Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 976248d83610eec6ae8b166e9e9f991b9016f544 / . / docs / ReleaseNotes-2.6.html
blob: 1ac46e8c36cf83aa4ecc9e0f4760c65f1bd03146 [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>
54
Chris Lattnerc758fec2009-10-05 02:12:39 +0000[diff] [blame]55 Many new papers added to /pubs/
56 Machine LICM, hoists things like constant pool loads, loads from readonly stubs, vector constant synthesization code, etc.
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]57 Machine Sinking
Chris Lattnerc758fec2009-10-05 02:12:39 +0000[diff] [blame]58 Regalloc improvements for commuting, various spiller peephole optimizations, cross-class coalescing.
59 Support for debug line numbers when optimization enabled
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]60 gold lto plugin
Chris Lattnerc758fec2009-10-05 02:12:39 +0000[diff] [blame]61
62 target-specific intrinsics (r63765)
63 <tt>llc -enable-value-prop</tt>, propagation of value info (sign/zero ext info) from one MBB to another
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]64 interpreter + libffi
Chris Lattnerc758fec2009-10-05 02:12:39 +0000[diff] [blame]65 LLVMContext
66 Preliminary support for addrspace 256 -> GS, 257 -> FS, known problems: CodeGenerator.html#x86_memory
67
68Add support for the PowerPC 64-bit SVR4 ABI.
69
70 NSW/NUW/exact div
71 Inbounds for GEP
72 SRoA improvements for vector unions, memset, arbitrary weird bitfield accesses etc. It now produces "strange" sized integers.
73 pre-alloc splitter??
Chris Lattnerc758fec2009-10-05 02:12:39 +0000[diff] [blame]74 X86: Support for softfloat modes, typically used by OS kernels.
75 MC:
76 MCSection, MCAsmInfo
77 MCInstPrinter did it make it in?
78 MCInst (X86 using it so far)
79 Rewrite of X86 GV selection logic: TargetOperand flags on ExternalSymbol, GV, etc operands.
80 Can parse and re-print out an darwin-x86 .s file.
81 TargetLoweringObjectFile, MCSectionKind
82 Verrrry early start of a macho writer.
83
84 ELF Writer? How stable?
85
86 LSR promotes int induction variables to 64-bit on 64-bit targets, major perf boost for numerical code.
87 LSR now analyzes pointer expressions (e.g. getelementptrs), not just integers.
88 Stack slot coloring for register spills (denser stack frames)
89 SelectionDAGS: New BuildVectorSDNode (r65296), and ISD::VECTOR_SHUFFLE (r69952 / PR2957)
90 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.
91 new linkage types linkonce_odr, weak_odr, linker_private, and available_externally.
92 Inliner reuse stack space when inlining arrays?
93
94 Regalloc hints for allocation stuff: Evan r73381/r73671. Finished/enabled?
95 API Cleanup:
96 no use of hash_set/hash_map, no more llvm::OStream
97 Use raw_ostream for everything, killed off llvm/Streams.h and DOUT
98
99 Mips now supports O32 Calling Convention
100
101 StringRef class, Twine class.
102 New BlackFin backend.
103
104 Shrink wrapping support in PEI, what is the state of it?
105
106 X86-64: better modeling of implicit zero extensions, eliminates a lot of redundant zexts
107 X86-64 TLS support for local exec and initial exec.
108 X86 - Better modeling of H registerts as subregs.
109 Getelementpr instruction now allows any integer type for array/pointer indexes.
110
111 include/llvm/Analysis/LiveValues.h => dead??
112 lib/Analysis/LoopVR.cpp ==> dead??
113 include/llvm/CodeGen/LazyLiveness.h ==> dead?
114 lib/Transforms/IPO/MergeFunctions.cpp ==> dead?
115 llvm/Analysis/PointerTracking.h ==> dead??
116 PRedSimplify -> Ask vmkit if it is still useful and for testcases.
117
118 ARM calling convention code is now tblgen generated instead of manual.
119 ARM: NEON support. neonfp for doing single precision fp with neon instead of VFP.
120 Tblgen now supports multiclass inheritance.
121 Unladen swallow as user?
122 Loop index split disabled by default?
123
124 New WeakVH and AssertingVH and CallbackVH classes.
125 New llvm/ADT/Triple class.
126 Removed the IA64 backend.
127 Profile info improvements by Andreas Neustifter.
128 PostRA scheduler improvements David Goodwin.
129
130 New MSP430 and SystemZ backends.
131 llvm-gcc now supports a new TCE target.
132 klee web page at klee.llvm.org
133 New llvm/System/Atomic.h, llvm/System/RWMutex.h for portable atomic ops, rw locks.
134 llvm_start_multithreaded: ProgrammersMAnual.html#threading
135
136 Tablegen now supports a number of new string and list operations like
137 !(subst), !(foreach), !car, !cdr, !null, !if, !cast.
138 New fadd, fsub, fmul instructions and classes.
139 New MachineVerifier pass.
140 Enabled GVN Load PRE.
141 ARM AAPCS-VFP hard float ABI is supported.
142
143 LLVM build now builds all libraries as .a files instead of some
144 libraries as relinked .o files. This requires some APIs like
145 InitializeAllTargets.h. TargetRegistry!
146
147 ARM Thumb2 support: status?
148 CBE status: not part of the release criteria.
149
150 New SourceMgr, SMLoc classes for simple parsers with caret diagnostics and #include support, (used by
151 tablegen, llvm-mc, the .ll parser, FileCheck, etc)
152 FileCheck! + CHECK-NEXT
153 New compiler-rt project.
154 New Static Single Information (SSI) construction pass (not used by anything yet, experimental).
155
156 llvm_report_error() error handling API (llvm/Support/ErrorHandling.h)
157
158 x86: Vector icmp/fcmp now work with SSE codegen.
159 X86: all global variable reference logic is now in ClassifyGlobalReference.
160 JIT support for oprofile (r75279), configure with --with-oprofile. Now we get line # and function info for JIT'd functions.
161 Mention gcc plugin.
162
163 New EngineBuilder class for creating JITs: r76276 Reid Kleckner <reid at kleckner.net>
164
165 -asm-verbose now prints location info (with -g) and loop nest info.
166 JIT now supports generating more than 16M of code.
167
168removed the BigBlock register allocator, it had bitrotted.
169Target intrinsics can now return multiple results.
170
171 SSE 4.2 support.
172 Ada bindings for LLVM IR.
173 Many extensions to the C APIs.
174
175<!-- Unfinished features in 2.6:
176 Mention gcc plugin.
177 strong phi elim
178 variable debug info for optimized code
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]179 postalloc scheduler: anti dependence breaking, hazard recognizer?
Chris Lattnerc758fec2009-10-05 02:12:39 +0000[diff] [blame]180 metadata
181 loop dependence analysis
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]182 -->
183
184 <!-- for announcement email:
Chris Lattnerc758fec2009-10-05 02:12:39 +0000[diff] [blame]185 Logo web page.
186 llvm devmtg
187 compiler_rt
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]188 -->
189
190<!-- *********************************************************************** -->
191<div class="doc_section">
192 <a name="subproj">Sub-project Status Update</a>
193</div>
194<!-- *********************************************************************** -->
195
196<div class="doc_text">
197<p>
198The LLVM 2.6 distribution currently consists of code from the core LLVM
199repository &mdash;which roughly includes the LLVM optimizers, code generators
200and supporting tools &mdash; and the llvm-gcc repository. In addition to this
201code, the LLVM Project includes other sub-projects that are in development. The
202two which are the most actively developed are the <a href="#clang">Clang
203Project</a> and the <a href="#vmkit">VMKit Project</a>.
204</p>
205
206</div>
207
208
209<!--=========================================================================-->
210<div class="doc_subsection">
211<a name="clang">Clang: C/C++/Objective-C Frontend Toolkit</a>
212</div>
213
214<div class="doc_text">
215
216<p>The <a href="http://clang.llvm.org/">Clang project</a> is an effort to build
217a set of new 'LLVM native' front-end technologies for the LLVM optimizer and
218code generator. While Clang is not included in the LLVM 2.6 release, it is
219continuing to make major strides forward in all areas. Its C and Objective-C
220parsing and code generation support is now very solid. For example, it is
221capable of successfully building many real-world applications for X86-32
222and X86-64,
223including the <a href="http://wiki.freebsd.org/BuildingFreeBSDWithClang">FreeBSD
224kernel</a> and <a href="http://gcc.gnu.org/gcc-4.2/">gcc 4.2</a>. C++ is also
225making <a href="http://clang.llvm.org/cxx_status.html">incredible progress</a>,
226and work on templates has recently started. If you are
227interested in fast compiles and good diagnostics, we encourage you to try it out
228by <a href="http://clang.llvm.org/get_started.html">building from mainline</a>
229and reporting any issues you hit to the <a
230href="http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev">Clang front-end mailing
231list</a>.</p>
232
233<p>In the LLVM 2.6 time-frame, the Clang team has made many improvements:</p>
234
235<ul>
236<li>Something wonderful!</li>
Chris Lattnerc32a5322009-09-30 06:27:22 +0000[diff] [blame]237<li>AuroraUX / FreeBSD &amp; OpenBSD Toolchain support.</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]238<li>Many many bugs are fixed and many features have been added.</li>
239</ul>
240</div>
241
242<!--=========================================================================-->
243<div class="doc_subsection">
244<a name="clangsa">Clang Static Analyzer</a>
245</div>
246
247<div class="doc_text">
248
249<p>Previously announced in the 2.4 LLVM release, the Clang project also
250includes an early stage static source code analysis tool for <a
251href="http://clang.llvm.org/StaticAnalysis.html">automatically finding bugs</a>
252in C and Objective-C programs. The tool performs a growing set of checks to find
253bugs that occur on a specific path within a program.</p>
254
255<p>In the LLVM 2.6 time-frame there have been many significant improvements to
256XYZ.</p>
257
258<p>The set of checks performed by the static analyzer continues to expand, and
259future plans for the tool include full source-level inter-procedural analysis
260and deeper checks such as buffer overrun detection. There are many opportunities
261to extend and enhance the static analyzer, and anyone interested in working on
262this project is encouraged to get involved!</p>
263
264</div>
265
266<!--=========================================================================-->
267<div class="doc_subsection">
268<a name="vmkit">VMKit: JVM/CLI Virtual Machine Implementation</a>
269</div>
270
271<div class="doc_text">
272<p>
273The <a href="http://vmkit.llvm.org/">VMKit project</a> is an implementation of
274a JVM and a CLI Virtual Machines (Microsoft .NET is an
275implementation of the CLI) using the Just-In-Time compiler of LLVM.</p>
276
277<p>Following LLVM 2.6, VMKit has its XYZ release that you can find on its
278<a href="http://vmkit.llvm.org/releases/">webpage</a>. The release includes
279bug fixes, cleanup and new features. The major changes are:</p>
280
281<ul>
282
283<li>Something wonderful!</li>
284
285</ul>
286</div>
287
288<!-- *********************************************************************** -->
289<div class="doc_section">
290 <a name="externalproj">External Projects Using LLVM 2.6</a>
291</div>
292<!-- *********************************************************************** -->
293
Chris Lattnerc32a5322009-09-30 06:27:22 +0000[diff] [blame]294
295<!--=========================================================================-->
296<div class="doc_subsection">
297<a name="macruby">MacRuby</a>
298</div>
299
300<div class="doc_text">
301
302<p>
303<a href="http://macruby.org">MacRuby</a> is an implementation of Ruby on top of
304core Mac OS X technologies, such as the Objective-C common runtime and garbage
305collector, and the CoreFoundation framework. It is principally developed by
306Apple and aims at enabling the creation of full-fledged Mac OS X applications.
307</p>
308
309<p>
310MacRuby uses LLVM for optimization passes, JIT and AOT compilation of Ruby
311expressions. It also uses zero-cost DWARF exceptions to implement Ruby exception
312handling.</p>
313
314</div>
315
316
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]317<!--=========================================================================-->
318<div class="doc_subsection">
319<a name="pure">Pure</a>
320</div>
321
322<div class="doc_text">
323<p>
324<a href="http://pure-lang.googlecode.com/">Pure</a>
325is an algebraic/functional programming language based on term rewriting.
326Programs are collections of equations which are used to evaluate expressions in
327a symbolic fashion. Pure offers dynamic typing, eager and lazy evaluation,
328lexical closures, a hygienic macro system (also based on term rewriting),
329built-in list and matrix support (including list and matrix comprehensions) and
330an easy-to-use C interface. The interpreter uses LLVM as a backend to
331 JIT-compile Pure programs to fast native code.</p>
332
333<p>In addition to the usual algebraic data structures, Pure also has
334MATLAB-style matrices in order to support numeric computations and signal
335processing in an efficient way. Pure is mainly aimed at mathematical
336applications right now, but it has been designed as a general purpose language.
337The dynamic interpreter environment and the C interface make it possible to use
338it as a kind of functional scripting language for many application areas.
339</p>
340</div>
341
342
343<!--=========================================================================-->
344<div class="doc_subsection">
345<a name="ldc">LLVM D Compiler</a>
346</div>
347
348<div class="doc_text">
349<p>
350<a href="http://www.dsource.org/projects/ldc">LDC</a> is an implementation of
351the D Programming Language using the LLVM optimizer and code generator.
352The LDC project works great with the LLVM 2.6 release. General improvements in
353this
354cycle have included new inline asm constraint handling, better debug info
355support, general bugfixes, and better x86-64 support. This has allowed
356some major improvements in LDC, getting us much closer to being as
357fully featured as the original DMD compiler from DigitalMars.
358</p>
359</div>
360
361<!--=========================================================================-->
362<div class="doc_subsection">
363<a name="RoadsendPHP">Roadsend PHP</a>
364</div>
365
366<div class="doc_text">
367<p><a href="http://code.roadsend.com/rphp">Roadsend PHP</a> (rphp) is an open
368source implementation of the PHP programming
369language that uses LLVM for its optimizer, JIT, and static compiler. This is a
370reimplementation of an earlier project that is now based on LLVM.</p>
371</div>
372
Jeffrey Yasskinba2aa782009-06-24 21:09:13 +0000[diff] [blame]373<!--=========================================================================-->
374<div class="doc_subsection">
375<a name="Unladen Swallow">Unladen Swallow</a>
376</div>
377
378<div class="doc_text">
379<p><a href="http://code.google.com/p/unladen-swallow/">Unladen Swallow</a> is a
380branch of <a href="http://python.org/">Python</a> intended to be fully
381compatible and significantly faster. It uses LLVM's optimization passes and JIT
382compiler.</p>
383</div>
384
Jeffrey Yasskin5e98cb72009-06-24 21:26:42 +0000[diff] [blame]385<!--=========================================================================-->
386<div class="doc_subsection">
387<a name="Rubinius">Rubinius</a>
388</div>
389
390<div class="doc_text">
391<p><a href="http://github.com/evanphx/rubinius">Rubinius</a> is a new virtual
392machine for Ruby. It leverages LLVM to dynamically compile Ruby code down to
393machine code using LLVM's JIT.</p>
394</div>
395
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]396
397<!-- *********************************************************************** -->
398<div class="doc_section">
399 <a name="whatsnew">What's New in LLVM 2.6?</a>
400</div>
401<!-- *********************************************************************** -->
402
403<div class="doc_text">
404
405<p>This release includes a huge number of bug fixes, performance tweaks, and
406minor improvements. Some of the major improvements and new features are listed
407in this section.
408</p>
409</div>
410
411<!--=========================================================================-->
412<div class="doc_subsection">
413<a name="majorfeatures">Major New Features</a>
414</div>
415
416<div class="doc_text">
417
418<p>LLVM 2.6 includes several major new capabilities:</p>
419
420<ul>
421<li>Something wonderful!</li>
Edward O'Callaghan50310f92009-08-22 02:17:22 +0000[diff] [blame]422<li>LLVM 2.6 includes a brand new experimental LLVM bindings to the Ada2005 programming language.</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]423</ul>
424
425</div>
426
427
428<!--=========================================================================-->
429<div class="doc_subsection">
430<a name="llvm-gcc">llvm-gcc 4.2 Improvements</a>
431</div>
432
433<div class="doc_text">
434
435<p>LLVM fully supports the llvm-gcc 4.2 front-end, which marries the GCC
436front-ends and driver with the LLVM optimizer and code generator. It currently
437includes support for the C, C++, Objective-C, Ada, and Fortran front-ends.</p>
438
439<ul>
440<li>Something wonderful!</li>
441</ul>
442
443</div>
444
445
446<!--=========================================================================-->
447<div class="doc_subsection">
448<a name="coreimprovements">LLVM IR and Core Improvements</a>
449</div>
450
451<div class="doc_text">
452<p>LLVM IR has several new features that are used by our existing front-ends and
453can be useful if you are writing a front-end for LLVM:</p>
454
455<ul>
456<li>Something wonderful!</li>
457</ul>
458
459</div>
460
461<!--=========================================================================-->
462<div class="doc_subsection">
463<a name="optimizer">Optimizer Improvements</a>
464</div>
465
466<div class="doc_text">
467
468<p>In addition to a large array of bug fixes and minor performance tweaks, this
469release includes a few major enhancements and additions to the optimizers:</p>
470
471<ul>
472
473<li>Something wonderful!</li>
474
475</ul>
476
477</div>
478
479<!--=========================================================================-->
480<div class="doc_subsection">
481<a name="codegen">Target Independent Code Generator Improvements</a>
482</div>
483
484<div class="doc_text">
485
486<p>We have put a significant amount of work into the code generator
487infrastructure, which allows us to implement more aggressive algorithms and make
488it run faster:</p>
489
490<ul>
491
492<li>Something wonderful!</li>
493</ul>
494</div>
495
496<!--=========================================================================-->
497<div class="doc_subsection">
498<a name="x86">X86-32 and X86-64 Target Improvements</a>
499</div>
500
501<div class="doc_text">
502<p>New features of the X86 target include:
503</p>
504
505<ul>
506
507<li>Something wonderful!</li>
508</ul>
509
510</div>
511
512<!--=========================================================================-->
513<div class="doc_subsection">
514<a name="pic16">PIC16 Target Improvements</a>
515</div>
516
517<div class="doc_text">
518<p>New features of the PIC16 target include:
519</p>
520
521<ul>
522<li>Something wonderful!</li>
523</ul>
524
525<p>Things not yet supported:</p>
526
527<ul>
528<li>Floating point.</li>
529<li>Passing/returning aggregate types to and from functions.</li>
530<li>Variable arguments.</li>
531<li>Indirect function calls.</li>
532<li>Interrupts/programs.</li>
533<li>Debug info.</li>
534</ul>
535
536</div>
537
Bob Wilson755cbe02009-08-12 21:19:49 +0000[diff] [blame]538<!--=========================================================================-->
539<div class="doc_subsection">
540<a name="ARM">ARM Target Improvements</a>
541</div>
542
543<div class="doc_text">
544<p>New features of the ARM target include:
545</p>
546
547<ul>
548
549<li>Preliminary support for processors, such as the Cortex-A8 and Cortex-A9,
Sandeep Patel64e39612009-08-20 15:01:16 +0000[diff] [blame]550that implement version v7-A of the ARM architecture. The ARM backend now
551supports both the Thumb2 and Advanced SIMD (Neon) instruction sets. The
552AAPCS-VFP "hard float" calling conventions are also supported with the
553<tt>-float-abi=hard</tt> flag. These features are still somewhat experimental
554and subject to change. The Neon intrinsics, in particular, may change in future
555releases of LLVM.
Bob Wilson755cbe02009-08-12 21:19:49 +0000[diff] [blame]556</li>
557</ul>
558
559</div>
560
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]561
562<!--=========================================================================-->
563<div class="doc_subsection">
564<a name="llvmc">Improvements in LLVMC</a>
565</div>
566
567<div class="doc_text">
568<p>New features include:</p>
569
570<ul>
571<li>Something wonderful!</li>
572</ul>
573
574</div>
575
576
577<!--=========================================================================-->
578<div class="doc_subsection">
579<a name="changes">Major Changes and Removed Features</a>
580</div>
581
582<div class="doc_text">
583
584<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]585on LLVM 2.5, this section lists some "gotchas" that you may run into upgrading
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]586from the previous release.</p>
587
588<ul>
589
590<li>Something horrible!</li>
591
592</ul>
593
594
595<p>In addition, many APIs have changed in this release. Some of the major LLVM
596API changes are:</p>
597
598<ul>
Owen Andersone27be3a2009-07-02 16:48:38 +0000[diff] [blame]599<li>LLVM's global uniquing tables for <tt>Type</tt>s and <tt>Constant</tt>s have
600 been privatized into members of an <tt>LLVMContext</tt>. A number of APIs
601 now take an <tt>LLVMContext</tt> as a parameter. To smooth the transition
602 for clients that will only ever use a single context, the new
603 <tt>getGlobalContext()</tt> API can be used to access a default global
604 context which can be passed in any and all cases where a context is
605 required.
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]606<li>The <tt>getABITypeSize</tt> methods are now called <tt>getAllocSize</tt>.</li>
Dan Gohmanf667e7e2009-07-07 20:05:15 +0000[diff] [blame]607<li>The <tt>Add</tt>, <tt>Sub</tt>, and <tt>Mul</tt> operators are no longer
608 overloaded for floating-point types. Floating-point addition, subtraction,
609 and multiplication are now represented with new operators <tt>FAdd</tt>,
610 <tt>FSub</tt>, and <tt>FMul</tt>. In the <tt>IRBuilder</tt> API,
611 <tt>CreateAdd</tt>, <tt>CreateSub</tt>, <tt>CreateMul</tt>, and
612 <tt>CreateNeg</tt> should only be used for integer arithmetic now;
613 <tt>CreateFAdd</tt>, <tt>CreateFSub</tt>, <tt>CreateFMul</tt>, and
614 <tt>CreateFNeg</tt> should now be used for floating-point arithmetic.</li>
Daniel Dunbar02e73132009-07-12 20:41:27 +0000[diff] [blame]615<li>The DynamicLibrary class can no longer be constructed, its functionality has
616 moved to static member functions.</li>
Dan Gohmane4d54d72009-07-15 19:59:19 +0000[diff] [blame]617<li><tt>raw_fd_ostream</tt>'s constructor for opening a given filename now
618 takes an extra <tt>Force</tt> argument. If <tt>Force</tt> is set to
619 <tt>false</tt>, an error will be reported if a file with the given name
620 already exists. If <tt>Force</tt> is set to <tt>true</tt>, the file will
621 be silently truncated (which is the behavior before this flag was
622 added).</li>
Torok Edwinf6fa8ae2009-07-21 20:27:10 +0000[diff] [blame]623<li><tt>SCEVHandle</tt> no longer exists, because reference counting is no
624longer done for <tt>SCEV*</tt> objects, instead <tt>const SCEV*</tt> should be
625used.</li>
Daniel Dunbar8b5ee822009-07-25 05:26:53 +0000[diff] [blame]626
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]627<li>Many APIs, notably <tt>llvm::Value</tt>, now use the <tt>StringRef</tt>
628and <tt>Twine</tt> classes instead of passing <tt>const char*</tt>
629or <tt>std::string</tt>, as described in
630the <a href="ProgrammersManual.html#string_apis">Programmer's Manual</a>. Most
Daniel Dunbar8b5ee822009-07-25 05:26:53 +0000[diff] [blame]631clients 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:
632 <ul>
633 <li><tt>getNameStr()</tt> is still available, and matches the old
634 behavior. Replacing <tt>getName()</tt> calls with this is an safe option,
635 although more efficient alternatives are now possible.</li>
636
637 <li>If you were just relying on <tt>getName()</tt> being able to be sent to
638 a <tt>std::ostream</tt>, consider migrating
639 to <tt>llvm::raw_ostream</tt>.</li>
640
641 <li>If you were using <tt>getName().c_str()</tt> to get a <tt>const
642 char*</tt> pointer to the name, you can use <tt>getName().data()</tt>.
643 Note that this string (as before), may not be the entire name if the
644 name containts embedded null characters.</li>
645
646 <li>If you were using operator plus on the result of <tt>getName()</tt> and
647 treating the result as an <tt>std::string</tt>, you can either
648 uses <tt>Twine::str</tt> to get the result as an <tt>std::string</tt>, or
649 could move to a <tt>Twine</tt> based design.</li>
Daniel Dunbar03d76512009-07-25 23:55:21 +0000[diff] [blame]650
651 <li><tt>isName()</tt> should be replaced with comparison
Benjamin Kramere15192b2009-08-05 15:42:44 +0000[diff] [blame]652 against <tt>getName()</tt> (this is now efficient).
Daniel Dunbar8b5ee822009-07-25 05:26:53 +0000[diff] [blame]653 </ul>
654</li>
655
Daniel Dunbara5881e32009-07-26 02:12:58 +0000[diff] [blame]656<li>The registration interfaces for backend Targets has changed (what was
Daniel Dunbard6b06b12009-07-26 05:41:39 +0000[diff] [blame]657previously TargetMachineRegistry). For backend authors, see the <a href="WritingAnLLVMBackend.html#TargetRegistration">Writing An LLVM Backend</a> guide. For clients, the notable API changes are:
658 <ul>
659 <li><tt>TargetMachineRegistry</tt> has been renamed
660 to <tt>TargetRegistry</tt>.</li>
661
662 <li>Clients should move to using the <tt>TargetRegistry::lookupTarget()</tt>
663 function to find targets.</li>
664 </ul>
665</li>
Daniel Dunbar8b5ee822009-07-25 05:26:53 +0000[diff] [blame]666
Torok Edwinf6fa8ae2009-07-21 20:27:10 +0000[diff] [blame]667<li>llvm-dis now fails if output file exists, instead of dumping to stdout.
Daniel Dunbar6e0d1cb2009-07-25 04:41:11 +0000[diff] [blame]668FIXME: describe any other tool changes due to the raw_fd_ostream change. FIXME:
669This is not an API change, maybe there should be a tool changes section?</li>
Torok Edwinf6fa8ae2009-07-21 20:27:10 +0000[diff] [blame]670<li>temporarely due to Context API change passes should call doInitialization()
671method of the pass they inherit from, otherwise Context is NULL.
672FIXME: remove this entry when this is no longer needed.<li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]673</ul>
674
675</div>
676
677
678
679<!-- *********************************************************************** -->
680<div class="doc_section">
681 <a name="portability">Portability and Supported Platforms</a>
682</div>
683<!-- *********************************************************************** -->
684
685<div class="doc_text">
686
687<p>LLVM is known to work on the following platforms:</p>
688
689<ul>
690<li>Intel and AMD machines (IA32, X86-64, AMD64, EMT-64) running Red Hat
Chris Lattner3e7b5ca2009-07-21 22:47:03 +0000[diff] [blame]691Linux, Fedora Core, FreeBSD and AuroraUX (and probably other unix-like systems).</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]692<li>PowerPC and X86-based Mac OS X systems, running 10.3 and above in 32-bit
693and 64-bit modes.</li>
694<li>Intel and AMD machines running on Win32 using MinGW libraries (native).</li>
695<li>Intel and AMD machines running on Win32 with the Cygwin libraries (limited
696 support is available for native builds with Visual C++).</li>
697<li>Sun UltraSPARC workstations running Solaris 10.</li>
698<li>Alpha-based machines running Debian GNU/Linux.</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]699</ul>
700
701<p>The core LLVM infrastructure uses GNU autoconf to adapt itself
702to the machine and operating system on which it is built. However, minor
703porting may be required to get LLVM to work on new platforms. We welcome your
704portability patches and reports of successful builds or error messages.</p>
705
706</div>
707
708<!-- *********************************************************************** -->
709<div class="doc_section">
710 <a name="knownproblems">Known Problems</a>
711</div>
712<!-- *********************************************************************** -->
713
714<div class="doc_text">
715
716<p>This section contains significant known problems with the LLVM system,
717listed by component. If you run into a problem, please check the <a
718href="http://llvm.org/bugs/">LLVM bug database</a> and submit a bug if
719there isn't already one.</p>
720
Chris Lattner3e7b5ca2009-07-21 22:47:03 +0000[diff] [blame]721<ul>
Chris Lattneraedb59a2009-07-21 23:17:26 +0000[diff] [blame]722<li>LLVM will not correctly compile on Solaris and/or OpenSolaris
Chris Lattner3e7b5ca2009-07-21 22:47:03 +0000[diff] [blame]723using the stock GCC 3.x.x series 'out the box',
724See: <a href="#brokengcc">Broken versions of GCC and other tools</a>.
725However, A <a href="http://pkg.auroraux.org/GCC">Modern GCC Build</a>
726for x86/x64 has been made available from the third party AuroraUX Project
727that has been meticulously tested for bootstrapping LLVM & Clang.</li>
728</ul>
729
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]730</div>
731
732<!-- ======================================================================= -->
733<div class="doc_subsection">
734 <a name="experimental">Experimental features included with this release</a>
735</div>
736
737<div class="doc_text">
738
739<p>The following components of this LLVM release are either untested, known to
740be broken or unreliable, or are in early development. These components should
741not be relied on, and bugs should not be filed against them, but they may be
742useful to some people. In particular, if you would like to work on one of these
743components, please contact us on the <a
744href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVMdev list</a>.</p>
745
746<ul>
Dan Gohmand2cb3d22009-07-24 00:30:09 +0000[diff] [blame]747<li>The MSIL, Alpha, SPU, MIPS, and PIC16 backends are experimental.</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]748<li>The <tt>llc</tt> "<tt>-filetype=asm</tt>" (the default) is the only
749 supported value for this option.</li>
750</ul>
751
752</div>
753
754<!-- ======================================================================= -->
755<div class="doc_subsection">
756 <a name="x86-be">Known problems with the X86 back-end</a>
757</div>
758
759<div class="doc_text">
760
761<ul>
762 <li>The X86 backend does not yet support
763 all <a href="http://llvm.org/PR879">inline assembly that uses the X86
764 floating point stack</a>. It supports the 'f' and 't' constraints, but not
765 'u'.</li>
766 <li>The X86 backend generates inefficient floating point code when configured
767 to generate code for systems that don't have SSE2.</li>
768 <li>Win64 code generation wasn't widely tested. Everything should work, but we
769 expect small issues to happen. Also, llvm-gcc cannot build the mingw64
770 runtime currently due
771 to <a href="http://llvm.org/PR2255">several</a>
772 <a href="http://llvm.org/PR2257">bugs</a> and due to lack of support for
773 the
774 'u' inline assembly constraint and for X87 floating point inline assembly.</li>
775 <li>The X86-64 backend does not yet support the LLVM IR instruction
776 <tt>va_arg</tt>. Currently, the llvm-gcc and front-ends support variadic
777 argument constructs on X86-64 by lowering them manually.</li>
778</ul>
779
780</div>
781
782<!-- ======================================================================= -->
783<div class="doc_subsection">
784 <a name="ppc-be">Known problems with the PowerPC back-end</a>
785</div>
786
787<div class="doc_text">
788
789<ul>
790<li>The Linux PPC32/ABI support needs testing for the interpreter and static
791compilation, and lacks support for debug information.</li>
792</ul>
793
794</div>
795
796<!-- ======================================================================= -->
797<div class="doc_subsection">
798 <a name="arm-be">Known problems with the ARM back-end</a>
799</div>
800
801<div class="doc_text">
802
803<ul>
Bob Wilson755cbe02009-08-12 21:19:49 +0000[diff] [blame]804<li>Support for the Advanced SIMD (Neon) instruction set is still incomplete
805and not well tested. Some features may not work at all, and the code quality
806may be poor in some cases.</li>
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]807<li>Thumb mode works only on ARMv6 or higher processors. On sub-ARMv6
808processors, thumb programs can crash or produce wrong
809results (<a href="http://llvm.org/PR1388">PR1388</a>).</li>
810<li>Compilation for ARM Linux OABI (old ABI) is supported but not fully tested.
811</li>
812<li>There is a bug in QEMU-ARM (&lt;= 0.9.0) which causes it to incorrectly
813 execute
814programs compiled with LLVM. Please use more recent versions of QEMU.</li>
815</ul>
816
817</div>
818
819<!-- ======================================================================= -->
820<div class="doc_subsection">
821 <a name="sparc-be">Known problems with the SPARC back-end</a>
822</div>
823
824<div class="doc_text">
825
826<ul>
827<li>The SPARC backend only supports the 32-bit SPARC ABI (-m32); it does not
828 support the 64-bit SPARC ABI (-m64).</li>
829</ul>
830
831</div>
832
833<!-- ======================================================================= -->
834<div class="doc_subsection">
835 <a name="mips-be">Known problems with the MIPS back-end</a>
836</div>
837
838<div class="doc_text">
839
840<ul>
841<li>The O32 ABI is not fully supported.</li>
842<li>64-bit MIPS targets are not supported yet.</li>
843</ul>
844
845</div>
846
847<!-- ======================================================================= -->
848<div class="doc_subsection">
849 <a name="alpha-be">Known problems with the Alpha back-end</a>
850</div>
851
852<div class="doc_text">
853
854<ul>
855
856<li>On 21164s, some rare FP arithmetic sequences which may trap do not have the
857appropriate nops inserted to ensure restartability.</li>
858
859</ul>
860</div>
861
862<!-- ======================================================================= -->
863<div class="doc_subsection">
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]864 <a name="c-be">Known problems with the C back-end</a>
865</div>
866
867<div class="doc_text">
868
869<ul>
870<li><a href="http://llvm.org/PR802">The C backend has only basic support for
871 inline assembly code</a>.</li>
872<li><a href="http://llvm.org/PR1658">The C backend violates the ABI of common
873 C++ programs</a>, preventing intermixing between C++ compiled by the CBE and
874 C++ code compiled with <tt>llc</tt> or native compilers.</li>
875<li>The C backend does not support all exception handling constructs.</li>
876<li>The C backend does not support arbitrary precision integers.</li>
877</ul>
878
879</div>
880
881
882<!-- ======================================================================= -->
883<div class="doc_subsection">
884 <a name="c-fe">Known problems with the llvm-gcc C front-end</a>
885</div>
886
887<div class="doc_text">
888
889<p>llvm-gcc does not currently support <a href="http://llvm.org/PR869">Link-Time
890Optimization</a> on most platforms "out-of-the-box". Please inquire on the
891LLVMdev mailing list if you are interested.</p>
892
893<p>The only major language feature of GCC not supported by llvm-gcc is
894 the <tt>__builtin_apply</tt> family of builtins. However, some extensions
895 are only supported on some targets. For example, trampolines are only
896 supported on some targets (these are used when you take the address of a
897 nested function).</p>
898
899<p>If you run into GCC extensions which are not supported, please let us know.
900</p>
901
902</div>
903
904<!-- ======================================================================= -->
905<div class="doc_subsection">
906 <a name="c++-fe">Known problems with the llvm-gcc C++ front-end</a>
907</div>
908
909<div class="doc_text">
910
911<p>The C++ front-end is considered to be fully
912tested and works for a number of non-trivial programs, including LLVM
913itself, Qt, Mozilla, etc.</p>
914
915<ul>
916<li>Exception handling works well on the X86 and PowerPC targets. Currently
917 only Linux and Darwin targets are supported (both 32 and 64 bit).</li>
918</ul>
919
920</div>
921
922<!-- ======================================================================= -->
923<div class="doc_subsection">
924 <a name="fortran-fe">Known problems with the llvm-gcc Fortran front-end</a>
925</div>
926
927<div class="doc_text">
928<ul>
929<li>Fortran support generally works, but there are still several unresolved bugs
930 in Bugzilla. Please see the tools/gfortran component for details.</li>
931</ul>
932</div>
933
934<!-- ======================================================================= -->
935<div class="doc_subsection">
936 <a name="ada-fe">Known problems with the llvm-gcc Ada front-end</a>
937</div>
938
939<div class="doc_text">
940The llvm-gcc 4.2 Ada compiler works fairly well; however, this is not a mature
941technology, and problems should be expected.
942<ul>
943<li>The Ada front-end currently only builds on X86-32. This is mainly due
944to lack of trampoline support (pointers to nested functions) on other platforms.
945However, it <a href="http://llvm.org/PR2006">also fails to build on X86-64</a>
946which does support trampolines.</li>
947<li>The Ada front-end <a href="http://llvm.org/PR2007">fails to bootstrap</a>.
948This is due to lack of LLVM support for <tt>setjmp</tt>/<tt>longjmp</tt> style
949exception handling, which is used internally by the compiler.
950Workaround: configure with --disable-bootstrap.</li>
951<li>The c380004, <a href="http://llvm.org/PR2010">c393010</a>
952and <a href="http://llvm.org/PR2421">cxg2021</a> ACATS tests fail
953(c380004 also fails with gcc-4.2 mainline).
954If the compiler is built with checks disabled then <a href="http://llvm.org/PR2010">c393010</a>
955causes the compiler to go into an infinite loop, using up all system memory.</li>
956<li>Some GCC specific Ada tests continue to crash the compiler.</li>
957<li>The -E binder option (exception backtraces)
958<a href="http://llvm.org/PR1982">does not work</a> and will result in programs
959crashing if an exception is raised. Workaround: do not use -E.</li>
960<li>Only discrete types <a href="http://llvm.org/PR1981">are allowed to start
961or finish at a non-byte offset</a> in a record. Workaround: do not pack records
962or use representation clauses that result in a field of a non-discrete type
963starting or finishing in the middle of a byte.</li>
964<li>The <tt>lli</tt> interpreter <a href="http://llvm.org/PR2009">considers
965'main' as generated by the Ada binder to be invalid</a>.
966Workaround: hand edit the file to use pointers for <tt>argv</tt> and
967<tt>envp</tt> rather than integers.</li>
968<li>The <tt>-fstack-check</tt> option <a href="http://llvm.org/PR2008">is
969ignored</a>.</li>
970</ul>
971</div>
972
Erick Tryzelaar17167be2009-09-28 04:42:55 +0000[diff] [blame]973<!-- ======================================================================= -->
974<div class="doc_subsection">
975 <a name="ocaml-bindingse">Known problems with the O'Caml bindings</a>
976</div>
977
978<div class="doc_text">
979
980<p>The Llvm.Linkage module is broken, and has incorrect values. Only
981Llvm.Linkage.External, Llvm.Linkage.Available_externally, and
982Llvm.Linkage.Link_once will be correct. If you need any of the other linkage
983modes, you'll have to write an external C library in order to expose the
984functionality. This has been fixed in the trunk.</p>
985</div>
986
Duncan Sandse2e4a892009-06-24 08:38:48 +0000[diff] [blame]987<!-- *********************************************************************** -->
988<div class="doc_section">
989 <a name="additionalinfo">Additional Information</a>
990</div>
991<!-- *********************************************************************** -->
992
993<div class="doc_text">
994
995<p>A wide variety of additional information is available on the <a
996href="http://llvm.org">LLVM web page</a>, in particular in the <a
997href="http://llvm.org/docs/">documentation</a> section. The web page also
998contains versions of the API documentation which is up-to-date with the
999Subversion version of the source code.
1000You can access versions of these documents specific to this release by going
1001into the "<tt>llvm/doc/</tt>" directory in the LLVM tree.</p>
1002
1003<p>If you have any questions or comments about LLVM, please feel free to contact
1004us via the <a href="http://llvm.org/docs/#maillist"> mailing
1005lists</a>.</p>
1006
1007</div>
1008
1009<!-- *********************************************************************** -->
1010
1011<hr>
1012<address>
1013 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
1014 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
1015 <a href="http://validator.w3.org/check/referer"><img
1016 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
1017
1018 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
1019 Last modified: $Date$
1020</address>
1021
1022</body>
1023</html>