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>
59 MCInstPrinter did it make it in?
60 MCInst (X86 using it so far)
61 Rewrite of X86 GV selection logic: TargetOperand flags on ExternalSymbol, GV, etc operands.
62 Can parse and re-print out an darwin-x86 .s file.
63 TargetLoweringObjectFile, MCSectionKind
64 Verrrry early start of a macho writer.
66 ELF Writer? How stable?
70 Shrink wrapping support in PEI, what is the state of it?
74 include/llvm/Analysis/LiveValues.h => dan
75 lib/Analysis/LoopVR.cpp ==> dan/nick/owen
76 lib/Transforms/IPO/MergeFunctions.cpp
77 llvm/Analysis/PointerTracking.h
80 LLVM build now builds all libraries as .a files instead of some
81 libraries as relinked .o files. This requires some APIs like
82 InitializeAllTargets.h. TargetRegistry!
86 <!-- Unfinished features in 2.6:
89 variable debug info for optimized code
90 postalloc scheduler: anti dependence breaking, hazard recognizer?
92 loop dependence analysis
95 <!-- for announcement email:
99 klee web page at klee.llvm.org
100 Many new papers added to /pubs/
105 <!-- *********************************************************************** -->
106 <div class="doc_section">
107 <a name="subproj">Sub-project Status Update</a>
109 <!-- *********************************************************************** -->
111 <div class="doc_text">
113 The LLVM 2.6 distribution currently consists of code from the core LLVM
114 repository —which roughly includes the LLVM optimizers, code generators
115 and supporting tools — and the llvm-gcc repository. In addition to this
116 code, the LLVM Project includes other sub-projects that are in development. The
117 two which are the most actively developed are the <a href="#clang">Clang
118 Project</a> and the <a href="#vmkit">VMKit Project</a>.
124 <!--=========================================================================-->
125 <div class="doc_subsection">
126 <a name="clang">Clang: C/C++/Objective-C Frontend Toolkit</a>
129 <div class="doc_text">
131 <p>The <a href="http://clang.llvm.org/">Clang project</a> is an effort to build
132 a set of new 'LLVM native' front-end technologies for the LLVM optimizer and
133 code generator. While Clang is not included in the LLVM 2.6 release, it is
134 continuing to make major strides forward in all areas. Its C and Objective-C
135 parsing and code generation support is now very solid. For example, it is
136 capable of successfully building many real-world applications for X86-32
138 including the <a href="http://wiki.freebsd.org/BuildingFreeBSDWithClang">FreeBSD
139 kernel</a> and <a href="http://gcc.gnu.org/gcc-4.2/">gcc 4.2</a>. C++ is also
140 making <a href="http://clang.llvm.org/cxx_status.html">incredible progress</a>,
141 and work on templates has recently started. If you are
142 interested in fast compiles and good diagnostics, we encourage you to try it out
143 by <a href="http://clang.llvm.org/get_started.html">building from mainline</a>
144 and reporting any issues you hit to the <a
145 href="http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev">Clang front-end mailing
148 <p>In the LLVM 2.6 time-frame, the Clang team has made many improvements:</p>
151 <li>Something wonderful!</li>
152 <li>AuroraUX / FreeBSD & OpenBSD Toolchain support.</li>
153 <li>Many many bugs are fixed and many features have been added.</li>
157 <!--=========================================================================-->
158 <div class="doc_subsection">
159 <a name="clangsa">Clang Static Analyzer</a>
162 <div class="doc_text">
164 <p>Previously announced in the 2.4 LLVM release, the Clang project also
165 includes an early stage static source code analysis tool for <a
166 href="http://clang.llvm.org/StaticAnalysis.html">automatically finding bugs</a>
167 in C and Objective-C programs. The tool performs a growing set of checks to find
168 bugs that occur on a specific path within a program.</p>
170 <p>In the LLVM 2.6 time-frame there have been many significant improvements to
173 <p>The set of checks performed by the static analyzer continues to expand, and
174 future plans for the tool include full source-level inter-procedural analysis
175 and deeper checks such as buffer overrun detection. There are many opportunities
176 to extend and enhance the static analyzer, and anyone interested in working on
177 this project is encouraged to get involved!</p>
181 <!--=========================================================================-->
182 <div class="doc_subsection">
183 <a name="vmkit">VMKit: JVM/CLI Virtual Machine Implementation</a>
186 <div class="doc_text">
188 The <a href="http://vmkit.llvm.org/">VMKit project</a> is an implementation of
189 a JVM and a CLI Virtual Machines (Microsoft .NET is an
190 implementation of the CLI) using the Just-In-Time compiler of LLVM.</p>
192 <p>Following LLVM 2.6, VMKit has its XYZ release that you can find on its
193 <a href="http://vmkit.llvm.org/releases/">webpage</a>. The release includes
194 bug fixes, cleanup and new features. The major changes are:</p>
198 <li>Something wonderful!</li>
203 <!-- *********************************************************************** -->
204 <div class="doc_section">
205 <a name="externalproj">External Projects Using LLVM 2.6</a>
207 <!-- *********************************************************************** -->
210 <!--=========================================================================-->
211 <div class="doc_subsection">
212 <a name="macruby">MacRuby</a>
215 <div class="doc_text">
218 <a href="http://macruby.org">MacRuby</a> is an implementation of Ruby on top of
219 core Mac OS X technologies, such as the Objective-C common runtime and garbage
220 collector, and the CoreFoundation framework. It is principally developed by
221 Apple and aims at enabling the creation of full-fledged Mac OS X applications.
225 MacRuby uses LLVM for optimization passes, JIT and AOT compilation of Ruby
226 expressions. It also uses zero-cost DWARF exceptions to implement Ruby exception
232 <!--=========================================================================-->
233 <div class="doc_subsection">
234 <a name="pure">Pure</a>
237 <div class="doc_text">
239 <a href="http://pure-lang.googlecode.com/">Pure</a>
240 is an algebraic/functional programming language based on term rewriting.
241 Programs are collections of equations which are used to evaluate expressions in
242 a symbolic fashion. Pure offers dynamic typing, eager and lazy evaluation,
243 lexical closures, a hygienic macro system (also based on term rewriting),
244 built-in list and matrix support (including list and matrix comprehensions) and
245 an easy-to-use C interface. The interpreter uses LLVM as a backend to
246 JIT-compile Pure programs to fast native code.</p>
248 <p>In addition to the usual algebraic data structures, Pure also has
249 MATLAB-style matrices in order to support numeric computations and signal
250 processing in an efficient way. Pure is mainly aimed at mathematical
251 applications right now, but it has been designed as a general purpose language.
252 The dynamic interpreter environment and the C interface make it possible to use
253 it as a kind of functional scripting language for many application areas.
258 <!--=========================================================================-->
259 <div class="doc_subsection">
260 <a name="ldc">LLVM D Compiler</a>
263 <div class="doc_text">
265 <a href="http://www.dsource.org/projects/ldc">LDC</a> is an implementation of
266 the D Programming Language using the LLVM optimizer and code generator.
267 The LDC project works great with the LLVM 2.6 release. General improvements in
269 cycle have included new inline asm constraint handling, better debug info
270 support, general bugfixes, and better x86-64 support. This has allowed
271 some major improvements in LDC, getting us much closer to being as
272 fully featured as the original DMD compiler from DigitalMars.
276 <!--=========================================================================-->
277 <div class="doc_subsection">
278 <a name="RoadsendPHP">Roadsend PHP</a>
281 <div class="doc_text">
282 <p><a href="http://code.roadsend.com/rphp">Roadsend PHP</a> (rphp) is an open
283 source implementation of the PHP programming
284 language that uses LLVM for its optimizer, JIT, and static compiler. This is a
285 reimplementation of an earlier project that is now based on LLVM.</p>
288 <!--=========================================================================-->
289 <div class="doc_subsection">
290 <a name="Unladen Swallow">Unladen Swallow</a>
293 <div class="doc_text">
294 <p><a href="http://code.google.com/p/unladen-swallow/">Unladen Swallow</a> is a
295 branch of <a href="http://python.org/">Python</a> intended to be fully
296 compatible and significantly faster. It uses LLVM's optimization passes and JIT
300 <!--=========================================================================-->
301 <div class="doc_subsection">
302 <a name="Rubinius">Rubinius</a>
305 <div class="doc_text">
306 <p><a href="http://github.com/evanphx/rubinius">Rubinius</a> is a new virtual
307 machine for Ruby. It leverages LLVM to dynamically compile Ruby code down to
308 machine code using LLVM's JIT.</p>
312 <!-- *********************************************************************** -->
313 <div class="doc_section">
314 <a name="whatsnew">What's New in LLVM 2.6?</a>
316 <!-- *********************************************************************** -->
318 <div class="doc_text">
320 <p>This release includes a huge number of bug fixes, performance tweaks, and
321 minor improvements. Some of the major improvements and new features are listed
327 <!--=========================================================================-->
328 <div class="doc_subsection">
329 <a name="majorfeatures">Major New Features</a>
332 <div class="doc_text">
334 <p>LLVM 2.6 includes several major new capabilities:</p>
337 <li>Support for debug line numbers when optimization enabled</li>
338 <li>gold lto plugin</li>
339 <li>New MSP430 and SystemZ backends.</li>
340 <li>New BlackFin backend.</li>
341 <li>LLVMContext, llvm_start_multithreaded: ProgrammersManual.html#threading</li>
342 <li>Unladen swallow as user?</li>
343 <li>klee web page at klee.llvm.org</li>
345 <li>New compiler-rt project.</li>
346 <li>LLVM 2.6 includes a brand new experimental LLVM bindings to the Ada2005 programming language.</li>
352 <!--=========================================================================-->
353 <div class="doc_subsection">
354 <a name="llvm-gcc">llvm-gcc 4.2 Improvements</a>
357 <div class="doc_text">
359 <p>LLVM fully supports the llvm-gcc 4.2 front-end, which marries the GCC
360 front-ends and driver with the LLVM optimizer and code generator. It currently
361 includes support for the C, C++, Objective-C, Ada, and Fortran front-ends.</p>
364 <li>Something wonderful!</li>
370 <!--=========================================================================-->
371 <div class="doc_subsection">
372 <a name="coreimprovements">LLVM IR and Core Improvements</a>
375 <div class="doc_text">
376 <p>LLVM IR has several new features that are used by our existing front-ends and
377 can be useful if you are writing a front-end for LLVM:</p>
380 <li>Getelementpr instruction now allows any integer type for array/pointer indexes.</li>
381 <li>Inbounds for GEP</li>
382 <li>NSW/NUW/exact div</li>
383 <li>LSR promotes int induction variables to 64-bit on 64-bit targets, major perf boost for numerical code.</li>
384 <li>LSR now analyzes pointer expressions (e.g. getelementptrs), not just integers.</li>
385 <li>new linkage types linkonce_odr, weak_odr, linker_private, and available_externally.</li>
386 <li>New fadd, fsub, fmul instructions and classes. </li>
387 <li>Target intrinsics can now return multiple results.</li>
392 <!--=========================================================================-->
393 <div class="doc_subsection">
394 <a name="optimizer">Optimizer Improvements</a>
397 <div class="doc_text">
399 <p>In addition to a large array of bug fixes and minor performance tweaks, this
400 release includes a few major enhancements and additions to the optimizers:</p>
404 <li>SRoA improvements for vector unions, memset, arbitrary weird bitfield accesses etc. It now produces "strange" sized integers.</li>
405 <li>Inliner reuse stack space when inlining arrays?</li>
406 <li>Enabled GVN Load PRE.</li>
407 <li>New Static Single Information (SSI) construction pass (not used by anything yet, experimental).</li>
414 <!--=========================================================================-->
415 <div class="doc_subsection">
416 <a name="codegen">Target Independent Code Generator Improvements</a>
419 <div class="doc_text">
421 <p>We have put a significant amount of work into the code generator
422 infrastructure, which allows us to implement more aggressive algorithms and make
427 <li> -asm-verbose now prints location info (with -g) and loop nest info.</li>
428 <li>Tblgen now supports multiclass inheritance and a number of new string and
429 list operations like !(subst), !(foreach), !car, !cdr, !null, !if, !cast.
430 These make the .td files more expressive and allow more aggressive factoring
431 of duplication across instruction patterns.</li>
432 <li>New MachineVerifier pass.</li>
433 <li>Machine LICM, hoists things like constant pool loads, loads from readonly stubs, vector constant synthesization code, etc.</li>
434 <li>Machine Sinking</li>
435 <li>target-specific intrinsics (r63765)</li>
436 <li>Regalloc improvements for commuting, various spiller peephole optimizations, cross-class coalescing.</li>
437 <li><tt>llc -enable-value-prop</tt>, propagation of value info (sign/zero ext info) from one MBB to another</li>
438 <li>Regalloc hints for allocation stuff: Evan r73381/r73671. Finished/enabled?</li>
439 <li>Stack slot coloring for register spills (denser stack frames)</li>
440 <li>SelectionDAGS: New BuildVectorSDNode (r65296), and ISD::VECTOR_SHUFFLE (r69952 / PR2957)</li>
441 <li>PostRA scheduler improvements David Goodwin.</li>
446 <!--=========================================================================-->
447 <div class="doc_subsection">
448 <a name="x86">X86-32 and X86-64 Target Improvements</a>
451 <div class="doc_text">
452 <p>New features of the X86 target include:
457 <li>Preliminary support for addrspace 256 -> GS, 257 -> FS, known problems: CodeGenerator.html#x86_memory</li>
458 <li>Support for softfloat modes, typically used by OS kernels.</li>
460 <li>X86-64: better modeling of implicit zero extensions, eliminates a lot of redundant zexts</li>
461 <li>X86-64 TLS support for local exec and initial exec.</li>
462 <li>Better modeling of H registerts as subregs.</li>
463 <li>Vector icmp/fcmp now work with SSE codegen.</li>
464 <li>SSE 4.2 support.</li>
465 <li>all global variable reference logic is now in ClassifyGlobalReference.</li>
471 <!--=========================================================================-->
472 <div class="doc_subsection">
473 <a name="pic16">PIC16 Target Improvements</a>
476 <div class="doc_text">
477 <p>New features of the PIC16 target include:
481 <li>Support for floating-point, indirect function calls, and
482 passing/returning aggregate types to functions.
483 <li>The code generator is able to generate debug info into output COFF files.
484 <li>Support for placing an object into a specific section or at a specific
485 address in memory.</li>
488 <p>Things not yet supported:</p>
491 <li>Variable arguments.</li>
492 <li>Interrupts/programs.</li>
497 <!--=========================================================================-->
498 <div class="doc_subsection">
499 <a name="ARM">ARM Target Improvements</a>
502 <div class="doc_text">
503 <p>New features of the ARM target include:
508 <li>Preliminary support for processors, such as the Cortex-A8 and Cortex-A9,
509 that implement version v7-A of the ARM architecture. The ARM backend now
510 supports both the Thumb2 and Advanced SIMD (Neon) instruction sets. The
511 AAPCS-VFP "hard float" calling conventions are also supported with the
512 <tt>-float-abi=hard</tt> flag. These features are still somewhat experimental
513 and subject to change. The Neon intrinsics, in particular, may change in future
517 ARM AAPCS-VFP hard float ABI is supported.
518 ARM calling convention code is now tblgen generated instead of manual.
519 ARM: NEON support. neonfp for doing single precision fp with neon instead of VFP.
525 <!--=========================================================================-->
526 <div class="doc_subsection">
527 <a name="OtherTarget">Other Target Specific Improvements</a>
530 <div class="doc_text">
531 <p>New features of other targets include:
535 <li>Add support for the PowerPC 64-bit SVR4 ABI.</li>
536 <li>Mips now supports O32 Calling Convention.</li>
541 <!--=========================================================================-->
542 <div class="doc_subsection">
543 <a name="newapis">New Useful APIs</a>
546 <div class="doc_text">
549 <li>New EngineBuilder class for creating JITs: r76276</li>
550 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.
551 StringRef class, Twine class.
552 New WeakVH and AssertingVH and CallbackVH classes.
553 New llvm/ADT/Triple class.
554 llvm_report_error() error handling API (llvm/Support/ErrorHandling.h)
555 New llvm/System/Atomic.h, llvm/System/RWMutex.h for portable atomic ops, rw locks.
556 New SourceMgr, SMLoc classes for simple parsers with caret diagnostics and #include support, (used by
557 tablegen, llvm-mc, the .ll parser, FileCheck, etc)
564 <!--=========================================================================-->
565 <div class="doc_subsection">
566 <a name="otherimprovements">Other Improvements and New Features</a>
569 <div class="doc_text">
570 <p>Other miscellaneous features include:</p>
573 <li>interpreter + libffi</li>
574 <li> JIT now supports generating more than 16M of code.</li>
576 href="http://llvm.org/doxygen/classllvm_1_1ExecutionEngine.html#fcd2b4b92ca38812ca31640b0da14927">register</a>
577 a <a href="http://llvm.org/doxygen/classllvm_1_1JITEventListener.html">JITEventListener</a>
578 to receive callbacks when the JIT emits or frees machine code. The
579 OProfile support uses this mechanism.</li>
580 JIT support for oprofile (r75279), configure with --with-oprofile. Now we get line # and function info for JIT'd functions.
582 <li> Profile info improvements by Andreas Neustifter.</li>
583 <li> Many extensions to the C APIs.</li>
589 <!--=========================================================================-->
590 <div class="doc_subsection">
591 <a name="changes">Major Changes and Removed Features</a>
594 <div class="doc_text">
596 <p>If you're already an LLVM user or developer with out-of-tree changes based
597 on LLVM 2.5, this section lists some "gotchas" that you may run into upgrading
598 from the previous release.</p>
602 <li>The Itanium (IA64) backend has been removed. It was not supported and
604 <li>The BigBlock register allocator has been removed, it also bitrotted.</li>
605 <li>The C Backend (-march=c) is no longer considered part of the LLVM release
606 criteria. We still want it to work, but no one is maintaining it and it lacks
607 support for arbitrary precision integers and other important IR features.</li>
611 <p>In addition, many APIs have changed in this release. Some of the major LLVM
616 no use of hash_set/hash_map, no more llvm::OStream
617 Use raw_ostream for everything, killed off llvm/Streams.h and DOUT
620 <li>LLVM's global uniquing tables for <tt>Type</tt>s and <tt>Constant</tt>s have
621 been privatized into members of an <tt>LLVMContext</tt>. A number of APIs
622 now take an <tt>LLVMContext</tt> as a parameter. To smooth the transition
623 for clients that will only ever use a single context, the new
624 <tt>getGlobalContext()</tt> API can be used to access a default global
625 context which can be passed in any and all cases where a context is
627 <li>The <tt>getABITypeSize</tt> methods are now called <tt>getAllocSize</tt>.</li>
628 <li>The <tt>Add</tt>, <tt>Sub</tt>, and <tt>Mul</tt> operators are no longer
629 overloaded for floating-point types. Floating-point addition, subtraction,
630 and multiplication are now represented with new operators <tt>FAdd</tt>,
631 <tt>FSub</tt>, and <tt>FMul</tt>. In the <tt>IRBuilder</tt> API,
632 <tt>CreateAdd</tt>, <tt>CreateSub</tt>, <tt>CreateMul</tt>, and
633 <tt>CreateNeg</tt> should only be used for integer arithmetic now;
634 <tt>CreateFAdd</tt>, <tt>CreateFSub</tt>, <tt>CreateFMul</tt>, and
635 <tt>CreateFNeg</tt> should now be used for floating-point arithmetic.</li>
636 <li>The DynamicLibrary class can no longer be constructed, its functionality has
637 moved to static member functions.</li>
638 <li><tt>raw_fd_ostream</tt>'s constructor for opening a given filename now
639 takes an extra <tt>Force</tt> argument. If <tt>Force</tt> is set to
640 <tt>false</tt>, an error will be reported if a file with the given name
641 already exists. If <tt>Force</tt> is set to <tt>true</tt>, the file will
642 be silently truncated (which is the behavior before this flag was
644 <li><tt>SCEVHandle</tt> no longer exists, because reference counting is no
645 longer done for <tt>SCEV*</tt> objects, instead <tt>const SCEV*</tt> should be
648 <li>Many APIs, notably <tt>llvm::Value</tt>, now use the <tt>StringRef</tt>
649 and <tt>Twine</tt> classes instead of passing <tt>const char*</tt>
650 or <tt>std::string</tt>, as described in
651 the <a href="ProgrammersManual.html#string_apis">Programmer's Manual</a>. Most
652 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:
654 <li><tt>getNameStr()</tt> is still available, and matches the old
655 behavior. Replacing <tt>getName()</tt> calls with this is an safe option,
656 although more efficient alternatives are now possible.</li>
658 <li>If you were just relying on <tt>getName()</tt> being able to be sent to
659 a <tt>std::ostream</tt>, consider migrating
660 to <tt>llvm::raw_ostream</tt>.</li>
662 <li>If you were using <tt>getName().c_str()</tt> to get a <tt>const
663 char*</tt> pointer to the name, you can use <tt>getName().data()</tt>.
664 Note that this string (as before), may not be the entire name if the
665 name containts embedded null characters.</li>
667 <li>If you were using operator plus on the result of <tt>getName()</tt> and
668 treating the result as an <tt>std::string</tt>, you can either
669 uses <tt>Twine::str</tt> to get the result as an <tt>std::string</tt>, or
670 could move to a <tt>Twine</tt> based design.</li>
672 <li><tt>isName()</tt> should be replaced with comparison
673 against <tt>getName()</tt> (this is now efficient).
677 <li>The registration interfaces for backend Targets has changed (what was
678 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:
680 <li><tt>TargetMachineRegistry</tt> has been renamed
681 to <tt>TargetRegistry</tt>.</li>
683 <li>Clients should move to using the <tt>TargetRegistry::lookupTarget()</tt>
684 function to find targets.</li>
688 <li>llvm-dis now fails if output file exists, instead of dumping to stdout.
689 FIXME: describe any other tool changes due to the raw_fd_ostream change. FIXME:
690 This is not an API change, maybe there should be a tool changes section?</li>
691 <li>temporarely due to Context API change passes should call doInitialization()
692 method of the pass they inherit from, otherwise Context is NULL.
693 FIXME: remove this entry when this is no longer needed.<li>
700 <!-- *********************************************************************** -->
701 <div class="doc_section">
702 <a name="portability">Portability and Supported Platforms</a>
704 <!-- *********************************************************************** -->
706 <div class="doc_text">
708 <p>LLVM is known to work on the following platforms:</p>
711 <li>Intel and AMD machines (IA32, X86-64, AMD64, EMT-64) running Red Hat
712 Linux, Fedora Core, FreeBSD and AuroraUX (and probably other unix-like systems).</li>
713 <li>PowerPC and X86-based Mac OS X systems, running 10.3 and above in 32-bit
714 and 64-bit modes.</li>
715 <li>Intel and AMD machines running on Win32 using MinGW libraries (native).</li>
716 <li>Intel and AMD machines running on Win32 with the Cygwin libraries (limited
717 support is available for native builds with Visual C++).</li>
718 <li>Sun UltraSPARC workstations running Solaris 10.</li>
719 <li>Alpha-based machines running Debian GNU/Linux.</li>
722 <p>The core LLVM infrastructure uses GNU autoconf to adapt itself
723 to the machine and operating system on which it is built. However, minor
724 porting may be required to get LLVM to work on new platforms. We welcome your
725 portability patches and reports of successful builds or error messages.</p>
729 <!-- *********************************************************************** -->
730 <div class="doc_section">
731 <a name="knownproblems">Known Problems</a>
733 <!-- *********************************************************************** -->
735 <div class="doc_text">
737 <p>This section contains significant known problems with the LLVM system,
738 listed by component. If you run into a problem, please check the <a
739 href="http://llvm.org/bugs/">LLVM bug database</a> and submit a bug if
740 there isn't already one.</p>
743 <li>LLVM will not correctly compile on Solaris and/or OpenSolaris
744 using the stock GCC 3.x.x series 'out the box',
745 See: <a href="#brokengcc">Broken versions of GCC and other tools</a>.
746 However, A <a href="http://pkg.auroraux.org/GCC">Modern GCC Build</a>
747 for x86/x64 has been made available from the third party AuroraUX Project
748 that has been meticulously tested for bootstrapping LLVM & Clang.</li>
753 <!-- ======================================================================= -->
754 <div class="doc_subsection">
755 <a name="experimental">Experimental features included with this release</a>
758 <div class="doc_text">
760 <p>The following components of this LLVM release are either untested, known to
761 be broken or unreliable, or are in early development. These components should
762 not be relied on, and bugs should not be filed against them, but they may be
763 useful to some people. In particular, if you would like to work on one of these
764 components, please contact us on the <a
765 href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVMdev list</a>.</p>
768 <li>The MSIL, Alpha, SPU, MIPS, and PIC16 backends are experimental.</li>
769 <li>The <tt>llc</tt> "<tt>-filetype=asm</tt>" (the default) is the only
770 supported value for this option.</li>
775 <!-- ======================================================================= -->
776 <div class="doc_subsection">
777 <a name="x86-be">Known problems with the X86 back-end</a>
780 <div class="doc_text">
783 <li>The X86 backend does not yet support
784 all <a href="http://llvm.org/PR879">inline assembly that uses the X86
785 floating point stack</a>. It supports the 'f' and 't' constraints, but not
787 <li>The X86 backend generates inefficient floating point code when configured
788 to generate code for systems that don't have SSE2.</li>
789 <li>Win64 code generation wasn't widely tested. Everything should work, but we
790 expect small issues to happen. Also, llvm-gcc cannot build the mingw64
791 runtime currently due
792 to <a href="http://llvm.org/PR2255">several</a>
793 <a href="http://llvm.org/PR2257">bugs</a> and due to lack of support for
795 'u' inline assembly constraint and for X87 floating point inline assembly.</li>
796 <li>The X86-64 backend does not yet support the LLVM IR instruction
797 <tt>va_arg</tt>. Currently, the llvm-gcc and front-ends support variadic
798 argument constructs on X86-64 by lowering them manually.</li>
803 <!-- ======================================================================= -->
804 <div class="doc_subsection">
805 <a name="ppc-be">Known problems with the PowerPC back-end</a>
808 <div class="doc_text">
811 <li>The Linux PPC32/ABI support needs testing for the interpreter and static
812 compilation, and lacks support for debug information.</li>
817 <!-- ======================================================================= -->
818 <div class="doc_subsection">
819 <a name="arm-be">Known problems with the ARM back-end</a>
822 <div class="doc_text">
825 <li>Support for the Advanced SIMD (Neon) instruction set is still incomplete
826 and not well tested. Some features may not work at all, and the code quality
827 may be poor in some cases.</li>
828 <li>Thumb mode works only on ARMv6 or higher processors. On sub-ARMv6
829 processors, thumb programs can crash or produce wrong
830 results (<a href="http://llvm.org/PR1388">PR1388</a>).</li>
831 <li>Compilation for ARM Linux OABI (old ABI) is supported but not fully tested.
837 <!-- ======================================================================= -->
838 <div class="doc_subsection">
839 <a name="sparc-be">Known problems with the SPARC back-end</a>
842 <div class="doc_text">
845 <li>The SPARC backend only supports the 32-bit SPARC ABI (-m32); it does not
846 support the 64-bit SPARC ABI (-m64).</li>
851 <!-- ======================================================================= -->
852 <div class="doc_subsection">
853 <a name="mips-be">Known problems with the MIPS back-end</a>
856 <div class="doc_text">
859 <li>The O32 ABI is not fully supported.</li>
860 <li>64-bit MIPS targets are not supported yet.</li>
865 <!-- ======================================================================= -->
866 <div class="doc_subsection">
867 <a name="alpha-be">Known problems with the Alpha back-end</a>
870 <div class="doc_text">
874 <li>On 21164s, some rare FP arithmetic sequences which may trap do not have the
875 appropriate nops inserted to ensure restartability.</li>
880 <!-- ======================================================================= -->
881 <div class="doc_subsection">
882 <a name="c-be">Known problems with the C back-end</a>
885 <div class="doc_text">
888 <li><a href="http://llvm.org/PR802">The C backend has only basic support for
889 inline assembly code</a>.</li>
890 <li><a href="http://llvm.org/PR1658">The C backend violates the ABI of common
891 C++ programs</a>, preventing intermixing between C++ compiled by the CBE and
892 C++ code compiled with <tt>llc</tt> or native compilers.</li>
893 <li>The C backend does not support all exception handling constructs.</li>
894 <li>The C backend does not support arbitrary precision integers.</li>
900 <!-- ======================================================================= -->
901 <div class="doc_subsection">
902 <a name="c-fe">Known problems with the llvm-gcc C front-end</a>
905 <div class="doc_text">
907 <p>llvm-gcc does not currently support <a href="http://llvm.org/PR869">Link-Time
908 Optimization</a> on most platforms "out-of-the-box". Please inquire on the
909 LLVMdev mailing list if you are interested.</p>
911 <p>The only major language feature of GCC not supported by llvm-gcc is
912 the <tt>__builtin_apply</tt> family of builtins. However, some extensions
913 are only supported on some targets. For example, trampolines are only
914 supported on some targets (these are used when you take the address of a
915 nested function).</p>
917 <p>If you run into GCC extensions which are not supported, please let us know.
922 <!-- ======================================================================= -->
923 <div class="doc_subsection">
924 <a name="c++-fe">Known problems with the llvm-gcc C++ front-end</a>
927 <div class="doc_text">
929 <p>The C++ front-end is considered to be fully
930 tested and works for a number of non-trivial programs, including LLVM
931 itself, Qt, Mozilla, etc.</p>
934 <li>Exception handling works well on the X86 and PowerPC targets. Currently
935 only Linux and Darwin targets are supported (both 32 and 64 bit).</li>
940 <!-- ======================================================================= -->
941 <div class="doc_subsection">
942 <a name="fortran-fe">Known problems with the llvm-gcc Fortran front-end</a>
945 <div class="doc_text">
947 <li>Fortran support generally works, but there are still several unresolved bugs
948 in Bugzilla. Please see the tools/gfortran component for details.</li>
952 <!-- ======================================================================= -->
953 <div class="doc_subsection">
954 <a name="ada-fe">Known problems with the llvm-gcc Ada front-end</a>
957 <div class="doc_text">
958 The llvm-gcc 4.2 Ada compiler works fairly well; however, this is not a mature
959 technology, and problems should be expected.
961 <li>The Ada front-end currently only builds on X86-32. This is mainly due
962 to lack of trampoline support (pointers to nested functions) on other platforms.
963 However, it <a href="http://llvm.org/PR2006">also fails to build on X86-64</a>
964 which does support trampolines.</li>
965 <li>The Ada front-end <a href="http://llvm.org/PR2007">fails to bootstrap</a>.
966 This is due to lack of LLVM support for <tt>setjmp</tt>/<tt>longjmp</tt> style
967 exception handling, which is used internally by the compiler.
968 Workaround: configure with --disable-bootstrap.</li>
969 <li>The c380004, <a href="http://llvm.org/PR2010">c393010</a>
970 and <a href="http://llvm.org/PR2421">cxg2021</a> ACATS tests fail
971 (c380004 also fails with gcc-4.2 mainline).
972 If the compiler is built with checks disabled then <a href="http://llvm.org/PR2010">c393010</a>
973 causes the compiler to go into an infinite loop, using up all system memory.</li>
974 <li>Some GCC specific Ada tests continue to crash the compiler.</li>
975 <li>The -E binder option (exception backtraces)
976 <a href="http://llvm.org/PR1982">does not work</a> and will result in programs
977 crashing if an exception is raised. Workaround: do not use -E.</li>
978 <li>Only discrete types <a href="http://llvm.org/PR1981">are allowed to start
979 or finish at a non-byte offset</a> in a record. Workaround: do not pack records
980 or use representation clauses that result in a field of a non-discrete type
981 starting or finishing in the middle of a byte.</li>
982 <li>The <tt>lli</tt> interpreter <a href="http://llvm.org/PR2009">considers
983 'main' as generated by the Ada binder to be invalid</a>.
984 Workaround: hand edit the file to use pointers for <tt>argv</tt> and
985 <tt>envp</tt> rather than integers.</li>
986 <li>The <tt>-fstack-check</tt> option <a href="http://llvm.org/PR2008">is
991 <!-- ======================================================================= -->
992 <div class="doc_subsection">
993 <a name="ocaml-bindingse">Known problems with the O'Caml bindings</a>
996 <div class="doc_text">
998 <p>The Llvm.Linkage module is broken, and has incorrect values. Only
999 Llvm.Linkage.External, Llvm.Linkage.Available_externally, and
1000 Llvm.Linkage.Link_once will be correct. If you need any of the other linkage
1001 modes, you'll have to write an external C library in order to expose the
1002 functionality. This has been fixed in the trunk.</p>
1005 <!-- *********************************************************************** -->
1006 <div class="doc_section">
1007 <a name="additionalinfo">Additional Information</a>
1009 <!-- *********************************************************************** -->
1011 <div class="doc_text">
1013 <p>A wide variety of additional information is available on the <a
1014 href="http://llvm.org">LLVM web page</a>, in particular in the <a
1015 href="http://llvm.org/docs/">documentation</a> section. The web page also
1016 contains versions of the API documentation which is up-to-date with the
1017 Subversion version of the source code.
1018 You can access versions of these documents specific to this release by going
1019 into the "<tt>llvm/doc/</tt>" directory in the LLVM tree.</p>
1021 <p>If you have any questions or comments about LLVM, please feel free to contact
1022 us via the <a href="http://llvm.org/docs/#maillist"> mailing
1027 <!-- *********************************************************************** -->
1031 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
1032 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
1033 <a href="http://validator.w3.org/check/referer"><img
1034 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
1036 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
1037 Last modified: $Date$