1 // SPDX-License-Identifier: GPL-2.0-only
2 /* -*- linux-c -*- ------------------------------------------------------- *
3  *
4  *   Copyright (C) 1991, 1992 Linus Torvalds
5  *   Copyright 2007 rPath, Inc. - All Rights Reserved
6  *
7  * ----------------------------------------------------------------------- */
8 
9 /*
10  * Oh, it's a waste of space, but oh-so-yummy for debugging.
11  */
12 
13 #include <stdarg.h>
14 
15 #include <linux/compiler.h>
16 #include <linux/ctype.h>
17 #include <linux/kernel.h>
18 #include <linux/limits.h>
19 #include <linux/string.h>
20 #include <linux/types.h>
21 
22 static
23 int skip_atoi(const char **s)
24 {
25 	int i = 0;
26 
27 	while (isdigit(**s))
28 		i = i * 10 + *((*s)++) - '0';
29 	return i;
30 }
31 
32 /*
33  * put_dec_full4 handles numbers in the range 0 <= r < 10000.
34  * The multiplier 0xccd is round(2^15/10), and the approximation
35  * r/10 == (r * 0xccd) >> 15 is exact for all r < 16389.
36  */
37 static
38 void put_dec_full4(char *end, unsigned int r)
39 {
40 	int i;
41 
42 	for (i = 0; i < 3; i++) {
43 		unsigned int q = (r * 0xccd) >> 15;
44 		*--end = '0' + (r - q * 10);
45 		r = q;
46 	}
47 	*--end = '0' + r;
48 }
49 
50 /* put_dec is copied from lib/vsprintf.c with small modifications */
51 
52 /*
53  * Call put_dec_full4 on x % 10000, return x / 10000.
54  * The approximation x/10000 == (x * 0x346DC5D7) >> 43
55  * holds for all x < 1,128,869,999.  The largest value this
56  * helper will ever be asked to convert is 1,125,520,955.
57  * (second call in the put_dec code, assuming n is all-ones).
58  */
59 static
60 unsigned int put_dec_helper4(char *end, unsigned int x)
61 {
62 	unsigned int q = (x * 0x346DC5D7ULL) >> 43;
63 
64 	put_dec_full4(end, x - q * 10000);
65 	return q;
66 }
67 
68 /* Based on code by Douglas W. Jones found at
69  * <http://www.cs.uiowa.edu/~jones/bcd/decimal.html#sixtyfour>
70  * (with permission from the author).
71  * Performs no 64-bit division and hence should be fast on 32-bit machines.
72  */
73 static
74 char *put_dec(char *end, unsigned long long n)
75 {
76 	unsigned int d3, d2, d1, q, h;
77 	char *p = end;
78 
79 	d1  = ((unsigned int)n >> 16); /* implicit "& 0xffff" */
80 	h   = (n >> 32);
81 	d2  = (h      ) & 0xffff;
82 	d3  = (h >> 16); /* implicit "& 0xffff" */
83 
84 	/* n = 2^48 d3 + 2^32 d2 + 2^16 d1 + d0
85 	     = 281_4749_7671_0656 d3 + 42_9496_7296 d2 + 6_5536 d1 + d0 */
86 	q = 656 * d3 + 7296 * d2 + 5536 * d1 + ((unsigned int)n & 0xffff);
87 	q = put_dec_helper4(p, q);
88 	p -= 4;
89 
90 	q += 7671 * d3 + 9496 * d2 + 6 * d1;
91 	q = put_dec_helper4(p, q);
92 	p -= 4;
93 
94 	q += 4749 * d3 + 42 * d2;
95 	q = put_dec_helper4(p, q);
96 	p -= 4;
97 
98 	q += 281 * d3;
99 	q = put_dec_helper4(p, q);
100 	p -= 4;
101 
102 	put_dec_full4(p, q);
103 	p -= 4;
104 
105 	/* strip off the extra 0's we printed */
106 	while (p < end && *p == '0')
107 		++p;
108 
109 	return p;
110 }
111 
112 static
113 char *number(char *end, unsigned long long num, int base, char locase)
114 {
115 	/*
116 	 * locase = 0 or 0x20. ORing digits or letters with 'locase'
117 	 * produces same digits or (maybe lowercased) letters
118 	 */
119 
120 	/* we are called with base 8, 10 or 16, only, thus don't need "G..."  */
121 	static const char digits[16] = "0123456789ABCDEF"; /* "GHIJKLMNOPQRSTUVWXYZ"; */
122 
123 	switch (base) {
124 	case 10:
125 		if (num != 0)
126 			end = put_dec(end, num);
127 		break;
128 	case 8:
129 		for (; num != 0; num >>= 3)
130 			*--end = '0' + (num & 07);
131 		break;
132 	case 16:
133 		for (; num != 0; num >>= 4)
134 			*--end = digits[num & 0xf] | locase;
135 		break;
136 	default:
137 		unreachable();
138 	}
139 
140 	return end;
141 }
142 
143 #define ZEROPAD	1		/* pad with zero */
144 #define SIGN	2		/* unsigned/signed long */
145 #define PLUS	4		/* show plus */
146 #define SPACE	8		/* space if plus */
147 #define LEFT	16		/* left justified */
148 #define SMALL	32		/* Must be 32 == 0x20 */
149 #define SPECIAL	64		/* 0x */
150 #define WIDE	128		/* UTF-16 string */
151 
152 static
153 int get_flags(const char **fmt)
154 {
155 	int flags = 0;
156 
157 	do {
158 		switch (**fmt) {
159 		case '-':
160 			flags |= LEFT;
161 			break;
162 		case '+':
163 			flags |= PLUS;
164 			break;
165 		case ' ':
166 			flags |= SPACE;
167 			break;
168 		case '#':
169 			flags |= SPECIAL;
170 			break;
171 		case '0':
172 			flags |= ZEROPAD;
173 			break;
174 		default:
175 			return flags;
176 		}
177 		++(*fmt);
178 	} while (1);
179 }
180 
181 static
182 int get_int(const char **fmt, va_list *ap)
183 {
184 	if (isdigit(**fmt))
185 		return skip_atoi(fmt);
186 	if (**fmt == '*') {
187 		++(*fmt);
188 		/* it's the next argument */
189 		return va_arg(*ap, int);
190 	}
191 	return 0;
192 }
193 
194 static
195 unsigned long long get_number(int sign, int qualifier, va_list *ap)
196 {
197 	if (sign) {
198 		switch (qualifier) {
199 		case 'L':
200 			return va_arg(*ap, long long);
201 		case 'l':
202 			return va_arg(*ap, long);
203 		case 'h':
204 			return (short)va_arg(*ap, int);
205 		case 'H':
206 			return (signed char)va_arg(*ap, int);
207 		default:
208 			return va_arg(*ap, int);
209 		};
210 	} else {
211 		switch (qualifier) {
212 		case 'L':
213 			return va_arg(*ap, unsigned long long);
214 		case 'l':
215 			return va_arg(*ap, unsigned long);
216 		case 'h':
217 			return (unsigned short)va_arg(*ap, int);
218 		case 'H':
219 			return (unsigned char)va_arg(*ap, int);
220 		default:
221 			return va_arg(*ap, unsigned int);
222 		}
223 	}
224 }
225 
226 static
227 char get_sign(long long *num, int flags)
228 {
229 	if (!(flags & SIGN))
230 		return 0;
231 	if (*num < 0) {
232 		*num = -(*num);
233 		return '-';
234 	}
235 	if (flags & PLUS)
236 		return '+';
237 	if (flags & SPACE)
238 		return ' ';
239 	return 0;
240 }
241 
242 static
243 size_t utf16s_utf8nlen(const u16 *s16, size_t maxlen)
244 {
245 	size_t len, clen;
246 
247 	for (len = 0; len < maxlen && *s16; len += clen) {
248 		u16 c0 = *s16++;
249 
250 		/* First, get the length for a BMP character */
251 		clen = 1 + (c0 >= 0x80) + (c0 >= 0x800);
252 		if (len + clen > maxlen)
253 			break;
254 		/*
255 		 * If this is a high surrogate, and we're already at maxlen, we
256 		 * can't include the character if it's a valid surrogate pair.
257 		 * Avoid accessing one extra word just to check if it's valid
258 		 * or not.
259 		 */
260 		if ((c0 & 0xfc00) == 0xd800) {
261 			if (len + clen == maxlen)
262 				break;
263 			if ((*s16 & 0xfc00) == 0xdc00) {
264 				++s16;
265 				++clen;
266 			}
267 		}
268 	}
269 
270 	return len;
271 }
272 
273 static
274 u32 utf16_to_utf32(const u16 **s16)
275 {
276 	u16 c0, c1;
277 
278 	c0 = *(*s16)++;
279 	/* not a surrogate */
280 	if ((c0 & 0xf800) != 0xd800)
281 		return c0;
282 	/* invalid: low surrogate instead of high */
283 	if (c0 & 0x0400)
284 		return 0xfffd;
285 	c1 = **s16;
286 	/* invalid: missing low surrogate */
287 	if ((c1 & 0xfc00) != 0xdc00)
288 		return 0xfffd;
289 	/* valid surrogate pair */
290 	++(*s16);
291 	return (0x10000 - (0xd800 << 10) - 0xdc00) + (c0 << 10) + c1;
292 }
293 
294 #define PUTC(c) \
295 do {				\
296 	if (pos < size)		\
297 		buf[pos] = (c);	\
298 	++pos;			\
299 } while (0);
300 
301 int vsnprintf(char *buf, size_t size, const char *fmt, va_list ap)
302 {
303 	/* The maximum space required is to print a 64-bit number in octal */
304 	char tmp[(sizeof(unsigned long long) * 8 + 2) / 3];
305 	char *tmp_end = &tmp[ARRAY_SIZE(tmp)];
306 	long long num;
307 	int base;
308 	const char *s;
309 	size_t len, pos;
310 	char sign;
311 
312 	int flags;		/* flags to number() */
313 
314 	int field_width;	/* width of output field */
315 	int precision;		/* min. # of digits for integers; max
316 				   number of chars for from string */
317 	int qualifier;		/* 'h', 'hh', 'l' or 'll' for integer fields */
318 
319 	va_list args;
320 
321 	/*
322 	 * We want to pass our input va_list to helper functions by reference,
323 	 * but there's an annoying edge case. If va_list was originally passed
324 	 * to us by value, we could just pass &ap down to the helpers. This is
325 	 * the case on, for example, X86_32.
326 	 * However, on X86_64 (and possibly others), va_list is actually a
327 	 * size-1 array containing a structure. Our function parameter ap has
328 	 * decayed from T[1] to T*, and &ap has type T** rather than T(*)[1],
329 	 * which is what will be expected by a function taking a va_list *
330 	 * parameter.
331 	 * One standard way to solve this mess is by creating a copy in a local
332 	 * variable of type va_list and then passing a pointer to that local
333 	 * copy instead, which is what we do here.
334 	 */
335 	va_copy(args, ap);
336 
337 	for (pos = 0; *fmt; ++fmt) {
338 		if (*fmt != '%' || *++fmt == '%') {
339 			PUTC(*fmt);
340 			continue;
341 		}
342 
343 		/* process flags */
344 		flags = get_flags(&fmt);
345 
346 		/* get field width */
347 		field_width = get_int(&fmt, &args);
348 		if (field_width < 0) {
349 			field_width = -field_width;
350 			flags |= LEFT;
351 		}
352 
353 		if (flags & LEFT)
354 			flags &= ~ZEROPAD;
355 
356 		/* get the precision */
357 		precision = -1;
358 		if (*fmt == '.') {
359 			++fmt;
360 			precision = get_int(&fmt, &args);
361 			if (precision >= 0)
362 				flags &= ~ZEROPAD;
363 		}
364 
365 		/* get the conversion qualifier */
366 		qualifier = -1;
367 		if (*fmt == 'h' || *fmt == 'l') {
368 			qualifier = *fmt;
369 			++fmt;
370 			if (qualifier == *fmt) {
371 				qualifier -= 'a'-'A';
372 				++fmt;
373 			}
374 		}
375 
376 		sign = 0;
377 
378 		switch (*fmt) {
379 		case 'c':
380 			flags &= LEFT;
381 			s = tmp;
382 			if (qualifier == 'l') {
383 				((u16 *)tmp)[0] = (u16)va_arg(args, unsigned int);
384 				((u16 *)tmp)[1] = L'\0';
385 				precision = INT_MAX;
386 				goto wstring;
387 			} else {
388 				tmp[0] = (unsigned char)va_arg(args, int);
389 				precision = len = 1;
390 			}
391 			goto output;
392 
393 		case 's':
394 			flags &= LEFT;
395 			if (precision < 0)
396 				precision = INT_MAX;
397 			s = va_arg(args, void *);
398 			if (!s)
399 				s = precision < 6 ? "" : "(null)";
400 			else if (qualifier == 'l') {
401 		wstring:
402 				flags |= WIDE;
403 				precision = len = utf16s_utf8nlen((const u16 *)s, precision);
404 				goto output;
405 			}
406 			precision = len = strnlen(s, precision);
407 			goto output;
408 
409 			/* integer number formats - set up the flags and "break" */
410 		case 'o':
411 			base = 8;
412 			break;
413 
414 		case 'p':
415 			if (precision < 0)
416 				precision = 2 * sizeof(void *);
417 			fallthrough;
418 		case 'x':
419 			flags |= SMALL;
420 			fallthrough;
421 		case 'X':
422 			base = 16;
423 			break;
424 
425 		case 'd':
426 		case 'i':
427 			flags |= SIGN;
428 			fallthrough;
429 		case 'u':
430 			flags &= ~SPECIAL;
431 			base = 10;
432 			break;
433 
434 		default:
435 			/*
436 			 * Bail out if the conversion specifier is invalid.
437 			 * There's probably a typo in the format string and the
438 			 * remaining specifiers are unlikely to match up with
439 			 * the arguments.
440 			 */
441 			goto fail;
442 		}
443 		if (*fmt == 'p') {
444 			num = (unsigned long)va_arg(args, void *);
445 		} else {
446 			num = get_number(flags & SIGN, qualifier, &args);
447 		}
448 
449 		sign = get_sign(&num, flags);
450 		if (sign)
451 			--field_width;
452 
453 		s = number(tmp_end, num, base, flags & SMALL);
454 		len = tmp_end - s;
455 		/* default precision is 1 */
456 		if (precision < 0)
457 			precision = 1;
458 		/* precision is minimum number of digits to print */
459 		if (precision < len)
460 			precision = len;
461 		if (flags & SPECIAL) {
462 			/*
463 			 * For octal, a leading 0 is printed only if necessary,
464 			 * i.e. if it's not already there because of the
465 			 * precision.
466 			 */
467 			if (base == 8 && precision == len)
468 				++precision;
469 			/*
470 			 * For hexadecimal, the leading 0x is skipped if the
471 			 * output is empty, i.e. both the number and the
472 			 * precision are 0.
473 			 */
474 			if (base == 16 && precision > 0)
475 				field_width -= 2;
476 			else
477 				flags &= ~SPECIAL;
478 		}
479 		/*
480 		 * For zero padding, increase the precision to fill the field
481 		 * width.
482 		 */
483 		if ((flags & ZEROPAD) && field_width > precision)
484 			precision = field_width;
485 
486 output:
487 		/* Calculate the padding necessary */
488 		field_width -= precision;
489 		/* Leading padding with ' ' */
490 		if (!(flags & LEFT))
491 			while (field_width-- > 0)
492 				PUTC(' ');
493 		/* sign */
494 		if (sign)
495 			PUTC(sign);
496 		/* 0x/0X for hexadecimal */
497 		if (flags & SPECIAL) {
498 			PUTC('0');
499 			PUTC( 'X' | (flags & SMALL));
500 		}
501 		/* Zero padding and excess precision */
502 		while (precision-- > len)
503 			PUTC('0');
504 		/* Actual output */
505 		if (flags & WIDE) {
506 			const u16 *ws = (const u16 *)s;
507 
508 			while (len-- > 0) {
509 				u32 c32 = utf16_to_utf32(&ws);
510 				u8 *s8;
511 				size_t clen;
512 
513 				if (c32 < 0x80) {
514 					PUTC(c32);
515 					continue;
516 				}
517 
518 				/* Number of trailing octets */
519 				clen = 1 + (c32 >= 0x800) + (c32 >= 0x10000);
520 
521 				len -= clen;
522 				s8 = (u8 *)&buf[pos];
523 
524 				/* Avoid writing partial character */
525 				PUTC('\0');
526 				pos += clen;
527 				if (pos >= size)
528 					continue;
529 
530 				/* Set high bits of leading octet */
531 				*s8 = (0xf00 >> 1) >> clen;
532 				/* Write trailing octets in reverse order */
533 				for (s8 += clen; clen; --clen, c32 >>= 6)
534 					*s8-- = 0x80 | (c32 & 0x3f);
535 				/* Set low bits of leading octet */
536 				*s8 |= c32;
537 			}
538 		} else {
539 			while (len-- > 0)
540 				PUTC(*s++);
541 		}
542 		/* Trailing padding with ' ' */
543 		while (field_width-- > 0)
544 			PUTC(' ');
545 	}
546 fail:
547 	va_end(args);
548 
549 	if (size)
550 		buf[min(pos, size-1)] = '\0';
551 
552 	return pos;
553 }
554 
555 int snprintf(char *buf, size_t size, const char *fmt, ...)
556 {
557 	va_list args;
558 	int i;
559 
560 	va_start(args, fmt);
561 	i = vsnprintf(buf, size, fmt, args);
562 	va_end(args);
563 	return i;
564 }
565