blob: ae6504d07fd6123b6a9d984bec070ece4c59d613 [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" },
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040051};
52#define text_range_text (&text_ranges[0])
53#define text_range_inittext (&text_ranges[1])
54
Rusty Russellc6bda7c2014-03-17 14:05:46 +103055static struct addr_range percpu_range = {
56 "__per_cpu_start", "__per_cpu_end", -1ULL, 0
57};
58
Linus Torvalds1da177e2005-04-16 15:20:36 -070059static struct sym_entry *table;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070060static unsigned int table_size, table_cnt;
Linus Torvalds1da177e2005-04-16 15:20:36 -070061static int all_symbols = 0;
Rusty Russellc6bda7c2014-03-17 14:05:46 +103062static int absolute_percpu = 0;
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -070063static int base_relative = 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -070064
Masahiro Yamadaf43e9da2019-02-04 10:53:16 +090065static int token_profit[0x10000];
Linus Torvalds1da177e2005-04-16 15:20:36 -070066
67/* the table that holds the result of the compression */
Masahiro Yamadaf43e9da2019-02-04 10:53:16 +090068static unsigned char best_table[256][2];
69static unsigned char best_table_len[256];
Linus Torvalds1da177e2005-04-16 15:20:36 -070070
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] "
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -070075 "[--base-relative] < in.map > out.S\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -070076 exit(1);
77}
78
79/*
80 * This ignores the intensely annoying "mapping symbols" found
81 * in ARM ELF files: $a, $t and $d.
82 */
Masahiro Yamadaf43e9da2019-02-04 10:53:16 +090083static int is_arm_mapping_symbol(const char *str)
Linus Torvalds1da177e2005-04-16 15:20:36 -070084{
Kyle McMartin6c34f1f2014-09-16 22:37:18 +010085 return str[0] == '$' && strchr("axtd", str[1])
Linus Torvalds1da177e2005-04-16 15:20:36 -070086 && (str[2] == '\0' || str[2] == '.');
87}
88
Kees Cook78eb7152014-03-17 13:18:27 +103089static int check_symbol_range(const char *sym, unsigned long long addr,
90 struct addr_range *ranges, int entries)
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040091{
92 size_t i;
Kees Cook78eb7152014-03-17 13:18:27 +103093 struct addr_range *ar;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040094
Kees Cook78eb7152014-03-17 13:18:27 +103095 for (i = 0; i < entries; ++i) {
96 ar = &ranges[i];
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040097
Kees Cook78eb7152014-03-17 13:18:27 +103098 if (strcmp(sym, ar->start_sym) == 0) {
99 ar->start = addr;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400100 return 0;
Kees Cook78eb7152014-03-17 13:18:27 +1030101 } else if (strcmp(sym, ar->end_sym) == 0) {
102 ar->end = addr;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400103 return 0;
104 }
105 }
106
107 return 1;
108}
109
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700110static int read_symbol(FILE *in, struct sym_entry *s)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700111{
Masahiro Yamada534c9f22018-05-09 16:23:47 +0900112 char sym[500], stype;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700113 int rc;
114
Masahiro Yamada534c9f22018-05-09 16:23:47 +0900115 rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, sym);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700116 if (rc != 3) {
Masahiro Yamada534c9f22018-05-09 16:23:47 +0900117 if (rc != EOF && fgets(sym, 500, in) == NULL)
Jean Sacrenef894872010-09-10 23:13:33 -0600118 fprintf(stderr, "Read error or end of file.\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700119 return -1;
120 }
Eugene Loh6db29832019-01-17 14:46:00 -0800121 if (strlen(sym) >= KSYM_NAME_LEN) {
122 fprintf(stderr, "Symbol %s too long for kallsyms (%zu >= %d).\n"
Masahiro Yamadabb66fc62014-06-10 19:08:13 +0900123 "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n",
Masahiro Yamada534c9f22018-05-09 16:23:47 +0900124 sym, strlen(sym), KSYM_NAME_LEN);
Andi Kleenf3462aa2013-10-23 15:07:53 +0200125 return -1;
126 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700127
128 /* Ignore most absolute/undefined (?) symbols. */
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100129 if (strcmp(sym, "_text") == 0)
130 _text = s->addr;
Kees Cook78eb7152014-03-17 13:18:27 +1030131 else if (check_symbol_range(sym, s->addr, text_ranges,
132 ARRAY_SIZE(text_ranges)) == 0)
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400133 /* nothing to do */;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700134 else if (toupper(stype) == 'A')
Linus Torvalds1da177e2005-04-16 15:20:36 -0700135 {
136 /* Keep these useful absolute symbols */
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700137 if (strcmp(sym, "__kernel_syscall_via_break") &&
138 strcmp(sym, "__kernel_syscall_via_epc") &&
139 strcmp(sym, "__kernel_sigtramp") &&
140 strcmp(sym, "__gp"))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700141 return -1;
142
143 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700144 else if (toupper(stype) == 'U' ||
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700145 is_arm_mapping_symbol(sym))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700146 return -1;
Ralf Baechle6f00df22005-09-06 15:16:41 -0700147 /* exclude also MIPS ELF local symbols ($L123 instead of .L123) */
Masahiro Yamada534c9f22018-05-09 16:23:47 +0900148 else if (sym[0] == '$')
Ralf Baechle6f00df22005-09-06 15:16:41 -0700149 return -1;
Sam Ravnborgaab34ac2008-05-19 20:07:58 +0200150 /* exclude debugging symbols */
Guenter Roeck51962a92017-10-13 15:57:58 -0700151 else if (stype == 'N' || stype == 'n')
Sam Ravnborgaab34ac2008-05-19 20:07:58 +0200152 return -1;
Vasily Gorbik33177f012019-06-28 19:22:47 +0200153 /* exclude s390 kasan local symbols */
154 else if (!strncmp(sym, ".LASANPC", 8))
155 return -1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700156
157 /* include the type field in the symbol name, so that it gets
158 * compressed together */
Masahiro Yamada534c9f22018-05-09 16:23:47 +0900159 s->len = strlen(sym) + 1;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700160 s->sym = malloc(s->len + 1);
Jesper Juhlf1a136e2006-03-25 03:07:46 -0800161 if (!s->sym) {
162 fprintf(stderr, "kallsyms failure: "
163 "unable to allocate required amount of memory\n");
164 exit(EXIT_FAILURE);
165 }
Masahiro Yamada534c9f22018-05-09 16:23:47 +0900166 strcpy((char *)s->sym + 1, sym);
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700167 s->sym[0] = stype;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700168
Ard Biesheuvel8c996942016-03-15 14:58:15 -0700169 s->percpu_absolute = 0;
170
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030171 /* Record if we've found __per_cpu_start/end. */
172 check_symbol_range(sym, s->addr, &percpu_range, 1);
173
Linus Torvalds1da177e2005-04-16 15:20:36 -0700174 return 0;
175}
176
Kees Cook78eb7152014-03-17 13:18:27 +1030177static int symbol_in_range(struct sym_entry *s, struct addr_range *ranges,
178 int entries)
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400179{
180 size_t i;
Kees Cook78eb7152014-03-17 13:18:27 +1030181 struct addr_range *ar;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400182
Kees Cook78eb7152014-03-17 13:18:27 +1030183 for (i = 0; i < entries; ++i) {
184 ar = &ranges[i];
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400185
Kees Cook78eb7152014-03-17 13:18:27 +1030186 if (s->addr >= ar->start && s->addr <= ar->end)
Mike Frysingerac6ca5c2009-06-15 07:52:48 -0400187 return 1;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400188 }
189
Mike Frysingerac6ca5c2009-06-15 07:52:48 -0400190 return 0;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400191}
192
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700193static int symbol_valid(struct sym_entry *s)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700194{
195 /* Symbols which vary between passes. Passes 1 and 2 must have
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100196 * identical symbol lists. The kallsyms_* symbols below are only added
197 * after pass 1, they would be included in pass 2 when --all-symbols is
198 * specified so exclude them to get a stable symbol list.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700199 */
200 static char *special_symbols[] = {
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100201 "kallsyms_addresses",
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700202 "kallsyms_offsets",
203 "kallsyms_relative_base",
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100204 "kallsyms_num_syms",
205 "kallsyms_names",
206 "kallsyms_markers",
207 "kallsyms_token_table",
208 "kallsyms_token_index",
209
Linus Torvalds1da177e2005-04-16 15:20:36 -0700210 /* Exclude linker generated symbols which vary between passes */
211 "_SDA_BASE_", /* ppc */
212 "_SDA2_BASE_", /* ppc */
213 NULL };
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200214
Ard Biesheuvel56067812017-02-03 09:54:05 +0000215 static char *special_prefixes[] = {
216 "__crc_", /* modversions */
Ard Biesheuvel1212f7a2018-03-01 17:19:01 +0000217 "__efistub_", /* arm64 EFI stub namespace */
Ard Biesheuvel56067812017-02-03 09:54:05 +0000218 NULL };
219
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200220 static char *special_suffixes[] = {
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200221 "_veneer", /* arm */
Ard Biesheuvelb9b74be2016-04-01 14:33:48 +0100222 "_from_arm", /* arm */
223 "_from_thumb", /* arm */
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200224 NULL };
225
Linus Torvalds1da177e2005-04-16 15:20:36 -0700226 int i;
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200227 char *sym_name = (char *)s->sym + 1;
228
Linus Torvalds1da177e2005-04-16 15:20:36 -0700229 /* if --all-symbols is not specified, then symbols outside the text
230 * and inittext sections are discarded */
231 if (!all_symbols) {
Kees Cook78eb7152014-03-17 13:18:27 +1030232 if (symbol_in_range(s, text_ranges,
233 ARRAY_SIZE(text_ranges)) == 0)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700234 return 0;
235 /* Corner case. Discard any symbols with the same value as
Robin Getza3b81112008-02-06 01:36:26 -0800236 * _etext _einittext; they can move between pass 1 and 2 when
237 * the kallsyms data are added. If these symbols move then
238 * they may get dropped in pass 2, which breaks the kallsyms
239 * rules.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700240 */
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400241 if ((s->addr == text_range_text->end &&
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200242 strcmp(sym_name,
Kees Cook78eb7152014-03-17 13:18:27 +1030243 text_range_text->end_sym)) ||
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400244 (s->addr == text_range_inittext->end &&
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200245 strcmp(sym_name,
Kees Cook78eb7152014-03-17 13:18:27 +1030246 text_range_inittext->end_sym)))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700247 return 0;
248 }
249
250 /* Exclude symbols which vary between passes. */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700251 for (i = 0; special_symbols[i]; i++)
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200252 if (strcmp(sym_name, special_symbols[i]) == 0)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700253 return 0;
254
Ard Biesheuvel56067812017-02-03 09:54:05 +0000255 for (i = 0; special_prefixes[i]; i++) {
256 int l = strlen(special_prefixes[i]);
257
258 if (l <= strlen(sym_name) &&
259 strncmp(sym_name, special_prefixes[i], l) == 0)
260 return 0;
261 }
262
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200263 for (i = 0; special_suffixes[i]; i++) {
264 int l = strlen(sym_name) - strlen(special_suffixes[i]);
265
266 if (l >= 0 && strcmp(sym_name + l, special_suffixes[i]) == 0)
267 return 0;
268 }
269
Linus Torvalds1da177e2005-04-16 15:20:36 -0700270 return 1;
271}
272
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700273static void read_map(FILE *in)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700274{
275 while (!feof(in)) {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700276 if (table_cnt >= table_size) {
277 table_size += 10000;
278 table = realloc(table, sizeof(*table) * table_size);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700279 if (!table) {
280 fprintf(stderr, "out of memory\n");
281 exit (1);
282 }
283 }
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800284 if (read_symbol(in, &table[table_cnt]) == 0) {
285 table[table_cnt].start_pos = table_cnt;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700286 table_cnt++;
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800287 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700288 }
289}
290
291static void output_label(char *label)
292{
Masahiro Yamada534c9f22018-05-09 16:23:47 +0900293 printf(".globl %s\n", label);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700294 printf("\tALGN\n");
Masahiro Yamada534c9f22018-05-09 16:23:47 +0900295 printf("%s:\n", label);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700296}
297
298/* uncompress a compressed symbol. When this function is called, the best table
299 * might still be compressed itself, so the function needs to be recursive */
300static int expand_symbol(unsigned char *data, int len, char *result)
301{
302 int c, rlen, total=0;
303
304 while (len) {
305 c = *data;
306 /* if the table holds a single char that is the same as the one
307 * we are looking for, then end the search */
308 if (best_table[c][0]==c && best_table_len[c]==1) {
309 *result++ = c;
310 total++;
311 } else {
312 /* if not, recurse and expand */
313 rlen = expand_symbol(best_table[c], best_table_len[c], result);
314 total += rlen;
315 result += rlen;
316 }
317 data++;
318 len--;
319 }
320 *result=0;
321
322 return total;
323}
324
Kees Cook78eb7152014-03-17 13:18:27 +1030325static int symbol_absolute(struct sym_entry *s)
326{
Ard Biesheuvel8c996942016-03-15 14:58:15 -0700327 return s->percpu_absolute;
Kees Cook78eb7152014-03-17 13:18:27 +1030328}
329
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700330static void write_src(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700331{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700332 unsigned int i, k, off;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700333 unsigned int best_idx[256];
334 unsigned int *markers;
Tejun Heo9281ace2007-07-17 04:03:51 -0700335 char buf[KSYM_NAME_LEN];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700336
Masahiro Yamada500193e2019-02-04 10:53:18 +0900337 printf("#include <asm/bitsperlong.h>\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700338 printf("#if BITS_PER_LONG == 64\n");
339 printf("#define PTR .quad\n");
Mathias Krause72d3ebb2018-12-30 13:36:00 +0100340 printf("#define ALGN .balign 8\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700341 printf("#else\n");
342 printf("#define PTR .long\n");
Mathias Krause72d3ebb2018-12-30 13:36:00 +0100343 printf("#define ALGN .balign 4\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700344 printf("#endif\n");
345
Jan Beulichaad09472006-12-08 02:35:57 -0800346 printf("\t.section .rodata, \"a\"\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700347
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700348 /* Provide proper symbols relocatability by their relativeness
349 * to a fixed anchor point in the runtime image, either '_text'
350 * for absolute address tables, in which case the linker will
351 * emit the final addresses at build time. Otherwise, use the
352 * offset relative to the lowest value encountered of all relative
353 * symbols, and emit non-relocatable fixed offsets that will be fixed
354 * up at runtime.
355 *
356 * The symbol names cannot be used to construct normal symbol
357 * references as the list of symbols contains symbols that are
358 * declared static and are private to their .o files. This prevents
359 * .tmp_kallsyms.o or any other object from referencing them.
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100360 */
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700361 if (!base_relative)
362 output_label("kallsyms_addresses");
363 else
364 output_label("kallsyms_offsets");
365
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700366 for (i = 0; i < table_cnt; i++) {
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700367 if (base_relative) {
368 long long offset;
369 int overflow;
370
371 if (!absolute_percpu) {
372 offset = table[i].addr - relative_base;
373 overflow = (offset < 0 || offset > UINT_MAX);
374 } else if (symbol_absolute(&table[i])) {
375 offset = table[i].addr;
376 overflow = (offset < 0 || offset > INT_MAX);
377 } else {
378 offset = relative_base - table[i].addr - 1;
379 overflow = (offset < INT_MIN || offset >= 0);
380 }
381 if (overflow) {
382 fprintf(stderr, "kallsyms failure: "
383 "%s symbol value %#llx out of range in relative mode\n",
384 symbol_absolute(&table[i]) ? "absolute" : "relative",
385 table[i].addr);
386 exit(EXIT_FAILURE);
387 }
388 printf("\t.long\t%#x\n", (int)offset);
389 } else if (!symbol_absolute(&table[i])) {
Vivek Goyal2c22d8b2006-12-07 02:14:10 +0100390 if (_text <= table[i].addr)
391 printf("\tPTR\t_text + %#llx\n",
392 table[i].addr - _text);
393 else
Andrew Morton2930ffc2014-03-10 15:49:48 -0700394 printf("\tPTR\t_text - %#llx\n",
395 _text - table[i].addr);
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100396 } else {
397 printf("\tPTR\t%#llx\n", table[i].addr);
398 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700399 }
400 printf("\n");
401
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700402 if (base_relative) {
403 output_label("kallsyms_relative_base");
404 printf("\tPTR\t_text - %#llx\n", _text - relative_base);
405 printf("\n");
406 }
407
Linus Torvalds1da177e2005-04-16 15:20:36 -0700408 output_label("kallsyms_num_syms");
Jan Beulich80ffbaa2018-09-03 06:09:34 -0600409 printf("\t.long\t%u\n", table_cnt);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700410 printf("\n");
411
412 /* table of offset markers, that give the offset in the compressed stream
413 * every 256 symbols */
Jesper Juhlf1a136e2006-03-25 03:07:46 -0800414 markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
415 if (!markers) {
416 fprintf(stderr, "kallsyms failure: "
417 "unable to allocate required memory\n");
418 exit(EXIT_FAILURE);
419 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700420
421 output_label("kallsyms_names");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700422 off = 0;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700423 for (i = 0; i < table_cnt; i++) {
424 if ((i & 0xFF) == 0)
425 markers[i >> 8] = off;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700426
427 printf("\t.byte 0x%02x", table[i].len);
428 for (k = 0; k < table[i].len; k++)
429 printf(", 0x%02x", table[i].sym[k]);
430 printf("\n");
431
432 off += table[i].len + 1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700433 }
434 printf("\n");
435
436 output_label("kallsyms_markers");
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700437 for (i = 0; i < ((table_cnt + 255) >> 8); i++)
Jan Beulich80ffbaa2018-09-03 06:09:34 -0600438 printf("\t.long\t%u\n", markers[i]);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700439 printf("\n");
440
441 free(markers);
442
443 output_label("kallsyms_token_table");
444 off = 0;
445 for (i = 0; i < 256; i++) {
446 best_idx[i] = off;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700447 expand_symbol(best_table[i], best_table_len[i], buf);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700448 printf("\t.asciz\t\"%s\"\n", buf);
449 off += strlen(buf) + 1;
450 }
451 printf("\n");
452
453 output_label("kallsyms_token_index");
454 for (i = 0; i < 256; i++)
455 printf("\t.short\t%d\n", best_idx[i]);
456 printf("\n");
457}
458
459
460/* table lookup compression functions */
461
Linus Torvalds1da177e2005-04-16 15:20:36 -0700462/* count all the possible tokens in a symbol */
463static void learn_symbol(unsigned char *symbol, int len)
464{
465 int i;
466
467 for (i = 0; i < len - 1; i++)
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700468 token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700469}
470
471/* decrease the count for all the possible tokens in a symbol */
472static void forget_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
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700480/* remove all the invalid symbols from the table and do the initial token count */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700481static void build_initial_tok_table(void)
482{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700483 unsigned int i, pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700484
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700485 pos = 0;
486 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700487 if ( symbol_valid(&table[i]) ) {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700488 if (pos != i)
489 table[pos] = table[i];
490 learn_symbol(table[pos].sym, table[pos].len);
491 pos++;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700492 }
493 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700494 table_cnt = pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700495}
496
Paulo Marques7c5d2492007-06-20 18:09:00 +0100497static void *find_token(unsigned char *str, int len, unsigned char *token)
498{
499 int i;
500
501 for (i = 0; i < len - 1; i++) {
502 if (str[i] == token[0] && str[i+1] == token[1])
503 return &str[i];
504 }
505 return NULL;
506}
507
Linus Torvalds1da177e2005-04-16 15:20:36 -0700508/* replace a given token in all the valid symbols. Use the sampled symbols
509 * to update the counts */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700510static void compress_symbols(unsigned char *str, int idx)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700511{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700512 unsigned int i, len, size;
513 unsigned char *p1, *p2;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700514
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700515 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700516
517 len = table[i].len;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700518 p1 = table[i].sym;
519
520 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100521 p2 = find_token(p1, len, str);
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700522 if (!p2) continue;
523
524 /* decrease the counts for this symbol's tokens */
525 forget_symbol(table[i].sym, len);
526
527 size = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700528
529 do {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700530 *p2 = idx;
531 p2++;
532 size -= (p2 - p1);
533 memmove(p2, p2 + 1, size);
534 p1 = p2;
535 len--;
536
537 if (size < 2) break;
538
Linus Torvalds1da177e2005-04-16 15:20:36 -0700539 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100540 p2 = find_token(p1, size, str);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700541
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700542 } while (p2);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700543
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700544 table[i].len = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700545
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700546 /* increase the counts for this symbol's new tokens */
547 learn_symbol(table[i].sym, len);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700548 }
549}
550
551/* search the token with the maximum profit */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700552static int find_best_token(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700553{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700554 int i, best, bestprofit;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700555
556 bestprofit=-10000;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700557 best = 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700558
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700559 for (i = 0; i < 0x10000; i++) {
560 if (token_profit[i] > bestprofit) {
561 best = i;
562 bestprofit = token_profit[i];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700563 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700564 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700565 return best;
566}
567
568/* this is the core of the algorithm: calculate the "best" table */
569static void optimize_result(void)
570{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700571 int i, best;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700572
573 /* using the '\0' symbol last allows compress_symbols to use standard
574 * fast string functions */
575 for (i = 255; i >= 0; i--) {
576
577 /* if this table slot is empty (it is not used by an actual
578 * original char code */
579 if (!best_table_len[i]) {
580
Cao jincbf7a902018-02-27 16:16:19 +0800581 /* find the token with the best profit value */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700582 best = find_best_token();
Xiaochen Wange0a04b12011-05-01 11:41:41 +0800583 if (token_profit[best] == 0)
584 break;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700585
586 /* place it in the "best" table */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700587 best_table_len[i] = 2;
588 best_table[i][0] = best & 0xFF;
589 best_table[i][1] = (best >> 8) & 0xFF;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700590
591 /* replace this token in all the valid symbols */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700592 compress_symbols(best_table[i], i);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700593 }
594 }
595}
596
597/* start by placing the symbols that are actually used on the table */
598static void insert_real_symbols_in_table(void)
599{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700600 unsigned int i, j, c;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700601
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700602 for (i = 0; i < table_cnt; i++) {
603 for (j = 0; j < table[i].len; j++) {
604 c = table[i].sym[j];
605 best_table[c][0]=c;
606 best_table_len[c]=1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700607 }
608 }
609}
610
611static void optimize_token_table(void)
612{
Linus Torvalds1da177e2005-04-16 15:20:36 -0700613 build_initial_tok_table();
614
615 insert_real_symbols_in_table();
616
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700617 /* When valid symbol is not registered, exit to error */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700618 if (!table_cnt) {
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700619 fprintf(stderr, "No valid symbol.\n");
620 exit(1);
621 }
622
Linus Torvalds1da177e2005-04-16 15:20:36 -0700623 optimize_result();
624}
625
Lai Jiangshanb478b782009-03-13 15:10:26 +0800626/* guess for "linker script provide" symbol */
627static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
628{
629 const char *symbol = (char *)se->sym + 1;
630 int len = se->len - 1;
631
632 if (len < 8)
633 return 0;
634
635 if (symbol[0] != '_' || symbol[1] != '_')
636 return 0;
637
638 /* __start_XXXXX */
639 if (!memcmp(symbol + 2, "start_", 6))
640 return 1;
641
642 /* __stop_XXXXX */
643 if (!memcmp(symbol + 2, "stop_", 5))
644 return 1;
645
646 /* __end_XXXXX */
647 if (!memcmp(symbol + 2, "end_", 4))
648 return 1;
649
650 /* __XXXXX_start */
651 if (!memcmp(symbol + len - 6, "_start", 6))
652 return 1;
653
654 /* __XXXXX_end */
655 if (!memcmp(symbol + len - 4, "_end", 4))
656 return 1;
657
658 return 0;
659}
660
661static int prefix_underscores_count(const char *str)
662{
663 const char *tail = str;
664
Paul Mundta9ece532009-09-22 16:44:12 -0700665 while (*tail == '_')
Lai Jiangshanb478b782009-03-13 15:10:26 +0800666 tail++;
667
668 return tail - str;
669}
670
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800671static int compare_symbols(const void *a, const void *b)
672{
673 const struct sym_entry *sa;
674 const struct sym_entry *sb;
675 int wa, wb;
676
677 sa = a;
678 sb = b;
679
680 /* sort by address first */
681 if (sa->addr > sb->addr)
682 return 1;
683 if (sa->addr < sb->addr)
684 return -1;
685
686 /* sort by "weakness" type */
687 wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
688 wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
689 if (wa != wb)
690 return wa - wb;
691
Lai Jiangshanb478b782009-03-13 15:10:26 +0800692 /* sort by "linker script provide" type */
693 wa = may_be_linker_script_provide_symbol(sa);
694 wb = may_be_linker_script_provide_symbol(sb);
695 if (wa != wb)
696 return wa - wb;
697
698 /* sort by the number of prefix underscores */
699 wa = prefix_underscores_count((const char *)sa->sym + 1);
700 wb = prefix_underscores_count((const char *)sb->sym + 1);
701 if (wa != wb)
702 return wa - wb;
703
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800704 /* sort by initial order, so that other symbols are left undisturbed */
705 return sa->start_pos - sb->start_pos;
706}
707
708static void sort_symbols(void)
709{
710 qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
711}
Linus Torvalds1da177e2005-04-16 15:20:36 -0700712
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030713static void make_percpus_absolute(void)
714{
715 unsigned int i;
716
717 for (i = 0; i < table_cnt; i++)
Ard Biesheuvel8c996942016-03-15 14:58:15 -0700718 if (symbol_in_range(&table[i], &percpu_range, 1)) {
719 /*
720 * Keep the 'A' override for percpu symbols to
721 * ensure consistent behavior compared to older
722 * versions of this tool.
723 */
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030724 table[i].sym[0] = 'A';
Ard Biesheuvel8c996942016-03-15 14:58:15 -0700725 table[i].percpu_absolute = 1;
726 }
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030727}
728
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700729/* find the minimum non-absolute symbol address */
730static void record_relative_base(void)
731{
732 unsigned int i;
733
734 relative_base = -1ULL;
735 for (i = 0; i < table_cnt; i++)
736 if (!symbol_absolute(&table[i]) &&
737 table[i].addr < relative_base)
738 relative_base = table[i].addr;
739}
740
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700741int main(int argc, char **argv)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700742{
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700743 if (argc >= 2) {
744 int i;
745 for (i = 1; i < argc; i++) {
746 if(strcmp(argv[i], "--all-symbols") == 0)
747 all_symbols = 1;
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030748 else if (strcmp(argv[i], "--absolute-percpu") == 0)
749 absolute_percpu = 1;
Masahiro Yamada534c9f22018-05-09 16:23:47 +0900750 else if (strcmp(argv[i], "--base-relative") == 0)
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700751 base_relative = 1;
752 else
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700753 usage();
754 }
755 } else if (argc != 1)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700756 usage();
757
758 read_map(stdin);
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030759 if (absolute_percpu)
760 make_percpus_absolute();
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700761 if (base_relative)
762 record_relative_base();
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100763 sort_symbols();
764 optimize_token_table();
Linus Torvalds1da177e2005-04-16 15:20:36 -0700765 write_src();
766
767 return 0;
768}