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 2.0 Release Notes</title>
11 <div class="doc_title">LLVM 2.0 Release Notes</div>
14 <li><a href="#intro">Introduction</a></li>
15 <li><a href="#whatsnew">What's New?</a></li>
16 <li><a href="GettingStarted.html">Installation Instructions</a></li>
17 <li><a href="#portability">Portability and Supported Platforms</a></li>
18 <li><a href="#knownproblems">Known Problems</a>
19 <li><a href="#additionalinfo">Additional Information</a></li>
22 <div class="doc_author">
23 <p>Written by the <a href="http://llvm.org">LLVM Team</a><p>
26 <!-- *********************************************************************** -->
27 <div class="doc_section">
28 <a name="intro">Introduction</a>
30 <!-- *********************************************************************** -->
32 <div class="doc_text">
34 <p>This document contains the release notes for the LLVM compiler
35 infrastructure, release 2.0. Here we describe the status of LLVM, including
36 major improvements from the previous release and any known problems. All LLVM
37 releases may be downloaded from the <a href="http://llvm.org/releases/">LLVM
38 releases web site</a>.</p>
40 <p>For more information about LLVM, including information about the latest
41 release, please check out the <a href="http://llvm.org/">main LLVM
42 web site</a>. If you have questions or comments, the <a
43 href="http://mail.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVM developer's mailing
44 list</a> is a good place to send them.</p>
46 <p>Note that if you are reading this file from a Subversion checkout or the
48 this document applies to the <i>next</i> release, not the current one. To see
49 the release notes for the current or previous releases, see the <a
50 href="http://llvm.org/releases/">releases page</a>.</p>
54 <!-- *********************************************************************** -->
55 <div class="doc_section">
56 <a name="whatsnew">What's New?</a>
58 <!-- *********************************************************************** -->
60 <div class="doc_text">
62 <p>This is the eleventh public release of the LLVM Compiler Infrastructure.
63 Being the first major release since 1.0, this release is different in several
64 ways from our previous releases:</p>
67 <li>We took this as an opportunity to
68 break backwards compatibility with the LLVM 1.x bytecode and .ll file format.
69 If you have LLVM 1.9 .ll files that you would like to upgrade to LLVM 2.x, we
70 recommend the use of the stand alone <a href="#llvm-upgrade">llvm-upgrade</a>
71 tool (which is included with 2.0). We intend to keep compatibility with .ll
72 and .bc formats within the 2.x release series, like we did within the 1.x
74 <li>There are several significant change to the LLVM IR and internal APIs, such
75 as a major overhaul of the type system, the completely new bitcode file
76 format, etc (described below).</li>
77 <li>We designed the release around a 6 month release cycle instead of the usual
78 3-month cycle. This gave us extra time to develop and test some of the
79 more invasive features in this release.</li>
80 <li>LLVM 2.0 no longer supports the llvm-gcc3 front-end. Users are required to
81 upgrade to llvm-gcc4. llvm-gcc4 includes many features over
82 llvm-gcc3, is faster, and is <a href="CFEBuildInstrs.html">much easier to
83 build from source</a>.</li>
86 <p>Note that while this is a major version bump, this release has been
87 extensively tested on a wide range of software. It is easy to say that this
88 is our best release yet, in terms of both features and correctness. This is
89 the first LLVM release to correctly compile and optimize major software like
90 LLVM itself, Mozilla/Seamonkey, Qt 4.3rc1, kOffice, etc out of the box on
96 <!--=========================================================================-->
97 <div class="doc_subsection">
98 <a name="newfeatures">New Features in LLVM 2.0</a>
101 <!--_________________________________________________________________________-->
102 <div class="doc_subsubsection"><a name="majorchanges">Major Changes</a></div>
103 <div class="doc_text">
105 <p>Changes to the LLVM IR itself:</p>
109 <li>Integer types are now completely signless. This means that we
110 have types like i8/i16/i32 instead of ubyte/sbyte/short/ushort/int
111 etc. LLVM operations that depend on sign have been split up into
112 separate instructions (<a href="http://llvm.org/PR950">PR950</a>). This
113 eliminates cast instructions that just change the sign of the operands (e.g.
114 int -> uint), which reduces the size of the IR and makes optimizers
115 simpler to write.</li>
117 <li>Integer types with arbitrary bitwidths (e.g. i13, i36, i42, i1057, etc) are
118 now supported in the LLVM IR and optimizations (<a
119 href="http://llvm.org/PR1043">PR1043</a>). However, neither llvm-gcc
120 (<a href="http://llvm.org/PR1284">PR1284</a>) nor the native code generators
121 (<a href="http://llvm.org/PR1270">PR1270</a>) support non-standard width
124 <li>'Type planes' have been removed (<a href="http://llvm.org/PR411">PR411</a>).
125 It is no longer possible to have two values with the same name in the
126 same symbol table. This simplifies LLVM internals, allowing significant
129 <li>Global variables and functions in .ll files are now prefixed with
130 @ instead of % (<a href="http://llvm.org/PR645">PR645</a>).</li>
132 <li>The LLVM 1.x "bytecode" format has been replaced with a
133 completely new binary representation, named 'bitcode'. The <a
134 href="BitCodeFormat.html">Bitcode Format</a> brings a
135 number of advantages to the LLVM over the old bytecode format: it is denser
136 (files are smaller), more extensible, requires less memory to read,
137 is easier to keep backwards compatible (so LLVM 2.5 will read 2.0 .bc
138 files), and has many other nice features.</li>
140 <li>Load and store instructions now track the alignment of their pointer
141 (<a href="http://www.llvm.org/PR400">PR400</a>). This allows the IR to
142 express loads that are not sufficiently aligned (e.g. due to '<tt>#pragma
143 packed</tt>') or to capture extra alignment information.</li>
146 <p>Major new features:</p>
150 <li>A number of ELF features are now supported by LLVM, including 'visibility',
151 extern weak linkage, Thread Local Storage (TLS) with the <tt>__thread</tt>
152 keyword, and symbol aliases.
153 Among other things, this means that many of the special options needed to
154 configure llvm-gcc on linux are no longer needed, and special hacks to build
155 large C++ libraries like Qt are not needed.</li>
157 <li>LLVM now has a new MSIL backend. <tt>llc -march=msil</tt> will now turn LLVM
158 into MSIL (".net") bytecode. This is still fairly early development
159 with a number of limitations.</li>
161 <li>A new <a href="CommandGuide/html/llvm-upgrade.html">llvm-upgrade</a> tool
162 exists to migrates LLVM 1.9 .ll files to LLVM 2.0 syntax.</li>
168 <!--_________________________________________________________________________-->
169 <div class="doc_subsubsection"><a name="llvmgccfeatures">llvm-gcc
170 Improvements</a></div>
171 <div class="doc_text">
172 <p>New features include:
176 <li>Precompiled Headers (PCH) are now supported.</li>
178 <li>"<tt>#pragma packed</tt>" is now supported, as are the various features
179 described above (visibility, extern weak linkage, __thread, aliases,
182 <li>Tracking function parameter/result attributes is now possible.</li>
184 <li>Many internal enhancements have been added, such as improvements to
185 NON_LVALUE_EXPR, arrays with non-zero base, structs with variable sized
186 fields, VIEW_CONVERT_EXPR, CEIL_DIV_EXPR, nested functions, and many other
187 things. This is primarily to supports non-C GCC front-ends, like Ada.</li>
189 <li>It is simpler to configure llvm-gcc for linux.</li>
195 <!--_________________________________________________________________________-->
196 <div class="doc_subsubsection"><a name="optimizer">Optimizer
197 Improvements</a></div>
199 <div class="doc_text">
200 <p>New features include:
204 <li>The <a href="WritingAnLLVMPass.html">pass manager</a> has been entirely
205 rewritten, making it significantly smaller, simpler, and more extensible.
206 Support has been added to run <tt>FunctionPass</tt>es interlaced with
207 <tt>CallGraphSCCPass</tt>es, we now support loop transformations
208 explicitly with <tt>LoopPass</tt>, and <tt>ModulePass</tt>es may now use the
209 result of <tt>FunctionPass</tt>es.</li>
211 <li>LLVM 2.0 includes a new loop rotation pass, which converts "for loops" into
212 "do/while loops", where the condition is at the bottom of the loop.</li>
214 <li>The Loop Strength Reduction pass has been improved, and we now support
215 sinking expressions across blocks to reduce register pressure.</li>
217 <li>The <tt>-scalarrepl</tt> pass can now promote unions containing FP values
218 into a register, it can also handle unions of vectors of the same
221 <li>The [Post]DominatorSet classes have been removed from LLVM and clients
222 switched to use the more-efficient ETForest class instead.</li>
224 <li>The ImmediateDominator class has also been removed, and clients have been
225 switched to use DominatorTree instead.</li>
227 <li>The predicate simplifier pass has been improved, making it able to do
228 simple value range propagation and eliminate more conditionals. However,
229 note that predsimplify is not enabled by default in llvm-gcc.</li>
235 <!--_________________________________________________________________________-->
236 <div class="doc_subsubsection"><a name="codegen">Code
237 Generator Enhancements</a></div>
239 <div class="doc_text">
241 New features include:
246 <li>LLVM now supports software floating point, which allows LLVM to target
247 chips that don't have hardware FPUs (e.g. ARM thumb mode).</li>
249 <li>A new register scavenger has been implemented, which is useful for
250 finding free registers after register allocation. This is useful when
251 rewriting frame references on RISC targets, for example.</li>
253 <li>Heuristics have been added to avoid coalescing vregs with very large live
254 ranges to physregs. This was bad because it effectively pinned the physical
255 register for the entire lifetime of the virtual register (<a
256 href="http://llvm.org/PR711">PR711</a>).</li>
258 <li>Support now exists for very simple (but still very useful)
259 rematerialization the register allocator, enough to move
260 instructions like "load immediate" and constant pool loads.</li>
262 <li>Switch statement lowering is significantly better, improving codegen for
263 sparse switches that have dense subregions, and implemented support
264 for the shift/and trick.</li>
266 <li>LLVM now supports tracking physreg sub-registers and super-registers
267 in the code generator, and includes extensive register
268 allocator changes to track them.</li>
270 <li>There is initial support for virtreg sub-registers
271 (<a href="http://llvm.org/PR1350">PR1350</a>).</li>
276 Other improvements include:
281 <li>Inline assembly support is much more solid that before.
282 The two primary features still missing are support for 80-bit floating point
283 stack registers on X86 (<a href="http://llvm.org/PR879">PR879</a>), and
284 support for inline asm in the C backend (<a
285 href="http://llvm.org/PR802">PR802</a>).</li>
287 <li>DWARF debug information generation has been improved. LLVM now passes
288 most of the GDB testsuite on MacOS and debug info is more dense.</li>
290 <li>Codegen support for Zero-cost DWARF exception handling has been added (<a
291 href="http://llvm.org/PR592">PR592</a>). It is mostly
292 complete and just in need of continued bug fixes and optimizations at
293 this point. However, support in llvm-g++ is disabled with an
294 #ifdef for the 2.0 release (<a
295 href="http://llvm.org/PR870">PR870</a>).</li>
297 <li>The code generator now has more accurate and general hooks for
298 describing addressing modes ("isLegalAddressingMode") to
299 optimizations like loop strength reduction and code sinking.</li>
301 <li>Progress has been made on a direct Mach-o .o file writer. Many small
302 apps work, but it is still not quite complete.</li>
306 <p>In addition, the LLVM target description format has itself been extended in
310 <li>TargetData now supports better target parameterization in
311 the .ll/.bc files, eliminating the 'pointersize/endianness' attributes
312 in the files (<a href="http://llvm.org/PR761">PR761</a>).</li>
314 <li>TargetData was generalized for finer grained alignment handling,
315 handling of vector alignment, and handling of preferred alignment</li>
317 <li>LLVM now supports describing target calling conventions
318 explicitly in .td files, reducing the amount of C++ code that needs
319 to be written for a port.</li>
325 <!--_________________________________________________________________________-->
326 <div class="doc_subsubsection"><a name="specifictargets">Target-Specific
327 Improvements</a></div>
329 <div class="doc_text">
331 <p>X86-specific Code Generator Enhancements:
335 <li>The MMX instruction set is now supported through intrinsics.</li>
336 <li>The scheduler was improved to better reduce register pressure on
337 X86 and other targets that are register pressure sensitive.</li>
338 <li>Linux/x86-64 support is much better.</li>
339 <li>PIC support for linux/x86 has been added.</li>
340 <li>The X86 backend now supports the GCC regparm attribute.</li>
341 <li>LLVM now supports inline asm with multiple constraint letters per operand
342 (like "mri") which is common in X86 inline asms.</li>
345 <p>ARM-specific Code Generator Enhancements:</p>
348 <li>The ARM code generator is now stable and fully supported.</li>
350 <li>There are major new features, including support for ARM
351 v4-v6 chips, vfp support, soft float point support, pre/postinc support,
352 load/store multiple generation, constant pool entry motion (to support
353 large functions), inline asm support, weak linkage support, static
354 ctor/dtor support and many bug fixes.</li>
356 <li>Added support for Thumb code generation (<tt>llc -march=thumb</tt>).</li>
358 <li>The ARM backend now supports the ARM AAPCS/EABI ABI and PIC codegen on
361 <li>Several bugs were fixed for DWARF debug info generation on arm/linux.</li>
365 <p>PowerPC-specific Code Generator Enhancements:</p>
368 <li>The PowerPC 64 JIT now supports addressing code loaded above the 2G
371 <li>Improved support for the Linux/ppc ABI and the linux/ppc JIT is fully
372 functional now. llvm-gcc and static compilation are not fully supported
375 <li>Many PowerPC 64 bug fixes.</li>
382 <!--_________________________________________________________________________-->
383 <div class="doc_subsubsection"><a name="other">Other Improvements</a></div>
384 <div class="doc_text">
386 <p>More specific changes include:</p>
389 <li>LLVM no longer relies on static destructors to shut itself down. Instead,
390 it lazily initializes itself and shuts down when <tt>llvm_shutdown()</tt> is
391 explicitly called.</li>
393 <li>LLVM now has significantly fewer static constructors, reducing startup time.
396 <li>Several classes have been refactored to reduce the amount of code that
397 gets linked into apps that use the JIT.</li>
399 <li>Construction of intrinsic function declarations has been simplified.</li>
401 <li>The gccas/gccld tools have been replaced with small shell scripts.</li>
403 <li>Support has been added to llvm-test for running on low-memory
404 or slow machines (make SMALL_PROBLEM_SIZE=1).</li>
409 <!--_________________________________________________________________________-->
410 <div class="doc_subsubsection"><a name="apichanges">API Changes</a></div>
411 <div class="doc_text">
413 <p>LLVM 2.0 contains a revamp of the type system and several other significant
414 internal changes. If you are programming to the C++ API, be aware of the
415 following major changes:</p>
418 <li>Pass registration is slightly different in LLVM 2.0 (you now need an
419 <tt>intptr_t</tt> in your constructor), as explained in the <a
420 href="WritingAnLLVMPass.html#basiccode">Writing an LLVM Pass</a>
423 <li><tt>ConstantBool</tt>, <tt>ConstantIntegral</tt> and <tt>ConstantInt</tt>
424 classes have been merged together, we now just have
425 <tt>ConstantInt</tt>.</li>
427 <li><tt>Type::IntTy</tt>, <tt>Type::UIntTy</tt>, <tt>Type::SByteTy</tt>, ... are
428 replaced by <tt>Type::Int8Ty</tt>, <tt>Type::Int16Ty</tt>, etc. LLVM types
429 have always corresponded to fixed size types
430 (e.g. long was always 64-bits), but the type system no longer includes
431 information about the sign of the type. Also, the
432 <tt>Type::isPrimitiveType()</tt> method now returns false for integers.</li>
434 <li>Several classes (<tt>CallInst</tt>, <tt>GetElementPtrInst</tt>,
435 <tt>ConstantArray</tt>, etc), that once took <tt>std::vector</tt> as
436 arguments now take ranges instead. For example, you can create a
437 <tt>GetElementPtrInst</tt> with code like:
440 Value *Ops[] = { Op1, Op2, Op3 };
441 GEP = new GetElementPtrInst(BasePtr, Ops, 3);
444 This avoids creation of a temporary vector (and a call to malloc/free). If
445 you have an <tt>std::vector</tt>, use code like this:
447 std::vector<Value*> Ops = ...;
448 GEP = new GetElementPtrInst(BasePtr, &Ops[0], Ops.size());
453 <li><tt>CastInst</tt> is now abstract and its functionality is split into
454 several parts, one for each of the <a href="LangRef.html#convertops">new
455 cast instructions</a>.</li>
457 <li><tt>Instruction::getNext()/getPrev()</tt> are now private (along with
458 <tt>BasicBlock::getNext</tt>, etc), for efficiency reasons (they are now no
459 longer just simple pointers). Please use <tt>BasicBlock::iterator</tt>, etc
463 <li><tt>Module::getNamedFunction()</tt> is now called
464 <tt>Module::getFunction()</tt>.</li>
466 <li><tt>SymbolTable.h</tt> has been split into <tt>ValueSymbolTable.h</tt> and
467 <tt>TypeSymbolTable.h</tt>.</li>
472 <!-- *********************************************************************** -->
473 <div class="doc_section">
474 <a name="portability">Portability and Supported Platforms</a>
476 <!-- *********************************************************************** -->
478 <div class="doc_text">
480 <p>LLVM is known to work on the following platforms:</p>
483 <li>Intel and AMD machines running Red Hat Linux, Fedora Core and FreeBSD
484 (and probably other unix-like systems).</li>
485 <li>PowerPC and X86-based Mac OS X systems, running 10.2 and above in 32-bit and
487 <li>Intel and AMD machines running on Win32 using MinGW libraries (native)</li>
488 <li>Intel and AMD machines running on Win32 with the Cygwin libraries (limited
489 support is available for native builds with Visual C++).</li>
490 <li>Sun UltraSPARC workstations running Solaris 8.</li>
491 <li>Alpha-based machines running Debian GNU/Linux.</li>
492 <li>Itanium-based machines running Linux and HP-UX.</li>
495 <p>The core LLVM infrastructure uses
496 <a href="http://www.gnu.org/software/autoconf/">GNU autoconf</a> to adapt itself
497 to the machine and operating system on which it is built. However, minor
498 porting may be required to get LLVM to work on new platforms. We welcome your
499 portability patches and reports of successful builds or error messages.</p>
503 <!-- *********************************************************************** -->
504 <div class="doc_section">
505 <a name="knownproblems">Known Problems</a>
507 <!-- *********************************************************************** -->
509 <div class="doc_text">
511 <p>This section contains all known problems with the LLVM system, listed by
512 component. As new problems are discovered, they will be added to these
513 sections. If you run into a problem, please check the <a
514 href="http://llvm.org/bugs/">LLVM bug database</a> and submit a bug if
515 there isn't already one.</p>
519 <!-- ======================================================================= -->
520 <div class="doc_subsection">
521 <a name="experimental">Experimental features included with this release</a>
524 <div class="doc_text">
526 <p>The following components of this LLVM release are either untested, known to
527 be broken or unreliable, or are in early development. These components should
528 not be relied on, and bugs should not be filed against them, but they may be
529 useful to some people. In particular, if you would like to work on one of these
530 components, please contact us on the <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVMdev list</a>.</p>
533 <li>The <tt>-cee</tt> pass is known to be buggy, and may be removed in in a
535 <li>C++ EH support is disabled for this release.</li>
536 <li>The MSIL backend is experimental.</li>
537 <li>The IA64 code generator is experimental.</li>
538 <li>The Alpha JIT is experimental.</li>
539 <li>"<tt>-filetype=asm</tt>" (the default) is the only supported value for the
540 <tt>-filetype</tt> llc option.</li>
545 <!-- ======================================================================= -->
546 <div class="doc_subsection">
547 <a name="x86-be">Known problems with the X86 back-end</a>
550 <div class="doc_text">
553 <li>The X86 backend does not yet support <a href="http://llvm.org/PR879">inline
554 assembly that uses the X86 floating point stack</a>.</li>
559 <!-- ======================================================================= -->
560 <div class="doc_subsection">
561 <a name="ppc-be">Known problems with the PowerPC back-end</a>
564 <div class="doc_text">
567 <li><a href="http://llvm.org/PR642">PowerPC backend does not correctly
568 implement ordered FP comparisons</a>.</li>
569 <li>The Linux PPC32/ABI support needs testing for the interpreter and static
570 compilation, and lacks support for debug information.</li>
575 <!-- ======================================================================= -->
576 <div class="doc_subsection">
577 <a name="arm-be">Known problems with the ARM back-end</a>
580 <div class="doc_text">
583 <li>Thumb mode works only on ARMv6 or higher processors. On sub-ARMv6
584 processors, thumb program can crash or produces wrong
585 results (<a href="http://llvm.org/PR1388">PR1388</a>).</li>
586 <li>Compilation for ARM Linux OABI (old ABI) is supported, but not fully tested.
588 <li>There is a bug in QEMU-ARM (<= 0.9.0) which causes it to incorrectly execute
589 programs compiled with LLVM. Please use more recent versions of QEMU.</li>
594 <!-- ======================================================================= -->
595 <div class="doc_subsection">
596 <a name="sparc-be">Known problems with the SPARC back-end</a>
599 <div class="doc_text">
602 <li>The SPARC backend only supports the 32-bit SPARC ABI (-m32), it does not
603 support the 64-bit SPARC ABI (-m64).</li>
608 <!-- ======================================================================= -->
609 <div class="doc_subsection">
610 <a name="alpha-be">Known problems with the Alpha back-end</a>
613 <div class="doc_text">
617 <li>On 21164s, some rare FP arithmetic sequences which may trap do not have the
618 appropriate nops inserted to ensure restartability.</li>
623 <!-- ======================================================================= -->
624 <div class="doc_subsection">
625 <a name="ia64-be">Known problems with the IA64 back-end</a>
628 <div class="doc_text">
632 <li>C++ programs are likely to fail on IA64, as calls to <tt>setjmp</tt> are
633 made where the argument is not 16-byte aligned, as required on IA64. (Strictly
634 speaking this is not a bug in the IA64 back-end; it will also be encountered
635 when building C++ programs using the C back-end.)</li>
637 <li>The C++ front-end does not use <a href="http://llvm.org/PR406">IA64
638 ABI compliant layout of v-tables</a>. In particular, it just stores function
639 pointers instead of function descriptors in the vtable. This bug prevents
640 mixing C++ code compiled with LLVM with C++ objects compiled by other C++
643 <li>There are a few ABI violations which will lead to problems when mixing LLVM
644 output with code built with other compilers, particularly for floating-point
647 <li>Defining vararg functions is not supported (but calling them is ok).</li>
649 <li>The Itanium backend has bitrotted somewhat.</li>
654 <!-- ======================================================================= -->
655 <div class="doc_subsection">
656 <a name="c-be">Known problems with the C back-end</a>
659 <div class="doc_text">
662 <li><a href="http://llvm.org/PR802">The C backend does not support inline
663 assembly code</a>.</li>
669 <!-- ======================================================================= -->
670 <div class="doc_subsection">
671 <a name="c-fe">Known problems with the C front-end</a>
674 <!-- _______________________________________________________________________ -->
675 <div class="doc_subsubsection">Bugs</div>
677 <div class="doc_text">
679 <p>llvm-gcc4 does not currently support <a href="http://llvm.org/PR869">Link-Time
680 Optimization</a> on most platforms "out-of-the-box". Please inquire on the
681 llvmdev mailing list if you are interested.</p>
685 <!-- _______________________________________________________________________ -->
686 <div class="doc_subsubsection">
690 <div class="doc_text">
693 <li><p>"long double" is silently transformed by the front-end into "double". There
694 is no support for floating point data types of any size other than 32 and 64
697 <li><p>llvm-gcc does <b>not</b> support <tt>__builtin_apply</tt> yet.
698 See <a href="http://gcc.gnu.org/onlinedocs/gcc/Constructing-Calls.html#Constructing%20Calls">Constructing Calls</a>: Dispatching a call to another function.</p>
701 <li><p>llvm-gcc <b>partially</b> supports these GCC extensions:</p>
703 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Nested-Functions.html#Nested%20Functions">Nested Functions</a>: As in Algol and Pascal, lexical scoping of functions.<br>
704 Nested functions are supported, but llvm-gcc does not support non-local
705 gotos or taking the address of a nested function.</li>
707 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Function-Attributes.html#Function%20Attributes">Function Attributes</a>:
709 Declaring that functions have no side effects or that they can never
712 <b>Supported:</b> <tt>alias</tt>, <tt>always_inline</tt>, <tt>cdecl</tt>,
713 <tt>constructor</tt>, <tt>destructor</tt>,
714 <tt>deprecated</tt>, <tt>fastcall</tt>, <tt>format</tt>,
715 <tt>format_arg</tt>, <tt>non_null</tt>, <tt>noreturn</tt>, <tt>regparm</tt>
716 <tt>section</tt>, <tt>stdcall</tt>, <tt>unused</tt>, <tt>used</tt>,
717 <tt>visibility</tt>, <tt>warn_unused_result</tt>, <tt>weak</tt><br>
719 <b>Ignored:</b> <tt>noinline</tt>, <tt>pure</tt>, <tt>const</tt>, <tt>nothrow</tt>,
720 <tt>malloc</tt>, <tt>no_instrument_function</tt></li>
724 <li><p>llvm-gcc supports the vast majority of GCC extensions, including:</p>
727 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Pragmas.html#Pragmas">Pragmas</a>: Pragmas accepted by GCC.</li>
728 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Local-Labels.html#Local%20Labels">Local Labels</a>: Labels local to a block.</li>
729 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Other-Builtins.html#Other%20Builtins">Other Builtins</a>:
730 Other built-in functions.</li>
731 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Variable-Attributes.html#Variable%20Attributes">Variable Attributes</a>:
732 Specifying attributes of variables.</li>
733 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Type-Attributes.html#Type%20Attributes">Type Attributes</a>: Specifying attributes of types.</li>
734 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Thread_002dLocal.html">Thread-Local</a>: Per-thread variables.</li>
735 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Variable-Length.html#Variable%20Length">Variable Length</a>:
736 Arrays whose length is computed at run time.</li>
737 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Labels-as-Values.html#Labels%20as%20Values">Labels as Values</a>: Getting pointers to labels and computed gotos.</li>
738 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Statement-Exprs.html#Statement%20Exprs">Statement Exprs</a>: Putting statements and declarations inside expressions.</li>
739 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Typeof.html#Typeof">Typeof</a>: <code>typeof</code>: referring to the type of an expression.</li>
740 <li><a href="http://gcc.gnu.org/onlinedocs/gcc-3.4.0/gcc/Lvalues.html#Lvalues">Lvalues</a>: Using <code>?:</code>, "<code>,</code>" and casts in lvalues.</li>
741 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Conditionals.html#Conditionals">Conditionals</a>: Omitting the middle operand of a <code>?:</code> expression.</li>
742 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Long-Long.html#Long%20Long">Long Long</a>: Double-word integers.</li>
743 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Complex.html#Complex">Complex</a>: Data types for complex numbers.</li>
744 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Hex-Floats.html#Hex%20Floats">Hex Floats</a>:Hexadecimal floating-point constants.</li>
745 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html#Zero%20Length">Zero Length</a>: Zero-length arrays.</li>
746 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Empty-Structures.html#Empty%20Structures">Empty Structures</a>: Structures with no members.</li>
747 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Variadic-Macros.html#Variadic%20Macros">Variadic Macros</a>: Macros with a variable number of arguments.</li>
748 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Escaped-Newlines.html#Escaped%20Newlines">Escaped Newlines</a>: Slightly looser rules for escaped newlines.</li>
749 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html#Extended%20Asm">Extended Asm</a>: Assembler instructions with C expressions as operands.</li>
750 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Constraints.html#Constraints">Constraints</a>: Constraints for asm operands.</li>
751 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Asm-Labels.html#Asm%20Labels">Asm Labels</a>: Specifying the assembler name to use for a C symbol.</li>
752 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Explicit-Reg-Vars.html#Explicit%20Reg%20Vars">Explicit Reg Vars</a>: Defining variables residing in specified registers.</li>
753 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Vector-Extensions.html#Vector%20Extensions">Vector Extensions</a>: Using vector instructions through built-in functions.</li>
754 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Target-Builtins.html#Target%20Builtins">Target Builtins</a>: Built-in functions specific to particular targets.</li>
755 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Subscripting.html#Subscripting">Subscripting</a>: Any array can be subscripted, even if not an lvalue.</li>
756 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Pointer-Arith.html#Pointer%20Arith">Pointer Arith</a>: Arithmetic on <code>void</code>-pointers and function pointers.</li>
757 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Initializers.html#Initializers">Initializers</a>: Non-constant initializers.</li>
758 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Compound-Literals.html#Compound%20Literals">Compound Literals</a>: Compound literals give structures, unions,
759 or arrays as values.</li>
760 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Designated-Inits.html#Designated%20Inits">Designated Inits</a>: Labeling elements of initializers.</li>
761 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Cast-to-Union.html#Cast%20to%20Union">Cast to Union</a>: Casting to union type from any member of the union.</li>
762 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Case-Ranges.html#Case%20Ranges">Case Ranges</a>: `case 1 ... 9' and such.</li>
763 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Mixed-Declarations.html#Mixed%20Declarations">Mixed Declarations</a>: Mixing declarations and code.</li>
764 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Function-Prototypes.html#Function%20Prototypes">Function Prototypes</a>: Prototype declarations and old-style definitions.</li>
765 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/C_002b_002b-Comments.html#C_002b_002b-Comments">C++ Comments</a>: C++ comments are recognized.</li>
766 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Dollar-Signs.html#Dollar%20Signs">Dollar Signs</a>: Dollar sign is allowed in identifiers.</li>
767 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Character-Escapes.html#Character%20Escapes">Character Escapes</a>: <code>\e</code> stands for the character <ESC>.</li>
768 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Alignment.html#Alignment">Alignment</a>: Inquiring about the alignment of a type or variable.</li>
769 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Inline.html#Inline">Inline</a>: Defining inline functions (as fast as macros).</li>
770 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Alternate-Keywords.html#Alternate%20Keywords">Alternate Keywords</a>:<code>__const__</code>, <code>__asm__</code>, etc., for header files.</li>
771 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Incomplete-Enums.html#Incomplete%20Enums">Incomplete Enums</a>: <code>enum foo;</code>, with details to follow.</li>
772 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Function-Names.html#Function%20Names">Function Names</a>: Printable strings which are the name of the current function.</li>
773 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Return-Address.html#Return%20Address">Return Address</a>: Getting the return or frame address of a function.</li>
774 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Unnamed-Fields.html#Unnamed%20Fields">Unnamed Fields</a>: Unnamed struct/union fields within structs/unions.</li>
775 <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Attribute-Syntax.html#Attribute%20Syntax">Attribute Syntax</a>: Formal syntax for attributes.</li>
780 <p>If you run into GCC extensions which have not been included in any of these
781 lists, please let us know (also including whether or not they work).</p>
785 <!-- ======================================================================= -->
786 <div class="doc_subsection">
787 <a name="c++-fe">Known problems with the C++ front-end</a>
790 <div class="doc_text">
792 <p>The C++ front-end is considered to be fully
793 tested and works for a number of non-trivial programs, including LLVM
794 itself, Qt, Mozilla, etc.</p>
797 <li>llvm-gcc4 only has partial support for <a href="http://llvm.org/PR870">C++
798 Exception Handling</a>, and it is not enabled by default.</li>
802 <li>Destructors for local objects are not always run when a <tt>longjmp</tt> is
803 performed. In particular, destructors for objects in the <tt>longjmp</tt>ing
804 function and in the <tt>setjmp</tt> receiver function may not be run.
805 Objects in intervening stack frames will be destroyed, however (which is
806 better than most compilers).</li>
808 <li>The LLVM C++ front-end follows the <a
809 href="http://www.codesourcery.com/cxx-abi">Itanium C++ ABI</a>.
810 This document, which is not Itanium specific, specifies a standard for name
811 mangling, class layout, v-table layout, RTTI formats, and other C++
812 representation issues. Because we use this API, code generated by the LLVM
813 compilers should be binary compatible with machine code generated by other
814 Itanium ABI C++ compilers (such as G++, the Intel and HP compilers, etc).
815 <i>However</i>, the exception handling mechanism used by llvm-gcc3 is very
816 different from the model used in the Itanium ABI, so <b>exceptions will not
817 interact correctly</b>. </li>
825 <!-- *********************************************************************** -->
826 <div class="doc_section">
827 <a name="additionalinfo">Additional Information</a>
829 <!-- *********************************************************************** -->
831 <div class="doc_text">
833 <p>A wide variety of additional information is available on the <a
834 href="http://llvm.org">LLVM web page</a>, in particular in the <a
835 href="http://llvm.org/docs/">documentation</a> section. The web page also
836 contains versions of the API documentation which is up-to-date with the
837 Subversion version of the source code.
838 You can access versions of these documents specific to this release by going
839 into the "<tt>llvm/doc/</tt>" directory in the LLVM tree.</p>
841 <p>If you have any questions or comments about LLVM, please feel free to contact
842 us via the <a href="http://llvm.org/docs/#maillist"> mailing
847 <!-- *********************************************************************** -->
851 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
852 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
853 <a href="http://validator.w3.org/check/referer"><img
854 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
856 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
857 Last modified: $Date$