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         /* exclude debugging symbols */
 112         else if (stype == 'N')
 113                 return -1;
 114
 115         /* include the type field in the symbol name, so that it gets
 116          * compressed together */
 117         s->len = strlen(str) + 1;
 118         s->sym = malloc(s->len + 1);
 119         if (!s->sym) {
 120                 fprintf(stderr, "kallsyms failure: "
 121                         "unable to allocate required amount of memory\n");
 122                 exit(EXIT_FAILURE);
 123         }
 124         strcpy((char *)s->sym + 1, str);
 125         s->sym[0] = stype;
 126
 127         return 0;
 128 }
 129
 130 static int symbol_valid(struct sym_entry *s)
 131 {
 132         /* Symbols which vary between passes.  Passes 1 and 2 must have
 133          * identical symbol lists.
 134          */
 135         static char *special_symbols[] = {
 136         /* Exclude linker generated symbols which vary between passes */
 137                 "_SDA_BASE_",           /* ppc */
 138                 "_SDA2_BASE_",          /* ppc */
 139                 NULL };
 140         int i;
 141         int offset = 1;
 142
 143         /* skip prefix char */
 144         if (symbol_prefix_char && *(s->sym + 1) == symbol_prefix_char)
 145                 offset++;
 146
 147         /* if --all-symbols is not specified, then symbols outside the text
 148          * and inittext sections are discarded */
 149         if (!all_symbols) {
 150                 if ((s->addr < _stext || s->addr > _etext)
 151                     && (s->addr < _sinittext || s->addr > _einittext))
 152                         return 0;
 153                 /* Corner case.  Discard any symbols with the same value as
 154                  * _etext _einittext; they can move between pass 1 and 2 when
 155                  * the kallsyms data are added.  If these symbols move then
 156                  * they may get dropped in pass 2, which breaks the kallsyms
 157                  * rules.
 158                  */
 159                 if ((s->addr == _etext &&
 160                                 strcmp((char *)s->sym + offset, "_etext")) ||
 161                     (s->addr == _einittext &&
 162                                 strcmp((char *)s->sym + offset, "_einittext")))
 163                         return 0;
 164         }
 165
 166         /* Exclude symbols which vary between passes. */
 167         if (strstr((char *)s->sym + offset, "_compiled.") ||
 168             strncmp((char*)s->sym + offset, "__compound_literal.", 19) == 0 ||
 169             strncmp((char*)s->sym + offset, "__compound_literal$", 19) == 0)
 170                 return 0;
 171
 172         for (i = 0; special_symbols[i]; i++)
 173                 if( strcmp((char *)s->sym + offset, special_symbols[i]) == 0 )
 174                         return 0;
 175
 176         return 1;
 177 }
 178
 179 static void read_map(FILE *in)
 180 {
 181         while (!feof(in)) {
 182                 if (table_cnt >= table_size) {
 183                         table_size += 10000;
 184                         table = realloc(table, sizeof(*table) * table_size);
 185                         if (!table) {
 186                                 fprintf(stderr, "out of memory\n");
 187                                 exit (1);
 188                         }
 189                 }
 190                 if (read_symbol(in, &table[table_cnt]) == 0) {
 191                         table[table_cnt].start_pos = table_cnt;
 192                         table_cnt++;
 193                 }
 194         }
 195 }
 196
 197 static void output_label(char *label)
 198 {
 199         if (symbol_prefix_char)
 200                 printf(".globl %c%s\n", symbol_prefix_char, label);
 201         else
 202                 printf(".globl %s\n", label);
 203         printf("\tALGN\n");
 204         if (symbol_prefix_char)
 205                 printf("%c%s:\n", symbol_prefix_char, label);
 206         else
 207                 printf("%s:\n", label);
 208 }
 209
 210 /* uncompress a compressed symbol. When this function is called, the best table
 211  * might still be compressed itself, so the function needs to be recursive */
 212 static int expand_symbol(unsigned char *data, int len, char *result)
 213 {
 214         int c, rlen, total=0;
 215
 216         while (len) {
 217                 c = *data;
 218                 /* if the table holds a single char that is the same as the one
 219                  * we are looking for, then end the search */
 220                 if (best_table[c][0]==c && best_table_len[c]==1) {
 221                         *result++ = c;
 222                         total++;
 223                 } else {
 224                         /* if not, recurse and expand */
 225                         rlen = expand_symbol(best_table[c], best_table_len[c], result);
 226                         total += rlen;
 227                         result += rlen;
 228                 }
 229                 data++;
 230                 len--;
 231         }
 232         *result=0;
 233
 234         return total;
 235 }
 236
 237 static void write_src(void)
 238 {
 239         unsigned int i, k, off;
 240         unsigned int best_idx[256];
 241         unsigned int *markers;
 242         char buf[KSYM_NAME_LEN];
 243
 244         printf("#include <asm/types.h>\n");
 245         printf("#if BITS_PER_LONG == 64\n");
 246         printf("#define PTR .quad\n");
 247         printf("#define ALGN .align 8\n");
 248         printf("#else\n");
 249         printf("#define PTR .long\n");
 250         printf("#define ALGN .align 4\n");
 251         printf("#endif\n");
 252
 253         printf("\t.section .rodata, \"a\"\n");
 254
 255         /* Provide proper symbols relocatability by their '_text'
 256          * relativeness.  The symbol names cannot be used to construct
 257          * normal symbol references as the list of symbols contains
 258          * symbols that are declared static and are private to their
 259          * .o files.  This prevents .tmp_kallsyms.o or any other
 260          * object from referencing them.
 261          */
 262         output_label("kallsyms_addresses");
 263         for (i = 0; i < table_cnt; i++) {
 264                 if (toupper(table[i].sym[0]) != 'A') {
 265                         if (_text <= table[i].addr)
 266                                 printf("\tPTR\t_text + %#llx\n",
 267                                         table[i].addr - _text);
 268                         else
 269                                 printf("\tPTR\t_text - %#llx\n",
 270                                         _text - table[i].addr);
 271                 } else {
 272                         printf("\tPTR\t%#llx\n", table[i].addr);
 273                 }
 274         }
 275         printf("\n");
 276
 277         output_label("kallsyms_num_syms");
 278         printf("\tPTR\t%d\n", table_cnt);
 279         printf("\n");
 280
 281         /* table of offset markers, that give the offset in the compressed stream
 282          * every 256 symbols */
 283         markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
 284         if (!markers) {
 285                 fprintf(stderr, "kallsyms failure: "
 286                         "unable to allocate required memory\n");
 287                 exit(EXIT_FAILURE);
 288         }
 289
 290         output_label("kallsyms_names");
 291         off = 0;
 292         for (i = 0; i < table_cnt; i++) {
 293                 if ((i & 0xFF) == 0)
 294                         markers[i >> 8] = off;
 295
 296                 printf("\t.byte 0x%02x", table[i].len);
 297                 for (k = 0; k < table[i].len; k++)
 298                         printf(", 0x%02x", table[i].sym[k]);
 299                 printf("\n");
 300
 301                 off += table[i].len + 1;
 302         }
 303         printf("\n");
 304
 305         output_label("kallsyms_markers");
 306         for (i = 0; i < ((table_cnt + 255) >> 8); i++)
 307                 printf("\tPTR\t%d\n", markers[i]);
 308         printf("\n");
 309
 310         free(markers);
 311
 312         output_label("kallsyms_token_table");
 313         off = 0;
 314         for (i = 0; i < 256; i++) {
 315                 best_idx[i] = off;
 316                 expand_symbol(best_table[i], best_table_len[i], buf);
 317                 printf("\t.asciz\t\"%s\"\n", buf);
 318                 off += strlen(buf) + 1;
 319         }
 320         printf("\n");
 321
 322         output_label("kallsyms_token_index");
 323         for (i = 0; i < 256; i++)
 324                 printf("\t.short\t%d\n", best_idx[i]);
 325         printf("\n");
 326 }
 327
 328
 329 /* table lookup compression functions */
 330
 331 /* count all the possible tokens in a symbol */
 332 static void learn_symbol(unsigned char *symbol, int len)
 333 {
 334         int i;
 335
 336         for (i = 0; i < len - 1; i++)
 337                 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
 338 }
 339
 340 /* decrease the count for all the possible tokens in a symbol */
 341 static void forget_symbol(unsigned char *symbol, int len)
 342 {
 343         int i;
 344
 345         for (i = 0; i < len - 1; i++)
 346                 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
 347 }
 348
 349 /* remove all the invalid symbols from the table and do the initial token count */
 350 static void build_initial_tok_table(void)
 351 {
 352         unsigned int i, pos;
 353
 354         pos = 0;
 355         for (i = 0; i < table_cnt; i++) {
 356                 if ( symbol_valid(&table[i]) ) {
 357                         if (pos != i)
 358                                 table[pos] = table[i];
 359                         learn_symbol(table[pos].sym, table[pos].len);
 360                         pos++;
 361                 }
 362         }
 363         table_cnt = pos;
 364 }
 365
 366 static void *find_token(unsigned char *str, int len, unsigned char *token)
 367 {
 368         int i;
 369
 370         for (i = 0; i < len - 1; i++) {
 371                 if (str[i] == token[0] && str[i+1] == token[1])
 372                         return &str[i];
 373         }
 374         return NULL;
 375 }
 376
 377 /* replace a given token in all the valid symbols. Use the sampled symbols
 378  * to update the counts */
 379 static void compress_symbols(unsigned char *str, int idx)
 380 {
 381         unsigned int i, len, size;
 382         unsigned char *p1, *p2;
 383
 384         for (i = 0; i < table_cnt; i++) {
 385
 386                 len = table[i].len;
 387                 p1 = table[i].sym;
 388
 389                 /* find the token on the symbol */
 390                 p2 = find_token(p1, len, str);
 391                 if (!p2) continue;
 392
 393                 /* decrease the counts for this symbol's tokens */
 394                 forget_symbol(table[i].sym, len);
 395
 396                 size = len;
 397
 398                 do {
 399                         *p2 = idx;
 400                         p2++;
 401                         size -= (p2 - p1);
 402                         memmove(p2, p2 + 1, size);
 403                         p1 = p2;
 404                         len--;
 405
 406                         if (size < 2) break;
 407
 408                         /* find the token on the symbol */
 409                         p2 = find_token(p1, size, str);
 410
 411                 } while (p2);
 412
 413                 table[i].len = len;
 414
 415                 /* increase the counts for this symbol's new tokens */
 416                 learn_symbol(table[i].sym, len);
 417         }
 418 }
 419
 420 /* search the token with the maximum profit */
 421 static int find_best_token(void)
 422 {
 423         int i, best, bestprofit;
 424
 425         bestprofit=-10000;
 426         best = 0;
 427
 428         for (i = 0; i < 0x10000; i++) {
 429                 if (token_profit[i] > bestprofit) {
 430                         best = i;
 431                         bestprofit = token_profit[i];
 432                 }
 433         }
 434         return best;
 435 }
 436
 437 /* this is the core of the algorithm: calculate the "best" table */
 438 static void optimize_result(void)
 439 {
 440         int i, best;
 441
 442         /* using the '\0' symbol last allows compress_symbols to use standard
 443          * fast string functions */
 444         for (i = 255; i >= 0; i--) {
 445
 446                 /* if this table slot is empty (it is not used by an actual
 447                  * original char code */
 448                 if (!best_table_len[i]) {
 449
 450                         /* find the token with the breates profit value */
 451                         best = find_best_token();
 452
 453                         /* place it in the "best" table */
 454                         best_table_len[i] = 2;
 455                         best_table[i][0] = best & 0xFF;
 456                         best_table[i][1] = (best >> 8) & 0xFF;
 457
 458                         /* replace this token in all the valid symbols */
 459                         compress_symbols(best_table[i], i);
 460                 }
 461         }
 462 }
 463
 464 /* start by placing the symbols that are actually used on the table */
 465 static void insert_real_symbols_in_table(void)
 466 {
 467         unsigned int i, j, c;
 468
 469         memset(best_table, 0, sizeof(best_table));
 470         memset(best_table_len, 0, sizeof(best_table_len));
 471
 472         for (i = 0; i < table_cnt; i++) {
 473                 for (j = 0; j < table[i].len; j++) {
 474                         c = table[i].sym[j];
 475                         best_table[c][0]=c;
 476                         best_table_len[c]=1;
 477                 }
 478         }
 479 }
 480
 481 static void optimize_token_table(void)
 482 {
 483         build_initial_tok_table();
 484
 485         insert_real_symbols_in_table();
 486
 487         /* When valid symbol is not registered, exit to error */
 488         if (!table_cnt) {
 489                 fprintf(stderr, "No valid symbol.\n");
 490                 exit(1);
 491         }
 492
 493         optimize_result();
 494 }
 495
 496 static int compare_symbols(const void *a, const void *b)
 497 {
 498         const struct sym_entry *sa;
 499         const struct sym_entry *sb;
 500         int wa, wb;
 501
 502         sa = a;
 503         sb = b;
 504
 505         /* sort by address first */
 506         if (sa->addr > sb->addr)
 507                 return 1;
 508         if (sa->addr < sb->addr)
 509                 return -1;
 510
 511         /* sort by "weakness" type */
 512         wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
 513         wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
 514         if (wa != wb)
 515                 return wa - wb;
 516
 517         /* sort by initial order, so that other symbols are left undisturbed */
 518         return sa->start_pos - sb->start_pos;
 519 }
 520
 521 static void sort_symbols(void)
 522 {
 523         qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
 524 }
 525
 526 int main(int argc, char **argv)
 527 {
 528         if (argc >= 2) {
 529                 int i;
 530                 for (i = 1; i < argc; i++) {
 531                         if(strcmp(argv[i], "--all-symbols") == 0)
 532                                 all_symbols = 1;
 533                         else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
 534                                 char *p = &argv[i][16];
 535                                 /* skip quote */
 536                                 if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
 537                                         p++;
 538                                 symbol_prefix_char = *p;
 539                         } else
 540                                 usage();
 541                 }
 542         } else if (argc != 1)
 543                 usage();
 544
 545         read_map(stdin);
 546         if (table_cnt) {
 547                 sort_symbols();
 548                 optimize_token_table();
 549         }
 550         write_src();
 551
 552         return 0;
 553 }