Line data    Source code

       1             : //===- llvm/MC/MCDisassembler.h - Disassembler interface --------*- C++ -*-===//
       2             : //
       3             : //                     The LLVM Compiler Infrastructure
       4             : //
       5             : // This file is distributed under the University of Illinois Open Source
       6             : // License. See LICENSE.TXT for details.
       7             : //
       8             : //===----------------------------------------------------------------------===//
       9             : 
      10             : #ifndef LLVM_MC_MCDISASSEMBLER_MCDISASSEMBLER_H
      11             : #define LLVM_MC_MCDISASSEMBLER_MCDISASSEMBLER_H
      12             : 
      13             : #include "llvm/MC/MCDisassembler/MCSymbolizer.h"
      14             : #include <cstdint>
      15             : #include <memory>
      16             : 
      17             : namespace llvm {
      18             : 
      19             : template <typename T> class ArrayRef;
      20             : class MCContext;
      21             : class MCInst;
      22             : class MCSubtargetInfo;
      23             : class raw_ostream;
      24             : 
      25             : /// Superclass for all disassemblers. Consumes a memory region and provides an
      26             : /// array of assembly instructions.
      27        2305 : class MCDisassembler {
      28             : public:
      29             :   /// Ternary decode status. Most backends will just use Fail and
      30             :   /// Success, however some have a concept of an instruction with
      31             :   /// understandable semantics but which is architecturally
      32             :   /// incorrect. An example of this is ARM UNPREDICTABLE instructions
      33             :   /// which are disassemblable but cause undefined behaviour.
      34             :   ///
      35             :   /// Because it makes sense to disassemble these instructions, there
      36             :   /// is a "soft fail" failure mode that indicates the MCInst& is
      37             :   /// valid but architecturally incorrect.
      38             :   ///
      39             :   /// The enum numbers are deliberately chosen such that reduction
      40             :   /// from Success->SoftFail ->Fail can be done with a simple
      41             :   /// bitwise-AND:
      42             :   ///
      43             :   ///   LEFT & TOP =  | Success       Unpredictable   Fail
      44             :   ///   --------------+-----------------------------------
      45             :   ///   Success       | Success       Unpredictable   Fail
      46             :   ///   Unpredictable | Unpredictable Unpredictable   Fail
      47             :   ///   Fail          | Fail          Fail            Fail
      48             :   ///
      49             :   /// An easy way of encoding this is as 0b11, 0b01, 0b00 for
      50             :   /// Success, SoftFail, Fail respectively.
      51             :   enum DecodeStatus {
      52             :     Fail = 0,
      53             :     SoftFail = 1,
      54             :     Success = 3
      55             :   };
      56             : 
      57             :   MCDisassembler(const MCSubtargetInfo &STI, MCContext &Ctx)
      58        4616 :     : Ctx(Ctx), STI(STI) {}
      59             : 
      60             :   virtual ~MCDisassembler();
      61             : 
      62             :   /// Returns the disassembly of a single instruction.
      63             :   ///
      64             :   /// \param Instr    - An MCInst to populate with the contents of the
      65             :   ///                   instruction.
      66             :   /// \param Size     - A value to populate with the size of the instruction, or
      67             :   ///                   the number of bytes consumed while attempting to decode
      68             :   ///                   an invalid instruction.
      69             :   /// \param Address  - The address, in the memory space of region, of the first
      70             :   ///                   byte of the instruction.
      71             :   /// \param Bytes    - A reference to the actual bytes of the instruction.
      72             :   /// \param VStream  - The stream to print warnings and diagnostic messages on.
      73             :   /// \param CStream  - The stream to print comments and annotations on.
      74             :   /// \return         - MCDisassembler::Success if the instruction is valid,
      75             :   ///                   MCDisassembler::SoftFail if the instruction was
      76             :   ///                                            disassemblable but invalid,
      77             :   ///                   MCDisassembler::Fail if the instruction was invalid.
      78             :   virtual DecodeStatus getInstruction(MCInst &Instr, uint64_t &Size,
      79             :                                       ArrayRef<uint8_t> Bytes, uint64_t Address,
      80             :                                       raw_ostream &VStream,
      81             :                                       raw_ostream &CStream) const = 0;
      82             : 
      83             : private:
      84             :   MCContext &Ctx;
      85             : 
      86             : protected:
      87             :   // Subtarget information, for instruction decoding predicates if required.
      88             :   const MCSubtargetInfo &STI;
      89             :   std::unique_ptr<MCSymbolizer> Symbolizer;
      90             : 
      91             : public:
      92             :   // Helpers around MCSymbolizer
      93             :   bool tryAddingSymbolicOperand(MCInst &Inst,
      94             :                                 int64_t Value,
      95             :                                 uint64_t Address, bool IsBranch,
      96             :                                 uint64_t Offset, uint64_t InstSize) const;
      97             : 
      98             :   void tryAddingPcLoadReferenceComment(int64_t Value, uint64_t Address) const;
      99             : 
     100             :   /// Set \p Symzer as the current symbolizer.
     101             :   /// This takes ownership of \p Symzer, and deletes the previously set one.
     102             :   void setSymbolizer(std::unique_ptr<MCSymbolizer> Symzer);
     103             : 
     104           0 :   MCContext& getContext() const { return Ctx; }
     105             : 
     106           0 :   const MCSubtargetInfo& getSubtargetInfo() const { return STI; }
     107             : 
     108             :   // Marked mutable because we cache it inside the disassembler, rather than
     109             :   // having to pass it around as an argument through all the autogenerated code.
     110             :   mutable raw_ostream *CommentStream = nullptr;
     111             : };
     112             : 
     113             : } // end namespace llvm
     114             : 
     115             : #endif // LLVM_MC_MCDISASSEMBLER_MCDISASSEMBLER_H