/*
* decode - Reads the appropriate instruction table to obtain the unique ID of
* an instruction.
*
* @param type - See modRMRequired().
* @param insnContext - See modRMRequired().
* @param opcode - See modRMRequired().
* @param modRM - The ModR/M byte if required, or any value if not.
* @return - The UID of the instruction, or 0 on failure.
*/
static InstrUID decode(OpcodeType type,
InstructionContext insnContext,
uint8_t opcode,
uint8_t modRM) {
const struct ModRMDecision* dec = 0;
switch (type) {
case ONEBYTE:
dec = &ONEBYTE_SYM.opcodeDecisions[insnContext].modRMDecisions[opcode];
break;
case TWOBYTE:
dec = &TWOBYTE_SYM.opcodeDecisions[insnContext].modRMDecisions[opcode];
break;
case THREEBYTE_38:
dec = &THREEBYTE38_SYM.opcodeDecisions[insnContext].modRMDecisions[opcode];
break;
case THREEBYTE_3A:
dec = &THREEBYTE3A_SYM.opcodeDecisions[insnContext].modRMDecisions[opcode];
break;
case THREEBYTE_A6:
dec = &THREEBYTEA6_SYM.opcodeDecisions[insnContext].modRMDecisions[opcode];
break;
case THREEBYTE_A7:
dec = &THREEBYTEA7_SYM.opcodeDecisions[insnContext].modRMDecisions[opcode];
break;
}
switch (dec->modrm_type) {
default:
debug("Corrupt table! Unknown modrm_type");
return 0;
case MODRM_ONEENTRY:
return modRMTable[dec->instructionIDs];
case MODRM_SPLITRM:
if (modFromModRM(modRM) == 0x3)
return modRMTable[dec->instructionIDs+1];
return modRMTable[dec->instructionIDs];
case MODRM_SPLITREG:
if (modFromModRM(modRM) == 0x3)
return modRMTable[dec->instructionIDs+((modRM & 0x38) >> 3)+8];
return modRMTable[dec->instructionIDs+((modRM & 0x38) >> 3)];
case MODRM_FULL:
return modRMTable[dec->instructionIDs+modRM];
}
}
/*
* decode - Reads the appropriate instruction table to obtain the unique ID of
* an instruction.
*
* @param type - See modRMRequired().
* @param insnContext - See modRMRequired().
* @param opcode - See modRMRequired().
* @param modRM - The ModR/M byte if required, or any value if not.
*/
static InstrUID decode(OpcodeType type,
InstructionContext insnContext,
uint8_t opcode,
uint8_t modRM) {
struct ModRMDecision* dec;
switch (type) {
default:
unreachable("Unknown opcode type");
case ONEBYTE:
dec = &ONEBYTE_SYM.opcodeDecisions[insnContext].modRMDecisions[opcode];
break;
case TWOBYTE:
dec = &TWOBYTE_SYM.opcodeDecisions[insnContext].modRMDecisions[opcode];
break;
case THREEBYTE_38:
dec = &THREEBYTE38_SYM.opcodeDecisions[insnContext].modRMDecisions[opcode];
break;
case THREEBYTE_3A:
dec = &THREEBYTE3A_SYM.opcodeDecisions[insnContext].modRMDecisions[opcode];
break;
}
switch (dec->modrm_type) {
default:
unreachable("Corrupt table! Unknown modrm_type");
case MODRM_ONEENTRY:
return dec->instructionIDs[0];
case MODRM_SPLITRM:
if (modFromModRM(modRM) == 0x3)
return dec->instructionIDs[1];
else
return dec->instructionIDs[0];
case MODRM_FULL:
return dec->instructionIDs[modRM];
}
return 0;
}
/*
* readSIB - Consumes the SIB byte to determine addressing information for an
* instruction.
*
* @param insn - The instruction whose SIB byte is to be read.
* @return - 0 if the SIB byte was successfully read; nonzero otherwise.
*/
static int readSIB(struct InternalInstruction* insn) {
SIBIndex sibIndexBase = 0;
SIBBase sibBaseBase = 0;
uint8_t index, base;
dbgprintf(insn, "readSIB()");
if (insn->consumedSIB)
return 0;
insn->consumedSIB = TRUE;
switch (insn->addressSize) {
case 2:
dbgprintf(insn, "SIB-based addressing doesn't work in 16-bit mode");
return -1;
break;
case 4:
sibIndexBase = SIB_INDEX_EAX;
sibBaseBase = SIB_BASE_EAX;
break;
case 8:
sibIndexBase = SIB_INDEX_RAX;
sibBaseBase = SIB_BASE_RAX;
break;
}
if (consumeByte(insn, &insn->sib))
return -1;
index = indexFromSIB(insn->sib) | (xFromREX(insn->rexPrefix) << 3);
switch (index) {
case 0x4:
insn->sibIndex = SIB_INDEX_NONE;
break;
default:
insn->sibIndex = (SIBIndex)(sibIndexBase + index);
if (insn->sibIndex == SIB_INDEX_sib ||
insn->sibIndex == SIB_INDEX_sib64)
insn->sibIndex = SIB_INDEX_NONE;
break;
}
switch (scaleFromSIB(insn->sib)) {
case 0:
insn->sibScale = 1;
break;
case 1:
insn->sibScale = 2;
break;
case 2:
insn->sibScale = 4;
break;
case 3:
insn->sibScale = 8;
break;
}
base = baseFromSIB(insn->sib) | (bFromREX(insn->rexPrefix) << 3);
switch (base) {
case 0x5:
switch (modFromModRM(insn->modRM)) {
case 0x0:
insn->eaDisplacement = EA_DISP_32;
insn->sibBase = SIB_BASE_NONE;
break;
case 0x1:
insn->eaDisplacement = EA_DISP_8;
insn->sibBase = (insn->addressSize == 4 ?
SIB_BASE_EBP : SIB_BASE_RBP);
break;
case 0x2:
insn->eaDisplacement = EA_DISP_32;
insn->sibBase = (insn->addressSize == 4 ?
SIB_BASE_EBP : SIB_BASE_RBP);
break;
case 0x3:
debug("Cannot have Mod = 0b11 and a SIB byte");
return -1;
}
break;
default:
insn->sibBase = (SIBBase)(sibBaseBase + base);
break;
}
return 0;
}
/*
* readModRM - Consumes all addressing information (ModR/M byte, SIB byte, and
* displacement) for an instruction and interprets it.
*
* @param insn - The instruction whose addressing information is to be read.
* @return - 0 if the information was successfully read; nonzero otherwise.
*/
static int readModRM(struct InternalInstruction* insn) {
uint8_t mod, rm, reg;
dbgprintf(insn, "readModRM()");
if (insn->consumedModRM)
return 0;
if (consumeByte(insn, &insn->modRM))
return -1;
insn->consumedModRM = TRUE;
mod = modFromModRM(insn->modRM);
rm = rmFromModRM(insn->modRM);
reg = regFromModRM(insn->modRM);
/*
* This goes by insn->registerSize to pick the correct register, which messes
* up if we're using (say) XMM or 8-bit register operands. That gets fixed in
* fixupReg().
*/
switch (insn->registerSize) {
case 2:
insn->regBase = MODRM_REG_AX;
insn->eaRegBase = EA_REG_AX;
break;
case 4:
insn->regBase = MODRM_REG_EAX;
insn->eaRegBase = EA_REG_EAX;
break;
case 8:
insn->regBase = MODRM_REG_RAX;
insn->eaRegBase = EA_REG_RAX;
break;
}
reg |= rFromREX(insn->rexPrefix) << 3;
rm |= bFromREX(insn->rexPrefix) << 3;
insn->reg = (Reg)(insn->regBase + reg);
switch (insn->addressSize) {
case 2:
insn->eaBaseBase = EA_BASE_BX_SI;
switch (mod) {
case 0x0:
if (rm == 0x6) {
insn->eaBase = EA_BASE_NONE;
insn->eaDisplacement = EA_DISP_16;
if (readDisplacement(insn))
return -1;
} else {
insn->eaBase = (EABase)(insn->eaBaseBase + rm);
insn->eaDisplacement = EA_DISP_NONE;
}
break;
case 0x1:
insn->eaBase = (EABase)(insn->eaBaseBase + rm);
insn->eaDisplacement = EA_DISP_8;
if (readDisplacement(insn))
return -1;
break;
case 0x2:
insn->eaBase = (EABase)(insn->eaBaseBase + rm);
insn->eaDisplacement = EA_DISP_16;
if (readDisplacement(insn))
return -1;
break;
case 0x3:
insn->eaBase = (EABase)(insn->eaRegBase + rm);
if (readDisplacement(insn))
return -1;
break;
}
break;
case 4:
case 8:
insn->eaBaseBase = (insn->addressSize == 4 ? EA_BASE_EAX : EA_BASE_RAX);
switch (mod) {
case 0x0:
insn->eaDisplacement = EA_DISP_NONE; /* readSIB may override this */
switch (rm) {
case 0x4:
case 0xc: /* in case REXW.b is set */
insn->eaBase = (insn->addressSize == 4 ?
EA_BASE_sib : EA_BASE_sib64);
readSIB(insn);
if (readDisplacement(insn))
return -1;
break;
case 0x5:
insn->eaBase = EA_BASE_NONE;
insn->eaDisplacement = EA_DISP_32;
if (readDisplacement(insn))
return -1;
break;
default:
insn->eaBase = (EABase)(insn->eaBaseBase + rm);
break;
}
break;
case 0x1:
case 0x2:
insn->eaDisplacement = (mod == 0x1 ? EA_DISP_8 : EA_DISP_32);
switch (rm) {
case 0x4:
case 0xc: /* in case REXW.b is set */
insn->eaBase = EA_BASE_sib;
readSIB(insn);
if (readDisplacement(insn))
return -1;
break;
default:
insn->eaBase = (EABase)(insn->eaBaseBase + rm);
if (readDisplacement(insn))
return -1;
break;
}
break;
case 0x3:
insn->eaDisplacement = EA_DISP_NONE;
insn->eaBase = (EABase)(insn->eaRegBase + rm);
break;
}
break;
} /* switch (insn->addressSize) */
return 0;
}