]>
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> | |
2213e9a6 | 25 | #include <limits.h> |
1da177e4 | 26 | |
17b1f0de | 27 | #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof(arr[0])) |
17b1f0de | 28 | |
9281acea | 29 | #define KSYM_NAME_LEN 128 |
1da177e4 | 30 | |
1da177e4 LT |
31 | struct sym_entry { |
32 | unsigned long long addr; | |
b3dbb4ec | 33 | unsigned int len; |
f2df3f65 | 34 | unsigned int start_pos; |
1da177e4 | 35 | unsigned char *sym; |
8c996940 | 36 | unsigned int percpu_absolute; |
1da177e4 LT |
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; | |
2213e9a6 | 45 | static unsigned long long relative_base; |
78eb7159 | 46 | static struct addr_range text_ranges[] = { |
17b1f0de MF |
47 | { "_stext", "_etext" }, |
48 | { "_sinittext", "_einittext" }, | |
17b1f0de MF |
49 | }; |
50 | #define text_range_text (&text_ranges[0]) | |
51 | #define text_range_inittext (&text_ranges[1]) | |
52 | ||
c6bda7c9 RR |
53 | static struct addr_range percpu_range = { |
54 | "__per_cpu_start", "__per_cpu_end", -1ULL, 0 | |
55 | }; | |
56 | ||
1da177e4 | 57 | static struct sym_entry *table; |
b3dbb4ec | 58 | static unsigned int table_size, table_cnt; |
1da177e4 | 59 | static int all_symbols = 0; |
c6bda7c9 | 60 | static int absolute_percpu = 0; |
2213e9a6 | 61 | static int base_relative = 0; |
1da177e4 | 62 | |
f43e9daa | 63 | static int token_profit[0x10000]; |
1da177e4 LT |
64 | |
65 | /* the table that holds the result of the compression */ | |
f43e9daa MY |
66 | static unsigned char best_table[256][2]; |
67 | static unsigned char best_table_len[256]; | |
1da177e4 LT |
68 | |
69 | ||
b3dbb4ec | 70 | static void usage(void) |
1da177e4 | 71 | { |
f6537f2f | 72 | fprintf(stderr, "Usage: kallsyms [--all-symbols] " |
2213e9a6 | 73 | "[--base-relative] < in.map > out.S\n"); |
1da177e4 LT |
74 | exit(1); |
75 | } | |
76 | ||
29e55ad3 MY |
77 | static char *sym_name(const struct sym_entry *s) |
78 | { | |
79 | return (char *)s->sym + 1; | |
80 | } | |
81 | ||
78eb7159 KC |
82 | static int check_symbol_range(const char *sym, unsigned long long addr, |
83 | struct addr_range *ranges, int entries) | |
17b1f0de MF |
84 | { |
85 | size_t i; | |
78eb7159 | 86 | struct addr_range *ar; |
17b1f0de | 87 | |
78eb7159 KC |
88 | for (i = 0; i < entries; ++i) { |
89 | ar = &ranges[i]; | |
17b1f0de | 90 | |
78eb7159 KC |
91 | if (strcmp(sym, ar->start_sym) == 0) { |
92 | ar->start = addr; | |
17b1f0de | 93 | return 0; |
78eb7159 KC |
94 | } else if (strcmp(sym, ar->end_sym) == 0) { |
95 | ar->end = addr; | |
17b1f0de MF |
96 | return 0; |
97 | } | |
98 | } | |
99 | ||
100 | return 1; | |
101 | } | |
102 | ||
b3dbb4ec | 103 | static int read_symbol(FILE *in, struct sym_entry *s) |
1da177e4 | 104 | { |
534c9f2e | 105 | char sym[500], stype; |
1da177e4 LT |
106 | int rc; |
107 | ||
534c9f2e | 108 | rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, sym); |
1da177e4 | 109 | if (rc != 3) { |
534c9f2e | 110 | if (rc != EOF && fgets(sym, 500, in) == NULL) |
ef894870 | 111 | fprintf(stderr, "Read error or end of file.\n"); |
1da177e4 LT |
112 | return -1; |
113 | } | |
6db2983c EL |
114 | if (strlen(sym) >= KSYM_NAME_LEN) { |
115 | fprintf(stderr, "Symbol %s too long for kallsyms (%zu >= %d).\n" | |
bb66fc67 | 116 | "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n", |
534c9f2e | 117 | sym, strlen(sym), KSYM_NAME_LEN); |
f3462aa9 AK |
118 | return -1; |
119 | } | |
1da177e4 LT |
120 | |
121 | /* Ignore most absolute/undefined (?) symbols. */ | |
fd593d12 EB |
122 | if (strcmp(sym, "_text") == 0) |
123 | _text = s->addr; | |
78eb7159 KC |
124 | else if (check_symbol_range(sym, s->addr, text_ranges, |
125 | ARRAY_SIZE(text_ranges)) == 0) | |
17b1f0de | 126 | /* nothing to do */; |
b3dbb4ec | 127 | else if (toupper(stype) == 'A') |
1da177e4 LT |
128 | { |
129 | /* Keep these useful absolute symbols */ | |
41f11a4f YS |
130 | if (strcmp(sym, "__kernel_syscall_via_break") && |
131 | strcmp(sym, "__kernel_syscall_via_epc") && | |
132 | strcmp(sym, "__kernel_sigtramp") && | |
133 | strcmp(sym, "__gp")) | |
1da177e4 LT |
134 | return -1; |
135 | ||
136 | } | |
e0109042 | 137 | else if (toupper(stype) == 'U') |
1da177e4 | 138 | return -1; |
e0109042 MY |
139 | /* |
140 | * Ignore generated symbols such as: | |
141 | * - mapping symbols in ARM ELF files ($a, $t, and $d) | |
142 | * - MIPS ELF local symbols ($L123 instead of .L123) | |
143 | */ | |
534c9f2e | 144 | else if (sym[0] == '$') |
6f00df24 | 145 | return -1; |
aab34ac8 | 146 | /* exclude debugging symbols */ |
51962a9d | 147 | else if (stype == 'N' || stype == 'n') |
aab34ac8 | 148 | return -1; |
33177f01 VG |
149 | /* exclude s390 kasan local symbols */ |
150 | else if (!strncmp(sym, ".LASANPC", 8)) | |
151 | return -1; | |
1da177e4 LT |
152 | |
153 | /* include the type field in the symbol name, so that it gets | |
154 | * compressed together */ | |
534c9f2e | 155 | s->len = strlen(sym) + 1; |
b3dbb4ec | 156 | s->sym = malloc(s->len + 1); |
f1a136e0 JJ |
157 | if (!s->sym) { |
158 | fprintf(stderr, "kallsyms failure: " | |
159 | "unable to allocate required amount of memory\n"); | |
160 | exit(EXIT_FAILURE); | |
161 | } | |
29e55ad3 | 162 | strcpy(sym_name(s), sym); |
b3dbb4ec | 163 | s->sym[0] = stype; |
1da177e4 | 164 | |
8c996940 AB |
165 | s->percpu_absolute = 0; |
166 | ||
c6bda7c9 RR |
167 | /* Record if we've found __per_cpu_start/end. */ |
168 | check_symbol_range(sym, s->addr, &percpu_range, 1); | |
169 | ||
1da177e4 LT |
170 | return 0; |
171 | } | |
172 | ||
78eb7159 KC |
173 | static int symbol_in_range(struct sym_entry *s, struct addr_range *ranges, |
174 | int entries) | |
17b1f0de MF |
175 | { |
176 | size_t i; | |
78eb7159 | 177 | struct addr_range *ar; |
17b1f0de | 178 | |
78eb7159 KC |
179 | for (i = 0; i < entries; ++i) { |
180 | ar = &ranges[i]; | |
17b1f0de | 181 | |
78eb7159 | 182 | if (s->addr >= ar->start && s->addr <= ar->end) |
ac6ca5c8 | 183 | return 1; |
17b1f0de MF |
184 | } |
185 | ||
ac6ca5c8 | 186 | return 0; |
17b1f0de MF |
187 | } |
188 | ||
b3dbb4ec | 189 | static int symbol_valid(struct sym_entry *s) |
1da177e4 LT |
190 | { |
191 | /* Symbols which vary between passes. Passes 1 and 2 must have | |
2ea03891 SR |
192 | * identical symbol lists. The kallsyms_* symbols below are only added |
193 | * after pass 1, they would be included in pass 2 when --all-symbols is | |
194 | * specified so exclude them to get a stable symbol list. | |
1da177e4 LT |
195 | */ |
196 | static char *special_symbols[] = { | |
2ea03891 | 197 | "kallsyms_addresses", |
2213e9a6 AB |
198 | "kallsyms_offsets", |
199 | "kallsyms_relative_base", | |
2ea03891 SR |
200 | "kallsyms_num_syms", |
201 | "kallsyms_names", | |
202 | "kallsyms_markers", | |
203 | "kallsyms_token_table", | |
204 | "kallsyms_token_index", | |
205 | ||
1da177e4 LT |
206 | /* Exclude linker generated symbols which vary between passes */ |
207 | "_SDA_BASE_", /* ppc */ | |
208 | "_SDA2_BASE_", /* ppc */ | |
209 | NULL }; | |
bd8b22d2 | 210 | |
56067812 AB |
211 | static char *special_prefixes[] = { |
212 | "__crc_", /* modversions */ | |
1212f7a1 | 213 | "__efistub_", /* arm64 EFI stub namespace */ |
56067812 AB |
214 | NULL }; |
215 | ||
bd8b22d2 | 216 | static char *special_suffixes[] = { |
bd8b22d2 | 217 | "_veneer", /* arm */ |
b9b74be1 AB |
218 | "_from_arm", /* arm */ |
219 | "_from_thumb", /* arm */ | |
bd8b22d2 AB |
220 | NULL }; |
221 | ||
1da177e4 | 222 | int i; |
29e55ad3 | 223 | const char *name = sym_name(s); |
bd8b22d2 | 224 | |
1da177e4 LT |
225 | /* if --all-symbols is not specified, then symbols outside the text |
226 | * and inittext sections are discarded */ | |
227 | if (!all_symbols) { | |
78eb7159 KC |
228 | if (symbol_in_range(s, text_ranges, |
229 | ARRAY_SIZE(text_ranges)) == 0) | |
1da177e4 LT |
230 | return 0; |
231 | /* Corner case. Discard any symbols with the same value as | |
a3b81113 RG |
232 | * _etext _einittext; they can move between pass 1 and 2 when |
233 | * the kallsyms data are added. If these symbols move then | |
234 | * they may get dropped in pass 2, which breaks the kallsyms | |
235 | * rules. | |
1da177e4 | 236 | */ |
17b1f0de | 237 | if ((s->addr == text_range_text->end && |
29e55ad3 | 238 | strcmp(name, text_range_text->end_sym)) || |
17b1f0de | 239 | (s->addr == text_range_inittext->end && |
29e55ad3 | 240 | strcmp(name, text_range_inittext->end_sym))) |
1da177e4 LT |
241 | return 0; |
242 | } | |
243 | ||
244 | /* Exclude symbols which vary between passes. */ | |
1da177e4 | 245 | for (i = 0; special_symbols[i]; i++) |
29e55ad3 | 246 | if (strcmp(name, special_symbols[i]) == 0) |
1da177e4 LT |
247 | return 0; |
248 | ||
56067812 AB |
249 | for (i = 0; special_prefixes[i]; i++) { |
250 | int l = strlen(special_prefixes[i]); | |
251 | ||
29e55ad3 | 252 | if (strncmp(name, special_prefixes[i], l) == 0) |
56067812 AB |
253 | return 0; |
254 | } | |
255 | ||
bd8b22d2 | 256 | for (i = 0; special_suffixes[i]; i++) { |
29e55ad3 | 257 | int l = strlen(name) - strlen(special_suffixes[i]); |
bd8b22d2 | 258 | |
29e55ad3 | 259 | if (l >= 0 && strcmp(name + l, special_suffixes[i]) == 0) |
bd8b22d2 AB |
260 | return 0; |
261 | } | |
262 | ||
1da177e4 LT |
263 | return 1; |
264 | } | |
265 | ||
5e5c4fa7 MY |
266 | /* remove all the invalid symbols from the table */ |
267 | static void shrink_table(void) | |
268 | { | |
269 | unsigned int i, pos; | |
270 | ||
271 | pos = 0; | |
272 | for (i = 0; i < table_cnt; i++) { | |
273 | if (symbol_valid(&table[i])) { | |
274 | if (pos != i) | |
275 | table[pos] = table[i]; | |
276 | pos++; | |
277 | } else { | |
278 | free(table[i].sym); | |
279 | } | |
280 | } | |
281 | table_cnt = pos; | |
282 | ||
283 | /* When valid symbol is not registered, exit to error */ | |
284 | if (!table_cnt) { | |
285 | fprintf(stderr, "No valid symbol.\n"); | |
286 | exit(1); | |
287 | } | |
288 | } | |
289 | ||
b3dbb4ec | 290 | static void read_map(FILE *in) |
1da177e4 LT |
291 | { |
292 | while (!feof(in)) { | |
b3dbb4ec PM |
293 | if (table_cnt >= table_size) { |
294 | table_size += 10000; | |
295 | table = realloc(table, sizeof(*table) * table_size); | |
1da177e4 LT |
296 | if (!table) { |
297 | fprintf(stderr, "out of memory\n"); | |
298 | exit (1); | |
299 | } | |
300 | } | |
f2df3f65 PM |
301 | if (read_symbol(in, &table[table_cnt]) == 0) { |
302 | table[table_cnt].start_pos = table_cnt; | |
b3dbb4ec | 303 | table_cnt++; |
f2df3f65 | 304 | } |
1da177e4 LT |
305 | } |
306 | } | |
307 | ||
308 | static void output_label(char *label) | |
309 | { | |
534c9f2e | 310 | printf(".globl %s\n", label); |
1da177e4 | 311 | printf("\tALGN\n"); |
534c9f2e | 312 | printf("%s:\n", label); |
1da177e4 LT |
313 | } |
314 | ||
315 | /* uncompress a compressed symbol. When this function is called, the best table | |
316 | * might still be compressed itself, so the function needs to be recursive */ | |
317 | static int expand_symbol(unsigned char *data, int len, char *result) | |
318 | { | |
319 | int c, rlen, total=0; | |
320 | ||
321 | while (len) { | |
322 | c = *data; | |
323 | /* if the table holds a single char that is the same as the one | |
324 | * we are looking for, then end the search */ | |
325 | if (best_table[c][0]==c && best_table_len[c]==1) { | |
326 | *result++ = c; | |
327 | total++; | |
328 | } else { | |
329 | /* if not, recurse and expand */ | |
330 | rlen = expand_symbol(best_table[c], best_table_len[c], result); | |
331 | total += rlen; | |
332 | result += rlen; | |
333 | } | |
334 | data++; | |
335 | len--; | |
336 | } | |
337 | *result=0; | |
338 | ||
339 | return total; | |
340 | } | |
341 | ||
78eb7159 KC |
342 | static int symbol_absolute(struct sym_entry *s) |
343 | { | |
8c996940 | 344 | return s->percpu_absolute; |
78eb7159 KC |
345 | } |
346 | ||
b3dbb4ec | 347 | static void write_src(void) |
1da177e4 | 348 | { |
b3dbb4ec | 349 | unsigned int i, k, off; |
1da177e4 LT |
350 | unsigned int best_idx[256]; |
351 | unsigned int *markers; | |
9281acea | 352 | char buf[KSYM_NAME_LEN]; |
1da177e4 | 353 | |
500193ec | 354 | printf("#include <asm/bitsperlong.h>\n"); |
1da177e4 LT |
355 | printf("#if BITS_PER_LONG == 64\n"); |
356 | printf("#define PTR .quad\n"); | |
72d3ebb9 | 357 | printf("#define ALGN .balign 8\n"); |
1da177e4 LT |
358 | printf("#else\n"); |
359 | printf("#define PTR .long\n"); | |
72d3ebb9 | 360 | printf("#define ALGN .balign 4\n"); |
1da177e4 LT |
361 | printf("#endif\n"); |
362 | ||
aad09470 | 363 | printf("\t.section .rodata, \"a\"\n"); |
1da177e4 | 364 | |
2213e9a6 AB |
365 | /* Provide proper symbols relocatability by their relativeness |
366 | * to a fixed anchor point in the runtime image, either '_text' | |
367 | * for absolute address tables, in which case the linker will | |
368 | * emit the final addresses at build time. Otherwise, use the | |
369 | * offset relative to the lowest value encountered of all relative | |
370 | * symbols, and emit non-relocatable fixed offsets that will be fixed | |
371 | * up at runtime. | |
372 | * | |
373 | * The symbol names cannot be used to construct normal symbol | |
374 | * references as the list of symbols contains symbols that are | |
375 | * declared static and are private to their .o files. This prevents | |
376 | * .tmp_kallsyms.o or any other object from referencing them. | |
fd593d12 | 377 | */ |
2213e9a6 AB |
378 | if (!base_relative) |
379 | output_label("kallsyms_addresses"); | |
380 | else | |
381 | output_label("kallsyms_offsets"); | |
382 | ||
b3dbb4ec | 383 | for (i = 0; i < table_cnt; i++) { |
2213e9a6 AB |
384 | if (base_relative) { |
385 | long long offset; | |
386 | int overflow; | |
387 | ||
388 | if (!absolute_percpu) { | |
389 | offset = table[i].addr - relative_base; | |
390 | overflow = (offset < 0 || offset > UINT_MAX); | |
391 | } else if (symbol_absolute(&table[i])) { | |
392 | offset = table[i].addr; | |
393 | overflow = (offset < 0 || offset > INT_MAX); | |
394 | } else { | |
395 | offset = relative_base - table[i].addr - 1; | |
396 | overflow = (offset < INT_MIN || offset >= 0); | |
397 | } | |
398 | if (overflow) { | |
399 | fprintf(stderr, "kallsyms failure: " | |
400 | "%s symbol value %#llx out of range in relative mode\n", | |
401 | symbol_absolute(&table[i]) ? "absolute" : "relative", | |
402 | table[i].addr); | |
403 | exit(EXIT_FAILURE); | |
404 | } | |
405 | printf("\t.long\t%#x\n", (int)offset); | |
406 | } else if (!symbol_absolute(&table[i])) { | |
2c22d8ba VG |
407 | if (_text <= table[i].addr) |
408 | printf("\tPTR\t_text + %#llx\n", | |
409 | table[i].addr - _text); | |
410 | else | |
2930ffc7 AM |
411 | printf("\tPTR\t_text - %#llx\n", |
412 | _text - table[i].addr); | |
fd593d12 EB |
413 | } else { |
414 | printf("\tPTR\t%#llx\n", table[i].addr); | |
415 | } | |
1da177e4 LT |
416 | } |
417 | printf("\n"); | |
418 | ||
2213e9a6 AB |
419 | if (base_relative) { |
420 | output_label("kallsyms_relative_base"); | |
421 | printf("\tPTR\t_text - %#llx\n", _text - relative_base); | |
422 | printf("\n"); | |
423 | } | |
424 | ||
1da177e4 | 425 | output_label("kallsyms_num_syms"); |
80ffbaa5 | 426 | printf("\t.long\t%u\n", table_cnt); |
1da177e4 LT |
427 | printf("\n"); |
428 | ||
429 | /* table of offset markers, that give the offset in the compressed stream | |
430 | * every 256 symbols */ | |
f1a136e0 JJ |
431 | markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256)); |
432 | if (!markers) { | |
433 | fprintf(stderr, "kallsyms failure: " | |
434 | "unable to allocate required memory\n"); | |
435 | exit(EXIT_FAILURE); | |
436 | } | |
1da177e4 LT |
437 | |
438 | output_label("kallsyms_names"); | |
1da177e4 | 439 | off = 0; |
b3dbb4ec PM |
440 | for (i = 0; i < table_cnt; i++) { |
441 | if ((i & 0xFF) == 0) | |
442 | markers[i >> 8] = off; | |
1da177e4 LT |
443 | |
444 | printf("\t.byte 0x%02x", table[i].len); | |
445 | for (k = 0; k < table[i].len; k++) | |
446 | printf(", 0x%02x", table[i].sym[k]); | |
447 | printf("\n"); | |
448 | ||
449 | off += table[i].len + 1; | |
1da177e4 LT |
450 | } |
451 | printf("\n"); | |
452 | ||
453 | output_label("kallsyms_markers"); | |
b3dbb4ec | 454 | for (i = 0; i < ((table_cnt + 255) >> 8); i++) |
80ffbaa5 | 455 | printf("\t.long\t%u\n", markers[i]); |
1da177e4 LT |
456 | printf("\n"); |
457 | ||
458 | free(markers); | |
459 | ||
460 | output_label("kallsyms_token_table"); | |
461 | off = 0; | |
462 | for (i = 0; i < 256; i++) { | |
463 | best_idx[i] = off; | |
b3dbb4ec | 464 | expand_symbol(best_table[i], best_table_len[i], buf); |
1da177e4 LT |
465 | printf("\t.asciz\t\"%s\"\n", buf); |
466 | off += strlen(buf) + 1; | |
467 | } | |
468 | printf("\n"); | |
469 | ||
470 | output_label("kallsyms_token_index"); | |
471 | for (i = 0; i < 256; i++) | |
472 | printf("\t.short\t%d\n", best_idx[i]); | |
473 | printf("\n"); | |
474 | } | |
475 | ||
476 | ||
477 | /* table lookup compression functions */ | |
478 | ||
1da177e4 LT |
479 | /* count all the possible tokens in a symbol */ |
480 | static void learn_symbol(unsigned char *symbol, int len) | |
481 | { | |
482 | int i; | |
483 | ||
484 | for (i = 0; i < len - 1; i++) | |
b3dbb4ec | 485 | token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++; |
1da177e4 LT |
486 | } |
487 | ||
488 | /* decrease the count for all the possible tokens in a symbol */ | |
489 | static void forget_symbol(unsigned char *symbol, int len) | |
490 | { | |
491 | int i; | |
492 | ||
493 | for (i = 0; i < len - 1; i++) | |
b3dbb4ec | 494 | token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--; |
1da177e4 LT |
495 | } |
496 | ||
5e5c4fa7 | 497 | /* do the initial token count */ |
1da177e4 LT |
498 | static void build_initial_tok_table(void) |
499 | { | |
5e5c4fa7 | 500 | unsigned int i; |
1da177e4 | 501 | |
5e5c4fa7 MY |
502 | for (i = 0; i < table_cnt; i++) |
503 | learn_symbol(table[i].sym, table[i].len); | |
1da177e4 LT |
504 | } |
505 | ||
7c5d249a PM |
506 | static void *find_token(unsigned char *str, int len, unsigned char *token) |
507 | { | |
508 | int i; | |
509 | ||
510 | for (i = 0; i < len - 1; i++) { | |
511 | if (str[i] == token[0] && str[i+1] == token[1]) | |
512 | return &str[i]; | |
513 | } | |
514 | return NULL; | |
515 | } | |
516 | ||
1da177e4 LT |
517 | /* replace a given token in all the valid symbols. Use the sampled symbols |
518 | * to update the counts */ | |
b3dbb4ec | 519 | static void compress_symbols(unsigned char *str, int idx) |
1da177e4 | 520 | { |
b3dbb4ec PM |
521 | unsigned int i, len, size; |
522 | unsigned char *p1, *p2; | |
1da177e4 | 523 | |
b3dbb4ec | 524 | for (i = 0; i < table_cnt; i++) { |
1da177e4 LT |
525 | |
526 | len = table[i].len; | |
b3dbb4ec PM |
527 | p1 = table[i].sym; |
528 | ||
529 | /* find the token on the symbol */ | |
7c5d249a | 530 | p2 = find_token(p1, len, str); |
b3dbb4ec PM |
531 | if (!p2) continue; |
532 | ||
533 | /* decrease the counts for this symbol's tokens */ | |
534 | forget_symbol(table[i].sym, len); | |
535 | ||
536 | size = len; | |
1da177e4 LT |
537 | |
538 | do { | |
b3dbb4ec PM |
539 | *p2 = idx; |
540 | p2++; | |
541 | size -= (p2 - p1); | |
542 | memmove(p2, p2 + 1, size); | |
543 | p1 = p2; | |
544 | len--; | |
545 | ||
546 | if (size < 2) break; | |
547 | ||
1da177e4 | 548 | /* find the token on the symbol */ |
7c5d249a | 549 | p2 = find_token(p1, size, str); |
1da177e4 | 550 | |
b3dbb4ec | 551 | } while (p2); |
1da177e4 | 552 | |
b3dbb4ec | 553 | table[i].len = len; |
1da177e4 | 554 | |
b3dbb4ec PM |
555 | /* increase the counts for this symbol's new tokens */ |
556 | learn_symbol(table[i].sym, len); | |
1da177e4 LT |
557 | } |
558 | } | |
559 | ||
560 | /* search the token with the maximum profit */ | |
b3dbb4ec | 561 | static int find_best_token(void) |
1da177e4 | 562 | { |
b3dbb4ec | 563 | int i, best, bestprofit; |
1da177e4 LT |
564 | |
565 | bestprofit=-10000; | |
b3dbb4ec | 566 | best = 0; |
1da177e4 | 567 | |
b3dbb4ec PM |
568 | for (i = 0; i < 0x10000; i++) { |
569 | if (token_profit[i] > bestprofit) { | |
570 | best = i; | |
571 | bestprofit = token_profit[i]; | |
1da177e4 | 572 | } |
1da177e4 | 573 | } |
1da177e4 LT |
574 | return best; |
575 | } | |
576 | ||
577 | /* this is the core of the algorithm: calculate the "best" table */ | |
578 | static void optimize_result(void) | |
579 | { | |
b3dbb4ec | 580 | int i, best; |
1da177e4 LT |
581 | |
582 | /* using the '\0' symbol last allows compress_symbols to use standard | |
583 | * fast string functions */ | |
584 | for (i = 255; i >= 0; i--) { | |
585 | ||
586 | /* if this table slot is empty (it is not used by an actual | |
587 | * original char code */ | |
588 | if (!best_table_len[i]) { | |
589 | ||
cbf7a90e | 590 | /* find the token with the best profit value */ |
1da177e4 | 591 | best = find_best_token(); |
e0a04b11 XW |
592 | if (token_profit[best] == 0) |
593 | break; | |
1da177e4 LT |
594 | |
595 | /* place it in the "best" table */ | |
b3dbb4ec PM |
596 | best_table_len[i] = 2; |
597 | best_table[i][0] = best & 0xFF; | |
598 | best_table[i][1] = (best >> 8) & 0xFF; | |
1da177e4 LT |
599 | |
600 | /* replace this token in all the valid symbols */ | |
b3dbb4ec | 601 | compress_symbols(best_table[i], i); |
1da177e4 LT |
602 | } |
603 | } | |
604 | } | |
605 | ||
606 | /* start by placing the symbols that are actually used on the table */ | |
607 | static void insert_real_symbols_in_table(void) | |
608 | { | |
b3dbb4ec | 609 | unsigned int i, j, c; |
1da177e4 | 610 | |
b3dbb4ec PM |
611 | for (i = 0; i < table_cnt; i++) { |
612 | for (j = 0; j < table[i].len; j++) { | |
613 | c = table[i].sym[j]; | |
614 | best_table[c][0]=c; | |
615 | best_table_len[c]=1; | |
1da177e4 LT |
616 | } |
617 | } | |
618 | } | |
619 | ||
620 | static void optimize_token_table(void) | |
621 | { | |
1da177e4 LT |
622 | build_initial_tok_table(); |
623 | ||
624 | insert_real_symbols_in_table(); | |
625 | ||
626 | optimize_result(); | |
627 | } | |
628 | ||
b478b782 LJ |
629 | /* guess for "linker script provide" symbol */ |
630 | static int may_be_linker_script_provide_symbol(const struct sym_entry *se) | |
631 | { | |
29e55ad3 | 632 | const char *symbol = sym_name(se); |
b478b782 LJ |
633 | int len = se->len - 1; |
634 | ||
635 | if (len < 8) | |
636 | return 0; | |
637 | ||
638 | if (symbol[0] != '_' || symbol[1] != '_') | |
639 | return 0; | |
640 | ||
641 | /* __start_XXXXX */ | |
642 | if (!memcmp(symbol + 2, "start_", 6)) | |
643 | return 1; | |
644 | ||
645 | /* __stop_XXXXX */ | |
646 | if (!memcmp(symbol + 2, "stop_", 5)) | |
647 | return 1; | |
648 | ||
649 | /* __end_XXXXX */ | |
650 | if (!memcmp(symbol + 2, "end_", 4)) | |
651 | return 1; | |
652 | ||
653 | /* __XXXXX_start */ | |
654 | if (!memcmp(symbol + len - 6, "_start", 6)) | |
655 | return 1; | |
656 | ||
657 | /* __XXXXX_end */ | |
658 | if (!memcmp(symbol + len - 4, "_end", 4)) | |
659 | return 1; | |
660 | ||
661 | return 0; | |
662 | } | |
663 | ||
f2df3f65 PM |
664 | static int compare_symbols(const void *a, const void *b) |
665 | { | |
666 | const struct sym_entry *sa; | |
667 | const struct sym_entry *sb; | |
668 | int wa, wb; | |
669 | ||
670 | sa = a; | |
671 | sb = b; | |
672 | ||
673 | /* sort by address first */ | |
674 | if (sa->addr > sb->addr) | |
675 | return 1; | |
676 | if (sa->addr < sb->addr) | |
677 | return -1; | |
678 | ||
679 | /* sort by "weakness" type */ | |
680 | wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W'); | |
681 | wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W'); | |
682 | if (wa != wb) | |
683 | return wa - wb; | |
684 | ||
b478b782 LJ |
685 | /* sort by "linker script provide" type */ |
686 | wa = may_be_linker_script_provide_symbol(sa); | |
687 | wb = may_be_linker_script_provide_symbol(sb); | |
688 | if (wa != wb) | |
689 | return wa - wb; | |
690 | ||
691 | /* sort by the number of prefix underscores */ | |
aa915245 MY |
692 | wa = strspn(sym_name(sa), "_"); |
693 | wb = strspn(sym_name(sb), "_"); | |
b478b782 LJ |
694 | if (wa != wb) |
695 | return wa - wb; | |
696 | ||
f2df3f65 PM |
697 | /* sort by initial order, so that other symbols are left undisturbed */ |
698 | return sa->start_pos - sb->start_pos; | |
699 | } | |
700 | ||
701 | static void sort_symbols(void) | |
702 | { | |
703 | qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols); | |
704 | } | |
1da177e4 | 705 | |
c6bda7c9 RR |
706 | static void make_percpus_absolute(void) |
707 | { | |
708 | unsigned int i; | |
709 | ||
710 | for (i = 0; i < table_cnt; i++) | |
8c996940 AB |
711 | if (symbol_in_range(&table[i], &percpu_range, 1)) { |
712 | /* | |
713 | * Keep the 'A' override for percpu symbols to | |
714 | * ensure consistent behavior compared to older | |
715 | * versions of this tool. | |
716 | */ | |
c6bda7c9 | 717 | table[i].sym[0] = 'A'; |
8c996940 AB |
718 | table[i].percpu_absolute = 1; |
719 | } | |
c6bda7c9 RR |
720 | } |
721 | ||
2213e9a6 AB |
722 | /* find the minimum non-absolute symbol address */ |
723 | static void record_relative_base(void) | |
724 | { | |
725 | unsigned int i; | |
726 | ||
2213e9a6 | 727 | for (i = 0; i < table_cnt; i++) |
f34ea029 MY |
728 | if (!symbol_absolute(&table[i])) { |
729 | /* | |
730 | * The table is sorted by address. | |
731 | * Take the first non-absolute symbol value. | |
732 | */ | |
2213e9a6 | 733 | relative_base = table[i].addr; |
f34ea029 MY |
734 | return; |
735 | } | |
2213e9a6 AB |
736 | } |
737 | ||
b3dbb4ec | 738 | int main(int argc, char **argv) |
1da177e4 | 739 | { |
41f11a4f YS |
740 | if (argc >= 2) { |
741 | int i; | |
742 | for (i = 1; i < argc; i++) { | |
743 | if(strcmp(argv[i], "--all-symbols") == 0) | |
744 | all_symbols = 1; | |
c6bda7c9 RR |
745 | else if (strcmp(argv[i], "--absolute-percpu") == 0) |
746 | absolute_percpu = 1; | |
534c9f2e | 747 | else if (strcmp(argv[i], "--base-relative") == 0) |
2213e9a6 AB |
748 | base_relative = 1; |
749 | else | |
41f11a4f YS |
750 | usage(); |
751 | } | |
752 | } else if (argc != 1) | |
1da177e4 LT |
753 | usage(); |
754 | ||
755 | read_map(stdin); | |
5e5c4fa7 | 756 | shrink_table(); |
c6bda7c9 RR |
757 | if (absolute_percpu) |
758 | make_percpus_absolute(); | |
f34ea029 | 759 | sort_symbols(); |
2213e9a6 AB |
760 | if (base_relative) |
761 | record_relative_base(); | |
2ea03891 | 762 | optimize_token_table(); |
1da177e4 LT |
763 | write_src(); |
764 | ||
765 | return 0; | |
766 | } |