-/*===- X86DisassemblerDecoder.c - Disassembler decoder -------------*- C -*-==*
+/*===-- X86DisassemblerDecoder.c - Disassembler decoder ------------*- C -*-===*
*
* The LLVM Compiler Infrastructure
*
#define TRUE 1
#define FALSE 0
-typedef int8_t bool;
-
#ifndef NDEBUG
#define debug(s) do { x86DisassemblerDebug(__FILE__, __LINE__, s); } while (0)
#else
* @return - The InstructionContext to use when looking up an
* an instruction with these attributes.
*/
-static InstructionContext contextForAttrs(uint8_t attrMask) {
+static InstructionContext contextForAttrs(uint16_t attrMask) {
return CONTEXTS_SYM[attrMask];
}
* @return - TRUE if the ModR/M byte is required, FALSE otherwise.
*/
static int modRMRequired(OpcodeType type,
- InstructionContext insnContext,
- uint8_t opcode) {
+ InstructionContext insnContext,
+ uint16_t opcode) {
const struct ContextDecision* decision = 0;
-
+
switch (type) {
case ONEBYTE:
decision = &ONEBYTE_SYM;
case THREEBYTE_3A:
decision = &THREEBYTE3A_SYM;
break;
+ case THREEBYTE_A6:
+ decision = &THREEBYTEA6_SYM;
+ break;
+ case THREEBYTE_A7:
+ decision = &THREEBYTEA7_SYM;
+ break;
+ case XOP8_MAP:
+ decision = &XOP8_MAP_SYM;
+ break;
+ case XOP9_MAP:
+ decision = &XOP9_MAP_SYM;
+ break;
+ case XOPA_MAP:
+ decision = &XOPA_MAP_SYM;
+ break;
}
-
+
return decision->opcodeDecisions[insnContext].modRMDecisions[opcode].
modrm_type != MODRM_ONEENTRY;
-
- return 0;
}
/*
InstructionContext insnContext,
uint8_t opcode,
uint8_t modRM) {
- struct ModRMDecision* dec;
-
+ const struct ModRMDecision* dec = 0;
+
switch (type) {
- default:
- debug("Unknown opcode type");
- return 0;
case ONEBYTE:
dec = &ONEBYTE_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;
+ case XOP8_MAP:
+ dec = &XOP8_MAP_SYM.opcodeDecisions[insnContext].modRMDecisions[opcode];
+ break;
+ case XOP9_MAP:
+ dec = &XOP9_MAP_SYM.opcodeDecisions[insnContext].modRMDecisions[opcode];
+ break;
+ case XOPA_MAP:
+ dec = &XOPA_MAP_SYM.opcodeDecisions[insnContext].modRMDecisions[opcode];
+ break;
}
-
+
switch (dec->modrm_type) {
default:
debug("Corrupt table! Unknown modrm_type");
return 0;
case MODRM_ONEENTRY:
- return dec->instructionIDs[0];
+ return modRMTable[dec->instructionIDs];
case MODRM_SPLITRM:
if (modFromModRM(modRM) == 0x3)
- return dec->instructionIDs[1];
- else
- return dec->instructionIDs[0];
+ 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_SPLITMISC:
+ if (modFromModRM(modRM) == 0x3)
+ return modRMTable[dec->instructionIDs+(modRM & 0x3f)+8];
+ return modRMTable[dec->instructionIDs+((modRM & 0x38) >> 3)];
case MODRM_FULL:
- return dec->instructionIDs[modRM];
+ return modRMTable[dec->instructionIDs+modRM];
}
}
* decode(); specifierForUID will not check bounds.
* @return - A pointer to the specification for that instruction.
*/
-static struct InstructionSpecifier* specifierForUID(InstrUID uid) {
+static const struct InstructionSpecifier *specifierForUID(InstrUID uid) {
return &INSTRUCTIONS_SYM[uid];
}
*/
static int consumeByte(struct InternalInstruction* insn, uint8_t* byte) {
int ret = insn->reader(insn->readerArg, byte, insn->readerCursor);
-
+
if (!ret)
++(insn->readerCursor);
-
+
return ret;
}
insn->readerCursor + offset); \
if (ret) \
return ret; \
- combined = combined | ((type)byte << ((type)offset * 8)); \
+ combined = combined | ((uint64_t)byte << (offset * 8)); \
} \
*ptr = combined; \
insn->readerCursor += sizeof(type); \
*/
static void dbgprintf(struct InternalInstruction* insn,
const char* format,
- ...) {
+ ...) {
char buffer[256];
va_list ap;
-
+
if (!insn->dlog)
return;
-
+
va_start(ap, format);
(void)vsnprintf(buffer, sizeof(buffer), format, ap);
va_end(ap);
-
+
insn->dlog(insn->dlogArg, buffer);
-
+
return;
}
BOOL isPrefix = TRUE;
BOOL prefixGroups[4] = { FALSE };
uint64_t prefixLocation;
- uint8_t byte;
-
+ uint8_t byte = 0;
+ uint8_t nextByte;
+
BOOL hasAdSize = FALSE;
BOOL hasOpSize = FALSE;
-
+
dbgprintf(insn, "readPrefixes()");
-
+
while (isPrefix) {
prefixLocation = insn->readerCursor;
-
+
+ /* If we fail reading prefixes, just stop here and let the opcode reader deal with it */
if (consumeByte(insn, &byte))
- return -1;
-
+ break;
+
+ /*
+ * If the byte is a LOCK/REP/REPNE prefix and not a part of the opcode, then
+ * break and let it be disassembled as a normal "instruction".
+ */
+ if (insn->readerCursor - 1 == insn->startLocation && byte == 0xf0)
+ break;
+
+ if (insn->readerCursor - 1 == insn->startLocation
+ && (byte == 0xf2 || byte == 0xf3)
+ && !lookAtByte(insn, &nextByte))
+ {
+ /*
+ * If the byte is 0xf2 or 0xf3, and any of the following conditions are
+ * met:
+ * - it is followed by a LOCK (0xf0) prefix
+ * - it is followed by an xchg instruction
+ * then it should be disassembled as a xacquire/xrelease not repne/rep.
+ */
+ if ((byte == 0xf2 || byte == 0xf3) &&
+ ((nextByte == 0xf0) |
+ ((nextByte & 0xfe) == 0x86 || (nextByte & 0xf8) == 0x90)))
+ insn->xAcquireRelease = TRUE;
+ /*
+ * Also if the byte is 0xf3, and the following condition is met:
+ * - it is followed by a "mov mem, reg" (opcode 0x88/0x89) or
+ * "mov mem, imm" (opcode 0xc6/0xc7) instructions.
+ * then it should be disassembled as an xrelease not rep.
+ */
+ if (byte == 0xf3 &&
+ (nextByte == 0x88 || nextByte == 0x89 ||
+ nextByte == 0xc6 || nextByte == 0xc7))
+ insn->xAcquireRelease = TRUE;
+ if (insn->mode == MODE_64BIT && (nextByte & 0xf0) == 0x40) {
+ if (consumeByte(insn, &nextByte))
+ return -1;
+ if (lookAtByte(insn, &nextByte))
+ return -1;
+ unconsumeByte(insn);
+ }
+ if (nextByte != 0x0f && nextByte != 0x90)
+ break;
+ }
+
switch (byte) {
case 0xf0: /* LOCK */
case 0xf2: /* REPNE/REPNZ */
isPrefix = FALSE;
break;
}
-
+
if (isPrefix)
dbgprintf(insn, "Found prefix 0x%hhx", byte);
}
-
- if (insn->mode == MODE_64BIT) {
- if ((byte & 0xf0) == 0x40) {
- uint8_t opcodeByte;
-
- if (lookAtByte(insn, &opcodeByte) || ((opcodeByte & 0xf0) == 0x40)) {
- dbgprintf(insn, "Redundant REX prefix");
+
+ insn->vectorExtensionType = TYPE_NO_VEX_XOP;
+
+ if (byte == 0x62) {
+ uint8_t byte1, byte2;
+
+ if (consumeByte(insn, &byte1)) {
+ dbgprintf(insn, "Couldn't read second byte of EVEX prefix");
+ return -1;
+ }
+
+ if (lookAtByte(insn, &byte2)) {
+ dbgprintf(insn, "Couldn't read third byte of EVEX prefix");
+ return -1;
+ }
+
+ if ((insn->mode == MODE_64BIT || (byte1 & 0xc0) == 0xc0) &&
+ ((~byte1 & 0xc) == 0xc) && ((byte2 & 0x4) == 0x4)) {
+ insn->vectorExtensionType = TYPE_EVEX;
+ }
+ else {
+ unconsumeByte(insn); /* unconsume byte1 */
+ unconsumeByte(insn); /* unconsume byte */
+ insn->necessaryPrefixLocation = insn->readerCursor - 2;
+ }
+
+ if (insn->vectorExtensionType == TYPE_EVEX) {
+ insn->vectorExtensionPrefix[0] = byte;
+ insn->vectorExtensionPrefix[1] = byte1;
+ if (consumeByte(insn, &insn->vectorExtensionPrefix[2])) {
+ dbgprintf(insn, "Couldn't read third byte of EVEX prefix");
return -1;
}
-
- insn->rexPrefix = byte;
- insn->necessaryPrefixLocation = insn->readerCursor - 2;
-
- dbgprintf(insn, "Found REX prefix 0x%hhx", byte);
- } else {
+ if (consumeByte(insn, &insn->vectorExtensionPrefix[3])) {
+ dbgprintf(insn, "Couldn't read fourth byte of EVEX prefix");
+ return -1;
+ }
+
+ /* We simulate the REX prefix for simplicity's sake */
+ if (insn->mode == MODE_64BIT) {
+ insn->rexPrefix = 0x40
+ | (wFromEVEX3of4(insn->vectorExtensionPrefix[2]) << 3)
+ | (rFromEVEX2of4(insn->vectorExtensionPrefix[1]) << 2)
+ | (xFromEVEX2of4(insn->vectorExtensionPrefix[1]) << 1)
+ | (bFromEVEX2of4(insn->vectorExtensionPrefix[1]) << 0);
+ }
+
+ dbgprintf(insn, "Found EVEX prefix 0x%hhx 0x%hhx 0x%hhx 0x%hhx",
+ insn->vectorExtensionPrefix[0], insn->vectorExtensionPrefix[1],
+ insn->vectorExtensionPrefix[2], insn->vectorExtensionPrefix[3]);
+ }
+ }
+ else if (byte == 0xc4) {
+ uint8_t byte1;
+
+ if (lookAtByte(insn, &byte1)) {
+ dbgprintf(insn, "Couldn't read second byte of VEX");
+ return -1;
+ }
+
+ if (insn->mode == MODE_64BIT || (byte1 & 0xc0) == 0xc0) {
+ insn->vectorExtensionType = TYPE_VEX_3B;
+ insn->necessaryPrefixLocation = insn->readerCursor - 1;
+ }
+ else {
unconsumeByte(insn);
insn->necessaryPrefixLocation = insn->readerCursor - 1;
}
- } else {
- unconsumeByte(insn);
+
+ if (insn->vectorExtensionType == TYPE_VEX_3B) {
+ insn->vectorExtensionPrefix[0] = byte;
+ consumeByte(insn, &insn->vectorExtensionPrefix[1]);
+ consumeByte(insn, &insn->vectorExtensionPrefix[2]);
+
+ /* We simulate the REX prefix for simplicity's sake */
+
+ if (insn->mode == MODE_64BIT) {
+ insn->rexPrefix = 0x40
+ | (wFromVEX3of3(insn->vectorExtensionPrefix[2]) << 3)
+ | (rFromVEX2of3(insn->vectorExtensionPrefix[1]) << 2)
+ | (xFromVEX2of3(insn->vectorExtensionPrefix[1]) << 1)
+ | (bFromVEX2of3(insn->vectorExtensionPrefix[1]) << 0);
+ }
+
+ dbgprintf(insn, "Found VEX prefix 0x%hhx 0x%hhx 0x%hhx",
+ insn->vectorExtensionPrefix[0], insn->vectorExtensionPrefix[1],
+ insn->vectorExtensionPrefix[2]);
+ }
}
-
+ else if (byte == 0xc5) {
+ uint8_t byte1;
+
+ if (lookAtByte(insn, &byte1)) {
+ dbgprintf(insn, "Couldn't read second byte of VEX");
+ return -1;
+ }
+
+ if (insn->mode == MODE_64BIT || (byte1 & 0xc0) == 0xc0) {
+ insn->vectorExtensionType = TYPE_VEX_2B;
+ }
+ else {
+ unconsumeByte(insn);
+ }
+
+ if (insn->vectorExtensionType == TYPE_VEX_2B) {
+ insn->vectorExtensionPrefix[0] = byte;
+ consumeByte(insn, &insn->vectorExtensionPrefix[1]);
+
+ if (insn->mode == MODE_64BIT) {
+ insn->rexPrefix = 0x40
+ | (rFromVEX2of2(insn->vectorExtensionPrefix[1]) << 2);
+ }
+
+ switch (ppFromVEX2of2(insn->vectorExtensionPrefix[1]))
+ {
+ default:
+ break;
+ case VEX_PREFIX_66:
+ hasOpSize = TRUE;
+ break;
+ }
+
+ dbgprintf(insn, "Found VEX prefix 0x%hhx 0x%hhx",
+ insn->vectorExtensionPrefix[0],
+ insn->vectorExtensionPrefix[1]);
+ }
+ }
+ else if (byte == 0x8f) {
+ uint8_t byte1;
+
+ if (lookAtByte(insn, &byte1)) {
+ dbgprintf(insn, "Couldn't read second byte of XOP");
+ return -1;
+ }
+
+ if ((byte1 & 0x38) != 0x0) { /* 0 in these 3 bits is a POP instruction. */
+ insn->vectorExtensionType = TYPE_XOP;
+ insn->necessaryPrefixLocation = insn->readerCursor - 1;
+ }
+ else {
+ unconsumeByte(insn);
+ insn->necessaryPrefixLocation = insn->readerCursor - 1;
+ }
+
+ if (insn->vectorExtensionType == TYPE_XOP) {
+ insn->vectorExtensionPrefix[0] = byte;
+ consumeByte(insn, &insn->vectorExtensionPrefix[1]);
+ consumeByte(insn, &insn->vectorExtensionPrefix[2]);
+
+ /* We simulate the REX prefix for simplicity's sake */
+
+ if (insn->mode == MODE_64BIT) {
+ insn->rexPrefix = 0x40
+ | (wFromXOP3of3(insn->vectorExtensionPrefix[2]) << 3)
+ | (rFromXOP2of3(insn->vectorExtensionPrefix[1]) << 2)
+ | (xFromXOP2of3(insn->vectorExtensionPrefix[1]) << 1)
+ | (bFromXOP2of3(insn->vectorExtensionPrefix[1]) << 0);
+ }
+
+ switch (ppFromXOP3of3(insn->vectorExtensionPrefix[2]))
+ {
+ default:
+ break;
+ case VEX_PREFIX_66:
+ hasOpSize = TRUE;
+ break;
+ }
+
+ dbgprintf(insn, "Found XOP prefix 0x%hhx 0x%hhx 0x%hhx",
+ insn->vectorExtensionPrefix[0], insn->vectorExtensionPrefix[1],
+ insn->vectorExtensionPrefix[2]);
+ }
+ }
+ else {
+ if (insn->mode == MODE_64BIT) {
+ if ((byte & 0xf0) == 0x40) {
+ uint8_t opcodeByte;
+
+ if (lookAtByte(insn, &opcodeByte) || ((opcodeByte & 0xf0) == 0x40)) {
+ dbgprintf(insn, "Redundant REX prefix");
+ return -1;
+ }
+
+ insn->rexPrefix = byte;
+ insn->necessaryPrefixLocation = insn->readerCursor - 2;
+
+ dbgprintf(insn, "Found REX prefix 0x%hhx", byte);
+ } else {
+ unconsumeByte(insn);
+ insn->necessaryPrefixLocation = insn->readerCursor - 1;
+ }
+ } else {
+ unconsumeByte(insn);
+ insn->necessaryPrefixLocation = insn->readerCursor - 1;
+ }
+ }
+
if (insn->mode == MODE_16BIT) {
insn->registerSize = (hasOpSize ? 4 : 2);
insn->addressSize = (hasAdSize ? 4 : 2);
insn->immediateSize = (hasOpSize ? 2 : 4);
}
}
-
+
return 0;
}
* @param insn - The instruction whose opcode is to be read.
* @return - 0 if the opcode could be read successfully; nonzero otherwise.
*/
-static int readOpcode(struct InternalInstruction* insn) {
+static int readOpcode(struct InternalInstruction* insn) {
/* Determine the length of the primary opcode */
-
+
uint8_t current;
-
+
dbgprintf(insn, "readOpcode()");
-
+
insn->opcodeType = ONEBYTE;
+
+ if (insn->vectorExtensionType == TYPE_EVEX)
+ {
+ switch (mmFromEVEX2of4(insn->vectorExtensionPrefix[1])) {
+ default:
+ dbgprintf(insn, "Unhandled mm field for instruction (0x%hhx)",
+ mmFromEVEX2of4(insn->vectorExtensionPrefix[1]));
+ return -1;
+ case VEX_LOB_0F:
+ insn->opcodeType = TWOBYTE;
+ return consumeByte(insn, &insn->opcode);
+ case VEX_LOB_0F38:
+ insn->opcodeType = THREEBYTE_38;
+ return consumeByte(insn, &insn->opcode);
+ case VEX_LOB_0F3A:
+ insn->opcodeType = THREEBYTE_3A;
+ return consumeByte(insn, &insn->opcode);
+ }
+ }
+ else if (insn->vectorExtensionType == TYPE_VEX_3B) {
+ switch (mmmmmFromVEX2of3(insn->vectorExtensionPrefix[1])) {
+ default:
+ dbgprintf(insn, "Unhandled m-mmmm field for instruction (0x%hhx)",
+ mmmmmFromVEX2of3(insn->vectorExtensionPrefix[1]));
+ return -1;
+ case VEX_LOB_0F:
+ insn->opcodeType = TWOBYTE;
+ return consumeByte(insn, &insn->opcode);
+ case VEX_LOB_0F38:
+ insn->opcodeType = THREEBYTE_38;
+ return consumeByte(insn, &insn->opcode);
+ case VEX_LOB_0F3A:
+ insn->opcodeType = THREEBYTE_3A;
+ return consumeByte(insn, &insn->opcode);
+ }
+ }
+ else if (insn->vectorExtensionType == TYPE_VEX_2B) {
+ insn->opcodeType = TWOBYTE;
+ return consumeByte(insn, &insn->opcode);
+ }
+ else if (insn->vectorExtensionType == TYPE_XOP) {
+ switch (mmmmmFromXOP2of3(insn->vectorExtensionPrefix[1])) {
+ default:
+ dbgprintf(insn, "Unhandled m-mmmm field for instruction (0x%hhx)",
+ mmmmmFromVEX2of3(insn->vectorExtensionPrefix[1]));
+ return -1;
+ case XOP_MAP_SELECT_8:
+ insn->opcodeType = XOP8_MAP;
+ return consumeByte(insn, &insn->opcode);
+ case XOP_MAP_SELECT_9:
+ insn->opcodeType = XOP9_MAP;
+ return consumeByte(insn, &insn->opcode);
+ case XOP_MAP_SELECT_A:
+ insn->opcodeType = XOPA_MAP;
+ return consumeByte(insn, &insn->opcode);
+ }
+ }
+
if (consumeByte(insn, ¤t))
return -1;
-
+
if (current == 0x0f) {
dbgprintf(insn, "Found a two-byte escape prefix (0x%hhx)", current);
-
- insn->twoByteEscape = current;
-
+
if (consumeByte(insn, ¤t))
return -1;
-
+
if (current == 0x38) {
dbgprintf(insn, "Found a three-byte escape prefix (0x%hhx)", current);
-
- insn->threeByteEscape = current;
-
+
if (consumeByte(insn, ¤t))
return -1;
-
+
insn->opcodeType = THREEBYTE_38;
} else if (current == 0x3a) {
dbgprintf(insn, "Found a three-byte escape prefix (0x%hhx)", current);
-
- insn->threeByteEscape = current;
-
+
if (consumeByte(insn, ¤t))
return -1;
-
+
insn->opcodeType = THREEBYTE_3A;
+ } else if (current == 0xa6) {
+ dbgprintf(insn, "Found a three-byte escape prefix (0x%hhx)", current);
+
+ if (consumeByte(insn, ¤t))
+ return -1;
+
+ insn->opcodeType = THREEBYTE_A6;
+ } else if (current == 0xa7) {
+ dbgprintf(insn, "Found a three-byte escape prefix (0x%hhx)", current);
+
+ if (consumeByte(insn, ¤t))
+ return -1;
+
+ insn->opcodeType = THREEBYTE_A7;
} else {
dbgprintf(insn, "Didn't find a three-byte escape prefix");
-
+
insn->opcodeType = TWOBYTE;
}
}
-
+
/*
* At this point we have consumed the full opcode.
* Anything we consume from here on must be unconsumed.
*/
-
+
insn->opcode = current;
-
+
return 0;
}
*/
static int getIDWithAttrMask(uint16_t* instructionID,
struct InternalInstruction* insn,
- uint8_t attrMask) {
+ uint16_t attrMask) {
BOOL hasModRMExtension;
-
- uint8_t instructionClass;
+
+ uint16_t instructionClass;
instructionClass = contextForAttrs(attrMask);
-
+
hasModRMExtension = modRMRequired(insn->opcodeType,
instructionClass,
insn->opcode);
-
+
if (hasModRMExtension) {
- readModRM(insn);
-
+ if (readModRM(insn))
+ return -1;
+
*instructionID = decode(insn->opcodeType,
instructionClass,
insn->opcode,
insn->opcode,
0);
}
-
+
return 0;
}
* @param orig - The instruction that is not 16-bit
* @param equiv - The instruction that is 16-bit
*/
-static BOOL is16BitEquvalent(const char* orig, const char* equiv) {
+static BOOL is16BitEquivalent(const char* orig, const char* equiv) {
off_t i;
-
+
for (i = 0;; i++) {
if (orig[i] == '\0' && equiv[i] == '\0')
return TRUE;
}
/*
- * is64BitEquivalent - Determines whether two instruction names refer to
- * equivalent instructions but one is 64-bit whereas the other is not.
- *
- * @param orig - The instruction that is not 64-bit
- * @param equiv - The instruction that is 64-bit
- */
-static BOOL is64BitEquivalent(const char* orig, const char* equiv) {
- off_t i;
-
- for (i = 0;; i++) {
- if (orig[i] == '\0' && equiv[i] == '\0')
- return TRUE;
- if (orig[i] == '\0' || equiv[i] == '\0')
- return FALSE;
- if (orig[i] != equiv[i]) {
- if ((orig[i] == 'W' || orig[i] == 'L') && equiv[i] == 'Q')
- continue;
- if ((orig[i] == '1' || orig[i] == '3') && equiv[i] == '6')
- continue;
- if ((orig[i] == '6' || orig[i] == '2') && equiv[i] == '4')
- continue;
- return FALSE;
- }
- }
-}
-
-
-/*
- * getID - Determines the ID of an instruction, consuming the ModR/M byte as
- * appropriate for extended and escape opcodes. Determines the attributes and
+ * getID - Determines the ID of an instruction, consuming the ModR/M byte as
+ * appropriate for extended and escape opcodes. Determines the attributes and
* context for the instruction before doing so.
*
* @param insn - The instruction whose ID is to be determined.
* @return - 0 if the ModR/M could be read when needed or was not needed;
* nonzero otherwise.
*/
-static int getID(struct InternalInstruction* insn) {
- uint8_t attrMask;
+static int getID(struct InternalInstruction* insn, const void *miiArg) {
+ uint16_t attrMask;
uint16_t instructionID;
-
+
dbgprintf(insn, "getID()");
-
+
attrMask = ATTR_NONE;
-
+
if (insn->mode == MODE_64BIT)
attrMask |= ATTR_64BIT;
-
+
+ if (insn->vectorExtensionType != TYPE_NO_VEX_XOP) {
+ attrMask |= (insn->vectorExtensionType == TYPE_EVEX) ? ATTR_EVEX : ATTR_VEX;
+
+ if (insn->vectorExtensionType == TYPE_EVEX) {
+ switch (ppFromEVEX3of4(insn->vectorExtensionPrefix[2])) {
+ case VEX_PREFIX_66:
+ attrMask |= ATTR_OPSIZE;
+ break;
+ case VEX_PREFIX_F3:
+ attrMask |= ATTR_XS;
+ break;
+ case VEX_PREFIX_F2:
+ attrMask |= ATTR_XD;
+ break;
+ }
+
+ if (zFromEVEX4of4(insn->vectorExtensionPrefix[3]))
+ attrMask |= ATTR_EVEXKZ;
+ if (bFromEVEX4of4(insn->vectorExtensionPrefix[3]))
+ attrMask |= ATTR_EVEXB;
+ if (aaaFromEVEX4of4(insn->vectorExtensionPrefix[3]))
+ attrMask |= ATTR_EVEXK;
+ if (lFromEVEX4of4(insn->vectorExtensionPrefix[3]))
+ attrMask |= ATTR_EVEXL;
+ if (l2FromEVEX4of4(insn->vectorExtensionPrefix[3]))
+ attrMask |= ATTR_EVEXL2;
+ }
+ else if (insn->vectorExtensionType == TYPE_VEX_3B) {
+ switch (ppFromVEX3of3(insn->vectorExtensionPrefix[2])) {
+ case VEX_PREFIX_66:
+ attrMask |= ATTR_OPSIZE;
+ break;
+ case VEX_PREFIX_F3:
+ attrMask |= ATTR_XS;
+ break;
+ case VEX_PREFIX_F2:
+ attrMask |= ATTR_XD;
+ break;
+ }
+
+ if (lFromVEX3of3(insn->vectorExtensionPrefix[2]))
+ attrMask |= ATTR_VEXL;
+ }
+ else if (insn->vectorExtensionType == TYPE_VEX_2B) {
+ switch (ppFromVEX2of2(insn->vectorExtensionPrefix[1])) {
+ case VEX_PREFIX_66:
+ attrMask |= ATTR_OPSIZE;
+ break;
+ case VEX_PREFIX_F3:
+ attrMask |= ATTR_XS;
+ break;
+ case VEX_PREFIX_F2:
+ attrMask |= ATTR_XD;
+ break;
+ }
+
+ if (lFromVEX2of2(insn->vectorExtensionPrefix[1]))
+ attrMask |= ATTR_VEXL;
+ }
+ else if (insn->vectorExtensionType == TYPE_XOP) {
+ switch (ppFromXOP3of3(insn->vectorExtensionPrefix[2])) {
+ case VEX_PREFIX_66:
+ attrMask |= ATTR_OPSIZE;
+ break;
+ case VEX_PREFIX_F3:
+ attrMask |= ATTR_XS;
+ break;
+ case VEX_PREFIX_F2:
+ attrMask |= ATTR_XD;
+ break;
+ }
+
+ if (lFromXOP3of3(insn->vectorExtensionPrefix[2]))
+ attrMask |= ATTR_VEXL;
+ }
+ else {
+ return -1;
+ }
+ }
+ else {
+ if (insn->mode != MODE_16BIT && isPrefixAtLocation(insn, 0x66, insn->necessaryPrefixLocation))
+ attrMask |= ATTR_OPSIZE;
+ else if (isPrefixAtLocation(insn, 0x67, insn->necessaryPrefixLocation))
+ attrMask |= ATTR_ADSIZE;
+ else if (isPrefixAtLocation(insn, 0xf3, insn->necessaryPrefixLocation))
+ attrMask |= ATTR_XS;
+ else if (isPrefixAtLocation(insn, 0xf2, insn->necessaryPrefixLocation))
+ attrMask |= ATTR_XD;
+ }
+
if (insn->rexPrefix & 0x08)
attrMask |= ATTR_REXW;
-
- if (isPrefixAtLocation(insn, 0x66, insn->necessaryPrefixLocation))
- attrMask |= ATTR_OPSIZE;
- else if (isPrefixAtLocation(insn, 0xf3, insn->necessaryPrefixLocation))
- attrMask |= ATTR_XS;
- else if (isPrefixAtLocation(insn, 0xf2, insn->necessaryPrefixLocation))
- attrMask |= ATTR_XD;
-
+
if (getIDWithAttrMask(&instructionID, insn, attrMask))
return -1;
-
- /* The following clauses compensate for limitations of the tables. */
-
- if ((attrMask & ATTR_XD) && (attrMask & ATTR_REXW)) {
+
+ /*
+ * JCXZ/JECXZ need special handling for 16-bit mode because the meaning
+ * of the AdSize prefix is inverted w.r.t. 32-bit mode.
+ */
+ if (insn->mode == MODE_16BIT && insn->opcode == 0xE3) {
+ const struct InstructionSpecifier *spec;
+ spec = specifierForUID(instructionID);
+
/*
- * Although for SSE instructions it is usually necessary to treat REX.W+F2
- * as F2 for decode (in the absence of a 64BIT_REXW_XD category) there is
- * an occasional instruction where F2 is incidental and REX.W is the more
- * significant. If the decoded instruction is 32-bit and adding REX.W
- * instead of F2 changes a 32 to a 64, we adopt the new encoding.
+ * Check for Ii8PCRel instructions. We could alternatively do a
+ * string-compare on the names, but this is probably cheaper.
*/
-
- struct InstructionSpecifier* spec;
- uint16_t instructionIDWithREXw;
- struct InstructionSpecifier* specWithREXw;
-
- spec = specifierForUID(instructionID);
-
- if (getIDWithAttrMask(&instructionIDWithREXw,
- insn,
- attrMask & (~ATTR_XD))) {
- /*
- * Decoding with REX.w would yield nothing; give up and return original
- * decode.
- */
-
- insn->instructionID = instructionID;
- insn->spec = spec;
- return 0;
- }
-
- specWithREXw = specifierForUID(instructionIDWithREXw);
-
- if (is64BitEquivalent(spec->name, specWithREXw->name)) {
- insn->instructionID = instructionIDWithREXw;
- insn->spec = specWithREXw;
- } else {
- insn->instructionID = instructionID;
- insn->spec = spec;
+ if (x86OperandSets[spec->operands][0].type == TYPE_REL8) {
+ attrMask ^= ATTR_ADSIZE;
+ if (getIDWithAttrMask(&instructionID, insn, attrMask))
+ return -1;
}
- return 0;
}
-
- if (insn->prefixPresent[0x66] && !(attrMask & ATTR_OPSIZE)) {
+
+ /* The following clauses compensate for limitations of the tables. */
+
+ if ((insn->mode == MODE_16BIT || insn->prefixPresent[0x66]) &&
+ !(attrMask & ATTR_OPSIZE)) {
/*
* The instruction tables make no distinction between instructions that
* allow OpSize anywhere (i.e., 16-bit operations) and that need it in a
* conservative, but in the specific case where OpSize is present but not
* in the right place we check if there's a 16-bit operation.
*/
-
- struct InstructionSpecifier* spec;
+
+ const struct InstructionSpecifier *spec;
uint16_t instructionIDWithOpsize;
- struct InstructionSpecifier* specWithOpsize;
-
+ const char *specName, *specWithOpSizeName;
+
spec = specifierForUID(instructionID);
-
+
if (getIDWithAttrMask(&instructionIDWithOpsize,
insn,
attrMask | ATTR_OPSIZE)) {
- /*
+ /*
* ModRM required with OpSize but not present; give up and return version
* without OpSize set
*/
-
+
insn->instructionID = instructionID;
insn->spec = spec;
return 0;
}
-
- specWithOpsize = specifierForUID(instructionIDWithOpsize);
-
- if (is16BitEquvalent(spec->name, specWithOpsize->name)) {
+
+ specName = x86DisassemblerGetInstrName(instructionID, miiArg);
+ specWithOpSizeName =
+ x86DisassemblerGetInstrName(instructionIDWithOpsize, miiArg);
+
+ if (is16BitEquivalent(specName, specWithOpSizeName) &&
+ (insn->mode == MODE_16BIT) ^ insn->prefixPresent[0x66]) {
insn->instructionID = instructionIDWithOpsize;
- insn->spec = specWithOpsize;
+ insn->spec = specifierForUID(instructionIDWithOpsize);
} else {
insn->instructionID = instructionID;
insn->spec = spec;
}
return 0;
}
-
+
+ if (insn->opcodeType == ONEBYTE && insn->opcode == 0x90 &&
+ insn->rexPrefix & 0x01) {
+ /*
+ * NOOP shouldn't decode as NOOP if REX.b is set. Instead
+ * it should decode as XCHG %r8, %eax.
+ */
+
+ const struct InstructionSpecifier *spec;
+ uint16_t instructionIDWithNewOpcode;
+ const struct InstructionSpecifier *specWithNewOpcode;
+
+ spec = specifierForUID(instructionID);
+
+ /* Borrow opcode from one of the other XCHGar opcodes */
+ insn->opcode = 0x91;
+
+ if (getIDWithAttrMask(&instructionIDWithNewOpcode,
+ insn,
+ attrMask)) {
+ insn->opcode = 0x90;
+
+ insn->instructionID = instructionID;
+ insn->spec = spec;
+ return 0;
+ }
+
+ specWithNewOpcode = specifierForUID(instructionIDWithNewOpcode);
+
+ /* Change back */
+ insn->opcode = 0x90;
+
+ insn->instructionID = instructionIDWithNewOpcode;
+ insn->spec = specWithNewOpcode;
+
+ return 0;
+ }
+
insn->instructionID = instructionID;
insn->spec = specifierForUID(insn->instructionID);
-
+
return 0;
}
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");
if (consumeByte(insn, &insn->sib))
return -1;
-
+
index = indexFromSIB(insn->sib) | (xFromREX(insn->rexPrefix) << 3);
-
+ if (insn->vectorExtensionType == TYPE_EVEX)
+ index |= v2FromEVEX4of4(insn->vectorExtensionPrefix[3]) << 4;
+
switch (index) {
case 0x4:
insn->sibIndex = SIB_INDEX_NONE;
break;
default:
- insn->sibIndex = (EABase)(sibIndexBase + index);
+ 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;
insn->sibScale = 8;
break;
}
-
+
base = baseFromSIB(insn->sib) | (bFromREX(insn->rexPrefix) << 3);
-
+
switch (base) {
case 0x5:
switch (modFromModRM(insn->modRM)) {
break;
case 0x1:
insn->eaDisplacement = EA_DISP_8;
- insn->sibBase = (insn->addressSize == 4 ?
+ insn->sibBase = (insn->addressSize == 4 ?
SIB_BASE_EBP : SIB_BASE_RBP);
break;
case 0x2:
insn->eaDisplacement = EA_DISP_32;
- insn->sibBase = (insn->addressSize == 4 ?
+ insn->sibBase = (insn->addressSize == 4 ?
SIB_BASE_EBP : SIB_BASE_RBP);
break;
case 0x3:
}
break;
default:
- insn->sibBase = (EABase)(sibBaseBase + base);
+ insn->sibBase = (SIBBase)(sibBaseBase + base);
break;
}
-
+
return 0;
}
* readDisplacement - Consumes the displacement of an instruction.
*
* @param insn - The instruction whose displacement is to be read.
- * @return - 0 if the displacement byte was successfully read; nonzero
+ * @return - 0 if the displacement byte was successfully read; nonzero
* otherwise.
*/
-static int readDisplacement(struct InternalInstruction* insn) {
+static int readDisplacement(struct InternalInstruction* insn) {
int8_t d8;
int16_t d16;
int32_t d32;
-
+
dbgprintf(insn, "readDisplacement()");
-
+
if (insn->consumedDisplacement)
return 0;
-
+
insn->consumedDisplacement = TRUE;
-
+ insn->displacementOffset = insn->readerCursor - insn->startLocation;
+
switch (insn->eaDisplacement) {
case EA_DISP_NONE:
insn->consumedDisplacement = FALSE;
insn->displacement = d32;
break;
}
-
+
insn->consumedDisplacement = TRUE;
return 0;
}
* @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) {
+static int readModRM(struct InternalInstruction* insn) {
uint8_t mod, rm, reg;
-
+
dbgprintf(insn, "readModRM()");
-
+
if (insn->consumedModRM)
return 0;
-
- consumeByte(insn, &insn->modRM);
+
+ 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
insn->eaRegBase = EA_REG_RAX;
break;
}
-
+
reg |= rFromREX(insn->rexPrefix) << 3;
rm |= bFromREX(insn->rexPrefix) << 3;
-
+ if (insn->vectorExtensionType == TYPE_EVEX) {
+ reg |= r2FromEVEX2of4(insn->vectorExtensionPrefix[1]) << 4;
+ rm |= xFromEVEX2of4(insn->vectorExtensionPrefix[1]) << 4;
+ }
+
insn->reg = (Reg)(insn->regBase + reg);
-
+
switch (insn->addressSize) {
case 2:
insn->eaBaseBase = EA_BASE_BX_SI;
-
+
switch (mod) {
case 0x0:
if (rm == 0x6) {
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 0x14:
case 0x4:
case 0xc: /* in case REXW.b is set */
- insn->eaBase = (insn->addressSize == 4 ?
+ insn->eaBase = (insn->addressSize == 4 ?
EA_BASE_sib : EA_BASE_sib64);
readSIB(insn);
if (readDisplacement(insn))
case 0x2:
insn->eaDisplacement = (mod == 0x1 ? EA_DISP_8 : EA_DISP_32);
switch (rm) {
+ case 0x14:
case 0x4:
case 0xc: /* in case REXW.b is set */
insn->eaBase = EA_BASE_sib;
}
break;
} /* switch (insn->addressSize) */
-
+
return 0;
}
return prefix##_EAX + index; \
case TYPE_R64: \
return prefix##_RAX + index; \
+ case TYPE_XMM512: \
+ return prefix##_ZMM0 + index; \
+ case TYPE_XMM256: \
+ return prefix##_YMM0 + index; \
case TYPE_XMM128: \
case TYPE_XMM64: \
case TYPE_XMM32: \
case TYPE_XMM: \
return prefix##_XMM0 + index; \
+ case TYPE_VK1: \
+ case TYPE_VK8: \
+ case TYPE_VK16: \
+ return prefix##_K0 + index; \
case TYPE_MM64: \
case TYPE_MM32: \
case TYPE_MM: \
* @return - 0 if fixup was successful; -1 if the register returned was
* invalid for its class.
*/
-static int fixupReg(struct InternalInstruction *insn,
- struct OperandSpecifier *op) {
+static int fixupReg(struct InternalInstruction *insn,
+ const struct OperandSpecifier *op) {
uint8_t valid;
-
+
dbgprintf(insn, "fixupReg()");
-
+
switch ((OperandEncoding)op->encoding) {
default:
debug("Expected a REG or R/M encoding in fixupReg");
return -1;
+ case ENCODING_VVVV:
+ insn->vvvv = (Reg)fixupRegValue(insn,
+ (OperandType)op->type,
+ insn->vvvv,
+ &valid);
+ if (!valid)
+ return -1;
+ break;
case ENCODING_REG:
insn->reg = (Reg)fixupRegValue(insn,
(OperandType)op->type,
}
break;
}
-
- return 0;
-}
-/*
- * readOpcodeModifier - Reads an operand from the opcode field of an
- * instruction. Handles AddRegFrm instructions.
- *
- * @param insn - The instruction whose opcode field is to be read.
- * @param inModRM - Indicates that the opcode field is to be read from the
- * ModR/M extension; useful for escape opcodes
- * @return - 0 on success; nonzero otherwise.
- */
-static int readOpcodeModifier(struct InternalInstruction* insn) {
- dbgprintf(insn, "readOpcodeModifier()");
-
- if (insn->consumedOpcodeModifier)
- return 0;
-
- insn->consumedOpcodeModifier = TRUE;
-
- switch (insn->spec->modifierType) {
- default:
- debug("Unknown modifier type.");
- return -1;
- case MODIFIER_NONE:
- debug("No modifier but an operand expects one.");
- return -1;
- case MODIFIER_OPCODE:
- insn->opcodeModifier = insn->opcode - insn->spec->modifierBase;
- return 0;
- case MODIFIER_MODRM:
- insn->opcodeModifier = insn->modRM - insn->spec->modifierBase;
- return 0;
- }
+ return 0;
}
/*
- * readOpcodeRegister - Reads an operand from the opcode field of an
+ * readOpcodeRegister - Reads an operand from the opcode field of an
* instruction and interprets it appropriately given the operand width.
* Handles AddRegFrm instructions.
*
- * @param insn - See readOpcodeModifier().
+ * @param insn - the instruction whose opcode field is to be read.
* @param size - The width (in bytes) of the register being specified.
* 1 means AL and friends, 2 means AX, 4 means EAX, and 8 means
* RAX.
static int readOpcodeRegister(struct InternalInstruction* insn, uint8_t size) {
dbgprintf(insn, "readOpcodeRegister()");
- if (readOpcodeModifier(insn))
- return -1;
-
if (size == 0)
size = insn->registerSize;
-
+
switch (size) {
case 1:
- insn->opcodeRegister = (Reg)(MODRM_REG_AL + ((bFromREX(insn->rexPrefix) << 3)
- | insn->opcodeModifier));
- if (insn->rexPrefix &&
+ insn->opcodeRegister = (Reg)(MODRM_REG_AL + ((bFromREX(insn->rexPrefix) << 3)
+ | (insn->opcode & 7)));
+ if (insn->rexPrefix &&
insn->opcodeRegister >= MODRM_REG_AL + 0x4 &&
insn->opcodeRegister < MODRM_REG_AL + 0x8) {
insn->opcodeRegister = (Reg)(MODRM_REG_SPL
+ (insn->opcodeRegister - MODRM_REG_AL - 4));
}
-
+
break;
case 2:
insn->opcodeRegister = (Reg)(MODRM_REG_AX
- + ((bFromREX(insn->rexPrefix) << 3)
- | insn->opcodeModifier));
+ + ((bFromREX(insn->rexPrefix) << 3)
+ | (insn->opcode & 7)));
break;
case 4:
insn->opcodeRegister = (Reg)(MODRM_REG_EAX
- + ((bFromREX(insn->rexPrefix) << 3)
- | insn->opcodeModifier));
+ + ((bFromREX(insn->rexPrefix) << 3)
+ | (insn->opcode & 7)));
break;
case 8:
- insn->opcodeRegister = (Reg)(MODRM_REG_RAX
- + ((bFromREX(insn->rexPrefix) << 3)
- | insn->opcodeModifier));
+ insn->opcodeRegister = (Reg)(MODRM_REG_RAX
+ + ((bFromREX(insn->rexPrefix) << 3)
+ | (insn->opcode & 7)));
break;
}
-
+
return 0;
}
uint16_t imm16;
uint32_t imm32;
uint64_t imm64;
-
+
dbgprintf(insn, "readImmediate()");
-
+
if (insn->numImmediatesConsumed == 2) {
debug("Already consumed two immediates");
return -1;
}
-
+
if (size == 0)
size = insn->immediateSize;
else
insn->immediateSize = size;
-
+ insn->immediateOffset = insn->readerCursor - insn->startLocation;
+
switch (size) {
case 1:
if (consumeByte(insn, &imm8))
insn->immediates[insn->numImmediatesConsumed] = imm64;
break;
}
-
+
insn->numImmediatesConsumed++;
-
+
+ return 0;
+}
+
+/*
+ * readVVVV - Consumes vvvv from an instruction if it has a VEX prefix.
+ *
+ * @param insn - The instruction whose operand is to be read.
+ * @return - 0 if the vvvv was successfully consumed; nonzero
+ * otherwise.
+ */
+static int readVVVV(struct InternalInstruction* insn) {
+ dbgprintf(insn, "readVVVV()");
+
+ if (insn->vectorExtensionType == TYPE_EVEX)
+ insn->vvvv = vvvvFromEVEX3of4(insn->vectorExtensionPrefix[2]);
+ else if (insn->vectorExtensionType == TYPE_VEX_3B)
+ insn->vvvv = vvvvFromVEX3of3(insn->vectorExtensionPrefix[2]);
+ else if (insn->vectorExtensionType == TYPE_VEX_2B)
+ insn->vvvv = vvvvFromVEX2of2(insn->vectorExtensionPrefix[1]);
+ else if (insn->vectorExtensionType == TYPE_XOP)
+ insn->vvvv = vvvvFromXOP3of3(insn->vectorExtensionPrefix[2]);
+ else
+ return -1;
+
+ if (insn->mode != MODE_64BIT)
+ insn->vvvv &= 0x7;
+
+ return 0;
+}
+
+/*
+ * readMaskRegister - Reads an mask register from the opcode field of an
+ * instruction.
+ *
+ * @param insn - The instruction whose opcode field is to be read.
+ * @return - 0 on success; nonzero otherwise.
+ */
+static int readMaskRegister(struct InternalInstruction* insn) {
+ dbgprintf(insn, "readMaskRegister()");
+
+ if (insn->vectorExtensionType != TYPE_EVEX)
+ return -1;
+
+ insn->writemask = aaaFromEVEX4of4(insn->vectorExtensionPrefix[3]);
return 0;
}
*/
static int readOperands(struct InternalInstruction* insn) {
int index;
-
+ int hasVVVV, needVVVV;
+ int sawRegImm = 0;
+
dbgprintf(insn, "readOperands()");
-
+
+ /* If non-zero vvvv specified, need to make sure one of the operands
+ uses it. */
+ hasVVVV = !readVVVV(insn);
+ needVVVV = hasVVVV && (insn->vvvv != 0);
+
for (index = 0; index < X86_MAX_OPERANDS; ++index) {
- switch (insn->spec->operands[index].encoding) {
+ switch (x86OperandSets[insn->spec->operands][index].encoding) {
case ENCODING_NONE:
+ case ENCODING_SI:
+ case ENCODING_DI:
break;
case ENCODING_REG:
case ENCODING_RM:
if (readModRM(insn))
return -1;
- if (fixupReg(insn, &insn->spec->operands[index]))
+ if (fixupReg(insn, &x86OperandSets[insn->spec->operands][index]))
return -1;
break;
case ENCODING_CB:
dbgprintf(insn, "We currently don't hande code-offset encodings");
return -1;
case ENCODING_IB:
+ if (sawRegImm) {
+ /* Saw a register immediate so don't read again and instead split the
+ previous immediate. FIXME: This is a hack. */
+ insn->immediates[insn->numImmediatesConsumed] =
+ insn->immediates[insn->numImmediatesConsumed - 1] & 0xf;
+ ++insn->numImmediatesConsumed;
+ break;
+ }
if (readImmediate(insn, 1))
return -1;
- if (insn->spec->operands[index].type == TYPE_IMM3 &&
+ if (x86OperandSets[insn->spec->operands][index].type == TYPE_IMM3 &&
insn->immediates[insn->numImmediatesConsumed - 1] > 7)
return -1;
+ if (x86OperandSets[insn->spec->operands][index].type == TYPE_IMM5 &&
+ insn->immediates[insn->numImmediatesConsumed - 1] > 31)
+ return -1;
+ if (x86OperandSets[insn->spec->operands][index].type == TYPE_XMM128 ||
+ x86OperandSets[insn->spec->operands][index].type == TYPE_XMM256)
+ sawRegImm = 1;
break;
case ENCODING_IW:
if (readImmediate(insn, 2))
if (readOpcodeRegister(insn, 0))
return -1;
break;
- case ENCODING_I:
- if (readOpcodeModifier(insn))
+ case ENCODING_FP:
+ break;
+ case ENCODING_VVVV:
+ needVVVV = 0; /* Mark that we have found a VVVV operand. */
+ if (!hasVVVV)
+ return -1;
+ if (fixupReg(insn, &x86OperandSets[insn->spec->operands][index]))
+ return -1;
+ break;
+ case ENCODING_WRITEMASK:
+ if (readMaskRegister(insn))
return -1;
+ break;
case ENCODING_DUP:
break;
default:
return -1;
}
}
-
+
+ /* If we didn't find ENCODING_VVVV operand, but non-zero vvvv present, fail */
+ if (needVVVV) return -1;
+
return 0;
}
* decodeInstruction - Reads and interprets a full instruction provided by the
* user.
*
- * @param insn - A pointer to the instruction to be populated. Must be
+ * @param insn - A pointer to the instruction to be populated. Must be
* pre-allocated.
* @param reader - The function to be used to read the instruction's bytes.
* @param readerArg - A generic argument to be passed to the reader to store
*/
int decodeInstruction(struct InternalInstruction* insn,
byteReader_t reader,
- void* readerArg,
+ const void* readerArg,
dlog_t logger,
void* loggerArg,
+ const void* miiArg,
uint64_t startLoc,
DisassemblerMode mode) {
memset(insn, 0, sizeof(struct InternalInstruction));
-
+
insn->reader = reader;
insn->readerArg = readerArg;
insn->dlog = logger;
insn->readerCursor = startLoc;
insn->mode = mode;
insn->numImmediatesConsumed = 0;
-
+
if (readPrefixes(insn) ||
readOpcode(insn) ||
- getID(insn) ||
+ getID(insn, miiArg) ||
insn->instructionID == 0 ||
readOperands(insn))
return -1;
-
+
+ insn->operands = &x86OperandSets[insn->spec->operands][0];
+
insn->length = insn->readerCursor - insn->startLocation;
-
+
dbgprintf(insn, "Read from 0x%llx to 0x%llx: length %zu",
startLoc, insn->readerCursor, insn->length);
-
+
if (insn->length > 15)
dbgprintf(insn, "Instruction exceeds 15-byte limit");
-
+
return 0;
}