disasm.c

   1// SPDX-License-Identifier: GPL-2.0-only
   2#include <ctype.h>
   3#include <errno.h>
   4#include <fcntl.h>
   5#include <inttypes.h>
   6#include <libgen.h>
   7#include <regex.h>
   8#include <stdlib.h>
   9#include <unistd.h>
  10
  11#include <linux/string.h>
  12#include <subcmd/run-command.h>
  13
  14#include "annotate.h"
  15#include "annotate-data.h"
  16#include "build-id.h"
  17#include "capstone.h"
  18#include "debug.h"
  19#include "disasm.h"
  20#include "dso.h"
  21#include "dwarf-regs.h"
  22#include "env.h"
  23#include "evsel.h"
  24#include "libbfd.h"
  25#include "llvm.h"
  26#include "map.h"
  27#include "maps.h"
  28#include "namespaces.h"
  29#include "srcline.h"
  30#include "symbol.h"
  31#include "util.h"
  32
  33static regex_t	 file_lineno;
  34
  35/* These can be referred from the arch-dependent code */
  36static struct ins_ops call_ops;
  37static struct ins_ops dec_ops;
  38static struct ins_ops jump_ops;
  39static struct ins_ops mov_ops;
  40static struct ins_ops nop_ops;
  41static struct ins_ops lock_ops;
  42static struct ins_ops ret_ops;
  43static struct ins_ops load_store_ops;
  44static struct ins_ops arithmetic_ops;
  45
  46static int jump__scnprintf(struct ins *ins, char *bf, size_t size,
  47			   struct ins_operands *ops, int max_ins_name);
  48static int call__scnprintf(struct ins *ins, char *bf, size_t size,
  49			   struct ins_operands *ops, int max_ins_name);
  50
  51static void ins__sort(struct arch *arch);
  52static int disasm_line__parse(char *line, const char **namep, char **rawp);
  53static int disasm_line__parse_powerpc(struct disasm_line *dl, struct annotate_args *args);
  54
  55static __attribute__((constructor)) void symbol__init_regexpr(void)
  56{
  57	regcomp(&file_lineno, "^/[^:]+:([0-9]+)", REG_EXTENDED);
  58}
  59
  60static int arch__grow_instructions(struct arch *arch)
  61{
  62	struct ins *new_instructions;
  63	size_t new_nr_allocated;
  64
  65	if (arch->nr_instructions_allocated == 0 && arch->instructions)
  66		goto grow_from_non_allocated_table;
  67
  68	new_nr_allocated = arch->nr_instructions_allocated + 128;
  69	new_instructions = realloc(arch->instructions, new_nr_allocated * sizeof(struct ins));
  70	if (new_instructions == NULL)
  71		return -1;
  72
  73out_update_instructions:
  74	arch->instructions = new_instructions;
  75	arch->nr_instructions_allocated = new_nr_allocated;
  76	return 0;
  77
  78grow_from_non_allocated_table:
  79	new_nr_allocated = arch->nr_instructions + 128;
  80	new_instructions = calloc(new_nr_allocated, sizeof(struct ins));
  81	if (new_instructions == NULL)
  82		return -1;
  83
  84	memcpy(new_instructions, arch->instructions, arch->nr_instructions * sizeof(struct ins));
  85	goto out_update_instructions;
  86}
  87
  88static int arch__associate_ins_ops(struct arch* arch, const char *name, struct ins_ops *ops)
  89{
  90	struct ins *ins;
  91
  92	if (arch->nr_instructions == arch->nr_instructions_allocated &&
  93	    arch__grow_instructions(arch))
  94		return -1;
  95
  96	ins = &arch->instructions[arch->nr_instructions];
  97	ins->name = strdup(name);
  98	if (!ins->name)
  99		return -1;
 100
 101	ins->ops  = ops;
 102	arch->nr_instructions++;
 103
 104	ins__sort(arch);
 105	return 0;
 106}
 107
 108#include "arch/arc/annotate/instructions.c"
 109#include "arch/arm/annotate/instructions.c"
 110#include "arch/arm64/annotate/instructions.c"
 111#include "arch/csky/annotate/instructions.c"
 112#include "arch/loongarch/annotate/instructions.c"
 113#include "arch/mips/annotate/instructions.c"
 114#include "arch/x86/annotate/instructions.c"
 115#include "arch/powerpc/annotate/instructions.c"
 116#include "arch/riscv64/annotate/instructions.c"
 117#include "arch/s390/annotate/instructions.c"
 118#include "arch/sparc/annotate/instructions.c"
 119
 120static struct arch architectures[] = {
 121	{
 122		.name = "arc",
 123		.init = arc__annotate_init,
 124	},
 125	{
 126		.name = "arm",
 127		.init = arm__annotate_init,
 128	},
 129	{
 130		.name = "arm64",
 131		.init = arm64__annotate_init,
 132	},
 133	{
 134		.name = "csky",
 135		.init = csky__annotate_init,
 136	},
 137	{
 138		.name = "mips",
 139		.init = mips__annotate_init,
 140		.objdump = {
 141			.comment_char = '#',
 142		},
 143	},
 144	{
 145		.name = "x86",
 146		.init = x86__annotate_init,
 147		.instructions = x86__instructions,
 148		.nr_instructions = ARRAY_SIZE(x86__instructions),
 149		.insn_suffix = "bwlq",
 150		.objdump =  {
 151			.comment_char = '#',
 152			.register_char = '%',
 153			.memory_ref_char = '(',
 154			.imm_char = '$',
 155		},
 156#ifdef HAVE_LIBDW_SUPPORT
 157		.update_insn_state = update_insn_state_x86,
 158#endif
 159	},
 160	{
 161		.name = "powerpc",
 162		.init = powerpc__annotate_init,
 163#ifdef HAVE_LIBDW_SUPPORT
 164		.update_insn_state = update_insn_state_powerpc,
 165#endif
 166	},
 167	{
 168		.name = "riscv64",
 169		.init = riscv64__annotate_init,
 170	},
 171	{
 172		.name = "s390",
 173		.init = s390__annotate_init,
 174		.objdump =  {
 175			.comment_char = '#',
 176		},
 177	},
 178	{
 179		.name = "sparc",
 180		.init = sparc__annotate_init,
 181		.objdump = {
 182			.comment_char = '#',
 183		},
 184	},
 185	{
 186		.name = "loongarch",
 187		.init = loongarch__annotate_init,
 188		.objdump = {
 189			.comment_char = '#',
 190		},
 191	},
 192};
 193
 194static int arch__key_cmp(const void *name, const void *archp)
 195{
 196	const struct arch *arch = archp;
 197
 198	return strcmp(name, arch->name);
 199}
 200
 201static int arch__cmp(const void *a, const void *b)
 202{
 203	const struct arch *aa = a;
 204	const struct arch *ab = b;
 205
 206	return strcmp(aa->name, ab->name);
 207}
 208
 209static void arch__sort(void)
 210{
 211	const int nmemb = ARRAY_SIZE(architectures);
 212
 213	qsort(architectures, nmemb, sizeof(struct arch), arch__cmp);
 214}
 215
 216struct arch *arch__find(const char *name)
 217{
 218	const int nmemb = ARRAY_SIZE(architectures);
 219	static bool sorted;
 220
 221	if (!sorted) {
 222		arch__sort();
 223		sorted = true;
 224	}
 225
 226	return bsearch(name, architectures, nmemb, sizeof(struct arch), arch__key_cmp);
 227}
 228
 229bool arch__is(struct arch *arch, const char *name)
 230{
 231	return !strcmp(arch->name, name);
 232}
 233
 234static void ins_ops__delete(struct ins_operands *ops)
 235{
 236	if (ops == NULL)
 237		return;
 238	zfree(&ops->source.raw);
 239	zfree(&ops->source.name);
 240	zfree(&ops->target.raw);
 241	zfree(&ops->target.name);
 242}
 243
 244static int ins__raw_scnprintf(struct ins *ins, char *bf, size_t size,
 245			      struct ins_operands *ops, int max_ins_name)
 246{
 247	return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name, ops->raw);
 248}
 249
 250static int ins__scnprintf(struct ins *ins, char *bf, size_t size,
 251			  struct ins_operands *ops, int max_ins_name)
 252{
 253	if (ins->ops->scnprintf)
 254		return ins->ops->scnprintf(ins, bf, size, ops, max_ins_name);
 255
 256	return ins__raw_scnprintf(ins, bf, size, ops, max_ins_name);
 257}
 258
 259bool ins__is_fused(struct arch *arch, const char *ins1, const char *ins2)
 260{
 261	if (!arch || !arch->ins_is_fused)
 262		return false;
 263
 264	return arch->ins_is_fused(arch, ins1, ins2);
 265}
 266
 267static int call__parse(struct arch *arch, struct ins_operands *ops, struct map_symbol *ms,
 268		struct disasm_line *dl __maybe_unused)
 269{
 270	char *endptr, *tok, *name;
 271	struct map *map = ms->map;
 272	struct addr_map_symbol target;
 273
 274	ops->target.addr = strtoull(ops->raw, &endptr, 16);
 275
 276	name = strchr(endptr, '<');
 277	if (name == NULL)
 278		goto indirect_call;
 279
 280	name++;
 281
 282	if (arch->objdump.skip_functions_char &&
 283	    strchr(name, arch->objdump.skip_functions_char))
 284		return -1;
 285
 286	tok = strchr(name, '>');
 287	if (tok == NULL)
 288		return -1;
 289
 290	*tok = '\0';
 291	ops->target.name = strdup(name);
 292	*tok = '>';
 293
 294	if (ops->target.name == NULL)
 295		return -1;
 296find_target:
 297	target = (struct addr_map_symbol) {
 298		.ms = { .map = map__get(map), },
 299		.addr = map__objdump_2mem(map, ops->target.addr),
 300	};
 301
 302	if (maps__find_ams(ms->maps, &target) == 0 &&
 303	    map__rip_2objdump(target.ms.map, map__map_ip(target.ms.map, target.addr)) == ops->target.addr)
 304		ops->target.sym = target.ms.sym;
 305
 306	addr_map_symbol__exit(&target);
 307	return 0;
 308
 309indirect_call:
 310	tok = strchr(endptr, '*');
 311	if (tok != NULL) {
 312		endptr++;
 313
 314		/* Indirect call can use a non-rip register and offset: callq  *0x8(%rbx).
 315		 * Do not parse such instruction.  */
 316		if (strstr(endptr, "(%r") == NULL)
 317			ops->target.addr = strtoull(endptr, NULL, 16);
 318	}
 319	goto find_target;
 320}
 321
 322static int call__scnprintf(struct ins *ins, char *bf, size_t size,
 323			   struct ins_operands *ops, int max_ins_name)
 324{
 325	if (ops->target.sym)
 326		return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name, ops->target.sym->name);
 327
 328	if (ops->target.addr == 0)
 329		return ins__raw_scnprintf(ins, bf, size, ops, max_ins_name);
 330
 331	if (ops->target.name)
 332		return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name, ops->target.name);
 333
 334	return scnprintf(bf, size, "%-*s *%" PRIx64, max_ins_name, ins->name, ops->target.addr);
 335}
 336
 337static struct ins_ops call_ops = {
 338	.parse	   = call__parse,
 339	.scnprintf = call__scnprintf,
 340};
 341
 342bool ins__is_call(const struct ins *ins)
 343{
 344	return ins->ops == &call_ops || ins->ops == &s390_call_ops || ins->ops == &loongarch_call_ops;
 345}
 346
 347/*
 348 * Prevents from matching commas in the comment section, e.g.:
 349 * ffff200008446e70:       b.cs    ffff2000084470f4 <generic_exec_single+0x314>  // b.hs, b.nlast
 350 *
 351 * and skip comma as part of function arguments, e.g.:
 352 * 1d8b4ac <linemap_lookup(line_maps const*, unsigned int)+0xcc>
 353 */
 354static inline const char *validate_comma(const char *c, struct ins_operands *ops)
 355{
 356	if (ops->jump.raw_comment && c > ops->jump.raw_comment)
 357		return NULL;
 358
 359	if (ops->jump.raw_func_start && c > ops->jump.raw_func_start)
 360		return NULL;
 361
 362	return c;
 363}
 364
 365static int jump__parse(struct arch *arch, struct ins_operands *ops, struct map_symbol *ms,
 366		struct disasm_line *dl __maybe_unused)
 367{
 368	struct map *map = ms->map;
 369	struct symbol *sym = ms->sym;
 370	struct addr_map_symbol target = {
 371		.ms = { .map = map__get(map), },
 372	};
 373	const char *c = strchr(ops->raw, ',');
 374	u64 start, end;
 375
 376	ops->jump.raw_comment = strchr(ops->raw, arch->objdump.comment_char);
 377	ops->jump.raw_func_start = strchr(ops->raw, '<');
 378
 379	c = validate_comma(c, ops);
 380
 381	/*
 382	 * Examples of lines to parse for the _cpp_lex_token@@Base
 383	 * function:
 384	 *
 385	 * 1159e6c: jne    115aa32 <_cpp_lex_token@@Base+0xf92>
 386	 * 1159e8b: jne    c469be <cpp_named_operator2name@@Base+0xa72>
 387	 *
 388	 * The first is a jump to an offset inside the same function,
 389	 * the second is to another function, i.e. that 0xa72 is an
 390	 * offset in the cpp_named_operator2name@@base function.
 391	 */
 392	/*
 393	 * skip over possible up to 2 operands to get to address, e.g.:
 394	 * tbnz	 w0, #26, ffff0000083cd190 <security_file_permission+0xd0>
 395	 */
 396	if (c != NULL) {
 397		c++;
 398		ops->target.addr = strtoull(c, NULL, 16);
 399		if (!ops->target.addr) {
 400			c = strchr(c, ',');
 401			c = validate_comma(c, ops);
 402			if (c != NULL) {
 403				c++;
 404				ops->target.addr = strtoull(c, NULL, 16);
 405			}
 406		}
 407	} else {
 408		ops->target.addr = strtoull(ops->raw, NULL, 16);
 409	}
 410
 411	target.addr = map__objdump_2mem(map, ops->target.addr);
 412	start = map__unmap_ip(map, sym->start);
 413	end = map__unmap_ip(map, sym->end);
 414
 415	ops->target.outside = target.addr < start || target.addr >= end;
 416
 417	/*
 418	 * FIXME: things like this in _cpp_lex_token (gcc's cc1 program):
 419
 420		cpp_named_operator2name@@Base+0xa72
 421
 422	 * Point to a place that is after the cpp_named_operator2name
 423	 * boundaries, i.e.  in the ELF symbol table for cc1
 424	 * cpp_named_operator2name is marked as being 32-bytes long, but it in
 425	 * fact is much larger than that, so we seem to need a symbols__find()
 426	 * routine that looks for >= current->start and  < next_symbol->start,
 427	 * possibly just for C++ objects?
 428	 *
 429	 * For now lets just make some progress by marking jumps to outside the
 430	 * current function as call like.
 431	 *
 432	 * Actual navigation will come next, with further understanding of how
 433	 * the symbol searching and disassembly should be done.
 434	 */
 435	if (maps__find_ams(ms->maps, &target) == 0 &&
 436	    map__rip_2objdump(target.ms.map, map__map_ip(target.ms.map, target.addr)) == ops->target.addr)
 437		ops->target.sym = target.ms.sym;
 438
 439	if (!ops->target.outside) {
 440		ops->target.offset = target.addr - start;
 441		ops->target.offset_avail = true;
 442	} else {
 443		ops->target.offset_avail = false;
 444	}
 445	addr_map_symbol__exit(&target);
 446	return 0;
 447}
 448
 449static int jump__scnprintf(struct ins *ins, char *bf, size_t size,
 450			   struct ins_operands *ops, int max_ins_name)
 451{
 452	const char *c;
 453
 454	if (!ops->target.addr || ops->target.offset < 0)
 455		return ins__raw_scnprintf(ins, bf, size, ops, max_ins_name);
 456
 457	if (ops->target.outside && ops->target.sym != NULL)
 458		return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name, ops->target.sym->name);
 459
 460	c = strchr(ops->raw, ',');
 461	c = validate_comma(c, ops);
 462
 463	if (c != NULL) {
 464		const char *c2 = strchr(c + 1, ',');
 465
 466		c2 = validate_comma(c2, ops);
 467		/* check for 3-op insn */
 468		if (c2 != NULL)
 469			c = c2;
 470		c++;
 471
 472		/* mirror arch objdump's space-after-comma style */
 473		if (*c == ' ')
 474			c++;
 475	}
 476
 477	return scnprintf(bf, size, "%-*s %.*s%" PRIx64, max_ins_name,
 478			 ins->name, c ? c - ops->raw : 0, ops->raw,
 479			 ops->target.offset);
 480}
 481
 482static void jump__delete(struct ins_operands *ops __maybe_unused)
 483{
 484	/*
 485	 * The ops->jump.raw_comment and ops->jump.raw_func_start belong to the
 486	 * raw string, don't free them.
 487	 */
 488}
 489
 490static struct ins_ops jump_ops = {
 491	.free	   = jump__delete,
 492	.parse	   = jump__parse,
 493	.scnprintf = jump__scnprintf,
 494};
 495
 496bool ins__is_jump(const struct ins *ins)
 497{
 498	return ins->ops == &jump_ops || ins->ops == &loongarch_jump_ops;
 499}
 500
 501static int comment__symbol(char *raw, char *comment, u64 *addrp, char **namep)
 502{
 503	char *endptr, *name, *t;
 504
 505	if (strstr(raw, "(%rip)") == NULL)
 506		return 0;
 507
 508	*addrp = strtoull(comment, &endptr, 16);
 509	if (endptr == comment)
 510		return 0;
 511	name = strchr(endptr, '<');
 512	if (name == NULL)
 513		return -1;
 514
 515	name++;
 516
 517	t = strchr(name, '>');
 518	if (t == NULL)
 519		return 0;
 520
 521	*t = '\0';
 522	*namep = strdup(name);
 523	*t = '>';
 524
 525	return 0;
 526}
 527
 528static int lock__parse(struct arch *arch, struct ins_operands *ops, struct map_symbol *ms,
 529		struct disasm_line *dl __maybe_unused)
 530{
 531	ops->locked.ops = zalloc(sizeof(*ops->locked.ops));
 532	if (ops->locked.ops == NULL)
 533		return 0;
 534
 535	if (disasm_line__parse(ops->raw, &ops->locked.ins.name, &ops->locked.ops->raw) < 0)
 536		goto out_free_ops;
 537
 538	ops->locked.ins.ops = ins__find(arch, ops->locked.ins.name, 0);
 539
 540	if (ops->locked.ins.ops == NULL)
 541		goto out_free_ops;
 542
 543	if (ops->locked.ins.ops->parse &&
 544	    ops->locked.ins.ops->parse(arch, ops->locked.ops, ms, NULL) < 0)
 545		goto out_free_ops;
 546
 547	return 0;
 548
 549out_free_ops:
 550	zfree(&ops->locked.ops);
 551	return 0;
 552}
 553
 554static int lock__scnprintf(struct ins *ins, char *bf, size_t size,
 555			   struct ins_operands *ops, int max_ins_name)
 556{
 557	int printed;
 558
 559	if (ops->locked.ins.ops == NULL)
 560		return ins__raw_scnprintf(ins, bf, size, ops, max_ins_name);
 561
 562	printed = scnprintf(bf, size, "%-*s ", max_ins_name, ins->name);
 563	return printed + ins__scnprintf(&ops->locked.ins, bf + printed,
 564					size - printed, ops->locked.ops, max_ins_name);
 565}
 566
 567static void lock__delete(struct ins_operands *ops)
 568{
 569	struct ins *ins = &ops->locked.ins;
 570
 571	if (ins->ops && ins->ops->free)
 572		ins->ops->free(ops->locked.ops);
 573	else
 574		ins_ops__delete(ops->locked.ops);
 575
 576	zfree(&ops->locked.ops);
 577	zfree(&ops->locked.ins.name);
 578	zfree(&ops->target.raw);
 579	zfree(&ops->target.name);
 580}
 581
 582static struct ins_ops lock_ops = {
 583	.free	   = lock__delete,
 584	.parse	   = lock__parse,
 585	.scnprintf = lock__scnprintf,
 586};
 587
 588/*
 589 * Check if the operand has more than one registers like x86 SIB addressing:
 590 *   0x1234(%rax, %rbx, 8)
 591 *
 592 * But it doesn't care segment selectors like %gs:0x5678(%rcx), so just check
 593 * the input string after 'memory_ref_char' if exists.
 594 */
 595static bool check_multi_regs(struct arch *arch, const char *op)
 596{
 597	int count = 0;
 598
 599	if (arch->objdump.register_char == 0)
 600		return false;
 601
 602	if (arch->objdump.memory_ref_char) {
 603		op = strchr(op, arch->objdump.memory_ref_char);
 604		if (op == NULL)
 605			return false;
 606	}
 607
 608	while ((op = strchr(op, arch->objdump.register_char)) != NULL) {
 609		count++;
 610		op++;
 611	}
 612
 613	return count > 1;
 614}
 615
 616static int mov__parse(struct arch *arch, struct ins_operands *ops, struct map_symbol *ms __maybe_unused,
 617		struct disasm_line *dl __maybe_unused)
 618{
 619	char *s = strchr(ops->raw, ','), *target, *comment, prev;
 620
 621	if (s == NULL)
 622		return -1;
 623
 624	*s = '\0';
 625
 626	/*
 627	 * x86 SIB addressing has something like 0x8(%rax, %rcx, 1)
 628	 * then it needs to have the closing parenthesis.
 629	 */
 630	if (strchr(ops->raw, '(')) {
 631		*s = ',';
 632		s = strchr(ops->raw, ')');
 633		if (s == NULL || s[1] != ',')
 634			return -1;
 635		*++s = '\0';
 636	}
 637
 638	ops->source.raw = strdup(ops->raw);
 639	*s = ',';
 640
 641	if (ops->source.raw == NULL)
 642		return -1;
 643
 644	ops->source.multi_regs = check_multi_regs(arch, ops->source.raw);
 645
 646	target = skip_spaces(++s);
 647	comment = strchr(s, arch->objdump.comment_char);
 648
 649	if (comment != NULL)
 650		s = comment - 1;
 651	else
 652		s = strchr(s, '\0') - 1;
 653
 654	while (s > target && isspace(s[0]))
 655		--s;
 656	s++;
 657	prev = *s;
 658	*s = '\0';
 659
 660	ops->target.raw = strdup(target);
 661	*s = prev;
 662
 663	if (ops->target.raw == NULL)
 664		goto out_free_source;
 665
 666	ops->target.multi_regs = check_multi_regs(arch, ops->target.raw);
 667
 668	if (comment == NULL)
 669		return 0;
 670
 671	comment = skip_spaces(comment);
 672	comment__symbol(ops->source.raw, comment + 1, &ops->source.addr, &ops->source.name);
 673	comment__symbol(ops->target.raw, comment + 1, &ops->target.addr, &ops->target.name);
 674
 675	return 0;
 676
 677out_free_source:
 678	zfree(&ops->source.raw);
 679	return -1;
 680}
 681
 682static int mov__scnprintf(struct ins *ins, char *bf, size_t size,
 683			   struct ins_operands *ops, int max_ins_name)
 684{
 685	return scnprintf(bf, size, "%-*s %s,%s", max_ins_name, ins->name,
 686			 ops->source.name ?: ops->source.raw,
 687			 ops->target.name ?: ops->target.raw);
 688}
 689
 690static struct ins_ops mov_ops = {
 691	.parse	   = mov__parse,
 692	.scnprintf = mov__scnprintf,
 693};
 694
 695#define PPC_22_30(R)    (((R) >> 1) & 0x1ff)
 696#define MINUS_EXT_XO_FORM	234
 697#define SUB_EXT_XO_FORM		232
 698#define	ADD_ZERO_EXT_XO_FORM	202
 699#define	SUB_ZERO_EXT_XO_FORM	200
 700
 701static int arithmetic__scnprintf(struct ins *ins, char *bf, size_t size,
 702		struct ins_operands *ops, int max_ins_name)
 703{
 704	return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name,
 705			ops->raw);
 706}
 707
 708/*
 709 * Sets the fields: multi_regs and "mem_ref".
 710 * "mem_ref" is set for ops->source which is later used to
 711 * fill the objdump->memory_ref-char field. This ops is currently
 712 * used by powerpc and since binary instruction code is used to
 713 * extract opcode, regs and offset, no other parsing is needed here.
 714 *
 715 * Dont set multi regs for 4 cases since it has only one operand
 716 * for source:
 717 * - Add to Minus One Extended XO-form ( Ex: addme, addmeo )
 718 * - Subtract From Minus One Extended XO-form ( Ex: subfme )
 719 * - Add to Zero Extended XO-form ( Ex: addze, addzeo )
 720 * - Subtract From Zero Extended XO-form ( Ex: subfze )
 721 */
 722static int arithmetic__parse(struct arch *arch __maybe_unused, struct ins_operands *ops,
 723		struct map_symbol *ms __maybe_unused, struct disasm_line *dl)
 724{
 725	int opcode = PPC_OP(dl->raw.raw_insn);
 726
 727	ops->source.mem_ref = false;
 728	if (opcode == 31) {
 729		if ((opcode != MINUS_EXT_XO_FORM) && (opcode != SUB_EXT_XO_FORM) \
 730				&& (opcode != ADD_ZERO_EXT_XO_FORM) && (opcode != SUB_ZERO_EXT_XO_FORM))
 731			ops->source.multi_regs = true;
 732	}
 733
 734	ops->target.mem_ref = false;
 735	ops->target.multi_regs = false;
 736
 737	return 0;
 738}
 739
 740static struct ins_ops arithmetic_ops = {
 741	.parse     = arithmetic__parse,
 742	.scnprintf = arithmetic__scnprintf,
 743};
 744
 745static int load_store__scnprintf(struct ins *ins, char *bf, size_t size,
 746		struct ins_operands *ops, int max_ins_name)
 747{
 748	return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name,
 749			ops->raw);
 750}
 751
 752/*
 753 * Sets the fields: multi_regs and "mem_ref".
 754 * "mem_ref" is set for ops->source which is later used to
 755 * fill the objdump->memory_ref-char field. This ops is currently
 756 * used by powerpc and since binary instruction code is used to
 757 * extract opcode, regs and offset, no other parsing is needed here
 758 */
 759static int load_store__parse(struct arch *arch __maybe_unused, struct ins_operands *ops,
 760		struct map_symbol *ms __maybe_unused, struct disasm_line *dl __maybe_unused)
 761{
 762	ops->source.mem_ref = true;
 763	ops->source.multi_regs = false;
 764	/* opcode 31 is of X form */
 765	if (PPC_OP(dl->raw.raw_insn) == 31)
 766		ops->source.multi_regs = true;
 767
 768	ops->target.mem_ref = false;
 769	ops->target.multi_regs = false;
 770
 771	return 0;
 772}
 773
 774static struct ins_ops load_store_ops = {
 775	.parse     = load_store__parse,
 776	.scnprintf = load_store__scnprintf,
 777};
 778
 779static int dec__parse(struct arch *arch __maybe_unused, struct ins_operands *ops, struct map_symbol *ms __maybe_unused,
 780		struct disasm_line *dl __maybe_unused)
 781{
 782	char *target, *comment, *s, prev;
 783
 784	target = s = ops->raw;
 785
 786	while (s[0] != '\0' && !isspace(s[0]))
 787		++s;
 788	prev = *s;
 789	*s = '\0';
 790
 791	ops->target.raw = strdup(target);
 792	*s = prev;
 793
 794	if (ops->target.raw == NULL)
 795		return -1;
 796
 797	comment = strchr(s, arch->objdump.comment_char);
 798	if (comment == NULL)
 799		return 0;
 800
 801	comment = skip_spaces(comment);
 802	comment__symbol(ops->target.raw, comment + 1, &ops->target.addr, &ops->target.name);
 803
 804	return 0;
 805}
 806
 807static int dec__scnprintf(struct ins *ins, char *bf, size_t size,
 808			   struct ins_operands *ops, int max_ins_name)
 809{
 810	return scnprintf(bf, size, "%-*s %s", max_ins_name, ins->name,
 811			 ops->target.name ?: ops->target.raw);
 812}
 813
 814static struct ins_ops dec_ops = {
 815	.parse	   = dec__parse,
 816	.scnprintf = dec__scnprintf,
 817};
 818
 819static int nop__scnprintf(struct ins *ins __maybe_unused, char *bf, size_t size,
 820			  struct ins_operands *ops __maybe_unused, int max_ins_name)
 821{
 822	return scnprintf(bf, size, "%-*s", max_ins_name, "nop");
 823}
 824
 825static struct ins_ops nop_ops = {
 826	.scnprintf = nop__scnprintf,
 827};
 828
 829static struct ins_ops ret_ops = {
 830	.scnprintf = ins__raw_scnprintf,
 831};
 832
 833static bool ins__is_nop(const struct ins *ins)
 834{
 835	return ins->ops == &nop_ops;
 836}
 837
 838bool ins__is_ret(const struct ins *ins)
 839{
 840	return ins->ops == &ret_ops;
 841}
 842
 843bool ins__is_lock(const struct ins *ins)
 844{
 845	return ins->ops == &lock_ops;
 846}
 847
 848static int ins__key_cmp(const void *name, const void *insp)
 849{
 850	const struct ins *ins = insp;
 851
 852	return strcmp(name, ins->name);
 853}
 854
 855static int ins__cmp(const void *a, const void *b)
 856{
 857	const struct ins *ia = a;
 858	const struct ins *ib = b;
 859
 860	return strcmp(ia->name, ib->name);
 861}
 862
 863static void ins__sort(struct arch *arch)
 864{
 865	const int nmemb = arch->nr_instructions;
 866
 867	qsort(arch->instructions, nmemb, sizeof(struct ins), ins__cmp);
 868}
 869
 870static struct ins_ops *__ins__find(struct arch *arch, const char *name, struct disasm_line *dl)
 871{
 872	struct ins *ins;
 873	const int nmemb = arch->nr_instructions;
 874
 875	if (arch__is(arch, "powerpc")) {
 876		/*
 877		 * For powerpc, identify the instruction ops
 878		 * from the opcode using raw_insn.
 879		 */
 880		struct ins_ops *ops;
 881
 882		ops = check_ppc_insn(dl);
 883		if (ops)
 884			return ops;
 885	}
 886
 887	if (!arch->sorted_instructions) {
 888		ins__sort(arch);
 889		arch->sorted_instructions = true;
 890	}
 891
 892	ins = bsearch(name, arch->instructions, nmemb, sizeof(struct ins), ins__key_cmp);
 893	if (ins)
 894		return ins->ops;
 895
 896	if (arch->insn_suffix) {
 897		char tmp[32];
 898		char suffix;
 899		size_t len = strlen(name);
 900
 901		if (len == 0 || len >= sizeof(tmp))
 902			return NULL;
 903
 904		suffix = name[len - 1];
 905		if (strchr(arch->insn_suffix, suffix) == NULL)
 906			return NULL;
 907
 908		strcpy(tmp, name);
 909		tmp[len - 1] = '\0'; /* remove the suffix and check again */
 910
 911		ins = bsearch(tmp, arch->instructions, nmemb, sizeof(struct ins), ins__key_cmp);
 912	}
 913	return ins ? ins->ops : NULL;
 914}
 915
 916struct ins_ops *ins__find(struct arch *arch, const char *name, struct disasm_line *dl)
 917{
 918	struct ins_ops *ops = __ins__find(arch, name, dl);
 919
 920	if (!ops && arch->associate_instruction_ops)
 921		ops = arch->associate_instruction_ops(arch, name);
 922
 923	return ops;
 924}
 925
 926static void disasm_line__init_ins(struct disasm_line *dl, struct arch *arch, struct map_symbol *ms)
 927{
 928	dl->ins.ops = ins__find(arch, dl->ins.name, dl);
 929
 930	if (!dl->ins.ops)
 931		return;
 932
 933	if (dl->ins.ops->parse && dl->ins.ops->parse(arch, &dl->ops, ms, dl) < 0)
 934		dl->ins.ops = NULL;
 935}
 936
 937static int disasm_line__parse(char *line, const char **namep, char **rawp)
 938{
 939	char tmp, *name = skip_spaces(line);
 940
 941	if (name[0] == '\0')
 942		return -1;
 943
 944	*rawp = name + 1;
 945
 946	while ((*rawp)[0] != '\0' && !isspace((*rawp)[0]))
 947		++*rawp;
 948
 949	tmp = (*rawp)[0];
 950	(*rawp)[0] = '\0';
 951	*namep = strdup(name);
 952
 953	if (*namep == NULL)
 954		goto out;
 955
 956	(*rawp)[0] = tmp;
 957	*rawp = strim(*rawp);
 958
 959	return 0;
 960
 961out:
 962	return -1;
 963}
 964
 965/*
 966 * Parses the result captured from symbol__disassemble_*
 967 * Example, line read from DSO file in powerpc:
 968 * line:    38 01 81 e8
 969 * opcode: fetched from arch specific get_opcode_insn
 970 * rawp_insn: e8810138
 971 *
 972 * rawp_insn is used later to extract the reg/offset fields
 973 */
 974#define	PPC_OP(op)	(((op) >> 26) & 0x3F)
 975#define	RAW_BYTES	11
 976
 977static int disasm_line__parse_powerpc(struct disasm_line *dl, struct annotate_args *args)
 978{
 979	char *line = dl->al.line;
 980	const char **namep = &dl->ins.name;
 981	char **rawp = &dl->ops.raw;
 982	char *tmp_raw_insn, *name_raw_insn = skip_spaces(line);
 983	char *name = skip_spaces(name_raw_insn + RAW_BYTES);
 984	int disasm = 0;
 985	int ret = 0;
 986
 987	if (args->options->disassembler_used)
 988		disasm = 1;
 989
 990	if (name_raw_insn[0] == '\0')
 991		return -1;
 992
 993	if (disasm)
 994		ret = disasm_line__parse(name, namep, rawp);
 995	else
 996		*namep = "";
 997
 998	tmp_raw_insn = strndup(name_raw_insn, 11);
 999	if (tmp_raw_insn == NULL)
1000		return -1;
1001
1002	remove_spaces(tmp_raw_insn);
1003
1004	sscanf(tmp_raw_insn, "%x", &dl->raw.raw_insn);
1005	if (disasm)
1006		dl->raw.raw_insn = be32_to_cpu(dl->raw.raw_insn);
1007
1008	return ret;
1009}
1010
1011static void annotation_line__init(struct annotation_line *al,
1012				  struct annotate_args *args,
1013				  int nr)
1014{
1015	al->offset = args->offset;
1016	al->line = strdup(args->line);
1017	al->line_nr = args->line_nr;
1018	al->fileloc = args->fileloc;
1019	al->data_nr = nr;
1020}
1021
1022static void annotation_line__exit(struct annotation_line *al)
1023{
1024	zfree_srcline(&al->path);
1025	zfree(&al->line);
1026	zfree(&al->cycles);
1027	zfree(&al->br_cntr);
1028}
1029
1030static size_t disasm_line_size(int nr)
1031{
1032	struct annotation_line *al;
1033
1034	return (sizeof(struct disasm_line) + (sizeof(al->data[0]) * nr));
1035}
1036
1037/*
1038 * Allocating the disasm annotation line data with
1039 * following structure:
1040 *
1041 *    -------------------------------------------
1042 *    struct disasm_line | struct annotation_line
1043 *    -------------------------------------------
1044 *
1045 * We have 'struct annotation_line' member as last member
1046 * of 'struct disasm_line' to have an easy access.
1047 */
1048struct disasm_line *disasm_line__new(struct annotate_args *args)
1049{
1050	struct disasm_line *dl = NULL;
1051	struct annotation *notes = symbol__annotation(args->ms->sym);
1052	int nr = notes->src->nr_events;
1053
1054	dl = zalloc(disasm_line_size(nr));
1055	if (!dl)
1056		return NULL;
1057
1058	annotation_line__init(&dl->al, args, nr);
1059	if (dl->al.line == NULL)
1060		goto out_delete;
1061
1062	if (args->offset != -1) {
1063		if (arch__is(args->arch, "powerpc")) {
1064			if (disasm_line__parse_powerpc(dl, args) < 0)
1065				goto out_free_line;
1066		} else if (disasm_line__parse(dl->al.line, &dl->ins.name, &dl->ops.raw) < 0)
1067			goto out_free_line;
1068
1069		disasm_line__init_ins(dl, args->arch, args->ms);
1070	}
1071
1072	return dl;
1073
1074out_free_line:
1075	zfree(&dl->al.line);
1076out_delete:
1077	free(dl);
1078	return NULL;
1079}
1080
1081void disasm_line__free(struct disasm_line *dl)
1082{
1083	if (dl->ins.ops && dl->ins.ops->free)
1084		dl->ins.ops->free(&dl->ops);
1085	else
1086		ins_ops__delete(&dl->ops);
1087	zfree(&dl->ins.name);
1088	annotation_line__exit(&dl->al);
1089	free(dl);
1090}
1091
1092int disasm_line__scnprintf(struct disasm_line *dl, char *bf, size_t size, bool raw, int max_ins_name)
1093{
1094	if (raw || !dl->ins.ops)
1095		return scnprintf(bf, size, "%-*s %s", max_ins_name, dl->ins.name, dl->ops.raw);
1096
1097	return ins__scnprintf(&dl->ins, bf, size, &dl->ops, max_ins_name);
1098}
1099
1100/*
1101 * symbol__parse_objdump_line() parses objdump output (with -d --no-show-raw)
1102 * which looks like following
1103 *
1104 *  0000000000415500 <_init>:
1105 *    415500:       sub    $0x8,%rsp
1106 *    415504:       mov    0x2f5ad5(%rip),%rax        # 70afe0 <_DYNAMIC+0x2f8>
1107 *    41550b:       test   %rax,%rax
1108 *    41550e:       je     415515 <_init+0x15>
1109 *    415510:       callq  416e70 <__gmon_start__@plt>
1110 *    415515:       add    $0x8,%rsp
1111 *    415519:       retq
1112 *
1113 * it will be parsed and saved into struct disasm_line as
1114 *  <offset>       <name>  <ops.raw>
1115 *
1116 * The offset will be a relative offset from the start of the symbol and -1
1117 * means that it's not a disassembly line so should be treated differently.
1118 * The ops.raw part will be parsed further according to type of the instruction.
1119 */
1120static int symbol__parse_objdump_line(struct symbol *sym,
1121				      struct annotate_args *args,
1122				      char *parsed_line, int *line_nr, char **fileloc)
1123{
1124	struct map *map = args->ms->map;
1125	struct annotation *notes = symbol__annotation(sym);
1126	struct disasm_line *dl;
1127	char *tmp;
1128	s64 line_ip, offset = -1;
1129	regmatch_t match[2];
1130
1131	/* /filename:linenr ? Save line number and ignore. */
1132	if (regexec(&file_lineno, parsed_line, 2, match, 0) == 0) {
1133		*line_nr = atoi(parsed_line + match[1].rm_so);
1134		free(*fileloc);
1135		*fileloc = strdup(parsed_line);
1136		return 0;
1137	}
1138
1139	/* Process hex address followed by ':'. */
1140	line_ip = strtoull(parsed_line, &tmp, 16);
1141	if (parsed_line != tmp && tmp[0] == ':' && tmp[1] != '\0') {
1142		u64 start = map__rip_2objdump(map, sym->start),
1143		    end = map__rip_2objdump(map, sym->end);
1144
1145		offset = line_ip - start;
1146		if ((u64)line_ip < start || (u64)line_ip >= end)
1147			offset = -1;
1148		else
1149			parsed_line = tmp + 1;
1150	}
1151
1152	args->offset  = offset;
1153	args->line    = parsed_line;
1154	args->line_nr = *line_nr;
1155	args->fileloc = *fileloc;
1156	args->ms->sym  = sym;
1157
1158	dl = disasm_line__new(args);
1159	(*line_nr)++;
1160
1161	if (dl == NULL)
1162		return -1;
1163
1164	if (!disasm_line__has_local_offset(dl)) {
1165		dl->ops.target.offset = dl->ops.target.addr -
1166					map__rip_2objdump(map, sym->start);
1167		dl->ops.target.offset_avail = true;
1168	}
1169
1170	/* kcore has no symbols, so add the call target symbol */
1171	if (dl->ins.ops && ins__is_call(&dl->ins) && !dl->ops.target.sym) {
1172		struct addr_map_symbol target = {
1173			.addr = dl->ops.target.addr,
1174			.ms = { .map = map__get(map), },
1175		};
1176
1177		if (!maps__find_ams(args->ms->maps, &target) &&
1178		    target.ms.sym->start == target.al_addr)
1179			dl->ops.target.sym = target.ms.sym;
1180
1181		addr_map_symbol__exit(&target);
1182	}
1183
1184	annotation_line__add(&dl->al, &notes->src->source);
1185	return 0;
1186}
1187
1188static void delete_last_nop(struct symbol *sym)
1189{
1190	struct annotation *notes = symbol__annotation(sym);
1191	struct list_head *list = &notes->src->source;
1192	struct disasm_line *dl;
1193
1194	while (!list_empty(list)) {
1195		dl = list_entry(list->prev, struct disasm_line, al.node);
1196
1197		if (dl->ins.ops) {
1198			if (!ins__is_nop(&dl->ins))
1199				return;
1200		} else {
1201			if (!strstr(dl->al.line, " nop ") &&
1202			    !strstr(dl->al.line, " nopl ") &&
1203			    !strstr(dl->al.line, " nopw "))
1204				return;
1205		}
1206
1207		list_del_init(&dl->al.node);
1208		disasm_line__free(dl);
1209	}
1210}
1211
1212int symbol__strerror_disassemble(struct map_symbol *ms, int errnum, char *buf, size_t buflen)
1213{
1214	struct dso *dso = map__dso(ms->map);
1215
1216	BUG_ON(buflen == 0);
1217
1218	if (errnum >= 0) {
1219		str_error_r(errnum, buf, buflen);
1220		return 0;
1221	}
1222
1223	switch (errnum) {
1224	case SYMBOL_ANNOTATE_ERRNO__NO_VMLINUX: {
1225		char bf[SBUILD_ID_SIZE + 15] = " with build id ";
1226		char *build_id_msg = NULL;
1227
1228		if (dso__has_build_id(dso)) {
1229			build_id__snprintf(dso__bid(dso), bf + 15, sizeof(bf) - 15);
1230			build_id_msg = bf;
1231		}
1232		scnprintf(buf, buflen,
1233			  "No vmlinux file%s\nwas found in the path.\n\n"
1234			  "Note that annotation using /proc/kcore requires CAP_SYS_RAWIO capability.\n\n"
1235			  "Please use:\n\n"
1236			  "  perf buildid-cache -vu vmlinux\n\n"
1237			  "or:\n\n"
1238			  "  --vmlinux vmlinux\n", build_id_msg ?: "");
1239	}
1240		break;
1241	case SYMBOL_ANNOTATE_ERRNO__NO_LIBOPCODES_FOR_BPF:
1242		scnprintf(buf, buflen, "Please link with binutils's libopcode to enable BPF annotation");
1243		break;
1244	case SYMBOL_ANNOTATE_ERRNO__ARCH_INIT_REGEXP:
1245		scnprintf(buf, buflen, "Problems with arch specific instruction name regular expressions.");
1246		break;
1247	case SYMBOL_ANNOTATE_ERRNO__ARCH_INIT_CPUID_PARSING:
1248		scnprintf(buf, buflen, "Problems while parsing the CPUID in the arch specific initialization.");
1249		break;
1250	case SYMBOL_ANNOTATE_ERRNO__BPF_INVALID_FILE:
1251		scnprintf(buf, buflen, "Invalid BPF file: %s.", dso__long_name(dso));
1252		break;
1253	case SYMBOL_ANNOTATE_ERRNO__BPF_MISSING_BTF:
1254		scnprintf(buf, buflen, "The %s BPF file has no BTF section, compile with -g or use pahole -J.",
1255			  dso__long_name(dso));
1256		break;
1257	case SYMBOL_ANNOTATE_ERRNO__COULDNT_DETERMINE_FILE_TYPE:
1258		scnprintf(buf, buflen, "Couldn't determine the file %s type.", dso__long_name(dso));
1259		break;
1260	default:
1261		scnprintf(buf, buflen, "Internal error: Invalid %d error code\n", errnum);
1262		break;
1263	}
1264
1265	return 0;
1266}
1267
1268static int dso__disassemble_filename(struct dso *dso, char *filename, size_t filename_size)
1269{
1270	char linkname[PATH_MAX];
1271	char *build_id_filename;
1272	char *build_id_path = NULL;
1273	char *pos;
1274	int len;
1275
1276	if (dso__symtab_type(dso) == DSO_BINARY_TYPE__KALLSYMS &&
1277	    !dso__is_kcore(dso))
1278		return SYMBOL_ANNOTATE_ERRNO__NO_VMLINUX;
1279
1280	build_id_filename = dso__build_id_filename(dso, NULL, 0, false);
1281	if (build_id_filename) {
1282		__symbol__join_symfs(filename, filename_size, build_id_filename);
1283		free(build_id_filename);
1284	} else {
1285		if (dso__has_build_id(dso))
1286			return ENOMEM;
1287		goto fallback;
1288	}
1289
1290	build_id_path = strdup(filename);
1291	if (!build_id_path)
1292		return ENOMEM;
1293
1294	/*
1295	 * old style build-id cache has name of XX/XXXXXXX.. while
1296	 * new style has XX/XXXXXXX../{elf,kallsyms,vdso}.
1297	 * extract the build-id part of dirname in the new style only.
1298	 */
1299	pos = strrchr(build_id_path, '/');
1300	if (pos && strlen(pos) < SBUILD_ID_SIZE - 2)
1301		dirname(build_id_path);
1302
1303	if (dso__is_kcore(dso))
1304		goto fallback;
1305
1306	len = readlink(build_id_path, linkname, sizeof(linkname) - 1);
1307	if (len < 0)
1308		goto fallback;
1309
1310	linkname[len] = '\0';
1311	if (strstr(linkname, DSO__NAME_KALLSYMS) ||
1312		access(filename, R_OK)) {
1313fallback:
1314		/*
1315		 * If we don't have build-ids or the build-id file isn't in the
1316		 * cache, or is just a kallsyms file, well, lets hope that this
1317		 * DSO is the same as when 'perf record' ran.
1318		 */
1319		if (dso__kernel(dso) && dso__long_name(dso)[0] == '/')
1320			snprintf(filename, filename_size, "%s", dso__long_name(dso));
1321		else
1322			__symbol__join_symfs(filename, filename_size, dso__long_name(dso));
1323
1324		mutex_lock(dso__lock(dso));
1325		if (access(filename, R_OK) && errno == ENOENT && dso__nsinfo(dso)) {
1326			char *new_name = dso__filename_with_chroot(dso, filename);
1327			if (new_name) {
1328				strlcpy(filename, new_name, filename_size);
1329				free(new_name);
1330			}
1331		}
1332		mutex_unlock(dso__lock(dso));
1333	} else if (dso__binary_type(dso) == DSO_BINARY_TYPE__NOT_FOUND) {
1334		dso__set_binary_type(dso, DSO_BINARY_TYPE__BUILD_ID_CACHE);
1335	}
1336
1337	free(build_id_path);
1338	return 0;
1339}
1340
1341static int symbol__disassemble_raw(char *filename, struct symbol *sym,
1342					struct annotate_args *args)
1343{
1344	struct annotation *notes = symbol__annotation(sym);
1345	struct map *map = args->ms->map;
1346	struct dso *dso = map__dso(map);
1347	u64 start = map__rip_2objdump(map, sym->start);
1348	u64 end = map__rip_2objdump(map, sym->end);
1349	u64 len = end - start;
1350	u64 offset;
1351	int i, count;
1352	u8 *buf = NULL;
1353	char disasm_buf[512];
1354	struct disasm_line *dl;
1355	u32 *line;
1356
1357	/* Return if objdump is specified explicitly */
1358	if (args->options->objdump_path)
1359		return -1;
1360
1361	pr_debug("Reading raw instruction from : %s using dso__data_read_offset\n", filename);
1362
1363	buf = malloc(len);
1364	if (buf == NULL)
1365		goto err;
1366
1367	count = dso__data_read_offset(dso, NULL, sym->start, buf, len);
1368
1369	line = (u32 *)buf;
1370
1371	if ((u64)count != len)
1372		goto err;
1373
1374	/* add the function address and name */
1375	scnprintf(disasm_buf, sizeof(disasm_buf), "%#"PRIx64" <%s>:",
1376		  start, sym->name);
1377
1378	args->offset = -1;
1379	args->line = disasm_buf;
1380	args->line_nr = 0;
1381	args->fileloc = NULL;
1382	args->ms->sym = sym;
1383
1384	dl = disasm_line__new(args);
1385	if (dl == NULL)
1386		goto err;
1387
1388	annotation_line__add(&dl->al, &notes->src->source);
1389
1390	/* Each raw instruction is 4 byte */
1391	count = len/4;
1392
1393	for (i = 0, offset = 0; i < count; i++) {
1394		args->offset = offset;
1395		sprintf(args->line, "%x", line[i]);
1396		dl = disasm_line__new(args);
1397		if (dl == NULL)
1398			break;
1399
1400		annotation_line__add(&dl->al, &notes->src->source);
1401		offset += 4;
1402	}
1403
1404	/* It failed in the middle */
1405	if (offset != len) {
1406		struct list_head *list = &notes->src->source;
1407
1408		/* Discard all lines and fallback to objdump */
1409		while (!list_empty(list)) {
1410			dl = list_first_entry(list, struct disasm_line, al.node);
1411
1412			list_del_init(&dl->al.node);
1413			disasm_line__free(dl);
1414		}
1415		count = -1;
1416	}
1417
1418out:
1419	free(buf);
1420	return count < 0 ? count : 0;
1421
1422err:
1423	count = -1;
1424	goto out;
1425}
1426
1427/*
1428 * Possibly create a new version of line with tabs expanded. Returns the
1429 * existing or new line, storage is updated if a new line is allocated. If
1430 * allocation fails then NULL is returned.
1431 */
1432char *expand_tabs(char *line, char **storage, size_t *storage_len)
1433{
1434	size_t i, src, dst, len, new_storage_len, num_tabs;
1435	char *new_line;
1436	size_t line_len = strlen(line);
1437
1438	for (num_tabs = 0, i = 0; i < line_len; i++)
1439		if (line[i] == '\t')
1440			num_tabs++;
1441
1442	if (num_tabs == 0)
1443		return line;
1444
1445	/*
1446	 * Space for the line and '\0', less the leading and trailing
1447	 * spaces. Each tab may introduce 7 additional spaces.
1448	 */
1449	new_storage_len = line_len + 1 + (num_tabs * 7);
1450
1451	new_line = malloc(new_storage_len);
1452	if (new_line == NULL) {
1453		pr_err("Failure allocating memory for tab expansion\n");
1454		return NULL;
1455	}
1456
1457	/*
1458	 * Copy regions starting at src and expand tabs. If there are two
1459	 * adjacent tabs then 'src == i', the memcpy is of size 0 and the spaces
1460	 * are inserted.
1461	 */
1462	for (i = 0, src = 0, dst = 0; i < line_len && num_tabs; i++) {
1463		if (line[i] == '\t') {
1464			len = i - src;
1465			memcpy(&new_line[dst], &line[src], len);
1466			dst += len;
1467			new_line[dst++] = ' ';
1468			while (dst % 8 != 0)
1469				new_line[dst++] = ' ';
1470			src = i + 1;
1471			num_tabs--;
1472		}
1473	}
1474
1475	/* Expand the last region. */
1476	len = line_len - src;
1477	memcpy(&new_line[dst], &line[src], len);
1478	dst += len;
1479	new_line[dst] = '\0';
1480
1481	free(*storage);
1482	*storage = new_line;
1483	*storage_len = new_storage_len;
1484	return new_line;
1485}
1486
1487static int symbol__disassemble_bpf_image(struct symbol *sym, struct annotate_args *args)
1488{
1489	struct annotation *notes = symbol__annotation(sym);
1490	struct disasm_line *dl;
1491
1492	args->offset = -1;
1493	args->line = strdup("to be implemented");
1494	args->line_nr = 0;
1495	args->fileloc = NULL;
1496	dl = disasm_line__new(args);
1497	if (dl)
1498		annotation_line__add(&dl->al, &notes->src->source);
1499
1500	zfree(&args->line);
1501	return 0;
1502}
1503
1504static int symbol__disassemble_objdump(const char *filename, struct symbol *sym,
1505				       struct annotate_args *args)
1506{
1507	struct annotation_options *opts = &annotate_opts;
1508	struct map *map = args->ms->map;
1509	struct dso *dso = map__dso(map);
1510	char *command;
1511	FILE *file;
1512	int lineno = 0;
1513	char *fileloc = NULL;
1514	int nline;
1515	char *line;
1516	size_t line_len;
1517	const char *objdump_argv[] = {
1518		"/bin/sh",
1519		"-c",
1520		NULL, /* Will be the objdump command to run. */
1521		"--",
1522		NULL, /* Will be the symfs path. */
1523		NULL,
1524	};
1525	struct child_process objdump_process;
1526	int err;
1527
1528	if (dso__binary_type(dso) == DSO_BINARY_TYPE__BPF_PROG_INFO)
1529		return symbol__disassemble_bpf_libbfd(sym, args);
1530
1531	if (dso__binary_type(dso) == DSO_BINARY_TYPE__BPF_IMAGE)
1532		return symbol__disassemble_bpf_image(sym, args);
1533
1534	err = asprintf(&command,
1535		 "%s %s%s --start-address=0x%016" PRIx64
1536		 " --stop-address=0x%016" PRIx64
1537		 " %s -d %s %s %s %c%s%c %s%s -C \"$1\"",
1538		 opts->objdump_path ?: "objdump",
1539		 opts->disassembler_style ? "-M " : "",
1540		 opts->disassembler_style ?: "",
1541		 map__rip_2objdump(map, sym->start),
1542		 map__rip_2objdump(map, sym->end),
1543		 opts->show_linenr ? "-l" : "",
1544		 opts->show_asm_raw ? "" : "--no-show-raw-insn",
1545		 opts->annotate_src ? "-S" : "",
1546		 opts->prefix ? "--prefix " : "",
1547		 opts->prefix ? '"' : ' ',
1548		 opts->prefix ?: "",
1549		 opts->prefix ? '"' : ' ',
1550		 opts->prefix_strip ? "--prefix-strip=" : "",
1551		 opts->prefix_strip ?: "");
1552
1553	if (err < 0) {
1554		pr_err("Failure allocating memory for the command to run\n");
1555		return err;
1556	}
1557
1558	pr_debug("Executing: %s\n", command);
1559
1560	objdump_argv[2] = command;
1561	objdump_argv[4] = filename;
1562
1563	/* Create a pipe to read from for stdout */
1564	memset(&objdump_process, 0, sizeof(objdump_process));
1565	objdump_process.argv = objdump_argv;
1566	objdump_process.out = -1;
1567	objdump_process.err = -1;
1568	objdump_process.no_stderr = 1;
1569	if (start_command(&objdump_process)) {
1570		pr_err("Failure starting to run %s\n", command);
1571		err = -1;
1572		goto out_free_command;
1573	}
1574
1575	file = fdopen(objdump_process.out, "r");
1576	if (!file) {
1577		pr_err("Failure creating FILE stream for %s\n", command);
1578		/*
1579		 * If we were using debug info should retry with
1580		 * original binary.
1581		 */
1582		err = -1;
1583		goto out_close_stdout;
1584	}
1585
1586	/* Storage for getline. */
1587	line = NULL;
1588	line_len = 0;
1589
1590	nline = 0;
1591	while (!feof(file)) {
1592		const char *match;
1593		char *expanded_line;
1594
1595		if (getline(&line, &line_len, file) < 0 || !line)
1596			break;
1597
1598		/* Skip lines containing "filename:" */
1599		match = strstr(line, filename);
1600		if (match && match[strlen(filename)] == ':')
1601			continue;
1602
1603		expanded_line = strim(line);
1604		expanded_line = expand_tabs(expanded_line, &line, &line_len);
1605		if (!expanded_line)
1606			break;
1607
1608		/*
1609		 * The source code line number (lineno) needs to be kept in
1610		 * across calls to symbol__parse_objdump_line(), so that it
1611		 * can associate it with the instructions till the next one.
1612		 * See disasm_line__new() and struct disasm_line::line_nr.
1613		 */
1614		if (symbol__parse_objdump_line(sym, args, expanded_line,
1615					       &lineno, &fileloc) < 0)
1616			break;
1617		nline++;
1618	}
1619	free(line);
1620	free(fileloc);
1621
1622	err = finish_command(&objdump_process);
1623	if (err)
1624		pr_err("Error running %s\n", command);
1625
1626	if (nline == 0) {
1627		err = -1;
1628		pr_err("No output from %s\n", command);
1629	}
1630
1631	/*
1632	 * kallsyms does not have symbol sizes so there may a nop at the end.
1633	 * Remove it.
1634	 */
1635	if (dso__is_kcore(dso))
1636		delete_last_nop(sym);
1637
1638	fclose(file);
1639
1640out_close_stdout:
1641	close(objdump_process.out);
1642
1643out_free_command:
1644	free(command);
1645	return err;
1646}
1647
1648int symbol__disassemble(struct symbol *sym, struct annotate_args *args)
1649{
1650	struct annotation_options *options = args->options;
1651	struct map *map = args->ms->map;
1652	struct dso *dso = map__dso(map);
1653	char symfs_filename[PATH_MAX];
1654	bool delete_extract = false;
1655	struct kcore_extract kce;
1656	bool decomp = false;
1657	int err = dso__disassemble_filename(dso, symfs_filename, sizeof(symfs_filename));
1658
1659	if (err)
1660		return err;
1661
1662	pr_debug("%s: filename=%s, sym=%s, start=%#" PRIx64 ", end=%#" PRIx64 "\n", __func__,
1663		 symfs_filename, sym->name, map__unmap_ip(map, sym->start),
1664		 map__unmap_ip(map, sym->end));
1665
1666	pr_debug("annotating [%p] %30s : [%p] %30s\n", dso, dso__long_name(dso), sym, sym->name);
1667
1668	if (dso__binary_type(dso) == DSO_BINARY_TYPE__NOT_FOUND) {
1669		return SYMBOL_ANNOTATE_ERRNO__COULDNT_DETERMINE_FILE_TYPE;
1670	} else if (dso__is_kcore(dso)) {
1671		kce.addr = map__rip_2objdump(map, sym->start);
1672		kce.kcore_filename = symfs_filename;
1673		kce.len = sym->end - sym->start;
1674		kce.offs = sym->start;
1675
1676		if (!kcore_extract__create(&kce)) {
1677			delete_extract = true;
1678			strlcpy(symfs_filename, kce.extract_filename, sizeof(symfs_filename));
1679		}
1680	} else if (dso__needs_decompress(dso)) {
1681		char tmp[KMOD_DECOMP_LEN];
1682
1683		if (dso__decompress_kmodule_path(dso, symfs_filename, tmp, sizeof(tmp)) < 0)
1684			return -1;
1685
1686		decomp = true;
1687		strcpy(symfs_filename, tmp);
1688	}
1689
1690	/*
1691	 * For powerpc data type profiling, use the dso__data_read_offset to
1692	 * read raw instruction directly and interpret the binary code to
1693	 * understand instructions and register fields. For sort keys as type
1694	 * and typeoff, disassemble to mnemonic notation is not required in
1695	 * case of powerpc.
1696	 */
1697	if (arch__is(args->arch, "powerpc")) {
1698		extern const char *sort_order;
1699
1700		if (sort_order && !strstr(sort_order, "sym")) {
1701			err = symbol__disassemble_raw(symfs_filename, sym, args);
1702			if (err == 0)
1703				goto out_remove_tmp;
1704
1705			err = symbol__disassemble_capstone_powerpc(symfs_filename, sym, args);
1706			if (err == 0)
1707				goto out_remove_tmp;
1708		}
1709	}
1710
1711	/* FIXME: LLVM and CAPSTONE should support source code */
1712	if (options->annotate_src && !options->hide_src_code) {
1713		err = symbol__disassemble_objdump(symfs_filename, sym, args);
1714		if (err == 0)
1715			goto out_remove_tmp;
1716	}
1717
1718	err = -1;
1719	for (u8 i = 0; i < ARRAY_SIZE(options->disassemblers) && err != 0; i++) {
1720		enum perf_disassembler dis = options->disassemblers[i];
1721
1722		switch (dis) {
1723		case PERF_DISASM_LLVM:
1724			args->options->disassembler_used = PERF_DISASM_LLVM;
1725			err = symbol__disassemble_llvm(symfs_filename, sym, args);
1726			break;
1727		case PERF_DISASM_CAPSTONE:
1728			args->options->disassembler_used = PERF_DISASM_CAPSTONE;
1729			err = symbol__disassemble_capstone(symfs_filename, sym, args);
1730			break;
1731		case PERF_DISASM_OBJDUMP:
1732			args->options->disassembler_used = PERF_DISASM_OBJDUMP;
1733			err = symbol__disassemble_objdump(symfs_filename, sym, args);
1734			break;
1735		case PERF_DISASM_UNKNOWN: /* End of disassemblers. */
1736		default:
1737			args->options->disassembler_used = PERF_DISASM_UNKNOWN;
1738			goto out_remove_tmp;
1739		}
1740		if (err == 0)
1741			pr_debug("Disassembled with %s\n", perf_disassembler__strs[dis]);
1742	}
1743out_remove_tmp:
1744	if (decomp)
1745		unlink(symfs_filename);
1746
1747	if (delete_extract)
1748		kcore_extract__delete(&kce);
1749
1750	return err;
1751}