gcm128.c

/* ====================================================================
 * Copyright (c) 2010 The OpenSSL Project.  All rights reserved.
 *
 * 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, 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.
 *
 * 3. All advertising materials mentioning features or use of this
 *    software must display the following acknowledgment:
 *    "This product includes software developed by the OpenSSL Project
 *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
 *
 * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
 *    endorse or promote products derived from this software without
 *    prior written permission. For written permission, please contact
 *    openssl-core@openssl.org.
 *
 * 5. Products derived from this software may not be called "OpenSSL"
 *    nor may "OpenSSL" appear in their names without prior written
 *    permission of the OpenSSL Project.
 *
 * 6. Redistributions of any form whatsoever must retain the following
 *    acknowledgment:
 *    "This product includes software developed by the OpenSSL Project
 *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
 *
 * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
 * EXPRESSED 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 OpenSSL PROJECT OR
 * ITS 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.
 * ====================================================================
 */

#include "modes.h"
#include <string.h>

#ifndef MODES_DEBUG
# ifndef NDEBUG
#  define NDEBUG
# endif
#endif
#include <assert.h>

#if (defined(_WIN32) || defined(_WIN64)) && !defined(__MINGW32__)
typedef __int64 i64;
typedef unsigned __int64 u64;
#define U64(C) C##UI64
#elif defined(__arch64__)
typedef long i64;
typedef unsigned long u64;
#define U64(C) C##UL
#else
typedef long long i64;
typedef unsigned long long u64;
#define U64(C) C##ULL
#endif

typedef unsigned int u32;
typedef unsigned char u8;
typedef struct { u64 hi,lo; } u128;

#define STRICT_ALIGNMENT
#if defined(__i386)	|| defined(__i386__)	|| \
    defined(__x86_64)	|| defined(__x86_64__)	|| \
    defined(_M_IX86)	|| defined(_M_AMD64)	|| defined(_M_X64) || \
    defined(__s390__)	|| defined(__s390x__)
# undef STRICT_ALIGNMENT
#endif

#if defined(__GNUC__) && __GNUC__>=2
# if defined(__x86_64) || defined(__x86_64__)
#  define BSWAP8(x) ({	u64 ret=(x);			\
			asm volatile ("bswapq %0"	\
			: "+r"(ret));	ret;		})
#  define BSWAP4(x) ({	u32 ret=(x);			\
			asm volatile ("bswapl %0"	\
			: "+r"(ret));	ret;		})
# elif defined(__i386) || defined(__i386__)
#  define BSWAP8(x) ({	u32 lo=(u64)(x)>>32,hi=(x);	\
			asm volatile ("bswapl %0; bswapl %1"	\
			: "+r"(hi),"+r"(lo));		\
			(u64)hi<<32|lo;			})
#  define BSWAP4(x) ({	u32 ret=(x);			\
			asm volatile ("bswapl %0"	\
			: "+r"(ret));	ret;		})
# endif
#elif defined(_MSC_VER)
# if _MSC_VER>=1300
#  pragma intrinsic(_byteswap_uint64,_byteswap_ulong)
#  define BSWAP8(x)	_byteswap_uint64((u64)(x))
#  define BSWAP4(x)	_byteswap_ulong((u32)(x))
# elif defined(_M_IX86)
# endif
#endif

#ifdef BSWAP4
#define GETU32(p)	BSWAP4(*(const u32 *)(p))
#define PUTU32(p,v)	*(u32 *)(p) = BSWAP4(v)
#else
#define GETU32(p)	((u32)(p)[0]<<24|(u32)(p)[1]<<16|(u32)(p)[2]<<8|(u32)(p)[3])
#define PUTU32(p,v)	((p)[0]=(u8)((v)>>24),(p)[1]=(u8)((v)>>16),(p)[2]=(u8)((v)>>8),(p)[3]=(u8)(v))
#endif

#define PACK(s) ((size_t)(s)<<(sizeof(size_t)*8-16))

#if 0
static void gcm_init_8bit(u128 Htable[256], u64 H[2])
{
	int  i, j;
	u128 V;

	Htable[0].hi = 0;
	Htable[0].lo = 0;
	V.hi = H[0];
	V.lo = H[1];

	for (Htable[128]=V, i=64; i>0; i>>=1) {
		if (sizeof(size_t)==8) {
			u64 T = U64(0xe100000000000000) & (0-(V.lo&1));
			V.lo  = (V.hi<<63)|(V.lo>>1);
			V.hi  = (V.hi>>1 )^T;
		}
		else {
			u32 T = 0xe1000000U & (0-(u32)(V.lo&1));
			V.lo  = (V.hi<<63)|(V.lo>>1);
			V.hi  = (V.hi>>1) ^((u64)T<<32);
		}
		Htable[i] = V;
	}

	for (i=2; i<256; i<<=1) {
		u128 *Hi = Htable+i, H0 = *Hi;
		for (j=1; j<i; ++j) {
			Hi[j].hi = H0.hi^Htable[j].hi;
			Hi[j].lo = H0.lo^Htable[j].lo;
		}
	}
}

static void gcm_mul_8bit(u64 Xi[2], u128 Htable[256])
{
	u128 Z = { 0, 0};
	const u8 *xi = (const u8 *)Xi+15;
	size_t rem, n = *xi;
	const union { long one; char little; } is_endian = {1};
	static const size_t rem_8bit[256] = {
		PACK(0x0000), PACK(0x01C2), PACK(0x0384), PACK(0x0246),
		PACK(0x0708), PACK(0x06CA), PACK(0x048C), PACK(0x054E),
		PACK(0x0E10), PACK(0x0FD2), PACK(0x0D94), PACK(0x0C56),
		PACK(0x0918), PACK(0x08DA), PACK(0x0A9C), PACK(0x0B5E),
		PACK(0x1C20), PACK(0x1DE2), PACK(0x1FA4), PACK(0x1E66),
		PACK(0x1B28), PACK(0x1AEA), PACK(0x18AC), PACK(0x196E),
		PACK(0x1230), PACK(0x13F2), PACK(0x11B4), PACK(0x1076),
		PACK(0x1538), PACK(0x14FA), PACK(0x16BC), PACK(0x177E),
		PACK(0x3840), PACK(0x3982), PACK(0x3BC4), PACK(0x3A06),
		PACK(0x3F48), PACK(0x3E8A), PACK(0x3CCC), PACK(0x3D0E),
		PACK(0x3650), PACK(0x3792), PACK(0x35D4), PACK(0x3416),
		PACK(0x3158), PACK(0x309A), PACK(0x32DC), PACK(0x331E),
		PACK(0x2460), PACK(0x25A2), PACK(0x27E4), PACK(0x2626),
		PACK(0x2368), PACK(0x22AA), PACK(0x20EC), PACK(0x212E),
		PACK(0x2A70), PACK(0x2BB2), PACK(0x29F4), PACK(0x2836),
		PACK(0x2D78), PACK(0x2CBA), PACK(0x2EFC), PACK(0x2F3E),
		PACK(0x7080), PACK(0x7142), PACK(0x7304), PACK(0x72C6),
		PACK(0x7788), PACK(0x764A), PACK(0x740C), PACK(0x75CE),
		PACK(0x7E90), PACK(0x7F52), PACK(0x7D14), PACK(0x7CD6),
		PACK(0x7998), PACK(0x785A), PACK(0x7A1C), PACK(0x7BDE),
		PACK(0x6CA0), PACK(0x6D62), PACK(0x6F24), PACK(0x6EE6),
		PACK(0x6BA8), PACK(0x6A6A), PACK(0x682C), PACK(0x69EE),
		PACK(0x62B0), PACK(0x6372), PACK(0x6134), PACK(0x60F6),
		PACK(0x65B8), PACK(0x647A), PACK(0x663C), PACK(0x67FE),
		PACK(0x48C0), PACK(0x4902), PACK(0x4B44), PACK(0x4A86),
		PACK(0x4FC8), PACK(0x4E0A), PACK(0x4C4C), PACK(0x4D8E),
		PACK(0x46D0), PACK(0x4712), PACK(0x4554), PACK(0x4496),
		PACK(0x41D8), PACK(0x401A), PACK(0x425C), PACK(0x439E),
		PACK(0x54E0), PACK(0x5522), PACK(0x5764), PACK(0x56A6),
		PACK(0x53E8), PACK(0x522A), PACK(0x506C), PACK(0x51AE),
		PACK(0x5AF0), PACK(0x5B32), PACK(0x5974), PACK(0x58B6),
		PACK(0x5DF8), PACK(0x5C3A), PACK(0x5E7C), PACK(0x5FBE),
		PACK(0xE100), PACK(0xE0C2), PACK(0xE284), PACK(0xE346),
		PACK(0xE608), PACK(0xE7CA), PACK(0xE58C), PACK(0xE44E),
		PACK(0xEF10), PACK(0xEED2), PACK(0xEC94), PACK(0xED56),
		PACK(0xE818), PACK(0xE9DA), PACK(0xEB9C), PACK(0xEA5E),
		PACK(0xFD20), PACK(0xFCE2), PACK(0xFEA4), PACK(0xFF66),
		PACK(0xFA28), PACK(0xFBEA), PACK(0xF9AC), PACK(0xF86E),
		PACK(0xF330), PACK(0xF2F2), PACK(0xF0B4), PACK(0xF176),
		PACK(0xF438), PACK(0xF5FA), PACK(0xF7BC), PACK(0xF67E),
		PACK(0xD940), PACK(0xD882), PACK(0xDAC4), PACK(0xDB06),
		PACK(0xDE48), PACK(0xDF8A), PACK(0xDDCC), PACK(0xDC0E),
		PACK(0xD750), PACK(0xD692), PACK(0xD4D4), PACK(0xD516),
		PACK(0xD058), PACK(0xD19A), PACK(0xD3DC), PACK(0xD21E),
		PACK(0xC560), PACK(0xC4A2), PACK(0xC6E4), PACK(0xC726),
		PACK(0xC268), PACK(0xC3AA), PACK(0xC1EC), PACK(0xC02E),
		PACK(0xCB70), PACK(0xCAB2), PACK(0xC8F4), PACK(0xC936),
		PACK(0xCC78), PACK(0xCDBA), PACK(0xCFFC), PACK(0xCE3E),
		PACK(0x9180), PACK(0x9042), PACK(0x9204), PACK(0x93C6),
		PACK(0x9688), PACK(0x974A), PACK(0x950C), PACK(0x94CE),
		PACK(0x9F90), PACK(0x9E52), PACK(0x9C14), PACK(0x9DD6),
		PACK(0x9898), PACK(0x995A), PACK(0x9B1C), PACK(0x9ADE),
		PACK(0x8DA0), PACK(0x8C62), PACK(0x8E24), PACK(0x8FE6),
		PACK(0x8AA8), PACK(0x8B6A), PACK(0x892C), PACK(0x88EE),
		PACK(0x83B0), PACK(0x8272), PACK(0x8034), PACK(0x81F6),
		PACK(0x84B8), PACK(0x857A), PACK(0x873C), PACK(0x86FE),
		PACK(0xA9C0), PACK(0xA802), PACK(0xAA44), PACK(0xAB86),
		PACK(0xAEC8), PACK(0xAF0A), PACK(0xAD4C), PACK(0xAC8E),
		PACK(0xA7D0), PACK(0xA612), PACK(0xA454), PACK(0xA596),
		PACK(0xA0D8), PACK(0xA11A), PACK(0xA35C), PACK(0xA29E),
		PACK(0xB5E0), PACK(0xB422), PACK(0xB664), PACK(0xB7A6),
		PACK(0xB2E8), PACK(0xB32A), PACK(0xB16C), PACK(0xB0AE),
		PACK(0xBBF0), PACK(0xBA32), PACK(0xB874), PACK(0xB9B6),
		PACK(0xBCF8), PACK(0xBD3A), PACK(0xBF7C), PACK(0xBEBE) };

	while (1) {
		Z.hi ^= Htable[n].hi;
		Z.lo ^= Htable[n].lo;

		if ((u8 *)Xi==xi)	break;

		n = *(--xi);

		rem  = (size_t)Z.lo&0xff;
		Z.lo = (Z.hi<<56)|(Z.lo>>8);
		Z.hi = (Z.hi>>8);
		if (sizeof(size_t)==8)
			Z.hi ^= rem_8bit[rem];
		else
			Z.hi ^= (u64)rem_8bit[rem]<<32;
	}

	if (is_endian.little) {
#ifdef BSWAP8
		Xi[0] = BSWAP8(Z.hi);
		Xi[1] = BSWAP8(Z.lo);
#else
		u8 *p = (u8 *)Xi;
		u32 v;
		v = (u32)(Z.hi>>32);	PUTU32(p,v);
		v = (u32)(Z.hi);	PUTU32(p+4,v);
		v = (u32)(Z.lo>>32);	PUTU32(p+8,v);
		v = (u32)(Z.lo);	PUTU32(p+12,v);
#endif
	}
	else {
		Xi[0] = Z.hi;
		Xi[1] = Z.lo;
	}
}
#endif

static void gcm_init_4bit(u128 Htable[16], u64 H[2])
{
	int  i, j;
	u128 V;

	Htable[0].hi = 0;
	Htable[0].lo = 0;
	V.hi = H[0];
	V.lo = H[1];

	for (Htable[8]=V, i=4; i>0; i>>=1) {
		if (sizeof(size_t)==8) {
			u64 T = U64(0xe100000000000000) & (0-(V.lo&1));
			V.lo  = (V.hi<<63)|(V.lo>>1);
			V.hi  = (V.hi>>1 )^T;
		}
		else {
			u32 T = 0xe1000000U & (0-(u32)(V.lo&1));
			V.lo  = (V.hi<<63)|(V.lo>>1);
			V.hi  = (V.hi>>1 )^((u64)T<<32);
		}
		Htable[i] = V;
	}

	for (i=2; i<16; i<<=1) {
		u128 *Hi = Htable+i, H0 = *Hi;
		for (j=1; j<i; ++j) {
			Hi[j].hi = H0.hi^Htable[j].hi;
			Hi[j].lo = H0.lo^Htable[j].lo;
		}
	}
}

static void gcm_mul_4bit(u64 Xi[2], u128 Htable[16])
{
	u128 Z = { 0, 0};
	const u8 *xi = (const u8 *)Xi+15;
	size_t rem, nlo = *xi, nhi;
	const union { long one; char little; } is_endian = {1};
	static const size_t rem_4bit[16] = {
		PACK(0x0000), PACK(0x1C20), PACK(0x3840), PACK(0x2460),
		PACK(0x7080), PACK(0x6CA0), PACK(0x48C0), PACK(0x54E0),
		PACK(0xE100), PACK(0xFD20), PACK(0xD940), PACK(0xC560),
		PACK(0x9180), PACK(0x8DA0), PACK(0xA9C0), PACK(0xB5E0) };

	while (1) {
		nhi  = nlo>>4;
		nlo &= 0xf;

		Z.hi ^= Htable[nlo].hi;
		Z.lo ^= Htable[nlo].lo;

		rem  = (size_t)Z.lo&0xf;
		Z.lo = (Z.hi<<60)|(Z.lo>>4);
		Z.hi = (Z.hi>>4);
		if (sizeof(size_t)==8)
			Z.hi ^= rem_4bit[rem];
		else
			Z.hi ^= (u64)rem_4bit[rem]<<32;

		Z.hi ^= Htable[nhi].hi;
		Z.lo ^= Htable[nhi].lo;

		if ((u8 *)Xi==xi)	break;

		nlo = *(--xi);

		rem  = (size_t)Z.lo&0xf;
		Z.lo = (Z.hi<<60)|(Z.lo>>4);
		Z.hi = (Z.hi>>4);
		if (sizeof(size_t)==8)
			Z.hi ^= rem_4bit[rem];
		else
			Z.hi ^= (u64)rem_4bit[rem]<<32;
	}

	if (is_endian.little) {
#ifdef BSWAP8
		Xi[0] = BSWAP8(Z.hi);
		Xi[1] = BSWAP8(Z.lo);
#else
		u8 *p = (u8 *)Xi;
		u32 v;
		v = (u32)(Z.hi>>32);	PUTU32(p,v);
		v = (u32)(Z.hi);	PUTU32(p+4,v);
		v = (u32)(Z.lo>>32);	PUTU32(p+8,v);
		v = (u32)(Z.lo);	PUTU32(p+12,v);
#endif
	}
	else {
		Xi[0] = Z.hi;
		Xi[1] = Z.lo;
	}
}

static void gcm_mul_1bit(u64 Xi[2],const u64 H[2])
{
	u128 V,Z = { 0,0 };
	long X;
	int  i,j;
	const long *xi = (const long *)Xi;
	const union { long one; char little; } is_endian = {1};

	V.hi = H[0];	/* h is in host byte order, no byte swaping */
	V.lo = H[1];

	for (j=0; j<16/sizeof(long); ++j) {
		if (is_endian.little) {
			if (sizeof(long)==8) {
#ifdef BSWAP8
				X = (long)(BSWAP8(xi[j]));
#else
				const u8 *p = (const u8 *)(xi+j);
				X = (long)((u64)GETU32(p)<<32|GETU32(p+4));
#endif
			}
			else {
				const u8 *p = (const u8 *)(xi+j);
				X = (long)GETU32(p);
			}
		}
		else
			X = xi[j];

		for (i=0; i<8*sizeof(long); ++i, X<<=1) {
			u64 M = (u64)(X>>(8*sizeof(long)-1));
			Z.hi ^= V.hi&M;
			Z.lo ^= V.lo&M;

			if (sizeof(size_t)==8) {
				u64 T = U64(0xe100000000000000) & (0-(V.lo&1));
				V.lo  = (V.hi<<63)|(V.lo>>1);
				V.hi  = (V.hi>>1 )^T;
			}
			else {
				u32 T = 0xe1000000U & (0-(u32)(V.lo&1));
				V.lo  = (V.hi<<63)|(V.lo>>1);
				V.hi  = (V.hi>>1 )^((u64)T<<32);
			}
				
		}
	}

	if (is_endian.little) {
#ifdef BSWAP8
		Xi[0] = BSWAP8(Z.hi);
		Xi[1] = BSWAP8(Z.lo);
#else
		u8 *p = (u8 *)Xi;
		u32 v;
		v = (u32)(Z.hi>>32);	PUTU32(p,v);
		v = (u32)(Z.hi);	PUTU32(p+4,v);
		v = (u32)(Z.lo>>32);	PUTU32(p+8,v);
		v = (u32)(Z.lo);	PUTU32(p+12,v);
#endif
	}
	else {
		Xi[0] = Z.hi;
		Xi[1] = Z.lo;
	}
}

#if 0
#define GCM_MUL(ctx,Xi)	gcm_mul_1bit(ctx->Xi.u,ctx->H.u)
#else
#define GCM_MUL(ctx,Xi)	gcm_mul_4bit(ctx->Xi.u,ctx->Htable)
#endif

typedef struct {
	/* Following 6 names follow names in GCM specification */
	union { u64 u[2]; u32 d[4]; u8 c[16]; }	Yi,EKi,EK0,
						Xi,H,
						len;
	/* Pre-computed table used by gcm_mul_4bit */
	u128 Htable[16];
	unsigned int res, ctr;
	block128_f block;
	void *key;
} GCM128_CONTEXT;

void CRYPTO_gcm128_init(GCM128_CONTEXT *ctx,void *key,block128_f block)
{
	const union { long one; char little; } is_endian = {1};

	memset(ctx,0,sizeof(*ctx));
	ctx->block = block;
	ctx->key   = key;

	(*block)(ctx->H.c,ctx->H.c,key);

	if (is_endian.little) {
		/* H is stored in host byte order */
#ifdef BSWAP8
		ctx->H.u[0] = BSWAP8(ctx->H.u[0]);
		ctx->H.u[1] = BSWAP8(ctx->H.u[1]);
#else
		u8 *p = ctx->H.c;
		u64 hi,lo;
		hi = (u64)GETU32(p)  <<32|GETU32(p+4);
		lo = (u64)GETU32(p+8)<<32|GETU32(p+12);
		ctx->H.u[0] = hi;
		ctx->H.u[1] = lo;
#endif
	}

	gcm_init_4bit(ctx->Htable,ctx->H.u);
}

void CRYPTO_gcm128_setiv(GCM128_CONTEXT *ctx,const unsigned char *iv,size_t len)
{
	const union { long one; char little; } is_endian = {1};

	ctx->Yi.u[0]  = 0;
	ctx->Yi.u[1]  = 0;
	ctx->Xi.u[0]  = 0;
	ctx->Xi.u[1]  = 0;
	ctx->len.u[0] = 0;
	ctx->len.u[1] = 0;
	ctx->res = 0;

	if (len==12) {
		memcpy(ctx->Yi.c,iv,12);
		ctx->Yi.c[15]=1;
		ctx->ctr=1;
	}
	else {
		size_t i;
		u64 len0 = len;

		while (len>=16) {
			for (i=0; i<16; ++i) ctx->Yi.c[i] ^= iv[i];
			GCM_MUL(ctx,Yi);
			iv += 16;
			len -= 16;
		}
		if (len) {
			for (i=0; i<len; ++i) ctx->Yi.c[i] ^= iv[i];
			GCM_MUL(ctx,Yi);
		}
		len0 <<= 3;
		if (is_endian.little) {
#ifdef BSWAP8
			ctx->Yi.u[1]  ^= BSWAP8(len0);
#else
			ctx->Yi.c[8]  ^= (u8)(len0>>56);
			ctx->Yi.c[9]  ^= (u8)(len0>>48);
			ctx->Yi.c[10] ^= (u8)(len0>>40);
			ctx->Yi.c[11] ^= (u8)(len0>>32);
			ctx->Yi.c[12] ^= (u8)(len0>>24);
			ctx->Yi.c[13] ^= (u8)(len0>>16);
			ctx->Yi.c[14] ^= (u8)(len0>>8);
			ctx->Yi.c[15] ^= (u8)(len0);
#endif
		}
		else
			ctx->Yi.u[1]  ^= len0;

		GCM_MUL(ctx,Yi);

		if (is_endian.little)
			ctx->ctr = GETU32(ctx->Yi.c+12);
		else
			ctx->ctr = ctx->Yi.d[3];
	}

	(*ctx->block)(ctx->Yi.c,ctx->EK0.c,ctx->key);
}

void CRYPTO_gcm128_aad(GCM128_CONTEXT *ctx,const unsigned char *aad,size_t len)
{
	size_t i;

	ctx->len.u[0] += len;

	while (len>=16) {
		for (i=0; i<16; ++i) ctx->Xi.c[i] ^= aad[i];
		GCM_MUL(ctx,Xi);
		aad += 16;
		len -= 16;
	}

	if (len) {
		for (i=0; i<len; ++i) ctx->Xi.c[i] ^= aad[i];
		GCM_MUL(ctx,Xi);
	}
}

void CRYPTO_gcm128_encrypt(GCM128_CONTEXT *ctx,
		const unsigned char *in, unsigned char *out,
		size_t len)
{
	const union { long one; char little; } is_endian = {1};
	unsigned int n, ctr;
	size_t i;

	ctx->len.u[1] += len;
	n   = ctx->res;
	ctr = ctx->ctr;

#if !defined(OPENSSL_SMALL_FOOTPRINT)
	if (16%sizeof(size_t) == 0) do {	/* always true actually */
		if (n) {
			while (n && len) {
				ctx->Xi.c[n] ^= *(out++) = *(in++)^ctx->EKi.c[n];
				--len;
				n = (n+1)%16;
			}
			if (n==0) GCM_MUL(ctx,Xi);
			else {
				ctx->res = n;
				return;
			}
		}

#if defined(STRICT_ALIGNMENT)
		if (((size_t)in|(size_t)out)%sizeof(size_t) != 0)
			break;
#endif
		while (len>=16) {
			++ctr;
			if (is_endian.little)
				PUTU32(ctx->Yi.c+12,ctr);
			else
				ctx->Yi.d[3] = ctr;
			(*ctx->block)(ctx->Yi.c,ctx->EKi.c,ctx->key);
			for (i=0; i<16; i+=sizeof(size_t))
				*(size_t *)(ctx->Xi.c+i) ^=
				*(size_t *)(out+i) =
				*(size_t *)(in+i)^*(size_t *)(ctx->EKi.c+i);
			GCM_MUL(ctx,Xi);
			out += 16;
			in  += 16;
			len -= 16;
		}

		if (len) {
			++ctr;
			if (is_endian.little)
				PUTU32(ctx->Yi.c+12,ctr);
			else
				ctx->Yi.d[3] = ctr;
			(*ctx->block)(ctx->Yi.c,ctx->EKi.c,ctx->key);
			while (len--) {
				ctx->Xi.c[n] ^= out[n] = in[n]^ctx->EKi.c[n];
				++n;
			}
		}

		ctx->res = n;
		ctx->ctr = ctr;
		return;
	} while(0);
#endif
	for (i=0;i<len;++i) {
		if (n==0) {
			++ctr;
			if (is_endian.little)
				PUTU32(ctx->Yi.c+12,ctr);
			else
				ctx->Yi.d[3] = ctr;
			(*ctx->block)(ctx->Yi.c,ctx->EKi.c,ctx->key);
		}
		ctx->Xi.c[n] ^= out[i] = in[i]^ctx->EKi.c[n];
		n = (n+1)%16;
		if (n==0)
			GCM_MUL(ctx,Xi);
	}

	ctx->res = n;
	ctx->ctr = ctr;
}

void CRYPTO_gcm128_decrypt(GCM128_CONTEXT *ctx,
		const unsigned char *in, unsigned char *out,
		size_t len)
{
	const union { long one; char little; } is_endian = {1};
	unsigned int n, ctr;
	size_t i;

	ctx->len.u[1] += len;
	n   = ctx->res;
	ctr = ctx->ctr;

#if !defined(OPENSSL_SMALL_FOOTPRINT)
	if (16%sizeof(size_t) == 0) do {	/* always true actually */
		if (n) {
			while (n && len) {
				u8 c = *(in++);
				*(out++) = c^ctx->EKi.c[n];
				ctx->Xi.c[n] ^= c;
				--len;
				n = (n+1)%16;
			}
			if (n==0) GCM_MUL (ctx,Xi);
			else {
				ctx->res = n;
				return;
			}
		}

#if defined(STRICT_ALIGNMENT)
		if (((size_t)in|(size_t)out)%sizeof(size_t) != 0)
			break;
#endif
		while (len>=16) {
			++ctr;
			if (is_endian.little)
				PUTU32(ctx->Yi.c+12,ctr);
			else
				ctx->Yi.d[3] = ctr;
			(*ctx->block)(ctx->Yi.c,ctx->EKi.c,ctx->key);
			for (i=0; i<16; i+=sizeof(size_t)) {
				size_t c = *(size_t *)(in+i);
				*(size_t *)(out+i) = c^*(size_t *)(ctx->EKi.c+i);
				*(size_t *)(ctx->Xi.c+i) ^= c;
			}
			GCM_MUL (ctx,Xi);
			out += 16;
			in  += 16;
			len -= 16;
		}

		if (len) {
			++ctr;
			if (is_endian.little)
				PUTU32(ctx->Yi.c+12,ctr);
			else
				ctx->Yi.d[3] = ctr;
			(*ctx->block)(ctx->Yi.c,ctx->EKi.c,ctx->key);
			while (len--) {
				u8 c = in[n];
				ctx->Xi.c[n] ^= c;
				out[n] = c^ctx->EKi.c[n];
				++n;
			}
		}

		ctx->res = n;
		ctx->ctr = ctr;
		return;
	} while(0);
#endif
	for (i=0;i<len;++i) {
		u8 c;
		if (n==0) {
			++ctr;
			if (is_endian.little)
				PUTU32(ctx->Yi.c+12,ctr);
			else
				ctx->Yi.d[3] = ctr;
			(*ctx->block)(ctx->Yi.c,ctx->EKi.c,ctx->key);
		}
		c = in[i];
		out[i] ^= ctx->EKi.c[n];
		ctx->Xi.c[n] ^= c;
		n = (n+1)%16;
		if (n==0)
			GCM_MUL(ctx,Xi);
	}

	ctx->res = n;
	ctx->ctr = ctr;
}

void CRYPTO_gcm128_finish(GCM128_CONTEXT *ctx)
{
	const union { long one; char little; } is_endian = {1};
	u64 alen = ctx->len.u[0]<<3;
	u64 clen = ctx->len.u[1]<<3;

	if (ctx->res)
		GCM_MUL(ctx,Xi);

	if (is_endian.little) {
#ifdef BSWAP8
		alen = BSWAP8(alen);
		clen = BSWAP8(clen);
#else
		u8 *p = ctx->len.c;

		ctx->len.u[0] = alen;
		ctx->len.u[1] = clen;

		alen = (u64)GETU32(p)  <<32|GETU32(p+4);
		clen = (u64)GETU32(p+8)<<32|GETU32(p+12);
#endif
	}

	ctx->Xi.u[0] ^= alen;
	ctx->Xi.u[1] ^= clen;
	GCM_MUL(ctx,Xi);

	ctx->Xi.u[0] ^= ctx->EK0.u[0];
	ctx->Xi.u[1] ^= ctx->EK0.u[1];
}

#if defined(SELFTEST)
#include <stdio.h>
#include <openssl/aes.h>

/* Test Case 1 */
static const u8	K1[16],
		*P1=NULL,
		*A1=NULL,
		IV1[12],
		*C1=NULL,
		T1[]=  {0x58,0xe2,0xfc,0xce,0xfa,0x7e,0x30,0x61,0x36,0x7f,0x1d,0x57,0xa4,0xe7,0x45,0x5a};
/* Test Case 2 */
#define K2 K1
#define A2 A1
#define IV2 IV1
static const u8	P2[16],
		C2[]=  {0x03,0x88,0xda,0xce,0x60,0xb6,0xa3,0x92,0xf3,0x28,0xc2,0xb9,0x71,0xb2,0xfe,0x78},
		T2[]=  {0xab,0x6e,0x47,0xd4,0x2c,0xec,0x13,0xbd,0xf5,0x3a,0x67,0xb2,0x12,0x57,0xbd,0xdf};

/* Test Case 3 */
#define A3 A2
static const u8	K3[]=  {0xfe,0xff,0xe9,0x92,0x86,0x65,0x73,0x1c,0x6d,0x6a,0x8f,0x94,0x67,0x30,0x83,0x08},
		P3[]=  {0xd9,0x31,0x32,0x25,0xf8,0x84,0x06,0xe5,0xa5,0x59,0x09,0xc5,0xaf,0xf5,0x26,0x9a,
			0x86,0xa7,0xa9,0x53,0x15,0x34,0xf7,0xda,0x2e,0x4c,0x30,0x3d,0x8a,0x31,0x8a,0x72,
			0x1c,0x3c,0x0c,0x95,0x95,0x68,0x09,0x53,0x2f,0xcf,0x0e,0x24,0x49,0xa6,0xb5,0x25,
			0xb1,0x6a,0xed,0xf5,0xaa,0x0d,0xe6,0x57,0xba,0x63,0x7b,0x39,0x1a,0xaf,0xd2,0x55},
		IV3[]= {0xca,0xfe,0xba,0xbe,0xfa,0xce,0xdb,0xad,0xde,0xca,0xf8,0x88},
		C3[]=  {0x42,0x83,0x1e,0xc2,0x21,0x77,0x74,0x24,0x4b,0x72,0x21,0xb7,0x84,0xd0,0xd4,0x9c,
			0xe3,0xaa,0x21,0x2f,0x2c,0x02,0xa4,0xe0,0x35,0xc1,0x7e,0x23,0x29,0xac,0xa1,0x2e,
			0x21,0xd5,0x14,0xb2,0x54,0x66,0x93,0x1c,0x7d,0x8f,0x6a,0x5a,0xac,0x84,0xaa,0x05,
			0x1b,0xa3,0x0b,0x39,0x6a,0x0a,0xac,0x97,0x3d,0x58,0xe0,0x91,0x47,0x3f,0x59,0x85},
		T3[]=  {0x4d,0x5c,0x2a,0xf3,0x27,0xcd,0x64,0xa6,0x2c,0xf3,0x5a,0xbd,0x2b,0xa6,0xfa,0xb4,};

/* Test Case 4 */
#define K4 K3
#define IV4 IV3
static const u8	P4[]=  {0xd9,0x31,0x32,0x25,0xf8,0x84,0x06,0xe5,0xa5,0x59,0x09,0xc5,0xaf,0xf5,0x26,0x9a,
			0x86,0xa7,0xa9,0x53,0x15,0x34,0xf7,0xda,0x2e,0x4c,0x30,0x3d,0x8a,0x31,0x8a,0x72,
			0x1c,0x3c,0x0c,0x95,0x95,0x68,0x09,0x53,0x2f,0xcf,0x0e,0x24,0x49,0xa6,0xb5,0x25,
			0xb1,0x6a,0xed,0xf5,0xaa,0x0d,0xe6,0x57,0xba,0x63,0x7b,0x39},
		A4[]=  {0xfe,0xed,0xfa,0xce,0xde,0xad,0xbe,0xef,0xfe,0xed,0xfa,0xce,0xde,0xad,0xbe,0xef,
			0xab,0xad,0xda,0xd2},
		C4[]=  {0x42,0x83,0x1e,0xc2,0x21,0x77,0x74,0x24,0x4b,0x72,0x21,0xb7,0x84,0xd0,0xd4,0x9c,
			0xe3,0xaa,0x21,0x2f,0x2c,0x02,0xa4,0xe0,0x35,0xc1,0x7e,0x23,0x29,0xac,0xa1,0x2e,
			0x21,0xd5,0x14,0xb2,0x54,0x66,0x93,0x1c,0x7d,0x8f,0x6a,0x5a,0xac,0x84,0xaa,0x05,
			0x1b,0xa3,0x0b,0x39,0x6a,0x0a,0xac,0x97,0x3d,0x58,0xe0,0x91},
		T4[]=  {0x5b,0xc9,0x4f,0xbc,0x32,0x21,0xa5,0xdb,0x94,0xfa,0xe9,0x5a,0xe7,0x12,0x1a,0x47};

/* Test Case 5 */
#define K5 K4
#define P5 P4
static const u8	A5[]=  {0xfe,0xed,0xfa,0xce,0xde,0xad,0xbe,0xef,0xfe,0xed,0xfa,0xce,0xde,0xad,0xbe,0xef,
			0xab,0xad,0xda,0xd2},
		IV5[]= {0xca,0xfe,0xba,0xbe,0xfa,0xce,0xdb,0xad},
		C5[]=  {0x61,0x35,0x3b,0x4c,0x28,0x06,0x93,0x4a,0x77,0x7f,0xf5,0x1f,0xa2,0x2a,0x47,0x55,
			0x69,0x9b,0x2a,0x71,0x4f,0xcd,0xc6,0xf8,0x37,0x66,0xe5,0xf9,0x7b,0x6c,0x74,0x23,
			0x73,0x80,0x69,0x00,0xe4,0x9f,0x24,0xb2,0x2b,0x09,0x75,0x44,0xd4,0x89,0x6b,0x42,
			0x49,0x89,0xb5,0xe1,0xeb,0xac,0x0f,0x07,0xc2,0x3f,0x45,0x98},
		T5[]=  {0x36,0x12,0xd2,0xe7,0x9e,0x3b,0x07,0x85,0x56,0x1b,0xe1,0x4a,0xac,0xa2,0xfc,0xcb};
/* Test Case 6 */
#define K6 K5
#define P6 P5
#define A6 A5
static const u8	IV6[]= {0x93,0x13,0x22,0x5d,0xf8,0x84,0x06,0xe5,0x55,0x90,0x9c,0x5a,0xff,0x52,0x69,0xaa,
			0x6a,0x7a,0x95,0x38,0x53,0x4f,0x7d,0xa1,0xe4,0xc3,0x03,0xd2,0xa3,0x18,0xa7,0x28,
			0xc3,0xc0,0xc9,0x51,0x56,0x80,0x95,0x39,0xfc,0xf0,0xe2,0x42,0x9a,0x6b,0x52,0x54,
			0x16,0xae,0xdb,0xf5,0xa0,0xde,0x6a,0x57,0xa6,0x37,0xb3,0x9b},
		C6[]=  {0x8c,0xe2,0x49,0x98,0x62,0x56,0x15,0xb6,0x03,0xa0,0x33,0xac,0xa1,0x3f,0xb8,0x94,
			0xbe,0x91,0x12,0xa5,0xc3,0xa2,0x11,0xa8,0xba,0x26,0x2a,0x3c,0xca,0x7e,0x2c,0xa7,
			0x01,0xe4,0xa9,0xa4,0xfb,0xa4,0x3c,0x90,0xcc,0xdc,0xb2,0x81,0xd4,0x8c,0x7c,0x6f,
			0xd6,0x28,0x75,0xd2,0xac,0xa4,0x17,0x03,0x4c,0x34,0xae,0xe5},
		T6[]=  {0x61,0x9c,0xc5,0xae,0xff,0xfe,0x0b,0xfa,0x46,0x2a,0xf4,0x3c,0x16,0x99,0xd0,0x50};

/* Test Case 7 */
static const u8 K7[24],
		*P7=NULL,
		*A7=NULL,
		IV7[12],
		*C7=NULL,
		T7[]=  {0xcd,0x33,0xb2,0x8a,0xc7,0x73,0xf7,0x4b,0xa0,0x0e,0xd1,0xf3,0x12,0x57,0x24,0x35};

/* Test Case 8 */
#define K8 K7
#define IV8 IV7
#define A8 A7
static const u8	P8[16],
		C8[]=  {0x98,0xe7,0x24,0x7c,0x07,0xf0,0xfe,0x41,0x1c,0x26,0x7e,0x43,0x84,0xb0,0xf6,0x00},
		T8[]=  {0x2f,0xf5,0x8d,0x80,0x03,0x39,0x27,0xab,0x8e,0xf4,0xd4,0x58,0x75,0x14,0xf0,0xfb};

/* Test Case 9 */
#define A9 A8
static const u8	K9[]=  {0xfe,0xff,0xe9,0x92,0x86,0x65,0x73,0x1c,0x6d,0x6a,0x8f,0x94,0x67,0x30,0x83,0x08,
			0xfe,0xff,0xe9,0x92,0x86,0x65,0x73,0x1c},
		P9[]=  {0xd9,0x31,0x32,0x25,0xf8,0x84,0x06,0xe5,0xa5,0x59,0x09,0xc5,0xaf,0xf5,0x26,0x9a,
			0x86,0xa7,0xa9,0x53,0x15,0x34,0xf7,0xda,0x2e,0x4c,0x30,0x3d,0x8a,0x31,0x8a,0x72,
			0x1c,0x3c,0x0c,0x95,0x95,0x68,0x09,0x53,0x2f,0xcf,0x0e,0x24,0x49,0xa6,0xb5,0x25,
			0xb1,0x6a,0xed,0xf5,0xaa,0x0d,0xe6,0x57,0xba,0x63,0x7b,0x39,0x1a,0xaf,0xd2,0x55},
		IV9[]= {0xca,0xfe,0xba,0xbe,0xfa,0xce,0xdb,0xad,0xde,0xca,0xf8,0x88},
		C9[]=  {0x39,0x80,0xca,0x0b,0x3c,0x00,0xe8,0x41,0xeb,0x06,0xfa,0xc4,0x87,0x2a,0x27,0x57,
			0x85,0x9e,0x1c,0xea,0xa6,0xef,0xd9,0x84,0x62,0x85,0x93,0xb4,0x0c,0xa1,0xe1,0x9c,
			0x7d,0x77,0x3d,0x00,0xc1,0x44,0xc5,0x25,0xac,0x61,0x9d,0x18,0xc8,0x4a,0x3f,0x47,
			0x18,0xe2,0x44,0x8b,0x2f,0xe3,0x24,0xd9,0xcc,0xda,0x27,0x10,0xac,0xad,0xe2,0x56},
		T9[]=  {0x99,0x24,0xa7,0xc8,0x58,0x73,0x36,0xbf,0xb1,0x18,0x02,0x4d,0xb8,0x67,0x4a,0x14};

/* Test Case 10 */
#define K10 K9
#define IV10 IV9
static const u8	P10[]= {0xd9,0x31,0x32,0x25,0xf8,0x84,0x06,0xe5,0xa5,0x59,0x09,0xc5,0xaf,0xf5,0x26,0x9a,
			0x86,0xa7,0xa9,0x53,0x15,0x34,0xf7,0xda,0x2e,0x4c,0x30,0x3d,0x8a,0x31,0x8a,0x72,
			0x1c,0x3c,0x0c,0x95,0x95,0x68,0x09,0x53,0x2f,0xcf,0x0e,0x24,0x49,0xa6,0xb5,0x25,
			0xb1,0x6a,0xed,0xf5,0xaa,0x0d,0xe6,0x57,0xba,0x63,0x7b,0x39},
		A10[]= {0xfe,0xed,0xfa,0xce,0xde,0xad,0xbe,0xef,0xfe,0xed,0xfa,0xce,0xde,0xad,0xbe,0xef,
			0xab,0xad,0xda,0xd2},
		C10[]= {0x39,0x80,0xca,0x0b,0x3c,0x00,0xe8,0x41,0xeb,0x06,0xfa,0xc4,0x87,0x2a,0x27,0x57,
			0x85,0x9e,0x1c,0xea,0xa6,0xef,0xd9,0x84,0x62,0x85,0x93,0xb4,0x0c,0xa1,0xe1,0x9c,
			0x7d,0x77,0x3d,0x00,0xc1,0x44,0xc5,0x25,0xac,0x61,0x9d,0x18,0xc8,0x4a,0x3f,0x47,
			0x18,0xe2,0x44,0x8b,0x2f,0xe3,0x24,0xd9,0xcc,0xda,0x27,0x10},
		T10[]= {0x25,0x19,0x49,0x8e,0x80,0xf1,0x47,0x8f,0x37,0xba,0x55,0xbd,0x6d,0x27,0x61,0x8c};

/* Test Case 11 */
#define K11 K10
#define P11 P10
#define A11 A10
static const u8	IV11[]={0xca,0xfe,0xba,0xbe,0xfa,0xce,0xdb,0xad},
		C11[]= {0x0f,0x10,0xf5,0x99,0xae,0x14,0xa1,0x54,0xed,0x24,0xb3,0x6e,0x25,0x32,0x4d,0xb8,
			0xc5,0x66,0x63,0x2e,0xf2,0xbb,0xb3,0x4f,0x83,0x47,0x28,0x0f,0xc4,0x50,0x70,0x57,
			0xfd,0xdc,0x29,0xdf,0x9a,0x47,0x1f,0x75,0xc6,0x65,0x41,0xd4,0xd4,0xda,0xd1,0xc9,
			0xe9,0x3a,0x19,0xa5,0x8e,0x8b,0x47,0x3f,0xa0,0xf0,0x62,0xf7},
		T11[]= {0x65,0xdc,0xc5,0x7f,0xcf,0x62,0x3a,0x24,0x09,0x4f,0xcc,0xa4,0x0d,0x35,0x33,0xf8};

/* Test Case 12 */
#define K12 K11
#define P12 P11
#define A12 A11
static const u8	IV12[]={0x93,0x13,0x22,0x5d,0xf8,0x84,0x06,0xe5,0x55,0x90,0x9c,0x5a,0xff,0x52,0x69,0xaa,
			0x6a,0x7a,0x95,0x38,0x53,0x4f,0x7d,0xa1,0xe4,0xc3,0x03,0xd2,0xa3,0x18,0xa7,0x28,
			0xc3,0xc0,0xc9,0x51,0x56,0x80,0x95,0x39,0xfc,0xf0,0xe2,0x42,0x9a,0x6b,0x52,0x54,
			0x16,0xae,0xdb,0xf5,0xa0,0xde,0x6a,0x57,0xa6,0x37,0xb3,0x9b},
		C12[]= {0xd2,0x7e,0x88,0x68,0x1c,0xe3,0x24,0x3c,0x48,0x30,0x16,0x5a,0x8f,0xdc,0xf9,0xff,
			0x1d,0xe9,0xa1,0xd8,0xe6,0xb4,0x47,0xef,0x6e,0xf7,0xb7,0x98,0x28,0x66,0x6e,0x45,
			0x81,0xe7,0x90,0x12,0xaf,0x34,0xdd,0xd9,0xe2,0xf0,0x37,0x58,0x9b,0x29,0x2d,0xb3,
			0xe6,0x7c,0x03,0x67,0x45,0xfa,0x22,0xe7,0xe9,0xb7,0x37,0x3b},
		T12[]= {0xdc,0xf5,0x66,0xff,0x29,0x1c,0x25,0xbb,0xb8,0x56,0x8f,0xc3,0xd3,0x76,0xa6,0xd9};

/* Test Case 13 */
static const u8	K13[32],
		*P13=NULL,
		*A13=NULL,
		IV13[12],
		*C13=NULL,
		T13[]={0x53,0x0f,0x8a,0xfb,0xc7,0x45,0x36,0xb9,0xa9,0x63,0xb4,0xf1,0xc4,0xcb,0x73,0x8b};

/* Test Case 14 */
#define K14 K13
#define A14 A13
static const u8	P14[16],
		IV14[12],
		C14[]= {0xce,0xa7,0x40,0x3d,0x4d,0x60,0x6b,0x6e,0x07,0x4e,0xc5,0xd3,0xba,0xf3,0x9d,0x18},
		T14[]= {0xd0,0xd1,0xc8,0xa7,0x99,0x99,0x6b,0xf0,0x26,0x5b,0x98,0xb5,0xd4,0x8a,0xb9,0x19};

/* Test Case 15 */
#define A15 A14
static const u8	K15[]= {0xfe,0xff,0xe9,0x92,0x86,0x65,0x73,0x1c,0x6d,0x6a,0x8f,0x94,0x67,0x30,0x83,0x08,
			0xfe,0xff,0xe9,0x92,0x86,0x65,0x73,0x1c,0x6d,0x6a,0x8f,0x94,0x67,0x30,0x83,0x08},
		P15[]= {0xd9,0x31,0x32,0x25,0xf8,0x84,0x06,0xe5,0xa5,0x59,0x09,0xc5,0xaf,0xf5,0x26,0x9a,
			0x86,0xa7,0xa9,0x53,0x15,0x34,0xf7,0xda,0x2e,0x4c,0x30,0x3d,0x8a,0x31,0x8a,0x72,
			0x1c,0x3c,0x0c,0x95,0x95,0x68,0x09,0x53,0x2f,0xcf,0x0e,0x24,0x49,0xa6,0xb5,0x25,
			0xb1,0x6a,0xed,0xf5,0xaa,0x0d,0xe6,0x57,0xba,0x63,0x7b,0x39,0x1a,0xaf,0xd2,0x55},
		IV15[]={0xca,0xfe,0xba,0xbe,0xfa,0xce,0xdb,0xad,0xde,0xca,0xf8,0x88},
		C15[]= {0x52,0x2d,0xc1,0xf0,0x99,0x56,0x7d,0x07,0xf4,0x7f,0x37,0xa3,0x2a,0x84,0x42,0x7d,
			0x64,0x3a,0x8c,0xdc,0xbf,0xe5,0xc0,0xc9,0x75,0x98,0xa2,0xbd,0x25,0x55,0xd1,0xaa,
			0x8c,0xb0,0x8e,0x48,0x59,0x0d,0xbb,0x3d,0xa7,0xb0,0x8b,0x10,0x56,0x82,0x88,0x38,
			0xc5,0xf6,0x1e,0x63,0x93,0xba,0x7a,0x0a,0xbc,0xc9,0xf6,0x62,0x89,0x80,0x15,0xad},
		T15[]= {0xb0,0x94,0xda,0xc5,0xd9,0x34,0x71,0xbd,0xec,0x1a,0x50,0x22,0x70,0xe3,0xcc,0x6c};

/* Test Case 16 */
#define K16 K15
#define IV16 IV15
static const u8	P16[]= {0xd9,0x31,0x32,0x25,0xf8,0x84,0x06,0xe5,0xa5,0x59,0x09,0xc5,0xaf,0xf5,0x26,0x9a,
			0x86,0xa7,0xa9,0x53,0x15,0x34,0xf7,0xda,0x2e,0x4c,0x30,0x3d,0x8a,0x31,0x8a,0x72,
			0x1c,0x3c,0x0c,0x95,0x95,0x68,0x09,0x53,0x2f,0xcf,0x0e,0x24,0x49,0xa6,0xb5,0x25,
			0xb1,0x6a,0xed,0xf5,0xaa,0x0d,0xe6,0x57,0xba,0x63,0x7b,0x39},
		A16[]= {0xfe,0xed,0xfa,0xce,0xde,0xad,0xbe,0xef,0xfe,0xed,0xfa,0xce,0xde,0xad,0xbe,0xef,
			0xab,0xad,0xda,0xd2},
		C16[]= {0x52,0x2d,0xc1,0xf0,0x99,0x56,0x7d,0x07,0xf4,0x7f,0x37,0xa3,0x2a,0x84,0x42,0x7d,
			0x64,0x3a,0x8c,0xdc,0xbf,0xe5,0xc0,0xc9,0x75,0x98,0xa2,0xbd,0x25,0x55,0xd1,0xaa,
			0x8c,0xb0,0x8e,0x48,0x59,0x0d,0xbb,0x3d,0xa7,0xb0,0x8b,0x10,0x56,0x82,0x88,0x38,
			0xc5,0xf6,0x1e,0x63,0x93,0xba,0x7a,0x0a,0xbc,0xc9,0xf6,0x62},
		T16[]= {0x76,0xfc,0x6e,0xce,0x0f,0x4e,0x17,0x68,0xcd,0xdf,0x88,0x53,0xbb,0x2d,0x55,0x1b};

/* Test Case 17 */
#define K17 K16
#define P17 P16
#define A17 A16
static const u8	IV17[]={0xca,0xfe,0xba,0xbe,0xfa,0xce,0xdb,0xad},
		C17[]= {0xc3,0x76,0x2d,0xf1,0xca,0x78,0x7d,0x32,0xae,0x47,0xc1,0x3b,0xf1,0x98,0x44,0xcb,
			0xaf,0x1a,0xe1,0x4d,0x0b,0x97,0x6a,0xfa,0xc5,0x2f,0xf7,0xd7,0x9b,0xba,0x9d,0xe0,
			0xfe,0xb5,0x82,0xd3,0x39,0x34,0xa4,0xf0,0x95,0x4c,0xc2,0x36,0x3b,0xc7,0x3f,0x78,
			0x62,0xac,0x43,0x0e,0x64,0xab,0xe4,0x99,0xf4,0x7c,0x9b,0x1f},
		T17[]= {0x3a,0x33,0x7d,0xbf,0x46,0xa7,0x92,0xc4,0x5e,0x45,0x49,0x13,0xfe,0x2e,0xa8,0xf2};

/* Test Case 18 */
#define K18 K17
#define P18 P17
#define A18 A17
static const u8	IV18[]={0x93,0x13,0x22,0x5d,0xf8,0x84,0x06,0xe5,0x55,0x90,0x9c,0x5a,0xff,0x52,0x69,0xaa,
			0x6a,0x7a,0x95,0x38,0x53,0x4f,0x7d,0xa1,0xe4,0xc3,0x03,0xd2,0xa3,0x18,0xa7,0x28,
			0xc3,0xc0,0xc9,0x51,0x56,0x80,0x95,0x39,0xfc,0xf0,0xe2,0x42,0x9a,0x6b,0x52,0x54,
			0x16,0xae,0xdb,0xf5,0xa0,0xde,0x6a,0x57,0xa6,0x37,0xb3,0x9b},
		C18[]= {0x5a,0x8d,0xef,0x2f,0x0c,0x9e,0x53,0xf1,0xf7,0x5d,0x78,0x53,0x65,0x9e,0x2a,0x20,
			0xee,0xb2,0xb2,0x2a,0xaf,0xde,0x64,0x19,0xa0,0x58,0xab,0x4f,0x6f,0x74,0x6b,0xf4,
			0x0f,0xc0,0xc3,0xb7,0x80,0xf2,0x44,0x45,0x2d,0xa3,0xeb,0xf1,0xc5,0xd8,0x2c,0xde,
			0xa2,0x41,0x89,0x97,0x20,0x0e,0xf8,0x2e,0x44,0xae,0x7e,0x3f},
		T18[]= {0xa4,0x4a,0x82,0x66,0xee,0x1c,0x8e,0xb0,0xc8,0xb5,0xd4,0xcf,0x5a,0xe9,0xf1,0x9a};

#define TEST_CASE(n)	do {					\
	u8 out[sizeof(P##n)];					\
	AES_set_encrypt_key(K##n,sizeof(K##n)*8,&key);		\
	CRYPTO_gcm128_init(&ctx,&key,(block128_f)AES_encrypt);	\
	CRYPTO_gcm128_setiv(&ctx,IV##n,sizeof(IV##n));		\
	if (A##n) CRYPTO_gcm128_aad(&ctx,A##n,sizeof(A##n));	\
	if (P##n) CRYPTO_gcm128_encrypt(&ctx,P##n,out,sizeof(out));	\
	CRYPTO_gcm128_finish(&ctx);				\
	if (memcmp(ctx.Xi.c,T##n,16) || (C##n && memcmp(out,C##n,sizeof(out))))	\
		ret++, printf ("encrypt test#%d failed.\n",##n);\
	CRYPTO_gcm128_setiv(&ctx,IV##n,sizeof(IV##n));		\
	if (A##n) CRYPTO_gcm128_aad(&ctx,A##n,sizeof(A##n));	\
	if (C##n) CRYPTO_gcm128_decrypt(&ctx,C##n,out,sizeof(out));	\
	CRYPTO_gcm128_finish(&ctx);				\
	if (memcmp(ctx.Xi.c,T##n,16) || (P##n && memcmp(out,P##n,sizeof(out))))	\
		ret++, printf ("decrypt test#%d failed.\n",##n);\
	} while(0)

int main()
{
	GCM128_CONTEXT ctx;
	AES_KEY key;
	int ret=0;

	TEST_CASE(1);
	TEST_CASE(2);
	TEST_CASE(3);
	TEST_CASE(4);
	TEST_CASE(5);
	TEST_CASE(6);
	TEST_CASE(7);
	TEST_CASE(8);
	TEST_CASE(9);
	TEST_CASE(10);
	TEST_CASE(11);
	TEST_CASE(12);
	TEST_CASE(13);
	TEST_CASE(14);
	TEST_CASE(15);
	TEST_CASE(16);
	TEST_CASE(17);
	TEST_CASE(18);

	return ret;
}
#endif