* Changed all the SHA code over to the standard openssl code (no changes)

  Next to convert over the MD5 functions
  (This is done for portability for other architectures/oses)


svn: 1914

autotools/configure.ac

@@ -98,7 +98,7 @@ AC_TYPE_UID_T
 
 #AC_C_CONST
 AC_C_VOLATILE
-AC_C_BIGENDIAN
+AC_C_BIGENDIAN(AC_DEFINE(B_ENDIAN, 1, [big endian]),AC_DEFINE(L_ENDIAN, 1, [little endian]))
 #AC_C_INLINE
 AC_C_CHAR_UNSIGNED
 #AC_C_STRINGIZE

autotools/includes/acinclude.m4

@@ -289,6 +289,9 @@ case "$egg_cv_var_system_type" in
   ;;
   IRIX)
   ;;
+  HP-UX)
+    AC_DEFINE(MD32_XARRAY, 1, [Define under HPUX])
+  ;;
   Ultrix)
     SHELL=/bin/sh5
   ;;

config.h.in

@@ -7,6 +7,9 @@
    */
 #undef BROKEN_SNPRINTF
 
+/* big endian */
+#undef B_ENDIAN
+
 /* Define if running under cygwin */
 #undef CYGWIN_HACKS
 
@@ -227,6 +230,12 @@
    slash. */
 #undef LSTAT_FOLLOWS_SLASHED_SYMLINK
 
+/* little endian */
+#undef L_ENDIAN
+
+/* Define under HPUX */
+#undef MD32_XARRAY
+
 /* Define if running on OSF/1 platform */
 #undef OSF1_HACKS
 
@@ -281,10 +290,6 @@
 /* Define if windows */
 #undef WIN32_LEAN_AND_MEAN
 
-/* Define to 1 if your processor stores words with the most significant byte
-   first (like Motorola and SPARC, unlike Intel and VAX). */
-#undef WORDS_BIGENDIAN
-
 /* Define to 1 if on AIX 3.
    System headers sometimes define this.
    We just want to avoid a redefinition error message.  */

configure

@@ -1,5 +1,5 @@
 #! /bin/sh
-# From configure.ac Revision: 1.42 .
+# From configure.ac Revision: 1.43 .
 # Guess values for system-dependent variables and create Makefiles.
 # Generated by GNU Autoconf 2.59.
 #
@@ -3546,6 +3546,13 @@ _ACEOF
 
   ;;
   IRIX)
+  ;;
+  HP-UX)
+
+cat >>confdefs.h <<\_ACEOF
+#define MD32_XARRAY 1
+_ACEOF
+
   ;;
   Ultrix)
     SHELL=/bin/sh5
@@ -5454,11 +5461,15 @@ case $ac_cv_c_bigendian in
   yes)
 
 cat >>confdefs.h <<\_ACEOF
-#define WORDS_BIGENDIAN 1
+#define B_ENDIAN 1
 _ACEOF
  ;;
   no)
-     ;;
+
+cat >>confdefs.h <<\_ACEOF
+#define L_ENDIAN 1
+_ACEOF
+ ;;
   *)
     { { echo "$as_me:$LINENO: error: unknown endianness
 presetting ac_cv_c_bigendian=no (or yes) will help" >&5

src/crypto/md32_common.h

@@ -0,0 +1,588 @@
+/* crypto/md32_common.h */
+
+/*
+ * This is a generic 32 bit "collector" for message digest algorithms.
+ * Whenever needed it collects input character stream into chunks of
+ * 32 bit values and invokes a block function that performs actual hash
+ * calculations.
+ *
+ * Porting guide.
+ *
+ * Obligatory macros:
+ *
+ * DATA_ORDER_IS_BIG_ENDIAN or DATA_ORDER_IS_LITTLE_ENDIAN
+ *	this macro defines byte order of input stream.
+ * HASH_CBLOCK
+ *	size of a unit chunk HASH_BLOCK operates on.
+ * HASH_LONG
+ *	has to be at lest 32 bit wide, if it's wider, then
+ *	HASH_LONG_LOG2 *has to* be defined along
+ * HASH_CTX
+ *	context structure that at least contains following
+ *	members:
+ *		typedef struct {
+ *			...
+ *			HASH_LONG	Nl,Nh;
+ *			HASH_LONG	data[HASH_LBLOCK];
+ *			int		num;
+ *			...
+ *			} HASH_CTX;
+ * HASH_UPDATE
+ *	name of "Update" function, implemented here.
+ * HASH_TRANSFORM
+ *	name of "Transform" function, implemented here.
+ * HASH_FINAL
+ *	name of "Final" function, implemented here.
+ * HASH_BLOCK_HOST_ORDER
+ *	name of "block" function treating *aligned* input message
+ *	in host byte order, implemented externally.
+ * HASH_BLOCK_DATA_ORDER
+ *	name of "block" function treating *unaligned* input message
+ *	in original (data) byte order, implemented externally (it
+ *	actually is optional if data and host are of the same
+ *	"endianess").
+ * HASH_MAKE_STRING
+ *	macro convering context variables to an ASCII hash string.
+ *
+ * Optional macros:
+ *
+ * B_ENDIAN or L_ENDIAN
+ *	defines host byte-order.
+ * HASH_LONG_LOG2
+ *	defaults to 2 if not states otherwise.
+ * HASH_LBLOCK
+ *	assumed to be HASH_CBLOCK/4 if not stated otherwise.
+ * HASH_BLOCK_DATA_ORDER_ALIGNED
+ *	alternative "block" function capable of treating
+ *	aligned input message in original (data) order,
+ *	implemented externally.
+ *
+ * MD5 example:
+ *
+ *	#define DATA_ORDER_IS_LITTLE_ENDIAN
+ *
+ *	#define HASH_LONG		MD5_LONG
+ *	#define HASH_LONG_LOG2		MD5_LONG_LOG2
+ *	#define HASH_CTX		MD5_CTX
+ *	#define HASH_CBLOCK		MD5_CBLOCK
+ *	#define HASH_LBLOCK		MD5_LBLOCK
+ *	#define HASH_UPDATE		MD5_Update
+ *	#define HASH_TRANSFORM		MD5_Transform
+ *	#define HASH_FINAL		MD5_Final
+ *	#define HASH_BLOCK_HOST_ORDER	md5_block_host_order
+ *	#define HASH_BLOCK_DATA_ORDER	md5_block_data_order
+ *
+ *					<appro@fy.chalmers.se>
+ */
+
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif /* HAVE_CONFIG_H */
+
+#if !defined(DATA_ORDER_IS_BIG_ENDIAN) && !defined(DATA_ORDER_IS_LITTLE_ENDIAN)
+#error "DATA_ORDER must be defined!"
+#endif
+
+#ifndef HASH_CBLOCK
+#error "HASH_CBLOCK must be defined!"
+#endif
+#ifndef HASH_LONG
+#error "HASH_LONG must be defined!"
+#endif
+#ifndef HASH_CTX
+#error "HASH_CTX must be defined!"
+#endif
+
+#ifndef HASH_UPDATE
+#error "HASH_UPDATE must be defined!"
+#endif
+#ifndef HASH_TRANSFORM
+#error "HASH_TRANSFORM must be defined!"
+#endif
+#ifndef HASH_FINAL
+#error "HASH_FINAL must be defined!"
+#endif
+
+#ifndef HASH_BLOCK_HOST_ORDER
+#error "HASH_BLOCK_HOST_ORDER must be defined!"
+#endif
+
+#if 0
+/*
+ * Moved below as it's required only if HASH_BLOCK_DATA_ORDER_ALIGNED
+ * isn't defined.
+ */
+#ifndef HASH_BLOCK_DATA_ORDER
+#error "HASH_BLOCK_DATA_ORDER must be defined!"
+#endif
+#endif
+
+#ifndef HASH_LBLOCK
+#define HASH_LBLOCK	(HASH_CBLOCK/4)
+#endif
+
+#ifndef HASH_LONG_LOG2
+#define HASH_LONG_LOG2	2
+#endif
+
+/*
+ * Engage compiler specific rotate intrinsic function if available.
+ */
+#undef ROTATE
+#ifndef PEDANTIC
+# if 0 /* defined(_MSC_VER) */
+#  define ROTATE(a,n)	_lrotl(a,n)
+# elif defined(__MWERKS__)
+#  if defined(__POWERPC__)
+#   define ROTATE(a,n)	__rlwinm(a,n,0,31)
+#  elif defined(__MC68K__)
+    /* Motorola specific tweak. <appro@fy.chalmers.se> */
+#   define ROTATE(a,n)	( n<24 ? __rol(a,n) : __ror(a,32-n) )
+#  else
+#   define ROTATE(a,n)	__rol(a,n)
+#  endif
+# elif defined(__GNUC__) && __GNUC__>=2 
+  /*
+   * Some GNU C inline assembler templates. Note that these are
+   * rotates by *constant* number of bits! But that's exactly
+   * what we need here...
+   *
+   * 					<appro@fy.chalmers.se>
+   */
+#  if defined(__i386) || defined(__i386__) || defined(__x86_64) || defined(__x86_64__)
+#   define ROTATE(a,n)	({ register unsigned int ret;	\
+				asm (			\
+				"roll %1,%0"		\
+				: "=r"(ret)		\
+				: "I"(n), "0"(a)	\
+				: "cc");		\
+			   ret;				\
+			})
+#  elif defined(__powerpc) || defined(__ppc)
+#   define ROTATE(a,n)	({ register unsigned int ret;	\
+				asm (			\
+				"rlwinm %0,%1,%2,0,31"	\
+				: "=r"(ret)		\
+				: "r"(a), "I"(n));	\
+			   ret;				\
+			})
+#  endif
+# endif
+
+/*
+ * Engage compiler specific "fetch in reverse byte order"
+ * intrinsic function if available.
+ */
+# if defined(__GNUC__) && __GNUC__>=2
+  /* some GNU C inline assembler templates by <appro@fy.chalmers.se> */
+#  if (defined(__i386) || defined(__i386__) || defined(__x86_64) || defined(__x86_64__)) && !defined(I386_ONLY)
+#   define BE_FETCH32(a)	({ register unsigned int l=(a);\
+				asm (			\
+				"bswapl %0"		\
+				: "=r"(l) : "0"(l));	\
+			  l;				\
+			})
+#  elif defined(__powerpc)
+#   define LE_FETCH32(a)	({ register unsigned int l;	\
+				asm (			\
+				"lwbrx %0,0,%1"		\
+				: "=r"(l)		\
+				: "r"(a));		\
+			   l;				\
+			})
+
+#  elif defined(__sparc)
+#  define LE_FETCH32(a)	({ register unsigned int l;		\
+				asm (				\
+				"lda [%1]#ASI_PRIMARY_LITTLE,%0"\
+				: "=r"(l)			\
+				: "r"(a));			\
+			   l;					\
+			})
+#  endif
+# endif
+#endif /* PEDANTIC */
+
+#if HASH_LONG_LOG2==2	/* Engage only if sizeof(HASH_LONG)== 4 */
+/* A nice byte order reversal from Wei Dai <weidai@eskimo.com> */
+#ifdef ROTATE
+/* 5 instructions with rotate instruction, else 9 */
+#define REVERSE_FETCH32(a,l)	(					\
+		l=*(const HASH_LONG *)(a),				\
+		((ROTATE(l,8)&0x00FF00FF)|(ROTATE((l&0x00FF00FF),24)))	\
+				)
+#else
+/* 6 instructions with rotate instruction, else 8 */
+#define REVERSE_FETCH32(a,l)	(				\
+		l=*(const HASH_LONG *)(a),			\
+		l=(((l>>8)&0x00FF00FF)|((l&0x00FF00FF)<<8)),	\
+		ROTATE(l,16)					\
+				)
+/*
+ * Originally the middle line started with l=(((l&0xFF00FF00)>>8)|...
+ * It's rewritten as above for two reasons:
+ *	- RISCs aren't good at long constants and have to explicitely
+ *	  compose 'em with several (well, usually 2) instructions in a
+ *	  register before performing the actual operation and (as you
+ *	  already realized:-) having same constant should inspire the
+ *	  compiler to permanently allocate the only register for it;
+ *	- most modern CPUs have two ALUs, but usually only one has
+ *	  circuitry for shifts:-( this minor tweak inspires compiler
+ *	  to schedule shift instructions in a better way...
+ *
+ *				<appro@fy.chalmers.se>
+ */
+#endif
+#endif
+
+#ifndef ROTATE
+#define ROTATE(a,n)     (((a)<<(n))|(((a)&0xffffffff)>>(32-(n))))
+#endif
+
+/*
+ * Make some obvious choices. E.g., HASH_BLOCK_DATA_ORDER_ALIGNED
+ * and HASH_BLOCK_HOST_ORDER ought to be the same if input data
+ * and host are of the same "endianess". It's possible to mask
+ * this with blank #define HASH_BLOCK_DATA_ORDER though...
+ *
+ *				<appro@fy.chalmers.se>
+ */
+#if defined(B_ENDIAN)
+#  if defined(DATA_ORDER_IS_BIG_ENDIAN)
+#    if !defined(HASH_BLOCK_DATA_ORDER_ALIGNED) && HASH_LONG_LOG2==2
+#      define HASH_BLOCK_DATA_ORDER_ALIGNED	HASH_BLOCK_HOST_ORDER
+#    endif
+#  elif defined(DATA_ORDER_IS_LITTLE_ENDIAN)
+#    ifndef HOST_FETCH32
+#      ifdef LE_FETCH32
+#        define HOST_FETCH32(p,l)	LE_FETCH32(p)
+#      elif defined(REVERSE_FETCH32)
+#        define HOST_FETCH32(p,l)	REVERSE_FETCH32(p,l)
+#      endif
+#    endif
+#  endif
+#elif defined(L_ENDIAN)
+#  if defined(DATA_ORDER_IS_LITTLE_ENDIAN)
+#    if !defined(HASH_BLOCK_DATA_ORDER_ALIGNED) && HASH_LONG_LOG2==2
+#      define HASH_BLOCK_DATA_ORDER_ALIGNED	HASH_BLOCK_HOST_ORDER
+#    endif
+#  elif defined(DATA_ORDER_IS_BIG_ENDIAN)
+#    ifndef HOST_FETCH32
+#      ifdef BE_FETCH32
+#        define HOST_FETCH32(p,l)	BE_FETCH32(p)
+#      elif defined(REVERSE_FETCH32)
+#        define HOST_FETCH32(p,l)	REVERSE_FETCH32(p,l)
+#      endif
+#    endif
+#  endif
+#endif
+
+#if !defined(HASH_BLOCK_DATA_ORDER_ALIGNED)
+#ifndef HASH_BLOCK_DATA_ORDER
+#error "HASH_BLOCK_DATA_ORDER must be defined!"
+#endif
+#endif
+
+#if defined(DATA_ORDER_IS_BIG_ENDIAN)
+
+#define HOST_c2l(c,l)	(l =(((unsigned long)(*((c)++)))<<24),		\
+			 l|=(((unsigned long)(*((c)++)))<<16),		\
+			 l|=(((unsigned long)(*((c)++)))<< 8),		\
+			 l|=(((unsigned long)(*((c)++)))    ),		\
+			 l)
+#define HOST_p_c2l(c,l,n)	{					\
+			switch (n) {					\
+			case 0: l =((unsigned long)(*((c)++)))<<24;	\
+			case 1: l|=((unsigned long)(*((c)++)))<<16;	\
+			case 2: l|=((unsigned long)(*((c)++)))<< 8;	\
+			case 3: l|=((unsigned long)(*((c)++)));		\
+				} }
+#define HOST_p_c2l_p(c,l,sc,len) {					\
+			switch (sc) {					\
+			case 0: l =((unsigned long)(*((c)++)))<<24;	\
+				if (--len == 0) break;			\
+			case 1: l|=((unsigned long)(*((c)++)))<<16;	\
+				if (--len == 0) break;			\
+			case 2: l|=((unsigned long)(*((c)++)))<< 8;	\
+				} }
+/* NOTE the pointer is not incremented at the end of this */
+#define HOST_c2l_p(c,l,n)	{					\
+			l=0; (c)+=n;					\
+			switch (n) {					\
+			case 3: l =((unsigned long)(*(--(c))))<< 8;	\
+			case 2: l|=((unsigned long)(*(--(c))))<<16;	\
+			case 1: l|=((unsigned long)(*(--(c))))<<24;	\
+				} }
+#define HOST_l2c(l,c)	(*((c)++)=(unsigned char)(((l)>>24)&0xff),	\
+			 *((c)++)=(unsigned char)(((l)>>16)&0xff),	\
+			 *((c)++)=(unsigned char)(((l)>> 8)&0xff),	\
+			 *((c)++)=(unsigned char)(((l)    )&0xff),	\
+			 l)
+
+#elif defined(DATA_ORDER_IS_LITTLE_ENDIAN)
+
+#define HOST_c2l(c,l)	(l =(((unsigned long)(*((c)++)))    ),		\
+			 l|=(((unsigned long)(*((c)++)))<< 8),		\
+			 l|=(((unsigned long)(*((c)++)))<<16),		\
+			 l|=(((unsigned long)(*((c)++)))<<24),		\
+			 l)
+#define HOST_p_c2l(c,l,n)	{					\
+			switch (n) {					\
+			case 0: l =((unsigned long)(*((c)++)));		\
+			case 1: l|=((unsigned long)(*((c)++)))<< 8;	\
+			case 2: l|=((unsigned long)(*((c)++)))<<16;	\
+			case 3: l|=((unsigned long)(*((c)++)))<<24;	\
+				} }
+#define HOST_p_c2l_p(c,l,sc,len) {					\
+			switch (sc) {					\
+			case 0: l =((unsigned long)(*((c)++)));		\
+				if (--len == 0) break;			\
+			case 1: l|=((unsigned long)(*((c)++)))<< 8;	\
+				if (--len == 0) break;			\
+			case 2: l|=((unsigned long)(*((c)++)))<<16;	\
+				} }
+/* NOTE the pointer is not incremented at the end of this */
+#define HOST_c2l_p(c,l,n)	{					\
+			l=0; (c)+=n;					\
+			switch (n) {					\
+			case 3: l =((unsigned long)(*(--(c))))<<16;	\
+			case 2: l|=((unsigned long)(*(--(c))))<< 8;	\
+			case 1: l|=((unsigned long)(*(--(c))));		\
+				} }
+#define HOST_l2c(l,c)	(*((c)++)=(unsigned char)(((l)    )&0xff),	\
+			 *((c)++)=(unsigned char)(((l)>> 8)&0xff),	\
+			 *((c)++)=(unsigned char)(((l)>>16)&0xff),	\
+			 *((c)++)=(unsigned char)(((l)>>24)&0xff),	\
+			 l)
+
+#endif
+
+/*
+ * Time for some action:-)
+ */
+
+int HASH_UPDATE (HASH_CTX *c, const void *data_, unsigned long len)
+	{
+	const unsigned char *data = (const unsigned char *) data_;
+	register HASH_LONG * p;
+	register unsigned long l;
+	int sw,sc,ew,ec;
+
+	if (len==0) return 1;
+
+	l=(c->Nl+(len<<3))&0xffffffffL;
+	/* 95-05-24 eay Fixed a bug with the overflow handling, thanks to
+	 * Wei Dai <weidai@eskimo.com> for pointing it out. */
+	if (l < c->Nl) /* overflow */
+		c->Nh++;
+	c->Nh+=(len>>29);
+	c->Nl=l;
+
+	if (c->num != 0)
+		{
+		p=c->data;
+		sw=c->num>>2;
+		sc=c->num&0x03;
+
+		if ((c->num+len) >= HASH_CBLOCK)
+			{
+			l=p[sw]; HOST_p_c2l(data,l,sc); p[sw++]=l;
+			for (; sw<HASH_LBLOCK; sw++)
+				{
+				HOST_c2l(data,l); p[sw]=l;
+				}
+			HASH_BLOCK_HOST_ORDER (c,p,1);
+			len-=(HASH_CBLOCK-c->num);
+			c->num=0;
+			/* drop through and do the rest */
+			}
+		else
+			{
+			c->num+=len;
+			if ((sc+len) < 4) /* ugly, add char's to a word */
+				{
+				l=p[sw]; HOST_p_c2l_p(data,l,sc,len); p[sw]=l;
+				}
+			else
+				{
+				ew=(c->num>>2);
+				ec=(c->num&0x03);
+				if (sc)
+					l=p[sw];
+				HOST_p_c2l(data,l,sc);
+				p[sw++]=l;
+				for (; sw < ew; sw++)
+					{
+					HOST_c2l(data,l); p[sw]=l;
+					}
+				if (ec)
+					{
+					HOST_c2l_p(data,l,ec); p[sw]=l;
+					}
+				}
+			return 1;
+			}
+		}
+
+	sw=len/HASH_CBLOCK;
+	if (sw > 0)
+		{
+#if defined(HASH_BLOCK_DATA_ORDER_ALIGNED)
+		/*
+		 * Note that HASH_BLOCK_DATA_ORDER_ALIGNED gets defined
+		 * only if sizeof(HASH_LONG)==4.
+		 */
+		if ((((unsigned long)data)%4) == 0)
+			{
+			/* data is properly aligned so that we can cast it: */
+			HASH_BLOCK_DATA_ORDER_ALIGNED (c,(HASH_LONG *)data,sw);
+			sw*=HASH_CBLOCK;
+			data+=sw;
+			len-=sw;
+			}
+		else
+#if !defined(HASH_BLOCK_DATA_ORDER)
+			while (sw--)
+				{
+				memcpy (p=c->data,data,HASH_CBLOCK);
+				HASH_BLOCK_DATA_ORDER_ALIGNED(c,p,1);
+				data+=HASH_CBLOCK;
+				len-=HASH_CBLOCK;
+				}
+#endif
+#endif
+#if defined(HASH_BLOCK_DATA_ORDER)
+			{
+			HASH_BLOCK_DATA_ORDER(c,data,sw);
+			sw*=HASH_CBLOCK;
+			data+=sw;
+			len-=sw;
+			}
+#endif
+		}
+
+	if (len!=0)
+		{
+		p = c->data;
+		c->num = len;
+		ew=len>>2;	/* words to copy */
+		ec=len&0x03;
+		for (; ew; ew--,p++)
+			{
+			HOST_c2l(data,l); *p=l;
+			}
+		HOST_c2l_p(data,l,ec);
+		*p=l;
+		}
+	return 1;
+	}
+
+
+void HASH_TRANSFORM (HASH_CTX *c, const unsigned char *data)
+	{
+#if defined(HASH_BLOCK_DATA_ORDER_ALIGNED)
+	if ((((unsigned long)data)%4) == 0)
+		/* data is properly aligned so that we can cast it: */
+		HASH_BLOCK_DATA_ORDER_ALIGNED (c,(HASH_LONG *)data,1);
+	else
+#if !defined(HASH_BLOCK_DATA_ORDER)
+		{
+		memcpy (c->data,data,HASH_CBLOCK);
+		HASH_BLOCK_DATA_ORDER_ALIGNED (c,c->data,1);
+		}
+#endif
+#endif
+#if defined(HASH_BLOCK_DATA_ORDER)
+	HASH_BLOCK_DATA_ORDER (c,data,1);
+#endif
+	}
+
+
+int HASH_FINAL (unsigned char *md, HASH_CTX *c)
+	{
+	register HASH_LONG *p;
+	register unsigned long l;
+	register int i,j;
+	static const unsigned char end[4]={0x80,0x00,0x00,0x00};
+	const unsigned char *cp=end;
+
+	/* c->num should definitly have room for at least one more byte. */
+	p=c->data;
+	i=c->num>>2;
+	j=c->num&0x03;
+
+#if 0
+	/* purify often complains about the following line as an
+	 * Uninitialized Memory Read.  While this can be true, the
+	 * following p_c2l macro will reset l when that case is true.
+	 * This is because j&0x03 contains the number of 'valid' bytes
+	 * already in p[i].  If and only if j&0x03 == 0, the UMR will
+	 * occur but this is also the only time p_c2l will do
+	 * l= *(cp++) instead of l|= *(cp++)
+	 * Many thanks to Alex Tang <altitude@cic.net> for pickup this
+	 * 'potential bug' */
+#ifdef PURIFY
+	if (j==0) p[i]=0; /* Yeah, but that's not the way to fix it:-) */
+#endif
+	l=p[i];
+#else
+	l = (j==0) ? 0 : p[i];
+#endif
+	HOST_p_c2l(cp,l,j); p[i++]=l; /* i is the next 'undefined word' */
+
+	if (i>(HASH_LBLOCK-2)) /* save room for Nl and Nh */
+		{
+		if (i<HASH_LBLOCK) p[i]=0;
+		HASH_BLOCK_HOST_ORDER (c,p,1);
+		i=0;
+		}
+	for (; i<(HASH_LBLOCK-2); i++)
+		p[i]=0;
+
+#if   defined(DATA_ORDER_IS_BIG_ENDIAN)
+	p[HASH_LBLOCK-2]=c->Nh;
+	p[HASH_LBLOCK-1]=c->Nl;
+#elif defined(DATA_ORDER_IS_LITTLE_ENDIAN)
+	p[HASH_LBLOCK-2]=c->Nl;
+	p[HASH_LBLOCK-1]=c->Nh;
+#endif
+	HASH_BLOCK_HOST_ORDER (c,p,1);
+
+#ifndef HASH_MAKE_STRING
+#error "HASH_MAKE_STRING must be defined!"
+#else
+	HASH_MAKE_STRING(c,md);
+#endif
+
+	c->num=0;
+	/* clear stuff, HASH_BLOCK may be leaving some stuff on the stack
+	 * but I'm not worried :-)
+	OPENSSL_cleanse((void *)c,sizeof(HASH_CTX));
+	 */
+	return 1;
+	}
+
+#ifndef MD32_REG_T
+#define MD32_REG_T long
+/*
+ * This comment was originaly written for MD5, which is why it
+ * discusses A-D. But it basically applies to all 32-bit digests,
+ * which is why it was moved to common header file.
+ *
+ * In case you wonder why A-D are declared as long and not
+ * as MD5_LONG. Doing so results in slight performance
+ * boost on LP64 architectures. The catch is we don't
+ * really care if 32 MSBs of a 64-bit register get polluted
+ * with eventual overflows as we *save* only 32 LSBs in
+ * *either* case. Now declaring 'em long excuses the compiler
+ * from keeping 32 MSBs zeroed resulting in 13% performance
+ * improvement under SPARC Solaris7/64 and 5% under AlphaLinux.
+ * Well, to be honest it should say that this *prevents* 
+ * performance degradation.
+ *				<appro@fy.chalmers.se>
+ * Apparently there're LP64 compilers that generate better
+ * code if A-D are declared int. Most notably GCC-x86_64
+ * generates better code.
+ *				<appro@fy.chalmers.se>
+ */
+#endif

src/crypto/sha.c

@@ -3,486 +3,9 @@
 
 
 
-#include "sha.h"
+#define SHA_1 1
+
+#include "sha_locl.h"
 #include "cleanse.h"
 #include <stdlib.h>
 
-#define Xupdate(a,ix,ia,ib,ic,id)     ( (a)=(ia^ib^ic^id),    \
-                                        ix=(a)=ROTATE((a),1)  \
-                                      )
-#define K_00_19 0x5a827999UL
-#define K_20_39 0x6ed9eba1UL
-#define K_40_59 0x8f1bbcdcUL
-#define K_60_79 0xca62c1d6UL
-
-#define F_00_19(b,c,d)  ((((c) ^ (d)) & (b)) ^ (d))
-#define F_20_39(b,c,d)  ((b) ^ (c) ^ (d))
-#define F_40_59(b,c,d)  (((b) & (c)) | (((b)|(c)) & (d)))
-#define F_60_79(b,c,d)  F_20_39(b,c,d)
-
-#define BODY_00_15(i,a,b,c,d,e,f,xi) \
-        (f)=xi+(e)+K_00_19+ROTATE((a),5)+F_00_19((b),(c),(d)); \
-        (b)=ROTATE((b),30);
-
-#define BODY_16_19(i,a,b,c,d,e,f,xi,xa,xb,xc,xd) \
-        Xupdate(f,xi,xa,xb,xc,xd); \
-        (f)+=(e)+K_00_19+ROTATE((a),5)+F_00_19((b),(c),(d)); \
-        (b)=ROTATE((b),30);
-
-#define BODY_20_31(i,a,b,c,d,e,f,xi,xa,xb,xc,xd) \
-        Xupdate(f,xi,xa,xb,xc,xd); \
-        (f)+=(e)+K_20_39+ROTATE((a),5)+F_20_39((b),(c),(d)); \
-        (b)=ROTATE((b),30);
-
-#define BODY_32_39(i,a,b,c,d,e,f,xa,xb,xc,xd) \
-        Xupdate(f,xa,xa,xb,xc,xd); \
-        (f)+=(e)+K_20_39+ROTATE((a),5)+F_20_39((b),(c),(d)); \
-        (b)=ROTATE((b),30);
-
-#define BODY_40_59(i,a,b,c,d,e,f,xa,xb,xc,xd) \
-        Xupdate(f,xa,xa,xb,xc,xd); \
-        (f)+=(e)+K_40_59+ROTATE((a),5)+F_40_59((b),(c),(d)); \
-        (b)=ROTATE((b),30);
-
-#define BODY_60_79(i,a,b,c,d,e,f,xa,xb,xc,xd) \
-        Xupdate(f,xa,xa,xb,xc,xd); \
-        (f)=xa+(e)+K_60_79+ROTATE((a),5)+F_60_79((b),(c),(d)); \
-        (b)=ROTATE((b),30);
-
-
-#define ROTATE(a,n)  ({ register unsigned int ret;   \
-                                __asm__ (               \
-                                "roll %1,%0"            \
-                                : "=r"(ret)             \
-                                : "I"(n), "0"(a)        \
-                                : "cc");                \
-                           ret;                         \
-                        })
-
-#define BE_FETCH32(a)        ({ register unsigned int l=(a);\
-                                __asm__ (               \
-                                "bswapl %0"             \
-                                : "=r"(l) : "0"(l));    \
-                          l;                            \
-                        })
-
-
-#define REVERSE_FETCH32(a,l)    (                                       \
-                l=*(const HASH_LONG *)(a),                              \
-                ((ROTATE(l,8)&0x00FF00FF)|(ROTATE((l&0x00FF00FF),24)))  \
-
-
-#define X(i)   XX##i
-
-void sha1_block_host_order (SHA_CTX *c, const void *d, int num)
-        {
-        const SHA_LONG *W = (const SHA_LONG *) d;
-        register unsigned long A,B,C,D,E,T;
-/* #ifndef MD32_XARRAY */
-        unsigned long     XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
-                                XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
-/*#else
-        SHA_LONG        XX[16];
-#endif*/
-
-        A=c->h0;
-        B=c->h1;
-        C=c->h2;
-        D=c->h3;
-        E=c->h4;
-
-        for (;;)
-                {
-        BODY_00_15( 0,A,B,C,D,E,T,W[ 0]);
-        BODY_00_15( 1,T,A,B,C,D,E,W[ 1]);
-        BODY_00_15( 2,E,T,A,B,C,D,W[ 2]);
-        BODY_00_15( 3,D,E,T,A,B,C,W[ 3]);
-        BODY_00_15( 4,C,D,E,T,A,B,W[ 4]);
-        BODY_00_15( 5,B,C,D,E,T,A,W[ 5]);
-        BODY_00_15( 6,A,B,C,D,E,T,W[ 6]);
-        BODY_00_15( 7,T,A,B,C,D,E,W[ 7]);
-        BODY_00_15( 8,E,T,A,B,C,D,W[ 8]);
-        BODY_00_15( 9,D,E,T,A,B,C,W[ 9]);
-        BODY_00_15(10,C,D,E,T,A,B,W[10]);
-        BODY_00_15(11,B,C,D,E,T,A,W[11]);
-        BODY_00_15(12,A,B,C,D,E,T,W[12]);
-        BODY_00_15(13,T,A,B,C,D,E,W[13]);
-        BODY_00_15(14,E,T,A,B,C,D,W[14]);
-        BODY_00_15(15,D,E,T,A,B,C,W[15]);
-
-        BODY_16_19(16,C,D,E,T,A,B,X( 0),W[ 0],W[ 2],W[ 8],W[13]);
-        BODY_16_19(17,B,C,D,E,T,A,X( 1),W[ 1],W[ 3],W[ 9],W[14]);
-        BODY_16_19(18,A,B,C,D,E,T,X( 2),W[ 2],W[ 4],W[10],W[15]);
-        BODY_16_19(19,T,A,B,C,D,E,X( 3),W[ 3],W[ 5],W[11],X( 0));
-
-        BODY_20_31(20,E,T,A,B,C,D,X( 4),W[ 4],W[ 6],W[12],X( 1));
-        BODY_20_31(21,D,E,T,A,B,C,X( 5),W[ 5],W[ 7],W[13],X( 2));
-        BODY_20_31(22,C,D,E,T,A,B,X( 6),W[ 6],W[ 8],W[14],X( 3));
-        BODY_20_31(23,B,C,D,E,T,A,X( 7),W[ 7],W[ 9],W[15],X( 4));
-        BODY_20_31(24,A,B,C,D,E,T,X( 8),W[ 8],W[10],X( 0),X( 5));
-        BODY_20_31(25,T,A,B,C,D,E,X( 9),W[ 9],W[11],X( 1),X( 6));
-        BODY_20_31(26,E,T,A,B,C,D,X(10),W[10],W[12],X( 2),X( 7));
-        BODY_20_31(27,D,E,T,A,B,C,X(11),W[11],W[13],X( 3),X( 8));
-        BODY_20_31(28,C,D,E,T,A,B,X(12),W[12],W[14],X( 4),X( 9));
-        BODY_20_31(29,B,C,D,E,T,A,X(13),W[13],W[15],X( 5),X(10));
-        BODY_20_31(30,A,B,C,D,E,T,X(14),W[14],X( 0),X( 6),X(11));
-        BODY_20_31(31,T,A,B,C,D,E,X(15),W[15],X( 1),X( 7),X(12));
-
-        BODY_32_39(32,E,T,A,B,C,D,X( 0),X( 2),X( 8),X(13));
-        BODY_32_39(33,D,E,T,A,B,C,X( 1),X( 3),X( 9),X(14));
-        BODY_32_39(34,C,D,E,T,A,B,X( 2),X( 4),X(10),X(15));
-        BODY_32_39(35,B,C,D,E,T,A,X( 3),X( 5),X(11),X( 0));
-        BODY_32_39(36,A,B,C,D,E,T,X( 4),X( 6),X(12),X( 1));
-        BODY_32_39(37,T,A,B,C,D,E,X( 5),X( 7),X(13),X( 2));
-        BODY_32_39(38,E,T,A,B,C,D,X( 6),X( 8),X(14),X( 3));
-        BODY_32_39(39,D,E,T,A,B,C,X( 7),X( 9),X(15),X( 4));
-
-        BODY_40_59(40,C,D,E,T,A,B,X( 8),X(10),X( 0),X( 5));
-        BODY_40_59(41,B,C,D,E,T,A,X( 9),X(11),X( 1),X( 6));
-        BODY_40_59(42,A,B,C,D,E,T,X(10),X(12),X( 2),X( 7));
-        BODY_40_59(43,T,A,B,C,D,E,X(11),X(13),X( 3),X( 8));
-        BODY_40_59(44,E,T,A,B,C,D,X(12),X(14),X( 4),X( 9));
-        BODY_40_59(45,D,E,T,A,B,C,X(13),X(15),X( 5),X(10));
-        BODY_40_59(46,C,D,E,T,A,B,X(14),X( 0),X( 6),X(11));
-        BODY_40_59(47,B,C,D,E,T,A,X(15),X( 1),X( 7),X(12));
-        BODY_40_59(48,A,B,C,D,E,T,X( 0),X( 2),X( 8),X(13));
-        BODY_40_59(49,T,A,B,C,D,E,X( 1),X( 3),X( 9),X(14));
-        BODY_40_59(50,E,T,A,B,C,D,X( 2),X( 4),X(10),X(15));
-        BODY_40_59(51,D,E,T,A,B,C,X( 3),X( 5),X(11),X( 0));
-        BODY_40_59(52,C,D,E,T,A,B,X( 4),X( 6),X(12),X( 1));
-        BODY_40_59(53,B,C,D,E,T,A,X( 5),X( 7),X(13),X( 2));
-        BODY_40_59(54,A,B,C,D,E,T,X( 6),X( 8),X(14),X( 3));
-        BODY_40_59(55,T,A,B,C,D,E,X( 7),X( 9),X(15),X( 4));
-        BODY_40_59(56,E,T,A,B,C,D,X( 8),X(10),X( 0),X( 5));
-        BODY_40_59(57,D,E,T,A,B,C,X( 9),X(11),X( 1),X( 6));
-        BODY_40_59(58,C,D,E,T,A,B,X(10),X(12),X( 2),X( 7));
-        BODY_40_59(59,B,C,D,E,T,A,X(11),X(13),X( 3),X( 8));
-
-        BODY_60_79(60,A,B,C,D,E,T,X(12),X(14),X( 4),X( 9));
-        BODY_60_79(61,T,A,B,C,D,E,X(13),X(15),X( 5),X(10));
-        BODY_60_79(62,E,T,A,B,C,D,X(14),X( 0),X( 6),X(11));
-        BODY_60_79(63,D,E,T,A,B,C,X(15),X( 1),X( 7),X(12));
-        BODY_60_79(64,C,D,E,T,A,B,X( 0),X( 2),X( 8),X(13));
-        BODY_60_79(65,B,C,D,E,T,A,X( 1),X( 3),X( 9),X(14));
-        BODY_60_79(66,A,B,C,D,E,T,X( 2),X( 4),X(10),X(15));
-        BODY_60_79(67,T,A,B,C,D,E,X( 3),X( 5),X(11),X( 0));
-        BODY_60_79(68,E,T,A,B,C,D,X( 4),X( 6),X(12),X( 1));
-        BODY_60_79(69,D,E,T,A,B,C,X( 5),X( 7),X(13),X( 2));
-        BODY_60_79(70,C,D,E,T,A,B,X( 6),X( 8),X(14),X( 3));
-
-        BODY_60_79(71,B,C,D,E,T,A,X( 7),X( 9),X(15),X( 4));
-        BODY_60_79(72,A,B,C,D,E,T,X( 8),X(10),X( 0),X( 5));
-        BODY_60_79(73,T,A,B,C,D,E,X( 9),X(11),X( 1),X( 6));
-        BODY_60_79(74,E,T,A,B,C,D,X(10),X(12),X( 2),X( 7));
-        BODY_60_79(75,D,E,T,A,B,C,X(11),X(13),X( 3),X( 8));
-        BODY_60_79(76,C,D,E,T,A,B,X(12),X(14),X( 4),X( 9));
-        BODY_60_79(77,B,C,D,E,T,A,X(13),X(15),X( 5),X(10));
-        BODY_60_79(78,A,B,C,D,E,T,X(14),X( 0),X( 6),X(11));
-        BODY_60_79(79,T,A,B,C,D,E,X(15),X( 1),X( 7),X(12));
-
-        c->h0=(c->h0+E)&0xffffffffL;
-        c->h1=(c->h1+T)&0xffffffffL;
-        c->h2=(c->h2+A)&0xffffffffL;
-        c->h3=(c->h3+B)&0xffffffffL;
-        c->h4=(c->h4+C)&0xffffffffL;
-
-        if (--num <= 0) break;
-
-        A=c->h0;
-        B=c->h1;
-        C=c->h2;
-        D=c->h3;
-        E=c->h4;
-
-        W+=SHA_LBLOCK;
-                }
-        }
-
-
-void sha1_block_data_order (SHA_CTX *c, const void *p, int num)
-        {
-        const unsigned char *data = (const unsigned char *) p;
-        register unsigned long A,B,C,D,E,T,l;
-/* #ifndef MD32_XARRAY */
-        unsigned long     XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
-                                XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
-/*#else
-        SHA_LONG        XX[16];
-#endif*/
-
-        A=c->h0;
-        B=c->h1;
-        C=c->h2;
-        D=c->h3;
-        E=c->h4;
-
-        for (;;)
-                {
-
-        HOST_c2l(data,l); X( 0)=l;              HOST_c2l(data,l); X( 1)=l;
-        BODY_00_15( 0,A,B,C,D,E,T,X( 0));       HOST_c2l(data,l); X( 2)=l;
-        BODY_00_15( 1,T,A,B,C,D,E,X( 1));       HOST_c2l(data,l); X( 3)=l;
-        BODY_00_15( 2,E,T,A,B,C,D,X( 2));       HOST_c2l(data,l); X( 4)=l;
-        BODY_00_15( 3,D,E,T,A,B,C,X( 3));       HOST_c2l(data,l); X( 5)=l;
-        BODY_00_15( 4,C,D,E,T,A,B,X( 4));       HOST_c2l(data,l); X( 6)=l;
-        BODY_00_15( 5,B,C,D,E,T,A,X( 5));       HOST_c2l(data,l); X( 7)=l;
-        BODY_00_15( 6,A,B,C,D,E,T,X( 6));       HOST_c2l(data,l); X( 8)=l;
-        BODY_00_15( 7,T,A,B,C,D,E,X( 7));       HOST_c2l(data,l); X( 9)=l;
-        BODY_00_15( 8,E,T,A,B,C,D,X( 8));       HOST_c2l(data,l); X(10)=l;
-        BODY_00_15( 9,D,E,T,A,B,C,X( 9));       HOST_c2l(data,l); X(11)=l;
-        BODY_00_15(10,C,D,E,T,A,B,X(10));       HOST_c2l(data,l); X(12)=l;
-        BODY_00_15(11,B,C,D,E,T,A,X(11));       HOST_c2l(data,l); X(13)=l;
-        BODY_00_15(12,A,B,C,D,E,T,X(12));       HOST_c2l(data,l); X(14)=l;
-        BODY_00_15(13,T,A,B,C,D,E,X(13));       HOST_c2l(data,l); X(15)=l;
-        BODY_00_15(14,E,T,A,B,C,D,X(14));
-        BODY_00_15(15,D,E,T,A,B,C,X(15));
-
-        BODY_16_19(16,C,D,E,T,A,B,X( 0),X( 0),X( 2),X( 8),X(13));
-        BODY_16_19(17,B,C,D,E,T,A,X( 1),X( 1),X( 3),X( 9),X(14));
-        BODY_16_19(18,A,B,C,D,E,T,X( 2),X( 2),X( 4),X(10),X(15));
-        BODY_16_19(19,T,A,B,C,D,E,X( 3),X( 3),X( 5),X(11),X( 0));
-
-        BODY_20_31(20,E,T,A,B,C,D,X( 4),X( 4),X( 6),X(12),X( 1));
-        BODY_20_31(21,D,E,T,A,B,C,X( 5),X( 5),X( 7),X(13),X( 2));
-        BODY_20_31(22,C,D,E,T,A,B,X( 6),X( 6),X( 8),X(14),X( 3));
-        BODY_20_31(23,B,C,D,E,T,A,X( 7),X( 7),X( 9),X(15),X( 4));
-        BODY_20_31(24,A,B,C,D,E,T,X( 8),X( 8),X(10),X( 0),X( 5));
-        BODY_20_31(25,T,A,B,C,D,E,X( 9),X( 9),X(11),X( 1),X( 6));
-        BODY_20_31(26,E,T,A,B,C,D,X(10),X(10),X(12),X( 2),X( 7));
-        BODY_20_31(27,D,E,T,A,B,C,X(11),X(11),X(13),X( 3),X( 8));
-        BODY_20_31(28,C,D,E,T,A,B,X(12),X(12),X(14),X( 4),X( 9));
-        BODY_20_31(29,B,C,D,E,T,A,X(13),X(13),X(15),X( 5),X(10));
-        BODY_20_31(30,A,B,C,D,E,T,X(14),X(14),X( 0),X( 6),X(11));
-        BODY_20_31(31,T,A,B,C,D,E,X(15),X(15),X( 1),X( 7),X(12));
-
-        BODY_32_39(32,E,T,A,B,C,D,X( 0),X( 2),X( 8),X(13));
-        BODY_32_39(33,D,E,T,A,B,C,X( 1),X( 3),X( 9),X(14));
-        BODY_32_39(34,C,D,E,T,A,B,X( 2),X( 4),X(10),X(15));
-        BODY_32_39(35,B,C,D,E,T,A,X( 3),X( 5),X(11),X( 0));
-        BODY_32_39(36,A,B,C,D,E,T,X( 4),X( 6),X(12),X( 1));
-        BODY_32_39(37,T,A,B,C,D,E,X( 5),X( 7),X(13),X( 2));
-        BODY_32_39(38,E,T,A,B,C,D,X( 6),X( 8),X(14),X( 3));
-        BODY_32_39(39,D,E,T,A,B,C,X( 7),X( 9),X(15),X( 4));
-
-        BODY_40_59(40,C,D,E,T,A,B,X( 8),X(10),X( 0),X( 5));
-        BODY_40_59(41,B,C,D,E,T,A,X( 9),X(11),X( 1),X( 6));
-        BODY_40_59(42,A,B,C,D,E,T,X(10),X(12),X( 2),X( 7));
-        BODY_40_59(43,T,A,B,C,D,E,X(11),X(13),X( 3),X( 8));
-        BODY_40_59(44,E,T,A,B,C,D,X(12),X(14),X( 4),X( 9));
-        BODY_40_59(45,D,E,T,A,B,C,X(13),X(15),X( 5),X(10));
-        BODY_40_59(46,C,D,E,T,A,B,X(14),X( 0),X( 6),X(11));
-        BODY_40_59(47,B,C,D,E,T,A,X(15),X( 1),X( 7),X(12));
-        BODY_40_59(48,A,B,C,D,E,T,X( 0),X( 2),X( 8),X(13));
-        BODY_40_59(49,T,A,B,C,D,E,X( 1),X( 3),X( 9),X(14));
-        BODY_40_59(50,E,T,A,B,C,D,X( 2),X( 4),X(10),X(15));
-        BODY_40_59(51,D,E,T,A,B,C,X( 3),X( 5),X(11),X( 0));
-        BODY_40_59(52,C,D,E,T,A,B,X( 4),X( 6),X(12),X( 1));
-        BODY_40_59(53,B,C,D,E,T,A,X( 5),X( 7),X(13),X( 2));
-        BODY_40_59(54,A,B,C,D,E,T,X( 6),X( 8),X(14),X( 3));
-        BODY_40_59(55,T,A,B,C,D,E,X( 7),X( 9),X(15),X( 4));
-        BODY_40_59(56,E,T,A,B,C,D,X( 8),X(10),X( 0),X( 5));
-        BODY_40_59(57,D,E,T,A,B,C,X( 9),X(11),X( 1),X( 6));
-        BODY_40_59(58,C,D,E,T,A,B,X(10),X(12),X( 2),X( 7));
-        BODY_40_59(59,B,C,D,E,T,A,X(11),X(13),X( 3),X( 8));
-
-        BODY_60_79(60,A,B,C,D,E,T,X(12),X(14),X( 4),X( 9));
-        BODY_60_79(61,T,A,B,C,D,E,X(13),X(15),X( 5),X(10));
-        BODY_60_79(62,E,T,A,B,C,D,X(14),X( 0),X( 6),X(11));
-        BODY_60_79(63,D,E,T,A,B,C,X(15),X( 1),X( 7),X(12));
-        BODY_60_79(64,C,D,E,T,A,B,X( 0),X( 2),X( 8),X(13));
-        BODY_60_79(65,B,C,D,E,T,A,X( 1),X( 3),X( 9),X(14));
-        BODY_60_79(66,A,B,C,D,E,T,X( 2),X( 4),X(10),X(15));
-        BODY_60_79(67,T,A,B,C,D,E,X( 3),X( 5),X(11),X( 0));
-        BODY_60_79(68,E,T,A,B,C,D,X( 4),X( 6),X(12),X( 1));
-        BODY_60_79(69,D,E,T,A,B,C,X( 5),X( 7),X(13),X( 2));
-        BODY_60_79(70,C,D,E,T,A,B,X( 6),X( 8),X(14),X( 3));
-        BODY_60_79(71,B,C,D,E,T,A,X( 7),X( 9),X(15),X( 4));
-        BODY_60_79(72,A,B,C,D,E,T,X( 8),X(10),X( 0),X( 5));
-        BODY_60_79(73,T,A,B,C,D,E,X( 9),X(11),X( 1),X( 6));
-        BODY_60_79(74,E,T,A,B,C,D,X(10),X(12),X( 2),X( 7));
-        BODY_60_79(75,D,E,T,A,B,C,X(11),X(13),X( 3),X( 8));
-        BODY_60_79(76,C,D,E,T,A,B,X(12),X(14),X( 4),X( 9));
-        BODY_60_79(77,B,C,D,E,T,A,X(13),X(15),X( 5),X(10));
-        BODY_60_79(78,A,B,C,D,E,T,X(14),X( 0),X( 6),X(11));
-        BODY_60_79(79,T,A,B,C,D,E,X(15),X( 1),X( 7),X(12));
-
-        c->h0=(c->h0+E)&0xffffffffL;
-        c->h1=(c->h1+T)&0xffffffffL;
-        c->h2=(c->h2+A)&0xffffffffL;
-        c->h3=(c->h3+B)&0xffffffffL;
-        c->h4=(c->h4+C)&0xffffffffL;
-
-        if (--num <= 0) break;
-
-        A=c->h0;
-        B=c->h1;
-        C=c->h2;
-        D=c->h3;
-        E=c->h4;
-
-                }
-        }
-
-
-int SHA1_Init (SHA_CTX *c)
-        {
-        c->h0=0x67452301UL;
-        c->h1=0xefcdab89UL;
-        c->h2=0x98badcfeUL;
-        c->h3=0x10325476UL;
-        c->h4=0xc3d2e1f0UL;
-        c->Nl=0;
-        c->Nh=0;
-        c->num=0;
-        return 1;
-        }
-
-int SHA1_Update (SHA_CTX *c, const void *data_, unsigned long len)
-        {
-        const unsigned char *data = (const unsigned char *) data_;
-        register SHA_LONG * p;
-        register unsigned long l;
-        int sw,sc,ew,ec;
-
-        if (len==0) return 1;
-
-        l=(c->Nl+(len<<3))&0xffffffffL;
-        /* 95-05-24 eay Fixed a bug with the overflow handling, thanks to
-         * Wei Dai <weidai@eskimo.com> for pointing it out. */
-        if (l < c->Nl) /* overflow */
-                c->Nh++;
-        c->Nh+=(len>>29);
-        c->Nl=l;
-
-        if (c->num != 0)
-                {
-                p=c->data;
-                sw=c->num>>2;
-                sc=c->num&0x03;
-
-                if ((c->num+len) >= SHA_CBLOCK)
-                        {
-                        l=p[sw]; HOST_p_c2l(data,l,sc); p[sw++]=l;
-                        for (; sw<SHA_LBLOCK; sw++)
-                                {
-                                HOST_c2l(data,l); p[sw]=l;
-                                }
-                        sha1_block_host_order (c,p,1);
-                        len-=(SHA_CBLOCK-c->num);
-                        c->num=0;
-                        /* drop through and do the rest */
-                        }
-                else
-                        {
-                        c->num+=len;
-                        if ((sc+len) < 4) /* ugly, add char's to a word */
-                                {
-                                l=p[sw]; HOST_p_c2l_p(data,l,sc,len); p[sw]=l;
-                                }
-                        else
-                                {
-                                ew=(c->num>>2);
-                                ec=(c->num&0x03);
-                                if (sc)
-                                        l=p[sw];
-                                HOST_p_c2l(data,l,sc);
-                                p[sw++]=l;
-                                for (; sw < ew; sw++)
-                                        {
-                                        HOST_c2l(data,l); p[sw]=l;
-                                        }
-                                if (ec)
-                                        {
-                                        HOST_c2l_p(data,l,ec); p[sw]=l;
-                                        }
-                                }
-                        return 1;
-                        }
-                }
-
-        sw=len/SHA_CBLOCK;
-        if (sw > 0)
-                {
-                /*
-                 * Note that HASH_BLOCK_DATA_ORDER_ALIGNED gets defined
-                 * only if sizeof(HASH_LONG)==4.
-                 */
-                if ((((unsigned long)data)%4) == 0)
-                        {
-                        /* data is properly aligned so that we can cast it: */
-                        sha1_block_host_order (c,(SHA_LONG *)data,sw);
-                        sw*=SHA_CBLOCK;
-                        data+=sw;
-                        len-=sw;
-                        }
-                else
-                  {
-                  sha1_block_data_order(c,data,sw);
-                  sw*=SHA_CBLOCK;
-                  data+=sw;
-                  len-=sw;
-                  }
-                }
-
-        if (len!=0)
-                {
-                p = c->data;
-                c->num = len;
-                ew=len>>2;      /* words to copy */
-                ec=len&0x03;
-                for (; ew; ew--,p++)
-                        {
-                        HOST_c2l(data,l); *p=l;
-                        }
-                HOST_c2l_p(data,l,ec);
-                *p=l;
-                }
-        return 1;
-        }
-
-int SHA1_Final (unsigned char *md, SHA_CTX *c)
-        {
-        register SHA_LONG *p;
-        register unsigned long l;
-        register int i,j;
-        static const unsigned char end[4]={0x80,0x00,0x00,0x00};
-        const unsigned char *cp=end;
-
-        /* c->num should definitly have room for at least one more byte. */
-        p=c->data;
-        i=c->num>>2;
-        j=c->num&0x03;
-
-        l = (j==0) ? 0 : p[i];
-
-        HOST_p_c2l(cp,l,j); p[i++]=l; /* i is the next 'undefined word' */
-
-        if (i>(SHA_LBLOCK-2)) /* save room for Nl and Nh */
-                {
-                if (i<SHA_LBLOCK) p[i]=0;
-                sha1_block_host_order (c,p,1);
-                i=0;
-                }
-        for (; i<(SHA_LBLOCK-2); i++)
-                p[i]=0;
-
-        p[SHA_LBLOCK-2]=c->Nh;
-        p[SHA_LBLOCK-1]=c->Nl;
-        sha1_block_host_order (c,p,1);
-
-        SHA_MAKE_STRING(c,md);
-
-        c->num=0;
-        /* clear stuff, HASH_BLOCK may be leaving some stuff on the stack
-         * but I'm not worried :-)
-         */
-        OPENSSL_cleanse((void *)c,sizeof(SHA_CTX));
-        return 1;
-        }
-
-
-void SHA1_Transform (SHA_CTX *c, const unsigned char *data)
-        {
-        sha1_block_data_order (c,data,1);
-        }
-

src/crypto/sha.h

@@ -1,7 +1,14 @@
 #ifndef _SHA_H
 #define _SHA_H
 
+#if defined(OPENSSL_SYS_WIN16) || defined(__LP32__)
+#define SHA_LONG unsigned long
+#elif defined(OPENSSL_SYS_CRAY) || defined(__ILP64__)
+#define SHA_LONG unsigned long
+#define SHA_LONG_LOG2 3
+#else
 #define SHA_LONG unsigned int
+#endif
 
 #define SHA_LBLOCK      16
 #define SHA_CBLOCK      (SHA_LBLOCK*4)  /* SHA treats input data as a
@@ -11,50 +18,6 @@
 #define SHA_DIGEST_LENGTH 20
 
 
-#define HOST_c2l(c,l)   (l =(((unsigned long)(*((c)++)))<<24),          \
-                         l|=(((unsigned long)(*((c)++)))<<16),          \
-                         l|=(((unsigned long)(*((c)++)))<< 8),          \
-                         l|=(((unsigned long)(*((c)++)))    ),          \
-                         l)
-#define HOST_p_c2l(c,l,n)       {                                       \
-                        switch (n) {                                    \
-                        case 0: l =((unsigned long)(*((c)++)))<<24;     \
-                        case 1: l|=((unsigned long)(*((c)++)))<<16;     \
-                        case 2: l|=((unsigned long)(*((c)++)))<< 8;     \
-                        case 3: l|=((unsigned long)(*((c)++)));         \
-                                } }
-#define HOST_p_c2l_p(c,l,sc,len) {                                      \
-                        switch (sc) {                                   \
-                        case 0: l =((unsigned long)(*((c)++)))<<24;     \
-                                if (--len == 0) break;                  \
-                        case 1: l|=((unsigned long)(*((c)++)))<<16;     \
-                                if (--len == 0) break;                  \
-                        case 2: l|=((unsigned long)(*((c)++)))<< 8;     \
-                                } }
-/* NOTE the pointer is not incremented at the end of this */
-#define HOST_c2l_p(c,l,n)       {                                       \
-                        l=0; (c)+=n;                                    \
-                        switch (n) {                                    \
-                        case 3: l =((unsigned long)(*(--(c))))<< 8;     \
-                        case 2: l|=((unsigned long)(*(--(c))))<<16;     \
-                        case 1: l|=((unsigned long)(*(--(c))))<<24;     \
-                                } }
-#define HOST_l2c(l,c)   (*((c)++)=(unsigned char)(((l)>>24)&0xff),      \
-                         *((c)++)=(unsigned char)(((l)>>16)&0xff),      \
-                         *((c)++)=(unsigned char)(((l)>> 8)&0xff),      \
-                         *((c)++)=(unsigned char)(((l)    )&0xff),      \
-                         l)
-
-
-#define SHA_MAKE_STRING(c,s)   do {    \
-        unsigned long ll;               \
-        ll=(c)->h0; HOST_l2c(ll,(s));   \
-        ll=(c)->h1; HOST_l2c(ll,(s));   \
-        ll=(c)->h2; HOST_l2c(ll,(s));   \
-        ll=(c)->h3; HOST_l2c(ll,(s));   \
-        ll=(c)->h4; HOST_l2c(ll,(s));   \
-        } while (0)
-
 typedef struct SHAstate_st
         {
         SHA_LONG h0,h1,h2,h3,h4;

src/crypto/sha_locl.h

@@ -0,0 +1,402 @@
+/* crypto/sha/sha_locl.h */
+
+#include <stdlib.h>
+#include <string.h>
+
+#include "sha.h"
+
+#ifndef SHA_LONG_LOG2
+#define SHA_LONG_LOG2	2	/* default to 32 bits */
+#endif
+
+#define DATA_ORDER_IS_BIG_ENDIAN
+
+#define HASH_LONG               SHA_LONG
+#define HASH_LONG_LOG2          SHA_LONG_LOG2
+#define HASH_CTX                SHA_CTX
+#define HASH_CBLOCK             SHA_CBLOCK
+#define HASH_LBLOCK             SHA_LBLOCK
+#define HASH_MAKE_STRING(c,s)   do {	\
+	unsigned long ll;		\
+	ll=(c)->h0; HOST_l2c(ll,(s));	\
+	ll=(c)->h1; HOST_l2c(ll,(s));	\
+	ll=(c)->h2; HOST_l2c(ll,(s));	\
+	ll=(c)->h3; HOST_l2c(ll,(s));	\
+	ll=(c)->h4; HOST_l2c(ll,(s));	\
+	} while (0)
+
+#if defined(SHA_1)
+
+# define HASH_UPDATE             	SHA1_Update
+# define HASH_TRANSFORM          	SHA1_Transform
+# define HASH_FINAL              	SHA1_Final
+# define HASH_INIT			SHA1_Init
+# define HASH_BLOCK_HOST_ORDER   	sha1_block_host_order
+# define HASH_BLOCK_DATA_ORDER   	sha1_block_data_order
+# if defined(__MWERKS__) && defined(__MC68K__)
+   /* Metrowerks for Motorola fails otherwise:-( <appro@fy.chalmers.se> */
+#  define Xupdate(a,ix,ia,ib,ic,id)	do { (a)=(ia^ib^ic^id);		\
+					     ix=(a)=ROTATE((a),1);	\
+					} while (0)
+# else
+#  define Xupdate(a,ix,ia,ib,ic,id)	( (a)=(ia^ib^ic^id),	\
+					  ix=(a)=ROTATE((a),1)	\
+					)
+# endif
+
+# ifdef SHA1_ASM
+#  if defined(__i386) || defined(__i386__) || defined(_M_IX86) || defined(__INTEL__)
+#   define sha1_block_host_order		sha1_block_asm_host_order
+#   define DONT_IMPLEMENT_BLOCK_HOST_ORDER
+#   define sha1_block_data_order		sha1_block_asm_data_order
+#   define DONT_IMPLEMENT_BLOCK_DATA_ORDER
+#   define HASH_BLOCK_DATA_ORDER_ALIGNED	sha1_block_asm_data_order
+#  endif
+# endif
+  void sha1_block_host_order (SHA_CTX *c, const void *p,int num);
+  void sha1_block_data_order (SHA_CTX *c, const void *p,int num);
+
+#else
+# error "Either SHA_0 or SHA_1 must be defined."
+#endif
+
+#include "md32_common.h"
+
+#define INIT_DATA_h0 0x67452301UL
+#define INIT_DATA_h1 0xefcdab89UL
+#define INIT_DATA_h2 0x98badcfeUL
+#define INIT_DATA_h3 0x10325476UL
+#define INIT_DATA_h4 0xc3d2e1f0UL
+
+int HASH_INIT (SHA_CTX *c)
+	{
+	c->h0=INIT_DATA_h0;
+	c->h1=INIT_DATA_h1;
+	c->h2=INIT_DATA_h2;
+	c->h3=INIT_DATA_h3;
+	c->h4=INIT_DATA_h4;
+	c->Nl=0;
+	c->Nh=0;
+	c->num=0;
+	return 1;
+	}
+
+#define K_00_19	0x5a827999UL
+#define K_20_39 0x6ed9eba1UL
+#define K_40_59 0x8f1bbcdcUL
+#define K_60_79 0xca62c1d6UL
+
+/* As  pointed out by Wei Dai <weidai@eskimo.com>, F() below can be
+ * simplified to the code in F_00_19.  Wei attributes these optimisations
+ * to Peter Gutmann's SHS code, and he attributes it to Rich Schroeppel.
+ * #define F(x,y,z) (((x) & (y))  |  ((~(x)) & (z)))
+ * I've just become aware of another tweak to be made, again from Wei Dai,
+ * in F_40_59, (x&a)|(y&a) -> (x|y)&a
+ */
+#define	F_00_19(b,c,d)	((((c) ^ (d)) & (b)) ^ (d)) 
+#define	F_20_39(b,c,d)	((b) ^ (c) ^ (d))
+#define F_40_59(b,c,d)	(((b) & (c)) | (((b)|(c)) & (d))) 
+#define	F_60_79(b,c,d)	F_20_39(b,c,d)
+
+#define BODY_00_15(i,a,b,c,d,e,f,xi) \
+	(f)=xi+(e)+K_00_19+ROTATE((a),5)+F_00_19((b),(c),(d)); \
+	(b)=ROTATE((b),30);
+
+#define BODY_16_19(i,a,b,c,d,e,f,xi,xa,xb,xc,xd) \
+	Xupdate(f,xi,xa,xb,xc,xd); \
+	(f)+=(e)+K_00_19+ROTATE((a),5)+F_00_19((b),(c),(d)); \
+	(b)=ROTATE((b),30);
+
+#define BODY_20_31(i,a,b,c,d,e,f,xi,xa,xb,xc,xd) \
+	Xupdate(f,xi,xa,xb,xc,xd); \
+	(f)+=(e)+K_20_39+ROTATE((a),5)+F_20_39((b),(c),(d)); \
+	(b)=ROTATE((b),30);
+
+#define BODY_32_39(i,a,b,c,d,e,f,xa,xb,xc,xd) \
+	Xupdate(f,xa,xa,xb,xc,xd); \
+	(f)+=(e)+K_20_39+ROTATE((a),5)+F_20_39((b),(c),(d)); \
+	(b)=ROTATE((b),30);
+
+#define BODY_40_59(i,a,b,c,d,e,f,xa,xb,xc,xd) \
+	Xupdate(f,xa,xa,xb,xc,xd); \
+	(f)+=(e)+K_40_59+ROTATE((a),5)+F_40_59((b),(c),(d)); \
+	(b)=ROTATE((b),30);
+
+#define BODY_60_79(i,a,b,c,d,e,f,xa,xb,xc,xd) \
+	Xupdate(f,xa,xa,xb,xc,xd); \
+	(f)=xa+(e)+K_60_79+ROTATE((a),5)+F_60_79((b),(c),(d)); \
+	(b)=ROTATE((b),30);
+
+#ifdef X
+#undef X
+#endif
+#ifndef MD32_XARRAY
+  /*
+   * Originally X was an array. As it's automatic it's natural
+   * to expect RISC compiler to accomodate at least part of it in
+   * the register bank, isn't it? Unfortunately not all compilers
+   * "find" this expectation reasonable:-( On order to make such
+   * compilers generate better code I replace X[] with a bunch of
+   * X0, X1, etc. See the function body below...
+   *					<appro@fy.chalmers.se>
+   */
+# define X(i)	XX##i
+#else
+  /*
+   * However! Some compilers (most notably HP C) get overwhelmed by
+   * that many local variables so that we have to have the way to
+   * fall down to the original behavior.
+   */
+# define X(i)	XX[i]
+#endif
+
+#ifndef DONT_IMPLEMENT_BLOCK_HOST_ORDER
+void HASH_BLOCK_HOST_ORDER (SHA_CTX *c, const void *d, int num)
+	{
+	const SHA_LONG *W = (SHA_LONG *) d;
+	register unsigned MD32_REG_T A,B,C,D,E,T;
+#ifndef MD32_XARRAY
+	unsigned MD32_REG_T	XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
+				XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
+#else
+	SHA_LONG	XX[16];
+#endif
+
+	A=c->h0;
+	B=c->h1;
+	C=c->h2;
+	D=c->h3;
+	E=c->h4;
+
+	for (;;)
+		{
+	BODY_00_15( 0,A,B,C,D,E,T,W[ 0]);
+	BODY_00_15( 1,T,A,B,C,D,E,W[ 1]);
+	BODY_00_15( 2,E,T,A,B,C,D,W[ 2]);
+	BODY_00_15( 3,D,E,T,A,B,C,W[ 3]);
+	BODY_00_15( 4,C,D,E,T,A,B,W[ 4]);
+	BODY_00_15( 5,B,C,D,E,T,A,W[ 5]);
+	BODY_00_15( 6,A,B,C,D,E,T,W[ 6]);
+	BODY_00_15( 7,T,A,B,C,D,E,W[ 7]);
+	BODY_00_15( 8,E,T,A,B,C,D,W[ 8]);
+	BODY_00_15( 9,D,E,T,A,B,C,W[ 9]);
+	BODY_00_15(10,C,D,E,T,A,B,W[10]);
+	BODY_00_15(11,B,C,D,E,T,A,W[11]);
+	BODY_00_15(12,A,B,C,D,E,T,W[12]);
+	BODY_00_15(13,T,A,B,C,D,E,W[13]);
+	BODY_00_15(14,E,T,A,B,C,D,W[14]);
+	BODY_00_15(15,D,E,T,A,B,C,W[15]);
+
+	BODY_16_19(16,C,D,E,T,A,B,X( 0),W[ 0],W[ 2],W[ 8],W[13]);
+	BODY_16_19(17,B,C,D,E,T,A,X( 1),W[ 1],W[ 3],W[ 9],W[14]);
+	BODY_16_19(18,A,B,C,D,E,T,X( 2),W[ 2],W[ 4],W[10],W[15]);
+	BODY_16_19(19,T,A,B,C,D,E,X( 3),W[ 3],W[ 5],W[11],X( 0));
+
+	BODY_20_31(20,E,T,A,B,C,D,X( 4),W[ 4],W[ 6],W[12],X( 1));
+	BODY_20_31(21,D,E,T,A,B,C,X( 5),W[ 5],W[ 7],W[13],X( 2));
+	BODY_20_31(22,C,D,E,T,A,B,X( 6),W[ 6],W[ 8],W[14],X( 3));
+	BODY_20_31(23,B,C,D,E,T,A,X( 7),W[ 7],W[ 9],W[15],X( 4));
+	BODY_20_31(24,A,B,C,D,E,T,X( 8),W[ 8],W[10],X( 0),X( 5));
+	BODY_20_31(25,T,A,B,C,D,E,X( 9),W[ 9],W[11],X( 1),X( 6));
+	BODY_20_31(26,E,T,A,B,C,D,X(10),W[10],W[12],X( 2),X( 7));
+	BODY_20_31(27,D,E,T,A,B,C,X(11),W[11],W[13],X( 3),X( 8));
+	BODY_20_31(28,C,D,E,T,A,B,X(12),W[12],W[14],X( 4),X( 9));
+	BODY_20_31(29,B,C,D,E,T,A,X(13),W[13],W[15],X( 5),X(10));
+	BODY_20_31(30,A,B,C,D,E,T,X(14),W[14],X( 0),X( 6),X(11));
+	BODY_20_31(31,T,A,B,C,D,E,X(15),W[15],X( 1),X( 7),X(12));
+
+	BODY_32_39(32,E,T,A,B,C,D,X( 0),X( 2),X( 8),X(13));
+	BODY_32_39(33,D,E,T,A,B,C,X( 1),X( 3),X( 9),X(14));
+	BODY_32_39(34,C,D,E,T,A,B,X( 2),X( 4),X(10),X(15));
+	BODY_32_39(35,B,C,D,E,T,A,X( 3),X( 5),X(11),X( 0));
+	BODY_32_39(36,A,B,C,D,E,T,X( 4),X( 6),X(12),X( 1));
+	BODY_32_39(37,T,A,B,C,D,E,X( 5),X( 7),X(13),X( 2));
+	BODY_32_39(38,E,T,A,B,C,D,X( 6),X( 8),X(14),X( 3));
+	BODY_32_39(39,D,E,T,A,B,C,X( 7),X( 9),X(15),X( 4));
+
+	BODY_40_59(40,C,D,E,T,A,B,X( 8),X(10),X( 0),X( 5));
+	BODY_40_59(41,B,C,D,E,T,A,X( 9),X(11),X( 1),X( 6));
+	BODY_40_59(42,A,B,C,D,E,T,X(10),X(12),X( 2),X( 7));
+	BODY_40_59(43,T,A,B,C,D,E,X(11),X(13),X( 3),X( 8));
+	BODY_40_59(44,E,T,A,B,C,D,X(12),X(14),X( 4),X( 9));
+	BODY_40_59(45,D,E,T,A,B,C,X(13),X(15),X( 5),X(10));
+	BODY_40_59(46,C,D,E,T,A,B,X(14),X( 0),X( 6),X(11));
+	BODY_40_59(47,B,C,D,E,T,A,X(15),X( 1),X( 7),X(12));
+	BODY_40_59(48,A,B,C,D,E,T,X( 0),X( 2),X( 8),X(13));
+	BODY_40_59(49,T,A,B,C,D,E,X( 1),X( 3),X( 9),X(14));
+	BODY_40_59(50,E,T,A,B,C,D,X( 2),X( 4),X(10),X(15));
+	BODY_40_59(51,D,E,T,A,B,C,X( 3),X( 5),X(11),X( 0));
+	BODY_40_59(52,C,D,E,T,A,B,X( 4),X( 6),X(12),X( 1));
+	BODY_40_59(53,B,C,D,E,T,A,X( 5),X( 7),X(13),X( 2));
+	BODY_40_59(54,A,B,C,D,E,T,X( 6),X( 8),X(14),X( 3));
+	BODY_40_59(55,T,A,B,C,D,E,X( 7),X( 9),X(15),X( 4));
+	BODY_40_59(56,E,T,A,B,C,D,X( 8),X(10),X( 0),X( 5));
+	BODY_40_59(57,D,E,T,A,B,C,X( 9),X(11),X( 1),X( 6));
+	BODY_40_59(58,C,D,E,T,A,B,X(10),X(12),X( 2),X( 7));
+	BODY_40_59(59,B,C,D,E,T,A,X(11),X(13),X( 3),X( 8));
+
+	BODY_60_79(60,A,B,C,D,E,T,X(12),X(14),X( 4),X( 9));
+	BODY_60_79(61,T,A,B,C,D,E,X(13),X(15),X( 5),X(10));
+	BODY_60_79(62,E,T,A,B,C,D,X(14),X( 0),X( 6),X(11));
+	BODY_60_79(63,D,E,T,A,B,C,X(15),X( 1),X( 7),X(12));
+	BODY_60_79(64,C,D,E,T,A,B,X( 0),X( 2),X( 8),X(13));
+	BODY_60_79(65,B,C,D,E,T,A,X( 1),X( 3),X( 9),X(14));
+	BODY_60_79(66,A,B,C,D,E,T,X( 2),X( 4),X(10),X(15));
+	BODY_60_79(67,T,A,B,C,D,E,X( 3),X( 5),X(11),X( 0));
+	BODY_60_79(68,E,T,A,B,C,D,X( 4),X( 6),X(12),X( 1));
+	BODY_60_79(69,D,E,T,A,B,C,X( 5),X( 7),X(13),X( 2));
+	BODY_60_79(70,C,D,E,T,A,B,X( 6),X( 8),X(14),X( 3));
+	BODY_60_79(71,B,C,D,E,T,A,X( 7),X( 9),X(15),X( 4));
+	BODY_60_79(72,A,B,C,D,E,T,X( 8),X(10),X( 0),X( 5));
+	BODY_60_79(73,T,A,B,C,D,E,X( 9),X(11),X( 1),X( 6));
+	BODY_60_79(74,E,T,A,B,C,D,X(10),X(12),X( 2),X( 7));
+	BODY_60_79(75,D,E,T,A,B,C,X(11),X(13),X( 3),X( 8));
+	BODY_60_79(76,C,D,E,T,A,B,X(12),X(14),X( 4),X( 9));
+	BODY_60_79(77,B,C,D,E,T,A,X(13),X(15),X( 5),X(10));
+	BODY_60_79(78,A,B,C,D,E,T,X(14),X( 0),X( 6),X(11));
+	BODY_60_79(79,T,A,B,C,D,E,X(15),X( 1),X( 7),X(12));
+	
+	c->h0=(c->h0+E)&0xffffffffL; 
+	c->h1=(c->h1+T)&0xffffffffL;
+	c->h2=(c->h2+A)&0xffffffffL;
+	c->h3=(c->h3+B)&0xffffffffL;
+	c->h4=(c->h4+C)&0xffffffffL;
+
+	if (--num <= 0) break;
+
+	A=c->h0;
+	B=c->h1;
+	C=c->h2;
+	D=c->h3;
+	E=c->h4;
+
+	W+=SHA_LBLOCK;
+		}
+	}
+#endif
+
+#ifndef DONT_IMPLEMENT_BLOCK_DATA_ORDER
+void HASH_BLOCK_DATA_ORDER (SHA_CTX *c, const void *p, int num)
+	{
+	const unsigned char *data = (const unsigned char *) p;
+	register unsigned MD32_REG_T A,B,C,D,E,T,l;
+#ifndef MD32_XARRAY
+	unsigned MD32_REG_T	XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
+				XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
+#else
+	SHA_LONG	XX[16];
+#endif
+
+	A=c->h0;
+	B=c->h1;
+	C=c->h2;
+	D=c->h3;
+	E=c->h4;
+
+	for (;;)
+		{
+
+	HOST_c2l(data,l); X( 0)=l;		HOST_c2l(data,l); X( 1)=l;
+	BODY_00_15( 0,A,B,C,D,E,T,X( 0));	HOST_c2l(data,l); X( 2)=l;
+	BODY_00_15( 1,T,A,B,C,D,E,X( 1));	HOST_c2l(data,l); X( 3)=l;
+	BODY_00_15( 2,E,T,A,B,C,D,X( 2));	HOST_c2l(data,l); X( 4)=l;
+	BODY_00_15( 3,D,E,T,A,B,C,X( 3));	HOST_c2l(data,l); X( 5)=l;
+	BODY_00_15( 4,C,D,E,T,A,B,X( 4));	HOST_c2l(data,l); X( 6)=l;
+	BODY_00_15( 5,B,C,D,E,T,A,X( 5));	HOST_c2l(data,l); X( 7)=l;
+	BODY_00_15( 6,A,B,C,D,E,T,X( 6));	HOST_c2l(data,l); X( 8)=l;
+	BODY_00_15( 7,T,A,B,C,D,E,X( 7));	HOST_c2l(data,l); X( 9)=l;
+	BODY_00_15( 8,E,T,A,B,C,D,X( 8));	HOST_c2l(data,l); X(10)=l;
+	BODY_00_15( 9,D,E,T,A,B,C,X( 9));	HOST_c2l(data,l); X(11)=l;
+	BODY_00_15(10,C,D,E,T,A,B,X(10));	HOST_c2l(data,l); X(12)=l;
+	BODY_00_15(11,B,C,D,E,T,A,X(11));	HOST_c2l(data,l); X(13)=l;
+	BODY_00_15(12,A,B,C,D,E,T,X(12));	HOST_c2l(data,l); X(14)=l;
+	BODY_00_15(13,T,A,B,C,D,E,X(13));	HOST_c2l(data,l); X(15)=l;
+	BODY_00_15(14,E,T,A,B,C,D,X(14));
+	BODY_00_15(15,D,E,T,A,B,C,X(15));
+
+	BODY_16_19(16,C,D,E,T,A,B,X( 0),X( 0),X( 2),X( 8),X(13));
+	BODY_16_19(17,B,C,D,E,T,A,X( 1),X( 1),X( 3),X( 9),X(14));
+	BODY_16_19(18,A,B,C,D,E,T,X( 2),X( 2),X( 4),X(10),X(15));
+	BODY_16_19(19,T,A,B,C,D,E,X( 3),X( 3),X( 5),X(11),X( 0));
+
+	BODY_20_31(20,E,T,A,B,C,D,X( 4),X( 4),X( 6),X(12),X( 1));
+	BODY_20_31(21,D,E,T,A,B,C,X( 5),X( 5),X( 7),X(13),X( 2));
+	BODY_20_31(22,C,D,E,T,A,B,X( 6),X( 6),X( 8),X(14),X( 3));
+	BODY_20_31(23,B,C,D,E,T,A,X( 7),X( 7),X( 9),X(15),X( 4));
+	BODY_20_31(24,A,B,C,D,E,T,X( 8),X( 8),X(10),X( 0),X( 5));
+	BODY_20_31(25,T,A,B,C,D,E,X( 9),X( 9),X(11),X( 1),X( 6));
+	BODY_20_31(26,E,T,A,B,C,D,X(10),X(10),X(12),X( 2),X( 7));
+	BODY_20_31(27,D,E,T,A,B,C,X(11),X(11),X(13),X( 3),X( 8));
+	BODY_20_31(28,C,D,E,T,A,B,X(12),X(12),X(14),X( 4),X( 9));
+	BODY_20_31(29,B,C,D,E,T,A,X(13),X(13),X(15),X( 5),X(10));
+	BODY_20_31(30,A,B,C,D,E,T,X(14),X(14),X( 0),X( 6),X(11));
+	BODY_20_31(31,T,A,B,C,D,E,X(15),X(15),X( 1),X( 7),X(12));
+
+	BODY_32_39(32,E,T,A,B,C,D,X( 0),X( 2),X( 8),X(13));
+	BODY_32_39(33,D,E,T,A,B,C,X( 1),X( 3),X( 9),X(14));
+	BODY_32_39(34,C,D,E,T,A,B,X( 2),X( 4),X(10),X(15));
+	BODY_32_39(35,B,C,D,E,T,A,X( 3),X( 5),X(11),X( 0));
+	BODY_32_39(36,A,B,C,D,E,T,X( 4),X( 6),X(12),X( 1));
+	BODY_32_39(37,T,A,B,C,D,E,X( 5),X( 7),X(13),X( 2));
+	BODY_32_39(38,E,T,A,B,C,D,X( 6),X( 8),X(14),X( 3));
+	BODY_32_39(39,D,E,T,A,B,C,X( 7),X( 9),X(15),X( 4));
+
+	BODY_40_59(40,C,D,E,T,A,B,X( 8),X(10),X( 0),X( 5));
+	BODY_40_59(41,B,C,D,E,T,A,X( 9),X(11),X( 1),X( 6));
+	BODY_40_59(42,A,B,C,D,E,T,X(10),X(12),X( 2),X( 7));
+	BODY_40_59(43,T,A,B,C,D,E,X(11),X(13),X( 3),X( 8));
+	BODY_40_59(44,E,T,A,B,C,D,X(12),X(14),X( 4),X( 9));
+	BODY_40_59(45,D,E,T,A,B,C,X(13),X(15),X( 5),X(10));
+	BODY_40_59(46,C,D,E,T,A,B,X(14),X( 0),X( 6),X(11));
+	BODY_40_59(47,B,C,D,E,T,A,X(15),X( 1),X( 7),X(12));
+	BODY_40_59(48,A,B,C,D,E,T,X( 0),X( 2),X( 8),X(13));
+	BODY_40_59(49,T,A,B,C,D,E,X( 1),X( 3),X( 9),X(14));
+	BODY_40_59(50,E,T,A,B,C,D,X( 2),X( 4),X(10),X(15));
+	BODY_40_59(51,D,E,T,A,B,C,X( 3),X( 5),X(11),X( 0));
+	BODY_40_59(52,C,D,E,T,A,B,X( 4),X( 6),X(12),X( 1));
+	BODY_40_59(53,B,C,D,E,T,A,X( 5),X( 7),X(13),X( 2));
+	BODY_40_59(54,A,B,C,D,E,T,X( 6),X( 8),X(14),X( 3));
+	BODY_40_59(55,T,A,B,C,D,E,X( 7),X( 9),X(15),X( 4));
+	BODY_40_59(56,E,T,A,B,C,D,X( 8),X(10),X( 0),X( 5));
+	BODY_40_59(57,D,E,T,A,B,C,X( 9),X(11),X( 1),X( 6));
+	BODY_40_59(58,C,D,E,T,A,B,X(10),X(12),X( 2),X( 7));
+	BODY_40_59(59,B,C,D,E,T,A,X(11),X(13),X( 3),X( 8));
+
+	BODY_60_79(60,A,B,C,D,E,T,X(12),X(14),X( 4),X( 9));
+	BODY_60_79(61,T,A,B,C,D,E,X(13),X(15),X( 5),X(10));
+	BODY_60_79(62,E,T,A,B,C,D,X(14),X( 0),X( 6),X(11));
+	BODY_60_79(63,D,E,T,A,B,C,X(15),X( 1),X( 7),X(12));
+	BODY_60_79(64,C,D,E,T,A,B,X( 0),X( 2),X( 8),X(13));
+	BODY_60_79(65,B,C,D,E,T,A,X( 1),X( 3),X( 9),X(14));
+	BODY_60_79(66,A,B,C,D,E,T,X( 2),X( 4),X(10),X(15));
+	BODY_60_79(67,T,A,B,C,D,E,X( 3),X( 5),X(11),X( 0));
+	BODY_60_79(68,E,T,A,B,C,D,X( 4),X( 6),X(12),X( 1));
+	BODY_60_79(69,D,E,T,A,B,C,X( 5),X( 7),X(13),X( 2));
+	BODY_60_79(70,C,D,E,T,A,B,X( 6),X( 8),X(14),X( 3));
+	BODY_60_79(71,B,C,D,E,T,A,X( 7),X( 9),X(15),X( 4));
+	BODY_60_79(72,A,B,C,D,E,T,X( 8),X(10),X( 0),X( 5));
+	BODY_60_79(73,T,A,B,C,D,E,X( 9),X(11),X( 1),X( 6));
+	BODY_60_79(74,E,T,A,B,C,D,X(10),X(12),X( 2),X( 7));
+	BODY_60_79(75,D,E,T,A,B,C,X(11),X(13),X( 3),X( 8));
+	BODY_60_79(76,C,D,E,T,A,B,X(12),X(14),X( 4),X( 9));
+	BODY_60_79(77,B,C,D,E,T,A,X(13),X(15),X( 5),X(10));
+	BODY_60_79(78,A,B,C,D,E,T,X(14),X( 0),X( 6),X(11));
+	BODY_60_79(79,T,A,B,C,D,E,X(15),X( 1),X( 7),X(12));
+	
+	c->h0=(c->h0+E)&0xffffffffL; 
+	c->h1=(c->h1+T)&0xffffffffL;
+	c->h2=(c->h2+A)&0xffffffffL;
+	c->h3=(c->h3+B)&0xffffffffL;
+	c->h4=(c->h4+C)&0xffffffffL;
+
+	if (--num <= 0) break;
+
+	A=c->h0;
+	B=c->h1;
+	C=c->h2;
+	D=c->h3;
+	E=c->h4;
+
+		}
+	}
+#endif