Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / llvm / de8ec5b3afd5cfeeb7401bd817c82d7beb5a7171 / . / docs / Passes.html
blob: 63b2b3661a73ef6c906296707f89d0bf2364d570 [file] [log] [blame]
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<html>
4<head>
5 <title>LLVM's Analysis and Transform Passes</title>
6 <link rel="stylesheet" href="llvm.css" type="text/css">
Gordon Henriksenddaa61d2007-10-25 08:58:56 +0000[diff] [blame]7 <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]8</head>
9<body>
10
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]11<!--
12
13If Passes.html is up to date, the following "one-liner" should print
14an empty diff.
15
16egrep -e '^<tr><td><a href="#.*">-.*</a></td><td>.*</td></tr>$' \
17 -e '^ <a name=".*">.*</a>$' < Passes.html >html; \
18perl >help <<'EOT' && diff -u help html; rm -f help html
19open HTML, "<Passes.html" or die "open: Passes.html: $!\n";
20while (<HTML>) {
21 m:^<tr><td><a href="#(.*)">-.*</a></td><td>.*</td></tr>$: or next;
22 $order{$1} = sprintf("%03d", 1 + int %order);
23}
Gordon Henriksenddaa61d2007-10-25 08:58:56 +0000[diff] [blame]24open HELP, "../Release/bin/opt -help|" or die "open: opt -help: $!\n";
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]25while (<HELP>) {
26 m:^ -([^ ]+) +- (.*)$: or next;
27 my $o = $order{$1};
28 $o = "000" unless defined $o;
29 push @x, "$o<tr><td><a href=\"#$1\">-$1</a></td><td>$2</td></tr>\n";
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]30 push @y, "$o <a name=\"$1\">-$1: $2</a>\n";
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]31}
32@x = map { s/^\d\d\d//; $_ } sort @x;
33@y = map { s/^\d\d\d//; $_ } sort @y;
34print @x, @y;
35EOT
36
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]37This (real) one-liner can also be helpful when converting comments to HTML:
38
39perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print " <p>\n" if !$on && $_ =~ /\S/; print " </p>\n" if $on && $_ =~ /^\s*$/; print " $_\n"; $on = ($_ =~ /\S/); } print " </p>\n" if $on'
40
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]41 -->
42
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]43<div class="doc_title">LLVM's Analysis and Transform Passes</div>
44
45<ol>
46 <li><a href="#intro">Introduction</a></li>
47 <li><a href="#analyses">Analysis Passes</a>
48 <li><a href="#transforms">Transform Passes</a></li>
49 <li><a href="#utilities">Utility Passes</a></li>
50</ol>
51
52<div class="doc_author">
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]53 <p>Written by <a href="mailto:rspencer@x10sys.com">Reid Spencer</a>
54 and Gordon Henriksen</p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]55</div>
56
57<!-- ======================================================================= -->
58<div class="doc_section"> <a name="intro">Introduction</a> </div>
59<div class="doc_text">
60 <p>This document serves as a high level summary of the optimization features
61 that LLVM provides. Optimizations are implemented as Passes that traverse some
62 portion of a program to either collect information or transform the program.
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]63 The table below divides the passes that LLVM provides into three categories.
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]64 Analysis passes compute information that other passes can use or for debugging
65 or program visualization purposes. Transform passes can use (or invalidate)
66 the analysis passes. Transform passes all mutate the program in some way.
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]67 Utility passes provides some utility but don't otherwise fit categorization.
Gabor Greif04367bf2007-07-06 22:07:22 +0000[diff] [blame]68 For example passes to extract functions to bitcode or write a module to
69 bitcode are neither analysis nor transform passes.
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]70 <p>The table below provides a quick summary of each pass and links to the more
71 complete pass description later in the document.</p>
72</div>
73<div class="doc_text" >
74<table>
Gordon Henriksenddaa61d2007-10-25 08:58:56 +0000[diff] [blame]75<tr><th colspan="2"><b>ANALYSIS PASSES</b></th></tr>
76<tr><th>Option</th><th>Name</th></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]77<tr><td><a href="#aa-eval">-aa-eval</a></td><td>Exhaustive Alias Analysis Precision Evaluator</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]78<tr><td><a href="#basicaa">-basicaa</a></td><td>Basic Alias Analysis (stateless AA impl)</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]79<tr><td><a href="#basiccg">-basiccg</a></td><td>Basic CallGraph Construction</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]80<tr><td><a href="#count-aa">-count-aa</a></td><td>Count Alias Analysis Query Responses</td></tr>
81<tr><td><a href="#debug-aa">-debug-aa</a></td><td>AA use debugger</td></tr>
82<tr><td><a href="#domfrontier">-domfrontier</a></td><td>Dominance Frontier Construction</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]83<tr><td><a href="#domtree">-domtree</a></td><td>Dominator Tree Construction</td></tr>
Duncan Sands3ee8fc92008-09-23 12:47:39 +0000[diff] [blame]84<tr><td><a href="#dot-callgraph">-dot-callgraph</a></td><td>Print Call Graph to 'dot' file</td></tr>
85<tr><td><a href="#dot-cfg">-dot-cfg</a></td><td>Print CFG of function to 'dot' file</td></tr>
86<tr><td><a href="#dot-cfg-only">-dot-cfg-only</a></td><td>Print CFG of function to 'dot' file (with no function bodies)</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]87<tr><td><a href="#dot-dom">-dot-dom</a></td><td>Print dominance tree of function to 'dot' file</td></tr>
88<tr><td><a href="#dot-dom-only">-dot-dom-only</a></td><td>Print dominance tree of function to 'dot' file (with no function bodies)</td></tr>
89<tr><td><a href="#dot-postdom">-dot-postdom</a></td><td>Print postdominance tree of function to 'dot' file</td></tr>
90<tr><td><a href="#dot-postdom-only">-dot-postdom-only</a></td><td>Print postdominance tree of function to 'dot' file (with no function bodies)</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]91<tr><td><a href="#globalsmodref-aa">-globalsmodref-aa</a></td><td>Simple mod/ref analysis for globals</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]92<tr><td><a href="#instcount">-instcount</a></td><td>Counts the various types of Instructions</td></tr>
93<tr><td><a href="#intervals">-intervals</a></td><td>Interval Partition Construction</td></tr>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]94<tr><td><a href="#iv-users">-iv-users</a></td><td>Induction Variable Users</td></tr>
95<tr><td><a href="#lazy-value-info">-lazy-value-info</a></td><td>Lazy Value Information Analysis</td></tr>
96<tr><td><a href="#lda">-lda</a></td><td>Loop Dependence Analysis</td></tr>
97<tr><td><a href="#libcall-aa">-libcall-aa</a></td><td>LibCall Alias Analysis</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]98<tr><td><a href="#lint">-lint</a></td><td>Statically lint-checks LLVM IR</td></tr>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]99<tr><td><a href="#loops">-loops</a></td><td>Natural Loop Information</td></tr>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]100<tr><td><a href="#memdep">-memdep</a></td><td>Memory Dependence Analysis</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]101<tr><td><a href="#module-debuginfo">-module-debuginfo</a></td><td>Decodes module-level debug info</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]102<tr><td><a href="#no-aa">-no-aa</a></td><td>No Alias Analysis (always returns 'may' alias)</td></tr>
103<tr><td><a href="#no-profile">-no-profile</a></td><td>No Profile Information</td></tr>
104<tr><td><a href="#postdomfrontier">-postdomfrontier</a></td><td>Post-Dominance Frontier Construction</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]105<tr><td><a href="#postdomtree">-postdomtree</a></td><td>Post-Dominator Tree Construction</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]106<tr><td><a href="#print-alias-sets">-print-alias-sets</a></td><td>Alias Set Printer</td></tr>
Duncan Sands3ee8fc92008-09-23 12:47:39 +0000[diff] [blame]107<tr><td><a href="#print-callgraph">-print-callgraph</a></td><td>Print a call graph</td></tr>
108<tr><td><a href="#print-callgraph-sccs">-print-callgraph-sccs</a></td><td>Print SCCs of the Call Graph</td></tr>
109<tr><td><a href="#print-cfg-sccs">-print-cfg-sccs</a></td><td>Print SCCs of each function CFG</td></tr>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]110<tr><td><a href="#print-dbginfo">-print-dbginfo</a></td><td>Print debug info in human readable form</td></tr>
111<tr><td><a href="#print-dom-info">-print-dom-info</a></td><td>Dominator Info Printer</td></tr>
Duncan Sands3ee8fc92008-09-23 12:47:39 +0000[diff] [blame]112<tr><td><a href="#print-externalfnconstants">-print-externalfnconstants</a></td><td>Print external fn callsites passed constants</td></tr>
113<tr><td><a href="#print-function">-print-function</a></td><td>Print function to stderr</td></tr>
114<tr><td><a href="#print-module">-print-module</a></td><td>Print module to stderr</td></tr>
115<tr><td><a href="#print-used-types">-print-used-types</a></td><td>Find Used Types</td></tr>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]116<tr><td><a href="#profile-estimator">-profile-estimator</a></td><td>Estimate profiling information</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]117<tr><td><a href="#profile-loader">-profile-loader</a></td><td>Load profile information from llvmprof.out</td></tr>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]118<tr><td><a href="#profile-verifier">-profile-verifier</a></td><td>Verify profiling information</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]119<tr><td><a href="#regions">-regions</a></td><td>Detect single entry single exit regions</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]120<tr><td><a href="#scalar-evolution">-scalar-evolution</a></td><td>Scalar Evolution Analysis</td></tr>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]121<tr><td><a href="#scev-aa">-scev-aa</a></td><td>ScalarEvolution-based Alias Analysis</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]122<tr><td><a href="#targetdata">-targetdata</a></td><td>Target Data Layout</td></tr>
123
124
Gordon Henriksenddaa61d2007-10-25 08:58:56 +0000[diff] [blame]125<tr><th colspan="2"><b>TRANSFORM PASSES</b></th></tr>
126<tr><th>Option</th><th>Name</th></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]127<tr><td><a href="#adce">-adce</a></td><td>Aggressive Dead Code Elimination</td></tr>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]128<tr><td><a href="#always-inline">-always-inline</a></td><td>Inliner for always_inline functions</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]129<tr><td><a href="#argpromotion">-argpromotion</a></td><td>Promote 'by reference' arguments to scalars</td></tr>
130<tr><td><a href="#block-placement">-block-placement</a></td><td>Profile Guided Basic Block Placement</td></tr>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]131<tr><td><a href="#break-crit-edges">-break-crit-edges</a></td><td>Break critical edges in CFG</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]132<tr><td><a href="#codegenprepare">-codegenprepare</a></td><td>Optimize for code generation</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]133<tr><td><a href="#constmerge">-constmerge</a></td><td>Merge Duplicate Global Constants</td></tr>
134<tr><td><a href="#constprop">-constprop</a></td><td>Simple constant propagation</td></tr>
135<tr><td><a href="#dce">-dce</a></td><td>Dead Code Elimination</td></tr>
136<tr><td><a href="#deadargelim">-deadargelim</a></td><td>Dead Argument Elimination</td></tr>
137<tr><td><a href="#deadtypeelim">-deadtypeelim</a></td><td>Dead Type Elimination</td></tr>
138<tr><td><a href="#die">-die</a></td><td>Dead Instruction Elimination</td></tr>
139<tr><td><a href="#dse">-dse</a></td><td>Dead Store Elimination</td></tr>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]140<tr><td><a href="#functionattrs">-functionattrs</a></td><td>Deduce function attributes</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]141<tr><td><a href="#globaldce">-globaldce</a></td><td>Dead Global Elimination</td></tr>
142<tr><td><a href="#globalopt">-globalopt</a></td><td>Global Variable Optimizer</td></tr>
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]143<tr><td><a href="#gvn">-gvn</a></td><td>Global Value Numbering</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]144<tr><td><a href="#indvars">-indvars</a></td><td>Canonicalize Induction Variables</td></tr>
145<tr><td><a href="#inline">-inline</a></td><td>Function Integration/Inlining</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]146<tr><td><a href="#insert-edge-profiling">-insert-edge-profiling</a></td><td>Insert instrumentation for edge profiling</td></tr>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]147<tr><td><a href="#insert-optimal-edge-profiling">-insert-optimal-edge-profiling</a></td><td>Insert optimal instrumentation for edge profiling</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]148<tr><td><a href="#instcombine">-instcombine</a></td><td>Combine redundant instructions</td></tr>
149<tr><td><a href="#internalize">-internalize</a></td><td>Internalize Global Symbols</td></tr>
150<tr><td><a href="#ipconstprop">-ipconstprop</a></td><td>Interprocedural constant propagation</td></tr>
151<tr><td><a href="#ipsccp">-ipsccp</a></td><td>Interprocedural Sparse Conditional Constant Propagation</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]152<tr><td><a href="#jump-threading">-jump-threading</a></td><td>Jump Threading</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]153<tr><td><a href="#lcssa">-lcssa</a></td><td>Loop-Closed SSA Form Pass</td></tr>
154<tr><td><a href="#licm">-licm</a></td><td>Loop Invariant Code Motion</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]155<tr><td><a href="#loop-deletion">-loop-deletion</a></td><td>Delete dead loops</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]156<tr><td><a href="#loop-extract">-loop-extract</a></td><td>Extract loops into new functions</td></tr>
157<tr><td><a href="#loop-extract-single">-loop-extract-single</a></td><td>Extract at most one loop into a new function</td></tr>
158<tr><td><a href="#loop-reduce">-loop-reduce</a></td><td>Loop Strength Reduction</td></tr>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]159<tr><td><a href="#loop-rotate">-loop-rotate</a></td><td>Rotate Loops</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]160<tr><td><a href="#loop-simplify">-loop-simplify</a></td><td>Canonicalize natural loops</td></tr>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]161<tr><td><a href="#loop-unroll">-loop-unroll</a></td><td>Unroll loops</td></tr>
162<tr><td><a href="#loop-unswitch">-loop-unswitch</a></td><td>Unswitch loops</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]163<tr><td><a href="#loweratomic">-loweratomic</a></td><td>Lower atomic intrinsics to non-atomic form</td></tr>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]164<tr><td><a href="#lowerinvoke">-lowerinvoke</a></td><td>Lower invoke and unwind, for unwindless code generators</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]165<tr><td><a href="#lowersetjmp">-lowersetjmp</a></td><td>Lower Set Jump</td></tr>
166<tr><td><a href="#lowerswitch">-lowerswitch</a></td><td>Lower SwitchInst's to branches</td></tr>
167<tr><td><a href="#mem2reg">-mem2reg</a></td><td>Promote Memory to Register</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]168<tr><td><a href="#memcpyopt">-memcpyopt</a></td><td>MemCpy Optimization</td></tr>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]169<tr><td><a href="#mergefunc">-mergefunc</a></td><td>Merge Functions</td></tr>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]170<tr><td><a href="#mergereturn">-mergereturn</a></td><td>Unify function exit nodes</td></tr>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]171<tr><td><a href="#partial-inliner">-partial-inliner</a></td><td>Partial Inliner</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]172<tr><td><a href="#prune-eh">-prune-eh</a></td><td>Remove unused exception handling info</td></tr>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]173<tr><td><a href="#reassociate">-reassociate</a></td><td>Reassociate expressions</td></tr>
174<tr><td><a href="#reg2mem">-reg2mem</a></td><td>Demote all values to stack slots</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]175<tr><td><a href="#scalarrepl">-scalarrepl</a></td><td>Scalar Replacement of Aggregates (DT)</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]176<tr><td><a href="#sccp">-sccp</a></td><td>Sparse Conditional Constant Propagation</td></tr>
177<tr><td><a href="#simplify-libcalls">-simplify-libcalls</a></td><td>Simplify well-known library calls</td></tr>
178<tr><td><a href="#simplifycfg">-simplifycfg</a></td><td>Simplify the CFG</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]179<tr><td><a href="#sink">-sink</a></td><td>Code sinking</td></tr>
180<tr><td><a href="#sretpromotion">-sretpromotion</a></td><td>Promote sret arguments to multiple ret values</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]181<tr><td><a href="#strip">-strip</a></td><td>Strip all symbols from a module</td></tr>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]182<tr><td><a href="#strip-dead-debug-info">-strip-dead-debug-info</a></td><td>Strip debug info for unused symbols</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]183<tr><td><a href="#strip-dead-prototypes">-strip-dead-prototypes</a></td><td>Strip Unused Function Prototypes</td></tr>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]184<tr><td><a href="#strip-debug-declare">-strip-debug-declare</a></td><td>Strip all llvm.dbg.declare intrinsics</td></tr>
185<tr><td><a href="#strip-nondebug">-strip-nondebug</a></td><td>Strip all symbols, except dbg symbols, from a module</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]186<tr><td><a href="#tailcallelim">-tailcallelim</a></td><td>Tail Call Elimination</td></tr>
187<tr><td><a href="#tailduplicate">-tailduplicate</a></td><td>Tail Duplication</td></tr>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]188
189
Gordon Henriksenddaa61d2007-10-25 08:58:56 +0000[diff] [blame]190<tr><th colspan="2"><b>UTILITY PASSES</b></th></tr>
191<tr><th>Option</th><th>Name</th></tr>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]192<tr><td><a href="#deadarghaX0r">-deadarghaX0r</a></td><td>Dead Argument Hacking (BUGPOINT USE ONLY; DO NOT USE)</td></tr>
193<tr><td><a href="#extract-blocks">-extract-blocks</a></td><td>Extract Basic Blocks From Module (for bugpoint use)</td></tr>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]194<tr><td><a href="#instnamer">-instnamer</a></td><td>Assign names to anonymous instructions</td></tr>
Gordon Henriksen90a52142007-11-05 02:05:35 +0000[diff] [blame]195<tr><td><a href="#preverify">-preverify</a></td><td>Preliminary module verification</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]196<tr><td><a href="#verify">-verify</a></td><td>Module Verifier</td></tr>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]197<tr><td><a href="#view-cfg">-view-cfg</a></td><td>View CFG of function</td></tr>
198<tr><td><a href="#view-cfg-only">-view-cfg-only</a></td><td>View CFG of function (with no function bodies)</td></tr>
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]199<tr><td><a href="#view-dom">-view-dom</a></td><td>View dominance tree of function</td></tr>
200<tr><td><a href="#view-dom-only">-view-dom-only</a></td><td>View dominance tree of function (with no function bodies)</td></tr>
201<tr><td><a href="#view-postdom">-view-postdom</a></td><td>View postdominance tree of function</td></tr>
202<tr><td><a href="#view-postdom-only">-view-postdom-only</a></td><td>View postdominance tree of function (with no function bodies)</td></tr>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]203</table>
204</div>
205
206<!-- ======================================================================= -->
207<div class="doc_section"> <a name="example">Analysis Passes</a></div>
208<div class="doc_text">
209 <p>This section describes the LLVM Analysis Passes.</p>
210</div>
211
212<!-------------------------------------------------------------------------- -->
213<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]214 <a name="aa-eval">-aa-eval: Exhaustive Alias Analysis Precision Evaluator</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]215</div>
216<div class="doc_text">
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]217 <p>This is a simple N^2 alias analysis accuracy evaluator.
218 Basically, for each function in the program, it simply queries to see how the
219 alias analysis implementation answers alias queries between each pair of
220 pointers in the function.</p>
221
222 <p>This is inspired and adapted from code by: Naveen Neelakantam, Francesco
223 Spadini, and Wojciech Stryjewski.</p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]224</div>
225
226<!-------------------------------------------------------------------------- -->
227<div class="doc_subsection">
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]228 <a name="basicaa">-basicaa: Basic Alias Analysis (stateless AA impl)</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]229</div>
230<div class="doc_text">
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]231 <p>
232 This is the default implementation of the Alias Analysis interface
233 that simply implements a few identities (two different globals cannot alias,
234 etc), but otherwise does no analysis.
235 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]236</div>
237
238<!-------------------------------------------------------------------------- -->
239<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]240 <a name="basiccg">-basiccg: Basic CallGraph Construction</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]241</div>
242<div class="doc_text">
243 <p>Yet to be written.</p>
244</div>
245
246<!-------------------------------------------------------------------------- -->
247<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]248 <a name="count-aa">-count-aa: Count Alias Analysis Query Responses</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]249</div>
250<div class="doc_text">
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]251 <p>
252 A pass which can be used to count how many alias queries
253 are being made and how the alias analysis implementation being used responds.
254 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]255</div>
256
257<!-------------------------------------------------------------------------- -->
258<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]259 <a name="debug-aa">-debug-aa: AA use debugger</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]260</div>
261<div class="doc_text">
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]262 <p>
263 This simple pass checks alias analysis users to ensure that if they
264 create a new value, they do not query AA without informing it of the value.
265 It acts as a shim over any other AA pass you want.
266 </p>
267
268 <p>
269 Yes keeping track of every value in the program is expensive, but this is
270 a debugging pass.
271 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]272</div>
273
274<!-------------------------------------------------------------------------- -->
275<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]276 <a name="domfrontier">-domfrontier: Dominance Frontier Construction</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]277</div>
278<div class="doc_text">
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]279 <p>
280 This pass is a simple dominator construction algorithm for finding forward
281 dominator frontiers.
282 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]283</div>
284
285<!-------------------------------------------------------------------------- -->
286<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]287 <a name="domtree">-domtree: Dominator Tree Construction</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]288</div>
289<div class="doc_text">
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]290 <p>
291 This pass is a simple dominator construction algorithm for finding forward
292 dominators.
293 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]294</div>
295
296<!-------------------------------------------------------------------------- -->
297<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]298 <a name="dot-callgraph">-dot-callgraph: Print Call Graph to 'dot' file</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]299</div>
300<div class="doc_text">
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]301 <p>
Duncan Sands3ee8fc92008-09-23 12:47:39 +0000[diff] [blame]302 This pass, only available in <code>opt</code>, prints the call graph into a
303 <code>.dot</code> graph. This graph can then be processed with the "dot" tool
304 to convert it to postscript or some other suitable format.
305 </p>
306</div>
307
308<!-------------------------------------------------------------------------- -->
309<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]310 <a name="dot-cfg">-dot-cfg: Print CFG of function to 'dot' file</a>
Duncan Sands3ee8fc92008-09-23 12:47:39 +0000[diff] [blame]311</div>
312<div class="doc_text">
313 <p>
314 This pass, only available in <code>opt</code>, prints the control flow graph
315 into a <code>.dot</code> graph. This graph can then be processed with the
316 "dot" tool to convert it to postscript or some other suitable format.
317 </p>
318</div>
319
320<!-------------------------------------------------------------------------- -->
321<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]322 <a name="dot-cfg-only">-dot-cfg-only: Print CFG of function to 'dot' file (with no function bodies)</a>
Duncan Sands3ee8fc92008-09-23 12:47:39 +0000[diff] [blame]323</div>
324<div class="doc_text">
325 <p>
326 This pass, only available in <code>opt</code>, prints the control flow graph
327 into a <code>.dot</code> graph, omitting the function bodies. This graph can
328 then be processed with the "dot" tool to convert it to postscript or some
329 other suitable format.
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]330 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]331</div>
332
333<!-------------------------------------------------------------------------- -->
334<div class="doc_subsection">
Eli Friedmande8ec5b2011-03-19 04:55:29 +0000[diff] [blame^]335 <a name="dot-dom">-dot-dom: Print dominance tree of function to 'dot' file</a>
Tobias Grosser733783b2010-05-07 09:33:18 +0000[diff] [blame]336</div>
337<div class="doc_text">
338 <p>
339 This pass, only available in <code>opt</code>, prints the dominator tree
340 into a <code>.dot</code> graph. This graph can then be processed with the
341 "dot" tool to convert it to postscript or some other suitable format.
342 </p>
343</div>
344
345<!-------------------------------------------------------------------------- -->
346<div class="doc_subsection">
Eli Friedmande8ec5b2011-03-19 04:55:29 +0000[diff] [blame^]347 <a name="dot-dom-only">-dot-dom-only: Print dominance tree of function to 'dot' file (with no function bodies)</a>
Tobias Grosser733783b2010-05-07 09:33:18 +0000[diff] [blame]348</div>
349<div class="doc_text">
350 <p>
351 This pass, only available in <code>opt</code>, prints the dominator tree
352 into a <code>.dot</code> graph, omitting the function bodies. This graph can
353 then be processed with the "dot" tool to convert it to postscript or some
354 other suitable format.
355 </p>
356</div>
357
358<!-------------------------------------------------------------------------- -->
359<div class="doc_subsection">
Eli Friedmande8ec5b2011-03-19 04:55:29 +0000[diff] [blame^]360 <a name="dot-postdom">-dot-postdom: Print postdominance tree of function to 'dot' file</a>
Tobias Grosser733783b2010-05-07 09:33:18 +0000[diff] [blame]361</div>
362<div class="doc_text">
363 <p>
364 This pass, only available in <code>opt</code>, prints the post dominator tree
365 into a <code>.dot</code> graph. This graph can then be processed with the
366 "dot" tool to convert it to postscript or some other suitable format.
367 </p>
368</div>
369
370<!-------------------------------------------------------------------------- -->
371<div class="doc_subsection">
Eli Friedmande8ec5b2011-03-19 04:55:29 +0000[diff] [blame^]372 <a name="dot-postdom-only">-dot-postdom-only: Print postdominance tree of function to 'dot' file (with no function bodies)</a>
Tobias Grosser733783b2010-05-07 09:33:18 +0000[diff] [blame]373</div>
374<div class="doc_text">
375 <p>
376 This pass, only available in <code>opt</code>, prints the post dominator tree
377 into a <code>.dot</code> graph, omitting the function bodies. This graph can
378 then be processed with the "dot" tool to convert it to postscript or some
379 other suitable format.
380 </p>
381</div>
382
383<!-------------------------------------------------------------------------- -->
384<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]385 <a name="globalsmodref-aa">-globalsmodref-aa: Simple mod/ref analysis for globals</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]386</div>
387<div class="doc_text">
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]388 <p>
389 This simple pass provides alias and mod/ref information for global values
390 that do not have their address taken, and keeps track of whether functions
391 read or write memory (are "pure"). For this simple (but very common) case,
392 we can provide pretty accurate and useful information.
393 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]394</div>
395
396<!-------------------------------------------------------------------------- -->
397<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]398 <a name="instcount">-instcount: Counts the various types of Instructions</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]399</div>
400<div class="doc_text">
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]401 <p>
402 This pass collects the count of all instructions and reports them
403 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]404</div>
405
406<!-------------------------------------------------------------------------- -->
407<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]408 <a name="intervals">-intervals: Interval Partition Construction</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]409</div>
410<div class="doc_text">
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]411 <p>
412 This analysis calculates and represents the interval partition of a function,
413 or a preexisting interval partition.
414 </p>
415
416 <p>
417 In this way, the interval partition may be used to reduce a flow graph down
418 to its degenerate single node interval partition (unless it is irreducible).
419 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]420</div>
421
422<!-------------------------------------------------------------------------- -->
423<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]424 <a name="iv-users">-iv-users: Induction Variable Users</a>
425</div>
426<div class="doc_text">
427 <p>Bookkeeping for "interesting" users of expressions computed from
428 induction variables.</p>
429</div>
430
431<!-------------------------------------------------------------------------- -->
432<div class="doc_subsection">
433 <a name="lazy-value-info">-lazy-value-info: Lazy Value Information Analysis</a>
434</div>
435<div class="doc_text">
436 <p>Interface for lazy computation of value constraint information.</p>
437</div>
438
439<!-------------------------------------------------------------------------- -->
440<div class="doc_subsection">
441 <a name="lda">-lda: Loop Dependence Analysis</a>
442</div>
443<div class="doc_text">
444 <p>Loop dependence analysis framework, which is used to detect dependences in
445 memory accesses in loops.</p>
446</div>
447
448<!-------------------------------------------------------------------------- -->
449<div class="doc_subsection">
450 <a name="libcall-aa">-libcall-aa: LibCall Alias Analysis</a>
451</div>
452<div class="doc_text">
453 <p>LibCall Alias Analysis.</p>
454</div>
455
456<!-------------------------------------------------------------------------- -->
457<div class="doc_subsection">
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]458 <a name="lint">-lint: Statically lint-checks LLVM IR</a>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]459</div>
460<div class="doc_text">
461 <p>This pass statically checks for common and easily-identified constructs
462 which produce undefined or likely unintended behavior in LLVM IR.</p>
463
464 <p>It is not a guarantee of correctness, in two ways. First, it isn't
465 comprehensive. There are checks which could be done statically which are
466 not yet implemented. Some of these are indicated by TODO comments, but
467 those aren't comprehensive either. Second, many conditions cannot be
468 checked statically. This pass does no dynamic instrumentation, so it
469 can't check for all possible problems.</p>
470
471 <p>Another limitation is that it assumes all code will be executed. A store
472 through a null pointer in a basic block which is never reached is harmless,
473 but this pass will warn about it anyway.</p>
474
475 <p>Optimization passes may make conditions that this pass checks for more or
476 less obvious. If an optimization pass appears to be introducing a warning,
477 it may be that the optimization pass is merely exposing an existing
478 condition in the code.</p>
479
480 <p>This code may be run before instcombine. In many cases, instcombine checks
481 for the same kinds of things and turns instructions with undefined behavior
482 into unreachable (or equivalent). Because of this, this pass makes some
483 effort to look through bitcasts and so on.
484 </p>
485</div>
486
487<!-------------------------------------------------------------------------- -->
488<div class="doc_subsection">
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]489 <a name="loops">-loops: Natural Loop Information</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]490</div>
491<div class="doc_text">
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]492 <p>
493 This analysis is used to identify natural loops and determine the loop depth
494 of various nodes of the CFG. Note that the loops identified may actually be
495 several natural loops that share the same header node... not just a single
496 natural loop.
497 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]498</div>
499
500<!-------------------------------------------------------------------------- -->
501<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]502 <a name="memdep">-memdep: Memory Dependence Analysis</a>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]503</div>
504<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]505 <p>
506 An analysis that determines, for a given memory operation, what preceding
507 memory operations it depends on. It builds on alias analysis information, and
508 tries to provide a lazy, caching interface to a common kind of alias
509 information query.
510 </p>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]511</div>
512
513<!-------------------------------------------------------------------------- -->
514<div class="doc_subsection">
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]515 <a name="module-debuginfo">-module-debuginfo: Decodes module-level debug info</a>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]516</div>
517<div class="doc_text">
518 <p>This pass decodes the debug info metadata in a module and prints in a
519 (sufficiently-prepared-) human-readable form.
520
521 For example, run this pass from opt along with the -analyze option, and
522 it'll print to standard output.
523 </p>
524</div>
525
526<!-------------------------------------------------------------------------- -->
527<div class="doc_subsection">
528 <a name="no-aa">-no-aa: No Alias Analysis (always returns 'may' alias)</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]529</div>
530<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]531 <p>
532 Always returns "I don't know" for alias queries. NoAA is unlike other alias
533 analysis implementations, in that it does not chain to a previous analysis. As
534 such it doesn't follow many of the rules that other alias analyses must.
535 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]536</div>
537
538<!-------------------------------------------------------------------------- -->
539<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]540 <a name="no-profile">-no-profile: No Profile Information</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]541</div>
542<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]543 <p>
544 The default "no profile" implementation of the abstract
545 <code>ProfileInfo</code> interface.
546 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]547</div>
548
549<!-------------------------------------------------------------------------- -->
550<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]551 <a name="postdomfrontier">-postdomfrontier: Post-Dominance Frontier Construction</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]552</div>
553<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]554 <p>
555 This pass is a simple post-dominator construction algorithm for finding
556 post-dominator frontiers.
557 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]558</div>
559
560<!-------------------------------------------------------------------------- -->
561<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]562 <a name="postdomtree">-postdomtree: Post-Dominator Tree Construction</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]563</div>
564<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]565 <p>
566 This pass is a simple post-dominator construction algorithm for finding
567 post-dominators.
568 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]569</div>
570
571<!-------------------------------------------------------------------------- -->
572<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]573 <a name="print-alias-sets">-print-alias-sets: Alias Set Printer</a>
Duncan Sands3ee8fc92008-09-23 12:47:39 +0000[diff] [blame]574</div>
575<div class="doc_text">
576 <p>Yet to be written.</p>
577</div>
578
579<!-------------------------------------------------------------------------- -->
580<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]581 <a name="print-callgraph">-print-callgraph: Print a call graph</a>
Duncan Sands3ee8fc92008-09-23 12:47:39 +0000[diff] [blame]582</div>
583<div class="doc_text">
584 <p>
585 This pass, only available in <code>opt</code>, prints the call graph to
Dan Gohman52fdaed2010-08-20 01:03:44 +0000[diff] [blame]586 standard error in a human-readable form.
Duncan Sands3ee8fc92008-09-23 12:47:39 +0000[diff] [blame]587 </p>
588</div>
589
590<!-------------------------------------------------------------------------- -->
591<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]592 <a name="print-callgraph-sccs">-print-callgraph-sccs: Print SCCs of the Call Graph</a>
Duncan Sands3ee8fc92008-09-23 12:47:39 +0000[diff] [blame]593</div>
594<div class="doc_text">
595 <p>
596 This pass, only available in <code>opt</code>, prints the SCCs of the call
Dan Gohman52fdaed2010-08-20 01:03:44 +0000[diff] [blame]597 graph to standard error in a human-readable form.
Duncan Sands3ee8fc92008-09-23 12:47:39 +0000[diff] [blame]598 </p>
599</div>
600
601<!-------------------------------------------------------------------------- -->
602<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]603 <a name="print-cfg-sccs">-print-cfg-sccs: Print SCCs of each function CFG</a>
Duncan Sands3ee8fc92008-09-23 12:47:39 +0000[diff] [blame]604</div>
605<div class="doc_text">
606 <p>
607 This pass, only available in <code>opt</code>, prints the SCCs of each
Dan Gohman52fdaed2010-08-20 01:03:44 +0000[diff] [blame]608 function CFG to standard error in a human-readable form.
Duncan Sands3ee8fc92008-09-23 12:47:39 +0000[diff] [blame]609 </p>
610</div>
611
612<!-------------------------------------------------------------------------- -->
613<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]614 <a name="print-dbginfo">-print-dbginfo: Print debug info in human readable form</a>
615</div>
616<div class="doc_text">
Peter Collingbournec3086ba2010-08-06 02:13:25 +0000[diff] [blame]617 <p>Pass that prints instructions, and associated debug info:</p>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]618 <ul>
619
620 <li>source/line/col information</li>
621 <li>original variable name</li>
622 <li>original type name</li>
623 </ul>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]624</div>
625
626<!-------------------------------------------------------------------------- -->
627<div class="doc_subsection">
628 <a name="print-dom-info">-print-dom-info: Dominator Info Printer</a>
629</div>
630<div class="doc_text">
631 <p>Dominator Info Printer.</p>
632</div>
633
634<!-------------------------------------------------------------------------- -->
635<div class="doc_subsection">
636 <a name="print-externalfnconstants">-print-externalfnconstants: Print external fn callsites passed constants</a>
Duncan Sands3ee8fc92008-09-23 12:47:39 +0000[diff] [blame]637</div>
638<div class="doc_text">
639 <p>
640 This pass, only available in <code>opt</code>, prints out call sites to
641 external functions that are called with constant arguments. This can be
642 useful when looking for standard library functions we should constant fold
643 or handle in alias analyses.
644 </p>
645</div>
646
647<!-------------------------------------------------------------------------- -->
648<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]649 <a name="print-function">-print-function: Print function to stderr</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]650</div>
651<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]652 <p>
653 The <code>PrintFunctionPass</code> class is designed to be pipelined with
654 other <code>FunctionPass</code>es, and prints out the functions of the module
655 as they are processed.
656 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]657</div>
658
659<!-------------------------------------------------------------------------- -->
660<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]661 <a name="print-module">-print-module: Print module to stderr</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]662</div>
663<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]664 <p>
665 This pass simply prints out the entire module when it is executed.
666 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]667</div>
668
669<!-------------------------------------------------------------------------- -->
670<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]671 <a name="print-used-types">-print-used-types: Find Used Types</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]672</div>
673<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]674 <p>
675 This pass is used to seek out all of the types in use by the program. Note
676 that this analysis explicitly does not include types only used by the symbol
677 table.
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]678</div>
679
680<!-------------------------------------------------------------------------- -->
681<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]682 <a name="profile-estimator">-profile-estimator: Estimate profiling information</a>
683</div>
684<div class="doc_text">
685 <p>Profiling information that estimates the profiling information
686 in a very crude and unimaginative way.
687 </p>
688</div>
689
690<!-------------------------------------------------------------------------- -->
691<div class="doc_subsection">
692 <a name="profile-loader">-profile-loader: Load profile information from llvmprof.out</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]693</div>
694<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]695 <p>
696 A concrete implementation of profiling information that loads the information
697 from a profile dump file.
698 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]699</div>
700
701<!-------------------------------------------------------------------------- -->
702<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]703 <a name="profile-verifier">-profile-verifier: Verify profiling information</a>
704</div>
705<div class="doc_text">
706 <p>Pass that checks profiling information for plausibility.</p>
707</div>
Tobias Grosserf96b0062010-07-22 07:46:31 +0000[diff] [blame]708<div class="doc_subsection">
Eli Friedmande8ec5b2011-03-19 04:55:29 +0000[diff] [blame^]709 <a name="regions">-regions: Detect single entry single exit regions</a>
Tobias Grosserf96b0062010-07-22 07:46:31 +0000[diff] [blame]710</div>
711<div class="doc_text">
712 <p>
713 The <code>RegionInfo</code> pass detects single entry single exit regions in a
714 function, where a region is defined as any subgraph that is connected to the
715 remaining graph at only two spots. Furthermore, an hierarchical region tree is
716 built.
717 </p>
718</div>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]719
720<!-------------------------------------------------------------------------- -->
721<div class="doc_subsection">
722 <a name="scalar-evolution">-scalar-evolution: Scalar Evolution Analysis</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]723</div>
724<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]725 <p>
726 The <code>ScalarEvolution</code> analysis can be used to analyze and
727 catagorize scalar expressions in loops. It specializes in recognizing general
728 induction variables, representing them with the abstract and opaque
729 <code>SCEV</code> class. Given this analysis, trip counts of loops and other
730 important properties can be obtained.
731 </p>
732
733 <p>
734 This analysis is primarily useful for induction variable substitution and
735 strength reduction.
736 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]737</div>
738
739<!-------------------------------------------------------------------------- -->
740<div class="doc_subsection">
Eli Friedmande8ec5b2011-03-19 04:55:29 +0000[diff] [blame^]741 <a name="scev-aa">-scev-aa: ScalarEvolution-based Alias Analysis</a>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]742</div>
743<div class="doc_text">
744 <p>Simple alias analysis implemented in terms of ScalarEvolution queries.
745
746 This differs from traditional loop dependence analysis in that it tests
747 for dependencies within a single iteration of a loop, rather than
748 dependencies between different iterations.
749
750 ScalarEvolution has a more complete understanding of pointer arithmetic
751 than BasicAliasAnalysis' collection of ad-hoc analyses.
752 </p>
753</div>
754
755<!-------------------------------------------------------------------------- -->
756<div class="doc_subsection">
757 <a name="strip-dead-debug-info">-strip-dead-debug-info: Strip debug info for unused symbols</a>
758</div>
759<div class="doc_text">
760 <p>
761 performs code stripping. this transformation can delete:
762 </p>
763
764 <ol>
765 <li>names for virtual registers</li>
766 <li>symbols for internal globals and functions</li>
767 <li>debug information</li>
768 </ol>
769
770 <p>
771 note that this transformation makes code much less readable, so it should
772 only be used in situations where the <tt>strip</tt> utility would be used,
773 such as reducing code size or making it harder to reverse engineer code.
774 </p>
775</div>
776
777<!-------------------------------------------------------------------------- -->
778<div class="doc_subsection">
779 <a name="targetdata">-targetdata: Target Data Layout</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]780</div>
781<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]782 <p>Provides other passes access to information on how the size and alignment
783 required by the the target ABI for various data types.</p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]784</div>
785
786<!-- ======================================================================= -->
787<div class="doc_section"> <a name="transform">Transform Passes</a></div>
788<div class="doc_text">
789 <p>This section describes the LLVM Transform Passes.</p>
790</div>
791
792<!-------------------------------------------------------------------------- -->
793<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]794 <a name="adce">-adce: Aggressive Dead Code Elimination</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]795</div>
796<div class="doc_text">
Reid Spenceraf4af3a2007-03-27 02:49:31 +0000[diff] [blame]797 <p>ADCE aggressively tries to eliminate code. This pass is similar to
798 <a href="#dce">DCE</a> but it assumes that values are dead until proven
799 otherwise. This is similar to <a href="#sccp">SCCP</a>, except applied to
800 the liveness of values.</p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]801</div>
802
803<!-------------------------------------------------------------------------- -->
804<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]805 <a name="always-inline">-always-inline: Inliner for always_inline functions</a>
806</div>
807<div class="doc_text">
808 <p>A custom inliner that handles only functions that are marked as
809 "always inline".</p>
810</div>
811
812<!-------------------------------------------------------------------------- -->
813<div class="doc_subsection">
814 <a name="argpromotion">-argpromotion: Promote 'by reference' arguments to scalars</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]815</div>
816<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]817 <p>
818 This pass promotes "by reference" arguments to be "by value" arguments. In
819 practice, this means looking for internal functions that have pointer
820 arguments. If it can prove, through the use of alias analysis, that an
821 argument is *only* loaded, then it can pass the value into the function
822 instead of the address of the value. This can cause recursive simplification
823 of code and lead to the elimination of allocas (especially in C++ template
824 code like the STL).
825 </p>
826
827 <p>
828 This pass also handles aggregate arguments that are passed into a function,
829 scalarizing them if the elements of the aggregate are only loaded. Note that
830 it refuses to scalarize aggregates which would require passing in more than
831 three operands to the function, because passing thousands of operands for a
832 large array or structure is unprofitable!
833 </p>
834
835 <p>
836 Note that this transformation could also be done for arguments that are only
837 stored to (returning the value instead), but does not currently. This case
838 would be best handled when and if LLVM starts supporting multiple return
839 values from functions.
840 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]841</div>
842
843<!-------------------------------------------------------------------------- -->
844<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]845 <a name="block-placement">-block-placement: Profile Guided Basic Block Placement</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]846</div>
847<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]848 <p>This pass is a very simple profile guided basic block placement algorithm.
849 The idea is to put frequently executed blocks together at the start of the
850 function and hopefully increase the number of fall-through conditional
851 branches. If there is no profile information for a particular function, this
852 pass basically orders blocks in depth-first order.</p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]853</div>
854
855<!-------------------------------------------------------------------------- -->
856<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]857 <a name="break-crit-edges">-break-crit-edges: Break critical edges in CFG</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]858</div>
859<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]860 <p>
861 Break all of the critical edges in the CFG by inserting a dummy basic block.
862 It may be "required" by passes that cannot deal with critical edges. This
863 transformation obviously invalidates the CFG, but can update forward dominator
864 (set, immediate dominators, tree, and frontier) information.
865 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]866</div>
867
868<!-------------------------------------------------------------------------- -->
869<div class="doc_subsection">
Eli Friedmande8ec5b2011-03-19 04:55:29 +0000[diff] [blame^]870 <a name="codegenprepare">-codegenprepare: Optimize for code generation</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]871</div>
872<div class="doc_text">
Gordon Henriksena8a118b2008-05-08 17:46:35 +0000[diff] [blame]873 This pass munges the code in the input function to better prepare it for
874 SelectionDAG-based code generation. This works around limitations in it's
875 basic-block-at-a-time approach. It should eventually be removed.
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]876</div>
877
878<!-------------------------------------------------------------------------- -->
879<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]880 <a name="constmerge">-constmerge: Merge Duplicate Global Constants</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]881</div>
882<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]883 <p>
884 Merges duplicate global constants together into a single constant that is
885 shared. This is useful because some passes (ie TraceValues) insert a lot of
886 string constants into the program, regardless of whether or not an existing
887 string is available.
888 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]889</div>
890
891<!-------------------------------------------------------------------------- -->
892<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]893 <a name="constprop">-constprop: Simple constant propagation</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]894</div>
895<div class="doc_text">
Reid Spenceraf4af3a2007-03-27 02:49:31 +0000[diff] [blame]896 <p>This file implements constant propagation and merging. It looks for
897 instructions involving only constant operands and replaces them with a
Gordon Henriksenddaa61d2007-10-25 08:58:56 +0000[diff] [blame]898 constant value instead of an instruction. For example:</p>
899 <blockquote><pre>add i32 1, 2</pre></blockquote>
900 <p>becomes</p>
901 <blockquote><pre>i32 3</pre></blockquote>
Reid Spenceraf4af3a2007-03-27 02:49:31 +0000[diff] [blame]902 <p>NOTE: this pass has a habit of making definitions be dead. It is a good
903 idea to to run a <a href="#die">DIE</a> (Dead Instruction Elimination) pass
904 sometime after running this pass.</p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]905</div>
906
907<!-------------------------------------------------------------------------- -->
908<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]909 <a name="dce">-dce: Dead Code Elimination</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]910</div>
911<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]912 <p>
913 Dead code elimination is similar to <a href="#die">dead instruction
914 elimination</a>, but it rechecks instructions that were used by removed
915 instructions to see if they are newly dead.
916 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]917</div>
918
919<!-------------------------------------------------------------------------- -->
920<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]921 <a name="deadargelim">-deadargelim: Dead Argument Elimination</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]922</div>
923<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]924 <p>
925 This pass deletes dead arguments from internal functions. Dead argument
926 elimination removes arguments which are directly dead, as well as arguments
927 only passed into function calls as dead arguments of other functions. This
928 pass also deletes dead arguments in a similar way.
929 </p>
930
931 <p>
932 This pass is often useful as a cleanup pass to run after aggressive
933 interprocedural passes, which add possibly-dead arguments.
934 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]935</div>
936
937<!-------------------------------------------------------------------------- -->
938<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]939 <a name="deadtypeelim">-deadtypeelim: Dead Type Elimination</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]940</div>
941<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]942 <p>
943 This pass is used to cleanup the output of GCC. It eliminate names for types
944 that are unused in the entire translation unit, using the <a
945 href="#findusedtypes">find used types</a> pass.
946 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]947</div>
948
949<!-------------------------------------------------------------------------- -->
950<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]951 <a name="die">-die: Dead Instruction Elimination</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]952</div>
953<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]954 <p>
955 Dead instruction elimination performs a single pass over the function,
956 removing instructions that are obviously dead.
957 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]958</div>
959
960<!-------------------------------------------------------------------------- -->
961<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]962 <a name="dse">-dse: Dead Store Elimination</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]963</div>
964<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]965 <p>
966 A trivial dead store elimination that only considers basic-block local
967 redundant stores.
968 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]969</div>
970
971<!-------------------------------------------------------------------------- -->
972<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]973 <a name="functionattrs">-functionattrs: Deduce function attributes</a>
974</div>
975<div class="doc_text">
976 <p>A simple interprocedural pass which walks the call-graph, looking for
977 functions which do not access or only read non-local memory, and marking them
978 readnone/readonly. In addition, it marks function arguments (of pointer type)
979 'nocapture' if a call to the function does not create any copies of the pointer
980 value that outlive the call. This more or less means that the pointer is only
981 dereferenced, and not returned from the function or stored in a global.
982 This pass is implemented as a bottom-up traversal of the call-graph.
983 </p>
984</div>
985
986<!-------------------------------------------------------------------------- -->
987<div class="doc_subsection">
988 <a name="globaldce">-globaldce: Dead Global Elimination</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]989</div>
990<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]991 <p>
992 This transform is designed to eliminate unreachable internal globals from the
993 program. It uses an aggressive algorithm, searching out globals that are
994 known to be alive. After it finds all of the globals which are needed, it
995 deletes whatever is left over. This allows it to delete recursive chunks of
996 the program which are unreachable.
997 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]998</div>
999
1000<!-------------------------------------------------------------------------- -->
1001<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1002 <a name="globalopt">-globalopt: Global Variable Optimizer</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1003</div>
1004<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]1005 <p>
1006 This pass transforms simple global variables that never have their address
1007 taken. If obviously true, it marks read/write globals as constant, deletes
1008 variables only stored to, etc.
1009 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1010</div>
1011
1012<!-------------------------------------------------------------------------- -->
1013<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1014 <a name="gvn">-gvn: Global Value Numbering</a>
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]1015</div>
1016<div class="doc_text">
1017 <p>
Chris Lattner60f03402009-10-10 18:40:48 +0000[diff] [blame]1018 This pass performs global value numbering to eliminate fully and partially
1019 redundant instructions. It also performs redundant load elimination.
Matthijs Kooijman845f5242008-06-05 07:55:49 +0000[diff] [blame]1020 </p>
Gordon Henriksen0e15dc22007-10-25 10:18:27 +0000[diff] [blame]1021</div>
1022
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]1023<!-------------------------------------------------------------------------- -->
1024<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1025 <a name="indvars">-indvars: Canonicalize Induction Variables</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1026</div>
1027<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]1028 <p>
1029 This transformation analyzes and transforms the induction variables (and
1030 computations derived from them) into simpler forms suitable for subsequent
1031 analysis and transformation.
1032 </p>
1033
1034 <p>
1035 This transformation makes the following changes to each loop with an
1036 identifiable induction variable:
1037 </p>
1038
1039 <ol>
1040 <li>All loops are transformed to have a <em>single</em> canonical
1041 induction variable which starts at zero and steps by one.</li>
1042 <li>The canonical induction variable is guaranteed to be the first PHI node
1043 in the loop header block.</li>
1044 <li>Any pointer arithmetic recurrences are raised to use array
1045 subscripts.</li>
1046 </ol>
1047
1048 <p>
1049 If the trip count of a loop is computable, this pass also makes the following
1050 changes:
1051 </p>
1052
1053 <ol>
1054 <li>The exit condition for the loop is canonicalized to compare the
1055 induction value against the exit value. This turns loops like:
1056 <blockquote><pre>for (i = 7; i*i < 1000; ++i)</pre></blockquote>
1057 into
1058 <blockquote><pre>for (i = 0; i != 25; ++i)</pre></blockquote></li>
1059 <li>Any use outside of the loop of an expression derived from the indvar
1060 is changed to compute the derived value outside of the loop, eliminating
1061 the dependence on the exit value of the induction variable. If the only
1062 purpose of the loop is to compute the exit value of some derived
1063 expression, this transformation will make the loop dead.</li>
Gordon Henriksene626bbe2007-11-04 16:17:00 +0000[diff] [blame]1064 </ol>
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]1065
1066 <p>
1067 This transformation should be followed by strength reduction after all of the
1068 desired loop transformations have been performed. Additionally, on targets
1069 where it is profitable, the loop could be transformed to count down to zero
1070 (the "do loop" optimization).
1071 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1072</div>
1073
1074<!-------------------------------------------------------------------------- -->
1075<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1076 <a name="inline">-inline: Function Integration/Inlining</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1077</div>
1078<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]1079 <p>
1080 Bottom-up inlining of functions into callees.
1081 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1082</div>
1083
1084<!-------------------------------------------------------------------------- -->
1085<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1086 <a name="insert-edge-profiling">-insert-edge-profiling: Insert instrumentation for edge profiling</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1087</div>
1088<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]1089 <p>
1090 This pass instruments the specified program with counters for edge profiling.
1091 Edge profiling can give a reasonable approximation of the hot paths through a
1092 program, and is used for a wide variety of program transformations.
1093 </p>
1094
1095 <p>
1096 Note that this implementation is very naïve. It inserts a counter for
1097 <em>every</em> edge in the program, instead of using control flow information
1098 to prune the number of counters inserted.
1099 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1100</div>
1101
1102<!-------------------------------------------------------------------------- -->
1103<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1104 <a name="insert-optimal-edge-profiling">-insert-optimal-edge-profiling: Insert optimal instrumentation for edge profiling</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1105</div>
1106<div class="doc_text">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1107 <p>This pass instruments the specified program with counters for edge profiling.
1108 Edge profiling can give a reasonable approximation of the hot paths through a
1109 program, and is used for a wide variety of program transformations.
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]1110 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1111</div>
1112
1113<!-------------------------------------------------------------------------- -->
1114<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1115 <a name="instcombine">-instcombine: Combine redundant instructions</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1116</div>
1117<div class="doc_text">
Gordon Henriksen55cbec32007-10-26 03:03:51 +0000[diff] [blame]1118 <p>
1119 Combine instructions to form fewer, simple
1120 instructions. This pass does not modify the CFG This pass is where algebraic
1121 simplification happens.
1122 </p>
1123
1124 <p>
1125 This pass combines things like:
1126 </p>
1127
1128<blockquote><pre
1129>%Y = add i32 %X, 1
1130%Z = add i32 %Y, 1</pre></blockquote>
1131
1132 <p>
1133 into:
1134 </p>
1135
1136<blockquote><pre
1137>%Z = add i32 %X, 2</pre></blockquote>
1138
1139 <p>
1140 This is a simple worklist driven algorithm.
1141 </p>
1142
1143 <p>
1144 This pass guarantees that the following canonicalizations are performed on
1145 the program:
1146 </p>
1147
1148 <ul>
1149 <li>If a binary operator has a constant operand, it is moved to the right-
1150 hand side.</li>
1151 <li>Bitwise operators with constant operands are always grouped so that
1152 shifts are performed first, then <code>or</code>s, then
1153 <code>and</code>s, then <code>xor</code>s.</li>
1154 <li>Compare instructions are converted from <code>&lt;</code>,
1155 <code>&gt;</code>, <code>≤</code>, or <code>≥</code> to
1156 <code>=</code> or <code>≠</code> if possible.</li>
1157 <li>All <code>cmp</code> instructions on boolean values are replaced with
1158 logical operations.</li>
1159 <li><code>add <var>X</var>, <var>X</var></code> is represented as
1160 <code>mul <var>X</var>, 2</code> ⇒ <code>shl <var>X</var>, 1</code></li>
1161 <li>Multiplies with a constant power-of-two argument are transformed into
1162 shifts.</li>
1163 <li>… etc.</li>
1164 </ul>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1165</div>
1166
1167<!-------------------------------------------------------------------------- -->
1168<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1169 <a name="internalize">-internalize: Internalize Global Symbols</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1170</div>
1171<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1172 <p>
1173 This pass loops over all of the functions in the input module, looking for a
1174 main function. If a main function is found, all other functions and all
1175 global variables with initializers are marked as internal.
1176 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1177</div>
1178
1179<!-------------------------------------------------------------------------- -->
1180<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1181 <a name="ipconstprop">-ipconstprop: Interprocedural constant propagation</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1182</div>
1183<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1184 <p>
1185 This pass implements an <em>extremely</em> simple interprocedural constant
1186 propagation pass. It could certainly be improved in many different ways,
1187 like using a worklist. This pass makes arguments dead, but does not remove
1188 them. The existing dead argument elimination pass should be run after this
1189 to clean up the mess.
1190 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1191</div>
1192
1193<!-------------------------------------------------------------------------- -->
1194<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1195 <a name="ipsccp">-ipsccp: Interprocedural Sparse Conditional Constant Propagation</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1196</div>
1197<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1198 <p>
1199 An interprocedural variant of <a href="#sccp">Sparse Conditional Constant
1200 Propagation</a>.
1201 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1202</div>
1203
1204<!-------------------------------------------------------------------------- -->
1205<div class="doc_subsection">
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]1206 <a name="jump-threading">-jump-threading: Jump Threading</a>
Gordon Henriksena8a118b2008-05-08 17:46:35 +0000[diff] [blame]1207</div>
1208<div class="doc_text">
1209 <p>
1210 Jump threading tries to find distinct threads of control flow running through
1211 a basic block. This pass looks at blocks that have multiple predecessors and
1212 multiple successors. If one or more of the predecessors of the block can be
1213 proven to always cause a jump to one of the successors, we forward the edge
1214 from the predecessor to the successor by duplicating the contents of this
1215 block.
1216 </p>
1217 <p>
1218 An example of when this can occur is code like this:
1219 </p>
1220
1221 <pre
1222>if () { ...
1223 X = 4;
1224}
1225if (X &lt; 3) {</pre>
1226
1227 <p>
1228 In this case, the unconditional branch at the end of the first if can be
1229 revectored to the false side of the second if.
1230 </p>
1231</div>
1232
1233<!-------------------------------------------------------------------------- -->
1234<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1235 <a name="lcssa">-lcssa: Loop-Closed SSA Form Pass</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1236</div>
1237<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1238 <p>
1239 This pass transforms loops by placing phi nodes at the end of the loops for
1240 all values that are live across the loop boundary. For example, it turns
1241 the left into the right code:
1242 </p>
1243
1244 <pre
1245>for (...) for (...)
1246 if (c) if (c)
1247 X1 = ... X1 = ...
1248 else else
1249 X2 = ... X2 = ...
1250 X3 = phi(X1, X2) X3 = phi(X1, X2)
1251... = X3 + 4 X4 = phi(X3)
1252 ... = X4 + 4</pre>
1253
1254 <p>
1255 This is still valid LLVM; the extra phi nodes are purely redundant, and will
1256 be trivially eliminated by <code>InstCombine</code>. The major benefit of
1257 this transformation is that it makes many other loop optimizations, such as
1258 LoopUnswitching, simpler.
1259 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1260</div>
1261
1262<!-------------------------------------------------------------------------- -->
1263<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1264 <a name="licm">-licm: Loop Invariant Code Motion</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1265</div>
1266<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1267 <p>
1268 This pass performs loop invariant code motion, attempting to remove as much
1269 code from the body of a loop as possible. It does this by either hoisting
1270 code into the preheader block, or by sinking code to the exit blocks if it is
1271 safe. This pass also promotes must-aliased memory locations in the loop to
1272 live in registers, thus hoisting and sinking "invariant" loads and stores.
1273 </p>
1274
1275 <p>
1276 This pass uses alias analysis for two purposes:
1277 </p>
1278
1279 <ul>
1280 <li>Moving loop invariant loads and calls out of loops. If we can determine
1281 that a load or call inside of a loop never aliases anything stored to,
1282 we can hoist it or sink it like any other instruction.</li>
1283 <li>Scalar Promotion of Memory - If there is a store instruction inside of
1284 the loop, we try to move the store to happen AFTER the loop instead of
1285 inside of the loop. This can only happen if a few conditions are true:
1286 <ul>
1287 <li>The pointer stored through is loop invariant.</li>
1288 <li>There are no stores or loads in the loop which <em>may</em> alias
1289 the pointer. There are no calls in the loop which mod/ref the
1290 pointer.</li>
1291 </ul>
1292 If these conditions are true, we can promote the loads and stores in the
1293 loop of the pointer to use a temporary alloca'd variable. We then use
1294 the mem2reg functionality to construct the appropriate SSA form for the
1295 variable.</li>
1296 </ul>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1297</div>
Gordon Henriksena8a118b2008-05-08 17:46:35 +0000[diff] [blame]1298<!-------------------------------------------------------------------------- -->
1299<div class="doc_subsection">
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]1300 <a name="loop-deletion">-loop-deletion: Delete dead loops</a>
Gordon Henriksena8a118b2008-05-08 17:46:35 +0000[diff] [blame]1301</div>
1302<div class="doc_text">
1303 <p>
1304 This file implements the Dead Loop Deletion Pass. This pass is responsible
1305 for eliminating loops with non-infinite computable trip counts that have no
1306 side effects or volatile instructions, and do not contribute to the
1307 computation of the function's return value.
1308 </p>
1309</div>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1310
1311<!-------------------------------------------------------------------------- -->
1312<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1313 <a name="loop-extract">-loop-extract: Extract loops into new functions</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1314</div>
1315<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1316 <p>
1317 A pass wrapper around the <code>ExtractLoop()</code> scalar transformation to
1318 extract each top-level loop into its own new function. If the loop is the
1319 <em>only</em> loop in a given function, it is not touched. This is a pass most
1320 useful for debugging via bugpoint.
1321 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1322</div>
1323
1324<!-------------------------------------------------------------------------- -->
1325<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1326 <a name="loop-extract-single">-loop-extract-single: Extract at most one loop into a new function</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1327</div>
1328<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1329 <p>
1330 Similar to <a href="#loop-extract">Extract loops into new functions</a>,
1331 this pass extracts one natural loop from the program into a function if it
1332 can. This is used by bugpoint.
1333 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1334</div>
1335
1336<!-------------------------------------------------------------------------- -->
1337<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1338 <a name="loop-reduce">-loop-reduce: Loop Strength Reduction</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1339</div>
1340<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1341 <p>
1342 This pass performs a strength reduction on array references inside loops that
1343 have as one or more of their components the loop induction variable. This is
1344 accomplished by creating a new value to hold the initial value of the array
1345 access for the first iteration, and then creating a new GEP instruction in
1346 the loop to increment the value by the appropriate amount.
1347 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1348</div>
1349
1350<!-------------------------------------------------------------------------- -->
1351<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1352 <a name="loop-rotate">-loop-rotate: Rotate Loops</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1353</div>
1354<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1355 <p>A simple loop rotation transformation.</p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1356</div>
1357
1358<!-------------------------------------------------------------------------- -->
1359<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1360 <a name="loop-unroll">-loop-unroll: Unroll loops</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1361</div>
1362<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1363 <p>
1364 This pass implements a simple loop unroller. It works best when loops have
1365 been canonicalized by the <a href="#indvars"><tt>-indvars</tt></a> pass,
1366 allowing it to determine the trip counts of loops easily.
1367 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1368</div>
1369
1370<!-------------------------------------------------------------------------- -->
1371<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1372 <a name="loop-unswitch">-loop-unswitch: Unswitch loops</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1373</div>
1374<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1375 <p>
1376 This pass transforms loops that contain branches on loop-invariant conditions
1377 to have multiple loops. For example, it turns the left into the right code:
1378 </p>
1379
1380 <pre
1381>for (...) if (lic)
1382 A for (...)
1383 if (lic) A; B; C
1384 B else
1385 C for (...)
1386 A; C</pre>
1387
1388 <p>
1389 This can increase the size of the code exponentially (doubling it every time
1390 a loop is unswitched) so we only unswitch if the resultant code will be
1391 smaller than a threshold.
1392 </p>
1393
1394 <p>
1395 This pass expects LICM to be run before it to hoist invariant conditions out
1396 of the loop, to make the unswitching opportunity obvious.
1397 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1398</div>
1399
1400<!-------------------------------------------------------------------------- -->
1401<div class="doc_subsection">
Tobias Grossereafbe652011-02-13 20:57:25 +0000[diff] [blame]1402 <a name="loop-simplify">-loop-simplify: Canonicalize natural loops</a>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]1403</div>
1404<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1405 <p>
1406 This pass performs several transformations to transform natural loops into a
1407 simpler form, which makes subsequent analyses and transformations simpler and
1408 more effective.
1409 </p>
1410
1411 <p>
1412 Loop pre-header insertion guarantees that there is a single, non-critical
1413 entry edge from outside of the loop to the loop header. This simplifies a
1414 number of analyses and transformations, such as LICM.
1415 </p>
1416
1417 <p>
1418 Loop exit-block insertion guarantees that all exit blocks from the loop
1419 (blocks which are outside of the loop that have predecessors inside of the
1420 loop) only have predecessors from inside of the loop (and are thus dominated
1421 by the loop header). This simplifies transformations such as store-sinking
1422 that are built into LICM.
1423 </p>
1424
1425 <p>
1426 This pass also guarantees that loops will have exactly one backedge.
1427 </p>
1428
1429 <p>
1430 Note that the simplifycfg pass will clean up blocks which are split out but
1431 end up being unnecessary, so usage of this pass should not pessimize
1432 generated code.
1433 </p>
1434
1435 <p>
1436 This pass obviously modifies the CFG, but updates loop information and
1437 dominator information.
1438 </p>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]1439</div>
1440
1441<!-------------------------------------------------------------------------- -->
1442<div class="doc_subsection">
Eli Friedmande8ec5b2011-03-19 04:55:29 +0000[diff] [blame^]1443 <a name="loweratomic">-loweratomic: Lower atomic intrinsics to non-atomic form</a>
Peter Collingbourne3bababf2010-08-03 16:19:16 +0000[diff] [blame]1444</div>
1445<div class="doc_text">
1446 <p>
1447 This pass lowers atomic intrinsics to non-atomic form for use in a known
1448 non-preemptible environment.
1449 </p>
1450
1451 <p>
1452 The pass does not verify that the environment is non-preemptible (in
1453 general this would require knowledge of the entire call graph of the
1454 program including any libraries which may not be available in bitcode form);
1455 it simply lowers every atomic intrinsic.
1456 </p>
1457</div>
1458
1459<!-------------------------------------------------------------------------- -->
1460<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1461 <a name="lowerinvoke">-lowerinvoke: Lower invoke and unwind, for unwindless code generators</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1462</div>
1463<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1464 <p>
1465 This transformation is designed for use by code generators which do not yet
1466 support stack unwinding. This pass supports two models of exception handling
1467 lowering, the 'cheap' support and the 'expensive' support.
1468 </p>
1469
1470 <p>
1471 'Cheap' exception handling support gives the program the ability to execute
1472 any program which does not "throw an exception", by turning 'invoke'
1473 instructions into calls and by turning 'unwind' instructions into calls to
1474 abort(). If the program does dynamically use the unwind instruction, the
1475 program will print a message then abort.
1476 </p>
1477
1478 <p>
1479 'Expensive' exception handling support gives the full exception handling
1480 support to the program at the cost of making the 'invoke' instruction
1481 really expensive. It basically inserts setjmp/longjmp calls to emulate the
1482 exception handling as necessary.
1483 </p>
1484
1485 <p>
1486 Because the 'expensive' support slows down programs a lot, and EH is only
1487 used for a subset of the programs, it must be specifically enabled by the
1488 <tt>-enable-correct-eh-support</tt> option.
1489 </p>
1490
1491 <p>
1492 Note that after this pass runs the CFG is not entirely accurate (exceptional
1493 control flow edges are not correct anymore) so only very simple things should
1494 be done after the lowerinvoke pass has run (like generation of native code).
1495 This should not be used as a general purpose "my LLVM-to-LLVM pass doesn't
1496 support the invoke instruction yet" lowering pass.
1497 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1498</div>
1499
1500<!-------------------------------------------------------------------------- -->
1501<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1502 <a name="lowersetjmp">-lowersetjmp: Lower Set Jump</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1503</div>
1504<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1505 <p>
1506 Lowers <tt>setjmp</tt> and <tt>longjmp</tt> to use the LLVM invoke and unwind
1507 instructions as necessary.
1508 </p>
1509
1510 <p>
1511 Lowering of <tt>longjmp</tt> is fairly trivial. We replace the call with a
1512 call to the LLVM library function <tt>__llvm_sjljeh_throw_longjmp()</tt>.
1513 This unwinds the stack for us calling all of the destructors for
1514 objects allocated on the stack.
1515 </p>
1516
1517 <p>
1518 At a <tt>setjmp</tt> call, the basic block is split and the <tt>setjmp</tt>
1519 removed. The calls in a function that have a <tt>setjmp</tt> are converted to
1520 invoke where the except part checks to see if it's a <tt>longjmp</tt>
1521 exception and, if so, if it's handled in the function. If it is, then it gets
1522 the value returned by the <tt>longjmp</tt> and goes to where the basic block
1523 was split. <tt>invoke</tt> instructions are handled in a similar fashion with
1524 the original except block being executed if it isn't a <tt>longjmp</tt>
1525 except that is handled by that function.
1526 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1527</div>
1528
1529<!-------------------------------------------------------------------------- -->
1530<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1531 <a name="lowerswitch">-lowerswitch: Lower SwitchInst's to branches</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1532</div>
1533<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1534 <p>
1535 Rewrites <tt>switch</tt> instructions with a sequence of branches, which
1536 allows targets to get away with not implementing the switch instruction until
1537 it is convenient.
1538 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1539</div>
1540
1541<!-------------------------------------------------------------------------- -->
1542<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1543 <a name="mem2reg">-mem2reg: Promote Memory to Register</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1544</div>
1545<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1546 <p>
1547 This file promotes memory references to be register references. It promotes
1548 <tt>alloca</tt> instructions which only have <tt>load</tt>s and
1549 <tt>store</tt>s as uses. An <tt>alloca</tt> is transformed by using dominator
1550 frontiers to place <tt>phi</tt> nodes, then traversing the function in
1551 depth-first order to rewrite <tt>load</tt>s and <tt>store</tt>s as
1552 appropriate. This is just the standard SSA construction algorithm to construct
1553 "pruned" SSA form.
1554 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1555</div>
1556
1557<!-------------------------------------------------------------------------- -->
1558<div class="doc_subsection">
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]1559 <a name="memcpyopt">-memcpyopt: MemCpy Optimization</a>
Gordon Henriksena8a118b2008-05-08 17:46:35 +0000[diff] [blame]1560</div>
1561<div class="doc_text">
1562 <p>
1563 This pass performs various transformations related to eliminating memcpy
1564 calls, or transforming sets of stores into memset's.
1565 </p>
1566</div>
1567
1568<!-------------------------------------------------------------------------- -->
1569<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1570 <a name="mergefunc">-mergefunc: Merge Functions</a>
1571</div>
1572<div class="doc_text">
1573 <p>This pass looks for equivalent functions that are mergable and folds them.
1574
1575 A hash is computed from the function, based on its type and number of
1576 basic blocks.
1577
1578 Once all hashes are computed, we perform an expensive equality comparison
1579 on each function pair. This takes n^2/2 comparisons per bucket, so it's
1580 important that the hash function be high quality. The equality comparison
1581 iterates through each instruction in each basic block.
1582
1583 When a match is found the functions are folded. If both functions are
1584 overridable, we move the functionality into a new internal function and
1585 leave two overridable thunks to it.
1586 </p>
1587</div>
1588
1589<!-------------------------------------------------------------------------- -->
1590<div class="doc_subsection">
1591 <a name="mergereturn">-mergereturn: Unify function exit nodes</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1592</div>
1593<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1594 <p>
1595 Ensure that functions have at most one <tt>ret</tt> instruction in them.
1596 Additionally, it keeps track of which node is the new exit node of the CFG.
1597 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1598</div>
1599
1600<!-------------------------------------------------------------------------- -->
1601<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1602 <a name="partial-inliner">-partial-inliner: Partial Inliner</a>
1603</div>
1604<div class="doc_text">
1605 <p>This pass performs partial inlining, typically by inlining an if
1606 statement that surrounds the body of the function.
1607 </p>
1608</div>
1609
1610<!-------------------------------------------------------------------------- -->
1611<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1612 <a name="prune-eh">-prune-eh: Remove unused exception handling info</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1613</div>
1614<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1615 <p>
1616 This file implements a simple interprocedural pass which walks the call-graph,
1617 turning <tt>invoke</tt> instructions into <tt>call</tt> instructions if and
1618 only if the callee cannot throw an exception. It implements this as a
1619 bottom-up traversal of the call-graph.
1620 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1621</div>
1622
1623<!-------------------------------------------------------------------------- -->
1624<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1625 <a name="reassociate">-reassociate: Reassociate expressions</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1626</div>
1627<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1628 <p>
1629 This pass reassociates commutative expressions in an order that is designed
1630 to promote better constant propagation, GCSE, LICM, PRE, etc.
1631 </p>
1632
1633 <p>
1634 For example: 4 + (<var>x</var> + 5) ⇒ <var>x</var> + (4 + 5)
1635 </p>
1636
1637 <p>
1638 In the implementation of this algorithm, constants are assigned rank = 0,
1639 function arguments are rank = 1, and other values are assigned ranks
1640 corresponding to the reverse post order traversal of current function
1641 (starting at 2), which effectively gives values in deep loops higher rank
1642 than values not in loops.
1643 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1644</div>
1645
1646<!-------------------------------------------------------------------------- -->
1647<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1648 <a name="reg2mem">-reg2mem: Demote all values to stack slots</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1649</div>
1650<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1651 <p>
1652 This file demotes all registers to memory references. It is intented to be
1653 the inverse of <a href="#mem2reg"><tt>-mem2reg</tt></a>. By converting to
Benjamin Kramer8040cd32009-10-12 14:46:08 +0000[diff] [blame]1654 <tt>load</tt> instructions, the only values live across basic blocks are
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1655 <tt>alloca</tt> instructions and <tt>load</tt> instructions before
1656 <tt>phi</tt> nodes. It is intended that this should make CFG hacking much
1657 easier. To make later hacking easier, the entry block is split into two, such
1658 that all introduced <tt>alloca</tt> instructions (and nothing else) are in the
1659 entry block.
1660 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1661</div>
1662
1663<!-------------------------------------------------------------------------- -->
1664<div class="doc_subsection">
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]1665 <a name="scalarrepl">-scalarrepl: Scalar Replacement of Aggregates (DT)</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1666</div>
1667<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1668 <p>
1669 The well-known scalar replacement of aggregates transformation. This
1670 transform breaks up <tt>alloca</tt> instructions of aggregate type (structure
1671 or array) into individual <tt>alloca</tt> instructions for each member if
1672 possible. Then, if possible, it transforms the individual <tt>alloca</tt>
1673 instructions into nice clean scalar SSA form.
1674 </p>
1675
1676 <p>
1677 This combines a simple scalar replacement of aggregates algorithm with the <a
1678 href="#mem2reg"><tt>mem2reg</tt></a> algorithm because often interact,
1679 especially for C++ programs. As such, iterating between <tt>scalarrepl</tt>,
1680 then <a href="#mem2reg"><tt>mem2reg</tt></a> until we run out of things to
1681 promote works well.
1682 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1683</div>
1684
1685<!-------------------------------------------------------------------------- -->
1686<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1687 <a name="sccp">-sccp: Sparse Conditional Constant Propagation</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1688</div>
1689<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1690 <p>
1691 Sparse conditional constant propagation and merging, which can be summarized
1692 as:
1693 </p>
1694
1695 <ol>
1696 <li>Assumes values are constant unless proven otherwise</li>
1697 <li>Assumes BasicBlocks are dead unless proven otherwise</li>
1698 <li>Proves values to be constant, and replaces them with constants</li>
1699 <li>Proves conditional branches to be unconditional</li>
1700 </ol>
1701
1702 <p>
1703 Note that this pass has a habit of making definitions be dead. It is a good
1704 idea to to run a DCE pass sometime after running this pass.
1705 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1706</div>
1707
1708<!-------------------------------------------------------------------------- -->
1709<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1710 <a name="sink">-sink: Code Sinking</a>
1711</div>
1712<div class="doc_text">
1713 <p>This pass moves instructions into successor blocks, when possible, so that
1714 they aren't executed on paths where their results aren't needed.
1715 </p>
1716</div>
1717
1718<!-------------------------------------------------------------------------- -->
1719<div class="doc_subsection">
1720 <a name="simplify-libcalls">-simplify-libcalls: Simplify well-known library calls</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1721</div>
1722<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1723 <p>
1724 Applies a variety of small optimizations for calls to specific well-known
1725 function calls (e.g. runtime library functions). For example, a call
1726 <tt>exit(3)</tt> that occurs within the <tt>main()</tt> function can be
1727 transformed into simply <tt>return 3</tt>.
1728 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1729</div>
1730
1731<!-------------------------------------------------------------------------- -->
1732<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1733 <a name="simplifycfg">-simplifycfg: Simplify the CFG</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1734</div>
1735<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1736 <p>
1737 Performs dead code elimination and basic block merging. Specifically:
1738 </p>
1739
1740 <ol>
1741 <li>Removes basic blocks with no predecessors.</li>
1742 <li>Merges a basic block into its predecessor if there is only one and the
1743 predecessor only has one successor.</li>
1744 <li>Eliminates PHI nodes for basic blocks with a single predecessor.</li>
1745 <li>Eliminates a basic block that only contains an unconditional
1746 branch.</li>
1747 </ol>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1748</div>
1749
1750<!-------------------------------------------------------------------------- -->
1751<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1752 <a name="strip">-strip: Strip all symbols from a module</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1753</div>
1754<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1755 <p>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1756 performs code stripping. this transformation can delete:
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1757 </p>
1758
1759 <ol>
1760 <li>names for virtual registers</li>
1761 <li>symbols for internal globals and functions</li>
1762 <li>debug information</li>
1763 </ol>
1764
1765 <p>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1766 note that this transformation makes code much less readable, so it should
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1767 only be used in situations where the <tt>strip</tt> utility would be used,
1768 such as reducing code size or making it harder to reverse engineer code.
1769 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1770</div>
1771
1772<!-------------------------------------------------------------------------- -->
1773<div class="doc_subsection">
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]1774 <a name="strip-dead-prototypes">-strip-dead-prototypes: Strip Unused Function Prototypes</a>
Gordon Henriksena8a118b2008-05-08 17:46:35 +0000[diff] [blame]1775</div>
1776<div class="doc_text">
1777 <p>
1778 This pass loops over all of the functions in the input module, looking for
1779 dead declarations and removes them. Dead declarations are declarations of
1780 functions for which no implementation is available (i.e., declarations for
1781 unused library functions).
1782 </p>
1783</div>
1784
1785<!-------------------------------------------------------------------------- -->
1786<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1787 <a name="strip-debug-declare">-strip-debug-declare: Strip all llvm.dbg.declare intrinsics</a>
1788</div>
1789<div class="doc_text">
Peter Collingbournec3086ba2010-08-06 02:13:25 +0000[diff] [blame]1790 <p>This pass implements code stripping. Specifically, it can delete:</p>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1791 <ul>
1792 <li>names for virtual registers</li>
1793 <li>symbols for internal globals and functions</li>
1794 <li>debug information</li>
1795 </ul>
Peter Collingbournec3086ba2010-08-06 02:13:25 +0000[diff] [blame]1796 <p>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1797 Note that this transformation makes code much less readable, so it should
1798 only be used in situations where the 'strip' utility would be used, such as
1799 reducing code size or making it harder to reverse engineer code.
1800 </p>
1801</div>
1802
1803<!-------------------------------------------------------------------------- -->
1804<div class="doc_subsection">
1805 <a name="strip-nondebug">-strip-nondebug: Strip all symbols, except dbg symbols, from a module</a>
1806</div>
1807<div class="doc_text">
Peter Collingbournec3086ba2010-08-06 02:13:25 +0000[diff] [blame]1808 <p>This pass implements code stripping. Specifically, it can delete:</p>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1809 <ul>
1810 <li>names for virtual registers</li>
1811 <li>symbols for internal globals and functions</li>
1812 <li>debug information</li>
1813 </ul>
Peter Collingbournec3086ba2010-08-06 02:13:25 +0000[diff] [blame]1814 <p>
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1815 Note that this transformation makes code much less readable, so it should
1816 only be used in situations where the 'strip' utility would be used, such as
1817 reducing code size or making it harder to reverse engineer code.
1818 </p>
1819</div>
1820
1821<!-------------------------------------------------------------------------- -->
1822<div class="doc_subsection">
1823 <a name="sretpromotion">-sretpromotion: Promote sret arguments</a>
Gordon Henriksena8a118b2008-05-08 17:46:35 +0000[diff] [blame]1824</div>
1825<div class="doc_text">
1826 <p>
1827 This pass finds functions that return a struct (using a pointer to the struct
1828 as the first argument of the function, marked with the '<tt>sret</tt>' attribute) and
1829 replaces them with a new function that simply returns each of the elements of
1830 that struct (using multiple return values).
1831 </p>
1832
1833 <p>
1834 This pass works under a number of conditions:
1835 </p>
1836
1837 <ul>
1838 <li>The returned struct must not contain other structs</li>
1839 <li>The returned struct must only be used to load values from</li>
1840 <li>The placeholder struct passed in is the result of an <tt>alloca</tt></li>
1841 </ul>
1842</div>
1843
1844<!-------------------------------------------------------------------------- -->
1845<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1846 <a name="tailcallelim">-tailcallelim: Tail Call Elimination</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1847</div>
1848<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1849 <p>
1850 This file transforms calls of the current function (self recursion) followed
1851 by a return instruction with a branch to the entry of the function, creating
1852 a loop. This pass also implements the following extensions to the basic
1853 algorithm:
1854 </p>
1855
1856 <ul>
1857 <li>Trivial instructions between the call and return do not prevent the
1858 transformation from taking place, though currently the analysis cannot
1859 support moving any really useful instructions (only dead ones).
1860 <li>This pass transforms functions that are prevented from being tail
1861 recursive by an associative expression to use an accumulator variable,
1862 thus compiling the typical naive factorial or <tt>fib</tt> implementation
1863 into efficient code.
1864 <li>TRE is performed if the function returns void, if the return
1865 returns the result returned by the call, or if the function returns a
1866 run-time constant on all exits from the function. It is possible, though
1867 unlikely, that the return returns something else (like constant 0), and
1868 can still be TRE'd. It can be TRE'd if <em>all other</em> return
1869 instructions in the function return the exact same value.
1870 <li>If it can prove that callees do not access theier caller stack frame,
1871 they are marked as eligible for tail call elimination (by the code
1872 generator).
1873 </ul>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1874</div>
1875
1876<!-------------------------------------------------------------------------- -->
1877<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1878 <a name="tailduplicate">-tailduplicate: Tail Duplication</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1879</div>
1880<div class="doc_text">
Gordon Henriksenc86b6772007-11-04 16:15:04 +0000[diff] [blame]1881 <p>
1882 This pass performs a limited form of tail duplication, intended to simplify
1883 CFGs by removing some unconditional branches. This pass is necessary to
1884 straighten out loops created by the C front-end, but also is capable of
1885 making other code nicer. After this pass is run, the CFG simplify pass
1886 should be run to clean up the mess.
1887 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1888</div>
1889
1890<!-- ======================================================================= -->
1891<div class="doc_section"> <a name="transform">Utility Passes</a></div>
1892<div class="doc_text">
1893 <p>This section describes the LLVM Utility Passes.</p>
1894</div>
1895
1896<!-------------------------------------------------------------------------- -->
1897<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1898 <a name="deadarghaX0r">-deadarghaX0r: Dead Argument Hacking (BUGPOINT USE ONLY; DO NOT USE)</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1899</div>
1900<div class="doc_text">
Gordon Henriksen75ff18e2007-11-04 18:10:18 +0000[diff] [blame]1901 <p>
1902 Same as dead argument elimination, but deletes arguments to functions which
1903 are external. This is only for use by <a
1904 href="Bugpoint.html">bugpoint</a>.</p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1905</div>
1906
1907<!-------------------------------------------------------------------------- -->
1908<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1909 <a name="extract-blocks">-extract-blocks: Extract Basic Blocks From Module (for bugpoint use)</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1910</div>
1911<div class="doc_text">
Gordon Henriksen75ff18e2007-11-04 18:10:18 +0000[diff] [blame]1912 <p>
1913 This pass is used by bugpoint to extract all blocks from the module into their
1914 own functions.</p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1915</div>
1916
1917<!-------------------------------------------------------------------------- -->
1918<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1919 <a name="instnamer">-instnamer: Assign names to anonymous instructions</a>
1920</div>
1921<div class="doc_text">
1922 <p>This is a little utility pass that gives instructions names, this is mostly
1923 useful when diffing the effect of an optimization because deleting an
1924 unnamed instruction can change all other instruction numbering, making the
1925 diff very noisy.
1926 </p>
1927</div>
1928
1929<!-------------------------------------------------------------------------- -->
1930<div class="doc_subsection">
1931 <a name="preverify">-preverify: Preliminary module verification</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1932</div>
1933<div class="doc_text">
Gordon Henriksen90a52142007-11-05 02:05:35 +0000[diff] [blame]1934 <p>
1935 Ensures that the module is in the form required by the <a
1936 href="#verifier">Module Verifier</a> pass.
1937 </p>
1938
1939 <p>
1940 Running the verifier runs this pass automatically, so there should be no need
1941 to use it directly.
1942 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1943</div>
1944
1945<!-------------------------------------------------------------------------- -->
1946<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1947 <a name="verify">-verify: Module Verifier</a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1948</div>
1949<div class="doc_text">
Gordon Henriksen75ff18e2007-11-04 18:10:18 +0000[diff] [blame]1950 <p>
1951 Verifies an LLVM IR code. This is useful to run after an optimization which is
1952 undergoing testing. Note that <tt>llvm-as</tt> verifies its input before
1953 emitting bitcode, and also that malformed bitcode is likely to make LLVM
1954 crash. All language front-ends are therefore encouraged to verify their output
1955 before performing optimizing transformations.
1956 </p>
1957
Gordon Henriksen23a8ce52007-11-04 18:14:08 +0000[diff] [blame]1958 <ul>
Gordon Henriksen75ff18e2007-11-04 18:10:18 +0000[diff] [blame]1959 <li>Both of a binary operator's parameters are of the same type.</li>
1960 <li>Verify that the indices of mem access instructions match other
1961 operands.</li>
1962 <li>Verify that arithmetic and other things are only performed on
1963 first-class types. Verify that shifts and logicals only happen on
1964 integrals f.e.</li>
1965 <li>All of the constants in a switch statement are of the correct type.</li>
1966 <li>The code is in valid SSA form.</li>
Chris Lattner46b3abc2009-10-28 04:47:06 +0000[diff] [blame]1967 <li>It is illegal to put a label into any other type (like a structure) or
1968 to return one.</li>
Nick Lewycky0c78ac12008-03-28 06:46:51 +0000[diff] [blame]1969 <li>Only phi nodes can be self referential: <tt>%x = add i32 %x, %x</tt> is
Gordon Henriksen873390e2007-11-04 18:17:58 +0000[diff] [blame]1970 invalid.</li>
Gordon Henriksen75ff18e2007-11-04 18:10:18 +0000[diff] [blame]1971 <li>PHI nodes must have an entry for each predecessor, with no extras.</li>
1972 <li>PHI nodes must be the first thing in a basic block, all grouped
1973 together.</li>
1974 <li>PHI nodes must have at least one entry.</li>
1975 <li>All basic blocks should only end with terminator insts, not contain
1976 them.</li>
1977 <li>The entry node to a function must not have predecessors.</li>
1978 <li>All Instructions must be embedded into a basic block.</li>
1979 <li>Functions cannot take a void-typed parameter.</li>
1980 <li>Verify that a function's argument list agrees with its declared
1981 type.</li>
1982 <li>It is illegal to specify a name for a void value.</li>
1983 <li>It is illegal to have a internal global value with no initializer.</li>
1984 <li>It is illegal to have a ret instruction that returns a value that does
1985 not agree with the function return value type.</li>
1986 <li>Function call argument types match the function prototype.</li>
1987 <li>All other things that are tested by asserts spread about the code.</li>
Gordon Henriksen23a8ce52007-11-04 18:14:08 +0000[diff] [blame]1988 </ul>
Gordon Henriksen75ff18e2007-11-04 18:10:18 +0000[diff] [blame]1989
1990 <p>
1991 Note that this does not provide full security verification (like Java), but
1992 instead just tries to ensure that code is well-formed.
1993 </p>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]1994</div>
1995
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]1996<!-------------------------------------------------------------------------- -->
1997<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]1998 <a name="view-cfg">-view-cfg: View CFG of function</a>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]1999</div>
2000<div class="doc_text">
Gordon Henriksen75ff18e2007-11-04 18:10:18 +0000[diff] [blame]2001 <p>
2002 Displays the control flow graph using the GraphViz tool.
2003 </p>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]2004</div>
2005
2006<!-------------------------------------------------------------------------- -->
2007<div class="doc_subsection">
Duncan Sands5c603862010-07-06 15:52:15 +0000[diff] [blame]2008 <a name="view-cfg-only">-view-cfg-only: View CFG of function (with no function bodies)</a>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]2009</div>
2010<div class="doc_text">
Gordon Henriksen75ff18e2007-11-04 18:10:18 +0000[diff] [blame]2011 <p>
2012 Displays the control flow graph using the GraphViz tool, but omitting function
2013 bodies.
2014 </p>
Gordon Henriksen1f5cce02007-10-25 08:46:12 +0000[diff] [blame]2015</div>
2016
Tobias Grosser733783b2010-05-07 09:33:18 +0000[diff] [blame]2017<!-------------------------------------------------------------------------- -->
2018<div class="doc_subsection">
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]2019 <a name="view-dom">-view-dom: View dominance tree of function</a>
Tobias Grosser733783b2010-05-07 09:33:18 +0000[diff] [blame]2020</div>
2021<div class="doc_text">
2022 <p>
2023 Displays the dominator tree using the GraphViz tool.
2024 </p>
2025</div>
2026
2027<!-------------------------------------------------------------------------- -->
2028<div class="doc_subsection">
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]2029 <a name="view-dom-only">-view-dom-only: View dominance tree of function (with no function
Tobias Grosser733783b2010-05-07 09:33:18 +0000[diff] [blame]2030 bodies)
2031 </a>
2032</div>
2033<div class="doc_text">
2034 <p>
2035 Displays the dominator tree using the GraphViz tool, but omitting function
2036 bodies.
2037 </p>
2038</div>
2039
2040<!-------------------------------------------------------------------------- -->
2041<div class="doc_subsection">
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]2042 <a name="view-postdom">-view-postdom: View postdominance tree of function</a>
Tobias Grosser733783b2010-05-07 09:33:18 +0000[diff] [blame]2043</div>
2044<div class="doc_text">
2045 <p>
2046 Displays the post dominator tree using the GraphViz tool.
2047 </p>
2048</div>
2049
2050<!-------------------------------------------------------------------------- -->
2051<div class="doc_subsection">
Eli Friedmane6ed15b2011-03-19 04:47:52 +0000[diff] [blame]2052 <a name="view-postdom-only">-view-postdom-only: View postdominance tree of function (with no
Tobias Grosser733783b2010-05-07 09:33:18 +0000[diff] [blame]2053 function bodies)
2054 </a>
2055</div>
2056<div class="doc_text">
2057 <p>
2058 Displays the post dominator tree using the GraphViz tool, but omitting
2059 function bodies.
2060 </p>
2061</div>
2062
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]2063<!-- *********************************************************************** -->
2064
2065<hr>
2066<address>
2067 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]2068 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]2069 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]2070 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Reid Spencerd9aac122007-03-26 09:32:31 +0000[diff] [blame]2071
2072 <a href="mailto:rspencer@x10sys.com">Reid Spencer</a><br>
2073 <a href="http://llvm.org">LLVM Compiler Infrastructure</a><br>
2074 Last modified: $Date$
2075</address>
2076
2077</body>
2078</html>