diff options
Diffstat (limited to 'mac-x86/crypto/bn/x86-mont.S')
| -rw-r--r-- | mac-x86/crypto/bn/x86-mont.S | 176 |
1 files changed, 150 insertions, 26 deletions
diff --git a/mac-x86/crypto/bn/x86-mont.S b/mac-x86/crypto/bn/x86-mont.S index 1b79c5f..234034b 100644 --- a/mac-x86/crypto/bn/x86-mont.S +++ b/mac-x86/crypto/bn/x86-mont.S @@ -43,6 +43,126 @@ L_bn_mul_mont_begin: movl %esi,20(%esp) leal -3(%edi),%ebx movl %ebp,24(%esp) + call L001PIC_me_up +L001PIC_me_up: + popl %eax + movl L_OPENSSL_ia32cap_P$non_lazy_ptr-L001PIC_me_up(%eax),%eax + btl $26,(%eax) + jnc L002non_sse2 + movl $-1,%eax + movd %eax,%mm7 + movl 8(%esp),%esi + movl 12(%esp),%edi + movl 16(%esp),%ebp + xorl %edx,%edx + xorl %ecx,%ecx + movd (%edi),%mm4 + movd (%esi),%mm5 + movd (%ebp),%mm3 + pmuludq %mm4,%mm5 + movq %mm5,%mm2 + movq %mm5,%mm0 + pand %mm7,%mm0 + pmuludq 20(%esp),%mm5 + pmuludq %mm5,%mm3 + paddq %mm0,%mm3 + movd 4(%ebp),%mm1 + movd 4(%esi),%mm0 + psrlq $32,%mm2 + psrlq $32,%mm3 + incl %ecx +.align 4,0x90 +L0031st: + pmuludq %mm4,%mm0 + pmuludq %mm5,%mm1 + paddq %mm0,%mm2 + paddq %mm1,%mm3 + movq %mm2,%mm0 + pand %mm7,%mm0 + movd 4(%ebp,%ecx,4),%mm1 + paddq %mm0,%mm3 + movd 4(%esi,%ecx,4),%mm0 + psrlq $32,%mm2 + movd %mm3,28(%esp,%ecx,4) + psrlq $32,%mm3 + leal 1(%ecx),%ecx + cmpl %ebx,%ecx + jl L0031st + pmuludq %mm4,%mm0 + pmuludq %mm5,%mm1 + paddq %mm0,%mm2 + paddq %mm1,%mm3 + movq %mm2,%mm0 + pand %mm7,%mm0 + paddq %mm0,%mm3 + movd %mm3,28(%esp,%ecx,4) + psrlq $32,%mm2 + psrlq $32,%mm3 + paddq %mm2,%mm3 + movq %mm3,32(%esp,%ebx,4) + incl %edx +L004outer: + xorl %ecx,%ecx + movd (%edi,%edx,4),%mm4 + movd (%esi),%mm5 + movd 32(%esp),%mm6 + movd (%ebp),%mm3 + pmuludq %mm4,%mm5 + paddq %mm6,%mm5 + movq %mm5,%mm0 + movq %mm5,%mm2 + pand %mm7,%mm0 + pmuludq 20(%esp),%mm5 + pmuludq %mm5,%mm3 + paddq %mm0,%mm3 + movd 36(%esp),%mm6 + movd 4(%ebp),%mm1 + movd 4(%esi),%mm0 + psrlq $32,%mm2 + psrlq $32,%mm3 + paddq %mm6,%mm2 + incl %ecx + decl %ebx +L005inner: + pmuludq %mm4,%mm0 + pmuludq %mm5,%mm1 + paddq %mm0,%mm2 + paddq %mm1,%mm3 + movq %mm2,%mm0 + movd 36(%esp,%ecx,4),%mm6 + pand %mm7,%mm0 + movd 4(%ebp,%ecx,4),%mm1 + paddq %mm0,%mm3 + movd 4(%esi,%ecx,4),%mm0 + psrlq $32,%mm2 + movd %mm3,28(%esp,%ecx,4) + psrlq $32,%mm3 + paddq %mm6,%mm2 + decl %ebx + leal 1(%ecx),%ecx + jnz L005inner + movl %ecx,%ebx + pmuludq %mm4,%mm0 + pmuludq %mm5,%mm1 + paddq %mm0,%mm2 + paddq %mm1,%mm3 + movq %mm2,%mm0 + pand %mm7,%mm0 + paddq %mm0,%mm3 + movd %mm3,28(%esp,%ecx,4) + psrlq $32,%mm2 + psrlq $32,%mm3 + movd 36(%esp,%ebx,4),%mm6 + paddq %mm2,%mm3 + paddq %mm6,%mm3 + movq %mm3,32(%esp,%ebx,4) + leal 1(%edx),%edx + cmpl %ebx,%edx + jle L004outer + emms + jmp L006common_tail +.align 4,0x90 +L002non_sse2: movl 8(%esp),%esi leal 1(%ebx),%ebp movl 12(%esp),%edi @@ -53,12 +173,12 @@ L_bn_mul_mont_begin: leal 4(%edi,%ebx,4),%eax orl %edx,%ebp movl (%edi),%edi - jz L001bn_sqr_mont + jz L007bn_sqr_mont movl %eax,28(%esp) movl (%esi),%eax xorl %edx,%edx .align 4,0x90 -L002mull: +L008mull: movl %edx,%ebp mull %edi addl %eax,%ebp @@ -67,7 +187,7 @@ L002mull: movl (%esi,%ecx,4),%eax cmpl %ebx,%ecx movl %ebp,28(%esp,%ecx,4) - jl L002mull + jl L008mull movl %edx,%ebp mull %edi movl 20(%esp),%edi @@ -85,9 +205,9 @@ L002mull: movl 4(%esi),%eax adcl $0,%edx incl %ecx - jmp L0032ndmadd + jmp L0092ndmadd .align 4,0x90 -L0041stmadd: +L0101stmadd: movl %edx,%ebp mull %edi addl 32(%esp,%ecx,4),%ebp @@ -98,7 +218,7 @@ L0041stmadd: adcl $0,%edx cmpl %ebx,%ecx movl %ebp,28(%esp,%ecx,4) - jl L0041stmadd + jl L0101stmadd movl %edx,%ebp mull %edi addl 32(%esp,%ebx,4),%eax @@ -121,7 +241,7 @@ L0041stmadd: adcl $0,%edx movl $1,%ecx .align 4,0x90 -L0032ndmadd: +L0092ndmadd: movl %edx,%ebp mull %edi addl 32(%esp,%ecx,4),%ebp @@ -132,7 +252,7 @@ L0032ndmadd: adcl $0,%edx cmpl %ebx,%ecx movl %ebp,24(%esp,%ecx,4) - jl L0032ndmadd + jl L0092ndmadd movl %edx,%ebp mull %edi addl 32(%esp,%ebx,4),%ebp @@ -148,16 +268,16 @@ L0032ndmadd: movl %edx,32(%esp,%ebx,4) cmpl 28(%esp),%ecx movl %eax,36(%esp,%ebx,4) - je L005common_tail + je L006common_tail movl (%ecx),%edi movl 8(%esp),%esi movl %ecx,12(%esp) xorl %ecx,%ecx xorl %edx,%edx movl (%esi),%eax - jmp L0041stmadd + jmp L0101stmadd .align 4,0x90 -L001bn_sqr_mont: +L007bn_sqr_mont: movl %ebx,(%esp) movl %ecx,12(%esp) movl %edi,%eax @@ -168,7 +288,7 @@ L001bn_sqr_mont: andl $1,%ebx incl %ecx .align 4,0x90 -L006sqr: +L011sqr: movl (%esi,%ecx,4),%eax movl %edx,%ebp mull %edi @@ -180,7 +300,7 @@ L006sqr: cmpl (%esp),%ecx movl %eax,%ebx movl %ebp,28(%esp,%ecx,4) - jl L006sqr + jl L011sqr movl (%esi,%ecx,4),%eax movl %edx,%ebp mull %edi @@ -204,7 +324,7 @@ L006sqr: movl 4(%esi),%eax movl $1,%ecx .align 4,0x90 -L0073rdmadd: +L0123rdmadd: movl %edx,%ebp mull %edi addl 32(%esp,%ecx,4),%ebp @@ -223,7 +343,7 @@ L0073rdmadd: adcl $0,%edx cmpl %ebx,%ecx movl %ebp,24(%esp,%ecx,4) - jl L0073rdmadd + jl L0123rdmadd movl %edx,%ebp mull %edi addl 32(%esp,%ebx,4),%ebp @@ -239,7 +359,7 @@ L0073rdmadd: movl %edx,32(%esp,%ebx,4) cmpl %ebx,%ecx movl %eax,36(%esp,%ebx,4) - je L005common_tail + je L006common_tail movl 4(%esi,%ecx,4),%edi leal 1(%ecx),%ecx movl %edi,%eax @@ -251,12 +371,12 @@ L0073rdmadd: xorl %ebp,%ebp cmpl %ebx,%ecx leal 1(%ecx),%ecx - je L008sqrlast + je L013sqrlast movl %edx,%ebx shrl $1,%edx andl $1,%ebx .align 4,0x90 -L009sqradd: +L014sqradd: movl (%esi,%ecx,4),%eax movl %edx,%ebp mull %edi @@ -272,13 +392,13 @@ L009sqradd: cmpl (%esp),%ecx movl %ebp,28(%esp,%ecx,4) movl %eax,%ebx - jle L009sqradd + jle L014sqradd movl %edx,%ebp addl %edx,%edx shrl $31,%ebp addl %ebx,%edx adcl $0,%ebp -L008sqrlast: +L013sqrlast: movl 20(%esp),%edi movl 16(%esp),%esi imull 32(%esp),%edi @@ -293,9 +413,9 @@ L008sqrlast: adcl $0,%edx movl $1,%ecx movl 4(%esi),%eax - jmp L0073rdmadd + jmp L0123rdmadd .align 4,0x90 -L005common_tail: +L006common_tail: movl 16(%esp),%ebp movl 4(%esp),%edi leal 32(%esp),%esi @@ -303,16 +423,16 @@ L005common_tail: movl %ebx,%ecx xorl %edx,%edx .align 4,0x90 -L010sub: +L015sub: sbbl (%ebp,%edx,4),%eax movl %eax,(%edi,%edx,4) decl %ecx movl 4(%esi,%edx,4),%eax leal 1(%edx),%edx - jge L010sub + jge L015sub sbbl $0,%eax .align 4,0x90 -L011copy: +L016copy: movl (%esi,%ebx,4),%edx movl (%edi,%ebx,4),%ebp xorl %ebp,%edx @@ -321,7 +441,7 @@ L011copy: movl %ecx,(%esi,%ebx,4) movl %edx,(%edi,%ebx,4) decl %ebx - jge L011copy + jge L016copy movl 24(%esp),%esp movl $1,%eax L000just_leave: @@ -335,4 +455,8 @@ L000just_leave: .byte 54,44,32,67,82,89,80,84,79,71,65,77,83,32,98,121 .byte 32,60,97,112,112,114,111,64,111,112,101,110,115,115,108,46 .byte 111,114,103,62,0 +.section __IMPORT,__pointers,non_lazy_symbol_pointers +L_OPENSSL_ia32cap_P$non_lazy_ptr: +.indirect_symbol _OPENSSL_ia32cap_P +.long 0 #endif |