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 | ||
78eb7159 KC |
77 | static int check_symbol_range(const char *sym, unsigned long long addr, |
78 | struct addr_range *ranges, int entries) | |
17b1f0de MF |
79 | { |
80 | size_t i; | |
78eb7159 | 81 | struct addr_range *ar; |
17b1f0de | 82 | |
78eb7159 KC |
83 | for (i = 0; i < entries; ++i) { |
84 | ar = &ranges[i]; | |
17b1f0de | 85 | |
78eb7159 KC |
86 | if (strcmp(sym, ar->start_sym) == 0) { |
87 | ar->start = addr; | |
17b1f0de | 88 | return 0; |
78eb7159 KC |
89 | } else if (strcmp(sym, ar->end_sym) == 0) { |
90 | ar->end = addr; | |
17b1f0de MF |
91 | return 0; |
92 | } | |
93 | } | |
94 | ||
95 | return 1; | |
96 | } | |
97 | ||
b3dbb4ec | 98 | static int read_symbol(FILE *in, struct sym_entry *s) |
1da177e4 | 99 | { |
534c9f2e | 100 | char sym[500], stype; |
1da177e4 LT |
101 | int rc; |
102 | ||
534c9f2e | 103 | rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, sym); |
1da177e4 | 104 | if (rc != 3) { |
534c9f2e | 105 | if (rc != EOF && fgets(sym, 500, in) == NULL) |
ef894870 | 106 | fprintf(stderr, "Read error or end of file.\n"); |
1da177e4 LT |
107 | return -1; |
108 | } | |
6db2983c EL |
109 | if (strlen(sym) >= KSYM_NAME_LEN) { |
110 | fprintf(stderr, "Symbol %s too long for kallsyms (%zu >= %d).\n" | |
bb66fc67 | 111 | "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n", |
534c9f2e | 112 | sym, strlen(sym), KSYM_NAME_LEN); |
f3462aa9 AK |
113 | return -1; |
114 | } | |
1da177e4 LT |
115 | |
116 | /* Ignore most absolute/undefined (?) symbols. */ | |
fd593d12 EB |
117 | if (strcmp(sym, "_text") == 0) |
118 | _text = s->addr; | |
78eb7159 KC |
119 | else if (check_symbol_range(sym, s->addr, text_ranges, |
120 | ARRAY_SIZE(text_ranges)) == 0) | |
17b1f0de | 121 | /* nothing to do */; |
b3dbb4ec | 122 | else if (toupper(stype) == 'A') |
1da177e4 LT |
123 | { |
124 | /* Keep these useful absolute symbols */ | |
41f11a4f YS |
125 | if (strcmp(sym, "__kernel_syscall_via_break") && |
126 | strcmp(sym, "__kernel_syscall_via_epc") && | |
127 | strcmp(sym, "__kernel_sigtramp") && | |
128 | strcmp(sym, "__gp")) | |
1da177e4 LT |
129 | return -1; |
130 | ||
131 | } | |
e0109042 | 132 | else if (toupper(stype) == 'U') |
1da177e4 | 133 | return -1; |
e0109042 MY |
134 | /* |
135 | * Ignore generated symbols such as: | |
136 | * - mapping symbols in ARM ELF files ($a, $t, and $d) | |
137 | * - MIPS ELF local symbols ($L123 instead of .L123) | |
138 | */ | |
534c9f2e | 139 | else if (sym[0] == '$') |
6f00df24 | 140 | return -1; |
aab34ac8 | 141 | /* exclude debugging symbols */ |
51962a9d | 142 | else if (stype == 'N' || stype == 'n') |
aab34ac8 | 143 | return -1; |
33177f01 VG |
144 | /* exclude s390 kasan local symbols */ |
145 | else if (!strncmp(sym, ".LASANPC", 8)) | |
146 | return -1; | |
1da177e4 LT |
147 | |
148 | /* include the type field in the symbol name, so that it gets | |
149 | * compressed together */ | |
534c9f2e | 150 | s->len = strlen(sym) + 1; |
b3dbb4ec | 151 | s->sym = malloc(s->len + 1); |
f1a136e0 JJ |
152 | if (!s->sym) { |
153 | fprintf(stderr, "kallsyms failure: " | |
154 | "unable to allocate required amount of memory\n"); | |
155 | exit(EXIT_FAILURE); | |
156 | } | |
534c9f2e | 157 | strcpy((char *)s->sym + 1, sym); |
b3dbb4ec | 158 | s->sym[0] = stype; |
1da177e4 | 159 | |
8c996940 AB |
160 | s->percpu_absolute = 0; |
161 | ||
c6bda7c9 RR |
162 | /* Record if we've found __per_cpu_start/end. */ |
163 | check_symbol_range(sym, s->addr, &percpu_range, 1); | |
164 | ||
1da177e4 LT |
165 | return 0; |
166 | } | |
167 | ||
78eb7159 KC |
168 | static int symbol_in_range(struct sym_entry *s, struct addr_range *ranges, |
169 | int entries) | |
17b1f0de MF |
170 | { |
171 | size_t i; | |
78eb7159 | 172 | struct addr_range *ar; |
17b1f0de | 173 | |
78eb7159 KC |
174 | for (i = 0; i < entries; ++i) { |
175 | ar = &ranges[i]; | |
17b1f0de | 176 | |
78eb7159 | 177 | if (s->addr >= ar->start && s->addr <= ar->end) |
ac6ca5c8 | 178 | return 1; |
17b1f0de MF |
179 | } |
180 | ||
ac6ca5c8 | 181 | return 0; |
17b1f0de MF |
182 | } |
183 | ||
b3dbb4ec | 184 | static int symbol_valid(struct sym_entry *s) |
1da177e4 LT |
185 | { |
186 | /* Symbols which vary between passes. Passes 1 and 2 must have | |
2ea03891 SR |
187 | * identical symbol lists. The kallsyms_* symbols below are only added |
188 | * after pass 1, they would be included in pass 2 when --all-symbols is | |
189 | * specified so exclude them to get a stable symbol list. | |
1da177e4 LT |
190 | */ |
191 | static char *special_symbols[] = { | |
2ea03891 | 192 | "kallsyms_addresses", |
2213e9a6 AB |
193 | "kallsyms_offsets", |
194 | "kallsyms_relative_base", | |
2ea03891 SR |
195 | "kallsyms_num_syms", |
196 | "kallsyms_names", | |
197 | "kallsyms_markers", | |
198 | "kallsyms_token_table", | |
199 | "kallsyms_token_index", | |
200 | ||
1da177e4 LT |
201 | /* Exclude linker generated symbols which vary between passes */ |
202 | "_SDA_BASE_", /* ppc */ | |
203 | "_SDA2_BASE_", /* ppc */ | |
204 | NULL }; | |
bd8b22d2 | 205 | |
56067812 AB |
206 | static char *special_prefixes[] = { |
207 | "__crc_", /* modversions */ | |
1212f7a1 | 208 | "__efistub_", /* arm64 EFI stub namespace */ |
56067812 AB |
209 | NULL }; |
210 | ||
bd8b22d2 | 211 | static char *special_suffixes[] = { |
bd8b22d2 | 212 | "_veneer", /* arm */ |
b9b74be1 AB |
213 | "_from_arm", /* arm */ |
214 | "_from_thumb", /* arm */ | |
bd8b22d2 AB |
215 | NULL }; |
216 | ||
1da177e4 | 217 | int i; |
bd8b22d2 AB |
218 | char *sym_name = (char *)s->sym + 1; |
219 | ||
1da177e4 LT |
220 | /* if --all-symbols is not specified, then symbols outside the text |
221 | * and inittext sections are discarded */ | |
222 | if (!all_symbols) { | |
78eb7159 KC |
223 | if (symbol_in_range(s, text_ranges, |
224 | ARRAY_SIZE(text_ranges)) == 0) | |
1da177e4 LT |
225 | return 0; |
226 | /* Corner case. Discard any symbols with the same value as | |
a3b81113 RG |
227 | * _etext _einittext; they can move between pass 1 and 2 when |
228 | * the kallsyms data are added. If these symbols move then | |
229 | * they may get dropped in pass 2, which breaks the kallsyms | |
230 | * rules. | |
1da177e4 | 231 | */ |
17b1f0de | 232 | if ((s->addr == text_range_text->end && |
bd8b22d2 | 233 | strcmp(sym_name, |
78eb7159 | 234 | text_range_text->end_sym)) || |
17b1f0de | 235 | (s->addr == text_range_inittext->end && |
bd8b22d2 | 236 | strcmp(sym_name, |
78eb7159 | 237 | text_range_inittext->end_sym))) |
1da177e4 LT |
238 | return 0; |
239 | } | |
240 | ||
241 | /* Exclude symbols which vary between passes. */ | |
1da177e4 | 242 | for (i = 0; special_symbols[i]; i++) |
bd8b22d2 | 243 | if (strcmp(sym_name, special_symbols[i]) == 0) |
1da177e4 LT |
244 | return 0; |
245 | ||
56067812 AB |
246 | for (i = 0; special_prefixes[i]; i++) { |
247 | int l = strlen(special_prefixes[i]); | |
248 | ||
c5e5002f | 249 | if (strncmp(sym_name, special_prefixes[i], l) == 0) |
56067812 AB |
250 | return 0; |
251 | } | |
252 | ||
bd8b22d2 AB |
253 | for (i = 0; special_suffixes[i]; i++) { |
254 | int l = strlen(sym_name) - strlen(special_suffixes[i]); | |
255 | ||
256 | if (l >= 0 && strcmp(sym_name + l, special_suffixes[i]) == 0) | |
257 | return 0; | |
258 | } | |
259 | ||
1da177e4 LT |
260 | return 1; |
261 | } | |
262 | ||
5e5c4fa7 MY |
263 | /* remove all the invalid symbols from the table */ |
264 | static void shrink_table(void) | |
265 | { | |
266 | unsigned int i, pos; | |
267 | ||
268 | pos = 0; | |
269 | for (i = 0; i < table_cnt; i++) { | |
270 | if (symbol_valid(&table[i])) { | |
271 | if (pos != i) | |
272 | table[pos] = table[i]; | |
273 | pos++; | |
274 | } else { | |
275 | free(table[i].sym); | |
276 | } | |
277 | } | |
278 | table_cnt = pos; | |
279 | ||
280 | /* When valid symbol is not registered, exit to error */ | |
281 | if (!table_cnt) { | |
282 | fprintf(stderr, "No valid symbol.\n"); | |
283 | exit(1); | |
284 | } | |
285 | } | |
286 | ||
b3dbb4ec | 287 | static void read_map(FILE *in) |
1da177e4 LT |
288 | { |
289 | while (!feof(in)) { | |
b3dbb4ec PM |
290 | if (table_cnt >= table_size) { |
291 | table_size += 10000; | |
292 | table = realloc(table, sizeof(*table) * table_size); | |
1da177e4 LT |
293 | if (!table) { |
294 | fprintf(stderr, "out of memory\n"); | |
295 | exit (1); | |
296 | } | |
297 | } | |
f2df3f65 PM |
298 | if (read_symbol(in, &table[table_cnt]) == 0) { |
299 | table[table_cnt].start_pos = table_cnt; | |
b3dbb4ec | 300 | table_cnt++; |
f2df3f65 | 301 | } |
1da177e4 LT |
302 | } |
303 | } | |
304 | ||
305 | static void output_label(char *label) | |
306 | { | |
534c9f2e | 307 | printf(".globl %s\n", label); |
1da177e4 | 308 | printf("\tALGN\n"); |
534c9f2e | 309 | printf("%s:\n", label); |
1da177e4 LT |
310 | } |
311 | ||
312 | /* uncompress a compressed symbol. When this function is called, the best table | |
313 | * might still be compressed itself, so the function needs to be recursive */ | |
314 | static int expand_symbol(unsigned char *data, int len, char *result) | |
315 | { | |
316 | int c, rlen, total=0; | |
317 | ||
318 | while (len) { | |
319 | c = *data; | |
320 | /* if the table holds a single char that is the same as the one | |
321 | * we are looking for, then end the search */ | |
322 | if (best_table[c][0]==c && best_table_len[c]==1) { | |
323 | *result++ = c; | |
324 | total++; | |
325 | } else { | |
326 | /* if not, recurse and expand */ | |
327 | rlen = expand_symbol(best_table[c], best_table_len[c], result); | |
328 | total += rlen; | |
329 | result += rlen; | |
330 | } | |
331 | data++; | |
332 | len--; | |
333 | } | |
334 | *result=0; | |
335 | ||
336 | return total; | |
337 | } | |
338 | ||
78eb7159 KC |
339 | static int symbol_absolute(struct sym_entry *s) |
340 | { | |
8c996940 | 341 | return s->percpu_absolute; |
78eb7159 KC |
342 | } |
343 | ||
b3dbb4ec | 344 | static void write_src(void) |
1da177e4 | 345 | { |
b3dbb4ec | 346 | unsigned int i, k, off; |
1da177e4 LT |
347 | unsigned int best_idx[256]; |
348 | unsigned int *markers; | |
9281acea | 349 | char buf[KSYM_NAME_LEN]; |
1da177e4 | 350 | |
500193ec | 351 | printf("#include <asm/bitsperlong.h>\n"); |
1da177e4 LT |
352 | printf("#if BITS_PER_LONG == 64\n"); |
353 | printf("#define PTR .quad\n"); | |
72d3ebb9 | 354 | printf("#define ALGN .balign 8\n"); |
1da177e4 LT |
355 | printf("#else\n"); |
356 | printf("#define PTR .long\n"); | |
72d3ebb9 | 357 | printf("#define ALGN .balign 4\n"); |
1da177e4 LT |
358 | printf("#endif\n"); |
359 | ||
aad09470 | 360 | printf("\t.section .rodata, \"a\"\n"); |
1da177e4 | 361 | |
2213e9a6 AB |
362 | /* Provide proper symbols relocatability by their relativeness |
363 | * to a fixed anchor point in the runtime image, either '_text' | |
364 | * for absolute address tables, in which case the linker will | |
365 | * emit the final addresses at build time. Otherwise, use the | |
366 | * offset relative to the lowest value encountered of all relative | |
367 | * symbols, and emit non-relocatable fixed offsets that will be fixed | |
368 | * up at runtime. | |
369 | * | |
370 | * The symbol names cannot be used to construct normal symbol | |
371 | * references as the list of symbols contains symbols that are | |
372 | * declared static and are private to their .o files. This prevents | |
373 | * .tmp_kallsyms.o or any other object from referencing them. | |
fd593d12 | 374 | */ |
2213e9a6 AB |
375 | if (!base_relative) |
376 | output_label("kallsyms_addresses"); | |
377 | else | |
378 | output_label("kallsyms_offsets"); | |
379 | ||
b3dbb4ec | 380 | for (i = 0; i < table_cnt; i++) { |
2213e9a6 AB |
381 | if (base_relative) { |
382 | long long offset; | |
383 | int overflow; | |
384 | ||
385 | if (!absolute_percpu) { | |
386 | offset = table[i].addr - relative_base; | |
387 | overflow = (offset < 0 || offset > UINT_MAX); | |
388 | } else if (symbol_absolute(&table[i])) { | |
389 | offset = table[i].addr; | |
390 | overflow = (offset < 0 || offset > INT_MAX); | |
391 | } else { | |
392 | offset = relative_base - table[i].addr - 1; | |
393 | overflow = (offset < INT_MIN || offset >= 0); | |
394 | } | |
395 | if (overflow) { | |
396 | fprintf(stderr, "kallsyms failure: " | |
397 | "%s symbol value %#llx out of range in relative mode\n", | |
398 | symbol_absolute(&table[i]) ? "absolute" : "relative", | |
399 | table[i].addr); | |
400 | exit(EXIT_FAILURE); | |
401 | } | |
402 | printf("\t.long\t%#x\n", (int)offset); | |
403 | } else if (!symbol_absolute(&table[i])) { | |
2c22d8ba VG |
404 | if (_text <= table[i].addr) |
405 | printf("\tPTR\t_text + %#llx\n", | |
406 | table[i].addr - _text); | |
407 | else | |
2930ffc7 AM |
408 | printf("\tPTR\t_text - %#llx\n", |
409 | _text - table[i].addr); | |
fd593d12 EB |
410 | } else { |
411 | printf("\tPTR\t%#llx\n", table[i].addr); | |
412 | } | |
1da177e4 LT |
413 | } |
414 | printf("\n"); | |
415 | ||
2213e9a6 AB |
416 | if (base_relative) { |
417 | output_label("kallsyms_relative_base"); | |
418 | printf("\tPTR\t_text - %#llx\n", _text - relative_base); | |
419 | printf("\n"); | |
420 | } | |
421 | ||
1da177e4 | 422 | output_label("kallsyms_num_syms"); |
80ffbaa5 | 423 | printf("\t.long\t%u\n", table_cnt); |
1da177e4 LT |
424 | printf("\n"); |
425 | ||
426 | /* table of offset markers, that give the offset in the compressed stream | |
427 | * every 256 symbols */ | |
f1a136e0 JJ |
428 | markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256)); |
429 | if (!markers) { | |
430 | fprintf(stderr, "kallsyms failure: " | |
431 | "unable to allocate required memory\n"); | |
432 | exit(EXIT_FAILURE); | |
433 | } | |
1da177e4 LT |
434 | |
435 | output_label("kallsyms_names"); | |
1da177e4 | 436 | off = 0; |
b3dbb4ec PM |
437 | for (i = 0; i < table_cnt; i++) { |
438 | if ((i & 0xFF) == 0) | |
439 | markers[i >> 8] = off; | |
1da177e4 LT |
440 | |
441 | printf("\t.byte 0x%02x", table[i].len); | |
442 | for (k = 0; k < table[i].len; k++) | |
443 | printf(", 0x%02x", table[i].sym[k]); | |
444 | printf("\n"); | |
445 | ||
446 | off += table[i].len + 1; | |
1da177e4 LT |
447 | } |
448 | printf("\n"); | |
449 | ||
450 | output_label("kallsyms_markers"); | |
b3dbb4ec | 451 | for (i = 0; i < ((table_cnt + 255) >> 8); i++) |
80ffbaa5 | 452 | printf("\t.long\t%u\n", markers[i]); |
1da177e4 LT |
453 | printf("\n"); |
454 | ||
455 | free(markers); | |
456 | ||
457 | output_label("kallsyms_token_table"); | |
458 | off = 0; | |
459 | for (i = 0; i < 256; i++) { | |
460 | best_idx[i] = off; | |
b3dbb4ec | 461 | expand_symbol(best_table[i], best_table_len[i], buf); |
1da177e4 LT |
462 | printf("\t.asciz\t\"%s\"\n", buf); |
463 | off += strlen(buf) + 1; | |
464 | } | |
465 | printf("\n"); | |
466 | ||
467 | output_label("kallsyms_token_index"); | |
468 | for (i = 0; i < 256; i++) | |
469 | printf("\t.short\t%d\n", best_idx[i]); | |
470 | printf("\n"); | |
471 | } | |
472 | ||
473 | ||
474 | /* table lookup compression functions */ | |
475 | ||
1da177e4 LT |
476 | /* count all the possible tokens in a symbol */ |
477 | static void learn_symbol(unsigned char *symbol, int len) | |
478 | { | |
479 | int i; | |
480 | ||
481 | for (i = 0; i < len - 1; i++) | |
b3dbb4ec | 482 | token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++; |
1da177e4 LT |
483 | } |
484 | ||
485 | /* decrease the count for all the possible tokens in a symbol */ | |
486 | static void forget_symbol(unsigned char *symbol, int len) | |
487 | { | |
488 | int i; | |
489 | ||
490 | for (i = 0; i < len - 1; i++) | |
b3dbb4ec | 491 | token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--; |
1da177e4 LT |
492 | } |
493 | ||
5e5c4fa7 | 494 | /* do the initial token count */ |
1da177e4 LT |
495 | static void build_initial_tok_table(void) |
496 | { | |
5e5c4fa7 | 497 | unsigned int i; |
1da177e4 | 498 | |
5e5c4fa7 MY |
499 | for (i = 0; i < table_cnt; i++) |
500 | learn_symbol(table[i].sym, table[i].len); | |
1da177e4 LT |
501 | } |
502 | ||
7c5d249a PM |
503 | static void *find_token(unsigned char *str, int len, unsigned char *token) |
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 */ | |
b3dbb4ec | 516 | static void compress_symbols(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 | { | |
629 | const char *symbol = (char *)se->sym + 1; | |
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 | ||
661 | static int prefix_underscores_count(const char *str) | |
662 | { | |
663 | const char *tail = str; | |
664 | ||
a9ece53c | 665 | while (*tail == '_') |
b478b782 LJ |
666 | tail++; |
667 | ||
668 | return tail - str; | |
669 | } | |
670 | ||
f2df3f65 PM |
671 | static int compare_symbols(const void *a, const void *b) |
672 | { | |
673 | const struct sym_entry *sa; | |
674 | const struct sym_entry *sb; | |
675 | int wa, wb; | |
676 | ||
677 | sa = a; | |
678 | sb = b; | |
679 | ||
680 | /* sort by address first */ | |
681 | if (sa->addr > sb->addr) | |
682 | return 1; | |
683 | if (sa->addr < sb->addr) | |
684 | return -1; | |
685 | ||
686 | /* sort by "weakness" type */ | |
687 | wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W'); | |
688 | wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W'); | |
689 | if (wa != wb) | |
690 | return wa - wb; | |
691 | ||
b478b782 LJ |
692 | /* sort by "linker script provide" type */ |
693 | wa = may_be_linker_script_provide_symbol(sa); | |
694 | wb = may_be_linker_script_provide_symbol(sb); | |
695 | if (wa != wb) | |
696 | return wa - wb; | |
697 | ||
698 | /* sort by the number of prefix underscores */ | |
699 | wa = prefix_underscores_count((const char *)sa->sym + 1); | |
700 | wb = prefix_underscores_count((const char *)sb->sym + 1); | |
701 | if (wa != wb) | |
702 | return wa - wb; | |
703 | ||
f2df3f65 PM |
704 | /* sort by initial order, so that other symbols are left undisturbed */ |
705 | return sa->start_pos - sb->start_pos; | |
706 | } | |
707 | ||
708 | static void sort_symbols(void) | |
709 | { | |
710 | qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols); | |
711 | } | |
1da177e4 | 712 | |
c6bda7c9 RR |
713 | static void make_percpus_absolute(void) |
714 | { | |
715 | unsigned int i; | |
716 | ||
717 | for (i = 0; i < table_cnt; i++) | |
8c996940 AB |
718 | if (symbol_in_range(&table[i], &percpu_range, 1)) { |
719 | /* | |
720 | * Keep the 'A' override for percpu symbols to | |
721 | * ensure consistent behavior compared to older | |
722 | * versions of this tool. | |
723 | */ | |
c6bda7c9 | 724 | table[i].sym[0] = 'A'; |
8c996940 AB |
725 | table[i].percpu_absolute = 1; |
726 | } | |
c6bda7c9 RR |
727 | } |
728 | ||
2213e9a6 AB |
729 | /* find the minimum non-absolute symbol address */ |
730 | static void record_relative_base(void) | |
731 | { | |
732 | unsigned int i; | |
733 | ||
2213e9a6 | 734 | for (i = 0; i < table_cnt; i++) |
f34ea029 MY |
735 | if (!symbol_absolute(&table[i])) { |
736 | /* | |
737 | * The table is sorted by address. | |
738 | * Take the first non-absolute symbol value. | |
739 | */ | |
2213e9a6 | 740 | relative_base = table[i].addr; |
f34ea029 MY |
741 | return; |
742 | } | |
2213e9a6 AB |
743 | } |
744 | ||
b3dbb4ec | 745 | int main(int argc, char **argv) |
1da177e4 | 746 | { |
41f11a4f YS |
747 | if (argc >= 2) { |
748 | int i; | |
749 | for (i = 1; i < argc; i++) { | |
750 | if(strcmp(argv[i], "--all-symbols") == 0) | |
751 | all_symbols = 1; | |
c6bda7c9 RR |
752 | else if (strcmp(argv[i], "--absolute-percpu") == 0) |
753 | absolute_percpu = 1; | |
534c9f2e | 754 | else if (strcmp(argv[i], "--base-relative") == 0) |
2213e9a6 AB |
755 | base_relative = 1; |
756 | else | |
41f11a4f YS |
757 | usage(); |
758 | } | |
759 | } else if (argc != 1) | |
1da177e4 LT |
760 | usage(); |
761 | ||
762 | read_map(stdin); | |
5e5c4fa7 | 763 | shrink_table(); |
c6bda7c9 RR |
764 | if (absolute_percpu) |
765 | make_percpus_absolute(); | |
f34ea029 | 766 | sort_symbols(); |
2213e9a6 AB |
767 | if (base_relative) |
768 | record_relative_base(); | |
2ea03891 | 769 | optimize_token_table(); |
1da177e4 LT |
770 | write_src(); |
771 | ||
772 | return 0; | |
773 | } |