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