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 <stdio.h>
 22#include <stdlib.h>
 23#include <string.h>
 24#include <ctype.h>
 25#include <limits.h>
 26
 27#ifndef ARRAY_SIZE
 28#define ARRAY_SIZE(arr) (sizeof(arr) / sizeof(arr[0]))
 29#endif
 30
 31#define KSYM_NAME_LEN		128
 32
 33struct sym_entry {
 34	unsigned long long addr;
 35	unsigned int len;
 36	unsigned int start_pos;
 37	unsigned char *sym;
 38	unsigned int percpu_absolute;
 39};
 40
 41struct addr_range {
 42	const char *start_sym, *end_sym;
 43	unsigned long long start, end;
 44};
 45
 46static unsigned long long _text;
 47static unsigned long long relative_base;
 48static struct addr_range text_ranges[] = {
 49	{ "_stext",     "_etext"     },
 50	{ "_sinittext", "_einittext" },
 
 
 51};
 52#define text_range_text     (&text_ranges[0])
 53#define text_range_inittext (&text_ranges[1])
 54
 55static struct addr_range percpu_range = {
 56	"__per_cpu_start", "__per_cpu_end", -1ULL, 0
 57};
 58
 59static struct sym_entry *table;
 60static unsigned int table_size, table_cnt;
 61static int all_symbols = 0;
 62static int absolute_percpu = 0;
 63static int base_relative = 0;
 64
 65static int token_profit[0x10000];
 66
 67/* the table that holds the result of the compression */
 68static unsigned char best_table[256][2];
 69static unsigned char best_table_len[256];
 70
 71
 72static void usage(void)
 73{
 74	fprintf(stderr, "Usage: kallsyms [--all-symbols] "
 75			"[--base-relative] < in.map > out.S\n");
 76	exit(1);
 77}
 78
 79/*
 80 * This ignores the intensely annoying "mapping symbols" found
 81 * in ARM ELF files: $a, $t and $d.
 82 */
 83static int is_arm_mapping_symbol(const char *str)
 84{
 85	return str[0] == '$' && strchr("axtd", str[1])
 86	       && (str[2] == '\0' || str[2] == '.');
 87}
 88
 89static int check_symbol_range(const char *sym, unsigned long long addr,
 90			      struct addr_range *ranges, int entries)
 91{
 92	size_t i;
 93	struct addr_range *ar;
 94
 95	for (i = 0; i < entries; ++i) {
 96		ar = &ranges[i];
 97
 98		if (strcmp(sym, ar->start_sym) == 0) {
 99			ar->start = addr;
100			return 0;
101		} else if (strcmp(sym, ar->end_sym) == 0) {
102			ar->end = addr;
103			return 0;
104		}
105	}
106
107	return 1;
108}
109
110static int read_symbol(FILE *in, struct sym_entry *s)
111{
112	char sym[500], stype;
 
113	int rc;
114
115	rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, sym);
116	if (rc != 3) {
117		if (rc != EOF && fgets(sym, 500, in) == NULL)
118			fprintf(stderr, "Read error or end of file.\n");
119		return -1;
120	}
121	if (strlen(sym) >= KSYM_NAME_LEN) {
122		fprintf(stderr, "Symbol %s too long for kallsyms (%zu >= %d).\n"
123				"Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n",
124			sym, strlen(sym), KSYM_NAME_LEN);
125		return -1;
126	}
127
128	/* Ignore most absolute/undefined (?) symbols. */
129	if (strcmp(sym, "_text") == 0)
130		_text = s->addr;
131	else if (check_symbol_range(sym, s->addr, text_ranges,
132				    ARRAY_SIZE(text_ranges)) == 0)
133		/* nothing to do */;
134	else if (toupper(stype) == 'A')
135	{
136		/* Keep these useful absolute symbols */
137		if (strcmp(sym, "__kernel_syscall_via_break") &&
138		    strcmp(sym, "__kernel_syscall_via_epc") &&
139		    strcmp(sym, "__kernel_sigtramp") &&
140		    strcmp(sym, "__gp"))
141			return -1;
142
143	}
144	else if (toupper(stype) == 'U' ||
145		 is_arm_mapping_symbol(sym))
146		return -1;
147	/* exclude also MIPS ELF local symbols ($L123 instead of .L123) */
148	else if (sym[0] == '$')
149		return -1;
150	/* exclude debugging symbols */
151	else if (stype == 'N' || stype == 'n')
152		return -1;
153	/* exclude s390 kasan local symbols */
154	else if (!strncmp(sym, ".LASANPC", 8))
155		return -1;
156
157	/* include the type field in the symbol name, so that it gets
158	 * compressed together */
159	s->len = strlen(sym) + 1;
160	s->sym = malloc(s->len + 1);
161	if (!s->sym) {
162		fprintf(stderr, "kallsyms failure: "
163			"unable to allocate required amount of memory\n");
164		exit(EXIT_FAILURE);
165	}
166	strcpy((char *)s->sym + 1, sym);
167	s->sym[0] = stype;
168
169	s->percpu_absolute = 0;
170
171	/* Record if we've found __per_cpu_start/end. */
172	check_symbol_range(sym, s->addr, &percpu_range, 1);
173
174	return 0;
175}
176
177static int symbol_in_range(struct sym_entry *s, struct addr_range *ranges,
178			   int entries)
179{
180	size_t i;
181	struct addr_range *ar;
182
183	for (i = 0; i < entries; ++i) {
184		ar = &ranges[i];
185
186		if (s->addr >= ar->start && s->addr <= ar->end)
187			return 1;
188	}
189
190	return 0;
191}
192
193static int symbol_valid(struct sym_entry *s)
194{
195	/* Symbols which vary between passes.  Passes 1 and 2 must have
196	 * 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.
199	 */
200	static char *special_symbols[] = {
201		"kallsyms_addresses",
202		"kallsyms_offsets",
203		"kallsyms_relative_base",
204		"kallsyms_num_syms",
205		"kallsyms_names",
206		"kallsyms_markers",
207		"kallsyms_token_table",
208		"kallsyms_token_index",
209
210	/* Exclude linker generated symbols which vary between passes */
211		"_SDA_BASE_",		/* ppc */
212		"_SDA2_BASE_",		/* ppc */
213		NULL };
214
215	static char *special_prefixes[] = {
216		"__crc_",		/* modversions */
217		"__efistub_",		/* arm64 EFI stub namespace */
218		NULL };
219
220	static char *special_suffixes[] = {
221		"_veneer",		/* arm */
222		"_from_arm",		/* arm */
223		"_from_thumb",		/* arm */
224		NULL };
225
226	int i;
227	char *sym_name = (char *)s->sym + 1;
 
 
 
 
228
229	/* if --all-symbols is not specified, then symbols outside the text
230	 * and inittext sections are discarded */
231	if (!all_symbols) {
232		if (symbol_in_range(s, text_ranges,
233				    ARRAY_SIZE(text_ranges)) == 0)
234			return 0;
235		/* Corner case.  Discard any symbols with the same value as
236		 * _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.
240		 */
241		if ((s->addr == text_range_text->end &&
242				strcmp(sym_name,
243				       text_range_text->end_sym)) ||
244		    (s->addr == text_range_inittext->end &&
245				strcmp(sym_name,
246				       text_range_inittext->end_sym)))
247			return 0;
248	}
249
250	/* Exclude symbols which vary between passes. */
251	for (i = 0; special_symbols[i]; i++)
252		if (strcmp(sym_name, special_symbols[i]) == 0)
253			return 0;
254
255	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
263	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
270	return 1;
271}
272
273static void read_map(FILE *in)
274{
275	while (!feof(in)) {
276		if (table_cnt >= table_size) {
277			table_size += 10000;
278			table = realloc(table, sizeof(*table) * table_size);
279			if (!table) {
280				fprintf(stderr, "out of memory\n");
281				exit (1);
282			}
283		}
284		if (read_symbol(in, &table[table_cnt]) == 0) {
285			table[table_cnt].start_pos = table_cnt;
286			table_cnt++;
287		}
288	}
289}
290
291static void output_label(char *label)
292{
293	printf(".globl %s\n", label);
 
 
 
294	printf("\tALGN\n");
295	printf("%s:\n", label);
 
 
 
296}
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
325static int symbol_absolute(struct sym_entry *s)
326{
327	return s->percpu_absolute;
328}
329
330static void write_src(void)
331{
332	unsigned int i, k, off;
333	unsigned int best_idx[256];
334	unsigned int *markers;
335	char buf[KSYM_NAME_LEN];
336
337	printf("#include <asm/bitsperlong.h>\n");
338	printf("#if BITS_PER_LONG == 64\n");
339	printf("#define PTR .quad\n");
340	printf("#define ALGN .balign 8\n");
341	printf("#else\n");
342	printf("#define PTR .long\n");
343	printf("#define ALGN .balign 4\n");
344	printf("#endif\n");
345
346	printf("\t.section .rodata, \"a\"\n");
347
348	/* 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.
360	 */
361	if (!base_relative)
362		output_label("kallsyms_addresses");
363	else
364		output_label("kallsyms_offsets");
365
366	for (i = 0; i < table_cnt; i++) {
367		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])) {
390			if (_text <= table[i].addr)
391				printf("\tPTR\t_text + %#llx\n",
392					table[i].addr - _text);
393			else
394				printf("\tPTR\t_text - %#llx\n",
395					_text - table[i].addr);
396		} else {
397			printf("\tPTR\t%#llx\n", table[i].addr);
398		}
399	}
400	printf("\n");
401
402	if (base_relative) {
403		output_label("kallsyms_relative_base");
404		printf("\tPTR\t_text - %#llx\n", _text - relative_base);
405		printf("\n");
406	}
407
408	output_label("kallsyms_num_syms");
409	printf("\t.long\t%u\n", table_cnt);
410	printf("\n");
411
412	/* table of offset markers, that give the offset in the compressed stream
413	 * every 256 symbols */
414	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	}
420
421	output_label("kallsyms_names");
422	off = 0;
423	for (i = 0; i < table_cnt; i++) {
424		if ((i & 0xFF) == 0)
425			markers[i >> 8] = off;
426
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;
433	}
434	printf("\n");
435
436	output_label("kallsyms_markers");
437	for (i = 0; i < ((table_cnt + 255) >> 8); i++)
438		printf("\t.long\t%u\n", markers[i]);
439	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;
447		expand_symbol(best_table[i], best_table_len[i], buf);
448		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
462/* 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++)
468		token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
469}
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++)
477		token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
478}
479
480/* remove all the invalid symbols from the table and do the initial token count */
481static void build_initial_tok_table(void)
482{
483	unsigned int i, pos;
484
485	pos = 0;
486	for (i = 0; i < table_cnt; i++) {
487		if ( symbol_valid(&table[i]) ) {
488			if (pos != i)
489				table[pos] = table[i];
490			learn_symbol(table[pos].sym, table[pos].len);
491			pos++;
492		}
493	}
494	table_cnt = pos;
495}
496
497static 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
508/* replace a given token in all the valid symbols. Use the sampled symbols
509 * to update the counts */
510static void compress_symbols(unsigned char *str, int idx)
511{
512	unsigned int i, len, size;
513	unsigned char *p1, *p2;
514
515	for (i = 0; i < table_cnt; i++) {
516
517		len = table[i].len;
518		p1 = table[i].sym;
519
520		/* find the token on the symbol */
521		p2 = find_token(p1, len, str);
522		if (!p2) continue;
523
524		/* decrease the counts for this symbol's tokens */
525		forget_symbol(table[i].sym, len);
526
527		size = len;
528
529		do {
530			*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
539			/* find the token on the symbol */
540			p2 = find_token(p1, size, str);
541
542		} while (p2);
543
544		table[i].len = len;
545
546		/* increase the counts for this symbol's new tokens */
547		learn_symbol(table[i].sym, len);
548	}
549}
550
551/* search the token with the maximum profit */
552static int find_best_token(void)
553{
554	int i, best, bestprofit;
555
556	bestprofit=-10000;
557	best = 0;
558
559	for (i = 0; i < 0x10000; i++) {
560		if (token_profit[i] > bestprofit) {
561			best = i;
562			bestprofit = token_profit[i];
563		}
564	}
565	return best;
566}
567
568/* this is the core of the algorithm: calculate the "best" table */
569static void optimize_result(void)
570{
571	int i, best;
572
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
581			/* find the token with the best profit value */
582			best = find_best_token();
583			if (token_profit[best] == 0)
584				break;
585
586			/* place it in the "best" table */
587			best_table_len[i] = 2;
588			best_table[i][0] = best & 0xFF;
589			best_table[i][1] = (best >> 8) & 0xFF;
590
591			/* replace this token in all the valid symbols */
592			compress_symbols(best_table[i], i);
593		}
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{
600	unsigned int i, j, c;
601
 
 
 
602	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;
607		}
608	}
609}
610
611static void optimize_token_table(void)
612{
613	build_initial_tok_table();
614
615	insert_real_symbols_in_table();
616
617	/* When valid symbol is not registered, exit to error */
618	if (!table_cnt) {
619		fprintf(stderr, "No valid symbol.\n");
620		exit(1);
621	}
622
623	optimize_result();
624}
625
626/* 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
665	while (*tail == '_')
666		tail++;
667
668	return tail - str;
669}
670
671static 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
692	/* 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
704	/* 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}
712
713static void make_percpus_absolute(void)
714{
715	unsigned int i;
716
717	for (i = 0; i < table_cnt; i++)
718		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			 */
724			table[i].sym[0] = 'A';
725			table[i].percpu_absolute = 1;
726		}
727}
728
729/* 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
741int main(int argc, char **argv)
742{
743	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;
748			else if (strcmp(argv[i], "--absolute-percpu") == 0)
749				absolute_percpu = 1;
750			else if (strcmp(argv[i], "--base-relative") == 0)
751				base_relative = 1;
752			else
 
 
753				usage();
754		}
755	} else if (argc != 1)
756		usage();
757
758	read_map(stdin);
759	if (absolute_percpu)
760		make_percpus_absolute();
761	if (base_relative)
762		record_relative_base();
763	sort_symbols();
764	optimize_token_table();
765	write_src();
766
767	return 0;
768}