Whirlpool hash implementation. The fact that subdirectory and .h file are

called whrlpool is not a typo, but a way to keep the names shorter than
8 characters. Remaining TODO list comprises adding OID, EVP, corresponding
flag to apps/openssl dgst, benchmark, engage assembler...

Makefile.org

@@ -106,7 +106,7 @@ SHLIBDIRS= crypto ssl
 # dirs in crypto to build
 SDIRS=  \
 	objects \
-	md2 md4 md5 sha mdc2 hmac ripemd \
+	md2 md4 md5 sha mdc2 hmac ripemd whrlpool \
 	des aes rc2 rc4 rc5 idea bf cast \
 	bn ec rsa dsa ecdsa dh ecdh dso engine \
 	buffer bio stack lhash rand err \

crypto/perlasm/x86ms.pl

@@ -416,6 +416,11 @@ sub main'set_label
 		}
 	}
 
+sub main'data_byte
+	{
+	push(@out,"\tDB\t".join(',',@_)."\n");
+	}
+
 sub main'data_word
 	{
 	push(@out,"\tDD\t".join(',',@_)."\n");

crypto/perlasm/x86nasm.pl

@@ -401,6 +401,11 @@ sub main'set_label
 	push(@out,"$label{$_[0]}:\n");
 	}
 
+sub main'data_byte
+	{
+	push(@out,(($main'mwerks)?".byte\t":"DB\t").join(',',@_)."\n");
+	}
+
 sub main'data_word
 	{
 	push(@out,(($main'mwerks)?".long\t":"DD\t").join(',',@_)."\n");

crypto/perlasm/x86unix.pl

@@ -594,6 +594,11 @@ ___
 		}
 	}
 
+sub main'data_byte
+	{
+	push(@out,"\t.byte\t".join(',',@_)."\n");
+	}
+
 sub main'data_word
 	{
 	push(@out,"\t.long\t".join(',',@_)."\n");

crypto/whrlpool/asm/wp-mmx.pl

+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. Rights for redistribution and usage in source and binary
+# forms are granted according to the OpenSSL license.
+# ====================================================================
+#
+# whirlpool_block_mmx implementation.
+#
+*SCALE=\(2); # 2 or 8, that is the question:-) Value of 8 results
+# in 16KB large table, which is tough on L1 cache, but eliminates
+# unaligned references to it. Value of 2 results in 4KB table, but
+# 7/8 of references to it are unaligned. AMD cores seem to be
+# allergic to the latter, while Intel ones - to former [see the
+# table]. I stick to value of 2 for two reasons: 1. smaller table
+# minimizes cache trashing and thus mitigates the hazard of side-
+# channel leakage similar to AES cache-timing one; 2. performance
+# gap among different -archs is smaller.
+#
+# Performance table lists rounded amounts of CPU cycles spent by
+# whirlpool_block_mmx routine on single 64 byte input block, i.e.
+# smaller is better and asymptotic throughput can be estimated by
+# multiplying 64 by CPU clock frequency and dividing by relevant
+# value from the given table:
+#
+#		$SCALE=2/8	icc8	gcc3	
+# Intel P4	3200/4600	4600(*)	6400
+# Intel PIII	2900/3000	4900	5400
+# AMD K[78]	2500/1800	9900	8200(**)
+#
+# (*)	I've sketched even non-MMX assembler, but for the record
+#	I've failed to beat the Intel compiler on P4, without using
+*	MMX that is...
+# (**)	... on AMD on the other hand non-MMX assembler was observed
+#	to perform significantly better, but I figured this MMX
+#	implementation is even faster anyway, so why bother? As for
+#	pre-MMX AMD core[s], the improvement coefficient is more
+#	than likely to vary anyway and I don't know how. But the
+#	least I know is that gcc-generated code compiled with
+#	-DL_ENDIAN and -DOPENSSL_SMALL_FOOTPRINT [see C module for
+#	details] and optimized for Pentium was observed to perform
+#	*better* on Pentium 100 than unrolled non-MMX assembler
+#	loop... So we just say that I don't know if maintaining
+#	non-MMX implementation would actually pay off, but till
+#	opposite is proved "unlikely" is assumed.
+
+push(@INC,"../CVS/HEAD/openssl/crypto/perlasm","../../perlasm");
+require "x86asm.pl";
+
+&asm_init($ARGV[0],"wp-mmx.pl");
+
+sub L()  { &data_byte(@_); }
+sub LL()
+{	if	($SCALE==2)	{ &data_byte(@_); &data_byte(@_); }
+	elsif	($SCALE==8)	{ for ($i=0;$i<8;$i++) {
+					&data_byte(@_);
+					unshift(@_,pop(@_));
+				  }
+				}
+	else			{ die "unvalid SCALE value"; }
+}
+
+sub scale()
+{	if	($SCALE==2)	{ &lea(@_[0],&DWP(0,@_[1],@_[1])); }
+	elsif	($SCALE==8)	{ &lea(@_[0],&DWP(0,"",@_[1],8));  }
+	else			{ die "unvalid SCALE value";       }
+}
+
+sub row()
+{	if	($SCALE==2)	{ ((8-shift)&7); }
+	elsif	($SCALE==8)	{ (8*shift);     }
+	else			{ die "unvalid SCALE value"; }
+}
+
+$tbl="ebp";
+@mm=("mm0","mm1","mm2","mm3","mm4","mm5","mm6","mm7");
+
+&function_begin_B("whirlpool_block_mmx");
+	&push	("ebp");
+	&push	("ebx");
+	&push	("esi");
+	&push	("edi");
+
+	&mov	("esi",&wparam(0));		# hash value
+	&mov	("edi",&wparam(1));		# input data stream
+	&mov	("ebp",&wparam(2));		# number of chunks in input
+
+	&mov	("eax","esp");			# copy stack pointer
+	&sub	("esp",128+20);			# allocate frame
+	&and	("esp",-64);			# align for cache-line
+
+	&lea	("ebx",&DWP(128,"esp"));
+	&mov	(&DWP(0,"ebx"),"esi");		# save parameter block
+	&mov	(&DWP(4,"ebx"),"edi");
+	&mov	(&DWP(8,"ebx"),"ebp");
+	&mov	(&DWP(16,"ebx"),"eax");		# saved stack pointer
+
+	&call	(&label("pic_point"));
+&set_label("pic_point");
+	&blindpop($tbl);
+        &lea    ($tbl,&DWP(&label("table")."-".&label("pic_point"),$tbl));
+
+	&xor	("ecx","ecx");
+	&xor	("edx","edx");
+
+	for($i=0;$i<8;$i++) { &movq(@mm[$i],&QWP($i*8,"esi")); }    # L=H
+&set_label("outerloop");
+	for($i=0;$i<8;$i++) { &movq(&QWP($i*8,"esp"),@mm[$i]); }    # K=H
+	for($i=0;$i<8;$i++) { &pxor(@mm[$i],&QWP($i*8,"edi")); }    # L^=inp
+	for($i=0;$i<8;$i++) { &movq(&QWP(64+$i*8,"esp"),@mm[$i]); } #S=L
+
+	&xor	("esi","esi");
+	&mov	(&DWP(12,"ebx"),"esi");		# zero round counter
+
+&set_label("round",16);
+	&movq	(@mm[0],&DWP(2048*$SCALE,$tbl,"esi",8));	# rc[r]
+	&mov	("eax",&DWP(0,"esp"));
+	&mov	("ebx",&DWP(4,"esp"));
+for($i=0;$i<8;$i++) {
+    my $func = ($i==0)? movq : pxor;
+	&movb	(&LB("ecx"),&LB("eax"));
+	&movb	(&LB("edx"),&HB("eax"));
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&shr	("eax",16);
+	&pxor	(@mm[0],&QWP(&row(0),$tbl,"esi",8));
+	&$func	(@mm[1],&QWP(&row(1),$tbl,"edi",8));
+	&movb	(&LB("ecx"),&LB("eax"));
+	&movb	(&LB("edx"),&HB("eax"));
+	&mov	("eax",&DWP(($i+1)*8,"esp"));
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&$func	(@mm[2],&QWP(&row(2),$tbl,"esi",8));
+	&$func	(@mm[3],&QWP(&row(3),$tbl,"edi",8));
+	&movb	(&LB("ecx"),&LB("ebx"));
+	&movb	(&LB("edx"),&HB("ebx"));
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&shr	("ebx",16);
+	&$func	(@mm[4],&QWP(&row(4),$tbl,"esi",8));
+	&$func	(@mm[5],&QWP(&row(5),$tbl,"edi",8));
+	&movb	(&LB("ecx"),&LB("ebx"));
+	&movb	(&LB("edx"),&HB("ebx"));
+	&mov	("ebx",&DWP(($i+1)*8+4,"esp"));
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&$func	(@mm[6],&QWP(&row(6),$tbl,"esi",8));
+	&$func	(@mm[7],&QWP(&row(7),$tbl,"edi",8));
+    push(@mm,shift(@mm));
+}
+
+	for($i=0;$i<8;$i++) { &movq(&QWP($i*8,"esp"),@mm[$i]); }    # K=L
+
+for($i=0;$i<8;$i++) {
+	&movb	(&LB("ecx"),&LB("eax"));
+	&movb	(&LB("edx"),&HB("eax"));
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&shr	("eax",16);
+	&pxor	(@mm[0],&QWP(&row(0),$tbl,"esi",8));
+	&pxor	(@mm[1],&QWP(&row(1),$tbl,"edi",8));
+	&movb	(&LB("ecx"),&LB("eax"));
+	&movb	(&LB("edx"),&HB("eax"));
+	&mov	("eax",&DWP(64+($i+1)*8,"esp"))		if ($i<7);
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&pxor	(@mm[2],&QWP(&row(2),$tbl,"esi",8));
+	&pxor	(@mm[3],&QWP(&row(3),$tbl,"edi",8));
+	&movb	(&LB("ecx"),&LB("ebx"));
+	&movb	(&LB("edx"),&HB("ebx"));
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&shr	("ebx",16);
+	&pxor	(@mm[4],&QWP(&row(4),$tbl,"esi",8));
+	&pxor	(@mm[5],&QWP(&row(5),$tbl,"edi",8));
+	&movb	(&LB("ecx"),&LB("ebx"));
+	&movb	(&LB("edx"),&HB("ebx"));
+	&mov	("ebx",&DWP(64+($i+1)*8+4,"esp"))	if ($i<7);
+	&scale	("esi","ecx");
+	&scale	("edi","edx");
+	&pxor	(@mm[6],&QWP(&row(6),$tbl,"esi",8));
+	&pxor	(@mm[7],&QWP(&row(7),$tbl,"edi",8));
+    push(@mm,shift(@mm));
+}
+	&lea	("ebx",&DWP(128,"esp"));
+	&mov	("esi",&DWP(12,"ebx"));		# pull round counter
+	&add	("esi",1);
+	&cmp	("esi",10);
+	&je	(&label("roundsdone"));
+
+	&mov	(&DWP(12,"ebx"),"esi");		# update round counter
+	for($i=0;$i<8;$i++) { &movq(&QWP(64+$i*8,"esp"),@mm[$i]); } # S=L
+	&jmp	(&label("round"));
+
+&set_label("roundsdone",16);
+	&mov	("esi",&DWP(0,"ebx"));		# reload argument block
+	&mov	("edi",&DWP(4,"ebx"));
+	&mov	("eax",&DWP(8,"ebx"));
+
+	for($i=0;$i<8;$i++) { &pxor(@mm[$i],&QWP($i*8,"edi")); }    # L^=inp
+	for($i=0;$i<8;$i++) { &pxor(@mm[$i],&QWP($i*8,"esi")); }    # L^=H
+	for($i=0;$i<8;$i++) { &movq(&QWP($i*8,"esi"),@mm[$i]); }    # H=L
+
+	&lea	("edi",&DWP(64,"edi"));		# inp+=64
+	&sub	("eax",1);			# num--
+	&jz	(&label("alldone"));
+	&mov	(&DWP(4,"ebx"),"edi");		# update argument block
+	&mov	(&DWP(8,"ebx"),"eax");
+	&jmp	(&label("outerloop"));
+
+&set_label("alldone");
+	&emms	();
+	&mov	("esp",&DWP(16,"ebx"));		# restore saved stack pointer
+	&pop	("edi");
+	&pop	("esi");
+	&pop	("ebx");
+	&pop	("ebp");
+	&ret	();
+
+&align(64);
+&set_label("table");
+	&LL(0x18,0x18,0x60,0x18,0xc0,0x78,0x30,0xd8);
+	&LL(0x23,0x23,0x8c,0x23,0x05,0xaf,0x46,0x26);
+	&LL(0xc6,0xc6,0x3f,0xc6,0x7e,0xf9,0x91,0xb8);
+	&LL(0xe8,0xe8,0x87,0xe8,0x13,0x6f,0xcd,0xfb);
+	&LL(0x87,0x87,0x26,0x87,0x4c,0xa1,0x13,0xcb);
+	&LL(0xb8,0xb8,0xda,0xb8,0xa9,0x62,0x6d,0x11);
+	&LL(0x01,0x01,0x04,0x01,0x08,0x05,0x02,0x09);
+	&LL(0x4f,0x4f,0x21,0x4f,0x42,0x6e,0x9e,0x0d);
+	&LL(0x36,0x36,0xd8,0x36,0xad,0xee,0x6c,0x9b);
+	&LL(0xa6,0xa6,0xa2,0xa6,0x59,0x04,0x51,0xff);
+	&LL(0xd2,0xd2,0x6f,0xd2,0xde,0xbd,0xb9,0x0c);
+	&LL(0xf5,0xf5,0xf3,0xf5,0xfb,0x06,0xf7,0x0e);
+	&LL(0x79,0x79,0xf9,0x79,0xef,0x80,0xf2,0x96);
+	&LL(0x6f,0x6f,0xa1,0x6f,0x5f,0xce,0xde,0x30);
+	&LL(0x91,0x91,0x7e,0x91,0xfc,0xef,0x3f,0x6d);
+	&LL(0x52,0x52,0x55,0x52,0xaa,0x07,0xa4,0xf8);
+	&LL(0x60,0x60,0x9d,0x60,0x27,0xfd,0xc0,0x47);
+	&LL(0xbc,0xbc,0xca,0xbc,0x89,0x76,0x65,0x35);
+	&LL(0x9b,0x9b,0x56,0x9b,0xac,0xcd,0x2b,0x37);
+	&LL(0x8e,0x8e,0x02,0x8e,0x04,0x8c,0x01,0x8a);
+	&LL(0xa3,0xa3,0xb6,0xa3,0x71,0x15,0x5b,0xd2);
+	&LL(0x0c,0x0c,0x30,0x0c,0x60,0x3c,0x18,0x6c);
+	&LL(0x7b,0x7b,0xf1,0x7b,0xff,0x8a,0xf6,0x84);
+	&LL(0x35,0x35,0xd4,0x35,0xb5,0xe1,0x6a,0x80);
+	&LL(0x1d,0x1d,0x74,0x1d,0xe8,0x69,0x3a,0xf5);
+	&LL(0xe0,0xe0,0xa7,0xe0,0x53,0x47,0xdd,0xb3);
+	&LL(0xd7,0xd7,0x7b,0xd7,0xf6,0xac,0xb3,0x21);
+	&LL(0xc2,0xc2,0x2f,0xc2,0x5e,0xed,0x99,0x9c);
+	&LL(0x2e,0x2e,0xb8,0x2e,0x6d,0x96,0x5c,0x43);
+	&LL(0x4b,0x4b,0x31,0x4b,0x62,0x7a,0x96,0x29);
+	&LL(0xfe,0xfe,0xdf,0xfe,0xa3,0x21,0xe1,0x5d);
+	&LL(0x57,0x57,0x41,0x57,0x82,0x16,0xae,0xd5);
+	&LL(0x15,0x15,0x54,0x15,0xa8,0x41,0x2a,0xbd);
+	&LL(0x77,0x77,0xc1,0x77,0x9f,0xb6,0xee,0xe8);
+	&LL(0x37,0x37,0xdc,0x37,0xa5,0xeb,0x6e,0x92);
+	&LL(0xe5,0xe5,0xb3,0xe5,0x7b,0x56,0xd7,0x9e);
+	&LL(0x9f,0x9f,0x46,0x9f,0x8c,0xd9,0x23,0x13);
+	&LL(0xf0,0xf0,0xe7,0xf0,0xd3,0x17,0xfd,0x23);
+	&LL(0x4a,0x4a,0x35,0x4a,0x6a,0x7f,0x94,0x20);
+	&LL(0xda,0xda,0x4f,0xda,0x9e,0x95,0xa9,0x44);
+	&LL(0x58,0x58,0x7d,0x58,0xfa,0x25,0xb0,0xa2);
+	&LL(0xc9,0xc9,0x03,0xc9,0x06,0xca,0x8f,0xcf);
+	&LL(0x29,0x29,0xa4,0x29,0x55,0x8d,0x52,0x7c);
+	&LL(0x0a,0x0a,0x28,0x0a,0x50,0x22,0x14,0x5a);
+	&LL(0xb1,0xb1,0xfe,0xb1,0xe1,0x4f,0x7f,0x50);
+	&LL(0xa0,0xa0,0xba,0xa0,0x69,0x1a,0x5d,0xc9);
+	&LL(0x6b,0x6b,0xb1,0x6b,0x7f,0xda,0xd6,0x14);
+	&LL(0x85,0x85,0x2e,0x85,0x5c,0xab,0x17,0xd9);
+	&LL(0xbd,0xbd,0xce,0xbd,0x81,0x73,0x67,0x3c);
+	&LL(0x5d,0x5d,0x69,0x5d,0xd2,0x34,0xba,0x8f);
+	&LL(0x10,0x10,0x40,0x10,0x80,0x50,0x20,0x90);
+	&LL(0xf4,0xf4,0xf7,0xf4,0xf3,0x03,0xf5,0x07);
+	&LL(0xcb,0xcb,0x0b,0xcb,0x16,0xc0,0x8b,0xdd);
+	&LL(0x3e,0x3e,0xf8,0x3e,0xed,0xc6,0x7c,0xd3);
+	&LL(0x05,0x05,0x14,0x05,0x28,0x11,0x0a,0x2d);
+	&LL(0x67,0x67,0x81,0x67,0x1f,0xe6,0xce,0x78);
+	&LL(0xe4,0xe4,0xb7,0xe4,0x73,0x53,0xd5,0x97);
+	&LL(0x27,0x27,0x9c,0x27,0x25,0xbb,0x4e,0x02);
+	&LL(0x41,0x41,0x19,0x41,0x32,0x58,0x82,0x73);
+	&LL(0x8b,0x8b,0x16,0x8b,0x2c,0x9d,0x0b,0xa7);
+	&LL(0xa7,0xa7,0xa6,0xa7,0x51,0x01,0x53,0xf6);
+	&LL(0x7d,0x7d,0xe9,0x7d,0xcf,0x94,0xfa,0xb2);
+	&LL(0x95,0x95,0x6e,0x95,0xdc,0xfb,0x37,0x49);
+	&LL(0xd8,0xd8,0x47,0xd8,0x8e,0x9f,0xad,0x56);
+	&LL(0xfb,0xfb,0xcb,0xfb,0x8b,0x30,0xeb,0x70);
+	&LL(0xee,0xee,0x9f,0xee,0x23,0x71,0xc1,0xcd);
+	&LL(0x7c,0x7c,0xed,0x7c,0xc7,0x91,0xf8,0xbb);
+	&LL(0x66,0x66,0x85,0x66,0x17,0xe3,0xcc,0x71);
+	&LL(0xdd,0xdd,0x53,0xdd,0xa6,0x8e,0xa7,0x7b);
+	&LL(0x17,0x17,0x5c,0x17,0xb8,0x4b,0x2e,0xaf);
+	&LL(0x47,0x47,0x01,0x47,0x02,0x46,0x8e,0x45);
+	&LL(0x9e,0x9e,0x42,0x9e,0x84,0xdc,0x21,0x1a);
+	&LL(0xca,0xca,0x0f,0xca,0x1e,0xc5,0x89,0xd4);
+	&LL(0x2d,0x2d,0xb4,0x2d,0x75,0x99,0x5a,0x58);
+	&LL(0xbf,0xbf,0xc6,0xbf,0x91,0x79,0x63,0x2e);
+	&LL(0x07,0x07,0x1c,0x07,0x38,0x1b,0x0e,0x3f);
+	&LL(0xad,0xad,0x8e,0xad,0x01,0x23,0x47,0xac);
+	&LL(0x5a,0x5a,0x75,0x5a,0xea,0x2f,0xb4,0xb0);
+	&LL(0x83,0x83,0x36,0x83,0x6c,0xb5,0x1b,0xef);
+	&LL(0x33,0x33,0xcc,0x33,0x85,0xff,0x66,0xb6);
+	&LL(0x63,0x63,0x91,0x63,0x3f,0xf2,0xc6,0x5c);
+	&LL(0x02,0x02,0x08,0x02,0x10,0x0a,0x04,0x12);
+	&LL(0xaa,0xaa,0x92,0xaa,0x39,0x38,0x49,0x93);
+	&LL(0x71,0x71,0xd9,0x71,0xaf,0xa8,0xe2,0xde);
+	&LL(0xc8,0xc8,0x07,0xc8,0x0e,0xcf,0x8d,0xc6);
+	&LL(0x19,0x19,0x64,0x19,0xc8,0x7d,0x32,0xd1);
+	&LL(0x49,0x49,0x39,0x49,0x72,0x70,0x92,0x3b);
+	&LL(0xd9,0xd9,0x43,0xd9,0x86,0x9a,0xaf,0x5f);
+	&LL(0xf2,0xf2,0xef,0xf2,0xc3,0x1d,0xf9,0x31);
+	&LL(0xe3,0xe3,0xab,0xe3,0x4b,0x48,0xdb,0xa8);
+	&LL(0x5b,0x5b,0x71,0x5b,0xe2,0x2a,0xb6,0xb9);
+	&LL(0x88,0x88,0x1a,0x88,0x34,0x92,0x0d,0xbc);
+	&LL(0x9a,0x9a,0x52,0x9a,0xa4,0xc8,0x29,0x3e);
+	&LL(0x26,0x26,0x98,0x26,0x2d,0xbe,0x4c,0x0b);
+	&LL(0x32,0x32,0xc8,0x32,0x8d,0xfa,0x64,0xbf);
+	&LL(0xb0,0xb0,0xfa,0xb0,0xe9,0x4a,0x7d,0x59);
+	&LL(0xe9,0xe9,0x83,0xe9,0x1b,0x6a,0xcf,0xf2);
+	&LL(0x0f,0x0f,0x3c,0x0f,0x78,0x33,0x1e,0x77);
+	&LL(0xd5,0xd5,0x73,0xd5,0xe6,0xa6,0xb7,0x33);
+	&LL(0x80,0x80,0x3a,0x80,0x74,0xba,0x1d,0xf4);
+	&LL(0xbe,0xbe,0xc2,0xbe,0x99,0x7c,0x61,0x27);
+	&LL(0xcd,0xcd,0x13,0xcd,0x26,0xde,0x87,0xeb);
+	&LL(0x34,0x34,0xd0,0x34,0xbd,0xe4,0x68,0x89);
+	&LL(0x48,0x48,0x3d,0x48,0x7a,0x75,0x90,0x32);
+	&LL(0xff,0xff,0xdb,0xff,0xab,0x24,0xe3,0x54);
+	&LL(0x7a,0x7a,0xf5,0x7a,0xf7,0x8f,0xf4,0x8d);
+	&LL(0x90,0x90,0x7a,0x90,0xf4,0xea,0x3d,0x64);
+	&LL(0x5f,0x5f,0x61,0x5f,0xc2,0x3e,0xbe,0x9d);
+	&LL(0x20,0x20,0x80,0x20,0x1d,0xa0,0x40,0x3d);
+	&LL(0x68,0x68,0xbd,0x68,0x67,0xd5,0xd0,0x0f);
+	&LL(0x1a,0x1a,0x68,0x1a,0xd0,0x72,0x34,0xca);
+	&LL(0xae,0xae,0x82,0xae,0x19,0x2c,0x41,0xb7);
+	&LL(0xb4,0xb4,0xea,0xb4,0xc9,0x5e,0x75,0x7d);
+	&LL(0x54,0x54,0x4d,0x54,0x9a,0x19,0xa8,0xce);
+	&LL(0x93,0x93,0x76,0x93,0xec,0xe5,0x3b,0x7f);
+	&LL(0x22,0x22,0x88,0x22,0x0d,0xaa,0x44,0x2f);
+	&LL(0x64,0x64,0x8d,0x64,0x07,0xe9,0xc8,0x63);
+	&LL(0xf1,0xf1,0xe3,0xf1,0xdb,0x12,0xff,0x2a);
+	&LL(0x73,0x73,0xd1,0x73,0xbf,0xa2,0xe6,0xcc);
+	&LL(0x12,0x12,0x48,0x12,0x90,0x5a,0x24,0x82);
+	&LL(0x40,0x40,0x1d,0x40,0x3a,0x5d,0x80,0x7a);
+	&LL(0x08,0x08,0x20,0x08,0x40,0x28,0x10,0x48);
+	&LL(0xc3,0xc3,0x2b,0xc3,0x56,0xe8,0x9b,0x95);
+	&LL(0xec,0xec,0x97,0xec,0x33,0x7b,0xc5,0xdf);
+	&LL(0xdb,0xdb,0x4b,0xdb,0x96,0x90,0xab,0x4d);
+	&LL(0xa1,0xa1,0xbe,0xa1,0x61,0x1f,0x5f,0xc0);
+	&LL(0x8d,0x8d,0x0e,0x8d,0x1c,0x83,0x07,0x91);
+	&LL(0x3d,0x3d,0xf4,0x3d,0xf5,0xc9,0x7a,0xc8);
+	&LL(0x97,0x97,0x66,0x97,0xcc,0xf1,0x33,0x5b);
+	&LL(0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00);
+	&LL(0xcf,0xcf,0x1b,0xcf,0x36,0xd4,0x83,0xf9);
+	&LL(0x2b,0x2b,0xac,0x2b,0x45,0x87,0x56,0x6e);
+	&LL(0x76,0x76,0xc5,0x76,0x97,0xb3,0xec,0xe1);
+	&LL(0x82,0x82,0x32,0x82,0x64,0xb0,0x19,0xe6);
+	&LL(0xd6,0xd6,0x7f,0xd6,0xfe,0xa9,0xb1,0x28);
+	&LL(0x1b,0x1b,0x6c,0x1b,0xd8,0x77,0x36,0xc3);
+	&LL(0xb5,0xb5,0xee,0xb5,0xc1,0x5b,0x77,0x74);
+	&LL(0xaf,0xaf,0x86,0xaf,0x11,0x29,0x43,0xbe);
+	&LL(0x6a,0x6a,0xb5,0x6a,0x77,0xdf,0xd4,0x1d);
+	&LL(0x50,0x50,0x5d,0x50,0xba,0x0d,0xa0,0xea);
+	&LL(0x45,0x45,0x09,0x45,0x12,0x4c,0x8a,0x57);
+	&LL(0xf3,0xf3,0xeb,0xf3,0xcb,0x18,0xfb,0x38);
+	&LL(0x30,0x30,0xc0,0x30,0x9d,0xf0,0x60,0xad);
+	&LL(0xef,0xef,0x9b,0xef,0x2b,0x74,0xc3,0xc4);
+	&LL(0x3f,0x3f,0xfc,0x3f,0xe5,0xc3,0x7e,0xda);
+	&LL(0x55,0x55,0x49,0x55,0x92,0x1c,0xaa,0xc7);
+	&LL(0xa2,0xa2,0xb2,0xa2,0x79,0x10,0x59,0xdb);
+	&LL(0xea,0xea,0x8f,0xea,0x03,0x65,0xc9,0xe9);
+	&LL(0x65,0x65,0x89,0x65,0x0f,0xec,0xca,0x6a);
+	&LL(0xba,0xba,0xd2,0xba,0xb9,0x68,0x69,0x03);
+	&LL(0x2f,0x2f,0xbc,0x2f,0x65,0x93,0x5e,0x4a);
+	&LL(0xc0,0xc0,0x27,0xc0,0x4e,0xe7,0x9d,0x8e);
+	&LL(0xde,0xde,0x5f,0xde,0xbe,0x81,0xa1,0x60);
+	&LL(0x1c,0x1c,0x70,0x1c,0xe0,0x6c,0x38,0xfc);
+	&LL(0xfd,0xfd,0xd3,0xfd,0xbb,0x2e,0xe7,0x46);
+	&LL(0x4d,0x4d,0x29,0x4d,0x52,0x64,0x9a,0x1f);
+	&LL(0x92,0x92,0x72,0x92,0xe4,0xe0,0x39,0x76);
+	&LL(0x75,0x75,0xc9,0x75,0x8f,0xbc,0xea,0xfa);
+	&LL(0x06,0x06,0x18,0x06,0x30,0x1e,0x0c,0x36);
+	&LL(0x8a,0x8a,0x12,0x8a,0x24,0x98,0x09,0xae);
+	&LL(0xb2,0xb2,0xf2,0xb2,0xf9,0x40,0x79,0x4b);
+	&LL(0xe6,0xe6,0xbf,0xe6,0x63,0x59,0xd1,0x85);
+	&LL(0x0e,0x0e,0x38,0x0e,0x70,0x36,0x1c,0x7e);
+	&LL(0x1f,0x1f,0x7c,0x1f,0xf8,0x63,0x3e,0xe7);
+	&LL(0x62,0x62,0x95,0x62,0x37,0xf7,0xc4,0x55);
+	&LL(0xd4,0xd4,0x77,0xd4,0xee,0xa3,0xb5,0x3a);
+	&LL(0xa8,0xa8,0x9a,0xa8,0x29,0x32,0x4d,0x81);
+	&LL(0x96,0x96,0x62,0x96,0xc4,0xf4,0x31,0x52);
+	&LL(0xf9,0xf9,0xc3,0xf9,0x9b,0x3a,0xef,0x62);
+	&LL(0xc5,0xc5,0x33,0xc5,0x66,0xf6,0x97,0xa3);
+	&LL(0x25,0x25,0x94,0x25,0x35,0xb1,0x4a,0x10);
+	&LL(0x59,0x59,0x79,0x59,0xf2,0x20,0xb2,0xab);
+	&LL(0x84,0x84,0x2a,0x84,0x54,0xae,0x15,0xd0);
+	&LL(0x72,0x72,0xd5,0x72,0xb7,0xa7,0xe4,0xc5);
+	&LL(0x39,0x39,0xe4,0x39,0xd5,0xdd,0x72,0xec);
+	&LL(0x4c,0x4c,0x2d,0x4c,0x5a,0x61,0x98,0x16);
+	&LL(0x5e,0x5e,0x65,0x5e,0xca,0x3b,0xbc,0x94);
+	&LL(0x78,0x78,0xfd,0x78,0xe7,0x85,0xf0,0x9f);
+	&LL(0x38,0x38,0xe0,0x38,0xdd,0xd8,0x70,0xe5);
+	&LL(0x8c,0x8c,0x0a,0x8c,0x14,0x86,0x05,0x98);
+	&LL(0xd1,0xd1,0x63,0xd1,0xc6,0xb2,0xbf,0x17);
+	&LL(0xa5,0xa5,0xae,0xa5,0x41,0x0b,0x57,0xe4);
+	&LL(0xe2,0xe2,0xaf,0xe2,0x43,0x4d,0xd9,0xa1);
+	&LL(0x61,0x61,0x99,0x61,0x2f,0xf8,0xc2,0x4e);
+	&LL(0xb3,0xb3,0xf6,0xb3,0xf1,0x45,0x7b,0x42);
+	&LL(0x21,0x21,0x84,0x21,0x15,0xa5,0x42,0x34);
+	&LL(0x9c,0x9c,0x4a,0x9c,0x94,0xd6,0x25,0x08);
+	&LL(0x1e,0x1e,0x78,0x1e,0xf0,0x66,0x3c,0xee);
+	&LL(0x43,0x43,0x11,0x43,0x22,0x52,0x86,0x61);
+	&LL(0xc7,0xc7,0x3b,0xc7,0x76,0xfc,0x93,0xb1);
+	&LL(0xfc,0xfc,0xd7,0xfc,0xb3,0x2b,0xe5,0x4f);
+	&LL(0x04,0x04,0x10,0x04,0x20,0x14,0x08,0x24);
+	&LL(0x51,0x51,0x59,0x51,0xb2,0x08,0xa2,0xe3);
+	&LL(0x99,0x99,0x5e,0x99,0xbc,0xc7,0x2f,0x25);
+	&LL(0x6d,0x6d,0xa9,0x6d,0x4f,0xc4,0xda,0x22);
+	&LL(0x0d,0x0d,0x34,0x0d,0x68,0x39,0x1a,0x65);
+	&LL(0xfa,0xfa,0xcf,0xfa,0x83,0x35,0xe9,0x79);
+	&LL(0xdf,0xdf,0x5b,0xdf,0xb6,0x84,0xa3,0x69);
+	&LL(0x7e,0x7e,0xe5,0x7e,0xd7,0x9b,0xfc,0xa9);
+	&LL(0x24,0x24,0x90,0x24,0x3d,0xb4,0x48,0x19);
+	&LL(0x3b,0x3b,0xec,0x3b,0xc5,0xd7,0x76,0xfe);
+	&LL(0xab,0xab,0x96,0xab,0x31,0x3d,0x4b,0x9a);
+	&LL(0xce,0xce,0x1f,0xce,0x3e,0xd1,0x81,0xf0);
+	&LL(0x11,0x11,0x44,0x11,0x88,0x55,0x22,0x99);
+	&LL(0x8f,0x8f,0x06,0x8f,0x0c,0x89,0x03,0x83);
+	&LL(0x4e,0x4e,0x25,0x4e,0x4a,0x6b,0x9c,0x04);
+	&LL(0xb7,0xb7,0xe6,0xb7,0xd1,0x51,0x73,0x66);
+	&LL(0xeb,0xeb,0x8b,0xeb,0x0b,0x60,0xcb,0xe0);
+	&LL(0x3c,0x3c,0xf0,0x3c,0xfd,0xcc,0x78,0xc1);
+	&LL(0x81,0x81,0x3e,0x81,0x7c,0xbf,0x1f,0xfd);
+	&LL(0x94,0x94,0x6a,0x94,0xd4,0xfe,0x35,0x40);
+	&LL(0xf7,0xf7,0xfb,0xf7,0xeb,0x0c,0xf3,0x1c);
+	&LL(0xb9,0xb9,0xde,0xb9,0xa1,0x67,0x6f,0x18);
+	&LL(0x13,0x13,0x4c,0x13,0x98,0x5f,0x26,0x8b);
+	&LL(0x2c,0x2c,0xb0,0x2c,0x7d,0x9c,0x58,0x51);
+	&LL(0xd3,0xd3,0x6b,0xd3,0xd6,0xb8,0xbb,0x05);
+	&LL(0xe7,0xe7,0xbb,0xe7,0x6b,0x5c,0xd3,0x8c);
+	&LL(0x6e,0x6e,0xa5,0x6e,0x57,0xcb,0xdc,0x39);
+	&LL(0xc4,0xc4,0x37,0xc4,0x6e,0xf3,0x95,0xaa);
+	&LL(0x03,0x03,0x0c,0x03,0x18,0x0f,0x06,0x1b);
+	&LL(0x56,0x56,0x45,0x56,0x8a,0x13,0xac,0xdc);
+	&LL(0x44,0x44,0x0d,0x44,0x1a,0x49,0x88,0x5e);
+	&LL(0x7f,0x7f,0xe1,0x7f,0xdf,0x9e,0xfe,0xa0);
+	&LL(0xa9,0xa9,0x9e,0xa9,0x21,0x37,0x4f,0x88);
+	&LL(0x2a,0x2a,0xa8,0x2a,0x4d,0x82,0x54,0x67);
+	&LL(0xbb,0xbb,0xd6,0xbb,0xb1,0x6d,0x6b,0x0a);
+	&LL(0xc1,0xc1,0x23,0xc1,0x46,0xe2,0x9f,0x87);
+	&LL(0x53,0x53,0x51,0x53,0xa2,0x02,0xa6,0xf1);
+	&LL(0xdc,0xdc,0x57,0xdc,0xae,0x8b,0xa5,0x72);
+	&LL(0x0b,0x0b,0x2c,0x0b,0x58,0x27,0x16,0x53);
+	&LL(0x9d,0x9d,0x4e,0x9d,0x9c,0xd3,0x27,0x01);
+	&LL(0x6c,0x6c,0xad,0x6c,0x47,0xc1,0xd8,0x2b);
+	&LL(0x31,0x31,0xc4,0x31,0x95,0xf5,0x62,0xa4);
+	&LL(0x74,0x74,0xcd,0x74,0x87,0xb9,0xe8,0xf3);
+	&LL(0xf6,0xf6,0xff,0xf6,0xe3,0x09,0xf1,0x15);
+	&LL(0x46,0x46,0x05,0x46,0x0a,0x43,0x8c,0x4c);
+	&LL(0xac,0xac,0x8a,0xac,0x09,0x26,0x45,0xa5);
+	&LL(0x89,0x89,0x1e,0x89,0x3c,0x97,0x0f,0xb5);
+	&LL(0x14,0x14,0x50,0x14,0xa0,0x44,0x28,0xb4);
+	&LL(0xe1,0xe1,0xa3,0xe1,0x5b,0x42,0xdf,0xba);
+	&LL(0x16,0x16,0x58,0x16,0xb0,0x4e,0x2c,0xa6);
+	&LL(0x3a,0x3a,0xe8,0x3a,0xcd,0xd2,0x74,0xf7);
+	&LL(0x69,0x69,0xb9,0x69,0x6f,0xd0,0xd2,0x06);
+	&LL(0x09,0x09,0x24,0x09,0x48,0x2d,0x12,0x41);
+	&LL(0x70,0x70,0xdd,0x70,0xa7,0xad,0xe0,0xd7);
+	&LL(0xb6,0xb6,0xe2,0xb6,0xd9,0x54,0x71,0x6f);
+	&LL(0xd0,0xd0,0x67,0xd0,0xce,0xb7,0xbd,0x1e);
+	&LL(0xed,0xed,0x93,0xed,0x3b,0x7e,0xc7,0xd6);
+	&LL(0xcc,0xcc,0x17,0xcc,0x2e,0xdb,0x85,0xe2);
+	&LL(0x42,0x42,0x15,0x42,0x2a,0x57,0x84,0x68);
+	&LL(0x98,0x98,0x5a,0x98,0xb4,0xc2,0x2d,0x2c);
+	&LL(0xa4,0xa4,0xaa,0xa4,0x49,0x0e,0x55,0xed);
+	&LL(0x28,0x28,0xa0,0x28,0x5d,0x88,0x50,0x75);
+	&LL(0x5c,0x5c,0x6d,0x5c,0xda,0x31,0xb8,0x86);
+	&LL(0xf8,0xf8,0xc7,0xf8,0x93,0x3f,0xed,0x6b);
+	&LL(0x86,0x86,0x22,0x86,0x44,0xa4,0x11,0xc2);
+
+	&L(0x18,0x23,0xc6,0xe8,0x87,0xb8,0x01,0x4f);	# rc[ROUNDS]
+	&L(0x36,0xa6,0xd2,0xf5,0x79,0x6f,0x91,0x52);
+	&L(0x60,0xbc,0x9b,0x8e,0xa3,0x0c,0x7b,0x35);
+	&L(0x1d,0xe0,0xd7,0xc2,0x2e,0x4b,0xfe,0x57);
+	&L(0x15,0x77,0x37,0xe5,0x9f,0xf0,0x4a,0xda);
+	&L(0x58,0xc9,0x29,0x0a,0xb1,0xa0,0x6b,0x85);
+	&L(0xbd,0x5d,0x10,0xf4,0xcb,0x3e,0x05,0x67);
+	&L(0xe4,0x27,0x41,0x8b,0xa7,0x7d,0x95,0xd8);
+	&L(0xfb,0xee,0x7c,0x66,0xdd,0x17,0x47,0x9e);
+	&L(0xca,0x2d,0xbf,0x07,0xad,0x5a,0x83,0x33);
+
+&function_end_B("whrilpool_block_mmx");
+&asm_finish(); 

crypto/whrlpool/whrlpool.h

+#ifndef HEADER_WHRLPOOL_H
+#define HEADER_WHRLPOOL_H
+
+#include <openssl/e_os2.h>
+#include <stddef.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#define WHIRLPOOL_DIGEST_LENGTH	(512/8)
+#define WHIRLPOOL_BBLOCK	512
+#define WHIRLPOOL_COUNTER	(256/8)
+
+typedef struct	{
+	union	{
+		unsigned char	c[WHIRLPOOL_DIGEST_LENGTH];
+		/* double q is here to ensures 64-bit alignment */
+		double		q[WHIRLPOOL_DIGEST_LENGTH/sizeof(double)];
+		}	H;
+	unsigned char	data[WHIRLPOOL_BBLOCK/8];
+	unsigned int	bitoff;
+	size_t		bitlen[WHIRLPOOL_COUNTER/sizeof(size_t)];
+	} WHIRLPOOL_CTX;
+
+#ifndef OPENSSL_NO_WHIRLPOOL
+void WHIRLPOOL_Init	(WHIRLPOOL_CTX *c);
+void WHIRLPOOL_Update	(WHIRLPOOL_CTX *c,const void *inp,size_t bytes);
+void WHIRLPOOL_BitUpdate(WHIRLPOOL_CTX *c,const void *inp,size_t bits);
+void WHIRLPOOL_Final	(unsigned char *md,WHIRLPOOL_CTX *c);
+unsigned char *WHIRLPOOL(const void *inp,size_t bytes,unsigned char *md);
+#endif
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif

crypto/whrlpool/wp_dgst.c

+/**
+ * The Whirlpool hashing function.
+ *
+ * <P>
+ * <b>References</b>
+ *
+ * <P>
+ * The Whirlpool algorithm was developed by
+ * <a href="mailto:pbarreto@scopus.com.br">Paulo S. L. M. Barreto</a> and
+ * <a href="mailto:vincent.rijmen@cryptomathic.com">Vincent Rijmen</a>.
+ *
+ * See
+ *      P.S.L.M. Barreto, V. Rijmen,
+ *      ``The Whirlpool hashing function,''
+ *      NESSIE submission, 2000 (tweaked version, 2001),
+ *      <https://www.cosic.esat.kuleuven.ac.be/nessie/workshop/submissions/whirlpool.zip>
+ *
+ * Based on "@version 3.0 (2003.03.12)" by Paulo S.L.M. Barreto and
+ * Vincent Rijmen. Lookup "reference implementations" on
+ * <http://planeta.terra.com.br/informatica/paulobarreto/>
+ *
+ * =============================================================================
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHORS ''AS IS'' AND ANY EXPRESS
+ * OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
+ * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
+ * OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
+ * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ */
+
+/*
+ * OpenSSL-specific implementation notes.
+ *
+ * WHIRLPOOL_Update as well as one-stroke WHIRLPOOL both expect
+ * number of *bytes* as input length argument. Bit-oriented routine
+ * as specified by authors is called WHIRLPOOL_BitUpdate[!] and
+ * does not have one-stroke counterpart.
+ *
+ * WHIRLPOOL_BitUpdate implements byte-oriented loop, essentially
+ * to serve WHIRLPOOL_Update. This is done for performance.
+ *
+ * Unlike authors' reference implementation, block processing
+ * routine whirlpool_block is designed to operate on multi-block
+ * input. This is done for perfomance.
+ */
+
+#include "wp_locl.h"
+#include <string.h>
+
+void WHIRLPOOL_Init	(WHIRLPOOL_CTX *c) { memset (c,0,sizeof(*c)); }
+
+void WHIRLPOOL_Update	(WHIRLPOOL_CTX *c,const void *_inp,size_t bytes)
+	{
+	/* Well, largest suitable chunk size actually is
+	 * (1<<(sizeof(size_t)*8-3))-64, but below number
+	 * is large enough for not to care about excessive
+	 * calls to WHIRLPOOL_BitUpdate... */
+	size_t chunk = ((size_t)1)<<(sizeof(size_t)*8-4);
+	const unsigned char *inp = _inp;
+
+	while (bytes>=chunk)
+		{
+		WHIRLPOOL_BitUpdate(c,inp,chunk*8);
+		bytes -= chunk;
+		inp   += chunk;
+		}
+	if (bytes)
+		WHIRLPOOL_BitUpdate(c,inp,bytes*8);
+	}
+
+void WHIRLPOOL_BitUpdate(WHIRLPOOL_CTX *c,const void *_inp,size_t bits)
+	{
+	size_t		n;
+	unsigned int	bitoff = c->bitoff,
+			bitrem = bitoff%8,
+			inpgap = (8-(unsigned int)bits%8)&7;
+	const unsigned char *inp=_inp;
+
+	/* This 256-bit increment procedure relies on the size_t
+	 * being natural size of CPU register, so that we don't
+	 * have to mask the value in order to detect overflows. */
+	c->bitlen[0] += bits;
+	if (c->bitlen[0] < bits)	/* overflow */
+		{
+		n = 1;
+		do  { 	c->bitlen[n]++;
+		    } while(c->bitlen[n]==0
+		   	    && ++n<(WHIRLPOOL_COUNTER/sizeof(size_t)));
+		}
+
+#ifndef OPENSSL_SMALL_FOOTPRINT
+	reconsider:
+	if (inpgap==0 && bitrem==0)	/* byte-oriented loop */
+		{
+		while (bits)
+			{
+			if (bitoff==0 && (n=bits/WHIRLPOOL_BBLOCK))
+				{
+				whirlpool_block(c,inp,n);
+				inp  += n*WHIRLPOOL_BBLOCK/8;
+				bits %= WHIRLPOOL_BBLOCK;
+				}
+			else
+				{
+				unsigned int byteoff = bitoff/8;
+
+				bitrem = WHIRLPOOL_BBLOCK - bitoff;/* re-use bitrem */
+				if (bits >= bitrem)
+					{
+					bits -= bitrem;
+					bitrem /= 8;
+					memcpy(c->data+byteoff,inp,bitrem);
+					inp  += bitrem;
+					whirlpool_block(c,c->data,1);
+					bitoff = 0;
+					}
+				else
+					{
+					memcpy(c->data+byteoff,inp,bits/8);
+					bitoff += bits;
+					bits = 0;
+					}
+				c->bitoff = bitoff;
+				}
+			}
+		}
+	else				/* bit-oriented loop */
+#endif
+		{
+		/*
+			   inp
+			   |
+			   +-------+-------+-------
+			      |||||||||||||||||||||
+			   +-------+-------+-------
+		+-------+-------+-------+-------+-------
+		||||||||||||||				c->data
+		+-------+-------+-------+-------+-------
+			|
+			c->bitoff/8
+		*/
+		while (bits)
+			{
+			unsigned int	byteoff	= bitoff/8;
+			unsigned char	b;
+
+#ifndef OPENSSL_SMALL_FOOTPRINT
+			if (bitrem==inpgap)
+				{
+				c->data[byteoff++] |= inp[0] & (0xff>>inpgap);
+				inpgap = 8-inpgap;
+				bitoff += inpgap;  bitrem = 0;	/* bitoff%8 */
+				bits   -= inpgap;  inpgap = 0;	/* bits%8   */
+				inp++;
+				if (bitoff==WHIRLPOOL_BBLOCK)
+					{
+					whirlpool_block(c,c->data,1);
+					bitoff = 0;
+					}
+				c->bitoff = bitoff;
+				goto reconsider;
+				}
+			else
+#endif
+			if (bits>=8)
+				{
+				b  = ((inp[0]<<inpgap) | (inp[1]>>(8-inpgap)));
+				b &= 0xff;
+				if (bitrem)	c->data[byteoff++] |= b>>bitrem;
+				else		c->data[byteoff++]  = b;
+				bitoff += 8;
+				bits   -= 8;
+				inp++;
+				if (bitoff>=WHIRLPOOL_BBLOCK)
+					{
+					whirlpool_block(c,c->data,1);
+					byteoff  = 0;
+					bitoff  %= WHIRLPOOL_BBLOCK;
+					}
+				if (bitrem)	c->data[byteoff] = b<<(8-bitrem);
+				}
+			else	/* remaining less than 8 bits */
+				{
+				b = (inp[0]<<inpgap)&0xff;
+				if (bitrem)	c->data[byteoff++] |= b>>bitrem;
+				else		c->data[byteoff++]  = b;
+				bitoff += bits;
+				if (bitoff==WHIRLPOOL_BBLOCK)
+					{
+					whirlpool_block(c,c->data,1);
+					byteoff  = 0;
+			        	bitoff  %= WHIRLPOOL_BBLOCK;
+					}
+				if (bitrem)	c->data[byteoff] = b<<(8-bitrem);
+				bits = 0;
+				}
+			c->bitoff = bitoff;
+			}
+		}
+	}
+
+void WHIRLPOOL_Final	(unsigned char *md,WHIRLPOOL_CTX *c)
+	{
+	unsigned int	bitoff  = c->bitoff,
+			byteoff = bitoff/8;
+	int		i,j;
+	size_t		v;
+	unsigned char  *p;
+
+	bitoff %= 8;
+	if (bitoff)	c->data[byteoff] |= 0x80>>bitoff;
+	else		c->data[byteoff]  = 0x80;
+	byteoff++;
+
+	/* pad with zeros */
+	if (byteoff > (WHIRLPOOL_BBLOCK/8-WHIRLPOOL_COUNTER))
+		{
+		if (byteoff<WHIRLPOOL_BBLOCK/8)
+			memset(&c->data[byteoff],0,WHIRLPOOL_BBLOCK/8-byteoff);
+		whirlpool_block(c,c->data,1);
+		byteoff = 0;
+		}
+	if (byteoff < (WHIRLPOOL_BBLOCK/8-WHIRLPOOL_COUNTER))
+		memset(&c->data[byteoff],0,
+			(WHIRLPOOL_BBLOCK/8-WHIRLPOOL_COUNTER)-byteoff);
+	/* smash 256-bit c->bitlen in big-endian order */
+	p = &c->data[WHIRLPOOL_BBLOCK/8-1];	/* last byte in c->data */
+	for(i=0;i<WHIRLPOOL_COUNTER/sizeof(size_t);i++)
+		for(v=c->bitlen[i],j=0;j<sizeof(size_t);j++,v>>=8)
+			*p-- = (unsigned char)(v&0xff);
+
+	whirlpool_block(c,c->data,1);
+
+	memcpy(md,c->H.c,WHIRLPOOL_DIGEST_LENGTH);
+	memset(c,0,sizeof(*c));
+	}
+
+unsigned char *WHIRLPOOL(const void *inp, size_t bytes,unsigned char *md)
+	{
+	WHIRLPOOL_CTX ctx;
+	static unsigned char m[WHIRLPOOL_DIGEST_LENGTH];
+
+	if (md == NULL) md=m;
+	WHIRLPOOL_Init(&ctx);
+	WHIRLPOOL_Update(&ctx,inp,bytes);
+	WHIRLPOOL_Final(md,&ctx);
+	return(md);
+	}

crypto/whrlpool/wp_locl.h

+#include <openssl/whrlpool.h>
+
+void whirlpool_block(WHIRLPOOL_CTX *,const void *,size_t);

crypto/whrlpool/wp_test.c

+/* ====================================================================
+ * Copyright (c) 2005 The OpenSSL Project.  All rights reserved.
+ * ====================================================================
+ */
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+
+#include <openssl/whrlpool.h>
+#include <openssl/crypto.h>
+
+#if defined(OPENSSL_NO_WHIRLPOOL)
+int main(int argc, char *argv[])
+{
+    printf("No Whirlpool support\n");
+    return(0);
+}
+#else
+
+/* ISO/IEC 10118-3 test vector set */
+unsigned char iso_test_1[WHIRLPOOL_DIGEST_LENGTH] = {
+	0x19,0xFA,0x61,0xD7,0x55,0x22,0xA4,0x66,
+	0x9B,0x44,0xE3,0x9C,0x1D,0x2E,0x17,0x26,
+	0xC5,0x30,0x23,0x21,0x30,0xD4,0x07,0xF8,
+	0x9A,0xFE,0xE0,0x96,0x49,0x97,0xF7,0xA7,
+	0x3E,0x83,0xBE,0x69,0x8B,0x28,0x8F,0xEB,
+	0xCF,0x88,0xE3,0xE0,0x3C,0x4F,0x07,0x57,
+	0xEA,0x89,0x64,0xE5,0x9B,0x63,0xD9,0x37,
+	0x08,0xB1,0x38,0xCC,0x42,0xA6,0x6E,0xB3 };
+
+unsigned char iso_test_2[WHIRLPOOL_DIGEST_LENGTH] = {
+	0x8A,0xCA,0x26,0x02,0x79,0x2A,0xEC,0x6F,
+	0x11,0xA6,0x72,0x06,0x53,0x1F,0xB7,0xD7,
+	0xF0,0xDF,0xF5,0x94,0x13,0x14,0x5E,0x69,
+	0x73,0xC4,0x50,0x01,0xD0,0x08,0x7B,0x42,
+	0xD1,0x1B,0xC6,0x45,0x41,0x3A,0xEF,0xF6,
+	0x3A,0x42,0x39,0x1A,0x39,0x14,0x5A,0x59,
+	0x1A,0x92,0x20,0x0D,0x56,0x01,0x95,0xE5,
+	0x3B,0x47,0x85,0x84,0xFD,0xAE,0x23,0x1A };
+
+unsigned char iso_test_3[WHIRLPOOL_DIGEST_LENGTH] = {
+	0x4E,0x24,0x48,0xA4,0xC6,0xF4,0x86,0xBB,
+	0x16,0xB6,0x56,0x2C,0x73,0xB4,0x02,0x0B,
+	0xF3,0x04,0x3E,0x3A,0x73,0x1B,0xCE,0x72,
+	0x1A,0xE1,0xB3,0x03,0xD9,0x7E,0x6D,0x4C,
+	0x71,0x81,0xEE,0xBD,0xB6,0xC5,0x7E,0x27,
+	0x7D,0x0E,0x34,0x95,0x71,0x14,0xCB,0xD6,
+	0xC7,0x97,0xFC,0x9D,0x95,0xD8,0xB5,0x82,
+	0xD2,0x25,0x29,0x20,0x76,0xD4,0xEE,0xF5 };
+
+unsigned char iso_test_4[WHIRLPOOL_DIGEST_LENGTH] = {
+	0x37,0x8C,0x84,0xA4,0x12,0x6E,0x2D,0xC6,
+	0xE5,0x6D,0xCC,0x74,0x58,0x37,0x7A,0xAC,
+	0x83,0x8D,0x00,0x03,0x22,0x30,0xF5,0x3C,
+	0xE1,0xF5,0x70,0x0C,0x0F,0xFB,0x4D,0x3B,
+	0x84,0x21,0x55,0x76,0x59,0xEF,0x55,0xC1,
+	0x06,0xB4,0xB5,0x2A,0xC5,0xA4,0xAA,0xA6,
+	0x92,0xED,0x92,0x00,0x52,0x83,0x8F,0x33,
+	0x62,0xE8,0x6D,0xBD,0x37,0xA8,0x90,0x3E };
+
+unsigned char iso_test_5[WHIRLPOOL_DIGEST_LENGTH] = {
+	0xF1,0xD7,0x54,0x66,0x26,0x36,0xFF,0xE9,
+	0x2C,0x82,0xEB,0xB9,0x21,0x2A,0x48,0x4A,
+	0x8D,0x38,0x63,0x1E,0xAD,0x42,0x38,0xF5,
+	0x44,0x2E,0xE1,0x3B,0x80,0x54,0xE4,0x1B,
+	0x08,0xBF,0x2A,0x92,0x51,0xC3,0x0B,0x6A,
+	0x0B,0x8A,0xAE,0x86,0x17,0x7A,0xB4,0xA6,
+	0xF6,0x8F,0x67,0x3E,0x72,0x07,0x86,0x5D,
+	0x5D,0x98,0x19,0xA3,0xDB,0xA4,0xEB,0x3B };
+
+unsigned char iso_test_6[WHIRLPOOL_DIGEST_LENGTH] = {
+	0xDC,0x37,0xE0,0x08,0xCF,0x9E,0xE6,0x9B,
+	0xF1,0x1F,0x00,0xED,0x9A,0xBA,0x26,0x90,
+	0x1D,0xD7,0xC2,0x8C,0xDE,0xC0,0x66,0xCC,
+	0x6A,0xF4,0x2E,0x40,0xF8,0x2F,0x3A,0x1E,
+	0x08,0xEB,0xA2,0x66,0x29,0x12,0x9D,0x8F,
+	0xB7,0xCB,0x57,0x21,0x1B,0x92,0x81,0xA6,
+	0x55,0x17,0xCC,0x87,0x9D,0x7B,0x96,0x21,
+	0x42,0xC6,0x5F,0x5A,0x7A,0xF0,0x14,0x67 };
+
+unsigned char iso_test_7[WHIRLPOOL_DIGEST_LENGTH] = {
+	0x46,0x6E,0xF1,0x8B,0xAB,0xB0,0x15,0x4D,
+	0x25,0xB9,0xD3,0x8A,0x64,0x14,0xF5,0xC0,
+	0x87,0x84,0x37,0x2B,0xCC,0xB2,0x04,0xD6,
+	0x54,0x9C,0x4A,0xFA,0xDB,0x60,0x14,0x29,
+	0x4D,0x5B,0xD8,0xDF,0x2A,0x6C,0x44,0xE5,
+	0x38,0xCD,0x04,0x7B,0x26,0x81,0xA5,0x1A,
+	0x2C,0x60,0x48,0x1E,0x88,0xC5,0xA2,0x0B,
+	0x2C,0x2A,0x80,0xCF,0x3A,0x9A,0x08,0x3B };
+
+unsigned char iso_test_8[WHIRLPOOL_DIGEST_LENGTH] = {
+	0x2A,0x98,0x7E,0xA4,0x0F,0x91,0x70,0x61,
+	0xF5,0xD6,0xF0,0xA0,0xE4,0x64,0x4F,0x48,
+	0x8A,0x7A,0x5A,0x52,0xDE,0xEE,0x65,0x62,
+	0x07,0xC5,0x62,0xF9,0x88,0xE9,0x5C,0x69,
+	0x16,0xBD,0xC8,0x03,0x1B,0xC5,0xBE,0x1B,
+	0x7B,0x94,0x76,0x39,0xFE,0x05,0x0B,0x56,
+	0x93,0x9B,0xAA,0xA0,0xAD,0xFF,0x9A,0xE6,
+	0x74,0x5B,0x7B,0x18,0x1C,0x3B,0xE3,0xFD };
+
+unsigned char iso_test_9[WHIRLPOOL_DIGEST_LENGTH] = {
+	0x0C,0x99,0x00,0x5B,0xEB,0x57,0xEF,0xF5,
+	0x0A,0x7C,0xF0,0x05,0x56,0x0D,0xDF,0x5D,
+	0x29,0x05,0x7F,0xD8,0x6B,0x20,0xBF,0xD6,
+	0x2D,0xEC,0xA0,0xF1,0xCC,0xEA,0x4A,0xF5,
+	0x1F,0xC1,0x54,0x90,0xED,0xDC,0x47,0xAF,
+	0x32,0xBB,0x2B,0x66,0xC3,0x4F,0xF9,0xAD,
+	0x8C,0x60,0x08,0xAD,0x67,0x7F,0x77,0x12,
+	0x69,0x53,0xB2,0x26,0xE4,0xED,0x8B,0x01 };
+
+int main (int argc,char *argv[])
+{ unsigned char md[WHIRLPOOL_DIGEST_LENGTH];
+  int		i;
+  WHIRLPOOL_CTX	ctx;
+
+#ifdef OPENSSL_IA32_SSE2
+    /* Alternative to this is to call OpenSSL_add_all_algorithms...
+     * The below code is retained exclusively for debugging purposes. */
+    { char      *env;
+
+	if ((env=getenv("OPENSSL_ia32cap")))
+	    OPENSSL_ia32cap = strtoul (env,NULL,0);
+    }
+#endif
+
+    fprintf(stdout,"Testing Whirlpool ");
+
+    WHIRLPOOL("",0,md);
+    if (memcmp(md,iso_test_1,sizeof(iso_test_1)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 1 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    WHIRLPOOL("a",1,md);
+    if (memcmp(md,iso_test_2,sizeof(iso_test_2)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 2 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    WHIRLPOOL("abc",3,md);
+    if (memcmp(md,iso_test_3,sizeof(iso_test_3)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 3 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    WHIRLPOOL("message digest",14,md);
+    if (memcmp(md,iso_test_4,sizeof(iso_test_4)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 4 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    WHIRLPOOL("abcdefghijklmnopqrstuvwxyz",26,md);
+    if (memcmp(md,iso_test_5,sizeof(iso_test_5)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 5 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    WHIRLPOOL(	"ABCDEFGHIJKLMNOPQRSTUVWXYZ"
+		"abcdefghijklmnopqrstuvwxyz"
+		"0123456789",62,md);
+    if (memcmp(md,iso_test_6,sizeof(iso_test_6)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 6 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    WHIRLPOOL(	"1234567890""1234567890""1234567890""1234567890"
+		"1234567890""1234567890""1234567890""1234567890",80,md);
+    if (memcmp(md,iso_test_7,sizeof(iso_test_7)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 7 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    WHIRLPOOL("abcdbcdecdefdefgefghfghighijhijk",32,md);
+    if (memcmp(md,iso_test_8,sizeof(iso_test_8)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 8 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+ 
+    WHIRLPOOL_Init (&ctx);
+    for (i=0;i<1000000;i+=288)
+	WHIRLPOOL_Update (&ctx,	"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa",
+				(1000000-i)<288?1000000-i:288);
+    WHIRLPOOL_Final (md,&ctx);
+    if (memcmp(md,iso_test_9,sizeof(iso_test_9)))
+    {   fflush(stdout);
+	fprintf(stderr,"\nTEST 9 of 9 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    fprintf(stdout," passed.\n"); fflush(stdout);
+
+  return 0;
+}
+#endif

test/Makefile

@@ -44,6 +44,7 @@ MD2TEST=	md2test
 MD4TEST=	md4test
 MD5TEST=	md5test
 HMACTEST=	hmactest
+WPTEST=		wp_test
 RC2TEST=	rc2test
 RC4TEST=	rc4test
 RC5TEST=	rc5test
@@ -62,7 +63,7 @@ EVPTEST=	evp_test
 TESTS=		alltests
 
 EXE=	$(BNTEST)$(EXE_EXT) $(ECTEST)$(EXE_EXT)  $(ECDSATEST)$(EXE_EXT) $(ECDHTEST)$(EXE_EXT) $(IDEATEST)$(EXE_EXT) \
-	$(MD2TEST)$(EXE_EXT)  $(MD4TEST)$(EXE_EXT) $(MD5TEST)$(EXE_EXT) $(HMACTEST)$(EXE_EXT) \
+	$(MD2TEST)$(EXE_EXT)  $(MD4TEST)$(EXE_EXT) $(MD5TEST)$(EXE_EXT) $(HMACTEST)$(EXE_EXT) $(WPTEST)$(EXE_EXT) \
 	$(RC2TEST)$(EXE_EXT) $(RC4TEST)$(EXE_EXT) $(RC5TEST)$(EXE_EXT) \
 	$(DESTEST)$(EXE_EXT) $(SHATEST)$(EXE_EXT) $(SHA1TEST)$(EXE_EXT) $(SHA256TEST)$(EXE_EXT) $(SHA512TEST)$(EXE_EXT) \
 	$(MDC2TEST)$(EXE_EXT) $(RMDTEST)$(EXE_EXT) \
@@ -74,7 +75,7 @@ EXE=	$(BNTEST)$(EXE_EXT) $(ECTEST)$(EXE_EXT)  $(ECDSATEST)$(EXE_EXT) $(ECDHTEST)
 
 OBJ=	$(BNTEST).o $(ECTEST).o  $(ECDSATEST).o $(ECDHTEST).o $(IDEATEST).o \
 	$(MD2TEST).o $(MD4TEST).o $(MD5TEST).o \
-	$(HMACTEST).o \
+	$(HMACTEST).o $(WPTEST).o \
 	$(RC2TEST).o $(RC4TEST).o $(RC5TEST).o \
 	$(DESTEST).o $(SHATEST).o $(SHA1TEST).o $(SHA256TEST).o $(SHA512TEST).o \
 	$(MDC2TEST).o $(RMDTEST).o \
@@ -83,7 +84,7 @@ OBJ=	$(BNTEST).o $(ECTEST).o  $(ECDSATEST).o $(ECDHTEST).o $(IDEATEST).o \
 	$(EVPTEST).o
 SRC=	$(BNTEST).c $(ECTEST).c  $(ECDSATEST).c $(ECDHTEST).c $(IDEATEST).c \
 	$(MD2TEST).c  $(MD4TEST).c $(MD5TEST).c \
-	$(HMACTEST).c \
+	$(HMACTEST).c $(WPTEST).c \
 	$(RC2TEST).c $(RC4TEST).c $(RC5TEST).c \
 	$(DESTEST).c $(SHATEST).c $(SHA1TEST).c $(MDC2TEST).c $(RMDTEST).c \
 	$(RANDTEST).c $(DHTEST).c $(ENGINETEST).c $(CASTTEST).c \
@@ -125,7 +126,7 @@ apps:
 
 alltests: \
 	test_des test_idea test_sha test_md4 test_md5 test_hmac \
-	test_md2 test_mdc2 \
+	test_md2 test_mdc2 test_wp \
 	test_rmd test_rc2 test_rc4 test_rc5 test_bf test_cast test_aes \
 	test_rand test_bn test_ec test_ecdsa test_ecdh \
 	test_enc test_x509 test_rsa test_crl test_sid \
@@ -159,6 +160,9 @@ test_md4:
 test_hmac:
 	../util/shlib_wrap.sh ./$(HMACTEST)
 
+test_wp:
+	../util/shlib_wrap.sh ./$(WPTEST)
+
 test_md2:
 	../util/shlib_wrap.sh ./$(MD2TEST)
 
@@ -363,6 +367,9 @@ $(MD5TEST)$(EXE_EXT): $(MD5TEST).o $(DLIBCRYPTO)
 $(HMACTEST)$(EXE_EXT): $(HMACTEST).o $(DLIBCRYPTO)
 	@target=$(HMACTEST); $(BUILD_CMD)
 
+$(WPTEST)$(EXE_EXT): $(WPTEST).o $(DLIBCRYPTO)
+	@target=$(WPTEST); $(BUILD_CMD)
+
 $(RC2TEST)$(EXE_EXT): $(RC2TEST).o $(DLIBCRYPTO)
 	@target=$(RC2TEST); $(BUILD_CMD)
 
@@ -613,3 +620,8 @@ ssltest.o: ../include/openssl/ssl23.h ../include/openssl/ssl3.h
 ssltest.o: ../include/openssl/stack.h ../include/openssl/symhacks.h
 ssltest.o: ../include/openssl/tls1.h ../include/openssl/x509.h
 ssltest.o: ../include/openssl/x509_vfy.h ../include/openssl/x509v3.h ssltest.c
+wp_test.o: ../include/openssl/crypto.h ../include/openssl/e_os2.h
+wp_test.o: ../include/openssl/opensslconf.h ../include/openssl/opensslv.h
+wp_test.o: ../include/openssl/ossl_typ.h ../include/openssl/safestack.h
+wp_test.o: ../include/openssl/stack.h ../include/openssl/symhacks.h
+wp_test.o: ../include/openssl/whrlpool.h wp_test.c