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 text_range {
  40         const char *stext, *etext;
  41         unsigned long long start, end;
  42 };
  43
  44 static unsigned long long _text;
  45 static struct text_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 sym_entry *table;
  55 static unsigned int table_size, table_cnt;
  56 static int all_symbols = 0;
  57 static char symbol_prefix_char = '\0';
  58 static unsigned long long kernel_start_addr = 0;
  59
  60 int token_profit[0x10000];
  61
  62 /* the table that holds the result of the compression */
  63 unsigned char best_table[256][2];
  64 unsigned char best_table_len[256];
  65
  66
  67 static void usage(void)
  68 {
  69         fprintf(stderr, "Usage: kallsyms [--all-symbols] "
  70                         "[--symbol-prefix=<prefix char>] "
  71                         "[--page-offset=<CONFIG_PAGE_OFFSET>] "
  72                         "< in.map > out.S\n");
  73         exit(1);
  74 }
  75
  76 /*
  77  * This ignores the intensely annoying "mapping symbols" found
  78  * in ARM ELF files: $a, $t and $d.
  79  */
  80 static inline int is_arm_mapping_symbol(const char *str)
  81 {
  82         return str[0] == '$' && strchr("atd", str[1])
  83                && (str[2] == '\0' || str[2] == '.');
  84 }
  85
  86 static int read_symbol_tr(const char *sym, unsigned long long addr)
  87 {
  88         size_t i;
  89         struct text_range *tr;
  90
  91         for (i = 0; i < ARRAY_SIZE(text_ranges); ++i) {
  92                 tr = &text_ranges[i];
  93
  94                 if (strcmp(sym, tr->stext) == 0) {
  95                         tr->start = addr;
  96                         return 0;
  97                 } else if (strcmp(sym, tr->etext) == 0) {
  98                         tr->end = addr;
  99                         return 0;
 100                 }
 101         }
 102
 103         return 1;
 104 }
 105
 106 static int read_symbol(FILE *in, struct sym_entry *s)
 107 {
 108         char str[500];
 109         char *sym, stype;
 110         int rc;
 111
 112         rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, str);
 113         if (rc != 3) {
 114                 if (rc != EOF && fgets(str, 500, in) == NULL)
 115                         fprintf(stderr, "Read error or end of file.\n");
 116                 return -1;
 117         }
 118         if (strlen(str) > KSYM_NAME_LEN) {
 119                 fprintf(stderr, "Symbol %s too long for kallsyms (%zu vs %d).\n"
 120                                 "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n",
 121                         str, strlen(str), KSYM_NAME_LEN);
 122                 return -1;
 123         }
 124
 125         sym = str;
 126         /* skip prefix char */
 127         if (symbol_prefix_char && str[0] == symbol_prefix_char)
 128                 sym++;
 129
 130         /* Ignore most absolute/undefined (?) symbols. */
 131         if (strcmp(sym, "_text") == 0)
 132                 _text = s->addr;
 133         else if (read_symbol_tr(sym, s->addr) == 0)
 134                 /* nothing to do */;
 135         else if (toupper(stype) == 'A')
 136         {
 137                 /* Keep these useful absolute symbols */
 138                 if (strcmp(sym, "__kernel_syscall_via_break") &&
 139                     strcmp(sym, "__kernel_syscall_via_epc") &&
 140                     strcmp(sym, "__kernel_sigtramp") &&
 141                     strcmp(sym, "__gp"))
 142                         return -1;
 143
 144         }
 145         else if (toupper(stype) == 'U' ||
 146                  is_arm_mapping_symbol(sym))
 147                 return -1;
 148         /* exclude also MIPS ELF local symbols ($L123 instead of .L123) */
 149         else if (str[0] == '$')
 150                 return -1;
 151         /* exclude debugging symbols */
 152         else if (stype == 'N')
 153                 return -1;
 154
 155         /* include the type field in the symbol name, so that it gets
 156          * compressed together */
 157         s->len = strlen(str) + 1;
 158         s->sym = malloc(s->len + 1);
 159         if (!s->sym) {
 160                 fprintf(stderr, "kallsyms failure: "
 161                         "unable to allocate required amount of memory\n");
 162                 exit(EXIT_FAILURE);
 163         }
 164         strcpy((char *)s->sym + 1, str);
 165         s->sym[0] = stype;
 166
 167         return 0;
 168 }
 169
 170 static int symbol_valid_tr(struct sym_entry *s)
 171 {
 172         size_t i;
 173         struct text_range *tr;
 174
 175         for (i = 0; i < ARRAY_SIZE(text_ranges); ++i) {
 176                 tr = &text_ranges[i];
 177
 178                 if (s->addr >= tr->start && s->addr <= tr->end)
 179                         return 1;
 180         }
 181
 182         return 0;
 183 }
 184
 185 static int symbol_valid(struct sym_entry *s)
 186 {
 187         /* Symbols which vary between passes.  Passes 1 and 2 must have
 188          * identical symbol lists.  The kallsyms_* symbols below are only added
 189          * after pass 1, they would be included in pass 2 when --all-symbols is
 190          * specified so exclude them to get a stable symbol list.
 191          */
 192         static char *special_symbols[] = {
 193                 "kallsyms_addresses",
 194                 "kallsyms_num_syms",
 195                 "kallsyms_names",
 196                 "kallsyms_markers",
 197                 "kallsyms_token_table",
 198                 "kallsyms_token_index",
 199
 200         /* Exclude linker generated symbols which vary between passes */
 201                 "_SDA_BASE_",           /* ppc */
 202                 "_SDA2_BASE_",          /* ppc */
 203                 NULL };
 204         int i;
 205         int offset = 1;
 206
 207         if (s->addr < kernel_start_addr)
 208                 return 0;
 209
 210         /* skip prefix char */
 211         if (symbol_prefix_char && *(s->sym + 1) == symbol_prefix_char)
 212                 offset++;
 213
 214         /* if --all-symbols is not specified, then symbols outside the text
 215          * and inittext sections are discarded */
 216         if (!all_symbols) {
 217                 if (symbol_valid_tr(s) == 0)
 218                         return 0;
 219                 /* Corner case.  Discard any symbols with the same value as
 220                  * _etext _einittext; they can move between pass 1 and 2 when
 221                  * the kallsyms data are added.  If these symbols move then
 222                  * they may get dropped in pass 2, which breaks the kallsyms
 223                  * rules.
 224                  */
 225                 if ((s->addr == text_range_text->end &&
 226                                 strcmp((char *)s->sym + offset, text_range_text->etext)) ||
 227                     (s->addr == text_range_inittext->end &&
 228                                 strcmp((char *)s->sym + offset, text_range_inittext->etext)))
 229                         return 0;
 230         }
 231
 232         /* Exclude symbols which vary between passes. */
 233         if (strstr((char *)s->sym + offset, "_compiled."))
 234                 return 0;
 235
 236         for (i = 0; special_symbols[i]; i++)
 237                 if( strcmp((char *)s->sym + offset, special_symbols[i]) == 0 )
 238                         return 0;
 239
 240         return 1;
 241 }
 242
 243 static void read_map(FILE *in)
 244 {
 245         while (!feof(in)) {
 246                 if (table_cnt >= table_size) {
 247                         table_size += 10000;
 248                         table = realloc(table, sizeof(*table) * table_size);
 249                         if (!table) {
 250                                 fprintf(stderr, "out of memory\n");
 251                                 exit (1);
 252                         }
 253                 }
 254                 if (read_symbol(in, &table[table_cnt]) == 0) {
 255                         table[table_cnt].start_pos = table_cnt;
 256                         table_cnt++;
 257                 }
 258         }
 259 }
 260
 261 static void output_label(char *label)
 262 {
 263         if (symbol_prefix_char)
 264                 printf(".globl %c%s\n", symbol_prefix_char, label);
 265         else
 266                 printf(".globl %s\n", label);
 267         printf("\tALGN\n");
 268         if (symbol_prefix_char)
 269                 printf("%c%s:\n", symbol_prefix_char, label);
 270         else
 271                 printf("%s:\n", label);
 272 }
 273
 274 /* uncompress a compressed symbol. When this function is called, the best table
 275  * might still be compressed itself, so the function needs to be recursive */
 276 static int expand_symbol(unsigned char *data, int len, char *result)
 277 {
 278         int c, rlen, total=0;
 279
 280         while (len) {
 281                 c = *data;
 282                 /* if the table holds a single char that is the same as the one
 283                  * we are looking for, then end the search */
 284                 if (best_table[c][0]==c && best_table_len[c]==1) {
 285                         *result++ = c;
 286                         total++;
 287                 } else {
 288                         /* if not, recurse and expand */
 289                         rlen = expand_symbol(best_table[c], best_table_len[c], result);
 290                         total += rlen;
 291                         result += rlen;
 292                 }
 293                 data++;
 294                 len--;
 295         }
 296         *result=0;
 297
 298         return total;
 299 }
 300
 301 static void write_src(void)
 302 {
 303         unsigned int i, k, off;
 304         unsigned int best_idx[256];
 305         unsigned int *markers;
 306         char buf[KSYM_NAME_LEN];
 307
 308         printf("#include <asm/types.h>\n");
 309         printf("#if BITS_PER_LONG == 64\n");
 310         printf("#define PTR .quad\n");
 311         printf("#define ALGN .align 8\n");
 312         printf("#else\n");
 313         printf("#define PTR .long\n");
 314         printf("#define ALGN .align 4\n");
 315         printf("#endif\n");
 316
 317         printf("\t.section .rodata, \"a\"\n");
 318
 319         /* Provide proper symbols relocatability by their '_text'
 320          * relativeness.  The symbol names cannot be used to construct
 321          * normal symbol references as the list of symbols contains
 322          * symbols that are declared static and are private to their
 323          * .o files.  This prevents .tmp_kallsyms.o or any other
 324          * object from referencing them.
 325          */
 326         output_label("kallsyms_addresses");
 327         for (i = 0; i < table_cnt; i++) {
 328                 if (toupper(table[i].sym[0]) != 'A') {
 329                         if (_text <= table[i].addr)
 330                                 printf("\tPTR\t_text + %#llx\n",
 331                                         table[i].addr - _text);
 332                         else
 333                                 printf("\tPTR\t%#llx\n", table[i].addr);
 334                 } else {
 335                         printf("\tPTR\t%#llx\n", table[i].addr);
 336                 }
 337         }
 338         printf("\n");
 339
 340         output_label("kallsyms_num_syms");
 341         printf("\tPTR\t%d\n", table_cnt);
 342         printf("\n");
 343
 344         /* table of offset markers, that give the offset in the compressed stream
 345          * every 256 symbols */
 346         markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
 347         if (!markers) {
 348                 fprintf(stderr, "kallsyms failure: "
 349                         "unable to allocate required memory\n");
 350                 exit(EXIT_FAILURE);
 351         }
 352
 353         output_label("kallsyms_names");
 354         off = 0;
 355         for (i = 0; i < table_cnt; i++) {
 356                 if ((i & 0xFF) == 0)
 357                         markers[i >> 8] = off;
 358
 359                 printf("\t.byte 0x%02x", table[i].len);
 360                 for (k = 0; k < table[i].len; k++)
 361                         printf(", 0x%02x", table[i].sym[k]);
 362                 printf("\n");
 363
 364                 off += table[i].len + 1;
 365         }
 366         printf("\n");
 367
 368         output_label("kallsyms_markers");
 369         for (i = 0; i < ((table_cnt + 255) >> 8); i++)
 370                 printf("\tPTR\t%d\n", markers[i]);
 371         printf("\n");
 372
 373         free(markers);
 374
 375         output_label("kallsyms_token_table");
 376         off = 0;
 377         for (i = 0; i < 256; i++) {
 378                 best_idx[i] = off;
 379                 expand_symbol(best_table[i], best_table_len[i], buf);
 380                 printf("\t.asciz\t\"%s\"\n", buf);
 381                 off += strlen(buf) + 1;
 382         }
 383         printf("\n");
 384
 385         output_label("kallsyms_token_index");
 386         for (i = 0; i < 256; i++)
 387                 printf("\t.short\t%d\n", best_idx[i]);
 388         printf("\n");
 389 }
 390
 391
 392 /* table lookup compression functions */
 393
 394 /* count all the possible tokens in a symbol */
 395 static void learn_symbol(unsigned char *symbol, int len)
 396 {
 397         int i;
 398
 399         for (i = 0; i < len - 1; i++)
 400                 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
 401 }
 402
 403 /* decrease the count for all the possible tokens in a symbol */
 404 static void forget_symbol(unsigned char *symbol, int len)
 405 {
 406         int i;
 407
 408         for (i = 0; i < len - 1; i++)
 409                 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
 410 }
 411
 412 /* remove all the invalid symbols from the table and do the initial token count */
 413 static void build_initial_tok_table(void)
 414 {
 415         unsigned int i, pos;
 416
 417         pos = 0;
 418         for (i = 0; i < table_cnt; i++) {
 419                 if ( symbol_valid(&table[i]) ) {
 420                         if (pos != i)
 421                                 table[pos] = table[i];
 422                         learn_symbol(table[pos].sym, table[pos].len);
 423                         pos++;
 424                 }
 425         }
 426         table_cnt = pos;
 427 }
 428
 429 static void *find_token(unsigned char *str, int len, unsigned char *token)
 430 {
 431         int i;
 432
 433         for (i = 0; i < len - 1; i++) {
 434                 if (str[i] == token[0] && str[i+1] == token[1])
 435                         return &str[i];
 436         }
 437         return NULL;
 438 }
 439
 440 /* replace a given token in all the valid symbols. Use the sampled symbols
 441  * to update the counts */
 442 static void compress_symbols(unsigned char *str, int idx)
 443 {
 444         unsigned int i, len, size;
 445         unsigned char *p1, *p2;
 446
 447         for (i = 0; i < table_cnt; i++) {
 448
 449                 len = table[i].len;
 450                 p1 = table[i].sym;
 451
 452                 /* find the token on the symbol */
 453                 p2 = find_token(p1, len, str);
 454                 if (!p2) continue;
 455
 456                 /* decrease the counts for this symbol's tokens */
 457                 forget_symbol(table[i].sym, len);
 458
 459                 size = len;
 460
 461                 do {
 462                         *p2 = idx;
 463                         p2++;
 464                         size -= (p2 - p1);
 465                         memmove(p2, p2 + 1, size);
 466                         p1 = p2;
 467                         len--;
 468
 469                         if (size < 2) break;
 470
 471                         /* find the token on the symbol */
 472                         p2 = find_token(p1, size, str);
 473
 474                 } while (p2);
 475
 476                 table[i].len = len;
 477
 478                 /* increase the counts for this symbol's new tokens */
 479                 learn_symbol(table[i].sym, len);
 480         }
 481 }
 482
 483 /* search the token with the maximum profit */
 484 static int find_best_token(void)
 485 {
 486         int i, best, bestprofit;
 487
 488         bestprofit=-10000;
 489         best = 0;
 490
 491         for (i = 0; i < 0x10000; i++) {
 492                 if (token_profit[i] > bestprofit) {
 493                         best = i;
 494                         bestprofit = token_profit[i];
 495                 }
 496         }
 497         return best;
 498 }
 499
 500 /* this is the core of the algorithm: calculate the "best" table */
 501 static void optimize_result(void)
 502 {
 503         int i, best;
 504
 505         /* using the '\0' symbol last allows compress_symbols to use standard
 506          * fast string functions */
 507         for (i = 255; i >= 0; i--) {
 508
 509                 /* if this table slot is empty (it is not used by an actual
 510                  * original char code */
 511                 if (!best_table_len[i]) {
 512
 513                         /* find the token with the breates profit value */
 514                         best = find_best_token();
 515                         if (token_profit[best] == 0)
 516                                 break;
 517
 518                         /* place it in the "best" table */
 519                         best_table_len[i] = 2;
 520                         best_table[i][0] = best & 0xFF;
 521                         best_table[i][1] = (best >> 8) & 0xFF;
 522
 523                         /* replace this token in all the valid symbols */
 524                         compress_symbols(best_table[i], i);
 525                 }
 526         }
 527 }
 528
 529 /* start by placing the symbols that are actually used on the table */
 530 static void insert_real_symbols_in_table(void)
 531 {
 532         unsigned int i, j, c;
 533
 534         memset(best_table, 0, sizeof(best_table));
 535         memset(best_table_len, 0, sizeof(best_table_len));
 536
 537         for (i = 0; i < table_cnt; i++) {
 538                 for (j = 0; j < table[i].len; j++) {
 539                         c = table[i].sym[j];
 540                         best_table[c][0]=c;
 541                         best_table_len[c]=1;
 542                 }
 543         }
 544 }
 545
 546 static void optimize_token_table(void)
 547 {
 548         build_initial_tok_table();
 549
 550         insert_real_symbols_in_table();
 551
 552         /* When valid symbol is not registered, exit to error */
 553         if (!table_cnt) {
 554                 fprintf(stderr, "No valid symbol.\n");
 555                 exit(1);
 556         }
 557
 558         optimize_result();
 559 }
 560
 561 /* guess for "linker script provide" symbol */
 562 static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
 563 {
 564         const char *symbol = (char *)se->sym + 1;
 565         int len = se->len - 1;
 566
 567         if (len < 8)
 568                 return 0;
 569
 570         if (symbol[0] != '_' || symbol[1] != '_')
 571                 return 0;
 572
 573         /* __start_XXXXX */
 574         if (!memcmp(symbol + 2, "start_", 6))
 575                 return 1;
 576
 577         /* __stop_XXXXX */
 578         if (!memcmp(symbol + 2, "stop_", 5))
 579                 return 1;
 580
 581         /* __end_XXXXX */
 582         if (!memcmp(symbol + 2, "end_", 4))
 583                 return 1;
 584
 585         /* __XXXXX_start */
 586         if (!memcmp(symbol + len - 6, "_start", 6))
 587                 return 1;
 588
 589         /* __XXXXX_end */
 590         if (!memcmp(symbol + len - 4, "_end", 4))
 591                 return 1;
 592
 593         return 0;
 594 }
 595
 596 static int prefix_underscores_count(const char *str)
 597 {
 598         const char *tail = str;
 599
 600         while (*tail == '_')
 601                 tail++;
 602
 603         return tail - str;
 604 }
 605
 606 static int compare_symbols(const void *a, const void *b)
 607 {
 608         const struct sym_entry *sa;
 609         const struct sym_entry *sb;
 610         int wa, wb;
 611
 612         sa = a;
 613         sb = b;
 614
 615         /* sort by address first */
 616         if (sa->addr > sb->addr)
 617                 return 1;
 618         if (sa->addr < sb->addr)
 619                 return -1;
 620
 621         /* sort by "weakness" type */
 622         wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
 623         wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
 624         if (wa != wb)
 625                 return wa - wb;
 626
 627         /* sort by "linker script provide" type */
 628         wa = may_be_linker_script_provide_symbol(sa);
 629         wb = may_be_linker_script_provide_symbol(sb);
 630         if (wa != wb)
 631                 return wa - wb;
 632
 633         /* sort by the number of prefix underscores */
 634         wa = prefix_underscores_count((const char *)sa->sym + 1);
 635         wb = prefix_underscores_count((const char *)sb->sym + 1);
 636         if (wa != wb)
 637                 return wa - wb;
 638
 639         /* sort by initial order, so that other symbols are left undisturbed */
 640         return sa->start_pos - sb->start_pos;
 641 }
 642
 643 static void sort_symbols(void)
 644 {
 645         qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
 646 }
 647
 648 int main(int argc, char **argv)
 649 {
 650         if (argc >= 2) {
 651                 int i;
 652                 for (i = 1; i < argc; i++) {
 653                         if(strcmp(argv[i], "--all-symbols") == 0)
 654                                 all_symbols = 1;
 655                         else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
 656                                 char *p = &argv[i][16];
 657                                 /* skip quote */
 658                                 if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
 659                                         p++;
 660                                 symbol_prefix_char = *p;
 661                         } else if (strncmp(argv[i], "--page-offset=", 14) == 0) {
 662                                 const char *p = &argv[i][14];
 663                                 kernel_start_addr = strtoull(p, NULL, 16);
 664                         } else
 665                                 usage();
 666                 }
 667         } else if (argc != 1)
 668                 usage();
 669
 670         read_map(stdin);
 671         sort_symbols();
 672         optimize_token_table();
 673         write_src();
 674
 675         return 0;
 676 }