<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40">
<head>
<meta http-equiv=Content-Type content="text/html; charset=us-ascii">
<meta name=Generator content="Microsoft Word 12 (filtered medium)">
<style>
<!--
/* Font Definitions */
@font-face
{font-family:"Cambria Math";
panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
{font-family:Calibri;
panose-1:2 15 5 2 2 2 4 3 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
{margin:0cm;
margin-bottom:.0001pt;
font-size:11.0pt;
font-family:"Calibri","sans-serif";}
a:link, span.MsoHyperlink
{mso-style-priority:99;
color:blue;
text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
{mso-style-priority:99;
color:purple;
text-decoration:underline;}
span.EmailStyle17
{mso-style-type:personal-compose;
font-family:"Calibri","sans-serif";
color:windowtext;}
.MsoChpDefault
{mso-style-type:export-only;}
@page Section1
{size:612.0pt 792.0pt;
margin:72.0pt 72.0pt 72.0pt 72.0pt;}
div.Section1
{page:Section1;}
-->
</style>
<!--[if gte mso 9]><xml>
<o:shapedefaults v:ext="edit" spidmax="1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext="edit">
<o:idmap v:ext="edit" data="1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang=EN-US link=blue vlink=purple>
<div class=Section1>
<p class=MsoNormal>Hi all,<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>I think I might have found a potential bug when using SSE intrinsic
and unaligned memory. Here’s the code to reproduce it:<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New";color:blue'>#include</span><span style='font-size:
10.0pt;font-family:"Courier New"'> <span style='color:#A31515'>"llvm/Module.h"<o:p></o:p></span></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New";color:blue'>#include</span><span style='font-size:
10.0pt;font-family:"Courier New"'> <span style='color:#A31515'>"llvm/Intrinsics.h"<o:p></o:p></span></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New";color:blue'>#include</span><span style='font-size:
10.0pt;font-family:"Courier New"'> <span style='color:#A31515'>"llvm/Instructions.h"<o:p></o:p></span></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New";color:blue'>#include</span><span style='font-size:
10.0pt;font-family:"Courier New"'> <span style='color:#A31515'>"llvm/ModuleProvider.h"<o:p></o:p></span></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New";color:blue'>#include</span><span style='font-size:
10.0pt;font-family:"Courier New"'> <span style='color:#A31515'>"llvm/ExecutionEngine/JIT.h"<o:p></o:p></span></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New";color:blue'>#include</span><span style='font-size:
10.0pt;font-family:"Courier New"'> <span style='color:#A31515'>"llvm/Support/LLVMBuilder.h"<o:p></o:p></span></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New";color:blue'>using</span><span style='font-size:10.0pt;
font-family:"Courier New"'> <span style='color:blue'>namespace</span> llvm;<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New";color:blue'>static</span><span style='font-size:10.0pt;
font-family:"Courier New"'> Function *createFunction(Module *module)<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>{<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> VectorType
*float4Type = VectorType::get(Type::FloatTy, 4);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> PointerType
*float4PointerType = PointerType::get(float4Type, 0);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> std::vector<<span
style='color:blue'>const</span> Type*> functionArguments;<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> functionArguments.push_back(float4PointerType);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> functionArguments.push_back(float4PointerType);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> FunctionType
*functionType = FunctionType::get(Type::VoidTy, functionArguments, <span
style='color:blue'>false</span>);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> Function *function = <span
style='color:blue'>new</span> Function(functionType,
GlobalValue::InternalLinkage, <span style='color:#A31515'>""</span>,
module);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> Function::arg_iterator
args = function->arg_begin();<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> BasicBlock
*entryBlock = <span style='color:blue'>new</span> BasicBlock(<span
style='color:#A31515'>""</span>, function, 0);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> LLVMBuilder *builder
= <span style='color:blue'>new</span> LLVMBuilder(entryBlock);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> {<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> Value
*out_arg = args++;<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> Value
*in_arg = args++;<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> Function
*rcpps = Intrinsic::getDeclaration(module, Intrinsic::x86_sse_rcp_ps);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> <o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> Value
*in = builder->Insert(<span style='color:blue'>new</span> LoadInst(in_arg, <span
style='color:#A31515'>""</span>, <span style='color:blue'>false</span>,
1));<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> Value
*out = builder->CreateCall(rcpps, in);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> builder->Insert(<span
style='color:blue'>new</span> StoreInst(out, out_arg, <span style='color:blue'>false</span>,
1));<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> builder->CreateRetVoid();<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> }<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> <span
style='color:blue'>delete</span> builder;<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> <span
style='color:blue'>return</span> function;<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>}<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New";color:blue'>int</span><span style='font-size:10.0pt;
font-family:"Courier New"'> main(<span style='color:blue'>int</span> argc, <span
style='color:blue'>char</span> **argv)<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>{<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> Module *module = <span
style='color:blue'>new</span> Module(<span style='color:#A31515'>"test"</span>);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> Function *function =
createFunction(module);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> ExistingModuleProvider
*moduleProvider = <span style='color:blue'>new</span>
ExistingModuleProvider(module);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> ExecutionEngine
*executionEngine = ExecutionEngine::create(moduleProvider, <span
style='color:blue'>false</span>);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> <span
style='color:blue'>float</span> <span style='color:green'>/*__declspec(align(16))*/</span>
in[4] = {2, 2, 2, 2};<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> <span
style='color:blue'>float</span> <span style='color:green'>/*__declspec(align(16))*/</span>
out[4] = {1, 1, 1, 1};<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> <span
style='color:blue'>void</span> (*func)(<span style='color:blue'>float</span>*,<span
style='color:blue'>float</span>*) = (<span style='color:blue'>void</span>(*)(<span
style='color:blue'>float</span>*,<span style='color:blue'>float</span>*))executionEngine->getPointerToFunction(function);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> func(out, in);<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> <span
style='color:blue'>delete</span> executionEngine;<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'><o:p> </o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'> <span
style='color:blue'>return</span> 0;<o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:10.0pt;
font-family:"Courier New"'>}<o:p></o:p></span></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>It generates the following assembly code:<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:8.0pt;
font-family:"Courier New";color:gray'>mov
eax,dword ptr [esp+8] <o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:8.0pt;
font-family:"Courier New";color:gray'>rcpps
xmm0,xmmword ptr [eax] <o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:8.0pt;
font-family:"Courier New";color:gray'>mov
eax,dword ptr [esp+4] <o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:8.0pt;
font-family:"Courier New";color:gray'>movups
xmmword ptr [eax],xmm0 <o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:8.0pt;font-family:"Courier New";
color:gray'>ret</span><o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>Note that even though the LoadInst is specified to have an
alignment of 1 (in fact no alignment), the rcpps tries to reference the memory
directly, but it expects aligned memory. If “in” happens to not be
16-byte aligned, an exception will be thrown. What really should have been
generated is something like this:<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:8.0pt;
font-family:"Courier New";color:gray'>mov
eax,dword ptr [esp+8] <o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:8.0pt;
font-family:"Courier New";color:gray'>movups
xmm0,xmmword ptr [eax] <o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:8.0pt;
font-family:"Courier New";color:gray'>rcpps
xmm1,xmmword ptr [eax] <o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:8.0pt;
font-family:"Courier New";color:gray'>mov
eax,dword ptr [esp+4] <o:p></o:p></span></p>
<p class=MsoNormal style='text-autospace:none'><span style='font-size:8.0pt;
font-family:"Courier New";color:gray'>movups
xmmword ptr [eax],xmm1 <o:p></o:p></span></p>
<p class=MsoNormal><span style='font-size:8.0pt;font-family:"Courier New";
color:gray'>ret</span><o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>Since I’m fairly new to LLVM I’m not entirely
sure if this is really a bug or something I’m not doing correctly, or
whether it’s already being addressed. The following thread appears to
talk about something similar: <a
href="http://thread.gmane.org/gmane.comp.compilers.llvm.devel/9476/focus=9478">http://thread.gmane.org/gmane.comp.compilers.llvm.devel/9476/focus=9478</a><o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>Thank you,<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
<p class=MsoNormal>Nicolas Capens<o:p></o:p></p>
</div>
</body>
</html>