[AMDGPU] Remove FeatureVGPRSpilling

[scott.linder]

Summary

This feature is only relevant to shaders, and is no longer used. When disabled, lowering of reserved registers for shaders causes a compiler crash.

Remove the feature and add a test for compilation of shaders at OptNone.

[arsenm]

LGTM

[scott.linder]

Remove unused AMDGPUSubtarget::EnableVGPRSpilling; I missed this in the first patch.

[arsenm]

LGTM

llvm/trunk/lib/Target/AMDGPU/AMDGPU.td

 
 def FeatureEnableHugePrivateBuffer : SubtargetFeature<
   "huge-private-buffer",
   "EnableHugePrivateBuffer",
   "true",
   "Enable private/scratch buffer sizes greater than 128 GB"
 >;
 
-def FeatureVGPRSpilling : SubtargetFeature<"vgpr-spilling",
-  "EnableVGPRSpilling",
-  "true",
-  "Enable spilling of VGPRs to scratch memory"
->;
-
 def FeatureDumpCode : SubtargetFeature <"DumpCode",
   "DumpCode",
   "true",
   "Dump MachineInstrs in the CodeEmitter"
 >;
 
 def FeatureDumpCodeLower : SubtargetFeature <"dumpcode",
   "DumpCode",

llvm/trunk/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp

static unsigned getRsrcReg(CallingConv::ID CallConv) {
   case CallingConv::AMDGPU_GS: return R_00B228_SPI_SHADER_PGM_RSRC1_GS;
   case CallingConv::AMDGPU_VS: return R_00B128_SPI_SHADER_PGM_RSRC1_VS;
   case CallingConv::AMDGPU_PS: return R_00B028_SPI_SHADER_PGM_RSRC1_PS;
   }
 }
 
 void AMDGPUAsmPrinter::EmitProgramInfoSI(const MachineFunction &MF,
                                          const SIProgramInfo &CurrentProgramInfo) {
-  const GCNSubtarget &STM = MF.getSubtarget<GCNSubtarget>();
   const SIMachineFunctionInfo *MFI = MF.getInfo<SIMachineFunctionInfo>();
   unsigned RsrcReg = getRsrcReg(MF.getFunction().getCallingConv());
 
   if (AMDGPU::isCompute(MF.getFunction().getCallingConv())) {
     OutStreamer->EmitIntValue(R_00B848_COMPUTE_PGM_RSRC1, 4);
 
     OutStreamer->EmitIntValue(CurrentProgramInfo.ComputePGMRSrc1, 4);
 
     OutStreamer->EmitIntValue(R_00B84C_COMPUTE_PGM_RSRC2, 4);
     OutStreamer->EmitIntValue(CurrentProgramInfo.ComputePGMRSrc2, 4);
 
     OutStreamer->EmitIntValue(R_00B860_COMPUTE_TMPRING_SIZE, 4);
     OutStreamer->EmitIntValue(S_00B860_WAVESIZE(CurrentProgramInfo.ScratchBlocks), 4);
 
     // TODO: Should probably note flat usage somewhere. SC emits a "FlatPtr32 =
     // 0" comment but I don't see a corresponding field in the register spec.
   } else {
     OutStreamer->EmitIntValue(RsrcReg, 4);
     OutStreamer->EmitIntValue(S_00B028_VGPRS(CurrentProgramInfo.VGPRBlocks) |
                               S_00B028_SGPRS(CurrentProgramInfo.SGPRBlocks), 4);
-    if (STM.isVGPRSpillingEnabled(MF.getFunction())) {
-      OutStreamer->EmitIntValue(R_0286E8_SPI_TMPRING_SIZE, 4);
-      OutStreamer->EmitIntValue(S_0286E8_WAVESIZE(CurrentProgramInfo.ScratchBlocks), 4);
-    }
+    OutStreamer->EmitIntValue(
+        S_0286E8_WAVESIZE(CurrentProgramInfo.ScratchBlocks), 4);
   }
 
   if (MF.getFunction().getCallingConv() == CallingConv::AMDGPU_PS) {
     OutStreamer->EmitIntValue(R_00B02C_SPI_SHADER_PGM_RSRC2_PS, 4);
     OutStreamer->EmitIntValue(S_00B02C_EXTRA_LDS_SIZE(CurrentProgramInfo.LDSBlocks), 4);
     OutStreamer->EmitIntValue(R_0286CC_SPI_PS_INPUT_ENA, 4);
     OutStreamer->EmitIntValue(MFI->getPSInputEnable(), 4);
     OutStreamer->EmitIntValue(R_0286D0_SPI_PS_INPUT_ADDR, 4);

llvm/trunk/lib/Target/AMDGPU/AMDGPUSubtarget.h

protected:
   bool HasApertureRegs;
   bool EnableXNACK;
   bool TrapHandler;
   bool DebuggerInsertNops;
   bool DebuggerEmitPrologue;
 
   // Used as options.
   bool EnableHugePrivateBuffer;
-  bool EnableVGPRSpilling;
   bool EnableLoadStoreOpt;
   bool EnableUnsafeDSOffsetFolding;
   bool EnableSIScheduler;
   bool EnableDS128;
   bool DumpCode;
 
   // Subtarget statically properties set by tablegen
   bool FP64;
public:
   // XXX - Why is this here if it isn't in the default pass set?
   bool enableEarlyIfConversion() const override {
     return true;
   }
 
   void overrideSchedPolicy(MachineSchedPolicy &Policy,
                            unsigned NumRegionInstrs) const override;
 
-  bool isVGPRSpillingEnabled(const Function &F) const;
-
   unsigned getMaxNumUserSGPRs() const {
     return 16;
   }
 
   bool hasSMemRealTime() const {
     return HasSMemRealTime;
   }
 

llvm/trunk/lib/Target/AMDGPU/AMDGPUSubtarget.cpp

GCNSubtarget::GCNSubtarget(const Triple &TT, StringRef GPU, StringRef FS,
 
     HasApertureRegs(false),
     EnableXNACK(false),
     TrapHandler(false),
     DebuggerInsertNops(false),
     DebuggerEmitPrologue(false),
 
     EnableHugePrivateBuffer(false),
-    EnableVGPRSpilling(false),
     EnableLoadStoreOpt(false),
     EnableUnsafeDSOffsetFolding(false),
     EnableSIScheduler(false),
     EnableDS128(false),
     DumpCode(false),
 
     FP64(false),
     GCN3Encoding(false),
void GCNSubtarget::overrideSchedPolicy(MachineSchedPolicy &Policy,
   Policy.OnlyTopDown = false;
   Policy.OnlyBottomUp = false;
 
   // Enabling ShouldTrackLaneMasks crashes the SI Machine Scheduler.
   if (!enableSIScheduler())
     Policy.ShouldTrackLaneMasks = true;
 }
 
-bool GCNSubtarget::isVGPRSpillingEnabled(const Function& F) const {
-  return EnableVGPRSpilling || !AMDGPU::isShader(F.getCallingConv());
-}
-
 unsigned GCNSubtarget::getOccupancyWithNumSGPRs(unsigned SGPRs) const {
   if (getGeneration() >= AMDGPUSubtarget::VOLCANIC_ISLANDS) {
     if (SGPRs <= 80)
       return 10;
     if (SGPRs <= 88)
       return 9;
     if (SGPRs <= 100)
       return 8;

llvm/trunk/lib/Target/AMDGPU/SIInstrInfo.cpp

if (RI.isSGPRClass(RC)) {
     if (ST.hasScalarStores()) {
       // m0 is used for offset to scalar stores if used to spill.
       Spill.addReg(AMDGPU::M0, RegState::ImplicitDefine | RegState::Dead);
     }
 
     return;
   }
 
-  if (!ST.isVGPRSpillingEnabled(MF->getFunction())) {
-    LLVMContext &Ctx = MF->getFunction().getContext();
-    Ctx.emitError("SIInstrInfo::storeRegToStackSlot - Do not know how to"
-                  " spill register");
-    BuildMI(MBB, MI, DL, get(AMDGPU::KILL))
-      .addReg(SrcReg);
-
-    return;
-  }
-
   assert(RI.hasVGPRs(RC) && "Only VGPR spilling expected");
 
   unsigned Opcode = getVGPRSpillSaveOpcode(SpillSize);
   MFI->setHasSpilledVGPRs();
   BuildMI(MBB, MI, DL, get(Opcode))
     .addReg(SrcReg, getKillRegState(isKill)) // data
     .addFrameIndex(FrameIndex)               // addr
     .addReg(MFI->getScratchRSrcReg())        // scratch_rsrc
if (RI.isSGPRClass(RC)) {
     if (ST.hasScalarStores()) {
       // m0 is used for offset to scalar stores if used to spill.
       Spill.addReg(AMDGPU::M0, RegState::ImplicitDefine | RegState::Dead);
     }
 
     return;
   }
 
-  if (!ST.isVGPRSpillingEnabled(MF->getFunction())) {
-    LLVMContext &Ctx = MF->getFunction().getContext();
-    Ctx.emitError("SIInstrInfo::loadRegFromStackSlot - Do not know how to"
-                  " restore register");
-    BuildMI(MBB, MI, DL, get(AMDGPU::IMPLICIT_DEF), DestReg);
-
-    return;
-  }
-
   assert(RI.hasVGPRs(RC) && "Only VGPR spilling expected");
 
   unsigned Opcode = getVGPRSpillRestoreOpcode(SpillSize);
   BuildMI(MBB, MI, DL, get(Opcode), DestReg)
     .addFrameIndex(FrameIndex)        // vaddr
     .addReg(MFI->getScratchRSrcReg()) // scratch_rsrc
     .addReg(MFI->getFrameOffsetReg()) // scratch_offset
     .addImm(0)                        // offset

llvm/trunk/lib/Target/AMDGPU/SIMachineFunctionInfo.cpp

if (ST.debuggerEmitPrologue()) {
     if (F.hasFnAttribute("amdgpu-work-item-id-y"))
       WorkItemIDY = true;
 
     if (F.hasFnAttribute("amdgpu-work-item-id-z"))
       WorkItemIDZ = true;
   }
 
   const MachineFrameInfo &FrameInfo = MF.getFrameInfo();
-  bool MaySpill = ST.isVGPRSpillingEnabled(F);
   bool HasStackObjects = FrameInfo.hasStackObjects();
 
   if (isEntryFunction()) {
     // X, XY, and XYZ are the only supported combinations, so make sure Y is
     // enabled if Z is.
     if (WorkItemIDZ)
       WorkItemIDY = true;
 
-    if (HasStackObjects || MaySpill) {
-      PrivateSegmentWaveByteOffset = true;
 
     // HS and GS always have the scratch wave offset in SGPR5 on GFX9.
     if (ST.getGeneration() >= AMDGPUSubtarget::GFX9 &&
         (CC == CallingConv::AMDGPU_HS || CC == CallingConv::AMDGPU_GS))
-      ArgInfo.PrivateSegmentWaveByteOffset
-        = ArgDescriptor::createRegister(AMDGPU::SGPR5);
-    }
+      ArgInfo.PrivateSegmentWaveByteOffset =
+          ArgDescriptor::createRegister(AMDGPU::SGPR5);
   }
 
   bool isAmdHsaOrMesa = ST.isAmdHsaOrMesa(F);
   if (isAmdHsaOrMesa) {
-    if (HasStackObjects || MaySpill)
-      PrivateSegmentBuffer = true;
 
     if (F.hasFnAttribute("amdgpu-dispatch-ptr"))
       DispatchPtr = true;
 
     if (F.hasFnAttribute("amdgpu-queue-ptr"))
       QueuePtr = true;
 
     if (F.hasFnAttribute("amdgpu-dispatch-id"))
       DispatchID = true;
   } else if (ST.isMesaGfxShader(F)) {
-    if (HasStackObjects || MaySpill)
-      ImplicitBufferPtr = true;
   }
 
   if (F.hasFnAttribute("amdgpu-kernarg-segment-ptr"))
     KernargSegmentPtr = true;
 
   if (ST.hasFlatAddressSpace() && isEntryFunction() && isAmdHsaOrMesa) {
     // TODO: This could be refined a lot. The attribute is a poor way of
     // detecting calls that may require it before argument lowering.

llvm/trunk/test/CodeGen/AMDGPU/local-stack-slot-offset.ll

-;RUN: llc < %s -march=amdgcn -mcpu=verde -mattr=+vgpr-spilling -mattr=-promote-alloca -verify-machineinstrs | FileCheck %s -check-prefix=CHECK
-;RUN: llc < %s -march=amdgcn -mcpu=tonga -mattr=-flat-for-global -mattr=+vgpr-spilling -mattr=-promote-alloca -verify-machineinstrs | FileCheck %s -check-prefix=CHECK
+;RUN: llc < %s -march=amdgcn -mcpu=verde -mattr=-promote-alloca -verify-machineinstrs | FileCheck %s -check-prefix=CHECK
+;RUN: llc < %s -march=amdgcn -mcpu=tonga -mattr=-flat-for-global -mattr=-promote-alloca -verify-machineinstrs | FileCheck %s -check-prefix=CHECK
 
 ; Allocate two stack slots of 2052 bytes each requiring a total of 4104 bytes.
 ; Extracting the last element of each does not fit into the offset field of
 ; MUBUF instructions, so a new base register is needed. This used to not
 ; happen, leading to an assertion.
 
 ; CHECK-LABEL: {{^}}main:
 ; CHECK: buffer_store_dword

llvm/trunk/test/CodeGen/AMDGPU/noop-shader-O0.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -O0 -mtriple=amdgcn-amd-amdpal -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+
+; Ensure NOOP shaders compile at OptNone.
+
+; Confirm registers reserved in SIMachineFunctionInfo are those expected during
+; lowering, even when e.g. spilling is required due to being at OptNone.
+
+target datalayout = "e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5"
+target triple = "amdgcn-amd-amdpal"
+
+define amdgpu_vs void @noop_vs() {
+; GCN-LABEL: noop_vs:
+; GCN:       ; %bb.0: ; %entry
+; GCN-NEXT:    s_endpgm
+entry:
+  ret void
+}
+
+define amdgpu_ls void @noop_ls() {
+; GCN-LABEL: noop_ls:
+; GCN:       ; %bb.0: ; %entry
+; GCN-NEXT:    s_endpgm
+entry:
+  ret void
+}
+
+define amdgpu_hs void @noop_hs() {
+; GCN-LABEL: noop_hs:
+; GCN:       ; %bb.0: ; %entry
+; GCN-NEXT:    s_endpgm
+entry:
+  ret void
+}
+
+define amdgpu_es void @noop_es() {
+; GCN-LABEL: noop_es:
+; GCN:       ; %bb.0: ; %entry
+; GCN-NEXT:    s_endpgm
+entry:
+  ret void
+}
+
+define amdgpu_gs void @noop_gs() {
+; GCN-LABEL: noop_gs:
+; GCN:       ; %bb.0: ; %entry
+; GCN-NEXT:    s_endpgm
+entry:
+  ret void
+}
+
+define amdgpu_ps void @noop_ps() {
+; GCN-LABEL: noop_ps:
+; GCN:       ; %bb.0: ; %entry
+; GCN-NEXT:    s_endpgm
+entry:
+  ret void
+}
+
+define amdgpu_cs void @noop_cs() {
+; GCN-LABEL: noop_cs:
+; GCN:       ; %bb.0: ; %entry
+; GCN-NEXT:    s_endpgm
+entry:
+  ret void
+}

llvm/trunk/test/CodeGen/AMDGPU/schedule-vs-if-nested-loop-failure.ll

-; RUN: llc -O0 -march=amdgcn -verify-machineinstrs -mattr=+vgpr-spilling < %s | FileCheck -check-prefix=GCN %s
-; RUN: llc -O0 -march=amdgcn -mcpu=tonga -verify-machineinstrs -mattr=+vgpr-spilling < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -O0 -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -O0 -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
 
 declare void @llvm.amdgcn.s.barrier() nounwind convergent
 
 ; GCN-LABEL: {{^}}main:
 define amdgpu_vs void @main(<4 x float> inreg %reg0, <4 x float> inreg %reg1) {
 main_body:
   %0 = extractelement <4 x float> %reg1, i32 0
   %1 = extractelement <4 x float> %reg1, i32 2

llvm/trunk/test/CodeGen/AMDGPU/scratch-simple.ll

-; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=verde -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=SI %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx803 -mattr=-flat-for-global -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=SI %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -mattr=-flat-for-global -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX9 %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=verde -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=SI %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx803 -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=SI %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX9 %s
 
 ; This used to fail due to a v_add_i32 instruction with an illegal immediate
 ; operand that was created during Local Stack Slot Allocation. Test case derived
 ; from https://bugs.freedesktop.org/show_bug.cgi?id=96602
 ;
 ; GCN-LABEL: {{^}}ps_main:
 
 ; GCN-DAG: s_mov_b32 s4, SCRATCH_RSRC_DWORD0

llvm/trunk/test/CodeGen/AMDGPU/selected-stack-object.ll

-; "Assertion failure" should be caught with both XFAIL * and +Asserts.
-; XFAIL: *
-; REQUIRES: asserts
-
-; RUN: llc -march=amdgcn -mcpu=verde -verify-machineinstrs < %s | FileCheck %s
-
-; See also local-stack-slot-bug.ll
-; This fails because a stack object is created during instruction selection.
-
-; CHECK-LABEL: {{^}}main:
-define amdgpu_ps float @main(i32 %idx) {
-main_body:
-  %v1 = extractelement <81 x float> <float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float undef, float 0x3FE41CFEA0000000, float 0xBFE7A693C0000000, float 0xBFEA477C60000000, float 0xBFEBE5DC60000000, float 0xBFEC71C720000000, float 0xBFEBE5DC60000000, float 0xBFEA477C60000000, float 0xBFE7A693C0000000, float 0xBFE41CFEA0000000, float 0x3FDF9B13E0000000, float 0x3FDF9B1380000000, float 0x3FD5C53B80000000, float 0x3FD5C53B00000000, float 0x3FC6326AC0000000, float 0x3FC63269E0000000, float 0xBEE05CEB00000000, float 0xBEE086A320000000, float 0xBFC63269E0000000, float 0xBFC6326AC0000000, float 0xBFD5C53B80000000, float 0xBFD5C53B80000000, float 0xBFDF9B13E0000000, float 0xBFDF9B1460000000, float 0xBFE41CFE80000000, float 0x3FE7A693C0000000, float 0x3FEA477C20000000, float 0x3FEBE5DC40000000, float 0x3FEC71C6E0000000, float 0x3FEBE5DC40000000, float 0x3FEA477C20000000, float 0x3FE7A693C0000000, float 0xBFE41CFE80000000>, i32 %idx
-  ret float %v1
-}

llvm/trunk/test/CodeGen/AMDGPU/si-sgpr-spill.ll

-; RUN: llc -march=amdgcn -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=TOVGPR %s
-; RUN: llc -march=amdgcn -mcpu=tonga -mattr=+vgpr-spilling,-mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=TOVGPR %s
+; RUN: llc -march=amdgcn -mcpu=tonga -mattr=-mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
 
 ; These tests check that the compiler won't crash when it needs to spill
 ; SGPRs.
 
 @ddxy_lds = external addrspace(3) global [64 x i32]
 
 ; GCN-LABEL: {{^}}main:
 ; GCN: s_wqm

llvm/trunk/test/CodeGen/AMDGPU/spill-m0.ll

-; RUN: llc -O0 -amdgpu-spill-sgpr-to-vgpr=1 -march=amdgcn -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=TOVGPR -check-prefix=GCN %s
-; RUN: llc -O0 -amdgpu-spill-sgpr-to-vgpr=1 -amdgpu-spill-sgpr-to-smem=0 -march=amdgcn -mcpu=tonga -mattr=+vgpr-spilling  -verify-machineinstrs < %s | FileCheck -check-prefix=TOVGPR -check-prefix=GCN %s
-; RUN: llc -O0 -amdgpu-spill-sgpr-to-vgpr=0 -march=amdgcn -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=TOVMEM -check-prefix=GCN %s
-; RUN: llc -O0 -amdgpu-spill-sgpr-to-vgpr=0 -amdgpu-spill-sgpr-to-smem=0 -march=amdgcn -mcpu=tonga -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=TOVMEM -check-prefix=GCN %s
-; RUN: llc -O0 -amdgpu-spill-sgpr-to-vgpr=0 -amdgpu-spill-sgpr-to-smem=1 -march=amdgcn -mcpu=tonga -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=TOSMEM -check-prefix=GCN %s
+; RUN: llc -O0 -amdgpu-spill-sgpr-to-vgpr=1 -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=TOVGPR -check-prefix=GCN %s
+; RUN: llc -O0 -amdgpu-spill-sgpr-to-vgpr=1 -amdgpu-spill-sgpr-to-smem=0 -march=amdgcn -mcpu=tonga  -verify-machineinstrs < %s | FileCheck -check-prefix=TOVGPR -check-prefix=GCN %s
+; RUN: llc -O0 -amdgpu-spill-sgpr-to-vgpr=0 -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=TOVMEM -check-prefix=GCN %s
+; RUN: llc -O0 -amdgpu-spill-sgpr-to-vgpr=0 -amdgpu-spill-sgpr-to-smem=0 -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=TOVMEM -check-prefix=GCN %s
+; RUN: llc -O0 -amdgpu-spill-sgpr-to-vgpr=0 -amdgpu-spill-sgpr-to-smem=1 -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=TOSMEM -check-prefix=GCN %s
 
 ; XXX - Why does it like to use vcc?
 
 ; GCN-LABEL: {{^}}spill_m0:
 ; TOSMEM: s_mov_b32 s[[LO:[0-9]+]], SCRATCH_RSRC_DWORD0
 ; TOSMEM: s_mov_b32 s[[HI:[0-9]+]], 0xe80000
 
 ; GCN-DAG: s_cmp_lg_u32

llvm/trunk/test/CodeGen/AMDGPU/vgpr-spill-emergency-stack-slot-compute.ll

-; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tahiti -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCNMESA -check-prefix=SIMESA %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=fiji -mattr=+vgpr-spilling,-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCNMESA -check-prefix=VIMESA %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -mattr=+vgpr-spilling,-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCNMESA -check-prefix=GFX9MESA %s
-; RUN: llc -march=amdgcn  -mcpu=hawaii -mtriple=amdgcn-unknown-amdhsa -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=CIHSA -check-prefix=HSA %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tahiti -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCNMESA -check-prefix=SIMESA %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=fiji -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCNMESA -check-prefix=VIMESA %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCNMESA -check-prefix=GFX9MESA %s
+; RUN: llc -march=amdgcn  -mcpu=hawaii -mtriple=amdgcn-unknown-amdhsa -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=CIHSA -check-prefix=HSA %s
 ; RUN: llc -march=amdgcn  -mcpu=fiji -mtriple=amdgcn-unknown-amdhsa -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VIHSA -check-prefix=HSA %s
 
 ; This ends up using all 256 registers and requires register
 ; scavenging which will fail to find an unsued register.
 
 ; Check the ScratchSize to avoid regressions from spilling
 ; intermediate register class copies.
 

llvm/trunk/test/CodeGen/AMDGPU/vgpr-spill-emergency-stack-slot.ll

-; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tahiti -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SI %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=fiji -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
-; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -mattr=+vgpr-spilling -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GFX9 %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=tahiti -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SI %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=fiji -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
+; RUN: llc -march=amdgcn -mtriple=amdgcn-- -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GFX9 %s
 
 ; This ends up using all 255 registers and requires register
 ; scavenging which will fail to find an unsued register.
 
 ; Check the ScratchSize to avoid regressions from spilling
 ; intermediate register class copies.
 
 ; FIXME: The same register is initialized to 0 for every spill.