[llvm] r208934 - Implement global merge optimization for global variables.
Rafael Espíndola
rafael.espindola at gmail.com
Thu May 15 20:43:20 PDT 2014
Sorry I missed this, but this uses a aliases pointing to GEPs. That is
not supported and is currently fully broken:
$ cat test.ll
@foo = global [1 x i32] zeroinitializer
@bar = alias getelementptr ([1 x i32]* @foo, i32 1, i32 1)
$ llvm-as test.ll -o - | llvm-dis -o -
@foo = global [1 x i32] zeroinitializer
@bar = alias getelementptr ([1 x i32]* @foo, i64 2, i32 0)
The good news is that I found this while rebasing my fix for pr10367
which just got LGTMed and should allow this to be correctly supported
by adding and explicit offset for aliases.
Since this is off by default anyway, I hope you don't mind If I revert
this tomorrow morning. I volunteer to help you port it to the new
alias :-)
Cheers,
Rafael
On 15 May 2014 19:45, Jiangning Liu <jiangning.liu at arm.com> wrote:
> Author: jiangning
> Date: Thu May 15 18:45:42 2014
> New Revision: 208934
>
> URL: http://llvm.org/viewvc/llvm-project?rev=208934&view=rev
> Log:
> Implement global merge optimization for global variables.
>
> This commit implements two command line switches -global-merge-on-external
> and -global-merge-aligned, and both of them are false by default, so this
> optimization is disabled by default for all targets.
>
> For ARM64, some back-end behaviors need to be tuned to get this optimization
> further enabled.
>
>
> Added:
> llvm/trunk/test/CodeGen/AArch64/global-merge.ll
> llvm/trunk/test/CodeGen/ARM/global-merge-1.ll
> llvm/trunk/test/CodeGen/ARM64/global-merge.ll
> llvm/trunk/test/Transforms/GlobalMerge/AArch64/
> llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-1.ll
> llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-2.ll
> llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-3.ll
> llvm/trunk/test/Transforms/GlobalMerge/AArch64/lit.local.cfg
> Removed:
> llvm/trunk/test/CodeGen/AArch64/global_merge_1.ll
> Modified:
> llvm/trunk/include/llvm/IR/GlobalAlias.h
> llvm/trunk/include/llvm/Target/TargetLowering.h
> llvm/trunk/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
> llvm/trunk/lib/IR/Globals.cpp
> llvm/trunk/lib/Target/AArch64/AArch64ISelLowering.cpp
> llvm/trunk/lib/Target/AArch64/AArch64ISelLowering.h
> llvm/trunk/lib/Target/ARM64/ARM64ISelLowering.cpp
> llvm/trunk/lib/Target/ARM64/ARM64ISelLowering.h
> llvm/trunk/lib/Transforms/Scalar/GlobalMerge.cpp
> llvm/trunk/lib/Transforms/Scalar/Scalar.cpp
> llvm/trunk/test/Transforms/GlobalMerge/ARM/arm.ll
> llvm/trunk/test/Transforms/GlobalMerge/ARM64/arm64.ll
>
> Modified: llvm/trunk/include/llvm/IR/GlobalAlias.h
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/include/llvm/IR/GlobalAlias.h?rev=208934&r1=208933&r2=208934&view=diff
> ==============================================================================
> --- llvm/trunk/include/llvm/IR/GlobalAlias.h (original)
> +++ llvm/trunk/include/llvm/IR/GlobalAlias.h Thu May 15 18:45:42 2014
> @@ -82,6 +82,10 @@ public:
> static inline bool classof(const Value *V) {
> return V->getValueID() == Value::GlobalAliasVal;
> }
> +
> + // return the constant offset of an expression, with which this global var
> + // has alias.
> + uint64_t calculateOffset(const DataLayout &DL) const;
> };
>
> template <>
>
> Modified: llvm/trunk/include/llvm/Target/TargetLowering.h
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/include/llvm/Target/TargetLowering.h?rev=208934&r1=208933&r2=208934&view=diff
> ==============================================================================
> --- llvm/trunk/include/llvm/Target/TargetLowering.h (original)
> +++ llvm/trunk/include/llvm/Target/TargetLowering.h Thu May 15 18:45:42 2014
> @@ -32,6 +32,7 @@
> #include "llvm/IR/CallingConv.h"
> #include "llvm/IR/InlineAsm.h"
> #include "llvm/IR/IRBuilder.h"
> +#include "llvm/IR/GlobalVariable.h"
> #include "llvm/MC/MCRegisterInfo.h"
> #include "llvm/Target/TargetCallingConv.h"
> #include "llvm/Target/TargetMachine.h"
> @@ -883,6 +884,12 @@ public:
> return 0;
> }
>
> + /// Returns the alignment required by global merge on external symbols.
> + /// By default, returns the natural alignment of merged data structure.
> + virtual unsigned getGlobalMergeAlignment(StructType *MergedTy) const {
> + return getDataLayout()->getABITypeAlignment(MergedTy);
> + }
> +
> /// Returns true if a cast between SrcAS and DestAS is a noop.
> virtual bool isNoopAddrSpaceCast(unsigned SrcAS, unsigned DestAS) const {
> return false;
>
> Modified: llvm/trunk/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/AsmPrinter/AsmPrinter.cpp?rev=208934&r1=208933&r2=208934&view=diff
> ==============================================================================
> --- llvm/trunk/lib/CodeGen/AsmPrinter/AsmPrinter.cpp (original)
> +++ llvm/trunk/lib/CodeGen/AsmPrinter/AsmPrinter.cpp Thu May 15 18:45:42 2014
> @@ -946,8 +946,11 @@ bool AsmPrinter::doFinalization(Module &
> EmitVisibility(Name, Alias.getVisibility());
>
> // Emit the directives as assignments aka .set:
> - OutStreamer.EmitAssignment(Name,
> - MCSymbolRefExpr::Create(Target, OutContext));
> + const MCExpr *Expr = MCSymbolRefExpr::Create(Target, OutContext);
> + if (uint64_t Offset = Alias.calculateOffset(*TM.getDataLayout()))
> + Expr = MCBinaryExpr::CreateAdd(Expr,
> + MCConstantExpr::Create(Offset, OutContext), OutContext);
> + OutStreamer.EmitAssignment(Name, Expr);
> }
> }
>
>
> Modified: llvm/trunk/lib/IR/Globals.cpp
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/IR/Globals.cpp?rev=208934&r1=208933&r2=208934&view=diff
> ==============================================================================
> --- llvm/trunk/lib/IR/Globals.cpp (original)
> +++ llvm/trunk/lib/IR/Globals.cpp Thu May 15 18:45:42 2014
> @@ -15,6 +15,7 @@
> #include "llvm/IR/GlobalValue.h"
> #include "llvm/ADT/SmallPtrSet.h"
> #include "llvm/IR/Constants.h"
> +#include "llvm/IR/DataLayout.h"
> #include "llvm/IR/DerivedTypes.h"
> #include "llvm/IR/GlobalAlias.h"
> #include "llvm/IR/GlobalVariable.h"
> @@ -282,3 +283,27 @@ GlobalObject *GlobalAlias::getAliasedGlo
> return cast<GlobalObject>(GV);
> }
> }
> +
> +uint64_t GlobalAlias::calculateOffset(const DataLayout &DL) const {
> + uint64_t Offset = 0;
> + const Constant *C = this;
> + while (C) {
> + if (const GlobalAlias *GA = dyn_cast<GlobalAlias>(C)) {
> + C = GA->getAliasee();
> + } else if (const ConstantExpr *CE = dyn_cast<ConstantExpr>(C)) {
> + if (CE->getOpcode() == Instruction::GetElementPtr) {
> + std::vector<Value*> Args;
> + for (unsigned I = 1; I < CE->getNumOperands(); ++I)
> + Args.push_back(CE->getOperand(I));
> + Offset += DL.getIndexedOffset(CE->getOperand(0)->getType(), Args);
> + }
> + C = CE->getOperand(0);
> + } else if (isa<GlobalValue>(C)) {
> + return Offset;
> + } else {
> + assert(0 && "Unexpected type in alias chain!");
> + return 0;
> + }
> + }
> + return Offset;
> +}
>
> Modified: llvm/trunk/lib/Target/AArch64/AArch64ISelLowering.cpp
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AArch64/AArch64ISelLowering.cpp?rev=208934&r1=208933&r2=208934&view=diff
> ==============================================================================
> --- llvm/trunk/lib/Target/AArch64/AArch64ISelLowering.cpp (original)
> +++ llvm/trunk/lib/Target/AArch64/AArch64ISelLowering.cpp Thu May 15 18:45:42 2014
> @@ -5560,3 +5560,17 @@ unsigned AArch64TargetLowering::getMaxim
> return 4095;
> }
>
> +/// getGlobalMergeAlignment - Set alignment to be the max size of merged
> +/// global variable data structure, and make it aligned up to power of 2.
> +/// This way, we could guarantee the merged global variable data structure
> +/// doesn't cross page boundary, because usually OS always allocates page at
> +/// 4096-byte aligned boundary.
> +unsigned AArch64TargetLowering::getGlobalMergeAlignment(
> + StructType *MergedTy) const {
> + unsigned Align = getDataLayout()->getTypeAllocSize(MergedTy);
> + if (Align & (Align - 1))
> + Align = llvm::NextPowerOf2(Align);
> +
> + return Align;
> +}
> +
>
> Modified: llvm/trunk/lib/Target/AArch64/AArch64ISelLowering.h
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AArch64/AArch64ISelLowering.h?rev=208934&r1=208933&r2=208934&view=diff
> ==============================================================================
> --- llvm/trunk/lib/Target/AArch64/AArch64ISelLowering.h (original)
> +++ llvm/trunk/lib/Target/AArch64/AArch64ISelLowering.h Thu May 15 18:45:42 2014
> @@ -386,6 +386,10 @@ public:
> /// be used for loads / stores from the global.
> unsigned getMaximalGlobalOffset() const override;
>
> + /// getGlobalMergeAlignment - Set alignment to be the max size of merged
> + /// global variable data structure, and make it aligned up to power of 2.
> + unsigned getGlobalMergeAlignment(StructType *MergedTy) const override;
> +
> protected:
> std::pair<const TargetRegisterClass*, uint8_t>
> findRepresentativeClass(MVT VT) const override;
>
> Modified: llvm/trunk/lib/Target/ARM64/ARM64ISelLowering.cpp
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/ARM64/ARM64ISelLowering.cpp?rev=208934&r1=208933&r2=208934&view=diff
> ==============================================================================
> --- llvm/trunk/lib/Target/ARM64/ARM64ISelLowering.cpp (original)
> +++ llvm/trunk/lib/Target/ARM64/ARM64ISelLowering.cpp Thu May 15 18:45:42 2014
> @@ -630,6 +630,20 @@ unsigned ARM64TargetLowering::getMaximal
> return 4095;
> }
>
> +/// getGlobalMergeAlignment - Set alignment to be the max size of merged
> +/// global variable data structure, and make it aligned up to power of 2.
> +/// This way, we could guarantee the merged global variable data structure
> +/// doesn't cross page boundary, because usually OS always allocates page at
> +/// 4096-byte aligned boundary.
> +unsigned ARM64TargetLowering::getGlobalMergeAlignment(
> + StructType *MergedTy) const {
> + unsigned Align = getDataLayout()->getTypeAllocSize(MergedTy);
> + if (Align & (Align - 1))
> + Align = llvm::NextPowerOf2(Align);
> +
> + return Align;
> +}
> +
> FastISel *
> ARM64TargetLowering::createFastISel(FunctionLoweringInfo &funcInfo,
> const TargetLibraryInfo *libInfo) const {
>
> Modified: llvm/trunk/lib/Target/ARM64/ARM64ISelLowering.h
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/ARM64/ARM64ISelLowering.h?rev=208934&r1=208933&r2=208934&view=diff
> ==============================================================================
> --- llvm/trunk/lib/Target/ARM64/ARM64ISelLowering.h (original)
> +++ llvm/trunk/lib/Target/ARM64/ARM64ISelLowering.h Thu May 15 18:45:42 2014
> @@ -236,6 +236,10 @@ public:
> /// be used for loads / stores from the global.
> unsigned getMaximalGlobalOffset() const override;
>
> + /// getGlobalMergeAlignment - Set alignment to be the max size of merged
> + /// global variable data structure, and make it aligned up to power of 2.
> + unsigned getGlobalMergeAlignment(StructType *MergedTy) const override;
> +
> /// Returns true if a cast between SrcAS and DestAS is a noop.
> bool isNoopAddrSpaceCast(unsigned SrcAS, unsigned DestAS) const override {
> // Addrspacecasts are always noops.
>
> Modified: llvm/trunk/lib/Transforms/Scalar/GlobalMerge.cpp
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Scalar/GlobalMerge.cpp?rev=208934&r1=208933&r2=208934&view=diff
> ==============================================================================
> --- llvm/trunk/lib/Transforms/Scalar/GlobalMerge.cpp (original)
> +++ llvm/trunk/lib/Transforms/Scalar/GlobalMerge.cpp Thu May 15 18:45:42 2014
> @@ -72,7 +72,7 @@ using namespace llvm;
> #define DEBUG_TYPE "global-merge"
>
> static cl::opt<bool>
> -EnableGlobalMerge("global-merge", cl::Hidden,
> +EnableGlobalMerge("enable-global-merge", cl::NotHidden,
> cl::desc("Enable global merge pass"),
> cl::init(true));
>
> @@ -81,6 +81,16 @@ EnableGlobalMergeOnConst("global-merge-o
> cl::desc("Enable global merge pass on constants"),
> cl::init(false));
>
> +static cl::opt<bool>
> +EnableGlobalMergeOnExternal("global-merge-on-external", cl::Hidden,
> + cl::desc("Enable global merge pass on external linkage"),
> + cl::init(false));
> +
> +static cl::opt<bool>
> +EnableGlobalMergeAligned("global-merge-aligned", cl::Hidden,
> + cl::desc("Set target specific alignment for global merge pass"),
> + cl::init(false));
> +
> STATISTIC(NumMerged , "Number of globals merged");
> namespace {
> class GlobalMerge : public FunctionPass {
> @@ -129,9 +139,21 @@ namespace {
> } // end anonymous namespace
>
> char GlobalMerge::ID = 0;
> -INITIALIZE_PASS(GlobalMerge, "global-merge",
> - "Global Merge", false, false)
>
> +static void *initializeGlobalMergePassOnce(PassRegistry &Registry) {
> + PassInfo *PI = new PassInfo(
> + "Merge global variables",
> + "global-merge", &GlobalMerge::ID,
> + PassInfo::NormalCtor_t(callDefaultCtor<GlobalMerge>), false,
> + false, PassInfo::TargetMachineCtor_t(
> + callTargetMachineCtor<GlobalMerge>));
> + Registry.registerPass(*PI, true);
> + return PI;
> +}
> +
> +void llvm::initializeGlobalMergePass(PassRegistry &Registry) {
> + CALL_ONCE_INITIALIZATION(initializeGlobalMergePassOnce)
> +}
>
> bool GlobalMerge::doMerge(SmallVectorImpl<GlobalVariable*> &Globals,
> Module &M, bool isConst, unsigned AddrSpace) const {
> @@ -154,11 +176,16 @@ bool GlobalMerge::doMerge(SmallVectorImp
>
> Type *Int32Ty = Type::getInt32Ty(M.getContext());
>
> + assert (Globals.size() > 1);
> +
> for (size_t i = 0, e = Globals.size(); i != e; ) {
> size_t j = 0;
> uint64_t MergedSize = 0;
> std::vector<Type*> Tys;
> std::vector<Constant*> Inits;
> +
> + bool HasExternal = false;
> + GlobalVariable *TheFirstExternal = 0;
> for (j = i; j != e; ++j) {
> Type *Ty = Globals[j]->getType()->getElementType();
> MergedSize += DL->getTypeAllocSize(Ty);
> @@ -167,17 +194,45 @@ bool GlobalMerge::doMerge(SmallVectorImp
> }
> Tys.push_back(Ty);
> Inits.push_back(Globals[j]->getInitializer());
> +
> + if (Globals[j]->hasExternalLinkage() && !HasExternal) {
> + HasExternal = true;
> + TheFirstExternal = Globals[j];
> + }
> }
>
> + // If merged variables doesn't have external linkage, we needn't to expose
> + // the symbol after merging.
> + GlobalValue::LinkageTypes Linkage = HasExternal ?
> + GlobalValue::ExternalLinkage :
> + GlobalValue::InternalLinkage ;
> +
> + // If merged variables have external linkage, we use symbol name of the
> + // first variable merged as the suffix of global symbol name. This would
> + // be able to avoid the link-time naming conflict for globalm symbols.
> + Twine MergedGVName = HasExternal ?
> + "_MergedGlobals_" + TheFirstExternal->getName() :
> + "_MergedGlobals" ;
> +
> StructType *MergedTy = StructType::get(M.getContext(), Tys);
> Constant *MergedInit = ConstantStruct::get(MergedTy, Inits);
> +
> GlobalVariable *MergedGV = new GlobalVariable(M, MergedTy, isConst,
> - GlobalValue::InternalLinkage,
> - MergedInit, "_MergedGlobals",
> - nullptr,
> - GlobalVariable::NotThreadLocal,
> - AddrSpace);
> + Linkage, MergedInit, MergedGVName,
> + nullptr, GlobalVariable::NotThreadLocal,
> + AddrSpace);
> +
> + if (EnableGlobalMergeAligned) {
> + unsigned Align = TLI->getGlobalMergeAlignment(MergedTy);
> + assert(((Align % DL->getABITypeAlignment(MergedTy)) == 0) &&
> + "Specified alignment doesn't meet natural alignment requirement.");
> + MergedGV->setAlignment(Align);
> + }
> +
> for (size_t k = i; k < j; ++k) {
> + GlobalValue::LinkageTypes Linkage = Globals[k]->getLinkage();
> + std::string Name = Globals[k]->getName();
> +
> Constant *Idx[2] = {
> ConstantInt::get(Int32Ty, 0),
> ConstantInt::get(Int32Ty, k-i)
> @@ -185,6 +240,12 @@ bool GlobalMerge::doMerge(SmallVectorImp
> Constant *GEP = ConstantExpr::getInBoundsGetElementPtr(MergedGV, Idx);
> Globals[k]->replaceAllUsesWith(GEP);
> Globals[k]->eraseFromParent();
> +
> + if (Linkage != GlobalValue::InternalLinkage) {
> + // Generate a new alias...
> + new GlobalAlias(GEP->getType(), Linkage, Name, GEP, &M);
> + }
> +
> NumMerged++;
> }
> i = j;
> @@ -245,8 +306,12 @@ bool GlobalMerge::doInitialization(Modul
> // Grab all non-const globals.
> for (Module::global_iterator I = M.global_begin(),
> E = M.global_end(); I != E; ++I) {
> - // Merge is safe for "normal" internal globals only
> - if (!I->hasLocalLinkage() || I->isThreadLocal() || I->hasSection())
> + // Merge is safe for "normal" internal or external globals only
> + if (I->isDeclaration() || I->isThreadLocal() || I->hasSection())
> + continue;
> +
> + if (!(EnableGlobalMergeOnExternal && I->hasExternalLinkage())
> + && !I->hasInternalLinkage())
> continue;
>
> PointerType *PT = dyn_cast<PointerType>(I->getType());
>
> Modified: llvm/trunk/lib/Transforms/Scalar/Scalar.cpp
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Scalar/Scalar.cpp?rev=208934&r1=208933&r2=208934&view=diff
> ==============================================================================
> --- llvm/trunk/lib/Transforms/Scalar/Scalar.cpp (original)
> +++ llvm/trunk/lib/Transforms/Scalar/Scalar.cpp Thu May 15 18:45:42 2014
> @@ -38,6 +38,7 @@ void llvm::initializeScalarOpts(PassRegi
> initializeDSEPass(Registry);
> initializeGVNPass(Registry);
> initializeEarlyCSEPass(Registry);
> + initializeGlobalMergePass(Registry);
> initializeIndVarSimplifyPass(Registry);
> initializeJumpThreadingPass(Registry);
> initializeLICMPass(Registry);
>
> Added: llvm/trunk/test/CodeGen/AArch64/global-merge.ll
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/AArch64/global-merge.ll?rev=208934&view=auto
> ==============================================================================
> --- llvm/trunk/test/CodeGen/AArch64/global-merge.ll (added)
> +++ llvm/trunk/test/CodeGen/AArch64/global-merge.ll Thu May 15 18:45:42 2014
> @@ -0,0 +1,34 @@
> +; RUN: llc < %s -mtriple=aarch64-none-linux-gnu -O0 | FileCheck --check-prefix=NO-MERGE %s
> +; RUN: llc < %s -mtriple=aarch64-none-linux-gnu -O0 -global-merge-on-external=true -global-merge-aligned=true | FileCheck --check-prefix=NO-MERGE %s
> +
> +; RUN: llc < %s -mtriple=arm64-apple-ios -O0 | FileCheck %s --check-prefix=CHECK-APPLE-IOS-NO-MERGE
> +; RUN: llc < %s -mtriple=arm64-apple-ios -O0 -global-merge-on-external=true -global-merge-aligned=false | FileCheck %s --check-prefix=CHECK-APPLE-IOS-NO-MERGE
> +
> +; RUN: llc < %s -mtriple=aarch64-none-linux-gnu -O1 | FileCheck %s
> +; RUN: llc < %s -mtriple=aarch64-none-linux-gnu -O1 -global-merge-on-external=true -global-merge-aligned=true | FileCheck %s
> +
> +; RUN: llc < %s -mtriple=arm64-apple-ios -O1 | FileCheck %s --check-prefix=CHECK-APPLE-IOS
> +; RUN: llc < %s -mtriple=arm64-apple-ios -O1 -global-merge-on-external=true -global-merge-aligned=false | FileCheck %s --check-prefix=CHECK-APPLE-IOS
> +
> + at m = internal global i32 0, align 4
> + at n = internal global i32 0, align 4
> +
> +define void @f1(i32 %a1, i32 %a2) {
> +; CHECK-LABEL: f1:
> +; CHECK: adrp x{{[0-9]+}}, _MergedGlobals
> +; CHECK-NOT: adrp
> +
> +; CHECK-APPLE-IOS-LABEL: f1:
> +; CHECK-APPLE-IOS: adrp x{{[0-9]+}}, __MergedGlobals
> +; CHECK-APPLE-IOS-NOT: adrp
> + store i32 %a1, i32* @m, align 4
> + store i32 %a2, i32* @n, align 4
> + ret void
> +}
> +
> +; CHECK: .local _MergedGlobals
> +; CHECK: .comm _MergedGlobals,8,8
> +; NO-MERGE-NOT: .local _MergedGlobals
> +
> +; CHECK-APPLE-IOS: .zerofill __DATA,__bss,__MergedGlobals,8,3
> +; CHECK-APPLE-IOS-NO-MERGE-NOT: .zerofill __DATA,__bss,__MergedGlobals,8,3
>
> Removed: llvm/trunk/test/CodeGen/AArch64/global_merge_1.ll
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/AArch64/global_merge_1.ll?rev=208933&view=auto
> ==============================================================================
> --- llvm/trunk/test/CodeGen/AArch64/global_merge_1.ll (original)
> +++ llvm/trunk/test/CodeGen/AArch64/global_merge_1.ll (removed)
> @@ -1,17 +0,0 @@
> -; RUN: llc < %s -mtriple=aarch64-none-linux-gnu | FileCheck %s
> -
> - at m = internal global i32 0, align 4
> - at n = internal global i32 0, align 4
> -
> -define void @f1(i32 %a1, i32 %a2) {
> -; CHECK-LABEL: f1:
> -; CHECK: adrp x{{[0-9]+}}, _MergedGlobals
> -; CHECK-NOT: adrp
> - store i32 %a1, i32* @m, align 4
> - store i32 %a2, i32* @n, align 4
> - ret void
> -}
> -
> -; CHECK: .local _MergedGlobals
> -; CHECK: .comm _MergedGlobals,8,8
> -
>
> Added: llvm/trunk/test/CodeGen/ARM/global-merge-1.ll
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/ARM/global-merge-1.ll?rev=208934&view=auto
> ==============================================================================
> --- llvm/trunk/test/CodeGen/ARM/global-merge-1.ll (added)
> +++ llvm/trunk/test/CodeGen/ARM/global-merge-1.ll Thu May 15 18:45:42 2014
> @@ -0,0 +1,85 @@
> +; RUN: llc %s -O0 -o - | FileCheck -check-prefix=NO-MERGE %s
> +; RUN: llc %s -O0 -o - -enable-global-merge=false | FileCheck -check-prefix=NO-MERGE %s
> +; RUN: llc %s -O0 -o - -enable-global-merge=true | FileCheck -check-prefix=NO-MERGE %s
> +; RUN: llc %s -O1 -o - | FileCheck -check-prefix=MERGE %s
> +; RUN: llc %s -O1 -o - -enable-global-merge=false | FileCheck -check-prefix=NO-MERGE %s
> +; RUN: llc %s -O1 -o - -enable-global-merge=true | FileCheck -check-prefix=MERGE %s
> +
> +; MERGE-NOT: .zerofill __DATA,__bss,_bar,20,2
> +; MERGE-NOT: .zerofill __DATA,__bss,_baz,20,2
> +; MERGE-NOT: .zerofill __DATA,__bss,_foo,20,2
> +; MERGE: .zerofill __DATA,__bss,__MergedGlobals,60,4
> +; MERGE-NOT: .zerofill __DATA,__bss,_bar,20,2
> +; MERGE-NOT: .zerofill __DATA,__bss,_baz,20,2
> +; MERGE-NOT: .zerofill __DATA,__bss,_foo,20,2
> +
> +; NO-MERGE-NOT: .zerofill __DATA,__bss,__MergedGlobals,60,4
> +; NO-MERGE: .zerofill __DATA,__bss,_bar,20,2
> +; NO-MERGE: .zerofill __DATA,__bss,_baz,20,2
> +; NO-MERGE: .zerofill __DATA,__bss,_foo,20,2
> +; NO-MERGE-NOT: .zerofill __DATA,__bss,__MergedGlobals,60,4
> +
> +target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:32:64-v128:32:128-a0:0:32-n32-S32"
> +target triple = "thumbv7-apple-ios3.0.0"
> +
> + at bar = internal global [5 x i32] zeroinitializer, align 4
> + at baz = internal global [5 x i32] zeroinitializer, align 4
> + at foo = internal global [5 x i32] zeroinitializer, align 4
> +
> +; Function Attrs: nounwind ssp
> +define internal void @initialize() #0 {
> + %1 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #3
> + store i32 %1, i32* getelementptr inbounds ([5 x i32]* @bar, i32 0, i32 0), align 4, !tbaa !1
> + %2 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #3
> + store i32 %2, i32* getelementptr inbounds ([5 x i32]* @baz, i32 0, i32 0), align 4, !tbaa !1
> + %3 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #3
> + store i32 %3, i32* getelementptr inbounds ([5 x i32]* @bar, i32 0, i32 1), align 4, !tbaa !1
> + %4 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #3
> + store i32 %4, i32* getelementptr inbounds ([5 x i32]* @baz, i32 0, i32 1), align 4, !tbaa !1
> + %5 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #3
> + store i32 %5, i32* getelementptr inbounds ([5 x i32]* @bar, i32 0, i32 2), align 4, !tbaa !1
> + %6 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #3
> + store i32 %6, i32* getelementptr inbounds ([5 x i32]* @baz, i32 0, i32 2), align 4, !tbaa !1
> + %7 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #3
> + store i32 %7, i32* getelementptr inbounds ([5 x i32]* @bar, i32 0, i32 3), align 4, !tbaa !1
> + %8 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #3
> + store i32 %8, i32* getelementptr inbounds ([5 x i32]* @baz, i32 0, i32 3), align 4, !tbaa !1
> + %9 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #3
> + store i32 %9, i32* getelementptr inbounds ([5 x i32]* @bar, i32 0, i32 4), align 4, !tbaa !1
> + %10 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #3
> + store i32 %10, i32* getelementptr inbounds ([5 x i32]* @baz, i32 0, i32 4), align 4, !tbaa !1
> + ret void
> +}
> +
> +declare i32 @calc(...) #1
> +
> +; Function Attrs: nounwind ssp
> +define internal void @calculate() #0 {
> + %1 = load <4 x i32>* bitcast ([5 x i32]* @bar to <4 x i32>*), align 4
> + %2 = load <4 x i32>* bitcast ([5 x i32]* @baz to <4 x i32>*), align 4
> + %3 = mul <4 x i32> %2, %1
> + store <4 x i32> %3, <4 x i32>* bitcast ([5 x i32]* @foo to <4 x i32>*), align 4
> + %4 = load i32* getelementptr inbounds ([5 x i32]* @bar, i32 0, i32 4), align 4, !tbaa !1
> + %5 = load i32* getelementptr inbounds ([5 x i32]* @baz, i32 0, i32 4), align 4, !tbaa !1
> + %6 = mul nsw i32 %5, %4
> + store i32 %6, i32* getelementptr inbounds ([5 x i32]* @foo, i32 0, i32 4), align 4, !tbaa !1
> + ret void
> +}
> +
> +; Function Attrs: nounwind readnone ssp
> +define internal i32* @returnFoo() #2 {
> + ret i32* getelementptr inbounds ([5 x i32]* @foo, i32 0, i32 0)
> +}
> +
> +attributes #0 = { nounwind ssp "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
> +attributes #1 = { "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
> +attributes #2 = { nounwind readnone ssp "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
> +attributes #3 = { nounwind }
> +
> +!llvm.ident = !{!0}
> +
> +!0 = metadata !{metadata !"LLVM version 3.4 "}
> +!1 = metadata !{metadata !2, metadata !2, i64 0}
> +!2 = metadata !{metadata !"int", metadata !3, i64 0}
> +!3 = metadata !{metadata !"omnipotent char", metadata !4, i64 0}
> +!4 = metadata !{metadata !"Simple C/C++ TBAA"}
>
> Added: llvm/trunk/test/CodeGen/ARM64/global-merge.ll
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/ARM64/global-merge.ll?rev=208934&view=auto
> ==============================================================================
> --- llvm/trunk/test/CodeGen/ARM64/global-merge.ll (added)
> +++ llvm/trunk/test/CodeGen/ARM64/global-merge.ll Thu May 15 18:45:42 2014
> @@ -0,0 +1,88 @@
> +; RUN: llc %s -O0 -o - | FileCheck -check-prefix=NO-MERGE %s
> +; RUN: llc %s -O0 -o - -enable-global-merge=false | FileCheck -check-prefix=NO-MERGE %s
> +; RUN: llc %s -O0 -o - -enable-global-merge=true | FileCheck -check-prefix=NO-MERGE %s
> +; RUN: llc %s -O1 -o - | FileCheck -check-prefix=MERGE %s
> +; RUN: llc %s -O1 -o - -enable-global-merge=false | FileCheck -check-prefix=NO-MERGE %s
> +; RUN: llc %s -O1 -o - -enable-global-merge=true | FileCheck -check-prefix=MERGE %s
> +
> +; MERGE-NOT: .zerofill __DATA,__bss,_bar,20,2
> +; MERGE-NOT: .zerofill __DATA,__bss,_baz,20,2
> +; MERGE-NOT: .zerofill __DATA,__bss,_foo,20,2
> +; MERGE: .zerofill __DATA,__bss,__MergedGlobals,60,4
> +; MERGE-NOT: .zerofill __DATA,__bss,_bar,20,2
> +; MERGE-NOT: .zerofill __DATA,__bss,_baz,20,2
> +; MERGE-NOT: .zerofill __DATA,__bss,_foo,20,2
> +
> +; NO-MERGE-NOT: .zerofill __DATA,__bss,__MergedGlobals,60,4
> +; NO-MERGE: .zerofill __DATA,__bss,_bar,20,2
> +; NO-MERGE: .zerofill __DATA,__bss,_baz,20,2
> +; NO-MERGE: .zerofill __DATA,__bss,_foo,20,2
> +; NO-MERGE-NOT: .zerofill __DATA,__bss,__MergedGlobals,60,4
> +
> +target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-n32:64-S128"
> +target triple = "arm64-apple-ios7.0.0"
> +
> + at bar = internal global [5 x i32] zeroinitializer, align 4
> + at baz = internal global [5 x i32] zeroinitializer, align 4
> + at foo = internal global [5 x i32] zeroinitializer, align 4
> +
> +; Function Attrs: nounwind ssp
> +define internal void @initialize() #0 {
> + %1 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #2
> + store i32 %1, i32* getelementptr inbounds ([5 x i32]* @bar, i64 0, i64 0), align 4
> + %2 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #2
> + store i32 %2, i32* getelementptr inbounds ([5 x i32]* @baz, i64 0, i64 0), align 4
> + %3 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #2
> + store i32 %3, i32* getelementptr inbounds ([5 x i32]* @bar, i64 0, i64 1), align 4
> + %4 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #2
> + store i32 %4, i32* getelementptr inbounds ([5 x i32]* @baz, i64 0, i64 1), align 4
> + %5 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #2
> + store i32 %5, i32* getelementptr inbounds ([5 x i32]* @bar, i64 0, i64 2), align 4
> + %6 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #2
> + store i32 %6, i32* getelementptr inbounds ([5 x i32]* @baz, i64 0, i64 2), align 4
> + %7 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #2
> + store i32 %7, i32* getelementptr inbounds ([5 x i32]* @bar, i64 0, i64 3), align 4
> + %8 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #2
> + store i32 %8, i32* getelementptr inbounds ([5 x i32]* @baz, i64 0, i64 3), align 4
> + %9 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #2
> + store i32 %9, i32* getelementptr inbounds ([5 x i32]* @bar, i64 0, i64 4), align 4
> + %10 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #2
> + store i32 %10, i32* getelementptr inbounds ([5 x i32]* @baz, i64 0, i64 4), align 4
> + ret void
> +}
> +
> +declare i32 @calc(...)
> +
> +; Function Attrs: nounwind ssp
> +define internal void @calculate() #0 {
> + %1 = load i32* getelementptr inbounds ([5 x i32]* @bar, i64 0, i64 0), align 4
> + %2 = load i32* getelementptr inbounds ([5 x i32]* @baz, i64 0, i64 0), align 4
> + %3 = mul nsw i32 %2, %1
> + store i32 %3, i32* getelementptr inbounds ([5 x i32]* @foo, i64 0, i64 0), align 4
> + %4 = load i32* getelementptr inbounds ([5 x i32]* @bar, i64 0, i64 1), align 4
> + %5 = load i32* getelementptr inbounds ([5 x i32]* @baz, i64 0, i64 1), align 4
> + %6 = mul nsw i32 %5, %4
> + store i32 %6, i32* getelementptr inbounds ([5 x i32]* @foo, i64 0, i64 1), align 4
> + %7 = load i32* getelementptr inbounds ([5 x i32]* @bar, i64 0, i64 2), align 4
> + %8 = load i32* getelementptr inbounds ([5 x i32]* @baz, i64 0, i64 2), align 4
> + %9 = mul nsw i32 %8, %7
> + store i32 %9, i32* getelementptr inbounds ([5 x i32]* @foo, i64 0, i64 2), align 4
> + %10 = load i32* getelementptr inbounds ([5 x i32]* @bar, i64 0, i64 3), align 4
> + %11 = load i32* getelementptr inbounds ([5 x i32]* @baz, i64 0, i64 3), align 4
> + %12 = mul nsw i32 %11, %10
> + store i32 %12, i32* getelementptr inbounds ([5 x i32]* @foo, i64 0, i64 3), align 4
> + %13 = load i32* getelementptr inbounds ([5 x i32]* @bar, i64 0, i64 4), align 4
> + %14 = load i32* getelementptr inbounds ([5 x i32]* @baz, i64 0, i64 4), align 4
> + %15 = mul nsw i32 %14, %13
> + store i32 %15, i32* getelementptr inbounds ([5 x i32]* @foo, i64 0, i64 4), align 4
> + ret void
> +}
> +
> +; Function Attrs: nounwind readnone ssp
> +define internal i32* @returnFoo() #1 {
> + ret i32* getelementptr inbounds ([5 x i32]* @foo, i64 0, i64 0)
> +}
> +
> +attributes #0 = { nounwind ssp }
> +attributes #1 = { nounwind readnone ssp }
> +attributes #2 = { nounwind }
>
> Added: llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-1.ll
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-1.ll?rev=208934&view=auto
> ==============================================================================
> --- llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-1.ll (added)
> +++ llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-1.ll Thu May 15 18:45:42 2014
> @@ -0,0 +1,22 @@
> +; RUN: opt %s -mtriple=aarch64-none-linux-gnu -global-merge -S -o - | FileCheck %s
> +; RUN: opt %s -mtriple=aarch64-none-linux-gnu -global-merge -global-merge-on-external -global-merge-aligned -S -o - | FileCheck %s
> +
> +; RUN: opt %s -mtriple=arm64-linux-gnuabi -global-merge -S -o - | FileCheck %s
> +; RUN: opt %s -mtriple=arm64-linux-gnuabi -global-merge -global-merge-on-external -global-merge-aligned -S -o - | FileCheck %s
> +
> +; RUN: opt %s -mtriple=arm64-apple-ios -global-merge -S -o - | FileCheck %s
> +; RUN: opt %s -mtriple=arm64-apple-ios -global-merge -global-merge-on-external -global-merge-aligned -S -o - | FileCheck %s
> +
> + at m = internal global i32 0, align 4
> + at n = internal global i32 0, align 4
> +
> +; CHECK: @_MergedGlobals = internal global { i32, i32 } zeroinitializer
> +
> +define void @f1(i32 %a1, i32 %a2) {
> +; CHECK-LABEL: @f1
> +; CHECK: getelementptr inbounds ({ i32, i32 }* @_MergedGlobals, i32 0, i32 0)
> +; CHECK: getelementptr inbounds ({ i32, i32 }* @_MergedGlobals, i32 0, i32 1)
> + store i32 %a1, i32* @m, align 4
> + store i32 %a2, i32* @n, align 4
> + ret void
> +}
>
> Added: llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-2.ll
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-2.ll?rev=208934&view=auto
> ==============================================================================
> --- llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-2.ll (added)
> +++ llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-2.ll Thu May 15 18:45:42 2014
> @@ -0,0 +1,30 @@
> +; RUN: opt %s -mtriple=aarch64-none-linux-gnu -global-merge -global-merge-on-external -global-merge-aligned -S -o - | FileCheck %s
> +; RUN: opt %s -mtriple=arm64-linux-gnuabi -global-merge -global-merge-on-external -global-merge-aligned -S -o - | FileCheck %s
> +; RUN: opt %s -mtriple=arm64-apple-ios -global-merge -global-merge-on-external -global-merge-aligned -S -o - | FileCheck %s
> +
> + at x = global i32 0, align 4
> + at y = global i32 0, align 4
> + at z = global i32 0, align 4
> +
> +; CHECK: @_MergedGlobals_x = global { i32, i32, i32 } zeroinitializer, align 16
> +; CHECK: @x = alias getelementptr inbounds ({ i32, i32, i32 }* @_MergedGlobals_x, i32 0, i32 0)
> +; CHECK: @y = alias getelementptr inbounds ({ i32, i32, i32 }* @_MergedGlobals_x, i32 0, i32 1)
> +; CHECK: @z = alias getelementptr inbounds ({ i32, i32, i32 }* @_MergedGlobals_x, i32 0, i32 2)
> +
> +define void @f1(i32 %a1, i32 %a2) {
> +; CHECK-LABEL: @f1
> +; CHECK: getelementptr inbounds ({ i32, i32, i32 }* @_MergedGlobals_x, i32 0, i32 0)
> +; CHECK: getelementptr inbounds ({ i32, i32, i32 }* @_MergedGlobals_x, i32 0, i32 1)
> + store i32 %a1, i32* @x, align 4
> + store i32 %a2, i32* @y, align 4
> + ret void
> +}
> +
> +define void @g1(i32 %a1, i32 %a2) {
> +; CHECK-LABEL: @g1
> +; CHECK: getelementptr inbounds ({ i32, i32, i32 }* @_MergedGlobals_x, i32 0, i32 1)
> +; CHECK: getelementptr inbounds ({ i32, i32, i32 }* @_MergedGlobals_x, i32 0, i32 2)
> + store i32 %a1, i32* @y, align 4
> + store i32 %a2, i32* @z, align 4
> + ret void
> +}
>
> Added: llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-3.ll
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-3.ll?rev=208934&view=auto
> ==============================================================================
> --- llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-3.ll (added)
> +++ llvm/trunk/test/Transforms/GlobalMerge/AArch64/global-merge-3.ll Thu May 15 18:45:42 2014
> @@ -0,0 +1,27 @@
> +; RUN: opt %s -mtriple=aarch64-none-linux-gnu -global-merge -global-merge-on-external -global-merge-aligned -S -o - | FileCheck %s
> +; RUN: opt %s -mtriple=arm64-linux-gnuabi -global-merge -global-merge-on-external -global-merge-aligned -S -o - | FileCheck %s
> +; RUN: opt %s -mtriple=arm64-apple-ios -global-merge -global-merge-on-external -global-merge-aligned -S -o - | FileCheck %s
> +
> + at x = global [1000 x i32] zeroinitializer, align 1
> + at y = global [1000 x i32] zeroinitializer, align 1
> + at z = internal global i32 1, align 4
> +
> +; CHECK: @_MergedGlobals_x = global { i32, [1000 x i32] } { i32 1, [1000 x i32] zeroinitializer }, align 4096
> +; CHECK: @_MergedGlobals_y = global { [1000 x i32] } zeroinitializer, align 4096
> +
> +; CHECK: @x = alias getelementptr inbounds ({ i32, [1000 x i32] }* @_MergedGlobals_x, i32 0, i32 1)
> +; CHECK: @y = alias getelementptr inbounds ({ [1000 x i32] }* @_MergedGlobals_y, i32 0, i32 0)
> +
> +define void @f1(i32 %a1, i32 %a2, i32 %a3) {
> +; CHECK-LABEL: @f1
> +; CHECK: %x3 = getelementptr inbounds [1000 x i32]* getelementptr inbounds ({ i32, [1000 x i32] }* @_MergedGlobals_x, i32 0, i32 1), i32 0, i64 3
> +; CHECK: %y3 = getelementptr inbounds [1000 x i32]* getelementptr inbounds ({ [1000 x i32] }* @_MergedGlobals_y, i32 0, i32 0), i32 0, i64 3
> +; CHECK: store i32 %a3, i32* getelementptr inbounds ({ i32, [1000 x i32] }* @_MergedGlobals_x, i32 0, i32 0), align 4
> +
> + %x3 = getelementptr inbounds [1000 x i32]* @x, i32 0, i64 3
> + %y3 = getelementptr inbounds [1000 x i32]* @y, i32 0, i64 3
> + store i32 %a1, i32* %x3, align 4
> + store i32 %a2, i32* %y3, align 4
> + store i32 %a3, i32* @z, align 4
> + ret void
> +}
>
> Added: llvm/trunk/test/Transforms/GlobalMerge/AArch64/lit.local.cfg
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/GlobalMerge/AArch64/lit.local.cfg?rev=208934&view=auto
> ==============================================================================
> --- llvm/trunk/test/Transforms/GlobalMerge/AArch64/lit.local.cfg (added)
> +++ llvm/trunk/test/Transforms/GlobalMerge/AArch64/lit.local.cfg Thu May 15 18:45:42 2014
> @@ -0,0 +1,4 @@
> +targets = set(config.root.targets_to_build.split())
> +if not 'AArch64' in targets:
> + config.unsupported = True
> +
>
> Modified: llvm/trunk/test/Transforms/GlobalMerge/ARM/arm.ll
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/GlobalMerge/ARM/arm.ll?rev=208934&r1=208933&r2=208934&view=diff
> ==============================================================================
> --- llvm/trunk/test/Transforms/GlobalMerge/ARM/arm.ll (original)
> +++ llvm/trunk/test/Transforms/GlobalMerge/ARM/arm.ll Thu May 15 18:45:42 2014
> @@ -1,23 +1,4 @@
> -; RUN: llc %s -O0 -o - | FileCheck -check-prefix=NO-MERGE %s
> -; RUN: llc %s -O0 -o - -global-merge=false | FileCheck -check-prefix=NO-MERGE %s
> -; RUN: llc %s -O0 -o - -global-merge=true | FileCheck -check-prefix=NO-MERGE %s
> -; RUN: llc %s -O1 -o - | FileCheck -check-prefix=MERGE %s
> -; RUN: llc %s -O1 -o - -global-merge=false | FileCheck -check-prefix=NO-MERGE %s
> -; RUN: llc %s -O1 -o - -global-merge=true | FileCheck -check-prefix=MERGE %s
> -
> -; MERGE-NOT: .zerofill __DATA,__bss,_bar,20,2
> -; MERGE-NOT: .zerofill __DATA,__bss,_baz,20,2
> -; MERGE-NOT: .zerofill __DATA,__bss,_foo,20,2
> -; MERGE: .zerofill __DATA,__bss,__MergedGlobals,60,4
> -; MERGE-NOT: .zerofill __DATA,__bss,_bar,20,2
> -; MERGE-NOT: .zerofill __DATA,__bss,_baz,20,2
> -; MERGE-NOT: .zerofill __DATA,__bss,_foo,20,2
> -
> -; NO-MERGE-NOT: .zerofill __DATA,__bss,__MergedGlobals,60,4
> -; NO-MERGE: .zerofill __DATA,__bss,_bar,20,2
> -; NO-MERGE: .zerofill __DATA,__bss,_baz,20,2
> -; NO-MERGE: .zerofill __DATA,__bss,_foo,20,2
> -; NO-MERGE-NOT: .zerofill __DATA,__bss,__MergedGlobals,60,4
> +; RUN: opt %s -mtriple=arm-linux-gnuabi -global-merge -S -o - | FileCheck %s
>
> target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:32:64-v128:32:128-a0:0:32-n32-S32"
> target triple = "thumbv7-apple-ios3.0.0"
> @@ -26,6 +7,8 @@ target triple = "thumbv7-apple-ios3.0.0"
> @baz = internal global [5 x i32] zeroinitializer, align 4
> @foo = internal global [5 x i32] zeroinitializer, align 4
>
> +; CHECK: @_MergedGlobals = internal global { [5 x i32], [5 x i32], [5 x i32] } zeroinitializer
> +
> ; Function Attrs: nounwind ssp
> define internal void @initialize() #0 {
> %1 = tail call i32 bitcast (i32 (...)* @calc to i32 ()*)() #3
>
> Modified: llvm/trunk/test/Transforms/GlobalMerge/ARM64/arm64.ll
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/GlobalMerge/ARM64/arm64.ll?rev=208934&r1=208933&r2=208934&view=diff
> ==============================================================================
> --- llvm/trunk/test/Transforms/GlobalMerge/ARM64/arm64.ll (original)
> +++ llvm/trunk/test/Transforms/GlobalMerge/ARM64/arm64.ll Thu May 15 18:45:42 2014
> @@ -1,23 +1,6 @@
> -; RUN: llc %s -O0 -o - | FileCheck -check-prefix=NO-MERGE %s
> -; RUN: llc %s -O0 -o - -global-merge=false | FileCheck -check-prefix=NO-MERGE %s
> -; RUN: llc %s -O0 -o - -global-merge=true | FileCheck -check-prefix=NO-MERGE %s
> -; RUN: llc %s -O1 -o - | FileCheck -check-prefix=MERGE %s
> -; RUN: llc %s -O1 -o - -global-merge=false | FileCheck -check-prefix=NO-MERGE %s
> -; RUN: llc %s -O1 -o - -global-merge=true | FileCheck -check-prefix=MERGE %s
> +; RUN: opt %s -mtriple=arm64-linux-gnuabi -global-merge -S -o - | FileCheck %s
>
> -; MERGE-NOT: .zerofill __DATA,__bss,_bar,20,2
> -; MERGE-NOT: .zerofill __DATA,__bss,_baz,20,2
> -; MERGE-NOT: .zerofill __DATA,__bss,_foo,20,2
> -; MERGE: .zerofill __DATA,__bss,__MergedGlobals,60,4
> -; MERGE-NOT: .zerofill __DATA,__bss,_bar,20,2
> -; MERGE-NOT: .zerofill __DATA,__bss,_baz,20,2
> -; MERGE-NOT: .zerofill __DATA,__bss,_foo,20,2
> -
> -; NO-MERGE-NOT: .zerofill __DATA,__bss,__MergedGlobals,60,4
> -; NO-MERGE: .zerofill __DATA,__bss,_bar,20,2
> -; NO-MERGE: .zerofill __DATA,__bss,_baz,20,2
> -; NO-MERGE: .zerofill __DATA,__bss,_foo,20,2
> -; NO-MERGE-NOT: .zerofill __DATA,__bss,__MergedGlobals,60,4
> +; CHECK: @_MergedGlobals = internal global { [5 x i32], [5 x i32], [5 x i32] } zeroinitializer
>
> target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-n32:64-S128"
> target triple = "arm64-apple-ios7.0.0"
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
More information about the llvm-commits
mailing list