xref: /dports/security/john/john-1.9.0-jumbo-1/src/misc.c

/*
 * This file is part of John the Ripper password cracker,
 * Copyright (c) 1996-99,2003 by Solar Designer
 *
 * ...with changes in the jumbo patch, by JimF and magnum.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted.
 *
 * There's ABSOLUTELY NO WARRANTY, express or implied.
 */

#include <stdio.h>
#define NEED_OS_FORK
#include "os.h"
#if (!AC_BUILT || HAVE_UNISTD_H) && !_MSC_VER
#include <unistd.h>
#endif
#ifdef _MSC_VER
#include <io.h>
#pragma warning ( disable : 4996 )
#endif
#include <stdlib.h>
#include <string.h>
#include <stdarg.h>
#include <errno.h>
#include <assert.h>

#include "memory.h"
#include "logger.h"
#include "params.h"
#include "misc.h"
#include "options.h"

#include "john_mpi.h"

void real_error(const char *file, int line)
{
#ifndef _JOHN_MISC_NO_LOG
	log_event("Terminating on error, %s:%d", file, line);
	log_done();
#else
	fprintf(stderr, "Terminating on error, %s:%d\n", file, line);
#endif

	exit(1);
}

void real_error_msg(const char *file, int line,  const char *format, ...)
{
	va_list args;

#if defined(HAVE_MPI) && !defined(_JOHN_MISC_NO_LOG)
	if (mpi_p > 1)
		fprintf(stderr, "%u@%s: ", mpi_id + 1, mpi_name);
	else
#elif OS_FORK && !defined(_JOHN_MISC_NO_LOG)
	if (options.fork)
		fprintf(stderr, "%u: ", options.node_min);
#endif
	va_start(args, format);
	vfprintf(stderr, format, args);
	va_end(args);

	real_error(file, line);
}

void real_pexit(const char *file, int line, const char *format, ...)
{
	va_list args;

#if !defined(_JOHN_MISC_NO_LOG)
#if HAVE_MPI
	if (mpi_p > 1)
		fprintf(stderr, "%u@%s: ", mpi_id + 1, mpi_name);
#endif
#if HAVE_MPI && OS_FORK
	else
#endif
#if OS_FORK
	if (options.fork)
		fprintf(stderr, "%u: ", options.node_min);
#endif
#endif

	va_start(args, format);
	vfprintf(stderr, format, args);
	va_end(args);

	fprintf(stderr, ": %s\n", strerror(errno));

	real_error(file, line);
}

int write_loop(int fd, const char *buffer, int count)
{
	int offset, block;

	offset = 0;
	while (count > 0) {
		block = write(fd, &buffer[offset], count);

/* If any write(2) fails, we consider that the entire write_loop() has
 * failed to do its job, unless we were interrupted by a signal. */
		if (block < 0) {
			if (errno == EINTR) continue;
			return block;
		}

		offset += block;
		count -= block;
	}

/* Should be equal to the requested size, unless our kernel got crazy. */
	return offset;
}

char *fgetl(char *s, int size, FILE *stream)
{
	char *res, *pos;
	int c;

	if ((res = fgets(s, size, stream))) {
		if (!*res) return res;

		pos = res + strlen(res) - 1;
		if (*pos == '\n') {
			*pos = 0;
			if (pos > res)
			if (*--pos == '\r') *pos = 0;
			return res;
		}
		if ((pos-res) + 1 < size) {
			/* There was a NULL byte in this line.
			   Look for \n past the 'read' null byte */
			while ((++pos - res) + 1 < size)
				if (*pos == '\n')
					return res; /* found it read no more */
		}
		if ((c = getc(stream)) == '\n') {
			if (*pos == '\r') *pos = 0;
		} else
		while (c != EOF && c != '\n')
			/* Line was longer than our buffer. discard extra */
			c = getc(stream);
	}

	return res;
}

char *fgetll(char *s, size_t size, FILE *stream)
{
	size_t len;
	int c;
	char *cp;

	/* fgets' size arg is a signed int! */
	assert(size <= INT32_MAX);

	if (!fgets(s, size, stream))
		return NULL;

	len = strlen(s);

	if (!len)
		return s;

	if (s[len-1] == '\n') {
		s[--len] = 0;
		while (len && (s[len-1] == '\n' || s[len-1] == '\r'))
			s[--len] = 0;
		return s;
	}
	else if (s[len-1] == '\r') {
		s[--len] = 0;
		while (len && (s[len-1] == '\n' || s[len-1] == '\r'))
			s[--len] = 0;
		/* we may have gotten the first byte of \r\n */
		c = getc(stream);
		if (c == EOF)
			return s;
		if (c != '\n')
			ungetc(c, stream);
		return s;
	}
	else if ((len + 1) < size) { /* We read a null byte */
		/* first check past the null byte, looking for the \n */
		while (++len < size)
			if (s[len] == '\n')
				return s; /* found it. Read no more. */
		/* did not find the \n, so read and discard rest of line */
		do {
			c = getc(stream);
		} while (c != EOF && c != '\n');
		return s;
	}

	cp = strdup(s);

	while (1) {
		int increase = MIN((((len >> 12) + 1) << 12), 0x40000000);
		size_t chunk_len;
		void *new_cp;

		new_cp = realloc(cp, len + increase);

		while (!new_cp) {
			increase >>= 2;
			if (increase < 0x10000)
				pexit("realloc");
			new_cp = realloc(cp, len + increase);
		}

		cp = new_cp;

		/* We get an EOF if there is no trailing \n on the last line */
		if (!fgets(&cp[len], increase, stream))
			return cp;

		chunk_len = strlen(&cp[len]);
		len += chunk_len;

		if (cp[len-1] == '\n') {
			cp[--len] = 0;
			while (len && (cp[len-1] == '\n' || cp[len-1] == '\r'))
				cp[--len] = 0;
			return cp;
		}
		else if (cp[len-1] == '\r') {
			cp[--len] = 0;
			while (len && (cp[len-1] == '\n' || cp[len-1] == '\r'))
				cp[--len] = 0;
			/* we may have gotten the first byte of \r\n */
			c = getc(stream);
			if (c == EOF)
				return cp;
			if (c != '\n')
				ungetc(c, stream);
			return cp;
		}
		else if ((chunk_len + 1) < increase) { /* We read a null byte */
			/* first check past the null byte, looking for the \n */
			while (++chunk_len < increase)
				if (cp[++len] == '\n')
					return cp; /* found it. read no more*/
			/* did not find the \n, so read and discard rest of line */
			do {
				c = getc(stream);
			} while (c != EOF && c != '\n');
			return cp;
		}
	}
}

void *strncpy_pad(void *dst, const void *src, size_t size, uint8_t pad)
{
	uint8_t *d = dst;
	const uint8_t *s = src;

	// logically the same as:  if (size < 1) IF size were signed.
	if (!size || size > (((size_t)-1) >> 1))
		return dst;

	while (*s && size) {
		*d++ = *s++;
		--size;
	}
	while (size--)
		*d++ = pad;

	return dst;
}

char *strnfcpy(char *dst, const char *src, int size)
{
	char *dptr;

	if (size < 1)
		return dst;
	dptr = dst;

	while (size--)
		if (!(*dptr++ = *src++))
			break;

	return dst;
}

char *strnzcpy(char *dst, const char *src, int size)
{
	char *dptr;

	if (size < 1)
		return dst;
	dptr = dst;

	while (--size)
		if (!(*dptr++ = *src++))
			return dst;
	*dptr = 0;

	return dst;
}

char *strnzcpylwr(char *dst, const char *src, int size)
{
	char *dptr;

	if (size < 1)
		return dst;
	dptr = dst;

	while (--size) {
		if (*src >= 'A' && *src <= 'Z') {
			*dptr = *src | 0x20;
		} else {
			*dptr = *src;
			if (!*src)
				return dst;
		}
		dptr++;
		src++;
	}
	*dptr = 0;

	return dst;
}

int strnzcpyn(char *dst, const char *src, int size)
{
	char *dptr;

	if (size < 1)
		return 0;
	dptr = dst;

	while (--size)
		if (!(*dptr++ = *src++))
			return (dptr-dst)-1;
	*dptr = 0;

	return (dptr-dst);
}

int strnzcpylwrn(char *dst, const char *src, int size)
{
	char *dptr;

	if (size < 1)
		return 0;
	dptr = dst;

	while (--size) {
		if (*src >= 'A' && *src <= 'Z') {
			*dptr = *src | 0x20;
		} else {
			*dptr = *src;
			if (!*src)
				return (dptr-dst);
		}
		dptr++;
		src++;
	}
	*dptr = 0;

	return (dptr-dst);

}

char *strnzcat(char *dst, const char *src, int size)
{
	char *dptr;

	if (size < 1)
		return dst;
	dptr = dst;

	while (size && *dptr) {
		size--; dptr++;
	}
	if (size)
	while (--size)
		if (!(*dptr++ = *src++))
			break;
	*dptr = 0;

	return dst;
}

/*
 * similar to strncat, but this one protects the dst buffer, AND it
 * assures that dst is properly NULL terminated upon completion.
 */
char *strnzcatn(char *dst, int size, const char *src, int src_max)
{
	char *dptr;

	if (size < 1)
		return dst;
	dptr = dst;

	while (size && *dptr)
		size--, dptr++;
	if (size)
	while (--size && src_max--)
		if (!(*dptr++ = *src++))
			break;
	*dptr = 0;
	return dst;
}


/*
 * strtok code, BUT returns empty token "" for adjacent delimiters. It also
 * returns leading and trailing tokens for leading and trailing delimiters
 * (strtok strips them away and does not return them). Several other issues
 * in strtok also impact this code
 *  - static pointer, not thread safe
 *  - mangled input string (requires a copy)
 * These 'bugs' were left in, so that this function is a straight drop in for
 * strtok, with the caveat of returning empty tokens for the 3 conditions.
 * Other than not being able to properly remove multiple adjacent tokens in
 * data such as arbitrary white space removal of text files, this is really
 * is what strtok should have been written to do from the beginning (IMHO).
 * A strtokm_r() should be trivial to write if we need thread safety, or need
 * to have multiple strtokm calls working at the same time, by just passing
 * in the *last pointer.
 * JimF, March 2015.
 */
char *strtokm(char *s1, const char *delims)
{
	static char *last = NULL;
	char *endp;

	if (!s1)
		s1 = last;
	if (!s1 || *s1 == 0)
		return last = NULL;
	endp = strpbrk(s1, delims);
	if (endp) {
		*endp = '\0';
		last = endp + 1;
	} else
		last = NULL;
	return s1;
}

unsigned int atou(const char *src)
{
	unsigned val;

	sscanf(src, "%u", &val);
	return val;
}

/*
 * _itoa replacement(s) but smarter/safer/better. _itoa is super useful, BUT
 * not a standard C function, and it does not protect buffer, and also has
 * what I deem strange behavior for negative numbers in non-base10 radix,
 * however, that output does 'mimic' what is often seen in some programming
 * type calculators.
 *
 * NOTE, differences between this function, and the _itoa found in visual C:
 *       1. we pass in buffer size, and protect the buffer (not in the VC interface of itoa)
 *       2. we handle all integers up to 63 bits
 *       3. vc only - signs return on base 10.  So _ltoa -666 base 16 return -29a, while
 *          vc's _itoa -666 base 16 returns fffffd66 (only 32 bit number for vc). I find
 *          our return much more in line, since the VC return is really forcing _itoa to
 *          return unsigned extension of the signed value, and not overly helpful. Both
 *          versions would return -666 for num of -666 if base was 10.
 */
MAYBE_INLINE const char *_lltoa(int64_t num, char *ret, int ret_sz, int base)
{
	char *p = ret, *p1 = ret;
	int64_t t;
	// first 35 bytes handle neg digits. byte 36 handles 0, and last 35 handle positive digits.
	const char bc[] = "zyxwvutsrqponmlkjihgfedcba987654321"
	                  "0"
	                  "123456789abcdefghijklmnopqrstuvwxyz";

	if (--ret_sz < 1)	// reduce ret_sz by 1 to handle the null
		return "";	// we can not touch ret, it is 0 bytes long.
	*ret = 0;
	// if we can not handle this base, bail.
	if (base < 2 || base > 36) return ret;
	// handle the possible '-' char also. (reduces ret_sz to fit that char)
	if (num < 0 && --ret_sz < 1)
		return ret;
	do {
		// build our string reversed.
		t = num;
		num /= base;
		*p++ = bc[35 + (t - num * base)];
		if (num && p - ret == ret_sz)
			break; // truncated MSB's but safe of buffer overflow.
	} while (num);

	if (t < 0) *p++ = '-'; // Apply negative sign
	*p-- = 0;
	// fast string-rev
	for (; p > p1; ++p1, --p) {
		*p1 ^= *p; *p ^= *p1; *p1 ^= *p;
	}
	return ret;
}

/*
 * similar to _lltoa, but for unsigned types. there were enough changes that I did not
 * want to make a single 'common' function.  Would have added many more if's to
 * and already semi-complex function.
 */
MAYBE_INLINE const char *_ulltoa(uint64_t num, char *ret, int ret_sz, int base)
{
	char *p = ret, *p1 = ret;
	uint64_t t;
	const char bc[] = "0123456789abcdefghijklmnopqrstuvwxyz";

	if (--ret_sz < 1)
		return "";
	*ret = 0;
	if (base < 2 || base > 36) return ret;
	do {
		t = num;
		num /= base;
		*p++ = bc[t - num * base];
		if (num && p - ret == ret_sz)
			break; // truncated MSB's but safe of buffer overflow.
	} while (num);

	*p-- = 0;
	// fast string-rev
	for (; p > p1; ++p1, --p) {
		*p1 ^= *p; *p ^= *p1; *p1 ^= *p;
	}
	return ret;
}

/*
 * these are the functions 'external' that other code in JtR can use. These
 * just call the 'common' code in the 2 inline functions.
 */
const char *jtr_itoa(int val, char *result, int rlen, int base)
{
	return _lltoa(val, result, rlen, base);
}

const char *jtr_utoa(unsigned int val, char *result, int rlen, int base)
{
	return _ulltoa((uint64_t)val, result, rlen, base);
}

const char *jtr_lltoa(int64_t val, char *result, int rlen, int base)
{
	return _lltoa(val, result, rlen, base);
}

const char *jtr_ulltoa(uint64_t val, char *result, int rlen, int base)
{
	return _ulltoa(val, result, rlen, base);
}

char *human_prefix(uint64_t num)
{
	char *out = mem_alloc_tiny(16, MEM_ALIGN_NONE);
	char prefixes[] = "\0KMGTPEZY";
	char *p = prefixes;

	while (p[1] && (num >= (100 << 10) ||
	                (!(num & 1023) && num >= (1 << 10)))) {
		num = (num >> 10);
		p++;
	}

	if (*p)
		snprintf(out, 16, "%u %c", (uint32_t)num, *p);
	else
		snprintf(out, 16, "%u ", (uint32_t)num);

	return out;
}

unsigned int lcm(unsigned int x, unsigned int y)
{
	unsigned int tmp, a, b;

	a = MAX(x, y);
	b = MIN(x, y);

	while (b) {
		tmp = b;
		b = a % b;
		a = tmp;
	}
	return x / a * y;
}

char *ltrim(char *str)
{
	char *out = str;

	while (*out == ' ' || *out == '\t')
		out++;

	return out;
}

char *rtrim(char *str)
{
	char *out = str + strlen(str) - 1;

	while (out >= str && (*out == ' ' || *out == '\t'))
		out--;

	*(out+1) = '\0';
	return str;
}