diff options
85 files changed, 1390 insertions, 4255 deletions
diff --git a/lib/VMCore/AutoUpgrade.cpp b/lib/VMCore/AutoUpgrade.cpp index 9d4543d..3ffcfdc 100644 --- a/lib/VMCore/AutoUpgrade.cpp +++ b/lib/VMCore/AutoUpgrade.cpp @@ -42,248 +42,7 @@ static bool UpgradeIntrinsicFunction1(Function *F, Function *&NewFn) { Module *M = F->getParent(); switch (Name[5]) { default: break; - case 'a': - // This upgrades the llvm.atomic.lcs, llvm.atomic.las, llvm.atomic.lss, - // and atomics with default address spaces to their new names to their new - // function name (e.g. llvm.atomic.add.i32 => llvm.atomic.add.i32.p0i32) - if (Name.compare(5,7,"atomic.",7) == 0) { - if (Name.compare(12,3,"lcs",3) == 0) { - std::string::size_type delim = Name.find('.',12); - F->setName("llvm.atomic.cmp.swap" + Name.substr(delim) + - ".p0" + Name.substr(delim+1)); - NewFn = F; - return true; - } - else if (Name.compare(12,3,"las",3) == 0) { - std::string::size_type delim = Name.find('.',12); - F->setName("llvm.atomic.load.add"+Name.substr(delim) - + ".p0" + Name.substr(delim+1)); - NewFn = F; - return true; - } - else if (Name.compare(12,3,"lss",3) == 0) { - std::string::size_type delim = Name.find('.',12); - F->setName("llvm.atomic.load.sub"+Name.substr(delim) - + ".p0" + Name.substr(delim+1)); - NewFn = F; - return true; - } - else if (Name.rfind(".p") == std::string::npos) { - // We don't have an address space qualifier so this has be upgraded - // to the new name. Copy the type name at the end of the intrinsic - // and add to it - std::string::size_type delim = Name.find_last_of('.'); - assert(delim != std::string::npos && "can not find type"); - F->setName(Name + ".p0" + Name.substr(delim+1)); - NewFn = F; - return true; - } - } else if (Name.compare(5, 9, "arm.neon.", 9) == 0) { - if (((Name.compare(14, 5, "vmovl", 5) == 0 || - Name.compare(14, 5, "vaddl", 5) == 0 || - Name.compare(14, 5, "vsubl", 5) == 0 || - Name.compare(14, 5, "vaddw", 5) == 0 || - Name.compare(14, 5, "vsubw", 5) == 0 || - Name.compare(14, 5, "vmlal", 5) == 0 || - Name.compare(14, 5, "vmlsl", 5) == 0 || - Name.compare(14, 5, "vabdl", 5) == 0 || - Name.compare(14, 5, "vabal", 5) == 0) && - (Name.compare(19, 2, "s.", 2) == 0 || - Name.compare(19, 2, "u.", 2) == 0)) || - - (Name.compare(14, 4, "vaba", 4) == 0 && - (Name.compare(18, 2, "s.", 2) == 0 || - Name.compare(18, 2, "u.", 2) == 0)) || - - (Name.compare(14, 6, "vmovn.", 6) == 0)) { - - // Calls to these are transformed into IR without intrinsics. - NewFn = 0; - return true; - } - // Old versions of NEON ld/st intrinsics are missing alignment arguments. - bool isVLd = (Name.compare(14, 3, "vld", 3) == 0); - bool isVSt = (Name.compare(14, 3, "vst", 3) == 0); - if (isVLd || isVSt) { - unsigned NumVecs = Name.at(17) - '0'; - if (NumVecs == 0 || NumVecs > 4) - return false; - bool isLaneOp = (Name.compare(18, 5, "lane.", 5) == 0); - if (!isLaneOp && Name.at(18) != '.') - return false; - unsigned ExpectedArgs = 2; // for the address and alignment - if (isVSt || isLaneOp) - ExpectedArgs += NumVecs; - if (isLaneOp) - ExpectedArgs += 1; // for the lane number - unsigned NumP = FTy->getNumParams(); - if (NumP != ExpectedArgs - 1) - return false; - - // Change the name of the old (bad) intrinsic, because - // its type is incorrect, but we cannot overload that name. - F->setName(""); - - // One argument is missing: add the alignment argument. - std::vector<const Type*> NewParams; - for (unsigned p = 0; p < NumP; ++p) - NewParams.push_back(FTy->getParamType(p)); - NewParams.push_back(Type::getInt32Ty(F->getContext())); - FunctionType *NewFTy = FunctionType::get(FTy->getReturnType(), - NewParams, false); - NewFn = cast<Function>(M->getOrInsertFunction(Name, NewFTy)); - return true; - } - } - break; - case 'b': - // This upgrades the name of the llvm.bswap intrinsic function to only use - // a single type name for overloading. We only care about the old format - // 'llvm.bswap.i*.i*', so check for 'bswap.' and then for there being - // a '.' after 'bswap.' - if (Name.compare(5,6,"bswap.",6) == 0) { - std::string::size_type delim = Name.find('.',11); - - if (delim != std::string::npos) { - // Construct the new name as 'llvm.bswap' + '.i*' - F->setName(Name.substr(0,10)+Name.substr(delim)); - NewFn = F; - return true; - } - } - break; - - case 'c': - // We only want to fix the 'llvm.ct*' intrinsics which do not have the - // correct return type, so we check for the name, and then check if the - // return type does not match the parameter type. - if ( (Name.compare(5,5,"ctpop",5) == 0 || - Name.compare(5,4,"ctlz",4) == 0 || - Name.compare(5,4,"cttz",4) == 0) && - FTy->getReturnType() != FTy->getParamType(0)) { - // We first need to change the name of the old (bad) intrinsic, because - // its type is incorrect, but we cannot overload that name. We - // arbitrarily unique it here allowing us to construct a correctly named - // and typed function below. - F->setName(""); - - // Now construct the new intrinsic with the correct name and type. We - // leave the old function around in order to query its type, whatever it - // may be, and correctly convert up to the new type. - NewFn = cast<Function>(M->getOrInsertFunction(Name, - FTy->getParamType(0), - FTy->getParamType(0), - (Type *)0)); - return true; - } - break; - - case 'e': - // The old llvm.eh.selector.i32 is equivalent to the new llvm.eh.selector. - if (Name.compare("llvm.eh.selector.i32") == 0) { - F->setName("llvm.eh.selector"); - NewFn = F; - return true; - } - // The old llvm.eh.typeid.for.i32 is equivalent to llvm.eh.typeid.for. - if (Name.compare("llvm.eh.typeid.for.i32") == 0) { - F->setName("llvm.eh.typeid.for"); - NewFn = F; - return true; - } - // Convert the old llvm.eh.selector.i64 to a call to llvm.eh.selector. - if (Name.compare("llvm.eh.selector.i64") == 0) { - NewFn = Intrinsic::getDeclaration(M, Intrinsic::eh_selector); - return true; - } - // Convert the old llvm.eh.typeid.for.i64 to a call to llvm.eh.typeid.for. - if (Name.compare("llvm.eh.typeid.for.i64") == 0) { - NewFn = Intrinsic::getDeclaration(M, Intrinsic::eh_typeid_for); - return true; - } - break; - - case 'm': { - // This upgrades the llvm.memcpy, llvm.memmove, and llvm.memset to the - // new format that allows overloading the pointer for different address - // space (e.g., llvm.memcpy.i16 => llvm.memcpy.p0i8.p0i8.i16) - const char* NewFnName = NULL; - if (Name.compare(5,8,"memcpy.i",8) == 0) { - if (Name[13] == '8') - NewFnName = "llvm.memcpy.p0i8.p0i8.i8"; - else if (Name.compare(13,2,"16") == 0) - NewFnName = "llvm.memcpy.p0i8.p0i8.i16"; - else if (Name.compare(13,2,"32") == 0) - NewFnName = "llvm.memcpy.p0i8.p0i8.i32"; - else if (Name.compare(13,2,"64") == 0) - NewFnName = "llvm.memcpy.p0i8.p0i8.i64"; - } else if (Name.compare(5,9,"memmove.i",9) == 0) { - if (Name[14] == '8') - NewFnName = "llvm.memmove.p0i8.p0i8.i8"; - else if (Name.compare(14,2,"16") == 0) - NewFnName = "llvm.memmove.p0i8.p0i8.i16"; - else if (Name.compare(14,2,"32") == 0) - NewFnName = "llvm.memmove.p0i8.p0i8.i32"; - else if (Name.compare(14,2,"64") == 0) - NewFnName = "llvm.memmove.p0i8.p0i8.i64"; - } - else if (Name.compare(5,8,"memset.i",8) == 0) { - if (Name[13] == '8') - NewFnName = "llvm.memset.p0i8.i8"; - else if (Name.compare(13,2,"16") == 0) - NewFnName = "llvm.memset.p0i8.i16"; - else if (Name.compare(13,2,"32") == 0) - NewFnName = "llvm.memset.p0i8.i32"; - else if (Name.compare(13,2,"64") == 0) - NewFnName = "llvm.memset.p0i8.i64"; - } - if (NewFnName) { - NewFn = cast<Function>(M->getOrInsertFunction(NewFnName, - FTy->getReturnType(), - FTy->getParamType(0), - FTy->getParamType(1), - FTy->getParamType(2), - FTy->getParamType(3), - Type::getInt1Ty(F->getContext()), - (Type *)0)); - return true; - } - break; - } case 'p': - // This upgrades the llvm.part.select overloaded intrinsic names to only - // use one type specifier in the name. We only care about the old format - // 'llvm.part.select.i*.i*', and solve as above with bswap. - if (Name.compare(5,12,"part.select.",12) == 0) { - std::string::size_type delim = Name.find('.',17); - - if (delim != std::string::npos) { - // Construct a new name as 'llvm.part.select' + '.i*' - F->setName(Name.substr(0,16)+Name.substr(delim)); - NewFn = F; - return true; - } - break; - } - - // This upgrades the llvm.part.set intrinsics similarly as above, however - // we care about 'llvm.part.set.i*.i*.i*', but only the first two types - // must match. There is an additional type specifier after these two - // matching types that we must retain when upgrading. Thus, we require - // finding 2 periods, not just one, after the intrinsic name. - if (Name.compare(5,9,"part.set.",9) == 0) { - std::string::size_type delim = Name.find('.',14); - - if (delim != std::string::npos && - Name.find('.',delim+1) != std::string::npos) { - // Construct a new name as 'llvm.part.select' + '.i*.i*' - F->setName(Name.substr(0,13)+Name.substr(delim)); - NewFn = F; - return true; - } - break; - } - // This upgrades the llvm.prefetch intrinsic to accept one more parameter, // which is a instruction / data cache identifier. The old version only // implicitly accepted the data version. @@ -336,270 +95,21 @@ static bool UpgradeIntrinsicFunction1(Function *F, Function *&NewFn) { } } - // This fixes all MMX shift intrinsic instructions to take a - // x86_mmx instead of a v1i64, v2i32, v4i16, or v8i8. - if (Name.compare(5, 8, "x86.mmx.", 8) == 0) { - const Type *X86_MMXTy = VectorType::getX86_MMXTy(FTy->getContext()); - - if (Name.compare(13, 4, "padd", 4) == 0 || - Name.compare(13, 4, "psub", 4) == 0 || - Name.compare(13, 4, "pmul", 4) == 0 || - Name.compare(13, 5, "pmadd", 5) == 0 || - Name.compare(13, 4, "pand", 4) == 0 || - Name.compare(13, 3, "por", 3) == 0 || - Name.compare(13, 4, "pxor", 4) == 0 || - Name.compare(13, 4, "pavg", 4) == 0 || - Name.compare(13, 4, "pmax", 4) == 0 || - Name.compare(13, 4, "pmin", 4) == 0 || - Name.compare(13, 4, "psad", 4) == 0 || - Name.compare(13, 4, "psll", 4) == 0 || - Name.compare(13, 4, "psrl", 4) == 0 || - Name.compare(13, 4, "psra", 4) == 0 || - Name.compare(13, 4, "pack", 4) == 0 || - Name.compare(13, 6, "punpck", 6) == 0 || - Name.compare(13, 4, "pcmp", 4) == 0) { - assert(FTy->getNumParams() == 2 && "MMX intrinsic takes 2 args!"); - const Type *SecondParamTy = X86_MMXTy; - - if (Name.compare(13, 5, "pslli", 5) == 0 || - Name.compare(13, 5, "psrli", 5) == 0 || - Name.compare(13, 5, "psrai", 5) == 0) - SecondParamTy = FTy->getParamType(1); - - // Don't do anything if it has the correct types. - if (FTy->getReturnType() == X86_MMXTy && - FTy->getParamType(0) == X86_MMXTy && - FTy->getParamType(1) == SecondParamTy) - break; - - // We first need to change the name of the old (bad) intrinsic, because - // its type is incorrect, but we cannot overload that name. We - // arbitrarily unique it here allowing us to construct a correctly named - // and typed function below. - F->setName(""); - - // Now construct the new intrinsic with the correct name and type. We - // leave the old function around in order to query its type, whatever it - // may be, and correctly convert up to the new type. - NewFn = cast<Function>(M->getOrInsertFunction(Name, - X86_MMXTy, X86_MMXTy, - SecondParamTy, (Type*)0)); - return true; - } - - if (Name.compare(13, 8, "maskmovq", 8) == 0) { - // Don't do anything if it has the correct types. - if (FTy->getParamType(0) == X86_MMXTy && - FTy->getParamType(1) == X86_MMXTy) - break; - - F->setName(""); - NewFn = cast<Function>(M->getOrInsertFunction(Name, - FTy->getReturnType(), - X86_MMXTy, - X86_MMXTy, - FTy->getParamType(2), - (Type*)0)); - return true; - } - - if (Name.compare(13, 8, "pmovmskb", 8) == 0) { - if (FTy->getParamType(0) == X86_MMXTy) - break; - - F->setName(""); - NewFn = cast<Function>(M->getOrInsertFunction(Name, - FTy->getReturnType(), - X86_MMXTy, - (Type*)0)); - return true; - } - - if (Name.compare(13, 5, "movnt", 5) == 0) { - if (FTy->getParamType(1) == X86_MMXTy) - break; - - F->setName(""); - NewFn = cast<Function>(M->getOrInsertFunction(Name, - FTy->getReturnType(), - FTy->getParamType(0), - X86_MMXTy, - (Type*)0)); - return true; - } - - if (Name.compare(13, 7, "palignr", 7) == 0) { - if (FTy->getReturnType() == X86_MMXTy && - FTy->getParamType(0) == X86_MMXTy && - FTy->getParamType(1) == X86_MMXTy) - break; - - F->setName(""); - NewFn = cast<Function>(M->getOrInsertFunction(Name, - X86_MMXTy, - X86_MMXTy, - X86_MMXTy, - FTy->getParamType(2), - (Type*)0)); - return true; - } - - if (Name.compare(13, 5, "pextr", 5) == 0) { - if (FTy->getParamType(0) == X86_MMXTy) - break; - - F->setName(""); - NewFn = cast<Function>(M->getOrInsertFunction(Name, - FTy->getReturnType(), - X86_MMXTy, - FTy->getParamType(1), - (Type*)0)); - return true; - } - - if (Name.compare(13, 5, "pinsr", 5) == 0) { - if (FTy->getReturnType() == X86_MMXTy && - FTy->getParamType(0) == X86_MMXTy) - break; - - F->setName(""); - NewFn = cast<Function>(M->getOrInsertFunction(Name, - X86_MMXTy, - X86_MMXTy, - FTy->getParamType(1), - FTy->getParamType(2), - (Type*)0)); - return true; - } - - if (Name.compare(13, 12, "cvtsi32.si64", 12) == 0) { - if (FTy->getReturnType() == X86_MMXTy) - break; - - F->setName(""); - NewFn = cast<Function>(M->getOrInsertFunction(Name, - X86_MMXTy, - FTy->getParamType(0), - (Type*)0)); - return true; - } - - if (Name.compare(13, 12, "cvtsi64.si32", 12) == 0) { - if (FTy->getParamType(0) == X86_MMXTy) - break; - - F->setName(""); - NewFn = cast<Function>(M->getOrInsertFunction(Name, - FTy->getReturnType(), - X86_MMXTy, - (Type*)0)); - return true; - } - - if (Name.compare(13, 8, "vec.init", 8) == 0) { - if (FTy->getReturnType() == X86_MMXTy) - break; - - F->setName(""); - - if (Name.compare(21, 2, ".b", 2) == 0) - NewFn = cast<Function>(M->getOrInsertFunction(Name, - X86_MMXTy, - FTy->getParamType(0), - FTy->getParamType(1), - FTy->getParamType(2), - FTy->getParamType(3), - FTy->getParamType(4), - FTy->getParamType(5), - FTy->getParamType(6), - FTy->getParamType(7), - (Type*)0)); - else if (Name.compare(21, 2, ".w", 2) == 0) - NewFn = cast<Function>(M->getOrInsertFunction(Name, - X86_MMXTy, - FTy->getParamType(0), - FTy->getParamType(1), - FTy->getParamType(2), - FTy->getParamType(3), - (Type*)0)); - else if (Name.compare(21, 2, ".d", 2) == 0) - NewFn = cast<Function>(M->getOrInsertFunction(Name, - X86_MMXTy, - FTy->getParamType(0), - FTy->getParamType(1), - (Type*)0)); - return true; - } - - - if (Name.compare(13, 9, "vec.ext.d", 9) == 0) { - if (FTy->getReturnType() == X86_MMXTy && - FTy->getParamType(0) == X86_MMXTy) - break; - - F->setName(""); - NewFn = cast<Function>(M->getOrInsertFunction(Name, - X86_MMXTy, - X86_MMXTy, - FTy->getParamType(1), - (Type*)0)); - return true; - } - - if (Name.compare(13, 9, "emms", 4) == 0 || - Name.compare(13, 9, "femms", 5) == 0) { - NewFn = 0; - break; - } - - // We really shouldn't get here ever. - assert(0 && "Invalid MMX intrinsic!"); - break; - } else if (Name.compare(5,17,"x86.sse2.loadh.pd",17) == 0 || - Name.compare(5,17,"x86.sse2.loadl.pd",17) == 0 || - Name.compare(5,16,"x86.sse2.movl.dq",16) == 0 || - Name.compare(5,15,"x86.sse2.movs.d",15) == 0 || - Name.compare(5,16,"x86.sse2.shuf.pd",16) == 0 || - Name.compare(5,18,"x86.sse2.unpckh.pd",18) == 0 || - Name.compare(5,18,"x86.sse2.unpckl.pd",18) == 0 || - Name.compare(5,20,"x86.sse2.punpckh.qdq",20) == 0 || - Name.compare(5,20,"x86.sse2.punpckl.qdq",20) == 0) { - // Calls to these intrinsics are transformed into ShuffleVector's. - NewFn = 0; - return true; - } else if (Name.compare(5, 16, "x86.sse41.pmulld", 16) == 0) { - // Calls to these intrinsics are transformed into vector multiplies. - NewFn = 0; - return true; - } else if (Name.compare(5, 18, "x86.ssse3.palign.r", 18) == 0 || - Name.compare(5, 22, "x86.ssse3.palign.r.128", 22) == 0) { - // Calls to these intrinsics are transformed into vector shuffles, shifts, - // or 0. - NewFn = 0; - return true; - } else if (Name.compare(5, 16, "x86.sse.loadu.ps", 16) == 0 || - Name.compare(5, 17, "x86.sse2.loadu.dq", 17) == 0 || - Name.compare(5, 17, "x86.sse2.loadu.pd", 17) == 0) { + if (Name.compare(5, 16, "x86.sse.loadu.ps", 16) == 0 || + Name.compare(5, 17, "x86.sse2.loadu.dq", 17) == 0 || + Name.compare(5, 17, "x86.sse2.loadu.pd", 17) == 0) { // Calls to these instructions are transformed into unaligned loads. NewFn = 0; return true; - } else if (Name.compare(5, 16, "x86.sse.movnt.ps", 16) == 0 || - Name.compare(5, 17, "x86.sse2.movnt.dq", 17) == 0 || - Name.compare(5, 17, "x86.sse2.movnt.pd", 17) == 0 || - Name.compare(5, 17, "x86.sse2.movnt.i", 16) == 0) { + } + + if (Name.compare(5, 16, "x86.sse.movnt.ps", 16) == 0 || + Name.compare(5, 17, "x86.sse2.movnt.dq", 17) == 0 || + Name.compare(5, 17, "x86.sse2.movnt.pd", 17) == 0 || + Name.compare(5, 17, "x86.sse2.movnt.i", 16) == 0) { // Calls to these instructions are transformed into nontemporal stores. NewFn = 0; return true; - } else if (Name.compare(5, 17, "x86.ssse3.pshuf.w", 17) == 0) { - // This is an SSE/MMX instruction. - const Type *X86_MMXTy = VectorType::getX86_MMXTy(FTy->getContext()); - NewFn = - cast<Function>(M->getOrInsertFunction("llvm.x86.sse.pshuf.w", - X86_MMXTy, - X86_MMXTy, - Type::getInt8Ty(F->getContext()), - (Type*)0)); - return true; } break; @@ -625,105 +135,10 @@ bool llvm::UpgradeIntrinsicFunction(Function *F, Function *&NewFn) { } bool llvm::UpgradeGlobalVariable(GlobalVariable *GV) { - StringRef Name(GV->getName()); - - // We are only upgrading one symbol here. - if (Name == ".llvm.eh.catch.all.value") { - GV->setName("llvm.eh.catch.all.value"); - return true; - } - + // Nothing to do yet. return false; } -/// ExtendNEONArgs - For NEON "long" and "wide" operations, where the results -/// have vector elements twice as big as one or both source operands, do the -/// sign- or zero-extension that used to be handled by intrinsics. The -/// extended values are returned via V0 and V1. -static void ExtendNEONArgs(CallInst *CI, Value *Arg0, Value *Arg1, - Value *&V0, Value *&V1) { - Function *F = CI->getCalledFunction(); - const std::string& Name = F->getName(); - bool isLong = (Name.at(18) == 'l'); - bool isSigned = (Name.at(19) == 's'); - - if (isSigned) { - if (isLong) - V0 = new SExtInst(Arg0, CI->getType(), "", CI); - else - V0 = Arg0; - V1 = new SExtInst(Arg1, CI->getType(), "", CI); - } else { - if (isLong) - V0 = new ZExtInst(Arg0, CI->getType(), "", CI); - else - V0 = Arg0; - V1 = new ZExtInst(Arg1, CI->getType(), "", CI); - } -} - -/// CallVABD - As part of expanding a call to one of the old NEON vabdl, vaba, -/// or vabal intrinsics, construct a call to a vabd intrinsic. Examine the -/// name of the old intrinsic to determine whether to use a signed or unsigned -/// vabd intrinsic. Get the type from the old call instruction, adjusted for -/// half-size vector elements if the old intrinsic was vabdl or vabal. -static Instruction *CallVABD(CallInst *CI, Value *Arg0, Value *Arg1) { - Function *F = CI->getCalledFunction(); - const std::string& Name = F->getName(); - bool isLong = (Name.at(18) == 'l'); - bool isSigned = (Name.at(isLong ? 19 : 18) == 's'); - - Intrinsic::ID intID; - if (isSigned) - intID = Intrinsic::arm_neon_vabds; - else - intID = Intrinsic::arm_neon_vabdu; - - const Type *Ty = CI->getType(); - if (isLong) - Ty = VectorType::getTruncatedElementVectorType(cast<const VectorType>(Ty)); - - Function *VABD = Intrinsic::getDeclaration(F->getParent(), intID, &Ty, 1); - Value *Operands[2]; - Operands[0] = Arg0; - Operands[1] = Arg1; - return CallInst::Create(VABD, Operands, Operands+2, - "upgraded."+CI->getName(), CI); -} - -/// ConstructNewCallInst - Construct a new CallInst with the signature of NewFn. -static void ConstructNewCallInst(Function *NewFn, CallInst *OldCI, - Value **Operands, unsigned NumOps, - bool AssignName = true) { - // Construct a new CallInst. - CallInst *NewCI = - CallInst::Create(NewFn, Operands, Operands + NumOps, - AssignName ? "upgraded." + OldCI->getName() : "", OldCI); - - NewCI->setTailCall(OldCI->isTailCall()); - NewCI->setCallingConv(OldCI->getCallingConv()); - - // Handle any uses of the old CallInst. If the type has changed, add a cast. - if (!OldCI->use_empty()) { - if (OldCI->getType() != NewCI->getType()) { - Function *OldFn = OldCI->getCalledFunction(); - CastInst *RetCast = - CastInst::Create(CastInst::getCastOpcode(NewCI, true, - OldFn->getReturnType(), true), - NewCI, OldFn->getReturnType(), NewCI->getName(),OldCI); - - // Replace all uses of the old call with the new cast which has the - // correct type. - OldCI->replaceAllUsesWith(RetCast); - } else { - OldCI->replaceAllUsesWith(NewCI); - } - } - - // Clean up the old call now that it has been completely upgraded. - OldCI->eraseFromParent(); -} - // UpgradeIntrinsicCall - Upgrade a call to an old intrinsic to be a call the // upgraded intrinsic. All argument and return casting must be provided in // order to seamlessly integrate with existing context. @@ -735,284 +150,9 @@ void llvm::UpgradeIntrinsicCall(CallInst *CI, Function *NewFn) { assert(F && "CallInst has no function associated with it."); if (!NewFn) { - // Get the Function's name. - const std::string& Name = F->getName(); - - // Upgrade ARM NEON intrinsics. - if (Name.compare(5, 9, "arm.neon.", 9) == 0) { - Instruction *NewI; - Value *V0, *V1; - if (Name.compare(14, 7, "vmovls.", 7) == 0) { - NewI = new SExtInst(CI->getArgOperand(0), CI->getType(), - "upgraded." + CI->getName(), CI); - } else if (Name.compare(14, 7, "vmovlu.", 7) == 0) { - NewI = new ZExtInst(CI->getArgOperand(0), CI->getType(), - "upgraded." + CI->getName(), CI); - } else if (Name.compare(14, 4, "vadd", 4) == 0) { - ExtendNEONArgs(CI, CI->getArgOperand(0), CI->getArgOperand(1), V0, V1); - NewI = BinaryOperator::CreateAdd(V0, V1, "upgraded."+CI->getName(), CI); - } else if (Name.compare(14, 4, "vsub", 4) == 0) { - ExtendNEONArgs(CI, CI->getArgOperand(0), CI->getArgOperand(1), V0, V1); - NewI = BinaryOperator::CreateSub(V0, V1,"upgraded."+CI->getName(),CI); - } else if (Name.compare(14, 4, "vmul", 4) == 0) { - ExtendNEONArgs(CI, CI->getArgOperand(0), CI->getArgOperand(1), V0, V1); - NewI = BinaryOperator::CreateMul(V0, V1,"upgraded."+CI->getName(),CI); - } else if (Name.compare(14, 4, "vmla", 4) == 0) { - ExtendNEONArgs(CI, CI->getArgOperand(1), CI->getArgOperand(2), V0, V1); - Instruction *MulI = BinaryOperator::CreateMul(V0, V1, "", CI); - NewI = BinaryOperator::CreateAdd(CI->getArgOperand(0), MulI, - "upgraded."+CI->getName(), CI); - } else if (Name.compare(14, 4, "vmls", 4) == 0) { - ExtendNEONArgs(CI, CI->getArgOperand(1), CI->getArgOperand(2), V0, V1); - Instruction *MulI = BinaryOperator::CreateMul(V0, V1, "", CI); - NewI = BinaryOperator::CreateSub(CI->getArgOperand(0), MulI, - "upgraded."+CI->getName(), CI); - } else if (Name.compare(14, 4, "vabd", 4) == 0) { - NewI = CallVABD(CI, CI->getArgOperand(0), CI->getArgOperand(1)); - NewI = new ZExtInst(NewI, CI->getType(), "upgraded."+CI->getName(), CI); - } else if (Name.compare(14, 4, "vaba", 4) == 0) { - NewI = CallVABD(CI, CI->getArgOperand(1), CI->getArgOperand(2)); - if (Name.at(18) == 'l') - NewI = new ZExtInst(NewI, CI->getType(), "", CI); - NewI = BinaryOperator::CreateAdd(CI->getArgOperand(0), NewI, - "upgraded."+CI->getName(), CI); - } else if (Name.compare(14, 6, "vmovn.", 6) == 0) { - NewI = new TruncInst(CI->getArgOperand(0), CI->getType(), - "upgraded." + CI->getName(), CI); - } else { - llvm_unreachable("Unknown arm.neon function for CallInst upgrade."); - } - // Replace any uses of the old CallInst. - if (!CI->use_empty()) - CI->replaceAllUsesWith(NewI); - CI->eraseFromParent(); - return; - } - - bool isLoadH = false, isLoadL = false, isMovL = false; - bool isMovSD = false, isShufPD = false; - bool isUnpckhPD = false, isUnpcklPD = false; - bool isPunpckhQPD = false, isPunpcklQPD = false; - if (F->getName() == "llvm.x86.sse2.loadh.pd") - isLoadH = true; - else if (F->getName() == "llvm.x86.sse2.loadl.pd") - isLoadL = true; - else if (F->getName() == "llvm.x86.sse2.movl.dq") - isMovL = true; - else if (F->getName() == "llvm.x86.sse2.movs.d") - isMovSD = true; - else if (F->getName() == "llvm.x86.sse2.shuf.pd") - isShufPD = true; - else if (F->getName() == "llvm.x86.sse2.unpckh.pd") - isUnpckhPD = true; - else if (F->getName() == "llvm.x86.sse2.unpckl.pd") - isUnpcklPD = true; - else if (F->getName() == "llvm.x86.sse2.punpckh.qdq") - isPunpckhQPD = true; - else if (F->getName() == "llvm.x86.sse2.punpckl.qdq") - isPunpcklQPD = true; - - if (isLoadH || isLoadL || isMovL || isMovSD || isShufPD || - isUnpckhPD || isUnpcklPD || isPunpckhQPD || isPunpcklQPD) { - std::vector<Constant*> Idxs; - Value *Op0 = CI->getArgOperand(0); - ShuffleVectorInst *SI = NULL; - if (isLoadH || isLoadL) { - Value *Op1 = UndefValue::get(Op0->getType()); - Value *Addr = new BitCastInst(CI->getArgOperand(1), - Type::getDoublePtrTy(C), - "upgraded.", CI); - Value *Load = new LoadInst(Addr, "upgraded.", false, 8, CI); - Value *Idx = ConstantInt::get(Type::getInt32Ty(C), 0); - Op1 = InsertElementInst::Create(Op1, Load, Idx, "upgraded.", CI); - - if (isLoadH) { - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), 0)); - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), 2)); - } else { - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), 2)); - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), 1)); - } - Value *Mask = ConstantVector::get(Idxs); - SI = new ShuffleVectorInst(Op0, Op1, Mask, "upgraded.", CI); - } else if (isMovL) { - Constant *Zero = ConstantInt::get(Type::getInt32Ty(C), 0); - Idxs.push_back(Zero); - Idxs.push_back(Zero); - Idxs.push_back(Zero); - Idxs.push_back(Zero); - Value *ZeroV = ConstantVector::get(Idxs); - - Idxs.clear(); - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), 4)); - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), 5)); - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), 2)); - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), 3)); - Value *Mask = ConstantVector::get(Idxs); - SI = new ShuffleVectorInst(ZeroV, Op0, Mask, "upgraded.", CI); - } else if (isMovSD || - isUnpckhPD || isUnpcklPD || isPunpckhQPD || isPunpcklQPD) { - Value *Op1 = CI->getArgOperand(1); - if (isMovSD) { - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), 2)); - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), 1)); - } else if (isUnpckhPD || isPunpckhQPD) { - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), 1)); - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), 3)); - } else { - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), 0)); - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), 2)); - } - Value *Mask = ConstantVector::get(Idxs); - SI = new ShuffleVectorInst(Op0, Op1, Mask, "upgraded.", CI); - } else if (isShufPD) { - Value *Op1 = CI->getArgOperand(1); - unsigned MaskVal = - cast<ConstantInt>(CI->getArgOperand(2))->getZExtValue(); - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), MaskVal & 1)); - Idxs.push_back(ConstantInt::get(Type::getInt32Ty(C), - ((MaskVal >> 1) & 1)+2)); - Value *Mask = ConstantVector::get(Idxs); - SI = new ShuffleVectorInst(Op0, Op1, Mask, "upgraded.", CI); - } - - assert(SI && "Unexpected!"); - - // Handle any uses of the old CallInst. - if (!CI->use_empty()) - // Replace all uses of the old call with the new cast which has the - // correct type. - CI->replaceAllUsesWith(SI); - - // Clean up the old call now that it has been completely upgraded. - CI->eraseFromParent(); - } else if (F->getName() == "llvm.x86.sse41.pmulld") { - // Upgrade this set of intrinsics into vector multiplies. - Instruction *Mul = BinaryOperator::CreateMul(CI->getArgOperand(0), - CI->getArgOperand(1), - CI->getName(), - CI); - // Fix up all the uses with our new multiply. - if (!CI->use_empty()) - CI->replaceAllUsesWith(Mul); - - // Remove upgraded multiply. - CI->eraseFromParent(); - } else if (F->getName() == "llvm.x86.ssse3.palign.r") { - Value *Op1 = CI->getArgOperand(0); - Value *Op2 = CI->getArgOperand(1); - Value *Op3 = CI->getArgOperand(2); - unsigned shiftVal = cast<ConstantInt>(Op3)->getZExtValue(); - Value *Rep; - IRBuilder<> Builder(C); - Builder.SetInsertPoint(CI->getParent(), CI); - - // If palignr is shifting the pair of input vectors less than 9 bytes, - // emit a shuffle instruction. - if (shiftVal <= 8) { - const Type *IntTy = Type::getInt32Ty(C); - const Type *EltTy = Type::getInt8Ty(C); - const Type *VecTy = VectorType::get(EltTy, 8); - - Op2 = Builder.CreateBitCast(Op2, VecTy); - Op1 = Builder.CreateBitCast(Op1, VecTy); - - llvm::SmallVector<llvm::Constant*, 8> Indices; - for (unsigned i = 0; i != 8; ++i) - Indices.push_back(ConstantInt::get(IntTy, shiftVal + i)); - - Value *SV = ConstantVector::get(Indices); - Rep = Builder.CreateShuffleVector(Op2, Op1, SV, "palignr"); - Rep = Builder.CreateBitCast(Rep, F->getReturnType()); - } - - // If palignr is shifting the pair of input vectors more than 8 but less - // than 16 bytes, emit a logical right shift of the destination. - else if (shiftVal < 16) { - // MMX has these as 1 x i64 vectors for some odd optimization reasons. - const Type *EltTy = Type::getInt64Ty(C); - const Type *VecTy = VectorType::get(EltTy, 1); - - Op1 = Builder.CreateBitCast(Op1, VecTy, "cast"); - Op2 = ConstantInt::get(VecTy, (shiftVal-8) * 8); - - // create i32 constant - Function *I = - Intrinsic::getDeclaration(F->getParent(), Intrinsic::x86_mmx_psrl_q); - Rep = Builder.CreateCall2(I, Op1, Op2, "palignr"); - } - - // If palignr is shifting the pair of vectors more than 32 bytes, emit zero. - else { - Rep = Constant::getNullValue(F->getReturnType()); - } - - // Replace any uses with our new instruction. - if (!CI->use_empty()) - CI->replaceAllUsesWith(Rep); - - // Remove upgraded instruction. - CI->eraseFromParent(); - - } else if (F->getName() == "llvm.x86.ssse3.palign.r.128") { - Value *Op1 = CI->getArgOperand(0); - Value *Op2 = CI->getArgOperand(1); - Value *Op3 = CI->getArgOperand(2); - unsigned shiftVal = cast<ConstantInt>(Op3)->getZExtValue(); - Value *Rep; - IRBuilder<> Builder(C); - Builder.SetInsertPoint(CI->getParent(), CI); - - // If palignr is shifting the pair of input vectors less than 17 bytes, - // emit a shuffle instruction. - if (shiftVal <= 16) { - const Type *IntTy = Type::getInt32Ty(C); - const Type *EltTy = Type::getInt8Ty(C); - const Type *VecTy = VectorType::get(EltTy, 16); - - Op2 = Builder.CreateBitCast(Op2, VecTy); - Op1 = Builder.CreateBitCast(Op1, VecTy); - - llvm::SmallVector<llvm::Constant*, 16> Indices; - for (unsigned i = 0; i != 16; ++i) - Indices.push_back(ConstantInt::get(IntTy, shiftVal + i)); - - Value *SV = ConstantVector::get(Indices); - Rep = Builder.CreateShuffleVector(Op2, Op1, SV, "palignr"); - Rep = Builder.CreateBitCast(Rep, F->getReturnType()); - } - - // If palignr is shifting the pair of input vectors more than 16 but less - // than 32 bytes, emit a logical right shift of the destination. - else if (shiftVal < 32) { - const Type *EltTy = Type::getInt64Ty(C); - const Type *VecTy = VectorType::get(EltTy, 2); - const Type *IntTy = Type::getInt32Ty(C); - - Op1 = Builder.CreateBitCast(Op1, VecTy, "cast"); - Op2 = ConstantInt::get(IntTy, (shiftVal-16) * 8); - - // create i32 constant - Function *I = - Intrinsic::getDeclaration(F->getParent(), Intrinsic::x86_sse2_psrl_dq); - Rep = Builder.CreateCall2(I, Op1, Op2, "palignr"); - } - - // If palignr is shifting the pair of vectors more than 32 bytes, emit zero. - else { - Rep = Constant::getNullValue(F->getReturnType()); - } - - // Replace any uses with our new instruction. - if (!CI->use_empty()) - CI->replaceAllUsesWith(Rep); - - // Remove upgraded instruction. - CI->eraseFromParent(); - - } else if (F->getName() == "llvm.x86.sse.loadu.ps" || - F->getName() == "llvm.x86.sse2.loadu.dq" || - F->getName() == "llvm.x86.sse2.loadu.pd") { + if (F->getName() == "llvm.x86.sse.loadu.ps" || + F->getName() == "llvm.x86.sse2.loadu.dq" || + F->getName() == "llvm.x86.sse2.loadu.pd") { // Convert to a native, unaligned load. const Type *VecTy = CI->getType(); const Type *IntTy = IntegerType::get(C, 128); @@ -1064,310 +204,6 @@ void llvm::UpgradeIntrinsicCall(CallInst *CI, Function *NewFn) { } switch (NewFn->getIntrinsicID()) { - default: llvm_unreachable("Unknown function for CallInst upgrade."); - case Intrinsic::arm_neon_vld1: - case Intrinsic::arm_neon_vld2: - case Intrinsic::arm_neon_vld3: - case Intrinsic::arm_neon_vld4: - case Intrinsic::arm_neon_vst1: - case Intrinsic::arm_neon_vst2: - case Intrinsic::arm_neon_vst3: - case Intrinsic::arm_neon_vst4: - case Intrinsic::arm_neon_vld2lane: - case Intrinsic::arm_neon_vld3lane: - case Intrinsic::arm_neon_vld4lane: - case Intrinsic::arm_neon_vst2lane: - case Intrinsic::arm_neon_vst3lane: - case Intrinsic::arm_neon_vst4lane: { - // Add a default alignment argument of 1. - SmallVector<Value*, 8> Operands(CS.arg_begin(), CS.arg_end()); - Operands.push_back(ConstantInt::get(Type::getInt32Ty(C), 1)); - CallInst *NewCI = CallInst::Create(NewFn, Operands.begin(), Operands.end(), - CI->getName(), CI); - NewCI->setTailCall(CI->isTailCall()); - NewCI->setCallingConv(CI->getCallingConv()); - - // Handle any uses of the old CallInst. - if (!CI->use_empty()) - // Replace all uses of the old call with the new cast which has the - // correct type. - CI->replaceAllUsesWith(NewCI); - - // Clean up the old call now that it has been completely upgraded. - CI->eraseFromParent(); - break; - } - - case Intrinsic::x86_mmx_padd_b: - case Intrinsic::x86_mmx_padd_w: - case Intrinsic::x86_mmx_padd_d: - case Intrinsic::x86_mmx_padd_q: - case Intrinsic::x86_mmx_padds_b: - case Intrinsic::x86_mmx_padds_w: - case Intrinsic::x86_mmx_paddus_b: - case Intrinsic::x86_mmx_paddus_w: - case Intrinsic::x86_mmx_psub_b: - case Intrinsic::x86_mmx_psub_w: - case Intrinsic::x86_mmx_psub_d: - case Intrinsic::x86_mmx_psub_q: - case Intrinsic::x86_mmx_psubs_b: - case Intrinsic::x86_mmx_psubs_w: - case Intrinsic::x86_mmx_psubus_b: - case Intrinsic::x86_mmx_psubus_w: - case Intrinsic::x86_mmx_pmulh_w: - case Intrinsic::x86_mmx_pmull_w: - case Intrinsic::x86_mmx_pmulhu_w: - case Intrinsic::x86_mmx_pmulu_dq: - case Intrinsic::x86_mmx_pmadd_wd: - case Intrinsic::x86_mmx_pand: - case Intrinsic::x86_mmx_pandn: - case Intrinsic::x86_mmx_por: - case Intrinsic::x86_mmx_pxor: - case Intrinsic::x86_mmx_pavg_b: - case Intrinsic::x86_mmx_pavg_w: - case Intrinsic::x86_mmx_pmaxu_b: - case Intrinsic::x86_mmx_pmaxs_w: - case Intrinsic::x86_mmx_pminu_b: - case Intrinsic::x86_mmx_pmins_w: - case Intrinsic::x86_mmx_psad_bw: - case Intrinsic::x86_mmx_psll_w: - case Intrinsic::x86_mmx_psll_d: - case Intrinsic::x86_mmx_psll_q: - case Intrinsic::x86_mmx_pslli_w: - case Intrinsic::x86_mmx_pslli_d: - case Intrinsic::x86_mmx_pslli_q: - case Intrinsic::x86_mmx_psrl_w: - case Intrinsic::x86_mmx_psrl_d: - case Intrinsic::x86_mmx_psrl_q: - case Intrinsic::x86_mmx_psrli_w: - case Intrinsic::x86_mmx_psrli_d: - case Intrinsic::x86_mmx_psrli_q: - case Intrinsic::x86_mmx_psra_w: - case Intrinsic::x86_mmx_psra_d: - case Intrinsic::x86_mmx_psrai_w: - case Intrinsic::x86_mmx_psrai_d: - case Intrinsic::x86_mmx_packsswb: - case Intrinsic::x86_mmx_packssdw: - case Intrinsic::x86_mmx_packuswb: - case Intrinsic::x86_mmx_punpckhbw: - case Intrinsic::x86_mmx_punpckhwd: - case Intrinsic::x86_mmx_punpckhdq: - case Intrinsic::x86_mmx_punpcklbw: - case Intrinsic::x86_mmx_punpcklwd: - case Intrinsic::x86_mmx_punpckldq: - case Intrinsic::x86_mmx_pcmpeq_b: - case Intrinsic::x86_mmx_pcmpeq_w: - case Intrinsic::x86_mmx_pcmpeq_d: - case Intrinsic::x86_mmx_pcmpgt_b: - case Intrinsic::x86_mmx_pcmpgt_w: - case Intrinsic::x86_mmx_pcmpgt_d: { - Value *Operands[2]; - - // Cast the operand to the X86 MMX type. - Operands[0] = new BitCastInst(CI->getArgOperand(0), - NewFn->getFunctionType()->getParamType(0), - "upgraded.", CI); - - switch (NewFn->getIntrinsicID()) { - default: - // Cast to the X86 MMX type. - Operands[1] = new BitCastInst(CI->getArgOperand(1), - NewFn->getFunctionType()->getParamType(1), - "upgraded.", CI); - break; - case Intrinsic::x86_mmx_pslli_w: - case Intrinsic::x86_mmx_pslli_d: - case Intrinsic::x86_mmx_pslli_q: - case Intrinsic::x86_mmx_psrli_w: - case Intrinsic::x86_mmx_psrli_d: - case Intrinsic::x86_mmx_psrli_q: - case Intrinsic::x86_mmx_psrai_w: - case Intrinsic::x86_mmx_psrai_d: - // These take an i32 as their second parameter. - Operands[1] = CI->getArgOperand(1); - break; - } - - ConstructNewCallInst(NewFn, CI, Operands, 2); - break; - } - case Intrinsic::x86_mmx_maskmovq: { - Value *Operands[3]; - - // Cast the operands to the X86 MMX type. - Operands[0] = new BitCastInst(CI->getArgOperand(0), - NewFn->getFunctionType()->getParamType(0), - "upgraded.", CI); - Operands[1] = new BitCastInst(CI->getArgOperand(1), - NewFn->getFunctionType()->getParamType(1), - "upgraded.", CI); - Operands[2] = CI->getArgOperand(2); - - ConstructNewCallInst(NewFn, CI, Operands, 3, false); - break; - } - case Intrinsic::x86_mmx_pmovmskb: { - Value *Operands[1]; - - // Cast the operand to the X86 MMX type. - Operands[0] = new BitCastInst(CI->getArgOperand(0), - NewFn->getFunctionType()->getParamType(0), - "upgraded.", CI); - - ConstructNewCallInst(NewFn, CI, Operands, 1); - break; - } - case Intrinsic::x86_mmx_movnt_dq: { - Value *Operands[2]; - - Operands[0] = CI->getArgOperand(0); - - // Cast the operand to the X86 MMX type. - Operands[1] = new BitCastInst(CI->getArgOperand(1), - NewFn->getFunctionType()->getParamType(1), - "upgraded.", CI); - - ConstructNewCallInst(NewFn, CI, Operands, 2, false); - break; - } - case Intrinsic::x86_mmx_palignr_b: { - Value *Operands[3]; - - // Cast the operands to the X86 MMX type. - Operands[0] = new BitCastInst(CI->getArgOperand(0), - NewFn->getFunctionType()->getParamType(0), - "upgraded.", CI); - Operands[1] = new BitCastInst(CI->getArgOperand(1), - NewFn->getFunctionType()->getParamType(1), - "upgraded.", CI); - Operands[2] = CI->getArgOperand(2); - - ConstructNewCallInst(NewFn, CI, Operands, 3); - break; - } - case Intrinsic::x86_mmx_pextr_w: { - Value *Operands[2]; - - // Cast the operands to the X86 MMX type. - Operands[0] = new BitCastInst(CI->getArgOperand(0), - NewFn->getFunctionType()->getParamType(0), - "upgraded.", CI); - Operands[1] = CI->getArgOperand(1); - - ConstructNewCallInst(NewFn, CI, Operands, 2); - break; - } - case Intrinsic::x86_mmx_pinsr_w: { - Value *Operands[3]; - - // Cast the operands to the X86 MMX type. - Operands[0] = new BitCastInst(CI->getArgOperand(0), - NewFn->getFunctionType()->getParamType(0), - "upgraded.", CI); - Operands[1] = CI->getArgOperand(1); - Operands[2] = CI->getArgOperand(2); - - ConstructNewCallInst(NewFn, CI, Operands, 3); - break; - } - case Intrinsic::x86_sse_pshuf_w: { - IRBuilder<> Builder(C); - Builder.SetInsertPoint(CI->getParent(), CI); - - // Cast the operand to the X86 MMX type. - Value *Operands[2]; - Operands[0] = - Builder.CreateBitCast(CI->getArgOperand(0), - NewFn->getFunctionType()->getParamType(0), - "upgraded."); - Operands[1] = - Builder.CreateTrunc(CI->getArgOperand(1), - Type::getInt8Ty(C), - "upgraded."); - - ConstructNewCallInst(NewFn, CI, Operands, 2); - break; - } - - case Intrinsic::ctlz: - case Intrinsic::ctpop: - case Intrinsic::cttz: { - // Build a small vector of the original arguments. - SmallVector<Value*, 8> Operands(CS.arg_begin(), CS.arg_end()); - - // Construct a new CallInst - CallInst *NewCI = CallInst::Create(NewFn, Operands.begin(), Operands.end(), - "upgraded."+CI->getName(), CI); - NewCI->setTailCall(CI->isTailCall()); - NewCI->setCallingConv(CI->getCallingConv()); - - // Handle any uses of the old CallInst. - if (!CI->use_empty()) { - // Check for sign extend parameter attributes on the return values. - bool SrcSExt = NewFn->getAttributes().paramHasAttr(0, Attribute::SExt); - bool DestSExt = F->getAttributes().paramHasAttr(0, Attribute::SExt); - - // Construct an appropriate cast from the new return type to the old. - CastInst *RetCast = CastInst::Create( - CastInst::getCastOpcode(NewCI, SrcSExt, - F->getReturnType(), - DestSExt), - NewCI, F->getReturnType(), - NewCI->getName(), CI); - NewCI->moveBefore(RetCast); - - // Replace all uses of the old call with the new cast which has the - // correct type. - CI->replaceAllUsesWith(RetCast); - } - - // Clean up the old call now that it has been completely upgraded. - CI->eraseFromParent(); - } - break; - case Intrinsic::eh_selector: - case Intrinsic::eh_typeid_for: { - // Only the return type changed. - SmallVector<Value*, 8> Operands(CS.arg_begin(), CS.arg_end()); - CallInst *NewCI = CallInst::Create(NewFn, Operands.begin(), Operands.end(), - "upgraded." + CI->getName(), CI); - NewCI->setTailCall(CI->isTailCall()); - NewCI->setCallingConv(CI->getCallingConv()); - - // Handle any uses of the old CallInst. - if (!CI->use_empty()) { - // Construct an appropriate cast from the new return type to the old. - CastInst *RetCast = - CastInst::Create(CastInst::getCastOpcode(NewCI, true, - F->getReturnType(), true), - NewCI, F->getReturnType(), NewCI->getName(), CI); - CI->replaceAllUsesWith(RetCast); - } - CI->eraseFromParent(); - } - break; - case Intrinsic::memcpy: - case Intrinsic::memmove: - case Intrinsic::memset: { - // Add isVolatile - const llvm::Type *I1Ty = llvm::Type::getInt1Ty(CI->getContext()); - Value *Operands[5] = { CI->getArgOperand(0), CI->getArgOperand(1), - CI->getArgOperand(2), CI->getArgOperand(3), - llvm::ConstantInt::get(I1Ty, 0) }; - CallInst *NewCI = CallInst::Create(NewFn, Operands, Operands+5, - CI->getName(), CI); - NewCI->setTailCall(CI->isTailCall()); - NewCI->setCallingConv(CI->getCallingConv()); - // Handle any uses of the old CallInst. - if (!CI->use_empty()) - // Replace all uses of the old call with the new cast which has the - // correct type. - CI->replaceAllUsesWith(NewCI); - - // Clean up the old call now that it has been completely upgraded. - CI->eraseFromParent(); - break; - } case Intrinsic::prefetch: { IRBuilder<> Builder(C); Builder.SetInsertPoint(CI->getParent(), CI); @@ -1401,13 +237,13 @@ void llvm::UpgradeCallsToIntrinsic(Function* F) { assert(F && "Illegal attempt to upgrade a non-existent intrinsic."); // Upgrade the function and check if it is a totaly new function. - Function* NewFn; + Function *NewFn; if (UpgradeIntrinsicFunction(F, NewFn)) { if (NewFn != F) { // Replace all uses to the old function with the new one if necessary. for (Value::use_iterator UI = F->use_begin(), UE = F->use_end(); UI != UE; ) { - if (CallInst* CI = dyn_cast<CallInst>(*UI++)) + if (CallInst *CI = dyn_cast<CallInst>(*UI++)) UpgradeIntrinsicCall(CI, NewFn); } // Remove old function, no longer used, from the module. @@ -1420,37 +256,27 @@ void llvm::UpgradeCallsToIntrinsic(Function* F) { /// If an llvm.dbg.declare intrinsic is invalid, then this function simply /// strips that use. void llvm::CheckDebugInfoIntrinsics(Module *M) { - - if (Function *FuncStart = M->getFunction("llvm.dbg.func.start")) { - while (!FuncStart->use_empty()) { - CallInst *CI = cast<CallInst>(FuncStart->use_back()); - CI->eraseFromParent(); - } + while (!FuncStart->use_empty()) + cast<CallInst>(FuncStart->use_back())->eraseFromParent(); FuncStart->eraseFromParent(); } if (Function *StopPoint = M->getFunction("llvm.dbg.stoppoint")) { - while (!StopPoint->use_empty()) { - CallInst *CI = cast<CallInst>(StopPoint->use_back()); - CI->eraseFromParent(); - } + while (!StopPoint->use_empty()) + cast<CallInst>(StopPoint->use_back())->eraseFromParent(); StopPoint->eraseFromParent(); } if (Function *RegionStart = M->getFunction("llvm.dbg.region.start")) { - while (!RegionStart->use_empty()) { - CallInst *CI = cast<CallInst>(RegionStart->use_back()); - CI->eraseFromParent(); - } + while (!RegionStart->use_empty()) + cast<CallInst>(RegionStart->use_back())->eraseFromParent(); RegionStart->eraseFromParent(); } if (Function *RegionEnd = M->getFunction("llvm.dbg.region.end")) { - while (!RegionEnd->use_empty()) { - CallInst *CI = cast<CallInst>(RegionEnd->use_back()); - CI->eraseFromParent(); - } + while (!RegionEnd->use_empty()) + cast<CallInst>(RegionEnd->use_back())->eraseFromParent(); RegionEnd->eraseFromParent(); } diff --git a/test/Analysis/BasicAA/getmodrefinfo-cs-cs.ll b/test/Analysis/BasicAA/getmodrefinfo-cs-cs.ll index 062ea59..ebd349a 100644 --- a/test/Analysis/BasicAA/getmodrefinfo-cs-cs.ll +++ b/test/Analysis/BasicAA/getmodrefinfo-cs-cs.ll @@ -15,12 +15,12 @@ define void @test0() { ; CHECK: NoModRef: call void @llvm.memset.p0i8.i64(i8* @A, i8 0, i64 1, i32 1, i1 false) <-> call void @llvm.memset.p0i8.i64(i8* @B, i8 0, i64 1, i32 1, i1 false) ; CHECK: NoModRef: call void @llvm.memset.p0i8.i64(i8* @B, i8 0, i64 1, i32 1, i1 false) <-> call void @llvm.memset.p0i8.i64(i8* @A, i8 0, i64 1, i32 1, i1 false) -declare void @llvm.memset.i64(i8*, i8, i64, i32) +declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) nounwind @A = external global i8 @B = external global i8 define void @test1() { - call void @llvm.memset.i64(i8* @A, i8 0, i64 1, i32 1) - call void @llvm.memset.i64(i8* @B, i8 0, i64 1, i32 1) + call void @llvm.memset.p0i8.i64(i8* @A, i8 0, i64 1, i32 1, i1 false) + call void @llvm.memset.p0i8.i64(i8* @B, i8 0, i64 1, i32 1, i1 false) ret void } diff --git a/test/Analysis/BasicAA/modref.ll b/test/Analysis/BasicAA/modref.ll index 7318a89..233396b 100644 --- a/test/Analysis/BasicAA/modref.ll +++ b/test/Analysis/BasicAA/modref.ll @@ -1,10 +1,6 @@ ; RUN: opt < %s -basicaa -gvn -dse -S | FileCheck %s target datalayout = "E-p:64:64:64-a0:0:8-f32:32:32-f64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-v64:64:64-v128:128:128" -declare void @llvm.memset.i32(i8*, i8, i32, i32) -declare void @llvm.memset.i8(i8*, i8, i8, i32) -declare void @llvm.memcpy.i8(i8*, i8*, i8, i32) -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) declare void @llvm.lifetime.end(i64, i8* nocapture) declare void @external(i32*) @@ -15,7 +11,7 @@ define i32 @test0(i8* %P) { store i32 0, i32* %A - call void @llvm.memset.i32(i8* %P, i8 0, i32 42, i32 1) + call void @llvm.memset.p0i8.i32(i8* %P, i8 0, i32 42, i32 1, i1 false) %B = load i32* %A ret i32 %B @@ -31,7 +27,7 @@ define i8 @test1() { store i8 2, i8* %B ;; Not written to by memcpy - call void @llvm.memcpy.i8(i8* %A, i8* %B, i8 -1, i32 0) + call void @llvm.memcpy.p0i8.p0i8.i8(i8* %A, i8* %B, i8 -1, i32 0, i1 false) %C = load i8* %B ret i8 %C @@ -42,7 +38,7 @@ define i8 @test2(i8* %P) { ; CHECK: @test2 %P2 = getelementptr i8* %P, i32 127 store i8 1, i8* %P2 ;; Not dead across memset - call void @llvm.memset.i8(i8* %P, i8 2, i8 127, i32 0) + call void @llvm.memset.p0i8.i8(i8* %P, i8 2, i8 127, i32 0, i1 false) %A = load i8* %P2 ret i8 %A ; CHECK: ret i8 1 @@ -55,7 +51,7 @@ define i8 @test2a(i8* %P) { ;; FIXME: DSE isn't zapping this dead store. store i8 1, i8* %P2 ;; Dead, clobbered by memset. - call void @llvm.memset.i8(i8* %P, i8 2, i8 127, i32 0) + call void @llvm.memset.p0i8.i8(i8* %P, i8 2, i8 127, i32 0, i1 false) %A = load i8* %P2 ret i8 %A ; CHECK-NOT: load @@ -95,7 +91,7 @@ define void @test3a(i8* %P, i8 %X) { define i32 @test4(i8* %P) { %tmp = load i32* @G1 - call void @llvm.memset.i32(i8* bitcast ([4000 x i32]* @G2 to i8*), i8 0, i32 4000, i32 1) + call void @llvm.memset.p0i8.i32(i8* bitcast ([4000 x i32]* @G2 to i8*), i8 0, i32 4000, i32 1, i1 false) %tmp2 = load i32* @G1 %sub = sub i32 %tmp2, %tmp ret i32 %sub @@ -110,7 +106,7 @@ define i32 @test4(i8* %P) { ; write to G1. define i32 @test5(i8* %P, i32 %Len) { %tmp = load i32* @G1 - call void @llvm.memcpy.i32(i8* bitcast ([4000 x i32]* @G2 to i8*), i8* bitcast (i32* @G1 to i8*), i32 %Len, i32 1) + call void @llvm.memcpy.p0i8.p0i8.i32(i8* bitcast ([4000 x i32]* @G2 to i8*), i8* bitcast (i32* @G1 to i8*), i32 %Len, i32 1, i1 false) %tmp2 = load i32* @G1 %sub = sub i32 %tmp2, %tmp ret i32 %sub @@ -132,3 +128,9 @@ define i8 @test6(i8* %p, i8* noalias %a) { ; CHECK-NOT: load ; CHECK: ret } + +declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1) nounwind +declare void @llvm.memset.p0i8.i8(i8* nocapture, i8, i8, i32, i1) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i8(i8* nocapture, i8* nocapture, i8, i32, i1) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind + diff --git a/test/Analysis/ScalarEvolution/avoid-smax-1.ll b/test/Analysis/ScalarEvolution/avoid-smax-1.ll index 89e8b98..474d564 100644 --- a/test/Analysis/ScalarEvolution/avoid-smax-1.ll +++ b/test/Analysis/ScalarEvolution/avoid-smax-1.ll @@ -173,7 +173,7 @@ bb23: ; preds = %bb24, %bb.nph %55 = mul i32 %y.21, %w ; <i32> [#uses=1] %.sum5 = add i32 %55, %.sum3 ; <i32> [#uses=1] %56 = getelementptr i8* %j, i32 %.sum5 ; <i8*> [#uses=1] - tail call void @llvm.memcpy.i32(i8* %56, i8* %54, i32 %w, i32 1) + tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* %56, i8* %54, i32 %w, i32 1, i1 false) %57 = add i32 %y.21, 1 ; <i32> [#uses=2] br label %bb24 @@ -190,7 +190,7 @@ bb26: ; preds = %bb24.bb26_crit_edge, %bb22 %60 = getelementptr i8* %j, i32 %.sum4 ; <i8*> [#uses=1] %61 = mul i32 %x, %w ; <i32> [#uses=1] %62 = sdiv i32 %61, 2 ; <i32> [#uses=1] - tail call void @llvm.memset.i32(i8* %60, i8 -128, i32 %62, i32 1) + tail call void @llvm.memset.p0i8.i32(i8* %60, i8 -128, i32 %62, i32 1, i1 false) ret void bb29: ; preds = %bb20, %entry @@ -208,7 +208,7 @@ bb30: ; preds = %bb31, %bb.nph11 %67 = getelementptr i8* %r, i32 %66 ; <i8*> [#uses=1] %68 = mul i32 %y.310, %w ; <i32> [#uses=1] %69 = getelementptr i8* %j, i32 %68 ; <i8*> [#uses=1] - tail call void @llvm.memcpy.i32(i8* %69, i8* %67, i32 %w, i32 1) + tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* %69, i8* %67, i32 %w, i32 1, i1 false) %70 = add i32 %y.310, 1 ; <i32> [#uses=2] br label %bb31 @@ -224,13 +224,12 @@ bb33: ; preds = %bb31.bb33_crit_edge, %bb29 %73 = getelementptr i8* %j, i32 %72 ; <i8*> [#uses=1] %74 = mul i32 %x, %w ; <i32> [#uses=1] %75 = sdiv i32 %74, 2 ; <i32> [#uses=1] - tail call void @llvm.memset.i32(i8* %73, i8 -128, i32 %75, i32 1) + tail call void @llvm.memset.p0i8.i32(i8* %73, i8 -128, i32 %75, i32 1, i1 false) ret void return: ; preds = %bb20 ret void } -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) nounwind - -declare void @llvm.memset.i32(i8*, i8, i32, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind +declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1) nounwind diff --git a/test/Analysis/ScalarEvolution/trip-count3.ll b/test/Analysis/ScalarEvolution/trip-count3.ll index 10b798b..1bf86ae 100644 --- a/test/Analysis/ScalarEvolution/trip-count3.ll +++ b/test/Analysis/ScalarEvolution/trip-count3.ll @@ -5,15 +5,14 @@ ; dividing by the stride will have a remainder. This could theoretically ; be teaching it how to use a more elaborate trip count computation. -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-s0:64:64-f80:128:128" -target triple = "x86_64-unknown-linux-gnu" - %struct.FILE = type { i32, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, %struct._IO_marker*, %struct.FILE*, i32, i32, i64, i16, i8, [1 x i8], i8*, i64, i8*, i8*, i8*, i8*, i64, i32, [20 x i8] } - %struct.SHA_INFO = type { [5 x i32], i32, i32, [16 x i32] } - %struct._IO_marker = type { %struct._IO_marker*, %struct.FILE*, i32 } -@_2E_str = external constant [26 x i8] ; <[26 x i8]*> [#uses=0] -@stdin = external global %struct.FILE* ; <%struct.FILE**> [#uses=0] -@_2E_str1 = external constant [3 x i8] ; <[3 x i8]*> [#uses=0] -@_2E_str12 = external constant [30 x i8] ; <[30 x i8]*> [#uses=0] +%struct.FILE = type { i32, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, %struct._IO_marker*, %struct.FILE*, i32, i32, i64, i16, i8, [1 x i8], i8*, i64, i8*, i8*, i8*, i8*, i64, i32, [20 x i8] } +%struct.SHA_INFO = type { [5 x i32], i32, i32, [16 x i32] } +%struct._IO_marker = type { %struct._IO_marker*, %struct.FILE*, i32 } + +@_2E_str = external constant [26 x i8] +@stdin = external global %struct.FILE* +@_2E_str1 = external constant [3 x i8] +@_2E_str12 = external constant [30 x i8] declare void @sha_init(%struct.SHA_INFO* nocapture) nounwind @@ -25,12 +24,8 @@ declare i32 @printf(i8* nocapture, ...) nounwind declare void @sha_final(%struct.SHA_INFO* nocapture) nounwind -declare void @llvm.memset.i64(i8* nocapture, i8, i64, i32) nounwind - declare void @sha_update(%struct.SHA_INFO* nocapture, i8* nocapture, i32) nounwind -declare void @llvm.memcpy.i64(i8* nocapture, i8* nocapture, i64, i32) nounwind - declare i64 @fread(i8* noalias nocapture, i64, i64, %struct.FILE* noalias nocapture) nounwind declare i32 @main(i32, i8** nocapture) nounwind @@ -43,36 +38,41 @@ declare void @sha_stream(%struct.SHA_INFO* nocapture, %struct.FILE* nocapture) n define void @sha_stream_bb3_2E_i(%struct.SHA_INFO* %sha_info, i8* %data1, i32, i8** %buffer_addr.0.i.out, i32* %count_addr.0.i.out) nounwind { newFuncRoot: - br label %bb3.i - -sha_update.exit.exitStub: ; preds = %bb3.i - store i8* %buffer_addr.0.i, i8** %buffer_addr.0.i.out - store i32 %count_addr.0.i, i32* %count_addr.0.i.out - ret void - -bb2.i: ; preds = %bb3.i - %1 = getelementptr %struct.SHA_INFO* %sha_info, i64 0, i32 3 ; <[16 x i32]*> [#uses=1] - %2 = bitcast [16 x i32]* %1 to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i64(i8* %2, i8* %buffer_addr.0.i, i64 64, i32 1) nounwind - %3 = getelementptr %struct.SHA_INFO* %sha_info, i64 0, i32 3, i64 0 ; <i32*> [#uses=1] - %4 = bitcast i32* %3 to i8* ; <i8*> [#uses=1] - br label %codeRepl - -codeRepl: ; preds = %bb2.i - call void @sha_stream_bb3_2E_i_bb1_2E_i_2E_i(i8* %4) - br label %byte_reverse.exit.i - -byte_reverse.exit.i: ; preds = %codeRepl - call fastcc void @sha_transform(%struct.SHA_INFO* %sha_info) nounwind - %5 = getelementptr i8* %buffer_addr.0.i, i64 64 ; <i8*> [#uses=1] - %6 = add i32 %count_addr.0.i, -64 ; <i32> [#uses=1] - br label %bb3.i - -bb3.i: ; preds = %byte_reverse.exit.i, %newFuncRoot - %buffer_addr.0.i = phi i8* [ %data1, %newFuncRoot ], [ %5, %byte_reverse.exit.i ] ; <i8*> [#uses=3] - %count_addr.0.i = phi i32 [ %0, %newFuncRoot ], [ %6, %byte_reverse.exit.i ] ; <i32> [#uses=3] - %7 = icmp sgt i32 %count_addr.0.i, 63 ; <i1> [#uses=1] - br i1 %7, label %bb2.i, label %sha_update.exit.exitStub + br label %bb3.i + +sha_update.exit.exitStub: ; preds = %bb3.i + store i8* %buffer_addr.0.i, i8** %buffer_addr.0.i.out + store i32 %count_addr.0.i, i32* %count_addr.0.i.out + ret void + +bb2.i: ; preds = %bb3.i + %1 = getelementptr %struct.SHA_INFO* %sha_info, i64 0, i32 3 + %2 = bitcast [16 x i32]* %1 to i8* + call void @llvm.memcpy.p0i8.p0i8.i64(i8* %2, i8* %buffer_addr.0.i, i64 64, i32 1, i1 false) + %3 = getelementptr %struct.SHA_INFO* %sha_info, i64 0, i32 3, i64 0 + %4 = bitcast i32* %3 to i8* + br label %codeRepl + +codeRepl: ; preds = %bb2.i + call void @sha_stream_bb3_2E_i_bb1_2E_i_2E_i(i8* %4) + br label %byte_reverse.exit.i + +byte_reverse.exit.i: ; preds = %codeRepl + call fastcc void @sha_transform(%struct.SHA_INFO* %sha_info) nounwind + %5 = getelementptr i8* %buffer_addr.0.i, i64 64 + %6 = add i32 %count_addr.0.i, -64 + br label %bb3.i + +bb3.i: ; preds = %byte_reverse.exit.i, %newFuncRoot + %buffer_addr.0.i = phi i8* [ %data1, %newFuncRoot ], [ %5, %byte_reverse.exit.i ] + %count_addr.0.i = phi i32 [ %0, %newFuncRoot ], [ %6, %byte_reverse.exit.i ] + %7 = icmp sgt i32 %count_addr.0.i, 63 + br i1 %7, label %bb2.i, label %sha_update.exit.exitStub } declare void @sha_stream_bb3_2E_i_bb1_2E_i_2E_i(i8*) nounwind + +declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) nounwind + +declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture, i64, i32, i1) nounwind + diff --git a/test/Assembler/AutoUpgradeIntrinsics.ll b/test/Assembler/AutoUpgradeIntrinsics.ll index 20beb49..eb4ac76 100644 --- a/test/Assembler/AutoUpgradeIntrinsics.ll +++ b/test/Assembler/AutoUpgradeIntrinsics.ll @@ -1,87 +1,6 @@ ; Tests to make sure intrinsics are automatically upgraded. -; RUN: llvm-as < %s | llvm-dis | not grep {i32 @llvm\\.ct} -; RUN: llvm-as < %s | llvm-dis | \ -; RUN: not grep {llvm\\.part\\.set\\.i\[0-9\]*\\.i\[0-9\]*\\.i\[0-9\]*} -; RUN: llvm-as < %s | llvm-dis | \ -; RUN: not grep {llvm\\.part\\.select\\.i\[0-9\]*\\.i\[0-9\]*} -; RUN: llvm-as < %s | llvm-dis | \ -; RUN: not grep {llvm\\.bswap\\.i\[0-9\]*\\.i\[0-9\]*} -; RUN: llvm-as < %s | llvm-dis | \ -; RUN: not grep {llvm\\.x86\\.sse2\\.loadu} -; RUN: llvm-as < %s | llvm-dis | \ -; RUN: grep {llvm\\.x86\\.mmx\\.ps} | grep {x86_mmx} | count 16 ; RUN: llvm-as < %s | llvm-dis | FileCheck %s -declare i32 @llvm.ctpop.i28(i28 %val) -declare i32 @llvm.cttz.i29(i29 %val) -declare i32 @llvm.ctlz.i30(i30 %val) - -define i32 @test_ct(i32 %A) { - %c1 = call i32 @llvm.ctpop.i28(i28 1234) - %c2 = call i32 @llvm.cttz.i29(i29 2345) - %c3 = call i32 @llvm.ctlz.i30(i30 3456) - %r1 = add i32 %c1, %c2 - %r2 = add i32 %r1, %c3 - ret i32 %r2 -} - -declare i32 @llvm.part.set.i32.i32.i32(i32 %x, i32 %rep, i32 %hi, i32 %lo) -declare i16 @llvm.part.set.i16.i16.i16(i16 %x, i16 %rep, i32 %hi, i32 %lo) -define i32 @test_part_set(i32 %A, i16 %B) { - %a = call i32 @llvm.part.set.i32.i32.i32(i32 %A, i32 27, i32 8, i32 0) - %b = call i16 @llvm.part.set.i16.i16.i16(i16 %B, i16 27, i32 8, i32 0) - %c = zext i16 %b to i32 - %d = add i32 %a, %c - ret i32 %d -} - -declare i32 @llvm.part.select.i32.i32(i32 %x, i32 %hi, i32 %lo) -declare i16 @llvm.part.select.i16.i16(i16 %x, i32 %hi, i32 %lo) -define i32 @test_part_select(i32 %A, i16 %B) { - %a = call i32 @llvm.part.select.i32.i32(i32 %A, i32 8, i32 0) - %b = call i16 @llvm.part.select.i16.i16(i16 %B, i32 8, i32 0) - %c = zext i16 %b to i32 - %d = add i32 %a, %c - ret i32 %d -} - -declare i32 @llvm.bswap.i32.i32(i32 %x) -declare i16 @llvm.bswap.i16.i16(i16 %x) -define i32 @test_bswap(i32 %A, i16 %B) { - %a = call i32 @llvm.bswap.i32.i32(i32 %A) - %b = call i16 @llvm.bswap.i16.i16(i16 %B) - %c = zext i16 %b to i32 - %d = add i32 %a, %c - ret i32 %d -} - -declare <4 x i16> @llvm.x86.mmx.psra.w(<4 x i16>, <2 x i32>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.psll.w(<4 x i16>, <2 x i32>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.psrl.w(<4 x i16>, <2 x i32>) nounwind readnone -define void @sh16(<4 x i16> %A, <2 x i32> %B) { - %r1 = call <4 x i16> @llvm.x86.mmx.psra.w( <4 x i16> %A, <2 x i32> %B ) ; <<4 x i16>> [#uses=0] - %r2 = call <4 x i16> @llvm.x86.mmx.psll.w( <4 x i16> %A, <2 x i32> %B ) ; <<4 x i16>> [#uses=0] - %r3 = call <4 x i16> @llvm.x86.mmx.psrl.w( <4 x i16> %A, <2 x i32> %B ) ; <<4 x i16>> [#uses=0] - ret void -} - -declare <2 x i32> @llvm.x86.mmx.psra.d(<2 x i32>, <2 x i32>) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.psll.d(<2 x i32>, <2 x i32>) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.psrl.d(<2 x i32>, <2 x i32>) nounwind readnone -define void @sh32(<2 x i32> %A, <2 x i32> %B) { - %r1 = call <2 x i32> @llvm.x86.mmx.psra.d( <2 x i32> %A, <2 x i32> %B ) ; <<2 x i32>> [#uses=0] - %r2 = call <2 x i32> @llvm.x86.mmx.psll.d( <2 x i32> %A, <2 x i32> %B ) ; <<2 x i32>> [#uses=0] - %r3 = call <2 x i32> @llvm.x86.mmx.psrl.d( <2 x i32> %A, <2 x i32> %B ) ; <<2 x i32>> [#uses=0] - ret void -} - -declare <1 x i64> @llvm.x86.mmx.psll.q(<1 x i64>, <2 x i32>) nounwind readnone -declare <1 x i64> @llvm.x86.mmx.psrl.q(<1 x i64>, <2 x i32>) nounwind readnone -define void @sh64(<1 x i64> %A, <2 x i32> %B) { - %r1 = call <1 x i64> @llvm.x86.mmx.psll.q( <1 x i64> %A, <2 x i32> %B ) ; <<1 x i64>> [#uses=0] - %r2 = call <1 x i64> @llvm.x86.mmx.psrl.q( <1 x i64> %A, <2 x i32> %B ) ; <<1 x i64>> [#uses=0] - ret void -} declare <4 x float> @llvm.x86.sse.loadu.ps(i8*) nounwind readnone declare <16 x i8> @llvm.x86.sse2.loadu.dq(i8*) nounwind readnone @@ -90,6 +9,10 @@ define void @test_loadu(i8* %a, double* %b) { %v0 = call <4 x float> @llvm.x86.sse.loadu.ps(i8* %a) %v1 = call <16 x i8> @llvm.x86.sse2.loadu.dq(i8* %a) %v2 = call <2 x double> @llvm.x86.sse2.loadu.pd(double* %b) + +; CHECK: load i128* {{.*}}, align 1 +; CHECK: load i128* {{.*}}, align 1 +; CHECK: load i128* {{.*}}, align 1 ret void } diff --git a/test/Assembler/AutoUpgradeMMXIntrinsics.ll b/test/Assembler/AutoUpgradeMMXIntrinsics.ll deleted file mode 100644 index 54120ff..0000000 --- a/test/Assembler/AutoUpgradeMMXIntrinsics.ll +++ /dev/null @@ -1,223 +0,0 @@ -; Tests to make sure MMX intrinsics are automatically upgraded. -; RUN: llvm-as < %s | llvm-dis -o %t -; RUN: grep {llvm\\.x86\\.mmx} %t | not grep {\\\<1 x i64\\\>} -; RUN: grep {llvm\\.x86\\.mmx} %t | not grep {\\\<2 x i32\\\>} -; RUN: grep {llvm\\.x86\\.mmx} %t | not grep {\\\<4 x i16\\\>} -; RUN: grep {llvm\\.x86\\.mmx} %t | not grep {\\\<8 x i8\\\>} -; RUN: grep {llvm\\.x86\\.sse\\.pshuf\\.w} %t | not grep i32 - -; Addition -declare <8 x i8> @llvm.x86.mmx.padd.b(<8 x i8>, <8 x i8>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.padd.w(<4 x i16>, <4 x i16>) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.padd.d(<2 x i32>, <2 x i32>) nounwind readnone -declare <1 x i64> @llvm.x86.mmx.padd.q(<1 x i64>, <1 x i64>) nounwind readnone -declare <8 x i8> @llvm.x86.mmx.padds.b(<8 x i8>, <8 x i8>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.padds.w(<4 x i16>, <4 x i16>) nounwind readnone -declare <8 x i8> @llvm.x86.mmx.paddus.b(<8 x i8>, <8 x i8>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.paddus.w(<4 x i16>, <4 x i16>) nounwind readnone -define void @add(<8 x i8> %A, <8 x i8> %B, <4 x i16> %C, <4 x i16> %D, - <2 x i32> %E, <2 x i32> %F, <1 x i64> %G, <1 x i64> %H) { - %r1 = call <8 x i8> @llvm.x86.mmx.padd.b(<8 x i8> %A, <8 x i8> %B) - %r2 = call <4 x i16> @llvm.x86.mmx.padd.w(<4 x i16> %C, <4 x i16> %D) - %r3 = call <2 x i32> @llvm.x86.mmx.padd.d(<2 x i32> %E, <2 x i32> %F) - %r4 = call <1 x i64> @llvm.x86.mmx.padd.q(<1 x i64> %G, <1 x i64> %H) - %r5 = call <8 x i8> @llvm.x86.mmx.padds.b(<8 x i8> %A, <8 x i8> %B) - %r6 = call <4 x i16> @llvm.x86.mmx.padds.w(<4 x i16> %C, <4 x i16> %D) - %r7 = call <8 x i8> @llvm.x86.mmx.paddus.b(<8 x i8> %A, <8 x i8> %B) - %r8 = call <4 x i16> @llvm.x86.mmx.paddus.w(<4 x i16> %C, <4 x i16> %D) - ret void -} - -; Subtraction -declare <8 x i8> @llvm.x86.mmx.psub.b(<8 x i8>, <8 x i8>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.psub.w(<4 x i16>, <4 x i16>) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.psub.d(<2 x i32>, <2 x i32>) nounwind readnone -declare <1 x i64> @llvm.x86.mmx.psub.q(<1 x i64>, <1 x i64>) nounwind readnone -declare <8 x i8> @llvm.x86.mmx.psubs.b(<8 x i8>, <8 x i8>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.psubs.w(<4 x i16>, <4 x i16>) nounwind readnone -declare <8 x i8> @llvm.x86.mmx.psubus.b(<8 x i8>, <8 x i8>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.psubus.w(<4 x i16>, <4 x i16>) nounwind readnone -define void @sub(<8 x i8> %A, <8 x i8> %B, <4 x i16> %C, <4 x i16> %D, - <2 x i32> %E, <2 x i32> %F, <1 x i64> %G, <1 x i64> %H) { - %r1 = call <8 x i8> @llvm.x86.mmx.psub.b(<8 x i8> %A, <8 x i8> %B) - %r2 = call <4 x i16> @llvm.x86.mmx.psub.w(<4 x i16> %C, <4 x i16> %D) - %r3 = call <2 x i32> @llvm.x86.mmx.psub.d(<2 x i32> %E, <2 x i32> %F) - %r4 = call <1 x i64> @llvm.x86.mmx.psub.q(<1 x i64> %G, <1 x i64> %H) - %r5 = call <8 x i8> @llvm.x86.mmx.psubs.b(<8 x i8> %A, <8 x i8> %B) - %r6 = call <4 x i16> @llvm.x86.mmx.psubs.w(<4 x i16> %C, <4 x i16> %D) - %r7 = call <8 x i8> @llvm.x86.mmx.psubus.b(<8 x i8> %A, <8 x i8> %B) - %r8 = call <4 x i16> @llvm.x86.mmx.psubus.w(<4 x i16> %C, <4 x i16> %D) - ret void -} - -; Multiplication -declare <4 x i16> @llvm.x86.mmx.pmulh.w(<4 x i16>, <4 x i16>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.pmull.w(<4 x i16>, <4 x i16>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.pmulhu.w(<4 x i16>, <4 x i16>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.pmulu.dq(<4 x i16>, <4 x i16>) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.pmadd.wd(<4 x i16>, <4 x i16>) nounwind readnone -define void @mul(<4 x i16> %A, <4 x i16> %B) { - %r1 = call <4 x i16> @llvm.x86.mmx.pmulh.w(<4 x i16> %A, <4 x i16> %B) - %r2 = call <4 x i16> @llvm.x86.mmx.pmull.w(<4 x i16> %A, <4 x i16> %B) - %r3 = call <4 x i16> @llvm.x86.mmx.pmulhu.w(<4 x i16> %A, <4 x i16> %B) - %r4 = call <4 x i16> @llvm.x86.mmx.pmulu.dq(<4 x i16> %A, <4 x i16> %B) - %r5 = call <2 x i32> @llvm.x86.mmx.pmadd.wd(<4 x i16> %A, <4 x i16> %B) - ret void -} - -; Bitwise operations -declare <1 x i64> @llvm.x86.mmx.pand(<1 x i64>, <1 x i64>) nounwind readnone -declare <1 x i64> @llvm.x86.mmx.pandn(<1 x i64>, <1 x i64>) nounwind readnone -declare <1 x i64> @llvm.x86.mmx.por(<1 x i64>, <1 x i64>) nounwind readnone -declare <1 x i64> @llvm.x86.mmx.pxor(<1 x i64>, <1 x i64>) nounwind readnone -define void @bit(<1 x i64> %A, <1 x i64> %B) { - %r1 = call <1 x i64> @llvm.x86.mmx.pand(<1 x i64> %A, <1 x i64> %B) - %r2 = call <1 x i64> @llvm.x86.mmx.pandn(<1 x i64> %A, <1 x i64> %B) - %r3 = call <1 x i64> @llvm.x86.mmx.por(<1 x i64> %A, <1 x i64> %B) - %r4 = call <1 x i64> @llvm.x86.mmx.pxor(<1 x i64> %A, <1 x i64> %B) - ret void -} - -; Averages -declare <8 x i8> @llvm.x86.mmx.pavg.b(<8 x i8>, <8 x i8>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.pavg.w(<4 x i16>, <4 x i16>) nounwind readnone -define void @avg(<8 x i8> %A, <8 x i8> %B, <4 x i16> %C, <4 x i16> %D) { - %r1 = call <8 x i8> @llvm.x86.mmx.pavg.b(<8 x i8> %A, <8 x i8> %B) - %r2 = call <4 x i16> @llvm.x86.mmx.pavg.w(<4 x i16> %C, <4 x i16> %D) - ret void -} - -; Maximum -declare <8 x i8> @llvm.x86.mmx.pmaxu.b(<8 x i8>, <8 x i8>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.pmaxs.w(<4 x i16>, <4 x i16>) nounwind readnone -define void @max(<8 x i8> %A, <8 x i8> %B, <4 x i16> %C, <4 x i16> %D) { - %r1 = call <8 x i8> @llvm.x86.mmx.pmaxu.b(<8 x i8> %A, <8 x i8> %B) - %r2 = call <4 x i16> @llvm.x86.mmx.pmaxs.w(<4 x i16> %C, <4 x i16> %D) - ret void -} - -; Minimum -declare <8 x i8> @llvm.x86.mmx.pminu.b(<8 x i8>, <8 x i8>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.pmins.w(<4 x i16>, <4 x i16>) nounwind readnone -define void @min(<8 x i8> %A, <8 x i8> %B, <4 x i16> %C, <4 x i16> %D) { - %r1 = call <8 x i8> @llvm.x86.mmx.pminu.b(<8 x i8> %A, <8 x i8> %B) - %r2 = call <4 x i16> @llvm.x86.mmx.pmins.w(<4 x i16> %C, <4 x i16> %D) - ret void -} - -; Packed sum of absolute differences -declare <4 x i16> @llvm.x86.mmx.psad.bw(<8 x i8>, <8 x i8>) nounwind readnone -define void @psad(<8 x i8> %A, <8 x i8> %B) { - %r1 = call <4 x i16> @llvm.x86.mmx.psad.bw(<8 x i8> %A, <8 x i8> %B) - ret void -} - -; Shift left -declare <4 x i16> @llvm.x86.mmx.psll.w(<4 x i16>, <1 x i64>) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.psll.d(<2 x i32>, <1 x i64>) nounwind readnone -declare <1 x i64> @llvm.x86.mmx.psll.q(<1 x i64>, <1 x i64>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.pslli.w(<4 x i16>, i32) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.pslli.d(<2 x i32>, i32) nounwind readnone -declare <1 x i64> @llvm.x86.mmx.pslli.q(<1 x i64>, i32) nounwind readnone -define void @shl(<4 x i16> %A, <2 x i32> %B, <1 x i64> %C, i32 %D) { - %r1 = call <4 x i16> @llvm.x86.mmx.psll.w(<4 x i16> %A, <1 x i64> %C) - %r2 = call <2 x i32> @llvm.x86.mmx.psll.d(<2 x i32> %B, <1 x i64> %C) - %r3 = call <1 x i64> @llvm.x86.mmx.psll.q(<1 x i64> %C, <1 x i64> %C) - %r4 = call <4 x i16> @llvm.x86.mmx.pslli.w(<4 x i16> %A, i32 %D) - %r5 = call <2 x i32> @llvm.x86.mmx.pslli.d(<2 x i32> %B, i32 %D) - %r6 = call <1 x i64> @llvm.x86.mmx.pslli.q(<1 x i64> %C, i32 %D) - ret void -} - -; Shift right logical -declare <4 x i16> @llvm.x86.mmx.psrl.w(<4 x i16>, <1 x i64>) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.psrl.d(<2 x i32>, <1 x i64>) nounwind readnone -declare <1 x i64> @llvm.x86.mmx.psrl.q(<1 x i64>, <1 x i64>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.psrli.w(<4 x i16>, i32) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.psrli.d(<2 x i32>, i32) nounwind readnone -declare <1 x i64> @llvm.x86.mmx.psrli.q(<1 x i64>, i32) nounwind readnone -define void @shr(<4 x i16> %A, <2 x i32> %B, <1 x i64> %C, i32 %D) { - %r1 = call <4 x i16> @llvm.x86.mmx.psrl.w(<4 x i16> %A, <1 x i64> %C) - %r2 = call <2 x i32> @llvm.x86.mmx.psrl.d(<2 x i32> %B, <1 x i64> %C) - %r3 = call <1 x i64> @llvm.x86.mmx.psrl.q(<1 x i64> %C, <1 x i64> %C) - %r4 = call <4 x i16> @llvm.x86.mmx.psrli.w(<4 x i16> %A, i32 %D) - %r5 = call <2 x i32> @llvm.x86.mmx.psrli.d(<2 x i32> %B, i32 %D) - %r6 = call <1 x i64> @llvm.x86.mmx.psrli.q(<1 x i64> %C, i32 %D) - ret void -} - -; Shift right arithmetic -declare <4 x i16> @llvm.x86.mmx.psra.w(<4 x i16>, <1 x i64>) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.psra.d(<2 x i32>, <1 x i64>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.psrai.w(<4 x i16>, i32) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.psrai.d(<2 x i32>, i32) nounwind readnone -define void @sra(<4 x i16> %A, <2 x i32> %B, <1 x i64> %C, i32 %D) { - %r1 = call <4 x i16> @llvm.x86.mmx.psra.w(<4 x i16> %A, <1 x i64> %C) - %r2 = call <2 x i32> @llvm.x86.mmx.psra.d(<2 x i32> %B, <1 x i64> %C) - %r3 = call <4 x i16> @llvm.x86.mmx.psrai.w(<4 x i16> %A, i32 %D) - %r4 = call <2 x i32> @llvm.x86.mmx.psrai.d(<2 x i32> %B, i32 %D) - ret void -} - -; Pack/Unpack ops -declare <8 x i8> @llvm.x86.mmx.packsswb(<4 x i16>, <4 x i16>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.packssdw(<2 x i32>, <2 x i32>) nounwind readnone -declare <8 x i8> @llvm.x86.mmx.packuswb(<4 x i16>, <4 x i16>) nounwind readnone -declare <8 x i8> @llvm.x86.mmx.punpckhbw(<8 x i8>, <8 x i8>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.punpckhwd(<4 x i16>, <4 x i16>) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.punpckhdq(<2 x i32>, <2 x i32>) nounwind readnone -declare <8 x i8> @llvm.x86.mmx.punpcklbw(<8 x i8>, <8 x i8>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.punpcklwd(<4 x i16>, <4 x i16>) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.punpckldq(<2 x i32>, <2 x i32>) nounwind readnone -define void @pack_unpack(<8 x i8> %A, <8 x i8> %B, <4 x i16> %C, <4 x i16> %D, - <2 x i32> %E, <2 x i32> %F) { - %r1 = call <8 x i8> @llvm.x86.mmx.packsswb(<4 x i16> %C, <4 x i16> %D) - %r2 = call <4 x i16> @llvm.x86.mmx.packssdw(<2 x i32> %E, <2 x i32> %F) - %r3 = call <8 x i8> @llvm.x86.mmx.packuswb(<4 x i16> %C, <4 x i16> %D) - %r4 = call <8 x i8> @llvm.x86.mmx.punpckhbw(<8 x i8> %A, <8 x i8> %B) - %r5 = call <4 x i16> @llvm.x86.mmx.punpckhwd(<4 x i16> %C, <4 x i16> %D) - %r6 = call <2 x i32> @llvm.x86.mmx.punpckhdq(<2 x i32> %E, <2 x i32> %F) - %r7 = call <8 x i8> @llvm.x86.mmx.punpcklbw(<8 x i8> %A, <8 x i8> %B) - %r8 = call <4 x i16> @llvm.x86.mmx.punpcklwd(<4 x i16> %C, <4 x i16> %D) - %r9 = call <2 x i32> @llvm.x86.mmx.punpckldq(<2 x i32> %E, <2 x i32> %F) - ret void -} - -; Integer comparison ops -declare <8 x i8> @llvm.x86.mmx.pcmpeq.b(<8 x i8>, <8 x i8>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.pcmpeq.w(<4 x i16>, <4 x i16>) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.pcmpeq.d(<2 x i32>, <2 x i32>) nounwind readnone -declare <8 x i8> @llvm.x86.mmx.pcmpgt.b(<8 x i8>, <8 x i8>) nounwind readnone -declare <4 x i16> @llvm.x86.mmx.pcmpgt.w(<4 x i16>, <4 x i16>) nounwind readnone -declare <2 x i32> @llvm.x86.mmx.pcmpgt.d(<2 x i32>, <2 x i32>) nounwind readnone -define void @cmp(<8 x i8> %A, <8 x i8> %B, <4 x i16> %C, <4 x i16> %D, - <2 x i32> %E, <2 x i32> %F) { - %r1 = call <8 x i8> @llvm.x86.mmx.pcmpeq.b(<8 x i8> %A, <8 x i8> %B) - %r2 = call <4 x i16> @llvm.x86.mmx.pcmpeq.w(<4 x i16> %C, <4 x i16> %D) - %r3 = call <2 x i32> @llvm.x86.mmx.pcmpeq.d(<2 x i32> %E, <2 x i32> %F) - %r4 = call <8 x i8> @llvm.x86.mmx.pcmpgt.b(<8 x i8> %A, <8 x i8> %B) - %r5 = call <4 x i16> @llvm.x86.mmx.pcmpgt.w(<4 x i16> %C, <4 x i16> %D) - %r6 = call <2 x i32> @llvm.x86.mmx.pcmpgt.d(<2 x i32> %E, <2 x i32> %F) - ret void -} - -; Miscellaneous -declare void @llvm.x86.mmx.maskmovq(<8 x i8>, <8 x i8>, i32*) nounwind readnone -declare i32 @llvm.x86.mmx.pmovmskb(<8 x i8>) nounwind readnone -declare void @llvm.x86.mmx.movnt.dq(i32*, <1 x i64>) nounwind readnone -declare <1 x i64> @llvm.x86.mmx.palignr.b(<1 x i64>, <1 x i64>, i8) nounwind readnone -declare i32 @llvm.x86.mmx.pextr.w(<1 x i64>, i32) nounwind readnone -declare <1 x i64> @llvm.x86.mmx.pinsr.w(<1 x i64>, i32, i32) nounwind readnone -declare <4 x i16> @llvm.x86.ssse3.pshuf.w(<4 x i16>, i32) nounwind readnone -define void @misc(<8 x i8> %A, <8 x i8> %B, <4 x i16> %C, <4 x i16> %D, - <2 x i32> %E, <2 x i32> %F, <1 x i64> %G, <1 x i64> %H, - i32* %I, i8 %J, i16 %K, i32 %L) { - call void @llvm.x86.mmx.maskmovq(<8 x i8> %A, <8 x i8> %B, i32* %I) - %r1 = call i32 @llvm.x86.mmx.pmovmskb(<8 x i8> %A) - call void @llvm.x86.mmx.movnt.dq(i32* %I, <1 x i64> %G) - %r2 = call <1 x i64> @llvm.x86.mmx.palignr.b(<1 x i64> %G, <1 x i64> %H, i8 %J) - %r3 = call i32 @llvm.x86.mmx.pextr.w(<1 x i64> %G, i32 37) - %r4 = call <1 x i64> @llvm.x86.mmx.pinsr.w(<1 x i64> %G, i32 37, i32 927) - %r5 = call <4 x i16> @llvm.x86.ssse3.pshuf.w(<4 x i16> %C, i32 37) - ret void -} diff --git a/test/Bitcode/memcpy.ll b/test/Bitcode/memcpy.ll deleted file mode 100644 index 299eb1e..0000000 --- a/test/Bitcode/memcpy.ll +++ /dev/null @@ -1,23 +0,0 @@ -; RUN: llvm-as %s -o /dev/null - -define void @test(i32* %P, i32* %Q) { -entry: - %tmp.1 = bitcast i32* %P to i8* ; <i8*> [#uses=3] - %tmp.3 = bitcast i32* %Q to i8* ; <i8*> [#uses=4] - tail call void @llvm.memcpy.i32( i8* %tmp.1, i8* %tmp.3, i32 100000, i32 1 ) - tail call void @llvm.memcpy.i64( i8* %tmp.1, i8* %tmp.3, i64 100000, i32 1 ) - tail call void @llvm.memset.i32( i8* %tmp.3, i8 14, i32 10000, i32 0 ) - tail call void @llvm.memmove.i32( i8* %tmp.1, i8* %tmp.3, i32 123124, i32 1 ) - tail call void @llvm.memmove.i64( i8* %tmp.1, i8* %tmp.3, i64 123124, i32 1 ) - ret void -} - -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) - -declare void @llvm.memcpy.i64(i8*, i8*, i64, i32) - -declare void @llvm.memset.i32(i8*, i8, i32, i32) - -declare void @llvm.memmove.i32(i8*, i8*, i32, i32) - -declare void @llvm.memmove.i64(i8*, i8*, i64, i32) diff --git a/test/CodeGen/ARM/2007-05-07-jumptoentry.ll b/test/CodeGen/ARM/2007-05-07-jumptoentry.ll deleted file mode 100644 index 26864f1..0000000 --- a/test/CodeGen/ARM/2007-05-07-jumptoentry.ll +++ /dev/null @@ -1,58 +0,0 @@ -; RUN: llc < %s | not grep 1_0 -; This used to create an extra branch to 'entry', LBB1_0. - -; ModuleID = 'bug.bc' -target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:64" -target triple = "arm-apple-darwin8" - %struct.HexxagonMove = type { i8, i8, i32 } - %struct.HexxagonMoveList = type { i32, %struct.HexxagonMove* } - -define void @_ZN16HexxagonMoveList8sortListEv(%struct.HexxagonMoveList* %this) { -entry: - %tmp51 = getelementptr %struct.HexxagonMoveList* %this, i32 0, i32 0 ; <i32*> [#uses=1] - %tmp2 = getelementptr %struct.HexxagonMoveList* %this, i32 0, i32 1 ; <%struct.HexxagonMove**> [#uses=2] - br label %bb49 - -bb1: ; preds = %bb49 - %tmp3 = load %struct.HexxagonMove** %tmp2 ; <%struct.HexxagonMove*> [#uses=5] - %tmp6 = getelementptr %struct.HexxagonMove* %tmp3, i32 %i.1, i32 2 ; <i32*> [#uses=1] - %tmp7 = load i32* %tmp6 ; <i32> [#uses=2] - %tmp12 = add i32 %i.1, 1 ; <i32> [#uses=7] - %tmp14 = getelementptr %struct.HexxagonMove* %tmp3, i32 %tmp12, i32 2 ; <i32*> [#uses=1] - %tmp15 = load i32* %tmp14 ; <i32> [#uses=1] - %tmp16 = icmp slt i32 %tmp7, %tmp15 ; <i1> [#uses=1] - br i1 %tmp16, label %cond_true, label %bb49 - -cond_true: ; preds = %bb1 - %tmp23.0 = getelementptr %struct.HexxagonMove* %tmp3, i32 %i.1, i32 0 ; <i8*> [#uses=2] - %tmp67 = load i8* %tmp23.0 ; <i8> [#uses=1] - %tmp23.1 = getelementptr %struct.HexxagonMove* %tmp3, i32 %i.1, i32 1 ; <i8*> [#uses=1] - %tmp68 = load i8* %tmp23.1 ; <i8> [#uses=1] - %tmp3638 = getelementptr %struct.HexxagonMove* %tmp3, i32 %tmp12, i32 0 ; <i8*> [#uses=1] - tail call void @llvm.memcpy.i32( i8* %tmp23.0, i8* %tmp3638, i32 8, i32 4 ) - %tmp41 = load %struct.HexxagonMove** %tmp2 ; <%struct.HexxagonMove*> [#uses=3] - %tmp44.0 = getelementptr %struct.HexxagonMove* %tmp41, i32 %tmp12, i32 0 ; <i8*> [#uses=1] - store i8 %tmp67, i8* %tmp44.0 - %tmp44.1 = getelementptr %struct.HexxagonMove* %tmp41, i32 %tmp12, i32 1 ; <i8*> [#uses=1] - store i8 %tmp68, i8* %tmp44.1 - %tmp44.2 = getelementptr %struct.HexxagonMove* %tmp41, i32 %tmp12, i32 2 ; <i32*> [#uses=1] - store i32 %tmp7, i32* %tmp44.2 - br label %bb49 - -bb49: ; preds = %bb59, %cond_true, %bb1, %entry - %i.1 = phi i32 [ 0, %entry ], [ %tmp12, %bb1 ], [ %tmp12, %cond_true ], [ 0, %bb59 ] ; <i32> [#uses=5] - %move.2 = phi i32 [ 0, %entry ], [ 1, %cond_true ], [ %move.2, %bb1 ], [ 0, %bb59 ] ; <i32> [#uses=2] - %tmp52 = load i32* %tmp51 ; <i32> [#uses=1] - %tmp53 = add i32 %tmp52, -1 ; <i32> [#uses=1] - %tmp55 = icmp sgt i32 %tmp53, %i.1 ; <i1> [#uses=1] - br i1 %tmp55, label %bb1, label %bb59 - -bb59: ; preds = %bb49 - %tmp61 = icmp eq i32 %move.2, 0 ; <i1> [#uses=1] - br i1 %tmp61, label %return, label %bb49 - -return: ; preds = %bb59 - ret void -} - -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) diff --git a/test/CodeGen/ARM/2007-05-31-RegScavengerInfiniteLoop.ll b/test/CodeGen/ARM/2007-05-31-RegScavengerInfiniteLoop.ll deleted file mode 100644 index 7ba2a19..0000000 --- a/test/CodeGen/ARM/2007-05-31-RegScavengerInfiniteLoop.ll +++ /dev/null @@ -1,237 +0,0 @@ -; RUN: llc < %s -; PR1424 - -target datalayout = "e-p:32:32:32-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" -target triple = "arm-unknown-linux-gnueabi" - %struct.AVClass = type { i8*, i8* (i8*)*, %struct.AVOption* } - %struct.AVCodec = type { i8*, i32, i32, i32, i32 (%struct.AVCodecContext*)*, i32 (%struct.AVCodecContext*, i8*, i32, i8*)*, i32 (%struct.AVCodecContext*)*, i32 (%struct.AVCodecContext*, i8*, i32*, i8*, i32)*, i32, %struct.AVCodec*, void (%struct.AVCodecContext*)*, %struct.AVRational*, i32* } - %struct.AVCodecContext = type { %struct.AVClass*, i32, i32, i32, i32, i32, i8*, i32, %struct.AVRational, i32, i32, i32, i32, i32, void (%struct.AVCodecContext*, %struct.AVFrame*, i32*, i32, i32, i32)*, i32, i32, i32, i32, i32, i32, i32, float, float, i32, i32, i32, i32, float, i32, i32, i32, %struct.AVCodec*, i8*, i32, i32, void (%struct.AVCodecContext*, i8*, i32, i32)*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i8*, [32 x i8], i32, i32, i32, i32, i32, i32, i32, float, i32, i32 (%struct.AVCodecContext*, %struct.AVFrame*)*, void (%struct.AVCodecContext*, %struct.AVFrame*)*, i32, i32, i32, i32, i8*, i8*, float, float, i32, %struct.RcOverride*, i32, i8*, i32, i32, i32, float, float, float, float, i32, float, float, float, float, float, i32, i32, i32, i32*, i32, i32, i32, i32, %struct.AVRational, %struct.AVFrame*, i32, i32, [4 x i64], i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32 (%struct.AVCodecContext*, i32*)*, i32, i32, i32, i32, i32, i32, i8*, i32, i32, i32, i32, i32, i32, i16*, i16*, i32, i32, i32, i32, %struct.AVPaletteControl*, i32, i32 (%struct.AVCodecContext*, %struct.AVFrame*)*, i32, i32, i32, i32, i32, i32, i32, i32 (%struct.AVCodecContext*, i32 (%struct.AVCodecContext*, i8*)*, i8**, i32*, i32)*, i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i32, i32, i32, float, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i64 } - %struct.AVEvalExpr = type opaque - %struct.AVFrame = type { [4 x i8*], [4 x i32], [4 x i8*], i32, i32, i64, i32, i32, i32, i32, i32, i8*, i32, i8*, [2 x [2 x i16]*], i32*, i8, i8*, [4 x i64], i32, i32, i32, i32, i32, %struct.AVPanScan*, i32, i32, i16*, [2 x i8*] } - %struct.AVOption = type opaque - %struct.AVPaletteControl = type { i32, [256 x i32] } - %struct.AVPanScan = type { i32, i32, i32, [3 x [2 x i16]] } - %struct.AVRational = type { i32, i32 } - %struct.BlockNode = type { i16, i16, i8, [3 x i8], i8, i8 } - %struct.DSPContext = type { void (i16*, i8*, i32)*, void (i16*, i8*, i8*, i32)*, void (i16*, i8*, i32)*, void (i16*, i8*, i32)*, void (i16*, i8*, i32)*, void (i8*, i16*, i32)*, void (i8*, i16*, i32)*, i32 (i16*)*, void (i8*, i8*, i32, i32, i32, i32, i32)*, void (i8*, i8*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32)*, void (i16*)*, i32 (i8*, i32)*, i32 (i8*, i32)*, [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], [5 x i32 (i8*, i8*, i8*, i32, i32)*], i32 (i8*, i16*, i32)*, [4 x [4 x void (i8*, i8*, i32, i32)*]], [4 x [4 x void (i8*, i8*, i32, i32)*]], [4 x [4 x void (i8*, i8*, i32, i32)*]], [4 x [4 x void (i8*, i8*, i32, i32)*]], [2 x void (i8*, i8*, i8*, i32, i32)*], [11 x void (i8*, i8*, i32, i32, i32)*], [11 x void (i8*, i8*, i32, i32, i32)*], [2 x [16 x void (i8*, i8*, i32)*]], [2 x [16 x void (i8*, i8*, i32)*]], [2 x [16 x void (i8*, i8*, i32)*]], [2 x [16 x void (i8*, i8*, i32)*]], [8 x void (i8*, i8*, i32)*], [3 x void (i8*, i8*, i32, i32, i32, i32)*], [3 x void (i8*, i8*, i32, i32, i32, i32)*], [3 x void (i8*, i8*, i32, i32, i32, i32)*], [4 x [16 x void (i8*, i8*, i32)*]], [4 x [16 x void (i8*, i8*, i32)*]], [4 x [16 x void (i8*, i8*, i32)*]], [4 x [16 x void (i8*, i8*, i32)*]], [10 x void (i8*, i32, i32, i32, i32)*], [10 x void (i8*, i8*, i32, i32, i32, i32, i32)*], [2 x [16 x void (i8*, i8*, i32)*]], [2 x [16 x void (i8*, i8*, i32)*]], void (i8*, i32, i32, i32, i32, i32, i32)*, void (i8*, i32, i32, i32, i32, i32, i32)*, void (i8*, i32, i32, i32, i32, i32, i32)*, void (i8*, i32, i32, i32, i32, i32, i32)*, void (i8*, i16*, i32)*, [2 x [4 x i32 (i8*, i8*, i8*, i32, i32)*]], void (i8*, i8*, i32)*, void (i8*, i8*, i8*, i32)*, void (i8*, i8*, i8*, i32, i32*, i32*)*, void (i32*, i32*, i32)*, void (i8*, i32, i32, i32, i8*)*, void (i8*, i32, i32, i32, i8*)*, void (i8*, i32, i32, i32, i8*)*, void (i8*, i32, i32, i32, i8*)*, void (i8*, i32, i32, i32)*, void (i8*, i32, i32, i32)*, void ([4 x [4 x i16]]*, i8*, [40 x i8]*, [40 x [2 x i16]]*, i32, i32, i32, i32, i32)*, void (i8*, i32, i32)*, void (i8*, i32, i32)*, void (i8*, i32)*, void (float*, float*, i32)*, void (float*, float*, i32)*, void (float*, float*, float*, i32)*, void (float*, float*, float*, float*, i32, i32, i32)*, void (i16*, float*, i32)*, void (i16*)*, void (i16*)*, void (i16*)*, void (i8*, i32, i16*)*, void (i8*, i32, i16*)*, [64 x i8], i32, i32 (i16*, i16*, i16*, i32)*, void (i16*, i16*, i32)*, void (i8*, i16*, i32)*, void (i8*, i16*, i32)*, void (i8*, i16*, i32)*, void (i8*, i16*, i32)*, void ([4 x i16]*)*, void (i32*, i32*, i32*, i32*, i32*, i32*, i32)*, void (i32*, i32)*, void (i8*, i32, i8**, i32, i32, i32, i32, i32, %struct.slice_buffer*, i32, i8*)*, void (i8*, i32, i32)*, [4 x void (i8*, i32, i8*, i32, i32, i32)*], void (i16*)*, void (i16*, i32)*, void (i16*, i32)*, void (i16*, i32)*, void (i8*, i32)*, void (i8*, i32)*, [16 x void (i8*, i8*, i32, i32)*] } - %struct.FILE = type { i32, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, %struct._IO_marker*, %struct.FILE*, i32, i32, i32, i16, i8, [1 x i8], i8*, i64, i8*, i8*, i8*, i8*, i32, i32, [40 x i8] } - %struct.GetBitContext = type { i8*, i8*, i32*, i32, i32, i32, i32 } - %struct.MJpegContext = type opaque - %struct.MotionEstContext = type { %struct.AVCodecContext*, i32, [4 x [2 x i32]], [4 x [2 x i32]], i8*, i8*, [2 x i8*], i8*, i32, i32*, i32*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, [4 x [4 x i8*]], [4 x [4 x i8*]], i32, i32, i32, i32, i32, [4 x void (i8*, i8*, i32, i32)*]*, [4 x void (i8*, i8*, i32, i32)*]*, [16 x void (i8*, i8*, i32)*]*, [16 x void (i8*, i8*, i32)*]*, [4097 x i8]*, i8*, i32 (%struct.MpegEncContext*, i32*, i32*, i32, i32, i32, i32, i32)* } - %struct.MpegEncContext = type { %struct.AVCodecContext*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, %struct.PutBitContext, i32, i32, i32, i32, i32, i32, i64, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, %struct.Picture*, %struct.Picture**, %struct.Picture**, i32, i32, [8 x %struct.MpegEncContext*], %struct.Picture, %struct.Picture, %struct.Picture, %struct.Picture, %struct.Picture*, %struct.Picture*, %struct.Picture*, [3 x i8*], [3 x i32], i16*, [3 x i16*], [20 x i16], i32, i32, i8*, i8*, i8*, i8*, i8*, [16 x i16]*, [3 x [16 x i16]*], i32, i8*, i32, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i32, i32, i32, i32, i32*, i32, i32, i32, i32, i32, i32, i32, [5 x i32], i32, i32, i32, i32, %struct.DSPContext, i32, i32, [2 x i16]*, [2 x i16]*, [2 x i16]*, [2 x i16]*, [2 x i16]*, [2 x i16]*, [2 x [2 x [2 x i16]*]], [2 x [2 x [2 x [2 x i16]*]]], [2 x i16]*, [2 x i16]*, [2 x i16]*, [2 x i16]*, [2 x i16]*, [2 x i16]*, [2 x [2 x [2 x i16]*]], [2 x [2 x [2 x [2 x i16]*]]], [2 x i8*], [2 x [2 x i8*]], i32, i32, i32, [2 x [4 x [2 x i32]]], [2 x [2 x i32]], [2 x [2 x [2 x i32]]], i8*, [2 x [64 x i16]], %struct.MotionEstContext, i32, i32, i32, i32, i32, i32, i16*, [6 x i32], [6 x i32], [3 x i8*], i32*, [64 x i16], [64 x i16], [64 x i16], [64 x i16], i32, i32, i32, i32, i32, i8*, i8*, i8*, i8*, i8*, i8*, [8 x i32], [64 x i32]*, [64 x i32]*, [2 x [64 x i16]]*, [2 x [64 x i16]]*, [12 x i32], %struct.ScanTable, %struct.ScanTable, %struct.ScanTable, %struct.ScanTable, [64 x i32]*, [2 x i32], [64 x i16]*, i8*, i64, i64, i32, i32, %struct.RateControlContext, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i8*, i32, i32, %struct.GetBitContext, i32, i32, i32, %struct.ParseContext, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i64, i64, i16, i16, i16, i16, i32, i32, i32, i32, i32, i32, i32, i32, i32, [2 x [2 x i32]], [2 x [2 x i32]], [2 x i32], i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, %struct.PutBitContext, %struct.PutBitContext, i32, i32, i32, i32, i32, i32, i8*, i32, i32, i32, i32, i32, [3 x i32], %struct.MJpegContext*, [3 x i32], [3 x i32], i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, [2 x [65 x [65 x [2 x i32]]]]*, i32, i32, %struct.GetBitContext, i32, i32, i32, i8*, i32, [2 x [2 x i32]], i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, [2 x i32], i32, i32, i32, i32, i8*, i32, [12 x i16*], [64 x i16]*, [8 x [64 x i16]]*, i32 (%struct.MpegEncContext*, [64 x i16]*)*, void (%struct.MpegEncContext*, i16*, i32, i32)*, void (%struct.MpegEncContext*, i16*, i32, i32)*, void (%struct.MpegEncContext*, i16*, i32, i32)*, void (%struct.MpegEncContext*, i16*, i32, i32)*, void (%struct.MpegEncContext*, i16*, i32, i32)*, void (%struct.MpegEncContext*, i16*, i32, i32)*, void (%struct.MpegEncContext*, i16*, i32, i32)*, void (%struct.MpegEncContext*, i16*, i32, i32)*, void (%struct.MpegEncContext*, i16*, i32, i32)*, void (%struct.MpegEncContext*, i16*, i32, i32)*, i32 (%struct.MpegEncContext*, i16*, i32, i32, i32*)*, i32 (%struct.MpegEncContext*, i16*, i32, i32, i32*)*, void (%struct.MpegEncContext*, i16*)* } - %struct.ParseContext = type { i8*, i32, i32, i32, i32, i32, i32, i32 } - %struct.Picture = type { [4 x i8*], [4 x i32], [4 x i8*], i32, i32, i64, i32, i32, i32, i32, i32, i8*, i32, i8*, [2 x [2 x i16]*], i32*, i8, i8*, [4 x i64], i32, i32, i32, i32, i32, %struct.AVPanScan*, i32, i32, i16*, [2 x i8*], [3 x i8*], [2 x [2 x i16]*], i32*, [2 x i32], i32, i32, i32, i32, [2 x [16 x i32]], [2 x i32], i32, i32, i16*, i16*, i8*, i32*, i32 } - %struct.Plane = type { i32, i32, [8 x [4 x %struct.SubBand]] } - %struct.Predictor = type { double, double, double } - %struct.PutBitContext = type { i32, i32, i8*, i8*, i8* } - %struct.RangeCoder = type { i32, i32, i32, i32, [256 x i8], [256 x i8], i8*, i8*, i8* } - %struct.RateControlContext = type { %struct.FILE*, i32, %struct.RateControlEntry*, double, [5 x %struct.Predictor], double, double, double, double, double, [5 x double], i32, i32, [5 x i64], [5 x i64], [5 x i64], [5 x i64], [5 x i32], i32, i8*, float, i32, %struct.AVEvalExpr* } - %struct.RateControlEntry = type { i32, float, i32, i32, i32, i32, i32, i64, i32, float, i32, i32, i32, i32, i32, i32 } - %struct.RcOverride = type { i32, i32, i32, float } - %struct.ScanTable = type { i8*, [64 x i8], [64 x i8] } - %struct.SnowContext = type { %struct.AVCodecContext*, %struct.RangeCoder, %struct.DSPContext, %struct.AVFrame, %struct.AVFrame, %struct.AVFrame, [8 x %struct.AVFrame], %struct.AVFrame, [32 x i8], [4224 x i8], i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, [8 x [2 x i16]*], [8 x i32*], i32*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, [4 x %struct.Plane], %struct.BlockNode*, [1024 x i32], i32, %struct.slice_buffer, %struct.MpegEncContext } - %struct.SubBand = type { i32, i32, i32, i32, i32, i32*, i32, i32, i32, %struct.x_and_coeff*, %struct.SubBand*, [519 x [32 x i8]] } - %struct._IO_marker = type { %struct._IO_marker*, %struct.FILE*, i32 } - %struct.slice_buffer = type { i32**, i32**, i32, i32, i32, i32, i32* } - %struct.x_and_coeff = type { i16, i16 } - -define fastcc void @iterative_me(%struct.SnowContext* %s) { -entry: - %state = alloca [4224 x i8], align 8 ; <[4224 x i8]*> [#uses=0] - %best_rd4233 = alloca i32, align 4 ; <i32*> [#uses=0] - %tmp21 = getelementptr %struct.SnowContext* %s, i32 0, i32 36 ; <i32*> [#uses=2] - br label %bb4198 - -bb79: ; preds = %bb4189.preheader - br i1 false, label %cond_next239, label %cond_true - -cond_true: ; preds = %bb79 - ret void - -cond_next239: ; preds = %bb79 - %tmp286 = alloca i8, i32 0 ; <i8*> [#uses=0] - ret void - -bb4198: ; preds = %bb4189.preheader, %entry - br i1 false, label %bb4189.preheader, label %bb4204 - -bb4189.preheader: ; preds = %bb4198 - br i1 false, label %bb79, label %bb4198 - -bb4204: ; preds = %bb4198 - br i1 false, label %bb4221, label %cond_next4213 - -cond_next4213: ; preds = %bb4204 - ret void - -bb4221: ; preds = %bb4204 - br i1 false, label %bb5242.preheader, label %UnifiedReturnBlock - -bb5242.preheader: ; preds = %bb4221 - br label %bb5242 - -bb4231: ; preds = %bb5233 - %tmp4254.sum = add i32 0, 1 ; <i32> [#uses=2] - br i1 false, label %bb4559, label %cond_next4622 - -bb4559: ; preds = %bb4231 - ret void - -cond_next4622: ; preds = %bb4231 - %tmp4637 = load i16* null ; <i16> [#uses=1] - %tmp46374638 = sext i16 %tmp4637 to i32 ; <i32> [#uses=1] - %tmp4642 = load i16* null ; <i16> [#uses=1] - %tmp46424643 = sext i16 %tmp4642 to i32 ; <i32> [#uses=1] - %tmp4648 = load i16* null ; <i16> [#uses=1] - %tmp46484649 = sext i16 %tmp4648 to i32 ; <i32> [#uses=1] - %tmp4653 = getelementptr %struct.BlockNode* null, i32 %tmp4254.sum, i32 0 ; <i16*> [#uses=1] - %tmp4654 = load i16* %tmp4653 ; <i16> [#uses=1] - %tmp46544655 = sext i16 %tmp4654 to i32 ; <i32> [#uses=1] - %tmp4644 = add i32 %tmp46374638, 2 ; <i32> [#uses=1] - %tmp4650 = add i32 %tmp4644, %tmp46424643 ; <i32> [#uses=1] - %tmp4656 = add i32 %tmp4650, %tmp46484649 ; <i32> [#uses=1] - %tmp4657 = add i32 %tmp4656, %tmp46544655 ; <i32> [#uses=2] - %tmp4658 = ashr i32 %tmp4657, 2 ; <i32> [#uses=1] - %tmp4662 = load i16* null ; <i16> [#uses=1] - %tmp46624663 = sext i16 %tmp4662 to i32 ; <i32> [#uses=1] - %tmp4672 = getelementptr %struct.BlockNode* null, i32 0, i32 1 ; <i16*> [#uses=1] - %tmp4673 = load i16* %tmp4672 ; <i16> [#uses=1] - %tmp46734674 = sext i16 %tmp4673 to i32 ; <i32> [#uses=1] - %tmp4678 = getelementptr %struct.BlockNode* null, i32 %tmp4254.sum, i32 1 ; <i16*> [#uses=1] - %tmp4679 = load i16* %tmp4678 ; <i16> [#uses=1] - %tmp46794680 = sext i16 %tmp4679 to i32 ; <i32> [#uses=1] - %tmp4669 = add i32 %tmp46624663, 2 ; <i32> [#uses=1] - %tmp4675 = add i32 %tmp4669, 0 ; <i32> [#uses=1] - %tmp4681 = add i32 %tmp4675, %tmp46734674 ; <i32> [#uses=1] - %tmp4682 = add i32 %tmp4681, %tmp46794680 ; <i32> [#uses=2] - %tmp4683 = ashr i32 %tmp4682, 2 ; <i32> [#uses=1] - %tmp4703 = load i32* %tmp21 ; <i32> [#uses=1] - %tmp4707 = shl i32 %tmp4703, 0 ; <i32> [#uses=4] - %tmp4710 = load %struct.BlockNode** null ; <%struct.BlockNode*> [#uses=6] - %tmp4713 = mul i32 %tmp4707, %mb_y.4 ; <i32> [#uses=1] - %tmp4715 = add i32 %tmp4713, %mb_x.7 ; <i32> [#uses=7] - store i8 0, i8* null - store i8 0, i8* null - %tmp47594761 = bitcast %struct.BlockNode* null to i8* ; <i8*> [#uses=2] - call void @llvm.memcpy.i32( i8* null, i8* %tmp47594761, i32 10, i32 0 ) - %tmp4716.sum5775 = add i32 %tmp4715, 1 ; <i32> [#uses=1] - %tmp4764 = getelementptr %struct.BlockNode* %tmp4710, i32 %tmp4716.sum5775 ; <%struct.BlockNode*> [#uses=1] - %tmp47644766 = bitcast %struct.BlockNode* %tmp4764 to i8* ; <i8*> [#uses=1] - %tmp4716.sum5774 = add i32 %tmp4715, %tmp4707 ; <i32> [#uses=0] - %tmp47704772 = bitcast %struct.BlockNode* null to i8* ; <i8*> [#uses=1] - %tmp4774 = add i32 %tmp4707, 1 ; <i32> [#uses=1] - %tmp4716.sum5773 = add i32 %tmp4774, %tmp4715 ; <i32> [#uses=1] - %tmp4777 = getelementptr %struct.BlockNode* %tmp4710, i32 %tmp4716.sum5773 ; <%struct.BlockNode*> [#uses=1] - %tmp47774779 = bitcast %struct.BlockNode* %tmp4777 to i8* ; <i8*> [#uses=1] - %tmp4781 = icmp slt i32 %mb_x.7, 0 ; <i1> [#uses=1] - %tmp4788 = or i1 %tmp4781, %tmp4784 ; <i1> [#uses=2] - br i1 %tmp4788, label %cond_true4791, label %cond_next4794 - -cond_true4791: ; preds = %cond_next4622 - unreachable - -cond_next4794: ; preds = %cond_next4622 - %tmp4797 = icmp slt i32 %mb_x.7, %tmp4707 ; <i1> [#uses=1] - br i1 %tmp4797, label %cond_next4803, label %cond_true4800 - -cond_true4800: ; preds = %cond_next4794 - unreachable - -cond_next4803: ; preds = %cond_next4794 - %tmp4825 = ashr i32 %tmp4657, 12 ; <i32> [#uses=1] - shl i32 %tmp4682, 4 ; <i32>:0 [#uses=1] - %tmp4828 = and i32 %0, -64 ; <i32> [#uses=1] - %tmp4831 = getelementptr %struct.BlockNode* %tmp4710, i32 %tmp4715, i32 2 ; <i8*> [#uses=0] - %tmp4826 = add i32 %tmp4828, %tmp4825 ; <i32> [#uses=1] - %tmp4829 = add i32 %tmp4826, 0 ; <i32> [#uses=1] - %tmp4835 = add i32 %tmp4829, 0 ; <i32> [#uses=1] - store i32 %tmp4835, i32* null - %tmp48534854 = trunc i32 %tmp4658 to i16 ; <i16> [#uses=1] - %tmp4856 = getelementptr %struct.BlockNode* %tmp4710, i32 %tmp4715, i32 0 ; <i16*> [#uses=1] - store i16 %tmp48534854, i16* %tmp4856 - %tmp48574858 = trunc i32 %tmp4683 to i16 ; <i16> [#uses=1] - %tmp4860 = getelementptr %struct.BlockNode* %tmp4710, i32 %tmp4715, i32 1 ; <i16*> [#uses=1] - store i16 %tmp48574858, i16* %tmp4860 - %tmp4866 = getelementptr %struct.BlockNode* %tmp4710, i32 %tmp4715, i32 4 ; <i8*> [#uses=0] - br i1 false, label %bb4933, label %cond_false4906 - -cond_false4906: ; preds = %cond_next4803 - call void @llvm.memcpy.i32( i8* %tmp47594761, i8* null, i32 10, i32 0 ) - call void @llvm.memcpy.i32( i8* %tmp47644766, i8* null, i32 10, i32 0 ) - call void @llvm.memcpy.i32( i8* %tmp47704772, i8* null, i32 10, i32 0 ) - call void @llvm.memcpy.i32( i8* %tmp47774779, i8* null, i32 10, i32 0 ) - br label %bb5215 - -bb4933: ; preds = %bb5215, %cond_next4803 - br i1 false, label %cond_true4944, label %bb5215 - -cond_true4944: ; preds = %bb4933 - %tmp4982 = load i32* %tmp21 ; <i32> [#uses=1] - %tmp4986 = shl i32 %tmp4982, 0 ; <i32> [#uses=2] - %tmp4992 = mul i32 %tmp4986, %mb_y.4 ; <i32> [#uses=1] - %tmp4994 = add i32 %tmp4992, %mb_x.7 ; <i32> [#uses=5] - %tmp4995.sum5765 = add i32 %tmp4994, 1 ; <i32> [#uses=1] - %tmp5043 = getelementptr %struct.BlockNode* null, i32 %tmp4995.sum5765 ; <%struct.BlockNode*> [#uses=1] - %tmp50435045 = bitcast %struct.BlockNode* %tmp5043 to i8* ; <i8*> [#uses=2] - call void @llvm.memcpy.i32( i8* null, i8* %tmp50435045, i32 10, i32 0 ) - %tmp4995.sum5764 = add i32 %tmp4994, %tmp4986 ; <i32> [#uses=1] - %tmp5049 = getelementptr %struct.BlockNode* null, i32 %tmp4995.sum5764 ; <%struct.BlockNode*> [#uses=1] - %tmp50495051 = bitcast %struct.BlockNode* %tmp5049 to i8* ; <i8*> [#uses=2] - call void @llvm.memcpy.i32( i8* null, i8* %tmp50495051, i32 10, i32 0 ) - %tmp4995.sum5763 = add i32 0, %tmp4994 ; <i32> [#uses=1] - %tmp5056 = getelementptr %struct.BlockNode* null, i32 %tmp4995.sum5763 ; <%struct.BlockNode*> [#uses=1] - %tmp50565058 = bitcast %struct.BlockNode* %tmp5056 to i8* ; <i8*> [#uses=1] - br i1 %tmp4788, label %cond_true5070, label %cond_next5073 - -cond_true5070: ; preds = %cond_true4944 - unreachable - -cond_next5073: ; preds = %cond_true4944 - %tmp5139 = getelementptr %struct.BlockNode* null, i32 %tmp4994, i32 1 ; <i16*> [#uses=0] - %tmp5145 = getelementptr %struct.BlockNode* null, i32 %tmp4994, i32 4 ; <i8*> [#uses=0] - call void @llvm.memcpy.i32( i8* %tmp50435045, i8* null, i32 10, i32 0 ) - call void @llvm.memcpy.i32( i8* %tmp50495051, i8* null, i32 10, i32 0 ) - call void @llvm.memcpy.i32( i8* %tmp50565058, i8* null, i32 10, i32 0 ) - br label %bb5215 - -bb5215: ; preds = %cond_next5073, %bb4933, %cond_false4906 - %i4232.3 = phi i32 [ 0, %cond_false4906 ], [ 0, %cond_next5073 ], [ 0, %bb4933 ] ; <i32> [#uses=1] - %tmp5217 = icmp slt i32 %i4232.3, 4 ; <i1> [#uses=1] - br i1 %tmp5217, label %bb4933, label %bb5220 - -bb5220: ; preds = %bb5215 - br i1 false, label %bb5230, label %cond_true5226 - -cond_true5226: ; preds = %bb5220 - ret void - -bb5230: ; preds = %bb5220 - %indvar.next = add i32 %indvar, 1 ; <i32> [#uses=1] - br label %bb5233 - -bb5233: ; preds = %bb5233.preheader, %bb5230 - %indvar = phi i32 [ 0, %bb5233.preheader ], [ %indvar.next, %bb5230 ] ; <i32> [#uses=2] - %mb_x.7 = shl i32 %indvar, 1 ; <i32> [#uses=4] - br i1 false, label %bb4231, label %bb5239 - -bb5239: ; preds = %bb5233 - %indvar.next37882 = add i32 %indvar37881, 1 ; <i32> [#uses=1] - br label %bb5242 - -bb5242: ; preds = %bb5239, %bb5242.preheader - %indvar37881 = phi i32 [ 0, %bb5242.preheader ], [ %indvar.next37882, %bb5239 ] ; <i32> [#uses=2] - %mb_y.4 = shl i32 %indvar37881, 1 ; <i32> [#uses=3] - br i1 false, label %bb5233.preheader, label %bb5248 - -bb5233.preheader: ; preds = %bb5242 - %tmp4784 = icmp slt i32 %mb_y.4, 0 ; <i1> [#uses=1] - br label %bb5233 - -bb5248: ; preds = %bb5242 - ret void - -UnifiedReturnBlock: ; preds = %bb4221 - ret void -} - -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) diff --git a/test/CodeGen/ARM/2009-03-07-SpillerBug.ll b/test/CodeGen/ARM/2009-03-07-SpillerBug.ll index 0ec17ae..377bbd2 100644 --- a/test/CodeGen/ARM/2009-03-07-SpillerBug.ll +++ b/test/CodeGen/ARM/2009-03-07-SpillerBug.ll @@ -1,16 +1,15 @@ ; RUN: llc < %s -mtriple=armv6-apple-darwin9 -mattr=+vfp2 ; rdar://6653182 - %struct.ggBRDF = type { i32 (...)** } - %struct.ggPoint2 = type { [2 x double] } - %struct.ggPoint3 = type { [3 x double] } - %struct.ggSpectrum = type { [8 x float] } - %struct.ggSphere = type { %struct.ggPoint3, double } - %struct.mrDiffuseAreaSphereLuminaire = type { %struct.mrSphere, %struct.ggSpectrum } - %struct.mrDiffuseCosineSphereLuminaire = type { %struct.mrDiffuseAreaSphereLuminaire } - %struct.mrSphere = type { %struct.ggBRDF, %struct.ggSphere } -declare void @llvm.memcpy.i32(i8* nocapture, i8* nocapture, i32, i32) nounwind +%struct.ggBRDF = type { i32 (...)** } +%struct.ggPoint2 = type { [2 x double] } +%struct.ggPoint3 = type { [3 x double] } +%struct.ggSpectrum = type { [8 x float] } +%struct.ggSphere = type { %struct.ggPoint3, double } +%struct.mrDiffuseAreaSphereLuminaire = type { %struct.mrSphere, %struct.ggSpectrum } +%struct.mrDiffuseCosineSphereLuminaire = type { %struct.mrDiffuseAreaSphereLuminaire } +%struct.mrSphere = type { %struct.ggBRDF, %struct.ggSphere } declare double @llvm.sqrt.f64(double) nounwind readonly @@ -20,59 +19,61 @@ declare double @acos(double) nounwind readonly define i32 @_ZNK34mrDiffuseSolidAngleSphereLuminaire18selectVisiblePointERK8ggPoint3RK9ggVector3RK8ggPoint2dRS0_Rd(%struct.mrDiffuseCosineSphereLuminaire* nocapture %this, %struct.ggPoint3* nocapture %x, %struct.ggPoint3* nocapture %unnamed_arg, %struct.ggPoint2* nocapture %uv, double %unnamed_arg2, %struct.ggPoint3* nocapture %on_light, double* nocapture %invProb) nounwind { entry: - %0 = call double @llvm.sqrt.f64(double 0.000000e+00) nounwind ; <double> [#uses=4] - %1 = fcmp ult double 0.000000e+00, %0 ; <i1> [#uses=1] - br i1 %1, label %bb3, label %bb7 + %0 = call double @llvm.sqrt.f64(double 0.000000e+00) nounwind + %1 = fcmp ult double 0.000000e+00, %0 + br i1 %1, label %bb3, label %bb7 -bb3: ; preds = %entry - %2 = fdiv double 1.000000e+00, 0.000000e+00 ; <double> [#uses=1] - %3 = fmul double 0.000000e+00, %2 ; <double> [#uses=2] - %4 = call double @llvm.sqrt.f64(double 0.000000e+00) nounwind ; <double> [#uses=1] - %5 = fdiv double 1.000000e+00, %4 ; <double> [#uses=2] - %6 = fmul double %3, %5 ; <double> [#uses=2] - %7 = fmul double 0.000000e+00, %5 ; <double> [#uses=2] - %8 = fmul double %3, %7 ; <double> [#uses=1] - %9 = fsub double %8, 0.000000e+00 ; <double> [#uses=1] - %10 = fmul double 0.000000e+00, %6 ; <double> [#uses=1] - %11 = fsub double 0.000000e+00, %10 ; <double> [#uses=1] - %12 = fsub double -0.000000e+00, %11 ; <double> [#uses=1] - %13 = fmul double %0, %0 ; <double> [#uses=2] - %14 = fsub double %13, 0.000000e+00 ; <double> [#uses=1] - %15 = call double @llvm.sqrt.f64(double %14) ; <double> [#uses=1] - %16 = fmul double 0.000000e+00, %15 ; <double> [#uses=1] - %17 = fdiv double %16, %0 ; <double> [#uses=1] - %18 = fadd double 0.000000e+00, %17 ; <double> [#uses=1] - %19 = call double @acos(double %18) nounwind readonly ; <double> [#uses=1] - %20 = load double* null, align 4 ; <double> [#uses=1] - %21 = fmul double %20, 0x401921FB54442D18 ; <double> [#uses=1] - %22 = call double @sin(double %19) nounwind readonly ; <double> [#uses=2] - %23 = fmul double %22, 0.000000e+00 ; <double> [#uses=2] - %24 = fmul double %6, %23 ; <double> [#uses=1] - %25 = fmul double %7, %23 ; <double> [#uses=1] - %26 = call double @sin(double %21) nounwind readonly ; <double> [#uses=1] - %27 = fmul double %22, %26 ; <double> [#uses=2] - %28 = fmul double %9, %27 ; <double> [#uses=1] - %29 = fmul double %27, %12 ; <double> [#uses=1] - %30 = fadd double %24, %28 ; <double> [#uses=1] - %31 = fadd double 0.000000e+00, %29 ; <double> [#uses=1] - %32 = fadd double %25, 0.000000e+00 ; <double> [#uses=1] - %33 = fadd double %30, 0.000000e+00 ; <double> [#uses=1] - %34 = fadd double %31, 0.000000e+00 ; <double> [#uses=1] - %35 = fadd double %32, 0.000000e+00 ; <double> [#uses=1] - %36 = bitcast %struct.ggPoint3* %x to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32(i8* null, i8* %36, i32 24, i32 4) nounwind - store double %33, double* null, align 8 - br i1 false, label %_Z20ggRaySphereIntersectRK6ggRay3RK8ggSphereddRd.exit, label %bb5.i.i.i +bb3: ; preds = %entry + %2 = fdiv double 1.000000e+00, 0.000000e+00 + %3 = fmul double 0.000000e+00, %2 + %4 = call double @llvm.sqrt.f64(double 0.000000e+00) nounwind + %5 = fdiv double 1.000000e+00, %4 + %6 = fmul double %3, %5 + %7 = fmul double 0.000000e+00, %5 + %8 = fmul double %3, %7 + %9 = fsub double %8, 0.000000e+00 + %10 = fmul double 0.000000e+00, %6 + %11 = fsub double 0.000000e+00, %10 + %12 = fsub double -0.000000e+00, %11 + %13 = fmul double %0, %0 + %14 = fsub double %13, 0.000000e+00 + %15 = call double @llvm.sqrt.f64(double %14) + %16 = fmul double 0.000000e+00, %15 + %17 = fdiv double %16, %0 + %18 = fadd double 0.000000e+00, %17 + %19 = call double @acos(double %18) nounwind readonly + %20 = load double* null, align 4 + %21 = fmul double %20, 0x401921FB54442D18 + %22 = call double @sin(double %19) nounwind readonly + %23 = fmul double %22, 0.000000e+00 + %24 = fmul double %6, %23 + %25 = fmul double %7, %23 + %26 = call double @sin(double %21) nounwind readonly + %27 = fmul double %22, %26 + %28 = fmul double %9, %27 + %29 = fmul double %27, %12 + %30 = fadd double %24, %28 + %31 = fadd double 0.000000e+00, %29 + %32 = fadd double %25, 0.000000e+00 + %33 = fadd double %30, 0.000000e+00 + %34 = fadd double %31, 0.000000e+00 + %35 = fadd double %32, 0.000000e+00 + %36 = bitcast %struct.ggPoint3* %x to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* null, i8* %36, i32 24, i32 4, i1 false) + store double %33, double* null, align 8 + br i1 false, label %_Z20ggRaySphereIntersectRK6ggRay3RK8ggSphereddRd.exit, label %bb5.i.i.i -bb5.i.i.i: ; preds = %bb3 - unreachable +bb5.i.i.i: ; preds = %bb3 + unreachable -_Z20ggRaySphereIntersectRK6ggRay3RK8ggSphereddRd.exit: ; preds = %bb3 - %37 = fsub double %13, 0.000000e+00 ; <double> [#uses=0] - %38 = fsub double -0.000000e+00, %34 ; <double> [#uses=0] - %39 = fsub double -0.000000e+00, %35 ; <double> [#uses=0] - ret i32 1 +_Z20ggRaySphereIntersectRK6ggRay3RK8ggSphereddRd.exit: ; preds = %bb3 + %37 = fsub double %13, 0.000000e+00 + %38 = fsub double -0.000000e+00, %34 + %39 = fsub double -0.000000e+00, %35 + ret i32 1 -bb7: ; preds = %entry - ret i32 0 +bb7: ; preds = %entry + ret i32 0 } + +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/CodeGen/ARM/2009-06-12-RegScavengerAssert.ll b/test/CodeGen/ARM/2009-06-12-RegScavengerAssert.ll deleted file mode 100644 index 27888d7..0000000 --- a/test/CodeGen/ARM/2009-06-12-RegScavengerAssert.ll +++ /dev/null @@ -1,77 +0,0 @@ -; RUN: llc < %s -mtriple=armv6-apple-darwin - - type { i32, i32, %struct.D_Sym**, [3 x %struct.D_Sym*] } ; type %0 - type { i32, %struct.D_Reduction** } ; type %1 - type { i32, %struct.D_RightEpsilonHint* } ; type %2 - type { i32, %struct.D_ErrorRecoveryHint* } ; type %3 - type { i32, i32, %struct.D_Reduction**, [3 x %struct.D_Reduction*] } ; type %4 - %struct.D_ErrorRecoveryHint = type { i16, i16, i8* } - %struct.D_ParseNode = type { i32, %struct.d_loc_t, i8*, i8*, %struct.D_Scope*, void (%struct.D_Parser*, %struct.d_loc_t*, i8**)*, i8*, i8* } - %struct.D_Parser = type { i8*, void (%struct.D_Parser*, %struct.d_loc_t*, i8**)*, %struct.D_Scope*, void (%struct.D_Parser*)*, %struct.D_ParseNode* (%struct.D_Parser*, i32, %struct.D_ParseNode**)*, void (%struct.D_ParseNode*)*, %struct.d_loc_t, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32 } - %struct.D_ParserTables = type { i32, %struct.D_State*, i16*, i32, i32, %struct.D_Symbol*, void (%struct.D_Parser*, %struct.d_loc_t*, i8**)*, i32, %struct.D_Pass*, i32 } - %struct.D_Pass = type { i8*, i32, i32, i32 } - %struct.D_Reduction = type { i16, i16, i32 (i8*, i8**, i32, i32, %struct.D_Parser*)*, i32 (i8*, i8**, i32, i32, %struct.D_Parser*)*, i16, i16, i32, i32, i32, i32, i32 (i8*, i8**, i32, i32, %struct.D_Parser*)** } - %struct.D_RightEpsilonHint = type { i16, i16, %struct.D_Reduction* } - %struct.D_Scope = type { i8, %struct.D_Sym*, %struct.D_SymHash*, %struct.D_Sym*, %struct.D_Scope*, %struct.D_Scope*, %struct.D_Scope*, %struct.D_Scope*, %struct.D_Scope* } - %struct.D_Shift = type { i16, i8, i8, i32, i32, i32 (i8*, i8**, i32, i32, %struct.D_Parser*)* } - %struct.D_State = type { i8*, i32, %1, %2, %3, %struct.D_Shift**, i32 (i8**, i32*, i32*, i16*, i32*, i8*, i32*)*, i8*, i8, i8, i8, i8*, %struct.D_Shift***, i32 } - %struct.D_Sym = type { i8*, i32, i32, %struct.D_Sym*, %struct.D_Sym*, i32 } - %struct.D_SymHash = type { i32, i32, %0 } - %struct.D_Symbol = type { i32, i8*, i32 } - %struct.PNode = type { i32, i32, i32, i32, %struct.D_Reduction*, %struct.D_Shift*, i32, %struct.VecPNode, i32, i8, i8, %struct.PNode*, %struct.PNode*, %struct.PNode*, %struct.PNode*, i8*, i8*, %struct.D_Scope*, i8*, %struct.D_ParseNode } - %struct.PNodeHash = type { %struct.PNode**, i32, i32, i32, %struct.PNode* } - %struct.Parser = type { %struct.D_Parser, i8*, i8*, %struct.D_ParserTables*, i32, i32, i32, i32, i32, i32, i32, %struct.PNodeHash, %struct.SNodeHash, %struct.Reduction*, %struct.Shift*, %struct.D_Scope*, %struct.SNode*, i32, %struct.Reduction*, %struct.Shift*, i32, %struct.PNode*, %struct.SNode*, %struct.ZNode*, %4, %struct.ShiftResult*, %struct.D_Shift, %struct.Parser*, i8* } - %struct.Reduction = type { %struct.ZNode*, %struct.SNode*, %struct.D_Reduction*, %struct.SNode*, i32, %struct.Reduction* } - %struct.SNode = type { %struct.D_State*, %struct.D_Scope*, i8*, %struct.d_loc_t, i32, %struct.PNode*, %struct.VecZNode, i32, %struct.SNode*, %struct.SNode* } - %struct.SNodeHash = type { %struct.SNode**, i32, i32, i32, %struct.SNode*, %struct.SNode* } - %struct.Shift = type { %struct.SNode*, %struct.Shift* } - %struct.ShiftResult = type { %struct.D_Shift*, %struct.d_loc_t } - %struct.VecPNode = type { i32, i32, %struct.PNode**, [3 x %struct.PNode*] } - %struct.VecSNode = type { i32, i32, %struct.SNode**, [3 x %struct.SNode*] } - %struct.VecZNode = type { i32, i32, %struct.ZNode**, [3 x %struct.ZNode*] } - %struct.ZNode = type { %struct.PNode*, %struct.VecSNode } - %struct.d_loc_t = type { i8*, i8*, i32, i32, i32 } - -declare void @llvm.memcpy.i32(i8* nocapture, i8* nocapture, i32, i32) nounwind - -define fastcc i32 @exhaustive_parse(%struct.Parser* %p, i32 %state) nounwind { -entry: - store i8* undef, i8** undef, align 4 - %0 = getelementptr %struct.Parser* %p, i32 0, i32 0, i32 6 ; <%struct.d_loc_t*> [#uses=1] - %1 = bitcast %struct.d_loc_t* %0 to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32(i8* undef, i8* %1, i32 20, i32 4) - br label %bb10 - -bb10: ; preds = %bb30, %bb29, %bb26, %entry - br i1 undef, label %bb18, label %bb20 - -bb18: ; preds = %bb10 - br i1 undef, label %bb20, label %bb19 - -bb19: ; preds = %bb18 - br label %bb20 - -bb20: ; preds = %bb19, %bb18, %bb10 - br i1 undef, label %bb21, label %bb22 - -bb21: ; preds = %bb20 - unreachable - -bb22: ; preds = %bb20 - br i1 undef, label %bb24, label %bb26 - -bb24: ; preds = %bb22 - unreachable - -bb26: ; preds = %bb22 - br i1 undef, label %bb10, label %bb29 - -bb29: ; preds = %bb26 - br i1 undef, label %bb10, label %bb30 - -bb30: ; preds = %bb29 - br i1 undef, label %bb31, label %bb10 - -bb31: ; preds = %bb30 - unreachable -} diff --git a/test/CodeGen/ARM/2009-08-31-LSDA-Name.ll b/test/CodeGen/ARM/2009-08-31-LSDA-Name.ll index 397eba4..8bde748 100644 --- a/test/CodeGen/ARM/2009-08-31-LSDA-Name.ll +++ b/test/CodeGen/ARM/2009-08-31-LSDA-Name.ll @@ -1,32 +1,35 @@ ; RUN: llc < %s -mtriple=arm-apple-darwin9 -march=arm | FileCheck %s +; CHECK: L_LSDA_0: + + %struct.A = type { i32* } define void @"\01-[MyFunction Name:]"() { entry: - %save_filt.1 = alloca i32 ; <i32*> [#uses=2] - %save_eptr.0 = alloca i8* ; <i8**> [#uses=2] - %a = alloca %struct.A ; <%struct.A*> [#uses=3] - %eh_exception = alloca i8* ; <i8**> [#uses=5] - %eh_selector = alloca i32 ; <i32*> [#uses=3] - %"alloca point" = bitcast i32 0 to i32 ; <i32> [#uses=0] - call void @_ZN1AC1Ev(%struct.A* %a) - invoke void @_Z3barv() + %save_filt.1 = alloca i32 + %save_eptr.0 = alloca i8* + %a = alloca %struct.A + %eh_exception = alloca i8* + %eh_selector = alloca i32 + %"alloca point" = bitcast i32 0 to i32 + call void @_ZN1AC1Ev(%struct.A* %a) + invoke void @_Z3barv() to label %invcont unwind label %lpad invcont: ; preds = %entry - call void @_ZN1AD1Ev(%struct.A* %a) nounwind + call void @_ZN1AD1Ev(%struct.A* %a) nounwind br label %return bb: ; preds = %ppad - %eh_select = load i32* %eh_selector ; <i32> [#uses=1] + %eh_select = load i32* %eh_selector store i32 %eh_select, i32* %save_filt.1, align 4 - %eh_value = load i8** %eh_exception ; <i8*> [#uses=1] + %eh_value = load i8** %eh_exception store i8* %eh_value, i8** %save_eptr.0, align 4 - call void @_ZN1AD1Ev(%struct.A* %a) nounwind - %0 = load i8** %save_eptr.0, align 4 ; <i8*> [#uses=1] + call void @_ZN1AD1Ev(%struct.A* %a) nounwind + %0 = load i8** %save_eptr.0, align 4 store i8* %0, i8** %eh_exception, align 4 - %1 = load i32* %save_filt.1, align 4 ; <i32> [#uses=1] + %1 = load i32* %save_filt.1, align 4 store i32 %1, i32* %eh_selector, align 4 br label %Unwind @@ -34,10 +37,10 @@ return: ; preds = %invcont ret void lpad: ; preds = %entry - %eh_ptr = call i8* @llvm.eh.exception() ; <i8*> [#uses=1] + %eh_ptr = call i8* @llvm.eh.exception() store i8* %eh_ptr, i8** %eh_exception - %eh_ptr1 = load i8** %eh_exception ; <i8*> [#uses=1] - %eh_select2 = call i32 (i8*, i8*, ...)* @llvm.eh.selector.i32(i8* %eh_ptr1, i8* bitcast (i32 (...)* @__gxx_personality_sj0 to i8*), i32 0) ; <i32> [#uses=1] + %eh_ptr1 = load i8** %eh_exception + %eh_select2 = call i32 (i8*, i8*, ...)* @llvm.eh.selector(i8* %eh_ptr1, i8* bitcast (i32 (...)* @__gxx_personality_sj0 to i8*), i32 0) store i32 %eh_select2, i32* %eh_selector br label %ppad @@ -45,20 +48,20 @@ ppad: ; preds = %lpad br label %bb Unwind: ; preds = %bb - %eh_ptr3 = load i8** %eh_exception ; <i8*> [#uses=1] - call void @_Unwind_SjLj_Resume(i8* %eh_ptr3) + %eh_ptr3 = load i8** %eh_exception + call void @_Unwind_SjLj_Resume(i8* %eh_ptr3) unreachable } define linkonce_odr void @_ZN1AC1Ev(%struct.A* %this) { entry: - %this_addr = alloca %struct.A* ; <%struct.A**> [#uses=2] - %"alloca point" = bitcast i32 0 to i32 ; <i32> [#uses=0] + %this_addr = alloca %struct.A* + %"alloca point" = bitcast i32 0 to i32 store %struct.A* %this, %struct.A** %this_addr - %0 = call i8* @_Znwm(i32 4) ; <i8*> [#uses=1] - %1 = bitcast i8* %0 to i32* ; <i32*> [#uses=1] - %2 = load %struct.A** %this_addr, align 4 ; <%struct.A*> [#uses=1] - %3 = getelementptr inbounds %struct.A* %2, i32 0, i32 0 ; <i32**> [#uses=1] + %0 = call i8* @_Znwm(i32 4) + %1 = bitcast i8* %0 to i32* + %2 = load %struct.A** %this_addr, align 4 + %3 = getelementptr inbounds %struct.A* %2, i32 0, i32 0 store i32* %1, i32** %3, align 4 br label %return @@ -70,14 +73,14 @@ declare i8* @_Znwm(i32) define linkonce_odr void @_ZN1AD1Ev(%struct.A* %this) nounwind { entry: - %this_addr = alloca %struct.A* ; <%struct.A**> [#uses=2] - %"alloca point" = bitcast i32 0 to i32 ; <i32> [#uses=0] + %this_addr = alloca %struct.A* + %"alloca point" = bitcast i32 0 to i32 store %struct.A* %this, %struct.A** %this_addr - %0 = load %struct.A** %this_addr, align 4 ; <%struct.A*> [#uses=1] - %1 = getelementptr inbounds %struct.A* %0, i32 0, i32 0 ; <i32**> [#uses=1] - %2 = load i32** %1, align 4 ; <i32*> [#uses=1] - %3 = bitcast i32* %2 to i8* ; <i8*> [#uses=1] - call void @_ZdlPv(i8* %3) nounwind + %0 = load %struct.A** %this_addr, align 4 + %1 = getelementptr inbounds %struct.A* %0, i32 0, i32 0 + %2 = load i32** %1, align 4 + %3 = bitcast i32* %2 to i8* + call void @_ZdlPv(i8* %3) nounwind br label %bb bb: ; preds = %entry @@ -86,17 +89,16 @@ bb: ; preds = %entry return: ; preds = %bb ret void } -;CHECK: L_LSDA_0: declare void @_ZdlPv(i8*) nounwind declare void @_Z3barv() -declare i8* @llvm.eh.exception() nounwind +declare i8* @llvm.eh.exception() nounwind readonly -declare i32 @llvm.eh.selector.i32(i8*, i8*, ...) nounwind +declare i32 @llvm.eh.selector(i8*, i8*, ...) nounwind -declare i32 @llvm.eh.typeid.for.i32(i8*) nounwind +declare i32 @llvm.eh.typeid.for(i8*) nounwind declare i32 @__gxx_personality_sj0(...) diff --git a/test/CodeGen/ARM/dyn-stackalloc.ll b/test/CodeGen/ARM/dyn-stackalloc.ll index 92e2d13..de2820e 100644 --- a/test/CodeGen/ARM/dyn-stackalloc.ll +++ b/test/CodeGen/ARM/dyn-stackalloc.ll @@ -1,48 +1,48 @@ ; RUN: llc < %s -march=arm - %struct.state = type { i32, %struct.info*, float**, i32, i32, i32, i32, i32, i32, i32, i32, i32, i64, i64, i64, i64, i64, i64, i8* } - %struct.info = type { i32, i32, i32, i32, i32, i32, i32, i8* } +%struct.comment = type { i8**, i32*, i32, i8* } +%struct.info = type { i32, i32, i32, i32, i32, i32, i32, i8* } +%struct.state = type { i32, %struct.info*, float**, i32, i32, i32, i32, i32, i32, i32, i32, i32, i64, i64, i64, i64, i64, i64, i8* } + +@str215 = external global [2 x i8] define void @t1(%struct.state* %v) { - %tmp6 = load i32* null - %tmp8 = alloca float, i32 %tmp6 - store i32 1, i32* null - br i1 false, label %bb123.preheader, label %return - -bb123.preheader: - br i1 false, label %bb43, label %return - -bb43: - call fastcc void @f1( float* %tmp8, float* null, i32 0 ) - %tmp70 = load i32* null - %tmp85 = getelementptr float* %tmp8, i32 0 - call fastcc void @f2( float* null, float* null, float* %tmp85, i32 %tmp70 ) - ret void - -return: - ret void + %tmp6 = load i32* null + %tmp8 = alloca float, i32 %tmp6 + store i32 1, i32* null + br i1 false, label %bb123.preheader, label %return + +bb123.preheader: ; preds = %0 + br i1 false, label %bb43, label %return + +bb43: ; preds = %bb123.preheader + call fastcc void @f1(float* %tmp8, float* null, i32 0) + %tmp70 = load i32* null + %tmp85 = getelementptr float* %tmp8, i32 0 + call fastcc void @f2(float* null, float* null, float* %tmp85, i32 %tmp70) + ret void + +return: ; preds = %bb123.preheader, %0 + ret void } declare fastcc void @f1(float*, float*, i32) declare fastcc void @f2(float*, float*, float*, i32) - %struct.comment = type { i8**, i32*, i32, i8* } -@str215 = external global [2 x i8] - define void @t2(%struct.comment* %vc, i8* %tag, i8* %contents) { - %tmp1 = call i32 @strlen( i8* %tag ) - %tmp3 = call i32 @strlen( i8* %contents ) - %tmp4 = add i32 %tmp1, 2 - %tmp5 = add i32 %tmp4, %tmp3 - %tmp6 = alloca i8, i32 %tmp5 - %tmp9 = call i8* @strcpy( i8* %tmp6, i8* %tag ) - %tmp6.len = call i32 @strlen( i8* %tmp6 ) - %tmp6.indexed = getelementptr i8* %tmp6, i32 %tmp6.len - call void @llvm.memcpy.i32( i8* %tmp6.indexed, i8* getelementptr ([2 x i8]* @str215, i32 0, i32 0), i32 2, i32 1 ) - %tmp15 = call i8* @strcat( i8* %tmp6, i8* %contents ) - call fastcc void @comment_add( %struct.comment* %vc, i8* %tmp6 ) - ret void + %tmp1 = call i32 @strlen(i8* %tag) + %tmp3 = call i32 @strlen(i8* %contents) + %tmp4 = add i32 %tmp1, 2 + %tmp5 = add i32 %tmp4, %tmp3 + %tmp6 = alloca i8, i32 %tmp5 + %tmp9 = call i8* @strcpy(i8* %tmp6, i8* %tag) + %tmp6.len = call i32 @strlen(i8* %tmp6) + %tmp6.indexed = getelementptr i8* %tmp6, i32 %tmp6.len + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp6.indexed, i8* getelementptr inbounds ([2 x i8]* @str215, i32 0, i32 0), i32 2, i32 1, i1 false) + %tmp15 = call i8* @strcat(i8* %tmp6, i8* %contents) + call fastcc void @comment_add(%struct.comment* %vc, i8* %tmp6) + ret void } declare i32 @strlen(i8*) @@ -51,6 +51,6 @@ declare i8* @strcat(i8*, i8*) declare fastcc void @comment_add(%struct.comment*, i8*) -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) - declare i8* @strcpy(i8*, i8*) + +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/CodeGen/ARM/memcpy-inline.ll b/test/CodeGen/ARM/memcpy-inline.ll index 5bae037..30b9f59 100644 --- a/test/CodeGen/ARM/memcpy-inline.ll +++ b/test/CodeGen/ARM/memcpy-inline.ll @@ -6,13 +6,14 @@ ; CHECK: ldrb %struct.x = type { i8, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8 } + @src = external global %struct.x @dst = external global %struct.x define i32 @t() { entry: - call void @llvm.memcpy.i32( i8* getelementptr (%struct.x* @dst, i32 0, i32 0), i8* getelementptr (%struct.x* @src, i32 0, i32 0), i32 11, i32 8 ) - ret i32 0 + call void @llvm.memcpy.p0i8.p0i8.i32(i8* getelementptr inbounds (%struct.x* @dst, i32 0, i32 0), i8* getelementptr inbounds (%struct.x* @src, i32 0, i32 0), i32 11, i32 8, i1 false) + ret i32 0 } -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/CodeGen/ARM/memfunc.ll b/test/CodeGen/ARM/memfunc.ll index 757364b..aeda022 100644 --- a/test/CodeGen/ARM/memfunc.ll +++ b/test/CodeGen/ARM/memfunc.ll @@ -9,24 +9,21 @@ entry: ; CHECK: memmove ; EABI: __aeabi_memmove - call void @llvm.memmove.i32( i8* bitcast ([500 x i32]* @from to i8*), i8* bitcast ([500 x i32]* @to to i8*), i32 500, i32 0 ) + call void @llvm.memmove.p0i8.p0i8.i32(i8* bitcast ([500 x i32]* @from to i8*), i8* bitcast ([500 x i32]* @to to i8*), i32 500, i32 0, i1 false) ; CHECK: memcpy ; EABI: __aeabi_memcpy - call void @llvm.memcpy.i32( i8* bitcast ([500 x i32]* @from to i8*), i8* bitcast ([500 x i32]* @to to i8*), i32 500, i32 0 ) + call void @llvm.memcpy.p0i8.p0i8.i32(i8* bitcast ([500 x i32]* @from to i8*), i8* bitcast ([500 x i32]* @to to i8*), i32 500, i32 0, i1 false) ; EABI memset swaps arguments ; CHECK: mov r1, #0 ; CHECK: memset ; EABI: mov r2, #0 ; EABI: __aeabi_memset - call void @llvm.memset.i32( i8* bitcast ([500 x i32]* @from to i8*), i8 0, i32 500, i32 0 ) + call void @llvm.memset.p0i8.i32(i8* bitcast ([500 x i32]* @from to i8*), i8 0, i32 500, i32 0, i1 false) unreachable } -declare void @llvm.memmove.i32(i8*, i8*, i32, i32) - -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) - -declare void @llvm.memset.i32(i8*, i8, i32, i32) - +declare void @llvm.memmove.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind +declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1) nounwind diff --git a/test/CodeGen/Generic/2005-07-12-memcpy-i64-length.ll b/test/CodeGen/Generic/2005-07-12-memcpy-i64-length.ll deleted file mode 100644 index 733202c..0000000 --- a/test/CodeGen/Generic/2005-07-12-memcpy-i64-length.ll +++ /dev/null @@ -1,11 +0,0 @@ -; RUN: llc < %s -; Test that llvm.memcpy works with a i64 length operand on all targets. - -declare void @llvm.memcpy.i64(i8*, i8*, i64, i32) - -define void @l12_l94_bc_divide_endif_2E_3_2E_ce() { -newFuncRoot: - tail call void @llvm.memcpy.i64( i8* null, i8* null, i64 0, i32 1 ) - unreachable -} - diff --git a/test/CodeGen/Generic/2007-11-21-UndeadIllegalNode.ll b/test/CodeGen/Generic/2007-11-21-UndeadIllegalNode.ll deleted file mode 100644 index 3090857..0000000 --- a/test/CodeGen/Generic/2007-11-21-UndeadIllegalNode.ll +++ /dev/null @@ -1,157 +0,0 @@ -; RUN: llc < %s -o - - - %struct.RETURN = type { i32, i32 } - %struct.ada__finalization__controlled = type { %struct.system__finalization_root__root_controlled } - %struct.ada__streams__root_stream_type = type { %struct.ada__tags__dispatch_table* } - %struct.ada__strings__unbounded__string_access = type { i8*, %struct.RETURN* } - %struct.ada__strings__unbounded__unbounded_string = type { %struct.ada__finalization__controlled, %struct.ada__strings__unbounded__string_access, i32 } - %struct.ada__tags__dispatch_table = type { [1 x i32] } - %struct.exception = type { i8, i8, i32, i8*, i8*, i32, i8* } - %struct.system__finalization_root__root_controlled = type { %struct.ada__streams__root_stream_type, %struct.system__finalization_root__root_controlled*, %struct.system__finalization_root__root_controlled* } - %struct.system__standard_library__exception_data = type { i8, i8, i32, i32, %struct.system__standard_library__exception_data*, i32, void ()* } -@C.495.7639 = internal constant %struct.RETURN { i32 1, i32 16 } ; <%struct.RETURN*> [#uses=1] -@ada__strings__index_error = external global %struct.exception ; <%struct.exception*> [#uses=1] -@.str5 = internal constant [16 x i8] c"a-strunb.adb:690" ; <[16 x i8]*> [#uses=1] - -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) - -declare void @ada__strings__unbounded__realloc_for_chunk(%struct.ada__strings__unbounded__unbounded_string*, i32) - -declare void @__gnat_raise_exception(%struct.system__standard_library__exception_data*, i64) - -define void @ada__strings__unbounded__insert__2(%struct.ada__strings__unbounded__unbounded_string* %source, i32 %before, i64 %new_item.0.0) { -entry: - %tmp24636 = lshr i64 %new_item.0.0, 32 ; <i64> [#uses=1] - %tmp24637 = trunc i64 %tmp24636 to i32 ; <i32> [#uses=1] - %tmp24638 = inttoptr i32 %tmp24637 to %struct.RETURN* ; <%struct.RETURN*> [#uses=2] - %tmp25 = getelementptr %struct.RETURN* %tmp24638, i32 0, i32 0 ; <i32*> [#uses=1] - %tmp26 = load i32* %tmp25, align 4 ; <i32> [#uses=1] - %tmp29 = getelementptr %struct.RETURN* %tmp24638, i32 0, i32 1 ; <i32*> [#uses=1] - %tmp30 = load i32* %tmp29, align 4 ; <i32> [#uses=1] - %tmp63 = getelementptr %struct.ada__strings__unbounded__unbounded_string* %source, i32 0, i32 1, i32 1 ; <%struct.RETURN**> [#uses=5] - %tmp64 = load %struct.RETURN** %tmp63, align 4 ; <%struct.RETURN*> [#uses=1] - %tmp65 = getelementptr %struct.RETURN* %tmp64, i32 0, i32 0 ; <i32*> [#uses=1] - %tmp66 = load i32* %tmp65, align 4 ; <i32> [#uses=1] - %tmp67 = icmp sgt i32 %tmp66, %before ; <i1> [#uses=1] - br i1 %tmp67, label %bb77, label %bb - -bb: ; preds = %entry - %tmp71 = getelementptr %struct.ada__strings__unbounded__unbounded_string* %source, i32 0, i32 2 ; <i32*> [#uses=4] - %tmp72 = load i32* %tmp71, align 4 ; <i32> [#uses=1] - %tmp73 = add i32 %tmp72, 1 ; <i32> [#uses=1] - %tmp74 = icmp slt i32 %tmp73, %before ; <i1> [#uses=1] - br i1 %tmp74, label %bb77, label %bb84 - -bb77: ; preds = %bb, %entry - tail call void @__gnat_raise_exception( %struct.system__standard_library__exception_data* bitcast (%struct.exception* @ada__strings__index_error to %struct.system__standard_library__exception_data*), i64 or (i64 zext (i32 ptrtoint ([16 x i8]* @.str5 to i32) to i64), i64 shl (i64 zext (i32 ptrtoint (%struct.RETURN* @C.495.7639 to i32) to i64), i64 32)) ) - unreachable - -bb84: ; preds = %bb - %tmp93 = sub i32 %tmp30, %tmp26 ; <i32> [#uses=2] - %tmp9394 = sext i32 %tmp93 to i36 ; <i36> [#uses=1] - %tmp95 = shl i36 %tmp9394, 3 ; <i36> [#uses=1] - %tmp96 = add i36 %tmp95, 8 ; <i36> [#uses=2] - %tmp97 = icmp sgt i36 %tmp96, -1 ; <i1> [#uses=1] - %tmp100 = select i1 %tmp97, i36 %tmp96, i36 0 ; <i36> [#uses=2] - %tmp101 = icmp slt i36 %tmp100, 17179869177 ; <i1> [#uses=1] - %tmp100.cast = trunc i36 %tmp100 to i32 ; <i32> [#uses=1] - %min102 = select i1 %tmp101, i32 %tmp100.cast, i32 -8 ; <i32> [#uses=1] - tail call void @ada__strings__unbounded__realloc_for_chunk( %struct.ada__strings__unbounded__unbounded_string* %source, i32 %min102 ) - %tmp148 = load i32* %tmp71, align 4 ; <i32> [#uses=4] - %tmp152 = add i32 %tmp93, 1 ; <i32> [#uses=2] - %tmp153 = icmp sgt i32 %tmp152, -1 ; <i1> [#uses=1] - %max154 = select i1 %tmp153, i32 %tmp152, i32 0 ; <i32> [#uses=5] - %tmp155 = add i32 %tmp148, %max154 ; <i32> [#uses=5] - %tmp315 = getelementptr %struct.ada__strings__unbounded__unbounded_string* %source, i32 0, i32 1, i32 0 ; <i8**> [#uses=4] - %tmp328 = load %struct.RETURN** %tmp63, align 4 ; <%struct.RETURN*> [#uses=1] - %tmp329 = getelementptr %struct.RETURN* %tmp328, i32 0, i32 0 ; <i32*> [#uses=1] - %tmp330 = load i32* %tmp329, align 4 ; <i32> [#uses=4] - %tmp324 = add i32 %max154, %before ; <i32> [#uses=3] - %tmp331 = sub i32 %tmp324, %tmp330 ; <i32> [#uses=1] - %tmp349 = sub i32 %before, %tmp330 ; <i32> [#uses=1] - %tmp356 = icmp sgt i32 %tmp331, %tmp349 ; <i1> [#uses=1] - %tmp431 = icmp sgt i32 %tmp324, %tmp155 ; <i1> [#uses=2] - br i1 %tmp356, label %bb420, label %bb359 - -bb359: ; preds = %bb84 - br i1 %tmp431, label %bb481, label %bb382 - -bb382: ; preds = %bb382, %bb359 - %indvar = phi i32 [ 0, %bb359 ], [ %indvar.next, %bb382 ] ; <i32> [#uses=2] - %max379.pn = phi i32 [ %max154, %bb359 ], [ %L492b.0, %bb382 ] ; <i32> [#uses=1] - %before.pn = phi i32 [ %before, %bb359 ], [ 1, %bb382 ] ; <i32> [#uses=1] - %L492b.0 = add i32 %before.pn, %max379.pn ; <i32> [#uses=3] - %tmp386 = load %struct.RETURN** %tmp63, align 4 ; <%struct.RETURN*> [#uses=1] - %tmp387 = getelementptr %struct.RETURN* %tmp386, i32 0, i32 0 ; <i32*> [#uses=1] - %tmp388 = load i32* %tmp387, align 4 ; <i32> [#uses=2] - %tmp392 = load i8** %tmp315, align 4 ; <i8*> [#uses=2] - %R493b.0 = add i32 %indvar, %before ; <i32> [#uses=1] - %tmp405 = sub i32 %R493b.0, %tmp388 ; <i32> [#uses=1] - %tmp406 = getelementptr i8* %tmp392, i32 %tmp405 ; <i8*> [#uses=1] - %tmp407 = load i8* %tmp406, align 1 ; <i8> [#uses=1] - %tmp408 = sub i32 %L492b.0, %tmp388 ; <i32> [#uses=1] - %tmp409 = getelementptr i8* %tmp392, i32 %tmp408 ; <i8*> [#uses=1] - store i8 %tmp407, i8* %tmp409, align 1 - %tmp414 = icmp eq i32 %L492b.0, %tmp155 ; <i1> [#uses=1] - %indvar.next = add i32 %indvar, 1 ; <i32> [#uses=1] - br i1 %tmp414, label %bb481, label %bb382 - -bb420: ; preds = %bb84 - br i1 %tmp431, label %bb481, label %bb436.preheader - -bb436.preheader: ; preds = %bb420 - %tmp4468 = load i8** %tmp315, align 4 ; <i8*> [#uses=2] - %tmp4599 = sub i32 %tmp148, %tmp330 ; <i32> [#uses=1] - %tmp46010 = getelementptr i8* %tmp4468, i32 %tmp4599 ; <i8*> [#uses=1] - %tmp46111 = load i8* %tmp46010, align 1 ; <i8> [#uses=1] - %tmp46212 = sub i32 %tmp155, %tmp330 ; <i32> [#uses=1] - %tmp46313 = getelementptr i8* %tmp4468, i32 %tmp46212 ; <i8*> [#uses=1] - store i8 %tmp46111, i8* %tmp46313, align 1 - %exitcond14 = icmp eq i32 %tmp155, %tmp324 ; <i1> [#uses=1] - br i1 %exitcond14, label %bb481, label %bb.nph - -bb.nph: ; preds = %bb436.preheader - %tmp5 = sub i32 %tmp148, %before ; <i32> [#uses=1] - br label %bb478 - -bb478: ; preds = %bb478, %bb.nph - %indvar6422 = phi i32 [ 0, %bb.nph ], [ %indvar.next643, %bb478 ] ; <i32> [#uses=1] - %indvar.next643 = add i32 %indvar6422, 1 ; <i32> [#uses=4] - %L490b.0 = sub i32 %tmp155, %indvar.next643 ; <i32> [#uses=1] - %R491b.0 = sub i32 %tmp148, %indvar.next643 ; <i32> [#uses=1] - %tmp440 = load %struct.RETURN** %tmp63, align 4 ; <%struct.RETURN*> [#uses=1] - %tmp441 = getelementptr %struct.RETURN* %tmp440, i32 0, i32 0 ; <i32*> [#uses=1] - %tmp442 = load i32* %tmp441, align 4 ; <i32> [#uses=2] - %tmp446 = load i8** %tmp315, align 4 ; <i8*> [#uses=2] - %tmp459 = sub i32 %R491b.0, %tmp442 ; <i32> [#uses=1] - %tmp460 = getelementptr i8* %tmp446, i32 %tmp459 ; <i8*> [#uses=1] - %tmp461 = load i8* %tmp460, align 1 ; <i8> [#uses=1] - %tmp462 = sub i32 %L490b.0, %tmp442 ; <i32> [#uses=1] - %tmp463 = getelementptr i8* %tmp446, i32 %tmp462 ; <i8*> [#uses=1] - store i8 %tmp461, i8* %tmp463, align 1 - %exitcond = icmp eq i32 %indvar.next643, %tmp5 ; <i1> [#uses=1] - br i1 %exitcond, label %bb481, label %bb478 - -bb481: ; preds = %bb478, %bb436.preheader, %bb420, %bb382, %bb359 - %tmp577 = add i32 %before, -1 ; <i32> [#uses=3] - %tmp578 = add i32 %max154, %tmp577 ; <i32> [#uses=2] - %tmp581 = icmp sge i32 %tmp578, %tmp577 ; <i1> [#uses=1] - %max582 = select i1 %tmp581, i32 %tmp578, i32 %tmp577 ; <i32> [#uses=1] - %tmp584 = sub i32 %max582, %before ; <i32> [#uses=1] - %tmp585 = add i32 %tmp584, 1 ; <i32> [#uses=2] - %tmp586 = icmp sgt i32 %tmp585, -1 ; <i1> [#uses=1] - %max587 = select i1 %tmp586, i32 %tmp585, i32 0 ; <i32> [#uses=1] - %tmp591 = load %struct.RETURN** %tmp63, align 4 ; <%struct.RETURN*> [#uses=1] - %tmp592 = getelementptr %struct.RETURN* %tmp591, i32 0, i32 0 ; <i32*> [#uses=1] - %tmp593 = load i32* %tmp592, align 4 ; <i32> [#uses=1] - %tmp597 = load i8** %tmp315, align 4 ; <i8*> [#uses=1] - %tmp600621 = trunc i64 %new_item.0.0 to i32 ; <i32> [#uses=1] - %tmp600622 = inttoptr i32 %tmp600621 to i8* ; <i8*> [#uses=1] - %tmp601 = sub i32 %before, %tmp593 ; <i32> [#uses=1] - %tmp602 = getelementptr i8* %tmp597, i32 %tmp601 ; <i8*> [#uses=1] - tail call void @llvm.memcpy.i32( i8* %tmp602, i8* %tmp600622, i32 %max587, i32 1 ) - %tmp606 = load i32* %tmp71, align 4 ; <i32> [#uses=1] - %tmp613 = add i32 %tmp606, %max154 ; <i32> [#uses=1] - store i32 %tmp613, i32* %tmp71, align 4 - ret void -} diff --git a/test/CodeGen/Generic/invalid-memcpy.ll b/test/CodeGen/Generic/invalid-memcpy.ll index 8448565..2dfa28b 100644 --- a/test/CodeGen/Generic/invalid-memcpy.ll +++ b/test/CodeGen/Generic/invalid-memcpy.ll @@ -4,16 +4,14 @@ ; greater than the alignment guaranteed for Qux or C.0.1173), but it ; should compile, not crash the code generator. -@C.0.1173 = external constant [33 x i8] ; <[33 x i8]*> [#uses=1] +@C.0.1173 = external constant [33 x i8] define void @Bork() { entry: - %Qux = alloca [33 x i8] ; <[33 x i8]*> [#uses=1] - %Qux1 = bitcast [33 x i8]* %Qux to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i64( i8* %Qux1, i8* getelementptr ([33 x i8]* @C.0.1173, i32 0, i32 0), i64 33, i32 8 ) - ret void + %Qux = alloca [33 x i8] + %Qux1 = bitcast [33 x i8]* %Qux to i8* + call void @llvm.memcpy.p0i8.p0i8.i64(i8* %Qux1, i8* getelementptr inbounds ([33 x i8]* @C.0.1173, i32 0, i32 0), i64 33, i32 8, i1 false) + ret void } -declare void @llvm.memcpy.i64(i8*, i8*, i64, i32) - - +declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture, i64, i32, i1) nounwind diff --git a/test/CodeGen/PowerPC/2008-10-30-IllegalShift.ll b/test/CodeGen/PowerPC/2008-10-30-IllegalShift.ll deleted file mode 100644 index 83f3f6f..0000000 --- a/test/CodeGen/PowerPC/2008-10-30-IllegalShift.ll +++ /dev/null @@ -1,14 +0,0 @@ -; RUN: llc < %s -march=ppc32 -; PR2986 -@argc = external global i32 ; <i32*> [#uses=1] -@buffer = external global [32 x i8], align 4 ; <[32 x i8]*> [#uses=1] - -define void @test1() nounwind noinline { -entry: - %0 = load i32* @argc, align 4 ; <i32> [#uses=1] - %1 = trunc i32 %0 to i8 ; <i8> [#uses=1] - tail call void @llvm.memset.i32(i8* getelementptr ([32 x i8]* @buffer, i32 0, i32 0), i8 %1, i32 17, i32 4) - unreachable -} - -declare void @llvm.memset.i32(i8*, i8, i32, i32) nounwind diff --git a/test/CodeGen/PowerPC/2009-11-15-ReMatBug.ll b/test/CodeGen/PowerPC/2009-11-15-ReMatBug.ll deleted file mode 100644 index 54f4b2e..0000000 --- a/test/CodeGen/PowerPC/2009-11-15-ReMatBug.ll +++ /dev/null @@ -1,155 +0,0 @@ -; RUN: llc < %s -mtriple=powerpc-apple-darwin8 - -%struct.FILE = type { i8*, i32, i32, i16, i16, %struct.__sbuf, i32, i8*, i32 (i8*)*, i32 (i8*, i8*, i32)*, i64 (i8*, i64, i32)*, i32 (i8*, i8*, i32)*, %struct.__sbuf, %struct.__sFILEX*, i32, [3 x i8], [1 x i8], %struct.__sbuf, i32, i64 } -%struct.__gcov_var = type { %struct.FILE*, i32, i32, i32, i32, i32, i32, [1025 x i32] } -%struct.__sFILEX = type opaque -%struct.__sbuf = type { i8*, i32 } -%struct.gcov_ctr_info = type { i32, i64*, void (i64*, i32)* } -%struct.gcov_ctr_summary = type { i32, i32, i64, i64, i64 } -%struct.gcov_fn_info = type { i32, i32, [0 x i32] } -%struct.gcov_info = type { i32, %struct.gcov_info*, i32, i8*, i32, %struct.gcov_fn_info*, i32, [0 x %struct.gcov_ctr_info] } -%struct.gcov_summary = type { i32, [1 x %struct.gcov_ctr_summary] } - -@__gcov_var = external global %struct.__gcov_var ; <%struct.__gcov_var*> [#uses=1] -@__sF = external global [0 x %struct.FILE] ; <[0 x %struct.FILE]*> [#uses=1] -@.str = external constant [56 x i8], align 4 ; <[56 x i8]*> [#uses=1] -@gcov_list = external global %struct.gcov_info* ; <%struct.gcov_info**> [#uses=1] -@.str7 = external constant [35 x i8], align 4 ; <[35 x i8]*> [#uses=1] -@.str8 = external constant [9 x i8], align 4 ; <[9 x i8]*> [#uses=1] -@.str9 = external constant [10 x i8], align 4 ; <[10 x i8]*> [#uses=1] -@.str10 = external constant [36 x i8], align 4 ; <[36 x i8]*> [#uses=1] - -declare i32 @"\01_fprintf$LDBL128"(%struct.FILE*, i8*, ...) nounwind - -define void @gcov_exit() nounwind { -entry: - %gi_ptr.0357 = load %struct.gcov_info** @gcov_list, align 4 ; <%struct.gcov_info*> [#uses=1] - %0 = alloca i8, i32 undef, align 1 ; <i8*> [#uses=3] - br i1 undef, label %return, label %bb.nph341 - -bb.nph341: ; preds = %entry - %object27 = bitcast %struct.gcov_summary* undef to i8* ; <i8*> [#uses=1] - br label %bb25 - -bb25: ; preds = %read_fatal, %bb.nph341 - %gi_ptr.1329 = phi %struct.gcov_info* [ %gi_ptr.0357, %bb.nph341 ], [ undef, %read_fatal ] ; <%struct.gcov_info*> [#uses=1] - call void @llvm.memset.i32(i8* %object27, i8 0, i32 36, i32 8) - br i1 undef, label %bb49.1, label %bb48 - -bb48: ; preds = %bb25 - br label %bb49.1 - -bb51: ; preds = %bb48.4, %bb49.3 - switch i32 undef, label %bb58 [ - i32 0, label %rewrite - i32 1734567009, label %bb59 - ] - -bb58: ; preds = %bb51 - %1 = call i32 (%struct.FILE*, i8*, ...)* @"\01_fprintf$LDBL128"(%struct.FILE* getelementptr inbounds ([0 x %struct.FILE]* @__sF, i32 0, i32 2), i8* getelementptr inbounds ([35 x i8]* @.str7, i32 0, i32 0), i8* %0) nounwind ; <i32> [#uses=0] - br label %read_fatal - -bb59: ; preds = %bb51 - br i1 undef, label %bb60, label %bb3.i156 - -bb3.i156: ; preds = %bb59 - store i8 52, i8* undef, align 1 - store i8 42, i8* undef, align 1 - %2 = call i32 (%struct.FILE*, i8*, ...)* @"\01_fprintf$LDBL128"(%struct.FILE* getelementptr inbounds ([0 x %struct.FILE]* @__sF, i32 0, i32 2), i8* getelementptr inbounds ([56 x i8]* @.str, i32 0, i32 0), i8* %0, i8* undef, i8* undef) nounwind ; <i32> [#uses=0] - br label %read_fatal - -bb60: ; preds = %bb59 - br i1 undef, label %bb78.preheader, label %rewrite - -bb78.preheader: ; preds = %bb60 - br i1 undef, label %bb62, label %bb80 - -bb62: ; preds = %bb78.preheader - br i1 undef, label %bb64, label %read_mismatch - -bb64: ; preds = %bb62 - br i1 undef, label %bb65, label %read_mismatch - -bb65: ; preds = %bb64 - br i1 undef, label %bb75, label %read_mismatch - -read_mismatch: ; preds = %bb98, %bb119.preheader, %bb72, %bb71, %bb65, %bb64, %bb62 - %3 = icmp eq i32 undef, -1 ; <i1> [#uses=1] - %iftmp.11.0 = select i1 %3, i8* getelementptr inbounds ([10 x i8]* @.str9, i32 0, i32 0), i8* getelementptr inbounds ([9 x i8]* @.str8, i32 0, i32 0) ; <i8*> [#uses=1] - %4 = call i32 (%struct.FILE*, i8*, ...)* @"\01_fprintf$LDBL128"(%struct.FILE* getelementptr inbounds ([0 x %struct.FILE]* @__sF, i32 0, i32 2), i8* getelementptr inbounds ([36 x i8]* @.str10, i32 0, i32 0), i8* %0, i8* %iftmp.11.0) nounwind ; <i32> [#uses=0] - br label %read_fatal - -bb71: ; preds = %bb75 - %5 = load i32* undef, align 4 ; <i32> [#uses=1] - %6 = getelementptr inbounds %struct.gcov_info* %gi_ptr.1329, i32 0, i32 7, i32 undef, i32 2 ; <void (i64*, i32)**> [#uses=1] - %7 = load void (i64*, i32)** %6, align 4 ; <void (i64*, i32)*> [#uses=1] - %8 = call i32 @__gcov_read_unsigned() nounwind ; <i32> [#uses=1] - %9 = call i32 @__gcov_read_unsigned() nounwind ; <i32> [#uses=1] - %10 = icmp eq i32 %tmp386, %8 ; <i1> [#uses=1] - br i1 %10, label %bb72, label %read_mismatch - -bb72: ; preds = %bb71 - %11 = icmp eq i32 undef, %9 ; <i1> [#uses=1] - br i1 %11, label %bb73, label %read_mismatch - -bb73: ; preds = %bb72 - call void %7(i64* null, i32 %5) nounwind - unreachable - -bb74: ; preds = %bb75 - %12 = add i32 %13, 1 ; <i32> [#uses=1] - br label %bb75 - -bb75: ; preds = %bb74, %bb65 - %13 = phi i32 [ %12, %bb74 ], [ 0, %bb65 ] ; <i32> [#uses=2] - %tmp386 = add i32 0, 27328512 ; <i32> [#uses=1] - %14 = shl i32 1, %13 ; <i32> [#uses=1] - %15 = load i32* undef, align 4 ; <i32> [#uses=1] - %16 = and i32 %15, %14 ; <i32> [#uses=1] - %17 = icmp eq i32 %16, 0 ; <i1> [#uses=1] - br i1 %17, label %bb74, label %bb71 - -bb80: ; preds = %bb78.preheader - unreachable - -read_fatal: ; preds = %read_mismatch, %bb3.i156, %bb58 - br i1 undef, label %return, label %bb25 - -rewrite: ; preds = %bb60, %bb51 - store i32 -1, i32* getelementptr inbounds (%struct.__gcov_var* @__gcov_var, i32 0, i32 6), align 4 - br i1 undef, label %bb94, label %bb119.preheader - -bb94: ; preds = %rewrite - unreachable - -bb119.preheader: ; preds = %rewrite - br i1 undef, label %read_mismatch, label %bb98 - -bb98: ; preds = %bb119.preheader - br label %read_mismatch - -return: ; preds = %read_fatal, %entry - ret void - -bb49.1: ; preds = %bb48, %bb25 - br i1 undef, label %bb49.2, label %bb48.2 - -bb49.2: ; preds = %bb48.2, %bb49.1 - br i1 undef, label %bb49.3, label %bb48.3 - -bb48.2: ; preds = %bb49.1 - br label %bb49.2 - -bb49.3: ; preds = %bb48.3, %bb49.2 - br i1 undef, label %bb51, label %bb48.4 - -bb48.3: ; preds = %bb49.2 - br label %bb49.3 - -bb48.4: ; preds = %bb49.3 - br label %bb51 -} - -declare i32 @__gcov_read_unsigned() nounwind - -declare void @llvm.memset.i32(i8* nocapture, i8, i32, i32) nounwind diff --git a/test/CodeGen/PowerPC/atomic-1.ll b/test/CodeGen/PowerPC/atomic-1.ll index ec4e42d..a2cf170 100644 --- a/test/CodeGen/PowerPC/atomic-1.ll +++ b/test/CodeGen/PowerPC/atomic-1.ll @@ -1,21 +1,23 @@ ; RUN: llc < %s -march=ppc32 | grep lwarx | count 3 ; RUN: llc < %s -march=ppc32 | grep stwcx. | count 4 -define i32 @exchange_and_add(i32* %mem, i32 %val) nounwind { - %tmp = call i32 @llvm.atomic.load.add.i32( i32* %mem, i32 %val ) - ret i32 %tmp +define i32 @exchange_and_add(i32* %mem, i32 %val) nounwind { + %tmp = call i32 @llvm.atomic.load.add.i32.p0i32(i32* %mem, i32 %val) + ret i32 %tmp } -define i32 @exchange_and_cmp(i32* %mem) nounwind { - %tmp = call i32 @llvm.atomic.cmp.swap.i32( i32* %mem, i32 0, i32 1 ) - ret i32 %tmp +define i32 @exchange_and_cmp(i32* %mem) nounwind { + %tmp = call i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* %mem, i32 0, i32 1) + ret i32 %tmp } -define i32 @exchange(i32* %mem, i32 %val) nounwind { - %tmp = call i32 @llvm.atomic.swap.i32( i32* %mem, i32 1 ) - ret i32 %tmp +define i32 @exchange(i32* %mem, i32 %val) nounwind { + %tmp = call i32 @llvm.atomic.swap.i32.p0i32(i32* %mem, i32 1) + ret i32 %tmp } -declare i32 @llvm.atomic.load.add.i32(i32*, i32) nounwind -declare i32 @llvm.atomic.cmp.swap.i32(i32*, i32, i32) nounwind -declare i32 @llvm.atomic.swap.i32(i32*, i32) nounwind +declare i32 @llvm.atomic.load.add.i32.p0i32(i32* nocapture, i32) nounwind + +declare i32 @llvm.atomic.cmp.swap.i32.p0i32(i32* nocapture, i32, i32) nounwind + +declare i32 @llvm.atomic.swap.i32.p0i32(i32* nocapture, i32) nounwind diff --git a/test/CodeGen/PowerPC/atomic-2.ll b/test/CodeGen/PowerPC/atomic-2.ll index 6d9daef..0fa2a29 100644 --- a/test/CodeGen/PowerPC/atomic-2.ll +++ b/test/CodeGen/PowerPC/atomic-2.ll @@ -1,21 +1,23 @@ ; RUN: llc < %s -march=ppc64 | grep ldarx | count 3 ; RUN: llc < %s -march=ppc64 | grep stdcx. | count 4 -define i64 @exchange_and_add(i64* %mem, i64 %val) nounwind { - %tmp = call i64 @llvm.atomic.load.add.i64( i64* %mem, i64 %val ) - ret i64 %tmp +define i64 @exchange_and_add(i64* %mem, i64 %val) nounwind { + %tmp = call i64 @llvm.atomic.load.add.i64.p0i64(i64* %mem, i64 %val) + ret i64 %tmp } -define i64 @exchange_and_cmp(i64* %mem) nounwind { - %tmp = call i64 @llvm.atomic.cmp.swap.i64( i64* %mem, i64 0, i64 1 ) - ret i64 %tmp +define i64 @exchange_and_cmp(i64* %mem) nounwind { + %tmp = call i64 @llvm.atomic.cmp.swap.i64.p0i64(i64* %mem, i64 0, i64 1) + ret i64 %tmp } -define i64 @exchange(i64* %mem, i64 %val) nounwind { - %tmp = call i64 @llvm.atomic.swap.i64( i64* %mem, i64 1 ) - ret i64 %tmp +define i64 @exchange(i64* %mem, i64 %val) nounwind { + %tmp = call i64 @llvm.atomic.swap.i64.p0i64(i64* %mem, i64 1) + ret i64 %tmp } -declare i64 @llvm.atomic.load.add.i64(i64*, i64) nounwind -declare i64 @llvm.atomic.cmp.swap.i64(i64*, i64, i64) nounwind -declare i64 @llvm.atomic.swap.i64(i64*, i64) nounwind +declare i64 @llvm.atomic.load.add.i64.p0i64(i64* nocapture, i64) nounwind + +declare i64 @llvm.atomic.cmp.swap.i64.p0i64(i64* nocapture, i64, i64) nounwind + +declare i64 @llvm.atomic.swap.i64.p0i64(i64* nocapture, i64) nounwind diff --git a/test/CodeGen/PowerPC/invalid-memcpy.ll b/test/CodeGen/PowerPC/invalid-memcpy.ll deleted file mode 100644 index 3b1f306..0000000 --- a/test/CodeGen/PowerPC/invalid-memcpy.ll +++ /dev/null @@ -1,20 +0,0 @@ -; RUN: llc < %s -march=ppc32 -; RUN: llc < %s -march=ppc64 - -; This testcase is invalid (the alignment specified for memcpy is -; greater than the alignment guaranteed for Qux or C.0.1173, but it -; should compile, not crash the code generator. - -@C.0.1173 = external constant [33 x i8] ; <[33 x i8]*> [#uses=1] - -define void @Bork() { -entry: - %Qux = alloca [33 x i8] ; <[33 x i8]*> [#uses=1] - %Qux1 = bitcast [33 x i8]* %Qux to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i64( i8* %Qux1, i8* getelementptr ([33 x i8]* @C.0.1173, i32 0, i32 0), i64 33, i32 8 ) - ret void -} - -declare void @llvm.memcpy.i64(i8*, i8*, i64, i32) - - diff --git a/test/CodeGen/Thumb/dyn-stackalloc.ll b/test/CodeGen/Thumb/dyn-stackalloc.ll index cd35be6..f3f0834 100644 --- a/test/CodeGen/Thumb/dyn-stackalloc.ll +++ b/test/CodeGen/Thumb/dyn-stackalloc.ll @@ -60,7 +60,7 @@ define void @t2(%struct.comment* %vc, i8* %tag, i8* %contents) { %tmp9 = call i8* @strcpy( i8* %tmp6, i8* %tag ) %tmp6.len = call i32 @strlen( i8* %tmp6 ) %tmp6.indexed = getelementptr i8* %tmp6, i32 %tmp6.len - call void @llvm.memcpy.i32( i8* %tmp6.indexed, i8* getelementptr ([2 x i8]* @str215, i32 0, i32 0), i32 2, i32 1 ) + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp6.indexed, i8* getelementptr inbounds ([2 x i8]* @str215, i32 0, i32 0), i32 2, i32 1, i1 false) %tmp15 = call i8* @strcat( i8* %tmp6, i8* %contents ) call fastcc void @comment_add( %struct.comment* %vc, i8* %tmp6 ) ret void @@ -72,6 +72,6 @@ declare i8* @strcat(i8*, i8*) declare fastcc void @comment_add(%struct.comment*, i8*) -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind declare i8* @strcpy(i8*, i8*) diff --git a/test/CodeGen/Thumb2/2009-08-04-SubregLoweringBug.ll b/test/CodeGen/Thumb2/2009-08-04-SubregLoweringBug.ll index bfea003..779e100 100644 --- a/test/CodeGen/Thumb2/2009-08-04-SubregLoweringBug.ll +++ b/test/CodeGen/Thumb2/2009-08-04-SubregLoweringBug.ll @@ -1,4 +1,3 @@ -; RUN: llc < %s -mtriple=thumbv7-apple-darwin9 -mcpu=cortex-a8 ; RUN: llc < %s -mtriple=thumbv7-apple-darwin9 -mcpu=cortex-a8 | not grep fcpys ; rdar://7117307 @@ -13,7 +12,7 @@ entry: br i1 undef, label %bb, label %bb6.preheader bb6.preheader: ; preds = %entry - call void @llvm.memcpy.i32(i8* undef, i8* undef, i32 12, i32 4) + call void @llvm.memcpy.p0i8.p0i8.i32(i8* undef, i8* undef, i32 12, i32 4, i1 false) br i1 undef, label %bb15, label %bb13 bb: ; preds = %entry @@ -31,4 +30,4 @@ bb15: ; preds = %bb13, %bb6.preheader ret void } -declare void @llvm.memcpy.i32(i8* nocapture, i8* nocapture, i32, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/CodeGen/X86/2004-02-12-Memcpy.ll b/test/CodeGen/X86/2004-02-12-Memcpy.ll deleted file mode 100644 index f15a1b4..0000000 --- a/test/CodeGen/X86/2004-02-12-Memcpy.ll +++ /dev/null @@ -1,25 +0,0 @@ -; RUN: llc < %s -march=x86 -mtriple=i686-pc-linux-gnu | grep movs | count 1 - -@A = global [32 x i32] zeroinitializer -@B = global [32 x i32] zeroinitializer - -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) - -define void @main() nounwind { - ; dword copy - call void @llvm.memcpy.i32(i8* bitcast ([32 x i32]* @A to i8*), - i8* bitcast ([32 x i32]* @B to i8*), - i32 128, i32 4 ) - - ; word copy - call void @llvm.memcpy.i32( i8* bitcast ([32 x i32]* @A to i8*), - i8* bitcast ([32 x i32]* @B to i8*), - i32 128, i32 2 ) - - ; byte copy - call void @llvm.memcpy.i32( i8* bitcast ([32 x i32]* @A to i8*), - i8* bitcast ([32 x i32]* @B to i8*), - i32 128, i32 1 ) - - ret void -} diff --git a/test/CodeGen/X86/2006-11-28-Memcpy.ll b/test/CodeGen/X86/2006-11-28-Memcpy.ll deleted file mode 100644 index 8c1573f..0000000 --- a/test/CodeGen/X86/2006-11-28-Memcpy.ll +++ /dev/null @@ -1,34 +0,0 @@ -; PR1022, PR1023 -; RUN: llc < %s -march=x86 | grep -- -573785174 | count 2 -; RUN: llc < %s -march=x86 | grep -E {movl _?bytes2} | count 1 - -@fmt = constant [4 x i8] c"%x\0A\00" ; <[4 x i8]*> [#uses=2] -@bytes = constant [4 x i8] c"\AA\BB\CC\DD" ; <[4 x i8]*> [#uses=1] -@bytes2 = global [4 x i8] c"\AA\BB\CC\DD" ; <[4 x i8]*> [#uses=1] - -define i32 @test1() nounwind { - %y = alloca i32 ; <i32*> [#uses=2] - %c = bitcast i32* %y to i8* ; <i8*> [#uses=1] - %z = getelementptr [4 x i8]* @bytes, i32 0, i32 0 ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %c, i8* %z, i32 4, i32 1 ) - %r = load i32* %y ; <i32> [#uses=1] - %t = bitcast [4 x i8]* @fmt to i8* ; <i8*> [#uses=1] - %tmp = call i32 (i8*, ...)* @printf( i8* %t, i32 %r ) ; <i32> [#uses=0] - ret i32 0 -} - -define void @test2() nounwind { - %y = alloca i32 ; <i32*> [#uses=2] - %c = bitcast i32* %y to i8* ; <i8*> [#uses=1] - %z = getelementptr [4 x i8]* @bytes2, i32 0, i32 0 ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %c, i8* %z, i32 4, i32 1 ) - %r = load i32* %y ; <i32> [#uses=1] - %t = bitcast [4 x i8]* @fmt to i8* ; <i8*> [#uses=1] - %tmp = call i32 (i8*, ...)* @printf( i8* %t, i32 %r ) ; <i32> [#uses=0] - ret void -} - -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) - -declare i32 @printf(i8*, ...) - diff --git a/test/CodeGen/X86/2007-05-05-Personality.ll b/test/CodeGen/X86/2007-05-05-Personality.ll index 0f49d2e..d1fc70d 100644 --- a/test/CodeGen/X86/2007-05-05-Personality.ll +++ b/test/CodeGen/X86/2007-05-05-Personality.ll @@ -3,35 +3,35 @@ ; CHECK: .cfi_personality 0, __gnat_eh_personality ; CHECK: .cfi_lsda 0, .Lexception0 -@error = external global i8 ; <i8*> [#uses=2] +@error = external global i8 define void @_ada_x() { entry: - invoke void @raise( ) - to label %eh_then unwind label %unwind - -unwind: ; preds = %entry - %eh_ptr = tail call i8* @llvm.eh.exception( ) ; <i8*> [#uses=2] - %eh_select = tail call i32 (i8*, i8*, ...)* @llvm.eh.selector.i32( i8* %eh_ptr, i8* bitcast (i32 (...)* @__gnat_eh_personality to i8*), i8* @error ) ; <i32> [#uses=1] - %eh_typeid = tail call i32 @llvm.eh.typeid.for.i32( i8* @error ) ; <i32> [#uses=1] - %tmp2 = icmp eq i32 %eh_select, %eh_typeid ; <i1> [#uses=1] - br i1 %tmp2, label %eh_then, label %Unwind - -eh_then: ; preds = %unwind, %entry - ret void - -Unwind: ; preds = %unwind - tail call i32 (...)* @_Unwind_Resume( i8* %eh_ptr ) ; <i32>:0 [#uses=0] - unreachable + invoke void @raise() + to label %eh_then unwind label %unwind + +unwind: ; preds = %entry + %eh_ptr = tail call i8* @llvm.eh.exception() + %eh_select = tail call i32 (i8*, i8*, ...)* @llvm.eh.selector(i8* %eh_ptr, i8* bitcast (i32 (...)* @__gnat_eh_personality to i8*), i8* @error) + %eh_typeid = tail call i32 @llvm.eh.typeid.for(i8* @error) + %tmp2 = icmp eq i32 %eh_select, %eh_typeid + br i1 %tmp2, label %eh_then, label %Unwind + +eh_then: ; preds = %unwind, %entry + ret void + +Unwind: ; preds = %unwind + %0 = tail call i32 (...)* @_Unwind_Resume(i8* %eh_ptr) + unreachable } declare void @raise() -declare i8* @llvm.eh.exception() +declare i8* @llvm.eh.exception() nounwind readonly -declare i32 @llvm.eh.selector.i32(i8*, i8*, ...) +declare i32 @llvm.eh.selector(i8*, i8*, ...) nounwind -declare i32 @llvm.eh.typeid.for.i32(i8*) +declare i32 @llvm.eh.typeid.for(i8*) nounwind declare i32 @__gnat_eh_personality(...) diff --git a/test/CodeGen/X86/2007-06-05-LSR-Dominator.ll b/test/CodeGen/X86/2007-06-05-LSR-Dominator.ll deleted file mode 100644 index 36a97ef..0000000 --- a/test/CodeGen/X86/2007-06-05-LSR-Dominator.ll +++ /dev/null @@ -1,129 +0,0 @@ -; PR1495 -; RUN: llc < %s -march=x86 - -target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64" -target triple = "i686-pc-linux-gnu" - %struct.AVRational = type { i32, i32 } - %struct.FFTComplex = type { float, float } - %struct.FFTContext = type { i32, i32, i16*, %struct.FFTComplex*, %struct.FFTComplex*, void (%struct.FFTContext*, %struct.FFTComplex*)*, void (%struct.MDCTContext*, float*, float*, float*)* } - %struct.MDCTContext = type { i32, i32, float*, float*, %struct.FFTContext } - %struct.Minima = type { i32, i32, i32, i32 } - %struct.codebook_t = type { i32, i8*, i32*, i32, float, float, i32, i32, i32*, float*, float* } - %struct.floor_class_t = type { i32, i32, i32, i32* } - %struct.floor_t = type { i32, i32*, i32, %struct.floor_class_t*, i32, i32, i32, %struct.Minima* } - %struct.mapping_t = type { i32, i32*, i32*, i32*, i32, i32*, i32* } - %struct.residue_t = type { i32, i32, i32, i32, i32, i32, [8 x i8]*, [2 x float]* } - %struct.venc_context_t = type { i32, i32, [2 x i32], [2 x %struct.MDCTContext], [2 x float*], i32, float*, float*, float*, float*, float, i32, %struct.codebook_t*, i32, %struct.floor_t*, i32, %struct.residue_t*, i32, %struct.mapping_t*, i32, %struct.AVRational* } - -define fastcc i32 @put_main_header(%struct.venc_context_t* %venc, i8** %out) { -entry: - br i1 false, label %bb1820, label %bb288.bb148_crit_edge - -bb288.bb148_crit_edge: ; preds = %entry - ret i32 0 - -cond_next1712: ; preds = %bb1820.bb1680_crit_edge - ret i32 0 - -bb1817: ; preds = %bb1820.bb1680_crit_edge - br label %bb1820 - -bb1820: ; preds = %bb1817, %entry - %pb.1.50 = phi i32 [ %tmp1693, %bb1817 ], [ 8, %entry ] ; <i32> [#uses=3] - br i1 false, label %bb2093, label %bb1820.bb1680_crit_edge - -bb1820.bb1680_crit_edge: ; preds = %bb1820 - %tmp1693 = add i32 %pb.1.50, 8 ; <i32> [#uses=2] - %tmp1702 = icmp slt i32 %tmp1693, 0 ; <i1> [#uses=1] - br i1 %tmp1702, label %cond_next1712, label %bb1817 - -bb2093: ; preds = %bb1820 - %tmp2102 = add i32 %pb.1.50, 65 ; <i32> [#uses=0] - %tmp2236 = add i32 %pb.1.50, 72 ; <i32> [#uses=1] - %tmp2237 = sdiv i32 %tmp2236, 8 ; <i32> [#uses=2] - br i1 false, label %bb2543, label %bb2536.bb2396_crit_edge - -bb2536.bb2396_crit_edge: ; preds = %bb2093 - ret i32 0 - -bb2543: ; preds = %bb2093 - br i1 false, label %cond_next2576, label %bb2690 - -cond_next2576: ; preds = %bb2543 - ret i32 0 - -bb2682: ; preds = %bb2690 - ret i32 0 - -bb2690: ; preds = %bb2543 - br i1 false, label %bb2682, label %bb2698 - -bb2698: ; preds = %bb2690 - br i1 false, label %cond_next2726, label %bb2831 - -cond_next2726: ; preds = %bb2698 - ret i32 0 - -bb2831: ; preds = %bb2698 - br i1 false, label %cond_next2859, label %bb2964 - -cond_next2859: ; preds = %bb2831 - br i1 false, label %bb2943, label %cond_true2866 - -cond_true2866: ; preds = %cond_next2859 - br i1 false, label %cond_true2874, label %cond_false2897 - -cond_true2874: ; preds = %cond_true2866 - ret i32 0 - -cond_false2897: ; preds = %cond_true2866 - ret i32 0 - -bb2943: ; preds = %cond_next2859 - ret i32 0 - -bb2964: ; preds = %bb2831 - br i1 false, label %cond_next2997, label %bb4589 - -cond_next2997: ; preds = %bb2964 - ret i32 0 - -bb3103: ; preds = %bb4589 - ret i32 0 - -bb4589: ; preds = %bb2964 - br i1 false, label %bb3103, label %bb4597 - -bb4597: ; preds = %bb4589 - br i1 false, label %cond_next4630, label %bb4744 - -cond_next4630: ; preds = %bb4597 - br i1 false, label %bb4744, label %cond_true4724 - -cond_true4724: ; preds = %cond_next4630 - br i1 false, label %bb4736, label %bb7531 - -bb4736: ; preds = %cond_true4724 - ret i32 0 - -bb4744: ; preds = %cond_next4630, %bb4597 - ret i32 0 - -bb7531: ; preds = %cond_true4724 - %v_addr.023.0.i6 = add i32 %tmp2237, -255 ; <i32> [#uses=1] - br label %bb.i14 - -bb.i14: ; preds = %bb.i14, %bb7531 - %n.021.0.i8 = phi i32 [ 0, %bb7531 ], [ %indvar.next, %bb.i14 ] ; <i32> [#uses=2] - %tmp..i9 = mul i32 %n.021.0.i8, -255 ; <i32> [#uses=1] - %tmp5.i11 = add i32 %v_addr.023.0.i6, %tmp..i9 ; <i32> [#uses=1] - %tmp10.i12 = icmp ugt i32 %tmp5.i11, 254 ; <i1> [#uses=1] - %indvar.next = add i32 %n.021.0.i8, 1 ; <i32> [#uses=1] - br i1 %tmp10.i12, label %bb.i14, label %bb12.loopexit.i18 - -bb12.loopexit.i18: ; preds = %bb.i14 - call void @llvm.memcpy.i32( i8* null, i8* null, i32 %tmp2237, i32 1 ) - ret i32 0 -} - -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) diff --git a/test/CodeGen/X86/2008-06-04-MemCpyLoweringBug.ll b/test/CodeGen/X86/2008-06-04-MemCpyLoweringBug.ll deleted file mode 100644 index 236b7cd..0000000 --- a/test/CodeGen/X86/2008-06-04-MemCpyLoweringBug.ll +++ /dev/null @@ -1,19 +0,0 @@ -; RUN: llc < %s -mtriple=i386-apple-darwin -mattr=+sse2 -disable-fp-elim | grep subl | grep 24 - - %struct.argument_t = type { i8*, %struct.argument_t*, i32, %struct.ipc_type_t*, i32, void (...)*, void (...)*, void (...)*, void (...)*, void (...)*, i8*, i8*, i8*, i8*, i8*, i32, i32, i32, %struct.routine*, %struct.argument_t*, %struct.argument_t*, %struct.argument_t*, %struct.argument_t*, %struct.argument_t*, %struct.argument_t*, %struct.argument_t*, i32, i32, i32, i32, i32, i32 } - %struct.ipc_type_t = type { i8*, %struct.ipc_type_t*, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i8*, i8*, i32, i32, i32, i32, i32, i32, %struct.ipc_type_t*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8* } - %struct.routine = type opaque -@"\01LC" = external constant [11 x i8] ; <[11 x i8]*> [#uses=1] - -define i8* @InArgMsgField(%struct.argument_t* %arg, i8* %str) nounwind { -entry: - %who = alloca [20 x i8] ; <[20 x i8]*> [#uses=1] - %who1 = getelementptr [20 x i8]* %who, i32 0, i32 0 ; <i8*> [#uses=2] - call void @llvm.memset.i32( i8* %who1, i8 0, i32 20, i32 1 ) - call void @llvm.memcpy.i32( i8* %who1, i8* getelementptr ([11 x i8]* @"\01LC", i32 0, i32 0), i32 11, i32 1 ) - unreachable -} - -declare void @llvm.memset.i32(i8*, i8, i32, i32) nounwind - -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) nounwind diff --git a/test/CodeGen/X86/2008-10-27-StackRealignment.ll b/test/CodeGen/X86/2008-10-27-StackRealignment.ll index 3d0766c..a57f716 100644 --- a/test/CodeGen/X86/2008-10-27-StackRealignment.ll +++ b/test/CodeGen/X86/2008-10-27-StackRealignment.ll @@ -11,12 +11,12 @@ target triple = "i386-pc-linux-gnu" define void @foo(i32 %t) nounwind { %tmp1210 = alloca i8, i32 32, align 4 - call void @llvm.memset.i64(i8* %tmp1210, i8 0, i64 32, i32 4) - + call void @llvm.memset.p0i8.i64(i8* %tmp1210, i8 0, i64 32, i32 4, i1 false) %x = alloca i8, i32 %t call void @dummy(i8* %x) ret void } -declare void @dummy(i8* %x) -declare void @llvm.memset.i64(i8*, i8, i64, i32) nounwind +declare void @dummy(i8*) + +declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) nounwind diff --git a/test/CodeGen/X86/2009-01-25-NoSSE.ll b/test/CodeGen/X86/2009-01-25-NoSSE.ll index 0583ef1..8406c4a 100644 --- a/test/CodeGen/X86/2009-01-25-NoSSE.ll +++ b/test/CodeGen/X86/2009-01-25-NoSSE.ll @@ -3,18 +3,18 @@ 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-s0:64:64-f80:128:128" target triple = "x86_64-unknown-linux-gnu" - %struct.ktermios = type { i32, i32, i32, i32, i8, [19 x i8], i32, i32 } + +%struct.ktermios = type { i32, i32, i32, i32, i8, [19 x i8], i32, i32 } define void @foo() nounwind { entry: - %termios = alloca %struct.ktermios, align 8 - %termios1 = bitcast %struct.ktermios* %termios to i8* - call void @llvm.memset.i64(i8* %termios1, i8 0, i64 44, i32 8) - call void @bar(%struct.ktermios* %termios) nounwind - ret void + %termios = alloca %struct.ktermios, align 8 + %termios1 = bitcast %struct.ktermios* %termios to i8* + call void @llvm.memset.p0i8.i64(i8* %termios1, i8 0, i64 44, i32 8, i1 false) + call void @bar(%struct.ktermios* %termios) nounwind + ret void } -declare void @llvm.memset.i64(i8* nocapture, i8, i64, i32) nounwind - declare void @bar(%struct.ktermios*) +declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) nounwind diff --git a/test/CodeGen/X86/2009-04-27-LiveIntervalsBug.ll b/test/CodeGen/X86/2009-04-27-LiveIntervalsBug.ll deleted file mode 100644 index 0a2fcdb..0000000 --- a/test/CodeGen/X86/2009-04-27-LiveIntervalsBug.ll +++ /dev/null @@ -1,165 +0,0 @@ -; RUN: llc < %s -mtriple=i386-unknown-linux-gnu | grep cmpxchgl | not grep eax -; PR4076 - - type { i8, i8, i8 } ; type %0 - type { i32, i8** } ; type %1 - type { %3* } ; type %2 - type { %4 } ; type %3 - type { %5 } ; type %4 - type { %6, i32, %7 } ; type %5 - type { i8* } ; type %6 - type { i32, [12 x i8] } ; type %7 - type { %9 } ; type %8 - type { %10, %11*, i8 } ; type %9 - type { %11* } ; type %10 - type { i32, %6, i8*, %12, %13*, i8, i32, %28, %29, i32, %30, i32, i32, i32, i8*, i8*, i8, i8 } ; type %11 - type { %13* } ; type %12 - type { %14, i32, %13*, %21 } ; type %13 - type { %15, %16 } ; type %14 - type { i32 (...)** } ; type %15 - type { %17, i8* (i32)*, void (i8*)*, i8 } ; type %16 - type { i32 (...)**, i8*, i8*, i8*, i8*, i8*, i8*, %18 } ; type %17 - type { %19* } ; type %18 - type { i32, %20**, i32, %20**, i8** } ; type %19 - type { i32 (...)**, i32 } ; type %20 - type { %22, %25*, i8, i8, %17*, %26*, %27*, %27* } ; type %21 - type { i32 (...)**, i32, i32, i32, i32, i32, %23*, %24, [8 x %24], i32, %24*, %18 } ; type %22 - type { %23*, void (i32, %22*, i32)*, i32, i32 } ; type %23 - type { i8*, i32 } ; type %24 - type { i32 (...)**, %21 } ; type %25 - type { %20, i32*, i8, i32*, i32*, i16*, i8, [256 x i8], [256 x i8], i8 } ; type %26 - type { %20 } ; type %27 - type { void (%9*)*, i32 } ; type %28 - type { %15* } ; type %29 - type { i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i8* } ; type %30 -@AtomicOps_Internalx86CPUFeatures = external global %0 ; <%0*> [#uses=1] -internal constant [19 x i8] c"xxxxxxxxxxxxxxxxxx\00" ; <[19 x i8]*>:0 [#uses=1] -internal constant [47 x i8] c"xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx\00" ; <[47 x i8]*>:1 [#uses=1] - -define i8** @func6(i8 zeroext, i32, i32, %1*) nounwind { -; <label>:4 - %5 = alloca i32, align 4 ; <i32*> [#uses=2] - %6 = alloca i32, align 4 ; <i32*> [#uses=2] - %7 = alloca %2, align 8 ; <%2*> [#uses=3] - %8 = alloca %8, align 8 ; <%8*> [#uses=2] - br label %17 - -; <label>:9 ; preds = %17 - %10 = getelementptr %1* %3, i32 %19, i32 0 ; <i32*> [#uses=1] - %11 = load i32* %10, align 4 ; <i32> [#uses=1] - %12 = icmp eq i32 %11, %2 ; <i1> [#uses=1] - br i1 %12, label %13, label %16 - -; <label>:13 ; preds = %9 - %14 = getelementptr %1* %3, i32 %19, i32 1 ; <i8***> [#uses=1] - %15 = load i8*** %14, align 4 ; <i8**> [#uses=1] - ret i8** %15 - -; <label>:16 ; preds = %9 - %indvar.next13 = add i32 %18, 1 ; <i32> [#uses=1] - br label %17 - -; <label>:17 ; preds = %16, %4 - %18 = phi i32 [ 0, %4 ], [ %indvar.next13, %16 ] ; <i32> [#uses=2] - %19 = add i32 %18, %1 ; <i32> [#uses=3] - %20 = icmp sgt i32 %19, 3 ; <i1> [#uses=1] - br i1 %20, label %21, label %9 - -; <label>:21 ; preds = %17 - call void @func5() - %22 = getelementptr %1* %3, i32 0, i32 0 ; <i32*> [#uses=1] - %23 = load i32* %22, align 4 ; <i32> [#uses=1] - %24 = icmp eq i32 %23, 0 ; <i1> [#uses=1] - br i1 %24, label %._crit_edge, label %._crit_edge1 - -._crit_edge1: ; preds = %._crit_edge1, %21 - %25 = phi i32 [ 0, %21 ], [ %26, %._crit_edge1 ] ; <i32> [#uses=1] - %26 = add i32 %25, 1 ; <i32> [#uses=4] - %27 = getelementptr %1* %3, i32 %26, i32 0 ; <i32*> [#uses=1] - %28 = load i32* %27, align 4 ; <i32> [#uses=1] - %29 = icmp ne i32 %28, 0 ; <i1> [#uses=1] - %30 = icmp ne i32 %26, 4 ; <i1> [#uses=1] - %31 = and i1 %29, %30 ; <i1> [#uses=1] - br i1 %31, label %._crit_edge1, label %._crit_edge - -._crit_edge: ; preds = %._crit_edge1, %21 - %32 = phi i32 [ 0, %21 ], [ %26, %._crit_edge1 ] ; <i32> [#uses=3] - %33 = call i8* @pthread_getspecific(i32 0) nounwind ; <i8*> [#uses=2] - %34 = icmp ne i8* %33, null ; <i1> [#uses=1] - %35 = icmp eq i8 %0, 0 ; <i1> [#uses=1] - %36 = or i1 %34, %35 ; <i1> [#uses=1] - br i1 %36, label %._crit_edge4, label %37 - -; <label>:37 ; preds = %._crit_edge - %38 = call i8* @func2(i32 2048) ; <i8*> [#uses=4] - call void @llvm.memset.i32(i8* %38, i8 0, i32 2048, i32 4) - %39 = call i32 @pthread_setspecific(i32 0, i8* %38) nounwind ; <i32> [#uses=2] - store i32 %39, i32* %5 - store i32 0, i32* %6 - %40 = icmp eq i32 %39, 0 ; <i1> [#uses=1] - br i1 %40, label %41, label %43 - -; <label>:41 ; preds = %37 - %42 = getelementptr %2* %7, i32 0, i32 0 ; <%3**> [#uses=1] - store %3* null, %3** %42, align 8 - br label %._crit_edge4 - -; <label>:43 ; preds = %37 - %44 = call %3* @func1(i32* %5, i32* %6, i8* getelementptr ([47 x i8]* @1, i32 0, i32 0)) ; <%3*> [#uses=2] - %45 = getelementptr %2* %7, i32 0, i32 0 ; <%3**> [#uses=1] - store %3* %44, %3** %45, align 8 - %46 = icmp eq %3* %44, null ; <i1> [#uses=1] - br i1 %46, label %._crit_edge4, label %47 - -; <label>:47 ; preds = %43 - call void @func4(%8* %8, i8* getelementptr ([19 x i8]* @0, i32 0, i32 0), i32 165, %2* %7) - call void @func3(%8* %8) noreturn - unreachable - -._crit_edge4: ; preds = %43, %41, %._crit_edge - %48 = phi i8* [ %38, %41 ], [ %33, %._crit_edge ], [ %38, %43 ] ; <i8*> [#uses=2] - %49 = bitcast i8* %48 to i8** ; <i8**> [#uses=3] - %50 = icmp ne i8* %48, null ; <i1> [#uses=1] - %51 = icmp slt i32 %32, 4 ; <i1> [#uses=1] - %52 = and i1 %50, %51 ; <i1> [#uses=1] - br i1 %52, label %53, label %._crit_edge6 - -; <label>:53 ; preds = %._crit_edge4 - %54 = getelementptr %1* %3, i32 %32, i32 0 ; <i32*> [#uses=1] - %55 = call i32 asm sideeffect "lock; cmpxchgl $1,$2", "={ax},q,*m,0,~{dirflag},~{fpsr},~{flags},~{memory}"(i32 %2, i32* %54, i32 0) nounwind ; <i32> [#uses=1] - %56 = load i8* getelementptr (%0* @AtomicOps_Internalx86CPUFeatures, i32 0, i32 0), align 8 ; <i8> [#uses=1] - %57 = icmp eq i8 %56, 0 ; <i1> [#uses=1] - br i1 %57, label %._crit_edge7, label %58 - -; <label>:58 ; preds = %53 - call void asm sideeffect "lfence", "~{dirflag},~{fpsr},~{flags},~{memory}"() nounwind - br label %._crit_edge7 - -._crit_edge7: ; preds = %58, %53 - %59 = icmp eq i32 %55, 0 ; <i1> [#uses=1] - br i1 %59, label %60, label %._crit_edge6 - -._crit_edge6: ; preds = %._crit_edge7, %._crit_edge4 - ret i8** %49 - -; <label>:60 ; preds = %._crit_edge7 - %61 = getelementptr %1* %3, i32 %32, i32 1 ; <i8***> [#uses=1] - store i8** %49, i8*** %61, align 4 - ret i8** %49 -} - -declare %3* @func1(i32* nocapture, i32* nocapture, i8*) - -declare void @func5() - -declare void @func4(%8*, i8*, i32, %2*) - -declare void @func3(%8*) noreturn - -declare i8* @pthread_getspecific(i32) nounwind - -declare i8* @func2(i32) - -declare void @llvm.memset.i32(i8* nocapture, i8, i32, i32) nounwind - -declare i32 @pthread_setspecific(i32, i8*) nounwind diff --git a/test/CodeGen/X86/2009-11-16-UnfoldMemOpBug.ll b/test/CodeGen/X86/2009-11-16-UnfoldMemOpBug.ll index d33f93e..94075e7 100644 --- a/test/CodeGen/X86/2009-11-16-UnfoldMemOpBug.ll +++ b/test/CodeGen/X86/2009-11-16-UnfoldMemOpBug.ll @@ -16,7 +16,7 @@ bb1: ; CHECK: LBB0_1: ; CHECK: movaps %xmm0, (%rsp) %tmp2 = phi i32 [ %tmp3, %bb1 ], [ 0, %entry ] - call void @llvm.memcpy.i64(i8* %tmp1, i8* getelementptr inbounds ([28 x i8]* @str, i64 0, i64 0), i64 28, i32 1) + call void @llvm.memcpy.p0i8.p0i8.i64(i8* %tmp1, i8* getelementptr inbounds ([28 x i8]* @str, i64 0, i64 0), i64 28, i32 1, i1 false) %tmp3 = add i32 %tmp2, 1 %tmp4 = icmp eq i32 %tmp3, %count br i1 %tmp4, label %bb2, label %bb1 @@ -25,4 +25,4 @@ bb2: ret void } -declare void @llvm.memcpy.i64(i8* nocapture, i8* nocapture, i64, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture, i64, i32, i1) nounwind diff --git a/test/CodeGen/X86/darwin-bzero.ll b/test/CodeGen/X86/darwin-bzero.ll index a9573cf..3099526 100644 --- a/test/CodeGen/X86/darwin-bzero.ll +++ b/test/CodeGen/X86/darwin-bzero.ll @@ -1,8 +1,8 @@ ; RUN: llc < %s -mtriple=i386-apple-darwin10 | grep __bzero -declare void @llvm.memset.i32(i8*, i8, i32, i32) +declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1) nounwind define void @foo(i8* %p, i32 %len) { - call void @llvm.memset.i32(i8* %p, i8 0, i32 %len, i32 1) + call void @llvm.memset.p0i8.i32(i8* %p, i8 0, i32 %len, i32 1, i1 false) ret void } diff --git a/test/CodeGen/X86/memcpy-2.ll b/test/CodeGen/X86/memcpy-2.ll index 17cd8e8..eae2e70 100644 --- a/test/CodeGen/X86/memcpy-2.ll +++ b/test/CodeGen/X86/memcpy-2.ll @@ -38,7 +38,7 @@ entry: ; X86-64: movq $0 %tmp1 = alloca [25 x i8] %tmp2 = bitcast [25 x i8]* %tmp1 to i8* - call void @llvm.memcpy.i32( i8* %tmp2, i8* getelementptr ([25 x i8]* @.str, i32 0, i32 0), i32 25, i32 1 ) nounwind + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp2, i8* getelementptr inbounds ([25 x i8]* @.str, i32 0, i32 0), i32 25, i32 1, i1 false) unreachable } @@ -72,7 +72,7 @@ entry: ; X86-64: movaps %xmm0, (%rdi) %tmp2 = bitcast %struct.s0* %a to i8* ; <i8*> [#uses=1] %tmp3 = bitcast %struct.s0* %b to i8* ; <i8*> [#uses=1] - tail call void @llvm.memcpy.i32(i8* %tmp2, i8* %tmp3, i32 16, i32 16) + tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp2, i8* %tmp3, i32 16, i32 16, i1 false) ret void } @@ -115,7 +115,7 @@ entry: ; X86-64: movq %rax, (%rdi) %tmp2 = bitcast %struct.s0* %a to i8* ; <i8*> [#uses=1] %tmp3 = bitcast %struct.s0* %b to i8* ; <i8*> [#uses=1] - tail call void @llvm.memcpy.i32(i8* %tmp2, i8* %tmp3, i32 16, i32 8) + tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp2, i8* %tmp3, i32 16, i32 8, i1 false) ret void } @@ -160,8 +160,8 @@ entry: ; X86-64: movl $2021161080 %tmp1 = alloca [30 x i8] %tmp2 = bitcast [30 x i8]* %tmp1 to i8* - call void @llvm.memcpy.i32(i8* %tmp2, i8* getelementptr inbounds ([30 x i8]* @.str2, i32 0, i32 0), i32 30, i32 1) + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp2, i8* getelementptr inbounds ([30 x i8]* @.str2, i32 0, i32 0), i32 30, i32 1, i1 false) unreachable } -declare void @llvm.memcpy.i32(i8* nocapture, i8* nocapture, i32, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/CodeGen/X86/memmove-4.ll b/test/CodeGen/X86/memmove-4.ll deleted file mode 100644 index 027db1f..0000000 --- a/test/CodeGen/X86/memmove-4.ll +++ /dev/null @@ -1,12 +0,0 @@ -; RUN: llc < %s | not grep call - -target triple = "i686-pc-linux-gnu" - -define void @a(i8* %a, i8* %b) nounwind { - %tmp2 = bitcast i8* %a to i8* - %tmp3 = bitcast i8* %b to i8* - tail call void @llvm.memmove.i32( i8* %tmp2, i8* %tmp3, i32 12, i32 4 ) - ret void -} - -declare void @llvm.memmove.i32(i8*, i8*, i32, i32) diff --git a/test/CodeGen/X86/memset-2.ll b/test/CodeGen/X86/memset-2.ll index 993583b..b2bd72b 100644 --- a/test/CodeGen/X86/memset-2.ll +++ b/test/CodeGen/X86/memset-2.ll @@ -6,7 +6,7 @@ define fastcc void @t1() nounwind { entry: ; CHECK: t1: ; CHECK: calll _memset - call void @llvm.memset.i32( i8* null, i8 0, i32 188, i32 1 ) nounwind + call void @llvm.memset.p0i8.i32(i8* null, i8 0, i32 188, i32 1, i1 false) unreachable } @@ -14,7 +14,7 @@ define fastcc void @t2(i8 signext %c) nounwind { entry: ; CHECK: t2: ; CHECK: calll _memset - call void @llvm.memset.i32( i8* undef, i8 %c, i32 76, i32 1 ) nounwind + call void @llvm.memset.p0i8.i32(i8* undef, i8 %c, i32 76, i32 1, i1 false) unreachable } diff --git a/test/CodeGen/X86/memset-3.ll b/test/CodeGen/X86/memset-3.ll index 9b20ad5..29febfa 100644 --- a/test/CodeGen/X86/memset-3.ll +++ b/test/CodeGen/X86/memset-3.ll @@ -5,8 +5,8 @@ define void @t() nounwind ssp { entry: %buf = alloca [512 x i8], align 1 %ptr = getelementptr inbounds [512 x i8]* %buf, i32 0, i32 0 - call void @llvm.memset.i32(i8* %ptr, i8 undef, i32 512, i32 1) + call void @llvm.memset.p0i8.i32(i8* %ptr, i8 undef, i32 512, i32 1, i1 false) unreachable } -declare void @llvm.memset.i32(i8* nocapture, i8, i32, i32) nounwind +declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1) nounwind diff --git a/test/CodeGen/X86/memset.ll b/test/CodeGen/X86/memset.ll index cf7464d..72b3e0f 100644 --- a/test/CodeGen/X86/memset.ll +++ b/test/CodeGen/X86/memset.ll @@ -8,11 +8,11 @@ entry: %up_mvd = alloca [8 x %struct.x] ; <[8 x %struct.x]*> [#uses=2] %up_mvd116 = getelementptr [8 x %struct.x]* %up_mvd, i32 0, i32 0 ; <%struct.x*> [#uses=1] %tmp110117 = bitcast [8 x %struct.x]* %up_mvd to i8* ; <i8*> [#uses=1] - call void @llvm.memset.i64( i8* %tmp110117, i8 0, i64 32, i32 8 ) + call void @llvm.memset.p0i8.i64(i8* %tmp110117, i8 0, i64 32, i32 8, i1 false) call void @foo( %struct.x* %up_mvd116 ) nounwind ret void } declare void @foo(%struct.x*) -declare void @llvm.memset.i64(i8*, i8, i64, i32) nounwind +declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) nounwind diff --git a/test/CodeGen/X86/memset64-on-x86-32.ll b/test/CodeGen/X86/memset64-on-x86-32.ll index 3f069b4..e20fce1 100644 --- a/test/CodeGen/X86/memset64-on-x86-32.ll +++ b/test/CodeGen/X86/memset64-on-x86-32.ll @@ -4,9 +4,8 @@ define void @bork() nounwind { entry: - call void @llvm.memset.i64( i8* null, i8 0, i64 80, i32 4 ) - ret void + call void @llvm.memset.p0i8.i64(i8* null, i8 0, i64 80, i32 4, i1 false) + ret void } -declare void @llvm.memset.i64(i8*, i8, i64, i32) nounwind - +declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) nounwind diff --git a/test/CodeGen/X86/optimize-max-0.ll b/test/CodeGen/X86/optimize-max-0.ll index 162c7a5..981a16a 100644 --- a/test/CodeGen/X86/optimize-max-0.ll +++ b/test/CodeGen/X86/optimize-max-0.ll @@ -8,454 +8,454 @@ target triple = "i386-apple-darwin9" define void @foo(i8* %r, i32 %s, i32 %w, i32 %x, i8* %j, i32 %d) nounwind { entry: - %0 = mul i32 %x, %w ; <i32> [#uses=2] - %1 = mul i32 %x, %w ; <i32> [#uses=1] - %2 = sdiv i32 %1, 4 ; <i32> [#uses=1] - %.sum2 = add i32 %2, %0 ; <i32> [#uses=2] - %cond = icmp eq i32 %d, 1 ; <i1> [#uses=1] - br i1 %cond, label %bb29, label %bb10.preheader - -bb10.preheader: ; preds = %entry - %3 = icmp sgt i32 %x, 0 ; <i1> [#uses=1] - br i1 %3, label %bb.nph9, label %bb18.loopexit - -bb.nph7: ; preds = %bb7.preheader - %4 = mul i32 %y.08, %w ; <i32> [#uses=1] - %5 = mul i32 %y.08, %s ; <i32> [#uses=1] - %6 = add i32 %5, 1 ; <i32> [#uses=1] - %tmp8 = icmp sgt i32 1, %w ; <i1> [#uses=1] - %smax9 = select i1 %tmp8, i32 1, i32 %w ; <i32> [#uses=1] - br label %bb6 - -bb6: ; preds = %bb7, %bb.nph7 - %x.06 = phi i32 [ 0, %bb.nph7 ], [ %indvar.next7, %bb7 ] ; <i32> [#uses=3] - %7 = add i32 %x.06, %4 ; <i32> [#uses=1] - %8 = shl i32 %x.06, 1 ; <i32> [#uses=1] - %9 = add i32 %6, %8 ; <i32> [#uses=1] - %10 = getelementptr i8* %r, i32 %9 ; <i8*> [#uses=1] - %11 = load i8* %10, align 1 ; <i8> [#uses=1] - %12 = getelementptr i8* %j, i32 %7 ; <i8*> [#uses=1] - store i8 %11, i8* %12, align 1 - br label %bb7 - -bb7: ; preds = %bb6 - %indvar.next7 = add i32 %x.06, 1 ; <i32> [#uses=2] - %exitcond10 = icmp ne i32 %indvar.next7, %smax9 ; <i1> [#uses=1] - br i1 %exitcond10, label %bb6, label %bb7.bb9_crit_edge - -bb7.bb9_crit_edge: ; preds = %bb7 - br label %bb9 - -bb9: ; preds = %bb7.preheader, %bb7.bb9_crit_edge - br label %bb10 - -bb10: ; preds = %bb9 - %indvar.next11 = add i32 %y.08, 1 ; <i32> [#uses=2] - %exitcond12 = icmp ne i32 %indvar.next11, %x ; <i1> [#uses=1] - br i1 %exitcond12, label %bb7.preheader, label %bb10.bb18.loopexit_crit_edge - -bb10.bb18.loopexit_crit_edge: ; preds = %bb10 - br label %bb10.bb18.loopexit_crit_edge.split - -bb10.bb18.loopexit_crit_edge.split: ; preds = %bb.nph9, %bb10.bb18.loopexit_crit_edge - br label %bb18.loopexit - -bb.nph9: ; preds = %bb10.preheader - %13 = icmp sgt i32 %w, 0 ; <i1> [#uses=1] - br i1 %13, label %bb.nph9.split, label %bb10.bb18.loopexit_crit_edge.split - -bb.nph9.split: ; preds = %bb.nph9 - br label %bb7.preheader - -bb7.preheader: ; preds = %bb.nph9.split, %bb10 - %y.08 = phi i32 [ 0, %bb.nph9.split ], [ %indvar.next11, %bb10 ] ; <i32> [#uses=3] - br i1 true, label %bb.nph7, label %bb9 - -bb.nph5: ; preds = %bb18.loopexit - %14 = sdiv i32 %w, 2 ; <i32> [#uses=1] - %15 = icmp slt i32 %w, 2 ; <i1> [#uses=1] - %16 = sdiv i32 %x, 2 ; <i32> [#uses=2] - br i1 %15, label %bb18.bb20_crit_edge.split, label %bb.nph5.split - -bb.nph5.split: ; preds = %bb.nph5 - %tmp2 = icmp sgt i32 1, %16 ; <i1> [#uses=1] - %smax3 = select i1 %tmp2, i32 1, i32 %16 ; <i32> [#uses=1] - br label %bb13 - -bb13: ; preds = %bb18, %bb.nph5.split - %y.14 = phi i32 [ 0, %bb.nph5.split ], [ %indvar.next1, %bb18 ] ; <i32> [#uses=4] - %17 = mul i32 %14, %y.14 ; <i32> [#uses=2] - %18 = shl i32 %y.14, 1 ; <i32> [#uses=1] - %19 = srem i32 %y.14, 2 ; <i32> [#uses=1] - %20 = add i32 %19, %18 ; <i32> [#uses=1] - %21 = mul i32 %20, %s ; <i32> [#uses=2] - br i1 true, label %bb.nph3, label %bb17 - -bb.nph3: ; preds = %bb13 - %22 = add i32 %17, %0 ; <i32> [#uses=1] - %23 = add i32 %17, %.sum2 ; <i32> [#uses=1] - %24 = sdiv i32 %w, 2 ; <i32> [#uses=2] - %tmp = icmp sgt i32 1, %24 ; <i1> [#uses=1] - %smax = select i1 %tmp, i32 1, i32 %24 ; <i32> [#uses=1] - br label %bb14 - -bb14: ; preds = %bb15, %bb.nph3 - %x.12 = phi i32 [ 0, %bb.nph3 ], [ %indvar.next, %bb15 ] ; <i32> [#uses=5] - %25 = shl i32 %x.12, 2 ; <i32> [#uses=1] - %26 = add i32 %25, %21 ; <i32> [#uses=1] - %27 = getelementptr i8* %r, i32 %26 ; <i8*> [#uses=1] - %28 = load i8* %27, align 1 ; <i8> [#uses=1] - %.sum = add i32 %22, %x.12 ; <i32> [#uses=1] - %29 = getelementptr i8* %j, i32 %.sum ; <i8*> [#uses=1] - store i8 %28, i8* %29, align 1 - %30 = shl i32 %x.12, 2 ; <i32> [#uses=1] - %31 = or i32 %30, 2 ; <i32> [#uses=1] - %32 = add i32 %31, %21 ; <i32> [#uses=1] - %33 = getelementptr i8* %r, i32 %32 ; <i8*> [#uses=1] - %34 = load i8* %33, align 1 ; <i8> [#uses=1] - %.sum6 = add i32 %23, %x.12 ; <i32> [#uses=1] - %35 = getelementptr i8* %j, i32 %.sum6 ; <i8*> [#uses=1] - store i8 %34, i8* %35, align 1 - br label %bb15 - -bb15: ; preds = %bb14 - %indvar.next = add i32 %x.12, 1 ; <i32> [#uses=2] - %exitcond = icmp ne i32 %indvar.next, %smax ; <i1> [#uses=1] - br i1 %exitcond, label %bb14, label %bb15.bb17_crit_edge - -bb15.bb17_crit_edge: ; preds = %bb15 - br label %bb17 - -bb17: ; preds = %bb15.bb17_crit_edge, %bb13 - br label %bb18 - -bb18.loopexit: ; preds = %bb10.bb18.loopexit_crit_edge.split, %bb10.preheader - %36 = icmp slt i32 %x, 2 ; <i1> [#uses=1] - br i1 %36, label %bb20, label %bb.nph5 - -bb18: ; preds = %bb17 - %indvar.next1 = add i32 %y.14, 1 ; <i32> [#uses=2] - %exitcond4 = icmp ne i32 %indvar.next1, %smax3 ; <i1> [#uses=1] - br i1 %exitcond4, label %bb13, label %bb18.bb20_crit_edge - -bb18.bb20_crit_edge: ; preds = %bb18 - br label %bb18.bb20_crit_edge.split - -bb18.bb20_crit_edge.split: ; preds = %bb18.bb20_crit_edge, %bb.nph5 - br label %bb20 - -bb20: ; preds = %bb18.bb20_crit_edge.split, %bb18.loopexit - switch i32 %d, label %return [ - i32 3, label %bb22 - i32 1, label %bb29 - ] - -bb22: ; preds = %bb20 - %37 = mul i32 %x, %w ; <i32> [#uses=1] - %38 = sdiv i32 %37, 4 ; <i32> [#uses=1] - %.sum3 = add i32 %38, %.sum2 ; <i32> [#uses=2] - %39 = add i32 %x, 15 ; <i32> [#uses=1] - %40 = and i32 %39, -16 ; <i32> [#uses=1] - %41 = add i32 %w, 15 ; <i32> [#uses=1] - %42 = and i32 %41, -16 ; <i32> [#uses=1] - %43 = mul i32 %40, %s ; <i32> [#uses=1] - %44 = icmp sgt i32 %x, 0 ; <i1> [#uses=1] - br i1 %44, label %bb.nph, label %bb26 - -bb.nph: ; preds = %bb22 - br label %bb23 - -bb23: ; preds = %bb24, %bb.nph - %y.21 = phi i32 [ 0, %bb.nph ], [ %indvar.next5, %bb24 ] ; <i32> [#uses=3] - %45 = mul i32 %y.21, %42 ; <i32> [#uses=1] - %.sum1 = add i32 %45, %43 ; <i32> [#uses=1] - %46 = getelementptr i8* %r, i32 %.sum1 ; <i8*> [#uses=1] - %47 = mul i32 %y.21, %w ; <i32> [#uses=1] - %.sum5 = add i32 %47, %.sum3 ; <i32> [#uses=1] - %48 = getelementptr i8* %j, i32 %.sum5 ; <i8*> [#uses=1] - tail call void @llvm.memcpy.i32(i8* %48, i8* %46, i32 %w, i32 1) - br label %bb24 - -bb24: ; preds = %bb23 - %indvar.next5 = add i32 %y.21, 1 ; <i32> [#uses=2] - %exitcond6 = icmp ne i32 %indvar.next5, %x ; <i1> [#uses=1] - br i1 %exitcond6, label %bb23, label %bb24.bb26_crit_edge - -bb24.bb26_crit_edge: ; preds = %bb24 - br label %bb26 - -bb26: ; preds = %bb24.bb26_crit_edge, %bb22 - %49 = mul i32 %x, %w ; <i32> [#uses=1] - %.sum4 = add i32 %.sum3, %49 ; <i32> [#uses=1] - %50 = getelementptr i8* %j, i32 %.sum4 ; <i8*> [#uses=1] - %51 = mul i32 %x, %w ; <i32> [#uses=1] - %52 = sdiv i32 %51, 2 ; <i32> [#uses=1] - tail call void @llvm.memset.i32(i8* %50, i8 -128, i32 %52, i32 1) - ret void - -bb29: ; preds = %bb20, %entry - %53 = add i32 %w, 15 ; <i32> [#uses=1] - %54 = and i32 %53, -16 ; <i32> [#uses=1] - %55 = icmp sgt i32 %x, 0 ; <i1> [#uses=1] - br i1 %55, label %bb.nph11, label %bb33 - -bb.nph11: ; preds = %bb29 - br label %bb30 - -bb30: ; preds = %bb31, %bb.nph11 - %y.310 = phi i32 [ 0, %bb.nph11 ], [ %indvar.next13, %bb31 ] ; <i32> [#uses=3] - %56 = mul i32 %y.310, %54 ; <i32> [#uses=1] - %57 = getelementptr i8* %r, i32 %56 ; <i8*> [#uses=1] - %58 = mul i32 %y.310, %w ; <i32> [#uses=1] - %59 = getelementptr i8* %j, i32 %58 ; <i8*> [#uses=1] - tail call void @llvm.memcpy.i32(i8* %59, i8* %57, i32 %w, i32 1) - br label %bb31 - -bb31: ; preds = %bb30 - %indvar.next13 = add i32 %y.310, 1 ; <i32> [#uses=2] - %exitcond14 = icmp ne i32 %indvar.next13, %x ; <i1> [#uses=1] - br i1 %exitcond14, label %bb30, label %bb31.bb33_crit_edge - -bb31.bb33_crit_edge: ; preds = %bb31 - br label %bb33 - -bb33: ; preds = %bb31.bb33_crit_edge, %bb29 - %60 = mul i32 %x, %w ; <i32> [#uses=1] - %61 = getelementptr i8* %j, i32 %60 ; <i8*> [#uses=1] - %62 = mul i32 %x, %w ; <i32> [#uses=1] - %63 = sdiv i32 %62, 2 ; <i32> [#uses=1] - tail call void @llvm.memset.i32(i8* %61, i8 -128, i32 %63, i32 1) - ret void - -return: ; preds = %bb20 - ret void + %0 = mul i32 %x, %w + %1 = mul i32 %x, %w + %2 = sdiv i32 %1, 4 + %.sum2 = add i32 %2, %0 + %cond = icmp eq i32 %d, 1 + br i1 %cond, label %bb29, label %bb10.preheader + +bb10.preheader: ; preds = %entry + %3 = icmp sgt i32 %x, 0 + br i1 %3, label %bb.nph9, label %bb18.loopexit + +bb.nph7: ; preds = %bb7.preheader + %4 = mul i32 %y.08, %w + %5 = mul i32 %y.08, %s + %6 = add i32 %5, 1 + %tmp8 = icmp sgt i32 1, %w + %smax9 = select i1 %tmp8, i32 1, i32 %w + br label %bb6 + +bb6: ; preds = %bb7, %bb.nph7 + %x.06 = phi i32 [ 0, %bb.nph7 ], [ %indvar.next7, %bb7 ] + %7 = add i32 %x.06, %4 + %8 = shl i32 %x.06, 1 + %9 = add i32 %6, %8 + %10 = getelementptr i8* %r, i32 %9 + %11 = load i8* %10, align 1 + %12 = getelementptr i8* %j, i32 %7 + store i8 %11, i8* %12, align 1 + br label %bb7 + +bb7: ; preds = %bb6 + %indvar.next7 = add i32 %x.06, 1 + %exitcond10 = icmp ne i32 %indvar.next7, %smax9 + br i1 %exitcond10, label %bb6, label %bb7.bb9_crit_edge + +bb7.bb9_crit_edge: ; preds = %bb7 + br label %bb9 + +bb9: ; preds = %bb7.preheader, %bb7.bb9_crit_edge + br label %bb10 + +bb10: ; preds = %bb9 + %indvar.next11 = add i32 %y.08, 1 + %exitcond12 = icmp ne i32 %indvar.next11, %x + br i1 %exitcond12, label %bb7.preheader, label %bb10.bb18.loopexit_crit_edge + +bb10.bb18.loopexit_crit_edge: ; preds = %bb10 + br label %bb10.bb18.loopexit_crit_edge.split + +bb10.bb18.loopexit_crit_edge.split: ; preds = %bb.nph9, %bb10.bb18.loopexit_crit_edge + br label %bb18.loopexit + +bb.nph9: ; preds = %bb10.preheader + %13 = icmp sgt i32 %w, 0 + br i1 %13, label %bb.nph9.split, label %bb10.bb18.loopexit_crit_edge.split + +bb.nph9.split: ; preds = %bb.nph9 + br label %bb7.preheader + +bb7.preheader: ; preds = %bb.nph9.split, %bb10 + %y.08 = phi i32 [ 0, %bb.nph9.split ], [ %indvar.next11, %bb10 ] + br i1 true, label %bb.nph7, label %bb9 + +bb.nph5: ; preds = %bb18.loopexit + %14 = sdiv i32 %w, 2 + %15 = icmp slt i32 %w, 2 + %16 = sdiv i32 %x, 2 + br i1 %15, label %bb18.bb20_crit_edge.split, label %bb.nph5.split + +bb.nph5.split: ; preds = %bb.nph5 + %tmp2 = icmp sgt i32 1, %16 + %smax3 = select i1 %tmp2, i32 1, i32 %16 + br label %bb13 + +bb13: ; preds = %bb18, %bb.nph5.split + %y.14 = phi i32 [ 0, %bb.nph5.split ], [ %indvar.next1, %bb18 ] + %17 = mul i32 %14, %y.14 + %18 = shl i32 %y.14, 1 + %19 = srem i32 %y.14, 2 + %20 = add i32 %19, %18 + %21 = mul i32 %20, %s + br i1 true, label %bb.nph3, label %bb17 + +bb.nph3: ; preds = %bb13 + %22 = add i32 %17, %0 + %23 = add i32 %17, %.sum2 + %24 = sdiv i32 %w, 2 + %tmp = icmp sgt i32 1, %24 + %smax = select i1 %tmp, i32 1, i32 %24 + br label %bb14 + +bb14: ; preds = %bb15, %bb.nph3 + %x.12 = phi i32 [ 0, %bb.nph3 ], [ %indvar.next, %bb15 ] + %25 = shl i32 %x.12, 2 + %26 = add i32 %25, %21 + %27 = getelementptr i8* %r, i32 %26 + %28 = load i8* %27, align 1 + %.sum = add i32 %22, %x.12 + %29 = getelementptr i8* %j, i32 %.sum + store i8 %28, i8* %29, align 1 + %30 = shl i32 %x.12, 2 + %31 = or i32 %30, 2 + %32 = add i32 %31, %21 + %33 = getelementptr i8* %r, i32 %32 + %34 = load i8* %33, align 1 + %.sum6 = add i32 %23, %x.12 + %35 = getelementptr i8* %j, i32 %.sum6 + store i8 %34, i8* %35, align 1 + br label %bb15 + +bb15: ; preds = %bb14 + %indvar.next = add i32 %x.12, 1 + %exitcond = icmp ne i32 %indvar.next, %smax + br i1 %exitcond, label %bb14, label %bb15.bb17_crit_edge + +bb15.bb17_crit_edge: ; preds = %bb15 + br label %bb17 + +bb17: ; preds = %bb15.bb17_crit_edge, %bb13 + br label %bb18 + +bb18.loopexit: ; preds = %bb10.bb18.loopexit_crit_edge.split, %bb10.preheader + %36 = icmp slt i32 %x, 2 + br i1 %36, label %bb20, label %bb.nph5 + +bb18: ; preds = %bb17 + %indvar.next1 = add i32 %y.14, 1 + %exitcond4 = icmp ne i32 %indvar.next1, %smax3 + br i1 %exitcond4, label %bb13, label %bb18.bb20_crit_edge + +bb18.bb20_crit_edge: ; preds = %bb18 + br label %bb18.bb20_crit_edge.split + +bb18.bb20_crit_edge.split: ; preds = %bb18.bb20_crit_edge, %bb.nph5 + br label %bb20 + +bb20: ; preds = %bb18.bb20_crit_edge.split, %bb18.loopexit + switch i32 %d, label %return [ + i32 3, label %bb22 + i32 1, label %bb29 + ] + +bb22: ; preds = %bb20 + %37 = mul i32 %x, %w + %38 = sdiv i32 %37, 4 + %.sum3 = add i32 %38, %.sum2 + %39 = add i32 %x, 15 + %40 = and i32 %39, -16 + %41 = add i32 %w, 15 + %42 = and i32 %41, -16 + %43 = mul i32 %40, %s + %44 = icmp sgt i32 %x, 0 + br i1 %44, label %bb.nph, label %bb26 + +bb.nph: ; preds = %bb22 + br label %bb23 + +bb23: ; preds = %bb24, %bb.nph + %y.21 = phi i32 [ 0, %bb.nph ], [ %indvar.next5, %bb24 ] + %45 = mul i32 %y.21, %42 + %.sum1 = add i32 %45, %43 + %46 = getelementptr i8* %r, i32 %.sum1 + %47 = mul i32 %y.21, %w + %.sum5 = add i32 %47, %.sum3 + %48 = getelementptr i8* %j, i32 %.sum5 + tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* %48, i8* %46, i32 %w, i32 1, i1 false) + br label %bb24 + +bb24: ; preds = %bb23 + %indvar.next5 = add i32 %y.21, 1 + %exitcond6 = icmp ne i32 %indvar.next5, %x + br i1 %exitcond6, label %bb23, label %bb24.bb26_crit_edge + +bb24.bb26_crit_edge: ; preds = %bb24 + br label %bb26 + +bb26: ; preds = %bb24.bb26_crit_edge, %bb22 + %49 = mul i32 %x, %w + %.sum4 = add i32 %.sum3, %49 + %50 = getelementptr i8* %j, i32 %.sum4 + %51 = mul i32 %x, %w + %52 = sdiv i32 %51, 2 + tail call void @llvm.memset.p0i8.i32(i8* %50, i8 -128, i32 %52, i32 1, i1 false) + ret void + +bb29: ; preds = %bb20, %entry + %53 = add i32 %w, 15 + %54 = and i32 %53, -16 + %55 = icmp sgt i32 %x, 0 + br i1 %55, label %bb.nph11, label %bb33 + +bb.nph11: ; preds = %bb29 + br label %bb30 + +bb30: ; preds = %bb31, %bb.nph11 + %y.310 = phi i32 [ 0, %bb.nph11 ], [ %indvar.next13, %bb31 ] + %56 = mul i32 %y.310, %54 + %57 = getelementptr i8* %r, i32 %56 + %58 = mul i32 %y.310, %w + %59 = getelementptr i8* %j, i32 %58 + tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* %59, i8* %57, i32 %w, i32 1, i1 false) + br label %bb31 + +bb31: ; preds = %bb30 + %indvar.next13 = add i32 %y.310, 1 + %exitcond14 = icmp ne i32 %indvar.next13, %x + br i1 %exitcond14, label %bb30, label %bb31.bb33_crit_edge + +bb31.bb33_crit_edge: ; preds = %bb31 + br label %bb33 + +bb33: ; preds = %bb31.bb33_crit_edge, %bb29 + %60 = mul i32 %x, %w + %61 = getelementptr i8* %j, i32 %60 + %62 = mul i32 %x, %w + %63 = sdiv i32 %62, 2 + tail call void @llvm.memset.p0i8.i32(i8* %61, i8 -128, i32 %63, i32 1, i1 false) + ret void + +return: ; preds = %bb20 + ret void } define void @bar(i8* %r, i32 %s, i32 %w, i32 %x, i8* %j, i32 %d) nounwind { entry: - %0 = mul i32 %x, %w ; <i32> [#uses=2] - %1 = mul i32 %x, %w ; <i32> [#uses=1] - %2 = udiv i32 %1, 4 ; <i32> [#uses=1] - %.sum2 = add i32 %2, %0 ; <i32> [#uses=2] - %cond = icmp eq i32 %d, 1 ; <i1> [#uses=1] - br i1 %cond, label %bb29, label %bb10.preheader - -bb10.preheader: ; preds = %entry - %3 = icmp ne i32 %x, 0 ; <i1> [#uses=1] - br i1 %3, label %bb.nph9, label %bb18.loopexit - -bb.nph7: ; preds = %bb7.preheader - %4 = mul i32 %y.08, %w ; <i32> [#uses=1] - %5 = mul i32 %y.08, %s ; <i32> [#uses=1] - %6 = add i32 %5, 1 ; <i32> [#uses=1] - %tmp8 = icmp ugt i32 1, %w ; <i1> [#uses=1] - %smax9 = select i1 %tmp8, i32 1, i32 %w ; <i32> [#uses=1] - br label %bb6 - -bb6: ; preds = %bb7, %bb.nph7 - %x.06 = phi i32 [ 0, %bb.nph7 ], [ %indvar.next7, %bb7 ] ; <i32> [#uses=3] - %7 = add i32 %x.06, %4 ; <i32> [#uses=1] - %8 = shl i32 %x.06, 1 ; <i32> [#uses=1] - %9 = add i32 %6, %8 ; <i32> [#uses=1] - %10 = getelementptr i8* %r, i32 %9 ; <i8*> [#uses=1] - %11 = load i8* %10, align 1 ; <i8> [#uses=1] - %12 = getelementptr i8* %j, i32 %7 ; <i8*> [#uses=1] - store i8 %11, i8* %12, align 1 - br label %bb7 - -bb7: ; preds = %bb6 - %indvar.next7 = add i32 %x.06, 1 ; <i32> [#uses=2] - %exitcond10 = icmp ne i32 %indvar.next7, %smax9 ; <i1> [#uses=1] - br i1 %exitcond10, label %bb6, label %bb7.bb9_crit_edge - -bb7.bb9_crit_edge: ; preds = %bb7 - br label %bb9 - -bb9: ; preds = %bb7.preheader, %bb7.bb9_crit_edge - br label %bb10 - -bb10: ; preds = %bb9 - %indvar.next11 = add i32 %y.08, 1 ; <i32> [#uses=2] - %exitcond12 = icmp ne i32 %indvar.next11, %x ; <i1> [#uses=1] - br i1 %exitcond12, label %bb7.preheader, label %bb10.bb18.loopexit_crit_edge - -bb10.bb18.loopexit_crit_edge: ; preds = %bb10 - br label %bb10.bb18.loopexit_crit_edge.split - -bb10.bb18.loopexit_crit_edge.split: ; preds = %bb.nph9, %bb10.bb18.loopexit_crit_edge - br label %bb18.loopexit - -bb.nph9: ; preds = %bb10.preheader - %13 = icmp ugt i32 %w, 0 ; <i1> [#uses=1] - br i1 %13, label %bb.nph9.split, label %bb10.bb18.loopexit_crit_edge.split - -bb.nph9.split: ; preds = %bb.nph9 - br label %bb7.preheader - -bb7.preheader: ; preds = %bb.nph9.split, %bb10 - %y.08 = phi i32 [ 0, %bb.nph9.split ], [ %indvar.next11, %bb10 ] ; <i32> [#uses=3] - br i1 true, label %bb.nph7, label %bb9 - -bb.nph5: ; preds = %bb18.loopexit - %14 = udiv i32 %w, 2 ; <i32> [#uses=1] - %15 = icmp ult i32 %w, 2 ; <i1> [#uses=1] - %16 = udiv i32 %x, 2 ; <i32> [#uses=2] - br i1 %15, label %bb18.bb20_crit_edge.split, label %bb.nph5.split - -bb.nph5.split: ; preds = %bb.nph5 - %tmp2 = icmp ugt i32 1, %16 ; <i1> [#uses=1] - %smax3 = select i1 %tmp2, i32 1, i32 %16 ; <i32> [#uses=1] - br label %bb13 - -bb13: ; preds = %bb18, %bb.nph5.split - %y.14 = phi i32 [ 0, %bb.nph5.split ], [ %indvar.next1, %bb18 ] ; <i32> [#uses=4] - %17 = mul i32 %14, %y.14 ; <i32> [#uses=2] - %18 = shl i32 %y.14, 1 ; <i32> [#uses=1] - %19 = urem i32 %y.14, 2 ; <i32> [#uses=1] - %20 = add i32 %19, %18 ; <i32> [#uses=1] - %21 = mul i32 %20, %s ; <i32> [#uses=2] - br i1 true, label %bb.nph3, label %bb17 - -bb.nph3: ; preds = %bb13 - %22 = add i32 %17, %0 ; <i32> [#uses=1] - %23 = add i32 %17, %.sum2 ; <i32> [#uses=1] - %24 = udiv i32 %w, 2 ; <i32> [#uses=2] - %tmp = icmp ugt i32 1, %24 ; <i1> [#uses=1] - %smax = select i1 %tmp, i32 1, i32 %24 ; <i32> [#uses=1] - br label %bb14 - -bb14: ; preds = %bb15, %bb.nph3 - %x.12 = phi i32 [ 0, %bb.nph3 ], [ %indvar.next, %bb15 ] ; <i32> [#uses=5] - %25 = shl i32 %x.12, 2 ; <i32> [#uses=1] - %26 = add i32 %25, %21 ; <i32> [#uses=1] - %27 = getelementptr i8* %r, i32 %26 ; <i8*> [#uses=1] - %28 = load i8* %27, align 1 ; <i8> [#uses=1] - %.sum = add i32 %22, %x.12 ; <i32> [#uses=1] - %29 = getelementptr i8* %j, i32 %.sum ; <i8*> [#uses=1] - store i8 %28, i8* %29, align 1 - %30 = shl i32 %x.12, 2 ; <i32> [#uses=1] - %31 = or i32 %30, 2 ; <i32> [#uses=1] - %32 = add i32 %31, %21 ; <i32> [#uses=1] - %33 = getelementptr i8* %r, i32 %32 ; <i8*> [#uses=1] - %34 = load i8* %33, align 1 ; <i8> [#uses=1] - %.sum6 = add i32 %23, %x.12 ; <i32> [#uses=1] - %35 = getelementptr i8* %j, i32 %.sum6 ; <i8*> [#uses=1] - store i8 %34, i8* %35, align 1 - br label %bb15 - -bb15: ; preds = %bb14 - %indvar.next = add i32 %x.12, 1 ; <i32> [#uses=2] - %exitcond = icmp ne i32 %indvar.next, %smax ; <i1> [#uses=1] - br i1 %exitcond, label %bb14, label %bb15.bb17_crit_edge - -bb15.bb17_crit_edge: ; preds = %bb15 - br label %bb17 - -bb17: ; preds = %bb15.bb17_crit_edge, %bb13 - br label %bb18 - -bb18.loopexit: ; preds = %bb10.bb18.loopexit_crit_edge.split, %bb10.preheader - %36 = icmp ult i32 %x, 2 ; <i1> [#uses=1] - br i1 %36, label %bb20, label %bb.nph5 - -bb18: ; preds = %bb17 - %indvar.next1 = add i32 %y.14, 1 ; <i32> [#uses=2] - %exitcond4 = icmp ne i32 %indvar.next1, %smax3 ; <i1> [#uses=1] - br i1 %exitcond4, label %bb13, label %bb18.bb20_crit_edge - -bb18.bb20_crit_edge: ; preds = %bb18 - br label %bb18.bb20_crit_edge.split - -bb18.bb20_crit_edge.split: ; preds = %bb18.bb20_crit_edge, %bb.nph5 - br label %bb20 - -bb20: ; preds = %bb18.bb20_crit_edge.split, %bb18.loopexit - switch i32 %d, label %return [ - i32 3, label %bb22 - i32 1, label %bb29 - ] - -bb22: ; preds = %bb20 - %37 = mul i32 %x, %w ; <i32> [#uses=1] - %38 = udiv i32 %37, 4 ; <i32> [#uses=1] - %.sum3 = add i32 %38, %.sum2 ; <i32> [#uses=2] - %39 = add i32 %x, 15 ; <i32> [#uses=1] - %40 = and i32 %39, -16 ; <i32> [#uses=1] - %41 = add i32 %w, 15 ; <i32> [#uses=1] - %42 = and i32 %41, -16 ; <i32> [#uses=1] - %43 = mul i32 %40, %s ; <i32> [#uses=1] - %44 = icmp ugt i32 %x, 0 ; <i1> [#uses=1] - br i1 %44, label %bb.nph, label %bb26 - -bb.nph: ; preds = %bb22 - br label %bb23 - -bb23: ; preds = %bb24, %bb.nph - %y.21 = phi i32 [ 0, %bb.nph ], [ %indvar.next5, %bb24 ] ; <i32> [#uses=3] - %45 = mul i32 %y.21, %42 ; <i32> [#uses=1] - %.sum1 = add i32 %45, %43 ; <i32> [#uses=1] - %46 = getelementptr i8* %r, i32 %.sum1 ; <i8*> [#uses=1] - %47 = mul i32 %y.21, %w ; <i32> [#uses=1] - %.sum5 = add i32 %47, %.sum3 ; <i32> [#uses=1] - %48 = getelementptr i8* %j, i32 %.sum5 ; <i8*> [#uses=1] - tail call void @llvm.memcpy.i32(i8* %48, i8* %46, i32 %w, i32 1) - br label %bb24 - -bb24: ; preds = %bb23 - %indvar.next5 = add i32 %y.21, 1 ; <i32> [#uses=2] - %exitcond6 = icmp ne i32 %indvar.next5, %x ; <i1> [#uses=1] - br i1 %exitcond6, label %bb23, label %bb24.bb26_crit_edge - -bb24.bb26_crit_edge: ; preds = %bb24 - br label %bb26 - -bb26: ; preds = %bb24.bb26_crit_edge, %bb22 - %49 = mul i32 %x, %w ; <i32> [#uses=1] - %.sum4 = add i32 %.sum3, %49 ; <i32> [#uses=1] - %50 = getelementptr i8* %j, i32 %.sum4 ; <i8*> [#uses=1] - %51 = mul i32 %x, %w ; <i32> [#uses=1] - %52 = udiv i32 %51, 2 ; <i32> [#uses=1] - tail call void @llvm.memset.i32(i8* %50, i8 -128, i32 %52, i32 1) - ret void - -bb29: ; preds = %bb20, %entry - %53 = add i32 %w, 15 ; <i32> [#uses=1] - %54 = and i32 %53, -16 ; <i32> [#uses=1] - %55 = icmp ugt i32 %x, 0 ; <i1> [#uses=1] - br i1 %55, label %bb.nph11, label %bb33 - -bb.nph11: ; preds = %bb29 - br label %bb30 - -bb30: ; preds = %bb31, %bb.nph11 - %y.310 = phi i32 [ 0, %bb.nph11 ], [ %indvar.next13, %bb31 ] ; <i32> [#uses=3] - %56 = mul i32 %y.310, %54 ; <i32> [#uses=1] - %57 = getelementptr i8* %r, i32 %56 ; <i8*> [#uses=1] - %58 = mul i32 %y.310, %w ; <i32> [#uses=1] - %59 = getelementptr i8* %j, i32 %58 ; <i8*> [#uses=1] - tail call void @llvm.memcpy.i32(i8* %59, i8* %57, i32 %w, i32 1) - br label %bb31 - -bb31: ; preds = %bb30 - %indvar.next13 = add i32 %y.310, 1 ; <i32> [#uses=2] - %exitcond14 = icmp ne i32 %indvar.next13, %x ; <i1> [#uses=1] - br i1 %exitcond14, label %bb30, label %bb31.bb33_crit_edge - -bb31.bb33_crit_edge: ; preds = %bb31 - br label %bb33 - -bb33: ; preds = %bb31.bb33_crit_edge, %bb29 - %60 = mul i32 %x, %w ; <i32> [#uses=1] - %61 = getelementptr i8* %j, i32 %60 ; <i8*> [#uses=1] - %62 = mul i32 %x, %w ; <i32> [#uses=1] - %63 = udiv i32 %62, 2 ; <i32> [#uses=1] - tail call void @llvm.memset.i32(i8* %61, i8 -128, i32 %63, i32 1) - ret void - -return: ; preds = %bb20 - ret void + %0 = mul i32 %x, %w + %1 = mul i32 %x, %w + %2 = udiv i32 %1, 4 + %.sum2 = add i32 %2, %0 + %cond = icmp eq i32 %d, 1 + br i1 %cond, label %bb29, label %bb10.preheader + +bb10.preheader: ; preds = %entry + %3 = icmp ne i32 %x, 0 + br i1 %3, label %bb.nph9, label %bb18.loopexit + +bb.nph7: ; preds = %bb7.preheader + %4 = mul i32 %y.08, %w + %5 = mul i32 %y.08, %s + %6 = add i32 %5, 1 + %tmp8 = icmp ugt i32 1, %w + %smax9 = select i1 %tmp8, i32 1, i32 %w + br label %bb6 + +bb6: ; preds = %bb7, %bb.nph7 + %x.06 = phi i32 [ 0, %bb.nph7 ], [ %indvar.next7, %bb7 ] + %7 = add i32 %x.06, %4 + %8 = shl i32 %x.06, 1 + %9 = add i32 %6, %8 + %10 = getelementptr i8* %r, i32 %9 + %11 = load i8* %10, align 1 + %12 = getelementptr i8* %j, i32 %7 + store i8 %11, i8* %12, align 1 + br label %bb7 + +bb7: ; preds = %bb6 + %indvar.next7 = add i32 %x.06, 1 + %exitcond10 = icmp ne i32 %indvar.next7, %smax9 + br i1 %exitcond10, label %bb6, label %bb7.bb9_crit_edge + +bb7.bb9_crit_edge: ; preds = %bb7 + br label %bb9 + +bb9: ; preds = %bb7.preheader, %bb7.bb9_crit_edge + br label %bb10 + +bb10: ; preds = %bb9 + %indvar.next11 = add i32 %y.08, 1 + %exitcond12 = icmp ne i32 %indvar.next11, %x + br i1 %exitcond12, label %bb7.preheader, label %bb10.bb18.loopexit_crit_edge + +bb10.bb18.loopexit_crit_edge: ; preds = %bb10 + br label %bb10.bb18.loopexit_crit_edge.split + +bb10.bb18.loopexit_crit_edge.split: ; preds = %bb.nph9, %bb10.bb18.loopexit_crit_edge + br label %bb18.loopexit + +bb.nph9: ; preds = %bb10.preheader + %13 = icmp ugt i32 %w, 0 + br i1 %13, label %bb.nph9.split, label %bb10.bb18.loopexit_crit_edge.split + +bb.nph9.split: ; preds = %bb.nph9 + br label %bb7.preheader + +bb7.preheader: ; preds = %bb.nph9.split, %bb10 + %y.08 = phi i32 [ 0, %bb.nph9.split ], [ %indvar.next11, %bb10 ] + br i1 true, label %bb.nph7, label %bb9 + +bb.nph5: ; preds = %bb18.loopexit + %14 = udiv i32 %w, 2 + %15 = icmp ult i32 %w, 2 + %16 = udiv i32 %x, 2 + br i1 %15, label %bb18.bb20_crit_edge.split, label %bb.nph5.split + +bb.nph5.split: ; preds = %bb.nph5 + %tmp2 = icmp ugt i32 1, %16 + %smax3 = select i1 %tmp2, i32 1, i32 %16 + br label %bb13 + +bb13: ; preds = %bb18, %bb.nph5.split + %y.14 = phi i32 [ 0, %bb.nph5.split ], [ %indvar.next1, %bb18 ] + %17 = mul i32 %14, %y.14 + %18 = shl i32 %y.14, 1 + %19 = urem i32 %y.14, 2 + %20 = add i32 %19, %18 + %21 = mul i32 %20, %s + br i1 true, label %bb.nph3, label %bb17 + +bb.nph3: ; preds = %bb13 + %22 = add i32 %17, %0 + %23 = add i32 %17, %.sum2 + %24 = udiv i32 %w, 2 + %tmp = icmp ugt i32 1, %24 + %smax = select i1 %tmp, i32 1, i32 %24 + br label %bb14 + +bb14: ; preds = %bb15, %bb.nph3 + %x.12 = phi i32 [ 0, %bb.nph3 ], [ %indvar.next, %bb15 ] + %25 = shl i32 %x.12, 2 + %26 = add i32 %25, %21 + %27 = getelementptr i8* %r, i32 %26 + %28 = load i8* %27, align 1 + %.sum = add i32 %22, %x.12 + %29 = getelementptr i8* %j, i32 %.sum + store i8 %28, i8* %29, align 1 + %30 = shl i32 %x.12, 2 + %31 = or i32 %30, 2 + %32 = add i32 %31, %21 + %33 = getelementptr i8* %r, i32 %32 + %34 = load i8* %33, align 1 + %.sum6 = add i32 %23, %x.12 + %35 = getelementptr i8* %j, i32 %.sum6 + store i8 %34, i8* %35, align 1 + br label %bb15 + +bb15: ; preds = %bb14 + %indvar.next = add i32 %x.12, 1 + %exitcond = icmp ne i32 %indvar.next, %smax + br i1 %exitcond, label %bb14, label %bb15.bb17_crit_edge + +bb15.bb17_crit_edge: ; preds = %bb15 + br label %bb17 + +bb17: ; preds = %bb15.bb17_crit_edge, %bb13 + br label %bb18 + +bb18.loopexit: ; preds = %bb10.bb18.loopexit_crit_edge.split, %bb10.preheader + %36 = icmp ult i32 %x, 2 + br i1 %36, label %bb20, label %bb.nph5 + +bb18: ; preds = %bb17 + %indvar.next1 = add i32 %y.14, 1 + %exitcond4 = icmp ne i32 %indvar.next1, %smax3 + br i1 %exitcond4, label %bb13, label %bb18.bb20_crit_edge + +bb18.bb20_crit_edge: ; preds = %bb18 + br label %bb18.bb20_crit_edge.split + +bb18.bb20_crit_edge.split: ; preds = %bb18.bb20_crit_edge, %bb.nph5 + br label %bb20 + +bb20: ; preds = %bb18.bb20_crit_edge.split, %bb18.loopexit + switch i32 %d, label %return [ + i32 3, label %bb22 + i32 1, label %bb29 + ] + +bb22: ; preds = %bb20 + %37 = mul i32 %x, %w + %38 = udiv i32 %37, 4 + %.sum3 = add i32 %38, %.sum2 + %39 = add i32 %x, 15 + %40 = and i32 %39, -16 + %41 = add i32 %w, 15 + %42 = and i32 %41, -16 + %43 = mul i32 %40, %s + %44 = icmp ugt i32 %x, 0 + br i1 %44, label %bb.nph, label %bb26 + +bb.nph: ; preds = %bb22 + br label %bb23 + +bb23: ; preds = %bb24, %bb.nph + %y.21 = phi i32 [ 0, %bb.nph ], [ %indvar.next5, %bb24 ] + %45 = mul i32 %y.21, %42 + %.sum1 = add i32 %45, %43 + %46 = getelementptr i8* %r, i32 %.sum1 + %47 = mul i32 %y.21, %w + %.sum5 = add i32 %47, %.sum3 + %48 = getelementptr i8* %j, i32 %.sum5 + tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* %48, i8* %46, i32 %w, i32 1, i1 false) + br label %bb24 + +bb24: ; preds = %bb23 + %indvar.next5 = add i32 %y.21, 1 + %exitcond6 = icmp ne i32 %indvar.next5, %x + br i1 %exitcond6, label %bb23, label %bb24.bb26_crit_edge + +bb24.bb26_crit_edge: ; preds = %bb24 + br label %bb26 + +bb26: ; preds = %bb24.bb26_crit_edge, %bb22 + %49 = mul i32 %x, %w + %.sum4 = add i32 %.sum3, %49 + %50 = getelementptr i8* %j, i32 %.sum4 + %51 = mul i32 %x, %w + %52 = udiv i32 %51, 2 + tail call void @llvm.memset.p0i8.i32(i8* %50, i8 -128, i32 %52, i32 1, i1 false) + ret void + +bb29: ; preds = %bb20, %entry + %53 = add i32 %w, 15 + %54 = and i32 %53, -16 + %55 = icmp ugt i32 %x, 0 + br i1 %55, label %bb.nph11, label %bb33 + +bb.nph11: ; preds = %bb29 + br label %bb30 + +bb30: ; preds = %bb31, %bb.nph11 + %y.310 = phi i32 [ 0, %bb.nph11 ], [ %indvar.next13, %bb31 ] + %56 = mul i32 %y.310, %54 + %57 = getelementptr i8* %r, i32 %56 + %58 = mul i32 %y.310, %w + %59 = getelementptr i8* %j, i32 %58 + tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* %59, i8* %57, i32 %w, i32 1, i1 false) + br label %bb31 + +bb31: ; preds = %bb30 + %indvar.next13 = add i32 %y.310, 1 + %exitcond14 = icmp ne i32 %indvar.next13, %x + br i1 %exitcond14, label %bb30, label %bb31.bb33_crit_edge + +bb31.bb33_crit_edge: ; preds = %bb31 + br label %bb33 + +bb33: ; preds = %bb31.bb33_crit_edge, %bb29 + %60 = mul i32 %x, %w + %61 = getelementptr i8* %j, i32 %60 + %62 = mul i32 %x, %w + %63 = udiv i32 %62, 2 + tail call void @llvm.memset.p0i8.i32(i8* %61, i8 -128, i32 %63, i32 1, i1 false) + ret void + +return: ; preds = %bb20 + ret void } -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind -declare void @llvm.memset.i32(i8*, i8, i32, i32) nounwind +declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1) nounwind diff --git a/test/CodeGen/X86/personality.ll b/test/CodeGen/X86/personality.ll index e952a9b..d3d8e3f 100644 --- a/test/CodeGen/X86/personality.ll +++ b/test/CodeGen/X86/personality.ll @@ -4,40 +4,41 @@ define void @_Z1fv() { entry: - invoke void @_Z1gv( ) - to label %return unwind label %unwind + invoke void @_Z1gv() + to label %return unwind label %unwind -unwind: ; preds = %entry - br i1 false, label %eh_then, label %cleanup20 +unwind: ; preds = %entry + br i1 false, label %eh_then, label %cleanup20 -eh_then: ; preds = %unwind - invoke void @__cxa_end_catch( ) - to label %return unwind label %unwind10 +eh_then: ; preds = %unwind + invoke void @__cxa_end_catch() + to label %return unwind label %unwind10 -unwind10: ; preds = %eh_then - %eh_select13 = tail call i64 (i8*, i8*, ...)* @llvm.eh.selector.i64( i8* null, i8* bitcast (void ()* @__gxx_personality_v0 to i8*), i32 1 ) ; <i32> [#uses=2] - %tmp18 = icmp slt i64 %eh_select13, 0 ; <i1> [#uses=1] - br i1 %tmp18, label %filter, label %cleanup20 +unwind10: ; preds = %eh_then + %upgraded.eh_select13 = tail call i32 (i8*, i8*, ...)* @llvm.eh.selector(i8* null, i8* bitcast (void ()* @__gxx_personality_v0 to i8*), i32 1) + %upgraded.eh_select131 = sext i32 %upgraded.eh_select13 to i64 + %tmp18 = icmp slt i64 %upgraded.eh_select131, 0 + br i1 %tmp18, label %filter, label %cleanup20 -filter: ; preds = %unwind10 - unreachable +filter: ; preds = %unwind10 + unreachable -cleanup20: ; preds = %unwind10, %unwind - %eh_selector.0 = phi i64 [ 0, %unwind ], [ %eh_select13, %unwind10 ] ; <i32> [#uses=0] - ret void +cleanup20: ; preds = %unwind10, %unwind + %eh_selector.0 = phi i64 [ 0, %unwind ], [ %upgraded.eh_select131, %unwind10 ] + ret void -return: ; preds = %eh_then, %entry - ret void +return: ; preds = %eh_then, %entry + ret void } declare void @_Z1gv() -declare i64 @llvm.eh.selector.i64(i8*, i8*, ...) - declare void @__gxx_personality_v0() declare void @__cxa_end_catch() +declare i32 @llvm.eh.selector(i8*, i8*, ...) nounwind + ; X64: zPLR ; X64: .byte 155 ; X64-NEXT: .long ___gxx_personality_v0@GOTPCREL+4 diff --git a/test/CodeGen/X86/pre-split2.ll b/test/CodeGen/X86/pre-split2.ll deleted file mode 100644 index 670737b..0000000 --- a/test/CodeGen/X86/pre-split2.ll +++ /dev/null @@ -1,26 +0,0 @@ -; RUN: llc < %s -march=x86 -mattr=+sse2 -pre-alloc-split -stats -regalloc=linearscan |& \ -; RUN: grep {pre-alloc-split} | count 2 - -define i32 @t(i32 %arg) { -entry: - br label %bb6 - -.noexc6: ; preds = %bb6 - %0 = and i32 %2, -8 ; <i32> [#uses=1] - tail call void @llvm.memmove.i32(i8* %3, i8* null, i32 %0, i32 1) nounwind - store double %1, double* null, align 8 - br label %bb6 - -bb6: ; preds = %.noexc6, %entry - %1 = uitofp i32 %arg to double ; <double> [#uses=1] - %2 = sub i32 0, 0 ; <i32> [#uses=1] - %3 = invoke i8* @_Znwm(i32 0) - to label %.noexc6 unwind label %lpad32 ; <i8*> [#uses=1] - -lpad32: ; preds = %bb6 - unreachable -} - -declare void @llvm.memmove.i32(i8*, i8*, i32, i32) nounwind - -declare i8* @_Znwm(i32) diff --git a/test/CodeGen/X86/pre-split3.ll b/test/CodeGen/X86/pre-split3.ll deleted file mode 100644 index 0c49a91..0000000 --- a/test/CodeGen/X86/pre-split3.ll +++ /dev/null @@ -1,26 +0,0 @@ -; RUN: llc < %s -march=x86 -mattr=+sse2 -pre-alloc-split -regalloc=linearscan -stats |& \ -; RUN: grep {pre-alloc-split} | grep {Number of intervals split} | grep 1 - -define i32 @t(i32 %arg) { -entry: - br label %bb6 - -.noexc6: ; preds = %bb6 - %0 = and i32 %2, -8 ; <i32> [#uses=1] - tail call void @llvm.memmove.i32(i8* %3, i8* null, i32 %0, i32 1) nounwind - store double %1, double* null, align 8 - br label %bb6 - -bb6: ; preds = %.noexc6, %entry - %1 = uitofp i32 %arg to double ; <double> [#uses=1] - %2 = sub i32 0, 0 ; <i32> [#uses=1] - %3 = invoke i8* @_Znwm(i32 0) - to label %.noexc6 unwind label %lpad32 ; <i8*> [#uses=1] - -lpad32: ; preds = %bb6 - unreachable -} - -declare void @llvm.memmove.i32(i8*, i8*, i32, i32) nounwind - -declare i8* @_Znwm(i32) diff --git a/test/CodeGen/X86/unaligned-load.ll b/test/CodeGen/X86/unaligned-load.ll index 9f70489..d8fffbe 100644 --- a/test/CodeGen/X86/unaligned-load.ll +++ b/test/CodeGen/X86/unaligned-load.ll @@ -10,9 +10,17 @@ entry: %String2Loc = alloca [31 x i8], align 1 br label %bb -bb: +bb: ; preds = %bb, %entry %String2Loc9 = getelementptr inbounds [31 x i8]* %String2Loc, i64 0, i64 0 - call void @llvm.memcpy.i64(i8* %String2Loc9, i8* getelementptr inbounds ([31 x i8]* @.str3, i64 0, i64 0), i64 31, i32 1) + call void @llvm.memcpy.p0i8.p0i8.i64(i8* %String2Loc9, i8* getelementptr inbounds ([31 x i8]* @.str3, i64 0, i64 0), i64 31, i32 1, i1 false) + br label %bb + +return: ; No predecessors! + ret void +} + +declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture, i64, i32, i1) nounwind + ; I386: calll {{_?}}memcpy ; CORE2: movabsq @@ -20,13 +28,6 @@ bb: ; CORE2: movabsq ; COREI7: movups _.str3 - br label %bb - -return: - ret void -} - -declare void @llvm.memcpy.i64(i8* nocapture, i8* nocapture, i64, i32) nounwind ; CORE2: .section ; CORE2: .align 3 diff --git a/test/CodeGen/X86/variable-sized-darwin-bzero.ll b/test/CodeGen/X86/variable-sized-darwin-bzero.ll index 4817db2..1e86d75 100644 --- a/test/CodeGen/X86/variable-sized-darwin-bzero.ll +++ b/test/CodeGen/X86/variable-sized-darwin-bzero.ll @@ -1,8 +1,8 @@ ; RUN: llc < %s -march=x86 -mtriple=i686-apple-darwin10 | grep __bzero -declare void @llvm.memset.i64(i8*, i8, i64, i32) - define void @foo(i8* %p, i64 %n) { - call void @llvm.memset.i64(i8* %p, i8 0, i64 %n, i32 4) + call void @llvm.memset.p0i8.i64(i8* %p, i8 0, i64 %n, i32 4, i1 false) ret void } + +declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) nounwind diff --git a/test/Transforms/DeadStoreElimination/crash.ll b/test/Transforms/DeadStoreElimination/crash.ll index bb279cd..148695f 100644 --- a/test/Transforms/DeadStoreElimination/crash.ll +++ b/test/Transforms/DeadStoreElimination/crash.ll @@ -36,11 +36,11 @@ bb14: ; preds = %bb4 %6 = getelementptr inbounds i16* %2, i64 undef ; <i16*> [#uses=1] store i16 undef, i16* %6, align 2 %7 = getelementptr inbounds i8* %5, i64 undef ; <i8*> [#uses=1] - call void @llvm.memcpy.i64(i8* %7, i8* undef, i64 undef, i32 1) nounwind + call void @llvm.memcpy.p0i8.p0i8.i64(i8* %7, i8* undef, i64 undef, i32 1, i1 false) unreachable } -declare void @llvm.memcpy.i64(i8* nocapture, i8* nocapture, i64, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture, i64, i32, i1) nounwind ; rdar://7635088 diff --git a/test/Transforms/DeadStoreElimination/lifetime.ll b/test/Transforms/DeadStoreElimination/lifetime.ll index 2b5cc5a..6785653 100644 --- a/test/Transforms/DeadStoreElimination/lifetime.ll +++ b/test/Transforms/DeadStoreElimination/lifetime.ll @@ -4,7 +4,7 @@ target datalayout = "E-p:64:64:64-a0:0:8-f32:32:32-f64:64:64-i1:8:8-i8:8:8-i16:1 declare void @llvm.lifetime.start(i64, i8* nocapture) nounwind declare void @llvm.lifetime.end(i64, i8* nocapture) nounwind -declare void @llvm.memset.i8(i8*, i8, i8, i32) +declare void @llvm.memset.p0i8.i8(i8* nocapture, i8, i8, i32, i1) nounwind define void @test1() { ; CHECK: @test1 @@ -14,7 +14,7 @@ define void @test1() { call void @llvm.lifetime.end(i64 1, i8* %A) ; CHECK: lifetime.end - call void @llvm.memset.i8(i8* %A, i8 0, i8 -1, i32 0) + call void @llvm.memset.p0i8.i8(i8* %A, i8 0, i8 -1, i32 0, i1 false) ; CHECK-NOT: memset ret void diff --git a/test/Transforms/DeadStoreElimination/memintrinsics.ll b/test/Transforms/DeadStoreElimination/memintrinsics.ll index e31e9fa..d5c5365 100644 --- a/test/Transforms/DeadStoreElimination/memintrinsics.ll +++ b/test/Transforms/DeadStoreElimination/memintrinsics.ll @@ -1,8 +1,8 @@ ; RUN: opt -S -dse < %s | FileCheck %s -declare void @llvm.memcpy.i8(i8*, i8*, i8, i32) -declare void @llvm.memmove.i8(i8*, i8*, i8, i32) -declare void @llvm.memset.i8(i8*, i8, i8, i32) +declare void @llvm.memcpy.p0i8.p0i8.i8(i8* nocapture, i8* nocapture, i8, i32, i1) nounwind +declare void @llvm.memmove.p0i8.p0i8.i8(i8* nocapture, i8* nocapture, i8, i32, i1) nounwind +declare void @llvm.memset.p0i8.i8(i8* nocapture, i8, i8, i32, i1) nounwind define void @test1() { ; CHECK: @test1 @@ -12,7 +12,7 @@ define void @test1() { store i8 0, i8* %A ;; Written to by memcpy ; CHECK-NOT: store - call void @llvm.memcpy.i8(i8* %A, i8* %B, i8 -1, i32 0) + call void @llvm.memcpy.p0i8.p0i8.i8(i8* %A, i8* %B, i8 -1, i32 0, i1 false) ret void ; CHECK: ret void @@ -26,7 +26,7 @@ define void @test2() { store i8 0, i8* %A ;; Written to by memmove ; CHECK-NOT: store - call void @llvm.memmove.i8(i8* %A, i8* %B, i8 -1, i32 0) + call void @llvm.memmove.p0i8.p0i8.i8(i8* %A, i8* %B, i8 -1, i32 0, i1 false) ret void ; CHECK: ret void @@ -40,7 +40,7 @@ define void @test3() { store i8 0, i8* %A ;; Written to by memset ; CHECK-NOT: store - call void @llvm.memset.i8(i8* %A, i8 0, i8 -1, i32 0) + call void @llvm.memset.p0i8.i8(i8* %A, i8 0, i8 -1, i32 0, i1 false) ret void ; CHECK: ret void diff --git a/test/Transforms/DeadStoreElimination/simple.ll b/test/Transforms/DeadStoreElimination/simple.ll index 23576da..5f143fc 100644 --- a/test/Transforms/DeadStoreElimination/simple.ll +++ b/test/Transforms/DeadStoreElimination/simple.ll @@ -3,8 +3,6 @@ target datalayout = "E-p:64:64:64-a0:0:8-f32:32:32-f64:64:64-i1:8:8-i8:8:8-i16:1 declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) nounwind declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture, i64, i32, i1) nounwind -declare void @llvm.memset.i64(i8*, i8, i64, i32) -declare void @llvm.memcpy.i64(i8*, i8*, i64, i32) declare i8* @llvm.init.trampoline(i8*, i8*, i8*) define void @test1(i32* %Q, i32* %P) { @@ -65,7 +63,7 @@ define void @test5(i32* %Q) { ; alias). define void @test6(i32 *%p, i8 *%q) { store i32 10, i32* %p, align 4 ;; dead. - call void @llvm.memset.i64(i8* %q, i8 42, i64 900, i32 1) + call void @llvm.memset.p0i8.i64(i8* %q, i8 42, i64 900, i32 1, i1 false) store i32 30, i32* %p, align 4 ret void ; CHECK: @test6 @@ -76,7 +74,7 @@ define void @test6(i32 *%p, i8 *%q) { ; alias). define void @test7(i32 *%p, i8 *%q, i8* noalias %r) { store i32 10, i32* %p, align 4 ;; dead. - call void @llvm.memcpy.i64(i8* %q, i8* %r, i64 900, i32 1) + call void @llvm.memcpy.p0i8.p0i8.i64(i8* %q, i8* %r, i64 900, i32 1, i1 false) store i32 30, i32* %p, align 4 ret void ; CHECK: @test7 @@ -184,8 +182,8 @@ define void @test14(i32* %Q) { ;; Fully dead overwrite of memcpy. define void @test15(i8* %P, i8* %Q) nounwind ssp { - tail call void @llvm.memcpy.i64(i8* %P, i8* %Q, i64 12, i32 1) - tail call void @llvm.memcpy.i64(i8* %P, i8* %Q, i64 12, i32 1) + tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* %P, i8* %Q, i64 12, i32 1, i1 false) + tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* %P, i8* %Q, i64 12, i32 1, i1 false) ret void ; CHECK: @test15 ; CHECK-NEXT: call void @llvm.memcpy @@ -194,8 +192,8 @@ define void @test15(i8* %P, i8* %Q) nounwind ssp { ;; Full overwrite of smaller memcpy. define void @test16(i8* %P, i8* %Q) nounwind ssp { - tail call void @llvm.memcpy.i64(i8* %P, i8* %Q, i64 8, i32 1) - tail call void @llvm.memcpy.i64(i8* %P, i8* %Q, i64 12, i32 1) + tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* %P, i8* %Q, i64 8, i32 1, i1 false) + tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* %P, i8* %Q, i64 12, i32 1, i1 false) ret void ; CHECK: @test16 ; CHECK-NEXT: call void @llvm.memcpy @@ -204,8 +202,8 @@ define void @test16(i8* %P, i8* %Q) nounwind ssp { ;; Overwrite of memset by memcpy. define void @test17(i8* %P, i8* noalias %Q) nounwind ssp { - tail call void @llvm.memset.i64(i8* %P, i8 42, i64 8, i32 1) - tail call void @llvm.memcpy.i64(i8* %P, i8* %Q, i64 12, i32 1) + tail call void @llvm.memset.p0i8.i64(i8* %P, i8 42, i64 8, i32 1, i1 false) + tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* %P, i8* %Q, i64 12, i32 1, i1 false) ret void ; CHECK: @test17 ; CHECK-NEXT: call void @llvm.memcpy @@ -228,8 +226,8 @@ define void @test17v(i8* %P, i8* %Q) nounwind ssp { ; A = B ; A = A define void @test18(i8* %P, i8* %Q, i8* %R) nounwind ssp { - tail call void @llvm.memcpy.i64(i8* %P, i8* %Q, i64 12, i32 1) - tail call void @llvm.memcpy.i64(i8* %P, i8* %R, i64 12, i32 1) + tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* %P, i8* %Q, i64 12, i32 1, i1 false) + tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* %P, i8* %R, i64 12, i32 1, i1 false) ret void ; CHECK: @test18 ; CHECK-NEXT: call void @llvm.memcpy diff --git a/test/Transforms/FunctionAttrs/2008-10-04-LocalMemory.ll b/test/Transforms/FunctionAttrs/2008-10-04-LocalMemory.ll deleted file mode 100644 index faac118..0000000 --- a/test/Transforms/FunctionAttrs/2008-10-04-LocalMemory.ll +++ /dev/null @@ -1,64 +0,0 @@ -; RUN: opt < %s -basicaa -functionattrs -S | FileCheck %s - -%struct.X = type { i32*, i32* } - -declare i32 @g(i32*) readnone - -define i32 @f() { -; CHECK: @f() readnone - %x = alloca i32 ; <i32*> [#uses=2] - store i32 0, i32* %x - %y = call i32 @g(i32* %x) ; <i32> [#uses=1] - ret i32 %y -} - -define i32 @foo() nounwind { -; CHECK: @foo() nounwind readonly -entry: - %y = alloca %struct.X ; <%struct.X*> [#uses=2] - %x = alloca %struct.X ; <%struct.X*> [#uses=2] - %j = alloca i32 ; <i32*> [#uses=2] - %i = alloca i32 ; <i32*> [#uses=2] - %"alloca point" = bitcast i32 0 to i32 ; <i32> [#uses=0] - store i32 0, i32* %i, align 4 - store i32 1, i32* %j, align 4 - %0 = getelementptr inbounds %struct.X* %y, i32 0, i32 0 ; <i32**> [#uses=1] - store i32* %i, i32** %0, align 8 - %1 = getelementptr inbounds %struct.X* %x, i32 0, i32 1 ; <i32**> [#uses=1] - store i32* %j, i32** %1, align 8 - %x1 = bitcast %struct.X* %x to i8* ; <i8*> [#uses=2] - %y2 = bitcast %struct.X* %y to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i64(i8* %x1, i8* %y2, i64 8, i32 1) - %2 = bitcast i8* %x1 to i32** ; <i32**> [#uses=1] - %3 = load i32** %2, align 8 ; <i32*> [#uses=1] - %4 = load i32* %3, align 4 ; <i32> [#uses=1] - br label %return - -return: ; preds = %entry - ret i32 %4 -} - -define i32 @t(i32 %a, i32 %b, i32 %c) nounwind { -; CHECK: @t(i32 %a, i32 %b, i32 %c) nounwind readnone -entry: - %a.addr = alloca i32 ; <i32*> [#uses=3] - %c.addr = alloca i32 ; <i32*> [#uses=2] - store i32 %a, i32* %a.addr - store i32 %c, i32* %c.addr - %tmp = load i32* %a.addr ; <i32> [#uses=1] - %tobool = icmp ne i32 %tmp, 0 ; <i1> [#uses=1] - br i1 %tobool, label %if.then, label %if.else - -if.then: ; preds = %entry - br label %if.end - -if.else: ; preds = %entry - br label %if.end - -if.end: ; preds = %if.else, %if.then - %p.0 = phi i32* [ %a.addr, %if.then ], [ %c.addr, %if.else ] ; <i32*> [#uses=1] - %tmp2 = load i32* %p.0 ; <i32> [#uses=1] - ret i32 %tmp2 -} - -declare void @llvm.memcpy.i64(i8* nocapture, i8* nocapture, i64, i32) nounwind diff --git a/test/Transforms/GVN/2008-02-24-NonDominatedMemcpy.ll b/test/Transforms/GVN/2008-02-24-NonDominatedMemcpy.ll deleted file mode 100644 index 9a75e1a..0000000 --- a/test/Transforms/GVN/2008-02-24-NonDominatedMemcpy.ll +++ /dev/null @@ -1,25 +0,0 @@ -; RUN: opt < %s -gvn -dse -S | grep {call.*memcpy} | count 1 - -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-s0:64:64-f80:128:128" -target triple = "x86_64-apple-darwin8" - %struct.ggFrame3 = type { %struct.ggPoint3, %struct.ggONB3 } - %struct.ggHMatrix3 = type { [4 x [4 x double]] } - %struct.ggONB3 = type { %struct.ggPoint3, %struct.ggPoint3, %struct.ggPoint3 } - %struct.ggPoint3 = type { [3 x double] } - %struct.ggQuaternion = type { [4 x double], i32, %struct.ggHMatrix3 } - -declare void @llvm.memcpy.i64(i8*, i8*, i64, i32) nounwind - -define void @_Z10ggCRSplineRK8ggFrame3S1_S1_S1_d(%struct.ggFrame3* noalias sret %agg.result, %struct.ggFrame3* %f0, %struct.ggFrame3* %f1, %struct.ggFrame3* %f2, %struct.ggFrame3* %f3, double %t) nounwind { -entry: - %qresult = alloca %struct.ggQuaternion ; <%struct.ggQuaternion*> [#uses=1] - %tmp = alloca %struct.ggONB3 ; <%struct.ggONB3*> [#uses=2] - call void @_ZN12ggQuaternion7getONB3Ev( %struct.ggONB3* noalias sret %tmp, %struct.ggQuaternion* %qresult ) nounwind - %tmp1.i = getelementptr %struct.ggFrame3* %agg.result, i32 0, i32 1 ; <%struct.ggONB3*> [#uses=1] - %tmp13.i = bitcast %struct.ggONB3* %tmp1.i to i8* ; <i8*> [#uses=1] - %tmp24.i = bitcast %struct.ggONB3* %tmp to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i64( i8* %tmp13.i, i8* %tmp24.i, i64 72, i32 8 ) nounwind - ret void -} - -declare void @_ZN12ggQuaternion7getONB3Ev(%struct.ggONB3* noalias sret , %struct.ggQuaternion*) nounwind diff --git a/test/Transforms/GVN/2008-02-26-MemCpySize.ll b/test/Transforms/GVN/2008-02-26-MemCpySize.ll deleted file mode 100644 index 6ed8a76..0000000 --- a/test/Transforms/GVN/2008-02-26-MemCpySize.ll +++ /dev/null @@ -1,46 +0,0 @@ -; RUN: opt < %s -gvn -dse -S | grep {call.*memcpy.*cell} | count 2 -; PR2099 - -target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" -target triple = "i686-apple-darwin9" - %struct.s = type { [11 x i8], i32 } -@.str = internal constant [11 x i8] c"0123456789\00" ; <[11 x i8]*> [#uses=1] -@cell = weak global %struct.s zeroinitializer ; <%struct.s*> [#uses=2] - -declare i32 @check(%struct.s* byval %p) nounwind - -declare i32 @strcmp(i8*, i8*) nounwind readonly - -define i32 @main() noreturn nounwind { -entry: - %p = alloca %struct.s, align 8 ; <%struct.s*> [#uses=2] - store i32 99, i32* getelementptr (%struct.s* @cell, i32 0, i32 1), align 4 - call void @llvm.memcpy.i32( i8* getelementptr (%struct.s* @cell, i32 0, i32 0, i32 0), i8* getelementptr ([11 x i8]* @.str, i32 0, i32 0), i32 11, i32 1 ) - %tmp = getelementptr %struct.s* %p, i32 0, i32 0, i32 0 ; <i8*> [#uses=2] - call void @llvm.memcpy.i64( i8* %tmp, i8* getelementptr (%struct.s* @cell, i32 0, i32 0, i32 0), i64 16, i32 8 ) - %tmp1.i = getelementptr %struct.s* %p, i32 0, i32 1 ; <i32*> [#uses=1] - %tmp2.i = load i32* %tmp1.i, align 4 ; <i32> [#uses=1] - %tmp3.i = icmp eq i32 %tmp2.i, 99 ; <i1> [#uses=1] - br i1 %tmp3.i, label %bb5.i, label %bb - -bb5.i: ; preds = %entry - %tmp91.i = call i32 @strcmp( i8* %tmp, i8* getelementptr ([11 x i8]* @.str, i32 0, i32 0) ) nounwind readonly ; <i32> [#uses=1] - %tmp53 = icmp eq i32 %tmp91.i, 0 ; <i1> [#uses=1] - br i1 %tmp53, label %bb7, label %bb - -bb: ; preds = %bb5.i, %entry - call void @abort( ) noreturn nounwind - unreachable - -bb7: ; preds = %bb5.i - call void @exit( i32 0 ) noreturn nounwind - unreachable -} - -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) nounwind - -declare void @abort() noreturn nounwind - -declare void @exit(i32) noreturn nounwind - -declare void @llvm.memcpy.i64(i8*, i8*, i64, i32) nounwind diff --git a/test/Transforms/GVN/nonescaping-malloc.ll b/test/Transforms/GVN/nonescaping-malloc.ll index 1d50205..dba9d81 100644 --- a/test/Transforms/GVN/nonescaping-malloc.ll +++ b/test/Transforms/GVN/nonescaping-malloc.ll @@ -20,13 +20,13 @@ target triple = "x86_64-apple-darwin10.0" define %"struct.llvm::StringMapEntry<void*>"* @_Z3fooRN4llvm9StringMapIPvNS_15MallocAllocatorEEEPKc(%"struct.llvm::StringMap<void*,llvm::MallocAllocator>"* %X, i8* %P) ssp { entry: - %tmp = alloca %"struct.llvm::StringRef", align 8 ; <%"struct.llvm::StringRef"*> [#uses=3] - %tmp.i = getelementptr inbounds %"struct.llvm::StringRef"* %tmp, i64 0, i32 0 ; <i8**> [#uses=1] + %tmp = alloca %"struct.llvm::StringRef", align 8 + %tmp.i = getelementptr inbounds %"struct.llvm::StringRef"* %tmp, i64 0, i32 0 store i8* %P, i8** %tmp.i, align 8 - %tmp1.i = call i64 @strlen(i8* %P) nounwind readonly ; <i64> [#uses=1] - %tmp2.i = getelementptr inbounds %"struct.llvm::StringRef"* %tmp, i64 0, i32 1 ; <i64*> [#uses=1] + %tmp1.i = call i64 @strlen(i8* %P) nounwind readonly + %tmp2.i = getelementptr inbounds %"struct.llvm::StringRef"* %tmp, i64 0, i32 1 store i64 %tmp1.i, i64* %tmp2.i, align 8 - %tmp1 = call %"struct.llvm::StringMapEntry<void*>"* @_ZN4llvm9StringMapIPvNS_15MallocAllocatorEE16GetOrCreateValueERKNS_9StringRefE(%"struct.llvm::StringMap<void*,llvm::MallocAllocator>"* %X, %"struct.llvm::StringRef"* %tmp) ssp ; <%"struct.llvm::StringMapEntry<void*>"*> [#uses=1] + %tmp1 = call %"struct.llvm::StringMapEntry<void*>"* @_ZN4llvm9StringMapIPvNS_15MallocAllocatorEE16GetOrCreateValueERKNS_9StringRefE(%"struct.llvm::StringMap<void*,llvm::MallocAllocator>"* %X, %"struct.llvm::StringRef"* %tmp) ssp ret %"struct.llvm::StringMapEntry<void*>"* %tmp1 } @@ -34,75 +34,75 @@ declare i64 @strlen(i8* nocapture) nounwind readonly declare noalias i8* @malloc(i64) nounwind -declare void @llvm.memcpy.i64(i8* nocapture, i8* nocapture, i64, i32) nounwind - declare i32 @_ZN4llvm13StringMapImpl15LookupBucketForENS_9StringRefE(%"struct.llvm::StringMapImpl"*, i64, i64) define linkonce_odr %"struct.llvm::StringMapEntry<void*>"* @_ZN4llvm9StringMapIPvNS_15MallocAllocatorEE16GetOrCreateValueERKNS_9StringRefE(%"struct.llvm::StringMap<void*,llvm::MallocAllocator>"* %this, %"struct.llvm::StringRef"* nocapture %Key) ssp align 2 { entry: - %elt = bitcast %"struct.llvm::StringRef"* %Key to i64* ; <i64*> [#uses=1] - %val = load i64* %elt ; <i64> [#uses=3] - %tmp = getelementptr inbounds %"struct.llvm::StringRef"* %Key, i64 0, i32 1 ; <i64*> [#uses=1] - %val2 = load i64* %tmp ; <i64> [#uses=2] - %tmp2.i = getelementptr inbounds %"struct.llvm::StringMap<void*,llvm::MallocAllocator>"* %this, i64 0, i32 0 ; <%"struct.llvm::StringMapImpl"*> [#uses=1] - %tmp3.i = tail call i32 @_ZN4llvm13StringMapImpl15LookupBucketForENS_9StringRefE(%"struct.llvm::StringMapImpl"* %tmp2.i, i64 %val, i64 %val2) ; <i32> [#uses=1] - %tmp4.i = getelementptr inbounds %"struct.llvm::StringMap<void*,llvm::MallocAllocator>"* %this, i64 0, i32 0, i32 0 ; <%"struct.llvm::StringMapImpl::ItemBucket"**> [#uses=1] - %tmp5.i = load %"struct.llvm::StringMapImpl::ItemBucket"** %tmp4.i, align 8 ; <%"struct.llvm::StringMapImpl::ItemBucket"*> [#uses=1] - %tmp6.i = zext i32 %tmp3.i to i64 ; <i64> [#uses=1] - %tmp7.i = getelementptr inbounds %"struct.llvm::StringMapImpl::ItemBucket"* %tmp5.i, i64 %tmp6.i, i32 1 ; <%"struct.llvm::StringMapEntryBase"**> [#uses=2] - %tmp8.i = load %"struct.llvm::StringMapEntryBase"** %tmp7.i, align 8 ; <%"struct.llvm::StringMapEntryBase"*> [#uses=3] - %tmp9.i = icmp eq %"struct.llvm::StringMapEntryBase"* %tmp8.i, null ; <i1> [#uses=1] - %tmp13.i = icmp eq %"struct.llvm::StringMapEntryBase"* %tmp8.i, inttoptr (i64 -1 to %"struct.llvm::StringMapEntryBase"*) ; <i1> [#uses=1] - %or.cond.i = or i1 %tmp9.i, %tmp13.i ; <i1> [#uses=1] + %elt = bitcast %"struct.llvm::StringRef"* %Key to i64* + %val = load i64* %elt + %tmp = getelementptr inbounds %"struct.llvm::StringRef"* %Key, i64 0, i32 1 + %val2 = load i64* %tmp + %tmp2.i = getelementptr inbounds %"struct.llvm::StringMap<void*,llvm::MallocAllocator>"* %this, i64 0, i32 0 + %tmp3.i = tail call i32 @_ZN4llvm13StringMapImpl15LookupBucketForENS_9StringRefE(%"struct.llvm::StringMapImpl"* %tmp2.i, i64 %val, i64 %val2) + %tmp4.i = getelementptr inbounds %"struct.llvm::StringMap<void*,llvm::MallocAllocator>"* %this, i64 0, i32 0, i32 0 + %tmp5.i = load %"struct.llvm::StringMapImpl::ItemBucket"** %tmp4.i, align 8 + %tmp6.i = zext i32 %tmp3.i to i64 + %tmp7.i = getelementptr inbounds %"struct.llvm::StringMapImpl::ItemBucket"* %tmp5.i, i64 %tmp6.i, i32 1 + %tmp8.i = load %"struct.llvm::StringMapEntryBase"** %tmp7.i, align 8 + %tmp9.i = icmp eq %"struct.llvm::StringMapEntryBase"* %tmp8.i, null + %tmp13.i = icmp eq %"struct.llvm::StringMapEntryBase"* %tmp8.i, inttoptr (i64 -1 to %"struct.llvm::StringMapEntryBase"*) + %or.cond.i = or i1 %tmp9.i, %tmp13.i br i1 %or.cond.i, label %bb4.i, label %bb6.i bb4.i: ; preds = %entry - %tmp41.i = inttoptr i64 %val to i8* ; <i8*> [#uses=2] - %tmp4.i35.i = getelementptr inbounds i8* %tmp41.i, i64 %val2 ; <i8*> [#uses=1] - %tmp.i.i = ptrtoint i8* %tmp4.i35.i to i64 ; <i64> [#uses=1] - %tmp1.i.i = trunc i64 %tmp.i.i to i32 ; <i32> [#uses=1] - %tmp3.i.i = trunc i64 %val to i32 ; <i32> [#uses=1] - %tmp4.i.i = sub i32 %tmp1.i.i, %tmp3.i.i ; <i32> [#uses=3] - %tmp5.i.i = add i32 %tmp4.i.i, 17 ; <i32> [#uses=1] - %tmp8.i.i = zext i32 %tmp5.i.i to i64 ; <i64> [#uses=1] - %tmp.i20.i.i = tail call noalias i8* @malloc(i64 %tmp8.i.i) nounwind ; <i8*> [#uses=7] - %tmp10.i.i = bitcast i8* %tmp.i20.i.i to %"struct.llvm::StringMapEntry<void*>"* ; <%"struct.llvm::StringMapEntry<void*>"*> [#uses=2] - %tmp12.i.i = icmp eq i8* %tmp.i20.i.i, null ; <i1> [#uses=1] + %tmp41.i = inttoptr i64 %val to i8* + %tmp4.i35.i = getelementptr inbounds i8* %tmp41.i, i64 %val2 + %tmp.i.i = ptrtoint i8* %tmp4.i35.i to i64 + %tmp1.i.i = trunc i64 %tmp.i.i to i32 + %tmp3.i.i = trunc i64 %val to i32 + %tmp4.i.i = sub i32 %tmp1.i.i, %tmp3.i.i + %tmp5.i.i = add i32 %tmp4.i.i, 17 + %tmp8.i.i = zext i32 %tmp5.i.i to i64 + %tmp.i20.i.i = tail call noalias i8* @malloc(i64 %tmp8.i.i) nounwind + %tmp10.i.i = bitcast i8* %tmp.i20.i.i to %"struct.llvm::StringMapEntry<void*>"* + %tmp12.i.i = icmp eq i8* %tmp.i20.i.i, null br i1 %tmp12.i.i, label %_ZN4llvm14StringMapEntryIPvE6CreateINS_15MallocAllocatorES1_EEPS2_PKcS7_RT_T0_.exit.i, label %bb.i.i bb.i.i: ; preds = %bb4.i - %tmp.i.i.i.i = bitcast i8* %tmp.i20.i.i to i32* ; <i32*> [#uses=1] + %tmp.i.i.i.i = bitcast i8* %tmp.i20.i.i to i32* store i32 %tmp4.i.i, i32* %tmp.i.i.i.i, align 4 - %tmp1.i19.i.i = getelementptr inbounds i8* %tmp.i20.i.i, i64 8 ; <i8*> [#uses=1] - %0 = bitcast i8* %tmp1.i19.i.i to i8** ; <i8**> [#uses=1] + %tmp1.i19.i.i = getelementptr inbounds i8* %tmp.i20.i.i, i64 8 + %0 = bitcast i8* %tmp1.i19.i.i to i8** store i8* null, i8** %0, align 8 br label %_ZN4llvm14StringMapEntryIPvE6CreateINS_15MallocAllocatorES1_EEPS2_PKcS7_RT_T0_.exit.i -_ZN4llvm14StringMapEntryIPvE6CreateINS_15MallocAllocatorES1_EEPS2_PKcS7_RT_T0_.exit.i: ; preds = %bb4.i, %bb.i.i - %tmp.i18.i.i = getelementptr inbounds i8* %tmp.i20.i.i, i64 16 ; <i8*> [#uses=1] - %tmp15.i.i = zext i32 %tmp4.i.i to i64 ; <i64> [#uses=2] - tail call void @llvm.memcpy.i64(i8* %tmp.i18.i.i, i8* %tmp41.i, i64 %tmp15.i.i, i32 1) nounwind - %tmp.i18.sum.i.i = add i64 %tmp15.i.i, 16 ; <i64> [#uses=1] - %tmp17.i.i = getelementptr inbounds i8* %tmp.i20.i.i, i64 %tmp.i18.sum.i.i ; <i8*> [#uses=1] +_ZN4llvm14StringMapEntryIPvE6CreateINS_15MallocAllocatorES1_EEPS2_PKcS7_RT_T0_.exit.i: ; preds = %bb.i.i, %bb4.i + %tmp.i18.i.i = getelementptr inbounds i8* %tmp.i20.i.i, i64 16 + %tmp15.i.i = zext i32 %tmp4.i.i to i64 + tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* %tmp.i18.i.i, i8* %tmp41.i, i64 %tmp15.i.i, i32 1, i1 false) + %tmp.i18.sum.i.i = add i64 %tmp15.i.i, 16 + %tmp17.i.i = getelementptr inbounds i8* %tmp.i20.i.i, i64 %tmp.i18.sum.i.i store i8 0, i8* %tmp17.i.i, align 1 - %tmp.i.i.i = getelementptr inbounds i8* %tmp.i20.i.i, i64 8 ; <i8*> [#uses=1] - %1 = bitcast i8* %tmp.i.i.i to i8** ; <i8**> [#uses=1] + %tmp.i.i.i = getelementptr inbounds i8* %tmp.i20.i.i, i64 8 + %1 = bitcast i8* %tmp.i.i.i to i8** store i8* null, i8** %1, align 8 - %tmp22.i = load %"struct.llvm::StringMapEntryBase"** %tmp7.i, align 8 ; <%"struct.llvm::StringMapEntryBase"*> [#uses=1] - %tmp24.i = icmp eq %"struct.llvm::StringMapEntryBase"* %tmp22.i, inttoptr (i64 -1 to %"struct.llvm::StringMapEntryBase"*) ; <i1> [#uses=1] + %tmp22.i = load %"struct.llvm::StringMapEntryBase"** %tmp7.i, align 8 + %tmp24.i = icmp eq %"struct.llvm::StringMapEntryBase"* %tmp22.i, inttoptr (i64 -1 to %"struct.llvm::StringMapEntryBase"*) br i1 %tmp24.i, label %bb9.i, label %_ZN4llvm9StringMapIPvNS_15MallocAllocatorEE16GetOrCreateValueIS1_EERNS_14StringMapEntryIS1_EENS_9StringRefET_.exit bb6.i: ; preds = %entry - %tmp16.i = bitcast %"struct.llvm::StringMapEntryBase"* %tmp8.i to %"struct.llvm::StringMapEntry<void*>"* ; <%"struct.llvm::StringMapEntry<void*>"*> [#uses=1] + %tmp16.i = bitcast %"struct.llvm::StringMapEntryBase"* %tmp8.i to %"struct.llvm::StringMapEntry<void*>"* ret %"struct.llvm::StringMapEntry<void*>"* %tmp16.i bb9.i: ; preds = %_ZN4llvm14StringMapEntryIPvE6CreateINS_15MallocAllocatorES1_EEPS2_PKcS7_RT_T0_.exit.i - %tmp25.i = getelementptr inbounds %"struct.llvm::StringMap<void*,llvm::MallocAllocator>"* %this, i64 0, i32 0, i32 3 ; <i32*> [#uses=2] - %tmp26.i = load i32* %tmp25.i, align 8 ; <i32> [#uses=1] - %tmp27.i = add i32 %tmp26.i, -1 ; <i32> [#uses=1] + %tmp25.i = getelementptr inbounds %"struct.llvm::StringMap<void*,llvm::MallocAllocator>"* %this, i64 0, i32 0, i32 3 + %tmp26.i = load i32* %tmp25.i, align 8 + %tmp27.i = add i32 %tmp26.i, -1 store i32 %tmp27.i, i32* %tmp25.i, align 8 ret %"struct.llvm::StringMapEntry<void*>"* %tmp10.i.i _ZN4llvm9StringMapIPvNS_15MallocAllocatorEE16GetOrCreateValueIS1_EERNS_14StringMapEntryIS1_EENS_9StringRefET_.exit: ; preds = %_ZN4llvm14StringMapEntryIPvE6CreateINS_15MallocAllocatorES1_EEPS2_PKcS7_RT_T0_.exit.i ret %"struct.llvm::StringMapEntry<void*>"* %tmp10.i.i } + +declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture, i64, i32, i1) nounwind diff --git a/test/Transforms/GVN/rle.ll b/test/Transforms/GVN/rle.ll index 28b1fc7..2f0d2eb 100644 --- a/test/Transforms/GVN/rle.ll +++ b/test/Transforms/GVN/rle.ll @@ -135,7 +135,7 @@ define i8* @coerce_mustalias7(i64 %V, i64* %P) { define signext i16 @memset_to_i16_local(i16* %A) nounwind ssp { entry: %conv = bitcast i16* %A to i8* - tail call void @llvm.memset.i64(i8* %conv, i8 1, i64 200, i32 1) + tail call void @llvm.memset.p0i8.i64(i8* %conv, i8 1, i64 200, i32 1, i1 false) %arrayidx = getelementptr inbounds i16* %A, i64 42 %tmp2 = load i16* %arrayidx ret i16 %tmp2 @@ -148,7 +148,7 @@ entry: define float @memset_to_float_local(float* %A, i8 %Val) nounwind ssp { entry: %conv = bitcast float* %A to i8* ; <i8*> [#uses=1] - tail call void @llvm.memset.i64(i8* %conv, i8 %Val, i64 400, i32 1) + tail call void @llvm.memset.p0i8.i64(i8* %conv, i8 %Val, i64 400, i32 1, i1 false) %arrayidx = getelementptr inbounds float* %A, i64 42 ; <float*> [#uses=1] %tmp2 = load float* %arrayidx ; <float> [#uses=1] ret float %tmp2 @@ -168,11 +168,11 @@ define i16 @memset_to_i16_nonlocal0(i16* %P, i1 %cond) { %P3 = bitcast i16* %P to i8* br i1 %cond, label %T, label %F T: - tail call void @llvm.memset.i64(i8* %P3, i8 1, i64 400, i32 1) + tail call void @llvm.memset.p0i8.i64(i8* %P3, i8 1, i64 400, i32 1, i1 false) br label %Cont F: - tail call void @llvm.memset.i64(i8* %P3, i8 2, i64 400, i32 1) + tail call void @llvm.memset.p0i8.i64(i8* %P3, i8 2, i64 400, i32 1, i1 false) br label %Cont Cont: @@ -193,7 +193,7 @@ Cont: define float @memcpy_to_float_local(float* %A) nounwind ssp { entry: %conv = bitcast float* %A to i8* ; <i8*> [#uses=1] - tail call void @llvm.memcpy.i64(i8* %conv, i8* bitcast ({i32, float, i32 }* @GCst to i8*), i64 12, i32 1) + tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* %conv, i8* bitcast ({i32, float, i32 }* @GCst to i8*), i64 12, i32 1, i1 false) %arrayidx = getelementptr inbounds float* %A, i64 1 ; <float*> [#uses=1] %tmp2 = load float* %arrayidx ; <float> [#uses=1] ret float %tmp2 @@ -203,11 +203,6 @@ entry: } -declare void @llvm.memset.i64(i8* nocapture, i8, i64, i32) nounwind -declare void @llvm.memcpy.i64(i8* nocapture, i8* nocapture, i64, i32) nounwind - - - ;; non-local i32/float -> i8 load forwarding. define i8 @coerce_mustalias_nonlocal0(i32* %P, i1 %cond) { @@ -539,7 +534,7 @@ define i32 @memset_to_load() nounwind readnone { entry: %x = alloca [256 x i32], align 4 ; <[256 x i32]*> [#uses=2] %tmp = bitcast [256 x i32]* %x to i8* ; <i8*> [#uses=1] - call void @llvm.memset.i64(i8* %tmp, i8 0, i64 1024, i32 4) + call void @llvm.memset.p0i8.i64(i8* %tmp, i8 0, i64 1024, i32 4, i1 false) %arraydecay = getelementptr inbounds [256 x i32]* %x, i32 0, i32 0 ; <i32*> %tmp1 = load i32* %arraydecay ; <i32> [#uses=1] ret i32 %tmp1 @@ -643,3 +638,7 @@ entry: ; CHECK-ret i32 } +declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) nounwind + +declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture, i64, i32, i1) nounwind + diff --git a/test/Transforms/GlobalOpt/memcpy.ll b/test/Transforms/GlobalOpt/memcpy.ll index 8f063a2..94e07a0 100644 --- a/test/Transforms/GlobalOpt/memcpy.ll +++ b/test/Transforms/GlobalOpt/memcpy.ll @@ -3,13 +3,11 @@ @G1 = internal global [58 x i8] c"asdlfkajsdlfkajsd;lfkajds;lfkjasd;flkajsd;lkfja;sdlkfjasd\00" ; <[58 x i8]*> [#uses=1] -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) - define void @foo() { - %Blah = alloca [58 x i8] ; <[58 x i8]*> [#uses=1] - %tmp.0 = getelementptr [58 x i8]* %Blah, i32 0, i32 0 ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %tmp.0, i8* getelementptr ([58 x i8]* @G1, i32 0, i32 0), i32 58, i32 1 ) - ret void + %Blah = alloca [58 x i8] + %tmp.0 = getelementptr [58 x i8]* %Blah, i32 0, i32 0 + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp.0, i8* getelementptr inbounds ([58 x i8]* @G1, i32 0, i32 0), i32 58, i32 1, i1 false) + ret void } - +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/GlobalOpt/memset.ll b/test/Transforms/GlobalOpt/memset.ll index a9b9d5e..3bb5ce9 100644 --- a/test/Transforms/GlobalOpt/memset.ll +++ b/test/Transforms/GlobalOpt/memset.ll @@ -1,21 +1,18 @@ ; both globals are write only, delete them. -; RUN: opt < %s -globalopt -S | \ -; RUN: not grep internal +; RUN: opt < %s -globalopt -S | not grep internal @G0 = internal global [58 x i8] c"asdlfkajsdlfkajsd;lfkajds;lfkjasd;flkajsd;lkfja;sdlkfjasd\00" ; <[58 x i8]*> [#uses=1] @G1 = internal global [4 x i32] [ i32 1, i32 2, i32 3, i32 4 ] ; <[4 x i32]*> [#uses=1] -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) - -declare void @llvm.memset.i32(i8*, i8, i32, i32) - define void @foo() { - %Blah = alloca [58 x i8] ; <[58 x i8]*> [#uses=1] - %tmp3 = bitcast [58 x i8]* %Blah to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* bitcast ([4 x i32]* @G1 to i8*), i8* %tmp3, i32 16, i32 1 ) - call void @llvm.memset.i32( i8* getelementptr ([58 x i8]* @G0, i32 0, i32 0), i8 17, i32 58, i32 1 ) - ret void + %Blah = alloca [58 x i8] + %tmp3 = bitcast [58 x i8]* %Blah to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* bitcast ([4 x i32]* @G1 to i8*), i8* %tmp3, i32 16, i32 1, i1 false) + call void @llvm.memset.p0i8.i32(i8* getelementptr inbounds ([58 x i8]* @G0, i32 0, i32 0), i8 17, i32 58, i32 1, i1 false) + ret void } +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind +declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1) nounwind diff --git a/test/Transforms/Inline/inline-invoke-tail.ll b/test/Transforms/Inline/inline-invoke-tail.ll index 961f678..462c29a 100644 --- a/test/Transforms/Inline/inline-invoke-tail.ll +++ b/test/Transforms/Inline/inline-invoke-tail.ll @@ -4,7 +4,7 @@ define internal void @foo(i32* %p, i32* %q) { %pp = bitcast i32* %p to i8* %qq = bitcast i32* %q to i8* - tail call void @llvm.memcpy.i32(i8* %pp, i8* %qq, i32 4, i32 1) + tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* %pp, i8* %qq, i32 4, i32 1, i1 false) ret void } @@ -24,12 +24,14 @@ invcont: lpad: %eh_ptr = call i8* @llvm.eh.exception() - %eh_select = call i32 (i8*, i8*, ...)* @llvm.eh.selector.i32(i8* %eh_ptr, i8* bitcast (i32 (...)* @__gxx_personality_v0 to i8*), i8* null) + %eh_select = call i32 (i8*, i8*, ...)* @llvm.eh.selector(i8* %eh_ptr, i8* bitcast (i32 (...)* @__gxx_personality_v0 to i8*), i8* null) unreachable } -declare i8* @llvm.eh.exception() nounwind +declare i8* @llvm.eh.exception() nounwind readonly -declare i32 @llvm.eh.selector.i32(i8*, i8*, ...) nounwind +declare i32 @llvm.eh.selector(i8*, i8*, ...) nounwind declare i32 @__gxx_personality_v0(...) + +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/InstCombine/2006-11-03-Memmove64.ll b/test/Transforms/InstCombine/2006-11-03-Memmove64.ll deleted file mode 100644 index 35bb45e..0000000 --- a/test/Transforms/InstCombine/2006-11-03-Memmove64.ll +++ /dev/null @@ -1,16 +0,0 @@ -; RUN: opt < %s -instcombine -S | \ -; RUN: not grep memmove.i32 -; Instcombine was trying to turn this into a memmove.i32 - -target datalayout = "e-p:64:64" -target triple = "alphaev67-unknown-linux-gnu" -@str10 = internal constant [1 x i8] zeroinitializer ; <[1 x i8]*> [#uses=1] - -define void @do_join(i8* %b) { -entry: - call void @llvm.memmove.i64( i8* %b, i8* getelementptr ([1 x i8]* @str10, i32 0, i64 0), i64 1, i32 1 ) - ret void -} - -declare void @llvm.memmove.i64(i8*, i8*, i64, i32) - diff --git a/test/Transforms/InstCombine/2007-05-04-Crash.ll b/test/Transforms/InstCombine/2007-05-04-Crash.ll deleted file mode 100644 index 9f50d8a..0000000 --- a/test/Transforms/InstCombine/2007-05-04-Crash.ll +++ /dev/null @@ -1,30 +0,0 @@ -; RUN: opt < %s -instcombine -disable-output -; PR1384 - -target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64" -target triple = "i686-apple-darwin8" - %struct.CFRuntimeBase = type { i32, [4 x i8] } - %struct.CGColor = type opaque - %struct.CGColorSpace = type { %struct.CFRuntimeBase, i8, i8, i8, i32, i32, i32, %struct.CGColor*, float*, %struct.CGMD5Signature, %struct.CGMD5Signature*, [0 x %struct.CGColorSpaceDescriptor] } - %struct.CGColorSpaceCalibratedRGBData = type { [3 x float], [3 x float], [3 x float], [9 x float] } - %struct.CGColorSpaceDescriptor = type { %struct.CGColorSpaceCalibratedRGBData } - %struct.CGColorSpaceLabData = type { [3 x float], [3 x float], [4 x float] } - %struct.CGMD5Signature = type { [16 x i8], i8 } - -declare fastcc %struct.CGColorSpace* @CGColorSpaceCreate(i32, i32) - -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) - -define %struct.CGColorSpace* @CGColorSpaceCreateLab(float* %whitePoint, float* %blackPoint, float* %range) { -entry: - %tmp17 = call fastcc %struct.CGColorSpace* @CGColorSpaceCreate( i32 5, i32 3 ) ; <%struct.CGColorSpace*> [#uses=2] - %tmp28 = getelementptr %struct.CGColorSpace* %tmp17, i32 0, i32 11 ; <[0 x %struct.CGColorSpaceDescriptor]*> [#uses=1] - %tmp29 = getelementptr [0 x %struct.CGColorSpaceDescriptor]* %tmp28, i32 0, i32 0 ; <%struct.CGColorSpaceDescriptor*> [#uses=1] - %tmp30 = getelementptr %struct.CGColorSpaceDescriptor* %tmp29, i32 0, i32 0 ; <%struct.CGColorSpaceCalibratedRGBData*> [#uses=1] - %tmp3031 = bitcast %struct.CGColorSpaceCalibratedRGBData* %tmp30 to %struct.CGColorSpaceLabData* ; <%struct.CGColorSpaceLabData*> [#uses=1] - %tmp45 = getelementptr %struct.CGColorSpaceLabData* %tmp3031, i32 0, i32 2 ; <[4 x float]*> [#uses=1] - %tmp46 = getelementptr [4 x float]* %tmp45, i32 0, i32 0 ; <float*> [#uses=1] - %tmp4648 = bitcast float* %tmp46 to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %tmp4648, i8* null, i32 16, i32 4 ) - ret %struct.CGColorSpace* %tmp17 -} diff --git a/test/Transforms/InstCombine/2007-10-10-EliminateMemCpy.ll b/test/Transforms/InstCombine/2007-10-10-EliminateMemCpy.ll index 710aff2..fe935f9 100644 --- a/test/Transforms/InstCombine/2007-10-10-EliminateMemCpy.ll +++ b/test/Transforms/InstCombine/2007-10-10-EliminateMemCpy.ll @@ -6,16 +6,15 @@ target datalayout = "E-p:64:64:64-a0:0:8-f32:32:32-f64:64:64-i1:8:8-i8:8:8-i16:1 define void @foo(i8* %P) { entry: - %P_addr = alloca i8* ; <i8**> [#uses=2] - %"alloca point" = bitcast i32 0 to i32 ; <i32> [#uses=0] - store i8* %P, i8** %P_addr - %tmp = load i8** %P_addr, align 4 ; <i8*> [#uses=1] - %tmp1 = getelementptr [4 x i8]* @.str, i32 0, i32 0 ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %tmp, i8* %tmp1, i32 4, i32 1 ) - br label %return + %P_addr = alloca i8* + store i8* %P, i8** %P_addr + %tmp = load i8** %P_addr, align 4 + %tmp1 = getelementptr [4 x i8]* @.str, i32 0, i32 0 + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp, i8* %tmp1, i32 4, i32 1, i1 false) + br label %return -return: ; preds = %entry - ret void +return: ; preds = %entry + ret void } -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/InstCombine/2009-02-20-InstCombine-SROA.ll b/test/Transforms/InstCombine/2009-02-20-InstCombine-SROA.ll index b29d8d2..a51c47d 100644 --- a/test/Transforms/InstCombine/2009-02-20-InstCombine-SROA.ll +++ b/test/Transforms/InstCombine/2009-02-20-InstCombine-SROA.ll @@ -4,275 +4,276 @@ target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" target triple = "i386-apple-darwin9.6" - %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >" = type { i32* } - %"struct.std::_Vector_base<int,std::allocator<int> >" = type { %"struct.std::_Vector_base<int,std::allocator<int> >::_Vector_impl" } - %"struct.std::_Vector_base<int,std::allocator<int> >::_Vector_impl" = type { i32*, i32*, i32* } - %"struct.std::bidirectional_iterator_tag" = type <{ i8 }> - %"struct.std::forward_iterator_tag" = type <{ i8 }> - %"struct.std::input_iterator_tag" = type <{ i8 }> - %"struct.std::random_access_iterator_tag" = type <{ i8 }> - %"struct.std::vector<int,std::allocator<int> >" = type { %"struct.std::_Vector_base<int,std::allocator<int> >" } + +%"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >" = type { i32* } +%"struct.std::_Vector_base<int,std::allocator<int> >" = type { %"struct.std::_Vector_base<int,std::allocator<int> >::_Vector_impl" } +%"struct.std::_Vector_base<int,std::allocator<int> >::_Vector_impl" = type { i32*, i32*, i32* } +%"struct.std::bidirectional_iterator_tag" = type <{ i8 }> +%"struct.std::forward_iterator_tag" = type <{ i8 }> +%"struct.std::input_iterator_tag" = type <{ i8 }> +%"struct.std::random_access_iterator_tag" = type <{ i8 }> +%"struct.std::vector<int,std::allocator<int> >" = type { %"struct.std::_Vector_base<int,std::allocator<int> >" } define i32* @_Z3fooRSt6vectorIiSaIiEE(%"struct.std::vector<int,std::allocator<int> >"* %X) { entry: - %0 = alloca %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >" ; <%"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"*> [#uses=2] - %__first_addr.i.i = alloca %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >" ; <%"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"*> [#uses=31] - %__last_addr.i.i = alloca %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >" ; <%"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"*> [#uses=4] - %unnamed_arg.i = alloca %"struct.std::bidirectional_iterator_tag", align 8 ; <%"struct.std::bidirectional_iterator_tag"*> [#uses=1] - %1 = alloca %"struct.std::bidirectional_iterator_tag" ; <%"struct.std::bidirectional_iterator_tag"*> [#uses=1] - %__first_addr.i = alloca %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >" ; <%"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"*> [#uses=2] - %2 = alloca %"struct.std::bidirectional_iterator_tag" ; <%"struct.std::bidirectional_iterator_tag"*> [#uses=2] - %3 = alloca %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >" ; <%"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"*> [#uses=2] - %4 = alloca i32 ; <i32*> [#uses=8] - %"alloca point" = bitcast i32 0 to i32 ; <i32> [#uses=0] - store i32 42, i32* %4, align 4 - %5 = getelementptr %"struct.std::vector<int,std::allocator<int> >"* %X, i32 0, i32 0 ; <%"struct.std::_Vector_base<int,std::allocator<int> >"*> [#uses=1] - %6 = getelementptr %"struct.std::_Vector_base<int,std::allocator<int> >"* %5, i32 0, i32 0 ; <%"struct.std::_Vector_base<int,std::allocator<int> >::_Vector_impl"*> [#uses=1] - %7 = getelementptr %"struct.std::_Vector_base<int,std::allocator<int> >::_Vector_impl"* %6, i32 0, i32 1 ; <i32**> [#uses=1] - %8 = load i32** %7, align 4 ; <i32*> [#uses=1] - %9 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %3, i32 0, i32 0 ; <i32**> [#uses=1] - store i32* %8, i32** %9, align 4 - %10 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %3, i32 0, i32 0 ; <i32**> [#uses=1] - %11 = load i32** %10, align 4 ; <i32*> [#uses=1] - %tmp2.i = ptrtoint i32* %11 to i32 ; <i32> [#uses=1] - %tmp1.i = inttoptr i32 %tmp2.i to i32* ; <i32*> [#uses=1] - %tmp3 = ptrtoint i32* %tmp1.i to i32 ; <i32> [#uses=1] - %tmp2 = inttoptr i32 %tmp3 to i32* ; <i32*> [#uses=1] - %12 = getelementptr %"struct.std::vector<int,std::allocator<int> >"* %X, i32 0, i32 0 ; <%"struct.std::_Vector_base<int,std::allocator<int> >"*> [#uses=1] - %13 = getelementptr %"struct.std::_Vector_base<int,std::allocator<int> >"* %12, i32 0, i32 0 ; <%"struct.std::_Vector_base<int,std::allocator<int> >::_Vector_impl"*> [#uses=1] - %14 = getelementptr %"struct.std::_Vector_base<int,std::allocator<int> >::_Vector_impl"* %13, i32 0, i32 0 ; <i32**> [#uses=1] - %15 = load i32** %14, align 4 ; <i32*> [#uses=1] - %16 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %0, i32 0, i32 0 ; <i32**> [#uses=1] - store i32* %15, i32** %16, align 4 - %17 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %0, i32 0, i32 0 ; <i32**> [#uses=1] - %18 = load i32** %17, align 4 ; <i32*> [#uses=1] - %tmp2.i17 = ptrtoint i32* %18 to i32 ; <i32> [#uses=1] - %tmp1.i18 = inttoptr i32 %tmp2.i17 to i32* ; <i32*> [#uses=1] - %tmp8 = ptrtoint i32* %tmp1.i18 to i32 ; <i32> [#uses=1] - %tmp6 = inttoptr i32 %tmp8 to i32* ; <i32*> [#uses=1] - %19 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i, i32 0, i32 0 ; <i32**> [#uses=1] - store i32* %tmp6, i32** %19 - %20 = getelementptr %"struct.std::bidirectional_iterator_tag"* %1, i32 0, i32 0 ; <i8*> [#uses=1] - %21 = load i8* %20, align 1 ; <i8> [#uses=1] - %22 = or i8 %21, 0 ; <i8> [#uses=1] - %23 = or i8 %22, 0 ; <i8> [#uses=1] - %24 = or i8 %23, 0 ; <i8> [#uses=0] - %25 = getelementptr %"struct.std::bidirectional_iterator_tag"* %2, i32 0, i32 0 ; <i8*> [#uses=1] - store i8 0, i8* %25, align 1 - %elt.i = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i, i32 0, i32 0 ; <i32**> [#uses=1] - %val.i = load i32** %elt.i ; <i32*> [#uses=1] - %tmp.i = bitcast %"struct.std::bidirectional_iterator_tag"* %unnamed_arg.i to i8* ; <i8*> [#uses=1] - %tmp9.i = bitcast %"struct.std::bidirectional_iterator_tag"* %2 to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i64(i8* %tmp.i, i8* %tmp9.i, i64 1, i32 1) - %26 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - store i32* %val.i, i32** %26 - %27 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__last_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - store i32* %tmp2, i32** %27 - %28 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__last_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %29 = load i32** %28, align 4 ; <i32*> [#uses=1] - %30 = ptrtoint i32* %29 to i32 ; <i32> [#uses=1] - %31 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %32 = load i32** %31, align 4 ; <i32*> [#uses=1] - %33 = ptrtoint i32* %32 to i32 ; <i32> [#uses=1] - %34 = sub i32 %30, %33 ; <i32> [#uses=1] - %35 = ashr i32 %34, 2 ; <i32> [#uses=1] - %36 = ashr i32 %35, 2 ; <i32> [#uses=1] - br label %bb12.i.i + %0 = alloca %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >" + %__first_addr.i.i = alloca %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >" + %__last_addr.i.i = alloca %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >" + %unnamed_arg.i = alloca %"struct.std::bidirectional_iterator_tag", align 8 + %1 = alloca %"struct.std::bidirectional_iterator_tag" + %__first_addr.i = alloca %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >" + %2 = alloca %"struct.std::bidirectional_iterator_tag" + %3 = alloca %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >" + %4 = alloca i32 + %"alloca point" = bitcast i32 0 to i32 + store i32 42, i32* %4, align 4 + %5 = getelementptr %"struct.std::vector<int,std::allocator<int> >"* %X, i32 0, i32 0 + %6 = getelementptr %"struct.std::_Vector_base<int,std::allocator<int> >"* %5, i32 0, i32 0 + %7 = getelementptr %"struct.std::_Vector_base<int,std::allocator<int> >::_Vector_impl"* %6, i32 0, i32 1 + %8 = load i32** %7, align 4 + %9 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %3, i32 0, i32 0 + store i32* %8, i32** %9, align 4 + %10 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %3, i32 0, i32 0 + %11 = load i32** %10, align 4 + %tmp2.i = ptrtoint i32* %11 to i32 + %tmp1.i = inttoptr i32 %tmp2.i to i32* + %tmp3 = ptrtoint i32* %tmp1.i to i32 + %tmp2 = inttoptr i32 %tmp3 to i32* + %12 = getelementptr %"struct.std::vector<int,std::allocator<int> >"* %X, i32 0, i32 0 + %13 = getelementptr %"struct.std::_Vector_base<int,std::allocator<int> >"* %12, i32 0, i32 0 + %14 = getelementptr %"struct.std::_Vector_base<int,std::allocator<int> >::_Vector_impl"* %13, i32 0, i32 0 + %15 = load i32** %14, align 4 + %16 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %0, i32 0, i32 0 + store i32* %15, i32** %16, align 4 + %17 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %0, i32 0, i32 0 + %18 = load i32** %17, align 4 + %tmp2.i17 = ptrtoint i32* %18 to i32 + %tmp1.i18 = inttoptr i32 %tmp2.i17 to i32* + %tmp8 = ptrtoint i32* %tmp1.i18 to i32 + %tmp6 = inttoptr i32 %tmp8 to i32* + %19 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i, i32 0, i32 0 + store i32* %tmp6, i32** %19 + %20 = getelementptr %"struct.std::bidirectional_iterator_tag"* %1, i32 0, i32 0 + %21 = load i8* %20, align 1 + %22 = or i8 %21, 0 + %23 = or i8 %22, 0 + %24 = or i8 %23, 0 + %25 = getelementptr %"struct.std::bidirectional_iterator_tag"* %2, i32 0, i32 0 + store i8 0, i8* %25, align 1 + %elt.i = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i, i32 0, i32 0 + %val.i = load i32** %elt.i + %tmp.i = bitcast %"struct.std::bidirectional_iterator_tag"* %unnamed_arg.i to i8* + %tmp9.i = bitcast %"struct.std::bidirectional_iterator_tag"* %2 to i8* + call void @llvm.memcpy.p0i8.p0i8.i64(i8* %tmp.i, i8* %tmp9.i, i64 1, i32 1, i1 false) + %26 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + store i32* %val.i, i32** %26 + %27 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__last_addr.i.i, i32 0, i32 0 + store i32* %tmp2, i32** %27 + %28 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__last_addr.i.i, i32 0, i32 0 + %29 = load i32** %28, align 4 + %30 = ptrtoint i32* %29 to i32 + %31 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %32 = load i32** %31, align 4 + %33 = ptrtoint i32* %32 to i32 + %34 = sub i32 %30, %33 + %35 = ashr i32 %34, 2 + %36 = ashr i32 %35, 2 + br label %bb12.i.i -bb.i.i: ; preds = %bb12.i.i - %37 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %38 = load i32** %37, align 4 ; <i32*> [#uses=1] - %39 = load i32* %38, align 4 ; <i32> [#uses=1] - %40 = load i32* %4, align 4 ; <i32> [#uses=1] - %41 = icmp eq i32 %39, %40 ; <i1> [#uses=1] - %42 = zext i1 %41 to i8 ; <i8> [#uses=1] - %toBool.i.i = icmp ne i8 %42, 0 ; <i1> [#uses=1] - br i1 %toBool.i.i, label %bb1.i.i, label %bb2.i.i +bb.i.i: ; preds = %bb12.i.i + %37 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %38 = load i32** %37, align 4 + %39 = load i32* %38, align 4 + %40 = load i32* %4, align 4 + %41 = icmp eq i32 %39, %40 + %42 = zext i1 %41 to i8 + %toBool.i.i = icmp ne i8 %42, 0 + br i1 %toBool.i.i, label %bb1.i.i, label %bb2.i.i -bb1.i.i: ; preds = %bb.i.i - %43 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %44 = load i32** %43, align 4 ; <i32*> [#uses=1] - br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit +bb1.i.i: ; preds = %bb.i.i + %43 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %44 = load i32** %43, align 4 + br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit -bb2.i.i: ; preds = %bb.i.i - %45 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %46 = load i32** %45, align 4 ; <i32*> [#uses=1] - %47 = getelementptr i32* %46, i64 1 ; <i32*> [#uses=1] - %48 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - store i32* %47, i32** %48, align 4 - %49 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %50 = load i32** %49, align 4 ; <i32*> [#uses=1] - %51 = load i32* %50, align 4 ; <i32> [#uses=1] - %52 = load i32* %4, align 4 ; <i32> [#uses=1] - %53 = icmp eq i32 %51, %52 ; <i1> [#uses=1] - %54 = zext i1 %53 to i8 ; <i8> [#uses=1] - %toBool3.i.i = icmp ne i8 %54, 0 ; <i1> [#uses=1] - br i1 %toBool3.i.i, label %bb4.i.i, label %bb5.i.i +bb2.i.i: ; preds = %bb.i.i + %45 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %46 = load i32** %45, align 4 + %47 = getelementptr i32* %46, i64 1 + %48 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + store i32* %47, i32** %48, align 4 + %49 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %50 = load i32** %49, align 4 + %51 = load i32* %50, align 4 + %52 = load i32* %4, align 4 + %53 = icmp eq i32 %51, %52 + %54 = zext i1 %53 to i8 + %toBool3.i.i = icmp ne i8 %54, 0 + br i1 %toBool3.i.i, label %bb4.i.i, label %bb5.i.i -bb4.i.i: ; preds = %bb2.i.i - %55 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %56 = load i32** %55, align 4 ; <i32*> [#uses=1] - br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit +bb4.i.i: ; preds = %bb2.i.i + %55 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %56 = load i32** %55, align 4 + br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit -bb5.i.i: ; preds = %bb2.i.i - %57 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %58 = load i32** %57, align 4 ; <i32*> [#uses=1] - %59 = getelementptr i32* %58, i64 1 ; <i32*> [#uses=1] - %60 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - store i32* %59, i32** %60, align 4 - %61 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %62 = load i32** %61, align 4 ; <i32*> [#uses=1] - %63 = load i32* %62, align 4 ; <i32> [#uses=1] - %64 = load i32* %4, align 4 ; <i32> [#uses=1] - %65 = icmp eq i32 %63, %64 ; <i1> [#uses=1] - %66 = zext i1 %65 to i8 ; <i8> [#uses=1] - %toBool6.i.i = icmp ne i8 %66, 0 ; <i1> [#uses=1] - br i1 %toBool6.i.i, label %bb7.i.i, label %bb8.i.i +bb5.i.i: ; preds = %bb2.i.i + %57 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %58 = load i32** %57, align 4 + %59 = getelementptr i32* %58, i64 1 + %60 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + store i32* %59, i32** %60, align 4 + %61 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %62 = load i32** %61, align 4 + %63 = load i32* %62, align 4 + %64 = load i32* %4, align 4 + %65 = icmp eq i32 %63, %64 + %66 = zext i1 %65 to i8 + %toBool6.i.i = icmp ne i8 %66, 0 + br i1 %toBool6.i.i, label %bb7.i.i, label %bb8.i.i -bb7.i.i: ; preds = %bb5.i.i - %67 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %68 = load i32** %67, align 4 ; <i32*> [#uses=1] - br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit +bb7.i.i: ; preds = %bb5.i.i + %67 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %68 = load i32** %67, align 4 + br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit -bb8.i.i: ; preds = %bb5.i.i - %69 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %70 = load i32** %69, align 4 ; <i32*> [#uses=1] - %71 = getelementptr i32* %70, i64 1 ; <i32*> [#uses=1] - %72 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - store i32* %71, i32** %72, align 4 - %73 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %74 = load i32** %73, align 4 ; <i32*> [#uses=1] - %75 = load i32* %74, align 4 ; <i32> [#uses=1] - %76 = load i32* %4, align 4 ; <i32> [#uses=1] - %77 = icmp eq i32 %75, %76 ; <i1> [#uses=1] - %78 = zext i1 %77 to i8 ; <i8> [#uses=1] - %toBool9.i.i = icmp ne i8 %78, 0 ; <i1> [#uses=1] - br i1 %toBool9.i.i, label %bb10.i.i, label %bb11.i.i +bb8.i.i: ; preds = %bb5.i.i + %69 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %70 = load i32** %69, align 4 + %71 = getelementptr i32* %70, i64 1 + %72 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + store i32* %71, i32** %72, align 4 + %73 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %74 = load i32** %73, align 4 + %75 = load i32* %74, align 4 + %76 = load i32* %4, align 4 + %77 = icmp eq i32 %75, %76 + %78 = zext i1 %77 to i8 + %toBool9.i.i = icmp ne i8 %78, 0 + br i1 %toBool9.i.i, label %bb10.i.i, label %bb11.i.i -bb10.i.i: ; preds = %bb8.i.i - %79 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %80 = load i32** %79, align 4 ; <i32*> [#uses=1] - br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit +bb10.i.i: ; preds = %bb8.i.i + %79 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %80 = load i32** %79, align 4 + br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit -bb11.i.i: ; preds = %bb8.i.i - %81 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %82 = load i32** %81, align 4 ; <i32*> [#uses=1] - %83 = getelementptr i32* %82, i64 1 ; <i32*> [#uses=1] - %84 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - store i32* %83, i32** %84, align 4 - %85 = sub i32 %__trip_count.0.i.i, 1 ; <i32> [#uses=1] - br label %bb12.i.i +bb11.i.i: ; preds = %bb8.i.i + %81 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %82 = load i32** %81, align 4 + %83 = getelementptr i32* %82, i64 1 + %84 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + store i32* %83, i32** %84, align 4 + %85 = sub i32 %__trip_count.0.i.i, 1 + br label %bb12.i.i -bb12.i.i: ; preds = %bb11.i.i, %entry - %__trip_count.0.i.i = phi i32 [ %36, %entry ], [ %85, %bb11.i.i ] ; <i32> [#uses=2] - %86 = icmp sgt i32 %__trip_count.0.i.i, 0 ; <i1> [#uses=1] - br i1 %86, label %bb.i.i, label %bb13.i.i +bb12.i.i: ; preds = %bb11.i.i, %entry + %__trip_count.0.i.i = phi i32 [ %36, %entry ], [ %85, %bb11.i.i ] + %86 = icmp sgt i32 %__trip_count.0.i.i, 0 + br i1 %86, label %bb.i.i, label %bb13.i.i -bb13.i.i: ; preds = %bb12.i.i - %87 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__last_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %88 = load i32** %87, align 4 ; <i32*> [#uses=1] - %89 = ptrtoint i32* %88 to i32 ; <i32> [#uses=1] - %90 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %91 = load i32** %90, align 4 ; <i32*> [#uses=1] - %92 = ptrtoint i32* %91 to i32 ; <i32> [#uses=1] - %93 = sub i32 %89, %92 ; <i32> [#uses=1] - %94 = ashr i32 %93, 2 ; <i32> [#uses=1] - switch i32 %94, label %bb26.i.i [ - i32 1, label %bb22.i.i - i32 2, label %bb18.i.i - i32 3, label %bb14.i.i - ] +bb13.i.i: ; preds = %bb12.i.i + %87 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__last_addr.i.i, i32 0, i32 0 + %88 = load i32** %87, align 4 + %89 = ptrtoint i32* %88 to i32 + %90 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %91 = load i32** %90, align 4 + %92 = ptrtoint i32* %91 to i32 + %93 = sub i32 %89, %92 + %94 = ashr i32 %93, 2 + switch i32 %94, label %bb26.i.i [ + i32 1, label %bb22.i.i + i32 2, label %bb18.i.i + i32 3, label %bb14.i.i + ] -bb14.i.i: ; preds = %bb13.i.i - %95 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %96 = load i32** %95, align 4 ; <i32*> [#uses=1] - %97 = load i32* %96, align 4 ; <i32> [#uses=1] - %98 = load i32* %4, align 4 ; <i32> [#uses=1] - %99 = icmp eq i32 %97, %98 ; <i1> [#uses=1] - %100 = zext i1 %99 to i8 ; <i8> [#uses=1] - %toBool15.i.i = icmp ne i8 %100, 0 ; <i1> [#uses=1] - br i1 %toBool15.i.i, label %bb16.i.i, label %bb17.i.i +bb14.i.i: ; preds = %bb13.i.i + %95 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %96 = load i32** %95, align 4 + %97 = load i32* %96, align 4 + %98 = load i32* %4, align 4 + %99 = icmp eq i32 %97, %98 + %100 = zext i1 %99 to i8 + %toBool15.i.i = icmp ne i8 %100, 0 + br i1 %toBool15.i.i, label %bb16.i.i, label %bb17.i.i -bb16.i.i: ; preds = %bb14.i.i - %101 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %102 = load i32** %101, align 4 ; <i32*> [#uses=1] - br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit +bb16.i.i: ; preds = %bb14.i.i + %101 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %102 = load i32** %101, align 4 + br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit -bb17.i.i: ; preds = %bb14.i.i - %103 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %104 = load i32** %103, align 4 ; <i32*> [#uses=1] - %105 = getelementptr i32* %104, i64 1 ; <i32*> [#uses=1] - %106 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - store i32* %105, i32** %106, align 4 - br label %bb18.i.i +bb17.i.i: ; preds = %bb14.i.i + %103 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %104 = load i32** %103, align 4 + %105 = getelementptr i32* %104, i64 1 + %106 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + store i32* %105, i32** %106, align 4 + br label %bb18.i.i -bb18.i.i: ; preds = %bb17.i.i, %bb13.i.i - %107 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %108 = load i32** %107, align 4 ; <i32*> [#uses=1] - %109 = load i32* %108, align 4 ; <i32> [#uses=1] - %110 = load i32* %4, align 4 ; <i32> [#uses=1] - %111 = icmp eq i32 %109, %110 ; <i1> [#uses=1] - %112 = zext i1 %111 to i8 ; <i8> [#uses=1] - %toBool19.i.i = icmp ne i8 %112, 0 ; <i1> [#uses=1] - br i1 %toBool19.i.i, label %bb20.i.i, label %bb21.i.i +bb18.i.i: ; preds = %bb17.i.i, %bb13.i.i + %107 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %108 = load i32** %107, align 4 + %109 = load i32* %108, align 4 + %110 = load i32* %4, align 4 + %111 = icmp eq i32 %109, %110 + %112 = zext i1 %111 to i8 + %toBool19.i.i = icmp ne i8 %112, 0 + br i1 %toBool19.i.i, label %bb20.i.i, label %bb21.i.i -bb20.i.i: ; preds = %bb18.i.i - %113 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %114 = load i32** %113, align 4 ; <i32*> [#uses=1] - br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit +bb20.i.i: ; preds = %bb18.i.i + %113 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %114 = load i32** %113, align 4 + br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit -bb21.i.i: ; preds = %bb18.i.i - %115 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %116 = load i32** %115, align 4 ; <i32*> [#uses=1] - %117 = getelementptr i32* %116, i64 1 ; <i32*> [#uses=1] - %118 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - store i32* %117, i32** %118, align 4 - br label %bb22.i.i +bb21.i.i: ; preds = %bb18.i.i + %115 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %116 = load i32** %115, align 4 + %117 = getelementptr i32* %116, i64 1 + %118 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + store i32* %117, i32** %118, align 4 + br label %bb22.i.i -bb22.i.i: ; preds = %bb21.i.i, %bb13.i.i - %119 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %120 = load i32** %119, align 4 ; <i32*> [#uses=1] - %121 = load i32* %120, align 4 ; <i32> [#uses=1] - %122 = load i32* %4, align 4 ; <i32> [#uses=1] - %123 = icmp eq i32 %121, %122 ; <i1> [#uses=1] - %124 = zext i1 %123 to i8 ; <i8> [#uses=1] - %toBool23.i.i = icmp ne i8 %124, 0 ; <i1> [#uses=1] - br i1 %toBool23.i.i, label %bb24.i.i, label %bb25.i.i +bb22.i.i: ; preds = %bb21.i.i, %bb13.i.i + %119 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %120 = load i32** %119, align 4 + %121 = load i32* %120, align 4 + %122 = load i32* %4, align 4 + %123 = icmp eq i32 %121, %122 + %124 = zext i1 %123 to i8 + %toBool23.i.i = icmp ne i8 %124, 0 + br i1 %toBool23.i.i, label %bb24.i.i, label %bb25.i.i -bb24.i.i: ; preds = %bb22.i.i - %125 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %126 = load i32** %125, align 4 ; <i32*> [#uses=1] - br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit +bb24.i.i: ; preds = %bb22.i.i + %125 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %126 = load i32** %125, align 4 + br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit -bb25.i.i: ; preds = %bb22.i.i - %127 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %128 = load i32** %127, align 4 ; <i32*> [#uses=1] - %129 = getelementptr i32* %128, i64 1 ; <i32*> [#uses=1] - %130 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - store i32* %129, i32** %130, align 4 - br label %bb26.i.i +bb25.i.i: ; preds = %bb22.i.i + %127 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + %128 = load i32** %127, align 4 + %129 = getelementptr i32* %128, i64 1 + %130 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__first_addr.i.i, i32 0, i32 0 + store i32* %129, i32** %130, align 4 + br label %bb26.i.i -bb26.i.i: ; preds = %bb25.i.i, %bb13.i.i - %131 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__last_addr.i.i, i32 0, i32 0 ; <i32**> [#uses=1] - %132 = load i32** %131, align 4 ; <i32*> [#uses=1] - br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit +bb26.i.i: ; preds = %bb25.i.i, %bb13.i.i + %131 = getelementptr %"struct.__gnu_cxx::__normal_iterator<int*,std::vector<int, std::allocator<int> > >"* %__last_addr.i.i, i32 0, i32 0 + %132 = load i32** %131, align 4 + br label %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit -_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit: ; preds = %bb26.i.i, %bb24.i.i, %bb20.i.i, %bb16.i.i, %bb10.i.i, %bb7.i.i, %bb4.i.i, %bb1.i.i - %.0.0.i.i = phi i32* [ %132, %bb26.i.i ], [ %126, %bb24.i.i ], [ %114, %bb20.i.i ], [ %102, %bb16.i.i ], [ %80, %bb10.i.i ], [ %68, %bb7.i.i ], [ %56, %bb4.i.i ], [ %44, %bb1.i.i ] ; <i32*> [#uses=1] - %tmp2.i.i = ptrtoint i32* %.0.0.i.i to i32 ; <i32> [#uses=1] - %tmp1.i.i = inttoptr i32 %tmp2.i.i to i32* ; <i32*> [#uses=1] - %tmp4.i = ptrtoint i32* %tmp1.i.i to i32 ; <i32> [#uses=1] - %tmp3.i = inttoptr i32 %tmp4.i to i32* ; <i32*> [#uses=1] - %tmp8.i = ptrtoint i32* %tmp3.i to i32 ; <i32> [#uses=1] - %tmp6.i = inttoptr i32 %tmp8.i to i32* ; <i32*> [#uses=1] - %tmp12 = ptrtoint i32* %tmp6.i to i32 ; <i32> [#uses=1] - %tmp10 = inttoptr i32 %tmp12 to i32* ; <i32*> [#uses=1] - %tmp16 = ptrtoint i32* %tmp10 to i32 ; <i32> [#uses=1] - br label %return +_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit: ; preds = %bb26.i.i, %bb24.i.i, %bb20.i.i, %bb16.i.i, %bb10.i.i, %bb7.i.i, %bb4.i.i, %bb1.i.i + %.0.0.i.i = phi i32* [ %132, %bb26.i.i ], [ %126, %bb24.i.i ], [ %114, %bb20.i.i ], [ %102, %bb16.i.i ], [ %80, %bb10.i.i ], [ %68, %bb7.i.i ], [ %56, %bb4.i.i ], [ %44, %bb1.i.i ] + %tmp2.i.i = ptrtoint i32* %.0.0.i.i to i32 + %tmp1.i.i = inttoptr i32 %tmp2.i.i to i32* + %tmp4.i = ptrtoint i32* %tmp1.i.i to i32 + %tmp3.i = inttoptr i32 %tmp4.i to i32* + %tmp8.i = ptrtoint i32* %tmp3.i to i32 + %tmp6.i = inttoptr i32 %tmp8.i to i32* + %tmp12 = ptrtoint i32* %tmp6.i to i32 + %tmp10 = inttoptr i32 %tmp12 to i32* + %tmp16 = ptrtoint i32* %tmp10 to i32 + br label %return -return: ; preds = %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit - %tmp14 = inttoptr i32 %tmp16 to i32* ; <i32*> [#uses=1] - ret i32* %tmp14 +return: ; preds = %_ZSt4findIN9__gnu_cxx17__normal_iteratorIPiSt6vectorIiSaIiEEEEiET_S7_S7_RKT0_.exit + %tmp14 = inttoptr i32 %tmp16 to i32* + ret i32* %tmp14 } -declare void @llvm.memcpy.i64(i8* nocapture, i8* nocapture, i64, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture, i64, i32, i1) nounwind diff --git a/test/Transforms/InstCombine/memcpy-to-load.ll b/test/Transforms/InstCombine/memcpy-to-load.ll index ebb8711..04aac98 100644 --- a/test/Transforms/InstCombine/memcpy-to-load.ll +++ b/test/Transforms/InstCombine/memcpy-to-load.ll @@ -4,11 +4,10 @@ target triple = "i686-apple-darwin8" define void @foo(double* %X, double* %Y) { entry: - %"alloca point" = bitcast i32 0 to i32 ; <i32> [#uses=0] - %tmp2 = bitcast double* %X to i8* ; <i8*> [#uses=1] - %tmp13 = bitcast double* %Y to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %tmp2, i8* %tmp13, i32 8, i32 1 ) - ret void + %tmp2 = bitcast double* %X to i8* + %tmp13 = bitcast double* %Y to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp2, i8* %tmp13, i32 8, i32 1, i1 false) + ret void } -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/InstCombine/memmove.ll b/test/Transforms/InstCombine/memmove.ll index 1806cfc..4602c12 100644 --- a/test/Transforms/InstCombine/memmove.ll +++ b/test/Transforms/InstCombine/memmove.ll @@ -1,24 +1,20 @@ ; This test makes sure that memmove instructions are properly eliminated. ; -; RUN: opt < %s -instcombine -S | \ -; RUN: not grep {call void @llvm.memmove} +; RUN: opt < %s -instcombine -S | not grep {call void @llvm.memmove} @S = internal constant [33 x i8] c"panic: restorelist inconsistency\00" ; <[33 x i8]*> [#uses=1] @h = constant [2 x i8] c"h\00" ; <[2 x i8]*> [#uses=1] @hel = constant [4 x i8] c"hel\00" ; <[4 x i8]*> [#uses=1] @hello_u = constant [8 x i8] c"hello_u\00" ; <[8 x i8]*> [#uses=1] - -declare void @llvm.memmove.i32(i8*, i8*, i32, i32) - define void @test1(i8* %A, i8* %B, i32 %N) { - call void @llvm.memmove.i32( i8* %A, i8* %B, i32 0, i32 1 ) + call void @llvm.memmove.p0i8.p0i8.i32(i8* %A, i8* %B, i32 0, i32 1, i1 false) ret void } define void @test2(i8* %A, i32 %N) { ;; dest can't alias source since we can't write to source! - call void @llvm.memmove.i32( i8* %A, i8* getelementptr ([33 x i8]* @S, i32 0, i32 0), i32 %N, i32 1 ) + call void @llvm.memmove.p0i8.p0i8.i32(i8* %A, i8* getelementptr inbounds ([33 x i8]* @S, i32 0, i32 0), i32 %N, i32 1, i1 false) ret void } @@ -28,15 +24,16 @@ define i32 @test3() { %hello_u_p = getelementptr [8 x i8]* @hello_u, i32 0, i32 0 ; <i8*> [#uses=1] %target = alloca [1024 x i8] ; <[1024 x i8]*> [#uses=1] %target_p = getelementptr [1024 x i8]* %target, i32 0, i32 0 ; <i8*> [#uses=3] - call void @llvm.memmove.i32( i8* %target_p, i8* %h_p, i32 2, i32 2 ) - call void @llvm.memmove.i32( i8* %target_p, i8* %hel_p, i32 4, i32 4 ) - call void @llvm.memmove.i32( i8* %target_p, i8* %hello_u_p, i32 8, i32 8 ) + call void @llvm.memmove.p0i8.p0i8.i32(i8* %target_p, i8* %h_p, i32 2, i32 2, i1 false) + call void @llvm.memmove.p0i8.p0i8.i32(i8* %target_p, i8* %hel_p, i32 4, i32 4, i1 false) + call void @llvm.memmove.p0i8.p0i8.i32(i8* %target_p, i8* %hello_u_p, i32 8, i32 8, i1 false) ret i32 0 } ; PR2370 define void @test4(i8* %a) { - tail call void @llvm.memmove.i32( i8* %a, i8* %a, i32 100, i32 1 ) - ret void + tail call void @llvm.memmove.p0i8.p0i8.i32(i8* %a, i8* %a, i32 100, i32 1, i1 false) + ret void } +declare void @llvm.memmove.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/InstCombine/memset.ll b/test/Transforms/InstCombine/memset.ll index 8e85694..7f7bc9f 100644 --- a/test/Transforms/InstCombine/memset.ll +++ b/test/Transforms/InstCombine/memset.ll @@ -1,15 +1,14 @@ ; RUN: opt < %s -instcombine -S | not grep {call.*llvm.memset} -declare void @llvm.memset.i32(i8*, i8, i32, i32) - define i32 @main() { - %target = alloca [1024 x i8] ; <[1024 x i8]*> [#uses=1] - %target_p = getelementptr [1024 x i8]* %target, i32 0, i32 0 ; <i8*> [#uses=5] - call void @llvm.memset.i32( i8* %target_p, i8 1, i32 0, i32 1 ) - call void @llvm.memset.i32( i8* %target_p, i8 1, i32 1, i32 1 ) - call void @llvm.memset.i32( i8* %target_p, i8 1, i32 2, i32 2 ) - call void @llvm.memset.i32( i8* %target_p, i8 1, i32 4, i32 4 ) - call void @llvm.memset.i32( i8* %target_p, i8 1, i32 8, i32 8 ) - ret i32 0 + %target = alloca [1024 x i8] + %target_p = getelementptr [1024 x i8]* %target, i32 0, i32 0 + call void @llvm.memset.p0i8.i32(i8* %target_p, i8 1, i32 0, i32 1, i1 false) + call void @llvm.memset.p0i8.i32(i8* %target_p, i8 1, i32 1, i32 1, i1 false) + call void @llvm.memset.p0i8.i32(i8* %target_p, i8 1, i32 2, i32 2, i1 false) + call void @llvm.memset.p0i8.i32(i8* %target_p, i8 1, i32 4, i32 4, i1 false) + call void @llvm.memset.p0i8.i32(i8* %target_p, i8 1, i32 8, i32 8, i1 false) + ret i32 0 } +declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1) nounwind diff --git a/test/Transforms/InstCombine/stack-overalign.ll b/test/Transforms/InstCombine/stack-overalign.ll index 88b4114..2fc8414 100644 --- a/test/Transforms/InstCombine/stack-overalign.ll +++ b/test/Transforms/InstCombine/stack-overalign.ll @@ -17,13 +17,13 @@ define void @foo() nounwind { entry: - %src = alloca [1024 x i8], align 1 - %src1 = getelementptr [1024 x i8]* %src, i32 0, i32 0 - call void @llvm.memcpy.i32(i8* getelementptr ([1024 x i8]* @dst, i32 0, i32 0), i8* %src1, i32 1024, i32 1) - call void @frob(i8* %src1) nounwind - ret void + %src = alloca [1024 x i8], align 1 + %src1 = getelementptr [1024 x i8]* %src, i32 0, i32 0 + call void @llvm.memcpy.p0i8.p0i8.i32(i8* getelementptr inbounds ([1024 x i8]* @dst, i32 0, i32 0), i8* %src1, i32 1024, i32 1, i1 false) + call void @frob(i8* %src1) nounwind + ret void } -declare void @llvm.memcpy.i32(i8* nocapture, i8* nocapture, i32, i32) nounwind - declare void @frob(i8*) + +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/MemCpyOpt/2008-02-24-MultipleUseofSRet.ll b/test/Transforms/MemCpyOpt/2008-02-24-MultipleUseofSRet.ll index 9f1e280..b95ad91 100644 --- a/test/Transforms/MemCpyOpt/2008-02-24-MultipleUseofSRet.ll +++ b/test/Transforms/MemCpyOpt/2008-02-24-MultipleUseofSRet.ll @@ -4,31 +4,33 @@ target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:32:32" target triple = "i386-pc-linux-gnu" -define internal fastcc void @initialize({ x86_fp80, x86_fp80 }* noalias sret %agg.result) nounwind { +%0 = type { x86_fp80, x86_fp80 } + +define internal fastcc void @initialize(%0* noalias sret %agg.result) nounwind { entry: - %agg.result.03 = getelementptr { x86_fp80, x86_fp80 }* %agg.result, i32 0, i32 0 ; <x86_fp80*> [#uses=1] - store x86_fp80 0xK00000000000000000000, x86_fp80* %agg.result.03 - %agg.result.15 = getelementptr { x86_fp80, x86_fp80 }* %agg.result, i32 0, i32 1 ; <x86_fp80*> [#uses=1] - store x86_fp80 0xK00000000000000000000, x86_fp80* %agg.result.15 - ret void + %agg.result.03 = getelementptr %0* %agg.result, i32 0, i32 0 + store x86_fp80 0xK00000000000000000000, x86_fp80* %agg.result.03 + %agg.result.15 = getelementptr %0* %agg.result, i32 0, i32 1 + store x86_fp80 0xK00000000000000000000, x86_fp80* %agg.result.15 + ret void } -declare fastcc x86_fp80 @passed_uninitialized({ x86_fp80, x86_fp80 }* %x) nounwind +declare fastcc x86_fp80 @passed_uninitialized(%0*) nounwind -define fastcc void @badly_optimized() nounwind { +define fastcc void @badly_optimized() nounwind { entry: - %z = alloca { x86_fp80, x86_fp80 } ; <{ x86_fp80, x86_fp80 }*> [#uses=2] - %tmp = alloca { x86_fp80, x86_fp80 } ; <{ x86_fp80, x86_fp80 }*> [#uses=2] - %memtmp = alloca { x86_fp80, x86_fp80 }, align 8 ; <{ x86_fp80, x86_fp80 }*> [#uses=2] - call fastcc void @initialize( { x86_fp80, x86_fp80 }* noalias sret %memtmp ) - %tmp1 = bitcast { x86_fp80, x86_fp80 }* %tmp to i8* ; <i8*> [#uses=1] - %memtmp2 = bitcast { x86_fp80, x86_fp80 }* %memtmp to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %tmp1, i8* %memtmp2, i32 24, i32 8 ) - %z3 = bitcast { x86_fp80, x86_fp80 }* %z to i8* ; <i8*> [#uses=1] - %tmp4 = bitcast { x86_fp80, x86_fp80 }* %tmp to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %z3, i8* %tmp4, i32 24, i32 8 ) - %tmp5 = call fastcc x86_fp80 @passed_uninitialized( { x86_fp80, x86_fp80 }* %z ) ; <x86_fp80> [#uses=0] - ret void + %z = alloca %0 + %tmp = alloca %0 + %memtmp = alloca %0, align 8 + call fastcc void @initialize(%0* noalias sret %memtmp) + %tmp1 = bitcast %0* %tmp to i8* + %memtmp2 = bitcast %0* %memtmp to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp1, i8* %memtmp2, i32 24, i32 8, i1 false) + %z3 = bitcast %0* %z to i8* + %tmp4 = bitcast %0* %tmp to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %z3, i8* %tmp4, i32 24, i32 8, i1 false) + %tmp5 = call fastcc x86_fp80 @passed_uninitialized(%0* %z) + ret void } -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/MemCpyOpt/2008-03-13-ReturnSlotBitcast.ll b/test/Transforms/MemCpyOpt/2008-03-13-ReturnSlotBitcast.ll index 418761e..24cf576 100644 --- a/test/Transforms/MemCpyOpt/2008-03-13-ReturnSlotBitcast.ll +++ b/test/Transforms/MemCpyOpt/2008-03-13-ReturnSlotBitcast.ll @@ -1,20 +1,22 @@ ; RUN: opt < %s -basicaa -memcpyopt -S | not grep {call.*memcpy.} target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64" - %a = type { i32 } - %b = type { float } -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) nounwind +%a = type { i32 } +%b = type { float } + declare void @g(%a*) define float @f() { entry: - %a_var = alloca %a - %b_var = alloca %b - call void @g(%a *%a_var) - %a_i8 = bitcast %a* %a_var to i8* - %b_i8 = bitcast %b* %b_var to i8* - call void @llvm.memcpy.i32(i8* %b_i8, i8* %a_i8, i32 4, i32 4) - %tmp1 = getelementptr %b* %b_var, i32 0, i32 0 - %tmp2 = load float* %tmp1 - ret float %tmp2 + %a_var = alloca %a + %b_var = alloca %b + call void @g(%a* %a_var) + %a_i8 = bitcast %a* %a_var to i8* + %b_i8 = bitcast %b* %b_var to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %b_i8, i8* %a_i8, i32 4, i32 4, i1 false) + %tmp1 = getelementptr %b* %b_var, i32 0, i32 0 + %tmp2 = load float* %tmp1 + ret float %tmp2 } + +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/MemCpyOpt/memcpy.ll b/test/Transforms/MemCpyOpt/memcpy.ll index fb97913..12519ef 100644 --- a/test/Transforms/MemCpyOpt/memcpy.ll +++ b/test/Transforms/MemCpyOpt/memcpy.ll @@ -3,17 +3,21 @@ target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" target triple = "i686-apple-darwin9" +%0 = type { x86_fp80, x86_fp80 } +%1 = type { i32, i32 } + define void @test1({ x86_fp80, x86_fp80 }* sret %agg.result, x86_fp80 %z.0, x86_fp80 %z.1) nounwind { entry: - %tmp2 = alloca { x86_fp80, x86_fp80 } ; <{ x86_fp80, x86_fp80 }*> [#uses=1] - %memtmp = alloca { x86_fp80, x86_fp80 }, align 16 ; <{ x86_fp80, x86_fp80 }*> [#uses=2] - %tmp5 = fsub x86_fp80 0xK80000000000000000000, %z.1 ; <x86_fp80> [#uses=1] - call void @ccoshl( { x86_fp80, x86_fp80 }* sret %memtmp, x86_fp80 %tmp5, x86_fp80 %z.0 ) nounwind - %tmp219 = bitcast { x86_fp80, x86_fp80 }* %tmp2 to i8* ; <i8*> [#uses=2] - %memtmp20 = bitcast { x86_fp80, x86_fp80 }* %memtmp to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %tmp219, i8* %memtmp20, i32 32, i32 16 ) - %agg.result21 = bitcast { x86_fp80, x86_fp80 }* %agg.result to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %agg.result21, i8* %tmp219, i32 32, i32 16 ) + %tmp2 = alloca %0 + %memtmp = alloca %0, align 16 + %tmp5 = fsub x86_fp80 0xK80000000000000000000, %z.1 + call void @ccoshl(%0* sret %memtmp, x86_fp80 %tmp5, x86_fp80 %z.0) nounwind + %tmp219 = bitcast %0* %tmp2 to i8* + %memtmp20 = bitcast %0* %memtmp to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp219, i8* %memtmp20, i32 32, i32 16, i1 false) + %agg.result21 = bitcast %0* %agg.result to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %agg.result21, i8* %tmp219, i32 32, i32 16, i1 false) + ret void ; Check that one of the memcpy's are removed. ;; FIXME: PR 8643 We should be able to eliminate the last memcpy here. @@ -23,22 +27,19 @@ entry: ; CHECK: call void @llvm.memcpy ; CHECK-NOT: llvm.memcpy ; CHECK: ret void - ret void } declare void @ccoshl({ x86_fp80, x86_fp80 }* sret , x86_fp80, x86_fp80) nounwind -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) nounwind - ; The intermediate alloca and one of the memcpy's should be eliminated, the ; other should be related with a memmove. define void @test2(i8* %P, i8* %Q) nounwind { - %memtmp = alloca { x86_fp80, x86_fp80 }, align 16 - %R = bitcast { x86_fp80, x86_fp80 }* %memtmp to i8* - call void @llvm.memcpy.i32( i8* %R, i8* %P, i32 32, i32 16 ) - call void @llvm.memcpy.i32( i8* %Q, i8* %R, i32 32, i32 16 ) - ret void + %memtmp = alloca %0, align 16 + %R = bitcast %0* %memtmp to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %R, i8* %P, i32 32, i32 16, i1 false) + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %Q, i8* %R, i32 32, i32 16, i1 false) + ret void ; CHECK: @test2 ; CHECK-NEXT: call void @llvm.memmove{{.*}}(i8* %Q, i8* %P @@ -51,12 +52,12 @@ define void @test2(i8* %P, i8* %Q) nounwind { @x = external global { x86_fp80, x86_fp80 } define void @test3({ x86_fp80, x86_fp80 }* noalias sret %agg.result) nounwind { - %x.0 = alloca { x86_fp80, x86_fp80 } - %x.01 = bitcast { x86_fp80, x86_fp80 }* %x.0 to i8* - call void @llvm.memcpy.i32( i8* %x.01, i8* bitcast ({ x86_fp80, x86_fp80 }* @x to i8*), i32 32, i32 16 ) - %agg.result2 = bitcast { x86_fp80, x86_fp80 }* %agg.result to i8* - call void @llvm.memcpy.i32( i8* %agg.result2, i8* %x.01, i32 32, i32 16 ) - ret void + %x.0 = alloca %0 + %x.01 = bitcast %0* %x.0 to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %x.01, i8* bitcast (%0* @x to i8*), i32 32, i32 16, i1 false) + %agg.result2 = bitcast %0* %agg.result to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %agg.result2, i8* %x.01, i32 32, i32 16, i1 false) + ret void ; CHECK: @test3 ; CHECK-NEXT: %agg.result2 = bitcast ; CHECK-NEXT: call void @llvm.memcpy @@ -66,10 +67,10 @@ define void @test3({ x86_fp80, x86_fp80 }* noalias sret %agg.result) nounwind { ; PR8644 define void @test4(i8 *%P) { - %A = alloca {i32, i32} - %a = bitcast {i32, i32}* %A to i8* + %A = alloca %1 + %a = bitcast %1* %A to i8* call void @llvm.memcpy.p0i8.p0i8.i64(i8* %a, i8* %P, i64 8, i32 4, i1 false) - call void @test4a(i8* byval align 1 %a) + call void @test4a(i8* byval align 1 %a) ret void ; CHECK: @test4 ; CHECK-NEXT: call void @test4a( @@ -127,4 +128,5 @@ entry: declare i32 @g(%struct.p* byval align 8) +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/MemCpyOpt/memmove.ll b/test/Transforms/MemCpyOpt/memmove.ll index 8d3fbd2..7f1667a 100644 --- a/test/Transforms/MemCpyOpt/memmove.ll +++ b/test/Transforms/MemCpyOpt/memmove.ll @@ -4,7 +4,7 @@ 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-s0:64:64-f80:128:128" target triple = "x86_64-apple-darwin9.0" -declare void @llvm.memmove.i64(i8* nocapture, i8* nocapture, i64, i32) nounwind +declare void @llvm.memmove.p0i8.p0i8.i64(i8* nocapture, i8* nocapture, i64, i32, i1) nounwind define i8* @test1(i8* nocapture %src) nounwind { entry: @@ -13,8 +13,8 @@ entry: %malloccall = tail call i8* @malloc(i32 trunc (i64 mul nuw (i64 ptrtoint (i8* getelementptr (i8* null, i32 1) to i64), i64 13) to i32)) %call3 = bitcast i8* %malloccall to [13 x i8]* - %call3.sub = getelementptr inbounds [13 x i8]* %call3, i64 0, i64 0 ; <i8*> [#uses=2] - tail call void @llvm.memmove.i64(i8* %call3.sub, i8* %src, i64 13, i32 1) + %call3.sub = getelementptr inbounds [13 x i8]* %call3, i64 0, i64 0 + tail call void @llvm.memmove.p0i8.p0i8.i64(i8* %call3.sub, i8* %src, i64 13, i32 1, i1 false) ret i8* %call3.sub } declare noalias i8* @malloc(i32) @@ -24,8 +24,8 @@ define void @test2(i8* %P) nounwind { entry: ; CHECK: @test2 ; CHECK: call void @llvm.memcpy - %add.ptr = getelementptr i8* %P, i64 16 ; <i8*> [#uses=1] - tail call void @llvm.memmove.i64(i8* %P, i8* %add.ptr, i64 16, i32 1) + %add.ptr = getelementptr i8* %P, i64 16 + tail call void @llvm.memmove.p0i8.p0i8.i64(i8* %P, i8* %add.ptr, i64 16, i32 1, i1 false) ret void } @@ -34,7 +34,7 @@ define void @test3(i8* %P) nounwind { entry: ; CHECK: @test3 ; CHECK: call void @llvm.memmove - %add.ptr = getelementptr i8* %P, i64 16 ; <i8*> [#uses=1] - tail call void @llvm.memmove.i64(i8* %P, i8* %add.ptr, i64 17, i32 1) + %add.ptr = getelementptr i8* %P, i64 16 + tail call void @llvm.memmove.p0i8.p0i8.i64(i8* %P, i8* %add.ptr, i64 17, i32 1, i1 false) ret void } diff --git a/test/Transforms/MemCpyOpt/sret.ll b/test/Transforms/MemCpyOpt/sret.ll index ddfd0fd..8eac7da 100644 --- a/test/Transforms/MemCpyOpt/sret.ll +++ b/test/Transforms/MemCpyOpt/sret.ll @@ -3,26 +3,28 @@ target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" target triple = "i686-apple-darwin9" -define void @ccosl({ x86_fp80, x86_fp80 }* noalias sret %agg.result, { x86_fp80, x86_fp80 }* byval align 8 %z) nounwind { +%0 = type { x86_fp80, x86_fp80 } + +define void @ccosl(%0* noalias sret %agg.result, %0* byval align 8 %z) nounwind { entry: - %iz = alloca { x86_fp80, x86_fp80 } ; <{ x86_fp80, x86_fp80 }*> [#uses=3] - %memtmp = alloca { x86_fp80, x86_fp80 }, align 16 ; <{ x86_fp80, x86_fp80 }*> [#uses=2] - %tmp1 = getelementptr { x86_fp80, x86_fp80 }* %z, i32 0, i32 1 ; <x86_fp80*> [#uses=1] - %tmp2 = load x86_fp80* %tmp1, align 16 ; <x86_fp80> [#uses=1] - %tmp3 = fsub x86_fp80 0xK80000000000000000000, %tmp2 ; <x86_fp80> [#uses=1] - %tmp4 = getelementptr { x86_fp80, x86_fp80 }* %iz, i32 0, i32 1 ; <x86_fp80*> [#uses=1] - %real = getelementptr { x86_fp80, x86_fp80 }* %iz, i32 0, i32 0 ; <x86_fp80*> [#uses=1] - %tmp7 = getelementptr { x86_fp80, x86_fp80 }* %z, i32 0, i32 0 ; <x86_fp80*> [#uses=1] - %tmp8 = load x86_fp80* %tmp7, align 16 ; <x86_fp80> [#uses=1] - store x86_fp80 %tmp3, x86_fp80* %real, align 16 - store x86_fp80 %tmp8, x86_fp80* %tmp4, align 16 - call void @ccoshl( { x86_fp80, x86_fp80 }* noalias sret %memtmp, { x86_fp80, x86_fp80 }* byval align 8 %iz ) nounwind - %memtmp14 = bitcast { x86_fp80, x86_fp80 }* %memtmp to i8* ; <i8*> [#uses=1] - %agg.result15 = bitcast { x86_fp80, x86_fp80 }* %agg.result to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %agg.result15, i8* %memtmp14, i32 32, i32 16 ) - ret void + %iz = alloca %0 + %memtmp = alloca %0, align 16 + %tmp1 = getelementptr %0* %z, i32 0, i32 1 + %tmp2 = load x86_fp80* %tmp1, align 16 + %tmp3 = fsub x86_fp80 0xK80000000000000000000, %tmp2 + %tmp4 = getelementptr %0* %iz, i32 0, i32 1 + %real = getelementptr %0* %iz, i32 0, i32 0 + %tmp7 = getelementptr %0* %z, i32 0, i32 0 + %tmp8 = load x86_fp80* %tmp7, align 16 + store x86_fp80 %tmp3, x86_fp80* %real, align 16 + store x86_fp80 %tmp8, x86_fp80* %tmp4, align 16 + call void @ccoshl(%0* noalias sret %memtmp, %0* byval align 8 %iz) nounwind + %memtmp14 = bitcast %0* %memtmp to i8* + %agg.result15 = bitcast %0* %agg.result to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %agg.result15, i8* %memtmp14, i32 32, i32 16, i1 false) + ret void } -declare void @ccoshl({ x86_fp80, x86_fp80 }* noalias sret , { x86_fp80, x86_fp80 }* byval ) nounwind +declare void @ccoshl(%0* noalias sret, %0* byval) nounwind -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/ScalarRepl/2007-05-24-LargeAggregate.ll b/test/Transforms/ScalarRepl/2007-05-24-LargeAggregate.ll deleted file mode 100644 index e67b610..0000000 --- a/test/Transforms/ScalarRepl/2007-05-24-LargeAggregate.ll +++ /dev/null @@ -1,27 +0,0 @@ -; RUN: opt < %s -scalarrepl -S | grep {alloca.*client_t} -; PR1446 -target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64" -target triple = "i686-pc-linux-gnu" - - %struct.clientSnapshot_t = type { i32, [32 x i8], %struct.playerState_t, i32, i32, i32, i32, i32 } - %struct.client_t = type { i32, [1024 x i8], [64 x [1024 x i8]], i32, i32, i32, i32, i32, i32, %struct.usercmd_t, i32, i32, [1024 x i8], %struct.sharedEntity_t*, [32 x i8], [64 x i8], i32, i32, i32, i32, i32, i32, [8 x i8*], [8 x i32], i32, i32, i32, i32, i32, i32, i32, i32, i32, [32 x %struct.clientSnapshot_t], i32, i32, i32, i32, i32, %struct.netchan_t, %struct.netchan_buffer_t*, %struct.netchan_buffer_t**, i32, [1025 x i32] } - %struct.entityShared_t = type { %struct.entityState_t, i32, i32, i32, i32, i32, [3 x float], [3 x float], i32, [3 x float], [3 x float], [3 x float], [3 x float], i32 } - %struct.entityState_t = type { i32, i32, i32, %struct.trajectory_t, %struct.trajectory_t, i32, i32, [3 x float], [3 x float], [3 x float], [3 x float], i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32, i32 } - %struct.msg_t = type { i32, i32, i32, i8*, i32, i32, i32, i32 } - %struct.netadr_t = type { i32, [4 x i8], [10 x i8], i16 } - %struct.netchan_buffer_t = type { %struct.msg_t, [16384 x i8], %struct.netchan_buffer_t* } - %struct.netchan_t = type { i32, i32, %struct.netadr_t, i32, i32, i32, i32, i32, [16384 x i8], i32, i32, i32, [16384 x i8] } - %struct.playerState_t = type { i32, i32, i32, i32, i32, [3 x float], [3 x float], i32, i32, i32, [3 x i32], i32, i32, i32, i32, i32, i32, [3 x float], i32, i32, [2 x i32], [2 x i32], i32, i32, i32, i32, i32, i32, [3 x float], i32, i32, i32, i32, i32, [16 x i32], [16 x i32], [16 x i32], [16 x i32], i32, i32, i32, i32, i32, i32, i32 } - %struct.sharedEntity_t = type { %struct.entityState_t, %struct.entityShared_t } - %struct.trajectory_t = type { i32, i32, i32, [3 x float], [3 x float] } - %struct.usercmd_t = type { i32, [3 x i32], i32, i8, i8, i8, i8 } - -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) - -define void @SV_DirectConnect(i64 %from.0.0, i64 %from.0.1, i32 %from.1) { -entry: - %temp = alloca %struct.client_t, align 16 ; <%struct.client_t*> [#uses=1] - %temp586 = bitcast %struct.client_t* %temp to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* null, i8* %temp586, i32 121596, i32 0 ) - unreachable -} diff --git a/test/Transforms/ScalarRepl/2007-05-29-MemcpyPreserve.ll b/test/Transforms/ScalarRepl/2007-05-29-MemcpyPreserve.ll index f1b8b80..cf96c4c 100644 --- a/test/Transforms/ScalarRepl/2007-05-29-MemcpyPreserve.ll +++ b/test/Transforms/ScalarRepl/2007-05-29-MemcpyPreserve.ll @@ -3,21 +3,22 @@ target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64" target triple = "i686-apple-darwin8" - %struct.LongestMember = type { i8, i32 } - %struct.MyString = type { i32 } - %struct.UnionType = type { %struct.LongestMember } + +%struct.LongestMember = type { i8, i32 } +%struct.MyString = type { i32 } +%struct.UnionType = type { %struct.LongestMember } define void @_Z4testP9UnionTypePS0_(%struct.UnionType* %p, %struct.UnionType** %pointerToUnion) { entry: - %tmp = alloca %struct.UnionType, align 8 ; <%struct.UnionType*> [#uses=2] - %tmp2 = getelementptr %struct.UnionType* %tmp, i32 0, i32 0, i32 0 ; <i8*> [#uses=1] - %tmp13 = getelementptr %struct.UnionType* %p, i32 0, i32 0, i32 0 ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %tmp2, i8* %tmp13, i32 8, i32 0 ) - %tmp5 = load %struct.UnionType** %pointerToUnion ; <%struct.UnionType*> [#uses=1] - %tmp56 = getelementptr %struct.UnionType* %tmp5, i32 0, i32 0, i32 0 ; <i8*> [#uses=1] - %tmp7 = getelementptr %struct.UnionType* %tmp, i32 0, i32 0, i32 0 ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %tmp56, i8* %tmp7, i32 8, i32 0 ) - ret void + %tmp = alloca %struct.UnionType, align 8 + %tmp2 = getelementptr %struct.UnionType* %tmp, i32 0, i32 0, i32 0 + %tmp13 = getelementptr %struct.UnionType* %p, i32 0, i32 0, i32 0 + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp2, i8* %tmp13, i32 8, i32 0, i1 false) + %tmp5 = load %struct.UnionType** %pointerToUnion + %tmp56 = getelementptr %struct.UnionType* %tmp5, i32 0, i32 0, i32 0 + %tmp7 = getelementptr %struct.UnionType* %tmp, i32 0, i32 0, i32 0 + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp56, i8* %tmp7, i32 8, i32 0, i1 false) + ret void } -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/ScalarRepl/2008-06-22-LargeArray.ll b/test/Transforms/ScalarRepl/2008-06-22-LargeArray.ll index b704727..71ba601 100644 --- a/test/Transforms/ScalarRepl/2008-06-22-LargeArray.ll +++ b/test/Transforms/ScalarRepl/2008-06-22-LargeArray.ll @@ -4,14 +4,14 @@ target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" target triple = "i386-apple-darwin8" -define void @memtest1(i8* %dst, i8* %src) nounwind { +define void @memtest1(i8* %dst, i8* %src) nounwind { entry: - %temp = alloca [200 x i8] ; <[100 x i8]*> [#uses=2] - %temp1 = bitcast [200 x i8]* %temp to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %temp1, i8* %src, i32 200, i32 1 ) - %temp3 = bitcast [200 x i8]* %temp to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %dst, i8* %temp3, i32 200, i32 1 ) - ret void + %temp = alloca [200 x i8] + %temp1 = bitcast [200 x i8]* %temp to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %temp1, i8* %src, i32 200, i32 1, i1 false) + %temp3 = bitcast [200 x i8]* %temp to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %dst, i8* %temp3, i32 200, i32 1, i1 false) + ret void } -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/ScalarRepl/2008-08-22-out-of-range-array-promote.ll b/test/Transforms/ScalarRepl/2008-08-22-out-of-range-array-promote.ll index 1df01c1..7cccb19 100644 --- a/test/Transforms/ScalarRepl/2008-08-22-out-of-range-array-promote.ll +++ b/test/Transforms/ScalarRepl/2008-08-22-out-of-range-array-promote.ll @@ -2,21 +2,22 @@ ; PR2423 target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128" target triple = "i386-apple-darwin8" - %struct.x = type { [1 x i32], i32, i32 } + +%struct.x = type { [1 x i32], i32, i32 } define i32 @b() nounwind { entry: - %s = alloca %struct.x ; <%struct.x*> [#uses=2] - %r = alloca %struct.x ; <%struct.x*> [#uses=2] - call i32 @a( %struct.x* %s ) nounwind ; <i32>:0 [#uses=0] - %r1 = bitcast %struct.x* %r to i8* ; <i8*> [#uses=1] - %s2 = bitcast %struct.x* %s to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %r1, i8* %s2, i32 12, i32 8 ) - getelementptr %struct.x* %r, i32 0, i32 0, i32 1 ; <i32*>:1 [#uses=1] - load i32* %1, align 4 ; <i32>:2 [#uses=1] - ret i32 %2 + %s = alloca %struct.x + %r = alloca %struct.x + %0 = call i32 @a(%struct.x* %s) nounwind + %r1 = bitcast %struct.x* %r to i8* + %s2 = bitcast %struct.x* %s to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %r1, i8* %s2, i32 12, i32 8, i1 false) + %1 = getelementptr %struct.x* %r, i32 0, i32 0, i32 1 + %2 = load i32* %1, align 4 + ret i32 %2 } declare i32 @a(%struct.x*) -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/ScalarRepl/2008-09-22-vector-gep.ll b/test/Transforms/ScalarRepl/2008-09-22-vector-gep.ll index e32e683..e7a58f1 100644 --- a/test/Transforms/ScalarRepl/2008-09-22-vector-gep.ll +++ b/test/Transforms/ScalarRepl/2008-09-22-vector-gep.ll @@ -18,8 +18,8 @@ entry: ; because the type of the first element in %struct.two is i8. %tmpS = getelementptr %struct.two* %S, i32 0, i32 0, i32 0 %tmpD = bitcast %struct.two* %D to i8* - call void @llvm.memmove.i32(i8* %tmpD, i8* %tmpS, i32 4, i32 1) + call void @llvm.memmove.p0i8.p0i8.i32(i8* %tmpD, i8* %tmpS, i32 4, i32 1, i1 false) ret void } -declare void @llvm.memmove.i32(i8*, i8*, i32, i32) nounwind +declare void @llvm.memmove.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind
\ No newline at end of file diff --git a/test/Transforms/ScalarRepl/2009-03-04-MemCpyAlign.ll b/test/Transforms/ScalarRepl/2009-03-04-MemCpyAlign.ll index 526457b..3218d59 100644 --- a/test/Transforms/ScalarRepl/2009-03-04-MemCpyAlign.ll +++ b/test/Transforms/ScalarRepl/2009-03-04-MemCpyAlign.ll @@ -12,9 +12,8 @@ entry: %0 = getelementptr %struct.st* %s, i32 0, i32 0 ; <i16*> [#uses=1] store i16 1, i16* %0, align 4 %s1 = bitcast %struct.st* %s to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32(i8* %p, i8* %s1, i32 2, i32 1) + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %p, i8* %s1, i32 2, i32 1, i1 false) ret void } -declare void @llvm.memcpy.i32(i8* nocapture, i8* nocapture, i32, i32) nounwind - +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/ScalarRepl/2009-12-11-NeonTypes.ll b/test/Transforms/ScalarRepl/2009-12-11-NeonTypes.ll index 31d9bae..1993e4f 100644 --- a/test/Transforms/ScalarRepl/2009-12-11-NeonTypes.ll +++ b/test/Transforms/ScalarRepl/2009-12-11-NeonTypes.ll @@ -13,54 +13,54 @@ define void @test(<8 x i16> %tmp.0, %struct.int16x8x2_t* %dst) nounwind { ; CHECK: @test ; CHECK-NOT: alloca ; CHECK: "alloca point" +; CHECK: store <8 x i16> +; CHECK: store <8 x i16> + entry: - %tmp_addr = alloca %struct.int16x8_t ; <%struct.int16x8_t*> [#uses=3] - %dst_addr = alloca %struct.int16x8x2_t* ; <%struct.int16x8x2_t**> [#uses=2] - %__rv = alloca %union..0anon ; <%union..0anon*> [#uses=2] - %__bx = alloca %struct.int16x8_t ; <%struct.int16x8_t*> [#uses=2] - %__ax = alloca %struct.int16x8_t ; <%struct.int16x8_t*> [#uses=2] - %tmp2 = alloca %struct.int16x8x2_t ; <%struct.int16x8x2_t*> [#uses=2] - %0 = alloca %struct.int16x8x2_t ; <%struct.int16x8x2_t*> [#uses=2] - %"alloca point" = bitcast i32 0 to i32 ; <i32> [#uses=0] - %1 = getelementptr inbounds %struct.int16x8_t* %tmp_addr, i32 0, i32 0 ; <<8 x i16>*> [#uses=1] + %tmp_addr = alloca %struct.int16x8_t + %dst_addr = alloca %struct.int16x8x2_t* + %__rv = alloca %union..0anon + %__bx = alloca %struct.int16x8_t + %__ax = alloca %struct.int16x8_t + %tmp2 = alloca %struct.int16x8x2_t + %0 = alloca %struct.int16x8x2_t + %"alloca point" = bitcast i32 0 to i32 + %1 = getelementptr inbounds %struct.int16x8_t* %tmp_addr, i32 0, i32 0 store <8 x i16> %tmp.0, <8 x i16>* %1 store %struct.int16x8x2_t* %dst, %struct.int16x8x2_t** %dst_addr - %2 = getelementptr inbounds %struct.int16x8_t* %__ax, i32 0, i32 0 ; <<8 x i16>*> [#uses=1] - %3 = getelementptr inbounds %struct.int16x8_t* %tmp_addr, i32 0, i32 0 ; <<8 x i16>*> [#uses=1] - %4 = load <8 x i16>* %3, align 16 ; <<8 x i16>> [#uses=1] + %2 = getelementptr inbounds %struct.int16x8_t* %__ax, i32 0, i32 0 + %3 = getelementptr inbounds %struct.int16x8_t* %tmp_addr, i32 0, i32 0 + %4 = load <8 x i16>* %3, align 16 store <8 x i16> %4, <8 x i16>* %2, align 16 - %5 = getelementptr inbounds %struct.int16x8_t* %__bx, i32 0, i32 0 ; <<8 x i16>*> [#uses=1] - %6 = getelementptr inbounds %struct.int16x8_t* %tmp_addr, i32 0, i32 0 ; <<8 x i16>*> [#uses=1] - %7 = load <8 x i16>* %6, align 16 ; <<8 x i16>> [#uses=1] + %5 = getelementptr inbounds %struct.int16x8_t* %__bx, i32 0, i32 0 + %6 = getelementptr inbounds %struct.int16x8_t* %tmp_addr, i32 0, i32 0 + %7 = load <8 x i16>* %6, align 16 store <8 x i16> %7, <8 x i16>* %5, align 16 - %8 = getelementptr inbounds %struct.int16x8_t* %__ax, i32 0, i32 0 ; <<8 x i16>*> [#uses=1] - %9 = load <8 x i16>* %8, align 16 ; <<8 x i16>> [#uses=2] - %10 = getelementptr inbounds %struct.int16x8_t* %__bx, i32 0, i32 0 ; <<8 x i16>*> [#uses=1] - %11 = load <8 x i16>* %10, align 16 ; <<8 x i16>> [#uses=2] - %12 = getelementptr inbounds %union..0anon* %__rv, i32 0, i32 0 ; <%struct.int16x8x2_t*> [#uses=1] - %13 = bitcast %struct.int16x8x2_t* %12 to %struct.__neon_int16x8x2_t* ; <%struct.__neon_int16x8x2_t*> [#uses=2] - %14 = shufflevector <8 x i16> %9, <8 x i16> %11, <8 x i32> <i32 0, i32 8, i32 2, i32 10, i32 4, i32 12, i32 6, i32 14> ; <<8 x i16>> [#uses=1] - %15 = getelementptr inbounds %struct.__neon_int16x8x2_t* %13, i32 0, i32 0 ; <<8 x i16>*> [#uses=1] + %8 = getelementptr inbounds %struct.int16x8_t* %__ax, i32 0, i32 0 + %9 = load <8 x i16>* %8, align 16 + %10 = getelementptr inbounds %struct.int16x8_t* %__bx, i32 0, i32 0 + %11 = load <8 x i16>* %10, align 16 + %12 = getelementptr inbounds %union..0anon* %__rv, i32 0, i32 0 + %13 = bitcast %struct.int16x8x2_t* %12 to %struct.__neon_int16x8x2_t* + %14 = shufflevector <8 x i16> %9, <8 x i16> %11, <8 x i32> <i32 0, i32 8, i32 2, i32 10, i32 4, i32 12, i32 6, i32 14> + %15 = getelementptr inbounds %struct.__neon_int16x8x2_t* %13, i32 0, i32 0 store <8 x i16> %14, <8 x i16>* %15 - %16 = shufflevector <8 x i16> %9, <8 x i16> %11, <8 x i32> <i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15> ; <<8 x i16>> [#uses=1] - %17 = getelementptr inbounds %struct.__neon_int16x8x2_t* %13, i32 0, i32 1 ; <<8 x i16>*> [#uses=1] + %16 = shufflevector <8 x i16> %9, <8 x i16> %11, <8 x i32> <i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15> + %17 = getelementptr inbounds %struct.__neon_int16x8x2_t* %13, i32 0, i32 1 store <8 x i16> %16, <8 x i16>* %17 - %18 = getelementptr inbounds %union..0anon* %__rv, i32 0, i32 0 ; <%struct.int16x8x2_t*> [#uses=1] - %19 = bitcast %struct.int16x8x2_t* %0 to i8* ; <i8*> [#uses=1] - %20 = bitcast %struct.int16x8x2_t* %18 to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32(i8* %19, i8* %20, i32 32, i32 16) - %tmp21 = bitcast %struct.int16x8x2_t* %tmp2 to i8* ; <i8*> [#uses=1] - %21 = bitcast %struct.int16x8x2_t* %0 to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32(i8* %tmp21, i8* %21, i32 32, i32 16) - %22 = load %struct.int16x8x2_t** %dst_addr, align 4 ; <%struct.int16x8x2_t*> [#uses=1] - %23 = bitcast %struct.int16x8x2_t* %22 to i8* ; <i8*> [#uses=1] - %tmp22 = bitcast %struct.int16x8x2_t* %tmp2 to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32(i8* %23, i8* %tmp22, i32 32, i32 16) + %18 = getelementptr inbounds %union..0anon* %__rv, i32 0, i32 0 + %19 = bitcast %struct.int16x8x2_t* %0 to i8* + %20 = bitcast %struct.int16x8x2_t* %18 to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %19, i8* %20, i32 32, i32 16, i1 false) + %tmp21 = bitcast %struct.int16x8x2_t* %tmp2 to i8* + %21 = bitcast %struct.int16x8x2_t* %0 to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp21, i8* %21, i32 32, i32 16, i1 false) + %22 = load %struct.int16x8x2_t** %dst_addr, align 4 + %23 = bitcast %struct.int16x8x2_t* %22 to i8* + %tmp22 = bitcast %struct.int16x8x2_t* %tmp2 to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %23, i8* %tmp22, i32 32, i32 16, i1 false) br label %return -; CHECK: store <8 x i16> -; CHECK: store <8 x i16> - return: ; preds = %entry ret void } @@ -69,21 +69,22 @@ return: ; preds = %entry %struct._NSRange = type { i64 } define void @test_memcpy_self() nounwind { -; CHECK: @test_memcpy_self -; CHECK-NOT: alloca -; CHECK: br i1 entry: - %range = alloca %struct._NSRange ; <%struct._NSRange*> [#uses=2] + %range = alloca %struct._NSRange br i1 undef, label %cond.true, label %cond.false cond.true: ; preds = %entry - %tmp3 = bitcast %struct._NSRange* %range to i8* ; <i8*> [#uses=1] - %tmp4 = bitcast %struct._NSRange* %range to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32(i8* %tmp3, i8* %tmp4, i32 8, i32 8) + %tmp3 = bitcast %struct._NSRange* %range to i8* + %tmp4 = bitcast %struct._NSRange* %range to i8* + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp3, i8* %tmp4, i32 8, i32 8, i1 false) ret void cond.false: ; preds = %entry ret void + +; CHECK: @test_memcpy_self +; CHECK-NOT: alloca +; CHECK: br i1 } -declare void @llvm.memcpy.i32(i8* nocapture, i8* nocapture, i32, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/ScalarRepl/2010-01-18-SelfCopy.ll b/test/Transforms/ScalarRepl/2010-01-18-SelfCopy.ll index 3aee399..52df6d5 100644 --- a/test/Transforms/ScalarRepl/2010-01-18-SelfCopy.ll +++ b/test/Transforms/ScalarRepl/2010-01-18-SelfCopy.ll @@ -3,7 +3,7 @@ target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128-n8:16:32" -%struct.test = type { [3 x double ] } +%struct.test = type { [3 x double] } define void @test_memcpy_self() nounwind { ; CHECK: @test_memcpy_self @@ -11,8 +11,8 @@ define void @test_memcpy_self() nounwind { ; CHECK: ret void %1 = alloca %struct.test %2 = bitcast %struct.test* %1 to i8* - call void @llvm.memcpy.i32(i8* %2, i8* %2, i32 24, i32 4) + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %2, i8* %2, i32 24, i32 4, i1 false) ret void } -declare void @llvm.memcpy.i32(i8* nocapture, i8* nocapture, i32, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind diff --git a/test/Transforms/ScalarRepl/badarray.ll b/test/Transforms/ScalarRepl/badarray.ll index 3ec3c01..768fec6 100644 --- a/test/Transforms/ScalarRepl/badarray.ll +++ b/test/Transforms/ScalarRepl/badarray.ll @@ -48,10 +48,10 @@ entry: %callret = call %padded *@test3f() ; <i32> [#uses=2] %callretcast = bitcast %padded* %callret to i8* ; <i8*> [#uses=1] %var_11 = bitcast %padded* %var_1 to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32(i8* %callretcast, i8* %var_11, i32 8, i32 4) + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %callretcast, i8* %var_11, i32 8, i32 4, i1 false) ret void } -declare void @llvm.memcpy.i32(i8* nocapture, i8* nocapture, i32, i32) nounwind +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind declare %padded* @test3f() diff --git a/test/Transforms/ScalarRepl/crash.ll b/test/Transforms/ScalarRepl/crash.ll index 83daaaf..e06d0c1 100644 --- a/test/Transforms/ScalarRepl/crash.ll +++ b/test/Transforms/ScalarRepl/crash.ll @@ -143,7 +143,6 @@ entry: %struct.anon = type { %struct.aal_spanarray_t } -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) define fastcc void @test7() { entry: @@ -158,7 +157,7 @@ cond_next114.i: ; preds = %cond_true cond_next: ; preds = %cond_true %SB19 = bitcast %struct.aal_spanbucket_t* %SB to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %SB19, i8* null, i32 12, i32 0 ) + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %SB19, i8* null, i32 12, i32 0, i1 false) br i1 false, label %cond_next34, label %cond_next79 cond_next34: ; preds = %cond_next @@ -196,7 +195,7 @@ entry: %.compoundliteral = alloca %0 %tmp228 = getelementptr %0* %.compoundliteral, i32 0, i32 7 %tmp229 = bitcast [0 x i16]* %tmp228 to i8* - call void @llvm.memset.i64(i8* %tmp229, i8 0, i64 0, i32 2) + call void @llvm.memset.p0i8.i64(i8* %tmp229, i8 0, i64 0, i32 2, i1 false) unreachable } @@ -260,3 +259,6 @@ entry: call void %0() nounwind ret void } + +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind +declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) nounwind diff --git a/test/Transforms/ScalarRepl/memset-aggregate.ll b/test/Transforms/ScalarRepl/memset-aggregate.ll index 5aeefcd..42e7a0f 100644 --- a/test/Transforms/ScalarRepl/memset-aggregate.ll +++ b/test/Transforms/ScalarRepl/memset-aggregate.ll @@ -14,31 +14,29 @@ entry: %L = alloca %struct.foo, align 8 ; <%struct.foo*> [#uses=2] %L2 = bitcast %struct.foo* %L to i8* ; <i8*> [#uses=1] %tmp13 = bitcast %struct.foo* %P to i8* ; <i8*> [#uses=1] - call void @llvm.memcpy.i32( i8* %L2, i8* %tmp13, i32 8, i32 4 ) + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %L2, i8* %tmp13, i32 8, i32 4, i1 false) %tmp4 = getelementptr %struct.foo* %L, i32 0, i32 0 ; <i32*> [#uses=1] %tmp5 = load i32* %tmp4 ; <i32> [#uses=1] ret i32 %tmp5 } -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) define i32 @test2() { entry: %L = alloca [4 x %struct.foo], align 16 ; <[4 x %struct.foo]*> [#uses=2] %L12 = bitcast [4 x %struct.foo]* %L to i8* ; <i8*> [#uses=1] - call void @llvm.memset.i32( i8* %L12, i8 0, i32 32, i32 16 ) + call void @llvm.memset.p0i8.i32(i8* %L12, i8 0, i32 32, i32 16, i1 false) %tmp4 = getelementptr [4 x %struct.foo]* %L, i32 0, i32 0, i32 0 ; <i32*> [#uses=1] %tmp5 = load i32* %tmp4 ; <i32> [#uses=1] ret i32 %tmp5 } -declare void @llvm.memset.i32(i8*, i8, i32, i32) define i32 @test3() { entry: %B = alloca %struct.bar, align 16 ; <%struct.bar*> [#uses=4] %B1 = bitcast %struct.bar* %B to i8* ; <i8*> [#uses=1] - call void @llvm.memset.i32( i8* %B1, i8 1, i32 24, i32 16 ) + call void @llvm.memset.p0i8.i32(i8* %B1, i8 1, i32 24, i32 16, i1 false) %tmp3 = getelementptr %struct.bar* %B, i32 0, i32 0, i32 0 ; <i32*> [#uses=1] store i32 1, i32* %tmp3 %tmp4 = getelementptr %struct.bar* %B, i32 0, i32 2 ; <double*> [#uses=1] @@ -58,9 +56,12 @@ entry: store i32 1, i32* %0, align 8 %1 = getelementptr %struct.f* %A, i32 0, i32 1 ; <i32*> [#uses=1] %2 = bitcast i32* %1 to i8* ; <i8*> [#uses=1] - call void @llvm.memset.i32(i8* %2, i8 2, i32 12, i32 4) + call void @llvm.memset.p0i8.i32(i8* %2, i8 2, i32 12, i32 4, i1 false) %3 = getelementptr %struct.f* %A, i32 0, i32 2 ; <i32*> [#uses=1] %4 = load i32* %3, align 8 ; <i32> [#uses=1] %retval12 = trunc i32 %4 to i16 ; <i16> [#uses=1] ret i16 %retval12 } +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind + +declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1) nounwind
\ No newline at end of file diff --git a/test/Transforms/SimplifyLibCalls/MemCpy.ll b/test/Transforms/SimplifyLibCalls/MemCpy.ll index 39662b1..c711178 100644 --- a/test/Transforms/SimplifyLibCalls/MemCpy.ll +++ b/test/Transforms/SimplifyLibCalls/MemCpy.ll @@ -4,17 +4,16 @@ @hel = constant [4 x i8] c"hel\00" ; <[4 x i8]*> [#uses=1] @hello_u = constant [8 x i8] c"hello_u\00" ; <[8 x i8]*> [#uses=1] -declare void @llvm.memcpy.i32(i8*, i8*, i32, i32) - define i32 @main() { - %h_p = getelementptr [2 x i8]* @h, i32 0, i32 0 ; <i8*> [#uses=1] - %hel_p = getelementptr [4 x i8]* @hel, i32 0, i32 0 ; <i8*> [#uses=1] - %hello_u_p = getelementptr [8 x i8]* @hello_u, i32 0, i32 0 ; <i8*> [#uses=1] - %target = alloca [1024 x i8] ; <[1024 x i8]*> [#uses=1] - %target_p = getelementptr [1024 x i8]* %target, i32 0, i32 0 ; <i8*> [#uses=3] - call void @llvm.memcpy.i32( i8* %target_p, i8* %h_p, i32 2, i32 2 ) - call void @llvm.memcpy.i32( i8* %target_p, i8* %hel_p, i32 4, i32 4 ) - call void @llvm.memcpy.i32( i8* %target_p, i8* %hello_u_p, i32 8, i32 8 ) - ret i32 0 + %h_p = getelementptr [2 x i8]* @h, i32 0, i32 0 + %hel_p = getelementptr [4 x i8]* @hel, i32 0, i32 0 + %hello_u_p = getelementptr [8 x i8]* @hello_u, i32 0, i32 0 + %target = alloca [1024 x i8] + %target_p = getelementptr [1024 x i8]* %target, i32 0, i32 0 + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %target_p, i8* %h_p, i32 2, i32 2, i1 false) + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %target_p, i8* %hel_p, i32 4, i32 4, i1 false) + call void @llvm.memcpy.p0i8.p0i8.i32(i8* %target_p, i8* %hello_u_p, i32 8, i32 8, i1 false) + ret i32 0 } +declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind |