1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
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>
11 <div class="doc_title">LLVM 2.6 Release Notes</div>
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>
24 <div class="doc_author">
25 <p>Written by the <a href="http://llvm.org">LLVM Team</a></p>
28 <!-- *********************************************************************** -->
29 <div class="doc_section">
30 <a name="intro">Introduction</a>
32 <!-- *********************************************************************** -->
34 <div class="doc_text">
36 <p>This document contains the release notes for the LLVM Compiler
37 Infrastructure, release 2.6. Here we describe the status of LLVM, including
38 major improvements from the previous release and significant known problems.
39 All LLVM releases may be downloaded from the <a
40 href="http://llvm.org/releases/">LLVM releases web site</a>.</p>
42 <p>For more information about LLVM, including information about the latest
43 release, please check out the <a href="http://llvm.org/">main LLVM
44 web site</a>. If you have questions or comments, the <a
45 href="http://mail.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVM Developer's Mailing
46 List</a> is a good place to send them.</p>
48 <p>Note that if you are reading this file from a Subversion checkout or the
49 main LLVM web page, this document applies to the <i>next</i> release, not the
50 current one. To see the release notes for a specific release, please see the
51 <a href="http://llvm.org/releases/">releases page</a>.</p>
58 include/llvm/Analysis/LiveValues.h => Dan
59 lib/Transforms/IPO/MergeFunctions.cpp => consider for 2.8.
60 llvm/Analysis/PointerTracking.h => Edwin wants this, consider for 2.8.
64 <!-- Unfinished features in 2.6:
67 variable debug info for optimized code
68 postalloc scheduler: anti dependence breaking, hazard recognizer?
70 loop dependence analysis
71 ELF Writer? How stable?
72 <li>PostRA scheduler improvements, ARM adoption (David Goodwin).</li>
76 <!-- for announcement email:
80 klee web page at klee.llvm.org
81 Many new papers added to /pubs/
86 <!-- *********************************************************************** -->
87 <div class="doc_section">
88 <a name="subproj">Sub-project Status Update</a>
90 <!-- *********************************************************************** -->
92 <div class="doc_text">
94 The LLVM 2.6 distribution currently consists of code from the core LLVM
95 repository —which roughly includes the LLVM optimizers, code generators
96 and supporting tools — and the llvm-gcc repository. In addition to this
97 code, the LLVM Project includes other sub-projects that are in development. The
98 two which are the most actively developed are the <a href="#clang">Clang
99 Project</a> and the <a href="#vmkit">VMKit Project</a>.
105 <!--=========================================================================-->
106 <div class="doc_subsection">
107 <a name="clang">Clang: C/C++/Objective-C Frontend Toolkit</a>
110 <div class="doc_text">
112 <p><b>UPDATE!</b> The <a href="http://clang.llvm.org/">Clang project</a> is an effort to build
113 a set of new 'LLVM native' front-end technologies for the LLVM optimizer and
114 code generator. While Clang is not included in the LLVM 2.6 release, it is
115 continuing to make major strides forward in all areas. Its C and Objective-C
116 parsing and code generation support is now very solid. For example, it is
117 capable of successfully building many real-world applications for X86-32
119 including the <a href="http://wiki.freebsd.org/BuildingFreeBSDWithClang">FreeBSD
120 kernel</a> and <a href="http://gcc.gnu.org/gcc-4.2/">gcc 4.2</a>. C++ is also
121 making <a href="http://clang.llvm.org/cxx_status.html">incredible progress</a>,
122 and work on templates has recently started. If you are
123 interested in fast compiles and good diagnostics, we encourage you to try it out
124 by <a href="http://clang.llvm.org/get_started.html">building from mainline</a>
125 and reporting any issues you hit to the <a
126 href="http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev">Clang front-end mailing
129 <p>In the LLVM 2.6 time-frame, the Clang team has made many improvements:</p>
132 <li>Something wonderful!</li>
133 <li>AuroraUX / FreeBSD & OpenBSD Toolchain support.</li>
134 <li>Many many bugs are fixed and many features have been added.</li>
138 <!--=========================================================================-->
139 <div class="doc_subsection">
140 <a name="clangsa">Clang Static Analyzer</a>
143 <div class="doc_text">
145 <p><b>UPDATE!</b> Previously announced in the 2.4 and 2.5 LLVM releases, the Clang project also
146 includes an early stage static source code analysis tool for <a
147 href="http://clang.llvm.org/StaticAnalysis.html">automatically finding bugs</a>
148 in C and Objective-C programs. The tool performs a growing set of checks to find
149 bugs that occur on a specific path within a program.</p>
151 <p>In the LLVM 2.6 time-frame there have been many significant improvements to
154 <p>The set of checks performed by the static analyzer continues to expand, and
155 future plans for the tool include full source-level inter-procedural analysis
156 and deeper checks such as buffer overrun detection. There are many opportunities
157 to extend and enhance the static analyzer, and anyone interested in working on
158 this project is encouraged to get involved!</p>
162 <!--=========================================================================-->
163 <div class="doc_subsection">
164 <a name="vmkit">VMKit: JVM/CLI Virtual Machine Implementation</a>
167 <div class="doc_text">
170 The <a href="http://vmkit.llvm.org/">VMKit project</a> is an implementation of
171 a JVM and a CLI Virtual Machines (Microsoft .NET is an
172 implementation of the CLI) using the Just-In-Time compiler of LLVM.</p>
174 <p>Following LLVM 2.6, VMKit has its XYZ release that you can find on its
175 <a href="http://vmkit.llvm.org/releases/">webpage</a>. The release includes
176 bug fixes, cleanup and new features. The major changes are:</p>
180 <li>Something wonderful!</li>
185 <!--=========================================================================-->
186 <div class="doc_subsection">
187 <a name="mc">llvm-mc: Machine Code Toolkit</a>
190 <div class="doc_text">
197 MCInstPrinter did it make it in?
198 MCInst (X86 using it so far)
199 Rewrite of X86 GV selection logic: TargetOperand flags on ExternalSymbol, GV, etc operands.
200 Can parse and re-print out an darwin-x86 .s file.
201 TargetLoweringObjectFile, MCSectionKind
202 Verrrry early start of a macho writer.
208 <!-- *********************************************************************** -->
209 <div class="doc_section">
210 <a name="externalproj">External Projects Using LLVM 2.6</a>
212 <!-- *********************************************************************** -->
214 <!--=========================================================================-->
215 <div class="doc_subsection">
216 <a name="Rubinius">Rubinius</a>
219 <div class="doc_text">
220 <p><a href="http://github.com/evanphx/rubinius">Rubinius</a> is an environment
221 for running Ruby code which strives to write as much of the core class
222 implementation in Ruby as possible. Combined with a bytecode interpreting VM, it
223 uses LLVM to optimize and compile ruby code down to machine code. Techniques
224 such as type feedback, method inlining, and uncommon traps are all used to
225 remove dynamism from ruby execution and increase performance.</p>
228 <!--=========================================================================-->
229 <div class="doc_subsection">
230 <a name="macruby">MacRuby</a>
233 <div class="doc_text">
236 <a href="http://macruby.org">MacRuby</a> is an implementation of Ruby on top of
237 core Mac OS X technologies, such as the Objective-C common runtime and garbage
238 collector, and the CoreFoundation framework. It is principally developed by
239 Apple and aims at enabling the creation of full-fledged Mac OS X applications.
243 MacRuby uses LLVM for optimization passes, JIT and AOT compilation of Ruby
244 expressions. It also uses zero-cost DWARF exceptions to implement Ruby exception
250 <!--=========================================================================-->
251 <div class="doc_subsection">
252 <a name="pure">Pure</a>
255 <div class="doc_text">
258 <a href="http://pure-lang.googlecode.com/">Pure</a>
259 is an algebraic/functional programming language based on term rewriting.
260 Programs are collections of equations which are used to evaluate expressions in
261 a symbolic fashion. Pure offers dynamic typing, eager and lazy evaluation,
262 lexical closures, a hygienic macro system (also based on term rewriting),
263 built-in list and matrix support (including list and matrix comprehensions) and
264 an easy-to-use C interface. The interpreter uses LLVM as a backend to
265 JIT-compile Pure programs to fast native code.</p>
267 <p>In addition to the usual algebraic data structures, Pure also has
268 MATLAB-style matrices in order to support numeric computations and signal
269 processing in an efficient way. Pure is mainly aimed at mathematical
270 applications right now, but it has been designed as a general purpose language.
271 The dynamic interpreter environment and the C interface make it possible to use
272 it as a kind of functional scripting language for many application areas.
277 <!--=========================================================================-->
278 <div class="doc_subsection">
279 <a name="ldc">LLVM D Compiler</a>
282 <div class="doc_text">
285 <a href="http://www.dsource.org/projects/ldc">LDC</a> is an implementation of
286 the D Programming Language using the LLVM optimizer and code generator.
287 The LDC project works great with the LLVM 2.6 release. General improvements in
289 cycle have included new inline asm constraint handling, better debug info
290 support, general bugfixes, and better x86-64 support. This has allowed
291 some major improvements in LDC, getting us much closer to being as
292 fully featured as the original DMD compiler from DigitalMars.
296 <!--=========================================================================-->
297 <div class="doc_subsection">
298 <a name="RoadsendPHP">Roadsend PHP</a>
301 <div class="doc_text">
304 <a href="http://code.roadsend.com/rphp">Roadsend PHP</a> (rphp) is an open
305 source implementation of the PHP programming
306 language that uses LLVM for its optimizer, JIT, and static compiler. This is a
307 reimplementation of an earlier project that is now based on LLVM.</p>
310 <!--=========================================================================-->
311 <div class="doc_subsection">
312 <a name="Unladen Swallow">Unladen Swallow</a>
315 <div class="doc_text">
318 <a href="http://code.google.com/p/unladen-swallow/">Unladen Swallow</a> is a
319 branch of <a href="http://python.org/">Python</a> intended to be fully
320 compatible and significantly faster. It uses LLVM's optimization passes and JIT
326 <!-- *********************************************************************** -->
327 <div class="doc_section">
328 <a name="whatsnew">What's New in LLVM 2.6?</a>
330 <!-- *********************************************************************** -->
332 <div class="doc_text">
334 <p>This release includes a huge number of bug fixes, performance tweaks, and
335 minor improvements. Some of the major improvements and new features are listed
341 <!--=========================================================================-->
342 <div class="doc_subsection">
343 <a name="majorfeatures">Major New Features</a>
346 <div class="doc_text">
348 <p>LLVM 2.6 includes several major new capabilities:</p>
351 <li>Support for debug line numbers when optimization enabled</li>
352 <li>gold lto plugin</li>
353 <li>New MSP430 and SystemZ backends.</li>
354 <li>New BlackFin backend.</li>
355 <li>LLVMContext, llvm_start_multithreaded: ProgrammersManual.html#threading</li>
356 <li>klee web page at klee.llvm.org</li>
358 <li>New compiler-rt project.</li>
359 <li>LLVM 2.6 includes a brand new experimental LLVM bindings to the Ada2005 programming language.</li>
364 <!--=========================================================================-->
365 <div class="doc_subsection">
366 <a name="coreimprovements">LLVM IR and Core Improvements</a>
369 <div class="doc_text">
370 <p>LLVM IR has several new features that are used by our existing front-ends and
371 can be useful if you are writing a front-end for LLVM:</p>
374 <li>Getelementpr instruction now allows any integer type for array/pointer indexes.</li>
375 <li>Inbounds for GEP</li>
376 <li>NSW/NUW/exact div</li>
377 <li>LSR promotes int induction variables to 64-bit on 64-bit targets, major perf boost for numerical code.</li>
378 <li>LSR now analyzes pointer expressions (e.g. getelementptrs), not just integers.</li>
379 <li>new linkage types linkonce_odr, weak_odr, linker_private, and available_externally.</li>
380 <li>New fadd, fsub, fmul instructions and classes. </li>
381 <li>Target intrinsics can now return multiple results.</li>
386 <!--=========================================================================-->
387 <div class="doc_subsection">
388 <a name="optimizer">Optimizer Improvements</a>
391 <div class="doc_text">
393 <p>In addition to a large array of bug fixes and minor performance tweaks, this
394 release includes a few major enhancements and additions to the optimizers:</p>
398 <li>SRoA improvements for vector unions, memset, arbitrary weird bitfield accesses etc. It now produces "strange" sized integers.</li>
399 <li>Inliner reuse stack space when inlining arrays?</li>
400 <li>Enabled GVN Load PRE.</li>
401 <li>New Static Single Information (SSI) construction pass (not used by anything yet, experimental).</li>
408 <!--=========================================================================-->
409 <div class="doc_subsection">
410 <a name="codegen">Target Independent Code Generator Improvements</a>
413 <div class="doc_text">
415 <p>We have put a significant amount of work into the code generator
416 infrastructure, which allows us to implement more aggressive algorithms and make
421 <li> -asm-verbose now prints location info (with -g) and loop nest info.</li>
422 <li>Tblgen now supports multiclass inheritance and a number of new string and
423 list operations like !(subst), !(foreach), !car, !cdr, !null, !if, !cast.
424 These make the .td files more expressive and allow more aggressive factoring
425 of duplication across instruction patterns.</li>
426 <li>New MachineVerifier pass.</li>
427 <li>Machine LICM, hoists things like constant pool loads, loads from readonly stubs, vector constant synthesization code, etc.</li>
428 <li>Machine Sinking</li>
429 <li>target-specific intrinsics (r63765)</li>
430 <li>Regalloc improvements for commuting, various spiller peephole optimizations, cross-class coalescing.</li>
431 <li><tt>llc -enable-value-prop</tt>, propagation of value info (sign/zero ext info) from one MBB to another</li>
432 <li>Regalloc hints for allocation stuff: Evan r73381/r73671. Finished/enabled?</li>
433 <li>Stack slot coloring for register spills (denser stack frames)</li>
434 <li>SelectionDAGS: New BuildVectorSDNode (r65296), and ISD::VECTOR_SHUFFLE (r69952 / PR2957)</li>
435 <li>Experimental support for shrink wrapping support in PEI.</li>
439 <!--=========================================================================-->
440 <div class="doc_subsection">
441 <a name="x86">X86-32 and X86-64 Target Improvements</a>
444 <div class="doc_text">
445 <p>New features of the X86 target include:
450 <li>Preliminary support for addrspace 256 -> GS, 257 -> FS, known problems: CodeGenerator.html#x86_memory</li>
451 <li>Support for softfloat modes, typically used by OS kernels.</li>
453 <li>X86-64: better modeling of implicit zero extensions, eliminates a lot of redundant zexts</li>
454 <li>X86-64 TLS support for local exec and initial exec.</li>
455 <li>Better modeling of H registerts as subregs.</li>
456 <li>Vector icmp/fcmp now work with SSE codegen.</li>
457 <li>SSE 4.2 support.</li>
458 <li>all global variable reference logic is now in ClassifyGlobalReference.</li>
464 <!--=========================================================================-->
465 <div class="doc_subsection">
466 <a name="pic16">PIC16 Target Improvements</a>
469 <div class="doc_text">
470 <p>New features of the PIC16 target include:
474 <li>Support for floating-point, indirect function calls, and
475 passing/returning aggregate types to functions.
476 <li>The code generator is able to generate debug info into output COFF files.
477 <li>Support for placing an object into a specific section or at a specific
478 address in memory.</li>
481 <p>Things not yet supported:</p>
484 <li>Variable arguments.</li>
485 <li>Interrupts/programs.</li>
490 <!--=========================================================================-->
491 <div class="doc_subsection">
492 <a name="ARM">ARM Target Improvements</a>
495 <div class="doc_text">
496 <p>New features of the ARM target include:
501 <li>Preliminary support for processors, such as the Cortex-A8 and Cortex-A9,
502 that implement version v7-A of the ARM architecture. The ARM backend now
503 supports both the Thumb2 and Advanced SIMD (Neon) instruction sets. The
504 AAPCS-VFP "hard float" calling conventions are also supported with the
505 <tt>-float-abi=hard</tt> flag. These features are still somewhat experimental
506 and subject to change. The Neon intrinsics, in particular, may change in future
510 ARM AAPCS-VFP hard float ABI is supported.
511 ARM calling convention code is now tblgen generated instead of manual.
512 ARM: NEON support. neonfp for doing single precision fp with neon instead of VFP.
518 <!--=========================================================================-->
519 <div class="doc_subsection">
520 <a name="OtherTarget">Other Target Specific Improvements</a>
523 <div class="doc_text">
524 <p>New features of other targets include:
528 <li>Add support for the PowerPC 64-bit SVR4 ABI.</li>
529 <li>Mips now supports O32 Calling Convention.</li>
534 <!--=========================================================================-->
535 <div class="doc_subsection">
536 <a name="newapis">New Useful APIs</a>
539 <div class="doc_text">
542 <li>New EngineBuilder class for creating JITs: r76276</li>
543 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.
544 StringRef class, Twine class.
545 New WeakVH and AssertingVH and CallbackVH classes.
546 New llvm/ADT/Triple class.
547 llvm_report_error() error handling API (llvm/Support/ErrorHandling.h)
548 New llvm/System/Atomic.h, llvm/System/RWMutex.h for portable atomic ops, rw locks.
549 New SourceMgr, SMLoc classes for simple parsers with caret diagnostics and #include support, (used by
550 tablegen, llvm-mc, the .ll parser, FileCheck, etc)
557 <!--=========================================================================-->
558 <div class="doc_subsection">
559 <a name="otherimprovements">Other Improvements and New Features</a>
562 <div class="doc_text">
563 <p>Other miscellaneous features include:</p>
566 <li>interpreter + libffi</li>
567 <li> JIT now supports generating more than 16M of code.</li>
569 href="http://llvm.org/doxygen/classllvm_1_1ExecutionEngine.html#fcd2b4b92ca38812ca31640b0da14927">register</a>
570 a <a href="http://llvm.org/doxygen/classllvm_1_1JITEventListener.html">JITEventListener</a>
571 to receive callbacks when the JIT emits or frees machine code. The
572 OProfile support uses this mechanism.</li>
573 JIT support for oprofile (r75279), configure with --with-oprofile. Now we get line # and function info for JIT'd functions.
575 <li>Profile info improvements by Andreas Neustifter.</li>
576 <li>Many extensions to the C APIs.</li>
580 * Dynamic plugins now work on Windows.
581 * New option property: init. Makes possible to provide default values for
582 options defined in plugins (interface to cl::init).
583 * New example: Skeleton, shows how to create a standalone LLVMC-based driver.
584 * New example: mcc16, a driver for the PIC16 toolchain.</li>
591 <!--=========================================================================-->
592 <div class="doc_subsection">
593 <a name="changes">Major Changes and Removed Features</a>
596 <div class="doc_text">
598 <p>If you're already an LLVM user or developer with out-of-tree changes based
599 on LLVM 2.5, this section lists some "gotchas" that you may run into upgrading
600 from the previous release.</p>
604 <li>The Itanium (IA64) backend has been removed. It was not supported and
606 <li>The BigBlock register allocator has been removed, it also bitrotted.</li>
607 <li>The C Backend (-march=c) is no longer considered part of the LLVM release
608 criteria. We still want it to work, but no one is maintaining it and it lacks
609 support for arbitrary precision integers and other important IR features.</li>
612 LLVM build now builds all libraries as .a files instead of some
613 libraries as relinked .o files. This requires some APIs like
614 InitializeAllTargets.h. TargetRegistry!
619 <p>In addition, many APIs have changed in this release. Some of the major LLVM
624 no use of hash_set/hash_map, no more llvm::OStream
625 Use raw_ostream for everything, killed off llvm/Streams.h and DOUT
628 <li>LLVM's global uniquing tables for <tt>Type</tt>s and <tt>Constant</tt>s have
629 been privatized into members of an <tt>LLVMContext</tt>. A number of APIs
630 now take an <tt>LLVMContext</tt> as a parameter. To smooth the transition
631 for clients that will only ever use a single context, the new
632 <tt>getGlobalContext()</tt> API can be used to access a default global
633 context which can be passed in any and all cases where a context is
635 <li>The <tt>getABITypeSize</tt> methods are now called <tt>getAllocSize</tt>.</li>
636 <li>The <tt>Add</tt>, <tt>Sub</tt>, and <tt>Mul</tt> operators are no longer
637 overloaded for floating-point types. Floating-point addition, subtraction,
638 and multiplication are now represented with new operators <tt>FAdd</tt>,
639 <tt>FSub</tt>, and <tt>FMul</tt>. In the <tt>IRBuilder</tt> API,
640 <tt>CreateAdd</tt>, <tt>CreateSub</tt>, <tt>CreateMul</tt>, and
641 <tt>CreateNeg</tt> should only be used for integer arithmetic now;
642 <tt>CreateFAdd</tt>, <tt>CreateFSub</tt>, <tt>CreateFMul</tt>, and
643 <tt>CreateFNeg</tt> should now be used for floating-point arithmetic.</li>
644 <li>The DynamicLibrary class can no longer be constructed, its functionality has
645 moved to static member functions.</li>
646 <li><tt>raw_fd_ostream</tt>'s constructor for opening a given filename now
647 takes an extra <tt>Force</tt> argument. If <tt>Force</tt> is set to
648 <tt>false</tt>, an error will be reported if a file with the given name
649 already exists. If <tt>Force</tt> is set to <tt>true</tt>, the file will
650 be silently truncated (which is the behavior before this flag was
652 <li><tt>SCEVHandle</tt> no longer exists, because reference counting is no
653 longer done for <tt>SCEV*</tt> objects, instead <tt>const SCEV*</tt> should be
656 <li>Many APIs, notably <tt>llvm::Value</tt>, now use the <tt>StringRef</tt>
657 and <tt>Twine</tt> classes instead of passing <tt>const char*</tt>
658 or <tt>std::string</tt>, as described in
659 the <a href="ProgrammersManual.html#string_apis">Programmer's Manual</a>. Most
660 clients 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:
662 <li><tt>getNameStr()</tt> is still available, and matches the old
663 behavior. Replacing <tt>getName()</tt> calls with this is an safe option,
664 although more efficient alternatives are now possible.</li>
666 <li>If you were just relying on <tt>getName()</tt> being able to be sent to
667 a <tt>std::ostream</tt>, consider migrating
668 to <tt>llvm::raw_ostream</tt>.</li>
670 <li>If you were using <tt>getName().c_str()</tt> to get a <tt>const
671 char*</tt> pointer to the name, you can use <tt>getName().data()</tt>.
672 Note that this string (as before), may not be the entire name if the
673 name containts embedded null characters.</li>
675 <li>If you were using operator plus on the result of <tt>getName()</tt> and
676 treating the result as an <tt>std::string</tt>, you can either
677 uses <tt>Twine::str</tt> to get the result as an <tt>std::string</tt>, or
678 could move to a <tt>Twine</tt> based design.</li>
680 <li><tt>isName()</tt> should be replaced with comparison
681 against <tt>getName()</tt> (this is now efficient).
685 <li>The registration interfaces for backend Targets has changed (what was
686 previously TargetMachineRegistry). For backend authors, see the <a href="WritingAnLLVMBackend.html#TargetRegistration">Writing An LLVM Backend</a> guide. For clients, the notable API changes are:
688 <li><tt>TargetMachineRegistry</tt> has been renamed
689 to <tt>TargetRegistry</tt>.</li>
691 <li>Clients should move to using the <tt>TargetRegistry::lookupTarget()</tt>
692 function to find targets.</li>
696 <li>llvm-dis now fails if output file exists, instead of dumping to stdout.
697 FIXME: describe any other tool changes due to the raw_fd_ostream change. FIXME:
698 This is not an API change, maybe there should be a tool changes section?</li>
699 <li>temporarely due to Context API change passes should call doInitialization()
700 method of the pass they inherit from, otherwise Context is NULL.
701 FIXME: remove this entry when this is no longer needed.<li>
708 <!-- *********************************************************************** -->
709 <div class="doc_section">
710 <a name="portability">Portability and Supported Platforms</a>
712 <!-- *********************************************************************** -->
714 <div class="doc_text">
716 <p>LLVM is known to work on the following platforms:</p>
719 <li>Intel and AMD machines (IA32, X86-64, AMD64, EMT-64) running Red Hat
720 Linux, Fedora Core, FreeBSD and AuroraUX (and probably other unix-like systems).</li>
721 <li>PowerPC and X86-based Mac OS X systems, running 10.3 and above in 32-bit
722 and 64-bit modes.</li>
723 <li>Intel and AMD machines running on Win32 using MinGW libraries (native).</li>
724 <li>Intel and AMD machines running on Win32 with the Cygwin libraries (limited
725 support is available for native builds with Visual C++).</li>
726 <li>Sun UltraSPARC workstations running Solaris 10.</li>
727 <li>Alpha-based machines running Debian GNU/Linux.</li>
730 <p>The core LLVM infrastructure uses GNU autoconf to adapt itself
731 to the machine and operating system on which it is built. However, minor
732 porting may be required to get LLVM to work on new platforms. We welcome your
733 portability patches and reports of successful builds or error messages.</p>
737 <!-- *********************************************************************** -->
738 <div class="doc_section">
739 <a name="knownproblems">Known Problems</a>
741 <!-- *********************************************************************** -->
743 <div class="doc_text">
745 <p>This section contains significant known problems with the LLVM system,
746 listed by component. If you run into a problem, please check the <a
747 href="http://llvm.org/bugs/">LLVM bug database</a> and submit a bug if
748 there isn't already one.</p>
751 <li>LLVM will not correctly compile on Solaris and/or OpenSolaris
752 using the stock GCC 3.x.x series 'out the box',
753 See: <a href="#brokengcc">Broken versions of GCC and other tools</a>.
754 However, A <a href="http://pkg.auroraux.org/GCC">Modern GCC Build</a>
755 for x86/x64 has been made available from the third party AuroraUX Project
756 that has been meticulously tested for bootstrapping LLVM & Clang.</li>
761 <!-- ======================================================================= -->
762 <div class="doc_subsection">
763 <a name="experimental">Experimental features included with this release</a>
766 <div class="doc_text">
768 <p>The following components of this LLVM release are either untested, known to
769 be broken or unreliable, or are in early development. These components should
770 not be relied on, and bugs should not be filed against them, but they may be
771 useful to some people. In particular, if you would like to work on one of these
772 components, please contact us on the <a
773 href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVMdev list</a>.</p>
776 <li>The MSIL, Alpha, SPU, MIPS, and PIC16 backends are experimental.</li>
777 <li>The <tt>llc</tt> "<tt>-filetype=asm</tt>" (the default) is the only
778 supported value for this option.</li>
783 <!-- ======================================================================= -->
784 <div class="doc_subsection">
785 <a name="x86-be">Known problems with the X86 back-end</a>
788 <div class="doc_text">
791 <li>The X86 backend does not yet support
792 all <a href="http://llvm.org/PR879">inline assembly that uses the X86
793 floating point stack</a>. It supports the 'f' and 't' constraints, but not
795 <li>The X86 backend generates inefficient floating point code when configured
796 to generate code for systems that don't have SSE2.</li>
797 <li>Win64 code generation wasn't widely tested. Everything should work, but we
798 expect small issues to happen. Also, llvm-gcc cannot build the mingw64
799 runtime currently due
800 to <a href="http://llvm.org/PR2255">several</a>
801 <a href="http://llvm.org/PR2257">bugs</a> and due to lack of support for
803 'u' inline assembly constraint and for X87 floating point inline assembly.</li>
804 <li>The X86-64 backend does not yet support the LLVM IR instruction
805 <tt>va_arg</tt>. Currently, the llvm-gcc and front-ends support variadic
806 argument constructs on X86-64 by lowering them manually.</li>
811 <!-- ======================================================================= -->
812 <div class="doc_subsection">
813 <a name="ppc-be">Known problems with the PowerPC back-end</a>
816 <div class="doc_text">
819 <li>The Linux PPC32/ABI support needs testing for the interpreter and static
820 compilation, and lacks support for debug information.</li>
825 <!-- ======================================================================= -->
826 <div class="doc_subsection">
827 <a name="arm-be">Known problems with the ARM back-end</a>
830 <div class="doc_text">
833 <li>Support for the Advanced SIMD (Neon) instruction set is still incomplete
834 and not well tested. Some features may not work at all, and the code quality
835 may be poor in some cases.</li>
836 <li>Thumb mode works only on ARMv6 or higher processors. On sub-ARMv6
837 processors, thumb programs can crash or produce wrong
838 results (<a href="http://llvm.org/PR1388">PR1388</a>).</li>
839 <li>Compilation for ARM Linux OABI (old ABI) is supported but not fully tested.
845 <!-- ======================================================================= -->
846 <div class="doc_subsection">
847 <a name="sparc-be">Known problems with the SPARC back-end</a>
850 <div class="doc_text">
853 <li>The SPARC backend only supports the 32-bit SPARC ABI (-m32); it does not
854 support the 64-bit SPARC ABI (-m64).</li>
859 <!-- ======================================================================= -->
860 <div class="doc_subsection">
861 <a name="mips-be">Known problems with the MIPS back-end</a>
864 <div class="doc_text">
867 <li>The O32 ABI is not fully supported.</li>
868 <li>64-bit MIPS targets are not supported yet.</li>
873 <!-- ======================================================================= -->
874 <div class="doc_subsection">
875 <a name="alpha-be">Known problems with the Alpha back-end</a>
878 <div class="doc_text">
882 <li>On 21164s, some rare FP arithmetic sequences which may trap do not have the
883 appropriate nops inserted to ensure restartability.</li>
888 <!-- ======================================================================= -->
889 <div class="doc_subsection">
890 <a name="c-be">Known problems with the C back-end</a>
893 <div class="doc_text">
896 <li><a href="http://llvm.org/PR802">The C backend has only basic support for
897 inline assembly code</a>.</li>
898 <li><a href="http://llvm.org/PR1658">The C backend violates the ABI of common
899 C++ programs</a>, preventing intermixing between C++ compiled by the CBE and
900 C++ code compiled with <tt>llc</tt> or native compilers.</li>
901 <li>The C backend does not support all exception handling constructs.</li>
902 <li>The C backend does not support arbitrary precision integers.</li>
908 <!-- ======================================================================= -->
909 <div class="doc_subsection">
910 <a name="c-fe">Known problems with the llvm-gcc C front-end</a>
913 <div class="doc_text">
915 <p>llvm-gcc does not currently support <a href="http://llvm.org/PR869">Link-Time
916 Optimization</a> on most platforms "out-of-the-box". Please inquire on the
917 LLVMdev mailing list if you are interested.</p>
919 <p>The only major language feature of GCC not supported by llvm-gcc is
920 the <tt>__builtin_apply</tt> family of builtins. However, some extensions
921 are only supported on some targets. For example, trampolines are only
922 supported on some targets (these are used when you take the address of a
923 nested function).</p>
925 <p>If you run into GCC extensions which are not supported, please let us know.
930 <!-- ======================================================================= -->
931 <div class="doc_subsection">
932 <a name="c++-fe">Known problems with the llvm-gcc C++ front-end</a>
935 <div class="doc_text">
937 <p>The C++ front-end is considered to be fully
938 tested and works for a number of non-trivial programs, including LLVM
939 itself, Qt, Mozilla, etc.</p>
942 <li>Exception handling works well on the X86 and PowerPC targets. Currently
943 only Linux and Darwin targets are supported (both 32 and 64 bit).</li>
948 <!-- ======================================================================= -->
949 <div class="doc_subsection">
950 <a name="fortran-fe">Known problems with the llvm-gcc Fortran front-end</a>
953 <div class="doc_text">
955 <li>Fortran support generally works, but there are still several unresolved bugs
956 in Bugzilla. Please see the tools/gfortran component for details.</li>
960 <!-- ======================================================================= -->
961 <div class="doc_subsection">
962 <a name="ada-fe">Known problems with the llvm-gcc Ada front-end</a>
965 <div class="doc_text">
966 The llvm-gcc 4.2 Ada compiler works fairly well; however, this is not a mature
967 technology, and problems should be expected.
969 <li>The Ada front-end currently only builds on X86-32. This is mainly due
970 to lack of trampoline support (pointers to nested functions) on other platforms.
971 However, it <a href="http://llvm.org/PR2006">also fails to build on X86-64</a>
972 which does support trampolines.</li>
973 <li>The Ada front-end <a href="http://llvm.org/PR2007">fails to bootstrap</a>.
974 This is due to lack of LLVM support for <tt>setjmp</tt>/<tt>longjmp</tt> style
975 exception handling, which is used internally by the compiler.
976 Workaround: configure with --disable-bootstrap.</li>
977 <li>The c380004, <a href="http://llvm.org/PR2010">c393010</a>
978 and <a href="http://llvm.org/PR2421">cxg2021</a> ACATS tests fail
979 (c380004 also fails with gcc-4.2 mainline).
980 If the compiler is built with checks disabled then <a href="http://llvm.org/PR2010">c393010</a>
981 causes the compiler to go into an infinite loop, using up all system memory.</li>
982 <li>Some GCC specific Ada tests continue to crash the compiler.</li>
983 <li>The -E binder option (exception backtraces)
984 <a href="http://llvm.org/PR1982">does not work</a> and will result in programs
985 crashing if an exception is raised. Workaround: do not use -E.</li>
986 <li>Only discrete types <a href="http://llvm.org/PR1981">are allowed to start
987 or finish at a non-byte offset</a> in a record. Workaround: do not pack records
988 or use representation clauses that result in a field of a non-discrete type
989 starting or finishing in the middle of a byte.</li>
990 <li>The <tt>lli</tt> interpreter <a href="http://llvm.org/PR2009">considers
991 'main' as generated by the Ada binder to be invalid</a>.
992 Workaround: hand edit the file to use pointers for <tt>argv</tt> and
993 <tt>envp</tt> rather than integers.</li>
994 <li>The <tt>-fstack-check</tt> option <a href="http://llvm.org/PR2008">is
999 <!-- ======================================================================= -->
1000 <div class="doc_subsection">
1001 <a name="ocaml-bindingse">Known problems with the O'Caml bindings</a>
1004 <div class="doc_text">
1006 <p>The Llvm.Linkage module is broken, and has incorrect values. Only
1007 Llvm.Linkage.External, Llvm.Linkage.Available_externally, and
1008 Llvm.Linkage.Link_once will be correct. If you need any of the other linkage
1009 modes, you'll have to write an external C library in order to expose the
1010 functionality. This has been fixed in the trunk.</p>
1013 <!-- *********************************************************************** -->
1014 <div class="doc_section">
1015 <a name="additionalinfo">Additional Information</a>
1017 <!-- *********************************************************************** -->
1019 <div class="doc_text">
1021 <p>A wide variety of additional information is available on the <a
1022 href="http://llvm.org">LLVM web page</a>, in particular in the <a
1023 href="http://llvm.org/docs/">documentation</a> section. The web page also
1024 contains versions of the API documentation which is up-to-date with the
1025 Subversion version of the source code.
1026 You can access versions of these documents specific to this release by going
1027 into the "<tt>llvm/doc/</tt>" directory in the LLVM tree.</p>
1029 <p>If you have any questions or comments about LLVM, please feel free to contact
1030 us via the <a href="http://llvm.org/docs/#maillist"> mailing
1035 <!-- *********************************************************************** -->
1039 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
1040 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
1041 <a href="http://validator.w3.org/check/referer"><img
1042 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
1044 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
1045 Last modified: $Date$