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 <stdio.h> 22#include <stdlib.h> 23#include <string.h> 24#include <ctype.h> 25 26#ifndef ARRAY_SIZE 27#define ARRAY_SIZE(arr) (sizeof(arr) / sizeof(arr[0])) 28#endif 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 char *sym; 37}; 38 39struct addr_range { 40 const char *start_sym, *end_sym; 41 unsigned long long start, end; 42}; 43 44static unsigned long long _text; 45static struct addr_range text_ranges[] = { 46 { "_stext", "_etext" }, 47 { "_sinittext", "_einittext" }, 48 { "_stext_l1", "_etext_l1" }, /* Blackfin on-chip L1 inst SRAM */ 49 { "_stext_l2", "_etext_l2" }, /* Blackfin on-chip L2 SRAM */ 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 = 0; 61static int absolute_percpu = 0; 62static char symbol_prefix_char = '\0'; 63static unsigned long long kernel_start_addr = 0; 64 65int token_profit[0x10000]; 66 67/* the table that holds the result of the compression */ 68unsigned char best_table[256][2]; 69unsigned char best_table_len[256]; 70 71 72static void usage(void) 73{ 74 fprintf(stderr, "Usage: kallsyms [--all-symbols] " 75 "[--symbol-prefix=<prefix char>] " 76 "[--page-offset=<CONFIG_PAGE_OFFSET>] " 77 "< in.map > out.S\n"); 78 exit(1); 79} 80 81/* 82 * This ignores the intensely annoying "mapping symbols" found 83 * in ARM ELF files: $a, $t and $d. 84 */ 85static inline int is_arm_mapping_symbol(const char *str) 86{ 87 return str[0] == '$' && strchr("axtd", str[1]) 88 && (str[2] == '\0' || str[2] == '.'); 89} 90 91static int check_symbol_range(const char *sym, unsigned long long addr, 92 struct addr_range *ranges, int entries) 93{ 94 size_t i; 95 struct addr_range *ar; 96 97 for (i = 0; i < entries; ++i) { 98 ar = &ranges[i]; 99 100 if (strcmp(sym, ar->start_sym) == 0) { 101 ar->start = addr; 102 return 0; 103 } else if (strcmp(sym, ar->end_sym) == 0) { 104 ar->end = addr; 105 return 0; 106 } 107 } 108 109 return 1; 110} 111 112static int read_symbol(FILE *in, struct sym_entry *s) 113{ 114 char str[500]; 115 char *sym, stype; 116 int rc; 117 118 rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, str); 119 if (rc != 3) { 120 if (rc != EOF && fgets(str, 500, in) == NULL) 121 fprintf(stderr, "Read error or end of file.\n"); 122 return -1; 123 } 124 if (strlen(str) > KSYM_NAME_LEN) { 125 fprintf(stderr, "Symbol %s too long for kallsyms (%zu vs %d).\n" 126 "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n", 127 str, strlen(str), KSYM_NAME_LEN); 128 return -1; 129 } 130 131 sym = str; 132 /* skip prefix char */ 133 if (symbol_prefix_char && str[0] == symbol_prefix_char) 134 sym++; 135 136 /* Ignore most absolute/undefined (?) symbols. */ 137 if (strcmp(sym, "_text") == 0) 138 _text = s->addr; 139 else if (check_symbol_range(sym, s->addr, text_ranges, 140 ARRAY_SIZE(text_ranges)) == 0) 141 /* nothing to do */; 142 else if (toupper(stype) == 'A') 143 { 144 /* Keep these useful absolute symbols */ 145 if (strcmp(sym, "__kernel_syscall_via_break") && 146 strcmp(sym, "__kernel_syscall_via_epc") && 147 strcmp(sym, "__kernel_sigtramp") && 148 strcmp(sym, "__gp")) 149 return -1; 150 151 } 152 else if (toupper(stype) == 'U' || 153 is_arm_mapping_symbol(sym)) 154 return -1; 155 /* exclude also MIPS ELF local symbols ($L123 instead of .L123) */ 156 else if (str[0] == '$') 157 return -1; 158 /* exclude debugging symbols */ 159 else if (stype == 'N') 160 return -1; 161 162 /* include the type field in the symbol name, so that it gets 163 * compressed together */ 164 s->len = strlen(str) + 1; 165 s->sym = malloc(s->len + 1); 166 if (!s->sym) { 167 fprintf(stderr, "kallsyms failure: " 168 "unable to allocate required amount of memory\n"); 169 exit(EXIT_FAILURE); 170 } 171 strcpy((char *)s->sym + 1, str); 172 s->sym[0] = stype; 173 174 /* Record if we've found __per_cpu_start/end. */ 175 check_symbol_range(sym, s->addr, &percpu_range, 1); 176 177 return 0; 178} 179 180static int symbol_in_range(struct sym_entry *s, struct addr_range *ranges, 181 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 (s->addr >= ar->start && s->addr <= ar->end) 190 return 1; 191 } 192 193 return 0; 194} 195 196static int symbol_valid(struct sym_entry *s) 197{ 198 /* Symbols which vary between passes. Passes 1 and 2 must have 199 * identical symbol lists. The kallsyms_* symbols below are only added 200 * after pass 1, they would be included in pass 2 when --all-symbols is 201 * specified so exclude them to get a stable symbol list. 202 */ 203 static char *special_symbols[] = { 204 "kallsyms_addresses", 205 "kallsyms_num_syms", 206 "kallsyms_names", 207 "kallsyms_markers", 208 "kallsyms_token_table", 209 "kallsyms_token_index", 210 211 /* Exclude linker generated symbols which vary between passes */ 212 "_SDA_BASE_", /* ppc */ 213 "_SDA2_BASE_", /* ppc */ 214 NULL }; 215 216 static char *special_suffixes[] = { 217 "_veneer", /* arm */ 218 NULL }; 219 220 int i; 221 char *sym_name = (char *)s->sym + 1; 222 223 224 if (s->addr < kernel_start_addr) 225 return 0; 226 227 /* skip prefix char */ 228 if (symbol_prefix_char && *sym_name == symbol_prefix_char) 229 sym_name++; 230 231 232 /* if --all-symbols is not specified, then symbols outside the text 233 * and inittext sections are discarded */ 234 if (!all_symbols) { 235 if (symbol_in_range(s, text_ranges, 236 ARRAY_SIZE(text_ranges)) == 0) 237 return 0; 238 /* Corner case. Discard any symbols with the same value as 239 * _etext _einittext; they can move between pass 1 and 2 when 240 * the kallsyms data are added. If these symbols move then 241 * they may get dropped in pass 2, which breaks the kallsyms 242 * rules. 243 */ 244 if ((s->addr == text_range_text->end && 245 strcmp(sym_name, 246 text_range_text->end_sym)) || 247 (s->addr == text_range_inittext->end && 248 strcmp(sym_name, 249 text_range_inittext->end_sym))) 250 return 0; 251 } 252 253 /* Exclude symbols which vary between passes. */ 254 for (i = 0; special_symbols[i]; i++) 255 if (strcmp(sym_name, special_symbols[i]) == 0) 256 return 0; 257 258 for (i = 0; special_suffixes[i]; i++) { 259 int l = strlen(sym_name) - strlen(special_suffixes[i]); 260 261 if (l >= 0 && strcmp(sym_name + l, special_suffixes[i]) == 0) 262 return 0; 263 } 264 265 return 1; 266} 267 268static void read_map(FILE *in) 269{ 270 while (!feof(in)) { 271 if (table_cnt >= table_size) { 272 table_size += 10000; 273 table = realloc(table, sizeof(*table) * table_size); 274 if (!table) { 275 fprintf(stderr, "out of memory\n"); 276 exit (1); 277 } 278 } 279 if (read_symbol(in, &table[table_cnt]) == 0) { 280 table[table_cnt].start_pos = table_cnt; 281 table_cnt++; 282 } 283 } 284} 285 286static void output_label(char *label) 287{ 288 if (symbol_prefix_char) 289 printf(".globl %c%s\n", symbol_prefix_char, label); 290 else 291 printf(".globl %s\n", label); 292 printf("\tALGN\n"); 293 if (symbol_prefix_char) 294 printf("%c%s:\n", symbol_prefix_char, label); 295 else 296 printf("%s:\n", label); 297} 298 299/* uncompress a compressed symbol. When this function is called, the best table 300 * might still be compressed itself, so the function needs to be recursive */ 301static int expand_symbol(unsigned char *data, int len, char *result) 302{ 303 int c, rlen, total=0; 304 305 while (len) { 306 c = *data; 307 /* if the table holds a single char that is the same as the one 308 * we are looking for, then end the search */ 309 if (best_table[c][0]==c && best_table_len[c]==1) { 310 *result++ = c; 311 total++; 312 } else { 313 /* if not, recurse and expand */ 314 rlen = expand_symbol(best_table[c], best_table_len[c], result); 315 total += rlen; 316 result += rlen; 317 } 318 data++; 319 len--; 320 } 321 *result=0; 322 323 return total; 324} 325 326static int symbol_absolute(struct sym_entry *s) 327{ 328 return toupper(s->sym[0]) == 'A'; 329} 330 331static void write_src(void) 332{ 333 unsigned int i, k, off; 334 unsigned int best_idx[256]; 335 unsigned int *markers; 336 char buf[KSYM_NAME_LEN]; 337 338 printf("#include <asm/types.h>\n"); 339 printf("#if BITS_PER_LONG == 64\n"); 340 printf("#define PTR .quad\n"); 341 printf("#define ALGN .align 8\n"); 342 printf("#else\n"); 343 printf("#define PTR .long\n"); 344 printf("#define ALGN .align 4\n"); 345 printf("#endif\n"); 346 347 printf("\t.section .rodata, \"a\"\n"); 348 349 /* Provide proper symbols relocatability by their '_text' 350 * relativeness. The symbol names cannot be used to construct 351 * normal symbol references as the list of symbols contains 352 * symbols that are declared static and are private to their 353 * .o files. This prevents .tmp_kallsyms.o or any other 354 * object from referencing them. 355 */ 356 output_label("kallsyms_addresses"); 357 for (i = 0; i < table_cnt; i++) { 358 if (!symbol_absolute(&table[i])) { 359 if (_text <= table[i].addr) 360 printf("\tPTR\t_text + %#llx\n", 361 table[i].addr - _text); 362 else 363 printf("\tPTR\t_text - %#llx\n", 364 _text - table[i].addr); 365 } else { 366 printf("\tPTR\t%#llx\n", table[i].addr); 367 } 368 } 369 printf("\n"); 370 371 output_label("kallsyms_num_syms"); 372 printf("\tPTR\t%d\n", table_cnt); 373 printf("\n"); 374 375 /* table of offset markers, that give the offset in the compressed stream 376 * every 256 symbols */ 377 markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256)); 378 if (!markers) { 379 fprintf(stderr, "kallsyms failure: " 380 "unable to allocate required memory\n"); 381 exit(EXIT_FAILURE); 382 } 383 384 output_label("kallsyms_names"); 385 off = 0; 386 for (i = 0; i < table_cnt; i++) { 387 if ((i & 0xFF) == 0) 388 markers[i >> 8] = off; 389 390 printf("\t.byte 0x%02x", table[i].len); 391 for (k = 0; k < table[i].len; k++) 392 printf(", 0x%02x", table[i].sym[k]); 393 printf("\n"); 394 395 off += table[i].len + 1; 396 } 397 printf("\n"); 398 399 output_label("kallsyms_markers"); 400 for (i = 0; i < ((table_cnt + 255) >> 8); i++) 401 printf("\tPTR\t%d\n", markers[i]); 402 printf("\n"); 403 404 free(markers); 405 406 output_label("kallsyms_token_table"); 407 off = 0; 408 for (i = 0; i < 256; i++) { 409 best_idx[i] = off; 410 expand_symbol(best_table[i], best_table_len[i], buf); 411 printf("\t.asciz\t\"%s\"\n", buf); 412 off += strlen(buf) + 1; 413 } 414 printf("\n"); 415 416 output_label("kallsyms_token_index"); 417 for (i = 0; i < 256; i++) 418 printf("\t.short\t%d\n", best_idx[i]); 419 printf("\n"); 420} 421 422 423/* table lookup compression functions */ 424 425/* count all the possible tokens in a symbol */ 426static void learn_symbol(unsigned char *symbol, int len) 427{ 428 int i; 429 430 for (i = 0; i < len - 1; i++) 431 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++; 432} 433 434/* decrease the count for all the possible tokens in a symbol */ 435static void forget_symbol(unsigned char *symbol, int len) 436{ 437 int i; 438 439 for (i = 0; i < len - 1; i++) 440 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--; 441} 442 443/* remove all the invalid symbols from the table and do the initial token count */ 444static void build_initial_tok_table(void) 445{ 446 unsigned int i, pos; 447 448 pos = 0; 449 for (i = 0; i < table_cnt; i++) { 450 if ( symbol_valid(&table[i]) ) { 451 if (pos != i) 452 table[pos] = table[i]; 453 learn_symbol(table[pos].sym, table[pos].len); 454 pos++; 455 } 456 } 457 table_cnt = pos; 458} 459 460static void *find_token(unsigned char *str, int len, unsigned char *token) 461{ 462 int i; 463 464 for (i = 0; i < len - 1; i++) { 465 if (str[i] == token[0] && str[i+1] == token[1]) 466 return &str[i]; 467 } 468 return NULL; 469} 470 471/* replace a given token in all the valid symbols. Use the sampled symbols 472 * to update the counts */ 473static void compress_symbols(unsigned char *str, int idx) 474{ 475 unsigned int i, len, size; 476 unsigned char *p1, *p2; 477 478 for (i = 0; i < table_cnt; i++) { 479 480 len = table[i].len; 481 p1 = table[i].sym; 482 483 /* find the token on the symbol */ 484 p2 = find_token(p1, len, str); 485 if (!p2) continue; 486 487 /* decrease the counts for this symbol's tokens */ 488 forget_symbol(table[i].sym, len); 489 490 size = len; 491 492 do { 493 *p2 = idx; 494 p2++; 495 size -= (p2 - p1); 496 memmove(p2, p2 + 1, size); 497 p1 = p2; 498 len--; 499 500 if (size < 2) break; 501 502 /* find the token on the symbol */ 503 p2 = find_token(p1, size, str); 504 505 } while (p2); 506 507 table[i].len = len; 508 509 /* increase the counts for this symbol's new tokens */ 510 learn_symbol(table[i].sym, len); 511 } 512} 513 514/* search the token with the maximum profit */ 515static int find_best_token(void) 516{ 517 int i, best, bestprofit; 518 519 bestprofit=-10000; 520 best = 0; 521 522 for (i = 0; i < 0x10000; i++) { 523 if (token_profit[i] > bestprofit) { 524 best = i; 525 bestprofit = token_profit[i]; 526 } 527 } 528 return best; 529} 530 531/* this is the core of the algorithm: calculate the "best" table */ 532static void optimize_result(void) 533{ 534 int i, best; 535 536 /* using the '\0' symbol last allows compress_symbols to use standard 537 * fast string functions */ 538 for (i = 255; i >= 0; i--) { 539 540 /* if this table slot is empty (it is not used by an actual 541 * original char code */ 542 if (!best_table_len[i]) { 543 544 /* find the token with the breates profit value */ 545 best = find_best_token(); 546 if (token_profit[best] == 0) 547 break; 548 549 /* place it in the "best" table */ 550 best_table_len[i] = 2; 551 best_table[i][0] = best & 0xFF; 552 best_table[i][1] = (best >> 8) & 0xFF; 553 554 /* replace this token in all the valid symbols */ 555 compress_symbols(best_table[i], i); 556 } 557 } 558} 559 560/* start by placing the symbols that are actually used on the table */ 561static void insert_real_symbols_in_table(void) 562{ 563 unsigned int i, j, c; 564 565 memset(best_table, 0, sizeof(best_table)); 566 memset(best_table_len, 0, sizeof(best_table_len)); 567 568 for (i = 0; i < table_cnt; i++) { 569 for (j = 0; j < table[i].len; j++) { 570 c = table[i].sym[j]; 571 best_table[c][0]=c; 572 best_table_len[c]=1; 573 } 574 } 575} 576 577static void optimize_token_table(void) 578{ 579 build_initial_tok_table(); 580 581 insert_real_symbols_in_table(); 582 583 /* When valid symbol is not registered, exit to error */ 584 if (!table_cnt) { 585 fprintf(stderr, "No valid symbol.\n"); 586 exit(1); 587 } 588 589 optimize_result(); 590} 591 592/* guess for "linker script provide" symbol */ 593static int may_be_linker_script_provide_symbol(const struct sym_entry *se) 594{ 595 const char *symbol = (char *)se->sym + 1; 596 int len = se->len - 1; 597 598 if (len < 8) 599 return 0; 600 601 if (symbol[0] != '_' || symbol[1] != '_') 602 return 0; 603 604 /* __start_XXXXX */ 605 if (!memcmp(symbol + 2, "start_", 6)) 606 return 1; 607 608 /* __stop_XXXXX */ 609 if (!memcmp(symbol + 2, "stop_", 5)) 610 return 1; 611 612 /* __end_XXXXX */ 613 if (!memcmp(symbol + 2, "end_", 4)) 614 return 1; 615 616 /* __XXXXX_start */ 617 if (!memcmp(symbol + len - 6, "_start", 6)) 618 return 1; 619 620 /* __XXXXX_end */ 621 if (!memcmp(symbol + len - 4, "_end", 4)) 622 return 1; 623 624 return 0; 625} 626 627static int prefix_underscores_count(const char *str) 628{ 629 const char *tail = str; 630 631 while (*tail == '_') 632 tail++; 633 634 return tail - str; 635} 636 637static int compare_symbols(const void *a, const void *b) 638{ 639 const struct sym_entry *sa; 640 const struct sym_entry *sb; 641 int wa, wb; 642 643 sa = a; 644 sb = b; 645 646 /* sort by address first */ 647 if (sa->addr > sb->addr) 648 return 1; 649 if (sa->addr < sb->addr) 650 return -1; 651 652 /* sort by "weakness" type */ 653 wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W'); 654 wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W'); 655 if (wa != wb) 656 return wa - wb; 657 658 /* sort by "linker script provide" type */ 659 wa = may_be_linker_script_provide_symbol(sa); 660 wb = may_be_linker_script_provide_symbol(sb); 661 if (wa != wb) 662 return wa - wb; 663 664 /* sort by the number of prefix underscores */ 665 wa = prefix_underscores_count((const char *)sa->sym + 1); 666 wb = prefix_underscores_count((const char *)sb->sym + 1); 667 if (wa != wb) 668 return wa - wb; 669 670 /* sort by initial order, so that other symbols are left undisturbed */ 671 return sa->start_pos - sb->start_pos; 672} 673 674static void sort_symbols(void) 675{ 676 qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols); 677} 678 679static void make_percpus_absolute(void) 680{ 681 unsigned int i; 682 683 for (i = 0; i < table_cnt; i++) 684 if (symbol_in_range(&table[i], &percpu_range, 1)) 685 table[i].sym[0] = 'A'; 686} 687 688int main(int argc, char **argv) 689{ 690 if (argc >= 2) { 691 int i; 692 for (i = 1; i < argc; i++) { 693 if(strcmp(argv[i], "--all-symbols") == 0) 694 all_symbols = 1; 695 else if (strcmp(argv[i], "--absolute-percpu") == 0) 696 absolute_percpu = 1; 697 else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) { 698 char *p = &argv[i][16]; 699 /* skip quote */ 700 if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\'')) 701 p++; 702 symbol_prefix_char = *p; 703 } else if (strncmp(argv[i], "--page-offset=", 14) == 0) { 704 const char *p = &argv[i][14]; 705 kernel_start_addr = strtoull(p, NULL, 16); 706 } else 707 usage(); 708 } 709 } else if (argc != 1) 710 usage(); 711 712 read_map(stdin); 713 if (absolute_percpu) 714 make_percpus_absolute(); 715 sort_symbols(); 716 optimize_token_table(); 717 write_src(); 718 719 return 0; 720} 721