Linux Audio

Check our new training course

Linux kernel drivers training

Mar 31-Apr 9, 2025, special US time zones
Register
Loading...
v6.8
  1// SPDX-License-Identifier: GPL-2.0-or-later
  2/*
  3 * x86 instruction analysis
  4 *
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  5 * Copyright (C) IBM Corporation, 2002, 2004, 2009
  6 */
  7
  8#include <linux/kernel.h>
  9#ifdef __KERNEL__
 10#include <linux/string.h>
 11#else
 12#include <string.h>
 13#endif
 14#include <asm/inat.h> /*__ignore_sync_check__ */
 15#include <asm/insn.h> /* __ignore_sync_check__ */
 16#include <asm/unaligned.h> /* __ignore_sync_check__ */
 17
 18#include <linux/errno.h>
 19#include <linux/kconfig.h>
 20
 21#include <asm/emulate_prefix.h> /* __ignore_sync_check__ */
 22
 23#define leXX_to_cpu(t, r)						\
 24({									\
 25	__typeof__(t) v;						\
 26	switch (sizeof(t)) {						\
 27	case 4: v = le32_to_cpu(r); break;				\
 28	case 2: v = le16_to_cpu(r); break;				\
 29	case 1:	v = r; break;						\
 30	default:							\
 31		BUILD_BUG(); break;					\
 32	}								\
 33	v;								\
 34})
 35
 36/* Verify next sizeof(t) bytes can be on the same instruction */
 37#define validate_next(t, insn, n)	\
 38	((insn)->next_byte + sizeof(t) + n <= (insn)->end_kaddr)
 39
 40#define __get_next(t, insn)	\
 41	({ t r = get_unaligned((t *)(insn)->next_byte); (insn)->next_byte += sizeof(t); leXX_to_cpu(t, r); })
 42
 43#define __peek_nbyte_next(t, insn, n)	\
 44	({ t r = get_unaligned((t *)(insn)->next_byte + n); leXX_to_cpu(t, r); })
 45
 46#define get_next(t, insn)	\
 47	({ if (unlikely(!validate_next(t, insn, 0))) goto err_out; __get_next(t, insn); })
 48
 49#define peek_nbyte_next(t, insn, n)	\
 50	({ if (unlikely(!validate_next(t, insn, n))) goto err_out; __peek_nbyte_next(t, insn, n); })
 51
 52#define peek_next(t, insn)	peek_nbyte_next(t, insn, 0)
 53
 54/**
 55 * insn_init() - initialize struct insn
 56 * @insn:	&struct insn to be initialized
 57 * @kaddr:	address (in kernel memory) of instruction (or copy thereof)
 58 * @buf_len:	length of the insn buffer at @kaddr
 59 * @x86_64:	!0 for 64-bit kernel or 64-bit app
 60 */
 61void insn_init(struct insn *insn, const void *kaddr, int buf_len, int x86_64)
 62{
 63	/*
 64	 * Instructions longer than MAX_INSN_SIZE (15 bytes) are invalid
 65	 * even if the input buffer is long enough to hold them.
 66	 */
 67	if (buf_len > MAX_INSN_SIZE)
 68		buf_len = MAX_INSN_SIZE;
 69
 70	memset(insn, 0, sizeof(*insn));
 71	insn->kaddr = kaddr;
 72	insn->end_kaddr = kaddr + buf_len;
 73	insn->next_byte = kaddr;
 74	insn->x86_64 = x86_64 ? 1 : 0;
 75	insn->opnd_bytes = 4;
 76	if (x86_64)
 77		insn->addr_bytes = 8;
 78	else
 79		insn->addr_bytes = 4;
 80}
 81
 82static const insn_byte_t xen_prefix[] = { __XEN_EMULATE_PREFIX };
 83static const insn_byte_t kvm_prefix[] = { __KVM_EMULATE_PREFIX };
 84
 85static int __insn_get_emulate_prefix(struct insn *insn,
 86				     const insn_byte_t *prefix, size_t len)
 87{
 88	size_t i;
 89
 90	for (i = 0; i < len; i++) {
 91		if (peek_nbyte_next(insn_byte_t, insn, i) != prefix[i])
 92			goto err_out;
 93	}
 94
 95	insn->emulate_prefix_size = len;
 96	insn->next_byte += len;
 97
 98	return 1;
 99
100err_out:
101	return 0;
102}
103
104static void insn_get_emulate_prefix(struct insn *insn)
105{
106	if (__insn_get_emulate_prefix(insn, xen_prefix, sizeof(xen_prefix)))
107		return;
108
109	__insn_get_emulate_prefix(insn, kvm_prefix, sizeof(kvm_prefix));
110}
111
112/**
113 * insn_get_prefixes - scan x86 instruction prefix bytes
114 * @insn:	&struct insn containing instruction
115 *
116 * Populates the @insn->prefixes bitmap, and updates @insn->next_byte
117 * to point to the (first) opcode.  No effect if @insn->prefixes.got
118 * is already set.
119 *
120 * * Returns:
121 * 0:  on success
122 * < 0: on error
123 */
124int insn_get_prefixes(struct insn *insn)
125{
126	struct insn_field *prefixes = &insn->prefixes;
127	insn_attr_t attr;
128	insn_byte_t b, lb;
129	int i, nb;
130
131	if (prefixes->got)
132		return 0;
133
134	insn_get_emulate_prefix(insn);
135
136	nb = 0;
137	lb = 0;
138	b = peek_next(insn_byte_t, insn);
139	attr = inat_get_opcode_attribute(b);
140	while (inat_is_legacy_prefix(attr)) {
141		/* Skip if same prefix */
142		for (i = 0; i < nb; i++)
143			if (prefixes->bytes[i] == b)
144				goto found;
145		if (nb == 4)
146			/* Invalid instruction */
147			break;
148		prefixes->bytes[nb++] = b;
149		if (inat_is_address_size_prefix(attr)) {
150			/* address size switches 2/4 or 4/8 */
151			if (insn->x86_64)
152				insn->addr_bytes ^= 12;
153			else
154				insn->addr_bytes ^= 6;
155		} else if (inat_is_operand_size_prefix(attr)) {
156			/* oprand size switches 2/4 */
157			insn->opnd_bytes ^= 6;
158		}
159found:
160		prefixes->nbytes++;
161		insn->next_byte++;
162		lb = b;
163		b = peek_next(insn_byte_t, insn);
164		attr = inat_get_opcode_attribute(b);
165	}
166	/* Set the last prefix */
167	if (lb && lb != insn->prefixes.bytes[3]) {
168		if (unlikely(insn->prefixes.bytes[3])) {
169			/* Swap the last prefix */
170			b = insn->prefixes.bytes[3];
171			for (i = 0; i < nb; i++)
172				if (prefixes->bytes[i] == lb)
173					insn_set_byte(prefixes, i, b);
174		}
175		insn_set_byte(&insn->prefixes, 3, lb);
176	}
177
178	/* Decode REX prefix */
179	if (insn->x86_64) {
180		b = peek_next(insn_byte_t, insn);
181		attr = inat_get_opcode_attribute(b);
182		if (inat_is_rex_prefix(attr)) {
183			insn_field_set(&insn->rex_prefix, b, 1);
 
184			insn->next_byte++;
185			if (X86_REX_W(b))
186				/* REX.W overrides opnd_size */
187				insn->opnd_bytes = 8;
188		}
189	}
190	insn->rex_prefix.got = 1;
191
192	/* Decode VEX prefix */
193	b = peek_next(insn_byte_t, insn);
194	attr = inat_get_opcode_attribute(b);
195	if (inat_is_vex_prefix(attr)) {
196		insn_byte_t b2 = peek_nbyte_next(insn_byte_t, insn, 1);
197		if (!insn->x86_64) {
198			/*
199			 * In 32-bits mode, if the [7:6] bits (mod bits of
200			 * ModRM) on the second byte are not 11b, it is
201			 * LDS or LES or BOUND.
202			 */
203			if (X86_MODRM_MOD(b2) != 3)
204				goto vex_end;
205		}
206		insn_set_byte(&insn->vex_prefix, 0, b);
207		insn_set_byte(&insn->vex_prefix, 1, b2);
208		if (inat_is_evex_prefix(attr)) {
209			b2 = peek_nbyte_next(insn_byte_t, insn, 2);
210			insn_set_byte(&insn->vex_prefix, 2, b2);
211			b2 = peek_nbyte_next(insn_byte_t, insn, 3);
212			insn_set_byte(&insn->vex_prefix, 3, b2);
213			insn->vex_prefix.nbytes = 4;
214			insn->next_byte += 4;
215			if (insn->x86_64 && X86_VEX_W(b2))
216				/* VEX.W overrides opnd_size */
217				insn->opnd_bytes = 8;
218		} else if (inat_is_vex3_prefix(attr)) {
219			b2 = peek_nbyte_next(insn_byte_t, insn, 2);
220			insn_set_byte(&insn->vex_prefix, 2, b2);
221			insn->vex_prefix.nbytes = 3;
222			insn->next_byte += 3;
223			if (insn->x86_64 && X86_VEX_W(b2))
224				/* VEX.W overrides opnd_size */
225				insn->opnd_bytes = 8;
226		} else {
227			/*
228			 * For VEX2, fake VEX3-like byte#2.
229			 * Makes it easier to decode vex.W, vex.vvvv,
230			 * vex.L and vex.pp. Masking with 0x7f sets vex.W == 0.
231			 */
232			insn_set_byte(&insn->vex_prefix, 2, b2 & 0x7f);
233			insn->vex_prefix.nbytes = 2;
234			insn->next_byte += 2;
235		}
236	}
237vex_end:
238	insn->vex_prefix.got = 1;
239
240	prefixes->got = 1;
241
242	return 0;
243
244err_out:
245	return -ENODATA;
246}
247
248/**
249 * insn_get_opcode - collect opcode(s)
250 * @insn:	&struct insn containing instruction
251 *
252 * Populates @insn->opcode, updates @insn->next_byte to point past the
253 * opcode byte(s), and set @insn->attr (except for groups).
254 * If necessary, first collects any preceding (prefix) bytes.
255 * Sets @insn->opcode.value = opcode1.  No effect if @insn->opcode.got
256 * is already 1.
257 *
258 * Returns:
259 * 0:  on success
260 * < 0: on error
261 */
262int insn_get_opcode(struct insn *insn)
263{
264	struct insn_field *opcode = &insn->opcode;
265	int pfx_id, ret;
266	insn_byte_t op;
267
268	if (opcode->got)
269		return 0;
270
271	if (!insn->prefixes.got) {
272		ret = insn_get_prefixes(insn);
273		if (ret)
274			return ret;
275	}
276
277	/* Get first opcode */
278	op = get_next(insn_byte_t, insn);
279	insn_set_byte(opcode, 0, op);
280	opcode->nbytes = 1;
281
282	/* Check if there is VEX prefix or not */
283	if (insn_is_avx(insn)) {
284		insn_byte_t m, p;
285		m = insn_vex_m_bits(insn);
286		p = insn_vex_p_bits(insn);
287		insn->attr = inat_get_avx_attribute(op, m, p);
288		if ((inat_must_evex(insn->attr) && !insn_is_evex(insn)) ||
289		    (!inat_accept_vex(insn->attr) &&
290		     !inat_is_group(insn->attr))) {
291			/* This instruction is bad */
292			insn->attr = 0;
293			return -EINVAL;
294		}
295		/* VEX has only 1 byte for opcode */
296		goto end;
297	}
298
299	insn->attr = inat_get_opcode_attribute(op);
300	while (inat_is_escape(insn->attr)) {
301		/* Get escaped opcode */
302		op = get_next(insn_byte_t, insn);
303		opcode->bytes[opcode->nbytes++] = op;
304		pfx_id = insn_last_prefix_id(insn);
305		insn->attr = inat_get_escape_attribute(op, pfx_id, insn->attr);
306	}
307
308	if (inat_must_vex(insn->attr)) {
309		/* This instruction is bad */
310		insn->attr = 0;
311		return -EINVAL;
312	}
313end:
314	opcode->got = 1;
315	return 0;
316
317err_out:
318	return -ENODATA;
319}
320
321/**
322 * insn_get_modrm - collect ModRM byte, if any
323 * @insn:	&struct insn containing instruction
324 *
325 * Populates @insn->modrm and updates @insn->next_byte to point past the
326 * ModRM byte, if any.  If necessary, first collects the preceding bytes
327 * (prefixes and opcode(s)).  No effect if @insn->modrm.got is already 1.
328 *
329 * Returns:
330 * 0:  on success
331 * < 0: on error
332 */
333int insn_get_modrm(struct insn *insn)
334{
335	struct insn_field *modrm = &insn->modrm;
336	insn_byte_t pfx_id, mod;
337	int ret;
338
339	if (modrm->got)
340		return 0;
341
342	if (!insn->opcode.got) {
343		ret = insn_get_opcode(insn);
344		if (ret)
345			return ret;
346	}
347
348	if (inat_has_modrm(insn->attr)) {
349		mod = get_next(insn_byte_t, insn);
350		insn_field_set(modrm, mod, 1);
 
351		if (inat_is_group(insn->attr)) {
352			pfx_id = insn_last_prefix_id(insn);
353			insn->attr = inat_get_group_attribute(mod, pfx_id,
354							      insn->attr);
355			if (insn_is_avx(insn) && !inat_accept_vex(insn->attr)) {
356				/* Bad insn */
357				insn->attr = 0;
358				return -EINVAL;
359			}
360		}
361	}
362
363	if (insn->x86_64 && inat_is_force64(insn->attr))
364		insn->opnd_bytes = 8;
365
366	modrm->got = 1;
367	return 0;
368
369err_out:
370	return -ENODATA;
371}
372
373
374/**
375 * insn_rip_relative() - Does instruction use RIP-relative addressing mode?
376 * @insn:	&struct insn containing instruction
377 *
378 * If necessary, first collects the instruction up to and including the
379 * ModRM byte.  No effect if @insn->x86_64 is 0.
380 */
381int insn_rip_relative(struct insn *insn)
382{
383	struct insn_field *modrm = &insn->modrm;
384	int ret;
385
386	if (!insn->x86_64)
387		return 0;
388
389	if (!modrm->got) {
390		ret = insn_get_modrm(insn);
391		if (ret)
392			return 0;
393	}
394	/*
395	 * For rip-relative instructions, the mod field (top 2 bits)
396	 * is zero and the r/m field (bottom 3 bits) is 0x5.
397	 */
398	return (modrm->nbytes && (modrm->bytes[0] & 0xc7) == 0x5);
399}
400
401/**
402 * insn_get_sib() - Get the SIB byte of instruction
403 * @insn:	&struct insn containing instruction
404 *
405 * If necessary, first collects the instruction up to and including the
406 * ModRM byte.
407 *
408 * Returns:
409 * 0: if decoding succeeded
410 * < 0: otherwise.
411 */
412int insn_get_sib(struct insn *insn)
413{
414	insn_byte_t modrm;
415	int ret;
416
417	if (insn->sib.got)
418		return 0;
419
420	if (!insn->modrm.got) {
421		ret = insn_get_modrm(insn);
422		if (ret)
423			return ret;
424	}
425
426	if (insn->modrm.nbytes) {
427		modrm = insn->modrm.bytes[0];
428		if (insn->addr_bytes != 2 &&
429		    X86_MODRM_MOD(modrm) != 3 && X86_MODRM_RM(modrm) == 4) {
430			insn_field_set(&insn->sib,
431				       get_next(insn_byte_t, insn), 1);
432		}
433	}
434	insn->sib.got = 1;
435
436	return 0;
437
438err_out:
439	return -ENODATA;
440}
441
442
443/**
444 * insn_get_displacement() - Get the displacement of instruction
445 * @insn:	&struct insn containing instruction
446 *
447 * If necessary, first collects the instruction up to and including the
448 * SIB byte.
449 * Displacement value is sign-expanded.
450 *
451 * * Returns:
452 * 0: if decoding succeeded
453 * < 0: otherwise.
454 */
455int insn_get_displacement(struct insn *insn)
456{
457	insn_byte_t mod, rm, base;
458	int ret;
459
460	if (insn->displacement.got)
461		return 0;
462
463	if (!insn->sib.got) {
464		ret = insn_get_sib(insn);
465		if (ret)
466			return ret;
467	}
468
469	if (insn->modrm.nbytes) {
470		/*
471		 * Interpreting the modrm byte:
472		 * mod = 00 - no displacement fields (exceptions below)
473		 * mod = 01 - 1-byte displacement field
474		 * mod = 10 - displacement field is 4 bytes, or 2 bytes if
475		 * 	address size = 2 (0x67 prefix in 32-bit mode)
476		 * mod = 11 - no memory operand
477		 *
478		 * If address size = 2...
479		 * mod = 00, r/m = 110 - displacement field is 2 bytes
480		 *
481		 * If address size != 2...
482		 * mod != 11, r/m = 100 - SIB byte exists
483		 * mod = 00, SIB base = 101 - displacement field is 4 bytes
484		 * mod = 00, r/m = 101 - rip-relative addressing, displacement
485		 * 	field is 4 bytes
486		 */
487		mod = X86_MODRM_MOD(insn->modrm.value);
488		rm = X86_MODRM_RM(insn->modrm.value);
489		base = X86_SIB_BASE(insn->sib.value);
490		if (mod == 3)
491			goto out;
492		if (mod == 1) {
493			insn_field_set(&insn->displacement,
494				       get_next(signed char, insn), 1);
495		} else if (insn->addr_bytes == 2) {
496			if ((mod == 0 && rm == 6) || mod == 2) {
497				insn_field_set(&insn->displacement,
498					       get_next(short, insn), 2);
 
499			}
500		} else {
501			if ((mod == 0 && rm == 5) || mod == 2 ||
502			    (mod == 0 && base == 5)) {
503				insn_field_set(&insn->displacement,
504					       get_next(int, insn), 4);
505			}
506		}
507	}
508out:
509	insn->displacement.got = 1;
510	return 0;
511
512err_out:
513	return -ENODATA;
514}
515
516/* Decode moffset16/32/64. Return 0 if failed */
517static int __get_moffset(struct insn *insn)
518{
519	switch (insn->addr_bytes) {
520	case 2:
521		insn_field_set(&insn->moffset1, get_next(short, insn), 2);
 
522		break;
523	case 4:
524		insn_field_set(&insn->moffset1, get_next(int, insn), 4);
 
525		break;
526	case 8:
527		insn_field_set(&insn->moffset1, get_next(int, insn), 4);
528		insn_field_set(&insn->moffset2, get_next(int, insn), 4);
 
 
529		break;
530	default:	/* opnd_bytes must be modified manually */
531		goto err_out;
532	}
533	insn->moffset1.got = insn->moffset2.got = 1;
534
535	return 1;
536
537err_out:
538	return 0;
539}
540
541/* Decode imm v32(Iz). Return 0 if failed */
542static int __get_immv32(struct insn *insn)
543{
544	switch (insn->opnd_bytes) {
545	case 2:
546		insn_field_set(&insn->immediate, get_next(short, insn), 2);
 
547		break;
548	case 4:
549	case 8:
550		insn_field_set(&insn->immediate, get_next(int, insn), 4);
 
551		break;
552	default:	/* opnd_bytes must be modified manually */
553		goto err_out;
554	}
555
556	return 1;
557
558err_out:
559	return 0;
560}
561
562/* Decode imm v64(Iv/Ov), Return 0 if failed */
563static int __get_immv(struct insn *insn)
564{
565	switch (insn->opnd_bytes) {
566	case 2:
567		insn_field_set(&insn->immediate1, get_next(short, insn), 2);
 
568		break;
569	case 4:
570		insn_field_set(&insn->immediate1, get_next(int, insn), 4);
571		insn->immediate1.nbytes = 4;
572		break;
573	case 8:
574		insn_field_set(&insn->immediate1, get_next(int, insn), 4);
575		insn_field_set(&insn->immediate2, get_next(int, insn), 4);
 
 
576		break;
577	default:	/* opnd_bytes must be modified manually */
578		goto err_out;
579	}
580	insn->immediate1.got = insn->immediate2.got = 1;
581
582	return 1;
583err_out:
584	return 0;
585}
586
587/* Decode ptr16:16/32(Ap) */
588static int __get_immptr(struct insn *insn)
589{
590	switch (insn->opnd_bytes) {
591	case 2:
592		insn_field_set(&insn->immediate1, get_next(short, insn), 2);
 
593		break;
594	case 4:
595		insn_field_set(&insn->immediate1, get_next(int, insn), 4);
 
596		break;
597	case 8:
598		/* ptr16:64 is not exist (no segment) */
599		return 0;
600	default:	/* opnd_bytes must be modified manually */
601		goto err_out;
602	}
603	insn_field_set(&insn->immediate2, get_next(unsigned short, insn), 2);
 
604	insn->immediate1.got = insn->immediate2.got = 1;
605
606	return 1;
607err_out:
608	return 0;
609}
610
611/**
612 * insn_get_immediate() - Get the immediate in an instruction
613 * @insn:	&struct insn containing instruction
614 *
615 * If necessary, first collects the instruction up to and including the
616 * displacement bytes.
617 * Basically, most of immediates are sign-expanded. Unsigned-value can be
618 * computed by bit masking with ((1 << (nbytes * 8)) - 1)
619 *
620 * Returns:
621 * 0:  on success
622 * < 0: on error
623 */
624int insn_get_immediate(struct insn *insn)
625{
626	int ret;
627
628	if (insn->immediate.got)
629		return 0;
630
631	if (!insn->displacement.got) {
632		ret = insn_get_displacement(insn);
633		if (ret)
634			return ret;
635	}
636
637	if (inat_has_moffset(insn->attr)) {
638		if (!__get_moffset(insn))
639			goto err_out;
640		goto done;
641	}
642
643	if (!inat_has_immediate(insn->attr))
644		/* no immediates */
645		goto done;
646
647	switch (inat_immediate_size(insn->attr)) {
648	case INAT_IMM_BYTE:
649		insn_field_set(&insn->immediate, get_next(signed char, insn), 1);
 
650		break;
651	case INAT_IMM_WORD:
652		insn_field_set(&insn->immediate, get_next(short, insn), 2);
 
653		break;
654	case INAT_IMM_DWORD:
655		insn_field_set(&insn->immediate, get_next(int, insn), 4);
 
656		break;
657	case INAT_IMM_QWORD:
658		insn_field_set(&insn->immediate1, get_next(int, insn), 4);
659		insn_field_set(&insn->immediate2, get_next(int, insn), 4);
 
 
660		break;
661	case INAT_IMM_PTR:
662		if (!__get_immptr(insn))
663			goto err_out;
664		break;
665	case INAT_IMM_VWORD32:
666		if (!__get_immv32(insn))
667			goto err_out;
668		break;
669	case INAT_IMM_VWORD:
670		if (!__get_immv(insn))
671			goto err_out;
672		break;
673	default:
674		/* Here, insn must have an immediate, but failed */
675		goto err_out;
676	}
677	if (inat_has_second_immediate(insn->attr)) {
678		insn_field_set(&insn->immediate2, get_next(signed char, insn), 1);
 
679	}
680done:
681	insn->immediate.got = 1;
682	return 0;
683
684err_out:
685	return -ENODATA;
686}
687
688/**
689 * insn_get_length() - Get the length of instruction
690 * @insn:	&struct insn containing instruction
691 *
692 * If necessary, first collects the instruction up to and including the
693 * immediates bytes.
694 *
695 * Returns:
696 *  - 0 on success
697 *  - < 0 on error
698*/
699int insn_get_length(struct insn *insn)
700{
701	int ret;
702
703	if (insn->length)
704		return 0;
705
706	if (!insn->immediate.got) {
707		ret = insn_get_immediate(insn);
708		if (ret)
709			return ret;
710	}
711
712	insn->length = (unsigned char)((unsigned long)insn->next_byte
713				     - (unsigned long)insn->kaddr);
714
715	return 0;
716}
717
718/* Ensure this instruction is decoded completely */
719static inline int insn_complete(struct insn *insn)
720{
721	return insn->opcode.got && insn->modrm.got && insn->sib.got &&
722		insn->displacement.got && insn->immediate.got;
723}
724
725/**
726 * insn_decode() - Decode an x86 instruction
727 * @insn:	&struct insn to be initialized
728 * @kaddr:	address (in kernel memory) of instruction (or copy thereof)
729 * @buf_len:	length of the insn buffer at @kaddr
730 * @m:		insn mode, see enum insn_mode
731 *
732 * Returns:
733 * 0: if decoding succeeded
734 * < 0: otherwise.
735 */
736int insn_decode(struct insn *insn, const void *kaddr, int buf_len, enum insn_mode m)
737{
738	int ret;
739
740/* #define INSN_MODE_KERN	-1 __ignore_sync_check__ mode is only valid in the kernel */
741
742	if (m == INSN_MODE_KERN)
743		insn_init(insn, kaddr, buf_len, IS_ENABLED(CONFIG_X86_64));
744	else
745		insn_init(insn, kaddr, buf_len, m == INSN_MODE_64);
746
747	ret = insn_get_length(insn);
748	if (ret)
749		return ret;
750
751	if (insn_complete(insn))
752		return 0;
753
754	return -EINVAL;
755}
v3.5.6
 
  1/*
  2 * x86 instruction analysis
  3 *
  4 * This program is free software; you can redistribute it and/or modify
  5 * it under the terms of the GNU General Public License as published by
  6 * the Free Software Foundation; either version 2 of the License, or
  7 * (at your option) any later version.
  8 *
  9 * This program is distributed in the hope that it will be useful,
 10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 12 * GNU General Public License for more details.
 13 *
 14 * You should have received a copy of the GNU General Public License
 15 * along with this program; if not, write to the Free Software
 16 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
 17 *
 18 * Copyright (C) IBM Corporation, 2002, 2004, 2009
 19 */
 20
 
 
 21#include <linux/string.h>
 22#include <asm/inat.h>
 23#include <asm/insn.h>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 24
 25/* Verify next sizeof(t) bytes can be on the same instruction */
 26#define validate_next(t, insn, n)	\
 27	((insn)->next_byte + sizeof(t) + n - (insn)->kaddr <= MAX_INSN_SIZE)
 28
 29#define __get_next(t, insn)	\
 30	({ t r = *(t*)insn->next_byte; insn->next_byte += sizeof(t); r; })
 31
 32#define __peek_nbyte_next(t, insn, n)	\
 33	({ t r = *(t*)((insn)->next_byte + n); r; })
 34
 35#define get_next(t, insn)	\
 36	({ if (unlikely(!validate_next(t, insn, 0))) goto err_out; __get_next(t, insn); })
 37
 38#define peek_nbyte_next(t, insn, n)	\
 39	({ if (unlikely(!validate_next(t, insn, n))) goto err_out; __peek_nbyte_next(t, insn, n); })
 40
 41#define peek_next(t, insn)	peek_nbyte_next(t, insn, 0)
 42
 43/**
 44 * insn_init() - initialize struct insn
 45 * @insn:	&struct insn to be initialized
 46 * @kaddr:	address (in kernel memory) of instruction (or copy thereof)
 
 47 * @x86_64:	!0 for 64-bit kernel or 64-bit app
 48 */
 49void insn_init(struct insn *insn, const void *kaddr, int x86_64)
 50{
 
 
 
 
 
 
 
 51	memset(insn, 0, sizeof(*insn));
 52	insn->kaddr = kaddr;
 
 53	insn->next_byte = kaddr;
 54	insn->x86_64 = x86_64 ? 1 : 0;
 55	insn->opnd_bytes = 4;
 56	if (x86_64)
 57		insn->addr_bytes = 8;
 58	else
 59		insn->addr_bytes = 4;
 60}
 61
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 62/**
 63 * insn_get_prefixes - scan x86 instruction prefix bytes
 64 * @insn:	&struct insn containing instruction
 65 *
 66 * Populates the @insn->prefixes bitmap, and updates @insn->next_byte
 67 * to point to the (first) opcode.  No effect if @insn->prefixes.got
 68 * is already set.
 
 
 
 
 69 */
 70void insn_get_prefixes(struct insn *insn)
 71{
 72	struct insn_field *prefixes = &insn->prefixes;
 73	insn_attr_t attr;
 74	insn_byte_t b, lb;
 75	int i, nb;
 76
 77	if (prefixes->got)
 78		return;
 
 
 79
 80	nb = 0;
 81	lb = 0;
 82	b = peek_next(insn_byte_t, insn);
 83	attr = inat_get_opcode_attribute(b);
 84	while (inat_is_legacy_prefix(attr)) {
 85		/* Skip if same prefix */
 86		for (i = 0; i < nb; i++)
 87			if (prefixes->bytes[i] == b)
 88				goto found;
 89		if (nb == 4)
 90			/* Invalid instruction */
 91			break;
 92		prefixes->bytes[nb++] = b;
 93		if (inat_is_address_size_prefix(attr)) {
 94			/* address size switches 2/4 or 4/8 */
 95			if (insn->x86_64)
 96				insn->addr_bytes ^= 12;
 97			else
 98				insn->addr_bytes ^= 6;
 99		} else if (inat_is_operand_size_prefix(attr)) {
100			/* oprand size switches 2/4 */
101			insn->opnd_bytes ^= 6;
102		}
103found:
104		prefixes->nbytes++;
105		insn->next_byte++;
106		lb = b;
107		b = peek_next(insn_byte_t, insn);
108		attr = inat_get_opcode_attribute(b);
109	}
110	/* Set the last prefix */
111	if (lb && lb != insn->prefixes.bytes[3]) {
112		if (unlikely(insn->prefixes.bytes[3])) {
113			/* Swap the last prefix */
114			b = insn->prefixes.bytes[3];
115			for (i = 0; i < nb; i++)
116				if (prefixes->bytes[i] == lb)
117					prefixes->bytes[i] = b;
118		}
119		insn->prefixes.bytes[3] = lb;
120	}
121
122	/* Decode REX prefix */
123	if (insn->x86_64) {
124		b = peek_next(insn_byte_t, insn);
125		attr = inat_get_opcode_attribute(b);
126		if (inat_is_rex_prefix(attr)) {
127			insn->rex_prefix.value = b;
128			insn->rex_prefix.nbytes = 1;
129			insn->next_byte++;
130			if (X86_REX_W(b))
131				/* REX.W overrides opnd_size */
132				insn->opnd_bytes = 8;
133		}
134	}
135	insn->rex_prefix.got = 1;
136
137	/* Decode VEX prefix */
138	b = peek_next(insn_byte_t, insn);
139	attr = inat_get_opcode_attribute(b);
140	if (inat_is_vex_prefix(attr)) {
141		insn_byte_t b2 = peek_nbyte_next(insn_byte_t, insn, 1);
142		if (!insn->x86_64) {
143			/*
144			 * In 32-bits mode, if the [7:6] bits (mod bits of
145			 * ModRM) on the second byte are not 11b, it is
146			 * LDS or LES.
147			 */
148			if (X86_MODRM_MOD(b2) != 3)
149				goto vex_end;
150		}
151		insn->vex_prefix.bytes[0] = b;
152		insn->vex_prefix.bytes[1] = b2;
153		if (inat_is_vex3_prefix(attr)) {
 
 
 
 
 
 
 
 
 
 
154			b2 = peek_nbyte_next(insn_byte_t, insn, 2);
155			insn->vex_prefix.bytes[2] = b2;
156			insn->vex_prefix.nbytes = 3;
157			insn->next_byte += 3;
158			if (insn->x86_64 && X86_VEX_W(b2))
159				/* VEX.W overrides opnd_size */
160				insn->opnd_bytes = 8;
161		} else {
 
 
 
 
 
 
162			insn->vex_prefix.nbytes = 2;
163			insn->next_byte += 2;
164		}
165	}
166vex_end:
167	insn->vex_prefix.got = 1;
168
169	prefixes->got = 1;
170
 
 
171err_out:
172	return;
173}
174
175/**
176 * insn_get_opcode - collect opcode(s)
177 * @insn:	&struct insn containing instruction
178 *
179 * Populates @insn->opcode, updates @insn->next_byte to point past the
180 * opcode byte(s), and set @insn->attr (except for groups).
181 * If necessary, first collects any preceding (prefix) bytes.
182 * Sets @insn->opcode.value = opcode1.  No effect if @insn->opcode.got
183 * is already 1.
 
 
 
 
184 */
185void insn_get_opcode(struct insn *insn)
186{
187	struct insn_field *opcode = &insn->opcode;
 
188	insn_byte_t op;
189	int pfx_id;
190	if (opcode->got)
191		return;
192	if (!insn->prefixes.got)
193		insn_get_prefixes(insn);
 
 
 
 
194
195	/* Get first opcode */
196	op = get_next(insn_byte_t, insn);
197	opcode->bytes[0] = op;
198	opcode->nbytes = 1;
199
200	/* Check if there is VEX prefix or not */
201	if (insn_is_avx(insn)) {
202		insn_byte_t m, p;
203		m = insn_vex_m_bits(insn);
204		p = insn_vex_p_bits(insn);
205		insn->attr = inat_get_avx_attribute(op, m, p);
206		if (!inat_accept_vex(insn->attr) && !inat_is_group(insn->attr))
207			insn->attr = 0;	/* This instruction is bad */
208		goto end;	/* VEX has only 1 byte for opcode */
 
 
 
 
 
 
209	}
210
211	insn->attr = inat_get_opcode_attribute(op);
212	while (inat_is_escape(insn->attr)) {
213		/* Get escaped opcode */
214		op = get_next(insn_byte_t, insn);
215		opcode->bytes[opcode->nbytes++] = op;
216		pfx_id = insn_last_prefix_id(insn);
217		insn->attr = inat_get_escape_attribute(op, pfx_id, insn->attr);
218	}
219	if (inat_must_vex(insn->attr))
220		insn->attr = 0;	/* This instruction is bad */
 
 
 
 
221end:
222	opcode->got = 1;
 
223
224err_out:
225	return;
226}
227
228/**
229 * insn_get_modrm - collect ModRM byte, if any
230 * @insn:	&struct insn containing instruction
231 *
232 * Populates @insn->modrm and updates @insn->next_byte to point past the
233 * ModRM byte, if any.  If necessary, first collects the preceding bytes
234 * (prefixes and opcode(s)).  No effect if @insn->modrm.got is already 1.
 
 
 
 
235 */
236void insn_get_modrm(struct insn *insn)
237{
238	struct insn_field *modrm = &insn->modrm;
239	insn_byte_t pfx_id, mod;
 
 
240	if (modrm->got)
241		return;
242	if (!insn->opcode.got)
243		insn_get_opcode(insn);
 
 
 
 
244
245	if (inat_has_modrm(insn->attr)) {
246		mod = get_next(insn_byte_t, insn);
247		modrm->value = mod;
248		modrm->nbytes = 1;
249		if (inat_is_group(insn->attr)) {
250			pfx_id = insn_last_prefix_id(insn);
251			insn->attr = inat_get_group_attribute(mod, pfx_id,
252							      insn->attr);
253			if (insn_is_avx(insn) && !inat_accept_vex(insn->attr))
254				insn->attr = 0;	/* This is bad */
 
 
 
255		}
256	}
257
258	if (insn->x86_64 && inat_is_force64(insn->attr))
259		insn->opnd_bytes = 8;
 
260	modrm->got = 1;
 
261
262err_out:
263	return;
264}
265
266
267/**
268 * insn_rip_relative() - Does instruction use RIP-relative addressing mode?
269 * @insn:	&struct insn containing instruction
270 *
271 * If necessary, first collects the instruction up to and including the
272 * ModRM byte.  No effect if @insn->x86_64 is 0.
273 */
274int insn_rip_relative(struct insn *insn)
275{
276	struct insn_field *modrm = &insn->modrm;
 
277
278	if (!insn->x86_64)
279		return 0;
280	if (!modrm->got)
281		insn_get_modrm(insn);
 
 
 
 
282	/*
283	 * For rip-relative instructions, the mod field (top 2 bits)
284	 * is zero and the r/m field (bottom 3 bits) is 0x5.
285	 */
286	return (modrm->nbytes && (modrm->value & 0xc7) == 0x5);
287}
288
289/**
290 * insn_get_sib() - Get the SIB byte of instruction
291 * @insn:	&struct insn containing instruction
292 *
293 * If necessary, first collects the instruction up to and including the
294 * ModRM byte.
 
 
 
 
295 */
296void insn_get_sib(struct insn *insn)
297{
298	insn_byte_t modrm;
 
299
300	if (insn->sib.got)
301		return;
302	if (!insn->modrm.got)
303		insn_get_modrm(insn);
 
 
 
 
 
304	if (insn->modrm.nbytes) {
305		modrm = (insn_byte_t)insn->modrm.value;
306		if (insn->addr_bytes != 2 &&
307		    X86_MODRM_MOD(modrm) != 3 && X86_MODRM_RM(modrm) == 4) {
308			insn->sib.value = get_next(insn_byte_t, insn);
309			insn->sib.nbytes = 1;
310		}
311	}
312	insn->sib.got = 1;
313
 
 
314err_out:
315	return;
316}
317
318
319/**
320 * insn_get_displacement() - Get the displacement of instruction
321 * @insn:	&struct insn containing instruction
322 *
323 * If necessary, first collects the instruction up to and including the
324 * SIB byte.
325 * Displacement value is sign-expanded.
 
 
 
 
326 */
327void insn_get_displacement(struct insn *insn)
328{
329	insn_byte_t mod, rm, base;
 
330
331	if (insn->displacement.got)
332		return;
333	if (!insn->sib.got)
334		insn_get_sib(insn);
 
 
 
 
 
335	if (insn->modrm.nbytes) {
336		/*
337		 * Interpreting the modrm byte:
338		 * mod = 00 - no displacement fields (exceptions below)
339		 * mod = 01 - 1-byte displacement field
340		 * mod = 10 - displacement field is 4 bytes, or 2 bytes if
341		 * 	address size = 2 (0x67 prefix in 32-bit mode)
342		 * mod = 11 - no memory operand
343		 *
344		 * If address size = 2...
345		 * mod = 00, r/m = 110 - displacement field is 2 bytes
346		 *
347		 * If address size != 2...
348		 * mod != 11, r/m = 100 - SIB byte exists
349		 * mod = 00, SIB base = 101 - displacement field is 4 bytes
350		 * mod = 00, r/m = 101 - rip-relative addressing, displacement
351		 * 	field is 4 bytes
352		 */
353		mod = X86_MODRM_MOD(insn->modrm.value);
354		rm = X86_MODRM_RM(insn->modrm.value);
355		base = X86_SIB_BASE(insn->sib.value);
356		if (mod == 3)
357			goto out;
358		if (mod == 1) {
359			insn->displacement.value = get_next(char, insn);
360			insn->displacement.nbytes = 1;
361		} else if (insn->addr_bytes == 2) {
362			if ((mod == 0 && rm == 6) || mod == 2) {
363				insn->displacement.value =
364					 get_next(short, insn);
365				insn->displacement.nbytes = 2;
366			}
367		} else {
368			if ((mod == 0 && rm == 5) || mod == 2 ||
369			    (mod == 0 && base == 5)) {
370				insn->displacement.value = get_next(int, insn);
371				insn->displacement.nbytes = 4;
372			}
373		}
374	}
375out:
376	insn->displacement.got = 1;
 
377
378err_out:
379	return;
380}
381
382/* Decode moffset16/32/64. Return 0 if failed */
383static int __get_moffset(struct insn *insn)
384{
385	switch (insn->addr_bytes) {
386	case 2:
387		insn->moffset1.value = get_next(short, insn);
388		insn->moffset1.nbytes = 2;
389		break;
390	case 4:
391		insn->moffset1.value = get_next(int, insn);
392		insn->moffset1.nbytes = 4;
393		break;
394	case 8:
395		insn->moffset1.value = get_next(int, insn);
396		insn->moffset1.nbytes = 4;
397		insn->moffset2.value = get_next(int, insn);
398		insn->moffset2.nbytes = 4;
399		break;
400	default:	/* opnd_bytes must be modified manually */
401		goto err_out;
402	}
403	insn->moffset1.got = insn->moffset2.got = 1;
404
405	return 1;
406
407err_out:
408	return 0;
409}
410
411/* Decode imm v32(Iz). Return 0 if failed */
412static int __get_immv32(struct insn *insn)
413{
414	switch (insn->opnd_bytes) {
415	case 2:
416		insn->immediate.value = get_next(short, insn);
417		insn->immediate.nbytes = 2;
418		break;
419	case 4:
420	case 8:
421		insn->immediate.value = get_next(int, insn);
422		insn->immediate.nbytes = 4;
423		break;
424	default:	/* opnd_bytes must be modified manually */
425		goto err_out;
426	}
427
428	return 1;
429
430err_out:
431	return 0;
432}
433
434/* Decode imm v64(Iv/Ov), Return 0 if failed */
435static int __get_immv(struct insn *insn)
436{
437	switch (insn->opnd_bytes) {
438	case 2:
439		insn->immediate1.value = get_next(short, insn);
440		insn->immediate1.nbytes = 2;
441		break;
442	case 4:
443		insn->immediate1.value = get_next(int, insn);
444		insn->immediate1.nbytes = 4;
445		break;
446	case 8:
447		insn->immediate1.value = get_next(int, insn);
448		insn->immediate1.nbytes = 4;
449		insn->immediate2.value = get_next(int, insn);
450		insn->immediate2.nbytes = 4;
451		break;
452	default:	/* opnd_bytes must be modified manually */
453		goto err_out;
454	}
455	insn->immediate1.got = insn->immediate2.got = 1;
456
457	return 1;
458err_out:
459	return 0;
460}
461
462/* Decode ptr16:16/32(Ap) */
463static int __get_immptr(struct insn *insn)
464{
465	switch (insn->opnd_bytes) {
466	case 2:
467		insn->immediate1.value = get_next(short, insn);
468		insn->immediate1.nbytes = 2;
469		break;
470	case 4:
471		insn->immediate1.value = get_next(int, insn);
472		insn->immediate1.nbytes = 4;
473		break;
474	case 8:
475		/* ptr16:64 is not exist (no segment) */
476		return 0;
477	default:	/* opnd_bytes must be modified manually */
478		goto err_out;
479	}
480	insn->immediate2.value = get_next(unsigned short, insn);
481	insn->immediate2.nbytes = 2;
482	insn->immediate1.got = insn->immediate2.got = 1;
483
484	return 1;
485err_out:
486	return 0;
487}
488
489/**
490 * insn_get_immediate() - Get the immediates of instruction
491 * @insn:	&struct insn containing instruction
492 *
493 * If necessary, first collects the instruction up to and including the
494 * displacement bytes.
495 * Basically, most of immediates are sign-expanded. Unsigned-value can be
496 * get by bit masking with ((1 << (nbytes * 8)) - 1)
 
 
 
 
497 */
498void insn_get_immediate(struct insn *insn)
499{
 
 
500	if (insn->immediate.got)
501		return;
502	if (!insn->displacement.got)
503		insn_get_displacement(insn);
 
 
 
 
504
505	if (inat_has_moffset(insn->attr)) {
506		if (!__get_moffset(insn))
507			goto err_out;
508		goto done;
509	}
510
511	if (!inat_has_immediate(insn->attr))
512		/* no immediates */
513		goto done;
514
515	switch (inat_immediate_size(insn->attr)) {
516	case INAT_IMM_BYTE:
517		insn->immediate.value = get_next(char, insn);
518		insn->immediate.nbytes = 1;
519		break;
520	case INAT_IMM_WORD:
521		insn->immediate.value = get_next(short, insn);
522		insn->immediate.nbytes = 2;
523		break;
524	case INAT_IMM_DWORD:
525		insn->immediate.value = get_next(int, insn);
526		insn->immediate.nbytes = 4;
527		break;
528	case INAT_IMM_QWORD:
529		insn->immediate1.value = get_next(int, insn);
530		insn->immediate1.nbytes = 4;
531		insn->immediate2.value = get_next(int, insn);
532		insn->immediate2.nbytes = 4;
533		break;
534	case INAT_IMM_PTR:
535		if (!__get_immptr(insn))
536			goto err_out;
537		break;
538	case INAT_IMM_VWORD32:
539		if (!__get_immv32(insn))
540			goto err_out;
541		break;
542	case INAT_IMM_VWORD:
543		if (!__get_immv(insn))
544			goto err_out;
545		break;
546	default:
547		/* Here, insn must have an immediate, but failed */
548		goto err_out;
549	}
550	if (inat_has_second_immediate(insn->attr)) {
551		insn->immediate2.value = get_next(char, insn);
552		insn->immediate2.nbytes = 1;
553	}
554done:
555	insn->immediate.got = 1;
 
556
557err_out:
558	return;
559}
560
561/**
562 * insn_get_length() - Get the length of instruction
563 * @insn:	&struct insn containing instruction
564 *
565 * If necessary, first collects the instruction up to and including the
566 * immediates bytes.
567 */
568void insn_get_length(struct insn *insn)
 
 
 
 
569{
 
 
570	if (insn->length)
571		return;
572	if (!insn->immediate.got)
573		insn_get_immediate(insn);
 
 
 
 
 
574	insn->length = (unsigned char)((unsigned long)insn->next_byte
575				     - (unsigned long)insn->kaddr);
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
576}