d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
17c6260f4c
freebsd-dev/contrib/file/softmagic.c
David E. O'Brien 17c6260f4c Virgin import of Christos Zoulas's FILE 4.21.
2007-05-24 21:59:38 +00:00

1640 lines
37 KiB
C
Raw Blame History

/*
* Copyright (c) Ian F. Darwin 1986-1995.
* Software written by Ian F. Darwin and others;
* maintained 1995-present by Christos Zoulas and others.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice immediately at the beginning of the file, without modification,
* this list of conditions, and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
* ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
/*
* softmagic - interpret variable magic from MAGIC
*/
#include "file.h"
#include "magic.h"
#include <string.h>
#include <ctype.h>
#include <stdlib.h>
#include <time.h>
#ifndef lint
FILE_RCSID("@(#)$File: softmagic.c,v 1.99 2007/05/08 14:44:18 christos Exp $")
#endif /* lint */
private int match(struct magic_set *, struct magic *, uint32_t,
const unsigned char *, size_t);
private int mget(struct magic_set *, const unsigned char *,
struct magic *, size_t, unsigned int);
private int magiccheck(struct magic_set *, struct magic *);
private int32_t mprint(struct magic_set *, struct magic *);
private void mdebug(uint32_t, const char *, size_t);
private int mcopy(struct magic_set *, union VALUETYPE *, int, int,
const unsigned char *, uint32_t, size_t, size_t);
private int mconvert(struct magic_set *, struct magic *);
private int print_sep(struct magic_set *, int);
private void cvt_8(union VALUETYPE *, const struct magic *);
private void cvt_16(union VALUETYPE *, const struct magic *);
private void cvt_32(union VALUETYPE *, const struct magic *);
private void cvt_64(union VALUETYPE *, const struct magic *);
/*
* softmagic - lookup one file in parsed, in-memory copy of database
* Passed the name and FILE * of one file to be typed.
*/
/*ARGSUSED1*/ /* nbytes passed for regularity, maybe need later */
protected int
file_softmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes)
{
struct mlist *ml;
int rv;
for (ml = ms->mlist->next; ml != ms->mlist; ml = ml->next)
if ((rv = match(ms, ml->magic, ml->nmagic, buf, nbytes)) != 0)
return rv;
return 0;
}
/*
* Go through the whole list, stopping if you find a match. Process all
* the continuations of that match before returning.
*
* We support multi-level continuations:
*
* At any time when processing a successful top-level match, there is a
* current continuation level; it represents the level of the last
* successfully matched continuation.
*
* Continuations above that level are skipped as, if we see one, it
* means that the continuation that controls them - i.e, the
* lower-level continuation preceding them - failed to match.
*
* Continuations below that level are processed as, if we see one,
* it means we've finished processing or skipping higher-level
* continuations under the control of a successful or unsuccessful
* lower-level continuation, and are now seeing the next lower-level
* continuation and should process it. The current continuation
* level reverts to the level of the one we're seeing.
*
* Continuations at the current level are processed as, if we see
* one, there's no lower-level continuation that may have failed.
*
* If a continuation matches, we bump the current continuation level
* so that higher-level continuations are processed.
*/
private int
match(struct magic_set *ms, struct magic *magic, uint32_t nmagic,
const unsigned char *s, size_t nbytes)
{
uint32_t magindex = 0;
unsigned int cont_level = 0;
int need_separator = 0;
int returnval = 0; /* if a match is found it is set to 1*/
int firstline = 1; /* a flag to print X\n X\n- X */
int printed_something = 0;
if (file_check_mem(ms, cont_level) == -1)
return -1;
for (magindex = 0; magindex < nmagic; magindex++) {
int flush;
ms->offset = magic[magindex].offset;
ms->line = magic[magindex].lineno;
/* if main entry matches, print it... */
flush = !mget(ms, s, &magic[magindex], nbytes, cont_level);
if (flush) {
if (magic[magindex].reln == '!')
flush = 0;
} else {
switch (magiccheck(ms, &magic[magindex])) {
case -1:
return -1;
case 0:
flush++;
break;
default:
break;
}
}
if (flush) {
/*
* main entry didn't match,
* flush its continuations
*/
while (magindex < nmagic - 1 &&
magic[magindex + 1].cont_level != 0)
magindex++;
continue;
}
/*
* If we are going to print something, we'll need to print
* a blank before we print something else.
*/
if (magic[magindex].desc[0]) {
need_separator = 1;
printed_something = 1;
if (print_sep(ms, firstline) == -1)
return -1;
}
if ((ms->c.li[cont_level].off = mprint(ms, &magic[magindex]))
== -1)
return -1;
/* and any continuations that match */
if (file_check_mem(ms, ++cont_level) == -1)
return -1;
while (magic[magindex+1].cont_level != 0 &&
++magindex < nmagic) {
ms->line = magic[magindex].lineno; /* for messages */
if (cont_level < magic[magindex].cont_level)
continue;
if (cont_level > magic[magindex].cont_level) {
/*
* We're at the end of the level
* "cont_level" continuations.
*/
cont_level = magic[magindex].cont_level;
}
ms->offset = magic[magindex].offset;
if (magic[magindex].flag & OFFADD) {
ms->offset +=
ms->c.li[cont_level - 1].off;
}
#ifdef ENABLE_CONDITIONALS
if (magic[magindex].cond == COND_ELSE ||
magic[magindex].cond == COND_ELIF) {
if (ms->c.li[cont_level].last_match == 1)
continue;
}
#endif
flush = !mget(ms, s, &magic[magindex], nbytes,
cont_level);
if (flush && magic[magindex].reln != '!')
continue;
switch (flush ? 1 : magiccheck(ms, &magic[magindex])) {
case -1:
return -1;
case 0:
#ifdef ENABLE_CONDITIONALS
ms->c.li[cont_level].last_match = 0;
#endif
break;
default:
#ifdef ENABLE_CONDITIONALS
ms->c.li[cont_level].last_match = 1;
#endif
if (magic[magindex].type != FILE_DEFAULT)
ms->c.li[cont_level].got_match = 1;
else if (ms->c.li[cont_level].got_match) {
ms->c.li[cont_level].got_match = 0;
break;
}
/*
* If we are going to print something,
* make sure that we have a separator first.
*/
if (magic[magindex].desc[0]) {
printed_something = 1;
if (print_sep(ms, firstline) == -1)
return -1;
}
/*
* This continuation matched. Print
* its message, with a blank before it
* if the previous item printed and
* this item isn't empty.
*/
/* space if previous printed */
if (need_separator
&& (magic[magindex].nospflag == 0)
&& (magic[magindex].desc[0] != '\0')) {
if (file_printf(ms, " ") == -1)
return -1;
need_separator = 0;
}
if ((ms->c.li[cont_level].off = mprint(ms, &magic[magindex])) == -1)
return -1;
if (magic[magindex].desc[0])
need_separator = 1;
/*
* If we see any continuations
* at a higher level,
* process them.
*/
if (file_check_mem(ms, ++cont_level) == -1)
return -1;
break;
}
}
firstline = 0;
if (printed_something)
returnval = 1;
if ((ms->flags & MAGIC_CONTINUE) == 0 && printed_something) {
return 1; /* don't keep searching */
}
}
return returnval; /* This is hit if -k is set or there is no match */
}
private int
check_fmt(struct magic_set *ms, struct magic *m)
{
regex_t rx;
int rc;
if (strchr(m->desc, '%') == NULL)
return 0;
rc = regcomp(&rx, "%[-0-9\\.]*s", REG_EXTENDED|REG_NOSUB);
if (rc) {
char errmsg[512];
(void)regerror(rc, &rx, errmsg, sizeof(errmsg));
file_magerror(ms, "regex error %d, (%s)", rc, errmsg);
return -1;
} else {
rc = regexec(&rx, m->desc, 0, 0, 0);
regfree(&rx);
return !rc;
}
}
#ifndef HAVE_STRNDUP
char * strndup(const char *, size_t);
char *
strndup(const char *str, size_t n)
{
size_t len;
char *copy;
len = strlen(str);
if (len > n)
len = n;
if (!(copy = malloc(len + 1)))
return (NULL);
(void) memcpy(copy, str, len + 1);
copy[len] = '\0';
return (copy);
}
#endif /* HAVE_STRNDUP */
private int32_t
mprint(struct magic_set *ms, struct magic *m)
{
uint64_t v;
int64_t t = 0;
char buf[512];
union VALUETYPE *p = &ms->ms_value;
switch (m->type) {
case FILE_BYTE:
v = file_signextend(ms, m, (uint64_t)p->b);
switch (check_fmt(ms, m)) {
case -1:
return -1;
case 1:
if (snprintf(buf, sizeof(buf), "%c",
(unsigned char)v) < 0)
return -1;
if (file_printf(ms, m->desc, buf) == -1)
return -1;
break;
default:
if (file_printf(ms, m->desc, (unsigned char) v) == -1)
return -1;
break;
}
t = ms->offset + sizeof(char);
break;
case FILE_SHORT:
case FILE_BESHORT:
case FILE_LESHORT:
v = file_signextend(ms, m, (uint64_t)p->h);
switch (check_fmt(ms, m)) {
case -1:
return -1;
case 1:
if (snprintf(buf, sizeof(buf), "%hu",
(unsigned short)v) < 0)
return -1;
if (file_printf(ms, m->desc, buf) == -1)
return -1;
break;
default:
if (file_printf(ms, m->desc, (unsigned short) v) == -1)
return -1;
break;
}
t = ms->offset + sizeof(short);
break;
case FILE_LONG:
case FILE_BELONG:
case FILE_LELONG:
case FILE_MELONG:
v = file_signextend(ms, m, (uint64_t)p->l);
switch (check_fmt(ms, m)) {
case -1:
return -1;
case 1:
if (snprintf(buf, sizeof(buf), "%u", (uint32_t)v) < 0)
return -1;
if (file_printf(ms, m->desc, buf) == -1)
return -1;
break;
default:
if (file_printf(ms, m->desc, (uint32_t) v) == -1)
return -1;
break;
}
t = ms->offset + sizeof(int32_t);
break;
case FILE_QUAD:
case FILE_BEQUAD:
case FILE_LEQUAD:
v = file_signextend(ms, m, p->q);
if (file_printf(ms, m->desc, (uint64_t) v) == -1)
return -1;
t = ms->offset + sizeof(int64_t);
break;
case FILE_STRING:
case FILE_PSTRING:
case FILE_BESTRING16:
case FILE_LESTRING16:
if (m->reln == '=' || m->reln == '!') {
if (file_printf(ms, m->desc, m->value.s) == -1)
return -1;
t = ms->offset + m->vallen;
}
else {
if (*m->value.s == '\0') {
char *cp = strchr(p->s,'\n');
if (cp)
*cp = '\0';
}
if (file_printf(ms, m->desc, p->s) == -1)
return -1;
t = ms->offset + strlen(p->s);
}
break;
case FILE_DATE:
case FILE_BEDATE:
case FILE_LEDATE:
case FILE_MEDATE:
if (file_printf(ms, m->desc, file_fmttime(p->l, 1)) == -1)
return -1;
t = ms->offset + sizeof(time_t);
break;
case FILE_LDATE:
case FILE_BELDATE:
case FILE_LELDATE:
case FILE_MELDATE:
if (file_printf(ms, m->desc, file_fmttime(p->l, 0)) == -1)
return -1;
t = ms->offset + sizeof(time_t);
break;
case FILE_QDATE:
case FILE_BEQDATE:
case FILE_LEQDATE:
if (file_printf(ms, m->desc, file_fmttime((uint32_t)p->q, 1))
== -1)
return -1;
t = ms->offset + sizeof(uint64_t);
break;
case FILE_QLDATE:
case FILE_BEQLDATE:
case FILE_LEQLDATE:
if (file_printf(ms, m->desc, file_fmttime((uint32_t)p->q, 0))
== -1)
return -1;
t = ms->offset + sizeof(uint64_t);
break;
case FILE_REGEX: {
char *cp;
int rval;
cp = strndup((const char *)ms->search.s, ms->search.rm_len);
if (cp == NULL) {
file_oomem(ms, ms->search.rm_len);
return -1;
}
rval = file_printf(ms, m->desc, cp);
free(cp);
if (rval == -1)
return -1;
if ((m->str_flags & REGEX_OFFSET_START))
t = ms->search.offset;
else
t = ms->search.offset + ms->search.rm_len;
break;
}
case FILE_SEARCH:
if (file_printf(ms, m->desc, m->value.s) == -1)
return -1;
if ((m->str_flags & REGEX_OFFSET_START))
t = ms->search.offset;
else
t = ms->search.offset + m->vallen;
break;
case FILE_DEFAULT:
if (file_printf(ms, m->desc, m->value.s) == -1)
return -1;
t = ms->offset;
break;
default:
file_magerror(ms, "invalid m->type (%d) in mprint()", m->type);
return -1;
}
return(t);
}
#define DO_CVT(fld, cast) \
if (m->num_mask) \
switch (m->mask_op & FILE_OPS_MASK) { \
case FILE_OPAND: \
p->fld &= cast m->num_mask; \
break; \
case FILE_OPOR: \
p->fld |= cast m->num_mask; \
break; \
case FILE_OPXOR: \
p->fld ^= cast m->num_mask; \
break; \
case FILE_OPADD: \
p->fld += cast m->num_mask; \
break; \
case FILE_OPMINUS: \
p->fld -= cast m->num_mask; \
break; \
case FILE_OPMULTIPLY: \
p->fld *= cast m->num_mask; \
break; \
case FILE_OPDIVIDE: \
p->fld /= cast m->num_mask; \
break; \
case FILE_OPMODULO: \
p->fld %= cast m->num_mask; \
break; \
} \
if (m->mask_op & FILE_OPINVERSE) \
p->fld = ~p->fld \
private void
cvt_8(union VALUETYPE *p, const struct magic *m)
{
DO_CVT(b, (uint8_t));
}
private void
cvt_16(union VALUETYPE *p, const struct magic *m)
{
DO_CVT(h, (uint16_t));
}
private void
cvt_32(union VALUETYPE *p, const struct magic *m)
{
DO_CVT(l, (uint32_t));
}
private void
cvt_64(union VALUETYPE *p, const struct magic *m)
{
DO_CVT(q, (uint64_t));
}
/*
* Convert the byte order of the data we are looking at
* While we're here, let's apply the mask operation
* (unless you have a better idea)
*/
private int
mconvert(struct magic_set *ms, struct magic *m)
{
union VALUETYPE *p = &ms->ms_value;
switch (m->type) {
case FILE_BYTE:
cvt_8(p, m);
return 1;
case FILE_SHORT:
cvt_16(p, m);
return 1;
case FILE_LONG:
case FILE_DATE:
case FILE_LDATE:
cvt_32(p, m);
return 1;
case FILE_QUAD:
case FILE_QDATE:
case FILE_QLDATE:
cvt_64(p, m);
return 1;
case FILE_STRING:
case FILE_BESTRING16:
case FILE_LESTRING16: {
size_t len;
/* Null terminate and eat *trailing* return */
p->s[sizeof(p->s) - 1] = '\0';
len = strlen(p->s);
if (len-- && p->s[len] == '\n')
p->s[len] = '\0';
return 1;
}
case FILE_PSTRING: {
char *ptr1 = p->s, *ptr2 = ptr1 + 1;
size_t len = *p->s;
if (len >= sizeof(p->s))
len = sizeof(p->s) - 1;
while (len--)
*ptr1++ = *ptr2++;
*ptr1 = '\0';
len = strlen(p->s);
if (len-- && p->s[len] == '\n')
p->s[len] = '\0';
return 1;
}
case FILE_BESHORT:
p->h = (short)((p->hs[0]<<8)|(p->hs[1]));
cvt_16(p, m);
return 1;
case FILE_BELONG:
case FILE_BEDATE:
case FILE_BELDATE:
p->l = (int32_t)
((p->hl[0]<<24)|(p->hl[1]<<16)|(p->hl[2]<<8)|(p->hl[3]));
cvt_32(p, m);
return 1;
case FILE_BEQUAD:
case FILE_BEQDATE:
case FILE_BEQLDATE:
p->q = (int64_t)
(((int64_t)p->hq[0]<<56)|((int64_t)p->hq[1]<<48)|
((int64_t)p->hq[2]<<40)|((int64_t)p->hq[3]<<32)|
(p->hq[4]<<24)|(p->hq[5]<<16)|(p->hq[6]<<8)|(p->hq[7]));
cvt_64(p, m);
return 1;
case FILE_LESHORT:
p->h = (short)((p->hs[1]<<8)|(p->hs[0]));
cvt_16(p, m);
return 1;
case FILE_LELONG:
case FILE_LEDATE:
case FILE_LELDATE:
p->l = (int32_t)
((p->hl[3]<<24)|(p->hl[2]<<16)|(p->hl[1]<<8)|(p->hl[0]));
cvt_32(p, m);
return 1;
case FILE_LEQUAD:
case FILE_LEQDATE:
case FILE_LEQLDATE:
p->q = (int64_t)
(((int64_t)p->hq[7]<<56)|((int64_t)p->hq[6]<<48)|
((int64_t)p->hq[5]<<40)|((int64_t)p->hq[4]<<32)|
(p->hq[3]<<24)|(p->hq[2]<<16)|(p->hq[1]<<8)|(p->hq[0]));
cvt_64(p, m);
return 1;
case FILE_MELONG:
case FILE_MEDATE:
case FILE_MELDATE:
p->l = (int32_t)
((p->hl[1]<<24)|(p->hl[0]<<16)|(p->hl[3]<<8)|(p->hl[2]));
cvt_32(p, m);
return 1;
case FILE_REGEX:
case FILE_SEARCH:
case FILE_DEFAULT:
return 1;
default:
file_magerror(ms, "invalid type %d in mconvert()", m->type);
return 0;
}
}
private void
mdebug(uint32_t offset, const char *str, size_t len)
{
(void) fprintf(stderr, "mget @%d: ", offset);
file_showstr(stderr, str, len);
(void) fputc('\n', stderr);
(void) fputc('\n', stderr);
}
private int
mcopy(struct magic_set *ms, union VALUETYPE *p, int type, int indir,
const unsigned char *s, uint32_t offset, size_t nbytes, size_t linecnt)
{
/*
* Note: FILE_SEARCH and FILE_REGEX do not actually copy
* anything, but setup pointers into the source
*/
if (indir == 0) {
switch (type) {
case FILE_SEARCH:
ms->search.s = (const char *)s + offset;
ms->search.s_len = nbytes - offset;
return 0;
case FILE_REGEX: {
/*
* offset is interpreted as last line to search,
* (starting at 1), not as bytes-from start-of-file
*/
const char *b;
const char *c;
const char *last; /* end of search region */
const char *buf; /* start of search region */
size_t lines;
if (s == NULL) {
ms->search.s_len = 0;
ms->search.s = NULL;
return 0;
}
buf = (const char *)s + offset;
last = (const char *)s + nbytes;
/* mget() guarantees buf <= last */
for (lines = linecnt, b = buf;
lines && ((b = strchr(c = b, '\n')) || (b = strchr(c, '\r')));
lines--, b++) {
last = b;
if (b[0] == '\r' && b[1] == '\n')
b++;
}
if (lines)
last = (const char *)s + nbytes;
ms->search.s = buf;
ms->search.s_len = last - buf;
ms->search.offset = offset;
ms->search.rm_len = 0;
return 0;
}
case FILE_BESTRING16:
case FILE_LESTRING16: {
const unsigned char *src = s + offset;
const unsigned char *esrc = s + nbytes;
char *dst = p->s;
char *edst = &p->s[sizeof(p->s) - 1];
if (type == FILE_BESTRING16)
src++;
/* check for pointer overflow */
if (src < s) {
file_magerror(ms, "invalid offset %zu in mcopy()",
offset);
return -1;
}
for (/*EMPTY*/; src < esrc; src++, dst++) {
if (dst < edst)
*dst = *src++;
else
break;
if (*dst == '\0')
*dst = ' ';
}
*edst = '\0';
return 0;
}
case FILE_STRING: /* XXX - these two should not need */
case FILE_PSTRING: /* to copy anything, but do anyway. */
default:
break;
}
}
if (offset >= nbytes) {
(void)memset(p, '\0', sizeof(*p));
return 0;
}
if (nbytes - offset < sizeof(*p))
nbytes = nbytes - offset;
else
nbytes = sizeof(*p);
(void)memcpy(p, s + offset, nbytes);
/*
* the usefulness of padding with zeroes eludes me, it
* might even cause problems
*/
if (nbytes < sizeof(*p))
(void)memset(((char *)(void *)p) + nbytes, '\0',
sizeof(*p) - nbytes);
return 0;
}
private int
mget(struct magic_set *ms, const unsigned char *s,
struct magic *m, size_t nbytes, unsigned int cont_level)
{
uint32_t offset = ms->offset;
uint32_t count = m->str_count;
union VALUETYPE *p = &ms->ms_value;
if (mcopy(ms, p, m->type, m->flag & INDIR, s, offset, nbytes, count) == -1)
return -1;
if ((ms->flags & MAGIC_DEBUG) != 0) {
mdebug(offset, (char *)(void *)p, sizeof(union VALUETYPE));
file_mdump(m);
}
if (m->flag & INDIR) {
int off = m->in_offset;
if (m->in_op & FILE_OPINDIRECT) {
const union VALUETYPE *q =
((const void *)(s + offset + off));
switch (m->in_type) {
case FILE_BYTE:
off = q->b;
break;
case FILE_SHORT:
off = q->h;
break;
case FILE_BESHORT:
off = (short)((q->hs[0]<<8)|(q->hs[1]));
break;
case FILE_LESHORT:
off = (short)((q->hs[1]<<8)|(q->hs[0]));
break;
case FILE_LONG:
off = q->l;
break;
case FILE_BELONG:
off = (int32_t)((q->hl[0]<<24)|(q->hl[1]<<16)|
(q->hl[2]<<8)|(q->hl[3]));
break;
case FILE_LELONG:
off = (int32_t)((q->hl[3]<<24)|(q->hl[2]<<16)|
(q->hl[1]<<8)|(q->hl[0]));
break;
case FILE_MELONG:
off = (int32_t)((q->hl[1]<<24)|(q->hl[0]<<16)|
(q->hl[3]<<8)|(q->hl[2]));
break;
}
}
switch (m->in_type) {
case FILE_BYTE:
if (nbytes < (offset + 1))
return 0;
if (off) {
switch (m->in_op & FILE_OPS_MASK) {
case FILE_OPAND:
offset = p->b & off;
break;
case FILE_OPOR:
offset = p->b | off;
break;
case FILE_OPXOR:
offset = p->b ^ off;
break;
case FILE_OPADD:
offset = p->b + off;
break;
case FILE_OPMINUS:
offset = p->b - off;
break;
case FILE_OPMULTIPLY:
offset = p->b * off;
break;
case FILE_OPDIVIDE:
offset = p->b / off;
break;
case FILE_OPMODULO:
offset = p->b % off;
break;
}
} else
offset = p->b;
if (m->in_op & FILE_OPINVERSE)
offset = ~offset;
break;
case FILE_BESHORT:
if (nbytes < (offset + 2))
return 0;
if (off) {
switch (m->in_op & FILE_OPS_MASK) {
case FILE_OPAND:
offset = (short)((p->hs[0]<<8)|
(p->hs[1])) &
off;
break;
case FILE_OPOR:
offset = (short)((p->hs[0]<<8)|
(p->hs[1])) |
off;
break;
case FILE_OPXOR:
offset = (short)((p->hs[0]<<8)|
(p->hs[1])) ^
off;
break;
case FILE_OPADD:
offset = (short)((p->hs[0]<<8)|
(p->hs[1])) +
off;
break;
case FILE_OPMINUS:
offset = (short)((p->hs[0]<<8)|
(p->hs[1])) -
off;
break;
case FILE_OPMULTIPLY:
offset = (short)((p->hs[0]<<8)|
(p->hs[1])) *
off;
break;
case FILE_OPDIVIDE:
offset = (short)((p->hs[0]<<8)|
(p->hs[1])) /
off;
break;
case FILE_OPMODULO:
offset = (short)((p->hs[0]<<8)|
(p->hs[1])) %
off;
break;
}
} else
offset = (short)((p->hs[0]<<8)|
(p->hs[1]));
if (m->in_op & FILE_OPINVERSE)
offset = ~offset;
break;
case FILE_LESHORT:
if (nbytes < (offset + 2))
return 0;
if (off) {
switch (m->in_op & FILE_OPS_MASK) {
case FILE_OPAND:
offset = (short)((p->hs[1]<<8)|
(p->hs[0])) &
off;
break;
case FILE_OPOR:
offset = (short)((p->hs[1]<<8)|
(p->hs[0])) |
off;
break;
case FILE_OPXOR:
offset = (short)((p->hs[1]<<8)|
(p->hs[0])) ^
off;
break;
case FILE_OPADD:
offset = (short)((p->hs[1]<<8)|
(p->hs[0])) +
off;
break;
case FILE_OPMINUS:
offset = (short)((p->hs[1]<<8)|
(p->hs[0])) -
off;
break;
case FILE_OPMULTIPLY:
offset = (short)((p->hs[1]<<8)|
(p->hs[0])) *
off;
break;
case FILE_OPDIVIDE:
offset = (short)((p->hs[1]<<8)|
(p->hs[0])) /
off;
break;
case FILE_OPMODULO:
offset = (short)((p->hs[1]<<8)|
(p->hs[0])) %
off;
break;
}
} else
offset = (short)((p->hs[1]<<8)|
(p->hs[0]));
if (m->in_op & FILE_OPINVERSE)
offset = ~offset;
break;
case FILE_SHORT:
if (nbytes < (offset + 2))
return 0;
if (off) {
switch (m->in_op & FILE_OPS_MASK) {
case FILE_OPAND:
offset = p->h & off;
break;
case FILE_OPOR:
offset = p->h | off;
break;
case FILE_OPXOR:
offset = p->h ^ off;
break;
case FILE_OPADD:
offset = p->h + off;
break;
case FILE_OPMINUS:
offset = p->h - off;
break;
case FILE_OPMULTIPLY:
offset = p->h * off;
break;
case FILE_OPDIVIDE:
offset = p->h / off;
break;
case FILE_OPMODULO:
offset = p->h % off;
break;
}
}
else
offset = p->h;
if (m->in_op & FILE_OPINVERSE)
offset = ~offset;
break;
case FILE_BELONG:
if (nbytes < (offset + 4))
return 0;
if (off) {
switch (m->in_op & FILE_OPS_MASK) {
case FILE_OPAND:
offset = (int32_t)((p->hl[0]<<24)|
(p->hl[1]<<16)|
(p->hl[2]<<8)|
(p->hl[3])) &
off;
break;
case FILE_OPOR:
offset = (int32_t)((p->hl[0]<<24)|
(p->hl[1]<<16)|
(p->hl[2]<<8)|
(p->hl[3])) |
off;
break;
case FILE_OPXOR:
offset = (int32_t)((p->hl[0]<<24)|
(p->hl[1]<<16)|
(p->hl[2]<<8)|
(p->hl[3])) ^
off;
break;
case FILE_OPADD:
offset = (int32_t)((p->hl[0]<<24)|
(p->hl[1]<<16)|
(p->hl[2]<<8)|
(p->hl[3])) +
off;
break;
case FILE_OPMINUS:
offset = (int32_t)((p->hl[0]<<24)|
(p->hl[1]<<16)|
(p->hl[2]<<8)|
(p->hl[3])) -
off;
break;
case FILE_OPMULTIPLY:
offset = (int32_t)((p->hl[0]<<24)|
(p->hl[1]<<16)|
(p->hl[2]<<8)|
(p->hl[3])) *
off;
break;
case FILE_OPDIVIDE:
offset = (int32_t)((p->hl[0]<<24)|
(p->hl[1]<<16)|
(p->hl[2]<<8)|
(p->hl[3])) /
off;
break;
case FILE_OPMODULO:
offset = (int32_t)((p->hl[0]<<24)|
(p->hl[1]<<16)|
(p->hl[2]<<8)|
(p->hl[3])) %
off;
break;
}
} else
offset = (int32_t)((p->hl[0]<<24)|
(p->hl[1]<<16)|
(p->hl[2]<<8)|
(p->hl[3]));
if (m->in_op & FILE_OPINVERSE)
offset = ~offset;
break;
case FILE_LELONG:
if (nbytes < (offset + 4))
return 0;
if (off) {
switch (m->in_op & FILE_OPS_MASK) {
case FILE_OPAND:
offset = (int32_t)((p->hl[3]<<24)|
(p->hl[2]<<16)|
(p->hl[1]<<8)|
(p->hl[0])) &
off;
break;
case FILE_OPOR:
offset = (int32_t)((p->hl[3]<<24)|
(p->hl[2]<<16)|
(p->hl[1]<<8)|
(p->hl[0])) |
off;
break;
case FILE_OPXOR:
offset = (int32_t)((p->hl[3]<<24)|
(p->hl[2]<<16)|
(p->hl[1]<<8)|
(p->hl[0])) ^
off;
break;
case FILE_OPADD:
offset = (int32_t)((p->hl[3]<<24)|
(p->hl[2]<<16)|
(p->hl[1]<<8)|
(p->hl[0])) +
off;
break;
case FILE_OPMINUS:
offset = (int32_t)((p->hl[3]<<24)|
(p->hl[2]<<16)|
(p->hl[1]<<8)|
(p->hl[0])) -
off;
break;
case FILE_OPMULTIPLY:
offset = (int32_t)((p->hl[3]<<24)|
(p->hl[2]<<16)|
(p->hl[1]<<8)|
(p->hl[0])) *
off;
break;
case FILE_OPDIVIDE:
offset = (int32_t)((p->hl[3]<<24)|
(p->hl[2]<<16)|
(p->hl[1]<<8)|
(p->hl[0])) /
off;
break;
case FILE_OPMODULO:
offset = (int32_t)((p->hl[3]<<24)|
(p->hl[2]<<16)|
(p->hl[1]<<8)|
(p->hl[0])) %
off;
break;
}
} else
offset = (int32_t)((p->hl[3]<<24)|
(p->hl[2]<<16)|
(p->hl[1]<<8)|
(p->hl[0]));
if (m->in_op & FILE_OPINVERSE)
offset = ~offset;
break;
case FILE_MELONG:
if (nbytes < (offset + 4))
return 0;
if (off) {
switch (m->in_op & FILE_OPS_MASK) {
case FILE_OPAND:
offset = (int32_t)((p->hl[1]<<24)|
(p->hl[0]<<16)|
(p->hl[3]<<8)|
(p->hl[2])) &
off;
break;
case FILE_OPOR:
offset = (int32_t)((p->hl[1]<<24)|
(p->hl[0]<<16)|
(p->hl[3]<<8)|
(p->hl[2])) |
off;
break;
case FILE_OPXOR:
offset = (int32_t)((p->hl[1]<<24)|
(p->hl[0]<<16)|
(p->hl[3]<<8)|
(p->hl[2])) ^
off;
break;
case FILE_OPADD:
offset = (int32_t)((p->hl[1]<<24)|
(p->hl[0]<<16)|
(p->hl[3]<<8)|
(p->hl[2])) +
off;
break;
case FILE_OPMINUS:
offset = (int32_t)((p->hl[1]<<24)|
(p->hl[0]<<16)|
(p->hl[3]<<8)|
(p->hl[2])) -
off;
break;
case FILE_OPMULTIPLY:
offset = (int32_t)((p->hl[1]<<24)|
(p->hl[0]<<16)|
(p->hl[3]<<8)|
(p->hl[2])) *
off;
break;
case FILE_OPDIVIDE:
offset = (int32_t)((p->hl[1]<<24)|
(p->hl[0]<<16)|
(p->hl[3]<<8)|
(p->hl[2])) /
off;
break;
case FILE_OPMODULO:
offset = (int32_t)((p->hl[1]<<24)|
(p->hl[0]<<16)|
(p->hl[3]<<8)|
(p->hl[2])) %
off;
break;
}
} else
offset = (int32_t)((p->hl[1]<<24)|
(p->hl[0]<<16)|
(p->hl[3]<<8)|
(p->hl[2]));
if (m->in_op & FILE_OPINVERSE)
offset = ~offset;
break;
case FILE_LONG:
if (nbytes < (offset + 4))
return 0;
if (off) {
switch (m->in_op & FILE_OPS_MASK) {
case FILE_OPAND:
offset = p->l & off;
break;
case FILE_OPOR:
offset = p->l | off;
break;
case FILE_OPXOR:
offset = p->l ^ off;
break;
case FILE_OPADD:
offset = p->l + off;
break;
case FILE_OPMINUS:
offset = p->l - off;
break;
case FILE_OPMULTIPLY:
offset = p->l * off;
break;
case FILE_OPDIVIDE:
offset = p->l / off;
break;
case FILE_OPMODULO:
offset = p->l % off;
break;
/* case TOOMANYSWITCHBLOCKS:
* ugh = p->eye % m->strain;
* rub;
* case BEER:
* off = p->tab & m->in_gest;
* sleep;
*/
}
} else
offset = p->l;
if (m->in_op & FILE_OPINVERSE)
offset = ~offset;
break;
}
if (m->flag & INDIROFFADD)
offset += ms->c.li[cont_level-1].off;
if (mcopy(ms, p, m->type, 0, s, offset, nbytes, count) == -1)
return -1;
ms->offset = offset;
if ((ms->flags & MAGIC_DEBUG) != 0) {
mdebug(offset, (char *)(void *)p,
sizeof(union VALUETYPE));
file_mdump(m);
}
}
/* Verify we have enough data to match magic type */
switch (m->type) {
case FILE_BYTE:
if (nbytes < (offset + 1)) /* should alway be true */
return 0;
break;
case FILE_SHORT:
case FILE_BESHORT:
case FILE_LESHORT:
if (nbytes < (offset + 2))
return 0;
break;
case FILE_LONG:
case FILE_BELONG:
case FILE_LELONG:
case FILE_MELONG:
case FILE_DATE:
case FILE_BEDATE:
case FILE_LEDATE:
case FILE_MEDATE:
case FILE_LDATE:
case FILE_BELDATE:
case FILE_LELDATE:
case FILE_MELDATE:
if (nbytes < (offset + 4))
return 0;
break;
case FILE_STRING:
case FILE_PSTRING:
case FILE_SEARCH:
if (nbytes < (offset + m->vallen))
return 0;
break;
case FILE_REGEX:
if (nbytes < offset)
return 0;
break;
case FILE_DEFAULT: /* nothing to check */
default:
break;
}
if (!mconvert(ms, m))
return 0;
return 1;
}
private uint64_t
file_strncmp(const char *s1, const char *s2, size_t len, uint32_t flags)
{
/*
* Convert the source args to unsigned here so that (1) the
* compare will be unsigned as it is in strncmp() and (2) so
* the ctype functions will work correctly without extra
* casting.
*/
const unsigned char *a = (const unsigned char *)s1;
const unsigned char *b = (const unsigned char *)s2;
uint64_t v;
/*
* What we want here is:
* v = strncmp(m->value.s, p->s, m->vallen);
* but ignoring any nulls. bcmp doesn't give -/+/0
* and isn't universally available anyway.
*/
v = 0;
if (0L == flags) { /* normal string: do it fast */
while (len-- > 0)
if ((v = *b++ - *a++) != '\0')
break;
}
else { /* combine the others */
while (len-- > 0) {
if ((flags & STRING_IGNORE_LOWERCASE) &&
islower(*a)) {
if ((v = tolower(*b++) - *a++) != '\0')
break;
}
else if ((flags & STRING_IGNORE_UPPERCASE) &&
isupper(*a)) {
if ((v = toupper(*b++) - *a++) != '\0')
break;
}
else if ((flags & STRING_COMPACT_BLANK) &&
isspace(*a)) {
a++;
if (isspace(*b++)) {
while (isspace(*b))
b++;
}
else {
v = 1;
break;
}
}
else if ((flags & STRING_COMPACT_OPTIONAL_BLANK) &&
isspace(*a)) {
a++;
while (isspace(*b))
b++;
}
else {
if ((v = *b++ - *a++) != '\0')
break;
}
}
}
return v;
}
private uint64_t
file_strncmp16(const char *a, const char *b, size_t len, uint32_t flags)
{
/*
* XXX - The 16-bit string compare probably needs to be done
* differently, especially if the flags are to be supported.
* At the moment, I am unsure.
*/
flags = 0;
return file_strncmp(a, b, len, flags);
}
private int
magiccheck(struct magic_set *ms, struct magic *m)
{
uint64_t l = m->value.q;
uint64_t v;
int matched;
union VALUETYPE *p = &ms->ms_value;
switch (m->type) {
case FILE_BYTE:
v = p->b;
break;
case FILE_SHORT:
case FILE_BESHORT:
case FILE_LESHORT:
v = p->h;
break;
case FILE_LONG:
case FILE_BELONG:
case FILE_LELONG:
case FILE_MELONG:
case FILE_DATE:
case FILE_BEDATE:
case FILE_LEDATE:
case FILE_MEDATE:
case FILE_LDATE:
case FILE_BELDATE:
case FILE_LELDATE:
case FILE_MELDATE:
v = p->l;
break;
case FILE_QUAD:
case FILE_LEQUAD:
case FILE_BEQUAD:
case FILE_QDATE:
case FILE_BEQDATE:
case FILE_LEQDATE:
case FILE_QLDATE:
case FILE_BEQLDATE:
case FILE_LEQLDATE:
v = p->q;
break;
case FILE_DEFAULT:
l = 0;
v = 0;
break;
case FILE_STRING:
case FILE_PSTRING:
l = 0;
v = file_strncmp(m->value.s, p->s, (size_t)m->vallen, m->str_flags);
break;
case FILE_BESTRING16:
case FILE_LESTRING16:
l = 0;
v = file_strncmp16(m->value.s, p->s, (size_t)m->vallen, m->str_flags);
break;
case FILE_SEARCH: { /* search ms->search.s for the string m->value.s */
size_t slen;
size_t idx;
if (ms->search.s == NULL)
return 0;
slen = MIN(m->vallen, sizeof(m->value.s));
l = 0;
v = 0;
ms->search.offset = m->offset;
for (idx = 0; m->str_count == 0 || idx < m->str_count; idx++) {
if (slen + idx > ms->search.s_len)
break;
v = file_strncmp(m->value.s, ms->search.s + idx, slen, m->str_flags);
if (v == 0) { /* found match */
ms->search.offset = m->offset + idx;
break;
}
}
break;
}
case FILE_REGEX: {
int rc;
regex_t rx;
char errmsg[512];
if (ms->search.s == NULL)
return 0;
l = 0;
rc = regcomp(&rx, m->value.s,
REG_EXTENDED|REG_NEWLINE|
((m->str_flags & STRING_IGNORE_CASE) ? REG_ICASE : 0));
if (rc) {
(void)regerror(rc, &rx, errmsg, sizeof(errmsg));
file_magerror(ms, "regex error %d, (%s)",
rc, errmsg);
v = (uint64_t)-1;
}
else {
regmatch_t pmatch[1];
#ifndef REG_STARTEND
#define REG_STARTEND 0
size_t l = ms->search.s_len - 1;
char c = ms->search.s[l];
((char *)(intptr_t)ms->search.s)[l] = '\0';
#else
pmatch[0].rm_so = 0;
pmatch[0].rm_eo = ms->search.s_len;
#endif
rc = regexec(&rx, (const char *)ms->search.s,
1, pmatch, REG_STARTEND);
#if REG_STARTEND == 0
((char *)(intptr_t)ms->search.s)[l] = c;
#endif
switch (rc) {
case 0:
ms->search.s += (int)pmatch[0].rm_so;
ms->search.offset += (size_t)pmatch[0].rm_so;
ms->search.rm_len =
(size_t)(pmatch[0].rm_eo - pmatch[0].rm_so);
v = 0;
break;
case REG_NOMATCH:
v = 1;
break;
default:
(void)regerror(rc, &rx, errmsg, sizeof(errmsg));
file_magerror(ms, "regexec error %d, (%s)",
rc, errmsg);
v = (uint64_t)-1;
break;
}
regfree(&rx);
}
if (v == (uint64_t)-1)
return -1;
break;
}
default:
file_magerror(ms, "invalid type %d in magiccheck()", m->type);
return -1;
}
v = file_signextend(ms, m, v);
switch (m->reln) {
case 'x':
if ((ms->flags & MAGIC_DEBUG) != 0)
(void) fprintf(stderr, "%llu == *any* = 1\n",
(unsigned long long)v);
matched = 1;
break;
case '!':
matched = v != l;
if ((ms->flags & MAGIC_DEBUG) != 0)
(void) fprintf(stderr, "%llu != %llu = %d\n",
(unsigned long long)v, (unsigned long long)l,
matched);
break;
case '=':
matched = v == l;
if ((ms->flags & MAGIC_DEBUG) != 0)
(void) fprintf(stderr, "%llu == %llu = %d\n",
(unsigned long long)v, (unsigned long long)l,
matched);
break;
case '>':
if (m->flag & UNSIGNED) {
matched = v > l;
if ((ms->flags & MAGIC_DEBUG) != 0)
(void) fprintf(stderr, "%llu > %llu = %d\n",
(unsigned long long)v,
(unsigned long long)l, matched);
}
else {
matched = (int64_t) v > (int64_t) l;
if ((ms->flags & MAGIC_DEBUG) != 0)
(void) fprintf(stderr, "%lld > %lld = %d\n",
(long long)v, (long long)l, matched);
}
break;
case '<':
if (m->flag & UNSIGNED) {
matched = v < l;
if ((ms->flags & MAGIC_DEBUG) != 0)
(void) fprintf(stderr, "%llu < %llu = %d\n",
(unsigned long long)v,
(unsigned long long)l, matched);
}
else {
matched = (int64_t) v < (int64_t) l;
if ((ms->flags & MAGIC_DEBUG) != 0)
(void) fprintf(stderr, "%lld < %lld = %d\n",
(long long)v, (long long)l, matched);
}
break;
case '&':
matched = (v & l) == l;
if ((ms->flags & MAGIC_DEBUG) != 0)
(void) fprintf(stderr, "((%llx & %llx) == %llx) = %d\n",
(unsigned long long)v, (unsigned long long)l,
(unsigned long long)l, matched);
break;
case '^':
matched = (v & l) != l;
if ((ms->flags & MAGIC_DEBUG) != 0)
(void) fprintf(stderr, "((%llx & %llx) != %llx) = %d\n",
(unsigned long long)v, (unsigned long long)l,
(unsigned long long)l, matched);
break;
default:
matched = 0;
file_magerror(ms, "cannot happen: invalid relation `%c'",
m->reln);
return -1;
}
return matched;
}
private int
print_sep(struct magic_set *ms, int firstline)
{
if (firstline)
return 0;
/*
* we found another match
* put a newline and '-' to do some simple formatting
*/
return file_printf(ms, "\n- ");
}
Reference in New Issue View Git Blame Copy Permalink