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 | { |
534c9f2e | 179 | char sym[500], stype; |
1da177e4 LT |
180 | int rc; |
181 | ||
534c9f2e | 182 | rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, sym); |
1da177e4 | 183 | if (rc != 3) { |
534c9f2e | 184 | if (rc != EOF && fgets(sym, 500, in) == NULL) |
ef894870 | 185 | fprintf(stderr, "Read error or end of file.\n"); |
1da177e4 LT |
186 | return -1; |
187 | } | |
6db2983c EL |
188 | if (strlen(sym) >= KSYM_NAME_LEN) { |
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", |
534c9f2e | 191 | sym, strlen(sym), KSYM_NAME_LEN); |
f3462aa9 AK |
192 | return -1; |
193 | } | |
1da177e4 | 194 | |
a41333e0 MY |
195 | if (is_ignored_symbol(sym, stype)) |
196 | return -1; | |
197 | ||
1da177e4 | 198 | /* Ignore most absolute/undefined (?) symbols. */ |
fd593d12 EB |
199 | if (strcmp(sym, "_text") == 0) |
200 | _text = s->addr; | |
b6233d0d MY |
201 | |
202 | check_symbol_range(sym, s->addr, text_ranges, ARRAY_SIZE(text_ranges)); | |
d44270fc | 203 | check_symbol_range(sym, s->addr, &percpu_range, 1); |
1da177e4 LT |
204 | |
205 | /* include the type field in the symbol name, so that it gets | |
206 | * compressed together */ | |
534c9f2e | 207 | s->len = strlen(sym) + 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 | } | |
29e55ad3 | 214 | strcpy(sym_name(s), sym); |
b3dbb4ec | 215 | s->sym[0] = stype; |
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 | ||
313 | /* uncompress a compressed symbol. When this function is called, the best table | |
314 | * might still be compressed itself, so the function needs to be recursive */ | |
4bfe2b78 | 315 | static int expand_symbol(const unsigned char *data, int len, char *result) |
1da177e4 LT |
316 | { |
317 | int c, rlen, total=0; | |
318 | ||
319 | while (len) { | |
320 | c = *data; | |
321 | /* if the table holds a single char that is the same as the one | |
322 | * we are looking for, then end the search */ | |
323 | if (best_table[c][0]==c && best_table_len[c]==1) { | |
324 | *result++ = c; | |
325 | total++; | |
326 | } else { | |
327 | /* if not, recurse and expand */ | |
328 | rlen = expand_symbol(best_table[c], best_table_len[c], result); | |
329 | total += rlen; | |
330 | result += rlen; | |
331 | } | |
332 | data++; | |
333 | len--; | |
334 | } | |
335 | *result=0; | |
336 | ||
337 | return total; | |
338 | } | |
339 | ||
4bfe2b78 | 340 | static int symbol_absolute(const struct sym_entry *s) |
78eb7159 | 341 | { |
8c996940 | 342 | return s->percpu_absolute; |
78eb7159 KC |
343 | } |
344 | ||
b3dbb4ec | 345 | static void write_src(void) |
1da177e4 | 346 | { |
b3dbb4ec | 347 | unsigned int i, k, off; |
1da177e4 LT |
348 | unsigned int best_idx[256]; |
349 | unsigned int *markers; | |
9281acea | 350 | char buf[KSYM_NAME_LEN]; |
1da177e4 | 351 | |
500193ec | 352 | printf("#include <asm/bitsperlong.h>\n"); |
1da177e4 LT |
353 | printf("#if BITS_PER_LONG == 64\n"); |
354 | printf("#define PTR .quad\n"); | |
72d3ebb9 | 355 | printf("#define ALGN .balign 8\n"); |
1da177e4 LT |
356 | printf("#else\n"); |
357 | printf("#define PTR .long\n"); | |
72d3ebb9 | 358 | printf("#define ALGN .balign 4\n"); |
1da177e4 LT |
359 | printf("#endif\n"); |
360 | ||
aad09470 | 361 | printf("\t.section .rodata, \"a\"\n"); |
1da177e4 | 362 | |
2213e9a6 AB |
363 | /* Provide proper symbols relocatability by their relativeness |
364 | * to a fixed anchor point in the runtime image, either '_text' | |
365 | * for absolute address tables, in which case the linker will | |
366 | * emit the final addresses at build time. Otherwise, use the | |
367 | * offset relative to the lowest value encountered of all relative | |
368 | * symbols, and emit non-relocatable fixed offsets that will be fixed | |
369 | * up at runtime. | |
370 | * | |
371 | * The symbol names cannot be used to construct normal symbol | |
372 | * references as the list of symbols contains symbols that are | |
373 | * declared static and are private to their .o files. This prevents | |
374 | * .tmp_kallsyms.o or any other object from referencing them. | |
fd593d12 | 375 | */ |
2213e9a6 AB |
376 | if (!base_relative) |
377 | output_label("kallsyms_addresses"); | |
378 | else | |
379 | output_label("kallsyms_offsets"); | |
380 | ||
b3dbb4ec | 381 | for (i = 0; i < table_cnt; i++) { |
2213e9a6 AB |
382 | if (base_relative) { |
383 | long long offset; | |
384 | int overflow; | |
385 | ||
386 | if (!absolute_percpu) { | |
387 | offset = table[i].addr - relative_base; | |
388 | overflow = (offset < 0 || offset > UINT_MAX); | |
389 | } else if (symbol_absolute(&table[i])) { | |
390 | offset = table[i].addr; | |
391 | overflow = (offset < 0 || offset > INT_MAX); | |
392 | } else { | |
393 | offset = relative_base - table[i].addr - 1; | |
394 | overflow = (offset < INT_MIN || offset >= 0); | |
395 | } | |
396 | if (overflow) { | |
397 | fprintf(stderr, "kallsyms failure: " | |
398 | "%s symbol value %#llx out of range in relative mode\n", | |
399 | symbol_absolute(&table[i]) ? "absolute" : "relative", | |
400 | table[i].addr); | |
401 | exit(EXIT_FAILURE); | |
402 | } | |
403 | printf("\t.long\t%#x\n", (int)offset); | |
404 | } else if (!symbol_absolute(&table[i])) { | |
2c22d8ba VG |
405 | if (_text <= table[i].addr) |
406 | printf("\tPTR\t_text + %#llx\n", | |
407 | table[i].addr - _text); | |
408 | else | |
2930ffc7 AM |
409 | printf("\tPTR\t_text - %#llx\n", |
410 | _text - table[i].addr); | |
fd593d12 EB |
411 | } else { |
412 | printf("\tPTR\t%#llx\n", table[i].addr); | |
413 | } | |
1da177e4 LT |
414 | } |
415 | printf("\n"); | |
416 | ||
2213e9a6 AB |
417 | if (base_relative) { |
418 | output_label("kallsyms_relative_base"); | |
419 | printf("\tPTR\t_text - %#llx\n", _text - relative_base); | |
420 | printf("\n"); | |
421 | } | |
422 | ||
1da177e4 | 423 | output_label("kallsyms_num_syms"); |
80ffbaa5 | 424 | printf("\t.long\t%u\n", table_cnt); |
1da177e4 LT |
425 | printf("\n"); |
426 | ||
427 | /* table of offset markers, that give the offset in the compressed stream | |
428 | * every 256 symbols */ | |
f1a136e0 JJ |
429 | markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256)); |
430 | if (!markers) { | |
431 | fprintf(stderr, "kallsyms failure: " | |
432 | "unable to allocate required memory\n"); | |
433 | exit(EXIT_FAILURE); | |
434 | } | |
1da177e4 LT |
435 | |
436 | output_label("kallsyms_names"); | |
1da177e4 | 437 | off = 0; |
b3dbb4ec PM |
438 | for (i = 0; i < table_cnt; i++) { |
439 | if ((i & 0xFF) == 0) | |
440 | markers[i >> 8] = off; | |
1da177e4 LT |
441 | |
442 | printf("\t.byte 0x%02x", table[i].len); | |
443 | for (k = 0; k < table[i].len; k++) | |
444 | printf(", 0x%02x", table[i].sym[k]); | |
445 | printf("\n"); | |
446 | ||
447 | off += table[i].len + 1; | |
1da177e4 LT |
448 | } |
449 | printf("\n"); | |
450 | ||
451 | output_label("kallsyms_markers"); | |
b3dbb4ec | 452 | for (i = 0; i < ((table_cnt + 255) >> 8); i++) |
80ffbaa5 | 453 | printf("\t.long\t%u\n", markers[i]); |
1da177e4 LT |
454 | printf("\n"); |
455 | ||
456 | free(markers); | |
457 | ||
458 | output_label("kallsyms_token_table"); | |
459 | off = 0; | |
460 | for (i = 0; i < 256; i++) { | |
461 | best_idx[i] = off; | |
b3dbb4ec | 462 | expand_symbol(best_table[i], best_table_len[i], buf); |
1da177e4 LT |
463 | printf("\t.asciz\t\"%s\"\n", buf); |
464 | off += strlen(buf) + 1; | |
465 | } | |
466 | printf("\n"); | |
467 | ||
468 | output_label("kallsyms_token_index"); | |
469 | for (i = 0; i < 256; i++) | |
470 | printf("\t.short\t%d\n", best_idx[i]); | |
471 | printf("\n"); | |
472 | } | |
473 | ||
474 | ||
475 | /* table lookup compression functions */ | |
476 | ||
1da177e4 | 477 | /* count all the possible tokens in a symbol */ |
4bfe2b78 | 478 | static void learn_symbol(const unsigned char *symbol, int len) |
1da177e4 LT |
479 | { |
480 | int i; | |
481 | ||
482 | for (i = 0; i < len - 1; i++) | |
b3dbb4ec | 483 | token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++; |
1da177e4 LT |
484 | } |
485 | ||
486 | /* decrease the count for all the possible tokens in a symbol */ | |
4bfe2b78 | 487 | static void forget_symbol(const unsigned char *symbol, int len) |
1da177e4 LT |
488 | { |
489 | int i; | |
490 | ||
491 | for (i = 0; i < len - 1; i++) | |
b3dbb4ec | 492 | token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--; |
1da177e4 LT |
493 | } |
494 | ||
5e5c4fa7 | 495 | /* do the initial token count */ |
1da177e4 LT |
496 | static void build_initial_tok_table(void) |
497 | { | |
5e5c4fa7 | 498 | unsigned int i; |
1da177e4 | 499 | |
5e5c4fa7 MY |
500 | for (i = 0; i < table_cnt; i++) |
501 | learn_symbol(table[i].sym, table[i].len); | |
1da177e4 LT |
502 | } |
503 | ||
2558c138 | 504 | static unsigned char *find_token(unsigned char *str, int len, |
4bfe2b78 | 505 | const unsigned char *token) |
7c5d249a PM |
506 | { |
507 | int i; | |
508 | ||
509 | for (i = 0; i < len - 1; i++) { | |
510 | if (str[i] == token[0] && str[i+1] == token[1]) | |
511 | return &str[i]; | |
512 | } | |
513 | return NULL; | |
514 | } | |
515 | ||
1da177e4 LT |
516 | /* replace a given token in all the valid symbols. Use the sampled symbols |
517 | * to update the counts */ | |
4bfe2b78 | 518 | static void compress_symbols(const unsigned char *str, int idx) |
1da177e4 | 519 | { |
b3dbb4ec PM |
520 | unsigned int i, len, size; |
521 | unsigned char *p1, *p2; | |
1da177e4 | 522 | |
b3dbb4ec | 523 | for (i = 0; i < table_cnt; i++) { |
1da177e4 LT |
524 | |
525 | len = table[i].len; | |
b3dbb4ec PM |
526 | p1 = table[i].sym; |
527 | ||
528 | /* find the token on the symbol */ | |
7c5d249a | 529 | p2 = find_token(p1, len, str); |
b3dbb4ec PM |
530 | if (!p2) continue; |
531 | ||
532 | /* decrease the counts for this symbol's tokens */ | |
533 | forget_symbol(table[i].sym, len); | |
534 | ||
535 | size = len; | |
1da177e4 LT |
536 | |
537 | do { | |
b3dbb4ec PM |
538 | *p2 = idx; |
539 | p2++; | |
540 | size -= (p2 - p1); | |
541 | memmove(p2, p2 + 1, size); | |
542 | p1 = p2; | |
543 | len--; | |
544 | ||
545 | if (size < 2) break; | |
546 | ||
1da177e4 | 547 | /* find the token on the symbol */ |
7c5d249a | 548 | p2 = find_token(p1, size, str); |
1da177e4 | 549 | |
b3dbb4ec | 550 | } while (p2); |
1da177e4 | 551 | |
b3dbb4ec | 552 | table[i].len = len; |
1da177e4 | 553 | |
b3dbb4ec PM |
554 | /* increase the counts for this symbol's new tokens */ |
555 | learn_symbol(table[i].sym, len); | |
1da177e4 LT |
556 | } |
557 | } | |
558 | ||
559 | /* search the token with the maximum profit */ | |
b3dbb4ec | 560 | static int find_best_token(void) |
1da177e4 | 561 | { |
b3dbb4ec | 562 | int i, best, bestprofit; |
1da177e4 LT |
563 | |
564 | bestprofit=-10000; | |
b3dbb4ec | 565 | best = 0; |
1da177e4 | 566 | |
b3dbb4ec PM |
567 | for (i = 0; i < 0x10000; i++) { |
568 | if (token_profit[i] > bestprofit) { | |
569 | best = i; | |
570 | bestprofit = token_profit[i]; | |
1da177e4 | 571 | } |
1da177e4 | 572 | } |
1da177e4 LT |
573 | return best; |
574 | } | |
575 | ||
576 | /* this is the core of the algorithm: calculate the "best" table */ | |
577 | static void optimize_result(void) | |
578 | { | |
b3dbb4ec | 579 | int i, best; |
1da177e4 LT |
580 | |
581 | /* using the '\0' symbol last allows compress_symbols to use standard | |
582 | * fast string functions */ | |
583 | for (i = 255; i >= 0; i--) { | |
584 | ||
585 | /* if this table slot is empty (it is not used by an actual | |
586 | * original char code */ | |
587 | if (!best_table_len[i]) { | |
588 | ||
cbf7a90e | 589 | /* find the token with the best profit value */ |
1da177e4 | 590 | best = find_best_token(); |
e0a04b11 XW |
591 | if (token_profit[best] == 0) |
592 | break; | |
1da177e4 LT |
593 | |
594 | /* place it in the "best" table */ | |
b3dbb4ec PM |
595 | best_table_len[i] = 2; |
596 | best_table[i][0] = best & 0xFF; | |
597 | best_table[i][1] = (best >> 8) & 0xFF; | |
1da177e4 LT |
598 | |
599 | /* replace this token in all the valid symbols */ | |
b3dbb4ec | 600 | compress_symbols(best_table[i], i); |
1da177e4 LT |
601 | } |
602 | } | |
603 | } | |
604 | ||
605 | /* start by placing the symbols that are actually used on the table */ | |
606 | static void insert_real_symbols_in_table(void) | |
607 | { | |
b3dbb4ec | 608 | unsigned int i, j, c; |
1da177e4 | 609 | |
b3dbb4ec PM |
610 | for (i = 0; i < table_cnt; i++) { |
611 | for (j = 0; j < table[i].len; j++) { | |
612 | c = table[i].sym[j]; | |
613 | best_table[c][0]=c; | |
614 | best_table_len[c]=1; | |
1da177e4 LT |
615 | } |
616 | } | |
617 | } | |
618 | ||
619 | static void optimize_token_table(void) | |
620 | { | |
1da177e4 LT |
621 | build_initial_tok_table(); |
622 | ||
623 | insert_real_symbols_in_table(); | |
624 | ||
625 | optimize_result(); | |
626 | } | |
627 | ||
b478b782 LJ |
628 | /* guess for "linker script provide" symbol */ |
629 | static int may_be_linker_script_provide_symbol(const struct sym_entry *se) | |
630 | { | |
29e55ad3 | 631 | const char *symbol = sym_name(se); |
b478b782 LJ |
632 | int len = se->len - 1; |
633 | ||
634 | if (len < 8) | |
635 | return 0; | |
636 | ||
637 | if (symbol[0] != '_' || symbol[1] != '_') | |
638 | return 0; | |
639 | ||
640 | /* __start_XXXXX */ | |
641 | if (!memcmp(symbol + 2, "start_", 6)) | |
642 | return 1; | |
643 | ||
644 | /* __stop_XXXXX */ | |
645 | if (!memcmp(symbol + 2, "stop_", 5)) | |
646 | return 1; | |
647 | ||
648 | /* __end_XXXXX */ | |
649 | if (!memcmp(symbol + 2, "end_", 4)) | |
650 | return 1; | |
651 | ||
652 | /* __XXXXX_start */ | |
653 | if (!memcmp(symbol + len - 6, "_start", 6)) | |
654 | return 1; | |
655 | ||
656 | /* __XXXXX_end */ | |
657 | if (!memcmp(symbol + len - 4, "_end", 4)) | |
658 | return 1; | |
659 | ||
660 | return 0; | |
661 | } | |
662 | ||
f2df3f65 PM |
663 | static int compare_symbols(const void *a, const void *b) |
664 | { | |
665 | const struct sym_entry *sa; | |
666 | const struct sym_entry *sb; | |
667 | int wa, wb; | |
668 | ||
669 | sa = a; | |
670 | sb = b; | |
671 | ||
672 | /* sort by address first */ | |
673 | if (sa->addr > sb->addr) | |
674 | return 1; | |
675 | if (sa->addr < sb->addr) | |
676 | return -1; | |
677 | ||
678 | /* sort by "weakness" type */ | |
679 | wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W'); | |
680 | wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W'); | |
681 | if (wa != wb) | |
682 | return wa - wb; | |
683 | ||
b478b782 LJ |
684 | /* sort by "linker script provide" type */ |
685 | wa = may_be_linker_script_provide_symbol(sa); | |
686 | wb = may_be_linker_script_provide_symbol(sb); | |
687 | if (wa != wb) | |
688 | return wa - wb; | |
689 | ||
690 | /* sort by the number of prefix underscores */ | |
aa915245 MY |
691 | wa = strspn(sym_name(sa), "_"); |
692 | wb = strspn(sym_name(sb), "_"); | |
b478b782 LJ |
693 | if (wa != wb) |
694 | return wa - wb; | |
695 | ||
f2df3f65 PM |
696 | /* sort by initial order, so that other symbols are left undisturbed */ |
697 | return sa->start_pos - sb->start_pos; | |
698 | } | |
699 | ||
700 | static void sort_symbols(void) | |
701 | { | |
702 | qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols); | |
703 | } | |
1da177e4 | 704 | |
c6bda7c9 RR |
705 | static void make_percpus_absolute(void) |
706 | { | |
707 | unsigned int i; | |
708 | ||
709 | for (i = 0; i < table_cnt; i++) | |
8c996940 AB |
710 | if (symbol_in_range(&table[i], &percpu_range, 1)) { |
711 | /* | |
712 | * Keep the 'A' override for percpu symbols to | |
713 | * ensure consistent behavior compared to older | |
714 | * versions of this tool. | |
715 | */ | |
c6bda7c9 | 716 | table[i].sym[0] = 'A'; |
8c996940 AB |
717 | table[i].percpu_absolute = 1; |
718 | } | |
c6bda7c9 RR |
719 | } |
720 | ||
2213e9a6 AB |
721 | /* find the minimum non-absolute symbol address */ |
722 | static void record_relative_base(void) | |
723 | { | |
724 | unsigned int i; | |
725 | ||
2213e9a6 | 726 | for (i = 0; i < table_cnt; i++) |
f34ea029 MY |
727 | if (!symbol_absolute(&table[i])) { |
728 | /* | |
729 | * The table is sorted by address. | |
730 | * Take the first non-absolute symbol value. | |
731 | */ | |
2213e9a6 | 732 | relative_base = table[i].addr; |
f34ea029 MY |
733 | return; |
734 | } | |
2213e9a6 AB |
735 | } |
736 | ||
b3dbb4ec | 737 | int main(int argc, char **argv) |
1da177e4 | 738 | { |
41f11a4f YS |
739 | if (argc >= 2) { |
740 | int i; | |
741 | for (i = 1; i < argc; i++) { | |
742 | if(strcmp(argv[i], "--all-symbols") == 0) | |
743 | all_symbols = 1; | |
c6bda7c9 RR |
744 | else if (strcmp(argv[i], "--absolute-percpu") == 0) |
745 | absolute_percpu = 1; | |
534c9f2e | 746 | else if (strcmp(argv[i], "--base-relative") == 0) |
2213e9a6 AB |
747 | base_relative = 1; |
748 | else | |
41f11a4f YS |
749 | usage(); |
750 | } | |
751 | } else if (argc != 1) | |
1da177e4 LT |
752 | usage(); |
753 | ||
754 | read_map(stdin); | |
5e5c4fa7 | 755 | shrink_table(); |
c6bda7c9 RR |
756 | if (absolute_percpu) |
757 | make_percpus_absolute(); | |
f34ea029 | 758 | sort_symbols(); |
2213e9a6 AB |
759 | if (base_relative) |
760 | record_relative_base(); | |
2ea03891 | 761 | optimize_token_table(); |
1da177e4 LT |
762 | write_src(); |
763 | ||
764 | return 0; | |
765 | } |