Commit | Line | Data |
---|---|---|
1da177e4 LT |
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 | * | |
adc40221 | 8 | * Usage: kallsyms [--all-symbols] [--absolute-percpu] |
79549da6 | 9 | * [--base-relative] [--lto-clang] in.map > out.S |
1da177e4 | 10 | * |
1da177e4 LT |
11 | * Table compression uses all the unused char codes on the symbols and |
12 | * maps these to the most used substrings (tokens). For instance, it might | |
13 | * map char code 0xF7 to represent "write_" and then in every symbol where | |
14 | * "write_" appears it can be replaced by 0xF7, saving 5 bytes. | |
15 | * The used codes themselves are also placed in the table so that the | |
16 | * decompresion can work without "special cases". | |
17 | * Applied to kernel symbols, this usually produces a compression ratio | |
18 | * of about 50%. | |
19 | * | |
20 | */ | |
21 | ||
1c975da5 | 22 | #include <errno.h> |
aa221f2e | 23 | #include <getopt.h> |
a41333e0 | 24 | #include <stdbool.h> |
1da177e4 LT |
25 | #include <stdio.h> |
26 | #include <stdlib.h> | |
27 | #include <string.h> | |
28 | #include <ctype.h> | |
2213e9a6 | 29 | #include <limits.h> |
1da177e4 | 30 | |
17b1f0de | 31 | #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof(arr[0])) |
17b1f0de | 32 | |
b8a94bfb | 33 | #define KSYM_NAME_LEN 512 |
1da177e4 | 34 | |
1da177e4 LT |
35 | struct sym_entry { |
36 | unsigned long long addr; | |
b3dbb4ec | 37 | unsigned int len; |
60443c88 | 38 | unsigned int seq; |
f2df3f65 | 39 | unsigned int start_pos; |
8c996940 | 40 | unsigned int percpu_absolute; |
9d82973e | 41 | unsigned char sym[]; |
1da177e4 LT |
42 | }; |
43 | ||
78eb7159 KC |
44 | struct addr_range { |
45 | const char *start_sym, *end_sym; | |
17b1f0de MF |
46 | unsigned long long start, end; |
47 | }; | |
48 | ||
49 | static unsigned long long _text; | |
2213e9a6 | 50 | static unsigned long long relative_base; |
78eb7159 | 51 | static struct addr_range text_ranges[] = { |
17b1f0de MF |
52 | { "_stext", "_etext" }, |
53 | { "_sinittext", "_einittext" }, | |
17b1f0de MF |
54 | }; |
55 | #define text_range_text (&text_ranges[0]) | |
56 | #define text_range_inittext (&text_ranges[1]) | |
57 | ||
c6bda7c9 RR |
58 | static struct addr_range percpu_range = { |
59 | "__per_cpu_start", "__per_cpu_end", -1ULL, 0 | |
60 | }; | |
61 | ||
8d605269 | 62 | static struct sym_entry **table; |
b3dbb4ec | 63 | static unsigned int table_size, table_cnt; |
831362fc MY |
64 | static int all_symbols; |
65 | static int absolute_percpu; | |
66 | static int base_relative; | |
010a0aad | 67 | static int lto_clang; |
1da177e4 | 68 | |
f43e9daa | 69 | static int token_profit[0x10000]; |
1da177e4 LT |
70 | |
71 | /* the table that holds the result of the compression */ | |
f43e9daa MY |
72 | static unsigned char best_table[256][2]; |
73 | static unsigned char best_table_len[256]; | |
1da177e4 LT |
74 | |
75 | ||
b3dbb4ec | 76 | static void usage(void) |
1da177e4 | 77 | { |
8d3a7507 | 78 | fprintf(stderr, "Usage: kallsyms [--all-symbols] [--absolute-percpu] " |
010a0aad | 79 | "[--base-relative] [--lto-clang] in.map > out.S\n"); |
1da177e4 LT |
80 | exit(1); |
81 | } | |
82 | ||
29e55ad3 MY |
83 | static char *sym_name(const struct sym_entry *s) |
84 | { | |
85 | return (char *)s->sym + 1; | |
86 | } | |
87 | ||
a41333e0 MY |
88 | static bool is_ignored_symbol(const char *name, char type) |
89 | { | |
a7b00a18 | 90 | if (type == 'u' || type == 'n') |
887df76d MY |
91 | return true; |
92 | ||
93 | if (toupper(type) == 'A') { | |
94 | /* Keep these useful absolute symbols */ | |
95 | if (strcmp(name, "__kernel_syscall_via_break") && | |
96 | strcmp(name, "__kernel_syscall_via_epc") && | |
97 | strcmp(name, "__kernel_sigtramp") && | |
98 | strcmp(name, "__gp")) | |
99 | return true; | |
100 | } | |
101 | ||
a41333e0 MY |
102 | return false; |
103 | } | |
104 | ||
b6233d0d MY |
105 | static void check_symbol_range(const char *sym, unsigned long long addr, |
106 | struct addr_range *ranges, int entries) | |
17b1f0de MF |
107 | { |
108 | size_t i; | |
78eb7159 | 109 | struct addr_range *ar; |
17b1f0de | 110 | |
78eb7159 KC |
111 | for (i = 0; i < entries; ++i) { |
112 | ar = &ranges[i]; | |
17b1f0de | 113 | |
78eb7159 KC |
114 | if (strcmp(sym, ar->start_sym) == 0) { |
115 | ar->start = addr; | |
b6233d0d | 116 | return; |
78eb7159 KC |
117 | } else if (strcmp(sym, ar->end_sym) == 0) { |
118 | ar->end = addr; | |
b6233d0d | 119 | return; |
17b1f0de MF |
120 | } |
121 | } | |
17b1f0de MF |
122 | } |
123 | ||
1c975da5 | 124 | static struct sym_entry *read_symbol(FILE *in, char **buf, size_t *buf_len) |
1da177e4 | 125 | { |
1c975da5 | 126 | char *name, type, *p; |
8d605269 | 127 | unsigned long long addr; |
1c975da5 MY |
128 | size_t len; |
129 | ssize_t readlen; | |
8d605269 | 130 | struct sym_entry *sym; |
1da177e4 | 131 | |
23835308 | 132 | errno = 0; |
1c975da5 MY |
133 | readlen = getline(buf, buf_len, in); |
134 | if (readlen < 0) { | |
135 | if (errno) { | |
136 | perror("read_symbol"); | |
137 | exit(EXIT_FAILURE); | |
138 | } | |
8d605269 | 139 | return NULL; |
1da177e4 | 140 | } |
1c975da5 MY |
141 | |
142 | if ((*buf)[readlen - 1] == '\n') | |
143 | (*buf)[readlen - 1] = 0; | |
144 | ||
145 | addr = strtoull(*buf, &p, 16); | |
146 | ||
147 | if (*buf == p || *p++ != ' ' || !isascii((type = *p++)) || *p++ != ' ') { | |
148 | fprintf(stderr, "line format error\n"); | |
149 | exit(EXIT_FAILURE); | |
150 | } | |
151 | ||
152 | name = p; | |
153 | len = strlen(name); | |
154 | ||
155 | if (len >= KSYM_NAME_LEN) { | |
6db2983c | 156 | fprintf(stderr, "Symbol %s too long for kallsyms (%zu >= %d).\n" |
bb66fc67 | 157 | "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n", |
1c975da5 | 158 | name, len, KSYM_NAME_LEN); |
8d605269 | 159 | return NULL; |
f3462aa9 | 160 | } |
1da177e4 | 161 | |
be9f6133 | 162 | if (strcmp(name, "_text") == 0) |
8d605269 | 163 | _text = addr; |
b6233d0d | 164 | |
7883a143 MP |
165 | /* Ignore most absolute/undefined (?) symbols. */ |
166 | if (is_ignored_symbol(name, type)) | |
167 | return NULL; | |
168 | ||
8d605269 MY |
169 | check_symbol_range(name, addr, text_ranges, ARRAY_SIZE(text_ranges)); |
170 | check_symbol_range(name, addr, &percpu_range, 1); | |
1da177e4 LT |
171 | |
172 | /* include the type field in the symbol name, so that it gets | |
173 | * compressed together */ | |
1c975da5 | 174 | len++; |
8d605269 | 175 | |
9d1b3895 | 176 | sym = malloc(sizeof(*sym) + len + 1); |
8d605269 | 177 | if (!sym) { |
f1a136e0 JJ |
178 | fprintf(stderr, "kallsyms failure: " |
179 | "unable to allocate required amount of memory\n"); | |
180 | exit(EXIT_FAILURE); | |
181 | } | |
8d605269 MY |
182 | sym->addr = addr; |
183 | sym->len = len; | |
184 | sym->sym[0] = type; | |
9d1b3895 | 185 | strcpy(sym_name(sym), name); |
8d605269 | 186 | sym->percpu_absolute = 0; |
8c996940 | 187 | |
8d605269 | 188 | return sym; |
1da177e4 LT |
189 | } |
190 | ||
4bfe2b78 MY |
191 | static int symbol_in_range(const struct sym_entry *s, |
192 | const struct addr_range *ranges, int entries) | |
17b1f0de MF |
193 | { |
194 | size_t i; | |
4bfe2b78 | 195 | const struct addr_range *ar; |
17b1f0de | 196 | |
78eb7159 KC |
197 | for (i = 0; i < entries; ++i) { |
198 | ar = &ranges[i]; | |
17b1f0de | 199 | |
78eb7159 | 200 | if (s->addr >= ar->start && s->addr <= ar->end) |
ac6ca5c8 | 201 | return 1; |
17b1f0de MF |
202 | } |
203 | ||
ac6ca5c8 | 204 | return 0; |
17b1f0de MF |
205 | } |
206 | ||
4bfe2b78 | 207 | static int symbol_valid(const struct sym_entry *s) |
1da177e4 | 208 | { |
29e55ad3 | 209 | const char *name = sym_name(s); |
bd8b22d2 | 210 | |
1da177e4 LT |
211 | /* if --all-symbols is not specified, then symbols outside the text |
212 | * and inittext sections are discarded */ | |
213 | if (!all_symbols) { | |
78eb7159 KC |
214 | if (symbol_in_range(s, text_ranges, |
215 | ARRAY_SIZE(text_ranges)) == 0) | |
1da177e4 LT |
216 | return 0; |
217 | /* Corner case. Discard any symbols with the same value as | |
a3b81113 RG |
218 | * _etext _einittext; they can move between pass 1 and 2 when |
219 | * the kallsyms data are added. If these symbols move then | |
220 | * they may get dropped in pass 2, which breaks the kallsyms | |
221 | * rules. | |
1da177e4 | 222 | */ |
17b1f0de | 223 | if ((s->addr == text_range_text->end && |
29e55ad3 | 224 | strcmp(name, text_range_text->end_sym)) || |
17b1f0de | 225 | (s->addr == text_range_inittext->end && |
29e55ad3 | 226 | strcmp(name, text_range_inittext->end_sym))) |
1da177e4 LT |
227 | return 0; |
228 | } | |
229 | ||
1da177e4 LT |
230 | return 1; |
231 | } | |
232 | ||
5e5c4fa7 MY |
233 | /* remove all the invalid symbols from the table */ |
234 | static void shrink_table(void) | |
235 | { | |
236 | unsigned int i, pos; | |
237 | ||
238 | pos = 0; | |
239 | for (i = 0; i < table_cnt; i++) { | |
8d605269 | 240 | if (symbol_valid(table[i])) { |
5e5c4fa7 MY |
241 | if (pos != i) |
242 | table[pos] = table[i]; | |
243 | pos++; | |
244 | } else { | |
8d605269 | 245 | free(table[i]); |
5e5c4fa7 MY |
246 | } |
247 | } | |
248 | table_cnt = pos; | |
249 | ||
250 | /* When valid symbol is not registered, exit to error */ | |
251 | if (!table_cnt) { | |
252 | fprintf(stderr, "No valid symbol.\n"); | |
253 | exit(1); | |
254 | } | |
255 | } | |
256 | ||
aa221f2e | 257 | static void read_map(const char *in) |
1da177e4 | 258 | { |
aa221f2e | 259 | FILE *fp; |
8d605269 | 260 | struct sym_entry *sym; |
1c975da5 MY |
261 | char *buf = NULL; |
262 | size_t buflen = 0; | |
8d605269 | 263 | |
aa221f2e MY |
264 | fp = fopen(in, "r"); |
265 | if (!fp) { | |
266 | perror(in); | |
267 | exit(1); | |
268 | } | |
269 | ||
270 | while (!feof(fp)) { | |
1c975da5 | 271 | sym = read_symbol(fp, &buf, &buflen); |
8d605269 MY |
272 | if (!sym) |
273 | continue; | |
274 | ||
275 | sym->start_pos = table_cnt; | |
276 | ||
b3dbb4ec PM |
277 | if (table_cnt >= table_size) { |
278 | table_size += 10000; | |
279 | table = realloc(table, sizeof(*table) * table_size); | |
1da177e4 LT |
280 | if (!table) { |
281 | fprintf(stderr, "out of memory\n"); | |
aa221f2e | 282 | fclose(fp); |
1da177e4 LT |
283 | exit (1); |
284 | } | |
285 | } | |
8d605269 MY |
286 | |
287 | table[table_cnt++] = sym; | |
1da177e4 | 288 | } |
aa221f2e | 289 | |
1c975da5 | 290 | free(buf); |
aa221f2e | 291 | fclose(fp); |
1da177e4 LT |
292 | } |
293 | ||
4bfe2b78 | 294 | static void output_label(const char *label) |
1da177e4 | 295 | { |
534c9f2e | 296 | printf(".globl %s\n", label); |
1da177e4 | 297 | printf("\tALGN\n"); |
534c9f2e | 298 | printf("%s:\n", label); |
1da177e4 LT |
299 | } |
300 | ||
fd2ab2f6 MY |
301 | /* Provide proper symbols relocatability by their '_text' relativeness. */ |
302 | static void output_address(unsigned long long addr) | |
303 | { | |
304 | if (_text <= addr) | |
305 | printf("\tPTR\t_text + %#llx\n", addr - _text); | |
306 | else | |
307 | printf("\tPTR\t_text - %#llx\n", _text - addr); | |
308 | } | |
309 | ||
1da177e4 LT |
310 | /* uncompress a compressed symbol. When this function is called, the best table |
311 | * might still be compressed itself, so the function needs to be recursive */ | |
4bfe2b78 | 312 | static int expand_symbol(const unsigned char *data, int len, char *result) |
1da177e4 LT |
313 | { |
314 | int c, rlen, total=0; | |
315 | ||
316 | while (len) { | |
317 | c = *data; | |
318 | /* if the table holds a single char that is the same as the one | |
319 | * we are looking for, then end the search */ | |
320 | if (best_table[c][0]==c && best_table_len[c]==1) { | |
321 | *result++ = c; | |
322 | total++; | |
323 | } else { | |
324 | /* if not, recurse and expand */ | |
325 | rlen = expand_symbol(best_table[c], best_table_len[c], result); | |
326 | total += rlen; | |
327 | result += rlen; | |
328 | } | |
329 | data++; | |
330 | len--; | |
331 | } | |
332 | *result=0; | |
333 | ||
334 | return total; | |
335 | } | |
336 | ||
4bfe2b78 | 337 | static int symbol_absolute(const struct sym_entry *s) |
78eb7159 | 338 | { |
8c996940 | 339 | return s->percpu_absolute; |
78eb7159 KC |
340 | } |
341 | ||
010a0aad ZL |
342 | static void cleanup_symbol_name(char *s) |
343 | { | |
344 | char *p; | |
345 | ||
010a0aad ZL |
346 | /* |
347 | * ASCII[.] = 2e | |
348 | * ASCII[0-9] = 30,39 | |
349 | * ASCII[A-Z] = 41,5a | |
350 | * ASCII[_] = 5f | |
351 | * ASCII[a-z] = 61,7a | |
352 | * | |
8cc32a9b YS |
353 | * As above, replacing the first '.' in ".llvm." with '\0' does not |
354 | * affect the main sorting, but it helps us with subsorting. | |
010a0aad | 355 | */ |
8cc32a9b | 356 | p = strstr(s, ".llvm."); |
010a0aad ZL |
357 | if (p) |
358 | *p = '\0'; | |
359 | } | |
360 | ||
60443c88 ZL |
361 | static int compare_names(const void *a, const void *b) |
362 | { | |
363 | int ret; | |
60443c88 ZL |
364 | const struct sym_entry *sa = *(const struct sym_entry **)a; |
365 | const struct sym_entry *sb = *(const struct sym_entry **)b; | |
366 | ||
dd1553b8 | 367 | ret = strcmp(sym_name(sa), sym_name(sb)); |
60443c88 ZL |
368 | if (!ret) { |
369 | if (sa->addr > sb->addr) | |
370 | return 1; | |
371 | else if (sa->addr < sb->addr) | |
372 | return -1; | |
373 | ||
374 | /* keep old order */ | |
375 | return (int)(sa->seq - sb->seq); | |
376 | } | |
377 | ||
378 | return ret; | |
379 | } | |
380 | ||
381 | static void sort_symbols_by_name(void) | |
382 | { | |
383 | qsort(table, table_cnt, sizeof(table[0]), compare_names); | |
384 | } | |
385 | ||
b3dbb4ec | 386 | static void write_src(void) |
1da177e4 | 387 | { |
b3dbb4ec | 388 | unsigned int i, k, off; |
1da177e4 LT |
389 | unsigned int best_idx[256]; |
390 | unsigned int *markers; | |
9281acea | 391 | char buf[KSYM_NAME_LEN]; |
1da177e4 | 392 | |
500193ec | 393 | printf("#include <asm/bitsperlong.h>\n"); |
1da177e4 LT |
394 | printf("#if BITS_PER_LONG == 64\n"); |
395 | printf("#define PTR .quad\n"); | |
72d3ebb9 | 396 | printf("#define ALGN .balign 8\n"); |
1da177e4 LT |
397 | printf("#else\n"); |
398 | printf("#define PTR .long\n"); | |
72d3ebb9 | 399 | printf("#define ALGN .balign 4\n"); |
1da177e4 LT |
400 | printf("#endif\n"); |
401 | ||
aad09470 | 402 | printf("\t.section .rodata, \"a\"\n"); |
1da177e4 | 403 | |
1da177e4 | 404 | output_label("kallsyms_num_syms"); |
80ffbaa5 | 405 | printf("\t.long\t%u\n", table_cnt); |
1da177e4 LT |
406 | printf("\n"); |
407 | ||
408 | /* table of offset markers, that give the offset in the compressed stream | |
409 | * every 256 symbols */ | |
f1a136e0 JJ |
410 | markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256)); |
411 | if (!markers) { | |
412 | fprintf(stderr, "kallsyms failure: " | |
413 | "unable to allocate required memory\n"); | |
414 | exit(EXIT_FAILURE); | |
415 | } | |
1da177e4 LT |
416 | |
417 | output_label("kallsyms_names"); | |
1da177e4 | 418 | off = 0; |
b3dbb4ec PM |
419 | for (i = 0; i < table_cnt; i++) { |
420 | if ((i & 0xFF) == 0) | |
421 | markers[i >> 8] = off; | |
60443c88 | 422 | table[i]->seq = i; |
1da177e4 | 423 | |
73bbb944 MO |
424 | /* There cannot be any symbol of length zero. */ |
425 | if (table[i]->len == 0) { | |
426 | fprintf(stderr, "kallsyms failure: " | |
427 | "unexpected zero symbol length\n"); | |
428 | exit(EXIT_FAILURE); | |
429 | } | |
430 | ||
431 | /* Only lengths that fit in up-to-two-byte ULEB128 are supported. */ | |
432 | if (table[i]->len > 0x3FFF) { | |
433 | fprintf(stderr, "kallsyms failure: " | |
434 | "unexpected huge symbol length\n"); | |
435 | exit(EXIT_FAILURE); | |
436 | } | |
437 | ||
438 | /* Encode length with ULEB128. */ | |
439 | if (table[i]->len <= 0x7F) { | |
440 | /* Most symbols use a single byte for the length. */ | |
441 | printf("\t.byte 0x%02x", table[i]->len); | |
442 | off += table[i]->len + 1; | |
443 | } else { | |
444 | /* "Big" symbols use two bytes. */ | |
445 | printf("\t.byte 0x%02x, 0x%02x", | |
446 | (table[i]->len & 0x7F) | 0x80, | |
447 | (table[i]->len >> 7) & 0x7F); | |
448 | off += table[i]->len + 2; | |
449 | } | |
8d605269 MY |
450 | for (k = 0; k < table[i]->len; k++) |
451 | printf(", 0x%02x", table[i]->sym[k]); | |
1da177e4 | 452 | printf("\n"); |
1da177e4 LT |
453 | } |
454 | printf("\n"); | |
455 | ||
dd1553b8 MY |
456 | /* |
457 | * Now that we wrote out the compressed symbol names, restore the | |
458 | * original names, which are needed in some of the later steps. | |
459 | */ | |
460 | for (i = 0; i < table_cnt; i++) { | |
461 | expand_symbol(table[i]->sym, table[i]->len, buf); | |
462 | strcpy((char *)table[i]->sym, buf); | |
463 | } | |
464 | ||
1da177e4 | 465 | output_label("kallsyms_markers"); |
b3dbb4ec | 466 | for (i = 0; i < ((table_cnt + 255) >> 8); i++) |
80ffbaa5 | 467 | printf("\t.long\t%u\n", markers[i]); |
1da177e4 LT |
468 | printf("\n"); |
469 | ||
470 | free(markers); | |
471 | ||
472 | output_label("kallsyms_token_table"); | |
473 | off = 0; | |
474 | for (i = 0; i < 256; i++) { | |
475 | best_idx[i] = off; | |
b3dbb4ec | 476 | expand_symbol(best_table[i], best_table_len[i], buf); |
1da177e4 LT |
477 | printf("\t.asciz\t\"%s\"\n", buf); |
478 | off += strlen(buf) + 1; | |
479 | } | |
480 | printf("\n"); | |
481 | ||
482 | output_label("kallsyms_token_index"); | |
483 | for (i = 0; i < 256; i++) | |
484 | printf("\t.short\t%d\n", best_idx[i]); | |
485 | printf("\n"); | |
404bad70 MY |
486 | |
487 | if (!base_relative) | |
488 | output_label("kallsyms_addresses"); | |
489 | else | |
490 | output_label("kallsyms_offsets"); | |
491 | ||
492 | for (i = 0; i < table_cnt; i++) { | |
493 | if (base_relative) { | |
494 | /* | |
495 | * Use the offset relative to the lowest value | |
496 | * encountered of all relative symbols, and emit | |
497 | * non-relocatable fixed offsets that will be fixed | |
498 | * up at runtime. | |
499 | */ | |
500 | ||
501 | long long offset; | |
502 | int overflow; | |
503 | ||
504 | if (!absolute_percpu) { | |
505 | offset = table[i]->addr - relative_base; | |
506 | overflow = (offset < 0 || offset > UINT_MAX); | |
507 | } else if (symbol_absolute(table[i])) { | |
508 | offset = table[i]->addr; | |
509 | overflow = (offset < 0 || offset > INT_MAX); | |
510 | } else { | |
511 | offset = relative_base - table[i]->addr - 1; | |
512 | overflow = (offset < INT_MIN || offset >= 0); | |
513 | } | |
514 | if (overflow) { | |
515 | fprintf(stderr, "kallsyms failure: " | |
516 | "%s symbol value %#llx out of range in relative mode\n", | |
517 | symbol_absolute(table[i]) ? "absolute" : "relative", | |
518 | table[i]->addr); | |
519 | exit(EXIT_FAILURE); | |
520 | } | |
dd1553b8 | 521 | printf("\t.long\t%#x /* %s */\n", (int)offset, table[i]->sym); |
404bad70 MY |
522 | } else if (!symbol_absolute(table[i])) { |
523 | output_address(table[i]->addr); | |
524 | } else { | |
525 | printf("\tPTR\t%#llx\n", table[i]->addr); | |
526 | } | |
527 | } | |
528 | printf("\n"); | |
529 | ||
530 | if (base_relative) { | |
531 | output_label("kallsyms_relative_base"); | |
532 | output_address(relative_base); | |
533 | printf("\n"); | |
534 | } | |
535 | ||
dd1553b8 MY |
536 | if (lto_clang) |
537 | for (i = 0; i < table_cnt; i++) | |
538 | cleanup_symbol_name((char *)table[i]->sym); | |
539 | ||
404bad70 MY |
540 | sort_symbols_by_name(); |
541 | output_label("kallsyms_seqs_of_names"); | |
542 | for (i = 0; i < table_cnt; i++) | |
543 | printf("\t.byte 0x%02x, 0x%02x, 0x%02x\n", | |
544 | (unsigned char)(table[i]->seq >> 16), | |
545 | (unsigned char)(table[i]->seq >> 8), | |
546 | (unsigned char)(table[i]->seq >> 0)); | |
547 | printf("\n"); | |
1da177e4 LT |
548 | } |
549 | ||
550 | ||
551 | /* table lookup compression functions */ | |
552 | ||
1da177e4 | 553 | /* count all the possible tokens in a symbol */ |
4bfe2b78 | 554 | static void learn_symbol(const unsigned char *symbol, int len) |
1da177e4 LT |
555 | { |
556 | int i; | |
557 | ||
558 | for (i = 0; i < len - 1; i++) | |
b3dbb4ec | 559 | token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++; |
1da177e4 LT |
560 | } |
561 | ||
562 | /* decrease the count for all the possible tokens in a symbol */ | |
4bfe2b78 | 563 | static void forget_symbol(const unsigned char *symbol, int len) |
1da177e4 LT |
564 | { |
565 | int i; | |
566 | ||
567 | for (i = 0; i < len - 1; i++) | |
b3dbb4ec | 568 | token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--; |
1da177e4 LT |
569 | } |
570 | ||
5e5c4fa7 | 571 | /* do the initial token count */ |
fcdf7197 | 572 | static void build_initial_token_table(void) |
1da177e4 | 573 | { |
5e5c4fa7 | 574 | unsigned int i; |
1da177e4 | 575 | |
5e5c4fa7 | 576 | for (i = 0; i < table_cnt; i++) |
8d605269 | 577 | learn_symbol(table[i]->sym, table[i]->len); |
1da177e4 LT |
578 | } |
579 | ||
2558c138 | 580 | static unsigned char *find_token(unsigned char *str, int len, |
4bfe2b78 | 581 | const unsigned char *token) |
7c5d249a PM |
582 | { |
583 | int i; | |
584 | ||
585 | for (i = 0; i < len - 1; i++) { | |
586 | if (str[i] == token[0] && str[i+1] == token[1]) | |
587 | return &str[i]; | |
588 | } | |
589 | return NULL; | |
590 | } | |
591 | ||
1da177e4 LT |
592 | /* replace a given token in all the valid symbols. Use the sampled symbols |
593 | * to update the counts */ | |
4bfe2b78 | 594 | static void compress_symbols(const unsigned char *str, int idx) |
1da177e4 | 595 | { |
b3dbb4ec PM |
596 | unsigned int i, len, size; |
597 | unsigned char *p1, *p2; | |
1da177e4 | 598 | |
b3dbb4ec | 599 | for (i = 0; i < table_cnt; i++) { |
1da177e4 | 600 | |
8d605269 MY |
601 | len = table[i]->len; |
602 | p1 = table[i]->sym; | |
b3dbb4ec PM |
603 | |
604 | /* find the token on the symbol */ | |
7c5d249a | 605 | p2 = find_token(p1, len, str); |
b3dbb4ec PM |
606 | if (!p2) continue; |
607 | ||
608 | /* decrease the counts for this symbol's tokens */ | |
8d605269 | 609 | forget_symbol(table[i]->sym, len); |
b3dbb4ec PM |
610 | |
611 | size = len; | |
1da177e4 LT |
612 | |
613 | do { | |
b3dbb4ec PM |
614 | *p2 = idx; |
615 | p2++; | |
616 | size -= (p2 - p1); | |
617 | memmove(p2, p2 + 1, size); | |
618 | p1 = p2; | |
619 | len--; | |
620 | ||
621 | if (size < 2) break; | |
622 | ||
1da177e4 | 623 | /* find the token on the symbol */ |
7c5d249a | 624 | p2 = find_token(p1, size, str); |
1da177e4 | 625 | |
b3dbb4ec | 626 | } while (p2); |
1da177e4 | 627 | |
8d605269 | 628 | table[i]->len = len; |
1da177e4 | 629 | |
b3dbb4ec | 630 | /* increase the counts for this symbol's new tokens */ |
8d605269 | 631 | learn_symbol(table[i]->sym, len); |
1da177e4 LT |
632 | } |
633 | } | |
634 | ||
635 | /* search the token with the maximum profit */ | |
b3dbb4ec | 636 | static int find_best_token(void) |
1da177e4 | 637 | { |
b3dbb4ec | 638 | int i, best, bestprofit; |
1da177e4 LT |
639 | |
640 | bestprofit=-10000; | |
b3dbb4ec | 641 | best = 0; |
1da177e4 | 642 | |
b3dbb4ec PM |
643 | for (i = 0; i < 0x10000; i++) { |
644 | if (token_profit[i] > bestprofit) { | |
645 | best = i; | |
646 | bestprofit = token_profit[i]; | |
1da177e4 | 647 | } |
1da177e4 | 648 | } |
1da177e4 LT |
649 | return best; |
650 | } | |
651 | ||
652 | /* this is the core of the algorithm: calculate the "best" table */ | |
653 | static void optimize_result(void) | |
654 | { | |
b3dbb4ec | 655 | int i, best; |
1da177e4 LT |
656 | |
657 | /* using the '\0' symbol last allows compress_symbols to use standard | |
658 | * fast string functions */ | |
659 | for (i = 255; i >= 0; i--) { | |
660 | ||
661 | /* if this table slot is empty (it is not used by an actual | |
662 | * original char code */ | |
663 | if (!best_table_len[i]) { | |
664 | ||
cbf7a90e | 665 | /* find the token with the best profit value */ |
1da177e4 | 666 | best = find_best_token(); |
e0a04b11 XW |
667 | if (token_profit[best] == 0) |
668 | break; | |
1da177e4 LT |
669 | |
670 | /* place it in the "best" table */ | |
b3dbb4ec PM |
671 | best_table_len[i] = 2; |
672 | best_table[i][0] = best & 0xFF; | |
673 | best_table[i][1] = (best >> 8) & 0xFF; | |
1da177e4 LT |
674 | |
675 | /* replace this token in all the valid symbols */ | |
b3dbb4ec | 676 | compress_symbols(best_table[i], i); |
1da177e4 LT |
677 | } |
678 | } | |
679 | } | |
680 | ||
681 | /* start by placing the symbols that are actually used on the table */ | |
682 | static void insert_real_symbols_in_table(void) | |
683 | { | |
b3dbb4ec | 684 | unsigned int i, j, c; |
1da177e4 | 685 | |
b3dbb4ec | 686 | for (i = 0; i < table_cnt; i++) { |
8d605269 MY |
687 | for (j = 0; j < table[i]->len; j++) { |
688 | c = table[i]->sym[j]; | |
b3dbb4ec PM |
689 | best_table[c][0]=c; |
690 | best_table_len[c]=1; | |
1da177e4 LT |
691 | } |
692 | } | |
693 | } | |
694 | ||
695 | static void optimize_token_table(void) | |
696 | { | |
fcdf7197 | 697 | build_initial_token_table(); |
1da177e4 LT |
698 | |
699 | insert_real_symbols_in_table(); | |
700 | ||
701 | optimize_result(); | |
702 | } | |
703 | ||
b478b782 LJ |
704 | /* guess for "linker script provide" symbol */ |
705 | static int may_be_linker_script_provide_symbol(const struct sym_entry *se) | |
706 | { | |
29e55ad3 | 707 | const char *symbol = sym_name(se); |
b478b782 LJ |
708 | int len = se->len - 1; |
709 | ||
710 | if (len < 8) | |
711 | return 0; | |
712 | ||
713 | if (symbol[0] != '_' || symbol[1] != '_') | |
714 | return 0; | |
715 | ||
716 | /* __start_XXXXX */ | |
717 | if (!memcmp(symbol + 2, "start_", 6)) | |
718 | return 1; | |
719 | ||
720 | /* __stop_XXXXX */ | |
721 | if (!memcmp(symbol + 2, "stop_", 5)) | |
722 | return 1; | |
723 | ||
724 | /* __end_XXXXX */ | |
725 | if (!memcmp(symbol + 2, "end_", 4)) | |
726 | return 1; | |
727 | ||
728 | /* __XXXXX_start */ | |
729 | if (!memcmp(symbol + len - 6, "_start", 6)) | |
730 | return 1; | |
731 | ||
732 | /* __XXXXX_end */ | |
733 | if (!memcmp(symbol + len - 4, "_end", 4)) | |
734 | return 1; | |
735 | ||
736 | return 0; | |
737 | } | |
738 | ||
f2df3f65 PM |
739 | static int compare_symbols(const void *a, const void *b) |
740 | { | |
8d605269 MY |
741 | const struct sym_entry *sa = *(const struct sym_entry **)a; |
742 | const struct sym_entry *sb = *(const struct sym_entry **)b; | |
f2df3f65 PM |
743 | int wa, wb; |
744 | ||
f2df3f65 PM |
745 | /* sort by address first */ |
746 | if (sa->addr > sb->addr) | |
747 | return 1; | |
748 | if (sa->addr < sb->addr) | |
749 | return -1; | |
750 | ||
751 | /* sort by "weakness" type */ | |
752 | wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W'); | |
753 | wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W'); | |
754 | if (wa != wb) | |
755 | return wa - wb; | |
756 | ||
b478b782 LJ |
757 | /* sort by "linker script provide" type */ |
758 | wa = may_be_linker_script_provide_symbol(sa); | |
759 | wb = may_be_linker_script_provide_symbol(sb); | |
760 | if (wa != wb) | |
761 | return wa - wb; | |
762 | ||
763 | /* sort by the number of prefix underscores */ | |
aa915245 MY |
764 | wa = strspn(sym_name(sa), "_"); |
765 | wb = strspn(sym_name(sb), "_"); | |
b478b782 LJ |
766 | if (wa != wb) |
767 | return wa - wb; | |
768 | ||
f2df3f65 PM |
769 | /* sort by initial order, so that other symbols are left undisturbed */ |
770 | return sa->start_pos - sb->start_pos; | |
771 | } | |
772 | ||
773 | static void sort_symbols(void) | |
774 | { | |
8d605269 | 775 | qsort(table, table_cnt, sizeof(table[0]), compare_symbols); |
f2df3f65 | 776 | } |
1da177e4 | 777 | |
c6bda7c9 RR |
778 | static void make_percpus_absolute(void) |
779 | { | |
780 | unsigned int i; | |
781 | ||
782 | for (i = 0; i < table_cnt; i++) | |
8d605269 | 783 | if (symbol_in_range(table[i], &percpu_range, 1)) { |
8c996940 AB |
784 | /* |
785 | * Keep the 'A' override for percpu symbols to | |
786 | * ensure consistent behavior compared to older | |
787 | * versions of this tool. | |
788 | */ | |
8d605269 MY |
789 | table[i]->sym[0] = 'A'; |
790 | table[i]->percpu_absolute = 1; | |
8c996940 | 791 | } |
c6bda7c9 RR |
792 | } |
793 | ||
2213e9a6 AB |
794 | /* find the minimum non-absolute symbol address */ |
795 | static void record_relative_base(void) | |
796 | { | |
797 | unsigned int i; | |
798 | ||
2213e9a6 | 799 | for (i = 0; i < table_cnt; i++) |
8d605269 | 800 | if (!symbol_absolute(table[i])) { |
f34ea029 MY |
801 | /* |
802 | * The table is sorted by address. | |
803 | * Take the first non-absolute symbol value. | |
804 | */ | |
8d605269 | 805 | relative_base = table[i]->addr; |
f34ea029 MY |
806 | return; |
807 | } | |
2213e9a6 AB |
808 | } |
809 | ||
b3dbb4ec | 810 | int main(int argc, char **argv) |
1da177e4 | 811 | { |
aa221f2e | 812 | while (1) { |
92e74fb6 | 813 | static const struct option long_options[] = { |
aa221f2e MY |
814 | {"all-symbols", no_argument, &all_symbols, 1}, |
815 | {"absolute-percpu", no_argument, &absolute_percpu, 1}, | |
816 | {"base-relative", no_argument, &base_relative, 1}, | |
010a0aad | 817 | {"lto-clang", no_argument, <o_clang, 1}, |
aa221f2e MY |
818 | {}, |
819 | }; | |
820 | ||
821 | int c = getopt_long(argc, argv, "", long_options, NULL); | |
822 | ||
823 | if (c == -1) | |
824 | break; | |
825 | if (c != 0) | |
826 | usage(); | |
827 | } | |
828 | ||
829 | if (optind >= argc) | |
1da177e4 LT |
830 | usage(); |
831 | ||
aa221f2e | 832 | read_map(argv[optind]); |
5e5c4fa7 | 833 | shrink_table(); |
c6bda7c9 RR |
834 | if (absolute_percpu) |
835 | make_percpus_absolute(); | |
f34ea029 | 836 | sort_symbols(); |
2213e9a6 AB |
837 | if (base_relative) |
838 | record_relative_base(); | |
2ea03891 | 839 | optimize_token_table(); |
1da177e4 LT |
840 | write_src(); |
841 | ||
842 | return 0; | |
843 | } |