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
 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         if (s->addr < kernel_start_addr)
 202                 return 0;
 203
 204         /* skip prefix char */
 205         if (symbol_prefix_char && *(s->sym + 1) == symbol_prefix_char)
 206                 offset++;
 207
 208         /* if --all-symbols is not specified, then symbols outside the text
 209          * and inittext sections are discarded */
 210         if (!all_symbols) {
 211                 if (symbol_valid_tr(s) == 0)
 212                         return 0;
 213                 /* Corner case.  Discard any symbols with the same value as
 214                  * _etext _einittext; they can move between pass 1 and 2 when
 215                  * the kallsyms data are added.  If these symbols move then
 216                  * they may get dropped in pass 2, which breaks the kallsyms
 217                  * rules.
 218                  */
 219                 if ((s->addr == text_range_text->end &&
 220                                 strcmp((char *)s->sym + offset, text_range_text->etext)) ||
 221                     (s->addr == text_range_inittext->end &&
 222                                 strcmp((char *)s->sym + offset, text_range_inittext->etext)))
 223                         return 0;
 224         }
 225
 226         /* Exclude symbols which vary between passes. */
 227         if (strstr((char *)s->sym + offset, "_compiled."))
 228                 return 0;
 229
 230         for (i = 0; special_symbols[i]; i++)
 231                 if( strcmp((char *)s->sym + offset, special_symbols[i]) == 0 )
 232                         return 0;
 233
 234         return 1;
 235 }
 236
 237 static void read_map(FILE *in)
 238 {
 239         while (!feof(in)) {
 240                 if (table_cnt >= table_size) {
 241                         table_size += 10000;
 242                         table = realloc(table, sizeof(*table) * table_size);
 243                         if (!table) {
 244                                 fprintf(stderr, "out of memory\n");
 245                                 exit (1);
 246                         }
 247                 }
 248                 if (read_symbol(in, &table[table_cnt]) == 0) {
 249                         table[table_cnt].start_pos = table_cnt;
 250                         table_cnt++;
 251                 }
 252         }
 253 }
 254
 255 static void output_label(char *label)
 256 {
 257         if (symbol_prefix_char)
 258                 printf(".globl %c%s\n", symbol_prefix_char, label);
 259         else
 260                 printf(".globl %s\n", label);
 261         printf("\tALGN\n");
 262         if (symbol_prefix_char)
 263                 printf("%c%s:\n", symbol_prefix_char, label);
 264         else
 265                 printf("%s:\n", label);
 266 }
 267
 268 /* uncompress a compressed symbol. When this function is called, the best table
 269  * might still be compressed itself, so the function needs to be recursive */
 270 static int expand_symbol(unsigned char *data, int len, char *result)
 271 {
 272         int c, rlen, total=0;
 273
 274         while (len) {
 275                 c = *data;
 276                 /* if the table holds a single char that is the same as the one
 277                  * we are looking for, then end the search */
 278                 if (best_table[c][0]==c && best_table_len[c]==1) {
 279                         *result++ = c;
 280                         total++;
 281                 } else {
 282                         /* if not, recurse and expand */
 283                         rlen = expand_symbol(best_table[c], best_table_len[c], result);
 284                         total += rlen;
 285                         result += rlen;
 286                 }
 287                 data++;
 288                 len--;
 289         }
 290         *result=0;
 291
 292         return total;
 293 }
 294
 295 static void write_src(void)
 296 {
 297         unsigned int i, k, off;
 298         unsigned int best_idx[256];
 299         unsigned int *markers;
 300         char buf[KSYM_NAME_LEN];
 301
 302         printf("#include <asm/types.h>\n");
 303         printf("#if BITS_PER_LONG == 64\n");
 304         printf("#define PTR .quad\n");
 305         printf("#define ALGN .align 8\n");
 306         printf("#else\n");
 307         printf("#define PTR .long\n");
 308         printf("#define ALGN .align 4\n");
 309         printf("#endif\n");
 310
 311         printf("\t.section .rodata, \"a\"\n");
 312
 313         /* Provide proper symbols relocatability by their '_text'
 314          * relativeness.  The symbol names cannot be used to construct
 315          * normal symbol references as the list of symbols contains
 316          * symbols that are declared static and are private to their
 317          * .o files.  This prevents .tmp_kallsyms.o or any other
 318          * object from referencing them.
 319          */
 320         output_label("kallsyms_addresses");
 321         for (i = 0; i < table_cnt; i++) {
 322                 if (toupper(table[i].sym[0]) != 'A') {
 323                         if (_text <= table[i].addr)
 324                                 printf("\tPTR\t_text + %#llx\n",
 325                                         table[i].addr - _text);
 326                         else
 327                                 printf("\tPTR\t_text - %#llx\n",
 328                                         _text - table[i].addr);
 329                 } else {
 330                         printf("\tPTR\t%#llx\n", table[i].addr);
 331                 }
 332         }
 333         printf("\n");
 334
 335         output_label("kallsyms_num_syms");
 336         printf("\tPTR\t%d\n", table_cnt);
 337         printf("\n");
 338
 339         /* table of offset markers, that give the offset in the compressed stream
 340          * every 256 symbols */
 341         markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
 342         if (!markers) {
 343                 fprintf(stderr, "kallsyms failure: "
 344                         "unable to allocate required memory\n");
 345                 exit(EXIT_FAILURE);
 346         }
 347
 348         output_label("kallsyms_names");
 349         off = 0;
 350         for (i = 0; i < table_cnt; i++) {
 351                 if ((i & 0xFF) == 0)
 352                         markers[i >> 8] = off;
 353
 354                 printf("\t.byte 0x%02x", table[i].len);
 355                 for (k = 0; k < table[i].len; k++)
 356                         printf(", 0x%02x", table[i].sym[k]);
 357                 printf("\n");
 358
 359                 off += table[i].len + 1;
 360         }
 361         printf("\n");
 362
 363         output_label("kallsyms_markers");
 364         for (i = 0; i < ((table_cnt + 255) >> 8); i++)
 365                 printf("\tPTR\t%d\n", markers[i]);
 366         printf("\n");
 367
 368         free(markers);
 369
 370         output_label("kallsyms_token_table");
 371         off = 0;
 372         for (i = 0; i < 256; i++) {
 373                 best_idx[i] = off;
 374                 expand_symbol(best_table[i], best_table_len[i], buf);
 375                 printf("\t.asciz\t\"%s\"\n", buf);
 376                 off += strlen(buf) + 1;
 377         }
 378         printf("\n");
 379
 380         output_label("kallsyms_token_index");
 381         for (i = 0; i < 256; i++)
 382                 printf("\t.short\t%d\n", best_idx[i]);
 383         printf("\n");
 384 }
 385
 386
 387 /* table lookup compression functions */
 388
 389 /* count all the possible tokens in a symbol */
 390 static void learn_symbol(unsigned char *symbol, int len)
 391 {
 392         int i;
 393
 394         for (i = 0; i < len - 1; i++)
 395                 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
 396 }
 397
 398 /* decrease the count for all the possible tokens in a symbol */
 399 static void forget_symbol(unsigned char *symbol, int len)
 400 {
 401         int i;
 402
 403         for (i = 0; i < len - 1; i++)
 404                 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
 405 }
 406
 407 /* remove all the invalid symbols from the table and do the initial token count */
 408 static void build_initial_tok_table(void)
 409 {
 410         unsigned int i, pos;
 411
 412         pos = 0;
 413         for (i = 0; i < table_cnt; i++) {
 414                 if ( symbol_valid(&table[i]) ) {
 415                         if (pos != i)
 416                                 table[pos] = table[i];
 417                         learn_symbol(table[pos].sym, table[pos].len);
 418                         pos++;
 419                 }
 420         }
 421         table_cnt = pos;
 422 }
 423
 424 static void *find_token(unsigned char *str, int len, unsigned char *token)
 425 {
 426         int i;
 427
 428         for (i = 0; i < len - 1; i++) {
 429                 if (str[i] == token[0] && str[i+1] == token[1])
 430                         return &str[i];
 431         }
 432         return NULL;
 433 }
 434
 435 /* replace a given token in all the valid symbols. Use the sampled symbols
 436  * to update the counts */
 437 static void compress_symbols(unsigned char *str, int idx)
 438 {
 439         unsigned int i, len, size;
 440         unsigned char *p1, *p2;
 441
 442         for (i = 0; i < table_cnt; i++) {
 443
 444                 len = table[i].len;
 445                 p1 = table[i].sym;
 446
 447                 /* find the token on the symbol */
 448                 p2 = find_token(p1, len, str);
 449                 if (!p2) continue;
 450
 451                 /* decrease the counts for this symbol's tokens */
 452                 forget_symbol(table[i].sym, len);
 453
 454                 size = len;
 455
 456                 do {
 457                         *p2 = idx;
 458                         p2++;
 459                         size -= (p2 - p1);
 460                         memmove(p2, p2 + 1, size);
 461                         p1 = p2;
 462                         len--;
 463
 464                         if (size < 2) break;
 465
 466                         /* find the token on the symbol */
 467                         p2 = find_token(p1, size, str);
 468
 469                 } while (p2);
 470
 471                 table[i].len = len;
 472
 473                 /* increase the counts for this symbol's new tokens */
 474                 learn_symbol(table[i].sym, len);
 475         }
 476 }
 477
 478 /* search the token with the maximum profit */
 479 static int find_best_token(void)
 480 {
 481         int i, best, bestprofit;
 482
 483         bestprofit=-10000;
 484         best = 0;
 485
 486         for (i = 0; i < 0x10000; i++) {
 487                 if (token_profit[i] > bestprofit) {
 488                         best = i;
 489                         bestprofit = token_profit[i];
 490                 }
 491         }
 492         return best;
 493 }
 494
 495 /* this is the core of the algorithm: calculate the "best" table */
 496 static void optimize_result(void)
 497 {
 498         int i, best;
 499
 500         /* using the '\0' symbol last allows compress_symbols to use standard
 501          * fast string functions */
 502         for (i = 255; i >= 0; i--) {
 503
 504                 /* if this table slot is empty (it is not used by an actual
 505                  * original char code */
 506                 if (!best_table_len[i]) {
 507
 508                         /* find the token with the breates profit value */
 509                         best = find_best_token();
 510                         if (token_profit[best] == 0)
 511                                 break;
 512
 513                         /* place it in the "best" table */
 514                         best_table_len[i] = 2;
 515                         best_table[i][0] = best & 0xFF;
 516                         best_table[i][1] = (best >> 8) & 0xFF;
 517
 518                         /* replace this token in all the valid symbols */
 519                         compress_symbols(best_table[i], i);
 520                 }
 521         }
 522 }
 523
 524 /* start by placing the symbols that are actually used on the table */
 525 static void insert_real_symbols_in_table(void)
 526 {
 527         unsigned int i, j, c;
 528
 529         memset(best_table, 0, sizeof(best_table));
 530         memset(best_table_len, 0, sizeof(best_table_len));
 531
 532         for (i = 0; i < table_cnt; i++) {
 533                 for (j = 0; j < table[i].len; j++) {
 534                         c = table[i].sym[j];
 535                         best_table[c][0]=c;
 536                         best_table_len[c]=1;
 537                 }
 538         }
 539 }
 540
 541 static void optimize_token_table(void)
 542 {
 543         build_initial_tok_table();
 544
 545         insert_real_symbols_in_table();
 546
 547         /* When valid symbol is not registered, exit to error */
 548         if (!table_cnt) {
 549                 fprintf(stderr, "No valid symbol.\n");
 550                 exit(1);
 551         }
 552
 553         optimize_result();
 554 }
 555
 556 /* guess for "linker script provide" symbol */
 557 static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
 558 {
 559         const char *symbol = (char *)se->sym + 1;
 560         int len = se->len - 1;
 561
 562         if (len < 8)
 563                 return 0;
 564
 565         if (symbol[0] != '_' || symbol[1] != '_')
 566                 return 0;
 567
 568         /* __start_XXXXX */
 569         if (!memcmp(symbol + 2, "start_", 6))
 570                 return 1;
 571
 572         /* __stop_XXXXX */
 573         if (!memcmp(symbol + 2, "stop_", 5))
 574                 return 1;
 575
 576         /* __end_XXXXX */
 577         if (!memcmp(symbol + 2, "end_", 4))
 578                 return 1;
 579
 580         /* __XXXXX_start */
 581         if (!memcmp(symbol + len - 6, "_start", 6))
 582                 return 1;
 583
 584         /* __XXXXX_end */
 585         if (!memcmp(symbol + len - 4, "_end", 4))
 586                 return 1;
 587
 588         return 0;
 589 }
 590
 591 static int prefix_underscores_count(const char *str)
 592 {
 593         const char *tail = str;
 594
 595         while (*tail == '_')
 596                 tail++;
 597
 598         return tail - str;
 599 }
 600
 601 static int compare_symbols(const void *a, const void *b)
 602 {
 603         const struct sym_entry *sa;
 604         const struct sym_entry *sb;
 605         int wa, wb;
 606
 607         sa = a;
 608         sb = b;
 609
 610         /* sort by address first */
 611         if (sa->addr > sb->addr)
 612                 return 1;
 613         if (sa->addr < sb->addr)
 614                 return -1;
 615
 616         /* sort by "weakness" type */
 617         wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
 618         wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
 619         if (wa != wb)
 620                 return wa - wb;
 621
 622         /* sort by "linker script provide" type */
 623         wa = may_be_linker_script_provide_symbol(sa);
 624         wb = may_be_linker_script_provide_symbol(sb);
 625         if (wa != wb)
 626                 return wa - wb;
 627
 628         /* sort by the number of prefix underscores */
 629         wa = prefix_underscores_count((const char *)sa->sym + 1);
 630         wb = prefix_underscores_count((const char *)sb->sym + 1);
 631         if (wa != wb)
 632                 return wa - wb;
 633
 634         /* sort by initial order, so that other symbols are left undisturbed */
 635         return sa->start_pos - sb->start_pos;
 636 }
 637
 638 static void sort_symbols(void)
 639 {
 640         qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
 641 }
 642
 643 int main(int argc, char **argv)
 644 {
 645         if (argc >= 2) {
 646                 int i;
 647                 for (i = 1; i < argc; i++) {
 648                         if(strcmp(argv[i], "--all-symbols") == 0)
 649                                 all_symbols = 1;
 650                         else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
 651                                 char *p = &argv[i][16];
 652                                 /* skip quote */
 653                                 if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
 654                                         p++;
 655                                 symbol_prefix_char = *p;
 656                         } else if (strncmp(argv[i], "--page-offset=", 14) == 0) {
 657                                 const char *p = &argv[i][14];
 658                                 kernel_start_addr = strtoull(p, NULL, 16);
 659                         } else
 660                                 usage();
 661                 }
 662         } else if (argc != 1)
 663                 usage();
 664
 665         read_map(stdin);
 666         sort_symbols();
 667         optimize_token_table();
 668         write_src();
 669
 670         return 0;
 671 }