scripts/kallsyms.c

   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
  32 struct sym_entry {
  33         unsigned long long addr;
  34         unsigned int len;
  35         unsigned int start_pos;
  36         unsigned char *sym;
  37 };
  38
  39 struct addr_range {
  40         const char *start_sym, *end_sym;
  41         unsigned long long start, end;
  42 };
  43
  44 static unsigned long long _text;
  45 static 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
  54 static struct addr_range percpu_range = {
  55         "__per_cpu_start", "__per_cpu_end", -1ULL, 0
  56 };
  57
  58 static struct sym_entry *table;
  59 static unsigned int table_size, table_cnt;
  60 static int all_symbols = 0;
  61 static int absolute_percpu = 0;
  62 static char symbol_prefix_char = '\0';
  63 static unsigned long long kernel_start_addr = 0;
  64
  65 int token_profit[0x10000];
  66
  67 /* the table that holds the result of the compression */
  68 unsigned char best_table[256][2];
  69 unsigned char best_table_len[256];
  70
  71
  72 static 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 $d.
  84  */
  85 static inline int is_arm_mapping_symbol(const char *str)
  86 {
  87         return str[0] == '$' && strchr("axtd", str[1])
  88                && (str[2] == '\0' || str[2] == '.');
  89 }
  90
  91 static 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
 112 static 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
 180 static 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
 196 static 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         int i;
 216         int offset = 1;
 217
 218         if (s->addr < kernel_start_addr)
 219                 return 0;
 220
 221         /* skip prefix char */
 222         if (symbol_prefix_char && *(s->sym + 1) == symbol_prefix_char)
 223                 offset++;
 224
 225         /* if --all-symbols is not specified, then symbols outside the text
 226          * and inittext sections are discarded */
 227         if (!all_symbols) {
 228                 if (symbol_in_range(s, text_ranges,
 229                                     ARRAY_SIZE(text_ranges)) == 0)
 230                         return 0;
 231                 /* Corner case.  Discard any symbols with the same value as
 232                  * _etext _einittext; they can move between pass 1 and 2 when
 233                  * the kallsyms data are added.  If these symbols move then
 234                  * they may get dropped in pass 2, which breaks the kallsyms
 235                  * rules.
 236                  */
 237                 if ((s->addr == text_range_text->end &&
 238                                 strcmp((char *)s->sym + offset,
 239                                        text_range_text->end_sym)) ||
 240                     (s->addr == text_range_inittext->end &&
 241                                 strcmp((char *)s->sym + offset,
 242                                        text_range_inittext->end_sym)))
 243                         return 0;
 244         }
 245
 246         /* Exclude symbols which vary between passes. */
 247         if (strstr((char *)s->sym + offset, "_compiled."))
 248                 return 0;
 249
 250         for (i = 0; special_symbols[i]; i++)
 251                 if( strcmp((char *)s->sym + offset, special_symbols[i]) == 0 )
 252                         return 0;
 253
 254         return 1;
 255 }
 256
 257 static void read_map(FILE *in)
 258 {
 259         while (!feof(in)) {
 260                 if (table_cnt >= table_size) {
 261                         table_size += 10000;
 262                         table = realloc(table, sizeof(*table) * table_size);
 263                         if (!table) {
 264                                 fprintf(stderr, "out of memory\n");
 265                                 exit (1);
 266                         }
 267                 }
 268                 if (read_symbol(in, &table[table_cnt]) == 0) {
 269                         table[table_cnt].start_pos = table_cnt;
 270                         table_cnt++;
 271                 }
 272         }
 273 }
 274
 275 static void output_label(char *label)
 276 {
 277         if (symbol_prefix_char)
 278                 printf(".globl %c%s\n", symbol_prefix_char, label);
 279         else
 280                 printf(".globl %s\n", label);
 281         printf("\tALGN\n");
 282         if (symbol_prefix_char)
 283                 printf("%c%s:\n", symbol_prefix_char, label);
 284         else
 285                 printf("%s:\n", label);
 286 }
 287
 288 /* uncompress a compressed symbol. When this function is called, the best table
 289  * might still be compressed itself, so the function needs to be recursive */
 290 static int expand_symbol(unsigned char *data, int len, char *result)
 291 {
 292         int c, rlen, total=0;
 293
 294         while (len) {
 295                 c = *data;
 296                 /* if the table holds a single char that is the same as the one
 297                  * we are looking for, then end the search */
 298                 if (best_table[c][0]==c && best_table_len[c]==1) {
 299                         *result++ = c;
 300                         total++;
 301                 } else {
 302                         /* if not, recurse and expand */
 303                         rlen = expand_symbol(best_table[c], best_table_len[c], result);
 304                         total += rlen;
 305                         result += rlen;
 306                 }
 307                 data++;
 308                 len--;
 309         }
 310         *result=0;
 311
 312         return total;
 313 }
 314
 315 static int symbol_absolute(struct sym_entry *s)
 316 {
 317         return toupper(s->sym[0]) == 'A';
 318 }
 319
 320 static void write_src(void)
 321 {
 322         unsigned int i, k, off;
 323         unsigned int best_idx[256];
 324         unsigned int *markers;
 325         char buf[KSYM_NAME_LEN];
 326
 327         printf("#include <asm/types.h>\n");
 328         printf("#if BITS_PER_LONG == 64\n");
 329         printf("#define PTR .quad\n");
 330         printf("#define ALGN .align 8\n");
 331         printf("#else\n");
 332         printf("#define PTR .long\n");
 333         printf("#define ALGN .align 4\n");
 334         printf("#endif\n");
 335
 336         printf("\t.section .rodata, \"a\"\n");
 337
 338         /* Provide proper symbols relocatability by their '_text'
 339          * relativeness.  The symbol names cannot be used to construct
 340          * normal symbol references as the list of symbols contains
 341          * symbols that are declared static and are private to their
 342          * .o files.  This prevents .tmp_kallsyms.o or any other
 343          * object from referencing them.
 344          */
 345         output_label("kallsyms_addresses");
 346         for (i = 0; i < table_cnt; i++) {
 347                 if (!symbol_absolute(&table[i])) {
 348                         if (_text <= table[i].addr)
 349                                 printf("\tPTR\t_text + %#llx\n",
 350                                         table[i].addr - _text);
 351                         else
 352                                 printf("\tPTR\t_text - %#llx\n",
 353                                         _text - table[i].addr);
 354                 } else {
 355                         printf("\tPTR\t%#llx\n", table[i].addr);
 356                 }
 357         }
 358         printf("\n");
 359
 360         output_label("kallsyms_num_syms");
 361         printf("\tPTR\t%d\n", table_cnt);
 362         printf("\n");
 363
 364         /* table of offset markers, that give the offset in the compressed stream
 365          * every 256 symbols */
 366         markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
 367         if (!markers) {
 368                 fprintf(stderr, "kallsyms failure: "
 369                         "unable to allocate required memory\n");
 370                 exit(EXIT_FAILURE);
 371         }
 372
 373         output_label("kallsyms_names");
 374         off = 0;
 375         for (i = 0; i < table_cnt; i++) {
 376                 if ((i & 0xFF) == 0)
 377                         markers[i >> 8] = off;
 378
 379                 printf("\t.byte 0x%02x", table[i].len);
 380                 for (k = 0; k < table[i].len; k++)
 381                         printf(", 0x%02x", table[i].sym[k]);
 382                 printf("\n");
 383
 384                 off += table[i].len + 1;
 385         }
 386         printf("\n");
 387
 388         output_label("kallsyms_markers");
 389         for (i = 0; i < ((table_cnt + 255) >> 8); i++)
 390                 printf("\tPTR\t%d\n", markers[i]);
 391         printf("\n");
 392
 393         free(markers);
 394
 395         output_label("kallsyms_token_table");
 396         off = 0;
 397         for (i = 0; i < 256; i++) {
 398                 best_idx[i] = off;
 399                 expand_symbol(best_table[i], best_table_len[i], buf);
 400                 printf("\t.asciz\t\"%s\"\n", buf);
 401                 off += strlen(buf) + 1;
 402         }
 403         printf("\n");
 404
 405         output_label("kallsyms_token_index");
 406         for (i = 0; i < 256; i++)
 407                 printf("\t.short\t%d\n", best_idx[i]);
 408         printf("\n");
 409 }
 410
 411
 412 /* table lookup compression functions */
 413
 414 /* count all the possible tokens in a symbol */
 415 static void learn_symbol(unsigned char *symbol, int len)
 416 {
 417         int i;
 418
 419         for (i = 0; i < len - 1; i++)
 420                 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
 421 }
 422
 423 /* decrease the count for all the possible tokens in a symbol */
 424 static void forget_symbol(unsigned char *symbol, int len)
 425 {
 426         int i;
 427
 428         for (i = 0; i < len - 1; i++)
 429                 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
 430 }
 431
 432 /* remove all the invalid symbols from the table and do the initial token count */
 433 static void build_initial_tok_table(void)
 434 {
 435         unsigned int i, pos;
 436
 437         pos = 0;
 438         for (i = 0; i < table_cnt; i++) {
 439                 if ( symbol_valid(&table[i]) ) {
 440                         if (pos != i)
 441                                 table[pos] = table[i];
 442                         learn_symbol(table[pos].sym, table[pos].len);
 443                         pos++;
 444                 }
 445         }
 446         table_cnt = pos;
 447 }
 448
 449 static void *find_token(unsigned char *str, int len, unsigned char *token)
 450 {
 451         int i;
 452
 453         for (i = 0; i < len - 1; i++) {
 454                 if (str[i] == token[0] && str[i+1] == token[1])
 455                         return &str[i];
 456         }
 457         return NULL;
 458 }
 459
 460 /* replace a given token in all the valid symbols. Use the sampled symbols
 461  * to update the counts */
 462 static void compress_symbols(unsigned char *str, int idx)
 463 {
 464         unsigned int i, len, size;
 465         unsigned char *p1, *p2;
 466
 467         for (i = 0; i < table_cnt; i++) {
 468
 469                 len = table[i].len;
 470                 p1 = table[i].sym;
 471
 472                 /* find the token on the symbol */
 473                 p2 = find_token(p1, len, str);
 474                 if (!p2) continue;
 475
 476                 /* decrease the counts for this symbol's tokens */
 477                 forget_symbol(table[i].sym, len);
 478
 479                 size = len;
 480
 481                 do {
 482                         *p2 = idx;
 483                         p2++;
 484                         size -= (p2 - p1);
 485                         memmove(p2, p2 + 1, size);
 486                         p1 = p2;
 487                         len--;
 488
 489                         if (size < 2) break;
 490
 491                         /* find the token on the symbol */
 492                         p2 = find_token(p1, size, str);
 493
 494                 } while (p2);
 495
 496                 table[i].len = len;
 497
 498                 /* increase the counts for this symbol's new tokens */
 499                 learn_symbol(table[i].sym, len);
 500         }
 501 }
 502
 503 /* search the token with the maximum profit */
 504 static int find_best_token(void)
 505 {
 506         int i, best, bestprofit;
 507
 508         bestprofit=-10000;
 509         best = 0;
 510
 511         for (i = 0; i < 0x10000; i++) {
 512                 if (token_profit[i] > bestprofit) {
 513                         best = i;
 514                         bestprofit = token_profit[i];
 515                 }
 516         }
 517         return best;
 518 }
 519
 520 /* this is the core of the algorithm: calculate the "best" table */
 521 static void optimize_result(void)
 522 {
 523         int i, best;
 524
 525         /* using the '\0' symbol last allows compress_symbols to use standard
 526          * fast string functions */
 527         for (i = 255; i >= 0; i--) {
 528
 529                 /* if this table slot is empty (it is not used by an actual
 530                  * original char code */
 531                 if (!best_table_len[i]) {
 532
 533                         /* find the token with the breates profit value */
 534                         best = find_best_token();
 535                         if (token_profit[best] == 0)
 536                                 break;
 537
 538                         /* place it in the "best" table */
 539                         best_table_len[i] = 2;
 540                         best_table[i][0] = best & 0xFF;
 541                         best_table[i][1] = (best >> 8) & 0xFF;
 542
 543                         /* replace this token in all the valid symbols */
 544                         compress_symbols(best_table[i], i);
 545                 }
 546         }
 547 }
 548
 549 /* start by placing the symbols that are actually used on the table */
 550 static void insert_real_symbols_in_table(void)
 551 {
 552         unsigned int i, j, c;
 553
 554         memset(best_table, 0, sizeof(best_table));
 555         memset(best_table_len, 0, sizeof(best_table_len));
 556
 557         for (i = 0; i < table_cnt; i++) {
 558                 for (j = 0; j < table[i].len; j++) {
 559                         c = table[i].sym[j];
 560                         best_table[c][0]=c;
 561                         best_table_len[c]=1;
 562                 }
 563         }
 564 }
 565
 566 static void optimize_token_table(void)
 567 {
 568         build_initial_tok_table();
 569
 570         insert_real_symbols_in_table();
 571
 572         /* When valid symbol is not registered, exit to error */
 573         if (!table_cnt) {
 574                 fprintf(stderr, "No valid symbol.\n");
 575                 exit(1);
 576         }
 577
 578         optimize_result();
 579 }
 580
 581 /* guess for "linker script provide" symbol */
 582 static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
 583 {
 584         const char *symbol = (char *)se->sym + 1;
 585         int len = se->len - 1;
 586
 587         if (len < 8)
 588                 return 0;
 589
 590         if (symbol[0] != '_' || symbol[1] != '_')
 591                 return 0;
 592
 593         /* __start_XXXXX */
 594         if (!memcmp(symbol + 2, "start_", 6))
 595                 return 1;
 596
 597         /* __stop_XXXXX */
 598         if (!memcmp(symbol + 2, "stop_", 5))
 599                 return 1;
 600
 601         /* __end_XXXXX */
 602         if (!memcmp(symbol + 2, "end_", 4))
 603                 return 1;
 604
 605         /* __XXXXX_start */
 606         if (!memcmp(symbol + len - 6, "_start", 6))
 607                 return 1;
 608
 609         /* __XXXXX_end */
 610         if (!memcmp(symbol + len - 4, "_end", 4))
 611                 return 1;
 612
 613         return 0;
 614 }
 615
 616 static int prefix_underscores_count(const char *str)
 617 {
 618         const char *tail = str;
 619
 620         while (*tail == '_')
 621                 tail++;
 622
 623         return tail - str;
 624 }
 625
 626 static int compare_symbols(const void *a, const void *b)
 627 {
 628         const struct sym_entry *sa;
 629         const struct sym_entry *sb;
 630         int wa, wb;
 631
 632         sa = a;
 633         sb = b;
 634
 635         /* sort by address first */
 636         if (sa->addr > sb->addr)
 637                 return 1;
 638         if (sa->addr < sb->addr)
 639                 return -1;
 640
 641         /* sort by "weakness" type */
 642         wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
 643         wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
 644         if (wa != wb)
 645                 return wa - wb;
 646
 647         /* sort by "linker script provide" type */
 648         wa = may_be_linker_script_provide_symbol(sa);
 649         wb = may_be_linker_script_provide_symbol(sb);
 650         if (wa != wb)
 651                 return wa - wb;
 652
 653         /* sort by the number of prefix underscores */
 654         wa = prefix_underscores_count((const char *)sa->sym + 1);
 655         wb = prefix_underscores_count((const char *)sb->sym + 1);
 656         if (wa != wb)
 657                 return wa - wb;
 658
 659         /* sort by initial order, so that other symbols are left undisturbed */
 660         return sa->start_pos - sb->start_pos;
 661 }
 662
 663 static void sort_symbols(void)
 664 {
 665         qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
 666 }
 667
 668 static void make_percpus_absolute(void)
 669 {
 670         unsigned int i;
 671
 672         for (i = 0; i < table_cnt; i++)
 673                 if (symbol_in_range(&table[i], &percpu_range, 1))
 674                         table[i].sym[0] = 'A';
 675 }
 676
 677 int main(int argc, char **argv)
 678 {
 679         if (argc >= 2) {
 680                 int i;
 681                 for (i = 1; i < argc; i++) {
 682                         if(strcmp(argv[i], "--all-symbols") == 0)
 683                                 all_symbols = 1;
 684                         else if (strcmp(argv[i], "--absolute-percpu") == 0)
 685                                 absolute_percpu = 1;
 686                         else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
 687                                 char *p = &argv[i][16];
 688                                 /* skip quote */
 689                                 if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
 690                                         p++;
 691                                 symbol_prefix_char = *p;
 692                         } else if (strncmp(argv[i], "--page-offset=", 14) == 0) {
 693                                 const char *p = &argv[i][14];
 694                                 kernel_start_addr = strtoull(p, NULL, 16);
 695                         } else
 696                                 usage();
 697                 }
 698         } else if (argc != 1)
 699                 usage();
 700
 701         read_map(stdin);
 702         if (absolute_percpu)
 703                 make_percpus_absolute();
 704         sort_symbols();
 705         optimize_token_table();
 706         write_src();
 707
 708         return 0;
 709 }