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 #if 0
 220 /*
 221  * "Bug 69021 - Nondeterministic output from kallsyms ... maybe?
 222  * Cause for inconsistent kallsyms data"
 223  *
 224  * https://bugzilla.kernel.org/show_bug.cgi?id=69021
 225  *
 226  * Dropping of a symbol " __stop___modver" in stage 2 causes
 227  * reduction of 'kallsyms_addresses' structure size by 8 bytes,
 228  * and the following comparison of stage 2 and stage 3 outputs
 229  * changes. Workaround by removing the check.
 230  *
 231  * MCST
 232  */
 233                 /* Corner case.  Discard any symbols with the same value as
 234                  * _etext _einittext; they can move between pass 1 and 2 when
 235                  * the kallsyms data are added.  If these symbols move then
 236                  * they may get dropped in pass 2, which breaks the kallsyms
 237                  * rules.
 238                  */
 239                 if ((s->addr == text_range_text->end &&
 240                                 strcmp((char *)s->sym + offset, text_range_text->etext)) ||
 241                     (s->addr == text_range_inittext->end &&
 242                                 strcmp((char *)s->sym + offset, text_range_inittext->etext)))
 243                         return 0;
 244 #endif
 245         }
 246
 247         /* Exclude symbols which vary between passes. */
 248         if (strstr((char *)s->sym + offset, "_compiled."))
 249                 return 0;
 250
 251         for (i = 0; special_symbols[i]; i++)
 252                 if( strcmp((char *)s->sym + offset, special_symbols[i]) == 0 )
 253                         return 0;
 254
 255         return 1;
 256 }
 257
 258 static void read_map(FILE *in)
 259 {
 260         while (!feof(in)) {
 261                 if (table_cnt >= table_size) {
 262                         table_size += 10000;
 263                         table = realloc(table, sizeof(*table) * table_size);
 264                         if (!table) {
 265                                 fprintf(stderr, "out of memory\n");
 266                                 exit (1);
 267                         }
 268                 }
 269                 if (read_symbol(in, &table[table_cnt]) == 0) {
 270                         table[table_cnt].start_pos = table_cnt;
 271                         table_cnt++;
 272                 }
 273         }
 274 }
 275
 276 static void output_label(char *label)
 277 {
 278         if (symbol_prefix_char)
 279                 printf(".globl %c%s\n", symbol_prefix_char, label);
 280         else
 281                 printf(".globl %s\n", label);
 282         printf("\tALGN\n");
 283         if (symbol_prefix_char)
 284                 printf("%c%s:\n", symbol_prefix_char, label);
 285         else
 286                 printf("%s:\n", label);
 287 }
 288
 289 /* uncompress a compressed symbol. When this function is called, the best table
 290  * might still be compressed itself, so the function needs to be recursive */
 291 static int expand_symbol(unsigned char *data, int len, char *result)
 292 {
 293         int c, rlen, total=0;
 294
 295         while (len) {
 296                 c = *data;
 297                 /* if the table holds a single char that is the same as the one
 298                  * we are looking for, then end the search */
 299                 if (best_table[c][0]==c && best_table_len[c]==1) {
 300                         *result++ = c;
 301                         total++;
 302                 } else {
 303                         /* if not, recurse and expand */
 304                         rlen = expand_symbol(best_table[c], best_table_len[c], result);
 305                         total += rlen;
 306                         result += rlen;
 307                 }
 308                 data++;
 309                 len--;
 310         }
 311         *result=0;
 312
 313         return total;
 314 }
 315
 316 static void write_src(void)
 317 {
 318         unsigned int i, k, off;
 319         unsigned int best_idx[256];
 320         unsigned int *markers;
 321         char buf[KSYM_NAME_LEN];
 322
 323         printf("#include <asm/types.h>\n");
 324         printf("#if BITS_PER_LONG == 64\n");
 325         printf("#define PTR .quad\n");
 326         printf("#define ALGN .align 8\n");
 327         printf("#else\n");
 328         printf("#define PTR .long\n");
 329         printf("#define ALGN .align 4\n");
 330         printf("#endif\n");
 331
 332         printf("\t.section .rodata, \"a\"\n");
 333
 334         /* Provide proper symbols relocatability by their '_text'
 335          * relativeness.  The symbol names cannot be used to construct
 336          * normal symbol references as the list of symbols contains
 337          * symbols that are declared static and are private to their
 338          * .o files.  This prevents .tmp_kallsyms.o or any other
 339          * object from referencing them.
 340          */
 341         output_label("kallsyms_addresses");
 342         for (i = 0; i < table_cnt; i++) {
 343                 if (toupper(table[i].sym[0]) != 'A') {
 344                         if (_text <= table[i].addr)
 345                                 printf("\tPTR\t_text + %#llx\n",
 346                                         table[i].addr - _text);
 347                         else
 348                                 printf("\tPTR\t_text - %#llx\n",
 349                                         _text - table[i].addr);
 350                 } else {
 351                         printf("\tPTR\t%#llx\n", table[i].addr);
 352                 }
 353         }
 354         printf("\n");
 355
 356         output_label("kallsyms_num_syms");
 357         printf("\tPTR\t%d\n", table_cnt);
 358         printf("\n");
 359
 360         /* table of offset markers, that give the offset in the compressed stream
 361          * every 256 symbols */
 362         markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
 363         if (!markers) {
 364                 fprintf(stderr, "kallsyms failure: "
 365                         "unable to allocate required memory\n");
 366                 exit(EXIT_FAILURE);
 367         }
 368
 369         output_label("kallsyms_names");
 370         off = 0;
 371         for (i = 0; i < table_cnt; i++) {
 372                 if ((i & 0xFF) == 0)
 373                         markers[i >> 8] = off;
 374
 375                 printf("\t.byte 0x%02x", table[i].len);
 376                 for (k = 0; k < table[i].len; k++)
 377                         printf(", 0x%02x", table[i].sym[k]);
 378                 printf("\n");
 379
 380                 off += table[i].len + 1;
 381         }
 382         printf("\n");
 383
 384         output_label("kallsyms_markers");
 385         for (i = 0; i < ((table_cnt + 255) >> 8); i++)
 386                 printf("\tPTR\t%d\n", markers[i]);
 387         printf("\n");
 388
 389         free(markers);
 390
 391         output_label("kallsyms_token_table");
 392         off = 0;
 393         for (i = 0; i < 256; i++) {
 394                 best_idx[i] = off;
 395                 expand_symbol(best_table[i], best_table_len[i], buf);
 396                 printf("\t.asciz\t\"%s\"\n", buf);
 397                 off += strlen(buf) + 1;
 398         }
 399         printf("\n");
 400
 401         output_label("kallsyms_token_index");
 402         for (i = 0; i < 256; i++)
 403                 printf("\t.short\t%d\n", best_idx[i]);
 404         printf("\n");
 405 }
 406
 407
 408 /* table lookup compression functions */
 409
 410 /* count all the possible tokens in a symbol */
 411 static void learn_symbol(unsigned char *symbol, int len)
 412 {
 413         int i;
 414
 415         for (i = 0; i < len - 1; i++)
 416                 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
 417 }
 418
 419 /* decrease the count for all the possible tokens in a symbol */
 420 static void forget_symbol(unsigned char *symbol, int len)
 421 {
 422         int i;
 423
 424         for (i = 0; i < len - 1; i++)
 425                 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
 426 }
 427
 428 /* remove all the invalid symbols from the table and do the initial token count */
 429 static void build_initial_tok_table(void)
 430 {
 431         unsigned int i, pos;
 432
 433         pos = 0;
 434         for (i = 0; i < table_cnt; i++) {
 435                 if ( symbol_valid(&table[i]) ) {
 436                         if (pos != i)
 437                                 table[pos] = table[i];
 438                         learn_symbol(table[pos].sym, table[pos].len);
 439                         pos++;
 440                 }
 441         }
 442         table_cnt = pos;
 443 }
 444
 445 static void *find_token(unsigned char *str, int len, unsigned char *token)
 446 {
 447         int i;
 448
 449         for (i = 0; i < len - 1; i++) {
 450                 if (str[i] == token[0] && str[i+1] == token[1])
 451                         return &str[i];
 452         }
 453         return NULL;
 454 }
 455
 456 /* replace a given token in all the valid symbols. Use the sampled symbols
 457  * to update the counts */
 458 static void compress_symbols(unsigned char *str, int idx)
 459 {
 460         unsigned int i, len, size;
 461         unsigned char *p1, *p2;
 462
 463         for (i = 0; i < table_cnt; i++) {
 464
 465                 len = table[i].len;
 466                 p1 = table[i].sym;
 467
 468                 /* find the token on the symbol */
 469                 p2 = find_token(p1, len, str);
 470                 if (!p2) continue;
 471
 472                 /* decrease the counts for this symbol's tokens */
 473                 forget_symbol(table[i].sym, len);
 474
 475                 size = len;
 476
 477                 do {
 478                         *p2 = idx;
 479                         p2++;
 480                         size -= (p2 - p1);
 481                         memmove(p2, p2 + 1, size);
 482                         p1 = p2;
 483                         len--;
 484
 485                         if (size < 2) break;
 486
 487                         /* find the token on the symbol */
 488                         p2 = find_token(p1, size, str);
 489
 490                 } while (p2);
 491
 492                 table[i].len = len;
 493
 494                 /* increase the counts for this symbol's new tokens */
 495                 learn_symbol(table[i].sym, len);
 496         }
 497 }
 498
 499 /* search the token with the maximum profit */
 500 static int find_best_token(void)
 501 {
 502         int i, best, bestprofit;
 503
 504         bestprofit=-10000;
 505         best = 0;
 506
 507         for (i = 0; i < 0x10000; i++) {
 508                 if (token_profit[i] > bestprofit) {
 509                         best = i;
 510                         bestprofit = token_profit[i];
 511                 }
 512         }
 513         return best;
 514 }
 515
 516 /* this is the core of the algorithm: calculate the "best" table */
 517 static void optimize_result(void)
 518 {
 519         int i, best;
 520
 521         /* using the '\0' symbol last allows compress_symbols to use standard
 522          * fast string functions */
 523         for (i = 255; i >= 0; i--) {
 524
 525                 /* if this table slot is empty (it is not used by an actual
 526                  * original char code */
 527                 if (!best_table_len[i]) {
 528
 529                         /* find the token with the breates profit value */
 530                         best = find_best_token();
 531                         if (token_profit[best] == 0)
 532                                 break;
 533
 534                         /* place it in the "best" table */
 535                         best_table_len[i] = 2;
 536                         best_table[i][0] = best & 0xFF;
 537                         best_table[i][1] = (best >> 8) & 0xFF;
 538
 539                         /* replace this token in all the valid symbols */
 540                         compress_symbols(best_table[i], i);
 541                 }
 542         }
 543 }
 544
 545 /* start by placing the symbols that are actually used on the table */
 546 static void insert_real_symbols_in_table(void)
 547 {
 548         unsigned int i, j, c;
 549
 550         memset(best_table, 0, sizeof(best_table));
 551         memset(best_table_len, 0, sizeof(best_table_len));
 552
 553         for (i = 0; i < table_cnt; i++) {
 554                 for (j = 0; j < table[i].len; j++) {
 555                         c = table[i].sym[j];
 556                         best_table[c][0]=c;
 557                         best_table_len[c]=1;
 558                 }
 559         }
 560 }
 561
 562 static void optimize_token_table(void)
 563 {
 564         build_initial_tok_table();
 565
 566         insert_real_symbols_in_table();
 567
 568         /* When valid symbol is not registered, exit to error */
 569         if (!table_cnt) {
 570                 fprintf(stderr, "No valid symbol.\n");
 571                 exit(1);
 572         }
 573
 574         optimize_result();
 575 }
 576
 577 /* guess for "linker script provide" symbol */
 578 static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
 579 {
 580         const char *symbol = (char *)se->sym + 1;
 581         int len = se->len - 1;
 582
 583         if (len < 8)
 584                 return 0;
 585
 586         if (symbol[0] != '_' || symbol[1] != '_')
 587                 return 0;
 588
 589         /* __start_XXXXX */
 590         if (!memcmp(symbol + 2, "start_", 6))
 591                 return 1;
 592
 593         /* __stop_XXXXX */
 594         if (!memcmp(symbol + 2, "stop_", 5))
 595                 return 1;
 596
 597         /* __end_XXXXX */
 598         if (!memcmp(symbol + 2, "end_", 4))
 599                 return 1;
 600
 601         /* __XXXXX_start */
 602         if (!memcmp(symbol + len - 6, "_start", 6))
 603                 return 1;
 604
 605         /* __XXXXX_end */
 606         if (!memcmp(symbol + len - 4, "_end", 4))
 607                 return 1;
 608
 609         return 0;
 610 }
 611
 612 static int prefix_underscores_count(const char *str)
 613 {
 614         const char *tail = str;
 615
 616         while (*tail == '_')
 617                 tail++;
 618
 619         return tail - str;
 620 }
 621
 622 static int compare_symbols(const void *a, const void *b)
 623 {
 624         const struct sym_entry *sa;
 625         const struct sym_entry *sb;
 626         int wa, wb;
 627
 628         sa = a;
 629         sb = b;
 630
 631         /* sort by address first */
 632         if (sa->addr > sb->addr)
 633                 return 1;
 634         if (sa->addr < sb->addr)
 635                 return -1;
 636
 637         /* sort by "weakness" type */
 638         wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
 639         wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
 640         if (wa != wb)
 641                 return wa - wb;
 642
 643         /* sort by "linker script provide" type */
 644         wa = may_be_linker_script_provide_symbol(sa);
 645         wb = may_be_linker_script_provide_symbol(sb);
 646         if (wa != wb)
 647                 return wa - wb;
 648
 649         /* sort by the number of prefix underscores */
 650         wa = prefix_underscores_count((const char *)sa->sym + 1);
 651         wb = prefix_underscores_count((const char *)sb->sym + 1);
 652         if (wa != wb)
 653                 return wa - wb;
 654
 655         /* sort by initial order, so that other symbols are left undisturbed */
 656         return sa->start_pos - sb->start_pos;
 657 }
 658
 659 static void sort_symbols(void)
 660 {
 661         qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
 662 }
 663
 664 int main(int argc, char **argv)
 665 {
 666         if (argc >= 2) {
 667                 int i;
 668                 for (i = 1; i < argc; i++) {
 669                         if(strcmp(argv[i], "--all-symbols") == 0)
 670                                 all_symbols = 1;
 671                         else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
 672                                 char *p = &argv[i][16];
 673                                 /* skip quote */
 674                                 if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
 675                                         p++;
 676                                 symbol_prefix_char = *p;
 677                         } else if (strncmp(argv[i], "--page-offset=", 14) == 0) {
 678                                 const char *p = &argv[i][14];
 679                                 kernel_start_addr = strtoull(p, NULL, 16);
 680                         } else
 681                                 usage();
 682                 }
 683         } else if (argc != 1)
 684                 usage();
 685
 686         read_map(stdin);
 687         sort_symbols();
 688         optimize_token_table();
 689         write_src();
 690
 691         return 0;
 692 }