]>
Commit | Line | Data |
---|---|---|
1da177e4 LT |
1 | /* Generate assembler source containing symbol information |
2 | * | |
3 | * Copyright 2002 by Kai Germaschewski | |
4 | * | |
5 | * This software may be used and distributed according to the terms | |
6 | * of the GNU General Public License, incorporated herein by reference. | |
7 | * | |
8 | * Usage: nm -n vmlinux | scripts/kallsyms [--all-symbols] > symbols.S | |
9 | * | |
1da177e4 LT |
10 | * Table compression uses all the unused char codes on the symbols and |
11 | * maps these to the most used substrings (tokens). For instance, it might | |
12 | * map char code 0xF7 to represent "write_" and then in every symbol where | |
13 | * "write_" appears it can be replaced by 0xF7, saving 5 bytes. | |
14 | * The used codes themselves are also placed in the table so that the | |
15 | * decompresion can work without "special cases". | |
16 | * Applied to kernel symbols, this usually produces a compression ratio | |
17 | * of about 50%. | |
18 | * | |
19 | */ | |
20 | ||
a41333e0 | 21 | #include <stdbool.h> |
1da177e4 LT |
22 | #include <stdio.h> |
23 | #include <stdlib.h> | |
24 | #include <string.h> | |
25 | #include <ctype.h> | |
2213e9a6 | 26 | #include <limits.h> |
1da177e4 | 27 | |
17b1f0de | 28 | #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof(arr[0])) |
17b1f0de | 29 | |
9281acea | 30 | #define KSYM_NAME_LEN 128 |
1da177e4 | 31 | |
1da177e4 LT |
32 | struct sym_entry { |
33 | unsigned long long addr; | |
b3dbb4ec | 34 | unsigned int len; |
f2df3f65 | 35 | unsigned int start_pos; |
8c996940 | 36 | unsigned int percpu_absolute; |
9d82973e | 37 | unsigned char sym[]; |
1da177e4 LT |
38 | }; |
39 | ||
78eb7159 KC |
40 | struct addr_range { |
41 | const char *start_sym, *end_sym; | |
17b1f0de MF |
42 | unsigned long long start, end; |
43 | }; | |
44 | ||
45 | static unsigned long long _text; | |
2213e9a6 | 46 | static unsigned long long relative_base; |
78eb7159 | 47 | static struct addr_range text_ranges[] = { |
17b1f0de MF |
48 | { "_stext", "_etext" }, |
49 | { "_sinittext", "_einittext" }, | |
17b1f0de MF |
50 | }; |
51 | #define text_range_text (&text_ranges[0]) | |
52 | #define text_range_inittext (&text_ranges[1]) | |
53 | ||
c6bda7c9 RR |
54 | static struct addr_range percpu_range = { |
55 | "__per_cpu_start", "__per_cpu_end", -1ULL, 0 | |
56 | }; | |
57 | ||
8d605269 | 58 | static struct sym_entry **table; |
b3dbb4ec | 59 | static unsigned int table_size, table_cnt; |
831362fc MY |
60 | static int all_symbols; |
61 | static int absolute_percpu; | |
62 | static int base_relative; | |
1da177e4 | 63 | |
f43e9daa | 64 | static int token_profit[0x10000]; |
1da177e4 LT |
65 | |
66 | /* the table that holds the result of the compression */ | |
f43e9daa MY |
67 | static unsigned char best_table[256][2]; |
68 | static unsigned char best_table_len[256]; | |
1da177e4 LT |
69 | |
70 | ||
b3dbb4ec | 71 | static void usage(void) |
1da177e4 | 72 | { |
8d3a7507 | 73 | fprintf(stderr, "Usage: kallsyms [--all-symbols] [--absolute-percpu] " |
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 | { | |
516d980f | 85 | /* Symbol names that exactly match to the following are ignored.*/ |
a41333e0 MY |
86 | static const char * const ignored_symbols[] = { |
87 | /* | |
88 | * Symbols which vary between passes. Passes 1 and 2 must have | |
89 | * identical symbol lists. The kallsyms_* symbols below are | |
90 | * only added after pass 1, they would be included in pass 2 | |
91 | * when --all-symbols is specified so exclude them to get a | |
92 | * stable symbol list. | |
93 | */ | |
94 | "kallsyms_addresses", | |
95 | "kallsyms_offsets", | |
96 | "kallsyms_relative_base", | |
97 | "kallsyms_num_syms", | |
98 | "kallsyms_names", | |
99 | "kallsyms_markers", | |
100 | "kallsyms_token_table", | |
101 | "kallsyms_token_index", | |
102 | /* Exclude linker generated symbols which vary between passes */ | |
103 | "_SDA_BASE_", /* ppc */ | |
104 | "_SDA2_BASE_", /* ppc */ | |
105 | NULL | |
106 | }; | |
107 | ||
516d980f | 108 | /* Symbol names that begin with the following are ignored.*/ |
a41333e0 | 109 | static const char * const ignored_prefixes[] = { |
a41333e0 | 110 | "__efistub_", /* arm64 EFI stub namespace */ |
6ccf9cb5 KS |
111 | "__kvm_nvhe_$", /* arm64 local symbols in non-VHE KVM namespace */ |
112 | "__kvm_nvhe_.L", /* arm64 local symbols in non-VHE KVM namespace */ | |
efe6e306 AB |
113 | "__AArch64ADRPThunk_", /* arm64 lld */ |
114 | "__ARMV5PILongThunk_", /* arm lld */ | |
115 | "__ARMV7PILongThunk_", | |
116 | "__ThumbV7PILongThunk_", | |
117 | "__LA25Thunk_", /* mips lld */ | |
118 | "__microLA25Thunk_", | |
a41333e0 MY |
119 | NULL |
120 | }; | |
121 | ||
516d980f | 122 | /* Symbol names that end with the following are ignored.*/ |
a41333e0 MY |
123 | static const char * const ignored_suffixes[] = { |
124 | "_from_arm", /* arm */ | |
125 | "_from_thumb", /* arm */ | |
126 | "_veneer", /* arm */ | |
127 | NULL | |
128 | }; | |
129 | ||
516d980f MY |
130 | /* Symbol names that contain the following are ignored.*/ |
131 | static const char * const ignored_matches[] = { | |
132 | ".long_branch.", /* ppc stub */ | |
133 | ".plt_branch.", /* ppc stub */ | |
134 | NULL | |
135 | }; | |
136 | ||
a41333e0 MY |
137 | const char * const *p; |
138 | ||
a41333e0 MY |
139 | for (p = ignored_symbols; *p; p++) |
140 | if (!strcmp(name, *p)) | |
141 | return true; | |
142 | ||
143 | for (p = ignored_prefixes; *p; p++) | |
144 | if (!strncmp(name, *p, strlen(*p))) | |
145 | return true; | |
146 | ||
147 | for (p = ignored_suffixes; *p; p++) { | |
148 | int l = strlen(name) - strlen(*p); | |
149 | ||
150 | if (l >= 0 && !strcmp(name + l, *p)) | |
151 | return true; | |
152 | } | |
153 | ||
516d980f MY |
154 | for (p = ignored_matches; *p; p++) { |
155 | if (strstr(name, *p)) | |
156 | return true; | |
157 | } | |
158 | ||
887df76d MY |
159 | if (type == 'U' || type == 'u') |
160 | return true; | |
161 | /* exclude debugging symbols */ | |
162 | if (type == 'N' || type == 'n') | |
163 | return true; | |
164 | ||
165 | if (toupper(type) == 'A') { | |
166 | /* Keep these useful absolute symbols */ | |
167 | if (strcmp(name, "__kernel_syscall_via_break") && | |
168 | strcmp(name, "__kernel_syscall_via_epc") && | |
169 | strcmp(name, "__kernel_sigtramp") && | |
170 | strcmp(name, "__gp")) | |
171 | return true; | |
172 | } | |
173 | ||
a41333e0 MY |
174 | return false; |
175 | } | |
176 | ||
b6233d0d MY |
177 | static void check_symbol_range(const char *sym, unsigned long long addr, |
178 | struct addr_range *ranges, int entries) | |
17b1f0de MF |
179 | { |
180 | size_t i; | |
78eb7159 | 181 | struct addr_range *ar; |
17b1f0de | 182 | |
78eb7159 KC |
183 | for (i = 0; i < entries; ++i) { |
184 | ar = &ranges[i]; | |
17b1f0de | 185 | |
78eb7159 KC |
186 | if (strcmp(sym, ar->start_sym) == 0) { |
187 | ar->start = addr; | |
b6233d0d | 188 | return; |
78eb7159 KC |
189 | } else if (strcmp(sym, ar->end_sym) == 0) { |
190 | ar->end = addr; | |
b6233d0d | 191 | return; |
17b1f0de MF |
192 | } |
193 | } | |
17b1f0de MF |
194 | } |
195 | ||
8d605269 | 196 | static struct sym_entry *read_symbol(FILE *in) |
1da177e4 | 197 | { |
be9f6133 | 198 | char name[500], type; |
8d605269 MY |
199 | unsigned long long addr; |
200 | unsigned int len; | |
201 | struct sym_entry *sym; | |
1da177e4 LT |
202 | int rc; |
203 | ||
8d605269 | 204 | rc = fscanf(in, "%llx %c %499s\n", &addr, &type, name); |
1da177e4 | 205 | if (rc != 3) { |
be9f6133 | 206 | if (rc != EOF && fgets(name, 500, in) == NULL) |
ef894870 | 207 | fprintf(stderr, "Read error or end of file.\n"); |
8d605269 | 208 | return NULL; |
1da177e4 | 209 | } |
be9f6133 | 210 | if (strlen(name) >= KSYM_NAME_LEN) { |
6db2983c | 211 | fprintf(stderr, "Symbol %s too long for kallsyms (%zu >= %d).\n" |
bb66fc67 | 212 | "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n", |
be9f6133 | 213 | name, strlen(name), KSYM_NAME_LEN); |
8d605269 | 214 | return NULL; |
f3462aa9 | 215 | } |
1da177e4 | 216 | |
be9f6133 | 217 | if (strcmp(name, "_text") == 0) |
8d605269 | 218 | _text = addr; |
b6233d0d | 219 | |
7883a143 MP |
220 | /* Ignore most absolute/undefined (?) symbols. */ |
221 | if (is_ignored_symbol(name, type)) | |
222 | return NULL; | |
223 | ||
8d605269 MY |
224 | check_symbol_range(name, addr, text_ranges, ARRAY_SIZE(text_ranges)); |
225 | check_symbol_range(name, addr, &percpu_range, 1); | |
1da177e4 LT |
226 | |
227 | /* include the type field in the symbol name, so that it gets | |
228 | * compressed together */ | |
8d605269 MY |
229 | |
230 | len = strlen(name) + 1; | |
231 | ||
9d1b3895 | 232 | sym = malloc(sizeof(*sym) + len + 1); |
8d605269 | 233 | if (!sym) { |
f1a136e0 JJ |
234 | fprintf(stderr, "kallsyms failure: " |
235 | "unable to allocate required amount of memory\n"); | |
236 | exit(EXIT_FAILURE); | |
237 | } | |
8d605269 MY |
238 | sym->addr = addr; |
239 | sym->len = len; | |
240 | sym->sym[0] = type; | |
9d1b3895 | 241 | strcpy(sym_name(sym), name); |
8d605269 | 242 | sym->percpu_absolute = 0; |
8c996940 | 243 | |
8d605269 | 244 | return sym; |
1da177e4 LT |
245 | } |
246 | ||
4bfe2b78 MY |
247 | static int symbol_in_range(const struct sym_entry *s, |
248 | const struct addr_range *ranges, int entries) | |
17b1f0de MF |
249 | { |
250 | size_t i; | |
4bfe2b78 | 251 | const struct addr_range *ar; |
17b1f0de | 252 | |
78eb7159 KC |
253 | for (i = 0; i < entries; ++i) { |
254 | ar = &ranges[i]; | |
17b1f0de | 255 | |
78eb7159 | 256 | if (s->addr >= ar->start && s->addr <= ar->end) |
ac6ca5c8 | 257 | return 1; |
17b1f0de MF |
258 | } |
259 | ||
ac6ca5c8 | 260 | return 0; |
17b1f0de MF |
261 | } |
262 | ||
4bfe2b78 | 263 | static int symbol_valid(const struct sym_entry *s) |
1da177e4 | 264 | { |
29e55ad3 | 265 | const char *name = sym_name(s); |
bd8b22d2 | 266 | |
1da177e4 LT |
267 | /* if --all-symbols is not specified, then symbols outside the text |
268 | * and inittext sections are discarded */ | |
269 | if (!all_symbols) { | |
78eb7159 KC |
270 | if (symbol_in_range(s, text_ranges, |
271 | ARRAY_SIZE(text_ranges)) == 0) | |
1da177e4 LT |
272 | return 0; |
273 | /* Corner case. Discard any symbols with the same value as | |
a3b81113 RG |
274 | * _etext _einittext; they can move between pass 1 and 2 when |
275 | * the kallsyms data are added. If these symbols move then | |
276 | * they may get dropped in pass 2, which breaks the kallsyms | |
277 | * rules. | |
1da177e4 | 278 | */ |
17b1f0de | 279 | if ((s->addr == text_range_text->end && |
29e55ad3 | 280 | strcmp(name, text_range_text->end_sym)) || |
17b1f0de | 281 | (s->addr == text_range_inittext->end && |
29e55ad3 | 282 | strcmp(name, text_range_inittext->end_sym))) |
1da177e4 LT |
283 | return 0; |
284 | } | |
285 | ||
1da177e4 LT |
286 | return 1; |
287 | } | |
288 | ||
5e5c4fa7 MY |
289 | /* remove all the invalid symbols from the table */ |
290 | static void shrink_table(void) | |
291 | { | |
292 | unsigned int i, pos; | |
293 | ||
294 | pos = 0; | |
295 | for (i = 0; i < table_cnt; i++) { | |
8d605269 | 296 | if (symbol_valid(table[i])) { |
5e5c4fa7 MY |
297 | if (pos != i) |
298 | table[pos] = table[i]; | |
299 | pos++; | |
300 | } else { | |
8d605269 | 301 | free(table[i]); |
5e5c4fa7 MY |
302 | } |
303 | } | |
304 | table_cnt = pos; | |
305 | ||
306 | /* When valid symbol is not registered, exit to error */ | |
307 | if (!table_cnt) { | |
308 | fprintf(stderr, "No valid symbol.\n"); | |
309 | exit(1); | |
310 | } | |
311 | } | |
312 | ||
b3dbb4ec | 313 | static void read_map(FILE *in) |
1da177e4 | 314 | { |
8d605269 MY |
315 | struct sym_entry *sym; |
316 | ||
1da177e4 | 317 | while (!feof(in)) { |
8d605269 MY |
318 | sym = read_symbol(in); |
319 | if (!sym) | |
320 | continue; | |
321 | ||
322 | sym->start_pos = table_cnt; | |
323 | ||
b3dbb4ec PM |
324 | if (table_cnt >= table_size) { |
325 | table_size += 10000; | |
326 | table = realloc(table, sizeof(*table) * table_size); | |
1da177e4 LT |
327 | if (!table) { |
328 | fprintf(stderr, "out of memory\n"); | |
329 | exit (1); | |
330 | } | |
331 | } | |
8d605269 MY |
332 | |
333 | table[table_cnt++] = sym; | |
1da177e4 LT |
334 | } |
335 | } | |
336 | ||
4bfe2b78 | 337 | static void output_label(const char *label) |
1da177e4 | 338 | { |
534c9f2e | 339 | printf(".globl %s\n", label); |
1da177e4 | 340 | printf("\tALGN\n"); |
534c9f2e | 341 | printf("%s:\n", label); |
1da177e4 LT |
342 | } |
343 | ||
fd2ab2f6 MY |
344 | /* Provide proper symbols relocatability by their '_text' relativeness. */ |
345 | static void output_address(unsigned long long addr) | |
346 | { | |
347 | if (_text <= addr) | |
348 | printf("\tPTR\t_text + %#llx\n", addr - _text); | |
349 | else | |
350 | printf("\tPTR\t_text - %#llx\n", _text - addr); | |
351 | } | |
352 | ||
1da177e4 LT |
353 | /* uncompress a compressed symbol. When this function is called, the best table |
354 | * might still be compressed itself, so the function needs to be recursive */ | |
4bfe2b78 | 355 | static int expand_symbol(const unsigned char *data, int len, char *result) |
1da177e4 LT |
356 | { |
357 | int c, rlen, total=0; | |
358 | ||
359 | while (len) { | |
360 | c = *data; | |
361 | /* if the table holds a single char that is the same as the one | |
362 | * we are looking for, then end the search */ | |
363 | if (best_table[c][0]==c && best_table_len[c]==1) { | |
364 | *result++ = c; | |
365 | total++; | |
366 | } else { | |
367 | /* if not, recurse and expand */ | |
368 | rlen = expand_symbol(best_table[c], best_table_len[c], result); | |
369 | total += rlen; | |
370 | result += rlen; | |
371 | } | |
372 | data++; | |
373 | len--; | |
374 | } | |
375 | *result=0; | |
376 | ||
377 | return total; | |
378 | } | |
379 | ||
4bfe2b78 | 380 | static int symbol_absolute(const struct sym_entry *s) |
78eb7159 | 381 | { |
8c996940 | 382 | return s->percpu_absolute; |
78eb7159 KC |
383 | } |
384 | ||
b3dbb4ec | 385 | static void write_src(void) |
1da177e4 | 386 | { |
b3dbb4ec | 387 | unsigned int i, k, off; |
1da177e4 LT |
388 | unsigned int best_idx[256]; |
389 | unsigned int *markers; | |
9281acea | 390 | char buf[KSYM_NAME_LEN]; |
1da177e4 | 391 | |
500193ec | 392 | printf("#include <asm/bitsperlong.h>\n"); |
1da177e4 LT |
393 | printf("#if BITS_PER_LONG == 64\n"); |
394 | printf("#define PTR .quad\n"); | |
72d3ebb9 | 395 | printf("#define ALGN .balign 8\n"); |
1da177e4 LT |
396 | printf("#else\n"); |
397 | printf("#define PTR .long\n"); | |
72d3ebb9 | 398 | printf("#define ALGN .balign 4\n"); |
1da177e4 LT |
399 | printf("#endif\n"); |
400 | ||
aad09470 | 401 | printf("\t.section .rodata, \"a\"\n"); |
1da177e4 | 402 | |
2213e9a6 AB |
403 | if (!base_relative) |
404 | output_label("kallsyms_addresses"); | |
405 | else | |
406 | output_label("kallsyms_offsets"); | |
407 | ||
b3dbb4ec | 408 | for (i = 0; i < table_cnt; i++) { |
2213e9a6 | 409 | if (base_relative) { |
fd2ab2f6 MY |
410 | /* |
411 | * Use the offset relative to the lowest value | |
412 | * encountered of all relative symbols, and emit | |
413 | * non-relocatable fixed offsets that will be fixed | |
414 | * up at runtime. | |
415 | */ | |
416 | ||
2213e9a6 AB |
417 | long long offset; |
418 | int overflow; | |
419 | ||
420 | if (!absolute_percpu) { | |
8d605269 | 421 | offset = table[i]->addr - relative_base; |
2213e9a6 | 422 | overflow = (offset < 0 || offset > UINT_MAX); |
8d605269 MY |
423 | } else if (symbol_absolute(table[i])) { |
424 | offset = table[i]->addr; | |
2213e9a6 AB |
425 | overflow = (offset < 0 || offset > INT_MAX); |
426 | } else { | |
8d605269 | 427 | offset = relative_base - table[i]->addr - 1; |
2213e9a6 AB |
428 | overflow = (offset < INT_MIN || offset >= 0); |
429 | } | |
430 | if (overflow) { | |
431 | fprintf(stderr, "kallsyms failure: " | |
432 | "%s symbol value %#llx out of range in relative mode\n", | |
8d605269 MY |
433 | symbol_absolute(table[i]) ? "absolute" : "relative", |
434 | table[i]->addr); | |
2213e9a6 AB |
435 | exit(EXIT_FAILURE); |
436 | } | |
437 | printf("\t.long\t%#x\n", (int)offset); | |
8d605269 MY |
438 | } else if (!symbol_absolute(table[i])) { |
439 | output_address(table[i]->addr); | |
fd593d12 | 440 | } else { |
8d605269 | 441 | printf("\tPTR\t%#llx\n", table[i]->addr); |
fd593d12 | 442 | } |
1da177e4 LT |
443 | } |
444 | printf("\n"); | |
445 | ||
2213e9a6 AB |
446 | if (base_relative) { |
447 | output_label("kallsyms_relative_base"); | |
fd2ab2f6 | 448 | output_address(relative_base); |
2213e9a6 AB |
449 | printf("\n"); |
450 | } | |
451 | ||
1da177e4 | 452 | output_label("kallsyms_num_syms"); |
80ffbaa5 | 453 | printf("\t.long\t%u\n", table_cnt); |
1da177e4 LT |
454 | printf("\n"); |
455 | ||
456 | /* table of offset markers, that give the offset in the compressed stream | |
457 | * every 256 symbols */ | |
f1a136e0 JJ |
458 | markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256)); |
459 | if (!markers) { | |
460 | fprintf(stderr, "kallsyms failure: " | |
461 | "unable to allocate required memory\n"); | |
462 | exit(EXIT_FAILURE); | |
463 | } | |
1da177e4 LT |
464 | |
465 | output_label("kallsyms_names"); | |
1da177e4 | 466 | off = 0; |
b3dbb4ec PM |
467 | for (i = 0; i < table_cnt; i++) { |
468 | if ((i & 0xFF) == 0) | |
469 | markers[i >> 8] = off; | |
1da177e4 | 470 | |
8d605269 MY |
471 | printf("\t.byte 0x%02x", table[i]->len); |
472 | for (k = 0; k < table[i]->len; k++) | |
473 | printf(", 0x%02x", table[i]->sym[k]); | |
1da177e4 LT |
474 | printf("\n"); |
475 | ||
8d605269 | 476 | off += table[i]->len + 1; |
1da177e4 LT |
477 | } |
478 | printf("\n"); | |
479 | ||
480 | output_label("kallsyms_markers"); | |
b3dbb4ec | 481 | for (i = 0; i < ((table_cnt + 255) >> 8); i++) |
80ffbaa5 | 482 | printf("\t.long\t%u\n", markers[i]); |
1da177e4 LT |
483 | printf("\n"); |
484 | ||
485 | free(markers); | |
486 | ||
487 | output_label("kallsyms_token_table"); | |
488 | off = 0; | |
489 | for (i = 0; i < 256; i++) { | |
490 | best_idx[i] = off; | |
b3dbb4ec | 491 | expand_symbol(best_table[i], best_table_len[i], buf); |
1da177e4 LT |
492 | printf("\t.asciz\t\"%s\"\n", buf); |
493 | off += strlen(buf) + 1; | |
494 | } | |
495 | printf("\n"); | |
496 | ||
497 | output_label("kallsyms_token_index"); | |
498 | for (i = 0; i < 256; i++) | |
499 | printf("\t.short\t%d\n", best_idx[i]); | |
500 | printf("\n"); | |
501 | } | |
502 | ||
503 | ||
504 | /* table lookup compression functions */ | |
505 | ||
1da177e4 | 506 | /* count all the possible tokens in a symbol */ |
4bfe2b78 | 507 | static void learn_symbol(const unsigned char *symbol, int len) |
1da177e4 LT |
508 | { |
509 | int i; | |
510 | ||
511 | for (i = 0; i < len - 1; i++) | |
b3dbb4ec | 512 | token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++; |
1da177e4 LT |
513 | } |
514 | ||
515 | /* decrease the count for all the possible tokens in a symbol */ | |
4bfe2b78 | 516 | static void forget_symbol(const unsigned char *symbol, int len) |
1da177e4 LT |
517 | { |
518 | int i; | |
519 | ||
520 | for (i = 0; i < len - 1; i++) | |
b3dbb4ec | 521 | token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--; |
1da177e4 LT |
522 | } |
523 | ||
5e5c4fa7 | 524 | /* do the initial token count */ |
1da177e4 LT |
525 | static void build_initial_tok_table(void) |
526 | { | |
5e5c4fa7 | 527 | unsigned int i; |
1da177e4 | 528 | |
5e5c4fa7 | 529 | for (i = 0; i < table_cnt; i++) |
8d605269 | 530 | learn_symbol(table[i]->sym, table[i]->len); |
1da177e4 LT |
531 | } |
532 | ||
2558c138 | 533 | static unsigned char *find_token(unsigned char *str, int len, |
4bfe2b78 | 534 | const unsigned char *token) |
7c5d249a PM |
535 | { |
536 | int i; | |
537 | ||
538 | for (i = 0; i < len - 1; i++) { | |
539 | if (str[i] == token[0] && str[i+1] == token[1]) | |
540 | return &str[i]; | |
541 | } | |
542 | return NULL; | |
543 | } | |
544 | ||
1da177e4 LT |
545 | /* replace a given token in all the valid symbols. Use the sampled symbols |
546 | * to update the counts */ | |
4bfe2b78 | 547 | static void compress_symbols(const unsigned char *str, int idx) |
1da177e4 | 548 | { |
b3dbb4ec PM |
549 | unsigned int i, len, size; |
550 | unsigned char *p1, *p2; | |
1da177e4 | 551 | |
b3dbb4ec | 552 | for (i = 0; i < table_cnt; i++) { |
1da177e4 | 553 | |
8d605269 MY |
554 | len = table[i]->len; |
555 | p1 = table[i]->sym; | |
b3dbb4ec PM |
556 | |
557 | /* find the token on the symbol */ | |
7c5d249a | 558 | p2 = find_token(p1, len, str); |
b3dbb4ec PM |
559 | if (!p2) continue; |
560 | ||
561 | /* decrease the counts for this symbol's tokens */ | |
8d605269 | 562 | forget_symbol(table[i]->sym, len); |
b3dbb4ec PM |
563 | |
564 | size = len; | |
1da177e4 LT |
565 | |
566 | do { | |
b3dbb4ec PM |
567 | *p2 = idx; |
568 | p2++; | |
569 | size -= (p2 - p1); | |
570 | memmove(p2, p2 + 1, size); | |
571 | p1 = p2; | |
572 | len--; | |
573 | ||
574 | if (size < 2) break; | |
575 | ||
1da177e4 | 576 | /* find the token on the symbol */ |
7c5d249a | 577 | p2 = find_token(p1, size, str); |
1da177e4 | 578 | |
b3dbb4ec | 579 | } while (p2); |
1da177e4 | 580 | |
8d605269 | 581 | table[i]->len = len; |
1da177e4 | 582 | |
b3dbb4ec | 583 | /* increase the counts for this symbol's new tokens */ |
8d605269 | 584 | learn_symbol(table[i]->sym, len); |
1da177e4 LT |
585 | } |
586 | } | |
587 | ||
588 | /* search the token with the maximum profit */ | |
b3dbb4ec | 589 | static int find_best_token(void) |
1da177e4 | 590 | { |
b3dbb4ec | 591 | int i, best, bestprofit; |
1da177e4 LT |
592 | |
593 | bestprofit=-10000; | |
b3dbb4ec | 594 | best = 0; |
1da177e4 | 595 | |
b3dbb4ec PM |
596 | for (i = 0; i < 0x10000; i++) { |
597 | if (token_profit[i] > bestprofit) { | |
598 | best = i; | |
599 | bestprofit = token_profit[i]; | |
1da177e4 | 600 | } |
1da177e4 | 601 | } |
1da177e4 LT |
602 | return best; |
603 | } | |
604 | ||
605 | /* this is the core of the algorithm: calculate the "best" table */ | |
606 | static void optimize_result(void) | |
607 | { | |
b3dbb4ec | 608 | int i, best; |
1da177e4 LT |
609 | |
610 | /* using the '\0' symbol last allows compress_symbols to use standard | |
611 | * fast string functions */ | |
612 | for (i = 255; i >= 0; i--) { | |
613 | ||
614 | /* if this table slot is empty (it is not used by an actual | |
615 | * original char code */ | |
616 | if (!best_table_len[i]) { | |
617 | ||
cbf7a90e | 618 | /* find the token with the best profit value */ |
1da177e4 | 619 | best = find_best_token(); |
e0a04b11 XW |
620 | if (token_profit[best] == 0) |
621 | break; | |
1da177e4 LT |
622 | |
623 | /* place it in the "best" table */ | |
b3dbb4ec PM |
624 | best_table_len[i] = 2; |
625 | best_table[i][0] = best & 0xFF; | |
626 | best_table[i][1] = (best >> 8) & 0xFF; | |
1da177e4 LT |
627 | |
628 | /* replace this token in all the valid symbols */ | |
b3dbb4ec | 629 | compress_symbols(best_table[i], i); |
1da177e4 LT |
630 | } |
631 | } | |
632 | } | |
633 | ||
634 | /* start by placing the symbols that are actually used on the table */ | |
635 | static void insert_real_symbols_in_table(void) | |
636 | { | |
b3dbb4ec | 637 | unsigned int i, j, c; |
1da177e4 | 638 | |
b3dbb4ec | 639 | for (i = 0; i < table_cnt; i++) { |
8d605269 MY |
640 | for (j = 0; j < table[i]->len; j++) { |
641 | c = table[i]->sym[j]; | |
b3dbb4ec PM |
642 | best_table[c][0]=c; |
643 | best_table_len[c]=1; | |
1da177e4 LT |
644 | } |
645 | } | |
646 | } | |
647 | ||
648 | static void optimize_token_table(void) | |
649 | { | |
1da177e4 LT |
650 | build_initial_tok_table(); |
651 | ||
652 | insert_real_symbols_in_table(); | |
653 | ||
654 | optimize_result(); | |
655 | } | |
656 | ||
b478b782 LJ |
657 | /* guess for "linker script provide" symbol */ |
658 | static int may_be_linker_script_provide_symbol(const struct sym_entry *se) | |
659 | { | |
29e55ad3 | 660 | const char *symbol = sym_name(se); |
b478b782 LJ |
661 | int len = se->len - 1; |
662 | ||
663 | if (len < 8) | |
664 | return 0; | |
665 | ||
666 | if (symbol[0] != '_' || symbol[1] != '_') | |
667 | return 0; | |
668 | ||
669 | /* __start_XXXXX */ | |
670 | if (!memcmp(symbol + 2, "start_", 6)) | |
671 | return 1; | |
672 | ||
673 | /* __stop_XXXXX */ | |
674 | if (!memcmp(symbol + 2, "stop_", 5)) | |
675 | return 1; | |
676 | ||
677 | /* __end_XXXXX */ | |
678 | if (!memcmp(symbol + 2, "end_", 4)) | |
679 | return 1; | |
680 | ||
681 | /* __XXXXX_start */ | |
682 | if (!memcmp(symbol + len - 6, "_start", 6)) | |
683 | return 1; | |
684 | ||
685 | /* __XXXXX_end */ | |
686 | if (!memcmp(symbol + len - 4, "_end", 4)) | |
687 | return 1; | |
688 | ||
689 | return 0; | |
690 | } | |
691 | ||
f2df3f65 PM |
692 | static int compare_symbols(const void *a, const void *b) |
693 | { | |
8d605269 MY |
694 | const struct sym_entry *sa = *(const struct sym_entry **)a; |
695 | const struct sym_entry *sb = *(const struct sym_entry **)b; | |
f2df3f65 PM |
696 | int wa, wb; |
697 | ||
f2df3f65 PM |
698 | /* sort by address first */ |
699 | if (sa->addr > sb->addr) | |
700 | return 1; | |
701 | if (sa->addr < sb->addr) | |
702 | return -1; | |
703 | ||
704 | /* sort by "weakness" type */ | |
705 | wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W'); | |
706 | wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W'); | |
707 | if (wa != wb) | |
708 | return wa - wb; | |
709 | ||
b478b782 LJ |
710 | /* sort by "linker script provide" type */ |
711 | wa = may_be_linker_script_provide_symbol(sa); | |
712 | wb = may_be_linker_script_provide_symbol(sb); | |
713 | if (wa != wb) | |
714 | return wa - wb; | |
715 | ||
716 | /* sort by the number of prefix underscores */ | |
aa915245 MY |
717 | wa = strspn(sym_name(sa), "_"); |
718 | wb = strspn(sym_name(sb), "_"); | |
b478b782 LJ |
719 | if (wa != wb) |
720 | return wa - wb; | |
721 | ||
f2df3f65 PM |
722 | /* sort by initial order, so that other symbols are left undisturbed */ |
723 | return sa->start_pos - sb->start_pos; | |
724 | } | |
725 | ||
726 | static void sort_symbols(void) | |
727 | { | |
8d605269 | 728 | qsort(table, table_cnt, sizeof(table[0]), compare_symbols); |
f2df3f65 | 729 | } |
1da177e4 | 730 | |
c6bda7c9 RR |
731 | static void make_percpus_absolute(void) |
732 | { | |
733 | unsigned int i; | |
734 | ||
735 | for (i = 0; i < table_cnt; i++) | |
8d605269 | 736 | if (symbol_in_range(table[i], &percpu_range, 1)) { |
8c996940 AB |
737 | /* |
738 | * Keep the 'A' override for percpu symbols to | |
739 | * ensure consistent behavior compared to older | |
740 | * versions of this tool. | |
741 | */ | |
8d605269 MY |
742 | table[i]->sym[0] = 'A'; |
743 | table[i]->percpu_absolute = 1; | |
8c996940 | 744 | } |
c6bda7c9 RR |
745 | } |
746 | ||
2213e9a6 AB |
747 | /* find the minimum non-absolute symbol address */ |
748 | static void record_relative_base(void) | |
749 | { | |
750 | unsigned int i; | |
751 | ||
2213e9a6 | 752 | for (i = 0; i < table_cnt; i++) |
8d605269 | 753 | if (!symbol_absolute(table[i])) { |
f34ea029 MY |
754 | /* |
755 | * The table is sorted by address. | |
756 | * Take the first non-absolute symbol value. | |
757 | */ | |
8d605269 | 758 | relative_base = table[i]->addr; |
f34ea029 MY |
759 | return; |
760 | } | |
2213e9a6 AB |
761 | } |
762 | ||
b3dbb4ec | 763 | int main(int argc, char **argv) |
1da177e4 | 764 | { |
41f11a4f YS |
765 | if (argc >= 2) { |
766 | int i; | |
767 | for (i = 1; i < argc; i++) { | |
768 | if(strcmp(argv[i], "--all-symbols") == 0) | |
769 | all_symbols = 1; | |
c6bda7c9 RR |
770 | else if (strcmp(argv[i], "--absolute-percpu") == 0) |
771 | absolute_percpu = 1; | |
534c9f2e | 772 | else if (strcmp(argv[i], "--base-relative") == 0) |
2213e9a6 AB |
773 | base_relative = 1; |
774 | else | |
41f11a4f YS |
775 | usage(); |
776 | } | |
777 | } else if (argc != 1) | |
1da177e4 LT |
778 | usage(); |
779 | ||
780 | read_map(stdin); | |
5e5c4fa7 | 781 | shrink_table(); |
c6bda7c9 RR |
782 | if (absolute_percpu) |
783 | make_percpus_absolute(); | |
f34ea029 | 784 | sort_symbols(); |
2213e9a6 AB |
785 | if (base_relative) |
786 | record_relative_base(); | |
2ea03891 | 787 | optimize_token_table(); |
1da177e4 LT |
788 | write_src(); |
789 | ||
790 | return 0; | |
791 | } |