812 through it.</dd>
Duncan Sands50f19f52007-07-27 19:57:41 +0000[diff] [blame]813 <dt><tt>nest</tt></dt>
814 <dd>This indicates that the parameter can be excised using the
815 <a href="#int_trampoline">trampoline intrinsics</a>.</dd>
Anton Korobeynikove925df72007-11-14 10:30:13 +0000[diff] [blame]816 <dt><tt>pure</tt></dt>
Duncan Sandsf04d5842007-11-14 21:14:02 +0000[diff] [blame^]817 <dd>This function attribute indicates that the function has no side-effects
818 except for producing a return value. The value returned must only depend on
819 the function arguments and/or global variables. It may use values obtained
820 by dereferencing pointers.</dd>
Anton Korobeynikove925df72007-11-14 10:30:13 +0000[diff] [blame]821 <dt><tt>const</tt></dt>
Duncan Sandsf04d5842007-11-14 21:14:02 +0000[diff] [blame^]822 <dd>A <tt>const</tt> function has the same restrictions as a <tt>pure</tt>
823 function, but in addition it is not allowed to dereference any pointer arguments
824 or global variables.
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]825 </dl>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]826
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]827</div>
828
829<!-- ======================================================================= -->
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]830<div class="doc_subsection">
Chris Lattner1eeeb0c2006-04-08 04:40:53 +0000[diff] [blame]831 <a name="moduleasm">Module-Level Inline Assembly</a>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]832</div>
833
834<div class="doc_text">
835<p>
836Modules may contain "module-level inline asm" blocks, which corresponds to the
837GCC "file scope inline asm" blocks. These blocks are internally concatenated by
838LLVM and treated as a single unit, but may be separated in the .ll file if
839desired. The syntax is very simple:
840</p>
841
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]842<div class="doc_code">
843<pre>
844module asm "inline asm code goes here"
845module asm "more can go here"
846</pre>
847</div>
Chris Lattner4e9aba72006-01-23 23:23:47 +0000[diff] [blame]848
849<p>The strings can contain any character by escaping non-printable characters.
850 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
851 for the number.
852</p>
853
854<p>
855 The inline asm code is simply printed to the machine code .s file when
856 assembly code is generated.
857</p>
858</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]859
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]860<!-- ======================================================================= -->
861<div class="doc_subsection">
862 <a name="datalayout">Data Layout</a>
863</div>
864
865<div class="doc_text">
866<p>A module may specify a target specific data layout string that specifies how
Reid Spencerc8910842007-04-11 23:49:50 +0000[diff] [blame]867data is to be laid out in memory. The syntax for the data layout is simply:</p>
868<pre> target datalayout = "<i>layout specification</i>"</pre>
869<p>The <i>layout specification</i> consists of a list of specifications
870separated by the minus sign character ('-'). Each specification starts with a
871letter and may include other information after the letter to define some
872aspect of the data layout. The specifications accepted are as follows: </p>
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]873<dl>
874 <dt><tt>E</tt></dt>
875 <dd>Specifies that the target lays out data in big-endian form. That is, the
876 bits with the most significance have the lowest address location.</dd>
877 <dt><tt>e</tt></dt>
878 <dd>Specifies that hte target lays out data in little-endian form. That is,
879 the bits with the least significance have the lowest address location.</dd>
880 <dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
881 <dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
882 <i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
883 alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
884 too.</dd>
885 <dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
886 <dd>This specifies the alignment for an integer type of a given bit
887 <i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
888 <dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
889 <dd>This specifies the alignment for a vector type of a given bit
890 <i>size</i>.</dd>
891 <dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
892 <dd>This specifies the alignment for a floating point type of a given bit
893 <i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
894 (double).</dd>
895 <dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
896 <dd>This specifies the alignment for an aggregate type of a given bit
897 <i>size</i>.</dd>
898</dl>
899<p>When constructing the data layout for a given target, LLVM starts with a
900default set of specifications which are then (possibly) overriden by the
901specifications in the <tt>datalayout</tt> keyword. The default specifications
902are given in this list:</p>
903<ul>
904 <li><tt>E</tt> - big endian</li>
905 <li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
906 <li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
907 <li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
908 <li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
909 <li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
910 <li><tt>i64:32:64</tt> - i64 has abi alignment of 32-bits but preferred
911 alignment of 64-bits</li>
912 <li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
913 <li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
914 <li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
915 <li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
916 <li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
917</ul>
918<p>When llvm is determining the alignment for a given type, it uses the
919following rules:
920<ol>
921 <li>If the type sought is an exact match for one of the specifications, that
922 specification is used.</li>
923 <li>If no match is found, and the type sought is an integer type, then the
924 smallest integer type that is larger than the bitwidth of the sought type is
925 used. If none of the specifications are larger than the bitwidth then the the
926 largest integer type is used. For example, given the default specifications
927 above, the i7 type will use the alignment of i8 (next largest) while both
928 i65 and i256 will use the alignment of i64 (largest specified).</li>
929 <li>If no match is found, and the type sought is a vector type, then the
930 largest vector type that is smaller than the sought vector type will be used
931 as a fall back. This happens because <128 x double> can be implemented in
932 terms of 64 <2 x double>, for example.</li>
933</ol>
934</div>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]935
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]936<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]937<div class="doc_section"> <a name="typesystem">Type System</a> </div>
938<!-- *********************************************************************** -->
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]939
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]940<div class="doc_text">
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]941
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]942<p>The LLVM type system is one of the most important features of the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]943intermediate representation. Being typed enables a number of
944optimizations to be performed on the IR directly, without having to do
945extra analyses on the side before the transformation. A strong type
946system makes it easier to read the generated code and enables novel
947analyses and transformations that are not feasible to perform on normal
948three address code representations.</p>
Chris Lattnerfa730212004-12-09 16:11:40 +0000[diff] [blame]949
950</div>
951
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]952<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]953<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]954<div class="doc_text">
John Criswell4457dc92004-04-09 16:48:45 +0000[diff] [blame]955<p>The primitive types are the fundamental building blocks of the LLVM
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000[diff] [blame]956system. The current set of primitive types is as follows:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]957
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]958<table class="layout">
959 <tr class="layout">
960 <td class="left">
961 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]962 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]963 <tr><th>Type</th><th>Description</th></tr>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]964 <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]965 <tr><td><tt>label</tt></td><td>Branch destination</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]966 </tbody>
967 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]968 </td>
969 <td class="right">
970 <table>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]971 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]972 <tr><th>Type</th><th>Description</th></tr>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]973 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]974 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]975 </tbody>
976 </table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]977 </td>
978 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]979</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]980</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]981
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]982<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]983<div class="doc_subsubsection"> <a name="t_classifications">Type
984Classifications</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]985<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]986<p>These different primitive types fall into a few useful
987classifications:</p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]988
989<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]990 <tbody>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]991 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]992 <tr>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]993 <td><a name="t_integer">integer</a></td>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]994 <td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]995 </tr>
996 <tr>
997 <td><a name="t_floating">floating point</a></td>
998 <td><tt>float, double</tt></td>
999 </tr>
1000 <tr>
1001 <td><a name="t_firstclass">first class</a></td>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1002 <td><tt>i1, ..., float, double, <br/>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1003 <a href="#t_pointer">pointer</a>,<a href="#t_vector">vector</a></tt>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1004 </td>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1005 </tr>
1006 </tbody>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1007</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1008
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1009<p>The <a href="#t_firstclass">first class</a> types are perhaps the
1010most important. Values of these types are the only ones which can be
1011produced by instructions, passed as arguments, or used as operands to
1012instructions. This means that all structures and arrays must be
1013manipulated either by pointer or by component.</p>
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<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1017<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1018
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1019<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1020
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1021<p>The real power in LLVM comes from the derived types in the system.
1022This is what allows a programmer to represent arrays, functions,
1023pointers, and other useful types. Note that these derived types may be
1024recursive: For example, it is possible to have a two dimensional array.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1025
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1026</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1027
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1028<!-- _______________________________________________________________________ -->
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1029<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
1030
1031<div class="doc_text">
1032
1033<h5>Overview:</h5>
1034<p>The integer type is a very simple derived type that simply specifies an
1035arbitrary bit width for the integer type desired. Any bit width from 1 bit to
10362^23-1 (about 8 million) can be specified.</p>
1037
1038<h5>Syntax:</h5>
1039
1040<pre>
1041 iN
1042</pre>
1043
1044<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
1045value.</p>
1046
1047<h5>Examples:</h5>
1048<table class="layout">
1049 <tr class="layout">
1050 <td class="left">
1051 <tt>i1</tt><br/>
1052 <tt>i4</tt><br/>
1053 <tt>i8</tt><br/>
1054 <tt>i16</tt><br/>
1055 <tt>i32</tt><br/>
1056 <tt>i42</tt><br/>
1057 <tt>i64</tt><br/>
1058 <tt>i1942652</tt><br/>
1059 </td>
1060 <td class="left">
1061 A boolean integer of 1 bit<br/>
1062 A nibble sized integer of 4 bits.<br/>
1063 A byte sized integer of 8 bits.<br/>
1064 A half word sized integer of 16 bits.<br/>
1065 A word sized integer of 32 bits.<br/>
1066 An integer whose bit width is the answer. <br/>
1067 A double word sized integer of 64 bits.<br/>
1068 A really big integer of over 1 million bits.<br/>
1069 </td>
1070 </tr>
1071</table>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1072</div>
Reid Spencer2b916312007-05-16 18:44:01 +0000[diff] [blame]1073
1074<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1075<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1076
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1077<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1078
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1079<h5>Overview:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1080
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1081<p>The array type is a very simple derived type that arranges elements
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1082sequentially in memory. The array type requires a size (number of
1083elements) and an underlying data type.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1084
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1085<h5>Syntax:</h5>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1086
1087<pre>
1088 [&lt;# elements&gt; x &lt;elementtype&gt;]
1089</pre>
1090
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]1091<p>The number of elements is a constant integer value; elementtype may
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1092be any type with a size.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1093
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1094<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1095<table class="layout">
1096 <tr class="layout">
1097 <td class="left">
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1098 <tt>[40 x i32 ]</tt><br/>
1099 <tt>[41 x i32 ]</tt><br/>
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1100 <tt>[40 x i8]</tt><br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1101 </td>
1102 <td class="left">
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1103 Array of 40 32-bit integer values.<br/>
1104 Array of 41 32-bit integer values.<br/>
1105 Array of 40 8-bit integer values.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1106 </td>
1107 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1108</table>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1109<p>Here are some examples of multidimensional arrays:</p>
1110<table class="layout">
1111 <tr class="layout">
1112 <td class="left">
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1113 <tt>[3 x [4 x i32]]</tt><br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1114 <tt>[12 x [10 x float]]</tt><br/>
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1115 <tt>[2 x [3 x [4 x i16]]]</tt><br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1116 </td>
1117 <td class="left">
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1118 3x4 array of 32-bit integer values.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1119 12x10 array of single precision floating point values.<br/>
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1120 2x3x4 array of 16-bit integer values.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1121 </td>
1122 </tr>
1123</table>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1124
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]1125<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
1126length array. Normally, accesses past the end of an array are undefined in
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1127LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
1128As a special case, however, zero length arrays are recognized to be variable
1129length. This allows implementation of 'pascal style arrays' with the LLVM
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1130type "{ i32, [0 x float]}", for example.</p>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]1131
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1132</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1133
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1134<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1135<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1136<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1137<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1138<p>The function type can be thought of as a function signature. It
1139consists of a return type and a list of formal parameter types.
John Criswell009900b2003-11-25 21:45:46 +0000[diff] [blame]1140Function types are usually used to build virtual function tables
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1141(which are structures of pointers to functions), for indirect function
1142calls, and when defining a function.</p>
John Criswell009900b2003-11-25 21:45:46 +0000[diff] [blame]1143<p>
1144The return type of a function type cannot be an aggregate type.
1145</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1146<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1147<pre> &lt;returntype&gt; (&lt;parameter list&gt;)<br></pre>
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]1148<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]1149specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1150which indicates that the function takes a variable number of arguments.
1151Variable argument functions can access their arguments with the <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1152 href="#int_varargs">variable argument handling intrinsic</a> functions.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1153<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1154<table class="layout">
1155 <tr class="layout">
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1156 <td class="left"><tt>i32 (i32)</tt></td>
1157 <td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1158 </td>
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1159 </tr><tr class="layout">
Reid Spencer9445e9a2007-07-19 23:13:04 +0000[diff] [blame]1160 <td class="left"><tt>float&nbsp;(i16&nbsp;signext,&nbsp;i32&nbsp;*)&nbsp;*
Reid Spencerf17a0b72006-12-31 07:20:23 +0000[diff] [blame]1161 </tt></td>
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1162 <td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
1163 an <tt>i16</tt> that should be sign extended and a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1164 <a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1165 <tt>float</tt>.
1166 </td>
1167 </tr><tr class="layout">
1168 <td class="left"><tt>i32 (i8*, ...)</tt></td>
1169 <td class="left">A vararg function that takes at least one
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1170 <a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
Reid Spencer92f82302006-12-31 07:18:34 +0000[diff] [blame]1171 which returns an integer. This is the signature for <tt>printf</tt> in
1172 LLVM.
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1173 </td>
1174 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1175</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]1176
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1177</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1178<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1179<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1180<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1181<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1182<p>The structure type is used to represent a collection of data members
1183together in memory. The packing of the field types is defined to match
1184the ABI of the underlying processor. The elements of a structure may
1185be any type that has a size.</p>
1186<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1187and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1188field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1189instruction.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1190<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1191<pre> { &lt;type list&gt; }<br></pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1192<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1193<table class="layout">
1194 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]1195 <td class="left"><tt>{ i32, i32, i32 }</tt></td>
1196 <td class="left">A triple of three <tt>i32</tt> values</td>
1197 </tr><tr class="layout">
1198 <td class="left"><tt>{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}</tt></td>
1199 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1200 second element is a <a href="#t_pointer">pointer</a> to a
1201 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1202 an <tt>i32</tt>.</td>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1203 </tr>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1204</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1205</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1206
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1207<!-- _______________________________________________________________________ -->
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]1208<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
1209</div>
1210<div class="doc_text">
1211<h5>Overview:</h5>
1212<p>The packed structure type is used to represent a collection of data members
1213together in memory. There is no padding between fields. Further, the alignment
1214of a packed structure is 1 byte. The elements of a packed structure may
1215be any type that has a size.</p>
1216<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
1217and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
1218field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
1219instruction.</p>
1220<h5>Syntax:</h5>
1221<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
1222<h5>Examples:</h5>
1223<table class="layout">
1224 <tr class="layout">
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]1225 <td class="left"><tt>&lt; { i32, i32, i32 } &gt;</tt></td>
1226 <td class="left">A triple of three <tt>i32</tt> values</td>
1227 </tr><tr class="layout">
1228 <td class="left"><tt>&lt;&nbsp;{&nbsp;float,&nbsp;i32&nbsp;(i32)&nbsp;*&nbsp;}&nbsp;&gt;</tt></td>
1229 <td class="left">A pair, where the first element is a <tt>float</tt> and the
1230 second element is a <a href="#t_pointer">pointer</a> to a
1231 <a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
1232 an <tt>i32</tt>.</td>
Andrew Lenharth75e10682006-12-08 17:13:00 +0000[diff] [blame]1233 </tr>
1234</table>
1235</div>
1236
1237<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1238<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1239<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1240<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1241<p>As in many languages, the pointer type represents a pointer or
1242reference to another object, which must live in memory.</p>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1243<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1244<pre> &lt;type&gt; *<br></pre>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1245<h5>Examples:</h5>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1246<table class="layout">
1247 <tr class="layout">
1248 <td class="left">
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1249 <tt>[4x i32]*</tt><br/>
1250 <tt>i32 (i32 *) *</tt><br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1251 </td>
1252 <td class="left">
1253 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]1254 four <tt>i32</tt> values<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1255 A <a href="#t_pointer">pointer</a> to a <a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1256 href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
1257 <tt>i32</tt>.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1258 </td>
1259 </tr>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1260</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1261</div>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1262
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1263<!-- _______________________________________________________________________ -->
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1264<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1265<div class="doc_text">
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1266
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1267<h5>Overview:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1268
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1269<p>A vector type is a simple derived type that represents a vector
1270of elements. Vector types are used when multiple primitive data
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1271are operated in parallel using a single instruction (SIMD).
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1272A vector type requires a size (number of
Chris Lattnerb8d172f2005-11-10 01:44:22 +0000[diff] [blame]1273elements) and an underlying primitive data type. Vectors must have a power
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1274of two length (1, 2, 4, 8, 16 ...). Vector types are
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1275considered <a href="#t_firstclass">first class</a>.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1276
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1277<h5>Syntax:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1278
1279<pre>
1280 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1281</pre>
1282
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1283<p>The number of elements is a constant integer value; elementtype may
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]1284be any integer or floating point type.</p>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1285
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1286<h5>Examples:</h5>
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1287
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1288<table class="layout">
1289 <tr class="layout">
1290 <td class="left">
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1291 <tt>&lt;4 x i32&gt;</tt><br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1292 <tt>&lt;8 x float&gt;</tt><br/>
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1293 <tt>&lt;2 x i64&gt;</tt><br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1294 </td>
1295 <td class="left">
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1296 Vector of 4 32-bit integer values.<br/>
1297 Vector of 8 floating-point values.<br/>
1298 Vector of 2 64-bit integer values.<br/>
Reid Spencerd3f876c2004-11-01 08:19:36 +0000[diff] [blame]1299 </td>
1300 </tr>
1301</table>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1302</div>
1303
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1304<!-- _______________________________________________________________________ -->
1305<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1306<div class="doc_text">
1307
1308<h5>Overview:</h5>
1309
1310<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksen8ac04ff2007-10-14 00:34:53 +0000[diff] [blame]1311corresponds (for example) to the C notion of a forward declared structure type.
Chris Lattner69c11bb2005-04-25 17:34:15 +0000[diff] [blame]1312In LLVM, opaque types can eventually be resolved to any type (not just a
1313structure type).</p>
1314
1315<h5>Syntax:</h5>
1316
1317<pre>
1318 opaque
1319</pre>
1320
1321<h5>Examples:</h5>
1322
1323<table class="layout">
1324 <tr class="layout">
1325 <td class="left">
1326 <tt>opaque</tt>
1327 </td>
1328 <td class="left">
1329 An opaque type.<br/>
1330 </td>
1331 </tr>
1332</table>
1333</div>
1334
1335
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1336<!-- *********************************************************************** -->
1337<div class="doc_section"> <a name="constants">Constants</a> </div>
1338<!-- *********************************************************************** -->
1339
1340<div class="doc_text">
1341
1342<p>LLVM has several different basic types of constants. This section describes
1343them all and their syntax.</p>
1344
1345</div>
1346
1347<!-- ======================================================================= -->
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]1348<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1349
1350<div class="doc_text">
1351
1352<dl>
1353 <dt><b>Boolean constants</b></dt>
1354
1355 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]1356 constants of the <tt><a href="#t_primitive">i1</a></tt> type.
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1357 </dd>
1358
1359 <dt><b>Integer constants</b></dt>
1360
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]1361 <dd>Standard integers (such as '4') are constants of the <a
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]1362 href="#t_integer">integer</a> type. Negative numbers may be used with
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1363 integer types.
1364 </dd>
1365
1366 <dt><b>Floating point constants</b></dt>
1367
1368 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1369 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1370 notation (see below). Floating point constants must have a <a
1371 href="#t_floating">floating point</a> type. </dd>
1372
1373 <dt><b>Null pointer constants</b></dt>
1374
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1375 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1376 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1377
1378</dl>
1379
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1380<p>The one non-intuitive notation for constants is the optional hexadecimal form
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1381of floating point constants. For example, the form '<tt>double
13820x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
13834.5e+15</tt>'. The only time hexadecimal floating point constants are required
Reid Spencercc16dc32004-12-09 18:02:53 +0000[diff] [blame]1384(and the only time that they are generated by the disassembler) is when a
1385floating point constant must be emitted but it cannot be represented as a
1386decimal floating point number. For example, NaN's, infinities, and other
1387special values are represented in their IEEE hexadecimal format so that
1388assembly and disassembly do not cause any bits to change in the constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1389
1390</div>
1391
1392<!-- ======================================================================= -->
1393<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1394</div>
1395
1396<div class="doc_text">
Chris Lattnerd4f6b172005-03-07 22:13:59 +0000[diff] [blame]1397<p>Aggregate constants arise from aggregation of simple constants
1398and smaller aggregate constants.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1399
1400<dl>
1401 <dt><b>Structure constants</b></dt>
1402
1403 <dd>Structure constants are represented with notation similar to structure
1404 type definitions (a comma separated list of elements, surrounded by braces
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1405 (<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* %G }</tt>",
Chris Lattner3e63a9d2007-07-13 20:01:46 +0000[diff] [blame]1406 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]1407 must have <a href="#t_struct">structure type</a>, and the number and
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1408 types of elements must match those specified by the type.
1409 </dd>
1410
1411 <dt><b>Array constants</b></dt>
1412
1413 <dd>Array constants are represented with notation similar to array type
1414 definitions (a comma separated list of elements, surrounded by square brackets
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1415 (<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1416 constants must have <a href="#t_array">array type</a>, and the number and
1417 types of elements must match those specified by the type.
1418 </dd>
1419
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1420 <dt><b>Vector constants</b></dt>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1421
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1422 <dd>Vector constants are represented with notation similar to vector type
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1423 definitions (a comma separated list of elements, surrounded by
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1424 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]1425 i32 11, i32 74, i32 100 &gt;</tt>". Vector constants must have <a
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1426 href="#t_vector">vector type</a>, and the number and types of elements must
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1427 match those specified by the type.
1428 </dd>
1429
1430 <dt><b>Zero initialization</b></dt>
1431
1432 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1433 value to zero of <em>any</em> type, including scalar and aggregate types.
1434 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]1435 large arrays) and is always exactly equivalent to using explicit zero
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1436 initializers.
1437 </dd>
1438</dl>
1439
1440</div>
1441
1442<!-- ======================================================================= -->
1443<div class="doc_subsection">
1444 <a name="globalconstants">Global Variable and Function Addresses</a>
1445</div>
1446
1447<div class="doc_text">
1448
1449<p>The addresses of <a href="#globalvars">global variables</a> and <a
1450href="#functionstructure">functions</a> are always implicitly valid (link-time)
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1451constants. These constants are explicitly referenced when the <a
1452href="#identifiers">identifier for the global</a> is used and always have <a
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1453href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1454file:</p>
1455
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1456<div class="doc_code">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1457<pre>
Chris Lattnera18a4242007-06-06 18:28:13 +0000[diff] [blame]1458@X = global i32 17
1459@Y = global i32 42
1460@Z = global [2 x i32*] [ i32* @X, i32* @Y ]
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1461</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1462</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1463
1464</div>
1465
1466<!-- ======================================================================= -->
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1467<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1468<div class="doc_text">
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1469 <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]1470 no specific value. Undefined values may be of any type and be used anywhere
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1471 a constant is permitted.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1472
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1473 <p>Undefined values indicate to the compiler that the program is well defined
1474 no matter what value is used, giving the compiler more freedom to optimize.
1475 </p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1476</div>
1477
1478<!-- ======================================================================= -->
1479<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1480</div>
1481
1482<div class="doc_text">
1483
1484<p>Constant expressions are used to allow expressions involving other constants
1485to be used as constants. Constant expressions may be of any <a
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1486href="#t_firstclass">first class</a> type and may involve any LLVM operation
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1487that does not have side effects (e.g. load and call are not supported). The
1488following is the syntax for constant expressions:</p>
1489
1490<dl>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1491 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1492 <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]1493 than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1494
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1495 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1496 <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]1497 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]1498
1499 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1500 <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]1501 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]1502
1503 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1504 <dd>Truncate a floating point constant to another floating point type. The
1505 size of CST must be larger than the size of TYPE. Both types must be
1506 floating point.</dd>
1507
1508 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1509 <dd>Floating point extend a constant to another type. The size of CST must be
1510 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1511
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]1512 <dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1513 <dd>Convert a floating point constant to the corresponding unsigned integer
1514 constant. TYPE must be an integer type. CST must be floating point. If the
1515 value won't fit in the integer type, the results are undefined.</dd>
1516
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]1517 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1518 <dd>Convert a floating point constant to the corresponding signed integer
1519 constant. TYPE must be an integer type. CST must be floating point. If the
1520 value won't fit in the integer type, the results are undefined.</dd>
1521
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]1522 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1523 <dd>Convert an unsigned integer constant to the corresponding floating point
1524 constant. TYPE must be floating point. CST must be of integer type. If the
Jeff Cohencb757312007-04-22 14:56:37 +0000[diff] [blame]1525 value won't fit in the floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1526
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]1527 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1528 <dd>Convert a signed integer constant to the corresponding floating point
1529 constant. TYPE must be floating point. CST must be of integer type. If the
Jeff Cohencb757312007-04-22 14:56:37 +0000[diff] [blame]1530 value won't fit in the floating point type, the results are undefined.</dd>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1531
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]1532 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1533 <dd>Convert a pointer typed constant to the corresponding integer constant
1534 TYPE must be an integer type. CST must be of pointer type. The CST value is
1535 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1536
1537 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1538 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1539 pointer type. CST must be of integer type. The CST value is zero extended,
1540 truncated, or unchanged to make it fit in a pointer size. This one is
1541 <i>really</i> dangerous!</dd>
1542
1543 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1544 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1545 identical (same number of bits). The conversion is done as if the CST value
1546 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]1547 with this operator, just the type. This can be used for conversion of
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1548 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]1549 pointers it is only valid to cast to another pointer type.
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]1550 </dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1551
1552 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1553
1554 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1555 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1556 instruction, the index list may have zero or more indexes, which are required
1557 to make sense for the type of "CSTPTR".</dd>
1558
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]1559 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1560
1561 <dd>Perform the <a href="#i_select">select operation</a> on
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]1562 constants.</dd>
1563
1564 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1565 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1566
1567 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1568 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino9fbe1452006-01-10 19:31:34 +0000[diff] [blame]1569
1570 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1571
1572 <dd>Perform the <a href="#i_extractelement">extractelement
1573 operation</a> on constants.
1574
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]1575 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1576
1577 <dd>Perform the <a href="#i_insertelement">insertelement
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]1578 operation</a> on constants.</dd>
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]1579
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]1580
1581 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1582
1583 <dd>Perform the <a href="#i_shufflevector">shufflevector
Reid Spencer01c42592006-12-04 19:23:19 +0000[diff] [blame]1584 operation</a> on constants.</dd>
Chris Lattnerc1989542006-04-08 00:13:41 +0000[diff] [blame]1585
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1586 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1587
Reid Spencer2dc45b82004-12-09 18:13:12 +0000[diff] [blame]1588 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1589 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1590 binary</a> operations. The constraints on operands are the same as those for
1591 the corresponding instruction (e.g. no bitwise operations on floating point
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]1592 values are allowed).</dd>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1593</dl>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1594</div>
Chris Lattner9ee5d222004-03-08 16:49:10 +0000[diff] [blame]1595
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1596<!-- *********************************************************************** -->
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1597<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1598<!-- *********************************************************************** -->
1599
1600<!-- ======================================================================= -->
1601<div class="doc_subsection">
1602<a name="inlineasm">Inline Assembler Expressions</a>
1603</div>
1604
1605<div class="doc_text">
1606
1607<p>
1608LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1609Module-Level Inline Assembly</a>) through the use of a special value. This
1610value represents the inline assembler as a string (containing the instructions
1611to emit), a list of operand constraints (stored as a string), and a flag that
1612indicates whether or not the inline asm expression has side effects. An example
1613inline assembler expression is:
1614</p>
1615
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1616<div class="doc_code">
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1617<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1618i32 (i32) asm "bswap $0", "=r,r"
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1619</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1620</div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1621
1622<p>
1623Inline assembler expressions may <b>only</b> be used as the callee operand of
1624a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
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]1629%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]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>
1634Inline asms with side effects not visible in the constraint list must be marked
1635as having side effects. This is done through the use of the
1636'<tt>sideeffect</tt>' keyword, like so:
1637</p>
1638
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1639<div class="doc_code">
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1640<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1641call void asm sideeffect "eieio", ""()
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1642</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]1643</div>
Chris Lattnere87d6532006-01-25 23:47:57 +0000[diff] [blame]1644
1645<p>TODO: The format of the asm and constraints string still need to be
1646documented here. Constraints on what can be done (e.g. duplication, moving, etc
1647need to be documented).
1648</p>
1649
1650</div>
1651
1652<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1653<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1654<!-- *********************************************************************** -->
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1655
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1656<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1657
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1658<p>The LLVM instruction set consists of several different
1659classifications of instructions: <a href="#terminators">terminator
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1660instructions</a>, <a href="#binaryops">binary instructions</a>,
1661<a href="#bitwiseops">bitwise binary instructions</a>, <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1662 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1663instructions</a>.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1664
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1665</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1666
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1667<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1668<div class="doc_subsection"> <a name="terminators">Terminator
1669Instructions</a> </div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1670
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1671<div class="doc_text">
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1672
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1673<p>As mentioned <a href="#functionstructure">previously</a>, every
1674basic block in a program ends with a "Terminator" instruction, which
1675indicates which block should be executed after the current block is
1676finished. These terminator instructions typically yield a '<tt>void</tt>'
1677value: they produce control flow, not values (the one exception being
1678the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1679<p>There are six different terminator instructions: the '<a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1680 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1681instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1682the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1683 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1684 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1685
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1686</div>
Chris Lattnerc3f59762004-12-09 17:30:23 +0000[diff] [blame]1687
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1688<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1689<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1690Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1691<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1692<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1693<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]1694 ret void <i>; Return from void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1695</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1696<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1697<p>The '<tt>ret</tt>' instruction is used to return control flow (and a
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]1698value) from a function back to the caller.</p>
John Criswell4457dc92004-04-09 16:48:45 +0000[diff] [blame]1699<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1700returns a value and then causes control flow, and one that just causes
1701control flow to occur.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1702<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1703<p>The '<tt>ret</tt>' instruction may return any '<a
1704 href="#t_firstclass">first class</a>' type. Notice that a function is
1705not <a href="#wellformed">well formed</a> if there exists a '<tt>ret</tt>'
1706instruction inside of the function that returns a value that does not
1707match the return type of the function.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1708<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1709<p>When the '<tt>ret</tt>' instruction is executed, control flow
1710returns back to the calling function's context. If the caller is a "<a
John Criswellfa081872004-06-25 15:16:57 +0000[diff] [blame]1711 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1712the instruction after the call. If the caller was an "<a
1713 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]1714at the beginning of the "normal" destination block. If the instruction
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1715returns a value, that value shall set the call or invoke instruction's
1716return value.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1717<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1718<pre> ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]1719 ret void <i>; Return from a void function</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1720</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1721</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1722<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1723<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1724<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1725<h5>Syntax:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]1726<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]1727</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1728<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1729<p>The '<tt>br</tt>' instruction is used to cause control flow to
1730transfer to a different basic block in the current function. There are
1731two forms of this instruction, corresponding to a conditional branch
1732and an unconditional branch.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1733<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1734<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]1735single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
Reid Spencerde151942007-02-19 23:54:10 +0000[diff] [blame]1736unconditional form of the '<tt>br</tt>' instruction takes a single
1737'<tt>label</tt>' value as a target.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1738<h5>Semantics:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]1739<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1740argument is evaluated. If the value is <tt>true</tt>, control flows
1741to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1742control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1743<h5>Example:</h5>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]1744<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]1745 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]1746</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1747<!-- _______________________________________________________________________ -->
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1748<div class="doc_subsubsection">
1749 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1750</div>
1751
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1752<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1753<h5>Syntax:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1754
1755<pre>
1756 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
1757</pre>
1758
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1759<h5>Overview:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1760
1761<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1762several different places. It is a generalization of the '<tt>br</tt>'
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1763instruction, allowing a branch to occur to one of many possible
1764destinations.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1765
1766
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1767<h5>Arguments:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1768
1769<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1770comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1771an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1772table is not allowed to contain duplicate constant entries.</p>
1773
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1774<h5>Semantics:</h5>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1775
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1776<p>The <tt>switch</tt> instruction specifies a table of values and
1777destinations. When the '<tt>switch</tt>' instruction is executed, this
John Criswell84114752004-06-25 16:05:06 +0000[diff] [blame]1778table is searched for the given value. If the value is found, control flow is
1779transfered to the corresponding destination; otherwise, control flow is
1780transfered to the default destination.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1781
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1782<h5>Implementation:</h5>
1783
1784<p>Depending on properties of the target machine and the particular
1785<tt>switch</tt> instruction, this instruction may be code generated in different
John Criswell84114752004-06-25 16:05:06 +0000[diff] [blame]1786ways. For example, it could be generated as a series of chained conditional
1787branches or with a lookup table.</p>
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1788
1789<h5>Example:</h5>
1790
1791<pre>
1792 <i>; Emulate a conditional br instruction</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]1793 %Val = <a href="#i_zext">zext</a> i1 %value to i32
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1794 switch i32 %Val, label %truedest [i32 0, label %falsedest ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1795
1796 <i>; Emulate an unconditional br instruction</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1797 switch i32 0, label %dest [ ]
Chris Lattnerc88c17b2004-02-24 04:54:45 +0000[diff] [blame]1798
1799 <i>; Implement a jump table:</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1800 switch i32 %val, label %otherwise [ i32 0, label %onzero
1801 i32 1, label %onone
1802 i32 2, label %ontwo ]
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1803</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1804</div>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1805
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1806<!-- _______________________________________________________________________ -->
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1807<div class="doc_subsubsection">
1808 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1809</div>
1810
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1811<div class="doc_text">
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1812
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1813<h5>Syntax:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1814
1815<pre>
1816 &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]1817 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1818</pre>
1819
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]1820<h5>Overview:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1821
1822<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1823function, with the possibility of control flow transfer to either the
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]1824'<tt>normal</tt>' label or the
1825'<tt>exception</tt>' label. If the callee function returns with the
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1826"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1827"normal" label. If the callee (or any indirect callees) returns with the "<a
John Criswelle4c57cc2005-05-12 16:52:32 +0000[diff] [blame]1828href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
1829continued at the dynamically nearest "exception" label.</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1830
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1831<h5>Arguments:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1832
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1833<p>This instruction requires several arguments:</p>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1834
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1835<ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1836 <li>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]1837 The optional "cconv" marker indicates which <a href="#callingconv">calling
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1838 convention</a> the call should use. If none is specified, the call defaults
1839 to using C calling conventions.
1840 </li>
1841 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1842 function value being invoked. In most cases, this is a direct function
1843 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1844 an arbitrary pointer to function value.
1845 </li>
1846
1847 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1848 function to be invoked. </li>
1849
1850 <li>'<tt>function args</tt>': argument list whose types match the function
1851 signature argument types. If the function signature indicates the function
1852 accepts a variable number of arguments, the extra arguments can be
1853 specified. </li>
1854
1855 <li>'<tt>normal label</tt>': the label reached when the called function
1856 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1857
1858 <li>'<tt>exception label</tt>': the label reached when a callee returns with
1859 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
1860
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1861</ol>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1862
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1863<h5>Semantics:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1864
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1865<p>This instruction is designed to operate as a standard '<tt><a
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1866href="#i_call">call</a></tt>' instruction in most regards. The primary
1867difference is that it establishes an association with a label, which is used by
1868the runtime library to unwind the stack.</p>
1869
1870<p>This instruction is used in languages with destructors to ensure that proper
1871cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
1872exception. Additionally, this is important for implementation of
1873'<tt>catch</tt>' clauses in high-level languages that support them.</p>
1874
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1875<h5>Example:</h5>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]1876<pre>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]1877 %retval = invoke i32 %Test(i32 15) to label %Continue
1878 unwind label %TestCleanup <i>; {i32}:retval set</i>
1879 %retval = invoke <a href="#callingconv">coldcc</a> i32 %Test(i32 15) to label %Continue
1880 unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1881</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1882</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1883
1884
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1885<!-- _______________________________________________________________________ -->
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1886
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1887<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
1888Instruction</a> </div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1889
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1890<div class="doc_text">
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1891
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1892<h5>Syntax:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1893<pre>
1894 unwind
1895</pre>
1896
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1897<h5>Overview:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1898
1899<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
1900at the first callee in the dynamic call stack which used an <a
1901href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
1902primarily used to implement exception handling.</p>
1903
Chris Lattner27f71f22003-09-03 00:41:47 +0000[diff] [blame]1904<h5>Semantics:</h5>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1905
1906<p>The '<tt>unwind</tt>' intrinsic causes execution of the current function to
1907immediately halt. The dynamic call stack is then searched for the first <a
1908href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
1909execution continues at the "exceptional" destination block specified by the
1910<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
1911dynamic call chain, undefined behavior results.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1912</div>
Chris Lattner35eca582004-10-16 18:04:13 +0000[diff] [blame]1913
1914<!-- _______________________________________________________________________ -->
1915
1916<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
1917Instruction</a> </div>
1918
1919<div class="doc_text">
1920
1921<h5>Syntax:</h5>
1922<pre>
1923 unreachable
1924</pre>
1925
1926<h5>Overview:</h5>
1927
1928<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
1929instruction is used to inform the optimizer that a particular portion of the
1930code is not reachable. This can be used to indicate that the code after a
1931no-return function cannot be reached, and other facts.</p>
1932
1933<h5>Semantics:</h5>
1934
1935<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
1936</div>
1937
1938
1939
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1940<!-- ======================================================================= -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1941<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1942<div class="doc_text">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1943<p>Binary operators are used to do most of the computation in a
1944program. They require two operands, execute an operation on them, and
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]1945produce a single value. The operands might represent
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1946multiple 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]1947The result value of a binary operator is not
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1948necessarily the same type as its operands.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1949<p>There are several different binary operators:</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1950</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1951<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1952<div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
1953Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1954<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1955<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1956<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]1957</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1958<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1959<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1960<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1961<p>The two arguments to the '<tt>add</tt>' instruction must be either <a
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1962 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1963 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]1964Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1965<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1966<p>The value produced is the integer or floating point sum of the two
1967operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1968<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1969<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]1970</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1971</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1972<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1973<div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
1974Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1975<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1976<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1977<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]1978</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1979<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1980<p>The '<tt>sub</tt>' instruction returns the difference of its two
1981operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1982<p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
1983instruction present in most other intermediate representations.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1984<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1985<p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1986 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1987values.
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]1988This instruction can also take <a href="#t_vector">vector</a> versions of the values.
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]1989Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1990<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]1991<p>The value produced is the integer or floating point difference of
1992the two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1993<h5>Example:</h5>
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]1994<pre>
1995 &lt;result&gt; = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]1996 &lt;result&gt; = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1997</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]1998</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]1999<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2000<div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
2001Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2002<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2003<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2004<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]2005</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2006<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2007<p>The '<tt>mul</tt>' instruction returns the product of its two
2008operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2009<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2010<p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2011 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2012values.
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2013This instruction can also take <a href="#t_vector">vector</a> versions of the values.
Chris Lattnera58561b2004-08-12 19:12:28 +0000[diff] [blame]2014Both arguments must have identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2015<h5>Semantics:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2016<p>The value produced is the integer or floating point product of the
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2017two operands.</p>
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]2018<p>Because the operands are the same width, the result of an integer
2019multiplication is the same whether the operands should be deemed unsigned or
2020signed.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2021<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2022<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]2023</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2024</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2025<!-- _______________________________________________________________________ -->
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2026<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
2027</a></div>
2028<div class="doc_text">
2029<h5>Syntax:</h5>
2030<pre> &lt;result&gt; = udiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2031</pre>
2032<h5>Overview:</h5>
2033<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
2034operands.</p>
2035<h5>Arguments:</h5>
2036<p>The two arguments to the '<tt>udiv</tt>' instruction must be
2037<a href="#t_integer">integer</a> values. Both arguments must have identical
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2038types. This instruction can also take <a href="#t_vector">vector</a> versions
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2039of the values in which case the elements must be integers.</p>
2040<h5>Semantics:</h5>
2041<p>The value produced is the unsigned integer quotient of the two operands. This
2042instruction always performs an unsigned division operation, regardless of
2043whether the arguments are unsigned or not.</p>
2044<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2045<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]2046</pre>
2047</div>
2048<!-- _______________________________________________________________________ -->
2049<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
2050</a> </div>
2051<div class="doc_text">
2052<h5>Syntax:</h5>
2053<pre> &lt;result&gt; = sdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2054</pre>
2055<h5>Overview:</h5>
2056<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
2057operands.</p>
2058<h5>Arguments:</h5>
2059<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
2060<a href="#t_integer">integer</a> values. Both arguments must have identical
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2061types. This instruction can also take <a href="#t_vector">vector</a> versions
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2062of the values in which case the elements must be integers.</p>
2063<h5>Semantics:</h5>
2064<p>The value produced is the signed integer quotient of the two operands. This
2065instruction always performs a signed division operation, regardless of whether
2066the arguments are signed or not.</p>
2067<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2068<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]2069</pre>
2070</div>
2071<!-- _______________________________________________________________________ -->
2072<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2073Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2074<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2075<h5>Syntax:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2076<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]2077</pre>
2078<h5>Overview:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2079<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2080operands.</p>
2081<h5>Arguments:</h5>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2082<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2083<a href="#t_floating">floating point</a> values. Both arguments must have
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2084identical types. This instruction can also take <a href="#t_vector">vector</a>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2085versions of floating point values.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2086<h5>Semantics:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2087<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2088<h5>Example:</h5>
Reid Spencer1628cec2006-10-26 06:15:43 +0000[diff] [blame]2089<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]2090</pre>
2091</div>
2092<!-- _______________________________________________________________________ -->
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]2093<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
2094</div>
2095<div class="doc_text">
2096<h5>Syntax:</h5>
2097<pre> &lt;result&gt; = urem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2098</pre>
2099<h5>Overview:</h5>
2100<p>The '<tt>urem</tt>' instruction returns the remainder from the
2101unsigned division of its two arguments.</p>
2102<h5>Arguments:</h5>
2103<p>The two arguments to the '<tt>urem</tt>' instruction must be
2104<a href="#t_integer">integer</a> values. Both arguments must have identical
Dan Gohman80176312007-11-05 23:35:22 +0000[diff] [blame]2105types. This instruction can also take <a href="#t_vector">vector</a> versions
2106of the values in which case the elements must be integers.</p>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]2107<h5>Semantics:</h5>
2108<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
2109This instruction always performs an unsigned division to get the remainder,
2110regardless of whether the arguments are unsigned or not.</p>
2111<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2112<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]2113</pre>
2114
2115</div>
2116<!-- _______________________________________________________________________ -->
2117<div class="doc_subsubsection"> <a name="i_srem">'<tt>srem</tt>'
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2118Instruction</a> </div>
2119<div class="doc_text">
2120<h5>Syntax:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]2121<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]2122</pre>
2123<h5>Overview:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]2124<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman80176312007-11-05 23:35:22 +0000[diff] [blame]2125signed division of its two operands. This instruction can also take
2126<a href="#t_vector">vector</a> versions of the values in which case
2127the elements must be integers.</p>
2128</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2129<h5>Arguments:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]2130<p>The two arguments to the '<tt>srem</tt>' instruction must be
2131<a href="#t_integer">integer</a> values. Both arguments must have identical
2132types.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2133<h5>Semantics:</h5>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]2134<p>This instruction returns the <i>remainder</i> of a division (where the result
Reid Spencerc9fdfc82007-03-24 22:23:39 +0000[diff] [blame]2135has the same sign as the dividend, <tt>var1</tt>), not the <i>modulo</i>
2136operator (where the result has the same sign as the divisor, <tt>var2</tt>) of
2137a value. For more information about the difference, see <a
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2138 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
Reid Spencerc9fdfc82007-03-24 22:23:39 +0000[diff] [blame]2139Math Forum</a>. For a table of how this is implemented in various languages,
Reid Spencer64f5c6c2007-03-24 22:40:44 +0000[diff] [blame]2140please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
Reid Spencerc9fdfc82007-03-24 22:23:39 +0000[diff] [blame]2141Wikipedia: modulo operation</a>.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2142<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2143<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]2144</pre>
2145
2146</div>
2147<!-- _______________________________________________________________________ -->
2148<div class="doc_subsubsection"> <a name="i_frem">'<tt>frem</tt>'
2149Instruction</a> </div>
2150<div class="doc_text">
2151<h5>Syntax:</h5>
2152<pre> &lt;result&gt; = frem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2153</pre>
2154<h5>Overview:</h5>
2155<p>The '<tt>frem</tt>' instruction returns the remainder from the
2156division of its two operands.</p>
2157<h5>Arguments:</h5>
2158<p>The two arguments to the '<tt>frem</tt>' instruction must be
2159<a href="#t_floating">floating point</a> values. Both arguments must have
Dan Gohman80176312007-11-05 23:35:22 +0000[diff] [blame]2160identical types. This instruction can also take <a href="#t_vector">vector</a>
2161versions of floating point values.</p>
Reid Spencer0a783f72006-11-02 01:53:59 +0000[diff] [blame]2162<h5>Semantics:</h5>
2163<p>This instruction returns the <i>remainder</i> of a division.</p>
2164<h5>Example:</h5>
2165<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]2166</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2167</div>
Robert Bocchino7b81c752006-02-17 21:18:08 +0000[diff] [blame]2168
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]2169<!-- ======================================================================= -->
2170<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
2171Operations</a> </div>
2172<div class="doc_text">
2173<p>Bitwise binary operators are used to do various forms of
2174bit-twiddling in a program. They are generally very efficient
2175instructions and can commonly be strength reduced from other
2176instructions. They require two operands, execute an operation on them,
2177and produce a single value. The resulting value of the bitwise binary
2178operators is always the same type as its first operand.</p>
2179</div>
2180
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2181<!-- _______________________________________________________________________ -->
2182<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2183Instruction</a> </div>
2184<div class="doc_text">
2185<h5>Syntax:</h5>
2186<pre> &lt;result&gt; = shl &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2187</pre>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]2188
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2189<h5>Overview:</h5>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]2190
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2191<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2192the left a specified number of bits.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]2193
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2194<h5>Arguments:</h5>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]2195
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2196<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
2197 href="#t_integer">integer</a> type.</p>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]2198
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2199<h5>Semantics:</h5>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]2200
2201<p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup>. If
2202<tt>var2</tt> is (statically or dynamically) equal to or larger than the number
2203of bits in <tt>var1</tt>, the result is undefined.</p>
2204
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2205<h5>Example:</h5><pre>
2206 &lt;result&gt; = shl i32 4, %var <i>; yields {i32}: 4 &lt;&lt; %var</i>
2207 &lt;result&gt; = shl i32 4, 2 <i>; yields {i32}: 16</i>
2208 &lt;result&gt; = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]2209 &lt;result&gt; = shl i32 1, 32 <i>; undefined</i>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2210</pre>
2211</div>
2212<!-- _______________________________________________________________________ -->
2213<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
2214Instruction</a> </div>
2215<div class="doc_text">
2216<h5>Syntax:</h5>
2217<pre> &lt;result&gt; = lshr &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2218</pre>
2219
2220<h5>Overview:</h5>
2221<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2222operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2223
2224<h5>Arguments:</h5>
2225<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
2226<a href="#t_integer">integer</a> type.</p>
2227
2228<h5>Semantics:</h5>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]2229
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2230<p>This instruction always performs a logical shift right operation. The most
2231significant bits of the result will be filled with zero bits after the
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]2232shift. If <tt>var2</tt> is (statically or dynamically) equal to or larger than
2233the number of bits in <tt>var1</tt>, the result is undefined.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2234
2235<h5>Example:</h5>
2236<pre>
2237 &lt;result&gt; = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
2238 &lt;result&gt; = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
2239 &lt;result&gt; = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
2240 &lt;result&gt; = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]2241 &lt;result&gt; = lshr i32 1, 32 <i>; undefined</i>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2242</pre>
2243</div>
2244
Reid Spencer8e11bf82007-02-02 13:57:07 +0000[diff] [blame]2245<!-- _______________________________________________________________________ -->
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2246<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2247Instruction</a> </div>
2248<div class="doc_text">
2249
2250<h5>Syntax:</h5>
2251<pre> &lt;result&gt; = ashr &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
2252</pre>
2253
2254<h5>Overview:</h5>
2255<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2256operand shifted to the right a specified number of bits with sign extension.</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2257
2258<h5>Arguments:</h5>
2259<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
2260<a href="#t_integer">integer</a> type.</p>
2261
2262<h5>Semantics:</h5>
2263<p>This instruction always performs an arithmetic shift right operation,
2264The most significant bits of the result will be filled with the sign bit
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]2265of <tt>var1</tt>. If <tt>var2</tt> is (statically or dynamically) equal to or
2266larger than the number of bits in <tt>var1</tt>, the result is undefined.
2267</p>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2268
2269<h5>Example:</h5>
2270<pre>
2271 &lt;result&gt; = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
2272 &lt;result&gt; = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
2273 &lt;result&gt; = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
2274 &lt;result&gt; = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattner6ccc2d52007-10-03 21:01:14 +0000[diff] [blame]2275 &lt;result&gt; = ashr i32 1, 32 <i>; undefined</i>
Reid Spencer569f2fa2007-01-31 21:39:12 +0000[diff] [blame]2276</pre>
2277</div>
2278
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2279<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2280<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
2281Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2282<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2283<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2284<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]2285</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2286<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2287<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
2288its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2289<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2290<p>The two arguments to the '<tt>and</tt>' instruction must be <a
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]2291 href="#t_integer">integer</a> values. Both arguments must have
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2292identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2293<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2294<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2295<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2296<div style="align: center">
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2297<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2298 <tbody>
2299 <tr>
2300 <td>In0</td>
2301 <td>In1</td>
2302 <td>Out</td>
2303 </tr>
2304 <tr>
2305 <td>0</td>
2306 <td>0</td>
2307 <td>0</td>
2308 </tr>
2309 <tr>
2310 <td>0</td>
2311 <td>1</td>
2312 <td>0</td>
2313 </tr>
2314 <tr>
2315 <td>1</td>
2316 <td>0</td>
2317 <td>0</td>
2318 </tr>
2319 <tr>
2320 <td>1</td>
2321 <td>1</td>
2322 <td>1</td>
2323 </tr>
2324 </tbody>
2325</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2326</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2327<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2328<pre> &lt;result&gt; = and i32 4, %var <i>; yields {i32}:result = 4 &amp; %var</i>
2329 &lt;result&gt; = and i32 15, 40 <i>; yields {i32}:result = 8</i>
2330 &lt;result&gt; = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2331</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2332</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2333<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2334<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2335<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2336<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2337<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]2338</pre>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2339<h5>Overview:</h5>
2340<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
2341or of its two operands.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2342<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2343<p>The two arguments to the '<tt>or</tt>' instruction must be <a
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]2344 href="#t_integer">integer</a> values. Both arguments must have
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2345identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2346<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2347<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2348<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2349<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2350<table border="1" cellspacing="0" cellpadding="4">
2351 <tbody>
2352 <tr>
2353 <td>In0</td>
2354 <td>In1</td>
2355 <td>Out</td>
2356 </tr>
2357 <tr>
2358 <td>0</td>
2359 <td>0</td>
2360 <td>0</td>
2361 </tr>
2362 <tr>
2363 <td>0</td>
2364 <td>1</td>
2365 <td>1</td>
2366 </tr>
2367 <tr>
2368 <td>1</td>
2369 <td>0</td>
2370 <td>1</td>
2371 </tr>
2372 <tr>
2373 <td>1</td>
2374 <td>1</td>
2375 <td>1</td>
2376 </tr>
2377 </tbody>
2378</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2379</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2380<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2381<pre> &lt;result&gt; = or i32 4, %var <i>; yields {i32}:result = 4 | %var</i>
2382 &lt;result&gt; = or i32 15, 40 <i>; yields {i32}:result = 47</i>
2383 &lt;result&gt; = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2384</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2385</div>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2386<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2387<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2388Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2389<div class="doc_text">
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2390<h5>Syntax:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2391<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]2392</pre>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2393<h5>Overview:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2394<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2395or of its two operands. The <tt>xor</tt> is used to implement the
2396"one's complement" operation, which is the "~" operator in C.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2397<h5>Arguments:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2398<p>The two arguments to the '<tt>xor</tt>' instruction must be <a
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]2399 href="#t_integer">integer</a> values. Both arguments must have
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2400identical types.</p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2401<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2402<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2403<p> </p>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2404<div style="align: center">
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2405<table border="1" cellspacing="0" cellpadding="4">
2406 <tbody>
2407 <tr>
2408 <td>In0</td>
2409 <td>In1</td>
2410 <td>Out</td>
2411 </tr>
2412 <tr>
2413 <td>0</td>
2414 <td>0</td>
2415 <td>0</td>
2416 </tr>
2417 <tr>
2418 <td>0</td>
2419 <td>1</td>
2420 <td>1</td>
2421 </tr>
2422 <tr>
2423 <td>1</td>
2424 <td>0</td>
2425 <td>1</td>
2426 </tr>
2427 <tr>
2428 <td>1</td>
2429 <td>1</td>
2430 <td>0</td>
2431 </tr>
2432 </tbody>
2433</table>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]2434</div>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2435<p> </p>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2436<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2437<pre> &lt;result&gt; = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
2438 &lt;result&gt; = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
2439 &lt;result&gt; = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
2440 &lt;result&gt; = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2441</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2442</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2443
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2444<!-- ======================================================================= -->
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2445<div class="doc_subsection">
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2446 <a name="vectorops">Vector Operations</a>
2447</div>
2448
2449<div class="doc_text">
2450
2451<p>LLVM supports several instructions to represent vector operations in a
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2452target-independent manner. These instructions cover the element-access and
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2453vector-specific operations needed to process vectors effectively. While LLVM
2454does directly support these vector operations, many sophisticated algorithms
2455will want to use target-specific intrinsics to take full advantage of a specific
2456target.</p>
2457
2458</div>
2459
2460<!-- _______________________________________________________________________ -->
2461<div class="doc_subsubsection">
2462 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2463</div>
2464
2465<div class="doc_text">
2466
2467<h5>Syntax:</h5>
2468
2469<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2470 &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]2471</pre>
2472
2473<h5>Overview:</h5>
2474
2475<p>
2476The '<tt>extractelement</tt>' instruction extracts a single scalar
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2477element from a vector at a specified index.
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2478</p>
2479
2480
2481<h5>Arguments:</h5>
2482
2483<p>
2484The first operand of an '<tt>extractelement</tt>' instruction is a
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2485value of <a href="#t_vector">vector</a> type. The second operand is
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2486an index indicating the position from which to extract the element.
2487The index may be a variable.</p>
2488
2489<h5>Semantics:</h5>
2490
2491<p>
2492The result is a scalar of the same type as the element type of
2493<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2494<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2495results are undefined.
2496</p>
2497
2498<h5>Example:</h5>
2499
2500<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2501 %result = extractelement &lt;4 x i32&gt; %vec, i32 0 <i>; yields i32</i>
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2502</pre>
2503</div>
2504
2505
2506<!-- _______________________________________________________________________ -->
2507<div class="doc_subsubsection">
2508 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2509</div>
2510
2511<div class="doc_text">
2512
2513<h5>Syntax:</h5>
2514
2515<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2516 &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]2517</pre>
2518
2519<h5>Overview:</h5>
2520
2521<p>
2522The '<tt>insertelement</tt>' instruction inserts a scalar
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2523element into a vector at a specified index.
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2524</p>
2525
2526
2527<h5>Arguments:</h5>
2528
2529<p>
2530The first operand of an '<tt>insertelement</tt>' instruction is a
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2531value of <a href="#t_vector">vector</a> type. The second operand is a
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2532scalar value whose type must equal the element type of the first
2533operand. The third operand is an index indicating the position at
2534which to insert the value. The index may be a variable.</p>
2535
2536<h5>Semantics:</h5>
2537
2538<p>
Reid Spencer485bad12007-02-15 03:07:05 +0000[diff] [blame]2539The result is a vector of the same type as <tt>val</tt>. Its
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2540element values are those of <tt>val</tt> except at position
2541<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2542exceeds the length of <tt>val</tt>, the results are undefined.
2543</p>
2544
2545<h5>Example:</h5>
2546
2547<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2548 %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]2549</pre>
2550</div>
2551
2552<!-- _______________________________________________________________________ -->
2553<div class="doc_subsubsection">
2554 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2555</div>
2556
2557<div class="doc_text">
2558
2559<h5>Syntax:</h5>
2560
2561<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2562 &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]2563</pre>
2564
2565<h5>Overview:</h5>
2566
2567<p>
2568The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
2569from two input vectors, returning a vector of the same type.
2570</p>
2571
2572<h5>Arguments:</h5>
2573
2574<p>
2575The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2576with types that match each other and types that match the result of the
2577instruction. The third argument is a shuffle mask, which has the same number
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2578of elements as the other vector type, but whose element type is always 'i32'.
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2579</p>
2580
2581<p>
2582The shuffle mask operand is required to be a constant vector with either
2583constant integer or undef values.
2584</p>
2585
2586<h5>Semantics:</h5>
2587
2588<p>
2589The elements of the two input vectors are numbered from left to right across
2590both of the vectors. The shuffle mask operand specifies, for each element of
2591the result vector, which element of the two input registers the result element
2592gets. The element selector may be undef (meaning "don't care") and the second
2593operand may be undef if performing a shuffle from only one vector.
2594</p>
2595
2596<h5>Example:</h5>
2597
2598<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2599 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; %v2,
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2600 &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]2601 %result = shufflevector &lt;4 x i32&gt; %v1, &lt;4 x i32&gt; undef,
2602 &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]2603</pre>
2604</div>
2605
Tanya Lattner09474292006-04-14 19:24:33 +0000[diff] [blame]2606
Chris Lattner3df241e2006-04-08 23:07:04 +0000[diff] [blame]2607<!-- ======================================================================= -->
2608<div class="doc_subsection">
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]2609 <a name="memoryops">Memory Access and Addressing Operations</a>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2610</div>
2611
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2612<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2613
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2614<p>A key design point of an SSA-based representation is how it
2615represents memory. In LLVM, no memory locations are in SSA form, which
2616makes things very simple. This section describes how to read, write,
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2617allocate, and free memory in LLVM.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2618
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2619</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2620
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2621<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2622<div class="doc_subsubsection">
2623 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
2624</div>
2625
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2626<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2627
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2628<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2629
2630<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2631 &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]2632</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2633
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2634<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2635
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2636<p>The '<tt>malloc</tt>' instruction allocates memory from the system
2637heap and returns a pointer to it.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2638
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2639<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2640
2641<p>The '<tt>malloc</tt>' instruction allocates
2642<tt>sizeof(&lt;type&gt;)*NumElements</tt>
John Criswell6e4ca612004-02-24 16:13:56 +0000[diff] [blame]2643bytes of memory from the operating system and returns a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2644appropriate type to the program. If "NumElements" is specified, it is the
2645number of elements allocated. If an alignment is specified, the value result
2646of the allocation is guaranteed to be aligned to at least that boundary. If
2647not specified, or if zero, the target can choose to align the allocation on any
2648convenient boundary.</p>
2649
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2650<p>'<tt>type</tt>' must be a sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2651
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2652<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2653
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2654<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
2655a pointer is returned.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2656
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2657<h5>Example:</h5>
2658
2659<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2660 %array = malloc [4 x i8 ] <i>; yields {[%4 x i8]*}:array</i>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2661
Bill Wendlingaac388b2007-05-29 09:42:13 +0000[diff] [blame]2662 %size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
2663 %array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
2664 %array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
2665 %array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
2666 %array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2667</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2668</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2669
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2670<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2671<div class="doc_subsubsection">
2672 <a name="i_free">'<tt>free</tt>' Instruction</a>
2673</div>
2674
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2675<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2676
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2677<h5>Syntax:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2678
2679<pre>
2680 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2681</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2682
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2683<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2684
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2685<p>The '<tt>free</tt>' instruction returns memory back to the unused
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]2686memory heap to be reallocated in the future.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2687
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2688<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2689
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2690<p>'<tt>value</tt>' shall be a pointer value that points to a value
2691that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
2692instruction.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2693
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2694<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2695
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2696<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2697after this instruction executes.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2698
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2699<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2700
2701<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2702 %array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
2703 free [4 x i8]* %array
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2704</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2705</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2706
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2707<!-- _______________________________________________________________________ -->
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2708<div class="doc_subsubsection">
2709 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
2710</div>
2711
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2712<div class="doc_text">
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2713
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2714<h5>Syntax:</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 &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]2718</pre>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2719
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2720<h5>Overview:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2721
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2722<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
2723currently executing function, to be automatically released when this function
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2724returns to its caller.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2725
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2726<h5>Arguments:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2727
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]2728<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2729bytes of memory on the runtime stack, returning a pointer of the
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2730appropriate type to the program. If "NumElements" is specified, it is the
2731number of elements allocated. If an alignment is specified, the value result
2732of the allocation is guaranteed to be aligned to at least that boundary. If
2733not specified, or if zero, the target can choose to align the allocation on any
2734convenient boundary.</p>
2735
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2736<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2737
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2738<h5>Semantics:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2739
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]2740<p>Memory is allocated; a pointer is returned. '<tt>alloca</tt>'d
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2741memory is automatically released when the function returns. The '<tt>alloca</tt>'
2742instruction is commonly used to represent automatic variables that must
2743have an address available. When the function returns (either with the <tt><a
John Criswelldae2e932005-05-12 16:55:34 +0000[diff] [blame]2744 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]2745instructions), the memory is reclaimed.</p>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2746
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2747<h5>Example:</h5>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2748
2749<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2750 %ptr = alloca i32 <i>; yields {i32*}:ptr</i>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2751 %ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
2752 %ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2753 %ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2754</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2755</div>
Chris Lattner2cbdc452005-11-06 08:02:57 +0000[diff] [blame]2756
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2757<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2758<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
2759Instruction</a> </div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2760<div class="doc_text">
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2761<h5>Syntax:</h5>
Christopher Lamb2330e4d2007-04-21 08:16:25 +0000[diff] [blame]2762<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]2763<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2764<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2765<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2766<p>The argument to the '<tt>load</tt>' instruction specifies the memory
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]2767address from which to load. The pointer must point to a <a
Chris Lattnere53e5082004-06-03 22:57:15 +0000[diff] [blame]2768 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
John Criswell0ec250c2005-10-24 16:17:18 +0000[diff] [blame]2769marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2770the number or order of execution of this <tt>load</tt> with other
2771volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
2772instructions. </p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2773<h5>Semantics:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2774<p>The location of memory pointed to is loaded.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2775<h5>Examples:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2776<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2777 <a
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2778 href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
2779 %val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2780</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2781</div>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2782<!-- _______________________________________________________________________ -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2783<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
2784Instruction</a> </div>
Reid Spencer035ab572006-11-09 21:18:01 +0000[diff] [blame]2785<div class="doc_text">
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2786<h5>Syntax:</h5>
Christopher Lamb2330e4d2007-04-21 08:16:25 +0000[diff] [blame]2787<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt;[, align &lt;alignment&gt;] <i>; yields {void}</i>
2788 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]2789</pre>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2790<h5>Overview:</h5>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2791<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2792<h5>Arguments:</h5>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2793<p>There are two arguments to the '<tt>store</tt>' instruction: a value
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2794to 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]2795operand must be a pointer to the type of the '<tt>&lt;value&gt;</tt>'
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]2796operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2797optimizer is not allowed to modify the number or order of execution of
2798this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
2799 href="#i_store">store</a></tt> instructions.</p>
2800<h5>Semantics:</h5>
2801<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
2802at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2803<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2804<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling8c6c72d2007-10-22 05:10:05 +0000[diff] [blame]2805 store i32 3, i32* %ptr <i>; yields {void}</i>
2806 %val = <a href="#i_load">load</a> i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2807</pre>
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]2808</div>
2809
Chris Lattner2b7d3202002-05-06 03:03:22 +0000[diff] [blame]2810<!-- _______________________________________________________________________ -->
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2811<div class="doc_subsubsection">
2812 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
2813</div>
2814
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2815<div class="doc_text">
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2816<h5>Syntax:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2817<pre>
2818 &lt;result&gt; = getelementptr &lt;ty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
2819</pre>
2820
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2821<h5>Overview:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2822
2823<p>
2824The '<tt>getelementptr</tt>' instruction is used to get the address of a
2825subelement of an aggregate data structure.</p>
2826
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2827<h5>Arguments:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2828
Reid Spencer85f5b5b2006-12-04 21:29:24 +0000[diff] [blame]2829<p>This instruction takes a list of integer operands that indicate what
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2830elements of the aggregate object to index to. The actual types of the arguments
2831provided depend on the type of the first pointer argument. The
2832'<tt>getelementptr</tt>' instruction is used to index down through the type
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]2833levels 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]2834structure, only <tt>i32</tt> integer constants are allowed. When indexing
Reid Spencer85f5b5b2006-12-04 21:29:24 +0000[diff] [blame]2835into an array or pointer, only integers of 32 or 64 bits are allowed, and will
2836be sign extended to 64-bit values.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2837
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2838<p>For example, let's consider a C code fragment and how it gets
2839compiled to LLVM:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2840
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2841<div class="doc_code">
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2842<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2843struct RT {
2844 char A;
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]2845 int B[10][20];
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2846 char C;
2847};
2848struct ST {
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]2849 int X;
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2850 double Y;
2851 struct RT Z;
2852};
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2853
Chris Lattnercabc8462007-05-29 15:43:56 +0000[diff] [blame]2854int *foo(struct ST *s) {
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2855 return &amp;s[1].Z.B[5][13];
2856}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2857</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2858</div>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2859
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2860<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2861
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2862<div class="doc_code">
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2863<pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2864%RT = type { i8 , [10 x [20 x i32]], i8 }
2865%ST = type { i32, double, %RT }
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2866
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2867define i32* %foo(%ST* %s) {
2868entry:
2869 %reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
2870 ret i32* %reg
2871}
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2872</pre>
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]2873</div>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2874
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2875<h5>Semantics:</h5>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2876
2877<p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
John Criswellc1f786c2005-05-13 22:25:59 +0000[diff] [blame]2878on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
Reid Spencer85f5b5b2006-12-04 21:29:24 +0000[diff] [blame]2879and <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]2880<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]2881to 64-bits. <a href="#t_struct">Structure</a> types require <tt>i32</tt>
Reid Spencer42ddd842006-12-03 16:53:48 +0000[diff] [blame]2882<b>constants</b>.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2883
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2884<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]2885type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2886}</tt>' type, a structure. The second index indexes into the third element of
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2887the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
2888i8 }</tt>' type, another structure. The third index indexes into the second
2889element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2890array. The two dimensions of the array are subscripted into, yielding an
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2891'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
2892to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2893
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]2894<p>Note that it is perfectly legal to index partially through a
2895structure, returning a pointer to an inner element. Because of this,
2896the LLVM code for the given testcase is equivalent to:</p>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2897
2898<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2899 define i32* %foo(%ST* %s) {
2900 %t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen6f1cc772007-04-22 01:17:39 +0000[diff] [blame]2901 %t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
2902 %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]2903 %t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
2904 %t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
2905 ret i32* %t5
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2906 }
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2907</pre>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]2908
2909<p>Note that it is undefined to access an array out of bounds: array and
2910pointer indexes must always be within the defined bounds of the array type.
2911The one exception for this rules is zero length arrays. These arrays are
2912defined to be accessible as variable length arrays, which requires access
2913beyond the zero'th element.</p>
2914
Chris Lattner884a9702006-08-15 00:45:58 +0000[diff] [blame]2915<p>The getelementptr instruction is often confusing. For some more insight
2916into how it works, see <a href="GetElementPtr.html">the getelementptr
2917FAQ</a>.</p>
2918
Chris Lattner7faa8832002-04-14 06:13:44 +0000[diff] [blame]2919<h5>Example:</h5>
Chris Lattnere67a9512005-06-24 17:22:57 +0000[diff] [blame]2920
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2921<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2922 <i>; yields [12 x i8]*:aptr</i>
2923 %aptr = getelementptr {i32, [12 x i8]}* %sptr, i64 0, i32 1
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2924</pre>
Chris Lattnerf74d5c72004-04-05 01:30:49 +0000[diff] [blame]2925</div>
Reid Spencer47ce1792006-11-09 21:15:49 +0000[diff] [blame]2926
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]2927<!-- ======================================================================= -->
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]2928<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2929</div>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2930<div class="doc_text">
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]2931<p>The instructions in this category are the conversion instructions (casting)
2932which all take a single operand and a type. They perform various bit conversions
2933on the operand.</p>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]2934</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2935
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]2936<!-- _______________________________________________________________________ -->
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]2937<div class="doc_subsubsection">
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2938 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
2939</div>
2940<div class="doc_text">
2941
2942<h5>Syntax:</h5>
2943<pre>
2944 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2945</pre>
2946
2947<h5>Overview:</h5>
2948<p>
2949The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
2950</p>
2951
2952<h5>Arguments:</h5>
2953<p>
2954The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
2955be 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]2956and type of the result, which must be an <a href="#t_integer">integer</a>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]2957type. The bit size of <tt>value</tt> must be larger than the bit size of
2958<tt>ty2</tt>. Equal sized types are not allowed.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2959
2960<h5>Semantics:</h5>
2961<p>
2962The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]2963and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
2964larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
2965It will always truncate bits.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2966
2967<h5>Example:</h5>
2968<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]2969 %X = trunc i32 257 to i8 <i>; yields i8:1</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]2970 %Y = trunc i32 123 to i1 <i>; yields i1:true</i>
2971 %Y = trunc i32 122 to i1 <i>; yields i1:false</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2972</pre>
2973</div>
2974
2975<!-- _______________________________________________________________________ -->
2976<div class="doc_subsubsection">
2977 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
2978</div>
2979<div class="doc_text">
2980
2981<h5>Syntax:</h5>
2982<pre>
2983 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2984</pre>
2985
2986<h5>Overview:</h5>
2987<p>The '<tt>zext</tt>' instruction zero extends its operand to type
2988<tt>ty2</tt>.</p>
2989
2990
2991<h5>Arguments:</h5>
2992<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]2993<a href="#t_integer">integer</a> type, and a type to cast it to, which must
2994also be of <a href="#t_integer">integer</a> type. The bit size of the
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]2995<tt>value</tt> must be smaller than the bit size of the destination type,
2996<tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]2997
2998<h5>Semantics:</h5>
2999<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]3000bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3001
Reid Spencerb5929522007-01-12 15:46:11 +0000[diff] [blame]3002<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]3003
3004<h5>Example:</h5>
3005<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3006 %X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3007 %Y = zext i1 true to i32 <i>; yields i32:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3008</pre>
3009</div>
3010
3011<!-- _______________________________________________________________________ -->
3012<div class="doc_subsubsection">
3013 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
3014</div>
3015<div class="doc_text">
3016
3017<h5>Syntax:</h5>
3018<pre>
3019 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3020</pre>
3021
3022<h5>Overview:</h5>
3023<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
3024
3025<h5>Arguments:</h5>
3026<p>
3027The '<tt>sext</tt>' instruction takes a value to cast, which must be of
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]3028<a href="#t_integer">integer</a> type, and a type to cast it to, which must
3029also be of <a href="#t_integer">integer</a> type. The bit size of the
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3030<tt>value</tt> must be smaller than the bit size of the destination type,
3031<tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3032
3033<h5>Semantics:</h5>
3034<p>
3035The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
3036bit (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]3037the type <tt>ty2</tt>.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3038
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3039<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]3040
3041<h5>Example:</h5>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3042<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3043 %X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3044 %Y = sext i1 true to i32 <i>; yields i32:-1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3045</pre>
3046</div>
3047
3048<!-- _______________________________________________________________________ -->
3049<div class="doc_subsubsection">
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]3050 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
3051</div>
3052
3053<div class="doc_text">
3054
3055<h5>Syntax:</h5>
3056
3057<pre>
3058 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3059</pre>
3060
3061<h5>Overview:</h5>
3062<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
3063<tt>ty2</tt>.</p>
3064
3065
3066<h5>Arguments:</h5>
3067<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
3068 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
3069cast it to. The size of <tt>value</tt> must be larger than the size of
3070<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
3071<i>no-op cast</i>.</p>
3072
3073<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3074<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
3075<a href="#t_floating">floating point</a> type to a smaller
3076<a href="#t_floating">floating point</a> type. If the value cannot fit within
3077the destination type, <tt>ty2</tt>, then the results are undefined.</p>
Reid Spencer3fa91b02006-11-09 21:48:10 +0000[diff] [blame]3078
3079<h5>Example:</h5>
3080<pre>
3081 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
3082 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
3083</pre>
3084</div>
3085
3086<!-- _______________________________________________________________________ -->
3087<div class="doc_subsubsection">
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3088 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
3089</div>
3090<div class="doc_text">
3091
3092<h5>Syntax:</h5>
3093<pre>
3094 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3095</pre>
3096
3097<h5>Overview:</h5>
3098<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
3099floating point value.</p>
3100
3101<h5>Arguments:</h5>
3102<p>The '<tt>fpext</tt>' instruction takes a
3103<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3104and a <a href="#t_floating">floating point</a> type to cast it to. The source
3105type must be smaller than the destination type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3106
3107<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3108<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3109<a href="#t_floating">floating point</a> type to a larger
3110<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3111used to make a <i>no-op cast</i> because it always changes bits. Use
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]3112<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]3113
3114<h5>Example:</h5>
3115<pre>
3116 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
3117 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
3118</pre>
3119</div>
3120
3121<!-- _______________________________________________________________________ -->
3122<div class="doc_subsubsection">
Reid Spencer24d6da52007-01-21 00:29:26 +0000[diff] [blame]3123 <a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3124</div>
3125<div class="doc_text">
3126
3127<h5>Syntax:</h5>
3128<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]3129 &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]3130</pre>
3131
3132<h5>Overview:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]3133<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3134unsigned integer equivalent of type <tt>ty2</tt>.
3135</p>
3136
3137<h5>Arguments:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]3138<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]3139<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]3140must be an <a href="#t_integer">integer</a> type.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3141
3142<h5>Semantics:</h5>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]3143<p> The '<tt>fptoui</tt>' instruction converts its
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3144<a href="#t_floating">floating point</a> operand into the nearest (rounding
3145towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
3146the results are undefined.</p>
3147
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3148<h5>Example:</h5>
3149<pre>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]3150 %X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]3151 %Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencer1539a1c2007-07-31 14:40:14 +0000[diff] [blame]3152 %X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3153</pre>
3154</div>
3155
3156<!-- _______________________________________________________________________ -->
3157<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3158 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3159</div>
3160<div class="doc_text">
3161
3162<h5>Syntax:</h5>
3163<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3164 &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]3165</pre>
3166
3167<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3168<p>The '<tt>fptosi</tt>' instruction converts
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3169<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]3170</p>
3171
3172
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3173<h5>Arguments:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3174<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]3175<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]3176must also be an <a href="#t_integer">integer</a> type.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3177
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3178<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3179<p>The '<tt>fptosi</tt>' instruction converts its
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3180<a href="#t_floating">floating point</a> operand into the nearest (rounding
3181towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
3182the results are undefined.</p>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3183
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]3184<h5>Example:</h5>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3185<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3186 %X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner88519042007-09-22 03:17:52 +0000[diff] [blame]3187 %Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3188 %X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3189</pre>
3190</div>
3191
3192<!-- _______________________________________________________________________ -->
3193<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3194 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3195</div>
3196<div class="doc_text">
3197
3198<h5>Syntax:</h5>
3199<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3200 &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]3201</pre>
3202
3203<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3204<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3205integer and converts that value to the <tt>ty2</tt> type.</p>
3206
3207
3208<h5>Arguments:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3209<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]3210<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]3211be a <a href="#t_floating">floating point</a> type.</p>
3212
3213<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3214<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3215integer quantity and converts it to the corresponding floating point value. If
Jeff Cohencb757312007-04-22 14:56:37 +0000[diff] [blame]3216the value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3217
3218
3219<h5>Example:</h5>
3220<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3221 %X = uitofp i32 257 to float <i>; yields float:257.0</i>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3222 %Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3223</pre>
3224</div>
3225
3226<!-- _______________________________________________________________________ -->
3227<div class="doc_subsubsection">
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3228 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3229</div>
3230<div class="doc_text">
3231
3232<h5>Syntax:</h5>
3233<pre>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3234 &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]3235</pre>
3236
3237<h5>Overview:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3238<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3239integer and converts that value to the <tt>ty2</tt> type.</p>
3240
3241<h5>Arguments:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3242<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]3243<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]3244a <a href="#t_floating">floating point</a> type.</p>
3245
3246<h5>Semantics:</h5>
Reid Spencerd4448792006-11-09 23:03:26 +0000[diff] [blame]3247<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3248integer quantity and converts it to the corresponding floating point value. If
Jeff Cohencb757312007-04-22 14:56:37 +0000[diff] [blame]3249the value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3250
3251<h5>Example:</h5>
3252<pre>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3253 %X = sitofp i32 257 to float <i>; yields float:257.0</i>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3254 %Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3255</pre>
3256</div>
3257
3258<!-- _______________________________________________________________________ -->
3259<div class="doc_subsubsection">
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]3260 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
3261</div>
3262<div class="doc_text">
3263
3264<h5>Syntax:</h5>
3265<pre>
3266 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3267</pre>
3268
3269<h5>Overview:</h5>
3270<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
3271the integer type <tt>ty2</tt>.</p>
3272
3273<h5>Arguments:</h5>
3274<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3275must 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]3276<tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.
3277
3278<h5>Semantics:</h5>
3279<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
3280<tt>ty2</tt> by interpreting the pointer value as an integer and either
3281truncating or zero extending that value to the size of the integer type. If
3282<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
3283<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]3284are the same size, then nothing is done (<i>no-op cast</i>) other than a type
3285change.</p>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]3286
3287<h5>Example:</h5>
3288<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3289 %X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
3290 %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]3291</pre>
3292</div>
3293
3294<!-- _______________________________________________________________________ -->
3295<div class="doc_subsubsection">
3296 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
3297</div>
3298<div class="doc_text">
3299
3300<h5>Syntax:</h5>
3301<pre>
3302 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3303</pre>
3304
3305<h5>Overview:</h5>
3306<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
3307a pointer type, <tt>ty2</tt>.</p>
3308
3309<h5>Arguments:</h5>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3310<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]3311value to cast, and a type to cast it to, which must be a
Anton Korobeynikov7f705592007-01-12 19:20:47 +0000[diff] [blame]3312<a href="#t_pointer">pointer</a> type.
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]3313
3314<h5>Semantics:</h5>
3315<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
3316<tt>ty2</tt> by applying either a zero extension or a truncation depending on
3317the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
3318size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
3319the size of a pointer then a zero extension is done. If they are the same size,
3320nothing is done (<i>no-op cast</i>).</p>
3321
3322<h5>Example:</h5>
3323<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3324 %X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
3325 %X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
3326 %Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]3327</pre>
3328</div>
3329
3330<!-- _______________________________________________________________________ -->
3331<div class="doc_subsubsection">
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]3332 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3333</div>
3334<div class="doc_text">
3335
3336<h5>Syntax:</h5>
3337<pre>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]3338 &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]3339</pre>
3340
3341<h5>Overview:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]3342<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3343<tt>ty2</tt> without changing any bits.</p>
3344
3345<h5>Arguments:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]3346<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3347a first class value, and a type to cast it to, which must also be a <a
3348 href="#t_firstclass">first class</a> type. The bit sizes of <tt>value</tt>
Reid Spencer19b569f2007-01-09 20:08:58 +0000[diff] [blame]3349and the destination type, <tt>ty2</tt>, must be identical. If the source
3350type is a pointer, the destination type must also be a pointer.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3351
3352<h5>Semantics:</h5>
Reid Spencer5c0ef472006-11-11 23:08:07 +0000[diff] [blame]3353<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Reid Spencer72679252006-11-11 21:00:47 +0000[diff] [blame]3354<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
3355this conversion. The conversion is done as if the <tt>value</tt> had been
3356stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
3357converted to other pointer types with this instruction. To convert pointers to
3358other types, use the <a href="#i_inttoptr">inttoptr</a> or
3359<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer9dee3ac2006-11-08 01:11:31 +0000[diff] [blame]3360
3361<h5>Example:</h5>
3362<pre>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3363 %X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3364 %Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
3365 %Z = bitcast <2xint> %V to i64; <i>; yields i64: %V</i>
Chris Lattner33ba0d92001-07-09 00:26:23 +0000[diff] [blame]3366</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3367</div>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3368
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]3369<!-- ======================================================================= -->
3370<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
3371<div class="doc_text">
3372<p>The instructions in this category are the "miscellaneous"
3373instructions, which defy better classification.</p>
3374</div>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3375
3376<!-- _______________________________________________________________________ -->
3377<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
3378</div>
3379<div class="doc_text">
3380<h5>Syntax:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3381<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]3382</pre>
3383<h5>Overview:</h5>
3384<p>The '<tt>icmp</tt>' instruction returns a boolean value based on comparison
3385of its two integer operands.</p>
3386<h5>Arguments:</h5>
3387<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3388the condition code indicating the kind of comparison to perform. It is not
3389a value, just a keyword. The possible condition code are:
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3390<ol>
3391 <li><tt>eq</tt>: equal</li>
3392 <li><tt>ne</tt>: not equal </li>
3393 <li><tt>ugt</tt>: unsigned greater than</li>
3394 <li><tt>uge</tt>: unsigned greater or equal</li>
3395 <li><tt>ult</tt>: unsigned less than</li>
3396 <li><tt>ule</tt>: unsigned less or equal</li>
3397 <li><tt>sgt</tt>: signed greater than</li>
3398 <li><tt>sge</tt>: signed greater or equal</li>
3399 <li><tt>slt</tt>: signed less than</li>
3400 <li><tt>sle</tt>: signed less or equal</li>
3401</ol>
Chris Lattner3b19d652007-01-15 01:54:13 +0000[diff] [blame]3402<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Reid Spencer350f8aa2007-01-04 05:19:58 +0000[diff] [blame]3403<a href="#t_pointer">pointer</a> typed. They must also be identical types.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3404<h5>Semantics:</h5>
3405<p>The '<tt>icmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3406the condition code given as <tt>cond</tt>. The comparison performed always
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3407yields a <a href="#t_primitive">i1</a> result, as follows:
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3408<ol>
3409 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
3410 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3411 </li>
3412 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
3413 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3414 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
3415 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3416 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
3417 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3418 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
3419 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3420 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
3421 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3422 <li><tt>sgt</tt>: interprets the operands as signed values and yields
3423 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3424 <li><tt>sge</tt>: interprets the operands as signed values and yields
3425 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3426 <li><tt>slt</tt>: interprets the operands as signed values and yields
3427 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3428 <li><tt>sle</tt>: interprets the operands as signed values and yields
3429 <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]3430</ol>
3431<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3432values are compared as if they were integers.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3433
3434<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3435<pre> &lt;result&gt; = icmp eq i32 4, 5 <i>; yields: result=false</i>
3436 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
3437 &lt;result&gt; = icmp ult i16 4, 5 <i>; yields: result=true</i>
3438 &lt;result&gt; = icmp sgt i16 4, 5 <i>; yields: result=false</i>
3439 &lt;result&gt; = icmp ule i16 -4, 5 <i>; yields: result=false</i>
3440 &lt;result&gt; = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3441</pre>
3442</div>
3443
3444<!-- _______________________________________________________________________ -->
3445<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
3446</div>
3447<div class="doc_text">
3448<h5>Syntax:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3449<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]3450</pre>
3451<h5>Overview:</h5>
3452<p>The '<tt>fcmp</tt>' instruction returns a boolean value based on comparison
3453of its floating point operands.</p>
3454<h5>Arguments:</h5>
3455<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3456the condition code indicating the kind of comparison to perform. It is not
3457a value, just a keyword. The possible condition code are:
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3458<ol>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3459 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3460 <li><tt>oeq</tt>: ordered and equal</li>
3461 <li><tt>ogt</tt>: ordered and greater than </li>
3462 <li><tt>oge</tt>: ordered and greater than or equal</li>
3463 <li><tt>olt</tt>: ordered and less than </li>
3464 <li><tt>ole</tt>: ordered and less than or equal</li>
3465 <li><tt>one</tt>: ordered and not equal</li>
3466 <li><tt>ord</tt>: ordered (no nans)</li>
3467 <li><tt>ueq</tt>: unordered or equal</li>
3468 <li><tt>ugt</tt>: unordered or greater than </li>
3469 <li><tt>uge</tt>: unordered or greater than or equal</li>
3470 <li><tt>ult</tt>: unordered or less than </li>
3471 <li><tt>ule</tt>: unordered or less than or equal</li>
3472 <li><tt>une</tt>: unordered or not equal</li>
3473 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3474 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3475</ol>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3476<p><i>Ordered</i> means that neither operand is a QNAN while
Reid Spencer93a49852006-12-06 07:08:07 +0000[diff] [blame]3477<i>unordered</i> means that either operand may be a QNAN.</p>
Reid Spencer350f8aa2007-01-04 05:19:58 +0000[diff] [blame]3478<p>The <tt>val1</tt> and <tt>val2</tt> arguments must be
3479<a href="#t_floating">floating point</a> typed. They must have identical
3480types.</p>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3481<h5>Semantics:</h5>
3482<p>The '<tt>fcmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3483the condition code given as <tt>cond</tt>. The comparison performed always
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3484yields a <a href="#t_primitive">i1</a> result, as follows:
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3485<ol>
3486 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3487 <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]3488 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3489 <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]3490 <tt>var1</tt> is greather than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3491 <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]3492 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3493 <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]3494 <tt>var1</tt> is less than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3495 <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]3496 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3497 <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]3498 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3499 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
3500 <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]3501 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3502 <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]3503 <tt>var1</tt> is greater than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3504 <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]3505 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3506 <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]3507 <tt>var1</tt> is less than <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3508 <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]3509 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3510 <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]3511 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
Reid Spencerb7f26282006-11-19 03:00:14 +0000[diff] [blame]3512 <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]3513 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
3514</ol>
Reid Spencerf3a70a62006-11-18 21:50:54 +0000[diff] [blame]3515
3516<h5>Example:</h5>
3517<pre> &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
3518 &lt;result&gt; = icmp one float 4.0, 5.0 <i>; yields: result=true</i>
3519 &lt;result&gt; = icmp olt float 4.0, 5.0 <i>; yields: result=true</i>
3520 &lt;result&gt; = icmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
3521</pre>
3522</div>
3523
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]3524<!-- _______________________________________________________________________ -->
3525<div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
3526Instruction</a> </div>
3527<div class="doc_text">
3528<h5>Syntax:</h5>
3529<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
3530<h5>Overview:</h5>
3531<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
3532the SSA graph representing the function.</p>
3533<h5>Arguments:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3534<p>The type of the incoming values is specified with the first type
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]3535field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
3536as arguments, with one pair for each predecessor basic block of the
3537current block. Only values of <a href="#t_firstclass">first class</a>
3538type may be used as the value arguments to the PHI node. Only labels
3539may be used as the label arguments.</p>
3540<p>There must be no non-phi instructions between the start of a basic
3541block and the PHI instructions: i.e. PHI instructions must be first in
3542a basic block.</p>
3543<h5>Semantics:</h5>
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3544<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
3545specified by the pair corresponding to the predecessor basic block that executed
3546just prior to the current block.</p>
Reid Spencer2fd21e62006-11-08 01:18:52 +0000[diff] [blame]3547<h5>Example:</h5>
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3548<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]3549</div>
3550
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3551<!-- _______________________________________________________________________ -->
3552<div class="doc_subsubsection">
3553 <a name="i_select">'<tt>select</tt>' Instruction</a>
3554</div>
3555
3556<div class="doc_text">
3557
3558<h5>Syntax:</h5>
3559
3560<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3561 &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]3562</pre>
3563
3564<h5>Overview:</h5>
3565
3566<p>
3567The '<tt>select</tt>' instruction is used to choose one value based on a
3568condition, without branching.
3569</p>
3570
3571
3572<h5>Arguments:</h5>
3573
3574<p>
3575The '<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.
3576</p>
3577
3578<h5>Semantics:</h5>
3579
3580<p>
3581If the boolean condition evaluates to true, the instruction returns the first
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]3582value argument; otherwise, it returns the second value argument.
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3583</p>
3584
3585<h5>Example:</h5>
3586
3587<pre>
Reid Spencerc78f3372007-01-12 03:35:51 +0000[diff] [blame]3588 %X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattnercc37aae2004-03-12 05:50:16 +0000[diff] [blame]3589</pre>
3590</div>
3591
Robert Bocchino05ccd702006-01-15 20:48:27 +0000[diff] [blame]3592
3593<!-- _______________________________________________________________________ -->
3594<div class="doc_subsubsection">
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3595 <a name="i_call">'<tt>call</tt>' Instruction</a>
3596</div>
3597
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3598<div class="doc_text">
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3599
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3600<h5>Syntax:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3601<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +0000[diff] [blame]3602 &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]3603</pre>
3604
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3605<h5>Overview:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3606
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3607<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3608
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3609<h5>Arguments:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3610
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3611<p>This instruction requires several arguments:</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3612
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3613<ol>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3614 <li>
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3615 <p>The optional "tail" marker indicates whether the callee function accesses
3616 any allocas or varargs in the caller. If the "tail" marker is present, the
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3617 function call is eligible for tail call optimization. Note that calls may
3618 be marked "tail" even if they do not occur before a <a
3619 href="#i_ret"><tt>ret</tt></a> instruction.
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3620 </li>
3621 <li>
Duncan Sands8036ca42007-03-30 12:22:09 +0000[diff] [blame]3622 <p>The optional "cconv" marker indicates which <a href="#callingconv">calling
Chris Lattnerbad10ee2005-05-06 22:57:40 +0000[diff] [blame]3623 convention</a> the call should use. If none is specified, the call defaults
3624 to using C calling conventions.
3625 </li>
3626 <li>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +0000[diff] [blame]3627 <p>'<tt>ty</tt>': the type of the call instruction itself which is also
3628 the type of the return value. Functions that return no value are marked
3629 <tt><a href="#t_void">void</a></tt>.</p>
3630 </li>
3631 <li>
3632 <p>'<tt>fnty</tt>': shall be the signature of the pointer to function
3633 value being invoked. The argument types must match the types implied by
3634 this signature. This type can be omitted if the function is not varargs
3635 and if the function type does not return a pointer to a function.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3636 </li>
3637 <li>
3638 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
3639 be invoked. In most cases, this is a direct function invocation, but
3640 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]3641 to function value.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3642 </li>
3643 <li>
3644 <p>'<tt>function args</tt>': argument list whose types match the
Reid Spencera7e302a2005-05-01 22:22:57 +0000[diff] [blame]3645 function signature argument types. All arguments must be of
3646 <a href="#t_firstclass">first class</a> type. If the function signature
3647 indicates the function accepts a variable number of arguments, the extra
3648 arguments can be specified.</p>
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3649 </li>
Chris Lattner6536cfe2002-05-06 22:08:29 +0000[diff] [blame]3650</ol>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3651
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3652<h5>Semantics:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3653
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3654<p>The '<tt>call</tt>' instruction is used to cause control flow to
3655transfer to a specified function, with its incoming arguments bound to
3656the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
3657instruction in the called function, control flow continues with the
3658instruction after the function call, and the return value of the
3659function is bound to the result argument. This is a simpler case of
3660the <a href="#i_invoke">invoke</a> instruction.</p>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3661
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]3662<h5>Example:</h5>
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3663
3664<pre>
Nick Lewyckydb7e3c92007-09-08 13:57:50 +0000[diff] [blame]3665 %retval = call i32 @test(i32 %argc)
3666 call i32 (i8 *, ...)* @printf(i8 * %msg, i32 12, i8 42);
3667 %X = tail call i32 @foo()
3668 %Y = tail call <a href="#callingconv">fastcc</a> i32 @foo()
3669 %Z = call void %foo(i8 97 signext)
Chris Lattner2bff5242005-05-06 05:47:36 +0000[diff] [blame]3670</pre>
3671
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3672</div>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3673
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3674<!-- _______________________________________________________________________ -->
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3675<div class="doc_subsubsection">
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]3676 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3677</div>
3678
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3679<div class="doc_text">
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3680
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]3681<h5>Syntax:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3682
3683<pre>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3684 &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]3685</pre>
3686
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]3687<h5>Overview:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3688
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3689<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3690the "variable argument" area of a function call. It is used to implement the
3691<tt>va_arg</tt> macro in C.</p>
3692
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]3693<h5>Arguments:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3694
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3695<p>This instruction takes a <tt>va_list*</tt> value and the type of
3696the argument. It returns a value of the specified argument type and
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3697increments the <tt>va_list</tt> to point to the next argument. The
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3698actual type of <tt>va_list</tt> is target specific.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3699
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]3700<h5>Semantics:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3701
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3702<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
3703type from the specified <tt>va_list</tt> and causes the
3704<tt>va_list</tt> to point to the next argument. For more information,
3705see the variable argument handling <a href="#int_varargs">Intrinsic
3706Functions</a>.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3707
3708<p>It is legal for this instruction to be called in a function which does not
3709take a variable number of arguments, for example, the <tt>vfprintf</tt>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3710function.</p>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3711
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3712<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]3713href="#intrinsics">intrinsic function</a> because it takes a type as an
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3714argument.</p>
3715
Chris Lattner8d1a81d2003-10-18 05:51:36 +0000[diff] [blame]3716<h5>Example:</h5>
Chris Lattnere19d7a72004-09-27 21:51:25 +0000[diff] [blame]3717
3718<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
3719
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3720</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3721
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3722<!-- *********************************************************************** -->
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3723<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
3724<!-- *********************************************************************** -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3725
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3726<div class="doc_text">
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3727
3728<p>LLVM supports the notion of an "intrinsic function". These functions have
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]3729well known names and semantics and are required to follow certain restrictions.
3730Overall, these intrinsics represent an extension mechanism for the LLVM
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3731language that does not require changing all of the transformations in LLVM when
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]3732adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3733
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]3734<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3735prefix is reserved in LLVM for intrinsic names; thus, function names may not
3736begin with this prefix. Intrinsic functions must always be external functions:
3737you cannot define the body of intrinsic functions. Intrinsic functions may
3738only be used in call or invoke instructions: it is illegal to take the address
3739of an intrinsic function. Additionally, because intrinsic functions are part
3740of the LLVM language, it is required if any are added that they be documented
3741here.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3742
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]3743<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
3744a family of functions that perform the same operation but on different data
3745types. Because LLVM can represent over 8 million different integer types,
3746overloading is used commonly to allow an intrinsic function to operate on any
3747integer type. One or more of the argument types or the result type can be
3748overloaded to accept any integer type. Argument types may also be defined as
3749exactly matching a previous argument's type or the result type. This allows an
3750intrinsic function which accepts multiple arguments, but needs all of them to
3751be of the same type, to only be overloaded with respect to a single argument or
3752the result.</p>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3753
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]3754<p>Overloaded intrinsics will have the names of its overloaded argument types
3755encoded into its function name, each preceded by a period. Only those types
3756which are overloaded result in a name suffix. Arguments whose type is matched
3757against another type do not. For example, the <tt>llvm.ctpop</tt> function can
3758take an integer of any width and returns an integer of exactly the same integer
3759width. This leads to a family of functions such as
3760<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
3761Only one type, the return type, is overloaded, and only one type suffix is
3762required. Because the argument's type is matched against the return type, it
3763does not require its own name suffix.</p>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]3764
3765<p>To learn how to add an intrinsic function, please see the
3766<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3767</p>
3768
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3769</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3770
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3771<!-- ======================================================================= -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3772<div class="doc_subsection">
3773 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
3774</div>
3775
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3776<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3777
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3778<p>Variable argument support is defined in LLVM with the <a
Chris Lattnerfb6977d2006-01-13 23:26:01 +0000[diff] [blame]3779 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3780intrinsic functions. These functions are related to the similarly
3781named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3782
Chris Lattner261efe92003-11-25 01:02:51 +0000[diff] [blame]3783<p>All of these functions operate on arguments that use a
3784target-specific value type "<tt>va_list</tt>". The LLVM assembly
3785language reference manual does not define what this type is, so all
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3786transformations should be prepared to handle these functions regardless of
3787the type used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3788
Chris Lattner374ab302006-05-15 17:26:46 +0000[diff] [blame]3789<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]3790instruction and the variable argument handling intrinsic functions are
3791used.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3792
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]3793<div class="doc_code">
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3794<pre>
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]3795define i32 @test(i32 %X, ...) {
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3796 ; Initialize variable argument processing
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3797 %ap = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3798 %ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]3799 call void @llvm.va_start(i8* %ap2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3800
3801 ; Read a single integer argument
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3802 %tmp = va_arg i8** %ap, i32
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3803
3804 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3805 %aq = alloca i8*
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3806 %aq2 = bitcast i8** %aq to i8*
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3807 call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]3808 call void @llvm.va_end(i8* %aq2)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3809
3810 ; Stop processing of arguments.
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]3811 call void @llvm.va_end(i8* %ap2)
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]3812 ret i32 %tmp
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3813}
Anton Korobeynikov5d522f32007-03-21 23:58:04 +0000[diff] [blame]3814
3815declare void @llvm.va_start(i8*)
3816declare void @llvm.va_copy(i8*, i8*)
3817declare void @llvm.va_end(i8*)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3818</pre>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3819</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3820
Bill Wendling2f7a8b02007-05-29 09:04:49 +0000[diff] [blame]3821</div>
3822
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3823<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3824<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3825 <a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3826</div>
3827
3828
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3829<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3830<h5>Syntax:</h5>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3831<pre> declare void %llvm.va_start(i8* &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3832<h5>Overview:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3833<P>The '<tt>llvm.va_start</tt>' intrinsic initializes
3834<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
3835href="#i_va_arg">va_arg</a></tt>.</p>
3836
3837<h5>Arguments:</h5>
3838
3839<P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
3840
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3841<h5>Semantics:</h5>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3842
3843<P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
3844macro available in C. In a target-dependent way, it initializes the
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3845<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]3846<tt>va_arg</tt> will produce the first variable argument passed to the function.
3847Unlike 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]3848last argument of the function as the compiler can figure that out.</p>
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3849
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3850</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3851
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3852<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3853<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3854 <a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3855</div>
3856
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3857<div class="doc_text">
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3858<h5>Syntax:</h5>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]3859<pre> declare void @llvm.va_end(i8* &lt;arglist&gt;)<br></pre>
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3860<h5>Overview:</h5>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3861
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3862<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*&lt;arglist&gt;</tt>,
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3863which 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]3864or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3865
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3866<h5>Arguments:</h5>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3867
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3868<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]3869
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3870<h5>Semantics:</h5>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3871
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3872<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]3873macro available in C. In a target-dependent way, it destroys the
3874<tt>va_list</tt> element to which the argument points. Calls to <a
3875href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
3876<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
3877<tt>llvm.va_end</tt>.</p>
Chris Lattnerb75137d2007-01-08 07:55:15 +0000[diff] [blame]3878
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3879</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3880
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3881<!-- _______________________________________________________________________ -->
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3882<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3883 <a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3884</div>
3885
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3886<div class="doc_text">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3887
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3888<h5>Syntax:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3889
3890<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]3891 declare void @llvm.va_copy(i8* &lt;destarglist&gt;, i8* &lt;srcarglist&gt;)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3892</pre>
3893
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3894<h5>Overview:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3895
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3896<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
3897from the source argument list to the destination argument list.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3898
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3899<h5>Arguments:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3900
Andrew Lenharth8bf607a2005-06-18 18:28:17 +0000[diff] [blame]3901<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]3902The 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]3903
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3904
Chris Lattnerd9ad5b32003-05-08 04:57:36 +0000[diff] [blame]3905<h5>Semantics:</h5>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3906
Jeff Cohenb627eab2007-04-29 01:07:00 +0000[diff] [blame]3907<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
3908macro available in C. In a target-dependent way, it copies the source
3909<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
3910intrinsic is necessary because the <tt><a href="#int_va_start">
3911llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
3912example, memory allocation.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3913
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]3914</div>
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]3915
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]3916<!-- ======================================================================= -->
3917<div class="doc_subsection">
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3918 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
3919</div>
3920
3921<div class="doc_text">
3922
3923<p>
3924LLVM support for <a href="GarbageCollection.html">Accurate Garbage
3925Collection</a> requires the implementation and generation of these intrinsics.
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3926These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3927stack</a>, as well as garbage collector implementations that require <a
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3928href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3929Front-ends for type-safe garbage collected languages should generate these
3930intrinsics to make use of the LLVM garbage collectors. For more details, see <a
3931href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
3932</p>
3933</div>
3934
3935<!-- _______________________________________________________________________ -->
3936<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3937 <a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3938</div>
3939
3940<div class="doc_text">
3941
3942<h5>Syntax:</h5>
3943
3944<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]3945 declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3946</pre>
3947
3948<h5>Overview:</h5>
3949
John Criswell9e2485c2004-12-10 15:51:16 +0000[diff] [blame]3950<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]3951the code generator, and allows some metadata to be associated with it.</p>
3952
3953<h5>Arguments:</h5>
3954
3955<p>The first argument specifies the address of a stack object that contains the
3956root pointer. The second pointer (which must be either a constant or a global
3957value address) contains the meta-data to be associated with the root.</p>
3958
3959<h5>Semantics:</h5>
3960
3961<p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
3962location. At compile-time, the code generator generates information to allow
3963the runtime to find the pointer at GC safe points.
3964</p>
3965
3966</div>
3967
3968
3969<!-- _______________________________________________________________________ -->
3970<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]3971 <a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3972</div>
3973
3974<div class="doc_text">
3975
3976<h5>Syntax:</h5>
3977
3978<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]3979 declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3980</pre>
3981
3982<h5>Overview:</h5>
3983
3984<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
3985locations, allowing garbage collector implementations that require read
3986barriers.</p>
3987
3988<h5>Arguments:</h5>
3989
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]3990<p>The second argument is the address to read from, which should be an address
3991allocated from the garbage collector. The first object is a pointer to the
3992start of the referenced object, if needed by the language runtime (otherwise
3993null).</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]3994
3995<h5>Semantics:</h5>
3996
3997<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
3998instruction, but may be replaced with substantially more complex code by the
3999garbage collector runtime, as needed.</p>
4000
4001</div>
4002
4003
4004<!-- _______________________________________________________________________ -->
4005<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4006 <a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]4007</div>
4008
4009<div class="doc_text">
4010
4011<h5>Syntax:</h5>
4012
4013<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]4014 declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]4015</pre>
4016
4017<h5>Overview:</h5>
4018
4019<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
4020locations, allowing garbage collector implementations that require write
4021barriers (such as generational or reference counting collectors).</p>
4022
4023<h5>Arguments:</h5>
4024
Chris Lattner80626e92006-03-14 20:02:51 +0000[diff] [blame]4025<p>The first argument is the reference to store, the second is the start of the
4026object to store it to, and the third is the address of the field of Obj to
4027store to. If the runtime does not require a pointer to the object, Obj may be
4028null.</p>
Chris Lattnerd7923912004-05-23 21:06:01 +0000[diff] [blame]4029
4030<h5>Semantics:</h5>
4031
4032<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
4033instruction, but may be replaced with substantially more complex code by the
4034garbage collector runtime, as needed.</p>
4035
4036</div>
4037
4038
4039
4040<!-- ======================================================================= -->
4041<div class="doc_subsection">
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4042 <a name="int_codegen">Code Generator Intrinsics</a>
4043</div>
4044
4045<div class="doc_text">
4046<p>
4047These intrinsics are provided by LLVM to expose special features that may only
4048be implemented with code generator support.
4049</p>
4050
4051</div>
4052
4053<!-- _______________________________________________________________________ -->
4054<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4055 <a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4056</div>
4057
4058<div class="doc_text">
4059
4060<h5>Syntax:</h5>
4061<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4062 declare i8 *@llvm.returnaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4063</pre>
4064
4065<h5>Overview:</h5>
4066
4067<p>
Chris Lattner32b5d712006-10-15 20:05:59 +0000[diff] [blame]4068The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
4069target-specific value indicating the return address of the current function
4070or one of its callers.
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4071</p>
4072
4073<h5>Arguments:</h5>
4074
4075<p>
4076The argument to this intrinsic indicates which function to return the address
4077for. Zero indicates the calling function, one indicates its caller, etc. The
4078argument is <b>required</b> to be a constant integer value.
4079</p>
4080
4081<h5>Semantics:</h5>
4082
4083<p>
4084The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
4085the return address of the specified call frame, or zero if it cannot be
4086identified. The value returned by this intrinsic is likely to be incorrect or 0
4087for arguments other than zero, so it should only be used for debugging purposes.
4088</p>
4089
4090<p>
4091Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +0000[diff] [blame]4092aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4093source-language caller.
4094</p>
4095</div>
4096
4097
4098<!-- _______________________________________________________________________ -->
4099<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4100 <a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4101</div>
4102
4103<div class="doc_text">
4104
4105<h5>Syntax:</h5>
4106<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]4107 declare i8 *@llvm.frameaddress(i32 &lt;level&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4108</pre>
4109
4110<h5>Overview:</h5>
4111
4112<p>
Chris Lattner32b5d712006-10-15 20:05:59 +0000[diff] [blame]4113The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
4114target-specific frame pointer value for the specified stack frame.
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4115</p>
4116
4117<h5>Arguments:</h5>
4118
4119<p>
4120The argument to this intrinsic indicates which function to return the frame
4121pointer for. Zero indicates the calling function, one indicates its caller,
4122etc. The argument is <b>required</b> to be a constant integer value.
4123</p>
4124
4125<h5>Semantics:</h5>
4126
4127<p>
4128The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
4129the frame address of the specified call frame, or zero if it cannot be
4130identified. The value returned by this intrinsic is likely to be incorrect or 0
4131for arguments other than zero, so it should only be used for debugging purposes.
4132</p>
4133
4134<p>
4135Note that calling this intrinsic does not prevent function inlining or other
Chris Lattnerb40bb382005-03-07 20:30:51 +0000[diff] [blame]4136aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4137source-language caller.
4138</p>
4139</div>
4140
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]4141<!-- _______________________________________________________________________ -->
4142<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4143 <a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4144</div>
4145
4146<div class="doc_text">
4147
4148<h5>Syntax:</h5>
4149<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]4150 declare i8 *@llvm.stacksave()
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4151</pre>
4152
4153<h5>Overview:</h5>
4154
4155<p>
4156The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4157the function stack, for use with <a href="#int_stackrestore">
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4158<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
4159features like scoped automatic variable sized arrays in C99.
4160</p>
4161
4162<h5>Semantics:</h5>
4163
4164<p>
4165This intrinsic returns a opaque pointer value that can be passed to <a
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4166href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4167<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
4168<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
4169state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
4170practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
4171that were allocated after the <tt>llvm.stacksave</tt> was executed.
4172</p>
4173
4174</div>
4175
4176<!-- _______________________________________________________________________ -->
4177<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4178 <a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4179</div>
4180
4181<div class="doc_text">
4182
4183<h5>Syntax:</h5>
4184<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4185 declare void @llvm.stackrestore(i8 * %ptr)
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4186</pre>
4187
4188<h5>Overview:</h5>
4189
4190<p>
4191The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
4192the function stack to the state it was in when the corresponding <a
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4193href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4194useful for implementing language features like scoped automatic variable sized
4195arrays in C99.
4196</p>
4197
4198<h5>Semantics:</h5>
4199
4200<p>
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4201See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
Chris Lattner57e1f392006-01-13 02:03:13 +0000[diff] [blame]4202</p>
4203
4204</div>
4205
4206
4207<!-- _______________________________________________________________________ -->
4208<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4209 <a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]4210</div>
4211
4212<div class="doc_text">
4213
4214<h5>Syntax:</h5>
4215<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]4216 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]4217</pre>
4218
4219<h5>Overview:</h5>
4220
4221
4222<p>
4223The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]4224a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
4225no
4226effect on the behavior of the program but can change its performance
Chris Lattner2a615362005-02-28 19:47:14 +0000[diff] [blame]4227characteristics.
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]4228</p>
4229
4230<h5>Arguments:</h5>
4231
4232<p>
4233<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
4234determining if the fetch should be for a read (0) or write (1), and
4235<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Chris Lattneraeffb4a2005-03-07 20:31:38 +0000[diff] [blame]4236locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
Chris Lattner9a9d7ac2005-02-28 19:24:19 +0000[diff] [blame]4237<tt>locality</tt> arguments must be constant integers.
4238</p>
4239
4240<h5>Semantics:</h5>
4241
4242<p>
4243This intrinsic does not modify the behavior of the program. In particular,
4244prefetches cannot trap and do not produce a value. On targets that support this
4245intrinsic, the prefetch can provide hints to the processor cache for better
4246performance.
4247</p>
4248
4249</div>
4250
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]4251<!-- _______________________________________________________________________ -->
4252<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4253 <a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]4254</div>
4255
4256<div class="doc_text">
4257
4258<h5>Syntax:</h5>
4259<pre>
Chris Lattner1df4f752007-09-21 17:30:40 +0000[diff] [blame]4260 declare void @llvm.pcmarker(i32 &lt;id&gt;)
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]4261</pre>
4262
4263<h5>Overview:</h5>
4264
4265
4266<p>
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]4267The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
4268(PC) in a region of
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]4269code to simulators and other tools. The method is target specific, but it is
4270expected that the marker will use exported symbols to transmit the PC of the marker.
Jeff Cohen25d4f7e2005-11-11 02:15:27 +0000[diff] [blame]4271The marker makes no guarantees that it will remain with any specific instruction
Chris Lattnerd07c3f42005-11-15 06:07:55 +0000[diff] [blame]4272after optimizations. It is possible that the presence of a marker will inhibit
Chris Lattnerb3e7afd2006-03-24 07:16:10 +0000[diff] [blame]4273optimizations. The intended use is to be inserted after optimizations to allow
John Criswellfc6b8952005-05-16 16:17:45 +0000[diff] [blame]4274correlations of simulation runs.
Andrew Lenharth7f4ec3b2005-03-28 20:05:49 +0000[diff] [blame]4275</p>
4276
4277<h5>Arguments:</h5>
4278
4279<p>
4280<tt>id</tt> is a numerical id identifying the marker.
4281</p>
4282
4283<h5>Semantics:</h5>
4284
4285<p>
4286This intrinsic does not modify the behavior of the program. Backends that do not
4287support this intrinisic may ignore it.
4288</p>
4289
4290</div>
4291
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]4292<!-- _______________________________________________________________________ -->
4293<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4294 <a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]4295</div>
4296
4297<div class="doc_text">
4298
4299<h5>Syntax:</h5>
4300<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4301 declare i64 @llvm.readcyclecounter( )
Andrew Lenharth51b8d542005-11-11 16:47:30 +0000[diff] [blame]4302</pre>
4303
4304<h5>Overview:</h5>
4305
4306
4307<p>
4308The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
4309counter register (or similar low latency, high accuracy clocks) on those targets
4310that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
4311As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
4312should only be used for small timings.
4313</p>
4314
4315<h5>Semantics:</h5>
4316
4317<p>
4318When directly supported, reading the cycle counter should not modify any memory.
4319Implementations are allowed to either return a application specific value or a
4320system wide value. On backends without support, this is lowered to a constant 0.
4321</p>
4322
4323</div>
4324
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4325<!-- ======================================================================= -->
4326<div class="doc_subsection">
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4327 <a name="int_libc">Standard C Library Intrinsics</a>
4328</div>
4329
4330<div class="doc_text">
4331<p>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4332LLVM provides intrinsics for a few important standard C library functions.
4333These intrinsics allow source-language front-ends to pass information about the
4334alignment of the pointer arguments to the code generator, providing opportunity
4335for more efficient code generation.
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4336</p>
4337
4338</div>
4339
4340<!-- _______________________________________________________________________ -->
4341<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4342 <a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4343</div>
4344
4345<div class="doc_text">
4346
4347<h5>Syntax:</h5>
4348<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4349 declare void @llvm.memcpy.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4350 i32 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4351 declare void @llvm.memcpy.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4352 i64 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4353</pre>
4354
4355<h5>Overview:</h5>
4356
4357<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4358The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4359location to the destination location.
4360</p>
4361
4362<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4363Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
4364intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4365</p>
4366
4367<h5>Arguments:</h5>
4368
4369<p>
4370The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4371the source. The third argument is an integer argument
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4372specifying the number of bytes to copy, and the fourth argument is the alignment
4373of the source and destination locations.
4374</p>
4375
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]4376<p>
4377If 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]4378the caller guarantees that both the source and destination pointers are aligned
4379to that boundary.
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]4380</p>
4381
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4382<h5>Semantics:</h5>
4383
4384<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4385The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner33aec9e2004-02-12 17:01:32 +0000[diff] [blame]4386location to the destination location, which are not allowed to overlap. It
4387copies "len" bytes of memory over. If the argument is known to be aligned to
4388some boundary, this can be specified as the fourth argument, otherwise it should
4389be set to 0 or 1.
4390</p>
4391</div>
4392
4393
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4394<!-- _______________________________________________________________________ -->
4395<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4396 <a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4397</div>
4398
4399<div class="doc_text">
4400
4401<h5>Syntax:</h5>
4402<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4403 declare void @llvm.memmove.i32(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4404 i32 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4405 declare void @llvm.memmove.i64(i8 * &lt;dest&gt;, i8 * &lt;src&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4406 i64 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4407</pre>
4408
4409<h5>Overview:</h5>
4410
4411<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4412The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
4413location to the destination location. It is similar to the
4414'<tt>llvm.memcmp</tt>' intrinsic but allows the two memory locations to overlap.
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4415</p>
4416
4417<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4418Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
4419intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4420</p>
4421
4422<h5>Arguments:</h5>
4423
4424<p>
4425The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4426the source. The third argument is an integer argument
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4427specifying the number of bytes to copy, and the fourth argument is the alignment
4428of the source and destination locations.
4429</p>
4430
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]4431<p>
4432If 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]4433the caller guarantees that the source and destination pointers are aligned to
4434that boundary.
Chris Lattner3301ced2004-02-12 21:18:15 +0000[diff] [blame]4435</p>
4436
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4437<h5>Semantics:</h5>
4438
4439<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4440The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner0eb51b42004-02-12 18:10:10 +0000[diff] [blame]4441location to the destination location, which may overlap. It
4442copies "len" bytes of memory over. If the argument is known to be aligned to
4443some boundary, this can be specified as the fourth argument, otherwise it should
4444be set to 0 or 1.
4445</p>
4446</div>
4447
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]4448
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4449<!-- _______________________________________________________________________ -->
4450<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4451 <a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4452</div>
4453
4454<div class="doc_text">
4455
4456<h5>Syntax:</h5>
4457<pre>
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4458 declare void @llvm.memset.i32(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4459 i32 &lt;len&gt;, i32 &lt;align&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4460 declare void @llvm.memset.i64(i8 * &lt;dest&gt;, i8 &lt;val&gt;,
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4461 i64 &lt;len&gt;, i32 &lt;align&gt;)
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4462</pre>
4463
4464<h5>Overview:</h5>
4465
4466<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4467The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4468byte value.
4469</p>
4470
4471<p>
4472Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
4473does not return a value, and takes an extra alignment argument.
4474</p>
4475
4476<h5>Arguments:</h5>
4477
4478<p>
4479The first argument is a pointer to the destination to fill, the second is the
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4480byte value to fill it with, the third argument is an integer
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4481argument specifying the number of bytes to fill, and the fourth argument is the
4482known alignment of destination location.
4483</p>
4484
4485<p>
4486If 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]4487the caller guarantees that the destination pointer is aligned to that boundary.
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4488</p>
4489
4490<h5>Semantics:</h5>
4491
4492<p>
Chris Lattner5b310c32006-03-03 00:07:20 +0000[diff] [blame]4493The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
4494the
Chris Lattner10610642004-02-14 04:08:35 +0000[diff] [blame]4495destination location. If the argument is known to be aligned to some boundary,
4496this can be specified as the fourth argument, otherwise it should be set to 0 or
44971.
4498</p>
4499</div>
4500
4501
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]4502<!-- _______________________________________________________________________ -->
4503<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4504 <a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]4505</div>
4506
4507<div class="doc_text">
4508
4509<h5>Syntax:</h5>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]4510<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]4511floating point or vector of floating point type. Not all targets support all
4512types however.
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]4513<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]4514 declare float @llvm.sqrt.f32(float %Val)
4515 declare double @llvm.sqrt.f64(double %Val)
4516 declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
4517 declare fp128 @llvm.sqrt.f128(fp128 %Val)
4518 declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]4519</pre>
4520
4521<h5>Overview:</h5>
4522
4523<p>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]4524The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]4525returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]4526<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
4527negative numbers (which allows for better optimization).
4528</p>
4529
4530<h5>Arguments:</h5>
4531
4532<p>
4533The argument and return value are floating point numbers of the same type.
4534</p>
4535
4536<h5>Semantics:</h5>
4537
4538<p>
Dan Gohmand6257fe2007-07-16 14:37:41 +0000[diff] [blame]4539This function returns the sqrt of the specified operand if it is a nonnegative
Chris Lattnera4d74142005-07-21 01:29:16 +0000[diff] [blame]4540floating point number.
4541</p>
4542</div>
4543
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]4544<!-- _______________________________________________________________________ -->
4545<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4546 <a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]4547</div>
4548
4549<div class="doc_text">
4550
4551<h5>Syntax:</h5>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]4552<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]4553floating point or vector of floating point type. Not all targets support all
4554types however.
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]4555<pre>
Dale Johannesen408f9c12007-10-02 17:47:38 +0000[diff] [blame]4556 declare float @llvm.powi.f32(float %Val, i32 %power)
4557 declare double @llvm.powi.f64(double %Val, i32 %power)
4558 declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
4559 declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
4560 declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]4561</pre>
4562
4563<h5>Overview:</h5>
4564
4565<p>
4566The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
4567specified (positive or negative) power. The order of evaluation of
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]4568multiplications is not defined. When a vector of floating point type is
4569used, the second argument remains a scalar integer value.
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]4570</p>
4571
4572<h5>Arguments:</h5>
4573
4574<p>
4575The second argument is an integer power, and the first is a value to raise to
4576that power.
4577</p>
4578
4579<h5>Semantics:</h5>
4580
4581<p>
4582This function returns the first value raised to the second power with an
4583unspecified sequence of rounding operations.</p>
4584</div>
4585
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]4586<!-- _______________________________________________________________________ -->
4587<div class="doc_subsubsection">
4588 <a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
4589</div>
4590
4591<div class="doc_text">
4592
4593<h5>Syntax:</h5>
4594<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
4595floating point or vector of floating point type. Not all targets support all
4596types however.
4597<pre>
4598 declare float @llvm.sin.f32(float %Val)
4599 declare double @llvm.sin.f64(double %Val)
4600 declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
4601 declare fp128 @llvm.sin.f128(fp128 %Val)
4602 declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
4603</pre>
4604
4605<h5>Overview:</h5>
4606
4607<p>
4608The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
4609</p>
4610
4611<h5>Arguments:</h5>
4612
4613<p>
4614The argument and return value are floating point numbers of the same type.
4615</p>
4616
4617<h5>Semantics:</h5>
4618
4619<p>
4620This function returns the sine of the specified operand, returning the
4621same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohmanba83b7e2007-10-17 18:05:13 +0000[diff] [blame]4622conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]4623</div>
4624
4625<!-- _______________________________________________________________________ -->
4626<div class="doc_subsubsection">
4627 <a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
4628</div>
4629
4630<div class="doc_text">
4631
4632<h5>Syntax:</h5>
4633<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
4634floating point or vector of floating point type. Not all targets support all
4635types however.
4636<pre>
4637 declare float @llvm.cos.f32(float %Val)
4638 declare double @llvm.cos.f64(double %Val)
4639 declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
4640 declare fp128 @llvm.cos.f128(fp128 %Val)
4641 declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
4642</pre>
4643
4644<h5>Overview:</h5>
4645
4646<p>
4647The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
4648</p>
4649
4650<h5>Arguments:</h5>
4651
4652<p>
4653The argument and return value are floating point numbers of the same type.
4654</p>
4655
4656<h5>Semantics:</h5>
4657
4658<p>
4659This function returns the cosine of the specified operand, returning the
4660same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohmanba83b7e2007-10-17 18:05:13 +0000[diff] [blame]4661conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]4662</div>
4663
4664<!-- _______________________________________________________________________ -->
4665<div class="doc_subsubsection">
4666 <a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
4667</div>
4668
4669<div class="doc_text">
4670
4671<h5>Syntax:</h5>
4672<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
4673floating point or vector of floating point type. Not all targets support all
4674types however.
4675<pre>
4676 declare float @llvm.pow.f32(float %Val, float %Power)
4677 declare double @llvm.pow.f64(double %Val, double %Power)
4678 declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
4679 declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
4680 declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
4681</pre>
4682
4683<h5>Overview:</h5>
4684
4685<p>
4686The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
4687specified (positive or negative) power.
4688</p>
4689
4690<h5>Arguments:</h5>
4691
4692<p>
4693The second argument is a floating point power, and the first is a value to
4694raise to that power.
4695</p>
4696
4697<h5>Semantics:</h5>
4698
4699<p>
4700This function returns the first value raised to the second power,
4701returning the
4702same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohmanba83b7e2007-10-17 18:05:13 +0000[diff] [blame]4703conditions in the same way.</p>
Dan Gohman91c284c2007-10-15 20:30:11 +0000[diff] [blame]4704</div>
4705
Chris Lattnerf4d252d2006-09-08 06:34:02 +0000[diff] [blame]4706
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4707<!-- ======================================================================= -->
4708<div class="doc_subsection">
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]4709 <a name="int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4710</div>
4711
4712<div class="doc_text">
4713<p>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]4714LLVM provides intrinsics for a few important bit manipulation operations.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4715These allow efficient code generation for some algorithms.
4716</p>
4717
4718</div>
4719
4720<!-- _______________________________________________________________________ -->
4721<div class="doc_subsubsection">
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4722 <a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]4723</div>
4724
4725<div class="doc_text">
4726
4727<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]4728<p>This is an overloaded intrinsic function. You can use bswap on any integer
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4729type that is an even number of bytes (i.e. BitWidth % 16 == 0).
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]4730<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4731 declare i16 @llvm.bswap.i16(i16 &lt;id&gt;)
4732 declare i32 @llvm.bswap.i32(i32 &lt;id&gt;)
4733 declare i64 @llvm.bswap.i64(i64 &lt;id&gt;)
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]4734</pre>
4735
4736<h5>Overview:</h5>
4737
4738<p>
Reid Spencer338ea092007-04-02 02:25:19 +0000[diff] [blame]4739The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]4740values with an even number of bytes (positive multiple of 16 bits). These are
4741useful for performing operations on data that is not in the target's native
4742byte order.
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]4743</p>
4744
4745<h5>Semantics:</h5>
4746
4747<p>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4748The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Reid Spencerca86e162006-12-31 07:07:53 +0000[diff] [blame]4749and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
4750intrinsic returns an i32 value that has the four bytes of the input i32
4751swapped, 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]4752i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
4753<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]4754additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
Nate Begeman7e36c472006-01-13 23:26:38 +0000[diff] [blame]4755</p>
4756
4757</div>
4758
4759<!-- _______________________________________________________________________ -->
4760<div class="doc_subsubsection">
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]4761 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4762</div>
4763
4764<div class="doc_text">
4765
4766<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]4767<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
4768width. Not all targets support all bit widths however.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4769<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4770 declare i8 @llvm.ctpop.i8 (i8 &lt;src&gt;)
4771 declare i16 @llvm.ctpop.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4772 declare i32 @llvm.ctpop.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4773 declare i64 @llvm.ctpop.i64(i64 &lt;src&gt;)
4774 declare i256 @llvm.ctpop.i256(i256 &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4775</pre>
4776
4777<h5>Overview:</h5>
4778
4779<p>
Chris Lattnerec6cb612006-01-16 22:38:59 +0000[diff] [blame]4780The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
4781value.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4782</p>
4783
4784<h5>Arguments:</h5>
4785
4786<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]4787The only argument is the value to be counted. The argument may be of any
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]4788integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4789</p>
4790
4791<h5>Semantics:</h5>
4792
4793<p>
4794The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
4795</p>
4796</div>
4797
4798<!-- _______________________________________________________________________ -->
4799<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]4800 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4801</div>
4802
4803<div class="doc_text">
4804
4805<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]4806<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
4807integer bit width. Not all targets support all bit widths however.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4808<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4809 declare i8 @llvm.ctlz.i8 (i8 &lt;src&gt;)
4810 declare i16 @llvm.ctlz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4811 declare i32 @llvm.ctlz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4812 declare i64 @llvm.ctlz.i64(i64 &lt;src&gt;)
4813 declare i256 @llvm.ctlz.i256(i256 &lt;src&gt;)
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4814</pre>
4815
4816<h5>Overview:</h5>
4817
4818<p>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]4819The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
4820leading zeros in a variable.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4821</p>
4822
4823<h5>Arguments:</h5>
4824
4825<p>
Chris Lattnercfe6b372005-05-07 01:46:40 +0000[diff] [blame]4826The only argument is the value to be counted. The argument may be of any
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]4827integer type. The return type must match the argument type.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4828</p>
4829
4830<h5>Semantics:</h5>
4831
4832<p>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]4833The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
4834in 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]4835of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
Andrew Lenharthec370fd2005-05-03 18:01:48 +0000[diff] [blame]4836</p>
4837</div>
Chris Lattner32006282004-06-11 02:28:03 +0000[diff] [blame]4838
4839
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]4840
4841<!-- _______________________________________________________________________ -->
4842<div class="doc_subsubsection">
Chris Lattner8a886be2006-01-16 22:34:14 +0000[diff] [blame]4843 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]4844</div>
4845
4846<div class="doc_text">
4847
4848<h5>Syntax:</h5>
Reid Spencer409e28f2007-04-01 08:04:23 +0000[diff] [blame]4849<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
4850integer bit width. Not all targets support all bit widths however.
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]4851<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4852 declare i8 @llvm.cttz.i8 (i8 &lt;src&gt;)
4853 declare i16 @llvm.cttz.i16(i16 &lt;src&gt;)
Anton Korobeynikovec43a062007-03-22 00:02:17 +0000[diff] [blame]4854 declare i32 @llvm.cttz.i32(i32 &lt;src&gt;)
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4855 declare i64 @llvm.cttz.i64(i64 &lt;src&gt;)
4856 declare i256 @llvm.cttz.i256(i256 &lt;src&gt;)
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]4857</pre>
4858
4859<h5>Overview:</h5>
4860
4861<p>
Reid Spencer0b118202006-01-16 21:12:35 +0000[diff] [blame]4862The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
4863trailing zeros.
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]4864</p>
4865
4866<h5>Arguments:</h5>
4867
4868<p>
4869The only argument is the value to be counted. The argument may be of any
Reid Spencera5173382007-01-04 16:43:23 +0000[diff] [blame]4870integer type. The return type must match the argument type.
Chris Lattnereff29ab2005-05-15 19:39:26 +0000[diff] [blame]4871</p>
4872
4873<h5>Semantics:</h5>
4874
4875<p>
4876The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
4877in a variable. If the src == 0 then the result is the size in bits of the type
4878of src. For example, <tt>llvm.cttz(2) = 1</tt>.
4879</p>
4880</div>
4881
Reid Spencer497d93e2007-04-01 08:27:01 +0000[diff] [blame]4882<!-- _______________________________________________________________________ -->
4883<div class="doc_subsubsection">
Reid Spencerbeacf662007-04-10 02:51:31 +0000[diff] [blame]4884 <a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
Reid Spencera13ba7d2007-04-01 19:00:37 +0000[diff] [blame]4885</div>
4886
4887<div class="doc_text">
4888
4889<h5>Syntax:</h5>
Reid Spencerbeacf662007-04-10 02:51:31 +0000[diff] [blame]4890<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
Reid Spencera13ba7d2007-04-01 19:00:37 +0000[diff] [blame]4891on any integer bit width.
4892<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4893 declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
4894 declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Reid Spencera13ba7d2007-04-01 19:00:37 +0000[diff] [blame]4895</pre>
4896
4897<h5>Overview:</h5>
Reid Spencerbeacf662007-04-10 02:51:31 +0000[diff] [blame]4898<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
Reid Spencera13ba7d2007-04-01 19:00:37 +0000[diff] [blame]4899range of bits from an integer value and returns them in the same bit width as
4900the original value.</p>
4901
4902<h5>Arguments:</h5>
4903<p>The first argument, <tt>%val</tt> and the result may be integer types of
4904any bit width but they must have the same bit width. The second and third
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4905arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
Reid Spencera13ba7d2007-04-01 19:00:37 +0000[diff] [blame]4906
4907<h5>Semantics:</h5>
Reid Spencerbeacf662007-04-10 02:51:31 +0000[diff] [blame]4908<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
Reid Spencera3e435f2007-04-04 02:42:35 +0000[diff] [blame]4909of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
4910<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
4911operates in forward mode.</p>
4912<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
4913right by <tt>%loBit</tt> bits and then ANDing it with a mask with
Reid Spencera13ba7d2007-04-01 19:00:37 +0000[diff] [blame]4914only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
4915<ol>
4916 <li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
4917 by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
4918 <li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
4919 to determine the number of bits to retain.</li>
4920 <li>A mask of the retained bits is created by shifting a -1 value.</li>
4921 <li>The mask is ANDed with <tt>%val</tt> to produce the result.
4922</ol>
Reid Spencerd6a85b52007-05-14 16:14:57 +0000[diff] [blame]4923<p>In reverse mode, a similar computation is made except that the bits are
4924returned in the reverse order. So, for example, if <tt>X</tt> has the value
4925<tt>i16 0x0ACF (101011001111)</tt> and we apply
4926<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
4927<tt>i16 0x0026 (000000100110)</tt>.</p>
Reid Spencera13ba7d2007-04-01 19:00:37 +0000[diff] [blame]4928</div>
4929
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]4930<div class="doc_subsubsection">
4931 <a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
4932</div>
4933
4934<div class="doc_text">
4935
4936<h5>Syntax:</h5>
4937<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
4938on any integer bit width.
4939<pre>
Chandler Carruth69940402007-08-04 01:51:18 +0000[diff] [blame]4940 declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
4941 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]4942</pre>
4943
4944<h5>Overview:</h5>
4945<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
4946of bits in an integer value with another integer value. It returns the integer
4947with the replaced bits.</p>
4948
4949<h5>Arguments:</h5>
4950<p>The first argument, <tt>%val</tt> and the result may be integer types of
4951any bit width but they must have the same bit width. <tt>%val</tt> is the value
4952whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
4953integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
4954type since they specify only a bit index.</p>
4955
4956<h5>Semantics:</h5>
4957<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
4958of operation: forwards and reverse. If <tt>%lo</tt> is greater than
4959<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
4960operates in forward mode.</p>
4961<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
4962truncating it down to the size of the replacement area or zero extending it
4963up to that size.</p>
4964<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
4965are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
4966in <tt>%repl</tt> replaces the <tt>%lo</tt>th bit in <tt>%val</tt> and etc. up
4967to the <tt>%hi</tt>th bit.
Reid Spencerc6749c42007-05-14 16:50:20 +0000[diff] [blame]4968<p>In reverse mode, a similar computation is made except that the bits are
4969reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
4970<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]4971<h5>Examples:</h5>
4972<pre>
Reid Spencerf0dbf642007-04-12 01:03:03 +0000[diff] [blame]4973 llvm.part.set(0xFFFF, 0, 4, 7) -&gt; 0xFF0F
Reid Spencerc6749c42007-05-14 16:50:20 +0000[diff] [blame]4974 llvm.part.set(0xFFFF, 0, 7, 4) -&gt; 0xFF0F
4975 llvm.part.set(0xFFFF, 1, 7, 4) -&gt; 0xFF8F
4976 llvm.part.set(0xFFFF, F, 8, 3) -&gt; 0xFFE7
Reid Spencerf0dbf642007-04-12 01:03:03 +0000[diff] [blame]4977 llvm.part.set(0xFFFF, 0, 3, 8) -&gt; 0xFE07
Reid Spencerc8910842007-04-11 23:49:50 +0000[diff] [blame]4978</pre>
Reid Spencerf86037f2007-04-11 23:23:49 +0000[diff] [blame]4979</div>
4980
Chris Lattner8ff75902004-01-06 05:31:32 +0000[diff] [blame]4981<!-- ======================================================================= -->
4982<div class="doc_subsection">
4983 <a name="int_debugger">Debugger Intrinsics</a>
4984</div>
4985
4986<div class="doc_text">
4987<p>
4988The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
4989are described in the <a
4990href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
4991Debugging</a> document.
4992</p>
4993</div>
4994
4995
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]4996<!-- ======================================================================= -->
4997<div class="doc_subsection">
4998 <a name="int_eh">Exception Handling Intrinsics</a>
4999</div>
5000
5001<div class="doc_text">
5002<p> The LLVM exception handling intrinsics (which all start with
5003<tt>llvm.eh.</tt> prefix), are described in the <a
5004href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
5005Handling</a> document. </p>
5006</div>
5007
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]5008<!-- ======================================================================= -->
5009<div class="doc_subsection">
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]5010 <a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]5011</div>
5012
5013<div class="doc_text">
5014<p>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]5015 This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]5016 the <tt>nest</tt> attribute, from a function. The result is a callable
5017 function pointer lacking the nest parameter - the caller does not need
5018 to provide a value for it. Instead, the value to use is stored in
5019 advance in a "trampoline", a block of memory usually allocated
5020 on the stack, which also contains code to splice the nest value into the
5021 argument list. This is used to implement the GCC nested function address
5022 extension.
5023</p>
5024<p>
5025 For example, if the function is
5026 <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]5027 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]5028<pre>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]5029 %tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
5030 %tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
5031 %p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8*), i8* %nval )
5032 %fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]5033</pre>
Bill Wendling03295ca2007-09-22 09:23:55 +0000[diff] [blame]5034 <p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
5035 to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]5036</div>
5037
5038<!-- _______________________________________________________________________ -->
5039<div class="doc_subsubsection">
5040 <a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
5041</div>
5042<div class="doc_text">
5043<h5>Syntax:</h5>
5044<pre>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]5045declare 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]5046</pre>
5047<h5>Overview:</h5>
5048<p>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]5049 This fills the memory pointed to by <tt>tramp</tt> with code
5050 and returns a function pointer suitable for executing it.
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]5051</p>
5052<h5>Arguments:</h5>
5053<p>
5054 The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
5055 pointers. The <tt>tramp</tt> argument must point to a sufficiently large
5056 and sufficiently aligned block of memory; this memory is written to by the
Duncan Sandsc00c2ba2007-08-22 23:39:54 +0000[diff] [blame]5057 intrinsic. Note that the size and the alignment are target-specific - LLVM
5058 currently provides no portable way of determining them, so a front-end that
5059 generates this intrinsic needs to have some target-specific knowledge.
5060 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]5061</p>
5062<h5>Semantics:</h5>
5063<p>
5064 The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]5065 dependent code, turning it into a function. A pointer to this function is
5066 returned, but needs to be bitcast to an
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]5067 <a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sandsf7331b32007-09-11 14:10:23 +0000[diff] [blame]5068 before being called. The new function's signature is the same as that of
5069 <tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
5070 removed. At most one such <tt>nest</tt> argument is allowed, and it must be
5071 of pointer type. Calling the new function is equivalent to calling
5072 <tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
5073 missing <tt>nest</tt> argument. If, after calling
5074 <tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
5075 modified, then the effect of any later call to the returned function pointer is
5076 undefined.
Duncan Sands36397f52007-07-27 12:58:54 +0000[diff] [blame]5077</p>
5078</div>
5079
5080<!-- ======================================================================= -->
5081<div class="doc_subsection">
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]5082 <a name="int_general">General Intrinsics</a>
5083</div>
5084
5085<div class="doc_text">
5086<p> This class of intrinsics is designed to be generic and has
5087no specific purpose. </p>
5088</div>
5089
5090<!-- _______________________________________________________________________ -->
5091<div class="doc_subsubsection">
5092 <a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
5093</div>
5094
5095<div class="doc_text">
5096
5097<h5>Syntax:</h5>
5098<pre>
Tanya Lattnerd2e84422007-06-18 23:42:37 +0000[diff] [blame]5099 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]5100</pre>
5101
5102<h5>Overview:</h5>
5103
5104<p>
5105The '<tt>llvm.var.annotation</tt>' intrinsic
5106</p>
5107
5108<h5>Arguments:</h5>
5109
5110<p>
Tanya Lattnerd2e84422007-06-18 23:42:37 +0000[diff] [blame]5111The first argument is a pointer to a value, the second is a pointer to a
5112global string, the third is a pointer to a global string which is the source
5113file name, and the last argument is the line number.
Tanya Lattner6d806e92007-06-15 20:50:54 +0000[diff] [blame]5114</p>
5115
5116<h5>Semantics:</h5>
5117
5118<p>
5119This intrinsic allows annotation of local variables with arbitrary strings.
5120This can be useful for special purpose optimizations that want to look for these
5121 annotations. These have no other defined use, they are ignored by code
5122 generation and optimization.
5123</div>
5124
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]5125<!-- _______________________________________________________________________ -->
5126<div class="doc_subsubsection">
Tanya Lattnere1a8da02007-09-21 23:57:59 +0000[diff] [blame]5127 <a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]5128</div>
5129
5130<div class="doc_text">
5131
5132<h5>Syntax:</h5>
Tanya Lattner39cfba62007-09-21 23:56:27 +0000[diff] [blame]5133<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
5134any integer bit width.
5135</p>
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]5136<pre>
Tanya Lattnerd3989a82007-09-22 00:03:01 +0000[diff] [blame]5137 declare i8 @llvm.annotation.i8(i8 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5138 declare i16 @llvm.annotation.i16(i16 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5139 declare i32 @llvm.annotation.i32(i32 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5140 declare i64 @llvm.annotation.i64(i64 &lt;val&gt;, i8* &lt;str&gt;, i8* &lt;str&gt;, i32 &lt;int&gt; )
5141 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]5142</pre>
5143
5144<h5>Overview:</h5>
Tanya Lattner39cfba62007-09-21 23:56:27 +0000[diff] [blame]5145
5146<p>
5147The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]5148</p>
5149
5150<h5>Arguments:</h5>
5151
5152<p>
5153The first argument is an integer value (result of some expression),
5154the second is a pointer to a global string, the third is a pointer to a global
5155string which is the source file name, and the last argument is the line number.
Tanya Lattner39cfba62007-09-21 23:56:27 +0000[diff] [blame]5156It returns the value of the first argument.
Tanya Lattnerb6367882007-09-21 22:59:12 +0000[diff] [blame]5157</p>
5158
5159<h5>Semantics:</h5>
5160
5161<p>
5162This intrinsic allows annotations to be put on arbitrary expressions
5163with arbitrary strings. This can be useful for special purpose optimizations
5164that want to look for these annotations. These have no other defined use, they
5165are ignored by code generation and optimization.
5166</div>
Jim Laskeydd4ef1b2007-03-14 19:31:19 +0000[diff] [blame]5167
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5168<!-- *********************************************************************** -->
Chris Lattner00950542001-06-06 20:29:01 +0000[diff] [blame]5169<hr>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]5170<address>
5171 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
5172 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
5173 <a href="http://validator.w3.org/check/referer"><img
5174 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
5175
5176 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer05fe4b02006-03-14 05:39:39 +0000[diff] [blame]5177 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukmandaa4cb02004-03-01 17:47:27 +0000[diff] [blame]5178 Last modified: $Date$
5179</address>
Misha Brukman9d0919f2003-11-08 01:05:38 +0000[diff] [blame]5180</body>
5181</html>