1 //===-- lib/MC/Disassembler.cpp - Disassembler Public C Interface ---------===//
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 #include "Disassembler.h"
11 #include "llvm-c/Disassembler.h"
12 
13 #include "llvm/MC/MCAsmInfo.h"
14 #include "llvm/MC/MCContext.h"
15 #include "llvm/MC/MCDisassembler.h"
16 #include "llvm/MC/MCInst.h"
17 #include "llvm/MC/MCInstPrinter.h"
18 #include "llvm/MC/MCRegisterInfo.h"
19 #include "llvm/Support/MemoryObject.h"
20 #include "llvm/Support/TargetRegistry.h"
21 
22 namespace llvm {
23 class Target;
24 } // namespace llvm
25 using namespace llvm;
26 
27 // LLVMCreateDisasm() creates a disassembler for the TripleName.  Symbolic
28 // disassembly is supported by passing a block of information in the DisInfo
29 // parameter and specifying the TagType and callback functions as described in
30 // the header llvm-c/Disassembler.h .  The pointer to the block and the
31 // functions can all be passed as NULL.  If successful, this returns a
32 // disassembler context.  If not, it returns NULL.
33 //
34 LLVMDisasmContextRef LLVMCreateDisasm(const char *TripleName, void *DisInfo,
35                                       int TagType, LLVMOpInfoCallback GetOpInfo,
36                                       LLVMSymbolLookupCallback SymbolLookUp) {
37   // Get the target.
38   std::string Error;
39   const Target *TheTarget = TargetRegistry::lookupTarget(TripleName, Error);
40   assert(TheTarget && "Unable to create target!");
41 
42   // Get the assembler info needed to setup the MCContext.
43   const MCAsmInfo *MAI = TheTarget->createMCAsmInfo(TripleName);
44   assert(MAI && "Unable to create target asm info!");
45 
46   const MCRegisterInfo *MRI = TheTarget->createMCRegInfo(TripleName);
47   assert(MRI && "Unable to create target register info!");
48 
49   // Package up features to be passed to target/subtarget
50   std::string FeaturesStr;
51   std::string CPU;
52 
53   const MCSubtargetInfo *STI = TheTarget->createMCSubtargetInfo(TripleName, CPU,
54                                                                 FeaturesStr);
55   assert(STI && "Unable to create subtarget info!");
56 
57   // Set up the MCContext for creating symbols and MCExpr's.
58   MCContext *Ctx = new MCContext(*MAI, *MRI, 0);
59   assert(Ctx && "Unable to create MCContext!");
60 
61   // Set up disassembler.
62   MCDisassembler *DisAsm = TheTarget->createMCDisassembler(*STI);
63   assert(DisAsm && "Unable to create disassembler!");
64   DisAsm->setupForSymbolicDisassembly(GetOpInfo, SymbolLookUp, DisInfo, Ctx);
65 
66   // Set up the instruction printer.
67   int AsmPrinterVariant = MAI->getAssemblerDialect();
68   MCInstPrinter *IP = TheTarget->createMCInstPrinter(AsmPrinterVariant,
69                                                      *MAI, *STI);
70   assert(IP && "Unable to create instruction printer!");
71 
72   LLVMDisasmContext *DC = new LLVMDisasmContext(TripleName, DisInfo, TagType,
73                                                 GetOpInfo, SymbolLookUp,
74                                                 TheTarget, MAI, MRI,
75                                                 Ctx, DisAsm, IP);
76   assert(DC && "Allocation failure!");
77 
78   return DC;
79 }
80 
81 //
82 // LLVMDisasmDispose() disposes of the disassembler specified by the context.
83 //
84 void LLVMDisasmDispose(LLVMDisasmContextRef DCR){
85   LLVMDisasmContext *DC = (LLVMDisasmContext *)DCR;
86   delete DC;
87 }
88 
89 namespace {
90 //
91 // The memory object created by LLVMDisasmInstruction().
92 //
93 class DisasmMemoryObject : public MemoryObject {
94   uint8_t *Bytes;
95   uint64_t Size;
96   uint64_t BasePC;
97 public:
98   DisasmMemoryObject(uint8_t *bytes, uint64_t size, uint64_t basePC) :
99                      Bytes(bytes), Size(size), BasePC(basePC) {}
100 
101   uint64_t getBase() const { return BasePC; }
102   uint64_t getExtent() const { return Size; }
103 
104   int readByte(uint64_t Addr, uint8_t *Byte) const {
105     if (Addr - BasePC >= Size)
106       return -1;
107     *Byte = Bytes[Addr - BasePC];
108     return 0;
109   }
110 };
111 } // end anonymous namespace
112 
113 //
114 // LLVMDisasmInstruction() disassembles a single instruction using the
115 // disassembler context specified in the parameter DC.  The bytes of the
116 // instruction are specified in the parameter Bytes, and contains at least
117 // BytesSize number of bytes.  The instruction is at the address specified by
118 // the PC parameter.  If a valid instruction can be disassembled its string is
119 // returned indirectly in OutString which whos size is specified in the
120 // parameter OutStringSize.  This function returns the number of bytes in the
121 // instruction or zero if there was no valid instruction.  If this function
122 // returns zero the caller will have to pick how many bytes they want to step
123 // over by printing a .byte, .long etc. to continue.
124 //
125 size_t LLVMDisasmInstruction(LLVMDisasmContextRef DCR, uint8_t *Bytes,
126                              uint64_t BytesSize, uint64_t PC, char *OutString,
127                              size_t OutStringSize){
128   LLVMDisasmContext *DC = (LLVMDisasmContext *)DCR;
129   // Wrap the pointer to the Bytes, BytesSize and PC in a MemoryObject.
130   DisasmMemoryObject MemoryObject(Bytes, BytesSize, PC);
131 
132   uint64_t Size;
133   MCInst Inst;
134   const MCDisassembler *DisAsm = DC->getDisAsm();
135   MCInstPrinter *IP = DC->getIP();
136   MCDisassembler::DecodeStatus S;
137   S = DisAsm->getInstruction(Inst, Size, MemoryObject, PC,
138                              /*REMOVE*/ nulls(), DC->CommentStream);
139   switch (S) {
140   case MCDisassembler::Fail:
141   case MCDisassembler::SoftFail:
142     // FIXME: Do something different for soft failure modes?
143     return 0;
144 
145   case MCDisassembler::Success: {
146     DC->CommentStream.flush();
147     StringRef Comments = DC->CommentsToEmit.str();
148 
149     SmallVector<char, 64> InsnStr;
150     raw_svector_ostream OS(InsnStr);
151     IP->printInst(&Inst, OS, Comments);
152     OS.flush();
153 
154     // Tell the comment stream that the vector changed underneath it.
155     DC->CommentsToEmit.clear();
156     DC->CommentStream.resync();
157 
158     assert(OutStringSize != 0 && "Output buffer cannot be zero size");
159     size_t OutputSize = std::min(OutStringSize-1, InsnStr.size());
160     std::memcpy(OutString, InsnStr.data(), OutputSize);
161     OutString[OutputSize] = '\0'; // Terminate string.
162 
163     return Size;
164   }
165   }
166   return 0;
167 }
168