Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / a73561c701b2a84659d600c7baafa83aaf9a1cb4 / . / docs / ReleaseNotes.html
blob: 3cb9f94bb320e407f046e496942c3b6ec8020546 [file] [log] [blame]
Anton Korobeynikov877b73b2008-06-08 10:23:46 +0000[diff] [blame]1
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]2<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
3 "http://www.w3.org/TR/html4/strict.dtd">
4<html>
5<head>
6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
7 <link rel="stylesheet" href="llvm.css" type="text/css">
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]8 <title>LLVM 2.3 Release Notes</title>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]9</head>
10<body>
11
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]12<div class="doc_title">LLVM 2.3 Release Notes</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]13
14<ol>
15 <li><a href="#intro">Introduction</a></li>
16 <li><a href="#whatsnew">What's New?</a></li>
17 <li><a href="GettingStarted.html">Installation Instructions</a></li>
18 <li><a href="#portability">Portability and Supported Platforms</a></li>
19 <li><a href="#knownproblems">Known Problems</a>
20 <li><a href="#additionalinfo">Additional Information</a></li>
21</ol>
22
23<div class="doc_author">
24 <p>Written by the <a href="http://llvm.org">LLVM Team</a><p>
25</div>
26
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]27<!-- Done through Week-of-Mon-20080324.txt -->
28
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]29<!-- *********************************************************************** -->
30<div class="doc_section">
31 <a name="intro">Introduction</a>
32</div>
33<!-- *********************************************************************** -->
34
35<div class="doc_text">
36
37<p>This document contains the release notes for the LLVM compiler
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]38infrastructure, release 2.3. Here we describe the status of LLVM, including
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]39major improvements from the previous release and any known problems. All LLVM
40releases may be downloaded from the <a href="http://llvm.org/releases/">LLVM
41releases web site</a>.</p>
42
43<p>For more information about LLVM, including information about the latest
44release, please check out the <a href="http://llvm.org/">main LLVM
45web site</a>. If you have questions or comments, the <a
46href="http://mail.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVM developer's mailing
47list</a> is a good place to send them.</p>
48
49<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]50main LLVM web page, this document applies to the <i>next</i> release, not the
51current one. To see the release notes for a specific releases, please see the
52<a href="http://llvm.org/releases/">releases page</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]53
54</div>
55
56<!-- *********************************************************************** -->
57<div class="doc_section">
58 <a name="whatsnew">What's New?</a>
59</div>
60<!-- *********************************************************************** -->
61
62<div class="doc_text">
63
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]64<p>This is the fourteenth public release of the LLVM Compiler Infrastructure.
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]65It includes a large number of features and refinements from LLVM 2.2.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]66
67</div>
68
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]69<!-- Unfinished features in 2.3:
Chris Lattner8fda4952008-02-10 07:46:44 +0000[diff] [blame]70 Machine LICM
71 Machine Sinking
72 LegalizeDAGTypes
73 -->
74
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]75<!--=========================================================================-->
76<div class="doc_subsection">
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]77<a name="changes">Major Changes in LLVM 2.3</a>
Chris Lattner41bf8e92008-02-10 07:04:35 +0000[diff] [blame]78</div>
79
80<div class="doc_text">
81
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]82<p>LLVM 2.2 was the last LLVM release to support llvm-gcc 4.0 and llvm-upgrade.
83llvm-gcc 4.0 has been replaced with llvm-gcc 4.2. llvm-upgrade was useful for
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]84upgrading LLVM 1.9 files to LLVM 2.x syntax, but you can always use a previous
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]85LLVM release to do this. One nice impact of this is that the LLVM regression
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]86test suite no longer depends on llvm-upgrade, which makes it run faster.</p>
Chris Lattnere0798082008-02-06 18:00:06 +0000[diff] [blame]87
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]88<p>LLVM API Changes:</p>
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]89
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]90<ul>
91<li>Several core LLVM IR classes have migrated to use 'FOOCLASS::Create(...)'
92 instead of 'new FOOCLASS(...)' (e.g. where FOOCLASS=BasicBlock). We hope to
93 standardize on FOOCLASS::Create for all IR classes in the future, but not
94 all of them have been moved over yet.</li>
95<li>LLVM 2.3 renames the LLVMBuilder and LLVMFoldingBuilder classes to
96 IRBuilder.</li>
97<li>MRegisterInfo was renamed to TargetRegisterInfo.</li>
98<li>The MappedFile class is gone, please use MemoryBuffer instead.</li>
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]99<li>The '-enable-eh' flag to llc has been removed. Now code should encode
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]100 whether it is safe to not generate unwind information for a function by
101 tagging the Function object with the 'nounwind' attribute.</li>
102
103</ul>
Chris Lattnere0798082008-02-06 18:00:06 +0000[diff] [blame]104</div>
105
106<!--=========================================================================-->
107<div class="doc_subsection">
108<a name="majorfeatures">Major New Features</a>
109</div>
110
111<div class="doc_text">
112
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]113<p>LLVM 2.3 includes several major new capabilities:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]114
115<ul>
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]116<li>Multiple Return Value Support.</li>
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]117
118
119<li><p>LLVM 2.3 includes a complete reimplementation of the "llvmc" tool. It is
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]120designed to overcome several problems with the original llvmc and to provide a
121superset of the features of the 'gcc' driver.</p>
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]122
123<p>The main features of llvmc2 is:</p>
124
125<ul>
126<li>Extended handling of command line options and smart rules for
127dispatching them to different tools.</li>
128<li>Flexible (and extensible) rules for defining different tools.</li>
129<li>The different intermediate steps performed by tools are represented
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]130as edges in the abstract graph.</li>
Duncan Sands8b4f3772008-06-08 19:38:43 +0000[diff] [blame]131<li>The 'language' for driver behavior definition is tablegen and thus
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]132it's relatively easy to add new features.</li>
133<li>The definition of driver is transformed into set of C++ classes, thus
134no runtime interpretation is needed.</li>
135</ul>
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]136</li>
137
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]138
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]139<li>Reimplemented <a href="LinkTimeOptimization.html">LTO interface</a> in
140 C.</li>
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]141
142
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]143<li>Kaleidoscope tutorial in Ocaml.</li>
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]144
Chris Lattner5ba16862008-02-10 08:18:42 +0000[diff] [blame]145</ul>
146
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]147</div>
148
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]149
150<!--=========================================================================-->
151<div class="doc_subsection">
152<a name="frontends">llvm-gcc 4.2 Improvements and Clang</a>
153</div>
154
155<div class="doc_text">
156
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]157<p>LLVM 2.3 fully supports the llvm-gcc 4.2 front-end.</p>
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]158
159<p>llvm-gcc 4.2 includes numerous fixes to better support the Objective-C
160front-end. Objective-C now works very well on Mac OS/X.</p>
161
162<p>llvm-gcc 4.2 includes many other fixes which improve conformance with the
163relevant parts of the GCC testsuite.</p>
164
165<p>The <a href="http://clang.llvm.org/">clang project</a> is an effort to build
166a set of new 'llvm native' front-end technologies for the LLVM optimizer
167and code generator. Currently, its C and Objective-C support is maturing
168nicely, and it has advanced source-to-source analysis and transformation
169capabilities. If you are interested in building source-level tools for C and
170Objective-C (and eventually C++), you should take a look. However, note that
171clang is not an official part of the LLVM 2.3 release. If you are interested in
172this project, please see its <a href="http://clang.llvm.org/">web site</a>.</p>
173
174</div>
175
176
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]177<!--=========================================================================-->
178<div class="doc_subsection">
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]179<a name="coreimprovements">LLVM Core Improvements</a>
180</div>
181
182<div class="doc_text">
183<p>New features include:
184</p>
185
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]186
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]187Common linkage?
188
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]189Atomic operation support, Alpha, X86, X86-64, PowerPC. "__sync_synchronize",
Anton Korobeynikov877b73b2008-06-08 10:23:46 +0000[diff] [blame]190"__sync_val_compare_and_swap", etc
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]191
192<ul>
Gordon Henriksenf9b229b2008-06-05 12:51:50 +0000[diff] [blame]193<li>The C and Ocaml bindings have received additional improvements. The
194bindings now cover pass managers, several transformation passes, iteration
195over the LLVM IR, target data, and parameter attribute lists.</li>
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]196</ul>
197
198</div>
199
200<!--=========================================================================-->
201<div class="doc_subsection">
Chris Lattner8fda4952008-02-10 07:46:44 +0000[diff] [blame]202<a name="optimizer">Optimizer Improvements</a>
203</div>
204
205<div class="doc_text">
206
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]207<p>In addition to a huge array of bug fixes and minor performance tweaks, the
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]208LLVM 2.3 optimizers support a few major enhancements:</p>
Chris Lattner8fda4952008-02-10 07:46:44 +0000[diff] [blame]209
210<ul>
211
Devang Patel67e46712008-06-05 21:44:00 +0000[diff] [blame]212<li>Loop index set splitting on by default.<p>
213This transformation hoists conditions from loop bodies and reduces loop's
214iteration space to improve performance. For example, <p>
215<pre>
216for (i = LB; i < UB; ++i)
217 if (i <= NV)
218 LOOP_BODY
219</pre>
220is transformed into
221<pre>
222NUB = min(NV+1, UB)
223for (i = LB; i < NUB; ++i)
224 LOOP_BODY
225</pre>
226</li>
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]227
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]228<li>LLVM includes a new <tt>memcpy</tt> optimization pass removes
229dead <tt>memcpy</tt> calls, unneeded copies of aggregates, and performs
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]230return slot optimization. The LLVM optimizer now notices long sequences of
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]231consecutive stores and merges them into <tt>memcpy</tt>'s where profitable.</li>
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]232
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]233<li>Alignment detection for vector memory references and for <tt>memcpy</tt> and
234<tt>memset</tt> is now more aggressive.</li>
Owen Anderson7f8ea062008-06-06 16:23:15 +0000[diff] [blame]235
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]236<li>The aggressive dead code elimination (ADCE) optimization has been rewritten
237to make it both faster and safer in the presence of code containing infinite
238loops. Some of its prior functionality has been factored out into the loop
239deletion pass, which <em>is</em> safe for infinite loops.</li>
240
241<li>Several optimizations have been sped up, leading to faster code generation
242 with the same code quality.</li>
243
244<li>The 'SimplifyLibCalls' pass, which optimizes calls to libc and libm
245 functions for C-based languages, has been rewritten to be a FunctionPass
246 instead a ModulePass. This allows it to be run more often and to be
247 included at -O1 in llvm-gcc. It was also extended to include more
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]248 optimizations and several corner case bugs were fixed.</li>
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]249
Duncan Sands8b4f3772008-06-08 19:38:43 +0000[diff] [blame]250<li>LLVM now includes a simple 'Jump Threading' pass, which attempts to simplify
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]251 conditional branches using information about predecessor blocks, simplifying
252 the control flow graph. This pass is pretty basic at this point, but catches
253 some important cases and provides a foundation to build on.</li>
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]254</ul>
255
256</div>
257
258<!--=========================================================================-->
259<div class="doc_subsection">
260<a name="codegen">Code Generator Improvements</a>
261</div>
262
263<div class="doc_text">
264
265<p>We put a significant amount of work into the code generator infrastructure,
266which allows us to implement more aggressive algorithms and make it run
267faster:</p>
268
269<ul>
270<li>MemOperand in the code generator: describe me!.</li>
271
272<li>The target-independent code generator infrastructure now uses LLVM's APInt
273 class to handle integer values, which allows it to support integer types
274 larger than 64 bits. Note that support for such types is also dependent on
275 target-specific support. Use of APInt is also a step toward support for
276 non-power-of-2 integer sizes.</li>
277
278<li>Several compile time speedups for code with large basic blocks.</li>
279
280<li>Several improvements which make llc's <tt>--view-sunit-dags</tt>
281 visualization of scheduling dependency graphs easier to understand.</li>
282
283<li>The code generator allows targets to write patterns that generate subreg
284 references directly in .td files now.</li>
285
286<li><tt>memcpy</tt> lowering in the backend is more aggressive, particularly for
287 <tt>memcpy</tt> calls introduced by the code generator when handling
288 pass-by-value structure argument copies.</li>
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]289</ul>
290
291</div>
292
293
294<!--=========================================================================-->
295<div class="doc_subsection">
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]296<a name="x86specific">X86/X86-64 Specific Improvements</a>
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]297</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]298
299<div class="doc_text">
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]300<p>New target-specific features include:
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]301</p>
302
303<ul>
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]304<li>llvm-gcc's X86-64 ABI conformance is far improved, particularly in the
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]305 area of passing and returning structures by value. llvm-gcc compiled code
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]306 now interoperates very well on X86-64 systems with other compilers.</li>
307
Duncan Sands8b4f3772008-06-08 19:38:43 +0000[diff] [blame]308<li>Support for Win64 was added. This includes code generation itself, JIT
Anton Korobeynikov77d15e22008-06-08 10:24:13 +0000[diff] [blame]309 support and necessary changes to llvm-gcc.</li>
310
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]311<li>The LLVM X86 backend now supports the support SSE 4.1 instruction set, and
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]312 the llvm-gcc 4.2 front-end supports the SSE 4.1 compiler builtins. Various
313 generic vector operations (insert/extract/shuffle) are much more efficient
314 when SSE 4.1 is enabled. The JIT automatically takes advantage of these
315 instructions, but llvm-gcc must be explicitly told to use them, e.g. with
316 <tt>-march=penryn</tt>.</li>
Anton Korobeynikov77d15e22008-06-08 10:24:13 +0000[diff] [blame]317
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]318<li>The X86 backend now does a number of optimizations that aim to avoid
319 converting numbers back and forth from SSE registers to the X87 floating
320 point stack.</li>
Anton Korobeynikov77d15e22008-06-08 10:24:13 +0000[diff] [blame]321
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]322<li>The X86 backend supports stack realignment, which is particularly useful for
323 vector code on OS's without 16-byte aligned stacks.</li>
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]324
325<li>The X86 backend now supports the "sseregparm" options in GCC, which allow
326 functions to be tagged as passing floating point values in SSE
327 registers.</li>
328
329<li>Trampolines (taking the address of a nested function) now work on
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]330 Linux/X86-64.</li>
Anton Korobeynikov77d15e22008-06-08 10:24:13 +0000[diff] [blame]331
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]332<li><tt>__builtin_prefetch</tt> is now compiled into the appropriate prefetch
333 instructions instead of being ignored.</li>
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]334
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]335<li>128-bit integers are now supported on X86-64 targets.</li>
Dan Gohman74e1ff52008-06-05 16:15:39 +0000[diff] [blame]336
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]337<li>The register allocator can now rematerialize PIC-base computations.</li>
338
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]339</ul>
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]340
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]341</div>
342
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]343<!--=========================================================================-->
344<div class="doc_subsection">
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]345<a name="targetspecific">Other Target Specific Improvements</a>
346</div>
347
348<div class="doc_text">
349<p>New target-specific features include:
350</p>
351
352<ul>
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]353<li>The LLVM C backend now supports vector code.</li>
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]354
355
356</ul>
357
358</div>
359
360
361
362<!--=========================================================================-->
363<div class="doc_subsection">
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]364<a name="otherimprovements">Other Improvements</a>
365</div>
366
367<div class="doc_text">
368<p>New features include:
369</p>
370
371<ul>
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]372<li>LLVM now builds with GCC 4.3.</li>
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame]373<li><tt>llvm2cpp</tt> tool has been folded into llc, use
374 <tt>llc -march=cpp</tt></li>
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]375</ul>
376
377</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]378
379<!-- *********************************************************************** -->
380<div class="doc_section">
381 <a name="portability">Portability and Supported Platforms</a>
382</div>
383<!-- *********************************************************************** -->
384
385<div class="doc_text">
386
387<p>LLVM is known to work on the following platforms:</p>
388
389<ul>
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]390<li>Intel and AMD machines (IA32) running Red Hat Linux, Fedora Core and FreeBSD
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]391 (and probably other unix-like systems).</li>
Chris Lattnerbd633de92008-02-06 06:30:34 +0000[diff] [blame]392<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]393 64-bit modes.</li>
Chris Lattnere0798082008-02-06 18:00:06 +0000[diff] [blame]394<li>Intel and AMD machines running on Win32 using MinGW libraries (native).</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]395<li>Intel and AMD machines running on Win32 with the Cygwin libraries (limited
396 support is available for native builds with Visual C++).</li>
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]397<li>Sun UltraSPARC workstations running Solaris 10.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]398<li>Alpha-based machines running Debian GNU/Linux.</li>
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]399<li>Itanium-based (IA64) machines running Linux and HP-UX.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]400</ul>
401
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]402<p>The core LLVM infrastructure uses GNU autoconf to adapt itself
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]403to the machine and operating system on which it is built. However, minor
404porting may be required to get LLVM to work on new platforms. We welcome your
405portability patches and reports of successful builds or error messages.</p>
406
407</div>
408
409<!-- *********************************************************************** -->
410<div class="doc_section">
411 <a name="knownproblems">Known Problems</a>
412</div>
413<!-- *********************************************************************** -->
414
415<div class="doc_text">
416
417<p>This section contains all known problems with the LLVM system, listed by
418component. As new problems are discovered, they will be added to these
419sections. If you run into a problem, please check the <a
420href="http://llvm.org/bugs/">LLVM bug database</a> and submit a bug if
421there isn't already one.</p>
422
423</div>
424
425<!-- ======================================================================= -->
426<div class="doc_subsection">
427 <a name="experimental">Experimental features included with this release</a>
428</div>
429
430<div class="doc_text">
431
432<p>The following components of this LLVM release are either untested, known to
433be broken or unreliable, or are in early development. These components should
434not be relied on, and bugs should not be filed against them, but they may be
435useful 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]436components, please contact us on the <a
437href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVMdev list</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]438
439<ul>
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]440<li>The MSIL, IA64, Alpha, SPU, and MIPS backends are experimental.</li>
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]441<li>The llc "<tt>-filetype=asm</tt>" (the default) is the only supported
Chris Lattnerbd633de92008-02-06 06:30:34 +0000[diff] [blame]442 value for this option.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]443</ul>
444
445</div>
446
447<!-- ======================================================================= -->
448<div class="doc_subsection">
449 <a name="x86-be">Known problems with the X86 back-end</a>
450</div>
451
452<div class="doc_text">
453
454<ul>
Anton Korobeynikov77d15e22008-06-08 10:24:13 +0000[diff] [blame]455 <li>The X86 backend does not yet support
456 all <a href="http://llvm.org/PR879">inline assembly that uses the X86
457 floating point stack</a>. It supports the 'f' and 't' constraints, but not
458 'u'.</li>
459 <li>The X86 backend generates inefficient floating point code when configured to
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]460 generate code for systems that don't have SSE2.</li>
Duncan Sands8b4f3772008-06-08 19:38:43 +0000[diff] [blame]461 <li>Win64 code generation wasn't widely tested. Everything should work, but we
Anton Korobeynikov77d15e22008-06-08 10:24:13 +0000[diff] [blame]462 expect small issues to happen. Also, llvm-gcc cannot build mingw64 runtime
463 currently due
464 to <a href="http://llvm.org/PR2255">several</a>
465 <a href="http://llvm.org/PR2257">bugs</a> in FP stackifier
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]466</ul>
467
468</div>
469
470<!-- ======================================================================= -->
471<div class="doc_subsection">
472 <a name="ppc-be">Known problems with the PowerPC back-end</a>
473</div>
474
475<div class="doc_text">
476
477<ul>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]478<li>The Linux PPC32/ABI support needs testing for the interpreter and static
479compilation, and lacks support for debug information.</li>
480</ul>
481
482</div>
483
484<!-- ======================================================================= -->
485<div class="doc_subsection">
486 <a name="arm-be">Known problems with the ARM back-end</a>
487</div>
488
489<div class="doc_text">
490
491<ul>
492<li>Thumb mode works only on ARMv6 or higher processors. On sub-ARMv6
Duncan Sands1827fa62007-09-26 15:59:54 +0000[diff] [blame]493processors, thumb programs can crash or produce wrong
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]494results (<a href="http://llvm.org/PR1388">PR1388</a>).</li>
495<li>Compilation for ARM Linux OABI (old ABI) is supported, but not fully tested.
496</li>
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]497<li>There is a bug in QEMU-ARM (&lt;= 0.9.0) which causes it to incorrectly execute
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]498programs compiled with LLVM. Please use more recent versions of QEMU.</li>
499</ul>
500
501</div>
502
503<!-- ======================================================================= -->
504<div class="doc_subsection">
505 <a name="sparc-be">Known problems with the SPARC back-end</a>
506</div>
507
508<div class="doc_text">
509
510<ul>
511<li>The SPARC backend only supports the 32-bit SPARC ABI (-m32), it does not
512 support the 64-bit SPARC ABI (-m64).</li>
513</ul>
514
515</div>
516
517<!-- ======================================================================= -->
518<div class="doc_subsection">
519 <a name="alpha-be">Known problems with the Alpha back-end</a>
520</div>
521
522<div class="doc_text">
523
524<ul>
525
526<li>On 21164s, some rare FP arithmetic sequences which may trap do not have the
527appropriate nops inserted to ensure restartability.</li>
528
529</ul>
530</div>
531
532<!-- ======================================================================= -->
533<div class="doc_subsection">
534 <a name="ia64-be">Known problems with the IA64 back-end</a>
535</div>
536
537<div class="doc_text">
538
539<ul>
540
541<li>C++ programs are likely to fail on IA64, as calls to <tt>setjmp</tt> are
542made where the argument is not 16-byte aligned, as required on IA64. (Strictly
543speaking this is not a bug in the IA64 back-end; it will also be encountered
544when building C++ programs using the C back-end.)</li>
545
546<li>The C++ front-end does not use <a href="http://llvm.org/PR406">IA64
547ABI compliant layout of v-tables</a>. In particular, it just stores function
548pointers instead of function descriptors in the vtable. This bug prevents
549mixing C++ code compiled with LLVM with C++ objects compiled by other C++
550compilers.</li>
551
552<li>There are a few ABI violations which will lead to problems when mixing LLVM
553output with code built with other compilers, particularly for floating-point
554programs.</li>
555
Duncan Sands8b4f3772008-06-08 19:38:43 +0000[diff] [blame]556<li>Defining vararg functions is not supported (but calling them is OK).</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]557
558<li>The Itanium backend has bitrotted somewhat.</li>
559</ul>
560
561</div>
562
563<!-- ======================================================================= -->
564<div class="doc_subsection">
565 <a name="c-be">Known problems with the C back-end</a>
566</div>
567
568<div class="doc_text">
569
570<ul>
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]571<li><a href="http://llvm.org/PR802">The C backend has only basic support for
572 inline assembly code</a>.</li>
Chris Lattnere0194462007-09-26 06:01:35 +0000[diff] [blame]573<li><a href="http://llvm.org/PR1658">The C backend violates the ABI of common
574 C++ programs</a>, preventing intermixing between C++ compiled by the CBE and
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]575 C++ code compiled with llc or native compilers.</li>
Duncan Sandsbd0ad6e2008-02-10 13:40:55 +0000[diff] [blame]576<li>The C backend does not support all exception handling constructs.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]577</ul>
578
579</div>
580
581
582<!-- ======================================================================= -->
583<div class="doc_subsection">
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]584 <a name="c-fe">Known problems with the llvm-gcc C front-end</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]585</div>
586
587<!-- _______________________________________________________________________ -->
588<div class="doc_subsubsection">Bugs</div>
589
590<div class="doc_text">
591
Chris Lattnere0798082008-02-06 18:00:06 +0000[diff] [blame]592<p>llvm-gcc does not currently support <a href="http://llvm.org/PR869">Link-Time
593Optimization</a> on most platforms "out-of-the-box". Please inquire on the
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]594llvmdev mailing list if you are interested.</p>
595
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]596<p>The only major language feature of GCC not supported by llvm-gcc is
597 the <tt>__builtin_apply</tt> family of builtins. However, some extensions
598 are only supported on some targets. For example, trampolines are only
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]599 supported on some targets (these are used when you take the address of a
600 nested function).</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]601
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]602<p>If you run into GCC extensions which are not supported, please let us know.
603</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]604
605</div>
606
607<!-- ======================================================================= -->
608<div class="doc_subsection">
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]609 <a name="c++-fe">Known problems with the llvm-gcc C++ front-end</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]610</div>
611
612<div class="doc_text">
613
614<p>The C++ front-end is considered to be fully
615tested and works for a number of non-trivial programs, including LLVM
616itself, Qt, Mozilla, etc.</p>
617
618<ul>
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]619<li>Exception handling works well on the X86 and PowerPC targets, including
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]620X86-64 darwin. This works when linking to a libstdc++ compiled by GCC. It is
621supported on X86-64 linux, but that is disabled by default in this release.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]622</ul>
623
624</div>
625
626
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]627<!-- ======================================================================= -->
628<div class="doc_subsection">
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]629 <a name="ada-fe">Known problems with the llvm-gcc Ada front-end</a>
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]630</div>
631
632<div class="doc_text">
633The llvm-gcc 4.2 Ada compiler works fairly well, however this is not a mature
634technology and problems should be expected.
635<ul>
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]636<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]637to lack of trampoline support (pointers to nested functions) on other platforms,
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]638however 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]639which does support trampolines.</li>
640<li>The Ada front-end <a href="http://llvm.org/PR2007">fails to bootstrap</a>.
641Workaround: configure with --disable-bootstrap.</li>
642<li>The c380004 and <a href="http://llvm.org/PR2010">c393010</a> ACATS tests
Duncan Sandsa73561c2008-06-08 20:18:35 +0000[diff] [blame^]643fail (c380004 also fails with gcc-4.2 mainline). When built at -O3, the
644<a href="http://llvm.org/PR2421">cxg2021</a> ACATS test also fails.</li>
645<li>Some gcc specific Ada tests continue to crash the compiler. The testsuite
646reports most tests as having failed even though they pass.</li>
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]647<li>The -E binder option (exception backtraces)
648<a href="http://llvm.org/PR1982">does not work</a> and will result in programs
649crashing if an exception is raised. Workaround: do not use -E.</li>
650<li>Only discrete types <a href="http://llvm.org/PR1981">are allowed to start
651or finish at a non-byte offset</a> in a record. Workaround: do not pack records
652or use representation clauses that result in a field of a non-discrete type
653starting or finishing in the middle of a byte.</li>
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]654<li>The <tt>lli</tt> interpreter <a href="http://llvm.org/PR2009">considers 'main'
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]655as generated by the Ada binder to be invalid</a>.
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]656Workaround: hand edit the file to use pointers for <tt>argv</tt> and <tt>envp</tt> rather than
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]657integers.</li>
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]658<li>The <tt>-fstack-check</tt> option <a href="http://llvm.org/PR2008">is ignored</a>.</li>
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]659</ul>
660</div>
661
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]662<!-- *********************************************************************** -->
663<div class="doc_section">
664 <a name="additionalinfo">Additional Information</a>
665</div>
666<!-- *********************************************************************** -->
667
668<div class="doc_text">
669
670<p>A wide variety of additional information is available on the <a
671href="http://llvm.org">LLVM web page</a>, in particular in the <a
672href="http://llvm.org/docs/">documentation</a> section. The web page also
673contains versions of the API documentation which is up-to-date with the
674Subversion version of the source code.
675You can access versions of these documents specific to this release by going
676into the "<tt>llvm/doc/</tt>" directory in the LLVM tree.</p>
677
678<p>If you have any questions or comments about LLVM, please feel free to contact
679us via the <a href="http://llvm.org/docs/#maillist"> mailing
680lists</a>.</p>
681
682</div>
683
684<!-- *********************************************************************** -->
685
686<hr>
687<address>
688 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
689 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
690 <a href="http://validator.w3.org/check/referer"><img
Chris Lattner5ba16862008-02-10 08:18:42 +0000[diff] [blame]691 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]692
693 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
694 Last modified: $Date$
695</address>
696
697</body>
698</html>