[AARCH64][LSE] Preliminary support for ARMv8.1 LSE Atomics.

Implemented support to AArch64 codegen for ARMv8.1 Large System
Extensions atomic instructions. Where supported, these instructions can
provide atomic operations with higher performance.

Currently supported operations include: fetch_add, fetch_or, fetch_xor,
fetch_smin, fetch_min/max (signed and unsigned), swap, and
compare_exchange.

This implementation implies sequential-consistency ordering, more
relaxed ordering is under development.

Subtarget->hasLSE is currently supported for Cavium ThunderX2T99.

Patch by Ananth Jasty.

Differential Revision: https://reviews.llvm.org/D33586

Change-Id: I82f6d3d64255622791ceb0715b7ab9f4dc4d4b2c

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@305893 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/AArch64/AArch64DeadRegisterDefinitionsPass.cpp b/lib/Target/AArch64/AArch64DeadRegisterDefinitionsPass.cpp
index 544f674..ee54550 100644 (file)
--- a/lib/Target/AArch64/AArch64DeadRegisterDefinitionsPass.cpp
+++ b/lib/Target/AArch64/AArch64DeadRegisterDefinitionsPass.cpp
@@ -13,7 +13,9 @@
 
 #include "AArch64.h"
 #include "AArch64RegisterInfo.h"
+#include "AArch64Subtarget.h"
 #include "llvm/ADT/Statistic.h"
+#include "llvm/CodeGen/ISDOpcodes.h"
 #include "llvm/CodeGen/MachineFunction.h"
 #include "llvm/CodeGen/MachineFunctionPass.h"
 #include "llvm/CodeGen/MachineInstr.h"
@@ -84,6 +86,51 @@ void AArch64DeadRegisterDefinitions::processMachineBasicBlock(
       DEBUG(dbgs() << "    Ignoring, XZR or WZR already used by the instruction\n");
       continue;
     }
+    if (MF.getSubtarget<AArch64Subtarget>().hasLSE()) {
+      // XZ/WZ for LSE can only be used when acquire semantics are not used,
+      // LDOPAL WZ is an invalid opcode.
+      switch (MI.getOpcode()) {
+      case AArch64::CASALb:
+      case AArch64::CASALh:
+      case AArch64::CASALs:
+      case AArch64::CASALd:
+      case AArch64::SWPALb:
+      case AArch64::SWPALh:
+      case AArch64::SWPALs:
+      case AArch64::SWPALd:
+      case AArch64::LDADDALb:
+      case AArch64::LDADDALh:
+      case AArch64::LDADDALs:
+      case AArch64::LDADDALd:
+      case AArch64::LDEORALb:
+      case AArch64::LDEORALh:
+      case AArch64::LDEORALs:
+      case AArch64::LDEORALd:
+      case AArch64::LDSETALb:
+      case AArch64::LDSETALh:
+      case AArch64::LDSETALs:
+      case AArch64::LDSETALd:
+      case AArch64::LDSMINALb:
+      case AArch64::LDSMINALh:
+      case AArch64::LDSMINALs:
+      case AArch64::LDSMINALd:
+      case AArch64::LDSMAXALb:
+      case AArch64::LDSMAXALh:
+      case AArch64::LDSMAXALs:
+      case AArch64::LDSMAXALd:
+      case AArch64::LDUMINALb:
+      case AArch64::LDUMINALh:
+      case AArch64::LDUMINALs:
+      case AArch64::LDUMINALd:
+      case AArch64::LDUMAXALb:
+      case AArch64::LDUMAXALh:
+      case AArch64::LDUMAXALs:
+      case AArch64::LDUMAXALd:
+        continue;
+      default:
+        break;
+      }
+    }
     const MCInstrDesc &Desc = MI.getDesc();
     for (int I = 0, E = Desc.getNumDefs(); I != E; ++I) {
       MachineOperand &MO = MI.getOperand(I);

diff --git a/lib/Target/AArch64/AArch64ISelDAGToDAG.cpp b/lib/Target/AArch64/AArch64ISelDAGToDAG.cpp
index 8c2c0a5..0468784 100644 (file)
--- a/lib/Target/AArch64/AArch64ISelDAGToDAG.cpp
+++ b/lib/Target/AArch64/AArch64ISelDAGToDAG.cpp
@@ -201,7 +201,7 @@ private:
 
   bool SelectCVTFixedPosOperand(SDValue N, SDValue &FixedPos, unsigned Width);
 
-  void SelectCMP_SWAP(SDNode *N);
+  bool SelectCMP_SWAP(SDNode *N);
 
 };
 } // end anonymous namespace
@@ -2609,9 +2609,13 @@ bool AArch64DAGToDAGISel::tryWriteRegister(SDNode *N) {
 }
 
 /// We've got special pseudo-instructions for these
-void AArch64DAGToDAGISel::SelectCMP_SWAP(SDNode *N) {
+bool AArch64DAGToDAGISel::SelectCMP_SWAP(SDNode *N) {
   unsigned Opcode;
   EVT MemTy = cast<MemSDNode>(N)->getMemoryVT();
+
+  // Leave IR for LSE if subtarget supports it.
+  if (Subtarget->hasLSE()) return false;
+
   if (MemTy == MVT::i8)
     Opcode = AArch64::CMP_SWAP_8;
   else if (MemTy == MVT::i16)
@@ -2637,6 +2641,8 @@ void AArch64DAGToDAGISel::SelectCMP_SWAP(SDNode *N) {
   ReplaceUses(SDValue(N, 0), SDValue(CmpSwap, 0));
   ReplaceUses(SDValue(N, 1), SDValue(CmpSwap, 2));
   CurDAG->RemoveDeadNode(N);
+
+  return true;
 }
 
 void AArch64DAGToDAGISel::Select(SDNode *Node) {
@@ -2660,8 +2666,9 @@ void AArch64DAGToDAGISel::Select(SDNode *Node) {
     break;
 
   case ISD::ATOMIC_CMP_SWAP:
-    SelectCMP_SWAP(Node);
-    return;
+    if (SelectCMP_SWAP(Node))
+      return;
+    break;
 
   case ISD::READ_REGISTER:
     if (tryReadRegister(Node))

diff --git a/lib/Target/AArch64/AArch64ISelLowering.cpp b/lib/Target/AArch64/AArch64ISelLowering.cpp
index 083ca21..2965106 100644 (file)
--- a/lib/Target/AArch64/AArch64ISelLowering.cpp
+++ b/lib/Target/AArch64/AArch64ISelLowering.cpp
@@ -10563,11 +10563,20 @@ AArch64TargetLowering::shouldExpandAtomicLoadInIR(LoadInst *LI) const {
 TargetLowering::AtomicExpansionKind
 AArch64TargetLowering::shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const {
   unsigned Size = AI->getType()->getPrimitiveSizeInBits();
-  return Size <= 128 ? AtomicExpansionKind::LLSC : AtomicExpansionKind::None;
+  if (Size > 128) return AtomicExpansionKind::None;
+  // Nand not supported in LSE.
+  if (AI->getOperation() == AtomicRMWInst::Nand) return AtomicExpansionKind::LLSC;
+  // Currently leaving And and Sub to LLSC
+  if ((AI->getOperation() == AtomicRMWInst::And) || (AI->getOperation() == AtomicRMWInst::Sub))
+    return AtomicExpansionKind::LLSC;
+  // Leave 128 bits to LLSC.
+  return (Subtarget->hasLSE() && Size < 128) ? AtomicExpansionKind::None : AtomicExpansionKind::LLSC;
 }
 
 bool AArch64TargetLowering::shouldExpandAtomicCmpXchgInIR(
     AtomicCmpXchgInst *AI) const {
+  // If subtarget has LSE, leave cmpxchg intact for codegen.
+  if (Subtarget->hasLSE()) return false;
   // At -O0, fast-regalloc cannot cope with the live vregs necessary to
   // implement cmpxchg without spilling. If the address being exchanged is also
   // on the stack and close enough to the spill slot, this can lead to a

diff --git a/lib/Target/AArch64/AArch64InstrAtomics.td b/lib/Target/AArch64/AArch64InstrAtomics.td
index 71826be..de283b7 100644 (file)
--- a/lib/Target/AArch64/AArch64InstrAtomics.td
+++ b/lib/Target/AArch64/AArch64InstrAtomics.td
@@ -405,3 +405,49 @@ def CMP_SWAP_128 : Pseudo<(outs GPR64:$RdLo, GPR64:$RdHi, GPR32:$scratch),
                           (ins GPR64:$addr, GPR64:$desiredLo, GPR64:$desiredHi,
                                GPR64:$newLo, GPR64:$newHi), []>,
                    Sched<[WriteAtomic]>;
+
+// v8.1 Atomic instructions:
+def : Pat<(atomic_load_add_8 GPR64:$Rn, GPR32:$Rs), (LDADDALb GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_add_16 GPR64:$Rn, GPR32:$Rs), (LDADDALh GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_add_32 GPR64:$Rn, GPR32:$Rs), (LDADDALs GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_add_64 GPR64:$Rn, GPR64:$Rs), (LDADDALd GPR64:$Rs, GPR64sp:$Rn)>;
+
+def : Pat<(atomic_load_or_8 GPR64:$Rn, GPR32:$Rs), (LDSETALb GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_or_16 GPR64:$Rn, GPR32:$Rs), (LDSETALh GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_or_32 GPR64:$Rn, GPR32:$Rs), (LDSETALs GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_or_64 GPR64:$Rn, GPR64:$Rs), (LDSETALd GPR64:$Rs, GPR64sp:$Rn)>;
+
+def : Pat<(atomic_load_xor_8 GPR64:$Rn, GPR32:$Rs), (LDEORALb GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_xor_16 GPR64:$Rn, GPR32:$Rs), (LDEORALh GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_xor_32 GPR64:$Rn, GPR32:$Rs), (LDEORALs GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_xor_64 GPR64:$Rn, GPR64:$Rs), (LDEORALd GPR64:$Rs, GPR64sp:$Rn)>;
+
+def : Pat<(atomic_load_max_8 GPR64:$Rn, GPR32:$Rs), (LDSMAXALb GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_max_16 GPR64:$Rn, GPR32:$Rs), (LDSMAXALh GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_max_32 GPR64:$Rn, GPR32:$Rs), (LDSMAXALs GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_max_64 GPR64:$Rn, GPR64:$Rs), (LDSMAXALd GPR64:$Rs, GPR64sp:$Rn)>;
+
+def : Pat<(atomic_load_umax_8 GPR64:$Rn, GPR32:$Rs), (LDUMAXALb GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_umax_16 GPR64:$Rn, GPR32:$Rs), (LDUMAXALh GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_umax_32 GPR64:$Rn, GPR32:$Rs), (LDUMAXALs GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_umax_64 GPR64:$Rn, GPR64:$Rs), (LDUMAXALd GPR64:$Rs, GPR64sp:$Rn)>;
+
+def : Pat<(atomic_load_min_8 GPR64:$Rn, GPR32:$Rs), (LDSMINALb GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_min_16 GPR64:$Rn, GPR32:$Rs), (LDSMINALh GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_min_32 GPR64:$Rn, GPR32:$Rs), (LDSMINALs GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_min_64 GPR64:$Rn, GPR64:$Rs), (LDSMINALd GPR64:$Rs, GPR64sp:$Rn)>;
+
+def : Pat<(atomic_load_umin_8 GPR64:$Rn, GPR32:$Rs), (LDUMINALb GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_umin_16 GPR64:$Rn, GPR32:$Rs), (LDUMINALh GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_umin_32 GPR64:$Rn, GPR32:$Rs), (LDUMINALs GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_load_umin_64 GPR64:$Rn, GPR64:$Rs), (LDUMINALd GPR64:$Rs, GPR64sp:$Rn)>;
+
+def : Pat<(atomic_cmp_swap_8 GPR64:$Rn, GPR32:$Rold, GPR32:$Rnew), (CASALb GPR32:$Rold, GPR32:$Rnew, GPR64sp:$Rn)>;
+def : Pat<(atomic_cmp_swap_16 GPR64:$Rn, GPR32:$Rold, GPR32:$Rnew), (CASALh GPR32:$Rold, GPR32:$Rnew, GPR64sp:$Rn)>;
+def : Pat<(atomic_cmp_swap_32 GPR64:$Rn, GPR32:$Rold, GPR32:$Rnew), (CASALs GPR32:$Rold, GPR32:$Rnew, GPR64sp:$Rn)>;
+def : Pat<(atomic_cmp_swap_64 GPR64:$Rn, GPR64:$Rold, GPR64:$Rnew), (CASALd GPR64:$Rold, GPR64:$Rnew, GPR64sp:$Rn)>;
+
+def : Pat<(atomic_swap_8 GPR64:$Rn, GPR32:$Rs), (SWPALb GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_swap_16 GPR64:$Rn, GPR32:$Rs), (SWPALh GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_swap_32 GPR64:$Rn, GPR32:$Rs), (SWPALs GPR32:$Rs, GPR64sp:$Rn)>;
+def : Pat<(atomic_swap_64 GPR64:$Rn, GPR64:$Rs), (SWPALd GPR64:$Rs, GPR64sp:$Rn)>;