Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / f3e35782a62cc388e267390ebd0007905fe3b694 / . / docs / ReleaseNotes.html
blob: 95674d73db80a29d800e28720d414ffc5db35f00 [file] [log] [blame]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<html>
4<head>
5 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
6 <link rel="stylesheet" href="llvm.css" type="text/css">
Chris Lattner2c3ef982008-10-11 18:18:28 +0000[diff] [blame]7 <title>LLVM 2.4 Release Notes</title>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]8</head>
9<body>
10
Chris Lattner2c3ef982008-10-11 18:18:28 +0000[diff] [blame]11<div class="doc_title">LLVM 2.4 Release Notes</div>
Mikhail Glushenkov808cd3e2008-10-13 02:08:34 +0000[diff] [blame]12
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]13<ol>
14 <li><a href="#intro">Introduction</a></li>
Chris Lattner78a130b2008-10-13 18:01:01 +0000[diff] [blame]15 <li><a href="#subproj">Sub-project Status Update</a></li>
16 <li><a href="#whatsnew">What's New in LLVM?</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]17 <li><a href="GettingStarted.html">Installation Instructions</a></li>
18 <li><a href="#portability">Portability and Supported Platforms</a></li>
Dan Gohman0de549c2008-10-14 16:23:02 +0000[diff] [blame]19 <li><a href="#knownproblems">Known Problems</a></li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]20 <li><a href="#additionalinfo">Additional Information</a></li>
21</ol>
22
23<div class="doc_author">
Dan Gohman0de549c2008-10-14 16:23:02 +0000[diff] [blame]24 <p>Written by the <a href="http://llvm.org">LLVM Team</a></p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]25</div>
26
27<!-- *********************************************************************** -->
28<div class="doc_section">
29 <a name="intro">Introduction</a>
30</div>
31<!-- *********************************************************************** -->
32
33<div class="doc_text">
34
Chris Lattner92eacf82008-10-13 17:57:36 +0000[diff] [blame]35<p>This document contains the release notes for the LLVM Compiler
36Infrastructure, release 2.4. Here we describe the status of LLVM, including
37major improvements from the previous release and significant known problems.
38All LLVM releases may be downloaded from the <a
39href="http://llvm.org/releases/">LLVM releases web site</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]40
41<p>For more information about LLVM, including information about the latest
42release, please check out the <a href="http://llvm.org/">main LLVM
43web site</a>. If you have questions or comments, the <a
Chris Lattner92eacf82008-10-13 17:57:36 +0000[diff] [blame]44href="http://mail.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVM Developer's Mailing
45List</a> is a good place to send them.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]46
Mikhail Glushenkov808cd3e2008-10-13 02:08:34 +0000[diff] [blame]47<p>Note that if you are reading this file from a Subversion checkout or the
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]48main LLVM web page, this document applies to the <i>next</i> release, not the
Gabor Greif7c65c4f2008-10-14 11:00:32 +0000[diff] [blame]49current one. To see the release notes for a specific release, please see the
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]50<a href="http://llvm.org/releases/">releases page</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]51
52</div>
53
Chris Lattner2c3ef982008-10-11 18:18:28 +0000[diff] [blame]54<!-- Unfinished features in 2.4:
Chris Lattner8fda4952008-02-10 07:46:44 +0000[diff] [blame]55 Machine LICM
56 Machine Sinking
57 LegalizeDAGTypes
Chris Lattner4f076402008-10-12 19:47:48 +0000[diff] [blame]58 llc -enable-value-prop, propagation of value info (sign/zero ext info) from
59 one MBB to another
Chris Lattner8fda4952008-02-10 07:46:44 +0000[diff] [blame]60 -->
Mikhail Glushenkov808cd3e2008-10-13 02:08:34 +0000[diff] [blame]61
Chris Lattner4f076402008-10-12 19:47:48 +0000[diff] [blame]62 <!-- for announcement email:
63 mention dev mtg
Chris Lattner548493b2008-10-12 22:57:58 +0000[diff] [blame]64 Xcode 3.1 and 3.1.1.
Chris Lattner4f076402008-10-12 19:47:48 +0000[diff] [blame]65 -->
Chris Lattner8fda4952008-02-10 07:46:44 +0000[diff] [blame]66
Chris Lattner78a130b2008-10-13 18:01:01 +0000[diff] [blame]67<!-- *********************************************************************** -->
68<div class="doc_section">
69 <a name="subproj">Sub-project Status Update</a>
Chris Lattnerf65224f2008-06-08 21:34:41 +0000[diff] [blame]70</div>
Chris Lattner78a130b2008-10-13 18:01:01 +0000[diff] [blame]71<!-- *********************************************************************** -->
Chris Lattnerf65224f2008-06-08 21:34:41 +0000[diff] [blame]72
73<div class="doc_text">
Chris Lattnerf6a0d382008-06-08 21:58:17 +0000[diff] [blame]74<p>
Chris Lattner78a130b2008-10-13 18:01:01 +0000[diff] [blame]75The LLVM 2.4 distribution currently consists of code from the core LLVM
Chris Lattnera3c683b2008-10-13 18:11:54 +0000[diff] [blame]76repository (which roughly includes the LLVM optimizers, code generators and
Chris Lattnerf6a0d382008-06-08 21:58:17 +0000[diff] [blame]77supporting tools) and the llvm-gcc repository. In addition to this code, the
78LLVM Project includes other sub-projects that are in development. The two which
Chris Lattner2c3ef982008-10-11 18:18:28 +0000[diff] [blame]79are the most actively developed are the <a href="#clang">Clang Project</a> and
Nicolas Geoffray2b139382008-10-14 19:23:04 +0000[diff] [blame]80the <a href="#vmkit">VMKit Project</a>.
Chris Lattnerf6a0d382008-06-08 21:58:17 +0000[diff] [blame]81</p>
Chris Lattner78a130b2008-10-13 18:01:01 +0000[diff] [blame]82
Chris Lattnerf6a0d382008-06-08 21:58:17 +0000[diff] [blame]83</div>
84
Chris Lattner78a130b2008-10-13 18:01:01 +0000[diff] [blame]85
Chris Lattnerf6a0d382008-06-08 21:58:17 +0000[diff] [blame]86<!--=========================================================================-->
Chris Lattner78a130b2008-10-13 18:01:01 +0000[diff] [blame]87<div class="doc_subsection">
Chris Lattnera3c683b2008-10-13 18:11:54 +0000[diff] [blame]88<a name="clang">Clang: C/C++/Objective-C Frontend Toolkit</a>
Chris Lattnerf6a0d382008-06-08 21:58:17 +0000[diff] [blame]89</div>
90
91<div class="doc_text">
92
93<p>The <a href="http://clang.llvm.org/">Clang project</a> is an effort to build
94a set of new 'LLVM native' front-end technologies for the LLVM optimizer
95and code generator. Clang is continuing to make major strides forward in all
96areas. Its C and Objective-C parsing support is very solid, and the code
97generation support is far enough along to build many C applications. While not
98yet production quality, it is progressing very nicely. In addition, C++
99front-end work has started to make significant progress.</p>
100
Daniel Dunbarf3e35782008-10-14 23:25:09 +0000[diff] [blame^]101Clang, in conjunction with the <tt>ccc</tt> driver, is now usable as a
102replacement for gcc for building some small- to medium-sized C applications.
103Additionally, Clang now has code generation support for Objective-C on Mac OS X
104platform. Major highlights include:
105<ul>
106 <li> Clang/ccc pass almost all of the LLVM test suite on Mac OS X and Linux
107on the 32-bit x86 architecture. This includes significant C
108applications such as <a href="http://www.sqlite.org">sqlite3</a>,
109<a href="http://www.lua.org">lua</a>, and
110<a href="http://www.clamav.net">Clam AntiVirus</a>.
111
112 <li> Clang can build the majority of Objective-C examples shipped with the
113Mac OS X Developer Tools.
114</ul>
115
116Clang code generation still needs considerable testing and development, however.
117Some areas under active development include:
118<ul>
119 <li> Improved support for C and Objective-C features, for example
120 variable-length arries, va_arg, exception handling (Obj-C), and garbage
121 collection (Obj-C).
122 <li> ABI compatibility, especially for platforms other than 32-bit x86.
123</ul>
Chris Lattner2c3ef982008-10-11 18:18:28 +0000[diff] [blame]124
Chris Lattnera3c683b2008-10-13 18:11:54 +0000[diff] [blame]125</div>
126
127<!--=========================================================================-->
128<div class="doc_subsection">
129<a name="clangsa">Clang Static Analyzer</a>
130</div>
131
132<div class="doc_text">
133
Ted Kremenek0437b2c2008-10-14 05:14:21 +0000[diff] [blame]134<p>The Clang project also includes an early stage static source code analysis
135tool for <a href="http://clang.llvm.org/StaticAnalysis.html">automatically
136finding bugs</a> in C and Objective-C programs. The tool performs a growing set
Chris Lattnerff170152008-10-14 06:02:29 +0000[diff] [blame]137of checks to find bugs that occur on a specific path within a program. Examples
Ted Kremenek0437b2c2008-10-14 05:14:21 +0000[diff] [blame]138of bugs the tool finds include logic errors such as null dereferences,
139violations of various API rules, dead code, and potential memory leaks in
Chris Lattnerff170152008-10-14 06:02:29 +0000[diff] [blame]140Objective-C programs. Since its inception, public feedback on the tool has been
141extremely positive, and conservative estimates put the number of real bugs it
142has found in industrial-quality software on the order of thousands.</p>
Chris Lattnerc2d84672008-10-14 00:52:49 +0000[diff] [blame]143
Ted Kremenek0437b2c2008-10-14 05:14:21 +0000[diff] [blame]144<p>The tool also provides a simple web GUI to inspect potential bugs found by
Chris Lattnerff170152008-10-14 06:02:29 +0000[diff] [blame]145the tool. While still early in development, the GUI illustrates some of the key
Ted Kremenek0437b2c2008-10-14 05:14:21 +0000[diff] [blame]146features of Clang: accurate source location information, which is used by the
147GUI to highlight specific code expressions that relate to a bug (including those
148that span multiple lines) and built-in knowledge of macros, which is used to
149perform inline expansion of macros within the GUI itself.</p>
150
Chris Lattnerff170152008-10-14 06:02:29 +0000[diff] [blame]151<p>The set of checks performed by the static analyzer is gradually expanding,
152and
Ted Kremenek0437b2c2008-10-14 05:14:21 +0000[diff] [blame]153future plans for the tool include full source-level inter-procedural analysis
154and deeper checks such as buffer overrun detection. There are many opportunities
155to extend and enhance the static analyzer, and anyone interested in working on
156this project is encouraged to get involved!</p>
Chris Lattnerf6a0d382008-06-08 21:58:17 +0000[diff] [blame]157
158</div>
159
Chris Lattner78a130b2008-10-13 18:01:01 +0000[diff] [blame]160<!--=========================================================================-->
161<div class="doc_subsection">
Nicolas Geoffray2b139382008-10-14 19:23:04 +0000[diff] [blame]162<a name="vmkit">VMKit: JVM/CLI Virtual Machine Implementation</a>
Chris Lattner78a130b2008-10-13 18:01:01 +0000[diff] [blame]163</div>
164
165<div class="doc_text">
166<p>
Nicolas Geoffray2b139382008-10-14 19:23:04 +0000[diff] [blame]167The <a href="http://vmkit.llvm.org/">VMKit project</a> is an implementation of
Chris Lattner78a130b2008-10-13 18:01:01 +0000[diff] [blame]168a JVM and a CLI Virtual Machines (Microsoft .NET is an
169implementation of the CLI) using the Just-In-Time compiler of LLVM.</p>
170
Nicolas Geoffray2b139382008-10-14 19:23:04 +0000[diff] [blame]171<p>Following LLVM 2.4, VMKit has its first release 0.24 that you can find on
172the <a href="http://llvm.org/releases/">release page</a>. The release includes
173bug fixes, cleanup and new features. The major changes include:</p>
Chris Lattner78a130b2008-10-13 18:01:01 +0000[diff] [blame]174
Nicolas Geoffray2b139382008-10-14 19:23:04 +0000[diff] [blame]175<ul>
176
177<li><p>Support for generics in the .Net virtual machine. This was implemented
178by Tilmann Scheller during his Google Summer of Code project.</li></p>
179<li><p>Initial support for the Mono class libraries.</p></li>
180<li><p>Support for MacOSX/x86, following LLVM's support for exceptions in
181JIT on MacOSX/x86.
182<li><p>A new vmkit driver: a program to run java or .net applications. The
183driver supports llvm command line arguments including the new "-fast" option.
184</p></li>
185<li><p>A new memory allocation scheme in the JVM that makes unloading a
186class loader very fast.</p></li>
187<li><p>VMKit now follows the LLVM Makefile machinery.</p></li>
188
189
190</ul>
Chris Lattner78a130b2008-10-13 18:01:01 +0000[diff] [blame]191</div>
192
Chris Lattnerf6a0d382008-06-08 21:58:17 +0000[diff] [blame]193
194<!-- *********************************************************************** -->
195<div class="doc_section">
Chris Lattner78a130b2008-10-13 18:01:01 +0000[diff] [blame]196 <a name="whatsnew">What's New in LLVM?</a>
Chris Lattnerf6a0d382008-06-08 21:58:17 +0000[diff] [blame]197</div>
198<!-- *********************************************************************** -->
199
Chris Lattner5cce7a52008-06-08 22:59:35 +0000[diff] [blame]200<div class="doc_text">
201
Chris Lattner92eacf82008-10-13 17:57:36 +0000[diff] [blame]202<p>This release includes a huge number of bug fixes, performance tweaks and
203minor improvements. Some of the major improvements and new features are listed
204in this section.
Chris Lattner5cce7a52008-06-08 22:59:35 +0000[diff] [blame]205</p>
206</div>
207
Chris Lattnerf6a0d382008-06-08 21:58:17 +0000[diff] [blame]208<!--=========================================================================-->
Chris Lattnerf65224f2008-06-08 21:34:41 +0000[diff] [blame]209<div class="doc_subsection">
Chris Lattnere0798082008-02-06 18:00:06 +0000[diff] [blame]210<a name="majorfeatures">Major New Features</a>
211</div>
212
213<div class="doc_text">
214
Chris Lattner2c3ef982008-10-11 18:18:28 +0000[diff] [blame]215<p>LLVM 2.4 includes several major new capabilities:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]216
217<ul>
Chris Lattner4352cc82008-10-13 20:47:20 +0000[diff] [blame]218<li><p>The most visible end-user change in LLVM 2.4 is that it includes many
Chris Lattnerea9c2242008-10-13 21:50:36 +0000[diff] [blame]219optimizations and changes to make -O0 compile times much faster. You should see
Gabor Greif7c65c4f2008-10-14 11:00:32 +0000[diff] [blame]220improvements on the order of 30% (or more) faster than LLVM 2.3. There are many
Chris Lattner4352cc82008-10-13 20:47:20 +0000[diff] [blame]221pieces to this change, described in more detail below. The speedups and new
222components can also be used for JIT compilers that want fast compilation as
223well.</p></li>
Chris Lattner2c3ef982008-10-11 18:18:28 +0000[diff] [blame]224
Chris Lattner4352cc82008-10-13 20:47:20 +0000[diff] [blame]225<li><p>The biggest change to the LLVM IR is that Multiple Return Values (which
226were introduced in LLVM 2.3) have been generalized to full support for "First
227Class Aggregate" values in LLVM 2.4. This means that LLVM IR supports using
228structs and arrays as values in a function. This capability is mostly useful
229for front-end authors, who prefer to treat things like complex numbers, simple
230tuples, dope vectors, etc as Value*'s instead of as a tuple of Value*'s or as
Gabor Greif7c65c4f2008-10-14 11:00:32 +0000[diff] [blame]231memory values. Bitcode files from LLVM 2.3 will automatically migrate to the
232general representation.</p></li>
Mikhail Glushenkov808cd3e2008-10-13 02:08:34 +0000[diff] [blame]233
Chris Lattner4352cc82008-10-13 20:47:20 +0000[diff] [blame]234<li><p>LLVM 2.4 also includes an initial port for the PIC16 microprocessor. This
Duncan Sandsc16c9242008-10-14 07:06:37 +0000[diff] [blame]235is the LLVM target that only has support for 8 bit registers, and a number of
Chris Lattner4352cc82008-10-13 20:47:20 +0000[diff] [blame]236other crazy constraints. While the port is still in early development stages,
237it shows some interesting things you can do with LLVM.</p></li>
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]238
Chris Lattner5ba16862008-02-10 08:18:42 +0000[diff] [blame]239</ul>
240
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]241</div>
242
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]243
244<!--=========================================================================-->
245<div class="doc_subsection">
Chris Lattnerf65224f2008-06-08 21:34:41 +0000[diff] [blame]246<a name="llvm-gcc">llvm-gcc 4.2 Improvements</a>
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]247</div>
248
249<div class="doc_text">
250
Chris Lattner4352cc82008-10-13 20:47:20 +0000[diff] [blame]251<p>LLVM fully supports the llvm-gcc 4.2 front-end, which marries the GCC
252front-ends and driver with the LLVM optimizer and code generator. It currently
253includes support for the C, C++, Objective-C, Ada, and Fortran front-ends.</p>
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]254
Chris Lattner5cce7a52008-06-08 22:59:35 +0000[diff] [blame]255<ul>
Chris Lattnerf7bae6d2008-10-13 20:59:32 +0000[diff] [blame]256<li>LLVM 2.4 supports the full set of atomic <tt>__sync_*</tt> builtins. LLVM
2572.3 only supported those used by OpenMP, but 2.4 supports them all. While
Chris Lattnerea9c2242008-10-13 21:50:36 +0000[diff] [blame]258llvm-gcc supports all of these builtins, note that not all targets do. X86
259support them all in both 32-bit and 64-bit mode and PowerPC supports them all
260except for the 64-bit operations when in 32-bit mode.</li>
Chris Lattner5cce7a52008-06-08 22:59:35 +0000[diff] [blame]261
Chris Lattnerf7bae6d2008-10-13 20:59:32 +0000[diff] [blame]262<li>llvm-gcc now supports an <tt>-flimited-precision</tt> option, which tells
263the compiler that it is ok to use low-precision approximations of certain libm
Chris Lattner4352cc82008-10-13 20:47:20 +0000[diff] [blame]264functions (like tan, log, etc). This allows you to get high performance if you
265only need (say) 14-bits of precision.</li>
266
267<li>llvm-gcc now supports a C language extension known as "<a
268href="http://lists.cs.uiuc.edu/pipermail/cfe-dev/2008-August/002670.html">Blocks
Gabor Greif7c65c4f2008-10-14 11:00:32 +0000[diff] [blame]269</a>". This feature is similar to nested functions and closures, but does not
Chris Lattner4352cc82008-10-13 20:47:20 +0000[diff] [blame]270require stack trampolines (with most ABIs) and supports returning closures
271from functions that define them. Note that actually <em>using</em> Blocks
272requires a small runtime that is not included with llvm-gcc.</li>
273
274<li>llvm-gcc now supports a new <tt>-flto</tt> option. On systems that support
275transparent Link Time Optimization (currently Darwin systems with Xcode 3.1 and
276later) this allows the use of LTO with other optimization levels like -Os.
277Previously, LTO could only be used with -O4, which implied optimizations in
278-O3 that can increase code size.</li>
Gabor Greif0e49cc82008-06-09 06:06:18 +0000[diff] [blame]279</ul>
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]280
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]281</div>
282
283
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]284<!--=========================================================================-->
285<div class="doc_subsection">
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]286<a name="coreimprovements">LLVM Core Improvements</a>
287</div>
288
289<div class="doc_text">
290<p>New features include:
291</p>
292
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]293<ul>
Gabor Greif7c65c4f2008-10-14 11:00:32 +0000[diff] [blame]294<li>A major change to the <tt>Use</tt> class landed, which shrank it by 25%. Since
Chris Lattnerf7bae6d2008-10-13 20:59:32 +0000[diff] [blame]295this is a pervasive part of the LLVM, it ended up reducing the memory use of
296LLVM IR in general by 15% for most programs.</li>
297
298<li>Values with no names are now pretty printed by <tt>llvm-dis</tt> more
299nicely. They now print as "<tt>%3 = add i32 %A, 4</tt>" instead of
300"<tt>add i32 %A, 4 ; &lt;i32&gt;:3</tt>", which makes it much easier to read.
301</li>
302
303<li>LLVM 2.4 includes some changes for better vector support. First, the shift
304operations (<tt>shl</tt>, <tt>ashr</tt>, <tt>lshr</tt>) now all support vectors
305and do an element-by-element shift (shifts of the whole vector can be
306accomplished by bitcasting the vector to &lt;1 x i128&gt; for example). Second,
Chris Lattnerc2d84672008-10-14 00:52:49 +0000[diff] [blame]307there is initial support in development for vector comparisons with the
308<a href="LangRef.html#i_fcmp">fcmp</a>/<a href="LangRef.html#i_icmp">icmp</a>
309instructions. These instructions compare two vectors and return a vector of
310i1's for each result. Note that there is very little codegen support available
311for any of these IR features though.</li>
Chris Lattnerf7bae6d2008-10-13 20:59:32 +0000[diff] [blame]312
313<li>A new <tt>DebugInfoBuilder</tt> class is available, which makes it much
314easier for front-ends to create debug info descriptors, similar to the way that
Gabor Greifaa53e562008-10-14 11:31:14 +0000[diff] [blame]315<tt>IRBuilder</tt> makes it easier to create LLVM IR.</li>
Chris Lattnerf7bae6d2008-10-13 20:59:32 +0000[diff] [blame]316
Dan Gohmand74867f2008-10-14 16:13:59 +0000[diff] [blame]317<li>The <tt>IRBuilder</tt> class is now parameterized by a class responsible
Gabor Greif7c65c4f2008-10-14 11:00:32 +0000[diff] [blame]318for constant folding. The default <tt>ConstantFolder</tt> class does target independent
319constant folding. The <tt>NoFolder</tt> class does no constant folding at all, which is
320useful when learning how LLVM works. The <tt>TargetFolder</tt> class folds the most,
Duncan Sandsc16c9242008-10-14 07:06:37 +0000[diff] [blame]321doing target dependent constant folding.</li>
322
323<li>LLVM now supports "function attributes", which allows us to separate return
Chris Lattnerf7bae6d2008-10-13 20:59:32 +0000[diff] [blame]324value attributes from function attributes. LLVM now supports attributes on a
325function itself, a return value, and its parameters. New supported function
326attributes include noinline/alwaysinline and the "opt-size" flag which says the
327function should be optimized for code size.</li>
328
Chris Lattnera5a49382008-10-12 18:30:33 +0000[diff] [blame]329<li>LLVM IR now directly represents "common" linkage, instead of
330 representing it as a form of weak linkage.</li>
Chris Lattnerf7bae6d2008-10-13 20:59:32 +0000[diff] [blame]331
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]332</ul>
Mikhail Glushenkov808cd3e2008-10-13 02:08:34 +0000[diff] [blame]333
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]334</div>
335
336<!--=========================================================================-->
337<div class="doc_subsection">
Chris Lattner8fda4952008-02-10 07:46:44 +0000[diff] [blame]338<a name="optimizer">Optimizer Improvements</a>
339</div>
340
341<div class="doc_text">
342
Chris Lattnerea9c2242008-10-13 21:50:36 +0000[diff] [blame]343<p>In addition to a huge array of bug fixes and minor performance tweaks, this
344release includes a few major enhancements and additions to the optimizers:</p>
Chris Lattner8fda4952008-02-10 07:46:44 +0000[diff] [blame]345
346<ul>
347
Chris Lattnerea9c2242008-10-13 21:50:36 +0000[diff] [blame]348<li>The Global Value Numbering (GVN) pass now does local Partial Redundancy
349Elimination (PRE) to eliminate some partially redundant expressions in cases
350where doing so won't grow code size.</li>
Chris Lattner4f076402008-10-12 19:47:48 +0000[diff] [blame]351
Chris Lattnerea9c2242008-10-13 21:50:36 +0000[diff] [blame]352<li>LLVM 2.4 includes a new loop deletion pass (which removes output-free
353provably-finite loops) and a rewritten Aggressive Dead Code Elimination (ADCE)
354pass that no longer uses control dependence information. These changes speed up
Gabor Greifaa53e562008-10-14 11:31:14 +0000[diff] [blame]355the optimizer and also prevent it from deleting output-free infinite
Chris Lattnerea9c2242008-10-13 21:50:36 +0000[diff] [blame]356loops.</li>
Chris Lattner4f076402008-10-12 19:47:48 +0000[diff] [blame]357
Duncan Sands2aecc3f2008-10-13 18:42:43 +0000[diff] [blame]358<li>The new AddReadAttrs pass works out which functions are read-only or
Gabor Greifaa53e562008-10-14 11:31:14 +0000[diff] [blame]359read-none (these correspond to 'pure' and 'const' in GCC) and marks them
Duncan Sands2aecc3f2008-10-13 18:42:43 +0000[diff] [blame]360with the appropriate attribute.</li>
Chris Lattnerea9c2242008-10-13 21:50:36 +0000[diff] [blame]361
362<li>LLVM 2.4 now includes a new SparsePropagation framework, which makes it
363trivial to build lattice-based dataflow solvers that operate over LLVM IR. Using
364this interface means that you just define objects to represent your lattice
365values and the transfer functions that operate on them. It handles the
366mechanics of worklist processing, liveness tracking, handling PHI nodes,
367etc.</li>
368
Chris Lattnerff170152008-10-14 06:02:29 +0000[diff] [blame]369<li>The Loop Strength Reduction and induction variable optimization passes have
370several improvements to avoid inserting MAX expressions, to optimize simple
371floating point induction variables and to analyze trip counts of more
372loops.</li>
373
Chris Lattnerea9c2242008-10-13 21:50:36 +0000[diff] [blame]374<li>Various helper functions (ComputeMaskedBits, ComputeNumSignBits, etc) were
375pulled out of the Instruction Combining pass and put into a new
376<tt>ValueTracking.h</tt> header, where they can be reused by other passes.</li>
377
378<li>The tail duplication pass has been removed from the standard optimizer
379sequence used by llvm-gcc. This pass still exists, but the benefits it once
380provided are now achieved by other passes.</li>
381
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]382</ul>
383
384</div>
385
386<!--=========================================================================-->
387<div class="doc_subsection">
388<a name="codegen">Code Generator Improvements</a>
389</div>
390
391<div class="doc_text">
392
Gabor Greifaa53e562008-10-14 11:31:14 +0000[diff] [blame]393<p>We have put a significant amount of work into the code generator infrastructure,
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]394which allows us to implement more aggressive algorithms and make it run
395faster:</p>
396
397<ul>
Chris Lattnerea9c2242008-10-13 21:50:36 +0000[diff] [blame]398<li>The target-independent code generator supports (and the X86 backend
399 currently implements) a new interface for "fast" instruction selection. This
400 interface is optimized to produce code as quickly as possible, sacrificing
401 code quality to do it. This is used by default at -O0 or when using
402 "llc -fast" on X86. It is straight-forward to add support for
403 other targets if faster -O0 compilation is desired.</li>
404
405<li>In addition to the new 'fast' instruction selection path, many existing
406 pieces of the code generator have been optimized in significant ways.
407 SelectionDAG's are now pool allocated and use better algorithms in many
Gabor Greifaa53e562008-10-14 11:31:14 +0000[diff] [blame]408 places, the ".s" file printers now use <tt>raw_ostream</tt> to emit text much faster,
Chris Lattnerea9c2242008-10-13 21:50:36 +0000[diff] [blame]409 etc. The end result of these improvements is that the compiler also takes
410 substantially less time to generate code that is just as good (and often
411 better) than before.</li>
412
Gabor Greifaa53e562008-10-14 11:31:14 +0000[diff] [blame]413<li>Each target has been split to separate the ".s" file printing logic from the
Chris Lattnerea9c2242008-10-13 21:50:36 +0000[diff] [blame]414 rest of the target. This enables JIT compilers that don't link in the
Gabor Greifaa53e562008-10-14 11:31:14 +0000[diff] [blame]415 (somewhat large) code and data tables used for printing a ".s" file.</li>
Chris Lattnerea9c2242008-10-13 21:50:36 +0000[diff] [blame]416
417<li>The code generator now includes a "stack slot coloring" pass, which packs
418 together individual spilled values into common stack slots. This reduces
419 the size of stack frames with many spills, which tends to increase L1 cache
420 effectiveness.</li>
421
422<li>Various pieces of the register allocator (e.g. the coalescer and two-address
423 operation elimination pass) now know how to rematerialize trivial operations
424 to avoid copies and include several other optimizations.</li>
425
426<li>The <a href="CodeGenerator.html#selectiondag_process">graphs</a> produced by
427 the <tt>llc -view-*-dags</tt> options are now significantly prettier and
428 easier to read.</li>
429
430<li>LLVM 2.4 includes a new register allocator based on Partitioned Boolean
431 Quadratic Programming (PBQP). This register allocator is still in
432 development, but is very simple and clean.</li>
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]433
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]434</ul>
435
436</div>
437
438
439<!--=========================================================================-->
440<div class="doc_subsection">
Chris Lattnercc64a322008-10-13 22:06:31 +0000[diff] [blame]441<a name="targetspecific">Target Specific Improvements</a>
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]442</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]443
444<div class="doc_text">
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]445<p>New target-specific features include:
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]446</p>
447
448<ul>
Chris Lattner4f076402008-10-12 19:47:48 +0000[diff] [blame]449<li>Exception handling is supported by default on Linux/x86-64.</li>
Duncan Sandsd8745a92008-10-13 17:32:15 +0000[diff] [blame]450<li>Position Independent Code (PIC) is now supported on Linux/x86-64.</li>
Chris Lattnerff170152008-10-14 06:02:29 +0000[diff] [blame]451<li>@llvm.frameaddress now supports getting the frame address of stack frames
452 &gt; 0 on x86/x86-64.</li>
Chris Lattnercc64a322008-10-13 22:06:31 +0000[diff] [blame]453<li>MIPS floating point support? [BRUNO]</li>
454<li>The PowerPC backend now supports trampolines.</li>
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]455</ul>
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]456
Mikhail Glushenkov808cd3e2008-10-13 02:08:34 +0000[diff] [blame]457</div>
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]458
459
460<!--=========================================================================-->
461<div class="doc_subsection">
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]462<a name="otherimprovements">Other Improvements</a>
463</div>
464
465<div class="doc_text">
466<p>New features include:
467</p>
468
469<ul>
Mikhail Glushenkovcc2d3012008-10-13 02:46:01 +0000[diff] [blame]470<li><tt>llvmc2</tt> (the generic compiler driver) gained plugin
Chris Lattnercc64a322008-10-13 22:06:31 +0000[diff] [blame]471 support. It is now easier to experiment with <tt>llvmc2</tt> and
472 build your own tools based on it.</li>
473
474<li>LLVM 2.4 includes a number of new generic algorithms and data structures,
Dan Gohman8d9af692008-10-14 15:14:55 +0000[diff] [blame]475 include a scoped hash table, 'immutable' data structures, a simple
476 free-list manager, and a <tt>raw_ostream</tt> class.
477 The <tt>raw_ostream</tt> class and
Chris Lattnercc64a322008-10-13 22:06:31 +0000[diff] [blame]478 <tt>format</tt> allow for efficient file output, and various pieces of LLVM
479 have switched over to use it. The eventual goal is to eliminate
480 std::ostream in favor of it.</li>
481
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]482</ul>
Mikhail Glushenkov808cd3e2008-10-13 02:08:34 +0000[diff] [blame]483
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]484</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]485
Chris Lattner92eacf82008-10-13 17:57:36 +0000[diff] [blame]486<!--=========================================================================-->
487<div class="doc_subsection">
488<a name="changes">Major Changes and Removed Features</a>
489</div>
490
491<div class="doc_text">
492
Chris Lattnercc64a322008-10-13 22:06:31 +0000[diff] [blame]493<p>If you're already an LLVM user or developer with out-of-tree changes based
494on LLVM 2.3, this section lists some "gotchas" that you may run into upgrading
495from the previous release.</p>
Chris Lattner92eacf82008-10-13 17:57:36 +0000[diff] [blame]496
497<ul>
498
Chris Lattnercc64a322008-10-13 22:06:31 +0000[diff] [blame]499<li>The LLVM IR generated by llvm-gcc no longer names all instructions. This
500 makes it run faster, but may be more confusing to some people. If you
501 prefer to have names, the '<tt>opt -instnamer</tt>' pass will add names to
502 all instructions.</li>
503
504<li>The LoadVN and GCSE passes have been removed from the tree. They are
505 obsolete and have been replaced with the GVN and MemoryDependence passes.
506 </li>
507</ul>
508
509
510<p>In addition, many APIs have changed in this release. Some of the major LLVM
511API changes are:</p>
512
513<ul>
514
Devang Patel0e3da1a2008-10-14 20:03:43 +0000[diff] [blame]515<li>Now, function attributes and return value attributes are managed
516sepearately. Interface exported by <tt>ParameterAttributes.h</tt> header is now
517experted by <tt>Attributes.h</tt> header. The new attributes interface changes are:
518<ul>
519<li><tt>getParamAttrs</tt> method is now replaced by
520<tt>getParamAttributes</tt>, <tt>getRetAttributes</tt> and
521<tt>getFnAttributes</tt> methods.</li>
522<li> Return value attributes are stored at index 0. Function attributes are
523stored at index ~0U. Parameter attributes are stored at index that matches
524parameter number.</li>
525<li> <tt>ParamAttr</tt> namespace is now renamed as <tt>Attribute</tt>.</li>
526<li> The name of the class</tt> that manages reference count of opaque
527attributes is changed from <tt>PAListPtr</tt> to <tt>AttrListPtr</tt>.</li>
528<li> <tt>ParamAttrsWithIndex</tt> is now renamed as <tt>AttributeWithIndex</tt>.
529</li>
530</ul>
531</li>
Chris Lattner92eacf82008-10-13 17:57:36 +0000[diff] [blame]532
533<li>The <tt>DbgStopPointInst</tt> methods <tt>getDirectory</tt> and
534<tt>getFileName</tt> now return <tt>Value*</tt> instead of strings. These can be
535converted to strings using <tt>llvm::GetConstantStringInfo</tt> defined via
Gabor Greifaa53e562008-10-14 11:31:14 +0000[diff] [blame]536"<tt>llvm/Analysis/ValueTracking.h</tt>".</li>
Chris Lattner92eacf82008-10-13 17:57:36 +0000[diff] [blame]537
Chris Lattnercc64a322008-10-13 22:06:31 +0000[diff] [blame]538<li>The APIs to create various instructions have changed from lower case
539 "create" methods to upper case "Create" methods (e.g.
540 <tt>BinaryOperator::create</tt>). LLVM 2.4 includes both cases, but the
541 lower case ones are removed in mainline, please migrate.</li>
542
Gabor Greifaa53e562008-10-14 11:31:14 +0000[diff] [blame]543<li>Various header files like "<tt>llvm/ADT/iterator</tt>" were given a ".h" suffix.
544 Change your code to #include "<tt>llvm/ADT/iterator.h</tt>" instead.</li>
Chris Lattner92eacf82008-10-13 17:57:36 +0000[diff] [blame]545
Gabor Greifaa53e562008-10-14 11:31:14 +0000[diff] [blame]546<li>In the code generator, many <tt>MachineOperand</tt> predicates were renamed to be
Chris Lattnerb7b69332008-10-14 06:37:11 +0000[diff] [blame]547 shorter (e.g. <tt>isFrameIndex()</tt> -&gt; <tt>isFI()</tt>),
548 <tt>SDOperand</tt> was renamed to <tt>SDValue</tt> (and the "<tt>Val</tt>"
549 member was changed to be the <tt>getNode()</tt> accessor), and the
550 <tt>MVT::ValueType</tt> enum has been replaced with an "<tt>MVT</tt>"
Dan Gohmanc16fad22008-10-14 17:06:44 +0000[diff] [blame]551 struct. The <tt>getSignExtended</tt> and <tt>getValue</tt> methods in the
552 ConstantSDNode class were renamed to <tt>getSExtValue</tt> and
553 <tt>getZExtValue</tt> respectively, to be more consistent with
554 the <tt>ConstantInt</tt> class.</li>
Chris Lattner92eacf82008-10-13 17:57:36 +0000[diff] [blame]555</ul>
556
557</div>
558
559
560
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]561<!-- *********************************************************************** -->
562<div class="doc_section">
563 <a name="portability">Portability and Supported Platforms</a>
564</div>
565<!-- *********************************************************************** -->
566
567<div class="doc_text">
568
569<p>LLVM is known to work on the following platforms:</p>
570
571<ul>
Mikhail Glushenkov808cd3e2008-10-13 02:08:34 +0000[diff] [blame]572<li>Intel and AMD machines (IA32) running Red Hat Linux, Fedora Core and FreeBSD
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]573 (and probably other unix-like systems).</li>
Chris Lattnerbd633de92008-02-06 06:30:34 +0000[diff] [blame]574<li>PowerPC and X86-based Mac OS X systems, running 10.3 and above in 32-bit and
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]575 64-bit modes.</li>
Chris Lattnere0798082008-02-06 18:00:06 +0000[diff] [blame]576<li>Intel and AMD machines running on Win32 using MinGW libraries (native).</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]577<li>Intel and AMD machines running on Win32 with the Cygwin libraries (limited
578 support is available for native builds with Visual C++).</li>
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]579<li>Sun UltraSPARC workstations running Solaris 10.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]580<li>Alpha-based machines running Debian GNU/Linux.</li>
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]581<li>Itanium-based (IA64) machines running Linux and HP-UX.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]582</ul>
583
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]584<p>The core LLVM infrastructure uses GNU autoconf to adapt itself
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]585to the machine and operating system on which it is built. However, minor
586porting may be required to get LLVM to work on new platforms. We welcome your
587portability patches and reports of successful builds or error messages.</p>
588
589</div>
590
591<!-- *********************************************************************** -->
592<div class="doc_section">
593 <a name="knownproblems">Known Problems</a>
594</div>
595<!-- *********************************************************************** -->
596
597<div class="doc_text">
598
599<p>This section contains all known problems with the LLVM system, listed by
600component. As new problems are discovered, they will be added to these
601sections. If you run into a problem, please check the <a
602href="http://llvm.org/bugs/">LLVM bug database</a> and submit a bug if
603there isn't already one.</p>
604
605</div>
606
607<!-- ======================================================================= -->
608<div class="doc_subsection">
609 <a name="experimental">Experimental features included with this release</a>
610</div>
611
612<div class="doc_text">
613
614<p>The following components of this LLVM release are either untested, known to
615be broken or unreliable, or are in early development. These components should
616not be relied on, and bugs should not be filed against them, but they may be
617useful to some people. In particular, if you would like to work on one of these
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]618components, please contact us on the <a
619href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVMdev list</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]620
621<ul>
Gabor Greifaa53e562008-10-14 11:31:14 +0000[diff] [blame]622<li>The MSIL, IA64, Alpha, SPU, MIPS, and PIC16 backends are experimental.</li>
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame]623<li>The llc "<tt>-filetype=asm</tt>" (the default) is the only supported
Chris Lattnerbd633de92008-02-06 06:30:34 +0000[diff] [blame]624 value for this option.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]625</ul>
626
627</div>
628
629<!-- ======================================================================= -->
630<div class="doc_subsection">
631 <a name="x86-be">Known problems with the X86 back-end</a>
632</div>
633
634<div class="doc_text">
635
636<ul>
Anton Korobeynikov77d15e22008-06-08 10:24:13 +0000[diff] [blame]637 <li>The X86 backend does not yet support
638 all <a href="http://llvm.org/PR879">inline assembly that uses the X86
639 floating point stack</a>. It supports the 'f' and 't' constraints, but not
640 'u'.</li>
Chris Lattnerd95b5962008-06-08 21:19:07 +0000[diff] [blame]641 <li>The X86 backend generates inefficient floating point code when configured
642 to generate code for systems that don't have SSE2.</li>
Duncan Sands8b4f3772008-06-08 19:38:43 +0000[diff] [blame]643 <li>Win64 code generation wasn't widely tested. Everything should work, but we
Anton Korobeynikov77d15e22008-06-08 10:24:13 +0000[diff] [blame]644 expect small issues to happen. Also, llvm-gcc cannot build mingw64 runtime
645 currently due
646 to <a href="http://llvm.org/PR2255">several</a>
Chris Lattnerf06141a2008-06-08 23:12:47 +0000[diff] [blame]647 <a href="http://llvm.org/PR2257">bugs</a> due to lack of support for the
648 'u' inline assembly constraint and X87 floating point inline assembly.</li>
Dan Gohman3882ecb2008-06-08 23:05:11 +0000[diff] [blame]649 <li>The X86-64 backend does not yet support the LLVM IR instruction
650 <tt>va_arg</tt>. Currently, the llvm-gcc front-end supports variadic
651 argument constructs on X86-64 by lowering them manually.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]652</ul>
653
654</div>
655
656<!-- ======================================================================= -->
657<div class="doc_subsection">
658 <a name="ppc-be">Known problems with the PowerPC back-end</a>
659</div>
660
661<div class="doc_text">
662
663<ul>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]664<li>The Linux PPC32/ABI support needs testing for the interpreter and static
665compilation, and lacks support for debug information.</li>
666</ul>
667
668</div>
669
670<!-- ======================================================================= -->
671<div class="doc_subsection">
672 <a name="arm-be">Known problems with the ARM back-end</a>
673</div>
674
675<div class="doc_text">
676
677<ul>
678<li>Thumb mode works only on ARMv6 or higher processors. On sub-ARMv6
Duncan Sands1827fa62007-09-26 15:59:54 +0000[diff] [blame]679processors, thumb programs can crash or produce wrong
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]680results (<a href="http://llvm.org/PR1388">PR1388</a>).</li>
681<li>Compilation for ARM Linux OABI (old ABI) is supported, but not fully tested.
682</li>
Chris Lattnerd95b5962008-06-08 21:19:07 +0000[diff] [blame]683<li>There is a bug in QEMU-ARM (&lt;= 0.9.0) which causes it to incorrectly
684 execute
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]685programs compiled with LLVM. Please use more recent versions of QEMU.</li>
686</ul>
687
688</div>
689
690<!-- ======================================================================= -->
691<div class="doc_subsection">
692 <a name="sparc-be">Known problems with the SPARC back-end</a>
693</div>
694
695<div class="doc_text">
696
697<ul>
698<li>The SPARC backend only supports the 32-bit SPARC ABI (-m32), it does not
699 support the 64-bit SPARC ABI (-m64).</li>
700</ul>
701
702</div>
703
704<!-- ======================================================================= -->
705<div class="doc_subsection">
706 <a name="alpha-be">Known problems with the Alpha back-end</a>
707</div>
708
709<div class="doc_text">
710
711<ul>
712
713<li>On 21164s, some rare FP arithmetic sequences which may trap do not have the
714appropriate nops inserted to ensure restartability.</li>
715
716</ul>
717</div>
718
719<!-- ======================================================================= -->
720<div class="doc_subsection">
721 <a name="ia64-be">Known problems with the IA64 back-end</a>
722</div>
723
724<div class="doc_text">
725
726<ul>
Chris Lattnerf06141a2008-06-08 23:12:47 +0000[diff] [blame]727<li>The Itanium backend is highly experimental, and has a number of known
728 issues. We are looking for a maintainer for the Itanium backend. If you
729 are interested, please contact the llvmdev mailing list.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]730</ul>
731
732</div>
733
734<!-- ======================================================================= -->
735<div class="doc_subsection">
736 <a name="c-be">Known problems with the C back-end</a>
737</div>
738
739<div class="doc_text">
740
741<ul>
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]742<li><a href="http://llvm.org/PR802">The C backend has only basic support for
743 inline assembly code</a>.</li>
Chris Lattnere0194462007-09-26 06:01:35 +0000[diff] [blame]744<li><a href="http://llvm.org/PR1658">The C backend violates the ABI of common
745 C++ programs</a>, preventing intermixing between C++ compiled by the CBE and
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame]746 C++ code compiled with llc or native compilers.</li>
Duncan Sandsbd0ad6e2008-02-10 13:40:55 +0000[diff] [blame]747<li>The C backend does not support all exception handling constructs.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]748</ul>
749
750</div>
751
752
753<!-- ======================================================================= -->
754<div class="doc_subsection">
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]755 <a name="c-fe">Known problems with the llvm-gcc C front-end</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]756</div>
757
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]758<div class="doc_text">
759
Chris Lattnere0798082008-02-06 18:00:06 +0000[diff] [blame]760<p>llvm-gcc does not currently support <a href="http://llvm.org/PR869">Link-Time
761Optimization</a> on most platforms "out-of-the-box". Please inquire on the
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]762llvmdev mailing list if you are interested.</p>
763
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]764<p>The only major language feature of GCC not supported by llvm-gcc is
765 the <tt>__builtin_apply</tt> family of builtins. However, some extensions
766 are only supported on some targets. For example, trampolines are only
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame]767 supported on some targets (these are used when you take the address of a
768 nested function).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]769
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]770<p>If you run into GCC extensions which are not supported, please let us know.
771</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]772
773</div>
774
775<!-- ======================================================================= -->
776<div class="doc_subsection">
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]777 <a name="c++-fe">Known problems with the llvm-gcc C++ front-end</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]778</div>
779
780<div class="doc_text">
781
782<p>The C++ front-end is considered to be fully
783tested and works for a number of non-trivial programs, including LLVM
784itself, Qt, Mozilla, etc.</p>
785
786<ul>
Anton Korobeynikovd9a17932008-10-11 18:27:16 +0000[diff] [blame]787<li>Exception handling works well on the X86 and PowerPC targets. Currently
788 only linux and darwin targets are supported (both 32 and 64 bit).</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]789</ul>
790
791</div>
792
793
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]794<!-- ======================================================================= -->
795<div class="doc_subsection">
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]796 <a name="ada-fe">Known problems with the llvm-gcc Ada front-end</a>
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]797</div>
798
799<div class="doc_text">
800The llvm-gcc 4.2 Ada compiler works fairly well, however this is not a mature
801technology and problems should be expected.
802<ul>
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame]803<li>The Ada front-end currently only builds on X86-32. This is mainly due
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]804to lack of trampoline support (pointers to nested functions) on other platforms,
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame]805however it <a href="http://llvm.org/PR2006">also fails to build on X86-64</a>
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]806which does support trampolines.</li>
807<li>The Ada front-end <a href="http://llvm.org/PR2007">fails to bootstrap</a>.
808Workaround: configure with --disable-bootstrap.</li>
Duncan Sands2dbc2e82008-10-13 17:27:23 +0000[diff] [blame]809<li>The c380004, <a href="http://llvm.org/PR2010">c393010</a>
810and <a href="http://llvm.org/PR2421">cxg2021</a> ACATS tests fail
811(c380004 also fails with gcc-4.2 mainline).</li>
812<li>Some gcc specific Ada tests continue to crash the compiler.</li>
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]813<li>The -E binder option (exception backtraces)
814<a href="http://llvm.org/PR1982">does not work</a> and will result in programs
815crashing if an exception is raised. Workaround: do not use -E.</li>
816<li>Only discrete types <a href="http://llvm.org/PR1981">are allowed to start
817or finish at a non-byte offset</a> in a record. Workaround: do not pack records
818or use representation clauses that result in a field of a non-discrete type
819starting or finishing in the middle of a byte.</li>
Chris Lattnerd95b5962008-06-08 21:19:07 +0000[diff] [blame]820<li>The <tt>lli</tt> interpreter <a href="http://llvm.org/PR2009">considers
821'main' as generated by the Ada binder to be invalid</a>.
822Workaround: hand edit the file to use pointers for <tt>argv</tt> and
823<tt>envp</tt> rather than integers.</li>
824<li>The <tt>-fstack-check</tt> option <a href="http://llvm.org/PR2008">is
825ignored</a>.</li>
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]826</ul>
827</div>
828
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]829<!-- *********************************************************************** -->
830<div class="doc_section">
831 <a name="additionalinfo">Additional Information</a>
832</div>
833<!-- *********************************************************************** -->
834
835<div class="doc_text">
836
837<p>A wide variety of additional information is available on the <a
838href="http://llvm.org">LLVM web page</a>, in particular in the <a
839href="http://llvm.org/docs/">documentation</a> section. The web page also
840contains versions of the API documentation which is up-to-date with the
841Subversion version of the source code.
842You can access versions of these documents specific to this release by going
843into the "<tt>llvm/doc/</tt>" directory in the LLVM tree.</p>
844
845<p>If you have any questions or comments about LLVM, please feel free to contact
846us via the <a href="http://llvm.org/docs/#maillist"> mailing
847lists</a>.</p>
848
849</div>
850
851<!-- *********************************************************************** -->
852
853<hr>
854<address>
855 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
856 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
857 <a href="http://validator.w3.org/check/referer"><img
Chris Lattner5ba16862008-02-10 08:18:42 +0000[diff] [blame]858 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]859
860 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
861 Last modified: $Date$
862</address>
863
864</body>
865</html>