1/*===-- llvm-c/Disassembler.h - 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|* This header provides a public interface to a disassembler library.         *|
11|* LLVM provides an implementation of this interface.                         *|
12|*                                                                            *|
13\*===----------------------------------------------------------------------===*/
14
15#ifndef LLVM_C_DISASSEMBLER_H
16#define LLVM_C_DISASSEMBLER_H
17
18#include "llvm/Support/DataTypes.h"
19#ifdef __cplusplus
20#include <cstddef>
21#else
22#include <stddef.h>
23#endif
24
25/**
26 * @defgroup LLVMCDisassembler Disassembler
27 * @ingroup LLVMC
28 *
29 * @{
30 */
31
32/**
33 * An opaque reference to a disassembler context.
34 */
35typedef void *LLVMDisasmContextRef;
36
37/**
38 * The type for the operand information call back function.  This is called to
39 * get the symbolic information for an operand of an instruction.  Typically
40 * this is from the relocation information, symbol table, etc.  That block of
41 * information is saved when the disassembler context is created and passed to
42 * the call back in the DisInfo parameter.  The instruction containing operand
43 * is at the PC parameter.  For some instruction sets, there can be more than
44 * one operand with symbolic information.  To determine the symbolic operand
45 * information for each operand, the bytes for the specific operand in the
46 * instruction are specified by the Offset parameter and its byte widith is the
47 * size parameter.  For instructions sets with fixed widths and one symbolic
48 * operand per instruction, the Offset parameter will be zero and Size parameter
49 * will be the instruction width.  The information is returned in TagBuf and is
50 * Triple specific with its specific information defined by the value of
51 * TagType for that Triple.  If symbolic information is returned the function
52 * returns 1, otherwise it returns 0.
53 */
54typedef int (*LLVMOpInfoCallback)(void *DisInfo, uint64_t PC,
55                                  uint64_t Offset, uint64_t Size,
56                                  int TagType, void *TagBuf);
57
58/**
59 * The initial support in LLVM MC for the most general form of a relocatable
60 * expression is "AddSymbol - SubtractSymbol + Offset".  For some Darwin targets
61 * this full form is encoded in the relocation information so that AddSymbol and
62 * SubtractSymbol can be link edited independent of each other.  Many other
63 * platforms only allow a relocatable expression of the form AddSymbol + Offset
64 * to be encoded.
65 *
66 * The LLVMOpInfoCallback() for the TagType value of 1 uses the struct
67 * LLVMOpInfo1.  The value of the relocatable expression for the operand,
68 * including any PC adjustment, is passed in to the call back in the Value
69 * field.  The symbolic information about the operand is returned using all
70 * the fields of the structure with the Offset of the relocatable expression
71 * returned in the Value field.  It is possible that some symbols in the
72 * relocatable expression were assembly temporary symbols, for example
73 * "Ldata - LpicBase + constant", and only the Values of the symbols without
74 * symbol names are present in the relocation information.  The VariantKind
75 * type is one of the Target specific #defines below and is used to print
76 * operands like "_foo@GOT", ":lower16:_foo", etc.
77 */
78struct LLVMOpInfoSymbol1 {
79  uint64_t Present;  /* 1 if this symbol is present */
80  const char *Name;  /* symbol name if not NULL */
81  uint64_t Value;    /* symbol value if name is NULL */
82};
83
84struct LLVMOpInfo1 {
85  struct LLVMOpInfoSymbol1 AddSymbol;
86  struct LLVMOpInfoSymbol1 SubtractSymbol;
87  uint64_t Value;
88  uint64_t VariantKind;
89};
90
91/**
92 * The operand VariantKinds for symbolic disassembly.
93 */
94#define LLVMDisassembler_VariantKind_None 0 /* all targets */
95
96/**
97 * The ARM target VariantKinds.
98 */
99#define LLVMDisassembler_VariantKind_ARM_HI16 1 /* :upper16: */
100#define LLVMDisassembler_VariantKind_ARM_LO16 2 /* :lower16: */
101
102/**
103 * The ARM64 target VariantKinds.
104 */
105#define LLVMDisassembler_VariantKind_ARM64_PAGE       1 /* @page */
106#define LLVMDisassembler_VariantKind_ARM64_PAGEOFF    2 /* @pageoff */
107#define LLVMDisassembler_VariantKind_ARM64_GOTPAGE    3 /* @gotpage */
108#define LLVMDisassembler_VariantKind_ARM64_GOTPAGEOFF 4 /* @gotpageoff */
109#define LLVMDisassembler_VariantKind_ARM64_TLVP       5 /* @tvlppage */
110#define LLVMDisassembler_VariantKind_ARM64_TLVOFF     6 /* @tvlppageoff */
111
112/**
113 * The type for the symbol lookup function.  This may be called by the
114 * disassembler for things like adding a comment for a PC plus a constant
115 * offset load instruction to use a symbol name instead of a load address value.
116 * It is passed the block information is saved when the disassembler context is
117 * created and the ReferenceValue to look up as a symbol.  If no symbol is found
118 * for the ReferenceValue NULL is returned.  The ReferenceType of the
119 * instruction is passed indirectly as is the PC of the instruction in
120 * ReferencePC.  If the output reference can be determined its type is returned
121 * indirectly in ReferenceType along with ReferenceName if any, or that is set
122 * to NULL.
123 */
124typedef const char *(*LLVMSymbolLookupCallback)(void *DisInfo,
125                                                uint64_t ReferenceValue,
126                                                uint64_t *ReferenceType,
127                                                uint64_t ReferencePC,
128                                                const char **ReferenceName);
129/**
130 * The reference types on input and output.
131 */
132/* No input reference type or no output reference type. */
133#define LLVMDisassembler_ReferenceType_InOut_None 0
134
135/* The input reference is from a branch instruction. */
136#define LLVMDisassembler_ReferenceType_In_Branch 1
137/* The input reference is from a PC relative load instruction. */
138#define LLVMDisassembler_ReferenceType_In_PCrel_Load 2
139
140/* The input reference is from an ARM64::ADRP instruction. */
141#define LLVMDisassembler_ReferenceType_In_ARM64_ADRP 0x100000001
142/* The input reference is from an ARM64::ADDXri instruction. */
143#define LLVMDisassembler_ReferenceType_In_ARM64_ADDXri 0x100000002
144/* The input reference is from an ARM64::LDRXui instruction. */
145#define LLVMDisassembler_ReferenceType_In_ARM64_LDRXui 0x100000003
146/* The input reference is from an ARM64::LDRXl instruction. */
147#define LLVMDisassembler_ReferenceType_In_ARM64_LDRXl 0x100000004
148/* The input reference is from an ARM64::ADR instruction. */
149#define LLVMDisassembler_ReferenceType_In_ARM64_ADR 0x100000005
150
151/* The output reference is to as symbol stub. */
152#define LLVMDisassembler_ReferenceType_Out_SymbolStub 1
153/* The output reference is to a symbol address in a literal pool. */
154#define LLVMDisassembler_ReferenceType_Out_LitPool_SymAddr 2
155/* The output reference is to a cstring address in a literal pool. */
156#define LLVMDisassembler_ReferenceType_Out_LitPool_CstrAddr 3
157
158/* The output reference is to a Objective-C CoreFoundation string. */
159#define LLVMDisassembler_ReferenceType_Out_Objc_CFString_Ref 4
160/* The output reference is to a Objective-C message. */
161#define LLVMDisassembler_ReferenceType_Out_Objc_Message 5
162/* The output reference is to a Objective-C message ref. */
163#define LLVMDisassembler_ReferenceType_Out_Objc_Message_Ref 6
164/* The output reference is to a Objective-C selector ref. */
165#define LLVMDisassembler_ReferenceType_Out_Objc_Selector_Ref 7
166/* The output reference is to a Objective-C class ref. */
167#define LLVMDisassembler_ReferenceType_Out_Objc_Class_Ref 8
168
169/* The output reference is to a C++ symbol name. */
170#define LLVMDisassembler_ReferenceType_DeMangled_Name 9
171
172#ifdef __cplusplus
173extern "C" {
174#endif /* !defined(__cplusplus) */
175
176/**
177 * Create a disassembler for the TripleName.  Symbolic disassembly is supported
178 * by passing a block of information in the DisInfo parameter and specifying the
179 * TagType and callback functions as described above.  These can all be passed
180 * as NULL.  If successful, this returns a disassembler context.  If not, it
181 * returns NULL. This function is equivalent to calling
182 * LLVMCreateDisasmCPUFeatures() with an empty CPU name and feature set.
183 */
184LLVMDisasmContextRef LLVMCreateDisasm(const char *TripleName, void *DisInfo,
185                                      int TagType, LLVMOpInfoCallback GetOpInfo,
186                                      LLVMSymbolLookupCallback SymbolLookUp);
187
188/**
189 * Create a disassembler for the TripleName and a specific CPU.  Symbolic
190 * disassembly is supported by passing a block of information in the DisInfo
191 * parameter and specifying the TagType and callback functions as described
192 * above.  These can all be passed * as NULL.  If successful, this returns a
193 * disassembler context.  If not, it returns NULL. This function is equivalent
194 * to calling LLVMCreateDisasmCPUFeatures() with an empty feature set.
195 */
196LLVMDisasmContextRef LLVMCreateDisasmCPU(const char *Triple, const char *CPU,
197                                         void *DisInfo, int TagType,
198                                         LLVMOpInfoCallback GetOpInfo,
199                                         LLVMSymbolLookupCallback SymbolLookUp);
200
201/**
202 * Create a disassembler for the TripleName, a specific CPU and specific feature
203 * string.  Symbolic disassembly is supported by passing a block of information
204 * in the DisInfo parameter and specifying the TagType and callback functions as
205 * described above.  These can all be passed * as NULL.  If successful, this
206 * returns a disassembler context.  If not, it returns NULL.
207 */
208LLVMDisasmContextRef
209LLVMCreateDisasmCPUFeatures(const char *Triple, const char *CPU,
210                            const char *Features, void *DisInfo, int TagType,
211                            LLVMOpInfoCallback GetOpInfo,
212                            LLVMSymbolLookupCallback SymbolLookUp);
213
214/**
215 * Set the disassembler's options.  Returns 1 if it can set the Options and 0
216 * otherwise.
217 */
218int LLVMSetDisasmOptions(LLVMDisasmContextRef DC, uint64_t Options);
219
220/* The option to produce marked up assembly. */
221#define LLVMDisassembler_Option_UseMarkup 1
222/* The option to print immediates as hex. */
223#define LLVMDisassembler_Option_PrintImmHex 2
224/* The option use the other assembler printer variant */
225#define LLVMDisassembler_Option_AsmPrinterVariant 4
226/* The option to set comment on instructions */
227#define LLVMDisassembler_Option_SetInstrComments 8
228  /* The option to print latency information alongside instructions */
229#define LLVMDisassembler_Option_PrintLatency 16
230
231/**
232 * Dispose of a disassembler context.
233 */
234void LLVMDisasmDispose(LLVMDisasmContextRef DC);
235
236/**
237 * Disassemble a single instruction using the disassembler context specified in
238 * the parameter DC.  The bytes of the instruction are specified in the
239 * parameter Bytes, and contains at least BytesSize number of bytes.  The
240 * instruction is at the address specified by the PC parameter.  If a valid
241 * instruction can be disassembled, its string is returned indirectly in
242 * OutString whose size is specified in the parameter OutStringSize.  This
243 * function returns the number of bytes in the instruction or zero if there was
244 * no valid instruction.
245 */
246size_t LLVMDisasmInstruction(LLVMDisasmContextRef DC, uint8_t *Bytes,
247                             uint64_t BytesSize, uint64_t PC,
248                             char *OutString, size_t OutStringSize);
249
250/**
251 * @}
252 */
253
254#ifdef __cplusplus
255}
256#endif /* !defined(__cplusplus) */
257
258#endif /* LLVM_C_DISASSEMBLER_H */
259