| 1 | //===- InstCombineAtomicRMW.cpp -------------------------------------------===// |
| 2 | // |
| 3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
| 4 | // See https://llvm.org/LICENSE.txt for license information. |
| 5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
| 6 | // |
| 7 | //===----------------------------------------------------------------------===// |
| 8 | // |
| 9 | // This file implements the visit functions for atomic rmw instructions. |
| 10 | // |
| 11 | //===----------------------------------------------------------------------===// |
| 12 | |
| 13 | #include "InstCombineInternal.h" |
| 14 | #include "llvm/IR/Instructions.h" |
| 15 | |
| 16 | using namespace llvm; |
| 17 | |
| 18 | namespace { |
| 19 | /// Return true if and only if the given instruction does not modify the memory |
| 20 | /// location referenced. Note that an idemptent atomicrmw may still have |
| 21 | /// ordering effects on nearby instructions, or be volatile. |
| 22 | /// TODO: Common w/ the version in AtomicExpandPass, and change the term used. |
| 23 | /// Idemptotent is confusing in this context. |
| 24 | bool isIdempotentRMW(AtomicRMWInst& RMWI) { |
| 25 | if (auto CF = dyn_cast<ConstantFP>(Val: RMWI.getValOperand())) |
| 26 | switch(RMWI.getOperation()) { |
| 27 | case AtomicRMWInst::FAdd: // -0.0 |
| 28 | return CF->isZero() && CF->isNegative(); |
| 29 | case AtomicRMWInst::FSub: // +0.0 |
| 30 | return CF->isZero() && !CF->isNegative(); |
| 31 | default: |
| 32 | return false; |
| 33 | }; |
| 34 | |
| 35 | auto C = dyn_cast<ConstantInt>(Val: RMWI.getValOperand()); |
| 36 | if(!C) |
| 37 | return false; |
| 38 | |
| 39 | switch(RMWI.getOperation()) { |
| 40 | case AtomicRMWInst::Add: |
| 41 | case AtomicRMWInst::Sub: |
| 42 | case AtomicRMWInst::Or: |
| 43 | case AtomicRMWInst::Xor: |
| 44 | return C->isZero(); |
| 45 | case AtomicRMWInst::And: |
| 46 | return C->isMinusOne(); |
| 47 | case AtomicRMWInst::Min: |
| 48 | return C->isMaxValue(IsSigned: true); |
| 49 | case AtomicRMWInst::Max: |
| 50 | return C->isMinValue(IsSigned: true); |
| 51 | case AtomicRMWInst::UMin: |
| 52 | return C->isMaxValue(IsSigned: false); |
| 53 | case AtomicRMWInst::UMax: |
| 54 | return C->isMinValue(IsSigned: false); |
| 55 | default: |
| 56 | return false; |
| 57 | } |
| 58 | } |
| 59 | |
| 60 | /// Return true if the given instruction always produces a value in memory |
| 61 | /// equivalent to its value operand. |
| 62 | bool isSaturating(AtomicRMWInst& RMWI) { |
| 63 | if (auto CF = dyn_cast<ConstantFP>(Val: RMWI.getValOperand())) |
| 64 | switch (RMWI.getOperation()) { |
| 65 | case AtomicRMWInst::FMax: |
| 66 | // maxnum(x, +inf) -> +inf |
| 67 | return !CF->isNegative() && CF->isInfinity(); |
| 68 | case AtomicRMWInst::FMin: |
| 69 | // minnum(x, -inf) -> +inf |
| 70 | return CF->isNegative() && CF->isInfinity(); |
| 71 | case AtomicRMWInst::FAdd: |
| 72 | case AtomicRMWInst::FSub: |
| 73 | return CF->isNaN(); |
| 74 | default: |
| 75 | return false; |
| 76 | }; |
| 77 | |
| 78 | auto C = dyn_cast<ConstantInt>(Val: RMWI.getValOperand()); |
| 79 | if(!C) |
| 80 | return false; |
| 81 | |
| 82 | switch(RMWI.getOperation()) { |
| 83 | default: |
| 84 | return false; |
| 85 | case AtomicRMWInst::Xchg: |
| 86 | return true; |
| 87 | case AtomicRMWInst::Or: |
| 88 | return C->isAllOnesValue(); |
| 89 | case AtomicRMWInst::And: |
| 90 | return C->isZero(); |
| 91 | case AtomicRMWInst::Min: |
| 92 | return C->isMinValue(IsSigned: true); |
| 93 | case AtomicRMWInst::Max: |
| 94 | return C->isMaxValue(IsSigned: true); |
| 95 | case AtomicRMWInst::UMin: |
| 96 | return C->isMinValue(IsSigned: false); |
| 97 | case AtomicRMWInst::UMax: |
| 98 | return C->isMaxValue(IsSigned: false); |
| 99 | }; |
| 100 | } |
| 101 | } // namespace |
| 102 | |
| 103 | Instruction *InstCombinerImpl::visitAtomicRMWInst(AtomicRMWInst &RMWI) { |
| 104 | |
| 105 | // Volatile RMWs perform a load and a store, we cannot replace this by just a |
| 106 | // load or just a store. We chose not to canonicalize out of general paranoia |
| 107 | // about user expectations around volatile. |
| 108 | if (RMWI.isVolatile()) |
| 109 | return nullptr; |
| 110 | |
| 111 | // Any atomicrmw op which produces a known result in memory can be |
| 112 | // replaced w/an atomicrmw xchg. |
| 113 | if (isSaturating(RMWI) && |
| 114 | RMWI.getOperation() != AtomicRMWInst::Xchg) { |
| 115 | RMWI.setOperation(AtomicRMWInst::Xchg); |
| 116 | return &RMWI; |
| 117 | } |
| 118 | |
| 119 | assert(RMWI.getOrdering() != AtomicOrdering::NotAtomic && |
| 120 | RMWI.getOrdering() != AtomicOrdering::Unordered && |
| 121 | "AtomicRMWs don't make sense with Unordered or NotAtomic" ); |
| 122 | |
| 123 | if (!isIdempotentRMW(RMWI)) |
| 124 | return nullptr; |
| 125 | |
| 126 | // We chose to canonicalize all idempotent operations to an single |
| 127 | // operation code and constant. This makes it easier for the rest of the |
| 128 | // optimizer to match easily. The choices of or w/0 and fadd w/-0.0 are |
| 129 | // arbitrary. |
| 130 | if (RMWI.getType()->isIntegerTy() && |
| 131 | RMWI.getOperation() != AtomicRMWInst::Or) { |
| 132 | RMWI.setOperation(AtomicRMWInst::Or); |
| 133 | return replaceOperand(I&: RMWI, OpNum: 1, V: ConstantInt::get(Ty: RMWI.getType(), V: 0)); |
| 134 | } else if (RMWI.getType()->isFloatingPointTy() && |
| 135 | RMWI.getOperation() != AtomicRMWInst::FAdd) { |
| 136 | RMWI.setOperation(AtomicRMWInst::FAdd); |
| 137 | return replaceOperand(I&: RMWI, OpNum: 1, V: ConstantFP::getNegativeZero(Ty: RMWI.getType())); |
| 138 | } |
| 139 | |
| 140 | return nullptr; |
| 141 | } |
| 142 | |