1251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan//===-- llvm/MC/MCDisassembler.h - Disassembler interface -------*- C++ -*-===//
2251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan//
3251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan//                     The LLVM Compiler Infrastructure
4251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan//
5251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan// This file is distributed under the University of Illinois Open Source
6251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan// License. See LICENSE.TXT for details.
7251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan//
8251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan//===----------------------------------------------------------------------===//
9de2d8694e25a814696358e95141f4b1aa4d8847ePirama Arumuga Nainar#ifndef LLVM_MC_MCDISASSEMBLER_MCDISASSEMBLER_H
10de2d8694e25a814696358e95141f4b1aa4d8847ePirama Arumuga Nainar#define LLVM_MC_MCDISASSEMBLER_MCDISASSEMBLER_H
11251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan
12bd3327654b5708f1ba92aff3ab25b1bbf5034797Kevin Enderby#include "llvm-c/Disassembler.h"
13de2d8694e25a814696358e95141f4b1aa4d8847ePirama Arumuga Nainar#include "llvm/MC/MCDisassembler/MCSymbolizer.h"
14255f89faee13dc491cb64fbeae3c763e7e2ea4e6Chandler Carruth#include "llvm/Support/DataTypes.h"
15251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan
16251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanannamespace llvm {
17639aa87bee77fe2d83f0978ae1eea53e49def324Jim Grosbach
18de2d8694e25a814696358e95141f4b1aa4d8847ePirama Arumuga Nainartemplate <typename T> class ArrayRef;
19251ef612a812ac99edeab6c08a752bf8ca220921Sean Callananclass MCInst;
20b950585cc5a0d665e9accfe5ce490cd269756f2eJames Molloyclass MCSubtargetInfo;
21251ef612a812ac99edeab6c08a752bf8ca220921Sean Callananclass raw_ostream;
22bd3327654b5708f1ba92aff3ab25b1bbf5034797Kevin Enderbyclass MCContext;
23639aa87bee77fe2d83f0978ae1eea53e49def324Jim Grosbach
2437ed9c199ca639565f6ce88105f9e39e898d82d0Stephen Hines/// Superclass for all disassemblers. Consumes a memory region and provides an
2537ed9c199ca639565f6ce88105f9e39e898d82d0Stephen Hines/// array of assembly instructions.
26251ef612a812ac99edeab6c08a752bf8ca220921Sean Callananclass MCDisassembler {
27251ef612a812ac99edeab6c08a752bf8ca220921Sean Callananpublic:
2883e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  /// Ternary decode status. Most backends will just use Fail and
2983e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  /// Success, however some have a concept of an instruction with
3083e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  /// understandable semantics but which is architecturally
3183e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  /// incorrect. An example of this is ARM UNPREDICTABLE instructions
3283e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  /// which are disassemblable but cause undefined behaviour.
3383e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  ///
3483e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  /// Because it makes sense to disassemble these instructions, there
3583e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  /// is a "soft fail" failure mode that indicates the MCInst& is
3683e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  /// valid but architecturally incorrect.
3783e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  ///
3883e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  /// The enum numbers are deliberately chosen such that reduction
3983e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  /// from Success->SoftFail ->Fail can be done with a simple
4083e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  /// bitwise-AND:
4183e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  ///
4283e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  ///   LEFT & TOP =  | Success       Unpredictable   Fail
4383e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  ///   --------------+-----------------------------------
4483e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  ///   Success       | Success       Unpredictable   Fail
4583e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  ///   Unpredictable | Unpredictable Unpredictable   Fail
4683e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  ///   Fail          | Fail          Fail            Fail
4783e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  ///
4883e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  /// An easy way of encoding this is as 0b11, 0b01, 0b00 for
4983e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  /// Success, SoftFail, Fail respectively.
5083e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  enum DecodeStatus {
5183e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson    Fail = 0,
5283e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson    SoftFail = 1,
5383e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson    Success = 3
5483e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  };
5583e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson
56dce4a407a24b04eebc6a376f8e62b41aaa7b071fStephen Hines  MCDisassembler(const MCSubtargetInfo &STI, MCContext &Ctx)
57dce4a407a24b04eebc6a376f8e62b41aaa7b071fStephen Hines    : Ctx(Ctx), STI(STI), Symbolizer(), CommentStream(nullptr) {}
58639aa87bee77fe2d83f0978ae1eea53e49def324Jim Grosbach
59251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan  virtual ~MCDisassembler();
60639aa87bee77fe2d83f0978ae1eea53e49def324Jim Grosbach
6137ed9c199ca639565f6ce88105f9e39e898d82d0Stephen Hines  /// Returns the disassembly of a single instruction.
62251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan  ///
636948897e478cbd66626159776a8017b3c18579b9Pirama Arumuga Nainar  /// \param Instr    - An MCInst to populate with the contents of the
64251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan  ///                   instruction.
656948897e478cbd66626159776a8017b3c18579b9Pirama Arumuga Nainar  /// \param Size     - A value to populate with the size of the instruction, or
66251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan  ///                   the number of bytes consumed while attempting to decode
67251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan  ///                   an invalid instruction.
686948897e478cbd66626159776a8017b3c18579b9Pirama Arumuga Nainar  /// \param Address  - The address, in the memory space of region, of the first
69251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan  ///                   byte of the instruction.
706948897e478cbd66626159776a8017b3c18579b9Pirama Arumuga Nainar  /// \param VStream  - The stream to print warnings and diagnostic messages on.
716948897e478cbd66626159776a8017b3c18579b9Pirama Arumuga Nainar  /// \param CStream  - The stream to print comments and annotations on.
726948897e478cbd66626159776a8017b3c18579b9Pirama Arumuga Nainar  /// \return         - MCDisassembler::Success if the instruction is valid,
73639aa87bee77fe2d83f0978ae1eea53e49def324Jim Grosbach  ///                   MCDisassembler::SoftFail if the instruction was
7483e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  ///                                            disassemblable but invalid,
7583e3f67fb68d497b600da83a62f000fcce7868a9Owen Anderson  ///                   MCDisassembler::Fail if the instruction was invalid.
7637ed9c199ca639565f6ce88105f9e39e898d82d0Stephen Hines  virtual DecodeStatus getInstruction(MCInst &Instr, uint64_t &Size,
7737ed9c199ca639565f6ce88105f9e39e898d82d0Stephen Hines                                      ArrayRef<uint8_t> Bytes, uint64_t Address,
7837ed9c199ca639565f6ce88105f9e39e898d82d0Stephen Hines                                      raw_ostream &VStream,
7937ed9c199ca639565f6ce88105f9e39e898d82d0Stephen Hines                                      raw_ostream &CStream) const = 0;
8037ed9c199ca639565f6ce88105f9e39e898d82d0Stephen Hines
81de7cbbfcce5c068f0699bdcb6dac093c0c91ba6fQuentin Colombetprivate:
82dce4a407a24b04eebc6a376f8e62b41aaa7b071fStephen Hines  MCContext &Ctx;
839899f70a7406d632c82849978bf6981f1ee4ccb5Sean Callanan
84b950585cc5a0d665e9accfe5ce490cd269756f2eJames Molloyprotected:
85b950585cc5a0d665e9accfe5ce490cd269756f2eJames Molloy  // Subtarget information, for instruction decoding predicates if required.
86b950585cc5a0d665e9accfe5ce490cd269756f2eJames Molloy  const MCSubtargetInfo &STI;
8736b56886974eae4f9c5ebc96befd3e7bfe5de338Stephen Hines  std::unique_ptr<MCSymbolizer> Symbolizer;
882c94d0faa0e1c268893d5e04dc77e8a35889db00Ahmed Bougacha
89bd3327654b5708f1ba92aff3ab25b1bbf5034797Kevin Enderbypublic:
902c94d0faa0e1c268893d5e04dc77e8a35889db00Ahmed Bougacha  // Helpers around MCSymbolizer
912c94d0faa0e1c268893d5e04dc77e8a35889db00Ahmed Bougacha  bool tryAddingSymbolicOperand(MCInst &Inst,
922c94d0faa0e1c268893d5e04dc77e8a35889db00Ahmed Bougacha                                int64_t Value,
932c94d0faa0e1c268893d5e04dc77e8a35889db00Ahmed Bougacha                                uint64_t Address, bool IsBranch,
942c94d0faa0e1c268893d5e04dc77e8a35889db00Ahmed Bougacha                                uint64_t Offset, uint64_t InstSize) const;
952c94d0faa0e1c268893d5e04dc77e8a35889db00Ahmed Bougacha
962c94d0faa0e1c268893d5e04dc77e8a35889db00Ahmed Bougacha  void tryAddingPcLoadReferenceComment(int64_t Value, uint64_t Address) const;
972c94d0faa0e1c268893d5e04dc77e8a35889db00Ahmed Bougacha
982c94d0faa0e1c268893d5e04dc77e8a35889db00Ahmed Bougacha  /// Set \p Symzer as the current symbolizer.
992c94d0faa0e1c268893d5e04dc77e8a35889db00Ahmed Bougacha  /// This takes ownership of \p Symzer, and deletes the previously set one.
10036b56886974eae4f9c5ebc96befd3e7bfe5de338Stephen Hines  void setSymbolizer(std::unique_ptr<MCSymbolizer> Symzer);
1012c94d0faa0e1c268893d5e04dc77e8a35889db00Ahmed Bougacha
102dce4a407a24b04eebc6a376f8e62b41aaa7b071fStephen Hines  MCContext& getContext() const { return Ctx; }
103de7cbbfcce5c068f0699bdcb6dac093c0c91ba6fQuentin Colombet
10462ab26548fee37422fb90daaac22fc1140be797dArtyom Skrobov  const MCSubtargetInfo& getSubtargetInfo() const { return STI; }
10562ab26548fee37422fb90daaac22fc1140be797dArtyom Skrobov
1069e5887b17e634b98f7c1cf0ee4f25c218097d08eKevin Enderby  // Marked mutable because we cache it inside the disassembler, rather than
1079e5887b17e634b98f7c1cf0ee4f25c218097d08eKevin Enderby  // having to pass it around as an argument through all the autogenerated code.
1089e5887b17e634b98f7c1cf0ee4f25c218097d08eKevin Enderby  mutable raw_ostream *CommentStream;
1099899f70a7406d632c82849978bf6981f1ee4ccb5Sean Callanan};
110251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan
111251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan} // namespace llvm
112251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan
113251ef612a812ac99edeab6c08a752bf8ca220921Sean Callanan#endif
114