diff options
Diffstat (limited to 'lib/Target/X86/Disassembler/X86DisassemblerDecoder.h')
| -rw-r--r-- | lib/Target/X86/Disassembler/X86DisassemblerDecoder.h | 283 |
1 files changed, 124 insertions, 159 deletions
diff --git a/lib/Target/X86/Disassembler/X86DisassemblerDecoder.h b/lib/Target/X86/Disassembler/X86DisassemblerDecoder.h index 4acf17a221..740ecce021 100644 --- a/lib/Target/X86/Disassembler/X86DisassemblerDecoder.h +++ b/lib/Target/X86/Disassembler/X86DisassemblerDecoder.h @@ -21,9 +21,7 @@ namespace llvm { namespace X86Disassembler { -/* - * Accessor functions for various fields of an Intel instruction - */ +// Accessor functions for various fields of an Intel instruction #define modFromModRM(modRM) (((modRM) & 0xc0) >> 6) #define regFromModRM(modRM) (((modRM) & 0x38) >> 3) #define rmFromModRM(modRM) ((modRM) & 0x7) @@ -73,10 +71,7 @@ namespace X86Disassembler { #define lFromXOP3of3(xop) (((xop) & 0x4) >> 2) #define ppFromXOP3of3(xop) ((xop) & 0x3) -/* - * These enums represent Intel registers for use by the decoder. - */ - +// These enums represent Intel registers for use by the decoder. #define REGS_8BIT \ ENTRY(AL) \ ENTRY(CL) \ @@ -382,13 +377,11 @@ namespace X86Disassembler { REGS_CONTROL \ ENTRY(RIP) -/* - * EABase - All possible values of the base field for effective-address - * computations, a.k.a. the Mod and R/M fields of the ModR/M byte. We - * distinguish between bases (EA_BASE_*) and registers that just happen to be - * referred to when Mod == 0b11 (EA_REG_*). - */ -typedef enum { +/// \brief All possible values of the base field for effective-address +/// computations, a.k.a. the Mod and R/M fields of the ModR/M byte. +/// We distinguish between bases (EA_BASE_*) and registers that just happen +/// to be referred to when Mod == 0b11 (EA_REG_*). +enum EABase { EA_BASE_NONE, #define ENTRY(x) EA_BASE_##x, ALL_EA_BASES @@ -397,15 +390,13 @@ typedef enum { ALL_REGS #undef ENTRY EA_max -} EABase; - -/* - * SIBIndex - All possible values of the SIB index field. - * Borrows entries from ALL_EA_BASES with the special case that - * sib is synonymous with NONE. - * Vector SIB: index can be XMM or YMM. - */ -typedef enum { +}; + +/// \brief All possible values of the SIB index field. +/// borrows entries from ALL_EA_BASES with the special case that +/// sib is synonymous with NONE. +/// Vector SIB: index can be XMM or YMM. +enum SIBIndex { SIB_INDEX_NONE, #define ENTRY(x) SIB_INDEX_##x, ALL_EA_BASES @@ -414,23 +405,18 @@ typedef enum { REGS_ZMM #undef ENTRY SIB_INDEX_max -} SIBIndex; +}; -/* - * SIBBase - All possible values of the SIB base field. - */ -typedef enum { +/// \brief All possible values of the SIB base field. +enum SIBBase { SIB_BASE_NONE, #define ENTRY(x) SIB_BASE_##x, ALL_SIB_BASES #undef ENTRY SIB_BASE_max -} SIBBase; +}; -/* - * EADisplacement - Possible displacement types for effective-address - * computations. - */ +/// \brief Possible displacement types for effective-address computations. typedef enum { EA_DISP_NONE, EA_DISP_8, @@ -438,20 +424,16 @@ typedef enum { EA_DISP_32 } EADisplacement; -/* - * Reg - All possible values of the reg field in the ModR/M byte. - */ -typedef enum { +/// \brief All possible values of the reg field in the ModR/M byte. +enum Reg { #define ENTRY(x) MODRM_REG_##x, ALL_REGS #undef ENTRY MODRM_REG_max -} Reg; +}; -/* - * SegmentOverride - All possible segment overrides. - */ -typedef enum { +/// \brief All possible segment overrides. +enum SegmentOverride { SEG_OVERRIDE_NONE, SEG_OVERRIDE_CS, SEG_OVERRIDE_SS, @@ -460,196 +442,182 @@ typedef enum { SEG_OVERRIDE_FS, SEG_OVERRIDE_GS, SEG_OVERRIDE_max -} SegmentOverride; - -/* - * VEXLeadingOpcodeByte - Possible values for the VEX.m-mmmm field - */ +}; -typedef enum { +/// \brief Possible values for the VEX.m-mmmm field +enum VEXLeadingOpcodeByte { VEX_LOB_0F = 0x1, VEX_LOB_0F38 = 0x2, VEX_LOB_0F3A = 0x3 -} VEXLeadingOpcodeByte; +}; -typedef enum { +enum XOPMapSelect { XOP_MAP_SELECT_8 = 0x8, XOP_MAP_SELECT_9 = 0x9, XOP_MAP_SELECT_A = 0xA -} XOPMapSelect; - -/* - * VEXPrefixCode - Possible values for the VEX.pp/EVEX.pp field - */ +}; -typedef enum { +/// \brief Possible values for the VEX.pp/EVEX.pp field +enum VEXPrefixCode { VEX_PREFIX_NONE = 0x0, VEX_PREFIX_66 = 0x1, VEX_PREFIX_F3 = 0x2, VEX_PREFIX_F2 = 0x3 -} VEXPrefixCode; +}; -typedef enum { +enum VectorExtensionType { TYPE_NO_VEX_XOP = 0x0, TYPE_VEX_2B = 0x1, TYPE_VEX_3B = 0x2, TYPE_EVEX = 0x3, TYPE_XOP = 0x4 -} VectorExtensionType; +}; typedef uint8_t BOOL; -/* - * byteReader_t - Type for the byte reader that the consumer must provide to - * the decoder. Reads a single byte from the instruction's address space. - * @param arg - A baton that the consumer can associate with any internal - * state that it needs. - * @param byte - A pointer to a single byte in memory that should be set to - * contain the value at address. - * @param address - The address in the instruction's address space that should - * be read from. - * @return - -1 if the byte cannot be read for any reason; 0 otherwise. - */ -typedef int (*byteReader_t)(const void* arg, uint8_t* byte, uint64_t address); - -/* - * dlog_t - Type for the logging function that the consumer can provide to - * get debugging output from the decoder. - * @param arg - A baton that the consumer can associate with any internal - * state that it needs. - * @param log - A string that contains the message. Will be reused after - * the logger returns. - */ -typedef void (*dlog_t)(void* arg, const char *log); - -/* - * The specification for how to extract and interpret a full instruction and - * its operands. - */ +/// \brief Type for the byte reader that the consumer must provide to +/// the decoder. Reads a single byte from the instruction's address space. +/// \param arg A baton that the consumer can associate with any internal +/// state that it needs. +/// \param byte A pointer to a single byte in memory that should be set to +/// contain the value at address. +/// \param address The address in the instruction's address space that should +/// be read from. +/// \return -1 if the byte cannot be read for any reason; 0 otherwise. +typedef int (*byteReader_t)(const void *arg, uint8_t *byte, uint64_t address); + +/// \brief Type for the logging function that the consumer can provide to +/// get debugging output from the decoder. +/// \param arg A baton that the consumer can associate with any internal +/// state that it needs. +/// \param log A string that contains the message. Will be reused after +/// the logger returns. +typedef void (*dlog_t)(void *arg, const char *log); + +/// The specification for how to extract and interpret a full instruction and +/// its operands. struct InstructionSpecifier { uint16_t operands; }; -/* - * The x86 internal instruction, which is produced by the decoder. - */ +/// The x86 internal instruction, which is produced by the decoder. struct InternalInstruction { - /* Reader interface (C) */ + // Reader interface (C) byteReader_t reader; - /* Opaque value passed to the reader */ + // Opaque value passed to the reader const void* readerArg; - /* The address of the next byte to read via the reader */ + // The address of the next byte to read via the reader uint64_t readerCursor; - /* Logger interface (C) */ + // Logger interface (C) dlog_t dlog; - /* Opaque value passed to the logger */ + // Opaque value passed to the logger void* dlogArg; - /* General instruction information */ + // General instruction information - /* The mode to disassemble for (64-bit, protected, real) */ + // The mode to disassemble for (64-bit, protected, real) DisassemblerMode mode; - /* The start of the instruction, usable with the reader */ + // The start of the instruction, usable with the reader uint64_t startLocation; - /* The length of the instruction, in bytes */ + // The length of the instruction, in bytes size_t length; - /* Prefix state */ + // Prefix state - /* 1 if the prefix byte corresponding to the entry is present; 0 if not */ + // 1 if the prefix byte corresponding to the entry is present; 0 if not uint8_t prefixPresent[0x100]; - /* contains the location (for use with the reader) of the prefix byte */ + // contains the location (for use with the reader) of the prefix byte uint64_t prefixLocations[0x100]; - /* The value of the vector extension prefix(EVEX/VEX/XOP), if present */ + // The value of the vector extension prefix(EVEX/VEX/XOP), if present uint8_t vectorExtensionPrefix[4]; - /* The type of the vector extension prefix */ + // The type of the vector extension prefix VectorExtensionType vectorExtensionType; - /* The value of the REX prefix, if present */ + // The value of the REX prefix, if present uint8_t rexPrefix; - /* The location where a mandatory prefix would have to be (i.e., right before - the opcode, or right before the REX prefix if one is present) */ + // The location where a mandatory prefix would have to be (i.e., right before + // the opcode, or right before the REX prefix if one is present). uint64_t necessaryPrefixLocation; - /* The segment override type */ + // The segment override type SegmentOverride segmentOverride; - /* 1 if the prefix byte, 0xf2 or 0xf3 is xacquire or xrelease */ + // 1 if the prefix byte, 0xf2 or 0xf3 is xacquire or xrelease BOOL xAcquireRelease; - /* Sizes of various critical pieces of data, in bytes */ + // Sizes of various critical pieces of data, in bytes uint8_t registerSize; uint8_t addressSize; uint8_t displacementSize; uint8_t immediateSize; - /* Offsets from the start of the instruction to the pieces of data, which is - needed to find relocation entries for adding symbolic operands */ + // Offsets from the start of the instruction to the pieces of data, which is + // needed to find relocation entries for adding symbolic operands. uint8_t displacementOffset; uint8_t immediateOffset; - /* opcode state */ + // opcode state - /* The last byte of the opcode, not counting any ModR/M extension */ + // The last byte of the opcode, not counting any ModR/M extension uint8_t opcode; - /* The ModR/M byte of the instruction, if it is an opcode extension */ + // The ModR/M byte of the instruction, if it is an opcode extension uint8_t modRMExtension; - /* decode state */ + // decode state - /* The type of opcode, used for indexing into the array of decode tables */ + // The type of opcode, used for indexing into the array of decode tables OpcodeType opcodeType; - /* The instruction ID, extracted from the decode table */ + // The instruction ID, extracted from the decode table uint16_t instructionID; - /* The specifier for the instruction, from the instruction info table */ + // The specifier for the instruction, from the instruction info table const InstructionSpecifier *spec; - /* state for additional bytes, consumed during operand decode. Pattern: - consumed___ indicates that the byte was already consumed and does not - need to be consumed again */ + // state for additional bytes, consumed during operand decode. Pattern: + // consumed___ indicates that the byte was already consumed and does not + // need to be consumed again. - /* The VEX.vvvv field, which contains a third register operand for some AVX - instructions */ + // The VEX.vvvv field, which contains a third register operand for some AVX + // instructions. Reg vvvv; - /* The writemask for AVX-512 instructions which is contained in EVEX.aaa */ + // The writemask for AVX-512 instructions which is contained in EVEX.aaa Reg writemask; - /* The ModR/M byte, which contains most register operands and some portion of - all memory operands */ + // The ModR/M byte, which contains most register operands and some portion of + // all memory operands. BOOL consumedModRM; uint8_t modRM; - /* The SIB byte, used for more complex 32- or 64-bit memory operands */ + // The SIB byte, used for more complex 32- or 64-bit memory operands BOOL consumedSIB; uint8_t sib; - /* The displacement, used for memory operands */ + // The displacement, used for memory operands BOOL consumedDisplacement; int32_t displacement; - /* Immediates. There can be two in some cases */ + // Immediates. There can be two in some cases uint8_t numImmediatesConsumed; uint8_t numImmediatesTranslated; uint64_t immediates[2]; - /* A register or immediate operand encoded into the opcode */ + // A register or immediate operand encoded into the opcode Reg opcodeRegister; - /* Portions of the ModR/M byte */ + // Portions of the ModR/M byte - /* These fields determine the allowable values for the ModR/M fields, which - depend on operand and address widths */ + // These fields determine the allowable values for the ModR/M fields, which + // depend on operand and address widths. EABase eaBaseBase; EABase eaRegBase; Reg regBase; - /* The Mod and R/M fields can encode a base for an effective address, or a - register. These are separated into two fields here */ + // The Mod and R/M fields can encode a base for an effective address, or a + // register. These are separated into two fields here. EABase eaBase; EADisplacement eaDisplacement; - /* The reg field always encodes a register */ + // The reg field always encodes a register Reg reg; - /* SIB state */ + // SIB state SIBIndex sibIndex; uint8_t sibScale; SIBBase sibBase; @@ -657,22 +625,21 @@ struct InternalInstruction { const OperandSpecifier *operands; }; -/* decodeInstruction - Decode one instruction and store the decoding results in - * a buffer provided by the consumer. - * @param insn - The buffer to store the instruction in. Allocated by the - * consumer. - * @param reader - The byteReader_t for the bytes to be read. - * @param readerArg - An argument to pass to the reader for storing context - * specific to the consumer. May be NULL. - * @param logger - The dlog_t to be used in printing status messages from the - * disassembler. May be NULL. - * @param loggerArg - An argument to pass to the logger for storing context - * specific to the logger. May be NULL. - * @param startLoc - The address (in the reader's address space) of the first - * byte in the instruction. - * @param mode - The mode (16-bit, 32-bit, 64-bit) to decode in. - * @return - Nonzero if there was an error during decode, 0 otherwise. - */ +/// \brief Decode one instruction and store the decoding results in +/// a buffer provided by the consumer. +/// \param insn The buffer to store the instruction in. Allocated by the +/// consumer. +/// \param reader The byteReader_t for the bytes to be read. +/// \param readerArg An argument to pass to the reader for storing context +/// specific to the consumer. May be NULL. +/// \param logger The dlog_t to be used in printing status messages from the +/// disassembler. May be NULL. +/// \param loggerArg An argument to pass to the logger for storing context +/// specific to the logger. May be NULL. +/// \param startLoc The address (in the reader's address space) of the first +/// byte in the instruction. +/// \param mode The mode (16-bit, 32-bit, 64-bit) to decode in. +/// \return Nonzero if there was an error during decode, 0 otherwise. int decodeInstruction(InternalInstruction *insn, byteReader_t reader, const void *readerArg, @@ -682,12 +649,10 @@ int decodeInstruction(InternalInstruction *insn, uint64_t startLoc, DisassemblerMode mode); -/* \brief Debug - Print a message to debugs() - * @param file - The name of the file printing the debug message. - * @param line - The line number that printed the debug message. - * @param s - The message to print. - */ - +/// \brief Print a message to debugs() +/// \param file The name of the file printing the debug message. +/// \param line The line number that printed the debug message. +/// \param s The message to print. void Debug(const char *file, unsigned line, const char *s); const char *GetInstrName(unsigned Opcode, const void *mii); |