From cad864d49e90b95de39d62f71fb6d1b7172b7498 Mon Sep 17 00:00:00 2001 From: Jessica Paquette Date: Tue, 13 Nov 2018 23:01:34 +0000 Subject: [PATCH] [MachineOutliner][NFC] Use MBB flags to avoid call checks in getOutliningInfo We already determine a bunch of information about an MBB in getMachineOutlinerMBBFlags. We can reuse that information to avoid calculating things that must be false/true. The first thing we can easily check is if an outlined sequence could ever contain calls. There's no reason to walk over the outlined range, checking for calls, if we already know that there are no calls in the block containing the sequence. llvm-svn: 346809 --- llvm/include/llvm/CodeGen/MachineOutliner.h | 7 +++-- llvm/lib/CodeGen/MachineOutliner.cpp | 11 +++++-- llvm/lib/Target/AArch64/AArch64InstrInfo.cpp | 46 +++++++++++++++------------- 3 files changed, 39 insertions(+), 25 deletions(-) diff --git a/llvm/include/llvm/CodeGen/MachineOutliner.h b/llvm/include/llvm/CodeGen/MachineOutliner.h index eaa7413..2bd39b2 100644 --- a/llvm/include/llvm/CodeGen/MachineOutliner.h +++ b/llvm/include/llvm/CodeGen/MachineOutliner.h @@ -82,6 +82,9 @@ public: /// been used across the sequence. LiveRegUnits UsedInSequence; + /// Target-specific flags for this Candidate's MBB. + unsigned Flags = 0x0; + /// Return the number of instructions in this Candidate. unsigned getLength() const { return Len; } @@ -120,9 +123,9 @@ public: Candidate(unsigned StartIdx, unsigned Len, MachineBasicBlock::iterator &FirstInst, MachineBasicBlock::iterator &LastInst, MachineBasicBlock *MBB, - unsigned FunctionIdx) + unsigned FunctionIdx, unsigned Flags) : StartIdx(StartIdx), Len(Len), FirstInst(FirstInst), LastInst(LastInst), - MBB(MBB), FunctionIdx(FunctionIdx) {} + MBB(MBB), FunctionIdx(FunctionIdx), Flags(Flags) {} Candidate() {} /// Used to ensure that \p Candidates are outlined in an order that diff --git a/llvm/lib/CodeGen/MachineOutliner.cpp b/llvm/lib/CodeGen/MachineOutliner.cpp index 6b9fadd..49d0893 100644 --- a/llvm/lib/CodeGen/MachineOutliner.cpp +++ b/llvm/lib/CodeGen/MachineOutliner.cpp @@ -625,6 +625,9 @@ struct InstructionMapper { /// Inverse of \p InstructionIntegerMap. DenseMap IntegerInstructionMap; + /// Correspondence between \p MachineBasicBlocks and target-defined flags. + DenseMap MBBFlagsMap; + /// The vector of unsigned integers that the module is mapped to. std::vector UnsignedVec; @@ -748,6 +751,9 @@ struct InstructionMapper { if (!TII.isMBBSafeToOutlineFrom(MBB, Flags)) return; + // Store info for the MBB for later outlining. + MBBFlagsMap[&MBB] = Flags; + MachineBasicBlock::iterator It = MBB.begin(); // The number of instructions in this block that will be considered for @@ -1106,10 +1112,11 @@ unsigned MachineOutliner::findCandidates( MachineBasicBlock::iterator StartIt = Mapper.InstrList[StartIdx]; MachineBasicBlock::iterator EndIt = Mapper.InstrList[EndIdx]; + MachineBasicBlock *MBB = StartIt->getParent(); CandidatesForRepeatedSeq.emplace_back(StartIdx, StringLen, StartIt, - EndIt, StartIt->getParent(), - FunctionList.size()); + EndIt, MBB, FunctionList.size(), + Mapper.MBBFlagsMap[MBB]); } } diff --git a/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp b/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp index 176514d..219dd4d 100644 --- a/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp +++ b/llvm/lib/Target/AArch64/AArch64InstrInfo.cpp @@ -5128,12 +5128,12 @@ AArch64InstrInfo::findRegisterToSaveLRTo(const outliner::Candidate &C) const { outliner::OutlinedFunction AArch64InstrInfo::getOutliningCandidateInfo( std::vector &RepeatedSequenceLocs) const { - unsigned SequenceSize = std::accumulate( - RepeatedSequenceLocs[0].front(), - std::next(RepeatedSequenceLocs[0].back()), - 0, [this](unsigned Sum, const MachineInstr &MI) { - return Sum + getInstSizeInBytes(MI); - }); + outliner::Candidate &FirstCand = RepeatedSequenceLocs[0]; + unsigned SequenceSize = + std::accumulate(FirstCand.front(), std::next(FirstCand.back()), 0, + [this](unsigned Sum, const MachineInstr &MI) { + return Sum + getInstSizeInBytes(MI); + }); // Compute liveness information for each candidate. const TargetRegisterInfo &TRI = getRegisterInfo(); @@ -5240,21 +5240,25 @@ AArch64InstrInfo::getOutliningCandidateInfo( } } - // Check if the range contains a call. These require a save + restore of the - // link register. - if (std::any_of(RepeatedSequenceLocs[0].front(), - RepeatedSequenceLocs[0].back(), - [](const MachineInstr &MI) { return MI.isCall(); })) - NumBytesToCreateFrame += 8; // Save + restore the link register. - - // Handle the last instruction separately. If this is a tail call, then the - // last instruction is a call. We don't want to save + restore in this case. - // However, it could be possible that the last instruction is a call without - // it being valid to tail call this sequence. We should consider this as well. - else if (FrameID != MachineOutlinerThunk && - FrameID != MachineOutlinerTailCall && - RepeatedSequenceLocs[0].back()->isCall()) - NumBytesToCreateFrame += 8; + // If the MBB containing the first candidate has calls, then it's possible + // that we have calls in the candidate. If there are no calls, then there's + // no way that any candidate could have any calls. + if (FirstCand.Flags & MachineOutlinerMBBFlags::HasCalls) { + // Check if the range contains a call. These require a save + restore of the + // link register. + if (std::any_of(FirstCand.front(), FirstCand.back(), + [](const MachineInstr &MI) { return MI.isCall(); })) + NumBytesToCreateFrame += 8; // Save + restore the link register. + + // Handle the last instruction separately. If this is a tail call, then the + // last instruction is a call. We don't want to save + restore in this case. + // However, it could be possible that the last instruction is a call without + // it being valid to tail call this sequence. We should consider this as + // well. + else if (FrameID != MachineOutlinerThunk && + FrameID != MachineOutlinerTailCall && FirstCand.back()->isCall()) + NumBytesToCreateFrame += 8; + } return outliner::OutlinedFunction(RepeatedSequenceLocs, SequenceSize, NumBytesToCreateFrame, FrameID); -- 2.7.4