<html><head><meta http-equiv="Content-Type" content="text/html charset=windows-1252"><base href="x-msg://638/"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">You can also use inline assembly to force spilling. See test/CodeGen/Thumb2/aligned-spill.ll <div><br></div><div>/jakob</div><div><br><div><div>On Mar 28, 2013, at 9:20 AM, "Gurd, Preston" <<a href="mailto:preston.gurd@intel.com">preston.gurd@intel.com</a>> wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><div lang="EN-US" link="blue" vlink="purple" style="font-family: Optima; font-size: medium; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: 2; text-align: -webkit-auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px; "><div class="WordSection1" style="page: WordSection1; "><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125); ">Thanks, Nadav. We will try to do what you suggest.<o:p></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125); "> </span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125); ">Preston<o:p></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125); "> </span></div><div><div style="border-style: solid none none; border-top-width: 1pt; border-top-color: rgb(181, 196, 223); padding: 3pt 0in 0in; "><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><b><span style="font-size: 10pt; font-family: Tahoma, sans-serif; ">From:</span></b><span style="font-size: 10pt; font-family: Tahoma, sans-serif; "><span class="Apple-converted-space"> </span>Nadav Rotem [mailto:nrotem@<a href="http://apple.com">apple.com</a>]<span class="Apple-converted-space"> </span><br><b>Sent:</b><span class="Apple-converted-space"> </span>Thursday, March 28, 2013 11:54 AM<br><b>To:</b><span class="Apple-converted-space"> </span>Gurd, Preston<br><b>Cc:</b><span class="Apple-converted-space"> </span><a href="mailto:llvm-commits@cs.uiuc.edu">llvm-commits@cs.uiuc.edu</a><br><b>Subject:</b><span class="Apple-converted-space"> </span>Re: [llvm] r178193 - This patch follows is a follow up to r178171, which uses the register<o:p></o:p></span></div></div></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><o:p> </o:p></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; ">Hi Preston,<o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><o:p> </o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; ">I appreciate your efforts to try to reduce the test case, but the test cases are still large. There are a few things that you can do to make them smaller. First, you can get rid of the complex types. A large part of the test cases are types and most of the type information is not used. You can replace some of the members in these structs with primitive types and this will allow you to delete some of the type definitions. Next, you can generate a new function call (to an external function) with lots of argument, this will ensure that all of the registers that are used to pass arguments to be spilled. If that does not work, you can create a small loop that uses variables that are declared externally, this will also make the variables to be live across the loop and your pointer will be spilled. Finally, you can insert assert(0) in the code that spills registers and use bugpoint to reduce the test case for you. This will allow you to create the smallest possible code that spills. As you can see there are many ways to reduce the test case. The big test cases are fragile and are not very helpful. The next time we would like to make major changes in the register allocator we would have to spend our time on fixing test cases that rely on specific heuristics of the current register allocator. It should not be too hard to come up with a synthetic test case that spills the pointer. Please reduce these test cases. <o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><o:p> </o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; ">Thanks,<o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; ">Nadav <o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><o:p> </o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><o:p> </o:p></div></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><o:p> </o:p></div><div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; ">On Mar 28, 2013, at 7:21 AM, "Gurd, Preston" <<a href="mailto:preston.gurd@intel.com" style="color: purple; text-decoration: underline; ">preston.gurd@intel.com</a>> wrote:<o:p></o:p></div></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><br><br><o:p></o:p></div><div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125); ">Hello Nadav,</span><o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125); "> </span><o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125); ">I agree that they are excessively large, but I submit that it can’t be helped. We did our best to minimize the code size, but the tests had to be complicated enough to force the spill and reload of an indirect call address.</span><o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125); "> </span><o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125); ">Preston</span><o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><span style="font-size: 11pt; font-family: Calibri, sans-serif; color: rgb(31, 73, 125); "> </span><o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><b><span style="font-size: 10pt; font-family: Tahoma, sans-serif; ">From:</span></b><span class="apple-converted-space"><span style="font-size: 10pt; font-family: Tahoma, sans-serif; "> </span></span><span style="font-size: 10pt; font-family: Tahoma, sans-serif; ">Nadav Rotem [<a href="mailto:nrotem@apple.com" style="color: purple; text-decoration: underline; ">mailto:nrotem@apple.com</a>]<span class="apple-converted-space"> </span><br><b>Sent:</b><span class="apple-converted-space"> </span>Wednesday, March 27, 2013 8:08 PM<br><b>To:</b><span class="apple-converted-space"> </span>Gurd, Preston;<span class="Apple-converted-space"> </span><a href="mailto:llvm-commits@cs.uiuc.edu" style="color: purple; text-decoration: underline; ">llvm-commits@cs.uiuc.edu</a><br><b>Subject:</b><span class="apple-converted-space"> </span>Re: [llvm] r178193 - This patch follows is a follow up to r178171, which uses the register</span><o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "> <o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; ">Hi Preston, <o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "> <o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; ">The test cases are really big. Is there a way to make them smaller and still be able to test the reloading optimization ?<o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "> <o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; ">Thanks,<o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; ">Nadav<o:p></o:p></div></div><div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><br><br>On 03/27/13,<span class="apple-converted-space"> </span><b>Preston Gurd<span class="apple-converted-space"> </span></b><<a href="mailto:preston.gurd@intel.com" style="color: purple; text-decoration: underline; "><span style="color: purple; ">preston.gurd@intel.com</span></a>> wrote:<o:p></o:p></div></div><div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; ">Author: pgurd<br>Date: Wed Mar 27 18:16:18 2013<br>New Revision: 178193<br><br>URL:<span class="apple-converted-space"> </span><a href="http://llvm.org/viewvc/llvm-project?rev=178193&view=rev" target="l" style="color: purple; text-decoration: underline; "><span style="color: purple; ">http://llvm.org/viewvc/llvm-project?rev=178193&view=rev</span></a><br>Log:<br>This patch follows is a follow up to r178171, which uses the register<span class="apple-converted-space"> </span><br>form of call in preference to memory indirect on Atom.<br><br>In this case, the patch applies the optimization to the code for reloading<br>spilled registers.<br><br>The patch also includes changes to sibcall.ll and movgs.ll, which were<br>failing on the Atom buildbot after the first patch was applied.<br><br>This patch by Sriram Murali.<br><br><br>Added:<br> llvm/trunk/test/CodeGen/X86/atom-call-reg-indirect-foldedreload32.ll<br> llvm/trunk/test/CodeGen/X86/atom-call-reg-indirect-foldedreload64.ll<br>Modified:<br> llvm/trunk/lib/Target/X86/X86InstrInfo.cpp<br> llvm/trunk/test/CodeGen/X86/movgs.ll<br> llvm/trunk/test/CodeGen/X86/sibcall.ll<br><br>Modified: llvm/trunk/lib/Target/X86/X86InstrInfo.cpp<br>URL:<span class="apple-converted-space"> </span><a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86InstrInfo.cpp?rev=178193&r1=178192&r2=178193&view=diff" target="l" style="color: purple; text-decoration: underline; "><span style="color: purple; ">http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86InstrInfo.cpp?rev=178193&r1=178192&r2=178193&view=diff</span></a><br>==============================================================================<br>--- llvm/trunk/lib/Target/X86/X86InstrInfo.cpp (original)<br>+++ llvm/trunk/lib/Target/X86/X86InstrInfo.cpp Wed Mar 27 18:16:18 2013<br>@@ -3655,7 +3655,16 @@ X86InstrInfo::foldMemoryOperandImpl(Mach<br> const SmallVectorImpl<MachineOperand> &MOs,<br> unsigned Size, unsigned Align) const {<br> const DenseMap<unsigned, std::pair<unsigned,unsigned> > *OpcodeTablePtr = 0;<br>+ bool isCallRegIndirect = TM.getSubtarget<X86Subtarget>().callRegIndirect();<br> bool isTwoAddrFold = false;<br>+<br>+ // Atom favors register form of call. So, we do not fold loads into calls<br>+ // when X86Subtarget is Atom.<br>+ if (isCallRegIndirect &&<br>+ (MI->getOpcode() == X86::CALL32r || MI->getOpcode() == X86::CALL64r)) {<br>+ return NULL;<br>+ }<br>+<br> unsigned NumOps = MI->getDesc().getNumOperands();<br> bool isTwoAddr = NumOps > 1 &&<br> MI->getDesc().getOperandConstraint(1, MCOI::TIED_TO) != -1;<br><br>Added: llvm/trunk/test/CodeGen/X86/atom-call-reg-indirect-foldedreload32.ll<br>URL:<span class="apple-converted-space"> </span><a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/atom-call-reg-indirect-foldedreload32.ll?rev=178193&view=auto" target="l" style="color: purple; text-decoration: underline; "><span style="color: purple; ">http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/atom-call-reg-indirect-foldedreload32.ll?rev=178193&view=auto</span></a><br>==============================================================================<br>--- llvm/trunk/test/CodeGen/X86/atom-call-reg-indirect-foldedreload32.ll (added)<br>+++ llvm/trunk/test/CodeGen/X86/atom-call-reg-indirect-foldedreload32.ll Wed Mar 27 18:16:18 2013<br>@@ -0,0 +1,204 @@<br>+; RUN: llc < %s -mtriple=i386-linux-gnu -mcpu=atom 2>&1 | \<br>+; RUN: grep "calll" | not grep "("<br>+; RUN: llc < %s -mtriple=i386-linux-gnu -mcpu=core2 2>&1 | \<br>+; RUN: grep "calll" | grep "4-byte Folded Reload"<br>+<br>+%struct.targettype = type {i32}<br>+%struct.op_ptr1 = type opaque<br>+%struct.op_ptr2 = type opaque<br>+%union.anon = type { [8 x i32], [48 x i8] }<br>+%struct.const1 = type { [64 x i16], i8 }<br>+%struct.const2 = type { [17 x i8], [256 x i8], i8 }<br>+<br>+%struct.ref1 = type { void (%struct.ref2*)*, i32 (%struct.ref2*)*, void (%struct.ref2*)*, i32 (%struct.ref2*, i8***)*, %struct.op_ptr2** }<br>+%struct.ref2 = type { %struct.localref13*, %struct.localref15*, %struct.localref12*, i8*, i8, i32, %struct.localref11*, i32, i32, i32, i32, i32, i32, i32, double, i8, i8, i32, i8, i8, i8, i32, i8, i32, i8, i8, i8, i32, i32, i32, i32, i32, i32, i8**, i32, i32, i32, i32, i32, [64 x i32]*, [4 x %struct.const1*], [4 x %struct.const2*], [4 x %struct.const2*], i32, %struct.ref3*, i8, i8, [16 x i8], [16 x i8], [16 x i8], i32, i8, i8, i8, i8, i16, i16, i8, i8, i8, %struct.localref10*, i32, i32, i32, i32, i8*, i32, [4 x %struct.ref3*], i32, i32, i32, [10 x i32], i32, i32, i32, i32, i32, %struct.localref8*, %struct.localref9*, %struct.ref1*, %struct.localref7*, %struct.localref6*, %struct.localref5*, %struct.localref1*, %struct.ref4*, %struct.localref2*, %struct.localref3*, %struct.localref4* }<br>+%struct.ref3 = type { i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i8, i32, i32, i32, i32, i32, i32, %struct.const1*, i8* }<br>+%struct.ref4 = type { void (%struct.ref2*)*, [5 x void (%struct.ref2*, %struct.ref3*, i16*, i8**, i32)*] }<br>+<br>+%struct.localref1 = type { void (%struct.ref2*)*, i8 (%struct.ref2*, [64 x i16]**)*, i8 }<br>+%struct.localref2 = type { void (%struct.ref2*)*, void (%struct.ref2*, i8***, i32*, i32, i8**, i32*, i32)*, i8 }<br>+%struct.localref3 = type { void (%struct.ref2*)*, void (%struct.ref2*, i8***, i32, i8**, i32)* }<br>+%struct.localref4 = type { {}*, void (%struct.ref2*, i8**, i8**, i32)*, void (%struct.ref2*)*, void (%struct.ref2*)* }<br>+%struct.localref5 = type { void (%struct.ref2*)*, i32 (%struct.ref2*)*, i8 (%struct.ref2*)*, i8, i8, i32, i32 }<br>+%struct.localref6 = type { i32 (%struct.ref2*)*, void (%struct.ref2*)*, void (%struct.ref2*)*, void (%struct.ref2*)*, i8, i8 }<br>+%struct.localref7 = type { void (%struct.ref2*, i32)*, void (%struct.ref2*, i8***, i32*, i32, i8**, i32*, i32)* }<br>+%struct.localref8 = type { void (%struct.ref2*)*, void (%struct.ref2*)*, i8 }<br>+%struct.localref9 = type { void (%struct.ref2*, i32)*, void (%struct.ref2*, i8**, i32*, i32)* }<br>+%struct.localref10 = type { %struct.localref10*, i8, i32, i32, i8* }<br>+%struct.localref11 = type { i8*, %struct.targettype, void (%struct.ref2*)*, i8 (%struct.ref2*)*, void (%struct.ref2*, %struct.targettype)*, i8 (%struct.ref2*, i32)*, void (%struct.ref2*)* }<br>+%struct.localref12 = type { {}*, %struct.targettype, %struct.targettype, i32, i32 }<br>+%struct.localref13 = type { void (%struct.localref14*)*, void (%struct.localref14*, i32)*, void (%struct.localref14*)*, void (%struct.localref14*, i8*)*, void (%struct.localref14*)*, i32, %union.anon, i32, %struct.targettype, i8**, i32, i8**, i32, i32 }<br>+%struct.localref14 = type { %struct.localref13*, %struct.localref15*, %struct.localref12*, i8*, i8, i32 }<br>+%struct.localref15 = type { i8* (%struct.localref14*, i32, %struct.targettype)*, i8* (%struct.localref14*, i32, %struct.targettype)*, i8** (%struct.localref14*, i32, i32, i32)*, [64 x i16]** (%struct.localref14*, i32, i32, i32)*, %struct.op_ptr1* (%struct.localref14*, i32, i8, i32, i32, i32)*, %struct.op_ptr2* (%struct.localref14*, i32, i8, i32, i32, i32)*, {}*, i8** (%struct.localref14*, %struct.op_ptr1*, i32, i32, i8)*, [64 x i16]** (%struct.localref14*, %struct.op_ptr2*, i32, i32, i8)*, void (%struct.localref14*, i32)*, {}*, %struct.targettype, %struct.targettype}<br>+<br>+define internal i32 @foldedreload(%struct.ref2* %cinfo, i8*** nocapture %output1) {<br>+ %1 = getelementptr inbounds %struct.ref2* %cinfo, i32 0, i32 79<br>+ %2 = load %struct.ref1** %1, align 4<br>+ %3 = getelementptr inbounds %struct.ref2* %cinfo, i32 0, i32 68<br>+ %4 = load i32* %3, align 4<br>+ %5 = add i32 %4, -1<br>+ %6 = getelementptr inbounds %struct.ref2* %cinfo, i32 0, i32 64<br>+ %7 = load i32* %6, align 4<br>+ %8 = add i32 %7, -1<br>+ %9 = getelementptr inbounds %struct.ref1* %2, i32 1, i32 1<br>+ %10 = bitcast i32 (%struct.ref2*)** %9 to i32*<br>+ %11 = load i32* %10, align 4<br>+ %12 = getelementptr inbounds %struct.ref1* %2, i32 1, i32 2<br>+ %13 = bitcast void (%struct.ref2*)** %12 to i32*<br>+ %14 = load i32* %13, align 4<br>+ %15 = icmp slt i32 %11, %14<br>+ br i1 %15, label %.lr.ph18, label %._crit_edge19<br>+<br>+.lr.ph18:<br>+ %16 = getelementptr inbounds %struct.ref1* %2, i32 1<br>+ %17 = bitcast %struct.ref1* %16 to i32*<br>+ %18 = getelementptr inbounds %struct.ref1* %16, i32 0, i32 0<br>+ %19 = getelementptr inbounds %struct.ref2* %cinfo, i32 0, i32 66<br>+ %20 = getelementptr inbounds %struct.ref2* %cinfo, i32 0, i32 84<br>+ %21 = getelementptr inbounds %struct.ref2* %cinfo, i32 0, i32 36<br>+ %22 = getelementptr inbounds %struct.ref1* %2, i32 1, i32 3<br>+ %23 = bitcast i32 (%struct.ref2*, i8***)** %22 to [10 x [64 x i16]*]*<br>+ %.pre = load i32* %17, align 4<br>+ br label %24<br>+<br>+; <label>:24<br>+ %25 = phi i32 [ %14, %.lr.ph18 ], [ %89, %88 ]<br>+ %26 = phi i32 [ %.pre, %.lr.ph18 ], [ 0, %88 ]<br>+ %var1.015 = phi i32 [ %11, %.lr.ph18 ], [ %90, %88 ]<br>+ %27 = icmp ugt i32 %26, %5<br>+ br i1 %27, label %88, label %.preheader7.lr.ph<br>+<br>+.preheader7.lr.ph:<br>+ %.pre24 = load i32* %19, align 4<br>+ br label %.preheader7<br>+<br>+.preheader7:<br>+ %28 = phi i32 [ %.pre24, %.preheader7.lr.ph ], [ %85, %._crit_edge11 ]<br>+ %var2.012 = phi i32 [ %26, %.preheader7.lr.ph ], [ %86, %._crit_edge11 ]<br>+ %29 = icmp sgt i32 %28, 0<br>+ br i1 %29, label %.lr.ph10, label %._crit_edge11<br>+<br>+.lr.ph10:<br>+ %30 = phi i32 [ %28, %.preheader7 ], [ %82, %81 ]<br>+ %var4.09 = phi i32 [ 0, %.preheader7 ], [ %var4.1.lcssa, %81 ]<br>+ %ci.08 = phi i32 [ 0, %.preheader7 ], [ %83, %81 ]<br>+ %31 = getelementptr inbounds %struct.ref2* %cinfo, i32 0, i32 67, i32 %ci.08<br>+ %32 = load %struct.ref3** %31, align 4<br>+ %33 = getelementptr inbounds %struct.ref3* %32, i32 0, i32 1<br>+ %34 = load i32* %33, align 4<br>+ %35 = load %struct.ref4** %20, align 4<br>+ %36 = getelementptr inbounds %struct.ref4* %35, i32 0, i32 1, i32 %34<br>+ %37 = load void (%struct.ref2*, %struct.ref3*, i16*, i8**, i32)** %36, align 4<br>+ %38 = getelementptr inbounds %struct.ref3* %32, i32 0, i32 17<br>+ %39 = load i32* %38, align 4<br>+ %40 = getelementptr inbounds %struct.ref3* %32, i32 0, i32 9<br>+ %41 = load i32* %40, align 4<br>+ %42 = getelementptr inbounds %struct.ref3* %32, i32 0, i32 16<br>+ %43 = load i32* %42, align 4<br>+ %44 = mul i32 %43, %var2.012<br>+ %45 = getelementptr inbounds %struct.ref3* %32, i32 0, i32 14<br>+ %46 = load i32* %45, align 4<br>+ %47 = icmp sgt i32 %46, 0<br>+ br i1 %47, label %.lr.ph6, label %81<br>+<br>+.lr.ph6:<br>+ %48 = getelementptr inbounds i8*** %output1, i32 %34<br>+ %49 = mul nsw i32 %41, %var1.015<br>+ %50 = load i8*** %48, align 4<br>+ %51 = getelementptr inbounds i8** %50, i32 %49<br>+ %52 = getelementptr inbounds %struct.ref3* %32, i32 0, i32 13<br>+ %53 = getelementptr inbounds %struct.ref3* %32, i32 0, i32 18<br>+ %54 = icmp sgt i32 %39, 0<br>+ br i1 %54, label %.<a href="http://lr.ph6.split.us" style="color: purple; text-decoration: underline; ">lr.ph6.split.us</a>, label %.lr.ph6..lr.ph6.split_crit_edge<br>+<br>+.lr.ph6..lr.ph6.split_crit_edge:<br>+ br label %._crit_edge26<br>+<br>+.<a href="http://lr.ph6.split.us" style="color: purple; text-decoration: underline; ">lr.ph6.split.us</a>:<br>+ %55 = phi i32 [ %63, %._crit_edge28 ], [ %46, %.lr.ph6 ]<br>+ %56 = phi i32 [ %64, %._crit_edge28 ], [ %41, %.lr.ph6 ]<br>+ %<a href="http://var4.15.us" style="color: purple; text-decoration: underline; ">var4.15.us</a><span class="Apple-converted-space"> </span>= phi i32 [ %66, %._crit_edge28 ], [ %var4.09, %.lr.ph6 ]<br>+ %<a href="http://output2.04.us" style="color: purple; text-decoration: underline; ">output2.04.us</a><span class="Apple-converted-space"> </span>= phi i8** [ %69, %._crit_edge28 ], [ %51, %.lr.ph6 ]<br>+ %<a href="http://var5.03.us" style="color: purple; text-decoration: underline; ">var5.03.us</a><span class="Apple-converted-space"> </span>= phi i32 [ %67, %._crit_edge28 ], [ 0, %.lr.ph6 ]<br>+ %57 = load i32* %21, align 4<br>+ %58 = icmp ult i32 %57, %8<br>+ br i1 %58, label %.<a href="http://lr.ph.us" style="color: purple; text-decoration: underline; ">lr.ph.us</a>, label %59<br>+<br>+; <label>:59<br>+ %60 = add nsw i32 %<a href="http://var5.03.us" style="color: purple; text-decoration: underline; ">var5.03.us</a>, %var1.015<br>+ %61 = load i32* %53, align 4<br>+ %62 = icmp slt i32 %60, %61<br>+ br i1 %62, label %.<a href="http://lr.ph.us" style="color: purple; text-decoration: underline; ">lr.ph.us</a>, label %._crit_edge27<br>+<br>+._crit_edge27:<br>+ %63 = phi i32 [ %.pre23.pre, %.<a href="http://loopexit.us" style="color: purple; text-decoration: underline; ">loopexit.us</a><span class="Apple-converted-space"> </span>], [ %55, %59 ]<br>+ %64 = phi i32 [ %74, %.<a href="http://loopexit.us" style="color: purple; text-decoration: underline; ">loopexit.us</a><span class="Apple-converted-space"> </span>], [ %56, %59 ]<br>+ %65 = load i32* %52, align 4<br>+ %66 = add nsw i32 %65, %<a href="http://var4.15.us" style="color: purple; text-decoration: underline; ">var4.15.us</a><br>+ %67 = add nsw i32 %<a href="http://var5.03.us" style="color: purple; text-decoration: underline; ">var5.03.us</a>, 1<br>+ %68 = icmp slt i32 %67, %63<br>+ br i1 %68, label %._crit_edge28, label %._crit_edge<br>+<br>+._crit_edge28:<br>+ %69 = getelementptr inbounds i8** %<a href="http://output2.04.us" style="color: purple; text-decoration: underline; ">output2.04.us</a>, i32 %64<br>+ br label %.<a href="http://lr.ph6.split.us" style="color: purple; text-decoration: underline; ">lr.ph6.split.us</a><br>+<br>+.<a href="http://lr.ph.us" style="color: purple; text-decoration: underline; ">lr.ph.us</a>:<br>+ %<a href="http://var3.02.us" style="color: purple; text-decoration: underline; ">var3.02.us</a><span class="Apple-converted-space"> </span>= phi i32 [ %75, %.<a href="http://lr.ph.us" style="color: purple; text-decoration: underline; ">lr.ph.us</a><span class="Apple-converted-space"> </span>], [ %44, %.<a href="http://lr.ph6.split.us" style="color: purple; text-decoration: underline; ">lr.ph6.split.us</a><span class="Apple-converted-space"> </span>], [ %44, %59 ]<br>+ %<a href="http://xindex.01.us" style="color: purple; text-decoration: underline; ">xindex.01.us</a><span class="Apple-converted-space"> </span>= phi i32 [ %76, %.<a href="http://lr.ph.us" style="color: purple; text-decoration: underline; ">lr.ph.us</a><span class="Apple-converted-space"> </span>], [ 0, %.<a href="http://lr.ph6.split.us" style="color: purple; text-decoration: underline; ">lr.ph6.split.us</a><span class="Apple-converted-space"> </span>], [ 0, %59 ]<br>+ %70 = add nsw i32 %<a href="http://xindex.01.us" style="color: purple; text-decoration: underline; ">xindex.01.us</a>, %<a href="http://var4.15.us" style="color: purple; text-decoration: underline; ">var4.15.us</a><br>+ %71 = getelementptr inbounds [10 x [64 x i16]*]* %23, i32 0, i32 %70<br>+ %72 = load [64 x i16]** %71, align 4<br>+ %73 = getelementptr inbounds [64 x i16]* %72, i32 0, i32 0<br>+ tail call void %37(%struct.ref2* %cinfo, %struct.ref3* %32, i16* %73, i8** %<a href="http://output2.04.us" style="color: purple; text-decoration: underline; ">output2.04.us</a>, i32 %<a href="http://var3.02.us" style="color: purple; text-decoration: underline; ">var3.02.us</a>) nounwind<br>+ %74 = load i32* %40, align 4<br>+ %75 = add i32 %74, %<a href="http://var3.02.us" style="color: purple; text-decoration: underline; ">var3.02.us</a><br>+ %76 = add nsw i32 %<a href="http://xindex.01.us" style="color: purple; text-decoration: underline; ">xindex.01.us</a>, 1<br>+ %exitcond = icmp eq i32 %76, %39<br>+ br i1 %exitcond, label %.<a href="http://loopexit.us" style="color: purple; text-decoration: underline; ">loopexit.us</a>, label %.<a href="http://lr.ph.us" style="color: purple; text-decoration: underline; ">lr.ph.us</a><br>+<br>+.<a href="http://loopexit.us" style="color: purple; text-decoration: underline; ">loopexit.us</a>:<br>+ %.pre23.pre = load i32* %45, align 4<br>+ br label %._crit_edge27<br>+<br>+._crit_edge26:<br>+ %var4.15 = phi i32 [ %var4.09, %.lr.ph6..lr.ph6.split_crit_edge ], [ %78, %._crit_edge26 ]<br>+ %var5.03 = phi i32 [ 0, %.lr.ph6..lr.ph6.split_crit_edge ], [ %79, %._crit_edge26 ]<br>+ %77 = load i32* %52, align 4<br>+ %78 = add nsw i32 %77, %var4.15<br>+ %79 = add nsw i32 %var5.03, 1<br>+ %80 = icmp slt i32 %79, %46<br>+ br i1 %80, label %._crit_edge26, label %._crit_edge<br>+<br>+._crit_edge:<br>+ %split = phi i32 [ %66, %._crit_edge27 ], [ %78, %._crit_edge26 ]<br>+ %.pre25 = load i32* %19, align 4<br>+ br label %81<br>+<br>+; <label>:81<br>+ %82 = phi i32 [ %.pre25, %._crit_edge ], [ %30, %.lr.ph10 ]<br>+ %var4.1.lcssa = phi i32 [ %split, %._crit_edge ], [ %var4.09, %.lr.ph10 ]<br>+ %83 = add nsw i32 %ci.08, 1<br>+ %84 = icmp slt i32 %83, %82<br>+ br i1 %84, label %.lr.ph10, label %._crit_edge11<br>+<br>+._crit_edge11:<br>+ %85 = phi i32 [ %28, %.preheader7 ], [ %82, %81 ]<br>+ %86 = add i32 %var2.012, 1<br>+ %87 = icmp ugt i32 %86, %5<br>+ br i1 %87, label %._crit_edge14, label %.preheader7<br>+<br>+._crit_edge14:<br>+ %.pre21 = load i32* %13, align 4<br>+ br label %88<br>+<br>+; <label>:88<br>+ %89 = phi i32 [ %.pre21, %._crit_edge14 ], [ %25, %24 ]<br>+ store void (%struct.ref2*)* null, void (%struct.ref2*)** %18, align 4<br>+ %90 = add nsw i32 %var1.015, 1<br>+ %91 = icmp slt i32 %90, %89<br>+ br i1 %91, label %24, label %._crit_edge19<br>+<br>+._crit_edge19:<br>+ ret i32 3<br>+}<br><br>Added: llvm/trunk/test/CodeGen/X86/atom-call-reg-indirect-foldedreload64.ll<br>URL:<span class="apple-converted-space"> </span><a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/atom-call-reg-indirect-foldedreload64.ll?rev=178193&view=auto" target="l" style="color: purple; text-decoration: underline; "><span style="color: purple; ">http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/atom-call-reg-indirect-foldedreload64.ll?rev=178193&view=auto</span></a><br>==============================================================================<br>--- llvm/trunk/test/CodeGen/X86/atom-call-reg-indirect-foldedreload64.ll (added)<br>+++ llvm/trunk/test/CodeGen/X86/atom-call-reg-indirect-foldedreload64.ll Wed Mar 27 18:16:18 2013<br>@@ -0,0 +1,213 @@<br>+; RUN: llc < %s -mtriple=x86_64-linux-gnu -mcpu=atom 2>&1 | \<br>+; RUN: grep "callq" | not grep "("<br>+; RUN: llc < %s -mtriple=x86_64-linux-gnu -mcpu=core2 2>&1 | \<br>+; RUN: grep "callq" | grep "8-byte Folded Reload"<br>+<br>+%struct.targettype = type {i64}<br>+%struct.op_ptr1 = type opaque<br>+%struct.op_ptr2 = type opaque<br>+%union.anon = type { [8 x i32], [48 x i8] }<br>+%struct.const1 = type { [64 x i16], i8 }<br>+%struct.const2 = type { [17 x i8], [256 x i8], i8 }<br>+%struct.coef1 = type { %struct.ref1, i32, i32, i32, [10 x [64 x i16]*] }<br>+<br>+%struct.ref1 = type { void (%struct.ref2*)*, i32 (%struct.ref2*)*, void (%struct.ref2*)*, i32 (%struct.ref2*, i8***)*, %struct.op_ptr2** }<br>+%struct.ref2 = type { %struct.localref13*, %struct.localref15*, %struct.localref12*, i8*, i8, i32, %struct.localref11*, i32, i32, i32, i32, i32, i32, i32, double, i8, i8, i32, i8, i8, i8, i32, i8, i32, i8, i8, i8, i32, i32, i32, i32, i32, i32, i8**, i32, i32, i32, i32, i32, [64 x i32]*, [4 x %struct.const1*], [4 x %struct.const2*], [4 x %struct.const2*], i32, %struct.ref3*, i8, i8, [16 x i8], [16 x i8], [16 x i8], i32, i8, i8, i8, i8, i16, i16, i8, i8, i8, %struct.localref10*, i32, i32, i32, i32, i8*, i32, [4 x %struct.ref3*], i32, i32, i32, [10 x i32], i32, i32, i32, i32, i32, %struct.localref8*, %struct.localref9*, %struct.ref1*, %struct.localref7*, %struct.localref6*, %struct.localref5*, %struct.localref1*, %struct.ref4*, %struct.localref2*, %struct.localref3*, %struct.localref4* }<br>+%struct.ref3 = type { i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i8, i32, i32, i32, i32, i32, i32, %struct.const1*, i8* }<br>+%struct.ref4 = type { void (%struct.ref2*)*, [5 x void (%struct.ref2*, %struct.ref3*, i16*, i8**, i32)*] }<br>+<br>+%struct.localref1 = type { void (%struct.ref2*)*, i8 (%struct.ref2*, [64 x i16]**)*, i8 }<br>+%struct.localref2 = type { void (%struct.ref2*)*, void (%struct.ref2*, i8***, i32*, i32, i8**, i32*, i32)*, i8 }<br>+%struct.localref3 = type { void (%struct.ref2*)*, void (%struct.ref2*, i8***, i32, i8**, i32)* }<br>+%struct.localref4 = type { {}*, void (%struct.ref2*, i8**, i8**, i32)*, void (%struct.ref2*)*, void (%struct.ref2*)* }<br>+%struct.localref5 = type { void (%struct.ref2*)*, i32 (%struct.ref2*)*, i8 (%struct.ref2*)*, i8, i8, i32, i32 }<br>+%struct.localref6 = type { i32 (%struct.ref2*)*, void (%struct.ref2*)*, void (%struct.ref2*)*, void (%struct.ref2*)*, i8, i8 }<br>+%struct.localref7 = type { void (%struct.ref2*, i32)*, void (%struct.ref2*, i8***, i32*, i32, i8**, i32*, i32)* }<br>+%struct.localref8 = type { void (%struct.ref2*)*, void (%struct.ref2*)*, i8 }<br>+%struct.localref9 = type { void (%struct.ref2*, i32)*, void (%struct.ref2*, i8**, i32*, i32)* }<br>+%struct.localref10 = type { %struct.localref10*, i8, i32, i32, i8* }<br>+%struct.localref11 = type { i8*, %struct.targettype, void (%struct.ref2*)*, i8 (%struct.ref2*)*, void (%struct.ref2*, %struct.targettype)*, i8 (%struct.ref2*, i32)*, void (%struct.ref2*)* }<br>+%struct.localref12 = type { {}*, %struct.targettype, %struct.targettype, i32, i32 }<br>+%struct.localref13 = type { void (%struct.localref14*)*, void (%struct.localref14*, i32)*, void (%struct.localref14*)*, void (%struct.localref14*, i8*)*, void (%struct.localref14*)*, i32, %union.anon, i32, %struct.targettype, i8**, i32, i8**, i32, i32 }<br>+%struct.localref14 = type { %struct.localref13*, %struct.localref15*, %struct.localref12*, i8*, i8, i32 }<br>+%struct.localref15 = type { i8* (%struct.localref14*, i32, %struct.targettype)*, i8* (%struct.localref14*, i32, %struct.targettype)*, i8** (%struct.localref14*, i32, i32, i32)*, [64 x i16]** (%struct.localref14*, i32, i32, i32)*, %struct.op_ptr1* (%struct.localref14*, i32, i8, i32, i32, i32)*, %struct.op_ptr2* (%struct.localref14*, i32, i8, i32, i32, i32)*, {}*, i8** (%struct.localref14*, %struct.op_ptr1*, i32, i32, i8)*, [64 x i16]** (%struct.localref14*, %struct.op_ptr2*, i32, i32, i8)*, void (%struct.localref14*, i32)*, {}*, %struct.targettype, %struct.targettype}<br>+<br>+define internal i32 @foldedreload(%struct.ref2* %cinfo, i8*** nocapture %output1) {<br>+ %1 = getelementptr inbounds %struct.ref2* %cinfo, i64 0, i32 79<br>+ %2 = load %struct.ref1** %1, align 8<br>+ %3 = bitcast %struct.ref1* %2 to %struct.coef1*<br>+ %4 = getelementptr inbounds %struct.ref2* %cinfo, i64 0, i32 68<br>+ %5 = load i32* %4, align 4<br>+ %6 = add i32 %5, -1<br>+ %7 = getelementptr inbounds %struct.ref2* %cinfo, i64 0, i32 64<br>+ %8 = load i32* %7, align 4<br>+ %9 = add i32 %8, -1<br>+ %10 = getelementptr inbounds %struct.coef1* %3, i64 0, i32 2<br>+ %11 = load i32* %10, align 4<br>+ %12 = getelementptr inbounds %struct.ref1* %2, i64 1, i32 1<br>+ %13 = bitcast i32 (%struct.ref2*)** %12 to i32*<br>+ %14 = load i32* %13, align 4<br>+ %15 = icmp slt i32 %11, %14<br>+ br i1 %15, label %.lr.ph18, label %._crit_edge19<br>+<br>+.lr.ph18:<br>+ %16 = getelementptr inbounds %struct.ref1* %2, i64 1<br>+ %17 = bitcast %struct.ref1* %16 to i32*<br>+ %18 = getelementptr inbounds %struct.ref2* %cinfo, i64 0, i32 66<br>+ %19 = getelementptr inbounds %struct.ref2* %cinfo, i64 0, i32 84<br>+ %20 = getelementptr inbounds %struct.ref2* %cinfo, i64 0, i32 36<br>+ %21 = getelementptr inbounds %struct.ref1* %2, i64 1, i32 2<br>+ %22 = bitcast void (%struct.ref2*)** %21 to [10 x [64 x i16]*]*<br>+ %.pre = load i32* %17, align 4<br>+ br label %23<br>+<br>+; <label>:23<br>+ %24 = phi i32 [ %14, %.lr.ph18 ], [ %92, %91 ]<br>+ %25 = phi i32 [ %.pre, %.lr.ph18 ], [ 0, %91 ]<br>+ %var1.015 = phi i32 [ %11, %.lr.ph18 ], [ %93, %91 ]<br>+ %26 = icmp ugt i32 %25, %6<br>+ br i1 %26, label %91, label %.preheader7.lr.ph<br>+<br>+.preheader7.lr.ph:<br>+ %.pre24 = load i32* %18, align 4<br>+ br label %.preheader7<br>+<br>+.preheader7:<br>+ %27 = phi i32 [ %.pre24, %.preheader7.lr.ph ], [ %88, %._crit_edge11 ]<br>+ %var2.012 = phi i32 [ %25, %.preheader7.lr.ph ], [ %89, %._crit_edge11 ]<br>+ %28 = icmp sgt i32 %27, 0<br>+ br i1 %28, label %.lr.ph10, label %._crit_edge11<br>+<br>+.lr.ph10:<br>+ %29 = phi i32 [ %27, %.preheader7 ], [ %85, %84 ]<br>+ %indvars.iv21 = phi i64 [ 0, %.preheader7 ], [ %indvars.iv.next22, %84 ]<br>+ %var4.09 = phi i32 [ 0, %.preheader7 ], [ %var4.1.lcssa, %84 ]<br>+ %30 = getelementptr inbounds %struct.ref2* %cinfo, i64 0, i32 67, i64 %indvars.iv21<br>+ %31 = load %struct.ref3** %30, align 8<br>+ %32 = getelementptr inbounds %struct.ref3* %31, i64 0, i32 1<br>+ %33 = load i32* %32, align 4<br>+ %34 = sext i32 %33 to i64<br>+ %35 = load %struct.ref4** %19, align 8<br>+ %36 = getelementptr inbounds %struct.ref4* %35, i64 0, i32 1, i64 %34<br>+ %37 = load void (%struct.ref2*, %struct.ref3*, i16*, i8**, i32)** %36, align 8<br>+ %38 = getelementptr inbounds %struct.ref3* %31, i64 0, i32 17<br>+ %39 = load i32* %38, align 4<br>+ %40 = getelementptr inbounds %struct.ref3* %31, i64 0, i32 9<br>+ %41 = load i32* %40, align 4<br>+ %42 = getelementptr inbounds %struct.ref3* %31, i64 0, i32 16<br>+ %43 = load i32* %42, align 4<br>+ %44 = mul i32 %43, %var2.012<br>+ %45 = getelementptr inbounds %struct.ref3* %31, i64 0, i32 14<br>+ %46 = load i32* %45, align 4<br>+ %47 = icmp sgt i32 %46, 0<br>+ br i1 %47, label %.lr.ph6, label %84<br>+<br>+.lr.ph6:<br>+ %48 = mul nsw i32 %41, %var1.015<br>+ %49 = getelementptr inbounds i8*** %output1, i64 %34<br>+ %50 = sext i32 %48 to i64<br>+ %51 = load i8*** %49, align 8<br>+ %52 = getelementptr inbounds i8** %51, i64 %50<br>+ %53 = getelementptr inbounds %struct.ref3* %31, i64 0, i32 13<br>+ %54 = getelementptr inbounds %struct.ref3* %31, i64 0, i32 18<br>+ %55 = icmp sgt i32 %39, 0<br>+ br i1 %55, label %.<a href="http://lr.ph6.split.us" style="color: purple; text-decoration: underline; ">lr.ph6.split.us</a>, label %.lr.ph6..lr.ph6.split_crit_edge<br>+<br>+.lr.ph6..lr.ph6.split_crit_edge:<br>+ br label %._crit_edge28<br>+<br>+.<a href="http://lr.ph6.split.us" style="color: purple; text-decoration: underline; ">lr.ph6.split.us</a>:<br>+ %56 = phi i32 [ %64, %._crit_edge30 ], [ %46, %.lr.ph6 ]<br>+ %57 = phi i32 [ %65, %._crit_edge30 ], [ %41, %.lr.ph6 ]<br>+ %<a href="http://var4.15.us" style="color: purple; text-decoration: underline; ">var4.15.us</a><span class="Apple-converted-space"> </span>= phi i32 [ %67, %._crit_edge30 ], [ %var4.09, %.lr.ph6 ]<br>+ %<a href="http://output2.04.us" style="color: purple; text-decoration: underline; ">output2.04.us</a><span class="Apple-converted-space"> </span>= phi i8** [ %71, %._crit_edge30 ], [ %52, %.lr.ph6 ]<br>+ %<a href="http://var5.03.us" style="color: purple; text-decoration: underline; ">var5.03.us</a><span class="Apple-converted-space"> </span>= phi i32 [ %68, %._crit_edge30 ], [ 0, %.lr.ph6 ]<br>+ %58 = load i32* %20, align 4<br>+ %59 = icmp ult i32 %58, %9<br>+ br i1 %59, label %.<a href="http://lr.ph.us" style="color: purple; text-decoration: underline; ">lr.ph.us</a>, label %60<br>+<br>+; <label>:60<br>+ %61 = add nsw i32 %<a href="http://var5.03.us" style="color: purple; text-decoration: underline; ">var5.03.us</a>, %var1.015<br>+ %62 = load i32* %54, align 4<br>+ %63 = icmp slt i32 %61, %62<br>+ br i1 %63, label %.<a href="http://lr.ph.us" style="color: purple; text-decoration: underline; ">lr.ph.us</a>, label %._crit_edge29<br>+<br>+._crit_edge29:<br>+ %64 = phi i32 [ %.pre25.pre, %.<a href="http://loopexit.us" style="color: purple; text-decoration: underline; ">loopexit.us</a><span class="Apple-converted-space"> </span>], [ %56, %60 ]<br>+ %65 = phi i32 [ %77, %.<a href="http://loopexit.us" style="color: purple; text-decoration: underline; ">loopexit.us</a><span class="Apple-converted-space"> </span>], [ %57, %60 ]<br>+ %66 = load i32* %53, align 4<br>+ %67 = add nsw i32 %66, %<a href="http://var4.15.us" style="color: purple; text-decoration: underline; ">var4.15.us</a><br>+ %68 = add nsw i32 %<a href="http://var5.03.us" style="color: purple; text-decoration: underline; ">var5.03.us</a>, 1<br>+ %69 = icmp slt i32 %68, %64<br>+ br i1 %69, label %._crit_edge30, label %._crit_edge<br>+<br>+._crit_edge30:<br>+ %70 = sext i32 %65 to i64<br>+ %71 = getelementptr inbounds i8** %<a href="http://output2.04.us" style="color: purple; text-decoration: underline; ">output2.04.us</a>, i64 %70<br>+ br label %.<a href="http://lr.ph6.split.us" style="color: purple; text-decoration: underline; ">lr.ph6.split.us</a><br>+<br>+; <label>:72<br>+ %indvars.iv = phi i64 [ 0, %.<a href="http://lr.ph.us" style="color: purple; text-decoration: underline; ">lr.ph.us</a><span class="Apple-converted-space"> </span>], [ %indvars.iv.next, %72 ]<br>+ %<a href="http://var3.02.us" style="color: purple; text-decoration: underline; ">var3.02.us</a><span class="Apple-converted-space"> </span>= phi i32 [ %44, %.<a href="http://lr.ph.us" style="color: purple; text-decoration: underline; ">lr.ph.us</a><span class="Apple-converted-space"> </span>], [ %78, %72 ]<br>+ %73 = add nsw i64 %indvars.iv, %79<br>+ %74 = getelementptr inbounds [10 x [64 x i16]*]* %22, i64 0, i64 %73<br>+ %75 = load [64 x i16]** %74, align 8<br>+ %76 = getelementptr inbounds [64 x i16]* %75, i64 0, i64 0<br>+ tail call void %37(%struct.ref2* %cinfo, %struct.ref3* %31, i16* %76, i8** %<a href="http://output2.04.us" style="color: purple; text-decoration: underline; ">output2.04.us</a>, i32 %<a href="http://var3.02.us" style="color: purple; text-decoration: underline; ">var3.02.us</a>) nounwind<br>+ %77 = load i32* %40, align 4<br>+ %78 = add i32 %77, %<a href="http://var3.02.us" style="color: purple; text-decoration: underline; ">var3.02.us</a><br>+ %indvars.iv.next = add i64 %indvars.iv, 1<br>+ %lftr.wideiv = trunc i64 %indvars.iv.next to i32<br>+ %exitcond = icmp eq i32 %lftr.wideiv, %39<br>+ br i1 %exitcond, label %.<a href="http://loopexit.us" style="color: purple; text-decoration: underline; ">loopexit.us</a>, label %72<br>+<br>+.<a href="http://loopexit.us" style="color: purple; text-decoration: underline; ">loopexit.us</a>:<br>+ %.pre25.pre = load i32* %45, align 4<br>+ br label %._crit_edge29<br>+<br>+.<a href="http://lr.ph.us" style="color: purple; text-decoration: underline; ">lr.ph.us</a>:<br>+ %79 = sext i32 %<a href="http://var4.15.us" style="color: purple; text-decoration: underline; ">var4.15.us</a><span class="Apple-converted-space"> </span>to i64<br>+ br label %72<br>+<br>+._crit_edge28:<br>+ %var4.15 = phi i32 [ %var4.09, %.lr.ph6..lr.ph6.split_crit_edge ], [ %81, %._crit_edge28 ]<br>+ %var5.03 = phi i32 [ 0, %.lr.ph6..lr.ph6.split_crit_edge ], [ %82, %._crit_edge28 ]<br>+ %80 = load i32* %53, align 4<br>+ %81 = add nsw i32 %80, %var4.15<br>+ %82 = add nsw i32 %var5.03, 1<br>+ %83 = icmp slt i32 %82, %46<br>+ br i1 %83, label %._crit_edge28, label %._crit_edge<br>+<br>+._crit_edge:<br>+ %split = phi i32 [ %67, %._crit_edge29 ], [ %81, %._crit_edge28 ]<br>+ %.pre27 = load i32* %18, align 4<br>+ br label %84<br>+<br>+; <label>:84<br>+ %85 = phi i32 [ %.pre27, %._crit_edge ], [ %29, %.lr.ph10 ]<br>+ %var4.1.lcssa = phi i32 [ %split, %._crit_edge ], [ %var4.09, %.lr.ph10 ]<br>+ %indvars.iv.next22 = add i64 %indvars.iv21, 1<br>+ %86 = trunc i64 %indvars.iv.next22 to i32<br>+ %87 = icmp slt i32 %86, %85<br>+ br i1 %87, label %.lr.ph10, label %._crit_edge11<br>+<br>+._crit_edge11:<br>+ %88 = phi i32 [ %27, %.preheader7 ], [ %85, %84 ]<br>+ %89 = add i32 %var2.012, 1<br>+ %90 = icmp ugt i32 %89, %6<br>+ br i1 %90, label %._crit_edge14, label %.preheader7<br>+<br>+._crit_edge14:<br>+ %.pre23 = load i32* %13, align 4<br>+ br label %91<br>+<br>+; <label>:91<br>+ %92 = phi i32 [ %.pre23, %._crit_edge14 ], [ %24, %23 ]<br>+ store i32 0, i32* %17, align 4<br>+ %93 = add nsw i32 %var1.015, 1<br>+ %94 = icmp slt i32 %93, %92<br>+ br i1 %94, label %23, label %._crit_edge19<br>+<br>+._crit_edge19:<br>+ ret i32 3<br>+}<br><br>Modified: llvm/trunk/test/CodeGen/X86/movgs.ll<br>URL:<span class="apple-converted-space"> </span><a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/movgs.ll?rev=178193&r1=178192&r2=178193&view=diff" target="l" style="color: purple; text-decoration: underline; "><span style="color: purple; ">http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/movgs.ll?rev=178193&r1=178192&r2=178193&view=diff</span></a><br>==============================================================================<br>--- llvm/trunk/test/CodeGen/X86/movgs.ll (original)<br>+++ llvm/trunk/test/CodeGen/X86/movgs.ll Wed Mar 27 18:16:18 2013<br>@@ -1,6 +1,6 @@<br>-; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu -mattr=sse41 | FileCheck %s --check-prefix=X32<br>-; RUN: llc < %s -mtriple=x86_64-linux -mattr=sse41 | FileCheck %s --check-prefix=X64<br>-; RUN: llc < %s -mtriple=x86_64-win32 -mattr=sse41 | FileCheck %s --check-prefix=X64<br>+; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu -mcpu=penryn -mattr=sse41 | FileCheck %s --check-prefix=X32<br>+; RUN: llc < %s -mtriple=x86_64-linux -mcpu=penryn -mattr=sse41 | FileCheck %s --check-prefix=X64<br>+; RUN: llc < %s -mtriple=x86_64-win32 -mcpu=penryn -mattr=sse41 | FileCheck %s --check-prefix=X64<br> <br> define i32 @test1() nounwind readonly {<br> entry:<br><br>Modified: llvm/trunk/test/CodeGen/X86/sibcall.ll<br>URL:<span class="apple-converted-space"> </span><a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/sibcall.ll?rev=178193&r1=178192&r2=178193&view=diff" target="l" style="color: purple; text-decoration: underline; "><span style="color: purple; ">http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/sibcall.ll?rev=178193&r1=178192&r2=178193&view=diff</span></a><br>==============================================================================<br>--- llvm/trunk/test/CodeGen/X86/sibcall.ll (original)<br>+++ llvm/trunk/test/CodeGen/X86/sibcall.ll Wed Mar 27 18:16:18 2013<br>@@ -1,5 +1,5 @@<br>-; RUN: llc < %s -mtriple=i686-linux -mattr=+sse2 -asm-verbose=false | FileCheck %s -check-prefix=32<br>-; RUN: llc < %s -mtriple=x86_64-linux -mattr=+sse2 -asm-verbose=false | FileCheck %s -check-prefix=64<br>+; RUN: llc < %s -mtriple=i686-linux -mcpu=core2 -mattr=+sse2 -asm-verbose=false | FileCheck %s -check-prefix=32<br>+; RUN: llc < %s -mtriple=x86_64-linux -mcpu=core2 -mattr=+sse2 -asm-verbose=false | FileCheck %s -check-prefix=64<br> <br> define void @t1(i32 %x) nounwind ssp {<br> entry:<br><br><br>_______________________________________________<br>llvm-commits mailing list<br><a href="mailto:llvm-commits@cs.uiuc.edu" style="color: purple; text-decoration: underline; "><span style="color: purple; ">llvm-commits@cs.uiuc.edu</span></a><br><a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits" target="l" style="color: purple; text-decoration: underline; "><span style="color: purple; ">http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits</span></a><o:p></o:p></div></div></div></div></div><div style="margin: 0in 0in 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif; "><o:p> </o:p></div></div>_______________________________________________<br>llvm-commits mailing list<br><a href="mailto:llvm-commits@cs.uiuc.edu">llvm-commits@cs.uiuc.edu</a><br>http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits</div></blockquote></div><br></div></body></html>