Blame - docs/LangRef.html - fp2-dev/platform/external/llvm

blob: c698e8a90a9d07b9193ae55ea83eff1aea434e5b [file] [log] [blame]

Misha Brukman	daa4cb0	2004-03-01 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 Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3	<html>
				4	<head>
				5	<title>LLVM Assembly Language Reference Manual</title>
Reid Spencer	3921c74	2004-08-26 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 Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	10	<link rel="stylesheet" href="llvm.css" type="text/css">
				11	</head>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	12
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	13	<body>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	14
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	15	<div class="doc_title"> LLVM Language Reference Manual </div>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	16	<ol>
Misha Brukman	9d0919f	2003-11-08 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 Lattner	fa73021	2004-12-09 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 Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	23	<li><a href="#linkage">Linkage Types</a></li>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	24	<li><a href="#callingconv">Calling Conventions</a></li>
Chris Lattner	e7886e4	2009-01-11 20:53:49 +0000	[diff] [blame]	25	<li><a href="#namedtypes">Named Types</a></li>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	26	<li><a href="#globalvars">Global Variables</a></li>
Chris Lattner	4e9aba7	2006-01-23 23:23:47 +0000	[diff] [blame]	27	<li><a href="#functionstructure">Functions</a></li>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	28	<li><a href="#aliasstructure">Aliases</a></li>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	29	<li><a href="#paramattrs">Parameter Attributes</a></li>
Devang Patel	2c9c3e7	2008-09-26 23:51:19 +0000	[diff] [blame]	30	<li><a href="#fnattrs">Function Attributes</a></li>
Gordon Henriksen	80a75bf	2007-12-10 03:18:06 +0000	[diff] [blame]	31	<li><a href="#gc">Garbage Collector Names</a></li>
Chris Lattner	4e9aba7	2006-01-23 23:23:47 +0000	[diff] [blame]	32	<li><a href="#moduleasm">Module-Level Inline Assembly</a></li>
Reid Spencer	de15194	2007-02-19 23:54:10 +0000	[diff] [blame]	33	<li><a href="#datalayout">Data Layout</a></li>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	34	</ol>
				35	</li>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	36	<li><a href="#typesystem">Type System</a>
				37	<ol>
Chris Lattner	4f69f46	2008-01-04 04:32:38 +0000	[diff] [blame]	38	<li><a href="#t_classifications">Type Classifications</a></li>
Robert Bocchino	7b81c75	2006-02-17 21:18:08 +0000	[diff] [blame]	39	<li><a href="#t_primitive">Primitive Types</a>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	40	<ol>
Chris Lattner	4f69f46	2008-01-04 04:32:38 +0000	[diff] [blame]	41	<li><a href="#t_floating">Floating Point Types</a></li>
				42	<li><a href="#t_void">Void Type</a></li>
				43	<li><a href="#t_label">Label Type</a></li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	44	</ol>
				45	</li>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	46	<li><a href="#t_derived">Derived Types</a>
				47	<ol>
Chris Lattner	b9488a6	2007-12-18 06:18:21 +0000	[diff] [blame]	48	<li><a href="#t_integer">Integer Type</a></li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	49	<li><a href="#t_array">Array Type</a></li>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	50	<li><a href="#t_function">Function Type</a></li>
				51	<li><a href="#t_pointer">Pointer Type</a></li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	52	<li><a href="#t_struct">Structure Type</a></li>
Andrew Lenharth	75e1068	2006-12-08 17:13:00 +0000	[diff] [blame]	53	<li><a href="#t_pstruct">Packed Structure Type</a></li>
Reid Spencer	485bad1	2007-02-15 03:07:05 +0000	[diff] [blame]	54	<li><a href="#t_vector">Vector Type</a></li>
Chris Lattner	69c11bb	2005-04-25 17:34:15 +0000	[diff] [blame]	55	<li><a href="#t_opaque">Opaque Type</a></li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	56	</ol>
				57	</li>
Chris Lattner	242d61d	2009-02-02 07:32:36 +0000	[diff] [blame]	58	<li><a href="#t_uprefs">Type Up-references</a></li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	59	</ol>
				60	</li>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	61	<li><a href="#constants">Constants</a>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	62	<ol>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	63	<li><a href="#simpleconstants">Simple Constants</a></li>
Chris Lattner	7088279	2009-02-28 18:32:25 +0000	[diff] [blame]	64	<li><a href="#complexconstants">Complex Constants</a></li>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	65	<li><a href="#globalconstants">Global Variable and Function Addresses</a></li>
				66	<li><a href="#undefvalues">Undefined Values</a></li>
				67	<li><a href="#constantexprs">Constant Expressions</a></li>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	68	</ol>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	69	</li>
Chris Lattner	e87d653	2006-01-25 23:47:57 +0000	[diff] [blame]	70	<li><a href="#othervalues">Other Values</a>
				71	<ol>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	72	<li><a href="#inlineasm">Inline Assembler Expressions</a></li>
Chris Lattner	e87d653	2006-01-25 23:47:57 +0000	[diff] [blame]	73	</ol>
				74	</li>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	75	<li><a href="#instref">Instruction Reference</a>
				76	<ol>
				77	<li><a href="#terminators">Terminator Instructions</a>
				78	<ol>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	79	<li><a href="#i_ret">'<tt>ret</tt>' Instruction</a></li>
				80	<li><a href="#i_br">'<tt>br</tt>' Instruction</a></li>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	81	<li><a href="#i_switch">'<tt>switch</tt>' Instruction</a></li>
				82	<li><a href="#i_invoke">'<tt>invoke</tt>' Instruction</a></li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	83	<li><a href="#i_unwind">'<tt>unwind</tt>' Instruction</a></li>
Chris Lattner	35eca58	2004-10-16 18:04:13 +0000	[diff] [blame]	84	<li><a href="#i_unreachable">'<tt>unreachable</tt>' Instruction</a></li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	85	</ol>
				86	</li>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	87	<li><a href="#binaryops">Binary Operations</a>
				88	<ol>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	89	<li><a href="#i_add">'<tt>add</tt>' Instruction</a></li>
				90	<li><a href="#i_sub">'<tt>sub</tt>' Instruction</a></li>
				91	<li><a href="#i_mul">'<tt>mul</tt>' Instruction</a></li>
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	92	<li><a href="#i_udiv">'<tt>udiv</tt>' Instruction</a></li>
				93	<li><a href="#i_sdiv">'<tt>sdiv</tt>' Instruction</a></li>
				94	<li><a href="#i_fdiv">'<tt>fdiv</tt>' Instruction</a></li>
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	95	<li><a href="#i_urem">'<tt>urem</tt>' Instruction</a></li>
				96	<li><a href="#i_srem">'<tt>srem</tt>' Instruction</a></li>
				97	<li><a href="#i_frem">'<tt>frem</tt>' Instruction</a></li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	98	</ol>
				99	</li>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	100	<li><a href="#bitwiseops">Bitwise Binary Operations</a>
				101	<ol>
Reid Spencer	8e11bf8	2007-02-02 13:57:07 +0000	[diff] [blame]	102	<li><a href="#i_shl">'<tt>shl</tt>' Instruction</a></li>
				103	<li><a href="#i_lshr">'<tt>lshr</tt>' Instruction</a></li>
				104	<li><a href="#i_ashr">'<tt>ashr</tt>' Instruction</a></li>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	105	<li><a href="#i_and">'<tt>and</tt>' Instruction</a></li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	106	<li><a href="#i_or">'<tt>or</tt>' Instruction</a></li>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	107	<li><a href="#i_xor">'<tt>xor</tt>' Instruction</a></li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	108	</ol>
				109	</li>
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	110	<li><a href="#vectorops">Vector Operations</a>
				111	<ol>
				112	<li><a href="#i_extractelement">'<tt>extractelement</tt>' Instruction</a></li>
				113	<li><a href="#i_insertelement">'<tt>insertelement</tt>' Instruction</a></li>
				114	<li><a href="#i_shufflevector">'<tt>shufflevector</tt>' Instruction</a></li>
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	115	</ol>
				116	</li>
Dan Gohman	a334d5f	2008-05-12 23:51:09 +0000	[diff] [blame]	117	<li><a href="#aggregateops">Aggregate Operations</a>
				118	<ol>
				119	<li><a href="#i_extractvalue">'<tt>extractvalue</tt>' Instruction</a></li>
				120	<li><a href="#i_insertvalue">'<tt>insertvalue</tt>' Instruction</a></li>
				121	</ol>
				122	</li>
Chris Lattner	884a970	2006-08-15 00:45:58 +0000	[diff] [blame]	123	<li><a href="#memoryops">Memory Access and Addressing Operations</a>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	124	<ol>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	125	<li><a href="#i_malloc">'<tt>malloc</tt>' Instruction</a></li>
				126	<li><a href="#i_free">'<tt>free</tt>' Instruction</a></li>
				127	<li><a href="#i_alloca">'<tt>alloca</tt>' Instruction</a></li>
Robert Bocchino	7b81c75	2006-02-17 21:18:08 +0000	[diff] [blame]	128	<li><a href="#i_load">'<tt>load</tt>' Instruction</a></li>
				129	<li><a href="#i_store">'<tt>store</tt>' Instruction</a></li>
				130	<li><a href="#i_getelementptr">'<tt>getelementptr</tt>' Instruction</a></li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	131	</ol>
				132	</li>
Reid Spencer	2fd21e6	2006-11-08 01:18:52 +0000	[diff] [blame]	133	<li><a href="#convertops">Conversion Operations</a>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	134	<ol>
				135	<li><a href="#i_trunc">'<tt>trunc .. to</tt>' Instruction</a></li>
				136	<li><a href="#i_zext">'<tt>zext .. to</tt>' Instruction</a></li>
				137	<li><a href="#i_sext">'<tt>sext .. to</tt>' Instruction</a></li>
				138	<li><a href="#i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a></li>
				139	<li><a href="#i_fpext">'<tt>fpext .. to</tt>' Instruction</a></li>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	140	<li><a href="#i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a></li>
				141	<li><a href="#i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a></li>
				142	<li><a href="#i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a></li>
				143	<li><a href="#i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a></li>
Reid Spencer	7267925	2006-11-11 21:00:47 +0000	[diff] [blame]	144	<li><a href="#i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a></li>
				145	<li><a href="#i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a></li>
Reid Spencer	5c0ef47	2006-11-11 23:08:07 +0000	[diff] [blame]	146	<li><a href="#i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a></li>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	147	</ol>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	148	</li>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	149	<li><a href="#otherops">Other Operations</a>
				150	<ol>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	151	<li><a href="#i_icmp">'<tt>icmp</tt>' Instruction</a></li>
				152	<li><a href="#i_fcmp">'<tt>fcmp</tt>' Instruction</a></li>
Nate Begeman	ac80ade	2008-05-12 19:01:56 +0000	[diff] [blame]	153	<li><a href="#i_vicmp">'<tt>vicmp</tt>' Instruction</a></li>
				154	<li><a href="#i_vfcmp">'<tt>vfcmp</tt>' Instruction</a></li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	155	<li><a href="#i_phi">'<tt>phi</tt>' Instruction</a></li>
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	156	<li><a href="#i_select">'<tt>select</tt>' Instruction</a></li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	157	<li><a href="#i_call">'<tt>call</tt>' Instruction</a></li>
Chris Lattner	fb6977d	2006-01-13 23:26:01 +0000	[diff] [blame]	158	<li><a href="#i_va_arg">'<tt>va_arg</tt>' Instruction</a></li>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	159	</ol>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	160	</li>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	161	</ol>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	162	</li>
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	163	<li><a href="#intrinsics">Intrinsic Functions</a>
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	164	<ol>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	165	<li><a href="#int_varargs">Variable Argument Handling Intrinsics</a>
				166	<ol>
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	167	<li><a href="#int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a></li>
				168	<li><a href="#int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a></li>
				169	<li><a href="#int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a></li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	170	</ol>
				171	</li>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	172	<li><a href="#int_gc">Accurate Garbage Collection Intrinsics</a>
				173	<ol>
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	174	<li><a href="#int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a></li>
				175	<li><a href="#int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a></li>
				176	<li><a href="#int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a></li>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	177	</ol>
				178	</li>
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	179	<li><a href="#int_codegen">Code Generator Intrinsics</a>
				180	<ol>
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	181	<li><a href="#int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a></li>
				182	<li><a href="#int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a></li>
				183	<li><a href="#int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a></li>
				184	<li><a href="#int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a></li>
				185	<li><a href="#int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a></li>
				186	<li><a href="#int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a></li>
				187	<li><a href="#int_readcyclecounter"><tt>llvm.readcyclecounter</tt>' Intrinsic</a></li>
John Criswell	7123e27	2004-04-09 16:43:20 +0000	[diff] [blame]	188	</ol>
				189	</li>
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	190	<li><a href="#int_libc">Standard C Library Intrinsics</a>
				191	<ol>
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	192	<li><a href="#int_memcpy">'<tt>llvm.memcpy.*</tt>' Intrinsic</a></li>
				193	<li><a href="#int_memmove">'<tt>llvm.memmove.*</tt>' Intrinsic</a></li>
				194	<li><a href="#int_memset">'<tt>llvm.memset.*</tt>' Intrinsic</a></li>
				195	<li><a href="#int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a></li>
				196	<li><a href="#int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a></li>
Dan Gohman	91c284c	2007-10-15 20:30:11 +0000	[diff] [blame]	197	<li><a href="#int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a></li>
				198	<li><a href="#int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a></li>
				199	<li><a href="#int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a></li>
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	200	</ol>
				201	</li>
Nate Begeman	7e36c47	2006-01-13 23:26:38 +0000	[diff] [blame]	202	<li><a href="#int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	203	<ol>
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	204	<li><a href="#int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a></li>
Chris Lattner	8a886be	2006-01-16 22:34:14 +0000	[diff] [blame]	205	<li><a href="#int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic </a></li>
				206	<li><a href="#int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic </a></li>
				207	<li><a href="#int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic </a></li>
Reid Spencer	f86037f	2007-04-11 23:23:49 +0000	[diff] [blame]	208	<li><a href="#int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic </a></li>
				209	<li><a href="#int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic </a></li>
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	210	</ol>
				211	</li>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	212	<li><a href="#int_overflow">Arithmetic with Overflow Intrinsics</a>
				213	<ol>
Bill Wendling	da01af7	2009-02-08 04:04:40 +0000	[diff] [blame]	214	<li><a href="#int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt> Intrinsics</a></li>
				215	<li><a href="#int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt> Intrinsics</a></li>
				216	<li><a href="#int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt> Intrinsics</a></li>
				217	<li><a href="#int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt> Intrinsics</a></li>
				218	<li><a href="#int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendling	41b485c	2009-02-08 23:00:09 +0000	[diff] [blame]	219	<li><a href="#int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt> Intrinsics</a></li>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	220	</ol>
				221	</li>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	222	<li><a href="#int_debugger">Debugger intrinsics</a></li>
Jim Laskey	dd4ef1b	2007-03-14 19:31:19 +0000	[diff] [blame]	223	<li><a href="#int_eh">Exception Handling intrinsics</a></li>
Duncan Sands	f7331b3	2007-09-11 14:10:23 +0000	[diff] [blame]	224	<li><a href="#int_trampoline">Trampoline Intrinsic</a>
Duncan Sands	36397f5	2007-07-27 12:58:54 +0000	[diff] [blame]	225	<ol>
				226	<li><a href="#int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a></li>
Duncan Sands	36397f5	2007-07-27 12:58:54 +0000	[diff] [blame]	227	</ol>
				228	</li>
Bill Wendling	3c44f5b	2008-11-18 22:10:53 +0000	[diff] [blame]	229	<li><a href="#int_atomics">Atomic intrinsics</a>
				230	<ol>
				231	<li><a href="#int_memory_barrier"><tt>llvm.memory_barrier</tt></a></li>
				232	<li><a href="#int_atomic_cmp_swap"><tt>llvm.atomic.cmp.swap</tt></a></li>
				233	<li><a href="#int_atomic_swap"><tt>llvm.atomic.swap</tt></a></li>
				234	<li><a href="#int_atomic_load_add"><tt>llvm.atomic.load.add</tt></a></li>
				235	<li><a href="#int_atomic_load_sub"><tt>llvm.atomic.load.sub</tt></a></li>
				236	<li><a href="#int_atomic_load_and"><tt>llvm.atomic.load.and</tt></a></li>
				237	<li><a href="#int_atomic_load_nand"><tt>llvm.atomic.load.nand</tt></a></li>
				238	<li><a href="#int_atomic_load_or"><tt>llvm.atomic.load.or</tt></a></li>
				239	<li><a href="#int_atomic_load_xor"><tt>llvm.atomic.load.xor</tt></a></li>
				240	<li><a href="#int_atomic_load_max"><tt>llvm.atomic.load.max</tt></a></li>
				241	<li><a href="#int_atomic_load_min"><tt>llvm.atomic.load.min</tt></a></li>
				242	<li><a href="#int_atomic_load_umax"><tt>llvm.atomic.load.umax</tt></a></li>
				243	<li><a href="#int_atomic_load_umin"><tt>llvm.atomic.load.umin</tt></a></li>
				244	</ol>
				245	</li>
Reid Spencer	2067764	2007-07-20 19:59:11 +0000	[diff] [blame]	246	<li><a href="#int_general">General intrinsics</a>
Tanya Lattner	6d806e9	2007-06-15 20:50:54 +0000	[diff] [blame]	247	<ol>
Reid Spencer	2067764	2007-07-20 19:59:11 +0000	[diff] [blame]	248	<li><a href="#int_var_annotation">
Bill Wendling	69e4adb	2008-11-19 05:56:17 +0000	[diff] [blame]	249	'<tt>llvm.var.annotation</tt>' Intrinsic</a></li>
Tanya Lattner	b636788	2007-09-21 22:59:12 +0000	[diff] [blame]	250	<li><a href="#int_annotation">
Bill Wendling	69e4adb	2008-11-19 05:56:17 +0000	[diff] [blame]	251	'<tt>llvm.annotation.*</tt>' Intrinsic</a></li>
Anton Korobeynikov	4cb8618	2008-01-15 22:31:34 +0000	[diff] [blame]	252	<li><a href="#int_trap">
Bill Wendling	69e4adb	2008-11-19 05:56:17 +0000	[diff] [blame]	253	'<tt>llvm.trap</tt>' Intrinsic</a></li>
				254	<li><a href="#int_stackprotector">
				255	'<tt>llvm.stackprotector</tt>' Intrinsic</a></li>
Tanya Lattner	b636788	2007-09-21 22:59:12 +0000	[diff] [blame]	256	</ol>
Tanya Lattner	6d806e9	2007-06-15 20:50:54 +0000	[diff] [blame]	257	</li>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	258	</ol>
				259	</li>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	260	</ol>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	261
				262	<div class="doc_author">
				263	<p>Written by <a href="mailto:sabre@nondot.org">Chris Lattner</a>
				264	and <a href="mailto:vadve@cs.uiuc.edu">Vikram Adve</a></p>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	265	</div>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	266
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	267	<!-- *********************************************************************** -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	268	<div class="doc_section"> <a name="abstract">Abstract </a></div>
				269	<!-- *********************************************************************** -->
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	270
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	271	<div class="doc_text">
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	272	<p>This document is a reference manual for the LLVM assembly language.
Bill Wendling	837f39b	2008-08-05 22:29:16 +0000	[diff] [blame]	273	LLVM is a Static Single Assignment (SSA) based representation that provides
Chris Lattner	d3eda89	2008-08-05 18:29:16 +0000	[diff] [blame]	274	type safety, low-level operations, flexibility, and the capability of
				275	representing 'all' high-level languages cleanly. It is the common code
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	276	representation used throughout all phases of the LLVM compilation
				277	strategy.</p>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	278	</div>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	279
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	280	<!-- *********************************************************************** -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	281	<div class="doc_section"> <a name="introduction">Introduction</a> </div>
				282	<!-- *********************************************************************** -->
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	283
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	284	<div class="doc_text">
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	285
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	286	<p>The LLVM code representation is designed to be used in three
Gabor Greif	04367bf	2007-07-06 22:07:22 +0000	[diff] [blame]	287	different forms: as an in-memory compiler IR, as an on-disk bitcode
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	288	representation (suitable for fast loading by a Just-In-Time compiler),
				289	and as a human readable assembly language representation. This allows
				290	LLVM to provide a powerful intermediate representation for efficient
				291	compiler transformations and analysis, while providing a natural means
				292	to debug and visualize the transformations. The three different forms
				293	of LLVM are all equivalent. This document describes the human readable
				294	representation and notation.</p>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	295
John Criswell	c1f786c	2005-05-13 22:25:59 +0000	[diff] [blame]	296	<p>The LLVM representation aims to be light-weight and low-level
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	297	while being expressive, typed, and extensible at the same time. It
				298	aims to be a "universal IR" of sorts, by being at a low enough level
				299	that high-level ideas may be cleanly mapped to it (similar to how
				300	microprocessors are "universal IR's", allowing many source languages to
				301	be mapped to them). By providing type information, LLVM can be used as
				302	the target of optimizations: for example, through pointer analysis, it
				303	can be proven that a C automatic variable is never accessed outside of
				304	the current function... allowing it to be promoted to a simple SSA
				305	value instead of a memory location.</p>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	306
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	307	</div>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	308
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	309	<!-- _______________________________________________________________________ -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	310	<div class="doc_subsubsection"> <a name="wellformed">Well-Formedness</a> </div>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	311
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	312	<div class="doc_text">
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	313
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	314	<p>It is important to note that this document describes 'well formed'
				315	LLVM assembly language. There is a difference between what the parser
				316	accepts and what is considered 'well formed'. For example, the
				317	following instruction is syntactically okay, but not well formed:</p>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	318
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	319	<div class="doc_code">
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	320	<pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	321	%x = <a href="#i_add">add</a> i32 1, %x
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	322	</pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	323	</div>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	324
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	325	<p>...because the definition of <tt>%x</tt> does not dominate all of
				326	its uses. The LLVM infrastructure provides a verification pass that may
				327	be used to verify that an LLVM module is well formed. This pass is
John Criswell	c1f786c	2005-05-13 22:25:59 +0000	[diff] [blame]	328	automatically run by the parser after parsing input assembly and by
Gabor Greif	04367bf	2007-07-06 22:07:22 +0000	[diff] [blame]	329	the optimizer before it outputs bitcode. The violations pointed out
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	330	by the verifier pass indicate bugs in transformation passes or input to
				331	the parser.</p>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	332	</div>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	333
Chris Lattner	cc68939	2007-10-03 17:34:29 +0000	[diff] [blame]	334	<!-- Describe the typesetting conventions here. -->
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	335
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	336	<!-- *********************************************************************** -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	337	<div class="doc_section"> <a name="identifiers">Identifiers</a> </div>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	338	<!-- *********************************************************************** -->
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	339
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	340	<div class="doc_text">
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	341
Reid Spencer	2c45228	2007-08-07 14:34:28 +0000	[diff] [blame]	342	<p>LLVM identifiers come in two basic types: global and local. Global
				343	identifiers (functions, global variables) begin with the @ character. Local
				344	identifiers (register names, types) begin with the % character. Additionally,
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	345	there are three different formats for identifiers, for different purposes:</p>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	346
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	347	<ol>
Reid Spencer	2c45228	2007-08-07 14:34:28 +0000	[diff] [blame]	348	<li>Named values are represented as a string of characters with their prefix.
				349	For example, %foo, @DivisionByZero, %a.really.long.identifier. The actual
				350	regular expression used is '<tt>[%@][a-zA-Z$._][a-zA-Z$._0-9]*</tt>'.
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	351	Identifiers which require other characters in their names can be surrounded
Daniel Dunbar	76dea95	2008-10-14 23:51:43 +0000	[diff] [blame]	352	with quotes. Special characters may be escaped using "\xx" where xx is the
				353	ASCII code for the character in hexadecimal. In this way, any character can
				354	be used in a name value, even quotes themselves.
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	355
Reid Spencer	2c45228	2007-08-07 14:34:28 +0000	[diff] [blame]	356	<li>Unnamed values are represented as an unsigned numeric value with their
				357	prefix. For example, %12, @2, %44.</li>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	358
Reid Spencer	cc16dc3	2004-12-09 18:02:53 +0000	[diff] [blame]	359	<li>Constants, which are described in a <a href="#constants">section about
				360	constants</a>, below.</li>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	361	</ol>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	362
Reid Spencer	2c45228	2007-08-07 14:34:28 +0000	[diff] [blame]	363	<p>LLVM requires that values start with a prefix for two reasons: Compilers
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	364	don't need to worry about name clashes with reserved words, and the set of
				365	reserved words may be expanded in the future without penalty. Additionally,
				366	unnamed identifiers allow a compiler to quickly come up with a temporary
				367	variable without having to avoid symbol table conflicts.</p>
				368
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	369	<p>Reserved words in LLVM are very similar to reserved words in other
Reid Spencer	5c0ef47	2006-11-11 23:08:07 +0000	[diff] [blame]	370	languages. There are keywords for different opcodes
				371	('<tt><a href="#i_add">add</a></tt>',
				372	'<tt><a href="#i_bitcast">bitcast</a></tt>',
				373	'<tt><a href="#i_ret">ret</a></tt>', etc...), for primitive type names ('<tt><a
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	374	href="#t_void">void</a></tt>', '<tt><a href="#t_primitive">i32</a></tt>', etc...),
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	375	and others. These reserved words cannot conflict with variable names, because
Reid Spencer	2c45228	2007-08-07 14:34:28 +0000	[diff] [blame]	376	none of them start with a prefix character ('%' or '@').</p>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	377
				378	<p>Here is an example of LLVM code to multiply the integer variable
				379	'<tt>%X</tt>' by 8:</p>
				380
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	381	<p>The easy way:</p>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	382
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	383	<div class="doc_code">
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	384	<pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	385	%result = <a href="#i_mul">mul</a> i32 %X, 8
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	386	</pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	387	</div>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	388
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	389	<p>After strength reduction:</p>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	390
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	391	<div class="doc_code">
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	392	<pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	393	%result = <a href="#i_shl">shl</a> i32 %X, i8 3
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	394	</pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	395	</div>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	396
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	397	<p>And the hard way:</p>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	398
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	399	<div class="doc_code">
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	400	<pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	401	<a href="#i_add">add</a> i32 %X, %X <i>; yields {i32}:%0</i>
				402	<a href="#i_add">add</a> i32 %0, %0 <i>; yields {i32}:%1</i>
				403	%result = <a href="#i_add">add</a> i32 %1, %1
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	404	</pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	405	</div>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	406
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	407	<p>This last way of multiplying <tt>%X</tt> by 8 illustrates several
				408	important lexical features of LLVM:</p>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	409
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	410	<ol>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	411
				412	<li>Comments are delimited with a '<tt>;</tt>' and go until the end of
				413	line.</li>
				414
				415	<li>Unnamed temporaries are created when the result of a computation is not
				416	assigned to a named value.</li>
				417
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	418	<li>Unnamed temporaries are numbered sequentially</li>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	419
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	420	</ol>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	421
John Criswell	e4c57cc	2005-05-12 16:52:32 +0000	[diff] [blame]	422	<p>...and it also shows a convention that we follow in this document. When
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	423	demonstrating instructions, we will follow an instruction with a comment that
				424	defines the type and name of value produced. Comments are shown in italic
				425	text.</p>
				426
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	427	</div>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	428
				429	<!-- *********************************************************************** -->
				430	<div class="doc_section"> <a name="highlevel">High Level Structure</a> </div>
				431	<!-- *********************************************************************** -->
				432
				433	<!-- ======================================================================= -->
				434	<div class="doc_subsection"> <a name="modulestructure">Module Structure</a>
				435	</div>
				436
				437	<div class="doc_text">
				438
				439	<p>LLVM programs are composed of "Module"s, each of which is a
				440	translation unit of the input programs. Each module consists of
				441	functions, global variables, and symbol table entries. Modules may be
				442	combined together with the LLVM linker, which merges function (and
				443	global variable) definitions, resolves forward declarations, and merges
				444	symbol table entries. Here is an example of the "hello world" module:</p>
				445
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	446	<div class="doc_code">
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	447	<pre><i>; Declare the string constant as a global constant...</i>
Chris Lattner	a89e5f1	2007-06-12 17:00:26 +0000	[diff] [blame]	448	<a href="#identifiers">@.LC0</a> = <a href="#linkage_internal">internal</a> <a
				449	href="#globalvars">constant</a> <a href="#t_array">[13 x i8]</a> c"hello world\0A\00" <i>; [13 x i8]*</i>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	450
				451	<i>; External declaration of the puts function</i>
Chris Lattner	a89e5f1	2007-06-12 17:00:26 +0000	[diff] [blame]	452	<a href="#functionstructure">declare</a> i32 @puts(i8 ) <i>; i32(i8 )* </i>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	453
				454	<i>; Definition of main function</i>
Chris Lattner	a89e5f1	2007-06-12 17:00:26 +0000	[diff] [blame]	455	define i32 @main() { <i>; i32()* </i>
Dan Gohman	2a08c53	2009-01-04 23:44:43 +0000	[diff] [blame]	456	<i>; Convert [13 x i8]* to i8 *...</i>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	457	%cast210 = <a
Dan Gohman	2a08c53	2009-01-04 23:44:43 +0000	[diff] [blame]	458	href="#i_getelementptr">getelementptr</a> [13 x i8]* @.LC0, i64 0, i64 0 <i>; i8 *</i>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	459
				460	<i>; Call puts function to write out the string to stdout...</i>
				461	<a
Chris Lattner	a89e5f1	2007-06-12 17:00:26 +0000	[diff] [blame]	462	href="#i_call">call</a> i32 @puts(i8 * %cast210) <i>; i32</i>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	463	<a
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	464	href="#i_ret">ret</a> i32 0<br>}<br>
				465	</pre>
				466	</div>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	467
				468	<p>This example is made up of a <a href="#globalvars">global variable</a>
				469	named "<tt>.LC0</tt>", an external declaration of the "<tt>puts</tt>"
				470	function, and a <a href="#functionstructure">function definition</a>
				471	for "<tt>main</tt>".</p>
				472
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	473	<p>In general, a module is made up of a list of global values,
				474	where both functions and global variables are global values. Global values are
				475	represented by a pointer to a memory location (in this case, a pointer to an
				476	array of char, and a pointer to a function), and have one of the following <a
				477	href="#linkage">linkage types</a>.</p>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	478
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	479	</div>
				480
				481	<!-- ======================================================================= -->
				482	<div class="doc_subsection">
				483	<a name="linkage">Linkage Types</a>
				484	</div>
				485
				486	<div class="doc_text">
				487
				488	<p>
				489	All Global Variables and Functions have one of the following types of linkage:
				490	</p>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	491
				492	<dl>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	493
Rafael Espindola	bb46f52	2009-01-15 20:18:42 +0000	[diff] [blame]	494	<dt><tt><b><a name="linkage_private">private</a></b></tt>: </dt>
				495
				496	<dd>Global values with private linkage are only directly accessible by
				497	objects in the current module. In particular, linking code into a module with
				498	an private global value may cause the private to be renamed as necessary to
				499	avoid collisions. Because the symbol is private to the module, all
				500	references can be updated. This doesn't show up in any symbol table in the
				501	object file.
				502	</dd>
				503
Dale Johannesen	2307a7f	2008-05-23 23:13:41 +0000	[diff] [blame]	504	<dt><tt><b><a name="linkage_internal">internal</a></b></tt>: </dt>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	505
Duncan Sands	81d05c2	2009-01-16 09:29:46 +0000	[diff] [blame]	506	<dd> Similar to private, but the value shows as a local symbol (STB_LOCAL in
Rafael Espindola	bb46f52	2009-01-15 20:18:42 +0000	[diff] [blame]	507	the case of ELF) in the object file. This corresponds to the notion of the
Chris Lattner	4887bd8	2007-01-14 06:51:48 +0000	[diff] [blame]	508	'<tt>static</tt>' keyword in C.
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	509	</dd>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	510
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	511	<dt><tt><b><a name="linkage_linkonce">linkonce</a></b></tt>: </dt>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	512
Chris Lattner	4887bd8	2007-01-14 06:51:48 +0000	[diff] [blame]	513	<dd>Globals with "<tt>linkonce</tt>" linkage are merged with other globals of
				514	the same name when linkage occurs. This is typically used to implement
				515	inline functions, templates, or other code which must be generated in each
				516	translation unit that uses it. Unreferenced <tt>linkonce</tt> globals are
				517	allowed to be discarded.
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	518	</dd>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	519
Dale Johannesen	2307a7f	2008-05-23 23:13:41 +0000	[diff] [blame]	520	<dt><tt><b><a name="linkage_common">common</a></b></tt>: </dt>
				521
				522	<dd>"<tt>common</tt>" linkage is exactly the same as <tt>linkonce</tt>
				523	linkage, except that unreferenced <tt>common</tt> globals may not be
				524	discarded. This is used for globals that may be emitted in multiple
				525	translation units, but that are not guaranteed to be emitted into every
				526	translation unit that uses them. One example of this is tentative
				527	definitions in C, such as "<tt>int X;</tt>" at global scope.
				528	</dd>
				529
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	530	<dt><tt><b><a name="linkage_weak">weak</a></b></tt>: </dt>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	531
Dale Johannesen	2307a7f	2008-05-23 23:13:41 +0000	[diff] [blame]	532	<dd>"<tt>weak</tt>" linkage is the same as <tt>common</tt> linkage, except
				533	that some targets may choose to emit different assembly sequences for them
				534	for target-dependent reasons. This is used for globals that are declared
				535	"weak" in C source code.
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	536	</dd>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	537
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	538	<dt><tt><b><a name="linkage_appending">appending</a></b></tt>: </dt>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	539
				540	<dd>"<tt>appending</tt>" linkage may only be applied to global variables of
				541	pointer to array type. When two global variables with appending linkage are
				542	linked together, the two global arrays are appended together. This is the
				543	LLVM, typesafe, equivalent of having the system linker append together
				544	"sections" with identical names when .o files are linked.
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	545	</dd>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	546
Anton Korobeynikov	7f70559	2007-01-12 19:20:47 +0000	[diff] [blame]	547	<dt><tt><b><a name="linkage_externweak">extern_weak</a></b></tt>: </dt>
Duncan Sands	667d4b8	2009-03-07 15:45:40 +0000	[diff] [blame]	548
Chris Lattner	d3eda89	2008-08-05 18:29:16 +0000	[diff] [blame]	549	<dd>The semantics of this linkage follow the ELF object file model: the
				550	symbol is weak until linked, if not linked, the symbol becomes null instead
				551	of being an undefined reference.
Anton Korobeynikov	7f70559	2007-01-12 19:20:47 +0000	[diff] [blame]	552	</dd>
Anton Korobeynikov	7f70559	2007-01-12 19:20:47 +0000	[diff] [blame]	553
Duncan Sands	667d4b8	2009-03-07 15:45:40 +0000	[diff] [blame]	554	<dt><tt><b><a name="linkage_linkonce">linkonce_odr</a></b></tt>: </dt>
				555	<dt><tt><b><a name="linkage_common">common_odr</a></b></tt>: </dt>
				556	<dt><tt><b><a name="linkage_weak">weak_odr</a></b></tt>: </dt>
Duncan Sands	667d4b8	2009-03-07 15:45:40 +0000	[diff] [blame]	557	<dd>Some languages allow inequivalent globals to be merged, such as two
				558	functions with different semantics. Other languages, such as <tt>C++</tt>,
				559	ensure that only equivalent globals are ever merged (the "one definition
				560	rule" - <tt>odr</tt>). Such languages can use the <tt>linkonce_odr</tt>,
Duncan Sands	5f4ee1f	2009-03-11 08:08:06 +0000	[diff] [blame]	561	<tt>common_odr</tt> and <tt>weak_odr</tt> linkage types to indicate that
				562	the global will only be merged with equivalent globals. These linkage
				563	types are otherwise the same as their non-<tt>odr</tt> versions.
Duncan Sands	667d4b8	2009-03-07 15:45:40 +0000	[diff] [blame]	564	</dd>
				565
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	566	<dt><tt><b><a name="linkage_external">externally visible</a></b></tt>:</dt>
Chris Lattner	e5d947b	2004-12-09 16:36:40 +0000	[diff] [blame]	567
				568	<dd>If none of the above identifiers are used, the global is externally
				569	visible, meaning that it participates in linkage and can be used to resolve
				570	external symbol references.
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	571	</dd>
Reid Spencer	c891084	2007-04-11 23:49:50 +0000	[diff] [blame]	572	</dl>
Anton Korobeynikov	b74ed07	2006-09-14 18:23:27 +0000	[diff] [blame]	573
Anton Korobeynikov	b74ed07	2006-09-14 18:23:27 +0000	[diff] [blame]	574	<p>
				575	The next two types of linkage are targeted for Microsoft Windows platform
				576	only. They are designed to support importing (exporting) symbols from (to)
Chris Lattner	d3eda89	2008-08-05 18:29:16 +0000	[diff] [blame]	577	DLLs (Dynamic Link Libraries).
Anton Korobeynikov	b74ed07	2006-09-14 18:23:27 +0000	[diff] [blame]	578	</p>
				579
Anton Korobeynikov	7f70559	2007-01-12 19:20:47 +0000	[diff] [blame]	580	<dl>
Anton Korobeynikov	b74ed07	2006-09-14 18:23:27 +0000	[diff] [blame]	581	<dt><tt><b><a name="linkage_dllimport">dllimport</a></b></tt>: </dt>
				582
				583	<dd>"<tt>dllimport</tt>" linkage causes the compiler to reference a function
				584	or variable via a global pointer to a pointer that is set up by the DLL
				585	exporting the symbol. On Microsoft Windows targets, the pointer name is
Dan Gohman	7956412	2009-01-12 21:35:55 +0000	[diff] [blame]	586	formed by combining <code>__imp_</code> and the function or variable name.
Anton Korobeynikov	b74ed07	2006-09-14 18:23:27 +0000	[diff] [blame]	587	</dd>
				588
				589	<dt><tt><b><a name="linkage_dllexport">dllexport</a></b></tt>: </dt>
				590
				591	<dd>"<tt>dllexport</tt>" linkage causes the compiler to provide a global
				592	pointer to a pointer in a DLL, so that it can be referenced with the
				593	<tt>dllimport</tt> attribute. On Microsoft Windows targets, the pointer
Dan Gohman	7956412	2009-01-12 21:35:55 +0000	[diff] [blame]	594	name is formed by combining <code>__imp_</code> and the function or variable
Anton Korobeynikov	b74ed07	2006-09-14 18:23:27 +0000	[diff] [blame]	595	name.
				596	</dd>
				597
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	598	</dl>
				599
Dan Gohman	f003276	2008-11-24 17:18:39 +0000	[diff] [blame]	600	<p>For example, since the "<tt>.LC0</tt>"
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	601	variable is defined to be internal, if another module defined a "<tt>.LC0</tt>"
				602	variable and was linked with this one, one of the two would be renamed,
				603	preventing a collision. Since "<tt>main</tt>" and "<tt>puts</tt>" are
				604	external (i.e., lacking any linkage declarations), they are accessible
Reid Spencer	ac8d276	2007-01-05 00:59:10 +0000	[diff] [blame]	605	outside of the current module.</p>
				606	<p>It is illegal for a function <i>declaration</i>
Duncan Sands	5f4ee1f	2009-03-11 08:08:06 +0000	[diff] [blame]	607	to have any linkage type other than "externally visible", <tt>dllimport</tt>
				608	or <tt>extern_weak</tt>.</p>
Duncan Sands	667d4b8	2009-03-07 15:45:40 +0000	[diff] [blame]	609	<p>Aliases can have only <tt>external</tt>, <tt>internal</tt>, <tt>weak</tt>
				610	or <tt>weak_odr</tt> linkages.</p>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	611	</div>
				612
				613	<!-- ======================================================================= -->
				614	<div class="doc_subsection">
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	615	<a name="callingconv">Calling Conventions</a>
				616	</div>
				617
				618	<div class="doc_text">
				619
				620	<p>LLVM <a href="#functionstructure">functions</a>, <a href="#i_call">calls</a>
				621	and <a href="#i_invoke">invokes</a> can all have an optional calling convention
				622	specified for the call. The calling convention of any pair of dynamic
				623	caller/callee must match, or the behavior of the program is undefined. The
				624	following calling conventions are supported by LLVM, and more may be added in
				625	the future:</p>
				626
				627	<dl>
				628	<dt><b>"<tt>ccc</tt>" - The C calling convention</b>:</dt>
				629
				630	<dd>This calling convention (the default if no other calling convention is
				631	specified) matches the target C calling conventions. This calling convention
John Criswell	e4c57cc	2005-05-12 16:52:32 +0000	[diff] [blame]	632	supports varargs function calls and tolerates some mismatch in the declared
Reid Spencer	c28d2bc	2006-12-31 21:30:18 +0000	[diff] [blame]	633	prototype and implemented declaration of the function (as does normal C).
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	634	</dd>
				635
				636	<dt><b>"<tt>fastcc</tt>" - The fast calling convention</b>:</dt>
				637
				638	<dd>This calling convention attempts to make calls as fast as possible
				639	(e.g. by passing things in registers). This calling convention allows the
				640	target to use whatever tricks it wants to produce fast code for the target,
Chris Lattner	d3eda89	2008-08-05 18:29:16 +0000	[diff] [blame]	641	without having to conform to an externally specified ABI (Application Binary
				642	Interface). Implementations of this convention should allow arbitrary
Arnold Schwaighofer	9097d14	2008-05-14 09:17:12 +0000	[diff] [blame]	643	<a href="CodeGenerator.html#tailcallopt">tail call optimization</a> to be
				644	supported. This calling convention does not support varargs and requires the
				645	prototype of all callees to exactly match the prototype of the function
				646	definition.
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	647	</dd>
				648
				649	<dt><b>"<tt>coldcc</tt>" - The cold calling convention</b>:</dt>
				650
				651	<dd>This calling convention attempts to make code in the caller as efficient
				652	as possible under the assumption that the call is not commonly executed. As
				653	such, these calls often preserve all registers so that the call does not break
				654	any live ranges in the caller side. This calling convention does not support
				655	varargs and requires the prototype of all callees to exactly match the
				656	prototype of the function definition.
				657	</dd>
				658
Chris Lattner	cfe6b37	2005-05-07 01:46:40 +0000	[diff] [blame]	659	<dt><b>"<tt>cc <<em>n</em>></tt>" - Numbered convention</b>:</dt>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	660
				661	<dd>Any calling convention may be specified by number, allowing
				662	target-specific calling conventions to be used. Target specific calling
				663	conventions start at 64.
				664	</dd>
Chris Lattner	cfe6b37	2005-05-07 01:46:40 +0000	[diff] [blame]	665	</dl>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	666
				667	<p>More calling conventions can be added/defined on an as-needed basis, to
				668	support pascal conventions or any other well-known target-independent
				669	convention.</p>
				670
				671	</div>
				672
				673	<!-- ======================================================================= -->
				674	<div class="doc_subsection">
Anton Korobeynikov	8cea37b	2007-01-23 12:35:46 +0000	[diff] [blame]	675	<a name="visibility">Visibility Styles</a>
				676	</div>
				677
				678	<div class="doc_text">
				679
				680	<p>
				681	All Global Variables and Functions have one of the following visibility styles:
				682	</p>
				683
				684	<dl>
				685	<dt><b>"<tt>default</tt>" - Default style</b>:</dt>
				686
Chris Lattner	d3eda89	2008-08-05 18:29:16 +0000	[diff] [blame]	687	<dd>On targets that use the ELF object file format, default visibility means
				688	that the declaration is visible to other
Anton Korobeynikov	8cea37b	2007-01-23 12:35:46 +0000	[diff] [blame]	689	modules and, in shared libraries, means that the declared entity may be
				690	overridden. On Darwin, default visibility means that the declaration is
				691	visible to other modules. Default visibility corresponds to "external
				692	linkage" in the language.
				693	</dd>
				694
				695	<dt><b>"<tt>hidden</tt>" - Hidden style</b>:</dt>
				696
				697	<dd>Two declarations of an object with hidden visibility refer to the same
				698	object if they are in the same shared object. Usually, hidden visibility
				699	indicates that the symbol will not be placed into the dynamic symbol table,
				700	so no other module (executable or shared library) can reference it
				701	directly.
				702	</dd>
				703
Anton Korobeynikov	6f9896f	2007-04-29 18:35:00 +0000	[diff] [blame]	704	<dt><b>"<tt>protected</tt>" - Protected style</b>:</dt>
				705
				706	<dd>On ELF, protected visibility indicates that the symbol will be placed in
				707	the dynamic symbol table, but that references within the defining module will
				708	bind to the local symbol. That is, the symbol cannot be overridden by another
				709	module.
				710	</dd>
Anton Korobeynikov	8cea37b	2007-01-23 12:35:46 +0000	[diff] [blame]	711	</dl>
				712
				713	</div>
				714
				715	<!-- ======================================================================= -->
				716	<div class="doc_subsection">
Chris Lattner	e7886e4	2009-01-11 20:53:49 +0000	[diff] [blame]	717	<a name="namedtypes">Named Types</a>
				718	</div>
				719
				720	<div class="doc_text">
				721
				722	<p>LLVM IR allows you to specify name aliases for certain types. This can make
				723	it easier to read the IR and make the IR more condensed (particularly when
				724	recursive types are involved). An example of a name specification is:
				725	</p>
				726
				727	<div class="doc_code">
				728	<pre>
				729	%mytype = type { %mytype*, i32 }
				730	</pre>
				731	</div>
				732
				733	<p>You may give a name to any <a href="#typesystem">type</a> except "<a
				734	href="t_void">void</a>". Type name aliases may be used anywhere a type is
				735	expected with the syntax "%mytype".</p>
				736
				737	<p>Note that type names are aliases for the structural type that they indicate,
				738	and that you can therefore specify multiple names for the same type. This often
				739	leads to confusing behavior when dumping out a .ll file. Since LLVM IR uses
				740	structural typing, the name is not part of the type. When printing out LLVM IR,
				741	the printer will pick <em>one name</em> to render all types of a particular
				742	shape. This means that if you have code where two different source types end up
				743	having the same LLVM type, that the dumper will sometimes print the "wrong" or
				744	unexpected type. This is an important design point and isn't going to
				745	change.</p>
				746
				747	</div>
				748
Chris Lattner	e7886e4	2009-01-11 20:53:49 +0000	[diff] [blame]	749	<!-- ======================================================================= -->
				750	<div class="doc_subsection">
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	751	<a name="globalvars">Global Variables</a>
				752	</div>
				753
				754	<div class="doc_text">
				755
Chris Lattner	3689a34	2005-02-12 19:30:21 +0000	[diff] [blame]	756	<p>Global variables define regions of memory allocated at compilation time
Chris Lattner	88f6c46	2005-11-12 00:45:07 +0000	[diff] [blame]	757	instead of run-time. Global variables may optionally be initialized, may have
Anton Korobeynikov	8b0a8c8	2007-04-25 14:27:10 +0000	[diff] [blame]	758	an explicit section to be placed in, and may have an optional explicit alignment
				759	specified. A variable may be defined as "thread_local", which means that it
				760	will not be shared by threads (each thread will have a separated copy of the
				761	variable). A variable may be defined as a global "constant," which indicates
				762	that the contents of the variable will <b>never</b> be modified (enabling better
Chris Lattner	3689a34	2005-02-12 19:30:21 +0000	[diff] [blame]	763	optimization, allowing the global data to be placed in the read-only section of
				764	an executable, etc). Note that variables that need runtime initialization
John Criswell	0ec250c	2005-10-24 16:17:18 +0000	[diff] [blame]	765	cannot be marked "constant" as there is a store to the variable.</p>
Chris Lattner	3689a34	2005-02-12 19:30:21 +0000	[diff] [blame]	766
				767	<p>
				768	LLVM explicitly allows <em>declarations</em> of global variables to be marked
				769	constant, even if the final definition of the global is not. This capability
				770	can be used to enable slightly better optimization of the program, but requires
				771	the language definition to guarantee that optimizations based on the
				772	'constantness' are valid for the translation units that do not include the
				773	definition.
				774	</p>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	775
				776	<p>As SSA values, global variables define pointer values that are in
				777	scope (i.e. they dominate) all basic blocks in the program. Global
				778	variables always define a pointer to their "content" type because they
				779	describe a region of memory, and all memory objects in LLVM are
				780	accessed through pointers.</p>
				781
Christopher Lamb	284d992	2007-12-11 09:31:00 +0000	[diff] [blame]	782	<p>A global variable may be declared to reside in a target-specifc numbered
				783	address space. For targets that support them, address spaces may affect how
				784	optimizations are performed and/or what target instructions are used to access
Christopher Lamb	d49e18d	2007-12-12 08:44:39 +0000	[diff] [blame]	785	the variable. The default address space is zero. The address space qualifier
				786	must precede any other attributes.</p>
Christopher Lamb	284d992	2007-12-11 09:31:00 +0000	[diff] [blame]	787
Chris Lattner	88f6c46	2005-11-12 00:45:07 +0000	[diff] [blame]	788	<p>LLVM allows an explicit section to be specified for globals. If the target
				789	supports it, it will emit globals to the section specified.</p>
				790
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	791	<p>An explicit alignment may be specified for a global. If not present, or if
				792	the alignment is set to zero, the alignment of the global is set by the target
				793	to whatever it feels convenient. If an explicit alignment is specified, the
				794	global is forced to have at least that much alignment. All alignments must be
				795	a power of 2.</p>
				796
Christopher Lamb	284d992	2007-12-11 09:31:00 +0000	[diff] [blame]	797	<p>For example, the following defines a global in a numbered address space with
				798	an initializer, section, and alignment:</p>
Chris Lattner	68027ea	2007-01-14 00:27:09 +0000	[diff] [blame]	799
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	800	<div class="doc_code">
Chris Lattner	68027ea	2007-01-14 00:27:09 +0000	[diff] [blame]	801	<pre>
Dan Gohman	398873c	2009-01-11 00:40:00 +0000	[diff] [blame]	802	@G = addrspace(5) constant float 1.0, section "foo", align 4
Chris Lattner	68027ea	2007-01-14 00:27:09 +0000	[diff] [blame]	803	</pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	804	</div>
Chris Lattner	68027ea	2007-01-14 00:27:09 +0000	[diff] [blame]	805
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	806	</div>
				807
				808
				809	<!-- ======================================================================= -->
				810	<div class="doc_subsection">
				811	<a name="functionstructure">Functions</a>
				812	</div>
				813
				814	<div class="doc_text">
				815
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	816	<p>LLVM function definitions consist of the "<tt>define</tt>" keyord,
				817	an optional <a href="#linkage">linkage type</a>, an optional
Anton Korobeynikov	8cea37b	2007-01-23 12:35:46 +0000	[diff] [blame]	818	<a href="#visibility">visibility style</a>, an optional
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	819	<a href="#callingconv">calling convention</a>, a return type, an optional
				820	<a href="#paramattrs">parameter attribute</a> for the return type, a function
				821	name, a (possibly empty) argument list (each with optional
Devang Patel	f642f47	2008-10-06 18:50:38 +0000	[diff] [blame]	822	<a href="#paramattrs">parameter attributes</a>), optional
				823	<a href="#fnattrs">function attributes</a>, an optional section,
				824	an optional alignment, an optional <a href="#gc">garbage collector name</a>,
Chris Lattner	0c46a7d	2008-10-04 18:10:21 +0000	[diff] [blame]	825	an opening curly brace, a list of basic blocks, and a closing curly brace.
Anton Korobeynikov	8cea37b	2007-01-23 12:35:46 +0000	[diff] [blame]	826
				827	LLVM function declarations consist of the "<tt>declare</tt>" keyword, an
				828	optional <a href="#linkage">linkage type</a>, an optional
				829	<a href="#visibility">visibility style</a>, an optional
				830	<a href="#callingconv">calling convention</a>, a return type, an optional
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	831	<a href="#paramattrs">parameter attribute</a> for the return type, a function
Gordon Henriksen	80a75bf	2007-12-10 03:18:06 +0000	[diff] [blame]	832	name, a possibly empty list of arguments, an optional alignment, and an optional
Gordon Henriksen	e754abe	2007-12-10 03:30:21 +0000	[diff] [blame]	833	<a href="#gc">garbage collector name</a>.</p>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	834
Chris Lattner	d3eda89	2008-08-05 18:29:16 +0000	[diff] [blame]	835	<p>A function definition contains a list of basic blocks, forming the CFG
				836	(Control Flow Graph) for
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	837	the function. Each basic block may optionally start with a label (giving the
				838	basic block a symbol table entry), contains a list of instructions, and ends
				839	with a <a href="#terminators">terminator</a> instruction (such as a branch or
				840	function return).</p>
				841
Chris Lattner	4a3c901	2007-06-08 16:52:14 +0000	[diff] [blame]	842	<p>The first basic block in a function is special in two ways: it is immediately
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	843	executed on entrance to the function, and it is not allowed to have predecessor
				844	basic blocks (i.e. there can not be any branches to the entry block of a
				845	function). Because the block can have no predecessors, it also cannot have any
				846	<a href="#i_phi">PHI nodes</a>.</p>
				847
Chris Lattner	88f6c46	2005-11-12 00:45:07 +0000	[diff] [blame]	848	<p>LLVM allows an explicit section to be specified for functions. If the target
				849	supports it, it will emit functions to the section specified.</p>
				850
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	851	<p>An explicit alignment may be specified for a function. If not present, or if
				852	the alignment is set to zero, the alignment of the function is set by the target
				853	to whatever it feels convenient. If an explicit alignment is specified, the
				854	function is forced to have at least that much alignment. All alignments must be
				855	a power of 2.</p>
				856
Devang Patel	307e8ab	2008-10-07 17:48:33 +0000	[diff] [blame]	857	<h5>Syntax:</h5>
				858
				859	<div class="doc_code">
Chris Lattner	50ad45c	2008-10-13 16:55:18 +0000	[diff] [blame]	860	<tt>
				861	define [<a href="#linkage">linkage</a>] [<a href="#visibility">visibility</a>]
				862	[<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>]
				863	<ResultType> @<FunctionName> ([argument list])
				864	[<a href="#fnattrs">fn Attrs</a>] [section "name"] [align N]
				865	[<a href="#gc">gc</a>] { ... }
				866	</tt>
Devang Patel	307e8ab	2008-10-07 17:48:33 +0000	[diff] [blame]	867	</div>
				868
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	869	</div>
				870
Anton Korobeynikov	8b0a8c8	2007-04-25 14:27:10 +0000	[diff] [blame]	871
				872	<!-- ======================================================================= -->
				873	<div class="doc_subsection">
				874	<a name="aliasstructure">Aliases</a>
				875	</div>
				876	<div class="doc_text">
				877	<p>Aliases act as "second name" for the aliasee value (which can be either
Anton Korobeynikov	726d45c	2008-03-22 08:36:14 +0000	[diff] [blame]	878	function, global variable, another alias or bitcast of global value). Aliases
				879	may have an optional <a href="#linkage">linkage type</a>, and an
Anton Korobeynikov	8b0a8c8	2007-04-25 14:27:10 +0000	[diff] [blame]	880	optional <a href="#visibility">visibility style</a>.</p>
				881
				882	<h5>Syntax:</h5>
				883
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	884	<div class="doc_code">
Bill Wendling	aac388b	2007-05-29 09:42:13 +0000	[diff] [blame]	885	<pre>
Duncan Sands	0b23ac1	2008-09-12 20:48:21 +0000	[diff] [blame]	886	@<Name> = alias [Linkage] [Visibility] <AliaseeTy> @<Aliasee>
Bill Wendling	aac388b	2007-05-29 09:42:13 +0000	[diff] [blame]	887	</pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	888	</div>
Anton Korobeynikov	8b0a8c8	2007-04-25 14:27:10 +0000	[diff] [blame]	889
				890	</div>
				891
				892
				893
Chris Lattner	4e9aba7	2006-01-23 23:23:47 +0000	[diff] [blame]	894	<!-- ======================================================================= -->
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	895	<div class="doc_subsection"><a name="paramattrs">Parameter Attributes</a></div>
				896	<div class="doc_text">
				897	<p>The return type and each parameter of a function type may have a set of
				898	<i>parameter attributes</i> associated with them. Parameter attributes are
				899	used to communicate additional information about the result or parameters of
Duncan Sands	dc02467	2007-11-27 13:23:08 +0000	[diff] [blame]	900	a function. Parameter attributes are considered to be part of the function,
				901	not of the function type, so functions with different parameter attributes
				902	can have the same function type.</p>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	903
Reid Spencer	950e9f8	2007-01-15 18:27:39 +0000	[diff] [blame]	904	<p>Parameter attributes are simple keywords that follow the type specified. If
				905	multiple parameter attributes are needed, they are space separated. For
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	906	example:</p>
				907
				908	<div class="doc_code">
				909	<pre>
Nick Lewycky	b6a7d25	2009-02-15 23:06:14 +0000	[diff] [blame]	910	declare i32 @printf(i8* noalias nocapture, ...)
Chris Lattner	66d922c	2008-10-04 18:33:34 +0000	[diff] [blame]	911	declare i32 @atoi(i8 zeroext)
				912	declare signext i8 @returns_signed_char()
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	913	</pre>
				914	</div>
				915
Duncan Sands	dc02467	2007-11-27 13:23:08 +0000	[diff] [blame]	916	<p>Note that any attributes for the function result (<tt>nounwind</tt>,
				917	<tt>readonly</tt>) come immediately after the argument list.</p>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	918
Anton Korobeynikov	7f70559	2007-01-12 19:20:47 +0000	[diff] [blame]	919	<p>Currently, only the following parameter attributes are defined:</p>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	920	<dl>
Reid Spencer	9445e9a	2007-07-19 23:13:04 +0000	[diff] [blame]	921	<dt><tt>zeroext</tt></dt>
Chris Lattner	66d922c	2008-10-04 18:33:34 +0000	[diff] [blame]	922	<dd>This indicates to the code generator that the parameter or return value
				923	should be zero-extended to a 32-bit value by the caller (for a parameter)
				924	or the callee (for a return value).</dd>
Chris Lattner	47507de	2008-01-11 06:20:47 +0000	[diff] [blame]	925
Reid Spencer	9445e9a	2007-07-19 23:13:04 +0000	[diff] [blame]	926	<dt><tt>signext</tt></dt>
Chris Lattner	66d922c	2008-10-04 18:33:34 +0000	[diff] [blame]	927	<dd>This indicates to the code generator that the parameter or return value
				928	should be sign-extended to a 32-bit value by the caller (for a parameter)
				929	or the callee (for a return value).</dd>
Chris Lattner	47507de	2008-01-11 06:20:47 +0000	[diff] [blame]	930
Anton Korobeynikov	34d6dea	2007-01-28 14:30:45 +0000	[diff] [blame]	931	<dt><tt>inreg</tt></dt>
Dale Johannesen	c9c6da6	2008-09-25 20:47:45 +0000	[diff] [blame]	932	<dd>This indicates that this parameter or return value should be treated
				933	in a special target-dependent fashion during while emitting code for a
				934	function call or return (usually, by putting it in a register as opposed
Chris Lattner	66d922c	2008-10-04 18:33:34 +0000	[diff] [blame]	935	to memory, though some targets use it to distinguish between two different
				936	kinds of registers). Use of this attribute is target-specific.</dd>
Chris Lattner	47507de	2008-01-11 06:20:47 +0000	[diff] [blame]	937
Duncan Sands	edb05df	2008-10-06 08:14:18 +0000	[diff] [blame]	938	<dt><tt><a name="byval">byval</a></tt></dt>
Chris Lattner	0747baa	2008-01-15 04:34:22 +0000	[diff] [blame]	939	<dd>This indicates that the pointer parameter should really be passed by
				940	value to the function. The attribute implies that a hidden copy of the
				941	pointee is made between the caller and the callee, so the callee is unable
Chris Lattner	ebec678	2008-08-05 18:21:08 +0000	[diff] [blame]	942	to modify the value in the callee. This attribute is only valid on LLVM
Chris Lattner	0747baa	2008-01-15 04:34:22 +0000	[diff] [blame]	943	pointer arguments. It is generally used to pass structs and arrays by
Duncan Sands	edb05df	2008-10-06 08:14:18 +0000	[diff] [blame]	944	value, but is also valid on pointers to scalars. The copy is considered to
				945	belong to the caller not the callee (for example,
				946	<tt><a href="#readonly">readonly</a></tt> functions should not write to
Devang Patel	f642f47	2008-10-06 18:50:38 +0000	[diff] [blame]	947	<tt>byval</tt> parameters). This is not a valid attribute for return
Chris Lattner	ce459b1	2009-02-05 05:42:28 +0000	[diff] [blame]	948	values. The byval attribute also supports specifying an alignment with the
				949	align attribute. This has a target-specific effect on the code generator
				950	that usually indicates a desired alignment for the synthesized stack
				951	slot.</dd>
Chris Lattner	47507de	2008-01-11 06:20:47 +0000	[diff] [blame]	952
Anton Korobeynikov	34d6dea	2007-01-28 14:30:45 +0000	[diff] [blame]	953	<dt><tt>sret</tt></dt>
Duncan Sands	e26dec6	2008-02-18 04:19:38 +0000	[diff] [blame]	954	<dd>This indicates that the pointer parameter specifies the address of a
				955	structure that is the return value of the function in the source program.
Chris Lattner	66d922c	2008-10-04 18:33:34 +0000	[diff] [blame]	956	This pointer must be guaranteed by the caller to be valid: loads and stores
				957	to the structure may be assumed by the callee to not to trap. This may only
Devang Patel	f642f47	2008-10-06 18:50:38 +0000	[diff] [blame]	958	be applied to the first parameter. This is not a valid attribute for
				959	return values. </dd>
Chris Lattner	47507de	2008-01-11 06:20:47 +0000	[diff] [blame]	960
Zhou Sheng	febca34	2007-06-05 05:28:26 +0000	[diff] [blame]	961	<dt><tt>noalias</tt></dt>
Nick Lewycky	02ff308	2008-11-24 03:41:24 +0000	[diff] [blame]	962	<dd>This indicates that the pointer does not alias any global or any other
				963	parameter. The caller is responsible for ensuring that this is the
Nick Lewycky	b2b32fd	2008-11-24 05:00:44 +0000	[diff] [blame]	964	case. On a function return value, <tt>noalias</tt> additionally indicates
				965	that the pointer does not alias any other pointers visible to the
Nick Lewycky	f23d0d3	2008-12-19 06:39:12 +0000	[diff] [blame]	966	caller. For further details, please see the discussion of the NoAlias
				967	response in
				968	<a href="http://llvm.org/docs/AliasAnalysis.html#MustMayNo">alias
				969	analysis</a>.</dd>
				970
				971	<dt><tt>nocapture</tt></dt>
				972	<dd>This indicates that the callee does not make any copies of the pointer
				973	that outlive the callee itself. This is not a valid attribute for return
				974	values.</dd>
Chris Lattner	47507de	2008-01-11 06:20:47 +0000	[diff] [blame]	975
Duncan Sands	50f19f5	2007-07-27 19:57:41 +0000	[diff] [blame]	976	<dt><tt>nest</tt></dt>
Duncan Sands	0789b8b	2008-07-08 09:27:25 +0000	[diff] [blame]	977	<dd>This indicates that the pointer parameter can be excised using the
Devang Patel	f642f47	2008-10-06 18:50:38 +0000	[diff] [blame]	978	<a href="#int_trampoline">trampoline intrinsics</a>. This is not a valid
				979	attribute for return values.</dd>
Anton Korobeynikov	7f70559	2007-01-12 19:20:47 +0000	[diff] [blame]	980	</dl>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	981
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	982	</div>
				983
				984	<!-- ======================================================================= -->
Chris Lattner	4e9aba7	2006-01-23 23:23:47 +0000	[diff] [blame]	985	<div class="doc_subsection">
Gordon Henriksen	80a75bf	2007-12-10 03:18:06 +0000	[diff] [blame]	986	<a name="gc">Garbage Collector Names</a>
				987	</div>
				988
				989	<div class="doc_text">
				990	<p>Each function may specify a garbage collector name, which is simply a
				991	string.</p>
				992
				993	<div class="doc_code"><pre
				994	>define void @f() gc "name" { ...</pre></div>
				995
				996	<p>The compiler declares the supported values of <i>name</i>. Specifying a
				997	collector which will cause the compiler to alter its output in order to support
				998	the named garbage collection algorithm.</p>
				999	</div>
				1000
				1001	<!-- ======================================================================= -->
				1002	<div class="doc_subsection">
Devang Patel	2c9c3e7	2008-09-26 23:51:19 +0000	[diff] [blame]	1003	<a name="fnattrs">Function Attributes</a>
Devang Patel	f8b9481	2008-09-04 23:05:13 +0000	[diff] [blame]	1004	</div>
				1005
				1006	<div class="doc_text">
Devang Patel	2c9c3e7	2008-09-26 23:51:19 +0000	[diff] [blame]	1007
				1008	<p>Function attributes are set to communicate additional information about
				1009	a function. Function attributes are considered to be part of the function,
				1010	not of the function type, so functions with different parameter attributes
				1011	can have the same function type.</p>
				1012
				1013	<p>Function attributes are simple keywords that follow the type specified. If
				1014	multiple attributes are needed, they are space separated. For
				1015	example:</p>
Devang Patel	f8b9481	2008-09-04 23:05:13 +0000	[diff] [blame]	1016
				1017	<div class="doc_code">
Bill Wendling	e36dccc	2008-09-07 10:26:33 +0000	[diff] [blame]	1018	<pre>
Devang Patel	2c9c3e7	2008-09-26 23:51:19 +0000	[diff] [blame]	1019	define void @f() noinline { ... }
				1020	define void @f() alwaysinline { ... }
				1021	define void @f() alwaysinline optsize { ... }
				1022	define void @f() optsize
Bill Wendling	e36dccc	2008-09-07 10:26:33 +0000	[diff] [blame]	1023	</pre>
Devang Patel	f8b9481	2008-09-04 23:05:13 +0000	[diff] [blame]	1024	</div>
				1025
Bill Wendling	e36dccc	2008-09-07 10:26:33 +0000	[diff] [blame]	1026	<dl>
Devang Patel	2c9c3e7	2008-09-26 23:51:19 +0000	[diff] [blame]	1027	<dt><tt>alwaysinline</tt></dt>
Chris Lattner	88d4b59	2008-10-04 18:23:17 +0000	[diff] [blame]	1028	<dd>This attribute indicates that the inliner should attempt to inline this
				1029	function into callers whenever possible, ignoring any active inlining size
				1030	threshold for this caller.</dd>
Bill Wendling	e36dccc	2008-09-07 10:26:33 +0000	[diff] [blame]	1031
Devang Patel	2c9c3e7	2008-09-26 23:51:19 +0000	[diff] [blame]	1032	<dt><tt>noinline</tt></dt>
Chris Lattner	88d4b59	2008-10-04 18:23:17 +0000	[diff] [blame]	1033	<dd>This attribute indicates that the inliner should never inline this function
Chris Lattner	94b5f7d	2008-10-05 17:14:59 +0000	[diff] [blame]	1034	in any situation. This attribute may not be used together with the
Chris Lattner	88d4b59	2008-10-04 18:23:17 +0000	[diff] [blame]	1035	<tt>alwaysinline</tt> attribute.</dd>
Bill Wendling	e36dccc	2008-09-07 10:26:33 +0000	[diff] [blame]	1036
Devang Patel	2c9c3e7	2008-09-26 23:51:19 +0000	[diff] [blame]	1037	<dt><tt>optsize</tt></dt>
Devang Patel	66c6c65	2008-09-29 18:34:44 +0000	[diff] [blame]	1038	<dd>This attribute suggests that optimization passes and code generator passes
Chris Lattner	88d4b59	2008-10-04 18:23:17 +0000	[diff] [blame]	1039	make choices that keep the code size of this function low, and otherwise do
				1040	optimizations specifically to reduce code size.</dd>
Bill Wendling	e36dccc	2008-09-07 10:26:33 +0000	[diff] [blame]	1041
Devang Patel	2c9c3e7	2008-09-26 23:51:19 +0000	[diff] [blame]	1042	<dt><tt>noreturn</tt></dt>
Chris Lattner	88d4b59	2008-10-04 18:23:17 +0000	[diff] [blame]	1043	<dd>This function attribute indicates that the function never returns normally.
				1044	This produces undefined behavior at runtime if the function ever does
				1045	dynamically return.</dd>
Devang Patel	2c9c3e7	2008-09-26 23:51:19 +0000	[diff] [blame]	1046
				1047	<dt><tt>nounwind</tt></dt>
Chris Lattner	88d4b59	2008-10-04 18:23:17 +0000	[diff] [blame]	1048	<dd>This function attribute indicates that the function never returns with an
				1049	unwind or exceptional control flow. If the function does unwind, its runtime
				1050	behavior is undefined.</dd>
				1051
				1052	<dt><tt>readnone</tt></dt>
Duncan Sands	edb05df	2008-10-06 08:14:18 +0000	[diff] [blame]	1053	<dd>This attribute indicates that the function computes its result (or the
				1054	exception it throws) based strictly on its arguments, without dereferencing any
				1055	pointer arguments or otherwise accessing any mutable state (e.g. memory, control
				1056	registers, etc) visible to caller functions. It does not write through any
				1057	pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments) and
				1058	never changes any state visible to callers.</dd>
Devang Patel	2c9c3e7	2008-09-26 23:51:19 +0000	[diff] [blame]	1059
Duncan Sands	edb05df	2008-10-06 08:14:18 +0000	[diff] [blame]	1060	<dt><tt><a name="readonly">readonly</a></tt></dt>
				1061	<dd>This attribute indicates that the function does not write through any
				1062	pointer arguments (including <tt><a href="#byval">byval</a></tt> arguments)
				1063	or otherwise modify any state (e.g. memory, control registers, etc) visible to
				1064	caller functions. It may dereference pointer arguments and read state that may
				1065	be set in the caller. A readonly function always returns the same value (or
				1066	throws the same exception) when called with the same set of arguments and global
				1067	state.</dd>
Bill Wendling	31359ba	2008-11-13 01:02:51 +0000	[diff] [blame]	1068
				1069	<dt><tt><a name="ssp">ssp</a></tt></dt>
Bill Wendling	baa39d8	2008-11-26 19:19:05 +0000	[diff] [blame]	1070	<dd>This attribute indicates that the function should emit a stack smashing
Bill Wendling	31359ba	2008-11-13 01:02:51 +0000	[diff] [blame]	1071	protector. It is in the form of a "canary"—a random value placed on the
				1072	stack before the local variables that's checked upon return from the function to
				1073	see if it has been overwritten. A heuristic is used to determine if a function
Bill Wendling	baa39d8	2008-11-26 19:19:05 +0000	[diff] [blame]	1074	needs stack protectors or not.
Bill Wendling	31359ba	2008-11-13 01:02:51 +0000	[diff] [blame]	1075
Bill Wendling	fbaa7ed	2008-11-26 19:07:40 +0000	[diff] [blame]	1076	<p>If a function that has an <tt>ssp</tt> attribute is inlined into a function
				1077	that doesn't have an <tt>ssp</tt> attribute, then the resulting function will
				1078	have an <tt>ssp</tt> attribute.</p></dd>
				1079
				1080	<dt><tt>sspreq</tt></dt>
Bill Wendling	baa39d8	2008-11-26 19:19:05 +0000	[diff] [blame]	1081	<dd>This attribute indicates that the function should <em>always</em> emit a
Bill Wendling	31359ba	2008-11-13 01:02:51 +0000	[diff] [blame]	1082	stack smashing protector. This overrides the <tt><a href="#ssp">ssp</a></tt>
Bill Wendling	baa39d8	2008-11-26 19:19:05 +0000	[diff] [blame]	1083	function attribute.
Bill Wendling	fbaa7ed	2008-11-26 19:07:40 +0000	[diff] [blame]	1084
				1085	<p>If a function that has an <tt>sspreq</tt> attribute is inlined into a
				1086	function that doesn't have an <tt>sspreq</tt> attribute or which has
				1087	an <tt>ssp</tt> attribute, then the resulting function will have
				1088	an <tt>sspreq</tt> attribute.</p></dd>
Bill Wendling	e36dccc	2008-09-07 10:26:33 +0000	[diff] [blame]	1089	</dl>
				1090
Devang Patel	f8b9481	2008-09-04 23:05:13 +0000	[diff] [blame]	1091	</div>
				1092
				1093	<!-- ======================================================================= -->
				1094	<div class="doc_subsection">
Chris Lattner	1eeeb0c	2006-04-08 04:40:53 +0000	[diff] [blame]	1095	<a name="moduleasm">Module-Level Inline Assembly</a>
Chris Lattner	4e9aba7	2006-01-23 23:23:47 +0000	[diff] [blame]	1096	</div>
				1097
				1098	<div class="doc_text">
				1099	<p>
				1100	Modules may contain "module-level inline asm" blocks, which corresponds to the
				1101	GCC "file scope inline asm" blocks. These blocks are internally concatenated by
				1102	LLVM and treated as a single unit, but may be separated in the .ll file if
				1103	desired. The syntax is very simple:
				1104	</p>
				1105
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	1106	<div class="doc_code">
				1107	<pre>
				1108	module asm "inline asm code goes here"
				1109	module asm "more can go here"
				1110	</pre>
				1111	</div>
Chris Lattner	4e9aba7	2006-01-23 23:23:47 +0000	[diff] [blame]	1112
				1113	<p>The strings can contain any character by escaping non-printable characters.
				1114	The escape sequence used is simply "\xx" where "xx" is the two digit hex code
				1115	for the number.
				1116	</p>
				1117
				1118	<p>
				1119	The inline asm code is simply printed to the machine code .s file when
				1120	assembly code is generated.
				1121	</p>
				1122	</div>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	1123
Reid Spencer	de15194	2007-02-19 23:54:10 +0000	[diff] [blame]	1124	<!-- ======================================================================= -->
				1125	<div class="doc_subsection">
				1126	<a name="datalayout">Data Layout</a>
				1127	</div>
				1128
				1129	<div class="doc_text">
				1130	<p>A module may specify a target specific data layout string that specifies how
Reid Spencer	c891084	2007-04-11 23:49:50 +0000	[diff] [blame]	1131	data is to be laid out in memory. The syntax for the data layout is simply:</p>
				1132	<pre> target datalayout = "<i>layout specification</i>"</pre>
				1133	<p>The <i>layout specification</i> consists of a list of specifications
				1134	separated by the minus sign character ('-'). Each specification starts with a
				1135	letter and may include other information after the letter to define some
				1136	aspect of the data layout. The specifications accepted are as follows: </p>
Reid Spencer	de15194	2007-02-19 23:54:10 +0000	[diff] [blame]	1137	<dl>
				1138	<dt><tt>E</tt></dt>
				1139	<dd>Specifies that the target lays out data in big-endian form. That is, the
				1140	bits with the most significance have the lowest address location.</dd>
				1141	<dt><tt>e</tt></dt>
Chris Lattner	d3eda89	2008-08-05 18:29:16 +0000	[diff] [blame]	1142	<dd>Specifies that the target lays out data in little-endian form. That is,
Reid Spencer	de15194	2007-02-19 23:54:10 +0000	[diff] [blame]	1143	the bits with the least significance have the lowest address location.</dd>
				1144	<dt><tt>p:<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
				1145	<dd>This specifies the <i>size</i> of a pointer and its <i>abi</i> and
				1146	<i>preferred</i> alignments. All sizes are in bits. Specifying the <i>pref</i>
				1147	alignment is optional. If omitted, the preceding <tt>:</tt> should be omitted
				1148	too.</dd>
				1149	<dt><tt>i<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
				1150	<dd>This specifies the alignment for an integer type of a given bit
				1151	<i>size</i>. The value of <i>size</i> must be in the range [1,2^23).</dd>
				1152	<dt><tt>v<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
				1153	<dd>This specifies the alignment for a vector type of a given bit
				1154	<i>size</i>.</dd>
				1155	<dt><tt>f<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
				1156	<dd>This specifies the alignment for a floating point type of a given bit
				1157	<i>size</i>. The value of <i>size</i> must be either 32 (float) or 64
				1158	(double).</dd>
				1159	<dt><tt>a<i>size</i>:<i>abi</i>:<i>pref</i></tt></dt>
				1160	<dd>This specifies the alignment for an aggregate type of a given bit
				1161	<i>size</i>.</dd>
				1162	</dl>
				1163	<p>When constructing the data layout for a given target, LLVM starts with a
				1164	default set of specifications which are then (possibly) overriden by the
				1165	specifications in the <tt>datalayout</tt> keyword. The default specifications
				1166	are given in this list:</p>
				1167	<ul>
				1168	<li><tt>E</tt> - big endian</li>
				1169	<li><tt>p:32:64:64</tt> - 32-bit pointers with 64-bit alignment</li>
				1170	<li><tt>i1:8:8</tt> - i1 is 8-bit (byte) aligned</li>
				1171	<li><tt>i8:8:8</tt> - i8 is 8-bit (byte) aligned</li>
				1172	<li><tt>i16:16:16</tt> - i16 is 16-bit aligned</li>
				1173	<li><tt>i32:32:32</tt> - i32 is 32-bit aligned</li>
Chris Lattner	d3eda89	2008-08-05 18:29:16 +0000	[diff] [blame]	1174	<li><tt>i64:32:64</tt> - i64 has ABI alignment of 32-bits but preferred
Reid Spencer	de15194	2007-02-19 23:54:10 +0000	[diff] [blame]	1175	alignment of 64-bits</li>
				1176	<li><tt>f32:32:32</tt> - float is 32-bit aligned</li>
				1177	<li><tt>f64:64:64</tt> - double is 64-bit aligned</li>
				1178	<li><tt>v64:64:64</tt> - 64-bit vector is 64-bit aligned</li>
				1179	<li><tt>v128:128:128</tt> - 128-bit vector is 128-bit aligned</li>
				1180	<li><tt>a0:0:1</tt> - aggregates are 8-bit aligned</li>
				1181	</ul>
Chris Lattner	ebec678	2008-08-05 18:21:08 +0000	[diff] [blame]	1182	<p>When LLVM is determining the alignment for a given type, it uses the
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	1183	following rules:</p>
Reid Spencer	de15194	2007-02-19 23:54:10 +0000	[diff] [blame]	1184	<ol>
				1185	<li>If the type sought is an exact match for one of the specifications, that
				1186	specification is used.</li>
				1187	<li>If no match is found, and the type sought is an integer type, then the
				1188	smallest integer type that is larger than the bitwidth of the sought type is
				1189	used. If none of the specifications are larger than the bitwidth then the the
				1190	largest integer type is used. For example, given the default specifications
				1191	above, the i7 type will use the alignment of i8 (next largest) while both
				1192	i65 and i256 will use the alignment of i64 (largest specified).</li>
				1193	<li>If no match is found, and the type sought is a vector type, then the
				1194	largest vector type that is smaller than the sought vector type will be used
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	1195	as a fall back. This happens because <128 x double> can be implemented
				1196	in terms of 64 <2 x double>, for example.</li>
Reid Spencer	de15194	2007-02-19 23:54:10 +0000	[diff] [blame]	1197	</ol>
				1198	</div>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	1199
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1200	<!-- *********************************************************************** -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1201	<div class="doc_section"> <a name="typesystem">Type System</a> </div>
				1202	<!-- *********************************************************************** -->
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	1203
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1204	<div class="doc_text">
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	1205
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1206	<p>The LLVM type system is one of the most important features of the
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1207	intermediate representation. Being typed enables a number of
Chris Lattner	d3eda89	2008-08-05 18:29:16 +0000	[diff] [blame]	1208	optimizations to be performed on the intermediate representation directly,
				1209	without having to do
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1210	extra analyses on the side before the transformation. A strong type
				1211	system makes it easier to read the generated code and enables novel
				1212	analyses and transformations that are not feasible to perform on normal
				1213	three address code representations.</p>
Chris Lattner	fa73021	2004-12-09 16:11:40 +0000	[diff] [blame]	1214
				1215	</div>
				1216
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1217	<!-- ======================================================================= -->
Chris Lattner	4f69f46	2008-01-04 04:32:38 +0000	[diff] [blame]	1218	<div class="doc_subsection"> <a name="t_classifications">Type
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1219	Classifications</a> </div>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1220	<div class="doc_text">
Chris Lattner	4f69f46	2008-01-04 04:32:38 +0000	[diff] [blame]	1221	<p>The types fall into a few useful
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1222	classifications:</p>
Misha Brukman	daa4cb0	2004-03-01 17:47:27 +0000	[diff] [blame]	1223
				1224	<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1225	<tbody>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1226	<tr><th>Classification</th><th>Types</th></tr>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1227	<tr>
Chris Lattner	4f69f46	2008-01-04 04:32:38 +0000	[diff] [blame]	1228	<td><a href="#t_integer">integer</a></td>
Reid Spencer	2b91631	2007-05-16 18:44:01 +0000	[diff] [blame]	1229	<td><tt>i1, i2, i3, ... i8, ... i16, ... i32, ... i64, ... </tt></td>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1230	</tr>
				1231	<tr>
Chris Lattner	4f69f46	2008-01-04 04:32:38 +0000	[diff] [blame]	1232	<td><a href="#t_floating">floating point</a></td>
				1233	<td><tt>float, double, x86_fp80, fp128, ppc_fp128</tt></td>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1234	</tr>
				1235	<tr>
				1236	<td><a name="t_firstclass">first class</a></td>
Chris Lattner	4f69f46	2008-01-04 04:32:38 +0000	[diff] [blame]	1237	<td><a href="#t_integer">integer</a>,
				1238	<a href="#t_floating">floating point</a>,
				1239	<a href="#t_pointer">pointer</a>,
Dan Gohman	0066db6	2008-06-18 18:42:13 +0000	[diff] [blame]	1240	<a href="#t_vector">vector</a>,
Dan Gohman	a334d5f	2008-05-12 23:51:09 +0000	[diff] [blame]	1241	<a href="#t_struct">structure</a>,
				1242	<a href="#t_array">array</a>,
Dan Gohman	ade5faa	2008-05-23 22:50:26 +0000	[diff] [blame]	1243	<a href="#t_label">label</a>.
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	1244	</td>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1245	</tr>
Chris Lattner	4f69f46	2008-01-04 04:32:38 +0000	[diff] [blame]	1246	<tr>
				1247	<td><a href="#t_primitive">primitive</a></td>
				1248	<td><a href="#t_label">label</a>,
				1249	<a href="#t_void">void</a>,
Chris Lattner	4f69f46	2008-01-04 04:32:38 +0000	[diff] [blame]	1250	<a href="#t_floating">floating point</a>.</td>
				1251	</tr>
				1252	<tr>
				1253	<td><a href="#t_derived">derived</a></td>
				1254	<td><a href="#t_integer">integer</a>,
				1255	<a href="#t_array">array</a>,
				1256	<a href="#t_function">function</a>,
				1257	<a href="#t_pointer">pointer</a>,
				1258	<a href="#t_struct">structure</a>,
				1259	<a href="#t_pstruct">packed structure</a>,
				1260	<a href="#t_vector">vector</a>,
				1261	<a href="#t_opaque">opaque</a>.
Dan Gohman	01ac101	2008-10-14 16:32:04 +0000	[diff] [blame]	1262	</td>
Chris Lattner	4f69f46	2008-01-04 04:32:38 +0000	[diff] [blame]	1263	</tr>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1264	</tbody>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1265	</table>
Misha Brukman	daa4cb0	2004-03-01 17:47:27 +0000	[diff] [blame]	1266
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1267	<p>The <a href="#t_firstclass">first class</a> types are perhaps the
				1268	most important. Values of these types are the only ones which can be
				1269	produced by instructions, passed as arguments, or used as operands to
Dan Gohman	c4b49eb	2008-05-23 21:53:15 +0000	[diff] [blame]	1270	instructions.</p>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1271	</div>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1272
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1273	<!-- ======================================================================= -->
Chris Lattner	4f69f46	2008-01-04 04:32:38 +0000	[diff] [blame]	1274	<div class="doc_subsection"> <a name="t_primitive">Primitive Types</a> </div>
Chris Lattner	8f8c7b7	2008-01-04 04:34:14 +0000	[diff] [blame]	1275
Chris Lattner	4f69f46	2008-01-04 04:32:38 +0000	[diff] [blame]	1276	<div class="doc_text">
				1277	<p>The primitive types are the fundamental building blocks of the LLVM
				1278	system.</p>
				1279
Chris Lattner	8f8c7b7	2008-01-04 04:34:14 +0000	[diff] [blame]	1280	</div>
				1281
Chris Lattner	4f69f46	2008-01-04 04:32:38 +0000	[diff] [blame]	1282	<!-- _______________________________________________________________________ -->
				1283	<div class="doc_subsubsection"> <a name="t_floating">Floating Point Types</a> </div>
				1284
				1285	<div class="doc_text">
				1286	<table>
				1287	<tbody>
				1288	<tr><th>Type</th><th>Description</th></tr>
				1289	<tr><td><tt>float</tt></td><td>32-bit floating point value</td></tr>
				1290	<tr><td><tt>double</tt></td><td>64-bit floating point value</td></tr>
				1291	<tr><td><tt>fp128</tt></td><td>128-bit floating point value (112-bit mantissa)</td></tr>
				1292	<tr><td><tt>x86_fp80</tt></td><td>80-bit floating point value (X87)</td></tr>
				1293	<tr><td><tt>ppc_fp128</tt></td><td>128-bit floating point value (two 64-bits)</td></tr>
				1294	</tbody>
				1295	</table>
				1296	</div>
				1297
				1298	<!-- _______________________________________________________________________ -->
				1299	<div class="doc_subsubsection"> <a name="t_void">Void Type</a> </div>
				1300
				1301	<div class="doc_text">
				1302	<h5>Overview:</h5>
				1303	<p>The void type does not represent any value and has no size.</p>
				1304
				1305	<h5>Syntax:</h5>
				1306
				1307	<pre>
				1308	void
				1309	</pre>
				1310	</div>
				1311
				1312	<!-- _______________________________________________________________________ -->
				1313	<div class="doc_subsubsection"> <a name="t_label">Label Type</a> </div>
				1314
				1315	<div class="doc_text">
				1316	<h5>Overview:</h5>
				1317	<p>The label type represents code labels.</p>
				1318
				1319	<h5>Syntax:</h5>
				1320
				1321	<pre>
				1322	label
				1323	</pre>
				1324	</div>
				1325
				1326
				1327	<!-- ======================================================================= -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1328	<div class="doc_subsection"> <a name="t_derived">Derived Types</a> </div>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1329
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1330	<div class="doc_text">
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1331
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1332	<p>The real power in LLVM comes from the derived types in the system.
				1333	This is what allows a programmer to represent arrays, functions,
				1334	pointers, and other useful types. Note that these derived types may be
				1335	recursive: For example, it is possible to have a two dimensional array.</p>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1336
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1337	</div>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1338
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1339	<!-- _______________________________________________________________________ -->
Reid Spencer	2b91631	2007-05-16 18:44:01 +0000	[diff] [blame]	1340	<div class="doc_subsubsection"> <a name="t_integer">Integer Type</a> </div>
				1341
				1342	<div class="doc_text">
				1343
				1344	<h5>Overview:</h5>
				1345	<p>The integer type is a very simple derived type that simply specifies an
				1346	arbitrary bit width for the integer type desired. Any bit width from 1 bit to
				1347	2^23-1 (about 8 million) can be specified.</p>
				1348
				1349	<h5>Syntax:</h5>
				1350
				1351	<pre>
				1352	iN
				1353	</pre>
				1354
				1355	<p>The number of bits the integer will occupy is specified by the <tt>N</tt>
				1356	value.</p>
				1357
				1358	<h5>Examples:</h5>
				1359	<table class="layout">
Chris Lattner	b9488a6	2007-12-18 06:18:21 +0000	[diff] [blame]	1360	<tbody>
				1361	<tr>
				1362	<td><tt>i1</tt></td>
				1363	<td>a single-bit integer.</td>
				1364	</tr><tr>
				1365	<td><tt>i32</tt></td>
				1366	<td>a 32-bit integer.</td>
				1367	</tr><tr>
				1368	<td><tt>i1942652</tt></td>
				1369	<td>a really big integer of over 1 million bits.</td>
Reid Spencer	2b91631	2007-05-16 18:44:01 +0000	[diff] [blame]	1370	</tr>
Chris Lattner	b9488a6	2007-12-18 06:18:21 +0000	[diff] [blame]	1371	</tbody>
Reid Spencer	2b91631	2007-05-16 18:44:01 +0000	[diff] [blame]	1372	</table>
Dan Gohman	d8791e5	2009-01-24 15:58:40 +0000	[diff] [blame]	1373
				1374	<p>Note that the code generator does not yet support large integer types
				1375	to be used as function return types. The specific limit on how large a
				1376	return type the code generator can currently handle is target-dependent;
				1377	currently it's often 64 bits for 32-bit targets and 128 bits for 64-bit
				1378	targets.</p>
				1379
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	1380	</div>
Reid Spencer	2b91631	2007-05-16 18:44:01 +0000	[diff] [blame]	1381
				1382	<!-- _______________________________________________________________________ -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1383	<div class="doc_subsubsection"> <a name="t_array">Array Type</a> </div>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1384
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1385	<div class="doc_text">
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1386
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1387	<h5>Overview:</h5>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1388
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1389	<p>The array type is a very simple derived type that arranges elements
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1390	sequentially in memory. The array type requires a size (number of
				1391	elements) and an underlying data type.</p>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1392
Chris Lattner	7faa883	2002-04-14 06:13:44 +0000	[diff] [blame]	1393	<h5>Syntax:</h5>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1394
				1395	<pre>
				1396	[<# elements> x <elementtype>]
				1397	</pre>
				1398
John Criswell	e4c57cc	2005-05-12 16:52:32 +0000	[diff] [blame]	1399	<p>The number of elements is a constant integer value; elementtype may
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1400	be any type with a size.</p>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1401
Chris Lattner	7faa883	2002-04-14 06:13:44 +0000	[diff] [blame]	1402	<h5>Examples:</h5>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1403	<table class="layout">
				1404	<tr class="layout">
Chris Lattner	23ff1f9	2007-12-19 05:04:11 +0000	[diff] [blame]	1405	<td class="left"><tt>[40 x i32]</tt></td>
				1406	<td class="left">Array of 40 32-bit integer values.</td>
				1407	</tr>
				1408	<tr class="layout">
				1409	<td class="left"><tt>[41 x i32]</tt></td>
				1410	<td class="left">Array of 41 32-bit integer values.</td>
				1411	</tr>
				1412	<tr class="layout">
				1413	<td class="left"><tt>[4 x i8]</tt></td>
				1414	<td class="left">Array of 4 8-bit integer values.</td>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1415	</tr>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1416	</table>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1417	<p>Here are some examples of multidimensional arrays:</p>
				1418	<table class="layout">
				1419	<tr class="layout">
Chris Lattner	23ff1f9	2007-12-19 05:04:11 +0000	[diff] [blame]	1420	<td class="left"><tt>[3 x [4 x i32]]</tt></td>
				1421	<td class="left">3x4 array of 32-bit integer values.</td>
				1422	</tr>
				1423	<tr class="layout">
				1424	<td class="left"><tt>[12 x [10 x float]]</tt></td>
				1425	<td class="left">12x10 array of single precision floating point values.</td>
				1426	</tr>
				1427	<tr class="layout">
				1428	<td class="left"><tt>[2 x [3 x [4 x i16]]]</tt></td>
				1429	<td class="left">2x3x4 array of 16-bit integer values.</td>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1430	</tr>
				1431	</table>
Chris Lattner	e67a951	2005-06-24 17:22:57 +0000	[diff] [blame]	1432
John Criswell	0ec250c	2005-10-24 16:17:18 +0000	[diff] [blame]	1433	<p>Note that 'variable sized arrays' can be implemented in LLVM with a zero
				1434	length array. Normally, accesses past the end of an array are undefined in
Chris Lattner	e67a951	2005-06-24 17:22:57 +0000	[diff] [blame]	1435	LLVM (e.g. it is illegal to access the 5th element of a 3 element array).
				1436	As a special case, however, zero length arrays are recognized to be variable
				1437	length. This allows implementation of 'pascal style arrays' with the LLVM
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	1438	type "{ i32, [0 x float]}", for example.</p>
Chris Lattner	e67a951	2005-06-24 17:22:57 +0000	[diff] [blame]	1439
Dan Gohman	d8791e5	2009-01-24 15:58:40 +0000	[diff] [blame]	1440	<p>Note that the code generator does not yet support large aggregate types
				1441	to be used as function return types. The specific limit on how large an
				1442	aggregate return type the code generator can currently handle is
				1443	target-dependent, and also dependent on the aggregate element types.</p>
				1444
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1445	</div>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1446
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1447	<!-- _______________________________________________________________________ -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1448	<div class="doc_subsubsection"> <a name="t_function">Function Type</a> </div>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1449	<div class="doc_text">
Chris Lattner	f4cde4e	2008-04-23 04:59:35 +0000	[diff] [blame]	1450
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1451	<h5>Overview:</h5>
Chris Lattner	f4cde4e	2008-04-23 04:59:35 +0000	[diff] [blame]	1452
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1453	<p>The function type can be thought of as a function signature. It
Devang Patel	a582f40	2008-03-24 05:35:41 +0000	[diff] [blame]	1454	consists of a return type and a list of formal parameter types. The
Chris Lattner	f4cde4e	2008-04-23 04:59:35 +0000	[diff] [blame]	1455	return type of a function type is a scalar type, a void type, or a struct type.
Devang Patel	7a3ad1a	2008-03-24 20:52:42 +0000	[diff] [blame]	1456	If the return type is a struct type then all struct elements must be of first
Chris Lattner	f4cde4e	2008-04-23 04:59:35 +0000	[diff] [blame]	1457	class types, and the struct must have at least one element.</p>
Devang Patel	c3fc6df	2008-03-10 20:49:15 +0000	[diff] [blame]	1458
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1459	<h5>Syntax:</h5>
Chris Lattner	f4cde4e	2008-04-23 04:59:35 +0000	[diff] [blame]	1460
				1461	<pre>
				1462	<returntype list> (<parameter list>)
				1463	</pre>
				1464
John Criswell	0ec250c	2005-10-24 16:17:18 +0000	[diff] [blame]	1465	<p>...where '<tt><parameter list></tt>' is a comma-separated list of type
Misha Brukman	c24b758	2004-08-12 20:16:08 +0000	[diff] [blame]	1466	specifiers. Optionally, the parameter list may include a type <tt>...</tt>,
Chris Lattner	27f71f2	2003-09-03 00:41:47 +0000	[diff] [blame]	1467	which indicates that the function takes a variable number of arguments.
				1468	Variable argument functions can access their arguments with the <a
Devang Patel	c3fc6df	2008-03-10 20:49:15 +0000	[diff] [blame]	1469	href="#int_varargs">variable argument handling intrinsic</a> functions.
				1470	'<tt><returntype list></tt>' is a comma-separated list of
				1471	<a href="#t_firstclass">first class</a> type specifiers.</p>
Chris Lattner	f4cde4e	2008-04-23 04:59:35 +0000	[diff] [blame]	1472
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1473	<h5>Examples:</h5>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1474	<table class="layout">
				1475	<tr class="layout">
Reid Spencer	92f8230	2006-12-31 07:18:34 +0000	[diff] [blame]	1476	<td class="left"><tt>i32 (i32)</tt></td>
				1477	<td class="left">function taking an <tt>i32</tt>, returning an <tt>i32</tt>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1478	</td>
Reid Spencer	92f8230	2006-12-31 07:18:34 +0000	[diff] [blame]	1479	</tr><tr class="layout">
Reid Spencer	9445e9a	2007-07-19 23:13:04 +0000	[diff] [blame]	1480	<td class="left"><tt>float (i16 signext, i32 )
Reid Spencer	f17a0b7	2006-12-31 07:20:23 +0000	[diff] [blame]	1481	</tt></td>
Reid Spencer	92f8230	2006-12-31 07:18:34 +0000	[diff] [blame]	1482	<td class="left"><a href="#t_pointer">Pointer</a> to a function that takes
				1483	an <tt>i16</tt> that should be sign extended and a
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	1484	<a href="#t_pointer">pointer</a> to <tt>i32</tt>, returning
Reid Spencer	92f8230	2006-12-31 07:18:34 +0000	[diff] [blame]	1485	<tt>float</tt>.
				1486	</td>
				1487	</tr><tr class="layout">
				1488	<td class="left"><tt>i32 (i8*, ...)</tt></td>
				1489	<td class="left">A vararg function that takes at least one
Reid Spencer	a517338	2007-01-04 16:43:23 +0000	[diff] [blame]	1490	<a href="#t_pointer">pointer</a> to <tt>i8 </tt> (char in C),
Reid Spencer	92f8230	2006-12-31 07:18:34 +0000	[diff] [blame]	1491	which returns an integer. This is the signature for <tt>printf</tt> in
				1492	LLVM.
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1493	</td>
Devang Patel	a582f40	2008-03-24 05:35:41 +0000	[diff] [blame]	1494	</tr><tr class="layout">
				1495	<td class="left"><tt>{i32, i32} (i32)</tt></td>
Misha Brukman	b0a57aa	2008-11-27 06:41:20 +0000	[diff] [blame]	1496	<td class="left">A function taking an <tt>i32</tt>, returning two
				1497	<tt>i32</tt> values as an aggregate of type <tt>{ i32, i32 }</tt>
Devang Patel	a582f40	2008-03-24 05:35:41 +0000	[diff] [blame]	1498	</td>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1499	</tr>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1500	</table>
Misha Brukman	daa4cb0	2004-03-01 17:47:27 +0000	[diff] [blame]	1501
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1502	</div>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1503	<!-- _______________________________________________________________________ -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1504	<div class="doc_subsubsection"> <a name="t_struct">Structure Type</a> </div>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1505	<div class="doc_text">
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1506	<h5>Overview:</h5>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1507	<p>The structure type is used to represent a collection of data members
				1508	together in memory. The packing of the field types is defined to match
				1509	the ABI of the underlying processor. The elements of a structure may
				1510	be any type that has a size.</p>
				1511	<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
				1512	and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
				1513	field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
				1514	instruction.</p>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1515	<h5>Syntax:</h5>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1516	<pre> { <type list> }<br></pre>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1517	<h5>Examples:</h5>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1518	<table class="layout">
				1519	<tr class="layout">
Jeff Cohen	6f1cc77	2007-04-22 01:17:39 +0000	[diff] [blame]	1520	<td class="left"><tt>{ i32, i32, i32 }</tt></td>
				1521	<td class="left">A triple of three <tt>i32</tt> values</td>
				1522	</tr><tr class="layout">
				1523	<td class="left"><tt>{ float, i32 (i32) * }</tt></td>
				1524	<td class="left">A pair, where the first element is a <tt>float</tt> and the
				1525	second element is a <a href="#t_pointer">pointer</a> to a
				1526	<a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
				1527	an <tt>i32</tt>.</td>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1528	</tr>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1529	</table>
Dan Gohman	d8791e5	2009-01-24 15:58:40 +0000	[diff] [blame]	1530
				1531	<p>Note that the code generator does not yet support large aggregate types
				1532	to be used as function return types. The specific limit on how large an
				1533	aggregate return type the code generator can currently handle is
				1534	target-dependent, and also dependent on the aggregate element types.</p>
				1535
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1536	</div>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1537
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	1538	<!-- _______________________________________________________________________ -->
Andrew Lenharth	75e1068	2006-12-08 17:13:00 +0000	[diff] [blame]	1539	<div class="doc_subsubsection"> <a name="t_pstruct">Packed Structure Type</a>
				1540	</div>
				1541	<div class="doc_text">
				1542	<h5>Overview:</h5>
				1543	<p>The packed structure type is used to represent a collection of data members
				1544	together in memory. There is no padding between fields. Further, the alignment
				1545	of a packed structure is 1 byte. The elements of a packed structure may
				1546	be any type that has a size.</p>
				1547	<p>Structures are accessed using '<tt><a href="#i_load">load</a></tt>
				1548	and '<tt><a href="#i_store">store</a></tt>' by getting a pointer to a
				1549	field with the '<tt><a href="#i_getelementptr">getelementptr</a></tt>'
				1550	instruction.</p>
				1551	<h5>Syntax:</h5>
				1552	<pre> < { <type list> } > <br></pre>
				1553	<h5>Examples:</h5>
				1554	<table class="layout">
				1555	<tr class="layout">
Jeff Cohen	6f1cc77	2007-04-22 01:17:39 +0000	[diff] [blame]	1556	<td class="left"><tt>< { i32, i32, i32 } ></tt></td>
				1557	<td class="left">A triple of three <tt>i32</tt> values</td>
				1558	</tr><tr class="layout">
Bill Wendling	e36dccc	2008-09-07 10:26:33 +0000	[diff] [blame]	1559	<td class="left">
				1560	<tt>< { float, i32 (i32)* } ></tt></td>
Jeff Cohen	6f1cc77	2007-04-22 01:17:39 +0000	[diff] [blame]	1561	<td class="left">A pair, where the first element is a <tt>float</tt> and the
				1562	second element is a <a href="#t_pointer">pointer</a> to a
				1563	<a href="#t_function">function</a> that takes an <tt>i32</tt>, returning
				1564	an <tt>i32</tt>.</td>
Andrew Lenharth	75e1068	2006-12-08 17:13:00 +0000	[diff] [blame]	1565	</tr>
				1566	</table>
				1567	</div>
				1568
				1569	<!-- _______________________________________________________________________ -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1570	<div class="doc_subsubsection"> <a name="t_pointer">Pointer Type</a> </div>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1571	<div class="doc_text">
Chris Lattner	7faa883	2002-04-14 06:13:44 +0000	[diff] [blame]	1572	<h5>Overview:</h5>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1573	<p>As in many languages, the pointer type represents a pointer or
Christopher Lamb	284d992	2007-12-11 09:31:00 +0000	[diff] [blame]	1574	reference to another object, which must live in memory. Pointer types may have
				1575	an optional address space attribute defining the target-specific numbered
				1576	address space where the pointed-to object resides. The default address space is
				1577	zero.</p>
Chris Lattner	0fd4a27	2009-02-08 19:53:29 +0000	[diff] [blame]	1578
				1579	<p>Note that LLVM does not permit pointers to void (<tt>void*</tt>) nor does
Chris Lattner	e220e8c	2009-02-08 22:21:28 +0000	[diff] [blame]	1580	it permit pointers to labels (<tt>label</tt>). Use <tt>i8</tt> instead.</p>
Chris Lattner	0fd4a27	2009-02-08 19:53:29 +0000	[diff] [blame]	1581
Chris Lattner	7faa883	2002-04-14 06:13:44 +0000	[diff] [blame]	1582	<h5>Syntax:</h5>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	1583	<pre> <type> *<br></pre>
Chris Lattner	7faa883	2002-04-14 06:13:44 +0000	[diff] [blame]	1584	<h5>Examples:</h5>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1585	<table class="layout">
				1586	<tr class="layout">
Dan Gohman	2a08c53	2009-01-04 23:44:43 +0000	[diff] [blame]	1587	<td class="left"><tt>[4 x i32]*</tt></td>
Chris Lattner	23ff1f9	2007-12-19 05:04:11 +0000	[diff] [blame]	1588	<td class="left">A <a href="#t_pointer">pointer</a> to <a
				1589	href="#t_array">array</a> of four <tt>i32</tt> values.</td>
				1590	</tr>
				1591	<tr class="layout">
				1592	<td class="left"><tt>i32 (i32 ) </tt></td>
				1593	<td class="left"> A <a href="#t_pointer">pointer</a> to a <a
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	1594	href="#t_function">function</a> that takes an <tt>i32*</tt>, returning an
Chris Lattner	23ff1f9	2007-12-19 05:04:11 +0000	[diff] [blame]	1595	<tt>i32</tt>.</td>
				1596	</tr>
				1597	<tr class="layout">
				1598	<td class="left"><tt>i32 addrspace(5)*</tt></td>
				1599	<td class="left">A <a href="#t_pointer">pointer</a> to an <tt>i32</tt> value
				1600	that resides in address space #5.</td>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1601	</tr>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1602	</table>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1603	</div>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1604
Chris Lattner	a58561b	2004-08-12 19:12:28 +0000	[diff] [blame]	1605	<!-- _______________________________________________________________________ -->
Reid Spencer	485bad1	2007-02-15 03:07:05 +0000	[diff] [blame]	1606	<div class="doc_subsubsection"> <a name="t_vector">Vector Type</a> </div>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1607	<div class="doc_text">
Chris Lattner	69c11bb	2005-04-25 17:34:15 +0000	[diff] [blame]	1608
Chris Lattner	a58561b	2004-08-12 19:12:28 +0000	[diff] [blame]	1609	<h5>Overview:</h5>
Chris Lattner	69c11bb	2005-04-25 17:34:15 +0000	[diff] [blame]	1610
Reid Spencer	485bad1	2007-02-15 03:07:05 +0000	[diff] [blame]	1611	<p>A vector type is a simple derived type that represents a vector
				1612	of elements. Vector types are used when multiple primitive data
Chris Lattner	a58561b	2004-08-12 19:12:28 +0000	[diff] [blame]	1613	are operated in parallel using a single instruction (SIMD).
Reid Spencer	485bad1	2007-02-15 03:07:05 +0000	[diff] [blame]	1614	A vector type requires a size (number of
Chris Lattner	b8d172f	2005-11-10 01:44:22 +0000	[diff] [blame]	1615	elements) and an underlying primitive data type. Vectors must have a power
Reid Spencer	485bad1	2007-02-15 03:07:05 +0000	[diff] [blame]	1616	of two length (1, 2, 4, 8, 16 ...). Vector types are
Chris Lattner	a58561b	2004-08-12 19:12:28 +0000	[diff] [blame]	1617	considered <a href="#t_firstclass">first class</a>.</p>
Chris Lattner	69c11bb	2005-04-25 17:34:15 +0000	[diff] [blame]	1618
Chris Lattner	a58561b	2004-08-12 19:12:28 +0000	[diff] [blame]	1619	<h5>Syntax:</h5>
Chris Lattner	69c11bb	2005-04-25 17:34:15 +0000	[diff] [blame]	1620
				1621	<pre>
				1622	< <# elements> x <elementtype> >
				1623	</pre>
				1624
John Criswell	c1f786c	2005-05-13 22:25:59 +0000	[diff] [blame]	1625	<p>The number of elements is a constant integer value; elementtype may
Chris Lattner	3b19d65	2007-01-15 01:54:13 +0000	[diff] [blame]	1626	be any integer or floating point type.</p>
Chris Lattner	69c11bb	2005-04-25 17:34:15 +0000	[diff] [blame]	1627
Chris Lattner	a58561b	2004-08-12 19:12:28 +0000	[diff] [blame]	1628	<h5>Examples:</h5>
Chris Lattner	69c11bb	2005-04-25 17:34:15 +0000	[diff] [blame]	1629
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1630	<table class="layout">
				1631	<tr class="layout">
Chris Lattner	23ff1f9	2007-12-19 05:04:11 +0000	[diff] [blame]	1632	<td class="left"><tt><4 x i32></tt></td>
				1633	<td class="left">Vector of 4 32-bit integer values.</td>
				1634	</tr>
				1635	<tr class="layout">
				1636	<td class="left"><tt><8 x float></tt></td>
				1637	<td class="left">Vector of 8 32-bit floating-point values.</td>
				1638	</tr>
				1639	<tr class="layout">
				1640	<td class="left"><tt><2 x i64></tt></td>
				1641	<td class="left">Vector of 2 64-bit integer values.</td>
Reid Spencer	d3f876c	2004-11-01 08:19:36 +0000	[diff] [blame]	1642	</tr>
				1643	</table>
Dan Gohman	d8791e5	2009-01-24 15:58:40 +0000	[diff] [blame]	1644
				1645	<p>Note that the code generator does not yet support large vector types
				1646	to be used as function return types. The specific limit on how large a
				1647	vector return type codegen can currently handle is target-dependent;
				1648	currently it's often a few times longer than a hardware vector register.</p>
				1649
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	1650	</div>
				1651
Chris Lattner	69c11bb	2005-04-25 17:34:15 +0000	[diff] [blame]	1652	<!-- _______________________________________________________________________ -->
				1653	<div class="doc_subsubsection"> <a name="t_opaque">Opaque Type</a> </div>
				1654	<div class="doc_text">
				1655
				1656	<h5>Overview:</h5>
				1657
				1658	<p>Opaque types are used to represent unknown types in the system. This
Gordon Henriksen	8ac04ff	2007-10-14 00:34:53 +0000	[diff] [blame]	1659	corresponds (for example) to the C notion of a forward declared structure type.
Chris Lattner	69c11bb	2005-04-25 17:34:15 +0000	[diff] [blame]	1660	In LLVM, opaque types can eventually be resolved to any type (not just a
				1661	structure type).</p>
				1662
				1663	<h5>Syntax:</h5>
				1664
				1665	<pre>
				1666	opaque
				1667	</pre>
				1668
				1669	<h5>Examples:</h5>
				1670
				1671	<table class="layout">
				1672	<tr class="layout">
Chris Lattner	23ff1f9	2007-12-19 05:04:11 +0000	[diff] [blame]	1673	<td class="left"><tt>opaque</tt></td>
				1674	<td class="left">An opaque type.</td>
Chris Lattner	69c11bb	2005-04-25 17:34:15 +0000	[diff] [blame]	1675	</tr>
				1676	</table>
				1677	</div>
				1678
Chris Lattner	242d61d	2009-02-02 07:32:36 +0000	[diff] [blame]	1679	<!-- ======================================================================= -->
				1680	<div class="doc_subsection">
				1681	<a name="t_uprefs">Type Up-references</a>
				1682	</div>
				1683
				1684	<div class="doc_text">
				1685	<h5>Overview:</h5>
				1686	<p>
				1687	An "up reference" allows you to refer to a lexically enclosing type without
				1688	requiring it to have a name. For instance, a structure declaration may contain a
				1689	pointer to any of the types it is lexically a member of. Example of up
				1690	references (with their equivalent as named type declarations) include:</p>
				1691
				1692	<pre>
Chris Lattner	3060f5b	2009-02-09 10:00:56 +0000	[diff] [blame]	1693	{ \2 * } %x = type { %x* }
Chris Lattner	242d61d	2009-02-02 07:32:36 +0000	[diff] [blame]	1694	{ \2 }* %y = type { %y }*
				1695	\1* %z = type %z*
				1696	</pre>
				1697
				1698	<p>
				1699	An up reference is needed by the asmprinter for printing out cyclic types when
				1700	there is no declared name for a type in the cycle. Because the asmprinter does
				1701	not want to print out an infinite type string, it needs a syntax to handle
				1702	recursive types that have no names (all names are optional in llvm IR).
				1703	</p>
				1704
				1705	<h5>Syntax:</h5>
				1706	<pre>
				1707	\<level>
				1708	</pre>
				1709
				1710	<p>
				1711	The level is the count of the lexical type that is being referred to.
				1712	</p>
				1713
				1714	<h5>Examples:</h5>
				1715
				1716	<table class="layout">
				1717	<tr class="layout">
				1718	<td class="left"><tt>\1*</tt></td>
				1719	<td class="left">Self-referential pointer.</td>
				1720	</tr>
				1721	<tr class="layout">
				1722	<td class="left"><tt>{ { \3*, i8 }, i32 }</tt></td>
				1723	<td class="left">Recursive structure where the upref refers to the out-most
				1724	structure.</td>
				1725	</tr>
				1726	</table>
				1727	</div>
				1728
Chris Lattner	69c11bb	2005-04-25 17:34:15 +0000	[diff] [blame]	1729
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1730	<!-- *********************************************************************** -->
				1731	<div class="doc_section"> <a name="constants">Constants</a> </div>
				1732	<!-- *********************************************************************** -->
				1733
				1734	<div class="doc_text">
				1735
				1736	<p>LLVM has several different basic types of constants. This section describes
				1737	them all and their syntax.</p>
				1738
				1739	</div>
				1740
				1741	<!-- ======================================================================= -->
Reid Spencer	cc16dc3	2004-12-09 18:02:53 +0000	[diff] [blame]	1742	<div class="doc_subsection"><a name="simpleconstants">Simple Constants</a></div>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1743
				1744	<div class="doc_text">
				1745
				1746	<dl>
				1747	<dt><b>Boolean constants</b></dt>
				1748
				1749	<dd>The two strings '<tt>true</tt>' and '<tt>false</tt>' are both valid
Reid Spencer	c78f337	2007-01-12 03:35:51 +0000	[diff] [blame]	1750	constants of the <tt><a href="#t_primitive">i1</a></tt> type.
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1751	</dd>
				1752
				1753	<dt><b>Integer constants</b></dt>
				1754
Reid Spencer	cc16dc3	2004-12-09 18:02:53 +0000	[diff] [blame]	1755	<dd>Standard integers (such as '4') are constants of the <a
Reid Spencer	a517338	2007-01-04 16:43:23 +0000	[diff] [blame]	1756	href="#t_integer">integer</a> type. Negative numbers may be used with
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1757	integer types.
				1758	</dd>
				1759
				1760	<dt><b>Floating point constants</b></dt>
				1761
				1762	<dd>Floating point constants use standard decimal notation (e.g. 123.421),
				1763	exponential notation (e.g. 1.23421e+2), or a more precise hexadecimal
Chris Lattner	a73afe0	2008-04-01 18:45:27 +0000	[diff] [blame]	1764	notation (see below). The assembler requires the exact decimal value of
				1765	a floating-point constant. For example, the assembler accepts 1.25 but
				1766	rejects 1.3 because 1.3 is a repeating decimal in binary. Floating point
				1767	constants must have a <a href="#t_floating">floating point</a> type. </dd>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1768
				1769	<dt><b>Null pointer constants</b></dt>
				1770
John Criswell	9e2485c	2004-12-10 15:51:16 +0000	[diff] [blame]	1771	<dd>The identifier '<tt>null</tt>' is recognized as a null pointer constant
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1772	and must be of <a href="#t_pointer">pointer type</a>.</dd>
				1773
				1774	</dl>
				1775
Dale Johannesen	bd5e5a8	2009-02-11 22:14:51 +0000	[diff] [blame]	1776	<p>The one non-intuitive notation for constants is the hexadecimal form
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1777	of floating point constants. For example, the form '<tt>double
				1778	0x432ff973cafa8000</tt>' is equivalent to (but harder to read than) '<tt>double
				1779	4.5e+15</tt>'. The only time hexadecimal floating point constants are required
Reid Spencer	cc16dc3	2004-12-09 18:02:53 +0000	[diff] [blame]	1780	(and the only time that they are generated by the disassembler) is when a
				1781	floating point constant must be emitted but it cannot be represented as a
Dale Johannesen	bd5e5a8	2009-02-11 22:14:51 +0000	[diff] [blame]	1782	decimal floating point number in a reasonable number of digits. For example,
				1783	NaN's, infinities, and other
Reid Spencer	cc16dc3	2004-12-09 18:02:53 +0000	[diff] [blame]	1784	special values are represented in their IEEE hexadecimal format so that
				1785	assembly and disassembly do not cause any bits to change in the constants.</p>
Dale Johannesen	bd5e5a8	2009-02-11 22:14:51 +0000	[diff] [blame]	1786	<p>When using the hexadecimal form, constants of types float and double are
				1787	represented using the 16-digit form shown above (which matches the IEEE754
				1788	representation for double); float values must, however, be exactly representable
				1789	as IEE754 single precision.
				1790	Hexadecimal format is always used for long
				1791	double, and there are three forms of long double. The 80-bit
				1792	format used by x86 is represented as <tt>0xK</tt>
				1793	followed by 20 hexadecimal digits.
				1794	The 128-bit format used by PowerPC (two adjacent doubles) is represented
				1795	by <tt>0xM</tt> followed by 32 hexadecimal digits. The IEEE 128-bit
				1796	format is represented
				1797	by <tt>0xL</tt> followed by 32 hexadecimal digits; no currently supported
				1798	target uses this format. Long doubles will only work if they match
				1799	the long double format on your target. All hexadecimal formats are big-endian
				1800	(sign bit at the left).</p>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1801	</div>
				1802
				1803	<!-- ======================================================================= -->
Chris Lattner	7088279	2009-02-28 18:32:25 +0000	[diff] [blame]	1804	<div class="doc_subsection">
				1805	<a name="aggregateconstants"> <!-- old anchor -->
				1806	<a name="complexconstants">Complex Constants</a></a>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1807	</div>
				1808
				1809	<div class="doc_text">
Chris Lattner	7088279	2009-02-28 18:32:25 +0000	[diff] [blame]	1810	<p>Complex constants are a (potentially recursive) combination of simple
				1811	constants and smaller complex constants.</p>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1812
				1813	<dl>
				1814	<dt><b>Structure constants</b></dt>
				1815
				1816	<dd>Structure constants are represented with notation similar to structure
				1817	type definitions (a comma separated list of elements, surrounded by braces
Chris Lattner	64910ee	2007-12-25 20:34:52 +0000	[diff] [blame]	1818	(<tt>{}</tt>)). For example: "<tt>{ i32 4, float 17.0, i32* @G }</tt>",
				1819	where "<tt>@G</tt>" is declared as "<tt>@G = external global i32</tt>". Structure constants
Chris Lattner	d4f6b17	2005-03-07 22:13:59 +0000	[diff] [blame]	1820	must have <a href="#t_struct">structure type</a>, and the number and
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1821	types of elements must match those specified by the type.
				1822	</dd>
				1823
				1824	<dt><b>Array constants</b></dt>
				1825
				1826	<dd>Array constants are represented with notation similar to array type
				1827	definitions (a comma separated list of elements, surrounded by square brackets
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	1828	(<tt>[]</tt>)). For example: "<tt>[ i32 42, i32 11, i32 74 ]</tt>". Array
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1829	constants must have <a href="#t_array">array type</a>, and the number and
				1830	types of elements must match those specified by the type.
				1831	</dd>
				1832
Reid Spencer	485bad1	2007-02-15 03:07:05 +0000	[diff] [blame]	1833	<dt><b>Vector constants</b></dt>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1834
Reid Spencer	485bad1	2007-02-15 03:07:05 +0000	[diff] [blame]	1835	<dd>Vector constants are represented with notation similar to vector type
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1836	definitions (a comma separated list of elements, surrounded by
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	1837	less-than/greater-than's (<tt><></tt>)). For example: "<tt>< i32 42,
Jeff Cohen	6f1cc77	2007-04-22 01:17:39 +0000	[diff] [blame]	1838	i32 11, i32 74, i32 100 ></tt>". Vector constants must have <a
Reid Spencer	485bad1	2007-02-15 03:07:05 +0000	[diff] [blame]	1839	href="#t_vector">vector type</a>, and the number and types of elements must
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1840	match those specified by the type.
				1841	</dd>
				1842
				1843	<dt><b>Zero initialization</b></dt>
				1844
				1845	<dd>The string '<tt>zeroinitializer</tt>' can be used to zero initialize a
				1846	value to zero of <em>any</em> type, including scalar and aggregate types.
				1847	This is often used to avoid having to print large zero initializers (e.g. for
John Criswell	0ec250c	2005-10-24 16:17:18 +0000	[diff] [blame]	1848	large arrays) and is always exactly equivalent to using explicit zero
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1849	initializers.
				1850	</dd>
				1851	</dl>
				1852
				1853	</div>
				1854
				1855	<!-- ======================================================================= -->
				1856	<div class="doc_subsection">
				1857	<a name="globalconstants">Global Variable and Function Addresses</a>
				1858	</div>
				1859
				1860	<div class="doc_text">
				1861
				1862	<p>The addresses of <a href="#globalvars">global variables</a> and <a
				1863	href="#functionstructure">functions</a> are always implicitly valid (link-time)
John Criswell	9e2485c	2004-12-10 15:51:16 +0000	[diff] [blame]	1864	constants. These constants are explicitly referenced when the <a
				1865	href="#identifiers">identifier for the global</a> is used and always have <a
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1866	href="#t_pointer">pointer</a> type. For example, the following is a legal LLVM
				1867	file:</p>
				1868
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	1869	<div class="doc_code">
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1870	<pre>
Chris Lattner	a18a424	2007-06-06 18:28:13 +0000	[diff] [blame]	1871	@X = global i32 17
				1872	@Y = global i32 42
				1873	@Z = global [2 x i32] [ i32 @X, i32* @Y ]
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1874	</pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	1875	</div>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1876
				1877	</div>
				1878
				1879	<!-- ======================================================================= -->
Reid Spencer	2dc45b8	2004-12-09 18:13:12 +0000	[diff] [blame]	1880	<div class="doc_subsection"><a name="undefvalues">Undefined Values</a></div>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1881	<div class="doc_text">
Reid Spencer	2dc45b8	2004-12-09 18:13:12 +0000	[diff] [blame]	1882	<p>The string '<tt>undef</tt>' is recognized as a type-less constant that has
John Criswell	c1f786c	2005-05-13 22:25:59 +0000	[diff] [blame]	1883	no specific value. Undefined values may be of any type and be used anywhere
Reid Spencer	2dc45b8	2004-12-09 18:13:12 +0000	[diff] [blame]	1884	a constant is permitted.</p>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1885
Reid Spencer	2dc45b8	2004-12-09 18:13:12 +0000	[diff] [blame]	1886	<p>Undefined values indicate to the compiler that the program is well defined
				1887	no matter what value is used, giving the compiler more freedom to optimize.
				1888	</p>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1889	</div>
				1890
				1891	<!-- ======================================================================= -->
				1892	<div class="doc_subsection"><a name="constantexprs">Constant Expressions</a>
				1893	</div>
				1894
				1895	<div class="doc_text">
				1896
				1897	<p>Constant expressions are used to allow expressions involving other constants
				1898	to be used as constants. Constant expressions may be of any <a
John Criswell	c1f786c	2005-05-13 22:25:59 +0000	[diff] [blame]	1899	href="#t_firstclass">first class</a> type and may involve any LLVM operation
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1900	that does not have side effects (e.g. load and call are not supported). The
				1901	following is the syntax for constant expressions:</p>
				1902
				1903	<dl>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	1904	<dt><b><tt>trunc ( CST to TYPE )</tt></b></dt>
				1905	<dd>Truncate a constant to another type. The bit size of CST must be larger
Chris Lattner	3b19d65	2007-01-15 01:54:13 +0000	[diff] [blame]	1906	than the bit size of TYPE. Both types must be integers.</dd>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1907
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	1908	<dt><b><tt>zext ( CST to TYPE )</tt></b></dt>
				1909	<dd>Zero extend a constant to another type. The bit size of CST must be
Chris Lattner	3b19d65	2007-01-15 01:54:13 +0000	[diff] [blame]	1910	smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	1911
				1912	<dt><b><tt>sext ( CST to TYPE )</tt></b></dt>
				1913	<dd>Sign extend a constant to another type. The bit size of CST must be
Chris Lattner	3b19d65	2007-01-15 01:54:13 +0000	[diff] [blame]	1914	smaller or equal to the bit size of TYPE. Both types must be integers.</dd>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	1915
				1916	<dt><b><tt>fptrunc ( CST to TYPE )</tt></b></dt>
				1917	<dd>Truncate a floating point constant to another floating point type. The
				1918	size of CST must be larger than the size of TYPE. Both types must be
				1919	floating point.</dd>
				1920
				1921	<dt><b><tt>fpext ( CST to TYPE )</tt></b></dt>
				1922	<dd>Floating point extend a constant to another type. The size of CST must be
				1923	smaller or equal to the size of TYPE. Both types must be floating point.</dd>
				1924
Reid Spencer	1539a1c	2007-07-31 14:40:14 +0000	[diff] [blame]	1925	<dt><b><tt>fptoui ( CST to TYPE )</tt></b></dt>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	1926	<dd>Convert a floating point constant to the corresponding unsigned integer
Nate Begeman	b348d18	2007-11-17 03:58:34 +0000	[diff] [blame]	1927	constant. TYPE must be a scalar or vector integer type. CST must be of scalar
				1928	or vector floating point type. Both CST and TYPE must be scalars, or vectors
				1929	of the same number of elements. If the value won't fit in the integer type,
				1930	the results are undefined.</dd>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	1931
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	1932	<dt><b><tt>fptosi ( CST to TYPE )</tt></b></dt>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	1933	<dd>Convert a floating point constant to the corresponding signed integer
Nate Begeman	b348d18	2007-11-17 03:58:34 +0000	[diff] [blame]	1934	constant. TYPE must be a scalar or vector integer type. CST must be of scalar
				1935	or vector floating point type. Both CST and TYPE must be scalars, or vectors
				1936	of the same number of elements. If the value won't fit in the integer type,
				1937	the results are undefined.</dd>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	1938
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	1939	<dt><b><tt>uitofp ( CST to TYPE )</tt></b></dt>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	1940	<dd>Convert an unsigned integer constant to the corresponding floating point
Nate Begeman	b348d18	2007-11-17 03:58:34 +0000	[diff] [blame]	1941	constant. TYPE must be a scalar or vector floating point type. CST must be of
				1942	scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
				1943	of the same number of elements. If the value won't fit in the floating point
				1944	type, the results are undefined.</dd>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	1945
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	1946	<dt><b><tt>sitofp ( CST to TYPE )</tt></b></dt>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	1947	<dd>Convert a signed integer constant to the corresponding floating point
Nate Begeman	b348d18	2007-11-17 03:58:34 +0000	[diff] [blame]	1948	constant. TYPE must be a scalar or vector floating point type. CST must be of
				1949	scalar or vector integer type. Both CST and TYPE must be scalars, or vectors
				1950	of the same number of elements. If the value won't fit in the floating point
				1951	type, the results are undefined.</dd>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	1952
Reid Spencer	5c0ef47	2006-11-11 23:08:07 +0000	[diff] [blame]	1953	<dt><b><tt>ptrtoint ( CST to TYPE )</tt></b></dt>
				1954	<dd>Convert a pointer typed constant to the corresponding integer constant
				1955	TYPE must be an integer type. CST must be of pointer type. The CST value is
				1956	zero extended, truncated, or unchanged to make it fit in TYPE.</dd>
				1957
				1958	<dt><b><tt>inttoptr ( CST to TYPE )</tt></b></dt>
				1959	<dd>Convert a integer constant to a pointer constant. TYPE must be a
				1960	pointer type. CST must be of integer type. The CST value is zero extended,
				1961	truncated, or unchanged to make it fit in a pointer size. This one is
				1962	<i>really</i> dangerous!</dd>
				1963
				1964	<dt><b><tt>bitcast ( CST to TYPE )</tt></b></dt>
Chris Lattner	03bbad6	2009-02-28 18:27:03 +0000	[diff] [blame]	1965	<dd>Convert a constant, CST, to another TYPE. The constraints of the operands
				1966	are the same as those for the <a href="#i_bitcast">bitcast
				1967	instruction</a>.</dd>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	1968
				1969	<dt><b><tt>getelementptr ( CSTPTR, IDX0, IDX1, ... )</tt></b></dt>
				1970
				1971	<dd>Perform the <a href="#i_getelementptr">getelementptr operation</a> on
				1972	constants. As with the <a href="#i_getelementptr">getelementptr</a>
				1973	instruction, the index list may have zero or more indexes, which are required
				1974	to make sense for the type of "CSTPTR".</dd>
				1975
Robert Bocchino	9fbe145	2006-01-10 19:31:34 +0000	[diff] [blame]	1976	<dt><b><tt>select ( COND, VAL1, VAL2 )</tt></b></dt>
				1977
				1978	<dd>Perform the <a href="#i_select">select operation</a> on
Reid Spencer	01c4259	2006-12-04 19:23:19 +0000	[diff] [blame]	1979	constants.</dd>
				1980
				1981	<dt><b><tt>icmp COND ( VAL1, VAL2 )</tt></b></dt>
				1982	<dd>Performs the <a href="#i_icmp">icmp operation</a> on constants.</dd>
				1983
				1984	<dt><b><tt>fcmp COND ( VAL1, VAL2 )</tt></b></dt>
				1985	<dd>Performs the <a href="#i_fcmp">fcmp operation</a> on constants.</dd>
Robert Bocchino	9fbe145	2006-01-10 19:31:34 +0000	[diff] [blame]	1986
Nate Begeman	ac80ade	2008-05-12 19:01:56 +0000	[diff] [blame]	1987	<dt><b><tt>vicmp COND ( VAL1, VAL2 )</tt></b></dt>
				1988	<dd>Performs the <a href="#i_vicmp">vicmp operation</a> on constants.</dd>
				1989
				1990	<dt><b><tt>vfcmp COND ( VAL1, VAL2 )</tt></b></dt>
				1991	<dd>Performs the <a href="#i_vfcmp">vfcmp operation</a> on constants.</dd>
				1992
Robert Bocchino	9fbe145	2006-01-10 19:31:34 +0000	[diff] [blame]	1993	<dt><b><tt>extractelement ( VAL, IDX )</tt></b></dt>
				1994
				1995	<dd>Perform the <a href="#i_extractelement">extractelement
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	1996	operation</a> on constants.</dd>
Robert Bocchino	9fbe145	2006-01-10 19:31:34 +0000	[diff] [blame]	1997
Robert Bocchino	05ccd70	2006-01-15 20:48:27 +0000	[diff] [blame]	1998	<dt><b><tt>insertelement ( VAL, ELT, IDX )</tt></b></dt>
				1999
				2000	<dd>Perform the <a href="#i_insertelement">insertelement
Reid Spencer	01c4259	2006-12-04 19:23:19 +0000	[diff] [blame]	2001	operation</a> on constants.</dd>
Robert Bocchino	05ccd70	2006-01-15 20:48:27 +0000	[diff] [blame]	2002
Chris Lattner	c198954	2006-04-08 00:13:41 +0000	[diff] [blame]	2003
				2004	<dt><b><tt>shufflevector ( VEC1, VEC2, IDXMASK )</tt></b></dt>
				2005
				2006	<dd>Perform the <a href="#i_shufflevector">shufflevector
Reid Spencer	01c4259	2006-12-04 19:23:19 +0000	[diff] [blame]	2007	operation</a> on constants.</dd>
Chris Lattner	c198954	2006-04-08 00:13:41 +0000	[diff] [blame]	2008
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	2009	<dt><b><tt>OPCODE ( LHS, RHS )</tt></b></dt>
				2010
Reid Spencer	2dc45b8	2004-12-09 18:13:12 +0000	[diff] [blame]	2011	<dd>Perform the specified operation of the LHS and RHS constants. OPCODE may
				2012	be any of the <a href="#binaryops">binary</a> or <a href="#bitwiseops">bitwise
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	2013	binary</a> operations. The constraints on operands are the same as those for
				2014	the corresponding instruction (e.g. no bitwise operations on floating point
John Criswell	e4c57cc	2005-05-12 16:52:32 +0000	[diff] [blame]	2015	values are allowed).</dd>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	2016	</dl>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	2017	</div>
Chris Lattner	9ee5d22	2004-03-08 16:49:10 +0000	[diff] [blame]	2018
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2019	<!-- *********************************************************************** -->
Chris Lattner	e87d653	2006-01-25 23:47:57 +0000	[diff] [blame]	2020	<div class="doc_section"> <a name="othervalues">Other Values</a> </div>
				2021	<!-- *********************************************************************** -->
				2022
				2023	<!-- ======================================================================= -->
				2024	<div class="doc_subsection">
				2025	<a name="inlineasm">Inline Assembler Expressions</a>
				2026	</div>
				2027
				2028	<div class="doc_text">
				2029
				2030	<p>
				2031	LLVM supports inline assembler expressions (as opposed to <a href="#moduleasm">
				2032	Module-Level Inline Assembly</a>) through the use of a special value. This
				2033	value represents the inline assembler as a string (containing the instructions
				2034	to emit), a list of operand constraints (stored as a string), and a flag that
				2035	indicates whether or not the inline asm expression has side effects. An example
				2036	inline assembler expression is:
				2037	</p>
				2038
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	2039	<div class="doc_code">
Chris Lattner	e87d653	2006-01-25 23:47:57 +0000	[diff] [blame]	2040	<pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	2041	i32 (i32) asm "bswap $0", "=r,r"
Chris Lattner	e87d653	2006-01-25 23:47:57 +0000	[diff] [blame]	2042	</pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	2043	</div>
Chris Lattner	e87d653	2006-01-25 23:47:57 +0000	[diff] [blame]	2044
				2045	<p>
				2046	Inline assembler expressions may <b>only</b> be used as the callee operand of
				2047	a <a href="#i_call"><tt>call</tt> instruction</a>. Thus, typically we have:
				2048	</p>
				2049
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	2050	<div class="doc_code">
Chris Lattner	e87d653	2006-01-25 23:47:57 +0000	[diff] [blame]	2051	<pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	2052	%X = call i32 asm "<a href="#int_bswap">bswap</a> $0", "=r,r"(i32 %Y)
Chris Lattner	e87d653	2006-01-25 23:47:57 +0000	[diff] [blame]	2053	</pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	2054	</div>
Chris Lattner	e87d653	2006-01-25 23:47:57 +0000	[diff] [blame]	2055
				2056	<p>
				2057	Inline asms with side effects not visible in the constraint list must be marked
				2058	as having side effects. This is done through the use of the
				2059	'<tt>sideeffect</tt>' keyword, like so:
				2060	</p>
				2061
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	2062	<div class="doc_code">
Chris Lattner	e87d653	2006-01-25 23:47:57 +0000	[diff] [blame]	2063	<pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	2064	call void asm sideeffect "eieio", ""()
Chris Lattner	e87d653	2006-01-25 23:47:57 +0000	[diff] [blame]	2065	</pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	2066	</div>
Chris Lattner	e87d653	2006-01-25 23:47:57 +0000	[diff] [blame]	2067
				2068	<p>TODO: The format of the asm and constraints string still need to be
				2069	documented here. Constraints on what can be done (e.g. duplication, moving, etc
Chris Lattner	4f99335	2008-10-04 18:36:02 +0000	[diff] [blame]	2070	need to be documented). This is probably best done by reference to another
				2071	document that covers inline asm from a holistic perspective.
Chris Lattner	e87d653	2006-01-25 23:47:57 +0000	[diff] [blame]	2072	</p>
				2073
				2074	</div>
				2075
				2076	<!-- *********************************************************************** -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2077	<div class="doc_section"> <a name="instref">Instruction Reference</a> </div>
				2078	<!-- *********************************************************************** -->
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	2079
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2080	<div class="doc_text">
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	2081
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2082	<p>The LLVM instruction set consists of several different
				2083	classifications of instructions: <a href="#terminators">terminator
John Criswell	c1f786c	2005-05-13 22:25:59 +0000	[diff] [blame]	2084	instructions</a>, <a href="#binaryops">binary instructions</a>,
				2085	<a href="#bitwiseops">bitwise binary instructions</a>, <a
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2086	href="#memoryops">memory instructions</a>, and <a href="#otherops">other
				2087	instructions</a>.</p>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	2088
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2089	</div>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	2090
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2091	<!-- ======================================================================= -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2092	<div class="doc_subsection"> <a name="terminators">Terminator
				2093	Instructions</a> </div>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	2094
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2095	<div class="doc_text">
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	2096
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2097	<p>As mentioned <a href="#functionstructure">previously</a>, every
				2098	basic block in a program ends with a "Terminator" instruction, which
				2099	indicates which block should be executed after the current block is
				2100	finished. These terminator instructions typically yield a '<tt>void</tt>'
				2101	value: they produce control flow, not values (the one exception being
				2102	the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction).</p>
John Criswell	9e2485c	2004-12-10 15:51:16 +0000	[diff] [blame]	2103	<p>There are six different terminator instructions: the '<a
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2104	href="#i_ret"><tt>ret</tt></a>' instruction, the '<a href="#i_br"><tt>br</tt></a>'
				2105	instruction, the '<a href="#i_switch"><tt>switch</tt></a>' instruction,
Chris Lattner	35eca58	2004-10-16 18:04:13 +0000	[diff] [blame]	2106	the '<a href="#i_invoke"><tt>invoke</tt></a>' instruction, the '<a
				2107	href="#i_unwind"><tt>unwind</tt></a>' instruction, and the '<a
				2108	href="#i_unreachable"><tt>unreachable</tt></a>' instruction.</p>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	2109
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2110	</div>
Chris Lattner	c3f5976	2004-12-09 17:30:23 +0000	[diff] [blame]	2111
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2112	<!-- _______________________________________________________________________ -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2113	<div class="doc_subsubsection"> <a name="i_ret">'<tt>ret</tt>'
				2114	Instruction</a> </div>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2115	<div class="doc_text">
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2116	<h5>Syntax:</h5>
Dan Gohman	b1e6b96	2008-10-04 19:00:07 +0000	[diff] [blame]	2117	<pre>
				2118	ret <type> <value> <i>; Return a value from a non-void function</i>
Chris Lattner	7faa883	2002-04-14 06:13:44 +0000	[diff] [blame]	2119	ret void <i>; Return from void function</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2120	</pre>
Chris Lattner	f4cde4e	2008-04-23 04:59:35 +0000	[diff] [blame]	2121
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2122	<h5>Overview:</h5>
Chris Lattner	f4cde4e	2008-04-23 04:59:35 +0000	[diff] [blame]	2123
Dan Gohman	b1e6b96	2008-10-04 19:00:07 +0000	[diff] [blame]	2124	<p>The '<tt>ret</tt>' instruction is used to return control flow (and
				2125	optionally a value) from a function back to the caller.</p>
John Criswell	4457dc9	2004-04-09 16:48:45 +0000	[diff] [blame]	2126	<p>There are two forms of the '<tt>ret</tt>' instruction: one that
Dan Gohman	b1e6b96	2008-10-04 19:00:07 +0000	[diff] [blame]	2127	returns a value and then causes control flow, and one that just causes
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2128	control flow to occur.</p>
Chris Lattner	f4cde4e	2008-04-23 04:59:35 +0000	[diff] [blame]	2129
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2130	<h5>Arguments:</h5>
Chris Lattner	f4cde4e	2008-04-23 04:59:35 +0000	[diff] [blame]	2131
Dan Gohman	b1e6b96	2008-10-04 19:00:07 +0000	[diff] [blame]	2132	<p>The '<tt>ret</tt>' instruction optionally accepts a single argument,
				2133	the return value. The type of the return value must be a
				2134	'<a href="#t_firstclass">first class</a>' type.</p>
				2135
				2136	<p>A function is not <a href="#wellformed">well formed</a> if
				2137	it it has a non-void return type and contains a '<tt>ret</tt>'
				2138	instruction with no return value or a return value with a type that
				2139	does not match its type, or if it has a void return type and contains
				2140	a '<tt>ret</tt>' instruction with a return value.</p>
Chris Lattner	f4cde4e	2008-04-23 04:59:35 +0000	[diff] [blame]	2141
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2142	<h5>Semantics:</h5>
Chris Lattner	f4cde4e	2008-04-23 04:59:35 +0000	[diff] [blame]	2143
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2144	<p>When the '<tt>ret</tt>' instruction is executed, control flow
				2145	returns back to the calling function's context. If the caller is a "<a
John Criswell	fa08187	2004-06-25 15:16:57 +0000	[diff] [blame]	2146	href="#i_call"><tt>call</tt></a>" instruction, execution continues at
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2147	the instruction after the call. If the caller was an "<a
				2148	href="#i_invoke"><tt>invoke</tt></a>" instruction, execution continues
John Criswell	e4c57cc	2005-05-12 16:52:32 +0000	[diff] [blame]	2149	at the beginning of the "normal" destination block. If the instruction
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2150	returns a value, that value shall set the call or invoke instruction's
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	2151	return value.</p>
Chris Lattner	f4cde4e	2008-04-23 04:59:35 +0000	[diff] [blame]	2152
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2153	<h5>Example:</h5>
Chris Lattner	f4cde4e	2008-04-23 04:59:35 +0000	[diff] [blame]	2154
				2155	<pre>
				2156	ret i32 5 <i>; Return an integer value of 5</i>
Chris Lattner	7faa883	2002-04-14 06:13:44 +0000	[diff] [blame]	2157	ret void <i>; Return from a void function</i>
Bill Wendling	0a4bbbf	2009-02-28 22:12:54 +0000	[diff] [blame]	2158	ret { i32, i8 } { i32 4, i8 2 } <i>; Return a struct of values 4 and 2</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2159	</pre>
Dan Gohman	f3e60bd	2009-01-12 23:12:39 +0000	[diff] [blame]	2160
Dan Gohman	d8791e5	2009-01-24 15:58:40 +0000	[diff] [blame]	2161	<p>Note that the code generator does not yet fully support large
				2162	return values. The specific sizes that are currently supported are
				2163	dependent on the target. For integers, on 32-bit targets the limit
				2164	is often 64 bits, and on 64-bit targets the limit is often 128 bits.
				2165	For aggregate types, the current limits are dependent on the element
				2166	types; for example targets are often limited to 2 total integer
				2167	elements and 2 total floating-point elements.</p>
Dan Gohman	f3e60bd	2009-01-12 23:12:39 +0000	[diff] [blame]	2168
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2169	</div>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2170	<!-- _______________________________________________________________________ -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2171	<div class="doc_subsubsection"> <a name="i_br">'<tt>br</tt>' Instruction</a> </div>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2172	<div class="doc_text">
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2173	<h5>Syntax:</h5>
Reid Spencer	c78f337	2007-01-12 03:35:51 +0000	[diff] [blame]	2174	<pre> br i1 <cond>, label <iftrue>, label <iffalse><br> br label <dest> <i>; Unconditional branch</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2175	</pre>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2176	<h5>Overview:</h5>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2177	<p>The '<tt>br</tt>' instruction is used to cause control flow to
				2178	transfer to a different basic block in the current function. There are
				2179	two forms of this instruction, corresponding to a conditional branch
				2180	and an unconditional branch.</p>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2181	<h5>Arguments:</h5>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2182	<p>The conditional branch form of the '<tt>br</tt>' instruction takes a
Reid Spencer	c78f337	2007-01-12 03:35:51 +0000	[diff] [blame]	2183	single '<tt>i1</tt>' value and two '<tt>label</tt>' values. The
Reid Spencer	de15194	2007-02-19 23:54:10 +0000	[diff] [blame]	2184	unconditional form of the '<tt>br</tt>' instruction takes a single
				2185	'<tt>label</tt>' value as a target.</p>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2186	<h5>Semantics:</h5>
Reid Spencer	c78f337	2007-01-12 03:35:51 +0000	[diff] [blame]	2187	<p>Upon execution of a conditional '<tt>br</tt>' instruction, the '<tt>i1</tt>'
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2188	argument is evaluated. If the value is <tt>true</tt>, control flows
				2189	to the '<tt>iftrue</tt>' <tt>label</tt> argument. If "cond" is <tt>false</tt>,
				2190	control flows to the '<tt>iffalse</tt>' <tt>label</tt> argument.</p>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2191	<h5>Example:</h5>
Reid Spencer	c78f337	2007-01-12 03:35:51 +0000	[diff] [blame]	2192	<pre>Test:<br> %cond = <a href="#i_icmp">icmp</a> eq, i32 %a, %b<br> br i1 %cond, label %IfEqual, label %IfUnequal<br>IfEqual:<br> <a
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	2193	href="#i_ret">ret</a> i32 1<br>IfUnequal:<br> <a href="#i_ret">ret</a> i32 0<br></pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2194	</div>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2195	<!-- _______________________________________________________________________ -->
Chris Lattner	c88c17b	2004-02-24 04:54:45 +0000	[diff] [blame]	2196	<div class="doc_subsubsection">
				2197	<a name="i_switch">'<tt>switch</tt>' Instruction</a>
				2198	</div>
				2199
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2200	<div class="doc_text">
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2201	<h5>Syntax:</h5>
Chris Lattner	c88c17b	2004-02-24 04:54:45 +0000	[diff] [blame]	2202
				2203	<pre>
				2204	switch <intty> <value>, label <defaultdest> [ <intty> <val>, label <dest> ... ]
				2205	</pre>
				2206
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2207	<h5>Overview:</h5>
Chris Lattner	c88c17b	2004-02-24 04:54:45 +0000	[diff] [blame]	2208
				2209	<p>The '<tt>switch</tt>' instruction is used to transfer control flow to one of
				2210	several different places. It is a generalization of the '<tt>br</tt>'
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2211	instruction, allowing a branch to occur to one of many possible
				2212	destinations.</p>
Chris Lattner	c88c17b	2004-02-24 04:54:45 +0000	[diff] [blame]	2213
				2214
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2215	<h5>Arguments:</h5>
Chris Lattner	c88c17b	2004-02-24 04:54:45 +0000	[diff] [blame]	2216
				2217	<p>The '<tt>switch</tt>' instruction uses three parameters: an integer
				2218	comparison value '<tt>value</tt>', a default '<tt>label</tt>' destination, and
				2219	an array of pairs of comparison value constants and '<tt>label</tt>'s. The
				2220	table is not allowed to contain duplicate constant entries.</p>
				2221
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2222	<h5>Semantics:</h5>
Chris Lattner	c88c17b	2004-02-24 04:54:45 +0000	[diff] [blame]	2223
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2224	<p>The <tt>switch</tt> instruction specifies a table of values and
				2225	destinations. When the '<tt>switch</tt>' instruction is executed, this
John Criswell	8411475	2004-06-25 16:05:06 +0000	[diff] [blame]	2226	table is searched for the given value. If the value is found, control flow is
				2227	transfered to the corresponding destination; otherwise, control flow is
				2228	transfered to the default destination.</p>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2229
Chris Lattner	c88c17b	2004-02-24 04:54:45 +0000	[diff] [blame]	2230	<h5>Implementation:</h5>
				2231
				2232	<p>Depending on properties of the target machine and the particular
				2233	<tt>switch</tt> instruction, this instruction may be code generated in different
John Criswell	8411475	2004-06-25 16:05:06 +0000	[diff] [blame]	2234	ways. For example, it could be generated as a series of chained conditional
				2235	branches or with a lookup table.</p>
Chris Lattner	c88c17b	2004-02-24 04:54:45 +0000	[diff] [blame]	2236
				2237	<h5>Example:</h5>
				2238
				2239	<pre>
				2240	<i>; Emulate a conditional br instruction</i>
Reid Spencer	c78f337	2007-01-12 03:35:51 +0000	[diff] [blame]	2241	%Val = <a href="#i_zext">zext</a> i1 %value to i32
Dan Gohman	2a08c53	2009-01-04 23:44:43 +0000	[diff] [blame]	2242	switch i32 %Val, label %truedest [ i32 0, label %falsedest ]
Chris Lattner	c88c17b	2004-02-24 04:54:45 +0000	[diff] [blame]	2243
				2244	<i>; Emulate an unconditional br instruction</i>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	2245	switch i32 0, label %dest [ ]
Chris Lattner	c88c17b	2004-02-24 04:54:45 +0000	[diff] [blame]	2246
				2247	<i>; Implement a jump table:</i>
Dan Gohman	2a08c53	2009-01-04 23:44:43 +0000	[diff] [blame]	2248	switch i32 %val, label %otherwise [ i32 0, label %onzero
				2249	i32 1, label %onone
				2250	i32 2, label %ontwo ]
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2251	</pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2252	</div>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2253
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2254	<!-- _______________________________________________________________________ -->
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2255	<div class="doc_subsubsection">
				2256	<a name="i_invoke">'<tt>invoke</tt>' Instruction</a>
				2257	</div>
				2258
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2259	<div class="doc_text">
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2260
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2261	<h5>Syntax:</h5>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2262
				2263	<pre>
Devang Patel	307e8ab	2008-10-07 17:48:33 +0000	[diff] [blame]	2264	<result> = invoke [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>] <ptr to function ty> <function ptr val>(<function args>) [<a href="#fnattrs">fn attrs</a>]
Chris Lattner	76b8a33	2006-05-14 18:23:06 +0000	[diff] [blame]	2265	to label <normal label> unwind label <exception label>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2266	</pre>
				2267
Chris Lattner	6536cfe	2002-05-06 22:08:29 +0000	[diff] [blame]	2268	<h5>Overview:</h5>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2269
				2270	<p>The '<tt>invoke</tt>' instruction causes control to transfer to a specified
				2271	function, with the possibility of control flow transfer to either the
John Criswell	e4c57cc	2005-05-12 16:52:32 +0000	[diff] [blame]	2272	'<tt>normal</tt>' label or the
				2273	'<tt>exception</tt>' label. If the callee function returns with the
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2274	"<tt><a href="#i_ret">ret</a></tt>" instruction, control flow will return to the
				2275	"normal" label. If the callee (or any indirect callees) returns with the "<a
John Criswell	e4c57cc	2005-05-12 16:52:32 +0000	[diff] [blame]	2276	href="#i_unwind"><tt>unwind</tt></a>" instruction, control is interrupted and
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	2277	continued at the dynamically nearest "exception" label.</p>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2278
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2279	<h5>Arguments:</h5>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2280
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2281	<p>This instruction requires several arguments:</p>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2282
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2283	<ol>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2284	<li>
Duncan Sands	8036ca4	2007-03-30 12:22:09 +0000	[diff] [blame]	2285	The optional "cconv" marker indicates which <a href="#callingconv">calling
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2286	convention</a> the call should use. If none is specified, the call defaults
				2287	to using C calling conventions.
				2288	</li>
Devang Patel	f642f47	2008-10-06 18:50:38 +0000	[diff] [blame]	2289
				2290	<li>The optional <a href="#paramattrs">Parameter Attributes</a> list for
				2291	return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
				2292	and '<tt>inreg</tt>' attributes are valid here.</li>
				2293
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2294	<li>'<tt>ptr to function ty</tt>': shall be the signature of the pointer to
				2295	function value being invoked. In most cases, this is a direct function
				2296	invocation, but indirect <tt>invoke</tt>s are just as possible, branching off
				2297	an arbitrary pointer to function value.
				2298	</li>
				2299
				2300	<li>'<tt>function ptr val</tt>': An LLVM value containing a pointer to a
				2301	function to be invoked. </li>
				2302
				2303	<li>'<tt>function args</tt>': argument list whose types match the function
				2304	signature argument types. If the function signature indicates the function
				2305	accepts a variable number of arguments, the extra arguments can be
				2306	specified. </li>
				2307
				2308	<li>'<tt>normal label</tt>': the label reached when the called function
				2309	executes a '<tt><a href="#i_ret">ret</a></tt>' instruction. </li>
				2310
				2311	<li>'<tt>exception label</tt>': the label reached when a callee returns with
				2312	the <a href="#i_unwind"><tt>unwind</tt></a> instruction. </li>
				2313
Devang Patel	307e8ab	2008-10-07 17:48:33 +0000	[diff] [blame]	2314	<li>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patel	f642f47	2008-10-06 18:50:38 +0000	[diff] [blame]	2315	'<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
				2316	'<tt>readnone</tt>' attributes are valid here.</li>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2317	</ol>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2318
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2319	<h5>Semantics:</h5>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2320
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2321	<p>This instruction is designed to operate as a standard '<tt><a
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2322	href="#i_call">call</a></tt>' instruction in most regards. The primary
				2323	difference is that it establishes an association with a label, which is used by
				2324	the runtime library to unwind the stack.</p>
				2325
				2326	<p>This instruction is used in languages with destructors to ensure that proper
				2327	cleanup is performed in the case of either a <tt>longjmp</tt> or a thrown
				2328	exception. Additionally, this is important for implementation of
				2329	'<tt>catch</tt>' clauses in high-level languages that support them.</p>
				2330
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2331	<h5>Example:</h5>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	2332	<pre>
Nick Lewycky	d703f65	2008-03-16 07:18:12 +0000	[diff] [blame]	2333	%retval = invoke i32 @Test(i32 15) to label %Continue
Jeff Cohen	6f1cc77	2007-04-22 01:17:39 +0000	[diff] [blame]	2334	unwind label %TestCleanup <i>; {i32}:retval set</i>
Nick Lewycky	d703f65	2008-03-16 07:18:12 +0000	[diff] [blame]	2335	%retval = invoke <a href="#callingconv">coldcc</a> i32 %Testfnptr(i32 15) to label %Continue
Jeff Cohen	6f1cc77	2007-04-22 01:17:39 +0000	[diff] [blame]	2336	unwind label %TestCleanup <i>; {i32}:retval set</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2337	</pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2338	</div>
Chris Lattner	35eca58	2004-10-16 18:04:13 +0000	[diff] [blame]	2339
				2340
Chris Lattner	27f71f2	2003-09-03 00:41:47 +0000	[diff] [blame]	2341	<!-- _______________________________________________________________________ -->
Chris Lattner	35eca58	2004-10-16 18:04:13 +0000	[diff] [blame]	2342
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2343	<div class="doc_subsubsection"> <a name="i_unwind">'<tt>unwind</tt>'
				2344	Instruction</a> </div>
Chris Lattner	35eca58	2004-10-16 18:04:13 +0000	[diff] [blame]	2345
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2346	<div class="doc_text">
Chris Lattner	35eca58	2004-10-16 18:04:13 +0000	[diff] [blame]	2347
Chris Lattner	27f71f2	2003-09-03 00:41:47 +0000	[diff] [blame]	2348	<h5>Syntax:</h5>
Chris Lattner	35eca58	2004-10-16 18:04:13 +0000	[diff] [blame]	2349	<pre>
				2350	unwind
				2351	</pre>
				2352
Chris Lattner	27f71f2	2003-09-03 00:41:47 +0000	[diff] [blame]	2353	<h5>Overview:</h5>
Chris Lattner	35eca58	2004-10-16 18:04:13 +0000	[diff] [blame]	2354
				2355	<p>The '<tt>unwind</tt>' instruction unwinds the stack, continuing control flow
				2356	at the first callee in the dynamic call stack which used an <a
				2357	href="#i_invoke"><tt>invoke</tt></a> instruction to perform the call. This is
				2358	primarily used to implement exception handling.</p>
				2359
Chris Lattner	27f71f2	2003-09-03 00:41:47 +0000	[diff] [blame]	2360	<h5>Semantics:</h5>
Chris Lattner	35eca58	2004-10-16 18:04:13 +0000	[diff] [blame]	2361
Chris Lattner	72ed200	2008-04-19 21:01:16 +0000	[diff] [blame]	2362	<p>The '<tt>unwind</tt>' instruction causes execution of the current function to
Chris Lattner	35eca58	2004-10-16 18:04:13 +0000	[diff] [blame]	2363	immediately halt. The dynamic call stack is then searched for the first <a
				2364	href="#i_invoke"><tt>invoke</tt></a> instruction on the call stack. Once found,
				2365	execution continues at the "exceptional" destination block specified by the
				2366	<tt>invoke</tt> instruction. If there is no <tt>invoke</tt> instruction in the
				2367	dynamic call chain, undefined behavior results.</p>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2368	</div>
Chris Lattner	35eca58	2004-10-16 18:04:13 +0000	[diff] [blame]	2369
				2370	<!-- _______________________________________________________________________ -->
				2371
				2372	<div class="doc_subsubsection"> <a name="i_unreachable">'<tt>unreachable</tt>'
				2373	Instruction</a> </div>
				2374
				2375	<div class="doc_text">
				2376
				2377	<h5>Syntax:</h5>
				2378	<pre>
				2379	unreachable
				2380	</pre>
				2381
				2382	<h5>Overview:</h5>
				2383
				2384	<p>The '<tt>unreachable</tt>' instruction has no defined semantics. This
				2385	instruction is used to inform the optimizer that a particular portion of the
				2386	code is not reachable. This can be used to indicate that the code after a
				2387	no-return function cannot be reached, and other facts.</p>
				2388
				2389	<h5>Semantics:</h5>
				2390
				2391	<p>The '<tt>unreachable</tt>' instruction has no defined semantics.</p>
				2392	</div>
				2393
				2394
				2395
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2396	<!-- ======================================================================= -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2397	<div class="doc_subsection"> <a name="binaryops">Binary Operations</a> </div>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2398	<div class="doc_text">
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2399	<p>Binary operators are used to do most of the computation in a
Chris Lattner	5a15814	2008-04-01 18:47:32 +0000	[diff] [blame]	2400	program. They require two operands of the same type, execute an operation on them, and
John Criswell	9e2485c	2004-12-10 15:51:16 +0000	[diff] [blame]	2401	produce a single value. The operands might represent
Reid Spencer	485bad1	2007-02-15 03:07:05 +0000	[diff] [blame]	2402	multiple data, as is the case with the <a href="#t_vector">vector</a> data type.
Chris Lattner	5a15814	2008-04-01 18:47:32 +0000	[diff] [blame]	2403	The result value has the same type as its operands.</p>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2404	<p>There are several different binary operators:</p>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2405	</div>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2406	<!-- _______________________________________________________________________ -->
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2407	<div class="doc_subsubsection">
				2408	<a name="i_add">'<tt>add</tt>' Instruction</a>
				2409	</div>
				2410
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2411	<div class="doc_text">
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2412
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2413	<h5>Syntax:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2414
				2415	<pre>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2416	<result> = add <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2417	</pre>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2418
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2419	<h5>Overview:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2420
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2421	<p>The '<tt>add</tt>' instruction returns the sum of its two operands.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2422
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2423	<h5>Arguments:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2424
				2425	<p>The two arguments to the '<tt>add</tt>' instruction must be <a
				2426	href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>, or
				2427	<a href="#t_vector">vector</a> values. Both arguments must have identical
				2428	types.</p>
				2429
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2430	<h5>Semantics:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2431
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2432	<p>The value produced is the integer or floating point sum of the two
				2433	operands.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2434
Chris Lattner	5ec8983	2008-01-28 00:36:27 +0000	[diff] [blame]	2435	<p>If an integer sum has unsigned overflow, the result returned is the
				2436	mathematical result modulo 2<sup>n</sup>, where n is the bit width of
				2437	the result.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2438
Chris Lattner	5ec8983	2008-01-28 00:36:27 +0000	[diff] [blame]	2439	<p>Because LLVM integers use a two's complement representation, this
				2440	instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2441
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2442	<h5>Example:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2443
				2444	<pre>
				2445	<result> = add i32 4, %var <i>; yields {i32}:result = 4 + %var</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2446	</pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2447	</div>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2448	<!-- _______________________________________________________________________ -->
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2449	<div class="doc_subsubsection">
				2450	<a name="i_sub">'<tt>sub</tt>' Instruction</a>
				2451	</div>
				2452
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2453	<div class="doc_text">
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2454
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2455	<h5>Syntax:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2456
				2457	<pre>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2458	<result> = sub <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2459	</pre>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2460
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2461	<h5>Overview:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2462
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2463	<p>The '<tt>sub</tt>' instruction returns the difference of its two
				2464	operands.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2465
				2466	<p>Note that the '<tt>sub</tt>' instruction is used to represent the
				2467	'<tt>neg</tt>' instruction present in most other intermediate
				2468	representations.</p>
				2469
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2470	<h5>Arguments:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2471
				2472	<p>The two arguments to the '<tt>sub</tt>' instruction must be <a
				2473	href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
				2474	or <a href="#t_vector">vector</a> values. Both arguments must have identical
				2475	types.</p>
				2476
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2477	<h5>Semantics:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2478
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2479	<p>The value produced is the integer or floating point difference of
				2480	the two operands.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2481
Chris Lattner	5ec8983	2008-01-28 00:36:27 +0000	[diff] [blame]	2482	<p>If an integer difference has unsigned overflow, the result returned is the
				2483	mathematical result modulo 2<sup>n</sup>, where n is the bit width of
				2484	the result.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2485
Chris Lattner	5ec8983	2008-01-28 00:36:27 +0000	[diff] [blame]	2486	<p>Because LLVM integers use a two's complement representation, this
				2487	instruction is appropriate for both signed and unsigned integers.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2488
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2489	<h5>Example:</h5>
Bill Wendling	aac388b	2007-05-29 09:42:13 +0000	[diff] [blame]	2490	<pre>
				2491	<result> = sub i32 4, %var <i>; yields {i32}:result = 4 - %var</i>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	2492	<result> = sub i32 0, %val <i>; yields {i32}:result = -%var</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2493	</pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2494	</div>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2495
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2496	<!-- _______________________________________________________________________ -->
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2497	<div class="doc_subsubsection">
				2498	<a name="i_mul">'<tt>mul</tt>' Instruction</a>
				2499	</div>
				2500
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2501	<div class="doc_text">
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2502
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2503	<h5>Syntax:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2504	<pre> <result> = mul <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2505	</pre>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2506	<h5>Overview:</h5>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2507	<p>The '<tt>mul</tt>' instruction returns the product of its two
				2508	operands.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2509
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2510	<h5>Arguments:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2511
				2512	<p>The two arguments to the '<tt>mul</tt>' instruction must be <a
				2513	href="#t_integer">integer</a>, <a href="#t_floating">floating point</a>,
				2514	or <a href="#t_vector">vector</a> values. Both arguments must have identical
				2515	types.</p>
				2516
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2517	<h5>Semantics:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2518
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2519	<p>The value produced is the integer or floating point product of the
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2520	two operands.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2521
Chris Lattner	5ec8983	2008-01-28 00:36:27 +0000	[diff] [blame]	2522	<p>If the result of an integer multiplication has unsigned overflow,
				2523	the result returned is the mathematical result modulo
				2524	2<sup>n</sup>, where n is the bit width of the result.</p>
				2525	<p>Because LLVM integers use a two's complement representation, and the
				2526	result is the same width as the operands, this instruction returns the
				2527	correct result for both signed and unsigned integers. If a full product
				2528	(e.g. <tt>i32</tt>x<tt>i32</tt>-><tt>i64</tt>) is needed, the operands
				2529	should be sign-extended or zero-extended as appropriate to the
				2530	width of the full product.</p>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2531	<h5>Example:</h5>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	2532	<pre> <result> = mul i32 4, %var <i>; yields {i32}:result = 4 * %var</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2533	</pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2534	</div>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2535
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2536	<!-- _______________________________________________________________________ -->
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2537	<div class="doc_subsubsection"> <a name="i_udiv">'<tt>udiv</tt>' Instruction
				2538	</a></div>
				2539	<div class="doc_text">
				2540	<h5>Syntax:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2541	<pre> <result> = udiv <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2542	</pre>
				2543	<h5>Overview:</h5>
				2544	<p>The '<tt>udiv</tt>' instruction returns the quotient of its two
				2545	operands.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2546
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2547	<h5>Arguments:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2548
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2549	<p>The two arguments to the '<tt>udiv</tt>' instruction must be
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2550	<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
				2551	values. Both arguments must have identical types.</p>
				2552
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2553	<h5>Semantics:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2554
Chris Lattner	5ec8983	2008-01-28 00:36:27 +0000	[diff] [blame]	2555	<p>The value produced is the unsigned integer quotient of the two operands.</p>
				2556	<p>Note that unsigned integer division and signed integer division are distinct
				2557	operations; for signed integer division, use '<tt>sdiv</tt>'.</p>
				2558	<p>Division by zero leads to undefined behavior.</p>
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2559	<h5>Example:</h5>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	2560	<pre> <result> = udiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2561	</pre>
				2562	</div>
				2563	<!-- _______________________________________________________________________ -->
				2564	<div class="doc_subsubsection"> <a name="i_sdiv">'<tt>sdiv</tt>' Instruction
				2565	</a> </div>
				2566	<div class="doc_text">
				2567	<h5>Syntax:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2568	<pre>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2569	<result> = sdiv <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2570	</pre>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2571
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2572	<h5>Overview:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2573
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2574	<p>The '<tt>sdiv</tt>' instruction returns the quotient of its two
				2575	operands.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2576
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2577	<h5>Arguments:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2578
				2579	<p>The two arguments to the '<tt>sdiv</tt>' instruction must be
				2580	<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
				2581	values. Both arguments must have identical types.</p>
				2582
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2583	<h5>Semantics:</h5>
Chris Lattner	a73afe0	2008-04-01 18:45:27 +0000	[diff] [blame]	2584	<p>The value produced is the signed integer quotient of the two operands rounded towards zero.</p>
Chris Lattner	5ec8983	2008-01-28 00:36:27 +0000	[diff] [blame]	2585	<p>Note that signed integer division and unsigned integer division are distinct
				2586	operations; for unsigned integer division, use '<tt>udiv</tt>'.</p>
				2587	<p>Division by zero leads to undefined behavior. Overflow also leads to
				2588	undefined behavior; this is a rare case, but can occur, for example,
				2589	by doing a 32-bit division of -2147483648 by -1.</p>
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2590	<h5>Example:</h5>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	2591	<pre> <result> = sdiv i32 4, %var <i>; yields {i32}:result = 4 / %var</i>
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2592	</pre>
				2593	</div>
				2594	<!-- _______________________________________________________________________ -->
				2595	<div class="doc_subsubsection"> <a name="i_fdiv">'<tt>fdiv</tt>'
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2596	Instruction</a> </div>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2597	<div class="doc_text">
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2598	<h5>Syntax:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2599	<pre>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2600	<result> = fdiv <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2601	</pre>
				2602	<h5>Overview:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2603
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2604	<p>The '<tt>fdiv</tt>' instruction returns the quotient of its two
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2605	operands.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2606
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2607	<h5>Arguments:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2608
Jeff Cohen	6f1cc77	2007-04-22 01:17:39 +0000	[diff] [blame]	2609	<p>The two arguments to the '<tt>fdiv</tt>' instruction must be
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2610	<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
				2611	of floating point values. Both arguments must have identical types.</p>
				2612
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2613	<h5>Semantics:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2614
Reid Spencer	1628cec	2006-10-26 06:15:43 +0000	[diff] [blame]	2615	<p>The value produced is the floating point quotient of the two operands.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2616
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2617	<h5>Example:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2618
				2619	<pre>
				2620	<result> = fdiv float 4.0, %var <i>; yields {float}:result = 4.0 / %var</i>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2621	</pre>
				2622	</div>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2623
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2624	<!-- _______________________________________________________________________ -->
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	2625	<div class="doc_subsubsection"> <a name="i_urem">'<tt>urem</tt>' Instruction</a>
				2626	</div>
				2627	<div class="doc_text">
				2628	<h5>Syntax:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2629	<pre> <result> = urem <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	2630	</pre>
				2631	<h5>Overview:</h5>
				2632	<p>The '<tt>urem</tt>' instruction returns the remainder from the
				2633	unsigned division of its two arguments.</p>
				2634	<h5>Arguments:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2635	<p>The two arguments to the '<tt>urem</tt>' instruction must be
				2636	<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
				2637	values. Both arguments must have identical types.</p>
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	2638	<h5>Semantics:</h5>
				2639	<p>This instruction returns the unsigned integer <i>remainder</i> of a division.
Chris Lattner	a73afe0	2008-04-01 18:45:27 +0000	[diff] [blame]	2640	This instruction always performs an unsigned division to get the remainder.</p>
Chris Lattner	5ec8983	2008-01-28 00:36:27 +0000	[diff] [blame]	2641	<p>Note that unsigned integer remainder and signed integer remainder are
				2642	distinct operations; for signed integer remainder, use '<tt>srem</tt>'.</p>
				2643	<p>Taking the remainder of a division by zero leads to undefined behavior.</p>
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	2644	<h5>Example:</h5>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	2645	<pre> <result> = urem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	2646	</pre>
				2647
				2648	</div>
				2649	<!-- _______________________________________________________________________ -->
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2650	<div class="doc_subsubsection">
				2651	<a name="i_srem">'<tt>srem</tt>' Instruction</a>
				2652	</div>
				2653
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2654	<div class="doc_text">
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2655
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2656	<h5>Syntax:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2657
				2658	<pre>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2659	<result> = srem <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2660	</pre>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2661
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2662	<h5>Overview:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2663
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	2664	<p>The '<tt>srem</tt>' instruction returns the remainder from the
Dan Gohman	8017631	2007-11-05 23:35:22 +0000	[diff] [blame]	2665	signed division of its two operands. This instruction can also take
				2666	<a href="#t_vector">vector</a> versions of the values in which case
				2667	the elements must be integers.</p>
Chris Lattner	c7d3ab3	2008-01-04 04:33:49 +0000	[diff] [blame]	2668
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2669	<h5>Arguments:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2670
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	2671	<p>The two arguments to the '<tt>srem</tt>' instruction must be
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2672	<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
				2673	values. Both arguments must have identical types.</p>
				2674
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2675	<h5>Semantics:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2676
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	2677	<p>This instruction returns the <i>remainder</i> of a division (where the result
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2678	has the same sign as the dividend, <tt>op1</tt>), not the <i>modulo</i>
				2679	operator (where the result has the same sign as the divisor, <tt>op2</tt>) of
Reid Spencer	c9fdfc8	2007-03-24 22:23:39 +0000	[diff] [blame]	2680	a value. For more information about the difference, see <a
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2681	href="http://mathforum.org/dr.math/problems/anne.4.28.99.html">The
Reid Spencer	c9fdfc8	2007-03-24 22:23:39 +0000	[diff] [blame]	2682	Math Forum</a>. For a table of how this is implemented in various languages,
Reid Spencer	64f5c6c	2007-03-24 22:40:44 +0000	[diff] [blame]	2683	please see <a href="http://en.wikipedia.org/wiki/Modulo_operation">
Reid Spencer	c9fdfc8	2007-03-24 22:23:39 +0000	[diff] [blame]	2684	Wikipedia: modulo operation</a>.</p>
Chris Lattner	5ec8983	2008-01-28 00:36:27 +0000	[diff] [blame]	2685	<p>Note that signed integer remainder and unsigned integer remainder are
				2686	distinct operations; for unsigned integer remainder, use '<tt>urem</tt>'.</p>
				2687	<p>Taking the remainder of a division by zero leads to undefined behavior.
				2688	Overflow also leads to undefined behavior; this is a rare case, but can occur,
				2689	for example, by taking the remainder of a 32-bit division of -2147483648 by -1.
				2690	(The remainder doesn't actually overflow, but this rule lets srem be
				2691	implemented using instructions that return both the result of the division
				2692	and the remainder.)</p>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2693	<h5>Example:</h5>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	2694	<pre> <result> = srem i32 4, %var <i>; yields {i32}:result = 4 % %var</i>
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	2695	</pre>
				2696
				2697	</div>
				2698	<!-- _______________________________________________________________________ -->
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2699	<div class="doc_subsubsection">
				2700	<a name="i_frem">'<tt>frem</tt>' Instruction</a> </div>
				2701
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	2702	<div class="doc_text">
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2703
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	2704	<h5>Syntax:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2705	<pre> <result> = frem <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	2706	</pre>
				2707	<h5>Overview:</h5>
				2708	<p>The '<tt>frem</tt>' instruction returns the remainder from the
				2709	division of its two operands.</p>
				2710	<h5>Arguments:</h5>
				2711	<p>The two arguments to the '<tt>frem</tt>' instruction must be
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2712	<a href="#t_floating">floating point</a> or <a href="#t_vector">vector</a>
				2713	of floating point values. Both arguments must have identical types.</p>
				2714
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	2715	<h5>Semantics:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2716
Chris Lattner	a73afe0	2008-04-01 18:45:27 +0000	[diff] [blame]	2717	<p>This instruction returns the <i>remainder</i> of a division.
				2718	The remainder has the same sign as the dividend.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2719
Reid Spencer	0a783f7	2006-11-02 01:53:59 +0000	[diff] [blame]	2720	<h5>Example:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2721
				2722	<pre>
				2723	<result> = frem float 4.0, %var <i>; yields {float}:result = 4.0 % %var</i>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2724	</pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2725	</div>
Robert Bocchino	7b81c75	2006-02-17 21:18:08 +0000	[diff] [blame]	2726
Reid Spencer	8e11bf8	2007-02-02 13:57:07 +0000	[diff] [blame]	2727	<!-- ======================================================================= -->
				2728	<div class="doc_subsection"> <a name="bitwiseops">Bitwise Binary
				2729	Operations</a> </div>
				2730	<div class="doc_text">
				2731	<p>Bitwise binary operators are used to do various forms of
				2732	bit-twiddling in a program. They are generally very efficient
				2733	instructions and can commonly be strength reduced from other
Chris Lattner	a73afe0	2008-04-01 18:45:27 +0000	[diff] [blame]	2734	instructions. They require two operands of the same type, execute an operation on them,
				2735	and produce a single value. The resulting value is the same type as its operands.</p>
Reid Spencer	8e11bf8	2007-02-02 13:57:07 +0000	[diff] [blame]	2736	</div>
				2737
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2738	<!-- _______________________________________________________________________ -->
				2739	<div class="doc_subsubsection"> <a name="i_shl">'<tt>shl</tt>'
				2740	Instruction</a> </div>
				2741	<div class="doc_text">
				2742	<h5>Syntax:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2743	<pre> <result> = shl <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2744	</pre>
Chris Lattner	6ccc2d5	2007-10-03 21:01:14 +0000	[diff] [blame]	2745
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2746	<h5>Overview:</h5>
Chris Lattner	6ccc2d5	2007-10-03 21:01:14 +0000	[diff] [blame]	2747
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2748	<p>The '<tt>shl</tt>' instruction returns the first operand shifted to
				2749	the left a specified number of bits.</p>
Chris Lattner	6ccc2d5	2007-10-03 21:01:14 +0000	[diff] [blame]	2750
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2751	<h5>Arguments:</h5>
Chris Lattner	6ccc2d5	2007-10-03 21:01:14 +0000	[diff] [blame]	2752
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2753	<p>Both arguments to the '<tt>shl</tt>' instruction must be the same <a
Nate Begeman	5bc1ea0	2008-07-29 15:49:41 +0000	[diff] [blame]	2754	href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2755	type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Chris Lattner	6ccc2d5	2007-10-03 21:01:14 +0000	[diff] [blame]	2756
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2757	<h5>Semantics:</h5>
Chris Lattner	6ccc2d5	2007-10-03 21:01:14 +0000	[diff] [blame]	2758
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2759	<p>The value produced is <tt>op1</tt> * 2<sup><tt>op2</tt></sup> mod 2<sup>n</sup>,
				2760	where n is the width of the result. If <tt>op2</tt> is (statically or dynamically) negative or
Mon P Wang	01f8d09	2008-12-10 08:55:09 +0000	[diff] [blame]	2761	equal to or larger than the number of bits in <tt>op1</tt>, the result is undefined.
				2762	If the arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
				2763	corresponding shift amount in <tt>op2</tt>.</p>
Chris Lattner	6ccc2d5	2007-10-03 21:01:14 +0000	[diff] [blame]	2764
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2765	<h5>Example:</h5><pre>
				2766	<result> = shl i32 4, %var <i>; yields {i32}: 4 << %var</i>
				2767	<result> = shl i32 4, 2 <i>; yields {i32}: 16</i>
				2768	<result> = shl i32 1, 10 <i>; yields {i32}: 1024</i>
Chris Lattner	6ccc2d5	2007-10-03 21:01:14 +0000	[diff] [blame]	2769	<result> = shl i32 1, 32 <i>; undefined</i>
Mon P Wang	e9f1015	2008-12-09 05:46:39 +0000	[diff] [blame]	2770	<result> = shl <2 x i32> < i32 1, i32 1>, < i32 1, i32 2> <i>; yields: result=<2 x i32> < i32 2, i32 4></i>
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2771	</pre>
				2772	</div>
				2773	<!-- _______________________________________________________________________ -->
				2774	<div class="doc_subsubsection"> <a name="i_lshr">'<tt>lshr</tt>'
				2775	Instruction</a> </div>
				2776	<div class="doc_text">
				2777	<h5>Syntax:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2778	<pre> <result> = lshr <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2779	</pre>
				2780
				2781	<h5>Overview:</h5>
				2782	<p>The '<tt>lshr</tt>' instruction (logical shift right) returns the first
Jeff Cohen	6f1cc77	2007-04-22 01:17:39 +0000	[diff] [blame]	2783	operand shifted to the right a specified number of bits with zero fill.</p>
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2784
				2785	<h5>Arguments:</h5>
				2786	<p>Both arguments to the '<tt>lshr</tt>' instruction must be the same
Nate Begeman	5bc1ea0	2008-07-29 15:49:41 +0000	[diff] [blame]	2787	<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2788	type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2789
				2790	<h5>Semantics:</h5>
Chris Lattner	6ccc2d5	2007-10-03 21:01:14 +0000	[diff] [blame]	2791
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2792	<p>This instruction always performs a logical shift right operation. The most
				2793	significant bits of the result will be filled with zero bits after the
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2794	shift. If <tt>op2</tt> is (statically or dynamically) equal to or larger than
Mon P Wang	01f8d09	2008-12-10 08:55:09 +0000	[diff] [blame]	2795	the number of bits in <tt>op1</tt>, the result is undefined. If the arguments are
				2796	vectors, each vector element of <tt>op1</tt> is shifted by the corresponding shift
				2797	amount in <tt>op2</tt>.</p>
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2798
				2799	<h5>Example:</h5>
				2800	<pre>
				2801	<result> = lshr i32 4, 1 <i>; yields {i32}:result = 2</i>
				2802	<result> = lshr i32 4, 2 <i>; yields {i32}:result = 1</i>
				2803	<result> = lshr i8 4, 3 <i>; yields {i8}:result = 0</i>
				2804	<result> = lshr i8 -2, 1 <i>; yields {i8}:result = 0x7FFFFFFF </i>
Chris Lattner	6ccc2d5	2007-10-03 21:01:14 +0000	[diff] [blame]	2805	<result> = lshr i32 1, 32 <i>; undefined</i>
Mon P Wang	e9f1015	2008-12-09 05:46:39 +0000	[diff] [blame]	2806	<result> = lshr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 2> <i>; yields: result=<2 x i32> < i32 0x7FFFFFFF, i32 1></i>
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2807	</pre>
				2808	</div>
				2809
Reid Spencer	8e11bf8	2007-02-02 13:57:07 +0000	[diff] [blame]	2810	<!-- _______________________________________________________________________ -->
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2811	<div class="doc_subsubsection"> <a name="i_ashr">'<tt>ashr</tt>'
				2812	Instruction</a> </div>
				2813	<div class="doc_text">
				2814
				2815	<h5>Syntax:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2816	<pre> <result> = ashr <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2817	</pre>
				2818
				2819	<h5>Overview:</h5>
				2820	<p>The '<tt>ashr</tt>' instruction (arithmetic shift right) returns the first
Jeff Cohen	6f1cc77	2007-04-22 01:17:39 +0000	[diff] [blame]	2821	operand shifted to the right a specified number of bits with sign extension.</p>
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2822
				2823	<h5>Arguments:</h5>
				2824	<p>Both arguments to the '<tt>ashr</tt>' instruction must be the same
Nate Begeman	5bc1ea0	2008-07-29 15:49:41 +0000	[diff] [blame]	2825	<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2826	type. '<tt>op2</tt>' is treated as an unsigned value.</p>
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2827
				2828	<h5>Semantics:</h5>
				2829	<p>This instruction always performs an arithmetic shift right operation,
				2830	The most significant bits of the result will be filled with the sign bit
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2831	of <tt>op1</tt>. If <tt>op2</tt> is (statically or dynamically) equal to or
Mon P Wang	01f8d09	2008-12-10 08:55:09 +0000	[diff] [blame]	2832	larger than the number of bits in <tt>op1</tt>, the result is undefined. If the
				2833	arguments are vectors, each vector element of <tt>op1</tt> is shifted by the
				2834	corresponding shift amount in <tt>op2</tt>.</p>
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2835
				2836	<h5>Example:</h5>
				2837	<pre>
				2838	<result> = ashr i32 4, 1 <i>; yields {i32}:result = 2</i>
				2839	<result> = ashr i32 4, 2 <i>; yields {i32}:result = 1</i>
				2840	<result> = ashr i8 4, 3 <i>; yields {i8}:result = 0</i>
				2841	<result> = ashr i8 -2, 1 <i>; yields {i8}:result = -1</i>
Chris Lattner	6ccc2d5	2007-10-03 21:01:14 +0000	[diff] [blame]	2842	<result> = ashr i32 1, 32 <i>; undefined</i>
Mon P Wang	e9f1015	2008-12-09 05:46:39 +0000	[diff] [blame]	2843	<result> = ashr <2 x i32> < i32 -2, i32 4>, < i32 1, i32 3> <i>; yields: result=<2 x i32> < i32 -1, i32 0></i>
Reid Spencer	569f2fa	2007-01-31 21:39:12 +0000	[diff] [blame]	2844	</pre>
				2845	</div>
				2846
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2847	<!-- _______________________________________________________________________ -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2848	<div class="doc_subsubsection"> <a name="i_and">'<tt>and</tt>'
				2849	Instruction</a> </div>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2850
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2851	<div class="doc_text">
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2852
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2853	<h5>Syntax:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2854
				2855	<pre>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2856	<result> = and <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2857	</pre>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2858
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2859	<h5>Overview:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2860
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2861	<p>The '<tt>and</tt>' instruction returns the bitwise logical and of
				2862	its two operands.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2863
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2864	<h5>Arguments:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2865
				2866	<p>The two arguments to the '<tt>and</tt>' instruction must be
				2867	<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
				2868	values. Both arguments must have identical types.</p>
				2869
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2870	<h5>Semantics:</h5>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2871	<p>The truth table used for the '<tt>and</tt>' instruction is:</p>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2872	<p> </p>
Bill Wendling	c7e4c4d	2008-09-07 10:29:20 +0000	[diff] [blame]	2873	<div>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2874	<table border="1" cellspacing="0" cellpadding="4">
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2875	<tbody>
				2876	<tr>
				2877	<td>In0</td>
				2878	<td>In1</td>
				2879	<td>Out</td>
				2880	</tr>
				2881	<tr>
				2882	<td>0</td>
				2883	<td>0</td>
				2884	<td>0</td>
				2885	</tr>
				2886	<tr>
				2887	<td>0</td>
				2888	<td>1</td>
				2889	<td>0</td>
				2890	</tr>
				2891	<tr>
				2892	<td>1</td>
				2893	<td>0</td>
				2894	<td>0</td>
				2895	</tr>
				2896	<tr>
				2897	<td>1</td>
				2898	<td>1</td>
				2899	<td>1</td>
				2900	</tr>
				2901	</tbody>
				2902	</table>
Misha Brukman	daa4cb0	2004-03-01 17:47:27 +0000	[diff] [blame]	2903	</div>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2904	<h5>Example:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2905	<pre>
				2906	<result> = and i32 4, %var <i>; yields {i32}:result = 4 & %var</i>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	2907	<result> = and i32 15, 40 <i>; yields {i32}:result = 8</i>
				2908	<result> = and i32 4, 8 <i>; yields {i32}:result = 0</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2909	</pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2910	</div>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2911	<!-- _______________________________________________________________________ -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2912	<div class="doc_subsubsection"> <a name="i_or">'<tt>or</tt>' Instruction</a> </div>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2913	<div class="doc_text">
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2914	<h5>Syntax:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2915	<pre> <result> = or <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2916	</pre>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2917	<h5>Overview:</h5>
				2918	<p>The '<tt>or</tt>' instruction returns the bitwise logical inclusive
				2919	or of its two operands.</p>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2920	<h5>Arguments:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2921
				2922	<p>The two arguments to the '<tt>or</tt>' instruction must be
				2923	<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
				2924	values. Both arguments must have identical types.</p>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2925	<h5>Semantics:</h5>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2926	<p>The truth table used for the '<tt>or</tt>' instruction is:</p>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2927	<p> </p>
Bill Wendling	c7e4c4d	2008-09-07 10:29:20 +0000	[diff] [blame]	2928	<div>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2929	<table border="1" cellspacing="0" cellpadding="4">
				2930	<tbody>
				2931	<tr>
				2932	<td>In0</td>
				2933	<td>In1</td>
				2934	<td>Out</td>
				2935	</tr>
				2936	<tr>
				2937	<td>0</td>
				2938	<td>0</td>
				2939	<td>0</td>
				2940	</tr>
				2941	<tr>
				2942	<td>0</td>
				2943	<td>1</td>
				2944	<td>1</td>
				2945	</tr>
				2946	<tr>
				2947	<td>1</td>
				2948	<td>0</td>
				2949	<td>1</td>
				2950	</tr>
				2951	<tr>
				2952	<td>1</td>
				2953	<td>1</td>
				2954	<td>1</td>
				2955	</tr>
				2956	</tbody>
				2957	</table>
Misha Brukman	daa4cb0	2004-03-01 17:47:27 +0000	[diff] [blame]	2958	</div>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2959	<h5>Example:</h5>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	2960	<pre> <result> = or i32 4, %var <i>; yields {i32}:result = 4 \| %var</i>
				2961	<result> = or i32 15, 40 <i>; yields {i32}:result = 47</i>
				2962	<result> = or i32 4, 8 <i>; yields {i32}:result = 12</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2963	</pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2964	</div>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2965	<!-- _______________________________________________________________________ -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2966	<div class="doc_subsubsection"> <a name="i_xor">'<tt>xor</tt>'
				2967	Instruction</a> </div>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2968	<div class="doc_text">
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2969	<h5>Syntax:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	2970	<pre> <result> = xor <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2971	</pre>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2972	<h5>Overview:</h5>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2973	<p>The '<tt>xor</tt>' instruction returns the bitwise logical exclusive
				2974	or of its two operands. The <tt>xor</tt> is used to implement the
				2975	"one's complement" operation, which is the "~" operator in C.</p>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2976	<h5>Arguments:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2977	<p>The two arguments to the '<tt>xor</tt>' instruction must be
				2978	<a href="#t_integer">integer</a> or <a href="#t_vector">vector</a> of integer
				2979	values. Both arguments must have identical types.</p>
				2980
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	2981	<h5>Semantics:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	2982
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	2983	<p>The truth table used for the '<tt>xor</tt>' instruction is:</p>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2984	<p> </p>
Bill Wendling	c7e4c4d	2008-09-07 10:29:20 +0000	[diff] [blame]	2985	<div>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	2986	<table border="1" cellspacing="0" cellpadding="4">
				2987	<tbody>
				2988	<tr>
				2989	<td>In0</td>
				2990	<td>In1</td>
				2991	<td>Out</td>
				2992	</tr>
				2993	<tr>
				2994	<td>0</td>
				2995	<td>0</td>
				2996	<td>0</td>
				2997	</tr>
				2998	<tr>
				2999	<td>0</td>
				3000	<td>1</td>
				3001	<td>1</td>
				3002	</tr>
				3003	<tr>
				3004	<td>1</td>
				3005	<td>0</td>
				3006	<td>1</td>
				3007	</tr>
				3008	<tr>
				3009	<td>1</td>
				3010	<td>1</td>
				3011	<td>0</td>
				3012	</tr>
				3013	</tbody>
				3014	</table>
Misha Brukman	daa4cb0	2004-03-01 17:47:27 +0000	[diff] [blame]	3015	</div>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3016	<p> </p>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3017	<h5>Example:</h5>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3018	<pre> <result> = xor i32 4, %var <i>; yields {i32}:result = 4 ^ %var</i>
				3019	<result> = xor i32 15, 40 <i>; yields {i32}:result = 39</i>
				3020	<result> = xor i32 4, 8 <i>; yields {i32}:result = 12</i>
				3021	<result> = xor i32 %V, -1 <i>; yields {i32}:result = ~%V</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3022	</pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3023	</div>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3024
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3025	<!-- ======================================================================= -->
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3026	<div class="doc_subsection">
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3027	<a name="vectorops">Vector Operations</a>
				3028	</div>
				3029
				3030	<div class="doc_text">
				3031
				3032	<p>LLVM supports several instructions to represent vector operations in a
Jeff Cohen	6f1cc77	2007-04-22 01:17:39 +0000	[diff] [blame]	3033	target-independent manner. These instructions cover the element-access and
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3034	vector-specific operations needed to process vectors effectively. While LLVM
				3035	does directly support these vector operations, many sophisticated algorithms
				3036	will want to use target-specific intrinsics to take full advantage of a specific
				3037	target.</p>
				3038
				3039	</div>
				3040
				3041	<!-- _______________________________________________________________________ -->
				3042	<div class="doc_subsubsection">
				3043	<a name="i_extractelement">'<tt>extractelement</tt>' Instruction</a>
				3044	</div>
				3045
				3046	<div class="doc_text">
				3047
				3048	<h5>Syntax:</h5>
				3049
				3050	<pre>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3051	<result> = extractelement <n x <ty>> <val>, i32 <idx> <i>; yields <ty></i>
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3052	</pre>
				3053
				3054	<h5>Overview:</h5>
				3055
				3056	<p>
				3057	The '<tt>extractelement</tt>' instruction extracts a single scalar
Reid Spencer	485bad1	2007-02-15 03:07:05 +0000	[diff] [blame]	3058	element from a vector at a specified index.
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3059	</p>
				3060
				3061
				3062	<h5>Arguments:</h5>
				3063
				3064	<p>
				3065	The first operand of an '<tt>extractelement</tt>' instruction is a
Reid Spencer	485bad1	2007-02-15 03:07:05 +0000	[diff] [blame]	3066	value of <a href="#t_vector">vector</a> type. The second operand is
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3067	an index indicating the position from which to extract the element.
				3068	The index may be a variable.</p>
				3069
				3070	<h5>Semantics:</h5>
				3071
				3072	<p>
				3073	The result is a scalar of the same type as the element type of
				3074	<tt>val</tt>. Its value is the value at position <tt>idx</tt> of
				3075	<tt>val</tt>. If <tt>idx</tt> exceeds the length of <tt>val</tt>, the
				3076	results are undefined.
				3077	</p>
				3078
				3079	<h5>Example:</h5>
				3080
				3081	<pre>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3082	%result = extractelement <4 x i32> %vec, i32 0 <i>; yields i32</i>
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3083	</pre>
				3084	</div>
				3085
				3086
				3087	<!-- _______________________________________________________________________ -->
				3088	<div class="doc_subsubsection">
				3089	<a name="i_insertelement">'<tt>insertelement</tt>' Instruction</a>
				3090	</div>
				3091
				3092	<div class="doc_text">
				3093
				3094	<h5>Syntax:</h5>
				3095
				3096	<pre>
Dan Gohman	f3480b9	2008-05-12 23:38:42 +0000	[diff] [blame]	3097	<result> = insertelement <n x <ty>> <val>, <ty> <elt>, i32 <idx> <i>; yields <n x <ty>></i>
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3098	</pre>
				3099
				3100	<h5>Overview:</h5>
				3101
				3102	<p>
				3103	The '<tt>insertelement</tt>' instruction inserts a scalar
Reid Spencer	485bad1	2007-02-15 03:07:05 +0000	[diff] [blame]	3104	element into a vector at a specified index.
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3105	</p>
				3106
				3107
				3108	<h5>Arguments:</h5>
				3109
				3110	<p>
				3111	The first operand of an '<tt>insertelement</tt>' instruction is a
Reid Spencer	485bad1	2007-02-15 03:07:05 +0000	[diff] [blame]	3112	value of <a href="#t_vector">vector</a> type. The second operand is a
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3113	scalar value whose type must equal the element type of the first
				3114	operand. The third operand is an index indicating the position at
				3115	which to insert the value. The index may be a variable.</p>
				3116
				3117	<h5>Semantics:</h5>
				3118
				3119	<p>
Reid Spencer	485bad1	2007-02-15 03:07:05 +0000	[diff] [blame]	3120	The result is a vector of the same type as <tt>val</tt>. Its
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3121	element values are those of <tt>val</tt> except at position
				3122	<tt>idx</tt>, where it gets the value <tt>elt</tt>. If <tt>idx</tt>
				3123	exceeds the length of <tt>val</tt>, the results are undefined.
				3124	</p>
				3125
				3126	<h5>Example:</h5>
				3127
				3128	<pre>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3129	%result = insertelement <4 x i32> %vec, i32 1, i32 0 <i>; yields <4 x i32></i>
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3130	</pre>
				3131	</div>
				3132
				3133	<!-- _______________________________________________________________________ -->
				3134	<div class="doc_subsubsection">
				3135	<a name="i_shufflevector">'<tt>shufflevector</tt>' Instruction</a>
				3136	</div>
				3137
				3138	<div class="doc_text">
				3139
				3140	<h5>Syntax:</h5>
				3141
				3142	<pre>
Mon P Wang	aeb06d2	2008-11-10 04:46:22 +0000	[diff] [blame]	3143	<result> = shufflevector <n x <ty>> <v1>, <n x <ty>> <v2>, <m x i32> <mask> <i>; yields <m x <ty>></i>
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3144	</pre>
				3145
				3146	<h5>Overview:</h5>
				3147
				3148	<p>
				3149	The '<tt>shufflevector</tt>' instruction constructs a permutation of elements
Mon P Wang	aeb06d2	2008-11-10 04:46:22 +0000	[diff] [blame]	3150	from two input vectors, returning a vector with the same element type as
				3151	the input and length that is the same as the shuffle mask.
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3152	</p>
				3153
				3154	<h5>Arguments:</h5>
				3155
				3156	<p>
Mon P Wang	aeb06d2	2008-11-10 04:46:22 +0000	[diff] [blame]	3157	The first two operands of a '<tt>shufflevector</tt>' instruction are vectors
				3158	with types that match each other. The third argument is a shuffle mask whose
				3159	element type is always 'i32'. The result of the instruction is a vector whose
				3160	length is the same as the shuffle mask and whose element type is the same as
				3161	the element type of the first two operands.
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3162	</p>
				3163
				3164	<p>
				3165	The shuffle mask operand is required to be a constant vector with either
				3166	constant integer or undef values.
				3167	</p>
				3168
				3169	<h5>Semantics:</h5>
				3170
				3171	<p>
				3172	The elements of the two input vectors are numbered from left to right across
				3173	both of the vectors. The shuffle mask operand specifies, for each element of
Mon P Wang	aeb06d2	2008-11-10 04:46:22 +0000	[diff] [blame]	3174	the result vector, which element of the two input vectors the result element
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3175	gets. The element selector may be undef (meaning "don't care") and the second
				3176	operand may be undef if performing a shuffle from only one vector.
				3177	</p>
				3178
				3179	<h5>Example:</h5>
				3180
				3181	<pre>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3182	%result = shufflevector <4 x i32> %v1, <4 x i32> %v2,
Jeff Cohen	6f1cc77	2007-04-22 01:17:39 +0000	[diff] [blame]	3183	<4 x i32> <i32 0, i32 4, i32 1, i32 5> <i>; yields <4 x i32></i>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3184	%result = shufflevector <4 x i32> %v1, <4 x i32> undef,
				3185	<4 x i32> <i32 0, i32 1, i32 2, i32 3> <i>; yields <4 x i32></i> - Identity shuffle.
Mon P Wang	aeb06d2	2008-11-10 04:46:22 +0000	[diff] [blame]	3186	%result = shufflevector <8 x i32> %v1, <8 x i32> undef,
				3187	<4 x i32> <i32 0, i32 1, i32 2, i32 3> <i>; yields <4 x i32></i>
				3188	%result = shufflevector <4 x i32> %v1, <4 x i32> %v2,
				3189	<8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7 > <i>; yields <8 x i32></i>
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3190	</pre>
				3191	</div>
				3192
Tanya Lattner	0947429	2006-04-14 19:24:33 +0000	[diff] [blame]	3193
Chris Lattner	3df241e	2006-04-08 23:07:04 +0000	[diff] [blame]	3194	<!-- ======================================================================= -->
				3195	<div class="doc_subsection">
Dan Gohman	a334d5f	2008-05-12 23:51:09 +0000	[diff] [blame]	3196	<a name="aggregateops">Aggregate Operations</a>
				3197	</div>
				3198
				3199	<div class="doc_text">
				3200
				3201	<p>LLVM supports several instructions for working with aggregate values.
				3202	</p>
				3203
				3204	</div>
				3205
				3206	<!-- _______________________________________________________________________ -->
				3207	<div class="doc_subsubsection">
				3208	<a name="i_extractvalue">'<tt>extractvalue</tt>' Instruction</a>
				3209	</div>
				3210
				3211	<div class="doc_text">
				3212
				3213	<h5>Syntax:</h5>
				3214
				3215	<pre>
				3216	<result> = extractvalue <aggregate type> <val>, <idx>{, <idx>}*
				3217	</pre>
				3218
				3219	<h5>Overview:</h5>
				3220
				3221	<p>
Dan Gohman	c3dac5c	2008-05-13 18:16:06 +0000	[diff] [blame]	3222	The '<tt>extractvalue</tt>' instruction extracts the value of a struct field
				3223	or array element from an aggregate value.
Dan Gohman	a334d5f	2008-05-12 23:51:09 +0000	[diff] [blame]	3224	</p>
				3225
				3226
				3227	<h5>Arguments:</h5>
				3228
				3229	<p>
				3230	The first operand of an '<tt>extractvalue</tt>' instruction is a
				3231	value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a>
Dan Gohman	c3dac5c	2008-05-13 18:16:06 +0000	[diff] [blame]	3232	type. The operands are constant indices to specify which value to extract
Dan Gohman	81a0c0b	2008-05-31 00:58:22 +0000	[diff] [blame]	3233	in a similar manner as indices in a
Dan Gohman	a334d5f	2008-05-12 23:51:09 +0000	[diff] [blame]	3234	'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
				3235	</p>
				3236
				3237	<h5>Semantics:</h5>
				3238
				3239	<p>
				3240	The result is the value at the position in the aggregate specified by
				3241	the index operands.
				3242	</p>
				3243
				3244	<h5>Example:</h5>
				3245
				3246	<pre>
Dan Gohman	81a0c0b	2008-05-31 00:58:22 +0000	[diff] [blame]	3247	%result = extractvalue {i32, float} %agg, 0 <i>; yields i32</i>
Dan Gohman	a334d5f	2008-05-12 23:51:09 +0000	[diff] [blame]	3248	</pre>
				3249	</div>
				3250
				3251
				3252	<!-- _______________________________________________________________________ -->
				3253	<div class="doc_subsubsection">
				3254	<a name="i_insertvalue">'<tt>insertvalue</tt>' Instruction</a>
				3255	</div>
				3256
				3257	<div class="doc_text">
				3258
				3259	<h5>Syntax:</h5>
				3260
				3261	<pre>
Dan Gohman	81a0c0b	2008-05-31 00:58:22 +0000	[diff] [blame]	3262	<result> = insertvalue <aggregate type> <val>, <ty> <val>, <idx> <i>; yields <n x <ty>></i>
Dan Gohman	a334d5f	2008-05-12 23:51:09 +0000	[diff] [blame]	3263	</pre>
				3264
				3265	<h5>Overview:</h5>
				3266
				3267	<p>
				3268	The '<tt>insertvalue</tt>' instruction inserts a value
Dan Gohman	c3dac5c	2008-05-13 18:16:06 +0000	[diff] [blame]	3269	into a struct field or array element in an aggregate.
Dan Gohman	a334d5f	2008-05-12 23:51:09 +0000	[diff] [blame]	3270	</p>
				3271
				3272
				3273	<h5>Arguments:</h5>
				3274
				3275	<p>
				3276	The first operand of an '<tt>insertvalue</tt>' instruction is a
				3277	value of <a href="#t_struct">struct</a> or <a href="#t_array">array</a> type.
				3278	The second operand is a first-class value to insert.
Dan Gohman	c4b49eb	2008-05-23 21:53:15 +0000	[diff] [blame]	3279	The following operands are constant indices
Dan Gohman	81a0c0b	2008-05-31 00:58:22 +0000	[diff] [blame]	3280	indicating the position at which to insert the value in a similar manner as
Dan Gohman	c3dac5c	2008-05-13 18:16:06 +0000	[diff] [blame]	3281	indices in a
Dan Gohman	a334d5f	2008-05-12 23:51:09 +0000	[diff] [blame]	3282	'<tt><a href="#i_getelementptr">getelementptr</a></tt>' instruction.
				3283	The value to insert must have the same type as the value identified
Dan Gohman	c3dac5c	2008-05-13 18:16:06 +0000	[diff] [blame]	3284	by the indices.
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	3285	</p>
Dan Gohman	a334d5f	2008-05-12 23:51:09 +0000	[diff] [blame]	3286
				3287	<h5>Semantics:</h5>
				3288
				3289	<p>
				3290	The result is an aggregate of the same type as <tt>val</tt>. Its
				3291	value is that of <tt>val</tt> except that the value at the position
Dan Gohman	c3dac5c	2008-05-13 18:16:06 +0000	[diff] [blame]	3292	specified by the indices is that of <tt>elt</tt>.
Dan Gohman	a334d5f	2008-05-12 23:51:09 +0000	[diff] [blame]	3293	</p>
				3294
				3295	<h5>Example:</h5>
				3296
				3297	<pre>
Dan Gohman	52bb2db	2008-06-23 15:26:37 +0000	[diff] [blame]	3298	%result = insertvalue {i32, float} %agg, i32 1, 0 <i>; yields {i32, float}</i>
Dan Gohman	a334d5f	2008-05-12 23:51:09 +0000	[diff] [blame]	3299	</pre>
				3300	</div>
				3301
				3302
				3303	<!-- ======================================================================= -->
				3304	<div class="doc_subsection">
Chris Lattner	884a970	2006-08-15 00:45:58 +0000	[diff] [blame]	3305	<a name="memoryops">Memory Access and Addressing Operations</a>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3306	</div>
				3307
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3308	<div class="doc_text">
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3309
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3310	<p>A key design point of an SSA-based representation is how it
				3311	represents memory. In LLVM, no memory locations are in SSA form, which
				3312	makes things very simple. This section describes how to read, write,
John Criswell	9e2485c	2004-12-10 15:51:16 +0000	[diff] [blame]	3313	allocate, and free memory in LLVM.</p>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3314
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3315	</div>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3316
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3317	<!-- _______________________________________________________________________ -->
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3318	<div class="doc_subsubsection">
				3319	<a name="i_malloc">'<tt>malloc</tt>' Instruction</a>
				3320	</div>
				3321
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3322	<div class="doc_text">
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3323
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3324	<h5>Syntax:</h5>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3325
				3326	<pre>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3327	<result> = malloc <type>[, i32 <NumElements>][, align <alignment>] <i>; yields {type*}:result</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3328	</pre>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3329
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3330	<h5>Overview:</h5>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3331
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3332	<p>The '<tt>malloc</tt>' instruction allocates memory from the system
Christopher Lamb	303dae9	2007-12-17 01:00:21 +0000	[diff] [blame]	3333	heap and returns a pointer to it. The object is always allocated in the generic
				3334	address space (address space zero).</p>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3335
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3336	<h5>Arguments:</h5>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3337
				3338	<p>The '<tt>malloc</tt>' instruction allocates
				3339	<tt>sizeof(<type>)*NumElements</tt>
John Criswell	6e4ca61	2004-02-24 16:13:56 +0000	[diff] [blame]	3340	bytes of memory from the operating system and returns a pointer of the
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3341	appropriate type to the program. If "NumElements" is specified, it is the
Gabor Greif	1acd2ee	2008-02-09 22:24:34 +0000	[diff] [blame]	3342	number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner	4316dec	2008-04-02 00:38:26 +0000	[diff] [blame]	3343	If a constant alignment is specified, the value result of the allocation is guaranteed to
Gabor Greif	1acd2ee	2008-02-09 22:24:34 +0000	[diff] [blame]	3344	be aligned to at least that boundary. If not specified, or if zero, the target can
				3345	choose to align the allocation on any convenient boundary.</p>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3346
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3347	<p>'<tt>type</tt>' must be a sized type.</p>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3348
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3349	<h5>Semantics:</h5>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3350
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3351	<p>Memory is allocated using the system "<tt>malloc</tt>" function, and
Nick Lewycky	02ff308	2008-11-24 03:41:24 +0000	[diff] [blame]	3352	a pointer is returned. The result of a zero byte allocation is undefined. The
Chris Lattner	72ed200	2008-04-19 21:01:16 +0000	[diff] [blame]	3353	result is null if there is insufficient memory available.</p>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3354
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3355	<h5>Example:</h5>
				3356
				3357	<pre>
Dan Gohman	81e2167	2009-01-04 23:49:44 +0000	[diff] [blame]	3358	%array = malloc [4 x i8] <i>; yields {[%4 x i8]*}:array</i>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3359
Bill Wendling	aac388b	2007-05-29 09:42:13 +0000	[diff] [blame]	3360	%size = <a href="#i_add">add</a> i32 2, 2 <i>; yields {i32}:size = i32 4</i>
				3361	%array1 = malloc i8, i32 4 <i>; yields {i8*}:array1</i>
				3362	%array2 = malloc [12 x i8], i32 %size <i>; yields {[12 x i8]*}:array2</i>
				3363	%array3 = malloc i32, i32 4, align 1024 <i>; yields {i32*}:array3</i>
				3364	%array4 = malloc i32, align 1024 <i>; yields {i32*}:array4</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3365	</pre>
Dan Gohman	f3e60bd	2009-01-12 23:12:39 +0000	[diff] [blame]	3366
				3367	<p>Note that the code generator does not yet respect the
				3368	alignment value.</p>
				3369
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3370	</div>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3371
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3372	<!-- _______________________________________________________________________ -->
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3373	<div class="doc_subsubsection">
				3374	<a name="i_free">'<tt>free</tt>' Instruction</a>
				3375	</div>
				3376
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3377	<div class="doc_text">
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3378
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3379	<h5>Syntax:</h5>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3380
				3381	<pre>
Dan Gohman	81e2167	2009-01-04 23:49:44 +0000	[diff] [blame]	3382	free <type> <value> <i>; yields {void}</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3383	</pre>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3384
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3385	<h5>Overview:</h5>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3386
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3387	<p>The '<tt>free</tt>' instruction returns memory back to the unused
John Criswell	c1f786c	2005-05-13 22:25:59 +0000	[diff] [blame]	3388	memory heap to be reallocated in the future.</p>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3389
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3390	<h5>Arguments:</h5>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3391
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3392	<p>'<tt>value</tt>' shall be a pointer value that points to a value
				3393	that was allocated with the '<tt><a href="#i_malloc">malloc</a></tt>'
				3394	instruction.</p>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3395
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3396	<h5>Semantics:</h5>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3397
John Criswell	9e2485c	2004-12-10 15:51:16 +0000	[diff] [blame]	3398	<p>Access to the memory pointed to by the pointer is no longer defined
Chris Lattner	e0db56d	2008-04-19 22:41:32 +0000	[diff] [blame]	3399	after this instruction executes. If the pointer is null, the operation
				3400	is a noop.</p>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3401
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3402	<h5>Example:</h5>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3403
				3404	<pre>
Dan Gohman	81e2167	2009-01-04 23:49:44 +0000	[diff] [blame]	3405	%array = <a href="#i_malloc">malloc</a> [4 x i8] <i>; yields {[4 x i8]*}:array</i>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3406	free [4 x i8]* %array
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3407	</pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3408	</div>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3409
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3410	<!-- _______________________________________________________________________ -->
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3411	<div class="doc_subsubsection">
				3412	<a name="i_alloca">'<tt>alloca</tt>' Instruction</a>
				3413	</div>
				3414
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3415	<div class="doc_text">
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3416
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3417	<h5>Syntax:</h5>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3418
				3419	<pre>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3420	<result> = alloca <type>[, i32 <NumElements>][, align <alignment>] <i>; yields {type*}:result</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3421	</pre>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3422
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3423	<h5>Overview:</h5>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3424
Jeff Cohen	6f1cc77	2007-04-22 01:17:39 +0000	[diff] [blame]	3425	<p>The '<tt>alloca</tt>' instruction allocates memory on the stack frame of the
				3426	currently executing function, to be automatically released when this function
Christopher Lamb	303dae9	2007-12-17 01:00:21 +0000	[diff] [blame]	3427	returns to its caller. The object is always allocated in the generic address
				3428	space (address space zero).</p>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3429
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3430	<h5>Arguments:</h5>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3431
John Criswell	9e2485c	2004-12-10 15:51:16 +0000	[diff] [blame]	3432	<p>The '<tt>alloca</tt>' instruction allocates <tt>sizeof(<type>)*NumElements</tt>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3433	bytes of memory on the runtime stack, returning a pointer of the
Gabor Greif	1acd2ee	2008-02-09 22:24:34 +0000	[diff] [blame]	3434	appropriate type to the program. If "NumElements" is specified, it is the
				3435	number of elements allocated, otherwise "NumElements" is defaulted to be one.
Chris Lattner	4316dec	2008-04-02 00:38:26 +0000	[diff] [blame]	3436	If a constant alignment is specified, the value result of the allocation is guaranteed
Gabor Greif	1acd2ee	2008-02-09 22:24:34 +0000	[diff] [blame]	3437	to be aligned to at least that boundary. If not specified, or if zero, the target
				3438	can choose to align the allocation on any convenient boundary.</p>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3439
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3440	<p>'<tt>type</tt>' may be any sized type.</p>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3441
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3442	<h5>Semantics:</h5>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3443
Chris Lattner	72ed200	2008-04-19 21:01:16 +0000	[diff] [blame]	3444	<p>Memory is allocated; a pointer is returned. The operation is undefiend if
				3445	there is insufficient stack space for the allocation. '<tt>alloca</tt>'d
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3446	memory is automatically released when the function returns. The '<tt>alloca</tt>'
				3447	instruction is commonly used to represent automatic variables that must
				3448	have an address available. When the function returns (either with the <tt><a
John Criswell	dae2e93	2005-05-12 16:55:34 +0000	[diff] [blame]	3449	href="#i_ret">ret</a></tt> or <tt><a href="#i_unwind">unwind</a></tt>
Chris Lattner	4316dec	2008-04-02 00:38:26 +0000	[diff] [blame]	3450	instructions), the memory is reclaimed. Allocating zero bytes
				3451	is legal, but the result is undefined.</p>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3452
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3453	<h5>Example:</h5>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3454
				3455	<pre>
Dan Gohman	81e2167	2009-01-04 23:49:44 +0000	[diff] [blame]	3456	%ptr = alloca i32 <i>; yields {i32*}:ptr</i>
				3457	%ptr = alloca i32, i32 4 <i>; yields {i32*}:ptr</i>
				3458	%ptr = alloca i32, i32 4, align 1024 <i>; yields {i32*}:ptr</i>
				3459	%ptr = alloca i32, align 1024 <i>; yields {i32*}:ptr</i>
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3460	</pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3461	</div>
Chris Lattner	2cbdc45	2005-11-06 08:02:57 +0000	[diff] [blame]	3462
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3463	<!-- _______________________________________________________________________ -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3464	<div class="doc_subsubsection"> <a name="i_load">'<tt>load</tt>'
				3465	Instruction</a> </div>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3466	<div class="doc_text">
Chris Lattner	2b7d320	2002-05-06 03:03:22 +0000	[diff] [blame]	3467	<h5>Syntax:</h5>
Christopher Lamb	2330e4d	2007-04-21 08:16:25 +0000	[diff] [blame]	3468	<pre> <result> = load <ty>* <pointer>[, align <alignment>]<br> <result> = volatile load <ty>* <pointer>[, align <alignment>]<br></pre>
Chris Lattner	2b7d320	2002-05-06 03:03:22 +0000	[diff] [blame]	3469	<h5>Overview:</h5>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3470	<p>The '<tt>load</tt>' instruction is used to read from memory.</p>
Chris Lattner	2b7d320	2002-05-06 03:03:22 +0000	[diff] [blame]	3471	<h5>Arguments:</h5>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3472	<p>The argument to the '<tt>load</tt>' instruction specifies the memory
John Criswell	0ec250c	2005-10-24 16:17:18 +0000	[diff] [blame]	3473	address from which to load. The pointer must point to a <a
Chris Lattner	e53e508	2004-06-03 22:57:15 +0000	[diff] [blame]	3474	href="#t_firstclass">first class</a> type. If the <tt>load</tt> is
John Criswell	0ec250c	2005-10-24 16:17:18 +0000	[diff] [blame]	3475	marked as <tt>volatile</tt>, then the optimizer is not allowed to modify
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3476	the number or order of execution of this <tt>load</tt> with other
				3477	volatile <tt>load</tt> and <tt><a href="#i_store">store</a></tt>
				3478	instructions. </p>
Chris Lattner	a31d1d7	2008-01-06 21:04:43 +0000	[diff] [blame]	3479	<p>
Chris Lattner	4316dec	2008-04-02 00:38:26 +0000	[diff] [blame]	3480	The optional constant "align" argument specifies the alignment of the operation
Chris Lattner	a31d1d7	2008-01-06 21:04:43 +0000	[diff] [blame]	3481	(that is, the alignment of the memory address). A value of 0 or an
				3482	omitted "align" argument means that the operation has the preferential
				3483	alignment for the target. It is the responsibility of the code emitter
				3484	to ensure that the alignment information is correct. Overestimating
				3485	the alignment results in an undefined behavior. Underestimating the
				3486	alignment may produce less efficient code. An alignment of 1 is always
				3487	safe.
				3488	</p>
Chris Lattner	2b7d320	2002-05-06 03:03:22 +0000	[diff] [blame]	3489	<h5>Semantics:</h5>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3490	<p>The location of memory pointed to is loaded.</p>
Chris Lattner	2b7d320	2002-05-06 03:03:22 +0000	[diff] [blame]	3491	<h5>Examples:</h5>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3492	<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3493	<a
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3494	href="#i_store">store</a> i32 3, i32* %ptr <i>; yields {void}</i>
				3495	%val = load i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner	2b7d320	2002-05-06 03:03:22 +0000	[diff] [blame]	3496	</pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3497	</div>
Chris Lattner	2b7d320	2002-05-06 03:03:22 +0000	[diff] [blame]	3498	<!-- _______________________________________________________________________ -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3499	<div class="doc_subsubsection"> <a name="i_store">'<tt>store</tt>'
				3500	Instruction</a> </div>
Reid Spencer	035ab57	2006-11-09 21:18:01 +0000	[diff] [blame]	3501	<div class="doc_text">
Chris Lattner	2b7d320	2002-05-06 03:03:22 +0000	[diff] [blame]	3502	<h5>Syntax:</h5>
Christopher Lamb	2330e4d	2007-04-21 08:16:25 +0000	[diff] [blame]	3503	<pre> store <ty> <value>, <ty>* <pointer>[, align <alignment>] <i>; yields {void}</i>
				3504	volatile store <ty> <value>, <ty>* <pointer>[, align <alignment>] <i>; yields {void}</i>
Chris Lattner	2b7d320	2002-05-06 03:03:22 +0000	[diff] [blame]	3505	</pre>
Chris Lattner	2b7d320	2002-05-06 03:03:22 +0000	[diff] [blame]	3506	<h5>Overview:</h5>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3507	<p>The '<tt>store</tt>' instruction is used to write to memory.</p>
Chris Lattner	2b7d320	2002-05-06 03:03:22 +0000	[diff] [blame]	3508	<h5>Arguments:</h5>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3509	<p>There are two arguments to the '<tt>store</tt>' instruction: a value
Jeff Cohen	6f1cc77	2007-04-22 01:17:39 +0000	[diff] [blame]	3510	to store and an address at which to store it. The type of the '<tt><pointer></tt>'
Chris Lattner	4316dec	2008-04-02 00:38:26 +0000	[diff] [blame]	3511	operand must be a pointer to the <a href="#t_firstclass">first class</a> type
				3512	of the '<tt><value></tt>'
John Criswell	c1f786c	2005-05-13 22:25:59 +0000	[diff] [blame]	3513	operand. If the <tt>store</tt> is marked as <tt>volatile</tt>, then the
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3514	optimizer is not allowed to modify the number or order of execution of
				3515	this <tt>store</tt> with other volatile <tt>load</tt> and <tt><a
				3516	href="#i_store">store</a></tt> instructions.</p>
Chris Lattner	a31d1d7	2008-01-06 21:04:43 +0000	[diff] [blame]	3517	<p>
Chris Lattner	4316dec	2008-04-02 00:38:26 +0000	[diff] [blame]	3518	The optional constant "align" argument specifies the alignment of the operation
Chris Lattner	a31d1d7	2008-01-06 21:04:43 +0000	[diff] [blame]	3519	(that is, the alignment of the memory address). A value of 0 or an
				3520	omitted "align" argument means that the operation has the preferential
				3521	alignment for the target. It is the responsibility of the code emitter
				3522	to ensure that the alignment information is correct. Overestimating
				3523	the alignment results in an undefined behavior. Underestimating the
				3524	alignment may produce less efficient code. An alignment of 1 is always
				3525	safe.
				3526	</p>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3527	<h5>Semantics:</h5>
				3528	<p>The contents of memory are updated to contain '<tt><value></tt>'
				3529	at the location specified by the '<tt><pointer></tt>' operand.</p>
Chris Lattner	2b7d320	2002-05-06 03:03:22 +0000	[diff] [blame]	3530	<h5>Example:</h5>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3531	<pre> %ptr = <a href="#i_alloca">alloca</a> i32 <i>; yields {i32*}:ptr</i>
Bill Wendling	8c6c72d	2007-10-22 05:10:05 +0000	[diff] [blame]	3532	store i32 3, i32* %ptr <i>; yields {void}</i>
				3533	%val = <a href="#i_load">load</a> i32* %ptr <i>; yields {i32}:val = i32 3</i>
Chris Lattner	2b7d320	2002-05-06 03:03:22 +0000	[diff] [blame]	3534	</pre>
Reid Spencer	47ce179	2006-11-09 21:15:49 +0000	[diff] [blame]	3535	</div>
				3536
Chris Lattner	2b7d320	2002-05-06 03:03:22 +0000	[diff] [blame]	3537	<!-- _______________________________________________________________________ -->
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3538	<div class="doc_subsubsection">
				3539	<a name="i_getelementptr">'<tt>getelementptr</tt>' Instruction</a>
				3540	</div>
				3541
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3542	<div class="doc_text">
Chris Lattner	7faa883	2002-04-14 06:13:44 +0000	[diff] [blame]	3543	<h5>Syntax:</h5>
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3544	<pre>
Matthijs Kooijman	e49d0bc	2008-10-13 13:44:15 +0000	[diff] [blame]	3545	<result> = getelementptr <pty>* <ptrval>{, <ty> <idx>}*
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3546	</pre>
				3547
Chris Lattner	7faa883	2002-04-14 06:13:44 +0000	[diff] [blame]	3548	<h5>Overview:</h5>
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3549
				3550	<p>
				3551	The '<tt>getelementptr</tt>' instruction is used to get the address of a
Matthijs Kooijman	e49d0bc	2008-10-13 13:44:15 +0000	[diff] [blame]	3552	subelement of an aggregate data structure. It performs address calculation only
				3553	and does not access memory.</p>
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3554
Chris Lattner	7faa883	2002-04-14 06:13:44 +0000	[diff] [blame]	3555	<h5>Arguments:</h5>
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3556
Matthijs Kooijman	e49d0bc	2008-10-13 13:44:15 +0000	[diff] [blame]	3557	<p>The first argument is always a pointer, and forms the basis of the
				3558	calculation. The remaining arguments are indices, that indicate which of the
				3559	elements of the aggregate object are indexed. The interpretation of each index
				3560	is dependent on the type being indexed into. The first index always indexes the
				3561	pointer value given as the first argument, the second index indexes a value of
				3562	the type pointed to (not necessarily the value directly pointed to, since the
				3563	first index can be non-zero), etc. The first type indexed into must be a pointer
				3564	value, subsequent types can be arrays, vectors and structs. Note that subsequent
				3565	types being indexed into can never be pointers, since that would require loading
				3566	the pointer before continuing calculation.</p>
				3567
				3568	<p>The type of each index argument depends on the type it is indexing into.
				3569	When indexing into a (packed) structure, only <tt>i32</tt> integer
				3570	<b>constants</b> are allowed. When indexing into an array, pointer or vector,
				3571	only integers of 32 or 64 bits are allowed (also non-constants). 32-bit values
				3572	will be sign extended to 64-bits if required.</p>
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3573
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3574	<p>For example, let's consider a C code fragment and how it gets
				3575	compiled to LLVM:</p>
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3576
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	3577	<div class="doc_code">
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3578	<pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	3579	struct RT {
				3580	char A;
Chris Lattner	cabc846	2007-05-29 15:43:56 +0000	[diff] [blame]	3581	int B[10][20];
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	3582	char C;
				3583	};
				3584	struct ST {
Chris Lattner	cabc846	2007-05-29 15:43:56 +0000	[diff] [blame]	3585	int X;
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	3586	double Y;
				3587	struct RT Z;
				3588	};
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3589
Chris Lattner	cabc846	2007-05-29 15:43:56 +0000	[diff] [blame]	3590	int foo(struct ST s) {
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	3591	return &s[1].Z.B[5][13];
				3592	}
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3593	</pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	3594	</div>
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3595
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3596	<p>The LLVM code generated by the GCC frontend is:</p>
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3597
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	3598	<div class="doc_code">
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3599	<pre>
Chris Lattner	e7886e4	2009-01-11 20:53:49 +0000	[diff] [blame]	3600	%RT = <a href="#namedtypes">type</a> { i8 , [10 x [20 x i32]], i8 }
				3601	%ST = <a href="#namedtypes">type</a> { i32, double, %RT }
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3602
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	3603	define i32* %foo(%ST* %s) {
				3604	entry:
				3605	%reg = getelementptr %ST* %s, i32 1, i32 2, i32 1, i32 5, i32 13
				3606	ret i32* %reg
				3607	}
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3608	</pre>
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	3609	</div>
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3610
Chris Lattner	7faa883	2002-04-14 06:13:44 +0000	[diff] [blame]	3611	<h5>Semantics:</h5>
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3612
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3613	<p>In the example above, the first index is indexing into the '<tt>%ST*</tt>'
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3614	type, which is a pointer, yielding a '<tt>%ST</tt>' = '<tt>{ i32, double, %RT
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3615	}</tt>' type, a structure. The second index indexes into the third element of
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3616	the structure, yielding a '<tt>%RT</tt>' = '<tt>{ i8 , [10 x [20 x i32]],
				3617	i8 }</tt>' type, another structure. The third index indexes into the second
				3618	element of the structure, yielding a '<tt>[10 x [20 x i32]]</tt>' type, an
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3619	array. The two dimensions of the array are subscripted into, yielding an
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3620	'<tt>i32</tt>' type. The '<tt>getelementptr</tt>' instruction returns a pointer
				3621	to this element, thus computing a value of '<tt>i32*</tt>' type.</p>
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3622
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	3623	<p>Note that it is perfectly legal to index partially through a
				3624	structure, returning a pointer to an inner element. Because of this,
				3625	the LLVM code for the given testcase is equivalent to:</p>
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3626
				3627	<pre>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3628	define i32* %foo(%ST* %s) {
				3629	%t1 = getelementptr %ST* %s, i32 1 <i>; yields %ST*:%t1</i>
Jeff Cohen	6f1cc77	2007-04-22 01:17:39 +0000	[diff] [blame]	3630	%t2 = getelementptr %ST* %t1, i32 0, i32 2 <i>; yields %RT*:%t2</i>
				3631	%t3 = getelementptr %RT* %t2, i32 0, i32 1 <i>; yields [10 x [20 x i32]]*:%t3</i>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3632	%t4 = getelementptr [10 x [20 x i32]]* %t3, i32 0, i32 5 <i>; yields [20 x i32]*:%t4</i>
				3633	%t5 = getelementptr [20 x i32]* %t4, i32 0, i32 13 <i>; yields i32*:%t5</i>
				3634	ret i32* %t5
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3635	}
Chris Lattner	6536cfe	2002-05-06 22:08:29 +0000	[diff] [blame]	3636	</pre>
Chris Lattner	e67a951	2005-06-24 17:22:57 +0000	[diff] [blame]	3637
Chris Lattner	8c0e62c	2009-03-09 20:55:18 +0000	[diff] [blame]	3638	<p>Note that it is undefined to access an array out of bounds: array
				3639	and pointer indexes must always be within the defined bounds of the
				3640	array type when accessed with an instruction that dereferences the
				3641	pointer (e.g. a load or store instruction). The one exception for
				3642	this rule is zero length arrays. These arrays are defined to be
				3643	accessible as variable length arrays, which requires access beyond the
				3644	zero'th element.</p>
Chris Lattner	e67a951	2005-06-24 17:22:57 +0000	[diff] [blame]	3645
Chris Lattner	884a970	2006-08-15 00:45:58 +0000	[diff] [blame]	3646	<p>The getelementptr instruction is often confusing. For some more insight
				3647	into how it works, see <a href="GetElementPtr.html">the getelementptr
				3648	FAQ</a>.</p>
				3649
Chris Lattner	7faa883	2002-04-14 06:13:44 +0000	[diff] [blame]	3650	<h5>Example:</h5>
Chris Lattner	e67a951	2005-06-24 17:22:57 +0000	[diff] [blame]	3651
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3652	<pre>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3653	<i>; yields [12 x i8]*:aptr</i>
Matthijs Kooijman	e49d0bc	2008-10-13 13:44:15 +0000	[diff] [blame]	3654	%aptr = getelementptr {i32, [12 x i8]}* %saptr, i64 0, i32 1
				3655	<i>; yields i8*:vptr</i>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	3656	%vptr = getelementptr {i32, <2 x i8>}* %svptr, i64 0, i32 1, i32 1
Matthijs Kooijman	e49d0bc	2008-10-13 13:44:15 +0000	[diff] [blame]	3657	<i>; yields i8*:eptr</i>
				3658	%eptr = getelementptr [12 x i8]* %aptr, i64 0, i32 1
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3659	</pre>
Chris Lattner	f74d5c7	2004-04-05 01:30:49 +0000	[diff] [blame]	3660	</div>
Reid Spencer	47ce179	2006-11-09 21:15:49 +0000	[diff] [blame]	3661
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	3662	<!-- ======================================================================= -->
Reid Spencer	2fd21e6	2006-11-08 01:18:52 +0000	[diff] [blame]	3663	<div class="doc_subsection"> <a name="convertops">Conversion Operations</a>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3664	</div>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3665	<div class="doc_text">
Reid Spencer	2fd21e6	2006-11-08 01:18:52 +0000	[diff] [blame]	3666	<p>The instructions in this category are the conversion instructions (casting)
				3667	which all take a single operand and a type. They perform various bit conversions
				3668	on the operand.</p>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	3669	</div>
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	3670
Chris Lattner	6536cfe	2002-05-06 22:08:29 +0000	[diff] [blame]	3671	<!-- _______________________________________________________________________ -->
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	3672	<div class="doc_subsubsection">
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3673	<a name="i_trunc">'<tt>trunc .. to</tt>' Instruction</a>
				3674	</div>
				3675	<div class="doc_text">
				3676
				3677	<h5>Syntax:</h5>
				3678	<pre>
				3679	<result> = trunc <ty> <value> to <ty2> <i>; yields ty2</i>
				3680	</pre>
				3681
				3682	<h5>Overview:</h5>
				3683	<p>
				3684	The '<tt>trunc</tt>' instruction truncates its operand to the type <tt>ty2</tt>.
				3685	</p>
				3686
				3687	<h5>Arguments:</h5>
				3688	<p>
				3689	The '<tt>trunc</tt>' instruction takes a <tt>value</tt> to trunc, which must
				3690	be an <a href="#t_integer">integer</a> type, and a type that specifies the size
Chris Lattner	3b19d65	2007-01-15 01:54:13 +0000	[diff] [blame]	3691	and type of the result, which must be an <a href="#t_integer">integer</a>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3692	type. The bit size of <tt>value</tt> must be larger than the bit size of
				3693	<tt>ty2</tt>. Equal sized types are not allowed.</p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3694
				3695	<h5>Semantics:</h5>
				3696	<p>
				3697	The '<tt>trunc</tt>' instruction truncates the high order bits in <tt>value</tt>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3698	and converts the remaining bits to <tt>ty2</tt>. Since the source size must be
				3699	larger than the destination size, <tt>trunc</tt> cannot be a <i>no-op cast</i>.
				3700	It will always truncate bits.</p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3701
				3702	<h5>Example:</h5>
				3703	<pre>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3704	%X = trunc i32 257 to i8 <i>; yields i8:1</i>
Reid Spencer	c78f337	2007-01-12 03:35:51 +0000	[diff] [blame]	3705	%Y = trunc i32 123 to i1 <i>; yields i1:true</i>
				3706	%Y = trunc i32 122 to i1 <i>; yields i1:false</i>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3707	</pre>
				3708	</div>
				3709
				3710	<!-- _______________________________________________________________________ -->
				3711	<div class="doc_subsubsection">
				3712	<a name="i_zext">'<tt>zext .. to</tt>' Instruction</a>
				3713	</div>
				3714	<div class="doc_text">
				3715
				3716	<h5>Syntax:</h5>
				3717	<pre>
				3718	<result> = zext <ty> <value> to <ty2> <i>; yields ty2</i>
				3719	</pre>
				3720
				3721	<h5>Overview:</h5>
				3722	<p>The '<tt>zext</tt>' instruction zero extends its operand to type
				3723	<tt>ty2</tt>.</p>
				3724
				3725
				3726	<h5>Arguments:</h5>
				3727	<p>The '<tt>zext</tt>' instruction takes a value to cast, which must be of
Chris Lattner	3b19d65	2007-01-15 01:54:13 +0000	[diff] [blame]	3728	<a href="#t_integer">integer</a> type, and a type to cast it to, which must
				3729	also be of <a href="#t_integer">integer</a> type. The bit size of the
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3730	<tt>value</tt> must be smaller than the bit size of the destination type,
				3731	<tt>ty2</tt>.</p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3732
				3733	<h5>Semantics:</h5>
				3734	<p>The <tt>zext</tt> fills the high order bits of the <tt>value</tt> with zero
Chris Lattner	d1d2517	2007-05-24 19:13:27 +0000	[diff] [blame]	3735	bits until it reaches the size of the destination type, <tt>ty2</tt>.</p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3736
Reid Spencer	b592952	2007-01-12 15:46:11 +0000	[diff] [blame]	3737	<p>When zero extending from i1, the result will always be either 0 or 1.</p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3738
				3739	<h5>Example:</h5>
				3740	<pre>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3741	%X = zext i32 257 to i64 <i>; yields i64:257</i>
Reid Spencer	c78f337	2007-01-12 03:35:51 +0000	[diff] [blame]	3742	%Y = zext i1 true to i32 <i>; yields i32:1</i>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3743	</pre>
				3744	</div>
				3745
				3746	<!-- _______________________________________________________________________ -->
				3747	<div class="doc_subsubsection">
				3748	<a name="i_sext">'<tt>sext .. to</tt>' Instruction</a>
				3749	</div>
				3750	<div class="doc_text">
				3751
				3752	<h5>Syntax:</h5>
				3753	<pre>
				3754	<result> = sext <ty> <value> to <ty2> <i>; yields ty2</i>
				3755	</pre>
				3756
				3757	<h5>Overview:</h5>
				3758	<p>The '<tt>sext</tt>' sign extends <tt>value</tt> to the type <tt>ty2</tt>.</p>
				3759
				3760	<h5>Arguments:</h5>
				3761	<p>
				3762	The '<tt>sext</tt>' instruction takes a value to cast, which must be of
Chris Lattner	3b19d65	2007-01-15 01:54:13 +0000	[diff] [blame]	3763	<a href="#t_integer">integer</a> type, and a type to cast it to, which must
				3764	also be of <a href="#t_integer">integer</a> type. The bit size of the
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3765	<tt>value</tt> must be smaller than the bit size of the destination type,
				3766	<tt>ty2</tt>.</p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3767
				3768	<h5>Semantics:</h5>
				3769	<p>
				3770	The '<tt>sext</tt>' instruction performs a sign extension by copying the sign
				3771	bit (highest order bit) of the <tt>value</tt> until it reaches the bit size of
Chris Lattner	d1d2517	2007-05-24 19:13:27 +0000	[diff] [blame]	3772	the type <tt>ty2</tt>.</p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3773
Reid Spencer	c78f337	2007-01-12 03:35:51 +0000	[diff] [blame]	3774	<p>When sign extending from i1, the extension always results in -1 or 0.</p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3775
				3776	<h5>Example:</h5>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3777	<pre>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3778	%X = sext i8 -1 to i16 <i>; yields i16 :65535</i>
Reid Spencer	c78f337	2007-01-12 03:35:51 +0000	[diff] [blame]	3779	%Y = sext i1 true to i32 <i>; yields i32:-1</i>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3780	</pre>
				3781	</div>
				3782
				3783	<!-- _______________________________________________________________________ -->
				3784	<div class="doc_subsubsection">
Reid Spencer	3fa91b0	2006-11-09 21:48:10 +0000	[diff] [blame]	3785	<a name="i_fptrunc">'<tt>fptrunc .. to</tt>' Instruction</a>
				3786	</div>
				3787
				3788	<div class="doc_text">
				3789
				3790	<h5>Syntax:</h5>
				3791
				3792	<pre>
				3793	<result> = fptrunc <ty> <value> to <ty2> <i>; yields ty2</i>
				3794	</pre>
				3795
				3796	<h5>Overview:</h5>
				3797	<p>The '<tt>fptrunc</tt>' instruction truncates <tt>value</tt> to type
				3798	<tt>ty2</tt>.</p>
				3799
				3800
				3801	<h5>Arguments:</h5>
				3802	<p>The '<tt>fptrunc</tt>' instruction takes a <a href="#t_floating">floating
				3803	point</a> value to cast and a <a href="#t_floating">floating point</a> type to
				3804	cast it to. The size of <tt>value</tt> must be larger than the size of
				3805	<tt>ty2</tt>. This implies that <tt>fptrunc</tt> cannot be used to make a
				3806	<i>no-op cast</i>.</p>
				3807
				3808	<h5>Semantics:</h5>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3809	<p> The '<tt>fptrunc</tt>' instruction truncates a <tt>value</tt> from a larger
				3810	<a href="#t_floating">floating point</a> type to a smaller
				3811	<a href="#t_floating">floating point</a> type. If the value cannot fit within
				3812	the destination type, <tt>ty2</tt>, then the results are undefined.</p>
Reid Spencer	3fa91b0	2006-11-09 21:48:10 +0000	[diff] [blame]	3813
				3814	<h5>Example:</h5>
				3815	<pre>
				3816	%X = fptrunc double 123.0 to float <i>; yields float:123.0</i>
				3817	%Y = fptrunc double 1.0E+300 to float <i>; yields undefined</i>
				3818	</pre>
				3819	</div>
				3820
				3821	<!-- _______________________________________________________________________ -->
				3822	<div class="doc_subsubsection">
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3823	<a name="i_fpext">'<tt>fpext .. to</tt>' Instruction</a>
				3824	</div>
				3825	<div class="doc_text">
				3826
				3827	<h5>Syntax:</h5>
				3828	<pre>
				3829	<result> = fpext <ty> <value> to <ty2> <i>; yields ty2</i>
				3830	</pre>
				3831
				3832	<h5>Overview:</h5>
				3833	<p>The '<tt>fpext</tt>' extends a floating point <tt>value</tt> to a larger
				3834	floating point value.</p>
				3835
				3836	<h5>Arguments:</h5>
				3837	<p>The '<tt>fpext</tt>' instruction takes a
				3838	<a href="#t_floating">floating point</a> <tt>value</tt> to cast,
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3839	and a <a href="#t_floating">floating point</a> type to cast it to. The source
				3840	type must be smaller than the destination type.</p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3841
				3842	<h5>Semantics:</h5>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3843	<p>The '<tt>fpext</tt>' instruction extends the <tt>value</tt> from a smaller
Duncan Sands	8036ca4	2007-03-30 12:22:09 +0000	[diff] [blame]	3844	<a href="#t_floating">floating point</a> type to a larger
				3845	<a href="#t_floating">floating point</a> type. The <tt>fpext</tt> cannot be
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3846	used to make a <i>no-op cast</i> because it always changes bits. Use
Reid Spencer	5c0ef47	2006-11-11 23:08:07 +0000	[diff] [blame]	3847	<tt>bitcast</tt> to make a <i>no-op cast</i> for a floating point cast.</p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3848
				3849	<h5>Example:</h5>
				3850	<pre>
				3851	%X = fpext float 3.1415 to double <i>; yields double:3.1415</i>
				3852	%Y = fpext float 1.0 to float <i>; yields float:1.0 (no-op)</i>
				3853	</pre>
				3854	</div>
				3855
				3856	<!-- _______________________________________________________________________ -->
				3857	<div class="doc_subsubsection">
Reid Spencer	24d6da5	2007-01-21 00:29:26 +0000	[diff] [blame]	3858	<a name="i_fptoui">'<tt>fptoui .. to</tt>' Instruction</a>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3859	</div>
				3860	<div class="doc_text">
				3861
				3862	<h5>Syntax:</h5>
				3863	<pre>
Reid Spencer	1539a1c	2007-07-31 14:40:14 +0000	[diff] [blame]	3864	<result> = fptoui <ty> <value> to <ty2> <i>; yields ty2</i>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3865	</pre>
				3866
				3867	<h5>Overview:</h5>
Reid Spencer	1539a1c	2007-07-31 14:40:14 +0000	[diff] [blame]	3868	<p>The '<tt>fptoui</tt>' converts a floating point <tt>value</tt> to its
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3869	unsigned integer equivalent of type <tt>ty2</tt>.
				3870	</p>
				3871
				3872	<h5>Arguments:</h5>
Reid Spencer	1539a1c	2007-07-31 14:40:14 +0000	[diff] [blame]	3873	<p>The '<tt>fptoui</tt>' instruction takes a value to cast, which must be a
Nate Begeman	b348d18	2007-11-17 03:58:34 +0000	[diff] [blame]	3874	scalar or vector <a href="#t_floating">floating point</a> value, and a type
				3875	to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
				3876	type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
				3877	vector integer type with the same number of elements as <tt>ty</tt></p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3878
				3879	<h5>Semantics:</h5>
Reid Spencer	1539a1c	2007-07-31 14:40:14 +0000	[diff] [blame]	3880	<p> The '<tt>fptoui</tt>' instruction converts its
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3881	<a href="#t_floating">floating point</a> operand into the nearest (rounding
				3882	towards zero) unsigned integer value. If the value cannot fit in <tt>ty2</tt>,
				3883	the results are undefined.</p>
				3884
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3885	<h5>Example:</h5>
				3886	<pre>
Reid Spencer	1539a1c	2007-07-31 14:40:14 +0000	[diff] [blame]	3887	%X = fptoui double 123.0 to i32 <i>; yields i32:123</i>
Chris Lattner	8851904	2007-09-22 03:17:52 +0000	[diff] [blame]	3888	%Y = fptoui float 1.0E+300 to i1 <i>; yields undefined:1</i>
Reid Spencer	1539a1c	2007-07-31 14:40:14 +0000	[diff] [blame]	3889	%X = fptoui float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3890	</pre>
				3891	</div>
				3892
				3893	<!-- _______________________________________________________________________ -->
				3894	<div class="doc_subsubsection">
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3895	<a name="i_fptosi">'<tt>fptosi .. to</tt>' Instruction</a>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3896	</div>
				3897	<div class="doc_text">
				3898
				3899	<h5>Syntax:</h5>
				3900	<pre>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3901	<result> = fptosi <ty> <value> to <ty2> <i>; yields ty2</i>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3902	</pre>
				3903
				3904	<h5>Overview:</h5>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3905	<p>The '<tt>fptosi</tt>' instruction converts
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3906	<a href="#t_floating">floating point</a> <tt>value</tt> to type <tt>ty2</tt>.
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	3907	</p>
				3908
Chris Lattner	6536cfe	2002-05-06 22:08:29 +0000	[diff] [blame]	3909	<h5>Arguments:</h5>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3910	<p> The '<tt>fptosi</tt>' instruction takes a value to cast, which must be a
Nate Begeman	b348d18	2007-11-17 03:58:34 +0000	[diff] [blame]	3911	scalar or vector <a href="#t_floating">floating point</a> value, and a type
				3912	to cast it to <tt>ty2</tt>, which must be an <a href="#t_integer">integer</a>
				3913	type. If <tt>ty</tt> is a vector floating point type, <tt>ty2</tt> must be a
				3914	vector integer type with the same number of elements as <tt>ty</tt></p>
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	3915
Chris Lattner	6536cfe	2002-05-06 22:08:29 +0000	[diff] [blame]	3916	<h5>Semantics:</h5>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3917	<p>The '<tt>fptosi</tt>' instruction converts its
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3918	<a href="#t_floating">floating point</a> operand into the nearest (rounding
				3919	towards zero) signed integer value. If the value cannot fit in <tt>ty2</tt>,
				3920	the results are undefined.</p>
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	3921
Chris Lattner	33ba0d9	2001-07-09 00:26:23 +0000	[diff] [blame]	3922	<h5>Example:</h5>
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	3923	<pre>
Reid Spencer	c78f337	2007-01-12 03:35:51 +0000	[diff] [blame]	3924	%X = fptosi double -123.0 to i32 <i>; yields i32:-123</i>
Chris Lattner	8851904	2007-09-22 03:17:52 +0000	[diff] [blame]	3925	%Y = fptosi float 1.0E-247 to i1 <i>; yields undefined:1</i>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3926	%X = fptosi float 1.04E+17 to i8 <i>; yields undefined:1</i>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3927	</pre>
				3928	</div>
				3929
				3930	<!-- _______________________________________________________________________ -->
				3931	<div class="doc_subsubsection">
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3932	<a name="i_uitofp">'<tt>uitofp .. to</tt>' Instruction</a>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3933	</div>
				3934	<div class="doc_text">
				3935
				3936	<h5>Syntax:</h5>
				3937	<pre>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3938	<result> = uitofp <ty> <value> to <ty2> <i>; yields ty2</i>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3939	</pre>
				3940
				3941	<h5>Overview:</h5>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3942	<p>The '<tt>uitofp</tt>' instruction regards <tt>value</tt> as an unsigned
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3943	integer and converts that value to the <tt>ty2</tt> type.</p>
				3944
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3945	<h5>Arguments:</h5>
Nate Begeman	b348d18	2007-11-17 03:58:34 +0000	[diff] [blame]	3946	<p>The '<tt>uitofp</tt>' instruction takes a value to cast, which must be a
				3947	scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
				3948	to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
				3949	type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
				3950	floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3951
				3952	<h5>Semantics:</h5>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3953	<p>The '<tt>uitofp</tt>' instruction interprets its operand as an unsigned
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3954	integer quantity and converts it to the corresponding floating point value. If
Jeff Cohen	cb75731	2007-04-22 14:56:37 +0000	[diff] [blame]	3955	the value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3956
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3957	<h5>Example:</h5>
				3958	<pre>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3959	%X = uitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	3960	%Y = uitofp i8 -1 to double <i>; yields double:255.0</i>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3961	</pre>
				3962	</div>
				3963
				3964	<!-- _______________________________________________________________________ -->
				3965	<div class="doc_subsubsection">
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3966	<a name="i_sitofp">'<tt>sitofp .. to</tt>' Instruction</a>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3967	</div>
				3968	<div class="doc_text">
				3969
				3970	<h5>Syntax:</h5>
				3971	<pre>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3972	<result> = sitofp <ty> <value> to <ty2> <i>; yields ty2</i>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3973	</pre>
				3974
				3975	<h5>Overview:</h5>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3976	<p>The '<tt>sitofp</tt>' instruction regards <tt>value</tt> as a signed
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3977	integer and converts that value to the <tt>ty2</tt> type.</p>
				3978
				3979	<h5>Arguments:</h5>
Nate Begeman	b348d18	2007-11-17 03:58:34 +0000	[diff] [blame]	3980	<p>The '<tt>sitofp</tt>' instruction takes a value to cast, which must be a
				3981	scalar or vector <a href="#t_integer">integer</a> value, and a type to cast it
				3982	to <tt>ty2</tt>, which must be an <a href="#t_floating">floating point</a>
				3983	type. If <tt>ty</tt> is a vector integer type, <tt>ty2</tt> must be a vector
				3984	floating point type with the same number of elements as <tt>ty</tt></p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3985
				3986	<h5>Semantics:</h5>
Reid Spencer	d444879	2006-11-09 23:03:26 +0000	[diff] [blame]	3987	<p>The '<tt>sitofp</tt>' instruction interprets its operand as a signed
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3988	integer quantity and converts it to the corresponding floating point value. If
Jeff Cohen	cb75731	2007-04-22 14:56:37 +0000	[diff] [blame]	3989	the value cannot fit in the floating point value, the results are undefined.</p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3990
				3991	<h5>Example:</h5>
				3992	<pre>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	3993	%X = sitofp i32 257 to float <i>; yields float:257.0</i>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	3994	%Y = sitofp i8 -1 to double <i>; yields double:-1.0</i>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	3995	</pre>
				3996	</div>
				3997
				3998	<!-- _______________________________________________________________________ -->
				3999	<div class="doc_subsubsection">
Reid Spencer	7267925	2006-11-11 21:00:47 +0000	[diff] [blame]	4000	<a name="i_ptrtoint">'<tt>ptrtoint .. to</tt>' Instruction</a>
				4001	</div>
				4002	<div class="doc_text">
				4003
				4004	<h5>Syntax:</h5>
				4005	<pre>
				4006	<result> = ptrtoint <ty> <value> to <ty2> <i>; yields ty2</i>
				4007	</pre>
				4008
				4009	<h5>Overview:</h5>
				4010	<p>The '<tt>ptrtoint</tt>' instruction converts the pointer <tt>value</tt> to
				4011	the integer type <tt>ty2</tt>.</p>
				4012
				4013	<h5>Arguments:</h5>
				4014	<p>The '<tt>ptrtoint</tt>' instruction takes a <tt>value</tt> to cast, which
Duncan Sands	8036ca4	2007-03-30 12:22:09 +0000	[diff] [blame]	4015	must be a <a href="#t_pointer">pointer</a> value, and a type to cast it to
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4016	<tt>ty2</tt>, which must be an <a href="#t_integer">integer</a> type.</p>
Reid Spencer	7267925	2006-11-11 21:00:47 +0000	[diff] [blame]	4017
				4018	<h5>Semantics:</h5>
				4019	<p>The '<tt>ptrtoint</tt>' instruction converts <tt>value</tt> to integer type
				4020	<tt>ty2</tt> by interpreting the pointer value as an integer and either
				4021	truncating or zero extending that value to the size of the integer type. If
				4022	<tt>value</tt> is smaller than <tt>ty2</tt> then a zero extension is done. If
				4023	<tt>value</tt> is larger than <tt>ty2</tt> then a truncation is done. If they
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4024	are the same size, then nothing is done (<i>no-op cast</i>) other than a type
				4025	change.</p>
Reid Spencer	7267925	2006-11-11 21:00:47 +0000	[diff] [blame]	4026
				4027	<h5>Example:</h5>
				4028	<pre>
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4029	%X = ptrtoint i32* %X to i8 <i>; yields truncation on 32-bit architecture</i>
				4030	%Y = ptrtoint i32* %x to i64 <i>; yields zero extension on 32-bit architecture</i>
Reid Spencer	7267925	2006-11-11 21:00:47 +0000	[diff] [blame]	4031	</pre>
				4032	</div>
				4033
				4034	<!-- _______________________________________________________________________ -->
				4035	<div class="doc_subsubsection">
				4036	<a name="i_inttoptr">'<tt>inttoptr .. to</tt>' Instruction</a>
				4037	</div>
				4038	<div class="doc_text">
				4039
				4040	<h5>Syntax:</h5>
				4041	<pre>
				4042	<result> = inttoptr <ty> <value> to <ty2> <i>; yields ty2</i>
				4043	</pre>
				4044
				4045	<h5>Overview:</h5>
				4046	<p>The '<tt>inttoptr</tt>' instruction converts an integer <tt>value</tt> to
				4047	a pointer type, <tt>ty2</tt>.</p>
				4048
				4049	<h5>Arguments:</h5>
Duncan Sands	8036ca4	2007-03-30 12:22:09 +0000	[diff] [blame]	4050	<p>The '<tt>inttoptr</tt>' instruction takes an <a href="#t_integer">integer</a>
Reid Spencer	7267925	2006-11-11 21:00:47 +0000	[diff] [blame]	4051	value to cast, and a type to cast it to, which must be a
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4052	<a href="#t_pointer">pointer</a> type.</p>
Reid Spencer	7267925	2006-11-11 21:00:47 +0000	[diff] [blame]	4053
				4054	<h5>Semantics:</h5>
				4055	<p>The '<tt>inttoptr</tt>' instruction converts <tt>value</tt> to type
				4056	<tt>ty2</tt> by applying either a zero extension or a truncation depending on
				4057	the size of the integer <tt>value</tt>. If <tt>value</tt> is larger than the
				4058	size of a pointer then a truncation is done. If <tt>value</tt> is smaller than
				4059	the size of a pointer then a zero extension is done. If they are the same size,
				4060	nothing is done (<i>no-op cast</i>).</p>
				4061
				4062	<h5>Example:</h5>
				4063	<pre>
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4064	%X = inttoptr i32 255 to i32* <i>; yields zero extension on 64-bit architecture</i>
				4065	%X = inttoptr i32 255 to i32* <i>; yields no-op on 32-bit architecture</i>
				4066	%Y = inttoptr i64 0 to i32* <i>; yields truncation on 32-bit architecture</i>
Reid Spencer	7267925	2006-11-11 21:00:47 +0000	[diff] [blame]	4067	</pre>
				4068	</div>
				4069
				4070	<!-- _______________________________________________________________________ -->
				4071	<div class="doc_subsubsection">
Reid Spencer	5c0ef47	2006-11-11 23:08:07 +0000	[diff] [blame]	4072	<a name="i_bitcast">'<tt>bitcast .. to</tt>' Instruction</a>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	4073	</div>
				4074	<div class="doc_text">
				4075
				4076	<h5>Syntax:</h5>
				4077	<pre>
Reid Spencer	5c0ef47	2006-11-11 23:08:07 +0000	[diff] [blame]	4078	<result> = bitcast <ty> <value> to <ty2> <i>; yields ty2</i>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	4079	</pre>
				4080
				4081	<h5>Overview:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4082
Reid Spencer	5c0ef47	2006-11-11 23:08:07 +0000	[diff] [blame]	4083	<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	4084	<tt>ty2</tt> without changing any bits.</p>
				4085
				4086	<h5>Arguments:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4087
Reid Spencer	5c0ef47	2006-11-11 23:08:07 +0000	[diff] [blame]	4088	<p>The '<tt>bitcast</tt>' instruction takes a value to cast, which must be
Dan Gohman	500233a	2008-09-08 16:45:59 +0000	[diff] [blame]	4089	a non-aggregate first class value, and a type to cast it to, which must also be
				4090	a non-aggregate <a href="#t_firstclass">first class</a> type. The bit sizes of
				4091	<tt>value</tt>
Reid Spencer	19b569f	2007-01-09 20:08:58 +0000	[diff] [blame]	4092	and the destination type, <tt>ty2</tt>, must be identical. If the source
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4093	type is a pointer, the destination type must also be a pointer. This
				4094	instruction supports bitwise conversion of vectors to integers and to vectors
				4095	of other types (as long as they have the same size).</p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	4096
				4097	<h5>Semantics:</h5>
Reid Spencer	5c0ef47	2006-11-11 23:08:07 +0000	[diff] [blame]	4098	<p>The '<tt>bitcast</tt>' instruction converts <tt>value</tt> to type
Reid Spencer	7267925	2006-11-11 21:00:47 +0000	[diff] [blame]	4099	<tt>ty2</tt>. It is always a <i>no-op cast</i> because no bits change with
				4100	this conversion. The conversion is done as if the <tt>value</tt> had been
				4101	stored to memory and read back as type <tt>ty2</tt>. Pointer types may only be
				4102	converted to other pointer types with this instruction. To convert pointers to
				4103	other types, use the <a href="#i_inttoptr">inttoptr</a> or
				4104	<a href="#i_ptrtoint">ptrtoint</a> instructions first.</p>
Reid Spencer	9dee3ac	2006-11-08 01:11:31 +0000	[diff] [blame]	4105
				4106	<h5>Example:</h5>
				4107	<pre>
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4108	%X = bitcast i8 255 to i8 <i>; yields i8 :-1</i>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	4109	%Y = bitcast i32* %x to sint* <i>; yields sint*:%x</i>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4110	%Z = bitcast <2 x int> %V to i64; <i>; yields i64: %V</i>
Chris Lattner	33ba0d9	2001-07-09 00:26:23 +0000	[diff] [blame]	4111	</pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4112	</div>
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	4113
Reid Spencer	2fd21e6	2006-11-08 01:18:52 +0000	[diff] [blame]	4114	<!-- ======================================================================= -->
				4115	<div class="doc_subsection"> <a name="otherops">Other Operations</a> </div>
				4116	<div class="doc_text">
				4117	<p>The instructions in this category are the "miscellaneous"
				4118	instructions, which defy better classification.</p>
				4119	</div>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4120
				4121	<!-- _______________________________________________________________________ -->
				4122	<div class="doc_subsubsection"><a name="i_icmp">'<tt>icmp</tt>' Instruction</a>
				4123	</div>
				4124	<div class="doc_text">
				4125	<h5>Syntax:</h5>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4126	<pre> <result> = icmp <cond> <ty> <op1>, <op2> <i>; yields {i1} or {<N x i1>}:result</i>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4127	</pre>
				4128	<h5>Overview:</h5>
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4129	<p>The '<tt>icmp</tt>' instruction returns a boolean value or
				4130	a vector of boolean values based on comparison
				4131	of its two integer, integer vector, or pointer operands.</p>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4132	<h5>Arguments:</h5>
				4133	<p>The '<tt>icmp</tt>' instruction takes three operands. The first operand is
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4134	the condition code indicating the kind of comparison to perform. It is not
				4135	a value, just a keyword. The possible condition code are:
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4136	</p>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4137	<ol>
				4138	<li><tt>eq</tt>: equal</li>
				4139	<li><tt>ne</tt>: not equal </li>
				4140	<li><tt>ugt</tt>: unsigned greater than</li>
				4141	<li><tt>uge</tt>: unsigned greater or equal</li>
				4142	<li><tt>ult</tt>: unsigned less than</li>
				4143	<li><tt>ule</tt>: unsigned less or equal</li>
				4144	<li><tt>sgt</tt>: signed greater than</li>
				4145	<li><tt>sge</tt>: signed greater or equal</li>
				4146	<li><tt>slt</tt>: signed less than</li>
				4147	<li><tt>sle</tt>: signed less or equal</li>
				4148	</ol>
Chris Lattner	3b19d65	2007-01-15 01:54:13 +0000	[diff] [blame]	4149	<p>The remaining two arguments must be <a href="#t_integer">integer</a> or
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4150	<a href="#t_pointer">pointer</a>
				4151	or integer <a href="#t_vector">vector</a> typed.
				4152	They must also be identical types.</p>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4153	<h5>Semantics:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4154	<p>The '<tt>icmp</tt>' compares <tt>op1</tt> and <tt>op2</tt> according to
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4155	the condition code given as <tt>cond</tt>. The comparison performed always
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4156	yields either an <a href="#t_primitive"><tt>i1</tt></a> or vector of <tt>i1</tt> result, as follows:
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4157	</p>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4158	<ol>
				4159	<li><tt>eq</tt>: yields <tt>true</tt> if the operands are equal,
				4160	<tt>false</tt> otherwise. No sign interpretation is necessary or performed.
				4161	</li>
				4162	<li><tt>ne</tt>: yields <tt>true</tt> if the operands are unequal,
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4163	<tt>false</tt> otherwise. No sign interpretation is necessary or performed.</li>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4164	<li><tt>ugt</tt>: interprets the operands as unsigned values and yields
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4165	<tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4166	<li><tt>uge</tt>: interprets the operands as unsigned values and yields
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4167	<tt>true</tt> if <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4168	<li><tt>ult</tt>: interprets the operands as unsigned values and yields
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4169	<tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4170	<li><tt>ule</tt>: interprets the operands as unsigned values and yields
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4171	<tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4172	<li><tt>sgt</tt>: interprets the operands as signed values and yields
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4173	<tt>true</tt> if <tt>op1</tt> is greater than <tt>op2</tt>.</li>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4174	<li><tt>sge</tt>: interprets the operands as signed values and yields
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4175	<tt>true</tt> if <tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4176	<li><tt>slt</tt>: interprets the operands as signed values and yields
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4177	<tt>true</tt> if <tt>op1</tt> is less than <tt>op2</tt>.</li>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4178	<li><tt>sle</tt>: interprets the operands as signed values and yields
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4179	<tt>true</tt> if <tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4180	</ol>
				4181	<p>If the operands are <a href="#t_pointer">pointer</a> typed, the pointer
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4182	values are compared as if they were integers.</p>
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4183	<p>If the operands are integer vectors, then they are compared
				4184	element by element. The result is an <tt>i1</tt> vector with
				4185	the same number of elements as the values being compared.
				4186	Otherwise, the result is an <tt>i1</tt>.
				4187	</p>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4188
				4189	<h5>Example:</h5>
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	4190	<pre> <result> = icmp eq i32 4, 5 <i>; yields: result=false</i>
				4191	<result> = icmp ne float* %X, %X <i>; yields: result=false</i>
				4192	<result> = icmp ult i16 4, 5 <i>; yields: result=true</i>
				4193	<result> = icmp sgt i16 4, 5 <i>; yields: result=false</i>
				4194	<result> = icmp ule i16 -4, 5 <i>; yields: result=false</i>
				4195	<result> = icmp sge i16 4, 5 <i>; yields: result=false</i>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4196	</pre>
Dan Gohman	c766f72	2009-01-22 01:39:38 +0000	[diff] [blame]	4197
				4198	<p>Note that the code generator does not yet support vector types with
				4199	the <tt>icmp</tt> instruction.</p>
				4200
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4201	</div>
				4202
				4203	<!-- _______________________________________________________________________ -->
				4204	<div class="doc_subsubsection"><a name="i_fcmp">'<tt>fcmp</tt>' Instruction</a>
				4205	</div>
				4206	<div class="doc_text">
				4207	<h5>Syntax:</h5>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4208	<pre> <result> = fcmp <cond> <ty> <op1>, <op2> <i>; yields {i1} or {<N x i1>}:result</i>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4209	</pre>
				4210	<h5>Overview:</h5>
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4211	<p>The '<tt>fcmp</tt>' instruction returns a boolean value
				4212	or vector of boolean values based on comparison
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4213	of its operands.</p>
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4214	<p>
				4215	If the operands are floating point scalars, then the result
				4216	type is a boolean (<a href="#t_primitive"><tt>i1</tt></a>).
				4217	</p>
				4218	<p>If the operands are floating point vectors, then the result type
				4219	is a vector of boolean with the same number of elements as the
				4220	operands being compared.</p>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4221	<h5>Arguments:</h5>
				4222	<p>The '<tt>fcmp</tt>' instruction takes three operands. The first operand is
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4223	the condition code indicating the kind of comparison to perform. It is not
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4224	a value, just a keyword. The possible condition code are:</p>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4225	<ol>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4226	<li><tt>false</tt>: no comparison, always returns false</li>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4227	<li><tt>oeq</tt>: ordered and equal</li>
				4228	<li><tt>ogt</tt>: ordered and greater than </li>
				4229	<li><tt>oge</tt>: ordered and greater than or equal</li>
				4230	<li><tt>olt</tt>: ordered and less than </li>
				4231	<li><tt>ole</tt>: ordered and less than or equal</li>
				4232	<li><tt>one</tt>: ordered and not equal</li>
				4233	<li><tt>ord</tt>: ordered (no nans)</li>
				4234	<li><tt>ueq</tt>: unordered or equal</li>
				4235	<li><tt>ugt</tt>: unordered or greater than </li>
				4236	<li><tt>uge</tt>: unordered or greater than or equal</li>
				4237	<li><tt>ult</tt>: unordered or less than </li>
				4238	<li><tt>ule</tt>: unordered or less than or equal</li>
				4239	<li><tt>une</tt>: unordered or not equal</li>
				4240	<li><tt>uno</tt>: unordered (either nans)</li>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4241	<li><tt>true</tt>: no comparison, always returns true</li>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4242	</ol>
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4243	<p><i>Ordered</i> means that neither operand is a QNAN while
Reid Spencer	93a4985	2006-12-06 07:08:07 +0000	[diff] [blame]	4244	<i>unordered</i> means that either operand may be a QNAN.</p>
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4245	<p>Each of <tt>val1</tt> and <tt>val2</tt> arguments must be
				4246	either a <a href="#t_floating">floating point</a> type
				4247	or a <a href="#t_vector">vector</a> of floating point type.
				4248	They must have identical types.</p>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4249	<h5>Semantics:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4250	<p>The '<tt>fcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4251	according to the condition code given as <tt>cond</tt>.
				4252	If the operands are vectors, then the vectors are compared
				4253	element by element.
				4254	Each comparison performed
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4255	always yields an <a href="#t_primitive">i1</a> result, as follows:</p>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4256	<ol>
				4257	<li><tt>false</tt>: always yields <tt>false</tt>, regardless of operands.</li>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4258	<li><tt>oeq</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4259	<tt>op1</tt> is equal to <tt>op2</tt>.</li>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4260	<li><tt>ogt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4261	<tt>op1</tt> is greather than <tt>op2</tt>.</li>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4262	<li><tt>oge</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4263	<tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4264	<li><tt>olt</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4265	<tt>op1</tt> is less than <tt>op2</tt>.</li>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4266	<li><tt>ole</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4267	<tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4268	<li><tt>one</tt>: yields <tt>true</tt> if both operands are not a QNAN and
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4269	<tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4270	<li><tt>ord</tt>: yields <tt>true</tt> if both operands are not a QNAN.</li>
				4271	<li><tt>ueq</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4272	<tt>op1</tt> is equal to <tt>op2</tt>.</li>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4273	<li><tt>ugt</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4274	<tt>op1</tt> is greater than <tt>op2</tt>.</li>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4275	<li><tt>uge</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4276	<tt>op1</tt> is greater than or equal to <tt>op2</tt>.</li>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4277	<li><tt>ult</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4278	<tt>op1</tt> is less than <tt>op2</tt>.</li>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4279	<li><tt>ule</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4280	<tt>op1</tt> is less than or equal to <tt>op2</tt>.</li>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4281	<li><tt>une</tt>: yields <tt>true</tt> if either operand is a QNAN or
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4282	<tt>op1</tt> is not equal to <tt>op2</tt>.</li>
Reid Spencer	b7f2628	2006-11-19 03:00:14 +0000	[diff] [blame]	4283	<li><tt>uno</tt>: yields <tt>true</tt> if either operand is a QNAN.</li>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4284	<li><tt>true</tt>: always yields <tt>true</tt>, regardless of operands.</li>
				4285	</ol>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4286
				4287	<h5>Example:</h5>
				4288	<pre> <result> = fcmp oeq float 4.0, 5.0 <i>; yields: result=false</i>
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4289	<result> = fcmp one float 4.0, 5.0 <i>; yields: result=true</i>
				4290	<result> = fcmp olt float 4.0, 5.0 <i>; yields: result=true</i>
				4291	<result> = fcmp ueq double 1.0, 2.0 <i>; yields: result=false</i>
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4292	</pre>
Dan Gohman	c766f72	2009-01-22 01:39:38 +0000	[diff] [blame]	4293
				4294	<p>Note that the code generator does not yet support vector types with
				4295	the <tt>fcmp</tt> instruction.</p>
				4296
Reid Spencer	f3a70a6	2006-11-18 21:50:54 +0000	[diff] [blame]	4297	</div>
				4298
Reid Spencer	2fd21e6	2006-11-08 01:18:52 +0000	[diff] [blame]	4299	<!-- _______________________________________________________________________ -->
Nate Begeman	ac80ade	2008-05-12 19:01:56 +0000	[diff] [blame]	4300	<div class="doc_subsubsection">
				4301	<a name="i_vicmp">'<tt>vicmp</tt>' Instruction</a>
				4302	</div>
				4303	<div class="doc_text">
				4304	<h5>Syntax:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4305	<pre> <result> = vicmp <cond> <ty> <op1>, <op2> <i>; yields {ty}:result</i>
Nate Begeman	ac80ade	2008-05-12 19:01:56 +0000	[diff] [blame]	4306	</pre>
				4307	<h5>Overview:</h5>
				4308	<p>The '<tt>vicmp</tt>' instruction returns an integer vector value based on
				4309	element-wise comparison of its two integer vector operands.</p>
				4310	<h5>Arguments:</h5>
				4311	<p>The '<tt>vicmp</tt>' instruction takes three operands. The first operand is
				4312	the condition code indicating the kind of comparison to perform. It is not
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4313	a value, just a keyword. The possible condition code are:</p>
Nate Begeman	ac80ade	2008-05-12 19:01:56 +0000	[diff] [blame]	4314	<ol>
				4315	<li><tt>eq</tt>: equal</li>
				4316	<li><tt>ne</tt>: not equal </li>
				4317	<li><tt>ugt</tt>: unsigned greater than</li>
				4318	<li><tt>uge</tt>: unsigned greater or equal</li>
				4319	<li><tt>ult</tt>: unsigned less than</li>
				4320	<li><tt>ule</tt>: unsigned less or equal</li>
				4321	<li><tt>sgt</tt>: signed greater than</li>
				4322	<li><tt>sge</tt>: signed greater or equal</li>
				4323	<li><tt>slt</tt>: signed less than</li>
				4324	<li><tt>sle</tt>: signed less or equal</li>
				4325	</ol>
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4326	<p>The remaining two arguments must be <a href="#t_vector">vector</a> or
Nate Begeman	ac80ade	2008-05-12 19:01:56 +0000	[diff] [blame]	4327	<a href="#t_integer">integer</a> typed. They must also be identical types.</p>
				4328	<h5>Semantics:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4329	<p>The '<tt>vicmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman	ac80ade	2008-05-12 19:01:56 +0000	[diff] [blame]	4330	according to the condition code given as <tt>cond</tt>. The comparison yields a
				4331	<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, of
				4332	identical type as the values being compared. The most significant bit in each
				4333	element is 1 if the element-wise comparison evaluates to true, and is 0
				4334	otherwise. All other bits of the result are undefined. The condition codes
				4335	are evaluated identically to the <a href="#i_icmp">'<tt>icmp</tt>'
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4336	instruction</a>.</p>
Nate Begeman	ac80ade	2008-05-12 19:01:56 +0000	[diff] [blame]	4337
				4338	<h5>Example:</h5>
				4339	<pre>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4340	<result> = vicmp eq <2 x i32> < i32 4, i32 0>, < i32 5, i32 0> <i>; yields: result=<2 x i32> < i32 0, i32 -1 ></i>
				4341	<result> = vicmp ult <2 x i8 > < i8 1, i8 2>, < i8 2, i8 2 > <i>; yields: result=<2 x i8> < i8 -1, i8 0 ></i>
Nate Begeman	ac80ade	2008-05-12 19:01:56 +0000	[diff] [blame]	4342	</pre>
				4343	</div>
				4344
				4345	<!-- _______________________________________________________________________ -->
				4346	<div class="doc_subsubsection">
				4347	<a name="i_vfcmp">'<tt>vfcmp</tt>' Instruction</a>
				4348	</div>
				4349	<div class="doc_text">
				4350	<h5>Syntax:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4351	<pre> <result> = vfcmp <cond> <ty> <op1>, <op2></pre>
Nate Begeman	ac80ade	2008-05-12 19:01:56 +0000	[diff] [blame]	4352	<h5>Overview:</h5>
				4353	<p>The '<tt>vfcmp</tt>' instruction returns an integer vector value based on
				4354	element-wise comparison of its two floating point vector operands. The output
				4355	elements have the same width as the input elements.</p>
				4356	<h5>Arguments:</h5>
				4357	<p>The '<tt>vfcmp</tt>' instruction takes three operands. The first operand is
				4358	the condition code indicating the kind of comparison to perform. It is not
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4359	a value, just a keyword. The possible condition code are:</p>
Nate Begeman	ac80ade	2008-05-12 19:01:56 +0000	[diff] [blame]	4360	<ol>
				4361	<li><tt>false</tt>: no comparison, always returns false</li>
				4362	<li><tt>oeq</tt>: ordered and equal</li>
				4363	<li><tt>ogt</tt>: ordered and greater than </li>
				4364	<li><tt>oge</tt>: ordered and greater than or equal</li>
				4365	<li><tt>olt</tt>: ordered and less than </li>
				4366	<li><tt>ole</tt>: ordered and less than or equal</li>
				4367	<li><tt>one</tt>: ordered and not equal</li>
				4368	<li><tt>ord</tt>: ordered (no nans)</li>
				4369	<li><tt>ueq</tt>: unordered or equal</li>
				4370	<li><tt>ugt</tt>: unordered or greater than </li>
				4371	<li><tt>uge</tt>: unordered or greater than or equal</li>
				4372	<li><tt>ult</tt>: unordered or less than </li>
				4373	<li><tt>ule</tt>: unordered or less than or equal</li>
				4374	<li><tt>une</tt>: unordered or not equal</li>
				4375	<li><tt>uno</tt>: unordered (either nans)</li>
				4376	<li><tt>true</tt>: no comparison, always returns true</li>
				4377	</ol>
				4378	<p>The remaining two arguments must be <a href="#t_vector">vector</a> of
				4379	<a href="#t_floating">floating point</a> typed. They must also be identical
				4380	types.</p>
				4381	<h5>Semantics:</h5>
Gabor Greif	fb224a2	2008-08-07 21:46:00 +0000	[diff] [blame]	4382	<p>The '<tt>vfcmp</tt>' instruction compares <tt>op1</tt> and <tt>op2</tt>
Nate Begeman	ac80ade	2008-05-12 19:01:56 +0000	[diff] [blame]	4383	according to the condition code given as <tt>cond</tt>. The comparison yields a
				4384	<a href="#t_vector">vector</a> of <a href="#t_integer">integer</a> result, with
				4385	an identical number of elements as the values being compared, and each element
				4386	having identical with to the width of the floating point elements. The most
				4387	significant bit in each element is 1 if the element-wise comparison evaluates to
				4388	true, and is 0 otherwise. All other bits of the result are undefined. The
				4389	condition codes are evaluated identically to the
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4390	<a href="#i_fcmp">'<tt>fcmp</tt>' instruction</a>.</p>
Nate Begeman	ac80ade	2008-05-12 19:01:56 +0000	[diff] [blame]	4391
				4392	<h5>Example:</h5>
				4393	<pre>
Chris Lattner	50ad45c	2008-10-13 16:55:18 +0000	[diff] [blame]	4394	<i>; yields: result=<2 x i32> < i32 0, i32 -1 ></i>
				4395	<result> = vfcmp oeq <2 x float> < float 4, float 0 >, < float 5, float 0 >
				4396
				4397	<i>; yields: result=<2 x i64> < i64 -1, i64 0 ></i>
				4398	<result> = vfcmp ult <2 x double> < double 1, double 2 >, < double 2, double 2>
Nate Begeman	ac80ade	2008-05-12 19:01:56 +0000	[diff] [blame]	4399	</pre>
				4400	</div>
				4401
				4402	<!-- _______________________________________________________________________ -->
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4403	<div class="doc_subsubsection">
				4404	<a name="i_phi">'<tt>phi</tt>' Instruction</a>
				4405	</div>
				4406
Reid Spencer	2fd21e6	2006-11-08 01:18:52 +0000	[diff] [blame]	4407	<div class="doc_text">
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4408
Reid Spencer	2fd21e6	2006-11-08 01:18:52 +0000	[diff] [blame]	4409	<h5>Syntax:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4410
Reid Spencer	2fd21e6	2006-11-08 01:18:52 +0000	[diff] [blame]	4411	<pre> <result> = phi <ty> [ <val0>, <label0>], ...<br></pre>
				4412	<h5>Overview:</h5>
				4413	<p>The '<tt>phi</tt>' instruction is used to implement the φ node in
				4414	the SSA graph representing the function.</p>
				4415	<h5>Arguments:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4416
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4417	<p>The type of the incoming values is specified with the first type
Reid Spencer	2fd21e6	2006-11-08 01:18:52 +0000	[diff] [blame]	4418	field. After this, the '<tt>phi</tt>' instruction takes a list of pairs
				4419	as arguments, with one pair for each predecessor basic block of the
				4420	current block. Only values of <a href="#t_firstclass">first class</a>
				4421	type may be used as the value arguments to the PHI node. Only labels
				4422	may be used as the label arguments.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4423
Reid Spencer	2fd21e6	2006-11-08 01:18:52 +0000	[diff] [blame]	4424	<p>There must be no non-phi instructions between the start of a basic
				4425	block and the PHI instructions: i.e. PHI instructions must be first in
				4426	a basic block.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4427
Reid Spencer	2fd21e6	2006-11-08 01:18:52 +0000	[diff] [blame]	4428	<h5>Semantics:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4429
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4430	<p>At runtime, the '<tt>phi</tt>' instruction logically takes on the value
				4431	specified by the pair corresponding to the predecessor basic block that executed
				4432	just prior to the current block.</p>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4433
Reid Spencer	2fd21e6	2006-11-08 01:18:52 +0000	[diff] [blame]	4434	<h5>Example:</h5>
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4435	<pre>
				4436	Loop: ; Infinite loop that counts from 0 on up...
				4437	%indvar = phi i32 [ 0, %LoopHeader ], [ %nextindvar, %Loop ]
				4438	%nextindvar = add i32 %indvar, 1
				4439	br label %Loop
				4440	</pre>
Reid Spencer	2fd21e6	2006-11-08 01:18:52 +0000	[diff] [blame]	4441	</div>
				4442
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	4443	<!-- _______________________________________________________________________ -->
				4444	<div class="doc_subsubsection">
				4445	<a name="i_select">'<tt>select</tt>' Instruction</a>
				4446	</div>
				4447
				4448	<div class="doc_text">
				4449
				4450	<h5>Syntax:</h5>
				4451
				4452	<pre>
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4453	<result> = select <i>selty</i> <cond>, <ty> <val1>, <ty> <val2> <i>; yields ty</i>
				4454
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4455	<i>selty</i> is either i1 or {<N x i1>}
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	4456	</pre>
				4457
				4458	<h5>Overview:</h5>
				4459
				4460	<p>
				4461	The '<tt>select</tt>' instruction is used to choose one value based on a
				4462	condition, without branching.
				4463	</p>
				4464
				4465
				4466	<h5>Arguments:</h5>
				4467
				4468	<p>
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4469	The '<tt>select</tt>' instruction requires an 'i1' value or
				4470	a vector of 'i1' values indicating the
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4471	condition, and two values of the same <a href="#t_firstclass">first class</a>
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4472	type. If the val1/val2 are vectors and
				4473	the condition is a scalar, then entire vectors are selected, not
Chris Lattner	5568e94	2008-05-20 20:48:21 +0000	[diff] [blame]	4474	individual elements.
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	4475	</p>
				4476
				4477	<h5>Semantics:</h5>
				4478
				4479	<p>
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4480	If the condition is an i1 and it evaluates to 1, the instruction returns the first
John Criswell	fc6b895	2005-05-16 16:17:45 +0000	[diff] [blame]	4481	value argument; otherwise, it returns the second value argument.
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	4482	</p>
Dan Gohman	f72fb67	2008-09-09 01:02:47 +0000	[diff] [blame]	4483	<p>
				4484	If the condition is a vector of i1, then the value arguments must
				4485	be vectors of the same size, and the selection is done element
				4486	by element.
				4487	</p>
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	4488
				4489	<h5>Example:</h5>
				4490
				4491	<pre>
Reid Spencer	c78f337	2007-01-12 03:35:51 +0000	[diff] [blame]	4492	%X = select i1 true, i8 17, i8 42 <i>; yields i8:17</i>
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	4493	</pre>
Dan Gohman	c766f72	2009-01-22 01:39:38 +0000	[diff] [blame]	4494
				4495	<p>Note that the code generator does not yet support conditions
				4496	with vector type.</p>
				4497
Chris Lattner	cc37aae	2004-03-12 05:50:16 +0000	[diff] [blame]	4498	</div>
				4499
Robert Bocchino	05ccd70	2006-01-15 20:48:27 +0000	[diff] [blame]	4500
				4501	<!-- _______________________________________________________________________ -->
				4502	<div class="doc_subsubsection">
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4503	<a name="i_call">'<tt>call</tt>' Instruction</a>
				4504	</div>
				4505
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4506	<div class="doc_text">
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4507
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	4508	<h5>Syntax:</h5>
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4509	<pre>
Devang Patel	307e8ab	2008-10-07 17:48:33 +0000	[diff] [blame]	4510	<result> = [tail] call [<a href="#callingconv">cconv</a>] [<a href="#paramattrs">ret attrs</a>] <ty> [<fnty>*] <fnptrval>(<function args>) [<a href="#fnattrs">fn attrs</a>]
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4511	</pre>
				4512
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	4513	<h5>Overview:</h5>
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4514
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4515	<p>The '<tt>call</tt>' instruction represents a simple function call.</p>
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4516
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	4517	<h5>Arguments:</h5>
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4518
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4519	<p>This instruction requires several arguments:</p>
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4520
Chris Lattner	6536cfe	2002-05-06 22:08:29 +0000	[diff] [blame]	4521	<ol>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	4522	<li>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	4523	<p>The optional "tail" marker indicates whether the callee function accesses
				4524	any allocas or varargs in the caller. If the "tail" marker is present, the
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4525	function call is eligible for tail call optimization. Note that calls may
				4526	be marked "tail" even if they do not occur before a <a
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4527	href="#i_ret"><tt>ret</tt></a> instruction.</p>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	4528	</li>
				4529	<li>
Duncan Sands	8036ca4	2007-03-30 12:22:09 +0000	[diff] [blame]	4530	<p>The optional "cconv" marker indicates which <a href="#callingconv">calling
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	4531	convention</a> the call should use. If none is specified, the call defaults
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4532	to using C calling conventions.</p>
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	4533	</li>
Devang Patel	f642f47	2008-10-06 18:50:38 +0000	[diff] [blame]	4534
				4535	<li>
				4536	<p>The optional <a href="#paramattrs">Parameter Attributes</a> list for
				4537	return values. Only '<tt>zeroext</tt>', '<tt>signext</tt>',
				4538	and '<tt>inreg</tt>' attributes are valid here.</p>
				4539	</li>
				4540
Chris Lattner	bad10ee	2005-05-06 22:57:40 +0000	[diff] [blame]	4541	<li>
Nick Lewycky	db7e3c9	2007-09-08 13:57:50 +0000	[diff] [blame]	4542	<p>'<tt>ty</tt>': the type of the call instruction itself which is also
				4543	the type of the return value. Functions that return no value are marked
				4544	<tt><a href="#t_void">void</a></tt>.</p>
				4545	</li>
				4546	<li>
				4547	<p>'<tt>fnty</tt>': shall be the signature of the pointer to function
				4548	value being invoked. The argument types must match the types implied by
				4549	this signature. This type can be omitted if the function is not varargs
				4550	and if the function type does not return a pointer to a function.</p>
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4551	</li>
				4552	<li>
				4553	<p>'<tt>fnptrval</tt>': An LLVM value containing a pointer to a function to
				4554	be invoked. In most cases, this is a direct function invocation, but
				4555	indirect <tt>call</tt>s are just as possible, calling an arbitrary pointer
John Criswell	fc6b895	2005-05-16 16:17:45 +0000	[diff] [blame]	4556	to function value.</p>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	4557	</li>
				4558	<li>
				4559	<p>'<tt>function args</tt>': argument list whose types match the
Reid Spencer	a7e302a	2005-05-01 22:22:57 +0000	[diff] [blame]	4560	function signature argument types. All arguments must be of
				4561	<a href="#t_firstclass">first class</a> type. If the function signature
				4562	indicates the function accepts a variable number of arguments, the extra
				4563	arguments can be specified.</p>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	4564	</li>
Devang Patel	f642f47	2008-10-06 18:50:38 +0000	[diff] [blame]	4565	<li>
Devang Patel	307e8ab	2008-10-07 17:48:33 +0000	[diff] [blame]	4566	<p>The optional <a href="#fnattrs">function attributes</a> list. Only
Devang Patel	f642f47	2008-10-06 18:50:38 +0000	[diff] [blame]	4567	'<tt>noreturn</tt>', '<tt>nounwind</tt>', '<tt>readonly</tt>' and
				4568	'<tt>readnone</tt>' attributes are valid here.</p>
				4569	</li>
Chris Lattner	6536cfe	2002-05-06 22:08:29 +0000	[diff] [blame]	4570	</ol>
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4571
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	4572	<h5>Semantics:</h5>
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4573
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	4574	<p>The '<tt>call</tt>' instruction is used to cause control flow to
				4575	transfer to a specified function, with its incoming arguments bound to
				4576	the specified values. Upon a '<tt><a href="#i_ret">ret</a></tt>'
				4577	instruction in the called function, control flow continues with the
				4578	instruction after the function call, and the return value of the
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4579	function is bound to the result argument.</p>
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4580
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	4581	<h5>Example:</h5>
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4582
				4583	<pre>
Nick Lewycky	db7e3c9	2007-09-08 13:57:50 +0000	[diff] [blame]	4584	%retval = call i32 @test(i32 %argc)
Chris Lattner	772fccf	2008-03-21 17:24:17 +0000	[diff] [blame]	4585	call i32 (i8 , ...) @printf(i8 * %msg, i32 12, i8 42) <i>; yields i32</i>
				4586	%X = tail call i32 @foo() <i>; yields i32</i>
				4587	%Y = tail call <a href="#callingconv">fastcc</a> i32 @foo() <i>; yields i32</i>
				4588	call void %foo(i8 97 signext)
Devang Patel	c3fc6df	2008-03-10 20:49:15 +0000	[diff] [blame]	4589
				4590	%struct.A = type { i32, i8 }
Devang Patel	f642f47	2008-10-06 18:50:38 +0000	[diff] [blame]	4591	%r = call %struct.A @foo() <i>; yields { 32, i8 }</i>
Dan Gohman	b1e6b96	2008-10-04 19:00:07 +0000	[diff] [blame]	4592	%gr = extractvalue %struct.A %r, 0 <i>; yields i32</i>
				4593	%gr1 = extractvalue %struct.A %r, 1 <i>; yields i8</i>
Chris Lattner	85a350f	2008-10-08 06:26:11 +0000	[diff] [blame]	4594	%Z = call void @foo() noreturn <i>; indicates that %foo never returns normally</i>
Matthijs Kooijman	cb73d19	2008-10-07 10:03:45 +0000	[diff] [blame]	4595	%ZZ = call zeroext i32 @bar() <i>; Return value is %zero extended</i>
Chris Lattner	2bff524	2005-05-06 05:47:36 +0000	[diff] [blame]	4596	</pre>
				4597
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4598	</div>
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4599
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4600	<!-- _______________________________________________________________________ -->
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4601	<div class="doc_subsubsection">
Chris Lattner	fb6977d	2006-01-13 23:26:01 +0000	[diff] [blame]	4602	<a name="i_va_arg">'<tt>va_arg</tt>' Instruction</a>
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4603	</div>
				4604
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4605	<div class="doc_text">
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4606
Chris Lattner	8d1a81d	2003-10-18 05:51:36 +0000	[diff] [blame]	4607	<h5>Syntax:</h5>
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4608
				4609	<pre>
Andrew Lenharth	8bf607a	2005-06-18 18:28:17 +0000	[diff] [blame]	4610	<resultval> = va_arg <va_list*> <arglist>, <argty>
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4611	</pre>
				4612
Chris Lattner	8d1a81d	2003-10-18 05:51:36 +0000	[diff] [blame]	4613	<h5>Overview:</h5>
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4614
Andrew Lenharth	8bf607a	2005-06-18 18:28:17 +0000	[diff] [blame]	4615	<p>The '<tt>va_arg</tt>' instruction is used to access arguments passed through
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4616	the "variable argument" area of a function call. It is used to implement the
				4617	<tt>va_arg</tt> macro in C.</p>
				4618
Chris Lattner	8d1a81d	2003-10-18 05:51:36 +0000	[diff] [blame]	4619	<h5>Arguments:</h5>
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4620
Andrew Lenharth	8bf607a	2005-06-18 18:28:17 +0000	[diff] [blame]	4621	<p>This instruction takes a <tt>va_list*</tt> value and the type of
				4622	the argument. It returns a value of the specified argument type and
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4623	increments the <tt>va_list</tt> to point to the next argument. The
Andrew Lenharth	8bf607a	2005-06-18 18:28:17 +0000	[diff] [blame]	4624	actual type of <tt>va_list</tt> is target specific.</p>
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4625
Chris Lattner	8d1a81d	2003-10-18 05:51:36 +0000	[diff] [blame]	4626	<h5>Semantics:</h5>
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4627
Andrew Lenharth	8bf607a	2005-06-18 18:28:17 +0000	[diff] [blame]	4628	<p>The '<tt>va_arg</tt>' instruction loads an argument of the specified
				4629	type from the specified <tt>va_list</tt> and causes the
				4630	<tt>va_list</tt> to point to the next argument. For more information,
				4631	see the variable argument handling <a href="#int_varargs">Intrinsic
				4632	Functions</a>.</p>
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4633
				4634	<p>It is legal for this instruction to be called in a function which does not
				4635	take a variable number of arguments, for example, the <tt>vfprintf</tt>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4636	function.</p>
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4637
Andrew Lenharth	8bf607a	2005-06-18 18:28:17 +0000	[diff] [blame]	4638	<p><tt>va_arg</tt> is an LLVM instruction instead of an <a
John Criswell	fc6b895	2005-05-16 16:17:45 +0000	[diff] [blame]	4639	href="#intrinsics">intrinsic function</a> because it takes a type as an
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4640	argument.</p>
				4641
Chris Lattner	8d1a81d	2003-10-18 05:51:36 +0000	[diff] [blame]	4642	<h5>Example:</h5>
Chris Lattner	e19d7a7	2004-09-27 21:51:25 +0000	[diff] [blame]	4643
				4644	<p>See the <a href="#int_varargs">variable argument processing</a> section.</p>
				4645
Dan Gohman	f3e60bd	2009-01-12 23:12:39 +0000	[diff] [blame]	4646	<p>Note that the code generator does not yet fully support va_arg
				4647	on many targets. Also, it does not currently support va_arg with
				4648	aggregate types on any target.</p>
				4649
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4650	</div>
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	4651
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4652	<!-- *********************************************************************** -->
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	4653	<div class="doc_section"> <a name="intrinsics">Intrinsic Functions</a> </div>
				4654	<!-- *********************************************************************** -->
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	4655
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4656	<div class="doc_text">
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	4657
				4658	<p>LLVM supports the notion of an "intrinsic function". These functions have
Reid Spencer	409e28f	2007-04-01 08:04:23 +0000	[diff] [blame]	4659	well known names and semantics and are required to follow certain restrictions.
				4660	Overall, these intrinsics represent an extension mechanism for the LLVM
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4661	language that does not require changing all of the transformations in LLVM when
Gabor Greif	04367bf	2007-07-06 22:07:22 +0000	[diff] [blame]	4662	adding to the language (or the bitcode reader/writer, the parser, etc...).</p>
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	4663
John Criswell	fc6b895	2005-05-16 16:17:45 +0000	[diff] [blame]	4664	<p>Intrinsic function names must all start with an "<tt>llvm.</tt>" prefix. This
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4665	prefix is reserved in LLVM for intrinsic names; thus, function names may not
				4666	begin with this prefix. Intrinsic functions must always be external functions:
				4667	you cannot define the body of intrinsic functions. Intrinsic functions may
				4668	only be used in call or invoke instructions: it is illegal to take the address
				4669	of an intrinsic function. Additionally, because intrinsic functions are part
				4670	of the LLVM language, it is required if any are added that they be documented
				4671	here.</p>
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	4672
Chandler Carruth	6994040	2007-08-04 01:51:18 +0000	[diff] [blame]	4673	<p>Some intrinsic functions can be overloaded, i.e., the intrinsic represents
				4674	a family of functions that perform the same operation but on different data
				4675	types. Because LLVM can represent over 8 million different integer types,
				4676	overloading is used commonly to allow an intrinsic function to operate on any
				4677	integer type. One or more of the argument types or the result type can be
				4678	overloaded to accept any integer type. Argument types may also be defined as
				4679	exactly matching a previous argument's type or the result type. This allows an
				4680	intrinsic function which accepts multiple arguments, but needs all of them to
				4681	be of the same type, to only be overloaded with respect to a single argument or
				4682	the result.</p>
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	4683
Chandler Carruth	6994040	2007-08-04 01:51:18 +0000	[diff] [blame]	4684	<p>Overloaded intrinsics will have the names of its overloaded argument types
				4685	encoded into its function name, each preceded by a period. Only those types
				4686	which are overloaded result in a name suffix. Arguments whose type is matched
				4687	against another type do not. For example, the <tt>llvm.ctpop</tt> function can
				4688	take an integer of any width and returns an integer of exactly the same integer
				4689	width. This leads to a family of functions such as
				4690	<tt>i8 @llvm.ctpop.i8(i8 %val)</tt> and <tt>i29 @llvm.ctpop.i29(i29 %val)</tt>.
				4691	Only one type, the return type, is overloaded, and only one type suffix is
				4692	required. Because the argument's type is matched against the return type, it
				4693	does not require its own name suffix.</p>
Reid Spencer	409e28f	2007-04-01 08:04:23 +0000	[diff] [blame]	4694
				4695	<p>To learn how to add an intrinsic function, please see the
				4696	<a href="ExtendingLLVM.html">Extending LLVM Guide</a>.
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	4697	</p>
				4698
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4699	</div>
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	4700
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4701	<!-- ======================================================================= -->
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	4702	<div class="doc_subsection">
				4703	<a name="int_varargs">Variable Argument Handling Intrinsics</a>
				4704	</div>
				4705
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4706	<div class="doc_text">
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4707
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4708	<p>Variable argument support is defined in LLVM with the <a
Chris Lattner	fb6977d	2006-01-13 23:26:01 +0000	[diff] [blame]	4709	href="#i_va_arg"><tt>va_arg</tt></a> instruction and these three
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	4710	intrinsic functions. These functions are related to the similarly
				4711	named macros defined in the <tt><stdarg.h></tt> header file.</p>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4712
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	4713	<p>All of these functions operate on arguments that use a
				4714	target-specific value type "<tt>va_list</tt>". The LLVM assembly
				4715	language reference manual does not define what this type is, so all
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4716	transformations should be prepared to handle these functions regardless of
				4717	the type used.</p>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4718
Chris Lattner	374ab30	2006-05-15 17:26:46 +0000	[diff] [blame]	4719	<p>This example shows how the <a href="#i_va_arg"><tt>va_arg</tt></a>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	4720	instruction and the variable argument handling intrinsic functions are
				4721	used.</p>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4722
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	4723	<div class="doc_code">
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	4724	<pre>
Anton Korobeynikov	5d522f3	2007-03-21 23:58:04 +0000	[diff] [blame]	4725	define i32 @test(i32 %X, ...) {
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	4726	; Initialize variable argument processing
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4727	%ap = alloca i8*
Chris Lattner	b75137d	2007-01-08 07:55:15 +0000	[diff] [blame]	4728	%ap2 = bitcast i8** %ap to i8*
Anton Korobeynikov	5d522f3	2007-03-21 23:58:04 +0000	[diff] [blame]	4729	call void @llvm.va_start(i8* %ap2)
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	4730
				4731	; Read a single integer argument
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4732	%tmp = va_arg i8** %ap, i32
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	4733
				4734	; Demonstrate usage of llvm.va_copy and llvm.va_end
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4735	%aq = alloca i8*
Chris Lattner	b75137d	2007-01-08 07:55:15 +0000	[diff] [blame]	4736	%aq2 = bitcast i8** %aq to i8*
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4737	call void @llvm.va_copy(i8* %aq2, i8* %ap2)
Anton Korobeynikov	5d522f3	2007-03-21 23:58:04 +0000	[diff] [blame]	4738	call void @llvm.va_end(i8* %aq2)
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	4739
				4740	; Stop processing of arguments.
Anton Korobeynikov	5d522f3	2007-03-21 23:58:04 +0000	[diff] [blame]	4741	call void @llvm.va_end(i8* %ap2)
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	4742	ret i32 %tmp
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	4743	}
Anton Korobeynikov	5d522f3	2007-03-21 23:58:04 +0000	[diff] [blame]	4744
				4745	declare void @llvm.va_start(i8*)
				4746	declare void @llvm.va_copy(i8, i8)
				4747	declare void @llvm.va_end(i8*)
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	4748	</pre>
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4749	</div>
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	4750
Bill Wendling	2f7a8b0	2007-05-29 09:04:49 +0000	[diff] [blame]	4751	</div>
				4752
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4753	<!-- _______________________________________________________________________ -->
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	4754	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	4755	<a name="int_va_start">'<tt>llvm.va_start</tt>' Intrinsic</a>
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	4756	</div>
				4757
				4758
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4759	<div class="doc_text">
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4760	<h5>Syntax:</h5>
Chris Lattner	b75137d	2007-01-08 07:55:15 +0000	[diff] [blame]	4761	<pre> declare void %llvm.va_start(i8* <arglist>)<br></pre>
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4762	<h5>Overview:</h5>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4763	<p>The '<tt>llvm.va_start</tt>' intrinsic initializes
Andrew Lenharth	8bf607a	2005-06-18 18:28:17 +0000	[diff] [blame]	4764	<tt>*<arglist></tt> for subsequent use by <tt><a
				4765	href="#i_va_arg">va_arg</a></tt>.</p>
				4766
				4767	<h5>Arguments:</h5>
				4768
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4769	<p>The argument is a pointer to a <tt>va_list</tt> element to initialize.</p>
Andrew Lenharth	8bf607a	2005-06-18 18:28:17 +0000	[diff] [blame]	4770
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4771	<h5>Semantics:</h5>
Andrew Lenharth	8bf607a	2005-06-18 18:28:17 +0000	[diff] [blame]	4772
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	4773	<p>The '<tt>llvm.va_start</tt>' intrinsic works just like the <tt>va_start</tt>
Andrew Lenharth	8bf607a	2005-06-18 18:28:17 +0000	[diff] [blame]	4774	macro available in C. In a target-dependent way, it initializes the
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4775	<tt>va_list</tt> element to which the argument points, so that the next call to
Andrew Lenharth	8bf607a	2005-06-18 18:28:17 +0000	[diff] [blame]	4776	<tt>va_arg</tt> will produce the first variable argument passed to the function.
				4777	Unlike the C <tt>va_start</tt> macro, this intrinsic does not need to know the
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4778	last argument of the function as the compiler can figure that out.</p>
Andrew Lenharth	8bf607a	2005-06-18 18:28:17 +0000	[diff] [blame]	4779
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4780	</div>
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	4781
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4782	<!-- _______________________________________________________________________ -->
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	4783	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	4784	<a name="int_va_end">'<tt>llvm.va_end</tt>' Intrinsic</a>
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	4785	</div>
				4786
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4787	<div class="doc_text">
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4788	<h5>Syntax:</h5>
Anton Korobeynikov	ec43a06	2007-03-22 00:02:17 +0000	[diff] [blame]	4789	<pre> declare void @llvm.va_end(i8* <arglist>)<br></pre>
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4790	<h5>Overview:</h5>
Chris Lattner	b75137d	2007-01-08 07:55:15 +0000	[diff] [blame]	4791
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4792	<p>The '<tt>llvm.va_end</tt>' intrinsic destroys <tt>*<arglist></tt>,
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	4793	which has been initialized previously with <tt><a href="#int_va_start">llvm.va_start</a></tt>
Chris Lattner	261efe9	2003-11-25 01:02:51 +0000	[diff] [blame]	4794	or <tt><a href="#i_va_copy">llvm.va_copy</a></tt>.</p>
Chris Lattner	b75137d	2007-01-08 07:55:15 +0000	[diff] [blame]	4795
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4796	<h5>Arguments:</h5>
Chris Lattner	b75137d	2007-01-08 07:55:15 +0000	[diff] [blame]	4797
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4798	<p>The argument is a pointer to a <tt>va_list</tt> to destroy.</p>
Chris Lattner	b75137d	2007-01-08 07:55:15 +0000	[diff] [blame]	4799
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4800	<h5>Semantics:</h5>
Chris Lattner	b75137d	2007-01-08 07:55:15 +0000	[diff] [blame]	4801
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4802	<p>The '<tt>llvm.va_end</tt>' intrinsic works just like the <tt>va_end</tt>
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4803	macro available in C. In a target-dependent way, it destroys the
				4804	<tt>va_list</tt> element to which the argument points. Calls to <a
				4805	href="#int_va_start"><tt>llvm.va_start</tt></a> and <a href="#int_va_copy">
				4806	<tt>llvm.va_copy</tt></a> must be matched exactly with calls to
				4807	<tt>llvm.va_end</tt>.</p>
Chris Lattner	b75137d	2007-01-08 07:55:15 +0000	[diff] [blame]	4808
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4809	</div>
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	4810
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4811	<!-- _______________________________________________________________________ -->
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	4812	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	4813	<a name="int_va_copy">'<tt>llvm.va_copy</tt>' Intrinsic</a>
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	4814	</div>
				4815
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4816	<div class="doc_text">
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4817
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4818	<h5>Syntax:</h5>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4819
				4820	<pre>
Anton Korobeynikov	ec43a06	2007-03-22 00:02:17 +0000	[diff] [blame]	4821	declare void @llvm.va_copy(i8* <destarglist>, i8* <srcarglist>)
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4822	</pre>
				4823
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4824	<h5>Overview:</h5>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4825
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4826	<p>The '<tt>llvm.va_copy</tt>' intrinsic copies the current argument position
				4827	from the source argument list to the destination argument list.</p>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4828
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4829	<h5>Arguments:</h5>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4830
Andrew Lenharth	8bf607a	2005-06-18 18:28:17 +0000	[diff] [blame]	4831	<p>The first argument is a pointer to a <tt>va_list</tt> element to initialize.
Andrew Lenharth	d0a4c62	2005-06-22 20:38:11 +0000	[diff] [blame]	4832	The second argument is a pointer to a <tt>va_list</tt> element to copy from.</p>
Andrew Lenharth	8bf607a	2005-06-18 18:28:17 +0000	[diff] [blame]	4833
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4834
Chris Lattner	d9ad5b3	2003-05-08 04:57:36 +0000	[diff] [blame]	4835	<h5>Semantics:</h5>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4836
Jeff Cohen	b627eab	2007-04-29 01:07:00 +0000	[diff] [blame]	4837	<p>The '<tt>llvm.va_copy</tt>' intrinsic works just like the <tt>va_copy</tt>
				4838	macro available in C. In a target-dependent way, it copies the source
				4839	<tt>va_list</tt> element into the destination <tt>va_list</tt> element. This
				4840	intrinsic is necessary because the <tt><a href="#int_va_start">
				4841	llvm.va_start</a></tt> intrinsic may be arbitrarily complex and require, for
				4842	example, memory allocation.</p>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4843
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	4844	</div>
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	4845
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	4846	<!-- ======================================================================= -->
				4847	<div class="doc_subsection">
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4848	<a name="int_gc">Accurate Garbage Collection Intrinsics</a>
				4849	</div>
				4850
				4851	<div class="doc_text">
				4852
				4853	<p>
				4854	LLVM support for <a href="GarbageCollection.html">Accurate Garbage
Chris Lattner	d3eda89	2008-08-05 18:29:16 +0000	[diff] [blame]	4855	Collection</a> (GC) requires the implementation and generation of these
				4856	intrinsics.
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	4857	These intrinsics allow identification of <a href="#int_gcroot">GC roots on the
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4858	stack</a>, as well as garbage collector implementations that require <a
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	4859	href="#int_gcread">read</a> and <a href="#int_gcwrite">write</a> barriers.
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4860	Front-ends for type-safe garbage collected languages should generate these
				4861	intrinsics to make use of the LLVM garbage collectors. For more details, see <a
				4862	href="GarbageCollection.html">Accurate Garbage Collection with LLVM</a>.
				4863	</p>
Christopher Lamb	303dae9	2007-12-17 01:00:21 +0000	[diff] [blame]	4864
				4865	<p>The garbage collection intrinsics only operate on objects in the generic
				4866	address space (address space zero).</p>
				4867
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4868	</div>
				4869
				4870	<!-- _______________________________________________________________________ -->
				4871	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	4872	<a name="int_gcroot">'<tt>llvm.gcroot</tt>' Intrinsic</a>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4873	</div>
				4874
				4875	<div class="doc_text">
				4876
				4877	<h5>Syntax:</h5>
				4878
				4879	<pre>
Chris Lattner	1df4f75	2007-09-21 17:30:40 +0000	[diff] [blame]	4880	declare void @llvm.gcroot(i8** %ptrloc, i8* %metadata)
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4881	</pre>
				4882
				4883	<h5>Overview:</h5>
				4884
John Criswell	9e2485c	2004-12-10 15:51:16 +0000	[diff] [blame]	4885	<p>The '<tt>llvm.gcroot</tt>' intrinsic declares the existence of a GC root to
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4886	the code generator, and allows some metadata to be associated with it.</p>
				4887
				4888	<h5>Arguments:</h5>
				4889
				4890	<p>The first argument specifies the address of a stack object that contains the
				4891	root pointer. The second pointer (which must be either a constant or a global
				4892	value address) contains the meta-data to be associated with the root.</p>
				4893
				4894	<h5>Semantics:</h5>
				4895
Chris Lattner	05d6709	2008-04-24 05:59:56 +0000	[diff] [blame]	4896	<p>At runtime, a call to this intrinsic stores a null pointer into the "ptrloc"
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4897	location. At compile-time, the code generator generates information to allow
Gordon Henriksen	e1433f2	2007-12-25 02:31:26 +0000	[diff] [blame]	4898	the runtime to find the pointer at GC safe points. The '<tt>llvm.gcroot</tt>'
				4899	intrinsic may only be used in a function which <a href="#gc">specifies a GC
				4900	algorithm</a>.</p>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4901
				4902	</div>
				4903
				4904
				4905	<!-- _______________________________________________________________________ -->
				4906	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	4907	<a name="int_gcread">'<tt>llvm.gcread</tt>' Intrinsic</a>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4908	</div>
				4909
				4910	<div class="doc_text">
				4911
				4912	<h5>Syntax:</h5>
				4913
				4914	<pre>
Chris Lattner	1df4f75	2007-09-21 17:30:40 +0000	[diff] [blame]	4915	declare i8* @llvm.gcread(i8* %ObjPtr, i8** %Ptr)
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4916	</pre>
				4917
				4918	<h5>Overview:</h5>
				4919
				4920	<p>The '<tt>llvm.gcread</tt>' intrinsic identifies reads of references from heap
				4921	locations, allowing garbage collector implementations that require read
				4922	barriers.</p>
				4923
				4924	<h5>Arguments:</h5>
				4925
Chris Lattner	80626e9	2006-03-14 20:02:51 +0000	[diff] [blame]	4926	<p>The second argument is the address to read from, which should be an address
				4927	allocated from the garbage collector. The first object is a pointer to the
				4928	start of the referenced object, if needed by the language runtime (otherwise
				4929	null).</p>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4930
				4931	<h5>Semantics:</h5>
				4932
				4933	<p>The '<tt>llvm.gcread</tt>' intrinsic has the same semantics as a load
				4934	instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen	e1433f2	2007-12-25 02:31:26 +0000	[diff] [blame]	4935	garbage collector runtime, as needed. The '<tt>llvm.gcread</tt>' intrinsic
				4936	may only be used in a function which <a href="#gc">specifies a GC
				4937	algorithm</a>.</p>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4938
				4939	</div>
				4940
				4941
				4942	<!-- _______________________________________________________________________ -->
				4943	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	4944	<a name="int_gcwrite">'<tt>llvm.gcwrite</tt>' Intrinsic</a>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4945	</div>
				4946
				4947	<div class="doc_text">
				4948
				4949	<h5>Syntax:</h5>
				4950
				4951	<pre>
Chris Lattner	1df4f75	2007-09-21 17:30:40 +0000	[diff] [blame]	4952	declare void @llvm.gcwrite(i8* %P1, i8* %Obj, i8** %P2)
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4953	</pre>
				4954
				4955	<h5>Overview:</h5>
				4956
				4957	<p>The '<tt>llvm.gcwrite</tt>' intrinsic identifies writes of references to heap
				4958	locations, allowing garbage collector implementations that require write
				4959	barriers (such as generational or reference counting collectors).</p>
				4960
				4961	<h5>Arguments:</h5>
				4962
Chris Lattner	80626e9	2006-03-14 20:02:51 +0000	[diff] [blame]	4963	<p>The first argument is the reference to store, the second is the start of the
				4964	object to store it to, and the third is the address of the field of Obj to
				4965	store to. If the runtime does not require a pointer to the object, Obj may be
				4966	null.</p>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4967
				4968	<h5>Semantics:</h5>
				4969
				4970	<p>The '<tt>llvm.gcwrite</tt>' intrinsic has the same semantics as a store
				4971	instruction, but may be replaced with substantially more complex code by the
Gordon Henriksen	e1433f2	2007-12-25 02:31:26 +0000	[diff] [blame]	4972	garbage collector runtime, as needed. The '<tt>llvm.gcwrite</tt>' intrinsic
				4973	may only be used in a function which <a href="#gc">specifies a GC
				4974	algorithm</a>.</p>
Chris Lattner	d792391	2004-05-23 21:06:01 +0000	[diff] [blame]	4975
				4976	</div>
				4977
				4978
				4979
				4980	<!-- ======================================================================= -->
				4981	<div class="doc_subsection">
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	4982	<a name="int_codegen">Code Generator Intrinsics</a>
				4983	</div>
				4984
				4985	<div class="doc_text">
				4986	<p>
				4987	These intrinsics are provided by LLVM to expose special features that may only
				4988	be implemented with code generator support.
				4989	</p>
				4990
				4991	</div>
				4992
				4993	<!-- _______________________________________________________________________ -->
				4994	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	4995	<a name="int_returnaddress">'<tt>llvm.returnaddress</tt>' Intrinsic</a>
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	4996	</div>
				4997
				4998	<div class="doc_text">
				4999
				5000	<h5>Syntax:</h5>
				5001	<pre>
Anton Korobeynikov	ec43a06	2007-03-22 00:02:17 +0000	[diff] [blame]	5002	declare i8 *@llvm.returnaddress(i32 <level>)
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5003	</pre>
				5004
				5005	<h5>Overview:</h5>
				5006
				5007	<p>
Chris Lattner	32b5d71	2006-10-15 20:05:59 +0000	[diff] [blame]	5008	The '<tt>llvm.returnaddress</tt>' intrinsic attempts to compute a
				5009	target-specific value indicating the return address of the current function
				5010	or one of its callers.
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5011	</p>
				5012
				5013	<h5>Arguments:</h5>
				5014
				5015	<p>
				5016	The argument to this intrinsic indicates which function to return the address
				5017	for. Zero indicates the calling function, one indicates its caller, etc. The
				5018	argument is <b>required</b> to be a constant integer value.
				5019	</p>
				5020
				5021	<h5>Semantics:</h5>
				5022
				5023	<p>
				5024	The '<tt>llvm.returnaddress</tt>' intrinsic either returns a pointer indicating
				5025	the return address of the specified call frame, or zero if it cannot be
				5026	identified. The value returned by this intrinsic is likely to be incorrect or 0
				5027	for arguments other than zero, so it should only be used for debugging purposes.
				5028	</p>
				5029
				5030	<p>
				5031	Note that calling this intrinsic does not prevent function inlining or other
Chris Lattner	b40bb38	2005-03-07 20:30:51 +0000	[diff] [blame]	5032	aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5033	source-language caller.
				5034	</p>
				5035	</div>
				5036
				5037
				5038	<!-- _______________________________________________________________________ -->
				5039	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5040	<a name="int_frameaddress">'<tt>llvm.frameaddress</tt>' Intrinsic</a>
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5041	</div>
				5042
				5043	<div class="doc_text">
				5044
				5045	<h5>Syntax:</h5>
				5046	<pre>
Chris Lattner	1df4f75	2007-09-21 17:30:40 +0000	[diff] [blame]	5047	declare i8 *@llvm.frameaddress(i32 <level>)
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5048	</pre>
				5049
				5050	<h5>Overview:</h5>
				5051
				5052	<p>
Chris Lattner	32b5d71	2006-10-15 20:05:59 +0000	[diff] [blame]	5053	The '<tt>llvm.frameaddress</tt>' intrinsic attempts to return the
				5054	target-specific frame pointer value for the specified stack frame.
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5055	</p>
				5056
				5057	<h5>Arguments:</h5>
				5058
				5059	<p>
				5060	The argument to this intrinsic indicates which function to return the frame
				5061	pointer for. Zero indicates the calling function, one indicates its caller,
				5062	etc. The argument is <b>required</b> to be a constant integer value.
				5063	</p>
				5064
				5065	<h5>Semantics:</h5>
				5066
				5067	<p>
				5068	The '<tt>llvm.frameaddress</tt>' intrinsic either returns a pointer indicating
				5069	the frame address of the specified call frame, or zero if it cannot be
				5070	identified. The value returned by this intrinsic is likely to be incorrect or 0
				5071	for arguments other than zero, so it should only be used for debugging purposes.
				5072	</p>
				5073
				5074	<p>
				5075	Note that calling this intrinsic does not prevent function inlining or other
Chris Lattner	b40bb38	2005-03-07 20:30:51 +0000	[diff] [blame]	5076	aggressive transformations, so the value returned may not be that of the obvious
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5077	source-language caller.
				5078	</p>
				5079	</div>
				5080
Chris Lattner	9a9d7ac	2005-02-28 19:24:19 +0000	[diff] [blame]	5081	<!-- _______________________________________________________________________ -->
				5082	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5083	<a name="int_stacksave">'<tt>llvm.stacksave</tt>' Intrinsic</a>
Chris Lattner	57e1f39	2006-01-13 02:03:13 +0000	[diff] [blame]	5084	</div>
				5085
				5086	<div class="doc_text">
				5087
				5088	<h5>Syntax:</h5>
				5089	<pre>
Chris Lattner	1df4f75	2007-09-21 17:30:40 +0000	[diff] [blame]	5090	declare i8 *@llvm.stacksave()
Chris Lattner	57e1f39	2006-01-13 02:03:13 +0000	[diff] [blame]	5091	</pre>
				5092
				5093	<h5>Overview:</h5>
				5094
				5095	<p>
				5096	The '<tt>llvm.stacksave</tt>' intrinsic is used to remember the current state of
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5097	the function stack, for use with <a href="#int_stackrestore">
Chris Lattner	57e1f39	2006-01-13 02:03:13 +0000	[diff] [blame]	5098	<tt>llvm.stackrestore</tt></a>. This is useful for implementing language
				5099	features like scoped automatic variable sized arrays in C99.
				5100	</p>
				5101
				5102	<h5>Semantics:</h5>
				5103
				5104	<p>
				5105	This intrinsic returns a opaque pointer value that can be passed to <a
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5106	href="#int_stackrestore"><tt>llvm.stackrestore</tt></a>. When an
Chris Lattner	57e1f39	2006-01-13 02:03:13 +0000	[diff] [blame]	5107	<tt>llvm.stackrestore</tt> intrinsic is executed with a value saved from
				5108	<tt>llvm.stacksave</tt>, it effectively restores the state of the stack to the
				5109	state it was in when the <tt>llvm.stacksave</tt> intrinsic executed. In
				5110	practice, this pops any <a href="#i_alloca">alloca</a> blocks from the stack
				5111	that were allocated after the <tt>llvm.stacksave</tt> was executed.
				5112	</p>
				5113
				5114	</div>
				5115
				5116	<!-- _______________________________________________________________________ -->
				5117	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5118	<a name="int_stackrestore">'<tt>llvm.stackrestore</tt>' Intrinsic</a>
Chris Lattner	57e1f39	2006-01-13 02:03:13 +0000	[diff] [blame]	5119	</div>
				5120
				5121	<div class="doc_text">
				5122
				5123	<h5>Syntax:</h5>
				5124	<pre>
Anton Korobeynikov	ec43a06	2007-03-22 00:02:17 +0000	[diff] [blame]	5125	declare void @llvm.stackrestore(i8 * %ptr)
Chris Lattner	57e1f39	2006-01-13 02:03:13 +0000	[diff] [blame]	5126	</pre>
				5127
				5128	<h5>Overview:</h5>
				5129
				5130	<p>
				5131	The '<tt>llvm.stackrestore</tt>' intrinsic is used to restore the state of
				5132	the function stack to the state it was in when the corresponding <a
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5133	href="#int_stacksave"><tt>llvm.stacksave</tt></a> intrinsic executed. This is
Chris Lattner	57e1f39	2006-01-13 02:03:13 +0000	[diff] [blame]	5134	useful for implementing language features like scoped automatic variable sized
				5135	arrays in C99.
				5136	</p>
				5137
				5138	<h5>Semantics:</h5>
				5139
				5140	<p>
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5141	See the description for <a href="#int_stacksave"><tt>llvm.stacksave</tt></a>.
Chris Lattner	57e1f39	2006-01-13 02:03:13 +0000	[diff] [blame]	5142	</p>
				5143
				5144	</div>
				5145
				5146
				5147	<!-- _______________________________________________________________________ -->
				5148	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5149	<a name="int_prefetch">'<tt>llvm.prefetch</tt>' Intrinsic</a>
Chris Lattner	9a9d7ac	2005-02-28 19:24:19 +0000	[diff] [blame]	5150	</div>
				5151
				5152	<div class="doc_text">
				5153
				5154	<h5>Syntax:</h5>
				5155	<pre>
Chris Lattner	1df4f75	2007-09-21 17:30:40 +0000	[diff] [blame]	5156	declare void @llvm.prefetch(i8* <address>, i32 <rw>, i32 <locality>)
Chris Lattner	9a9d7ac	2005-02-28 19:24:19 +0000	[diff] [blame]	5157	</pre>
				5158
				5159	<h5>Overview:</h5>
				5160
				5161
				5162	<p>
				5163	The '<tt>llvm.prefetch</tt>' intrinsic is a hint to the code generator to insert
John Criswell	fc6b895	2005-05-16 16:17:45 +0000	[diff] [blame]	5164	a prefetch instruction if supported; otherwise, it is a noop. Prefetches have
				5165	no
				5166	effect on the behavior of the program but can change its performance
Chris Lattner	2a61536	2005-02-28 19:47:14 +0000	[diff] [blame]	5167	characteristics.
Chris Lattner	9a9d7ac	2005-02-28 19:24:19 +0000	[diff] [blame]	5168	</p>
				5169
				5170	<h5>Arguments:</h5>
				5171
				5172	<p>
				5173	<tt>address</tt> is the address to be prefetched, <tt>rw</tt> is the specifier
				5174	determining if the fetch should be for a read (0) or write (1), and
				5175	<tt>locality</tt> is a temporal locality specifier ranging from (0) - no
Chris Lattner	aeffb4a	2005-03-07 20:31:38 +0000	[diff] [blame]	5176	locality, to (3) - extremely local keep in cache. The <tt>rw</tt> and
Chris Lattner	9a9d7ac	2005-02-28 19:24:19 +0000	[diff] [blame]	5177	<tt>locality</tt> arguments must be constant integers.
				5178	</p>
				5179
				5180	<h5>Semantics:</h5>
				5181
				5182	<p>
				5183	This intrinsic does not modify the behavior of the program. In particular,
				5184	prefetches cannot trap and do not produce a value. On targets that support this
				5185	intrinsic, the prefetch can provide hints to the processor cache for better
				5186	performance.
				5187	</p>
				5188
				5189	</div>
				5190
Andrew Lenharth	7f4ec3b	2005-03-28 20:05:49 +0000	[diff] [blame]	5191	<!-- _______________________________________________________________________ -->
				5192	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5193	<a name="int_pcmarker">'<tt>llvm.pcmarker</tt>' Intrinsic</a>
Andrew Lenharth	7f4ec3b	2005-03-28 20:05:49 +0000	[diff] [blame]	5194	</div>
				5195
				5196	<div class="doc_text">
				5197
				5198	<h5>Syntax:</h5>
				5199	<pre>
Chris Lattner	1df4f75	2007-09-21 17:30:40 +0000	[diff] [blame]	5200	declare void @llvm.pcmarker(i32 <id>)
Andrew Lenharth	7f4ec3b	2005-03-28 20:05:49 +0000	[diff] [blame]	5201	</pre>
				5202
				5203	<h5>Overview:</h5>
				5204
				5205
				5206	<p>
John Criswell	fc6b895	2005-05-16 16:17:45 +0000	[diff] [blame]	5207	The '<tt>llvm.pcmarker</tt>' intrinsic is a method to export a Program Counter
Chris Lattner	d3eda89	2008-08-05 18:29:16 +0000	[diff] [blame]	5208	(PC) in a region of
				5209	code to simulators and other tools. The method is target specific, but it is
				5210	expected that the marker will use exported symbols to transmit the PC of the
				5211	marker.
				5212	The marker makes no guarantees that it will remain with any specific instruction
				5213	after optimizations. It is possible that the presence of a marker will inhibit
Chris Lattner	b3e7afd	2006-03-24 07:16:10 +0000	[diff] [blame]	5214	optimizations. The intended use is to be inserted after optimizations to allow
John Criswell	fc6b895	2005-05-16 16:17:45 +0000	[diff] [blame]	5215	correlations of simulation runs.
Andrew Lenharth	7f4ec3b	2005-03-28 20:05:49 +0000	[diff] [blame]	5216	</p>
				5217
				5218	<h5>Arguments:</h5>
				5219
				5220	<p>
				5221	<tt>id</tt> is a numerical id identifying the marker.
				5222	</p>
				5223
				5224	<h5>Semantics:</h5>
				5225
				5226	<p>
				5227	This intrinsic does not modify the behavior of the program. Backends that do not
				5228	support this intrinisic may ignore it.
				5229	</p>
				5230
				5231	</div>
				5232
Andrew Lenharth	51b8d54	2005-11-11 16:47:30 +0000	[diff] [blame]	5233	<!-- _______________________________________________________________________ -->
				5234	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5235	<a name="int_readcyclecounter">'<tt>llvm.readcyclecounter</tt>' Intrinsic</a>
Andrew Lenharth	51b8d54	2005-11-11 16:47:30 +0000	[diff] [blame]	5236	</div>
				5237
				5238	<div class="doc_text">
				5239
				5240	<h5>Syntax:</h5>
				5241	<pre>
Anton Korobeynikov	ec43a06	2007-03-22 00:02:17 +0000	[diff] [blame]	5242	declare i64 @llvm.readcyclecounter( )
Andrew Lenharth	51b8d54	2005-11-11 16:47:30 +0000	[diff] [blame]	5243	</pre>
				5244
				5245	<h5>Overview:</h5>
				5246
				5247
				5248	<p>
				5249	The '<tt>llvm.readcyclecounter</tt>' intrinsic provides access to the cycle
				5250	counter register (or similar low latency, high accuracy clocks) on those targets
				5251	that support it. On X86, it should map to RDTSC. On Alpha, it should map to RPCC.
				5252	As the backing counters overflow quickly (on the order of 9 seconds on alpha), this
				5253	should only be used for small timings.
				5254	</p>
				5255
				5256	<h5>Semantics:</h5>
				5257
				5258	<p>
				5259	When directly supported, reading the cycle counter should not modify any memory.
				5260	Implementations are allowed to either return a application specific value or a
				5261	system wide value. On backends without support, this is lowered to a constant 0.
				5262	</p>
				5263
				5264	</div>
				5265
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5266	<!-- ======================================================================= -->
				5267	<div class="doc_subsection">
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	5268	<a name="int_libc">Standard C Library Intrinsics</a>
				5269	</div>
				5270
				5271	<div class="doc_text">
				5272	<p>
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5273	LLVM provides intrinsics for a few important standard C library functions.
				5274	These intrinsics allow source-language front-ends to pass information about the
				5275	alignment of the pointer arguments to the code generator, providing opportunity
				5276	for more efficient code generation.
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	5277	</p>
				5278
				5279	</div>
				5280
				5281	<!-- _______________________________________________________________________ -->
				5282	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5283	<a name="int_memcpy">'<tt>llvm.memcpy</tt>' Intrinsic</a>
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	5284	</div>
				5285
				5286	<div class="doc_text">
				5287
				5288	<h5>Syntax:</h5>
Chris Lattner	824b958	2008-11-21 16:42:48 +0000	[diff] [blame]	5289	<p>This is an overloaded intrinsic. You can use llvm.memcpy on any integer bit
				5290	width. Not all targets support all bit widths however.</p>
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	5291	<pre>
Chris Lattner	824b958	2008-11-21 16:42:48 +0000	[diff] [blame]	5292	declare void @llvm.memcpy.i8(i8 * <dest>, i8 * <src>,
				5293	i8 <len>, i32 <align>)
				5294	declare void @llvm.memcpy.i16(i8 * <dest>, i8 * <src>,
				5295	i16 <len>, i32 <align>)
Anton Korobeynikov	ec43a06	2007-03-22 00:02:17 +0000	[diff] [blame]	5296	declare void @llvm.memcpy.i32(i8 * <dest>, i8 * <src>,
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	5297	i32 <len>, i32 <align>)
Anton Korobeynikov	ec43a06	2007-03-22 00:02:17 +0000	[diff] [blame]	5298	declare void @llvm.memcpy.i64(i8 * <dest>, i8 * <src>,
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	5299	i64 <len>, i32 <align>)
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	5300	</pre>
				5301
				5302	<h5>Overview:</h5>
				5303
				5304	<p>
Chris Lattner	5b310c3	2006-03-03 00:07:20 +0000	[diff] [blame]	5305	The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	5306	location to the destination location.
				5307	</p>
				5308
				5309	<p>
Chris Lattner	5b310c3	2006-03-03 00:07:20 +0000	[diff] [blame]	5310	Note that, unlike the standard libc function, the <tt>llvm.memcpy.*</tt>
				5311	intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	5312	</p>
				5313
				5314	<h5>Arguments:</h5>
				5315
				5316	<p>
				5317	The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner	5b310c3	2006-03-03 00:07:20 +0000	[diff] [blame]	5318	the source. The third argument is an integer argument
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	5319	specifying the number of bytes to copy, and the fourth argument is the alignment
				5320	of the source and destination locations.
				5321	</p>
				5322
Chris Lattner	3301ced	2004-02-12 21:18:15 +0000	[diff] [blame]	5323	<p>
				5324	If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattner	f0afc2c	2006-03-04 00:02:10 +0000	[diff] [blame]	5325	the caller guarantees that both the source and destination pointers are aligned
				5326	to that boundary.
Chris Lattner	3301ced	2004-02-12 21:18:15 +0000	[diff] [blame]	5327	</p>
				5328
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	5329	<h5>Semantics:</h5>
				5330
				5331	<p>
Chris Lattner	5b310c3	2006-03-03 00:07:20 +0000	[diff] [blame]	5332	The '<tt>llvm.memcpy.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner	33aec9e	2004-02-12 17:01:32 +0000	[diff] [blame]	5333	location to the destination location, which are not allowed to overlap. It
				5334	copies "len" bytes of memory over. If the argument is known to be aligned to
				5335	some boundary, this can be specified as the fourth argument, otherwise it should
				5336	be set to 0 or 1.
				5337	</p>
				5338	</div>
				5339
				5340
Chris Lattner	0eb51b4	2004-02-12 18:10:10 +0000	[diff] [blame]	5341	<!-- _______________________________________________________________________ -->
				5342	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5343	<a name="int_memmove">'<tt>llvm.memmove</tt>' Intrinsic</a>
Chris Lattner	0eb51b4	2004-02-12 18:10:10 +0000	[diff] [blame]	5344	</div>
				5345
				5346	<div class="doc_text">
				5347
				5348	<h5>Syntax:</h5>
Chris Lattner	824b958	2008-11-21 16:42:48 +0000	[diff] [blame]	5349	<p>This is an overloaded intrinsic. You can use llvm.memmove on any integer bit
				5350	width. Not all targets support all bit widths however.</p>
Chris Lattner	0eb51b4	2004-02-12 18:10:10 +0000	[diff] [blame]	5351	<pre>
Chris Lattner	824b958	2008-11-21 16:42:48 +0000	[diff] [blame]	5352	declare void @llvm.memmove.i8(i8 * <dest>, i8 * <src>,
				5353	i8 <len>, i32 <align>)
				5354	declare void @llvm.memmove.i16(i8 * <dest>, i8 * <src>,
				5355	i16 <len>, i32 <align>)
Anton Korobeynikov	ec43a06	2007-03-22 00:02:17 +0000	[diff] [blame]	5356	declare void @llvm.memmove.i32(i8 * <dest>, i8 * <src>,
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	5357	i32 <len>, i32 <align>)
Anton Korobeynikov	ec43a06	2007-03-22 00:02:17 +0000	[diff] [blame]	5358	declare void @llvm.memmove.i64(i8 * <dest>, i8 * <src>,
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	5359	i64 <len>, i32 <align>)
Chris Lattner	0eb51b4	2004-02-12 18:10:10 +0000	[diff] [blame]	5360	</pre>
				5361
				5362	<h5>Overview:</h5>
				5363
				5364	<p>
Chris Lattner	5b310c3	2006-03-03 00:07:20 +0000	[diff] [blame]	5365	The '<tt>llvm.memmove.*</tt>' intrinsics move a block of memory from the source
				5366	location to the destination location. It is similar to the
Chris Lattner	4b2cbcf	2008-01-06 19:51:52 +0000	[diff] [blame]	5367	'<tt>llvm.memcpy</tt>' intrinsic but allows the two memory locations to overlap.
Chris Lattner	0eb51b4	2004-02-12 18:10:10 +0000	[diff] [blame]	5368	</p>
				5369
				5370	<p>
Chris Lattner	5b310c3	2006-03-03 00:07:20 +0000	[diff] [blame]	5371	Note that, unlike the standard libc function, the <tt>llvm.memmove.*</tt>
				5372	intrinsics do not return a value, and takes an extra alignment argument.
Chris Lattner	0eb51b4	2004-02-12 18:10:10 +0000	[diff] [blame]	5373	</p>
				5374
				5375	<h5>Arguments:</h5>
				5376
				5377	<p>
				5378	The first argument is a pointer to the destination, the second is a pointer to
Chris Lattner	5b310c3	2006-03-03 00:07:20 +0000	[diff] [blame]	5379	the source. The third argument is an integer argument
Chris Lattner	0eb51b4	2004-02-12 18:10:10 +0000	[diff] [blame]	5380	specifying the number of bytes to copy, and the fourth argument is the alignment
				5381	of the source and destination locations.
				5382	</p>
				5383
Chris Lattner	3301ced	2004-02-12 21:18:15 +0000	[diff] [blame]	5384	<p>
				5385	If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattner	f0afc2c	2006-03-04 00:02:10 +0000	[diff] [blame]	5386	the caller guarantees that the source and destination pointers are aligned to
				5387	that boundary.
Chris Lattner	3301ced	2004-02-12 21:18:15 +0000	[diff] [blame]	5388	</p>
				5389
Chris Lattner	0eb51b4	2004-02-12 18:10:10 +0000	[diff] [blame]	5390	<h5>Semantics:</h5>
				5391
				5392	<p>
Chris Lattner	5b310c3	2006-03-03 00:07:20 +0000	[diff] [blame]	5393	The '<tt>llvm.memmove.*</tt>' intrinsics copy a block of memory from the source
Chris Lattner	0eb51b4	2004-02-12 18:10:10 +0000	[diff] [blame]	5394	location to the destination location, which may overlap. It
				5395	copies "len" bytes of memory over. If the argument is known to be aligned to
				5396	some boundary, this can be specified as the fourth argument, otherwise it should
				5397	be set to 0 or 1.
				5398	</p>
				5399	</div>
				5400
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	5401
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5402	<!-- _______________________________________________________________________ -->
				5403	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5404	<a name="int_memset">'<tt>llvm.memset.*</tt>' Intrinsics</a>
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5405	</div>
				5406
				5407	<div class="doc_text">
				5408
				5409	<h5>Syntax:</h5>
Chris Lattner	824b958	2008-11-21 16:42:48 +0000	[diff] [blame]	5410	<p>This is an overloaded intrinsic. You can use llvm.memset on any integer bit
				5411	width. Not all targets support all bit widths however.</p>
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5412	<pre>
Chris Lattner	824b958	2008-11-21 16:42:48 +0000	[diff] [blame]	5413	declare void @llvm.memset.i8(i8 * <dest>, i8 <val>,
				5414	i8 <len>, i32 <align>)
				5415	declare void @llvm.memset.i16(i8 * <dest>, i8 <val>,
				5416	i16 <len>, i32 <align>)
Anton Korobeynikov	ec43a06	2007-03-22 00:02:17 +0000	[diff] [blame]	5417	declare void @llvm.memset.i32(i8 * <dest>, i8 <val>,
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	5418	i32 <len>, i32 <align>)
Anton Korobeynikov	ec43a06	2007-03-22 00:02:17 +0000	[diff] [blame]	5419	declare void @llvm.memset.i64(i8 * <dest>, i8 <val>,
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	5420	i64 <len>, i32 <align>)
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5421	</pre>
				5422
				5423	<h5>Overview:</h5>
				5424
				5425	<p>
Chris Lattner	5b310c3	2006-03-03 00:07:20 +0000	[diff] [blame]	5426	The '<tt>llvm.memset.*</tt>' intrinsics fill a block of memory with a particular
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5427	byte value.
				5428	</p>
				5429
				5430	<p>
				5431	Note that, unlike the standard libc function, the <tt>llvm.memset</tt> intrinsic
				5432	does not return a value, and takes an extra alignment argument.
				5433	</p>
				5434
				5435	<h5>Arguments:</h5>
				5436
				5437	<p>
				5438	The first argument is a pointer to the destination to fill, the second is the
Chris Lattner	5b310c3	2006-03-03 00:07:20 +0000	[diff] [blame]	5439	byte value to fill it with, the third argument is an integer
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5440	argument specifying the number of bytes to fill, and the fourth argument is the
				5441	known alignment of destination location.
				5442	</p>
				5443
				5444	<p>
				5445	If the call to this intrinisic has an alignment value that is not 0 or 1, then
Chris Lattner	f0afc2c	2006-03-04 00:02:10 +0000	[diff] [blame]	5446	the caller guarantees that the destination pointer is aligned to that boundary.
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5447	</p>
				5448
				5449	<h5>Semantics:</h5>
				5450
				5451	<p>
Chris Lattner	5b310c3	2006-03-03 00:07:20 +0000	[diff] [blame]	5452	The '<tt>llvm.memset.*</tt>' intrinsics fill "len" bytes of memory starting at
				5453	the
Chris Lattner	1061064	2004-02-14 04:08:35 +0000	[diff] [blame]	5454	destination location. If the argument is known to be aligned to some boundary,
				5455	this can be specified as the fourth argument, otherwise it should be set to 0 or
				5456	1.
				5457	</p>
				5458	</div>
				5459
				5460
Chris Lattner	3200628	2004-06-11 02:28:03 +0000	[diff] [blame]	5461	<!-- _______________________________________________________________________ -->
				5462	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5463	<a name="int_sqrt">'<tt>llvm.sqrt.*</tt>' Intrinsic</a>
Chris Lattner	a4d7414	2005-07-21 01:29:16 +0000	[diff] [blame]	5464	</div>
				5465
				5466	<div class="doc_text">
				5467
				5468	<h5>Syntax:</h5>
Dale Johannesen	408f9c1	2007-10-02 17:47:38 +0000	[diff] [blame]	5469	<p>This is an overloaded intrinsic. You can use <tt>llvm.sqrt</tt> on any
Dan Gohman	91c284c	2007-10-15 20:30:11 +0000	[diff] [blame]	5470	floating point or vector of floating point type. Not all targets support all
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	5471	types however.</p>
Chris Lattner	a4d7414	2005-07-21 01:29:16 +0000	[diff] [blame]	5472	<pre>
Dale Johannesen	408f9c1	2007-10-02 17:47:38 +0000	[diff] [blame]	5473	declare float @llvm.sqrt.f32(float %Val)
				5474	declare double @llvm.sqrt.f64(double %Val)
				5475	declare x86_fp80 @llvm.sqrt.f80(x86_fp80 %Val)
				5476	declare fp128 @llvm.sqrt.f128(fp128 %Val)
				5477	declare ppc_fp128 @llvm.sqrt.ppcf128(ppc_fp128 %Val)
Chris Lattner	a4d7414	2005-07-21 01:29:16 +0000	[diff] [blame]	5478	</pre>
				5479
				5480	<h5>Overview:</h5>
				5481
				5482	<p>
Reid Spencer	0b11820	2006-01-16 21:12:35 +0000	[diff] [blame]	5483	The '<tt>llvm.sqrt</tt>' intrinsics return the sqrt of the specified operand,
Dan Gohman	91c284c	2007-10-15 20:30:11 +0000	[diff] [blame]	5484	returning the same value as the libm '<tt>sqrt</tt>' functions would. Unlike
Chris Lattner	a4d7414	2005-07-21 01:29:16 +0000	[diff] [blame]	5485	<tt>sqrt</tt> in libm, however, <tt>llvm.sqrt</tt> has undefined behavior for
Chris Lattner	103e2d7	2008-01-29 07:00:44 +0000	[diff] [blame]	5486	negative numbers other than -0.0 (which allows for better optimization, because
				5487	there is no need to worry about errno being set). <tt>llvm.sqrt(-0.0)</tt> is
				5488	defined to return -0.0 like IEEE sqrt.
Chris Lattner	a4d7414	2005-07-21 01:29:16 +0000	[diff] [blame]	5489	</p>
				5490
				5491	<h5>Arguments:</h5>
				5492
				5493	<p>
				5494	The argument and return value are floating point numbers of the same type.
				5495	</p>
				5496
				5497	<h5>Semantics:</h5>
				5498
				5499	<p>
Dan Gohman	d6257fe	2007-07-16 14:37:41 +0000	[diff] [blame]	5500	This function returns the sqrt of the specified operand if it is a nonnegative
Chris Lattner	a4d7414	2005-07-21 01:29:16 +0000	[diff] [blame]	5501	floating point number.
				5502	</p>
				5503	</div>
				5504
Chris Lattner	f4d252d	2006-09-08 06:34:02 +0000	[diff] [blame]	5505	<!-- _______________________________________________________________________ -->
				5506	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5507	<a name="int_powi">'<tt>llvm.powi.*</tt>' Intrinsic</a>
Chris Lattner	f4d252d	2006-09-08 06:34:02 +0000	[diff] [blame]	5508	</div>
				5509
				5510	<div class="doc_text">
				5511
				5512	<h5>Syntax:</h5>
Dale Johannesen	408f9c1	2007-10-02 17:47:38 +0000	[diff] [blame]	5513	<p>This is an overloaded intrinsic. You can use <tt>llvm.powi</tt> on any
Dan Gohman	91c284c	2007-10-15 20:30:11 +0000	[diff] [blame]	5514	floating point or vector of floating point type. Not all targets support all
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	5515	types however.</p>
Chris Lattner	f4d252d	2006-09-08 06:34:02 +0000	[diff] [blame]	5516	<pre>
Dale Johannesen	408f9c1	2007-10-02 17:47:38 +0000	[diff] [blame]	5517	declare float @llvm.powi.f32(float %Val, i32 %power)
				5518	declare double @llvm.powi.f64(double %Val, i32 %power)
				5519	declare x86_fp80 @llvm.powi.f80(x86_fp80 %Val, i32 %power)
				5520	declare fp128 @llvm.powi.f128(fp128 %Val, i32 %power)
				5521	declare ppc_fp128 @llvm.powi.ppcf128(ppc_fp128 %Val, i32 %power)
Chris Lattner	f4d252d	2006-09-08 06:34:02 +0000	[diff] [blame]	5522	</pre>
				5523
				5524	<h5>Overview:</h5>
				5525
				5526	<p>
				5527	The '<tt>llvm.powi.*</tt>' intrinsics return the first operand raised to the
				5528	specified (positive or negative) power. The order of evaluation of
Dan Gohman	91c284c	2007-10-15 20:30:11 +0000	[diff] [blame]	5529	multiplications is not defined. When a vector of floating point type is
				5530	used, the second argument remains a scalar integer value.
Chris Lattner	f4d252d	2006-09-08 06:34:02 +0000	[diff] [blame]	5531	</p>
				5532
				5533	<h5>Arguments:</h5>
				5534
				5535	<p>
				5536	The second argument is an integer power, and the first is a value to raise to
				5537	that power.
				5538	</p>
				5539
				5540	<h5>Semantics:</h5>
				5541
				5542	<p>
				5543	This function returns the first value raised to the second power with an
				5544	unspecified sequence of rounding operations.</p>
				5545	</div>
				5546
Dan Gohman	91c284c	2007-10-15 20:30:11 +0000	[diff] [blame]	5547	<!-- _______________________________________________________________________ -->
				5548	<div class="doc_subsubsection">
				5549	<a name="int_sin">'<tt>llvm.sin.*</tt>' Intrinsic</a>
				5550	</div>
				5551
				5552	<div class="doc_text">
				5553
				5554	<h5>Syntax:</h5>
				5555	<p>This is an overloaded intrinsic. You can use <tt>llvm.sin</tt> on any
				5556	floating point or vector of floating point type. Not all targets support all
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	5557	types however.</p>
Dan Gohman	91c284c	2007-10-15 20:30:11 +0000	[diff] [blame]	5558	<pre>
				5559	declare float @llvm.sin.f32(float %Val)
				5560	declare double @llvm.sin.f64(double %Val)
				5561	declare x86_fp80 @llvm.sin.f80(x86_fp80 %Val)
				5562	declare fp128 @llvm.sin.f128(fp128 %Val)
				5563	declare ppc_fp128 @llvm.sin.ppcf128(ppc_fp128 %Val)
				5564	</pre>
				5565
				5566	<h5>Overview:</h5>
				5567
				5568	<p>
				5569	The '<tt>llvm.sin.*</tt>' intrinsics return the sine of the operand.
				5570	</p>
				5571
				5572	<h5>Arguments:</h5>
				5573
				5574	<p>
				5575	The argument and return value are floating point numbers of the same type.
				5576	</p>
				5577
				5578	<h5>Semantics:</h5>
				5579
				5580	<p>
				5581	This function returns the sine of the specified operand, returning the
				5582	same values as the libm <tt>sin</tt> functions would, and handles error
Dan Gohman	ba83b7e	2007-10-17 18:05:13 +0000	[diff] [blame]	5583	conditions in the same way.</p>
Dan Gohman	91c284c	2007-10-15 20:30:11 +0000	[diff] [blame]	5584	</div>
				5585
				5586	<!-- _______________________________________________________________________ -->
				5587	<div class="doc_subsubsection">
				5588	<a name="int_cos">'<tt>llvm.cos.*</tt>' Intrinsic</a>
				5589	</div>
				5590
				5591	<div class="doc_text">
				5592
				5593	<h5>Syntax:</h5>
				5594	<p>This is an overloaded intrinsic. You can use <tt>llvm.cos</tt> on any
				5595	floating point or vector of floating point type. Not all targets support all
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	5596	types however.</p>
Dan Gohman	91c284c	2007-10-15 20:30:11 +0000	[diff] [blame]	5597	<pre>
				5598	declare float @llvm.cos.f32(float %Val)
				5599	declare double @llvm.cos.f64(double %Val)
				5600	declare x86_fp80 @llvm.cos.f80(x86_fp80 %Val)
				5601	declare fp128 @llvm.cos.f128(fp128 %Val)
				5602	declare ppc_fp128 @llvm.cos.ppcf128(ppc_fp128 %Val)
				5603	</pre>
				5604
				5605	<h5>Overview:</h5>
				5606
				5607	<p>
				5608	The '<tt>llvm.cos.*</tt>' intrinsics return the cosine of the operand.
				5609	</p>
				5610
				5611	<h5>Arguments:</h5>
				5612
				5613	<p>
				5614	The argument and return value are floating point numbers of the same type.
				5615	</p>
				5616
				5617	<h5>Semantics:</h5>
				5618
				5619	<p>
				5620	This function returns the cosine of the specified operand, returning the
				5621	same values as the libm <tt>cos</tt> functions would, and handles error
Dan Gohman	ba83b7e	2007-10-17 18:05:13 +0000	[diff] [blame]	5622	conditions in the same way.</p>
Dan Gohman	91c284c	2007-10-15 20:30:11 +0000	[diff] [blame]	5623	</div>
				5624
				5625	<!-- _______________________________________________________________________ -->
				5626	<div class="doc_subsubsection">
				5627	<a name="int_pow">'<tt>llvm.pow.*</tt>' Intrinsic</a>
				5628	</div>
				5629
				5630	<div class="doc_text">
				5631
				5632	<h5>Syntax:</h5>
				5633	<p>This is an overloaded intrinsic. You can use <tt>llvm.pow</tt> on any
				5634	floating point or vector of floating point type. Not all targets support all
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	5635	types however.</p>
Dan Gohman	91c284c	2007-10-15 20:30:11 +0000	[diff] [blame]	5636	<pre>
				5637	declare float @llvm.pow.f32(float %Val, float %Power)
				5638	declare double @llvm.pow.f64(double %Val, double %Power)
				5639	declare x86_fp80 @llvm.pow.f80(x86_fp80 %Val, x86_fp80 %Power)
				5640	declare fp128 @llvm.pow.f128(fp128 %Val, fp128 %Power)
				5641	declare ppc_fp128 @llvm.pow.ppcf128(ppc_fp128 %Val, ppc_fp128 Power)
				5642	</pre>
				5643
				5644	<h5>Overview:</h5>
				5645
				5646	<p>
				5647	The '<tt>llvm.pow.*</tt>' intrinsics return the first operand raised to the
				5648	specified (positive or negative) power.
				5649	</p>
				5650
				5651	<h5>Arguments:</h5>
				5652
				5653	<p>
				5654	The second argument is a floating point power, and the first is a value to
				5655	raise to that power.
				5656	</p>
				5657
				5658	<h5>Semantics:</h5>
				5659
				5660	<p>
				5661	This function returns the first value raised to the second power,
				5662	returning the
				5663	same values as the libm <tt>pow</tt> functions would, and handles error
Dan Gohman	ba83b7e	2007-10-17 18:05:13 +0000	[diff] [blame]	5664	conditions in the same way.</p>
Dan Gohman	91c284c	2007-10-15 20:30:11 +0000	[diff] [blame]	5665	</div>
				5666
Chris Lattner	f4d252d	2006-09-08 06:34:02 +0000	[diff] [blame]	5667
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	5668	<!-- ======================================================================= -->
				5669	<div class="doc_subsection">
Nate Begeman	7e36c47	2006-01-13 23:26:38 +0000	[diff] [blame]	5670	<a name="int_manip">Bit Manipulation Intrinsics</a>
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	5671	</div>
				5672
				5673	<div class="doc_text">
				5674	<p>
Nate Begeman	7e36c47	2006-01-13 23:26:38 +0000	[diff] [blame]	5675	LLVM provides intrinsics for a few important bit manipulation operations.
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	5676	These allow efficient code generation for some algorithms.
				5677	</p>
				5678
				5679	</div>
				5680
				5681	<!-- _______________________________________________________________________ -->
				5682	<div class="doc_subsubsection">
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5683	<a name="int_bswap">'<tt>llvm.bswap.*</tt>' Intrinsics</a>
Nate Begeman	7e36c47	2006-01-13 23:26:38 +0000	[diff] [blame]	5684	</div>
				5685
				5686	<div class="doc_text">
				5687
				5688	<h5>Syntax:</h5>
Reid Spencer	409e28f	2007-04-01 08:04:23 +0000	[diff] [blame]	5689	<p>This is an overloaded intrinsic function. You can use bswap on any integer
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	5690	type that is an even number of bytes (i.e. BitWidth % 16 == 0).</p>
Nate Begeman	7e36c47	2006-01-13 23:26:38 +0000	[diff] [blame]	5691	<pre>
Chandler Carruth	6994040	2007-08-04 01:51:18 +0000	[diff] [blame]	5692	declare i16 @llvm.bswap.i16(i16 <id>)
				5693	declare i32 @llvm.bswap.i32(i32 <id>)
				5694	declare i64 @llvm.bswap.i64(i64 <id>)
Nate Begeman	7e36c47	2006-01-13 23:26:38 +0000	[diff] [blame]	5695	</pre>
				5696
				5697	<h5>Overview:</h5>
				5698
				5699	<p>
Reid Spencer	338ea09	2007-04-02 02:25:19 +0000	[diff] [blame]	5700	The '<tt>llvm.bswap</tt>' family of intrinsics is used to byte swap integer
Reid Spencer	409e28f	2007-04-01 08:04:23 +0000	[diff] [blame]	5701	values with an even number of bytes (positive multiple of 16 bits). These are
				5702	useful for performing operations on data that is not in the target's native
				5703	byte order.
Nate Begeman	7e36c47	2006-01-13 23:26:38 +0000	[diff] [blame]	5704	</p>
				5705
				5706	<h5>Semantics:</h5>
				5707
				5708	<p>
Chandler Carruth	6994040	2007-08-04 01:51:18 +0000	[diff] [blame]	5709	The <tt>llvm.bswap.i16</tt> intrinsic returns an i16 value that has the high
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	5710	and low byte of the input i16 swapped. Similarly, the <tt>llvm.bswap.i32</tt>
				5711	intrinsic returns an i32 value that has the four bytes of the input i32
				5712	swapped, so that if the input bytes are numbered 0, 1, 2, 3 then the returned
Chandler Carruth	6994040	2007-08-04 01:51:18 +0000	[diff] [blame]	5713	i32 will have its bytes in 3, 2, 1, 0 order. The <tt>llvm.bswap.i48</tt>,
				5714	<tt>llvm.bswap.i64</tt> and other intrinsics extend this concept to
Reid Spencer	409e28f	2007-04-01 08:04:23 +0000	[diff] [blame]	5715	additional even-byte lengths (6 bytes, 8 bytes and more, respectively).
Nate Begeman	7e36c47	2006-01-13 23:26:38 +0000	[diff] [blame]	5716	</p>
				5717
				5718	</div>
				5719
				5720	<!-- _______________________________________________________________________ -->
				5721	<div class="doc_subsubsection">
Reid Spencer	0b11820	2006-01-16 21:12:35 +0000	[diff] [blame]	5722	<a name="int_ctpop">'<tt>llvm.ctpop.*</tt>' Intrinsic</a>
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	5723	</div>
				5724
				5725	<div class="doc_text">
				5726
				5727	<h5>Syntax:</h5>
Reid Spencer	409e28f	2007-04-01 08:04:23 +0000	[diff] [blame]	5728	<p>This is an overloaded intrinsic. You can use llvm.ctpop on any integer bit
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	5729	width. Not all targets support all bit widths however.</p>
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	5730	<pre>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	5731	declare i8 @llvm.ctpop.i8(i8 <src>)
Chandler Carruth	6994040	2007-08-04 01:51:18 +0000	[diff] [blame]	5732	declare i16 @llvm.ctpop.i16(i16 <src>)
Anton Korobeynikov	ec43a06	2007-03-22 00:02:17 +0000	[diff] [blame]	5733	declare i32 @llvm.ctpop.i32(i32 <src>)
Chandler Carruth	6994040	2007-08-04 01:51:18 +0000	[diff] [blame]	5734	declare i64 @llvm.ctpop.i64(i64 <src>)
				5735	declare i256 @llvm.ctpop.i256(i256 <src>)
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	5736	</pre>
				5737
				5738	<h5>Overview:</h5>
				5739
				5740	<p>
Chris Lattner	ec6cb61	2006-01-16 22:38:59 +0000	[diff] [blame]	5741	The '<tt>llvm.ctpop</tt>' family of intrinsics counts the number of bits set in a
				5742	value.
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	5743	</p>
				5744
				5745	<h5>Arguments:</h5>
				5746
				5747	<p>
Chris Lattner	cfe6b37	2005-05-07 01:46:40 +0000	[diff] [blame]	5748	The only argument is the value to be counted. The argument may be of any
Reid Spencer	a517338	2007-01-04 16:43:23 +0000	[diff] [blame]	5749	integer type. The return type must match the argument type.
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	5750	</p>
				5751
				5752	<h5>Semantics:</h5>
				5753
				5754	<p>
				5755	The '<tt>llvm.ctpop</tt>' intrinsic counts the 1's in a variable.
				5756	</p>
				5757	</div>
				5758
				5759	<!-- _______________________________________________________________________ -->
				5760	<div class="doc_subsubsection">
Chris Lattner	8a886be	2006-01-16 22:34:14 +0000	[diff] [blame]	5761	<a name="int_ctlz">'<tt>llvm.ctlz.*</tt>' Intrinsic</a>
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	5762	</div>
				5763
				5764	<div class="doc_text">
				5765
				5766	<h5>Syntax:</h5>
Reid Spencer	409e28f	2007-04-01 08:04:23 +0000	[diff] [blame]	5767	<p>This is an overloaded intrinsic. You can use <tt>llvm.ctlz</tt> on any
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	5768	integer bit width. Not all targets support all bit widths however.</p>
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	5769	<pre>
Chandler Carruth	6994040	2007-08-04 01:51:18 +0000	[diff] [blame]	5770	declare i8 @llvm.ctlz.i8 (i8 <src>)
				5771	declare i16 @llvm.ctlz.i16(i16 <src>)
Anton Korobeynikov	ec43a06	2007-03-22 00:02:17 +0000	[diff] [blame]	5772	declare i32 @llvm.ctlz.i32(i32 <src>)
Chandler Carruth	6994040	2007-08-04 01:51:18 +0000	[diff] [blame]	5773	declare i64 @llvm.ctlz.i64(i64 <src>)
				5774	declare i256 @llvm.ctlz.i256(i256 <src>)
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	5775	</pre>
				5776
				5777	<h5>Overview:</h5>
				5778
				5779	<p>
Reid Spencer	0b11820	2006-01-16 21:12:35 +0000	[diff] [blame]	5780	The '<tt>llvm.ctlz</tt>' family of intrinsic functions counts the number of
				5781	leading zeros in a variable.
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	5782	</p>
				5783
				5784	<h5>Arguments:</h5>
				5785
				5786	<p>
Chris Lattner	cfe6b37	2005-05-07 01:46:40 +0000	[diff] [blame]	5787	The only argument is the value to be counted. The argument may be of any
Reid Spencer	a517338	2007-01-04 16:43:23 +0000	[diff] [blame]	5788	integer type. The return type must match the argument type.
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	5789	</p>
				5790
				5791	<h5>Semantics:</h5>
				5792
				5793	<p>
Chris Lattner	eff29ab	2005-05-15 19:39:26 +0000	[diff] [blame]	5794	The '<tt>llvm.ctlz</tt>' intrinsic counts the leading (most significant) zeros
				5795	in a variable. If the src == 0 then the result is the size in bits of the type
Reid Spencer	ca86e16	2006-12-31 07:07:53 +0000	[diff] [blame]	5796	of src. For example, <tt>llvm.ctlz(i32 2) = 30</tt>.
Andrew Lenharth	ec370fd	2005-05-03 18:01:48 +0000	[diff] [blame]	5797	</p>
				5798	</div>
Chris Lattner	3200628	2004-06-11 02:28:03 +0000	[diff] [blame]	5799
				5800
Chris Lattner	eff29ab	2005-05-15 19:39:26 +0000	[diff] [blame]	5801
				5802	<!-- _______________________________________________________________________ -->
				5803	<div class="doc_subsubsection">
Chris Lattner	8a886be	2006-01-16 22:34:14 +0000	[diff] [blame]	5804	<a name="int_cttz">'<tt>llvm.cttz.*</tt>' Intrinsic</a>
Chris Lattner	eff29ab	2005-05-15 19:39:26 +0000	[diff] [blame]	5805	</div>
				5806
				5807	<div class="doc_text">
				5808
				5809	<h5>Syntax:</h5>
Reid Spencer	409e28f	2007-04-01 08:04:23 +0000	[diff] [blame]	5810	<p>This is an overloaded intrinsic. You can use <tt>llvm.cttz</tt> on any
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	5811	integer bit width. Not all targets support all bit widths however.</p>
Chris Lattner	eff29ab	2005-05-15 19:39:26 +0000	[diff] [blame]	5812	<pre>
Chandler Carruth	6994040	2007-08-04 01:51:18 +0000	[diff] [blame]	5813	declare i8 @llvm.cttz.i8 (i8 <src>)
				5814	declare i16 @llvm.cttz.i16(i16 <src>)
Anton Korobeynikov	ec43a06	2007-03-22 00:02:17 +0000	[diff] [blame]	5815	declare i32 @llvm.cttz.i32(i32 <src>)
Chandler Carruth	6994040	2007-08-04 01:51:18 +0000	[diff] [blame]	5816	declare i64 @llvm.cttz.i64(i64 <src>)
				5817	declare i256 @llvm.cttz.i256(i256 <src>)
Chris Lattner	eff29ab	2005-05-15 19:39:26 +0000	[diff] [blame]	5818	</pre>
				5819
				5820	<h5>Overview:</h5>
				5821
				5822	<p>
Reid Spencer	0b11820	2006-01-16 21:12:35 +0000	[diff] [blame]	5823	The '<tt>llvm.cttz</tt>' family of intrinsic functions counts the number of
				5824	trailing zeros.
Chris Lattner	eff29ab	2005-05-15 19:39:26 +0000	[diff] [blame]	5825	</p>
				5826
				5827	<h5>Arguments:</h5>
				5828
				5829	<p>
				5830	The only argument is the value to be counted. The argument may be of any
Reid Spencer	a517338	2007-01-04 16:43:23 +0000	[diff] [blame]	5831	integer type. The return type must match the argument type.
Chris Lattner	eff29ab	2005-05-15 19:39:26 +0000	[diff] [blame]	5832	</p>
				5833
				5834	<h5>Semantics:</h5>
				5835
				5836	<p>
				5837	The '<tt>llvm.cttz</tt>' intrinsic counts the trailing (least significant) zeros
				5838	in a variable. If the src == 0 then the result is the size in bits of the type
				5839	of src. For example, <tt>llvm.cttz(2) = 1</tt>.
				5840	</p>
				5841	</div>
				5842
Reid Spencer	497d93e	2007-04-01 08:27:01 +0000	[diff] [blame]	5843	<!-- _______________________________________________________________________ -->
				5844	<div class="doc_subsubsection">
Reid Spencer	beacf66	2007-04-10 02:51:31 +0000	[diff] [blame]	5845	<a name="int_part_select">'<tt>llvm.part.select.*</tt>' Intrinsic</a>
Reid Spencer	a13ba7d	2007-04-01 19:00:37 +0000	[diff] [blame]	5846	</div>
				5847
				5848	<div class="doc_text">
				5849
				5850	<h5>Syntax:</h5>
Reid Spencer	beacf66	2007-04-10 02:51:31 +0000	[diff] [blame]	5851	<p>This is an overloaded intrinsic. You can use <tt>llvm.part.select</tt>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	5852	on any integer bit width.</p>
Reid Spencer	a13ba7d	2007-04-01 19:00:37 +0000	[diff] [blame]	5853	<pre>
Chandler Carruth	6994040	2007-08-04 01:51:18 +0000	[diff] [blame]	5854	declare i17 @llvm.part.select.i17 (i17 %val, i32 %loBit, i32 %hiBit)
				5855	declare i29 @llvm.part.select.i29 (i29 %val, i32 %loBit, i32 %hiBit)
Reid Spencer	a13ba7d	2007-04-01 19:00:37 +0000	[diff] [blame]	5856	</pre>
				5857
				5858	<h5>Overview:</h5>
Reid Spencer	beacf66	2007-04-10 02:51:31 +0000	[diff] [blame]	5859	<p>The '<tt>llvm.part.select</tt>' family of intrinsic functions selects a
Reid Spencer	a13ba7d	2007-04-01 19:00:37 +0000	[diff] [blame]	5860	range of bits from an integer value and returns them in the same bit width as
				5861	the original value.</p>
				5862
				5863	<h5>Arguments:</h5>
				5864	<p>The first argument, <tt>%val</tt> and the result may be integer types of
				5865	any bit width but they must have the same bit width. The second and third
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5866	arguments must be <tt>i32</tt> type since they specify only a bit index.</p>
Reid Spencer	a13ba7d	2007-04-01 19:00:37 +0000	[diff] [blame]	5867
				5868	<h5>Semantics:</h5>
Reid Spencer	beacf66	2007-04-10 02:51:31 +0000	[diff] [blame]	5869	<p>The operation of the '<tt>llvm.part.select</tt>' intrinsic has two modes
Reid Spencer	a3e435f	2007-04-04 02:42:35 +0000	[diff] [blame]	5870	of operation: forwards and reverse. If <tt>%loBit</tt> is greater than
				5871	<tt>%hiBits</tt> then the intrinsic operates in reverse mode. Otherwise it
				5872	operates in forward mode.</p>
				5873	<p>In forward mode, this intrinsic is the equivalent of shifting <tt>%val</tt>
				5874	right by <tt>%loBit</tt> bits and then ANDing it with a mask with
Reid Spencer	a13ba7d	2007-04-01 19:00:37 +0000	[diff] [blame]	5875	only the <tt>%hiBit - %loBit</tt> bits set, as follows:</p>
				5876	<ol>
				5877	<li>The <tt>%val</tt> is shifted right (LSHR) by the number of bits specified
				5878	by <tt>%loBits</tt>. This normalizes the value to the low order bits.</li>
				5879	<li>The <tt>%loBits</tt> value is subtracted from the <tt>%hiBits</tt> value
				5880	to determine the number of bits to retain.</li>
				5881	<li>A mask of the retained bits is created by shifting a -1 value.</li>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	5882	<li>The mask is ANDed with <tt>%val</tt> to produce the result.</li>
Reid Spencer	a13ba7d	2007-04-01 19:00:37 +0000	[diff] [blame]	5883	</ol>
Reid Spencer	d6a85b5	2007-05-14 16:14:57 +0000	[diff] [blame]	5884	<p>In reverse mode, a similar computation is made except that the bits are
				5885	returned in the reverse order. So, for example, if <tt>X</tt> has the value
				5886	<tt>i16 0x0ACF (101011001111)</tt> and we apply
				5887	<tt>part.select(i16 X, 8, 3)</tt> to it, we get back the value
				5888	<tt>i16 0x0026 (000000100110)</tt>.</p>
Reid Spencer	a13ba7d	2007-04-01 19:00:37 +0000	[diff] [blame]	5889	</div>
				5890
Reid Spencer	f86037f	2007-04-11 23:23:49 +0000	[diff] [blame]	5891	<div class="doc_subsubsection">
				5892	<a name="int_part_set">'<tt>llvm.part.set.*</tt>' Intrinsic</a>
				5893	</div>
				5894
				5895	<div class="doc_text">
				5896
				5897	<h5>Syntax:</h5>
				5898	<p>This is an overloaded intrinsic. You can use <tt>llvm.part.set</tt>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	5899	on any integer bit width.</p>
Reid Spencer	f86037f	2007-04-11 23:23:49 +0000	[diff] [blame]	5900	<pre>
Chandler Carruth	6994040	2007-08-04 01:51:18 +0000	[diff] [blame]	5901	declare i17 @llvm.part.set.i17.i9 (i17 %val, i9 %repl, i32 %lo, i32 %hi)
				5902	declare i29 @llvm.part.set.i29.i9 (i29 %val, i9 %repl, i32 %lo, i32 %hi)
Reid Spencer	f86037f	2007-04-11 23:23:49 +0000	[diff] [blame]	5903	</pre>
				5904
				5905	<h5>Overview:</h5>
				5906	<p>The '<tt>llvm.part.set</tt>' family of intrinsic functions replaces a range
				5907	of bits in an integer value with another integer value. It returns the integer
				5908	with the replaced bits.</p>
				5909
				5910	<h5>Arguments:</h5>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	5911	<p>The first argument, <tt>%val</tt>, and the result may be integer types of
				5912	any bit width, but they must have the same bit width. <tt>%val</tt> is the value
Reid Spencer	f86037f	2007-04-11 23:23:49 +0000	[diff] [blame]	5913	whose bits will be replaced. The second argument, <tt>%repl</tt> may be an
				5914	integer of any bit width. The third and fourth arguments must be <tt>i32</tt>
				5915	type since they specify only a bit index.</p>
				5916
				5917	<h5>Semantics:</h5>
				5918	<p>The operation of the '<tt>llvm.part.set</tt>' intrinsic has two modes
				5919	of operation: forwards and reverse. If <tt>%lo</tt> is greater than
				5920	<tt>%hi</tt> then the intrinsic operates in reverse mode. Otherwise it
				5921	operates in forward mode.</p>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	5922
Reid Spencer	f86037f	2007-04-11 23:23:49 +0000	[diff] [blame]	5923	<p>For both modes, the <tt>%repl</tt> value is prepared for use by either
				5924	truncating it down to the size of the replacement area or zero extending it
				5925	up to that size.</p>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	5926
Reid Spencer	f86037f	2007-04-11 23:23:49 +0000	[diff] [blame]	5927	<p>In forward mode, the bits between <tt>%lo</tt> and <tt>%hi</tt> (inclusive)
				5928	are replaced with corresponding bits from <tt>%repl</tt>. That is the 0th bit
				5929	in <tt>%repl</tt> replaces the <tt>%lo</tt>th bit in <tt>%val</tt> and etc. up
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	5930	to the <tt>%hi</tt>th bit.</p>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	5931
Reid Spencer	c6749c4	2007-05-14 16:50:20 +0000	[diff] [blame]	5932	<p>In reverse mode, a similar computation is made except that the bits are
				5933	reversed. That is, the <tt>0</tt>th bit in <tt>%repl</tt> replaces the
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	5934	<tt>%hi</tt> bit in <tt>%val</tt> and etc. down to the <tt>%lo</tt>th bit.</p>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	5935
Reid Spencer	f86037f	2007-04-11 23:23:49 +0000	[diff] [blame]	5936	<h5>Examples:</h5>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	5937
Reid Spencer	f86037f	2007-04-11 23:23:49 +0000	[diff] [blame]	5938	<pre>
Reid Spencer	f0dbf64	2007-04-12 01:03:03 +0000	[diff] [blame]	5939	llvm.part.set(0xFFFF, 0, 4, 7) -> 0xFF0F
Reid Spencer	c6749c4	2007-05-14 16:50:20 +0000	[diff] [blame]	5940	llvm.part.set(0xFFFF, 0, 7, 4) -> 0xFF0F
				5941	llvm.part.set(0xFFFF, 1, 7, 4) -> 0xFF8F
				5942	llvm.part.set(0xFFFF, F, 8, 3) -> 0xFFE7
Reid Spencer	f0dbf64	2007-04-12 01:03:03 +0000	[diff] [blame]	5943	llvm.part.set(0xFFFF, 0, 3, 8) -> 0xFE07
Reid Spencer	c891084	2007-04-11 23:49:50 +0000	[diff] [blame]	5944	</pre>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	5945
				5946	</div>
				5947
Bill Wendling	da01af7	2009-02-08 04:04:40 +0000	[diff] [blame]	5948	<!-- ======================================================================= -->
				5949	<div class="doc_subsection">
				5950	<a name="int_overflow">Arithmetic with Overflow Intrinsics</a>
				5951	</div>
				5952
				5953	<div class="doc_text">
				5954	<p>
				5955	LLVM provides intrinsics for some arithmetic with overflow operations.
				5956	</p>
				5957
				5958	</div>
				5959
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	5960	<!-- _______________________________________________________________________ -->
				5961	<div class="doc_subsubsection">
Bill Wendling	da01af7	2009-02-08 04:04:40 +0000	[diff] [blame]	5962	<a name="int_sadd_overflow">'<tt>llvm.sadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	5963	</div>
				5964
				5965	<div class="doc_text">
				5966
				5967	<h5>Syntax:</h5>
				5968
				5969	<p>This is an overloaded intrinsic. You can use <tt>llvm.sadd.with.overflow</tt>
Bill Wendling	da01af7	2009-02-08 04:04:40 +0000	[diff] [blame]	5970	on any integer bit width.</p>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	5971
				5972	<pre>
				5973	declare {i16, i1} @llvm.sadd.with.overflow.i16(i16 %a, i16 %b)
				5974	declare {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
				5975	declare {i64, i1} @llvm.sadd.with.overflow.i64(i64 %a, i64 %b)
				5976	</pre>
				5977
				5978	<h5>Overview:</h5>
				5979
				5980	<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
				5981	a signed addition of the two arguments, and indicate whether an overflow
				5982	occurred during the signed summation.</p>
				5983
				5984	<h5>Arguments:</h5>
				5985
				5986	<p>The arguments (%a and %b) and the first element of the result structure may
				5987	be of integer types of any bit width, but they must have the same bit width. The
				5988	second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
				5989	and <tt>%b</tt> are the two values that will undergo signed addition.</p>
				5990
				5991	<h5>Semantics:</h5>
				5992
				5993	<p>The '<tt>llvm.sadd.with.overflow</tt>' family of intrinsic functions perform
				5994	a signed addition of the two variables. They return a structure — the
				5995	first element of which is the signed summation, and the second element of which
				5996	is a bit specifying if the signed summation resulted in an overflow.</p>
				5997
				5998	<h5>Examples:</h5>
				5999	<pre>
				6000	%res = call {i32, i1} @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
				6001	%sum = extractvalue {i32, i1} %res, 0
				6002	%obit = extractvalue {i32, i1} %res, 1
				6003	br i1 %obit, label %overflow, label %normal
				6004	</pre>
				6005
				6006	</div>
				6007
				6008	<!-- _______________________________________________________________________ -->
				6009	<div class="doc_subsubsection">
Bill Wendling	da01af7	2009-02-08 04:04:40 +0000	[diff] [blame]	6010	<a name="int_uadd_overflow">'<tt>llvm.uadd.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	6011	</div>
				6012
				6013	<div class="doc_text">
				6014
				6015	<h5>Syntax:</h5>
				6016
				6017	<p>This is an overloaded intrinsic. You can use <tt>llvm.uadd.with.overflow</tt>
Bill Wendling	da01af7	2009-02-08 04:04:40 +0000	[diff] [blame]	6018	on any integer bit width.</p>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	6019
				6020	<pre>
				6021	declare {i16, i1} @llvm.uadd.with.overflow.i16(i16 %a, i16 %b)
				6022	declare {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
				6023	declare {i64, i1} @llvm.uadd.with.overflow.i64(i64 %a, i64 %b)
				6024	</pre>
				6025
				6026	<h5>Overview:</h5>
				6027
				6028	<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
				6029	an unsigned addition of the two arguments, and indicate whether a carry occurred
				6030	during the unsigned summation.</p>
				6031
				6032	<h5>Arguments:</h5>
				6033
				6034	<p>The arguments (%a and %b) and the first element of the result structure may
				6035	be of integer types of any bit width, but they must have the same bit width. The
				6036	second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
				6037	and <tt>%b</tt> are the two values that will undergo unsigned addition.</p>
				6038
				6039	<h5>Semantics:</h5>
				6040
				6041	<p>The '<tt>llvm.uadd.with.overflow</tt>' family of intrinsic functions perform
				6042	an unsigned addition of the two arguments. They return a structure — the
				6043	first element of which is the sum, and the second element of which is a bit
				6044	specifying if the unsigned summation resulted in a carry.</p>
				6045
				6046	<h5>Examples:</h5>
				6047	<pre>
				6048	%res = call {i32, i1} @llvm.uadd.with.overflow.i32(i32 %a, i32 %b)
				6049	%sum = extractvalue {i32, i1} %res, 0
				6050	%obit = extractvalue {i32, i1} %res, 1
				6051	br i1 %obit, label %carry, label %normal
				6052	</pre>
				6053
				6054	</div>
				6055
				6056	<!-- _______________________________________________________________________ -->
				6057	<div class="doc_subsubsection">
Bill Wendling	da01af7	2009-02-08 04:04:40 +0000	[diff] [blame]	6058	<a name="int_ssub_overflow">'<tt>llvm.ssub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	6059	</div>
				6060
				6061	<div class="doc_text">
				6062
				6063	<h5>Syntax:</h5>
				6064
				6065	<p>This is an overloaded intrinsic. You can use <tt>llvm.ssub.with.overflow</tt>
Bill Wendling	da01af7	2009-02-08 04:04:40 +0000	[diff] [blame]	6066	on any integer bit width.</p>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	6067
				6068	<pre>
				6069	declare {i16, i1} @llvm.ssub.with.overflow.i16(i16 %a, i16 %b)
				6070	declare {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
				6071	declare {i64, i1} @llvm.ssub.with.overflow.i64(i64 %a, i64 %b)
				6072	</pre>
				6073
				6074	<h5>Overview:</h5>
				6075
				6076	<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
				6077	a signed subtraction of the two arguments, and indicate whether an overflow
				6078	occurred during the signed subtraction.</p>
				6079
				6080	<h5>Arguments:</h5>
				6081
				6082	<p>The arguments (%a and %b) and the first element of the result structure may
				6083	be of integer types of any bit width, but they must have the same bit width. The
				6084	second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
				6085	and <tt>%b</tt> are the two values that will undergo signed subtraction.</p>
				6086
				6087	<h5>Semantics:</h5>
				6088
				6089	<p>The '<tt>llvm.ssub.with.overflow</tt>' family of intrinsic functions perform
				6090	a signed subtraction of the two arguments. They return a structure — the
				6091	first element of which is the subtraction, and the second element of which is a bit
				6092	specifying if the signed subtraction resulted in an overflow.</p>
				6093
				6094	<h5>Examples:</h5>
				6095	<pre>
				6096	%res = call {i32, i1} @llvm.ssub.with.overflow.i32(i32 %a, i32 %b)
				6097	%sum = extractvalue {i32, i1} %res, 0
				6098	%obit = extractvalue {i32, i1} %res, 1
				6099	br i1 %obit, label %overflow, label %normal
				6100	</pre>
				6101
				6102	</div>
				6103
				6104	<!-- _______________________________________________________________________ -->
				6105	<div class="doc_subsubsection">
Bill Wendling	da01af7	2009-02-08 04:04:40 +0000	[diff] [blame]	6106	<a name="int_usub_overflow">'<tt>llvm.usub.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	6107	</div>
				6108
				6109	<div class="doc_text">
				6110
				6111	<h5>Syntax:</h5>
				6112
				6113	<p>This is an overloaded intrinsic. You can use <tt>llvm.usub.with.overflow</tt>
Bill Wendling	da01af7	2009-02-08 04:04:40 +0000	[diff] [blame]	6114	on any integer bit width.</p>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	6115
				6116	<pre>
				6117	declare {i16, i1} @llvm.usub.with.overflow.i16(i16 %a, i16 %b)
				6118	declare {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
				6119	declare {i64, i1} @llvm.usub.with.overflow.i64(i64 %a, i64 %b)
				6120	</pre>
				6121
				6122	<h5>Overview:</h5>
				6123
				6124	<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
				6125	an unsigned subtraction of the two arguments, and indicate whether an overflow
				6126	occurred during the unsigned subtraction.</p>
				6127
				6128	<h5>Arguments:</h5>
				6129
				6130	<p>The arguments (%a and %b) and the first element of the result structure may
				6131	be of integer types of any bit width, but they must have the same bit width. The
				6132	second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
				6133	and <tt>%b</tt> are the two values that will undergo unsigned subtraction.</p>
				6134
				6135	<h5>Semantics:</h5>
				6136
				6137	<p>The '<tt>llvm.usub.with.overflow</tt>' family of intrinsic functions perform
				6138	an unsigned subtraction of the two arguments. They return a structure — the
				6139	first element of which is the subtraction, and the second element of which is a bit
				6140	specifying if the unsigned subtraction resulted in an overflow.</p>
				6141
				6142	<h5>Examples:</h5>
				6143	<pre>
				6144	%res = call {i32, i1} @llvm.usub.with.overflow.i32(i32 %a, i32 %b)
				6145	%sum = extractvalue {i32, i1} %res, 0
				6146	%obit = extractvalue {i32, i1} %res, 1
				6147	br i1 %obit, label %overflow, label %normal
				6148	</pre>
				6149
				6150	</div>
				6151
				6152	<!-- _______________________________________________________________________ -->
				6153	<div class="doc_subsubsection">
Bill Wendling	da01af7	2009-02-08 04:04:40 +0000	[diff] [blame]	6154	<a name="int_smul_overflow">'<tt>llvm.smul.with.overflow.*</tt>' Intrinsics</a>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	6155	</div>
				6156
				6157	<div class="doc_text">
				6158
				6159	<h5>Syntax:</h5>
				6160
				6161	<p>This is an overloaded intrinsic. You can use <tt>llvm.smul.with.overflow</tt>
Bill Wendling	da01af7	2009-02-08 04:04:40 +0000	[diff] [blame]	6162	on any integer bit width.</p>
Bill Wendling	ac1df8e	2009-02-08 01:40:31 +0000	[diff] [blame]	6163
				6164	<pre>
				6165	declare {i16, i1} @llvm.smul.with.overflow.i16(i16 %a, i16 %b)
				6166	declare {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
				6167	declare {i64, i1} @llvm.smul.with.overflow.i64(i64 %a, i64 %b)
				6168	</pre>
				6169
				6170	<h5>Overview:</h5>
				6171
				6172	<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
				6173	a signed multiplication of the two arguments, and indicate whether an overflow
				6174	occurred during the signed multiplication.</p>
				6175
				6176	<h5>Arguments:</h5>
				6177
				6178	<p>The arguments (%a and %b) and the first element of the result structure may
				6179	be of integer types of any bit width, but they must have the same bit width. The
				6180	second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
				6181	and <tt>%b</tt> are the two values that will undergo signed multiplication.</p>
				6182
				6183	<h5>Semantics:</h5>
				6184
				6185	<p>The '<tt>llvm.smul.with.overflow</tt>' family of intrinsic functions perform
				6186	a signed multiplication of the two arguments. They return a structure —
				6187	the first element of which is the multiplication, and the second element of
				6188	which is a bit specifying if the signed multiplication resulted in an
				6189	overflow.</p>
				6190
				6191	<h5>Examples:</h5>
				6192	<pre>
				6193	%res = call {i32, i1} @llvm.smul.with.overflow.i32(i32 %a, i32 %b)
				6194	%sum = extractvalue {i32, i1} %res, 0
				6195	%obit = extractvalue {i32, i1} %res, 1
				6196	br i1 %obit, label %overflow, label %normal
				6197	</pre>
				6198
Reid Spencer	f86037f	2007-04-11 23:23:49 +0000	[diff] [blame]	6199	</div>
				6200
Bill Wendling	41b485c	2009-02-08 23:00:09 +0000	[diff] [blame]	6201	<!-- _______________________________________________________________________ -->
				6202	<div class="doc_subsubsection">
				6203	<a name="int_umul_overflow">'<tt>llvm.umul.with.overflow.*</tt>' Intrinsics</a>
				6204	</div>
				6205
				6206	<div class="doc_text">
				6207
				6208	<h5>Syntax:</h5>
				6209
				6210	<p>This is an overloaded intrinsic. You can use <tt>llvm.umul.with.overflow</tt>
				6211	on any integer bit width.</p>
				6212
				6213	<pre>
				6214	declare {i16, i1} @llvm.umul.with.overflow.i16(i16 %a, i16 %b)
				6215	declare {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
				6216	declare {i64, i1} @llvm.umul.with.overflow.i64(i64 %a, i64 %b)
				6217	</pre>
				6218
				6219	<h5>Overview:</h5>
				6220
				6221	<p><i><b>Warning:</b> '<tt>llvm.umul.with.overflow</tt>' is badly broken. It is
				6222	actively being fixed, but it should not currently be used!</i></p>
				6223
				6224	<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
				6225	a unsigned multiplication of the two arguments, and indicate whether an overflow
				6226	occurred during the unsigned multiplication.</p>
				6227
				6228	<h5>Arguments:</h5>
				6229
				6230	<p>The arguments (%a and %b) and the first element of the result structure may
				6231	be of integer types of any bit width, but they must have the same bit width. The
				6232	second element of the result structure must be of type <tt>i1</tt>. <tt>%a</tt>
				6233	and <tt>%b</tt> are the two values that will undergo unsigned
				6234	multiplication.</p>
				6235
				6236	<h5>Semantics:</h5>
				6237
				6238	<p>The '<tt>llvm.umul.with.overflow</tt>' family of intrinsic functions perform
				6239	an unsigned multiplication of the two arguments. They return a structure —
				6240	the first element of which is the multiplication, and the second element of
				6241	which is a bit specifying if the unsigned multiplication resulted in an
				6242	overflow.</p>
				6243
				6244	<h5>Examples:</h5>
				6245	<pre>
				6246	%res = call {i32, i1} @llvm.umul.with.overflow.i32(i32 %a, i32 %b)
				6247	%sum = extractvalue {i32, i1} %res, 0
				6248	%obit = extractvalue {i32, i1} %res, 1
				6249	br i1 %obit, label %overflow, label %normal
				6250	</pre>
				6251
				6252	</div>
				6253
Chris Lattner	8ff7590	2004-01-06 05:31:32 +0000	[diff] [blame]	6254	<!-- ======================================================================= -->
				6255	<div class="doc_subsection">
				6256	<a name="int_debugger">Debugger Intrinsics</a>
				6257	</div>
				6258
				6259	<div class="doc_text">
				6260	<p>
				6261	The LLVM debugger intrinsics (which all start with <tt>llvm.dbg.</tt> prefix),
				6262	are described in the <a
				6263	href="SourceLevelDebugging.html#format_common_intrinsics">LLVM Source Level
				6264	Debugging</a> document.
				6265	</p>
				6266	</div>
				6267
				6268
Jim Laskey	dd4ef1b	2007-03-14 19:31:19 +0000	[diff] [blame]	6269	<!-- ======================================================================= -->
				6270	<div class="doc_subsection">
				6271	<a name="int_eh">Exception Handling Intrinsics</a>
				6272	</div>
				6273
				6274	<div class="doc_text">
				6275	<p> The LLVM exception handling intrinsics (which all start with
				6276	<tt>llvm.eh.</tt> prefix), are described in the <a
				6277	href="ExceptionHandling.html#format_common_intrinsics">LLVM Exception
				6278	Handling</a> document. </p>
				6279	</div>
				6280
Tanya Lattner	6d806e9	2007-06-15 20:50:54 +0000	[diff] [blame]	6281	<!-- ======================================================================= -->
				6282	<div class="doc_subsection">
Duncan Sands	f7331b3	2007-09-11 14:10:23 +0000	[diff] [blame]	6283	<a name="int_trampoline">Trampoline Intrinsic</a>
Duncan Sands	36397f5	2007-07-27 12:58:54 +0000	[diff] [blame]	6284	</div>
				6285
				6286	<div class="doc_text">
				6287	<p>
Duncan Sands	f7331b3	2007-09-11 14:10:23 +0000	[diff] [blame]	6288	This intrinsic makes it possible to excise one parameter, marked with
Duncan Sands	36397f5	2007-07-27 12:58:54 +0000	[diff] [blame]	6289	the <tt>nest</tt> attribute, from a function. The result is a callable
				6290	function pointer lacking the nest parameter - the caller does not need
				6291	to provide a value for it. Instead, the value to use is stored in
				6292	advance in a "trampoline", a block of memory usually allocated
				6293	on the stack, which also contains code to splice the nest value into the
				6294	argument list. This is used to implement the GCC nested function address
				6295	extension.
				6296	</p>
				6297	<p>
				6298	For example, if the function is
				6299	<tt>i32 f(i8* nest %c, i32 %x, i32 %y)</tt> then the resulting function
Bill Wendling	03295ca	2007-09-22 09:23:55 +0000	[diff] [blame]	6300	pointer has signature <tt>i32 (i32, i32)*</tt>. It can be created as follows:</p>
Duncan Sands	36397f5	2007-07-27 12:58:54 +0000	[diff] [blame]	6301	<pre>
Duncan Sands	f7331b3	2007-09-11 14:10:23 +0000	[diff] [blame]	6302	%tramp = alloca [10 x i8], align 4 ; size and alignment only correct for X86
				6303	%tramp1 = getelementptr [10 x i8]* %tramp, i32 0, i32 0
				6304	%p = call i8* @llvm.init.trampoline( i8* %tramp1, i8* bitcast (i32 (i8* nest , i32, i32)* @f to i8), i8 %nval )
				6305	%fp = bitcast i8* %p to i32 (i32, i32)*
Duncan Sands	36397f5	2007-07-27 12:58:54 +0000	[diff] [blame]	6306	</pre>
Bill Wendling	03295ca	2007-09-22 09:23:55 +0000	[diff] [blame]	6307	<p>The call <tt>%val = call i32 %fp( i32 %x, i32 %y )</tt> is then equivalent
				6308	to <tt>%val = call i32 %f( i8* %nval, i32 %x, i32 %y )</tt>.</p>
Duncan Sands	36397f5	2007-07-27 12:58:54 +0000	[diff] [blame]	6309	</div>
				6310
				6311	<!-- _______________________________________________________________________ -->
				6312	<div class="doc_subsubsection">
				6313	<a name="int_it">'<tt>llvm.init.trampoline</tt>' Intrinsic</a>
				6314	</div>
				6315	<div class="doc_text">
				6316	<h5>Syntax:</h5>
				6317	<pre>
Duncan Sands	f7331b3	2007-09-11 14:10:23 +0000	[diff] [blame]	6318	declare i8* @llvm.init.trampoline(i8* <tramp>, i8* <func>, i8* <nval>)
Duncan Sands	36397f5	2007-07-27 12:58:54 +0000	[diff] [blame]	6319	</pre>
				6320	<h5>Overview:</h5>
				6321	<p>
Duncan Sands	f7331b3	2007-09-11 14:10:23 +0000	[diff] [blame]	6322	This fills the memory pointed to by <tt>tramp</tt> with code
				6323	and returns a function pointer suitable for executing it.
Duncan Sands	36397f5	2007-07-27 12:58:54 +0000	[diff] [blame]	6324	</p>
				6325	<h5>Arguments:</h5>
				6326	<p>
				6327	The <tt>llvm.init.trampoline</tt> intrinsic takes three arguments, all
				6328	pointers. The <tt>tramp</tt> argument must point to a sufficiently large
				6329	and sufficiently aligned block of memory; this memory is written to by the
Duncan Sands	c00c2ba	2007-08-22 23:39:54 +0000	[diff] [blame]	6330	intrinsic. Note that the size and the alignment are target-specific - LLVM
				6331	currently provides no portable way of determining them, so a front-end that
				6332	generates this intrinsic needs to have some target-specific knowledge.
				6333	The <tt>func</tt> argument must hold a function bitcast to an <tt>i8*</tt>.
Duncan Sands	36397f5	2007-07-27 12:58:54 +0000	[diff] [blame]	6334	</p>
				6335	<h5>Semantics:</h5>
				6336	<p>
				6337	The block of memory pointed to by <tt>tramp</tt> is filled with target
Duncan Sands	f7331b3	2007-09-11 14:10:23 +0000	[diff] [blame]	6338	dependent code, turning it into a function. A pointer to this function is
				6339	returned, but needs to be bitcast to an
Duncan Sands	36397f5	2007-07-27 12:58:54 +0000	[diff] [blame]	6340	<a href="#int_trampoline">appropriate function pointer type</a>
Duncan Sands	f7331b3	2007-09-11 14:10:23 +0000	[diff] [blame]	6341	before being called. The new function's signature is the same as that of
				6342	<tt>func</tt> with any arguments marked with the <tt>nest</tt> attribute
				6343	removed. At most one such <tt>nest</tt> argument is allowed, and it must be
				6344	of pointer type. Calling the new function is equivalent to calling
				6345	<tt>func</tt> with the same argument list, but with <tt>nval</tt> used for the
				6346	missing <tt>nest</tt> argument. If, after calling
				6347	<tt>llvm.init.trampoline</tt>, the memory pointed to by <tt>tramp</tt> is
				6348	modified, then the effect of any later call to the returned function pointer is
				6349	undefined.
Duncan Sands	36397f5	2007-07-27 12:58:54 +0000	[diff] [blame]	6350	</p>
				6351	</div>
				6352
				6353	<!-- ======================================================================= -->
				6354	<div class="doc_subsection">
Andrew Lenharth	22c5c1b	2008-02-16 01:24:58 +0000	[diff] [blame]	6355	<a name="int_atomics">Atomic Operations and Synchronization Intrinsics</a>
				6356	</div>
				6357
				6358	<div class="doc_text">
				6359	<p>
				6360	These intrinsic functions expand the "universal IR" of LLVM to represent
				6361	hardware constructs for atomic operations and memory synchronization. This
				6362	provides an interface to the hardware, not an interface to the programmer. It
Chris Lattner	d3eda89	2008-08-05 18:29:16 +0000	[diff] [blame]	6363	is aimed at a low enough level to allow any programming models or APIs
				6364	(Application Programming Interfaces) which
Andrew Lenharth	22c5c1b	2008-02-16 01:24:58 +0000	[diff] [blame]	6365	need atomic behaviors to map cleanly onto it. It is also modeled primarily on
				6366	hardware behavior. Just as hardware provides a "universal IR" for source
				6367	languages, it also provides a starting point for developing a "universal"
				6368	atomic operation and synchronization IR.
				6369	</p>
				6370	<p>
				6371	These do <em>not</em> form an API such as high-level threading libraries,
				6372	software transaction memory systems, atomic primitives, and intrinsic
				6373	functions as found in BSD, GNU libc, atomic_ops, APR, and other system and
				6374	application libraries. The hardware interface provided by LLVM should allow
				6375	a clean implementation of all of these APIs and parallel programming models.
				6376	No one model or paradigm should be selected above others unless the hardware
				6377	itself ubiquitously does so.
				6378
				6379	</p>
				6380	</div>
				6381
				6382	<!-- _______________________________________________________________________ -->
				6383	<div class="doc_subsubsection">
				6384	<a name="int_memory_barrier">'<tt>llvm.memory.barrier</tt>' Intrinsic</a>
				6385	</div>
				6386	<div class="doc_text">
				6387	<h5>Syntax:</h5>
				6388	<pre>
				6389	declare void @llvm.memory.barrier( i1 <ll>, i1 <ls>, i1 <sl>, i1 <ss>,
				6390	i1 <device> )
				6391
				6392	</pre>
				6393	<h5>Overview:</h5>
				6394	<p>
				6395	The <tt>llvm.memory.barrier</tt> intrinsic guarantees ordering between
				6396	specific pairs of memory access types.
				6397	</p>
				6398	<h5>Arguments:</h5>
				6399	<p>
				6400	The <tt>llvm.memory.barrier</tt> intrinsic requires five boolean arguments.
				6401	The first four arguments enables a specific barrier as listed below. The fith
				6402	argument specifies that the barrier applies to io or device or uncached memory.
				6403
				6404	</p>
				6405	<ul>
				6406	<li><tt>ll</tt>: load-load barrier</li>
				6407	<li><tt>ls</tt>: load-store barrier</li>
				6408	<li><tt>sl</tt>: store-load barrier</li>
				6409	<li><tt>ss</tt>: store-store barrier</li>
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	6410	<li><tt>device</tt>: barrier applies to device and uncached memory also.</li>
Andrew Lenharth	22c5c1b	2008-02-16 01:24:58 +0000	[diff] [blame]	6411	</ul>
				6412	<h5>Semantics:</h5>
				6413	<p>
				6414	This intrinsic causes the system to enforce some ordering constraints upon
				6415	the loads and stores of the program. This barrier does not indicate
				6416	<em>when</em> any events will occur, it only enforces an <em>order</em> in
				6417	which they occur. For any of the specified pairs of load and store operations
				6418	(f.ex. load-load, or store-load), all of the first operations preceding the
				6419	barrier will complete before any of the second operations succeeding the
				6420	barrier begin. Specifically the semantics for each pairing is as follows:
				6421	</p>
				6422	<ul>
				6423	<li><tt>ll</tt>: All loads before the barrier must complete before any load
				6424	after the barrier begins.</li>
				6425
				6426	<li><tt>ls</tt>: All loads before the barrier must complete before any
				6427	store after the barrier begins.</li>
				6428	<li><tt>ss</tt>: All stores before the barrier must complete before any
				6429	store after the barrier begins.</li>
				6430	<li><tt>sl</tt>: All stores before the barrier must complete before any
				6431	load after the barrier begins.</li>
				6432	</ul>
				6433	<p>
				6434	These semantics are applied with a logical "and" behavior when more than one
				6435	is enabled in a single memory barrier intrinsic.
				6436	</p>
				6437	<p>
				6438	Backends may implement stronger barriers than those requested when they do not
				6439	support as fine grained a barrier as requested. Some architectures do not
				6440	need all types of barriers and on such architectures, these become noops.
				6441	</p>
				6442	<h5>Example:</h5>
				6443	<pre>
				6444	%ptr = malloc i32
				6445	store i32 4, %ptr
				6446
				6447	%result1 = load i32* %ptr <i>; yields {i32}:result1 = 4</i>
				6448	call void @llvm.memory.barrier( i1 false, i1 true, i1 false, i1 false )
				6449	<i>; guarantee the above finishes</i>
				6450	store i32 8, %ptr <i>; before this begins</i>
				6451	</pre>
				6452	</div>
				6453
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6454	<!-- _______________________________________________________________________ -->
				6455	<div class="doc_subsubsection">
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6456	<a name="int_atomic_cmp_swap">'<tt>llvm.atomic.cmp.swap.*</tt>' Intrinsic</a>
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6457	</div>
				6458	<div class="doc_text">
				6459	<h5>Syntax:</h5>
				6460	<p>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6461	This is an overloaded intrinsic. You can use <tt>llvm.atomic.cmp.swap</tt> on
				6462	any integer bit width and for different address spaces. Not all targets
				6463	support all bit widths however.</p>
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6464
				6465	<pre>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6466	declare i8 @llvm.atomic.cmp.swap.i8.p0i8( i8* <ptr>, i8 <cmp>, i8 <val> )
				6467	declare i16 @llvm.atomic.cmp.swap.i16.p0i16( i16* <ptr>, i16 <cmp>, i16 <val> )
				6468	declare i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* <ptr>, i32 <cmp>, i32 <val> )
				6469	declare i64 @llvm.atomic.cmp.swap.i64.p0i64( i64* <ptr>, i64 <cmp>, i64 <val> )
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6470
				6471	</pre>
				6472	<h5>Overview:</h5>
				6473	<p>
				6474	This loads a value in memory and compares it to a given value. If they are
				6475	equal, it stores a new value into the memory.
				6476	</p>
				6477	<h5>Arguments:</h5>
				6478	<p>
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6479	The <tt>llvm.atomic.cmp.swap</tt> intrinsic takes three arguments. The result as
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6480	well as both <tt>cmp</tt> and <tt>val</tt> must be integer values with the
				6481	same bit width. The <tt>ptr</tt> argument must be a pointer to a value of
				6482	this integer type. While any bit width integer may be used, targets may only
				6483	lower representations they support in hardware.
				6484
				6485	</p>
				6486	<h5>Semantics:</h5>
				6487	<p>
				6488	This entire intrinsic must be executed atomically. It first loads the value
				6489	in memory pointed to by <tt>ptr</tt> and compares it with the value
				6490	<tt>cmp</tt>. If they are equal, <tt>val</tt> is stored into the memory. The
				6491	loaded value is yielded in all cases. This provides the equivalent of an
				6492	atomic compare-and-swap operation within the SSA framework.
				6493	</p>
				6494	<h5>Examples:</h5>
				6495
				6496	<pre>
				6497	%ptr = malloc i32
				6498	store i32 4, %ptr
				6499
				6500	%val1 = add i32 4, 4
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6501	%result1 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 4, %val1 )
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6502	<i>; yields {i32}:result1 = 4</i>
				6503	%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
				6504	%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
				6505
				6506	%val2 = add i32 1, 1
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6507	%result2 = call i32 @llvm.atomic.cmp.swap.i32.p0i32( i32* %ptr, i32 5, %val2 )
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6508	<i>; yields {i32}:result2 = 8</i>
				6509	%stored2 = icmp eq i32 %result2, 5 <i>; yields {i1}:stored2 = false</i>
				6510
				6511	%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 8</i>
				6512	</pre>
				6513	</div>
				6514
				6515	<!-- _______________________________________________________________________ -->
				6516	<div class="doc_subsubsection">
				6517	<a name="int_atomic_swap">'<tt>llvm.atomic.swap.*</tt>' Intrinsic</a>
				6518	</div>
				6519	<div class="doc_text">
				6520	<h5>Syntax:</h5>
				6521
				6522	<p>
				6523	This is an overloaded intrinsic. You can use <tt>llvm.atomic.swap</tt> on any
				6524	integer bit width. Not all targets support all bit widths however.</p>
				6525	<pre>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6526	declare i8 @llvm.atomic.swap.i8.p0i8( i8* <ptr>, i8 <val> )
				6527	declare i16 @llvm.atomic.swap.i16.p0i16( i16* <ptr>, i16 <val> )
				6528	declare i32 @llvm.atomic.swap.i32.p0i32( i32* <ptr>, i32 <val> )
				6529	declare i64 @llvm.atomic.swap.i64.p0i64( i64* <ptr>, i64 <val> )
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6530
				6531	</pre>
				6532	<h5>Overview:</h5>
				6533	<p>
				6534	This intrinsic loads the value stored in memory at <tt>ptr</tt> and yields
				6535	the value from memory. It then stores the value in <tt>val</tt> in the memory
				6536	at <tt>ptr</tt>.
				6537	</p>
				6538	<h5>Arguments:</h5>
				6539
				6540	<p>
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6541	The <tt>llvm.atomic.swap</tt> intrinsic takes two arguments. Both the
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6542	<tt>val</tt> argument and the result must be integers of the same bit width.
				6543	The first argument, <tt>ptr</tt>, must be a pointer to a value of this
				6544	integer type. The targets may only lower integer representations they
				6545	support.
				6546	</p>
				6547	<h5>Semantics:</h5>
				6548	<p>
				6549	This intrinsic loads the value pointed to by <tt>ptr</tt>, yields it, and
				6550	stores <tt>val</tt> back into <tt>ptr</tt> atomically. This provides the
				6551	equivalent of an atomic swap operation within the SSA framework.
				6552
				6553	</p>
				6554	<h5>Examples:</h5>
				6555	<pre>
				6556	%ptr = malloc i32
				6557	store i32 4, %ptr
				6558
				6559	%val1 = add i32 4, 4
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6560	%result1 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val1 )
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6561	<i>; yields {i32}:result1 = 4</i>
				6562	%stored1 = icmp eq i32 %result1, 4 <i>; yields {i1}:stored1 = true</i>
				6563	%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 8</i>
				6564
				6565	%val2 = add i32 1, 1
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6566	%result2 = call i32 @llvm.atomic.swap.i32.p0i32( i32* %ptr, i32 %val2 )
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6567	<i>; yields {i32}:result2 = 8</i>
				6568
				6569	%stored2 = icmp eq i32 %result2, 8 <i>; yields {i1}:stored2 = true</i>
				6570	%memval2 = load i32* %ptr <i>; yields {i32}:memval2 = 2</i>
				6571	</pre>
				6572	</div>
				6573
				6574	<!-- _______________________________________________________________________ -->
				6575	<div class="doc_subsubsection">
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6576	<a name="int_atomic_load_add">'<tt>llvm.atomic.load.add.*</tt>' Intrinsic</a>
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6577
				6578	</div>
				6579	<div class="doc_text">
				6580	<h5>Syntax:</h5>
				6581	<p>
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6582	This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.add</tt> on any
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6583	integer bit width. Not all targets support all bit widths however.</p>
				6584	<pre>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6585	declare i8 @llvm.atomic.load.add.i8..p0i8( i8* <ptr>, i8 <delta> )
				6586	declare i16 @llvm.atomic.load.add.i16..p0i16( i16* <ptr>, i16 <delta> )
				6587	declare i32 @llvm.atomic.load.add.i32..p0i32( i32* <ptr>, i32 <delta> )
				6588	declare i64 @llvm.atomic.load.add.i64..p0i64( i64* <ptr>, i64 <delta> )
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6589
				6590	</pre>
				6591	<h5>Overview:</h5>
				6592	<p>
				6593	This intrinsic adds <tt>delta</tt> to the value stored in memory at
				6594	<tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
				6595	</p>
				6596	<h5>Arguments:</h5>
				6597	<p>
				6598
				6599	The intrinsic takes two arguments, the first a pointer to an integer value
				6600	and the second an integer value. The result is also an integer value. These
				6601	integer types can have any bit width, but they must all have the same bit
				6602	width. The targets may only lower integer representations they support.
				6603	</p>
				6604	<h5>Semantics:</h5>
				6605	<p>
				6606	This intrinsic does a series of operations atomically. It first loads the
				6607	value stored at <tt>ptr</tt>. It then adds <tt>delta</tt>, stores the result
				6608	to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
				6609	</p>
				6610
				6611	<h5>Examples:</h5>
				6612	<pre>
				6613	%ptr = malloc i32
				6614	store i32 4, %ptr
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6615	%result1 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 4 )
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6616	<i>; yields {i32}:result1 = 4</i>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6617	%result2 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 2 )
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6618	<i>; yields {i32}:result2 = 8</i>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6619	%result3 = call i32 @llvm.atomic.load.add.i32.p0i32( i32* %ptr, i32 5 )
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6620	<i>; yields {i32}:result3 = 10</i>
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6621	%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 15</i>
Andrew Lenharth	ab0b949	2008-02-21 06:45:13 +0000	[diff] [blame]	6622	</pre>
				6623	</div>
				6624
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6625	<!-- _______________________________________________________________________ -->
				6626	<div class="doc_subsubsection">
				6627	<a name="int_atomic_load_sub">'<tt>llvm.atomic.load.sub.*</tt>' Intrinsic</a>
				6628
				6629	</div>
				6630	<div class="doc_text">
				6631	<h5>Syntax:</h5>
				6632	<p>
				6633	This is an overloaded intrinsic. You can use <tt>llvm.atomic.load.sub</tt> on
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6634	any integer bit width and for different address spaces. Not all targets
				6635	support all bit widths however.</p>
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6636	<pre>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6637	declare i8 @llvm.atomic.load.sub.i8.p0i32( i8* <ptr>, i8 <delta> )
				6638	declare i16 @llvm.atomic.load.sub.i16.p0i32( i16* <ptr>, i16 <delta> )
				6639	declare i32 @llvm.atomic.load.sub.i32.p0i32( i32* <ptr>, i32 <delta> )
				6640	declare i64 @llvm.atomic.load.sub.i64.p0i32( i64* <ptr>, i64 <delta> )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6641
				6642	</pre>
				6643	<h5>Overview:</h5>
				6644	<p>
				6645	This intrinsic subtracts <tt>delta</tt> to the value stored in memory at
				6646	<tt>ptr</tt>. It yields the original value at <tt>ptr</tt>.
				6647	</p>
				6648	<h5>Arguments:</h5>
				6649	<p>
				6650
				6651	The intrinsic takes two arguments, the first a pointer to an integer value
				6652	and the second an integer value. The result is also an integer value. These
				6653	integer types can have any bit width, but they must all have the same bit
				6654	width. The targets may only lower integer representations they support.
				6655	</p>
				6656	<h5>Semantics:</h5>
				6657	<p>
				6658	This intrinsic does a series of operations atomically. It first loads the
				6659	value stored at <tt>ptr</tt>. It then subtracts <tt>delta</tt>, stores the
				6660	result to <tt>ptr</tt>. It yields the original value stored at <tt>ptr</tt>.
				6661	</p>
				6662
				6663	<h5>Examples:</h5>
				6664	<pre>
				6665	%ptr = malloc i32
				6666	store i32 8, %ptr
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6667	%result1 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 4 )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6668	<i>; yields {i32}:result1 = 8</i>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6669	%result2 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 2 )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6670	<i>; yields {i32}:result2 = 4</i>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6671	%result3 = call i32 @llvm.atomic.load.sub.i32.p0i32( i32* %ptr, i32 5 )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6672	<i>; yields {i32}:result3 = 2</i>
				6673	%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = -3</i>
				6674	</pre>
				6675	</div>
				6676
				6677	<!-- _______________________________________________________________________ -->
				6678	<div class="doc_subsubsection">
				6679	<a name="int_atomic_load_and">'<tt>llvm.atomic.load.and.*</tt>' Intrinsic</a><br>
				6680	<a name="int_atomic_load_nand">'<tt>llvm.atomic.load.nand.*</tt>' Intrinsic</a><br>
				6681	<a name="int_atomic_load_or">'<tt>llvm.atomic.load.or.*</tt>' Intrinsic</a><br>
				6682	<a name="int_atomic_load_xor">'<tt>llvm.atomic.load.xor.*</tt>' Intrinsic</a><br>
				6683
				6684	</div>
				6685	<div class="doc_text">
				6686	<h5>Syntax:</h5>
				6687	<p>
				6688	These are overloaded intrinsics. You can use <tt>llvm.atomic.load_and</tt>,
				6689	<tt>llvm.atomic.load_nand</tt>, <tt>llvm.atomic.load_or</tt>, and
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6690	<tt>llvm.atomic.load_xor</tt> on any integer bit width and for different
				6691	address spaces. Not all targets support all bit widths however.</p>
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6692	<pre>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6693	declare i8 @llvm.atomic.load.and.i8.p0i8( i8* <ptr>, i8 <delta> )
				6694	declare i16 @llvm.atomic.load.and.i16.p0i16( i16* <ptr>, i16 <delta> )
				6695	declare i32 @llvm.atomic.load.and.i32.p0i32( i32* <ptr>, i32 <delta> )
				6696	declare i64 @llvm.atomic.load.and.i64.p0i64( i64* <ptr>, i64 <delta> )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6697
				6698	</pre>
				6699
				6700	<pre>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6701	declare i8 @llvm.atomic.load.or.i8.p0i8( i8* <ptr>, i8 <delta> )
				6702	declare i16 @llvm.atomic.load.or.i16.p0i16( i16* <ptr>, i16 <delta> )
				6703	declare i32 @llvm.atomic.load.or.i32.p0i32( i32* <ptr>, i32 <delta> )
				6704	declare i64 @llvm.atomic.load.or.i64.p0i64( i64* <ptr>, i64 <delta> )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6705
				6706	</pre>
				6707
				6708	<pre>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6709	declare i8 @llvm.atomic.load.nand.i8.p0i32( i8* <ptr>, i8 <delta> )
				6710	declare i16 @llvm.atomic.load.nand.i16.p0i32( i16* <ptr>, i16 <delta> )
				6711	declare i32 @llvm.atomic.load.nand.i32.p0i32( i32* <ptr>, i32 <delta> )
				6712	declare i64 @llvm.atomic.load.nand.i64.p0i32( i64* <ptr>, i64 <delta> )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6713
				6714	</pre>
				6715
				6716	<pre>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6717	declare i8 @llvm.atomic.load.xor.i8.p0i32( i8* <ptr>, i8 <delta> )
				6718	declare i16 @llvm.atomic.load.xor.i16.p0i32( i16* <ptr>, i16 <delta> )
				6719	declare i32 @llvm.atomic.load.xor.i32.p0i32( i32* <ptr>, i32 <delta> )
				6720	declare i64 @llvm.atomic.load.xor.i64.p0i32( i64* <ptr>, i64 <delta> )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6721
				6722	</pre>
				6723	<h5>Overview:</h5>
				6724	<p>
				6725	These intrinsics bitwise the operation (and, nand, or, xor) <tt>delta</tt> to
				6726	the value stored in memory at <tt>ptr</tt>. It yields the original value
				6727	at <tt>ptr</tt>.
				6728	</p>
				6729	<h5>Arguments:</h5>
				6730	<p>
				6731
				6732	These intrinsics take two arguments, the first a pointer to an integer value
				6733	and the second an integer value. The result is also an integer value. These
				6734	integer types can have any bit width, but they must all have the same bit
				6735	width. The targets may only lower integer representations they support.
				6736	</p>
				6737	<h5>Semantics:</h5>
				6738	<p>
				6739	These intrinsics does a series of operations atomically. They first load the
				6740	value stored at <tt>ptr</tt>. They then do the bitwise operation
				6741	<tt>delta</tt>, store the result to <tt>ptr</tt>. They yield the original
				6742	value stored at <tt>ptr</tt>.
				6743	</p>
				6744
				6745	<h5>Examples:</h5>
				6746	<pre>
				6747	%ptr = malloc i32
				6748	store i32 0x0F0F, %ptr
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6749	%result0 = call i32 @llvm.atomic.load.nand.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6750	<i>; yields {i32}:result0 = 0x0F0F</i>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6751	%result1 = call i32 @llvm.atomic.load.and.i32.p0i32( i32* %ptr, i32 0xFF )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6752	<i>; yields {i32}:result1 = 0xFFFFFFF0</i>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6753	%result2 = call i32 @llvm.atomic.load.or.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6754	<i>; yields {i32}:result2 = 0xF0</i>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6755	%result3 = call i32 @llvm.atomic.load.xor.i32.p0i32( i32* %ptr, i32 0F )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6756	<i>; yields {i32}:result3 = FF</i>
				6757	%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = F0</i>
				6758	</pre>
				6759	</div>
				6760
				6761
				6762	<!-- _______________________________________________________________________ -->
				6763	<div class="doc_subsubsection">
				6764	<a name="int_atomic_load_max">'<tt>llvm.atomic.load.max.*</tt>' Intrinsic</a><br>
				6765	<a name="int_atomic_load_min">'<tt>llvm.atomic.load.min.*</tt>' Intrinsic</a><br>
				6766	<a name="int_atomic_load_umax">'<tt>llvm.atomic.load.umax.*</tt>' Intrinsic</a><br>
				6767	<a name="int_atomic_load_umin">'<tt>llvm.atomic.load.umin.*</tt>' Intrinsic</a><br>
				6768
				6769	</div>
				6770	<div class="doc_text">
				6771	<h5>Syntax:</h5>
				6772	<p>
				6773	These are overloaded intrinsics. You can use <tt>llvm.atomic.load_max</tt>,
				6774	<tt>llvm.atomic.load_min</tt>, <tt>llvm.atomic.load_umax</tt>, and
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6775	<tt>llvm.atomic.load_umin</tt> on any integer bit width and for different
				6776	address spaces. Not all targets
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6777	support all bit widths however.</p>
				6778	<pre>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6779	declare i8 @llvm.atomic.load.max.i8.p0i8( i8* <ptr>, i8 <delta> )
				6780	declare i16 @llvm.atomic.load.max.i16.p0i16( i16* <ptr>, i16 <delta> )
				6781	declare i32 @llvm.atomic.load.max.i32.p0i32( i32* <ptr>, i32 <delta> )
				6782	declare i64 @llvm.atomic.load.max.i64.p0i64( i64* <ptr>, i64 <delta> )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6783
				6784	</pre>
				6785
				6786	<pre>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6787	declare i8 @llvm.atomic.load.min.i8.p0i8( i8* <ptr>, i8 <delta> )
				6788	declare i16 @llvm.atomic.load.min.i16.p0i16( i16* <ptr>, i16 <delta> )
				6789	declare i32 @llvm.atomic.load.min.i32..p0i32( i32* <ptr>, i32 <delta> )
				6790	declare i64 @llvm.atomic.load.min.i64..p0i64( i64* <ptr>, i64 <delta> )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6791
				6792	</pre>
				6793
				6794	<pre>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6795	declare i8 @llvm.atomic.load.umax.i8.p0i8( i8* <ptr>, i8 <delta> )
				6796	declare i16 @llvm.atomic.load.umax.i16.p0i16( i16* <ptr>, i16 <delta> )
				6797	declare i32 @llvm.atomic.load.umax.i32.p0i32( i32* <ptr>, i32 <delta> )
				6798	declare i64 @llvm.atomic.load.umax.i64.p0i64( i64* <ptr>, i64 <delta> )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6799
				6800	</pre>
				6801
				6802	<pre>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6803	declare i8 @llvm.atomic.load.umin.i8..p0i8( i8* <ptr>, i8 <delta> )
				6804	declare i16 @llvm.atomic.load.umin.i16.p0i16( i16* <ptr>, i16 <delta> )
				6805	declare i32 @llvm.atomic.load.umin.i32..p0i32( i32* <ptr>, i32 <delta> )
				6806	declare i64 @llvm.atomic.load.umin.i64..p0i64( i64* <ptr>, i64 <delta> )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6807
				6808	</pre>
				6809	<h5>Overview:</h5>
				6810	<p>
				6811	These intrinsics takes the signed or unsigned minimum or maximum of
				6812	<tt>delta</tt> and the value stored in memory at <tt>ptr</tt>. It yields the
				6813	original value at <tt>ptr</tt>.
				6814	</p>
				6815	<h5>Arguments:</h5>
				6816	<p>
				6817
				6818	These intrinsics take two arguments, the first a pointer to an integer value
				6819	and the second an integer value. The result is also an integer value. These
				6820	integer types can have any bit width, but they must all have the same bit
				6821	width. The targets may only lower integer representations they support.
				6822	</p>
				6823	<h5>Semantics:</h5>
				6824	<p>
				6825	These intrinsics does a series of operations atomically. They first load the
				6826	value stored at <tt>ptr</tt>. They then do the signed or unsigned min or max
				6827	<tt>delta</tt> and the value, store the result to <tt>ptr</tt>. They yield
				6828	the original value stored at <tt>ptr</tt>.
				6829	</p>
				6830
				6831	<h5>Examples:</h5>
				6832	<pre>
				6833	%ptr = malloc i32
				6834	store i32 7, %ptr
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6835	%result0 = call i32 @llvm.atomic.load.min.i32.p0i32( i32* %ptr, i32 -2 )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6836	<i>; yields {i32}:result0 = 7</i>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6837	%result1 = call i32 @llvm.atomic.load.max.i32.p0i32( i32* %ptr, i32 8 )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6838	<i>; yields {i32}:result1 = -2</i>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6839	%result2 = call i32 @llvm.atomic.load.umin.i32.p0i32( i32* %ptr, i32 10 )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6840	<i>; yields {i32}:result2 = 8</i>
Mon P Wang	e3b3a72	2008-07-30 04:36:53 +0000	[diff] [blame]	6841	%result3 = call i32 @llvm.atomic.load.umax.i32.p0i32( i32* %ptr, i32 30 )
Mon P Wang	2887310	2008-06-25 08:15:39 +0000	[diff] [blame]	6842	<i>; yields {i32}:result3 = 8</i>
				6843	%memval1 = load i32* %ptr <i>; yields {i32}:memval1 = 30</i>
				6844	</pre>
				6845	</div>
Andrew Lenharth	22c5c1b	2008-02-16 01:24:58 +0000	[diff] [blame]	6846
				6847	<!-- ======================================================================= -->
				6848	<div class="doc_subsection">
Tanya Lattner	6d806e9	2007-06-15 20:50:54 +0000	[diff] [blame]	6849	<a name="int_general">General Intrinsics</a>
				6850	</div>
				6851
				6852	<div class="doc_text">
				6853	<p> This class of intrinsics is designed to be generic and has
				6854	no specific purpose. </p>
				6855	</div>
				6856
				6857	<!-- _______________________________________________________________________ -->
				6858	<div class="doc_subsubsection">
				6859	<a name="int_var_annotation">'<tt>llvm.var.annotation</tt>' Intrinsic</a>
				6860	</div>
				6861
				6862	<div class="doc_text">
				6863
				6864	<h5>Syntax:</h5>
				6865	<pre>
Tanya Lattner	d2e8442	2007-06-18 23:42:37 +0000	[diff] [blame]	6866	declare void @llvm.var.annotation(i8* <val>, i8* <str>, i8* <str>, i32 <int> )
Tanya Lattner	6d806e9	2007-06-15 20:50:54 +0000	[diff] [blame]	6867	</pre>
				6868
				6869	<h5>Overview:</h5>
				6870
				6871	<p>
				6872	The '<tt>llvm.var.annotation</tt>' intrinsic
				6873	</p>
				6874
				6875	<h5>Arguments:</h5>
				6876
				6877	<p>
Tanya Lattner	d2e8442	2007-06-18 23:42:37 +0000	[diff] [blame]	6878	The first argument is a pointer to a value, the second is a pointer to a
				6879	global string, the third is a pointer to a global string which is the source
				6880	file name, and the last argument is the line number.
Tanya Lattner	6d806e9	2007-06-15 20:50:54 +0000	[diff] [blame]	6881	</p>
				6882
				6883	<h5>Semantics:</h5>
				6884
				6885	<p>
Anton Korobeynikov	4cb8618	2008-01-15 22:31:34 +0000	[diff] [blame]	6886	This intrinsic allows annotation of local variables with arbitrary strings.
Tanya Lattner	6d806e9	2007-06-15 20:50:54 +0000	[diff] [blame]	6887	This can be useful for special purpose optimizations that want to look for these
Anton Korobeynikov	4cb8618	2008-01-15 22:31:34 +0000	[diff] [blame]	6888	annotations. These have no other defined use, they are ignored by code
				6889	generation and optimization.
				6890	</p>
Tanya Lattner	6d806e9	2007-06-15 20:50:54 +0000	[diff] [blame]	6891	</div>
				6892
Tanya Lattner	b636788	2007-09-21 22:59:12 +0000	[diff] [blame]	6893	<!-- _______________________________________________________________________ -->
				6894	<div class="doc_subsubsection">
Tanya Lattner	e1a8da0	2007-09-21 23:57:59 +0000	[diff] [blame]	6895	<a name="int_annotation">'<tt>llvm.annotation.*</tt>' Intrinsic</a>
Tanya Lattner	b636788	2007-09-21 22:59:12 +0000	[diff] [blame]	6896	</div>
				6897
				6898	<div class="doc_text">
				6899
				6900	<h5>Syntax:</h5>
Tanya Lattner	39cfba6	2007-09-21 23:56:27 +0000	[diff] [blame]	6901	<p>This is an overloaded intrinsic. You can use '<tt>llvm.annotation</tt>' on
				6902	any integer bit width.
				6903	</p>
Tanya Lattner	b636788	2007-09-21 22:59:12 +0000	[diff] [blame]	6904	<pre>
Tanya Lattner	d3989a8	2007-09-22 00:03:01 +0000	[diff] [blame]	6905	declare i8 @llvm.annotation.i8(i8 <val>, i8* <str>, i8* <str>, i32 <int> )
				6906	declare i16 @llvm.annotation.i16(i16 <val>, i8* <str>, i8* <str>, i32 <int> )
				6907	declare i32 @llvm.annotation.i32(i32 <val>, i8* <str>, i8* <str>, i32 <int> )
				6908	declare i64 @llvm.annotation.i64(i64 <val>, i8* <str>, i8* <str>, i32 <int> )
				6909	declare i256 @llvm.annotation.i256(i256 <val>, i8* <str>, i8* <str>, i32 <int> )
Tanya Lattner	b636788	2007-09-21 22:59:12 +0000	[diff] [blame]	6910	</pre>
				6911
				6912	<h5>Overview:</h5>
Tanya Lattner	39cfba6	2007-09-21 23:56:27 +0000	[diff] [blame]	6913
				6914	<p>
				6915	The '<tt>llvm.annotation</tt>' intrinsic.
Tanya Lattner	b636788	2007-09-21 22:59:12 +0000	[diff] [blame]	6916	</p>
				6917
				6918	<h5>Arguments:</h5>
				6919
				6920	<p>
				6921	The first argument is an integer value (result of some expression),
				6922	the second is a pointer to a global string, the third is a pointer to a global
				6923	string which is the source file name, and the last argument is the line number.
Tanya Lattner	39cfba6	2007-09-21 23:56:27 +0000	[diff] [blame]	6924	It returns the value of the first argument.
Tanya Lattner	b636788	2007-09-21 22:59:12 +0000	[diff] [blame]	6925	</p>
				6926
				6927	<h5>Semantics:</h5>
				6928
				6929	<p>
				6930	This intrinsic allows annotations to be put on arbitrary expressions
				6931	with arbitrary strings. This can be useful for special purpose optimizations
				6932	that want to look for these annotations. These have no other defined use, they
				6933	are ignored by code generation and optimization.
Dan Gohman	0e451ce	2008-10-14 16:51:45 +0000	[diff] [blame]	6934	</p>
Tanya Lattner	b636788	2007-09-21 22:59:12 +0000	[diff] [blame]	6935	</div>
Jim Laskey	dd4ef1b	2007-03-14 19:31:19 +0000	[diff] [blame]	6936
Anton Korobeynikov	4cb8618	2008-01-15 22:31:34 +0000	[diff] [blame]	6937	<!-- _______________________________________________________________________ -->
				6938	<div class="doc_subsubsection">
				6939	<a name="int_trap">'<tt>llvm.trap</tt>' Intrinsic</a>
				6940	</div>
				6941
				6942	<div class="doc_text">
				6943
				6944	<h5>Syntax:</h5>
				6945	<pre>
				6946	declare void @llvm.trap()
				6947	</pre>
				6948
				6949	<h5>Overview:</h5>
				6950
				6951	<p>
				6952	The '<tt>llvm.trap</tt>' intrinsic
				6953	</p>
				6954
				6955	<h5>Arguments:</h5>
				6956
				6957	<p>
				6958	None
				6959	</p>
				6960
				6961	<h5>Semantics:</h5>
				6962
				6963	<p>
				6964	This intrinsics is lowered to the target dependent trap instruction. If the
				6965	target does not have a trap instruction, this intrinsic will be lowered to the
				6966	call of the abort() function.
				6967	</p>
				6968	</div>
				6969
Bill Wendling	69e4adb	2008-11-19 05:56:17 +0000	[diff] [blame]	6970	<!-- _______________________________________________________________________ -->
				6971	<div class="doc_subsubsection">
Misha Brukman	dccb025	2008-11-22 23:55:29 +0000	[diff] [blame]	6972	<a name="int_stackprotector">'<tt>llvm.stackprotector</tt>' Intrinsic</a>
Bill Wendling	69e4adb	2008-11-19 05:56:17 +0000	[diff] [blame]	6973	</div>
				6974	<div class="doc_text">
				6975	<h5>Syntax:</h5>
				6976	<pre>
				6977	declare void @llvm.stackprotector( i8* <guard>, i8** <slot> )
				6978
				6979	</pre>
				6980	<h5>Overview:</h5>
				6981	<p>
				6982	The <tt>llvm.stackprotector</tt> intrinsic takes the <tt>guard</tt> and stores
				6983	it onto the stack at <tt>slot</tt>. The stack slot is adjusted to ensure that
				6984	it is placed on the stack before local variables.
				6985	</p>
				6986	<h5>Arguments:</h5>
				6987	<p>
				6988	The <tt>llvm.stackprotector</tt> intrinsic requires two pointer arguments. The
				6989	first argument is the value loaded from the stack guard
				6990	<tt>@__stack_chk_guard</tt>. The second variable is an <tt>alloca</tt> that
				6991	has enough space to hold the value of the guard.
				6992	</p>
				6993	<h5>Semantics:</h5>
				6994	<p>
				6995	This intrinsic causes the prologue/epilogue inserter to force the position of
				6996	the <tt>AllocaInst</tt> stack slot to be before local variables on the
				6997	stack. This is to ensure that if a local variable on the stack is overwritten,
				6998	it will destroy the value of the guard. When the function exits, the guard on
				6999	the stack is checked against the original guard. If they're different, then
				7000	the program aborts by calling the <tt>__stack_chk_fail()</tt> function.
				7001	</p>
				7002	</div>
				7003
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	7004	<!-- *********************************************************************** -->
Chris Lattner	0095054	2001-06-06 20:29:01 +0000	[diff] [blame]	7005	<hr>
Misha Brukman	daa4cb0	2004-03-01 17:47:27 +0000	[diff] [blame]	7006	<address>
				7007	<a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman	4440870	2008-12-11 17:34:48 +0000	[diff] [blame]	7008	src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukman	daa4cb0	2004-03-01 17:47:27 +0000	[diff] [blame]	7009	<a href="http://validator.w3.org/check/referer"><img
Misha Brukman	4440870	2008-12-11 17:34:48 +0000	[diff] [blame]	7010	src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Misha Brukman	daa4cb0	2004-03-01 17:47:27 +0000	[diff] [blame]	7011
				7012	<a href="mailto:sabre@nondot.org">Chris Lattner</a><br>
Reid Spencer	05fe4b0	2006-03-14 05:39:39 +0000	[diff] [blame]	7013	<a href="http://llvm.org">The LLVM Compiler Infrastructure</a><br>
Misha Brukman	daa4cb0	2004-03-01 17:47:27 +0000	[diff] [blame]	7014	Last modified: $Date$
				7015	</address>
Chris Lattner	c7d3ab3	2008-01-04 04:33:49 +0000	[diff] [blame]	7016
Misha Brukman	9d0919f	2003-11-08 01:05:38 +0000	[diff] [blame]	7017	</body>
				7018	</html>