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	/* Symbol names that exactly match to the following are ignored.*/
 86	static const char * const ignored_symbols[] = {
 87		/*
 88		 * Symbols which vary between passes. Passes 1 and 2 must have
 89		 * identical symbol lists. The kallsyms_* symbols below are
 90		 * only added after pass 1, they would be included in pass 2
 91		 * when --all-symbols is specified so exclude them to get a
 92		 * stable symbol list.
 93		 */
 94		"kallsyms_addresses",
 95		"kallsyms_offsets",
 96		"kallsyms_relative_base",
 97		"kallsyms_num_syms",
 98		"kallsyms_names",
 99		"kallsyms_markers",
100		"kallsyms_token_table",
101		"kallsyms_token_index",
102		/* Exclude linker generated symbols which vary between passes */
103		"_SDA_BASE_",		/* ppc */
104		"_SDA2_BASE_",		/* ppc */
105		NULL
106	};
107
108	/* Symbol names that begin with the following are ignored.*/
109	static const char * const ignored_prefixes[] = {
110		"$",			/* local symbols for ARM, MIPS, etc. */
111		".LASANPC",		/* s390 kasan local symbols */
112		"__crc_",		/* modversions */
113		"__efistub_",		/* arm64 EFI stub namespace */
114		"__kvm_nvhe_",		/* arm64 non-VHE KVM namespace */
115		NULL
116	};
117
118	/* Symbol names that end with the following are ignored.*/
119	static const char * const ignored_suffixes[] = {
120		"_from_arm",		/* arm */
121		"_from_thumb",		/* arm */
122		"_veneer",		/* arm */
123		NULL
124	};
125
126	/* Symbol names that contain the following are ignored.*/
127	static const char * const ignored_matches[] = {
128		".long_branch.",	/* ppc stub */
129		".plt_branch.",		/* ppc stub */
130		NULL
131	};
132
133	const char * const *p;
134
135	for (p = ignored_symbols; *p; p++)
136		if (!strcmp(name, *p))
137			return true;
138
139	for (p = ignored_prefixes; *p; p++)
140		if (!strncmp(name, *p, strlen(*p)))
141			return true;
142
143	for (p = ignored_suffixes; *p; p++) {
144		int l = strlen(name) - strlen(*p);
145
146		if (l >= 0 && !strcmp(name + l, *p))
147			return true;
148	}
149
150	for (p = ignored_matches; *p; p++) {
151		if (strstr(name, *p))
152			return true;
153	}
154
155	if (type == 'U' || type == 'u')
156		return true;
157	/* exclude debugging symbols */
158	if (type == 'N' || type == 'n')
159		return true;
160
161	if (toupper(type) == 'A') {
162		/* Keep these useful absolute symbols */
163		if (strcmp(name, "__kernel_syscall_via_break") &&
164		    strcmp(name, "__kernel_syscall_via_epc") &&
165		    strcmp(name, "__kernel_sigtramp") &&
166		    strcmp(name, "__gp"))
167			return true;
168	}
169
170	return false;
171}
172
173static void check_symbol_range(const char *sym, unsigned long long addr,
174			       struct addr_range *ranges, int entries)
175{
176	size_t i;
177	struct addr_range *ar;
178
179	for (i = 0; i < entries; ++i) {
180		ar = &ranges[i];
181
182		if (strcmp(sym, ar->start_sym) == 0) {
183			ar->start = addr;
184			return;
185		} else if (strcmp(sym, ar->end_sym) == 0) {
186			ar->end = addr;
187			return;
188		}
189	}
 
 
190}
191
192static struct sym_entry *read_symbol(FILE *in)
193{
194	char name[500], type;
195	unsigned long long addr;
196	unsigned int len;
197	struct sym_entry *sym;
198	int rc;
199
200	rc = fscanf(in, "%llx %c %499s\n", &addr, &type, name);
201	if (rc != 3) {
202		if (rc != EOF && fgets(name, 500, in) == NULL)
203			fprintf(stderr, "Read error or end of file.\n");
204		return NULL;
205	}
206	if (strlen(name) >= KSYM_NAME_LEN) {
207		fprintf(stderr, "Symbol %s too long for kallsyms (%zu >= %d).\n"
208				"Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n",
209			name, strlen(name), KSYM_NAME_LEN);
210		return NULL;
211	}
212
213	if (strcmp(name, "_text") == 0)
214		_text = addr;
 
 
215
216	/* Ignore most absolute/undefined (?) symbols. */
217	if (is_ignored_symbol(name, type))
218		return NULL;
 
 
 
 
 
 
 
 
 
 
219
220	check_symbol_range(name, addr, text_ranges, ARRAY_SIZE(text_ranges));
221	check_symbol_range(name, addr, &percpu_range, 1);
 
 
 
 
 
 
 
 
222
223	/* include the type field in the symbol name, so that it gets
224	 * compressed together */
225
226	len = strlen(name) + 1;
227
228	sym = malloc(sizeof(*sym) + len + 1);
229	if (!sym) {
230		fprintf(stderr, "kallsyms failure: "
231			"unable to allocate required amount of memory\n");
232		exit(EXIT_FAILURE);
233	}
234	sym->addr = addr;
235	sym->len = len;
236	sym->sym[0] = type;
237	strcpy(sym_name(sym), name);
238	sym->percpu_absolute = 0;
239
240	return sym;
241}
242
243static int symbol_in_range(const struct sym_entry *s,
244			   const struct addr_range *ranges, int entries)
245{
246	size_t i;
247	const struct addr_range *ar;
248
249	for (i = 0; i < entries; ++i) {
250		ar = &ranges[i];
251
252		if (s->addr >= ar->start && s->addr <= ar->end)
253			return 1;
254	}
255
256	return 0;
257}
258
259static int symbol_valid(const struct sym_entry *s)
260{
261	const char *name = sym_name(s);
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
262
263	/* if --all-symbols is not specified, then symbols outside the text
264	 * and inittext sections are discarded */
265	if (!all_symbols) {
266		if (symbol_in_range(s, text_ranges,
267				    ARRAY_SIZE(text_ranges)) == 0)
268			return 0;
269		/* Corner case.  Discard any symbols with the same value as
270		 * _etext _einittext; they can move between pass 1 and 2 when
271		 * the kallsyms data are added.  If these symbols move then
272		 * they may get dropped in pass 2, which breaks the kallsyms
273		 * rules.
274		 */
275		if ((s->addr == text_range_text->end &&
276		     strcmp(name, text_range_text->end_sym)) ||
277		    (s->addr == text_range_inittext->end &&
278		     strcmp(name, text_range_inittext->end_sym)))
279			return 0;
280	}
281
282	return 1;
283}
284
285/* remove all the invalid symbols from the table */
286static void shrink_table(void)
287{
288	unsigned int i, pos;
289
290	pos = 0;
291	for (i = 0; i < table_cnt; i++) {
292		if (symbol_valid(table[i])) {
293			if (pos != i)
294				table[pos] = table[i];
295			pos++;
296		} else {
297			free(table[i]);
298		}
299	}
300	table_cnt = pos;
301
302	/* When valid symbol is not registered, exit to error */
303	if (!table_cnt) {
304		fprintf(stderr, "No valid symbol.\n");
305		exit(1);
306	}
307}
308
309static void read_map(FILE *in)
310{
311	struct sym_entry *sym;
312
313	while (!feof(in)) {
314		sym = read_symbol(in);
315		if (!sym)
316			continue;
317
318		sym->start_pos = table_cnt;
319
320		if (table_cnt >= table_size) {
321			table_size += 10000;
322			table = realloc(table, sizeof(*table) * table_size);
323			if (!table) {
324				fprintf(stderr, "out of memory\n");
325				exit (1);
326			}
327		}
328
329		table[table_cnt++] = sym;
 
 
330	}
331}
332
333static void output_label(const char *label)
334{
335	printf(".globl %s\n", label);
 
 
 
336	printf("\tALGN\n");
337	printf("%s:\n", label);
338}
339
340/* Provide proper symbols relocatability by their '_text' relativeness. */
341static void output_address(unsigned long long addr)
342{
343	if (_text <= addr)
344		printf("\tPTR\t_text + %#llx\n", addr - _text);
345	else
346		printf("\tPTR\t_text - %#llx\n", _text - addr);
347}
348
349/* uncompress a compressed symbol. When this function is called, the best table
350 * might still be compressed itself, so the function needs to be recursive */
351static int expand_symbol(const unsigned char *data, int len, char *result)
352{
353	int c, rlen, total=0;
354
355	while (len) {
356		c = *data;
357		/* if the table holds a single char that is the same as the one
358		 * we are looking for, then end the search */
359		if (best_table[c][0]==c && best_table_len[c]==1) {
360			*result++ = c;
361			total++;
362		} else {
363			/* if not, recurse and expand */
364			rlen = expand_symbol(best_table[c], best_table_len[c], result);
365			total += rlen;
366			result += rlen;
367		}
368		data++;
369		len--;
370	}
371	*result=0;
372
373	return total;
374}
375
376static int symbol_absolute(const struct sym_entry *s)
377{
378	return s->percpu_absolute;
379}
380
381static void write_src(void)
382{
383	unsigned int i, k, off;
384	unsigned int best_idx[256];
385	unsigned int *markers;
386	char buf[KSYM_NAME_LEN];
387
388	printf("#include <asm/bitsperlong.h>\n");
389	printf("#if BITS_PER_LONG == 64\n");
390	printf("#define PTR .quad\n");
391	printf("#define ALGN .balign 8\n");
392	printf("#else\n");
393	printf("#define PTR .long\n");
394	printf("#define ALGN .balign 4\n");
395	printf("#endif\n");
396
397	printf("\t.section .rodata, \"a\"\n");
398
399	if (!base_relative)
400		output_label("kallsyms_addresses");
401	else
402		output_label("kallsyms_offsets");
403
 
 
 
404	for (i = 0; i < table_cnt; i++) {
405		if (base_relative) {
406			/*
407			 * Use the offset relative to the lowest value
408			 * encountered of all relative symbols, and emit
409			 * non-relocatable fixed offsets that will be fixed
410			 * up at runtime.
411			 */
412
413			long long offset;
414			int overflow;
415
416			if (!absolute_percpu) {
417				offset = table[i]->addr - relative_base;
418				overflow = (offset < 0 || offset > UINT_MAX);
419			} else if (symbol_absolute(table[i])) {
420				offset = table[i]->addr;
421				overflow = (offset < 0 || offset > INT_MAX);
422			} else {
423				offset = relative_base - table[i]->addr - 1;
424				overflow = (offset < INT_MIN || offset >= 0);
425			}
426			if (overflow) {
427				fprintf(stderr, "kallsyms failure: "
428					"%s symbol value %#llx out of range in relative mode\n",
429					symbol_absolute(table[i]) ? "absolute" : "relative",
430					table[i]->addr);
431				exit(EXIT_FAILURE);
432			}
433			printf("\t.long\t%#x\n", (int)offset);
434		} else if (!symbol_absolute(table[i])) {
435			output_address(table[i]->addr);
436		} else {
437			printf("\tPTR\t%#llx\n", table[i]->addr);
438		}
439	}
440	printf("\n");
441
442	if (base_relative) {
443		output_label("kallsyms_relative_base");
444		output_address(relative_base);
445		printf("\n");
446	}
447
448	output_label("kallsyms_num_syms");
449	printf("\t.long\t%u\n", table_cnt);
450	printf("\n");
451
452	/* table of offset markers, that give the offset in the compressed stream
453	 * every 256 symbols */
454	markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
455	if (!markers) {
456		fprintf(stderr, "kallsyms failure: "
457			"unable to allocate required memory\n");
458		exit(EXIT_FAILURE);
459	}
460
461	output_label("kallsyms_names");
462	off = 0;
463	for (i = 0; i < table_cnt; i++) {
464		if ((i & 0xFF) == 0)
465			markers[i >> 8] = off;
466
467		printf("\t.byte 0x%02x", table[i]->len);
468		for (k = 0; k < table[i]->len; k++)
469			printf(", 0x%02x", table[i]->sym[k]);
470		printf("\n");
471
472		off += table[i]->len + 1;
473	}
474	printf("\n");
475
476	output_label("kallsyms_markers");
477	for (i = 0; i < ((table_cnt + 255) >> 8); i++)
478		printf("\t.long\t%u\n", markers[i]);
479	printf("\n");
480
481	free(markers);
482
483	output_label("kallsyms_token_table");
484	off = 0;
485	for (i = 0; i < 256; i++) {
486		best_idx[i] = off;
487		expand_symbol(best_table[i], best_table_len[i], buf);
488		printf("\t.asciz\t\"%s\"\n", buf);
489		off += strlen(buf) + 1;
490	}
491	printf("\n");
492
493	output_label("kallsyms_token_index");
494	for (i = 0; i < 256; i++)
495		printf("\t.short\t%d\n", best_idx[i]);
496	printf("\n");
497}
498
499
500/* table lookup compression functions */
501
502/* count all the possible tokens in a symbol */
503static void learn_symbol(const unsigned char *symbol, int len)
504{
505	int i;
506
507	for (i = 0; i < len - 1; i++)
508		token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
509}
510
511/* decrease the count for all the possible tokens in a symbol */
512static void forget_symbol(const unsigned char *symbol, int len)
513{
514	int i;
515
516	for (i = 0; i < len - 1; i++)
517		token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
518}
519
520/* do the initial token count */
521static void build_initial_tok_table(void)
522{
523	unsigned int i;
524
525	for (i = 0; i < table_cnt; i++)
526		learn_symbol(table[i]->sym, table[i]->len);
 
 
 
 
 
 
 
 
527}
528
529static unsigned char *find_token(unsigned char *str, int len,
530				 const unsigned char *token)
531{
532	int i;
533
534	for (i = 0; i < len - 1; i++) {
535		if (str[i] == token[0] && str[i+1] == token[1])
536			return &str[i];
537	}
538	return NULL;
539}
540
541/* replace a given token in all the valid symbols. Use the sampled symbols
542 * to update the counts */
543static void compress_symbols(const unsigned char *str, int idx)
544{
545	unsigned int i, len, size;
546	unsigned char *p1, *p2;
547
548	for (i = 0; i < table_cnt; i++) {
549
550		len = table[i]->len;
551		p1 = table[i]->sym;
552
553		/* find the token on the symbol */
554		p2 = find_token(p1, len, str);
555		if (!p2) continue;
556
557		/* decrease the counts for this symbol's tokens */
558		forget_symbol(table[i]->sym, len);
559
560		size = len;
561
562		do {
563			*p2 = idx;
564			p2++;
565			size -= (p2 - p1);
566			memmove(p2, p2 + 1, size);
567			p1 = p2;
568			len--;
569
570			if (size < 2) break;
571
572			/* find the token on the symbol */
573			p2 = find_token(p1, size, str);
574
575		} while (p2);
576
577		table[i]->len = len;
578
579		/* increase the counts for this symbol's new tokens */
580		learn_symbol(table[i]->sym, len);
581	}
582}
583
584/* search the token with the maximum profit */
585static int find_best_token(void)
586{
587	int i, best, bestprofit;
588
589	bestprofit=-10000;
590	best = 0;
591
592	for (i = 0; i < 0x10000; i++) {
593		if (token_profit[i] > bestprofit) {
594			best = i;
595			bestprofit = token_profit[i];
596		}
597	}
598	return best;
599}
600
601/* this is the core of the algorithm: calculate the "best" table */
602static void optimize_result(void)
603{
604	int i, best;
605
606	/* using the '\0' symbol last allows compress_symbols to use standard
607	 * fast string functions */
608	for (i = 255; i >= 0; i--) {
609
610		/* if this table slot is empty (it is not used by an actual
611		 * original char code */
612		if (!best_table_len[i]) {
613
614			/* find the token with the best profit value */
615			best = find_best_token();
616			if (token_profit[best] == 0)
617				break;
618
619			/* place it in the "best" table */
620			best_table_len[i] = 2;
621			best_table[i][0] = best & 0xFF;
622			best_table[i][1] = (best >> 8) & 0xFF;
623
624			/* replace this token in all the valid symbols */
625			compress_symbols(best_table[i], i);
626		}
627	}
628}
629
630/* start by placing the symbols that are actually used on the table */
631static void insert_real_symbols_in_table(void)
632{
633	unsigned int i, j, c;
634
 
 
 
635	for (i = 0; i < table_cnt; i++) {
636		for (j = 0; j < table[i]->len; j++) {
637			c = table[i]->sym[j];
638			best_table[c][0]=c;
639			best_table_len[c]=1;
640		}
641	}
642}
643
644static void optimize_token_table(void)
645{
646	build_initial_tok_table();
647
648	insert_real_symbols_in_table();
649
 
 
 
 
 
 
650	optimize_result();
651}
652
653/* guess for "linker script provide" symbol */
654static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
655{
656	const char *symbol = sym_name(se);
657	int len = se->len - 1;
658
659	if (len < 8)
660		return 0;
661
662	if (symbol[0] != '_' || symbol[1] != '_')
663		return 0;
664
665	/* __start_XXXXX */
666	if (!memcmp(symbol + 2, "start_", 6))
667		return 1;
668
669	/* __stop_XXXXX */
670	if (!memcmp(symbol + 2, "stop_", 5))
671		return 1;
672
673	/* __end_XXXXX */
674	if (!memcmp(symbol + 2, "end_", 4))
675		return 1;
676
677	/* __XXXXX_start */
678	if (!memcmp(symbol + len - 6, "_start", 6))
679		return 1;
680
681	/* __XXXXX_end */
682	if (!memcmp(symbol + len - 4, "_end", 4))
683		return 1;
684
685	return 0;
686}
687
 
 
 
 
 
 
 
 
 
 
688static int compare_symbols(const void *a, const void *b)
689{
690	const struct sym_entry *sa = *(const struct sym_entry **)a;
691	const struct sym_entry *sb = *(const struct sym_entry **)b;
692	int wa, wb;
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
706	/* 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 = strspn(sym_name(sa), "_");
714	wb = strspn(sym_name(sb), "_");
715	if (wa != wb)
716		return wa - wb;
717
718	/* 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(table[0]), compare_symbols);
725}
726
727static void make_percpus_absolute(void)
728{
729	unsigned int i;
730
731	for (i = 0; i < table_cnt; i++)
732		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			 */
738			table[i]->sym[0] = 'A';
739			table[i]->percpu_absolute = 1;
740		}
741}
742
743/* find the minimum non-absolute symbol address */
744static void record_relative_base(void)
745{
746	unsigned int i;
747
748	for (i = 0; i < table_cnt; i++)
749		if (!symbol_absolute(table[i])) {
750			/*
751			 * The table is sorted by address.
752			 * Take the first non-absolute symbol value.
753			 */
754			relative_base = table[i]->addr;
755			return;
756		}
757}
758
759int main(int argc, char **argv)
760{
761	if (argc >= 2) {
762		int i;
763		for (i = 1; i < argc; i++) {
764			if(strcmp(argv[i], "--all-symbols") == 0)
765				all_symbols = 1;
766			else if (strcmp(argv[i], "--absolute-percpu") == 0)
767				absolute_percpu = 1;
768			else if (strcmp(argv[i], "--base-relative") == 0)
769				base_relative = 1;
770			else
 
 
771				usage();
772		}
773	} else if (argc != 1)
774		usage();
775
776	read_map(stdin);
777	shrink_table();
778	if (absolute_percpu)
779		make_percpus_absolute();
780	sort_symbols();
781	if (base_relative)
782		record_relative_base();
783	optimize_token_table();
784	write_src();
785
786	return 0;
787}