Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 524a5ddbb346dbdd29437b17d86d2be2aaec36dd / . / docs / ReleaseNotes.html
blob: 2bc3f0f533fffad65793be01977f09ce5e77df0c [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 Lattner38ab04c2009-02-25 04:41:31 +0000[diff] [blame]7 <title>LLVM 2.5 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 Lattner38ab04c2009-02-25 04:41:31 +0000[diff] [blame]11<div class="doc_title">LLVM 2.5 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>
Chris Lattnerab68e9e2009-02-26 22:33:38 +0000[diff] [blame]16 <li><a href="#externalproj">External Projects Using LLVM 2.5</a></li>
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]17 <li><a href="#whatsnew">What's New in LLVM 2.5?</a></li>
Chris Lattner4b538b92004-04-30 22:17:12 +0000[diff] [blame]18 <li><a href="GettingStarted.html">Installation Instructions</a></li>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]19 <li><a href="#portability">Portability and Supported Platforms</a></li>
Dan Gohman44aa9212008-10-14 16:23:02 +0000[diff] [blame]20 <li><a href="#knownproblems">Known Problems</a></li>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]21 <li><a href="#additionalinfo">Additional Information</a></li>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]22</ol>
23
Chris Lattner7911ce22004-05-23 21:07:27 +0000[diff] [blame]24<div class="doc_author">
Dan Gohman44aa9212008-10-14 16:23:02 +0000[diff] [blame]25 <p>Written by the <a href="http://llvm.org">LLVM Team</a></p>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]26</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]27
28<!-- *********************************************************************** -->
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]29<div class="doc_section">
30 <a name="intro">Introduction</a>
31</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]32<!-- *********************************************************************** -->
33
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]34<div class="doc_text">
35
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]36<p>This document contains the release notes for the LLVM Compiler
Chris Lattner38ab04c2009-02-25 04:41:31 +0000[diff] [blame]37Infrastructure, release 2.5. Here we describe the status of LLVM, including
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]38major improvements from the previous release and significant known problems.
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]39All LLVM releases may be downloaded from the <a
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]40href="http://llvm.org/releases/">LLVM releases web site</a>.</p>
Chris Lattner19092612003-10-02 16:38:05 +0000[diff] [blame]41
Chris Lattner7506b1d2004-12-07 08:04:13 +0000[diff] [blame]42<p>For more information about LLVM, including information about the latest
Chris Lattnerc463b272005-10-29 07:07:09 +0000[diff] [blame]43release, please check out the <a href="http://llvm.org/">main LLVM
Chris Lattner47ad72c2003-10-07 21:38:31 +0000[diff] [blame]44web site</a>. If you have questions or comments, the <a
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]45href="http://mail.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVM Developer's Mailing
46List</a> is a good place to send them.</p>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]47
Mikhail Glushenkovea65d7d2008-10-13 02:08:34 +0000[diff] [blame]48<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]49main LLVM web page, this document applies to the <i>next</i> release, not the
Gabor Greiffa933f82008-10-14 11:00:32 +0000[diff] [blame]50current one. To see the release notes for a specific release, please see the
Chris Lattner84977642007-09-21 03:54:09 +0000[diff] [blame]51<a href="http://llvm.org/releases/">releases page</a>.</p>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]52
53</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]54
Chris Lattner38ab04c2009-02-25 04:41:31 +0000[diff] [blame]55<!-- Unfinished features in 2.5:
Chris Lattneracce85d2008-02-10 07:46:44 +0000[diff] [blame]56 Machine LICM
57 Machine Sinking
Chris Lattner74c80df2009-02-25 06:34:50 +0000[diff] [blame]58 target-specific intrinsics
59 gold lto plugin
Chris Lattner99f375b2009-02-25 07:32:53 +0000[diff] [blame]60 pre-alloc splitter, strong phi elim
Gabor Greif4906abe2009-03-02 12:02:51 +0000[diff] [blame]61 <tt>llc -enable-value-prop</tt>, propagation of value info
62 (sign/zero ext info) from one MBB to another
Chris Lattner74c80df2009-02-25 06:34:50 +0000[diff] [blame]63 debug info for optimized code
Chris Lattnerc1771e32009-02-26 07:29:17 +0000[diff] [blame]64 interpreter + libffi
Chris Lattnerab68e9e2009-02-26 22:33:38 +0000[diff] [blame]65 postalloc scheduler: anti dependence breaking, hazard recognizer?
66
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]67initial support for debug line numbers when optimization enabled, not useful in
68 2.5 but will be for 2.6.
69
Chris Lattneracce85d2008-02-10 07:46:44 +0000[diff] [blame]70 -->
Mikhail Glushenkovea65d7d2008-10-13 02:08:34 +0000[diff] [blame]71
Chris Lattner547a3912008-10-12 19:47:48 +0000[diff] [blame]72 <!-- for announcement email:
Chris Lattner74c80df2009-02-25 06:34:50 +0000[diff] [blame]73 -->
Chris Lattneracce85d2008-02-10 07:46:44 +0000[diff] [blame]74
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]75<!-- *********************************************************************** -->
76<div class="doc_section">
77 <a name="subproj">Sub-project Status Update</a>
Chris Lattnerea34f642008-06-08 21:34:41 +0000[diff] [blame]78</div>
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]79<!-- *********************************************************************** -->
Chris Lattnerea34f642008-06-08 21:34:41 +0000[diff] [blame]80
81<div class="doc_text">
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]82<p>
Chris Lattner38ab04c2009-02-25 04:41:31 +0000[diff] [blame]83The LLVM 2.5 distribution currently consists of code from the core LLVM
Bill Wendling6bc15282009-03-02 04:28:18 +0000[diff] [blame]84repository &mdash;which roughly includes the LLVM optimizers, code generators
85and supporting tools &mdash; and the llvm-gcc repository. In addition to this
86code, the LLVM Project includes other sub-projects that are in development. The
87two which are the most actively developed are the <a href="#clang">Clang
88Project</a> and the <a href="#vmkit">VMKit Project</a>.
Bill Wendling63d8c552009-03-02 04:28:57 +0000[diff] [blame]89</p>
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]90
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]91</div>
92
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]93
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]94<!--=========================================================================-->
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]95<div class="doc_subsection">
Chris Lattnerfb97b2d2008-10-13 18:11:54 +0000[diff] [blame]96<a name="clang">Clang: C/C++/Objective-C Frontend Toolkit</a>
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]97</div>
98
99<div class="doc_text">
100
101<p>The <a href="http://clang.llvm.org/">Clang project</a> is an effort to build
Bill Wendling6bc15282009-03-02 04:28:18 +0000[diff] [blame]102a set of new 'LLVM native' front-end technologies for the LLVM optimizer and
103code generator. While Clang is not included in the LLVM 2.5 release, it is
104continuing to make major strides forward in all areas. Its C and Objective-C
Chris Lattnerf40c40b2009-02-25 05:09:54 +0000[diff] [blame]105parsing and code generation support is now very solid. For example, it is
Bill Wendling6bc15282009-03-02 04:28:18 +0000[diff] [blame]106capable of successfully building many real-world applications for X86-32
107andX86-64,
Chris Lattnerab68e9e2009-02-26 22:33:38 +0000[diff] [blame]108including <a href="http://wiki.freebsd.org/BuildingFreeBSDWithClang">the FreeBSD
Bill Wendling6bc15282009-03-02 04:28:18 +0000[diff] [blame]109kernel</a>. C++ is also
110making <a href="http://clang.llvm.org/cxx_status.html">incredible progress</a>,
111and work on templates has recently started.</p>
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]112
Chris Lattnerf40c40b2009-02-25 05:09:54 +0000[diff] [blame]113<p>While Clang is not yet production quality, it is progressing very nicely and
114is quite usable for building many C and Objective-C applications. If you are
115interested in fast compiles and good diagnostics, we encourage you to try it out
Chris Lattner74c80df2009-02-25 06:34:50 +0000[diff] [blame]116by <a href="http://clang.llvm.org/get_started.html">building from mainline</a>
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]117and reporting any issues you hit to the <a
Chris Lattnerf40c40b2009-02-25 05:09:54 +0000[diff] [blame]118href="http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev">Clang front-end mailing
119list</a>.</p>
120
121<p>In the LLVM 2.5 time-frame, the Clang team has made many improvements:</p>
Bill Wendling741748a2008-10-27 09:27:33 +0000[diff] [blame]122
Daniel Dunbar13739432008-10-14 23:25:09 +0000[diff] [blame]123<ul>
Chris Lattnerf40c40b2009-02-25 05:09:54 +0000[diff] [blame]124<li>Clang now has a new driver, which is focused on providing a GCC-compatible
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]125 interface.</li>
126<li>The X86-64 ABI is now supported.</li>
127<li>Precompiled header support is now implemented.</li>
Chris Lattnerf40c40b2009-02-25 05:09:54 +0000[diff] [blame]128<li>Objective-C support is significantly improved beyond LLVM 2.4, supporting
129 many features, such as Objective-C Garbage Collection.</li>
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]130<li>Many many bugs are fixed and many features have been added.</li>
Bill Wendling6bc15282009-03-02 04:28:18 +0000[diff] [blame]131</ul>
Chris Lattnerfb97b2d2008-10-13 18:11:54 +0000[diff] [blame]132</div>
133
134<!--=========================================================================-->
135<div class="doc_subsection">
136<a name="clangsa">Clang Static Analyzer</a>
137</div>
138
139<div class="doc_text">
140
Ted Kremenek8b3894c2009-02-27 07:01:32 +0000[diff] [blame]141<p>Previously announced in the last LLVM release, the Clang project also
142includes an early stage static source code analysis tool for <a
143href="http://clang.llvm.org/StaticAnalysis.html">automatically finding bugs</a>
144in C and Objective-C programs. The tool performs a growing set of checks to find
145bugs that occur on a specific path within a program.</p>
Chris Lattnercc042612008-10-14 00:52:49 +0000[diff] [blame]146
Ted Kremenek8b3894c2009-02-27 07:01:32 +0000[diff] [blame]147<p>In the LLVM 2.5 time-frame there have been many significant improvements to
148the analyzer's core path simulation engine and machinery for generating
149path-based bug reports to end-users. Particularly noteworthy improvements
150include experimental support for full field-sensitivity and reasoning about heap
151objects as well as an improved value-constraints subengine that does a much
152better job of reasoning about inequality relationships (e.g., <tt>x &gt; 2</tt>)
153between variables and constants.
Ted Kremenek3c3ec0c2008-10-14 05:14:21 +0000[diff] [blame]154
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]155<p>The set of checks performed by the static analyzer continues to expand, and
Ted Kremenek8b3894c2009-02-27 07:01:32 +0000[diff] [blame]156future plans for the tool include full source-level inter-procedural analysis
157and deeper checks such as buffer overrun detection. There are many opportunities
158to extend and enhance the static analyzer, and anyone interested in working on
159this project is encouraged to get involved!</p>
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]160
161</div>
162
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]163<!--=========================================================================-->
164<div class="doc_subsection">
Nicolas Geoffray99a4d302008-10-14 19:23:04 +0000[diff] [blame]165<a name="vmkit">VMKit: JVM/CLI Virtual Machine Implementation</a>
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]166</div>
167
168<div class="doc_text">
169<p>
Nicolas Geoffray99a4d302008-10-14 19:23:04 +0000[diff] [blame]170The <a href="http://vmkit.llvm.org/">VMKit project</a> is an implementation of
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]171a JVM and a CLI Virtual Machines (Microsoft .NET is an
172implementation of the CLI) using the Just-In-Time compiler of LLVM.</p>
173
Nicolas Geoffray09c9ed42009-03-02 09:25:38 +0000[diff] [blame]174<p>Following LLVM 2.5, VMKit has its second release that you can find on its
Nicolas Geoffray56d313d2008-10-15 20:25:04 +0000[diff] [blame]175<a href="http://vmkit.llvm.org/releases/">webpage</a>. The release includes
176bug fixes, cleanup and new features. The major changes are:</p>
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]177
Nicolas Geoffray99a4d302008-10-14 19:23:04 +0000[diff] [blame]178<ul>
179
Chris Lattner75e26ca2009-02-26 05:20:36 +0000[diff] [blame]180<li>Ahead of Time compiler: compiles .class files to llvm .bc. VMKit uses this
Gabor Greif29ecc1c2009-03-02 11:34:51 +0000[diff] [blame]181functionality to native compile the standard classes (e.g. java.lang.String).
Gabor Greif4906abe2009-03-02 12:02:51 +0000[diff] [blame]182Users can compile AoT .class files into dynamic libraries and run them with the
Chris Lattner75e26ca2009-02-26 05:20:36 +0000[diff] [blame]183help of VMKit.</li>
184
185<li>New exception model: the dwarf exception model is very slow for
186exception-intensive applications, so the JVM has had a new implementation of
187exceptions which check at each function call if an exception happened. There is
188a low performance penalty on applications without exceptions, but it is a big
189gain for exception-intensive applications. For example the jack benchmark in
190Spec JVM98 is 6x faster (performance gain of 83%).</li>
191
Nicolas Geoffray09c9ed42009-03-02 09:25:38 +0000[diff] [blame]192<li>User-level management of thread stacks, so that thread local data access
193at runtime is fast and portable. </li>
194
195<li>Implementation of biased locking for faster object synchronizations at
196runtime.</li>
197
198<li>New support for OSX/X64, Linux/X64 (with the Boehm GC) and Linux/ppc32.</li>
Nicolas Geoffray99a4d302008-10-14 19:23:04 +0000[diff] [blame]199
200</ul>
Chris Lattner96a445e2008-10-13 18:01:01 +0000[diff] [blame]201</div>
202
Chris Lattnerab68e9e2009-02-26 22:33:38 +0000[diff] [blame]203<!-- *********************************************************************** -->
204<div class="doc_section">
205 <a name="externalproj">External Projects Using LLVM 2.5</a>
206</div>
207<!-- *********************************************************************** -->
208
209<!--=========================================================================-->
210<div class="doc_subsection">
211<a name="pure">Pure</a>
212</div>
213
214<div class="doc_text">
215<p>
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]216<a href="http://pure-lang.googlecode.com/">Pure</a>
217is an algebraic/functional programming language based on term rewriting.
Chris Lattnerab68e9e2009-02-26 22:33:38 +0000[diff] [blame]218Programs are collections of equations which are used to evaluate expressions in
219a symbolic fashion. Pure offers dynamic typing, eager and lazy evaluation,
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]220lexical closures, a hygienic macro system (also based on term rewriting),
Chris Lattnerab68e9e2009-02-26 22:33:38 +0000[diff] [blame]221built-in list and matrix support (including list and matrix comprehensions) and
222an easy-to-use C interface. The interpreter uses LLVM as a backend to
223 JIT-compile Pure programs to fast native code.</p>
224
225<p>In addition to the usual algebraic data structures, Pure also has
226MATLAB-style matrices in order to support numeric computations and signal
227processing in an efficient way. Pure is mainly aimed at mathematical
228applications right now, but it has been designed as a general purpose language.
229The dynamic interpreter environment and the C interface make it possible to use
230it as a kind of functional scripting language for many application areas.
231</p>
232</div>
233
234
235<!--=========================================================================-->
236<div class="doc_subsection">
237<a name="ldc">LLVM D Compiler</a>
238</div>
239
240<div class="doc_text">
241<p>
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]242<a href="http://www.dsource.org/projects/ldc">LDC</a> is an implementation of
243the D Programming Language using the LLVM optimizer and code generator.
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]244The LDC project works great with the LLVM 2.5 release. General improvements in
245this
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]246cycle have included new inline asm constraint handling, better debug info
247support, general bugfixes, and better x86-64 support. This has allowed
Chris Lattnerab68e9e2009-02-26 22:33:38 +0000[diff] [blame]248some major improvements in LDC, getting us much closer to being as
249fully featured as the original DMD compiler from DigitalMars.
250</p>
251</div>
252
Chris Lattnerbc31caf2009-02-28 18:58:01 +0000[diff] [blame]253<!--=========================================================================-->
254<div class="doc_subsection">
255<a name="RoadsendPHP">Roadsend PHP</a>
256</div>
257
258<div class="doc_text">
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]259<p><a href="http://code.roadsend.com/rphp">Roadsend PHP</a> (rphp) is an open
260source compiler for the PHP programming language that uses LLVM for its
261optimizer, JIT, and static compiler. This is a reimplementation of an earlier
262project that is now based on the LLVM.</p>
Chris Lattnerbc31caf2009-02-28 18:58:01 +0000[diff] [blame]263</div>
Chris Lattnerab68e9e2009-02-26 22:33:38 +0000[diff] [blame]264
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]265
266<!-- *********************************************************************** -->
267<div class="doc_section">
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]268 <a name="whatsnew">What's New in LLVM 2.5?</a>
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]269</div>
270<!-- *********************************************************************** -->
271
Chris Lattnerf8e0b4e2008-06-08 22:59:35 +0000[diff] [blame]272<div class="doc_text">
273
Bill Wendling741748a2008-10-27 09:27:33 +0000[diff] [blame]274<p>This release includes a huge number of bug fixes, performance tweaks, and
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]275minor improvements. Some of the major improvements and new features are listed
276in this section.
Chris Lattnerf8e0b4e2008-06-08 22:59:35 +0000[diff] [blame]277</p>
278</div>
279
Chris Lattner8348b472008-06-08 21:58:17 +0000[diff] [blame]280<!--=========================================================================-->
Chris Lattnerea34f642008-06-08 21:34:41 +0000[diff] [blame]281<div class="doc_subsection">
Chris Lattner252b83d2008-02-06 18:00:06 +0000[diff] [blame]282<a name="majorfeatures">Major New Features</a>
283</div>
284
285<div class="doc_text">
286
Chris Lattner38ab04c2009-02-25 04:41:31 +0000[diff] [blame]287<p>LLVM 2.5 includes several major new capabilities:</p>
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]288
289<ul>
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]290<li>LLVM 2.5 includes a brand new <a
291href="http://en.wikipedia.org/wiki/XCore">XCore</a> backend.</li>
Chris Lattner74c80df2009-02-25 06:34:50 +0000[diff] [blame]292
Bill Wendlingdc1350e2009-03-02 07:42:15 +0000[diff] [blame]293<li>llvm-gcc now generally supports the GFortran front-end, and the precompiled
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]294release binaries now support Fortran, even on Mac OS/X.</li>
Chris Lattner74c80df2009-02-25 06:34:50 +0000[diff] [blame]295
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]296<li>CMake is now used by the <a href="GettingStartedVS.html">LLVM build process
297on Windows</a>. It automatically generates Visual Studio project files (and
298more) from a set of simple text files. This makes it much easier to
299maintain. In time, we'd like to standardize on CMake for everything.</li>
Chris Lattnerc1771e32009-02-26 07:29:17 +0000[diff] [blame]300
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]301<li>LLVM 2.5 now uses (and includes) Google Test for unit testing.</li>
Chris Lattnerc1771e32009-02-26 07:29:17 +0000[diff] [blame]302
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]303<li>The LLVM native code generator now supports arbitrary precision integers.
304Types like <tt>i33</tt> have long been valid in the LLVM IR, but were previously
305only supported by the interpreter. Note that the C backend still does not
306support these.</li>
Chris Lattner99f375b2009-02-25 07:32:53 +0000[diff] [blame]307
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]308<li>LLVM 2.5 no longer uses 'bison,' so it is easier to build on Windows.</li>
Chris Lattner8170c102008-02-10 08:18:42 +0000[diff] [blame]309</ul>
310
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]311</div>
312
Chris Lattner791f77b2008-06-05 06:25:56 +0000[diff] [blame]313
314<!--=========================================================================-->
315<div class="doc_subsection">
Chris Lattnerea34f642008-06-08 21:34:41 +0000[diff] [blame]316<a name="llvm-gcc">llvm-gcc 4.2 Improvements</a>
Chris Lattner791f77b2008-06-05 06:25:56 +0000[diff] [blame]317</div>
318
319<div class="doc_text">
320
Chris Lattnerb82eb062008-10-13 20:47:20 +0000[diff] [blame]321<p>LLVM fully supports the llvm-gcc 4.2 front-end, which marries the GCC
322front-ends and driver with the LLVM optimizer and code generator. It currently
323includes support for the C, C++, Objective-C, Ada, and Fortran front-ends.</p>
Chris Lattner791f77b2008-06-05 06:25:56 +0000[diff] [blame]324
Chris Lattnerf8e0b4e2008-06-08 22:59:35 +0000[diff] [blame]325<ul>
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]326<li>In this release, the GCC inliner is completely disabled. Previously the GCC
327inliner was used to handle always-inline functions and other cases. This caused
328problems with code size growth, and it is completely disabled in this
329release.</li>
330
331<li>llvm-gcc (and LLVM in general) now support code generation for stack
332canaries, which is an effective form of <a
333href="http://en.wikipedia.org/wiki/Stack-smashing_protection">buffer overflow
334protection</a>. llvm-gcc supports this with the <tt>-fstack-protector</tt>
335command line option (just like GCC). In LLVM IR, you can request code
336generation for stack canaries with function attributes.
337</li>
Gabor Greifa3bd11d2008-06-09 06:06:18 +0000[diff] [blame]338</ul>
Chris Lattner791f77b2008-06-05 06:25:56 +0000[diff] [blame]339
Chris Lattner791f77b2008-06-05 06:25:56 +0000[diff] [blame]340</div>
341
342
Chris Lattner84977642007-09-21 03:54:09 +0000[diff] [blame]343<!--=========================================================================-->
344<div class="doc_subsection">
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]345<a name="coreimprovements">LLVM IR and Core Improvements</a>
Chris Lattnerf304ffc2008-02-10 08:17:19 +0000[diff] [blame]346</div>
347
348<div class="doc_text">
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]349<p>LLVM IR has several new features that are used by our existing front-ends and
350can be useful if you are writing a front-end for LLVM:</p>
Chris Lattnerf304ffc2008-02-10 08:17:19 +0000[diff] [blame]351
Chris Lattner791f77b2008-06-05 06:25:56 +0000[diff] [blame]352<ul>
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]353<li>The <a href="LangRef.html#i_shufflevector">shufflevector</a> instruction
354has been generalized to allow different shuffle mask width than its input
355vectors. This allows you to use shufflevector to combine two
356"&lt;4 x float&gt;" vectors into a "&lt;8 x float&gt;" for example.</li>
357
358<li>LLVM IR now supports new intrinsics for computing and acting on <a
359href="LangRef.html#int_overflow">overflow of integer operations</a>. This allows
360efficient code generation for languages that must trap or throw an exception on
361overflow. While these intrinsics work on all targets, they only generate
362efficient code on X86 so far.</li>
363
364<li>LLVM IR now supports a new <a href="LangRef.html#linkage">private
365linkage</a> type to produce labels that are stripped by the assembler before it
366produces a .o file (thus they are invisible to the linker).</li>
367
368<li>LLVM IR supports two new attributes for better alias analysis. The <a
369href="LangRef.html#paramattrs">noalias</a> attribute can now be used on the
370return value of a function to indicate that it returns new memory (e.g.
371'malloc', 'calloc', etc).</li>
372
373<li>The new <a href="LangRef.html#paramattrs">nocapture</a> attribute can be
374used on pointer arguments to functions that access through but do not return the
375pointer in a data structure that out lives the call (e.g. 'strlen', 'memcpy',
376and many others). The simplifylibcalls pass applies these attributes to
377standard libc functions.</li>
378
379<li>The parser for ".ll" files in lib/AsmParser is now completely rewritten as a
380recursive descent parser. This parser produces better error messages (including
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]381caret diagnostics), is less fragile (less likely to crash on strange things),
382does not leak memory, is more efficient, and eliminates LLVM's last use of the
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]383'bison' tool.</li>
384
385<li>Debug information representation and manipulation internals have been
386 consolidated to use a new set of classes in
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]387 <tt>llvm/Analysis/DebugInfo.h</tt>. These routines are more
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]388 efficient, robust, and extensible and replace the older mechanisms.
389 llvm-gcc, clang, and the code generator now use them to create and process
390 debug information.</li>
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]391
Chris Lattnerf304ffc2008-02-10 08:17:19 +0000[diff] [blame]392</ul>
Mikhail Glushenkovea65d7d2008-10-13 02:08:34 +0000[diff] [blame]393
Chris Lattnerf304ffc2008-02-10 08:17:19 +0000[diff] [blame]394</div>
395
396<!--=========================================================================-->
397<div class="doc_subsection">
Chris Lattneracce85d2008-02-10 07:46:44 +0000[diff] [blame]398<a name="optimizer">Optimizer Improvements</a>
399</div>
400
401<div class="doc_text">
402
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]403<p>In addition to a large array of bug fixes and minor performance tweaks, this
Chris Lattnerf3013872008-10-13 21:50:36 +0000[diff] [blame]404release includes a few major enhancements and additions to the optimizers:</p>
Chris Lattneracce85d2008-02-10 07:46:44 +0000[diff] [blame]405
406<ul>
407
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]408<li>The loop optimizer now improves floating point induction variables in
409several ways, including adding shadow induction variables to avoid
410"integer &lt;-&gt; floating point" conversions in loops when safe.</li>
411
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]412<li>The "-mem2reg" pass is now much faster on code with large basic blocks.</li>
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]413
414<li>The "-jump-threading" pass is more powerful: it is iterative
415 and handles threading based on values with fully and partially redundant
416 loads.</li>
417
418<li>The "-memdep" memory dependence analysis pass (used by GVN and memcpyopt) is
419 both faster and more aggressive.</li>
420
421<li>The "-scalarrepl" scalar replacement of aggregates pass is more aggressive
422 about promoting unions to registers.</li>
Chris Lattnerf3013872008-10-13 21:50:36 +0000[diff] [blame]423
Chris Lattner0b832202008-06-08 02:45:07 +0000[diff] [blame]424</ul>
425
426</div>
427
428<!--=========================================================================-->
429<div class="doc_subsection">
Chris Lattner511433e2009-03-02 03:24:11 +0000[diff] [blame]430<a name="codegen">Target Independent Code Generator Improvements</a>
Chris Lattner0b832202008-06-08 02:45:07 +0000[diff] [blame]431</div>
432
433<div class="doc_text">
434
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]435<p>We have put a significant amount of work into the code generator
436infrastructure, which allows us to implement more aggressive algorithms and make
437it run faster:</p>
Chris Lattner0b832202008-06-08 02:45:07 +0000[diff] [blame]438
439<ul>
Chris Lattner511433e2009-03-02 03:24:11 +0000[diff] [blame]440<li>The <a href="WritingAnLLVMBackend.html">Writing an LLVM Compiler
441Backend</a> document has been greatly expanded and is substantially more
442complete.</li>
443
444<li>The SelectionDAG type legalization logic has been completely rewritten, is
445now more powerful (it supports arbitrary precision integer types for example),
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]446and is more correct in several corner cases. The type legalizer converts
Chris Lattner511433e2009-03-02 03:24:11 +0000[diff] [blame]447operations on types that are not natively supported by the target machine into
448equivalent code sequences that only use natively supported types. The old type
449legalizer is still available (for now) and will be used if
450<tt>-disable-legalize-types</tt> is passed to the code generator.
Duncan Sands07636ea2009-02-25 11:50:08 +0000[diff] [blame]451</li>
Chris Lattner0b832202008-06-08 02:45:07 +0000[diff] [blame]452
Chris Lattner511433e2009-03-02 03:24:11 +0000[diff] [blame]453<li>The code generator now supports widening illegal vectors to larger legal
454ones (for example, converting operations on &lt;3 x float&gt; to work on
455&lt;4 x float&gt;) which is very important for common graphics
456applications.</li>
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]457
Chris Lattner511433e2009-03-02 03:24:11 +0000[diff] [blame]458<li>The assembly printers for each target are now split out into their own
459libraries that are separate from the main code generation logic. This reduces
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]460the code size of JIT compilers by not requiring them to be linked in.</li>
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]461
Chris Lattner511433e2009-03-02 03:24:11 +0000[diff] [blame]462<li>The 'fast' instruction selection path (used at -O0 and for fast JIT
463 compilers) now supports accelerating codegen for code that uses exception
464 handling constructs.</li>
465
466<li>The optional PBQP register allocator now supports register coalescing.</li>
Chris Lattner84977642007-09-21 03:54:09 +0000[diff] [blame]467</ul>
Chris Lattner84977642007-09-21 03:54:09 +0000[diff] [blame]468</div>
469
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]470<!--=========================================================================-->
471<div class="doc_subsection">
472<a name="x86">X86-32 and X86-64 Target Improvements</a>
473</div>
474
475<div class="doc_text">
476<p>New features of the X86 target include:
477</p>
478
479<ul>
Bill Wendling99dac472009-03-02 07:54:14 +0000[diff] [blame]480<li>The <tt><a href="LangRef.html#int_returnaddress">llvm.returnaddress</a></tt>
481intrinsic (which is used to implement <tt>__builtin_return_address</tt>) now
482supports non-zero stack depths on X86.</li>
Chris Lattner511433e2009-03-02 03:24:11 +0000[diff] [blame]483
484<li>The X86 backend now supports code generation of vector shift operations
485using SSE instructions.</li>
486
Bill Wendling99dac472009-03-02 07:54:14 +0000[diff] [blame]487<li>X86-64 code generation now takes advantage of red zone, unless the
488<tt>-mno-red-zone</tt> option is specified.</li>
Chris Lattner511433e2009-03-02 03:24:11 +0000[diff] [blame]489
490<li>The X86 backend now supports using address space #256 in LLVM IR as a way of
491performing memory references off the GS segment register. This allows a
492front-end to take advantage of very low-level programming techniques when
Gabor Greif29ecc1c2009-03-02 11:34:51 +0000[diff] [blame]493targeting X86 CPUs. See <tt>test/CodeGen/X86/movgs.ll</tt> for a simple
Bill Wendling99dac472009-03-02 07:54:14 +0000[diff] [blame]494example.</li>
Chris Lattner511433e2009-03-02 03:24:11 +0000[diff] [blame]495
496<li>The X86 backend now supports a <tt>-disable-mmx</tt> command line option to
497 prevent use of MMX even on chips that support it. This is important for cases
Bill Wendling99dac472009-03-02 07:54:14 +0000[diff] [blame]498 where code does not contain the proper <tt>llvm.x86.mmx.emms</tt>
499 intrinsics.</li>
Chris Lattner511433e2009-03-02 03:24:11 +0000[diff] [blame]500
Bill Wendling99dac472009-03-02 07:54:14 +0000[diff] [blame]501<li>The X86 JIT now detects the new Intel <a
502 href="http://en.wikipedia.org/wiki/Intel_Core_i7">Core i7</a> and <a
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]503 href="http://en.wikipedia.org/wiki/Intel_Atom">Atom</a> chips and
504 auto-configures itself appropriately for the features of these chips.</li>
Chris Lattner511433e2009-03-02 03:24:11 +0000[diff] [blame]505
506<li>The JIT now supports exception handling constructs on Linux/X86-64 and
507 Darwin/x86-64.</li>
Bill Wendling99dac472009-03-02 07:54:14 +0000[diff] [blame]508
Chris Lattner511433e2009-03-02 03:24:11 +0000[diff] [blame]509<li>The JIT supports Thread Local Storage (TLS) on Linux/X86-32 but not yet on
510 X86-64.</li>
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]511</ul>
512
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]513</div>
Chris Lattner84977642007-09-21 03:54:09 +0000[diff] [blame]514
515<!--=========================================================================-->
516<div class="doc_subsection">
Chris Lattner11398992009-02-26 07:32:11 +0000[diff] [blame]517<a name="pic16">PIC16 Target Improvements</a>
518</div>
519
520<div class="doc_text">
521<p>New features of the PIC16 target include:
522</p>
523
524<ul>
525<li>Both direct and indirect load/stores work now.</li>
526<li>Logical, bitwise and conditional operations now work for integer data
527types.</li>
528<li>Function calls involving basic types work now.</li>
529<li>Support for integer arrays.</li>
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]530<li>The compiler can now emit libcalls for operations not supported by m/c
Bill Wendling99dac472009-03-02 07:54:14 +0000[diff] [blame]531instructions.</li>
532<li>Support for both data and ROM address spaces.</li>
Chris Lattner11398992009-02-26 07:32:11 +0000[diff] [blame]533</ul>
534
535<p>Things not yet supported:</p>
536
537<ul>
538<li>Floating point.</li>
Bill Wendling99dac472009-03-02 07:54:14 +0000[diff] [blame]539<li>Passing/returning aggregate types to and from functions.</li>
Chris Lattner11398992009-02-26 07:32:11 +0000[diff] [blame]540<li>Variable arguments.</li>
541<li>Indirect function calls.</li>
Chris Lattner511433e2009-03-02 03:24:11 +0000[diff] [blame]542<li>Interrupts/programs.</li>
Chris Lattner11398992009-02-26 07:32:11 +0000[diff] [blame]543<li>Debug info.</li>
Chris Lattner11398992009-02-26 07:32:11 +0000[diff] [blame]544</ul>
545
546</div>
547
548
549<!--=========================================================================-->
550<div class="doc_subsection">
Chris Lattnerc441fb82009-03-01 02:30:21 +0000[diff] [blame]551<a name="llvmc">Improvements in LLVMC</a>
552</div>
553
554<div class="doc_text">
555<p>New features include:</p>
556
557<ul>
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]558<li>Beginning with LLVM 2.5, <tt>llvmc2</tt> is known as
559 just <tt>llvmc</tt>. The old <tt>llvmc</tt> driver was removed.</li>
Chris Lattnerc441fb82009-03-01 02:30:21 +0000[diff] [blame]560
561<li>The Clang plugin was substantially improved and is now enabled
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]562 by default. The command <tt>llvmc --clang</tt> can be now used as a
563 synonym to <tt>ccc</tt>.</li>
Chris Lattnerc441fb82009-03-01 02:30:21 +0000[diff] [blame]564
Bill Wendling99dac472009-03-02 07:54:14 +0000[diff] [blame]565<li>There is now a <tt>--check-graph</tt> option, which is supposed to catch
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]566 common errors like multiple default edges, mismatched output/input language
567 names and cycles. In general, these checks can't be done at compile-time
568 because of the need to support plugins.</li>
Chris Lattnerc441fb82009-03-01 02:30:21 +0000[diff] [blame]569
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]570<li>Plugins are now more flexible and can refer to compilation graph nodes and
571 options defined in other plugins. To manage dependencies, a priority-sorting
Bill Wendling99dac472009-03-02 07:54:14 +0000[diff] [blame]572 mechanism was introduced. This change affects the TableGen file syntax. See the
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]573 documentation for details.</li>
Chris Lattnerc441fb82009-03-01 02:30:21 +0000[diff] [blame]574
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]575<li>Hooks can now be provided with arguments. The syntax is "<tt>$CALL(MyHook,
Bill Wendling99dac472009-03-02 07:54:14 +0000[diff] [blame]576 'Arg1', 'Arg2', 'Arg3')</tt>".</li>
Chris Lattnerc441fb82009-03-01 02:30:21 +0000[diff] [blame]577
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]578<li>A new option type: multi-valued option, for options that take more than one
579 argument (for example, "<tt>-foo a b c</tt>").</li>
Chris Lattnerc441fb82009-03-01 02:30:21 +0000[diff] [blame]580
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]581<li>New option properties: '<tt>one_or_more</tt>', '<tt>zero_or_more</tt>',
582'<tt>hidden</tt>' and '<tt>really_hidden</tt>'.</li>
Chris Lattnerc441fb82009-03-01 02:30:21 +0000[diff] [blame]583
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]584<li>The '<tt>case</tt>' expression gained an '<tt>error</tt>' action and
585 an '<tt>empty</tt>' test (equivalent to "<tt>(not (not_empty ...))</tt>").</li>
Chris Lattnerc441fb82009-03-01 02:30:21 +0000[diff] [blame]586
587<li>Documentation now looks more consistent to the rest of the LLVM
588 docs. There is also a man page now.</li>
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]589
Chris Lattnerc441fb82009-03-01 02:30:21 +0000[diff] [blame]590</ul>
591
592</div>
593
Chris Lattner77d29b12008-06-05 08:02:49 +0000[diff] [blame]594
595<!--=========================================================================-->
596<div class="doc_subsection">
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]597<a name="changes">Major Changes and Removed Features</a>
598</div>
599
600<div class="doc_text">
601
Chris Lattnereeb4da02008-10-13 22:06:31 +0000[diff] [blame]602<p>If you're already an LLVM user or developer with out-of-tree changes based
Chris Lattner38ab04c2009-02-25 04:41:31 +0000[diff] [blame]603on LLVM 2.4, this section lists some "gotchas" that you may run into upgrading
Chris Lattnereeb4da02008-10-13 22:06:31 +0000[diff] [blame]604from the previous release.</p>
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]605
606<ul>
607
Bill Wendling99dac472009-03-02 07:54:14 +0000[diff] [blame]608<li>llvm-gcc defaults to <tt>-fno-math-errno</tt> on all X86 targets.</li>
609
Chris Lattnereeb4da02008-10-13 22:06:31 +0000[diff] [blame]610</ul>
611
612
613<p>In addition, many APIs have changed in this release. Some of the major LLVM
614API changes are:</p>
615
616<ul>
Gabor Greifcc465502009-03-02 12:23:55 +0000[diff] [blame]617<li>Some deprecated interfaces to create <tt>Instruction</tt> subclasses, that
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]618 were spelled with lower case "create," have been removed.</li>
Devang Patelb34dd132008-10-14 20:03:43 +0000[diff] [blame]619</ul>
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]620
Chris Lattnerf6662f92008-10-13 17:57:36 +0000[diff] [blame]621</div>
622
623
624
Chris Lattner19092612003-10-02 16:38:05 +0000[diff] [blame]625<!-- *********************************************************************** -->
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]626<div class="doc_section">
627 <a name="portability">Portability and Supported Platforms</a>
628</div>
Chris Lattner19092612003-10-02 16:38:05 +0000[diff] [blame]629<!-- *********************************************************************** -->
630
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]631<div class="doc_text">
632
John Criswell0b5b5e92004-12-08 20:35:47 +0000[diff] [blame]633<p>LLVM is known to work on the following platforms:</p>
Chris Lattner4654bdb2004-06-01 18:22:41 +0000[diff] [blame]634
635<ul>
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]636<li>Intel and AMD machines (IA32, X86-64, AMD64, EMT-64) running Red Hat
Chris Lattnere18b32e2008-11-10 05:40:34 +0000[diff] [blame]637Linux, Fedora Core and FreeBSD (and probably other unix-like systems).</li>
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]638<li>PowerPC and X86-based Mac OS X systems, running 10.3 and above in 32-bit
Chris Lattnere18b32e2008-11-10 05:40:34 +0000[diff] [blame]639and 64-bit modes.</li>
Chris Lattner252b83d2008-02-06 18:00:06 +0000[diff] [blame]640<li>Intel and AMD machines running on Win32 using MinGW libraries (native).</li>
Chris Lattner57a460e2007-05-23 04:39:32 +0000[diff] [blame]641<li>Intel and AMD machines running on Win32 with the Cygwin libraries (limited
642 support is available for native builds with Visual C++).</li>
Gabor Greif96a89c72008-06-05 18:39:01 +0000[diff] [blame]643<li>Sun UltraSPARC workstations running Solaris 10.</li>
John Criswell9321fa82005-05-13 20:28:15 +0000[diff] [blame]644<li>Alpha-based machines running Debian GNU/Linux.</li>
Gabor Greif96a89c72008-06-05 18:39:01 +0000[diff] [blame]645<li>Itanium-based (IA64) machines running Linux and HP-UX.</li>
Chris Lattner4654bdb2004-06-01 18:22:41 +0000[diff] [blame]646</ul>
647
Chris Lattnerbc5786b2008-06-05 06:57:39 +0000[diff] [blame]648<p>The core LLVM infrastructure uses GNU autoconf to adapt itself
Brian Gaekeb0fd7612004-05-09 05:28:35 +0000[diff] [blame]649to the machine and operating system on which it is built. However, minor
650porting may be required to get LLVM to work on new platforms. We welcome your
651portability patches and reports of successful builds or error messages.</p>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]652
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]653</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]654
655<!-- *********************************************************************** -->
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]656<div class="doc_section">
657 <a name="knownproblems">Known Problems</a>
658</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]659<!-- *********************************************************************** -->
660
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]661<div class="doc_text">
662
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]663<p>This section contains significant known problems with the LLVM system,
Chris Lattnere18b32e2008-11-10 05:40:34 +0000[diff] [blame]664listed by component. If you run into a problem, please check the <a
Chris Lattnerc463b272005-10-29 07:07:09 +0000[diff] [blame]665href="http://llvm.org/bugs/">LLVM bug database</a> and submit a bug if
Chris Lattner5eccca42003-12-12 21:22:16 +0000[diff] [blame]666there isn't already one.</p>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]667
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]668</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]669
Chris Lattnerf5ee1702004-03-14 02:03:02 +0000[diff] [blame]670<!-- ======================================================================= -->
671<div class="doc_subsection">
672 <a name="experimental">Experimental features included with this release</a>
673</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]674
Chris Lattnerf5ee1702004-03-14 02:03:02 +0000[diff] [blame]675<div class="doc_text">
676
Misha Brukman6df9e2c2004-05-12 21:46:05 +0000[diff] [blame]677<p>The following components of this LLVM release are either untested, known to
678be broken or unreliable, or are in early development. These components should
679not be relied on, and bugs should not be filed against them, but they may be
680useful 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]681components, please contact us on the <a
682href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVMdev list</a>.</p>
Chris Lattnerf5ee1702004-03-14 02:03:02 +0000[diff] [blame]683
684<ul>
Gabor Greifa8b84902008-10-14 11:31:14 +0000[diff] [blame]685<li>The MSIL, IA64, Alpha, SPU, MIPS, and PIC16 backends are experimental.</li>
Bill Wendling99dac472009-03-02 07:54:14 +0000[diff] [blame]686<li>The <tt>llc</tt> "<tt>-filetype=asm</tt>" (the default) is the only
687 supported value for this option.</li>
Chris Lattnerf5ee1702004-03-14 02:03:02 +0000[diff] [blame]688</ul>
689
690</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]691
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]692<!-- ======================================================================= -->
693<div class="doc_subsection">
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]694 <a name="x86-be">Known problems with the X86 back-end</a>
John Criswellc0c186d2005-11-08 21:11:33 +0000[diff] [blame]695</div>
696
697<div class="doc_text">
698
699<ul>
Anton Korobeynikova6094be2008-06-08 10:24:13 +0000[diff] [blame]700 <li>The X86 backend does not yet support
701 all <a href="http://llvm.org/PR879">inline assembly that uses the X86
702 floating point stack</a>. It supports the 'f' and 't' constraints, but not
703 'u'.</li>
Chris Lattnere6e1b352008-06-08 21:19:07 +0000[diff] [blame]704 <li>The X86 backend generates inefficient floating point code when configured
705 to generate code for systems that don't have SSE2.</li>
Duncan Sands47eff2b2008-06-08 19:38:43 +0000[diff] [blame]706 <li>Win64 code generation wasn't widely tested. Everything should work, but we
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]707 expect small issues to happen. Also, llvm-gcc cannot build the mingw64
708 runtime currently due
Anton Korobeynikova6094be2008-06-08 10:24:13 +0000[diff] [blame]709 to <a href="http://llvm.org/PR2255">several</a>
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]710 <a href="http://llvm.org/PR2257">bugs</a> and due to lack of support for
711 the
712 'u' inline assembly constraint and for X87 floating point inline assembly.</li>
Dan Gohman8207ba92008-06-08 23:05:11 +0000[diff] [blame]713 <li>The X86-64 backend does not yet support the LLVM IR instruction
Chris Lattner917cc712009-03-02 02:37:32 +0000[diff] [blame]714 <tt>va_arg</tt>. Currently, the llvm-gcc and front-ends support variadic
Dan Gohman8207ba92008-06-08 23:05:11 +0000[diff] [blame]715 argument constructs on X86-64 by lowering them manually.</li>
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]716</ul>
717
718</div>
719
720<!-- ======================================================================= -->
721<div class="doc_subsection">
722 <a name="ppc-be">Known problems with the PowerPC back-end</a>
723</div>
724
725<div class="doc_text">
726
727<ul>
Nicolas Geoffraye4285dc2007-05-15 09:21:28 +0000[diff] [blame]728<li>The Linux PPC32/ABI support needs testing for the interpreter and static
Chris Lattner57a460e2007-05-23 04:39:32 +0000[diff] [blame]729compilation, and lacks support for debug information.</li>
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]730</ul>
731
732</div>
733
734<!-- ======================================================================= -->
735<div class="doc_subsection">
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]736 <a name="arm-be">Known problems with the ARM back-end</a>
737</div>
738
739<div class="doc_text">
740
741<ul>
Chris Lattner57a460e2007-05-23 04:39:32 +0000[diff] [blame]742<li>Thumb mode works only on ARMv6 or higher processors. On sub-ARMv6
Duncan Sandsc90d68b2007-09-26 15:59:54 +0000[diff] [blame]743processors, thumb programs can crash or produce wrong
Chris Lattner57a460e2007-05-23 04:39:32 +0000[diff] [blame]744results (<a href="http://llvm.org/PR1388">PR1388</a>).</li>
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]745<li>Compilation for ARM Linux OABI (old ABI) is supported but not fully tested.
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]746</li>
Chris Lattnere6e1b352008-06-08 21:19:07 +0000[diff] [blame]747<li>There is a bug in QEMU-ARM (&lt;= 0.9.0) which causes it to incorrectly
748 execute
Chris Lattner57a460e2007-05-23 04:39:32 +0000[diff] [blame]749programs compiled with LLVM. Please use more recent versions of QEMU.</li>
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]750</ul>
751
752</div>
753
754<!-- ======================================================================= -->
755<div class="doc_subsection">
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]756 <a name="sparc-be">Known problems with the SPARC back-end</a>
757</div>
758
759<div class="doc_text">
760
761<ul>
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]762<li>The SPARC backend only supports the 32-bit SPARC ABI (-m32); it does not
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]763 support the 64-bit SPARC ABI (-m64).</li>
764</ul>
765
766</div>
767
768<!-- ======================================================================= -->
769<div class="doc_subsection">
Bruno Cardoso Lopesb7e1a4f2008-10-25 14:56:26 +0000[diff] [blame]770 <a name="mips-be">Known problems with the MIPS back-end</a>
771</div>
772
773<div class="doc_text">
774
775<ul>
776<li>The O32 ABI is not fully supported.</li>
777<li>64-bit MIPS targets are not supported yet.</li>
778</ul>
779
780</div>
781
782<!-- ======================================================================= -->
783<div class="doc_subsection">
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]784 <a name="alpha-be">Known problems with the Alpha back-end</a>
785</div>
786
787<div class="doc_text">
788
789<ul>
790
791<li>On 21164s, some rare FP arithmetic sequences which may trap do not have the
792appropriate nops inserted to ensure restartability.</li>
793
John Criswellc0c186d2005-11-08 21:11:33 +0000[diff] [blame]794</ul>
795</div>
796
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]797<!-- ======================================================================= -->
798<div class="doc_subsection">
799 <a name="ia64-be">Known problems with the IA64 back-end</a>
800</div>
801
802<div class="doc_text">
803
804<ul>
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]805<li>The Itanium backend is highly experimental and has a number of known
Chris Lattner04af7cb2008-06-08 23:12:47 +0000[diff] [blame]806 issues. We are looking for a maintainer for the Itanium backend. If you
Gabor Greif3bd3a262008-10-15 10:47:24 +0000[diff] [blame]807 are interested, please contact the LLVMdev mailing list.</li>
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]808</ul>
809
810</div>
811
812<!-- ======================================================================= -->
813<div class="doc_subsection">
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]814 <a name="c-be">Known problems with the C back-end</a>
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]815</div>
816
817<div class="doc_text">
818
819<ul>
Chris Lattner5733b272008-06-05 06:35:40 +0000[diff] [blame]820<li><a href="http://llvm.org/PR802">The C backend has only basic support for
821 inline assembly code</a>.</li>
Chris Lattner725a0d82007-09-26 06:01:35 +0000[diff] [blame]822<li><a href="http://llvm.org/PR1658">The C backend violates the ABI of common
823 C++ programs</a>, preventing intermixing between C++ compiled by the CBE and
Gabor Greif4906abe2009-03-02 12:02:51 +0000[diff] [blame]824 C++ code compiled with <tt>llc</tt> or native compilers.</li>
Duncan Sandsf74c0cc2008-02-10 13:40:55 +0000[diff] [blame]825<li>The C backend does not support all exception handling constructs.</li>
Duncan Sands50723a92009-02-25 11:51:54 +0000[diff] [blame]826<li>The C backend does not support arbitrary precision integers.</li>
Chris Lattner26299222006-11-18 07:51:14 +0000[diff] [blame]827</ul>
828
829</div>
John Criswellc0c186d2005-11-08 21:11:33 +0000[diff] [blame]830
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]831
832<!-- ======================================================================= -->
833<div class="doc_subsection">
Chris Lattner5733b272008-06-05 06:35:40 +0000[diff] [blame]834 <a name="c-fe">Known problems with the llvm-gcc C front-end</a>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]835</div>
Chris Lattner47588f92003-10-02 05:07:23 +0000[diff] [blame]836
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]837<div class="doc_text">
Chris Lattnerc5d658a2006-03-03 00:34:26 +0000[diff] [blame]838
Chris Lattner252b83d2008-02-06 18:00:06 +0000[diff] [blame]839<p>llvm-gcc does not currently support <a href="http://llvm.org/PR869">Link-Time
840Optimization</a> on most platforms "out-of-the-box". Please inquire on the
Gabor Greif3bd3a262008-10-15 10:47:24 +0000[diff] [blame]841LLVMdev mailing list if you are interested.</p>
Chris Lattnerd9ea0172006-08-08 17:27:28 +0000[diff] [blame]842
Chris Lattner5733b272008-06-05 06:35:40 +0000[diff] [blame]843<p>The only major language feature of GCC not supported by llvm-gcc is
844 the <tt>__builtin_apply</tt> family of builtins. However, some extensions
845 are only supported on some targets. For example, trampolines are only
Duncan Sands27aff872008-06-08 20:18:35 +0000[diff] [blame]846 supported on some targets (these are used when you take the address of a
847 nested function).</p>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]848
Chris Lattner5733b272008-06-05 06:35:40 +0000[diff] [blame]849<p>If you run into GCC extensions which are not supported, please let us know.
850</p>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]851
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]852</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]853
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]854<!-- ======================================================================= -->
855<div class="doc_subsection">
Chris Lattner5733b272008-06-05 06:35:40 +0000[diff] [blame]856 <a name="c++-fe">Known problems with the llvm-gcc C++ front-end</a>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]857</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]858
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]859<div class="doc_text">
860
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]861<p>The C++ front-end is considered to be fully
Chris Lattner7506b1d2004-12-07 08:04:13 +0000[diff] [blame]862tested and works for a number of non-trivial programs, including LLVM
Chris Lattnerf3e5bc62007-05-14 06:56:09 +0000[diff] [blame]863itself, Qt, Mozilla, etc.</p>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]864
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]865<ul>
Anton Korobeynikov0021fc12008-10-11 18:27:16 +0000[diff] [blame]866<li>Exception handling works well on the X86 and PowerPC targets. Currently
Chris Lattnere18b32e2008-11-10 05:40:34 +0000[diff] [blame]867 only Linux and Darwin targets are supported (both 32 and 64 bit).</li>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]868</ul>
Chris Lattnerfcc54b32003-10-07 22:14:37 +0000[diff] [blame]869
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]870</div>
871
Chris Lattner1eb4df62008-10-30 03:58:13 +0000[diff] [blame]872<!-- ======================================================================= -->
873<div class="doc_subsection">
874 <a name="fortran-fe">Known problems with the llvm-gcc Fortran front-end</a>
875</div>
876
877<div class="doc_text">
Gabor Greifba10fe02008-11-04 21:50:59 +0000[diff] [blame]878<ul>
Chris Lattner1eb4df62008-10-30 03:58:13 +0000[diff] [blame]879<li>Fortran support generally works, but there are still several unresolved bugs
880 in Bugzilla. Please see the tools/gfortran component for details.</li>
Mikhail Glushenkovf795ef02009-03-01 18:09:47 +0000[diff] [blame]881
Chris Lattner1eb4df62008-10-30 03:58:13 +0000[diff] [blame]882<li>The Fortran front-end currently does not build on Darwin (without tweaks)
883 due to unresolved dependencies on the C front-end.</li>
Gabor Greifba10fe02008-11-04 21:50:59 +0000[diff] [blame]884</ul>
Chris Lattner1eb4df62008-10-30 03:58:13 +0000[diff] [blame]885</div>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]886
Chris Lattner2b659ef2008-02-12 06:29:45 +0000[diff] [blame]887<!-- ======================================================================= -->
888<div class="doc_subsection">
Chris Lattner5733b272008-06-05 06:35:40 +0000[diff] [blame]889 <a name="ada-fe">Known problems with the llvm-gcc Ada front-end</a>
Chris Lattner2b659ef2008-02-12 06:29:45 +0000[diff] [blame]890</div>
891
892<div class="doc_text">
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]893The llvm-gcc 4.2 Ada compiler works fairly well; however, this is not a mature
894technology, and problems should be expected.
Chris Lattner2b659ef2008-02-12 06:29:45 +0000[diff] [blame]895<ul>
Duncan Sands27aff872008-06-08 20:18:35 +0000[diff] [blame]896<li>The Ada front-end currently only builds on X86-32. This is mainly due
John Criswell524a5dd2009-03-02 15:28:15 +0000[diff] [blame^]897to lack of trampoline support (pointers to nested functions) on other platforms.
898However, 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]899which does support trampolines.</li>
900<li>The Ada front-end <a href="http://llvm.org/PR2007">fails to bootstrap</a>.
Duncan Sands326a4982009-02-25 11:59:06 +0000[diff] [blame]901This is due to lack of LLVM support for <tt>setjmp</tt>/<tt>longjmp</tt> style
902exception handling, which is used internally by the compiler.
Chris Lattner2b659ef2008-02-12 06:29:45 +0000[diff] [blame]903Workaround: configure with --disable-bootstrap.</li>
Duncan Sands978bcee2008-10-13 17:27:23 +0000[diff] [blame]904<li>The c380004, <a href="http://llvm.org/PR2010">c393010</a>
905and <a href="http://llvm.org/PR2421">cxg2021</a> ACATS tests fail
Duncan Sands326a4982009-02-25 11:59:06 +0000[diff] [blame]906(c380004 also fails with gcc-4.2 mainline).
907If the compiler is built with checks disabled then <a href="http://llvm.org/PR2010">c393010</a>
908causes the compiler to go into an infinite loop, using up all system memory.</li>
Duncan Sands978bcee2008-10-13 17:27:23 +0000[diff] [blame]909<li>Some gcc specific Ada tests continue to crash the compiler.</li>
Chris Lattner2b659ef2008-02-12 06:29:45 +0000[diff] [blame]910<li>The -E binder option (exception backtraces)
911<a href="http://llvm.org/PR1982">does not work</a> and will result in programs
912crashing if an exception is raised. Workaround: do not use -E.</li>
913<li>Only discrete types <a href="http://llvm.org/PR1981">are allowed to start
914or finish at a non-byte offset</a> in a record. Workaround: do not pack records
915or use representation clauses that result in a field of a non-discrete type
916starting or finishing in the middle of a byte.</li>
Chris Lattnere6e1b352008-06-08 21:19:07 +0000[diff] [blame]917<li>The <tt>lli</tt> interpreter <a href="http://llvm.org/PR2009">considers
918'main' as generated by the Ada binder to be invalid</a>.
919Workaround: hand edit the file to use pointers for <tt>argv</tt> and
920<tt>envp</tt> rather than integers.</li>
921<li>The <tt>-fstack-check</tt> option <a href="http://llvm.org/PR2008">is
922ignored</a>.</li>
Chris Lattner2b659ef2008-02-12 06:29:45 +0000[diff] [blame]923</ul>
924</div>
925
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]926<!-- *********************************************************************** -->
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]927<div class="doc_section">
928 <a name="additionalinfo">Additional Information</a>
929</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]930<!-- *********************************************************************** -->
931
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]932<div class="doc_text">
933
Chris Lattner416db102005-05-16 17:13:10 +0000[diff] [blame]934<p>A wide variety of additional information is available on the <a
Chris Lattnerb4b0ce72007-05-18 00:44:29 +0000[diff] [blame]935href="http://llvm.org">LLVM web page</a>, in particular in the <a
936href="http://llvm.org/docs/">documentation</a> section. The web page also
Reid Spencer669ed452007-07-09 08:04:31 +0000[diff] [blame]937contains versions of the API documentation which is up-to-date with the
938Subversion version of the source code.
Misha Brukman109d9e82005-03-30 19:14:24 +0000[diff] [blame]939You can access versions of these documents specific to this release by going
940into the "<tt>llvm/doc/</tt>" directory in the LLVM tree.</p>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]941
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]942<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]943us via the <a href="http://llvm.org/docs/#maillist"> mailing
Chris Lattner5eccca42003-12-12 21:22:16 +0000[diff] [blame]944lists</a>.</p>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]945
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]946</div>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]947
948<!-- *********************************************************************** -->
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]949
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]950<hr>
Misha Brukman2061e892003-11-22 01:23:39 +0000[diff] [blame]951<address>
Misha Brukman38847d52003-12-21 22:53:21 +0000[diff] [blame]952 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]953 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukman2061e892003-11-22 01:23:39 +0000[diff] [blame]954 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman44408702008-12-11 17:34:48 +0000[diff] [blame]955 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Misha Brukman2061e892003-11-22 01:23:39 +0000[diff] [blame]956
Chris Lattnerb4b0ce72007-05-18 00:44:29 +0000[diff] [blame]957 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]958 Last modified: $Date$
Misha Brukman2061e892003-11-22 01:23:39 +0000[diff] [blame]959</address>
Chris Lattner79c3fe12003-10-02 04:57:28 +0000[diff] [blame]960
Misha Brukman500bc302003-11-22 00:38:41 +0000[diff] [blame]961</body>
962</html>