AMDGPU: Fold CI-specific complex SMRD patterns into existing complex patterns

This is just a cleanup. Also, it adds checking that ByteCount is aligned to 4.

Reviewers: arsenm, nhaehnle, tstellarAMD

Subscribers: kzhuravl, wdng, yaxunl, tony-tye

Differential Revision: https://reviews.llvm.org/D28994

llvm-svn: 303658

4 files changed, 18 insertions, 51 deletions

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp b/llvm/lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp
index 19fce064783..5eb2324c79c 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp
@@ -145,10 +145,8 @@ private:
   bool SelectSMRD(SDValue Addr, SDValue &SBase, SDValue &Offset,
                   bool &Imm) const;
   bool SelectSMRDImm(SDValue Addr, SDValue &SBase, SDValue &Offset) const;
-  bool SelectSMRDImm32(SDValue Addr, SDValue &SBase, SDValue &Offset) const;
   bool SelectSMRDSgpr(SDValue Addr, SDValue &SBase, SDValue &Offset) const;
   bool SelectSMRDBufferImm(SDValue Addr, SDValue &Offset) const;
-  bool SelectSMRDBufferImm32(SDValue Addr, SDValue &Offset) const;
   bool SelectSMRDBufferSgpr(SDValue Addr, SDValue &Offset) const;
   bool SelectMOVRELOffset(SDValue Index, SDValue &Base, SDValue &Offset) const;
 
@@ -1330,7 +1328,6 @@ bool AMDGPUDAGToDAGISel::SelectSMRDOffset(SDValue ByteOffsetNode,
     return false;
 
   SDLoc SL(ByteOffsetNode);
-  AMDGPUSubtarget::Generation Gen = Subtarget->getGeneration();
   int64_t ByteOffset = C->getSExtValue();
   int64_t EncodedOffset = AMDGPU::getSMRDEncodedOffset(*Subtarget, ByteOffset);
 
@@ -1343,8 +1340,8 @@ bool AMDGPUDAGToDAGISel::SelectSMRDOffset(SDValue ByteOffsetNode,
   if (!isUInt<32>(EncodedOffset) || !isUInt<32>(ByteOffset))
     return false;
 
-  if (Gen == AMDGPUSubtarget::SEA_ISLANDS && isUInt<32>(EncodedOffset)) {
-    // 32-bit Immediates are supported on Sea Islands.
+  if (Subtarget->has32BitLiteralSMRDOffset() &&
+      ByteOffset % 4 == 0 && isUInt<32>(EncodedOffset)) {
     Offset = CurDAG->getTargetConstant(EncodedOffset, SL, MVT::i32);
   } else {
     SDValue C32Bit = CurDAG->getTargetConstant(ByteOffset, SL, MVT::i32);
@@ -1376,20 +1373,15 @@ bool AMDGPUDAGToDAGISel::SelectSMRD(SDValue Addr, SDValue &SBase,
 bool AMDGPUDAGToDAGISel::SelectSMRDImm(SDValue Addr, SDValue &SBase,
                                        SDValue &Offset) const {
   bool Imm;
-  return SelectSMRD(Addr, SBase, Offset, Imm) && Imm;
-}
-
-bool AMDGPUDAGToDAGISel::SelectSMRDImm32(SDValue Addr, SDValue &SBase,
-                                         SDValue &Offset) const {
-
-  if (Subtarget->getGeneration() != AMDGPUSubtarget::SEA_ISLANDS)
-    return false;
 
-  bool Imm;
   if (!SelectSMRD(Addr, SBase, Offset, Imm))
     return false;
 
-  return !Imm && isa<ConstantSDNode>(Offset);
+  if (Subtarget->has32BitLiteralSMRDOffset() &&
+      isa<ConstantSDNode>(Offset))
+    return true;
+
+  return Imm;
 }
 
 bool AMDGPUDAGToDAGISel::SelectSMRDSgpr(SDValue Addr, SDValue &SBase,
@@ -1402,19 +1394,15 @@ bool AMDGPUDAGToDAGISel::SelectSMRDSgpr(SDValue Addr, SDValue &SBase,
 bool AMDGPUDAGToDAGISel::SelectSMRDBufferImm(SDValue Addr,
                                              SDValue &Offset) const {
   bool Imm;
-  return SelectSMRDOffset(Addr, Offset, Imm) && Imm;
-}
-
-bool AMDGPUDAGToDAGISel::SelectSMRDBufferImm32(SDValue Addr,
-                                               SDValue &Offset) const {
-  if (Subtarget->getGeneration() != AMDGPUSubtarget::SEA_ISLANDS)
-    return false;
 
-  bool Imm;
   if (!SelectSMRDOffset(Addr, Offset, Imm))
     return false;
 
-  return !Imm && isa<ConstantSDNode>(Offset);
+  if (Subtarget->has32BitLiteralSMRDOffset() &&
+      isa<ConstantSDNode>(Offset))
+    return true;
+
+  return Imm;
 }
 
 bool AMDGPUDAGToDAGISel::SelectSMRDBufferSgpr(SDValue Addr,
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.h b/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.h
index e543cae07ad..d122af0e27b 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.h
+++ b/llvm/lib/Target/AMDGPU/AMDGPUSubtarget.h
@@ -399,6 +399,10 @@ public:
     return FlatScratchInsts;
   }
 
+  bool has32BitLiteralSMRDOffset() const {
+    return getGeneration() == SEA_ISLANDS;
+  }
+
   bool isMesaKernel(const MachineFunction &MF) const {
     return isMesa3DOS() && !AMDGPU::isShader(MF.getFunction()->getCallingConv());
   }
diff --git a/llvm/lib/Target/AMDGPU/SMInstructions.td b/llvm/lib/Target/AMDGPU/SMInstructions.td
index 5b840a14dbc..828fd009eeb 100644
--- a/llvm/lib/Target/AMDGPU/SMInstructions.td
+++ b/llvm/lib/Target/AMDGPU/SMInstructions.td
@@ -234,10 +234,8 @@ def smrd_load : PatFrag <(ops node:$ptr), (load node:$ptr), [{
 }]>;
 
 def SMRDImm         : ComplexPattern<i64, 2, "SelectSMRDImm">;
-def SMRDImm32       : ComplexPattern<i64, 2, "SelectSMRDImm32">;
 def SMRDSgpr        : ComplexPattern<i64, 2, "SelectSMRDSgpr">;
 def SMRDBufferImm   : ComplexPattern<i32, 1, "SelectSMRDBufferImm">;
-def SMRDBufferImm32 : ComplexPattern<i32, 1, "SelectSMRDBufferImm32">;
 def SMRDBufferSgpr  : ComplexPattern<i32, 1, "SelectSMRDBufferSgpr">;
 
 let Predicates = [isGCN] in {
@@ -276,7 +274,7 @@ defm : SMRD_Pattern <"S_LOAD_DWORDX8",  v8i32>;
 defm : SMRD_Pattern <"S_LOAD_DWORDX16", v16i32>;
 
 // 1. Offset as an immediate
-def SM_LOAD_PATTERN : Pat <  // name this pattern to reuse AddedComplexity on CI
+def : Pat <
   (SIload_constant v4i32:$sbase, (SMRDBufferImm i32:$offset)),
   (S_BUFFER_LOAD_DWORD_IMM $sbase, $offset, 0)
 >;
@@ -504,26 +502,3 @@ class SMRD_Real_ci <bits<5> op, SM_Pseudo ps>
 }
 
 def S_DCACHE_INV_VOL_ci : SMRD_Real_ci <0x1d, S_DCACHE_INV_VOL>;
-
-let AddedComplexity = SM_LOAD_PATTERN.AddedComplexity in {
-
-class SMRD_Pattern_ci <string Instr, ValueType vt> : Pat <
-  (smrd_load (SMRDImm32 i64:$sbase, i32:$offset)),
-  (vt (!cast<SM_Pseudo>(Instr#"_IMM_ci") $sbase, $offset, 0))> {
-  let Predicates = [isCIOnly];
-}
-
-def : SMRD_Pattern_ci <"S_LOAD_DWORD",    i32>;
-def : SMRD_Pattern_ci <"S_LOAD_DWORDX2",  v2i32>;
-def : SMRD_Pattern_ci <"S_LOAD_DWORDX4",  v4i32>;
-def : SMRD_Pattern_ci <"S_LOAD_DWORDX8",  v8i32>;
-def : SMRD_Pattern_ci <"S_LOAD_DWORDX16", v16i32>;
-
-def : Pat <
-  (SIload_constant v4i32:$sbase, (SMRDBufferImm32 i32:$offset)),
-  (S_BUFFER_LOAD_DWORD_IMM_ci $sbase, $offset, 0)> {
-  let Predicates = [isCI]; // should this be isCIOnly?
-}
-
-} // End let AddedComplexity = SM_LOAD_PATTERN.AddedComplexity
-
diff --git a/llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp b/llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp
index 630f469eabf..e14587e413a 100644
--- a/llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp
+++ b/llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp
@@ -763,7 +763,7 @@ int64_t getSMRDEncodedOffset(const MCSubtargetInfo &ST, int64_t ByteOffset) {
 
 bool isLegalSMRDImmOffset(const MCSubtargetInfo &ST, int64_t ByteOffset) {
   int64_t EncodedOffset = getSMRDEncodedOffset(ST, ByteOffset);
-  return isSI(ST) || isCI(ST) ? isUInt<8>(EncodedOffset) :
+  return isSI(ST) || isCI(ST) ? ByteOffset % 4 == 0 && isUInt<8>(EncodedOffset) :
                                 isUInt<20>(EncodedOffset);
 }
 } // end namespace AMDGPU