diff options
author | Tanya Lattner <tonic@nondot.org> | 2006-11-20 05:20:40 +0000 |
---|---|---|
committer | Tanya Lattner <tonic@nondot.org> | 2006-11-20 05:20:40 +0000 |
commit | 76e76afab4dd464a68966d0b93e4fa63d617c8cd (patch) | |
tree | d603343020b05521ab3bbef37110d45fe4c428b1 | |
parent | bb869226b839b6ea973fd389e969ba1d31d0cd57 (diff) | |
download | llvm-76e76afab4dd464a68966d0b93e4fa63d617c8cd.zip llvm-76e76afab4dd464a68966d0b93e4fa63d617c8cd.tar.gz llvm-76e76afab4dd464a68966d0b93e4fa63d617c8cd.tar.bz2 |
Merging from mainline
llvm-svn: 31868
-rw-r--r-- | llvm/docs/ReleaseNotes.html | 493 |
1 files changed, 262 insertions, 231 deletions
diff --git a/llvm/docs/ReleaseNotes.html b/llvm/docs/ReleaseNotes.html index c557069..172eba8 100644 --- a/llvm/docs/ReleaseNotes.html +++ b/llvm/docs/ReleaseNotes.html @@ -62,7 +62,8 @@ href="http://llvm.org/releases/">releases page</a>.</p> <div class="doc_text"> <p>This is the tenth public release of the LLVM Compiler Infrastructure. This -release incorporates a large number of enhancements and new features. +release incorporates a large number of enhancements, new features, and bug +fixes. We recommend that all users of previous LLVM versions upgrade. </p> </div> @@ -73,41 +74,106 @@ release incorporates a large number of enhancements and new features. </div> <!--_________________________________________________________________________--> -<div class="doc_subsubsection"><a name="elfdwarf">DWARF debugging -support for X86/ELF</a></div> +<div class="doc_subsubsection"><a name="x86-64">New X86-64 Backend</a></div> <div class="doc_text"> -<p>The llvm-gcc4 C front-end now generates debugging info for C and C++ for -X86/ELF platforms (Linux). This extends the PPC/Darwin and X86/Darwin debugging -support available in release 18.8 DWARF is a standard debugging format used on -many platforms.</p> +<p>LLVM 1.9 now fully supports the x86-64 instruction set on Mac OS/X, and +supports it on Linux (and other operating systems) when compiling in -static +mode. LLVM includes JIT support for X86-64, and supports both Intel EMT-64T +and AMD-64 architectures. The X86-64 instruction set permits addressing a +64-bit addressing space and provides the compiler with twice the +number of integer registers to use.</p> </div> <!--_________________________________________________________________________--> -<div class="doc_subsubsection"><a name="signedinst">Signed Instructions</a></div> +<div class="doc_subsubsection"><a name="lto">Link-Time Optimization integration +with native linkers</a></div> <div class="doc_text"> -<p>As a step towards making LLVM's integer types signless, several new -instructions have been added to LLVM. The DIV instruction has become UDIV, SDIV, -and FDIV. The REM instruction has become UREM, SREM and FREM. The SHR -instruction has become ASHR and LSHR. See the <a href="LangRef.html">Language - Reference</a> for details on these new instructions.</p> +<p>LLVM now includes <a href="LinkTimeOptimization.html">liblto</a> which can +be used to integrate LLVM Link-Time Optimization support into a native linker. +This allows LLVM .bc to transparently participate with linking an application, +even when some .o files are in LLVM form and some are not.</p> </div> <!--_________________________________________________________________________--> -<div class="doc_subsubsection"><a name="featureA">New Feature C</a></div> +<div class="doc_subsubsection"><a name="dwarf">DWARF debugging +support for Linux, Cygwin and MinGW on X86</a></div> <div class="doc_text"> -<p>Describe feature C here.</p> +<p>llvm-gcc4 now supports generating debugging info for Linux, Cygwin and MinGW. +This extends the PPC/Darwin and X86/Darwin debugging support available in the +1.8 release. DWARF is a standard debugging format used on many platforms.</p> </div> <!--_________________________________________________________________________--> -<div class="doc_subsubsection"><a name="featureB">New Feature D</a></div> +<div class="doc_subsubsection"><a name="optimizer">Optimizer +Improvements</a></div> <div class="doc_text"> -<p>Describe feature D here.</p> +<p>The mid-level optimizer is now faster and produces better code in many cases. + Significant changes include:</p> + +<ul> +<li>LLVM includes a new 'predicate simplifier' pass, which + currently performs dominator tree-based optimizations.</li> +<li>The complete loop unroll pass now supports unrolling of + multiple basic block loops.</li> +<li>The 'globalopt' pass can now perform the scalar replacement of + aggregates transformation on some heap allocations.</li> +<li>The globalsmodref-aa alias analysis can now track 'indirect pointer + globals' more accurately.</li> +<li>The instruction combiner can now perform element propagation +analysis of vector expressions, eliminating computation of vector elements +that are not used.</li> +</ul> + </div> <!--_________________________________________________________________________--> -<div class="doc_subsubsection"><a name="jitrelease">New Feature E</a></div> +<div class="doc_subsubsection"><a name="codegen">Code +Generator Enhancements</a></div> + <div class="doc_text"> -<p>Describe feature E here.</p> +<p> +The LLVM Target-Independent code generator now supports more target features and +optimizes many cases more aggressively. New features include: +</p> + +<ul> +<li>LLVM now includes a late branch folding pass which optimizes code + layout, performs several branch optzns, and deletes unreachable code.</li> +<li>The code generator now support targets that have pre/post-increment + addressing modes.</li> +<li>LLVM now supports dynamically-loadable register allocators and + schedulers.</li> +<li>LLVM 1.9 includes several improvements to inline asm support, + including support for new constraints and modifiers.</li> +<li>The register coalescer is now more aggressive than before, + allowing it to eliminate more copies.</li> +</ul> + +<p>In addition, the LLVM target description format has itself been extended in + several ways:</p> + +<ul> +<li>tblgen now allows definition of '<a + href="TableGenFundamentals.html#multiclass">multiclasses</a>' which can be + used to factor instruction patterns more aggressively in .td files.</li> +<li>LLVM has a new TargetAsmInfo class which captures a variety of + information about the target assembly language format.</li> +<li>.td files now support "<tt>${:foo}</tt>" syntax for encoding + subtarget-specific assembler syntax into instruction descriptions.</li> +</ul> + +<p>Further, several significant target-specific enhancements are included in +LLVM 1.9:</p> + +<ul> +<li>The LLVM ARM backend now supports more instructions + and the use of a frame pointer. It is now possible to build + libgcc and a simple cross compiler, but it is not considered "complete" yet. + </li> +<li>LLVM supports the Win32 dllimport/dllexport linkage and + stdcall/fastcall calling conventions.</li> +</ul> + </div> <!--_________________________________________________________________________--> @@ -121,38 +187,41 @@ instruction has become ASHR and LSHR. See the <a href="LangRef.html">Language <p>More specific changes include:</p> <ul> -<li>LLVM 1.8 includes an initial ARM backend. This backend is in early - development stages.</li> -<li>LLVM 1.8 now includes significantly better support for mingw and - cygwin.</li> -<li>The <a href="CommandGuide/html/llvm-config.html">llvm-config</a> tool is - now built by default and has several new features.</li> -<li>The X86 and PPC backends now use the correct platform ABI for passing - vectors as arguments to functions.</li> -<li>The X86 backend now includes support for the Microsoft ML assembler - ("MASM").</li> -<li>The PowerPC backend now pattern matches the 'rlwimi' instruction more - aggressively.</li> -<li>Most of LLVM is now built with "-pedantic", ensuring better portability - to more C++ Compilers.</li> -<li>The PowerPC backend now includes initial 64-bit support. The JIT is not - complete, and the static compiler has a couple of known bugs, but support - is mostly in place. LLVM 1.9 will include completed PPC-64 support. </li> - +<li>The llvm-test framework now supports SPEC2006.</li> +<li>LLVM now includes a <a href="GetElementPtr.html">FAQ about the +<tt>getelementptr</tt> instruction</a>.</li> +<li>Bugpoint now supports a new "<tt>-find-bugs</tt>" mode. This mode makes + bugpoint permute pass sequences to try to expose bugs due to pass + sequencing.</li> +<li>The JIT now supports lazily streaming code from multiple modules at a + time, implicitly linking the code as it goes.</li> </ul> </div> <!--=========================================================================--> <div class="doc_subsection"> -<a name="changes">Significant Changes in LLVM 1.8</a> +<a name="apichanges">Significant API Changes in LLVM 1.9</a> </div> <div class="doc_text"> + +<p>Several significant API changes have been made. If you are maintaining +out-of-tree code, please be aware that:</p> + <ul> -<li>The LLVM "SparcV9" backend (deprecated in LLVM 1.7) has been removed in -LLVM 1.8. The LLVM "Sparc" backend replaces it.</li> -<li>The --version option now prints more useful information, including the - build configuration for the tool.</li> +<li>The ConstantSInt and ConstantUInt classes have been merged into the + ConstantInt class.</li> +<li><p>As a step towards making LLVM's integer types signless, several new +instructions have been added to LLVM. The <tt>Div</tt> instruction is now +<tt>UDiv</tt>, <tt>SDiv</tt>, and <tt>FDiv</tt>. The <tt>Rem</tt> instruction +is now <tt>URem</tt>, <tt>SRem</tt> and <tt>FRem</tt>. See the +<a href="LangRef.html">Language Reference</a> for details on these new +instructions.</p> +<li><p><tt>ConstantBool::True</tt> and <tt>ConstantBool::False</tt> have been + renamed to <tt>ConstantBool::getTrue()</tt> and + <tt>ConstantBool::getFalse()</tt>.</p></li> +<li>The 'analyze' tool has been merged into the 'opt' tool.</li> + </ul> </div> @@ -174,7 +243,8 @@ LLVM 1.8. The LLVM "Sparc" backend replaces it.</li> <li>Sun UltraSPARC workstations running Solaris 8.</li> <li>Intel and AMD machines running on Win32 with the Cygwin libraries (limited support is available for native builds with Visual C++).</li> -<li>PowerPC and X86-based Mac OS X systems, running 10.2 and above.</li> +<li>PowerPC and X86-based Mac OS X systems, running 10.2 and above in 32-bit and + 64-bit modes.</li> <li>Alpha-based machines running Debian GNU/Linux.</li> <li>Itanium-based machines running Linux and HP-UX.</li> </ul> @@ -220,6 +290,7 @@ components, please contact us on the <a href="http://lists.cs.uiuc.edu/mailman/l <li>The <tt>-cee</tt> pass is known to be buggy, and may be removed in in a future release.</li> <li>The IA64 code generator is experimental.</li> +<li>The ARM code generator is experimental.</li> <li>The Alpha JIT is experimental.</li> <li>"<tt>-filetype=asm</tt>" (the default) is the only supported value for the <tt>-filetype</tt> llc option.</li> @@ -229,16 +300,138 @@ components, please contact us on the <a href="http://lists.cs.uiuc.edu/mailman/l <!-- ======================================================================= --> <div class="doc_subsection"> - <a name="build">Known problems with the Build System</a> + <a name="x86-be">Known problems with the X86 back-end</a> +</div> + +<div class="doc_text"> + +<ul> +<li>The X86 backend does not yet support <a href="http://llvm.org/PR879">inline + assembly that uses the X86 floating point stack</a>. See the <a + href="http://llvm.org/PR879">bug</a> for details on workarounds on + Linux.</li> +</ul> + +</div> + +<!-- ======================================================================= --> +<div class="doc_subsection"> + <a name="ppc-be">Known problems with the PowerPC back-end</a> +</div> + +<div class="doc_text"> + +<ul> +<li><a href="http://llvm.org/PR642">PowerPC backend does not correctly +implement ordered FP comparisons</a>.</li> +<li>The 64-bit PowerPC backend is not fully stable. If you desire PPC64 support, + please use mainline CVS LLVM, which has several important bug fixes.</li> +</ul> + +</div> + +<!-- ======================================================================= --> +<div class="doc_subsection"> + <a name="sparc-be">Known problems with the SPARC back-end</a> +</div> + +<div class="doc_text"> + +<ul> +<li>The SPARC backend only supports the 32-bit SPARC ABI (-m32), it does not + support the 64-bit SPARC ABI (-m64).</li> +</ul> + +</div> + +<!-- ======================================================================= --> +<div class="doc_subsection"> + <a name="c-be">Known problems with the C back-end</a> +</div> + +<div class="doc_text"> + +<ul> + +<li>The C back-end produces code that violates the ANSI C Type-Based Alias +Analysis rules. As such, special options may be necessary to compile the code +(for example, GCC requires the <tt>-fno-strict-aliasing</tt> option). This +problem probably cannot be fixed.</li> + +<li><a href="http://llvm.org/PR56">Zero arg vararg functions are not +supported</a>. This should not affect LLVM produced by the C or C++ +frontends.</li> + +<li>The C backend does not correctly implement the <a +href="LangRef.html#i_stacksave"><tt>llvm.stacksave</tt></a> or +<a href="LangRef.html#i_stackrestore"><tt>llvm.stackrestore</tt></a> +intrinsics. This means that some code compiled by it can run out of stack +space if they depend on these (e.g. C99 varargs).</li> + +<li><a href="http://llvm.org/PR802">The C backend does not support inline + assembly code</a>.</li> +</ul> + +</div> + +<!-- ======================================================================= --> +<div class="doc_subsection"> + <a name="alpha-be">Known problems with the Alpha back-end</a> </div> <div class="doc_text"> <ul> -<li>none yet</li> + +<li>On 21164s, some rare FP arithmetic sequences which may trap do not have the +appropriate nops inserted to ensure restartability.</li> + +</ul> +</div> + +<!-- ======================================================================= --> +<div class="doc_subsection"> + <a name="ia64-be">Known problems with the IA64 back-end</a> +</div> + +<div class="doc_text"> + +<ul> + +<li>C++ programs are likely to fail on IA64, as calls to <tt>setjmp</tt> are +made where the argument is not 16-byte aligned, as required on IA64. (Strictly +speaking this is not a bug in the IA64 back-end; it will also be encountered +when building C++ programs using the C back-end.)</li> + +<li>The C++ front-end does not use <a href="http://llvm.org/PR406">IA64 +ABI compliant layout of v-tables</a>. In particular, it just stores function +pointers instead of function descriptors in the vtable. This bug prevents +mixing C++ code compiled with LLVM with C++ objects compiled by other C++ +compilers.</li> + +<li>There are a few ABI violations which will lead to problems when mixing LLVM +output with code built with other compilers, particularly for floating-point +programs.</li> + +<li>Defining vararg functions is not supported (but calling them is ok).</li> + </ul> + </div> +<!-- ======================================================================= --> +<div class="doc_subsection"> + <a name="arm-be">Known problems with the ARM back-end</a> +</div> + +<div class="doc_text"> + +<ul> +<li>The ARM backend is currently in early development stages, it is not +ready for production use.</li> +</ul> + +</div> <!-- ======================================================================= --> <div class="doc_subsection"> @@ -264,29 +457,15 @@ components, please contact us on the <a href="http://lists.cs.uiuc.edu/mailman/l <div class="doc_text"> <p> -llvm-gcc3 has many significant problems that are fixed by llvm-gcc4. -Two major ones include:</p> - -<ul> -<li>With llvm-gcc3, - C99 variable sized arrays do not release stack memory when they go out of - scope. Thus, the following program may run out of stack space: -<pre> - for (i = 0; i != 1000000; ++i) { - int X[n]; - foo(X); - } -</pre></li> - -<li>With llvm-gcc3, Initialization of global union variables can only be done <a -href="http://llvm.org/PR162">with the largest union member</a>.</li> - -</ul> <p>llvm-gcc4 is far more stable and produces better code than llvm-gcc3, but -does not currently support Link-Time-Optimization or C++ Exception Handling, +does not currently support <a href="http://llvm.org/PR869">Link-Time +Optimization</a> or <a href="http://llvm.org/PR870">C++ Exception Handling</a>, which llvm-gcc3 does.</p> +<p>llvm-gcc4 does not support the <a href="http://llvm.org/PR947">GCC indirect +goto extension</a>, but llvm-gcc3 does.</p> + </div> <!-- _______________________________________________________________________ --> @@ -302,28 +481,12 @@ which llvm-gcc3 does.</p> support for floating point data types of any size other than 32 and 64 bits.</li> -<li>The following Unix system functionality has not been tested and may not -work: - <ol> - <li><tt>sigsetjmp</tt>, <tt>siglongjmp</tt> - These are not turned into the - appropriate <tt>invoke</tt>/<tt>unwind</tt> instructions. Note that - <tt>setjmp</tt> and <tt>longjmp</tt> <em>are</em> compiled correctly. - <li><tt>getcontext</tt>, <tt>setcontext</tt>, <tt>makecontext</tt> - - These functions have not been tested. - </ol></li> - <li>Although many GCC extensions are supported, some are not. In particular, the following extensions are known to <b>not be</b> supported: <ol> <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Local-Labels.html#Local%20Labels">Local Labels</a>: Labels local to a block.</li> <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.</li> <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Constructing-Calls.html#Constructing%20Calls">Constructing Calls</a>: Dispatching a call to another function.</li> - <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> - <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Constraints.html#Constraints">Constraints</a>: Constraints for asm operands.</li> - <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> - <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> - <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> - <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> <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Thread_002dLocal.html">Thread-Local</a>: Per-thread variables.</li> <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Pragmas.html#Pragmas">Pragmas</a>: Pragmas accepted by GCC.</li> </ol> @@ -344,29 +507,31 @@ work: Declaring that functions have no side effects or that they can never return.<br> - <b>Supported:</b> <tt>format</tt>, <tt>format_arg</tt>, <tt>non_null</tt>, - <tt>noreturn</tt>, <tt>constructor</tt>, <tt>destructor</tt>, - <tt>unused</tt>, <tt>used</tt>, - <tt>deprecated</tt>, <tt>warn_unused_result</tt>, <tt>weak</tt><br> + <b>Supported:</b> <tt>constructor</tt>, <tt>destructor</tt>, + <tt>deprecated</tt>, <tt>fastcall</tt>, <tt>format</tt>, + <tt>format_arg</tt>, <tt>non_null</tt>, <tt>noreturn</tt>, + <tt>stdcall</tt>, <tt>unused</tt>, <tt>used</tt>, + <tt>warn_unused_result</tt>, <tt>weak</tt><br> <b>Ignored:</b> <tt>noinline</tt>, <tt>always_inline</tt>, <tt>pure</tt>, <tt>const</tt>, <tt>nothrow</tt>, <tt>malloc</tt>, <tt>no_instrument_function</tt>, <tt>cdecl</tt><br> <b>Unsupported:</b> <tt>section</tt>, <tt>alias</tt>, - <tt>visibility</tt>, <tt>regparm</tt>, <tt>stdcall</tt>, - <tt>fastcall</tt>, all other target specific attributes</li> + <tt>visibility</tt>, <tt>regparm</tt>, all other target specific + attributes</li> <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Variable-Attributes.html#Variable%20Attributes">Variable Attributes</a>: Specifying attributes of variables.<br> <b>Supported:</b> <tt>cleanup</tt>, <tt>common</tt>, <tt>nocommon</tt>, - <tt>deprecated</tt>, <tt>transparent_union</tt>, + <tt>deprecated</tt>, <tt>dllimport</tt>, + <tt>dllexport</tt>, <tt>transparent_union</tt>, <tt>unused</tt>, <tt>used</tt>, <tt>weak</tt><br> <b>Unsupported:</b> <tt>aligned</tt>, <tt>mode</tt>, <tt>packed</tt>, <tt>section</tt>, <tt>shared</tt>, <tt>tls_model</tt>, - <tt>vector_size</tt>, <tt>dllimport</tt>, - <tt>dllexport</tt>, all target specific attributes.</li> + <tt>vector_size</tt>, all target specific attributes. + </li> <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Type-Attributes.html#Type%20Attributes">Type Attributes</a>: Specifying attributes of types.<br> <b>Supported:</b> <tt>transparent_union</tt>, <tt>unused</tt>, @@ -402,6 +567,12 @@ work: <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Empty-Structures.html#Empty%20Structures">Empty Structures</a>: Structures with no members.</li> <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> <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Escaped-Newlines.html#Escaped%20Newlines">Escaped Newlines</a>: Slightly looser rules for escaped newlines.</li> + <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> + <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Constraints.html#Constraints">Constraints</a>: Constraints for asm operands.</li> + <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> + <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> + <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> + <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> <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> <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> <li><a href="http://gcc.gnu.org/onlinedocs/gcc/Initializers.html#Initializers">Initializers</a>: Non-constant initializers.</li> @@ -446,26 +617,13 @@ itself.</p> </div> <!-- _______________________________________________________________________ --> -<div class="doc_subsubsection">Bugs</div> - -<div class="doc_text"> - -<ul> -<li>The C++ front-end inherits all problems afflicting the <a href="#c-fe">C - front-end</a>.</li> - -</ul> - -</div> - -<!-- _______________________________________________________________________ --> <div class="doc_subsubsection"> Notes </div> <div class="doc_text"> - <ul> +<li>llvm-gcc4 does not support C++ exception handling at all yet.</li> <li>Destructors for local objects are not always run when a <tt>longjmp</tt> is performed. In particular, destructors for objects in the <tt>longjmp</tt>ing @@ -480,7 +638,7 @@ itself.</p> representation issues. Because we use this API, code generated by the LLVM compilers should be binary compatible with machine code generated by other Itanium ABI C++ compilers (such as G++, the Intel and HP compilers, etc). - <i>However</i>, the exception handling mechanism used by LLVM is very + <i>However</i>, the exception handling mechanism used by llvm-gcc3 is very different from the model used in the Itanium ABI, so <b>exceptions will not interact correctly</b>. </li> @@ -488,134 +646,7 @@ itself.</p> </div> -<!-- ======================================================================= --> -<div class="doc_subsection"> - <a name="c-be">Known problems with the C back-end</a> -</div> - -<div class="doc_text"> - -<ul> - -<li>The C back-end produces code that violates the ANSI C Type-Based Alias -Analysis rules. As such, special options may be necessary to compile the code -(for example, GCC requires the <tt>-fno-strict-aliasing</tt> option). This -problem probably cannot be fixed.</li> - -<li><a href="http://llvm.org/PR56">Zero arg vararg functions are not -supported</a>. This should not affect LLVM produced by the C or C++ -frontends.</li> - -<li>The C backend does not correctly implement the <a -href="LangRef.html#i_stacksave"><tt>llvm.stacksave</tt></a> or -<a href="LangRef.html#i_stackrestore"><tt>llvm.stackrestore</tt></a> -intrinsics. This means that some code compiled by it can run out of stack -space if they depend on these (e.g. C99 varargs).</li> - -</ul> - -</div> - -<!-- ======================================================================= --> -<div class="doc_subsection"> - <a name="x86-be">Known problems with the X86 back-end</a> -</div> - -<div class="doc_text"> - -<ul> -<li>none yet.</li> -</ul> - -</div> - -<!-- ======================================================================= --> -<div class="doc_subsection"> - <a name="ppc-be">Known problems with the PowerPC back-end</a> -</div> - -<div class="doc_text"> - -<ul> -<li><a href="http://llvm.org/PR642">PowerPC backend does not correctly -implement ordered FP comparisons</a>.</li> -</ul> - -</div> - -<!-- ======================================================================= --> -<div class="doc_subsection"> - <a name="alpha-be">Known problems with the Alpha back-end</a> -</div> - -<div class="doc_text"> - -<ul> - -<li>On 21164s, some rare FP arithmetic sequences which may trap do not have the -appropriate nops inserted to ensure restartability.</li> - -</ul> - -</div> - -<!-- ======================================================================= --> -<div class="doc_subsection"> - <a name="ia64-be">Known problems with the IA64 back-end</a> -</div> - -<div class="doc_text"> - -<ul> - -<li>C++ programs are likely to fail on IA64, as calls to <tt>setjmp</tt> are -made where the argument is not 16-byte aligned, as required on IA64. (Strictly -speaking this is not a bug in the IA64 back-end; it will also be encountered -when building C++ programs using the C back-end.)</li> - -<li>The C++ front-end does not use <a href="http://llvm.org/PR406">IA64 -ABI compliant layout of v-tables</a>. In particular, it just stores function -pointers instead of function descriptors in the vtable. This bug prevents -mixing C++ code compiled with LLVM with C++ objects compiled by other C++ -compilers.</li> - -<li>There are a few ABI violations which will lead to problems when mixing LLVM -output with code built with other compilers, particularly for floating-point -programs.</li> - -<li>Defining vararg functions is not supported (but calling them is ok).</li> - -</ul> - -</div> - -<!-- ======================================================================= --> -<div class="doc_subsection"> - <a name="sparc-be">Known problems with the SPARC back-end</a> -</div> -<div class="doc_text"> - -<ul> -<li>The SPARC backend only supports the 32-bit SPARC ABI (-m32), it does not - support the 64-bit SPARC ABI (-m64).</li> -</ul> - -</div> - -<!-- ======================================================================= --> -<div class="doc_subsection"> - <a name="arm-be">Known problems with the ARM back-end</a> -</div> - -<div class="doc_text"> - -<ul> -<li>The ARM backend is currently in early development stages, it is not -ready for production use.</li> -</ul> - -</div> <!-- *********************************************************************** --> <div class="doc_section"> |