+ mov -48(%rsi),%r15
+ mov -40(%rsi),%r14
+ mov -32(%rsi),%r13
+ mov -24(%rsi),%r12
+ mov -16(%rsi),%rbp
+ mov -8(%rsi),%rbx
+ lea (%rsi),%rsp
+.Lsqr8x_epilogue:
+ ret
+.size bn_sqr8x_mont,.-bn_sqr8x_mont
+___
+}}}
+\f
+if ($addx) {{{
+my $bp="%rdx"; # original value
+
+$code.=<<___;
+.type bn_mulx4x_mont,\@function,6
+.align 32
+bn_mulx4x_mont:
+ mov %rsp,%rax
+.Lmulx4x_enter:
+ push %rbx
+ push %rbp
+ push %r12
+ push %r13
+ push %r14
+ push %r15
+.Lmulx4x_prologue:
+
+ shl \$3,${num}d # convert $num to bytes
+ xor %r10,%r10
+ sub $num,%r10 # -$num
+ mov ($n0),$n0 # *n0
+ lea -72(%rsp,%r10),%rbp # future alloca(frame+$num+8)
+ and \$-128,%rbp
+ mov %rsp,%r11
+ sub %rbp,%r11
+ and \$-4096,%r11
+ lea (%rbp,%r11),%rsp
+ mov (%rsp),%r10
+ cmp %rbp,%rsp
+ ja .Lmulx4x_page_walk
+ jmp .Lmulx4x_page_walk_done
+
+.align 16
+.Lmulx4x_page_walk:
+ lea -4096(%rsp),%rsp
+ mov (%rsp),%r10
+ cmp %rbp,%rsp
+ ja .Lmulx4x_page_walk
+.Lmulx4x_page_walk_done:
+
+ lea ($bp,$num),%r10
+ ##############################################################
+ # Stack layout
+ # +0 num
+ # +8 off-loaded &b[i]
+ # +16 end of b[num]
+ # +24 saved n0
+ # +32 saved rp
+ # +40 saved %rsp
+ # +48 inner counter
+ # +56
+ # +64 tmp[num+1]
+ #
+ mov $num,0(%rsp) # save $num
+ shr \$5,$num
+ mov %r10,16(%rsp) # end of b[num]
+ sub \$1,$num
+ mov $n0, 24(%rsp) # save *n0
+ mov $rp, 32(%rsp) # save $rp
+ mov %rax,40(%rsp) # save original %rsp
+ mov $num,48(%rsp) # inner counter
+ jmp .Lmulx4x_body
+
+.align 32
+.Lmulx4x_body:
+___
+my ($aptr, $bptr, $nptr, $tptr, $mi, $bi, $zero, $num)=
+ ("%rsi","%rdi","%rcx","%rbx","%r8","%r9","%rbp","%rax");
+my $rptr=$bptr;
+$code.=<<___;
+ lea 8($bp),$bptr
+ mov ($bp),%rdx # b[0], $bp==%rdx actually
+ lea 64+32(%rsp),$tptr
+ mov %rdx,$bi
+
+ mulx 0*8($aptr),$mi,%rax # a[0]*b[0]
+ mulx 1*8($aptr),%r11,%r14 # a[1]*b[0]
+ add %rax,%r11
+ mov $bptr,8(%rsp) # off-load &b[i]
+ mulx 2*8($aptr),%r12,%r13 # ...
+ adc %r14,%r12
+ adc \$0,%r13
+
+ mov $mi,$bptr # borrow $bptr
+ imulq 24(%rsp),$mi # "t[0]"*n0
+ xor $zero,$zero # cf=0, of=0
+
+ mulx 3*8($aptr),%rax,%r14
+ mov $mi,%rdx
+ lea 4*8($aptr),$aptr
+ adcx %rax,%r13
+ adcx $zero,%r14 # cf=0
+
+ mulx 0*8($nptr),%rax,%r10
+ adcx %rax,$bptr # discarded
+ adox %r11,%r10
+ mulx 1*8($nptr),%rax,%r11
+ adcx %rax,%r10
+ adox %r12,%r11
+ .byte 0xc4,0x62,0xfb,0xf6,0xa1,0x10,0x00,0x00,0x00 # mulx 2*8($nptr),%rax,%r12
+ mov 48(%rsp),$bptr # counter value
+ mov %r10,-4*8($tptr)
+ adcx %rax,%r11
+ adox %r13,%r12
+ mulx 3*8($nptr),%rax,%r15
+ mov $bi,%rdx
+ mov %r11,-3*8($tptr)
+ adcx %rax,%r12
+ adox $zero,%r15 # of=0
+ lea 4*8($nptr),$nptr
+ mov %r12,-2*8($tptr)
+
+ jmp .Lmulx4x_1st
+
+.align 32
+.Lmulx4x_1st:
+ adcx $zero,%r15 # cf=0, modulo-scheduled
+ mulx 0*8($aptr),%r10,%rax # a[4]*b[0]
+ adcx %r14,%r10
+ mulx 1*8($aptr),%r11,%r14 # a[5]*b[0]
+ adcx %rax,%r11
+ mulx 2*8($aptr),%r12,%rax # ...
+ adcx %r14,%r12
+ mulx 3*8($aptr),%r13,%r14
+ .byte 0x67,0x67
+ mov $mi,%rdx
+ adcx %rax,%r13
+ adcx $zero,%r14 # cf=0
+ lea 4*8($aptr),$aptr
+ lea 4*8($tptr),$tptr
+
+ adox %r15,%r10
+ mulx 0*8($nptr),%rax,%r15
+ adcx %rax,%r10
+ adox %r15,%r11
+ mulx 1*8($nptr),%rax,%r15
+ adcx %rax,%r11
+ adox %r15,%r12
+ mulx 2*8($nptr),%rax,%r15
+ mov %r10,-5*8($tptr)
+ adcx %rax,%r12
+ mov %r11,-4*8($tptr)
+ adox %r15,%r13
+ mulx 3*8($nptr),%rax,%r15
+ mov $bi,%rdx
+ mov %r12,-3*8($tptr)
+ adcx %rax,%r13
+ adox $zero,%r15
+ lea 4*8($nptr),$nptr
+ mov %r13,-2*8($tptr)
+
+ dec $bptr # of=0, pass cf
+ jnz .Lmulx4x_1st
+
+ mov 0(%rsp),$num # load num
+ mov 8(%rsp),$bptr # re-load &b[i]
+ adc $zero,%r15 # modulo-scheduled
+ add %r15,%r14
+ sbb %r15,%r15 # top-most carry
+ mov %r14,-1*8($tptr)
+ jmp .Lmulx4x_outer
+
+.align 32
+.Lmulx4x_outer:
+ mov ($bptr),%rdx # b[i]
+ lea 8($bptr),$bptr # b++
+ sub $num,$aptr # rewind $aptr
+ mov %r15,($tptr) # save top-most carry
+ lea 64+4*8(%rsp),$tptr
+ sub $num,$nptr # rewind $nptr
+
+ mulx 0*8($aptr),$mi,%r11 # a[0]*b[i]
+ xor %ebp,%ebp # xor $zero,$zero # cf=0, of=0
+ mov %rdx,$bi
+ mulx 1*8($aptr),%r14,%r12 # a[1]*b[i]
+ adox -4*8($tptr),$mi
+ adcx %r14,%r11
+ mulx 2*8($aptr),%r15,%r13 # ...
+ adox -3*8($tptr),%r11
+ adcx %r15,%r12
+ adox -2*8($tptr),%r12
+ adcx $zero,%r13
+ adox $zero,%r13
+
+ mov $bptr,8(%rsp) # off-load &b[i]
+ mov $mi,%r15
+ imulq 24(%rsp),$mi # "t[0]"*n0
+ xor %ebp,%ebp # xor $zero,$zero # cf=0, of=0
+
+ mulx 3*8($aptr),%rax,%r14
+ mov $mi,%rdx
+ adcx %rax,%r13
+ adox -1*8($tptr),%r13
+ adcx $zero,%r14
+ lea 4*8($aptr),$aptr
+ adox $zero,%r14
+
+ mulx 0*8($nptr),%rax,%r10
+ adcx %rax,%r15 # discarded
+ adox %r11,%r10
+ mulx 1*8($nptr),%rax,%r11
+ adcx %rax,%r10
+ adox %r12,%r11
+ mulx 2*8($nptr),%rax,%r12
+ mov %r10,-4*8($tptr)
+ adcx %rax,%r11
+ adox %r13,%r12
+ mulx 3*8($nptr),%rax,%r15
+ mov $bi,%rdx
+ mov %r11,-3*8($tptr)
+ lea 4*8($nptr),$nptr
+ adcx %rax,%r12
+ adox $zero,%r15 # of=0
+ mov 48(%rsp),$bptr # counter value
+ mov %r12,-2*8($tptr)
+
+ jmp .Lmulx4x_inner
+
+.align 32
+.Lmulx4x_inner:
+ mulx 0*8($aptr),%r10,%rax # a[4]*b[i]
+ adcx $zero,%r15 # cf=0, modulo-scheduled
+ adox %r14,%r10
+ mulx 1*8($aptr),%r11,%r14 # a[5]*b[i]
+ adcx 0*8($tptr),%r10
+ adox %rax,%r11
+ mulx 2*8($aptr),%r12,%rax # ...
+ adcx 1*8($tptr),%r11
+ adox %r14,%r12
+ mulx 3*8($aptr),%r13,%r14
+ mov $mi,%rdx
+ adcx 2*8($tptr),%r12
+ adox %rax,%r13
+ adcx 3*8($tptr),%r13
+ adox $zero,%r14 # of=0
+ lea 4*8($aptr),$aptr
+ lea 4*8($tptr),$tptr
+ adcx $zero,%r14 # cf=0
+
+ adox %r15,%r10
+ mulx 0*8($nptr),%rax,%r15
+ adcx %rax,%r10
+ adox %r15,%r11
+ mulx 1*8($nptr),%rax,%r15
+ adcx %rax,%r11
+ adox %r15,%r12
+ mulx 2*8($nptr),%rax,%r15
+ mov %r10,-5*8($tptr)
+ adcx %rax,%r12
+ adox %r15,%r13
+ mulx 3*8($nptr),%rax,%r15
+ mov $bi,%rdx
+ mov %r11,-4*8($tptr)
+ mov %r12,-3*8($tptr)
+ adcx %rax,%r13
+ adox $zero,%r15
+ lea 4*8($nptr),$nptr
+ mov %r13,-2*8($tptr)
+
+ dec $bptr # of=0, pass cf
+ jnz .Lmulx4x_inner
+
+ mov 0(%rsp),$num # load num
+ mov 8(%rsp),$bptr # re-load &b[i]
+ adc $zero,%r15 # modulo-scheduled
+ sub 0*8($tptr),$zero # pull top-most carry
+ adc %r15,%r14
+ sbb %r15,%r15 # top-most carry
+ mov %r14,-1*8($tptr)
+
+ cmp 16(%rsp),$bptr
+ jne .Lmulx4x_outer
+
+ lea 64(%rsp),$tptr
+ sub $num,$nptr # rewind $nptr
+ neg %r15
+ mov $num,%rdx
+ shr \$3+2,$num # %cf=0
+ mov 32(%rsp),$rptr # restore rp
+ jmp .Lmulx4x_sub
+
+.align 32
+.Lmulx4x_sub:
+ mov 8*0($tptr),%r11
+ mov 8*1($tptr),%r12
+ mov 8*2($tptr),%r13
+ mov 8*3($tptr),%r14
+ lea 8*4($tptr),$tptr
+ sbb 8*0($nptr),%r11
+ sbb 8*1($nptr),%r12
+ sbb 8*2($nptr),%r13
+ sbb 8*3($nptr),%r14
+ lea 8*4($nptr),$nptr
+ mov %r11,8*0($rptr)
+ mov %r12,8*1($rptr)
+ mov %r13,8*2($rptr)
+ mov %r14,8*3($rptr)
+ lea 8*4($rptr),$rptr
+ dec $num # preserves %cf
+ jnz .Lmulx4x_sub
+
+ sbb \$0,%r15 # top-most carry
+ lea 64(%rsp),$tptr
+ sub %rdx,$rptr # rewind
+
+ movq %r15,%xmm1
+ pxor %xmm0,%xmm0
+ pshufd \$0,%xmm1,%xmm1
+ mov 40(%rsp),%rsi # restore %rsp
+ jmp .Lmulx4x_cond_copy
+
+.align 32
+.Lmulx4x_cond_copy:
+ movdqa 16*0($tptr),%xmm2
+ movdqa 16*1($tptr),%xmm3
+ lea 16*2($tptr),$tptr
+ movdqu 16*0($rptr),%xmm4
+ movdqu 16*1($rptr),%xmm5
+ lea 16*2($rptr),$rptr
+ movdqa %xmm0,-16*2($tptr) # zero tp
+ movdqa %xmm0,-16*1($tptr)
+ pcmpeqd %xmm1,%xmm0
+ pand %xmm1,%xmm2
+ pand %xmm1,%xmm3
+ pand %xmm0,%xmm4
+ pand %xmm0,%xmm5
+ pxor %xmm0,%xmm0
+ por %xmm2,%xmm4
+ por %xmm3,%xmm5
+ movdqu %xmm4,-16*2($rptr)
+ movdqu %xmm5,-16*1($rptr)
+ sub \$32,%rdx
+ jnz .Lmulx4x_cond_copy
+
+ mov %rdx,($tptr)
+
+ mov \$1,%rax
+ mov -48(%rsi),%r15
+ mov -40(%rsi),%r14
+ mov -32(%rsi),%r13
+ mov -24(%rsi),%r12
+ mov -16(%rsi),%rbp
+ mov -8(%rsi),%rbx
+ lea (%rsi),%rsp
+.Lmulx4x_epilogue:
+ ret
+.size bn_mulx4x_mont,.-bn_mulx4x_mont
+___
+}}}
+$code.=<<___;
+.asciz "Montgomery Multiplication for x86_64, CRYPTOGAMS by <appro\@openssl.org>"
+.align 16
+___
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+# CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern __imp_RtlVirtualUnwind
+.type mul_handler,\@abi-omnipotent
+.align 16
+mul_handler:
+ push %rsi
+ push %rdi
+ push %rbx
+ push %rbp
+ push %r12
+ push %r13
+ push %r14
+ push %r15
+ pushfq
+ sub \$64,%rsp
+
+ mov 120($context),%rax # pull context->Rax
+ mov 248($context),%rbx # pull context->Rip
+
+ mov 8($disp),%rsi # disp->ImageBase
+ mov 56($disp),%r11 # disp->HandlerData
+
+ mov 0(%r11),%r10d # HandlerData[0]
+ lea (%rsi,%r10),%r10 # end of prologue label
+ cmp %r10,%rbx # context->Rip<end of prologue label
+ jb .Lcommon_seh_tail
+
+ mov 152($context),%rax # pull context->Rsp
+
+ mov 4(%r11),%r10d # HandlerData[1]
+ lea (%rsi,%r10),%r10 # epilogue label
+ cmp %r10,%rbx # context->Rip>=epilogue label
+ jae .Lcommon_seh_tail
+
+ mov 192($context),%r10 # pull $num
+ mov 8(%rax,%r10,8),%rax # pull saved stack pointer
+
+ jmp .Lcommon_pop_regs
+.size mul_handler,.-mul_handler
+
+.type sqr_handler,\@abi-omnipotent
+.align 16
+sqr_handler:
+ push %rsi
+ push %rdi
+ push %rbx
+ push %rbp
+ push %r12
+ push %r13
+ push %r14
+ push %r15
+ pushfq
+ sub \$64,%rsp
+
+ mov 120($context),%rax # pull context->Rax
+ mov 248($context),%rbx # pull context->Rip
+
+ mov 8($disp),%rsi # disp->ImageBase
+ mov 56($disp),%r11 # disp->HandlerData
+
+ mov 0(%r11),%r10d # HandlerData[0]
+ lea (%rsi,%r10),%r10 # end of prologue label
+ cmp %r10,%rbx # context->Rip<.Lsqr_body
+ jb .Lcommon_seh_tail
+
+ mov 4(%r11),%r10d # HandlerData[1]
+ lea (%rsi,%r10),%r10 # body label
+ cmp %r10,%rbx # context->Rip>=.Lsqr_epilogue
+ jb .Lcommon_pop_regs
+
+ mov 152($context),%rax # pull context->Rsp
+
+ mov 8(%r11),%r10d # HandlerData[2]
+ lea (%rsi,%r10),%r10 # epilogue label
+ cmp %r10,%rbx # context->Rip>=.Lsqr_epilogue
+ jae .Lcommon_seh_tail
+
+ mov 40(%rax),%rax # pull saved stack pointer
+
+.Lcommon_pop_regs:
+ mov -8(%rax),%rbx
+ mov -16(%rax),%rbp
+ mov -24(%rax),%r12
+ mov -32(%rax),%r13
+ mov -40(%rax),%r14
+ mov -48(%rax),%r15
+ mov %rbx,144($context) # restore context->Rbx
+ mov %rbp,160($context) # restore context->Rbp
+ mov %r12,216($context) # restore context->R12
+ mov %r13,224($context) # restore context->R13
+ mov %r14,232($context) # restore context->R14
+ mov %r15,240($context) # restore context->R15
+
+.Lcommon_seh_tail:
+ mov 8(%rax),%rdi
+ mov 16(%rax),%rsi
+ mov %rax,152($context) # restore context->Rsp
+ mov %rsi,168($context) # restore context->Rsi
+ mov %rdi,176($context) # restore context->Rdi
+
+ mov 40($disp),%rdi # disp->ContextRecord
+ mov $context,%rsi # context
+ mov \$154,%ecx # sizeof(CONTEXT)
+ .long 0xa548f3fc # cld; rep movsq
+
+ mov $disp,%rsi
+ xor %rcx,%rcx # arg1, UNW_FLAG_NHANDLER
+ mov 8(%rsi),%rdx # arg2, disp->ImageBase
+ mov 0(%rsi),%r8 # arg3, disp->ControlPc
+ mov 16(%rsi),%r9 # arg4, disp->FunctionEntry
+ mov 40(%rsi),%r10 # disp->ContextRecord
+ lea 56(%rsi),%r11 # &disp->HandlerData
+ lea 24(%rsi),%r12 # &disp->EstablisherFrame
+ mov %r10,32(%rsp) # arg5
+ mov %r11,40(%rsp) # arg6
+ mov %r12,48(%rsp) # arg7
+ mov %rcx,56(%rsp) # arg8, (NULL)
+ call *__imp_RtlVirtualUnwind(%rip)
+
+ mov \$1,%eax # ExceptionContinueSearch
+ add \$64,%rsp
+ popfq