glibc/iconvdata/iso-2022-cn-ext.c
Charles Fol f9dc609e06 iconv: ISO-2022-CN-EXT: fix out-of-bound writes when writing escape sequence (CVE-2024-2961)
ISO-2022-CN-EXT uses escape sequences to indicate character set changes
(as specified by RFC 1922).  While the SOdesignation has the expected
bounds checks, neither SS2designation nor SS3designation have its;
allowing a write overflow of 1, 2, or 3 bytes with fixed values:
'$+I', '$+J', '$+K', '$+L', '$+M', or '$*H'.

Checked on aarch64-linux-gnu.

Co-authored-by: Adhemerval Zanella  <adhemerval.zanella@linaro.org>
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
2024-04-17 09:21:40 -03:00

687 lines
24 KiB
C

/* Conversion module for ISO-2022-CN-EXT.
Copyright (C) 2000-2024 Free Software Foundation, Inc.
This file is part of the GNU C Library.
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<https://www.gnu.org/licenses/>. */
#include <dlfcn.h>
#include <gconv.h>
#include <stdint.h>
#include <stdlib.h>
#include <string.h>
#include "gb2312.h"
#include "iso-ir-165.h"
#include "cns11643.h"
#include "cns11643l1.h"
#include "cns11643l2.h"
#include <libc-diag.h>
#include <assert.h>
/* This makes obvious what everybody knows: 0x1b is the Esc character. */
#define ESC 0x1b
/* We have single-byte shift-in and shift-out sequences, and the single
shift sequences SS2 and SS3 which replaces the SS2/SS3 designation for
the next two bytes. */
#define SI 0x0f
#define SO 0x0e
#define SS2_0 ESC
#define SS2_1 0x4e
#define SS3_0 ESC
#define SS3_1 0x4f
/* Definitions used in the body of the `gconv' function. */
#define CHARSET_NAME "ISO-2022-CN-EXT//"
#define DEFINE_INIT 1
#define DEFINE_FINI 1
#define ONE_DIRECTION 0
#define FROM_LOOP from_iso2022cn_ext_loop
#define TO_LOOP to_iso2022cn_ext_loop
#define FROM_LOOP_MIN_NEEDED_FROM 1
#define FROM_LOOP_MAX_NEEDED_FROM 4
#define FROM_LOOP_MIN_NEEDED_TO 4
#define FROM_LOOP_MAX_NEEDED_TO 4
#define TO_LOOP_MIN_NEEDED_FROM 4
#define TO_LOOP_MAX_NEEDED_FROM 4
#define TO_LOOP_MIN_NEEDED_TO 1
#define TO_LOOP_MAX_NEEDED_TO 6
#define PREPARE_LOOP \
int save_set; \
int *setp = &data->__statep->__count;
#define EXTRA_LOOP_ARGS , setp
/* The charsets GB/T 12345-90, GB 7589-87, GB/T 13131-9X, GB 7590-87,
and GB/T 13132-9X are not registered to the best of my knowledge and
therefore have no escape sequence assigned. We cannot handle them
for this reason. Tell the implementation about this. */
#define X12345 '\0'
#define X7589 '\0'
#define X13131 '\0'
#define X7590 '\0'
#define X13132 '\0'
/* The COUNT element of the state keeps track of the currently selected
character set. The possible values are: */
enum
{
ASCII_set = 0,
GB2312_set,
GB12345_set,
CNS11643_1_set,
ISO_IR_165_set,
SO_mask = 7,
GB7589_set = 1 << 3,
GB13131_set = 2 << 3,
CNS11643_2_set = 3 << 3,
SS2_mask = 3 << 3,
GB7590_set = 1 << 5,
GB13132_set = 2 << 5,
CNS11643_3_set = 3 << 5,
CNS11643_4_set = 4 << 5,
CNS11643_5_set = 5 << 5,
CNS11643_6_set = 6 << 5,
CNS11643_7_set = 7 << 5,
SS3_mask = 7 << 5,
#define CURRENT_MASK (SO_mask | SS2_mask | SS3_mask)
GB2312_ann = 1 << 8,
GB12345_ann = 2 << 8,
CNS11643_1_ann = 3 << 8,
ISO_IR_165_ann = 4 << 8,
SO_ann = 7 << 8,
GB7589_ann = 1 << 11,
GB13131_ann = 2 << 11,
CNS11643_2_ann = 3 << 11,
SS2_ann = 3 << 11,
GB7590_ann = 1 << 13,
GB13132_ann = 2 << 13,
CNS11643_3_ann = 3 << 13,
CNS11643_4_ann = 4 << 13,
CNS11643_5_ann = 5 << 13,
CNS11643_6_ann = 6 << 13,
CNS11643_7_ann = 7 << 13,
SS3_ann = 7 << 13
};
/* Since this is a stateful encoding we have to provide code which resets
the output state to the initial state. This has to be done during the
flushing. */
#define EMIT_SHIFT_TO_INIT \
if (data->__statep->__count >> 3 != ASCII_set) \
{ \
if (FROM_DIRECTION) \
/* It's easy, we don't have to emit anything, we just reset the \
state for the input. */ \
data->__statep->__count = ASCII_set << 3; \
else \
{ \
/* We are not in the initial state. To switch back we have \
to emit `SI'. */ \
if (__glibc_unlikely (outbuf == outend)) \
/* We don't have enough room in the output buffer. */ \
status = __GCONV_FULL_OUTPUT; \
else \
{ \
/* Write out the shift sequence. */ \
*outbuf++ = SI; \
if (data->__flags & __GCONV_IS_LAST) \
*irreversible += 1; \
data->__statep->__count = ASCII_set << 3; \
} \
} \
}
/* Since we might have to reset input pointer we must be able to save
and restore the state. */
#define SAVE_RESET_STATE(Save) \
if (Save) \
save_set = *setp; \
else \
*setp = save_set
/* First define the conversion function from ISO-2022-CN to UCS4. */
#define MIN_NEEDED_INPUT FROM_LOOP_MIN_NEEDED_FROM
#define MAX_NEEDED_INPUT FROM_LOOP_MAX_NEEDED_FROM
#define MIN_NEEDED_OUTPUT FROM_LOOP_MIN_NEEDED_TO
#define MAX_NEEDED_OUTPUT FROM_LOOP_MAX_NEEDED_TO
#define LOOPFCT FROM_LOOP
#define BODY \
{ \
uint32_t ch = *inptr; \
\
/* This is a 7bit character set, disallow all 8bit characters. */ \
if (ch > 0x7f) \
STANDARD_FROM_LOOP_ERR_HANDLER (1); \
\
/* Recognize escape sequences. */ \
if (ch == ESC) \
{ \
/* There are three kinds of escape sequences we have to handle: \
- those announcing the use of GB and CNS characters on the \
line; we can simply ignore them \
- the initial byte of the SS2 sequence. \
- the initial byte of the SS3 sequence. \
*/ \
if (inptr + 2 > inend \
|| (inptr[1] == '$' \
&& (inptr + 3 > inend \
|| (inptr[2] == ')' && inptr + 4 > inend) \
|| (inptr[2] == '*' && inptr + 4 > inend) \
|| (inptr[2] == '+' && inptr + 4 > inend))) \
|| (inptr[1] == SS2_1 && inptr + 4 > inend) \
|| (inptr[1] == SS3_1 && inptr + 4 > inend)) \
{ \
result = __GCONV_INCOMPLETE_INPUT; \
break; \
} \
if (inptr[1] == '$' \
&& ((inptr[2] == ')' \
&& (inptr[3] == 'A' \
|| (X12345 != '\0' && inptr[3] == X12345) \
|| inptr[3] == 'E' || inptr[3] == 'G')) \
|| (inptr[2] == '*' \
&& ((X7589 != '\0' && inptr[3] == X7589) \
|| (X13131 != '\0' && inptr[3] == X13131) \
|| inptr[3] == 'H')) \
|| (inptr[2] == '+' \
&& ((X7590 != '\0' && inptr[3] == X7590) \
|| (X13132 != '\0' && inptr[3] == X13132) \
|| inptr[3] == 'I' || inptr[3] == 'J' \
|| inptr[3] == 'K' || inptr[3] == 'L' \
|| inptr[3] == 'M')))) \
{ \
/* OK, we accept those character sets. */ \
if (inptr[3] == 'A') \
ann = (ann & ~SO_ann) | GB2312_ann; \
else if (inptr[3] == 'G') \
ann = (ann & ~SO_ann) | CNS11643_1_ann; \
else if (inptr[3] == 'E') \
ann = (ann & ~SO_ann) | ISO_IR_165_ann; \
else if (X12345 != '\0' && inptr[3] == X12345) \
ann = (ann & ~SO_ann) | GB12345_ann; \
else if (inptr[3] == 'H') \
ann = (ann & ~SS2_ann) | CNS11643_2_ann; \
else if (X7589 != '\0' && inptr[3] == X7589) \
ann = (ann & ~SS2_ann) | GB7589_ann; \
else if (X13131 != '\0' && inptr[3] == X13131) \
ann = (ann & ~SS2_ann) | GB13131_ann; \
else if (inptr[3] == 'I') \
ann = (ann & ~SS3_ann) | CNS11643_3_ann; \
else if (inptr[3] == 'J') \
ann = (ann & ~SS3_ann) | CNS11643_4_ann; \
else if (inptr[3] == 'K') \
ann = (ann & ~SS3_ann) | CNS11643_5_ann; \
else if (inptr[3] == 'L') \
ann = (ann & ~SS3_ann) | CNS11643_6_ann; \
else if (inptr[3] == 'M') \
ann = (ann & ~SS3_ann) | CNS11643_7_ann; \
else if (X7590 != '\0' && inptr[3] == X7590) \
ann = (ann & ~SS3_ann) | GB7590_ann; \
else if (X13132 != '\0' && inptr[3] == X13132) \
ann = (ann & ~SS3_ann) | GB13132_ann; \
inptr += 4; \
continue; \
} \
} \
else if (ch == SO) \
{ \
/* Switch to use GB2312, GB12345, CNS 11643 plane 1, or ISO-IR-165, \
depending on which S0 designation came last. The only problem \
is what to do with faulty input files where no designator came. \
XXX For now I'll default to use GB2312. If this is not the \
best behavior (e.g., we should flag an error) let me know. */ \
++inptr; \
if ((ann & SO_ann) != 0) \
switch (ann & SO_ann) \
{ \
case GB2312_ann: \
set = GB2312_set; \
break; \
case GB12345_ann: \
set = GB12345_set; \
break; \
case CNS11643_1_ann: \
set = CNS11643_1_set; \
break; \
case ISO_IR_165_ann: \
set = ISO_IR_165_set; \
break; \
default: \
abort (); \
} \
else \
{ \
STANDARD_FROM_LOOP_ERR_HANDLER (1); \
} \
continue; \
} \
else if (ch == SI) \
{ \
/* Switch to use ASCII. */ \
++inptr; \
set = ASCII_set; \
continue; \
} \
\
if (ch == ESC && inptr[1] == SS2_1) \
{ \
/* This is a character from CNS 11643 plane 2. \
XXX We could test here whether the use of this character \
set was announced. \
XXX Currently GB7589 and GB13131 are not supported. */ \
inptr += 2; \
ch = cns11643l2_to_ucs4 (&inptr, 2, 0); \
if (ch == __UNKNOWN_10646_CHAR) \
STANDARD_FROM_LOOP_ERR_HANDLER (2); \
} \
/* Note that we can assume here that at least 4 bytes are available if \
the first byte is ESC since otherwise the first if would have been \
true. */ \
else if (ch == ESC && inptr[1] == SS3_1) \
{ \
/* This is a character from CNS 11643 plane 3 or higher. \
XXX Currently GB7590 and GB13132 are not supported. */ \
unsigned char buf[3]; \
const unsigned char *tmp = buf; \
\
buf[1] = inptr[2]; \
buf[2] = inptr[3]; \
switch (ann & SS3_ann) \
{ \
case CNS11643_3_ann: \
buf[0] = 0x23; \
ch = cns11643_to_ucs4 (&tmp, 3, 0); \
break; \
case CNS11643_4_ann: \
buf[0] = 0x24; \
ch = cns11643_to_ucs4 (&tmp, 3, 0); \
break; \
case CNS11643_5_ann: \
buf[0] = 0x25; \
ch = cns11643_to_ucs4 (&tmp, 3, 0); \
break; \
case CNS11643_6_ann: \
buf[0] = 0x26; \
ch = cns11643_to_ucs4 (&tmp, 3, 0); \
break; \
case CNS11643_7_ann: \
buf[0] = 0x27; \
ch = cns11643_to_ucs4 (&tmp, 3, 0); \
break; \
default: \
/* XXX Currently GB7590 and GB13132 are not supported. */ \
ch = __UNKNOWN_10646_CHAR; \
break; \
} \
if (ch == __UNKNOWN_10646_CHAR) \
{ \
STANDARD_FROM_LOOP_ERR_HANDLER (4); \
} \
assert (tmp == buf + 3); \
inptr += 4; \
} \
else if (set == ASCII_set) \
{ \
/* Almost done, just advance the input pointer. */ \
++inptr; \
} \
else \
{ \
/* That's pretty easy, we have a dedicated functions for this. */ \
if (inend - inptr < 2) \
{ \
result = __GCONV_INCOMPLETE_INPUT; \
break; \
} \
if (set == GB2312_set) \
ch = gb2312_to_ucs4 (&inptr, inend - inptr, 0); \
else if (set == ISO_IR_165_set) \
ch = isoir165_to_ucs4 (&inptr, inend - inptr); \
else \
{ \
assert (set == CNS11643_1_set); \
ch = cns11643l1_to_ucs4 (&inptr, inend - inptr, 0); \
} \
\
if (ch == 0) \
{ \
result = __GCONV_INCOMPLETE_INPUT; \
break; \
} \
else if (ch == __UNKNOWN_10646_CHAR) \
{ \
STANDARD_FROM_LOOP_ERR_HANDLER (2); \
} \
} \
\
*((uint32_t *) outptr) = ch; \
outptr += sizeof (uint32_t); \
}
#define EXTRA_LOOP_DECLS , int *setp
#define INIT_PARAMS int set = (*setp >> 3) & CURRENT_MASK; \
int ann = (*setp >> 3) & ~CURRENT_MASK
#define UPDATE_PARAMS *setp = (set | ann) << 3
#define LOOP_NEED_FLAGS
#include <iconv/loop.c>
/* Next, define the other direction. */
#define MIN_NEEDED_INPUT TO_LOOP_MIN_NEEDED_FROM
#define MAX_NEEDED_INPUT TO_LOOP_MAX_NEEDED_FROM
#define MIN_NEEDED_OUTPUT TO_LOOP_MIN_NEEDED_TO
#define MAX_NEEDED_OUTPUT TO_LOOP_MAX_NEEDED_TO
#define LOOPFCT TO_LOOP
/* With GCC 5.3 when compiling with -Os the compiler emits a warning
that buf[0] and buf[1] may be used uninitialized. This can only
happen in the case where tmpbuf[3] is used, and in that case the
write to the tmpbuf[1] and tmpbuf[2] was assured because
ucs4_to_cns11643 would have filled in those entries. The difficulty
is in getting the compiler to see this logic because tmpbuf[0] is
involved in determining the code page and is the indicator that
tmpbuf[2] is initialized. */
DIAG_PUSH_NEEDS_COMMENT;
DIAG_IGNORE_Os_NEEDS_COMMENT (5, "-Wmaybe-uninitialized");
#define BODY \
{ \
uint32_t ch; \
size_t written = 0; \
\
ch = *((const uint32_t *) inptr); \
\
/* First see whether we can write the character using the currently \
selected character set. */ \
if (ch < 0x80) \
{ \
if (set != ASCII_set) \
{ \
*outptr++ = SI; \
set = ASCII_set; \
if (outptr == outend) \
{ \
result = __GCONV_FULL_OUTPUT; \
break; \
} \
} \
\
*outptr++ = ch; \
written = 1; \
\
/* At the end of the line we have to clear the `ann' flags since \
every line must contain this information again. */ \
if (ch == L'\n') \
ann = 0; \
} \
else \
{ \
unsigned char buf[2] = { 0, 0 }; \
int used; \
\
if (set == GB2312_set || ((ann & SO_ann) != CNS11643_1_ann \
&& (ann & SO_ann) != ISO_IR_165_ann)) \
{ \
written = ucs4_to_gb2312 (ch, buf, 2); \
used = GB2312_set; \
} \
else if (set == ISO_IR_165_set || (ann & SO_ann) == ISO_IR_165_set) \
{ \
written = ucs4_to_isoir165 (ch, buf, 2); \
used = ISO_IR_165_set; \
} \
else \
{ \
written = ucs4_to_cns11643l1 (ch, buf, 2); \
used = CNS11643_1_set; \
} \
\
if (written == __UNKNOWN_10646_CHAR) \
{ \
/* Cannot convert it using the currently selected SO set. \
Next try the SS2 set. */ \
written = ucs4_to_cns11643l2 (ch, buf, 2); \
if (written != __UNKNOWN_10646_CHAR) \
/* Yep, that worked. */ \
used = CNS11643_2_set; \
else \
{ \
unsigned char tmpbuf[3]; \
\
switch (0) \
{ \
default: \
/* Well, see whether we have to change the SO set. */ \
\
if (used != GB2312_set) \
{ \
written = ucs4_to_gb2312 (ch, buf, 2); \
if (written != __UNKNOWN_10646_CHAR) \
{ \
used = GB2312_set; \
break; \
} \
} \
\
if (used != ISO_IR_165_set) \
{ \
written = ucs4_to_isoir165 (ch, buf, 2); \
if (written != __UNKNOWN_10646_CHAR) \
{ \
used = ISO_IR_165_set; \
break; \
} \
} \
\
if (used != CNS11643_1_set) \
{ \
written = ucs4_to_cns11643l1 (ch, buf, 2); \
if (written != __UNKNOWN_10646_CHAR) \
{ \
used = CNS11643_1_set; \
break; \
} \
} \
\
written = ucs4_to_cns11643 (ch, tmpbuf, 3); \
if (written == 3 && tmpbuf[0] >= 3 && tmpbuf[0] <= 7) \
{ \
buf[0] = tmpbuf[1]; \
buf[1] = tmpbuf[2]; \
switch (tmpbuf[0]) \
{ \
case 3: \
used = CNS11643_3_set; \
break; \
case 4: \
used = CNS11643_4_set; \
break; \
case 5: \
used = CNS11643_5_set; \
break; \
case 6: \
used = CNS11643_6_set; \
break; \
case 7: \
used = CNS11643_7_set; \
break; \
default: \
abort (); \
} \
written = 2; \
break; \
} \
\
/* XXX Currently GB7590 and GB13132 are not supported. */\
\
/* Even this does not work. Error. */ \
used = ASCII_set; \
} \
if (used == ASCII_set) \
{ \
UNICODE_TAG_HANDLER (ch, 4); \
STANDARD_TO_LOOP_ERR_HANDLER (4); \
} \
} \
} \
assert (written == 2); \
\
/* See whether we have to emit an escape sequence. */ \
if (set != used) \
{ \
/* First see whether we announced that we use this \
character set. */ \
if ((used & SO_mask) != 0 && (ann & SO_ann) != (used << 8)) \
{ \
const char *escseq; \
\
if (outptr + 4 > outend) \
{ \
result = __GCONV_FULL_OUTPUT; \
break; \
} \
\
assert (used >= 1 && used <= 4); \
escseq = ")A\0\0)G)E" + (used - 1) * 2; \
*outptr++ = ESC; \
*outptr++ = '$'; \
*outptr++ = *escseq++; \
*outptr++ = *escseq++; \
\
ann = (ann & ~SO_ann) | (used << 8); \
} \
else if ((used & SS2_mask) != 0 && (ann & SS2_ann) != (used << 8))\
{ \
const char *escseq; \
\
if (outptr + 4 > outend) \
{ \
result = __GCONV_FULL_OUTPUT; \
break; \
} \
\
assert (used == CNS11643_2_set); /* XXX */ \
escseq = "*H"; \
*outptr++ = ESC; \
*outptr++ = '$'; \
*outptr++ = *escseq++; \
*outptr++ = *escseq++; \
\
ann = (ann & ~SS2_ann) | (used << 8); \
} \
else if ((used & SS3_mask) != 0 && (ann & SS3_ann) != (used << 8))\
{ \
const char *escseq; \
\
if (outptr + 4 > outend) \
{ \
result = __GCONV_FULL_OUTPUT; \
break; \
} \
\
assert ((used >> 5) >= 3 && (used >> 5) <= 7); \
escseq = "+I+J+K+L+M" + ((used >> 5) - 3) * 2; \
*outptr++ = ESC; \
*outptr++ = '$'; \
*outptr++ = *escseq++; \
*outptr++ = *escseq++; \
\
ann = (ann & ~SS3_ann) | (used << 8); \
} \
\
if (used == CNS11643_2_set) \
{ \
if (outptr + 2 > outend) \
{ \
result = __GCONV_FULL_OUTPUT; \
break; \
} \
*outptr++ = SS2_0; \
*outptr++ = SS2_1; \
} \
else if (used >= CNS11643_3_set && used <= CNS11643_7_set) \
{ \
if (outptr + 2 > outend) \
{ \
result = __GCONV_FULL_OUTPUT; \
break; \
} \
*outptr++ = SS3_0; \
*outptr++ = SS3_1; \
} \
else \
{ \
/* We only have to emit something if currently ASCII is \
selected. Otherwise we are switching within the \
SO charset. */ \
if (set == ASCII_set) \
{ \
if (outptr + 1 > outend) \
{ \
result = __GCONV_FULL_OUTPUT; \
break; \
} \
*outptr++ = SO; \
} \
} \
\
/* Always test the length here since we have used up all the \
guaranteed output buffer slots. */ \
if (outptr + 2 > outend) \
{ \
result = __GCONV_FULL_OUTPUT; \
break; \
} \
} \
else if (outptr + 2 > outend) \
{ \
result = __GCONV_FULL_OUTPUT; \
break; \
} \
\
*outptr++ = buf[0]; \
*outptr++ = buf[1]; \
set = used; \
} \
\
/* Now that we wrote the output increment the input pointer. */ \
inptr += 4; \
}
DIAG_POP_NEEDS_COMMENT;
#define EXTRA_LOOP_DECLS , int *setp
#define INIT_PARAMS int set = (*setp >> 3) & CURRENT_MASK; \
int ann = (*setp >> 3) & ~CURRENT_MASK
#define REINIT_PARAMS do \
{ \
set = (*setp >> 3) & CURRENT_MASK; \
ann = (*setp >> 3) & ~CURRENT_MASK; \
} \
while (0)
#define UPDATE_PARAMS *setp = (set | ann) << 3
#define LOOP_NEED_FLAGS
#include <iconv/loop.c>
/* Now define the toplevel functions. */
#include <iconv/skeleton.c>