1/* Generate assembler source containing symbol information 2 * 3 * Copyright 2002 by Kai Germaschewski 4 * 5 * This software may be used and distributed according to the terms 6 * of the GNU General Public License, incorporated herein by reference. 7 * 8 * Usage: nm -n vmlinux | scripts/kallsyms [--all-symbols] > symbols.S 9 * 10 * Table compression uses all the unused char codes on the symbols and 11 * maps these to the most used substrings (tokens). For instance, it might 12 * map char code 0xF7 to represent "write_" and then in every symbol where 13 * "write_" appears it can be replaced by 0xF7, saving 5 bytes. 14 * The used codes themselves are also placed in the table so that the 15 * decompresion can work without "special cases". 16 * Applied to kernel symbols, this usually produces a compression ratio 17 * of about 50%. 18 * 19 */ 20 21#include <stdbool.h> 22#include <stdio.h> 23#include <stdlib.h> 24#include <string.h> 25#include <ctype.h> 26#include <limits.h> 27 28#define ARRAY_SIZE(arr) (sizeof(arr) / sizeof(arr[0])) 29 30#define KSYM_NAME_LEN 128 31 32struct sym_entry { 33 unsigned long long addr; 34 unsigned int len; 35 unsigned int start_pos; 36 unsigned int percpu_absolute; 37 unsigned char sym[]; 38}; 39 40struct addr_range { 41 const char *start_sym, *end_sym; 42 unsigned long long start, end; 43}; 44 45static unsigned long long _text; 46static unsigned long long relative_base; 47static struct addr_range text_ranges[] = { 48 { "_stext", "_etext" }, 49 { "_sinittext", "_einittext" }, 50}; 51#define text_range_text (&text_ranges[0]) 52#define text_range_inittext (&text_ranges[1]) 53 54static struct addr_range percpu_range = { 55 "__per_cpu_start", "__per_cpu_end", -1ULL, 0 56}; 57 58static struct sym_entry **table; 59static unsigned int table_size, table_cnt; 60static int all_symbols; 61static int absolute_percpu; 62static int base_relative; 63 64static int token_profit[0x10000]; 65 66/* the table that holds the result of the compression */ 67static unsigned char best_table[256][2]; 68static unsigned char best_table_len[256]; 69 70 71static void usage(void) 72{ 73 fprintf(stderr, "Usage: kallsyms [--all-symbols] " 74 "[--base-relative] < in.map > out.S\n"); 75 exit(1); 76} 77 78static char *sym_name(const struct sym_entry *s) 79{ 80 return (char *)s->sym + 1; 81} 82 83static bool is_ignored_symbol(const char *name, char type) 84{ 85 /* Symbol names that exactly match to the following are ignored.*/ 86 static const char * const ignored_symbols[] = { 87 /* 88 * Symbols which vary between passes. Passes 1 and 2 must have 89 * identical symbol lists. The kallsyms_* symbols below are 90 * only added after pass 1, they would be included in pass 2 91 * when --all-symbols is specified so exclude them to get a 92 * stable symbol list. 93 */ 94 "kallsyms_addresses", 95 "kallsyms_offsets", 96 "kallsyms_relative_base", 97 "kallsyms_num_syms", 98 "kallsyms_names", 99 "kallsyms_markers", 100 "kallsyms_token_table", 101 "kallsyms_token_index", 102 /* Exclude linker generated symbols which vary between passes */ 103 "_SDA_BASE_", /* ppc */ 104 "_SDA2_BASE_", /* ppc */ 105 NULL 106 }; 107 108 /* Symbol names that begin with the following are ignored.*/ 109 static const char * const ignored_prefixes[] = { 110 "$", /* local symbols for ARM, MIPS, etc. */ 111 "L0", /* LoongArch local symbols */ 112 ".L", /* LoongArch/S390 local symbols */ 113 "__crc_", /* modversions */ 114 "__efistub_", /* arm64 EFI stub namespace */ 115 "__kvm_nvhe_", /* arm64 non-VHE KVM namespace */ 116 "__AArch64ADRPThunk_", /* arm64 lld */ 117 "__ARMV5PILongThunk_", /* arm lld */ 118 "__ARMV7PILongThunk_", 119 "__ThumbV7PILongThunk_", 120 "__LA25Thunk_", /* mips lld */ 121 "__microLA25Thunk_", 122 NULL 123 }; 124 125 /* Symbol names that end with the following are ignored.*/ 126 static const char * const ignored_suffixes[] = { 127 "_from_arm", /* arm */ 128 "_from_thumb", /* arm */ 129 "_veneer", /* arm */ 130 NULL 131 }; 132 133 /* Symbol names that contain the following are ignored.*/ 134 static const char * const ignored_matches[] = { 135 ".long_branch.", /* ppc stub */ 136 ".plt_branch.", /* ppc stub */ 137 NULL 138 }; 139 140 const char * const *p; 141 142 for (p = ignored_symbols; *p; p++) 143 if (!strcmp(name, *p)) 144 return true; 145 146 for (p = ignored_prefixes; *p; p++) 147 if (!strncmp(name, *p, strlen(*p))) 148 return true; 149 150 for (p = ignored_suffixes; *p; p++) { 151 int l = strlen(name) - strlen(*p); 152 153 if (l >= 0 && !strcmp(name + l, *p)) 154 return true; 155 } 156 157 for (p = ignored_matches; *p; p++) { 158 if (strstr(name, *p)) 159 return true; 160 } 161 162 if (type == 'U' || type == 'u') 163 return true; 164 /* exclude debugging symbols */ 165 if (type == 'N' || type == 'n') 166 return true; 167 168 if (toupper(type) == 'A') { 169 /* Keep these useful absolute symbols */ 170 if (strcmp(name, "__kernel_syscall_via_break") && 171 strcmp(name, "__kernel_syscall_via_epc") && 172 strcmp(name, "__kernel_sigtramp") && 173 strcmp(name, "__gp")) 174 return true; 175 } 176 177 return false; 178} 179 180static void check_symbol_range(const char *sym, unsigned long long addr, 181 struct addr_range *ranges, int entries) 182{ 183 size_t i; 184 struct addr_range *ar; 185 186 for (i = 0; i < entries; ++i) { 187 ar = &ranges[i]; 188 189 if (strcmp(sym, ar->start_sym) == 0) { 190 ar->start = addr; 191 return; 192 } else if (strcmp(sym, ar->end_sym) == 0) { 193 ar->end = addr; 194 return; 195 } 196 } 197} 198 199static struct sym_entry *read_symbol(FILE *in) 200{ 201 char name[500], type; 202 unsigned long long addr; 203 unsigned int len; 204 struct sym_entry *sym; 205 int rc; 206 207 rc = fscanf(in, "%llx %c %499s\n", &addr, &type, name); 208 if (rc != 3) { 209 if (rc != EOF && fgets(name, 500, in) == NULL) 210 fprintf(stderr, "Read error or end of file.\n"); 211 return NULL; 212 } 213 if (strlen(name) >= KSYM_NAME_LEN) { 214 fprintf(stderr, "Symbol %s too long for kallsyms (%zu >= %d).\n" 215 "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n", 216 name, strlen(name), KSYM_NAME_LEN); 217 return NULL; 218 } 219 220 if (strcmp(name, "_text") == 0) 221 _text = addr; 222 223 /* Ignore most absolute/undefined (?) symbols. */ 224 if (is_ignored_symbol(name, type)) 225 return NULL; 226 227 check_symbol_range(name, addr, text_ranges, ARRAY_SIZE(text_ranges)); 228 check_symbol_range(name, addr, &percpu_range, 1); 229 230 /* include the type field in the symbol name, so that it gets 231 * compressed together */ 232 233 len = strlen(name) + 1; 234 235 sym = malloc(sizeof(*sym) + len + 1); 236 if (!sym) { 237 fprintf(stderr, "kallsyms failure: " 238 "unable to allocate required amount of memory\n"); 239 exit(EXIT_FAILURE); 240 } 241 sym->addr = addr; 242 sym->len = len; 243 sym->sym[0] = type; 244 strcpy(sym_name(sym), name); 245 sym->percpu_absolute = 0; 246 247 return sym; 248} 249 250static int symbol_in_range(const struct sym_entry *s, 251 const struct addr_range *ranges, int entries) 252{ 253 size_t i; 254 const struct addr_range *ar; 255 256 for (i = 0; i < entries; ++i) { 257 ar = &ranges[i]; 258 259 if (s->addr >= ar->start && s->addr <= ar->end) 260 return 1; 261 } 262 263 return 0; 264} 265 266static int symbol_valid(const struct sym_entry *s) 267{ 268 const char *name = sym_name(s); 269 270 /* if --all-symbols is not specified, then symbols outside the text 271 * and inittext sections are discarded */ 272 if (!all_symbols) { 273 if (symbol_in_range(s, text_ranges, 274 ARRAY_SIZE(text_ranges)) == 0) 275 return 0; 276 /* Corner case. Discard any symbols with the same value as 277 * _etext _einittext; they can move between pass 1 and 2 when 278 * the kallsyms data are added. If these symbols move then 279 * they may get dropped in pass 2, which breaks the kallsyms 280 * rules. 281 */ 282 if ((s->addr == text_range_text->end && 283 strcmp(name, text_range_text->end_sym)) || 284 (s->addr == text_range_inittext->end && 285 strcmp(name, text_range_inittext->end_sym))) 286 return 0; 287 } 288 289 return 1; 290} 291 292/* remove all the invalid symbols from the table */ 293static void shrink_table(void) 294{ 295 unsigned int i, pos; 296 297 pos = 0; 298 for (i = 0; i < table_cnt; i++) { 299 if (symbol_valid(table[i])) { 300 if (pos != i) 301 table[pos] = table[i]; 302 pos++; 303 } else { 304 free(table[i]); 305 } 306 } 307 table_cnt = pos; 308 309 /* When valid symbol is not registered, exit to error */ 310 if (!table_cnt) { 311 fprintf(stderr, "No valid symbol.\n"); 312 exit(1); 313 } 314} 315 316static void read_map(FILE *in) 317{ 318 struct sym_entry *sym; 319 320 while (!feof(in)) { 321 sym = read_symbol(in); 322 if (!sym) 323 continue; 324 325 sym->start_pos = table_cnt; 326 327 if (table_cnt >= table_size) { 328 table_size += 10000; 329 table = realloc(table, sizeof(*table) * table_size); 330 if (!table) { 331 fprintf(stderr, "out of memory\n"); 332 exit (1); 333 } 334 } 335 336 table[table_cnt++] = sym; 337 } 338} 339 340static void output_label(const char *label) 341{ 342 printf(".globl %s\n", label); 343 printf("\tALGN\n"); 344 printf("%s:\n", label); 345} 346 347/* Provide proper symbols relocatability by their '_text' relativeness. */ 348static void output_address(unsigned long long addr) 349{ 350 if (_text <= addr) 351 printf("\tPTR\t_text + %#llx\n", addr - _text); 352 else 353 printf("\tPTR\t_text - %#llx\n", _text - addr); 354} 355 356/* uncompress a compressed symbol. When this function is called, the best table 357 * might still be compressed itself, so the function needs to be recursive */ 358static int expand_symbol(const unsigned char *data, int len, char *result) 359{ 360 int c, rlen, total=0; 361 362 while (len) { 363 c = *data; 364 /* if the table holds a single char that is the same as the one 365 * we are looking for, then end the search */ 366 if (best_table[c][0]==c && best_table_len[c]==1) { 367 *result++ = c; 368 total++; 369 } else { 370 /* if not, recurse and expand */ 371 rlen = expand_symbol(best_table[c], best_table_len[c], result); 372 total += rlen; 373 result += rlen; 374 } 375 data++; 376 len--; 377 } 378 *result=0; 379 380 return total; 381} 382 383static int symbol_absolute(const struct sym_entry *s) 384{ 385 return s->percpu_absolute; 386} 387 388static void write_src(void) 389{ 390 unsigned int i, k, off; 391 unsigned int best_idx[256]; 392 unsigned int *markers; 393 char buf[KSYM_NAME_LEN]; 394 395 printf("#include <asm/bitsperlong.h>\n"); 396 printf("#if BITS_PER_LONG == 64\n"); 397 printf("#define PTR .quad\n"); 398 printf("#define ALGN .balign 8\n"); 399 printf("#else\n"); 400 printf("#define PTR .long\n"); 401 printf("#define ALGN .balign 4\n"); 402 printf("#endif\n"); 403 404 printf("\t.section .rodata, \"a\"\n"); 405 406 if (!base_relative) 407 output_label("kallsyms_addresses"); 408 else 409 output_label("kallsyms_offsets"); 410 411 for (i = 0; i < table_cnt; i++) { 412 if (base_relative) { 413 /* 414 * Use the offset relative to the lowest value 415 * encountered of all relative symbols, and emit 416 * non-relocatable fixed offsets that will be fixed 417 * up at runtime. 418 */ 419 420 long long offset; 421 int overflow; 422 423 if (!absolute_percpu) { 424 offset = table[i]->addr - relative_base; 425 overflow = (offset < 0 || offset > UINT_MAX); 426 } else if (symbol_absolute(table[i])) { 427 offset = table[i]->addr; 428 overflow = (offset < 0 || offset > INT_MAX); 429 } else { 430 offset = relative_base - table[i]->addr - 1; 431 overflow = (offset < INT_MIN || offset >= 0); 432 } 433 if (overflow) { 434 fprintf(stderr, "kallsyms failure: " 435 "%s symbol value %#llx out of range in relative mode\n", 436 symbol_absolute(table[i]) ? "absolute" : "relative", 437 table[i]->addr); 438 exit(EXIT_FAILURE); 439 } 440 printf("\t.long\t%#x\n", (int)offset); 441 } else if (!symbol_absolute(table[i])) { 442 output_address(table[i]->addr); 443 } else { 444 printf("\tPTR\t%#llx\n", table[i]->addr); 445 } 446 } 447 printf("\n"); 448 449 if (base_relative) { 450 output_label("kallsyms_relative_base"); 451 output_address(relative_base); 452 printf("\n"); 453 } 454 455 output_label("kallsyms_num_syms"); 456 printf("\t.long\t%u\n", table_cnt); 457 printf("\n"); 458 459 /* table of offset markers, that give the offset in the compressed stream 460 * every 256 symbols */ 461 markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256)); 462 if (!markers) { 463 fprintf(stderr, "kallsyms failure: " 464 "unable to allocate required memory\n"); 465 exit(EXIT_FAILURE); 466 } 467 468 output_label("kallsyms_names"); 469 off = 0; 470 for (i = 0; i < table_cnt; i++) { 471 if ((i & 0xFF) == 0) 472 markers[i >> 8] = off; 473 474 printf("\t.byte 0x%02x", table[i]->len); 475 for (k = 0; k < table[i]->len; k++) 476 printf(", 0x%02x", table[i]->sym[k]); 477 printf("\n"); 478 479 off += table[i]->len + 1; 480 } 481 printf("\n"); 482 483 output_label("kallsyms_markers"); 484 for (i = 0; i < ((table_cnt + 255) >> 8); i++) 485 printf("\t.long\t%u\n", markers[i]); 486 printf("\n"); 487 488 free(markers); 489 490 output_label("kallsyms_token_table"); 491 off = 0; 492 for (i = 0; i < 256; i++) { 493 best_idx[i] = off; 494 expand_symbol(best_table[i], best_table_len[i], buf); 495 printf("\t.asciz\t\"%s\"\n", buf); 496 off += strlen(buf) + 1; 497 } 498 printf("\n"); 499 500 output_label("kallsyms_token_index"); 501 for (i = 0; i < 256; i++) 502 printf("\t.short\t%d\n", best_idx[i]); 503 printf("\n"); 504} 505 506 507/* table lookup compression functions */ 508 509/* count all the possible tokens in a symbol */ 510static void learn_symbol(const unsigned char *symbol, int len) 511{ 512 int i; 513 514 for (i = 0; i < len - 1; i++) 515 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++; 516} 517 518/* decrease the count for all the possible tokens in a symbol */ 519static void forget_symbol(const unsigned char *symbol, int len) 520{ 521 int i; 522 523 for (i = 0; i < len - 1; i++) 524 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--; 525} 526 527/* do the initial token count */ 528static void build_initial_tok_table(void) 529{ 530 unsigned int i; 531 532 for (i = 0; i < table_cnt; i++) 533 learn_symbol(table[i]->sym, table[i]->len); 534} 535 536static unsigned char *find_token(unsigned char *str, int len, 537 const unsigned char *token) 538{ 539 int i; 540 541 for (i = 0; i < len - 1; i++) { 542 if (str[i] == token[0] && str[i+1] == token[1]) 543 return &str[i]; 544 } 545 return NULL; 546} 547 548/* replace a given token in all the valid symbols. Use the sampled symbols 549 * to update the counts */ 550static void compress_symbols(const unsigned char *str, int idx) 551{ 552 unsigned int i, len, size; 553 unsigned char *p1, *p2; 554 555 for (i = 0; i < table_cnt; i++) { 556 557 len = table[i]->len; 558 p1 = table[i]->sym; 559 560 /* find the token on the symbol */ 561 p2 = find_token(p1, len, str); 562 if (!p2) continue; 563 564 /* decrease the counts for this symbol's tokens */ 565 forget_symbol(table[i]->sym, len); 566 567 size = len; 568 569 do { 570 *p2 = idx; 571 p2++; 572 size -= (p2 - p1); 573 memmove(p2, p2 + 1, size); 574 p1 = p2; 575 len--; 576 577 if (size < 2) break; 578 579 /* find the token on the symbol */ 580 p2 = find_token(p1, size, str); 581 582 } while (p2); 583 584 table[i]->len = len; 585 586 /* increase the counts for this symbol's new tokens */ 587 learn_symbol(table[i]->sym, len); 588 } 589} 590 591/* search the token with the maximum profit */ 592static int find_best_token(void) 593{ 594 int i, best, bestprofit; 595 596 bestprofit=-10000; 597 best = 0; 598 599 for (i = 0; i < 0x10000; i++) { 600 if (token_profit[i] > bestprofit) { 601 best = i; 602 bestprofit = token_profit[i]; 603 } 604 } 605 return best; 606} 607 608/* this is the core of the algorithm: calculate the "best" table */ 609static void optimize_result(void) 610{ 611 int i, best; 612 613 /* using the '\0' symbol last allows compress_symbols to use standard 614 * fast string functions */ 615 for (i = 255; i >= 0; i--) { 616 617 /* if this table slot is empty (it is not used by an actual 618 * original char code */ 619 if (!best_table_len[i]) { 620 621 /* find the token with the best profit value */ 622 best = find_best_token(); 623 if (token_profit[best] == 0) 624 break; 625 626 /* place it in the "best" table */ 627 best_table_len[i] = 2; 628 best_table[i][0] = best & 0xFF; 629 best_table[i][1] = (best >> 8) & 0xFF; 630 631 /* replace this token in all the valid symbols */ 632 compress_symbols(best_table[i], i); 633 } 634 } 635} 636 637/* start by placing the symbols that are actually used on the table */ 638static void insert_real_symbols_in_table(void) 639{ 640 unsigned int i, j, c; 641 642 for (i = 0; i < table_cnt; i++) { 643 for (j = 0; j < table[i]->len; j++) { 644 c = table[i]->sym[j]; 645 best_table[c][0]=c; 646 best_table_len[c]=1; 647 } 648 } 649} 650 651static void optimize_token_table(void) 652{ 653 build_initial_tok_table(); 654 655 insert_real_symbols_in_table(); 656 657 optimize_result(); 658} 659 660/* guess for "linker script provide" symbol */ 661static int may_be_linker_script_provide_symbol(const struct sym_entry *se) 662{ 663 const char *symbol = sym_name(se); 664 int len = se->len - 1; 665 666 if (len < 8) 667 return 0; 668 669 if (symbol[0] != '_' || symbol[1] != '_') 670 return 0; 671 672 /* __start_XXXXX */ 673 if (!memcmp(symbol + 2, "start_", 6)) 674 return 1; 675 676 /* __stop_XXXXX */ 677 if (!memcmp(symbol + 2, "stop_", 5)) 678 return 1; 679 680 /* __end_XXXXX */ 681 if (!memcmp(symbol + 2, "end_", 4)) 682 return 1; 683 684 /* __XXXXX_start */ 685 if (!memcmp(symbol + len - 6, "_start", 6)) 686 return 1; 687 688 /* __XXXXX_end */ 689 if (!memcmp(symbol + len - 4, "_end", 4)) 690 return 1; 691 692 return 0; 693} 694 695static int compare_symbols(const void *a, const void *b) 696{ 697 const struct sym_entry *sa = *(const struct sym_entry **)a; 698 const struct sym_entry *sb = *(const struct sym_entry **)b; 699 int wa, wb; 700 701 /* sort by address first */ 702 if (sa->addr > sb->addr) 703 return 1; 704 if (sa->addr < sb->addr) 705 return -1; 706 707 /* sort by "weakness" type */ 708 wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W'); 709 wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W'); 710 if (wa != wb) 711 return wa - wb; 712 713 /* sort by "linker script provide" type */ 714 wa = may_be_linker_script_provide_symbol(sa); 715 wb = may_be_linker_script_provide_symbol(sb); 716 if (wa != wb) 717 return wa - wb; 718 719 /* sort by the number of prefix underscores */ 720 wa = strspn(sym_name(sa), "_"); 721 wb = strspn(sym_name(sb), "_"); 722 if (wa != wb) 723 return wa - wb; 724 725 /* sort by initial order, so that other symbols are left undisturbed */ 726 return sa->start_pos - sb->start_pos; 727} 728 729static void sort_symbols(void) 730{ 731 qsort(table, table_cnt, sizeof(table[0]), compare_symbols); 732} 733 734static void make_percpus_absolute(void) 735{ 736 unsigned int i; 737 738 for (i = 0; i < table_cnt; i++) 739 if (symbol_in_range(table[i], &percpu_range, 1)) { 740 /* 741 * Keep the 'A' override for percpu symbols to 742 * ensure consistent behavior compared to older 743 * versions of this tool. 744 */ 745 table[i]->sym[0] = 'A'; 746 table[i]->percpu_absolute = 1; 747 } 748} 749 750/* find the minimum non-absolute symbol address */ 751static void record_relative_base(void) 752{ 753 unsigned int i; 754 755 for (i = 0; i < table_cnt; i++) 756 if (!symbol_absolute(table[i])) { 757 /* 758 * The table is sorted by address. 759 * Take the first non-absolute symbol value. 760 */ 761 relative_base = table[i]->addr; 762 return; 763 } 764} 765 766int main(int argc, char **argv) 767{ 768 if (argc >= 2) { 769 int i; 770 for (i = 1; i < argc; i++) { 771 if(strcmp(argv[i], "--all-symbols") == 0) 772 all_symbols = 1; 773 else if (strcmp(argv[i], "--absolute-percpu") == 0) 774 absolute_percpu = 1; 775 else if (strcmp(argv[i], "--base-relative") == 0) 776 base_relative = 1; 777 else 778 usage(); 779 } 780 } else if (argc != 1) 781 usage(); 782 783 read_map(stdin); 784 shrink_table(); 785 if (absolute_percpu) 786 make_percpus_absolute(); 787 sort_symbols(); 788 if (base_relative) 789 record_relative_base(); 790 optimize_token_table(); 791 write_src(); 792 793 return 0; 794} 795