Return to add_n.S CVS log

Up to [local] / OpenXM_contrib / gmp / mpn / x86

Annotation of OpenXM_contrib/gmp/mpn/x86/add_n.S, Revision 1.1.1.1

1.1       maekawa     1: /* i80386 __mpn_add_n -- Add two limb vectors of the same length > 0 and store
                      2: sum in a third limb vector.
                      3:
                      4: Copyright (C) 1992, 1994, 1995 Free Software Foundation, Inc.
                      5:
                      6: This file is part of the GNU MP Library.
                      7:
                      8: The GNU MP Library is free software; you can redistribute it and/or modify
                      9: it under the terms of the GNU Library General Public License as published by
                     10: the Free Software Foundation; either version 2 of the License, or (at your
                     11: option) any later version.
                     12:
                     13: The GNU MP Library is distributed in the hope that it will be useful, but
                     14: WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
                     15: or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU Library General Public
                     16: License for more details.
                     17:
                     18: You should have received a copy of the GNU Library General Public License
                     19: along with the GNU MP Library; see the file COPYING.LIB.  If not, write to
                     20: the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
                     21: MA 02111-1307, USA. */
                     22:
                     23: /*
                     24:   INPUT PARAMETERS
                     25:   res_ptr      (sp + 4)
                     26:   s1_ptr       (sp + 8)
                     27:   s2_ptr       (sp + 12)
                     28:   size         (sp + 16)
                     29: */
                     30:
                     31: #include "sysdep.h"
                     32: #include "asm-syntax.h"
                     33:
                     34: .text
                     35:        ALIGN (3)
                     36:        .globl C_SYMBOL_NAME(__mpn_add_n)
                     37: C_SYMBOL_NAME(__mpn_add_n:)
                     38:        pushl %edi
                     39:        pushl %esi
                     40:
                     41:        movl 12(%esp),%edi              /* res_ptr */
                     42:        movl 16(%esp),%esi              /* s1_ptr */
                     43:        movl 20(%esp),%edx              /* s2_ptr */
                     44:        movl 24(%esp),%ecx              /* size */
                     45:
                     46:        movl    %ecx,%eax
                     47:        shrl    $3,%ecx                 /* compute count for unrolled loop */
                     48:        negl    %eax
                     49:        andl    $7,%eax                 /* get index where to start loop */
                     50:        jz      Loop                    /* necessary special case for 0 */
                     51:        incl    %ecx                    /* adjust loop count */
                     52:        shll    $2,%eax                 /* adjustment for pointers... */
                     53:        subl    %eax,%edi               /* ... since they are offset ... */
                     54:        subl    %eax,%esi               /* ... by a constant when we ... */
                     55:        subl    %eax,%edx               /* ... enter the loop */
                     56:        shrl    $2,%eax                 /* restore previous value */
                     57: #ifdef PIC
                     58: /* Calculate start address in loop for PIC.  Due to limitations in some
                     59:    assemblers, Loop-L0-3 cannot be put into the leal */
                     60:        call    L0
                     61: L0:    leal    (%eax,%eax,8),%eax
                     62:        addl    (%esp),%eax
                     63:        addl    $(Loop-L0-3),%eax
                     64:        addl    $4,%esp
                     65: #else
                     66: /* Calculate start address in loop for non-PIC.  */
                     67:        leal    (Loop - 3)(%eax,%eax,8),%eax
                     68: #endif
                     69:        jmp     *%eax                   /* jump into loop */
                     70:        ALIGN (3)
                     71: Loop:  movl    (%esi),%eax
                     72:        adcl    (%edx),%eax
                     73:        movl    %eax,(%edi)
                     74:        movl    4(%esi),%eax
                     75:        adcl    4(%edx),%eax
                     76:        movl    %eax,4(%edi)
                     77:        movl    8(%esi),%eax
                     78:        adcl    8(%edx),%eax
                     79:        movl    %eax,8(%edi)
                     80:        movl    12(%esi),%eax
                     81:        adcl    12(%edx),%eax
                     82:        movl    %eax,12(%edi)
                     83:        movl    16(%esi),%eax
                     84:        adcl    16(%edx),%eax
                     85:        movl    %eax,16(%edi)
                     86:        movl    20(%esi),%eax
                     87:        adcl    20(%edx),%eax
                     88:        movl    %eax,20(%edi)
                     89:        movl    24(%esi),%eax
                     90:        adcl    24(%edx),%eax
                     91:        movl    %eax,24(%edi)
                     92:        movl    28(%esi),%eax
                     93:        adcl    28(%edx),%eax
                     94:        movl    %eax,28(%edi)
                     95:        leal    32(%edi),%edi
                     96:        leal    32(%esi),%esi
                     97:        leal    32(%edx),%edx
                     98:        decl    %ecx
                     99:        jnz     Loop
                    100:
                    101:        sbbl    %eax,%eax
                    102:        negl    %eax
                    103:
                    104:        popl %esi
                    105:        popl %edi
                    106:        ret