sysdeps/x86_64/lshift.S

   1 /* x86-64 __mpn_lshift --
   2    Copyright (C) 2007, 2009 Free Software Foundation, Inc.
   3    This file is part of the GNU MP Library.
   4
   5    The GNU MP Library is free software; you can redistribute it and/or modify
   6    it under the terms of the GNU Lesser General Public License as published by
   7    the Free Software Foundation; either version 2.1 of the License, or (at your
   8    option) any later version.
   9
  10    The GNU MP Library is distributed in the hope that it will be useful, but
  11    WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
  12    or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public
  13    License for more details.
  14
  15    You should have received a copy of the GNU Lesser General Public License
  16    along with the GNU MP Library; see the file COPYING.LIB.  If not, write to
  17    the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
  18    MA 02111-1307, USA. */
  19
  20 #include "sysdep.h"
  21 #include "asm-syntax.h"
  22
  23 #define rp      %rdi
  24 #define up      %rsi
  25 #define n       %rdx
  26 #define cnt     %cl
  27
  28         .text
  29 ENTRY (__mpn_lshift)
  30         lea     -8(rp,n,8), rp
  31         lea     -8(up,n,8), up
  32
  33         mov     %edx, %eax
  34         and     $3, %eax
  35         jne     L(nb00)
  36 L(b00): /* n = 4, 8, 12, ... */
  37         mov     (up), %r10
  38         mov     -8(up), %r11
  39         xor     %eax, %eax
  40         shld    %cl, %r10, %rax
  41         mov     -16(up), %r8
  42         lea     24(rp), rp
  43         sub     $4, n
  44         jmp     L(00)
  45
  46 L(nb00):/* n = 1, 5, 9, ... */
  47         cmp     $2, %eax
  48         jae     L(nb01)
  49 L(b01): mov     (up), %r9
  50         xor     %eax, %eax
  51         shld    %cl, %r9, %rax
  52         sub     $2, n
  53         jb      L(le1)
  54         mov     -8(up), %r10
  55         mov     -16(up), %r11
  56         lea     -8(up), up
  57         lea     16(rp), rp
  58         jmp     L(01)
  59 L(le1): shl     %cl, %r9
  60         mov     %r9, (rp)
  61         ret
  62
  63 L(nb01):/* n = 2, 6, 10, ... */
  64         jne     L(b11)
  65 L(b10): mov     (up), %r8
  66         mov     -8(up), %r9
  67         xor     %eax, %eax
  68         shld    %cl, %r8, %rax
  69         sub     $3, n
  70         jb      L(le2)
  71         mov     -16(up), %r10
  72         lea     -16(up), up
  73         lea     8(rp), rp
  74         jmp     L(10)
  75 L(le2): shld    %cl, %r9, %r8
  76         mov     %r8, (rp)
  77         shl     %cl, %r9
  78         mov     %r9, -8(rp)
  79         ret
  80
  81         .p2align 4              /* performance critical! */
  82 L(b11): /* n = 3, 7, 11, ... */
  83         mov     (up), %r11
  84         mov     -8(up), %r8
  85         xor     %eax, %eax
  86         shld    %cl, %r11, %rax
  87         mov     -16(up), %r9
  88         lea     -24(up), up
  89         sub     $4, n
  90         jb      L(end)
  91
  92         .p2align 4
  93 L(top): shld    %cl, %r8, %r11
  94         mov     (up), %r10
  95         mov     %r11, (rp)
  96 L(10):  shld    %cl, %r9, %r8
  97         mov     -8(up), %r11
  98         mov     %r8, -8(rp)
  99 L(01):  shld    %cl, %r10, %r9
 100         mov     -16(up), %r8
 101         mov     %r9, -16(rp)
 102 L(00):  shld    %cl, %r11, %r10
 103         mov     -24(up), %r9
 104         mov     %r10, -24(rp)
 105         add     $-32, up
 106         lea     -32(rp), rp
 107         sub     $4, n
 108         jnc     L(top)
 109
 110 L(end): shld    %cl, %r8, %r11
 111         mov     %r11, (rp)
 112         shld    %cl, %r9, %r8
 113         mov     %r8, -8(rp)
 114         shl     %cl, %r9
 115         mov     %r9, -16(rp)
 116         ret
 117 END (__mpn_lshift)