blob: 64343cc084b41f2f9f813742c381717e41ab77ac [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) {
111 /* skip line */
112 fgets(str, 500, in);
113 }
114 return -1;
115 }
116
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700117 sym = str;
118 /* skip prefix char */
119 if (symbol_prefix_char && str[0] == symbol_prefix_char)
120 sym++;
121
Linus Torvalds1da177e2005-04-16 15:20:36 -0700122 /* Ignore most absolute/undefined (?) symbols. */
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100123 if (strcmp(sym, "_text") == 0)
124 _text = s->addr;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400125 else if (read_symbol_tr(sym, s->addr) == 0)
126 /* nothing to do */;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700127 else if (toupper(stype) == 'A')
Linus Torvalds1da177e2005-04-16 15:20:36 -0700128 {
129 /* Keep these useful absolute symbols */
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700130 if (strcmp(sym, "__kernel_syscall_via_break") &&
131 strcmp(sym, "__kernel_syscall_via_epc") &&
132 strcmp(sym, "__kernel_sigtramp") &&
133 strcmp(sym, "__gp"))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700134 return -1;
135
136 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700137 else if (toupper(stype) == 'U' ||
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700138 is_arm_mapping_symbol(sym))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700139 return -1;
Ralf Baechle6f00df22005-09-06 15:16:41 -0700140 /* exclude also MIPS ELF local symbols ($L123 instead of .L123) */
141 else if (str[0] == '$')
142 return -1;
Sam Ravnborgaab34ac2008-05-19 20:07:58 +0200143 /* exclude debugging symbols */
144 else if (stype == 'N')
145 return -1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700146
147 /* include the type field in the symbol name, so that it gets
148 * compressed together */
149 s->len = strlen(str) + 1;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700150 s->sym = malloc(s->len + 1);
Jesper Juhlf1a136e2006-03-25 03:07:46 -0800151 if (!s->sym) {
152 fprintf(stderr, "kallsyms failure: "
153 "unable to allocate required amount of memory\n");
154 exit(EXIT_FAILURE);
155 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700156 strcpy((char *)s->sym + 1, str);
157 s->sym[0] = stype;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700158
159 return 0;
160}
161
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400162static int symbol_valid_tr(struct sym_entry *s)
163{
164 size_t i;
165 struct text_range *tr;
166
167 for (i = 0; i < ARRAY_SIZE(text_ranges); ++i) {
168 tr = &text_ranges[i];
169
Mike Frysingerac6ca5c2009-06-15 07:52:48 -0400170 if (s->addr >= tr->start && s->addr <= tr->end)
171 return 1;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400172 }
173
Mike Frysingerac6ca5c2009-06-15 07:52:48 -0400174 return 0;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400175}
176
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700177static int symbol_valid(struct sym_entry *s)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700178{
179 /* Symbols which vary between passes. Passes 1 and 2 must have
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100180 * identical symbol lists. The kallsyms_* symbols below are only added
181 * after pass 1, they would be included in pass 2 when --all-symbols is
182 * specified so exclude them to get a stable symbol list.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700183 */
184 static char *special_symbols[] = {
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100185 "kallsyms_addresses",
186 "kallsyms_num_syms",
187 "kallsyms_names",
188 "kallsyms_markers",
189 "kallsyms_token_table",
190 "kallsyms_token_index",
191
Linus Torvalds1da177e2005-04-16 15:20:36 -0700192 /* Exclude linker generated symbols which vary between passes */
193 "_SDA_BASE_", /* ppc */
194 "_SDA2_BASE_", /* ppc */
195 NULL };
196 int i;
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700197 int offset = 1;
198
199 /* skip prefix char */
200 if (symbol_prefix_char && *(s->sym + 1) == symbol_prefix_char)
201 offset++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700202
203 /* if --all-symbols is not specified, then symbols outside the text
204 * and inittext sections are discarded */
205 if (!all_symbols) {
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400206 if (symbol_valid_tr(s) == 0)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700207 return 0;
208 /* Corner case. Discard any symbols with the same value as
Robin Getza3b81112008-02-06 01:36:26 -0800209 * _etext _einittext; they can move between pass 1 and 2 when
210 * the kallsyms data are added. If these symbols move then
211 * they may get dropped in pass 2, which breaks the kallsyms
212 * rules.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700213 */
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400214 if ((s->addr == text_range_text->end &&
215 strcmp((char *)s->sym + offset, text_range_text->etext)) ||
216 (s->addr == text_range_inittext->end &&
217 strcmp((char *)s->sym + offset, text_range_inittext->etext)))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700218 return 0;
219 }
220
221 /* Exclude symbols which vary between passes. */
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100222 if (strstr((char *)s->sym + offset, "_compiled."))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700223 return 0;
224
225 for (i = 0; special_symbols[i]; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700226 if( strcmp((char *)s->sym + offset, special_symbols[i]) == 0 )
Linus Torvalds1da177e2005-04-16 15:20:36 -0700227 return 0;
228
229 return 1;
230}
231
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700232static void read_map(FILE *in)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700233{
234 while (!feof(in)) {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700235 if (table_cnt >= table_size) {
236 table_size += 10000;
237 table = realloc(table, sizeof(*table) * table_size);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700238 if (!table) {
239 fprintf(stderr, "out of memory\n");
240 exit (1);
241 }
242 }
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800243 if (read_symbol(in, &table[table_cnt]) == 0) {
244 table[table_cnt].start_pos = table_cnt;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700245 table_cnt++;
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800246 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700247 }
248}
249
250static void output_label(char *label)
251{
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700252 if (symbol_prefix_char)
253 printf(".globl %c%s\n", symbol_prefix_char, label);
254 else
255 printf(".globl %s\n", label);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700256 printf("\tALGN\n");
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700257 if (symbol_prefix_char)
258 printf("%c%s:\n", symbol_prefix_char, label);
259 else
260 printf("%s:\n", label);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700261}
262
263/* uncompress a compressed symbol. When this function is called, the best table
264 * might still be compressed itself, so the function needs to be recursive */
265static int expand_symbol(unsigned char *data, int len, char *result)
266{
267 int c, rlen, total=0;
268
269 while (len) {
270 c = *data;
271 /* if the table holds a single char that is the same as the one
272 * we are looking for, then end the search */
273 if (best_table[c][0]==c && best_table_len[c]==1) {
274 *result++ = c;
275 total++;
276 } else {
277 /* if not, recurse and expand */
278 rlen = expand_symbol(best_table[c], best_table_len[c], result);
279 total += rlen;
280 result += rlen;
281 }
282 data++;
283 len--;
284 }
285 *result=0;
286
287 return total;
288}
289
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700290static void write_src(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700291{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700292 unsigned int i, k, off;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700293 unsigned int best_idx[256];
294 unsigned int *markers;
Tejun Heo9281ace2007-07-17 04:03:51 -0700295 char buf[KSYM_NAME_LEN];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700296
297 printf("#include <asm/types.h>\n");
298 printf("#if BITS_PER_LONG == 64\n");
299 printf("#define PTR .quad\n");
300 printf("#define ALGN .align 8\n");
301 printf("#else\n");
302 printf("#define PTR .long\n");
303 printf("#define ALGN .align 4\n");
304 printf("#endif\n");
305
Jan Beulichaad09472006-12-08 02:35:57 -0800306 printf("\t.section .rodata, \"a\"\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700307
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100308 /* Provide proper symbols relocatability by their '_text'
309 * relativeness. The symbol names cannot be used to construct
310 * normal symbol references as the list of symbols contains
311 * symbols that are declared static and are private to their
312 * .o files. This prevents .tmp_kallsyms.o or any other
313 * object from referencing them.
314 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700315 output_label("kallsyms_addresses");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700316 for (i = 0; i < table_cnt; i++) {
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100317 if (toupper(table[i].sym[0]) != 'A') {
Vivek Goyal2c22d8b2006-12-07 02:14:10 +0100318 if (_text <= table[i].addr)
319 printf("\tPTR\t_text + %#llx\n",
320 table[i].addr - _text);
321 else
322 printf("\tPTR\t_text - %#llx\n",
323 _text - table[i].addr);
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100324 } else {
325 printf("\tPTR\t%#llx\n", table[i].addr);
326 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700327 }
328 printf("\n");
329
330 output_label("kallsyms_num_syms");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700331 printf("\tPTR\t%d\n", table_cnt);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700332 printf("\n");
333
334 /* table of offset markers, that give the offset in the compressed stream
335 * every 256 symbols */
Jesper Juhlf1a136e2006-03-25 03:07:46 -0800336 markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
337 if (!markers) {
338 fprintf(stderr, "kallsyms failure: "
339 "unable to allocate required memory\n");
340 exit(EXIT_FAILURE);
341 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700342
343 output_label("kallsyms_names");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700344 off = 0;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700345 for (i = 0; i < table_cnt; i++) {
346 if ((i & 0xFF) == 0)
347 markers[i >> 8] = off;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700348
349 printf("\t.byte 0x%02x", table[i].len);
350 for (k = 0; k < table[i].len; k++)
351 printf(", 0x%02x", table[i].sym[k]);
352 printf("\n");
353
354 off += table[i].len + 1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700355 }
356 printf("\n");
357
358 output_label("kallsyms_markers");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700359 for (i = 0; i < ((table_cnt + 255) >> 8); i++)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700360 printf("\tPTR\t%d\n", markers[i]);
361 printf("\n");
362
363 free(markers);
364
365 output_label("kallsyms_token_table");
366 off = 0;
367 for (i = 0; i < 256; i++) {
368 best_idx[i] = off;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700369 expand_symbol(best_table[i], best_table_len[i], buf);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700370 printf("\t.asciz\t\"%s\"\n", buf);
371 off += strlen(buf) + 1;
372 }
373 printf("\n");
374
375 output_label("kallsyms_token_index");
376 for (i = 0; i < 256; i++)
377 printf("\t.short\t%d\n", best_idx[i]);
378 printf("\n");
379}
380
381
382/* table lookup compression functions */
383
Linus Torvalds1da177e2005-04-16 15:20:36 -0700384/* count all the possible tokens in a symbol */
385static void learn_symbol(unsigned char *symbol, int len)
386{
387 int i;
388
389 for (i = 0; i < len - 1; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700390 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700391}
392
393/* decrease the count for all the possible tokens in a symbol */
394static void forget_symbol(unsigned char *symbol, int len)
395{
396 int i;
397
398 for (i = 0; i < len - 1; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700399 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700400}
401
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700402/* remove all the invalid symbols from the table and do the initial token count */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700403static void build_initial_tok_table(void)
404{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700405 unsigned int i, pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700406
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700407 pos = 0;
408 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700409 if ( symbol_valid(&table[i]) ) {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700410 if (pos != i)
411 table[pos] = table[i];
412 learn_symbol(table[pos].sym, table[pos].len);
413 pos++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700414 }
415 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700416 table_cnt = pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700417}
418
Paulo Marques7c5d2492007-06-20 18:09:00 +0100419static void *find_token(unsigned char *str, int len, unsigned char *token)
420{
421 int i;
422
423 for (i = 0; i < len - 1; i++) {
424 if (str[i] == token[0] && str[i+1] == token[1])
425 return &str[i];
426 }
427 return NULL;
428}
429
Linus Torvalds1da177e2005-04-16 15:20:36 -0700430/* replace a given token in all the valid symbols. Use the sampled symbols
431 * to update the counts */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700432static void compress_symbols(unsigned char *str, int idx)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700433{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700434 unsigned int i, len, size;
435 unsigned char *p1, *p2;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700436
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700437 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700438
439 len = table[i].len;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700440 p1 = table[i].sym;
441
442 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100443 p2 = find_token(p1, len, str);
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700444 if (!p2) continue;
445
446 /* decrease the counts for this symbol's tokens */
447 forget_symbol(table[i].sym, len);
448
449 size = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700450
451 do {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700452 *p2 = idx;
453 p2++;
454 size -= (p2 - p1);
455 memmove(p2, p2 + 1, size);
456 p1 = p2;
457 len--;
458
459 if (size < 2) break;
460
Linus Torvalds1da177e2005-04-16 15:20:36 -0700461 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100462 p2 = find_token(p1, size, str);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700463
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700464 } while (p2);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700465
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700466 table[i].len = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700467
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700468 /* increase the counts for this symbol's new tokens */
469 learn_symbol(table[i].sym, len);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700470 }
471}
472
473/* search the token with the maximum profit */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700474static int find_best_token(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700475{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700476 int i, best, bestprofit;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700477
478 bestprofit=-10000;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700479 best = 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700480
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700481 for (i = 0; i < 0x10000; i++) {
482 if (token_profit[i] > bestprofit) {
483 best = i;
484 bestprofit = token_profit[i];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700485 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700486 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700487 return best;
488}
489
490/* this is the core of the algorithm: calculate the "best" table */
491static void optimize_result(void)
492{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700493 int i, best;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700494
495 /* using the '\0' symbol last allows compress_symbols to use standard
496 * fast string functions */
497 for (i = 255; i >= 0; i--) {
498
499 /* if this table slot is empty (it is not used by an actual
500 * original char code */
501 if (!best_table_len[i]) {
502
503 /* find the token with the breates profit value */
504 best = find_best_token();
505
506 /* place it in the "best" table */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700507 best_table_len[i] = 2;
508 best_table[i][0] = best & 0xFF;
509 best_table[i][1] = (best >> 8) & 0xFF;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700510
511 /* replace this token in all the valid symbols */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700512 compress_symbols(best_table[i], i);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700513 }
514 }
515}
516
517/* start by placing the symbols that are actually used on the table */
518static void insert_real_symbols_in_table(void)
519{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700520 unsigned int i, j, c;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700521
522 memset(best_table, 0, sizeof(best_table));
523 memset(best_table_len, 0, sizeof(best_table_len));
524
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700525 for (i = 0; i < table_cnt; i++) {
526 for (j = 0; j < table[i].len; j++) {
527 c = table[i].sym[j];
528 best_table[c][0]=c;
529 best_table_len[c]=1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700530 }
531 }
532}
533
534static void optimize_token_table(void)
535{
Linus Torvalds1da177e2005-04-16 15:20:36 -0700536 build_initial_tok_table();
537
538 insert_real_symbols_in_table();
539
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700540 /* When valid symbol is not registered, exit to error */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700541 if (!table_cnt) {
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700542 fprintf(stderr, "No valid symbol.\n");
543 exit(1);
544 }
545
Linus Torvalds1da177e2005-04-16 15:20:36 -0700546 optimize_result();
547}
548
Lai Jiangshanb478b7822009-03-13 15:10:26 +0800549/* guess for "linker script provide" symbol */
550static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
551{
552 const char *symbol = (char *)se->sym + 1;
553 int len = se->len - 1;
554
555 if (len < 8)
556 return 0;
557
558 if (symbol[0] != '_' || symbol[1] != '_')
559 return 0;
560
561 /* __start_XXXXX */
562 if (!memcmp(symbol + 2, "start_", 6))
563 return 1;
564
565 /* __stop_XXXXX */
566 if (!memcmp(symbol + 2, "stop_", 5))
567 return 1;
568
569 /* __end_XXXXX */
570 if (!memcmp(symbol + 2, "end_", 4))
571 return 1;
572
573 /* __XXXXX_start */
574 if (!memcmp(symbol + len - 6, "_start", 6))
575 return 1;
576
577 /* __XXXXX_end */
578 if (!memcmp(symbol + len - 4, "_end", 4))
579 return 1;
580
581 return 0;
582}
583
584static int prefix_underscores_count(const char *str)
585{
586 const char *tail = str;
587
588 while (*tail != '_')
589 tail++;
590
591 return tail - str;
592}
593
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800594static int compare_symbols(const void *a, const void *b)
595{
596 const struct sym_entry *sa;
597 const struct sym_entry *sb;
598 int wa, wb;
599
600 sa = a;
601 sb = b;
602
603 /* sort by address first */
604 if (sa->addr > sb->addr)
605 return 1;
606 if (sa->addr < sb->addr)
607 return -1;
608
609 /* sort by "weakness" type */
610 wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
611 wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
612 if (wa != wb)
613 return wa - wb;
614
Lai Jiangshanb478b7822009-03-13 15:10:26 +0800615 /* sort by "linker script provide" type */
616 wa = may_be_linker_script_provide_symbol(sa);
617 wb = may_be_linker_script_provide_symbol(sb);
618 if (wa != wb)
619 return wa - wb;
620
621 /* sort by the number of prefix underscores */
622 wa = prefix_underscores_count((const char *)sa->sym + 1);
623 wb = prefix_underscores_count((const char *)sb->sym + 1);
624 if (wa != wb)
625 return wa - wb;
626
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800627 /* sort by initial order, so that other symbols are left undisturbed */
628 return sa->start_pos - sb->start_pos;
629}
630
631static void sort_symbols(void)
632{
633 qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
634}
Linus Torvalds1da177e2005-04-16 15:20:36 -0700635
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700636int main(int argc, char **argv)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700637{
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700638 if (argc >= 2) {
639 int i;
640 for (i = 1; i < argc; i++) {
641 if(strcmp(argv[i], "--all-symbols") == 0)
642 all_symbols = 1;
643 else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
644 char *p = &argv[i][16];
645 /* skip quote */
646 if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
647 p++;
648 symbol_prefix_char = *p;
649 } else
650 usage();
651 }
652 } else if (argc != 1)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700653 usage();
654
655 read_map(stdin);
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100656 sort_symbols();
657 optimize_token_table();
Linus Torvalds1da177e2005-04-16 15:20:36 -0700658 write_src();
659
660 return 0;
661}