[thirdparty/glibc.git] / iconvdata / utf-16.c

/* Conversion module for UTF-16.
   Copyright (C) 1999-2019 Free Software Foundation, Inc.
   This file is part of the GNU C Library.
   Contributed by Ulrich Drepper <drepper@cygnus.com>, 1999.

   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 <byteswap.h>
#include <dlfcn.h>
#include <gconv.h>
#include <stddef.h>
#include <stdint.h>
#include <stdlib.h>
#include <string.h>

/* This is the Byte Order Mark character (BOM).  */
#define BOM	0xfeff
/* And in the other byte order.  */
#define BOM_OE	0xfffe


/* Definitions used in the body of the `gconv' function.  */
#define FROM_LOOP		from_utf16_loop
#define TO_LOOP			to_utf16_loop
#define DEFINE_INIT		0
#define DEFINE_FINI		0
#define MIN_NEEDED_FROM		2
#define MAX_NEEDED_FROM		4
#define MIN_NEEDED_TO		4
#define ONE_DIRECTION		0
#define FROM_DIRECTION		(dir == from_utf16)
#define PREPARE_LOOP \
  enum direction dir = ((struct utf16_data *) step->__data)->dir;	      \
  enum variant var = ((struct utf16_data *) step->__data)->var;		      \
  if (__glibc_unlikely (data->__invocation_counter == 0))		      \
    {									      \
      if (var == UTF_16)						      \
	{								      \
	  if (FROM_DIRECTION)						      \
	    {								      \
	      /* We have to find out which byte order the file is	      \
		 encoded in.  */					      \
	      if (inptr + 2 > inend)					      \
		return (inptr == inend					      \
			? __GCONV_EMPTY_INPUT : __GCONV_INCOMPLETE_INPUT);    \
									      \
	      if (get16u (inptr) == BOM)				      \
		/* Simply ignore the BOM character.  */			      \
		*inptrp = inptr += 2;					      \
	      else if (get16u (inptr) == BOM_OE)			      \
		{							      \
		  data->__flags |= __GCONV_SWAP;			      \
		  *inptrp = inptr += 2;					      \
		}							      \
	    }								      \
	  else if (!FROM_DIRECTION && !data->__internal_use)		      \
	    {								      \
	      /* Emit the Byte Order Mark.  */				      \
	      if (__glibc_unlikely (outbuf + 2 > outend))		      \
		return __GCONV_FULL_OUTPUT;				      \
									      \
	      put16u (outbuf, BOM);					      \
	      outbuf += 2;						      \
	    }								      \
	}								      \
      else if ((var == UTF_16LE && BYTE_ORDER == BIG_ENDIAN)		      \
	       || (var == UTF_16BE && BYTE_ORDER == LITTLE_ENDIAN))	      \
	data->__flags |= __GCONV_SWAP;					      \
    }									      \
  const int swap = data->__flags & __GCONV_SWAP;
#define EXTRA_LOOP_ARGS		, swap


/* Direction of the transformation.  */
enum direction
{
  illegal_dir,
  to_utf16,
  from_utf16
};

enum variant
{
  illegal_var,
  UTF_16,
  UTF_16LE,
  UTF_16BE
};

struct utf16_data
{
  enum direction dir;
  enum variant var;
};


extern int gconv_init (struct __gconv_step *step);
int
gconv_init (struct __gconv_step *step)
{
  /* Determine which direction.  */
  struct utf16_data *new_data;
  enum direction dir = illegal_dir;
  enum variant var = illegal_var;
  int result;

  if (__strcasecmp (step->__from_name, "UTF-16//") == 0)
    {
      dir = from_utf16;
      var = UTF_16;
    }
  else if (__strcasecmp (step->__to_name, "UTF-16//") == 0)
    {
      dir = to_utf16;
      var = UTF_16;
    }
  else if (__strcasecmp (step->__from_name, "UTF-16BE//") == 0)
    {
      dir = from_utf16;
      var = UTF_16BE;
    }
  else if (__strcasecmp (step->__to_name, "UTF-16BE//") == 0)
    {
      dir = to_utf16;
      var = UTF_16BE;
    }
  else if (__strcasecmp (step->__from_name, "UTF-16LE//") == 0)
    {
      dir = from_utf16;
      var = UTF_16LE;
    }
  else if (__strcasecmp (step->__to_name, "UTF-16LE//") == 0)
    {
      dir = to_utf16;
      var = UTF_16LE;
    }

  result = __GCONV_NOCONV;
  if (__builtin_expect (dir, to_utf16) != illegal_dir)
    {
      new_data = (struct utf16_data *) malloc (sizeof (struct utf16_data));

      result = __GCONV_NOMEM;
      if (new_data != NULL)
	{
	  new_data->dir = dir;
	  new_data->var = var;
	  step->__data = new_data;

	  if (dir == from_utf16)
	    {
	      step->__min_needed_from = MIN_NEEDED_FROM;
	      step->__max_needed_from = MAX_NEEDED_FROM;
	      step->__min_needed_to = MIN_NEEDED_TO;
	      step->__max_needed_to = MIN_NEEDED_TO;
	    }
	  else
	    {
	      step->__min_needed_from = MIN_NEEDED_TO;
	      step->__max_needed_from = MIN_NEEDED_TO;
	      step->__min_needed_to = MIN_NEEDED_FROM;
	      step->__max_needed_to = MAX_NEEDED_FROM;
	    }

	  step->__stateful = 0;

	  result = __GCONV_OK;
	}
    }

  return result;
}


extern void gconv_end (struct __gconv_step *data);
void
gconv_end (struct __gconv_step *data)
{
  free (data->__data);
}


/* Convert from the internal (UCS4-like) format to UTF-16.  */
#define MIN_NEEDED_INPUT	MIN_NEEDED_TO
#define MIN_NEEDED_OUTPUT	MIN_NEEDED_FROM
#define MAX_NEEDED_OUTPUT	MAX_NEEDED_FROM
#define LOOPFCT			TO_LOOP
#define BODY \
  {									      \
    uint32_t c = get32 (inptr);						      \
									      \
    if (__glibc_unlikely (c >= 0xd800 && c < 0xe000))			      \
      {									      \
	/* Surrogate characters in UCS-4 input are not valid.		      \
	   We must catch this.  If we let surrogates pass through,	      \
	   attackers could make a security hole exploit by		      \
	   synthesizing any desired plane 1-16 character.  */		      \
	result = __GCONV_ILLEGAL_INPUT;					      \
	if (! ignore_errors_p ())					      \
	  break;							      \
	inptr += 4;							      \
	++*irreversible;						      \
	continue;							      \
      }									      \
									      \
    if (swap)								      \
      {									      \
	if (__glibc_unlikely (c >= 0x10000))				      \
	  {								      \
	    if (__glibc_unlikely (c >= 0x110000))			      \
	      {								      \
		STANDARD_TO_LOOP_ERR_HANDLER (4);			      \
	      }								      \
									      \
	    /* Generate a surrogate character.  */			      \
	    if (__glibc_unlikely (outptr + 4 > outend))			      \
	      {								      \
		/* Overflow in the output buffer.  */			      \
		result = __GCONV_FULL_OUTPUT;				      \
		break;							      \
	      }								      \
									      \
	    put16 (outptr, bswap_16 (0xd7c0 + (c >> 10)));		      \
	    outptr += 2;						      \
	    put16 (outptr, bswap_16 (0xdc00 + (c & 0x3ff)));		      \
	  }								      \
	else								      \
	  put16 (outptr, bswap_16 (c));					      \
      }									      \
    else								      \
      {									      \
	if (__glibc_unlikely (c >= 0x10000))				      \
	  {								      \
	    if (__glibc_unlikely (c >= 0x110000))			      \
	      {								      \
		STANDARD_TO_LOOP_ERR_HANDLER (4);			      \
	      }								      \
									      \
	    /* Generate a surrogate character.  */			      \
	    if (__glibc_unlikely (outptr + 4 > outend))			      \
	      {								      \
		/* Overflow in the output buffer.  */			      \
		result = __GCONV_FULL_OUTPUT;				      \
		break;							      \
	      }								      \
									      \
	    put16 (outptr, 0xd7c0 + (c >> 10));				      \
	    outptr += 2;						      \
	    put16 (outptr, 0xdc00 + (c & 0x3ff));			      \
	  }								      \
	else								      \
	  put16 (outptr, c);						      \
      }									      \
    outptr += 2;							      \
    inptr += 4;								      \
  }
#define LOOP_NEED_FLAGS
#define EXTRA_LOOP_DECLS \
	, int swap
#include <iconv/loop.c>


/* Convert from UTF-16 to the internal (UCS4-like) format.  */
#define MIN_NEEDED_INPUT	MIN_NEEDED_FROM
#define MAX_NEEDED_INPUT	MAX_NEEDED_FROM
#define MIN_NEEDED_OUTPUT	MIN_NEEDED_TO
#define LOOPFCT			FROM_LOOP
#define BODY \
  {									      \
    uint16_t u1 = get16 (inptr);					      \
									      \
    if (swap)								      \
      {									      \
	u1 = bswap_16 (u1);						      \
									      \
	if (__builtin_expect (u1 < 0xd800, 1) || u1 > 0xdfff)		      \
	  {								      \
	    /* No surrogate.  */					      \
	    put32 (outptr, u1);						      \
	    inptr += 2;							      \
	  }								      \
	else								      \
	  {								      \
	    uint16_t u2;						      \
									      \
	    if (__glibc_unlikely (u1 >= 0xdc00))			      \
	      {								      \
		/* This is no valid first word for a surrogate.  */	      \
		STANDARD_FROM_LOOP_ERR_HANDLER (2);			      \
	      }								      \
									      \
	    /* It's a surrogate character.  At least the first word says      \
	       it is.  */						      \
	    if (__glibc_unlikely (inptr + 4 > inend))			      \
	      {								      \
		/* We don't have enough input for another complete input      \
		   character.  */					      \
		result = __GCONV_INCOMPLETE_INPUT;			      \
		break;							      \
	      }								      \
									      \
	    inptr += 2;							      \
	    u2 = bswap_16 (get16 (inptr));				      \
	    if (__builtin_expect (u2 < 0xdc00, 0)			      \
		|| __builtin_expect (u2 > 0xdfff, 0))			      \
	      {								      \
		/* This is no valid second word for a surrogate.  */	      \
		inptr -= 2;						      \
		STANDARD_FROM_LOOP_ERR_HANDLER (2);			      \
	      }								      \
									      \
	    put32 (outptr, ((u1 - 0xd7c0) << 10) + (u2 - 0xdc00));	      \
	    inptr += 2;							      \
	  }								      \
      }									      \
    else								      \
      {									      \
	if (__builtin_expect (u1 < 0xd800, 1) || u1 > 0xdfff)		      \
	  {								      \
	    /* No surrogate.  */					      \
	    put32 (outptr, u1);						      \
	    inptr += 2;							      \
	  }								      \
	else								      \
	  {								      \
	    if (__glibc_unlikely (u1 >= 0xdc00))			      \
	      {								      \
		/* This is no valid first word for a surrogate.  */	      \
		STANDARD_FROM_LOOP_ERR_HANDLER (2);			      \
	      }								      \
									      \
	    /* It's a surrogate character.  At least the first word says      \
	       it is.  */						      \
	    if (__glibc_unlikely (inptr + 4 > inend))			      \
	      {								      \
		/* We don't have enough input for another complete input      \
		   character.  */					      \
		result = __GCONV_INCOMPLETE_INPUT;			      \
		break;							      \
	      }								      \
									      \
	    inptr += 2;							      \
	    uint16_t u2 = get16 (inptr);				      \
	    if (__builtin_expect (u2 < 0xdc00, 0)			      \
		|| __builtin_expect (u2 > 0xdfff, 0))			      \
	      {								      \
		/* This is no valid second word for a surrogate.  */	      \
		inptr -= 2;						      \
		STANDARD_FROM_LOOP_ERR_HANDLER (2);			      \
	      }								      \
									      \
	    put32 (outptr, ((u1 - 0xd7c0) << 10) + (u2 - 0xdc00));	      \
	    inptr += 2;							      \
	  }								      \
      }									      \
    outptr += 4;							      \
  }
#define LOOP_NEED_FLAGS
#define EXTRA_LOOP_DECLS \
	, int swap
#include <iconv/loop.c>


/* Now define the toplevel functions.  */
#include <iconv/skeleton.c>
Commit	Line	Data
7cdd956e	1	/* Conversion module for UTF-16.
04277e02	2	Copyright (C) 1999-2019 Free Software Foundation, Inc.
7cdd956e UD	3	This file is part of the GNU C Library.
	4	Contributed by Ulrich Drepper <drepper@cygnus.com>, 1999.
	5
	6	The GNU C Library is free software; you can redistribute it and/or
41bdb6e2 AJ	7	modify it under the terms of the GNU Lesser General Public
	8	License as published by the Free Software Foundation; either
	9	version 2.1 of the License, or (at your option) any later version.
7cdd956e UD	10
	11	The GNU C Library is distributed in the hope that it will be useful,
	12	but WITHOUT ANY WARRANTY; without even the implied warranty of
	13	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
41bdb6e2	14	Lesser General Public License for more details.
7cdd956e	15
41bdb6e2	16	You should have received a copy of the GNU Lesser General Public
59ba27a6	17	License along with the GNU C Library; if not, see
5a82c748	18	<https://www.gnu.org/licenses/>. */
7cdd956e UD	19
7cdd956e UD	20	#include <byteswap.h>
55985355	21	#include <dlfcn.h>
7cdd956e UD	22	#include <gconv.h>
	23	#include <stddef.h>
	24	#include <stdint.h>
	25	#include <stdlib.h>
	26	#include <string.h>
	27
	28	/* This is the Byte Order Mark character (BOM). */
	29	#define BOM 0xfeff
8d617a71 UD	30	/* And in the other byte order. */
8d617a71 UD	31	#define BOM_OE 0xfffe
7cdd956e UD	32
	33
	34	/* Definitions used in the body of the `gconv' function. */
	35	#define FROM_LOOP from_utf16_loop
	36	#define TO_LOOP to_utf16_loop
	37	#define DEFINE_INIT 0
	38	#define DEFINE_FINI 0
	39	#define MIN_NEEDED_FROM 2
	40	#define MAX_NEEDED_FROM 4
	41	#define MIN_NEEDED_TO 4
13e402e7	42	#define ONE_DIRECTION 0
7cdd956e UD	43	#define FROM_DIRECTION (dir == from_utf16)
	44	#define PREPARE_LOOP \
	45	enum direction dir = ((struct utf16_data *) step->__data)->dir; \
	46	enum variant var = ((struct utf16_data *) step->__data)->var; \
a1ffb40e	47	if (__glibc_unlikely (data->__invocation_counter == 0)) \
8d617a71	48	{ \
ee190f67	49	if (var == UTF_16) \
8d617a71	50	{ \
ee190f67	51	if (FROM_DIRECTION) \
8d617a71	52	{ \
ee190f67 UD	53	/* We have to find out which byte order the file is \
	54	encoded in. */ \
	55	if (inptr + 2 > inend) \
	56	return (inptr == inend \
	57	? __GCONV_EMPTY_INPUT : __GCONV_INCOMPLETE_INPUT); \
	58	\
	59	if (get16u (inptr) == BOM) \
	60	/* Simply ignore the BOM character. */ \
	61	*inptrp = inptr += 2; \
	62	else if (get16u (inptr) == BOM_OE) \
	63	{ \
	64	data->__flags \|= __GCONV_SWAP; \
	65	*inptrp = inptr += 2; \
	66	} \
8d617a71	67	} \
ee190f67 UD	68	else if (!FROM_DIRECTION && !data->__internal_use) \
	69	{ \
	70	/* Emit the Byte Order Mark. */ \
a1ffb40e	71	if (__glibc_unlikely (outbuf + 2 > outend)) \
ee190f67	72	return __GCONV_FULL_OUTPUT; \
7cdd956e	73	\
ee190f67 UD	74	put16u (outbuf, BOM); \
	75	outbuf += 2; \
	76	} \
dc4bb1c2	77	} \
ee190f67 UD	78	else if ((var == UTF_16LE && BYTE_ORDER == BIG_ENDIAN) \
	79	\|\| (var == UTF_16BE && BYTE_ORDER == LITTLE_ENDIAN)) \
	80	data->__flags \|= __GCONV_SWAP; \
b721a2c0	81	} \
ee190f67	82	const int swap = data->__flags & __GCONV_SWAP;
dc4bb1c2	83	#define EXTRA_LOOP_ARGS , swap
7cdd956e UD	84
	85
	86	/* Direction of the transformation. */
	87	enum direction
	88	{
	89	illegal_dir,
	90	to_utf16,
	91	from_utf16
	92	};
	93
	94	enum variant
	95	{
	96	illegal_var,
	97	UTF_16,
	98	UTF_16LE,
	99	UTF_16BE
	100	};
	101
	102	struct utf16_data
	103	{
	104	enum direction dir;
	105	enum variant var;
	106	};
	107
	108
8c0b7170	109	extern int gconv_init (struct __gconv_step *step);
7cdd956e UD	110	int
	111	gconv_init (struct __gconv_step *step)
	112	{
	113	/* Determine which direction. */
	114	struct utf16_data *new_data;
	115	enum direction dir = illegal_dir;
	116	enum variant var = illegal_var;
	117	int result;
	118
755104ed	119	if (__strcasecmp (step->__from_name, "UTF-16//") == 0)
7cdd956e UD	120	{
	121	dir = from_utf16;
	122	var = UTF_16;
	123	}
755104ed	124	else if (__strcasecmp (step->__to_name, "UTF-16//") == 0)
7cdd956e UD	125	{
	126	dir = to_utf16;
	127	var = UTF_16;
	128	}
755104ed	129	else if (__strcasecmp (step->__from_name, "UTF-16BE//") == 0)
7cdd956e UD	130	{
	131	dir = from_utf16;
	132	var = UTF_16BE;
	133	}
755104ed	134	else if (__strcasecmp (step->__to_name, "UTF-16BE//") == 0)
7cdd956e UD	135	{
	136	dir = to_utf16;
	137	var = UTF_16BE;
	138	}
755104ed	139	else if (__strcasecmp (step->__from_name, "UTF-16LE//") == 0)
7cdd956e UD	140	{
	141	dir = from_utf16;
	142	var = UTF_16LE;
	143	}
755104ed	144	else if (__strcasecmp (step->__to_name, "UTF-16LE//") == 0)
7cdd956e UD	145	{
	146	dir = to_utf16;
	147	var = UTF_16LE;
	148	}
	149
	150	result = __GCONV_NOCONV;
ee4ce289	151	if (__builtin_expect (dir, to_utf16) != illegal_dir)
7cdd956e UD	152	{
	153	new_data = (struct utf16_data *) malloc (sizeof (struct utf16_data));
	154
	155	result = __GCONV_NOMEM;
	156	if (new_data != NULL)
	157	{
	158	new_data->dir = dir;
	159	new_data->var = var;
	160	step->__data = new_data;
	161
bc4831b9	162	if (dir == from_utf16)
7cdd956e UD	163	{
7cdd956e UD	164	step->__min_needed_from = MIN_NEEDED_FROM;
b721a2c0	165	step->__max_needed_from = MAX_NEEDED_FROM;
7cdd956e UD	166	step->__min_needed_to = MIN_NEEDED_TO;
	167	step->__max_needed_to = MIN_NEEDED_TO;
	168	}
	169	else
	170	{
	171	step->__min_needed_from = MIN_NEEDED_TO;
	172	step->__max_needed_from = MIN_NEEDED_TO;
	173	step->__min_needed_to = MIN_NEEDED_FROM;
b721a2c0	174	step->__max_needed_to = MAX_NEEDED_FROM;
7cdd956e UD	175	}
	176
	177	step->__stateful = 0;
	178
	179	result = __GCONV_OK;
	180	}
	181	}
	182
	183	return result;
	184	}
	185
	186
8c0b7170	187	extern void gconv_end (struct __gconv_step *data);
7cdd956e UD	188	void
	189	gconv_end (struct __gconv_step *data)
	190	{
	191	free (data->__data);
	192	}
	193
	194
	195	/* Convert from the internal (UCS4-like) format to UTF-16. */
	196	#define MIN_NEEDED_INPUT MIN_NEEDED_TO
	197	#define MIN_NEEDED_OUTPUT MIN_NEEDED_FROM
	198	#define MAX_NEEDED_OUTPUT MAX_NEEDED_FROM
	199	#define LOOPFCT TO_LOOP
	200	#define BODY \
	201	{ \
77e1d15a	202	uint32_t c = get32 (inptr); \
7cdd956e	203	\
a1ffb40e	204	if (__glibc_unlikely (c >= 0xd800 && c < 0xe000)) \
755104ed UD	205	{ \
	206	/* Surrogate characters in UCS-4 input are not valid. \
	207	We must catch this. If we let surrogates pass through, \
	208	attackers could make a security hole exploit by \
	209	synthesizing any desired plane 1-16 character. */ \
e438a468	210	result = __GCONV_ILLEGAL_INPUT; \
755104ed	211	if (! ignore_errors_p ()) \
e438a468	212	break; \
755104ed UD	213	inptr += 4; \
	214	++*irreversible; \
	215	continue; \
	216	} \
	217	\
8d617a71	218	if (swap) \
7cdd956e	219	{ \
a1ffb40e	220	if (__glibc_unlikely (c >= 0x10000)) \
7cdd956e	221	{ \
a1ffb40e	222	if (__glibc_unlikely (c >= 0x110000)) \
7cdd956e	223	{ \
e438a468	224	STANDARD_TO_LOOP_ERR_HANDLER (4); \
7cdd956e UD	225	} \
	226	\
	227	/* Generate a surrogate character. */ \
a1ffb40e	228	if (__glibc_unlikely (outptr + 4 > outend)) \
7cdd956e UD	229	{ \
	230	/* Overflow in the output buffer. */ \
	231	result = __GCONV_FULL_OUTPUT; \
	232	break; \
	233	} \
	234	\
77e1d15a	235	put16 (outptr, bswap_16 (0xd7c0 + (c >> 10))); \
7cdd956e	236	outptr += 2; \
77e1d15a	237	put16 (outptr, bswap_16 (0xdc00 + (c & 0x3ff))); \
7cdd956e UD	238	} \
7cdd956e UD	239	else \
77e1d15a	240	put16 (outptr, bswap_16 (c)); \
7cdd956e UD	241	} \
	242	else \
	243	{ \
a1ffb40e	244	if (__glibc_unlikely (c >= 0x10000)) \
7cdd956e	245	{ \
a1ffb40e	246	if (__glibc_unlikely (c >= 0x110000)) \
7cdd956e	247	{ \
e438a468	248	STANDARD_TO_LOOP_ERR_HANDLER (4); \
7cdd956e UD	249	} \
	250	\
	251	/* Generate a surrogate character. */ \
a1ffb40e	252	if (__glibc_unlikely (outptr + 4 > outend)) \
7cdd956e UD	253	{ \
	254	/* Overflow in the output buffer. */ \
	255	result = __GCONV_FULL_OUTPUT; \
	256	break; \
	257	} \
	258	\
77e1d15a	259	put16 (outptr, 0xd7c0 + (c >> 10)); \
7cdd956e	260	outptr += 2; \
77e1d15a	261	put16 (outptr, 0xdc00 + (c & 0x3ff)); \
7cdd956e UD	262	} \
7cdd956e UD	263	else \
77e1d15a	264	put16 (outptr, c); \
7cdd956e UD	265	} \
	266	outptr += 2; \
	267	inptr += 4; \
	268	}
55985355	269	#define LOOP_NEED_FLAGS
7cdd956e	270	#define EXTRA_LOOP_DECLS \
dc4bb1c2	271	, int swap
7cdd956e UD	272	#include <iconv/loop.c>
	273
	274
	275	/* Convert from UTF-16 to the internal (UCS4-like) format. */
	276	#define MIN_NEEDED_INPUT MIN_NEEDED_FROM
	277	#define MAX_NEEDED_INPUT MAX_NEEDED_FROM
	278	#define MIN_NEEDED_OUTPUT MIN_NEEDED_TO
	279	#define LOOPFCT FROM_LOOP
	280	#define BODY \
	281	{ \
77e1d15a	282	uint16_t u1 = get16 (inptr); \
7cdd956e	283	\
8d617a71	284	if (swap) \
7cdd956e UD	285	{ \
	286	u1 = bswap_16 (u1); \
	287	\
db2d05f9	288	if (__builtin_expect (u1 < 0xd800, 1) \|\| u1 > 0xdfff) \
7cdd956e UD	289	{ \
7cdd956e UD	290	/* No surrogate. */ \
ee4ce289	291	put32 (outptr, u1); \
7cdd956e UD	292	inptr += 2; \
	293	} \
	294	else \
	295	{ \
	296	uint16_t u2; \
	297	\
7ab1de21 SL	298	if (__glibc_unlikely (u1 >= 0xdc00)) \
	299	{ \
	300	/* This is no valid first word for a surrogate. */ \
	301	STANDARD_FROM_LOOP_ERR_HANDLER (2); \
	302	} \
	303	\
7cdd956e UD	304	/* It's a surrogate character. At least the first word says \
7cdd956e UD	305	it is. */ \
a1ffb40e	306	if (__glibc_unlikely (inptr + 4 > inend)) \
7cdd956e UD	307	{ \
	308	/* We don't have enough input for another complete input \
	309	character. */ \
	310	result = __GCONV_INCOMPLETE_INPUT; \
	311	break; \
	312	} \
	313	\
77e1d15a UD	314	inptr += 2; \
77e1d15a UD	315	u2 = bswap_16 (get16 (inptr)); \
db2d05f9	316	if (__builtin_expect (u2 < 0xdc00, 0) \
5be072b6	317	\|\| __builtin_expect (u2 > 0xdfff, 0)) \
7cdd956e UD	318	{ \
7cdd956e UD	319	/* This is no valid second word for a surrogate. */ \
e438a468 UD	320	inptr -= 2; \
e438a468 UD	321	STANDARD_FROM_LOOP_ERR_HANDLER (2); \
7cdd956e UD	322	} \
7cdd956e UD	323	\
77e1d15a UD	324	put32 (outptr, ((u1 - 0xd7c0) << 10) + (u2 - 0xdc00)); \
77e1d15a UD	325	inptr += 2; \
7cdd956e UD	326	} \
	327	} \
	328	else \
	329	{ \
db2d05f9	330	if (__builtin_expect (u1 < 0xd800, 1) \|\| u1 > 0xdfff) \
7cdd956e UD	331	{ \
7cdd956e UD	332	/* No surrogate. */ \
77e1d15a	333	put32 (outptr, u1); \
7cdd956e UD	334	inptr += 2; \
	335	} \
	336	else \
	337	{ \
7ab1de21 SL	338	if (__glibc_unlikely (u1 >= 0xdc00)) \
	339	{ \
	340	/* This is no valid first word for a surrogate. */ \
	341	STANDARD_FROM_LOOP_ERR_HANDLER (2); \
	342	} \
	343	\
7cdd956e UD	344	/* It's a surrogate character. At least the first word says \
7cdd956e UD	345	it is. */ \
a1ffb40e	346	if (__glibc_unlikely (inptr + 4 > inend)) \
7cdd956e UD	347	{ \
	348	/* We don't have enough input for another complete input \
	349	character. */ \
	350	result = __GCONV_INCOMPLETE_INPUT; \
	351	break; \
	352	} \
	353	\
77e1d15a	354	inptr += 2; \
dc4bb1c2	355	uint16_t u2 = get16 (inptr); \
db2d05f9	356	if (__builtin_expect (u2 < 0xdc00, 0) \
5be072b6	357	\|\| __builtin_expect (u2 > 0xdfff, 0)) \
7cdd956e UD	358	{ \
7cdd956e UD	359	/* This is no valid second word for a surrogate. */ \
e438a468 UD	360	inptr -= 2; \
e438a468 UD	361	STANDARD_FROM_LOOP_ERR_HANDLER (2); \
7cdd956e UD	362	} \
7cdd956e UD	363	\
77e1d15a UD	364	put32 (outptr, ((u1 - 0xd7c0) << 10) + (u2 - 0xdc00)); \
77e1d15a UD	365	inptr += 2; \
7cdd956e UD	366	} \
	367	} \
	368	outptr += 4; \
	369	}
55985355	370	#define LOOP_NEED_FLAGS
7cdd956e	371	#define EXTRA_LOOP_DECLS \
dc4bb1c2	372	, int swap
7cdd956e UD	373	#include <iconv/loop.c>
	374
	375
	376	/* Now define the toplevel functions. */
	377	#include <iconv/skeleton.c>