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 | * | |
8 | * Usage: nm -n vmlinux | scripts/kallsyms [--all-symbols] > symbols.S | |
9 | * | |
1da177e4 LT |
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 | ||
17b1f0de MF |
26 | #ifndef ARRAY_SIZE |
27 | #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof(arr[0])) | |
28 | #endif | |
29 | ||
9281acea | 30 | #define KSYM_NAME_LEN 128 |
1da177e4 | 31 | |
1da177e4 LT |
32 | struct sym_entry { |
33 | unsigned long long addr; | |
b3dbb4ec | 34 | unsigned int len; |
f2df3f65 | 35 | unsigned int start_pos; |
1da177e4 LT |
36 | unsigned char *sym; |
37 | }; | |
38 | ||
78eb7159 KC |
39 | struct addr_range { |
40 | const char *start_sym, *end_sym; | |
17b1f0de MF |
41 | unsigned long long start, end; |
42 | }; | |
43 | ||
44 | static unsigned long long _text; | |
78eb7159 | 45 | static struct addr_range text_ranges[] = { |
17b1f0de MF |
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 | ||
c6bda7c9 RR |
54 | static struct addr_range percpu_range = { |
55 | "__per_cpu_start", "__per_cpu_end", -1ULL, 0 | |
56 | }; | |
57 | ||
1da177e4 | 58 | static struct sym_entry *table; |
b3dbb4ec | 59 | static unsigned int table_size, table_cnt; |
1da177e4 | 60 | static int all_symbols = 0; |
c6bda7c9 | 61 | static int absolute_percpu = 0; |
41f11a4f | 62 | static char symbol_prefix_char = '\0'; |
f6537f2f | 63 | static unsigned long long kernel_start_addr = 0; |
1da177e4 | 64 | |
b3dbb4ec | 65 | int token_profit[0x10000]; |
1da177e4 LT |
66 | |
67 | /* the table that holds the result of the compression */ | |
b3dbb4ec | 68 | unsigned char best_table[256][2]; |
1da177e4 LT |
69 | unsigned char best_table_len[256]; |
70 | ||
71 | ||
b3dbb4ec | 72 | static void usage(void) |
1da177e4 | 73 | { |
f6537f2f ML |
74 | fprintf(stderr, "Usage: kallsyms [--all-symbols] " |
75 | "[--symbol-prefix=<prefix char>] " | |
76 | "[--page-offset=<CONFIG_PAGE_OFFSET>] " | |
77 | "< in.map > out.S\n"); | |
1da177e4 LT |
78 | exit(1); |
79 | } | |
80 | ||
81 | /* | |
82 | * This ignores the intensely annoying "mapping symbols" found | |
83 | * in ARM ELF files: $a, $t and $d. | |
84 | */ | |
b3dbb4ec | 85 | static inline int is_arm_mapping_symbol(const char *str) |
1da177e4 | 86 | { |
6c34f1f5 | 87 | return str[0] == '$' && strchr("axtd", str[1]) |
1da177e4 LT |
88 | && (str[2] == '\0' || str[2] == '.'); |
89 | } | |
90 | ||
78eb7159 KC |
91 | static int check_symbol_range(const char *sym, unsigned long long addr, |
92 | struct addr_range *ranges, int entries) | |
17b1f0de MF |
93 | { |
94 | size_t i; | |
78eb7159 | 95 | struct addr_range *ar; |
17b1f0de | 96 | |
78eb7159 KC |
97 | for (i = 0; i < entries; ++i) { |
98 | ar = &ranges[i]; | |
17b1f0de | 99 | |
78eb7159 KC |
100 | if (strcmp(sym, ar->start_sym) == 0) { |
101 | ar->start = addr; | |
17b1f0de | 102 | return 0; |
78eb7159 KC |
103 | } else if (strcmp(sym, ar->end_sym) == 0) { |
104 | ar->end = addr; | |
17b1f0de MF |
105 | return 0; |
106 | } | |
107 | } | |
108 | ||
109 | return 1; | |
110 | } | |
111 | ||
b3dbb4ec | 112 | static int read_symbol(FILE *in, struct sym_entry *s) |
1da177e4 LT |
113 | { |
114 | char str[500]; | |
b3dbb4ec | 115 | char *sym, stype; |
1da177e4 LT |
116 | int rc; |
117 | ||
b3dbb4ec | 118 | rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, str); |
1da177e4 | 119 | if (rc != 3) { |
ef894870 JS |
120 | if (rc != EOF && fgets(str, 500, in) == NULL) |
121 | fprintf(stderr, "Read error or end of file.\n"); | |
1da177e4 LT |
122 | return -1; |
123 | } | |
f3462aa9 | 124 | if (strlen(str) > KSYM_NAME_LEN) { |
6f62259b | 125 | fprintf(stderr, "Symbol %s too long for kallsyms (%zu vs %d).\n" |
bb66fc67 | 126 | "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n", |
f3462aa9 AK |
127 | str, strlen(str), KSYM_NAME_LEN); |
128 | return -1; | |
129 | } | |
1da177e4 | 130 | |
41f11a4f YS |
131 | sym = str; |
132 | /* skip prefix char */ | |
133 | if (symbol_prefix_char && str[0] == symbol_prefix_char) | |
134 | sym++; | |
135 | ||
1da177e4 | 136 | /* Ignore most absolute/undefined (?) symbols. */ |
fd593d12 EB |
137 | if (strcmp(sym, "_text") == 0) |
138 | _text = s->addr; | |
78eb7159 KC |
139 | else if (check_symbol_range(sym, s->addr, text_ranges, |
140 | ARRAY_SIZE(text_ranges)) == 0) | |
17b1f0de | 141 | /* nothing to do */; |
b3dbb4ec | 142 | else if (toupper(stype) == 'A') |
1da177e4 LT |
143 | { |
144 | /* Keep these useful absolute symbols */ | |
41f11a4f YS |
145 | if (strcmp(sym, "__kernel_syscall_via_break") && |
146 | strcmp(sym, "__kernel_syscall_via_epc") && | |
147 | strcmp(sym, "__kernel_sigtramp") && | |
148 | strcmp(sym, "__gp")) | |
1da177e4 LT |
149 | return -1; |
150 | ||
151 | } | |
b3dbb4ec | 152 | else if (toupper(stype) == 'U' || |
41f11a4f | 153 | is_arm_mapping_symbol(sym)) |
1da177e4 | 154 | return -1; |
6f00df24 RB |
155 | /* exclude also MIPS ELF local symbols ($L123 instead of .L123) */ |
156 | else if (str[0] == '$') | |
157 | return -1; | |
aab34ac8 SR |
158 | /* exclude debugging symbols */ |
159 | else if (stype == 'N') | |
160 | return -1; | |
1da177e4 LT |
161 | |
162 | /* include the type field in the symbol name, so that it gets | |
163 | * compressed together */ | |
164 | s->len = strlen(str) + 1; | |
b3dbb4ec | 165 | s->sym = malloc(s->len + 1); |
f1a136e0 JJ |
166 | if (!s->sym) { |
167 | fprintf(stderr, "kallsyms failure: " | |
168 | "unable to allocate required amount of memory\n"); | |
169 | exit(EXIT_FAILURE); | |
170 | } | |
b3dbb4ec PM |
171 | strcpy((char *)s->sym + 1, str); |
172 | s->sym[0] = stype; | |
1da177e4 | 173 | |
c6bda7c9 RR |
174 | /* Record if we've found __per_cpu_start/end. */ |
175 | check_symbol_range(sym, s->addr, &percpu_range, 1); | |
176 | ||
1da177e4 LT |
177 | return 0; |
178 | } | |
179 | ||
78eb7159 KC |
180 | static int symbol_in_range(struct sym_entry *s, struct addr_range *ranges, |
181 | int entries) | |
17b1f0de MF |
182 | { |
183 | size_t i; | |
78eb7159 | 184 | struct addr_range *ar; |
17b1f0de | 185 | |
78eb7159 KC |
186 | for (i = 0; i < entries; ++i) { |
187 | ar = &ranges[i]; | |
17b1f0de | 188 | |
78eb7159 | 189 | if (s->addr >= ar->start && s->addr <= ar->end) |
ac6ca5c8 | 190 | return 1; |
17b1f0de MF |
191 | } |
192 | ||
ac6ca5c8 | 193 | return 0; |
17b1f0de MF |
194 | } |
195 | ||
b3dbb4ec | 196 | static int symbol_valid(struct sym_entry *s) |
1da177e4 LT |
197 | { |
198 | /* Symbols which vary between passes. Passes 1 and 2 must have | |
2ea03891 SR |
199 | * identical symbol lists. The kallsyms_* symbols below are only added |
200 | * after pass 1, they would be included in pass 2 when --all-symbols is | |
201 | * specified so exclude them to get a stable symbol list. | |
1da177e4 LT |
202 | */ |
203 | static char *special_symbols[] = { | |
2ea03891 SR |
204 | "kallsyms_addresses", |
205 | "kallsyms_num_syms", | |
206 | "kallsyms_names", | |
207 | "kallsyms_markers", | |
208 | "kallsyms_token_table", | |
209 | "kallsyms_token_index", | |
210 | ||
1da177e4 LT |
211 | /* Exclude linker generated symbols which vary between passes */ |
212 | "_SDA_BASE_", /* ppc */ | |
213 | "_SDA2_BASE_", /* ppc */ | |
214 | NULL }; | |
bd8b22d2 AB |
215 | |
216 | static char *special_suffixes[] = { | |
bd8b22d2 AB |
217 | "_veneer", /* arm */ |
218 | NULL }; | |
219 | ||
1da177e4 | 220 | int i; |
bd8b22d2 AB |
221 | char *sym_name = (char *)s->sym + 1; |
222 | ||
41f11a4f | 223 | |
f6537f2f ML |
224 | if (s->addr < kernel_start_addr) |
225 | return 0; | |
226 | ||
41f11a4f | 227 | /* skip prefix char */ |
bd8b22d2 AB |
228 | if (symbol_prefix_char && *sym_name == symbol_prefix_char) |
229 | sym_name++; | |
230 | ||
1da177e4 LT |
231 | |
232 | /* if --all-symbols is not specified, then symbols outside the text | |
233 | * and inittext sections are discarded */ | |
234 | if (!all_symbols) { | |
78eb7159 KC |
235 | if (symbol_in_range(s, text_ranges, |
236 | ARRAY_SIZE(text_ranges)) == 0) | |
1da177e4 LT |
237 | return 0; |
238 | /* Corner case. Discard any symbols with the same value as | |
a3b81113 RG |
239 | * _etext _einittext; they can move between pass 1 and 2 when |
240 | * the kallsyms data are added. If these symbols move then | |
241 | * they may get dropped in pass 2, which breaks the kallsyms | |
242 | * rules. | |
1da177e4 | 243 | */ |
17b1f0de | 244 | if ((s->addr == text_range_text->end && |
bd8b22d2 | 245 | strcmp(sym_name, |
78eb7159 | 246 | text_range_text->end_sym)) || |
17b1f0de | 247 | (s->addr == text_range_inittext->end && |
bd8b22d2 | 248 | strcmp(sym_name, |
78eb7159 | 249 | text_range_inittext->end_sym))) |
1da177e4 LT |
250 | return 0; |
251 | } | |
252 | ||
253 | /* Exclude symbols which vary between passes. */ | |
1da177e4 | 254 | for (i = 0; special_symbols[i]; i++) |
bd8b22d2 | 255 | if (strcmp(sym_name, special_symbols[i]) == 0) |
1da177e4 LT |
256 | return 0; |
257 | ||
bd8b22d2 AB |
258 | for (i = 0; special_suffixes[i]; i++) { |
259 | int l = strlen(sym_name) - strlen(special_suffixes[i]); | |
260 | ||
261 | if (l >= 0 && strcmp(sym_name + l, special_suffixes[i]) == 0) | |
262 | return 0; | |
263 | } | |
264 | ||
1da177e4 LT |
265 | return 1; |
266 | } | |
267 | ||
b3dbb4ec | 268 | static void read_map(FILE *in) |
1da177e4 LT |
269 | { |
270 | while (!feof(in)) { | |
b3dbb4ec PM |
271 | if (table_cnt >= table_size) { |
272 | table_size += 10000; | |
273 | table = realloc(table, sizeof(*table) * table_size); | |
1da177e4 LT |
274 | if (!table) { |
275 | fprintf(stderr, "out of memory\n"); | |
276 | exit (1); | |
277 | } | |
278 | } | |
f2df3f65 PM |
279 | if (read_symbol(in, &table[table_cnt]) == 0) { |
280 | table[table_cnt].start_pos = table_cnt; | |
b3dbb4ec | 281 | table_cnt++; |
f2df3f65 | 282 | } |
1da177e4 LT |
283 | } |
284 | } | |
285 | ||
286 | static void output_label(char *label) | |
287 | { | |
41f11a4f YS |
288 | if (symbol_prefix_char) |
289 | printf(".globl %c%s\n", symbol_prefix_char, label); | |
290 | else | |
291 | printf(".globl %s\n", label); | |
1da177e4 | 292 | printf("\tALGN\n"); |
41f11a4f YS |
293 | if (symbol_prefix_char) |
294 | printf("%c%s:\n", symbol_prefix_char, label); | |
295 | else | |
296 | printf("%s:\n", label); | |
1da177e4 LT |
297 | } |
298 | ||
299 | /* uncompress a compressed symbol. When this function is called, the best table | |
300 | * might still be compressed itself, so the function needs to be recursive */ | |
301 | static int expand_symbol(unsigned char *data, int len, char *result) | |
302 | { | |
303 | int c, rlen, total=0; | |
304 | ||
305 | while (len) { | |
306 | c = *data; | |
307 | /* if the table holds a single char that is the same as the one | |
308 | * we are looking for, then end the search */ | |
309 | if (best_table[c][0]==c && best_table_len[c]==1) { | |
310 | *result++ = c; | |
311 | total++; | |
312 | } else { | |
313 | /* if not, recurse and expand */ | |
314 | rlen = expand_symbol(best_table[c], best_table_len[c], result); | |
315 | total += rlen; | |
316 | result += rlen; | |
317 | } | |
318 | data++; | |
319 | len--; | |
320 | } | |
321 | *result=0; | |
322 | ||
323 | return total; | |
324 | } | |
325 | ||
78eb7159 KC |
326 | static int symbol_absolute(struct sym_entry *s) |
327 | { | |
328 | return toupper(s->sym[0]) == 'A'; | |
329 | } | |
330 | ||
b3dbb4ec | 331 | static void write_src(void) |
1da177e4 | 332 | { |
b3dbb4ec | 333 | unsigned int i, k, off; |
1da177e4 LT |
334 | unsigned int best_idx[256]; |
335 | unsigned int *markers; | |
9281acea | 336 | char buf[KSYM_NAME_LEN]; |
1da177e4 LT |
337 | |
338 | printf("#include <asm/types.h>\n"); | |
339 | printf("#if BITS_PER_LONG == 64\n"); | |
340 | printf("#define PTR .quad\n"); | |
341 | printf("#define ALGN .align 8\n"); | |
342 | printf("#else\n"); | |
343 | printf("#define PTR .long\n"); | |
344 | printf("#define ALGN .align 4\n"); | |
345 | printf("#endif\n"); | |
346 | ||
aad09470 | 347 | printf("\t.section .rodata, \"a\"\n"); |
1da177e4 | 348 | |
fd593d12 EB |
349 | /* Provide proper symbols relocatability by their '_text' |
350 | * relativeness. The symbol names cannot be used to construct | |
351 | * normal symbol references as the list of symbols contains | |
352 | * symbols that are declared static and are private to their | |
353 | * .o files. This prevents .tmp_kallsyms.o or any other | |
354 | * object from referencing them. | |
355 | */ | |
1da177e4 | 356 | output_label("kallsyms_addresses"); |
b3dbb4ec | 357 | for (i = 0; i < table_cnt; i++) { |
78eb7159 | 358 | if (!symbol_absolute(&table[i])) { |
2c22d8ba VG |
359 | if (_text <= table[i].addr) |
360 | printf("\tPTR\t_text + %#llx\n", | |
361 | table[i].addr - _text); | |
362 | else | |
2930ffc7 AM |
363 | printf("\tPTR\t_text - %#llx\n", |
364 | _text - table[i].addr); | |
fd593d12 EB |
365 | } else { |
366 | printf("\tPTR\t%#llx\n", table[i].addr); | |
367 | } | |
1da177e4 LT |
368 | } |
369 | printf("\n"); | |
370 | ||
371 | output_label("kallsyms_num_syms"); | |
b3dbb4ec | 372 | printf("\tPTR\t%d\n", table_cnt); |
1da177e4 LT |
373 | printf("\n"); |
374 | ||
375 | /* table of offset markers, that give the offset in the compressed stream | |
376 | * every 256 symbols */ | |
f1a136e0 JJ |
377 | markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256)); |
378 | if (!markers) { | |
379 | fprintf(stderr, "kallsyms failure: " | |
380 | "unable to allocate required memory\n"); | |
381 | exit(EXIT_FAILURE); | |
382 | } | |
1da177e4 LT |
383 | |
384 | output_label("kallsyms_names"); | |
1da177e4 | 385 | off = 0; |
b3dbb4ec PM |
386 | for (i = 0; i < table_cnt; i++) { |
387 | if ((i & 0xFF) == 0) | |
388 | markers[i >> 8] = off; | |
1da177e4 LT |
389 | |
390 | printf("\t.byte 0x%02x", table[i].len); | |
391 | for (k = 0; k < table[i].len; k++) | |
392 | printf(", 0x%02x", table[i].sym[k]); | |
393 | printf("\n"); | |
394 | ||
395 | off += table[i].len + 1; | |
1da177e4 LT |
396 | } |
397 | printf("\n"); | |
398 | ||
399 | output_label("kallsyms_markers"); | |
b3dbb4ec | 400 | for (i = 0; i < ((table_cnt + 255) >> 8); i++) |
1da177e4 LT |
401 | printf("\tPTR\t%d\n", markers[i]); |
402 | printf("\n"); | |
403 | ||
404 | free(markers); | |
405 | ||
406 | output_label("kallsyms_token_table"); | |
407 | off = 0; | |
408 | for (i = 0; i < 256; i++) { | |
409 | best_idx[i] = off; | |
b3dbb4ec | 410 | expand_symbol(best_table[i], best_table_len[i], buf); |
1da177e4 LT |
411 | printf("\t.asciz\t\"%s\"\n", buf); |
412 | off += strlen(buf) + 1; | |
413 | } | |
414 | printf("\n"); | |
415 | ||
416 | output_label("kallsyms_token_index"); | |
417 | for (i = 0; i < 256; i++) | |
418 | printf("\t.short\t%d\n", best_idx[i]); | |
419 | printf("\n"); | |
420 | } | |
421 | ||
422 | ||
423 | /* table lookup compression functions */ | |
424 | ||
1da177e4 LT |
425 | /* count all the possible tokens in a symbol */ |
426 | static void learn_symbol(unsigned char *symbol, int len) | |
427 | { | |
428 | int i; | |
429 | ||
430 | for (i = 0; i < len - 1; i++) | |
b3dbb4ec | 431 | token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++; |
1da177e4 LT |
432 | } |
433 | ||
434 | /* decrease the count for all the possible tokens in a symbol */ | |
435 | static void forget_symbol(unsigned char *symbol, int len) | |
436 | { | |
437 | int i; | |
438 | ||
439 | for (i = 0; i < len - 1; i++) | |
b3dbb4ec | 440 | token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--; |
1da177e4 LT |
441 | } |
442 | ||
b3dbb4ec | 443 | /* remove all the invalid symbols from the table and do the initial token count */ |
1da177e4 LT |
444 | static void build_initial_tok_table(void) |
445 | { | |
b3dbb4ec | 446 | unsigned int i, pos; |
1da177e4 | 447 | |
b3dbb4ec PM |
448 | pos = 0; |
449 | for (i = 0; i < table_cnt; i++) { | |
1da177e4 | 450 | if ( symbol_valid(&table[i]) ) { |
b3dbb4ec PM |
451 | if (pos != i) |
452 | table[pos] = table[i]; | |
453 | learn_symbol(table[pos].sym, table[pos].len); | |
454 | pos++; | |
1da177e4 | 455 | } |
1da177e4 | 456 | } |
b3dbb4ec | 457 | table_cnt = pos; |
1da177e4 LT |
458 | } |
459 | ||
7c5d249a PM |
460 | static void *find_token(unsigned char *str, int len, unsigned char *token) |
461 | { | |
462 | int i; | |
463 | ||
464 | for (i = 0; i < len - 1; i++) { | |
465 | if (str[i] == token[0] && str[i+1] == token[1]) | |
466 | return &str[i]; | |
467 | } | |
468 | return NULL; | |
469 | } | |
470 | ||
1da177e4 LT |
471 | /* replace a given token in all the valid symbols. Use the sampled symbols |
472 | * to update the counts */ | |
b3dbb4ec | 473 | static void compress_symbols(unsigned char *str, int idx) |
1da177e4 | 474 | { |
b3dbb4ec PM |
475 | unsigned int i, len, size; |
476 | unsigned char *p1, *p2; | |
1da177e4 | 477 | |
b3dbb4ec | 478 | for (i = 0; i < table_cnt; i++) { |
1da177e4 LT |
479 | |
480 | len = table[i].len; | |
b3dbb4ec PM |
481 | p1 = table[i].sym; |
482 | ||
483 | /* find the token on the symbol */ | |
7c5d249a | 484 | p2 = find_token(p1, len, str); |
b3dbb4ec PM |
485 | if (!p2) continue; |
486 | ||
487 | /* decrease the counts for this symbol's tokens */ | |
488 | forget_symbol(table[i].sym, len); | |
489 | ||
490 | size = len; | |
1da177e4 LT |
491 | |
492 | do { | |
b3dbb4ec PM |
493 | *p2 = idx; |
494 | p2++; | |
495 | size -= (p2 - p1); | |
496 | memmove(p2, p2 + 1, size); | |
497 | p1 = p2; | |
498 | len--; | |
499 | ||
500 | if (size < 2) break; | |
501 | ||
1da177e4 | 502 | /* find the token on the symbol */ |
7c5d249a | 503 | p2 = find_token(p1, size, str); |
1da177e4 | 504 | |
b3dbb4ec | 505 | } while (p2); |
1da177e4 | 506 | |
b3dbb4ec | 507 | table[i].len = len; |
1da177e4 | 508 | |
b3dbb4ec PM |
509 | /* increase the counts for this symbol's new tokens */ |
510 | learn_symbol(table[i].sym, len); | |
1da177e4 LT |
511 | } |
512 | } | |
513 | ||
514 | /* search the token with the maximum profit */ | |
b3dbb4ec | 515 | static int find_best_token(void) |
1da177e4 | 516 | { |
b3dbb4ec | 517 | int i, best, bestprofit; |
1da177e4 LT |
518 | |
519 | bestprofit=-10000; | |
b3dbb4ec | 520 | best = 0; |
1da177e4 | 521 | |
b3dbb4ec PM |
522 | for (i = 0; i < 0x10000; i++) { |
523 | if (token_profit[i] > bestprofit) { | |
524 | best = i; | |
525 | bestprofit = token_profit[i]; | |
1da177e4 | 526 | } |
1da177e4 | 527 | } |
1da177e4 LT |
528 | return best; |
529 | } | |
530 | ||
531 | /* this is the core of the algorithm: calculate the "best" table */ | |
532 | static void optimize_result(void) | |
533 | { | |
b3dbb4ec | 534 | int i, best; |
1da177e4 LT |
535 | |
536 | /* using the '\0' symbol last allows compress_symbols to use standard | |
537 | * fast string functions */ | |
538 | for (i = 255; i >= 0; i--) { | |
539 | ||
540 | /* if this table slot is empty (it is not used by an actual | |
541 | * original char code */ | |
542 | if (!best_table_len[i]) { | |
543 | ||
544 | /* find the token with the breates profit value */ | |
545 | best = find_best_token(); | |
e0a04b11 XW |
546 | if (token_profit[best] == 0) |
547 | break; | |
1da177e4 LT |
548 | |
549 | /* place it in the "best" table */ | |
b3dbb4ec PM |
550 | best_table_len[i] = 2; |
551 | best_table[i][0] = best & 0xFF; | |
552 | best_table[i][1] = (best >> 8) & 0xFF; | |
1da177e4 LT |
553 | |
554 | /* replace this token in all the valid symbols */ | |
b3dbb4ec | 555 | compress_symbols(best_table[i], i); |
1da177e4 LT |
556 | } |
557 | } | |
558 | } | |
559 | ||
560 | /* start by placing the symbols that are actually used on the table */ | |
561 | static void insert_real_symbols_in_table(void) | |
562 | { | |
b3dbb4ec | 563 | unsigned int i, j, c; |
1da177e4 LT |
564 | |
565 | memset(best_table, 0, sizeof(best_table)); | |
566 | memset(best_table_len, 0, sizeof(best_table_len)); | |
567 | ||
b3dbb4ec PM |
568 | for (i = 0; i < table_cnt; i++) { |
569 | for (j = 0; j < table[i].len; j++) { | |
570 | c = table[i].sym[j]; | |
571 | best_table[c][0]=c; | |
572 | best_table_len[c]=1; | |
1da177e4 LT |
573 | } |
574 | } | |
575 | } | |
576 | ||
577 | static void optimize_token_table(void) | |
578 | { | |
1da177e4 LT |
579 | build_initial_tok_table(); |
580 | ||
581 | insert_real_symbols_in_table(); | |
582 | ||
41f11a4f | 583 | /* When valid symbol is not registered, exit to error */ |
b3dbb4ec | 584 | if (!table_cnt) { |
41f11a4f YS |
585 | fprintf(stderr, "No valid symbol.\n"); |
586 | exit(1); | |
587 | } | |
588 | ||
1da177e4 LT |
589 | optimize_result(); |
590 | } | |
591 | ||
b478b782 LJ |
592 | /* guess for "linker script provide" symbol */ |
593 | static int may_be_linker_script_provide_symbol(const struct sym_entry *se) | |
594 | { | |
595 | const char *symbol = (char *)se->sym + 1; | |
596 | int len = se->len - 1; | |
597 | ||
598 | if (len < 8) | |
599 | return 0; | |
600 | ||
601 | if (symbol[0] != '_' || symbol[1] != '_') | |
602 | return 0; | |
603 | ||
604 | /* __start_XXXXX */ | |
605 | if (!memcmp(symbol + 2, "start_", 6)) | |
606 | return 1; | |
607 | ||
608 | /* __stop_XXXXX */ | |
609 | if (!memcmp(symbol + 2, "stop_", 5)) | |
610 | return 1; | |
611 | ||
612 | /* __end_XXXXX */ | |
613 | if (!memcmp(symbol + 2, "end_", 4)) | |
614 | return 1; | |
615 | ||
616 | /* __XXXXX_start */ | |
617 | if (!memcmp(symbol + len - 6, "_start", 6)) | |
618 | return 1; | |
619 | ||
620 | /* __XXXXX_end */ | |
621 | if (!memcmp(symbol + len - 4, "_end", 4)) | |
622 | return 1; | |
623 | ||
624 | return 0; | |
625 | } | |
626 | ||
627 | static int prefix_underscores_count(const char *str) | |
628 | { | |
629 | const char *tail = str; | |
630 | ||
a9ece53c | 631 | while (*tail == '_') |
b478b782 LJ |
632 | tail++; |
633 | ||
634 | return tail - str; | |
635 | } | |
636 | ||
f2df3f65 PM |
637 | static int compare_symbols(const void *a, const void *b) |
638 | { | |
639 | const struct sym_entry *sa; | |
640 | const struct sym_entry *sb; | |
641 | int wa, wb; | |
642 | ||
643 | sa = a; | |
644 | sb = b; | |
645 | ||
646 | /* sort by address first */ | |
647 | if (sa->addr > sb->addr) | |
648 | return 1; | |
649 | if (sa->addr < sb->addr) | |
650 | return -1; | |
651 | ||
652 | /* sort by "weakness" type */ | |
653 | wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W'); | |
654 | wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W'); | |
655 | if (wa != wb) | |
656 | return wa - wb; | |
657 | ||
b478b782 LJ |
658 | /* sort by "linker script provide" type */ |
659 | wa = may_be_linker_script_provide_symbol(sa); | |
660 | wb = may_be_linker_script_provide_symbol(sb); | |
661 | if (wa != wb) | |
662 | return wa - wb; | |
663 | ||
664 | /* sort by the number of prefix underscores */ | |
665 | wa = prefix_underscores_count((const char *)sa->sym + 1); | |
666 | wb = prefix_underscores_count((const char *)sb->sym + 1); | |
667 | if (wa != wb) | |
668 | return wa - wb; | |
669 | ||
f2df3f65 PM |
670 | /* sort by initial order, so that other symbols are left undisturbed */ |
671 | return sa->start_pos - sb->start_pos; | |
672 | } | |
673 | ||
674 | static void sort_symbols(void) | |
675 | { | |
676 | qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols); | |
677 | } | |
1da177e4 | 678 | |
c6bda7c9 RR |
679 | static void make_percpus_absolute(void) |
680 | { | |
681 | unsigned int i; | |
682 | ||
683 | for (i = 0; i < table_cnt; i++) | |
684 | if (symbol_in_range(&table[i], &percpu_range, 1)) | |
685 | table[i].sym[0] = 'A'; | |
686 | } | |
687 | ||
b3dbb4ec | 688 | int main(int argc, char **argv) |
1da177e4 | 689 | { |
41f11a4f YS |
690 | if (argc >= 2) { |
691 | int i; | |
692 | for (i = 1; i < argc; i++) { | |
693 | if(strcmp(argv[i], "--all-symbols") == 0) | |
694 | all_symbols = 1; | |
c6bda7c9 RR |
695 | else if (strcmp(argv[i], "--absolute-percpu") == 0) |
696 | absolute_percpu = 1; | |
41f11a4f YS |
697 | else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) { |
698 | char *p = &argv[i][16]; | |
699 | /* skip quote */ | |
700 | if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\'')) | |
701 | p++; | |
702 | symbol_prefix_char = *p; | |
f6537f2f ML |
703 | } else if (strncmp(argv[i], "--page-offset=", 14) == 0) { |
704 | const char *p = &argv[i][14]; | |
705 | kernel_start_addr = strtoull(p, NULL, 16); | |
41f11a4f YS |
706 | } else |
707 | usage(); | |
708 | } | |
709 | } else if (argc != 1) | |
1da177e4 LT |
710 | usage(); |
711 | ||
712 | read_map(stdin); | |
c6bda7c9 RR |
713 | if (absolute_percpu) |
714 | make_percpus_absolute(); | |
2ea03891 SR |
715 | sort_symbols(); |
716 | optimize_token_table(); | |
1da177e4 LT |
717 | write_src(); |
718 | ||
719 | return 0; | |
720 | } |