include/llvm/Bitcode/BitCodes.h

   1 //===- BitCodes.h - Enum values for the bitcode format ----------*- C++ -*-===//
   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 header Bitcode enum values.
  10 //
  11 // The enum values defined in this file should be considered permanent.  If
  12 // new features are added, they should have values added at the end of the
  13 // respective lists.
  14 //
  15 //===----------------------------------------------------------------------===//
  16
  17 #ifndef LLVM_BITCODE_BITCODES_H
  18 #define LLVM_BITCODE_BITCODES_H
  19
  20 #include "llvm/ADT/SmallVector.h"
  21 #include "llvm/Support/DataTypes.h"
  22 #include "llvm/Support/ErrorHandling.h"
  23 #include <cassert>
  24
  25 namespace llvm {
  26 /// Offsets of the 32-bit fields of bitcode wrapper header.
  27 static const unsigned BWH_MagicField = 0 * 4;
  28 static const unsigned BWH_VersionField = 1 * 4;
  29 static const unsigned BWH_OffsetField = 2 * 4;
  30 static const unsigned BWH_SizeField = 3 * 4;
  31 static const unsigned BWH_CPUTypeField = 4 * 4;
  32 static const unsigned BWH_HeaderSize = 5 * 4;
  33
  34 namespace bitc {
  35   enum StandardWidths {
  36     BlockIDWidth   = 8,  // We use VBR-8 for block IDs.
  37     CodeLenWidth   = 4,  // Codelen are VBR-4.
  38     BlockSizeWidth = 32  // BlockSize up to 2^32 32-bit words = 16GB per block.
  39   };
  40
  41   // The standard abbrev namespace always has a way to exit a block, enter a
  42   // nested block, define abbrevs, and define an unabbreviated record.
  43   enum FixedAbbrevIDs {
  44     END_BLOCK = 0,  // Must be zero to guarantee termination for broken bitcode.
  45     ENTER_SUBBLOCK = 1,
  46
  47     /// DEFINE_ABBREV - Defines an abbrev for the current block.  It consists
  48     /// of a vbr5 for # operand infos.  Each operand info is emitted with a
  49     /// single bit to indicate if it is a literal encoding.  If so, the value is
  50     /// emitted with a vbr8.  If not, the encoding is emitted as 3 bits followed
  51     /// by the info value as a vbr5 if needed.
  52     DEFINE_ABBREV = 2,
  53
  54     // UNABBREV_RECORDs are emitted with a vbr6 for the record code, followed by
  55     // a vbr6 for the # operands, followed by vbr6's for each operand.
  56     UNABBREV_RECORD = 3,
  57
  58     // This is not a code, this is a marker for the first abbrev assignment.
  59     FIRST_APPLICATION_ABBREV = 4
  60   };
  61
  62   /// StandardBlockIDs - All bitcode files can optionally include a BLOCKINFO
  63   /// block, which contains metadata about other blocks in the file.
  64   enum StandardBlockIDs {
  65     /// BLOCKINFO_BLOCK is used to define metadata about blocks, for example,
  66     /// standard abbrevs that should be available to all blocks of a specified
  67     /// ID.
  68     BLOCKINFO_BLOCK_ID = 0,
  69
  70     // Block IDs 1-7 are reserved for future expansion.
  71     FIRST_APPLICATION_BLOCKID = 8
  72   };
  73
  74   /// BlockInfoCodes - The blockinfo block contains metadata about user-defined
  75   /// blocks.
  76   enum BlockInfoCodes {
  77     // DEFINE_ABBREV has magic semantics here, applying to the current SETBID'd
  78     // block, instead of the BlockInfo block.
  79
  80     BLOCKINFO_CODE_SETBID        = 1, // SETBID: [blockid#]
  81     BLOCKINFO_CODE_BLOCKNAME     = 2, // BLOCKNAME: [name]
  82     BLOCKINFO_CODE_SETRECORDNAME = 3  // BLOCKINFO_CODE_SETRECORDNAME:
  83                                       //                             [id, name]
  84   };
  85
  86 } // End bitc namespace
  87
  88 /// BitCodeAbbrevOp - This describes one or more operands in an abbreviation.
  89 /// This is actually a union of two different things:
  90 ///   1. It could be a literal integer value ("the operand is always 17").
  91 ///   2. It could be an encoding specification ("this operand encoded like so").
  92 ///
  93 class BitCodeAbbrevOp {
  94   uint64_t Val;           // A literal value or data for an encoding.
  95   bool IsLiteral : 1;     // Indicate whether this is a literal value or not.
  96   unsigned Enc   : 3;     // The encoding to use.
  97 public:
  98   enum Encoding {
  99     Fixed = 1,  // A fixed width field, Val specifies number of bits.
 100     VBR   = 2,  // A VBR field where Val specifies the width of each chunk.
 101     Array = 3,  // A sequence of fields, next field species elt encoding.
 102     Char6 = 4,  // A 6-bit fixed field which maps to [a-zA-Z0-9._].
 103     Blob  = 5   // 32-bit aligned array of 8-bit characters.
 104   };
 105
 106   explicit BitCodeAbbrevOp(uint64_t V) :  Val(V), IsLiteral(true) {}
 107   explicit BitCodeAbbrevOp(Encoding E, uint64_t Data = 0)
 108     : Val(Data), IsLiteral(false), Enc(E) {}
 109
 110   bool isLiteral() const  { return IsLiteral; }
 111   bool isEncoding() const { return !IsLiteral; }
 112
 113   // Accessors for literals.
 114   uint64_t getLiteralValue() const { assert(isLiteral()); return Val; }
 115
 116   // Accessors for encoding info.
 117   Encoding getEncoding() const { assert(isEncoding()); return (Encoding)Enc; }
 118   uint64_t getEncodingData() const {
 119     assert(isEncoding() && hasEncodingData());
 120     return Val;
 121   }
 122
 123   bool hasEncodingData() const { return hasEncodingData(getEncoding()); }
 124   static bool hasEncodingData(Encoding E) {
 125     switch (E) {
 126     case Fixed:
 127     case VBR:
 128       return true;
 129     case Array:
 130     case Char6:
 131     case Blob:
 132       return false;
 133     }
 134     report_fatal_error("Invalid encoding");
 135   }
 136
 137   /// isChar6 - Return true if this character is legal in the Char6 encoding.
 138   static bool isChar6(char C) {
 139     if (C >= 'a' && C <= 'z') return true;
 140     if (C >= 'A' && C <= 'Z') return true;
 141     if (C >= '0' && C <= '9') return true;
 142     if (C == '.' || C == '_') return true;
 143     return false;
 144   }
 145   static unsigned EncodeChar6(char C) {
 146     if (C >= 'a' && C <= 'z') return C-'a';
 147     if (C >= 'A' && C <= 'Z') return C-'A'+26;
 148     if (C >= '0' && C <= '9') return C-'0'+26+26;
 149     if (C == '.')             return 62;
 150     if (C == '_')             return 63;
 151     llvm_unreachable("Not a value Char6 character!");
 152   }
 153
 154   static char DecodeChar6(unsigned V) {
 155     assert((V & ~63) == 0 && "Not a Char6 encoded character!");
 156     return "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789._"
 157         [V];
 158   }
 159
 160 };
 161
 162 /// BitCodeAbbrev - This class represents an abbreviation record.  An
 163 /// abbreviation allows a complex record that has redundancy to be stored in a
 164 /// specialized format instead of the fully-general, fully-vbr, format.
 165 class BitCodeAbbrev {
 166   SmallVector<BitCodeAbbrevOp, 32> OperandList;
 167
 168 public:
 169   unsigned getNumOperandInfos() const {
 170     return static_cast<unsigned>(OperandList.size());
 171   }
 172   const BitCodeAbbrevOp &getOperandInfo(unsigned N) const {
 173     return OperandList[N];
 174   }
 175
 176   void Add(const BitCodeAbbrevOp &OpInfo) {
 177     OperandList.push_back(OpInfo);
 178   }
 179 };
 180 } // End llvm namespace
 181
 182 #endif