Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / a8f42149127c37a6c2c4ef13ecaf724967dff5ba / . / docs / ReleaseNotes-2.6.html
blob: 9f548c4d09faf5be7d1e621c691a559d4eb3ed1e [file] [log] [blame]
Duncan Sandse2e4a892009-06-24 08:38:48 +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">
7 <title>LLVM 2.6 Release Notes</title>
8</head>
9<body>
10
11<div class="doc_title">LLVM 2.6 Release Notes</div>
12
13<ol>
14 <li><a href="#intro">Introduction</a></li>
15 <li><a href="#subproj">Sub-project Status Update</a></li>
16 <li><a href="#externalproj">External Projects Using LLVM 2.6</a></li>
17 <li><a href="#whatsnew">What's New in LLVM 2.6?</a></li>
18 <li><a href="GettingStarted.html">Installation Instructions</a></li>
19 <li><a href="#portability">Portability and Supported Platforms</a></li>
20 <li><a href="#knownproblems">Known Problems</a></li>
21 <li><a href="#additionalinfo">Additional Information</a></li>
22</ol>
23
24<div class="doc_author">
25 <p>Written by the <a href="http://llvm.org">LLVM Team</a></p>
26</div>
27
28<!-- *********************************************************************** -->
29<div class="doc_section">
30 <a name="intro">Introduction</a>
31</div>
32<!-- *********************************************************************** -->
33
34<div class="doc_text">
35
36<p>This document contains the release notes for the LLVM Compiler
37Infrastructure, release 2.6. Here we describe the status of LLVM, including
38major improvements from the previous release and significant known problems.
39All LLVM releases may be downloaded from the <a
40href="http://llvm.org/releases/">LLVM releases web site</a>.</p>
41
42<p>For more information about LLVM, including information about the latest
43release, please check out the <a href="http://llvm.org/">main LLVM
44web site</a>. If you have questions or comments, the <a
45href="http://mail.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVM Developer's Mailing
46List</a> is a good place to send them.</p>
47
48<p>Note that if you are reading this file from a Subversion checkout or the
49main LLVM web page, this document applies to the <i>next</i> release, not the
50current one. To see the release notes for a specific release, please see the
51<a href="http://llvm.org/releases/">releases page</a>.</p>
52
53</div>
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]54
Chris Lattnerc758fec2009-10-05 02:12:39 +0000[diff] [blame]55
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]56<!--
57Almost dead code.
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]58 include/llvm/Analysis/LiveValues.h => Dan
59 lib/Transforms/IPO/MergeFunctions.cpp => consider for 2.8.
60 llvm/Analysis/PointerTracking.h => Edwin wants this, consider for 2.8.
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]61-->
Chris Lattnerc758fec2009-10-05 02:12:39 +0000[diff] [blame]62
Chris Lattnerc758fec2009-10-05 02:12:39 +0000[diff] [blame]63
64<!-- Unfinished features in 2.6:
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]65 gcc plugin.
Chris Lattnerc758fec2009-10-05 02:12:39 +0000[diff] [blame]66 strong phi elim
67 variable debug info for optimized code
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]68 postalloc scheduler: anti dependence breaking, hazard recognizer?
Chris Lattnerc758fec2009-10-05 02:12:39 +0000[diff] [blame]69 metadata
70 loop dependence analysis
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]71 ELF Writer? How stable?
72 <li>PostRA scheduler improvements, ARM adoption (David Goodwin).</li>
Chris Lattner45178502009-10-09 06:24:25 +0000[diff] [blame]73 2.7 supports the GDB 7.0 jit interfaces for debug info.
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]74 -->
75
76 <!-- for announcement email:
Chris Lattnerc758fec2009-10-05 02:12:39 +0000[diff] [blame]77 Logo web page.
78 llvm devmtg
79 compiler_rt
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]80 klee web page at klee.llvm.org
81 Many new papers added to /pubs/
82 Mention gcc plugin.
83
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]84 -->
85
86<!-- *********************************************************************** -->
87<div class="doc_section">
88 <a name="subproj">Sub-project Status Update</a>
89</div>
90<!-- *********************************************************************** -->
91
92<div class="doc_text">
93<p>
94The LLVM 2.6 distribution currently consists of code from the core LLVM
95repository &mdash;which roughly includes the LLVM optimizers, code generators
96and supporting tools &mdash; and the llvm-gcc repository. In addition to this
97code, the LLVM Project includes other sub-projects that are in development. The
98two which are the most actively developed are the <a href="#clang">Clang
99Project</a> and the <a href="#vmkit">VMKit Project</a>.
100</p>
101
102</div>
103
104
105<!--=========================================================================-->
106<div class="doc_subsection">
107<a name="clang">Clang: C/C++/Objective-C Frontend Toolkit</a>
108</div>
109
110<div class="doc_text">
111
Chris Lattner28d02742009-10-09 05:01:15 +0000[diff] [blame]112<p>The <a href="http://clang.llvm.org/">Clang project</a> is an effort to build
113a set of new 'LLVM native' front-end technologies for the C family of languages.
114LLVM 2.6 is the first release to officially include Clang, and it provides a
115production quality C and Objective-C compiler. If you are interested in fast
116compiles and good diagnostics, we encourage you to try it out.</p>
117
118<p>In addition to supporting these languages, C++ support is also <a
119href="http://clang.llvm.org/cxx_status.html">well under way</a>, and mainline
120Clang is able to parse the libstdc++ 4.2 headers and even codegen simple apps.
121If you are interested in Clang C++ support or any other Clang feature, we
122strongly encourage you to get involved on the <a
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]123href="http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev">Clang front-end mailing
124list</a>.</p>
125
126<p>In the LLVM 2.6 time-frame, the Clang team has made many improvements:</p>
127
128<ul>
Chris Lattner28d02742009-10-09 05:01:15 +0000[diff] [blame]129<li>C and Objective-C support are now considered production quality.</li>
Chris Lattnerc32a5322009-09-30 06:27:22 +0000[diff] [blame]130<li>AuroraUX / FreeBSD &amp; OpenBSD Toolchain support.</li>
Chris Lattner28d02742009-10-09 05:01:15 +0000[diff] [blame]131<li>Most of Objective-C 2.0 is now supported with the GNU runtime.</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]132<li>Many many bugs are fixed and many features have been added.</li>
133</ul>
134</div>
135
136<!--=========================================================================-->
137<div class="doc_subsection">
138<a name="clangsa">Clang Static Analyzer</a>
139</div>
140
141<div class="doc_text">
142
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]143<p><b>UPDATE!</b> Previously announced in the 2.4 and 2.5 LLVM releases, the Clang project also
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]144includes an early stage static source code analysis tool for <a
145href="http://clang.llvm.org/StaticAnalysis.html">automatically finding bugs</a>
146in C and Objective-C programs. The tool performs a growing set of checks to find
147bugs that occur on a specific path within a program.</p>
148
149<p>In the LLVM 2.6 time-frame there have been many significant improvements to
150XYZ.</p>
151
152<p>The set of checks performed by the static analyzer continues to expand, and
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>
157
158</div>
159
160<!--=========================================================================-->
161<div class="doc_subsection">
162<a name="vmkit">VMKit: JVM/CLI Virtual Machine Implementation</a>
163</div>
164
165<div class="doc_text">
166<p>
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]167<b>UPDATE!</b>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]168The <a href="http://vmkit.llvm.org/">VMKit project</a> is an implementation of
169a JVM and a CLI Virtual Machines (Microsoft .NET is an
170implementation of the CLI) using the Just-In-Time compiler of LLVM.</p>
171
172<p>Following LLVM 2.6, VMKit has its XYZ release that you can find on its
173<a href="http://vmkit.llvm.org/releases/">webpage</a>. The release includes
174bug fixes, cleanup and new features. The major changes are:</p>
175
176<ul>
177
178<li>Something wonderful!</li>
179
180</ul>
181</div>
182
Chris Lattner6ee62f82009-10-09 05:55:04 +0000[diff] [blame]183
184<!--=========================================================================-->
185<div class="doc_subsection">
186<a name="compiler-rt">compiler-rt: Compiler Runtime Library</a>
187</div>
188
189<div class="doc_text">
190<p>
191The new LLVM <a href="http://compiler-rt.llvm.org/">compiler-rt project</a>
192is a simple library that provides an implementation of the low-level
193target-specific hooks required by code generation and other runtime components.
194For example, when compiling for a 32-bit target, converting a double to a 64-bit
195unsigned integer is compiling into a runtime call to the "__fixunsdfdi"
196function. The compiler-rt library provides optimized implementations of this and
197other low-level routines.</p>
198
199<p>
200All of the code in the compiler-rt project is available under the standard LLVM
201License, a "BSD-style" license.</p>
202
203</div>
204
205<!--=========================================================================-->
206<div class="doc_subsection">
207<a name="klee">klee: Symbolic Execution and Automatic Test Case Generator</a>
208</div>
209
210<div class="doc_text">
211<p>
212The new LLVM <a href="http://klee.llvm.org/">klee project</a> is a symbolic
213execution framework for programs in LLVM bitcode form. Klee tries to
214symbolically evaluate "all" paths through the application and records state
215transitions that lead to fault states. This allows it to construct testcases
216that lead to faults and can even be used to verify algorithms. For more
217details, please see the <a
218href="http://llvm.org/pubs/2008-12-OSDI-KLEE.html">OSDI 2008 paper</a> about
219Klee.</p>
220
221</div>
222
223<!--=========================================================================-->
224<div class="doc_subsection">
225<a name="dragonegg">Dragon Egg: An LLVM backend plugin for GCC</a>
226</div>
227
228<div class="doc_text">
229<p>
230<b>Duncan needs to write me</b>.
231</p>
232
233</div>
234
235
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]236<!--=========================================================================-->
237<div class="doc_subsection">
238<a name="mc">llvm-mc: Machine Code Toolkit</a>
239</div>
240
241<div class="doc_text">
242<p>
Chris Lattner45178502009-10-09 06:24:25 +0000[diff] [blame]243The LLVM Machine Code (MC) Toolkit project is a (very early) effort to build
244better tools for dealing with machine code, object file formats, etc. The idea
245is to be able to generate most of the target specific details of assemblers and
246disassemblers from existing LLVM target .td files (with suitable enhancements),
247and to build infrastructure for reading and writing common object file formats.
248One of the first deliverables is to build a full assembler and integrate it into
249the compiler, which is predicted to substantially reduce compile time in some
250scenarios.
251</p>
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]252
Chris Lattner45178502009-10-09 06:24:25 +0000[diff] [blame]253<p>In the LLVM 2.6 timeframe, the MC framework has grown to the point where it
254can reliably parse and pretty print (with some encoding information) a
255darwin/x86 .s file successfully, and has the very early phases of a Mach-O
256assembler in progress. Beyond the MC framework itself, major refactoring of the
257LLVM code generator has started. The idea is to make the code generator reason
258about the code it is producing in a much more semantic way, rather than a
259textual way. For example, the code generator now uses MCSection objects to
260represent section assignments, instead of text strings that print to .section
261directives.</p>
262
263<p>MC is an early and ongoing project that will hopefully continue to lead to
264many improvements in the code generator and build infrastructure useful for many
265other situations.
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]266</p>
267
268</div>
269
270
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]271<!-- *********************************************************************** -->
272<div class="doc_section">
273 <a name="externalproj">External Projects Using LLVM 2.6</a>
274</div>
275<!-- *********************************************************************** -->
276
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]277<!--=========================================================================-->
278<div class="doc_subsection">
279<a name="Rubinius">Rubinius</a>
280</div>
281
282<div class="doc_text">
283<p><a href="http://github.com/evanphx/rubinius">Rubinius</a> is an environment
284for running Ruby code which strives to write as much of the core class
285implementation in Ruby as possible. Combined with a bytecode interpreting VM, it
286uses LLVM to optimize and compile ruby code down to machine code. Techniques
287such as type feedback, method inlining, and uncommon traps are all used to
288remove dynamism from ruby execution and increase performance.</p>
Chris Lattnerd4a537b2009-10-08 16:01:33 +0000[diff] [blame]289
290<p>Since LLVM 2.5, Rubinius has made several major leaps forward, implementing
291a counter based JIT, type feedback, and speculative method inlining.
292</p>
293
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]294</div>
Chris Lattnerc32a5322009-09-30 06:27:22 +0000[diff] [blame]295
296<!--=========================================================================-->
297<div class="doc_subsection">
298<a name="macruby">MacRuby</a>
299</div>
300
301<div class="doc_text">
302
303<p>
304<a href="http://macruby.org">MacRuby</a> is an implementation of Ruby on top of
305core Mac OS X technologies, such as the Objective-C common runtime and garbage
306collector, and the CoreFoundation framework. It is principally developed by
307Apple and aims at enabling the creation of full-fledged Mac OS X applications.
308</p>
309
310<p>
311MacRuby uses LLVM for optimization passes, JIT and AOT compilation of Ruby
312expressions. It also uses zero-cost DWARF exceptions to implement Ruby exception
313handling.</p>
314
315</div>
316
317
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]318<!--=========================================================================-->
319<div class="doc_subsection">
320<a name="pure">Pure</a>
321</div>
322
323<div class="doc_text">
324<p>
325<a href="http://pure-lang.googlecode.com/">Pure</a>
326is an algebraic/functional programming language based on term rewriting.
327Programs are collections of equations which are used to evaluate expressions in
328a symbolic fashion. Pure offers dynamic typing, eager and lazy evaluation,
329lexical closures, a hygienic macro system (also based on term rewriting),
330built-in list and matrix support (including list and matrix comprehensions) and
331an easy-to-use C interface. The interpreter uses LLVM as a backend to
332 JIT-compile Pure programs to fast native code.</p>
333
Chris Lattnerd4a537b2009-10-08 16:01:33 +0000[diff] [blame]334<p>Pure versions 0.31 and later have been tested and are known to work with
335LLVM 2.6 (and continue to work with older LLVM releases >= 2.3 as well).
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]336</p>
337</div>
338
339
340<!--=========================================================================-->
341<div class="doc_subsection">
342<a name="ldc">LLVM D Compiler</a>
343</div>
344
345<div class="doc_text">
346<p>
347<a href="http://www.dsource.org/projects/ldc">LDC</a> is an implementation of
348the D Programming Language using the LLVM optimizer and code generator.
349The LDC project works great with the LLVM 2.6 release. General improvements in
350this
351cycle have included new inline asm constraint handling, better debug info
352support, general bugfixes, and better x86-64 support. This has allowed
353some major improvements in LDC, getting us much closer to being as
354fully featured as the original DMD compiler from DigitalMars.
355</p>
356</div>
357
358<!--=========================================================================-->
359<div class="doc_subsection">
360<a name="RoadsendPHP">Roadsend PHP</a>
361</div>
362
363<div class="doc_text">
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]364<p>
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]365<a href="http://code.roadsend.com/rphp">Roadsend PHP</a> (rphp) is an open
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]366source implementation of the PHP programming
367language that uses LLVM for its optimizer, JIT, and static compiler. This is a
368reimplementation of an earlier project that is now based on LLVM.</p>
369</div>
370
Jeffrey Yasskinba2aa782009-06-24 21:09:13 +0000[diff] [blame]371<!--=========================================================================-->
372<div class="doc_subsection">
Chris Lattner6ee62f82009-10-09 05:55:04 +0000[diff] [blame]373<a name="UnladenSwallow">Unladen Swallow</a>
Jeffrey Yasskinba2aa782009-06-24 21:09:13 +0000[diff] [blame]374</div>
375
376<div class="doc_text">
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]377<p>
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]378<a href="http://code.google.com/p/unladen-swallow/">Unladen Swallow</a> is a
Jeffrey Yasskinba2aa782009-06-24 21:09:13 +0000[diff] [blame]379branch of <a href="http://python.org/">Python</a> intended to be fully
380compatible and significantly faster. It uses LLVM's optimization passes and JIT
381compiler.</p>
382</div>
383
Chris Lattner6ee62f82009-10-09 05:55:04 +0000[diff] [blame]384<!--=========================================================================-->
385<div class="doc_subsection">
386<a name="llvm-lua">llvm-lua</a>
387</div>
388
389<div class="doc_text">
390<p>
391<a href="http://code.google.com/p/llvm-lua/">LLVM-Lua</a> uses LLVM to add JIT
392&amp; static compiling support to the Lua VM. Lua bytecode is analyzed to
393remove type checks, then LLVM is used to compile those bytecodes down to machine
394code.</p>
395</div>
396
Jeffrey Yasskin5e98cb72009-06-24 21:26:42 +0000[diff] [blame]397
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]398
399<!-- *********************************************************************** -->
400<div class="doc_section">
401 <a name="whatsnew">What's New in LLVM 2.6?</a>
402</div>
403<!-- *********************************************************************** -->
404
405<div class="doc_text">
406
407<p>This release includes a huge number of bug fixes, performance tweaks, and
408minor improvements. Some of the major improvements and new features are listed
409in this section.
410</p>
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]411
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]412</div>
413
414<!--=========================================================================-->
415<div class="doc_subsection">
416<a name="majorfeatures">Major New Features</a>
417</div>
418
419<div class="doc_text">
420
421<p>LLVM 2.6 includes several major new capabilities:</p>
422
423<ul>
Chris Lattner6ee62f82009-10-09 05:55:04 +0000[diff] [blame]424<li>New <a href="#compiler-rt">compiler-rt</a>, <A href="#klee">klee</a>,
Chris Lattner45178502009-10-09 06:24:25 +0000[diff] [blame]425 and <a href="#mc">machine code toolkit</a> sub-projects.</li>
426<li>Debug information now includes line numbers when optimizations are enabled.
427 This allows statistical sampling tools like oprofile and Shark to map
428 samples back to source lines.</li>
429<li>LLVM now includes new experimental backends to support the MSP430, SystemZ,
Chris Lattner6ee62f82009-10-09 05:55:04 +0000[diff] [blame]430 and BlackFin architectures.</li>
431<li>LLVM supports a new <a href="GoldPlugin.html">Gold Linker Plugin</a> which
432 enables support for <a href="LinkTimeOptimization.html">transparent
433 link-time optimization</a> on ELF targets when used with the Gold binutils
434 linker.</li>
Chris Lattner45178502009-10-09 06:24:25 +0000[diff] [blame]435<li>LLVM now supports doing optimization and code generation on multiple threads
436 by allowing multiple "LLVMContext" objects to exist. Please see the <a
437 href="ProgrammersManual.html#threading">threading entry in the Programmer's
438 Manual</a> for more information.</li>
Chris Lattnera8f42142009-10-09 06:36:25 +0000[diff] [blame^]439<li>LLVM now has experimental support for <a
440 href="http://nondot.org/~sabre/LLVMNotes/EmbeddedMetadata.txt">embedded
441 metadata</a> in LLVM IR, though the implementation is not guaranteed to be
442 final and the .bc file format may change in future releases. Debug info
443 does not yet use this format in LLVM 2.6.</p>
444
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]445</ul>
446
447</div>
448
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]449<!--=========================================================================-->
450<div class="doc_subsection">
451<a name="coreimprovements">LLVM IR and Core Improvements</a>
452</div>
453
454<div class="doc_text">
455<p>LLVM IR has several new features that are used by our existing front-ends and
456can be useful if you are writing a front-end for LLVM:</p>
457
458<ul>
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]459<li>Getelementpr instruction now allows any integer type for array/pointer indexes.</li>
460<li>Inbounds for GEP</li>
461<li>NSW/NUW/exact div</li>
462<li>LSR promotes int induction variables to 64-bit on 64-bit targets, major perf boost for numerical code.</li>
463<li>LSR now analyzes pointer expressions (e.g. getelementptrs), not just integers.</li>
464<li>new linkage types linkonce_odr, weak_odr, linker_private, and available_externally.</li>
465<li>New fadd, fsub, fmul instructions and classes. </li>
466<li>Target intrinsics can now return multiple results.</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]467</ul>
468
469</div>
470
471<!--=========================================================================-->
472<div class="doc_subsection">
473<a name="optimizer">Optimizer Improvements</a>
474</div>
475
476<div class="doc_text">
477
478<p>In addition to a large array of bug fixes and minor performance tweaks, this
479release includes a few major enhancements and additions to the optimizers:</p>
480
481<ul>
482
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]483<li>SRoA improvements for vector unions, memset, arbitrary weird bitfield accesses etc. It now produces "strange" sized integers.</li>
484<li>Inliner reuse stack space when inlining arrays?</li>
485<li>Enabled GVN Load PRE.</li>
486<li>New Static Single Information (SSI) construction pass (not used by anything yet, experimental).</li>
487</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]488
489</ul>
490
491</div>
492
493<!--=========================================================================-->
494<div class="doc_subsection">
495<a name="codegen">Target Independent Code Generator Improvements</a>
496</div>
497
498<div class="doc_text">
499
500<p>We have put a significant amount of work into the code generator
501infrastructure, which allows us to implement more aggressive algorithms and make
502it run faster:</p>
503
504<ul>
505
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]506<li> -asm-verbose now prints location info (with -g) and loop nest info.</li>
507<li>Tblgen now supports multiclass inheritance and a number of new string and
508 list operations like !(subst), !(foreach), !car, !cdr, !null, !if, !cast.
509 These make the .td files more expressive and allow more aggressive factoring
510 of duplication across instruction patterns.</li>
511<li>New MachineVerifier pass.</li>
512<li>Machine LICM, hoists things like constant pool loads, loads from readonly stubs, vector constant synthesization code, etc.</li>
513<li>Machine Sinking</li>
514<li>target-specific intrinsics (r63765)</li>
515<li>Regalloc improvements for commuting, various spiller peephole optimizations, cross-class coalescing.</li>
516<li><tt>llc -enable-value-prop</tt>, propagation of value info (sign/zero ext info) from one MBB to another</li>
517<li>Regalloc hints for allocation stuff: Evan r73381/r73671. Finished/enabled?</li>
518<li>Stack slot coloring for register spills (denser stack frames)</li>
519<li>SelectionDAGS: New BuildVectorSDNode (r65296), and ISD::VECTOR_SHUFFLE (r69952 / PR2957)</li>
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]520<li>Experimental support for shrink wrapping support in PEI.</li>
Chris Lattner6ee62f82009-10-09 05:55:04 +0000[diff] [blame]521<li>Experimental support for writing ELF .o files directly from the compiler,
522 it works well for many simple C testcases, but doesn't support exception
523 handling, debug info, inline assembly, etc.</li>
524<li>Targets can now specify register allocation hints through
525 MachineRegisterInfo:: setRegAllocationHint. A regalloc hint consists 1) hint
526 type, 2) physical register number. A hint type of zero specifies a register
527 allocation preference. Other hint type values are target specific which are
528 resolved by TargetRegisterInfo::ResolveRegAllocHint. An example of which is
529 the ARM target can uses register hint to request that the register allocator
530 provide an even / odd register pair to two virtual registers. It is
531 important to note the register allocation hints are just hints. There is no
532 guarantee the register allocators will be able to satisfy the hints.</li>
533
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]534</ul>
535</div>
536
537<!--=========================================================================-->
538<div class="doc_subsection">
539<a name="x86">X86-32 and X86-64 Target Improvements</a>
540</div>
541
542<div class="doc_text">
543<p>New features of the X86 target include:
544</p>
545
546<ul>
547
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]548<li>Preliminary support for addrspace 256 -> GS, 257 -> FS, known problems: CodeGenerator.html#x86_memory</li>
549<li>Support for softfloat modes, typically used by OS kernels.</li>
550
551<li>X86-64: better modeling of implicit zero extensions, eliminates a lot of redundant zexts</li>
552<li>X86-64 TLS support for local exec and initial exec.</li>
553<li>Better modeling of H registerts as subregs.</li>
554<li>Vector icmp/fcmp now work with SSE codegen.</li>
555<li>SSE 4.2 support.</li>
556<li>all global variable reference logic is now in ClassifyGlobalReference.</li>
557</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]558</ul>
559
560</div>
561
562<!--=========================================================================-->
563<div class="doc_subsection">
564<a name="pic16">PIC16 Target Improvements</a>
565</div>
566
567<div class="doc_text">
568<p>New features of the PIC16 target include:
569</p>
570
571<ul>
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]572<li>Support for floating-point, indirect function calls, and
573 passing/returning aggregate types to functions.
574<li>The code generator is able to generate debug info into output COFF files.
575<li>Support for placing an object into a specific section or at a specific
576 address in memory.</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]577</ul>
578
579<p>Things not yet supported:</p>
580
581<ul>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]582<li>Variable arguments.</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]583<li>Interrupts/programs.</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]584</ul>
585
586</div>
587
Bob Wilson755cbe02009-08-12 21:19:49 +0000[diff] [blame]588<!--=========================================================================-->
589<div class="doc_subsection">
590<a name="ARM">ARM Target Improvements</a>
591</div>
592
593<div class="doc_text">
594<p>New features of the ARM target include:
595</p>
596
597<ul>
598
599<li>Preliminary support for processors, such as the Cortex-A8 and Cortex-A9,
Sandeep Patel64e39612009-08-20 15:01:16 +0000[diff] [blame]600that implement version v7-A of the ARM architecture. The ARM backend now
601supports both the Thumb2 and Advanced SIMD (Neon) instruction sets. The
602AAPCS-VFP "hard float" calling conventions are also supported with the
603<tt>-float-abi=hard</tt> flag. These features are still somewhat experimental
604and subject to change. The Neon intrinsics, in particular, may change in future
605releases of LLVM.
Bob Wilson755cbe02009-08-12 21:19:49 +0000[diff] [blame]606</li>
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]607
608 ARM AAPCS-VFP hard float ABI is supported.
609 ARM calling convention code is now tblgen generated instead of manual.
610 ARM: NEON support. neonfp for doing single precision fp with neon instead of VFP.
611
Bob Wilson755cbe02009-08-12 21:19:49 +0000[diff] [blame]612</ul>
613
614</div>
615
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]616<!--=========================================================================-->
617<div class="doc_subsection">
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]618<a name="OtherTarget">Other Target Specific Improvements</a>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]619</div>
620
621<div class="doc_text">
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]622<p>New features of other targets include:
623</p>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]624
625<ul>
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]626<li>Mips now supports O32 Calling Convention.</li>
Chris Lattnera8f42142009-10-09 06:36:25 +0000[diff] [blame^]627<li>Many improvements to the 32-bit PowerPC SVR4 ABI (used on powerpc-linux)
628 support, lots of bugs fixed.</li>
629<li>Added support for the 64-bit PowerPC SVR4 ABI (used on powerpc64-linux).
630 Needs more testing.</li>
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]631</ul>
632
633</div>
634
635<!--=========================================================================-->
636<div class="doc_subsection">
Chris Lattner45178502009-10-09 06:24:25 +0000[diff] [blame]637<a name="executionengine">Interpreter and JIT Improvements</a>
638</div>
639
640<div class="doc_text">
641
642<ul>
643<li>The JIT now supports generating more than 16M of code.</li>
644<li>When configured with --with-oprofile, the JIT can now inform oprofile about
645 JIT'd code, allowing oprofile to get line number and function name
646 information for JIT'd functions.</li>
647<li>When "libffi" is available, the LLVM interpreter now uses it, which supports
648 calling almost arbitrary external (natively compiled) functions.</li>
649<li>Clients of the JIT can now register a 'JITEventListener' object to receive
650 callbacks when the JIT emits or frees machine code. The OProfile support
651 uses this mechanism.</li>
652</ul>
653
654</div>
655
656
657<!--=========================================================================-->
658<div class="doc_subsection">
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]659<a name="newapis">New Useful APIs</a>
660</div>
661
662<div class="doc_text">
663
664<ul>
665<li>New EngineBuilder class for creating JITs: r76276</li>
666 New PrettyStackTrace, crashes of llvm tools should give some indication of what the compiler was doing at the time of the crash (e.g. running a pass), and print out command line arguments.
667 StringRef class, Twine class.
668 New WeakVH and AssertingVH and CallbackVH classes.
669 New llvm/ADT/Triple class.
670 llvm_report_error() error handling API (llvm/Support/ErrorHandling.h)
671 New llvm/System/Atomic.h, llvm/System/RWMutex.h for portable atomic ops, rw locks.
672 New SourceMgr, SMLoc classes for simple parsers with caret diagnostics and #include support, (used by
673 tablegen, llvm-mc, the .ll parser, FileCheck, etc)
674
Chris Lattner45178502009-10-09 06:24:25 +0000[diff] [blame]675</ul>
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]676
677
678</div>
679
680<!--=========================================================================-->
681<div class="doc_subsection">
682<a name="otherimprovements">Other Improvements and New Features</a>
683</div>
684
685<div class="doc_text">
686<p>Other miscellaneous features include:</p>
687
688<ul>
Chris Lattner45178502009-10-09 06:24:25 +0000[diff] [blame]689<li>LLVM now includes a new internal '<a
690 href="http://llvm.org/cmds/FileCheck.html">FileCheck</a>' tool which allows
691 writing much more accurate regression tests that run faster. Please see the
692 <a href="TestingGuide.html#FileCheck">FileCheck section of the Testing
693 Guide</a> for more information.</li>
Chris Lattner6ee62f82009-10-09 05:55:04 +0000[diff] [blame]694<li>LLVM profile information support has been significantly improved to produce
695correct use counts, and has support for edge profiling with reduced runtime
696overhead. Combined, the generated profile information is both more correct and
697imposes about half as much overhead (2.6. from 12% to 6% overhead on SPEC
698CPU2000).</li>
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]699<li>Many extensions to the C APIs.</li>
Chris Lattner45178502009-10-09 06:24:25 +0000[diff] [blame]700<li>LLVM 2.6 includes a brand new experimental LLVM bindings to the Ada2005
701programming language.</li>
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]702
703<li>LLVMC:
704
705* Dynamic plugins now work on Windows.
706* New option property: init. Makes possible to provide default values for
707 options defined in plugins (interface to cl::init).
708* New example: Skeleton, shows how to create a standalone LLVMC-based driver.
709* New example: mcc16, a driver for the PIC16 toolchain.</li>
710
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]711</ul>
712
713</div>
714
715
716<!--=========================================================================-->
717<div class="doc_subsection">
718<a name="changes">Major Changes and Removed Features</a>
719</div>
720
721<div class="doc_text">
722
723<p>If you're already an LLVM user or developer with out-of-tree changes based
Owen Andersone27be3a2009-07-02 16:48:38 +0000[diff] [blame]724on LLVM 2.5, this section lists some "gotchas" that you may run into upgrading
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]725from the previous release.</p>
726
727<ul>
728
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]729<li>The Itanium (IA64) backend has been removed. It was not supported and
730 bitrotted.</li>
731<li>The BigBlock register allocator has been removed, it also bitrotted.</li>
732<li>The C Backend (-march=c) is no longer considered part of the LLVM release
733criteria. We still want it to work, but no one is maintaining it and it lacks
734support for arbitrary precision integers and other important IR features.</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]735</ul>
736
Chris Lattner80ed2552009-10-08 07:01:46 +0000[diff] [blame]737 LLVM build now builds all libraries as .a files instead of some
738 libraries as relinked .o files. This requires some APIs like
739 InitializeAllTargets.h. TargetRegistry!
740
741
742
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]743
744<p>In addition, many APIs have changed in this release. Some of the major LLVM
745API changes are:</p>
746
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]747
748 API Cleanup:
749 no use of hash_set/hash_map, no more llvm::OStream
750 Use raw_ostream for everything, killed off llvm/Streams.h and DOUT
751
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]752<ul>
Owen Andersone27be3a2009-07-02 16:48:38 +0000[diff] [blame]753<li>LLVM's global uniquing tables for <tt>Type</tt>s and <tt>Constant</tt>s have
754 been privatized into members of an <tt>LLVMContext</tt>. A number of APIs
755 now take an <tt>LLVMContext</tt> as a parameter. To smooth the transition
756 for clients that will only ever use a single context, the new
757 <tt>getGlobalContext()</tt> API can be used to access a default global
758 context which can be passed in any and all cases where a context is
759 required.
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]760<li>The <tt>getABITypeSize</tt> methods are now called <tt>getAllocSize</tt>.</li>
Dan Gohmanf667e7e2009-07-07 20:05:15 +0000[diff] [blame]761<li>The <tt>Add</tt>, <tt>Sub</tt>, and <tt>Mul</tt> operators are no longer
762 overloaded for floating-point types. Floating-point addition, subtraction,
763 and multiplication are now represented with new operators <tt>FAdd</tt>,
764 <tt>FSub</tt>, and <tt>FMul</tt>. In the <tt>IRBuilder</tt> API,
765 <tt>CreateAdd</tt>, <tt>CreateSub</tt>, <tt>CreateMul</tt>, and
766 <tt>CreateNeg</tt> should only be used for integer arithmetic now;
767 <tt>CreateFAdd</tt>, <tt>CreateFSub</tt>, <tt>CreateFMul</tt>, and
768 <tt>CreateFNeg</tt> should now be used for floating-point arithmetic.</li>
Daniel Dunbar02e73132009-07-12 20:41:27 +0000[diff] [blame]769<li>The DynamicLibrary class can no longer be constructed, its functionality has
770 moved to static member functions.</li>
Dan Gohmane4d54d72009-07-15 19:59:19 +0000[diff] [blame]771<li><tt>raw_fd_ostream</tt>'s constructor for opening a given filename now
772 takes an extra <tt>Force</tt> argument. If <tt>Force</tt> is set to
773 <tt>false</tt>, an error will be reported if a file with the given name
774 already exists. If <tt>Force</tt> is set to <tt>true</tt>, the file will
775 be silently truncated (which is the behavior before this flag was
776 added).</li>
Torok Edwinf6fa8ae2009-07-21 20:27:10 +0000[diff] [blame]777<li><tt>SCEVHandle</tt> no longer exists, because reference counting is no
778longer done for <tt>SCEV*</tt> objects, instead <tt>const SCEV*</tt> should be
779used.</li>
Daniel Dunbar8b5ee822009-07-25 05:26:53 +0000[diff] [blame]780
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]781<li>Many APIs, notably <tt>llvm::Value</tt>, now use the <tt>StringRef</tt>
782and <tt>Twine</tt> classes instead of passing <tt>const char*</tt>
783or <tt>std::string</tt>, as described in
784the <a href="ProgrammersManual.html#string_apis">Programmer's Manual</a>. Most
Daniel Dunbar8b5ee822009-07-25 05:26:53 +0000[diff] [blame]785clients should be unaffected by this transition, unless they are used to <tt>Value::getName()</tt> returning a string. Here are some tips on updating to 2.6:
786 <ul>
787 <li><tt>getNameStr()</tt> is still available, and matches the old
788 behavior. Replacing <tt>getName()</tt> calls with this is an safe option,
789 although more efficient alternatives are now possible.</li>
790
791 <li>If you were just relying on <tt>getName()</tt> being able to be sent to
792 a <tt>std::ostream</tt>, consider migrating
793 to <tt>llvm::raw_ostream</tt>.</li>
794
795 <li>If you were using <tt>getName().c_str()</tt> to get a <tt>const
796 char*</tt> pointer to the name, you can use <tt>getName().data()</tt>.
797 Note that this string (as before), may not be the entire name if the
798 name containts embedded null characters.</li>
799
800 <li>If you were using operator plus on the result of <tt>getName()</tt> and
801 treating the result as an <tt>std::string</tt>, you can either
802 uses <tt>Twine::str</tt> to get the result as an <tt>std::string</tt>, or
803 could move to a <tt>Twine</tt> based design.</li>
Daniel Dunbar03d76512009-07-25 23:55:21 +0000[diff] [blame]804
805 <li><tt>isName()</tt> should be replaced with comparison
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]806 against <tt>getName()</tt> (this is now efficient).
Daniel Dunbar8b5ee822009-07-25 05:26:53 +0000[diff] [blame]807 </ul>
808</li>
809
Daniel Dunbara5881e32009-07-26 02:12:58 +0000[diff] [blame]810<li>The registration interfaces for backend Targets has changed (what was
Daniel Dunbard6b06b12009-07-26 05:41:39 +0000[diff] [blame]811previously TargetMachineRegistry). For backend authors, see the <a href="WritingAnLLVMBackend.html#TargetRegistration">Writing An LLVM Backend</a> guide. For clients, the notable API changes are:
812 <ul>
813 <li><tt>TargetMachineRegistry</tt> has been renamed
814 to <tt>TargetRegistry</tt>.</li>
815
816 <li>Clients should move to using the <tt>TargetRegistry::lookupTarget()</tt>
817 function to find targets.</li>
818 </ul>
819</li>
Daniel Dunbar8b5ee822009-07-25 05:26:53 +0000[diff] [blame]820
Torok Edwinf6fa8ae2009-07-21 20:27:10 +0000[diff] [blame]821<li>llvm-dis now fails if output file exists, instead of dumping to stdout.
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]822FIXME: describe any other tool changes due to the raw_fd_ostream change. FIXME:
823This is not an API change, maybe there should be a tool changes section?</li>
Torok Edwinf6fa8ae2009-07-21 20:27:10 +0000[diff] [blame]824<li>temporarely due to Context API change passes should call doInitialization()
825method of the pass they inherit from, otherwise Context is NULL.
826FIXME: remove this entry when this is no longer needed.<li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]827</ul>
828
829</div>
830
831
832
833<!-- *********************************************************************** -->
834<div class="doc_section">
835 <a name="portability">Portability and Supported Platforms</a>
836</div>
837<!-- *********************************************************************** -->
838
839<div class="doc_text">
840
841<p>LLVM is known to work on the following platforms:</p>
842
843<ul>
844<li>Intel and AMD machines (IA32, X86-64, AMD64, EMT-64) running Red Hat
Chris Lattner3e7b5ca2009-07-21 22:47:03 +0000[diff] [blame]845Linux, Fedora Core, FreeBSD and AuroraUX (and probably other unix-like systems).</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]846<li>PowerPC and X86-based Mac OS X systems, running 10.3 and above in 32-bit
847and 64-bit modes.</li>
848<li>Intel and AMD machines running on Win32 using MinGW libraries (native).</li>
849<li>Intel and AMD machines running on Win32 with the Cygwin libraries (limited
850 support is available for native builds with Visual C++).</li>
851<li>Sun UltraSPARC workstations running Solaris 10.</li>
852<li>Alpha-based machines running Debian GNU/Linux.</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]853</ul>
854
855<p>The core LLVM infrastructure uses GNU autoconf to adapt itself
856to the machine and operating system on which it is built. However, minor
857porting may be required to get LLVM to work on new platforms. We welcome your
858portability patches and reports of successful builds or error messages.</p>
859
860</div>
861
862<!-- *********************************************************************** -->
863<div class="doc_section">
864 <a name="knownproblems">Known Problems</a>
865</div>
866<!-- *********************************************************************** -->
867
868<div class="doc_text">
869
870<p>This section contains significant known problems with the LLVM system,
871listed by component. If you run into a problem, please check the <a
872href="http://llvm.org/bugs/">LLVM bug database</a> and submit a bug if
873there isn't already one.</p>
874
Chris Lattner3e7b5ca2009-07-21 22:47:03 +0000[diff] [blame]875<ul>
Chris Lattneraedb59a2009-07-21 23:17:26 +0000[diff] [blame]876<li>LLVM will not correctly compile on Solaris and/or OpenSolaris
Chris Lattner3e7b5ca2009-07-21 22:47:03 +0000[diff] [blame]877using the stock GCC 3.x.x series 'out the box',
878See: <a href="#brokengcc">Broken versions of GCC and other tools</a>.
879However, A <a href="http://pkg.auroraux.org/GCC">Modern GCC Build</a>
880for x86/x64 has been made available from the third party AuroraUX Project
Chris Lattner7b0a6812009-10-08 06:27:53 +0000[diff] [blame]881that has been meticulously tested for bootstrapping LLVM &amp; Clang.</li>
Chris Lattner3e7b5ca2009-07-21 22:47:03 +0000[diff] [blame]882</ul>
883
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]884</div>
885
886<!-- ======================================================================= -->
887<div class="doc_subsection">
888 <a name="experimental">Experimental features included with this release</a>
889</div>
890
891<div class="doc_text">
892
893<p>The following components of this LLVM release are either untested, known to
894be broken or unreliable, or are in early development. These components should
895not be relied on, and bugs should not be filed against them, but they may be
896useful to some people. In particular, if you would like to work on one of these
897components, please contact us on the <a
898href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVMdev list</a>.</p>
899
900<ul>
Dan Gohmand2cb3d22009-07-24 00:30:09 +0000[diff] [blame]901<li>The MSIL, Alpha, SPU, MIPS, and PIC16 backends are experimental.</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]902<li>The <tt>llc</tt> "<tt>-filetype=asm</tt>" (the default) is the only
903 supported value for this option.</li>
904</ul>
905
906</div>
907
908<!-- ======================================================================= -->
909<div class="doc_subsection">
910 <a name="x86-be">Known problems with the X86 back-end</a>
911</div>
912
913<div class="doc_text">
914
915<ul>
916 <li>The X86 backend does not yet support
917 all <a href="http://llvm.org/PR879">inline assembly that uses the X86
918 floating point stack</a>. It supports the 'f' and 't' constraints, but not
919 'u'.</li>
920 <li>The X86 backend generates inefficient floating point code when configured
921 to generate code for systems that don't have SSE2.</li>
922 <li>Win64 code generation wasn't widely tested. Everything should work, but we
923 expect small issues to happen. Also, llvm-gcc cannot build the mingw64
924 runtime currently due
925 to <a href="http://llvm.org/PR2255">several</a>
926 <a href="http://llvm.org/PR2257">bugs</a> and due to lack of support for
927 the
928 'u' inline assembly constraint and for X87 floating point inline assembly.</li>
929 <li>The X86-64 backend does not yet support the LLVM IR instruction
930 <tt>va_arg</tt>. Currently, the llvm-gcc and front-ends support variadic
931 argument constructs on X86-64 by lowering them manually.</li>
932</ul>
933
934</div>
935
936<!-- ======================================================================= -->
937<div class="doc_subsection">
938 <a name="ppc-be">Known problems with the PowerPC back-end</a>
939</div>
940
941<div class="doc_text">
942
943<ul>
944<li>The Linux PPC32/ABI support needs testing for the interpreter and static
945compilation, and lacks support for debug information.</li>
946</ul>
947
948</div>
949
950<!-- ======================================================================= -->
951<div class="doc_subsection">
952 <a name="arm-be">Known problems with the ARM back-end</a>
953</div>
954
955<div class="doc_text">
956
957<ul>
Bob Wilson755cbe02009-08-12 21:19:49 +0000[diff] [blame]958<li>Support for the Advanced SIMD (Neon) instruction set is still incomplete
959and not well tested. Some features may not work at all, and the code quality
960may be poor in some cases.</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]961<li>Thumb mode works only on ARMv6 or higher processors. On sub-ARMv6
962processors, thumb programs can crash or produce wrong
963results (<a href="http://llvm.org/PR1388">PR1388</a>).</li>
964<li>Compilation for ARM Linux OABI (old ABI) is supported but not fully tested.
965</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]966</ul>
967
968</div>
969
970<!-- ======================================================================= -->
971<div class="doc_subsection">
972 <a name="sparc-be">Known problems with the SPARC back-end</a>
973</div>
974
975<div class="doc_text">
976
977<ul>
978<li>The SPARC backend only supports the 32-bit SPARC ABI (-m32); it does not
979 support the 64-bit SPARC ABI (-m64).</li>
980</ul>
981
982</div>
983
984<!-- ======================================================================= -->
985<div class="doc_subsection">
986 <a name="mips-be">Known problems with the MIPS back-end</a>
987</div>
988
989<div class="doc_text">
990
991<ul>
992<li>The O32 ABI is not fully supported.</li>
993<li>64-bit MIPS targets are not supported yet.</li>
994</ul>
995
996</div>
997
998<!-- ======================================================================= -->
999<div class="doc_subsection">
1000 <a name="alpha-be">Known problems with the Alpha back-end</a>
1001</div>
1002
1003<div class="doc_text">
1004
1005<ul>
1006
1007<li>On 21164s, some rare FP arithmetic sequences which may trap do not have the
1008appropriate nops inserted to ensure restartability.</li>
1009
1010</ul>
1011</div>
1012
1013<!-- ======================================================================= -->
1014<div class="doc_subsection">
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]1015 <a name="c-be">Known problems with the C back-end</a>
1016</div>
1017
1018<div class="doc_text">
1019
1020<ul>
1021<li><a href="http://llvm.org/PR802">The C backend has only basic support for
1022 inline assembly code</a>.</li>
1023<li><a href="http://llvm.org/PR1658">The C backend violates the ABI of common
1024 C++ programs</a>, preventing intermixing between C++ compiled by the CBE and
1025 C++ code compiled with <tt>llc</tt> or native compilers.</li>
1026<li>The C backend does not support all exception handling constructs.</li>
1027<li>The C backend does not support arbitrary precision integers.</li>
1028</ul>
1029
1030</div>
1031
1032
1033<!-- ======================================================================= -->
1034<div class="doc_subsection">
1035 <a name="c-fe">Known problems with the llvm-gcc C front-end</a>
1036</div>
1037
1038<div class="doc_text">
1039
1040<p>llvm-gcc does not currently support <a href="http://llvm.org/PR869">Link-Time
1041Optimization</a> on most platforms "out-of-the-box". Please inquire on the
1042LLVMdev mailing list if you are interested.</p>
1043
1044<p>The only major language feature of GCC not supported by llvm-gcc is
1045 the <tt>__builtin_apply</tt> family of builtins. However, some extensions
1046 are only supported on some targets. For example, trampolines are only
1047 supported on some targets (these are used when you take the address of a
1048 nested function).</p>
1049
1050<p>If you run into GCC extensions which are not supported, please let us know.
1051</p>
1052
1053</div>
1054
1055<!-- ======================================================================= -->
1056<div class="doc_subsection">
1057 <a name="c++-fe">Known problems with the llvm-gcc C++ front-end</a>
1058</div>
1059
1060<div class="doc_text">
1061
1062<p>The C++ front-end is considered to be fully
1063tested and works for a number of non-trivial programs, including LLVM
1064itself, Qt, Mozilla, etc.</p>
1065
1066<ul>
1067<li>Exception handling works well on the X86 and PowerPC targets. Currently
1068 only Linux and Darwin targets are supported (both 32 and 64 bit).</li>
1069</ul>
1070
1071</div>
1072
1073<!-- ======================================================================= -->
1074<div class="doc_subsection">
1075 <a name="fortran-fe">Known problems with the llvm-gcc Fortran front-end</a>
1076</div>
1077
1078<div class="doc_text">
1079<ul>
1080<li>Fortran support generally works, but there are still several unresolved bugs
1081 in Bugzilla. Please see the tools/gfortran component for details.</li>
1082</ul>
1083</div>
1084
1085<!-- ======================================================================= -->
1086<div class="doc_subsection">
1087 <a name="ada-fe">Known problems with the llvm-gcc Ada front-end</a>
1088</div>
1089
1090<div class="doc_text">
1091The llvm-gcc 4.2 Ada compiler works fairly well; however, this is not a mature
1092technology, and problems should be expected.
1093<ul>
1094<li>The Ada front-end currently only builds on X86-32. This is mainly due
1095to lack of trampoline support (pointers to nested functions) on other platforms.
1096However, it <a href="http://llvm.org/PR2006">also fails to build on X86-64</a>
1097which does support trampolines.</li>
1098<li>The Ada front-end <a href="http://llvm.org/PR2007">fails to bootstrap</a>.
1099This is due to lack of LLVM support for <tt>setjmp</tt>/<tt>longjmp</tt> style
1100exception handling, which is used internally by the compiler.
1101Workaround: configure with --disable-bootstrap.</li>
1102<li>The c380004, <a href="http://llvm.org/PR2010">c393010</a>
1103and <a href="http://llvm.org/PR2421">cxg2021</a> ACATS tests fail
1104(c380004 also fails with gcc-4.2 mainline).
1105If the compiler is built with checks disabled then <a href="http://llvm.org/PR2010">c393010</a>
1106causes the compiler to go into an infinite loop, using up all system memory.</li>
1107<li>Some GCC specific Ada tests continue to crash the compiler.</li>
1108<li>The -E binder option (exception backtraces)
1109<a href="http://llvm.org/PR1982">does not work</a> and will result in programs
1110crashing if an exception is raised. Workaround: do not use -E.</li>
1111<li>Only discrete types <a href="http://llvm.org/PR1981">are allowed to start
1112or finish at a non-byte offset</a> in a record. Workaround: do not pack records
1113or use representation clauses that result in a field of a non-discrete type
1114starting or finishing in the middle of a byte.</li>
1115<li>The <tt>lli</tt> interpreter <a href="http://llvm.org/PR2009">considers
1116'main' as generated by the Ada binder to be invalid</a>.
1117Workaround: hand edit the file to use pointers for <tt>argv</tt> and
1118<tt>envp</tt> rather than integers.</li>
1119<li>The <tt>-fstack-check</tt> option <a href="http://llvm.org/PR2008">is
1120ignored</a>.</li>
1121</ul>
1122</div>
1123
Erick Tryzelaar17167be2009-09-28 04:42:55 +0000[diff] [blame]1124<!-- ======================================================================= -->
1125<div class="doc_subsection">
1126 <a name="ocaml-bindingse">Known problems with the O'Caml bindings</a>
1127</div>
1128
1129<div class="doc_text">
1130
1131<p>The Llvm.Linkage module is broken, and has incorrect values. Only
1132Llvm.Linkage.External, Llvm.Linkage.Available_externally, and
1133Llvm.Linkage.Link_once will be correct. If you need any of the other linkage
1134modes, you'll have to write an external C library in order to expose the
1135functionality. This has been fixed in the trunk.</p>
1136</div>
1137
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]1138<!-- *********************************************************************** -->
1139<div class="doc_section">
1140 <a name="additionalinfo">Additional Information</a>
1141</div>
1142<!-- *********************************************************************** -->
1143
1144<div class="doc_text">
1145
1146<p>A wide variety of additional information is available on the <a
1147href="http://llvm.org">LLVM web page</a>, in particular in the <a
1148href="http://llvm.org/docs/">documentation</a> section. The web page also
1149contains versions of the API documentation which is up-to-date with the
1150Subversion version of the source code.
1151You can access versions of these documents specific to this release by going
1152into the "<tt>llvm/doc/</tt>" directory in the LLVM tree.</p>
1153
1154<p>If you have any questions or comments about LLVM, please feel free to contact
1155us via the <a href="http://llvm.org/docs/#maillist"> mailing
1156lists</a>.</p>
1157
1158</div>
1159
1160<!-- *********************************************************************** -->
1161
1162<hr>
1163<address>
1164 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
1165 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
1166 <a href="http://validator.w3.org/check/referer"><img
1167 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
1168
1169 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
1170 Last modified: $Date$
1171</address>
1172
1173</body>
1174</html>