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