dis.h revision 13f7948dbc7ae32c3e30086604ad6ade2ead3671
/** @file
* DIS - The VirtualBox Disassembler.
*/
/*
* Copyright (C) 2006-2012 Oracle Corporation
*
* This file is part of VirtualBox Open Source Edition (OSE), as
* available from http://www.virtualbox.org. This file is free software;
* General Public License (GPL) as published by the Free Software
* Foundation, in version 2 as it comes in the "COPYING" file of the
* VirtualBox OSE distribution. VirtualBox OSE is distributed in the
* hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
*
* The contents of this file may alternatively be used under the terms
* of the Common Development and Distribution License Version 1.0
* (CDDL) only, as it comes in the "COPYING.CDDL" file of the
* VirtualBox OSE distribution, in which case the provisions of the
* CDDL are applicable instead of those of the GPL.
*
* You may elect to license modified versions of this file under the
* terms and conditions of either the GPL or the CDDL or both.
*/
#ifndef ___VBox_disasm_h
#define ___VBox_disasm_h
#include <VBox/disopcode.h>
#if defined(__L4ENV__)
# include <setjmp.h>
#endif
/** CPU mode flags (DISCPUSTATE::mode).
* @{
*/
typedef enum
{
CPUMODE_16BIT = 1,
CPUMODE_32BIT = 2,
CPUMODE_64BIT = 3,
/** hack forcing the size of the enum to 32-bits. */
CPUMODE_MAKE_32BIT_HACK = 0x7fffffff
} DISCPUMODE;
/** @} */
/** Prefix byte flags
* @{
*/
#define PREFIX_NONE 0
/** non-default address size. */
#define PREFIX_ADDRSIZE RT_BIT(0)
/** non-default operand size. */
/** lock prefix. */
/** segment prefix. */
/** rep(e) prefix (not a prefix, but we'll treat is as one). */
/** rep(e) prefix (not a prefix, but we'll treat is as one). */
/** REX prefix (64 bits) */
/** @} */
/** 64 bits prefix byte flags
* @{
*/
#define PREFIX_REX_OP_2_FLAGS(a) (a - OP_PARM_REX_START)
/** @} */
/**
* Operand type.
*/
#define OPTYPE_INVALID RT_BIT(0)
#define OPTYPE_RRM_DANGEROUS RT_BIT(14) /**< Some additional dangerous ones when recompiling raw r0. */
#define OPTYPE_RRM_DANGEROUS_16 RT_BIT(15) /**< Some additional dangerous ones when recompiling 16-bit raw r0. */
#define OPTYPE_FORCED_64_OP_SIZE RT_BIT(22) /**< Forced 64 bits operand size; regardless of prefix bytes */
#define OPTYPE_REXB_EXTENDS_OPREG RT_BIT(23) /**< REX.B extends the register field in the opcode byte */
#define OPTYPE_FORCED_32_OP_SIZE_X86 RT_BIT(25) /**< Forced 32 bits operand size; regardless of prefix bytes (only in 16 & 32 bits mode!) */
#define OPTYPE_ALL (0xffffffff)
/** Parameter usage flags.
* @{
*/
/** DS:ESI */
/** ES:EDI */
#define USE_IMMEDIATE (USE_IMMEDIATE8|USE_IMMEDIATE16|USE_IMMEDIATE32|USE_IMMEDIATE64|USE_IMMEDIATE8_REL|USE_IMMEDIATE16_REL|USE_IMMEDIATE32_REL|USE_IMMEDIATE64_REL|USE_IMMEDIATE_ADDR_0_32|USE_IMMEDIATE_ADDR_16_32|USE_IMMEDIATE_ADDR_0_16|USE_IMMEDIATE_ADDR_16_16|USE_IMMEDIATE16_SX8|USE_IMMEDIATE32_SX8|USE_IMMEDIATE64_SX8)
#define DIS_IS_EFFECTIVE_ADDR(flags) !!((flags) & (USE_BASE|USE_INDEX|USE_DISPLACEMENT32|USE_DISPLACEMENT64|USE_DISPLACEMENT16|USE_DISPLACEMENT8|USE_RIPDISPLACEMENT32))
/** @} */
/** index in {"RAX", "RCX", "RDX", "RBX", "RSP", "RBP", "RSI", "RDI", "R8", "R9", "R10", "R11", "R12", "R13", "R14", "R15"}
* @{
*/
#define USE_REG_RAX 0
#define USE_REG_RCX 1
#define USE_REG_RDX 2
#define USE_REG_RBX 3
#define USE_REG_RSP 4
#define USE_REG_RBP 5
#define USE_REG_RSI 6
#define USE_REG_RDI 7
#define USE_REG_R8 8
#define USE_REG_R9 9
#define USE_REG_R10 10
#define USE_REG_R11 11
#define USE_REG_R12 12
#define USE_REG_R13 13
#define USE_REG_R14 14
#define USE_REG_R15 15
/** @} */
/** index in {"EAX", "ECX", "EDX", "EBX", "ESP", "EBP", "ESI", "EDI", "R8D", "R9D", "R10D", "R11D", "R12D", "R13D", "R14D", "R15D"}
* @{
*/
#define USE_REG_EAX 0
#define USE_REG_ECX 1
#define USE_REG_EDX 2
#define USE_REG_EBX 3
#define USE_REG_ESP 4
#define USE_REG_EBP 5
#define USE_REG_ESI 6
#define USE_REG_EDI 7
#define USE_REG_R8D 8
#define USE_REG_R9D 9
#define USE_REG_R10D 10
#define USE_REG_R11D 11
#define USE_REG_R12D 12
#define USE_REG_R13D 13
#define USE_REG_R14D 14
#define USE_REG_R15D 15
/** @} */
/** index in {"AX", "CX", "DX", "BX", "SP", "BP", "SI", "DI", "R8W", "R9W", "R10W", "R11W", "R12W", "R13W", "R14W", "R15W"}
* @{
*/
#define USE_REG_AX 0
#define USE_REG_CX 1
#define USE_REG_DX 2
#define USE_REG_BX 3
#define USE_REG_SP 4
#define USE_REG_BP 5
#define USE_REG_SI 6
#define USE_REG_DI 7
#define USE_REG_R8W 8
#define USE_REG_R9W 9
#define USE_REG_R10W 10
#define USE_REG_R11W 11
#define USE_REG_R12W 12
#define USE_REG_R13W 13
#define USE_REG_R14W 14
#define USE_REG_R15W 15
/** @} */
/** index in {"AL", "CL", "DL", "BL", "AH", "CH", "DH", "BH", "R8B", "R9B", "R10B", "R11B", "R12B", "R13B", "R14B", "R15B", "SPL", "BPL", "SIL", "DIL"}
* @{
*/
#define USE_REG_AL 0
#define USE_REG_CL 1
#define USE_REG_DL 2
#define USE_REG_BL 3
#define USE_REG_AH 4
#define USE_REG_CH 5
#define USE_REG_DH 6
#define USE_REG_BH 7
#define USE_REG_R8B 8
#define USE_REG_R9B 9
#define USE_REG_R10B 10
#define USE_REG_R11B 11
#define USE_REG_R12B 12
#define USE_REG_R13B 13
#define USE_REG_R14B 14
#define USE_REG_R15B 15
#define USE_REG_SPL 16
#define USE_REG_BPL 17
#define USE_REG_SIL 18
#define USE_REG_DIL 19
/** @} */
/** index in {ES, CS, SS, DS, FS, GS}
* @{
*/
typedef enum
{
DIS_SELREG_ES = 0,
DIS_SELREG_CS = 1,
DIS_SELREG_SS = 2,
DIS_SELREG_DS = 3,
DIS_SELREG_FS = 4,
DIS_SELREG_GS = 5,
/** The usual 32-bit paranoia. */
DIS_SEGREG_32BIT_HACK = 0x7fffffff
} DIS_SELREG;
/** @} */
#define USE_REG_FP0 0
#define USE_REG_FP1 1
#define USE_REG_FP2 2
#define USE_REG_FP3 3
#define USE_REG_FP4 4
#define USE_REG_FP5 5
#define USE_REG_FP6 6
#define USE_REG_FP7 7
#define USE_REG_CR0 0
#define USE_REG_CR1 1
#define USE_REG_CR2 2
#define USE_REG_CR3 3
#define USE_REG_CR4 4
#define USE_REG_CR8 8
#define USE_REG_DR0 0
#define USE_REG_DR1 1
#define USE_REG_DR2 2
#define USE_REG_DR3 3
#define USE_REG_DR4 4
#define USE_REG_DR5 5
#define USE_REG_DR6 6
#define USE_REG_DR7 7
#define USE_REG_MMX0 0
#define USE_REG_MMX1 1
#define USE_REG_MMX2 2
#define USE_REG_MMX3 3
#define USE_REG_MMX4 4
#define USE_REG_MMX5 5
#define USE_REG_MMX6 6
#define USE_REG_MMX7 7
#define USE_REG_XMM0 0
#define USE_REG_XMM1 1
#define USE_REG_XMM2 2
#define USE_REG_XMM3 3
#define USE_REG_XMM4 4
#define USE_REG_XMM5 5
#define USE_REG_XMM6 6
#define USE_REG_XMM7 7
/** Used by DISQueryParamVal & EMIQueryParamVal
* @{
*/
#define PARAM_VAL8 RT_BIT(0)
#define PARMTYPE_REGISTER 1
#define PARMTYPE_ADDRESS 2
#define PARMTYPE_IMMEDIATE 3
typedef struct
{
union
{
struct
{
} farptr;
} val;
} OP_PARAMVAL;
/** Pointer to opcode parameter value. */
typedef OP_PARAMVAL *POP_PARAMVAL;
typedef enum
{
} PARAM_TYPE;
/** @} */
/**
* Operand Parameter.
*/
typedef struct OP_PARAMETER
{
union
{
} uDisp;
union
{
/** ST(0) - ST(7) */
/** MMX0 - MMX7 */
/** XMM0 - XMM7 */
/** {ES, CS, SS, DS, FS, GS} */
/** TR0-TR7 (?) */
/** CR0-CR4 */
/** DR0-DR7 */
} base;
union
{
} index;
/** 2, 4 or 8. */
/** Parameter size. */
#ifndef DIS_SEPARATE_FORMATTER
char szParam[32];
#endif
} OP_PARAMETER;
/** Pointer to opcode parameter. */
typedef OP_PARAMETER *POP_PARAMETER;
/** Pointer to opcode parameter. */
typedef const OP_PARAMETER *PCOP_PARAMETER;
/** Pointer to opcode. */
/** Pointer to const opcode. */
/**
* Callback for reading opcode bytes.
*
* @param pDisState Pointer to the CPU state. The primary user argument
* can be retrived from DISCPUSTATE::apvUserData[0]. If
* more is required these can be passed in the
* subsequent slots.
* @param pbDst Pointer to output buffer.
* @param uSrcAddr The address to start reading at.
* @param cbToRead The number of bytes to read.
*/
typedef DECLCALLBACK(int) FNDISREADBYTES(PDISCPUSTATE pDisState, uint8_t *pbDst, RTUINTPTR uSrcAddr, uint32_t cbToRead);
/** Pointer to a opcode byte reader. */
typedef FNDISREADBYTES *PFNDISREADBYTES;
/** Parser callback.
* @remark no DECLCALLBACK() here because it's considered to be internal (really, I'm too lazy to update all the functions). */
typedef unsigned FNDISPARSE(RTUINTPTR pu8CodeBlock, PCOPCODE pOp, POP_PARAMETER pParam, PDISCPUSTATE pCpu);
typedef FNDISPARSE *PFNDISPARSE;
typedef struct DISCPUSTATE
{
/** Global setting. */
/** Per instruction prefix settings. */
/** segment prefix value. */
/** rex prefix value (64 bits only */
/** addressing mode (16 or 32 bits). (CPUMODE_*) */
/** operand mode (16 or 32 bits). (CPUMODE_*) */
/** ModRM fields. */
union
{
/** Bitfield view */
struct
{
unsigned Rm : 4;
unsigned Reg : 4;
unsigned Mod : 2;
} Bits;
/** unsigned view */
unsigned u;
} ModRM;
/** SIB fields. */
union
{
/** Bitfield view */
struct
{
unsigned Base : 4;
unsigned Index : 4;
unsigned Scale : 2;
} Bits;
/** unsigned view */
unsigned u;
} SIB;
/** The instruction size. */
/** The address of the instruction. */
/** The offsetted address of the instruction. */
/** The size of the prefix bytes. */
/** First opcode byte of instruction. */
/** Last prefix byte (for SSE2 extension tables). */
/** The instruction bytes. */
/** Internal: pointer to disassembly function table */
/** Internal: instruction filter */
/** Return code set by a worker function like the opcode bytes readers. */
/** Pointer to the current instruction. */
#ifndef DIS_CORE_ONLY
/** Opcode format string for current instruction. */
const char *pszOpcode;
#endif
/** Optional read function */
/** User data slots for the read callback. The first entry is used for the
* pvUser argument, the rest are up for grabs.
* @remarks This must come last so that we can memset everything before this. */
void *apvUserData[3];
} DISCPUSTATE;
/** The storage padding sufficient to hold the largest DISCPUSTATE in all
* contexts (R3, R0 and RC). Used various places in the VMM internals. */
/** Opcode. */
#pragma pack(4)
typedef struct OPCODE
{
#ifndef DIS_CORE_ONLY
const char *pszOpcode;
#endif
unsigned optype;
} OPCODE;
#pragma pack()
DISDECL(int) DISInstr(RTUINTPTR uInstrAddr, DISCPUMODE enmCpuMode, PDISCPUSTATE pCpu, uint32_t *pcbSize, char *pszOutput);
DISDECL(int) DISInstrWithOff(PDISCPUSTATE pCpu, RTUINTPTR uInstrAddr, RTUINTPTR offRealAddr, uint32_t *pcbSize, char *pszOutput);
DISDECL(int) DISInstrWithReader(RTUINTPTR uInstrAddr, DISCPUMODE enmCpuMode, PFNDISREADBYTES pfnReadBytes, void *pvUser,
DISDECL(int) DISCoreOneEx(RTUINTPTR InstructionAddr, DISCPUMODE enmCpuMode, PFNDISREADBYTES pfnReadBytes, void *pvUser,
DISDECL(int) DISQueryParamVal(PCPUMCTXCORE pCtx, PDISCPUSTATE pCpu, POP_PARAMETER pParam, POP_PARAMVAL pParamVal, PARAM_TYPE parmtype);
DISDECL(int) DISQueryParamRegPtr(PCPUMCTXCORE pCtx, PDISCPUSTATE pCpu, POP_PARAMETER pParam, void **ppReg, size_t *pcbSize);
DISDECL(int) DISFetchRegSegEx(PCCPUMCTXCORE pCtx, DIS_SELREG sel, RTSEL *pVal, PCPUMSELREGHID *ppSelHidReg);
/**
* Try resolve an address into a symbol name.
*
* For use with DISFormatYasmEx(), DISFormatMasmEx() and DISFormatGasEx().
*
* @returns VBox status code.
* @retval VINF_SUCCESS on success, pszBuf contains the full symbol name.
* @retval VINF_BUFFER_OVERFLOW if pszBuf is too small the symbol name. The
* content of pszBuf is truncated and zero terminated.
* @retval VERR_SYMBOL_NOT_FOUND if no matching symbol was found for the address.
*
* @param pCpu Pointer to the disassembler CPU state.
* @param u32Sel The selector value. Use DIS_FMT_SEL_IS_REG, DIS_FMT_SEL_GET_VALUE,
* DIS_FMT_SEL_GET_REG to access this.
* @param uAddress The segment address.
* @param pszBuf Where to store the symbol name
* @param cchBuf The size of the buffer.
* @param poff If not a perfect match, then this is where the offset from the return
* symbol to the specified address is returned.
* @param pvUser The user argument.
*/
typedef DECLCALLBACK(int) FNDISGETSYMBOL(PCDISCPUSTATE pCpu, uint32_t u32Sel, RTUINTPTR uAddress, char *pszBuf, size_t cchBuf, RTINTPTR *poff, void *pvUser);
/** Pointer to a FNDISGETSYMBOL(). */
typedef FNDISGETSYMBOL *PFNDISGETSYMBOL;
/**
* Checks if the FNDISGETSYMBOL argument u32Sel is a register or not.
*/
/**
* Extracts the selector value from the FNDISGETSYMBOL argument u32Sel.
* @returns Selector value.
*/
/**
* Extracts the register number from the FNDISGETSYMBOL argument u32Sel.
* @returns USE_REG_CS, USE_REG_SS, USE_REG_DS, USE_REG_ES, USE_REG_FS or USE_REG_FS.
*/
/** @internal */
/** @internal */
/** @name Flags for use with DISFormatYasmEx(), DISFormatMasmEx() and DISFormatGasEx().
* @{
*/
/** Put the address to the right. */
#define DIS_FMT_FLAGS_ADDR_RIGHT RT_BIT_32(0)
/** Put the address to the left. */
/** Put the address in comments.
* For some assemblers this implies placing it to the right. */
/** Put the instruction bytes to the right of the disassembly. */
/** Put the instruction bytes to the left of the disassembly. */
/** Put the instruction bytes in comments.
* For some assemblers this implies placing the bytes to the right. */
/** Put the bytes in square brackets. */
/** Put spaces between the bytes. */
/** Display the relative +/- offset of branch instructions that uses relative addresses,
* and put the target address in parenthesis. */
/** Strict assembly. The assembly should, when ever possible, make the
* assembler reproduce the exact same binary. (Refers to the yasm
* strict keyword.) */
/** Checks if the given flags are a valid combination. */
#define DIS_FMT_FLAGS_IS_VALID(fFlags) \
&& ((fFlags) & (DIS_FMT_FLAGS_ADDR_RIGHT | DIS_FMT_FLAGS_ADDR_LEFT)) != (DIS_FMT_FLAGS_ADDR_RIGHT | DIS_FMT_FLAGS_ADDR_LEFT) \
&& ( !((fFlags) & DIS_FMT_FLAGS_ADDR_COMMENT) \
&& ((fFlags) & (DIS_FMT_FLAGS_BYTES_RIGHT | DIS_FMT_FLAGS_BYTES_LEFT)) != (DIS_FMT_FLAGS_BYTES_RIGHT | DIS_FMT_FLAGS_BYTES_LEFT) \
)
/** @} */
DISDECL(size_t) DISFormatYasmEx(PCDISCPUSTATE pCpu, char *pszBuf, size_t cchBuf, uint32_t fFlags, PFNDISGETSYMBOL pfnGetSymbol, void *pvUser);
DISDECL(size_t) DISFormatMasmEx(PCDISCPUSTATE pCpu, char *pszBuf, size_t cchBuf, uint32_t fFlags, PFNDISGETSYMBOL pfnGetSymbol, void *pvUser);
DISDECL(size_t) DISFormatGasEx( PCDISCPUSTATE pCpu, char *pszBuf, size_t cchBuf, uint32_t fFlags, PFNDISGETSYMBOL pfnGetSymbol, void *pvUser);
/** @todo DISAnnotate(PCDISCPUSTATE pCpu, char *pszBuf, size_t cchBuf, register reader, memory reader); */
#endif