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')
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		NULL };
225
226	static char *special_suffixes[] = {
227		"_veneer",		/* arm */
228		"_from_arm",		/* arm */
229		"_from_thumb",		/* arm */
230		NULL };
231
232	int i;
233	char *sym_name = (char *)s->sym + 1;
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_prefixes[i]; i++) {
267		int l = strlen(special_prefixes[i]);
268
269		if (l <= strlen(sym_name) &&
270		    strncmp(sym_name, special_prefixes[i], l) == 0)
271			return 0;
272	}
273
274	for (i = 0; special_suffixes[i]; i++) {
275		int l = strlen(sym_name) - strlen(special_suffixes[i]);
276
277		if (l >= 0 && strcmp(sym_name + l, special_suffixes[i]) == 0)
278			return 0;
279	}
280
281	return 1;
282}
283
284static void read_map(FILE *in)
285{
286	while (!feof(in)) {
287		if (table_cnt >= table_size) {
288			table_size += 10000;
289			table = realloc(table, sizeof(*table) * table_size);
290			if (!table) {
291				fprintf(stderr, "out of memory\n");
292				exit (1);
293			}
294		}
295		if (read_symbol(in, &table[table_cnt]) == 0) {
296			table[table_cnt].start_pos = table_cnt;
297			table_cnt++;
298		}
299	}
300}
301
302static void output_label(char *label)
303{
304	if (symbol_prefix_char)
305		printf(".globl %c%s\n", symbol_prefix_char, label);
306	else
307		printf(".globl %s\n", label);
308	printf("\tALGN\n");
309	if (symbol_prefix_char)
310		printf("%c%s:\n", symbol_prefix_char, label);
311	else
312		printf("%s:\n", label);
313}
314
315/* uncompress a compressed symbol. When this function is called, the best table
316 * might still be compressed itself, so the function needs to be recursive */
317static int expand_symbol(unsigned char *data, int len, char *result)
318{
319	int c, rlen, total=0;
320
321	while (len) {
322		c = *data;
323		/* if the table holds a single char that is the same as the one
324		 * we are looking for, then end the search */
325		if (best_table[c][0]==c && best_table_len[c]==1) {
326			*result++ = c;
327			total++;
328		} else {
329			/* if not, recurse and expand */
330			rlen = expand_symbol(best_table[c], best_table_len[c], result);
331			total += rlen;
332			result += rlen;
333		}
334		data++;
335		len--;
336	}
337	*result=0;
338
339	return total;
340}
341
342static int symbol_absolute(struct sym_entry *s)
343{
344	return s->percpu_absolute;
345}
346
347static void write_src(void)
348{
349	unsigned int i, k, off;
350	unsigned int best_idx[256];
351	unsigned int *markers;
352	char buf[KSYM_NAME_LEN];
353
354	printf("#include <asm/types.h>\n");
355	printf("#if BITS_PER_LONG == 64\n");
356	printf("#define PTR .quad\n");
357	printf("#define ALGN .align 8\n");
358	printf("#else\n");
359	printf("#define PTR .long\n");
360	printf("#define ALGN .align 4\n");
361	printf("#endif\n");
362
363	printf("\t.section .rodata, \"a\"\n");
364
365	/* Provide proper symbols relocatability by their relativeness
366	 * to a fixed anchor point in the runtime image, either '_text'
367	 * for absolute address tables, in which case the linker will
368	 * emit the final addresses at build time. Otherwise, use the
369	 * offset relative to the lowest value encountered of all relative
370	 * symbols, and emit non-relocatable fixed offsets that will be fixed
371	 * up at runtime.
372	 *
373	 * The symbol names cannot be used to construct normal symbol
374	 * references as the list of symbols contains symbols that are
375	 * declared static and are private to their .o files.  This prevents
376	 * .tmp_kallsyms.o or any other object from referencing them.
377	 */
378	if (!base_relative)
379		output_label("kallsyms_addresses");
380	else
381		output_label("kallsyms_offsets");
382
383	for (i = 0; i < table_cnt; i++) {
384		if (base_relative) {
385			long long offset;
386			int overflow;
387
388			if (!absolute_percpu) {
389				offset = table[i].addr - relative_base;
390				overflow = (offset < 0 || offset > UINT_MAX);
391			} else if (symbol_absolute(&table[i])) {
392				offset = table[i].addr;
393				overflow = (offset < 0 || offset > INT_MAX);
394			} else {
395				offset = relative_base - table[i].addr - 1;
396				overflow = (offset < INT_MIN || offset >= 0);
397			}
398			if (overflow) {
399				fprintf(stderr, "kallsyms failure: "
400					"%s symbol value %#llx out of range in relative mode\n",
401					symbol_absolute(&table[i]) ? "absolute" : "relative",
402					table[i].addr);
403				exit(EXIT_FAILURE);
404			}
405			printf("\t.long\t%#x\n", (int)offset);
406		} else if (!symbol_absolute(&table[i])) {
407			if (_text <= table[i].addr)
408				printf("\tPTR\t_text + %#llx\n",
409					table[i].addr - _text);
410			else
411				printf("\tPTR\t_text - %#llx\n",
412					_text - table[i].addr);
413		} else {
414			printf("\tPTR\t%#llx\n", table[i].addr);
415		}
416	}
417	printf("\n");
418
419	if (base_relative) {
420		output_label("kallsyms_relative_base");
421		printf("\tPTR\t_text - %#llx\n", _text - relative_base);
422		printf("\n");
423	}
424
425	output_label("kallsyms_num_syms");
426	printf("\tPTR\t%d\n", table_cnt);
427	printf("\n");
428
429	/* table of offset markers, that give the offset in the compressed stream
430	 * every 256 symbols */
431	markers = malloc(sizeof(unsigned int) * ((table_cnt + 255) / 256));
432	if (!markers) {
433		fprintf(stderr, "kallsyms failure: "
434			"unable to allocate required memory\n");
435		exit(EXIT_FAILURE);
436	}
437
438	output_label("kallsyms_names");
439	off = 0;
440	for (i = 0; i < table_cnt; i++) {
441		if ((i & 0xFF) == 0)
442			markers[i >> 8] = off;
443
444		printf("\t.byte 0x%02x", table[i].len);
445		for (k = 0; k < table[i].len; k++)
446			printf(", 0x%02x", table[i].sym[k]);
447		printf("\n");
448
449		off += table[i].len + 1;
450	}
451	printf("\n");
452
453	output_label("kallsyms_markers");
454	for (i = 0; i < ((table_cnt + 255) >> 8); i++)
455		printf("\tPTR\t%d\n", markers[i]);
456	printf("\n");
457
458	free(markers);
459
460	output_label("kallsyms_token_table");
461	off = 0;
462	for (i = 0; i < 256; i++) {
463		best_idx[i] = off;
464		expand_symbol(best_table[i], best_table_len[i], buf);
465		printf("\t.asciz\t\"%s\"\n", buf);
466		off += strlen(buf) + 1;
467	}
468	printf("\n");
469
470	output_label("kallsyms_token_index");
471	for (i = 0; i < 256; i++)
472		printf("\t.short\t%d\n", best_idx[i]);
473	printf("\n");
474}
475
476
477/* table lookup compression functions */
478
479/* count all the possible tokens in a symbol */
480static void learn_symbol(unsigned char *symbol, int len)
481{
482	int i;
483
484	for (i = 0; i < len - 1; i++)
485		token_profit[ symbol[i] + (symbol[i + 1] << 8) ]++;
486}
487
488/* decrease the count for all the possible tokens in a symbol */
489static void forget_symbol(unsigned char *symbol, int len)
490{
491	int i;
492
493	for (i = 0; i < len - 1; i++)
494		token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
495}
496
497/* remove all the invalid symbols from the table and do the initial token count */
498static void build_initial_tok_table(void)
499{
500	unsigned int i, pos;
501
502	pos = 0;
503	for (i = 0; i < table_cnt; i++) {
504		if ( symbol_valid(&table[i]) ) {
505			if (pos != i)
506				table[pos] = table[i];
507			learn_symbol(table[pos].sym, table[pos].len);
508			pos++;
509		}
510	}
511	table_cnt = pos;
512}
513
514static void *find_token(unsigned char *str, int len, unsigned char *token)
515{
516	int i;
517
518	for (i = 0; i < len - 1; i++) {
519		if (str[i] == token[0] && str[i+1] == token[1])
520			return &str[i];
521	}
522	return NULL;
523}
524
525/* replace a given token in all the valid symbols. Use the sampled symbols
526 * to update the counts */
527static void compress_symbols(unsigned char *str, int idx)
528{
529	unsigned int i, len, size;
530	unsigned char *p1, *p2;
531
532	for (i = 0; i < table_cnt; i++) {
533
534		len = table[i].len;
535		p1 = table[i].sym;
536
537		/* find the token on the symbol */
538		p2 = find_token(p1, len, str);
539		if (!p2) continue;
540
541		/* decrease the counts for this symbol's tokens */
542		forget_symbol(table[i].sym, len);
543
544		size = len;
545
546		do {
547			*p2 = idx;
548			p2++;
549			size -= (p2 - p1);
550			memmove(p2, p2 + 1, size);
551			p1 = p2;
552			len--;
553
554			if (size < 2) break;
555
556			/* find the token on the symbol */
557			p2 = find_token(p1, size, str);
558
559		} while (p2);
560
561		table[i].len = len;
562
563		/* increase the counts for this symbol's new tokens */
564		learn_symbol(table[i].sym, len);
565	}
566}
567
568/* search the token with the maximum profit */
569static int find_best_token(void)
570{
571	int i, best, bestprofit;
572
573	bestprofit=-10000;
574	best = 0;
575
576	for (i = 0; i < 0x10000; i++) {
577		if (token_profit[i] > bestprofit) {
578			best = i;
579			bestprofit = token_profit[i];
580		}
581	}
582	return best;
583}
584
585/* this is the core of the algorithm: calculate the "best" table */
586static void optimize_result(void)
587{
588	int i, best;
589
590	/* using the '\0' symbol last allows compress_symbols to use standard
591	 * fast string functions */
592	for (i = 255; i >= 0; i--) {
593
594		/* if this table slot is empty (it is not used by an actual
595		 * original char code */
596		if (!best_table_len[i]) {
597
598			/* find the token with the breates profit value */
599			best = find_best_token();
600			if (token_profit[best] == 0)
601				break;
602
603			/* place it in the "best" table */
604			best_table_len[i] = 2;
605			best_table[i][0] = best & 0xFF;
606			best_table[i][1] = (best >> 8) & 0xFF;
607
608			/* replace this token in all the valid symbols */
609			compress_symbols(best_table[i], i);
610		}
611	}
612}
613
614/* start by placing the symbols that are actually used on the table */
615static void insert_real_symbols_in_table(void)
616{
617	unsigned int i, j, c;
618
619	memset(best_table, 0, sizeof(best_table));
620	memset(best_table_len, 0, sizeof(best_table_len));
621
622	for (i = 0; i < table_cnt; i++) {
623		for (j = 0; j < table[i].len; j++) {
624			c = table[i].sym[j];
625			best_table[c][0]=c;
626			best_table_len[c]=1;
627		}
628	}
629}
630
631static void optimize_token_table(void)
632{
633	build_initial_tok_table();
634
635	insert_real_symbols_in_table();
636
637	/* When valid symbol is not registered, exit to error */
638	if (!table_cnt) {
639		fprintf(stderr, "No valid symbol.\n");
640		exit(1);
641	}
642
643	optimize_result();
644}
645
646/* guess for "linker script provide" symbol */
647static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
648{
649	const char *symbol = (char *)se->sym + 1;
650	int len = se->len - 1;
651
652	if (len < 8)
653		return 0;
654
655	if (symbol[0] != '_' || symbol[1] != '_')
656		return 0;
657
658	/* __start_XXXXX */
659	if (!memcmp(symbol + 2, "start_", 6))
660		return 1;
661
662	/* __stop_XXXXX */
663	if (!memcmp(symbol + 2, "stop_", 5))
664		return 1;
665
666	/* __end_XXXXX */
667	if (!memcmp(symbol + 2, "end_", 4))
668		return 1;
669
670	/* __XXXXX_start */
671	if (!memcmp(symbol + len - 6, "_start", 6))
672		return 1;
673
674	/* __XXXXX_end */
675	if (!memcmp(symbol + len - 4, "_end", 4))
676		return 1;
677
678	return 0;
679}
680
681static int prefix_underscores_count(const char *str)
682{
683	const char *tail = str;
684
685	while (*tail == '_')
686		tail++;
687
688	return tail - str;
689}
690
691static int compare_symbols(const void *a, const void *b)
692{
693	const struct sym_entry *sa;
694	const struct sym_entry *sb;
695	int wa, wb;
696
697	sa = a;
698	sb = b;
699
700	/* sort by address first */
701	if (sa->addr > sb->addr)
702		return 1;
703	if (sa->addr < sb->addr)
704		return -1;
705
706	/* sort by "weakness" type */
707	wa = (sa->sym[0] == 'w') || (sa->sym[0] == 'W');
708	wb = (sb->sym[0] == 'w') || (sb->sym[0] == 'W');
709	if (wa != wb)
710		return wa - wb;
711
712	/* sort by "linker script provide" type */
713	wa = may_be_linker_script_provide_symbol(sa);
714	wb = may_be_linker_script_provide_symbol(sb);
715	if (wa != wb)
716		return wa - wb;
717
718	/* sort by the number of prefix underscores */
719	wa = prefix_underscores_count((const char *)sa->sym + 1);
720	wb = prefix_underscores_count((const char *)sb->sym + 1);
721	if (wa != wb)
722		return wa - wb;
723
724	/* sort by initial order, so that other symbols are left undisturbed */
725	return sa->start_pos - sb->start_pos;
726}
727
728static void sort_symbols(void)
729{
730	qsort(table, table_cnt, sizeof(struct sym_entry), compare_symbols);
731}
732
733static void make_percpus_absolute(void)
734{
735	unsigned int i;
736
737	for (i = 0; i < table_cnt; i++)
738		if (symbol_in_range(&table[i], &percpu_range, 1)) {
739			/*
740			 * Keep the 'A' override for percpu symbols to
741			 * ensure consistent behavior compared to older
742			 * versions of this tool.
743			 */
744			table[i].sym[0] = 'A';
745			table[i].percpu_absolute = 1;
746		}
747}
748
749/* find the minimum non-absolute symbol address */
750static void record_relative_base(void)
751{
752	unsigned int i;
753
754	relative_base = -1ULL;
755	for (i = 0; i < table_cnt; i++)
756		if (!symbol_absolute(&table[i]) &&
757		    table[i].addr < relative_base)
758			relative_base = table[i].addr;
759}
760
761int main(int argc, char **argv)
762{
763	if (argc >= 2) {
764		int i;
765		for (i = 1; i < argc; i++) {
766			if(strcmp(argv[i], "--all-symbols") == 0)
767				all_symbols = 1;
768			else if (strcmp(argv[i], "--absolute-percpu") == 0)
769				absolute_percpu = 1;
770			else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0) {
771				char *p = &argv[i][16];
772				/* skip quote */
773				if ((*p == '"' && *(p+2) == '"') || (*p == '\'' && *(p+2) == '\''))
774					p++;
775				symbol_prefix_char = *p;
776			} else if (strcmp(argv[i], "--base-relative") == 0)
777				base_relative = 1;
778			else
779				usage();
780		}
781	} else if (argc != 1)
782		usage();
783
784	read_map(stdin);
785	if (absolute_percpu)
786		make_percpus_absolute();
787	if (base_relative)
788		record_relative_base();
789	sort_symbols();
790	optimize_token_table();
791	write_src();
792
793	return 0;
794}