diff options
Diffstat (limited to 'arch/x86/math-emu/shr_Xsig.S')
| -rw-r--r-- | arch/x86/math-emu/shr_Xsig.S | 87 |
1 files changed, 87 insertions, 0 deletions
diff --git a/arch/x86/math-emu/shr_Xsig.S b/arch/x86/math-emu/shr_Xsig.S new file mode 100644 index 000000000..31cdd118e --- /dev/null +++ b/arch/x86/math-emu/shr_Xsig.S @@ -0,0 +1,87 @@ + .file "shr_Xsig.S" +/*---------------------------------------------------------------------------+ + | shr_Xsig.S | + | | + | 12 byte right shift function | + | | + | Copyright (C) 1992,1994,1995 | + | W. Metzenthen, 22 Parker St, Ormond, Vic 3163, | + | Australia. E-mail billm@jacobi.maths.monash.edu.au | + | | + | Call from C as: | + | void shr_Xsig(Xsig *arg, unsigned nr) | + | | + | Extended shift right function. | + | Fastest for small shifts. | + | Shifts the 12 byte quantity pointed to by the first arg (arg) | + | right by the number of bits specified by the second arg (nr). | + | | + +---------------------------------------------------------------------------*/ + +#include "fpu_emu.h" + +.text +ENTRY(shr_Xsig) + push %ebp + movl %esp,%ebp + pushl %esi + movl PARAM2,%ecx + movl PARAM1,%esi + cmpl $32,%ecx /* shrd only works for 0..31 bits */ + jnc L_more_than_31 + +/* less than 32 bits */ + pushl %ebx + movl (%esi),%eax /* lsl */ + movl 4(%esi),%ebx /* midl */ + movl 8(%esi),%edx /* msl */ + shrd %cl,%ebx,%eax + shrd %cl,%edx,%ebx + shr %cl,%edx + movl %eax,(%esi) + movl %ebx,4(%esi) + movl %edx,8(%esi) + popl %ebx + popl %esi + leave + ret + +L_more_than_31: + cmpl $64,%ecx + jnc L_more_than_63 + + subb $32,%cl + movl 4(%esi),%eax /* midl */ + movl 8(%esi),%edx /* msl */ + shrd %cl,%edx,%eax + shr %cl,%edx + movl %eax,(%esi) + movl %edx,4(%esi) + movl $0,8(%esi) + popl %esi + leave + ret + +L_more_than_63: + cmpl $96,%ecx + jnc L_more_than_95 + + subb $64,%cl + movl 8(%esi),%eax /* msl */ + shr %cl,%eax + xorl %edx,%edx + movl %eax,(%esi) + movl %edx,4(%esi) + movl %edx,8(%esi) + popl %esi + leave + ret + +L_more_than_95: + xorl %eax,%eax + movl %eax,(%esi) + movl %eax,4(%esi) + movl %eax,8(%esi) + popl %esi + leave + ret |