Gitiles
Code Review Sign In
gerrit-public.fairphone.software / toolchain / llvm-project / 8df88e2cc9bf474123cc7ee37bc57a040efd3497 / . / llvm / docs / LangRef.html
blob: 35585f252b8a156dc75f1a512515710aea7b1f8f [file] [log] [blame]
Misha Brukmanc501f552004-03-01 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 Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3<html>
4<head>
5 <title>LLVM Assembly Language Reference Manual</title>
Reid Spencercb84e432004-08-26 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 Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]10 <link rel="stylesheet" href="llvm.css" type="text/css">
11</head>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]12
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]13<body>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]14
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]15<div class="doc_title"> LLVM Language Reference Manual </div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]16<ol>
Misha Brukman76307852003-11-08 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 Lattner6af02f32004-12-09 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 Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]23 <li><a href="#linkage">Linkage Types</a></li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]24 <li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]25 <li><a href="#globalvars">Global Variables</a></li>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]26 <li><a href="#functionstructure">Functions</a></li>
27 <li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]28 </ol>
29 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]30 <li><a href="#typesystem">Type System</a>
31 <ol>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000[diff] [blame]32 <li><a href="#t_primitive">Primitive Types</a>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]33 <ol>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]34 <li><a href="#t_classifications">Type Classifications</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]35 </ol>
36 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]37 <li><a href="#t_derived">Derived Types</a>
38 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]39 <li><a href="#t_array">Array Type</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]40 <li><a href="#t_function">Function Type</a></li>
41 <li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]42 <li><a href="#t_struct">Structure Type</a></li>
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame^]43 <li><a href="#t_pstruct">Packed Structure Type</a></li>
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]44 <li><a href="#t_packed">Packed Type</a></li>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]45 <li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]46 </ol>
47 </li>
48 </ol>
49 </li>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]50 <li><a href="#constants">Constants</a>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]51 <ol>
52 <li><a href="#simpleconstants">Simple Constants</a>
53 <li><a href="#aggregateconstants">Aggregate Constants</a>
54 <li><a href="#globalconstants">Global Variable and Function Addresses</a>
55 <li><a href="#undefvalues">Undefined Values</a>
56 <li><a href="#constantexprs">Constant Expressions</a>
57 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]58 </li>
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]59 <li><a href="#othervalues">Other Values</a>
60 <ol>
61 <li><a href="#inlineasm">Inline Assembler Expressions</a>
62 </ol>
63 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]64 <li><a href="#instref">Instruction Reference</a>
65 <ol>
66 <li><a href="#terminators">Terminator Instructions</a>
67 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]68 <li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
69 <li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]70 <li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
71 <li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]72 <li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]73 <li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]74 </ol>
75 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]76 <li><a href="#binaryops">Binary Operations</a>
77 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]78 <li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
79 <li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
80 <li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]81 <li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
82 <li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
83 <li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]84 <li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
85 <li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
86 <li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]87 </ol>
88 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]89 <li><a href="#bitwiseops">Bitwise Binary Operations</a>
90 <ol>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]91 <li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]92 <li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]93 <li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
94 <li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
Reid Spencerfdff9382006-11-08 06:47:33 +0000[diff] [blame]95 <li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
96 <li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]97 </ol>
98 </li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]99 <li><a href="#vectorops">Vector Operations</a>
100 <ol>
101 <li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
102 <li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
103 <li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]104 </ol>
105 </li>
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]106 <li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]107 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]108 <li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
109 <li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
110 <li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000[diff] [blame]111 <li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
112 <li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
113 <li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]114 </ol>
115 </li>
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]116 <li><a href="#convertops">Conversion Operations</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]117 <ol>
118 <li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
119 <li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
120 <li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
121 <li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
122 <li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]123 <li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
124 <li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
125 <li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
126 <li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]127 <li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
128 <li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]129 <li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]130 </ol>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]131 <li><a href="#otherops">Other Operations</a>
132 <ol>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]133 <li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
134 <li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]135 <li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]136 <li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]137 <li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattner33337472006-01-13 23:26:01 +0000[diff] [blame]138 <li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]139 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]140 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]141 </ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]142 </li>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]143 <li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]144 <ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]145 <li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
146 <ol>
147 <li><a href="#i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
148 <li><a href="#i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
149 <li><a href="#i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
150 </ol>
151 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]152 <li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
153 <ol>
154 <li><a href="#i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
155 <li><a href="#i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
156 <li><a href="#i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
157 </ol>
158 </li>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]159 <li><a href="#int_codegen">Code Generator Intrinsics</a>
160 <ol>
161 <li><a href="#i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
162 <li><a href="#i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]163 <li><a href="#i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
164 <li><a href="#i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]165 <li><a href="#i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]166 <li><a href="#i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]167 <li><a href="#i_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswellaa1c3c12004-04-09 16:43:20 +0000[diff] [blame]168 </ol>
169 </li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]170 <li><a href="#int_libc">Standard C Library Intrinsics</a>
171 <ol>
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]172 <li><a href="#i_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
173 <li><a href="#i_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
174 <li><a href="#i_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
Chris Lattner069b5bd2006-01-16 22:38:59 +0000[diff] [blame]175 <li><a href="#i_isunordered">'<tt>llvm.isunordered.*</tt>' Intrinsic</a></li>
176 <li><a href="#i_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]177 <li><a href="#i_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]178 </ol>
179 </li>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]180 <li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]181 <ol>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]182 <li><a href="#i_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]183 <li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
184 <li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
185 <li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]186 </ol>
187 </li>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]188 <li><a href="#int_debugger">Debugger intrinsics</a></li>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]189 </ol>
190 </li>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]191</ol>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]192
193<div class="doc_author">
194 <p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
195 and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]196</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]197
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]198<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]199<div class="doc_section"> <a name="abstract">Abstract </a></div>
200<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]201
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]202<div class="doc_text">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]203<p>This document is a reference manual for the LLVM assembly language.
204LLVM is an SSA based representation that provides type safety,
205low-level operations, flexibility, and the capability of representing
206'all' high-level languages cleanly. It is the common code
207representation used throughout all phases of the LLVM compilation
208strategy.</p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]209</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]210
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]211<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]212<div class="doc_section"> <a name="introduction">Introduction</a> </div>
213<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]214
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]215<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]216
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]217<p>The LLVM code representation is designed to be used in three
218different forms: as an in-memory compiler IR, as an on-disk bytecode
219representation (suitable for fast loading by a Just-In-Time compiler),
220and as a human readable assembly language representation. This allows
221LLVM to provide a powerful intermediate representation for efficient
222compiler transformations and analysis, while providing a natural means
223to debug and visualize the transformations. The three different forms
224of LLVM are all equivalent. This document describes the human readable
225representation and notation.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]226
John Criswell4a3327e2005-05-13 22:25:59 +0000[diff] [blame]227<p>The LLVM representation aims to be light-weight and low-level
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]228while being expressive, typed, and extensible at the same time. It
229aims to be a "universal IR" of sorts, by being at a low enough level
230that high-level ideas may be cleanly mapped to it (similar to how
231microprocessors are "universal IR's", allowing many source languages to
232be mapped to them). By providing type information, LLVM can be used as
233the target of optimizations: for example, through pointer analysis, it
234can be proven that a C automatic variable is never accessed outside of
235the current function... allowing it to be promoted to a simple SSA
236value instead of a memory location.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]237
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]238</div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]239
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]240<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]241<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]242
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]243<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]244
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]245<p>It is important to note that this document describes 'well formed'
246LLVM assembly language. There is a difference between what the parser
247accepts and what is considered 'well formed'. For example, the
248following instruction is syntactically okay, but not well formed:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]249
250<pre>
251 %x = <a href="#i_add">add</a> int 1, %x
252</pre>
253
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]254<p>...because the definition of <tt>%x</tt> does not dominate all of
255its uses. The LLVM infrastructure provides a verification pass that may
256be used to verify that an LLVM module is well formed. This pass is
John Criswell4a3327e2005-05-13 22:25:59 +0000[diff] [blame]257automatically run by the parser after parsing input assembly and by
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]258the optimizer before it outputs bytecode. The violations pointed out
259by the verifier pass indicate bugs in transformation passes or input to
260the parser.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]261
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]262<!-- Describe the typesetting conventions here. --> </div>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]263
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]264<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]265<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]266<!-- *********************************************************************** -->
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]267
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]268<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]269
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]270<p>LLVM uses three different forms of identifiers, for different
271purposes:</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]272
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]273<ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]274 <li>Named values are represented as a string of characters with a '%' prefix.
275 For example, %foo, %DivisionByZero, %a.really.long.identifier. The actual
276 regular expression used is '<tt>%[a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
277 Identifiers which require other characters in their names can be surrounded
278 with quotes. In this way, anything except a <tt>"</tt> character can be used
279 in a name.</li>
280
281 <li>Unnamed values are represented as an unsigned numeric value with a '%'
282 prefix. For example, %12, %2, %44.</li>
283
Reid Spencer8f08d802004-12-09 18:02:53 +0000[diff] [blame]284 <li>Constants, which are described in a <a href="#constants">section about
285 constants</a>, below.</li>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]286</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]287
288<p>LLVM requires that values start with a '%' sign for two reasons: Compilers
289don't need to worry about name clashes with reserved words, and the set of
290reserved words may be expanded in the future without penalty. Additionally,
291unnamed identifiers allow a compiler to quickly come up with a temporary
292variable without having to avoid symbol table conflicts.</p>
293
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]294<p>Reserved words in LLVM are very similar to reserved words in other
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]295languages. There are keywords for different opcodes
296('<tt><a href="#i_add">add</a></tt>',
297 '<tt><a href="#i_bitcast">bitcast</a></tt>',
298 '<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]299href="#t_void">void</a></tt>', '<tt><a href="#t_uint">uint</a></tt>', etc...),
300and others. These reserved words cannot conflict with variable names, because
301none of them start with a '%' character.</p>
302
303<p>Here is an example of LLVM code to multiply the integer variable
304'<tt>%X</tt>' by 8:</p>
305
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]306<p>The easy way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]307
308<pre>
309 %result = <a href="#i_mul">mul</a> uint %X, 8
310</pre>
311
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]312<p>After strength reduction:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]313
314<pre>
315 %result = <a href="#i_shl">shl</a> uint %X, ubyte 3
316</pre>
317
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]318<p>And the hard way:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]319
320<pre>
321 <a href="#i_add">add</a> uint %X, %X <i>; yields {uint}:%0</i>
322 <a href="#i_add">add</a> uint %0, %0 <i>; yields {uint}:%1</i>
323 %result = <a href="#i_add">add</a> uint %1, %1
324</pre>
325
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]326<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
327important lexical features of LLVM:</p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]328
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]329<ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]330
331 <li>Comments are delimited with a '<tt>;</tt>' and go until the end of
332 line.</li>
333
334 <li>Unnamed temporaries are created when the result of a computation is not
335 assigned to a named value.</li>
336
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]337 <li>Unnamed temporaries are numbered sequentially</li>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]338
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]339</ol>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]340
John Criswell02fdc6f2005-05-12 16:52:32 +0000[diff] [blame]341<p>...and it also shows a convention that we follow in this document. When
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]342demonstrating instructions, we will follow an instruction with a comment that
343defines the type and name of value produced. Comments are shown in italic
344text.</p>
345
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]346</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]347
348<!-- *********************************************************************** -->
349<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
350<!-- *********************************************************************** -->
351
352<!-- ======================================================================= -->
353<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
354</div>
355
356<div class="doc_text">
357
358<p>LLVM programs are composed of "Module"s, each of which is a
359translation unit of the input programs. Each module consists of
360functions, global variables, and symbol table entries. Modules may be
361combined together with the LLVM linker, which merges function (and
362global variable) definitions, resolves forward declarations, and merges
363symbol table entries. Here is an example of the "hello world" module:</p>
364
365<pre><i>; Declare the string constant as a global constant...</i>
366<a href="#identifiers">%.LC0</a> = <a href="#linkage_internal">internal</a> <a
367 href="#globalvars">constant</a> <a href="#t_array">[13 x sbyte]</a> c"hello world\0A\00" <i>; [13 x sbyte]*</i>
368
369<i>; External declaration of the puts function</i>
370<a href="#functionstructure">declare</a> int %puts(sbyte*) <i>; int(sbyte*)* </i>
371
Chris Lattnerd2d29a02006-06-13 03:05:47 +0000[diff] [blame]372<i>; Global variable / Function body section separator</i>
373implementation
374
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]375<i>; Definition of main function</i>
376int %main() { <i>; int()* </i>
377 <i>; Convert [13x sbyte]* to sbyte *...</i>
378 %cast210 = <a
379 href="#i_getelementptr">getelementptr</a> [13 x sbyte]* %.LC0, long 0, long 0 <i>; sbyte*</i>
380
381 <i>; Call puts function to write out the string to stdout...</i>
382 <a
383 href="#i_call">call</a> int %puts(sbyte* %cast210) <i>; int</i>
384 <a
385 href="#i_ret">ret</a> int 0<br>}<br></pre>
386
387<p>This example is made up of a <a href="#globalvars">global variable</a>
388named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
389function, and a <a href="#functionstructure">function definition</a>
390for "<tt>main</tt>".</p>
391
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]392<p>In general, a module is made up of a list of global values,
393where both functions and global variables are global values. Global values are
394represented by a pointer to a memory location (in this case, a pointer to an
395array of char, and a pointer to a function), and have one of the following <a
396href="#linkage">linkage types</a>.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]397
Chris Lattnerd2d29a02006-06-13 03:05:47 +0000[diff] [blame]398<p>Due to a limitation in the current LLVM assembly parser (it is limited by
399one-token lookahead), modules are split into two pieces by the "implementation"
400keyword. Global variable prototypes and definitions must occur before the
401keyword, and function definitions must occur after it. Function prototypes may
402occur either before or after it. In the future, the implementation keyword may
403become a noop, if the parser gets smarter.</p>
404
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]405</div>
406
407<!-- ======================================================================= -->
408<div class="doc_subsection">
409 <a name="linkage">Linkage Types</a>
410</div>
411
412<div class="doc_text">
413
414<p>
415All Global Variables and Functions have one of the following types of linkage:
416</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]417
418<dl>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]419
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]420 <dt><tt><b><a name="linkage_internal">internal</a></b></tt> </dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]421
422 <dd>Global values with internal linkage are only directly accessible by
423 objects in the current module. In particular, linking code into a module with
424 an internal global value may cause the internal to be renamed as necessary to
425 avoid collisions. Because the symbol is internal to the module, all
426 references can be updated. This corresponds to the notion of the
427 '<tt>static</tt>' keyword in C, or the idea of "anonymous namespaces" in C++.
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]428 </dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]429
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]430 <dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]431
432 <dd>"<tt>linkonce</tt>" linkage is similar to <tt>internal</tt> linkage, with
433 the twist that linking together two modules defining the same
434 <tt>linkonce</tt> globals will cause one of the globals to be discarded. This
435 is typically used to implement inline functions. Unreferenced
436 <tt>linkonce</tt> globals are allowed to be discarded.
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]437 </dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]438
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]439 <dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]440
441 <dd>"<tt>weak</tt>" linkage is exactly the same as <tt>linkonce</tt> linkage,
442 except that unreferenced <tt>weak</tt> globals may not be discarded. This is
443 used to implement constructs in C such as "<tt>int X;</tt>" at global scope.
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]444 </dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]445
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]446 <dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]447
448 <dd>"<tt>appending</tt>" linkage may only be applied to global variables of
449 pointer to array type. When two global variables with appending linkage are
450 linked together, the two global arrays are appended together. This is the
451 LLVM, typesafe, equivalent of having the system linker append together
452 "sections" with identical names when .o files are linked.
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]453 </dd>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]454
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]455 <dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]456
457 <dd>If none of the above identifiers are used, the global is externally
458 visible, meaning that it participates in linkage and can be used to resolve
459 external symbol references.
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]460 </dd>
Anton Korobeynikovd61d39e2006-09-14 18:23:27 +0000[diff] [blame]461
462 <dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
463
464 <dd>"<tt>extern_weak</tt>" TBD
465 </dd>
466
467 <p>
468 The next two types of linkage are targeted for Microsoft Windows platform
469 only. They are designed to support importing (exporting) symbols from (to)
470 DLLs.
471 </p>
472
473 <dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
474
475 <dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
476 or variable via a global pointer to a pointer that is set up by the DLL
477 exporting the symbol. On Microsoft Windows targets, the pointer name is
478 formed by combining <code>_imp__</code> and the function or variable name.
479 </dd>
480
481 <dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
482
483 <dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
484 pointer to a pointer in a DLL, so that it can be referenced with the
485 <tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
486 name is formed by combining <code>_imp__</code> and the function or variable
487 name.
488 </dd>
489
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]490</dl>
491
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]492<p><a name="linkage_external">For example, since the "<tt>.LC0</tt>"
493variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
494variable and was linked with this one, one of the two would be renamed,
495preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
496external (i.e., lacking any linkage declarations), they are accessible
497outside of the current module. It is illegal for a function <i>declaration</i>
498to have any linkage type other than "externally visible".</a></p>
Chris Lattnerd79749a2004-12-09 16:36:40 +0000[diff] [blame]499
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]500</div>
501
502<!-- ======================================================================= -->
503<div class="doc_subsection">
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]504 <a name="callingconv">Calling Conventions</a>
505</div>
506
507<div class="doc_text">
508
509<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
510and <a href="#i_invoke">invokes</a> can all have an optional calling convention
511specified for the call. The calling convention of any pair of dynamic
512caller/callee must match, or the behavior of the program is undefined. The
513following calling conventions are supported by LLVM, and more may be added in
514the future:</p>
515
516<dl>
517 <dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
518
519 <dd>This calling convention (the default if no other calling convention is
520 specified) matches the target C calling conventions. This calling convention
John Criswell02fdc6f2005-05-12 16:52:32 +0000[diff] [blame]521 supports varargs function calls and tolerates some mismatch in the declared
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]522 prototype and implemented declaration of the function (as does normal C).
523 </dd>
524
Chris Lattner95ff1952006-05-19 21:15:36 +0000[diff] [blame]525 <dt><b>"<tt>csretcc</tt>" - The C struct return calling convention</b>:</dt>
526
527 <dd>This calling convention matches the target C calling conventions, except
528 that functions with this convention are required to take a pointer as their
529 first argument, and the return type of the function must be void. This is
530 used for C functions that return aggregates by-value. In this case, the
531 function has been transformed to take a pointer to the struct as the first
532 argument to the function. For targets where the ABI specifies specific
533 behavior for structure-return calls, the calling convention can be used to
534 distinguish between struct return functions and other functions that take a
535 pointer to a struct as the first argument.
536 </dd>
537
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]538 <dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
539
540 <dd>This calling convention attempts to make calls as fast as possible
541 (e.g. by passing things in registers). This calling convention allows the
542 target to use whatever tricks it wants to produce fast code for the target,
Chris Lattnerc792eb32005-05-06 23:08:23 +0000[diff] [blame]543 without having to conform to an externally specified ABI. Implementations of
544 this convention should allow arbitrary tail call optimization to be supported.
545 This calling convention does not support varargs and requires the prototype of
546 all callees to exactly match the prototype of the function definition.
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]547 </dd>
548
549 <dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
550
551 <dd>This calling convention attempts to make code in the caller as efficient
552 as possible under the assumption that the call is not commonly executed. As
553 such, these calls often preserve all registers so that the call does not break
554 any live ranges in the caller side. This calling convention does not support
555 varargs and requires the prototype of all callees to exactly match the
556 prototype of the function definition.
557 </dd>
558
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]559 <dt><b>"<tt>cc &lt;<em>n</em>&gt;</tt>" - Numbered convention</b>:</dt>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]560
561 <dd>Any calling convention may be specified by number, allowing
562 target-specific calling conventions to be used. Target specific calling
563 conventions start at 64.
564 </dd>
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]565</dl>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]566
567<p>More calling conventions can be added/defined on an as-needed basis, to
568support pascal conventions or any other well-known target-independent
569convention.</p>
570
571</div>
572
573<!-- ======================================================================= -->
574<div class="doc_subsection">
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]575 <a name="globalvars">Global Variables</a>
576</div>
577
578<div class="doc_text">
579
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]580<p>Global variables define regions of memory allocated at compilation time
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]581instead of run-time. Global variables may optionally be initialized, may have
582an explicit section to be placed in, and may
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]583have an optional explicit alignment specified. A
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]584variable may be defined as a global "constant," which indicates that the
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]585contents of the variable will <b>never</b> be modified (enabling better
586optimization, allowing the global data to be placed in the read-only section of
587an executable, etc). Note that variables that need runtime initialization
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]588cannot be marked "constant" as there is a store to the variable.</p>
Chris Lattner5d5aede2005-02-12 19:30:21 +0000[diff] [blame]589
590<p>
591LLVM explicitly allows <em>declarations</em> of global variables to be marked
592constant, even if the final definition of the global is not. This capability
593can be used to enable slightly better optimization of the program, but requires
594the language definition to guarantee that optimizations based on the
595'constantness' are valid for the translation units that do not include the
596definition.
597</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]598
599<p>As SSA values, global variables define pointer values that are in
600scope (i.e. they dominate) all basic blocks in the program. Global
601variables always define a pointer to their "content" type because they
602describe a region of memory, and all memory objects in LLVM are
603accessed through pointers.</p>
604
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]605<p>LLVM allows an explicit section to be specified for globals. If the target
606supports it, it will emit globals to the section specified.</p>
607
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]608<p>An explicit alignment may be specified for a global. If not present, or if
609the alignment is set to zero, the alignment of the global is set by the target
610to whatever it feels convenient. If an explicit alignment is specified, the
611global is forced to have at least that much alignment. All alignments must be
612a power of 2.</p>
613
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]614</div>
615
616
617<!-- ======================================================================= -->
618<div class="doc_subsection">
619 <a name="functionstructure">Functions</a>
620</div>
621
622<div class="doc_text">
623
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]624<p>LLVM function definitions consist of an optional <a href="#linkage">linkage
625type</a>, an optional <a href="#callingconv">calling convention</a>, a return
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]626type, a function name, a (possibly empty) argument list, an optional section,
627an optional alignment, an opening curly brace,
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]628a list of basic blocks, and a closing curly brace. LLVM function declarations
629are defined with the "<tt>declare</tt>" keyword, an optional <a
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]630href="#callingconv">calling convention</a>, a return type, a function name,
631a possibly empty list of arguments, and an optional alignment.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]632
633<p>A function definition contains a list of basic blocks, forming the CFG for
634the function. Each basic block may optionally start with a label (giving the
635basic block a symbol table entry), contains a list of instructions, and ends
636with a <a href="#terminators">terminator</a> instruction (such as a branch or
637function return).</p>
638
John Criswell02fdc6f2005-05-12 16:52:32 +0000[diff] [blame]639<p>The first basic block in a program is special in two ways: it is immediately
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]640executed on entrance to the function, and it is not allowed to have predecessor
641basic blocks (i.e. there can not be any branches to the entry block of a
642function). Because the block can have no predecessors, it also cannot have any
643<a href="#i_phi">PHI nodes</a>.</p>
644
645<p>LLVM functions are identified by their name and type signature. Hence, two
646functions with the same name but different parameter lists or return values are
Chris Lattner455fc8c2005-03-07 22:13:59 +0000[diff] [blame]647considered different functions, and LLVM will resolve references to each
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]648appropriately.</p>
649
Chris Lattner662c8722005-11-12 00:45:07 +0000[diff] [blame]650<p>LLVM allows an explicit section to be specified for functions. If the target
651supports it, it will emit functions to the section specified.</p>
652
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]653<p>An explicit alignment may be specified for a function. If not present, or if
654the alignment is set to zero, the alignment of the function is set by the target
655to whatever it feels convenient. If an explicit alignment is specified, the
656function is forced to have at least that much alignment. All alignments must be
657a power of 2.</p>
658
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]659</div>
660
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]661<!-- ======================================================================= -->
662<div class="doc_subsection">
Chris Lattner93564892006-04-08 04:40:53 +0000[diff] [blame]663 <a name="moduleasm">Module-Level Inline Assembly</a>
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]664</div>
665
666<div class="doc_text">
667<p>
668Modules may contain "module-level inline asm" blocks, which corresponds to the
669GCC "file scope inline asm" blocks. These blocks are internally concatenated by
670LLVM and treated as a single unit, but may be separated in the .ll file if
671desired. The syntax is very simple:
672</p>
673
674<div class="doc_code"><pre>
Chris Lattnera1280ad2006-01-24 00:37:20 +0000[diff] [blame]675 module asm "inline asm code goes here"
676 module asm "more can go here"
Chris Lattner91c15c42006-01-23 23:23:47 +0000[diff] [blame]677</pre></div>
678
679<p>The strings can contain any character by escaping non-printable characters.
680 The escape sequence used is simply "\xx" where "xx" is the two digit hex code
681 for the number.
682</p>
683
684<p>
685 The inline asm code is simply printed to the machine code .s file when
686 assembly code is generated.
687</p>
688</div>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]689
690
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]691<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]692<div class="doc_section"> <a name="typesystem">Type System</a> </div>
693<!-- *********************************************************************** -->
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]694
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]695<div class="doc_text">
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]696
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]697<p>The LLVM type system is one of the most important features of the
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]698intermediate representation. Being typed enables a number of
699optimizations to be performed on the IR directly, without having to do
700extra analyses on the side before the transformation. A strong type
701system makes it easier to read the generated code and enables novel
702analyses and transformations that are not feasible to perform on normal
703three address code representations.</p>
Chris Lattner6af02f32004-12-09 16:11:40 +0000[diff] [blame]704
705</div>
706
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]707<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]708<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]709<div class="doc_text">
John Criswell417228d2004-04-09 16:48:45 +0000[diff] [blame]710<p>The primitive types are the fundamental building blocks of the LLVM
Chris Lattner455fc8c2005-03-07 22:13:59 +0000[diff] [blame]711system. The current set of primitive types is as follows:</p>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]712
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]713<table class="layout">
714 <tr class="layout">
715 <td class="left">
716 <table>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]717 <tbody>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]718 <tr><th>Type</th><th>Description</th></tr>
719 <tr><td><tt>void</tt></td><td>No value</td></tr>
Misha Brukman36c6bc12005-04-22 18:02:52 +0000[diff] [blame]720 <tr><td><tt>ubyte</tt></td><td>Unsigned 8-bit value</td></tr>
721 <tr><td><tt>ushort</tt></td><td>Unsigned 16-bit value</td></tr>
722 <tr><td><tt>uint</tt></td><td>Unsigned 32-bit value</td></tr>
723 <tr><td><tt>ulong</tt></td><td>Unsigned 64-bit value</td></tr>
724 <tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]725 <tr><td><tt>label</tt></td><td>Branch destination</td></tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]726 </tbody>
727 </table>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]728 </td>
729 <td class="right">
730 <table>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]731 <tbody>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]732 <tr><th>Type</th><th>Description</th></tr>
733 <tr><td><tt>bool</tt></td><td>True or False value</td></tr>
Misha Brukman36c6bc12005-04-22 18:02:52 +0000[diff] [blame]734 <tr><td><tt>sbyte</tt></td><td>Signed 8-bit value</td></tr>
735 <tr><td><tt>short</tt></td><td>Signed 16-bit value</td></tr>
736 <tr><td><tt>int</tt></td><td>Signed 32-bit value</td></tr>
737 <tr><td><tt>long</tt></td><td>Signed 64-bit value</td></tr>
738 <tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]739 </tbody>
740 </table>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]741 </td>
742 </tr>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]743</table>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]744</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]745
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]746<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]747<div class="doc_subsubsection"> <a name="t_classifications">Type
748Classifications</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]749<div class="doc_text">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]750<p>These different primitive types fall into a few useful
751classifications:</p>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]752
753<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]754 <tbody>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]755 <tr><th>Classification</th><th>Types</th></tr>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]756 <tr>
757 <td><a name="t_signed">signed</a></td>
758 <td><tt>sbyte, short, int, long, float, double</tt></td>
759 </tr>
760 <tr>
761 <td><a name="t_unsigned">unsigned</a></td>
762 <td><tt>ubyte, ushort, uint, ulong</tt></td>
763 </tr>
764 <tr>
765 <td><a name="t_integer">integer</a></td>
766 <td><tt>ubyte, sbyte, ushort, short, uint, int, ulong, long</tt></td>
767 </tr>
768 <tr>
769 <td><a name="t_integral">integral</a></td>
Misha Brukman20f9a622004-08-12 20:16:08 +0000[diff] [blame]770 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long</tt>
771 </td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]772 </tr>
773 <tr>
774 <td><a name="t_floating">floating point</a></td>
775 <td><tt>float, double</tt></td>
776 </tr>
777 <tr>
778 <td><a name="t_firstclass">first class</a></td>
Misha Brukman20f9a622004-08-12 20:16:08 +0000[diff] [blame]779 <td><tt>bool, ubyte, sbyte, ushort, short, uint, int, ulong, long,<br>
780 float, double, <a href="#t_pointer">pointer</a>,
781 <a href="#t_packed">packed</a></tt></td>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]782 </tr>
783 </tbody>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]784</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]785
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]786<p>The <a href="#t_firstclass">first class</a> types are perhaps the
787most important. Values of these types are the only ones which can be
788produced by instructions, passed as arguments, or used as operands to
789instructions. This means that all structures and arrays must be
790manipulated either by pointer or by component.</p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]791</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]792
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]793<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]794<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]795
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]796<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]797
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]798<p>The real power in LLVM comes from the derived types in the system.
799This is what allows a programmer to represent arrays, functions,
800pointers, and other useful types. Note that these derived types may be
801recursive: For example, it is possible to have a two dimensional array.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]802
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]803</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]804
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]805<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]806<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]807
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]808<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]809
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]810<h5>Overview:</h5>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]811
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]812<p>The array type is a very simple derived type that arranges elements
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]813sequentially in memory. The array type requires a size (number of
814elements) and an underlying data type.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]815
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]816<h5>Syntax:</h5>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]817
818<pre>
819 [&lt;# elements&gt; x &lt;elementtype&gt;]
820</pre>
821
John Criswell02fdc6f2005-05-12 16:52:32 +0000[diff] [blame]822<p>The number of elements is a constant integer value; elementtype may
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]823be any type with a size.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]824
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]825<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]826<table class="layout">
827 <tr class="layout">
828 <td class="left">
829 <tt>[40 x int ]</tt><br/>
830 <tt>[41 x int ]</tt><br/>
831 <tt>[40 x uint]</tt><br/>
832 </td>
833 <td class="left">
834 Array of 40 integer values.<br/>
835 Array of 41 integer values.<br/>
836 Array of 40 unsigned integer values.<br/>
837 </td>
838 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]839</table>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]840<p>Here are some examples of multidimensional arrays:</p>
841<table class="layout">
842 <tr class="layout">
843 <td class="left">
844 <tt>[3 x [4 x int]]</tt><br/>
845 <tt>[12 x [10 x float]]</tt><br/>
846 <tt>[2 x [3 x [4 x uint]]]</tt><br/>
847 </td>
848 <td class="left">
John Criswell4a3327e2005-05-13 22:25:59 +0000[diff] [blame]849 3x4 array of integer values.<br/>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]850 12x10 array of single precision floating point values.<br/>
851 2x3x4 array of unsigned integer values.<br/>
852 </td>
853 </tr>
854</table>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]855
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]856<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
857length array. Normally, accesses past the end of an array are undefined in
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]858LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
859As a special case, however, zero length arrays are recognized to be variable
860length. This allows implementation of 'pascal style arrays' with the LLVM
861type "{ int, [0 x float]}", for example.</p>
862
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]863</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]864
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]865<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]866<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]867<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]868<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]869<p>The function type can be thought of as a function signature. It
870consists of a return type and a list of formal parameter types.
John Criswella0d50d22003-11-25 21:45:46 +0000[diff] [blame]871Function types are usually used to build virtual function tables
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]872(which are structures of pointers to functions), for indirect function
873calls, and when defining a function.</p>
John Criswella0d50d22003-11-25 21:45:46 +0000[diff] [blame]874<p>
875The return type of a function type cannot be an aggregate type.
876</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]877<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]878<pre> &lt;returntype&gt; (&lt;parameter list&gt;)<br></pre>
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]879<p>...where '<tt>&lt;parameter list&gt;</tt>' is a comma-separated list of type
Misha Brukman20f9a622004-08-12 20:16:08 +0000[diff] [blame]880specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]881which indicates that the function takes a variable number of arguments.
882Variable argument functions can access their arguments with the <a
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]883 href="#int_varargs">variable argument handling intrinsic</a> functions.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]884<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]885<table class="layout">
886 <tr class="layout">
887 <td class="left">
888 <tt>int (int)</tt> <br/>
889 <tt>float (int, int *) *</tt><br/>
890 <tt>int (sbyte *, ...)</tt><br/>
891 </td>
892 <td class="left">
893 function taking an <tt>int</tt>, returning an <tt>int</tt><br/>
894 <a href="#t_pointer">Pointer</a> to a function that takes an
Misha Brukman20f9a622004-08-12 20:16:08 +0000[diff] [blame]895 <tt>int</tt> and a <a href="#t_pointer">pointer</a> to <tt>int</tt>,
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]896 returning <tt>float</tt>.<br/>
897 A vararg function that takes at least one <a href="#t_pointer">pointer</a>
898 to <tt>sbyte</tt> (signed char in C), which returns an integer. This is
899 the signature for <tt>printf</tt> in LLVM.<br/>
900 </td>
901 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]902</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]903
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]904</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]905<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]906<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]907<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]908<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]909<p>The structure type is used to represent a collection of data members
910together in memory. The packing of the field types is defined to match
911the ABI of the underlying processor. The elements of a structure may
912be any type that has a size.</p>
913<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
914and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
915field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
916instruction.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]917<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]918<pre> { &lt;type list&gt; }<br></pre>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]919<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]920<table class="layout">
921 <tr class="layout">
922 <td class="left">
923 <tt>{ int, int, int }</tt><br/>
924 <tt>{ float, int (int) * }</tt><br/>
925 </td>
926 <td class="left">
927 a triple of three <tt>int</tt> values<br/>
928 A pair, where the first element is a <tt>float</tt> and the second element
929 is a <a href="#t_pointer">pointer</a> to a <a href="#t_function">function</a>
930 that takes an <tt>int</tt>, returning an <tt>int</tt>.<br/>
931 </td>
932 </tr>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]933</table>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]934</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]935
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]936<!-- _______________________________________________________________________ -->
Andrew Lenharth8df88e22006-12-08 17:13:00 +0000[diff] [blame^]937<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
938</div>
939<div class="doc_text">
940<h5>Overview:</h5>
941<p>The packed structure type is used to represent a collection of data members
942together in memory. There is no padding between fields. Further, the alignment
943of a packed structure is 1 byte. The elements of a packed structure may
944be any type that has a size.</p>
945<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
946and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
947field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
948instruction.</p>
949<h5>Syntax:</h5>
950<pre> &lt; { &lt;type list&gt; } &gt; <br></pre>
951<h5>Examples:</h5>
952<table class="layout">
953 <tr class="layout">
954 <td class="left">
955 <tt> &lt; { int, int, int } &gt; </tt><br/>
956 <tt> &lt; { float, int (int) * } &gt; </tt><br/>
957 </td>
958 <td class="left">
959 a triple of three <tt>int</tt> values<br/>
960 A pair, where the first element is a <tt>float</tt> and the second element
961 is a <a href="#t_pointer">pointer</a> to a <a href="#t_function">function</a>
962 that takes an <tt>int</tt>, returning an <tt>int</tt>.<br/>
963 </td>
964 </tr>
965</table>
966</div>
967
968<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]969<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]970<div class="doc_text">
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]971<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]972<p>As in many languages, the pointer type represents a pointer or
973reference to another object, which must live in memory.</p>
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]974<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]975<pre> &lt;type&gt; *<br></pre>
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]976<h5>Examples:</h5>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]977<table class="layout">
978 <tr class="layout">
979 <td class="left">
980 <tt>[4x int]*</tt><br/>
981 <tt>int (int *) *</tt><br/>
982 </td>
983 <td class="left">
984 A <a href="#t_pointer">pointer</a> to <a href="#t_array">array</a> of
985 four <tt>int</tt> values<br/>
986 A <a href="#t_pointer">pointer</a> to a <a
Chris Lattner16fb0032005-02-19 02:22:14 +0000[diff] [blame]987 href="#t_function">function</a> that takes an <tt>int*</tt>, returning an
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]988 <tt>int</tt>.<br/>
989 </td>
990 </tr>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]991</table>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]992</div>
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]993
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]994<!-- _______________________________________________________________________ -->
995<div class="doc_subsubsection"> <a name="t_packed">Packed Type</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]996<div class="doc_text">
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]997
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]998<h5>Overview:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]999
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1000<p>A packed type is a simple derived type that represents a vector
1001of elements. Packed types are used when multiple primitive data
1002are operated in parallel using a single instruction (SIMD).
1003A packed type requires a size (number of
Chris Lattner330ce692005-11-10 01:44:22 +0000[diff] [blame]1004elements) and an underlying primitive data type. Vectors must have a power
1005of two length (1, 2, 4, 8, 16 ...). Packed types are
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1006considered <a href="#t_firstclass">first class</a>.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1007
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1008<h5>Syntax:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1009
1010<pre>
1011 &lt; &lt;# elements&gt; x &lt;elementtype&gt; &gt;
1012</pre>
1013
John Criswell4a3327e2005-05-13 22:25:59 +0000[diff] [blame]1014<p>The number of elements is a constant integer value; elementtype may
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1015be any integral or floating point type.</p>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1016
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1017<h5>Examples:</h5>
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1018
Reid Spencerc3c4c4f2004-11-01 08:19:36 +0000[diff] [blame]1019<table class="layout">
1020 <tr class="layout">
1021 <td class="left">
1022 <tt>&lt;4 x int&gt;</tt><br/>
1023 <tt>&lt;8 x float&gt;</tt><br/>
1024 <tt>&lt;2 x uint&gt;</tt><br/>
1025 </td>
1026 <td class="left">
1027 Packed vector of 4 integer values.<br/>
1028 Packed vector of 8 floating-point values.<br/>
1029 Packed vector of 2 unsigned integer values.<br/>
1030 </td>
1031 </tr>
1032</table>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1033</div>
1034
Chris Lattner37b6b092005-04-25 17:34:15 +0000[diff] [blame]1035<!-- _______________________________________________________________________ -->
1036<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
1037<div class="doc_text">
1038
1039<h5>Overview:</h5>
1040
1041<p>Opaque types are used to represent unknown types in the system. This
1042corresponds (for example) to the C notion of a foward declared structure type.
1043In LLVM, opaque types can eventually be resolved to any type (not just a
1044structure type).</p>
1045
1046<h5>Syntax:</h5>
1047
1048<pre>
1049 opaque
1050</pre>
1051
1052<h5>Examples:</h5>
1053
1054<table class="layout">
1055 <tr class="layout">
1056 <td class="left">
1057 <tt>opaque</tt>
1058 </td>
1059 <td class="left">
1060 An opaque type.<br/>
1061 </td>
1062 </tr>
1063</table>
1064</div>
1065
1066
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1067<!-- *********************************************************************** -->
1068<div class="doc_section"> <a name="constants">Constants</a> </div>
1069<!-- *********************************************************************** -->
1070
1071<div class="doc_text">
1072
1073<p>LLVM has several different basic types of constants. This section describes
1074them all and their syntax.</p>
1075
1076</div>
1077
1078<!-- ======================================================================= -->
Reid Spencer8f08d802004-12-09 18:02:53 +0000[diff] [blame]1079<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1080
1081<div class="doc_text">
1082
1083<dl>
1084 <dt><b>Boolean constants</b></dt>
1085
1086 <dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
1087 constants of the <tt><a href="#t_primitive">bool</a></tt> type.
1088 </dd>
1089
1090 <dt><b>Integer constants</b></dt>
1091
Reid Spencer8f08d802004-12-09 18:02:53 +0000[diff] [blame]1092 <dd>Standard integers (such as '4') are constants of the <a
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1093 href="#t_integer">integer</a> type. Negative numbers may be used with signed
1094 integer types.
1095 </dd>
1096
1097 <dt><b>Floating point constants</b></dt>
1098
1099 <dd>Floating point constants use standard decimal notation (e.g. 123.421),
1100 exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1101 notation (see below). Floating point constants must have a <a
1102 href="#t_floating">floating point</a> type. </dd>
1103
1104 <dt><b>Null pointer constants</b></dt>
1105
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]1106 <dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1107 and must be of <a href="#t_pointer">pointer type</a>.</dd>
1108
1109</dl>
1110
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]1111<p>The one non-intuitive notation for constants is the optional hexadecimal form
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1112of floating point constants. For example, the form '<tt>double
11130x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
11144.5e+15</tt>'. The only time hexadecimal floating point constants are required
Reid Spencer8f08d802004-12-09 18:02:53 +0000[diff] [blame]1115(and the only time that they are generated by the disassembler) is when a
1116floating point constant must be emitted but it cannot be represented as a
1117decimal floating point number. For example, NaN's, infinities, and other
1118special values are represented in their IEEE hexadecimal format so that
1119assembly and disassembly do not cause any bits to change in the constants.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1120
1121</div>
1122
1123<!-- ======================================================================= -->
1124<div class="doc_subsection"><a name="aggregateconstants">Aggregate Constants</a>
1125</div>
1126
1127<div class="doc_text">
Chris Lattner455fc8c2005-03-07 22:13:59 +0000[diff] [blame]1128<p>Aggregate constants arise from aggregation of simple constants
1129and smaller aggregate constants.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1130
1131<dl>
1132 <dt><b>Structure constants</b></dt>
1133
1134 <dd>Structure constants are represented with notation similar to structure
1135 type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner455fc8c2005-03-07 22:13:59 +0000[diff] [blame]1136 (<tt>{}</tt>)). For example: "<tt>{ int 4, float 17.0, int* %G }</tt>",
1137 where "<tt>%G</tt>" is declared as "<tt>%G = external global int</tt>". Structure constants
1138 must have <a href="#t_struct">structure type</a>, and the number and
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1139 types of elements must match those specified by the type.
1140 </dd>
1141
1142 <dt><b>Array constants</b></dt>
1143
1144 <dd>Array constants are represented with notation similar to array type
1145 definitions (a comma separated list of elements, surrounded by square brackets
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]1146 (<tt>[]</tt>)). For example: "<tt>[ int 42, int 11, int 74 ]</tt>". Array
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1147 constants must have <a href="#t_array">array type</a>, and the number and
1148 types of elements must match those specified by the type.
1149 </dd>
1150
1151 <dt><b>Packed constants</b></dt>
1152
1153 <dd>Packed constants are represented with notation similar to packed type
1154 definitions (a comma separated list of elements, surrounded by
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]1155 less-than/greater-than's (<tt>&lt;&gt;</tt>)). For example: "<tt>&lt; int 42,
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1156 int 11, int 74, int 100 &gt;</tt>". Packed constants must have <a
1157 href="#t_packed">packed type</a>, and the number and types of elements must
1158 match those specified by the type.
1159 </dd>
1160
1161 <dt><b>Zero initialization</b></dt>
1162
1163 <dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
1164 value to zero of <em>any</em> type, including scalar and aggregate types.
1165 This is often used to avoid having to print large zero initializers (e.g. for
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]1166 large arrays) and is always exactly equivalent to using explicit zero
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1167 initializers.
1168 </dd>
1169</dl>
1170
1171</div>
1172
1173<!-- ======================================================================= -->
1174<div class="doc_subsection">
1175 <a name="globalconstants">Global Variable and Function Addresses</a>
1176</div>
1177
1178<div class="doc_text">
1179
1180<p>The addresses of <a href="#globalvars">global variables</a> and <a
1181href="#functionstructure">functions</a> are always implicitly valid (link-time)
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]1182constants. These constants are explicitly referenced when the <a
1183href="#identifiers">identifier for the global</a> is used and always have <a
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1184href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
1185file:</p>
1186
1187<pre>
1188 %X = global int 17
1189 %Y = global int 42
1190 %Z = global [2 x int*] [ int* %X, int* %Y ]
1191</pre>
1192
1193</div>
1194
1195<!-- ======================================================================= -->
Reid Spencer641f5c92004-12-09 18:13:12 +0000[diff] [blame]1196<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1197<div class="doc_text">
Reid Spencer641f5c92004-12-09 18:13:12 +0000[diff] [blame]1198 <p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
John Criswell4a3327e2005-05-13 22:25:59 +0000[diff] [blame]1199 no specific value. Undefined values may be of any type and be used anywhere
Reid Spencer641f5c92004-12-09 18:13:12 +0000[diff] [blame]1200 a constant is permitted.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1201
Reid Spencer641f5c92004-12-09 18:13:12 +0000[diff] [blame]1202 <p>Undefined values indicate to the compiler that the program is well defined
1203 no matter what value is used, giving the compiler more freedom to optimize.
1204 </p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1205</div>
1206
1207<!-- ======================================================================= -->
1208<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
1209</div>
1210
1211<div class="doc_text">
1212
1213<p>Constant expressions are used to allow expressions involving other constants
1214to be used as constants. Constant expressions may be of any <a
John Criswell4a3327e2005-05-13 22:25:59 +0000[diff] [blame]1215href="#t_firstclass">first class</a> type and may involve any LLVM operation
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1216that does not have side effects (e.g. load and call are not supported). The
1217following is the syntax for constant expressions:</p>
1218
1219<dl>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]1220 <dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
1221 <dd>Truncate a constant to another type. The bit size of CST must be larger
1222 than the bit size of TYPE. Both types must be integral.</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1223
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]1224 <dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
1225 <dd>Zero extend a constant to another type. The bit size of CST must be
1226 smaller or equal to the bit size of TYPE. Both types must be integral.</dd>
1227
1228 <dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
1229 <dd>Sign extend a constant to another type. The bit size of CST must be
1230 smaller or equal to the bit size of TYPE. Both types must be integral.</dd>
1231
1232 <dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
1233 <dd>Truncate a floating point constant to another floating point type. The
1234 size of CST must be larger than the size of TYPE. Both types must be
1235 floating point.</dd>
1236
1237 <dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
1238 <dd>Floating point extend a constant to another type. The size of CST must be
1239 smaller or equal to the size of TYPE. Both types must be floating point.</dd>
1240
1241 <dt><b><tt>fp2uint ( CST to TYPE )</tt></b></dt>
1242 <dd>Convert a floating point constant to the corresponding unsigned integer
1243 constant. TYPE must be an integer type. CST must be floating point. If the
1244 value won't fit in the integer type, the results are undefined.</dd>
1245
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]1246 <dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]1247 <dd>Convert a floating point constant to the corresponding signed integer
1248 constant. TYPE must be an integer type. CST must be floating point. If the
1249 value won't fit in the integer type, the results are undefined.</dd>
1250
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]1251 <dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]1252 <dd>Convert an unsigned integer constant to the corresponding floating point
1253 constant. TYPE must be floating point. CST must be of integer type. If the
1254 value won't fit in the floating point type, the results are undefined.</dd>
1255
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]1256 <dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]1257 <dd>Convert a signed integer constant to the corresponding floating point
1258 constant. TYPE must be floating point. CST must be of integer type. If the
1259 value won't fit in the floating point type, the results are undefined.</dd>
1260
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]1261 <dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
1262 <dd>Convert a pointer typed constant to the corresponding integer constant
1263 TYPE must be an integer type. CST must be of pointer type. The CST value is
1264 zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
1265
1266 <dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
1267 <dd>Convert a integer constant to a pointer constant. TYPE must be a
1268 pointer type. CST must be of integer type. The CST value is zero extended,
1269 truncated, or unchanged to make it fit in a pointer size. This one is
1270 <i>really</i> dangerous!</dd>
1271
1272 <dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]1273 <dd>Convert a constant, CST, to another TYPE. The size of CST and TYPE must be
1274 identical (same number of bits). The conversion is done as if the CST value
1275 was stored to memory and read back as TYPE. In other words, no bits change
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]1276 with this operator, just the type. This can be used for conversion of
1277 packed types to any other type, as long as they have the same bit width. For
1278 pointers it is only valid to cast to another pointer type.
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]1279 </dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1280
1281 <dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
1282
1283 <dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
1284 constants. As with the <a href="#i_getelementptr">getelementptr</a>
1285 instruction, the index list may have zero or more indexes, which are required
1286 to make sense for the type of "CSTPTR".</dd>
1287
Robert Bocchino7e97a6d2006-01-10 19:31:34 +0000[diff] [blame]1288 <dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
1289
1290 <dd>Perform the <a href="#i_select">select operation</a> on
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]1291 constants.</dd>
1292
1293 <dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
1294 <dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
1295
1296 <dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
1297 <dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino7e97a6d2006-01-10 19:31:34 +0000[diff] [blame]1298
1299 <dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
1300
1301 <dd>Perform the <a href="#i_extractelement">extractelement
1302 operation</a> on constants.
1303
Robert Bocchinof72fdfe2006-01-15 20:48:27 +0000[diff] [blame]1304 <dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
1305
1306 <dd>Perform the <a href="#i_insertelement">insertelement
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]1307 operation</a> on constants.</dd>
Robert Bocchinof72fdfe2006-01-15 20:48:27 +0000[diff] [blame]1308
Chris Lattner016a0e52006-04-08 00:13:41 +0000[diff] [blame]1309
1310 <dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
1311
1312 <dd>Perform the <a href="#i_shufflevector">shufflevector
Reid Spencer9965ee72006-12-04 19:23:19 +0000[diff] [blame]1313 operation</a> on constants.</dd>
Chris Lattner016a0e52006-04-08 00:13:41 +0000[diff] [blame]1314
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1315 <dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
1316
Reid Spencer641f5c92004-12-09 18:13:12 +0000[diff] [blame]1317 <dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
1318 be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1319 binary</a> operations. The constraints on operands are the same as those for
1320 the corresponding instruction (e.g. no bitwise operations on floating point
John Criswell02fdc6f2005-05-12 16:52:32 +0000[diff] [blame]1321 values are allowed).</dd>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1322</dl>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1323</div>
Chris Lattnerb1652612004-03-08 16:49:10 +0000[diff] [blame]1324
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1325<!-- *********************************************************************** -->
Chris Lattner98f013c2006-01-25 23:47:57 +0000[diff] [blame]1326<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
1327<!-- *********************************************************************** -->
1328
1329<!-- ======================================================================= -->
1330<div class="doc_subsection">
1331<a name="inlineasm">Inline Assembler Expressions</a>
1332</div>
1333
1334<div class="doc_text">
1335
1336<p>
1337LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
1338Module-Level Inline Assembly</a>) through the use of a special value. This
1339value represents the inline assembler as a string (containing the instructions
1340to emit), a list of operand constraints (stored as a string), and a flag that
1341indicates whether or not the inline asm expression has side effects. An example
1342inline assembler expression is:
1343</p>
1344
1345<pre>
1346 int(int) asm "bswap $0", "=r,r"
1347</pre>
1348
1349<p>
1350Inline assembler expressions may <b>only</b> be used as the callee operand of
1351a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
1352</p>
1353
1354<pre>
1355 %X = call int asm "<a href="#i_bswap">bswap</a> $0", "=r,r"(int %Y)
1356</pre>
1357
1358<p>
1359Inline asms with side effects not visible in the constraint list must be marked
1360as having side effects. This is done through the use of the
1361'<tt>sideeffect</tt>' keyword, like so:
1362</p>
1363
1364<pre>
1365 call void asm sideeffect "eieio", ""()
1366</pre>
1367
1368<p>TODO: The format of the asm and constraints string still need to be
1369documented here. Constraints on what can be done (e.g. duplication, moving, etc
1370need to be documented).
1371</p>
1372
1373</div>
1374
1375<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1376<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
1377<!-- *********************************************************************** -->
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1378
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1379<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1380
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1381<p>The LLVM instruction set consists of several different
1382classifications of instructions: <a href="#terminators">terminator
John Criswell4a3327e2005-05-13 22:25:59 +0000[diff] [blame]1383instructions</a>, <a href="#binaryops">binary instructions</a>,
1384<a href="#bitwiseops">bitwise binary instructions</a>, <a
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1385 href="#memoryops">memory instructions</a>, and <a href="#otherops">other
1386instructions</a>.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1387
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1388</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1389
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1390<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1391<div class="doc_subsection"> <a name="terminators">Terminator
1392Instructions</a> </div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1393
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1394<div class="doc_text">
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1395
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1396<p>As mentioned <a href="#functionstructure">previously</a>, every
1397basic block in a program ends with a "Terminator" instruction, which
1398indicates which block should be executed after the current block is
1399finished. These terminator instructions typically yield a '<tt>void</tt>'
1400value: they produce control flow, not values (the one exception being
1401the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]1402<p>There are six different terminator instructions: the '<a
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1403 href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
1404instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]1405the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
1406 href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
1407 href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1408
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1409</div>
Chris Lattner74d3f822004-12-09 17:30:23 +0000[diff] [blame]1410
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1411<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1412<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
1413Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1414<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1415<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1416<pre> ret &lt;type&gt; &lt;value&gt; <i>; Return a value from a non-void function</i>
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1417 ret void <i>; Return from void function</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1418</pre>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1419<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1420<p>The '<tt>ret</tt>' instruction is used to return control flow (and a
John Criswell4a3327e2005-05-13 22:25:59 +0000[diff] [blame]1421value) from a function back to the caller.</p>
John Criswell417228d2004-04-09 16:48:45 +0000[diff] [blame]1422<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1423returns a value and then causes control flow, and one that just causes
1424control flow to occur.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1425<h5>Arguments:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1426<p>The '<tt>ret</tt>' instruction may return any '<a
1427 href="#t_firstclass">first class</a>' type. Notice that a function is
1428not <a href="#wellformed">well formed</a> if there exists a '<tt>ret</tt>'
1429instruction inside of the function that returns a value that does not
1430match the return type of the function.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1431<h5>Semantics:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1432<p>When the '<tt>ret</tt>' instruction is executed, control flow
1433returns back to the calling function's context. If the caller is a "<a
John Criswell40db33f2004-06-25 15:16:57 +0000[diff] [blame]1434 href="#i_call"><tt>call</tt></a>" instruction, execution continues at
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1435the instruction after the call. If the caller was an "<a
1436 href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
John Criswell02fdc6f2005-05-12 16:52:32 +0000[diff] [blame]1437at the beginning of the "normal" destination block. If the instruction
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1438returns a value, that value shall set the call or invoke instruction's
1439return value.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1440<h5>Example:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1441<pre> ret int 5 <i>; Return an integer value of 5</i>
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]1442 ret void <i>; Return from a void function</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1443</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1444</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1445<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1446<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1447<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1448<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1449<pre> br bool &lt;cond&gt;, label &lt;iftrue&gt;, label &lt;iffalse&gt;<br> br label &lt;dest&gt; <i>; Unconditional branch</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1450</pre>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1451<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1452<p>The '<tt>br</tt>' instruction is used to cause control flow to
1453transfer to a different basic block in the current function. There are
1454two forms of this instruction, corresponding to a conditional branch
1455and an unconditional branch.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1456<h5>Arguments:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1457<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
1458single '<tt>bool</tt>' value and two '<tt>label</tt>' values. The
1459unconditional form of the '<tt>br</tt>' instruction takes a single '<tt>label</tt>'
1460value as a target.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1461<h5>Semantics:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1462<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>bool</tt>'
1463argument is evaluated. If the value is <tt>true</tt>, control flows
1464to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
1465control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1466<h5>Example:</h5>
Reid Spencere0083262006-11-18 21:55:45 +0000[diff] [blame]1467<pre>Test:<br> %cond = <a href="#i_icmp">icmp</a> eq, int %a, %b<br> br bool %cond, label %IfEqual, label %IfUnequal<br>IfEqual:<br> <a
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1468 href="#i_ret">ret</a> int 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> int 0<br></pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1469</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1470<!-- _______________________________________________________________________ -->
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]1471<div class="doc_subsubsection">
1472 <a name="i_switch">'<tt>switch</tt>' Instruction</a>
1473</div>
1474
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1475<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1476<h5>Syntax:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]1477
1478<pre>
1479 switch &lt;intty&gt; &lt;value&gt;, label &lt;defaultdest&gt; [ &lt;intty&gt; &lt;val&gt;, label &lt;dest&gt; ... ]
1480</pre>
1481
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1482<h5>Overview:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]1483
1484<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
1485several different places. It is a generalization of the '<tt>br</tt>'
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1486instruction, allowing a branch to occur to one of many possible
1487destinations.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]1488
1489
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1490<h5>Arguments:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]1491
1492<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
1493comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
1494an array of pairs of comparison value constants and '<tt>label</tt>'s. The
1495table is not allowed to contain duplicate constant entries.</p>
1496
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1497<h5>Semantics:</h5>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]1498
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1499<p>The <tt>switch</tt> instruction specifies a table of values and
1500destinations. When the '<tt>switch</tt>' instruction is executed, this
John Criswellbcbb18c2004-06-25 16:05:06 +0000[diff] [blame]1501table is searched for the given value. If the value is found, control flow is
1502transfered to the corresponding destination; otherwise, control flow is
1503transfered to the default destination.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1504
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]1505<h5>Implementation:</h5>
1506
1507<p>Depending on properties of the target machine and the particular
1508<tt>switch</tt> instruction, this instruction may be code generated in different
John Criswellbcbb18c2004-06-25 16:05:06 +0000[diff] [blame]1509ways. For example, it could be generated as a series of chained conditional
1510branches or with a lookup table.</p>
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]1511
1512<h5>Example:</h5>
1513
1514<pre>
1515 <i>; Emulate a conditional br instruction</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]1516 %Val = <a href="#i_zext">zext</a> bool %value to int
Chris Lattnercf96c6c2004-02-24 04:54:45 +0000[diff] [blame]1517 switch int %Val, label %truedest [int 0, label %falsedest ]
1518
1519 <i>; Emulate an unconditional br instruction</i>
1520 switch uint 0, label %dest [ ]
1521
1522 <i>; Implement a jump table:</i>
1523 switch uint %val, label %otherwise [ uint 0, label %onzero
1524 uint 1, label %onone
1525 uint 2, label %ontwo ]
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1526</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1527</div>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1528
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1529<!-- _______________________________________________________________________ -->
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1530<div class="doc_subsubsection">
1531 <a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
1532</div>
1533
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1534<div class="doc_text">
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1535
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1536<h5>Syntax:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1537
1538<pre>
1539 &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 Lattner6b7a0082006-05-14 18:23:06 +0000[diff] [blame]1540 to label &lt;normal label&gt; unwind label &lt;exception label&gt;
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1541</pre>
1542
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]1543<h5>Overview:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1544
1545<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
1546function, with the possibility of control flow transfer to either the
John Criswell02fdc6f2005-05-12 16:52:32 +0000[diff] [blame]1547'<tt>normal</tt>' label or the
1548'<tt>exception</tt>' label. If the callee function returns with the
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1549"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
1550"normal" label. If the callee (or any indirect callees) returns with the "<a
John Criswell02fdc6f2005-05-12 16:52:32 +0000[diff] [blame]1551href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
1552continued at the dynamically nearest "exception" label.</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1553
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1554<h5>Arguments:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1555
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1556<p>This instruction requires several arguments:</p>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1557
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1558<ol>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1559 <li>
John Criswell4a3327e2005-05-13 22:25:59 +0000[diff] [blame]1560 The optional "cconv" marker indicates which <a href="callingconv">calling
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1561 convention</a> the call should use. If none is specified, the call defaults
1562 to using C calling conventions.
1563 </li>
1564 <li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
1565 function value being invoked. In most cases, this is a direct function
1566 invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
1567 an arbitrary pointer to function value.
1568 </li>
1569
1570 <li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
1571 function to be invoked. </li>
1572
1573 <li>'<tt>function args</tt>': argument list whose types match the function
1574 signature argument types. If the function signature indicates the function
1575 accepts a variable number of arguments, the extra arguments can be
1576 specified. </li>
1577
1578 <li>'<tt>normal label</tt>': the label reached when the called function
1579 executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
1580
1581 <li>'<tt>exception label</tt>': the label reached when a callee returns with
1582 the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
1583
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1584</ol>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1585
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1586<h5>Semantics:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1587
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1588<p>This instruction is designed to operate as a standard '<tt><a
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1589href="#i_call">call</a></tt>' instruction in most regards. The primary
1590difference is that it establishes an association with a label, which is used by
1591the runtime library to unwind the stack.</p>
1592
1593<p>This instruction is used in languages with destructors to ensure that proper
1594cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
1595exception. Additionally, this is important for implementation of
1596'<tt>catch</tt>' clauses in high-level languages that support them.</p>
1597
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1598<h5>Example:</h5>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1599<pre>
1600 %retval = invoke int %Test(int 15) to label %Continue
Chris Lattner6b7a0082006-05-14 18:23:06 +0000[diff] [blame]1601 unwind label %TestCleanup <i>; {int}:retval set</i>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]1602 %retval = invoke <a href="#callingconv">coldcc</a> int %Test(int 15) to label %Continue
Chris Lattner6b7a0082006-05-14 18:23:06 +0000[diff] [blame]1603 unwind label %TestCleanup <i>; {int}:retval set</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1604</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1605</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]1606
1607
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]1608<!-- _______________________________________________________________________ -->
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]1609
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1610<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
1611Instruction</a> </div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]1612
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1613<div class="doc_text">
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]1614
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]1615<h5>Syntax:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]1616<pre>
1617 unwind
1618</pre>
1619
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]1620<h5>Overview:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]1621
1622<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
1623at the first callee in the dynamic call stack which used an <a
1624href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
1625primarily used to implement exception handling.</p>
1626
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]1627<h5>Semantics:</h5>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]1628
1629<p>The '<tt>unwind</tt>' intrinsic causes execution of the current function to
1630immediately halt. The dynamic call stack is then searched for the first <a
1631href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
1632execution continues at the "exceptional" destination block specified by the
1633<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
1634dynamic call chain, undefined behavior results.</p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1635</div>
Chris Lattner08b7d5b2004-10-16 18:04:13 +0000[diff] [blame]1636
1637<!-- _______________________________________________________________________ -->
1638
1639<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
1640Instruction</a> </div>
1641
1642<div class="doc_text">
1643
1644<h5>Syntax:</h5>
1645<pre>
1646 unreachable
1647</pre>
1648
1649<h5>Overview:</h5>
1650
1651<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
1652instruction is used to inform the optimizer that a particular portion of the
1653code is not reachable. This can be used to indicate that the code after a
1654no-return function cannot be reached, and other facts.</p>
1655
1656<h5>Semantics:</h5>
1657
1658<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
1659</div>
1660
1661
1662
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1663<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1664<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1665<div class="doc_text">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1666<p>Binary operators are used to do most of the computation in a
1667program. They require two operands, execute an operation on them, and
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]1668produce a single value. The operands might represent
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1669multiple data, as is the case with the <a href="#t_packed">packed</a> data type.
1670The result value of a binary operator is not
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1671necessarily the same type as its operands.</p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1672<p>There are several different binary operators:</p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1673</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1674<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1675<div class="doc_subsubsection"> <a name="i_add">'<tt>add</tt>'
1676Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1677<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1678<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1679<pre> &lt;result&gt; = add &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1680</pre>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1681<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1682<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1683<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1684<p>The two arguments to the '<tt>add</tt>' instruction must be either <a
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1685 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a> values.
1686 This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1687Both arguments must have identical types.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1688<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1689<p>The value produced is the integer or floating point sum of the two
1690operands.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1691<h5>Example:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1692<pre> &lt;result&gt; = add int 4, %var <i>; yields {int}:result = 4 + %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1693</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1694</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1695<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1696<div class="doc_subsubsection"> <a name="i_sub">'<tt>sub</tt>'
1697Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1698<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1699<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1700<pre> &lt;result&gt; = sub &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1701</pre>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1702<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1703<p>The '<tt>sub</tt>' instruction returns the difference of its two
1704operands.</p>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1705<p>Note that the '<tt>sub</tt>' instruction is used to represent the '<tt>neg</tt>'
1706instruction present in most other intermediate representations.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1707<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1708<p>The two arguments to the '<tt>sub</tt>' instruction must be either <a
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1709 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1710values.
1711This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1712Both arguments must have identical types.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1713<h5>Semantics:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1714<p>The value produced is the integer or floating point difference of
1715the two operands.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1716<h5>Example:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1717<pre> &lt;result&gt; = sub int 4, %var <i>; yields {int}:result = 4 - %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1718 &lt;result&gt; = sub int 0, %val <i>; yields {int}:result = -%var</i>
1719</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1720</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1721<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1722<div class="doc_subsubsection"> <a name="i_mul">'<tt>mul</tt>'
1723Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1724<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1725<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1726<pre> &lt;result&gt; = mul &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1727</pre>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1728<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1729<p>The '<tt>mul</tt>' instruction returns the product of its two
1730operands.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1731<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1732<p>The two arguments to the '<tt>mul</tt>' instruction must be either <a
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1733 href="#t_integer">integer</a> or <a href="#t_floating">floating point</a>
Chris Lattnerc8cb6952004-08-12 19:12:28 +0000[diff] [blame]1734values.
1735This instruction can also take <a href="#t_packed">packed</a> versions of the values.
1736Both arguments must have identical types.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1737<h5>Semantics:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1738<p>The value produced is the integer or floating point product of the
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1739two operands.</p>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1740<p>There is no signed vs unsigned multiplication. The appropriate
1741action is taken based on the type of the operand.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1742<h5>Example:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1743<pre> &lt;result&gt; = mul int 4, %var <i>; yields {int}:result = 4 * %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1744</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1745</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1746<!-- _______________________________________________________________________ -->
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]1747<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
1748</a></div>
1749<div class="doc_text">
1750<h5>Syntax:</h5>
1751<pre> &lt;result&gt; = udiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1752</pre>
1753<h5>Overview:</h5>
1754<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
1755operands.</p>
1756<h5>Arguments:</h5>
1757<p>The two arguments to the '<tt>udiv</tt>' instruction must be
1758<a href="#t_integer">integer</a> values. Both arguments must have identical
1759types. This instruction can also take <a href="#t_packed">packed</a> versions
1760of the values in which case the elements must be integers.</p>
1761<h5>Semantics:</h5>
1762<p>The value produced is the unsigned integer quotient of the two operands. This
1763instruction always performs an unsigned division operation, regardless of
1764whether the arguments are unsigned or not.</p>
1765<h5>Example:</h5>
1766<pre> &lt;result&gt; = udiv uint 4, %var <i>; yields {uint}:result = 4 / %var</i>
1767</pre>
1768</div>
1769<!-- _______________________________________________________________________ -->
1770<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
1771</a> </div>
1772<div class="doc_text">
1773<h5>Syntax:</h5>
1774<pre> &lt;result&gt; = sdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1775</pre>
1776<h5>Overview:</h5>
1777<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
1778operands.</p>
1779<h5>Arguments:</h5>
1780<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
1781<a href="#t_integer">integer</a> values. Both arguments must have identical
1782types. This instruction can also take <a href="#t_packed">packed</a> versions
1783of the values in which case the elements must be integers.</p>
1784<h5>Semantics:</h5>
1785<p>The value produced is the signed integer quotient of the two operands. This
1786instruction always performs a signed division operation, regardless of whether
1787the arguments are signed or not.</p>
1788<h5>Example:</h5>
1789<pre> &lt;result&gt; = sdiv int 4, %var <i>; yields {int}:result = 4 / %var</i>
1790</pre>
1791</div>
1792<!-- _______________________________________________________________________ -->
1793<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1794Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1795<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1796<h5>Syntax:</h5>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]1797<pre> &lt;result&gt; = fdiv &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1798</pre>
1799<h5>Overview:</h5>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]1800<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1801operands.</p>
1802<h5>Arguments:</h5>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]1803<p>The two arguments to the '<tt>div</tt>' instruction must be
1804<a href="#t_floating">floating point</a> values. Both arguments must have
1805identical types. This instruction can also take <a href="#t_packed">packed</a>
1806versions of the values in which case the elements must be floating point.</p>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1807<h5>Semantics:</h5>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]1808<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1809<h5>Example:</h5>
Reid Spencer7e80b0b2006-10-26 06:15:43 +0000[diff] [blame]1810<pre> &lt;result&gt; = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1811</pre>
1812</div>
1813<!-- _______________________________________________________________________ -->
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]1814<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
1815</div>
1816<div class="doc_text">
1817<h5>Syntax:</h5>
1818<pre> &lt;result&gt; = urem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1819</pre>
1820<h5>Overview:</h5>
1821<p>The '<tt>urem</tt>' instruction returns the remainder from the
1822unsigned division of its two arguments.</p>
1823<h5>Arguments:</h5>
1824<p>The two arguments to the '<tt>urem</tt>' instruction must be
1825<a href="#t_integer">integer</a> values. Both arguments must have identical
1826types.</p>
1827<h5>Semantics:</h5>
1828<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
1829This instruction always performs an unsigned division to get the remainder,
1830regardless of whether the arguments are unsigned or not.</p>
1831<h5>Example:</h5>
1832<pre> &lt;result&gt; = urem uint 4, %var <i>; yields {uint}:result = 4 % %var</i>
1833</pre>
1834
1835</div>
1836<!-- _______________________________________________________________________ -->
1837<div class="doc_subsubsection"> <a name="i_srem">'<tt>srem</tt>'
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1838Instruction</a> </div>
1839<div class="doc_text">
1840<h5>Syntax:</h5>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]1841<pre> &lt;result&gt; = srem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1842</pre>
1843<h5>Overview:</h5>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]1844<p>The '<tt>srem</tt>' instruction returns the remainder from the
1845signed division of its two operands.</p>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1846<h5>Arguments:</h5>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]1847<p>The two arguments to the '<tt>srem</tt>' instruction must be
1848<a href="#t_integer">integer</a> values. Both arguments must have identical
1849types.</p>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1850<h5>Semantics:</h5>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]1851<p>This instruction returns the <i>remainder</i> of a division (where the result
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1852has the same sign as the divisor), not the <i>modulus</i> (where the
1853result has the same sign as the dividend) of a value. For more
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]1854information about the difference, see <a
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1855 href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
1856Math Forum</a>.</p>
1857<h5>Example:</h5>
Reid Spencer7eb55b32006-11-02 01:53:59 +0000[diff] [blame]1858<pre> &lt;result&gt; = srem int 4, %var <i>; yields {int}:result = 4 % %var</i>
1859</pre>
1860
1861</div>
1862<!-- _______________________________________________________________________ -->
1863<div class="doc_subsubsection"> <a name="i_frem">'<tt>frem</tt>'
1864Instruction</a> </div>
1865<div class="doc_text">
1866<h5>Syntax:</h5>
1867<pre> &lt;result&gt; = frem &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
1868</pre>
1869<h5>Overview:</h5>
1870<p>The '<tt>frem</tt>' instruction returns the remainder from the
1871division of its two operands.</p>
1872<h5>Arguments:</h5>
1873<p>The two arguments to the '<tt>frem</tt>' instruction must be
1874<a href="#t_floating">floating point</a> values. Both arguments must have
1875identical types.</p>
1876<h5>Semantics:</h5>
1877<p>This instruction returns the <i>remainder</i> of a division.</p>
1878<h5>Example:</h5>
1879<pre> &lt;result&gt; = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1880</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1881</div>
Robert Bocchino820bc75b2006-02-17 21:18:08 +0000[diff] [blame]1882
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1883<!-- ======================================================================= -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1884<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
1885Operations</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1886<div class="doc_text">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1887<p>Bitwise binary operators are used to do various forms of
1888bit-twiddling in a program. They are generally very efficient
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]1889instructions and can commonly be strength reduced from other
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1890instructions. They require two operands, execute an operation on them,
1891and produce a single value. The resulting value of the bitwise binary
1892operators is always the same type as its first operand.</p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1893</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1894<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1895<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
1896Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1897<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1898<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1899<pre> &lt;result&gt; = and &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1900</pre>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1901<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1902<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
1903its two operands.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1904<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1905<p>The two arguments to the '<tt>and</tt>' instruction must be <a
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1906 href="#t_integral">integral</a> values. Both arguments must have
1907identical types.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1908<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1909<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1910<p> </p>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1911<div style="align: center">
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1912<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1913 <tbody>
1914 <tr>
1915 <td>In0</td>
1916 <td>In1</td>
1917 <td>Out</td>
1918 </tr>
1919 <tr>
1920 <td>0</td>
1921 <td>0</td>
1922 <td>0</td>
1923 </tr>
1924 <tr>
1925 <td>0</td>
1926 <td>1</td>
1927 <td>0</td>
1928 </tr>
1929 <tr>
1930 <td>1</td>
1931 <td>0</td>
1932 <td>0</td>
1933 </tr>
1934 <tr>
1935 <td>1</td>
1936 <td>1</td>
1937 <td>1</td>
1938 </tr>
1939 </tbody>
1940</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1941</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1942<h5>Example:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1943<pre> &lt;result&gt; = and int 4, %var <i>; yields {int}:result = 4 &amp; %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1944 &lt;result&gt; = and int 15, 40 <i>; yields {int}:result = 8</i>
1945 &lt;result&gt; = and int 4, 8 <i>; yields {int}:result = 0</i>
1946</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1947</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1948<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1949<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1950<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1951<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1952<pre> &lt;result&gt; = or &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1953</pre>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1954<h5>Overview:</h5>
1955<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
1956or of its two operands.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1957<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1958<p>The two arguments to the '<tt>or</tt>' instruction must be <a
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1959 href="#t_integral">integral</a> values. Both arguments must have
1960identical types.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1961<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]1962<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1963<p> </p>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1964<div style="align: center">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1965<table border="1" cellspacing="0" cellpadding="4">
1966 <tbody>
1967 <tr>
1968 <td>In0</td>
1969 <td>In1</td>
1970 <td>Out</td>
1971 </tr>
1972 <tr>
1973 <td>0</td>
1974 <td>0</td>
1975 <td>0</td>
1976 </tr>
1977 <tr>
1978 <td>0</td>
1979 <td>1</td>
1980 <td>1</td>
1981 </tr>
1982 <tr>
1983 <td>1</td>
1984 <td>0</td>
1985 <td>1</td>
1986 </tr>
1987 <tr>
1988 <td>1</td>
1989 <td>1</td>
1990 <td>1</td>
1991 </tr>
1992 </tbody>
1993</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]1994</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1995<h5>Example:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]1996<pre> &lt;result&gt; = or int 4, %var <i>; yields {int}:result = 4 | %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]1997 &lt;result&gt; = or int 15, 40 <i>; yields {int}:result = 47</i>
1998 &lt;result&gt; = or int 4, 8 <i>; yields {int}:result = 12</i>
1999</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2000</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2001<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2002<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
2003Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2004<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2005<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2006<pre> &lt;result&gt; = xor &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2007</pre>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2008<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2009<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
2010or of its two operands. The <tt>xor</tt> is used to implement the
2011"one's complement" operation, which is the "~" operator in C.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2012<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2013<p>The two arguments to the '<tt>xor</tt>' instruction must be <a
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2014 href="#t_integral">integral</a> values. Both arguments must have
2015identical types.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2016<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2017<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2018<p> </p>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]2019<div style="align: center">
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2020<table border="1" cellspacing="0" cellpadding="4">
2021 <tbody>
2022 <tr>
2023 <td>In0</td>
2024 <td>In1</td>
2025 <td>Out</td>
2026 </tr>
2027 <tr>
2028 <td>0</td>
2029 <td>0</td>
2030 <td>0</td>
2031 </tr>
2032 <tr>
2033 <td>0</td>
2034 <td>1</td>
2035 <td>1</td>
2036 </tr>
2037 <tr>
2038 <td>1</td>
2039 <td>0</td>
2040 <td>1</td>
2041 </tr>
2042 <tr>
2043 <td>1</td>
2044 <td>1</td>
2045 <td>0</td>
2046 </tr>
2047 </tbody>
2048</table>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]2049</div>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2050<p> </p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2051<h5>Example:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2052<pre> &lt;result&gt; = xor int 4, %var <i>; yields {int}:result = 4 ^ %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2053 &lt;result&gt; = xor int 15, 40 <i>; yields {int}:result = 39</i>
2054 &lt;result&gt; = xor int 4, 8 <i>; yields {int}:result = 12</i>
Chris Lattner5ed60612003-09-03 00:41:47 +0000[diff] [blame]2055 &lt;result&gt; = xor int %V, -1 <i>; yields {int}:result = ~%V</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2056</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2057</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2058<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2059<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
2060Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2061<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2062<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2063<pre> &lt;result&gt; = shl &lt;ty&gt; &lt;var1&gt;, ubyte &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2064</pre>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2065<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2066<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
2067the left a specified number of bits.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2068<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2069<p>The first argument to the '<tt>shl</tt>' instruction must be an <a
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2070 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>'
2071type.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2072<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2073<p>The value produced is <tt>var1</tt> * 2<sup><tt>var2</tt></sup>.</p>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2074<h5>Example:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2075<pre> &lt;result&gt; = shl int 4, ubyte %var <i>; yields {int}:result = 4 &lt;&lt; %var</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2076 &lt;result&gt; = shl int 4, ubyte 2 <i>; yields {int}:result = 16</i>
2077 &lt;result&gt; = shl int 1, ubyte 10 <i>; yields {int}:result = 1024</i>
2078</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2079</div>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2080<!-- _______________________________________________________________________ -->
Reid Spencerfdff9382006-11-08 06:47:33 +0000[diff] [blame]2081<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2082Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2083<div class="doc_text">
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2084<h5>Syntax:</h5>
Reid Spencerfdff9382006-11-08 06:47:33 +0000[diff] [blame]2085<pre> &lt;result&gt; = lshr &lt;ty&gt; &lt;var1&gt;, ubyte &lt;var2&gt; <i>; yields {ty}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2086</pre>
Reid Spencerfdff9382006-11-08 06:47:33 +0000[diff] [blame]2087
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2088<h5>Overview:</h5>
Reid Spencerfdff9382006-11-08 06:47:33 +0000[diff] [blame]2089<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
2090operand shifted to the right a specified number of bits.</p>
2091
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2092<h5>Arguments:</h5>
Reid Spencerfdff9382006-11-08 06:47:33 +0000[diff] [blame]2093<p>The first argument to the '<tt>lshr</tt>' instruction must be an <a
2094 href="#t_integer">integer</a> type. The second argument must be an '<tt>ubyte</tt>' type.</p>
2095
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2096<h5>Semantics:</h5>
Reid Spencerfdff9382006-11-08 06:47:33 +0000[diff] [blame]2097<p>This instruction always performs a logical shift right operation, regardless
2098of whether the arguments are unsigned or not. The <tt>var2</tt> most significant
2099bits will be filled with zero bits after the shift.</p>
2100
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2101<h5>Example:</h5>
Reid Spencerfdff9382006-11-08 06:47:33 +0000[diff] [blame]2102<pre>
2103 &lt;result&gt; = lshr uint 4, ubyte 1 <i>; yields {uint}:result = 2</i>
2104 &lt;result&gt; = lshr int 4, ubyte 2 <i>; yields {uint}:result = 1</i>
2105 &lt;result&gt; = lshr sbyte 4, ubyte 3 <i>; yields {sbyte}:result = 0</i>
2106 &lt;result&gt; = lshr sbyte -2, ubyte 1 <i>; yields {sbyte}:result = 0x7FFFFFFF </i>
2107</pre>
2108</div>
2109
2110<!-- ======================================================================= -->
2111<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
2112Instruction</a> </div>
2113<div class="doc_text">
2114
2115<h5>Syntax:</h5>
2116<pre> &lt;result&gt; = ashr &lt;ty&gt; &lt;var1&gt;, ubyte &lt;var2&gt; <i>; yields {ty}:result</i>
2117</pre>
2118
2119<h5>Overview:</h5>
2120<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
2121operand shifted to the right a specified number of bits.</p>
2122
2123<h5>Arguments:</h5>
2124<p>The first argument to the '<tt>ashr</tt>' instruction must be an
2125<a href="#t_integer">integer</a> type. The second argument must be an
2126'<tt>ubyte</tt>' type.</p>
2127
2128<h5>Semantics:</h5>
2129<p>This instruction always performs an arithmetic shift right operation,
2130regardless of whether the arguments are signed or not. The <tt>var2</tt> most
2131significant bits will be filled with the sign bit of <tt>var1</tt>.</p>
2132
2133<h5>Example:</h5>
2134<pre>
2135 &lt;result&gt; = ashr uint 4, ubyte 1 <i>; yields {uint}:result = 2</i>
2136 &lt;result&gt; = ashr int 4, ubyte 2 <i>; yields {int}:result = 1</i>
2137 &lt;result&gt; = ashr ubyte 4, ubyte 3 <i>; yields {ubyte}:result = 0</i>
2138 &lt;result&gt; = ashr sbyte -2, ubyte 1 <i>; yields {sbyte}:result = -1</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2139</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2140</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2141
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2142<!-- ======================================================================= -->
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2143<div class="doc_subsection">
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]2144 <a name="vectorops">Vector Operations</a>
2145</div>
2146
2147<div class="doc_text">
2148
2149<p>LLVM supports several instructions to represent vector operations in a
2150target-independent manner. This instructions cover the element-access and
2151vector-specific operations needed to process vectors effectively. While LLVM
2152does directly support these vector operations, many sophisticated algorithms
2153will want to use target-specific intrinsics to take full advantage of a specific
2154target.</p>
2155
2156</div>
2157
2158<!-- _______________________________________________________________________ -->
2159<div class="doc_subsubsection">
2160 <a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
2161</div>
2162
2163<div class="doc_text">
2164
2165<h5>Syntax:</h5>
2166
2167<pre>
2168 &lt;result&gt; = extractelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, uint &lt;idx&gt; <i>; yields &lt;ty&gt;</i>
2169</pre>
2170
2171<h5>Overview:</h5>
2172
2173<p>
2174The '<tt>extractelement</tt>' instruction extracts a single scalar
2175element from a packed vector at a specified index.
2176</p>
2177
2178
2179<h5>Arguments:</h5>
2180
2181<p>
2182The first operand of an '<tt>extractelement</tt>' instruction is a
2183value of <a href="#t_packed">packed</a> type. The second operand is
2184an index indicating the position from which to extract the element.
2185The index may be a variable.</p>
2186
2187<h5>Semantics:</h5>
2188
2189<p>
2190The result is a scalar of the same type as the element type of
2191<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
2192<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
2193results are undefined.
2194</p>
2195
2196<h5>Example:</h5>
2197
2198<pre>
2199 %result = extractelement &lt;4 x int&gt; %vec, uint 0 <i>; yields int</i>
2200</pre>
2201</div>
2202
2203
2204<!-- _______________________________________________________________________ -->
2205<div class="doc_subsubsection">
2206 <a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
2207</div>
2208
2209<div class="doc_text">
2210
2211<h5>Syntax:</h5>
2212
2213<pre>
2214 &lt;result&gt; = insertelement &lt;n x &lt;ty&gt;&gt; &lt;val&gt;, &lt;ty&gt; &lt;elt&gt, uint &lt;idx&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
2215</pre>
2216
2217<h5>Overview:</h5>
2218
2219<p>
2220The '<tt>insertelement</tt>' instruction inserts a scalar
2221element into a packed vector at a specified index.
2222</p>
2223
2224
2225<h5>Arguments:</h5>
2226
2227<p>
2228The first operand of an '<tt>insertelement</tt>' instruction is a
2229value of <a href="#t_packed">packed</a> type. The second operand is a
2230scalar value whose type must equal the element type of the first
2231operand. The third operand is an index indicating the position at
2232which to insert the value. The index may be a variable.</p>
2233
2234<h5>Semantics:</h5>
2235
2236<p>
2237The result is a packed vector of the same type as <tt>val</tt>. Its
2238element values are those of <tt>val</tt> except at position
2239<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
2240exceeds the length of <tt>val</tt>, the results are undefined.
2241</p>
2242
2243<h5>Example:</h5>
2244
2245<pre>
2246 %result = insertelement &lt;4 x int&gt; %vec, int 1, uint 0 <i>; yields &lt;4 x int&gt;</i>
2247</pre>
2248</div>
2249
2250<!-- _______________________________________________________________________ -->
2251<div class="doc_subsubsection">
2252 <a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
2253</div>
2254
2255<div class="doc_text">
2256
2257<h5>Syntax:</h5>
2258
2259<pre>
2260 &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 uint&gt; &lt;mask&gt; <i>; yields &lt;n x &lt;ty&gt;&gt;</i>
2261</pre>
2262
2263<h5>Overview:</h5>
2264
2265<p>
2266The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
2267from two input vectors, returning a vector of the same type.
2268</p>
2269
2270<h5>Arguments:</h5>
2271
2272<p>
2273The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
2274with types that match each other and types that match the result of the
2275instruction. The third argument is a shuffle mask, which has the same number
2276of elements as the other vector type, but whose element type is always 'uint'.
2277</p>
2278
2279<p>
2280The shuffle mask operand is required to be a constant vector with either
2281constant integer or undef values.
2282</p>
2283
2284<h5>Semantics:</h5>
2285
2286<p>
2287The elements of the two input vectors are numbered from left to right across
2288both of the vectors. The shuffle mask operand specifies, for each element of
2289the result vector, which element of the two input registers the result element
2290gets. The element selector may be undef (meaning "don't care") and the second
2291operand may be undef if performing a shuffle from only one vector.
2292</p>
2293
2294<h5>Example:</h5>
2295
2296<pre>
2297 %result = shufflevector &lt;4 x int&gt; %v1, &lt;4 x int&gt; %v2,
2298 &lt;4 x uint&gt; &lt;uint 0, uint 4, uint 1, uint 5&gt; <i>; yields &lt;4 x int&gt;</i>
2299 %result = shufflevector &lt;4 x int&gt; %v1, &lt;4 x int&gt; undef,
2300 &lt;4 x uint&gt; &lt;uint 0, uint 1, uint 2, uint 3&gt; <i>; yields &lt;4 x int&gt;</i> - Identity shuffle.
2301</pre>
2302</div>
2303
Tanya Lattnerb138bbe2006-04-14 19:24:33 +0000[diff] [blame]2304
Chris Lattnerce83bff2006-04-08 23:07:04 +0000[diff] [blame]2305<!-- ======================================================================= -->
2306<div class="doc_subsection">
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]2307 <a name="memoryops">Memory Access and Addressing Operations</a>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2308</div>
2309
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2310<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2311
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2312<p>A key design point of an SSA-based representation is how it
2313represents memory. In LLVM, no memory locations are in SSA form, which
2314makes things very simple. This section describes how to read, write,
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]2315allocate, and free memory in LLVM.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2316
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2317</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2318
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2319<!-- _______________________________________________________________________ -->
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2320<div class="doc_subsubsection">
2321 <a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
2322</div>
2323
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2324<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2325
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2326<h5>Syntax:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2327
2328<pre>
2329 &lt;result&gt; = malloc &lt;type&gt;[, uint &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2330</pre>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2331
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2332<h5>Overview:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2333
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2334<p>The '<tt>malloc</tt>' instruction allocates memory from the system
2335heap and returns a pointer to it.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2336
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2337<h5>Arguments:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2338
2339<p>The '<tt>malloc</tt>' instruction allocates
2340<tt>sizeof(&lt;type&gt;)*NumElements</tt>
John Criswella92e5862004-02-24 16:13:56 +0000[diff] [blame]2341bytes of memory from the operating system and returns a pointer of the
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2342appropriate type to the program. If "NumElements" is specified, it is the
2343number of elements allocated. If an alignment is specified, the value result
2344of the allocation is guaranteed to be aligned to at least that boundary. If
2345not specified, or if zero, the target can choose to align the allocation on any
2346convenient boundary.</p>
2347
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2348<p>'<tt>type</tt>' must be a sized type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2349
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2350<h5>Semantics:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2351
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2352<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
2353a pointer is returned.</p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2354
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2355<h5>Example:</h5>
2356
2357<pre>
2358 %array = malloc [4 x ubyte ] <i>; yields {[%4 x ubyte]*}:array</i>
2359
2360 %size = <a href="#i_add">add</a> uint 2, 2 <i>; yields {uint}:size = uint 4</i>
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2361 %array1 = malloc ubyte, uint 4 <i>; yields {ubyte*}:array1</i>
2362 %array2 = malloc [12 x ubyte], uint %size <i>; yields {[12 x ubyte]*}:array2</i>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2363 %array3 = malloc int, uint 4, align 1024 <i>; yields {int*}:array3</i>
2364 %array4 = malloc int, align 1024 <i>; yields {int*}:array4</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2365</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2366</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2367
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2368<!-- _______________________________________________________________________ -->
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2369<div class="doc_subsubsection">
2370 <a name="i_free">'<tt>free</tt>' Instruction</a>
2371</div>
2372
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2373<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2374
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2375<h5>Syntax:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2376
2377<pre>
2378 free &lt;type&gt; &lt;value&gt; <i>; yields {void}</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2379</pre>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2380
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2381<h5>Overview:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2382
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2383<p>The '<tt>free</tt>' instruction returns memory back to the unused
John Criswell4a3327e2005-05-13 22:25:59 +0000[diff] [blame]2384memory heap to be reallocated in the future.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2385
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2386<h5>Arguments:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2387
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2388<p>'<tt>value</tt>' shall be a pointer value that points to a value
2389that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
2390instruction.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2391
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2392<h5>Semantics:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2393
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]2394<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2395after this instruction executes.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2396
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2397<h5>Example:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2398
2399<pre>
2400 %array = <a href="#i_malloc">malloc</a> [4 x ubyte] <i>; yields {[4 x ubyte]*}:array</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2401 free [4 x ubyte]* %array
2402</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2403</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2404
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2405<!-- _______________________________________________________________________ -->
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2406<div class="doc_subsubsection">
2407 <a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
2408</div>
2409
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2410<div class="doc_text">
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2411
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2412<h5>Syntax:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2413
2414<pre>
2415 &lt;result&gt; = alloca &lt;type&gt;[, uint &lt;NumElements&gt;][, align &lt;alignment&gt;] <i>; yields {type*}:result</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2416</pre>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2417
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2418<h5>Overview:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2419
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2420<p>The '<tt>alloca</tt>' instruction allocates memory on the current
2421stack frame of the procedure that is live until the current function
2422returns to its caller.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2423
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2424<h5>Arguments:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2425
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]2426<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(&lt;type&gt;)*NumElements</tt>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2427bytes of memory on the runtime stack, returning a pointer of the
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2428appropriate type to the program. If "NumElements" is specified, it is the
2429number of elements allocated. If an alignment is specified, the value result
2430of the allocation is guaranteed to be aligned to at least that boundary. If
2431not specified, or if zero, the target can choose to align the allocation on any
2432convenient boundary.</p>
2433
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2434<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2435
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2436<h5>Semantics:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2437
John Criswell4a3327e2005-05-13 22:25:59 +0000[diff] [blame]2438<p>Memory is allocated; a pointer is returned. '<tt>alloca</tt>'d
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2439memory is automatically released when the function returns. The '<tt>alloca</tt>'
2440instruction is commonly used to represent automatic variables that must
2441have an address available. When the function returns (either with the <tt><a
John Criswellc932bef2005-05-12 16:55:34 +0000[diff] [blame]2442 href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2443instructions), the memory is reclaimed.</p>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2444
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2445<h5>Example:</h5>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2446
2447<pre>
2448 %ptr = alloca int <i>; yields {int*}:ptr</i>
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2449 %ptr = alloca int, uint 4 <i>; yields {int*}:ptr</i>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2450 %ptr = alloca int, uint 4, align 1024 <i>; yields {int*}:ptr</i>
2451 %ptr = alloca int, align 1024 <i>; yields {int*}:ptr</i>
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2452</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2453</div>
Chris Lattner54611b42005-11-06 08:02:57 +0000[diff] [blame]2454
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2455<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2456<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
2457Instruction</a> </div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2458<div class="doc_text">
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]2459<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2460<pre> &lt;result&gt; = load &lt;ty&gt;* &lt;pointer&gt;<br> &lt;result&gt; = volatile load &lt;ty&gt;* &lt;pointer&gt;<br></pre>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]2461<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2462<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]2463<h5>Arguments:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2464<p>The argument to the '<tt>load</tt>' instruction specifies the memory
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]2465address from which to load. The pointer must point to a <a
Chris Lattner10ee9652004-06-03 22:57:15 +0000[diff] [blame]2466 href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]2467marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2468the number or order of execution of this <tt>load</tt> with other
2469volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
2470instructions. </p>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]2471<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2472<p>The location of memory pointed to is loaded.</p>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]2473<h5>Examples:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2474<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
2475 <a
2476 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]2477 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
2478</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2479</div>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]2480<!-- _______________________________________________________________________ -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2481<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
2482Instruction</a> </div>
Reid Spencera89fb182006-11-09 21:18:01 +0000[diff] [blame]2483<div class="doc_text">
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]2484<h5>Syntax:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2485<pre> store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt; <i>; yields {void}</i>
Chris Lattner12d456c2003-09-08 18:27:49 +0000[diff] [blame]2486 volatile store &lt;ty&gt; &lt;value&gt;, &lt;ty&gt;* &lt;pointer&gt; <i>; yields {void}</i>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]2487</pre>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]2488<h5>Overview:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2489<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]2490<h5>Arguments:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2491<p>There are two arguments to the '<tt>store</tt>' instruction: a value
John Criswell4c0cf7f2005-10-24 16:17:18 +0000[diff] [blame]2492to store and an address in which to store it. The type of the '<tt>&lt;pointer&gt;</tt>'
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2493operand must be a pointer to the type of the '<tt>&lt;value&gt;</tt>'
John Criswell4a3327e2005-05-13 22:25:59 +0000[diff] [blame]2494operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2495optimizer is not allowed to modify the number or order of execution of
2496this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
2497 href="#i_store">store</a></tt> instructions.</p>
2498<h5>Semantics:</h5>
2499<p>The contents of memory are updated to contain '<tt>&lt;value&gt;</tt>'
2500at the location specified by the '<tt>&lt;pointer&gt;</tt>' operand.</p>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]2501<h5>Example:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2502<pre> %ptr = <a href="#i_alloca">alloca</a> int <i>; yields {int*}:ptr</i>
2503 <a
2504 href="#i_store">store</a> int 3, int* %ptr <i>; yields {void}</i>
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]2505 %val = load int* %ptr <i>; yields {int}:val = int 3</i>
2506</pre>
Reid Spencer443460a2006-11-09 21:15:49 +0000[diff] [blame]2507</div>
2508
Chris Lattner095735d2002-05-06 03:03:22 +0000[diff] [blame]2509<!-- _______________________________________________________________________ -->
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2510<div class="doc_subsubsection">
2511 <a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
2512</div>
2513
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2514<div class="doc_text">
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2515<h5>Syntax:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2516<pre>
2517 &lt;result&gt; = getelementptr &lt;ty&gt;* &lt;ptrval&gt;{, &lt;ty&gt; &lt;idx&gt;}*
2518</pre>
2519
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2520<h5>Overview:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2521
2522<p>
2523The '<tt>getelementptr</tt>' instruction is used to get the address of a
2524subelement of an aggregate data structure.</p>
2525
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2526<h5>Arguments:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2527
Reid Spencercee005c2006-12-04 21:29:24 +0000[diff] [blame]2528<p>This instruction takes a list of integer operands that indicate what
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2529elements of the aggregate object to index to. The actual types of the arguments
2530provided depend on the type of the first pointer argument. The
2531'<tt>getelementptr</tt>' instruction is used to index down through the type
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]2532levels of a structure or to a specific index in an array. When indexing into a
Reid Spencerc0312692006-12-03 16:53:48 +0000[diff] [blame]2533structure, only <tt>uint</tt> integer constants are allowed. When indexing
Reid Spencercee005c2006-12-04 21:29:24 +0000[diff] [blame]2534into an array or pointer, only integers of 32 or 64 bits are allowed, and will
2535be sign extended to 64-bit values.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2536
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2537<p>For example, let's consider a C code fragment and how it gets
2538compiled to LLVM:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2539
2540<pre>
2541 struct RT {
2542 char A;
2543 int B[10][20];
2544 char C;
2545 };
2546 struct ST {
2547 int X;
2548 double Y;
2549 struct RT Z;
2550 };
2551
2552 int *foo(struct ST *s) {
2553 return &amp;s[1].Z.B[5][13];
2554 }
2555</pre>
2556
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2557<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2558
2559<pre>
2560 %RT = type { sbyte, [10 x [20 x int]], sbyte }
2561 %ST = type { int, double, %RT }
2562
Brian Gaeke317ef962004-07-02 21:08:14 +0000[diff] [blame]2563 implementation
2564
2565 int* %foo(%ST* %s) {
2566 entry:
2567 %reg = getelementptr %ST* %s, int 1, uint 2, uint 1, int 5, int 13
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2568 ret int* %reg
2569 }
2570</pre>
2571
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2572<h5>Semantics:</h5>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2573
2574<p>The index types specified for the '<tt>getelementptr</tt>' instruction depend
John Criswell4a3327e2005-05-13 22:25:59 +0000[diff] [blame]2575on the pointer type that is being indexed into. <a href="#t_pointer">Pointer</a>
Reid Spencercee005c2006-12-04 21:29:24 +0000[diff] [blame]2576and <a href="#t_array">array</a> types can use a 32-bit or 64-bit
Reid Spencerc0312692006-12-03 16:53:48 +0000[diff] [blame]2577<a href="#t_integer">integer</a> type but the value will always be sign extended
2578to 64-bits. <a href="#t_struct">Structure</a> types, require <tt>uint</tt>
2579<b>constants</b>.</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2580
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2581<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2582type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ int, double, %RT
2583}</tt>' type, a structure. The second index indexes into the third element of
2584the structure, yielding a '<tt>%RT</tt>' = '<tt>{ sbyte, [10 x [20 x int]],
2585sbyte }</tt>' type, another structure. The third index indexes into the second
2586element of the structure, yielding a '<tt>[10 x [20 x int]]</tt>' type, an
2587array. The two dimensions of the array are subscripted into, yielding an
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]2588'<tt>int</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2589to this element, thus computing a value of '<tt>int*</tt>' type.</p>
2590
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]2591<p>Note that it is perfectly legal to index partially through a
2592structure, returning a pointer to an inner element. Because of this,
2593the LLVM code for the given testcase is equivalent to:</p>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2594
2595<pre>
Chris Lattner455fc8c2005-03-07 22:13:59 +0000[diff] [blame]2596 int* %foo(%ST* %s) {
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2597 %t1 = getelementptr %ST* %s, int 1 <i>; yields %ST*:%t1</i>
2598 %t2 = getelementptr %ST* %t1, int 0, uint 2 <i>; yields %RT*:%t2</i>
2599 %t3 = getelementptr %RT* %t2, int 0, uint 1 <i>; yields [10 x [20 x int]]*:%t3</i>
2600 %t4 = getelementptr [10 x [20 x int]]* %t3, int 0, int 5 <i>; yields [20 x int]*:%t4</i>
2601 %t5 = getelementptr [20 x int]* %t4, int 0, int 13 <i>; yields int*:%t5</i>
2602 ret int* %t5
2603 }
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]2604</pre>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]2605
2606<p>Note that it is undefined to access an array out of bounds: array and
2607pointer indexes must always be within the defined bounds of the array type.
2608The one exception for this rules is zero length arrays. These arrays are
2609defined to be accessible as variable length arrays, which requires access
2610beyond the zero'th element.</p>
2611
Chris Lattner6ab66722006-08-15 00:45:58 +0000[diff] [blame]2612<p>The getelementptr instruction is often confusing. For some more insight
2613into how it works, see <a href="GetElementPtr.html">the getelementptr
2614FAQ</a>.</p>
2615
Chris Lattner590645f2002-04-14 06:13:44 +0000[diff] [blame]2616<h5>Example:</h5>
Chris Lattnerc0ad71e2005-06-24 17:22:57 +0000[diff] [blame]2617
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2618<pre>
2619 <i>; yields [12 x ubyte]*:aptr</i>
2620 %aptr = getelementptr {int, [12 x ubyte]}* %sptr, long 0, uint 1
2621</pre>
Chris Lattner33fd7022004-04-05 01:30:49 +0000[diff] [blame]2622</div>
Reid Spencer443460a2006-11-09 21:15:49 +0000[diff] [blame]2623
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]2624<!-- ======================================================================= -->
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]2625<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2626</div>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2627<div class="doc_text">
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]2628<p>The instructions in this category are the conversion instructions (casting)
2629which all take a single operand and a type. They perform various bit conversions
2630on the operand.</p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]2631</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]2632
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]2633<!-- _______________________________________________________________________ -->
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]2634<div class="doc_subsubsection">
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2635 <a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
2636</div>
2637<div class="doc_text">
2638
2639<h5>Syntax:</h5>
2640<pre>
2641 &lt;result&gt; = trunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2642</pre>
2643
2644<h5>Overview:</h5>
2645<p>
2646The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
2647</p>
2648
2649<h5>Arguments:</h5>
2650<p>
2651The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
2652be an <a href="#t_integer">integer</a> type, and a type that specifies the size
2653and type of the result, which must be an <a href="#t_integral">integral</a>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2654type. The bit size of <tt>value</tt> must be larger than the bit size of
2655<tt>ty2</tt>. Equal sized types are not allowed.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2656
2657<h5>Semantics:</h5>
2658<p>
2659The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2660and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
2661larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
2662It will always truncate bits.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2663
2664<h5>Example:</h5>
2665<pre>
2666 %X = trunc int 257 to ubyte <i>; yields ubyte:1</i>
2667 %Y = trunc int 123 to bool <i>; yields bool:true</i>
2668</pre>
2669</div>
2670
2671<!-- _______________________________________________________________________ -->
2672<div class="doc_subsubsection">
2673 <a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
2674</div>
2675<div class="doc_text">
2676
2677<h5>Syntax:</h5>
2678<pre>
2679 &lt;result&gt; = zext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2680</pre>
2681
2682<h5>Overview:</h5>
2683<p>The '<tt>zext</tt>' instruction zero extends its operand to type
2684<tt>ty2</tt>.</p>
2685
2686
2687<h5>Arguments:</h5>
2688<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
2689<a href="#t_integral">integral</a> type, and a type to cast it to, which must
2690also be of <a href="#t_integral">integral</a> type. The bit size of the
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2691<tt>value</tt> must be smaller than the bit size of the destination type,
2692<tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2693
2694<h5>Semantics:</h5>
2695<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
2696bits until it reaches the size of the destination type, <tt>ty2</tt>. When the
2697the operand and the type are the same size, no bit filling is done and the
2698cast is considered a <i>no-op cast</i> because no bits change (only the type
2699changes).</p>
2700
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2701<p>When zero extending from bool, the result will alwasy be either 0 or 1.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2702
2703<h5>Example:</h5>
2704<pre>
2705 %X = zext int 257 to ulong <i>; yields ulong:257</i>
2706 %Y = zext bool true to int <i>; yields int:1</i>
2707</pre>
2708</div>
2709
2710<!-- _______________________________________________________________________ -->
2711<div class="doc_subsubsection">
2712 <a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
2713</div>
2714<div class="doc_text">
2715
2716<h5>Syntax:</h5>
2717<pre>
2718 &lt;result&gt; = sext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2719</pre>
2720
2721<h5>Overview:</h5>
2722<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
2723
2724<h5>Arguments:</h5>
2725<p>
2726The '<tt>sext</tt>' instruction takes a value to cast, which must be of
2727<a href="#t_integral">integral</a> type, and a type to cast it to, which must
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2728also be of <a href="#t_integral">integral</a> type. The bit size of the
2729<tt>value</tt> must be smaller than the bit size of the destination type,
2730<tt>ty2</tt>.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2731
2732<h5>Semantics:</h5>
2733<p>
2734The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
2735bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
2736the type <tt>ty2</tt>. When the the operand and the type are the same size,
2737no bit filling is done and the cast is considered a <i>no-op cast</i> because
2738no bits change (only the type changes).</p>
2739
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2740<p>When sign extending from bool, the extension always results in -1 or 0.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2741
2742<h5>Example:</h5>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2743<pre>
2744 %X = sext sbyte -1 to ushort <i>; yields ushort:65535</i>
2745 %Y = sext bool true to int <i>; yields int:-1</i>
2746</pre>
2747</div>
2748
2749<!-- _______________________________________________________________________ -->
2750<div class="doc_subsubsection">
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]2751 <a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
2752</div>
2753
2754<div class="doc_text">
2755
2756<h5>Syntax:</h5>
2757
2758<pre>
2759 &lt;result&gt; = fptrunc &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2760</pre>
2761
2762<h5>Overview:</h5>
2763<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
2764<tt>ty2</tt>.</p>
2765
2766
2767<h5>Arguments:</h5>
2768<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
2769 point</a> value to cast and a <a href="#t_floating">floating point</a> type to
2770cast it to. The size of <tt>value</tt> must be larger than the size of
2771<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
2772<i>no-op cast</i>.</p>
2773
2774<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2775<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
2776<a href="#t_floating">floating point</a> type to a smaller
2777<a href="#t_floating">floating point</a> type. If the value cannot fit within
2778the destination type, <tt>ty2</tt>, then the results are undefined.</p>
Reid Spencer2e2740d2006-11-09 21:48:10 +0000[diff] [blame]2779
2780<h5>Example:</h5>
2781<pre>
2782 %X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
2783 %Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
2784</pre>
2785</div>
2786
2787<!-- _______________________________________________________________________ -->
2788<div class="doc_subsubsection">
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2789 <a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
2790</div>
2791<div class="doc_text">
2792
2793<h5>Syntax:</h5>
2794<pre>
2795 &lt;result&gt; = fpext &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2796</pre>
2797
2798<h5>Overview:</h5>
2799<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
2800floating point value.</p>
2801
2802<h5>Arguments:</h5>
2803<p>The '<tt>fpext</tt>' instruction takes a
2804<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2805and a <a href="#t_floating">floating point</a> type to cast it to. The source
2806type must be smaller than the destination type.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2807
2808<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2809<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
2810<a href="t_floating">floating point</a> type to a larger
2811<a href="t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
2812used to make a <i>no-op cast</i> because it always changes bits. Use
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]2813<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2814
2815<h5>Example:</h5>
2816<pre>
2817 %X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
2818 %Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
2819</pre>
2820</div>
2821
2822<!-- _______________________________________________________________________ -->
2823<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2824 <a name="i_fp2uint">'<tt>fptoui .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2825</div>
2826<div class="doc_text">
2827
2828<h5>Syntax:</h5>
2829<pre>
2830 &lt;result&gt; = fp2uint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2831</pre>
2832
2833<h5>Overview:</h5>
2834<p>The '<tt>fp2uint</tt>' converts a floating point <tt>value</tt> to its
2835unsigned integer equivalent of type <tt>ty2</tt>.
2836</p>
2837
2838<h5>Arguments:</h5>
2839<p>The '<tt>fp2uint</tt>' instruction takes a value to cast, which must be a
2840<a href="#t_floating">floating point</a> value, and a type to cast it to, which
2841must be an <a href="#t_integral">integral</a> type.</p>
2842
2843<h5>Semantics:</h5>
2844<p> The '<tt>fp2uint</tt>' instruction converts its
2845<a href="#t_floating">floating point</a> operand into the nearest (rounding
2846towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
2847the results are undefined.</p>
2848
2849<p>When converting to bool, the conversion is done as a comparison against
2850zero. If the <tt>value</tt> was zero, the bool result will be <tt>false</tt>.
2851If the <tt>value</tt> was non-zero, the bool result will be <tt>true</tt>.</p>
2852
2853<h5>Example:</h5>
2854<pre>
2855 %X = fp2uint double 123.0 to int <i>; yields int:123</i>
2856 %Y = fp2uint float 1.0E+300 to bool <i>; yields bool:true</i>
2857 %X = fp2uint float 1.04E+17 to ubyte <i>; yields undefined:1</i>
2858</pre>
2859</div>
2860
2861<!-- _______________________________________________________________________ -->
2862<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2863 <a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2864</div>
2865<div class="doc_text">
2866
2867<h5>Syntax:</h5>
2868<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2869 &lt;result&gt; = fptosi &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2870</pre>
2871
2872<h5>Overview:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2873<p>The '<tt>fptosi</tt>' instruction converts
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2874<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]2875</p>
2876
2877
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]2878<h5>Arguments:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2879<p> The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2880<a href="#t_floating">floating point</a> value, and a type to cast it to, which
2881must also be an <a href="#t_integral">integral</a> type.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]2882
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]2883<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2884<p>The '<tt>fptosi</tt>' instruction converts its
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2885<a href="#t_floating">floating point</a> operand into the nearest (rounding
2886towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
2887the results are undefined.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]2888
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2889<p>When converting to bool, the conversion is done as a comparison against
2890zero. If the <tt>value</tt> was zero, the bool result will be <tt>false</tt>.
2891If the <tt>value</tt> was non-zero, the bool result will be <tt>true</tt>.</p>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]2892
Chris Lattner70de6632001-07-09 00:26:23 +0000[diff] [blame]2893<h5>Example:</h5>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]2894<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2895 %X = fptosi double -123.0 to int <i>; yields int:-123</i>
2896 %Y = fptosi float 1.0E-247 to bool <i>; yields bool:true</i>
2897 %X = fptosi float 1.04E+17 to sbyte <i>; yields undefined:1</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2898</pre>
2899</div>
2900
2901<!-- _______________________________________________________________________ -->
2902<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2903 <a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2904</div>
2905<div class="doc_text">
2906
2907<h5>Syntax:</h5>
2908<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2909 &lt;result&gt; = uitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2910</pre>
2911
2912<h5>Overview:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2913<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2914integer and converts that value to the <tt>ty2</tt> type.</p>
2915
2916
2917<h5>Arguments:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2918<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be an
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2919<a href="#t_integral">integral</a> value, and a type to cast it to, which must
2920be a <a href="#t_floating">floating point</a> type.</p>
2921
2922<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2923<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2924integer quantity and converts it to the corresponding floating point value. If
2925the value cannot fit in the floating point value, the results are undefined.</p>
2926
2927
2928<h5>Example:</h5>
2929<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2930 %X = uitofp int 257 to float <i>; yields float:257.0</i>
2931 %Y = uitofp sbyte -1 to double <i>; yields double:255.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2932</pre>
2933</div>
2934
2935<!-- _______________________________________________________________________ -->
2936<div class="doc_subsubsection">
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2937 <a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2938</div>
2939<div class="doc_text">
2940
2941<h5>Syntax:</h5>
2942<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2943 &lt;result&gt; = sitofp &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2944</pre>
2945
2946<h5>Overview:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2947<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2948integer and converts that value to the <tt>ty2</tt> type.</p>
2949
2950<h5>Arguments:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2951<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be an
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2952<a href="#t_integral">integral</a> value, and a type to cast it to, which must be
2953a <a href="#t_floating">floating point</a> type.</p>
2954
2955<h5>Semantics:</h5>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2956<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2957integer quantity and converts it to the corresponding floating point value. If
2958the value cannot fit in the floating point value, the results are undefined.</p>
2959
2960<h5>Example:</h5>
2961<pre>
Reid Spencer51b07252006-11-09 23:03:26 +0000[diff] [blame]2962 %X = sitofp int 257 to float <i>; yields float:257.0</i>
2963 %Y = sitofp sbyte -1 to double <i>; yields double:-1.0</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]2964</pre>
2965</div>
2966
2967<!-- _______________________________________________________________________ -->
2968<div class="doc_subsubsection">
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]2969 <a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
2970</div>
2971<div class="doc_text">
2972
2973<h5>Syntax:</h5>
2974<pre>
2975 &lt;result&gt; = ptrtoint &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
2976</pre>
2977
2978<h5>Overview:</h5>
2979<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
2980the integer type <tt>ty2</tt>.</p>
2981
2982<h5>Arguments:</h5>
2983<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
2984must be a <a href="t_pointer">pointer</a> value, and a type to cast it to
2985<tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.
2986
2987<h5>Semantics:</h5>
2988<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
2989<tt>ty2</tt> by interpreting the pointer value as an integer and either
2990truncating or zero extending that value to the size of the integer type. If
2991<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
2992<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
2993are the same size, then nothing is done (<i>no-op cast</i>).</p>
2994
2995<h5>Example:</h5>
2996<pre>
2997 %X = ptrtoint int* %X to sbyte <i>; yields truncation on 32-bit</i>
2998 %Y = ptrtoint int* %x to ulong <i>; yields zero extend on 32-bit</i>
2999</pre>
3000</div>
3001
3002<!-- _______________________________________________________________________ -->
3003<div class="doc_subsubsection">
3004 <a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
3005</div>
3006<div class="doc_text">
3007
3008<h5>Syntax:</h5>
3009<pre>
3010 &lt;result&gt; = inttoptr &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
3011</pre>
3012
3013<h5>Overview:</h5>
3014<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
3015a pointer type, <tt>ty2</tt>.</p>
3016
3017<h5>Arguments:</h5>
3018<p>The '<tt>inttoptr</tt>' instruction takes an <a href="i_integer">integer</a>
3019value to cast, and a type to cast it to, which must be a
3020<a href="#t_pointer">pointer</a> type. </tt>
3021
3022<h5>Semantics:</h5>
3023<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
3024<tt>ty2</tt> by applying either a zero extension or a truncation depending on
3025the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
3026size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
3027the size of a pointer then a zero extension is done. If they are the same size,
3028nothing is done (<i>no-op cast</i>).</p>
3029
3030<h5>Example:</h5>
3031<pre>
3032 %X = inttoptr int 255 to int* <i>; yields zero extend on 64-bit</i>
3033 %X = inttoptr int 255 to int* <i>; yields no-op on 32-bit </i>
3034 %Y = inttoptr short 0 to int* <i>; yields zero extend on 32-bit</i>
3035</pre>
3036</div>
3037
3038<!-- _______________________________________________________________________ -->
3039<div class="doc_subsubsection">
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]3040 <a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]3041</div>
3042<div class="doc_text">
3043
3044<h5>Syntax:</h5>
3045<pre>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]3046 &lt;result&gt; = bitcast &lt;ty&gt; &lt;value&gt; to &lt;ty2&gt; <i>; yields ty2</i>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]3047</pre>
3048
3049<h5>Overview:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]3050<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]3051<tt>ty2</tt> without changing any bits.</p>
3052
3053<h5>Arguments:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]3054<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]3055a first class value, and a type to cast it to, which must also be a <a
3056 href="#t_firstclass">first class</a> type. The bit sizes of <tt>value</tt>
3057and the destination type, <tt>ty2</tt>, must be identical.</p>
3058
3059<h5>Semantics:</h5>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]3060<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Reid Spencerb7344ff2006-11-11 21:00:47 +0000[diff] [blame]3061<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
3062this conversion. The conversion is done as if the <tt>value</tt> had been
3063stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
3064converted to other pointer types with this instruction. To convert pointers to
3065other types, use the <a href="#i_inttoptr">inttoptr</a> or
3066<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer59b6b7d2006-11-08 01:11:31 +0000[diff] [blame]3067
3068<h5>Example:</h5>
3069<pre>
Reid Spencer5b950642006-11-11 23:08:07 +0000[diff] [blame]3070 %X = bitcast ubyte 255 to sbyte <i>; yields sbyte:-1</i>
3071 %Y = bitcast uint* %x to sint* <i>; yields sint*:%x</i>
3072 %Z = bitcast <2xint> %V to long; <i>; yields long: %V</i>
Chris Lattner70de6632001-07-09 00:26:23 +0000[diff] [blame]3073</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3074</div>
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]3075
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]3076<!-- ======================================================================= -->
3077<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
3078<div class="doc_text">
3079<p>The instructions in this category are the "miscellaneous"
3080instructions, which defy better classification.</p>
3081</div>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3082
3083<!-- _______________________________________________________________________ -->
3084<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
3085</div>
3086<div class="doc_text">
3087<h5>Syntax:</h5>
3088<pre> &lt;result&gt; = icmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
3089</pre>
3090<h5>Overview:</h5>
3091<p>The '<tt>icmp</tt>' instruction returns a boolean value based on comparison
3092of its two integer operands.</p>
3093<h5>Arguments:</h5>
3094<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
3095the condition code which indicates the kind of comparison to perform. It is not
3096a value, just a keyword. The possibilities for the condition code are:
3097<ol>
3098 <li><tt>eq</tt>: equal</li>
3099 <li><tt>ne</tt>: not equal </li>
3100 <li><tt>ugt</tt>: unsigned greater than</li>
3101 <li><tt>uge</tt>: unsigned greater or equal</li>
3102 <li><tt>ult</tt>: unsigned less than</li>
3103 <li><tt>ule</tt>: unsigned less or equal</li>
3104 <li><tt>sgt</tt>: signed greater than</li>
3105 <li><tt>sge</tt>: signed greater or equal</li>
3106 <li><tt>slt</tt>: signed less than</li>
3107 <li><tt>sle</tt>: signed less or equal</li>
3108</ol>
3109<p>The remaining two arguments must be of <a href="#t_integral">integral</a>,
3110<a href="#t_pointer">pointer</a> or a <a href="#t_packed">packed</a> integral
3111type. They must have identical types.</p>
3112<h5>Semantics:</h5>
3113<p>The '<tt>icmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3114the condition code given as <tt>cond</tt>. The comparison performed always
3115yields a <a href="#t_bool">bool</a> result, as follows:
3116<ol>
3117 <li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
3118 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3119 </li>
3120 <li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
3121 <tt>false</tt> otherwise. No sign interpretation is necessary or performed.
3122 <li><tt>ugt</tt>: interprets the operands as unsigned values and yields
3123 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3124 <li><tt>uge</tt>: interprets the operands as unsigned values and yields
3125 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3126 <li><tt>ult</tt>: interprets the operands as unsigned values and yields
3127 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3128 <li><tt>ule</tt>: interprets the operands as unsigned values and yields
3129 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3130 <li><tt>sgt</tt>: interprets the operands as signed values and yields
3131 <tt>true</tt> if <tt>var1</tt> is greater than <tt>var2</tt>.</li>
3132 <li><tt>sge</tt>: interprets the operands as signed values and yields
3133 <tt>true</tt> if <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
3134 <li><tt>slt</tt>: interprets the operands as signed values and yields
3135 <tt>true</tt> if <tt>var1</tt> is less than <tt>var2</tt>.</li>
3136 <li><tt>sle</tt>: interprets the operands as signed values and yields
3137 <tt>true</tt> if <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
3138 </li>
3139</ol>
3140<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
3141values are treated as integers and then compared.</p>
3142<p>If the operands are <a href="#t_packed">packed</a> typed, the elements of
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3143the vector are compared in turn and the predicate must hold for all
3144elements.</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3145
3146<h5>Example:</h5>
3147<pre> &lt;result&gt; = icmp eq int 4, 5 <i>; yields: result=false</i>
3148 &lt;result&gt; = icmp ne float* %X, %X <i>; yields: result=false</i>
3149 &lt;result&gt; = icmp ult short 4, 5 <i>; yields: result=true</i>
3150 &lt;result&gt; = icmp sgt sbyte 4, 5 <i>; yields: result=false</i>
3151 &lt;result&gt; = icmp ule sbyte -4, 5 <i>; yields: result=false</i>
3152 &lt;result&gt; = icmp sge sbyte 4, 5 <i>; yields: result=false</i>
3153</pre>
3154</div>
3155
3156<!-- _______________________________________________________________________ -->
3157<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
3158</div>
3159<div class="doc_text">
3160<h5>Syntax:</h5>
3161<pre> &lt;result&gt; = fcmp &lt;cond&gt; &lt;ty&gt; &lt;var1&gt;, &lt;var2&gt; <i>; yields {bool}:result</i>
3162</pre>
3163<h5>Overview:</h5>
3164<p>The '<tt>fcmp</tt>' instruction returns a boolean value based on comparison
3165of its floating point operands.</p>
3166<h5>Arguments:</h5>
3167<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
3168the condition code which indicates the kind of comparison to perform. It is not
3169a value, just a keyword. The possibilities for the condition code are:
3170<ol>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3171 <li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3172 <li><tt>oeq</tt>: ordered and equal</li>
3173 <li><tt>ogt</tt>: ordered and greater than </li>
3174 <li><tt>oge</tt>: ordered and greater than or equal</li>
3175 <li><tt>olt</tt>: ordered and less than </li>
3176 <li><tt>ole</tt>: ordered and less than or equal</li>
3177 <li><tt>one</tt>: ordered and not equal</li>
3178 <li><tt>ord</tt>: ordered (no nans)</li>
3179 <li><tt>ueq</tt>: unordered or equal</li>
3180 <li><tt>ugt</tt>: unordered or greater than </li>
3181 <li><tt>uge</tt>: unordered or greater than or equal</li>
3182 <li><tt>ult</tt>: unordered or less than </li>
3183 <li><tt>ule</tt>: unordered or less than or equal</li>
3184 <li><tt>une</tt>: unordered or not equal</li>
3185 <li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3186 <li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3187</ol>
Reid Spencer02e0d1d2006-12-06 07:08:07 +0000[diff] [blame]3188<p>In the preceding, <i>ordered</i> means that neither operand is a QNAN while
3189<i>unordered</i> means that either operand may be a QNAN.</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3190<p>The <tt>val1</tt> and <tt>val2</tt> arguments must be of
3191<a href="#t_floating">floating point</a>, or a <a href="#t_packed">packed</a>
3192floating point type. They must have identical types.</p>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3193<p>In the foregoing, <i>ordered</i> means that neither operand is a QNAN and
3194<i>unordered</i> means that either operand is a QNAN.</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3195<h5>Semantics:</h5>
3196<p>The '<tt>fcmp</tt>' compares <tt>var1</tt> and <tt>var2</tt> according to
3197the condition code given as <tt>cond</tt>. The comparison performed always
3198yields a <a href="#t_bool">bool</a> result, as follows:
3199<ol>
3200 <li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3201 <li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3202 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3203 <li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3204 <tt>var1</tt> is greather than <tt>var2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3205 <li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3206 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3207 <li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3208 <tt>var1</tt> is less than <tt>var2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3209 <li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3210 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3211 <li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3212 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3213 <li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
3214 <li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3215 <tt>var1</tt> is equal to <tt>var2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3216 <li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3217 <tt>var1</tt> is greater than <tt>var2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3218 <li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3219 <tt>var1</tt> is greater than or equal to <tt>var2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3220 <li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3221 <tt>var1</tt> is less than <tt>var2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3222 <li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3223 <tt>var1</tt> is less than or equal to <tt>var2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3224 <li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3225 <tt>var1</tt> is not equal to <tt>var2</tt>.</li>
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3226 <li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3227 <li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
3228</ol>
3229<p>If the operands are <a href="#t_packed">packed</a> typed, the elements of
3230the vector are compared in turn and the predicate must hold for all elements.
Reid Spencerf69acf32006-11-19 03:00:14 +0000[diff] [blame]3231</p>
Reid Spencerc828a0e2006-11-18 21:50:54 +0000[diff] [blame]3232
3233<h5>Example:</h5>
3234<pre> &lt;result&gt; = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
3235 &lt;result&gt; = icmp one float 4.0, 5.0 <i>; yields: result=true</i>
3236 &lt;result&gt; = icmp olt float 4.0, 5.0 <i>; yields: result=true</i>
3237 &lt;result&gt; = icmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
3238</pre>
3239</div>
3240
Reid Spencer97c5fa42006-11-08 01:18:52 +0000[diff] [blame]3241<!-- _______________________________________________________________________ -->
3242<div class="doc_subsubsection"> <a name="i_phi">'<tt>phi</tt>'
3243Instruction</a> </div>
3244<div class="doc_text">
3245<h5>Syntax:</h5>
3246<pre> &lt;result&gt; = phi &lt;ty&gt; [ &lt;val0&gt;, &lt;label0&gt;], ...<br></pre>
3247<h5>Overview:</h5>
3248<p>The '<tt>phi</tt>' instruction is used to implement the &#966; node in
3249the SSA graph representing the function.</p>
3250<h5>Arguments:</h5>
3251<p>The type of the incoming values are specified with the first type
3252field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
3253as arguments, with one pair for each predecessor basic block of the
3254current block. Only values of <a href="#t_firstclass">first class</a>
3255type may be used as the value arguments to the PHI node. Only labels
3256may be used as the label arguments.</p>
3257<p>There must be no non-phi instructions between the start of a basic
3258block and the PHI instructions: i.e. PHI instructions must be first in
3259a basic block.</p>
3260<h5>Semantics:</h5>
3261<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the
3262value specified by the parameter, depending on which basic block we
3263came from in the last <a href="#terminators">terminator</a> instruction.</p>
3264<h5>Example:</h5>
3265<pre>Loop: ; Infinite loop that counts from 0 on up...<br> %indvar = phi uint [ 0, %LoopHeader ], [ %nextindvar, %Loop ]<br> %nextindvar = add uint %indvar, 1<br> br label %Loop<br></pre>
3266</div>
3267
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]3268<!-- _______________________________________________________________________ -->
3269<div class="doc_subsubsection">
3270 <a name="i_select">'<tt>select</tt>' Instruction</a>
3271</div>
3272
3273<div class="doc_text">
3274
3275<h5>Syntax:</h5>
3276
3277<pre>
3278 &lt;result&gt; = select bool &lt;cond&gt;, &lt;ty&gt; &lt;val1&gt;, &lt;ty&gt; &lt;val2&gt; <i>; yields ty</i>
3279</pre>
3280
3281<h5>Overview:</h5>
3282
3283<p>
3284The '<tt>select</tt>' instruction is used to choose one value based on a
3285condition, without branching.
3286</p>
3287
3288
3289<h5>Arguments:</h5>
3290
3291<p>
3292The '<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.
3293</p>
3294
3295<h5>Semantics:</h5>
3296
3297<p>
3298If the boolean condition evaluates to true, the instruction returns the first
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]3299value argument; otherwise, it returns the second value argument.
Chris Lattnerb53c28d2004-03-12 05:50:16 +0000[diff] [blame]3300</p>
3301
3302<h5>Example:</h5>
3303
3304<pre>
3305 %X = select bool true, ubyte 17, ubyte 42 <i>; yields ubyte:17</i>
3306</pre>
3307</div>
3308
Robert Bocchinof72fdfe2006-01-15 20:48:27 +0000[diff] [blame]3309
3310<!-- _______________________________________________________________________ -->
3311<div class="doc_subsubsection">
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3312 <a name="i_call">'<tt>call</tt>' Instruction</a>
3313</div>
3314
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3315<div class="doc_text">
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3316
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3317<h5>Syntax:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3318<pre>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3319 &lt;result&gt; = [tail] call [<a href="#callingconv">cconv</a>] &lt;ty&gt;* &lt;fnptrval&gt;(&lt;param list&gt;)
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3320</pre>
3321
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3322<h5>Overview:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3323
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3324<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3325
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3326<h5>Arguments:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3327
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3328<p>This instruction requires several arguments:</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3329
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]3330<ol>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3331 <li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3332 <p>The optional "tail" marker indicates whether the callee function accesses
3333 any allocas or varargs in the caller. If the "tail" marker is present, the
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3334 function call is eligible for tail call optimization. Note that calls may
3335 be marked "tail" even if they do not occur before a <a
3336 href="#i_ret"><tt>ret</tt></a> instruction.
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3337 </li>
3338 <li>
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3339 <p>The optional "cconv" marker indicates which <a href="callingconv">calling
3340 convention</a> the call should use. If none is specified, the call defaults
3341 to using C calling conventions.
3342 </li>
3343 <li>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3344 <p>'<tt>ty</tt>': shall be the signature of the pointer to function value
3345 being invoked. The argument types must match the types implied by this
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]3346 signature. This type can be omitted if the function is not varargs and
3347 if the function type does not return a pointer to a function.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3348 </li>
3349 <li>
3350 <p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
3351 be invoked. In most cases, this is a direct function invocation, but
3352 indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]3353 to function value.</p>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3354 </li>
3355 <li>
3356 <p>'<tt>function args</tt>': argument list whose types match the
Reid Spencerd845d162005-05-01 22:22:57 +0000[diff] [blame]3357 function signature argument types. All arguments must be of
3358 <a href="#t_firstclass">first class</a> type. If the function signature
3359 indicates the function accepts a variable number of arguments, the extra
3360 arguments can be specified.</p>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3361 </li>
Chris Lattnera8292f32002-05-06 22:08:29 +0000[diff] [blame]3362</ol>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3363
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3364<h5>Semantics:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3365
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3366<p>The '<tt>call</tt>' instruction is used to cause control flow to
3367transfer to a specified function, with its incoming arguments bound to
3368the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
3369instruction in the called function, control flow continues with the
3370instruction after the function call, and the return value of the
3371function is bound to the result argument. This is a simpler case of
3372the <a href="#i_invoke">invoke</a> instruction.</p>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3373
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]3374<h5>Example:</h5>
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3375
3376<pre>
3377 %retval = call int %test(int %argc)
3378 call int(sbyte*, ...) *%printf(sbyte* %msg, int 12, sbyte 42);
3379 %X = tail call int %foo()
Chris Lattner0132aff2005-05-06 22:57:40 +0000[diff] [blame]3380 %Y = tail call <a href="#callingconv">fastcc</a> int %foo()
Chris Lattnere23c1392005-05-06 05:47:36 +0000[diff] [blame]3381</pre>
3382
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3383</div>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3384
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3385<!-- _______________________________________________________________________ -->
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3386<div class="doc_subsubsection">
Chris Lattner33337472006-01-13 23:26:01 +0000[diff] [blame]3387 <a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3388</div>
3389
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3390<div class="doc_text">
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3391
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]3392<h5>Syntax:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3393
3394<pre>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3395 &lt;resultval&gt; = va_arg &lt;va_list*&gt; &lt;arglist&gt;, &lt;argty&gt;
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3396</pre>
3397
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]3398<h5>Overview:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3399
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3400<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3401the "variable argument" area of a function call. It is used to implement the
3402<tt>va_arg</tt> macro in C.</p>
3403
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]3404<h5>Arguments:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3405
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3406<p>This instruction takes a <tt>va_list*</tt> value and the type of
3407the argument. It returns a value of the specified argument type and
Jeff Cohendc6bfea2005-11-11 02:15:27 +0000[diff] [blame]3408increments the <tt>va_list</tt> to point to the next argument. Again, the
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3409actual type of <tt>va_list</tt> is target specific.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3410
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]3411<h5>Semantics:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3412
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3413<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
3414type from the specified <tt>va_list</tt> and causes the
3415<tt>va_list</tt> to point to the next argument. For more information,
3416see the variable argument handling <a href="#int_varargs">Intrinsic
3417Functions</a>.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3418
3419<p>It is legal for this instruction to be called in a function which does not
3420take a variable number of arguments, for example, the <tt>vfprintf</tt>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3421function.</p>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3422
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3423<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]3424href="#intrinsics">intrinsic function</a> because it takes a type as an
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3425argument.</p>
3426
Chris Lattner26ca62e2003-10-18 05:51:36 +0000[diff] [blame]3427<h5>Example:</h5>
Chris Lattner6a4a0492004-09-27 21:51:25 +0000[diff] [blame]3428
3429<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
3430
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3431</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]3432
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3433<!-- *********************************************************************** -->
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3434<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
3435<!-- *********************************************************************** -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]3436
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3437<div class="doc_text">
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]3438
3439<p>LLVM supports the notion of an "intrinsic function". These functions have
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]3440well known names and semantics and are required to follow certain
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]3441restrictions. Overall, these instructions represent an extension mechanism for
3442the LLVM language that does not require changing all of the transformations in
3443LLVM to add to the language (or the bytecode reader/writer, the parser,
3444etc...).</p>
3445
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]3446<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
3447prefix is reserved in LLVM for intrinsic names; thus, functions may not be named
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]3448this. Intrinsic functions must always be external functions: you cannot define
3449the body of intrinsic functions. Intrinsic functions may only be used in call
3450or invoke instructions: it is illegal to take the address of an intrinsic
3451function. Additionally, because intrinsic functions are part of the LLVM
3452language, it is required that they all be documented here if any are added.</p>
3453
3454
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]3455<p>To learn how to add an intrinsic function, please see the <a
Chris Lattner90391c12005-05-11 03:35:57 +0000[diff] [blame]3456href="ExtendingLLVM.html">Extending LLVM Guide</a>.
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]3457</p>
3458
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3459</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]3460
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3461<!-- ======================================================================= -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]3462<div class="doc_subsection">
3463 <a name="int_varargs">Variable Argument Handling Intrinsics</a>
3464</div>
3465
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3466<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3467
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3468<p>Variable argument support is defined in LLVM with the <a
Chris Lattner33337472006-01-13 23:26:01 +0000[diff] [blame]3469 href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3470intrinsic functions. These functions are related to the similarly
3471named macros defined in the <tt>&lt;stdarg.h&gt;</tt> header file.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3472
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3473<p>All of these functions operate on arguments that use a
3474target-specific value type "<tt>va_list</tt>". The LLVM assembly
3475language reference manual does not define what this type is, so all
3476transformations should be prepared to handle intrinsics with any type
3477used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3478
Chris Lattner30b868d2006-05-15 17:26:46 +0000[diff] [blame]3479<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3480instruction and the variable argument handling intrinsic functions are
3481used.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3482
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]3483<pre>
3484int %test(int %X, ...) {
3485 ; Initialize variable argument processing
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3486 %ap = alloca sbyte*
3487 call void %<a href="#i_va_start">llvm.va_start</a>(sbyte** %ap)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]3488
3489 ; Read a single integer argument
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3490 %tmp = va_arg sbyte** %ap, int
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]3491
3492 ; Demonstrate usage of llvm.va_copy and llvm.va_end
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3493 %aq = alloca sbyte*
Andrew Lenharth5305ea52005-06-22 20:38:11 +0000[diff] [blame]3494 call void %<a href="#i_va_copy">llvm.va_copy</a>(sbyte** %aq, sbyte** %ap)
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3495 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %aq)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]3496
3497 ; Stop processing of arguments.
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3498 call void %<a href="#i_va_end">llvm.va_end</a>(sbyte** %ap)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]3499 ret int %tmp
3500}
3501</pre>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3502</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]3503
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3504<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]3505<div class="doc_subsubsection">
3506 <a name="i_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
3507</div>
3508
3509
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3510<div class="doc_text">
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3511<h5>Syntax:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3512<pre> declare void %llvm.va_start(&lt;va_list&gt;* &lt;arglist&gt;)<br></pre>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3513<h5>Overview:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3514<P>The '<tt>llvm.va_start</tt>' intrinsic initializes
3515<tt>*&lt;arglist&gt;</tt> for subsequent use by <tt><a
3516href="#i_va_arg">va_arg</a></tt>.</p>
3517
3518<h5>Arguments:</h5>
3519
3520<P>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
3521
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3522<h5>Semantics:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3523
3524<P>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
3525macro available in C. In a target-dependent way, it initializes the
3526<tt>va_list</tt> element the argument points to, so that the next call to
3527<tt>va_arg</tt> will produce the first variable argument passed to the function.
3528Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
3529last argument of the function, the compiler can figure that out.</p>
3530
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3531</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]3532
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3533<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]3534<div class="doc_subsubsection">
3535 <a name="i_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
3536</div>
3537
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3538<div class="doc_text">
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3539<h5>Syntax:</h5>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3540<pre> declare void %llvm.va_end(&lt;va_list*&gt; &lt;arglist&gt;)<br></pre>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3541<h5>Overview:</h5>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3542<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>&lt;arglist&gt;</tt>
3543which has been initialized previously with <tt><a href="#i_va_start">llvm.va_start</a></tt>
3544or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3545<h5>Arguments:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3546<p>The argument is a <tt>va_list</tt> to destroy.</p>
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3547<h5>Semantics:</h5>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3548<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Chris Lattner48b383b02003-11-25 01:02:51 +0000[diff] [blame]3549macro available in C. In a target-dependent way, it destroys the <tt>va_list</tt>.
3550Calls to <a href="#i_va_start"><tt>llvm.va_start</tt></a> and <a
3551 href="#i_va_copy"><tt>llvm.va_copy</tt></a> must be matched exactly
3552with calls to <tt>llvm.va_end</tt>.</p>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3553</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]3554
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3555<!-- _______________________________________________________________________ -->
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]3556<div class="doc_subsubsection">
3557 <a name="i_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
3558</div>
3559
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3560<div class="doc_text">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3561
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3562<h5>Syntax:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3563
3564<pre>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3565 declare void %llvm.va_copy(&lt;va_list&gt;* &lt;destarglist&gt;,
Andrew Lenharth5305ea52005-06-22 20:38:11 +0000[diff] [blame]3566 &lt;va_list&gt;* &lt;srcarglist&gt;)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3567</pre>
3568
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3569<h5>Overview:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3570
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3571<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position from
3572the source argument list to the destination argument list.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3573
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3574<h5>Arguments:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3575
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3576<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Andrew Lenharth5305ea52005-06-22 20:38:11 +0000[diff] [blame]3577The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3578
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3579
Chris Lattnerbd64b4e2003-05-08 04:57:36 +0000[diff] [blame]3580<h5>Semantics:</h5>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3581
Andrew Lenharth5fb787c2005-06-18 18:28:17 +0000[diff] [blame]3582<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt> macro
3583available in C. In a target-dependent way, it copies the source
3584<tt>va_list</tt> element into the destination list. This intrinsic is necessary
3585because the <tt><a href="i_va_begin">llvm.va_begin</a></tt> intrinsic may be
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3586arbitrarily complex and require memory allocation, for example.</p>
3587
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]3588</div>
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]3589
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]3590<!-- ======================================================================= -->
3591<div class="doc_subsection">
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3592 <a name="int_gc">Accurate Garbage Collection Intrinsics</a>
3593</div>
3594
3595<div class="doc_text">
3596
3597<p>
3598LLVM support for <a href="GarbageCollection.html">Accurate Garbage
3599Collection</a> requires the implementation and generation of these intrinsics.
3600These intrinsics allow identification of <a href="#i_gcroot">GC roots on the
3601stack</a>, as well as garbage collector implementations that require <a
3602href="#i_gcread">read</a> and <a href="#i_gcwrite">write</a> barriers.
3603Front-ends for type-safe garbage collected languages should generate these
3604intrinsics to make use of the LLVM garbage collectors. For more details, see <a
3605href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
3606</p>
3607</div>
3608
3609<!-- _______________________________________________________________________ -->
3610<div class="doc_subsubsection">
3611 <a name="i_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
3612</div>
3613
3614<div class="doc_text">
3615
3616<h5>Syntax:</h5>
3617
3618<pre>
Reid Spencer7821d062005-04-26 20:50:44 +0000[diff] [blame]3619 declare void %llvm.gcroot(&lt;ty&gt;** %ptrloc, &lt;ty2&gt;* %metadata)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3620</pre>
3621
3622<h5>Overview:</h5>
3623
John Criswelldfe6a862004-12-10 15:51:16 +0000[diff] [blame]3624<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3625the code generator, and allows some metadata to be associated with it.</p>
3626
3627<h5>Arguments:</h5>
3628
3629<p>The first argument specifies the address of a stack object that contains the
3630root pointer. The second pointer (which must be either a constant or a global
3631value address) contains the meta-data to be associated with the root.</p>
3632
3633<h5>Semantics:</h5>
3634
3635<p>At runtime, a call to this intrinsics stores a null pointer into the "ptrloc"
3636location. At compile-time, the code generator generates information to allow
3637the runtime to find the pointer at GC safe points.
3638</p>
3639
3640</div>
3641
3642
3643<!-- _______________________________________________________________________ -->
3644<div class="doc_subsubsection">
3645 <a name="i_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
3646</div>
3647
3648<div class="doc_text">
3649
3650<h5>Syntax:</h5>
3651
3652<pre>
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]3653 declare sbyte* %llvm.gcread(sbyte* %ObjPtr, sbyte** %Ptr)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3654</pre>
3655
3656<h5>Overview:</h5>
3657
3658<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
3659locations, allowing garbage collector implementations that require read
3660barriers.</p>
3661
3662<h5>Arguments:</h5>
3663
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]3664<p>The second argument is the address to read from, which should be an address
3665allocated from the garbage collector. The first object is a pointer to the
3666start of the referenced object, if needed by the language runtime (otherwise
3667null).</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3668
3669<h5>Semantics:</h5>
3670
3671<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
3672instruction, but may be replaced with substantially more complex code by the
3673garbage collector runtime, as needed.</p>
3674
3675</div>
3676
3677
3678<!-- _______________________________________________________________________ -->
3679<div class="doc_subsubsection">
3680 <a name="i_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
3681</div>
3682
3683<div class="doc_text">
3684
3685<h5>Syntax:</h5>
3686
3687<pre>
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]3688 declare void %llvm.gcwrite(sbyte* %P1, sbyte* %Obj, sbyte** %P2)
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3689</pre>
3690
3691<h5>Overview:</h5>
3692
3693<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
3694locations, allowing garbage collector implementations that require write
3695barriers (such as generational or reference counting collectors).</p>
3696
3697<h5>Arguments:</h5>
3698
Chris Lattnerf9228072006-03-14 20:02:51 +0000[diff] [blame]3699<p>The first argument is the reference to store, the second is the start of the
3700object to store it to, and the third is the address of the field of Obj to
3701store to. If the runtime does not require a pointer to the object, Obj may be
3702null.</p>
Chris Lattner757528b0b2004-05-23 21:06:01 +0000[diff] [blame]3703
3704<h5>Semantics:</h5>
3705
3706<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
3707instruction, but may be replaced with substantially more complex code by the
3708garbage collector runtime, as needed.</p>
3709
3710</div>
3711
3712
3713
3714<!-- ======================================================================= -->
3715<div class="doc_subsection">
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]3716 <a name="int_codegen">Code Generator Intrinsics</a>
3717</div>
3718
3719<div class="doc_text">
3720<p>
3721These intrinsics are provided by LLVM to expose special features that may only
3722be implemented with code generator support.
3723</p>
3724
3725</div>
3726
3727<!-- _______________________________________________________________________ -->
3728<div class="doc_subsubsection">
3729 <a name="i_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
3730</div>
3731
3732<div class="doc_text">
3733
3734<h5>Syntax:</h5>
3735<pre>
Chris Lattnerb3d430e2006-01-13 01:20:27 +0000[diff] [blame]3736 declare sbyte *%llvm.returnaddress(uint &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]3737</pre>
3738
3739<h5>Overview:</h5>
3740
3741<p>
Chris Lattnerc1fb4262006-10-15 20:05:59 +0000[diff] [blame]3742The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
3743target-specific value indicating the return address of the current function
3744or one of its callers.
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]3745</p>
3746
3747<h5>Arguments:</h5>
3748
3749<p>
3750The argument to this intrinsic indicates which function to return the address
3751for. Zero indicates the calling function, one indicates its caller, etc. The
3752argument is <b>required</b> to be a constant integer value.
3753</p>
3754
3755<h5>Semantics:</h5>
3756
3757<p>
3758The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
3759the return address of the specified call frame, or zero if it cannot be
3760identified. The value returned by this intrinsic is likely to be incorrect or 0
3761for arguments other than zero, so it should only be used for debugging purposes.
3762</p>
3763
3764<p>
3765Note that calling this intrinsic does not prevent function inlining or other
Chris Lattner2e6eb5f2005-03-07 20:30:51 +0000[diff] [blame]3766aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]3767source-language caller.
3768</p>
3769</div>
3770
3771
3772<!-- _______________________________________________________________________ -->
3773<div class="doc_subsubsection">
3774 <a name="i_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
3775</div>
3776
3777<div class="doc_text">
3778
3779<h5>Syntax:</h5>
3780<pre>
Chris Lattnerb3d430e2006-01-13 01:20:27 +0000[diff] [blame]3781 declare sbyte *%llvm.frameaddress(uint &lt;level&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]3782</pre>
3783
3784<h5>Overview:</h5>
3785
3786<p>
Chris Lattnerc1fb4262006-10-15 20:05:59 +0000[diff] [blame]3787The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
3788target-specific frame pointer value for the specified stack frame.
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]3789</p>
3790
3791<h5>Arguments:</h5>
3792
3793<p>
3794The argument to this intrinsic indicates which function to return the frame
3795pointer for. Zero indicates the calling function, one indicates its caller,
3796etc. The argument is <b>required</b> to be a constant integer value.
3797</p>
3798
3799<h5>Semantics:</h5>
3800
3801<p>
3802The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
3803the frame address of the specified call frame, or zero if it cannot be
3804identified. The value returned by this intrinsic is likely to be incorrect or 0
3805for arguments other than zero, so it should only be used for debugging purposes.
3806</p>
3807
3808<p>
3809Note that calling this intrinsic does not prevent function inlining or other
Chris Lattner2e6eb5f2005-03-07 20:30:51 +0000[diff] [blame]3810aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]3811source-language caller.
3812</p>
3813</div>
3814
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]3815<!-- _______________________________________________________________________ -->
3816<div class="doc_subsubsection">
Chris Lattner2f0f0012006-01-13 02:03:13 +0000[diff] [blame]3817 <a name="i_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
3818</div>
3819
3820<div class="doc_text">
3821
3822<h5>Syntax:</h5>
3823<pre>
3824 declare sbyte *%llvm.stacksave()
3825</pre>
3826
3827<h5>Overview:</h5>
3828
3829<p>
3830The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
3831the function stack, for use with <a href="#i_stackrestore">
3832<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
3833features like scoped automatic variable sized arrays in C99.
3834</p>
3835
3836<h5>Semantics:</h5>
3837
3838<p>
3839This intrinsic returns a opaque pointer value that can be passed to <a
3840href="#i_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
3841<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
3842<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
3843state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
3844practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
3845that were allocated after the <tt>llvm.stacksave</tt> was executed.
3846</p>
3847
3848</div>
3849
3850<!-- _______________________________________________________________________ -->
3851<div class="doc_subsubsection">
3852 <a name="i_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
3853</div>
3854
3855<div class="doc_text">
3856
3857<h5>Syntax:</h5>
3858<pre>
3859 declare void %llvm.stackrestore(sbyte* %ptr)
3860</pre>
3861
3862<h5>Overview:</h5>
3863
3864<p>
3865The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
3866the function stack to the state it was in when the corresponding <a
3867href="#llvm.stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
3868useful for implementing language features like scoped automatic variable sized
3869arrays in C99.
3870</p>
3871
3872<h5>Semantics:</h5>
3873
3874<p>
3875See the description for <a href="#i_stacksave"><tt>llvm.stacksave</tt></a>.
3876</p>
3877
3878</div>
3879
3880
3881<!-- _______________________________________________________________________ -->
3882<div class="doc_subsubsection">
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]3883 <a name="i_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
3884</div>
3885
3886<div class="doc_text">
3887
3888<h5>Syntax:</h5>
3889<pre>
Reid Spencer7821d062005-04-26 20:50:44 +0000[diff] [blame]3890 declare void %llvm.prefetch(sbyte * &lt;address&gt;,
3891 uint &lt;rw&gt;, uint &lt;locality&gt;)
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]3892</pre>
3893
3894<h5>Overview:</h5>
3895
3896
3897<p>
3898The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]3899a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
3900no
3901effect on the behavior of the program but can change its performance
Chris Lattnerff851072005-02-28 19:47:14 +0000[diff] [blame]3902characteristics.
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]3903</p>
3904
3905<h5>Arguments:</h5>
3906
3907<p>
3908<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
3909determining if the fetch should be for a read (0) or write (1), and
3910<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Chris Lattnerd3e641c2005-03-07 20:31:38 +0000[diff] [blame]3911locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
Chris Lattnerc8a2c222005-02-28 19:24:19 +0000[diff] [blame]3912<tt>locality</tt> arguments must be constant integers.
3913</p>
3914
3915<h5>Semantics:</h5>
3916
3917<p>
3918This intrinsic does not modify the behavior of the program. In particular,
3919prefetches cannot trap and do not produce a value. On targets that support this
3920intrinsic, the prefetch can provide hints to the processor cache for better
3921performance.
3922</p>
3923
3924</div>
3925
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]3926<!-- _______________________________________________________________________ -->
3927<div class="doc_subsubsection">
3928 <a name="i_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
3929</div>
3930
3931<div class="doc_text">
3932
3933<h5>Syntax:</h5>
3934<pre>
Reid Spencer7821d062005-04-26 20:50:44 +0000[diff] [blame]3935 declare void %llvm.pcmarker( uint &lt;id&gt; )
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]3936</pre>
3937
3938<h5>Overview:</h5>
3939
3940
3941<p>
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]3942The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
3943(PC) in a region of
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]3944code to simulators and other tools. The method is target specific, but it is
3945expected that the marker will use exported symbols to transmit the PC of the marker.
Jeff Cohendc6bfea2005-11-11 02:15:27 +0000[diff] [blame]3946The marker makes no guarantees that it will remain with any specific instruction
Chris Lattnere64d41d2005-11-15 06:07:55 +0000[diff] [blame]3947after optimizations. It is possible that the presence of a marker will inhibit
Chris Lattnerb40261e2006-03-24 07:16:10 +0000[diff] [blame]3948optimizations. The intended use is to be inserted after optimizations to allow
John Criswell88190562005-05-16 16:17:45 +0000[diff] [blame]3949correlations of simulation runs.
Andrew Lenharthb4427912005-03-28 20:05:49 +0000[diff] [blame]3950</p>
3951
3952<h5>Arguments:</h5>
3953
3954<p>
3955<tt>id</tt> is a numerical id identifying the marker.
3956</p>
3957
3958<h5>Semantics:</h5>
3959
3960<p>
3961This intrinsic does not modify the behavior of the program. Backends that do not
3962support this intrinisic may ignore it.
3963</p>
3964
3965</div>
3966
Andrew Lenharth01aa5632005-11-11 16:47:30 +0000[diff] [blame]3967<!-- _______________________________________________________________________ -->
3968<div class="doc_subsubsection">
3969 <a name="i_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
3970</div>
3971
3972<div class="doc_text">
3973
3974<h5>Syntax:</h5>
3975<pre>
3976 declare ulong %llvm.readcyclecounter( )
3977</pre>
3978
3979<h5>Overview:</h5>
3980
3981
3982<p>
3983The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
3984counter register (or similar low latency, high accuracy clocks) on those targets
3985that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
3986As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
3987should only be used for small timings.
3988</p>
3989
3990<h5>Semantics:</h5>
3991
3992<p>
3993When directly supported, reading the cycle counter should not modify any memory.
3994Implementations are allowed to either return a application specific value or a
3995system wide value. On backends without support, this is lowered to a constant 0.
3996</p>
3997
3998</div>
3999
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]4000<!-- ======================================================================= -->
4001<div class="doc_subsection">
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]4002 <a name="int_libc">Standard C Library Intrinsics</a>
4003</div>
4004
4005<div class="doc_text">
4006<p>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]4007LLVM provides intrinsics for a few important standard C library functions.
4008These intrinsics allow source-language front-ends to pass information about the
4009alignment of the pointer arguments to the code generator, providing opportunity
4010for more efficient code generation.
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]4011</p>
4012
4013</div>
4014
4015<!-- _______________________________________________________________________ -->
4016<div class="doc_subsubsection">
4017 <a name="i_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
4018</div>
4019
4020<div class="doc_text">
4021
4022<h5>Syntax:</h5>
4023<pre>
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4024 declare void %llvm.memcpy.i32(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
4025 uint &lt;len&gt;, uint &lt;align&gt;)
4026 declare void %llvm.memcpy.i64(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
4027 ulong &lt;len&gt;, uint &lt;align&gt;)
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]4028</pre>
4029
4030<h5>Overview:</h5>
4031
4032<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4033The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]4034location to the destination location.
4035</p>
4036
4037<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4038Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
4039intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]4040</p>
4041
4042<h5>Arguments:</h5>
4043
4044<p>
4045The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4046the source. The third argument is an integer argument
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]4047specifying the number of bytes to copy, and the fourth argument is the alignment
4048of the source and destination locations.
4049</p>
4050
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]4051<p>
4052If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattner5316e5d2006-03-04 00:02:10 +0000[diff] [blame]4053the caller guarantees that both the source and destination pointers are aligned
4054to that boundary.
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]4055</p>
4056
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]4057<h5>Semantics:</h5>
4058
4059<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4060The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattnerfee11462004-02-12 17:01:32 +0000[diff] [blame]4061location to the destination location, which are not allowed to overlap. It
4062copies "len" bytes of memory over. If the argument is known to be aligned to
4063some boundary, this can be specified as the fourth argument, otherwise it should
4064be set to 0 or 1.
4065</p>
4066</div>
4067
4068
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]4069<!-- _______________________________________________________________________ -->
4070<div class="doc_subsubsection">
4071 <a name="i_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
4072</div>
4073
4074<div class="doc_text">
4075
4076<h5>Syntax:</h5>
4077<pre>
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4078 declare void %llvm.memmove.i32(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
4079 uint &lt;len&gt;, uint &lt;align&gt;)
4080 declare void %llvm.memmove.i64(sbyte* &lt;dest&gt;, sbyte* &lt;src&gt;,
4081 ulong &lt;len&gt;, uint &lt;align&gt;)
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]4082</pre>
4083
4084<h5>Overview:</h5>
4085
4086<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4087The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
4088location to the destination location. It is similar to the
4089'<tt>llvm.memcmp</tt>' intrinsic but allows the two memory locations to overlap.
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]4090</p>
4091
4092<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4093Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
4094intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]4095</p>
4096
4097<h5>Arguments:</h5>
4098
4099<p>
4100The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4101the source. The third argument is an integer argument
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]4102specifying the number of bytes to copy, and the fourth argument is the alignment
4103of the source and destination locations.
4104</p>
4105
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]4106<p>
4107If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattner5316e5d2006-03-04 00:02:10 +0000[diff] [blame]4108the caller guarantees that the source and destination pointers are aligned to
4109that boundary.
Chris Lattner4c67c482004-02-12 21:18:15 +0000[diff] [blame]4110</p>
4111
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]4112<h5>Semantics:</h5>
4113
4114<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4115The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
Chris Lattnerf30152e2004-02-12 18:10:10 +0000[diff] [blame]4116location to the destination location, which may overlap. It
4117copies "len" bytes of memory over. If the argument is known to be aligned to
4118some boundary, this can be specified as the fourth argument, otherwise it should
4119be set to 0 or 1.
4120</p>
4121</div>
4122
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]4123
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]4124<!-- _______________________________________________________________________ -->
4125<div class="doc_subsubsection">
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4126 <a name="i_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]4127</div>
4128
4129<div class="doc_text">
4130
4131<h5>Syntax:</h5>
4132<pre>
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4133 declare void %llvm.memset.i32(sbyte* &lt;dest&gt;, ubyte &lt;val&gt;,
4134 uint &lt;len&gt;, uint &lt;align&gt;)
4135 declare void %llvm.memset.i64(sbyte* &lt;dest&gt;, ubyte &lt;val&gt;,
4136 ulong &lt;len&gt;, uint &lt;align&gt;)
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]4137</pre>
4138
4139<h5>Overview:</h5>
4140
4141<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4142The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]4143byte value.
4144</p>
4145
4146<p>
4147Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
4148does not return a value, and takes an extra alignment argument.
4149</p>
4150
4151<h5>Arguments:</h5>
4152
4153<p>
4154The first argument is a pointer to the destination to fill, the second is the
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4155byte value to fill it with, the third argument is an integer
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]4156argument specifying the number of bytes to fill, and the fourth argument is the
4157known alignment of destination location.
4158</p>
4159
4160<p>
4161If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattner5316e5d2006-03-04 00:02:10 +0000[diff] [blame]4162the caller guarantees that the destination pointer is aligned to that boundary.
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]4163</p>
4164
4165<h5>Semantics:</h5>
4166
4167<p>
Chris Lattner0c8b2592006-03-03 00:07:20 +0000[diff] [blame]4168The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
4169the
Chris Lattner3649c3a2004-02-14 04:08:35 +0000[diff] [blame]4170destination location. If the argument is known to be aligned to some boundary,
4171this can be specified as the fourth argument, otherwise it should be set to 0 or
41721.
4173</p>
4174</div>
4175
4176
Chris Lattner3b4f4372004-06-11 02:28:03 +0000[diff] [blame]4177<!-- _______________________________________________________________________ -->
4178<div class="doc_subsubsection">
Reid Spencerb4f9a6f2006-01-16 21:12:35 +0000[diff] [blame]4179 <a name="i_isunordered">'<tt>llvm.isunordered.*</tt>' Intrinsic</a>
Alkis Evlogimenos0fa39232004-06-13 01:16:15 +0000[diff] [blame]4180</div>
4181
4182<div class="doc_text">
4183
4184<h5>Syntax:</h5>
4185<pre>
Reid Spencerb4f9a6f2006-01-16 21:12:35 +0000[diff] [blame]4186 declare bool %llvm.isunordered.f32(float Val1, float Val2)
4187 declare bool %llvm.isunordered.f64(double Val1, double Val2)
Alkis Evlogimenos0fa39232004-06-13 01:16:15 +0000[diff] [blame]4188</pre>
4189
4190<h5>Overview:</h5>
4191
4192<p>
Reid Spencerb4f9a6f2006-01-16 21:12:35 +0000[diff] [blame]4193The '<tt>llvm.isunordered</tt>' intrinsics return true if either or both of the
Alkis Evlogimenos0fa39232004-06-13 01:16:15 +0000[diff] [blame]4194specified floating point values is a NAN.
4195</p>
4196
4197<h5>Arguments:</h5>
4198
4199<p>
4200The arguments are floating point numbers of the same type.
4201</p>
4202
4203<h5>Semantics:</h5>
4204
4205<p>
4206If either or both of the arguments is a SNAN or QNAN, it returns true, otherwise
4207false.
4208</p>
4209</div>
4210
4211
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]4212<!-- _______________________________________________________________________ -->
4213<div class="doc_subsubsection">
Chris Lattner069b5bd2006-01-16 22:38:59 +0000[diff] [blame]4214 <a name="i_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]4215</div>
4216
4217<div class="doc_text">
4218
4219<h5>Syntax:</h5>
4220<pre>
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]4221 declare float %llvm.sqrt.f32(float %Val)
4222 declare double %llvm.sqrt.f64(double %Val)
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]4223</pre>
4224
4225<h5>Overview:</h5>
4226
4227<p>
Reid Spencerb4f9a6f2006-01-16 21:12:35 +0000[diff] [blame]4228The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Chris Lattner8a8f2e52005-07-21 01:29:16 +0000[diff] [blame]4229returning the same value as the libm '<tt>sqrt</tt>' function would. Unlike
4230<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
4231negative numbers (which allows for better optimization).
4232</p>
4233
4234<h5>Arguments:</h5>
4235
4236<p>
4237The argument and return value are floating point numbers of the same type.
4238</p>
4239
4240<h5>Semantics:</h5>
4241
4242<p>
4243This function returns the sqrt of the specified operand if it is a positive
4244floating point number.
4245</p>
4246</div>
4247
Chris Lattner33b73f92006-09-08 06:34:02 +0000[diff] [blame]4248<!-- _______________________________________________________________________ -->
4249<div class="doc_subsubsection">
4250 <a name="i_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
4251</div>
4252
4253<div class="doc_text">
4254
4255<h5>Syntax:</h5>
4256<pre>
4257 declare float %llvm.powi.f32(float %Val, int %power)
4258 declare double %llvm.powi.f64(double %Val, int %power)
4259</pre>
4260
4261<h5>Overview:</h5>
4262
4263<p>
4264The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
4265specified (positive or negative) power. The order of evaluation of
4266multiplications is not defined.
4267</p>
4268
4269<h5>Arguments:</h5>
4270
4271<p>
4272The second argument is an integer power, and the first is a value to raise to
4273that power.
4274</p>
4275
4276<h5>Semantics:</h5>
4277
4278<p>
4279This function returns the first value raised to the second power with an
4280unspecified sequence of rounding operations.</p>
4281</div>
4282
4283
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]4284<!-- ======================================================================= -->
4285<div class="doc_subsection">
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]4286 <a name="int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]4287</div>
4288
4289<div class="doc_text">
4290<p>
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]4291LLVM provides intrinsics for a few important bit manipulation operations.
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]4292These allow efficient code generation for some algorithms.
4293</p>
4294
4295</div>
4296
4297<!-- _______________________________________________________________________ -->
4298<div class="doc_subsubsection">
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]4299 <a name="i_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
4300</div>
4301
4302<div class="doc_text">
4303
4304<h5>Syntax:</h5>
4305<pre>
Chris Lattner069b5bd2006-01-16 22:38:59 +0000[diff] [blame]4306 declare ushort %llvm.bswap.i16(ushort &lt;id&gt;)
4307 declare uint %llvm.bswap.i32(uint &lt;id&gt;)
4308 declare ulong %llvm.bswap.i64(ulong &lt;id&gt;)
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]4309</pre>
4310
4311<h5>Overview:</h5>
4312
4313<p>
4314The '<tt>llvm.bwsap</tt>' family of intrinsics is used to byteswap a 16, 32 or
431564 bit quantity. These are useful for performing operations on data that is not
4316in the target's native byte order.
4317</p>
4318
4319<h5>Semantics:</h5>
4320
4321<p>
Chris Lattner069b5bd2006-01-16 22:38:59 +0000[diff] [blame]4322The <tt>llvm.bswap.16</tt> intrinsic returns a ushort value that has the high and low
4323byte of the input ushort swapped. Similarly, the <tt>llvm.bswap.i32</tt> intrinsic
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]4324returns a uint value that has the four bytes of the input uint swapped, so that
4325if the input bytes are numbered 0, 1, 2, 3 then the returned uint will have its
Chris Lattner069b5bd2006-01-16 22:38:59 +0000[diff] [blame]4326bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i64</tt> intrinsic extends this concept
Nate Begeman0f223bb2006-01-13 23:26:38 +0000[diff] [blame]4327to 64 bits.
4328</p>
4329
4330</div>
4331
4332<!-- _______________________________________________________________________ -->
4333<div class="doc_subsubsection">
Reid Spencerb4f9a6f2006-01-16 21:12:35 +0000[diff] [blame]4334 <a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]4335</div>
4336
4337<div class="doc_text">
4338
4339<h5>Syntax:</h5>
4340<pre>
Chris Lattner069b5bd2006-01-16 22:38:59 +0000[diff] [blame]4341 declare ubyte %llvm.ctpop.i8 (ubyte &lt;src&gt;)
4342 declare ushort %llvm.ctpop.i16(ushort &lt;src&gt;)
4343 declare uint %llvm.ctpop.i32(uint &lt;src&gt;)
4344 declare ulong %llvm.ctpop.i64(ulong &lt;src&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]4345</pre>
4346
4347<h5>Overview:</h5>
4348
4349<p>
Chris Lattner069b5bd2006-01-16 22:38:59 +0000[diff] [blame]4350The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
4351value.
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]4352</p>
4353
4354<h5>Arguments:</h5>
4355
4356<p>
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]4357The only argument is the value to be counted. The argument may be of any
Chris Lattner069b5bd2006-01-16 22:38:59 +0000[diff] [blame]4358unsigned integer type. The return type must match the argument type.
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]4359</p>
4360
4361<h5>Semantics:</h5>
4362
4363<p>
4364The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
4365</p>
4366</div>
4367
4368<!-- _______________________________________________________________________ -->
4369<div class="doc_subsubsection">
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]4370 <a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]4371</div>
4372
4373<div class="doc_text">
4374
4375<h5>Syntax:</h5>
4376<pre>
Chris Lattner069b5bd2006-01-16 22:38:59 +0000[diff] [blame]4377 declare ubyte %llvm.ctlz.i8 (ubyte &lt;src&gt;)
4378 declare ushort %llvm.ctlz.i16(ushort &lt;src&gt;)
4379 declare uint %llvm.ctlz.i32(uint &lt;src&gt;)
4380 declare ulong %llvm.ctlz.i64(ulong &lt;src&gt;)
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]4381</pre>
4382
4383<h5>Overview:</h5>
4384
4385<p>
Reid Spencerb4f9a6f2006-01-16 21:12:35 +0000[diff] [blame]4386The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
4387leading zeros in a variable.
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]4388</p>
4389
4390<h5>Arguments:</h5>
4391
4392<p>
Chris Lattner573f64e2005-05-07 01:46:40 +0000[diff] [blame]4393The only argument is the value to be counted. The argument may be of any
Chris Lattner069b5bd2006-01-16 22:38:59 +0000[diff] [blame]4394unsigned integer type. The return type must match the argument type.
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]4395</p>
4396
4397<h5>Semantics:</h5>
4398
4399<p>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]4400The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
4401in a variable. If the src == 0 then the result is the size in bits of the type
Chris Lattner905bd172006-04-21 21:37:40 +0000[diff] [blame]4402of src. For example, <tt>llvm.ctlz(int 2) = 30</tt>.
Andrew Lenharth1d463522005-05-03 18:01:48 +0000[diff] [blame]4403</p>
4404</div>
Chris Lattner3b4f4372004-06-11 02:28:03 +0000[diff] [blame]4405
4406
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]4407
4408<!-- _______________________________________________________________________ -->
4409<div class="doc_subsubsection">
Chris Lattnerb748c672006-01-16 22:34:14 +0000[diff] [blame]4410 <a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]4411</div>
4412
4413<div class="doc_text">
4414
4415<h5>Syntax:</h5>
4416<pre>
Chris Lattner069b5bd2006-01-16 22:38:59 +0000[diff] [blame]4417 declare ubyte %llvm.cttz.i8 (ubyte &lt;src&gt;)
4418 declare ushort %llvm.cttz.i16(ushort &lt;src&gt;)
4419 declare uint %llvm.cttz.i32(uint &lt;src&gt;)
4420 declare ulong %llvm.cttz.i64(ulong &lt;src&gt;)
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]4421</pre>
4422
4423<h5>Overview:</h5>
4424
4425<p>
Reid Spencerb4f9a6f2006-01-16 21:12:35 +0000[diff] [blame]4426The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
4427trailing zeros.
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]4428</p>
4429
4430<h5>Arguments:</h5>
4431
4432<p>
4433The only argument is the value to be counted. The argument may be of any
Chris Lattner069b5bd2006-01-16 22:38:59 +0000[diff] [blame]4434unsigned integer type. The return type must match the argument type.
Chris Lattnerefa20fa2005-05-15 19:39:26 +0000[diff] [blame]4435</p>
4436
4437<h5>Semantics:</h5>
4438
4439<p>
4440The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
4441in a variable. If the src == 0 then the result is the size in bits of the type
4442of src. For example, <tt>llvm.cttz(2) = 1</tt>.
4443</p>
4444</div>
4445
Chris Lattner941515c2004-01-06 05:31:32 +0000[diff] [blame]4446<!-- ======================================================================= -->
4447<div class="doc_subsection">
4448 <a name="int_debugger">Debugger Intrinsics</a>
4449</div>
4450
4451<div class="doc_text">
4452<p>
4453The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
4454are described in the <a
4455href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
4456Debugging</a> document.
4457</p>
4458</div>
4459
4460
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4461<!-- *********************************************************************** -->
Chris Lattner2f7c9632001-06-06 20:29:01 +0000[diff] [blame]4462<hr>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]4463<address>
4464 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
4465 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
4466 <a href="http://validator.w3.org/check/referer"><img
4467 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
4468
4469 <a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencerca058542006-03-14 05:39:39 +0000[diff] [blame]4470 <a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukmanc501f552004-03-01 17:47:27 +0000[diff] [blame]4471 Last modified: $Date$
4472</address>
Misha Brukman76307852003-11-08 01:05:38 +0000[diff] [blame]4473</body>
4474</html>