1 //===-- lib/MC/Disassembler.cpp - Disassembler Public C 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 #include "Disassembler.h"
11 #include "llvm-c/Disassembler.h"
12
13 #include "llvm/MC/MCAsmInfo.h"
14 #include "llvm/MC/MCDisassembler.h"
15 #include "llvm/MC/MCInst.h"
16 #include "llvm/MC/MCInstPrinter.h"
17 #include "llvm/MC/MCContext.h"
18 #include "llvm/Target/TargetRegistry.h"
19 #include "llvm/Target/TargetAsmInfo.h" // FIXME.
20 #include "llvm/Target/TargetMachine.h" // FIXME.
21 #include "llvm/Target/TargetSelect.h"
22 #include "llvm/Support/MemoryObject.h"
23
24 namespace llvm {
25 class Target;
26 } // namespace llvm
27 using namespace llvm;
28
29 // LLVMCreateDisasm() creates a disassembler for the TripleName. Symbolic
30 // disassembly is supported by passing a block of information in the DisInfo
31 // parameter and specifying the TagType and callback functions as described in
32 // the header llvm-c/Disassembler.h . The pointer to the block and the
33 // functions can all be passed as NULL. If successful, this returns a
34 // disassembler context. If not, it returns NULL.
35 //
LLVMCreateDisasm(const char * TripleName,void * DisInfo,int TagType,LLVMOpInfoCallback GetOpInfo,LLVMSymbolLookupCallback SymbolLookUp)36 LLVMDisasmContextRef LLVMCreateDisasm(const char *TripleName, void *DisInfo,
37 int TagType, LLVMOpInfoCallback GetOpInfo,
38 LLVMSymbolLookupCallback SymbolLookUp) {
39 // Initialize targets and assembly printers/parsers.
40 llvm::InitializeAllTargetInfos();
41 // FIXME: We shouldn't need to initialize the Target(Machine)s.
42 llvm::InitializeAllTargets();
43 llvm::InitializeAllMCAsmInfos();
44 llvm::InitializeAllMCCodeGenInfos();
45 llvm::InitializeAllMCRegisterInfos();
46 llvm::InitializeAllAsmPrinters();
47 llvm::InitializeAllAsmParsers();
48 llvm::InitializeAllDisassemblers();
49
50 // Get the target.
51 std::string Error;
52 const Target *TheTarget = TargetRegistry::lookupTarget(TripleName, Error);
53 assert(TheTarget && "Unable to create target!");
54
55 // Get the assembler info needed to setup the MCContext.
56 const MCAsmInfo *MAI = TheTarget->createMCAsmInfo(TripleName);
57 assert(MAI && "Unable to create target asm info!");
58
59 const MCRegisterInfo *MRI = TheTarget->createMCRegInfo(TripleName);
60 assert(MRI && "Unable to create target register info!");
61
62 // Package up features to be passed to target/subtarget
63 std::string FeaturesStr;
64 std::string CPU;
65
66 // FIXME: We shouldn't need to do this (and link in codegen).
67 // When we split this out, we should do it in a way that makes
68 // it straightforward to switch subtargets on the fly.
69 TargetMachine *TM = TheTarget->createTargetMachine(TripleName, CPU,
70 FeaturesStr);
71 assert(TM && "Unable to create target machine!");
72
73 // Get the target assembler info needed to setup the context.
74 const TargetAsmInfo *tai = new TargetAsmInfo(*TM);
75 assert(tai && "Unable to create target assembler!");
76
77 // Set up the MCContext for creating symbols and MCExpr's.
78 MCContext *Ctx = new MCContext(*MAI, *MRI, 0, tai);
79 assert(Ctx && "Unable to create MCContext!");
80
81 // Set up disassembler.
82 MCDisassembler *DisAsm = TheTarget->createMCDisassembler();
83 assert(DisAsm && "Unable to create disassembler!");
84 DisAsm->setupForSymbolicDisassembly(GetOpInfo, DisInfo, Ctx);
85
86 // Set up the instruction printer.
87 int AsmPrinterVariant = MAI->getAssemblerDialect();
88 MCInstPrinter *IP = TheTarget->createMCInstPrinter(AsmPrinterVariant,
89 *MAI);
90 assert(IP && "Unable to create instruction printer!");
91
92 LLVMDisasmContext *DC = new LLVMDisasmContext(TripleName, DisInfo, TagType,
93 GetOpInfo, SymbolLookUp,
94 TheTarget, MAI, MRI, TM, tai,
95 Ctx, DisAsm, IP);
96 assert(DC && "Allocation failure!");
97 return DC;
98 }
99
100 //
101 // LLVMDisasmDispose() disposes of the disassembler specified by the context.
102 //
LLVMDisasmDispose(LLVMDisasmContextRef DCR)103 void LLVMDisasmDispose(LLVMDisasmContextRef DCR){
104 LLVMDisasmContext *DC = (LLVMDisasmContext *)DCR;
105 delete DC;
106 }
107
108 namespace {
109 //
110 // The memory object created by LLVMDisasmInstruction().
111 //
112 class DisasmMemoryObject : public MemoryObject {
113 uint8_t *Bytes;
114 uint64_t Size;
115 uint64_t BasePC;
116 public:
DisasmMemoryObject(uint8_t * bytes,uint64_t size,uint64_t basePC)117 DisasmMemoryObject(uint8_t *bytes, uint64_t size, uint64_t basePC) :
118 Bytes(bytes), Size(size), BasePC(basePC) {}
119
getBase() const120 uint64_t getBase() const { return BasePC; }
getExtent() const121 uint64_t getExtent() const { return Size; }
122
readByte(uint64_t Addr,uint8_t * Byte) const123 int readByte(uint64_t Addr, uint8_t *Byte) const {
124 if (Addr - BasePC >= Size)
125 return -1;
126 *Byte = Bytes[Addr - BasePC];
127 return 0;
128 }
129 };
130 } // end anonymous namespace
131
132 //
133 // LLVMDisasmInstruction() disassembles a single instruction using the
134 // disassembler context specified in the parameter DC. The bytes of the
135 // instruction are specified in the parameter Bytes, and contains at least
136 // BytesSize number of bytes. The instruction is at the address specified by
137 // the PC parameter. If a valid instruction can be disassembled its string is
138 // returned indirectly in OutString which whos size is specified in the
139 // parameter OutStringSize. This function returns the number of bytes in the
140 // instruction or zero if there was no valid instruction. If this function
141 // returns zero the caller will have to pick how many bytes they want to step
142 // over by printing a .byte, .long etc. to continue.
143 //
LLVMDisasmInstruction(LLVMDisasmContextRef DCR,uint8_t * Bytes,uint64_t BytesSize,uint64_t PC,char * OutString,size_t OutStringSize)144 size_t LLVMDisasmInstruction(LLVMDisasmContextRef DCR, uint8_t *Bytes,
145 uint64_t BytesSize, uint64_t PC, char *OutString,
146 size_t OutStringSize){
147 LLVMDisasmContext *DC = (LLVMDisasmContext *)DCR;
148 // Wrap the pointer to the Bytes, BytesSize and PC in a MemoryObject.
149 DisasmMemoryObject MemoryObject(Bytes, BytesSize, PC);
150
151 uint64_t Size;
152 MCInst Inst;
153 const MCDisassembler *DisAsm = DC->getDisAsm();
154 MCInstPrinter *IP = DC->getIP();
155 if (!DisAsm->getInstruction(Inst, Size, MemoryObject, PC, /*REMOVE*/ nulls()))
156 return 0;
157
158 SmallVector<char, 64> InsnStr;
159 raw_svector_ostream OS(InsnStr);
160 IP->printInst(&Inst, OS);
161 OS.flush();
162
163 assert(OutStringSize != 0 && "Output buffer cannot be zero size");
164 size_t OutputSize = std::min(OutStringSize-1, InsnStr.size());
165 std::memcpy(OutString, InsnStr.data(), OutputSize);
166 OutString[OutputSize] = '\0'; // Terminate string.
167
168 return Size;
169 }
170