Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 44aa9214bfe5cc0222f04a0718d40dbab133c5b4 / . / docs / ReleaseNotes.html
blob: 93a491f98943a0cd4bcdb1810c0e25c633dd7ac0 [file] [log] [blame]
Misha Brukman500bc302003-11-22 00:38:41 +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>
Reid Spencer6454ed32004-11-18 18:38:58 +0000[diff] [blame]5 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]6 <link rel="stylesheet" href="llvm.css" type="text/css">
Chris Lattner857fb132008-10-11 18:18:28 +0000[diff] [blame]7 <title>LLVM 2.4 Release Notes</title>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]8</head>
9<body>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]10
Chris Lattner857fb132008-10-11 18:18:28 +0000[diff] [blame]11<div class="doc_title">LLVM 2.4 Release Notes</div>
Mikhail Glushenkovea65d7d2008-10-13 02:08:34 +0000[diff] [blame]12
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]13<ol>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]14 <li><a href="#intro">Introduction</a></li>
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]15 <li><a href="#subproj">Sub-project Status Update</a></li>
16 <li><a href="#whatsnew">What's New in LLVM?</a></li>
Chris Lattner4b538b92004-04-30 22:17:12 +0000[diff] [blame]17 <li><a href="GettingStarted.html">Installation Instructions</a></li>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]18 <li><a href="#portability">Portability and Supported Platforms</a></li>
Dan Gohman44aa9212008-10-14 16:23:02 +0000[diff] [blame^]19 <li><a href="#knownproblems">Known Problems</a></li>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]20 <li><a href="#additionalinfo">Additional Information</a></li>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]21</ol>
22
Chris Lattner7911ce22004-05-23 21:07:27 +0000[diff] [blame]23<div class="doc_author">
Dan Gohman44aa9212008-10-14 16:23:02 +0000[diff] [blame^]24 <p>Written by the <a href="http://llvm.org">LLVM Team</a></p>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]25</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]26
27<!-- *********************************************************************** -->
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]28<div class="doc_section">
29 <a name="intro">Introduction</a>
30</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]31<!-- *********************************************************************** -->
32
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]33<div class="doc_text">
34
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]35<p>This document contains the release notes for the LLVM Compiler
36Infrastructure, release 2.4. Here we describe the status of LLVM, including
37major improvements from the previous release and significant known problems.
38All LLVM releases may be downloaded from the <a
39href="http://llvm.org/releases/">LLVM releases web site</a>.</p>
Chris Lattner19092612003-10-02 16:38:05 +0000[diff] [blame]40
Chris Lattner7506b1d2004-12-07 08:04:13 +0000[diff] [blame]41<p>For more information about LLVM, including information about the latest
Chris Lattnerc463b272005-10-29 07:07:09 +0000[diff] [blame]42release, please check out the <a href="http://llvm.org/">main LLVM
Chris Lattner47ad72c2003-10-07 21:38:31 +0000[diff] [blame]43web site</a>. If you have questions or comments, the <a
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]44href="http://mail.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVM Developer's Mailing
45List</a> is a good place to send them.</p>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]46
Mikhail Glushenkovea65d7d2008-10-13 02:08:34 +0000[diff] [blame]47<p>Note that if you are reading this file from a Subversion checkout or the
Chris Lattner84977642007-09-21 03:54:09 +0000[diff] [blame]48main LLVM web page, this document applies to the <i>next</i> release, not the
Gabor Greiffa933f82008-10-14 11:00:32 +0000[diff] [blame]49current one. To see the release notes for a specific release, please see the
Chris Lattner84977642007-09-21 03:54:09 +0000[diff] [blame]50<a href="http://llvm.org/releases/">releases page</a>.</p>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]51
52</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]53
Chris Lattner857fb132008-10-11 18:18:28 +0000[diff] [blame]54<!-- Unfinished features in 2.4:
Chris Lattneracce85d2008-02-10 07:46:44 +0000[diff] [blame]55 Machine LICM
56 Machine Sinking
57 LegalizeDAGTypes
Chris Lattner547a3912008-10-12 19:47:48 +0000[diff] [blame]58 llc -enable-value-prop, propagation of value info (sign/zero ext info) from
59 one MBB to another
Chris Lattneracce85d2008-02-10 07:46:44 +0000[diff] [blame]60 -->
Mikhail Glushenkovea65d7d2008-10-13 02:08:34 +0000[diff] [blame]61
Chris Lattner547a3912008-10-12 19:47:48 +0000[diff] [blame]62 <!-- for announcement email:
63 mention dev mtg
Chris Lattnercdb0e4c2008-10-12 22:57:58 +0000[diff] [blame]64 Xcode 3.1 and 3.1.1.
Chris Lattner547a3912008-10-12 19:47:48 +0000[diff] [blame]65 -->
Chris Lattneracce85d2008-02-10 07:46:44 +0000[diff] [blame]66
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]67<!-- *********************************************************************** -->
68<div class="doc_section">
69 <a name="subproj">Sub-project Status Update</a>
Chris Lattnerea34f642008-06-08 21:34:41 +0000[diff] [blame]70</div>
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]71<!-- *********************************************************************** -->
Chris Lattnerea34f642008-06-08 21:34:41 +0000[diff] [blame]72
73<div class="doc_text">
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]74<p>
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]75The LLVM 2.4 distribution currently consists of code from the core LLVM
Chris Lattnerfb97b2d2008-10-13 18:11:54 +0000[diff] [blame]76repository (which roughly includes the LLVM optimizers, code generators and
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]77supporting tools) and the llvm-gcc repository. In addition to this code, the
78LLVM Project includes other sub-projects that are in development. The two which
Chris Lattner857fb132008-10-11 18:18:28 +0000[diff] [blame]79are the most actively developed are the <a href="#clang">Clang Project</a> and
Duncan Sandsbf8c4f22008-10-13 18:33:28 +0000[diff] [blame]80the <a href="#vmkit">vmkit Project</a>.
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]81</p>
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]82
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]83</div>
84
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]85
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]86<!--=========================================================================-->
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]87<div class="doc_subsection">
Chris Lattnerfb97b2d2008-10-13 18:11:54 +0000[diff] [blame]88<a name="clang">Clang: C/C++/Objective-C Frontend Toolkit</a>
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]89</div>
90
91<div class="doc_text">
92
93<p>The <a href="http://clang.llvm.org/">Clang project</a> is an effort to build
94a set of new 'LLVM native' front-end technologies for the LLVM optimizer
95and code generator. Clang is continuing to make major strides forward in all
96areas. Its C and Objective-C parsing support is very solid, and the code
97generation support is far enough along to build many C applications. While not
98yet production quality, it is progressing very nicely. In addition, C++
99front-end work has started to make significant progress.</p>
100
Chris Lattnercc042612008-10-14 00:52:49 +0000[diff] [blame]101<p>Codegen progress/state [DANIEL]</p>
Chris Lattner857fb132008-10-11 18:18:28 +0000[diff] [blame]102
Chris Lattnerfb97b2d2008-10-13 18:11:54 +0000[diff] [blame]103</div>
104
105<!--=========================================================================-->
106<div class="doc_subsection">
107<a name="clangsa">Clang Static Analyzer</a>
108</div>
109
110<div class="doc_text">
111
Ted Kremenek3c3ec0c2008-10-14 05:14:21 +0000[diff] [blame]112<p>The Clang project also includes an early stage static source code analysis
113tool for <a href="http://clang.llvm.org/StaticAnalysis.html">automatically
114finding bugs</a> in C and Objective-C programs. The tool performs a growing set
Chris Lattnerc702d162008-10-14 06:02:29 +0000[diff] [blame]115of checks to find bugs that occur on a specific path within a program. Examples
Ted Kremenek3c3ec0c2008-10-14 05:14:21 +0000[diff] [blame]116of bugs the tool finds include logic errors such as null dereferences,
117violations of various API rules, dead code, and potential memory leaks in
Chris Lattnerc702d162008-10-14 06:02:29 +0000[diff] [blame]118Objective-C programs. Since its inception, public feedback on the tool has been
119extremely positive, and conservative estimates put the number of real bugs it
120has found in industrial-quality software on the order of thousands.</p>
Chris Lattnercc042612008-10-14 00:52:49 +0000[diff] [blame]121
Ted Kremenek3c3ec0c2008-10-14 05:14:21 +0000[diff] [blame]122<p>The tool also provides a simple web GUI to inspect potential bugs found by
Chris Lattnerc702d162008-10-14 06:02:29 +0000[diff] [blame]123the tool. While still early in development, the GUI illustrates some of the key
Ted Kremenek3c3ec0c2008-10-14 05:14:21 +0000[diff] [blame]124features of Clang: accurate source location information, which is used by the
125GUI to highlight specific code expressions that relate to a bug (including those
126that span multiple lines) and built-in knowledge of macros, which is used to
127perform inline expansion of macros within the GUI itself.</p>
128
Chris Lattnerc702d162008-10-14 06:02:29 +0000[diff] [blame]129<p>The set of checks performed by the static analyzer is gradually expanding,
130and
Ted Kremenek3c3ec0c2008-10-14 05:14:21 +0000[diff] [blame]131future plans for the tool include full source-level inter-procedural analysis
132and deeper checks such as buffer overrun detection. There are many opportunities
133to extend and enhance the static analyzer, and anyone interested in working on
134this project is encouraged to get involved!</p>
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]135
136</div>
137
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]138<!--=========================================================================-->
139<div class="doc_subsection">
Chris Lattnerfb97b2d2008-10-13 18:11:54 +0000[diff] [blame]140<a name="vmkit">vmkit: JVM/CLI Virtual Machine Implementation</a>
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]141</div>
142
143<div class="doc_text">
144<p>
Chris Lattnerfb97b2d2008-10-13 18:11:54 +0000[diff] [blame]145The <a href="http://vmkit.llvm.org/">vmkit project</a> is an implementation of
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]146a JVM and a CLI Virtual Machines (Microsoft .NET is an
147implementation of the CLI) using the Just-In-Time compiler of LLVM.</p>
148
Chris Lattnercc042612008-10-14 00:52:49 +0000[diff] [blame]149<p>[NICOLAS]</p>
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]150
151</div>
152
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]153
154<!-- *********************************************************************** -->
155<div class="doc_section">
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]156 <a name="whatsnew">What's New in LLVM?</a>
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]157</div>
158<!-- *********************************************************************** -->
159
Chris Lattnerf8e0b4e2008-06-08 22:59:35 +0000[diff] [blame]160<div class="doc_text">
161
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]162<p>This release includes a huge number of bug fixes, performance tweaks and
163minor improvements. Some of the major improvements and new features are listed
164in this section.
Chris Lattnerf8e0b4e2008-06-08 22:59:35 +0000[diff] [blame]165</p>
166</div>
167
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]168<!--=========================================================================-->
Chris Lattnerea34f642008-06-08 21:34:41 +0000[diff] [blame]169<div class="doc_subsection">
Chris Lattner252b83d2008-02-06 18:00:06 +0000[diff] [blame]170<a name="majorfeatures">Major New Features</a>
171</div>
172
173<div class="doc_text">
174
Chris Lattner857fb132008-10-11 18:18:28 +0000[diff] [blame]175<p>LLVM 2.4 includes several major new capabilities:</p>
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]176
177<ul>
Chris Lattnerb82eb062008-10-13 20:47:20 +0000[diff] [blame]178<li><p>The most visible end-user change in LLVM 2.4 is that it includes many
Chris Lattnerf3013872008-10-13 21:50:36 +0000[diff] [blame]179optimizations and changes to make -O0 compile times much faster. You should see
Gabor Greiffa933f82008-10-14 11:00:32 +0000[diff] [blame]180improvements on the order of 30% (or more) faster than LLVM 2.3. There are many
Chris Lattnerb82eb062008-10-13 20:47:20 +0000[diff] [blame]181pieces to this change, described in more detail below. The speedups and new
182components can also be used for JIT compilers that want fast compilation as
183well.</p></li>
Chris Lattner857fb132008-10-11 18:18:28 +0000[diff] [blame]184
Chris Lattnerb82eb062008-10-13 20:47:20 +0000[diff] [blame]185<li><p>The biggest change to the LLVM IR is that Multiple Return Values (which
186were introduced in LLVM 2.3) have been generalized to full support for "First
187Class Aggregate" values in LLVM 2.4. This means that LLVM IR supports using
188structs and arrays as values in a function. This capability is mostly useful
189for front-end authors, who prefer to treat things like complex numbers, simple
190tuples, dope vectors, etc as Value*'s instead of as a tuple of Value*'s or as
Gabor Greiffa933f82008-10-14 11:00:32 +0000[diff] [blame]191memory values. Bitcode files from LLVM 2.3 will automatically migrate to the
192general representation.</p></li>
Mikhail Glushenkovea65d7d2008-10-13 02:08:34 +0000[diff] [blame]193
Chris Lattnerb82eb062008-10-13 20:47:20 +0000[diff] [blame]194<li><p>LLVM 2.4 also includes an initial port for the PIC16 microprocessor. This
Duncan Sands9ba23792008-10-14 07:06:37 +0000[diff] [blame]195is the LLVM target that only has support for 8 bit registers, and a number of
Chris Lattnerb82eb062008-10-13 20:47:20 +0000[diff] [blame]196other crazy constraints. While the port is still in early development stages,
197it shows some interesting things you can do with LLVM.</p></li>
Chris Lattnerbc5786b2008-06-05 06:57:39 +0000[diff] [blame]198
Chris Lattner8170c102008-02-10 08:18:42 +0000[diff] [blame]199</ul>
200
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]201</div>
202
Chris Lattner791f77b2008-06-05 06:25:56 +0000[diff] [blame]203
204<!--=========================================================================-->
205<div class="doc_subsection">
Chris Lattnerea34f642008-06-08 21:34:41 +0000[diff] [blame]206<a name="llvm-gcc">llvm-gcc 4.2 Improvements</a>
Chris Lattner791f77b2008-06-05 06:25:56 +0000[diff] [blame]207</div>
208
209<div class="doc_text">
210
Chris Lattnerb82eb062008-10-13 20:47:20 +0000[diff] [blame]211<p>LLVM fully supports the llvm-gcc 4.2 front-end, which marries the GCC
212front-ends and driver with the LLVM optimizer and code generator. It currently
213includes support for the C, C++, Objective-C, Ada, and Fortran front-ends.</p>
Chris Lattner791f77b2008-06-05 06:25:56 +0000[diff] [blame]214
Chris Lattnerf8e0b4e2008-06-08 22:59:35 +0000[diff] [blame]215<ul>
Chris Lattner7752d1a2008-10-13 20:59:32 +0000[diff] [blame]216<li>LLVM 2.4 supports the full set of atomic <tt>__sync_*</tt> builtins. LLVM
2172.3 only supported those used by OpenMP, but 2.4 supports them all. While
Chris Lattnerf3013872008-10-13 21:50:36 +0000[diff] [blame]218llvm-gcc supports all of these builtins, note that not all targets do. X86
219support them all in both 32-bit and 64-bit mode and PowerPC supports them all
220except for the 64-bit operations when in 32-bit mode.</li>
Chris Lattnerf8e0b4e2008-06-08 22:59:35 +0000[diff] [blame]221
Chris Lattner7752d1a2008-10-13 20:59:32 +0000[diff] [blame]222<li>llvm-gcc now supports an <tt>-flimited-precision</tt> option, which tells
223the compiler that it is ok to use low-precision approximations of certain libm
Chris Lattnerb82eb062008-10-13 20:47:20 +0000[diff] [blame]224functions (like tan, log, etc). This allows you to get high performance if you
225only need (say) 14-bits of precision.</li>
226
227<li>llvm-gcc now supports a C language extension known as "<a
228href="http://lists.cs.uiuc.edu/pipermail/cfe-dev/2008-August/002670.html">Blocks
Gabor Greiffa933f82008-10-14 11:00:32 +0000[diff] [blame]229</a>". This feature is similar to nested functions and closures, but does not
Chris Lattnerb82eb062008-10-13 20:47:20 +0000[diff] [blame]230require stack trampolines (with most ABIs) and supports returning closures
231from functions that define them. Note that actually <em>using</em> Blocks
232requires a small runtime that is not included with llvm-gcc.</li>
233
234<li>llvm-gcc now supports a new <tt>-flto</tt> option. On systems that support
235transparent Link Time Optimization (currently Darwin systems with Xcode 3.1 and
236later) this allows the use of LTO with other optimization levels like -Os.
237Previously, LTO could only be used with -O4, which implied optimizations in
238-O3 that can increase code size.</li>
Gabor Greifa3bd11d2008-06-09 06:06:18 +0000[diff] [blame]239</ul>
Chris Lattner791f77b2008-06-05 06:25:56 +0000[diff] [blame]240
Chris Lattner791f77b2008-06-05 06:25:56 +0000[diff] [blame]241</div>
242
243
Chris Lattner84977642007-09-21 03:54:09 +0000[diff] [blame]244<!--=========================================================================-->
245<div class="doc_subsection">
Chris Lattnerf304ffc2008-02-10 08:17:19 +0000[diff] [blame]246<a name="coreimprovements">LLVM Core Improvements</a>
247</div>
248
249<div class="doc_text">
250<p>New features include:
251</p>
252
Chris Lattner791f77b2008-06-05 06:25:56 +0000[diff] [blame]253<ul>
Gabor Greiffa933f82008-10-14 11:00:32 +0000[diff] [blame]254<li>A major change to the <tt>Use</tt> class landed, which shrank it by 25%. Since
Chris Lattner7752d1a2008-10-13 20:59:32 +0000[diff] [blame]255this is a pervasive part of the LLVM, it ended up reducing the memory use of
256LLVM IR in general by 15% for most programs.</li>
257
258<li>Values with no names are now pretty printed by <tt>llvm-dis</tt> more
259nicely. They now print as "<tt>%3 = add i32 %A, 4</tt>" instead of
260"<tt>add i32 %A, 4 ; &lt;i32&gt;:3</tt>", which makes it much easier to read.
261</li>
262
263<li>LLVM 2.4 includes some changes for better vector support. First, the shift
264operations (<tt>shl</tt>, <tt>ashr</tt>, <tt>lshr</tt>) now all support vectors
265and do an element-by-element shift (shifts of the whole vector can be
266accomplished by bitcasting the vector to &lt;1 x i128&gt; for example). Second,
Chris Lattnercc042612008-10-14 00:52:49 +0000[diff] [blame]267there is initial support in development for vector comparisons with the
268<a href="LangRef.html#i_fcmp">fcmp</a>/<a href="LangRef.html#i_icmp">icmp</a>
269instructions. These instructions compare two vectors and return a vector of
270i1's for each result. Note that there is very little codegen support available
271for any of these IR features though.</li>
Chris Lattner7752d1a2008-10-13 20:59:32 +0000[diff] [blame]272
273<li>A new <tt>DebugInfoBuilder</tt> class is available, which makes it much
274easier for front-ends to create debug info descriptors, similar to the way that
Gabor Greifa8b84902008-10-14 11:31:14 +0000[diff] [blame]275<tt>IRBuilder</tt> makes it easier to create LLVM IR.</li>
Chris Lattner7752d1a2008-10-13 20:59:32 +0000[diff] [blame]276
Dan Gohman1bbddce2008-10-14 16:13:59 +0000[diff] [blame]277<li>The <tt>IRBuilder</tt> class is now parameterized by a class responsible
Gabor Greiffa933f82008-10-14 11:00:32 +0000[diff] [blame]278for constant folding. The default <tt>ConstantFolder</tt> class does target independent
279constant folding. The <tt>NoFolder</tt> class does no constant folding at all, which is
280useful when learning how LLVM works. The <tt>TargetFolder</tt> class folds the most,
Duncan Sands9ba23792008-10-14 07:06:37 +0000[diff] [blame]281doing target dependent constant folding.</li>
282
283<li>LLVM now supports "function attributes", which allows us to separate return
Chris Lattner7752d1a2008-10-13 20:59:32 +0000[diff] [blame]284value attributes from function attributes. LLVM now supports attributes on a
285function itself, a return value, and its parameters. New supported function
286attributes include noinline/alwaysinline and the "opt-size" flag which says the
287function should be optimized for code size.</li>
288
Chris Lattnercce6d4b2008-10-12 18:30:33 +0000[diff] [blame]289<li>LLVM IR now directly represents "common" linkage, instead of
290 representing it as a form of weak linkage.</li>
Chris Lattner7752d1a2008-10-13 20:59:32 +0000[diff] [blame]291
Chris Lattnerf304ffc2008-02-10 08:17:19 +0000[diff] [blame]292</ul>
Mikhail Glushenkovea65d7d2008-10-13 02:08:34 +0000[diff] [blame]293
Chris Lattnerf304ffc2008-02-10 08:17:19 +0000[diff] [blame]294</div>
295
296<!--=========================================================================-->
297<div class="doc_subsection">
Chris Lattneracce85d2008-02-10 07:46:44 +0000[diff] [blame]298<a name="optimizer">Optimizer Improvements</a>
299</div>
300
301<div class="doc_text">
302
Chris Lattnerf3013872008-10-13 21:50:36 +0000[diff] [blame]303<p>In addition to a huge array of bug fixes and minor performance tweaks, this
304release includes a few major enhancements and additions to the optimizers:</p>
Chris Lattneracce85d2008-02-10 07:46:44 +0000[diff] [blame]305
306<ul>
307
Chris Lattnerf3013872008-10-13 21:50:36 +0000[diff] [blame]308<li>The Global Value Numbering (GVN) pass now does local Partial Redundancy
309Elimination (PRE) to eliminate some partially redundant expressions in cases
310where doing so won't grow code size.</li>
Chris Lattner547a3912008-10-12 19:47:48 +0000[diff] [blame]311
Chris Lattnerf3013872008-10-13 21:50:36 +0000[diff] [blame]312<li>LLVM 2.4 includes a new loop deletion pass (which removes output-free
313provably-finite loops) and a rewritten Aggressive Dead Code Elimination (ADCE)
314pass that no longer uses control dependence information. These changes speed up
Gabor Greifa8b84902008-10-14 11:31:14 +0000[diff] [blame]315the optimizer and also prevent it from deleting output-free infinite
Chris Lattnerf3013872008-10-13 21:50:36 +0000[diff] [blame]316loops.</li>
Chris Lattner547a3912008-10-12 19:47:48 +0000[diff] [blame]317
Duncan Sands1050d732008-10-13 18:42:43 +0000[diff] [blame]318<li>The new AddReadAttrs pass works out which functions are read-only or
Gabor Greifa8b84902008-10-14 11:31:14 +0000[diff] [blame]319read-none (these correspond to 'pure' and 'const' in GCC) and marks them
Duncan Sands1050d732008-10-13 18:42:43 +0000[diff] [blame]320with the appropriate attribute.</li>
Chris Lattnerf3013872008-10-13 21:50:36 +0000[diff] [blame]321
322<li>LLVM 2.4 now includes a new SparsePropagation framework, which makes it
323trivial to build lattice-based dataflow solvers that operate over LLVM IR. Using
324this interface means that you just define objects to represent your lattice
325values and the transfer functions that operate on them. It handles the
326mechanics of worklist processing, liveness tracking, handling PHI nodes,
327etc.</li>
328
Chris Lattnerc702d162008-10-14 06:02:29 +0000[diff] [blame]329<li>The Loop Strength Reduction and induction variable optimization passes have
330several improvements to avoid inserting MAX expressions, to optimize simple
331floating point induction variables and to analyze trip counts of more
332loops.</li>
333
Chris Lattnerf3013872008-10-13 21:50:36 +0000[diff] [blame]334<li>Various helper functions (ComputeMaskedBits, ComputeNumSignBits, etc) were
335pulled out of the Instruction Combining pass and put into a new
336<tt>ValueTracking.h</tt> header, where they can be reused by other passes.</li>
337
338<li>The tail duplication pass has been removed from the standard optimizer
339sequence used by llvm-gcc. This pass still exists, but the benefits it once
340provided are now achieved by other passes.</li>
341
Chris Lattner0b832202008-06-08 02:45:07 +0000[diff] [blame]342</ul>
343
344</div>
345
346<!--=========================================================================-->
347<div class="doc_subsection">
348<a name="codegen">Code Generator Improvements</a>
349</div>
350
351<div class="doc_text">
352
Gabor Greifa8b84902008-10-14 11:31:14 +0000[diff] [blame]353<p>We have put a significant amount of work into the code generator infrastructure,
Chris Lattner0b832202008-06-08 02:45:07 +0000[diff] [blame]354which allows us to implement more aggressive algorithms and make it run
355faster:</p>
356
357<ul>
Chris Lattnerf3013872008-10-13 21:50:36 +0000[diff] [blame]358<li>The target-independent code generator supports (and the X86 backend
359 currently implements) a new interface for "fast" instruction selection. This
360 interface is optimized to produce code as quickly as possible, sacrificing
361 code quality to do it. This is used by default at -O0 or when using
362 "llc -fast" on X86. It is straight-forward to add support for
363 other targets if faster -O0 compilation is desired.</li>
364
365<li>In addition to the new 'fast' instruction selection path, many existing
366 pieces of the code generator have been optimized in significant ways.
367 SelectionDAG's are now pool allocated and use better algorithms in many
Gabor Greifa8b84902008-10-14 11:31:14 +0000[diff] [blame]368 places, the ".s" file printers now use <tt>raw_ostream</tt> to emit text much faster,
Chris Lattnerf3013872008-10-13 21:50:36 +0000[diff] [blame]369 etc. The end result of these improvements is that the compiler also takes
370 substantially less time to generate code that is just as good (and often
371 better) than before.</li>
372
Gabor Greifa8b84902008-10-14 11:31:14 +0000[diff] [blame]373<li>Each target has been split to separate the ".s" file printing logic from the
Chris Lattnerf3013872008-10-13 21:50:36 +0000[diff] [blame]374 rest of the target. This enables JIT compilers that don't link in the
Gabor Greifa8b84902008-10-14 11:31:14 +0000[diff] [blame]375 (somewhat large) code and data tables used for printing a ".s" file.</li>
Chris Lattnerf3013872008-10-13 21:50:36 +0000[diff] [blame]376
377<li>The code generator now includes a "stack slot coloring" pass, which packs
378 together individual spilled values into common stack slots. This reduces
379 the size of stack frames with many spills, which tends to increase L1 cache
380 effectiveness.</li>
381
382<li>Various pieces of the register allocator (e.g. the coalescer and two-address
383 operation elimination pass) now know how to rematerialize trivial operations
384 to avoid copies and include several other optimizations.</li>
385
386<li>The <a href="CodeGenerator.html#selectiondag_process">graphs</a> produced by
387 the <tt>llc -view-*-dags</tt> options are now significantly prettier and
388 easier to read.</li>
389
390<li>LLVM 2.4 includes a new register allocator based on Partitioned Boolean
391 Quadratic Programming (PBQP). This register allocator is still in
392 development, but is very simple and clean.</li>
Chris Lattner0b832202008-06-08 02:45:07 +0000[diff] [blame]393
Chris Lattner84977642007-09-21 03:54:09 +0000[diff] [blame]394</ul>
395
396</div>
397
398
399<!--=========================================================================-->
400<div class="doc_subsection">
Chris Lattnereeb4da02008-10-13 22:06:31 +0000[diff] [blame]401<a name="targetspecific">Target Specific Improvements</a>
Chris Lattner84977642007-09-21 03:54:09 +0000[diff] [blame]402</div>
Chris Lattner60893e52007-05-18 06:33:02 +0000[diff] [blame]403
Chris Lattner738bd302006-04-18 06:32:08 +0000[diff] [blame]404<div class="doc_text">
Chris Lattnerf304ffc2008-02-10 08:17:19 +0000[diff] [blame]405<p>New target-specific features include:
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]406</p>
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]407
408<ul>
Chris Lattner547a3912008-10-12 19:47:48 +0000[diff] [blame]409<li>Exception handling is supported by default on Linux/x86-64.</li>
Duncan Sandsbbf379a2008-10-13 17:32:15 +0000[diff] [blame]410<li>Position Independent Code (PIC) is now supported on Linux/x86-64.</li>
Chris Lattnerc702d162008-10-14 06:02:29 +0000[diff] [blame]411<li>@llvm.frameaddress now supports getting the frame address of stack frames
412 &gt; 0 on x86/x86-64.</li>
Chris Lattnereeb4da02008-10-13 22:06:31 +0000[diff] [blame]413<li>MIPS floating point support? [BRUNO]</li>
414<li>The PowerPC backend now supports trampolines.</li>
Chris Lattner77d29b12008-06-05 08:02:49 +0000[diff] [blame]415</ul>
Chris Lattner77d29b12008-06-05 08:02:49 +0000[diff] [blame]416
Mikhail Glushenkovea65d7d2008-10-13 02:08:34 +0000[diff] [blame]417</div>
Chris Lattner77d29b12008-06-05 08:02:49 +0000[diff] [blame]418
419
420<!--=========================================================================-->
421<div class="doc_subsection">
Chris Lattner84977642007-09-21 03:54:09 +0000[diff] [blame]422<a name="otherimprovements">Other Improvements</a>
423</div>
424
425<div class="doc_text">
426<p>New features include:
427</p>
428
429<ul>
Mikhail Glushenkov49feb452008-10-13 02:46:01 +0000[diff] [blame]430<li><tt>llvmc2</tt> (the generic compiler driver) gained plugin
Chris Lattnereeb4da02008-10-13 22:06:31 +0000[diff] [blame]431 support. It is now easier to experiment with <tt>llvmc2</tt> and
432 build your own tools based on it.</li>
433
434<li>LLVM 2.4 includes a number of new generic algorithms and data structures,
Dan Gohmanbd9f9af2008-10-14 15:14:55 +0000[diff] [blame]435 include a scoped hash table, 'immutable' data structures, a simple
436 free-list manager, and a <tt>raw_ostream</tt> class.
437 The <tt>raw_ostream</tt> class and
Chris Lattnereeb4da02008-10-13 22:06:31 +0000[diff] [blame]438 <tt>format</tt> allow for efficient file output, and various pieces of LLVM
439 have switched over to use it. The eventual goal is to eliminate
440 std::ostream in favor of it.</li>
441
Chris Lattner84977642007-09-21 03:54:09 +0000[diff] [blame]442</ul>
Mikhail Glushenkovea65d7d2008-10-13 02:08:34 +0000[diff] [blame]443
Chris Lattner84977642007-09-21 03:54:09 +0000[diff] [blame]444</div>
Chris Lattnerc5d658a2006-03-03 00:34:26 +0000[diff] [blame]445
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]446<!--=========================================================================-->
447<div class="doc_subsection">
448<a name="changes">Major Changes and Removed Features</a>
449</div>
450
451<div class="doc_text">
452
Chris Lattnereeb4da02008-10-13 22:06:31 +0000[diff] [blame]453<p>If you're already an LLVM user or developer with out-of-tree changes based
454on LLVM 2.3, this section lists some "gotchas" that you may run into upgrading
455from the previous release.</p>
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]456
457<ul>
458
Chris Lattnereeb4da02008-10-13 22:06:31 +0000[diff] [blame]459<li>The LLVM IR generated by llvm-gcc no longer names all instructions. This
460 makes it run faster, but may be more confusing to some people. If you
461 prefer to have names, the '<tt>opt -instnamer</tt>' pass will add names to
462 all instructions.</li>
463
464<li>The LoadVN and GCSE passes have been removed from the tree. They are
465 obsolete and have been replaced with the GVN and MemoryDependence passes.
466 </li>
467</ul>
468
469
470<p>In addition, many APIs have changed in this release. Some of the major LLVM
471API changes are:</p>
472
473<ul>
474
475<li>Attributes changes [DEVANG] </li>
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]476
477<li>The <tt>DbgStopPointInst</tt> methods <tt>getDirectory</tt> and
478<tt>getFileName</tt> now return <tt>Value*</tt> instead of strings. These can be
479converted to strings using <tt>llvm::GetConstantStringInfo</tt> defined via
Gabor Greifa8b84902008-10-14 11:31:14 +0000[diff] [blame]480"<tt>llvm/Analysis/ValueTracking.h</tt>".</li>
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]481
Chris Lattnereeb4da02008-10-13 22:06:31 +0000[diff] [blame]482<li>The APIs to create various instructions have changed from lower case
483 "create" methods to upper case "Create" methods (e.g.
484 <tt>BinaryOperator::create</tt>). LLVM 2.4 includes both cases, but the
485 lower case ones are removed in mainline, please migrate.</li>
486
Gabor Greifa8b84902008-10-14 11:31:14 +0000[diff] [blame]487<li>Various header files like "<tt>llvm/ADT/iterator</tt>" were given a ".h" suffix.
488 Change your code to #include "<tt>llvm/ADT/iterator.h</tt>" instead.</li>
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]489
Gabor Greifa8b84902008-10-14 11:31:14 +0000[diff] [blame]490<li>In the code generator, many <tt>MachineOperand</tt> predicates were renamed to be
Chris Lattnerf1796b62008-10-14 06:37:11 +0000[diff] [blame]491 shorter (e.g. <tt>isFrameIndex()</tt> -&gt; <tt>isFI()</tt>),
492 <tt>SDOperand</tt> was renamed to <tt>SDValue</tt> (and the "<tt>Val</tt>"
493 member was changed to be the <tt>getNode()</tt> accessor), and the
494 <tt>MVT::ValueType</tt> enum has been replaced with an "<tt>MVT</tt>"
495 struct.</li>
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]496</ul>
497
498</div>
499
500
501
Chris Lattner19092612003-10-02 16:38:05 +0000[diff] [blame]502<!-- *********************************************************************** -->
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]503<div class="doc_section">
504 <a name="portability">Portability and Supported Platforms</a>
505</div>
Chris Lattner19092612003-10-02 16:38:05 +0000[diff] [blame]506<!-- *********************************************************************** -->
507
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]508<div class="doc_text">
509
John Criswell0b5b5e92004-12-08 20:35:47 +0000[diff] [blame]510<p>LLVM is known to work on the following platforms:</p>
Chris Lattner4654bdb2004-06-01 18:22:41 +0000[diff] [blame]511
512<ul>
Mikhail Glushenkovea65d7d2008-10-13 02:08:34 +0000[diff] [blame]513<li>Intel and AMD machines (IA32) running Red Hat Linux, Fedora Core and FreeBSD
Reid Spencer00812e22005-05-17 02:47:27 +0000[diff] [blame]514 (and probably other unix-like systems).</li>
Chris Lattner000c73b2008-02-06 06:30:34 +0000[diff] [blame]515<li>PowerPC and X86-based Mac OS X systems, running 10.3 and above in 32-bit and
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]516 64-bit modes.</li>
Chris Lattner252b83d2008-02-06 18:00:06 +0000[diff] [blame]517<li>Intel and AMD machines running on Win32 using MinGW libraries (native).</li>
Chris Lattner57a460e2007-05-23 04:39:32 +0000[diff] [blame]518<li>Intel and AMD machines running on Win32 with the Cygwin libraries (limited
519 support is available for native builds with Visual C++).</li>
Gabor Greif96a89c72008-06-05 18:39:01 +0000[diff] [blame]520<li>Sun UltraSPARC workstations running Solaris 10.</li>
John Criswell9321fa82005-05-13 20:28:15 +0000[diff] [blame]521<li>Alpha-based machines running Debian GNU/Linux.</li>
Gabor Greif96a89c72008-06-05 18:39:01 +0000[diff] [blame]522<li>Itanium-based (IA64) machines running Linux and HP-UX.</li>
Chris Lattner4654bdb2004-06-01 18:22:41 +0000[diff] [blame]523</ul>
524
Chris Lattnerbc5786b2008-06-05 06:57:39 +0000[diff] [blame]525<p>The core LLVM infrastructure uses GNU autoconf to adapt itself
Brian Gaekeb0fd7612004-05-09 05:28:35 +0000[diff] [blame]526to the machine and operating system on which it is built. However, minor
527porting may be required to get LLVM to work on new platforms. We welcome your
528portability patches and reports of successful builds or error messages.</p>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]529
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]530</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]531
532<!-- *********************************************************************** -->
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]533<div class="doc_section">
534 <a name="knownproblems">Known Problems</a>
535</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]536<!-- *********************************************************************** -->
537
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]538<div class="doc_text">
539
540<p>This section contains all known problems with the LLVM system, listed by
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]541component. As new problems are discovered, they will be added to these
Chris Lattner5eccca42003-12-12 21:22:16 +0000[diff] [blame]542sections. If you run into a problem, please check the <a
Chris Lattnerc463b272005-10-29 07:07:09 +0000[diff] [blame]543href="http://llvm.org/bugs/">LLVM bug database</a> and submit a bug if
Chris Lattner5eccca42003-12-12 21:22:16 +0000[diff] [blame]544there isn't already one.</p>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]545
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]546</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]547
Chris Lattnerf5ee1702004-03-14 02:03:02 +0000[diff] [blame]548<!-- ======================================================================= -->
549<div class="doc_subsection">
550 <a name="experimental">Experimental features included with this release</a>
551</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]552
Chris Lattnerf5ee1702004-03-14 02:03:02 +0000[diff] [blame]553<div class="doc_text">
554
Misha Brukman6df9e2c2004-05-12 21:46:05 +0000[diff] [blame]555<p>The following components of this LLVM release are either untested, known to
556be broken or unreliable, or are in early development. These components should
557not be relied on, and bugs should not be filed against them, but they may be
558useful to some people. In particular, if you would like to work on one of these
Chris Lattner2b659ef2008-02-12 06:29:45 +0000[diff] [blame]559components, please contact us on the <a
560href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVMdev list</a>.</p>
Chris Lattnerf5ee1702004-03-14 02:03:02 +0000[diff] [blame]561
562<ul>
Gabor Greifa8b84902008-10-14 11:31:14 +0000[diff] [blame]563<li>The MSIL, IA64, Alpha, SPU, MIPS, and PIC16 backends are experimental.</li>
Duncan Sands27aff872008-06-08 20:18:35 +0000[diff] [blame]564<li>The llc "<tt>-filetype=asm</tt>" (the default) is the only supported
Chris Lattner000c73b2008-02-06 06:30:34 +0000[diff] [blame]565 value for this option.</li>
Chris Lattnerf5ee1702004-03-14 02:03:02 +0000[diff] [blame]566</ul>
567
568</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]569
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]570<!-- ======================================================================= -->
571<div class="doc_subsection">
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]572 <a name="x86-be">Known problems with the X86 back-end</a>
John Criswellc0c186d2005-11-08 21:11:33 +0000[diff] [blame]573</div>
574
575<div class="doc_text">
576
577<ul>
Anton Korobeynikova6094be2008-06-08 10:24:13 +0000[diff] [blame]578 <li>The X86 backend does not yet support
579 all <a href="http://llvm.org/PR879">inline assembly that uses the X86
580 floating point stack</a>. It supports the 'f' and 't' constraints, but not
581 'u'.</li>
Chris Lattnere6e1b352008-06-08 21:19:07 +0000[diff] [blame]582 <li>The X86 backend generates inefficient floating point code when configured
583 to generate code for systems that don't have SSE2.</li>
Duncan Sands47eff2b2008-06-08 19:38:43 +0000[diff] [blame]584 <li>Win64 code generation wasn't widely tested. Everything should work, but we
Anton Korobeynikova6094be2008-06-08 10:24:13 +0000[diff] [blame]585 expect small issues to happen. Also, llvm-gcc cannot build mingw64 runtime
586 currently due
587 to <a href="http://llvm.org/PR2255">several</a>
Chris Lattner04af7cb2008-06-08 23:12:47 +0000[diff] [blame]588 <a href="http://llvm.org/PR2257">bugs</a> due to lack of support for the
589 'u' inline assembly constraint and X87 floating point inline assembly.</li>
Dan Gohman8207ba92008-06-08 23:05:11 +0000[diff] [blame]590 <li>The X86-64 backend does not yet support the LLVM IR instruction
591 <tt>va_arg</tt>. Currently, the llvm-gcc front-end supports variadic
592 argument constructs on X86-64 by lowering them manually.</li>
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]593</ul>
594
595</div>
596
597<!-- ======================================================================= -->
598<div class="doc_subsection">
599 <a name="ppc-be">Known problems with the PowerPC back-end</a>
600</div>
601
602<div class="doc_text">
603
604<ul>
Nicolas Geoffraye4285dc2007-05-15 09:21:28 +0000[diff] [blame]605<li>The Linux PPC32/ABI support needs testing for the interpreter and static
Chris Lattner57a460e2007-05-23 04:39:32 +0000[diff] [blame]606compilation, and lacks support for debug information.</li>
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]607</ul>
608
609</div>
610
611<!-- ======================================================================= -->
612<div class="doc_subsection">
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]613 <a name="arm-be">Known problems with the ARM back-end</a>
614</div>
615
616<div class="doc_text">
617
618<ul>
Chris Lattner57a460e2007-05-23 04:39:32 +0000[diff] [blame]619<li>Thumb mode works only on ARMv6 or higher processors. On sub-ARMv6
Duncan Sandsc90d68b2007-09-26 15:59:54 +0000[diff] [blame]620processors, thumb programs can crash or produce wrong
Chris Lattner57a460e2007-05-23 04:39:32 +0000[diff] [blame]621results (<a href="http://llvm.org/PR1388">PR1388</a>).</li>
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]622<li>Compilation for ARM Linux OABI (old ABI) is supported, but not fully tested.
623</li>
Chris Lattnere6e1b352008-06-08 21:19:07 +0000[diff] [blame]624<li>There is a bug in QEMU-ARM (&lt;= 0.9.0) which causes it to incorrectly
625 execute
Chris Lattner57a460e2007-05-23 04:39:32 +0000[diff] [blame]626programs compiled with LLVM. Please use more recent versions of QEMU.</li>
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]627</ul>
628
629</div>
630
631<!-- ======================================================================= -->
632<div class="doc_subsection">
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]633 <a name="sparc-be">Known problems with the SPARC back-end</a>
634</div>
635
636<div class="doc_text">
637
638<ul>
639<li>The SPARC backend only supports the 32-bit SPARC ABI (-m32), it does not
640 support the 64-bit SPARC ABI (-m64).</li>
641</ul>
642
643</div>
644
645<!-- ======================================================================= -->
646<div class="doc_subsection">
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]647 <a name="alpha-be">Known problems with the Alpha back-end</a>
648</div>
649
650<div class="doc_text">
651
652<ul>
653
654<li>On 21164s, some rare FP arithmetic sequences which may trap do not have the
655appropriate nops inserted to ensure restartability.</li>
656
John Criswellc0c186d2005-11-08 21:11:33 +0000[diff] [blame]657</ul>
658</div>
659
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]660<!-- ======================================================================= -->
661<div class="doc_subsection">
662 <a name="ia64-be">Known problems with the IA64 back-end</a>
663</div>
664
665<div class="doc_text">
666
667<ul>
Chris Lattner04af7cb2008-06-08 23:12:47 +0000[diff] [blame]668<li>The Itanium backend is highly experimental, and has a number of known
669 issues. We are looking for a maintainer for the Itanium backend. If you
670 are interested, please contact the llvmdev mailing list.</li>
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]671</ul>
672
673</div>
674
675<!-- ======================================================================= -->
676<div class="doc_subsection">
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]677 <a name="c-be">Known problems with the C back-end</a>
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]678</div>
679
680<div class="doc_text">
681
682<ul>
Chris Lattner5733b272008-06-05 06:35:40 +0000[diff] [blame]683<li><a href="http://llvm.org/PR802">The C backend has only basic support for
684 inline assembly code</a>.</li>
Chris Lattner725a0d82007-09-26 06:01:35 +0000[diff] [blame]685<li><a href="http://llvm.org/PR1658">The C backend violates the ABI of common
686 C++ programs</a>, preventing intermixing between C++ compiled by the CBE and
Duncan Sands27aff872008-06-08 20:18:35 +0000[diff] [blame]687 C++ code compiled with llc or native compilers.</li>
Duncan Sandsf74c0cc2008-02-10 13:40:55 +0000[diff] [blame]688<li>The C backend does not support all exception handling constructs.</li>
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]689</ul>
690
691</div>
John Criswellc0c186d2005-11-08 21:11:33 +0000[diff] [blame]692
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]693
694<!-- ======================================================================= -->
695<div class="doc_subsection">
Chris Lattner5733b272008-06-05 06:35:40 +0000[diff] [blame]696 <a name="c-fe">Known problems with the llvm-gcc C front-end</a>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]697</div>
Chris Lattner47588f92003-10-02 05:07:23 +0000[diff] [blame]698
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]699<div class="doc_text">
Chris Lattnerc5d658a2006-03-03 00:34:26 +0000[diff] [blame]700
Chris Lattner252b83d2008-02-06 18:00:06 +0000[diff] [blame]701<p>llvm-gcc does not currently support <a href="http://llvm.org/PR869">Link-Time
702Optimization</a> on most platforms "out-of-the-box". Please inquire on the
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]703llvmdev mailing list if you are interested.</p>
Chris Lattnerd9ea0172006-08-08 17:27:28 +0000[diff] [blame]704
Chris Lattner5733b272008-06-05 06:35:40 +0000[diff] [blame]705<p>The only major language feature of GCC not supported by llvm-gcc is
706 the <tt>__builtin_apply</tt> family of builtins. However, some extensions
707 are only supported on some targets. For example, trampolines are only
Duncan Sands27aff872008-06-08 20:18:35 +0000[diff] [blame]708 supported on some targets (these are used when you take the address of a
709 nested function).</p>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]710
Chris Lattner5733b272008-06-05 06:35:40 +0000[diff] [blame]711<p>If you run into GCC extensions which are not supported, please let us know.
712</p>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]713
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]714</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]715
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]716<!-- ======================================================================= -->
717<div class="doc_subsection">
Chris Lattner5733b272008-06-05 06:35:40 +0000[diff] [blame]718 <a name="c++-fe">Known problems with the llvm-gcc C++ front-end</a>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]719</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]720
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]721<div class="doc_text">
722
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]723<p>The C++ front-end is considered to be fully
Chris Lattner7506b1d2004-12-07 08:04:13 +0000[diff] [blame]724tested and works for a number of non-trivial programs, including LLVM
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]725itself, Qt, Mozilla, etc.</p>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]726
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]727<ul>
Anton Korobeynikov0021fc12008-10-11 18:27:16 +0000[diff] [blame]728<li>Exception handling works well on the X86 and PowerPC targets. Currently
729 only linux and darwin targets are supported (both 32 and 64 bit).</li>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]730</ul>
Chris Lattnerfcc54b32003-10-07 22:14:37 +0000[diff] [blame]731
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]732</div>
733
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]734
Chris Lattner2b659ef2008-02-12 06:29:45 +0000[diff] [blame]735<!-- ======================================================================= -->
736<div class="doc_subsection">
Chris Lattner5733b272008-06-05 06:35:40 +0000[diff] [blame]737 <a name="ada-fe">Known problems with the llvm-gcc Ada front-end</a>
Chris Lattner2b659ef2008-02-12 06:29:45 +0000[diff] [blame]738</div>
739
740<div class="doc_text">
741The llvm-gcc 4.2 Ada compiler works fairly well, however this is not a mature
742technology and problems should be expected.
743<ul>
Duncan Sands27aff872008-06-08 20:18:35 +0000[diff] [blame]744<li>The Ada front-end currently only builds on X86-32. This is mainly due
Chris Lattner2b659ef2008-02-12 06:29:45 +0000[diff] [blame]745to lack of trampoline support (pointers to nested functions) on other platforms,
Duncan Sands27aff872008-06-08 20:18:35 +0000[diff] [blame]746however it <a href="http://llvm.org/PR2006">also fails to build on X86-64</a>
Chris Lattner2b659ef2008-02-12 06:29:45 +0000[diff] [blame]747which does support trampolines.</li>
748<li>The Ada front-end <a href="http://llvm.org/PR2007">fails to bootstrap</a>.
749Workaround: configure with --disable-bootstrap.</li>
Duncan Sands978bcee2008-10-13 17:27:23 +0000[diff] [blame]750<li>The c380004, <a href="http://llvm.org/PR2010">c393010</a>
751and <a href="http://llvm.org/PR2421">cxg2021</a> ACATS tests fail
752(c380004 also fails with gcc-4.2 mainline).</li>
753<li>Some gcc specific Ada tests continue to crash the compiler.</li>
Chris Lattner2b659ef2008-02-12 06:29:45 +0000[diff] [blame]754<li>The -E binder option (exception backtraces)
755<a href="http://llvm.org/PR1982">does not work</a> and will result in programs
756crashing if an exception is raised. Workaround: do not use -E.</li>
757<li>Only discrete types <a href="http://llvm.org/PR1981">are allowed to start
758or finish at a non-byte offset</a> in a record. Workaround: do not pack records
759or use representation clauses that result in a field of a non-discrete type
760starting or finishing in the middle of a byte.</li>
Chris Lattnere6e1b352008-06-08 21:19:07 +0000[diff] [blame]761<li>The <tt>lli</tt> interpreter <a href="http://llvm.org/PR2009">considers
762'main' as generated by the Ada binder to be invalid</a>.
763Workaround: hand edit the file to use pointers for <tt>argv</tt> and
764<tt>envp</tt> rather than integers.</li>
765<li>The <tt>-fstack-check</tt> option <a href="http://llvm.org/PR2008">is
766ignored</a>.</li>
Chris Lattner2b659ef2008-02-12 06:29:45 +0000[diff] [blame]767</ul>
768</div>
769
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]770<!-- *********************************************************************** -->
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]771<div class="doc_section">
772 <a name="additionalinfo">Additional Information</a>
773</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]774<!-- *********************************************************************** -->
775
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]776<div class="doc_text">
777
Chris Lattner416db102005-05-16 17:13:10 +0000[diff] [blame]778<p>A wide variety of additional information is available on the <a
Chris Lattnerb4b0ce72007-05-18 00:44:29 +0000[diff] [blame]779href="http://llvm.org">LLVM web page</a>, in particular in the <a
780href="http://llvm.org/docs/">documentation</a> section. The web page also
Reid Spencer669ed452007-07-09 08:04:31 +0000[diff] [blame]781contains versions of the API documentation which is up-to-date with the
782Subversion version of the source code.
Misha Brukman109d9e82005-03-30 19:14:24 +0000[diff] [blame]783You can access versions of these documents specific to this release by going
784into the "<tt>llvm/doc/</tt>" directory in the LLVM tree.</p>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]785
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]786<p>If you have any questions or comments about LLVM, please feel free to contact
Chris Lattnerc463b272005-10-29 07:07:09 +0000[diff] [blame]787us via the <a href="http://llvm.org/docs/#maillist"> mailing
Chris Lattner5eccca42003-12-12 21:22:16 +0000[diff] [blame]788lists</a>.</p>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]789
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]790</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]791
792<!-- *********************************************************************** -->
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]793
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]794<hr>
Misha Brukman2061e892003-11-22 01:23:39 +0000[diff] [blame]795<address>
Misha Brukman38847d52003-12-21 22:53:21 +0000[diff] [blame]796 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman2061e892003-11-22 01:23:39 +0000[diff] [blame]797 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
798 <a href="http://validator.w3.org/check/referer"><img
Chris Lattner8170c102008-02-10 08:18:42 +0000[diff] [blame]799 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
Misha Brukman2061e892003-11-22 01:23:39 +0000[diff] [blame]800
Chris Lattnerb4b0ce72007-05-18 00:44:29 +0000[diff] [blame]801 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]802 Last modified: $Date$
Misha Brukman2061e892003-11-22 01:23:39 +0000[diff] [blame]803</address>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]804
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]805</body>
806</html>