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	{ "_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
 57static struct addr_range percpu_range = {
 58	"__per_cpu_start", "__per_cpu_end", -1ULL, 0
 59};
 60
 61static struct sym_entry *table;
 62static unsigned int table_size, table_cnt;
 63static int all_symbols = 0;
 64static int absolute_percpu = 0;
 65static char symbol_prefix_char = '\0';
 66static int base_relative = 0;
 67
 68int token_profit[0x10000];
 69
 70/* the table that holds the result of the compression */
 71unsigned char best_table[256][2];
 72unsigned char best_table_len[256];
 73
 74
 75static void usage(void)
 76{
 77	fprintf(stderr, "Usage: kallsyms [--all-symbols] "
 78			"[--symbol-prefix=<prefix char>] "
 79			"[--base-relative] < in.map > out.S\n");
 80	exit(1);
 81}
 82
 83/*
 84 * This ignores the intensely annoying "mapping symbols" found
 85 * in ARM ELF files: $a, $t and $d.
 86 */
 87static inline int is_arm_mapping_symbol(const char *str)
 88{
 89	return str[0] == '$' && strchr("axtd", str[1])
 90	       && (str[2] == '\0' || str[2] == '.');
 91}
 92
 93static int check_symbol_range(const char *sym, unsigned long long addr,
 94			      struct addr_range *ranges, int entries)
 95{
 96	size_t i;
 97	struct addr_range *ar;
 98
 99	for (i = 0; i < entries; ++i) {
100		ar = &ranges[i];
101
102		if (strcmp(sym, ar->start_sym) == 0) {
103			ar->start = addr;
104			return 0;
105		} else if (strcmp(sym, ar->end_sym) == 0) {
106			ar->end = addr;
107			return 0;
108		}
109	}
110
111	return 1;
112}
113
114static int read_symbol(FILE *in, struct sym_entry *s)
115{
116	char str[500];
117	char *sym, stype;
118	int rc;
119
120	rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, str);
121	if (rc != 3) {
122		if (rc != EOF && fgets(str, 500, in) == NULL)
123			fprintf(stderr, "Read error or end of file.\n");
124		return -1;
125	}
126	if (strlen(str) > KSYM_NAME_LEN) {
127		fprintf(stderr, "Symbol %s too long for kallsyms (%zu vs %d).\n"
128				"Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n",
129			str, strlen(str), KSYM_NAME_LEN);
130		return -1;
131	}
132
133	sym = str;
134	/* skip prefix char */
135	if (symbol_prefix_char && str[0] == symbol_prefix_char)
136		sym++;
137
138	/* Ignore most absolute/undefined (?) symbols. */
139	if (strcmp(sym, "_text") == 0)
140		_text = s->addr;
141	else if (check_symbol_range(sym, s->addr, text_ranges,
142				    ARRAY_SIZE(text_ranges)) == 0)
143		/* nothing to do */;
144	else if (toupper(stype) == 'A')
145	{
146		/* Keep these useful absolute symbols */
147		if (strcmp(sym, "__kernel_syscall_via_break") &&
148		    strcmp(sym, "__kernel_syscall_via_epc") &&
149		    strcmp(sym, "__kernel_sigtramp") &&
150		    strcmp(sym, "__gp"))
151			return -1;
152
153	}
154	else if (toupper(stype) == 'U' ||
155		 is_arm_mapping_symbol(sym))
156		return -1;
157	/* exclude also MIPS ELF local symbols ($L123 instead of .L123) */
158	else if (str[0] == '$')
159		return -1;
160	/* exclude debugging symbols */
161	else if (stype == 'N' || stype == 'n')
162		return -1;
163
164	/* include the type field in the symbol name, so that it gets
165	 * compressed together */
166	s->len = strlen(str) + 1;
167	s->sym = malloc(s->len + 1);
168	if (!s->sym) {
169		fprintf(stderr, "kallsyms failure: "
170			"unable to allocate required amount of memory\n");
171		exit(EXIT_FAILURE);
172	}
173	strcpy((char *)s->sym + 1, str);
174	s->sym[0] = stype;
175
176	s->percpu_absolute = 0;
177
178	/* Record if we've found __per_cpu_start/end. */
179	check_symbol_range(sym, s->addr, &percpu_range, 1);
180
181	return 0;
182}
183
184static int symbol_in_range(struct sym_entry *s, struct addr_range *ranges,
185			   int entries)
186{
187	size_t i;
188	struct addr_range *ar;
189
190	for (i = 0; i < entries; ++i) {
191		ar = &ranges[i];
192
193		if (s->addr >= ar->start && s->addr <= ar->end)
194			return 1;
195	}
196
197	return 0;
198}
199
200static int symbol_valid(struct sym_entry *s)
201{
202	/* Symbols which vary between passes.  Passes 1 and 2 must have
203	 * identical symbol lists.  The kallsyms_* symbols below are only added
204	 * after pass 1, they would be included in pass 2 when --all-symbols is
205	 * specified so exclude them to get a stable symbol list.
206	 */
207	static char *special_symbols[] = {
208		"kallsyms_addresses",
209		"kallsyms_offsets",
210		"kallsyms_relative_base",
211		"kallsyms_num_syms",
212		"kallsyms_names",
213		"kallsyms_markers",
214		"kallsyms_token_table",
215		"kallsyms_token_index",
216
217	/* Exclude linker generated symbols which vary between passes */
218		"_SDA_BASE_",		/* ppc */
219		"_SDA2_BASE_",		/* ppc */
220		NULL };
221
222	static char *special_prefixes[] = {
223		"__crc_",		/* modversions */
224		"__efistub_",		/* arm64 EFI stub namespace */
225		NULL };
226
227	static char *special_suffixes[] = {
228		"_veneer",		/* arm */
229		"_from_arm",		/* arm */
230		"_from_thumb",		/* arm */
231		NULL };
232
233	int i;
234	char *sym_name = (char *)s->sym + 1;
235
236	/* skip prefix char */
237	if (symbol_prefix_char && *sym_name == symbol_prefix_char)
238		sym_name++;
239
240
241	/* if --all-symbols is not specified, then symbols outside the text
242	 * and inittext sections are discarded */
243	if (!all_symbols) {
244		if (symbol_in_range(s, text_ranges,
245				    ARRAY_SIZE(text_ranges)) == 0)
246			return 0;
247		/* Corner case.  Discard any symbols with the same value as
248		 * _etext _einittext; they can move between pass 1 and 2 when
249		 * the kallsyms data are added.  If these symbols move then
250		 * they may get dropped in pass 2, which breaks the kallsyms
251		 * rules.
252		 */
253		if ((s->addr == text_range_text->end &&
254				strcmp(sym_name,
255				       text_range_text->end_sym)) ||
256		    (s->addr == text_range_inittext->end &&
257				strcmp(sym_name,
258				       text_range_inittext->end_sym)))
259			return 0;
260	}
261
262	/* Exclude symbols which vary between passes. */
263	for (i = 0; special_symbols[i]; i++)
264		if (strcmp(sym_name, special_symbols[i]) == 0)
265			return 0;
266
267	for (i = 0; special_prefixes[i]; i++) {
268		int l = strlen(special_prefixes[i]);
269
270		if (l <= strlen(sym_name) &&
271		    strncmp(sym_name, special_prefixes[i], l) == 0)
272			return 0;
273	}
274
275	for (i = 0; special_suffixes[i]; i++) {
276		int l = strlen(sym_name) - strlen(special_suffixes[i]);
277
278		if (l >= 0 && strcmp(sym_name + l, special_suffixes[i]) == 0)
 
279			return 0;
280	}
281
282	return 1;
283}
284
285static void read_map(FILE *in)
286{
287	while (!feof(in)) {
288		if (table_cnt >= table_size) {
289			table_size += 10000;
290			table = realloc(table, sizeof(*table) * table_size);
291			if (!table) {
292				fprintf(stderr, "out of memory\n");
293				exit (1);
294			}
295		}
296		if (read_symbol(in, &table[table_cnt]) == 0) {
297			table[table_cnt].start_pos = table_cnt;
298			table_cnt++;
299		}
300	}
301}
302
303static void output_label(char *label)
304{
305	if (symbol_prefix_char)
306		printf(".globl %c%s\n", symbol_prefix_char, label);
307	else
308		printf(".globl %s\n", label);
309	printf("\tALGN\n");
310	if (symbol_prefix_char)
311		printf("%c%s:\n", symbol_prefix_char, label);
312	else
313		printf("%s:\n", label);
314}
315
316/* uncompress a compressed symbol. When this function is called, the best table
317 * might still be compressed itself, so the function needs to be recursive */
318static int expand_symbol(unsigned char *data, int len, char *result)
319{
320	int c, rlen, total=0;
321
322	while (len) {
323		c = *data;
324		/* if the table holds a single char that is the same as the one
325		 * we are looking for, then end the search */
326		if (best_table[c][0]==c && best_table_len[c]==1) {
327			*result++ = c;
328			total++;
329		} else {
330			/* if not, recurse and expand */
331			rlen = expand_symbol(best_table[c], best_table_len[c], result);
332			total += rlen;
333			result += rlen;
334		}
335		data++;
336		len--;
337	}
338	*result=0;
339
340	return total;
341}
342
343static int symbol_absolute(struct sym_entry *s)
344{
345	return s->percpu_absolute;
346}
347
348static void write_src(void)
349{
350	unsigned int i, k, off;
351	unsigned int best_idx[256];
352	unsigned int *markers;
353	char buf[KSYM_NAME_LEN];
354
355	printf("#include <asm/types.h>\n");
356	printf("#if BITS_PER_LONG == 64\n");
357	printf("#define PTR .quad\n");
358	printf("#define ALGN .align 8\n");
359	printf("#else\n");
360	printf("#define PTR .long\n");
361	printf("#define ALGN .align 4\n");
362	printf("#endif\n");
363
364	printf("\t.section .rodata, \"a\"\n");
365
366	/* Provide proper symbols relocatability by their relativeness
367	 * to a fixed anchor point in the runtime image, either '_text'
368	 * for absolute address tables, in which case the linker will
369	 * emit the final addresses at build time. Otherwise, use the
370	 * offset relative to the lowest value encountered of all relative
371	 * symbols, and emit non-relocatable fixed offsets that will be fixed
372	 * up at runtime.
373	 *
374	 * The symbol names cannot be used to construct normal symbol
375	 * references as the list of symbols contains symbols that are
376	 * declared static and are private to their .o files.  This prevents
377	 * .tmp_kallsyms.o or any other object from referencing them.
378	 */
379	if (!base_relative)
380		output_label("kallsyms_addresses");
381	else
382		output_label("kallsyms_offsets");
383
384	for (i = 0; i < table_cnt; i++) {
385		if (base_relative) {
386			long long offset;
387			int overflow;
388
389			if (!absolute_percpu) {
390				offset = table[i].addr - relative_base;
391				overflow = (offset < 0 || offset > UINT_MAX);
392			} else if (symbol_absolute(&table[i])) {
393				offset = table[i].addr;
394				overflow = (offset < 0 || offset > INT_MAX);
395			} else {
396				offset = relative_base - table[i].addr - 1;
397				overflow = (offset < INT_MIN || offset >= 0);
398			}
399			if (overflow) {
400				fprintf(stderr, "kallsyms failure: "
401					"%s symbol value %#llx out of range in relative mode\n",
402					symbol_absolute(&table[i]) ? "absolute" : "relative",
403					table[i].addr);
404				exit(EXIT_FAILURE);
405			}
406			printf("\t.long\t%#x\n", (int)offset);
407		} else if (!symbol_absolute(&table[i])) {
408			if (_text <= table[i].addr)
409				printf("\tPTR\t_text + %#llx\n",
410					table[i].addr - _text);
411			else
412				printf("\tPTR\t_text - %#llx\n",
413					_text - table[i].addr);
414		} else {
415			printf("\tPTR\t%#llx\n", table[i].addr);
416		}
417	}
418	printf("\n");
419
420	if (base_relative) {
421		output_label("kallsyms_relative_base");
422		printf("\tPTR\t_text - %#llx\n", _text - relative_base);
423		printf("\n");
424	}
425
426	output_label("kallsyms_num_syms");
427	printf("\tPTR\t%d\n", table_cnt);
428	printf("\n");
429
430	/* table of offset markers, that give the offset in the compressed stream
431	 * every 256 symbols */
432	markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
433	if (!markers) {
434		fprintf(stderr, "kallsyms failure: "
435			"unable to allocate required memory\n");
436		exit(EXIT_FAILURE);
437	}
438
439	output_label("kallsyms_names");
440	off = 0;
441	for (i = 0; i < table_cnt; i++) {
442		if ((i & 0xFF) == 0)
443			markers[i >> 8] = off;
444
445		printf("\t.byte 0x%02x", table[i].len);
446		for (k = 0; k < table[i].len; k++)
447			printf(", 0x%02x", table[i].sym[k]);
448		printf("\n");
449
450		off += table[i].len + 1;
451	}
452	printf("\n");
453
454	output_label("kallsyms_markers");
455	for (i = 0; i < ((table_cnt + 255) >> 8); i++)
456		printf("\tPTR\t%d\n", markers[i]);
457	printf("\n");
458
459	free(markers);
460
461	output_label("kallsyms_token_table");
462	off = 0;
463	for (i = 0; i < 256; i++) {
464		best_idx[i] = off;
465		expand_symbol(best_table[i], best_table_len[i], buf);
466		printf("\t.asciz\t\"%s\"\n", buf);
467		off += strlen(buf) + 1;
468	}
469	printf("\n");
470
471	output_label("kallsyms_token_index");
472	for (i = 0; i < 256; i++)
473		printf("\t.short\t%d\n", best_idx[i]);
474	printf("\n");
475}
476
477
478/* table lookup compression functions */
479
480/* count all the possible tokens in a symbol */
481static void learn_symbol(unsigned char *symbol, int len)
482{
483	int i;
484
485	for (i = 0; i < len - 1; i++)
486		token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
487}
488
489/* decrease the count for all the possible tokens in a symbol */
490static void forget_symbol(unsigned char *symbol, int len)
491{
492	int i;
493
494	for (i = 0; i < len - 1; i++)
495		token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
496}
497
498/* remove all the invalid symbols from the table and do the initial token count */
499static void build_initial_tok_table(void)
500{
501	unsigned int i, pos;
502
503	pos = 0;
504	for (i = 0; i < table_cnt; i++) {
505		if ( symbol_valid(&table[i]) ) {
506			if (pos != i)
507				table[pos] = table[i];
508			learn_symbol(table[pos].sym, table[pos].len);
509			pos++;
510		}
511	}
512	table_cnt = pos;
513}
514
515static void *find_token(unsigned char *str, int len, 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(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	memset(best_table, 0, sizeof(best_table));
621	memset(best_table_len, 0, sizeof(best_table_len));
622
623	for (i = 0; i < table_cnt; i++) {
624		for (j = 0; j < table[i].len; j++) {
625			c = table[i].sym[j];
626			best_table[c][0]=c;
627			best_table_len[c]=1;
628		}
629	}
630}
631
632static void optimize_token_table(void)
633{
634	build_initial_tok_table();
635
636	insert_real_symbols_in_table();
637
638	/* When valid symbol is not registered, exit to error */
639	if (!table_cnt) {
640		fprintf(stderr, "No valid symbol.\n");
641		exit(1);
642	}
643
644	optimize_result();
645}
646
647/* guess for "linker script provide" symbol */
648static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
649{
650	const char *symbol = (char *)se->sym + 1;
651	int len = se->len - 1;
652
653	if (len < 8)
654		return 0;
655
656	if (symbol[0] != '_' || symbol[1] != '_')
657		return 0;
658
659	/* __start_XXXXX */
660	if (!memcmp(symbol + 2, "start_", 6))
661		return 1;
662
663	/* __stop_XXXXX */
664	if (!memcmp(symbol + 2, "stop_", 5))
665		return 1;
666
667	/* __end_XXXXX */
668	if (!memcmp(symbol + 2, "end_", 4))
669		return 1;
670
671	/* __XXXXX_start */
672	if (!memcmp(symbol + len - 6, "_start", 6))
673		return 1;
674
675	/* __XXXXX_end */
676	if (!memcmp(symbol + len - 4, "_end", 4))
677		return 1;
678
679	return 0;
680}
681
682static int prefix_underscores_count(const char *str)
683{
684	const char *tail = str;
685
686	while (*tail == '_')
687		tail++;
688
689	return tail - str;
690}
691
692static int compare_symbols(const void *a, const void *b)
693{
694	const struct sym_entry *sa;
695	const struct sym_entry *sb;
696	int wa, wb;
697
698	sa = a;
699	sb = b;
700
701	/* sort by address first */
702	if (sa->addr > sb->addr)
703		return 1;
704	if (sa->addr < sb->addr)
705		return -1;
706
707	/* sort by "weakness" type */
708	wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
709	wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
710	if (wa != wb)
711		return wa - wb;
712
713	/* sort by "linker script provide" type */
714	wa = may_be_linker_script_provide_symbol(sa);
715	wb = may_be_linker_script_provide_symbol(sb);
716	if (wa != wb)
717		return wa - wb;
718
719	/* sort by the number of prefix underscores */
720	wa = prefix_underscores_count((const char *)sa->sym + 1);
721	wb = prefix_underscores_count((const char *)sb->sym + 1);
722	if (wa != wb)
723		return wa - wb;
724
725	/* sort by initial order, so that other symbols are left undisturbed */
726	return sa->start_pos - sb->start_pos;
727}
728
729static void sort_symbols(void)
730{
731	qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
732}
733
734static void make_percpus_absolute(void)
735{
736	unsigned int i;
737
738	for (i = 0; i < table_cnt; i++)
739		if (symbol_in_range(&table[i], &percpu_range, 1)) {
740			/*
741			 * Keep the 'A' override for percpu symbols to
742			 * ensure consistent behavior compared to older
743			 * versions of this tool.
744			 */
745			table[i].sym[0] = 'A';
746			table[i].percpu_absolute = 1;
747		}
748}
749
750/* find the minimum non-absolute symbol address */
751static void record_relative_base(void)
752{
753	unsigned int i;
754
755	relative_base = -1ULL;
756	for (i = 0; i < table_cnt; i++)
757		if (!symbol_absolute(&table[i]) &&
758		    table[i].addr < relative_base)
759			relative_base = table[i].addr;
760}
761
762int main(int argc, char **argv)
763{
764	if (argc >= 2) {
765		int i;
766		for (i = 1; i < argc; i++) {
767			if(strcmp(argv[i], "--all-symbols") == 0)
768				all_symbols = 1;
769			else if (strcmp(argv[i], "--absolute-percpu") == 0)
770				absolute_percpu = 1;
771			else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
772				char *p = &argv[i][16];
773				/* skip quote */
774				if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
775					p++;
776				symbol_prefix_char = *p;
777			} else if (strcmp(argv[i], "--base-relative") == 0)
778				base_relative = 1;
779			else
780				usage();
781		}
782	} else if (argc != 1)
783		usage();
784
785	read_map(stdin);
786	if (absolute_percpu)
787		make_percpus_absolute();
788	if (base_relative)
789		record_relative_base();
790	sort_symbols();
791	optimize_token_table();
792	write_src();
793
794	return 0;
795}