| 1 | //==-- loop_proto_to_llvm.cpp - Protobuf-C++ conversion |
| 2 | //---------------------==// |
| 3 | // |
| 4 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
| 5 | // See https://llvm.org/LICENSE.txt for license information. |
| 6 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
| 7 | // |
| 8 | //===----------------------------------------------------------------------===// |
| 9 | // |
| 10 | // Implements functions for converting between protobufs and LLVM IR. |
| 11 | // |
| 12 | // |
| 13 | //===----------------------------------------------------------------------===// |
| 14 | |
| 15 | #include "loop_proto_to_llvm.h" |
| 16 | #include "cxx_loop_proto.pb.h" |
| 17 | #include "../handle-llvm/input_arrays.h" |
| 18 | |
| 19 | // The following is needed to convert protos in human-readable form |
| 20 | #include <google/protobuf/text_format.h> |
| 21 | |
| 22 | #include <ostream> |
| 23 | #include <sstream> |
| 24 | |
| 25 | namespace clang_fuzzer { |
| 26 | |
| 27 | // Forward decls |
| 28 | std::string BinopToString(std::ostream &os, const BinaryOp &x); |
| 29 | std::string StateSeqToString(std::ostream &os, const StatementSeq &x); |
| 30 | |
| 31 | // Counter variable to generate new LLVM IR variable names and wrapper function |
| 32 | static std::string get_var() { |
| 33 | static int ctr = 0; |
| 34 | return "%var" + std::to_string(val: ctr++); |
| 35 | } |
| 36 | |
| 37 | static bool inner_loop = false; |
| 38 | class InnerLoop { |
| 39 | public: |
| 40 | InnerLoop() { |
| 41 | inner_loop = true; |
| 42 | } |
| 43 | ~InnerLoop() { |
| 44 | inner_loop = false; |
| 45 | } |
| 46 | }; |
| 47 | |
| 48 | |
| 49 | // Proto to LLVM. |
| 50 | |
| 51 | std::string ConstToString(const Const &x) { |
| 52 | return std::to_string(x.val()); |
| 53 | } |
| 54 | std::string VarRefToString(std::ostream &os, const VarRef &x) { |
| 55 | std::string which_loop = inner_loop ? "inner" : "outer" ; |
| 56 | std::string arr; |
| 57 | switch(x.arr()) { |
| 58 | case VarRef::ARR_A: |
| 59 | arr = "%a" ; |
| 60 | break; |
| 61 | case VarRef::ARR_B: |
| 62 | arr = "%b" ; |
| 63 | break; |
| 64 | case VarRef::ARR_C: |
| 65 | arr = "%c" ; |
| 66 | break; |
| 67 | } |
| 68 | std::string ptr_var = get_var(); |
| 69 | os << ptr_var << " = getelementptr inbounds i32, i32* " << arr |
| 70 | << ", i64 %" << which_loop << "_ct\n" ; |
| 71 | return ptr_var; |
| 72 | } |
| 73 | std::string RvalueToString(std::ostream &os, const Rvalue &x) { |
| 74 | if(x.has_cons()) |
| 75 | return ConstToString(x.cons()); |
| 76 | if(x.has_binop()) |
| 77 | return BinopToString(os, x.binop()); |
| 78 | if(x.has_varref()) { |
| 79 | std::string var_ref = VarRefToString(os, x.varref()); |
| 80 | std::string val_var = get_var(); |
| 81 | os << val_var << " = load i32, i32* " << var_ref << "\n" ; |
| 82 | return val_var; |
| 83 | } |
| 84 | return "1" ; |
| 85 | |
| 86 | } |
| 87 | std::string BinopToString(std::ostream &os, const BinaryOp &x) { |
| 88 | std::string left = RvalueToString(os, x.left()); |
| 89 | std::string right = RvalueToString(os, x.right()); |
| 90 | std::string op; |
| 91 | switch (x.op()) { |
| 92 | case BinaryOp::PLUS: |
| 93 | op = "add" ; |
| 94 | break; |
| 95 | case BinaryOp::MINUS: |
| 96 | op = "sub" ; |
| 97 | break; |
| 98 | case BinaryOp::MUL: |
| 99 | op = "mul" ; |
| 100 | break; |
| 101 | case BinaryOp::XOR: |
| 102 | op = "xor" ; |
| 103 | break; |
| 104 | case BinaryOp::AND: |
| 105 | op = "and" ; |
| 106 | break; |
| 107 | case BinaryOp::OR: |
| 108 | op = "or" ; |
| 109 | break; |
| 110 | // Support for Boolean operators will be added later |
| 111 | case BinaryOp::EQ: |
| 112 | case BinaryOp::NE: |
| 113 | case BinaryOp::LE: |
| 114 | case BinaryOp::GE: |
| 115 | case BinaryOp::LT: |
| 116 | case BinaryOp::GT: |
| 117 | op = "add" ; |
| 118 | break; |
| 119 | } |
| 120 | std::string val_var = get_var(); |
| 121 | os << val_var << " = " << op << " i32 " << left << ", " << right << "\n" ; |
| 122 | return val_var; |
| 123 | } |
| 124 | std::ostream &operator<<(std::ostream &os, const AssignmentStatement &x) { |
| 125 | std::string rvalue = RvalueToString(os, x.rvalue()); |
| 126 | std::string var_ref = VarRefToString(os, x.varref()); |
| 127 | return os << "store i32 " << rvalue << ", i32* " << var_ref << "\n" ; |
| 128 | } |
| 129 | std::ostream &operator<<(std::ostream &os, const Statement &x) { |
| 130 | return os << x.assignment(); |
| 131 | } |
| 132 | std::ostream &operator<<(std::ostream &os, const StatementSeq &x) { |
| 133 | for (auto &st : x.statements()) { |
| 134 | os << st; |
| 135 | } |
| 136 | return os; |
| 137 | } |
| 138 | void NestedLoopToString(std::ostream &os, const LoopFunction &x) { |
| 139 | os << "target triple = \"x86_64-unknown-linux-gnu\"\n" |
| 140 | << "define void @foo(i32* %a, i32* %b, i32* noalias %c, i64 %s) {\n" |
| 141 | << "outer_loop_start:\n" |
| 142 | << "%cmp = icmp sgt i64 %s, 0\n" |
| 143 | << "br i1 %cmp, label %inner_loop_start, label %end\n" |
| 144 | << "outer_loop:\n" |
| 145 | << x.outer_statements() |
| 146 | << "%o_ct_new = add i64 %outer_ct, 1\n" |
| 147 | << "%jmp_outer = icmp eq i64 %o_ct_new, %s\n" |
| 148 | << "br i1 %jmp_outer, label %end, label %inner_loop_start\n" |
| 149 | << "inner_loop_start:\n" |
| 150 | << "%outer_ct = phi i64 [%o_ct_new, %outer_loop], [0, %outer_loop_start]\n" |
| 151 | << "br label %inner_loop\n" |
| 152 | << "inner_loop:\n" |
| 153 | << "%inner_ct = phi i64 [0, %inner_loop_start], [%i_ct_new, %inner_loop]\n" ; |
| 154 | { |
| 155 | InnerLoop IL; |
| 156 | os << x.inner_statements(); |
| 157 | } |
| 158 | os << "%i_ct_new = add i64 %inner_ct, 1\n" |
| 159 | << "%jmp_inner = icmp eq i64 %i_ct_new, %s\n" |
| 160 | << "br i1 %jmp_inner, label %outer_loop, label %inner_loop, !llvm.loop !0\n" |
| 161 | << "end:\n" |
| 162 | << "ret void\n" |
| 163 | << "}\n" |
| 164 | << "!0 = distinct !{!0, !1, !2}\n" |
| 165 | << "!1 = !{!\"llvm.loop.vectorize.enable\", i1 true}\n" |
| 166 | << "!2 = !{!\"llvm.loop.vectorize.width\", i32 " << kArraySize << "}\n" ; |
| 167 | } |
| 168 | void SingleLoopToString(std::ostream &os, const LoopFunction &x) { |
| 169 | os << "target triple = \"x86_64-unknown-linux-gnu\"\n" |
| 170 | << "define void @foo(i32* %a, i32* %b, i32* noalias %c, i64 %s) {\n" |
| 171 | << "%cmp = icmp sgt i64 %s, 0\n" |
| 172 | << "br i1 %cmp, label %start, label %end\n" |
| 173 | << "start:\n" |
| 174 | << "br label %loop\n" |
| 175 | << "end:\n" |
| 176 | << "ret void\n" |
| 177 | << "loop:\n" |
| 178 | << "%outer_ct = phi i64 [ %ctnew, %loop ], [ 0, %start ]\n" |
| 179 | << x.outer_statements() |
| 180 | << "%ctnew = add i64 %outer_ct, 1\n" |
| 181 | << "%j = icmp eq i64 %ctnew, %s\n" |
| 182 | << "br i1 %j, label %end, label %loop, !llvm.loop !0\n}\n" |
| 183 | << "!0 = distinct !{!0, !1, !2}\n" |
| 184 | << "!1 = !{!\"llvm.loop.vectorize.enable\", i1 true}\n" |
| 185 | << "!2 = !{!\"llvm.loop.vectorize.width\", i32 " << kArraySize << "}\n" ; |
| 186 | } |
| 187 | std::ostream &operator<<(std::ostream &os, const LoopFunction &x) { |
| 188 | if (x.has_inner_statements()) |
| 189 | NestedLoopToString(os, x); |
| 190 | else |
| 191 | SingleLoopToString(os, x); |
| 192 | return os; |
| 193 | } |
| 194 | |
| 195 | // --------------------------------- |
| 196 | |
| 197 | std::string LoopFunctionToLLVMString(const LoopFunction &input) { |
| 198 | std::ostringstream os; |
| 199 | os << input; |
| 200 | return os.str(); |
| 201 | } |
| 202 | std::string LoopProtoToLLVM(const uint8_t *data, size_t size) { |
| 203 | LoopFunction message; |
| 204 | if (!message.ParsePartialFromArray(data, size)) |
| 205 | return "#error invalid proto\n" ; |
| 206 | return LoopFunctionToLLVMString(message); |
| 207 | } |
| 208 | |
| 209 | } // namespace clang_fuzzer |
| 210 | |