blob: e3902fb39afd1f2c53f63c0cfa9aaa1273956a02 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001/* 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 *
Linus Torvalds1da177e2005-04-16 15:20:36 -070010 * 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>
25
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040026#ifndef ARRAY_SIZE
27#define ARRAY_SIZE(arr) (sizeof(arr) / sizeof(arr[0]))
28#endif
29
Tejun Heo9281ace2007-07-17 04:03:51 -070030#define KSYM_NAME_LEN 128
Linus Torvalds1da177e2005-04-16 15:20:36 -070031
Linus Torvalds1da177e2005-04-16 15:20:36 -070032struct sym_entry {
33 unsigned long long addr;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070034 unsigned int len;
Paulo Marquesf2df3f62008-02-06 01:37:33 -080035 unsigned int start_pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -070036 unsigned char *sym;
37};
38
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040039struct text_range {
40 const char *stext, *etext;
41 unsigned long long start, end;
42};
43
44static unsigned long long _text;
45static struct text_range text_ranges[] = {
46 { "_stext", "_etext" },
47 { "_sinittext", "_einittext" },
48 { "_stext_l1", "_etext_l1" }, /* Blackfin on-chip L1 inst SRAM */
49 { "_stext_l2", "_etext_l2" }, /* Blackfin on-chip L2 SRAM */
50};
51#define text_range_text (&text_ranges[0])
52#define text_range_inittext (&text_ranges[1])
53
Linus Torvalds1da177e2005-04-16 15:20:36 -070054static struct sym_entry *table;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070055static unsigned int table_size, table_cnt;
Linus Torvalds1da177e2005-04-16 15:20:36 -070056static int all_symbols = 0;
Yoshinori Sato41f11a42005-05-01 08:59:06 -070057static char symbol_prefix_char = '\0';
Linus Torvalds1da177e2005-04-16 15:20:36 -070058
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070059int token_profit[0x10000];
Linus Torvalds1da177e2005-04-16 15:20:36 -070060
61/* the table that holds the result of the compression */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070062unsigned char best_table[256][2];
Linus Torvalds1da177e2005-04-16 15:20:36 -070063unsigned char best_table_len[256];
64
65
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070066static void usage(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -070067{
Yoshinori Sato41f11a42005-05-01 08:59:06 -070068 fprintf(stderr, "Usage: kallsyms [--all-symbols] [--symbol-prefix=<prefix char>] < in.map > out.S\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -070069 exit(1);
70}
71
72/*
73 * This ignores the intensely annoying "mapping symbols" found
74 * in ARM ELF files: $a, $t and $d.
75 */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070076static inline int is_arm_mapping_symbol(const char *str)
Linus Torvalds1da177e2005-04-16 15:20:36 -070077{
78 return str[0] == '$' && strchr("atd", str[1])
79 && (str[2] == '\0' || str[2] == '.');
80}
81
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040082static int read_symbol_tr(const char *sym, unsigned long long addr)
83{
84 size_t i;
85 struct text_range *tr;
86
87 for (i = 0; i < ARRAY_SIZE(text_ranges); ++i) {
88 tr = &text_ranges[i];
89
90 if (strcmp(sym, tr->stext) == 0) {
91 tr->start = addr;
92 return 0;
93 } else if (strcmp(sym, tr->etext) == 0) {
94 tr->end = addr;
95 return 0;
96 }
97 }
98
99 return 1;
100}
101
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700102static int read_symbol(FILE *in, struct sym_entry *s)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700103{
104 char str[500];
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700105 char *sym, stype;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700106 int rc;
107
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700108 rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, str);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700109 if (rc != 3) {
110 if (rc != EOF) {
Himanshu Chauhan71d41ae2010-01-27 16:53:20 -0800111 /* skip line. sym is used as dummy to
112 * shut of "warn_unused_result" warning.
113 */
114 sym = fgets(str, 500, in);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700115 }
116 return -1;
117 }
118
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700119 sym = str;
120 /* skip prefix char */
121 if (symbol_prefix_char && str[0] == symbol_prefix_char)
122 sym++;
123
Linus Torvalds1da177e2005-04-16 15:20:36 -0700124 /* Ignore most absolute/undefined (?) symbols. */
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100125 if (strcmp(sym, "_text") == 0)
126 _text = s->addr;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400127 else if (read_symbol_tr(sym, s->addr) == 0)
128 /* nothing to do */;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700129 else if (toupper(stype) == 'A')
Linus Torvalds1da177e2005-04-16 15:20:36 -0700130 {
131 /* Keep these useful absolute symbols */
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700132 if (strcmp(sym, "__kernel_syscall_via_break") &&
133 strcmp(sym, "__kernel_syscall_via_epc") &&
134 strcmp(sym, "__kernel_sigtramp") &&
135 strcmp(sym, "__gp"))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700136 return -1;
137
138 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700139 else if (toupper(stype) == 'U' ||
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700140 is_arm_mapping_symbol(sym))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700141 return -1;
Ralf Baechle6f00df22005-09-06 15:16:41 -0700142 /* exclude also MIPS ELF local symbols ($L123 instead of .L123) */
143 else if (str[0] == '$')
144 return -1;
Sam Ravnborgaab34ac2008-05-19 20:07:58 +0200145 /* exclude debugging symbols */
146 else if (stype == 'N')
147 return -1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700148
149 /* include the type field in the symbol name, so that it gets
150 * compressed together */
151 s->len = strlen(str) + 1;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700152 s->sym = malloc(s->len + 1);
Jesper Juhlf1a136e2006-03-25 03:07:46 -0800153 if (!s->sym) {
154 fprintf(stderr, "kallsyms failure: "
155 "unable to allocate required amount of memory\n");
156 exit(EXIT_FAILURE);
157 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700158 strcpy((char *)s->sym + 1, str);
159 s->sym[0] = stype;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700160
161 return 0;
162}
163
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400164static int symbol_valid_tr(struct sym_entry *s)
165{
166 size_t i;
167 struct text_range *tr;
168
169 for (i = 0; i < ARRAY_SIZE(text_ranges); ++i) {
170 tr = &text_ranges[i];
171
Mike Frysingerac6ca5c2009-06-15 07:52:48 -0400172 if (s->addr >= tr->start && s->addr <= tr->end)
173 return 1;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400174 }
175
Mike Frysingerac6ca5c2009-06-15 07:52:48 -0400176 return 0;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400177}
178
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700179static int symbol_valid(struct sym_entry *s)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700180{
181 /* Symbols which vary between passes. Passes 1 and 2 must have
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100182 * identical symbol lists. The kallsyms_* symbols below are only added
183 * after pass 1, they would be included in pass 2 when --all-symbols is
184 * specified so exclude them to get a stable symbol list.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700185 */
186 static char *special_symbols[] = {
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100187 "kallsyms_addresses",
188 "kallsyms_num_syms",
189 "kallsyms_names",
190 "kallsyms_markers",
191 "kallsyms_token_table",
192 "kallsyms_token_index",
193
Linus Torvalds1da177e2005-04-16 15:20:36 -0700194 /* Exclude linker generated symbols which vary between passes */
195 "_SDA_BASE_", /* ppc */
196 "_SDA2_BASE_", /* ppc */
197 NULL };
198 int i;
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700199 int offset = 1;
200
201 /* skip prefix char */
202 if (symbol_prefix_char && *(s->sym + 1) == symbol_prefix_char)
203 offset++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700204
205 /* if --all-symbols is not specified, then symbols outside the text
206 * and inittext sections are discarded */
207 if (!all_symbols) {
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400208 if (symbol_valid_tr(s) == 0)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700209 return 0;
210 /* Corner case. Discard any symbols with the same value as
Robin Getza3b81112008-02-06 01:36:26 -0800211 * _etext _einittext; they can move between pass 1 and 2 when
212 * the kallsyms data are added. If these symbols move then
213 * they may get dropped in pass 2, which breaks the kallsyms
214 * rules.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700215 */
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400216 if ((s->addr == text_range_text->end &&
217 strcmp((char *)s->sym + offset, text_range_text->etext)) ||
218 (s->addr == text_range_inittext->end &&
219 strcmp((char *)s->sym + offset, text_range_inittext->etext)))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700220 return 0;
221 }
222
223 /* Exclude symbols which vary between passes. */
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100224 if (strstr((char *)s->sym + offset, "_compiled."))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700225 return 0;
226
227 for (i = 0; special_symbols[i]; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700228 if( strcmp((char *)s->sym + offset, special_symbols[i]) == 0 )
Linus Torvalds1da177e2005-04-16 15:20:36 -0700229 return 0;
230
231 return 1;
232}
233
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700234static void read_map(FILE *in)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700235{
236 while (!feof(in)) {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700237 if (table_cnt >= table_size) {
238 table_size += 10000;
239 table = realloc(table, sizeof(*table) * table_size);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700240 if (!table) {
241 fprintf(stderr, "out of memory\n");
242 exit (1);
243 }
244 }
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800245 if (read_symbol(in, &table[table_cnt]) == 0) {
246 table[table_cnt].start_pos = table_cnt;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700247 table_cnt++;
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800248 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700249 }
250}
251
252static void output_label(char *label)
253{
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700254 if (symbol_prefix_char)
255 printf(".globl %c%s\n", symbol_prefix_char, label);
256 else
257 printf(".globl %s\n", label);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700258 printf("\tALGN\n");
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700259 if (symbol_prefix_char)
260 printf("%c%s:\n", symbol_prefix_char, label);
261 else
262 printf("%s:\n", label);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700263}
264
265/* uncompress a compressed symbol. When this function is called, the best table
266 * might still be compressed itself, so the function needs to be recursive */
267static int expand_symbol(unsigned char *data, int len, char *result)
268{
269 int c, rlen, total=0;
270
271 while (len) {
272 c = *data;
273 /* if the table holds a single char that is the same as the one
274 * we are looking for, then end the search */
275 if (best_table[c][0]==c && best_table_len[c]==1) {
276 *result++ = c;
277 total++;
278 } else {
279 /* if not, recurse and expand */
280 rlen = expand_symbol(best_table[c], best_table_len[c], result);
281 total += rlen;
282 result += rlen;
283 }
284 data++;
285 len--;
286 }
287 *result=0;
288
289 return total;
290}
291
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700292static void write_src(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700293{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700294 unsigned int i, k, off;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700295 unsigned int best_idx[256];
296 unsigned int *markers;
Tejun Heo9281ace2007-07-17 04:03:51 -0700297 char buf[KSYM_NAME_LEN];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700298
299 printf("#include <asm/types.h>\n");
300 printf("#if BITS_PER_LONG == 64\n");
301 printf("#define PTR .quad\n");
302 printf("#define ALGN .align 8\n");
303 printf("#else\n");
304 printf("#define PTR .long\n");
305 printf("#define ALGN .align 4\n");
306 printf("#endif\n");
307
Jan Beulichaad09472006-12-08 02:35:57 -0800308 printf("\t.section .rodata, \"a\"\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700309
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100310 /* Provide proper symbols relocatability by their '_text'
311 * relativeness. The symbol names cannot be used to construct
312 * normal symbol references as the list of symbols contains
313 * symbols that are declared static and are private to their
314 * .o files. This prevents .tmp_kallsyms.o or any other
315 * object from referencing them.
316 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700317 output_label("kallsyms_addresses");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700318 for (i = 0; i < table_cnt; i++) {
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100319 if (toupper(table[i].sym[0]) != 'A') {
Vivek Goyal2c22d8b2006-12-07 02:14:10 +0100320 if (_text <= table[i].addr)
321 printf("\tPTR\t_text + %#llx\n",
322 table[i].addr - _text);
323 else
324 printf("\tPTR\t_text - %#llx\n",
325 _text - table[i].addr);
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100326 } else {
327 printf("\tPTR\t%#llx\n", table[i].addr);
328 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700329 }
330 printf("\n");
331
332 output_label("kallsyms_num_syms");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700333 printf("\tPTR\t%d\n", table_cnt);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700334 printf("\n");
335
336 /* table of offset markers, that give the offset in the compressed stream
337 * every 256 symbols */
Jesper Juhlf1a136e2006-03-25 03:07:46 -0800338 markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
339 if (!markers) {
340 fprintf(stderr, "kallsyms failure: "
341 "unable to allocate required memory\n");
342 exit(EXIT_FAILURE);
343 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700344
345 output_label("kallsyms_names");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700346 off = 0;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700347 for (i = 0; i < table_cnt; i++) {
348 if ((i & 0xFF) == 0)
349 markers[i >> 8] = off;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700350
351 printf("\t.byte 0x%02x", table[i].len);
352 for (k = 0; k < table[i].len; k++)
353 printf(", 0x%02x", table[i].sym[k]);
354 printf("\n");
355
356 off += table[i].len + 1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700357 }
358 printf("\n");
359
360 output_label("kallsyms_markers");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700361 for (i = 0; i < ((table_cnt + 255) >> 8); i++)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700362 printf("\tPTR\t%d\n", markers[i]);
363 printf("\n");
364
365 free(markers);
366
367 output_label("kallsyms_token_table");
368 off = 0;
369 for (i = 0; i < 256; i++) {
370 best_idx[i] = off;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700371 expand_symbol(best_table[i], best_table_len[i], buf);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700372 printf("\t.asciz\t\"%s\"\n", buf);
373 off += strlen(buf) + 1;
374 }
375 printf("\n");
376
377 output_label("kallsyms_token_index");
378 for (i = 0; i < 256; i++)
379 printf("\t.short\t%d\n", best_idx[i]);
380 printf("\n");
381}
382
383
384/* table lookup compression functions */
385
Linus Torvalds1da177e2005-04-16 15:20:36 -0700386/* count all the possible tokens in a symbol */
387static void learn_symbol(unsigned char *symbol, int len)
388{
389 int i;
390
391 for (i = 0; i < len - 1; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700392 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700393}
394
395/* decrease the count for all the possible tokens in a symbol */
396static void forget_symbol(unsigned char *symbol, int len)
397{
398 int i;
399
400 for (i = 0; i < len - 1; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700401 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700402}
403
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700404/* remove all the invalid symbols from the table and do the initial token count */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700405static void build_initial_tok_table(void)
406{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700407 unsigned int i, pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700408
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700409 pos = 0;
410 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700411 if ( symbol_valid(&table[i]) ) {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700412 if (pos != i)
413 table[pos] = table[i];
414 learn_symbol(table[pos].sym, table[pos].len);
415 pos++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700416 }
417 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700418 table_cnt = pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700419}
420
Paulo Marques7c5d2492007-06-20 18:09:00 +0100421static void *find_token(unsigned char *str, int len, unsigned char *token)
422{
423 int i;
424
425 for (i = 0; i < len - 1; i++) {
426 if (str[i] == token[0] && str[i+1] == token[1])
427 return &str[i];
428 }
429 return NULL;
430}
431
Linus Torvalds1da177e2005-04-16 15:20:36 -0700432/* replace a given token in all the valid symbols. Use the sampled symbols
433 * to update the counts */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700434static void compress_symbols(unsigned char *str, int idx)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700435{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700436 unsigned int i, len, size;
437 unsigned char *p1, *p2;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700438
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700439 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700440
441 len = table[i].len;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700442 p1 = table[i].sym;
443
444 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100445 p2 = find_token(p1, len, str);
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700446 if (!p2) continue;
447
448 /* decrease the counts for this symbol's tokens */
449 forget_symbol(table[i].sym, len);
450
451 size = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700452
453 do {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700454 *p2 = idx;
455 p2++;
456 size -= (p2 - p1);
457 memmove(p2, p2 + 1, size);
458 p1 = p2;
459 len--;
460
461 if (size < 2) break;
462
Linus Torvalds1da177e2005-04-16 15:20:36 -0700463 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100464 p2 = find_token(p1, size, str);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700465
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700466 } while (p2);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700467
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700468 table[i].len = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700469
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700470 /* increase the counts for this symbol's new tokens */
471 learn_symbol(table[i].sym, len);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700472 }
473}
474
475/* search the token with the maximum profit */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700476static int find_best_token(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700477{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700478 int i, best, bestprofit;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700479
480 bestprofit=-10000;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700481 best = 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700482
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700483 for (i = 0; i < 0x10000; i++) {
484 if (token_profit[i] > bestprofit) {
485 best = i;
486 bestprofit = token_profit[i];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700487 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700488 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700489 return best;
490}
491
492/* this is the core of the algorithm: calculate the "best" table */
493static void optimize_result(void)
494{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700495 int i, best;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700496
497 /* using the '\0' symbol last allows compress_symbols to use standard
498 * fast string functions */
499 for (i = 255; i >= 0; i--) {
500
501 /* if this table slot is empty (it is not used by an actual
502 * original char code */
503 if (!best_table_len[i]) {
504
505 /* find the token with the breates profit value */
506 best = find_best_token();
507
508 /* place it in the "best" table */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700509 best_table_len[i] = 2;
510 best_table[i][0] = best & 0xFF;
511 best_table[i][1] = (best >> 8) & 0xFF;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700512
513 /* replace this token in all the valid symbols */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700514 compress_symbols(best_table[i], i);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700515 }
516 }
517}
518
519/* start by placing the symbols that are actually used on the table */
520static void insert_real_symbols_in_table(void)
521{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700522 unsigned int i, j, c;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700523
524 memset(best_table, 0, sizeof(best_table));
525 memset(best_table_len, 0, sizeof(best_table_len));
526
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700527 for (i = 0; i < table_cnt; i++) {
528 for (j = 0; j < table[i].len; j++) {
529 c = table[i].sym[j];
530 best_table[c][0]=c;
531 best_table_len[c]=1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700532 }
533 }
534}
535
536static void optimize_token_table(void)
537{
Linus Torvalds1da177e2005-04-16 15:20:36 -0700538 build_initial_tok_table();
539
540 insert_real_symbols_in_table();
541
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700542 /* When valid symbol is not registered, exit to error */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700543 if (!table_cnt) {
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700544 fprintf(stderr, "No valid symbol.\n");
545 exit(1);
546 }
547
Linus Torvalds1da177e2005-04-16 15:20:36 -0700548 optimize_result();
549}
550
Lai Jiangshanb478b782009-03-13 15:10:26 +0800551/* guess for "linker script provide" symbol */
552static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
553{
554 const char *symbol = (char *)se->sym + 1;
555 int len = se->len - 1;
556
557 if (len < 8)
558 return 0;
559
560 if (symbol[0] != '_' || symbol[1] != '_')
561 return 0;
562
563 /* __start_XXXXX */
564 if (!memcmp(symbol + 2, "start_", 6))
565 return 1;
566
567 /* __stop_XXXXX */
568 if (!memcmp(symbol + 2, "stop_", 5))
569 return 1;
570
571 /* __end_XXXXX */
572 if (!memcmp(symbol + 2, "end_", 4))
573 return 1;
574
575 /* __XXXXX_start */
576 if (!memcmp(symbol + len - 6, "_start", 6))
577 return 1;
578
579 /* __XXXXX_end */
580 if (!memcmp(symbol + len - 4, "_end", 4))
581 return 1;
582
583 return 0;
584}
585
586static int prefix_underscores_count(const char *str)
587{
588 const char *tail = str;
589
Paul Mundta9ece532009-09-22 16:44:12 -0700590 while (*tail == '_')
Lai Jiangshanb478b782009-03-13 15:10:26 +0800591 tail++;
592
593 return tail - str;
594}
595
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800596static int compare_symbols(const void *a, const void *b)
597{
598 const struct sym_entry *sa;
599 const struct sym_entry *sb;
600 int wa, wb;
601
602 sa = a;
603 sb = b;
604
605 /* sort by address first */
606 if (sa->addr > sb->addr)
607 return 1;
608 if (sa->addr < sb->addr)
609 return -1;
610
611 /* sort by "weakness" type */
612 wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
613 wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
614 if (wa != wb)
615 return wa - wb;
616
Lai Jiangshanb478b782009-03-13 15:10:26 +0800617 /* sort by "linker script provide" type */
618 wa = may_be_linker_script_provide_symbol(sa);
619 wb = may_be_linker_script_provide_symbol(sb);
620 if (wa != wb)
621 return wa - wb;
622
623 /* sort by the number of prefix underscores */
624 wa = prefix_underscores_count((const char *)sa->sym + 1);
625 wb = prefix_underscores_count((const char *)sb->sym + 1);
626 if (wa != wb)
627 return wa - wb;
628
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800629 /* sort by initial order, so that other symbols are left undisturbed */
630 return sa->start_pos - sb->start_pos;
631}
632
633static void sort_symbols(void)
634{
635 qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
636}
Linus Torvalds1da177e2005-04-16 15:20:36 -0700637
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700638int main(int argc, char **argv)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700639{
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700640 if (argc >= 2) {
641 int i;
642 for (i = 1; i < argc; i++) {
643 if(strcmp(argv[i], "--all-symbols") == 0)
644 all_symbols = 1;
645 else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
646 char *p = &argv[i][16];
647 /* skip quote */
648 if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
649 p++;
650 symbol_prefix_char = *p;
651 } else
652 usage();
653 }
654 } else if (argc != 1)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700655 usage();
656
657 read_map(stdin);
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100658 sort_symbols();
659 optimize_token_table();
Linus Torvalds1da177e2005-04-16 15:20:36 -0700660 write_src();
661
662 return 0;
663}