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 unsigned long long kernel_start_addr = 0;
 67static int base_relative = 0;
 68
 69int token_profit[0x10000];
 70
 71/* the table that holds the result of the compression */
 72unsigned char best_table[256][2];
 73unsigned char best_table_len[256];
 74
 75
 76static void usage(void)
 77{
 78	fprintf(stderr, "Usage: kallsyms [--all-symbols] "
 79			"[--symbol-prefix=<prefix char>] "
 80			"[--page-offset=<CONFIG_PAGE_OFFSET>] "
 81			"[--base-relative] < in.map > out.S\n");
 82	exit(1);
 83}
 84
 85/*
 86 * This ignores the intensely annoying "mapping symbols" found
 87 * in ARM ELF files: $a, $t and $d.
 88 */
 89static inline int is_arm_mapping_symbol(const char *str)
 90{
 91	return str[0] == '$' && strchr("axtd", str[1])
 92	       && (str[2] == '\0' || str[2] == '.');
 93}
 94
 95static int check_symbol_range(const char *sym, unsigned long long addr,
 96			      struct addr_range *ranges, int entries)
 97{
 98	size_t i;
 99	struct addr_range *ar;
100
101	for (i = 0; i < entries; ++i) {
102		ar = &ranges[i];
103
104		if (strcmp(sym, ar->start_sym) == 0) {
105			ar->start = addr;
106			return 0;
107		} else if (strcmp(sym, ar->end_sym) == 0) {
108			ar->end = addr;
109			return 0;
110		}
111	}
112
113	return 1;
114}
115
116static int read_symbol(FILE *in, struct sym_entry *s)
117{
118	char str[500];
119	char *sym, stype;
120	int rc;
121
122	rc = fscanf(in, "%llx %c %499s\n", &s->addr, &stype, str);
123	if (rc != 3) {
124		if (rc != EOF && fgets(str, 500, in) == NULL)
125			fprintf(stderr, "Read error or end of file.\n");
126		return -1;
127	}
128	if (strlen(str) > KSYM_NAME_LEN) {
129		fprintf(stderr, "Symbol %s too long for kallsyms (%zu vs %d).\n"
130				"Please increase KSYM_NAME_LEN both in kernel and kallsyms.c\n",
131			str, strlen(str), KSYM_NAME_LEN);
132		return -1;
133	}
134
135	sym = str;
136	/* skip prefix char */
137	if (symbol_prefix_char && str[0] == symbol_prefix_char)
138		sym++;
139
140	/* Ignore most absolute/undefined (?) symbols. */
141	if (strcmp(sym, "_text") == 0)
142		_text = s->addr;
143	else if (check_symbol_range(sym, s->addr, text_ranges,
144				    ARRAY_SIZE(text_ranges)) == 0)
145		/* nothing to do */;
146	else if (toupper(stype) == 'A')
147	{
148		/* Keep these useful absolute symbols */
149		if (strcmp(sym, "__kernel_syscall_via_break") &&
150		    strcmp(sym, "__kernel_syscall_via_epc") &&
151		    strcmp(sym, "__kernel_sigtramp") &&
152		    strcmp(sym, "__gp"))
153			return -1;
154
155	}
156	else if (toupper(stype) == 'U' ||
157		 is_arm_mapping_symbol(sym))
158		return -1;
159	/* exclude also MIPS ELF local symbols ($L123 instead of .L123) */
160	else if (str[0] == '$')
161		return -1;
162	/* exclude debugging symbols */
163	else if (stype == 'N')
164		return -1;
165
166	/* include the type field in the symbol name, so that it gets
167	 * compressed together */
168	s->len = strlen(str) + 1;
169	s->sym = malloc(s->len + 1);
170	if (!s->sym) {
171		fprintf(stderr, "kallsyms failure: "
172			"unable to allocate required amount of memory\n");
173		exit(EXIT_FAILURE);
174	}
175	strcpy((char *)s->sym + 1, str);
176	s->sym[0] = stype;
177
178	s->percpu_absolute = 0;
179
180	/* Record if we've found __per_cpu_start/end. */
181	check_symbol_range(sym, s->addr, &percpu_range, 1);
182
183	return 0;
184}
185
186static int symbol_in_range(struct sym_entry *s, struct addr_range *ranges,
187			   int entries)
188{
189	size_t i;
190	struct addr_range *ar;
191
192	for (i = 0; i < entries; ++i) {
193		ar = &ranges[i];
194
195		if (s->addr >= ar->start && s->addr <= ar->end)
196			return 1;
197	}
198
199	return 0;
200}
201
202static int symbol_valid(struct sym_entry *s)
203{
204	/* Symbols which vary between passes.  Passes 1 and 2 must have
205	 * identical symbol lists.  The kallsyms_* symbols below are only added
206	 * after pass 1, they would be included in pass 2 when --all-symbols is
207	 * specified so exclude them to get a stable symbol list.
208	 */
209	static char *special_symbols[] = {
210		"kallsyms_addresses",
211		"kallsyms_offsets",
212		"kallsyms_relative_base",
213		"kallsyms_num_syms",
214		"kallsyms_names",
215		"kallsyms_markers",
216		"kallsyms_token_table",
217		"kallsyms_token_index",
218
219	/* Exclude linker generated symbols which vary between passes */
220		"_SDA_BASE_",		/* ppc */
221		"_SDA2_BASE_",		/* ppc */
222		NULL };
223
224	static char *special_suffixes[] = {
225		"_veneer",		/* arm */
226		NULL };
227
228	int i;
229	char *sym_name = (char *)s->sym + 1;
230
231
232	if (s->addr < kernel_start_addr)
233		return 0;
234
235	/* skip prefix char */
236	if (symbol_prefix_char && *sym_name == symbol_prefix_char)
237		sym_name++;
238
239
240	/* if --all-symbols is not specified, then symbols outside the text
241	 * and inittext sections are discarded */
242	if (!all_symbols) {
243		if (symbol_in_range(s, text_ranges,
244				    ARRAY_SIZE(text_ranges)) == 0)
245			return 0;
246		/* Corner case.  Discard any symbols with the same value as
247		 * _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.
251		 */
252		if ((s->addr == text_range_text->end &&
253				strcmp(sym_name,
254				       text_range_text->end_sym)) ||
255		    (s->addr == text_range_inittext->end &&
256				strcmp(sym_name,
257				       text_range_inittext->end_sym)))
258			return 0;
259	}
260
261	/* Exclude symbols which vary between passes. */
262	for (i = 0; special_symbols[i]; i++)
263		if (strcmp(sym_name, special_symbols[i]) == 0)
264			return 0;
265
266	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
273	return 1;
274}
275
276static void read_map(FILE *in)
277{
278	while (!feof(in)) {
279		if (table_cnt >= table_size) {
280			table_size += 10000;
281			table = realloc(table, sizeof(*table) * table_size);
282			if (!table) {
283				fprintf(stderr, "out of memory\n");
284				exit (1);
285			}
286		}
287		if (read_symbol(in, &table[table_cnt]) == 0) {
288			table[table_cnt].start_pos = table_cnt;
289			table_cnt++;
290		}
291	}
292}
293
294static void output_label(char *label)
295{
296	if (symbol_prefix_char)
297		printf(".globl %c%s\n", symbol_prefix_char, label);
298	else
299		printf(".globl %s\n", label);
300	printf("\tALGN\n");
301	if (symbol_prefix_char)
302		printf("%c%s:\n", symbol_prefix_char, label);
303	else
304		printf("%s:\n", label);
305}
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
334static int symbol_absolute(struct sym_entry *s)
335{
336	return s->percpu_absolute;
337}
338
339static void write_src(void)
340{
341	unsigned int i, k, off;
342	unsigned int best_idx[256];
343	unsigned int *markers;
344	char buf[KSYM_NAME_LEN];
345
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
355	printf("\t.section .rodata, \"a\"\n");
356
357	/* 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.
369	 */
370	if (!base_relative)
371		output_label("kallsyms_addresses");
372	else
373		output_label("kallsyms_offsets");
374
375	for (i = 0; i < table_cnt; i++) {
376		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])) {
399			if (_text <= table[i].addr)
400				printf("\tPTR\t_text + %#llx\n",
401					table[i].addr - _text);
402			else
403				printf("\tPTR\t_text - %#llx\n",
404					_text - table[i].addr);
405		} else {
406			printf("\tPTR\t%#llx\n", table[i].addr);
407		}
408	}
409	printf("\n");
410
411	if (base_relative) {
412		output_label("kallsyms_relative_base");
413		printf("\tPTR\t_text - %#llx\n", _text - relative_base);
414		printf("\n");
415	}
416
417	output_label("kallsyms_num_syms");
418	printf("\tPTR\t%d\n", table_cnt);
419	printf("\n");
420
421	/* table of offset markers, that give the offset in the compressed stream
422	 * every 256 symbols */
423	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	}
429
430	output_label("kallsyms_names");
431	off = 0;
432	for (i = 0; i < table_cnt; i++) {
433		if ((i & 0xFF) == 0)
434			markers[i >> 8] = off;
435
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;
442	}
443	printf("\n");
444
445	output_label("kallsyms_markers");
446	for (i = 0; i < ((table_cnt + 255) >> 8); i++)
447		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;
456		expand_symbol(best_table[i], best_table_len[i], buf);
457		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
471/* 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++)
477		token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
478}
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++)
486		token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
487}
488
489/* remove all the invalid symbols from the table and do the initial token count */
490static void build_initial_tok_table(void)
491{
492	unsigned int i, pos;
493
494	pos = 0;
495	for (i = 0; i < table_cnt; i++) {
496		if ( symbol_valid(&table[i]) ) {
497			if (pos != i)
498				table[pos] = table[i];
499			learn_symbol(table[pos].sym, table[pos].len);
500			pos++;
501		}
502	}
503	table_cnt = pos;
504}
505
506static void *find_token(unsigned char *str, int len, unsigned char *token)
507{
508	int i;
509
510	for (i = 0; i < len - 1; i++) {
511		if (str[i] == token[0] && str[i+1] == token[1])
512			return &str[i];
513	}
514	return NULL;
515}
516
517/* replace a given token in all the valid symbols. Use the sampled symbols
518 * to update the counts */
519static void compress_symbols(unsigned char *str, int idx)
520{
521	unsigned int i, len, size;
522	unsigned char *p1, *p2;
523
524	for (i = 0; i < table_cnt; i++) {
525
526		len = table[i].len;
527		p1 = table[i].sym;
528
529		/* find the token on the symbol */
530		p2 = find_token(p1, len, str);
531		if (!p2) continue;
532
533		/* decrease the counts for this symbol's tokens */
534		forget_symbol(table[i].sym, len);
535
536		size = len;
537
538		do {
539			*p2 = idx;
540			p2++;
541			size -= (p2 - p1);
542			memmove(p2, p2 + 1, size);
543			p1 = p2;
544			len--;
545
546			if (size < 2) break;
547
548			/* find the token on the symbol */
549			p2 = find_token(p1, size, str);
550
551		} while (p2);
552
553		table[i].len = len;
554
555		/* increase the counts for this symbol's new tokens */
556		learn_symbol(table[i].sym, len);
557	}
558}
559
560/* search the token with the maximum profit */
561static int find_best_token(void)
562{
563	int i, best, bestprofit;
564
565	bestprofit=-10000;
566	best = 0;
567
568	for (i = 0; i < 0x10000; i++) {
569		if (token_profit[i] > bestprofit) {
570			best = i;
571			bestprofit = token_profit[i];
572		}
573	}
574	return best;
575}
576
577/* this is the core of the algorithm: calculate the "best" table */
578static void optimize_result(void)
579{
580	int i, best;
581
582	/* using the '\0' symbol last allows compress_symbols to use standard
583	 * fast string functions */
584	for (i = 255; i >= 0; i--) {
585
586		/* if this table slot is empty (it is not used by an actual
587		 * original char code */
588		if (!best_table_len[i]) {
589
590			/* find the token with the breates profit value */
591			best = find_best_token();
592			if (token_profit[best] == 0)
593				break;
594
595			/* place it in the "best" table */
596			best_table_len[i] = 2;
597			best_table[i][0] = best & 0xFF;
598			best_table[i][1] = (best >> 8) & 0xFF;
599
600			/* replace this token in all the valid symbols */
601			compress_symbols(best_table[i], i);
602		}
603	}
604}
605
606/* start by placing the symbols that are actually used on the table */
607static void insert_real_symbols_in_table(void)
608{
609	unsigned int i, j, c;
610
611	memset(best_table, 0, sizeof(best_table));
612	memset(best_table_len, 0, sizeof(best_table_len));
613
614	for (i = 0; i < table_cnt; i++) {
615		for (j = 0; j < table[i].len; j++) {
616			c = table[i].sym[j];
617			best_table[c][0]=c;
618			best_table_len[c]=1;
619		}
620	}
621}
622
623static void optimize_token_table(void)
624{
625	build_initial_tok_table();
626
627	insert_real_symbols_in_table();
628
629	/* When valid symbol is not registered, exit to error */
630	if (!table_cnt) {
631		fprintf(stderr, "No valid symbol.\n");
632		exit(1);
633	}
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 = (char *)se->sym + 1;
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 prefix_underscores_count(const char *str)
674{
675	const char *tail = str;
676
677	while (*tail == '_')
678		tail++;
679
680	return tail - str;
681}
682
683static int compare_symbols(const void *a, const void *b)
684{
685	const struct sym_entry *sa;
686	const struct sym_entry *sb;
687	int wa, wb;
688
689	sa = a;
690	sb = b;
691
692	/* sort by address first */
693	if (sa->addr > sb->addr)
694		return 1;
695	if (sa->addr < sb->addr)
696		return -1;
697
698	/* sort by "weakness" type */
699	wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
700	wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
701	if (wa != wb)
702		return wa - wb;
703
704	/* sort by "linker script provide" type */
705	wa = may_be_linker_script_provide_symbol(sa);
706	wb = may_be_linker_script_provide_symbol(sb);
707	if (wa != wb)
708		return wa - wb;
709
710	/* sort by the number of prefix underscores */
711	wa = prefix_underscores_count((const char *)sa->sym + 1);
712	wb = prefix_underscores_count((const char *)sb->sym + 1);
713	if (wa != wb)
714		return wa - wb;
715
716	/* sort by initial order, so that other symbols are left undisturbed */
717	return sa->start_pos - sb->start_pos;
718}
719
720static void sort_symbols(void)
721{
722	qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
723}
724
725static void make_percpus_absolute(void)
726{
727	unsigned int i;
728
729	for (i = 0; i < table_cnt; i++)
730		if (symbol_in_range(&table[i], &percpu_range, 1)) {
731			/*
732			 * Keep the 'A' override for percpu symbols to
733			 * ensure consistent behavior compared to older
734			 * versions of this tool.
735			 */
736			table[i].sym[0] = 'A';
737			table[i].percpu_absolute = 1;
738		}
739}
740
741/* find the minimum non-absolute symbol address */
742static void record_relative_base(void)
743{
744	unsigned int i;
745
746	relative_base = -1ULL;
747	for (i = 0; i < table_cnt; i++)
748		if (!symbol_absolute(&table[i]) &&
749		    table[i].addr < relative_base)
750			relative_base = table[i].addr;
751}
752
753int main(int argc, char **argv)
754{
755	if (argc >= 2) {
756		int i;
757		for (i = 1; i < argc; i++) {
758			if(strcmp(argv[i], "--all-symbols") == 0)
759				all_symbols = 1;
760			else if (strcmp(argv[i], "--absolute-percpu") == 0)
761				absolute_percpu = 1;
762			else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
763				char *p = &argv[i][16];
764				/* skip quote */
765				if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
766					p++;
767				symbol_prefix_char = *p;
768			} else if (strncmp(argv[i], "--page-offset=", 14) == 0) {
769				const char *p = &argv[i][14];
770				kernel_start_addr = strtoull(p, NULL, 16);
771			} else if (strcmp(argv[i], "--base-relative") == 0)
772				base_relative = 1;
773			else
774				usage();
775		}
776	} else if (argc != 1)
777		usage();
778
779	read_map(stdin);
780	if (absolute_percpu)
781		make_percpus_absolute();
782	if (base_relative)
783		record_relative_base();
784	sort_symbols();
785	optimize_token_table();
786	write_src();
787
788	return 0;
789}