blob: 10085de886fef49b78a12746a2e0a593545d56e0 [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';
Ming Leif6537f22013-11-02 09:11:33 +103058static unsigned long long kernel_start_addr = 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -070059
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070060int token_profit[0x10000];
Linus Torvalds1da177e2005-04-16 15:20:36 -070061
62/* the table that holds the result of the compression */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070063unsigned char best_table[256][2];
Linus Torvalds1da177e2005-04-16 15:20:36 -070064unsigned char best_table_len[256];
65
66
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070067static void usage(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -070068{
Ming Leif6537f22013-11-02 09:11:33 +103069 fprintf(stderr, "Usage: kallsyms [--all-symbols] "
70 "[--symbol-prefix=<prefix char>] "
71 "[--page-offset=<CONFIG_PAGE_OFFSET>] "
72 "< in.map > out.S\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -070073 exit(1);
74}
75
76/*
77 * This ignores the intensely annoying "mapping symbols" found
78 * in ARM ELF files: $a, $t and $d.
79 */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070080static inline int is_arm_mapping_symbol(const char *str)
Linus Torvalds1da177e2005-04-16 15:20:36 -070081{
82 return str[0] == '$' && strchr("atd", str[1])
83 && (str[2] == '\0' || str[2] == '.');
84}
85
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040086static int read_symbol_tr(const char *sym, unsigned long long addr)
87{
88 size_t i;
89 struct text_range *tr;
90
91 for (i = 0; i < ARRAY_SIZE(text_ranges); ++i) {
92 tr = &text_ranges[i];
93
94 if (strcmp(sym, tr->stext) == 0) {
95 tr->start = addr;
96 return 0;
97 } else if (strcmp(sym, tr->etext) == 0) {
98 tr->end = addr;
99 return 0;
100 }
101 }
102
103 return 1;
104}
105
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700106static int read_symbol(FILE *in, struct sym_entry *s)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700107{
108 char str[500];
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700109 char *sym, stype;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700110 int rc;
111
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700112 rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, str);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700113 if (rc != 3) {
Jean Sacrenef894872010-09-10 23:13:33 -0600114 if (rc != EOF && fgets(str, 500, in) == NULL)
115 fprintf(stderr, "Read error or end of file.\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700116 return -1;
117 }
Andi Kleenf3462aa2013-10-23 15:07:53 +0200118 if (strlen(str) > KSYM_NAME_LEN) {
Fabio Estevam6f622592013-11-08 00:45:01 -0200119 fprintf(stderr, "Symbol %s too long for kallsyms (%zu vs %d).\n"
Andi Kleenf3462aa2013-10-23 15:07:53 +0200120 "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n",
121 str, strlen(str), KSYM_NAME_LEN);
122 return -1;
123 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700124
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700125 sym = str;
126 /* skip prefix char */
127 if (symbol_prefix_char && str[0] == symbol_prefix_char)
128 sym++;
129
Linus Torvalds1da177e2005-04-16 15:20:36 -0700130 /* Ignore most absolute/undefined (?) symbols. */
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100131 if (strcmp(sym, "_text") == 0)
132 _text = s->addr;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400133 else if (read_symbol_tr(sym, s->addr) == 0)
134 /* nothing to do */;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700135 else if (toupper(stype) == 'A')
Linus Torvalds1da177e2005-04-16 15:20:36 -0700136 {
137 /* Keep these useful absolute symbols */
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700138 if (strcmp(sym, "__kernel_syscall_via_break") &&
139 strcmp(sym, "__kernel_syscall_via_epc") &&
140 strcmp(sym, "__kernel_sigtramp") &&
141 strcmp(sym, "__gp"))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700142 return -1;
143
144 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700145 else if (toupper(stype) == 'U' ||
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700146 is_arm_mapping_symbol(sym))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700147 return -1;
Ralf Baechle6f00df22005-09-06 15:16:41 -0700148 /* exclude also MIPS ELF local symbols ($L123 instead of .L123) */
149 else if (str[0] == '$')
150 return -1;
Sam Ravnborgaab34ac2008-05-19 20:07:58 +0200151 /* exclude debugging symbols */
152 else if (stype == 'N')
153 return -1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700154
155 /* include the type field in the symbol name, so that it gets
156 * compressed together */
157 s->len = strlen(str) + 1;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700158 s->sym = malloc(s->len + 1);
Jesper Juhlf1a136e2006-03-25 03:07:46 -0800159 if (!s->sym) {
160 fprintf(stderr, "kallsyms failure: "
161 "unable to allocate required amount of memory\n");
162 exit(EXIT_FAILURE);
163 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700164 strcpy((char *)s->sym + 1, str);
165 s->sym[0] = stype;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700166
167 return 0;
168}
169
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400170static int symbol_valid_tr(struct sym_entry *s)
171{
172 size_t i;
173 struct text_range *tr;
174
175 for (i = 0; i < ARRAY_SIZE(text_ranges); ++i) {
176 tr = &text_ranges[i];
177
Mike Frysingerac6ca5c2009-06-15 07:52:48 -0400178 if (s->addr >= tr->start && s->addr <= tr->end)
179 return 1;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400180 }
181
Mike Frysingerac6ca5c2009-06-15 07:52:48 -0400182 return 0;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400183}
184
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700185static int symbol_valid(struct sym_entry *s)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700186{
187 /* Symbols which vary between passes. Passes 1 and 2 must have
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100188 * identical symbol lists. The kallsyms_* symbols below are only added
189 * after pass 1, they would be included in pass 2 when --all-symbols is
190 * specified so exclude them to get a stable symbol list.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700191 */
192 static char *special_symbols[] = {
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100193 "kallsyms_addresses",
194 "kallsyms_num_syms",
195 "kallsyms_names",
196 "kallsyms_markers",
197 "kallsyms_token_table",
198 "kallsyms_token_index",
199
Linus Torvalds1da177e2005-04-16 15:20:36 -0700200 /* Exclude linker generated symbols which vary between passes */
201 "_SDA_BASE_", /* ppc */
202 "_SDA2_BASE_", /* ppc */
203 NULL };
204 int i;
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700205 int offset = 1;
206
Ming Leif6537f22013-11-02 09:11:33 +1030207 if (s->addr < kernel_start_addr)
208 return 0;
209
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700210 /* skip prefix char */
211 if (symbol_prefix_char && *(s->sym + 1) == symbol_prefix_char)
212 offset++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700213
214 /* if --all-symbols is not specified, then symbols outside the text
215 * and inittext sections are discarded */
216 if (!all_symbols) {
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400217 if (symbol_valid_tr(s) == 0)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700218 return 0;
219 /* Corner case. Discard any symbols with the same value as
Robin Getza3b81112008-02-06 01:36:26 -0800220 * _etext _einittext; they can move between pass 1 and 2 when
221 * the kallsyms data are added. If these symbols move then
222 * they may get dropped in pass 2, which breaks the kallsyms
223 * rules.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700224 */
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400225 if ((s->addr == text_range_text->end &&
226 strcmp((char *)s->sym + offset, text_range_text->etext)) ||
227 (s->addr == text_range_inittext->end &&
228 strcmp((char *)s->sym + offset, text_range_inittext->etext)))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700229 return 0;
230 }
231
232 /* Exclude symbols which vary between passes. */
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100233 if (strstr((char *)s->sym + offset, "_compiled."))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700234 return 0;
235
236 for (i = 0; special_symbols[i]; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700237 if( strcmp((char *)s->sym + offset, special_symbols[i]) == 0 )
Linus Torvalds1da177e2005-04-16 15:20:36 -0700238 return 0;
239
240 return 1;
241}
242
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700243static void read_map(FILE *in)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700244{
245 while (!feof(in)) {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700246 if (table_cnt >= table_size) {
247 table_size += 10000;
248 table = realloc(table, sizeof(*table) * table_size);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700249 if (!table) {
250 fprintf(stderr, "out of memory\n");
251 exit (1);
252 }
253 }
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800254 if (read_symbol(in, &table[table_cnt]) == 0) {
255 table[table_cnt].start_pos = table_cnt;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700256 table_cnt++;
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800257 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700258 }
259}
260
261static void output_label(char *label)
262{
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700263 if (symbol_prefix_char)
264 printf(".globl %c%s\n", symbol_prefix_char, label);
265 else
266 printf(".globl %s\n", label);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700267 printf("\tALGN\n");
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700268 if (symbol_prefix_char)
269 printf("%c%s:\n", symbol_prefix_char, label);
270 else
271 printf("%s:\n", label);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700272}
273
274/* uncompress a compressed symbol. When this function is called, the best table
275 * might still be compressed itself, so the function needs to be recursive */
276static int expand_symbol(unsigned char *data, int len, char *result)
277{
278 int c, rlen, total=0;
279
280 while (len) {
281 c = *data;
282 /* if the table holds a single char that is the same as the one
283 * we are looking for, then end the search */
284 if (best_table[c][0]==c && best_table_len[c]==1) {
285 *result++ = c;
286 total++;
287 } else {
288 /* if not, recurse and expand */
289 rlen = expand_symbol(best_table[c], best_table_len[c], result);
290 total += rlen;
291 result += rlen;
292 }
293 data++;
294 len--;
295 }
296 *result=0;
297
298 return total;
299}
300
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700301static void write_src(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700302{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700303 unsigned int i, k, off;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700304 unsigned int best_idx[256];
305 unsigned int *markers;
Tejun Heo9281ace2007-07-17 04:03:51 -0700306 char buf[KSYM_NAME_LEN];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700307
308 printf("#include <asm/types.h>\n");
309 printf("#if BITS_PER_LONG == 64\n");
310 printf("#define PTR .quad\n");
311 printf("#define ALGN .align 8\n");
312 printf("#else\n");
313 printf("#define PTR .long\n");
314 printf("#define ALGN .align 4\n");
315 printf("#endif\n");
316
Jan Beulichaad09472006-12-08 02:35:57 -0800317 printf("\t.section .rodata, \"a\"\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700318
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100319 /* Provide proper symbols relocatability by their '_text'
320 * relativeness. The symbol names cannot be used to construct
321 * normal symbol references as the list of symbols contains
322 * symbols that are declared static and are private to their
323 * .o files. This prevents .tmp_kallsyms.o or any other
324 * object from referencing them.
325 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700326 output_label("kallsyms_addresses");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700327 for (i = 0; i < table_cnt; i++) {
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100328 if (toupper(table[i].sym[0]) != 'A') {
Vivek Goyal2c22d8b2006-12-07 02:14:10 +0100329 if (_text <= table[i].addr)
330 printf("\tPTR\t_text + %#llx\n",
331 table[i].addr - _text);
332 else
333 printf("\tPTR\t_text - %#llx\n",
334 _text - table[i].addr);
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100335 } else {
336 printf("\tPTR\t%#llx\n", table[i].addr);
337 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700338 }
339 printf("\n");
340
341 output_label("kallsyms_num_syms");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700342 printf("\tPTR\t%d\n", table_cnt);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700343 printf("\n");
344
345 /* table of offset markers, that give the offset in the compressed stream
346 * every 256 symbols */
Jesper Juhlf1a136e2006-03-25 03:07:46 -0800347 markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
348 if (!markers) {
349 fprintf(stderr, "kallsyms failure: "
350 "unable to allocate required memory\n");
351 exit(EXIT_FAILURE);
352 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700353
354 output_label("kallsyms_names");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700355 off = 0;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700356 for (i = 0; i < table_cnt; i++) {
357 if ((i & 0xFF) == 0)
358 markers[i >> 8] = off;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700359
360 printf("\t.byte 0x%02x", table[i].len);
361 for (k = 0; k < table[i].len; k++)
362 printf(", 0x%02x", table[i].sym[k]);
363 printf("\n");
364
365 off += table[i].len + 1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700366 }
367 printf("\n");
368
369 output_label("kallsyms_markers");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700370 for (i = 0; i < ((table_cnt + 255) >> 8); i++)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700371 printf("\tPTR\t%d\n", markers[i]);
372 printf("\n");
373
374 free(markers);
375
376 output_label("kallsyms_token_table");
377 off = 0;
378 for (i = 0; i < 256; i++) {
379 best_idx[i] = off;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700380 expand_symbol(best_table[i], best_table_len[i], buf);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700381 printf("\t.asciz\t\"%s\"\n", buf);
382 off += strlen(buf) + 1;
383 }
384 printf("\n");
385
386 output_label("kallsyms_token_index");
387 for (i = 0; i < 256; i++)
388 printf("\t.short\t%d\n", best_idx[i]);
389 printf("\n");
390}
391
392
393/* table lookup compression functions */
394
Linus Torvalds1da177e2005-04-16 15:20:36 -0700395/* count all the possible tokens in a symbol */
396static void learn_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
404/* decrease the count for all the possible tokens in a symbol */
405static void forget_symbol(unsigned char *symbol, int len)
406{
407 int i;
408
409 for (i = 0; i < len - 1; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700410 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700411}
412
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700413/* remove all the invalid symbols from the table and do the initial token count */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700414static void build_initial_tok_table(void)
415{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700416 unsigned int i, pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700417
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700418 pos = 0;
419 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700420 if ( symbol_valid(&table[i]) ) {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700421 if (pos != i)
422 table[pos] = table[i];
423 learn_symbol(table[pos].sym, table[pos].len);
424 pos++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700425 }
426 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700427 table_cnt = pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700428}
429
Paulo Marques7c5d2492007-06-20 18:09:00 +0100430static void *find_token(unsigned char *str, int len, unsigned char *token)
431{
432 int i;
433
434 for (i = 0; i < len - 1; i++) {
435 if (str[i] == token[0] && str[i+1] == token[1])
436 return &str[i];
437 }
438 return NULL;
439}
440
Linus Torvalds1da177e2005-04-16 15:20:36 -0700441/* replace a given token in all the valid symbols. Use the sampled symbols
442 * to update the counts */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700443static void compress_symbols(unsigned char *str, int idx)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700444{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700445 unsigned int i, len, size;
446 unsigned char *p1, *p2;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700447
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700448 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700449
450 len = table[i].len;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700451 p1 = table[i].sym;
452
453 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100454 p2 = find_token(p1, len, str);
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700455 if (!p2) continue;
456
457 /* decrease the counts for this symbol's tokens */
458 forget_symbol(table[i].sym, len);
459
460 size = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700461
462 do {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700463 *p2 = idx;
464 p2++;
465 size -= (p2 - p1);
466 memmove(p2, p2 + 1, size);
467 p1 = p2;
468 len--;
469
470 if (size < 2) break;
471
Linus Torvalds1da177e2005-04-16 15:20:36 -0700472 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100473 p2 = find_token(p1, size, str);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700474
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700475 } while (p2);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700476
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700477 table[i].len = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700478
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700479 /* increase the counts for this symbol's new tokens */
480 learn_symbol(table[i].sym, len);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700481 }
482}
483
484/* search the token with the maximum profit */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700485static int find_best_token(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700486{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700487 int i, best, bestprofit;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700488
489 bestprofit=-10000;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700490 best = 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700491
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700492 for (i = 0; i < 0x10000; i++) {
493 if (token_profit[i] > bestprofit) {
494 best = i;
495 bestprofit = token_profit[i];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700496 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700497 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700498 return best;
499}
500
501/* this is the core of the algorithm: calculate the "best" table */
502static void optimize_result(void)
503{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700504 int i, best;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700505
506 /* using the '\0' symbol last allows compress_symbols to use standard
507 * fast string functions */
508 for (i = 255; i >= 0; i--) {
509
510 /* if this table slot is empty (it is not used by an actual
511 * original char code */
512 if (!best_table_len[i]) {
513
514 /* find the token with the breates profit value */
515 best = find_best_token();
Xiaochen Wange0a04b12011-05-01 11:41:41 +0800516 if (token_profit[best] == 0)
517 break;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700518
519 /* place it in the "best" table */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700520 best_table_len[i] = 2;
521 best_table[i][0] = best & 0xFF;
522 best_table[i][1] = (best >> 8) & 0xFF;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700523
524 /* replace this token in all the valid symbols */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700525 compress_symbols(best_table[i], i);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700526 }
527 }
528}
529
530/* start by placing the symbols that are actually used on the table */
531static void insert_real_symbols_in_table(void)
532{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700533 unsigned int i, j, c;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700534
535 memset(best_table, 0, sizeof(best_table));
536 memset(best_table_len, 0, sizeof(best_table_len));
537
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700538 for (i = 0; i < table_cnt; i++) {
539 for (j = 0; j < table[i].len; j++) {
540 c = table[i].sym[j];
541 best_table[c][0]=c;
542 best_table_len[c]=1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700543 }
544 }
545}
546
547static void optimize_token_table(void)
548{
Linus Torvalds1da177e2005-04-16 15:20:36 -0700549 build_initial_tok_table();
550
551 insert_real_symbols_in_table();
552
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700553 /* When valid symbol is not registered, exit to error */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700554 if (!table_cnt) {
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700555 fprintf(stderr, "No valid symbol.\n");
556 exit(1);
557 }
558
Linus Torvalds1da177e2005-04-16 15:20:36 -0700559 optimize_result();
560}
561
Lai Jiangshanb478b782009-03-13 15:10:26 +0800562/* guess for "linker script provide" symbol */
563static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
564{
565 const char *symbol = (char *)se->sym + 1;
566 int len = se->len - 1;
567
568 if (len < 8)
569 return 0;
570
571 if (symbol[0] != '_' || symbol[1] != '_')
572 return 0;
573
574 /* __start_XXXXX */
575 if (!memcmp(symbol + 2, "start_", 6))
576 return 1;
577
578 /* __stop_XXXXX */
579 if (!memcmp(symbol + 2, "stop_", 5))
580 return 1;
581
582 /* __end_XXXXX */
583 if (!memcmp(symbol + 2, "end_", 4))
584 return 1;
585
586 /* __XXXXX_start */
587 if (!memcmp(symbol + len - 6, "_start", 6))
588 return 1;
589
590 /* __XXXXX_end */
591 if (!memcmp(symbol + len - 4, "_end", 4))
592 return 1;
593
594 return 0;
595}
596
597static int prefix_underscores_count(const char *str)
598{
599 const char *tail = str;
600
Paul Mundta9ece532009-09-22 16:44:12 -0700601 while (*tail == '_')
Lai Jiangshanb478b782009-03-13 15:10:26 +0800602 tail++;
603
604 return tail - str;
605}
606
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800607static int compare_symbols(const void *a, const void *b)
608{
609 const struct sym_entry *sa;
610 const struct sym_entry *sb;
611 int wa, wb;
612
613 sa = a;
614 sb = b;
615
616 /* sort by address first */
617 if (sa->addr > sb->addr)
618 return 1;
619 if (sa->addr < sb->addr)
620 return -1;
621
622 /* sort by "weakness" type */
623 wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
624 wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
625 if (wa != wb)
626 return wa - wb;
627
Lai Jiangshanb478b782009-03-13 15:10:26 +0800628 /* sort by "linker script provide" type */
629 wa = may_be_linker_script_provide_symbol(sa);
630 wb = may_be_linker_script_provide_symbol(sb);
631 if (wa != wb)
632 return wa - wb;
633
634 /* sort by the number of prefix underscores */
635 wa = prefix_underscores_count((const char *)sa->sym + 1);
636 wb = prefix_underscores_count((const char *)sb->sym + 1);
637 if (wa != wb)
638 return wa - wb;
639
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800640 /* sort by initial order, so that other symbols are left undisturbed */
641 return sa->start_pos - sb->start_pos;
642}
643
644static void sort_symbols(void)
645{
646 qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
647}
Linus Torvalds1da177e2005-04-16 15:20:36 -0700648
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700649int main(int argc, char **argv)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700650{
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700651 if (argc >= 2) {
652 int i;
653 for (i = 1; i < argc; i++) {
654 if(strcmp(argv[i], "--all-symbols") == 0)
655 all_symbols = 1;
656 else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
657 char *p = &argv[i][16];
658 /* skip quote */
659 if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
660 p++;
661 symbol_prefix_char = *p;
Ming Leif6537f22013-11-02 09:11:33 +1030662 } else if (strncmp(argv[i], "--page-offset=", 14) == 0) {
663 const char *p = &argv[i][14];
664 kernel_start_addr = strtoull(p, NULL, 16);
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700665 } else
666 usage();
667 }
668 } else if (argc != 1)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700669 usage();
670
671 read_map(stdin);
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100672 sort_symbols();
673 optimize_token_table();
Linus Torvalds1da177e2005-04-16 15:20:36 -0700674 write_src();
675
676 return 0;
677}