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 #include <limits.h>
26
27 #ifndef ARRAY_SIZE
28 #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof(arr[0]))
29 #endif
30
31 #define KSYM_NAME_LEN 128
32
33 struct sym_entry {
34 unsigned long long addr;
35 unsigned int len;
36 unsigned int start_pos;
37 unsigned char *sym;
38 unsigned int percpu_absolute;
39 };
40
41 struct addr_range {
42 const char *start_sym, *end_sym;
43 unsigned long long start, end;
44 };
45
46 static unsigned long long _text;
47 static unsigned long long relative_base;
48 static struct addr_range text_ranges[] = {
49 { "_stext", "_etext" },
50 { "_sinittext", "_einittext" },
51 { "_stext_l1", "_etext_l1" }, /* Blackfin on-chip L1 inst SRAM */
52 { "_stext_l2", "_etext_l2" }, /* Blackfin on-chip L2 SRAM */
53 };
54 #define text_range_text (&text_ranges[0])
55 #define text_range_inittext (&text_ranges[1])
56
57 static struct addr_range percpu_range = {
58 "__per_cpu_start", "__per_cpu_end", -1ULL, 0
59 };
60
61 static struct sym_entry *table;
62 static unsigned int table_size, table_cnt;
63 static int all_symbols = 0;
64 static int absolute_percpu = 0;
65 static int base_relative = 0;
66
67 int token_profit[0x10000];
68
69 /* the table that holds the result of the compression */
70 unsigned char best_table[256][2];
71 unsigned char best_table_len[256];
72
73
usage(void)74 static void usage(void)
75 {
76 fprintf(stderr, "Usage: kallsyms [--all-symbols] "
77 "[--base-relative] < in.map > out.S\n");
78 exit(1);
79 }
80
81 /*
82 * This ignores the intensely annoying "mapping symbols" found
83 * in ARM ELF files: $a, $t and $d.
84 */
is_arm_mapping_symbol(const char * str)85 static inline int is_arm_mapping_symbol(const char *str)
86 {
87 return str[0] == '$' && strchr("axtd", str[1])
88 && (str[2] == '\0' || str[2] == '.');
89 }
90
check_symbol_range(const char * sym,unsigned long long addr,struct addr_range * ranges,int entries)91 static int check_symbol_range(const char *sym, unsigned long long addr,
92 struct addr_range *ranges, int entries)
93 {
94 size_t i;
95 struct addr_range *ar;
96
97 for (i = 0; i < entries; ++i) {
98 ar = &ranges[i];
99
100 if (strcmp(sym, ar->start_sym) == 0) {
101 ar->start = addr;
102 return 0;
103 } else if (strcmp(sym, ar->end_sym) == 0) {
104 ar->end = addr;
105 return 0;
106 }
107 }
108
109 return 1;
110 }
111
read_symbol(FILE * in,struct sym_entry * s)112 static int read_symbol(FILE *in, struct sym_entry *s)
113 {
114 char sym[500], stype;
115 int rc;
116
117 rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, sym);
118 if (rc != 3) {
119 if (rc != EOF && fgets(sym, 500, in) == NULL)
120 fprintf(stderr, "Read error or end of file.\n");
121 return -1;
122 }
123 if (strlen(sym) > KSYM_NAME_LEN) {
124 fprintf(stderr, "Symbol %s too long for kallsyms (%zu vs %d).\n"
125 "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n",
126 sym, strlen(sym), KSYM_NAME_LEN);
127 return -1;
128 }
129
130 /* Ignore most absolute/undefined (?) symbols. */
131 if (strcmp(sym, "_text") == 0)
132 _text = s->addr;
133 else if (check_symbol_range(sym, s->addr, text_ranges,
134 ARRAY_SIZE(text_ranges)) == 0)
135 /* nothing to do */;
136 else if (toupper(stype) == 'A')
137 {
138 /* Keep these useful absolute symbols */
139 if (strcmp(sym, "__kernel_syscall_via_break") &&
140 strcmp(sym, "__kernel_syscall_via_epc") &&
141 strcmp(sym, "__kernel_sigtramp") &&
142 strcmp(sym, "__gp"))
143 return -1;
144
145 }
146 else if (toupper(stype) == 'U' ||
147 is_arm_mapping_symbol(sym))
148 return -1;
149 /* exclude also MIPS ELF local symbols ($L123 instead of .L123) */
150 else if (sym[0] == '$')
151 return -1;
152 /* exclude debugging symbols */
153 else if (stype == 'N' || stype == 'n')
154 return -1;
155
156 /* include the type field in the symbol name, so that it gets
157 * compressed together */
158 s->len = strlen(sym) + 1;
159 s->sym = malloc(s->len + 1);
160 if (!s->sym) {
161 fprintf(stderr, "kallsyms failure: "
162 "unable to allocate required amount of memory\n");
163 exit(EXIT_FAILURE);
164 }
165 strcpy((char *)s->sym + 1, sym);
166 s->sym[0] = stype;
167
168 s->percpu_absolute = 0;
169
170 /* Record if we've found __per_cpu_start/end. */
171 check_symbol_range(sym, s->addr, &percpu_range, 1);
172
173 return 0;
174 }
175
symbol_in_range(struct sym_entry * s,struct addr_range * ranges,int entries)176 static int symbol_in_range(struct sym_entry *s, struct addr_range *ranges,
177 int entries)
178 {
179 size_t i;
180 struct addr_range *ar;
181
182 for (i = 0; i < entries; ++i) {
183 ar = &ranges[i];
184
185 if (s->addr >= ar->start && s->addr <= ar->end)
186 return 1;
187 }
188
189 return 0;
190 }
191
symbol_valid(struct sym_entry * s)192 static int symbol_valid(struct sym_entry *s)
193 {
194 /* Symbols which vary between passes. Passes 1 and 2 must have
195 * identical symbol lists. The kallsyms_* symbols below are only added
196 * after pass 1, they would be included in pass 2 when --all-symbols is
197 * specified so exclude them to get a stable symbol list.
198 */
199 static char *special_symbols[] = {
200 "kallsyms_addresses",
201 "kallsyms_offsets",
202 "kallsyms_relative_base",
203 "kallsyms_num_syms",
204 "kallsyms_names",
205 "kallsyms_markers",
206 "kallsyms_token_table",
207 "kallsyms_token_index",
208
209 /* Exclude linker generated symbols which vary between passes */
210 "_SDA_BASE_", /* ppc */
211 "_SDA2_BASE_", /* ppc */
212 NULL };
213
214 static char *special_prefixes[] = {
215 "__crc_", /* modversions */
216 "__efistub_", /* arm64 EFI stub namespace */
217 NULL };
218
219 static char *special_suffixes[] = {
220 "_veneer", /* arm */
221 "_from_arm", /* arm */
222 "_from_thumb", /* arm */
223 NULL };
224
225 int i;
226 char *sym_name = (char *)s->sym + 1;
227
228 /* if --all-symbols is not specified, then symbols outside the text
229 * and inittext sections are discarded */
230 if (!all_symbols) {
231 if (symbol_in_range(s, text_ranges,
232 ARRAY_SIZE(text_ranges)) == 0)
233 return 0;
234 /* Corner case. Discard any symbols with the same value as
235 * _etext _einittext; they can move between pass 1 and 2 when
236 * the kallsyms data are added. If these symbols move then
237 * they may get dropped in pass 2, which breaks the kallsyms
238 * rules.
239 */
240 if ((s->addr == text_range_text->end &&
241 strcmp(sym_name,
242 text_range_text->end_sym)) ||
243 (s->addr == text_range_inittext->end &&
244 strcmp(sym_name,
245 text_range_inittext->end_sym)))
246 return 0;
247 }
248
249 /* Exclude symbols which vary between passes. */
250 for (i = 0; special_symbols[i]; i++)
251 if (strcmp(sym_name, special_symbols[i]) == 0)
252 return 0;
253
254 for (i = 0; special_prefixes[i]; i++) {
255 int l = strlen(special_prefixes[i]);
256
257 if (l <= strlen(sym_name) &&
258 strncmp(sym_name, special_prefixes[i], l) == 0)
259 return 0;
260 }
261
262 for (i = 0; special_suffixes[i]; i++) {
263 int l = strlen(sym_name) - strlen(special_suffixes[i]);
264
265 if (l >= 0 && strcmp(sym_name + l, special_suffixes[i]) == 0)
266 return 0;
267 }
268
269 return 1;
270 }
271
read_map(FILE * in)272 static void read_map(FILE *in)
273 {
274 while (!feof(in)) {
275 if (table_cnt >= table_size) {
276 table_size += 10000;
277 table = realloc(table, sizeof(*table) * table_size);
278 if (!table) {
279 fprintf(stderr, "out of memory\n");
280 exit (1);
281 }
282 }
283 if (read_symbol(in, &table[table_cnt]) == 0) {
284 table[table_cnt].start_pos = table_cnt;
285 table_cnt++;
286 }
287 }
288 }
289
output_label(char * label)290 static void output_label(char *label)
291 {
292 printf(".globl %s\n", label);
293 printf("\tALGN\n");
294 printf("%s:\n", label);
295 }
296
297 /* uncompress a compressed symbol. When this function is called, the best table
298 * might still be compressed itself, so the function needs to be recursive */
expand_symbol(unsigned char * data,int len,char * result)299 static int expand_symbol(unsigned char *data, int len, char *result)
300 {
301 int c, rlen, total=0;
302
303 while (len) {
304 c = *data;
305 /* if the table holds a single char that is the same as the one
306 * we are looking for, then end the search */
307 if (best_table[c][0]==c && best_table_len[c]==1) {
308 *result++ = c;
309 total++;
310 } else {
311 /* if not, recurse and expand */
312 rlen = expand_symbol(best_table[c], best_table_len[c], result);
313 total += rlen;
314 result += rlen;
315 }
316 data++;
317 len--;
318 }
319 *result=0;
320
321 return total;
322 }
323
symbol_absolute(struct sym_entry * s)324 static int symbol_absolute(struct sym_entry *s)
325 {
326 return s->percpu_absolute;
327 }
328
write_src(void)329 static void write_src(void)
330 {
331 unsigned int i, k, off;
332 unsigned int best_idx[256];
333 unsigned int *markers;
334 char buf[KSYM_NAME_LEN];
335
336 printf("#include <asm/types.h>\n");
337 printf("#if BITS_PER_LONG == 64\n");
338 printf("#define PTR .quad\n");
339 printf("#define ALGN .align 8\n");
340 printf("#else\n");
341 printf("#define PTR .long\n");
342 printf("#define ALGN .align 4\n");
343 printf("#endif\n");
344
345 printf("\t.section .rodata, \"a\"\n");
346
347 /* Provide proper symbols relocatability by their relativeness
348 * to a fixed anchor point in the runtime image, either '_text'
349 * for absolute address tables, in which case the linker will
350 * emit the final addresses at build time. Otherwise, use the
351 * offset relative to the lowest value encountered of all relative
352 * symbols, and emit non-relocatable fixed offsets that will be fixed
353 * up at runtime.
354 *
355 * The symbol names cannot be used to construct normal symbol
356 * references as the list of symbols contains symbols that are
357 * declared static and are private to their .o files. This prevents
358 * .tmp_kallsyms.o or any other object from referencing them.
359 */
360 if (!base_relative)
361 output_label("kallsyms_addresses");
362 else
363 output_label("kallsyms_offsets");
364
365 for (i = 0; i < table_cnt; i++) {
366 if (base_relative) {
367 long long offset;
368 int overflow;
369
370 if (!absolute_percpu) {
371 offset = table[i].addr - relative_base;
372 overflow = (offset < 0 || offset > UINT_MAX);
373 } else if (symbol_absolute(&table[i])) {
374 offset = table[i].addr;
375 overflow = (offset < 0 || offset > INT_MAX);
376 } else {
377 offset = relative_base - table[i].addr - 1;
378 overflow = (offset < INT_MIN || offset >= 0);
379 }
380 if (overflow) {
381 fprintf(stderr, "kallsyms failure: "
382 "%s symbol value %#llx out of range in relative mode\n",
383 symbol_absolute(&table[i]) ? "absolute" : "relative",
384 table[i].addr);
385 exit(EXIT_FAILURE);
386 }
387 printf("\t.long\t%#x\n", (int)offset);
388 } else if (!symbol_absolute(&table[i])) {
389 if (_text <= table[i].addr)
390 printf("\tPTR\t_text + %#llx\n",
391 table[i].addr - _text);
392 else
393 printf("\tPTR\t_text - %#llx\n",
394 _text - table[i].addr);
395 } else {
396 printf("\tPTR\t%#llx\n", table[i].addr);
397 }
398 }
399 printf("\n");
400
401 if (base_relative) {
402 output_label("kallsyms_relative_base");
403 printf("\tPTR\t_text - %#llx\n", _text - relative_base);
404 printf("\n");
405 }
406
407 output_label("kallsyms_num_syms");
408 printf("\tPTR\t%u\n", table_cnt);
409 printf("\n");
410
411 /* table of offset markers, that give the offset in the compressed stream
412 * every 256 symbols */
413 markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
414 if (!markers) {
415 fprintf(stderr, "kallsyms failure: "
416 "unable to allocate required memory\n");
417 exit(EXIT_FAILURE);
418 }
419
420 output_label("kallsyms_names");
421 off = 0;
422 for (i = 0; i < table_cnt; i++) {
423 if ((i & 0xFF) == 0)
424 markers[i >> 8] = off;
425
426 printf("\t.byte 0x%02x", table[i].len);
427 for (k = 0; k < table[i].len; k++)
428 printf(", 0x%02x", table[i].sym[k]);
429 printf("\n");
430
431 off += table[i].len + 1;
432 }
433 printf("\n");
434
435 output_label("kallsyms_markers");
436 for (i = 0; i < ((table_cnt + 255) >> 8); i++)
437 printf("\tPTR\t%d\n", markers[i]);
438 printf("\n");
439
440 free(markers);
441
442 output_label("kallsyms_token_table");
443 off = 0;
444 for (i = 0; i < 256; i++) {
445 best_idx[i] = off;
446 expand_symbol(best_table[i], best_table_len[i], buf);
447 printf("\t.asciz\t\"%s\"\n", buf);
448 off += strlen(buf) + 1;
449 }
450 printf("\n");
451
452 output_label("kallsyms_token_index");
453 for (i = 0; i < 256; i++)
454 printf("\t.short\t%d\n", best_idx[i]);
455 printf("\n");
456 }
457
458
459 /* table lookup compression functions */
460
461 /* count all the possible tokens in a symbol */
learn_symbol(unsigned char * symbol,int len)462 static void learn_symbol(unsigned char *symbol, int len)
463 {
464 int i;
465
466 for (i = 0; i < len - 1; i++)
467 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
468 }
469
470 /* decrease the count for all the possible tokens in a symbol */
forget_symbol(unsigned char * symbol,int len)471 static void forget_symbol(unsigned char *symbol, int len)
472 {
473 int i;
474
475 for (i = 0; i < len - 1; i++)
476 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
477 }
478
479 /* remove all the invalid symbols from the table and do the initial token count */
build_initial_tok_table(void)480 static void build_initial_tok_table(void)
481 {
482 unsigned int i, pos;
483
484 pos = 0;
485 for (i = 0; i < table_cnt; i++) {
486 if ( symbol_valid(&table[i]) ) {
487 if (pos != i)
488 table[pos] = table[i];
489 learn_symbol(table[pos].sym, table[pos].len);
490 pos++;
491 }
492 }
493 table_cnt = pos;
494 }
495
find_token(unsigned char * str,int len,unsigned char * token)496 static void *find_token(unsigned char *str, int len, unsigned char *token)
497 {
498 int i;
499
500 for (i = 0; i < len - 1; i++) {
501 if (str[i] == token[0] && str[i+1] == token[1])
502 return &str[i];
503 }
504 return NULL;
505 }
506
507 /* replace a given token in all the valid symbols. Use the sampled symbols
508 * to update the counts */
compress_symbols(unsigned char * str,int idx)509 static void compress_symbols(unsigned char *str, int idx)
510 {
511 unsigned int i, len, size;
512 unsigned char *p1, *p2;
513
514 for (i = 0; i < table_cnt; i++) {
515
516 len = table[i].len;
517 p1 = table[i].sym;
518
519 /* find the token on the symbol */
520 p2 = find_token(p1, len, str);
521 if (!p2) continue;
522
523 /* decrease the counts for this symbol's tokens */
524 forget_symbol(table[i].sym, len);
525
526 size = len;
527
528 do {
529 *p2 = idx;
530 p2++;
531 size -= (p2 - p1);
532 memmove(p2, p2 + 1, size);
533 p1 = p2;
534 len--;
535
536 if (size < 2) break;
537
538 /* find the token on the symbol */
539 p2 = find_token(p1, size, str);
540
541 } while (p2);
542
543 table[i].len = len;
544
545 /* increase the counts for this symbol's new tokens */
546 learn_symbol(table[i].sym, len);
547 }
548 }
549
550 /* search the token with the maximum profit */
find_best_token(void)551 static int find_best_token(void)
552 {
553 int i, best, bestprofit;
554
555 bestprofit=-10000;
556 best = 0;
557
558 for (i = 0; i < 0x10000; i++) {
559 if (token_profit[i] > bestprofit) {
560 best = i;
561 bestprofit = token_profit[i];
562 }
563 }
564 return best;
565 }
566
567 /* this is the core of the algorithm: calculate the "best" table */
optimize_result(void)568 static void optimize_result(void)
569 {
570 int i, best;
571
572 /* using the '\0' symbol last allows compress_symbols to use standard
573 * fast string functions */
574 for (i = 255; i >= 0; i--) {
575
576 /* if this table slot is empty (it is not used by an actual
577 * original char code */
578 if (!best_table_len[i]) {
579
580 /* find the token with the best profit value */
581 best = find_best_token();
582 if (token_profit[best] == 0)
583 break;
584
585 /* place it in the "best" table */
586 best_table_len[i] = 2;
587 best_table[i][0] = best & 0xFF;
588 best_table[i][1] = (best >> 8) & 0xFF;
589
590 /* replace this token in all the valid symbols */
591 compress_symbols(best_table[i], i);
592 }
593 }
594 }
595
596 /* start by placing the symbols that are actually used on the table */
insert_real_symbols_in_table(void)597 static void insert_real_symbols_in_table(void)
598 {
599 unsigned int i, j, c;
600
601 memset(best_table, 0, sizeof(best_table));
602 memset(best_table_len, 0, sizeof(best_table_len));
603
604 for (i = 0; i < table_cnt; i++) {
605 for (j = 0; j < table[i].len; j++) {
606 c = table[i].sym[j];
607 best_table[c][0]=c;
608 best_table_len[c]=1;
609 }
610 }
611 }
612
optimize_token_table(void)613 static void optimize_token_table(void)
614 {
615 build_initial_tok_table();
616
617 insert_real_symbols_in_table();
618
619 /* When valid symbol is not registered, exit to error */
620 if (!table_cnt) {
621 fprintf(stderr, "No valid symbol.\n");
622 exit(1);
623 }
624
625 optimize_result();
626 }
627
628 /* guess for "linker script provide" symbol */
may_be_linker_script_provide_symbol(const struct sym_entry * se)629 static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
630 {
631 const char *symbol = (char *)se->sym + 1;
632 int len = se->len - 1;
633
634 if (len < 8)
635 return 0;
636
637 if (symbol[0] != '_' || symbol[1] != '_')
638 return 0;
639
640 /* __start_XXXXX */
641 if (!memcmp(symbol + 2, "start_", 6))
642 return 1;
643
644 /* __stop_XXXXX */
645 if (!memcmp(symbol + 2, "stop_", 5))
646 return 1;
647
648 /* __end_XXXXX */
649 if (!memcmp(symbol + 2, "end_", 4))
650 return 1;
651
652 /* __XXXXX_start */
653 if (!memcmp(symbol + len - 6, "_start", 6))
654 return 1;
655
656 /* __XXXXX_end */
657 if (!memcmp(symbol + len - 4, "_end", 4))
658 return 1;
659
660 return 0;
661 }
662
prefix_underscores_count(const char * str)663 static int prefix_underscores_count(const char *str)
664 {
665 const char *tail = str;
666
667 while (*tail == '_')
668 tail++;
669
670 return tail - str;
671 }
672
compare_symbols(const void * a,const void * b)673 static int compare_symbols(const void *a, const void *b)
674 {
675 const struct sym_entry *sa;
676 const struct sym_entry *sb;
677 int wa, wb;
678
679 sa = a;
680 sb = b;
681
682 /* sort by address first */
683 if (sa->addr > sb->addr)
684 return 1;
685 if (sa->addr < sb->addr)
686 return -1;
687
688 /* sort by "weakness" type */
689 wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
690 wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
691 if (wa != wb)
692 return wa - wb;
693
694 /* sort by "linker script provide" type */
695 wa = may_be_linker_script_provide_symbol(sa);
696 wb = may_be_linker_script_provide_symbol(sb);
697 if (wa != wb)
698 return wa - wb;
699
700 /* sort by the number of prefix underscores */
701 wa = prefix_underscores_count((const char *)sa->sym + 1);
702 wb = prefix_underscores_count((const char *)sb->sym + 1);
703 if (wa != wb)
704 return wa - wb;
705
706 /* sort by initial order, so that other symbols are left undisturbed */
707 return sa->start_pos - sb->start_pos;
708 }
709
sort_symbols(void)710 static void sort_symbols(void)
711 {
712 qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
713 }
714
make_percpus_absolute(void)715 static void make_percpus_absolute(void)
716 {
717 unsigned int i;
718
719 for (i = 0; i < table_cnt; i++)
720 if (symbol_in_range(&table[i], &percpu_range, 1)) {
721 /*
722 * Keep the 'A' override for percpu symbols to
723 * ensure consistent behavior compared to older
724 * versions of this tool.
725 */
726 table[i].sym[0] = 'A';
727 table[i].percpu_absolute = 1;
728 }
729 }
730
731 /* find the minimum non-absolute symbol address */
record_relative_base(void)732 static void record_relative_base(void)
733 {
734 unsigned int i;
735
736 relative_base = -1ULL;
737 for (i = 0; i < table_cnt; i++)
738 if (!symbol_absolute(&table[i]) &&
739 table[i].addr < relative_base)
740 relative_base = table[i].addr;
741 }
742
main(int argc,char ** argv)743 int main(int argc, char **argv)
744 {
745 if (argc >= 2) {
746 int i;
747 for (i = 1; i < argc; i++) {
748 if(strcmp(argv[i], "--all-symbols") == 0)
749 all_symbols = 1;
750 else if (strcmp(argv[i], "--absolute-percpu") == 0)
751 absolute_percpu = 1;
752 else if (strcmp(argv[i], "--base-relative") == 0)
753 base_relative = 1;
754 else
755 usage();
756 }
757 } else if (argc != 1)
758 usage();
759
760 read_map(stdin);
761 if (absolute_percpu)
762 make_percpus_absolute();
763 if (base_relative)
764 record_relative_base();
765 sort_symbols();
766 optimize_token_table();
767 write_src();
768
769 return 0;
770 }
771
