/scripts/kallsyms.c

  1/* 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 *
 10 *      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 <stdbool.h>
 22#include <stdio.h>
 23#include <stdlib.h>
 24#include <string.h>
 25#include <ctype.h>
 26#include <limits.h>
 27
 28#define ARRAY_SIZE(arr) (sizeof(arr) / sizeof(arr[0]))
 29
 30#define KSYM_NAME_LEN		128
 31
 32struct sym_entry {
 33	unsigned long long addr;
 34	unsigned int len;
 35	unsigned int start_pos;
 36	unsigned int percpu_absolute;
 37	unsigned char sym[];
 38};
 39
 40struct addr_range {
 41	const char *start_sym, *end_sym;
 42	unsigned long long start, end;
 43};
 44
 45static unsigned long long _text;
 46static unsigned long long relative_base;
 47static struct addr_range text_ranges[] = {
 48	{ "_stext",     "_etext"     },
 49	{ "_sinittext", "_einittext" },
 50};
 51#define text_range_text     (&text_ranges[0])
 52#define text_range_inittext (&text_ranges[1])
 53
 54static struct addr_range percpu_range = {
 55	"__per_cpu_start", "__per_cpu_end", -1ULL, 0
 56};
 57
 58static struct sym_entry **table;
 59static unsigned int table_size, table_cnt;
 60static int all_symbols;
 61static int absolute_percpu;
 62static int base_relative;
 63
 64static int token_profit[0x10000];
 65
 66/* the table that holds the result of the compression */
 67static unsigned char best_table[256][2];
 68static unsigned char best_table_len[256];
 69
 70
 71static void usage(void)
 72{
 73	fprintf(stderr, "Usage: kallsyms [--all-symbols] "
 74			"[--base-relative] < in.map > out.S\n");
 75	exit(1);
 76}
 77
 78static char *sym_name(const struct sym_entry *s)
 79{
 80	return (char *)s->sym + 1;
 81}
 82
 83static bool is_ignored_symbol(const char *name, char type)
 84{
 85	static const char * const ignored_symbols[] = {
 86		/*
 87		 * Symbols which vary between passes. Passes 1 and 2 must have
 88		 * identical symbol lists. The kallsyms_* symbols below are
 89		 * only added after pass 1, they would be included in pass 2
 90		 * when --all-symbols is specified so exclude them to get a
 91		 * stable symbol list.
 92		 */
 93		"kallsyms_addresses",
 94		"kallsyms_offsets",
 95		"kallsyms_relative_base",
 96		"kallsyms_num_syms",
 97		"kallsyms_names",
 98		"kallsyms_markers",
 99		"kallsyms_token_table",
100		"kallsyms_token_index",
101		/* Exclude linker generated symbols which vary between passes */
102		"_SDA_BASE_",		/* ppc */
103		"_SDA2_BASE_",		/* ppc */
104		NULL
105	};
106
107	static const char * const ignored_prefixes[] = {
108		"$",			/* local symbols for ARM, MIPS, etc. */
109		".LASANPC",		/* s390 kasan local symbols */
110		"__crc_",		/* modversions */
111		"__efistub_",		/* arm64 EFI stub namespace */
112		NULL
113	};
114
115	static const char * const ignored_suffixes[] = {
116		"_from_arm",		/* arm */
117		"_from_thumb",		/* arm */
118		"_veneer",		/* arm */
119		NULL
120	};
121
122	const char * const *p;
123
124	/* Exclude symbols which vary between passes. */
125	for (p = ignored_symbols; *p; p++)
126		if (!strcmp(name, *p))
127			return true;
128
129	for (p = ignored_prefixes; *p; p++)
130		if (!strncmp(name, *p, strlen(*p)))
131			return true;
132
133	for (p = ignored_suffixes; *p; p++) {
134		int l = strlen(name) - strlen(*p);
135
136		if (l >= 0 && !strcmp(name + l, *p))
137			return true;
138	}
139
140	if (type == 'U' || type == 'u')
141		return true;
142	/* exclude debugging symbols */
143	if (type == 'N' || type == 'n')
144		return true;
145
146	if (toupper(type) == 'A') {
147		/* Keep these useful absolute symbols */
148		if (strcmp(name, "__kernel_syscall_via_break") &&
149		    strcmp(name, "__kernel_syscall_via_epc") &&
150		    strcmp(name, "__kernel_sigtramp") &&
151		    strcmp(name, "__gp"))
152			return true;
153	}
154
155	return false;
156}
157
158static void check_symbol_range(const char *sym, unsigned long long addr,
159			       struct addr_range *ranges, int entries)
160{
161	size_t i;
162	struct addr_range *ar;
163
164	for (i = 0; i < entries; ++i) {
165		ar = &ranges[i];
166
167		if (strcmp(sym, ar->start_sym) == 0) {
168			ar->start = addr;
169			return;
170		} else if (strcmp(sym, ar->end_sym) == 0) {
171			ar->end = addr;
172			return;
173		}
174	}
175}
176
177static struct sym_entry *read_symbol(FILE *in)
178{
179	char name[500], type;
180	unsigned long long addr;
181	unsigned int len;
182	struct sym_entry *sym;
183	int rc;
184
185	rc = fscanf(in, "%llx %c %499s\n", &addr, &type, name);
186	if (rc != 3) {
187		if (rc != EOF && fgets(name, 500, in) == NULL)
188			fprintf(stderr, "Read error or end of file.\n");
189		return NULL;
190	}
191	if (strlen(name) >= KSYM_NAME_LEN) {
192		fprintf(stderr, "Symbol %s too long for kallsyms (%zu >= %d).\n"
193				"Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n",
194			name, strlen(name), KSYM_NAME_LEN);
195		return NULL;
196	}
197
198	if (strcmp(name, "_text") == 0)
199		_text = addr;
200
201	/* Ignore most absolute/undefined (?) symbols. */
202	if (is_ignored_symbol(name, type))
203		return NULL;
204
205	check_symbol_range(name, addr, text_ranges, ARRAY_SIZE(text_ranges));
206	check_symbol_range(name, addr, &percpu_range, 1);
207
208	/* include the type field in the symbol name, so that it gets
209	 * compressed together */
210
211	len = strlen(name) + 1;
212
213	sym = malloc(sizeof(*sym) + len + 1);
214	if (!sym) {
215		fprintf(stderr, "kallsyms failure: "
216			"unable to allocate required amount of memory\n");
217		exit(EXIT_FAILURE);
218	}
219	sym->addr = addr;
220	sym->len = len;
221	sym->sym[0] = type;
222	strcpy(sym_name(sym), name);
223	sym->percpu_absolute = 0;
224
225	return sym;
226}
227
228static int symbol_in_range(const struct sym_entry *s,
229			   const struct addr_range *ranges, int entries)
230{
231	size_t i;
232	const struct addr_range *ar;
233
234	for (i = 0; i < entries; ++i) {
235		ar = &ranges[i];
236
237		if (s->addr >= ar->start && s->addr <= ar->end)
238			return 1;
239	}
240
241	return 0;
242}
243
244static int symbol_valid(const struct sym_entry *s)
245{
246	const char *name = sym_name(s);
247
248	/* if --all-symbols is not specified, then symbols outside the text
249	 * and inittext sections are discarded */
250	if (!all_symbols) {
251		if (symbol_in_range(s, text_ranges,
252				    ARRAY_SIZE(text_ranges)) == 0)
253			return 0;
254		/* Corner case.  Discard any symbols with the same value as
255		 * _etext _einittext; they can move between pass 1 and 2 when
256		 * the kallsyms data are added.  If these symbols move then
257		 * they may get dropped in pass 2, which breaks the kallsyms
258		 * rules.
259		 */
260		if ((s->addr == text_range_text->end &&
261		     strcmp(name, text_range_text->end_sym)) ||
262		    (s->addr == text_range_inittext->end &&
263		     strcmp(name, text_range_inittext->end_sym)))
264			return 0;
265	}
266
267	return 1;
268}
269
270/* remove all the invalid symbols from the table */
271static void shrink_table(void)
272{
273	unsigned int i, pos;
274
275	pos = 0;
276	for (i = 0; i < table_cnt; i++) {
277		if (symbol_valid(table[i])) {
278			if (pos != i)
279				table[pos] = table[i];
280			pos++;
281		} else {
282			free(table[i]);
283		}
284	}
285	table_cnt = pos;
286
287	/* When valid symbol is not registered, exit to error */
288	if (!table_cnt) {
289		fprintf(stderr, "No valid symbol.\n");
290		exit(1);
291	}
292}
293
294static void read_map(FILE *in)
295{
296	struct sym_entry *sym;
297
298	while (!feof(in)) {
299		sym = read_symbol(in);
300		if (!sym)
301			continue;
302
303		sym->start_pos = table_cnt;
304
305		if (table_cnt >= table_size) {
306			table_size += 10000;
307			table = realloc(table, sizeof(*table) * table_size);
308			if (!table) {
309				fprintf(stderr, "out of memory\n");
310				exit (1);
311			}
312		}
313
314		table[table_cnt++] = sym;
315	}
316}
317
318static void output_label(const char *label)
319{
320	printf(".globl %s\n", label);
321	printf("\tALGN\n");
322	printf("%s:\n", label);
323}
324
325/* Provide proper symbols relocatability by their '_text' relativeness. */
326static void output_address(unsigned long long addr)
327{
328	if (_text <= addr)
329		printf("\tPTR\t_text + %#llx\n", addr - _text);
330	else
331		printf("\tPTR\t_text - %#llx\n", _text - addr);
332}
333
334/* uncompress a compressed symbol. When this function is called, the best table
335 * might still be compressed itself, so the function needs to be recursive */
336static int expand_symbol(const unsigned char *data, int len, char *result)
337{
338	int c, rlen, total=0;
339
340	while (len) {
341		c = *data;
342		/* if the table holds a single char that is the same as the one
343		 * we are looking for, then end the search */
344		if (best_table[c][0]==c && best_table_len[c]==1) {
345			*result++ = c;
346			total++;
347		} else {
348			/* if not, recurse and expand */
349			rlen = expand_symbol(best_table[c], best_table_len[c], result);
350			total += rlen;
351			result += rlen;
352		}
353		data++;
354		len--;
355	}
356	*result=0;
357
358	return total;
359}
360
361static int symbol_absolute(const struct sym_entry *s)
362{
363	return s->percpu_absolute;
364}
365
366static void write_src(void)
367{
368	unsigned int i, k, off;
369	unsigned int best_idx[256];
370	unsigned int *markers;
371	char buf[KSYM_NAME_LEN];
372
373	printf("#include <asm/bitsperlong.h>\n");
374	printf("#if BITS_PER_LONG == 64\n");
375	printf("#define PTR .quad\n");
376	printf("#define ALGN .balign 8\n");
377	printf("#else\n");
378	printf("#define PTR .long\n");
379	printf("#define ALGN .balign 4\n");
380	printf("#endif\n");
381
382	printf("\t.section .rodata, \"a\"\n");
383
384	if (!base_relative)
385		output_label("kallsyms_addresses");
386	else
387		output_label("kallsyms_offsets");
388
389	for (i = 0; i < table_cnt; i++) {
390		if (base_relative) {
391			/*
392			 * Use the offset relative to the lowest value
393			 * encountered of all relative symbols, and emit
394			 * non-relocatable fixed offsets that will be fixed
395			 * up at runtime.
396			 */
397
398			long long offset;
399			int overflow;
400
401			if (!absolute_percpu) {
402				offset = table[i]->addr - relative_base;
403				overflow = (offset < 0 || offset > UINT_MAX);
404			} else if (symbol_absolute(table[i])) {
405				offset = table[i]->addr;
406				overflow = (offset < 0 || offset > INT_MAX);
407			} else {
408				offset = relative_base - table[i]->addr - 1;
409				overflow = (offset < INT_MIN || offset >= 0);
410			}
411			if (overflow) {
412				fprintf(stderr, "kallsyms failure: "
413					"%s symbol value %#llx out of range in relative mode\n",
414					symbol_absolute(table[i]) ? "absolute" : "relative",
415					table[i]->addr);
416				exit(EXIT_FAILURE);
417			}
418			printf("\t.long\t%#x\n", (int)offset);
419		} else if (!symbol_absolute(table[i])) {
420			output_address(table[i]->addr);
421		} else {
422			printf("\tPTR\t%#llx\n", table[i]->addr);
423		}
424	}
425	printf("\n");
426
427	if (base_relative) {
428		output_label("kallsyms_relative_base");
429		output_address(relative_base);
430		printf("\n");
431	}
432
433	output_label("kallsyms_num_syms");
434	printf("\t.long\t%u\n", table_cnt);
435	printf("\n");
436
437	/* table of offset markers, that give the offset in the compressed stream
438	 * every 256 symbols */
439	markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
440	if (!markers) {
441		fprintf(stderr, "kallsyms failure: "
442			"unable to allocate required memory\n");
443		exit(EXIT_FAILURE);
444	}
445
446	output_label("kallsyms_names");
447	off = 0;
448	for (i = 0; i < table_cnt; i++) {
449		if ((i & 0xFF) == 0)
450			markers[i >> 8] = off;
451
452		printf("\t.byte 0x%02x", table[i]->len);
453		for (k = 0; k < table[i]->len; k++)
454			printf(", 0x%02x", table[i]->sym[k]);
455		printf("\n");
456
457		off += table[i]->len + 1;
458	}
459	printf("\n");
460
461	output_label("kallsyms_markers");
462	for (i = 0; i < ((table_cnt + 255) >> 8); i++)
463		printf("\t.long\t%u\n", markers[i]);
464	printf("\n");
465
466	free(markers);
467
468	output_label("kallsyms_token_table");
469	off = 0;
470	for (i = 0; i < 256; i++) {
471		best_idx[i] = off;
472		expand_symbol(best_table[i], best_table_len[i], buf);
473		printf("\t.asciz\t\"%s\"\n", buf);
474		off += strlen(buf) + 1;
475	}
476	printf("\n");
477
478	output_label("kallsyms_token_index");
479	for (i = 0; i < 256; i++)
480		printf("\t.short\t%d\n", best_idx[i]);
481	printf("\n");
482}
483
484
485/* table lookup compression functions */
486
487/* count all the possible tokens in a symbol */
488static void learn_symbol(const unsigned char *symbol, int len)
489{
490	int i;
491
492	for (i = 0; i < len - 1; i++)
493		token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
494}
495
496/* decrease the count for all the possible tokens in a symbol */
497static void forget_symbol(const unsigned char *symbol, int len)
498{
499	int i;
500
501	for (i = 0; i < len - 1; i++)
502		token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
503}
504
505/* do the initial token count */
506static void build_initial_tok_table(void)
507{
508	unsigned int i;
509
510	for (i = 0; i < table_cnt; i++)
511		learn_symbol(table[i]->sym, table[i]->len);
512}
513
514static unsigned char *find_token(unsigned char *str, int len,
515				 const unsigned char *token)
516{
517	int i;
518
519	for (i = 0; i < len - 1; i++) {
520		if (str[i] == token[0] && str[i+1] == token[1])
521			return &str[i];
522	}
523	return NULL;
524}
525
526/* replace a given token in all the valid symbols. Use the sampled symbols
527 * to update the counts */
528static void compress_symbols(const unsigned char *str, int idx)
529{
530	unsigned int i, len, size;
531	unsigned char *p1, *p2;
532
533	for (i = 0; i < table_cnt; i++) {
534
535		len = table[i]->len;
536		p1 = table[i]->sym;
537
538		/* find the token on the symbol */
539		p2 = find_token(p1, len, str);
540		if (!p2) continue;
541
542		/* decrease the counts for this symbol's tokens */
543		forget_symbol(table[i]->sym, len);
544
545		size = len;
546
547		do {
548			*p2 = idx;
549			p2++;
550			size -= (p2 - p1);
551			memmove(p2, p2 + 1, size);
552			p1 = p2;
553			len--;
554
555			if (size < 2) break;
556
557			/* find the token on the symbol */
558			p2 = find_token(p1, size, str);
559
560		} while (p2);
561
562		table[i]->len = len;
563
564		/* increase the counts for this symbol's new tokens */
565		learn_symbol(table[i]->sym, len);
566	}
567}
568
569/* search the token with the maximum profit */
570static int find_best_token(void)
571{
572	int i, best, bestprofit;
573
574	bestprofit=-10000;
575	best = 0;
576
577	for (i = 0; i < 0x10000; i++) {
578		if (token_profit[i] > bestprofit) {
579			best = i;
580			bestprofit = token_profit[i];
581		}
582	}
583	return best;
584}
585
586/* this is the core of the algorithm: calculate the "best" table */
587static void optimize_result(void)
588{
589	int i, best;
590
591	/* using the '\0' symbol last allows compress_symbols to use standard
592	 * fast string functions */
593	for (i = 255; i >= 0; i--) {
594
595		/* if this table slot is empty (it is not used by an actual
596		 * original char code */
597		if (!best_table_len[i]) {
598
599			/* find the token with the best profit value */
600			best = find_best_token();
601			if (token_profit[best] == 0)
602				break;
603
604			/* place it in the "best" table */
605			best_table_len[i] = 2;
606			best_table[i][0] = best & 0xFF;
607			best_table[i][1] = (best >> 8) & 0xFF;
608
609			/* replace this token in all the valid symbols */
610			compress_symbols(best_table[i], i);
611		}
612	}
613}
614
615/* start by placing the symbols that are actually used on the table */
616static void insert_real_symbols_in_table(void)
617{
618	unsigned int i, j, c;
619
620	for (i = 0; i < table_cnt; i++) {
621		for (j = 0; j < table[i]->len; j++) {
622			c = table[i]->sym[j];
623			best_table[c][0]=c;
624			best_table_len[c]=1;
625		}
626	}
627}
628
629static void optimize_token_table(void)
630{
631	build_initial_tok_table();
632
633	insert_real_symbols_in_table();
634
635	optimize_result();
636}
637
638/* guess for "linker script provide" symbol */
639static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
640{
641	const char *symbol = sym_name(se);
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 compare_symbols(const void *a, const void *b)
674{
675	const struct sym_entry *sa = *(const struct sym_entry **)a;
676	const struct sym_entry *sb = *(const struct sym_entry **)b;
677	int wa, wb;
678
679	/* sort by address first */
680	if (sa->addr > sb->addr)
681		return 1;
682	if (sa->addr < sb->addr)
683		return -1;
684
685	/* sort by "weakness" type */
686	wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
687	wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
688	if (wa != wb)
689		return wa - wb;
690
691	/* sort by "linker script provide" type */
692	wa = may_be_linker_script_provide_symbol(sa);
693	wb = may_be_linker_script_provide_symbol(sb);
694	if (wa != wb)
695		return wa - wb;
696
697	/* sort by the number of prefix underscores */
698	wa = strspn(sym_name(sa), "_");
699	wb = strspn(sym_name(sb), "_");
700	if (wa != wb)
701		return wa - wb;
702
703	/* sort by initial order, so that other symbols are left undisturbed */
704	return sa->start_pos - sb->start_pos;
705}
706
707static void sort_symbols(void)
708{
709	qsort(table, table_cnt, sizeof(table[0]), compare_symbols);
710}
711
712static void make_percpus_absolute(void)
713{
714	unsigned int i;
715
716	for (i = 0; i < table_cnt; i++)
717		if (symbol_in_range(table[i], &percpu_range, 1)) {
718			/*
719			 * Keep the 'A' override for percpu symbols to
720			 * ensure consistent behavior compared to older
721			 * versions of this tool.
722			 */
723			table[i]->sym[0] = 'A';
724			table[i]->percpu_absolute = 1;
725		}
726}
727
728/* find the minimum non-absolute symbol address */
729static void record_relative_base(void)
730{
731	unsigned int i;
732
733	for (i = 0; i < table_cnt; i++)
734		if (!symbol_absolute(table[i])) {
735			/*
736			 * The table is sorted by address.
737			 * Take the first non-absolute symbol value.
738			 */
739			relative_base = table[i]->addr;
740			return;
741		}
742}
743
744int main(int argc, char **argv)
745{
746	if (argc >= 2) {
747		int i;
748		for (i = 1; i < argc; i++) {
749			if(strcmp(argv[i], "--all-symbols") == 0)
750				all_symbols = 1;
751			else if (strcmp(argv[i], "--absolute-percpu") == 0)
752				absolute_percpu = 1;
753			else if (strcmp(argv[i], "--base-relative") == 0)
754				base_relative = 1;
755			else
756				usage();
757		}
758	} else if (argc != 1)
759		usage();
760
761	read_map(stdin);
762	shrink_table();
763	if (absolute_percpu)
764		make_percpus_absolute();
765	sort_symbols();
766	if (base_relative)
767		record_relative_base();
768	optimize_token_table();
769	write_src();
770
771	return 0;
772}