blob: 6402b0d36291414930932798347b1aa293f54057 [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';
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -070066static int base_relative = 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -070067
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070068int token_profit[0x10000];
Linus Torvalds1da177e2005-04-16 15:20:36 -070069
70/* the table that holds the result of the compression */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070071unsigned char best_table[256][2];
Linus Torvalds1da177e2005-04-16 15:20:36 -070072unsigned char best_table_len[256];
73
74
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070075static void usage(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -070076{
Ming Leif6537f22013-11-02 09:11:33 +103077 fprintf(stderr, "Usage: kallsyms [--all-symbols] "
78 "[--symbol-prefix=<prefix char>] "
79 "[--page-offset=<CONFIG_PAGE_OFFSET>] "
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -070080 "[--base-relative] < in.map > out.S\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -070081 exit(1);
82}
83
84/*
85 * This ignores the intensely annoying "mapping symbols" found
86 * in ARM ELF files: $a, $t and $d.
87 */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -070088static inline int is_arm_mapping_symbol(const char *str)
Linus Torvalds1da177e2005-04-16 15:20:36 -070089{
Kyle McMartin6c34f1f2014-09-16 22:37:18 +010090 return str[0] == '$' && strchr("axtd", str[1])
Linus Torvalds1da177e2005-04-16 15:20:36 -070091 && (str[2] == '\0' || str[2] == '.');
92}
93
Kees Cook78eb7152014-03-17 13:18:27 +103094static int check_symbol_range(const char *sym, unsigned long long addr,
95 struct addr_range *ranges, int entries)
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040096{
97 size_t i;
Kees Cook78eb7152014-03-17 13:18:27 +103098 struct addr_range *ar;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -040099
Kees Cook78eb7152014-03-17 13:18:27 +1030100 for (i = 0; i < entries; ++i) {
101 ar = &ranges[i];
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400102
Kees Cook78eb7152014-03-17 13:18:27 +1030103 if (strcmp(sym, ar->start_sym) == 0) {
104 ar->start = addr;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400105 return 0;
Kees Cook78eb7152014-03-17 13:18:27 +1030106 } else if (strcmp(sym, ar->end_sym) == 0) {
107 ar->end = addr;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400108 return 0;
109 }
110 }
111
112 return 1;
113}
114
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700115static int read_symbol(FILE *in, struct sym_entry *s)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700116{
117 char str[500];
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700118 char *sym, stype;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700119 int rc;
120
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700121 rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, str);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700122 if (rc != 3) {
Jean Sacrenef894872010-09-10 23:13:33 -0600123 if (rc != EOF && fgets(str, 500, in) == NULL)
124 fprintf(stderr, "Read error or end of file.\n");
Linus Torvalds1da177e2005-04-16 15:20:36 -0700125 return -1;
126 }
Andi Kleenf3462aa2013-10-23 15:07:53 +0200127 if (strlen(str) > KSYM_NAME_LEN) {
Fabio Estevam6f622592013-11-08 00:45:01 -0200128 fprintf(stderr, "Symbol %s too long for kallsyms (%zu vs %d).\n"
Masahiro Yamadabb66fc62014-06-10 19:08:13 +0900129 "Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n",
Andi Kleenf3462aa2013-10-23 15:07:53 +0200130 str, strlen(str), KSYM_NAME_LEN);
131 return -1;
132 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700133
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700134 sym = str;
135 /* skip prefix char */
136 if (symbol_prefix_char && str[0] == symbol_prefix_char)
137 sym++;
138
Linus Torvalds1da177e2005-04-16 15:20:36 -0700139 /* Ignore most absolute/undefined (?) symbols. */
Eric W. Biedermanfd593d12006-12-07 02:14:04 +0100140 if (strcmp(sym, "_text") == 0)
141 _text = s->addr;
Kees Cook78eb7152014-03-17 13:18:27 +1030142 else if (check_symbol_range(sym, s->addr, text_ranges,
143 ARRAY_SIZE(text_ranges)) == 0)
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400144 /* nothing to do */;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700145 else if (toupper(stype) == 'A')
Linus Torvalds1da177e2005-04-16 15:20:36 -0700146 {
147 /* Keep these useful absolute symbols */
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700148 if (strcmp(sym, "__kernel_syscall_via_break") &&
149 strcmp(sym, "__kernel_syscall_via_epc") &&
150 strcmp(sym, "__kernel_sigtramp") &&
151 strcmp(sym, "__gp"))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700152 return -1;
153
154 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700155 else if (toupper(stype) == 'U' ||
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700156 is_arm_mapping_symbol(sym))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700157 return -1;
Ralf Baechle6f00df22005-09-06 15:16:41 -0700158 /* exclude also MIPS ELF local symbols ($L123 instead of .L123) */
159 else if (str[0] == '$')
160 return -1;
Sam Ravnborgaab34ac2008-05-19 20:07:58 +0200161 /* exclude debugging symbols */
162 else if (stype == 'N')
163 return -1;
Vasily Gorbikc677e7a2019-06-28 19:22:47 +0200164 /* exclude s390 kasan local symbols */
165 else if (!strncmp(sym, ".LASANPC", 8))
166 return -1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700167
168 /* include the type field in the symbol name, so that it gets
169 * compressed together */
170 s->len = strlen(str) + 1;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700171 s->sym = malloc(s->len + 1);
Jesper Juhlf1a136e2006-03-25 03:07:46 -0800172 if (!s->sym) {
173 fprintf(stderr, "kallsyms failure: "
174 "unable to allocate required amount of memory\n");
175 exit(EXIT_FAILURE);
176 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700177 strcpy((char *)s->sym + 1, str);
178 s->sym[0] = stype;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700179
Ard Biesheuvel8c996942016-03-15 14:58:15 -0700180 s->percpu_absolute = 0;
181
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030182 /* Record if we've found __per_cpu_start/end. */
183 check_symbol_range(sym, s->addr, &percpu_range, 1);
184
Linus Torvalds1da177e2005-04-16 15:20:36 -0700185 return 0;
186}
187
Kees Cook78eb7152014-03-17 13:18:27 +1030188static int symbol_in_range(struct sym_entry *s, struct addr_range *ranges,
189 int entries)
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400190{
191 size_t i;
Kees Cook78eb7152014-03-17 13:18:27 +1030192 struct addr_range *ar;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400193
Kees Cook78eb7152014-03-17 13:18:27 +1030194 for (i = 0; i < entries; ++i) {
195 ar = &ranges[i];
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400196
Kees Cook78eb7152014-03-17 13:18:27 +1030197 if (s->addr >= ar->start && s->addr <= ar->end)
Mike Frysingerac6ca5c2009-06-15 07:52:48 -0400198 return 1;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400199 }
200
Mike Frysingerac6ca5c2009-06-15 07:52:48 -0400201 return 0;
Mike Frysinger17b1f0d2009-06-08 19:12:13 -0400202}
203
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700204static int symbol_valid(struct sym_entry *s)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700205{
206 /* Symbols which vary between passes. Passes 1 and 2 must have
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100207 * identical symbol lists. The kallsyms_* symbols below are only added
208 * after pass 1, they would be included in pass 2 when --all-symbols is
209 * specified so exclude them to get a stable symbol list.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700210 */
211 static char *special_symbols[] = {
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100212 "kallsyms_addresses",
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700213 "kallsyms_offsets",
214 "kallsyms_relative_base",
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100215 "kallsyms_num_syms",
216 "kallsyms_names",
217 "kallsyms_markers",
218 "kallsyms_token_table",
219 "kallsyms_token_index",
220
Linus Torvalds1da177e2005-04-16 15:20:36 -0700221 /* Exclude linker generated symbols which vary between passes */
222 "_SDA_BASE_", /* ppc */
223 "_SDA2_BASE_", /* ppc */
224 NULL };
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200225
226 static char *special_suffixes[] = {
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200227 "_veneer", /* arm */
Ard Biesheuvelb9b74be2016-04-01 14:33:48 +0100228 "_from_arm", /* arm */
229 "_from_thumb", /* arm */
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200230 NULL };
231
Linus Torvalds1da177e2005-04-16 15:20:36 -0700232 int i;
Ard Biesheuvelbd8b22d2015-03-30 15:20:31 +0200233 char *sym_name = (char *)s->sym + 1;
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++;
Masahiro Yamada9abb8f62019-11-24 01:04:30 +0900501 } else {
502 free(table[i].sym);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700503 }
504 }
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700505 table_cnt = pos;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700506}
507
Paulo Marques7c5d2492007-06-20 18:09:00 +0100508static void *find_token(unsigned char *str, int len, unsigned char *token)
509{
510 int i;
511
512 for (i = 0; i < len - 1; i++) {
513 if (str[i] == token[0] && str[i+1] == token[1])
514 return &str[i];
515 }
516 return NULL;
517}
518
Linus Torvalds1da177e2005-04-16 15:20:36 -0700519/* replace a given token in all the valid symbols. Use the sampled symbols
520 * to update the counts */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700521static void compress_symbols(unsigned char *str, int idx)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700522{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700523 unsigned int i, len, size;
524 unsigned char *p1, *p2;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700525
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700526 for (i = 0; i < table_cnt; i++) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700527
528 len = table[i].len;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700529 p1 = table[i].sym;
530
531 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100532 p2 = find_token(p1, len, str);
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700533 if (!p2) continue;
534
535 /* decrease the counts for this symbol's tokens */
536 forget_symbol(table[i].sym, len);
537
538 size = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700539
540 do {
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700541 *p2 = idx;
542 p2++;
543 size -= (p2 - p1);
544 memmove(p2, p2 + 1, size);
545 p1 = p2;
546 len--;
547
548 if (size < 2) break;
549
Linus Torvalds1da177e2005-04-16 15:20:36 -0700550 /* find the token on the symbol */
Paulo Marques7c5d2492007-06-20 18:09:00 +0100551 p2 = find_token(p1, size, str);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700552
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700553 } while (p2);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700554
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700555 table[i].len = len;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700556
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700557 /* increase the counts for this symbol's new tokens */
558 learn_symbol(table[i].sym, len);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700559 }
560}
561
562/* search the token with the maximum profit */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700563static int find_best_token(void)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700564{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700565 int i, best, bestprofit;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700566
567 bestprofit=-10000;
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700568 best = 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700569
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700570 for (i = 0; i < 0x10000; i++) {
571 if (token_profit[i] > bestprofit) {
572 best = i;
573 bestprofit = token_profit[i];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700574 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700575 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700576 return best;
577}
578
579/* this is the core of the algorithm: calculate the "best" table */
580static void optimize_result(void)
581{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700582 int i, best;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700583
584 /* using the '\0' symbol last allows compress_symbols to use standard
585 * fast string functions */
586 for (i = 255; i >= 0; i--) {
587
588 /* if this table slot is empty (it is not used by an actual
589 * original char code */
590 if (!best_table_len[i]) {
591
592 /* find the token with the breates profit value */
593 best = find_best_token();
Xiaochen Wange0a04b12011-05-01 11:41:41 +0800594 if (token_profit[best] == 0)
595 break;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700596
597 /* place it in the "best" table */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700598 best_table_len[i] = 2;
599 best_table[i][0] = best & 0xFF;
600 best_table[i][1] = (best >> 8) & 0xFF;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700601
602 /* replace this token in all the valid symbols */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700603 compress_symbols(best_table[i], i);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700604 }
605 }
606}
607
608/* start by placing the symbols that are actually used on the table */
609static void insert_real_symbols_in_table(void)
610{
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700611 unsigned int i, j, c;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700612
613 memset(best_table, 0, sizeof(best_table));
614 memset(best_table_len, 0, sizeof(best_table_len));
615
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700616 for (i = 0; i < table_cnt; i++) {
617 for (j = 0; j < table[i].len; j++) {
618 c = table[i].sym[j];
619 best_table[c][0]=c;
620 best_table_len[c]=1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700621 }
622 }
623}
624
625static void optimize_token_table(void)
626{
Linus Torvalds1da177e2005-04-16 15:20:36 -0700627 build_initial_tok_table();
628
629 insert_real_symbols_in_table();
630
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700631 /* When valid symbol is not registered, exit to error */
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700632 if (!table_cnt) {
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700633 fprintf(stderr, "No valid symbol.\n");
634 exit(1);
635 }
636
Linus Torvalds1da177e2005-04-16 15:20:36 -0700637 optimize_result();
638}
639
Lai Jiangshanb478b782009-03-13 15:10:26 +0800640/* guess for "linker script provide" symbol */
641static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
642{
643 const char *symbol = (char *)se->sym + 1;
644 int len = se->len - 1;
645
646 if (len < 8)
647 return 0;
648
649 if (symbol[0] != '_' || symbol[1] != '_')
650 return 0;
651
652 /* __start_XXXXX */
653 if (!memcmp(symbol + 2, "start_", 6))
654 return 1;
655
656 /* __stop_XXXXX */
657 if (!memcmp(symbol + 2, "stop_", 5))
658 return 1;
659
660 /* __end_XXXXX */
661 if (!memcmp(symbol + 2, "end_", 4))
662 return 1;
663
664 /* __XXXXX_start */
665 if (!memcmp(symbol + len - 6, "_start", 6))
666 return 1;
667
668 /* __XXXXX_end */
669 if (!memcmp(symbol + len - 4, "_end", 4))
670 return 1;
671
672 return 0;
673}
674
675static int prefix_underscores_count(const char *str)
676{
677 const char *tail = str;
678
Paul Mundta9ece532009-09-22 16:44:12 -0700679 while (*tail == '_')
Lai Jiangshanb478b782009-03-13 15:10:26 +0800680 tail++;
681
682 return tail - str;
683}
684
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800685static int compare_symbols(const void *a, const void *b)
686{
687 const struct sym_entry *sa;
688 const struct sym_entry *sb;
689 int wa, wb;
690
691 sa = a;
692 sb = b;
693
694 /* sort by address first */
695 if (sa->addr > sb->addr)
696 return 1;
697 if (sa->addr < sb->addr)
698 return -1;
699
700 /* sort by "weakness" type */
701 wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
702 wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
703 if (wa != wb)
704 return wa - wb;
705
Lai Jiangshanb478b782009-03-13 15:10:26 +0800706 /* sort by "linker script provide" type */
707 wa = may_be_linker_script_provide_symbol(sa);
708 wb = may_be_linker_script_provide_symbol(sb);
709 if (wa != wb)
710 return wa - wb;
711
712 /* sort by the number of prefix underscores */
713 wa = prefix_underscores_count((const char *)sa->sym + 1);
714 wb = prefix_underscores_count((const char *)sb->sym + 1);
715 if (wa != wb)
716 return wa - wb;
717
Paulo Marquesf2df3f62008-02-06 01:37:33 -0800718 /* sort by initial order, so that other symbols are left undisturbed */
719 return sa->start_pos - sb->start_pos;
720}
721
722static void sort_symbols(void)
723{
724 qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
725}
Linus Torvalds1da177e2005-04-16 15:20:36 -0700726
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030727static void make_percpus_absolute(void)
728{
729 unsigned int i;
730
731 for (i = 0; i < table_cnt; i++)
Ard Biesheuvel8c996942016-03-15 14:58:15 -0700732 if (symbol_in_range(&table[i], &percpu_range, 1)) {
733 /*
734 * Keep the 'A' override for percpu symbols to
735 * ensure consistent behavior compared to older
736 * versions of this tool.
737 */
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030738 table[i].sym[0] = 'A';
Ard Biesheuvel8c996942016-03-15 14:58:15 -0700739 table[i].percpu_absolute = 1;
740 }
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030741}
742
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700743/* find the minimum non-absolute symbol address */
744static void record_relative_base(void)
745{
746 unsigned int i;
747
748 relative_base = -1ULL;
749 for (i = 0; i < table_cnt; i++)
750 if (!symbol_absolute(&table[i]) &&
751 table[i].addr < relative_base)
752 relative_base = table[i].addr;
753}
754
Paulo Marquesb3dbb4e2005-09-06 15:16:31 -0700755int main(int argc, char **argv)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700756{
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700757 if (argc >= 2) {
758 int i;
759 for (i = 1; i < argc; i++) {
760 if(strcmp(argv[i], "--all-symbols") == 0)
761 all_symbols = 1;
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030762 else if (strcmp(argv[i], "--absolute-percpu") == 0)
763 absolute_percpu = 1;
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700764 else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
765 char *p = &argv[i][16];
766 /* skip quote */
767 if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
768 p++;
769 symbol_prefix_char = *p;
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700770 } else if (strcmp(argv[i], "--base-relative") == 0)
771 base_relative = 1;
772 else
Yoshinori Sato41f11a42005-05-01 08:59:06 -0700773 usage();
774 }
775 } else if (argc != 1)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700776 usage();
777
778 read_map(stdin);
Rusty Russellc6bda7c2014-03-17 14:05:46 +1030779 if (absolute_percpu)
780 make_percpus_absolute();
Ard Biesheuvel2213e9a2016-03-15 14:58:19 -0700781 if (base_relative)
782 record_relative_base();
Sam Ravnborg2ea03892009-01-14 21:38:20 +0100783 sort_symbols();
784 optimize_token_table();
Linus Torvalds1da177e2005-04-16 15:20:36 -0700785 write_src();
786
787 return 0;
788}