Vendor import of llvm-project main llvmorg-15-init-15358-g53dc0f107877.vendor/llvm-project/llvmorg-15-init-15358-g53dc0f107877

1 files changed, 9 insertions, 149 deletions

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.cpp b/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.cpp
index e82f9232b114..77816a783630 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.cpp
@@ -50,11 +50,6 @@ static cl::opt<bool> EnableVGPRIndexMode(
   cl::desc("Use GPR indexing mode instead of movrel for vector indexing"),
   cl::init(false));
 
-static cl::opt<bool> EnableFlatScratch(
-  "amdgpu-enable-flat-scratch",
-  cl::desc("Use flat scratch instructions"),
-  cl::init(false));
-
 static cl::opt<bool> UseAA("amdgpu-use-aa-in-codegen",
                            cl::desc("Enable the use of AA during codegen."),
                            cl::init(true));
@@ -159,26 +154,7 @@ GCNSubtarget::initializeSubtargetDependencies(const Triple &TT,
   return *this;
 }
 
-AMDGPUSubtarget::AMDGPUSubtarget(const Triple &TT) :
-  TargetTriple(TT),
-  GCN3Encoding(false),
-  Has16BitInsts(false),
-  HasMadMixInsts(false),
-  HasMadMacF32Insts(false),
-  HasDsSrc2Insts(false),
-  HasSDWA(false),
-  HasVOP3PInsts(false),
-  HasMulI24(true),
-  HasMulU24(true),
-  HasSMulHi(false),
-  HasInv2PiInlineImm(false),
-  HasFminFmaxLegacy(true),
-  EnablePromoteAlloca(false),
-  HasTrigReducedRange(false),
-  MaxWavesPerEU(10),
-  LocalMemorySize(0),
-  WavefrontSizeLog2(0)
-  { }
+AMDGPUSubtarget::AMDGPUSubtarget(const Triple &TT) : TargetTriple(TT) {}
 
 GCNSubtarget::GCNSubtarget(const Triple &TT, StringRef GPU, StringRef FS,
                            const GCNTargetMachine &TM)
@@ -187,120 +163,7 @@ GCNSubtarget::GCNSubtarget(const Triple &TT, StringRef GPU, StringRef FS,
     AMDGPUSubtarget(TT),
     TargetTriple(TT),
     TargetID(*this),
-    Gen(INVALID),
     InstrItins(getInstrItineraryForCPU(GPU)),
-    LDSBankCount(0),
-    MaxPrivateElementSize(0),
-
-    FastFMAF32(false),
-    FastDenormalF32(false),
-    HalfRate64Ops(false),
-    FullRate64Ops(false),
-
-    FlatForGlobal(false),
-    AutoWaitcntBeforeBarrier(false),
-    UnalignedScratchAccess(false),
-    UnalignedAccessMode(false),
-
-    HasApertureRegs(false),
-    SupportsXNACK(false),
-    EnableXNACK(false),
-    EnableTgSplit(false),
-    EnableCuMode(false),
-    TrapHandler(false),
-
-    EnableLoadStoreOpt(false),
-    EnableUnsafeDSOffsetFolding(false),
-    EnableSIScheduler(false),
-    EnableDS128(false),
-    EnablePRTStrictNull(false),
-    DumpCode(false),
-
-    FP64(false),
-    CIInsts(false),
-    GFX8Insts(false),
-    GFX9Insts(false),
-    GFX90AInsts(false),
-    GFX10Insts(false),
-    GFX10_3Insts(false),
-    GFX7GFX8GFX9Insts(false),
-    SGPRInitBug(false),
-    NegativeScratchOffsetBug(false),
-    NegativeUnalignedScratchOffsetBug(false),
-    HasSMemRealTime(false),
-    HasIntClamp(false),
-    HasFmaMixInsts(false),
-    HasMovrel(false),
-    HasVGPRIndexMode(false),
-    HasScalarStores(false),
-    HasScalarAtomics(false),
-    HasSDWAOmod(false),
-    HasSDWAScalar(false),
-    HasSDWASdst(false),
-    HasSDWAMac(false),
-    HasSDWAOutModsVOPC(false),
-    HasDPP(false),
-    HasDPP8(false),
-    Has64BitDPP(false),
-    HasPackedFP32Ops(false),
-    HasExtendedImageInsts(false),
-    HasR128A16(false),
-    HasGFX10A16(false),
-    HasG16(false),
-    HasNSAEncoding(false),
-    NSAMaxSize(0),
-    GFX10_AEncoding(false),
-    GFX10_BEncoding(false),
-    HasDLInsts(false),
-    HasDot1Insts(false),
-    HasDot2Insts(false),
-    HasDot3Insts(false),
-    HasDot4Insts(false),
-    HasDot5Insts(false),
-    HasDot6Insts(false),
-    HasDot7Insts(false),
-    HasMAIInsts(false),
-    HasPkFmacF16Inst(false),
-    HasAtomicFaddInsts(false),
-    SupportsSRAMECC(false),
-    EnableSRAMECC(false),
-    HasNoSdstCMPX(false),
-    HasVscnt(false),
-    HasGetWaveIdInst(false),
-    HasSMemTimeInst(false),
-    HasShaderCyclesRegister(false),
-    HasVOP3Literal(false),
-    HasNoDataDepHazard(false),
-    FlatAddressSpace(false),
-    FlatInstOffsets(false),
-    FlatGlobalInsts(false),
-    FlatScratchInsts(false),
-    ScalarFlatScratchInsts(false),
-    HasArchitectedFlatScratch(false),
-    AddNoCarryInsts(false),
-    HasUnpackedD16VMem(false),
-    LDSMisalignedBug(false),
-    HasMFMAInlineLiteralBug(false),
-    UnalignedBufferAccess(false),
-    UnalignedDSAccess(false),
-    HasPackedTID(false),
-
-    ScalarizeGlobal(false),
-
-    HasVcmpxPermlaneHazard(false),
-    HasVMEMtoScalarWriteHazard(false),
-    HasSMEMtoVectorWriteHazard(false),
-    HasInstFwdPrefetchBug(false),
-    HasVcmpxExecWARHazard(false),
-    HasLdsBranchVmemWARHazard(false),
-    HasNSAtoVMEMBug(false),
-    HasNSAClauseBug(false),
-    HasOffset3fBug(false),
-    HasFlatSegmentOffsetBug(false),
-    HasImageStoreD16Bug(false),
-    HasImageGather4D16Bug(false),
-
-    FeatureDisable(false),
     InstrInfo(initializeSubtargetDependencies(TT, GPU, FS)),
     TLInfo(TM, *this),
     FrameLowering(TargetFrameLowering::StackGrowsUp, getStackAlignment(), 0) {
@@ -314,11 +177,6 @@ GCNSubtarget::GCNSubtarget(const Triple &TT, StringRef GPU, StringRef FS,
   *this, *static_cast<AMDGPURegisterBankInfo *>(RegBankInfo.get()), TM));
 }
 
-bool GCNSubtarget::enableFlatScratch() const {
-  return flatScratchIsArchitected() ||
-         (EnableFlatScratch && hasFlatScratchInsts());
-}
-
 unsigned GCNSubtarget::getConstantBusLimit(unsigned Opcode) const {
   if (getGeneration() < GFX10)
     return 1;
@@ -326,12 +184,15 @@ unsigned GCNSubtarget::getConstantBusLimit(unsigned Opcode) const {
   switch (Opcode) {
   case AMDGPU::V_LSHLREV_B64_e64:
   case AMDGPU::V_LSHLREV_B64_gfx10:
+  case AMDGPU::V_LSHLREV_B64_e64_gfx11:
   case AMDGPU::V_LSHL_B64_e64:
   case AMDGPU::V_LSHRREV_B64_e64:
   case AMDGPU::V_LSHRREV_B64_gfx10:
+  case AMDGPU::V_LSHRREV_B64_e64_gfx11:
   case AMDGPU::V_LSHR_B64_e64:
   case AMDGPU::V_ASHRREV_I64_e64:
   case AMDGPU::V_ASHRREV_I64_gfx10:
+  case AMDGPU::V_ASHRREV_I64_e64_gfx11:
   case AMDGPU::V_ASHR_I64_e64:
     return 1;
   }
@@ -658,7 +519,8 @@ unsigned AMDGPUSubtarget::getImplicitArgNumBytes(const Function &F) const {
     return 16;
 
   // Assume all implicit inputs are used by default
-  return AMDGPU::getIntegerAttribute(F, "amdgpu-implicitarg-num-bytes", 56);
+  unsigned NBytes = (AMDGPU::getAmdhsaCodeObjectVersion() >= 5) ? 256 : 56;
+  return AMDGPU::getIntegerAttribute(F, "amdgpu-implicitarg-num-bytes", NBytes);
 }
 
 uint64_t AMDGPUSubtarget::getExplicitKernArgSize(const Function &F,
@@ -673,13 +535,11 @@ uint64_t AMDGPUSubtarget::getExplicitKernArgSize(const Function &F,
   for (const Argument &Arg : F.args()) {
     const bool IsByRef = Arg.hasByRefAttr();
     Type *ArgTy = IsByRef ? Arg.getParamByRefType() : Arg.getType();
-    MaybeAlign Alignment = IsByRef ? Arg.getParamAlign() : None;
-    if (!Alignment)
-      Alignment = DL.getABITypeAlign(ArgTy);
-
+    Align Alignment = DL.getValueOrABITypeAlignment(
+        IsByRef ? Arg.getParamAlign() : None, ArgTy);
     uint64_t AllocSize = DL.getTypeAllocSize(ArgTy);
     ExplicitArgBytes = alignTo(ExplicitArgBytes, Alignment) + AllocSize;
-    MaxAlign = max(MaxAlign, Alignment);
+    MaxAlign = std::max(MaxAlign, Alignment);
   }
 
   return ExplicitArgBytes;