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 #define KSYM_NAME_LEN           128
  27
  28 struct sym_entry {
  29         unsigned long long addr;
  30         unsigned int len;
  31         unsigned int start_pos;
  32         unsigned char *sym;
  33 };
  34
  35 static struct sym_entry *table;
  36 static unsigned int table_size, table_cnt;
  37 static unsigned long long _text, _stext, _etext, _sinittext, _einittext;
  38 static int all_symbols = 0;
  39 static char symbol_prefix_char = '\0';
  40
  41 int token_profit[0x10000];
  42
  43 /* the table that holds the result of the compression */
  44 unsigned char best_table[256][2];
  45 unsigned char best_table_len[256];
  46
  47
  48 static void usage(void)
  49 {
  50         fprintf(stderr, "Usage: kallsyms [--all-symbols] [--symbol-prefix=<prefix char>] < in.map > out.S\n");
  51         exit(1);
  52 }
  53
  54 /*
  55  * This ignores the intensely annoying "mapping symbols" found
  56  * in ARM ELF files: $a, $t and $d.
  57  */
  58 static inline int is_arm_mapping_symbol(const char *str)
  59 {
  60         return str[0] == '$' && strchr("atd", str[1])
  61                && (str[2] == '\0' || str[2] == '.');
  62 }
  63
  64 static int read_symbol(FILE *in, struct sym_entry *s)
  65 {
  66         char str[500];
  67         char *sym, stype;
  68         int rc;
  69
  70         rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, str);
  71         if (rc != 3) {
  72                 if (rc != EOF) {
  73                         /* skip line */
  74                         fgets(str, 500, in);
  75                 }
  76                 return -1;
  77         }
  78
  79         sym = str;
  80         /* skip prefix char */
  81         if (symbol_prefix_char && str[0] == symbol_prefix_char)
  82                 sym++;
  83
  84         /* Ignore most absolute/undefined (?) symbols. */
  85         if (strcmp(sym, "_text") == 0)
  86                 _text = s->addr;
  87         else if (strcmp(sym, "_stext") == 0)
  88                 _stext = s->addr;
  89         else if (strcmp(sym, "_etext") == 0)
  90                 _etext = s->addr;
  91         else if (strcmp(sym, "_sinittext") == 0)
  92                 _sinittext = s->addr;
  93         else if (strcmp(sym, "_einittext") == 0)
  94                 _einittext = s->addr;
  95         else if (toupper(stype) == 'A')
  96         {
  97                 /* Keep these useful absolute symbols */
  98                 if (strcmp(sym, "__kernel_syscall_via_break") &&
  99                     strcmp(sym, "__kernel_syscall_via_epc") &&
 100                     strcmp(sym, "__kernel_sigtramp") &&
 101                     strcmp(sym, "__gp"))
 102                         return -1;
 103
 104         }
 105         else if (toupper(stype) == 'U' ||
 106                  is_arm_mapping_symbol(sym))
 107                 return -1;
 108         /* exclude also MIPS ELF local symbols ($L123 instead of .L123) */
 109         else if (str[0] == '$')
 110                 return -1;
 111
 112         /* include the type field in the symbol name, so that it gets
 113          * compressed together */
 114         s->len = strlen(str) + 1;
 115         s->sym = malloc(s->len + 1);
 116         if (!s->sym) {
 117                 fprintf(stderr, "kallsyms failure: "
 118                         "unable to allocate required amount of memory\n");
 119                 exit(EXIT_FAILURE);
 120         }
 121         strcpy((char *)s->sym + 1, str);
 122         s->sym[0] = stype;
 123
 124         return 0;
 125 }
 126
 127 static int symbol_valid(struct sym_entry *s)
 128 {
 129         /* Symbols which vary between passes.  Passes 1 and 2 must have
 130          * identical symbol lists.  The kallsyms_* symbols below are only added
 131          * after pass 1, they would be included in pass 2 when --all-symbols is
 132          * specified so exclude them to get a stable symbol list.
 133          */
 134         static char *special_symbols[] = {
 135                 "kallsyms_addresses",
 136                 "kallsyms_num_syms",
 137                 "kallsyms_names",
 138                 "kallsyms_markers",
 139                 "kallsyms_token_table",
 140                 "kallsyms_token_index",
 141
 142         /* Exclude linker generated symbols which vary between passes */
 143                 "_SDA_BASE_",           /* ppc */
 144                 "_SDA2_BASE_",          /* ppc */
 145                 NULL };
 146         int i;
 147         int offset = 1;
 148
 149         /* skip prefix char */
 150         if (symbol_prefix_char && *(s->sym + 1) == symbol_prefix_char)
 151                 offset++;
 152
 153         /* if --all-symbols is not specified, then symbols outside the text
 154          * and inittext sections are discarded */
 155         if (!all_symbols) {
 156                 if ((s->addr < _stext || s->addr > _etext)
 157                     && (s->addr < _sinittext || s->addr > _einittext))
 158                         return 0;
 159                 /* Corner case.  Discard any symbols with the same value as
 160                  * _etext _einittext; they can move between pass 1 and 2 when
 161                  * the kallsyms data are added.  If these symbols move then
 162                  * they may get dropped in pass 2, which breaks the kallsyms
 163                  * rules.
 164                  */
 165                 if ((s->addr == _etext &&
 166                                 strcmp((char *)s->sym + offset, "_etext")) ||
 167                     (s->addr == _einittext &&
 168                                 strcmp((char *)s->sym + offset, "_einittext")))
 169                         return 0;
 170         }
 171
 172         /* Exclude symbols which vary between passes. */
 173         if (strstr((char *)s->sym + offset, "_compiled."))
 174                 return 0;
 175
 176         for (i = 0; special_symbols[i]; i++)
 177                 if( strcmp((char *)s->sym + offset, special_symbols[i]) == 0 )
 178                         return 0;
 179
 180         return 1;
 181 }
 182
 183 static void read_map(FILE *in)
 184 {
 185         while (!feof(in)) {
 186                 if (table_cnt >= table_size) {
 187                         table_size += 10000;
 188                         table = realloc(table, sizeof(*table) * table_size);
 189                         if (!table) {
 190                                 fprintf(stderr, "out of memory\n");
 191                                 exit (1);
 192                         }
 193                 }
 194                 if (read_symbol(in, &table[table_cnt]) == 0) {
 195                         table[table_cnt].start_pos = table_cnt;
 196                         table_cnt++;
 197                 }
 198         }
 199 }
 200
 201 static void output_label(char *label)
 202 {
 203         if (symbol_prefix_char)
 204                 printf(".globl %c%s\n", symbol_prefix_char, label);
 205         else
 206                 printf(".globl %s\n", label);
 207         printf("\tALGN\n");
 208         if (symbol_prefix_char)
 209                 printf("%c%s:\n", symbol_prefix_char, label);
 210         else
 211                 printf("%s:\n", label);
 212 }
 213
 214 /* uncompress a compressed symbol. When this function is called, the best table
 215  * might still be compressed itself, so the function needs to be recursive */
 216 static int expand_symbol(unsigned char *data, int len, char *result)
 217 {
 218         int c, rlen, total=0;
 219
 220         while (len) {
 221                 c = *data;
 222                 /* if the table holds a single char that is the same as the one
 223                  * we are looking for, then end the search */
 224                 if (best_table[c][0]==c && best_table_len[c]==1) {
 225                         *result++ = c;
 226                         total++;
 227                 } else {
 228                         /* if not, recurse and expand */
 229                         rlen = expand_symbol(best_table[c], best_table_len[c], result);
 230                         total += rlen;
 231                         result += rlen;
 232                 }
 233                 data++;
 234                 len--;
 235         }
 236         *result=0;
 237
 238         return total;
 239 }
 240
 241 static void write_src(void)
 242 {
 243         unsigned int i, k, off;
 244         unsigned int best_idx[256];
 245         unsigned int *markers;
 246         char buf[KSYM_NAME_LEN];
 247
 248         printf("#include <asm/types.h>\n");
 249         printf("#if BITS_PER_LONG == 64\n");
 250         printf("#define PTR .quad\n");
 251         printf("#define ALGN .align 8\n");
 252         printf("#else\n");
 253         printf("#define PTR .long\n");
 254         printf("#define ALGN .align 4\n");
 255         printf("#endif\n");
 256
 257         printf("\t.section .rodata, \"a\"\n");
 258
 259         /* Provide proper symbols relocatability by their '_text'
 260          * relativeness.  The symbol names cannot be used to construct
 261          * normal symbol references as the list of symbols contains
 262          * symbols that are declared static and are private to their
 263          * .o files.  This prevents .tmp_kallsyms.o or any other
 264          * object from referencing them.
 265          */
 266         output_label("kallsyms_addresses");
 267         for (i = 0; i < table_cnt; i++) {
 268                 if (toupper(table[i].sym[0]) != 'A') {
 269                         if (_text <= table[i].addr)
 270                                 printf("\tPTR\t_text + %#llx\n",
 271                                         table[i].addr - _text);
 272                         else
 273                                 printf("\tPTR\t_text - %#llx\n",
 274                                         _text - table[i].addr);
 275                 } else {
 276                         printf("\tPTR\t%#llx\n", table[i].addr);
 277                 }
 278         }
 279         printf("\n");
 280
 281         output_label("kallsyms_num_syms");
 282         printf("\tPTR\t%d\n", table_cnt);
 283         printf("\n");
 284
 285         /* table of offset markers, that give the offset in the compressed stream
 286          * every 256 symbols */
 287         markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
 288         if (!markers) {
 289                 fprintf(stderr, "kallsyms failure: "
 290                         "unable to allocate required memory\n");
 291                 exit(EXIT_FAILURE);
 292         }
 293
 294         output_label("kallsyms_names");
 295         off = 0;
 296         for (i = 0; i < table_cnt; i++) {
 297                 if ((i & 0xFF) == 0)
 298                         markers[i >> 8] = off;
 299
 300                 printf("\t.byte 0x%02x", table[i].len);
 301                 for (k = 0; k < table[i].len; k++)
 302                         printf(", 0x%02x", table[i].sym[k]);
 303                 printf("\n");
 304
 305                 off += table[i].len + 1;
 306         }
 307         printf("\n");
 308
 309         output_label("kallsyms_markers");
 310         for (i = 0; i < ((table_cnt + 255) >> 8); i++)
 311                 printf("\tPTR\t%d\n", markers[i]);
 312         printf("\n");
 313
 314         free(markers);
 315
 316         output_label("kallsyms_token_table");
 317         off = 0;
 318         for (i = 0; i < 256; i++) {
 319                 best_idx[i] = off;
 320                 expand_symbol(best_table[i], best_table_len[i], buf);
 321                 printf("\t.asciz\t\"%s\"\n", buf);
 322                 off += strlen(buf) + 1;
 323         }
 324         printf("\n");
 325
 326         output_label("kallsyms_token_index");
 327         for (i = 0; i < 256; i++)
 328                 printf("\t.short\t%d\n", best_idx[i]);
 329         printf("\n");
 330 }
 331
 332
 333 /* table lookup compression functions */
 334
 335 /* count all the possible tokens in a symbol */
 336 static void learn_symbol(unsigned char *symbol, int len)
 337 {
 338         int i;
 339
 340         for (i = 0; i < len - 1; i++)
 341                 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
 342 }
 343
 344 /* decrease the count for all the possible tokens in a symbol */
 345 static void forget_symbol(unsigned char *symbol, int len)
 346 {
 347         int i;
 348
 349         for (i = 0; i < len - 1; i++)
 350                 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
 351 }
 352
 353 /* remove all the invalid symbols from the table and do the initial token count */
 354 static void build_initial_tok_table(void)
 355 {
 356         unsigned int i, pos;
 357
 358         pos = 0;
 359         for (i = 0; i < table_cnt; i++) {
 360                 if ( symbol_valid(&table[i]) ) {
 361                         if (pos != i)
 362                                 table[pos] = table[i];
 363                         learn_symbol(table[pos].sym, table[pos].len);
 364                         pos++;
 365                 }
 366         }
 367         table_cnt = pos;
 368 }
 369
 370 static void *find_token(unsigned char *str, int len, unsigned char *token)
 371 {
 372         int i;
 373
 374         for (i = 0; i < len - 1; i++) {
 375                 if (str[i] == token[0] && str[i+1] == token[1])
 376                         return &str[i];
 377         }
 378         return NULL;
 379 }
 380
 381 /* replace a given token in all the valid symbols. Use the sampled symbols
 382  * to update the counts */
 383 static void compress_symbols(unsigned char *str, int idx)
 384 {
 385         unsigned int i, len, size;
 386         unsigned char *p1, *p2;
 387
 388         for (i = 0; i < table_cnt; i++) {
 389
 390                 len = table[i].len;
 391                 p1 = table[i].sym;
 392
 393                 /* find the token on the symbol */
 394                 p2 = find_token(p1, len, str);
 395                 if (!p2) continue;
 396
 397                 /* decrease the counts for this symbol's tokens */
 398                 forget_symbol(table[i].sym, len);
 399
 400                 size = len;
 401
 402                 do {
 403                         *p2 = idx;
 404                         p2++;
 405                         size -= (p2 - p1);
 406                         memmove(p2, p2 + 1, size);
 407                         p1 = p2;
 408                         len--;
 409
 410                         if (size < 2) break;
 411
 412                         /* find the token on the symbol */
 413                         p2 = find_token(p1, size, str);
 414
 415                 } while (p2);
 416
 417                 table[i].len = len;
 418
 419                 /* increase the counts for this symbol's new tokens */
 420                 learn_symbol(table[i].sym, len);
 421         }
 422 }
 423
 424 /* search the token with the maximum profit */
 425 static int find_best_token(void)
 426 {
 427         int i, best, bestprofit;
 428
 429         bestprofit=-10000;
 430         best = 0;
 431
 432         for (i = 0; i < 0x10000; i++) {
 433                 if (token_profit[i] > bestprofit) {
 434                         best = i;
 435                         bestprofit = token_profit[i];
 436                 }
 437         }
 438         return best;
 439 }
 440
 441 /* this is the core of the algorithm: calculate the "best" table */
 442 static void optimize_result(void)
 443 {
 444         int i, best;
 445
 446         /* using the '\0' symbol last allows compress_symbols to use standard
 447          * fast string functions */
 448         for (i = 255; i >= 0; i--) {
 449
 450                 /* if this table slot is empty (it is not used by an actual
 451                  * original char code */
 452                 if (!best_table_len[i]) {
 453
 454                         /* find the token with the breates profit value */
 455                         best = find_best_token();
 456
 457                         /* place it in the "best" table */
 458                         best_table_len[i] = 2;
 459                         best_table[i][0] = best & 0xFF;
 460                         best_table[i][1] = (best >> 8) & 0xFF;
 461
 462                         /* replace this token in all the valid symbols */
 463                         compress_symbols(best_table[i], i);
 464                 }
 465         }
 466 }
 467
 468 /* start by placing the symbols that are actually used on the table */
 469 static void insert_real_symbols_in_table(void)
 470 {
 471         unsigned int i, j, c;
 472
 473         memset(best_table, 0, sizeof(best_table));
 474         memset(best_table_len, 0, sizeof(best_table_len));
 475
 476         for (i = 0; i < table_cnt; i++) {
 477                 for (j = 0; j < table[i].len; j++) {
 478                         c = table[i].sym[j];
 479                         best_table[c][0]=c;
 480                         best_table_len[c]=1;
 481                 }
 482         }
 483 }
 484
 485 static void optimize_token_table(void)
 486 {
 487         build_initial_tok_table();
 488
 489         insert_real_symbols_in_table();
 490
 491         /* When valid symbol is not registered, exit to error */
 492         if (!table_cnt) {
 493                 fprintf(stderr, "No valid symbol.\n");
 494                 exit(1);
 495         }
 496
 497         optimize_result();
 498 }
 499
 500 static int compare_symbols(const void *a, const void *b)
 501 {
 502         const struct sym_entry *sa;
 503         const struct sym_entry *sb;
 504         int wa, wb;
 505
 506         sa = a;
 507         sb = b;
 508
 509         /* sort by address first */
 510         if (sa->addr > sb->addr)
 511                 return 1;
 512         if (sa->addr < sb->addr)
 513                 return -1;
 514
 515         /* sort by "weakness" type */
 516         wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
 517         wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
 518         if (wa != wb)
 519                 return wa - wb;
 520
 521         /* sort by initial order, so that other symbols are left undisturbed */
 522         return sa->start_pos - sb->start_pos;
 523 }
 524
 525 static void sort_symbols(void)
 526 {
 527         qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
 528 }
 529
 530 int main(int argc, char **argv)
 531 {
 532         if (argc >= 2) {
 533                 int i;
 534                 for (i = 1; i < argc; i++) {
 535                         if(strcmp(argv[i], "--all-symbols") == 0)
 536                                 all_symbols = 1;
 537                         else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
 538                                 char *p = &argv[i][16];
 539                                 /* skip quote */
 540                                 if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
 541                                         p++;
 542                                 symbol_prefix_char = *p;
 543                         } else
 544                                 usage();
 545                 }
 546         } else if (argc != 1)
 547                 usage();
 548
 549         read_map(stdin);
 550         sort_symbols();
 551         optimize_token_table();
 552         write_src();
 553
 554         return 0;
 555 }