Andrew Scull | 5e1ddfa | 2018-08-14 10:06:54 +0100 | [diff] [blame] | 1 | //===- llvm/Support/LEB128.h - [SU]LEB128 utility functions -----*- C++ -*-===// |
| 2 | // |
Andrew Walbran | 16937d0 | 2019-10-22 13:54:20 +0100 | [diff] [blame] | 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 |
Andrew Scull | 5e1ddfa | 2018-08-14 10:06:54 +0100 | [diff] [blame] | 6 | // |
| 7 | //===----------------------------------------------------------------------===// |
| 8 | // |
| 9 | // This file declares some utility functions for encoding SLEB128 and |
| 10 | // ULEB128 values. |
| 11 | // |
| 12 | //===----------------------------------------------------------------------===// |
| 13 | |
| 14 | #ifndef LLVM_SUPPORT_LEB128_H |
| 15 | #define LLVM_SUPPORT_LEB128_H |
| 16 | |
| 17 | #include "llvm/Support/raw_ostream.h" |
| 18 | |
| 19 | namespace llvm { |
| 20 | |
| 21 | /// Utility function to encode a SLEB128 value to an output stream. Returns |
| 22 | /// the length in bytes of the encoded value. |
| 23 | inline unsigned encodeSLEB128(int64_t Value, raw_ostream &OS, |
| 24 | unsigned PadTo = 0) { |
| 25 | bool More; |
| 26 | unsigned Count = 0; |
| 27 | do { |
| 28 | uint8_t Byte = Value & 0x7f; |
| 29 | // NOTE: this assumes that this signed shift is an arithmetic right shift. |
| 30 | Value >>= 7; |
| 31 | More = !((((Value == 0 ) && ((Byte & 0x40) == 0)) || |
| 32 | ((Value == -1) && ((Byte & 0x40) != 0)))); |
| 33 | Count++; |
| 34 | if (More || Count < PadTo) |
| 35 | Byte |= 0x80; // Mark this byte to show that more bytes will follow. |
| 36 | OS << char(Byte); |
| 37 | } while (More); |
| 38 | |
| 39 | // Pad with 0x80 and emit a terminating byte at the end. |
| 40 | if (Count < PadTo) { |
| 41 | uint8_t PadValue = Value < 0 ? 0x7f : 0x00; |
| 42 | for (; Count < PadTo - 1; ++Count) |
| 43 | OS << char(PadValue | 0x80); |
| 44 | OS << char(PadValue); |
| 45 | Count++; |
| 46 | } |
| 47 | return Count; |
| 48 | } |
| 49 | |
| 50 | /// Utility function to encode a SLEB128 value to a buffer. Returns |
| 51 | /// the length in bytes of the encoded value. |
| 52 | inline unsigned encodeSLEB128(int64_t Value, uint8_t *p, unsigned PadTo = 0) { |
| 53 | uint8_t *orig_p = p; |
| 54 | unsigned Count = 0; |
| 55 | bool More; |
| 56 | do { |
| 57 | uint8_t Byte = Value & 0x7f; |
| 58 | // NOTE: this assumes that this signed shift is an arithmetic right shift. |
| 59 | Value >>= 7; |
| 60 | More = !((((Value == 0 ) && ((Byte & 0x40) == 0)) || |
| 61 | ((Value == -1) && ((Byte & 0x40) != 0)))); |
| 62 | Count++; |
| 63 | if (More || Count < PadTo) |
| 64 | Byte |= 0x80; // Mark this byte to show that more bytes will follow. |
| 65 | *p++ = Byte; |
| 66 | } while (More); |
| 67 | |
| 68 | // Pad with 0x80 and emit a terminating byte at the end. |
| 69 | if (Count < PadTo) { |
| 70 | uint8_t PadValue = Value < 0 ? 0x7f : 0x00; |
| 71 | for (; Count < PadTo - 1; ++Count) |
| 72 | *p++ = (PadValue | 0x80); |
| 73 | *p++ = PadValue; |
| 74 | } |
| 75 | return (unsigned)(p - orig_p); |
| 76 | } |
| 77 | |
| 78 | /// Utility function to encode a ULEB128 value to an output stream. Returns |
| 79 | /// the length in bytes of the encoded value. |
| 80 | inline unsigned encodeULEB128(uint64_t Value, raw_ostream &OS, |
| 81 | unsigned PadTo = 0) { |
| 82 | unsigned Count = 0; |
| 83 | do { |
| 84 | uint8_t Byte = Value & 0x7f; |
| 85 | Value >>= 7; |
| 86 | Count++; |
| 87 | if (Value != 0 || Count < PadTo) |
| 88 | Byte |= 0x80; // Mark this byte to show that more bytes will follow. |
| 89 | OS << char(Byte); |
| 90 | } while (Value != 0); |
| 91 | |
| 92 | // Pad with 0x80 and emit a null byte at the end. |
| 93 | if (Count < PadTo) { |
| 94 | for (; Count < PadTo - 1; ++Count) |
| 95 | OS << '\x80'; |
| 96 | OS << '\x00'; |
| 97 | Count++; |
| 98 | } |
| 99 | return Count; |
| 100 | } |
| 101 | |
| 102 | /// Utility function to encode a ULEB128 value to a buffer. Returns |
| 103 | /// the length in bytes of the encoded value. |
| 104 | inline unsigned encodeULEB128(uint64_t Value, uint8_t *p, |
| 105 | unsigned PadTo = 0) { |
| 106 | uint8_t *orig_p = p; |
| 107 | unsigned Count = 0; |
| 108 | do { |
| 109 | uint8_t Byte = Value & 0x7f; |
| 110 | Value >>= 7; |
| 111 | Count++; |
| 112 | if (Value != 0 || Count < PadTo) |
| 113 | Byte |= 0x80; // Mark this byte to show that more bytes will follow. |
| 114 | *p++ = Byte; |
| 115 | } while (Value != 0); |
| 116 | |
| 117 | // Pad with 0x80 and emit a null byte at the end. |
| 118 | if (Count < PadTo) { |
| 119 | for (; Count < PadTo - 1; ++Count) |
| 120 | *p++ = '\x80'; |
| 121 | *p++ = '\x00'; |
| 122 | } |
| 123 | |
| 124 | return (unsigned)(p - orig_p); |
| 125 | } |
| 126 | |
| 127 | /// Utility function to decode a ULEB128 value. |
| 128 | inline uint64_t decodeULEB128(const uint8_t *p, unsigned *n = nullptr, |
| 129 | const uint8_t *end = nullptr, |
| 130 | const char **error = nullptr) { |
| 131 | const uint8_t *orig_p = p; |
| 132 | uint64_t Value = 0; |
| 133 | unsigned Shift = 0; |
| 134 | if (error) |
| 135 | *error = nullptr; |
| 136 | do { |
Olivier Deprez | f4ef2d0 | 2021-04-20 13:36:24 +0200 | [diff] [blame] | 137 | if (p == end) { |
Andrew Scull | 5e1ddfa | 2018-08-14 10:06:54 +0100 | [diff] [blame] | 138 | if (error) |
| 139 | *error = "malformed uleb128, extends past end"; |
| 140 | if (n) |
| 141 | *n = (unsigned)(p - orig_p); |
| 142 | return 0; |
| 143 | } |
| 144 | uint64_t Slice = *p & 0x7f; |
| 145 | if (Shift >= 64 || Slice << Shift >> Shift != Slice) { |
| 146 | if (error) |
| 147 | *error = "uleb128 too big for uint64"; |
| 148 | if (n) |
| 149 | *n = (unsigned)(p - orig_p); |
| 150 | return 0; |
| 151 | } |
| 152 | Value += uint64_t(*p & 0x7f) << Shift; |
| 153 | Shift += 7; |
| 154 | } while (*p++ >= 128); |
| 155 | if (n) |
| 156 | *n = (unsigned)(p - orig_p); |
| 157 | return Value; |
| 158 | } |
| 159 | |
| 160 | /// Utility function to decode a SLEB128 value. |
| 161 | inline int64_t decodeSLEB128(const uint8_t *p, unsigned *n = nullptr, |
| 162 | const uint8_t *end = nullptr, |
| 163 | const char **error = nullptr) { |
| 164 | const uint8_t *orig_p = p; |
| 165 | int64_t Value = 0; |
| 166 | unsigned Shift = 0; |
| 167 | uint8_t Byte; |
Andrew Walbran | 3d2c197 | 2020-04-07 12:24:26 +0100 | [diff] [blame] | 168 | if (error) |
| 169 | *error = nullptr; |
Andrew Scull | 5e1ddfa | 2018-08-14 10:06:54 +0100 | [diff] [blame] | 170 | do { |
Olivier Deprez | f4ef2d0 | 2021-04-20 13:36:24 +0200 | [diff] [blame] | 171 | if (p == end) { |
Andrew Scull | 5e1ddfa | 2018-08-14 10:06:54 +0100 | [diff] [blame] | 172 | if (error) |
| 173 | *error = "malformed sleb128, extends past end"; |
| 174 | if (n) |
| 175 | *n = (unsigned)(p - orig_p); |
| 176 | return 0; |
| 177 | } |
| 178 | Byte = *p++; |
Andrew Walbran | 3d2c197 | 2020-04-07 12:24:26 +0100 | [diff] [blame] | 179 | Value |= (uint64_t(Byte & 0x7f) << Shift); |
Andrew Scull | 5e1ddfa | 2018-08-14 10:06:54 +0100 | [diff] [blame] | 180 | Shift += 7; |
| 181 | } while (Byte >= 128); |
Andrew Walbran | 3d2c197 | 2020-04-07 12:24:26 +0100 | [diff] [blame] | 182 | // Sign extend negative numbers if needed. |
| 183 | if (Shift < 64 && (Byte & 0x40)) |
Andrew Scull | 5e1ddfa | 2018-08-14 10:06:54 +0100 | [diff] [blame] | 184 | Value |= (-1ULL) << Shift; |
| 185 | if (n) |
| 186 | *n = (unsigned)(p - orig_p); |
| 187 | return Value; |
| 188 | } |
| 189 | |
| 190 | /// Utility function to get the size of the ULEB128-encoded value. |
| 191 | extern unsigned getULEB128Size(uint64_t Value); |
| 192 | |
| 193 | /// Utility function to get the size of the SLEB128-encoded value. |
| 194 | extern unsigned getSLEB128Size(int64_t Value); |
| 195 | |
| 196 | } // namespace llvm |
| 197 | |
| 198 | #endif // LLVM_SYSTEM_LEB128_H |