cctools/include/llvm-c/Disassembler.h

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