8 files changed, 57 insertions, 11 deletions
diff --git a/llvm/lib/CodeGen/MIRParser/MIRParser.cpp b/llvm/lib/CodeGen/MIRParser/MIRParser.cpp
index 7d8e62736a3..e4e3fbbd75d 100644
--- a/llvm/lib/CodeGen/MIRParser/MIRParser.cpp
+++ b/llvm/lib/CodeGen/MIRParser/MIRParser.cpp
@@ -417,6 +417,8 @@ MIRParserImpl::initializeMachineFunction(const yaml::MachineFunction &YamlMF,
 
   computeFunctionProperties(MF);
 
+  MF.getSubtarget().mirFileLoaded(MF);
+
   MF.verify();
   return false;
 }
diff --git a/llvm/lib/CodeGen/TargetSubtargetInfo.cpp b/llvm/lib/CodeGen/TargetSubtargetInfo.cpp
index 1a317cd865f..7aab869f025 100644
--- a/llvm/lib/CodeGen/TargetSubtargetInfo.cpp
+++ b/llvm/lib/CodeGen/TargetSubtargetInfo.cpp
@@ -111,3 +111,6 @@ std::string TargetSubtargetInfo::getSchedInfoStr(MCInst const &MCI) const {
       TSchedModel.computeInstrRThroughput(MCI.getOpcode());
   return createSchedInfoStr(Latency, RThroughput);
 }
+
+void TargetSubtargetInfo::mirFileLoaded(MachineFunction &MF) const {
+}
diff --git a/llvm/lib/Target/AArch64/AArch64FrameLowering.cpp b/llvm/lib/Target/AArch64/AArch64FrameLowering.cpp
index d66f7b59a4b..ea4bfe7e8d9 100644
--- a/llvm/lib/Target/AArch64/AArch64FrameLowering.cpp
+++ b/llvm/lib/Target/AArch64/AArch64FrameLowering.cpp
@@ -142,6 +142,12 @@ static cl::opt<bool> EnableRedZone("aarch64-redzone",
 
 STATISTIC(NumRedZoneFunctions, "Number of functions using red zone");
 
+/// This is the biggest offset to the stack pointer we can encode in aarch64
+/// instructions (without using a separate calculation and a temp register).
+/// Note that the exception here are vector stores/loads which cannot encode any
+/// displacements (see estimateRSStackSizeLimit(), isAArch64FrameOffsetLegal()).
+static const unsigned DefaultSafeSPDisplacement = 255;
+
 /// Look at each instruction that references stack frames and return the stack
 /// size limit beyond which some of these instructions will require a scratch
 /// register during their expansion later.
@@ -167,7 +173,7 @@ static unsigned estimateRSStackSizeLimit(MachineFunction &MF) {
       }
     }
   }
-  return 255;
+  return DefaultSafeSPDisplacement;
 }
 
 bool AArch64FrameLowering::canUseRedZone(const MachineFunction &MF) const {
@@ -191,11 +197,25 @@ bool AArch64FrameLowering::hasFP(const MachineFunction &MF) const {
   const MachineFrameInfo &MFI = MF.getFrameInfo();
   const TargetRegisterInfo *RegInfo = MF.getSubtarget().getRegisterInfo();
   // Retain behavior of always omitting the FP for leaf functions when possible.
-  return (MFI.hasCalls() &&
-          MF.getTarget().Options.DisableFramePointerElim(MF)) ||
-         MFI.hasVarSizedObjects() || MFI.isFrameAddressTaken() ||
-         MFI.hasStackMap() || MFI.hasPatchPoint() ||
-         RegInfo->needsStackRealignment(MF);
+  if (MFI.hasCalls() && MF.getTarget().Options.DisableFramePointerElim(MF))
+    return true;
+  if (MFI.hasVarSizedObjects() || MFI.isFrameAddressTaken() ||
+      MFI.hasStackMap() || MFI.hasPatchPoint() ||
+      RegInfo->needsStackRealignment(MF))
+    return true;
+  // With large callframes around we may need to use FP to access the scavenging
+  // emergency spillslot.
+  //
+  // Unfortunately some calls to hasFP() like machine verifier ->
+  // getReservedReg() -> hasFP in the middle of global isel are too early
+  // to know the max call frame size. Hopefully conservatively returning "true"
+  // in those cases is fine.
+  // DefaultSafeSPDisplacement is fine as we only emergency spill GP regs.
+  if (!MFI.isMaxCallFrameSizeComputed() ||
+      MFI.getMaxCallFrameSize() > DefaultSafeSPDisplacement)
+    return true;
+
+  return false;
 }
 
 /// hasReservedCallFrame - Under normal circumstances, when a frame pointer is
diff --git a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
index f2c1f86a974..d721e7205c9 100644
--- a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
+++ b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
@@ -10982,3 +10982,8 @@ AArch64TargetLowering::getVaListSizeInBits(const DataLayout &DL) const {
 
   return 3 * getPointerTy(DL).getSizeInBits() + 2 * 32;
 }
+
+void AArch64TargetLowering::finalizeLowering(MachineFunction &MF) const {
+  MF.getFrameInfo().computeMaxCallFrameSize(MF);
+  TargetLoweringBase::finalizeLowering(MF);
+}
diff --git a/llvm/lib/Target/AArch64/AArch64ISelLowering.h b/llvm/lib/Target/AArch64/AArch64ISelLowering.h
index 0af05fff262..6018bc501db 100644
--- a/llvm/lib/Target/AArch64/AArch64ISelLowering.h
+++ b/llvm/lib/Target/AArch64/AArch64ISelLowering.h
@@ -648,6 +648,8 @@ private:
                           SelectionDAG &DAG) const override;
 
   bool shouldNormalizeToSelectSequence(LLVMContext &, EVT) const override;
+
+  void finalizeLowering(MachineFunction &MF) const override;
 };
 
 namespace AArch64 {
diff --git a/llvm/lib/Target/AArch64/AArch64RegisterInfo.cpp b/llvm/lib/Target/AArch64/AArch64RegisterInfo.cpp
index 88dd297e007..360b39125b7 100644
--- a/llvm/lib/Target/AArch64/AArch64RegisterInfo.cpp
+++ b/llvm/lib/Target/AArch64/AArch64RegisterInfo.cpp
@@ -225,11 +225,13 @@ bool AArch64RegisterInfo::requiresVirtualBaseRegisters(
 
 bool
 AArch64RegisterInfo::useFPForScavengingIndex(const MachineFunction &MF) const {
-  const MachineFrameInfo &MFI = MF.getFrameInfo();
-  // AArch64FrameLowering::resolveFrameIndexReference() can always fall back
-  // to the stack pointer, so only put the emergency spill slot next to the
-  // FP when there's no better way to access it (SP or base pointer).
-  return MFI.hasVarSizedObjects() && !hasBasePointer(MF);
+  // This function indicates whether the emergency spillslot should be placed
+  // close to the beginning of the stackframe (closer to FP) or the end
+  // (closer to SP).
+  //
+  // The beginning works most reliably if we have a frame pointer.
+  const AArch64FrameLowering &TFI = *getFrameLowering(MF);
+  return TFI.hasFP(MF);
 }
 
 bool AArch64RegisterInfo::requiresFrameIndexScavenging(
diff --git a/llvm/lib/Target/AArch64/AArch64Subtarget.cpp b/llvm/lib/Target/AArch64/AArch64Subtarget.cpp
index 688bb936d0c..eb9bb1498d6 100644
--- a/llvm/lib/Target/AArch64/AArch64Subtarget.cpp
+++ b/llvm/lib/Target/AArch64/AArch64Subtarget.cpp
@@ -250,3 +250,13 @@ std::unique_ptr<PBQPRAConstraint>
 AArch64Subtarget::getCustomPBQPConstraints() const {
   return balanceFPOps() ? llvm::make_unique<A57ChainingConstraint>() : nullptr;
 }
+
+void AArch64Subtarget::mirFileLoaded(MachineFunction &MF) const {
+  // We usually compute max call frame size after ISel. Do the computation now
+  // if the .mir file didn't specify it. Note that this will probably give you
+  // bogus values after PEI has eliminated the callframe setup/destroy pseudo
+  // instructions, specify explicitely if you need it to be correct.
+  MachineFrameInfo &MFI = MF.getFrameInfo();
+  if (!MFI.isMaxCallFrameSizeComputed())
+    MFI.computeMaxCallFrameSize(MF);
+}
diff --git a/llvm/lib/Target/AArch64/AArch64Subtarget.h b/llvm/lib/Target/AArch64/AArch64Subtarget.h
index 9245b2f396b..45a8eb16464 100644
--- a/llvm/lib/Target/AArch64/AArch64Subtarget.h
+++ b/llvm/lib/Target/AArch64/AArch64Subtarget.h
@@ -326,6 +326,8 @@ public:
       return false;
     }
   }
+
+  void mirFileLoaded(MachineFunction &MF) const override;
 };
 } // End llvm namespace