blob: f4b016782f0d9d0759a944f8baa6869dbc511bad [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[] = {
217 "_compiled.", /* gcc < 3.0: "gcc[0-9]_compiled." */
218 "_veneer", /* arm */
219 NULL };
220
Linus Torvalds1da177e2005-04-16 15:20:36 -0700221 int i;
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200222 char *sym_name = (char *)s->sym + 1;
223
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700224
Ming Leif6537f22013-11-02 09:11:33 +1030225 if (s->addr < kernel_start_addr)
226 return 0;
227
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700228 /* skip prefix char */
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200229 if (symbol_prefix_char && *sym_name == symbol_prefix_char)
230 sym_name++;
231
Linus Torvalds1da177e2005-04-16 15:20:36 -0700232
233 /* if --all-symbols is not specified, then symbols outside the text
234 * and inittext sections are discarded */
235 if (!all_symbols) {
Kees Cook78eb7152014-03-17 13:18:27 +1030236 if (symbol_in_range(s, text_ranges,
237 ARRAY_SIZE(text_ranges)) == 0)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700238 return 0;
239 /* Corner case. Discard any symbols with the same value as
Robin Getza3b81112008-02-06 01:36:26 -0800240 * _etext _einittext; they can move between pass 1 and 2 when
241 * the kallsyms data are added. If these symbols move then
242 * they may get dropped in pass 2, which breaks the kallsyms
243 * rules.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700244 */
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400245 if ((s->addr == text_range_text->end &&
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200246 strcmp(sym_name,
Kees Cook78eb7152014-03-17 13:18:27 +1030247 text_range_text->end_sym)) ||
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400248 (s->addr == text_range_inittext->end &&
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200249 strcmp(sym_name,
Kees Cook78eb7152014-03-17 13:18:27 +1030250 text_range_inittext->end_sym)))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700251 return 0;
252 }
253
254 /* Exclude symbols which vary between passes. */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700255 for (i = 0; special_symbols[i]; i++)
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200256 if (strcmp(sym_name, special_symbols[i]) == 0)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700257 return 0;
258
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200259 for (i = 0; special_suffixes[i]; i++) {
260 int l = strlen(sym_name) - strlen(special_suffixes[i]);
261
262 if (l >= 0 && strcmp(sym_name + l, special_suffixes[i]) == 0)
263 return 0;
264 }
265
Linus Torvalds1da177e2005-04-16 15:20:36 -0700266 return 1;
267}
268
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700269static void read_map(FILE *in)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700270{
271 while (!feof(in)) {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700272 if (table_cnt >= table_size) {
273 table_size += 10000;
274 table = realloc(table, sizeof(*table) * table_size);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700275 if (!table) {
276 fprintf(stderr, "out of memory\n");
277 exit (1);
278 }
279 }
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800280 if (read_symbol(in, &table[table_cnt]) == 0) {
281 table[table_cnt].start_pos = table_cnt;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700282 table_cnt++;
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800283 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700284 }
285}
286
287static void output_label(char *label)
288{
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700289 if (symbol_prefix_char)
290 printf(".globl %c%s\n", symbol_prefix_char, label);
291 else
292 printf(".globl %s\n", label);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700293 printf("\tALGN\n");
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700294 if (symbol_prefix_char)
295 printf("%c%s:\n", symbol_prefix_char, label);
296 else
297 printf("%s:\n", label);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700298}
299
300/* uncompress a compressed symbol. When this function is called, the best table
301 * might still be compressed itself, so the function needs to be recursive */
302static int expand_symbol(unsigned char *data, int len, char *result)
303{
304 int c, rlen, total=0;
305
306 while (len) {
307 c = *data;
308 /* if the table holds a single char that is the same as the one
309 * we are looking for, then end the search */
310 if (best_table[c][0]==c && best_table_len[c]==1) {
311 *result++ = c;
312 total++;
313 } else {
314 /* if not, recurse and expand */
315 rlen = expand_symbol(best_table[c], best_table_len[c], result);
316 total += rlen;
317 result += rlen;
318 }
319 data++;
320 len--;
321 }
322 *result=0;
323
324 return total;
325}
326
Kees Cook78eb7152014-03-17 13:18:27 +1030327static int symbol_absolute(struct sym_entry *s)
328{
329 return toupper(s->sym[0]) == 'A';
330}
331
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700332static void write_src(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700333{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700334 unsigned int i, k, off;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700335 unsigned int best_idx[256];
336 unsigned int *markers;
Tejun Heo9281ace2007-07-17 04:03:51 -0700337 char buf[KSYM_NAME_LEN];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700338
339 printf("#include <asm/types.h>\n");
340 printf("#if BITS_PER_LONG == 64\n");
341 printf("#define PTR .quad\n");
342 printf("#define ALGN .align 8\n");
343 printf("#else\n");
344 printf("#define PTR .long\n");
345 printf("#define ALGN .align 4\n");
346 printf("#endif\n");
347
Jan Beulichaad09472006-12-08 02:35:57 -0800348 printf("\t.section .rodata, \"a\"\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700349
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100350 /* Provide proper symbols relocatability by their '_text'
351 * relativeness. The symbol names cannot be used to construct
352 * normal symbol references as the list of symbols contains
353 * symbols that are declared static and are private to their
354 * .o files. This prevents .tmp_kallsyms.o or any other
355 * object from referencing them.
356 */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700357 output_label("kallsyms_addresses");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700358 for (i = 0; i < table_cnt; i++) {
Kees Cook78eb7152014-03-17 13:18:27 +1030359 if (!symbol_absolute(&table[i])) {
Vivek Goyal2c22d8b2006-12-07 02:14:10 +0100360 if (_text <= table[i].addr)
361 printf("\tPTR\t_text + %#llx\n",
362 table[i].addr - _text);
363 else
Andrew Morton2930ffc2014-03-10 15:49:48 -0700364 printf("\tPTR\t_text - %#llx\n",
365 _text - table[i].addr);
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100366 } else {
367 printf("\tPTR\t%#llx\n", table[i].addr);
368 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700369 }
370 printf("\n");
371
372 output_label("kallsyms_num_syms");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700373 printf("\tPTR\t%d\n", table_cnt);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700374 printf("\n");
375
376 /* table of offset markers, that give the offset in the compressed stream
377 * every 256 symbols */
Jesper Juhlf1a136e2006-03-25 03:07:46 -0800378 markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
379 if (!markers) {
380 fprintf(stderr, "kallsyms failure: "
381 "unable to allocate required memory\n");
382 exit(EXIT_FAILURE);
383 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700384
385 output_label("kallsyms_names");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700386 off = 0;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700387 for (i = 0; i < table_cnt; i++) {
388 if ((i & 0xFF) == 0)
389 markers[i >> 8] = off;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700390
391 printf("\t.byte 0x%02x", table[i].len);
392 for (k = 0; k < table[i].len; k++)
393 printf(", 0x%02x", table[i].sym[k]);
394 printf("\n");
395
396 off += table[i].len + 1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700397 }
398 printf("\n");
399
400 output_label("kallsyms_markers");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700401 for (i = 0; i < ((table_cnt + 255) >> 8); i++)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700402 printf("\tPTR\t%d\n", markers[i]);
403 printf("\n");
404
405 free(markers);
406
407 output_label("kallsyms_token_table");
408 off = 0;
409 for (i = 0; i < 256; i++) {
410 best_idx[i] = off;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700411 expand_symbol(best_table[i], best_table_len[i], buf);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700412 printf("\t.asciz\t\"%s\"\n", buf);
413 off += strlen(buf) + 1;
414 }
415 printf("\n");
416
417 output_label("kallsyms_token_index");
418 for (i = 0; i < 256; i++)
419 printf("\t.short\t%d\n", best_idx[i]);
420 printf("\n");
421}
422
423
424/* table lookup compression functions */
425
Linus Torvalds1da177e2005-04-16 15:20:36 -0700426/* count all the possible tokens in a symbol */
427static void learn_symbol(unsigned char *symbol, int len)
428{
429 int i;
430
431 for (i = 0; i < len - 1; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700432 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700433}
434
435/* decrease the count for all the possible tokens in a symbol */
436static void forget_symbol(unsigned char *symbol, int len)
437{
438 int i;
439
440 for (i = 0; i < len - 1; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700441 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700442}
443
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700444/* remove all the invalid symbols from the table and do the initial token count */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700445static void build_initial_tok_table(void)
446{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700447 unsigned int i, pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700448
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700449 pos = 0;
450 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700451 if ( symbol_valid(&table[i]) ) {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700452 if (pos != i)
453 table[pos] = table[i];
454 learn_symbol(table[pos].sym, table[pos].len);
455 pos++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700456 }
457 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700458 table_cnt = pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700459}
460
Paulo Marques7c5d2492007-06-20 18:09:00 +0100461static void *find_token(unsigned char *str, int len, unsigned char *token)
462{
463 int i;
464
465 for (i = 0; i < len - 1; i++) {
466 if (str[i] == token[0] && str[i+1] == token[1])
467 return &str[i];
468 }
469 return NULL;
470}
471
Linus Torvalds1da177e2005-04-16 15:20:36 -0700472/* replace a given token in all the valid symbols. Use the sampled symbols
473 * to update the counts */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700474static void compress_symbols(unsigned char *str, int idx)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700475{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700476 unsigned int i, len, size;
477 unsigned char *p1, *p2;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700478
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700479 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700480
481 len = table[i].len;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700482 p1 = table[i].sym;
483
484 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100485 p2 = find_token(p1, len, str);
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700486 if (!p2) continue;
487
488 /* decrease the counts for this symbol's tokens */
489 forget_symbol(table[i].sym, len);
490
491 size = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700492
493 do {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700494 *p2 = idx;
495 p2++;
496 size -= (p2 - p1);
497 memmove(p2, p2 + 1, size);
498 p1 = p2;
499 len--;
500
501 if (size < 2) break;
502
Linus Torvalds1da177e2005-04-16 15:20:36 -0700503 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100504 p2 = find_token(p1, size, str);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700505
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700506 } while (p2);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700507
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700508 table[i].len = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700509
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700510 /* increase the counts for this symbol's new tokens */
511 learn_symbol(table[i].sym, len);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700512 }
513}
514
515/* search the token with the maximum profit */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700516static int find_best_token(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700517{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700518 int i, best, bestprofit;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700519
520 bestprofit=-10000;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700521 best = 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700522
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700523 for (i = 0; i < 0x10000; i++) {
524 if (token_profit[i] > bestprofit) {
525 best = i;
526 bestprofit = token_profit[i];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700527 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700528 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700529 return best;
530}
531
532/* this is the core of the algorithm: calculate the "best" table */
533static void optimize_result(void)
534{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700535 int i, best;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700536
537 /* using the '\0' symbol last allows compress_symbols to use standard
538 * fast string functions */
539 for (i = 255; i >= 0; i--) {
540
541 /* if this table slot is empty (it is not used by an actual
542 * original char code */
543 if (!best_table_len[i]) {
544
545 /* find the token with the breates profit value */
546 best = find_best_token();
Xiaochen Wange0a04b12011-05-01 11:41:41 +0800547 if (token_profit[best] == 0)
548 break;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700549
550 /* place it in the "best" table */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700551 best_table_len[i] = 2;
552 best_table[i][0] = best & 0xFF;
553 best_table[i][1] = (best >> 8) & 0xFF;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700554
555 /* replace this token in all the valid symbols */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700556 compress_symbols(best_table[i], i);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700557 }
558 }
559}
560
561/* start by placing the symbols that are actually used on the table */
562static void insert_real_symbols_in_table(void)
563{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700564 unsigned int i, j, c;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700565
566 memset(best_table, 0, sizeof(best_table));
567 memset(best_table_len, 0, sizeof(best_table_len));
568
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700569 for (i = 0; i < table_cnt; i++) {
570 for (j = 0; j < table[i].len; j++) {
571 c = table[i].sym[j];
572 best_table[c][0]=c;
573 best_table_len[c]=1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700574 }
575 }
576}
577
578static void optimize_token_table(void)
579{
Linus Torvalds1da177e2005-04-16 15:20:36 -0700580 build_initial_tok_table();
581
582 insert_real_symbols_in_table();
583
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700584 /* When valid symbol is not registered, exit to error */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700585 if (!table_cnt) {
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700586 fprintf(stderr, "No valid symbol.\n");
587 exit(1);
588 }
589
Linus Torvalds1da177e2005-04-16 15:20:36 -0700590 optimize_result();
591}
592
Lai Jiangshanb478b7822009-03-13 15:10:26 +0800593/* guess for "linker script provide" symbol */
594static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
595{
596 const char *symbol = (char *)se->sym + 1;
597 int len = se->len - 1;
598
599 if (len < 8)
600 return 0;
601
602 if (symbol[0] != '_' || symbol[1] != '_')
603 return 0;
604
605 /* __start_XXXXX */
606 if (!memcmp(symbol + 2, "start_", 6))
607 return 1;
608
609 /* __stop_XXXXX */
610 if (!memcmp(symbol + 2, "stop_", 5))
611 return 1;
612
613 /* __end_XXXXX */
614 if (!memcmp(symbol + 2, "end_", 4))
615 return 1;
616
617 /* __XXXXX_start */
618 if (!memcmp(symbol + len - 6, "_start", 6))
619 return 1;
620
621 /* __XXXXX_end */
622 if (!memcmp(symbol + len - 4, "_end", 4))
623 return 1;
624
625 return 0;
626}
627
628static int prefix_underscores_count(const char *str)
629{
630 const char *tail = str;
631
Paul Mundta9ece532009-09-22 16:44:12 -0700632 while (*tail == '_')
Lai Jiangshanb478b7822009-03-13 15:10:26 +0800633 tail++;
634
635 return tail - str;
636}
637
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800638static int compare_symbols(const void *a, const void *b)
639{
640 const struct sym_entry *sa;
641 const struct sym_entry *sb;
642 int wa, wb;
643
644 sa = a;
645 sb = b;
646
647 /* sort by address first */
648 if (sa->addr > sb->addr)
649 return 1;
650 if (sa->addr < sb->addr)
651 return -1;
652
653 /* sort by "weakness" type */
654 wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
655 wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
656 if (wa != wb)
657 return wa - wb;
658
Lai Jiangshanb478b7822009-03-13 15:10:26 +0800659 /* sort by "linker script provide" type */
660 wa = may_be_linker_script_provide_symbol(sa);
661 wb = may_be_linker_script_provide_symbol(sb);
662 if (wa != wb)
663 return wa - wb;
664
665 /* sort by the number of prefix underscores */
666 wa = prefix_underscores_count((const char *)sa->sym + 1);
667 wb = prefix_underscores_count((const char *)sb->sym + 1);
668 if (wa != wb)
669 return wa - wb;
670
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800671 /* sort by initial order, so that other symbols are left undisturbed */
672 return sa->start_pos - sb->start_pos;
673}
674
675static void sort_symbols(void)
676{
677 qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
678}
Linus Torvalds1da177e2005-04-16 15:20:36 -0700679
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030680static void make_percpus_absolute(void)
681{
682 unsigned int i;
683
684 for (i = 0; i < table_cnt; i++)
685 if (symbol_in_range(&table[i], &percpu_range, 1))
686 table[i].sym[0] = 'A';
687}
688
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700689int main(int argc, char **argv)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700690{
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700691 if (argc >= 2) {
692 int i;
693 for (i = 1; i < argc; i++) {
694 if(strcmp(argv[i], "--all-symbols") == 0)
695 all_symbols = 1;
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030696 else if (strcmp(argv[i], "--absolute-percpu") == 0)
697 absolute_percpu = 1;
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700698 else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
699 char *p = &argv[i][16];
700 /* skip quote */
701 if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
702 p++;
703 symbol_prefix_char = *p;
Ming Leif6537f22013-11-02 09:11:33 +1030704 } else if (strncmp(argv[i], "--page-offset=", 14) == 0) {
705 const char *p = &argv[i][14];
706 kernel_start_addr = strtoull(p, NULL, 16);
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700707 } else
708 usage();
709 }
710 } else if (argc != 1)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700711 usage();
712
713 read_map(stdin);
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030714 if (absolute_percpu)
715 make_percpus_absolute();
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100716 sort_symbols();
717 optimize_token_table();
Linus Torvalds1da177e2005-04-16 15:20:36 -0700718 write_src();
719
720 return 0;
721}