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