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 | ||
a41333e0 | 21 | #include <stdbool.h> |
1da177e4 LT |
22 | #include <stdio.h> |
23 | #include <stdlib.h> | |
24 | #include <string.h> | |
25 | #include <ctype.h> | |
2213e9a6 | 26 | #include <limits.h> |
1da177e4 | 27 | |
17b1f0de | 28 | #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof(arr[0])) |
17b1f0de | 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; |
8c996940 | 36 | unsigned int percpu_absolute; |
9d82973e | 37 | unsigned char sym[]; |
1da177e4 LT |
38 | }; |
39 | ||
78eb7159 KC |
40 | struct addr_range { |
41 | const char *start_sym, *end_sym; | |
17b1f0de MF |
42 | unsigned long long start, end; |
43 | }; | |
44 | ||
45 | static unsigned long long _text; | |
2213e9a6 | 46 | static unsigned long long relative_base; |
78eb7159 | 47 | static struct addr_range text_ranges[] = { |
17b1f0de MF |
48 | { "_stext", "_etext" }, |
49 | { "_sinittext", "_einittext" }, | |
17b1f0de MF |
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 | ||
8d605269 | 58 | static struct sym_entry **table; |
b3dbb4ec | 59 | static unsigned int table_size, table_cnt; |
831362fc MY |
60 | static int all_symbols; |
61 | static int absolute_percpu; | |
62 | static int base_relative; | |
1da177e4 | 63 | |
f43e9daa | 64 | static int token_profit[0x10000]; |
1da177e4 LT |
65 | |
66 | /* the table that holds the result of the compression */ | |
f43e9daa MY |
67 | static unsigned char best_table[256][2]; |
68 | static unsigned char best_table_len[256]; | |
1da177e4 LT |
69 | |
70 | ||
b3dbb4ec | 71 | static void usage(void) |
1da177e4 | 72 | { |
f6537f2f | 73 | fprintf(stderr, "Usage: kallsyms [--all-symbols] " |
2213e9a6 | 74 | "[--base-relative] < in.map > out.S\n"); |
1da177e4 LT |
75 | exit(1); |
76 | } | |
77 | ||
29e55ad3 MY |
78 | static char *sym_name(const struct sym_entry *s) |
79 | { | |
80 | return (char *)s->sym + 1; | |
81 | } | |
82 | ||
a41333e0 MY |
83 | static bool is_ignored_symbol(const char *name, char type) |
84 | { | |
85 | static const char * const ignored_symbols[] = { | |
86 | /* | |
87 | * Symbols which vary between passes. Passes 1 and 2 must have | |
88 | * identical symbol lists. The kallsyms_* symbols below are | |
89 | * only added after pass 1, they would be included in pass 2 | |
90 | * when --all-symbols is specified so exclude them to get a | |
91 | * stable symbol list. | |
92 | */ | |
93 | "kallsyms_addresses", | |
94 | "kallsyms_offsets", | |
95 | "kallsyms_relative_base", | |
96 | "kallsyms_num_syms", | |
97 | "kallsyms_names", | |
98 | "kallsyms_markers", | |
99 | "kallsyms_token_table", | |
100 | "kallsyms_token_index", | |
101 | /* Exclude linker generated symbols which vary between passes */ | |
102 | "_SDA_BASE_", /* ppc */ | |
103 | "_SDA2_BASE_", /* ppc */ | |
104 | NULL | |
105 | }; | |
106 | ||
107 | static const char * const ignored_prefixes[] = { | |
97261e1e MY |
108 | "$", /* local symbols for ARM, MIPS, etc. */ |
109 | ".LASANPC", /* s390 kasan local symbols */ | |
a41333e0 MY |
110 | "__crc_", /* modversions */ |
111 | "__efistub_", /* arm64 EFI stub namespace */ | |
112 | NULL | |
113 | }; | |
114 | ||
115 | static const char * const ignored_suffixes[] = { | |
116 | "_from_arm", /* arm */ | |
117 | "_from_thumb", /* arm */ | |
118 | "_veneer", /* arm */ | |
119 | NULL | |
120 | }; | |
121 | ||
122 | const char * const *p; | |
123 | ||
124 | /* Exclude symbols which vary between passes. */ | |
125 | for (p = ignored_symbols; *p; p++) | |
126 | if (!strcmp(name, *p)) | |
127 | return true; | |
128 | ||
129 | for (p = ignored_prefixes; *p; p++) | |
130 | if (!strncmp(name, *p, strlen(*p))) | |
131 | return true; | |
132 | ||
133 | for (p = ignored_suffixes; *p; p++) { | |
134 | int l = strlen(name) - strlen(*p); | |
135 | ||
136 | if (l >= 0 && !strcmp(name + l, *p)) | |
137 | return true; | |
138 | } | |
139 | ||
887df76d MY |
140 | if (type == 'U' || type == 'u') |
141 | return true; | |
142 | /* exclude debugging symbols */ | |
143 | if (type == 'N' || type == 'n') | |
144 | return true; | |
145 | ||
146 | if (toupper(type) == 'A') { | |
147 | /* Keep these useful absolute symbols */ | |
148 | if (strcmp(name, "__kernel_syscall_via_break") && | |
149 | strcmp(name, "__kernel_syscall_via_epc") && | |
150 | strcmp(name, "__kernel_sigtramp") && | |
151 | strcmp(name, "__gp")) | |
152 | return true; | |
153 | } | |
154 | ||
a41333e0 MY |
155 | return false; |
156 | } | |
157 | ||
b6233d0d MY |
158 | static void check_symbol_range(const char *sym, unsigned long long addr, |
159 | struct addr_range *ranges, int entries) | |
17b1f0de MF |
160 | { |
161 | size_t i; | |
78eb7159 | 162 | struct addr_range *ar; |
17b1f0de | 163 | |
78eb7159 KC |
164 | for (i = 0; i < entries; ++i) { |
165 | ar = &ranges[i]; | |
17b1f0de | 166 | |
78eb7159 KC |
167 | if (strcmp(sym, ar->start_sym) == 0) { |
168 | ar->start = addr; | |
b6233d0d | 169 | return; |
78eb7159 KC |
170 | } else if (strcmp(sym, ar->end_sym) == 0) { |
171 | ar->end = addr; | |
b6233d0d | 172 | return; |
17b1f0de MF |
173 | } |
174 | } | |
17b1f0de MF |
175 | } |
176 | ||
8d605269 | 177 | static struct sym_entry *read_symbol(FILE *in) |
1da177e4 | 178 | { |
be9f6133 | 179 | char name[500], type; |
8d605269 MY |
180 | unsigned long long addr; |
181 | unsigned int len; | |
182 | struct sym_entry *sym; | |
1da177e4 LT |
183 | int rc; |
184 | ||
8d605269 | 185 | rc = fscanf(in, "%llx %c %499s\n", &addr, &type, name); |
1da177e4 | 186 | if (rc != 3) { |
be9f6133 | 187 | if (rc != EOF && fgets(name, 500, in) == NULL) |
ef894870 | 188 | fprintf(stderr, "Read error or end of file.\n"); |
8d605269 | 189 | return NULL; |
1da177e4 | 190 | } |
be9f6133 | 191 | if (strlen(name) >= KSYM_NAME_LEN) { |
6db2983c | 192 | fprintf(stderr, "Symbol %s too long for kallsyms (%zu >= %d).\n" |
bb66fc67 | 193 | "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n", |
be9f6133 | 194 | name, strlen(name), KSYM_NAME_LEN); |
8d605269 | 195 | return NULL; |
f3462aa9 | 196 | } |
1da177e4 | 197 | |
be9f6133 | 198 | if (strcmp(name, "_text") == 0) |
8d605269 | 199 | _text = addr; |
b6233d0d | 200 | |
7883a143 MP |
201 | /* Ignore most absolute/undefined (?) symbols. */ |
202 | if (is_ignored_symbol(name, type)) | |
203 | return NULL; | |
204 | ||
8d605269 MY |
205 | check_symbol_range(name, addr, text_ranges, ARRAY_SIZE(text_ranges)); |
206 | check_symbol_range(name, addr, &percpu_range, 1); | |
1da177e4 LT |
207 | |
208 | /* include the type field in the symbol name, so that it gets | |
209 | * compressed together */ | |
8d605269 MY |
210 | |
211 | len = strlen(name) + 1; | |
212 | ||
9d1b3895 | 213 | sym = malloc(sizeof(*sym) + len + 1); |
8d605269 | 214 | if (!sym) { |
f1a136e0 JJ |
215 | fprintf(stderr, "kallsyms failure: " |
216 | "unable to allocate required amount of memory\n"); | |
217 | exit(EXIT_FAILURE); | |
218 | } | |
8d605269 MY |
219 | sym->addr = addr; |
220 | sym->len = len; | |
221 | sym->sym[0] = type; | |
9d1b3895 | 222 | strcpy(sym_name(sym), name); |
8d605269 | 223 | sym->percpu_absolute = 0; |
8c996940 | 224 | |
8d605269 | 225 | return sym; |
1da177e4 LT |
226 | } |
227 | ||
4bfe2b78 MY |
228 | static int symbol_in_range(const struct sym_entry *s, |
229 | const struct addr_range *ranges, int entries) | |
17b1f0de MF |
230 | { |
231 | size_t i; | |
4bfe2b78 | 232 | const struct addr_range *ar; |
17b1f0de | 233 | |
78eb7159 KC |
234 | for (i = 0; i < entries; ++i) { |
235 | ar = &ranges[i]; | |
17b1f0de | 236 | |
78eb7159 | 237 | if (s->addr >= ar->start && s->addr <= ar->end) |
ac6ca5c8 | 238 | return 1; |
17b1f0de MF |
239 | } |
240 | ||
ac6ca5c8 | 241 | return 0; |
17b1f0de MF |
242 | } |
243 | ||
4bfe2b78 | 244 | static int symbol_valid(const struct sym_entry *s) |
1da177e4 | 245 | { |
29e55ad3 | 246 | const char *name = sym_name(s); |
bd8b22d2 | 247 | |
1da177e4 LT |
248 | /* if --all-symbols is not specified, then symbols outside the text |
249 | * and inittext sections are discarded */ | |
250 | if (!all_symbols) { | |
78eb7159 KC |
251 | if (symbol_in_range(s, text_ranges, |
252 | ARRAY_SIZE(text_ranges)) == 0) | |
1da177e4 LT |
253 | return 0; |
254 | /* Corner case. Discard any symbols with the same value as | |
a3b81113 RG |
255 | * _etext _einittext; they can move between pass 1 and 2 when |
256 | * the kallsyms data are added. If these symbols move then | |
257 | * they may get dropped in pass 2, which breaks the kallsyms | |
258 | * rules. | |
1da177e4 | 259 | */ |
17b1f0de | 260 | if ((s->addr == text_range_text->end && |
29e55ad3 | 261 | strcmp(name, text_range_text->end_sym)) || |
17b1f0de | 262 | (s->addr == text_range_inittext->end && |
29e55ad3 | 263 | strcmp(name, text_range_inittext->end_sym))) |
1da177e4 LT |
264 | return 0; |
265 | } | |
266 | ||
1da177e4 LT |
267 | return 1; |
268 | } | |
269 | ||
5e5c4fa7 MY |
270 | /* remove all the invalid symbols from the table */ |
271 | static void shrink_table(void) | |
272 | { | |
273 | unsigned int i, pos; | |
274 | ||
275 | pos = 0; | |
276 | for (i = 0; i < table_cnt; i++) { | |
8d605269 | 277 | if (symbol_valid(table[i])) { |
5e5c4fa7 MY |
278 | if (pos != i) |
279 | table[pos] = table[i]; | |
280 | pos++; | |
281 | } else { | |
8d605269 | 282 | free(table[i]); |
5e5c4fa7 MY |
283 | } |
284 | } | |
285 | table_cnt = pos; | |
286 | ||
287 | /* When valid symbol is not registered, exit to error */ | |
288 | if (!table_cnt) { | |
289 | fprintf(stderr, "No valid symbol.\n"); | |
290 | exit(1); | |
291 | } | |
292 | } | |
293 | ||
b3dbb4ec | 294 | static void read_map(FILE *in) |
1da177e4 | 295 | { |
8d605269 MY |
296 | struct sym_entry *sym; |
297 | ||
1da177e4 | 298 | while (!feof(in)) { |
8d605269 MY |
299 | sym = read_symbol(in); |
300 | if (!sym) | |
301 | continue; | |
302 | ||
303 | sym->start_pos = table_cnt; | |
304 | ||
b3dbb4ec PM |
305 | if (table_cnt >= table_size) { |
306 | table_size += 10000; | |
307 | table = realloc(table, sizeof(*table) * table_size); | |
1da177e4 LT |
308 | if (!table) { |
309 | fprintf(stderr, "out of memory\n"); | |
310 | exit (1); | |
311 | } | |
312 | } | |
8d605269 MY |
313 | |
314 | table[table_cnt++] = sym; | |
1da177e4 LT |
315 | } |
316 | } | |
317 | ||
4bfe2b78 | 318 | static void output_label(const char *label) |
1da177e4 | 319 | { |
534c9f2e | 320 | printf(".globl %s\n", label); |
1da177e4 | 321 | printf("\tALGN\n"); |
534c9f2e | 322 | printf("%s:\n", label); |
1da177e4 LT |
323 | } |
324 | ||
fd2ab2f6 MY |
325 | /* Provide proper symbols relocatability by their '_text' relativeness. */ |
326 | static void output_address(unsigned long long addr) | |
327 | { | |
328 | if (_text <= addr) | |
329 | printf("\tPTR\t_text + %#llx\n", addr - _text); | |
330 | else | |
331 | printf("\tPTR\t_text - %#llx\n", _text - addr); | |
332 | } | |
333 | ||
1da177e4 LT |
334 | /* uncompress a compressed symbol. When this function is called, the best table |
335 | * might still be compressed itself, so the function needs to be recursive */ | |
4bfe2b78 | 336 | static int expand_symbol(const unsigned char *data, int len, char *result) |
1da177e4 LT |
337 | { |
338 | int c, rlen, total=0; | |
339 | ||
340 | while (len) { | |
341 | c = *data; | |
342 | /* if the table holds a single char that is the same as the one | |
343 | * we are looking for, then end the search */ | |
344 | if (best_table[c][0]==c && best_table_len[c]==1) { | |
345 | *result++ = c; | |
346 | total++; | |
347 | } else { | |
348 | /* if not, recurse and expand */ | |
349 | rlen = expand_symbol(best_table[c], best_table_len[c], result); | |
350 | total += rlen; | |
351 | result += rlen; | |
352 | } | |
353 | data++; | |
354 | len--; | |
355 | } | |
356 | *result=0; | |
357 | ||
358 | return total; | |
359 | } | |
360 | ||
4bfe2b78 | 361 | static int symbol_absolute(const struct sym_entry *s) |
78eb7159 | 362 | { |
8c996940 | 363 | return s->percpu_absolute; |
78eb7159 KC |
364 | } |
365 | ||
b3dbb4ec | 366 | static void write_src(void) |
1da177e4 | 367 | { |
b3dbb4ec | 368 | unsigned int i, k, off; |
1da177e4 LT |
369 | unsigned int best_idx[256]; |
370 | unsigned int *markers; | |
9281acea | 371 | char buf[KSYM_NAME_LEN]; |
1da177e4 | 372 | |
500193ec | 373 | printf("#include <asm/bitsperlong.h>\n"); |
1da177e4 LT |
374 | printf("#if BITS_PER_LONG == 64\n"); |
375 | printf("#define PTR .quad\n"); | |
72d3ebb9 | 376 | printf("#define ALGN .balign 8\n"); |
1da177e4 LT |
377 | printf("#else\n"); |
378 | printf("#define PTR .long\n"); | |
72d3ebb9 | 379 | printf("#define ALGN .balign 4\n"); |
1da177e4 LT |
380 | printf("#endif\n"); |
381 | ||
aad09470 | 382 | printf("\t.section .rodata, \"a\"\n"); |
1da177e4 | 383 | |
2213e9a6 AB |
384 | if (!base_relative) |
385 | output_label("kallsyms_addresses"); | |
386 | else | |
387 | output_label("kallsyms_offsets"); | |
388 | ||
b3dbb4ec | 389 | for (i = 0; i < table_cnt; i++) { |
2213e9a6 | 390 | if (base_relative) { |
fd2ab2f6 MY |
391 | /* |
392 | * Use the offset relative to the lowest value | |
393 | * encountered of all relative symbols, and emit | |
394 | * non-relocatable fixed offsets that will be fixed | |
395 | * up at runtime. | |
396 | */ | |
397 | ||
2213e9a6 AB |
398 | long long offset; |
399 | int overflow; | |
400 | ||
401 | if (!absolute_percpu) { | |
8d605269 | 402 | offset = table[i]->addr - relative_base; |
2213e9a6 | 403 | overflow = (offset < 0 || offset > UINT_MAX); |
8d605269 MY |
404 | } else if (symbol_absolute(table[i])) { |
405 | offset = table[i]->addr; | |
2213e9a6 AB |
406 | overflow = (offset < 0 || offset > INT_MAX); |
407 | } else { | |
8d605269 | 408 | offset = relative_base - table[i]->addr - 1; |
2213e9a6 AB |
409 | overflow = (offset < INT_MIN || offset >= 0); |
410 | } | |
411 | if (overflow) { | |
412 | fprintf(stderr, "kallsyms failure: " | |
413 | "%s symbol value %#llx out of range in relative mode\n", | |
8d605269 MY |
414 | symbol_absolute(table[i]) ? "absolute" : "relative", |
415 | table[i]->addr); | |
2213e9a6 AB |
416 | exit(EXIT_FAILURE); |
417 | } | |
418 | printf("\t.long\t%#x\n", (int)offset); | |
8d605269 MY |
419 | } else if (!symbol_absolute(table[i])) { |
420 | output_address(table[i]->addr); | |
fd593d12 | 421 | } else { |
8d605269 | 422 | printf("\tPTR\t%#llx\n", table[i]->addr); |
fd593d12 | 423 | } |
1da177e4 LT |
424 | } |
425 | printf("\n"); | |
426 | ||
2213e9a6 AB |
427 | if (base_relative) { |
428 | output_label("kallsyms_relative_base"); | |
fd2ab2f6 | 429 | output_address(relative_base); |
2213e9a6 AB |
430 | printf("\n"); |
431 | } | |
432 | ||
1da177e4 | 433 | output_label("kallsyms_num_syms"); |
80ffbaa5 | 434 | printf("\t.long\t%u\n", table_cnt); |
1da177e4 LT |
435 | printf("\n"); |
436 | ||
437 | /* table of offset markers, that give the offset in the compressed stream | |
438 | * every 256 symbols */ | |
f1a136e0 JJ |
439 | markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256)); |
440 | if (!markers) { | |
441 | fprintf(stderr, "kallsyms failure: " | |
442 | "unable to allocate required memory\n"); | |
443 | exit(EXIT_FAILURE); | |
444 | } | |
1da177e4 LT |
445 | |
446 | output_label("kallsyms_names"); | |
1da177e4 | 447 | off = 0; |
b3dbb4ec PM |
448 | for (i = 0; i < table_cnt; i++) { |
449 | if ((i & 0xFF) == 0) | |
450 | markers[i >> 8] = off; | |
1da177e4 | 451 | |
8d605269 MY |
452 | printf("\t.byte 0x%02x", table[i]->len); |
453 | for (k = 0; k < table[i]->len; k++) | |
454 | printf(", 0x%02x", table[i]->sym[k]); | |
1da177e4 LT |
455 | printf("\n"); |
456 | ||
8d605269 | 457 | off += table[i]->len + 1; |
1da177e4 LT |
458 | } |
459 | printf("\n"); | |
460 | ||
461 | output_label("kallsyms_markers"); | |
b3dbb4ec | 462 | for (i = 0; i < ((table_cnt + 255) >> 8); i++) |
80ffbaa5 | 463 | printf("\t.long\t%u\n", markers[i]); |
1da177e4 LT |
464 | printf("\n"); |
465 | ||
466 | free(markers); | |
467 | ||
468 | output_label("kallsyms_token_table"); | |
469 | off = 0; | |
470 | for (i = 0; i < 256; i++) { | |
471 | best_idx[i] = off; | |
b3dbb4ec | 472 | expand_symbol(best_table[i], best_table_len[i], buf); |
1da177e4 LT |
473 | printf("\t.asciz\t\"%s\"\n", buf); |
474 | off += strlen(buf) + 1; | |
475 | } | |
476 | printf("\n"); | |
477 | ||
478 | output_label("kallsyms_token_index"); | |
479 | for (i = 0; i < 256; i++) | |
480 | printf("\t.short\t%d\n", best_idx[i]); | |
481 | printf("\n"); | |
482 | } | |
483 | ||
484 | ||
485 | /* table lookup compression functions */ | |
486 | ||
1da177e4 | 487 | /* count all the possible tokens in a symbol */ |
4bfe2b78 | 488 | static void learn_symbol(const unsigned char *symbol, int len) |
1da177e4 LT |
489 | { |
490 | int i; | |
491 | ||
492 | for (i = 0; i < len - 1; i++) | |
b3dbb4ec | 493 | token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++; |
1da177e4 LT |
494 | } |
495 | ||
496 | /* decrease the count for all the possible tokens in a symbol */ | |
4bfe2b78 | 497 | static void forget_symbol(const unsigned char *symbol, int len) |
1da177e4 LT |
498 | { |
499 | int i; | |
500 | ||
501 | for (i = 0; i < len - 1; i++) | |
b3dbb4ec | 502 | token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--; |
1da177e4 LT |
503 | } |
504 | ||
5e5c4fa7 | 505 | /* do the initial token count */ |
1da177e4 LT |
506 | static void build_initial_tok_table(void) |
507 | { | |
5e5c4fa7 | 508 | unsigned int i; |
1da177e4 | 509 | |
5e5c4fa7 | 510 | for (i = 0; i < table_cnt; i++) |
8d605269 | 511 | learn_symbol(table[i]->sym, table[i]->len); |
1da177e4 LT |
512 | } |
513 | ||
2558c138 | 514 | static unsigned char *find_token(unsigned char *str, int len, |
4bfe2b78 | 515 | const unsigned char *token) |
7c5d249a PM |
516 | { |
517 | int i; | |
518 | ||
519 | for (i = 0; i < len - 1; i++) { | |
520 | if (str[i] == token[0] && str[i+1] == token[1]) | |
521 | return &str[i]; | |
522 | } | |
523 | return NULL; | |
524 | } | |
525 | ||
1da177e4 LT |
526 | /* replace a given token in all the valid symbols. Use the sampled symbols |
527 | * to update the counts */ | |
4bfe2b78 | 528 | static void compress_symbols(const unsigned char *str, int idx) |
1da177e4 | 529 | { |
b3dbb4ec PM |
530 | unsigned int i, len, size; |
531 | unsigned char *p1, *p2; | |
1da177e4 | 532 | |
b3dbb4ec | 533 | for (i = 0; i < table_cnt; i++) { |
1da177e4 | 534 | |
8d605269 MY |
535 | len = table[i]->len; |
536 | p1 = table[i]->sym; | |
b3dbb4ec PM |
537 | |
538 | /* find the token on the symbol */ | |
7c5d249a | 539 | p2 = find_token(p1, len, str); |
b3dbb4ec PM |
540 | if (!p2) continue; |
541 | ||
542 | /* decrease the counts for this symbol's tokens */ | |
8d605269 | 543 | forget_symbol(table[i]->sym, len); |
b3dbb4ec PM |
544 | |
545 | size = len; | |
1da177e4 LT |
546 | |
547 | do { | |
b3dbb4ec PM |
548 | *p2 = idx; |
549 | p2++; | |
550 | size -= (p2 - p1); | |
551 | memmove(p2, p2 + 1, size); | |
552 | p1 = p2; | |
553 | len--; | |
554 | ||
555 | if (size < 2) break; | |
556 | ||
1da177e4 | 557 | /* find the token on the symbol */ |
7c5d249a | 558 | p2 = find_token(p1, size, str); |
1da177e4 | 559 | |
b3dbb4ec | 560 | } while (p2); |
1da177e4 | 561 | |
8d605269 | 562 | table[i]->len = len; |
1da177e4 | 563 | |
b3dbb4ec | 564 | /* increase the counts for this symbol's new tokens */ |
8d605269 | 565 | learn_symbol(table[i]->sym, len); |
1da177e4 LT |
566 | } |
567 | } | |
568 | ||
569 | /* search the token with the maximum profit */ | |
b3dbb4ec | 570 | static int find_best_token(void) |
1da177e4 | 571 | { |
b3dbb4ec | 572 | int i, best, bestprofit; |
1da177e4 LT |
573 | |
574 | bestprofit=-10000; | |
b3dbb4ec | 575 | best = 0; |
1da177e4 | 576 | |
b3dbb4ec PM |
577 | for (i = 0; i < 0x10000; i++) { |
578 | if (token_profit[i] > bestprofit) { | |
579 | best = i; | |
580 | bestprofit = token_profit[i]; | |
1da177e4 | 581 | } |
1da177e4 | 582 | } |
1da177e4 LT |
583 | return best; |
584 | } | |
585 | ||
586 | /* this is the core of the algorithm: calculate the "best" table */ | |
587 | static void optimize_result(void) | |
588 | { | |
b3dbb4ec | 589 | int i, best; |
1da177e4 LT |
590 | |
591 | /* using the '\0' symbol last allows compress_symbols to use standard | |
592 | * fast string functions */ | |
593 | for (i = 255; i >= 0; i--) { | |
594 | ||
595 | /* if this table slot is empty (it is not used by an actual | |
596 | * original char code */ | |
597 | if (!best_table_len[i]) { | |
598 | ||
cbf7a90e | 599 | /* find the token with the best profit value */ |
1da177e4 | 600 | best = find_best_token(); |
e0a04b11 XW |
601 | if (token_profit[best] == 0) |
602 | break; | |
1da177e4 LT |
603 | |
604 | /* place it in the "best" table */ | |
b3dbb4ec PM |
605 | best_table_len[i] = 2; |
606 | best_table[i][0] = best & 0xFF; | |
607 | best_table[i][1] = (best >> 8) & 0xFF; | |
1da177e4 LT |
608 | |
609 | /* replace this token in all the valid symbols */ | |
b3dbb4ec | 610 | compress_symbols(best_table[i], i); |
1da177e4 LT |
611 | } |
612 | } | |
613 | } | |
614 | ||
615 | /* start by placing the symbols that are actually used on the table */ | |
616 | static void insert_real_symbols_in_table(void) | |
617 | { | |
b3dbb4ec | 618 | unsigned int i, j, c; |
1da177e4 | 619 | |
b3dbb4ec | 620 | for (i = 0; i < table_cnt; i++) { |
8d605269 MY |
621 | for (j = 0; j < table[i]->len; j++) { |
622 | c = table[i]->sym[j]; | |
b3dbb4ec PM |
623 | best_table[c][0]=c; |
624 | best_table_len[c]=1; | |
1da177e4 LT |
625 | } |
626 | } | |
627 | } | |
628 | ||
629 | static void optimize_token_table(void) | |
630 | { | |
1da177e4 LT |
631 | build_initial_tok_table(); |
632 | ||
633 | insert_real_symbols_in_table(); | |
634 | ||
635 | optimize_result(); | |
636 | } | |
637 | ||
b478b782 LJ |
638 | /* guess for "linker script provide" symbol */ |
639 | static int may_be_linker_script_provide_symbol(const struct sym_entry *se) | |
640 | { | |
29e55ad3 | 641 | const char *symbol = sym_name(se); |
b478b782 LJ |
642 | int len = se->len - 1; |
643 | ||
644 | if (len < 8) | |
645 | return 0; | |
646 | ||
647 | if (symbol[0] != '_' || symbol[1] != '_') | |
648 | return 0; | |
649 | ||
650 | /* __start_XXXXX */ | |
651 | if (!memcmp(symbol + 2, "start_", 6)) | |
652 | return 1; | |
653 | ||
654 | /* __stop_XXXXX */ | |
655 | if (!memcmp(symbol + 2, "stop_", 5)) | |
656 | return 1; | |
657 | ||
658 | /* __end_XXXXX */ | |
659 | if (!memcmp(symbol + 2, "end_", 4)) | |
660 | return 1; | |
661 | ||
662 | /* __XXXXX_start */ | |
663 | if (!memcmp(symbol + len - 6, "_start", 6)) | |
664 | return 1; | |
665 | ||
666 | /* __XXXXX_end */ | |
667 | if (!memcmp(symbol + len - 4, "_end", 4)) | |
668 | return 1; | |
669 | ||
670 | return 0; | |
671 | } | |
672 | ||
f2df3f65 PM |
673 | static int compare_symbols(const void *a, const void *b) |
674 | { | |
8d605269 MY |
675 | const struct sym_entry *sa = *(const struct sym_entry **)a; |
676 | const struct sym_entry *sb = *(const struct sym_entry **)b; | |
f2df3f65 PM |
677 | int wa, wb; |
678 | ||
f2df3f65 PM |
679 | /* sort by address first */ |
680 | if (sa->addr > sb->addr) | |
681 | return 1; | |
682 | if (sa->addr < sb->addr) | |
683 | return -1; | |
684 | ||
685 | /* sort by "weakness" type */ | |
686 | wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W'); | |
687 | wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W'); | |
688 | if (wa != wb) | |
689 | return wa - wb; | |
690 | ||
b478b782 LJ |
691 | /* sort by "linker script provide" type */ |
692 | wa = may_be_linker_script_provide_symbol(sa); | |
693 | wb = may_be_linker_script_provide_symbol(sb); | |
694 | if (wa != wb) | |
695 | return wa - wb; | |
696 | ||
697 | /* sort by the number of prefix underscores */ | |
aa915245 MY |
698 | wa = strspn(sym_name(sa), "_"); |
699 | wb = strspn(sym_name(sb), "_"); | |
b478b782 LJ |
700 | if (wa != wb) |
701 | return wa - wb; | |
702 | ||
f2df3f65 PM |
703 | /* sort by initial order, so that other symbols are left undisturbed */ |
704 | return sa->start_pos - sb->start_pos; | |
705 | } | |
706 | ||
707 | static void sort_symbols(void) | |
708 | { | |
8d605269 | 709 | qsort(table, table_cnt, sizeof(table[0]), compare_symbols); |
f2df3f65 | 710 | } |
1da177e4 | 711 | |
c6bda7c9 RR |
712 | static void make_percpus_absolute(void) |
713 | { | |
714 | unsigned int i; | |
715 | ||
716 | for (i = 0; i < table_cnt; i++) | |
8d605269 | 717 | if (symbol_in_range(table[i], &percpu_range, 1)) { |
8c996940 AB |
718 | /* |
719 | * Keep the 'A' override for percpu symbols to | |
720 | * ensure consistent behavior compared to older | |
721 | * versions of this tool. | |
722 | */ | |
8d605269 MY |
723 | table[i]->sym[0] = 'A'; |
724 | table[i]->percpu_absolute = 1; | |
8c996940 | 725 | } |
c6bda7c9 RR |
726 | } |
727 | ||
2213e9a6 AB |
728 | /* find the minimum non-absolute symbol address */ |
729 | static void record_relative_base(void) | |
730 | { | |
731 | unsigned int i; | |
732 | ||
2213e9a6 | 733 | for (i = 0; i < table_cnt; i++) |
8d605269 | 734 | if (!symbol_absolute(table[i])) { |
f34ea029 MY |
735 | /* |
736 | * The table is sorted by address. | |
737 | * Take the first non-absolute symbol value. | |
738 | */ | |
8d605269 | 739 | relative_base = table[i]->addr; |
f34ea029 MY |
740 | return; |
741 | } | |
2213e9a6 AB |
742 | } |
743 | ||
b3dbb4ec | 744 | int main(int argc, char **argv) |
1da177e4 | 745 | { |
41f11a4f YS |
746 | if (argc >= 2) { |
747 | int i; | |
748 | for (i = 1; i < argc; i++) { | |
749 | if(strcmp(argv[i], "--all-symbols") == 0) | |
750 | all_symbols = 1; | |
c6bda7c9 RR |
751 | else if (strcmp(argv[i], "--absolute-percpu") == 0) |
752 | absolute_percpu = 1; | |
534c9f2e | 753 | else if (strcmp(argv[i], "--base-relative") == 0) |
2213e9a6 AB |
754 | base_relative = 1; |
755 | else | |
41f11a4f YS |
756 | usage(); |
757 | } | |
758 | } else if (argc != 1) | |
1da177e4 LT |
759 | usage(); |
760 | ||
761 | read_map(stdin); | |
5e5c4fa7 | 762 | shrink_table(); |
c6bda7c9 RR |
763 | if (absolute_percpu) |
764 | make_percpus_absolute(); | |
f34ea029 | 765 | sort_symbols(); |
2213e9a6 AB |
766 | if (base_relative) |
767 | record_relative_base(); | |
2ea03891 | 768 | optimize_token_table(); |
1da177e4 LT |
769 | write_src(); |
770 | ||
771 | return 0; | |
772 | } |