Gitiles
Code Review Sign In
gerrit-public.fairphone.software / toolchain / llvm-project / a3cc83bc8203a5f299f654a51ec6931ad10b50e7 / . / llvm / docs / ReleaseNotes.html
blob: d17d5e78fd818f6b755e46920d9e048f6450671d [file] [log] [blame]
Misha Brukman80731b92003-11-22 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 Spencere4190fa2004-11-18 18:38:58 +0000[diff] [blame]5 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]6 <link rel="stylesheet" href="llvm.css" type="text/css">
Jeffrey Yasskin0830b972010-01-28 01:14:43 +0000[diff] [blame]7 <title>LLVM 2.7 Release Notes</title>
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]8</head>
9<body>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]10
Jeffrey Yasskin0830b972010-01-28 01:14:43 +0000[diff] [blame]11<div class="doc_title">LLVM 2.7 Release Notes</div>
Mikhail Glushenkov024f7cf2008-10-13 02:08:34 +0000[diff] [blame]12
Chris Lattnerc871bac2010-03-17 04:02:39 +0000[diff] [blame]13<img align=right src="http://llvm.org/img/DragonSmall.png"
Gabor Greif27b166352010-04-22 10:21:43 +0000[diff] [blame]14 width="136" height="136" alt="LLVM Dragon Logo">
Chris Lattnerc871bac2010-03-17 04:02:39 +0000[diff] [blame]15
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]16<ol>
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]17 <li><a href="#intro">Introduction</a></li>
Chris Lattnerf5cd9862008-10-13 18:01:01 +0000[diff] [blame]18 <li><a href="#subproj">Sub-project Status Update</a></li>
Jeffrey Yasskin0830b972010-01-28 01:14:43 +0000[diff] [blame]19 <li><a href="#externalproj">External Projects Using LLVM 2.7</a></li>
20 <li><a href="#whatsnew">What's New in LLVM 2.7?</a></li>
Chris Lattner77a51732004-04-30 22:17:12 +0000[diff] [blame]21 <li><a href="GettingStarted.html">Installation Instructions</a></li>
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]22 <li><a href="#portability">Portability and Supported Platforms</a></li>
Dan Gohmanad888912008-10-14 16:23:02 +0000[diff] [blame]23 <li><a href="#knownproblems">Known Problems</a></li>
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]24 <li><a href="#additionalinfo">Additional Information</a></li>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]25</ol>
26
Chris Lattner020e1fc2004-05-23 21:07:27 +0000[diff] [blame]27<div class="doc_author">
Dan Gohmanad888912008-10-14 16:23:02 +0000[diff] [blame]28 <p>Written by the <a href="http://llvm.org">LLVM Team</a></p>
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]29</div>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]30
Chris Lattner06971182010-04-13 06:37:00 +0000[diff] [blame]31<!--
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]32<h1 style="color:red">These are in-progress notes for the upcoming LLVM 2.8
Jeffrey Yasskin0830b972010-01-28 01:14:43 +0000[diff] [blame]33release.<br>
34You may prefer the
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]35<a href="http://llvm.org/releases/2.6/docs/ReleaseNotes.html">LLVM 2.7
Chris Lattner06971182010-04-13 06:37:00 +0000[diff] [blame]36Release Notes</a>.</h1>-->
Jeffrey Yasskin0830b972010-01-28 01:14:43 +0000[diff] [blame]37
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]38<!-- *********************************************************************** -->
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]39<div class="doc_section">
40 <a name="intro">Introduction</a>
41</div>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]42<!-- *********************************************************************** -->
43
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]44<div class="doc_text">
45
Chris Lattner1e4d5bc2008-10-13 17:57:36 +0000[diff] [blame]46<p>This document contains the release notes for the LLVM Compiler
Jeffrey Yasskin0830b972010-01-28 01:14:43 +0000[diff] [blame]47Infrastructure, release 2.7. Here we describe the status of LLVM, including
Chris Lattner1e4d5bc2008-10-13 17:57:36 +0000[diff] [blame]48major improvements from the previous release and significant known problems.
Mikhail Glushenkov25422542009-03-01 18:09:47 +0000[diff] [blame]49All LLVM releases may be downloaded from the <a
Chris Lattner1e4d5bc2008-10-13 17:57:36 +0000[diff] [blame]50href="http://llvm.org/releases/">LLVM releases web site</a>.</p>
Chris Lattner62495762003-10-02 16:38:05 +0000[diff] [blame]51
Chris Lattnerb5bb5972004-12-07 08:04:13 +0000[diff] [blame]52<p>For more information about LLVM, including information about the latest
Chris Lattnera69595e2005-10-29 07:07:09 +0000[diff] [blame]53release, please check out the <a href="http://llvm.org/">main LLVM
Chris Lattnere7525b52003-10-07 21:38:31 +0000[diff] [blame]54web site</a>. If you have questions or comments, the <a
Chris Lattner0b1c9a52010-03-17 04:41:49 +0000[diff] [blame]55href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVM Developer's
56Mailing List</a> is a good place to send them.</p>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]57
Mikhail Glushenkov024f7cf2008-10-13 02:08:34 +0000[diff] [blame]58<p>Note that if you are reading this file from a Subversion checkout or the
Chris Lattner0a1fd102007-09-21 03:54:09 +0000[diff] [blame]59main LLVM web page, this document applies to the <i>next</i> release, not the
Gabor Greif355f81c2008-10-14 11:00:32 +0000[diff] [blame]60current one. To see the release notes for a specific release, please see the
Chris Lattner0a1fd102007-09-21 03:54:09 +0000[diff] [blame]61<a href="http://llvm.org/releases/">releases page</a>.</p>
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]62
63</div>
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]64
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]65
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]66<!--
67Almost dead code.
68 include/llvm/Analysis/LiveValues.h => Dan
69 lib/Transforms/IPO/MergeFunctions.cpp => consider for 2.8.
70 llvm/Analysis/PointerTracking.h => Edwin wants this, consider for 2.8.
Chris Lattner06971182010-04-13 06:37:00 +0000[diff] [blame]71 ABCD, GEPSplitterPass
Chris Lattner89ea59d2010-01-16 21:25:13 +0000[diff] [blame]72 MSIL backend?
Chris Lattner10a3cc42010-03-17 06:41:58 +0000[diff] [blame]73 lib/Transforms/Utils/SSI.cpp -> ABCD depends on it.
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]74-->
75
76
Jeffrey Yasskin0830b972010-01-28 01:14:43 +0000[diff] [blame]77<!-- Features that need text if they're finished for 2.7:
Chris Lattnera67df2d2010-04-22 06:28:20 +0000[diff] [blame]78 combiner-aa?
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]79 strong phi elim
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]80 llvm.dbg.value: variable debug info for optimized code
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]81 loop dependence analysis
Chris Lattner2b8a52e2008-02-10 07:46:44 +0000[diff] [blame]82 -->
Mikhail Glushenkov024f7cf2008-10-13 02:08:34 +0000[diff] [blame]83
Chris Lattnerc4d68542008-10-12 19:47:48 +0000[diff] [blame]84 <!-- for announcement email:
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]85 Logo web page.
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]86 Many new papers added to /pubs/
Chris Lattnerd02da042009-02-25 06:34:50 +0000[diff] [blame]87 -->
Chris Lattner2b8a52e2008-02-10 07:46:44 +0000[diff] [blame]88
Chris Lattnerf5cd9862008-10-13 18:01:01 +0000[diff] [blame]89<!-- *********************************************************************** -->
90<div class="doc_section">
91 <a name="subproj">Sub-project Status Update</a>
Chris Lattner625a3d82008-06-08 21:34:41 +0000[diff] [blame]92</div>
Chris Lattnerf5cd9862008-10-13 18:01:01 +0000[diff] [blame]93<!-- *********************************************************************** -->
Chris Lattner625a3d82008-06-08 21:34:41 +0000[diff] [blame]94
95<div class="doc_text">
Chris Lattnerc75fd522008-06-08 21:58:17 +0000[diff] [blame]96<p>
Jeffrey Yasskin0830b972010-01-28 01:14:43 +0000[diff] [blame]97The LLVM 2.7 distribution currently consists of code from the core LLVM
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]98repository (which roughly includes the LLVM optimizers, code generators
99and supporting tools), the Clang repository and the llvm-gcc repository. In
100addition to this code, the LLVM Project includes other sub-projects that are in
101development. Here we include updates on these subprojects.
Bill Wendlingf170d2e2009-03-02 04:28:57 +0000[diff] [blame]102</p>
Chris Lattnerf5cd9862008-10-13 18:01:01 +0000[diff] [blame]103
Chris Lattnerc75fd522008-06-08 21:58:17 +0000[diff] [blame]104</div>
105
Chris Lattnerf5cd9862008-10-13 18:01:01 +0000[diff] [blame]106
Chris Lattnerc75fd522008-06-08 21:58:17 +0000[diff] [blame]107<!--=========================================================================-->
Chris Lattnerf5cd9862008-10-13 18:01:01 +0000[diff] [blame]108<div class="doc_subsection">
Chris Lattner44c09cd2008-10-13 18:11:54 +0000[diff] [blame]109<a name="clang">Clang: C/C++/Objective-C Frontend Toolkit</a>
Chris Lattnerc75fd522008-06-08 21:58:17 +0000[diff] [blame]110</div>
111
112<div class="doc_text">
113
Chris Lattner5de7f6e2010-04-26 17:42:18 +0000[diff] [blame]114<p><a href="http://clang.llvm.org/">Clang</a> is an LLVM front end for the C,
115C++, and Objective-C languages. Clang aims to provide a better user experience
116through expressive diagnostics, a high level of conformance to language
117standards, fast compilation, and low memory use. Like LLVM, Clang provides a
118modular, library-based architecture that makes it suitable for creating or
119integrating with other development tools. Clang is considered a
120production-quality compiler for C and Objective-C on x86 (32- and 64-bit).</p>
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]121
Jeffrey Yasskin0830b972010-01-28 01:14:43 +0000[diff] [blame]122<p>In the LLVM 2.7 time-frame, the Clang team has made many improvements:</p>
Bill Wendlingef362462008-10-27 09:27:33 +0000[diff] [blame]123
Daniel Dunbarc11b40f2008-10-14 23:25:09 +0000[diff] [blame]124<ul>
Douglas Gregorc5b5c052010-04-22 20:42:40 +0000[diff] [blame]125
Chris Lattner5de7f6e2010-04-26 17:42:18 +0000[diff] [blame]126<li>C++ Support: Clang is now capable of self-hosting! While still
127alpha-quality, Clang's C++ support has matured enough to build LLVM and Clang,
128and C++ is now enabled by default. See the <a
129href="http://clang.llvm.org/cxx_compatibility.html">Clang C++ compatibility
130page</a> for common C++ migration issues.</li>
Douglas Gregorc5b5c052010-04-22 20:42:40 +0000[diff] [blame]131
Chris Lattner5de7f6e2010-04-26 17:42:18 +0000[diff] [blame]132<li>Objective-C: Clang now includes experimental support for an updated
133Objective-C ABI on non-Darwin platforms. This includes support for non-fragile
134instance variables and accelerated proxies, as well as greater potential for
135future optimisations. The new ABI is used when compiling with the
136-fobjc-nonfragile-abi and -fgnu-runtime options. Code compiled with these
137options may be mixed with code compiled with GCC or clang using the old GNU ABI,
138but requires the libobjc2 runtime from the GNUstep project.</li>
David Chisnall17dff472010-04-25 19:13:33 +0000[diff] [blame]139
Chris Lattner5de7f6e2010-04-26 17:42:18 +0000[diff] [blame]140<li>New warnings: Clang contains a number of new warnings, including
141control-flow warnings (unreachable code, missing return statements in a
142non-<code>void</code> function, etc.), sign-comparison warnings, and improved
143format-string warnings.</li>
Daniel Dunbareb4bc7f2010-03-25 16:09:18 +0000[diff] [blame]144
Daniel Dunbareb4bc7f2010-03-25 16:09:18 +0000[diff] [blame]145<li>CIndex API and Python bindings: Clang now includes a C API as part of the
Gabor Greifc1670312010-04-25 21:27:54 +0000[diff] [blame]146CIndex library. Although we may make some changes to the API in the future, it
Daniel Dunbareb4bc7f2010-03-25 16:09:18 +0000[diff] [blame]147is intended to be stable and has been designed for use by external projects. See
148the Clang
149doxygen <a href="http://clang.llvm.org/doxygen/group__CINDEX.html">CIndex</a>
Wesley Peck9bc5e8f2010-04-22 14:19:00 +0000[diff] [blame]150documentation for more details. The CIndex API also includes a preliminary
Daniel Dunbareb4bc7f2010-03-25 16:09:18 +0000[diff] [blame]151set of Python bindings.</li>
152
153<li>ARM Support: Clang now has ABI support for both the Darwin and Linux ARM
154ABIs. Coupled with many improvements to the LLVM ARM backend, Clang is now
Gabor Greifc1670312010-04-25 21:27:54 +0000[diff] [blame]155suitable for use as a beta quality ARM compiler.</li>
David Chisnall17dff472010-04-25 19:13:33 +0000[diff] [blame]156
Bill Wendling87adbe42009-03-02 04:28:18 +0000[diff] [blame]157</ul>
Chris Lattner44c09cd2008-10-13 18:11:54 +0000[diff] [blame]158</div>
159
160<!--=========================================================================-->
161<div class="doc_subsection">
162<a name="clangsa">Clang Static Analyzer</a>
163</div>
164
165<div class="doc_text">
166
Chris Lattner2284b6a2010-04-22 06:38:11 +0000[diff] [blame]167<p>The <a href="http://clang-analyzer.llvm.org/">Clang Static Analyzer</a>
168 project is an effort to use static source code analysis techniques to
169 automatically find bugs in C and Objective-C programs (and hopefully <a
170 href="http://clang-analyzer.llvm.org/dev_cxx.html">C++ in the
171 future</a>!). The tool is very good at finding bugs that occur on specific
172 paths through code, such as on error conditions.</p>
Chris Lattnerbcffa5a2008-10-14 00:52:49 +0000[diff] [blame]173
Chris Lattner2284b6a2010-04-22 06:38:11 +0000[diff] [blame]174<p>In the LLVM 2.7 time-frame, the analyzer core has made several major and
175 minor improvements, including better support for tracking the fields of
176 structures, initial support (not enabled by default yet) for doing
177 interprocedural (cross-function) analysis, and new checks have been added.
178</p>
Chris Lattnerc75fd522008-06-08 21:58:17 +0000[diff] [blame]179
180</div>
181
Chris Lattnerf5cd9862008-10-13 18:01:01 +0000[diff] [blame]182<!--=========================================================================-->
183<div class="doc_subsection">
Nicolas Geoffraye5ec9322008-10-14 19:23:04 +0000[diff] [blame]184<a name="vmkit">VMKit: JVM/CLI Virtual Machine Implementation</a>
Chris Lattnerf5cd9862008-10-13 18:01:01 +0000[diff] [blame]185</div>
186
187<div class="doc_text">
188<p>
Nicolas Geoffraye5ec9322008-10-14 19:23:04 +0000[diff] [blame]189The <a href="http://vmkit.llvm.org/">VMKit project</a> is an implementation of
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]190a JVM and a CLI Virtual Machine (Microsoft .NET is an
191implementation of the CLI) using LLVM for static and just-in-time
192compilation.</p>
Chris Lattnerf5cd9862008-10-13 18:01:01 +0000[diff] [blame]193
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]194<p>
Chris Lattneref2e5732010-03-29 18:34:13 +0000[diff] [blame]195With the release of LLVM 2.7, VMKit has shifted to a great framework for writing
196virtual machines. VMKit now offers precise and efficient garbage collection with
197multi-threading support, thanks to the MMTk memory management toolkit, as well
198as just in time and ahead of time compilation with LLVM. The major changes in
199VMKit 0.27 are:</p>
Chris Lattnerf5cd9862008-10-13 18:01:01 +0000[diff] [blame]200
Nicolas Geoffraye5ec9322008-10-14 19:23:04 +0000[diff] [blame]201<ul>
202
Chris Lattneref2e5732010-03-29 18:34:13 +0000[diff] [blame]203<li>Garbage collection: VMKit now uses the MMTk toolkit for garbage collectors.
204 The first collector to be ported is the MarkSweep collector, which is precise,
205 and drastically improves the performance of VMKit.</li>
206<li>Line number information in the JVM: by using the debug metadata of LLVM, the
207 JVM now supports precise line number information, useful when printing a stack
208 trace.</li>
209<li>Interface calls in the JVM: we implemented a variant of the Interface Method
210 Table technique for interface calls in the JVM.
211</li>
Nicolas Geoffraye5ec9322008-10-14 19:23:04 +0000[diff] [blame]212
213</ul>
Chris Lattnerf5cd9862008-10-13 18:01:01 +0000[diff] [blame]214</div>
215
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]216
217<!--=========================================================================-->
218<div class="doc_subsection">
219<a name="compiler-rt">compiler-rt: Compiler Runtime Library</a>
220</div>
221
222<div class="doc_text">
223<p>
224The new LLVM <a href="http://compiler-rt.llvm.org/">compiler-rt project</a>
225is a simple library that provides an implementation of the low-level
226target-specific hooks required by code generation and other runtime components.
227For example, when compiling for a 32-bit target, converting a double to a 64-bit
228unsigned integer is compiled into a runtime call to the "__fixunsdfdi"
229function. The compiler-rt library provides highly optimized implementations of
230this and other low-level routines (some are 3x faster than the equivalent
231libgcc routines).</p>
232
233<p>
234All of the code in the compiler-rt project is available under the standard LLVM
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]235License, a "BSD-style" license. New in LLVM 2.7: compiler_rt now
236supports ARM targets.</p>
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]237
238</div>
239
240<!--=========================================================================-->
241<div class="doc_subsection">
Duncan Sands92452b92010-04-02 09:23:15 +0000[diff] [blame]242<a name="dragonegg">DragonEgg: llvm-gcc ported to gcc-4.5</a>
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]243</div>
244
245<div class="doc_text">
246<p>
Duncan Sands92452b92010-04-02 09:23:15 +0000[diff] [blame]247<a href="http://dragonegg.llvm.org/">DragonEgg</a> is a port of llvm-gcc to
248gcc-4.5. Unlike llvm-gcc, which makes many intrusive changes to the underlying
249gcc-4.2 code, dragonegg in theory does not require any gcc-4.5 modifications
250whatsoever (currently one small patch is needed). This is thanks to the new
251<a href="http://gcc.gnu.org/wiki/plugins">gcc plugin architecture</a>, which
252makes it possible to modify the behaviour of gcc at runtime by loading a plugin,
253which is nothing more than a dynamic library which conforms to the gcc plugin
254interface. DragonEgg is a gcc plugin that causes the LLVM optimizers to be run
255instead of the gcc optimizers, and the LLVM code generators instead of the gcc
256code generators, just like llvm-gcc. To use it, you add
257"-fplugin=path/dragonegg.so" to the gcc-4.5 command line, and gcc-4.5 magically
258becomes llvm-gcc-4.5!
259</p>
260
261<p>
262DragonEgg is still a work in progress. Currently C works very well, while C++,
263Ada and Fortran work fairly well. All other languages either don't work at all,
264or only work poorly. For the moment only the x86-32 and x86-64 targets are
Duncan Sands4dd125e2010-04-20 19:40:58 +0000[diff] [blame]265supported, and only on linux and darwin (darwin needs an additional gcc patch).
Duncan Sands92452b92010-04-02 09:23:15 +0000[diff] [blame]266</p>
267
268<p>
Duncan Sandsfe291172010-04-21 13:51:48 +0000[diff] [blame]269DragonEgg is a new project which is seeing its first release with llvm-2.7.
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]270</p>
271
272</div>
273
274
275<!--=========================================================================-->
276<div class="doc_subsection">
277<a name="mc">llvm-mc: Machine Code Toolkit</a>
278</div>
279
280<div class="doc_text">
281<p>
Chris Lattner2284b6a2010-04-22 06:38:11 +0000[diff] [blame]282The LLVM Machine Code (aka MC) sub-project of LLVM was created to solve a number
283of problems in the realm of assembly, disassembly, object file format handling,
284and a number of other related areas that CPU instruction-set level tools work
285in. It is a sub-project of LLVM which provides it with a number of advantages
286over other compilers that do not have tightly integrated assembly-level tools.
287For a gentle introduction, please see the <a
288href="http://blog.llvm.org/2010/04/intro-to-llvm-mc-project.html">Intro to the
289LLVM MC Project Blog Post</a>.
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]290</p>
291
Chris Lattner2284b6a2010-04-22 06:38:11 +0000[diff] [blame]292<p>2.7 includes major parts of the work required by the new MC Project. A few
293 targets have been refactored to support it, and work is underway to support a
Gabor Greif60084b12010-04-25 21:30:22 +0000[diff] [blame]294 native assembler in LLVM. This work is not complete in LLVM 2.7, but it has
Chris Lattner2284b6a2010-04-22 06:38:11 +0000[diff] [blame]295 made substantially more progress on LLVM mainline.</p>
296
297<p>One minor example of what MC can do is to transcode an AT&amp;T syntax
Gabor Greifd80517c2010-04-22 10:25:23 +0000[diff] [blame]298 X86 .s file into intel syntax. You can do this with something like:</p>
Chris Lattner2284b6a2010-04-22 06:38:11 +0000[diff] [blame]299<pre>
300 llvm-mc foo.s -output-asm-variant=1 -o foo-intel.s
301</pre>
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]302
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]303</div>
304
305
Chris Lattner53e06f92009-02-26 22:33:38 +0000[diff] [blame]306<!-- *********************************************************************** -->
307<div class="doc_section">
Jeffrey Yasskin0830b972010-01-28 01:14:43 +0000[diff] [blame]308 <a name="externalproj">External Open Source Projects Using LLVM 2.7</a>
Chris Lattner53e06f92009-02-26 22:33:38 +0000[diff] [blame]309</div>
310<!-- *********************************************************************** -->
311
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]312<div class="doc_text">
313
314<p>An exciting aspect of LLVM is that it is used as an enabling technology for
315 a lot of other language and tools projects. This section lists some of the
Jeffrey Yasskin0830b972010-01-28 01:14:43 +0000[diff] [blame]316 projects that have already been updated to work with LLVM 2.7.</p>
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]317</div>
318
Chris Lattner53e06f92009-02-26 22:33:38 +0000[diff] [blame]319<!--=========================================================================-->
320<div class="doc_subsection">
321<a name="pure">Pure</a>
322</div>
323
324<div class="doc_text">
325<p>
Chris Lattnerd1094e02009-03-02 02:37:32 +0000[diff] [blame]326<a href="http://pure-lang.googlecode.com/">Pure</a>
327is an algebraic/functional programming language based on term rewriting.
Chris Lattner53e06f92009-02-26 22:33:38 +0000[diff] [blame]328Programs are collections of equations which are used to evaluate expressions in
329a symbolic fashion. Pure offers dynamic typing, eager and lazy evaluation,
Mikhail Glushenkov25422542009-03-01 18:09:47 +0000[diff] [blame]330lexical closures, a hygienic macro system (also based on term rewriting),
Chris Lattner53e06f92009-02-26 22:33:38 +0000[diff] [blame]331built-in list and matrix support (including list and matrix comprehensions) and
332an easy-to-use C interface. The interpreter uses LLVM as a backend to
333 JIT-compile Pure programs to fast native code.</p>
334
Chris Lattner6aba4862010-03-17 17:25:49 +0000[diff] [blame]335<p>Pure versions 0.43 and later have been tested and are known to work with
336LLVM 2.7 (and continue to work with older LLVM releases >= 2.5).</p>
337
Chris Lattner53e06f92009-02-26 22:33:38 +0000[diff] [blame]338</div>
339
Chris Lattnerfb23a962009-02-28 18:58:01 +0000[diff] [blame]340<!--=========================================================================-->
341<div class="doc_subsection">
342<a name="RoadsendPHP">Roadsend PHP</a>
343</div>
344
345<div class="doc_text">
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]346<p>
347<a href="http://code.roadsend.com/rphp">Roadsend PHP</a> (rphp) is an open
Chris Lattner0c72d8f2009-03-02 19:07:24 +0000[diff] [blame]348source implementation of the PHP programming
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]349language that uses LLVM for its optimizer, JIT and static compiler. This is a
Chris Lattnerc871bac2010-03-17 04:02:39 +0000[diff] [blame]350reimplementation of an earlier project that is now based on LLVM.
Chris Lattner6aba4862010-03-17 17:25:49 +0000[diff] [blame]351</p>
Chris Lattnerfb23a962009-02-28 18:58:01 +0000[diff] [blame]352</div>
Chris Lattner53e06f92009-02-26 22:33:38 +0000[diff] [blame]353
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]354<!--=========================================================================-->
355<div class="doc_subsection">
356<a name="UnladenSwallow">Unladen Swallow</a>
357</div>
358
359<div class="doc_text">
360<p>
361<a href="http://code.google.com/p/unladen-swallow/">Unladen Swallow</a> is a
362branch of <a href="http://python.org/">Python</a> intended to be fully
363compatible and significantly faster. It uses LLVM's optimization passes and JIT
Chris Lattner5b59a732010-03-18 06:52:15 +0000[diff] [blame]364compiler.
Chris Lattnerc871bac2010-03-17 04:02:39 +0000[diff] [blame]365</p>
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]366</div>
367
368<!--=========================================================================-->
369<div class="doc_subsection">
Chris Lattnerb4a02792010-03-29 17:50:39 +0000[diff] [blame]370<a name="tce">TTA-based Codesign Environment (TCE)</a>
371</div>
372
373<div class="doc_text">
374<p>
375<a href="http://tce.cs.tut.fi/">TCE</a> is a toolset for designing
376application-specific processors (ASP) based on the Transport triggered
377architecture (TTA). The toolset provides a complete co-design flow from C/C++
378programs down to synthesizable VHDL and parallel program binaries. Processor
379customization points include the register files, function units, supported
380operations, and the interconnection network.</p>
381
382<p>TCE uses llvm-gcc/Clang and LLVM for C/C++ language support, target
383independent optimizations and also for parts of code generation. It generates
384new LLVM-based code generators "on the fly" for the designed TTA processors and
385loads them in to the compiler backend as runtime libraries to avoid per-target
386recompilation of larger parts of the compiler chain.</p>
387
388</div>
Chris Lattnerc75fd522008-06-08 21:58:17 +0000[diff] [blame]389
John Criswell9003e0f2010-04-06 14:52:14 +0000[diff] [blame]390<!--=========================================================================-->
391<div class="doc_subsection">
392<a name="safecode">SAFECode Compiler</a>
393</div>
394
395<div class="doc_text">
396<p>
397<a href="http://safecode.cs.illinois.edu">SAFECode</a> is a memory safe C
398compiler built using LLVM. It takes standard, unannotated C code, analyzes the
399code to ensure that memory accesses and array indexing operations are safe, and
400instruments the code with run-time checks when safety cannot be proven
401statically.
402</p>
403</div>
404
Chris Lattnerad280412010-04-22 17:28:36 +0000[diff] [blame]405<!--=========================================================================-->
406<div class="doc_subsection">
407<a name="icedtea">IcedTea Java Virtual Machine Implementation</a>
408</div>
John Criswell9003e0f2010-04-06 14:52:14 +0000[diff] [blame]409
Chris Lattnerad280412010-04-22 17:28:36 +0000[diff] [blame]410<div class="doc_text">
411<p>
412<a href="http://icedtea.classpath.org/wiki/Main_Page">IcedTea</a> provides a
413harness to build OpenJDK using only free software build tools and to provide
414replacements for the not-yet free parts of OpenJDK. One of the extensions that
415IcedTea provides is a new JIT compiler named <a
416href="http://icedtea.classpath.org/wiki/ZeroSharkFaq">Shark</a> which uses LLVM
417to provide native code generation without introducing processor-dependent
418code.
419</p>
420<p>Icedtea6 1.8 and later have been tested and are known to work with
421LLVM 2.7 (and continue to work with older LLVM releases >= 2.6 as well).
422</p>
423</div>
424
425<!--=========================================================================-->
426<div class="doc_subsection">
427<a name="llvm-lua">LLVM-Lua</a>
428</div>
429
430<div class="doc_text">
431<p>
432<a href="http://code.google.com/p/llvm-lua/">LLVM-Lua</a> uses LLVM
433 to add JIT and static compiling support to the Lua VM. Lua
434bytecode is analyzed to remove type checks, then LLVM is used to compile the
435bytecode down to machine code.
436</p>
437<p>LLVM-Lua 1.2.0 have been tested and is known to work with LLVM 2.7.
438</p>
439</div>
Chris Lattner64694df2010-04-22 21:34:16 +0000[diff] [blame]440
441<!--=========================================================================-->
442<div class="doc_subsection">
443<a name="MacRuby">MacRuby</a>
444</div>
445
446<div class="doc_text">
447<p>
448<a href="http://macruby.org">MacRuby</a> is an implementation of Ruby based on
449core Mac OS technologies, sponsored by Apple Inc. It uses LLVM at runtime for
450optimization passes, JIT compilation and exception handling. It also allows
451static (ahead-of-time) compilation of Ruby code straight to machine code.
452</p>
453<p>The upcoming MacRuby 0.6 release works with LLVM 2.7.
454</p>
455</div>
456
Chris Lattnerdf85c892010-04-26 17:38:10 +0000[diff] [blame]457<!--=========================================================================-->
458<div class="doc_subsection">
459<a name="GHC">Glasgow Haskell Compiler (GHC)</a>
460</div>
461
462<div class="doc_text">
463<p>
464<a href="http://www.haskell.org/ghc/">GHC</a> is an open source,
465state-of-the-art programming suite for Haskell, a standard lazy
466functional programming language. It includes an optimizing static
467compiler generating good code for a variety of platforms, together
468with an interactive system for convenient, quick development.</p>
469
470<p>In addition to the existing C and native code generators, GHC now
471supports an <a
472href="http://hackage.haskell.org/trac/ghc/wiki/Commentary/Compiler/Backends/LLVM">LLVM
473code generator</a>. GHC supports LLVM 2.7.</p>
474
475</div>
476
Chris Lattner64694df2010-04-22 21:34:16 +0000[diff] [blame]477
Chris Lattnerc75fd522008-06-08 21:58:17 +0000[diff] [blame]478<!-- *********************************************************************** -->
479<div class="doc_section">
Jeffrey Yasskin0830b972010-01-28 01:14:43 +0000[diff] [blame]480 <a name="whatsnew">What's New in LLVM 2.7?</a>
Chris Lattnerc75fd522008-06-08 21:58:17 +0000[diff] [blame]481</div>
482<!-- *********************************************************************** -->
483
Chris Lattnerb7bc2aa2008-06-08 22:59:35 +0000[diff] [blame]484<div class="doc_text">
485
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]486<p>This release includes a huge number of bug fixes, performance tweaks and
Chris Lattner1e4d5bc2008-10-13 17:57:36 +0000[diff] [blame]487minor improvements. Some of the major improvements and new features are listed
488in this section.
Chris Lattnerb7bc2aa2008-06-08 22:59:35 +0000[diff] [blame]489</p>
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]490
Chris Lattnera67df2d2010-04-22 06:28:20 +0000[diff] [blame]491</div>
492
493<!--=========================================================================-->
494<div class="doc_subsection">
495<a name="orgchanges">LLVM Community Changes</a>
496</div>
497
498<div class="doc_text">
499
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]500<p>In addition to changes to the code, between LLVM 2.6 and 2.7, a number of
501organization changes have happened:
502</p>
503
504<ul>
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]505<li>LLVM has a new <a href="http://llvm.org/Logo.html">official logo</a>!</li>
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]506
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]507<li>Ted Kremenek and Doug Gregor have stepped forward as <a
508 href="http://llvm.org/docs/DeveloperPolicy.html#owners">Code Owners</a> of the
Chris Lattnera67df2d2010-04-22 06:28:20 +0000[diff] [blame]509 Clang static analyzer and the Clang frontend, respectively.</li>
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]510
511<li>LLVM now has an <a href="http://blog.llvm.org">official Blog</a> at
512 <a href="http://blog.llvm.org">http://blog.llvm.org</a>. This is a great way
513 to learn about new LLVM-related features as they are implemented. Several
514 features in this release are already explained on the blog.</li>
515
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]516<li>The LLVM web pages are now checked into the SVN server, in the "www",
517 "www-pubs" and "www-releases" SVN modules. Previously they were hidden in a
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]518 largely inaccessible old CVS server.</li>
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]519
520<li><a href="http://llvm.org">llvm.org</a> is now hosted on a new (and much
521 faster) server. It is still graciously hosted at the University of Illinois
522 of Urbana Champaign.</li>
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]523</ul>
Chris Lattnerb7bc2aa2008-06-08 22:59:35 +0000[diff] [blame]524</div>
525
Chris Lattnerc75fd522008-06-08 21:58:17 +0000[diff] [blame]526<!--=========================================================================-->
Chris Lattner625a3d82008-06-08 21:34:41 +0000[diff] [blame]527<div class="doc_subsection">
Chris Lattnercdc44ed2008-02-06 18:00:06 +0000[diff] [blame]528<a name="majorfeatures">Major New Features</a>
529</div>
530
531<div class="doc_text">
532
Jeffrey Yasskin0830b972010-01-28 01:14:43 +0000[diff] [blame]533<p>LLVM 2.7 includes several major new capabilities:</p>
Chris Lattner97beb512007-05-14 06:56:09 +0000[diff] [blame]534
535<ul>
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]536<li>2.7 includes initial support for the <a
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]537 href="http://en.wikipedia.org/wiki/MicroBlaze">MicroBlaze</a> target.
538 MicroBlaze is a soft processor core designed for Xilinx FPGAs.</li>
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]539
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]540<li>2.7 includes a new LLVM IR "extensible metadata" feature. This feature
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]541 supports many different use cases, including allowing front-end authors to
542 encode source level information into LLVM IR, which is consumed by later
543 language-specific passes. This is a great way to do high-level optimizations
544 like devirtualization, type-based alias analysis, etc. See the <a
545 href="http://blog.llvm.org/2010/04/extensible-metadata-in-llvm-ir.html">
546 Extensible Metadata Blog Post</a> for more information.</li>
547
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]548<li>2.7 encodes <a href="SourceLevelDebugging.html">debug information</a>
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]549in a completely new way, built on extensible metadata. The new implementation
550is much more memory efficient and paves the way for improvements to optimized
551code debugging experience.</li>
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]552
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]553<li>2.7 now directly supports taking the address of a label and doing an
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]554 indirect branch through a pointer. This is particularly useful for
555 interpreter loops, and is used to implement the GCC "address of label"
556 extension. For more information, see the <a
557href="http://blog.llvm.org/2010/01/address-of-label-and-indirect-branches.html">
558Address of Label and Indirect Branches in LLVM IR Blog Post</a>.
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]559
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]560<li>2.7 is the first release to start supporting APIs for assembling and
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]561 disassembling target machine code. These APIs are useful for a variety of
562 low level clients, and are surfaced in the new "enhanced disassembly" API.
563 For more information see the <a
564 href="http://blog.llvm.org/2010/01/x86-disassembler.html">The X86
565 Disassembler Blog Post</a> for more information.</li>
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]566
Chris Lattner2284b6a2010-04-22 06:38:11 +0000[diff] [blame]567<li>2.7 includes major parts of the work required by the new MC Project,
568 see the <a href="#mc">MC update above</a> for more information.</li>
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]569
Chris Lattner458e79f2008-02-10 08:18:42 +0000[diff] [blame]570</ul>
Chris Lattner10a3cc42010-03-17 06:41:58 +0000[diff] [blame]571
Chris Lattner97beb512007-05-14 06:56:09 +0000[diff] [blame]572</div>
573
Chris Lattner0a1fd102007-09-21 03:54:09 +0000[diff] [blame]574<!--=========================================================================-->
575<div class="doc_subsection">
Chris Lattnerd1094e02009-03-02 02:37:32 +0000[diff] [blame]576<a name="coreimprovements">LLVM IR and Core Improvements</a>
Chris Lattnerdd6acc02008-02-10 08:17:19 +0000[diff] [blame]577</div>
578
579<div class="doc_text">
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]580<p>LLVM IR has several new features for better support of new targets and that
581expose new optimization opportunities:</p>
Chris Lattnerdd6acc02008-02-10 08:17:19 +0000[diff] [blame]582
Chris Lattnerb7112222008-06-05 06:25:56 +0000[diff] [blame]583<ul>
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]584<li>LLVM IR now supports a 16-bit "half float" data type through <a
585 href="LangRef.html#int_fp16">two new intrinsics</a> and APFloat support.</li>
586<li>LLVM IR supports two new <a href="LangRef.html#fnattrs">function
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]587 attributes</a>: inlinehint and alignstack(n). The former is a hint to the
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]588 optimizer that a function was declared 'inline' and thus the inliner should
589 weight it higher when considering inlining it. The later
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]590 indicates to the code generator that the function diverges from the platform
591 ABI on stack alignment.</li>
592<li>The new <a href="LangRef.html#int_objectsize">llvm.objectsize</a> intrinsic
593 allows the optimizer to infer the sizes of memory objects in some cases.
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]594 This intrinsic is used to implement the GCC <tt>__builtin_object_size</tt>
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]595 extension.</li>
Chris Lattnera67df2d2010-04-22 06:28:20 +0000[diff] [blame]596<li>LLVM IR now supports marking load and store instructions with <a
597 href="LangRef.html#i_load">"non-temporal" hints</a> (building on the new
598 metadata feature). This hint encourages the code
599 generator to generate non-temporal accesses when possible, which are useful
600 for code that is carefully managing cache behavior. Currently, only the
601 X86 backend provides target support for this feature.</li>
602
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]603<li>LLVM 2.7 has pre-alpha support for <a
Gabor Greif862b3532010-04-22 10:11:24 +0000[diff] [blame]604 href="LangRef.html#t_union">unions in LLVM IR</a>.
Chris Lattnera67df2d2010-04-22 06:28:20 +0000[diff] [blame]605 Unfortunately, this support is not really usable in 2.7, so if you're
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]606 interested in pushing it forward, please help contribute to LLVM mainline.</li>
Gabor Greif27b166352010-04-22 10:21:43 +0000[diff] [blame]607
Chris Lattnerdd6acc02008-02-10 08:17:19 +0000[diff] [blame]608</ul>
Mikhail Glushenkov024f7cf2008-10-13 02:08:34 +0000[diff] [blame]609
Chris Lattnerdd6acc02008-02-10 08:17:19 +0000[diff] [blame]610</div>
611
612<!--=========================================================================-->
613<div class="doc_subsection">
Chris Lattner2b8a52e2008-02-10 07:46:44 +0000[diff] [blame]614<a name="optimizer">Optimizer Improvements</a>
615</div>
616
617<div class="doc_text">
618
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]619<p>In addition to a large array of minor performance tweaks and bug fixes, this
Chris Lattner25879d72008-10-13 21:50:36 +0000[diff] [blame]620release includes a few major enhancements and additions to the optimizers:</p>
Chris Lattner2b8a52e2008-02-10 07:46:44 +0000[diff] [blame]621
622<ul>
623
Chris Lattnera3cc83b2010-04-27 07:28:11 +0000[diff] [blame^]624<li>The inliner now merges arrays stack objects in different callees when
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]625 inlining multiple call sites into one function. This reduces the stack size
626 of the resultant function.</li>
627<li>The -basicaa alias analysis pass (which is the default) has been improved to
628 be less dependent on "type safe" pointers. It can now look through bitcasts
629 and other constructs more aggressively, allowing better load/store
630 optimization.</li>
631<li>The load elimination optimization in the GVN Pass [<a
632href="http://blog.llvm.org/2009/12/introduction-to-load-elimination-in-gvn.html">intro
633 blog post</a>] has been substantially improved to be more aggressive about
634 partial redundancy elimination and do more aggressive phi translation. Please
635 see the <a
636 href="http://blog.llvm.org/2009/12/advanced-topics-in-redundant-load.html">
637 Advanced Topics in Redundant Load Elimination with a Focus on PHI Translation
638 Blog Post</a> for more details.</li>
639<li>The module <a href="LangRef.html#datalayout">target data string</a> now
Wesley Peckc8459162010-04-22 13:28:34 +0000[diff] [blame]640 includes a notion of 'native' integer data types for the target. This
641 helps mid-level optimizations avoid promoting complex sequences of
642 operations to data types that are not natively supported (e.g. converting
643 i32 operations to i64 on 32-bit chips).</li>
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]644<li>The mid-level optimizer is now conservative when operating on a module with
645 no target data. Previously, it would default to SparcV9 settings, which is
646 not what most people expected.</li>
Chris Lattnera668b452010-04-21 06:42:24 +0000[diff] [blame]647<li>Jump threading is now much more aggressive at simplifying correlated
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]648 conditionals and threading blocks with otherwise complex logic. It has
649 subsumed the old "Conditional Propagation" pass, and -condprop has been
650 removed from LLVM 2.7.</li>
651<li>The -instcombine pass has been refactored from being one huge file to being
652 a library of its own. Internally, it uses a customized IRBuilder to clean
653 it up and simplify it.</li>
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]654
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]655<li>The optimal edge profiling pass is reliable and much more complete than in
656 2.6. It can be used with the llvm-prof tool but isn't wired up to the
657 llvm-gcc and clang command line options yet.</li>
Chris Lattnerd1094e02009-03-02 02:37:32 +0000[diff] [blame]658
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]659<li>A new experimental alias analysis implementation, -scev-aa, has been added.
660 It uses LLVM's Scalar Evolution implementation to do symbolic analysis of
661 pointer offset expressions to disambiguate pointers. It can catch a few
662 cases that basicaa cannot, particularly in complex loop nests.</li>
Chris Lattnera668b452010-04-21 06:42:24 +0000[diff] [blame]663
Chris Lattnera67df2d2010-04-22 06:28:20 +0000[diff] [blame]664<li>The default pass ordering has been tweaked for improved optimization
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]665 effectiveness.</li>
Chris Lattnera668b452010-04-21 06:42:24 +0000[diff] [blame]666
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]667</ul>
Chris Lattnerd1094e02009-03-02 02:37:32 +0000[diff] [blame]668
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]669</div>
Chris Lattnerd1094e02009-03-02 02:37:32 +0000[diff] [blame]670
Chris Lattnerd1094e02009-03-02 02:37:32 +0000[diff] [blame]671
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]672<!--=========================================================================-->
673<div class="doc_subsection">
674<a name="executionengine">Interpreter and JIT Improvements</a>
675</div>
Chris Lattner25879d72008-10-13 21:50:36 +0000[diff] [blame]676
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]677<div class="doc_text">
678
679<ul>
Wesley Peckba1ad202010-04-22 13:36:27 +0000[diff] [blame]680<li>The JIT now supports generating debug information and is compatible with
681the new GDB 7.0 (and later) interfaces for registering dynamically generated
682debug info.</li>
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]683
Chris Lattnera67df2d2010-04-22 06:28:20 +0000[diff] [blame]684<li>The JIT now <a href="http://llvm.org/PR5184">defaults
Jeffrey Yasskin9fb8ce82010-01-29 19:10:38 +0000[diff] [blame]685to compiling eagerly</a> to avoid a race condition in the lazy JIT.
686Clients that still want the lazy JIT can switch it on by calling
687<tt>ExecutionEngine::DisableLazyCompilation(false)</tt>.</li>
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]688
Jeffrey Yasskine0913882010-02-11 01:07:39 +0000[diff] [blame]689<li>It is now possible to create more than one JIT instance in the same process.
690These JITs can generate machine code in parallel,
691although <a href="http://llvm.org/docs/ProgrammersManual.html#jitthreading">you
692still have to obey the other threading restrictions</a>.</li>
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]693
Chris Lattner7795ea92008-06-08 02:45:07 +0000[diff] [blame]694</ul>
695
696</div>
697
698<!--=========================================================================-->
699<div class="doc_subsection">
Chris Lattnerd434bfb2009-03-02 03:24:11 +0000[diff] [blame]700<a name="codegen">Target Independent Code Generator Improvements</a>
Chris Lattner7795ea92008-06-08 02:45:07 +0000[diff] [blame]701</div>
702
703<div class="doc_text">
704
Mikhail Glushenkov25422542009-03-01 18:09:47 +0000[diff] [blame]705<p>We have put a significant amount of work into the code generator
706infrastructure, which allows us to implement more aggressive algorithms and make
707it run faster:</p>
Chris Lattner7795ea92008-06-08 02:45:07 +0000[diff] [blame]708
709<ul>
Chris Lattnera67df2d2010-04-22 06:28:20 +0000[diff] [blame]710<li>The 'llc -asm-verbose' option (which is now the default) has been enhanced
711 to emit many useful comments to .s files indicating information about spill
712 slots and loop nest structure. This should make it much easier to read and
713 understand assembly files. This is wired up in llvm-gcc and clang to
714 the <tt>-fverbose-asm</tt> option.</li>
715
Dan Gohman00365f92010-04-22 20:50:43 +0000[diff] [blame]716<li>New LSR with "full strength reduction" mode, which can reduce address
717 register pressure in loops where address generation is important.</li>
Chris Lattnera67df2d2010-04-22 06:28:20 +0000[diff] [blame]718
719<li>A new codegen level Common Subexpression Elimination pass (MachineCSE)
720 is available and enabled by default. It catches redundancies exposed by
721 lowering.</li>
722<li>A new pre-register-allocation tail duplication pass is available and enabled
723 by default, it can substantially improve branch prediction quality in some
724 cases.</li>
725<li>A new sign and zero extension optimization pass (OptimizeExtsPass)
726 is available and enabled by default. This pass can takes advantage
727 architecture features like x86-64 implicit zero extension behavior and
728 sub-registers.</li>
729<li>The code generator now supports a mode where it attempts to preserve the
730 order of instructions in the input code. This is important for source that
731 is hand scheduled and extremely sensitive to scheduling. It is compatible
732 with the GCC <tt>-fno-schedule-insns</tt> option.</li>
733<li>The target-independent code generator now supports generating code with
734 arbitrary numbers of result values. Returning more values than was
735 previously supported is handled by returning through a hidden pointer. In
736 2.7, only the X86 and XCore targets have adopted support for this
737 though.</li>
738<li>The code generator now supports generating code that follows the
739 <a href="LangRef.html#callingconv">Glasgow Haskell Compiler Calling
740 Convention</a> and ABI.</li>
741<li>The "<a href="CodeGenerator.html#selectiondag_select">DAG instruction
742 selection</a>" phase of the code generator has been largely rewritten for
743 2.7. Previously, tblgen spit out tons of C++ code which was compiled and
744 linked into the target to do the pattern matching, now it emits a much
745 smaller table which is read by the target-independent code. The primary
746 advantages of this approach is that the size and compile time of various
747 targets is much improved. The X86 code generator shrunk by 1.5MB of code,
748 for example.</li>
749<li>Almost the entire code generator has switched to emitting code through the
750 MC interfaces instead of printing textually to the .s file. This led to a
751 number of cleanups and speedups. In 2.7, debug an exception handling
752 information does not go through MC yet.</li>
Chris Lattner0a1fd102007-09-21 03:54:09 +0000[diff] [blame]753</ul>
Chris Lattner0a1fd102007-09-21 03:54:09 +0000[diff] [blame]754</div>
755
Chris Lattnerd1094e02009-03-02 02:37:32 +0000[diff] [blame]756<!--=========================================================================-->
757<div class="doc_subsection">
758<a name="x86">X86-32 and X86-64 Target Improvements</a>
759</div>
760
761<div class="doc_text">
762<p>New features of the X86 target include:
763</p>
764
765<ul>
Chris Lattnera668b452010-04-21 06:42:24 +0000[diff] [blame]766<li>The X86 backend now optimizes tails calls much more aggressively for
767 functions that use the standard C calling convention.</li>
768<li>The X86 backend now models scalar SSE registers as subregs of the SSE vector
769 registers, making the code generator more aggressive in cases where scalars
770 and vector types are mixed.</li>
Chris Lattnerd434bfb2009-03-02 03:24:11 +0000[diff] [blame]771
Chris Lattnerd1094e02009-03-02 02:37:32 +0000[diff] [blame]772</ul>
773
Chris Lattnerd1094e02009-03-02 02:37:32 +0000[diff] [blame]774</div>
Chris Lattner0a1fd102007-09-21 03:54:09 +0000[diff] [blame]775
776<!--=========================================================================-->
777<div class="doc_subsection">
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]778<a name="ARM">ARM Target Improvements</a>
Chris Lattnerc92d7692009-03-01 02:30:21 +0000[diff] [blame]779</div>
780
781<div class="doc_text">
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]782<p>New features of the ARM target include:
783</p>
Chris Lattnerc92d7692009-03-01 02:30:21 +0000[diff] [blame]784
785<ul>
Chris Lattnerc92d7692009-03-01 02:30:21 +0000[diff] [blame]786
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]787<li>The ARM backend now generates instructions in unified assembly syntax.</li>
788
789<li>llvm-gcc now has complete support for the ARM v7 NEON instruction set. This
790 support differs slightly from the GCC implementation. Please see the
791 <a
792href="http://blog.llvm.org/2010/04/arm-advanced-simd-neon-intrinsics-and.html">
793 ARM Advanced SIMD (NEON) Intrinsics and Types in LLVM Blog Post</a> for
794 helpful information if migrating code from GCC to LLVM-GCC.</li>
795
Wesley Peck16144fb2010-04-22 13:43:14 +0000[diff] [blame]796<li>The ARM and Thumb code generators now use register scavenging for stack
Jim Grosbach77f0f5b2010-04-22 18:28:43 +0000[diff] [blame]797 object address materialization. This allows the use of R3 as a general
798 purpose register in Thumb1 code, as it was previous reserved for use in
799 stack address materialization. Secondly, sequential uses of the same
800 value will now re-use the materialized constant.</li>
801
Wesley Peck16144fb2010-04-22 13:43:14 +0000[diff] [blame]802<li>The ARM backend now has good support for ARMv4 targets and has been tested
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]803 on StrongARM hardware. Previously, LLVM only supported ARMv4T and
804 newer chips.</li>
Jim Grosbach96284682010-04-22 18:33:31 +0000[diff] [blame]805
806<li>Atomic builtins are now supported for ARMv6 and ARMv7 (__sync_synchronize,
807 __sync_fetch_and_add, etc.).</li>
808
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]809</ul>
Chris Lattnerc92d7692009-03-01 02:30:21 +0000[diff] [blame]810
Chris Lattnerc92d7692009-03-01 02:30:21 +0000[diff] [blame]811
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]812</div>
Chris Lattnerc92d7692009-03-01 02:30:21 +0000[diff] [blame]813
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]814<!--=========================================================================-->
815<div class="doc_subsection">
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]816<a name="newapis">New Useful APIs</a>
817</div>
818
819<div class="doc_text">
820
821<p>This release includes a number of new APIs that are used internally, which
822 may also be useful for external clients.
823</p>
824
825<ul>
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]826<li>The optimizer uses the new CodeMetrics class to measure the size of code.
Chris Lattner77a23802010-04-22 17:39:38 +0000[diff] [blame]827 Various passes (like the inliner, loop unswitcher, etc) all use this to make
828 more accurate estimates of the code size impact of various
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]829 optimizations.</li>
830<li>A new <a href="http://llvm.org/doxygen/InstructionSimplify_8h-source.html">
Wesley Peck3edc1682010-04-22 13:50:46 +0000[diff] [blame]831 llvm/Analysis/InstructionSimplify.h</a> interface is available for doing
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]832 symbolic simplification of instructions (e.g. <tt>a+0</tt> -&gt; <tt>a</tt>)
833 without requiring the instruction to exist. This centralizes a lot of
834 ad-hoc symbolic manipulation code scattered in various passes.</li>
835<li>The optimizer now uses a new <a
836 href="http://llvm.org/doxygen/SSAUpdater_8h-source.html">SSAUpdater</a>
837 class which efficiently supports
838 doing unstructured SSA update operations. This centralized a bunch of code
Wesley Peck3edc1682010-04-22 13:50:46 +0000[diff] [blame]839 scattered throughout various passes (e.g. jump threading, lcssa,
840 loop rotate, etc) for doing this sort of thing. The code generator has a
841 similar <a href="http://llvm.org/doxygen/MachineSSAUpdater_8h-source.html">
Chris Lattner656db162010-04-22 05:41:35 +0000[diff] [blame]842 MachineSSAUpdater</a> class.</li>
843<li>The <a href="http://llvm.org/doxygen/Regex_8h-source.html">
844 llvm/Support/Regex.h</a> header exposes a platform independent regular
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]845 expression API. Building on this, the <a
846 href="TestingGuide.html#FileCheck">FileCheck</a> utility now supports
847 regular exressions.</li>
848<li>raw_ostream now supports a circular "debug stream" accessed with "dbgs()".
849 By default, this stream works the same way as "errs()", but if you pass
850 <tt>-debug-buffer-size=1000</tt> to opt, the debug stream is capped to a
851 fixed sized circular buffer and the output is printed at the end of the
852 program's execution. This is helpful if you have a long lived compiler
853 process and you're interested in seeing snapshots in time.</li>
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]854</ul>
855
856
857</div>
858
859<!--=========================================================================-->
860<div class="doc_subsection">
861<a name="otherimprovements">Other Improvements and New Features</a>
862</div>
863
864<div class="doc_text">
865<p>Other miscellaneous features include:</p>
866
867<ul>
Chris Lattner1dcca8b2010-04-21 06:23:40 +0000[diff] [blame]868<li>You can now build LLVM as a big dynamic library (e.g. "libllvm2.7.so"). To
869 get this, configure LLVM with the --enable-shared option.</li>
870
Wesley Peckcad613d2010-04-22 13:53:54 +0000[diff] [blame]871<li>LLVM command line tools now overwrite their output by default. Previously,
872 they would only do this with -f. This makes them more convenient to use, and
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]873 behave more like standard unix tools.</li>
874
875<li>The opt and llc tools now autodetect whether their input is a .ll or .bc
876 file, and automatically do the right thing. This means you don't need to
877 explicitly use the llvm-as tool for most things.</li>
Chris Lattnerc92d7692009-03-01 02:30:21 +0000[diff] [blame]878</ul>
879
880</div>
881
Chris Lattner6cb64032008-06-05 08:02:49 +0000[diff] [blame]882
883<!--=========================================================================-->
884<div class="doc_subsection">
Chris Lattner1e4d5bc2008-10-13 17:57:36 +0000[diff] [blame]885<a name="changes">Major Changes and Removed Features</a>
886</div>
887
888<div class="doc_text">
889
Chris Lattner934e2d42008-10-13 22:06:31 +0000[diff] [blame]890<p>If you're already an LLVM user or developer with out-of-tree changes based
Jeffrey Yasskin0830b972010-01-28 01:14:43 +0000[diff] [blame]891on LLVM 2.6, this section lists some "gotchas" that you may run into upgrading
Chris Lattner934e2d42008-10-13 22:06:31 +0000[diff] [blame]892from the previous release.</p>
Chris Lattner1e4d5bc2008-10-13 17:57:36 +0000[diff] [blame]893
894<ul>
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]895
896<li>
897The Andersen's alias analysis ("anders-aa") pass, the Predicate Simplifier
898("predsimplify") pass, the LoopVR pass, the GVNPRE pass, and the random sampling
899profiling ("rsprofiling") passes have all been removed. They were not being
900actively maintained and had substantial problems. If you are interested in
901these components, you are welcome to ressurect them from SVN, fix the
902correctness problems, and resubmit them to mainline.</li>
903
904<li>LLVM now defaults to building most libraries with RTTI turned off, providing
905a code size reduction. Packagers who are interested in building LLVM to support
906plugins that require RTTI information should build with "make REQUIRE_RTTI=1"
907and should read the new <a href="Packaging.html">Advice on Packaging LLVM</a>
908document.</li>
909
Jeffrey Yasskin914050b2010-02-09 23:03:44 +0000[diff] [blame]910<li>The LLVM interpreter now defaults to <em>not</em> using <tt>libffi</tt> even
911if you have it installed. This makes it more likely that an LLVM built on one
912system will work when copied to a similar system. To use <tt>libffi</tt>,
Chris Lattnera67df2d2010-04-22 06:28:20 +0000[diff] [blame]913configure with <tt>--enable-libffi</tt>.</li>
Chris Lattner934e2d42008-10-13 22:06:31 +0000[diff] [blame]914
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]915<li>Debug information uses a completely different representation, an LLVM 2.6
916.bc file should work with LLVM 2.7, but debug info won't come forward.</li>
917
918<li>The LLVM 2.6 (and earlier) "malloc" and "free" instructions got removed,
919 along with LowerAllocations pass. Now you should just use a call to the
920 malloc and free functions in libc. These calls are optimized as well as
921 the old instructions were.</li>
922</ul>
Chris Lattner934e2d42008-10-13 22:06:31 +0000[diff] [blame]923
924<p>In addition, many APIs have changed in this release. Some of the major LLVM
925API changes are:</p>
926
927<ul>
Gabor Greif137d90c2010-04-26 20:46:03 +0000[diff] [blame]928<li>Just about everything has been converted to use <tt>raw_ostream</tt> instead of
929 <tt>std::ostream</tt>.</li>
930<li><tt>llvm/ADT/iterator.h</tt> has been removed, just use <tt>&lt;iterator&gt;</tt>
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]931 instead.</li>
Gabor Greif137d90c2010-04-26 20:46:03 +0000[diff] [blame]932<li>The <tt>Streams.h</tt> file and <tt>DOUT</tt> got removed, use <tt>DEBUG(errs() &lt;&lt; ...);</tt>
Chris Lattner87e04952010-04-21 05:17:40 +0000[diff] [blame]933 instead.</li>
Gabor Greif137d90c2010-04-26 20:46:03 +0000[diff] [blame]934<li>The <tt>TargetAsmInfo</tt> interface was renamed to <tt>MCAsmInfo</tt>.</li>
Jeffrey Yasskin4a8b1da2010-01-28 01:41:20 +0000[diff] [blame]935<li><tt>ModuleProvider</tt> has been <a
Gabor Greif27b166352010-04-22 10:21:43 +0000[diff] [blame]936href="http://llvm.org/viewvc/llvm-project?view=rev&amp;revision=94686">removed</a>
Jeffrey Yasskin4a8b1da2010-01-28 01:41:20 +0000[diff] [blame]937and its methods moved to <tt>Module</tt> and <tt>GlobalValue</tt>.
938Most clients can remove uses of <tt>ExistingModuleProvider</tt>,
939replace <tt>getBitcodeModuleProvider</tt> with
940<tt>getLazyBitcodeModule</tt>, and pass their <tt>Module</tt> to
941functions that used to accept <tt>ModuleProvider</tt>. Clients who
942wrote their own <tt>ModuleProvider</tt>s will need to derive from
943<tt>GVMaterializer</tt> instead and use
944<tt>Module::setMaterializer</tt> to attach it to a
945<tt>Module</tt>.</li>
946
947<li><tt>GhostLinkage</tt> has given up the ghost.
948<tt>GlobalValue</tt>s that have not yet been read from their backing
949storage have the same linkage they will have after being read in.
950Clients must replace calls to
951<tt>GlobalValue::hasNotBeenReadFromBitcode</tt> with
952<tt>GlobalValue::isMaterializable</tt>.</li>
Daniel Dunbar1e96fca2010-02-10 04:09:52 +0000[diff] [blame]953
Duncan Sands80c430a2010-02-17 17:20:17 +0000[diff] [blame]954<li>The <tt>isInteger</tt>, <tt>isIntOrIntVector</tt>, <tt>isFloatingPoint</tt>,
955<tt>isFPOrFPVector</tt> and <tt>isFPOrFPVector</tt> methods have been renamed
956<tt>isIntegerTy</tt>, <tt>isIntOrIntVectorTy</tt>, <tt>isFloatingPointTy</tt>,
957<tt>isFPOrFPVectorTy</tt> and <tt>isFPOrFPVectorTy</tt> respectively.</li>
Chris Lattnercecb9232010-04-22 17:25:00 +0000[diff] [blame]958
959<li><tt>llvm::Instruction::clone()</tt> no longer takes argument.</li>
960<li><tt>raw_fd_ostream</tt>'s constructor now takes a flag argument, not individual
961 booleans (see <tt>include/llvm/Support/raw_ostream.h</tt> for details).</li>
962<li>Some header files have been renamed:
963<ul>
964 <li><tt>llvm/Support/AIXDataTypesFix.h</tt> to
965 <tt>llvm/System/AIXDataTypesFix.h</tt></li>
966 <li><tt>llvm/Support/DataTypes.h</tt> to <tt>llvm/System/DataTypes.h</tt></li>
967 <li><tt>llvm/Transforms/Utils/InlineCost.h</tt> to
968 <tt>llvm/Analysis/InlineCost.h</tt></li>
969 <li><tt>llvm/Support/Mangler.h</tt> to <tt>llvm/Target/Mangler.h</tt></li>
970 <li><tt>llvm/Analysis/Passes.h</tt> to <tt>llvm/CodeGen/Passes.h</tt></li>
971</ul></li>
Devang Pateldbf83832008-10-14 20:03:43 +0000[diff] [blame]972</ul>
Chris Lattner1e4d5bc2008-10-13 17:57:36 +0000[diff] [blame]973
Chris Lattner1e4d5bc2008-10-13 17:57:36 +0000[diff] [blame]974</div>
975
976
977
Chris Lattner62495762003-10-02 16:38:05 +0000[diff] [blame]978<!-- *********************************************************************** -->
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]979<div class="doc_section">
980 <a name="portability">Portability and Supported Platforms</a>
981</div>
Chris Lattner62495762003-10-02 16:38:05 +0000[diff] [blame]982<!-- *********************************************************************** -->
983
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]984<div class="doc_text">
985
John Criswell64307722004-12-08 20:35:47 +0000[diff] [blame]986<p>LLVM is known to work on the following platforms:</p>
Chris Lattner2c7b0872004-06-01 18:22:41 +0000[diff] [blame]987
988<ul>
Mikhail Glushenkov25422542009-03-01 18:09:47 +0000[diff] [blame]989<li>Intel and AMD machines (IA32, X86-64, AMD64, EMT-64) running Red Hat
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]990 Linux, Fedora Core, FreeBSD and AuroraUX (and probably other unix-like
991 systems).</li>
Chris Lattnera67df2d2010-04-22 06:28:20 +0000[diff] [blame]992<li>PowerPC and X86-based Mac OS X systems, running 10.4 and above in 32-bit
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]993 and 64-bit modes.</li>
Chris Lattnercdc44ed2008-02-06 18:00:06 +0000[diff] [blame]994<li>Intel and AMD machines running on Win32 using MinGW libraries (native).</li>
Chris Lattnerbee7b322007-05-23 04:39:32 +0000[diff] [blame]995<li>Intel and AMD machines running on Win32 with the Cygwin libraries (limited
996 support is available for native builds with Visual C++).</li>
Chris Lattner51948622009-10-16 16:30:58 +0000[diff] [blame]997<li>Sun x86 and AMD64 machines running Solaris 10, OpenSolaris 0906.</li>
John Criswelld5234922005-05-13 20:28:15 +0000[diff] [blame]998<li>Alpha-based machines running Debian GNU/Linux.</li>
Chris Lattner2c7b0872004-06-01 18:22:41 +0000[diff] [blame]999</ul>
1000
Chris Lattner618a99f2008-06-05 06:57:39 +0000[diff] [blame]1001<p>The core LLVM infrastructure uses GNU autoconf to adapt itself
Brian Gaekee3e557e2004-05-09 05:28:35 +0000[diff] [blame]1002to the machine and operating system on which it is built. However, minor
1003porting may be required to get LLVM to work on new platforms. We welcome your
1004portability patches and reports of successful builds or error messages.</p>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1005
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1006</div>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1007
1008<!-- *********************************************************************** -->
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1009<div class="doc_section">
1010 <a name="knownproblems">Known Problems</a>
1011</div>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1012<!-- *********************************************************************** -->
1013
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1014<div class="doc_text">
1015
Mikhail Glushenkov25422542009-03-01 18:09:47 +0000[diff] [blame]1016<p>This section contains significant known problems with the LLVM system,
Chris Lattner2a092392008-11-10 05:40:34 +0000[diff] [blame]1017listed by component. If you run into a problem, please check the <a
Chris Lattnera69595e2005-10-29 07:07:09 +0000[diff] [blame]1018href="http://llvm.org/bugs/">LLVM bug database</a> and submit a bug if
Chris Lattnerb84f3322003-12-12 21:22:16 +0000[diff] [blame]1019there isn't already one.</p>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1020
Chris Lattner6aba4862010-03-17 17:25:49 +0000[diff] [blame]1021<ul>
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]1022<li>LLVM will not correctly compile on Solaris and/or OpenSolaris
1023using the stock GCC 3.x.x series 'out the box',
Chris Lattner6e41c792009-11-03 21:50:09 +0000[diff] [blame]1024See: <a href="GettingStarted.html#brokengcc">Broken versions of GCC and other tools</a>.
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]1025However, A <a href="http://pkg.auroraux.org/GCC">Modern GCC Build</a>
1026for x86/x86-64 has been made available from the third party AuroraUX Project
1027that has been meticulously tested for bootstrapping LLVM &amp; Clang.</li>
1028</ul>
1029
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1030</div>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1031
Chris Lattnerb911de42004-03-14 02:03:02 +0000[diff] [blame]1032<!-- ======================================================================= -->
1033<div class="doc_subsection">
1034 <a name="experimental">Experimental features included with this release</a>
1035</div>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1036
Chris Lattnerb911de42004-03-14 02:03:02 +0000[diff] [blame]1037<div class="doc_text">
1038
Misha Brukmanfa50a222004-05-12 21:46:05 +0000[diff] [blame]1039<p>The following components of this LLVM release are either untested, known to
1040be broken or unreliable, or are in early development. These components should
1041not be relied on, and bugs should not be filed against them, but they may be
1042useful to some people. In particular, if you would like to work on one of these
Chris Lattnere38ac152008-02-12 06:29:45 +0000[diff] [blame]1043components, please contact us on the <a
1044href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVMdev list</a>.</p>
Chris Lattnerb911de42004-03-14 02:03:02 +0000[diff] [blame]1045
1046<ul>
Wesley Peck3aa84e52010-03-18 14:31:30 +0000[diff] [blame]1047<li>The MSIL, Alpha, SPU, MIPS, PIC16, Blackfin, MSP430, SystemZ and MicroBlaze
1048 backends are experimental.</li>
Chris Lattnera67df2d2010-04-22 06:28:20 +0000[diff] [blame]1049<li><tt>llc</tt> "<tt>-filetype=asm</tt>" (the default) is the only
Chris Lattner0b1c9a52010-03-17 04:41:49 +0000[diff] [blame]1050 supported value for this option. The MachO writer is experimental, and
1051 works much better in mainline SVN.</li>
Chris Lattnerb911de42004-03-14 02:03:02 +0000[diff] [blame]1052</ul>
1053
1054</div>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1055
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1056<!-- ======================================================================= -->
1057<div class="doc_subsection">
Chris Lattnerb81f10e2006-11-18 07:51:14 +0000[diff] [blame]1058 <a name="x86-be">Known problems with the X86 back-end</a>
John Criswell3bdbd302005-11-08 21:11:33 +0000[diff] [blame]1059</div>
1060
1061<div class="doc_text">
1062
1063<ul>
Anton Korobeynikov486c7d32008-06-08 10:24:13 +0000[diff] [blame]1064 <li>The X86 backend does not yet support
1065 all <a href="http://llvm.org/PR879">inline assembly that uses the X86
1066 floating point stack</a>. It supports the 'f' and 't' constraints, but not
1067 'u'.</li>
Chris Lattnerb076b002008-06-08 21:19:07 +0000[diff] [blame]1068 <li>The X86 backend generates inefficient floating point code when configured
1069 to generate code for systems that don't have SSE2.</li>
Duncan Sands0bc15262008-06-08 19:38:43 +0000[diff] [blame]1070 <li>Win64 code generation wasn't widely tested. Everything should work, but we
John Criswellea03c9d2009-03-02 15:28:15 +0000[diff] [blame]1071 expect small issues to happen. Also, llvm-gcc cannot build the mingw64
Chris Lattnera67df2d2010-04-22 06:28:20 +0000[diff] [blame]1072 runtime currently due to lack of support for the 'u' inline assembly
1073 constraint and for X87 floating point inline assembly.</li>
Dan Gohman721b3722008-06-08 23:05:11 +0000[diff] [blame]1074 <li>The X86-64 backend does not yet support the LLVM IR instruction
Chris Lattnera67df2d2010-04-22 06:28:20 +0000[diff] [blame]1075 <tt>va_arg</tt>. Currently, front-ends support variadic
Dan Gohman721b3722008-06-08 23:05:11 +0000[diff] [blame]1076 argument constructs on X86-64 by lowering them manually.</li>
Chris Lattnerb81f10e2006-11-18 07:51:14 +0000[diff] [blame]1077</ul>
1078
1079</div>
1080
1081<!-- ======================================================================= -->
1082<div class="doc_subsection">
1083 <a name="ppc-be">Known problems with the PowerPC back-end</a>
1084</div>
1085
1086<div class="doc_text">
1087
1088<ul>
Nicolas Geoffray77d99502007-05-15 09:21:28 +0000[diff] [blame]1089<li>The Linux PPC32/ABI support needs testing for the interpreter and static
Chris Lattnerbee7b322007-05-23 04:39:32 +0000[diff] [blame]1090compilation, and lacks support for debug information.</li>
Chris Lattnerb81f10e2006-11-18 07:51:14 +0000[diff] [blame]1091</ul>
1092
1093</div>
1094
1095<!-- ======================================================================= -->
1096<div class="doc_subsection">
Chris Lattner97beb512007-05-14 06:56:09 +0000[diff] [blame]1097 <a name="arm-be">Known problems with the ARM back-end</a>
1098</div>
1099
1100<div class="doc_text">
1101
1102<ul>
Chris Lattnerbee7b322007-05-23 04:39:32 +0000[diff] [blame]1103<li>Thumb mode works only on ARMv6 or higher processors. On sub-ARMv6
Duncan Sands47fc0a22007-09-26 15:59:54 +0000[diff] [blame]1104processors, thumb programs can crash or produce wrong
Chris Lattnerbee7b322007-05-23 04:39:32 +0000[diff] [blame]1105results (<a href="http://llvm.org/PR1388">PR1388</a>).</li>
John Criswellea03c9d2009-03-02 15:28:15 +0000[diff] [blame]1106<li>Compilation for ARM Linux OABI (old ABI) is supported but not fully tested.
Chris Lattner97beb512007-05-14 06:56:09 +0000[diff] [blame]1107</li>
Chris Lattner97beb512007-05-14 06:56:09 +0000[diff] [blame]1108</ul>
1109
1110</div>
1111
1112<!-- ======================================================================= -->
1113<div class="doc_subsection">
Chris Lattnerb81f10e2006-11-18 07:51:14 +0000[diff] [blame]1114 <a name="sparc-be">Known problems with the SPARC back-end</a>
1115</div>
1116
1117<div class="doc_text">
1118
1119<ul>
John Criswellea03c9d2009-03-02 15:28:15 +0000[diff] [blame]1120<li>The SPARC backend only supports the 32-bit SPARC ABI (-m32); it does not
Chris Lattnerb81f10e2006-11-18 07:51:14 +0000[diff] [blame]1121 support the 64-bit SPARC ABI (-m64).</li>
1122</ul>
1123
1124</div>
1125
1126<!-- ======================================================================= -->
1127<div class="doc_subsection">
Bruno Cardoso Lopes24eb3de2008-10-25 14:56:26 +0000[diff] [blame]1128 <a name="mips-be">Known problems with the MIPS back-end</a>
1129</div>
1130
1131<div class="doc_text">
1132
1133<ul>
Bruno Cardoso Lopes24eb3de2008-10-25 14:56:26 +0000[diff] [blame]1134<li>64-bit MIPS targets are not supported yet.</li>
1135</ul>
1136
1137</div>
1138
1139<!-- ======================================================================= -->
1140<div class="doc_subsection">
Chris Lattnerb81f10e2006-11-18 07:51:14 +0000[diff] [blame]1141 <a name="alpha-be">Known problems with the Alpha back-end</a>
1142</div>
1143
1144<div class="doc_text">
1145
1146<ul>
1147
1148<li>On 21164s, some rare FP arithmetic sequences which may trap do not have the
1149appropriate nops inserted to ensure restartability.</li>
1150
John Criswell3bdbd302005-11-08 21:11:33 +0000[diff] [blame]1151</ul>
1152</div>
1153
Chris Lattnerb81f10e2006-11-18 07:51:14 +0000[diff] [blame]1154<!-- ======================================================================= -->
1155<div class="doc_subsection">
Chris Lattner97beb512007-05-14 06:56:09 +0000[diff] [blame]1156 <a name="c-be">Known problems with the C back-end</a>
Chris Lattnerb81f10e2006-11-18 07:51:14 +0000[diff] [blame]1157</div>
1158
1159<div class="doc_text">
1160
1161<ul>
Chris Lattnera1a4c9a2008-06-05 06:35:40 +0000[diff] [blame]1162<li><a href="http://llvm.org/PR802">The C backend has only basic support for
1163 inline assembly code</a>.</li>
Chris Lattner8e061162007-09-26 06:01:35 +0000[diff] [blame]1164<li><a href="http://llvm.org/PR1658">The C backend violates the ABI of common
1165 C++ programs</a>, preventing intermixing between C++ compiled by the CBE and
Gabor Greif75b2f762009-03-02 12:02:51 +0000[diff] [blame]1166 C++ code compiled with <tt>llc</tt> or native compilers.</li>
Duncan Sandse09506a2008-02-10 13:40:55 +0000[diff] [blame]1167<li>The C backend does not support all exception handling constructs.</li>
Duncan Sands3aa36732009-02-25 11:51:54 +0000[diff] [blame]1168<li>The C backend does not support arbitrary precision integers.</li>
Chris Lattnerb81f10e2006-11-18 07:51:14 +0000[diff] [blame]1169</ul>
1170
1171</div>
John Criswell3bdbd302005-11-08 21:11:33 +0000[diff] [blame]1172
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1173
1174<!-- ======================================================================= -->
1175<div class="doc_subsection">
Chris Lattner0b1c9a52010-03-17 04:41:49 +0000[diff] [blame]1176 <a name="c-fe">Known problems with the llvm-gcc C and C++ front-end</a>
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1177</div>
Chris Lattner178f3db2003-10-02 05:07:23 +0000[diff] [blame]1178
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1179<div class="doc_text">
Chris Lattner72a269f2006-03-03 00:34:26 +0000[diff] [blame]1180
Chris Lattnera1a4c9a2008-06-05 06:35:40 +0000[diff] [blame]1181<p>The only major language feature of GCC not supported by llvm-gcc is
1182 the <tt>__builtin_apply</tt> family of builtins. However, some extensions
1183 are only supported on some targets. For example, trampolines are only
Duncan Sands56448462008-06-08 20:18:35 +0000[diff] [blame]1184 supported on some targets (these are used when you take the address of a
1185 nested function).</p>
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1186
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1187</div>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1188
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1189<!-- ======================================================================= -->
1190<div class="doc_subsection">
Chris Lattner7292fb12008-10-30 03:58:13 +0000[diff] [blame]1191 <a name="fortran-fe">Known problems with the llvm-gcc Fortran front-end</a>
1192</div>
1193
1194<div class="doc_text">
Gabor Greife5f9d672008-11-04 21:50:59 +0000[diff] [blame]1195<ul>
Chris Lattner7292fb12008-10-30 03:58:13 +0000[diff] [blame]1196<li>Fortran support generally works, but there are still several unresolved bugs
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]1197 in <a href="http://llvm.org/bugs/">Bugzilla</a>. Please see the
1198 tools/gfortran component for details.</li>
Gabor Greife5f9d672008-11-04 21:50:59 +0000[diff] [blame]1199</ul>
Chris Lattner7292fb12008-10-30 03:58:13 +0000[diff] [blame]1200</div>
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1201
Chris Lattnere38ac152008-02-12 06:29:45 +0000[diff] [blame]1202<!-- ======================================================================= -->
1203<div class="doc_subsection">
Chris Lattnera1a4c9a2008-06-05 06:35:40 +0000[diff] [blame]1204 <a name="ada-fe">Known problems with the llvm-gcc Ada front-end</a>
Chris Lattnere38ac152008-02-12 06:29:45 +0000[diff] [blame]1205</div>
1206
1207<div class="doc_text">
John Criswellea03c9d2009-03-02 15:28:15 +0000[diff] [blame]1208The llvm-gcc 4.2 Ada compiler works fairly well; however, this is not a mature
1209technology, and problems should be expected.
Chris Lattnere38ac152008-02-12 06:29:45 +0000[diff] [blame]1210<ul>
Duncan Sands56448462008-06-08 20:18:35 +0000[diff] [blame]1211<li>The Ada front-end currently only builds on X86-32. This is mainly due
John Criswellea03c9d2009-03-02 15:28:15 +0000[diff] [blame]1212to lack of trampoline support (pointers to nested functions) on other platforms.
1213However, it <a href="http://llvm.org/PR2006">also fails to build on X86-64</a>
Chris Lattnere38ac152008-02-12 06:29:45 +0000[diff] [blame]1214which does support trampolines.</li>
1215<li>The Ada front-end <a href="http://llvm.org/PR2007">fails to bootstrap</a>.
Duncan Sandsd34bf762009-02-25 11:59:06 +0000[diff] [blame]1216This is due to lack of LLVM support for <tt>setjmp</tt>/<tt>longjmp</tt> style
1217exception handling, which is used internally by the compiler.
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]1218Workaround: configure with <tt>--disable-bootstrap</tt>.</li>
Duncan Sandsaef747f2008-10-13 17:27:23 +0000[diff] [blame]1219<li>The c380004, <a href="http://llvm.org/PR2010">c393010</a>
1220and <a href="http://llvm.org/PR2421">cxg2021</a> ACATS tests fail
Duncan Sandsd34bf762009-02-25 11:59:06 +0000[diff] [blame]1221(c380004 also fails with gcc-4.2 mainline).
1222If the compiler is built with checks disabled then <a href="http://llvm.org/PR2010">c393010</a>
1223causes the compiler to go into an infinite loop, using up all system memory.</li>
Duncan Sands2dbb00e2009-03-02 16:35:57 +0000[diff] [blame]1224<li>Some GCC specific Ada tests continue to crash the compiler.</li>
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]1225<li>The <tt>-E</tt> binder option (exception backtraces)
Chris Lattnere38ac152008-02-12 06:29:45 +0000[diff] [blame]1226<a href="http://llvm.org/PR1982">does not work</a> and will result in programs
Chris Lattneraa61f412009-10-13 17:48:04 +0000[diff] [blame]1227crashing if an exception is raised. Workaround: do not use <tt>-E</tt>.</li>
Chris Lattnere38ac152008-02-12 06:29:45 +0000[diff] [blame]1228<li>Only discrete types <a href="http://llvm.org/PR1981">are allowed to start
1229or finish at a non-byte offset</a> in a record. Workaround: do not pack records
1230or use representation clauses that result in a field of a non-discrete type
1231starting or finishing in the middle of a byte.</li>
Chris Lattnerb076b002008-06-08 21:19:07 +0000[diff] [blame]1232<li>The <tt>lli</tt> interpreter <a href="http://llvm.org/PR2009">considers
1233'main' as generated by the Ada binder to be invalid</a>.
1234Workaround: hand edit the file to use pointers for <tt>argv</tt> and
1235<tt>envp</tt> rather than integers.</li>
1236<li>The <tt>-fstack-check</tt> option <a href="http://llvm.org/PR2008">is
1237ignored</a>.</li>
Chris Lattnere38ac152008-02-12 06:29:45 +0000[diff] [blame]1238</ul>
1239</div>
1240
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1241<!-- *********************************************************************** -->
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1242<div class="doc_section">
1243 <a name="additionalinfo">Additional Information</a>
1244</div>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1245<!-- *********************************************************************** -->
1246
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1247<div class="doc_text">
1248
Chris Lattnercb5596d2005-05-16 17:13:10 +0000[diff] [blame]1249<p>A wide variety of additional information is available on the <a
Chris Lattnere0c1df42007-05-18 00:44:29 +0000[diff] [blame]1250href="http://llvm.org">LLVM web page</a>, in particular in the <a
1251href="http://llvm.org/docs/">documentation</a> section. The web page also
Reid Spencerc7f87f22007-07-09 08:04:31 +0000[diff] [blame]1252contains versions of the API documentation which is up-to-date with the
1253Subversion version of the source code.
Misha Brukman96158092005-03-30 19:14:24 +0000[diff] [blame]1254You can access versions of these documents specific to this release by going
1255into the "<tt>llvm/doc/</tt>" directory in the LLVM tree.</p>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1256
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1257<p>If you have any questions or comments about LLVM, please feel free to contact
Chris Lattnera69595e2005-10-29 07:07:09 +0000[diff] [blame]1258us via the <a href="http://llvm.org/docs/#maillist"> mailing
Chris Lattnerb84f3322003-12-12 21:22:16 +0000[diff] [blame]1259lists</a>.</p>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1260
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1261</div>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1262
1263<!-- *********************************************************************** -->
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1264
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1265<hr>
Misha Brukman68aab3b2003-11-22 01:23:39 +0000[diff] [blame]1266<address>
Misha Brukman8834af92003-12-21 22:53:21 +0000[diff] [blame]1267 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +0000[diff] [blame]1268 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukman68aab3b2003-11-22 01:23:39 +0000[diff] [blame]1269 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +0000[diff] [blame]1270 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Misha Brukman68aab3b2003-11-22 01:23:39 +0000[diff] [blame]1271
Chris Lattnere0c1df42007-05-18 00:44:29 +0000[diff] [blame]1272 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1273 Last modified: $Date$
Misha Brukman68aab3b2003-11-22 01:23:39 +0000[diff] [blame]1274</address>
Chris Lattner3d482502003-10-02 04:57:28 +0000[diff] [blame]1275
Misha Brukman80731b92003-11-22 00:38:41 +0000[diff] [blame]1276</body>
1277</html>