Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 0eef6e4b957f4385f0789a8f3dbf4702757d094e / . / docs / ReleaseNotes.html
blob: 9d66b6d150670c500cdf4297be7e801758bd0cc8 [file] [log] [blame]
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<html>
4<head>
5 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
6 <link rel="stylesheet" href="llvm.css" type="text/css">
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]7 <title>LLVM 2.3 Release Notes</title>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]8</head>
9<body>
10
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]11<div class="doc_title">LLVM 2.3 Release Notes</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]12
13<ol>
14 <li><a href="#intro">Introduction</a></li>
15 <li><a href="#whatsnew">What's New?</a></li>
16 <li><a href="GettingStarted.html">Installation Instructions</a></li>
17 <li><a href="#portability">Portability and Supported Platforms</a></li>
18 <li><a href="#knownproblems">Known Problems</a>
19 <li><a href="#additionalinfo">Additional Information</a></li>
20</ol>
21
22<div class="doc_author">
23 <p>Written by the <a href="http://llvm.org">LLVM Team</a><p>
24</div>
25
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]26<!-- Done through Week-of-Mon-20080324.txt -->
27
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]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
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]37infrastructure, release 2.3. Here we describe the status of LLVM, including
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]38major improvements from the previous release and any known problems. All LLVM
39releases may be downloaded from the <a href="http://llvm.org/releases/">LLVM
40releases 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
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]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 releases, please see the
51<a href="http://llvm.org/releases/">releases page</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]52
53</div>
54
55<!-- *********************************************************************** -->
56<div class="doc_section">
57 <a name="whatsnew">What's New?</a>
58</div>
59<!-- *********************************************************************** -->
60
61<div class="doc_text">
62
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]63<p>This is the fourteenth public release of the LLVM Compiler Infrastructure.
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]64It includes a large number of features and refinements from LLVM 2.2.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]65
66</div>
67
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]68<!-- Unfinished features in 2.3:
Chris Lattner8fda4952008-02-10 07:46:44 +0000[diff] [blame]69 Machine LICM
70 Machine Sinking
71 LegalizeDAGTypes
72 -->
73
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]74<!--=========================================================================-->
75<div class="doc_subsection">
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]76<a name="changes">Major Changes in LLVM 2.3</a>
Chris Lattner41bf8e92008-02-10 07:04:35 +0000[diff] [blame]77</div>
78
79<div class="doc_text">
80
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]81<p>LLVM 2.2 was the last LLVM release to support llvm-gcc 4.0 and llvm-upgrade.
82llvm-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]83upgrading 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]84LLVM release to do this. One nice impact of this is that the LLVM regression
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]85test suite no longer depends on llvm-upgrade, which makes it run faster.</p>
Chris Lattnere0798082008-02-06 18:00:06 +0000[diff] [blame]86
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame^]87<p>LLVM API Changes:</p>
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]88
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame^]89<ul>
90<li>Several core LLVM IR classes have migrated to use 'FOOCLASS::Create(...)'
91 instead of 'new FOOCLASS(...)' (e.g. where FOOCLASS=BasicBlock). We hope to
92 standardize on FOOCLASS::Create for all IR classes in the future, but not
93 all of them have been moved over yet.</li>
94<li>LLVM 2.3 renames the LLVMBuilder and LLVMFoldingBuilder classes to
95 IRBuilder.</li>
96<li>MRegisterInfo was renamed to TargetRegisterInfo.</li>
97<li>The MappedFile class is gone, please use MemoryBuffer instead.</li>
98<li>The '-enable-eh' flag to LLC has been removed. Now code should encode
99 whether it is safe to not generate unwind information for a function by
100 tagging the Function object with the 'nounwind' attribute.</li>
101
102</ul>
Chris Lattnere0798082008-02-06 18:00:06 +0000[diff] [blame]103</div>
104
105<!--=========================================================================-->
106<div class="doc_subsection">
107<a name="majorfeatures">Major New Features</a>
108</div>
109
110<div class="doc_text">
111
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]112<p>LLVM 2.3 includes several major new capabilities:</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]113
114<ul>
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]115<li>Multiple Return Value Support</li>
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]116
117
118<li><p>LLVM 2.3 includes a complete reimplementation of the "llvmc" tool. It is
119designed to overcome several problems of the original llvmc and to provide a
120superset of the features as the 'gcc' driver.</p>
121
122<p>The main features of llvmc2 is:</p>
123
124<ul>
125<li>Extended handling of command line options and smart rules for
126dispatching them to different tools.</li>
127<li>Flexible (and extensible) rules for defining different tools.</li>
128<li>The different intermediate steps performed by tools are represented
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]129as edges in the abstract graph.</li>
130<li>The 'language' for driver behaviour definition is tablegen and thus
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]131it's relatively easy to add new features.</li>
132<li>The definition of driver is transformed into set of C++ classes, thus
133no runtime interpretation is needed.</li>
134</ul>
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]135</li>
136
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]137
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]138<li>Reimplemented <a href="LinkTimeOptimization.html">LTO interface</a> in
139 C.</li>
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]140
141
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]142<li>kaleidoscope tutorial in ocaml.</li>
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]143
Chris Lattner5ba16862008-02-10 08:18:42 +0000[diff] [blame]144</ul>
145
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]146</div>
147
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]148
149<!--=========================================================================-->
150<div class="doc_subsection">
151<a name="frontends">llvm-gcc 4.2 Improvements and Clang</a>
152</div>
153
154<div class="doc_text">
155
156<p>LLVM 2.3 fully supports llvm-gcc 4.2 front-end.</p>
157
158<p>llvm-gcc 4.2 includes numerous fixes to better support the Objective-C
159front-end. Objective-C now works very well on Mac OS/X.</p>
160
161<p>llvm-gcc 4.2 includes many other fixes which improve conformance with the
162relevant parts of the GCC testsuite.</p>
163
164<p>The <a href="http://clang.llvm.org/">clang project</a> is an effort to build
165a set of new 'llvm native' front-end technologies for the LLVM optimizer
166and code generator. Currently, its C and Objective-C support is maturing
167nicely, and it has advanced source-to-source analysis and transformation
168capabilities. If you are interested in building source-level tools for C and
169Objective-C (and eventually C++), you should take a look. However, note that
170clang is not an official part of the LLVM 2.3 release. If you are interested in
171this project, please see its <a href="http://clang.llvm.org/">web site</a>.</p>
172
173</div>
174
175
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]176<!--=========================================================================-->
177<div class="doc_subsection">
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]178<a name="coreimprovements">LLVM Core Improvements</a>
179</div>
180
181<div class="doc_text">
182<p>New features include:
183</p>
184
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]185
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]186Common linkage?
187
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame^]188Atomic operation support, Alpha, X86, X86-64, PowerPC. "__sync_synchronize",
189"__sync_val_compare_and_swap", etc"
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]190
191<ul>
Gordon Henriksenf9b229b2008-06-05 12:51:50 +0000[diff] [blame]192<li>The C and Ocaml bindings have received additional improvements. The
193bindings now cover pass managers, several transformation passes, iteration
194over the LLVM IR, target data, and parameter attribute lists.</li>
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]195</ul>
196
197</div>
198
199<!--=========================================================================-->
200<div class="doc_subsection">
Chris Lattner8fda4952008-02-10 07:46:44 +0000[diff] [blame]201<a name="optimizer">Optimizer Improvements</a>
202</div>
203
204<div class="doc_text">
205
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]206<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]207LLVM 2.3 optimizers support a few major enhancements:</p>
Chris Lattner8fda4952008-02-10 07:46:44 +0000[diff] [blame]208
209<ul>
210
Devang Patel67e46712008-06-05 21:44:00 +0000[diff] [blame]211<li>Loop index set splitting on by default.<p>
212This transformation hoists conditions from loop bodies and reduces loop's
213iteration space to improve performance. For example, <p>
214<pre>
215for (i = LB; i < UB; ++i)
216 if (i <= NV)
217 LOOP_BODY
218</pre>
219is transformed into
220<pre>
221NUB = min(NV+1, UB)
222for (i = LB; i < NUB; ++i)
223 LOOP_BODY
224</pre>
225</li>
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame^]226
227<li>LLVM includes a new <tt>memcpy</tt> optimization pass which optimizes out
228dead <tt>memcpy</tt> calls, unneeded copies of aggregates, and handles the
229return slot optimization. The LLVM optimizer now notices long sequences of
230consequtive stores and merges them into <tt>memcpy</tt>s where profitable.</li>
231
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]232<li>Alignment detection for vector memory references and for <tt>memcpy</tt> and
233<tt>memset</tt> is now more aggressive.</li>
Owen Anderson7f8ea062008-06-06 16:23:15 +0000[diff] [blame]234
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame^]235<li>The aggressive dead code elimination (ADCE) optimization has been rewritten
236to make it both faster and safer in the presence of code containing infinite
237loops. Some of its prior functionality has been factored out into the loop
238deletion pass, which <em>is</em> safe for infinite loops.</li>
239
240<li>Several optimizations have been sped up, leading to faster code generation
241 with the same code quality.</li>
242
243<li>The 'SimplifyLibCalls' pass, which optimizes calls to libc and libm
244 functions for C-based languages, has been rewritten to be a FunctionPass
245 instead a ModulePass. This allows it to be run more often and to be
246 included at -O1 in llvm-gcc. It was also extended to include more
247 optimizations and several corner case bugs are fixed.</li>
248
249<li>LLVM now includes a simple 'Jump Threading' pass, which attemps to simplify
250 the conditional branches with information about predecessor blocks. This
251 simplifies the control flow graph. This pass is pretty basic at this point,
252 but catches some important cases and provides a foundation to build off
253 of.</li>
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
308<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^]309 the llvm-gcc 4.2 front-end supports the SSE 4.1 compiler builtins. Various
310 generic vector operations (insert/extract/shuffle) are much more efficient
311 when SSE 4.1 is enabled. The JIT automatically takes advantage of these
312 instructions, but llvm-gcc must be explicitly told to use them, e.g. with
313 <tt>-march=penryn</tt>.</li>
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]314
315<li>The X86 backend now does a number of optimizations that aim to avoid
316 converting numbers back and forth from SSE registers to the X87 floating
317 point stack.</li>
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]318
319<li>The X86 backend supports stack realignment, which is particularly useful for
320 vector code on OS's without 16-byte aligned stacks.</li>
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]321
322<li>The X86 backend now supports the "sseregparm" options in GCC, which allow
323 functions to be tagged as passing floating point values in SSE
324 registers.</li>
325
326<li>Trampolines (taking the address of a nested function) now work on
327 Linux/x86-64.</li>
328
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]329<li><tt>__builtin_prefetch</tt> is now compiled into the appropriate prefetch
330 instructions instead of being ignored.</li>
Chris Lattnerc5a4e312008-06-05 06:25:56 +0000[diff] [blame]331
Dan Gohman74e1ff52008-06-05 16:15:39 +0000[diff] [blame]332<li>128-bit integers are now supported on x86-64 targets.</li>
333
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame^]334<li>The register allocator can now rematerialize PIC-base computations.</li>
335
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]336</ul>
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]337
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]338</div>
339
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]340<!--=========================================================================-->
341<div class="doc_subsection">
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]342<a name="targetspecific">Other Target Specific Improvements</a>
343</div>
344
345<div class="doc_text">
346<p>New target-specific features include:
347</p>
348
349<ul>
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]350<li>The LLVM C backend now supports vector code.</li>
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]351
352
353</ul>
354
355</div>
356
357
358
359<!--=========================================================================-->
360<div class="doc_subsection">
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]361<a name="otherimprovements">Other Improvements</a>
362</div>
363
364<div class="doc_text">
365<p>New features include:
366</p>
367
368<ul>
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]369<li>LLVM now builds with GCC 4.3.</li>
Chris Lattner0eef6e42008-06-08 02:45:07 +0000[diff] [blame^]370<li><tt>llvm2cpp</tt> tool has been folded into llc, use
371 <tt>llc -march=cpp</tt></li>
Chris Lattnerb0d6c502007-09-21 03:54:09 +0000[diff] [blame]372</ul>
373
374</div>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]375
376<!-- *********************************************************************** -->
377<div class="doc_section">
378 <a name="portability">Portability and Supported Platforms</a>
379</div>
380<!-- *********************************************************************** -->
381
382<div class="doc_text">
383
384<p>LLVM is known to work on the following platforms:</p>
385
386<ul>
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]387<li>Intel and AMD machines (IA32) running Red Hat Linux, Fedora Core and FreeBSD
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]388 (and probably other unix-like systems).</li>
Chris Lattnerbd633de92008-02-06 06:30:34 +0000[diff] [blame]389<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]390 64-bit modes.</li>
Chris Lattnere0798082008-02-06 18:00:06 +0000[diff] [blame]391<li>Intel and AMD machines running on Win32 using MinGW libraries (native).</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]392<li>Intel and AMD machines running on Win32 with the Cygwin libraries (limited
393 support is available for native builds with Visual C++).</li>
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]394<li>Sun UltraSPARC workstations running Solaris 10.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]395<li>Alpha-based machines running Debian GNU/Linux.</li>
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]396<li>Itanium-based (IA64) machines running Linux and HP-UX.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]397</ul>
398
Chris Lattnerba948522008-06-05 06:57:39 +0000[diff] [blame]399<p>The core LLVM infrastructure uses GNU autoconf to adapt itself
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]400to the machine and operating system on which it is built. However, minor
401porting may be required to get LLVM to work on new platforms. We welcome your
402portability patches and reports of successful builds or error messages.</p>
403
404</div>
405
406<!-- *********************************************************************** -->
407<div class="doc_section">
408 <a name="knownproblems">Known Problems</a>
409</div>
410<!-- *********************************************************************** -->
411
412<div class="doc_text">
413
414<p>This section contains all known problems with the LLVM system, listed by
415component. As new problems are discovered, they will be added to these
416sections. If you run into a problem, please check the <a
417href="http://llvm.org/bugs/">LLVM bug database</a> and submit a bug if
418there isn't already one.</p>
419
420</div>
421
422<!-- ======================================================================= -->
423<div class="doc_subsection">
424 <a name="experimental">Experimental features included with this release</a>
425</div>
426
427<div class="doc_text">
428
429<p>The following components of this LLVM release are either untested, known to
430be broken or unreliable, or are in early development. These components should
431not be relied on, and bugs should not be filed against them, but they may be
432useful 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]433components, please contact us on the <a
434href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVMdev list</a>.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]435
436<ul>
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]437<li>The MSIL, IA64, Alpha, SPU, and MIPS backends are experimental.</li>
Chris Lattnerbd633de92008-02-06 06:30:34 +0000[diff] [blame]438<li>The LLC "<tt>-filetype=asm</tt>" (the default) is the only supported
439 value for this option.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]440</ul>
441
442</div>
443
444<!-- ======================================================================= -->
445<div class="doc_subsection">
446 <a name="x86-be">Known problems with the X86 back-end</a>
447</div>
448
449<div class="doc_text">
450
451<ul>
Chris Lattnere45f18b2008-06-05 08:02:49 +0000[diff] [blame]452<li>The X86 backend does not yet support all <a
453 href="http://llvm.org/PR879">inline assembly that uses the X86 floating
454 point stack</a>. It supports the 'f' and 't' constraints, but not 'u'.</li>
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]455<li>The X86 backend generates inefficient floating point code when configured to
456 generate code for systems that don't have SSE2.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]457</ul>
458
459</div>
460
461<!-- ======================================================================= -->
462<div class="doc_subsection">
463 <a name="ppc-be">Known problems with the PowerPC back-end</a>
464</div>
465
466<div class="doc_text">
467
468<ul>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]469<li>The Linux PPC32/ABI support needs testing for the interpreter and static
470compilation, and lacks support for debug information.</li>
471</ul>
472
473</div>
474
475<!-- ======================================================================= -->
476<div class="doc_subsection">
477 <a name="arm-be">Known problems with the ARM back-end</a>
478</div>
479
480<div class="doc_text">
481
482<ul>
483<li>Thumb mode works only on ARMv6 or higher processors. On sub-ARMv6
Duncan Sands1827fa62007-09-26 15:59:54 +0000[diff] [blame]484processors, thumb programs can crash or produce wrong
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]485results (<a href="http://llvm.org/PR1388">PR1388</a>).</li>
486<li>Compilation for ARM Linux OABI (old ABI) is supported, but not fully tested.
487</li>
Chris Lattnerc3649452008-02-10 08:17:19 +0000[diff] [blame]488<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]489programs compiled with LLVM. Please use more recent versions of QEMU.</li>
490</ul>
491
492</div>
493
494<!-- ======================================================================= -->
495<div class="doc_subsection">
496 <a name="sparc-be">Known problems with the SPARC back-end</a>
497</div>
498
499<div class="doc_text">
500
501<ul>
502<li>The SPARC backend only supports the 32-bit SPARC ABI (-m32), it does not
503 support the 64-bit SPARC ABI (-m64).</li>
504</ul>
505
506</div>
507
508<!-- ======================================================================= -->
509<div class="doc_subsection">
510 <a name="alpha-be">Known problems with the Alpha back-end</a>
511</div>
512
513<div class="doc_text">
514
515<ul>
516
517<li>On 21164s, some rare FP arithmetic sequences which may trap do not have the
518appropriate nops inserted to ensure restartability.</li>
519
520</ul>
521</div>
522
523<!-- ======================================================================= -->
524<div class="doc_subsection">
525 <a name="ia64-be">Known problems with the IA64 back-end</a>
526</div>
527
528<div class="doc_text">
529
530<ul>
531
532<li>C++ programs are likely to fail on IA64, as calls to <tt>setjmp</tt> are
533made where the argument is not 16-byte aligned, as required on IA64. (Strictly
534speaking this is not a bug in the IA64 back-end; it will also be encountered
535when building C++ programs using the C back-end.)</li>
536
537<li>The C++ front-end does not use <a href="http://llvm.org/PR406">IA64
538ABI compliant layout of v-tables</a>. In particular, it just stores function
539pointers instead of function descriptors in the vtable. This bug prevents
540mixing C++ code compiled with LLVM with C++ objects compiled by other C++
541compilers.</li>
542
543<li>There are a few ABI violations which will lead to problems when mixing LLVM
544output with code built with other compilers, particularly for floating-point
545programs.</li>
546
547<li>Defining vararg functions is not supported (but calling them is ok).</li>
548
549<li>The Itanium backend has bitrotted somewhat.</li>
550</ul>
551
552</div>
553
554<!-- ======================================================================= -->
555<div class="doc_subsection">
556 <a name="c-be">Known problems with the C back-end</a>
557</div>
558
559<div class="doc_text">
560
561<ul>
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]562<li><a href="http://llvm.org/PR802">The C backend has only basic support for
563 inline assembly code</a>.</li>
Chris Lattnere0194462007-09-26 06:01:35 +0000[diff] [blame]564<li><a href="http://llvm.org/PR1658">The C backend violates the ABI of common
565 C++ programs</a>, preventing intermixing between C++ compiled by the CBE and
566 C++ code compiled with LLC or native compilers.</li>
Duncan Sandsbd0ad6e2008-02-10 13:40:55 +0000[diff] [blame]567<li>The C backend does not support all exception handling constructs.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]568</ul>
569
570</div>
571
572
573<!-- ======================================================================= -->
574<div class="doc_subsection">
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]575 <a name="c-fe">Known problems with the llvm-gcc C front-end</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]576</div>
577
578<!-- _______________________________________________________________________ -->
579<div class="doc_subsubsection">Bugs</div>
580
581<div class="doc_text">
582
Chris Lattnere0798082008-02-06 18:00:06 +0000[diff] [blame]583<p>llvm-gcc does not currently support <a href="http://llvm.org/PR869">Link-Time
584Optimization</a> on most platforms "out-of-the-box". Please inquire on the
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]585llvmdev mailing list if you are interested.</p>
586
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]587<p>The only major language feature of GCC not supported by llvm-gcc is
588 the <tt>__builtin_apply</tt> family of builtins. However, some extensions
589 are only supported on some targets. For example, trampolines are only
590 supported on some targets, which are used when you take the address of a
591 nested function.</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]592
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]593<p>If you run into GCC extensions which are not supported, please let us know.
594</p>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]595
596</div>
597
598<!-- ======================================================================= -->
599<div class="doc_subsection">
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]600 <a name="c++-fe">Known problems with the llvm-gcc C++ front-end</a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]601</div>
602
603<div class="doc_text">
604
605<p>The C++ front-end is considered to be fully
606tested and works for a number of non-trivial programs, including LLVM
607itself, Qt, Mozilla, etc.</p>
608
609<ul>
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]610<li>Exception handling works well on the X86 and PowerPC targets, including
611x86-64 darwin. This works when linking to a libstdc++ compiled by GCC. It is
612supported on x86-64 linux, but that is disabled by default in this release.</li>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]613</ul>
614
615</div>
616
617
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]618<!-- ======================================================================= -->
619<div class="doc_subsection">
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]620 <a name="ada-fe">Known problems with the llvm-gcc Ada front-end</a>
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]621</div>
622
623<div class="doc_text">
624The llvm-gcc 4.2 Ada compiler works fairly well, however this is not a mature
625technology and problems should be expected.
626<ul>
627<li>The Ada front-end currently only builds on x86-32. This is mainly due
628to lack of trampoline support (pointers to nested functions) on other platforms,
629however it <a href="http://llvm.org/PR2006">also fails to build on x86-64</a>
630which does support trampolines.</li>
631<li>The Ada front-end <a href="http://llvm.org/PR2007">fails to bootstrap</a>.
632Workaround: configure with --disable-bootstrap.</li>
633<li>The c380004 and <a href="http://llvm.org/PR2010">c393010</a> ACATS tests
634fail (c380004 also fails with gcc-4.2 mainline).</li>
635<li>Many gcc specific Ada tests continue to crash the compiler.</li>
636<li>The -E binder option (exception backtraces)
637<a href="http://llvm.org/PR1982">does not work</a> and will result in programs
638crashing if an exception is raised. Workaround: do not use -E.</li>
639<li>Only discrete types <a href="http://llvm.org/PR1981">are allowed to start
640or finish at a non-byte offset</a> in a record. Workaround: do not pack records
641or use representation clauses that result in a field of a non-discrete type
642starting or finishing in the middle of a byte.</li>
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]643<li>The <tt>lli</tt> interpreter <a href="http://llvm.org/PR2009">considers 'main'
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]644as generated by the Ada binder to be invalid</a>.
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]645Workaround: 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]646integers.</li>
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]647<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]648</ul>
649</div>
650
651<!-- ======================================================================= -->
652<div class="doc_subsection">
Chris Lattner499c07c2008-06-05 06:35:40 +0000[diff] [blame]653 <a name="fortran-fe">Known problems with the llvm-gcc Fortran front-end</a>
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]654</div>
655
656<div class="doc_text">
657
658<ul>
659<li>The llvm-gcc 4.2 gfortran front-end supports a broad range of Fortran code, but does
Gabor Greif5e75e652008-06-05 18:39:01 +0000[diff] [blame]660<a href="http://llvm.org/PR1971">not support <tt>EQUIVALENCE</tt> yet</a>.</li>
Chris Lattnerdaeb3232008-02-12 06:29:45 +0000[diff] [blame]661</ul>
662</div>
663
664
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]665
666<!-- *********************************************************************** -->
667<div class="doc_section">
668 <a name="additionalinfo">Additional Information</a>
669</div>
670<!-- *********************************************************************** -->
671
672<div class="doc_text">
673
674<p>A wide variety of additional information is available on the <a
675href="http://llvm.org">LLVM web page</a>, in particular in the <a
676href="http://llvm.org/docs/">documentation</a> section. The web page also
677contains versions of the API documentation which is up-to-date with the
678Subversion version of the source code.
679You can access versions of these documents specific to this release by going
680into the "<tt>llvm/doc/</tt>" directory in the LLVM tree.</p>
681
682<p>If you have any questions or comments about LLVM, please feel free to contact
683us via the <a href="http://llvm.org/docs/#maillist"> mailing
684lists</a>.</p>
685
686</div>
687
688<!-- *********************************************************************** -->
689
690<hr>
691<address>
692 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
693 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
694 <a href="http://validator.w3.org/check/referer"><img
Chris Lattner5ba16862008-02-10 08:18:42 +0000[diff] [blame]695 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!"></a>
Dan Gohmanf17a25c2007-07-18 16:29:46 +0000[diff] [blame]696
697 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
698 Last modified: $Date$
699</address>
700
701</body>
702</html>