blob: 8fa81e84e29510c053e5e52cefc5eb9e1da16887 [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
Kees Cook78eb7152014-03-17 13:18:27 +103039struct addr_range {
40 const char *start_sym, *end_sym;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040041 unsigned long long start, end;
42};
43
44static unsigned long long _text;
Kees Cook78eb7152014-03-17 13:18:27 +103045static struct addr_range text_ranges[] = {
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040046 { "_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
Rusty Russellc6bda7c2014-03-17 14:05:46 +103054static struct addr_range percpu_range = {
55 "__per_cpu_start", "__per_cpu_end", -1ULL, 0
56};
57
Linus Torvalds1da177e2005-04-16 15:20:36 -070058static struct sym_entry *table;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070059static unsigned int table_size, table_cnt;
Linus Torvalds1da177e2005-04-16 15:20:36 -070060static int all_symbols = 0;
Rusty Russellc6bda7c2014-03-17 14:05:46 +103061static int absolute_percpu = 0;
Yoshinori Sato41f11a42005-05-01 08:59:06 -070062static char symbol_prefix_char = '\0';
Ming Leif6537f22013-11-02 09:11:33 +103063static unsigned long long kernel_start_addr = 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -070064
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070065int token_profit[0x10000];
Linus Torvalds1da177e2005-04-16 15:20:36 -070066
67/* the table that holds the result of the compression */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070068unsigned char best_table[256][2];
Linus Torvalds1da177e2005-04-16 15:20:36 -070069unsigned char best_table_len[256];
70
71
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070072static void usage(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -070073{
Ming Leif6537f22013-11-02 09:11:33 +103074 fprintf(stderr, "Usage: kallsyms [--all-symbols] "
75 "[--symbol-prefix=<prefix char>] "
76 "[--page-offset=<CONFIG_PAGE_OFFSET>] "
77 "< in.map > out.S\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -070078 exit(1);
79}
80
81/*
82 * This ignores the intensely annoying "mapping symbols" found
83 * in ARM ELF files: $a, $t and $d.
84 */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070085static inline int is_arm_mapping_symbol(const char *str)
Linus Torvalds1da177e2005-04-16 15:20:36 -070086{
Kyle McMartin6c34f1f2014-09-16 22:37:18 +010087 return str[0] == '$' && strchr("axtd", str[1])
Linus Torvalds1da177e2005-04-16 15:20:36 -070088 && (str[2] == '\0' || str[2] == '.');
89}
90
Kees Cook78eb7152014-03-17 13:18:27 +103091static int check_symbol_range(const char *sym, unsigned long long addr,
92 struct addr_range *ranges, int entries)
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040093{
94 size_t i;
Kees Cook78eb7152014-03-17 13:18:27 +103095 struct addr_range *ar;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040096
Kees Cook78eb7152014-03-17 13:18:27 +103097 for (i = 0; i < entries; ++i) {
98 ar = &ranges[i];
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040099
Kees Cook78eb7152014-03-17 13:18:27 +1030100 if (strcmp(sym, ar->start_sym) == 0) {
101 ar->start = addr;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400102 return 0;
Kees Cook78eb7152014-03-17 13:18:27 +1030103 } else if (strcmp(sym, ar->end_sym) == 0) {
104 ar->end = addr;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400105 return 0;
106 }
107 }
108
109 return 1;
110}
111
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700112static int read_symbol(FILE *in, struct sym_entry *s)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700113{
114 char str[500];
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700115 char *sym, stype;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700116 int rc;
117
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700118 rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, str);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700119 if (rc != 3) {
Jean Sacrenef894872010-09-10 23:13:33 -0600120 if (rc != EOF && fgets(str, 500, in) == NULL)
121 fprintf(stderr, "Read error or end of file.\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700122 return -1;
123 }
Andi Kleenf3462aa2013-10-23 15:07:53 +0200124 if (strlen(str) > KSYM_NAME_LEN) {
Fabio Estevam6f622592013-11-08 00:45:01 -0200125 fprintf(stderr, "Symbol %s too long for kallsyms (%zu vs %d).\n"
Masahiro Yamadabb66fc62014-06-10 19:08:13 +0900126 "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n",
Andi Kleenf3462aa2013-10-23 15:07:53 +0200127 str, strlen(str), KSYM_NAME_LEN);
128 return -1;
129 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700130
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700131 sym = str;
132 /* skip prefix char */
133 if (symbol_prefix_char && str[0] == symbol_prefix_char)
134 sym++;
135
Linus Torvalds1da177e2005-04-16 15:20:36 -0700136 /* Ignore most absolute/undefined (?) symbols. */
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100137 if (strcmp(sym, "_text") == 0)
138 _text = s->addr;
Kees Cook78eb7152014-03-17 13:18:27 +1030139 else if (check_symbol_range(sym, s->addr, text_ranges,
140 ARRAY_SIZE(text_ranges)) == 0)
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400141 /* nothing to do */;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700142 else if (toupper(stype) == 'A')
Linus Torvalds1da177e2005-04-16 15:20:36 -0700143 {
144 /* Keep these useful absolute symbols */
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700145 if (strcmp(sym, "__kernel_syscall_via_break") &&
146 strcmp(sym, "__kernel_syscall_via_epc") &&
147 strcmp(sym, "__kernel_sigtramp") &&
148 strcmp(sym, "__gp"))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700149 return -1;
150
151 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700152 else if (toupper(stype) == 'U' ||
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700153 is_arm_mapping_symbol(sym))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700154 return -1;
Ralf Baechle6f00df22005-09-06 15:16:41 -0700155 /* exclude also MIPS ELF local symbols ($L123 instead of .L123) */
156 else if (str[0] == '$')
157 return -1;
Sam Ravnborgaab34ac2008-05-19 20:07:58 +0200158 /* exclude debugging symbols */
159 else if (stype == 'N')
160 return -1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700161
162 /* include the type field in the symbol name, so that it gets
163 * compressed together */
164 s->len = strlen(str) + 1;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700165 s->sym = malloc(s->len + 1);
Jesper Juhlf1a136e2006-03-25 03:07:46 -0800166 if (!s->sym) {
167 fprintf(stderr, "kallsyms failure: "
168 "unable to allocate required amount of memory\n");
169 exit(EXIT_FAILURE);
170 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700171 strcpy((char *)s->sym + 1, str);
172 s->sym[0] = stype;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700173
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030174 /* Record if we've found __per_cpu_start/end. */
175 check_symbol_range(sym, s->addr, &percpu_range, 1);
176
Linus Torvalds1da177e2005-04-16 15:20:36 -0700177 return 0;
178}
179
Kees Cook78eb7152014-03-17 13:18:27 +1030180static int symbol_in_range(struct sym_entry *s, struct addr_range *ranges,
181 int entries)
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400182{
183 size_t i;
Kees Cook78eb7152014-03-17 13:18:27 +1030184 struct addr_range *ar;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400185
Kees Cook78eb7152014-03-17 13:18:27 +1030186 for (i = 0; i < entries; ++i) {
187 ar = &ranges[i];
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400188
Kees Cook78eb7152014-03-17 13:18:27 +1030189 if (s->addr >= ar->start && s->addr <= ar->end)
Mike Frysingerac6ca5c2009-06-15 07:52:48 -0400190 return 1;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400191 }
192
Mike Frysingerac6ca5c2009-06-15 07:52:48 -0400193 return 0;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400194}
195
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700196static int symbol_valid(struct sym_entry *s)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700197{
198 /* Symbols which vary between passes. Passes 1 and 2 must have
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100199 * identical symbol lists. The kallsyms_* symbols below are only added
200 * after pass 1, they would be included in pass 2 when --all-symbols is
201 * specified so exclude them to get a stable symbol list.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700202 */
203 static char *special_symbols[] = {
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100204 "kallsyms_addresses",
205 "kallsyms_num_syms",
206 "kallsyms_names",
207 "kallsyms_markers",
208 "kallsyms_token_table",
209 "kallsyms_token_index",
210
Linus Torvalds1da177e2005-04-16 15:20:36 -0700211 /* Exclude linker generated symbols which vary between passes */
212 "_SDA_BASE_", /* ppc */
213 "_SDA2_BASE_", /* ppc */
214 NULL };
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200215
216 static char *special_suffixes[] = {
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200217 "_veneer", /* arm */
218 NULL };
219
Linus Torvalds1da177e2005-04-16 15:20:36 -0700220 int i;
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200221 char *sym_name = (char *)s->sym + 1;
222
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700223
Ming Leif6537f22013-11-02 09:11:33 +1030224 if (s->addr < kernel_start_addr)
225 return 0;
226
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700227 /* skip prefix char */
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200228 if (symbol_prefix_char && *sym_name == symbol_prefix_char)
229 sym_name++;
230
Linus Torvalds1da177e2005-04-16 15:20:36 -0700231
232 /* if --all-symbols is not specified, then symbols outside the text
233 * and inittext sections are discarded */
234 if (!all_symbols) {
Kees Cook78eb7152014-03-17 13:18:27 +1030235 if (symbol_in_range(s, text_ranges,
236 ARRAY_SIZE(text_ranges)) == 0)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700237 return 0;
238 /* Corner case. Discard any symbols with the same value as
Robin Getza3b81112008-02-06 01:36:26 -0800239 * _etext _einittext; they can move between pass 1 and 2 when
240 * the kallsyms data are added. If these symbols move then
241 * they may get dropped in pass 2, which breaks the kallsyms
242 * rules.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700243 */
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400244 if ((s->addr == text_range_text->end &&
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200245 strcmp(sym_name,
Kees Cook78eb7152014-03-17 13:18:27 +1030246 text_range_text->end_sym)) ||
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400247 (s->addr == text_range_inittext->end &&
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200248 strcmp(sym_name,
Kees Cook78eb7152014-03-17 13:18:27 +1030249 text_range_inittext->end_sym)))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700250 return 0;
251 }
252
253 /* Exclude symbols which vary between passes. */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700254 for (i = 0; special_symbols[i]; i++)
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200255 if (strcmp(sym_name, special_symbols[i]) == 0)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700256 return 0;
257
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200258 for (i = 0; special_suffixes[i]; i++) {
259 int l = strlen(sym_name) - strlen(special_suffixes[i]);
260
261 if (l >= 0 && strcmp(sym_name + l, special_suffixes[i]) == 0)
262 return 0;
263 }
264
Linus Torvalds1da177e2005-04-16 15:20:36 -0700265 return 1;
266}
267
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700268static void read_map(FILE *in)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700269{
270 while (!feof(in)) {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700271 if (table_cnt >= table_size) {
272 table_size += 10000;
273 table = realloc(table, sizeof(*table) * table_size);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700274 if (!table) {
275 fprintf(stderr, "out of memory\n");
276 exit (1);
277 }
278 }
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800279 if (read_symbol(in, &table[table_cnt]) == 0) {
280 table[table_cnt].start_pos = table_cnt;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700281 table_cnt++;
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800282 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700283 }
284}
285
286static void output_label(char *label)
287{
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700288 if (symbol_prefix_char)
289 printf(".globl %c%s\n", symbol_prefix_char, label);
290 else
291 printf(".globl %s\n", label);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700292 printf("\tALGN\n");
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700293 if (symbol_prefix_char)
294 printf("%c%s:\n", symbol_prefix_char, label);
295 else
296 printf("%s:\n", label);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700297}
298
299/* uncompress a compressed symbol. When this function is called, the best table
300 * might still be compressed itself, so the function needs to be recursive */
301static int expand_symbol(unsigned char *data, int len, char *result)
302{
303 int c, rlen, total=0;
304
305 while (len) {
306 c = *data;
307 /* if the table holds a single char that is the same as the one
308 * we are looking for, then end the search */
309 if (best_table[c][0]==c && best_table_len[c]==1) {
310 *result++ = c;
311 total++;
312 } else {
313 /* if not, recurse and expand */
314 rlen = expand_symbol(best_table[c], best_table_len[c], result);
315 total += rlen;
316 result += rlen;
317 }
318 data++;
319 len--;
320 }
321 *result=0;
322
323 return total;
324}
325
Kees Cook78eb7152014-03-17 13:18:27 +1030326static int symbol_absolute(struct sym_entry *s)
327{
328 return toupper(s->sym[0]) == 'A';
329}
330
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700331static void write_src(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700332{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700333 unsigned int i, k, off;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700334 unsigned int best_idx[256];
335 unsigned int *markers;
Tejun Heo9281ace2007-07-17 04:03:51 -0700336 char buf[KSYM_NAME_LEN];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700337
338 printf("#include <asm/types.h>\n");
339 printf("#if BITS_PER_LONG == 64\n");
340 printf("#define PTR .quad\n");
341 printf("#define ALGN .align 8\n");
342 printf("#else\n");
343 printf("#define PTR .long\n");
344 printf("#define ALGN .align 4\n");
345 printf("#endif\n");
346
Jan Beulichaad09472006-12-08 02:35:57 -0800347 printf("\t.section .rodata, \"a\"\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700348
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100349 /* Provide proper symbols relocatability by their '_text'
350 * relativeness. The symbol names cannot be used to construct
351 * normal symbol references as the list of symbols contains
352 * symbols that are declared static and are private to their
353 * .o files. This prevents .tmp_kallsyms.o or any other
354 * object from referencing them.
355 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700356 output_label("kallsyms_addresses");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700357 for (i = 0; i < table_cnt; i++) {
Kees Cook78eb7152014-03-17 13:18:27 +1030358 if (!symbol_absolute(&table[i])) {
Vivek Goyal2c22d8b2006-12-07 02:14:10 +0100359 if (_text <= table[i].addr)
360 printf("\tPTR\t_text + %#llx\n",
361 table[i].addr - _text);
362 else
Andrew Morton2930ffc2014-03-10 15:49:48 -0700363 printf("\tPTR\t_text - %#llx\n",
364 _text - table[i].addr);
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100365 } else {
366 printf("\tPTR\t%#llx\n", table[i].addr);
367 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700368 }
369 printf("\n");
370
371 output_label("kallsyms_num_syms");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700372 printf("\tPTR\t%d\n", table_cnt);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700373 printf("\n");
374
375 /* table of offset markers, that give the offset in the compressed stream
376 * every 256 symbols */
Jesper Juhlf1a136e2006-03-25 03:07:46 -0800377 markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
378 if (!markers) {
379 fprintf(stderr, "kallsyms failure: "
380 "unable to allocate required memory\n");
381 exit(EXIT_FAILURE);
382 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700383
384 output_label("kallsyms_names");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700385 off = 0;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700386 for (i = 0; i < table_cnt; i++) {
387 if ((i & 0xFF) == 0)
388 markers[i >> 8] = off;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700389
390 printf("\t.byte 0x%02x", table[i].len);
391 for (k = 0; k < table[i].len; k++)
392 printf(", 0x%02x", table[i].sym[k]);
393 printf("\n");
394
395 off += table[i].len + 1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700396 }
397 printf("\n");
398
399 output_label("kallsyms_markers");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700400 for (i = 0; i < ((table_cnt + 255) >> 8); i++)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700401 printf("\tPTR\t%d\n", markers[i]);
402 printf("\n");
403
404 free(markers);
405
406 output_label("kallsyms_token_table");
407 off = 0;
408 for (i = 0; i < 256; i++) {
409 best_idx[i] = off;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700410 expand_symbol(best_table[i], best_table_len[i], buf);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700411 printf("\t.asciz\t\"%s\"\n", buf);
412 off += strlen(buf) + 1;
413 }
414 printf("\n");
415
416 output_label("kallsyms_token_index");
417 for (i = 0; i < 256; i++)
418 printf("\t.short\t%d\n", best_idx[i]);
419 printf("\n");
420}
421
422
423/* table lookup compression functions */
424
Linus Torvalds1da177e2005-04-16 15:20:36 -0700425/* count all the possible tokens in a symbol */
426static void learn_symbol(unsigned char *symbol, int len)
427{
428 int i;
429
430 for (i = 0; i < len - 1; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700431 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700432}
433
434/* decrease the count for all the possible tokens in a symbol */
435static void forget_symbol(unsigned char *symbol, int len)
436{
437 int i;
438
439 for (i = 0; i < len - 1; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700440 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700441}
442
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700443/* remove all the invalid symbols from the table and do the initial token count */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700444static void build_initial_tok_table(void)
445{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700446 unsigned int i, pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700447
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700448 pos = 0;
449 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700450 if ( symbol_valid(&table[i]) ) {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700451 if (pos != i)
452 table[pos] = table[i];
453 learn_symbol(table[pos].sym, table[pos].len);
454 pos++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700455 }
456 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700457 table_cnt = pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700458}
459
Paulo Marques7c5d2492007-06-20 18:09:00 +0100460static void *find_token(unsigned char *str, int len, unsigned char *token)
461{
462 int i;
463
464 for (i = 0; i < len - 1; i++) {
465 if (str[i] == token[0] && str[i+1] == token[1])
466 return &str[i];
467 }
468 return NULL;
469}
470
Linus Torvalds1da177e2005-04-16 15:20:36 -0700471/* replace a given token in all the valid symbols. Use the sampled symbols
472 * to update the counts */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700473static void compress_symbols(unsigned char *str, int idx)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700474{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700475 unsigned int i, len, size;
476 unsigned char *p1, *p2;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700477
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700478 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700479
480 len = table[i].len;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700481 p1 = table[i].sym;
482
483 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100484 p2 = find_token(p1, len, str);
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700485 if (!p2) continue;
486
487 /* decrease the counts for this symbol's tokens */
488 forget_symbol(table[i].sym, len);
489
490 size = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700491
492 do {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700493 *p2 = idx;
494 p2++;
495 size -= (p2 - p1);
496 memmove(p2, p2 + 1, size);
497 p1 = p2;
498 len--;
499
500 if (size < 2) break;
501
Linus Torvalds1da177e2005-04-16 15:20:36 -0700502 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100503 p2 = find_token(p1, size, str);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700504
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700505 } while (p2);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700506
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700507 table[i].len = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700508
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700509 /* increase the counts for this symbol's new tokens */
510 learn_symbol(table[i].sym, len);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700511 }
512}
513
514/* search the token with the maximum profit */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700515static int find_best_token(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700516{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700517 int i, best, bestprofit;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700518
519 bestprofit=-10000;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700520 best = 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700521
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700522 for (i = 0; i < 0x10000; i++) {
523 if (token_profit[i] > bestprofit) {
524 best = i;
525 bestprofit = token_profit[i];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700526 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700527 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700528 return best;
529}
530
531/* this is the core of the algorithm: calculate the "best" table */
532static void optimize_result(void)
533{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700534 int i, best;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700535
536 /* using the '\0' symbol last allows compress_symbols to use standard
537 * fast string functions */
538 for (i = 255; i >= 0; i--) {
539
540 /* if this table slot is empty (it is not used by an actual
541 * original char code */
542 if (!best_table_len[i]) {
543
544 /* find the token with the breates profit value */
545 best = find_best_token();
Xiaochen Wange0a04b12011-05-01 11:41:41 +0800546 if (token_profit[best] == 0)
547 break;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700548
549 /* place it in the "best" table */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700550 best_table_len[i] = 2;
551 best_table[i][0] = best & 0xFF;
552 best_table[i][1] = (best >> 8) & 0xFF;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700553
554 /* replace this token in all the valid symbols */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700555 compress_symbols(best_table[i], i);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700556 }
557 }
558}
559
560/* start by placing the symbols that are actually used on the table */
561static void insert_real_symbols_in_table(void)
562{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700563 unsigned int i, j, c;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700564
565 memset(best_table, 0, sizeof(best_table));
566 memset(best_table_len, 0, sizeof(best_table_len));
567
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700568 for (i = 0; i < table_cnt; i++) {
569 for (j = 0; j < table[i].len; j++) {
570 c = table[i].sym[j];
571 best_table[c][0]=c;
572 best_table_len[c]=1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700573 }
574 }
575}
576
577static void optimize_token_table(void)
578{
Linus Torvalds1da177e2005-04-16 15:20:36 -0700579 build_initial_tok_table();
580
581 insert_real_symbols_in_table();
582
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700583 /* When valid symbol is not registered, exit to error */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700584 if (!table_cnt) {
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700585 fprintf(stderr, "No valid symbol.\n");
586 exit(1);
587 }
588
Linus Torvalds1da177e2005-04-16 15:20:36 -0700589 optimize_result();
590}
591
Lai Jiangshanb478b7822009-03-13 15:10:26 +0800592/* guess for "linker script provide" symbol */
593static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
594{
595 const char *symbol = (char *)se->sym + 1;
596 int len = se->len - 1;
597
598 if (len < 8)
599 return 0;
600
601 if (symbol[0] != '_' || symbol[1] != '_')
602 return 0;
603
604 /* __start_XXXXX */
605 if (!memcmp(symbol + 2, "start_", 6))
606 return 1;
607
608 /* __stop_XXXXX */
609 if (!memcmp(symbol + 2, "stop_", 5))
610 return 1;
611
612 /* __end_XXXXX */
613 if (!memcmp(symbol + 2, "end_", 4))
614 return 1;
615
616 /* __XXXXX_start */
617 if (!memcmp(symbol + len - 6, "_start", 6))
618 return 1;
619
620 /* __XXXXX_end */
621 if (!memcmp(symbol + len - 4, "_end", 4))
622 return 1;
623
624 return 0;
625}
626
627static int prefix_underscores_count(const char *str)
628{
629 const char *tail = str;
630
Paul Mundta9ece532009-09-22 16:44:12 -0700631 while (*tail == '_')
Lai Jiangshanb478b7822009-03-13 15:10:26 +0800632 tail++;
633
634 return tail - str;
635}
636
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800637static int compare_symbols(const void *a, const void *b)
638{
639 const struct sym_entry *sa;
640 const struct sym_entry *sb;
641 int wa, wb;
642
643 sa = a;
644 sb = b;
645
646 /* sort by address first */
647 if (sa->addr > sb->addr)
648 return 1;
649 if (sa->addr < sb->addr)
650 return -1;
651
652 /* sort by "weakness" type */
653 wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
654 wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
655 if (wa != wb)
656 return wa - wb;
657
Lai Jiangshanb478b7822009-03-13 15:10:26 +0800658 /* sort by "linker script provide" type */
659 wa = may_be_linker_script_provide_symbol(sa);
660 wb = may_be_linker_script_provide_symbol(sb);
661 if (wa != wb)
662 return wa - wb;
663
664 /* sort by the number of prefix underscores */
665 wa = prefix_underscores_count((const char *)sa->sym + 1);
666 wb = prefix_underscores_count((const char *)sb->sym + 1);
667 if (wa != wb)
668 return wa - wb;
669
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800670 /* sort by initial order, so that other symbols are left undisturbed */
671 return sa->start_pos - sb->start_pos;
672}
673
674static void sort_symbols(void)
675{
676 qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
677}
Linus Torvalds1da177e2005-04-16 15:20:36 -0700678
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030679static void make_percpus_absolute(void)
680{
681 unsigned int i;
682
683 for (i = 0; i < table_cnt; i++)
684 if (symbol_in_range(&table[i], &percpu_range, 1))
685 table[i].sym[0] = 'A';
686}
687
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700688int main(int argc, char **argv)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700689{
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700690 if (argc >= 2) {
691 int i;
692 for (i = 1; i < argc; i++) {
693 if(strcmp(argv[i], "--all-symbols") == 0)
694 all_symbols = 1;
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030695 else if (strcmp(argv[i], "--absolute-percpu") == 0)
696 absolute_percpu = 1;
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700697 else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
698 char *p = &argv[i][16];
699 /* skip quote */
700 if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
701 p++;
702 symbol_prefix_char = *p;
Ming Leif6537f22013-11-02 09:11:33 +1030703 } else if (strncmp(argv[i], "--page-offset=", 14) == 0) {
704 const char *p = &argv[i][14];
705 kernel_start_addr = strtoull(p, NULL, 16);
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700706 } else
707 usage();
708 }
709 } else if (argc != 1)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700710 usage();
711
712 read_map(stdin);
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030713 if (absolute_percpu)
714 make_percpus_absolute();
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100715 sort_symbols();
716 optimize_token_table();
Linus Torvalds1da177e2005-04-16 15:20:36 -0700717 write_src();
718
719 return 0;
720}