blob: 638b143ee60f4246cc16d67936bc62095a66c646 [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>
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -070025#include <limits.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -070026
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040027#ifndef ARRAY_SIZE
28#define ARRAY_SIZE(arr) (sizeof(arr) / sizeof(arr[0]))
29#endif
30
Tejun Heo9281ace2007-07-17 04:03:51 -070031#define KSYM_NAME_LEN 128
Linus Torvalds1da177e2005-04-16 15:20:36 -070032
Linus Torvalds1da177e2005-04-16 15:20:36 -070033struct sym_entry {
34 unsigned long long addr;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070035 unsigned int len;
Paulo Marquesf2df3f62008-02-06 01:37:33 -080036 unsigned int start_pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -070037 unsigned char *sym;
Ard Biesheuvel8c996942016-03-15 14:58:15 -070038 unsigned int percpu_absolute;
Linus Torvalds1da177e2005-04-16 15:20:36 -070039};
40
Kees Cook78eb7152014-03-17 13:18:27 +103041struct addr_range {
42 const char *start_sym, *end_sym;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040043 unsigned long long start, end;
44};
45
46static unsigned long long _text;
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -070047static unsigned long long relative_base;
Kees Cook78eb7152014-03-17 13:18:27 +103048static struct addr_range text_ranges[] = {
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040049 { "_stext", "_etext" },
50 { "_sinittext", "_einittext" },
51 { "_stext_l1", "_etext_l1" }, /* Blackfin on-chip L1 inst SRAM */
52 { "_stext_l2", "_etext_l2" }, /* Blackfin on-chip L2 SRAM */
53};
54#define text_range_text (&text_ranges[0])
55#define text_range_inittext (&text_ranges[1])
56
Rusty Russellc6bda7c2014-03-17 14:05:46 +103057static struct addr_range percpu_range = {
58 "__per_cpu_start", "__per_cpu_end", -1ULL, 0
59};
60
Linus Torvalds1da177e2005-04-16 15:20:36 -070061static struct sym_entry *table;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070062static unsigned int table_size, table_cnt;
Linus Torvalds1da177e2005-04-16 15:20:36 -070063static int all_symbols = 0;
Rusty Russellc6bda7c2014-03-17 14:05:46 +103064static int absolute_percpu = 0;
Yoshinori Sato41f11a42005-05-01 08:59:06 -070065static char symbol_prefix_char = '\0';
Ming Leif6537f22013-11-02 09:11:33 +103066static unsigned long long kernel_start_addr = 0;
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -070067static int base_relative = 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -070068
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070069int token_profit[0x10000];
Linus Torvalds1da177e2005-04-16 15:20:36 -070070
71/* the table that holds the result of the compression */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070072unsigned char best_table[256][2];
Linus Torvalds1da177e2005-04-16 15:20:36 -070073unsigned char best_table_len[256];
74
75
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070076static void usage(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -070077{
Ming Leif6537f22013-11-02 09:11:33 +103078 fprintf(stderr, "Usage: kallsyms [--all-symbols] "
79 "[--symbol-prefix=<prefix char>] "
80 "[--page-offset=<CONFIG_PAGE_OFFSET>] "
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -070081 "[--base-relative] < in.map > out.S\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -070082 exit(1);
83}
84
85/*
86 * This ignores the intensely annoying "mapping symbols" found
87 * in ARM ELF files: $a, $t and $d.
88 */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070089static inline int is_arm_mapping_symbol(const char *str)
Linus Torvalds1da177e2005-04-16 15:20:36 -070090{
Kyle McMartin6c34f1f2014-09-16 22:37:18 +010091 return str[0] == '$' && strchr("axtd", str[1])
Linus Torvalds1da177e2005-04-16 15:20:36 -070092 && (str[2] == '\0' || str[2] == '.');
93}
94
Kees Cook78eb7152014-03-17 13:18:27 +103095static int check_symbol_range(const char *sym, unsigned long long addr,
96 struct addr_range *ranges, int entries)
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040097{
98 size_t i;
Kees Cook78eb7152014-03-17 13:18:27 +103099 struct addr_range *ar;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400100
Kees Cook78eb7152014-03-17 13:18:27 +1030101 for (i = 0; i < entries; ++i) {
102 ar = &ranges[i];
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400103
Kees Cook78eb7152014-03-17 13:18:27 +1030104 if (strcmp(sym, ar->start_sym) == 0) {
105 ar->start = addr;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400106 return 0;
Kees Cook78eb7152014-03-17 13:18:27 +1030107 } else if (strcmp(sym, ar->end_sym) == 0) {
108 ar->end = addr;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400109 return 0;
110 }
111 }
112
113 return 1;
114}
115
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700116static int read_symbol(FILE *in, struct sym_entry *s)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700117{
118 char str[500];
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700119 char *sym, stype;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700120 int rc;
121
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700122 rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, str);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700123 if (rc != 3) {
Jean Sacrenef894872010-09-10 23:13:33 -0600124 if (rc != EOF && fgets(str, 500, in) == NULL)
125 fprintf(stderr, "Read error or end of file.\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700126 return -1;
127 }
Andi Kleenf3462aa2013-10-23 15:07:53 +0200128 if (strlen(str) > KSYM_NAME_LEN) {
Fabio Estevam6f622592013-11-08 00:45:01 -0200129 fprintf(stderr, "Symbol %s too long for kallsyms (%zu vs %d).\n"
Masahiro Yamadabb66fc62014-06-10 19:08:13 +0900130 "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n",
Andi Kleenf3462aa2013-10-23 15:07:53 +0200131 str, strlen(str), KSYM_NAME_LEN);
132 return -1;
133 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700134
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700135 sym = str;
136 /* skip prefix char */
137 if (symbol_prefix_char && str[0] == symbol_prefix_char)
138 sym++;
139
Linus Torvalds1da177e2005-04-16 15:20:36 -0700140 /* Ignore most absolute/undefined (?) symbols. */
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100141 if (strcmp(sym, "_text") == 0)
142 _text = s->addr;
Kees Cook78eb7152014-03-17 13:18:27 +1030143 else if (check_symbol_range(sym, s->addr, text_ranges,
144 ARRAY_SIZE(text_ranges)) == 0)
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400145 /* nothing to do */;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700146 else if (toupper(stype) == 'A')
Linus Torvalds1da177e2005-04-16 15:20:36 -0700147 {
148 /* Keep these useful absolute symbols */
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700149 if (strcmp(sym, "__kernel_syscall_via_break") &&
150 strcmp(sym, "__kernel_syscall_via_epc") &&
151 strcmp(sym, "__kernel_sigtramp") &&
152 strcmp(sym, "__gp"))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700153 return -1;
154
155 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700156 else if (toupper(stype) == 'U' ||
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700157 is_arm_mapping_symbol(sym))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700158 return -1;
Ralf Baechle6f00df22005-09-06 15:16:41 -0700159 /* exclude also MIPS ELF local symbols ($L123 instead of .L123) */
160 else if (str[0] == '$')
161 return -1;
Sam Ravnborgaab34ac2008-05-19 20:07:58 +0200162 /* exclude debugging symbols */
163 else if (stype == 'N')
164 return -1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700165
166 /* include the type field in the symbol name, so that it gets
167 * compressed together */
168 s->len = strlen(str) + 1;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700169 s->sym = malloc(s->len + 1);
Jesper Juhlf1a136e2006-03-25 03:07:46 -0800170 if (!s->sym) {
171 fprintf(stderr, "kallsyms failure: "
172 "unable to allocate required amount of memory\n");
173 exit(EXIT_FAILURE);
174 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700175 strcpy((char *)s->sym + 1, str);
176 s->sym[0] = stype;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700177
Ard Biesheuvel8c996942016-03-15 14:58:15 -0700178 s->percpu_absolute = 0;
179
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030180 /* Record if we've found __per_cpu_start/end. */
181 check_symbol_range(sym, s->addr, &percpu_range, 1);
182
Linus Torvalds1da177e2005-04-16 15:20:36 -0700183 return 0;
184}
185
Kees Cook78eb7152014-03-17 13:18:27 +1030186static int symbol_in_range(struct sym_entry *s, struct addr_range *ranges,
187 int entries)
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400188{
189 size_t i;
Kees Cook78eb7152014-03-17 13:18:27 +1030190 struct addr_range *ar;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400191
Kees Cook78eb7152014-03-17 13:18:27 +1030192 for (i = 0; i < entries; ++i) {
193 ar = &ranges[i];
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400194
Kees Cook78eb7152014-03-17 13:18:27 +1030195 if (s->addr >= ar->start && s->addr <= ar->end)
Mike Frysingerac6ca5c2009-06-15 07:52:48 -0400196 return 1;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400197 }
198
Mike Frysingerac6ca5c2009-06-15 07:52:48 -0400199 return 0;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400200}
201
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700202static int symbol_valid(struct sym_entry *s)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700203{
204 /* Symbols which vary between passes. Passes 1 and 2 must have
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100205 * identical symbol lists. The kallsyms_* symbols below are only added
206 * after pass 1, they would be included in pass 2 when --all-symbols is
207 * specified so exclude them to get a stable symbol list.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700208 */
209 static char *special_symbols[] = {
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100210 "kallsyms_addresses",
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700211 "kallsyms_offsets",
212 "kallsyms_relative_base",
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100213 "kallsyms_num_syms",
214 "kallsyms_names",
215 "kallsyms_markers",
216 "kallsyms_token_table",
217 "kallsyms_token_index",
218
Linus Torvalds1da177e2005-04-16 15:20:36 -0700219 /* Exclude linker generated symbols which vary between passes */
220 "_SDA_BASE_", /* ppc */
221 "_SDA2_BASE_", /* ppc */
222 NULL };
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200223
224 static char *special_suffixes[] = {
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200225 "_veneer", /* arm */
226 NULL };
227
Linus Torvalds1da177e2005-04-16 15:20:36 -0700228 int i;
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200229 char *sym_name = (char *)s->sym + 1;
230
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700231
Ming Leif6537f22013-11-02 09:11:33 +1030232 if (s->addr < kernel_start_addr)
233 return 0;
234
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700235 /* skip prefix char */
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200236 if (symbol_prefix_char && *sym_name == symbol_prefix_char)
237 sym_name++;
238
Linus Torvalds1da177e2005-04-16 15:20:36 -0700239
240 /* if --all-symbols is not specified, then symbols outside the text
241 * and inittext sections are discarded */
242 if (!all_symbols) {
Kees Cook78eb7152014-03-17 13:18:27 +1030243 if (symbol_in_range(s, text_ranges,
244 ARRAY_SIZE(text_ranges)) == 0)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700245 return 0;
246 /* Corner case. Discard any symbols with the same value as
Robin Getza3b81112008-02-06 01:36:26 -0800247 * _etext _einittext; they can move between pass 1 and 2 when
248 * the kallsyms data are added. If these symbols move then
249 * they may get dropped in pass 2, which breaks the kallsyms
250 * rules.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700251 */
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400252 if ((s->addr == text_range_text->end &&
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200253 strcmp(sym_name,
Kees Cook78eb7152014-03-17 13:18:27 +1030254 text_range_text->end_sym)) ||
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400255 (s->addr == text_range_inittext->end &&
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200256 strcmp(sym_name,
Kees Cook78eb7152014-03-17 13:18:27 +1030257 text_range_inittext->end_sym)))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700258 return 0;
259 }
260
261 /* Exclude symbols which vary between passes. */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700262 for (i = 0; special_symbols[i]; i++)
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200263 if (strcmp(sym_name, special_symbols[i]) == 0)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700264 return 0;
265
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200266 for (i = 0; special_suffixes[i]; i++) {
267 int l = strlen(sym_name) - strlen(special_suffixes[i]);
268
269 if (l >= 0 && strcmp(sym_name + l, special_suffixes[i]) == 0)
270 return 0;
271 }
272
Linus Torvalds1da177e2005-04-16 15:20:36 -0700273 return 1;
274}
275
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700276static void read_map(FILE *in)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700277{
278 while (!feof(in)) {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700279 if (table_cnt >= table_size) {
280 table_size += 10000;
281 table = realloc(table, sizeof(*table) * table_size);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700282 if (!table) {
283 fprintf(stderr, "out of memory\n");
284 exit (1);
285 }
286 }
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800287 if (read_symbol(in, &table[table_cnt]) == 0) {
288 table[table_cnt].start_pos = table_cnt;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700289 table_cnt++;
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800290 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700291 }
292}
293
294static void output_label(char *label)
295{
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700296 if (symbol_prefix_char)
297 printf(".globl %c%s\n", symbol_prefix_char, label);
298 else
299 printf(".globl %s\n", label);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700300 printf("\tALGN\n");
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700301 if (symbol_prefix_char)
302 printf("%c%s:\n", symbol_prefix_char, label);
303 else
304 printf("%s:\n", label);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700305}
306
307/* uncompress a compressed symbol. When this function is called, the best table
308 * might still be compressed itself, so the function needs to be recursive */
309static int expand_symbol(unsigned char *data, int len, char *result)
310{
311 int c, rlen, total=0;
312
313 while (len) {
314 c = *data;
315 /* if the table holds a single char that is the same as the one
316 * we are looking for, then end the search */
317 if (best_table[c][0]==c && best_table_len[c]==1) {
318 *result++ = c;
319 total++;
320 } else {
321 /* if not, recurse and expand */
322 rlen = expand_symbol(best_table[c], best_table_len[c], result);
323 total += rlen;
324 result += rlen;
325 }
326 data++;
327 len--;
328 }
329 *result=0;
330
331 return total;
332}
333
Kees Cook78eb7152014-03-17 13:18:27 +1030334static int symbol_absolute(struct sym_entry *s)
335{
Ard Biesheuvel8c996942016-03-15 14:58:15 -0700336 return s->percpu_absolute;
Kees Cook78eb7152014-03-17 13:18:27 +1030337}
338
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700339static void write_src(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700340{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700341 unsigned int i, k, off;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700342 unsigned int best_idx[256];
343 unsigned int *markers;
Tejun Heo9281ace2007-07-17 04:03:51 -0700344 char buf[KSYM_NAME_LEN];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700345
346 printf("#include <asm/types.h>\n");
347 printf("#if BITS_PER_LONG == 64\n");
348 printf("#define PTR .quad\n");
349 printf("#define ALGN .align 8\n");
350 printf("#else\n");
351 printf("#define PTR .long\n");
352 printf("#define ALGN .align 4\n");
353 printf("#endif\n");
354
Jan Beulichaad09472006-12-08 02:35:57 -0800355 printf("\t.section .rodata, \"a\"\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700356
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700357 /* Provide proper symbols relocatability by their relativeness
358 * to a fixed anchor point in the runtime image, either '_text'
359 * for absolute address tables, in which case the linker will
360 * emit the final addresses at build time. Otherwise, use the
361 * offset relative to the lowest value encountered of all relative
362 * symbols, and emit non-relocatable fixed offsets that will be fixed
363 * up at runtime.
364 *
365 * The symbol names cannot be used to construct normal symbol
366 * references as the list of symbols contains symbols that are
367 * declared static and are private to their .o files. This prevents
368 * .tmp_kallsyms.o or any other object from referencing them.
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100369 */
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700370 if (!base_relative)
371 output_label("kallsyms_addresses");
372 else
373 output_label("kallsyms_offsets");
374
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700375 for (i = 0; i < table_cnt; i++) {
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700376 if (base_relative) {
377 long long offset;
378 int overflow;
379
380 if (!absolute_percpu) {
381 offset = table[i].addr - relative_base;
382 overflow = (offset < 0 || offset > UINT_MAX);
383 } else if (symbol_absolute(&table[i])) {
384 offset = table[i].addr;
385 overflow = (offset < 0 || offset > INT_MAX);
386 } else {
387 offset = relative_base - table[i].addr - 1;
388 overflow = (offset < INT_MIN || offset >= 0);
389 }
390 if (overflow) {
391 fprintf(stderr, "kallsyms failure: "
392 "%s symbol value %#llx out of range in relative mode\n",
393 symbol_absolute(&table[i]) ? "absolute" : "relative",
394 table[i].addr);
395 exit(EXIT_FAILURE);
396 }
397 printf("\t.long\t%#x\n", (int)offset);
398 } else if (!symbol_absolute(&table[i])) {
Vivek Goyal2c22d8b2006-12-07 02:14:10 +0100399 if (_text <= table[i].addr)
400 printf("\tPTR\t_text + %#llx\n",
401 table[i].addr - _text);
402 else
Andrew Morton2930ffc2014-03-10 15:49:48 -0700403 printf("\tPTR\t_text - %#llx\n",
404 _text - table[i].addr);
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100405 } else {
406 printf("\tPTR\t%#llx\n", table[i].addr);
407 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700408 }
409 printf("\n");
410
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700411 if (base_relative) {
412 output_label("kallsyms_relative_base");
413 printf("\tPTR\t_text - %#llx\n", _text - relative_base);
414 printf("\n");
415 }
416
Linus Torvalds1da177e2005-04-16 15:20:36 -0700417 output_label("kallsyms_num_syms");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700418 printf("\tPTR\t%d\n", table_cnt);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700419 printf("\n");
420
421 /* table of offset markers, that give the offset in the compressed stream
422 * every 256 symbols */
Jesper Juhlf1a136e2006-03-25 03:07:46 -0800423 markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
424 if (!markers) {
425 fprintf(stderr, "kallsyms failure: "
426 "unable to allocate required memory\n");
427 exit(EXIT_FAILURE);
428 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700429
430 output_label("kallsyms_names");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700431 off = 0;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700432 for (i = 0; i < table_cnt; i++) {
433 if ((i & 0xFF) == 0)
434 markers[i >> 8] = off;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700435
436 printf("\t.byte 0x%02x", table[i].len);
437 for (k = 0; k < table[i].len; k++)
438 printf(", 0x%02x", table[i].sym[k]);
439 printf("\n");
440
441 off += table[i].len + 1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700442 }
443 printf("\n");
444
445 output_label("kallsyms_markers");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700446 for (i = 0; i < ((table_cnt + 255) >> 8); i++)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700447 printf("\tPTR\t%d\n", markers[i]);
448 printf("\n");
449
450 free(markers);
451
452 output_label("kallsyms_token_table");
453 off = 0;
454 for (i = 0; i < 256; i++) {
455 best_idx[i] = off;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700456 expand_symbol(best_table[i], best_table_len[i], buf);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700457 printf("\t.asciz\t\"%s\"\n", buf);
458 off += strlen(buf) + 1;
459 }
460 printf("\n");
461
462 output_label("kallsyms_token_index");
463 for (i = 0; i < 256; i++)
464 printf("\t.short\t%d\n", best_idx[i]);
465 printf("\n");
466}
467
468
469/* table lookup compression functions */
470
Linus Torvalds1da177e2005-04-16 15:20:36 -0700471/* count all the possible tokens in a symbol */
472static void learn_symbol(unsigned char *symbol, int len)
473{
474 int i;
475
476 for (i = 0; i < len - 1; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700477 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700478}
479
480/* decrease the count for all the possible tokens in a symbol */
481static void forget_symbol(unsigned char *symbol, int len)
482{
483 int i;
484
485 for (i = 0; i < len - 1; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700486 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700487}
488
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700489/* remove all the invalid symbols from the table and do the initial token count */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700490static void build_initial_tok_table(void)
491{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700492 unsigned int i, pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700493
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700494 pos = 0;
495 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700496 if ( symbol_valid(&table[i]) ) {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700497 if (pos != i)
498 table[pos] = table[i];
499 learn_symbol(table[pos].sym, table[pos].len);
500 pos++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700501 }
502 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700503 table_cnt = pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700504}
505
Paulo Marques7c5d2492007-06-20 18:09:00 +0100506static void *find_token(unsigned char *str, int len, unsigned char *token)
507{
508 int i;
509
510 for (i = 0; i < len - 1; i++) {
511 if (str[i] == token[0] && str[i+1] == token[1])
512 return &str[i];
513 }
514 return NULL;
515}
516
Linus Torvalds1da177e2005-04-16 15:20:36 -0700517/* replace a given token in all the valid symbols. Use the sampled symbols
518 * to update the counts */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700519static void compress_symbols(unsigned char *str, int idx)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700520{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700521 unsigned int i, len, size;
522 unsigned char *p1, *p2;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700523
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700524 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700525
526 len = table[i].len;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700527 p1 = table[i].sym;
528
529 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100530 p2 = find_token(p1, len, str);
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700531 if (!p2) continue;
532
533 /* decrease the counts for this symbol's tokens */
534 forget_symbol(table[i].sym, len);
535
536 size = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700537
538 do {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700539 *p2 = idx;
540 p2++;
541 size -= (p2 - p1);
542 memmove(p2, p2 + 1, size);
543 p1 = p2;
544 len--;
545
546 if (size < 2) break;
547
Linus Torvalds1da177e2005-04-16 15:20:36 -0700548 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100549 p2 = find_token(p1, size, str);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700550
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700551 } while (p2);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700552
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700553 table[i].len = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700554
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700555 /* increase the counts for this symbol's new tokens */
556 learn_symbol(table[i].sym, len);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700557 }
558}
559
560/* search the token with the maximum profit */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700561static int find_best_token(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700562{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700563 int i, best, bestprofit;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700564
565 bestprofit=-10000;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700566 best = 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700567
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700568 for (i = 0; i < 0x10000; i++) {
569 if (token_profit[i] > bestprofit) {
570 best = i;
571 bestprofit = token_profit[i];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700572 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700573 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700574 return best;
575}
576
577/* this is the core of the algorithm: calculate the "best" table */
578static void optimize_result(void)
579{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700580 int i, best;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700581
582 /* using the '\0' symbol last allows compress_symbols to use standard
583 * fast string functions */
584 for (i = 255; i >= 0; i--) {
585
586 /* if this table slot is empty (it is not used by an actual
587 * original char code */
588 if (!best_table_len[i]) {
589
590 /* find the token with the breates profit value */
591 best = find_best_token();
Xiaochen Wange0a04b12011-05-01 11:41:41 +0800592 if (token_profit[best] == 0)
593 break;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700594
595 /* place it in the "best" table */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700596 best_table_len[i] = 2;
597 best_table[i][0] = best & 0xFF;
598 best_table[i][1] = (best >> 8) & 0xFF;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700599
600 /* replace this token in all the valid symbols */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700601 compress_symbols(best_table[i], i);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700602 }
603 }
604}
605
606/* start by placing the symbols that are actually used on the table */
607static void insert_real_symbols_in_table(void)
608{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700609 unsigned int i, j, c;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700610
611 memset(best_table, 0, sizeof(best_table));
612 memset(best_table_len, 0, sizeof(best_table_len));
613
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700614 for (i = 0; i < table_cnt; i++) {
615 for (j = 0; j < table[i].len; j++) {
616 c = table[i].sym[j];
617 best_table[c][0]=c;
618 best_table_len[c]=1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700619 }
620 }
621}
622
623static void optimize_token_table(void)
624{
Linus Torvalds1da177e2005-04-16 15:20:36 -0700625 build_initial_tok_table();
626
627 insert_real_symbols_in_table();
628
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700629 /* When valid symbol is not registered, exit to error */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700630 if (!table_cnt) {
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700631 fprintf(stderr, "No valid symbol.\n");
632 exit(1);
633 }
634
Linus Torvalds1da177e2005-04-16 15:20:36 -0700635 optimize_result();
636}
637
Lai Jiangshanb478b7822009-03-13 15:10:26 +0800638/* guess for "linker script provide" symbol */
639static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
640{
641 const char *symbol = (char *)se->sym + 1;
642 int len = se->len - 1;
643
644 if (len < 8)
645 return 0;
646
647 if (symbol[0] != '_' || symbol[1] != '_')
648 return 0;
649
650 /* __start_XXXXX */
651 if (!memcmp(symbol + 2, "start_", 6))
652 return 1;
653
654 /* __stop_XXXXX */
655 if (!memcmp(symbol + 2, "stop_", 5))
656 return 1;
657
658 /* __end_XXXXX */
659 if (!memcmp(symbol + 2, "end_", 4))
660 return 1;
661
662 /* __XXXXX_start */
663 if (!memcmp(symbol + len - 6, "_start", 6))
664 return 1;
665
666 /* __XXXXX_end */
667 if (!memcmp(symbol + len - 4, "_end", 4))
668 return 1;
669
670 return 0;
671}
672
673static int prefix_underscores_count(const char *str)
674{
675 const char *tail = str;
676
Paul Mundta9ece532009-09-22 16:44:12 -0700677 while (*tail == '_')
Lai Jiangshanb478b7822009-03-13 15:10:26 +0800678 tail++;
679
680 return tail - str;
681}
682
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800683static int compare_symbols(const void *a, const void *b)
684{
685 const struct sym_entry *sa;
686 const struct sym_entry *sb;
687 int wa, wb;
688
689 sa = a;
690 sb = b;
691
692 /* sort by address first */
693 if (sa->addr > sb->addr)
694 return 1;
695 if (sa->addr < sb->addr)
696 return -1;
697
698 /* sort by "weakness" type */
699 wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
700 wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
701 if (wa != wb)
702 return wa - wb;
703
Lai Jiangshanb478b7822009-03-13 15:10:26 +0800704 /* sort by "linker script provide" type */
705 wa = may_be_linker_script_provide_symbol(sa);
706 wb = may_be_linker_script_provide_symbol(sb);
707 if (wa != wb)
708 return wa - wb;
709
710 /* sort by the number of prefix underscores */
711 wa = prefix_underscores_count((const char *)sa->sym + 1);
712 wb = prefix_underscores_count((const char *)sb->sym + 1);
713 if (wa != wb)
714 return wa - wb;
715
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800716 /* sort by initial order, so that other symbols are left undisturbed */
717 return sa->start_pos - sb->start_pos;
718}
719
720static void sort_symbols(void)
721{
722 qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
723}
Linus Torvalds1da177e2005-04-16 15:20:36 -0700724
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030725static void make_percpus_absolute(void)
726{
727 unsigned int i;
728
729 for (i = 0; i < table_cnt; i++)
Ard Biesheuvel8c996942016-03-15 14:58:15 -0700730 if (symbol_in_range(&table[i], &percpu_range, 1)) {
731 /*
732 * Keep the 'A' override for percpu symbols to
733 * ensure consistent behavior compared to older
734 * versions of this tool.
735 */
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030736 table[i].sym[0] = 'A';
Ard Biesheuvel8c996942016-03-15 14:58:15 -0700737 table[i].percpu_absolute = 1;
738 }
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030739}
740
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700741/* find the minimum non-absolute symbol address */
742static void record_relative_base(void)
743{
744 unsigned int i;
745
746 relative_base = -1ULL;
747 for (i = 0; i < table_cnt; i++)
748 if (!symbol_absolute(&table[i]) &&
749 table[i].addr < relative_base)
750 relative_base = table[i].addr;
751}
752
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700753int main(int argc, char **argv)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700754{
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700755 if (argc >= 2) {
756 int i;
757 for (i = 1; i < argc; i++) {
758 if(strcmp(argv[i], "--all-symbols") == 0)
759 all_symbols = 1;
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030760 else if (strcmp(argv[i], "--absolute-percpu") == 0)
761 absolute_percpu = 1;
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700762 else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
763 char *p = &argv[i][16];
764 /* skip quote */
765 if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
766 p++;
767 symbol_prefix_char = *p;
Ming Leif6537f22013-11-02 09:11:33 +1030768 } else if (strncmp(argv[i], "--page-offset=", 14) == 0) {
769 const char *p = &argv[i][14];
770 kernel_start_addr = strtoull(p, NULL, 16);
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700771 } else if (strcmp(argv[i], "--base-relative") == 0)
772 base_relative = 1;
773 else
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700774 usage();
775 }
776 } else if (argc != 1)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700777 usage();
778
779 read_map(stdin);
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030780 if (absolute_percpu)
781 make_percpus_absolute();
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700782 if (base_relative)
783 record_relative_base();
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100784 sort_symbols();
785 optimize_token_table();
Linus Torvalds1da177e2005-04-16 15:20:36 -0700786 write_src();
787
788 return 0;
789}