From ae12b879721b02b79a2a2457b36627d24147e74f Mon Sep 17 00:00:00 2001 From: wkj Date: Wed, 21 Apr 2004 04:52:27 +0000 Subject: MD5 and SHA1 in transliterated to x86 assembler with C preprocessor goo. Thanks to Brendan Connell for putting up with this nonsense. --- src/libsec/386/md5block.s | 241 +++++++++++++++++++++++++++++++++++++++++++++ src/libsec/386/mkfile | 16 +++ src/libsec/386/sha1block.s | 214 ++++++++++++++++++++++++++++++++++++++++ 3 files changed, 471 insertions(+) create mode 100644 src/libsec/386/md5block.s create mode 100644 src/libsec/386/mkfile create mode 100644 src/libsec/386/sha1block.s (limited to 'src/libsec/386') diff --git a/src/libsec/386/md5block.s b/src/libsec/386/md5block.s new file mode 100644 index 00000000..602c970d --- /dev/null +++ b/src/libsec/386/md5block.s @@ -0,0 +1,241 @@ +/* + * rfc1321 requires that I include this. The code is new. The constants + * all come from the rfc (hence the copyright). We trade a table for the + * macros in rfc. The total size is a lot less. -- presotto + * + * Copyright (C) 1991-2, RSA Data Security, Inc. Created 1991. All + * rights reserved. + * + * License to copy and use this software is granted provided that it + * is identified as the "RSA Data Security, Inc. MD5 Message-Digest + * Algorithm" in all material mentioning or referencing this software + * or this function. + * + * License is also granted to make and use derivative works provided + * that such works are identified as "derived from the RSA Data + * Security, Inc. MD5 Message-Digest Algorithm" in all material + * mentioning or referencing the derived work. + * + * RSA Data Security, Inc. makes no representations concerning either + * the merchantability of this software or the suitability of this + * software forany particular purpose. It is provided "as is" + * without express or implied warranty of any kind. + * These notices must be retained in any copies of any part of this + * documentation and/or software. + */ +#define S11 7 +#define S12 12 +#define S13 17 +#define S14 22 + +#define S21 5 +#define S22 9 +#define S23 14 +#define S24 20 + +#define S31 4 +#define S32 11 +#define S33 16 +#define S34 23 + +#define S41 6 +#define S42 10 +#define S43 15 +#define S44 21 + +#define PAYME(x) $##x + +/* + * SI is data + * a += FN(B,C,D); + * a += x[sh] + t[sh]; + * a = (a << S11) | (a >> (32 - S11)); + * a += b; + */ + +#define BODY1(off,V,FN,SH,A,B,C,D)\ + FN(B,C,D)\ + leal V(A, %edi, 1), A;\ + addl off(%ebp), A;\ + roll PAYME(SH), A;\ + addl B, A;\ + +#define BODY(off,V,FN,SH,A,B,C,D)\ + FN(B,C,D)\ + leal V(A, %edi, 1), A;\ + addl (off)(%ebp), A;\ + roll PAYME(SH), A;\ + addl B,A;\ + +/* + * fn1 = ((c ^ d) & b) ^ d + */ +#define FN1(B,C,D)\ + movl C, %edi;\ + xorl D, %edi;\ + andl B, %edi;\ + xorl D, %edi;\ + +/* + * fn2 = ((b ^ c) & d) ^ c; + */ +#define FN2(B,C,D)\ + movl B, %edi;\ + xorl C, %edi;\ + andl D, %edi;\ + xorl C, %edi;\ + +/* + * fn3 = b ^ c ^ d; + */ +#define FN3(B,C,D)\ + movl B, %edi;\ + xorl C, %edi;\ + xorl D, %edi;\ + +/* + * fn4 = c ^ (b | ~d); + */ +#define FN4(B,C,D)\ + movl D, %edi;\ + xorl $-1, %edi;\ + orl B, %edi;\ + xorl C, %edi;\ + +#define DATA 8 +#define LEN 12 +#define STATE 16 + +#define EDATA (-4) +#define OLDEBX (-8) +#define OLDESI (-12) +#define OLDEDI (-16) + + .text + + .p2align 2,0x90 + .globl _md5block + .type _md5block, @function + _md5block: + + /* Prelude */ + pushl %ebp + movl %ebx, OLDEBX(%esp) + movl %esi, OLDESI(%esp) + movl %edi, OLDEDI(%esp) + + movl DATA(%esp), %eax + addl LEN(%esp), %eax + movl %eax, EDATA(%esp) + + movl DATA(%esp), %ebp + +mainloop: + movl STATE(%esp), %esi + movl (%esi), %eax + movl 4(%esi), %ebx + movl 8(%esi), %ecx + movl 12(%esi), %edx + + BODY1( 0*4,0xd76aa478,FN1,S11,%eax,%ebx,%ecx,%edx) + BODY1( 1*4,0xe8c7b756,FN1,S12,%edx,%eax,%ebx,%ecx) + BODY1( 2*4,0x242070db,FN1,S13,%ecx,%edx,%eax,%ebx) + BODY1( 3*4,0xc1bdceee,FN1,S14,%ebx,%ecx,%edx,%eax) + + BODY1( 4*4,0xf57c0faf,FN1,S11,%eax,%ebx,%ecx,%edx) + BODY1( 5*4,0x4787c62a,FN1,S12,%edx,%eax,%ebx,%ecx) + BODY1( 6*4,0xa8304613,FN1,S13,%ecx,%edx,%eax,%ebx) + BODY1( 7*4,0xfd469501,FN1,S14,%ebx,%ecx,%edx,%eax) + + BODY1( 8*4,0x698098d8,FN1,S11,%eax,%ebx,%ecx,%edx) + BODY1( 9*4,0x8b44f7af,FN1,S12,%edx,%eax,%ebx,%ecx) + BODY1(10*4,0xffff5bb1,FN1,S13,%ecx,%edx,%eax,%ebx) + BODY1(11*4,0x895cd7be,FN1,S14,%ebx,%ecx,%edx,%eax) + + BODY1(12*4,0x6b901122,FN1,S11,%eax,%ebx,%ecx,%edx) + BODY1(13*4,0xfd987193,FN1,S12,%edx,%eax,%ebx,%ecx) + BODY1(14*4,0xa679438e,FN1,S13,%ecx,%edx,%eax,%ebx) + BODY1(15*4,0x49b40821,FN1,S14,%ebx,%ecx,%edx,%eax) + + + BODY( 1*4,0xf61e2562,FN2,S21,%eax,%ebx,%ecx,%edx) + BODY( 6*4,0xc040b340,FN2,S22,%edx,%eax,%ebx,%ecx) + BODY(11*4,0x265e5a51,FN2,S23,%ecx,%edx,%eax,%ebx) + BODY( 0*4,0xe9b6c7aa,FN2,S24,%ebx,%ecx,%edx,%eax) + + BODY( 5*4,0xd62f105d,FN2,S21,%eax,%ebx,%ecx,%edx) + BODY(10*4,0x02441453,FN2,S22,%edx,%eax,%ebx,%ecx) + BODY(15*4,0xd8a1e681,FN2,S23,%ecx,%edx,%eax,%ebx) + BODY( 4*4,0xe7d3fbc8,FN2,S24,%ebx,%ecx,%edx,%eax) + + BODY( 9*4,0x21e1cde6,FN2,S21,%eax,%ebx,%ecx,%edx) + BODY(14*4,0xc33707d6,FN2,S22,%edx,%eax,%ebx,%ecx) + BODY( 3*4,0xf4d50d87,FN2,S23,%ecx,%edx,%eax,%ebx) + BODY( 8*4,0x455a14ed,FN2,S24,%ebx,%ecx,%edx,%eax) + + BODY(13*4,0xa9e3e905,FN2,S21,%eax,%ebx,%ecx,%edx) + BODY( 2*4,0xfcefa3f8,FN2,S22,%edx,%eax,%ebx,%ecx) + BODY( 7*4,0x676f02d9,FN2,S23,%ecx,%edx,%eax,%ebx) + BODY(12*4,0x8d2a4c8a,FN2,S24,%ebx,%ecx,%edx,%eax) + + + BODY( 5*4,0xfffa3942,FN3,S31,%eax,%ebx,%ecx,%edx) + BODY( 8*4,0x8771f681,FN3,S32,%edx,%eax,%ebx,%ecx) + BODY(11*4,0x6d9d6122,FN3,S33,%ecx,%edx,%eax,%ebx) + BODY(14*4,0xfde5380c,FN3,S34,%ebx,%ecx,%edx,%eax) + + BODY( 1*4,0xa4beea44,FN3,S31,%eax,%ebx,%ecx,%edx) + BODY( 4*4,0x4bdecfa9,FN3,S32,%edx,%eax,%ebx,%ecx) + BODY( 7*4,0xf6bb4b60,FN3,S33,%ecx,%edx,%eax,%ebx) + BODY(10*4,0xbebfbc70,FN3,S34,%ebx,%ecx,%edx,%eax) + + BODY(13*4,0x289b7ec6,FN3,S31,%eax,%ebx,%ecx,%edx) + BODY( 0*4,0xeaa127fa,FN3,S32,%edx,%eax,%ebx,%ecx) + BODY( 3*4,0xd4ef3085,FN3,S33,%ecx,%edx,%eax,%ebx) + BODY( 6*4,0x04881d05,FN3,S34,%ebx,%ecx,%edx,%eax) + + BODY( 9*4,0xd9d4d039,FN3,S31,%eax,%ebx,%ecx,%edx) + BODY(12*4,0xe6db99e5,FN3,S32,%edx,%eax,%ebx,%ecx) + BODY(15*4,0x1fa27cf8,FN3,S33,%ecx,%edx,%eax,%ebx) + BODY( 2*4,0xc4ac5665,FN3,S34,%ebx,%ecx,%edx,%eax) + + + BODY( 0*4,0xf4292244,FN4,S41,%eax,%ebx,%ecx,%edx) + BODY( 7*4,0x432aff97,FN4,S42,%edx,%eax,%ebx,%ecx) + BODY(14*4,0xab9423a7,FN4,S43,%ecx,%edx,%eax,%ebx) + BODY( 5*4,0xfc93a039,FN4,S44,%ebx,%ecx,%edx,%eax) + + BODY(12*4,0x655b59c3,FN4,S41,%eax,%ebx,%ecx,%edx) + BODY( 3*4,0x8f0ccc92,FN4,S42,%edx,%eax,%ebx,%ecx) + BODY(10*4,0xffeff47d,FN4,S43,%ecx,%edx,%eax,%ebx) + BODY( 1*4,0x85845dd1,FN4,S44,%ebx,%ecx,%edx,%eax) + + BODY( 8*4,0x6fa87e4f,FN4,S41,%eax,%ebx,%ecx,%edx) + BODY(15*4,0xfe2ce6e0,FN4,S42,%edx,%eax,%ebx,%ecx) + BODY( 6*4,0xa3014314,FN4,S43,%ecx,%edx,%eax,%ebx) + BODY(13*4,0x4e0811a1,FN4,S44,%ebx,%ecx,%edx,%eax) + + BODY( 4*4,0xf7537e82,FN4,S41,%eax,%ebx,%ecx,%edx) + BODY(11*4,0xbd3af235,FN4,S42,%edx,%eax,%ebx,%ecx) + BODY( 2*4,0x2ad7d2bb,FN4,S43,%ecx,%edx,%eax,%ebx) + BODY( 9*4,0xeb86d391,FN4,S44,%ebx,%ecx,%edx,%eax) + + addl $(16*4), %ebp + movl STATE(%esp), %edi + addl %eax,0(%edi) + addl %ebx,4(%edi) + addl %ecx,8(%edi) + addl %edx,12(%edi) + + movl EDATA(%esp), %edi + cmpl %edi, %ebp + jb mainloop + + /* Postlude */ + movl OLDEBX(%esp), %ebx + movl OLDESI(%esp), %esi + movl OLDEDI(%esp), %edi + movl %esp, %ebp + leave + ret + diff --git a/src/libsec/386/mkfile b/src/libsec/386/mkfile new file mode 100644 index 00000000..b3d560ad --- /dev/null +++ b/src/libsec/386/mkfile @@ -0,0 +1,16 @@ +<$PLAN9/src/mkhdr + +LIB=libsec.a +SFILES=\ + md5block.s\ + sha1block.s\ + +HFILES=$PLAN9/include/libsec.h + +OFILES=${SFILES:%.s=%.$O} + +UPDATE=mkfile\ + $HFILES\ + $SFILES\ + +<$PLAN9/src/mksyslib diff --git a/src/libsec/386/sha1block.s b/src/libsec/386/sha1block.s new file mode 100644 index 00000000..7d0696d0 --- /dev/null +++ b/src/libsec/386/sha1block.s @@ -0,0 +1,214 @@ +.text + +.p2align 2,0x90 +.globl _sha1block + .type _sha1block, @function +_sha1block: + +/* x = (wp[off-f] ^ wp[off-8] ^ wp[off-14] ^ wp[off-16]) <<< 1; + * wp[off] = x; + * x += A <<< 5; + * E += 0xca62c1d6 + x; + * x = FN(B,C,D); + * E += x; + * B >>> 2 + */ +#define BSWAPDI BYTE $0x0f; BYTE $0xcf; + +#define BODY(off,FN,V,A,B,C,D,E)\ + movl (off-64)(%ebp), %edi;\ + xorl (off-56)(%ebp), %edi;\ + xorl (off-32)(%ebp), %edi;\ + xorl (off-12)(%ebp), %edi;\ + roll $1, %edi;\ + movl %edi, off(%ebp);\ + leal V(%edi, E, 1), E;\ + movl A, %edi;\ + roll $5, %edi;\ + addl %edi, E;\ + FN(B,C,D)\ + addl %edi, E;\ + rorl $2, B;\ + +#define BODY0(off,FN,V,A,B,C,D,E)\ + movl off(%ebx), %edi;\ + bswap %edi;\ + movl %edi, off(%ebp);\ + leal V(%edi,E,1), E;\ + movl A, %edi;\ + roll $5,%edi;\ + addl %edi,E;\ + FN(B,C,D)\ + addl %edi,E;\ + rorl $2,B;\ + +/* + * fn1 = (((C^D)&B)^D); + */ +#define FN1(B,C,D)\ + movl C, %edi;\ + xorl D, %edi;\ + andl B, %edi;\ + xorl D, %edi;\ + +/* + * fn24 = B ^ C ^ D + */ +#define FN24(B,C,D)\ + movl B, %edi;\ + xorl C, %edi;\ + xorl D, %edi;\ + +/* + * fn3 = ((B ^ C) & (D ^= B)) ^ B + * D ^= B to restore D + */ +#define FN3(B,C,D)\ + movl B, %edi;\ + xorl C, %edi;\ + xorl B, D;\ + andl D, %edi;\ + xorl B, %edi;\ + xorl B, D;\ + +/* + * stack offsets + * void sha1block(uchar *DATA, int LEN, ulong *STATE) + */ +#define DATA 8 +#define LEN 12 +#define STATE 16 + +/* + * stack offsets for locals + * ulong w[80]; + * uchar *edata; + * ulong *w15, *w40, *w60, *w80; + * register local + * ulong *wp = %ebp + * ulong a = eax, b = ebx, c = ecx, d = edx, e = esi + * ulong tmp = edi + */ +#define WARRAY (-4-(80*4)) +#define TMP1 (-8-(80*4)) +#define TMP2 (-12-(80*4)) +#define W15 (-16-(80*4)) +#define W40 (-20-(80*4)) +#define W60 (-24-(80*4)) +#define W80 (-28-(80*4)) +#define EDATA (-32-(80*4)) +#define OLDEBX (-36-(80*4)) +#define OLDESI (-40-(80*4)) +#define OLDEDI (-44-(80*4)) + + /* Prelude */ + pushl %ebp + mov %ebx, OLDEBX(%esp) + mov %esi, OLDESI(%esp) + mov %edi, OLDEDI(%esp) + + movl DATA(%esp), %eax + addl LEN(%esp), %eax + movl %eax, EDATA(%esp) + + leal (WARRAY+15*4)(%esp), %edi /* aw15 */ + movl %edi, W15(%esp) + leal (WARRAY+40*4)(%esp), %edx /* aw40 */ + movl %edx, W40(%esp) + leal (WARRAY+60*4)(%esp), %ecx /* aw60 */ + movl %ecx, W60(%esp) + leal (WARRAY+80*4)(%esp), %edi /* aw80 */ + movl %edi, W80(%esp) + +mainloop: + leal WARRAY(%esp), %ebp /* warray */ + + movl STATE(%esp), %edi /* state */ + movl (%edi),%eax + movl 4(%edi),%ebx + movl %ebx, TMP1(%esp) /* tmp1 */ + movl 8(%edi), %ecx + movl 12(%edi), %edx + movl 16(%edi), %esi + + movl DATA(%esp), %ebx /* data */ + +loop1: + BODY0(0,FN1,0x5a827999,%eax,TMP1(%esp),%ecx,%edx,%esi) + movl %esi,TMP2(%esp) + BODY0(4,FN1,0x5a827999,%esi,%eax,TMP1(%esp),%ecx,%edx) + movl TMP1(%esp),%esi + BODY0(8,FN1,0x5a827999,%edx,TMP2(%esp),%eax,%esi,%ecx) + BODY0(12,FN1,0x5a827999,%ecx,%edx,TMP2(%esp),%eax,%esi) + movl %esi,TMP1(%esp) + BODY0(16,FN1,0x5a827999,%esi,%ecx,%edx,TMP2(%esp),%eax) + movl TMP2(%esp),%esi + + addl $20, %ebx + addl $20, %ebp + cmpl W15(%esp), %ebp /* w15 */ + jb loop1 + + BODY0(0,FN1,0x5a827999,%eax,TMP1(%esp),%ecx,%edx,%esi) + addl $4, %ebx + MOVL %ebx, DATA(%esp) /* data */ + MOVL TMP1(%esp),%ebx + + BODY(4,FN1,0x5a827999,%esi,%eax,%ebx,%ecx,%edx) + BODY(8,FN1,0x5a827999,%edx,%esi,%eax,%ebx,%ecx) + BODY(12,FN1,0x5a827999,%ecx,%edx,%esi,%eax,%ebx) + BODY(16,FN1,0x5a827999,%ebx,%ecx,%edx,%esi,%eax) + + addl $20, %ebp + +loop2: + BODY(0,FN24,0x6ed9eba1,%eax,%ebx,%ecx,%edx,%esi) + BODY(4,FN24,0x6ed9eba1,%esi,%eax,%ebx,%ecx,%edx) + BODY(8,FN24,0x6ed9eba1,%edx,%esi,%eax,%ebx,%ecx) + BODY(12,FN24,0x6ed9eba1,%ecx,%edx,%esi,%eax,%ebx) + BODY(16,FN24,0x6ed9eba1,%ebx,%ecx,%edx,%esi,%eax) + + addl $20,%ebp + cmpl W40(%esp), %ebp + jb loop2 + +loop3: + BODY(0,FN3,0x8f1bbcdc,%eax,%ebx,%ecx,%edx,%esi) + BODY(4,FN3,0x8f1bbcdc,%esi,%eax,%ebx,%ecx,%edx) + BODY(8,FN3,0x8f1bbcdc,%edx,%esi,%eax,%ebx,%ecx) + BODY(12,FN3,0x8f1bbcdc,%ecx,%edx,%esi,%eax,%ebx) + BODY(16,FN3,0x8f1bbcdc,%ebx,%ecx,%edx,%esi,%eax) + + addl $20, %ebp + cmpl W60(%esp), %ebp /* w60 */ + jb loop3 + +loop4: + BODY(0,FN24,0xca62c1d6,%eax,%ebx,%ecx,%edx,%esi) + BODY(4,FN24,0xca62c1d6,%esi,%eax,%ebx,%ecx,%edx) + BODY(8,FN24,0xca62c1d6,%edx,%esi,%eax,%ebx,%ecx) + BODY(12,FN24,0xca62c1d6,%ecx,%edx,%esi,%eax,%ebx) + BODY(16,FN24,0xca62c1d6,%ebx,%ecx,%edx,%esi,%eax) + + addl $20, %ebp + cmpl W80(%esp), %ebp /* w80 */ + jb loop4 + + movl STATE(%esp), %edi /* state */ + addl %eax, 0(%edi) + addl %ebx, 4(%edi) + addl %ecx, 8(%edi) + addl %edx, 12(%edi) + addl %esi, 16(%edi) + + movl EDATA(%esp), %edi /* edata */ + cmpl %edi, DATA(%esp) /* data */ + jb mainloop + + /* Postlude */ + mov OLDEBX(%esp), %ebx + mov OLDESI(%esp), %esi + mov OLDEDI(%esp), %edi + movl %esp, %ebp + leave + ret -- cgit v1.2.3