bn_mul.h

一个用纯C写的RSA_SHA1数字签名库

/**
 * \file bn_mul.h
 */
/*
 *      Multiply source vector [s] with b, add result
 *       to destination vector [d] and set carry c.
 *
 *      Currently supports:
 *
 *         . IA-32 (386+)         . AMD64 / EM64T
 *         . IA-32 (SSE2)         . Motorola 68000
 *         . PowerPC, 32-bit      . MicroBlaze
 *         . PowerPC, 64-bit      . TriCore
 *         . SPARC v8             . ARM v3+
 *         . Alpha                . MIPS32
 *         . C, longlong          . C, generic
 */
#ifndef XYSSL_BN_MUL_H
#define XYSSL_BN_MUL_H

#include "config.h"

#if defined(XYSSL_HAVE_ASM)

#if defined(__GNUC__)
#if defined(__i386__)

#define MULADDC_INIT                            \
    asm( "movl   %%ebx, %0      " : "=m" (t));  \
    asm( "movl   %0, %%esi      " :: "m" (s));  \
    asm( "movl   %0, %%edi      " :: "m" (d));  \
    asm( "movl   %0, %%ecx      " :: "m" (c));  \
    asm( "movl   %0, %%ebx      " :: "m" (b));

#define MULADDC_CORE                            \
    asm( "lodsl                 " );            \
    asm( "mull   %ebx           " );            \
    asm( "addl   %ecx,   %eax   " );            \
    asm( "adcl   $0,     %edx   " );            \
    asm( "addl   (%edi), %eax   " );            \
    asm( "adcl   $0,     %edx   " );            \
    asm( "movl   %edx,   %ecx   " );            \
    asm( "stosl                 " );

#define MULADDC_STOP                            \
    asm( "movl   %0, %%ebx      " :: "m" (t));  \
    asm( "movl   %%ecx, %0      " : "=m" (c));  \
    asm( "movl   %%edi, %0      " : "=m" (d));  \
    asm( "movl   %%esi, %0      " : "=m" (s) :: \
    "eax", "ecx", "edx", "esi", "edi" );

#if defined(XYSSL_HAVE_SSE2)

#define MULADDC_HUIT                            \
    asm( "movd     %ecx,     %mm1     " );      \
    asm( "movd     %ebx,     %mm0     " );      \
    asm( "movd     (%edi),   %mm3     " );      \
    asm( "paddq    %mm3,     %mm1     " );      \
    asm( "movd     (%esi),   %mm2     " );      \
    asm( "pmuludq  %mm0,     %mm2     " );      \
    asm( "movd     4(%esi),  %mm4     " );      \
    asm( "pmuludq  %mm0,     %mm4     " );      \
    asm( "movd     8(%esi),  %mm6     " );      \
    asm( "pmuludq  %mm0,     %mm6     " );      \
    asm( "movd     12(%esi), %mm7     " );      \
    asm( "pmuludq  %mm0,     %mm7     " );      \
    asm( "paddq    %mm2,     %mm1     " );      \
    asm( "movd     4(%edi),  %mm3     " );      \
    asm( "paddq    %mm4,     %mm3     " );      \
    asm( "movd     8(%edi),  %mm5     " );      \
    asm( "paddq    %mm6,     %mm5     " );      \
    asm( "movd     12(%edi), %mm4     " );      \
    asm( "paddq    %mm4,     %mm7     " );      \
    asm( "movd     %mm1,     (%edi)   " );      \
    asm( "movd     16(%esi), %mm2     " );      \
    asm( "pmuludq  %mm0,     %mm2     " );      \
    asm( "psrlq    $32,      %mm1     " );      \
    asm( "movd     20(%esi), %mm4     " );      \
    asm( "pmuludq  %mm0,     %mm4     " );      \
    asm( "paddq    %mm3,     %mm1     " );      \
    asm( "movd     24(%esi), %mm6     " );      \
    asm( "pmuludq  %mm0,     %mm6     " );      \
    asm( "movd     %mm1,     4(%edi)  " );      \
    asm( "psrlq    $32,      %mm1     " );      \
    asm( "movd     28(%esi), %mm3     " );      \
    asm( "pmuludq  %mm0,     %mm3     " );      \
    asm( "paddq    %mm5,     %mm1     " );      \
    asm( "movd     16(%edi), %mm5     " );      \
    asm( "paddq    %mm5,     %mm2     " );      \
    asm( "movd     %mm1,     8(%edi)  " );      \
    asm( "psrlq    $32,      %mm1     " );      \
    asm( "paddq    %mm7,     %mm1     " );      \
    asm( "movd     20(%edi), %mm5     " );      \
    asm( "paddq    %mm5,     %mm4     " );      \
    asm( "movd     %mm1,     12(%edi) " );      \
    asm( "psrlq    $32,      %mm1     " );      \
    asm( "paddq    %mm2,     %mm1     " );      \
    asm( "movd     24(%edi), %mm5     " );      \
    asm( "paddq    %mm5,     %mm6     " );      \
    asm( "movd     %mm1,     16(%edi) " );      \
    asm( "psrlq    $32,      %mm1     " );      \
    asm( "paddq    %mm4,     %mm1     " );      \
    asm( "movd     28(%edi), %mm5     " );      \
    asm( "paddq    %mm5,     %mm3     " );      \
    asm( "movd     %mm1,     20(%edi) " );      \
    asm( "psrlq    $32,      %mm1     " );      \
    asm( "paddq    %mm6,     %mm1     " );      \
    asm( "movd     %mm1,     24(%edi) " );      \
    asm( "psrlq    $32,      %mm1     " );      \
    asm( "paddq    %mm3,     %mm1     " );      \
    asm( "movd     %mm1,     28(%edi) " );      \
    asm( "addl     $32,      %edi     " );      \
    asm( "addl     $32,      %esi     " );      \
    asm( "psrlq    $32,      %mm1     " );      \
    asm( "movd     %mm1,     %ecx     " );

#endif /* SSE2 */
#endif /* i386 */

#if defined(__amd64__) || defined (__x86_64__)

#define MULADDC_INIT                            \
    asm( "movq   %0, %%rsi      " :: "m" (s));  \
    asm( "movq   %0, %%rdi      " :: "m" (d));  \
    asm( "movq   %0, %%rcx      " :: "m" (c));  \
    asm( "movq   %0, %%rbx      " :: "m" (b));  \
    asm( "xorq   %r8, %r8       " );

#define MULADDC_CORE                            \
    asm( "movq  (%rsi),%rax     " );            \
    asm( "mulq   %rbx           " );            \
    asm( "addq   $8,   %rsi     " );            \
    asm( "addq   %rcx, %rax     " );            \
    asm( "movq   %r8,  %rcx     " );            \
    asm( "adcq   $0,   %rdx     " );            \
    asm( "nop                   " );            \
    asm( "addq   %rax, (%rdi)   " );            \
    asm( "adcq   %rdx, %rcx     " );            \
    asm( "addq   $8,   %rdi     " );

#define MULADDC_STOP                            \
    asm( "movq   %%rcx, %0      " : "=m" (c));  \
    asm( "movq   %%rdi, %0      " : "=m" (d));  \
    asm( "movq   %%rsi, %0      " : "=m" (s) :: \
    "rax", "rcx", "rdx", "rbx", "rsi", "rdi", "r8" );

#endif /* AMD64 */

#if defined(__mc68020__) || defined(__mcpu32__)

#define MULADDC_INIT                            \
    asm( "movl   %0, %%a2       " :: "m" (s));  \
    asm( "movl   %0, %%a3       " :: "m" (d));  \
    asm( "movl   %0, %%d3       " :: "m" (c));  \
    asm( "movl   %0, %%d2       " :: "m" (b));  \
    asm( "moveq  #0, %d0        " );

#define MULADDC_CORE                            \
    asm( "movel  %a2@+, %d1     " );            \
    asm( "mulul  %d2, %d4:%d1   " );            \
    asm( "addl   %d3, %d1       " );            \
    asm( "addxl  %d0, %d4       " );            \
    asm( "moveq  #0,  %d3       " );            \
    asm( "addl   %d1, %a3@+     " );            \
    asm( "addxl  %d4, %d3       " );

#define MULADDC_STOP                            \
    asm( "movl   %%d3, %0       " : "=m" (c));  \
    asm( "movl   %%a3, %0       " : "=m" (d));  \
    asm( "movl   %%a2, %0       " : "=m" (s) :: \
    "d0", "d1", "d2", "d3", "d4", "a2", "a3" );

#define MULADDC_HUIT                            \
    asm( "movel  %a2@+, %d1     " );            \
    asm( "mulul  %d2, %d4:%d1   " );            \
    asm( "addxl  %d3, %d1       " );            \
    asm( "addxl  %d0, %d4       " );            \
    asm( "addl   %d1, %a3@+     " );            \
    asm( "movel  %a2@+, %d1     " );            \
    asm( "mulul  %d2, %d3:%d1   " );            \
    asm( "addxl  %d4, %d1       " );            \
    asm( "addxl  %d0, %d3       " );            \
    asm( "addl   %d1, %a3@+     " );            \
    asm( "movel  %a2@+, %d1     " );            \
    asm( "mulul  %d2, %d4:%d1   " );            \
    asm( "addxl  %d3, %d1       " );            \
    asm( "addxl  %d0, %d4       " );            \
    asm( "addl   %d1, %a3@+     " );            \
    asm( "movel  %a2@+, %d1     " );            \
    asm( "mulul  %d2, %d3:%d1   " );            \
    asm( "addxl  %d4, %d1       " );            \
    asm( "addxl  %d0, %d3       " );            \
    asm( "addl   %d1, %a3@+     " );            \
    asm( "movel  %a2@+, %d1     " );            \
    asm( "mulul  %d2, %d4:%d1   " );            \
    asm( "addxl  %d3, %d1       " );            \
    asm( "addxl  %d0, %d4       " );            \
    asm( "addl   %d1, %a3@+     " );            \
    asm( "movel  %a2@+, %d1     " );            \
    asm( "mulul  %d2, %d3:%d1   " );            \
    asm( "addxl  %d4, %d1       " );            \
    asm( "addxl  %d0, %d3       " );            \
    asm( "addl   %d1, %a3@+     " );            \
    asm( "movel  %a2@+, %d1     " );            \
    asm( "mulul  %d2, %d4:%d1   " );            \
    asm( "addxl  %d3, %d1       " );            \
    asm( "addxl  %d0, %d4       " );            \
    asm( "addl   %d1, %a3@+     " );            \
    asm( "movel  %a2@+, %d1     " );            \
    asm( "mulul  %d2, %d3:%d1   " );            \
    asm( "addxl  %d4, %d1       " );            \
    asm( "addxl  %d0, %d3       " );            \
    asm( "addl   %d1, %a3@+     " );            \
    asm( "addxl  %d0, %d3       " );

#endif /* MC68000 */

#if defined(__powerpc__)   || defined(__ppc__)
#if defined(__powerpc64__) || defined(__ppc64__)

#if defined(__MACH__) && defined(__APPLE__)

#define MULADDC_INIT                            \
    asm( "ld     r3, %0         " :: "m" (s));  \
    asm( "ld     r4, %0         " :: "m" (d));  \
    asm( "ld     r5, %0         " :: "m" (c));  \
    asm( "ld     r6, %0         " :: "m" (b));  \
    asm( "addi   r3, r3, -8     " );            \
    asm( "addi   r4, r4, -8     " );            \
    asm( "addic  r5, r5,  0     " );

#define MULADDC_CORE                            \
    asm( "ldu    r7, 8(r3)      " );            \
    asm( "mulld  r8, r7, r6     " );            \
    asm( "mulhdu r9, r7, r6     " );            \
    asm( "adde   r8, r8, r5     " );            \
    asm( "ld     r7, 8(r4)      " );            \
    asm( "addze  r5, r9         " );            \
    asm( "addc   r8, r8, r7     " );            \
    asm( "stdu   r8, 8(r4)      " );

#define MULADDC_STOP                            \
    asm( "addze  r5, r5         " );            \
    asm( "addi   r4, r4, 8      " );            \
    asm( "addi   r3, r3, 8      " );            \
    asm( "std    r5, %0         " : "=m" (c));  \
    asm( "std    r4, %0         " : "=m" (d));  \
    asm( "std    r3, %0         " : "=m" (s) :: \
    "r3", "r4", "r5", "r6", "r7", "r8", "r9" );

#else

#define MULADDC_INIT                            \
    asm( "ld     %%r3, %0       " :: "m" (s));  \
    asm( "ld     %%r4, %0       " :: "m" (d));  \
    asm( "ld     %%r5, %0       " :: "m" (c));  \
    asm( "ld     %%r6, %0       " :: "m" (b));  \
    asm( "addi   %r3, %r3, -8   " );            \
    asm( "addi   %r4, %r4, -8   " );            \
    asm( "addic  %r5, %r5,  0   " );

#define MULADDC_CORE                            \
    asm( "ldu    %r7, 8(%r3)    " );            \
    asm( "mulld  %r8, %r7, %r6  " );            \
    asm( "mulhdu %r9, %r7, %r6  " );            \
    asm( "adde   %r8, %r8, %r5  " );            \
    asm( "ld     %r7, 8(%r4)    " );            \
    asm( "addze  %r5, %r9       " );            \
    asm( "addc   %r8, %r8, %r7  " );            \
    asm( "stdu   %r8, 8(%r4)    " );

#define MULADDC_STOP                            \
    asm( "addze  %r5, %r5       " );            \
    asm( "addi   %r4, %r4, 8    " );            \
    asm( "addi   %r3, %r3, 8    " );            \
    asm( "std    %%r5, %0       " : "=m" (c));  \
    asm( "std    %%r4, %0       " : "=m" (d));  \
    asm( "std    %%r3, %0       " : "=m" (s) :: \
    "r3", "r4", "r5", "r6", "r7", "r8", "r9" );

#endif

#else /* PPC32 */

#if defined(__MACH__) && defined(__APPLE__)

#define MULADDC_INIT                            \
    asm( "lwz    r3, %0         " :: "m" (s));  \
    asm( "lwz    r4, %0         " :: "m" (d));  \
    asm( "lwz    r5, %0         " :: "m" (c));  \
    asm( "lwz    r6, %0         " :: "m" (b));  \
    asm( "addi   r3, r3, -4     " );            \
    asm( "addi   r4, r4, -4     " );            \
    asm( "addic  r5, r5,  0     " );

#define MULADDC_CORE                            \
    asm( "lwzu   r7, 4(r3)      " );            \
    asm( "mullw  r8, r7, r6     " );            \
    asm( "mulhwu r9, r7, r6     " );            \
    asm( "adde   r8, r8, r5     " );            \
    asm( "lwz    r7, 4(r4)      " );            \
    asm( "addze  r5, r9         " );            \
    asm( "addc   r8, r8, r7     " );            \
    asm( "stwu   r8, 4(r4)      " );

#define MULADDC_STOP                            \
    asm( "addze  r5, r5         " );            \
    asm( "addi   r4, r4, 4      " );            \
    asm( "addi   r3, r3, 4      " );            \
    asm( "stw    r5, %0         " : "=m" (c));  \
    asm( "stw    r4, %0         " : "=m" (d));  \
    asm( "stw    r3, %0         " : "=m" (s) :: \
    "r3", "r4", "r5", "r6", "r7", "r8", "r9" );

#else

#define MULADDC_INIT                            \
    asm( "lwz    %%r3, %0       " :: "m" (s));  \
    asm( "lwz    %%r4, %0       " :: "m" (d));  \
    asm( "lwz    %%r5, %0       " :: "m" (c));  \
    asm( "lwz    %%r6, %0       " :: "m" (b));  \
    asm( "addi   %r3, %r3, -4   " );            \
    asm( "addi   %r4, %r4, -4   " );            \
    asm( "addic  %r5, %r5,  0   " );

#define MULADDC_CORE                            \
    asm( "lwzu   %r7, 4(%r3)    " );            \
    asm( "mullw  %r8, %r7, %r6  " );            \
    asm( "mulhwu %r9, %r7, %r6  " );            \
    asm( "adde   %r8, %r8, %r5  " );            \
    asm( "lwz    %r7, 4(%r4)    " );            \
    asm( "addze  %r5, %r9       " );            \
    asm( "addc   %r8, %r8, %r7  " );            \
    asm( "stwu   %r8, 4(%r4)    " );

#define MULADDC_STOP                            \
    asm( "addze  %r5, %r5       " );            \
    asm( "addi   %r4, %r4, 4    " );            \
    asm( "addi   %r3, %r3, 4    " );            \
    asm( "stw    %%r5, %0       " : "=m" (c));  \
    asm( "stw    %%r4, %0       " : "=m" (d));  \
    asm( "stw    %%r3, %0       " : "=m" (s) :: \